Handbook of Chemical Engineering Vol II

HANDBOOK
OF
CHEMICAL ENGINEERING
TkkQrawBJillBook Qx lm
II1II11II1I1IIII111II1I1II1III1IU11I1IIIIIIIII11I111I11IIIIIIIIIIIII1II
PUBLISHERS OF BOOKS F O R_•
Electrical World ^ Engineering News-Record

Power v Engineering and Mining Journal-Press
Chemical and Metallurgical Engineering
Electric Railway Journal v Coal Age
American Machinistv Ingenieria Intemacional
Electrical Merchandising v BusTransportation
Journal of Electricity and Western Industry
Industrial Engineer
HANDBOOK
OF
Prepared by a Staff of Specialists
DONALD M. LIDDELL, Editor-in-Chief

CONSULTING ENGINEER AND ECONOMIST, MEMBER, A. I. M. E.f M. M. S. A
AUTHOR “METALLURGIST AND CHEMISTS* HANDBOOK*’
IN TWO VOLUMES
VOL. II
First Edition
Third Impression
McGRAW-HILL BOOK COMPANY, Inc.

NEW YORK: 370 SEVENTH AVENUE
LONDON: 6 & 8 BOUVERIE ST., E. C. 4
1922
Copyright, 1922, by the
McGraw-Hill Book Company, Inc.
PRINTED IN THE UNITED STATES OF AMERICA
THE MAPLE PRESS - YORK PA

TABLE OF CONTENTS
VOL. II
Section Page
XIV. High Temperature Production.519
XV. Mixing and Kneading.529
XVI. Sampling.565
XVII. Fermentation.585
XVIII. Distillation—Fractional—Destructive.607
XIX. Refrigeration.655
XX. Oxidation and Reduction.683
XXI. Electrochemistry.731
XXII. Catalysis.749
XXIII. Colloidal Chemistry.765
XXIV. Smelting—Zinc—Lead—Copper—Iron—Roasting.793
XXV. Lutes and Cements.841
XXVI. Radio-active Elements.857
XXVII. Rare Metals.883
XXVIII. Rare Gases . ..911
XXIX. Materials of Construction.927
XXX. Plant Design.961
XXXI. Methods of Financing.979
Appendix.985
Index.993
HANDBOOK
OF
SECTION XIV
HIGH TEMPERATURE PRODUCTION

By Colin G. Fink1
A Modem Art.—One of the most distinctive features of modern metallurgy

as compared with that of two or three generations ago is the ease with which
temperatures of 2,000 to 4,000°C. are attainable by means of comparatively
simple apparatus. Metals such as platinum, tantalum and tungsten, or com¬
pounds such as lime, magnesia or alumina, heretofore rated as infusible are
today readily melted. Pure magnesia melts at 2,240°, pure tungsten at 3,260°C.
Not only have these high temperatures been brought about by the introduction of
the electric arc, but great strides have been made in the better utilization of the heating
value of carbonaceous fuels. Furthermore, the introduction of the art of alumino-
thermics has made it comparatively simple to produce metals and alloys at 3,000°C.,
metals and alloys which heretofore could not be produced as readily and as cheaply
and as free from carbon by any other process.
Before proceeding to outline briefly the various heating methods employed in
bringing about high temperature melting or reactions, we shall refer in a few words to
the subject of refractories for high-melting-point containers.
High-temperature Refractories.—Our choice of materials is very limited
especially for temperature work in excess of 2;000°C. Of the various refractories
readily obtainable, lime, magnesia and alumina are the most serviceable basic ones
at temperatures up to about 1,900°. The purer these oxides the higher their
melting points. In laboratory tests thoria crucibles have given very satis¬
factory results at temperatures as high as 2,400°C. Although pure tungsten
melts at 3,260°C., crucibles made from this metal, unless constructed with very
heavy walls, show marked tendency to collapse at about 2,800°C. In those
cases where carbon does not readily combine with the substance to be heated,
commercially pure graphite crucibles or supports are serviceable up to tempera¬
tures of about 3,000°C. The life of these graphite crucibles at these elevated
temperatures is however comparatively short since the amorphous carbon binder
used in the manufacture of the crucible rapidly loses its binding qualities as the
graphitization temperature is approached.
At the maximum temperatures of 3,000 to 4,000°C. commercially obtainable there
is, aside from carbon, no metal or compound known that will serve as lining or crucible
material for protracted heating. When operating at these extreme temperatures
resource is had to various special methods which can best be described by citing a few
definite examples. In order to melt pure tungsten metal into a solid coherent form
the practice today consists in compressing metallic tungsten powder into rod or ingot
form and then after a preliminary sintering connecting the ends to an electric low-
voltage, high-current terminal and passing enough current through the rod or ingot to
bring its temperature to almost the melting point. This operation is carried out in an
atmosphere of hydrogen or ammonia in order to avoid oxidation of the metal. Another
1 Consulting electro metallurgist, New York, N. Y.
519
520 CHEMICAL ENGINEERING
method, which is commercially applied to melting tantalum into buttons or ingots

consists in playing an arc between a compressed tantalum rod vertically suspended
and a small base of fused tantalum and allowing the metal globules, as they descend
from the upper electrode, to collect and build up into a button or rod. The operation
is carried out in vacuum on account of the marked tendency of tantalum to combine
with all of the ordinary gases.
Of particular interest on account of the increasing wide application is the alumino-
thermic or Goldschmidt method of attaining temperatures of 3,000°C. and above.
The success of the process is largely dependent upon the speed with which the entire
melting operation is carried out. Only a few minutes or seconds suffice to complete
the reaction and during this short time the magnesite lining of the pot does not reach
its melting point which is 1,000° below that of the reaction temperature. In place of
aluminum metal, magnesium-aluminum alloy or rare earth metals may be used.
Carbon-free ferro-vanadium, chromium and manganese metals are made exclusively
by Goldschmidt’s method.
Oxyacetylene Flame.—The temperature of the oxyacetylene flame is approxi¬
mately 3,50Q°C. This flame, as is well known, has a wide application in the
welding and cutting of iron and steel. Its use for melting metals and compounds
is rather limited on account of the marked tendency for the material being heated
to take up carbon.
Oxyhydrogen Flame.—The oxyhydrogen flame is well suited for the melting of
platinum and other precious metals. It is also used extensively in fusing and
working quartz and for fusing oxide of aluminum in the manufacture of synthetic
rubies and amethysts. The temperature of the flame is about 2,000°C.
Oxywater-gas Flame.—For the working of high-fusing glass such as Pyrex,
the oxywater-gas flame is very serviceable. The temperature of the flame is
several hundred degrees lower than that of the oxyhydrogen flame and therefore
will not melt platinum nor quartz.
Surface Combustion.—Although the phenomenon of surface combustion had
been observed and studied for a good many years, it was not until the beginning
of this century that the underlying principles were adapted in commercial
burners and heating apparatus. The practical success of surface combustion is
largely due to W. A. Bone1 and there are a large number of welding furnaces,
glass-melting furnaces, oil-cracking, carbonizing and heat-treating furnaces in use
today constructed on the ^Bonecourt” system.
In surface combustion “a homogenous explosive mixture of gas and air, in the
proper ratio for complete combustion is caused to burn without flame in contact
with a granular incandescent solid, whereby a large proportion of the potential energy
of the gas is immediately converted into radiant form.” One of the most familiar
cases of surface combustion is the Welsbaeh incandescent gas mantle. The tem¬
perature of this mantle under favorable conditions is as high as 2,300°C. whereas in the
ordinary Bunsen gas flame it is difficult to attain a temperature of 1,800°C. We are
dealing here with contact catalysis. As early as 1834 Faraday advanced the theory
that the function of the surface was to condense both the oxygen and the combustible
gas, “thus producing in the surface layers a condition comparable to that of high,
pressure.” In Bone’s early experiments (1902)2 he showed that (1) the power of
accelerating gaseous combustion is possessed by all surfaces at temperatures below the
1 Proc. Roy. Inst. Great Britain, Vol. 21, pp. 41—62, 1914; Engineering, Vol. 91, p. 487, 1911; Journ.
Franklin Inst. Vol. 173, p. 101, 1912; Journ. Soc. Chem. Ind. Vol. 38, p. 228 T, 1919.
* Proc. Roy. Inst. Great Britain l.c.
HIGH TEMPERATURE PRODUCTION 521
ignition point; (2) such an accelerated surface combustion is dependent upon an

absorption of' the combustible gas and probably also of the oxygen, by the surface,
whereby it becomes activated by association with the surface; and (3) the surface
itself becomes electrically charged during the process. Bone contended “that if hot
surfaces possess the power of accelerating gaseous combustion at temperatures below
the ignition point, the same power must also be manifested in even a greater degree
at higher temperatures and especially so when the surface itself becomes incandescent.”
J. J. Thomson suggested that the emission of electrons or charged particles from the
incandescent surface was a factor of primary importance.
The contact material may be a porous firebrick, granular quartz, granular alumina,
etc. Platinum is easily melted in a surface combustion furnace using ordinary coal
gas. In Bone’s high temperature test furnaces
pure alumina can be fused. The advantages Exit for Spent Gases
claimed for the Bonecourt surface combustion
system are: (1) The combustion is greatly
accelerated by the incandescent surface and if
so desired may be concentrated just where the
heat is required; (2) the combustion is perfect
with a minimum excess of air; (3) the attain¬
ment of very high temperatures is possible
without the aid of elaborate regenerative
devices; and (4) owing to the large amount of
radiant energy developed, transmission of heat
from the seat of combustion to the object to be
heated is very rapid. For many important
industrial heating applications surface com¬
bustion is not only “pre-eminently economical
but also easy of control.” With coal gas or Fig. 1.—Surface-combustion unit.
water gas, without regeneration, temperatures
of 1,900 to 2,000°C. are attained; with heat recuperation, still higher temperatures
are attained.1
Fuel Oil Furnaces.,2—The fuel oil is sprayed into the furnace by means of an
atomizer or burner which pulverizes the oil and delivers it in a gaseous vapor
or in small globules at the hottest place in the furnace. Air is admitted from
below and as soon as the temperature is raised to the ignition point, chemical
reaction takes place with the atomized fuel oil and thus heat is generated. The
heat is absorbed by the gases of the furnace and consequently their temperature
is at once raised oftentimes to 1,200 to 1,500°C. These furnace gases consist of
the inert nitrogen that partly constituted the entering air, the carbon dioxide
or carbon monoxide formed by the burning of the carbon, water vapor formed by
the burning of the hydrogen, sulphur dioxide formed by the burning of the
sulphur content, which latter ingredient is always small and a considerable quan¬
tity of free oxygen depending on the amount of excess air admitted to the furnace.
Colloidal Fuel.3—Colloidal fuel is a combination of liquid hydrocarbon with
finely pulverized carbonaceous substance, the components so combined and so
1 Additional references: A. E. Blake, in Bacon & Hamor’s “American Fuels,” McGraw-Hill, 1921;
W. D. Bancroft, Jour. Phys%c. Chem.t Vol. 21, p. 644, 1917; Carleton Ellis, Trans. A. I. M. E., Vol. 43,
p. 612, 1912.
2 “Fuel Oil and Steam Engineering” by Robert and C. H. Delany, McGraw-Hill, 1921.
8 An abstract of a detailed report by Lindon W. Bates, Canadian Chem. Jour. Vol. 4, p. 40, 1920;
treated as to form a stable fuel capable of being atomized and burned in a

furnace. It is made in three forms, a liquid, a gel and a mobile paste. The
name “colloidal55 is not scientifically exact since only part of the carbon present
is in the colloidal form. On the other hand the title is descriptive since the
composite possesses many important colloid-like characteristics. The propor¬
tions are usually 60 per cent oil and 40 per cent coal. Up to ratios of 25 per
cent oil and 75 per cent coal the fuel is a mobile paste. All kinds of oils and
solid carbons may be used. The cheap coal wastes and breakages are all avail¬
able. The “colloidal fuel55 is used in the selfsame way as oil fuel and with the
same apparatus. In general colloidal fuel is on a parity with oil, while certain
grades are superior.
In marine tests on the U. S. S. "Gem,” equipped with Normand express boilers,
colloidal fuel evaporative efficiency reached 91.5 per cent on a short run, while that of
the Navy oil did not on any run exceed 79.4 per cent and that of pulverized coal
72.6 per cent. The calorific value per unit volume is greater than that of straight oil.
For example, in a composite made up of 35 per cent by weight of pulverized coal of
14,000 B.t.u. per pound and 1.6 specific gravity, and 65 per cent oil of 18,200 B.t.u.
per pound, 0.96 specific gravity, a gallon of the composite has 165,000 B.t.u. while oil
has 146,000 B.t.u. per gallon. Colloidal fuel is heavier than water while oil is lighter;
colloidal fuel may be stored under a water seal and its fire may be quenched with
water. The gel is easily changed to the liquid form by heating. (See also p. 789).
Electric Furnaces.—The advantages of electric furnaces over fuel-fired fur¬
naces are: Ease of control of temperature, long life of linings, absence of con¬
taminating sulphur and other impurities found in ordinary fuels, rigid control of
composition of charge, ease of attaining very high temperatures in a short time,
absence of nitrogen, low losses due to volatilization of zinc and similar metals,
etc. Metals and compounds can be manufactured in the electric furnace that
cannot be made in a fuel-fired furnace.
There are three distinct types of electric furnaces now in commercial operation:
The resistance furnaces, the arc furnaces and the induction furnaces.
The resistor-type electric furnaces in present use employ as resistor element
either carbon, a carbide or a metal or alloy. In the manufacture of carborundum
from silica and anthracite the charge itself acts as resistor; similarly, in the graphitiza-
tion of carbon electrodes the electrodes themselves form the resistor.
Of the large capacity electric resistor furnaces, the Baily is one of the most widely
used. This furnace is suitable for melting brass, aluminum, copper and bronzes
and also used for holding molten ferro-manganese and steel at temperature. The
heating element consists of a circular trough of carborundum packed with granulated
carbon. By the circular form of the heating element and the shape of the furnace
lining a uniform heat is projected by radiation and reflection onto the materials to be
melted. The standard sizes are, 50 kw. with 500 lb. hearth capacity, 75 kw. with
750 lb. hearth capacity and 105 kw. with 1,500 lb. capacity. Among the advantages
of the Baily furnace are the low zinc losses when melting brass, simplicity of operation,
ready renewal of the resistor, absence of any moving parts and low labor cost. On the
other hand, like all resistor-type furnaces, the maximum attainable temperature of
the Baily is about 2,100°C. as compared with 3,500°C. in the arc furnace and therefore
ferro alloys such as ferro-tungsten or ferro-chromium cannot be produced in the
Baily.1
1 For details see T. F. Baily, Trans. Am. Electrochem. Soc., Vol. 40, 1921; ibid., Vol. 38, p. 411,
1919; ibid., Vol. 32, pp. 155-164, 1917; cf. also Frank Thornton, ibid., Vol. 32, pp. 141-154, 1917.
HIGH TEMPERATURE PRODUCTION r>23
Carbon resistor furnaces for laboratory use have been downed by Bmrhers
Tucker, Fitzgerald, Clawson and others. The highest temperatures are obtamed
in the Arsem vacuum furnace. Its heater is a graphite helix m an upnghl posdnm
or in another type, the resistor consists of grids cut, from graphite slabs. 1 he.
vacuum chamber is made of cast bronze. All joints are made tight, by lead gaskets.
After charging the furnace, the cover is replaced and the chamber is exhausted.
The vacuum pumps are kept in operation during the experiment. The. furnace lias
been installed in a large number of university and research laboratories. It, has been
found very serviceable in the study of alloys, high melting point determinations, high
temperature fractional distillation, etc.1 The maximum temperature attained in the
Arsem furnace is 3,100°C.
Resistance Furnaces with Metal Resistors.—Of the metal resistors commonly
used in large furnaces the nickel-chrome alloys such as “iuchromo” and a dim¬
mer5 are the most serviceable on account of their low temperature eoellieient
and their resistance to oxidation at temperatures as high as 1,200°C. The alloy
resistor is in the shape of ribbon or wire and is supported on frames with insulated
knobs or hooks. Large furnaces of this typo are widely employd for the heat
treatment of steel parts, enameling, japanning, etc.
In the laboratory and for small scale operation nickel, platinum, molybdenum and
tungsten are very serviceable resistors. Platinum ribbon-wound quartz tubes will
operate very satisfactorily up to temperatures of about 1,400 to l,500°tl At higher
temperatures, especially in the presence of reducing gases, the ribbon gradually
disintegrates and small platinum crystals will be found in the. silica packing surround¬
ing the tube. For low temperatures, say 000 to S()()°0., nickel wire or ribbon resists
oxidation remarkably well due largely to the adherent film of oxide, formed upon first
heating. Molybdenum and tungsten wire-wound alundum tube furnaces will operate
satisfactorily up to temperatures as high as l,80()°O., but since these two metals will
“burn” when heated in air at these temperatures, the tubes art; eneased in un airtight,
iron box in which an atmosphere of hydrogen or ammonia is maintained.3 A very
serviceable furnace of this type consists of a 24-in. alundum tube, 1 in. in diameter,
wound with 17 ft. of 0.020-in. diameter tungsten wire. The current consumption is
from 10 to 17 amp.; a rheostat of about 200 ohms is connected in series across the
220-volt service line. For the determination of melting points of pure metals and
alloys, such as nickel, cobalt, iron, copper, etc., this tungsten-hydrogen furnace is to bo
recommended. To cite but a single illustration, the melting point of cobalt deter¬
mined in the Arsem vacuum furnace is 1,494°, whereas, in the tungsten-hydrogen
furnace the melting point is 1,610°, as repeatedly determined by the writer. This
difference is due to the absence of every trace of carbon in the metal in the latter ease.
For best results the insulator packing between the; outer wall of the tube and the
inner walls of the iron casing is pure granular aluminum oxide*; wheat expending the*
furnace at temperatures of 1,500°C. and below pure; calcined magnesia is ve*ry
satisfactory.
For laboratory-scale operations at temperature's above the; softening point of
alumina or alundum, the tube itself is made of tungsten. A furnace* of this type* is
described by Wartenberg.3 The tungsten tube is directly e;e)nne*cte*d to the; terminals
from the transformer. The tube is encase;d in an airtight box.4 The*se* tungstem
tube furnaces are recommended for carrying out tests at temperatures up to about
1 w* C. Arsem, Jour. Ind. Eng. Ckem., Vol. 2, pp. 3 ft, lftlO; “Thu Klcctric Furniu* A If ml
Stansfield, McGraw-Hill; and General Electric, Bull., No. 48ftH,
2 For details, see Winne and Dantsizen, Tram. Am. Eleelrochcm Hoc.,
Vol. 20, p. “H7.
8 Zeit. Elektrochem., Vol. IS, p. 876, 1909.
4 See also Appelberg, U. S. Pat. 953,774, Apr. 5, 1910.
2,500°C. in an atmosphere absolutely free from carbon monoxide or hydrocarbons.

Rubies and sapphires can readily be made in the tungsten-tube furnace.
The Electric Arc.—The temperature of the electric arc has been variously
estimated at from 2,000 to 6,000°C., but the usually accepted temperature of the
ordinary arc is 3,500 to 4,000°C. W. R. Mott1 has carried out an exhaustive
research on the behavior of various elements and compounds in the arc. His
apparatus is very simple and inexpensive. Only small quantities of materials
are necessary and definite results can be obtained in a very short time. The out¬
standing feature of the Beck carbon arc is the very high concentration of energy.
A current density of over one ampere .per square millimeter is obtained and the
total crater of the positive carbon reaches a very high temperature. The positive
carbon is covered with rare earths The temperature of the incandescent gas
within the positive crater is between 5,000°C. to 5,500°C. (1,000 c.p. per sq. mm.
vs. 140 c.p. of an ordinary arc2). By increasing the gas pressure about the arc up
to 22 atmospheres, Lummer of Breslau has raised the temperature to 6,000°C.
or very close to that of the sun. The Beck arc is practically applied to high
intensity search lights. Another extreme temperature arc is the tungsten arc
operated in hydrogen gas. The tungsten arc is commercially applied in the
‘^pointolite/, operating in an enclosed atmosphere of argon. The lamp is used
for microscope illumination. On a large commercial scale the electric arc is the
basis of the electric steel furnaces, such as theHeroult, the Yom Baur, the Snyder
furnace, etc.3 In the Birkeland and Eyde process for the fixation of atmosphetic
nitrogen, air is passed through an electric arc. The gases are heated to very high
temperatures and then rapidly cooled to counteract the tendency of the nitrous
acid to dissociate again into its elements.
In using the electric arc for melting or high-temperature heating we can make
the material to be heated or melted the lower electrode and use a carbon rod as the
tipper electrode, Fig. 2; or, in the reflected arc heating method, we can incline two
carbon or metal electrodes at an angle of about 120° and bring the material to be
heated under the arc (Figs. 3 and 4) or as in the Heroult furnace operate two or three
arcs in series (Fig. 5).
The advantage of heating or melting by the arc processes is simplicity and ease of
manipulation. On the other hand when operating under ordinary atmospheric
conditions it is difficult to avoid contamination of the material to be melted or under
test by vapors from the arc.
Electric-arc Furnaces.—The electric-arc furnaces are usually subdivided into
two classes, those that are designed primarily for the manufacture or refining
of steel and ferroalloys and those used in non-ferrous metallurgy, or so-called
electric brass furnaces. A large number of different makes of electric steel
furnaces are being marketed in this country and abroad. To mention but a few
of these: Heroult, Snyder, Rennerfelt, Greaves-Etchells, Groenwall-Dixon,
Xudlum-Girod, Booth-Hall, Moore, Webb, Stassano, Greene, Yolta, Yom Baur
a,nd Wile.
The furnace most widely used in this country is the Heroult. It was the' first
arc furnace to be used commercially in the manufacture of steel. This furnace has a
1 Trans. Am. Electrochem. Soc., Vol. 37, p. 665.
* C. S. McDowell, Trans. A. I. E. E. Vol. 34, pp. 363-85, 1915; G. Gehlhoff, EhUrotechn, Z. Vol.
42, p. 1315, 1921.
8 Cf. “Electric Furnaces in the Iron and Steel Industry,” Rodenhauser, Schoniawa, and Vom
Bauer.
bowl shaped hearth, similar to an open hearth, with no electrical connections through
the bottom. The electrodes pass through the roof and are suspended vertically, so
as to have the arcs in series, electric current passing from one electrode to the bath and
out through another electrode. Three electrodes are used for three-phase current and
two electrodes for single phase current. The furnace is of rugged construction, and
has a powerful and reliable tilting mechanism. The circular shape gives the maxi-
Fig. 2.—Girod. Fig. 3.—Stassano. Fig. 4.—Rennerfelt. Fig. 5.—Heroult.

Figs. 2, 3, 4, 5.—Arc-furnace types.
mum mechanical strengh and minimum surface for heat radiation. Reactance in the
furnace and in the power circuits is reduced to a minimum, thus giving a power factor
of 0.90 or higher. It is possible to do this, and still secure a minimum disturbance on
the power line, due to employing low voltage and short arcs. The Heroult furnaces
have been constructed in sizes ranging from 1 ton to 40 tons capacity and are oper¬
ated with either basic or acid hearths. They are extensively employed in making
tool steel, high-speed steel, special grade alloy steels, the melting of high-priced non-
ferrous alloys, the melting of ferromanganese for addition to open-hearth steel, the
making of small steel castings and malleable iron castings. The charges consist of
either cold scrap or partly refined steel taken from Bessemer converters or open-hearth
furnaces, or of molten iron from cupolas.
Electric Furnaces for Non-ferrous Metals and Alloys.—Besides the Baily
resistance-type furnace above described the Booth rotating furnace, the Detroit
rocking electric furnace, the General Electric muffled-arc furnace, the Yon Schle-
gell repelling-arc furnace, the Hennerfelt, the Northrup high-frequency furnace
and the Ajax induction furnace are some of the more common installations
for melting non-ferrous metals.
The Booth rotating furnace and the Detroit rocking furnace are similar in general
outline: A cylindrical shell with an electrode entering from either end. The arc
plays between the two electrodes and the heating of the brass
or other metal bath is by radiation or indirect heating. In
the repelling-arc furnace the electrodes are suspended through
the top of the furnace. The electrodes are balanced so that
they touch and hang together when no current is flowing.
The arc is formed by the mutually repelling force when cur¬
rent flows in the electrodes. The extent of this action is
regulated by balance weights on the electrodes.
In the G. E. muffled-arc furnace there are two heating
Fig. 6.—Ajax-Wyatt
units, one at each end of the furnace. Each unit consists furnace.
of two vertical electrodes, a carbon wearing block for each,
and two cross electrodes, which lie in a duct, one on each side of the wearing
blocks. The duct is filled with granular coke or graphite which muffles the arc and
supplies the heat, as numerous small contact and resistance arcs are formed in it by
the passage of the current through it between the vertical electrodes and the wearing
blocks, and hence to the cross electrodes. Although this furnace is inherently two-
phase, it may be operated either single- or three-phase, the most satisfactory phase
voltage being from 60 to 70 volts. There is no contact between the charge and the
electrodes, hence the tendency of the brass to become volatilized or pick up impurities
is reduced to a minimum.
Induction Furnaces.—Electric furnaces in which the heat is generated by
induction have undergone a marked development since the appearance of the
Colby, Kjellin and Roechling-Rodenhauser furnaces. In steel practice the
induction furnace has been largely replaced by the arc furnaces. In the brass¬
melting industry, however, the induction furnace is one of the most satisfactory
types. The Ajax-Wyatt furnace (Fig. 6) based on patents by Hering and Wyatt
consists of a vertical ring hearth. The heating is internal in the metal itself
which acts as the resistor. Through the pinch effect and motor effect the brass
is kept in constant circulation.
The “pinch” effect derives its name thus. When a current passes through a
■
molten metal the accompanying electromagnetic field causes the surface layers to be
LID
PBCVFNTS RADIATION AND CONDUCTION OF HEAT TO THE OUT.
S.DE. CONFmiNG HEAT TO THE CRUCIBLE INTERIOR
(. N
fclCCYRlCAL — ■.—CRUCIBLE LINING

COIL—— — holds the metal
CARRIES THE and HELPS KEEP
CURRENT, WHICH THE HEAT IN THE
PRODUCES THE CRUCIBLE
GRUOBLE COIL
ITSELF IS WATER
COOLED. MAKING
THE FURNACE BOX
COOL AND COM*
portable to
WORK WITJi
^THE CRUCIBLE
ELECTRICAL WHERE IT IS
INSULATION— — CENERATEO
»OR PROTECTION.
AMO TO HELP KEEP
THE heat in the
CRUCIBLE
IOX-FOR, MECHANICAL RlCIDITV

AND EASE Or HANDLING
\ / I SUPPORTS THE ELECTRIC^RUClBLE AND PROVIDES I \ /

\ * I CONTACT TO THE ELECTRICAL POWER SUPPLV \ /
Fig. 7.—Crucible induction-furnace.
attracted to the center. This contracts the cross section, “the pinch,” and if the
channel be a long horizontal one the conductor is likely to be broken at some point.
If the conductor be a vertical one the centripetal pressure of the outer layers cause
differences in pressure that tend to drive the center of the column up and allow the
outside to flow down. There is no electrode consumption and the thermal efficiency
has been exceptionally high.1 The furnace has been adopted very widely by those
brass rolling mills that operate 24 hr. a day and need not often change from one alloy
to another—conditions vital to the successful operation of this furnace.
The Ajax-Northrup High-frequency Induction Furnace.—Of all the electric
furnaces invented and developed, this Northrup furnace is undoubtedly the
most ingenious. It is in successful commercial operation at the U. S. Mint and at
several silver foundries. On a small scale in laboratories some 40 Northrup fur¬
naces have been installed. Tungsten metal can be melted in the Northrup
1 G. H. Clamer, Jour. Franklin Inst., Vol. 190, p. 473, 1920; H. W. Gillett, Electric furnaces for
on-fevrous alloys; Trans. Am. Electrochem. Soc., Vol. 39, 1921,
furnace. The metal or alloy is heated without the use of an arc or other electrical
connection. The material to be melted or treated may be surrounded completely
with refractory material so that the heat losses are a minimum. A temperature
of 2,600°C. can be attained within 15 min. The larger size furnaces will hold
550 lb. of pure silver operating at 75 kw., three-phase, 60 cycle, 6,600 volts. 1 he
crucible or container is surrounded by a single layer solenoid carrying the elect ric
current of high frequency 10,000 to 12,000 cycles per second. The coil remains
cold, while the metal or alloy or other material under investigation attains a, very
high temperature due to eddy currents induced in it.
Unlike other induction furnaces there is no necessity for an interlinkage of a mag¬
netic circuit with the electric circuit. The high frequency converter system has no
moving parts and is highly dependable. If a non-conducting material is to be melted,
a conducting crucible is used, while non-conducting crucibles are used when conduct ivo
materials are to be melted. For melting in vacuum a long quartz tube is employed.
As there are no electrical connections to pass through the walls of the quartz tube, a
high vacuum is easily obtained. In melting gold 180 oz. arc melted per kilowat t hour.1
Platinum melting is carried on in lime, zirconia or magnesia crucibles ami .since the
heat is produced in the material itself, there is very little contamination. (JlaHH may
be melted in a nickel container where the temperature is not over 1,425°C.
1 Electrical Review, Vol. 77, p. 184, 1920.
SECTION XY
MIXING AND KNEADING
By Dr. L. C. Reese1
Introduction.—Mixing and kneading are among the oldest, the most fre¬
quently employed and the most important operations in household and industry.
The ancients prepared and mixed powders, lotions and ointments for medical use;
they smelted metals from mixtures of ores and flux; they made and blended the
beautiful colprs and paints the unexcelled brilliancy and durability of which we
still admire in the relics and ruins which have come down to our time, and they
kneaded flour, water and leaven into dough which, baked, gave the light, spongy,
wholesome and easily digested bread still forming the staple food of the civilized
races.
Mixing was mostly, and kneading up to about 50 years ago was always, done by
hand. The simple implements used for making the work of the hands more effective
were, for small quantities, the mixing bowl or trough and stirring rod, paddle, spoon,
or fork, and for larger amounts, the mixing board or floor and the shovel.
Before a mixture is ready for use, it is frequently necessary that its components
should be subjected to another treatment, especially to grinding and sifting; both
operations, mixing on the one hand, and grinding or sifting on the other, can often
be combined with the greatest advantage. This old expedient is used also with
modern machinery where in any way possible and profitable. For instance, coarse
materials offering the same or similar resistance to pulverizing and therefore capable
of being evenly ground in the same mill, as many cereals, seeds, mineral colors, ores,
drugs, dyes, chemicals, etc., are placed raw in the required proportion in the hopper
of the mill, ground together, and the resulting rough mixture is then passed through
one or more sifters or bolting reels, whereby usually a satisfactory product is obtained.
Substances already finely powdered, as flour, ground condiments, salts, chemicals,
colors, etc., were formerly mostly mixed by sifting them several times through a sieve
of sufficiently fine mesh, thus removing coarse particles and loosening the materials
to make them more active or more receptive to further treatment, while at the same
time producing an intimate mixture. At present the mixing machine is usually pre¬
ferred for this purpose, as its action is quicker and in general also more thorough; but
the housewife still mixes the flour, baking powder and condiments for her cake by
passing them several times through her little hand-sifter.
Oil paints, as well those for ordinary use as the finest for art paintings,were
generally prepared by the painter’s assistants—the color grinders—grinding and
rubbing the finely powdered and sifted pigments by means of the stone into the oil
spread out on the slab, until the required shade, consistency and covering capacity
had been attained. This laborious and tedious work has been supplanted by the
power-driven mixing machines and paint mills used in the large factories which sup¬
ply the ready-made paints and colors to the trade in handy cans and to artists in
neat tubes of tinfoil.
In the scientific laboratory the chemist always comminuted and mixed his sub¬
stances for analysis and his chemicals to assist reaction, and the pharmacist prepared
his powders, pills and ointments by means of the mortar and pestle which since the
i Consulting chemical engineer, 1121 Adams St., Saginaw West Side, Michigan.
34 529
oldest times rightly forms the symbol of their trade. Today the chemist buys his
reagents, and the pharmacist his wares from large chemical and pharmaceutical fac¬
tories in which the simple mortar and pestle has long been superseded by numerous
types of mixing machines. Not only in these few examples described, but in every
trade in which mixing and kneading are done, production was transferred in our
machine age from the craftsman, working for a limited number of his neighbors, to
the large factories caring for the millions, which change was in all cases greatly assisted,
and in some directly caused, by the invention of the power-driven mixing and kneading
machines, following closely the development of a suitable power plant. The advan¬
tage of these modern machines does not consist so much in their better action, in the
attainment of a more homogeneous mixture, but mainly in the production of much
larger quantities at a huge saving of manual labor and working time.
Divisions.—Mixing operations and the appliances used are appropriately
divided according to the nature of the materials to be treated into the following
three classes: (1) Mixing solids with solids (coarse and powdered materials);
(2) mixing liquids with liquids (also while heating, melting, evaporating, etc.);
(3) mixing liquids with solids (solutions, creams, pastes, ointments, polishes, inks,
concrete, etc.) To these main heads I add the following, dealing with special
kinds of mixing: (4) Mixing and kneading doughs for bread (also for cake, bis¬
cuits, crackers, noodles, macaroni, etc.); (5) mixing and incorporating plastics
(rubber, putty, linoleum, soap, etc.); (6) whisking and emulsifying.
In Class (1) which comprises finely pulverized substances like flour, fuller’s earth,
chemicals, china clay, explosives, baking powders, spice, face and tooth powders,
chocolate, cocoa and other food powders, etc., as well as materials of a larger size and
coarser shape, as minerals, drugs, tea, coffee, insulating materials, fertilizers, glass
mixes, graphites, etc., all the more generally used mixing contrivances, mixing ma¬
chines with their various stirrers, agitators and blades, mixing drums and tumblers,
etc., are employed to intermingle the ingredients of a mixture so intimately as to
form a completely uniform, homogeneous mass. I therefore first describe these
general devices, and then particular machines and apparatus adapted and constructed
for working special materials.
Basic Principles.—When the quantity of matter to be mixed in one batch by
means of a machine provided with a certain mixing device is increased, the size
of the container of the machine must, of course, be increased in accordance with
the increased bulk of the mass to be worked, and in three dimensions, but the
active surface of the mixing device while retaining the same principle of construc¬
tion can only be enlarged in two dimensions. It naturally follows that the
amount of power and time required to obtain the same result for a unit of mixture
does not remain the same for all sizes of the same construction, but increases at
a considerably quicker rate than does the number of units for which the machine
is intended. In general, the highest degree of homogeneous mixing can be
attained in a machine only up to a certain size; above that the result deteriorates
more and more.
When having to mix very large quantities of materials at the same time, it is
therefore more advantageous to employ a battery of smaller machines, which do the
work in the most effective and economic manner, than to try to construct for the
purpose over-large machines which can only give imperfect results. For the same
reasons, when trying out new mixing devices, we often find that they seem to show
very promising improvement in the experimental stage, but disappoint when the
machine is constructed on a commercial scale.
MIXING AND KNEADING f>a I
In order to produce in a machine several mixtures of exactly the same comixnsdion

nd degree, it is necessary accurately to weigh or measure the materials to bo mixed,
nd to treat them in exactly the same manner, especially for the same number ot
evolutions of the agitator and for the same length of time. Mixing opera! ions are
reatly facilitated and accelerated by charging the ingredients into the machine
radually, and as far as possible in the same proportion in which they are to form the
squired mixture. Continuously working mixing machines provided with *sopnrate
oppers for each ingredient, delivering them in the right proportion into the machine
ithout interruption, and with an outlet through which the mixed mass is discharged
1 a continuous stream, have been frequently constructed and tried, but in these
lachines neither the feeding nor the mixing has hitherto been of that, uniformity
hicli in general the accurate work of mixing and kneading requires; they can there-
>re be used only in exceptional oases where absolute uccuracy and homogeneity are
ot required.
In many mixing operations, for instance in the manufacture of chemicals, pharma-
eutical preparations such as pills, drops, tablets, etc., of explosives, of dry colors,
Dior pastes, artists’ oil paints, etc,, it. is imperative that, the mixing be ho thorough
nd continued for such a time that the ingredients form a completely homogeneous
mss, otherwise the chemical and pharmaceutical products would have unequal and
ndesired effects, the explosives give uneven discharges, the colors and paints shew
breaks and spots, etc. In these cases it is therefore necessary accurately to dt'termme
le end of the mixing operation.
The finished mixture has usually a characteristic appearance, color, cousin*omn,
ensity, etc., by means of which after some experience the eye or hand can remhlv
xognizc at any time during the mixing operation the stab' of the mass under treat
lent, and its completion. When the ingredients consist of fKiwdcrs of different eiders*,
nts or shades, a small sample is thinly and evenly spread by means of a small nputula
n very smooth white paper; the number of the various colored streaks and
ms exposed and their different shades indicate the degree of mixing, which is fimahed
rhen the whole shows a completely even color. When the substances to be mixed
re composed of fine particles having a different shape, often also a slightly different
ize and color characteristic to each ingredient, the homogeneity or intimatiuiei« of
tie mixture produced can be judged with fair accuracy by examining a sample under
magnifying glass or microscope. In many eases also a suitable chemical teat run
e devised to find the end-point of the mixing operation.
The value of a mixing machine does not depend solely on the efficiency, durst* ion
nd cost of the work it performs, but is influenced by the ease and quickness with
duch it can be charged, discharged and cleaned, its bearings oiled, its driving gear
barted and stopped, etc. The protection of tin* attendant from being caught f*y the
loving parts inside and outside' the machine, us black's, pulleys, gear wheels* belts,
te., is also an important consideration when choosing the general arrangement find
osition of the machine'.
All mixing and kneading machines “including those intended for oompamt ively
ght work—should bo strongly built amt rigidly fixed to their frame and KUpfairts
o that all vibrations which might damage the machine or its driving gc'ur or dim ifimh
he efficiency of its working are prevented. The bearings of the running pur tit of
he machine must be so arranged and protected that no lubricating oils or oilier
materials used therein can enter into or come into contact with and thus contaminate
he contents of the mixer, and that no particle's of the mixture can penetrate** into mnl
bus gradually destroy the hearings.
Mixing machine's are', usually made of iron and steed. In cast's where iron must
iot rub against iron, the mixing parts are made of brass, bronze, wood or the like,
Vhen materials are to be mixed that for chemical reasons must not come into cotn „«#
with iron, the interior of the machine including the mixing devices may be tinned,
lined with lead, silver, nickel or other metal, enameled or protected in some other
suitable way.
For heating or cooling materials while being mixed, the trough or tank of the
machine in which they are treated is surruonded with a jacket through which a cur¬
rent of steam or of cooling water is passed; also hollow mixing blades and agitators
have been constructed for increasing the effect of the jacket. In order to prevent
the particles of the mixture from caking on or adhering to the heated or cooled walls
of the vessel, wdiich would make their proper mixing impossible and at the same time
greatly diminish the effect of the jacket, suitable scrapers arranged at or forming
one piece with the mixing device and passing closely along the walls must be employed
to keep them clean and thus distribute the heat evenly in the mass.
Provided the duration of the mixing operation is in all cases the same, in general
each type of mixing device requires a different speed for producing from the same
materials the same kind of mixture, and every mixture composed of different ingredi¬
ents or of the same ingredients in different proportions, when treated in the same
mixer, makes a different speed necessary for obtaining the same degree of intimateness.
Therefore, in order to mix successfully, attention must be given that the driving gear
alwrays provides the correct speed, and the operation is continued for the proper time,
as determined for the mixture by experiment and experience.
The higher the speed, the greater is the amount of power required as well as the
wrear and tear of the machine. In some cases it has been found advantageous to
change the speed of the mixing device during the operation, for which a change-speed
gearing, like Reeves", allowing one to alter the speed at will within certain limits, is
usually employed. Machines, in which it is desired to make several mixtures requir¬
ing different speeds, are generally provided with interchangeable gearings, one for
each speed needed. To reverse the direction in which the agitators or blades rotate
is a means frequently used to accelerate the mixing operation, and to facilitate the
removal of the finished mixture from the machine. A convenient reversing device is
the well-known double-clutch, slide-mounted on the driving shaft of the machine
between two loose power pulleys or gear wheels capable of rotating the shaft in oppo¬
site directions.
Hoppers suitably constructed with inclined sides and arranged above the machine
are generally employed to charge it with solid materials; the outlet opening of these
hoppers is frequently provided with sliding shutters or similar means to cause the
substances to be mixed to enter the machine gradually and in an even stream. For
charging liquids into the mixer usually a feeding tank is placed above it; the outlet
pipe is fixed to its bottom or lower part and provided with a stopcock for regulating
the flow of the liquid therefrom; a gage glass serves for observing and measuring the
quantity of liquid contained in and leaving the tank.
A great assistance in accurately charging a mixing machine and a considerable
saving of labor and supervision are gained by employing automatic weighing hoppers
and tanks, in which these containers are attached to scales provided with automatic
means to shut off the current of material entering them as soon as the weight of the
mass contained therein is equal to that required and previously adjusted on the scale
beam. Mixing and kneading machines are generally emptied by turning the mouth
or charging opening of their receptacle downward; only in the rare cases where this
is impossible they are discharged from the fop or through doors or manholes arranged
at their bottom or sides. When the center of gravity of those parts of the machine
to be revolved and of the mixture contained therein lies in or near the axis round
which the former can be turned, little power is required for this purpose, so that it
can be tipped over by hand; but where this is not the case, suitable power gearing of
ample strength or the assistance of a counter-weight mechanism must be employed.
MIXING AND KNEADING ■>:a
When great, accuracy is desired, modern machines arc' equipped with revolution
counters combined with an adjustable device for automatically stopping the machine
as soon as it has made the number of revolutions lor which the' device has been set.
The most favorable construction and size of a mixing machine lor a given quantity
and proportion of its ingredients, the most advantageous duration ot tin* operation,
and the exact velocity of the rotating mixing devices for obtaining tin* best jM>smhle
result can only be ascertained for each case preliminarily by experiment and definitely
by experience. The movements of the particles and molecules of a mass being mixed
are—even when a simple device is used—too complicated for any mathematical cal-
culations which might serve as a basis for the construction of a mixing machine; they
have been attempted, but led only to disappointment. A thorough knowledge' of
the effect of the simple standard mixing devices alone and in combination and of
the properties of the substances to be mixed and of the mixture, obtained therefrom,
and carefully planned and executed experiments form tho. sound foundation on whteh
a designer must base the construction and arrangement of successful mixing and
kneading machines.
Mixing Solids with Solids. -The facility with which solid substances ran be
mixed depends on their physical and chemical qualities: if their particles are
coarse or fine, long and thin, or short and plump, round or angular-shaped; if
they are absolutely free from moisture or air dry or contain still a certain amount
of water; if they are hygroscopic or not; if their particles adhere to each other
readily, hardly or not at all; if their qualities are altered by a slight change
of their temperature, and if they chemically affect each other.
When solid substances having a different specific gravity are violently shaken ami
stirred in a receptacle, the heavier ones will collect in (lie hot tom and the lighter ones
in the top layers of the mass. If substances thus treat cal have aliout f lie same aisecifie
gravity, but arc different in sixe, the coarser particles will gather m the top and the
finer ones on the bottom, and if their particles are of different shape, partly spherical,
globular or plump and partly angular, polygonal or elongated, the first the round
ones will collect in the lower part, and the second the cornered ones, in flteupfter
part of the mass. For these reasons the mixing appliances for such matters must be
so constructed that the lower layers of the mixture are turned again and again on
top of the upper layers, and so gently that, no reshifting or rearrangement tmti take
place, other than the one designed.
.In order to make a mixture of different solid mat tern mi Homogeneous as {mam Me
it is necessary to isolate all the single part idea of the components from each other
to combine them again in a fresh order and to rejamt these operntionn unfit the
particles are distributed through the whole mass in a« even and regular a manner ?m
required. lor this purpose, the single particles are tumbled over each other, preaned
down between each other and thrown in crosswise ami opjsisite nidi*wane directions
by rotating either the receptacle containing the nmm to be mixed, or by rotating an
agitator stirrer, mixing blade or the like arranged in a stationary receptacle for the
mass to be treated, or both the receptacle and the stirrer. The velocity of the rotio
tion must be such that the centrifugal force cannot appreciably intluewm the mixing
operation and its result. A temporary reversing of the direction of the rotation dm •
mg the operation is sometimes of advantage,
. receptacle, by the rotation of which the mixture m to be produced, m tumidly
circular in cross section, consisting of a cylinder, drum or mink. If an agitator forms
the exclusive expedient for mixing, it m in general placed within a stationary frough
having a half cylindrical bottom and rovolvea in an mm identical with that of flu*
latter, the outer tips of the agitator passing along the inner surface of the mtm\ Tin*
square open upper part of the trough is usually formed as a hopper and serves for
charging and discharging the machine. In case two or more blades are used for a
common receptacle, the latter consists of two or more adjacent parallel half cylinders.
The amount of power required for mixing solid materials is comparatively small.
Mixing machines, in the narrower meaning of the designation, are apparatus
in which the materials to be treated are placed in a trough or tank and mixed
by a mechanism revolving therein. For solid substances usually a half cylindrical
trough is used, the sides of which are extended upward to a square hopper. The
materials are mixed by an agitator or mixing blades supported in stuffing boxes
fixed to the flat sides of the trough in such a manner that their axis of rotation
is identical with the axis of the trough cylinder.
The machine is charged with the material while the trough with its hopper is
upright, and discharged by tilting it over so that the finished mixture drops from
the cylindrical bottom part along the sides of the hopper into the receptacle placed
underneath. This easy and quick method of filling and emptying the mixing trough
is a great advantage and the reason for the nearly universal employment of this
construction. The bearings, in which rotate the ends of the agitator shaft protruding
from the trough, must be durable and tight so that no fine particles of the materials
being mixed can escape from and no extraneous matter, as lubricating grease or dirt,
can penetrate into the interior of the trough. Dents, pockets, corners, angles, cuts,
incisions, grooves and the like, in which parti cles of the substances to be mixed might
settle and be retained, thus escaping the treatment, must be avoided as well in the
cylinder or trough, as in the agitators or blades. The shape and surface of the
rotating devices used must be such that, while they pass through the mass being
mixed, all the particles of the latter are successively acted upon by direct touch and
then easily slip or drop away again into the intermingling mass. The mixing trough
and agitator are in general made of a suitable metal, usually iron which, if required,
is tinned, nickel-plated or enameled. In case rough materials, the particles of which
are so sharp and hard that they wear away the metal, are to be mixed, it is a good
plan to provide the interior of the trough with a replaceable lining of steel-plate. In
order to prevent dust arising during the mixing operation from entering the surround¬
ing atmosphere, the open top of the trough hopper should be provided with a well-
fitting lid.
Samples of the mass under treatment should be taken only after the machine has
been stopped. The eagerness of the attendant is often so great that he tries to test
or take a sample from the mass being mixed without interrupting the work of the
machine. This thoughtless want of precaution has led to frequent accidents. The
machine ought therefore to be arranged in such a manner and put into such a position
that the attendant cannot reach into it during the operation; and, where this arranges-
ment is not possible, a lid ought to be provided which cannot be opened while the
agitators are rotating. In some states the rules of safety prescribed by law require
well-locking safety lids for all mixing machines.
The agitators are usually belt-driven, in the simpler machines the driving pulley
being fixed directly to the part of the agitator shaft extending outside the trough;
in more complicated machines, where two or more agitators are revolved by means
of gearwheels meshing with each other, the driving pulley is mounted on a main shaft
arranged in the framework supporting the machine; a pinion fixed to this shaft and
meshing with the gearing of the agitators imparts then the motion to the latter.
Lately by reason of its convenience and lack of parts requiring frequent repairs like
belts which only too often need to be taken up or mended, the direct electric drive
of mixing machines by means of gearwheels has been generally introduced in factories
where this power is available at a reasonable price.
MIXING AND KNEADING 535
The speed of the agitators is usually expressed, as for all rotating devices, in revolu¬
tions per minute. When the diameter of an agitator is increased, the circular path
which the lengthened part describes will be longer than that traversed by the original
part; the effect of the former on the materials to be mixed will therefore be greater
than that of the latter; in order to equalize these conditions, it is necessary to decrease
the number of revolutions according to the increased diameter with due consideration
of the particular shape of the agitating device and of the nature of the materials under
treatment.
When the machine is charged when the agitator is at rest, the latter has at first to
overcome the resistance of the inert mass to be mixed, which requires a considerable
greater amount of power than is necessary during the operation itself when all the
particles are in lively motion. This excess of power required depends on the nature
of the materials to be mixed, the width of the mixing trough, the height to which it has
been filled, and the shape of the agitator, and must be taken into consideration when
installing the driving gear of the machine. In order to avoid the sudden, jerk when
starting a fully charged machine, most attendants prefer to have the agitators running
before introducing the materials into the mixing trough.
I have often been asked what I consider to be the most advantageous dimensions of
the trough of a mixing machine for solids in regard to its length and diameter, if an
absolutely homogeneous mixture is to be obtained. The area affected by any point
of an agitator at any moment under the most favorable circumstances cannot exceed
a right triangle the base of which is the axis of the agitator and the sides of which form
an angle of 45° with the vertical line drawn from the given point to the axis of the
agitator. If therefore that part of an agitator working in the middle between the flat
sides of the trough is to have any influence at all on the particles situated at these sides,
it is necessary that the length does not exceed the diameter of the trough. This
deduction is in complete conformity with experience and experiment. In cases where
the agitator consists of arms passing closely along the vertical ends of the trough and
so constructed as to drive the particles of the mixture from the sides towards the inner
part, the thickness of these arms, measured on a line parallel to the axis of the trough,
may be added to its diameter for its total length. The troughs of the now usually
employed various types of mixing machines, in which the mixtures are made batch-
wise, agree completely or very nearly with the condition laid down above; only for
machines in which agitators are used causing a predominating sidewise movement of
the mass under treatment, and which are usually intended for continuous work,
troughs of greater length have been tried. I describe in the following paragraphs,
first, the less elaborate forms of the agitators used in this kind of machines and their
effect, illustrating these devices by means of the diagrammatic Figs. 1 to 9 in all of
which a represents the lower, half-cylindrical part of the trough containing the mate¬
rials to be mixed, b the open hopper above a, c the agitator, d the agitator shaft and e
the axis round which the shaft d and the agitator c rotate, and which is identical with
the axis of the cylindrical part a.
When first constructing machines to perform the work done by their hands and
arms, men always try to imitate their shape and action. Thus we perceive in the
earlier mixing machine stirring rods, fixed to the revolving agitator shaft, replacing
the action of the human fingers. Figures 1 and 2 represent vertical sections of the
mixing trough and agitators of this simple machine, Fig. 1 in a plane vertical to the
axis e, and Fig. 2 in a plane through the length of the latter. The round or square
rods / are fixed to the shaft c which is supported by the bearings g arranged in the sides
of the trough a, and rotated by means of the pulley h and belt i. The rods / which
may be of any convenient number, are arranged in alternate rows and so long that
they pass along the circular bottom of the trough a which is as long as its diameter;
their length was usually 9 to 15 in.
In this and similar machines in which the shaft extends their whole length, it is
usually enlarged to a hub into which the rods or other mixing devices are screwed or
fixed in any other suitable way, or with which they are cast in one piece.
When this type of machine is charged above its shaft or hub, it does not mix well,
because the particles surrounding the latter are hardly affected even when the agitator
is rotated quickly; the trough ought therefore only to be filled up to a line fairly under¬
neath the shaft, as indicated in. Fig. 2 by the dotted line X. When the agitator c is
then rotated slowly—say 8 to 10 r.p.m. the rods / pass through the mass in the
trough a doing little more than disturbing it; no real mixing can take place, but, when
the speed is increased to 30 to 60 r.p.m., in accordance with the material under
treatment, the whole mass will be whipped into motion, the particles whirling in
different directions and at different speeds through, between and against each other,
whereby a thorough mixing is made possible.
When the substances to be mixed are partly coarse and partly fine, or of different
size, the mixing will remain rough, as the rods will always rake the coarser particles
upwards into the top layers again. In case all the ingredients are of powder form, but
-2.—Stirring-rod mixer.
differ greatly in regard to their specific gravity, the heavier ones are inclined to collect
mainly at and near the walls of the trough a, and their mixture will thus not be perfect.
But when the powdery substances have the same or nearly the same specific gravity,
they intermingle in this manner well, and the mixture will be practically homogeneous
after the operation has lasted its due time. This simple machine which requires only
a very small amount of power was formerly frequently used for mixing paints, different
kinds of flour, powders etc., but is today mostly replaced by more modern devices.
In order to accelerate and intensify the mixing, two or more rod agitators were
sometimes arranged in one trough for combined work, the rods of the one agitator
passing through between those of the next agitator with their tips near to their respec¬
tive hubs. Figure 3 represents a machine which I saw a good many years ago in an
ocher mill where it was successfully used for blending ochers, umbers and similar
pigments to a completely uniform mass of exactly the same even color. The three
agitators, c, cx and c2, were inclosed in the polygonal wooden box a serving as mixing
trough; their parallel shafts were placed at distances from each other somewhat
greater than the length of their rods. The driving gear was so arranged that, when the
agitator Ci, rotated in one direction, c2 revolved in the same and c in the opposite
direction. The rods/, /i and/2 were simply securely screwed into wooden hubs in rows
of four, in planes vertical to the axes, and / and /2 situated between the planes of two
adjacent rows fi. The box a was changed through the door k at its top nearly up to
the hub of the agitator clf as indicated by the line X, and discharged through the door
l at its bottom; both doors were tightly locked during the operation. The capacity of
the machine was 5 to 6 cwt., and the average time of mixing for obtaining an absolutely
blended color 4 to 5 min.
The general arrangement of the machine shown in Fig. 4 is the same as that of
Fig. 1 with the exception that the rods / are replaced by paddles m and mh the broad,
straight outer edges of which pass closely to
the inner surface of the trough a. The
width of the broad upper part of the pad¬
dles is somewhat greater than the distance
of their centers so that their alternately
arranged rows m and mi, cover more than
twice the whole length of the mixing trough
a when the agitator c is rotated. The power
required for the same speed is naturally
greater in the paddle- than in the previously
described rod-agitator. At a speed of about
10-20 r.p.m. the paddles m, and mi will
press the layers of materials at the bottom
of the trough a forward and upward along
its cylindrical wall, the portion retained, Fig. 3.—Pigment mixer,
when above the mass, dropping on to the
top of the latter, while the rest escapes through the open space between adjacent
paddles and at their back edges into the upper layers; the empty spaces produced
behind the moving paddles are at once filled out sidewise and from the inner and
upper layers; in this way all the ingredients are gradually and gently intermingled,
even when the difference in their specific gravity is considerable.
The device works well for finely pulverized substances and for coarser ones of small
Figs. 4-5.—Paddle mixer.
and practically equal size, but larger sized and fine materials can not be thoroughly
mixed together by flat paddles, as during their upward movement they will retain the
coarser particles, while the finer ones escape through the interstices of the latter and
are thus sifted out. In this case the paddles are constructed in the shape of scoops
or shovels, in the concavity of which, pointing in the direction of the rotation, the
finest particles will remain, until at the highest point they are ejected onto the top of
the mixing mass. Sometimes it is of advantage to arrange the scoops or shovels m
and mi pointing forward as illustrated in Fig. 5, their broad front edges acting as
scrapers and catching up everything in their way. Considering the construction, the
mixing effect is good and at all events satisfactory for the purposes for which this kind
of agitator is frequently used, as mixing of different minerals, cereals, chemicals, lumps

of color, clay, etc., ores with flux, glass mixes, etc., etc. The result is also here more
satisfactory, when the top of the mass to be mixed is kept somewhat underneath the
hub of the agitator c, as shown by line X in Figs. 4 and 5; a higher speed of the agi¬
tator—say 30 to 40 r.p.m. is of little or no advantage in this type of machine.
In order to counteract the friction of the particles of the mass under treatment
on the vertical sidewalls of the trough, the machine shown in Fig. 6 is provided with
crosswise arranged side scrapers o and oh which at the same time serve as main support
for the four narrow blades p, pi, which extend the whole length of the trough a. The
scrapers o and oi are fixed to the hub c and arranged so that their sharp front ends are-
drawn tightly along the surface of the side walls during the rotation of c; the blades
Figs. 6-9.—Mixers with scrapers.
are riveted or otherwise rigidly fixed to the outer ends of the scrapers in such a man¬
ner that their outer edges pass closely to the circular surface of the trough bottom a.
The crosspieces q and qi connecting the blades p and pi with the hub, and the cross¬
rods r connecting two opposite scrapers are intended to strengthen the rigidity and
increase the effectiveness of the device.
The agitator shown in Fig. 7 has only two blades; its side scrapers o are reinforced
to arm-like structures and cast in one piece with the broader and strengthened blades
p and the bosses s with the trunnions t which replace the ends of the agitator shaft d
in the devices described above. In this way the obstruction of the shaft and hub to
the mixing operation is avoided, and the trough may be charged up to the line y.
Often crossrods q connecting the arms o and similar, even more complicated contriv¬
ances are used with the intention to improve the mixing effect of the arm-blade agi¬
tator. Of great importance to the good working of the latter are the shapes of the
front and back surfaces which are in active touch with the mass in the trough. Figure
8 is a cross-section of the blade p in a plane vertical on the axis e of the device, and
Fig. 9 a cross-section of the side arms o in a plane parallel to the axis e. In Fig. 8, (1)
indicates the front surface of the blade acting on the material and situated in its whole
length in the plane (2) passing through the axis e; the back surface (3) of p is curved
so that the particles coming into contact with it easily slip and drop away again.
In Fig. 9, a denotes a sidewall of the trough of the machine, and o the sidearm
moving closely along a; the front surface 4 of o is inclined in the direction of the rota¬
tion indicated by the arrow, the particles at and near the wall a being thus forced
away from the same into the interior of the mass; the back surface 5 of o is curved
so that particles gathering in the angle formed by it and the sidewall a are not retained
there, but drop easily away during the upward movement of the arm o. At a slow
3pCCCl—Say 10-25 r.p.m.—powdery substances and coarse ones of moderate size an*
.veil mixed in this type of machine; the amount, of power required is not greater than
,hat necessary for the same work in the machine shown in Fig. 4.
The simple arm-blade agitator just, described is frequently used in a. modified
ionstruction for mixing liquids, doughs and pastelike masses; but. though sometimes
lseful for mixing solids, it has been in most eases superseded by its combination with
,he oblique-disc agitator and double-trough device and their numerous modifieations
tnd imitations. This particular construction was introduced about 40 years ago;
the belief that a good mixer is necessarily also a good knender for bread dough/
,he machine was at first especially intended for this heavy work ami therefore was
milt very strongly and solidly which is a great advantage* for any mixing machine.
Fioh. 10 13. Oblique mixing-dim* machine.
Che excellent result of the mixing, tlie. short, time and small amount of j>ower required
ogether with the simplicity, durability and practical freedom from the frequent,
epairs which are such a drawback in many other constructions and lighter budf
nachines, are the reasons that this particular machine is now employed in practically
'very trade in which mixing is done, and in every civilized country. Figures 10 to 1?
how the principles of the construction of the doubled rough and oblique disc-mixing
nachine and are based on the patent drawings'* describing the original ideas of the
nventors. Fig. 10 is a horizontal section of the machine through the axis of its
igitator, Fig. 11 a vertical section on the line (l t) and Fig. 12 on the line (2 21 of
fig. 10; Fig. 13 shows the elliptical circumference of the agitator blades. A indicatea
he mixing trough, H and C the two agitators and D their driving gear; a is the upper
topper-like part of the trough A; ti\ and a% are the two equal parallel half cylinders
orming its bottom; «a is the edge or saddle in which rq and meet, and <q and «8
1 See p. 550.
2U. S. Pat. No. 180,508, Aug. 1, 1870, Paul Fimvittmtitui for ** Knotting Machine;” F. S. Pat
ô. 254,042, Fab. 21, 1882, Paul PpLF,m»»«m for “ Mixing and Knotating Mm'hmtF, H. Put. No.
!4,908, Feb. 20, 1895, Paul Pflbimekb for ** Machine for Mixing, Kneading, etc.”
are the stuffing boxes in which the shafts of the agitators B and C rotate coaxially to
a\ and a2 which must be so arranged to each other and in regard to the saddle a 3 that
the tips of the agitators B and C, when rotating, pass as closely as possible to each
other above az in the center part of the trough A. b indicates the elliptical disc form¬
ing the mixing blade of the agitator B, hi its hub or stem.
The ellipse forming the circumference of the blades, Fig. 13, has as its minor axis
the diameter of the cylindrical part of the mixing trough, and as its major axis the
hypotenuse of a right-angled triangle the two sides of which are respectively equal to
the diameter and the length of the trough cylinder. The plane bounded by this
ellipse, which we may consider as the ideal working surface of the design, forms the
front of the blades b and c the centers of which are identical with the centers of the
trough cylinders a\ and az', b and c are arranged at their corresponding hubs 61 and Ci
and in regard to the shafts 62 and C2 revolving coaxially to ai and a2 respectively, in
such an inclined position that the tip ends of the major axis of the ellipse, describe
when rotated circles closely identical with the circles forming the circumference of the
two opposite flat ends of the cylindrical parts of the mixing trough ax and a2) as shown
in Fig. 11. In Fig. 10 the minor axis of the ellipse of the blade b is vertical to the plane
forming the open top of the half cylindrical bottom ax, while that of c lies in the corre¬
sponding plane of a2, appearing, seen from above, as a circle; in Fig. 12 c is in the same
position, but seen from the side. When the blades are revolved as indicated by the
arrows in Fig. 11, they act on the material to be mixed twice, first with their front and
then with their back surface, during each rotation. In order to give to the blades the
necessary strength, their back surfaces decline towards their hubs as shown in Figs. 10
and 12 and can therefore not be so effective as the even front surfaces, the hubs
naturally also obstructing the mixing at this side of the blades. Like the edges of the
two blades of the agitator shown in Fig. 7, those of the elliptical blade describe in
outline a cylinder corresponding to the half-cylinder of which the bottom of the trough
consists, but act in consequence of their greater length on a correspondingly greater
amount of particles. The elliptical blade not only pushes the material forward, then
upward, and turns it over, but causes two strong sidewise motions during each revolu¬
tion, first to the one and then to the opposite end of the trough.
This strong effect is still greatly enhanced by the use of two such blades B and C
rotating in the double trough A in the direction shown by the arrows in Fig. 11. At
each revolution each blade will twice press over a portion of the material contained in
the half cylinder, in which it moves, into the adjacent one, and will twice draw portions
from the latter into its own half, provided—of course
—that the trough is charged with material well above
the saddle between its two half-cylindrical bottom
parts. A different velocity of the two blades which
prevents them from working against and thus
obstructing each other, has proved advantageous.
In Fig. 10 the proportion of the size of the spurwheel
Fig. 14.—Cut-out blade d to that of di is 4:3; the refore the blade C driven
for dough. by di will make 4 revolutions in the same time in
which the blade B rotated by d makes 3.
It is obvious that a full size oval blade arranged and worked as described will
exercise a considerable outward pressure on the material being mixed; and, when the
latter is of a thick and heavy character or plastic like dough the particles of which
adhere strongly to each other, the strain on the machine and the power required to
drive it will be especially great. In order to avoid these strains and to facilitate the
working, gaps having usually curved outlines were provided in opposite sides of the
periphery of the disc blade, as shown in Fig. 14.
The construction of the agitator shown in Fig. 10, in which the shaft carrying the
disc blade is supported only at one end of the trough, is in most cases too weak for
work on any large scale. In commercial machines therefore the shaft revolves usually
in bearings provided in each end of the machine and, when extending the whole length
of the trough, is reinforced to a hub at the middle part of which the oblique blade is
arranged; the latter, the hub and the shaft are usually cast, in one piece. As this long
shaft or hub takes up a not inconsiderable portion of the active capacity of the mixing
trough without assisting—if not actually
obstructing—the mixing operation, it, is
omitted where the shape of the blade re¬
quired for the material to be treated, allows
to do so, and the blade is in that, case fixed,
as shown in Fig. 15 at t he ends of its major
axis to sidearms constructed like those
described in Figs. 7 ami 9.
In Fig. 14, a indicates the blade, b the
hub and c the shaft of the agitator; ax and
a% are the gaps cut from the ideal elliptical
plane of a. In Fig. 15 a and ax denote the
two half-cylindrical bottom parts of the
mixing trough A with a2 the saddle or cut¬
ting edge between them; B is one oblique
blade and hi the sidearms carrying it, while
c and Ci indicate the corresponding parts
for the blade C; d is one spur wheel, and
di the other; the proportion of the diameter of d to that of <h is 7:4. It will be
noted that, when looking at the agitator in the direction of the minor axis of its blade,
as at C, the blade C forms with the sidearms cx and the hubs the letter Z.
In the actual blades the gaps in the perimeter of the disc plane are constructed
according to the effect to be obtained, small or large, and either of the same or similar
size and shape as in Figs. 14 and 15, or of unequal and unsymmetrieal dimensions as
in Figs. 16 and 17. In Fig. 16 the smaller gaps of the two blades are arranged at, one
end and the larger ones at the opposite end of the double trough, while in Fig. 17 a
small gap of the one blade is shown opjyosite to the large gap of the second blade at
each end of the trough. 1 consider the last arrangement preferable in all ordinary
cases of mixing and kneading. In some eases only a comparatively narrow strip is
left at opposite sides of the eireumferenee of the blade besides the small middle part
connecting the two outer sections, as shown in dotted lines in Fig. Ki As the* iron or
other material of which the blade is made must-*™-of course—-be of a suflieient strength
and thickness for the work, and as the active surfaces lie in one plane, but at ommsite
sides of the blade, the latter will have a peculiar twisted shape and appearance. A
completely correct pattern for a blade of this class is obviously very difficult to make;
expediency in pattern-shop and foundry are probably the reasons that blades made in
later years often deviate to their detriment from the fine, correct forms produced by
the original makers. Angles and corners not quite avoidable in this construction must
be made as obtuse as possible or rounded off, as also all back surfaces curved, as
already explained in regard to Fig. 7. All parts of the blade and especially the corners
must have as smooth a surface as possible in order to get the best result.
A reversing of the direction of the rotating blade will quickly remove particles
from corners where they are not under the full mixing action, and thus greatly assist
in the mixing operation, especially when sticky, adhesive or plastic masses are treated.
When the direction of the two blades in a double trough is reversed, the material
will be drawn away from the middle part of the trough and pressed against and piled
up the sides of the trough and its prolongation, and will thus greatly facilitate the
passage of the finished mixture from the machine when the latter is tilted.
The strong side currents prevailing in the material treated in this class of mixer
make it possible, without appreciably impairing the effect, to lengthen the half-
cylindrical trough somewhat above the proportion to its width given above, page 535.
The ideal proportion of the length to the diameter of the trough for a single blade
machine would be 1:1, and it should in practice not be increased above 3:2. The
proportion 1:1 is shown in Fig. 10, for a two-blade machine, the length of the double
trough A being equal to the diameter of one half-cylinder, thus half of its total width.
The same proportion, but modified by the deduction of the combined widths of the
two sidearms, is given in the Figs. 15 and 16. When applying the above limit, 3:2,
of the one-blade to the two-blade machine, the proportion for its trough would be 3:4;
but, as this machine is more effective, the proportion may be still slightly increased;
for larger machines the trough is usually made about square, 1:1, and in some cases
even a shade longer than wide. It appears obvious that with the increase of the
length of the blade its mixing effect will be diminished and the time required for the
mixing prolonged.
The speed necessary for the elliptical agitator, when mixing solids, is a low one,
9-25 r.p.m., and the power consumed is small indeed. In a two-blade machine the
relation of the speed of one blade to that of the other one is arranged with advantage,
according to the materials to be mixed, between 2:3 and 1:2. For instance, I mixed
in a two-blade mixer 420 lb. of flour with 8 oz. tartaric acid, the speed of the one blade
being 8, and that of the other one 13 r.p.m., and obtained after 11 min. a mixture
which according to a number of chemical tests of the different portions of the mass
was completely homogeneous. The two-blade machine still mixes perfectly well
when charged to the top of the blades, which should just protrude from the mass when
at rest; this—in comparison to other types of mixers-—extraordinarily large working
capacity of the trough in proportion to its size is one of the greatest advantages of
this machine.
Figure 18 represents the usual form of a single-blade mixer, the trough being tilted
for emptying, which is effected by means of a wormwhecl and quadrant worked by
a handwheel. It mixes substances offering a slight resistance as light, dry powders,
thin pastes and doughs very well.
Figure 19 shows the best-known type of a two-blade mixer, the trough a being
completely tilted, thus showing the Z-shaped blades b and c. The trough is counter¬
balanced, d designating the rod and chains carrying the counter weight and e the
hand-driven gearing working the arrangement; / shows the reversing gear. This
machine is highly efficient in every kind of the ordinary mixing of solids, doughs,
pastes, etc.; it has the largest capacity possible for this type and is easily emptied
and cleaned; it is a good all-round mixer.
In Fig. 20 the wing-like blades of reduced elliptical shape are made of one piece
with the hub and shaft. This form mixes heavy powdery substances, the specific
gravities of which often greatly differ, very satisfactorily. Figure 21 shows in a
nearly square double-trough two agitators each of which carries two parallel winglike
Fig. IK. ~ Single-blade mixer. Fiu. 10.--Two-blado mixer.
blades arranged obliquely at the hub in such a maimer that the inner tips of these
blades rotate in a circle close to the inner surface of their half-cylindrical trough part
and in the middle of its length. It is used for very heavy work. Other character¬
istic blades of this kind are mentioned in later paragraphs.
Figures 1 to 17 describe only the simple, fundamental shapes and arrangements of
mixing blades; extraordinarily numerous are their mollifications, additions and com¬
binations designed either for a certain purpose, a particular kind of mixing or particu¬
lar materials to be mixed, or for general improvement of the mixing effect. These
improvements consist mostly in turnings, curvatures and inclinations of the active
planes of the blades, in bendings and angles in Hit's sidearms and cross connections, in
specially shaped and arranged paddles and rot Is, in little plates, plain as well m of an
intricate shape, fixed to the shaft or hub, the main parts of the blades or its sidearms,
and in wire-like rods crossing in any direction from one part of the blade to the other
one or spreading from one set of arms to the following ones. It would bo of no
advantage to describe all these more or less complicated minor devices; though some¬
times answering their special purposes well enough, it is only too often that on exactly
testing them, the unbiased exi>erimenter finds that if there is any improvement in
the mixing effect perceptible at all, it is infinitesimal, as the additional currents and
sides of the blade, the latter will have a peculiar twisted shape and appearance. A
completely correct pattern for a blade of this class is obviously very difficult to make;
expediency in pattern-shop and foundry are probably the reasons that blades made in
later years often deviate to their detriment from the fine, correct forms produced by
the original makers. Angles and corners not quite avoidable in this construction must
be made as obtuse as possible or rounded off, as also all back surfaces curved, as
already explained in regard to Fig. 7. All parts of the blade and especially the corners
must have as smooth a surface as possible in order to get the best result.
A reversing of the direction of the rotating blade will quickly remove particles
from corners where they are not under the full mixing action, and thus greatly assist
in the mixing operation, especially when sticky, adhesive or plastic masses are treated.
When the direction of the two blades in a double trough is reversed, the material
will be drawn away from the middle part of the trough and pressed against and piled
up the sides of the trough and its prolongation, and will thus greatly facilitate the
passage of the finished mixture from the machine when the latter is tilted.
The strong side currents prevailing in the material treated in this class of mixer
make it possible, without appreciably impairing the effect, to lengthen the half-
cylindrical trough somewhat above the proportion to its width given above, page 535.
The ideal proportion of the length to the diameter of the trough for a single blade
machine would be 1:1, and it should in practice not be increased above 3:2. The
proportion 1:1 is shown in Fig. 10, for a two-blade machine, the length of the double
trough A being equal to the diameter of one half-cylinder, thus half of its total width.
The same proportion, but modified by the deduction of the combined widths of the
two sidearms, is given in the Figs. 15 and 16. When applying the above limit, 3:2,
of the one-blade to the two-blade machine, the proportion for its trough would be 3:4;
but, as this machine is more effective, the proportion may be still slightly increased;
for larger machines the trough is usually made about square, 1:1, and in some cases
even a shade longer than wide. It appears obvious that with the increase of the
length of the blade its mixing effect will be diminished and the time required for the
mixing prolonged.
The speed necessary for the elliptical agitator, when mixing solids, is a low one,
9-25 r.p.m., and the power consumed is small indeed. In a two-blade machine the
relation of the speed of one blade to that of the other one is arranged with advantage,
according to the materials to be mixed, between 2:3 and 1:2. For instance, I mixed
in a two-blade mixer 420 lb. of flour with 8 oz. tartaric acid, the speed of the one blade
being 8, and that of the other one 13 r.p.m., and obtained after 11 min. a mixture
which according to a number of chemical tests of the different portions of the mass
was completely homogeneous. The two-blade machine still mixes perfectly well
when charged to the top of the blades, which should just protrude from the mass when
at rest; this—in comparison to other types of mixers—extraordinarily large working
capacity of the trough in proportion to its size is one of the greatest advantages of
this machine.
Figure 18 represents the usual form of a single-blade mixer, the trough being tilted
for emptying, which is effected by means of a wormwheel and quadrant worked by
a handwheel. It mixes substances offering a slight resistance as light, dry powders,
thin pastes and doughs very well.
Figure 19 shows the best-known type of a two-blade mixer, the trough a being
completely tilted, thus showing the Z-shaped blades b and c. The trough is counter¬
balanced, d designating the rod and chains carrying the counter weight and e the
hand-driven gearing working the arrangement; / shows the reversing gear. This
machine is highly efficient in every kind of the ordinary mixing of solids, doughs,
pastes, etc.; it has the largest capacity possible for this type and is easily emptied
and cleaned; it is a good all-round mixer.
In Fig. 20 the wing-like blades of reduced elliptical shape are made of one piece
xvith the hub and shaft. This form mixes heavy powdery substances, the specific
gravities of which often greatly differ, very satisfactorily. Figure 21 shows in a
yearly square double-trough two agitators each of which carries two parallel winglike
Fig. 18.—Single-blade mixer. Fig. 19.—Two-blade mixer.
blades arranged obliquely at the hub in such a manner that the inner tips of these
blades rotate in a circle close to the inner surface of their half-cylindrical trough part
and in the middle of its length. It is used for very heavy work. Other character¬
istic blades of this land are mentioned in later paragraphs.
Figures 1 to 17 describe only the simple, fundamental shapes and arrangements of
mixing blades; extraordinarily numerous are their modifications, additions and com¬
binations designed either for a certain purpose, a particular kind of mixing or particu¬
lar materials to be mixed, or for general improvement of the mixing effect. These
improvements consist mostly in turnings, curvatures and inclinations of the active
Figs. 20-21. Fig. 22.
planes of the blades, in bendings and angles in the sidearms and cross connections, in
specially shaped and arranged paddles and rods, in little plates, plain as well as of an
intricate shape, fixed to the shaft or hub, the main parts of the blades or its sidearms,
and in wire-like rods crossing in any direction from one part of the blade to the other
one or spreading from one set of arms to the following ones. It would be of no
advantage to describe all these more or less complicated minor devices; though some¬
times answering their special purposes well enough, it is only too often that on exactly
testing them, the unbiased experimenter finds that if there is any improvement in
the mixing effect perceptible at all, it is infinitesimal, as the additional currents and
movements caused by the modifications interfere with, derange, counteract or deflect

the main and decisive currents produced by the original blades, thus weakening and
diminishing their force and effect, while at the same time requiring a greater amount
of driving power.
The mortar and pestle—-already mentioned—are made of steel, brass, china,
clay, glass or agate according to the substances to be treated. Figure 22 represents
a vertical cross-section of a mortar A and its pestle B. A consists of the walls a,
of the mixing bowl b and its rim c. The mixing bowl forming the bottom of A
is circular in horizontal and elliptical in vertical cross section; its rim is slightly
inclined to the outside. The pestle B is composed of the rod d with handle e
and the spherical working surface /. For mixing and rubbing together powdery,
friable or soft substances, it is advantageous to have the radius of the surface /
small, while for crushing hard materials as ores, minerals, etc., the radius must be
large. For quick and efficient work the mixing bowl b should not be more than
half filled, as shown by line X.
When pressing the pestle A against the surface of the mixing bowl b and moving
it round and round therein, those particles of the substances to be mixed which are
compressed between the surface of b and that part of the surface / nearest to the same,
are rubbed together and thus intermingled, while those particles coming in touch
with the remainder of / and the lower part of d are pushed aside and forced between,
and above the particles not directly disturbed; those driven up the rim drop down
again onto the upper layer in the bowl, an action, which, as soon as required, is assisted
by means of a spatula which at the same time serves to loosen particles adhering to
the walls.
In factories of pharmaceutical products often mortars of a considerable size are
employed, the pestle receiving its gyratory movement from gearing driven by machine
power; frequently also the mortar is simultaneously rotated round its center; scrapers
following with their outer edges the inner outline of the mixing bowl and revolving
with the pestle serve for turning the material over, before it is again subjected to the
action of the latter. The largest mortar I saw at work had a diameter of a little
over 3 ft. 4 in. These large machine mortars are not infrequently used for mixing
powdered or dry substances, but more often for making pill masses, ointments, salves,
creams, etc., etc.
The mixers and mills especially designed for the latter work do it usually more
thoroughly and quickly than the large power-driven mortars which on the otheh with
have the advantage that all the parts of the mortar and pestle coming in touch with
the materials to be mixed can be made of any non-corroding material, capable of
withstanding any action of the substances thereon and any chemical reaction taking
place in the mass being mixed, and that they can be very easily and quickly cleaned
without leaving the least traces of former operations, thus being speedily ready for
a new mixture.
Drum Mixer.—An old expedient to mix dry solid substances of any kind,
coarse or fine, is to place them in a cask, filling it to about K to H of its volume,
close it firmly and then roll it slowly about on an even floor until the materials
therein, continuously tumbling over and through each other, are perfectly
intermingled. In the modern machines instead of the wooden cask, generally a
cylindrical drum of metal is used which receptacle is rotatably suspended between,
two uprights of the framework carrying the apparatus. To this end, the flat
sides of the drum or cask have in their centers rigidly attached thereto sockets in
which the shafts, one of which carries the driving pulley or gear wheel, are fixed
These shafts and the bearings provided in the uprights in which they revolve are
so arranged that the axis of rotation is horizontal and identical with the axis of
the receptacle. Another arrangement for rotating the drum, especially when
of a large size, is to surround it in planes vertical to its axis with toothed rings
meshing with gear wheels fixed to the driving shaft rotating in bearings provided
in the frame work carrying the machine; loose rollers mounted in blocks fixed to
the basis of the apparatus serve usually as main or additional supports for the
drum and its often very heavy contents.
The action of the drum mixer is always gentle and, if the operation is properly
arranged, very fragile substances, like dry leaves as tea, flowers, seeds, crystals, etc.,
can be thoroughly mixed without being broken up or damaged in the slightest way.
Fine powders, as well as coarser substances of great difference in specific gravity are
also easily and perfectly intermingled and blended.
A great advantage of this method of mixing consists therein, that, as no parts of
the apparatus can exert at any point any pressure on the mass under treatment, as is
the case in most other machines, the single particles of this mass cannot be compressed
to form small lumps or scales which could only be separated with difficulty from the
otherwise perfect mixture. For instance impalpably pulverized different pigments,
fillers and driers can be mixed in the drum to a uniform shade without requiring to be
resifted. For the same reason, and as the closed drum excludes the free access of
air, hygroscopic substances can be evenly distributed in other matters, as for example
dried malt extract in cereal flour.
The naturally slow action of the drum mixer is accelerated by arranging in its
interior rows of stirrers or lifting plates fixed to and extending along the circular walls
radially and parallel to its axis; in this manner a considerably larger proportion of the
material under treatment is lifted at every rotation of the drum than when the interior
of the latter is plain. According to the requirements of the case, the stirrers may be
arranged in sequence or alternately and may be thick or thin, narrow or broad paddles,
often slightly inclined towards the direction of the rotation, and the plates may extend
the whole length of the circular sidewalls or only for a part, in which latter case they
are arranged so that always the interstice between two plates in one row is followed
by a plate of similar width in the succeeding row. The radial length of the stirrers
must not exceed one-half of the length of the radius of the drum, and that of the
plates must be still less because otherwise their effectiveness suffers by carrying the
material too far.
It appears to be obvious that a long drum does not mix so well as a shorter one of
the same diameter, as the particles contained in the former intermingle sidewise to the
same limited degree as those contained in the latter. For obtaining a perfect mixture
the ideal construction is here also that the length of the mixing drum is equal to its
diameter. Quite a number of shapes differing from the described cylindrical one,
polygonal in cross-section, sometimes cube or boxlike, are employed for the rotating
mixing vessel; its axis is in some apparatus so arranged as to form an angle with the
axis of rotation, and in others one or both axes are inclined from the horizontal.
Convenient as these modifications may be in special cases, no real advantage accrues
or is likely to accrue from any of these alterations in regard to the mixing operation.
To place the ingredients to be mixed into the drum through an opening in its cir¬
cular side, for which purpose it is moved to the top, is easy, and to remove the ready
mixture through the same opening, when moved to the bottom, is equally so, in regard
to comparatively small apparatus, but this arrangement becomes less satisfactory
and more difficult with the increasing size of the drum. Several openings placed side
by side have been sometimes used, but the multiplied removal and accurate replace¬
ment of their covers for each mixing renders the charging and discharging operation
35
very tedious. In order to avoid this interruption of work many devices have been
designed for charging and emptying the mixing drum, while revolving, from its flat
end walls, one or both being provided with openings, and the drum being rotated by
the toothed-ring gearing mentioned above. For instance, for charging the apparatus
a funnel-like hopper is placed at its side, its bent neck reaching through the center
opening in one of the flat sides, into the interior of the rotating drum; for removing the
finished mixture, an inclined chute or trough suitably supported outside the drum and
wide open at its top is pushed with its closed end through the center opening in
the opposite flat side into the interior of the drum nearly through its whole length.
Scoop-shaped lifting plates serve to drop the material raised thereby, when the drum
rotates, into the inclined chute, from the open end of which outside the drum the mass
slides into the receptacle intended to collect it. In another arrangement the drum is
charged in the same way through the center opening in one sidewall, while it is emptied
through the completely open opposite end which is closed by means of a suitable cover
during the mixing operation; for quickly removing the finished mixture the interior
of the drum is provided with inclined plates or spiral rings or blades so arranged as to
act like a screw conveyer and thus drive the material towards and out of the uncovered
end.
The drum mixer is used on a large scale for mixing grain, flour, seeds, fertilizers,
ores, etc., etc. The mixing cylinder forms the rotary body of the modern continuous
driers for moist grain, seeds, sand, clay, chalk, graphite, roots, pyrites, pigments,
chemicals and similar materials which, by being continually intermingled, are
uniformly dried. A rotary drum, when closed and heated from the outside, serves for
mixing and evenly roasting coffee, cacao beans, chicory, etc., and, when provided with
openings in its flat sides and fired, for roasting ores, minerals, etc., and to fuse the
ingredients in the Le Blanc soda process, for cements, etc., in all cases equalizing the
temperature of and reaction in the roasting or melting mixture.
The tumble-box mixer consists of a long receptacle, containing the substances
to be mixed, carried on trunnions fixed to its long sides. The receptacle may be
round, square or polygonal in cross-section; its ends are flat, preferably in a
plane vertical to its long axis. One end is closed, while the other end is open, but
provided with a tightly closing lid. For efficient work the box must not be filled
above one-sixth to one-fifth of its length. The box is charged when upright;
the lid is then tightly put on, and the box slowly revolved, making about 12 to
25 r.p.m. When the end containing the materials to be mixed moves upwards,
they slide gradually down along that side wall which at the moment is in the lower
position, to the opposite end, and are thus turned over, the upper layer be¬
coming the lower one, while, as their movement is slanting, both layers will also
more or less intermingle. The mixture is fully turned over twice during each
full rotation, and, when finished, discharged by opening the cover and tilting
the box. This tumble box mixes small quantities, say 30 to 40 lb., of dry colors,
chemicals, drugs, powders, etc., very well, but I have never seen it used on a large
commercial scale.
Continuous mixing machines are sometimes used when very large quantities
of powdery products, for instance different kinds of cereal flour, the materials
for self-raising flour, baking, face and tooth powders, etc., are to be mixed in a
short time at lowest possible costs.
The mixing wheel, shown in Fig. 23, is an old device used in European mills.
It is arranged underneath an opening in the ceiling of the chamber in which the
flour mixture is collected. A indicates the chamber, a its ceiling, a\ the opening
JI/7A7A7/ AND K NBA 1)1 NO 547
therein, B the hopper, the neck l> of which passes through the opening ah C the
mixing wheel and D the driving gear ol (\ The wheel C consists of a stationary,
circular upper half c provided in the confer with an opening fitting over and fixed
to the neck b of the hopper B, and the rotatable, solid lower half fixed to the
shaft d of the driving gem* D; r and ct are provided with concentric rows of pegs
c2 alternately so arranged that those fixed to flic lower side of r and those fixed to
the upper side of c-t pass freely through between each other, when rx is revolved;
ci is carried by the brackets c;!, and the beam ci, supporting the lx‘a rings of the
vertical shaft d, by the brackets cf); c;, and are fixed to the ceiling a. The shaft
d is rotated by the gear wheels </, and is some!imes provided with stirrers
The opening of the neck h of the hopper B is usually regulated by shutfers bx.
Km. 'Jit. M ixing \\hiM*t,
The flour streams coming from flu* ditlerenf mills are led into the* hopper Bt
drop then through b into between the parts v and vx of f lie wheel C and are mixed,
while resting on the rotating lower half, by means of the pegs r*.*, whereupon the
mixture is projected by centrifugal force over fhe edge' of r{ info flic chamber
A. A homogeneous mixture can of course nol he produced in this manner;
but millers, who mixed by means of such a wheel, and bakers, who used the flour
thus mixed, averred that the* mixture produced in this manner was quite satis¬
factory for their purpose, which after all is the* main thing.
The blades of more modern continuous mixers are generally based on fhe principle
of the screw conveyor, but, as already mentioned on page* fuII, even in these accuracy
of composition and perfect mixing cannot be continuously obtained. The only con¬
tinuous mixing apparatus giving full satisfaction in these respeeta is that in
which two or more batehmixers are alternately charged by the same set of auto¬
matic weighing hoppers, and alternately and automatically discharged into the same
receptacle, as soon as the mixing is perfect.
Mixing Liquids with Liquids. In distinction from the stable cohesion of Urn
constituent molecules of solids, those of liquids, though still adhering to each
ot er, move freely among themselves; thus they easily separate and then combine
again. When two liquids are brought together in a vessel, they either mix, that
is they spread uniformly within each other, forming a completely homogeneous
mass, a solution, as for instance alcohol and water, or they do not mix, but
form separate layers, as water and oil. Two liquids are miscible with each other,
when the attraction of the molecules of one liquid to those of the other one is
greater than the attraction of the molecules of each single liquid to those of its
own kind, and they are non-miscible, when the opposite conditions prevail.
When two liquids of different specific gravity are carefully poured one after the
other into a vessel, they will stand in separate layers, the heavier one at the bottom
and the lighter one on the top of the heavier one. If these two liquids are miscible,
the molecules contained in the two layers will immediately begin to penetrate into
the other layer and continue to do so, until the two layers have completely dis¬
appeared, and a homogeneous solution is formed; the duration of this gradual inter¬
mingling—the diffusion—of the two liquids is the shorter, the smaller the difference in
their specific gravity and the greater the mutual attraction of their different molecules
as compared with the separate attractions. Violent agitation by means of a simple rod
or paddle will greatly accelerate the perfect mixture of the two liquids. In case the
two liquids having a different specific gravity are non-miscible, they will form two
clearly defined layers meeting in an even plane; very rapid and violent agitation will
convert them into a mixture in which minute particles in the shape of globules of the
one liquid are suspended in the bulk of the other one. The production of mixtures of
this class—emulsions—will be discussed later.
When two liquids—miscible or non-miscible with each other—are brought together
for the sake of their chemical reaction, the more violently acting agent is usually
poured very gradually into the liquid acted upon; by diligently stirring the mixing
mass during the process the reaction is equalized throughout the whole mass, its
intermediate phases become more clearly distinguishable, and the whole operation is
accelerated and made more effective in every way. In the laboratory we mix liquids
by means of the stirring rod or spatula, if contained in open vessels like a beaker or
dish, and by violent shaking, if contained in a closed flask, bottle or similar container.
On a commercial scale usually the first method—stirring—is employed; the agita¬
tor serving this purpose consists in general of a vertical rotatable shaft arranged
coaxially to the tank, kettle, drum, pan, montejus, cask, etc., containing the liquids
to be mixed, and of the actual stirring devices fixed to the shaft, as: horizontal or
inclined rods—square or round, straight or bent—vertical or inclined, plain, curved
or angular vanes, propellers, etc. For discharging the ready mixture the containers
are usually provided with an outlet at their bottom or the lower part of their sidewalls;
but, when this arrangement is not convenient, and the vessels are to be emptied from
the top by tilting or are removable from the machine, suitable lifting gear is attached
to the framework carrying the driving gear of the agitator in order to raise the same
out of the interior above the highest position of the mixing vessel.
When liquids are heated, which is done either by direct fire underneath or by steam
led through jackets surrounding or tubes arranged within their containers, the heat
is transmitted through the walls of the latter or the heating tubes therein to the liquids
—with rare exceptions—at a much quicker rate than it is distributed within the latter
by conduction or the currents caused by the change of their specific gravity or their
quiet evaporation. Boiling and melting kettles, evaporating pans, vacuum pans,
digesters, autoclaves, etc., are therefore also provided with agitators of the kind men¬
tioned above. In order to prevent components of the heated liquid from adhering
to the walls or even being scorched, these stirrers have branches—and in more severe
eases actual scrapers—following the outline of or moving closely to the heating sur¬
face of the vessels.
For accelerating the mixing action it is sometimes advantageous to supply the

driving gear with a reversing device so that it is possible to revolve the blades in
opposite directions at will. Denser liquids naturally not mixing so easily as more
fluent ones require accordingly a more powerful mixing device. The simplest way
to attain this purpose is to increase the number, size or whirling power of the blades,
arms or branches of the agitator. An effective construction is the double-motion
mixer which consists of two agitators rotating round the same axis, but in opposite
direction to each other. The inner shaft of this device is solid and carries the inner
blades or branches moving in the central part of the tank, while the outer shaft con¬
sists of two tubes rotatably mounted one at the lower and the other at the upper part
of the inner solid shaft and connected by the outer arms or scrapers moving along
the walls of the container. The upper tube of the outer agitator rotates in bearings
arranged in the lower part, and the upper end of the inner solid agitator protruding
from the former in bearings arranged in the upper part of the framework carrying the
driving gear of the apparatus. Between these bearings two pinions facing each
other are fixed, one to the top .of the outer agitator and the other to the inner one;
the driving gear wheel meshing with both pinions is arranged between them, thus
Fig. 25.—Simple agitator with direct fire.
rotating both agitators at the same time in opposite directions. The agitator must
be so constructed, and its speed so regulated that no excessive splashing of the liquid
can take place to the detriment of the operation, besides causing loss of the materials.
The great movability of the molecules of liquids is the reason that the mixing
vessel containing them may have any size desired, controlled only by constructional
possibilities and convenience. The design of the agitator and its blades and the
velocity of its rotation and the power required for the same depend mainly on the
consistency and specific gravity of the liquids to be mixed. When starting an agitator
in a stagnant liquid mass, the resistance to be overcome is considerable, but, when
once put into motion, the power necessary to revolve the agitator is very small indeed.
The various machines described in Figs. 1 to 21 may be used for the mixing of
liquids, but, instead of the trough with horizontal agitator axis, usually and preferably
upright tanks of a circular cross-section are employed, the shaft of their agitator
being arranged in their vertical axis. A very forceful mixing device for liquids is
the double spiral agitator,1 shown in Fig. 24. d and di are parallel cross rods fixed
vertically to the shaft a, and passing, in case a tank forms the mixing vessel, one along
its flat bottom and the other one in its upper, open part, while, if a trough is employed,
xMade by J. H. Day Co., Cincinnati, Ohio—Patented.
both move along its flat ends like the sidearms in Fig. 6. The ends of the spirals 6,
6i and c are mounted on the cross rods d and dh while they are supported in the
interior by the cross rods e. The theory is, that, while the outer spirals cause a current
in the mass to be mixed in one direction, the inner one does so in the opposite direction,
provided that the agitator is revolved with adequate velocity, effecting a criss-cross
whirling motion and thorough intermingling of the particles. The speed required
for obtaining the best results in working this agitator is rather great, 50 to 80 r.p.m.
The double-spiral agitator is also useful for the mixing of sifted, soft, light and very
Fig. 26.—Vacuum pan. Fig. 27.—Double-motion agitator.
fine powders, as flour, face and toilet powders, etc., the particles of which flow like a
fluid readily between each other when violently agitated.
Figure 25 represents in sectional elevation a simple and effective agitator frequently
used in chemical factories in boiling kettles heated by direct fire. The agitator
rotated at 15 to 25 r.p.m. mixes very well and keeps the bottom clean from dried up
and scorched material.
The vacuum pan, Fig. 26,1 is provided with a very effective sweep agitator
which greatly accelerates and equalizes the evaporation, a is the pan, b its
jacketed bottom for heating by steam, coils c adding to the heating surface, d
the outlet, e the shaft of the agitator, / its driving gear carried by the bracket gy
h a cross piece of the latter supporting the upper end of the agitator protruding
from a, i the bridge guiding the lower end of the latter, and k its sweep-arms pass¬
ing closely to the coils c, while l indicates the inclined paddles forming with k
the active part of the mixing device. If the sweep agitator is installed in a boiling
pan or kettle having inside no heating coils or other impediments, its outside
arms are usually provided with special scrapers following in outline the inner
surface of the vessel, and either arranged with intervals on the opposite arms of
1 Dopp apparatus, manufactured by Sowers Manufacturing Co., Buffalo, N. Y.
the sweep in staggered position and consisting of thin metal plates provided
with sharp, knife-like front ends, or running along the whole length of the arms
and formed of long, curved, knife-like narrow plates or more or less elastic rolls,
pads, brushes or the like of metal wire or straps pressing against the interior wall
of the vessel. It is very advantageous to provide the scrapers with devices
for regulating their distance from the surface which they are intended to keep
clean; but the utmost care must be taken that their construction is such that no
part can loosen or unscrew while the machine is at work.
Figure 27 represents a double-motion agitator as mentioned above, the apparatus1
being intended for the mixing of thinner liquids. The tank is provided with the out¬
let at its bottom and at its top with the
bracket c supporting the shaft of the
driving gear for the double-motion
agitator. The latter consists of the
shaft e on which the paddles/,/i are
mounted, and the sleeves g and gi which
carry the lateral arms h and hi connected
at their ends by the vertical sidepieces
or scrapers i which are strengthened by
the crosspiece k passing loosely round
the shaft e. The long upper sleeve g
protruding through a stuffing box inside
c carries at its top the gearwheel l,
arranged so that the weight of the
sleeves and the sweep is sustained Fig. 28.—Mixer for heavy liquids.
thereby. The end of the shaft e projects
through the stuffing box in the top of the bracket c and is held in position by
the nut m, while its lower end loosely revolves in a guide bracket. The shaft e
Fig. 29. Fig. 30.
has affixed thereto within the bracket c the gearwheel o, facing the gearwheel
l; the pinion p mounted on the driving shaft is arranged between o and Z, meshing
with both, and drives thus the shaft with paddles /, fi in one, and the sleeves g, gi
with the sweep consisting of the arms h, hi, the scraper i, and the crosspiece k in the
opposite direction.
1 Made by the Pfaudler Co., Rochester, N. Y,
A most powerful mixer for the heaviest liquid masses, while being heated, is shown
in Fig. 28;1 the kettle a is steam-jacketed and open at the top, the double-motion
agitator and driving gear being carried by the bracket extending across the top; the
agitator is constructed like that in Fig. 27, with the exception, that the inner shaft
has three paddles instead of two, that the side pieces are strengthened by two cross¬
pieces instead of one and provided with a special scraper at the bottom and smaller
staggered scraper plates at the sidewall and that the diameter of the gearwheel
for the sweep being double that of the gearwheel for the inner paddles, the velocity
of the revolution of the latter is double as great as that of the former.
For rapidly intermingling easily miscible liquids and for keeping them in
more or less gentle motion to prevent them from becoming stagnant, the propeller¬
shaped agitator is now generally used. This device—if rotated at a sufficiently high
speed—creates an intense whirling motion which spreads very widely in a liquid mass; •
it is very simple of construction, easily cleaned and takes up a very small space. A
small kettle of the type shown in Fig. 27 fitted with a propeller was used to mix malt
extract and other liquid foods, the mass being heated to and kept at 60°C. by direct
steam. The agitator made 80 r.p.m., and did its work very satisfactorily. Figure 29,
is an agitator shaft2 provided with several propeller blades which, when rotated at
up to 150 r.p.m. is very effective. Small agitators1 of the propeller type can also be
arranged horizontally in the sidewall of a comparatively large vessel used as butter¬
milk machine, ice-cream mixer, starter can, etc., in dairies. The thrust of the propeller
is directed across the tank and produces a thorough circulation at a speed of about
300 r.p.m.; only very little power is required.
In cases where it is desired to give an upward tendency to the circulation of the
liquids in the center of the mass, a propeller of the marine type is used and surrounded
as shown in Fig. 30, by an open cylinder or drum3 in the lower part of which the
former is situated. The agitator is especially adapted for rapid mixing of fight liquids
and requires for thorough work a speed of 25 to 50 r.p.m. For very thick liquids
instead of the propeller a screw conveyor rotating within the drum is employed with
advantage; in case the mass is to be heated at the same time, the drum is constructed
as a radiator forming a hollow shell or a cluster of steam tubes as for instance in the
crutcher for melting and mixing soap masses. Figure 314 represents a very powerful
apparatus of this class in which conveyor screw and drum are rotated in opposite
direction, a is the steam jacketed kettle; b is the conveyor screw fixed to the center
shaft, which rotates in the sleeves d and di carrying the drum e, sweeps / and inclined
cross pieces g stretching from e to /; screw b rotating in one direction and the drum e
with the sweeps / in the opposite direction.
Liquids which easily combine, as for instance hot and cold water, liquids plain and
containing chemicals dissolved, are frequently mixed in a continuous manner by
leading them in the required proportion through separate pipes into a wider common
tube or box in which they encounter cross bars or baffle plates often arranged in a
screw-like fashion.
Mixing Liquids with Solids.—An absolute homogeneous mixture of liquid
and solid substances can only be obtained if the latter are soluble in the former,
even then in most cases after standing we find the strongest and thus the heaviest
solution at the bottom, getting gradually weaker toward the surface. Water-
soluble chemicals, which are not affected by heat, are generally dissolved by
introducing into the water direct steam, which, entering with great velocity,
1 Made by Sowers Manufacturing Co.
2 These agitators are made of any metal or glass-enamelled by the Pfaudler Co., Rochester, N. Y.
8 Made by the Pfaudler Co., Rochester, N. Y.
< Dopp apparatus, manufactured by Sowers Manufacturing Co., Buffalo, N. Y.
causes a rapid movement in the mass, while heating it at the same time by being
condensed, thus greatly accelerating the dissolution. If more valuable liquids,
as alcohol, ether, acetone, chloroform, carbon disulphide, etc., are used as solvents
the operation is performed in a closed pan or kettle provided with an agitator,
and, in case heat is to be used for assisting in the operation, with a steam jacket or
steam coils and with a reflux condenser.
In cases where heat may cause damage to the product or is to be avoided for prac¬
tical or economical reasons, the vessel, in which the solids are to be dissolved, is
provided with an agitator adapted to keep the undissolved particles, prone to settle
at the bottom in continuous motion. An apparatus1 is especially designed by Werner
and Pfleiderer for this kind of work, in which the effect of the agitator, consisting of a
propeller or tangential wheel, is greatly enhanced by the peculiar egglike shape of the
mixing tank which in its lower part forms a hemisphere and in its upper part a trun¬
cated cone. The propeller revolving at a high speed forces the liquid and the solids
Fig. 31.—Double-motion kettle.
suspended therein with great violence from its center sidewise towards the spherical
wall, thus driving both upwards against the inclined wall where the current is broken,
(if needed, by baffle-bars), and directed towards the axis of the agitator, and then
draws the mass in again from this central part for repetition of the cycle, until the
operation is finished.
Solids which are to be intimately mixed with a liquid in which they are insoluble*,
must be in a finely pulverized state. In case that the adhesion of the molecules of tins
liquid to a solid, otherwise unaffected by the same, is greater than the cohesion of the
liquid molecules to each other, the solid is wetted by the liquid, and both are then easily
mixed. When a small quantity of such solids is to be suspended in a comparatively
large quantity of liquid, the simple mixers shown in Figs. 24, 26, 27 and 30 arc used
with advantage, the agitator blades being arranged so that their surfaces point in the
direction of the rotation, downwards when the solid particles are heavier, and upwards
when they are lighter than the liquid. 9
If only a small quantity of liquid is mixed with solid matter, creams, pastes or
pulps are formed of a stiffness and consistency corresponding to the proportion of the
latter to the former. The machines Figs. 4, 6, 7, 18, 19, 21, 26, 28 and 30 and similar
1 Made by the Balcer-Perkins Manufacturing Corp., Saginaw, Mich.
mixers usually serve this purpose very well; for making a fine mortar, sanitary pastes
or the like, and mixing paints, an edge mill is often used, in which a pair of stone or
metal rollers of large diameter, loosely revolving on a horizontal shaft turned round the
vertical axis of the machine, press, squeeze, flatten and, by thus crushing and dis¬
intergrating the smallest lumps and particles of solid matter, thoroughly mix the mass
which is contained in an annular pan arranged in the circular path of the rollers.
The pan may sometimes also be rotated opposite to the direction of the latter round
the same vertical shaft. In regard to pastes, etc., it may be said in general that a
certain composition and quantity of solid matters can only retain—without any
separation after standing for some time—a distinct maximum of a liquid, the amount
being mainly dependent on the fineness of the solid particles, the size of their inter¬
stices and their capillarity.
It is always advisable to charge powdery matters gradually into a liquid while the
agitator is working, and only to add a fresh portion, after the previous one has been
distributed therein; if the quantity of the liquid is large, and the solid particles are
inclined to combine, when wetted, to lumps and knots, the whole mass of the solid
matter is first mixed with a small portion of the liquid to form a thin paste which is
then gradually added to the rest of the liquid whirling round in the machine.
In cases where the solid materials are not wetted by the liquids with which they are
to be mixed, more energetic means must be employed, and the operation must be
continued for a longer time in order to convert them into a smooth mixture which
is as homogeneous as possible according to the nature of the ingredients. For inti¬
mately mixing and blending oils and fatty substances, varnishes and the like with
powders of any kind, with watery solutions of salts, acids and other chemicals, and
with similar materials in the manufacture of butter, margarine, cosmetics, sanitary
and color creams and pastes, rubber imitations, scouring soaps, shoe, stove and
metal polishes, printing inks, lead-pencil masses, ointments, salves, chocolates,
candy, chewing gums, etc., the two-blade mixer Fig. 19 and the double-motion agi¬
tator have proved very useful machines. Small quantities of these products are
frequently prepared in suitable containers which are changeable; after having been
charged, the single receptacle is placed on a rotating platform, and an energetic
agitator lowered into the mass to be mixed by means of a mechanism arranged at
the support of the platform and driving gear of the machine.
Pharmacists often prefer to prepare their ointments, creams, salves and pastes in a
machine mortar or in a small stone mill especially equipped for this purpose. It is
claimed by some experts, that such preparations, when mixed in that manner, are
smoother, softer and of a more homogeneous nature than those made in mixing
machines, because the molecules of the substances to be mixed—so different in their
nature from and even repelling each other—are finely ground together under pressure,
also the smallest knots and unevennesses being completely smoothed out, while in
the mixing machine only a forcible intermingling of the particles could take place.
For the same reason, the finer oil colors for coaches, furniture, art paintings, printing
inks, pastel colors, and similar products are usually made in stone mills, the under
runner of which makes about 100 r.p.m.; the feeding hopper is frequently constructed
as a preparatory mixer, being fitted with a suitable agitator; for rather stiff pastes
and colors it is sometimes advantageous to make arrangements to feed the mass to
be mixed into the mill under pressure. For the products described, but especially
for printing inks, often roller mills, usually having three water-cooled rollers of iron or
granite, are also employed. The mixture for chocolate mass, consisting of cacao
powder, fat, sugar and condiments, is generally produced in an edge mill the stationary
pan of which is heated by steam or hot water.
Much attention has been given to the development of concrete mixers. In the
manufacture of artificial stones, concrete building blocks, drain pipes etc., in which
fine sand, trass, gypsum, pigments and oilier filling materials are incorporated with
cement and water, it is—in order to obtain a, good result and to prevent the formation
of hard-lumps in and the premature hardening of the mass—the best plan first to
thoroughly mix and intermingle a,11 the solid ingredients while in the dry state
which may be done in any machine described above for pulverized materials, and then
quickly to introduce the water in shape of a strong spray covering the whole exposed
surface of the mixture while still being agitated. For mixing coarse sand, gravel,
pebbles, stone clappings, etc., with cement and water for producing concrete for roads,
dams, water reservoirs, etc., the most, frequently used machine is a rapidly revolving,
inclined drum mixer provided with means for quickly charging and discharging a
batch; in order to bring the concrete in the best possible state for hardening on its
bed, each batch must be comparatively small; it must be quickly handled, and the
operation rapidly repeated.
In the manufacture of explosives, smokeless powder, celluloid, etc., the two-blade
mixer, Fig. 20, in which each shaft carries t.wo symmetrically arranged, elliptical blades
has been prominently used for a considerable time; its advantage is that the peculiar
construction of these blades and the low speed at which they arc worked, contribute
considerably to the safety of the operation, as for instance mixing different kinds of
guncotton in a moist state or blending them dry with a solvent. The machine is
usually provided with a jacket for cold and hot water or steam, and, in case that, the
contents arc to he dried at. the necessary low temperature and the solvent is to he
regained at the same time, the top of the machine is closed by a tight lid connected to
a condenser and vacuum pump, the operation thus being performed in vacuo.
Mixing and Kneading Dough for Bread.—Ordinary dough is prepared by
fermentation from flour, salt, yeast, and water as main ingredients to which, in
order to stimulate the activity of the yeast and to improve the taste, flavor,
composition and appearance of the finished article, varying smaller quantities
of milk, dried milk, fats, sugar, malt-flour, malt-extract, cornflakes, yeast¬
feeding salts, spice, etc., are added. The quantity of water used for the dough
is between 50 and 05 per cent of the flour contained therein. The soluble
ingredients are usually dissolved in a small portion of the water, and the yeast
finely distributed in another one; all the components of the dough are then
combined and thoroughly mixed and kneaded, that is pounded, pulled, pushed,
drawn out, stretched and folded over again until the resulting dough has the
required consistency, elasticity, and homogeneity.
The dough is then put aside in a warm room for fermentation and, after having
risen, scaled off and molded into loaves, which, after again undergoing a short time
of fermentation, are baked. Wheat flour produces a white, very light, spongy, palat¬
able and easily digested bread which for these reasons is preferred and more generally
used in civilized countries. These good qualities are due to the wheat flour contain¬
ing a considerable quantity (0 to 12 per cent) of gluten in a quality characteristic of
this cereal. When dough is made of this kind of flour, the water dissolves its soluble
components, penetrates through and softens the starch grains and will be absorbed
by the gluten in a quantity equal to about double the weight of the latter. The
gluten is thereby converted into a soft, swollen and yet firm and cohesive moss of
very considerable elasticity. By kneading the dough the gluten is drawn out to
strands and then widened to bands, which, crossing the dough in all directions, will
finally form, when combined at their edges, a complete sponge-like network contain¬
ing all the other ingredients of the dough and possessed of such a strength and
resiliency as to retain therein the carbonic-acid gas developed by the subsequent
fermentation of the sugary components. In a perfectly, an ideally kneaded dough
these cavities will be of equal size, have walls of equal thickness and resiliency and
contain equal amounts of the other ingredients; it is obvious that, if only to approach
this perfection, the kneading must consist of drawing, stretching, pulling and sidewise
combining only, and every tearing and separating action must be avoided; in short,
it is necessary to knead the dough in bulk.
Some 50 years ago nearly all bread dough was mixed and kneaded by hand, and
the bread baked on the hearth; only in the bakeries for large garrisons, navy yards,
hospitals, etc., kneading machines were employed; they were usually more beaters
than mixers and required a very considerable amount of power for working the stiff
doughs necessary for hearth bread. The most successful kneader then used had an
agitator consisting of two four- or six-armed side pieces arranged at the shaft so as to
rotate closely to the flat ends of the half-cylindrical trough and connected at their
ends by diagonal rods, two crossing from each arm of a sidepiece, one to the next
arm at the left side and the other one to the next arm at the right side of the opposite
sidepiece. In consequence of this arrangement the rods could only approach closely
to the circular wall of the trough at their joints with the side arms. An agitator
Fig. 32.—Double mixer.
quite similar to that described is shown in the dough mixer, Fig. 32,1 which is espe¬
cially recommended for the development of gluten in heavy doughs. In those times
the bakeries were small shops, too small for a steam-power plant; in the seventies of
last century the small gas engine of 2 to 10 hp., so useful for numerous small establish¬
ments, came on the market and contributed greatly to the quick introduction in
Europe of the two elliptic-blade mixing and kneading machine, Fig. 19, which—con¬
trary to the machines already known—required comparatively little power. This
machine is—as has already been stated above—a quick and excellent mixer; it pro¬
duces a very good dough for hearth bread in a very short time, to 5 min., and at
a low speed, 7 to 10 and 12 to 20 r.p.m. respectively for the two agitators, the actual
power used in a two sack machine (for 560 lb. of flour) being at the most 5 to 6 hp.
But the dough maker must be an experienced man, capable of judging, when the
dough has reached its proper consistency for a good, well-risen loaf and stop the
machine at the right moment, the danger being that after longer kneading it would
become soft and putty-like and yield only a comparatively small and rodden loaf.
The reason for this injury to the dough by overkneading it in this class of machine
is that not only small bits are tom from the already finished dough when pressed
along the circular surfaces of the double trough by the blades rotating closely along
it, but also a considerable portion is cut away on the saddle while being passed over
1 Made by J. H. Day Co., Cincinnati, Ohio.

from one-half to the other one, twice at each revolution of the blades; the fully
extended gluten strands thus after some time become completely disintegrated and
give the dough the appearance as if a weak flour had been used. This explanation
is confirmed by the fact that, when a thus over-kneaded dough stands for some time,
it will partly regain its strength, apparently in the measure as the torn gluten ends
are able to join again, due to expansion by fermentation.
For hearth bread, only a limited quantity of water can be kneaded into the flour,
49 to 55 per cent according to its strength, because, if more were used, the dough,
though perhaps showing a good shape after molding, would flatten out in the oven
like a pancake; also any appreciable addition of fat (shortening) or sugary matter
must be avoided, because, if used in larger amounts, they naturally diminish the
strength of the dough. Doughs containing a large proportion of water (say 60 per
Figs. 33-34.—Corby dough-mixer.
cent), or of the last ingredients mentioned (say 2 per cent) are placed into pans the
sides of which, upholding the dough mass during baking, not only prevent it from
running, but cause it to creep upwards and thus produce a very light loaf. These
light doughs require, as compared to the heavier ones, in the two-blade machine,
Fig. 19, a higher speed, 10 to 18 and 20 to 36 r.p.m. of the two blades and a longer
time of kneading, 10 to 20 min., and the result is not so good as in a modem high¬
speed kneading machine which is now mostly used in the United States where the
bakers, having the best wheat flour in the world at their disposal, are bent on pro¬
ducing for us also the very best bread by adding a remarkable amount of fat and
sugary substances (about 2% per cent of each), milk powders, etc., and a correspond¬
ingly larger quantity of water.
According to the Corbys,1 the inventors of this method of dough making, their
kneader consists essentially of a receptacle, in which a beater comprising a pair of
spiders and a series of longitudinal bars extending between the latter revolves at a high
speed. Figure 33 represents a top view and Fig. 34 a cross-section of the original
design: a is the kneading trough having a half-cylindrical bottom; b and &i are the
spiders fixed to the trunnions c and Cj provided outside a with the gear wheels d and di
for driving the agitator. The four arms of the spiders b and bi are of different length,
carry a different number of beater bars e and are arranged in line at the trunnions c and
ci so that the bars e extending between them are parallel to the axis of rotation. It will
be noted that also the bars e nearest to the circular wall of a do not approach it closely.
The driving gear is so arranged that the beater is driven at a high speed which means
for this class of machine 45 to 80 r.p.m.
A short time after the materials have been placed in the trough and the beater
is running at a high speed, a thin dough-like mixture is formed, having tlitle tenacity
and a rough appearance; then the rapidly rotating beater bars separate small portions
1 W. H. Corby and Charles J. Corby, U. S. Pat. 029,239, July 18, 1899, of Process of Dough¬
making; 639,889, Dec. 26, 1899, for Machine used for this process.
from the main mass by centrifugal force and throw them about; with the increased
tenacity caused by the quick development of the glutenous condition of the dough,
these separate portions adhere to each other and to the main part more and more,
until the whole mass extends in strands and sheets from bar to bar of the beater; the
centrifugal force will drive all these sheets and strands always farther away from the
axis of rotation, so that almost the whole bulk of the dough is worked by the outer bars
and against the wall of the trough which tends to retard its passage. The dough will
thus be drawn out to thinner or thicker sheets between the bars and the wall according
to their smaller or greater distance from each other. Free from the wall in the upper
part of the trough, the sheets are folded up and then caught again between the bars
and wall to undergo the same treatment which is repeated until the dough is finished,
forming a perfectly smooth, creamlike, homogeneous and very tenacious mass which is
the foundation of a well fermented dough and a very light1 bread baked therefrom.
The operation is completed in 20 to 30 min. The improvements made in this class of
Figs. 35-36.—High-speed mixer.
kneader refer mainly to the strengthening of the beaters, the trough, gearing and frame
work supporting the machine, as they are naturally under immense stress and pressure,
especially during the later part of the operation.
Figure 35 shows the longitudinal and Fig. 36 the cross-section of the trough A and
dasher B of a modern high-speed kneader,2 in which the spiders b, hi of B consist each
of two pairs of arms 62 and 63 arranged vertically to each other, and are rigidly mounted
on the central driving shaft c passing through the whole length of the trough. The
longer arms b2 reach closely into the corners of the half-cylindrical bottom and the sides
of a and carry at their ends the strong beater rods d bent in a curve towards the center
of the driving shaft c, while the similarly curved beater rods di are fixed to the ends of
the shorter arms bz. The inwardly curved beater rods d and di cause the dough mass
to be drawn towards the middle part of the machine and thus greatly facilitate and
increase the effect of the working of the dough in bulk.
In another prominent machine of this kind, the agitator and trough of which are
shown in Fig. 373 two beaters a and are arranged one within the other, the beater
a being mounted on trunnions b and bx revolubly fitted in tube-like trunnions c and Ci
carrying the outer beater ai. To the end of the trunnion b protruding from the tubular
trunnion c is affixed the spurwheel d, and to the tubular trunnion cx on the opposite
side of the machine the spurwheel di which is somewhat larger in diameter than d.
The spurwheels d and di mesh with pinions e and ei fixed on the driving shaft / of the
machine so that the beaters a and ai are rotated in the same direction, but the inner
one at a higher speed than the outer one. The outer beater ai consists of two equal,
symmetrically arranged parts of the peculiar shape shown in the illustration, the inner
beater a is formed of two curved blades corresponding to the shape of the inner cir-
1 For illustrating this point I may state that I found by repeated careful experiments that the
proportion of the size of a good English Cottage loaf (hearth bread) to the volume of an equal weight of
water is =’ 2% :1, while the same proportion for a good class American loaf is = 4:1.
2 W. J. Tbevis's, U. S. Pat. 1,041,908, Oct. 22, 1912. Bakers-Perkins Mfg. Corp., Saginaw, Mich.
* F. Notz and J. McBriar's U. S. Pat. 831,372; Sept. 18, 1900. Manufactured by American Oven
& Machine Co , Chicago. Ill.
cumference of the curved part of «i, sufficient space being left between the two for
efficiently working the dough.
The proportion of the revolutions per minute of the inner to those of the outer
beater is about 3:2; for the slack pan-bread-dough of the large bread factories 60:40
r.p.m. is recommended. The peculiar construction and working of this agitator tends
to make the machine a somewhat better mixer; but its main point consists therein
that the stretching of the developing glutenous dough docs not take place between the
circular wall of the trough bottom and the outer blades, but in consequence of the more
rapid rotation of the inner beater chiefly between the latter and the outer one, a cir¬
cumstance which appears to be of decided advantage.
The working capacity of these high-speed kneaders is naturally much less than that
of low-speed machines; for in general, the higher the speed, the smaller is the working
capacity of the machine for the same mixture and result. The violent action on the
dough of beaters revolving at a high speed will raise its temperature, in hot weather
even to an injurious degree; some constructors seek to prevent this occurrence by
introducing cold air into the dough and the interior of the kneading trough, and others
by passing cold water through a jacket surrounding the latter, while bakers sometimes
use the rough but simple remedy of dropping a few pieces of ice into the mixing mass.
The time necessary for obtaining a perfect dough in high-speed kneaders is long, 15 to
20 min. and even more, and the amount of power required very great, 20 to 25 hp.
on the average for a machine mixing 1,000 to 1,200 lb. of dough at 50 to 60 r.p.m. of
the agitator.
In Europe where a low price of ordinary bread always was a paramount economical
necessity, the costs of producing the dough forms naturally an important factor; the
endeavor of the designers of kneading machines was therefore to devise a construction
which requires less power even than the two blade kneader, Fig. 19, and in which
the operation can be performed without the supervision of an experienced dough
maker. This problem seems to have been solved by the reciprocating arm-kneader of
which Fig. 38 shows an example. In this machine the dough is worked in the circular
rotating pan a by means of the reciprocating arm b ending in the crosspiece c which
consists of two lateral hook-like appendages as shown in Fig. 39.
The action of the arm b and the crossbar c the course of which is shown in Fig. 38
in dotted lines, is a remarkable imitation of the old-fashioned kneading by hand, a
fact which probably aided greatly in the introduction of the machine. The driving
gear d is supported by the strong frame-work e arranged at the back of the pan a
which, provided with the toothed ring /, is rotated by gearing g carried by a bracket
fixed to e. Being removable, the pan a is provided with wheels, and, when discon¬
nected from the machine, serves as a container for the dough during its fermentation,
thus saving the emptying of the kneading trough necessary in other machines. When
kneading is to be done, the pan is pushed, after the arm b has been raised, on to the
bed-plate h and into the hinged locking device % which holds it in its proper position
in regard to the driving gear. The arm b draws the dough out very gently, making
it loose, and works only on that portion which it reaches on its downward stroke,
leaving the rest for a short time undisturbed so as to allow the gluten of the flour to
contract and to connect, which is considered to contribute greatly to its development.
As the arm b does not throw or splash the dough about, the working capacity of
the machine is very great indeed, while the flat bottom of the pan and the peculiar
arrangement of the kneading arm allows also a very small amount to be efficiently
treated; thus the working capacity of a pan of 560 liters cubic contents would range
between 50 to 500 kg. of dough.
The kneading arm of this machine makes 24 to 30 strokes per minute, while the
pan does 4 to 6 revolutions in the same time. As only a portion of the dough is
kneaded at each stroke, the power required is naturally very small, even considerably
less than that for the two-blade mixer, Fig. 19; for a size suitable to work a batch of
about 1,200 lb. of dough 3 to 5 hp. are quite sufficient. Kneading a batch of dough
in this machine takes about 10 min., and, as the action of the agitator is very gentle,
the elasticity and tenacity of the mixture do not suffer by over kneading. In another
kind of reciprocating arm-kneader which is much used in England, only the inclined,
separate bottom of the pan is rotatable, while the upper part is stationary, the ready
dough being discharged through a door provided in the latter.
Very large kneaders intended to work 20 to 25 bbl. of flour in one batch have been
constructed and tried, but without success; for general practical and economical
reasons I consider in regard to the above described types of kneaders a working
capacity of 1,200 to 1,500 lb. of medium dough as the most favorable one.
For smaller bakeries where 1 to 2 bbl. of flour form a batch, simple single-blade
mixers, for instance that shown in Fig. 18, will usually do the work with satisfactory
result and at the costs of little power. Also a special drum mixer in which a number
of rods, extending parallel to the axis from one flat side to the opposite one of the
drum, act as beaters, is employed for making dough; it is only suitable for smaller
batches and slack doughs.
Cake doughs are made from flour with such large additions of fat, sugar,
raisins, currants, fruit jams, condiments, etc., that the qualities of the flour and
especially of its gluten loose their prominence. Any mixer, which is liquid-
tight and mixes quickly and well, will therefore do the work, for instance a small
two-blade mixer or a double-motion agitator of simple construction. In smaller
bakeries and confectioneries usually simple agitators rotating round a vertical
axis and adapted to be lowered into the mixing bowl or pan are employed, the
agitators as well as the mixing vessels being exchangeable. Agitator blades of
different type and size and mixing vessels of different shape and capacity are
Fig. 40.—Typical cake-mixing blades.
supplied, and the driving gear has three or four different speeds, so that every
probable kind and size of cake dough can be produced in the single machine.
Figure 401 shows a typical set of blades for this class of mixer; a is intended for
roll or pie dough, b for heavy cake, almond paste, sugar icings, etc., c for mashing
vegetables and fruits, for sponge and soft-cake batches and d for whipping eggs,
cream, mayonnaise, dressings, etc. The blades are rotated the quicker, the
thinner and more fluid the mixture to be treated.
Biscuits, crackers and similar flat and thin cakes which are intended to keep
well, contain, after having been quickly baked and, if still too moist, dried, only
3 to 9 per cent of water against about 40 per cent in bread which is eaten fresh;
thus it is not only advantageous, but necessary that the original dough is made
with as small a quantity of water as possible. The ingredients are first mixed in
a kneading machine suitable for heavy dough; in consequence of the absorption
1 Mixer manufactured by Bead Machinery Co., York, Pa.
36
of water by the starch and gluten of the flour the mixture will after a time be-
come so stiff and hard that its further treatment in the machine becomes useless
and even impossible; in this way neither the soluble and insoluble components,
as sugar, salt, fats, condiments, etc., can be thoroughly intermingled with the
water and flour, nor the glutenous condition of the latter fully developed.
This purpose is attained, and the kneading of the dough finished by pressing and
drawing it out to a long sheet between two heavy rollers the dough brake by then
folding it up, letting it pass through the brake again and repeating these operations,
until the resulting dough has the required evenness, smoothness, cohesiveness and
elasticity for its thin sheets to be cut and baked into perfect products. The dough-
brake is" an evolution of the rolling pin and consists essentially, as diagrammatieally
shown in Fig. 41, of two parallel, superimposed rollers a and b and the stationary
table c which contains in its middle part the opening d under which the roller 6, the
feed roller, protruding somewhat therefrom, is rotatably fixed, the long edges e of d
being formed to serve as scrapers for b.
The rollers rotate in opposite directions so that the dough sheet travels through
between them; in order to work from either side of the table, the driving gear is fitted
with devices for easily and quickly reversing the rotation of the rollers; their distance
is adjustable for producing a dough sheet of pre-determined thickness. The rollers
used in this machine are usually of steel, smoothly polished and have the same diame¬
ter which is made the larger—and thereby the weight of the roller the heavier—
the greater the difficulty with which the dough is worked. The power necessary for
driving the machine is comparatively small; a brake having rollers of 2 ft. 2 in. length
and 10 in. diameter (weight about 580 lb. each) requires only about 2 to 3 hp. The
output of the machine greatly depends on the skill and experience of the operator.
Macaroni, noodles, vermicelli and similar products contain also only a small
quantity of moisture, about 9 to 12 per cent, and therefore as little water as
possible for obtaining a good result is used in the dough. The flour for the same
must be very strong, and its gluten developed to its full viscidity and elasticity
so that the dough can be easily divided and shaped by means of dies under high
pressure in a screw, or hydraulic press, and that the products thus obtained, tubes,
strings, threads, bands, figures, etc., retain their shape during drying and do not
break or are damaged when packed. The dough can, of course, be worked like
that of biscuits first in a mixer and finally in a dough brake; but, as no sheeting
is required, usually the manufacturers of those articles employ instead a kind
of edge mill which, working the dough under high pressure automatically, saves
the considerable amount of manual labor and attention necessary for the dough
brake. This apparatus1—the rotary roller pan—consists of a flat circular pan
rotated round a vertical shaft, arranged in its center, by means of a toothed ring
fixed concentrically to the bottom of the pan and meshing with the driving gear
of the machine, and of one or two heavy rollers loosely mounted on shafts fixed
to the framework radially to the pan and rotated by the dough passing underneath
on the bottom of the latter when revolving. The dough is continuously turned
over by a plow-like scraper and crushed, ground together and kneaded under the
pressure of the very weighty rollers the pressure of which may be enhanced by
strong springs either pressing the vertically movable pan against the rollers
mounted on stationary supports or vice versa. The working capacity for a pan of
1 Made by the Baker-Perkins Mfg. Corpn., Saginaw, Mich.
3 bout 5 ft. 6 in. diameter is a batch of 250 to 300 lb. of flour, requiring about
<5 hp.
Mixing and Incorporating Plastics.—Substances which are of a plastic nature
and capable of imparting this quality to a limited amount of solid or liquid
■matter, have usually so high a degree of cohesiveness that they can only be
mixed with the other ingredients with great difficulty, considerable pressure being
necessary to overcome their toughness. Two-blade machines as shown in
Figs. 15, 17, 19 and 21, the double-motion agitator, the combined double motion
and screw agitator, Fig. 31, and similar machines are in most cases capable of doing
this work satisfactorily. Heat must frequently be employed for assisting in the
operation. In cases in which the pressure during mixing must be especially
great, for instance when substances, wrhich like india rubber possess besides great
cohesiveness an exceptional high degree of elasticity and tenacity, rollers like
those described for the dough brake, Fig. 41, are employed with advantage. A
two-blade machine, Fig. 17, especially constructed for this kind of work in which
the active surfaces of the blades, while retaining their diagonal position in the
double trough, form curved instead of even planes, has been found very useful in
the manufacture of all kinds of rubber goods, balata, lineloum, artificial ivory, art
stones, and similar products. When, as is frequently the case, fine powdery
substances are to be incorporated into tough rubber-like masses, it appears to
be advantageous to add the powder gradually in the measure as it is taken up,
spreading it widely over the whole mass exposed in the machine.
Whisking and Emulsifying.—By very rapid and violent agitation cream, eggs,
their mixture with sugar, thin batters made of flour, milk, eggs, etc., sauces and
mayonnaises consisting of egg yolks, oils, vinegar and flavorings and similar
mixtures comprising gelatinous and fatty ingredients and water, are whisked,
whipped or beaten into fight perfectly intermingled and blended masses which,
in case the gelatinous components are present in sufficient a quantity to be capable
of retaining the air whipped into the mixture in the shape of small bubbles, are
foamy or spongelike. Figure 4(M shows a simple eggwhisk made of a bunch of
wire loops which is rotated at 100 to 150 r.p.m. and gives for small as well as large
quantities a good result'.
Very effective mixers of this class are constructed according to the double-motion
principle, each of the agitators consisting, if the mixing vessel is a bowl or the like
having a spherical bottom, of half-circular bows of strong round wires symmetrically
arranged round the two hubs, or, if, that vessel is a trough with half-cylindrical bot¬
tom, of rectangularly bent strips of band steel connected by crosspieces, all pre¬
senting their narrow sides in the direction of the rotation. The agitators revolving
in opposite direction to each other are driven at 90 to 200 r.p.m.; the inner one fre¬
quently a little quicker than the outer one. The size of the mixing vessel must be
very large in regard to its real working capacity, as the materials under treatment
are much thrown and splashed about, and as the volume of the mass increases by the
inclusion of air to a remarkable degree; for instance 1 qt. of cream may form, after
having been properly whipped, 3 to 4 qt. of whipped product. The nature of the
ingredients prevents the employment of large-sized apparatus of this class; a whisk
capable of beating up 200 eggs with sugar will have a total capacity of about 10 gal.,
which is about the limit size employed.
The product of the even distribution of oleaginous matters in shape of minute
globules in a large quantity of water by means of still more forceful agitation than
that used in whisking forms an emulsion, the name being derived from the best-known
representative of this class, milk, which consists of a watery solution of milk sugar,
proteids and salts in which about 3}4 per cent of butterfat are finely suspended.
Mechanical means alone are not sufficient to produce an emulsion which would keep
any length of time; it is necessary that the water with which the fatty substances are
to be mixed contains in solution proteids, gelatinous compounds, sugar, glycerin,
soap or other colloid chemicals. The mixers used are the same as those employed
for whisking; only that the agitators are rotated at a higher speed, from 500 to 1,500
r.p.m. as the case may require. A most efficient emulsifier for commercial use is a
rather flat propeller-agitator, the blades of which form an angle of 5 to 10° with the
horizontal and are revolved at a speed of 1,000 to 1,500 r.p.m., the oily substance
being very gradually added through perforations in the hollow shaft.
SECTION XVI
SAMPLING
By Donald M. Liddell1
Definition.—Sampling is the process of obtaining a small quantity of material

which shall be representative of the average of some large amount. Theoreti¬
cally each portion of a sample however small should itself be representative of
the large body, but in practice this cannot be obtained owing to the impossibility
of reducing all of the sample to a molecular state of subdivision.
Owing to the difficulties of getting a small sample of a few ounces in weight to

represent a large amount of material, the sampler will ordinarily divide a large ship¬
ment into several lots, each to be sampled and settled for separately.
The importance of proper sampling is rarely correctly estimated. The utmost
care is ordinarily given to all the operations of a chemical laboratory. Large sums
are spent in its proper equipment and high salaries paid to the chemists employed
without regard to the fact that the accuracy of the analytical work is primarily
limited by the accuracy of the samples furnished to the laboratories. It appears to
be the customary assumption that any $3 a day man can prepare samples for a
$20,000 staff of chemists. As an interesting commentary in how little is ordinarily
thought of sampling, it was not until the eleventh volume of the transactions of the
Institute of Mining Engineers that any article appeared on sampling.
The ways in which the accuracy of a sample can be vitiated are many, but a
proper understanding of the underlying principles will enable one to see the mistakes
and to correct them. On the other hand, one of the easiest ways to point out the
underlying principles is to describe certain specific errors in sampling and to show the
cause and the cure, so that in large measure this paper will be a discussion of the errors
of sampling.
In the sampling of solid materials we come immediately upon two classes—

those which can be crushed and those which must be drilled, sawed or filed. In
sampling materials of the first class where payment must be made on the value of
the shipment the question of moisture is the first to present itself and is probably
one of the greatest sources of error. After weighing the material the moisture
sample must be taken as quickly as possible, and must be large enough to repre¬
sent the whole mass correctly yet small enough to permit of drying the entire
moisture sample. The reason for this latter stipulation is that it is impossible
to crush and mix a moisture sample because mechanical work performed on it
alters the moisture content. The moisture sample must not be representative of
fine material alone or of the coarse material only, but must be fairly representa¬
tive of both classes in their proper proportions. Usually a finely divided material
will carry a greater proportion of moisture than will coarse lumps, and it is ordi¬
narily the tendency of the sampler, owing to the small size of the moisture sample
(comparatively) to take an undue proportion of fines, since even one lump of ore
appears as a mountain in the moisture sample.
1 Weld <fe Liddell, Engineers and Economists.
565
Ordinarily a moisture sample is said to have been dried at 212°F., even though the
temperature of the drying oven may have been allowed to run up even to 240°. This
matter of drying temperature is one of importance for the material may contain water
of crystallization not completely given off at 212° or may contain salts where the acid
radical is more or less completely expelled at temperatures only slightly above 212°,
or the “moisture” may be some such substance as free sulphuric acid. At any rate
it is important that the moisture sample and the final sample for the laboratory
should be dried at the same temperature and if the material is hygroscopic care must
be taken to keep it in a stoppered bottle or other watertight receptacle. It should
also be realized that practically all crushed pulps are hygroscopic, and if any sample
is to be re-analyzed after it has stood for some time in paper or cloth bags, it should
be dried first.
In drawing the original moisture sample from any material whatever it is an
elementary proposition that the sample must not be taken from the immediate sur¬
face of the pile since there will be some alteration in the moisture content after any
exposure, however short. It can be proved fairly conclusively that if the moisture
sample be taken after the first crushing on material the size of which is fairly uniform,
the moisture thus determined will be from one-tenth to one-fourteenth lower than it
should be, depending on the season of the year. Thus in ordinary summer weather,
if a value of 5 per cent is obtained for the moisture, the true moisture is probably
close to 5.5 per cent.
In taking a sample for laboratory purposes, it is rare that all of any material is
either put through a sampling mill or hand sampled in its entirety. A certain pro¬
portion is ordinarily withdrawn at once and put through the sampling process. This
is primarily a matter of sampling cost but may have important commercial considera¬
tions also. For instance, in sampling ore for smelting it is usually desirable to keep
it as coarse as possible. On the other hand, ore to be milled can be crushed in toto
previous to sampling. In sampling crystallized chemicals, it is usually undesirable
to crush too finely. To determine this point, the sampler should know what is to be
done with the material, but it is preferable to take the selection of the material which
enters the sample away from the judgment of the sampler, so that machine sampling
is best. It is at least imperative to take every fifth or tenth or twentieth shovelful
rather than to allow a sampler to take a shovelful whenever his inclinations so move
him, though Brunton says that at best the process is that of “shovel selection”
rather than “shovel sampling.” The man who is doing the work should also be
watched to see that he takes shovelfuls of about the same size, both of the material
accepted for the sample and of that going into the reject, and here the working period
comes in, for at the close of a long period of shoveling, the shovelfuls will usually be
much smaller than at the start.
Other plans than taking shovelfuls at stated intervals are necessary where material
is not being unloaded by hand, but in any case the principle remains that the selec¬
tion of the sample should not be left to the caprice of the sampler—thus the United
States Steel Corporation in sampling ore arriving in hopper bottom cars specifies
that a network of cords shall be stretched over the top of the car and that equal
quantities of material shall be taken from under each of the points of intersection in
the net.1 In sampling certain homogeneous manganese ore arriving by steamer it
is perfectly proper to take every twentieth or even fortieth scoop as it comes up
through the hatch. The sample scoops are then unloaded into a car by themselves
and that car alone later goes to a sampling mill.
i The full paper on the methods of the Steel Corporation for sampling ore in bulk was published in
the Jour. Ind. and Eng. Chem., 1909, p. 107, and should be referred to by anyone interested in this
SAMPLING 567
It is also possible to take samples direct from the faces of ore in a vessel which is
being unloaded, but this method is not ordinarily to be recommended, as it leaves too
much to the judgment of the individual samplers.
Of course these fractional-selection methods introduce hand sampling, and the
more that can be eliminated, the better off we are. While the advocates of hand
sampling claim that with a trained gang of men the actions are practically mechanical,
this is untrue, for if they are trained they know what they are doing and why they are
doing it, which virtually is allowing the ore to be handled by interested parties. In
such cases it is either impossible for any man to submerge completely what he considers
are his employer’s interests, or else he consciously stands so straight as to lean over
backward. The greatest errors introduced by hand sampling are: (1) In fractional
shoveling, the shovelful is ordinarily not uniform, the ordinary tendency being to make
the sample shovelful the smallest. The size of the shovelfuls will also vary through
the day, as already mentioned. Moreover the place from which the sample is taken
is at the discretion of the workman. (2) In coning, the larger pieces of ore invariably
fall to the outside of the pile while the top retains the fines. Then when the pile is
spread out the fines go to the very outside. As the moisture sample is ordinarily
taken from near the surface of the pile and at this stage of the work and the fines
contain the more moisture, this operates in favor of the buyer. (The process of
“coning” is explained below.)
Cutting Down Samples.—Having obtained the original sample it is then sub¬

ject to alternate reduction in size of the individual pieces and in the amount of
the sample itself. Those cases are almost negligible where there should be two
successive reductions in amount of the sample without an intermediate reduction
in the size of the largest pieces making it up. It is not always necessary to crush
the entire sample. Much time and expense can often be saved by screening a
sample and crushing only the coarse portions, mixing them when crushed with
the finer pieces which have bypassed the crushers. (The question of coarse
pieces which will not crush will be discussed later under the heading of “Metal-
lics,” p. 581.) It is self evident that after this process pains must be taken to
mix the portions thoroughly.
Reduction of the bulk of the sample itself may be done either mechanically or by
hand or by a combination of the two. The hand operation is ordinarily known as
coning and quartering. The material is piled first into a cone and then the top of the
cone raked out so as to form the frustum of a cone.1 Diameters are drawn across this
at right angles and the opposite quarters are rejected. The remaining quarters are
then again shoveled into a cone being careful that shovelfuls are taken alternately
from the two quarters to go into the new cone. This cone is again flattened out and
the quartering process repeated.
The two great dangers in this process are the tendency of large lumps to segregate
at the bottom and outside of the cone, and unless great care is taken to place the
shovelfuls exactly on the apex of the heap there will be an undue concentration of
coarse material in some one quarter of the pile. The other is that in conducting
the quartering operations the workmen will either invariably take or reject all of the
large lumps which lie along the border line of the cut, thus causing the sample to contain
an undue amount of either coarse or fine. In raking out the cone into a frustum the
radial distribution of the material should not be disturbed. This question will be
referred to again under “salting,” p. 583, in discussing the operation known as “draw-
1 A detailed description of this process is given, by Jesse E, Soobey, Met. and Chevn. Eng., February,
1913.
ing the center.” Care must be taken not to lose dust from the material and it may be
necessary to sprinkle the pile from time to time in order to avoid this.
A piece of apparatus used in cutting down samples by hand work is the assayer’s
shovel, shown in the illustration. This consists in a number of parallel troughs
attached to a shovel handle, and separated by spaces of width equal to that of the
troughs. ■ When crushed ore is shoveled on to this shovel, a portion is retained in the
trough and a portion falls through, thereby splitting the sample.
A variation of this idea is found in the Jones sampler which consists of a number of
parallel triangular troughs facing in alternate directions (see illustration). One-half
of the material must go one way and one-half the other. For cutting down very small
and finely ground samples still another modification is the assay er’s riffle, also shown
in illnef.rflt.inn
SAMPLING 569
The mechanical samplers as ordinarily found are either the Vezin, Bridgman
or the Brunton type1 or some modification of them. In any case the sampling
principle is to cut across a falling stream of ore, taking either successive portions
of the stream as the sampler cuts across it or else diverting the entire falling
Fig. 3.—Assayers riffle.
stream for a certain period into the sample. In either case the danger is the
same, that the apparatus used should not be sufficiently large for the component
parts of the ore stream; that is, the large lumps of ore are driven into the large
Fig. 4.—Brunton sampler.
rejected portion leaving an undue proportion of fines in the sample. This is a

matter which should be carefully watched, as should also be the point that the
entire falling stream is diverted or is cut across and not merely a portion of it,
for such a stream tends to classify itself as it falls, and a sample cut from one
edge only will not be the same as one cut from the center, nor will the edges
themselves ordinarily be alike. In rotating samplers it is therefore important
that the cutting edges are radial planes.
The matter of relative size of piece and size of opening is illustrated very well
by the sampler’s shovel and the assayer’s riffle. It is easy to see that these pieces of
apparatus are entirely unreliable on material containing individual pieces of larger
diameter than the width of the openings, for then these larger pieces will all be retained
on the shovel or riffle. It does not seem safe to use these two pieces of apparatus with
material the maximum pieces of which exceed one-fifth the width of the opening, and
it seems probable that the same proportion would be a good one to apply to sizes of
the lumps and cutout holes in machine sampling.
The Brunton sampler consists either in a wedge-shaped deflector plate that can be
thrown back and forth in such manner as to send all the ore stream to a storage bin or
to the sample, or in a scythe-shaped casting which swings back and forth in the
ore steam. The drive is ordinarily by a connecting rod driven from a belted wheel,
and by altering the position of the connecting rod the arc of travel is altered.
In the latter form most of the surface slants toward the storage-bin hopper, the
remainder toward the sample bin and the stream of ore is directed according to which
portion of the casting it is hitting. The proportion of ore deflected is altered by adjust-
Fig. 5.—Vezin. sampler. Fig. 6.—Snyder automatic ore sampler.
ing the arc of travel of the casting. As the arc is cut down, the portion inclined toward
the sample bin constitutes a greater and greater portion of the part of the casting in
use. In the former type of apparatus, the wedge, the adjustment is made by cams
which may be set at varying angles. The sampler points one way during one part
of the revolution and in the other during the remainder, so that any adjustment is
possible.
Practically the same end as that attained by the oscillation of the Brunton sampler
is gained by rotation in the Snyder sampler (this is not the Charles Snyder sampler
spoken of on p. 571). This sampler consists merely of a circular casting shaped much
like a miner's gold pan, having one or more openings (according to per cent sample
desired) in its sloping flange, mounted on the end of a horizontal shaft provided with
tight and loose driving pulleys.
In operation the sampler revolves 10 to 30 times per minute, the speed depending
upon the size of the sampler itself and the size of the material being handled. The
material to be sampled is directed by a spout so as to fall inside of the sloping flange of
the sampler. The rejections slide off the flange into a receiving hopper. The sample
drops through the opening in the flange as the opening passes under the spout. It is
a most simple machine, and its only disadvantage is that the percentage of sample to
reject is fixed by the ratio of the arc of the opening to the full circumference or 300°.
The Vezin sampler is shown in the cut, and as will be seen consists in a rotating
cylinder bearing two or more scoops, which as they revolve, cut through a falling
stream of ore, delivering the portion they cut out (the sample) through the truncated
cone at the bottom, while the reject passes on the ore bin. Theoretically this sampler.
SAMPLING 571
as well as the Charles Snyder1 and Bridgman2 types have the objection that centrifugal
action disturbs the selection. In addition they are occasionally found made so that
the edges of the cutting scoop are not radial, which gives improper selection, and unless
they are rugged construction the edges of the scoops become bent, leading to improper
results. These machines are often made so that several reductions in bulk of the
sample are made without any intermediate crushing. This, I believe, is absolutely
wrong. On the other hand, they have great capacity, require little headroom and are
easily cleaned, with a hose, if necessary.
Crushing for sampling may be done in any apparatus that fulfills the following
conditions: It must be easy to clean; it must not wear greatly, otherwise the
material from the crusher will contaminate the sample; the material must not fly
about, but must all be delivered at one spot. Hence, gyratory crushers, jaw
crushers and rolls, are used for sample breaking, and ball mills, drug mills, small
breakers and small rolls for final crushing. The three above conditions are the
only essentials.
For feeding material to the crushing or to the conveying machinery, it should be
remembered that shaking belts or tables all classify the material fed and hence tend,
to inaccuracy unless pains be taken to offset this tendency.
Coal Sampling.—Probably coal sampling is subject to greater proportional

errors than is any other common sampling operation. This is not alone due to the
fact that it is a cheap substance which means that not much can be spent upon
sampling it, but is also owing to its deceptive appearance. The various parts
of a heap of coal look beautifully homogeneous; it usually comes already partly"
crushed, but the point remains that the assumption of uniformity leads to bad.
errors. One swallow will come nearer making a summer, than will one lump of
coal, picked out by the master mechanic, to making a good sample.
While the final preparation is a matter of ordinary coning, quartering and breaking:
down, the following, by C. E. Scott, is a good description of what ought to be done in-
taking the sample.
“The size of the sample shall not be less than the proportion of one part per 1,000,
and the minimum size of any one sample shall be 100 lb.; the maximum number of
tons one sample shall represent shall be 1,000, and preferably much smaller than,
this, say, 250 tons. All the impurities, such as slate and pyrites, should be broken
down in the original sample, and if this is done as each shovelful is taken, it will make
the work much easier. In collecting the original sample from a railroad car it should
be taken as the coal is being discharged; where the coal is delivered to the point of
sampling in wagons, one large shovelful should be taken from each wagon; where it is
delivered in a barge, the sample should be taken as the barge is discharging (if thro
discharging is accomplished by the aid of a bucket or belt conveyor it is advisable ho
take advantage of this and obtain a small shovelful from the buckets or off the belt
regular intervals); where the sampling must be taken from a pile, a shovelful should
be taken from all accessible parts of the pile, and the shovel should be filled from tlxo
coal underneath the surface. As the coal is collected it must be placed in a box, barrel,
or can, of sufficient size, which has been thoroughly cleaned out before using. A lid
should be kept on the container at all times to prevent dust, ashes, etc., from collecting
on the sample. When the total sample is completed it should be put through a cruslaer
or dumped on to a clean wood or cement floor, or bn a canvas cloth, and crushed down
,
i Trans. A.I.M.E., Vol. 40 p. 570.
a Trans. A.I.M.E., Vol. 20, p. 416.
by means of a hammer until no pieces larger than 1 in. remain. The sample should
then be mixed by shoveling it over itself into a conical pile.”
Size-weight Ratio.—Many attempts have been made to deduce a proper
ratio between the size of the largest pieces in the lot of material to be sampled
and the size of the sample. Probably one of the best of these was a paper on
“Accuracy in Sampling Coal.”1 This was an investigation into the cause of
inaccuracies in sampling, based on both laboratory experiment and theoretical
calculation. While the investigation was concerned with coal, the general
results and reasoning are applicable to all substances:
Size of Slate Contained in Coal, and Size of Original Sample Required to

Insure the Error of Sampling Being Less Than 1 Per Cent
Weight of largest piece of Original sample should

Size of slate, inches
slate, pounds weigh, pounds
4 6.7 39,000
3 2.5 12,500
2 0.75 3,800
lX 0.38 1,900
m 0.24 1,200
i 0.12 600
X 0.046 230
X 0.018 90
Size to Which Slate and Coal Should be Broken before Quartering Samples
of Various Weights
Weight of sample Weight of sample

Should be broken Should be broken
to be divided, to be divided,
to, inches to, inches
pounds pounds
7,500 2 40 2 mesh
3,800 ix 5 4 mesh
1,200 i X 8 mesh
460 H X 10 mesh
180 X
Limit Beyond Which Samples Should not be Divided when Crushed to

Different Sizes in Laboratory
Size of coal mesh Should not be divided to less than, grams

l
2 8,300
4 1,100
8 120 ) Should be pulverized
10 55 \ to at least
20 J
3 60 mesh
l T51. d ■Ra.ilfiv. Jour. Ind. Ena. Chem. . 1909. r>. 161.

SAMPLING 573
In order to obtain a sample which shall correctly represent any material it is

necessary to crush to such a fineness that a particle more or less of any component of
the material taken into the sample shall not materially affect its analysis or its value.
The maximum error is determined by the ratio of the weight of the largest particle of
metal or high grade mineral or impurity to the weight of the entire lot. The chance of
finding a full-size piece of highest grade material would be much greater in a lot of ore
crushed to 0.25-in. cubes than in a lot crushed to 1-in. cubes. Therefore accuracy
demands that the ratio between the weight of the largest particle and the entire lot
shall increase directly as the fineness.
As a consequence of these considerations, in gold ore sampling it is customary to
go a little beyond the above figures, both in size of sample and in fineness of crushing,
while in sampling chemicals and other more uniform products, the samples may be
smaller in weight. Tables for ore sampling follow:
Weights to be Taken in Sampling Ore1
Weights Diameters of largest particle
Very low Low Medium ores Rich Rich and

grade of grade spotty
ores,
Grams Pounds uniform ores, milli- ores, milli-
ores, milli- milli- Milli- Milli-
meters meters
meters meters meters meters
20,000.0 207.0 114.0 76.2 50.8 31.6 5.4

10,000.0 147.0 80.3 53.9 35.9 22.4 3.8
5,000.0 104.0 56.8 38.1 25.4 15.8 2.7
2,000.0 65.6 35.9 24.1 16.1 10.0 1.7
1,000.0 46.4 25.4 17.0 11.4 7.1 1.2
500.0 32.8 18.0 12.0 8.0 5.0 0.85
200.0 20.7 11.4 7.6 5.1 3.2 0.54
100.0 14.7 8.0 5.4 3.6 2.2 0.38
50.0 10.4 5.7 3.8 2.5 1.6 0.27
20.0 6.6 3.6 2.4 1.6 1.0 0.17
10.0 4.6 2.5 1.7 1.1 0.71 0.12
5.0 3.3 1.8 1.2 0.80 0.50
2.0 2.1 1.1 0.76 0.51 0.32
1.0 1.5 0.80 0.54 0.36 0.22
0.5 1.0 0.57 0.38 0.25 0.16
90.0 0.2 0.66 • 0.36 0.24 0.16 0.10
45.0 0.1 0.46 0.25 0.17 0.11
22.5 0.05 0.33 0.18 0.12
9.0 0.02 0.21 0.11
4.5 0.01 0.15
2.25 0.005 0.10
i Richards: “Ore Dressing,” Vol. 2.

Size-weight Ratio in Sampling1
Diameter of Minimum weight of

largest par¬ sample, pounds,
ticle, inches Colorado practice
0.04 0.0625
0.08 0.50
0.16 4.00
0.32 32.00
0.64 256.00
1.25 2,048.00
2.50 16,384.00
Smallest Permissible Weight for Samples of a Given Size2
Effect on value created

Ratio of weight of
Size, inches Weight of by one cube
largest cube to weight
cube or mesh sample, pounds assaying $100,000 per
of sample
ton, of sp. gr. 5
2 10,000 1 7,000 $14.42

m 5,000 1 8,300 12.17
1 2,000 1 11,000 9.00
H 1,000 1 13,000 7.50
X 400 1 18,000 5.62
X 300 1 31,000 3.17
X 200 1 71,000 1.40
He 100 1 83,000 1.20
X 75 1 220,000 0.44
6 mesh 50 1 430,000 0.23
10 mesh 25 1 930,000 0.107
18 mesh 10 1 1,900,000 0.051
30 mesh 4 1 4,200,000 0.023
50 mesh 1 1 5,500,000 0.018
Scheme for Sampling Rich Ores with Vezin Samplers3
Sample, per Pounds in 100

Inches
cent tons
Maximum size of cubes. 1.00 20.0000 40,000

Maximum size of cubes. 0.25 1.2500 2,500
6 mesh. 0.0625 0.0785 157
30 mesh.. 0.0171 0.0050 10
1 E. A. Smith: “Sampling and Assay of the Precious Metals.”

1 It. H. Richards: “Ore Dressing,” Vol. 3.
lR. H. Richards; “Ore Dressing” Vol. 3.
SAMPLING 575
Many experiments on the sampling of gold ores lead to the belief that carrying the
grinding beyond 100-mesh does not add to the accuracy of the sampling, and that at
this fineness equally good results will be obtained whether the material to be taken for
the assay be cut out of the sample by one cut of the spatula or taken by various cuts
from over the entire surface of the sample.1
We may, I think, therefore take that 100-mesh is as fine as it is necessary to reduce
any samples whatever, so far as accuracy of sampling is concerned. On the other
hand, since I have seen more concordant results obtained on a 60-mesh sample after
reducing it to 80-mesh, we may conclude that 80- and 100-mesh mark the limits of
desirable sampling. This totally ignores the question as to what are the best limits
for manipulation from the chemist’s and assayer’s side. It may safely be assumed
that no material requires finer grinding than does a spotty gold ore.
It is also of importance to note that contamination of the sample with material
abraded from the grinding machinery increases rapidly as the mesh of the sample is
cut down, so that this in itself offers a reason for keeping the mesh size as large as
possible.
The contamination of samples of those materials which can be easily crushed

by particles of the sampling tools, reference to which has previously been made,
may reach greater proportions than is ordinarily credited. In sampling coke
using a Braun pulverizer I have shown that the amount of iron introduced in the
final sample may equal 1 per cent of its entire weight, all of which will of course
show as ash when that constituent is determined, meaning that this determina¬
tion will be at least 6 per cent from the true result and may be as high as 20
per cent. From the ordinary bucking board and muller used for ore crushing
by a not over-industrious Mexican I have seen iron to the extent of 3 per cent of
the weight of the sample introduced into it. Of course if the sample is to be
assayed for copper, lead, silver or gold, the introduction of iron in this way saves
a .considerable amount of money in the course of a year in payments for these
valuable metals, but the inaccuracies introduced are fatal to good furnace prac¬
tice, even if one has chloroformed his conscience.2 For this reason the selection
of material for the grinding tools to be used in sampling is a most important,
question. Chrome steel and manganese steel should be used to replace cast iron
which is usually used, and incidentally, the greater wear obtained from these
harder substances will almost more than pay for the difference in original prices.
For this reason too bucking boards and mullers which have been cast with chilled
surfaces should be watched and when the chilled surface is worn through they
should be promptly discarded.
That iron can be introduced to the extent of 1 per cent of the weight of the coke
sample may possibly be questioned. However, there is no doubt that in the samples
spoken of this amount of iron was present in a metallic state as was determined by
agitating the sample with a neutral copper-sulphate solution, washing the resulting
mass free from the resultant copper and iron salts and determining the metallic copper
left precipitated on the coke dust. This iron equivalent of this amount of copper was
then calculated. The metallic iron thus shown to be present was not iron reduced
from the ashy constituents of the coke by the action of the incandescent carbon
1 Lewis V. Wright: Min. Mag., November, 1910.

2 The introduction of metallic iron into the sample also results in the assayer’s finding a heavy
black residue when the sample is panned that not infrequently results in all hands starting to look for
iridium. It also prohibits the use of evolution methods for the determination of COa in the ore.
during the coking process for when the same coke was ground in Abb<§ mills or
crushed in porcelain and agate mortars using porcelain and agate pestles the sample
thus produced did not react with neutral copper-sulphate solution, nor was the analy¬
sis of the ash obtained by burning off the coke at all the same, as the latter case
showed much more silica and much less iron. Reference to this subject of sample
contamination will be made again under copper-bullion sampling.
SAMPLING METALLIC MATERIALS
In sampling metals there are two main dangers to be guarded against. First,
that the outside of the bar or slab will ordinarily have a certain amount of dirt
on it of character entirely dissimilar from the inside of the bar and also that even
in the interior of the bar the metal is as a rule not homogeneous. In fact it is
the rarest exception when an alloy of two or more metals does not segregate upon
cooling. For instance, in a copper bar which averages 120 oz. of silver to the
ton individual samples can be drilled from it which will run anywhere from 80
to 200 oz.
As regards dirt and scale on the outside of the bar there are times that a por¬
tion of this belongs in the sample. This will be the case in sampling copper or
lead bullion where the metal is to be paid for upon assay. As the dirt is weighed
with the shipment a certain portion of it should appear in the sample. On the
other hand in handling bars of bullion of any sort it is a very easy matter to detach
more of this dirt or scale and get it into the sample pan than should properly be
there. Here then is a case in which the judgement of the representatives of
buyer and seller must be usjjd as to when the proper amount of dirt is getting into
the sample and when this is being exceeded.
The fact that there is dirt and scale on the outside of metallic bars gives rise to
some very curious consequences. It would appear at first sight that in the drilling
of copper bullion it would make no difference whether the drill was pushed through the
bar from top to bottom or vice versa, yet this is not the case, nor is it by any means
merely a theoretical difference. The gold and silver contents may be affected by more
than a per cent of their value and the copper content itself sensibly modified by the
direction in which the drilling is done. This is due not alone to the fact that ordinarily
the top of the copper bar will have more dirt and scale on it than will the lower surface,
but also to the fact that the interstices in the metal will retain some of the fine powder
first drilled off when the metal is being drilled top up, while all of this dirt will fall into
the sample when the bar is being drilled bottom up.1 In the sampling of lead bars a
punch is ordinarily used so that this question of dirt does not enter in any great degree.
On the other hand the punch is ordinarily driven half way through the bar so that it is
important in order to avoid the effect of segregation that the available length of the
punch should not be greater than half the thickness of the bar. For if it is, it results
in an undue amount being taken representative of the center of the bar as compared
with that representative of the outside. In punching lead bars it is sometimes the
practice of the sampler to wet the punches with oil. As the punchings are later melted
down and small sample bars cast from them, this results in inaccuracies for the oil will
reduce the oxide of lead present and give a higher assay than should be obtained. On
the other hand if the samples are allowed to remain uncovered too long before melting
the oil causes any flying dirt to stick to the punch cores and the sample will be con¬
taminated in this way.
1 Eng, and Min. Jour., Nov. 5, 1910, p, 897.
SAMPLING 577
The melting is ordinarily done in a graphite crucible. Of course this in itself tends
to a certain reduction of lead oxide but probably not more than sufficient to overcome
the effect of the oxygen of the air. This latter reagent makes it necessary that this
melting should be done as quickly as possibly and at the lowest temperature at which
the molten lead will be completely fluid, for if the temperature be too high or the
mass be kept too long in molten state the lead will scorify, resulting in the sample being
higher in gold and silver than the bars actually are.
It should be unnecessary to point out that when wooden blocks are used under
material that is being drilled, care should be taken not to drill into these blocks, but
too many samples are found containing wood chips to allow this caution to be omitted.
Samples of gold and silver are ordinarily poured from the molten metal and
consequently there is but little danger of contamination by particles of the samp¬
ling tools. Lead is so soft that it also is not affected. With copper, iron, steel
and some other metals the sample is invariably contaminated with particles
from the sampling tools. Thus in copper drillings which have been ground the
contamination may amount to as much as 0.4 of 1 per cent. It might appear that
in this case it would be comparatively easy to pass the drillings over a magnet
and remove all of this extraneous material. Unfortunately if this were done an
error would be introduced on the other side because blister copper will have a
certain amount of converter slag adhering to it some of which will enter the
sample, where it properly belongs.1 This slag is, like iron, paramagnetic and
will be removed by a magnet. In sampling steel or iron the case is even worse.
The only accurate method is to use sampling tools of a hardness greatly in excess
of that of the metal to be sampled and if possible to ascertain the loss experi¬
enced by the tools as compared with the total weight of the sample using these
weights and the analysis of the tools and of the sample as the basis for the correc¬
tion. Of course such refinement as this is happily seldom necessary.
It will occasionally happen that metallic material will be so dirty from adherent
slag, matte or other foreign material that it is necessary to melt the entire shipment
down, recast it and sample both metal and scoria separately (see paragraph on
“Metafiles,” p. 581).
Owing to the fact, already referred to, that molten metals segregate on cooling,
drillings from any one portion of the bar wifi ordinarily not be representative of that
bar. There is of course a point in going from center to periphery of a bar where the
sample will be representative of one or more elements, but not necessarily of all.
That is, in a lead-silver alloy the silver content wifi be at a maximum at the surface
of the bar and a minimum at the center, while the lead wifi be at its minimum at the
surface of the bar and its maximum at the center. While there will be some interme¬
diate locus at which the content of either silver or of lead represents the average of one
of these materials in the bar, whether these loci will be identical depends on the other
elements present.
As a consequence a sample must be taken which wifi represent all portions of the
bar. In copper this is obtained by drilling holes in successive bars by means of a
templet which covers all, one-half, or quarter of the bar (my own preference being for
the last).2 The templet is so arranged that it spaces holes evenly over all, one-half,
1 Eng. and Min. Jour. Oct. 15, 1910, p 752.
2 Theoretically the quarters are identical, and it takes fewer holes to cover a quarter templet than
a whole templet, spacing the holes on V/i in. squares. Consequently in lots of bullion where the
number of the bars are not integral multiples of the number of templet holes, a better distribution of
holes can usually be attained with quarter templets than with whole, as there will, on the doctrine of
chances, ordinarily be fewer holes to scatter, after the even rounds have been drilled.
37
or one-quarter of the bar and not over in. apart. While the more durable templets
are built of sheet iron, the best material, all things considered, is wood. Flanges, as
shown in the illustration of the templet, catch both edges of the pig or bar and serve
to fix the location of the hole. In the templet that is here illustrated, chalk is rubbed
on the surface of the templet over and around the desired hole and some of it falls
through the hole into the bar. This spot is then marked with a punch, the dust
brushed off the bar and drilling done by the punch mark. (With a sheet iron templet,
the punch is occasionally driven through the templet hole and no chalk used. The
drill holes must go completely through the bar and half the drilling should be done
from the top and half from the bottom. With lead, as a rule, the punchings are taken
on successive bars so that each hole is one-fifth the width of the bar further over and
one-fifth further along the length of the bar than was the preceding one. If a single
pig of iron is being sampled the sample should be obtained either over the whole
surface of the pig or by drillings taken from several points scattered over the entire
surface. In samples of gold and silver bars if they are not taken from a poured sample
the cuttings from the bar should be taken from both top and bottom.
Ladle samples taken by abstracting a ladleful of molten metal from the

charge and granulating it in the water are often recommended but the matter of
segregation applies to them likewise. No sooner does metal begin to chill in a
ladle than segregation begins to take place and if any skull forms inside a ladle it
will ordinarily differ markedly from the material poured out. For this reason
the ladle used in sampling any metal should be hot as compared with the metal
to be poured from it. The sample should be poured quickly and no more should
be poured after a skull has begun to form in the ladle.
Lead Sampling.—As can be seen from the character of the metal, lead can
be sampled with less likelihood of error than can any other metal. The process
is thus described by Raht.1
After melting the bullion bars down in a kettle holding about 20 tons, the coppery
dross is skimmed off and sent back to the blast furnace to be smelted over; with this
one of the worst sources of error in sampling is removed. The bullion is then
thoroughly stirred, say, for 5 min., and while this is still going on a number of samples
are taken by means of a small ladle holding approximately one-half an assay ton of
metal. It is important that this ladle should be at least as hot as the bullion, so that;
none of the metal would stick to its surface, otherwise the samples would be too low in
silver. The samples thus obtained are taken direct for assaying without any trimming;
or clipping. Instead of trimming the sample down to the exact weight of half an
assay ton, the exact weight of each sample is ascertained, whatever it may be, and from
1 Raht, A.: The distribution of silver in lead bullion, and the different methods of sampling, ildf-in..
Ind., Vol. 3, 1894, p. 414.
SAMPLING 579
the resulting silver button the contents are figured. Care is taken, however, in the
selection of the samples for assaying out of the great number originally taken that these
should be somewhere near the same weight, so that when they are cupeled side by side
they would finish about the same time.
It will be seen that all the objections to punch sampling are met in this method.
Still, another one might be raised against it, which is the uncertainty whether the
mixture was thorough. I must say that this does not appear to carry much weight,
but it can be overcome by taking samples at regular intervals out of the stream of
metal running from the kettle into the molds. These samples would have to be
remelted in a graphite crucible and out of this again one-half assay ton samples taken,
which are then treated as above. This method, of course, brings in again the objection
to the remelting. There cannot be the least doubt that either of these two dip samples
would be nearly absolutely correct, but they can be carried on only at the works of the
seller of the bullion, as it would be next to impossible to obtain a correct sample from
the dross skimmings. It is to be hoped that the dip sample will be generally intro¬
duced wherever practicable, and so do away with the annoying sample and assay
differences between the bullion producer and refiner.
SAMPLING PIG IRON
In the sampling of pig iron one of the optional methods of the United States
Steel Corporation is described as follows by the chemists’ committee of that
corporation:1
Plate or Pat Test.—With a suitable hand ladle a portion is dipped from the stream
of molten iron and, with the spoon of the ladle resting upon a clean, dry, iron plate, a
pat of such size as may be desired is poured thereon.
The use of the water-shot sample is to be considered as in violation of the standard
method.
Number of Samples to be Taken.—As tending to a reasonable degree of accuracy, it
is recommended that a portion be secured, timed as nearly as may be practicable for
the middle of each ladle of iron representing the cast. Equal portions from each of
these samples are conveniently combined for subsequent analysis.
Preparation of the Sample.—The tests are either drilled or crushed, as required. If
crushed, only that portion passing an 80-mesh sieve is used for the subsequent analysis.
In view of the segregation that takes place in any ladle, as has been commented on,
and in view of the differences that exist between the coarse and fines of cast iron, the
above may do for furnace control work, but for accurate sampling it is a capital set of
instructions on how not to do it.
Shot samples may be taken from a charge which is being poured by batting
through the falling stream of metal so that the resulting shot goes into a nearby
pan of water.
Moisture in Bullion.—In metals which are not completely solid and which
have been cooled under water, moisture is possible and this moisture is not driven
out even at 212°. This is of commercial importance in pig copper which has been
bosh cooled. Apparently the explanation is that the water gets into minute holes
in the metal, the necks of which close completely due to the contraction and
distortion on cooling and will not reopen when the material is heated to 212° only.
i The methods of the United States Steel Corporation for commercial sampling and analysis of
pig iron, Jour. Ind. and Eng. Chem., Vol. 4, 1912, p. 801.
As I have shown this moisture may amount to as much as 0.6 of 1 per cent and is
not completely expelled at temperatures below 240°F.1
In sampling pig iron it is ordinarily the grade of the pure pig which we are
concerned with and not the average analysis as affected by adhering scale and
sand. For this reason in taking a sample of pig iron either by drilling or filing
from the broken end of a bar it will be found of great assistance to wrap a strip
of paper tightly around the bar several times, thus preventing any scale or sand
from dropping off into the sample.
In sampling refined copper bars, which are probably as nearly homogeneous as any
commercial metal, most of the analyses can be made from borings taken anywhere
except just at the surface of the bar. For suboxide determina¬
tion filings or sawings are best. Teh Cananea Consolidated
Copper Co. formerly ran a set of six small parallel power-driven
saws for the preparation of such samples. These samples must
be freed from iron by means of a magnet and since the original
material is free from iron and iron oxides, no error is thereby
introduced.
Where the sample must be taken from an iron or steel
surface by planing, it is well to hammer off all scale, sand or
other extraneous material from the outside of the bar and then
Fig. 8.—Sampling set up the bar so that it projects up through a sheet of stiff
pig iron. paper as shown in the illustration. The shavings will then
be caught on this paper and can easily be transferred to a
sample bottle. Care must be taken not to run the planer too fast, for if this pre¬
caution be not observed the coarse shavings will fly away and be lost, leaving only
fines behind, which seriously vitiates the sample.
In general there is a great difference between the coarse and fine portions of any
sample. With iron any free carbon in the metal pulverizes readily and is found in the
fines. How great the difference between the coarse and fine is shown by the following
analyses of dark gray cast iron planings screened through an 80-mesh sieve:2
Coarse, Fine, Calculated

173 g, 16 g., composition,
per cent per cent per cent
Total carbon... 2.62 11.66 3.39

Graphite. 2.38 10.91 3.10
Combined carbon. 0.24 0.75 0.29
Silicon. 2.44 2.27 2.42
Phosphorus. 0.51 0.56 0.515
Sulphur. 0.08 0.073 0.079
Another case (taken from the same source) was a Swedish gray iron which was
planed over the entire cross-section giving 693 g. of shavings. These were screened so
as to obtain three portions: On 80-mesh; through 80-mesh and on 120-mesh; and
through 120-mesh. Analysis of the screened portions gave results as follows:
1 Eng. and Min. JoumDec. 3, 1910, p. 1095.

2 Bauer and Deiss: “Sampling and Analysis of Iron and Steel.”
SAMPLING 581
Coarse, Medium, Fine, Calculated

469 g., 193 g., 1 31 g, analysis,
per cent per cent per cent per cent
Total carbon. 3.53 4.04 22.00 4.50

Graphite. 2.12 2.53 20.94 3.07
Combined carbon. 1.41 1.51 1.06 1.43
Silicon. 0.81 0.83 0.76 0.80
Phosphorus. 0.051 0.044 0.03 0.048
Sulphur. 0.017 0.014 0.013 0.016
On a sample of comparatively clean converter copper, screened over a 12-mesh

screen, the portion on the screen assayed 98.94 per cent Cu, that which passed the
screen 97.21 per cent. The proportions were as 24:1.
There is no need of multiplying instances with analyses of other materials. The
sampler as well as the analyst must realize the variations that may exist between
the coarse and fines of a sample and the sampler must so arrange his methods that
the correct proportions of each may be turned into the laboratory. As to how the
analyst will treat the sample after it is delivered to him does not fall within the scope
of this book.
Hardened steels may be annealed before machining. Annealing for 15 min. at
750 to 800°C., followed by slow cooling until the pearlite point is reached (700°C.)
is sufficient to anneal carbon steels. Below 600°C. the steel may be cooled as quickly
as desired. In the case of certain alloys and self-hardening steels it will be necessary
to use chrome-tungsten and other alloy steels for the planing tool. With material
that cannot be worked even with special steels, it is necessary to knock off small pieces
from different parts of the sample and to pulverize these pieces. It is recommended
by Deiss and Bauer that the sample be wrapped in heavy linen cloth before pulverizing
with the remark that the broken pieces must be carefully cleaned to free them from
linen fibers. It should be noted that in cases of this class that carbon analyses are
always subject to some doubt, as it is next to impossible to remove all of the cloth.
As a very elementary precaution, in taking samples of iron or steel, it is well to look
at the drills or other cutting tools, to see that no large pieces have been broken out of
them. If there have been, these pieces can usually be found and taken out of the
METALLIC S
The limiting case of variation between coarse and fine material in the same
sample is that of metallic particles or masses occurring in material which other¬
wise is capable of being crushed. In this case the sampler should weigh the
metallic and non-metallic portions of the sample as soon as they can be properly
separated, sample each separately and report the original weights to the
laboratory.
There are often metallic substances occurring which can be ground by suitable
means and the powder added to the pulp sample. This, however, is a questionable
procedure, for this powder must be of different composition from the easily pulverized
material and its introduction into the sample makes a “spotty” sample. Much of
the trouble, for instance, with samples of gold ore that give discordant results is due to
forcing through the sampling screen small metallic particles that should have been
removed and weighed and reported as “metallies.”
It may be noted at this point, that the formula for computing the average value of
a sample containing metalUcs is as follows:
If W — weight of pulp
w — weight of met allies
A = assay of pulp
a — assay of metallics
™ , Tr , WA -j- wa
Then average value 7 equals ^ ^7“
Occasionally, as in sampling reverberatory slags, metallics will be taken out at each,

stage of the crushing. In this case a note should be made of the relative weights of
metallic and non-metallic portions at each stage, but it appears to involve no great
error to turn in but one sample of metallics along with the final sample of pulp.
The assay of the “fine” of the next to the last stage must be calculated from tlie
assays and weights of the coarse and fine of the last stage. The assay of the coarse of
the next to last stage is assumed to be the same as the assay of the coarse of the last
stage, and so on.
In the general sampling of slags and other fused products it should be noted that
these materials will decompose most easily when they have been suddenly chilled, and
it is, therefore, recommended, for the convenience of the analyst, that this be done
when possible, in sampling fused masses.
SAMPLING OF LIQUIDS
Concerning the sampling of liquids or semi-liquid material there is not much

to say except to point out the necessity of rapid agitation to insure uniformity
of the bulk to be sampled and drawing off the sample during the agitation or so
quickly thereafter that segregation has not had time to take place.
The same precautions are necessary in further laboratory tests of these samples.
Even though the liquids contain no sediment it must still be remembered that tempera¬
ture changes in the room in which a liquid is standing are sufficient to cause differences
in concentration of dissolved solids between the top and the bottom layers.
Errors are sometimes introduced through samples being taken at high temperatures
and crystallization occurring due to cooling down the samples before use is made of
them.
As to the matter of the proper treatment of samples where there is a sediment; tha
is, how to estimate correctly the influence of the sediment on the original sample and
to what extent the sediment should be washed and what is the influence of the washr
ing water upon the sample, no general directions can be given.1 This is a mattel
that must be determined by the chemist for each case independently. The genera
instruction can, however, be taken to heart that where these samples are not routine
matters and the procedure absolutely worked out it will invariably pay to recheck
these calculations after making them, or better still, to attempt to deduce the final
result by an entirely new chain of reasoning.
SAMPLING GASES
In sampling gases it is again difficult to give general directions. The use of

two bottles connected by a rubber tube attached to glass or metal tubes reaching
1
The classic tables bearing on samples and problems involving a pulp suspended in a liquid are
those of G. H. Clevenger, H. W. Young and T. N. Turner, Eng. and Min. Jour., Dec. 19, 1914.
SAMPLING 583
to the bottoms of the bottles and passing through two-hole corks, the other aper¬
ture being closed by another glass tube running into a rubber tube and stop
cock, as an aspirating apparatus is well known. The apparatus is simple, but
has the grave defect that the water over which the gas is usually collected absorbs
the easily soluble gases, such as C02, NH3, S02 and the like. This can be obvi¬
ated to some extent by trying to saturate the water in the bottles with the gas to
be collected but there still is danger either that the water will not be nearly
enough saturated to stop absorption or that it will contain so much gas that it
will be in unstable equilibrium and will be given off into the sample.
To a certain extent these errors can be avoided by the use of a 10 per cent sulphuric-
acid solution, as the solubility of most gases (except bases) is less in this than it is in
water. The errors can only be entirely done away with, however, by collection over
mercury, or by the use of previously exhausted vacuum tubes, the valve of which can
be opened in the gas to be sampled.
In sampling furnace gases the results are almost invariably vitiated by a rush of air
along the sampling tube down into the zone to be sampled. Leaky joints in the appa¬
ratus should always be carefully watched for also.
INTENTIONAL VITIATION OF SAMPLES
Salting.—Of almost equal importance with knowing what one must do to

obtain correct samples, is to know what may be intentionally done to render one’s
samples incorrect. This is, of course, of most importance to the mining engineer
who is sampling mines, or to those engaged generally in sampling for the precious
metals, but the sampler of any sort of material should know what may be done
to render his work misleading.
Many of the following methods are taken from a paper by George A. James on the
subject.1
Salting may, of course, be innocently or criminally done. Among the ways in
which the latter may occur is the use of old coin bags for carrying samples in. The
abraded particles left clinging to the bags may cause serious errors. Another way
of salting one’s own samples is to use one sack for two samples, separating them by a
string around the center of the bag. In such a case the samples must be removed by
cutting both ends of the sack, leaving the string tied. Inferior or porous sacks must
be avoided, or else the fines may be lost through the meshes of the bag. Canvas used
for catching the sample as broken in the mine may be contaminated, as may also the
tools. This last remark applies with equal force to everything used throughout the
sampling process. Proper allowance is not usually made for the pulp that may be
retained by the grinding machinery, nor for metallics which may stick to the muller,
bucking board, etc.
As a rule assayers do not give sufficient attention to drying the samples properly.
While the failure to consider 2 to 3 per cent of water may at times be a wise factor of
safety, omission of proper drying may sometimes lead to disastrous errors.
The use of the precaution commonly taken to overcome possible salting, washing
the ore with water, is not so efficient as is usually supposed. While gold chloride and
silver nitrate are both soluble, they are, after drying on the ore, usually reduced and
rendered insoluble. Seals are of little value, as a whole wax seal can be raised with a
knife, and later replaced intact. Cloth or paper bags yield to the hypodermic syringe
while even glass bottles will succumb to a hypodermic needle thrust through the cork.
i Min. and Eny. World, May 20, 1912, p. 1099.
The best means of avoiding the effects of salting is to render its determination easy.
This may be done by taking a few samples of utterly worthless waste rock in with one,
and later mislabeling them, or better still, by roughly dividing a sample in a mine, and
making the two portions of uneven weight. To salt such samples artistically enough
to raise the values of the two portions equally, can usually not be done in the time and
with the facilities at hand. Resampling is, of course, necessary when salting has been
discovered.
Salting in precious metal samples may also be done by use of gold under the finger
nails, or in the hair or beard, or by cigarette or pipe ashes carelessly (?) dropped into
the sample.
A method of salting practically impossible to detect, or rather a method of influenc¬
ing the sampling, is known as “drawing the center.” In this the men who are piling
up the cone purposely place the successive shovelfuls a little to one side of the apex
(always moving the new apex in the same direction) Eventually this gives a pile
with the original apex of the cone well to one side of the new and buried in the cone.
The quarter containing the original apex will then contain an undue proportion of fine
and if this quarter and its opposite are taken the sample contains too much fine;
if the adjacent quarter and its opposite are rejected the sample contains too much
coarse. These quarters will be either taken or rejected according to which way the
sample is to be thrown. A way of avoiding this is to place a small stick upright in
the center of the cone as soon as it is well started and seeing that the stick always is
kept in the center of the heap.
In general, in watching any sampling by the coning and quartering methods it is
advisable continually to compare the looks of the rejected heap with the accepted,
and if It does not look right, repile the entire sample. In general it should be remem¬
bered that in coning non-homogeneous material there is no mixing takes place, but
rather a classification, due to the large pieces rolling to the outside of the pile. Coning
and quartering is not a means of mixing, it is only a means of dividing a sample.
Ore sampling may seem unduly emphasized in the above remarks, but when
it is remembered how much money changes hands each year on samples of ore
and of copper bullion it can be seen that this part of the subject deserves the
space given to it.
Moreover, it is believed that the man who can take fair samples of ore and of
bullion, and can avoid being stuck with unfair samples when some one else is doing the
sampling, will be amply able to sample anything else that may be presented to him.
ADDITIONAL BIBLIOGRAPHY
Methods of the U. S. Steel Corporation for Iron Ore Sampling, Met. and Chem-
HNrtg.y February, 1909, p. 107.
Methods of the TJ. S. Steel Corporation for Technical Sampling and Analysis of
Gases, Met. and Chem. Eng., June, 1911 and July 1911.
Principles and Practice of Sampling Metallic Metallurgical Materials. By
Edward Keller, Bull, 122, IT. S. Bureau of Mines.
Theory and Practice of Ore Sampling. D. W. Brunton, Trans. A.I.M.E.,
VoL 25, p. 826.
Modem Practice of Ore Sampling. D. W. Brunton, Trans. A.I.M.E., Yol. 40,
p. 567.
SECTION XVII
FERMENTATION
By C. Glaser, Ph. D.1
Introduction.—While fermentation has been used since the days of barbarism

by men to alter the composition of liquids containing certain organic substances,
such as the juices of certain fruits, or artificial solutions extracted from grain,
no one until recent times had any idea of what was really the cause of changes
observed. Wine has been made for thousands of years and even beer has been
brewed for many centuries, but it was only with the advent of the microscope
that efforts to elucidate the mysteries became promising; and even then, the
explanations offered were not substantiated by the facts, as developed less than
70 years ago by Pasteur and his co-laborers. Between the years 1854 and 1871
Pasteur devoted a large portion of his time to these researches and his theory,
developed in opposition to the views held by other investigators (even a man like
Liebig being in the opposing camp), and slowly accepted after this period, has
stood the rigid tests of criticism.
It has been improved and enlarged upon since, by many investigators, but his
leading conclusions are as true to-day as they were then and his theories have enabled
mankind really to understand what is going on, when a sugar solution of medium
concentration commences to show a few gas-bubbles after it has been exposed to the
air, and in a few days goes over into an apparently boiling mass without much rise in
temperature, only to fall back into quiescence after a further lapse of time.
The gas given off had been found previously to consist mainly of carbon dioxide
and the liquid, left behind, to contain alcohol. A sediment, formed in the liquid,
called “yeast” was known to be capable of starting fermentation again in fresh sugar-
solutions and of repeating the cycle of actions above described.
Pasteur’s discoveries showed that certain organisms have a preference for certain
optically active organic acids, in so far as they would, for instance, consume dextro-
tartaric-acid, while they would not touch the levo-rotatory form. He developed from
this his theory of dis-symmetrical action of living microbes as a general principle.
In his report to the Faculty de Science at Lille, France, he was soon able to report the
systematic breaking down of potato-starch into sugar, alcohol (ether) and vinegar
by the actions of certain microbes, thus extending the previous realm of fermentation
processes. He also got a glimpse of the possibility of fermentation without living cells,
but ’with an extract therefrom, which was a forerunner of Buchner’s later discovery,
(about 1896-1897) that the extract of yeast cells, carefully freed from the cells and
constituting merely the liquid part thereof, would quickly decompose sugar in solu¬
tions into carbon dioxide and alcohol.
Fermentation a Catalytic Process.—This latest discovery enables us for the

first time to give a concise and exact definition of fermentation, as a catalytic decom¬
position of sugar by zymase into alcohol and carbon dioxide, zymase being the
enzyme extractable from yeast-cells. So far it has not been practicable to
1 Consulting chemist and biologist, Baltimore, Md.
585
isolate in quantity this enzyme and the discovery of it, while clearing up the
mystery of fermentation, has had no important practical results. We can state,
however, that the appearance of secondary products such as acetic, lactic and
butyric acids is due to impurities in the yeast in the shape of the bacteria corres¬
ponding to these products.
From the above researches, in which many other investigators were also active, it
has become possible to lay down positive rules for the production of alcohol on an
industrial scale from material either containing (1) fermentable sugar, or (2) com¬
pounds which can be converted by hydrolysis into fermentable sugars. To this
latter class belong the “grains” (wheat, corn, rye, barley, rice, potatoes) with their
considerable percentages of starch; also starch obtainable from other vegetable
sources, and to these materials has lately been added wood in the shape of shavings,
which can be partly hydrolyzed by dilute hydrochloric or sulphuric acid, acting at a
high temperature under pressure on all polysaccharides present, even when in the
shape of hemicelluloses and celluloses, but at a decreasing rate, as the last named
increase.
To class (1) belong the sugar cane, the sugar beet, wine grapes and many other
fruits. Sugar cane is the basic material for Jamaica rum; rice for arrac; while the
sugar beet is raised in European countries and also in recent years in the Western
States to a considerable degree for making the finer grades of alcohol. Potatoes as a
source of alcohol were proposed as early as 1682, but only since 1820 have they been
used for this purpose in large quantities, which however are lately reported to be
falling off again. The processes of converting starches and other polysaccharides into
fermentable sugar will be considered later in their most important modifications.
When we speak of “alcohol” we usually mean only the most important member of
a large group of chemical substances, comprised under that name as a class, namely
ethyl alcohol, which has the chemical formula C2H6OH and is also called “spirits of
wine.” It was prepared as early as the eighth century by distillation from wTine, as
described by Marcus Graecchus. However, only in 1796 was it first prepared free
from water bv Lowitz and by Richter.
THE PRODUCTS OF FERMENTATION
Enzymes and Ferments.—To the most important and interesting products oi

metabolism belong the enzymes or ferments. Since they are commonly ejectec
by the living cell, they are genuine “excreta.” But since they play the mosi
important part in the questions of nutrition of and addition of energy to the cell
they can not be compared with those elimination products which the cell discards
because it has no further use for them. They must rather be considered as £
tool with which the cells act upon their surroundings, in order to draw from then
the greatest possible advantage.
The following important functions are fulfilled by the enzymes: (1) to change tbu
character of the nutrious substrata in such a way that they are rendered fit for food
(2) to produce heat-energy by oxidations and exothermic decompositions, which ar
under the given relations a condition of life to certain microorganisms; or (3) they ma;
exercise both these functions together. Many things point to the fact that enzyme
participate in reductions and in syntheses. It may also be maintained that the sam
enzymes which produce exothermal decompositions, may under certain condition
carry on processes diametrically opposed, and that they are capable of rebuilding agab
by absorption of heat the molecules which they previously have decomposed.
FERMENTATION 587
So far we do not know what the enzymes are in a chemical sense. None of them
except diastase has been prepared in a state of purity and their constitution is yet
shrouded in mystery. We can only state that they are a part of certain living cells,
from which they can be separated by rupturing the cells and pressing out their proto¬
plasma. This latter substance is the ultimate carrier of the enzymes, as proven by
Buchner’s researches. Attempts to separate the enzyme from the cell-juice have not
led to encouraging results, for all attempts to purify a primary precipitate by repeated
precipitations with alcohol have led to a great weakening and finally complete dis¬
appearance from the product of its enzymatic powers. Pure enzymes are therefore
not yet known. The actions of enzymes are greatly influenced by temperature and,
as a rule, come to a stop only a few degrees above their respective optimum
temperatures.
Invertase, an enzyme occurring in many saccharomycetes and bacteria

splits the molecule of cane sugar by hydrolysis into two molecules of grape sugar.
C12H22O11 + H20 = C6Hi206 (dextrose) + C6Hi206 (lasvulose)
Its optimum temperature is between 52 and 57°C., while even at ordinary tem¬
perature some action takes place. The ferment is destroyed at about 70°C.
Maltase changes maltose into dextrose and malto dextrin. Its optimum
temperature is between 40 and 60°C. and it is destroyed at about 70°C.
Zymase.—Other, deeper splitting decompositions may be visualized by
considering its action on grape sugar: C6Hi206 = 2 C2H6OH + 2 C02. It is
Alcohol Carbon-
dioxide
very sensitive to a temperature approaching blood heat (35°C.). Its optimum is
about 10°C. The report has however been published by the Carlsberg laboratory
in Copenhagen that yeast cells, which had been washed and. dried in the air,
then dried in vacuo at 38°C. for several days, followed by 6 hr. drying at 100°C.
have maintained their power to cause fermentation without significant reduction
for 3 years, although all the cells were undoubtedly dead. It has thereby been
proven that one of the most typical and far reaching fermentations can take
place without the direct cooperation of living cells.
Diastase is the only enzyme which has been prepared in a high state of
purity. It is a white powder, soluble in water, which solution has the power
in a high degree to saccharify starch. According to Payers 1 part of it sacchari¬
fies 2,000 parts of starch, while according to Dubrunfaut even 100,000 to 200,000
parts will be so changed. In a dry condition it resists very high temperatures.
According to Krauch it may be heated for several hours to 120 to 125°C. without
loss of power; according to Hueppe it will even stand a short exposure to 159 to
162°C. Chemically, diastase appears to be an albumen of the nature of a peptone.
It does not pass by filtration through porous clay-cells. Diastase splits starch
into maltose and dextrin, it acts with great energy upon starch paste, which
becomes limpid and transparent in 1 to 3 min. even without application of heat,
and does not thereafter give a blue color with iodine; after 5 to 6 min. the red
coloration of erythrodextrin also disappears and after one hour the liquid contains
maltose and dextrin in the ratio of 80.4:19.6, which points to the following
decomposition:
C120H200O100 + 8H20 = 8C12H22O11 + 4C6Hi0O6

Starch Water Maltose Malto dextrin
While malto dextrin is not considered fermentable, Maercker states that in the
manufacture of alcohol this also will be fermented finally, because during the
long primary fermentation a modifying effect of diastase becomes appreciable,
by which the dextrin is hydrolyzed to maltose (or glucose).
The formation of certain acids, which always accompany, in small amounts the
products of fermentation, is probably also referable to the action of enzymes. The
following formulas will make this clear:
Glucose CsHÔe = 2 (CH2)2.OH.COOH = Lactic acid
or = 2 C02 + 2 H20 + C3H7.COOH = Butyric acid
or = 3CH3 COOH = Acetic acid.
. Considering the lethal temperatures for all organisms discussed that cause fermenta¬
tion, it will be seen that all fermentations can be arrested by heating the fermenting
solution to about 75°C. It is however also possible to stop fermentation by the addi¬
tion of strong acids, or alkalis, as yeast will stand only limited quantities of either.
Caustic alkalis particularly are very detrimental and kill yeasts quickly.
THE MALTING-PROCESS
The purpose of malting is the production of the enzyme diastase, which has
the power to convert starch into sugar and dextrin. Diastase is produced
during the process of germination of grain. In general malt may be defined as a
cereal, which has been subjected to the process of germination until the maximum
amount of diastase has been accumulated. Germination is a biological process,
during which the primary organs of the embryo (endosperm and acrospire)
develop at the expense of the nutrients stored in the grain. In order to start
growth it is necessary to supply water to the grain, which is done by steeping,
until 47 to 48 per cent has been absorbed. This condition of saturation must
be maintained during the process.
Germination may be compared to a slow process of combustion, inasmuch
as carbon dioxide and water are given up with the liberation of heat. The process
reminds us strongly of the process of respiration. The grain suffers thereby a loss
in weight of about 3 per cent of its contents of carbohydrates.
It is necessary to provide for good ventilation, in order to prevent the accumula¬

tion of too much of the products of respiration, as by a strong increase of carbon
dioxide in the malt pile the intensity of germination is considerably retarded.1 The tem¬
peratures should be regulated so that it does not exceed 18°C. at any time or place
inside of the pile of grain. The object is generally secu ed by taking up the malt by
a shovel and letting it fall through the air in such a way that the formerly lower grains
collect now on the surface and vice versa. If sprinkling should become necessary, not
more than half a pint of water per bushel should be applied on the fourth or fifth day.
The process is kept up until the acrospires show a length of growth equal to from
% to H of the grain. Usually in 7 to 8 days the desired results are obtained. The
grains are then dried in a current of air and are finally passed through a drying-kiln
where the grains are heated until the desired color is obtained on the outside, and alsc
the well-known malt-aroma is developed. These are particularly important when th<
malt is destined for brewer’s uses.
1 This is analogous fco the phenomena of “mass action” and equilibrium in ordinary chemical re
actions.—Editor.
n ii\u:\ l A T!uS rm
I h«’ m-unit icojjr »>i m At uma other cereals than barley or rye in carried union
r\ irfl* u,» -mu * ormeio!*“* and * 1*»*v-i no! require a separate descriptkm. Such malts
air •.••Mmu met AJh
CONVERSION OF STARCH INTO FERMENTABLE SUGAR
The Mushing Process, 1 nd*T this is understood (hr hydrolysis of

nfairli h\ f h** nc?a»u »*s di i.' fa o, In ?i well made malt flic* amount of diastase
pf *' * u* i> a a <■ * A «*!e •*» ft to c»r,\rrt all fin* starch present in if, hut it can
oum* ”4 ». rs *•' d * t ueh, d« mrd from other sources at (hr same* time*.
< *f *N. ’ * < f '<f* * o,r u ** i made in technical chemistry. Brewing with
*u iT md a n \ d''’o' wuh mu*»••»,* j\ he* described first, to bring out the* important
pi«m* »ils *’/ f a Or * »■«onpiu'ati«»jn
IN N« I* o!r i Hi u^'d ttifh the requisite quantity of water and the
!aOs r a on * d '** ih ^ u-pr* cun* t«t front do to I0T\ Ground malt, ahout
oia* 1 co*u : v ** cw <eh* . t t1 o i”. now started running into flu* kettle and
the *u j a . * • i'o*, , e in motion, to prevent tin* ground mult settling out
on f m i *•!''. 1 >. * i »* o ? n mi of tho kettle. When the whole charge is in, the*
feiij'f1 B*.*e m ** * ’ av t in- i Nought quickly tip to VTA \ and held there for
1 . hi I N o * * i 4no* m| fhe "in edi“ is now raised at tin* rate of \ p<»r
Him ♦** , m hi <u 1 t it k ‘bed, win st flie heat is checked, so as to hold tin* tern-
frtv,fio f kU nk Pi N‘pf o, nuf tl the M Break M occurs, wlnclt is a triidrury
of He o- * o * * « i* * s J* to ej»ii iti* from the wort (tin* solution! and w liich is
a ’ ;/n f‘t Po } t • a - * th J f l.o ronoi tun hi finished. An iodine test for starch
and rîN’N ? m N< AA be n»*g due mid thus verify tin* indication of the
*' NraL ’* I’i r \\"i * cmu? tin* now leimenltthie stignr and dc*\trin, hut uosturcli.
THE HOP-JACK
Ilrtwiiii, tit the brewing indii-Try tlitc next step after tlir production of
the M Wolf M i nee a hovel pt efferled in large, separate mpperdictf les for it two fold
purpose to fire tin* from atlniiidiiiiiH bodies, preeipifithle at boiling heat,
and to ifro- it tie* aeru-domed hop flavor and latter taste* by adding to flu* boiling
»«ff thr tcipn utr i|muitit\ of Incp'g ttlneh are the dried, ripe, female blossoms of
t|*f- enlinnf*‘d hop.plant For bretterv purjwer*i only the unfertilized blossoms
ran !«* no'd, a- ihr matur'd fertilized blossoms have it sharp, iu*rid taste*. Tlie
nioiiuif i»’ puinaph' $) uouuU roiitniiird in small rVi*llt*vv ghcbulen, itbout O.lh mm.
dmnnTnf hinHin ;e hopmenl or lufcnhu. din* female plant is propagated by
ruffing. In the* rnitutry it in iiiiisnly grown in the states of New York, Oregon,
thihioiiiia and W aslnngfon, while the very choieeif hops are still imported from
KtirojM* i HnUmm t
|fii|«#t ao* r«*4»gi4iiri| tw n preserving action on l«*crf nunh* Umn

llieir ft®mu funetiuii of giving the ehnnielerintte flavor* Tliis w nttributed to the prere
rnre *4 "hop" tnuii, wliirli linn fi bitter tiwte. Ilirre n nli+o a iumtm ihop-tanimt!
I»re»i«il# Ifthirh favor# fie* preei|nt.iiSi»*ii of albumen in the 1 coding wort. Attc*r this
tfeiitiiieiif the wml m e.#witrt| by |iw«ii»i it fiver iynteiiw of upright piiwti, niipplird
lm . dt t ^ 1 I mlas** ***** i# i** # Ik «!-«#« # *>** in *kf«*#*#»#** t #14 ♦ lui f s **•¥** mtf
When we now attempt to use the same principle in the conversion of the starches of
rice, grits, etc., we find, that it is necessary to cover the material with water and heat
nearly to boiling, stirring well, until its starchy contents have been changed to a
voluminous soft paste. In practice this is done under pressure in a separate converter,
followed by cooling, after which the product is transfered to the mashing-tun. About
one-fourth of the weight of the rice or grits in finely crushed malt is now added;
the temperature of the mass brought to about 40 or 42°C. and the whole intimately
mixed. It is often necessary at this stage to add more water, to thin the mixture to a
convenient consistency for stirring. After 1 hour’s standing, the temperature is
gradually raised 1°C. per minute until 70°C. is reached. Owing to the smaller
quantity of malt a little more time than in the case of malt alone will have to be
allowed on the finishing stretch. The break will also be somewhat less distinct and it
is advisable to check its progress by frequent iodine tests a few minutes apart, until
the starch has disappeared. The fact is clearly demonstrated, however, that the
diastase of malt has a much larger capacity for hydrolyzing starch to sugar than is
required for its own starch. The liquids (worts) produced by the two mashing-
processes1 described, contain the same classes of fermentable sugars and may be
indiscriminately used in the production of alcohol.
The last described mashing-process is however preferable for economic reasons,
since it represents a large saving of malt, the most expensive of the raw-materials used
for brewing purposes.
CONVERSION OF POLYSACCHARIDES INTO FERMENTABLE SUGARS
Acid Hydrolysis.—Of importance as a preparation for bacterial fermentation

is acid hydrolysis. If dilute acids, (hydrochloric, sulphuric, and also some
organic acids: oxalic, tartaric, citric) are allowed to act on starch at a tempera¬
ture close to the boiling point, a complete liquefaction of the starch takes place
in a short time; the solution, when cooled, gives a strong blue coloration with
iodine; alcohol produces a voluminous precipitate, which, separated by filtration
and washed with alcohol, will be found to be an isomeric modification of starch,
amidulin or soluble starch, which can be partly obtained in a crystallized form,
partly as an amorphous powder. Upon prolonged action of the dilute acid the
soluble starch disappears gradually and the liquid gives a burgundy-red coloration
with iodine, upon addition of alcohol a white precipitate is obtained; Fehling’s
solution shows reduction. At this stage the liquid contains ery thro dextrin and
a reducing sugar. When the heating is continued the amount of sugar increases
steadily, until finally no iodine-reaction can be obtained, the erythrodextrin has
passed over into acchroodextrin, which also gradually disappears and there
finally remains besides more or less sugar, according to the duration of the heat¬
ing, a body “Gallisin” (C12H24O10) (Schmitt and Cobling). The end product of
the action of the acid is always dextrose. The process may be represented by
the formula:
C120H200O100 + 2OH2O = 20C6Hi206
Starch Dextrose
The application of this process under pressure has become of commercial im¬
portance in the manufacture of alcohol from wood shavings, the high temperature
obtained permitting the hydrolysis of compounds closely related to cellulose,
with formation of fermentable sugar.
1 With malt alone, and a mixture of starch and malt.
FERMENTATION 591
ALCOHOLIC FERMENTATION
By fermentation is understood a sequence of highly characteristic processes,

which, with the aid of an active reagent present in relatively small quantity,
cause the decomposition or alteration of large quantities of organic substances.
The active reagent is called “ferment” but we know now that very different
kinds of ferments may enter into a reaction. They may be microorganisms or
bodies related to albumen. Therefore a distinction is made as between organized
ferments and enzymes. By fermentation in a narrower sense is now understood
such decomposition as may be caused by the activity of living microorganisms.
The process of fermentation involves three classes of bodies:
1. The substratum or substance to be acted upon;
2. The microorganisms, as the cause of the reaction by their life processes;
3. The products of fermentation.
The activation may be caused: (1) by budding yeast or blastomycetes; (2)
molds, or fungi, hyphomycetes; or (3) by bacteria or schizomycetes.
While all three classes of microorganisms may cause alcoholic fermentation, only
the class of blastomycetes is of controlling importance for the industrial production
of alcohol, while in acid-fermentations and in water purifications bacteria play the
most important part. We shall first consider alcoholic fermentation. The substrata
for alcoholic fermentation are various kinds of sugar, which by passage through the
bodies of blastomycetes, participating in their life-cycle, are decomposed into alcohol
and carbon dioxide and minute quantities of succinic acid and glycerin. The main
part of the process may therefore be presented by the following equations:
Ci2H220n(cane sugar) + H20 = 4C2II5OH + 4CO2
CeHisOeCglucose) = 2C2H6OH + 2C02
From 100 parts of cane sugar therefore 53.8 parts of alcohol and 51.46 parts of carbon
dioxide would be produced, if no other reactions took place, while from 100 parts of
glucose the corresponding figures are 51 parts alcohol and nearly 49 parts carbon
dioxide. These figures are, however, never reached. Jodtbaur gives the following
figures for actual results obtained.
Cane sugar.51.11 per cent Alcohol; 49.03 per cent C02
Maltose (anhydrous).51.08 per cent Alcohol; 49.04 per cent CO2
Dextrose.48.67 per cent Alcohol; 46.54 per cent C02
The figures for cane sugar agree with those found by Pasteur. The microorganisms
predominantly employed in alcoholic fermentation are known as “cultured yeast/’
which consists of immense numbers of cells of saccharomyces cerevisice I (Hansen).
Under the microscope this organism is shown to be a slightly elliptical cell, growing in
groups. In order to avoid untoward reactions and obtain constant results, the use of
this yeast, raised under proper conditions from one single cell, has been successfully
introduced in the technical field, particularly in the brewing industry. By rigidly
excluding acid-producing bacteria, the formation of more than traces of acetic, pro¬
pionic, lactic and succinic acids is successfully avoided.
In the distilling industry the same care in regard to yeast need not be exercised,
since the fermentation is followed by distillation, which is so regulated that most of
the impurities, including higher alcohols, produced by microbes, are eliminated.
Moreover, in this latter case the formation of certain esters is desirable, as they play an
important part in the development of the final aroma. A slight acidity of the freshly
distilled spirits is therefore necessary, to induce formation of esters, by combination

with alcohol during the storage of the distilled spirits in wood. Ethyl acetate is the
principal ester formed.
YEAST
All microorganisms are distinguished from other organisms as being of a

simple structure, which consists in one cell for the individual. In spite of the
apparent simplicity they possess as a class an astonishing amount of energy
which manifests itself in the causation of fermentation, in putrefaction and in
the production of infectious diseases, as the case may be. They are essentially
plantlike and may be divided, as has already been said, into three main classes:
1. Mold fungi or Hyphomycetes.
2. Yeast fungi or Blastomycetes.
3. Bacteria or Schizomycetes.
We shall here consider the second class, as used in the fermentation of worts, the
products of the mashing-process, which has been previously described. It is
the general aim in the industries here comprised, to have a reliable, powerful
yeast, which will do the work of producing alcohol from the sugars presented to
it, in the minimum of time and without leaving any sugar behind, i.e.} do it
economically. This problem has been of such vital importance to the industries
of brewing and distilling, that even long centuries ago conditions had been worked
out empirically, how to protect these industries against untoward (infectious or
other) influences, which might impair the efficiency of their most important
laborer. It has been left however to the ingenious studies and labors of scientists
of the second half of the last century, among whom I may name Pasteur,
Liebig and Buchner, to reduce the problem successfully to a solid basis which
has placed the technic of the industry in a position to regulate the processes
at will and at all times.
In the first instance a certain mechanical purity must be aimed at. A first class
yeast must only contain cells of saccharomyces cerevisix I to the exclusion of all of the
numerous varieties of this species and further it should not contain any other kinds of
microorganisms, particularly no bacteria. The only way to reach this point is to
raise the stock of yeast from one single cell of the required species, under exclusion of
any and every chance infection.1 This is accomplished by dividing a small quantity,
say 1 drop, of yeast in a large quantity of sterile water, say 1 quart, under aseptic
conditions. After closing the flask with a tight-fitting, sterilized glass stopper, the
mixture is shaken until homogeneous. Then small quantities of this water, 0.1 c.c.
each, are transferred by a sterilized platinum-loop to sterilized, cotton-plugged test-
tubes, containing wort-gelatin, which has been previously liquified by placing the
tubes in water of 30°C. just for a sufficient time to melt the gelatin. The technique
of this is as follows. The gelatin-tube is held between thumb and index finger of the
left hand in as nearly a horizontal position as the contents will permit; the cotton plug
removed and placed between the third and fourth fingers of the left hand. A platinum
loop (fused into a piece of glass tubing) is then grasped with the right hand, heated in
a gas or alcohol flame, and cooled. A helper now lifts for a moment the stopper of
the flask, containing the diluted yeast, the platinum loop is inserted with the right
1
The processes of producing a pure culture are described in detail, since it is obvious that the method
is of general application and that it is the basis of all fermentation industries.—Editor.
FERMENTATION 593
hand, the arop removed and transferred at once into the open gelatin-tube. The
helper immediately closes the yeast flask, while the operator replaces the cotton-plug
with the right hand in the gelatin-tube. These manipulations are repeated for every
gelatin-tube that it is desired to seed. The tubes are now shaken and each emptied
into separate, sterilized, cool, Petri dishes. The covers are replaced at once, the gela¬
tin distributed evenly by gentle tilting and the dishes removed to an icebox, where the
gelatin quickly solidifies.
The various cultures in the Petri dishes should be examined daily looking through
the cover, without raising it. After several, usually 3 to 4 days, one may see in some
or all of them little white dots. Each represents a culture from one cell. While
waiting for them to grow to distinct visibility, one must prepare a number of sterile
flasks, say six or a dozen, containing clear, sterile, sweet wort; or if such is not at hand,
a clean sterile sugar solution of about 10° Bal. density, containing some nutrient salts.
For this purpose dissolve in every 100 c.c. of sugar solution 1 g. ammonium nitrate;
0.5 g. potassium-hydrogen phosphate, HK2P04; 0.05 g. precipitated calcium phos¬
phate (Ca4HP04); and 0.5 g. magnesium sulphate, cryst. From the flasks are selected
a sufficient number to make cultures from the Petri dishes. A piece of platinum wire,
rather stiff and about 1 in long, is fused into a piece of glass tubing as a handle.
A Petri dish with a round dot or dots is placed near one of the flasks, containing nutri¬
ent solution. A lighted Bunsen-burner should be at hand. The platinum wire is
sterilized and cooled, the cover of the Petri dish lifted, the small culture selected
quickly which looks healthiest; touched with the point of the platinum wire, so that a
little bit of the colony will adhere to it, the wire withdrawn from the Petri dish and the
lid replaced quickly. The cotton plug is then removed from the sterile flask, the tip of
the platinum wire lowered into the liquid, stirred about for a few seconds, withdrawn,
and the cotton plug replaced quickly. With one assistant one can easily make a dozen
inoculations in a couple of minutes. The flasks should be kept in a room at about
60 to 70°F.; in hot weather placed in an icebox. If observed daily, one or another will
show signs of fermentation beginning after a day or two, which will increase and come
to an end in about a week. These are the first pure yeasts to start a supply from.
In order to raise it from small beginnings one must have a yeast-propagating apparatus
of sufficient capacity for one’s requirements. Before starting the yeast in it, it is
necessary to make a careful microscopical examination to make sure that it is really a
pure culture. If this should not be the case, another of the little fermenting flasks,
should be tested and if necessary another, until one is found which is surely a pure
culture of saccharomyces cervisicoe I. If one fail in this, one must discard the whole
work so far done and start all over again, until one surely has a pure culture. The
propagation of the pure culture on a commercial scale-must then take place in one of
the standard forms of propagation apparatus, of which Hansen’s is the prototype.
This will be described in detail in order to show the essentials in its construction.
He describes it in “ Practical Studies in Fermentation” as follows: “As shown in
the accompanying illustration the apparatus consists of two main parts and the
connecting tubes, namely: The fermenting cylinder C and the wort-cylinder D. The
details of the airpump and airholder are not shown in the drawing. The airpump A
is driven by machinery and draws air through a filter in order to effect a preliminary
purification. The airholder B is provided with a pressure gauge and a safety valve.
It is charged with air under a pressure of from 1 to 4 atmospheres. The pipes must be
fitted with cocks at suitable points for removing water, which will collect in them.
This is of especial importance in the case of the pipe between the airholder B and the
filters, g and m. These are best united by metal tubes with the air pipes. If metal
tubes are used they should naturally possess some degree of elasticity and must be so
arranged that the filters can be readily fitted and disconnected.
“Through the top of the fermenting-cylinder C passes a stirrer 5, the lower end of
which is fitted with two blades, one carrying a sheet of rubber in such a way that when
rotated it comes into contact with both the bottom and the sides of the cylinder.
From the top there passes a doubly bent tube c, and by opening its cock, connection
is made with the inside of the cylinder. The lower free end of the tube dips under
water in the vessel d.
“A little below the top is a horizontal tube e provided with a cock, by means of
which the inside of the cylinder is connected with the vertical glass tube /. This is
connected at its upper end with the filter g and at its lower end with a second cock and
similar horizontal tube h to that described above. The top mark on the glass tube
is 79 cm. from the bottom of the cylinder, the next 20 cm., and the lowest 10 cm.
When filled to the top mark, the cylinder holds about 1% barrels or 46.5 gal. or 176 1.
The glass tube is fixed into the cocks e and h by a packing of hemp or cotton wool with
vaseline; rubber is not suitable, as it is hardened by steam. The filter g consists of a
metaflic capsule containing a tightly packed plug of cotton wool 22 cm. long and 3 cm.
in diameter. The plug contains at least 35 g. of cotton wool; the addition of a little
more is immaterial. If firmly pressed in, the capsule will hold 50 g., but that much is
unnecessary. The filter is closed above by means of a cover, which is screwed on and
which is connected with the tube from the airholder. Before the fitter is screwed on, it
is sterilized by heating it for 2 hrs. at a temperature of about 150°C.
At the opposite side of the cylinder there is a small tube j scarcely 1.5 cm. long
fitted with rubber tubing, the latter being closed by a pinch cock and a glass stopper.
Passing from the bottom of the evlinder is a tube fc, through which connection can be
FERMENTATION 595
made with the wort cylinder D; this tube is made in two pieces to prevent too great
rigidity and in addition to the two large cocks shown it is provided with two smaller
ones, which are made use of during the process of steaming described below, partly for
running off the condensed water and partly for introducing steam.
The cock shown at l is or withdrawing the beer and the yeast. The valve is screwed
down in opening the cock and is screwed up when this is closed. In the figure it is
closed. Its construction prevents infection from occurring while the liquid is being
drawn off as the liquid cleanses the cock on passing through it. The pipe carrying the
cock is passed through the side of the cylinder and is bent toward the bottom, its end
being 3.5 cm. above the latter. It is, in short, so arranged that no air from without
can enter the cylinder, while its contents are being drawn off.
“The wort cylinder D must be raised somewhat above the level of the fermenting
cylinder. (The wort can, of course, also be forced into the fermenting cylinder by
means of compressed air, but in this case the wort cylinder must be provided with a
safety valve.) Its height is also greater than that of the latter, but its diameter is the
same. At the top is a filter m exactly as at g, and connected with it is a pipe (indicated
by dotted lines) passing inside the cylinder. The lower closed end of this pipe has
some small perforations through which the air finds an exit after passing through the
filter. The tube n corresponds with the tube c of the first cylinder, and like the latter
its open end dips into a vessel of water o. In the case of the wort cylinder it is very
important that the bore of the tube n, and of its cock should not be too small, in order
that they may not become choked by hops and other matter; a suitable diameter for
the tube is 1.5 cm. Around the upper portion of the cylinder, a little below the top,
there is a pipe in the form of a ring p, the inner side of which is provided with small
perforations. One end of this pipe is closed and the other is connected with a cold-
water tap. In addition to the cocks on the connecting pipe k between the two cylin¬
ders, the wort cylinder has three others q, r, $. The cock $ is for the introduction of
the wort and is put in connection with the wort main U between the copper and the
cooler. The cylinder stands in a shallow tray provided with an outlet t for the water,
which flows over the sides of the cylinder while the latter is being cooled. The
dotted lines at t show the bars on which the cylinder rests, and also the ring-like
portion and bottom of the cylinder.
“If the fermentation cylinder is not standing in a room with even temperature it is
necessary to arrange it in such a manner that the temperature of the liquid contained
in it can always be controlled, and that it can be lowered at will.
“This is done by means of the jacket, shown around (7, covering not only the sides,
but also the bottom of the cylinder; the bottom of the jacket is provided with screws
and can be removed, when it requires cleaning. For the introduction of a thermome¬
ter there is a tubular aperture through the jacket and the side of the cylinder.
The jacket is provided with a tap near the bottom, forming the inlet for cold water,
and another near the top and on the opposite side for its exit; a third tap at the
bottom serves for removing the sediment which is gradually deposited by the water.
The wort cylinder is here also provided with a jacket, which however can very well
be omitted, as the perforated ring serves the same purpose sufficiently well. Never¬
theless the jacket has the advantage that it encloses the water from the ring so that
the operator is not liable to receive splashings. It adds, however, considerably
to the cost of the cylinder and makes it more cumbersome to manipulate. The middle
portion of the cover is made of copper and is provided with a brass flange with 12
bolt holes. Between the cover and the collar' of the cylinder a rubber washer is
inserted and fits into a groove, a perfectly air-tight joint is thus insured.
“In order to prevent the stirrer from being raised out of its bed at the bottom of
the cylinder while in use a ball socket is provided. The axis ends in a ball which rests
in a hemispherical socket and two pieces, accurately fitting the upper portion of the
ball are bolted on; the axis can be rotated but cannot be raised from its socket. With
regard to the tinning of the cylinder, it must be pointed out that the tin should not
contain an appreciable amount of lead. If this should be the case, the yeast grown in
the apparatus will, according to Prior, be unsatisfactory.
“In putting up the apparatus it should be borne in mind that it should remain
undisturbed in its position. When possible, it will be best to place it in the fermenting
cellar. There is then, as a rule, no trouble with the regulation of temperature and in
drawing off the beer and the yeast there will also be less labor involved, for those
employed in the intervals do other work close at hand. If the temperature in the
cellar is below 6°C. it is advisable to have the fermenting cylinder jacketed.
In putting up the apparatus it is, of course, necessary at once to consider whether
one or two fermenting cylinders are to be employed; in any case a single wood cylin¬
der will suffice. The apparatus having been erected, it is necessary first to test
the cylinder in tightness. Steam is cautiously introduced through K, while all other
cocks are closed: also water under pressure may be employed.
Sterilizing the Apparatus.—Before the apparatus is set working, the two

cylinders must be thoroughly sterilized, also the connection between them and
the pipe through which the wort is to pass to the wort cylinder. This is accom¬
plished by passing a strong current of steam through the system. The filters
are sterilized, as previously described, in a sterilizing oven. The fermenting
cylinder is sterilized by steam through one of the cocks on pipe K. While high
pressure steam is passing the different cocks are alternately opened from time
to time, so that the steam can escape through them as well as through the bent
tube c; this operation requires half an hour. Shortly before this the filter is
screwed on and then all cocks are closed except the one in the bent tube c. Simul¬
taneously the cock of the filter is opened to allow air through it, g, and tube h
into the cylinder. The latter cools as the air enters and the steam is gradually
turned off. The cooling is thus effected by the current of air, which mixed with
steam escapes through bent tube c. As long as steam is seen to escape, the vessel
d with water is not required; this functions only at a later period as an indicator.
If the steam were shut off suddenly, there would be danger of the filter not
admitting sufficient air to prevent a reduction of pressure due to cooling and the
result would be the inspiration of impure air into the cylinder or the collapse of
the latter by the external pressure of the atmosphere. Under the conditions
mentioned and at the usual temperature of a fermenting cellar, the cooling
takes about 2 hrs. With regard to the small water vessels d and o surround¬
ing the bent tubes, their only function is to indicate the direction of the air
current, be it outward or inward.
Operation of the Apparatus.—The wort cylinder and its pipes s, q and k are
sterilized in the same manner, but the process of cooling is here omitted. When
the steaming is nearly finished, the cock of the air filter is opened and wort is
admitted. The latter is the ordinary hopped lager-beer wort, which has
been sterilized by boiling in the copper, and is run as hot as possible through
pipe u and cock s into the cylinder. Shortly before the steaming is finished the
pumping of the boiling wort on to the cooler is commenced and 10 min. later
cock s is opened. The wort is allowed to flow into the cylinder until it reaches
the upper cock q} when cock a is closed. It is advisable to place a small bucket
under cock q to catch any wort which may run out. When this occurs the
cylinder contains the desired volume of wort. The hot steam and air escape
FERMENTATION 597
partly through q, partly through bent tube n. It is advisable to run off the first
small quantity of wort, which enters the cylinder, through cock r, as it is mixed
with condensed water, which gives it a disagreeable taste. When the cylinder
is finally filled with wort as desired, cocks q and s are closed. Air stjrilized by
passing through the filter is now forced through the hot wort before cooling is
commenced, and this aeration is continued during cooling. Generally a pressure
of from 1 to 2 atmospheres in the air tank suffices. It is merely necessary that
the sterile air in the cylinder should always exert a slight pressure over that of the
atmosphere, thus preventing the drawing in of any impure air, and to insure the
taking up by the wort of the fullest amount of oxygen possible. It is evident
that the operator must not forget to first open the cock n. If this is not done,
there is a risk of damage to the apparatus.
“As soon as the wort is ready for cooling, the perforated ring p is connected with a
water tap and the sprinkler allowed to play against the outside of the cylinder until
the temperature of the wort is reduced to about 10°C. In an ordinary fermenting
cellar this takes about 1 hr.; further cooling must be effected by ice water. The
air is passed through the wort continuously, and in escaping through the bent
tube carries some wort along; the rousing of the wort produces much foam, but this
never gives rise to contamination. The aeration must, however, not be too vigorous
or there may be too heavy a loss of wort. It is only when the wort has cooled to
11 to 12°C., that foam passes through the tube, which may be moderated by pouring
warm water in vessel o. The wort, now ready for fermentation is transferred through
pipe k to the fermenting cylinder.
If it is desired-to free the wort from most of its suspended matter it may be allowed
to settle for an hour, before drawing it over. To guard against the entry of impure
air the filter must be left slightly open during this time, at the end of which a consider¬
able amount of solid matter will have settled out, of which only a small portion will
be carried over, as the mouth of pipe k is at a sufficient height from the bottom of the
wort-cylinder.
The wort introduced should not reach above the small tube j, through which the
yeast is introduced. The yeast is previously collected in large two-necked glass
flasks or tin cans and in the transferring operation a bunsen burner or a spirit lamp
may be made use of.
“The stirring apparatus is now set in motion and the yeast mixed well with the
wort. As soon as this is done any remainder of the wort is added until its level rises
to the upper mark on glass-tube/, the volume then measuring about 1H bbl. The
column of liquid in this tube is forced by pressure of air, passing through the filter into
the cylinder, the cock on the upper horizontal tube e being closed and the cock bn the
lower tube h opened. When it is not desired to continue aeration during fermentation
the latter cock is closed, but only after the cock above the filter has been closed.
“After about 10 days the desired portion of the newly formed yeast can be drawn
off. It is here assumed that the cylinder has been exposed to the ordinary temperature
of the fermenting cellar; if the temperature has been higher, the yeast will be ready
for removal in a shorter time.
“The beer is run off at cock l and when froth appears this is closed. Some wort
from the wort-cylinder—which by this time has been recharged with wort for a new
fermentation—is now passed in until the level rises to the second mark from the
bottom on glass tube /. The yeast is now well stirred up by means of the stirring
apparatus, and the mixture of yeast and wort is drawn off into a perfectly clean
vessel (cleansed with hot water and then steamed). When the level of the liquid has
Bunk to the lowest mark on the glass tube, the cock is closed and wort again run in to
the second mark. The yeast is again stirred up and drawn off to the lowest mark; tl
amount withdrawn now measures about 13 gal. The portion remaining behind
sufficient to start a new growth.
It is advisable to have two marks in the vessel, into which the yeast is drawn o
one indicating 6]4 gal. and the other 13 gal. Great accuracy is not required. Tl
yeast obtained is sufficient to pitch 8 bbl. of wort and a new fermentation is start*
as soon as possible in an ordinary, well cleaned, fermenting vessel. If this can not
done at once, the vessel containing the yeast must be covered over and set aside in
cool and clean place.”
ALCOHOLIC FERMENTATION
Practical Applications.—We have now discussed all the preliminary process

and their products up to the process proper of fermentation and shall n(
describe the latter on a broad basis. i.e.} under such conditions as are essent
for the production of alcohol, no matter what its final use may be.
In the fermenting cellar usually a temperature of from 1 to 2°C. is maintained,

warm weather by artificial refrigeration. We start with 10 to 12 lb. of pure culti
yeast and 8 bbl. of a wort of a gravity of 14°Balling (sp. gr. 1.0572) of which abc
10°Bal. represent its contents of fermentable sugar.
The wort is contained in a cylindrical wooden tub, about two to three times
high as its diameter and filled to about 6 in. from its top, which is open. We n
“pitch” the yeast, i.e., we introduce it, after diluting it with its own volume of wc
at the top and cause its thorough distribution throughout the wort in the tub, best
agitating with a lively current of filtered air. After 18 to 24 hr. some bubbles of
may be observed to rise to the surface. During this stage of the process a good dea
heat is generated in the fermenting tub, causing a gradual rise of temperature to ab
8°C. On top of the wort a head of material consisting largely of albuminoid mal
gathers and should be skimmed off. When removed, the surface of the wort shov
fine white froth. The fermentation now increases rapidly in strength and becoi
rather violent during the third and fourth day, when large masses of foam collect
the top. From about that time the activity diminishes and comes apparently to
end after 8 to 14 days, when the foam collapses and the liquid beneath becomes vis
first in patches, then generally. It is now cooled to 4°C.
This is the end of the main fermentation. When now, after the yeast has sett
the fermented product is removed to a storage vat or chip-cask, it will be found, t
at 1 to 2°C. a gentle after fermentation (of malt-dextrins) takes place, sustained
the floating yeasts.1 This is however only important in beer brewing, as it h£
beneficial influence upon the aroma and taste of the beer. When this after-fermei
tion is over an examination of the product will show a specific gravity of about 1
the matter in solution will be about 2.60 per cent and alcohol from 5.5. to 6 per c<
The product of the fermentation will measure about 230 gal., which will yield u
distillation about 25 to 27 gal. of “proof” spirit or its equivalent of any des
strength.
DISPOSITION OF YEAST, AFTER FERMENTATION IS FINISHED
The disposition of used yeasts in the breweries is a sore spot in their econo
Of course as long as yeast maintains a condition of relative purity, ascertai
by microscopical control, it may be used again until a point is reached whei
1 Of which sufficient always remain suspended to start the process again.
FERMENTATION 599
better be discarded, or until it is pushed out of line by the pressure of newer

cultures; for the multiplication of yeasts during fermentation progresses at an
enormous rate. It is finally disposed of by washing it down the sewer. If it
were not for the bitter hop taste, which clings to it tenaciously, it might find
use as an important adjunct to cattle foods, or it might conceivably even be used
as a concentrated human food. So far no practical way is in sight to make such
uses possible.
The case is different with yeast from distilleries, which is entirely free from this
objection and is consequently eagerly bought by manufactures of yeast-cakes for
baking-purposes.
FERMENTATION OF GRAPE-MUST
In the art of making wines from fruit juices the processes usually applied are
more primitive. The unit for a fermentation is the barrel which may be of
great diversity of size. The barrels are set up in the well known type of wine-
fermenting cellars, which also vary greatly in size and capacity. They must
have a fairly cool uniform temperature, but are not artificially cooled. No yeasts
are raised in cultures, but the germs are either allowed to enter with the must
coming from the fruit-press, having been derived from the skins of the fruit,
where they always are found; or the introduction may take place from the atmos¬
phere of the cellar, where frequent fermentations have previously taken place
and where yeasts and other microbes are always present. The process of fer¬
mentation is under such conditions as to be dependent largely upon accident,
time, character and strength of infection, and its- cycle usually represents the
time of several months, instead of about 2 weeks, as stated in the preceding
sections. The hazard of introduction of undesirable yeasts or dangerous bacteria
may prove a great danger to the product.
The fermentation is slow in starting and will only gradually assume livelier pro¬
portions, in so far as the development of a sufficient quantity of yeast in the casks
progresses more or less slowly. The predominating desirable yeast is saccharomyces
Fig. 2.—Typical cultures. X 500.
A. Saccharomyces ccrevisise I. (The two cells marked X are dead.) B. Sacchar-

myces apiculatus. C. Pediscocci and sarcinse, lactic- and acetic-acid bacteria, clostridum
butyrici and wild yeast.
apiculatus. Its cells are smaller on the average than saccharomyces cerevisice,
ellipsoidal (lemon shaped) in form, with the ends extending and frequently rounded
off, bulb-like. It ferments only dextrose. Since it does not form spores, it is really
not a true saccharomyces. It occurs commonly on ripe fruits and almost always in
wine-yeasts. Admixtures of saccharomyces ellipsoideus Hansen or allied species are
frequently present. Besides these yeasts generally a multitude of other micro¬

organisms are found, since anything which will settle on fruits or berries has a good
chance to get into the cellar with the must.
According to geographical location these mixtures of microorganisms will vary and
so will the product. This is perhaps the reason why wine fermented in certain cellars is
likely to show fairly constant characteristics of “bouquet” and “aroma” for many
years.
These observations may even be extended to all wine-producing countries, like
France Italy, Greece and many others, whose products differ materially from the
standpoint of the “connoisseur,” although their grapes do not vary much. Still more
strength is given to these deductions by the fact that, when foreign grapes are imported
into a country, to be there used to increase the wine production, they do not produce
the kind of wine usually produced in their home, but they become “naturalized” and
produce wines identical with or at least very similar to the wines of the country where
they are fermented. It is often quite impossible to distinguish the two products.
There is therefore a hint at the possibility of producing wines with characteristic
bouquets at will by using yeast-mixtures of known varieties, containing also known
species of bacteria.
Experiments on record, made on an industrial scale, a number of years ago, have
proven that it is even possible to produce by a suitable selection of yeast-mixtures,
from a wort made from grains, wine-like products, showing characteristics of a grape-
wine, produced with the same yeast from grape-must. Products offered in various
markets under the name of malt-wines belong to this class.
The conclusion may therefore be drawn in a general way, that, while the alcoholic
strength of a wine will correspond closely to the fermentable sugar furnished by the
grapes, its aroma (bouquet) depends upon a secondary action and is referable to the
building up of certain esters and related bodies, by a synthesising action of micro¬
organisms, not necessarily yeasts, but probably predominantly of certain acid-forming
bacteria, which in the presence of alcohol, can also form esters. This last process of
forming the bouquet is much slower than the process of fermentation and requires
several years at the relatively low temperature of wine cellars.
In order to control the ripening process the action of the acid-forming bacteria
must not be allowed to proceed without a check. As it is impossible to arrest their
activity by heat, the purpose is usually attained by introducing at the proper time a
moderate quantity of a solution of isinglass in a restricted quantity of water, so that it
forms a stiff jelly. It is allowed to spread over the level of the wine and to sink slowly
to the lowest portion of the horizontal cask, carrying along all matter in suspension.
The process, above described, may serve as a model for other fruit juices, like
apple-cider or pear-cider, to be converted into their respective “wines.”
Lactic Add Fermentation.—The spontaneous curdling of milk, which takes

place regularly upon the latter standing exposed to air, is a separation of casein
by lactic acid. Lactic acid, which is not contained in fresh milk, forms by
decomposition of lactose (sugar of milk) under the influence of organized ferments,
the most active of which has been named lactic acid bacteria. A number of
different bacteria may cause lactic acid fermentation. The organism shown
in the accompanying illustration however has this faculty in such preponderating
degree that it is known under the name of Bacillus acidi lactici.
As in the case of any fermentation the “optimum” conditions for the multiplica¬
tion of the fermenting organism must be supplied. A pure solution of milk sugar
exposed to the air does not ferment.
The lactic acid ferment requires for its nutrition, like yeast, albumen and certain
FERMENTATION 601
salts, mainly phosphates and chlorides of sodium, potassium and calcium, which are
offered in milk in abundant quantities.
As in the case of alcoholic fermentation, which takes place only slowly at low
temperatures, the lactic acid bacteria develop slowly at low temperatures, so that by
sufficient cooling the acid-formation can be much retarded. Like alcohol yeast the
lactic ferment is destroyed at a higher temperature. At 45.4°C. lactic-acid fermenta¬
tion is arrested.
Lactic-acid bacteria are also very sensitive to their own product, so much so, that
the formation of lactic acid will stop when a liquid contains 1.6 per cent of lactic acid
(Richet). Kueppe describes the bacillus acidi ladici as oval bodies, which under
Fig. 3.—Lactic-acid bacteria. X 400.
strong magnification resolve themselves into short, stout rods, 1 to 1.7/* long and
0.3 to 0.4/i thick. They are mostly found in pairs. Before partition they may
become as long as 2.8m- They form spores. Micrococci forms are never seen.
At temperatures below 10°C. the bacillus does not develop. The optimum is
between 35 and 42°C., at 44.5 to 44.8°C. retardation sets in and at 45.4°C. its activity
is arrested (death point).
Besides lactic acid and a little carbon dioxide no other products are formed.
According to Kueppe only 0.8 per cent lactic acid as a maximum is formed. Lintner
gives as the most favorable temperature for the development of the ferment and the
production of acid, 40 to 50°C. He gives a concentration of about 1 per cent of lactic
acid as the limit of growth and fermentation.
If continuous production of lactic acid is desired it is therefore necessary to keep
the medium neutral by the presence of an excess of calcium carbonate, so that all lactic
acid will be converted into the calcium salt, as fast as made.
Lactic acid bacilli are almost always present in the air and can be readily attracted
by any suitable medium, as wort or glucose solutions. The decomposition of glucose
by lactic acid bacilli may be represented by the following equation:
CflHiaOa + 2O2 = 2C2H603. -f- 2CO2

Gluoose Laotic acid
Byproducts except a little carbon dioxide do not appear in this process.

The process in its technical execution is simple, when once properly started. To
begin with a “pony” must be allowed to absorb lactic acid bacilli from the air, or
better, must be infected from a pure culture. After being kept for several days at a
temperature of 40 to 45°C., best 42°C., the sugar solution will be found in a lively
state of fermentation. It may be kept thus, until the acidity indicates about 1 per
cent, or a little over, of lactic acid.
Meanwhile a fermenting-tub is charged at the bottom with powdered calcium
carbonate, enough to neutralize the maximum of lactic acid expected, and is then filled
with wort or sugar solution (glucose) to about 6 in. from the top. The fermentation
room should have a temperature of not less than 40°C. and not over 45°C., best 42°C.,
and must be kept within these limits. The pony is then transferred into the ferment¬
ing-tub and well distributed by stirring.
The fermentation is allowed to proceed until it ends by exhaustion of the nutrient
solution, which will be the case in from 7 to 10 days. The lactate of calcium is then
separated from the waste liquor, of which a sufficient portion is set aside as the next
J‘pony,” while the balance is discarded.
The lactate of calcium is washed free from adhering mother liquor with a small
quantity of clean water and is then decomposed by a calculated quantity of dilute
sulphuric acid of about 25°B6. strength. To every hundred pounds of the lactate
176.6 lb. of 25°BA sulphuric acid will be required. The solution of lactic acid is
separated from the gypsum in a filter press and is further concentrated by evaporation
to any required strength. It has been usually the practice to concentrate the acid to a
specific gravity 1.24, but commercial lactic acid is often found in the market with a
specific gravity of 1.21 which corresponds to 75 per cent of lactic acid.
ACETIC ACID FERMENTATION
Vinegar from Alcohol.—It is not intended to discuss the possibilities of synthe¬

sizing acetic acid from all the substances wherefrom chemistry has succeeded in
producing it, but the practical standpoint will be adopted of explaining briefly
the chemical changes which the most important materials have to undergo in
order to furnish the acid and then to describe the production of acetic acid in
the usual, dilute form of vinegar by a peculiar process of fermentation, caused
by certain bacteria, mainly by bacillus aceticus. The age of the process does not
detract from its intrinsic interest, as even today some points exist in it which
would stand further elucidation. The starting point is the same class of material,
which has been before discussed in the manufacture of alcohol, to wit, saccharine
solutions whose sugar has been split by fermentation into alcohol and carbon
dioxide. The starting point in a narrower sense is alcohol and its continued
further fermentation may be visualized by the formula:
C2H5OH + 20 = CH3, COOH + H20

Alcohol Oxygen Acetic acid Water
In the destructive distillation of wood and similar organic material much acetic
acid is formed, which before purification is then known as pyroligneous acid. After
purification, usually by production of the calcium salt, which forms an important
article in trade, the purest kind of acetic acid is produced, which solidifies below
16.7°C. to a beautiful mass of crystals, whence it is known as glacial acetic ^cid.
Most of the rather dilute solutions of acetic acid known as vinegar are made from
grains or fruit juices, apple juice being a favorite among the latter. The grains are
first subjected to a mashing process and an alcoholic fermentation, while fruit juices
are directly subjected to the oxidizing process. The strength of the alcohol should
not be over 10 per cent, but is often much weaker. In practice from 4 to 7 per cent is
often found suitable.
FERMENT A TION 603
The above formula shows the conversion to be a process of slow and incomplete
oxidation caused by the life-process of mycoderma aceti or lately more frequently
called Bacillus aceticus. Mycoderma species are to be found everywhere in the air.
If they come in contact with substances which support their lives, they commence to
multiply rapidly. The best conditions for their growth are furnished by an aqueous
liquid containing about 4 to 5 per cent alcohol, 1 per cent acetic acid and some nutrient
salts (ammonium phosphate, potassium, phosphate and magnesium phosphate, about
0.1 per cent of each). Sometimes good results are obtained by adding infusions of
bread, malt, or raisins, instead of the salts.
The temperature of the fermenting room is kept at between 25 and 32°C. as an
optimum. During the process a great deal of heat is generated and it is advisable
not to allow the temperature of the solution to go beyond 35°C. as this will cause loss of
alcohol. At 45°C. the process stops by the killing of the active organism. Therefore
means must be provided to keep the temperature within above limits, in winter by
heating; in summer by lively ventilation. Losses caused by neglect in this direction
may be as high as 20 per cent. The duration of a vinegar-fermentation depends
largely upon the temperature. Between 22 and 30°C. it is given as 8 to 16 weeks, at
30 to 35°C. as 4 to 8 weeks and at 35 to 38°C. as 2 to 4 weeks.
The following description is given in order to show a simple apparatus as an effec¬
tive vinegar-former. It consists of a large tub, made of oakwood, about 10 ft. high,
measuring 4 ft. at the top and 3J^ ft. at the bottom. Eighteen inches from the
bottom is a perforated false bottom, resting on a strong rim of beechwood. Two
inches below this are bored 8 to 10 holes, distributed evenly around the circumference.
In order to prevent leakage the holes incline slightly towards the inside of the tub.
Twelve inches below the cover is the second strong rim of beechwood, which carries
another false bottom set in watertight and provided with several hundred narrow
perforations. These perforations are about 1 in. apart. This sieve-bottom is for the
purpose of dividing evenly the alcohol-vinegar-mixture over the chips below. Often
these perforations are further restricted by placing pieces of twine, 2 to 3 in. long with
a knot tied at the upper end as a support, so as to allow only single drops of the liquid
to fall on the chips. The solution is poured on this bottom, is gradually taken up by
the twine, and drops down in a regular distribution. The space between the two bot¬
toms is filled with wood-shavings. In the upper portion of the latter is placed a long
right-angle thermometer, whose bulb reaches to the axis of the tub, and whose scale is
on the outside. The upper perforated cover carries six 1-in. holes, in which 6-in.
glass-tubes are inserted, which serve as ventilators to remove exhausted air from
the chamber below. Holes below the lower false bottom admit fresh air. An oak
cover closes the top. It has a funnel-shaped opening in the center, which may be
more or less closed at will. Liquid is poured in through it and the draft is there
regulated to suit conditions.
After the converter is ready, hot vinegar is passed through it for several days to
soak the chips and the inside of the tub. The alcohol-vinegar-water-mixture is now
poured through the funnel-opening of the cover at a rate of to 2 gal. per hour. It
divides itself over the chips and flows gradually and slowly to the bottom. The
procedure is repeated in either the same tub or alternately through another one,
constructed in the same way, until the desired degree of acidity has been obtained.
In the beginning the working-room is heated to 38°C. and the charge to 50 to 52°C.
When working regularly, the room is kept at 21 °C. and the liquid at 26 to 27°C.
By the rapid oxidation inside of the tub the temperature rises to 38 to 42°C. The
thermometer furnishes the best control of the process. After two passages through the
tubs, the liquid may properly be fortified by addition of more alcohol. This may be
repeated after two more passages, and thus a stronger vinegar (vinegar-sprit) is
obtain d.
REMOVAL OF ORGANIC MATTER FROM WATER
Sewage Treatment and Factory Effluents.—Almost any effluents from human

dwellings or places of industry carry (as suspensions or in solution, or both) large
amounts of organic waste material, which renders such discharges dangerous to
higher life. With a better understanding of the life-cycles of microorganisms,
these are now generally employed under controlled conditions for the purpose of
decomposing and removing such waste matter, before allowing the waste-waters
to pass into natural waterways.
This object is attained by permitting certain species of bacteria to invade the
waste-waters and cause therein by their life-processes an attack upon and a destruc¬
tion of any organic material present which they can utilize. These processes are
identical with the processes, generally occurring in nature, called “putrefaction”
which all serve the purpose of opening the way to a restoration to purposes of life all
dead organic matter, by freeing the carbon contained therein, mainly as carbon
dioxide (carbonic-acid gas) though sometimes hydrocarbons of the methane class may
appear in a subordinate way.
The reduction of nitrates to nitrites and of sulphur in sulphates and albuminous
material to hydrogen sulphide are frequently connected with the activity of microbes,
among which the most powerful are the bacteria which liquefy gelatin, lactic- and
butyric-acid bacilli, bacillus coli, bacilli typhi and particularly certain anaerobic
species. Hydrolysis is caused by numerous bacteria and yeasts.
Cellulose is thus reduced by cytase to glucose, butyric acid or methane. An
anaerobic bacillus, raised in pure culture by Omeliansky, decomposes cellulose into
methane and carbon dioxide or hydrogen and carbon dioxide, with aldehyde and acetic
acid as byproducts. Butyric acid is produced from almost any carbohydrates by
Bacillus butyricus Hueppe, Bacillus amylobacter van Tieghem and Clostridium buty-
ricum Prazsnowski. The corresponding decompositions of starch and sugars have
been discussed in previous chapters.
Fats will give glycerin and free fatty acids by ferments called lipases. Urea is
changed by urase, which has been found in over 60 species of bacteria, to ammonium
carbonate; hippuric acid to benzoic acid and glycocoll. Albumes are reduced to
albumoses and peptones by peptic and tryptic enzymes. Glycerin (when diluted, as
is always the case in effluents) is fermented by many bacteria, yielding alcohol, acetic,
formic, and succinic acids; or butyl alcohol and butyric acid. A great many more
examples might be cited; but the above will suffice to show the complexity of the
processes involved.
The essentials of a systematic plant for the purification of sewage consist, aside
from the question of power for moving, of a sufficient number of tanks or, in large
plants, basins, of which the first sets into which the sewage enters serve the purpose of
settling out solid matter in suspension and to start the decomposition of organic
matter present by fermentation, which may set in spontaneously or may be started by
artificial infection with suitable material.
The end products of the fermentation are carbon dioxide and water, accompanied
under certain conditions by methane and bad smelling gases of putrefaction. When
this part of the process is finished, the liquid is transferred to other tanks or basins,
where it is freed from any remaining obnoxious odors by aeration, there it passes
usually through clarifying and settling tanks or basins, after which it is ready for
discharge. There is no pure culture required for sewage treatment. The process
can be initiated by ferments picked up from the air. The fermentation period in
this case may last almost a year. It can be materially shortened by using material
from a previously fermented tank, without any regard for “pure” cultures.
FERMENTATION 605
In conclusion it may be pointed out that fermentation should play an impor¬

tant part in the chemical industries of the future quite apart from the production
of ethyl alcohol for industrial purposes or (in other countries) the old familiar
arts of brewing and distilling. There seems to be no underlying reason why
ethyl alcohol should be the sole fermentation alcohol produced on a large scale.
N-propyl alcohol is among the fermentation products of fermenting starch with
Amylobacter butylicum and A. aeihylicum. Methyl alcohol is produced as one of
the products of the fermenting of glycerol with the Bacillus bodcopricus. There is
an endless field for work with various ferments and various substrata for the
production of all of the alcohols.
Suggestive of the fines along which fermentation processes may be extended is

the way in which chemistry and bacteriology work together in glycerin production.
Since glycerin can be produced by the reduction of glyceric anhydride and dioxy-
acetone, and since there is some reason for believing these substances are formed
as intermediate products in the ordinary alcoholic fermentation, if one could add a
strong reducing agent to alcoholic fermentation and still have the ferment function,
it seemed logical that glycerin would be the product. Certain yeasts were found that
would grow in the presence of large amounts of sodium sulphite and sodium bisul¬
phite, and glycerin was produced by fermenting in this way. (Br. Pat. 163,034 of
1921.)
The methods given above for the production of pure cultures both on a small and
on a large scale are entirely general. It is also a general proposition that up to a cer¬
tain point warmth hastens fermentation, until a certain temperature is reached, the
“optimum,” and that as a rule, it is only a little above this point that the bacteria are
killed by heat. On the other hand, intense cold, though it renders the bacteria
inactive, usually does not kill. As a rule also, acids and alkalis are alike fatal to the
fermenting organisms, and usually the end product of the fermentation is itself a
poison to the bacteria causing it. Some bacteria require air for their proper growth,
others require freedom from it (anaerobic). To some extent these means may be used
partially to purify cultures.
Apart from the interesting problems of fermentation with microorganisms the
chemist also has the practically untouched field of the extraction and use of the
enzymes, that is, of chemical fermentation.
SECTION XVIII
FRACTIONAL DISTILLATION
By W. F. Faragher1
The Field of Distillation.—It is frequently desirable to separate naturally
occurring mixtures, or those prepared by laboratory or factory processes, into
their components. The extent of the separation required may vary from a
slight enrichment of the products in one or several of the components, to the
isolation, with maximum yield, of each constituent in a state of practical purity.
In the event that the mixture is a liquid, the method of separation which is most
frequently effective and which is, therefore, most frequently applied is fractional
distillation.
The purpose of fractional distillation as given in substance above, is quite distinct
from that of simple distillation as exemplified in the preparation of distilled water, and
also from that of drying, as illustrated by the evaporation of water from soap or glue,
or of organic solvents from films of nitrocellulose or cellulose acetate. The last two
processes are closely akin in aim to the concentration of solutions of practically
non-volatile substances and are discussed in Section IX. The removal of volatile
constituents as a whole characterizes these processes and any separations or changes
in the concentrations of the liquids which occur are only incidental. Provisions are of
course made for the recovery of expensive solvents. However, the objects in view are
obviously distinct from those of fractional distillation.
In a given fractionation, the results obtained are controlled by the number of sucl
cessive distillations carried out and by the character of the apparatus used. Industria-
equipment has been brought to a high state of efficiency as a result of the pressure of
commercial demands. Empiricism has necessarily had a large part in this develop¬
ment, but an increasing insight into the processes which occur has been of inestimable
value. The basis of the latter factor is found in the numerous investigations of the
laws which govern the changes of vapor-pressures of mixtures of liquids with changes
of temperature, pressure and composition of the liquid. The abolition of empiricism
awaits the formulation of generally applicable quantitative laws.
Vapor-Pressure.—If a pure substance is introduced in successive small

quantities into an evacuated vessel which is maintained at a constant tempera¬
ture, it is found, in general, that certain of the first portions change completely
into vapor. After the evaporation of each portion, the vapor exerts a definite
pressure upon the walls of the vessel. The actual values of these pressures for a
given substance (so long as the liquid evaporates completely) depend only upon
the total mass of the substance which has been introduced and upon the volume of
the vessel. As the concentration of the vapor increases, the pressure in the
vessel increases in approximate agreement with the general gas law. Experi¬
mentally it is found that the concentration and therefore the pressure of the
vapor can not be increased indefinitely under the conditions stated above.
After the pressure reaches a certain value, the introduction of more liquid effects
1 Mellon Institute, Pittsburgh, Pa.
607
no further increase of pressure. In fact, the liquid remains unchanged and even
effects the condensation of some of the vapor. Otherwise the vapor, occupying
a smaller volume, would exert a higher pressure, {piVi = p2y2, and p2 >pi, since
v\ >v2. Pi and represent the pressure and the volume respectively before the
liquid was present as such, pi having its maximum value; and p2 and v2 represent
the values of the pressure and volume which would obtain if there were no
condensation of vapor after liquid accumulated in the system. By thus exclud¬
ing the possibility of the condensation of vapor and the further evaporation of
liquid, the mass of the vapor would remain constant, and Gay-Lussac’s law should
hold as presented.) As more and more liquid is added to the vessel, the volume
of liquid increases and that of the vapor decreases. The pressure exerted by the
vapor, however, remains constant. Liquid and vapor are in equilibrium, then,
when the vapor exerts a certain pressure which is called the vapor-pressure
(or vapor-tension) of the substance. When the volume of the liquid placed into
the vessel just equals the volume of the vessel, the vapor phase is lacking.1
Any further addition of liquid increases the pressure in the vessel very greatly on
account of the low compressibility of liquids.
At any other temperature than that of the above considerations, analogous

conditions are observed. At each temperature a pure substance and its vapor can
exist in equilibrium only at one fixed pressure (dissociations, pressure due to other
vapors or gases, etc., are excluded). Increase of vapor-pressure always accompanies
a rise in temperature.
The vapor-pressure of a pure substance is a function of the temperature, and
is independent of the relative or absolute volumes of the liquid and vapor phases.
On the other hand, we may regard the pressure as the independent variable, and
the temperature as the dependent variable. For a given substance, kept at a definite
pressure, ph liquid and vapor can coexist unchanged only at one fixed temperature,
T. At higher temperatures, the liquid changes wholly to vapor, and at lower
temperatures the vapor condenses more and more to liquid.
The case is covered fully by the phase rule, as it is ordinarily stated: P + F =
C + 2. (P = number of phases in the system; F — number of degrees of freedom;
and C = number of components of the system.) For a system having the two phases,
pure liquid and its vapor, we may make one choice of the variables, pressure and
temperature, which determine the condition of the system. The other variable then
assumes a fixed value.
In dealing with a fixed mass of a pure substance, we find that we may choose either
the temperature and the total volume or the pressure and the total volume, at which
liquid and vapor shall coexist. The total volume, however, may vary only from vh
the volume which the given mass of the substance occupies as liquid, to v2, the volume
which it occupies as saturated vapor at the temperature and pressure which exist.
We cannot, however, choose both the temperature and the pressure under these
conditions.
It is not essential that we consider the vessel, in which equilibrium between liquid
and vapor is established, to be evacuated. It may contain any indifferent gas at any
pressure, p, up to quite considerable values, without influencing greatly the value of
the pressure of the vapor at equilibrium. In this connection, an indifferent gas is one
which is not appreciably soluble in the liquid and which does not react chemically
with it. The chief difference in the two cases is the length of time required for the
1 Strictly speaking, this volume is measured at the temperature of the vessel and at the pressure of
the saturated vanor.
FRACTIONAL DISTILLATION 609
establishment of equilibrium. In the first case, the vapor from the liquid very quickly
fills the whole free space, while in the second case it must diffuse slowly through the
relatively highly concentrated, indifferent gas. The behavior of mixed vapor and
indifferent gas, as outlined above, is in agreement with Dalton’s law of partial pres¬
sures; P = pi + pi. P is the total pressure in the vessel; pi is the partial pres¬
sure of the vapor (the vapor-pressure of the substance); and p2 is the partial pressure
of the indifferent gas. Tor the establishment of equilibrium between liquid and
vapor, only pi is of consequence.
The form of the vapor-pressure-temperature curve is approximately the same for

all pure substances. With rising temperatures, the vapor-pressure increases reatively
quite rapidly, suggesting an exponential relationship between the variables. Curves
for various substances are shown in Fig. 1.
In the above discussion of vapor-pressure, we have stipulated that the vessel and
its contents are kept at a constant temperature, while a given vapor-pressure is being
realized. The evaporation of the required mass of liquid is effected, obviously, by the
absorption of heat from the surroundings. In order to convert one mole1 of the
substance into saturated vapor (at the conditions of the system), a quantity of heat
equal to the molar heat of evaporation of the substance must be added. Likewise, the
reverse change, when effected, is accompanied by an evolution of heat, the molar heat
of condensation. These two heat quantities are equal numerically. The spontaneous
establishment of equilibrium between vapor and liquid is controlled by these energy
relationships.
The behavior of the system in equilibrium, with changes of temperature or pres¬
sure, is predicted by the Le Chatelier-van’t Hoff law of mobile equilibrium. For
example, the addition of heat to the system, which tends to raise the temperature,
1A mole (or gram-molecule) of a substance is the weight in grams of the substance equal to its
molecular weight, thus; a mole of water would be 18 grams (Ha + 0 — 2 + 16).
is followed by the evaporation of additional liquid, a reaction which absorbs heat.

There is produced, as a net result, an increased partial pressure of the vapor at the
resulting higher temperature.
Form of the Vapor-Pressure Curve.—If the corresponding values of vapor-
pressure and temperature are plotted in a system of rectangular coordinates,
the former being ordinates and the latter abscissas, a curve which grows increas¬
ingly convex toward the axis of abscissas as the temperature rises is generally
obtained for pure substances. Since just above the critical point liquid can no
longer exist, the vapor-pressure curve must end there. For a pure liquid, the
curve may be considered to end at the freezing-point. However, in many cases
the curve is experimentally realizable at lower temperatures. At these tempera¬
tures, the liquid is supercooled and the curve obtained is found to be a continua¬
tion of the curve for the substance at temperatures above the freezing-point.
While the vapor-pressures of the liquid and the solid phases of a substance
are equal at the freezing-(or melting-) point, as is required by the phase
rule, the vapor-pressure curve of the solid phase, at lower temperatures, lies
below the continuation of the vapor-pressure curve of the liquid. This be¬
havior is to be expected, since at all of these temperatures the supercooled
liquid is unstable with respect to the solid phase. The unstable form has the
higher vapor-pressure, and the continuation of the vapor-pressure curve of the
liquid is simply the vapor-pressure curve of the supercooled liquid. The lower
limit of the vapor-pressure curve of a substance may, perhaps, be considered to
lie at the absolute zero, where vapors are theoretically incapable of existence.
The vapor-pressure curve as described above is also the boiling-point curve of the
substance. This statement follows from the fact that if the external pressure is made
equal to the vapor-pressure of the substance, p, at the corresponding temperature, T,
the substance will boil. In the same way, the vapor-pressure curve of a solid substance
is its sublimation curve.
Many attempts have been made to formulate a general law which will express with
the required accuracy, the relationship between vapor-pressure and temperature for
pure substances. Theoretically, the solution of the problem is given by the Clausius-
Clapeyron equation. This equation relates the heat of evaporation of the liquid, the
temperature, the specific volumes of the vapor and of the liquid, and the temperature
coefficient of the vapor-pressure. The derivation of this equation is accomplished by
utilizing the first two laws of thermodynamics.
Consider a mole of the substance to be contained in a cylinder fitted with a friction¬
less, weightless piston, which is contained in a reservoir of constant temperature, T.
The piston rests upon the liquid, and is loaded with weights so that the pressure on the
liquid is equal to the vapor-pressure, p, of the substance. The volume of the liquid,
»i, is the molar volume at the existing temperature and pressure. The whole of the
liquid is caused to evaporate reversibly and isothermally, by slowly drawing out
the piston. The final volume, 02, is the molar volume of the saturated vapor. During
the evaporation, a quantity of heat, Lv, which is the molar heat of evaporation,
passes from the reservoir to the contents of the cylinder and the vapor performs
the external work W\ = p(v2 — 01).
This change is represented by the line AB in Fig. 2. The mole of saturated
vapor is then allowed to undergo an adiabatic expansion, until the pressure reaches a
value p — dp, which corresponds to the vapor-pressure of the substance at a tempera-
^ In fbe absence of the liquid phase, this change is considered to occur
with no condensation of vapor. Graphically, the expansion is represented by the
curve BC. During this change, the volume increases to v2 + dv2, and the temperature
falls to T — dTij which may be higher or lower than T — dT, the equilibrium tempera¬
ture for the pressure p — dp. The work performed by the gas during this adiabatic
expansion is W2 = (p — dp)dv2, or neglecting the product of the two differentials,
W2 — pdv2. At the constant pressure p — dp, the gas is then brought from the
temperature T — dTi to the equilibrium temperature, T — dT. If dTi>dT, the
vapor must be raised in temperature by the addition of heat, and if dTi<dT heat must
be abstracted from the vapor. The quantity of this heat is given by the equation
Q = C(dT — dTi). C is the molar specific heat of the vapor under the existing condi¬
tions. A further amount of work, Wz, is done during this adjustment of temperature,
Fig. 2.—Curve illustrating Clausius-Clapeyron equation.
and its value is (p — dp)dv3, where dv3 is the volume change which occurs. In the
diagram, this change is represented by the line CD. Since the volume change shown
is negative, heat was removed from the vapor, and dTi was less than dT. The condi¬
tions represented at D are those of the saturated vapor at the temperature T — dT
and under the pressure p — dp. By condensing the vapor reversibly at this tempera¬
ture, the changes pictured by the line DE are effected. The work done upon the
4,
vapor, W is equal to (p — dp) (v% 4* dv2 — dvz — *>i). Also, heat to the value of
the molar heat of condensation under these conditions, Lv + dLv, is given up to the
surroundings. Any change of molar volume of the liquid, v1} caused by the change of
temperature, dT, is regarded to be negligible in comparison with the change in the
molar volume of the vapor.
After the condensation of the vapor is complete (point E), the liquid is restored to
its original conditions of temperature and pressure by the addition of heat. The
quantity of heat required is Qi = CidT, where C1 is the molar specific heat of the
liquid. [Regarding the work done by the vapor as positive, and that done upon it as
negative, the whole of the work gained is
dW = Wi + W% - Wz - W4
= p(v 2 — i>i) + pdv 2 — (p — dp)dvz — (p — dp)(v2 + dv2 — dv$ — v{)
= (v2 — vi)dp, if the products of differentials are disregarded.
Since, in the limiting case, the process has been conducted reversibly and the
system restored to its original condition, the essential feature consists in the transfer
of Lv heat units from the temperature T to the temperature T — dT. We know,
therefore, that the second law of thermodynamics is applicable.
jrrr * dT T dT , f >.
dW = = Lv ‘~y = dp(v2 - t>i).
or
dp ___ Ly
dT T(v 2 — Vi)
This is the equation known as the Clausius-Clapeyron equation.
If the molar heat of evaporation and the specific volume of the liquid are known
as functions of the temperature, and if, further, the molar volume of the vapor can be
calculated accurately from the temperature and the pressure by using a suitable
equation of state, the changes of vapor pressure with changes in temperature can be
calculated with great accuracy. However, data which are sufficiently extensive and
accurate for this purpose are not generally available.
At temperatures and pressures not too close to the critical values, the molar volume
of the liquid is negligible in comparison with that of the vapor. The equation then
assumes the simple form:
dp _ Ly
dT - Tv
By assuming that the gas laws are applicable to the vapor, the substitution of RTfp
for v may be made in the above equation.
dp/dT = Lvp/RT2, or d log p/dT = LV/RT2
Before this equation can be integrated, the molar heat of evaporation, Lv, must be
expressed as a function of the temperature, T. For temperatures not too widely
separated, Lv may be regarded as constant, and in this form the equation is frequently
useful.
Integrated between the pressures pi and P2, corresponding to the temperatures Ti
and Ti (Ti being the higher temperature), the solution is as follows:
logPx-logp^^-A)
By changing to Briggsian logarithms and substituting the value of R in calories, we

get the form:
_L*_ _ (±_±\
1.985 X 2.303 \T2 Tj
Many of the most important equations which have been proposed and tested for
calculating the observed vapor-pressure curves of various substances, are reducible
to an equation derived from the above. In the derived equations, the dependence of
Lv upon the temperature is calculated thermodynamically in terms of the molar
specific heats of the liquid and of the vapor. The molar volume of the vapor is
expressed by one of the equations of state.
Measurement of the Vapor-Pressure of a Pure Substance.—There are

two general methods for measuring the vapor-pressures of substances: the static
and the dynamic methods. In the former method after equilibrium has been
established between liquid (or solid) and vapor, the relative volumes of the
phases remain constant. In the various dynamic methods, on the other hand,
the relative volumes change steadily. The limitations and advantages of the
different methods are discussed in the standard guides for making physico¬
chemical measurements.
A suitable method for measuring the vapor-pressures of substances at temperatures
lower than their normal boiling-points, has been devised by Ramsay and Young
(Phil. Trans., Series A, Vol. 178, p. 57 (1887);also “Stoichiometry,” Sydney Young, p.
121, Longmans, Green & Co., London, 1918). The apparatus is shown in Fig. 3. The
tube on the left is a simple barometer, which is provided with a jacket so that it may
be kept at a known constant temperature by a stream of water. On the right is a

barometer tube of special form, which is also jacketed in order to maintain the tube
and its contents at any constant temperature by means of the saturated vapor of a
suitable liquid which is boiled at the requisite pressure.
Some of the liquid to be investigated is introduced into
the tube on the right (details, lx.) and after equilibrium
has been established, the vapor-pressure is read. It is
necessary to reduce both of the columns of mercury to
their values at 0°C., and to take into account the effect
of the small column of unevaporated substance.
Two very useful methods have been developed by
Alexander Smith and A. W . C. Menzies. The first of
these is a static method1 and the second a dynamic
method.2 The principle of a large number of dynamic
methods which have been proposed is the determination
of the boiling point of the substance at a series of pres¬
sures. The method of Smith and Mcnzics and that of
Ramsay and Young3 are representative, and both give
accurate results.
For substances which have relatively low vapor-
pressures at the temperatures employed, methods based
upon one proposed by James Walker4 are useful. A
current of an indifferent gas at a known pressure and the
desired temperature is passed over or through the sub¬
stance. Precautions are necessary to insure that the
gas-stream leaves the substance at the equilibrium tem¬
perature and that the partial pressure of the substance
reaches the value of its vapor-pressure. Measurements
of the volume of the indifferent gas, the total pressure,
the temperature and the weight of substance which has
evaporated, provide the data which are necessary for the
calculation of the vapor pressure. In the calculation, it
is assumed that the vapor of the substance obeys the
gas laws, and that the formula of the substance as
vapor, i.e.j its molar weight, is known.
Dependence of Vapor-Pressure upon the Total
Pressure upon the Liquid.—In our discussion of the
vapor-pressure of a pure substance, we have so far
disregarded the effect of the total pressure exerted apparatus
upon the liquid. The applicability of Dalton's law of
partial pressures was, however, expressly limited to moderate total pressures,
and might well have been limited to temperatures considerably distant from
the critical temperature.
That the vapor-pressure does depend upon the total pressure borne by the liquid,
may be shown by a simple thermodynamic calculation.5 The equation which
expresses the relationship is as follows:
/ dp\ _ vo
\dP) t Vo
1 Jour. Am. Chem. Soc., Vol. 82, pp. 1412-34.
%Jour. Am. Chem. Soc., Vol. 82, pp. 897-905, 907-14 and 1448-59 U01O).
,
8 Phil. Trans., Vol. 175 p. 37 (1884); also “Stoichiometry,” Sydney Young, p. 129.
,
4 Zeit. phys. Chem., Vol. 2 p. 602 (1888).
At any constant temperature, 7\ therefore, the rate at which the vapor-pressure, p,

increases with an increase of the pressure P, upon the liquid, is equal to the molar
volume of the substance as liquid v0, divided by the molar volume as saturated vapor,
Fo. Both molar volumes are, of course, to be measured under the conditions which
exist in the system. Analogous considerations give similar relationships between the
partial vapor pressures of substances in solution and their corresponding molar vol¬
umes as liquid and vapor. In this case, both the mole fraction, n, and the temperature
are considered to be constant. The partial molar volume as liquid, v0lA is the in¬
crease in volume which is produced by adding one mole of substance to a very large
volume of the given solution.
The equation has the form:
At temperatures which are sufficiently far removed from the critical temperature,
it is obvious that change of vapor-pressure is not great for moderate changes of P
Poland 0O,J is small in comparison with F0[or VoM]). With rising temperatures,
however, the ratio of the specific volumes increases, becoming unity at the critical
temperature.
Dependence of Vapor-Pressure upon the Form of the Liquid Surface.—In the
methods of experimentation assumed in the previous discussion of vapor-pressure, as
well as in the methods described for measuring vapor-pressure, the liquid surface is
either plane, or has a large radius of curvature. Under these conditions, the form of
the surface influences only very slightly the value of the vapor-pressure at a given
temperature. It has been shown by Lord Kelvin,1 however, that the curvature of the
liquid surface can not be neglected if the radius of curvature is very small. The
vapor-pressure of a substance with a concave surface is smaller, and that of the
substance with a convex surface is greater than the vapor-pressure of the substance
with a plane surface. These considerations are of moment in the event that the liquid
is present in the form of small droplets or of films.
Vapor-Pressure of Mixed Liquids.—Pairs of liquids show varying behaviors

when they are mixed. In some cases, there is practically zero solubility of the
liquids in each other, while in other cases the solubilities are limited. On the
other hand, many pairs of liquids mix in all proportions, the process of solution
being accompanied, as a rule, by contraction or expansion in volume, and by the
evolution or absorption of heat. Pairs of liquids which are related chemically,
such as bromobenzene and chlorobenzene, form solutions without undergoing
appreciable changes in total volume and in temperature.
These differences in behavior may be pictured in terms of the attractions of the like
molecules of each liquid for one another, and of the attractions of the unlike molecules
of the two liquids for each other. In the event that the former attractions are rela¬
tively large, the two liquids should show little or no tendency to form solutions. A
pair of liquids for which the attraction of the unlike molecules is relatively greater,
should be more miscible. Still other pairs of liquids with progressively larger values
of the attractions of the unlike molecules should be progressively more miscible,
reaching finally the condition of complete miscibility. In the case of the partially
miscible liquids, there should be, in general, an increase in the total volume. The
considerations for this conclusion are that strong attractions of like molecules, which
cause, supposedly, relatively small volumes, give way in part to weaker attractions
with their accompanying larger volumes. With increasing miscibility, this expansion
,
iPhil. Mag. (4), Vol. 42 p. 448 (1871).
upon mixing should become smaller, then disappear, and eventually be followed by
contraction in the total volume. In the formation of the dilute solutions of partly
miscible liquids, there should be absorption of heat, which should grow smaller and
finally change to an evolution of heat as the tendency to exhibit complete miscibility
increased. The reason for this behavior is the assumption that the work necessary
to overcome the relatively strong attractions of the like molecules is greater than
that gained when the weaker attractions of the unlike molecules are satisfied.
Although the changes outlined above are in general in harmony with experimental
data, exceptions in the case of completely miscible liquids are of frequent occurrence.
For example, the addition of propyl alcohol to water is at first accompanied by a
contraction in the total volume and an absorption of heat. Pairs of liquids which are
closely related chemically, probably have attractions between all of the molecules
which are approximately equal. Daniel Berthelot1 and also Galitzine2 have proposed
that in these cases, the following relationship holds: A1)2 — \/Ai. A 2. Misrepresents
the attraction of the unlike molecules, and A\ and A% the attractions of the like
molecules of the two liquids, Ai and A2 respectively.
An explanation of the volume changes on the basis of molecular attractions as
outlined above was suggested by Brown in 1881.3 If the attractions of the like mole¬
cules exceeded those of the unlike, the mean attraction should be decreased upon the
formation of a solution. In keeping with this change, the volume of the solution
should exceed the sum of the volumes of the liquids. Brown added the objection,
however, that in view of this decrease in the value of the mean attraction, it would be
difficult to understand why the solution should be formed at all.
More recently these ideas have been given definiteness by Washburn4 in his presen¬
tation of the conception of thermodynamic environment. Two liquids which are as
unlike chemically as mercury and benzene separate quickly into two layers after they
are shaken together. Each layer is practically a pure substance, since analytical
methods are not sufficiently accurate to detect the presence of a second substance in
either case. If the mercury is replaced by water, which does not resemble benzene
closely, but which is much more like benzene than is mercury, we find that each
substance, at a given temperature, dissolves slightly in the other. At ordinary
temperatures, the benzene layer contains about 0.03 per cent of water, and the water
layer about 0.01 per cent of benzene. In order to have benzene soluble in all propor¬
tions in a second liquid, it is necessary to choose a liquid which is still more like it
chemically. Ordinary alcohol is such a liquid. The attractive forces between the
different molecules are such that whatever the proportions in which they are present,
the molecules can accommodate themselves to the conditions which exist. The first
portions of alcohol which are added are in an environment essentially like that in
pure benzene. With increasing quantities of alcohol, the conditions in the solution
differ less and less from those in pure alcohol. Finally, the conditions approximate
those in pure alcohol.
The tendency for a given molecule of any substance to leave a solution is deter¬
mined by the conditions which exist in the solution. The sum total of all of these
conditions in respect to a given molecular species, constitute the thermodynamic
environment for that species. Some of these factors are the attractive and repulsive
forces of the molecules, and the effects resulting from the collisions of molecules. The
mole fractions and the natures of the substances constituting the solution, as well as
the temperature and the pressure, influence the thermodynamic environment.
The thermodynamic environment in mercury and even that in water, is so different
1 Compl. rend., Vol. 126, p. 1703 (.1898).
2 Wied. annalen, Vol. 41, p. 770 (.1890).
3 Jour. Cherry. Soc., London, Vol. 39, p. 212.
4 Lib. cit., p. 134.
from the thermodynamic environment in benzene, that solutions of appreciable

concentrations cannot be formed. Alcohol and benzene are not closely alike, but
their differences are not so great that the molecules of either are unable to enter into
the conditions existing in the solutions of both. That differences exist, however, is
indicated by the appreciable fall in temperature which occurs when equal molar
quantities of the two substances are mixed. (The volume change is negligible.)
In the formation of solutions of two substances, two opposite tendencies develop with
increasing concentrations of one of them, e.g., the solute. At very low concentrations,
the attractive forces of the molecules of the solute scarcely come into play, since these
molecules are separated by relatively great distances. As the concentration increases,
the forces become of importance, and may cause the molecules of the solute to form a
separate phase of more suitable thermodynamic environment. At the same time,
however, the opposite tendency develops, and may act as a check to such a separation:
with increasing concentration, the condition of the solution surrounding any molecule
of the solute becomes more like that in the pure solute. The tendency for the solute
to separate as a separate phase is consequently smaller.
In general, even those substances which are wholly miscible, differ so much in their
thermodynamic environments that the change to the new conditions in a solution is
accompanied by appreciable changes in volume and energy content. Non-associated
liquids which undergo no chemical reactions upon mixing may be considered to indicate
roughly the magnitude of their differences of condition by the amount of the changes in
volume and in energy content.
The importance of the internal pressures of non-polar1 liquids in relation to their
escaping tendencies from solutions, has been emphasized by Hildebrand.2 As meas¬
ured by the closeness with which Raoult’s law is followed, Hildebrand has shown that
in some cases, approximate equality of internal pressures is of more importance than
chemical similarity.
Vapor-Pressures of Non-miscible Liquids.—A mixed vapor from two or more

liquids which are practically insoluble in one another follows Dalton's law of
partial pressures: The vapor from each liquid behaves approximately as if it
were alone in the vapor-space. It follows, therefore, that the boiling-point of a
mixture of such liquids is lower than that of any of the single liquids. These
facts were recognized by Gay-Lussac,3 and were proved experimentally by
Renault.4 Since the partial pressure of each liquid remains constant during
the distillation of such a mixture it is possible to calculate the relative weights of
the liquids obtained in the distillate.
Let Pa and Pb represent the partial pressures of the two liquids, A and B, in the
mixture which is boiling at the temperature t. The molar weights of the liquids in the
vapor form are Ma and MB. In each liter of mixed vapor, there is one liter of A at a
pressure Pa and temperature t, and one liter of B at a pressure PB and temperature t.
The volume of each substance as vapor corrected to normal conditions and expressed in
liters is as follows, assuming that the gas laws are applicable:
273 XPaX 1
For A, Vo,a =
(273 +t) X 760
273 X Pb X 1
and for B, 70,a =
(273 At) X 760
i See p. 621.
-Jour. Am. Chem. Soc., Vol. 38, p. 1452 (1916); ibid., Vol. 39, p. 2297 (1917); and ibid., Vol. 41,
p. 1067 (1919).
3 Ann. Chim. Phys. (I), Vol. 49, p. 393 (1832); Pogg. Ann. Vol. 28, p. 498 (1832).
4 M6m. de VAcadimie de Vlnatitut de France, Vol. 26, p. 715 (1862).
The weight of each of these corrected volumes of vapor is obtianable from the
relation that one mole of a substance as vapor at normal conditions occupies a volume
of 22.4 liters.
273 X Pa X 1 X Ma
Weight of A: Wa
(273 + 0 X 760 X 22.4
273 X Ps X 1 X Mb
Weight of B: WB
(273 + t) X 760 X 22.4
The relative weights are, then:
Wa Pa.Ma
Wb FB‘Mb
In an experiment described by Young,1 80 g. of water and 110 g. of chlorobenzene

were distilled at a pressure of 740.2 mm. The temperature varied only from 90.25° to
90.35° while practically all of the chlorobenzene was distilling. The temperature then
rose rapidly to nearly 100°. The distillate was collected in five fractions, the average
composition being 71.4 per cent chlorobenzene. The percentage calculated from the
above formula is 71.2 per cent chlorobenzene.
In a distillation of a mixture of aniline and water, which dissolve one another
relatively more than the above pair of liquids, the composition of the distillate was
found to be independent of the relative weights of the two liquids. The percentage of
aniline in the distillate, however, was appreciably lower than that indicated by the
formula. This fact is not surprising, since the partial pressure of each liquid, con¬
sidered as a solvent, is lowered by the other substance, the solute, and the partial vapor
pressures from the solution are not necessarily in the same proportion as the vapor
pressures of the pure liquids at the same temperature.
The theory of the steam-distillation of pure organic liquids which are not very
soluble in water, is outlined in the above discussion. For the process to be utilizable
in a given instance, the organic substance must be stable towards water at the tempera¬
ture of distillation, which is usually only slightly below that of water, but considerably
below that of the pure organic substance. The relative weight of the organic sub¬
stance in the distillate may be satisfactory even for high-boiling substances, since the
unfavorable effect of low partial pressure of the organic compound is offset by the
relatively large value of its molar weight in the vapor form.
Miscible Liquids.—The lowering of the vapor-pressure of a substance by the
addition of some of another substance which yields a solution is predictable from
the van’t Hoff-Le Chatelier law of mobile equilibrium. The condensation of
some of the vapor above the solution dilutes the solution which was formed, and
results in a condition which is more like that in the original pure liquid. The
more dilute vapor of the original substance which remains above the solution
possesses a smaller pressure than the vapor-pressure of the pure liquid. If the
added liquid has an appreciable vapor-pressure at the temperature of the solution,
the vapor in equilibrium with the solution will be mixed and the total pressure
will be made up of the partial pressures of the two constituents. The partial
pressure of the added constituent will be less than the vapor-pressure of that
liquid at the same temperature, since the relationships of solvent and solute are
reciprocal.
The behavior of pairs of non-electrolytic liquids, which are completely miscible,
with regard to the total pressure of the equilibrium vapor, is illustrated by the three
curves in Fig. 4. These curves apply at a constant temperature.
1 “Fractional Distillation,” Sydney Young, MacMillan, 1903, p. 90.
The compositions of the solutions, expressed in mole fractions, are given on the axis
AB, and the total pressure of the mixed vapors by the perpendicular distances from the
curves to the axis AB. Experimentally, curves of these three types have been realized.
In Fig. 5, three experimentally determined types of partial-pressure curves1 of

completely miscible, non-electrolytic pairs of liquids are shown.
In the simplest type of total pressure curve, 1, (Fig. 4) the pressure rises in direct
proportion to the mole fraction of the more volatile constituent, A. The pressure for
Fig. 5.—Partial pressure curves.
a mixture of given composition is, therefore, expressed by the following formula:

p _ Pa-wia + Pb( 100 — mA)
100
P is the total pressure, PA and PB the partial pressures of pure A and pure B,
respectively, at the same temperature as that of the solution, and mA is the mole
fraction of A in the mixture.
For those liquids which yield curves of type 2, the total pressure rises more rapidly
than the mole fraction of A. The total pressure of liquids which furnish a curve of
type 3 increases more slowly than the mole fraction of A.
Similar considerations hold for the partial pressure curves of Fig. 5. According
to the summary presented by Kuenen2 about 160 pairs of liquids (non-electrolytes)
have been examined carefully. Of these, only about 10 per cent give curves of the
simple type 1. Of the remainder, about 10 per cent have curves of type 3, and the
other 80 per cent, therefore, correspond to type 2.
1 Zeit. phys. Chem64, 727.
2 “ Theorie der Verdampfung und Verflussigung von Gemischen.” “ Bredig’s Handbuch, No 4.
Leipzig, 1906, p. 133.
The general equations which express the relationships between the changes of the
partial pressures of the vapors of the constituents in a binary solution are as follows:
For A- ti = wand for * dil = *(r- *•>

These equations express the conditions where the mole fractions are changed from
ua and nB to tia + driA and nB + dus, respectively.1 The corresponding pressure
changes are from pA to pA + dp a and from pB to pB + dps- The right-hand members
of the equations show that the ratios of the small changes of partial pressures to the
small changes of the mole fractions are functions of the thermodynamic environment
(T. E.) in the solution. The temperature and the total pressure are considered to
remain constant during these changes.
In order to utilize the equations, i.e., to integrate them, it is necessary to know the
form of the functions on the right-hand side of the equations. No general formulation
which is applicable to all kinds of solutions has yet been made.
A limited number of solutions have thermodynamic environments which remain
practically constant for varying constitutions. For these solutions the above equa¬
tions assume a form which can be integrated directly:
dp a
Ua, and — kB
driA dnB
The integrated equations are as follows:

Pa = kAUA + CA, and pB = kBriB + CB
The constants of integration, Ca and CB} are seen to be equal to zero, since when
riA or 7iB is zero, the partial pressure pA or pB, respectively, is zero, as is also the
corresponding constant of integration.
The significance of the proportionality factors, kA and &#, is shown by making tia
and 7ib equal to unity. The partial pressures Pa and pB then equal the vapor-pressures
of the pure substances A and B, respectively. The equations for such solutions, which
are called ideal solutions, assume the following forms:
Pa = P<>,a.71a and pB = p0,B-nB

In these equations p0,A is the vapor-pressure of pure A, and p0}B that of pure B at the
temperature of the solution. If we consider the substance A as the solvent, the equa¬
tion pa = Pc, An a is simply a statement of Raoult’s law in its simplest form. In the
same way, considering B as the solute the equation pB = PotBnB is really a slightly
modified form of Henry's law. It has been shown by Lewis2 that if vapors obey the
gas laws, the law of Henry may be derived by thermo dynamic considerations from
Raoult's law. The former law, therefore, must hold for solutions which follow the
latter law. There can be no question that the most satisfactory criterion of an ideal or
perfect solution is agreement with Raoult’s law.
In binary solutions of liquids in general, Raoult's law expresses the existing
relationships when the mole fraction of the one constituent is quite large in comparison
with that of the other constituent. This statement is equivalent to saying that
in dilute solutions Itaoult's law is valid. The curves of types 2 and 3 (Fig. 5), then,
tend to become tangent to curve 1 at the extreme concentrations, as shown.
Two general methods of explaining the causes of the deviations from Raoult's law
have been advanced. In the one, supported in particular by Dolezalek,3 a chemical
reaction between the two constituents, or the formation of complex molecular species
of one or both of the constituents, is considered to occur. Raoult's law, however, is
1 Washburn, Lib. cit., pp. 132 and 143.
2 Jour. Am. Chem. Soc., Vol. 38, p. 674.
3 Zeit. phya. Chem., Vol. 64, p. 727 (1908); Vol. 71, p. 191 (1910); Vol. 83, p. 40 (1913).
still considered to hold for each molecular species in the solution. By an application of
the law of mass action, Dolezalek has utilized this method in studying a number of
systems which have been investigated. He has succeeded in obtaining calculated
values which agree satisfactorily with the measured values. In many cases, the
requirements of the method are in accordance with the general characteristics and
behavior of the substances. In other cases, however, the interpretations require
chemical reactions or associations which are at variance with the commonly accepted
nature of the constituents. For example, pairs of liquids which by all commonly
recognized tests are normal, i.e., non-associated, often have vapor-pressures in excess
of those required by Raoult’s law. According to Dolezalek, one of these liquids would
have to be considered to be associated in the solution. In many cases, the degree of
association indicated by the calculation is so great as to assign the substance definitely
to the class of associated liquids. Such pairs of liquids are represented by bromine
and carbon tetrachloride, carbon disulphide and methylol, benzene and hexane, and
benzene and ether. The pair of normal liquids, hexane and aniline, have been found
by Hildebrand and Keyes to deviate from Raoult’s law to the extent of forming two
phases at temperatures below 59.3°C.1 An explanation of this behavior on the basis
of association of the constituents, so that Raoult’s law is applicable, is difficult to
conceive.
The second method of explaining deviations from Raoult’s law, is that suggested
by van der Waals,2 and adopted by van Laar3 and Kohnstamm.4 This theory relates
the deviations from the law of Raoult to the constants of a modified van der Waals’
equation, which is applicable to pairs of completely miscible liquids. The possible
types of vapor-pressure curves are also predicted from the properties of the equation.
For example, van Laar shows that we should expect a deviation from Raoult’s law in
the case of normal liquids, if their critical pressures are different. The method of van
der Waals fails when one of the constituents is associated. Although the critical
pressures may be practically the same, deviations from Raoult’s law to the extent of
the formation of two liquid layers may occur.
Raoult’s law has a very simple kinetic basis. If the vapor of a binary mixture follows
the gas law, and if the solutions have a practically constant thermodynamic environ¬
ment, it is a natural conclusion that since only a fraction of the total number of mole¬
cules is that of one constituent, the partial pressure of that constituent should be less
than the vapor-pressure of the pure substance. Since the fraction of the total number
of molecules is expressed by the mole fraction of the constituent, it is to be expected
that the partial pressure should be the same fraction of the vapor-pressure of the
substance at the same temperature. Deviations from Raoult’s law have their kinetic
basis in a lack of constancy of the thermodynamic environment. According to Hilde¬
brand,5 the most satisfactory criterion for constancy of thermodynamic environment
is equality of internal pressures. The conception of internal pressure has been
employed frequently for other purposes, and several methods have been devised for
evaluating it for different liquids. Although the actual values yielded by the different
methods differ quite considerably the relative values show satisfactory agreement.
Internal pressure is the force which, together with external pressure, counteracts
the thermal pressure caused by the kinetic energy of the molecules of a liquid. The
tendency of molecules to escape from a solution would seem to be closely related to
this property, internal pressure. For normal liquids with approximately equal
internal pressures, Hildebrand finds that the observed partial pressures agree satisfac-
1Jour. Am. Ckem. Soc., Yol. 39, p. 2126.
* Die Kontinuitat,” van der Waals. See p. 190 for van der Waals’ constants.
8 Zeit. yhya. Chem., Vo! 72, p. 723 (1910).
, ,
4 Ibid., Vo! 36 p. 41 (1901) and Vo! 76 p. 527 (1910).
,
5 Jour. Am. Chem. Soc., Vo! 38 p. 1458 (1916).
torily with those required by Raoult’s law. Differences in internal pressures result in
deviations from the law in the direction of greater partial pressures than the law
indicates. It might be expected, naturally, that two kinds of molecules of different
sizes and attractive powers would give volumes larger than those calculated on the
basis of additivity. Furthermore, it might be expected that the attractions would
be smaller and the tendencies to escape into the vapor-phase greater than in the
case of the separate liquids. Preliminary calculations give promise that the discovery
of a formal relationship between these quantities may be expected.1
Another factor which is concerned in the deviations from Raoult’s law is the polar
nature of many liquids. This property has been discussed by Bray and Branch2
and by Lewis.3 Some of the properties of typical polar substances are chemical reac¬
tivity, existence in tautomeric forms, electrical conductivity, ionizing power, high
dielectric constant, and association of the molecules into complexes. Typical non¬
polar substances have the opposite properties.
As a result of polarity, there is a molecular attraction which yields a higher internal
pressure than would otherwise exist. Addition compounds are formed readily, and
there is a tendency for a substance to increase in polarity when it is introduced into
other polar substances.
A mixture of a strongly polar and a non-polar liquid generally results in the forma¬
tion of a double liquid layer, and in considerable deviations from Raoult’s law. Two
polar liquids may show agreement with the law, but in general, one of them will be
more electro-positive than the other and the vapor pressures will be smaller than the
law requires.
The bearing of internal pressure and of polarity would seem to be of real moment
in this field, although as yet the method is capable of indicating only the relative
variations from Raoult’s law. The effects of the two properties upon the deviations
have been summarized by Hildebrand as follows:
(1) Raoult’s law should be followed by binary mixtures of non-polar liquids which
have the same internal pressures; (2) non-polar liquids of unequal internal pressures
yield solutions with greater vapor-pressures than required by Raoult’s law; (3) a polar
liquid and a non-polar liquid behave in the same way as is indicated in (2); (4) mixtures
of polar liquids may show deviations in either direction, but as a rule smaller vapor-
pressures than are expected, are observed.
The usefulness of the above considerations in predicting the results to be obtained
upon distilling a mixture of liquids, is shown by the following example.4 The results
to be obtained bya fractional distillation of pairs of liquids which follow Raoult’s law
are predictable from the partial pressures of the constituents. Internal-pressure
relationships, by indicating the nature of the deviations from Raoult’s law, enable one
to judge of the results in a given case in comparison with other cases which have been
fully investigated. The deviations of oxygen and nitrogen have been investigated
fully.6 Since the differences in the internal pressures of oxygen and nitrogen arc
greater than those of nitrogen and argon, smaller deviations from Raoult’s law were
predicted for the latter pair of substances by Hildebrand. Measurements which
had been made previous to this prediction, but which were unknown to Hildebrand, are
in agreement with the prediction.
Partial and Total Vapor-Pressure Curves.—A system composed of a vapor

and a liquid phase of a binary mixture has, according to the phase rule, two degrees
1 Hildebrand, 1. c.
2 Jour. Am. Chem.

Soc., Vol. 82, p. 1440 (1913).
zIbid., , ,
Vol. 38 p. 1448 and Vol. 38 p. 762 (1916).
4 Hildebrand, Jour. Am. Chem. Soc., Vol. 41, p. 1076 (1919).
of freedom. At any given temperature, therefore, the vapor-pressure may have

a continuous series of values. Obviously, these vapor-pressures are those of
solutions of a continuous series of values of compositions, expressed preferably
as mole fractions.
In Fig. 6, the relationships between the partial-pressure curves and the total-
pressure curves for the three typical forms which have been considered, are shown.
Each total-pressure curve is the sum of the partial-pressure curves of the constituents.
The compositions given are those of the liquid, as has been the case for all curves so
far presented.
Besides the curves which have been considered, there are several other forms which
are possible. Some of these have been realized experimentally. Concerning the
existence of others, there is still a division of opinion. This aspect of the subject is
outside the scope of this article, and can be referred to only casually. The various
possible forms of partial-pressure curves, including those with points of inflection,
permit of a large variety of total-pressure curves, by their combination.1
Form of the Vapor-Composition Curve.—It is of course true, that besides

the curves showing the relationships between the compositions of the liquid
mixtures and the vapor-pressures, there are curves showing the relationships
between the composition of the saturated vapors and the vapor-pressures. We
have already seen that the vapor-pressure curve of ideal binary solutions is of
a very simple form, being expressed by the equation of a straight line. The curve
which shows the relationship between the composition of the saturated vapor'
and the vapor-pressure is, however, not so simple even for this case.
If it is assumed that the vapors follow the gas laws, the ratios of the mole fractions
of the constituents in the vapor phase are equal to the ratios of their partial pressures.
Let %i and 1— n\ represent the mole fractions of A and B respectively, in the liquid
phase, and n2 and l—n2 the mole fractions in the vapor phase. From Fig. 7, we see
that
n2 _ aG _ niPi _ rii Pi
1 - n2 ” bG “* (1 - nOP2 ~1-WiT2
* Nernst, Theoretical Chemistry,” MacMillan, 1916, p. 107; Ostwald, “Lehrbuch der allgemeinen
Chemie.” 2d Ed., II, p. 617 if.; Kuenbn, “Theorie der Verdampfung und Verfliissigung von Gemis-
chen,” Leipzig, 1906; Marshall, Jour. Chem. Soc., Vol. 89, p. 1350 (1906); Konosvalow, Jour. Chim.
Phys., Vol. 5, p. 237 (1§07); Zawidzki, Zeit. phys. Chem., Vol. 69, p. 630 (1909); Dolezalek, ibid.,
Vol. 64, p. 729 (1908); Bobe, ibid., Vol. 66, 458 (1908); Bancroft, “Phase Rule,” Ithiea, 1897, p. 116
ff.
On the other hand, since the total pressure, P, is equal to the sum of the partial
pressures we have the equation
P = mPi + (1 - nOPa
From these two equations is obtained the equation of the curve for the composition
of the vapor and the total pressure
p _• P1P2_
Pi(l — 112) -f- P2^2

This equation is that of a hyperbola. At any given pressure, the composition of
the vapor is higher in the more volatile constituent than that of the liquid. For the
special case that Pi = P2 (optical isomers), the curve is a single horizontal straight line
for both the liquid and the vapor. It is also evident that the form of the two curves
Pi
for pairs of liquids having equal values for the ratio -5- is the same, regardless of the
-t 2
absolute values of these quantities. The vapor-curve lies below the liquid-curve at
each composition.
As we have seen, most pairs of liquids do not yield these simple curves. The
general considerations outlined above are only qualitatively applicable to most cases.
Total-Pressure Curves with Maxima and Minima-—The total vapor-pressure

curves for certain pairs of liquids have maximum values. In relatively fewer
cases, the curves possess minima. Pairs of liquids which present these phenom¬
ena can not be separated directly by fractional distillation. It has been observed
that for pairs of liquids which yield a total vapor-pressure curve with a mini¬
mum, one of the liquids is an acid and the other a substance of basic character,
or that the liquids are water and certain acids. Some exceptions to these condi¬
tions are known. A list of binary mixtures which yield vapor-pressure curves
with maxima or minima is given by Young.1
In the case of these curves, the vapor composition curves lie below the liquid com¬
position curves, both having a common tangent at the maximum or the minimum
value. The vapor is not, however, uniformly richer in the more volatile component,
as can be seen readily by drawing the curves.
Theory of Fractional Distillation-—A quantitative basis for a theory of frac¬

tional distillation is furnished by the work of Konowalow. This theory, for the
sake of simplification, is limited to binary mixtures which do not yield a vapor-
pressure curve with a maximum or a minimum value. The latter mixtures
may, however, be included, if the solutions are considered as binary mixtures of
one or the other pure substance (the one present in excess of the composition of
the mixture at the maximum or minimum), and of the solution having the maxi¬
mum or the minimum vapor-pressure.
By a process which amounts to an application of the van’t Hoff-Le Chatelier law of
mobile equilibrium, Konowalow (l.c.) arrived at the following conclusion: The vapor
of a binary mixture has a larger content of that constituent by the addition of which
the vapor pressure of the liquid is increased, than does its equilibrium liquid.
Konowalow's course of reasoning is as follows: Upon increasing the volume of the
vapor phase of a binary mixture in equilibrium with its liquid phase (constant tempera¬
ture), some change must ensue which tends to counteract the change in volume.
This change is a decrease in vapor-pressure. Since for the pair of liquids under
1 “Fractional Distillation,” p. 67 ff.

consideration, the curve of liquid composition and vapor-pressure slopes downward

continuously from the point representing 100 per cent of the more volatile constituent
to the point representing 100 per cent of the less volatile constituent, this decrease in
vapor-pressure requires that the liquid have a higher percentage of the less volatile
constituent than originally. By applying the above reasoning to equilibrium systems
of all possible compositions, it is seen that the curve of vapor-compositions and vapor-
pressures lies below the curve of liquid compositions and vapor-pressures at all points
(except for the pure substances). The compositions of vapor and of liquid which are
in equilibrium are represented by the points of intersection of the two curves with a
horizontal line drawn through the point which shows the pressure of the system. In
every case, therefore, the vapor is richer than the liquid in the more volatile constit¬
uent. A construction of the curves makes this fact very apparent.
On the basis of the results obtained by Konowalow, Golodetz1 has outlined a theory
of fractional distillation. The theory is qualitative only, but is nevertheless of some
utility. If we consider A to represent the more volatile constituent and B the less
volatile, the following form may be given to the statement of Golodetz’s theory.
1. If a mixture of two liquids A and B is boiled, the temperature of the saturated
vapor lies between the boiling points of the constituents. This temperature approxi¬
mates the boiling point of one of the constituents the more nearly, the greater the
concentration of this constituent in the liquid phase.
2. The vapors are always richer in the more volatile constituent than the liquid
mixture.
Since the process is reversible, these principles apply in an analogous way to the
fractional condensation of a saturated vapor. It should be stated that the same
limitations are postulated as were stated above in the discussion of Konowalow's
work.
All mathematical formulations of these principles have the same defect: they lack
general applicability. After a consideration of some of the qualitative information
which may be gained from these principles, we shall consider some of the quantitative
treatments which have been proposed.
Measurement of Partial Pressures.—It is impossible to measure directly the

partial pressures of the constituents of a binary (or other) mixture. In all of the
methods employed for determining these partial pressures, the composition of
the vapor in equilibrium with the solution, under the chosen conditions, is
ascertained. It is then assumed that the partial pressures are proportional to
the relative molecular concentrations of the constituents in the vapor. The
measured total pressure and the assumed molecular weights of the constituents
make possible the calculation of the partial pressures.
In most of the methods employed for determining partial pressures, a small quan¬
tity of distillate is collected from' a relatively large quantity of the solution. It is the
aim of the method to secure sufficient of the distillate for analysis, without changing
appreciably the composition of the original solution. In the extensive investigations
of Zawidzki,2 about 1 c.c. of distillate was obtained from 100 to 120 c.c. of the solution.
In some cases, certainly, the composition of the vapor does change appreciably under
these conditions. Another objection to the method is the fact that water, if present in
the solution, tends to pass over with the first portions of the distillate, often as the
result of the formation of a solution of minimum boiling-point. Various investigators
have carried out their distillations under constant pressure, while others have made
them at constant temperature. An accurate method for determining partial pressures,
9-
1 Golodetz, Chem. IndVol. 35, pp. 102 and 141 (1912).
2 Zeit. phya. Chem., Vol. 35, p. 129 (1900).
which is free from the above objections and which can be made to yield any desired
volume of distillate having the composition of the equilibrium vapor, has been per¬
fected by Rosanoff and Easley.1 Rosanoff, Bacon and White2 have devised a simple,
effective laboratory method for partial vapor pressure determinations.
Boiling Point-Composition Curves.—In the case of normal liquids, the

boiling-points of binary mixtures can be calculated if the vapor-pressure curves
of the individual liquids are available. The molecular compositions of the
mixtures having a vapor-pressure of 760 mm. at a series of temperatures between
the boiling points of the pure liquids are calculated by applying the formula:
Pb - 760
m = 100'
Pb — Pa
m is the percentage molecular composition of the solution, for the less volatile
constituent. PB is the vapor-pressure of the more volatile constituent and PA
that of the less volatile constituent at the same temperature. Obviously, the
same formula may be used in constructing the boiling-point curve at another
pressure, by substituting the value of this pressure in place of 760 in the formula.
The values of m-thus obtained are plotted as abscissas against the corresponding
temperatures to obtain the boiling-point curve.
The curve obtained in this manner expresses the composition of the liquid, i.e.,
of the solution. For normal liquids, the curve is not a straight line, but rather lies
below the straight line joining the boiling-points of the pure substances. The devia¬
tion from this straight line is least for pairs of liquids which have boiling-points close
together.
The calculation of the composition of the vapor in equilibrium with the solution
at its boiling-point, is possible for normal liquids. In accordance with the principles of
Konowalow, the vapor-composition curve lies above the liquid-composition curve,
since the vapor contains a higher percentage of the lower-boiling constituent than its
equilibrium solution. The vapor and the
liquid which are in equilibrium are those
which have the same temperature. In the
usual diagram, therefore, in which the
boiling points are plotted as ordinates, and
the compositions as abscissas, any given
solution is in equilibrium with that vapor
which lies on the same horizontal tempera¬
ture line.
Distillation of a Binary Mixture.—

The behavior of a binary solution upon
distillation, under such conditions that
the vapor and solution remain in contact
so that they are continuously in equilib¬
rium, is shown by Fig. 8. The curve AqB
is the vapor-composition curve and AnB the liquid-composition curve. The
boiling-points are given as ordinates. Molecular compositions are shown as
abscissas. If a solution of the composition corresponding to the point n is
heated to its boiling-point, the first vapor which is formed is richer in the lower
i Jour. Am. Chem. Soc., Vol. 31, p. 953 (1909).
1Jour. Am. Chem. Soc., Vol. 36, p. 1803 (1914).
40
boiling constituent, jB, than the original solution. Its composition may be
found by reading the abscissa of q', the intersection of the vapor-composition
curve, AqB, and a line drawn through w, parallel to the axis AB. The remaining
solution has a lower percentage of the lower-boiling constituent and a higher
boiling-point than the original solution, and is represented, therefore, by some
point to the left of n on the curve AnB. Continued formation of vapor causes
the composition of the solution to become progressively higher in the higher¬
boiling constituent, A. The points representing these compositions lie farther
and farther to the left of P, on the curve AnB. When the composition of the
solution reaches the value corresponding to the point n'} the diagram shows that
the equilibrium vapor has the same composition as the original solution. Since
the total mass of vapor and liquid has remained constant throughout the distilla¬
tion, it is obvious that at the point n'} all of the liquid has been transformed into
vapor. Just before all of the liquid has evaporated, the solution has the compo¬
sition represented by a point just to the right of n'. It is obvious, therefore,
that when the distillation is conducted in this manner, the maximum enrichment
of the solution in the higher boiling constituent, A} is limited to a value given by
the abscissa of the point nf. The proportion of the solution having this composi¬
tion is, of course, very small.
The behavior of a vapor having the composition corresponding to the point q,

when it is condensed gradually in the presence of the condensate formed, is shown
in the same diagram. The first small portion of liquid formed is richer in the higher
boiling constituent than the vapor, its composition being given by the point n'. The
remaining vapor has, therefore, a higher percentage of the lower-boiling constituent,
and is represented by a point to the right of q on the curve AqB. Further condensa¬
tion of vapor enriches the vapor continuously in the lower-boiling constituent, the
points representing the condition of the vapor lying progressively to the right of q.
When the vapor reaches the condition represented by the point q% the liquid is seen to
have the same composition as the original vapor. Obviously, at this point, the vapor
has been completely condensed to liquid.
Furthermore, it is apparent that the
maximum enrichment of the vapor in the
lower boiling constituent by this process
is given by the point n': and also that
the yield of this vapor is very small.
Quite a different behavior is observed
if the vapor which is formed during the
distillation is removed continuously or
at intervals, and condensed apart. In
Fig. 9, the behavior upon distillation of
a solution which has the composition
corresponding to the point p, is shown.
When the liquid has changed in com¬
position from that represented by p to
that by plf the equilibrium vapor is
removed from the system. The remain¬
ing solution, boiling at a higher temperature, th gives off further vapor, and
changes in composition until the point p2 is reached, when * the vapor is again
removed. By repeating this process in successive steps, the composition of the
. solution is seen to approach that of practically pure A} the higher-boiling constituent.
Here, too, the yield of enriched solution is relatively very small. By this process,
however, it is possible to proceed to a practically pure constituent of the solution.
The fractional condensation of a binary vapor, with step-wise removal of the con¬
densate, is shown in the same diagram. Starting with a vapor of composition corre¬
sponding to q6, the first condensate is seen to be represented by the point p&. After
the vapor reaches the composition of the point qs, the condensate is removed and the
condensation and removal of condensate carried successively through the stages
etc. The composition of the final vapor can be made, in this way, practically
to reach that of the pure constituent B. The yield is again relatively very small.
By employing the preceding process of distillation, it is then possible to isolate the
higher-boiling constituent in pure form. Analogously, by fractionally condensing
the mixed vapors, it is possible to obtain a vapor which is the pure low-boiling con¬
stituent. Neither process alone can yield the two pure substances. Since in a given
operation the yield of pure substance is negligibly small, the processes of distillation
and condensation must be carried out simultaneously and repeatedly, in order to
be of practical use. The elementary theoretical aspects of the operation, are, however,
given in the above discussions.
Boiling-point Curves and Relative Weights of Vapor and Solution.—A

valuable feature of the boiling-point-composition curves is brought out in
Fig. 10. A solution of weight M and
original composition P, is distilled until ^
the solution has the composition p and ^ '
the equilibrium vapor the composition q.
If the weight of the solution is m, and that
of the vapor n} the following relationships
hold:
M = m + n and MP — mp + nq.
From these two equations, the following
is derived:
m __ q — P
n ~P — p
From the diagram, therefore, we see
that the ratio of the weight of the
solution to that of the vapor is given by the proportion ~~ - The ratio of
the weight of the vapor to that of the total is shown by pa/pq and the ratio
of the weight of the liquid to the total weight by aq/pq.
Principle of Fractionating Column.—The design of the usual laboratory
fractionating column or of the conventional column still for commercial use is
such that fractional distillation and condensation are combined and repeated a
number of times. Instead of passing directly to the condenser, the vapor which
leaves the flask or boiler, passes to a column with a number of compartments,
separated by perforated trays or otherwise, in which partial condensation is
effected by cooling. The condensate is trapped in part, so that the vapor from
the boiler or from a lower compartment must bubble through the layer of
condensate. In this way, an exchange of constituents between vapor and
condensate is effected, so that the vapor leaving a compartment is richer in the
lower boiling constituent than it was when it entered. The incoming vapor,
having a temperature higher than that of the condensate in the compartment,
heats it and causes it to boil. The vapor formed from the condensate is richer
in the lower-boiling constituent, as we have seen. The incoming vapor itself,
which is cooled in the compartment, is partially condensed, and as we have seen,
also yields a condensate which is richer in the higher-boiling constituent. The
portion of this vapor which remains is therefore enriched in the lower-boiling
constituent. The enriched vapor passes on to the next higher compartment and
the excess of condensate flows down to the next lower compartment. The num¬
ber of designs of apparatus for effecting fractional distillation both for laboratory
and for plant use, is very large. The fundamental principles of nearly all,
however, are those outlined above.
Changes Occurring in a Fractionating Column.—The method of operation
of a fractionating column with ten trays or compartments, is shown in Fig. II.1
The curve AqB is the composition
curve for the vapor, and ApxB that
of the liquid. Boiling-points are
plotted as ordinates. The vapor
from the boiler has the composition
corresponding to q, and is, therefore,'
already fairly rich in the lower-boil¬
ing constituent, B. Since the top
tray is to deliver nearly pure B to the
fractional condenser, the tempera¬
ture at this part of the column
must be only slightly higher than
Fig. IX.—Ten-tray fractionating column. the boiling point of pure B. The
temperatures on the successive
trays are maintained at values such that the differences are greater for the
lower pairs of trays. This provision, effected by structural features and heat-
insulation, insures that the efficiency of the column shall be good, as will be
shown later.
On the first tray, the vapor, q, is fractionally condensed, its composition being
changed to qx. If the weight of vapor, q, were M, the weight of vapor qh is seen to be
M ~~~ * This vapor then passes to the second tray, where its composition is altered
to that of 02. The weight of residual vapor is M In the same wav. we

Piqi V2q2 J*
could trace the vapor through each of the trays. Obviously, the weight M', of vapor
leaving the tenth tray is given by the following equation:
M'=M Pl™x. . p*m2

Vi <L M2..P10 qio
The factors of the right-hand member become progressively smaller, and it is
obvious that much of the original vapor is returned to lower trays as condensate.
In a similar way, we could trace the course of the condensates from tray to tray.
Thus, the condensate pi0 flows to tray 9 and yields a vapor of composition #9; and the
other condensates p», ps, • . . yield the vapors g8, qi, etc. The weights of these
secondary vapors are expressed by formulas analogous to those for the primary vapor.
If the curves are drawn for the case in which the temperature intervals between
each pair of trays is constant, instead of greater for the lower trays than for the upper,
1 Gubwitsch: " Wissenschaftliche Grundlagen der Erdolbearbeitung.” Julius Spbingbb, Berlin,
1913, p. 16 fF9.
the equation for the weight of primary vapor, M", leaving the tenth tray has the same
form as that given above for Mf. The first factors of the right-hand member,
are smaller in the M' equation than those of the M" equation. However, the last
factors of the M' equation are relatively much greater than those of the M" equation.
The result, therefore, is that the quantity of vapor, is greater than M". It is
advantageous, therefore, in this case, to make the temperature differences between
successive upper trays as small as possible. In the event that the vapor from the
boiler is lower in the lower-boiling constituent, the efficiency of the column is increased
by making the temperature differences between the lower trays small, also.
Since the requirement of small temperature differences between the trays requires
a large number of trays for an ordinary separation, it is customary to take the vapor
from the top tray of a column with a considerable number of trays to a fractional
condenser, where the removal of the final portions of the higher-boiling constituent is
effected without rectification, at this place, of the condensate (the condensate is returned
to the column). This fractional condenser is in effect a constant temperature still-
head, and should be maintained constantly at a temperature only slightly above that
of the boiling-point of the more volatile constituent, if a pure product is desired. The
vapor from the fractional condenser, which is the product of the column, is led to the
cold condenser.
Importance of Number of Trays in a Column.—Since the fractional condensa¬

tion and distillation are effected upon the trays, it would seem to be obvious that
the number of trays should be as large
as possible for the sake of efficiency. /[
The effect of decreasing the number
of trays is shown in Fig. 12. If the
first tray is left out, the vapor from
the boiler, q, is changed on the second
tray into the vapor q2. If the weight
of the vapor, q, was M, the weight of
q2 is M * In the event that the
vapor qi is altered on both the first and

second trays, the quantity of vapor,
02, is M ~—-• The .latter Fig. 12.—Effect of decreased trays.

Mi M2
quantity is greater than the former.1
Return of Condensate from Fractional Condenser.—The advantage of
returning all of the condensate from the fractional condenser to the top tray,
rather than distributing it to several trays, is made apparent in Fig. 13. Since
the vapor passing to the fractional condenser is generated from the solution on
the top tray, it follows that the condensate which is formed in this condenser
must be richer in the lower-boiling constituent than the liquid on the top tray
. P2W2' pmi Pirn*'
lif pm pm > PlWli
pmi __ P2W2' + pm
From Fig. 12,
PiQi pun? + Pin
Since -- is a nrooer fraction, the addition of the same quantity, pm, to its numerator and denomi-
ptmi *
. .
nator increases its value.
, .
Therefore, ~
pimi Piirii'
—
(Since it is in equilibrium with a vapor which is richer in the low-boiling con¬

stituent than the vapor which entered the condenser.) When the condensate is
led off at several levels from the condenser, even the first condensate is richer in
the low-boiling constituent than that
on the top tray. This consideration,
as well as the fact that the separation
is more efficient the higher the ratio
of the weight of the condensate with
which the vapor is brought in contact
on a tray to the weight of the vapor,
\& shows that the condensate from the
fractional condenser should pass down
through all of the trays.
In Fig. 13, suppose that the con¬
densate p% is passed around the seventh
tray to the sixth tray. It yields there
vapor of composition qG, whose weight is
Fig. 13.—Fractional condensation.
pGn
By the partial condensation
Ps-p&q&
of this vapor on the seventh tray, the vapor q7 remains, and its
weight is p& . however the condensate p& passes first to tray seven, it is
Pzqs-Vni ;
PlO
partially condensed, yielding vapor q7 with a weight p& and condensate p7 with a
Piqi
weight ps-~Y‘ The condensate p7 then passes to tray six, where it yields vapor qG
having a weight pa- The vapor qG condenses partially on tray seven. The
oq7-pGr-p7?n
part which remains, vapor q7, has a weight p% • By shunting out tray
VilrPiqvP'iq-i
seven, the quantity of vapor q7 obtained was found above to be , while by

VaQQm P7Q7
having both trays in operation the quantity of vapor, q7, was pr-7°- -f pH.—LPAL2lI!L
p iQi V7<lfP*<R'Pnt
The latter is the larger quantity,1 and the using of all of the trays is therefore desirable.
In the above discussion, it has been assumed that the rate of boiling of the original
solution has been kept constant, i.e., that the heat-consumption of the still has been
maintained constant. Obviously, if a column is to effect the separation of the
constituents of the solution, there must be a fall of temperature from the bottom tray
to the fractional condenser, otherwise the vapor would pass the trays unchanged. In
, pro . oq7-per-prm - pan-prm
1 If . p - + pa*--p8*-
p7qi p7q7-ptqz-p7q7 V^q^Piqi
p7o'piqrp7q7 + oqrp&r'p7m '>pm‘p7m‘p7q7. (By dividing out the common factor ps, making all
denominators the same and then multiplying out the denominators.)
From the figure, it is seen that the left-hand member p7o-peq6*p7q7 + oqrpG r-p7m = p7o-p7q7(p7m +
Per) + oq7.par.p7m = p7o-p7qrp7m -J- p7o-p7qrp&r + oqrpmprtn (Since petfG = p7m + per).
Also the right-hand member pm* p7m. p7q7 = p7m-p7q7{p7o -]- per) = p7m'p7qrp7o + p7m- p7qrptr
= p7m'p7qrp7o + p7m'pGr(p7o + 037) = p7m’p7qrp7o -j- p7m' pmp7o + p7m-pm’oq7. (Since pm =>
p7o + per, and p7q7 ■» p7o + oq7).
By comparing the expanded members, it is seen that two of the terms are alike. The second term of
the left-hand member, piqrpio'pm is greater than that of the right member, p7m'p7o-pzr. (Since
vo >n m. an other wo factors are the same in each).
practice, the proper temperature-differences between trays is maintained, for a given

rate of working, by suitably insulating the column. The loss of heat by radiation,
etc., beyond the quantities necessary to maintain the desired temperature-differences,
is objectionable, since such loss of heat cools the liquid on a tray below its boiling-
point, thereby causing a loss of heat energy, since the liquid must be reheated by the
rising vapors. Cooling of the walls in contact with the vapor effects a fractional
condensation and a slowing of the working of the column which can only be overcome
by an increased number of trays or by an increase of the heat units supplied to the
boiler.
A given column works satisfactorily only under a given rate of vapor supply, so
long as its heat insulation remains unchanged. This rate must be such that the
necessary fall of temperature to effect the desired fractionation can occur. If the
vapor supply is decreased, the temperature falls on any given tray to such an extent
that the condensation of a solution rich in the lower-boiling product is brought about,
and the yield of the column is greatly decreased. If the vapor-supply is too great, the
required temperature differences are not realized, and the vapor leaves the top tray
with too great a content of the high-boiling material. By providing for the removal
of more heat from the column, e.g., by decreasing the efficiency of the heat-insulation,
the desired temperature differences may be had with the increased supply of vapor.
In this way the product of the same composition is obtained in increased quantity.
Likewise, the ratio of the weight of the product to that of the liquid flowing back into
the boiler is the same as before.
Quantitative Treatment of the Problems of Fractional Distillation.—The

theory of fractional distillation of Golodetz, based upon the principles of Konowa-
low, is theoretically sound, but since it is purely qualitative, it can be used only in
a general way in connection with problems of fractional distillation. The funda¬
mental question to be answered by a quantitative theory of fractional distilla¬
tion is as follows: What percentage of a solution of a given composition, X0,
must be removed by a simple distillation, in order to change the composition of
the solution to X? The theory of Golodetz is not able to answer this question.
Earlier investigators limited their investigations to an examination of the relations

between the total vapor-pressures and the vapor-pressures of the pure constituents
at corresponding temperatures. The values of the partial pressures and the composi¬
tions of the equilibrium vapors were largely neglected. As a result, little progress
toward a theory of fractional distillation was realized. Beginning with the work of
Wanklyn1 and of Berthelot2 in 1863, many attempts have been made to discover a
rigid mathematical relationship between the composition of the vapor and that of the
solution. As has been stated before, none of these are generally applicable, although
all are of some utility. The formula proposed independently by Wanklyn and Berthe¬
lot is as follows:
Xi WiDiPt
X2 W2D2P2
Xi and X2 are the weights of the two substances in the vapor, Xi referring
to the more volatile constituent; W1 and W2 the weights in the liquid; Di
and D2 the vapor-densities under the existing conditions; and Pi and P2 the vapor-
pressures of the pure constituents at the temperature in question.
The extended researches of Brown3 showed that the above formula, is not generally
1 Proc. Roy. Soc., Vol. 12, p. 534.
2 Com-pt. rend., Vol. 57, p. 430.
,
3 Jour. Chem. Soc., Vol. 35 p.547 (1879); ibid., Vol. , ,
37 p. 49 (1880); ibid,, Vol. 39 p. 304 (188 1).
applicable. As the result of his work, Brown proposed the formula
which was satisfactory for a number of binary solutions. Still better agreement
between observed and calculated results was found by substituting a constant, C,
P
for the factor jt' la general, the value of the constant, (7, approximated that of the
-T2
Pi for normal liquids. The difference between the two quantities became
factor
P2
greater as the divergence of the vapor-pressure curve from the simple straight-line
form increased. In some cases, the substitution of C0 + am for the constant, C,
gave good agreement. C0 and a are constants, and m is the weight (or molecular)
percentage in the liquid of the constituent which is the more volatile. Since
Xl
it was observed by Brown that for some pairs of liquids the ratios tf and
A2 W2
were the same whether the distillations were performed at atmospheric pressure or at
considerably reduced pressure, he applied his formula successfully to the data of other
investigators who made their distillations at constant pressure rather than at constant
temperature. This constancy of ratios, however, is not generally true. Rosanoff
and Easley1 have made calculations on the basis of Kirchoff’s equation, which indicate
that for substances which do not have too great heats of dilution, the ratios are not
greatly affected by quite considerable variations of temperature. (No great change
jp
in the ratio y is postulated).
Brown’s simple formula is probably true for pairs of liquids which have a vapor-
pressure curve of the straight-line form. It is not generally true, however, as can be
shown by considering solutions which have a maximum or a minimum vapor-pressure.
Since these solutions have the same composition for the vapor and the liquid at the
pressure of the maximum or minimum value, the constant has the value of unity.2
This value of the constant would require that the relative compositions are the same
at other pressures also, which is, of course, not the case.
Adopting Brown’s equation, Barrell, Thomas and Young3 outlined a theory of
fractional distillation. The utility of this theory is quite limited.
A set of equations was proposed by Lord Rayleigh.4 Since the partial pressures of
the constituents are required for each case, and the equations are limited to dilute solu¬
tions, these formulas are not generally of use.
A rigidly accurate relationship between the partial pressures of the constituents of
a binary solution and the mole fractions of the substances in the solution, is given
by the Duhem-Margules equation. This relationship is based upon thermodynamic
principles, the assumption being made that the mixed vapors follow the gas laws.
Luther5 and Lewis6 have derived the equation thermodynamically. In its usual form,
the equation is written as follows:
d log pi _ d log p2
d log x d log (1 — x)
Pi and p2 are the partial pressures of the constituents A and B> and x and 1 - x the
mole fractions of A and B in the solution.
1Jour. Am. Chem. Soc., Vol. 31, p. 953 (1909).
*1 TV1 Xi Wi
If 1.
X2 Wi X2 W2
3 Phil. Mag. (5), Vol. 37, p. 8 (1894).
*Phil. Mag. (6), Vol. 4, p. 521 (1902).
6 Ostwald’s “Lehrbuch,” 2d Ed., Vol. 3, p. 639.
6 Jour. Am. Chem. Soc., Vol. 28, p. 569 (1906).
A pair of solution functions for this differential equation has been proposed by
Margules.1 These functions are as follows:
Pl = piXaOgai(l - z) + f (1 - 3)3 + f (1 - *)» +-
p2 = Pad-x)e°e?'x + % + f +-
Pi and p2, have the significance which they possess in the differential equation.
Pi and P2 are the vapor-pressures of the pure components. aQ} <y2, a3 . • , and
i30, 0i, 02, 03. . . are constants.
The following relationships between the constants are obtained from the solution
functions and the differential equation:
00 = O£o — 02 = Q£2 “b «3 + OLi + . . (1)

01 = — o-l 03 = — OLZ — 2a 4 — 3as • • (2)
If each of these solution functions is put into the logarithmic form, and then
differentiated with reference to x} the curves of the tangents to the partial pressure
curves are found to have the following form:
it ~ t ' !“• ~ “•* “ «4»(1 — *) I (3)
™ + j9i(1 - x) + ^2 X (1 - x) + ... 1 (4)
When x has the value 1 or 0, these equations reduce to these:
(Sr),.,--<»•-<6>
From Raoult’s law, which is valid for dilute solutions, we have the following
relations:
Equations (1), (2), (5), (6), (7) and (8) give the following values of the constants:
a.o = 0o = 1 and «i = 0i = 0
Substituting these values in the original solution functions, the forms in which
the functions are generally used are obtained:
l2 a - *)*+? a - *)»+—
pi = Pxxe 3
-
Pi - P,( 1 - x)e 3
In order to evaluate the constants, the experimentally determined partial pressure

curve of one of the components may be used. Margules (Z.c.) has devised another
method which involves the determination of the tangents to the total pressure curve
at the values x = 0 and x = 1.2 Both methods of evaluating the constants are
only approximate, but their accuracy is in general sufficient.
An investigation of thirteen pairs of liquids was made by Zawidzki (Z.c.), and the
data were compared with the requirements of Margules solution functions. Certain
of these pairs of liquids, e.g., benzene and ethylene dichloride, gave the value 0 for
the constants a2, a3, 02, 0s, etc., and under these conditions, the functions reduce
to Brown’s equation. Other pairs of normal liquids gave finite values for the con¬
stants, and the agreement between the observed and the calculated values of the
partial pressures were fairly good (occasional discrepancies of 4 to 5 per cent). Za-
1 Sitzungsber. Wiener Akad., Vol. 104, p. 1243 (1895).

2 See Zawidzki, Zeit. phys, Chem., Vol. 35, p. 165.
widzki applied the formula with fair success to pairs of associated liquids by using the
average molar weights of the associated substances (in the liquid or the gaseous
phase) in calculating the mole fractions, x and 1 — x.
The inadequacy of the Margules solution functions with a small number of con¬
stants, has been emphasized by Bose.1 * In the event that a large number of constants
is necessary, the use of the equations is cumbersome. More satisfactory results can
then be obtained by a graphical method devised by Bosea and Marshall.3 4 * Story4
has suggested a method for using the functions by employing more convergent series
than had been applied previously. His method is applicable to ternary and more
complex systems, as well.
The rigidly accurate equations for the partial pressure curves have been derived
by van Laar6 by using the thermodynamic potential function and the principles given
by van der Waals (“Die Continuitat”). These equations take account of the differ¬
ential heats of dilution, and reduce to the simpler equations only when these heats of
dilution are negligible. It is evident, therefore, that measured deviations from the
simple equations (and these are the rule) are only to be expected, since the heats of
dilution are not as a rule negligible. The assumption that compounds or associated
forms are produced in the mixtures, is superfluous. Unfortunately, the rigidly
accurate equations are not generally utilizable on account of the lack of the
necessary data.
A relationship between the total and the partial pressures of certain binary mix-
urles has been found by Rosanoff, Bacon and Schulze.6 The equation showing the
treationship is as follows:
^=
ax
p-p
logP i — logP 2
logVALn£)
pix
(A)
P is the total pressure; Pi and P2 the vapor pressures of the components; pi and p2
the partial pressures of the components; x and I — x the mole fractions of the
components in the liquid phase. This equation holds for isothermal measurements.
Pi - P2
The constant term, is calculated from the values of Pi and P2,
log Pi — logP 2
obtained by expressing the experimentally determined isothermal total-pressure curve
as a parabolic function in the fourth degree of x and substituting x = 0 and x = 1.
The parabolic function is then differentiated, and the differential coefficient, divided
by the above constant, is equated to log • From this equation and the
known total pressure, which is equal to pi + £>2, the partial pressures are calculated.
By an analysis of Margules’ solution functions and a generalization of Raoult’s law,
Rosanoff and Easley7 show how the equation derived from Equation (A) and the
parabolic function can be modified to make possible the calculation of the partial
pressures of binary mixtures over a range of temperatures. The combination of the
two solution functions, gives the following equation:
, Pi (1 ~ x) 1 Pi .
log;
p*x
= lOg 5- +
Jr 2
«2 [(!-*>+!] -f[(!-*)* -|] +f[(l -*>'
-a1 Zeit. phys. Chem., Vol. 71, p. 637.
* Phyeitc. Zeit., Vol. 8, p. 353 (1907).
,
* Jour. Chem. Soc., Vol. 89 p. 1350 (1906).
4 Zeit. phys. Chem., Vol. 71, p. 129.
,
8 Zeit. phys. Chem., Vol. 72 p. 723-9.
,
8 Jour. Am. Chem. Soc., Vol. 36 p. 1993 (1914).
,
7 Jour. Am. Chem. Soc., Vol. 31 p. 953 (1909).
In all cases where the heat of dilation is small, the constants <*2, a3 and a ( are independ¬
ent of the temperature over a considerable range. Changes of temperature influence
P
the-composition of the vapor only through the term “log It is only necessary,
A 2
therefore, to express this ratio as a function of the temperature for the desired range.
The equation for isothermal values derived as described above, has the following form:—
logîi-= b cx -\-dx2 + exz
P2X
If any such expression for a temperature within the range is worked out, it may be
converted as follows into an equation which holds throughout the range. From the
p
right-hand member, subtract the logarithm of the ratio, p-> for that temperature, and
add the logarithm of the function which shows how the ratio depends on the
temperature.
By employing this relationship, Rosanoff, Bacon and Schulze1 have devised a
theory of fractional distillation. For organic liquids, which have small heats of dilu¬
tion («2, a3, a4, etc., independent of temperature over a considerable range), the
necessary data are limited to the isothermal total pressures of the solutions at a
temperature not too different from the temperatures existing during the distillation
and the form of the function showing the relation between the ratio Pi/P2 and the
temperature.
The derivation of the equations for showing the course of a distillation without
reflux condensation is carried out as follows: If wi and w2 are the weights of the
constituents in the liquid, and dwi and dw2 the weights in the vapor, the following
relationship holds.
Pi (1 — x) __ w2 dw 1
P2X Wi dw2
or
d log w\ _ pi (I — x)
d log w2 px2 (1)
From equation (A), showing the relationship between the total and partial vapor-
pressures (supra),
log 5) = k dP (2)
& p2x dx
where h is the reciprocal of
Pi -P*
log Pi - log P2
This equation may be written
dP
Pi (1 ~ x) dx (3)
p2x
From (3) and (1) we obtain
dP
d log wi _ x dx
(4)
d log w2
If the molar weights of the constituents are Mi and M2,
Wi/M
-> which yields
1 — x w2/M2
d log (wi/w2) = or
x (1 — x)
dx
d log w2 — din Wi —
X (1 — x) (5)
* Jour. Am. Chem. Soc., Vpl. 36, p. 1993 (1914).
From (4) and (5)
e dx dx
dlogw i =-gp- .(6)
k-
(e - 1) (1 - x)x
If the initial weight and the initial mole fraction of the first component are Wx and
X0, respectively,
k dl
'S.
wi e dx
log dP
(7)
wi
(e <** - 1) (1 - x)x
For the second component, the solution is
/»(1 ~ *)
da: (8)
W2 dP
J (1(l *o) (e
- so) ( dx - 1) (1 - «)»
where W and 1 — x0 are the initial weight and mole fraction of the second
constituent.
The integration indicated maybe performed by any of the methods of approxima¬
1 — x Mj
tion. Since w2 = • wi} only one of the integrations needs to be performed.
x Mi
If the total and partial-pressure curves are straight lines, equation (4) reduces to
Brown's equation, and the integrals assume a simpler form.
Obviously, the theory is applicable to all cases in which the relationship (equation
A, p. 634,) between the total vapor-pressure and the partial vapor-pressures is valid.
Constant Temperature Still-heads.—Enrichment of a binary vapor in the

lower-boiling constituent may be effected by fractional condensation in a single
constant-temperature still-head or in a series of such still-heads. Warren used
such an apparatus for the fractionation of such complex mixtures as crude petro¬
leum, coal-tar, etc., in 1864. In the use of this still-head, the principle of cooling
is depended upon, as the washing of the vapor with the condensate is quite
incomplete.
The use of a constant-temperature still-head for separating binary mixtures, was
investigated by Brown.1 For the pairs of liquids used, Brown found a very simple
rule to hold: Whatever the composition of the liquid in the boiler, the vapor leaving
the still-head had a constant composition, depending only upon the temperature
of the still-head. This composition was the same as that of the vapor which was in
equilibrium with the particular binary mixture which boiled at the temperature of
the still-head (at the pressure of the distillation). By maintaining the temperature
of the still-head only slightly higher than the boiling-point of the lower-boiling con¬
stituent, Brown was able to affect the separation of quite pure products.
Young and Thomas2 have secured very good results in fractionating the low-boiling
fractions of Pennsylvania crude oil by using a constant-temperature still-head above
a Young and Thomas dephlegmator. The fractional condenser above certain of the
column stills is an example of this practice.
Rosanoff and Bacon3 have found that for binary mixtures which have maximum or
minimum boiling-point curves, Brown's rule must be modified. The composition of
the constant-boiling mixture is, of course, not altered by the still-head. In using the
1 Jour. Chem. Soc. Vol. 39, p. 517 (1881).
2 Jour. Chem. Soc., Vol. 71, p. 440 (1897).
boiling-point curves for ascertaining the effect of a constant-temperature still-head,

the usual diagram must be divided into two portions by passing an axis of ordinates
through the composition representing the mixture of maximum or minimum boiling-
point. The portion of the diagram to each side of this axis must be treated as a binary
mixture, the constituents being the constant-boiling mixture on the one hand, and one
of the pure constituents, on the other. As can be seen readily by drawing the curves
for liquids having a minimum-boiling mixture, each temperature between the boiling-
point of the lower-boiling constituent and that of the constant-boiling mixture is the
boiling-point of two solutions of quite different compositions. The vapors emitted
by these two solutions at the boiling-point are likewise quite different. Rosanoff
and Bacon find that the composition of the vapor leaving the constant-temperature
still-head is constant for the cases being considered also, but that it depends upon the
concentration of the solution charged into the boiler in this way: If the composition is
represented in that part of the diagram to the right of the axis drawn through the
point representing the solution of constant boiling-point, the vapor will have the
composition corresponding to that solution having the boiling-point equal to the still-
head temperature, as shown in this portion of the diagram. If the composition of the
original solution is found in the portion of the diagram to the left of the axis, the com¬
position of the vapor leaving the still-head is located in this portion of the diagram.
The behavior of the ternary system carbon tetrachloride, toluene and ethylene
bromide, when distilled through a constant temperature still-head, has been studied
by Hosanoff, Schulze and Dunphy.1 Solutions of these substances yield boiling-point
curves and a boiling-point surface free from maxima and minima. The authors state
their conclusions as follows:
“In distillations with a still-head maintained at a constant temperature, the
composition of the distillate is at every instant identical with that of a vapor evolved
by a mixture whose boiling-point equals the temperature of the still-head. If the
mixture is binary, the composition of the distillate is, in the course of a single distilla¬
tion, constant. In those cases in which the binary boiling-point curve passes through
a maximum or a minimum, the composition of the distillate depends on that of the
mixture originally placed in the still. If the number of substances in the mixture is
three or more, the composition of the distillate not only depends on that of the original
mixture, but varies in the course of a single distillation. This variation, however, is
moderate, and the nearer the constant temperature of the still-head is to the boiling-
point of the most volatile component, the more nearly constant is the composition of
the distillate.17
The conclusions embodied in the last sentence are based upon the results with the
one ternary system, and are probably too broad for some cases. Regulated tempera¬
ture still-heads are nevertheless interesting in their action, and offer a promising field
of investigation in connection with their commercial application.
Quantitative Treatment of Fractional-distillation Problems.—In HausbrandV
“Operation of Rectifying and Distilling Equipment77 complete sets of equations are
presented for solving the problems of fractional distillation as carried out in commer¬
cial equipment. Although a number of simplifying assumptions are made, the results
obtained with the various mixtures dealt with (binary) are sufficiently accurate for
purposes of operating or designing commercial equipment. Both the heat economy
and the degree of separation of the constituents obtained are considered by Hausbrand.
In order to use these equations, it is necessary that the liquid composition-boiling-
point curve, the vapor composition-boiling-point curve, and the heats of evaporation
of the constituents be accessible.
A discussion of these equations can not be undertaken here. They are, however,
2 “Rektifizier-und Destillier-Apparate,’* E. HaxrsBKAND, 3d ed., Julius Springer, Berlin, 1916,
approximate formulas, based upon the principles outlined above. Quantitative treat¬
ment of some of the problems of fractionation are also given by Lewis1 and by Gay.2
The basis of treatment is in both cases similar to that of Hausbrand.
Solutions containing more than two constituents may often be treated as binary
mixtures, by choosing the most volatile constituent and the sum total of the others
as the components.
Sublimation.—Solid substances possess definite vapor-pressures at given

temperatures, just as is the case with liquids. The process of vaporizing a solid
is called sublimation. The process may be carried out on a solid below its melt¬
ing point at the pressure employed, by continuously supplying heat. The
number of heat units required per gram of substance is the heat of sublimation,
which is analogous to the heat of evaporation of a liquid. The quantity of heat
liberated when a gram of vapor condenses is numerically equal to the heat of
sublimation of the same substance at any given temperature. Sublimation
curves are in all respects analogous to vapor-pressure curves. The molar heat of
sublimation, which is the number of heat units required to volatilize one gram
mole of a solid, is strictly analogous to the molar heat of evaporation, and prob¬
lems of sublimation are dealt with by methods which parallel those of
distillation.
The melting-points of certain substances are higher than the temperatures at which
the vapor-pressures reach the value of one atmosphere. When heated at pressures
below this value, therefore, it is impossible to melt these substances. All of the heat
furnished is used up in heating the solid to its “boiling-point,” i.e., the temperature at
which the vapor-pressure equals the external pressure, and in subliming tne solid. By
increasing the external pressure sufficiently, it is possible to raise the “boiling-point”
of such solids above the melting points at the pressures used. By this expedient, it is
possible to liquefy these substances.
In subliming mixed solids which do not form solid solutions, the relationships are
the same as those discussed above for immiscible liquids. In general, the processes
which are carried out practically, aim to separate one substance from others which
may be considered non-volatile under the conditions used. For example, crude sul¬
phur is sublimed in order to recover practically pure sulphur. The non-volatile
earthy residue is then discarded.
The sublimation of solid solutions has not received much consideration. From the
theoretical side, the treatment of the problems involved is in practically all ways the
same as that of the problems of miscible liquids. Experimentally and practically
there is the difference between the two cases, that in reestablishing equilibrium the
exchange of constituents between the vapor phase and the other phase is not so readily
effected with solid solutions as with liquid solutions.
DESTRUCTIVE DISTILLATION
By D. Drogin8
The process of destructive or dry distillation, also known as cracking as in the

case of petroleum, consists primarily of heating organic substances such as coal,
wood, shale, petroleum, tar, asphalt, etc., in a still, furnace or retort, in the
1 Journ. Ind. and Eng. Chem., Vol. 1, p. 522 (1909).
2 Chimie et Industrie, Vol. 3, p. 157 (1920).
* Petroleum engineer, 961 Frelinghuysen Ave., Newark, N. J.
DESTRUCTIVE DISTILLATION 639
absence of air or in the presence of a limited supply, with the object of making
from these substances products of industrial value, as well as to obtain residues
of commercial importance. The yield of the products obtained, their quantity
and quality, depend largely on several factors, the important ones being: (a)
the kind of still, furnace or retort used; (b) the volume, and in case of a solid, the
size of the particles constituting the charge; (c) the degree of heat employed in the
distillation; (d) the use of reduced, atmospheric or increased pressure; (e) whether
or not steam is used; (/) the presence of a catalytic agent, and finally; (g) on the
manner of condensation and purification of the ensuing vapors.
The chemical changes involved in the process of destructive distillation are based
on the fact that the subjecting of hydrocarbons, whether solid, liquid or gaseous
to a high temperature tends to break down the more complex molecules into simpler
molecules. Thus, in the destructive distillation of coal, the following reactions take
■place: Solid cool-^high molecular weight paraffins—>low molecular weight paraffins—*
acetylenes, naphthenes and polycyclic compounds—>benzene and its homologues.1
Oil Cracking.—The main products of cracking paraffin oils at low tempera¬

ture (400-500°C.) are paraffins and olefines. At about 700-800°C. aromatic
bodies like benzene, or toluene, are formed, together with olefines, diolefines and
small quantities of paraffins. The more complicated the molecule the more
readily it decomposes. This also holds true for unsaturated hydrocarbons.
At the highest temperature there is complete decomposition of hydrocarbons
into their elements. These changes are much influenced by alteration in tempera¬
ture, pressure and catalysis.
In general, it may be said that the paraffins begin to break up at much lower
temperatures than the aromatic compounds and that their decomposition is progres¬
sive, i.e., the high molecular weight paraffins go to low molecular weight paraffins
and higher olefines. The olefines then break up into carbon, methane, acetylene and
hydrogen. The paraffins again break to olefines and paraffins and the methane finally
breaks down in carbon and hydrogen.
Recent investigations by Rittman2 and others as to the thermal reactions of petro¬
leum and coal tar hydrocarbons show that the series of products formed by the
cracking reaction under considerable pressure are as follows:
Heavy hydrocarbons<=^light hydrocarbons, saturated and unsaturated^cymene^
xylene ^ toluene ^ benzene, naphthalene, diphenyl, etc. anthracene ^ carbon and
gas. Although reversible the reactions take place almost entirely in the forward
—►direction.
Egloff and Twomey8 showed that at atmospheric pressure the order of the reaction
is the same, namely, higher benzene homologues—^benzene (diphenyl)—► naphthalene
—►anthracene—^carbon and gas.
An increase of temperature favors the production of substances whose formation
is accompanied by the absorption of heat i.eendothermic bodies; conversely, the
formation of bodies whose formation is attended by the evolution of heat, i.e.j exo¬
thermic bodies, is checked by rise of temperature.
The use of steam in the process of destructive distillation is to aid the pro¬
duction of hydrocarbons at as low a temperature as possible. Ordinarily, these
hydrocarbons would come over at their boiling point at atmospheric pressure.
,
1 M. C. Whitaker and W. R. Crowell, Jour. Ind. Eng. Chem., Vol. 9 1917, pp. 261-269.
2 W. F. Rittman, C. B. Dutton and E. W. Dean, Journ. Ind. Eng. Chem., Vol. 8, 1916, pp. 351-361.
,
8G. Egloff and T. J. Twomey, Journ. Phya. Chem., Vol. 20 1916, pp. 121-150.
But, with the introduction of steam, the vapor pressure of the mixture, being
the sum of the partial pressures of the hydrocarbon vapor at that temperature
and the water vapor, is equal to the atmospheric pressure. Therefore, the hydro¬
carbons come over at a temperature below their boiling points. Either
superheated or wet steam is used.
The use of steam is recommended in the destructive distillation of shale in Scotland.
The steam equalizes the temperature in the retorts, preventing excessive cracking
of the oil vapors and reducing the amount of unsaturated hydrocarbons. It increases
the yield of ammonia and prevents the formation of nitrogen compounds of an unde¬
sirable nature, such as pyridine. According to C. W. Botkin,1 steam distillation
increases the yield of oil a few gallons per ton and the specific gravity of the oil thus
obtained is higher by 0.03 to 0.04 than when a dry distillation is made. The yield
of NH3 is increased two to three-fold when steam distillation is used.
Winchester,2 in six tests to determine the comparative values of steam and dry
distillation in the yield of oil from shale in Colorado and adjacent areas, found that
with four out of the six samples from the same lot of shale the use of steam causes
an increased yield of from 10-50 per cent of that obtained with dry distillation.
There are certain disadvantages in using steam in connection with shale distilla¬
tion. It increases the volume of volatile products formed. This requires the use of a
large condensing apparatus. The oil and water when condensed, form an emulsion
which is very slow in separating by settling. The heat units carried away from retort
by steam nearly equal what are required to decompose the shale.
In the destructive distillation of resinous wood, steam is introduced to extract
the turpentine already present in the wood. Steam also takes part in the production
of important gaseous fuels, such as water gas, which is formed by drawing air with
steam through a deep bed of coke in a closed retort. The reactions which take place
are three in number:
C + 2H20 = C02 + 2H2

C + H20 - CO + H2
CO 4- H20 « C02 4* H2
The course of the reaction between the steam and carbon depends chiefly upon
the temperature. At 600°C. (1,112°F.) the products are principally C02 4- H2; at
higher temperatures less C02 is formed and more CO; at 1000°C. (1,832°F.) the reaction
is nearly that represented by the equation: C 4- H20 = H2 4- CO.
The effect of pressure on petroleum cracking can be said to be as follows:

Increase of pressure favors the formation of products occupying a smaller volume,
conversely a diminution of pressure favors the production of substances occupying
an increased volume. Generally, an increase of pressure favors the formation of
unsaturated compounds and carbon.
Destructive distillation under pressure provides a considerable output of “cracked”

gasoline in the United States.
In thevacuum process of distillation a partial vacuum is created in the still by the
aid of a pump. The stills employed are generally of the same type as the ordinary
horizontal crude still, but smaller, heavier and strongly braced. The hydrocarbons
come over at a temperature considerably lower than their boiling points at atmospheric
pressure. There is an improvement both in the quantity and quality of the
distillation products when vacuum distillation is combined with steam distillation.
1 C. W. Botkin, Quarterly, Colo. School Mines, 4, No. 4, 16-17 (1919).

2D. E Winchester, U. S. G. S., Bull, 641-F, 139-98 (1916).
Catalysis.—Chemical changes can be made to take place more readily in the

presence of a catalyst. The temperature need not then be so high and the reac¬
tion can be conducted with a greater degree of ease and a smaller loss. The
vapors generated in the destructive distillation of heavy oils and petroleum
residues are caused to come in contact with catalytic materials heated to a high
temperature ranging from 400°C. up. These materials may consist of finely
divided metals, which are prepared by reducing the corresponding oxides with a
reducing gas at an elevated temperature;1 or heated graphite, aluminum chloride,
iron shavings, fullers earth, metallic oxides and coke. Some catalysts are readily
poisoned by the sulphur in the oils.
T. T. Gray2 has developed a very interesting process for the decolorization and
purification of the distillates obtained from cracking processes as well as those obtained
by the distillation of crude petroleum by the usual methods, in which the products
to be purified are passed in contact with fullers earth, bauxite, nickelized fullers earth
or other catalysts maintained at a temperature sufficiently high to keep the products
treated in vapor form. By this method undesirable compounds are recomposed or
polymerized and removed by the catalyst. It is probable that hydrogenation takes
place when the vapors containing cracked hydrocarbons and hydrogen come in contact
with the catalyst. By this process it is possible to make products of merchantable
color without treatment with sulphuric acid and it is said to increase the yield of gasoline
obtained by pressure or other such treatment from 12 to 14 per cent. It can be applied
to any cracking process or still operation and it is thought it will eliminate the acid
treatment of all these products.
Destructive Distillation of Shale.—If a shale is subjected to destructive

distillation the products of value obtained from it are oil, ammonia and inflam¬
mable gas. All shales do not give similar results. Those from Scotland furnish
more ammonia and less oil than shales from the United States. The form of
retort in which the shale is distilled and the conditions under which the distilla¬
tion is conducted are largely responsible for the yields obtained.
The distillation of shale is at present an important industry in Scotland. In forty-

five years, from 1871-1916, the production of shale oil rose from 593,310 bbl. to
1,965,000 bbl. and that of ammonium sulphate from 2,350 tons to 59,400* tons. The
average yield of the main products obtained from a short ton of shale retorted is
as follows: Crude oil, 24.5 U. S. gal.; ammonium sulphate, 35.7 lb.; gas, 9,800 cu. ft.
(240 B.t.u. per cubic foot); spent shale, 80-85 per cent.
In America the industry has not as yet gained a foothold, though an unusually
large number of processes exist for the retorting of the shale. In this country the
shales differ widely in physical and chemical properties. It is, therefore, not commer¬
cially feasible to adopt a single type of retort for all.
The oil in shale usually does not exist in the free state, but is produced in the
process of destructive distillation from an oil-forming substance known as “kerogen.”
The oil is highly unsaturated and is composed largely of hydrocarbons of the paraffip,
olefin and diolefin series, a small amount of naphthenes, benzenes and 1 per cent of
nitrogen.
For retorting shale in Scotland the Pumpherston type of retort used in the Pump-
herston works, gives the best results. Several advantages are combined in it, for it is
simple and economical in construction. The feed is continuous, likewise the removal
of the oil, gas and spent shale.
1 Sabatier and Mailbe, U. S. Pat. 1124333 (1915).
2 T. T. Gray, U. S. Patent 1340889 (1920).
41
G42 CHEMICAL ENGINEERING
The retort is circular throughout, vertical and tapers. The upper part is made
of cast iron about 12 ft. high, 2 ft. in diameter at the top and 3 ft. at the bottom. The
lower part is made of firebrick and is about 18 ft. high and about 3 ft. in diameter at
the top where it joins the iron portion. The average capacity per day is four and a
half short tons per retort.
The retort is heated by the permanent gases resulting from the destructive dis¬
tillation of the shale. This is supplemented by producer gas when there is a shortage
of permanent gas.
During the process of distillation, steam is introduced. About 750 to 1,000 lbs.
of steam are used for every ton (2,240 lbs.) of shale retorted. The advantage's of steam
are manifold. The temperature in the retort is rendered uniform by the transfer of
heat from the walls of the retort to the center of the charge. It minimizes the cracking
of the oil; it helps to sweep the oil vapors out of the retort; it cools the spent shale and
it aids in the production of ammonia and water gas.
The steam is decomposed when it comes in contact with the rod hot shal(‘. The
hydrogen unites with the nitrogen in the shale to form amlnonia, and the oxygen com¬
bines with the carbon to form water gas.
The distillation of shale is carried out in the following way. Pieces not larger
than 0 in. square are delivered to a hopper at the top of the. retort, and fed to the retort
by gravity. The upper part is heated externally to a dull red heat. The tempera¬
ture at this point ranges from 000-700°P.
In the lower part of the retort the temperature of the shah' is about 1,700 1,800”P.
In this part most of the ammonia and water gas is formed by the action of the steam.
The spent shale is discharged from the retort into a sealed hopper, into which steam is
admitted.
The oil and water vapors and the gases consisting of ammonia, hydrogen, carbon
monoxide, carbon dioxide and some free nitrogen, enter a common main about 30 in.
in diameter. The main leads the gases and vapors into a water heater, where water
for the steam boiler is heated in pipes, then through air-cooled vertical condensers.
The condensed oil and water vapors arc led into separators where tin* ammonia water
is separated from the oil. The gases are passed info a water scrubber to remove last
traces of ammonia and then into a naphtha scrubber where the gas is washed with
mineral oil to absorb the naphtha. The gases are then list'd as fuel for heating the
retort.
The ammonia water obtained from the separators and scrubbers is distilled, and the
ammonia is passed into sulphuric acid to form ammonium sulphate.
The crude shale oil as well as the scrubber oils, are fractionally distilled. Scotch
shale oils suffer an average refining loss of 25 per cent. Then' is only about 7 per
cent loss in completely refining petroleum.
The percentage yield and the products obtained from the shah' oil largely depend
on the nature of the shale treated and on the conditions of retorting and refining.
The products obtained from shale oil in a large, plant in Scotland an* as follows:
Vku (*kni*
Naphtha (including scrubber naphtha).... 9.9

Burning oil. 24.7
Gas and fuel oil. . .. 24,4
Lubricating oil...... . . 0.9
Wax...... . .... 9.5
Still coke.. . . 2.0
Loss. 22,9
100.0
a
o
UJ
o
ID
CC
U
<
5
<
2
z
z
UJ
0-
z
o
<
^J
K
tO
5
£3
ID
a:
h-
vO
UJ
o
o
tu
CO
o
o
ID
a
o
od
Q-
<
Destructive distillatioii of petroleum is employed when a large yield of gasoline

and illuminating oil is desired. This form of distillation is especially adapted to
crudes containing little gasoline or kerosene (California and Mexican) or those
unfit for the manufacture of cylinder stock (Mexican),
Horizontal underfired cylindrical stills having charging capacity of from 200 to

1,200 bbls. are generally used in the distillation of crude oil. A still with charging
capacity of 500 bbls. is 10 ft. in diameter and 40 ft. long. A 1,000-bbl. crude still is
14 ft. 6 in. in diameter and 40 ft. long. In cracking processes the upper part of the still is
generally left exposed to the air. In the ordinary process of distillation the stills are
completely bricked in, so that the vapors shall be kept fully heated until they escape
to the vapor pipe. The continuous operation in petroleum distillation does not per¬
mit of sufficient cracking to obtain highest yields of light products.
In case of a crude of paraffin base the distillation is usually carried out in a tower
still. This is an ordinary still provided with an aerial condenser through which the
vapors must pass before entering the water-cooled condenser. The oil is fire stilled
and the heavy vapors which condense in the top of the still and in the aerial condenser
fall back in the superheated oil and are thereby cracked or partially decomposed.
The distillation is carried down without interruption to coke, thus separating the
products as naphtha, illuminating oils and paraffin oils in one distillation.
In the accompanying charts are indicated the main products obtained in the
destructive distillation of a paraffin base crude (Pennsylvania) and a mixed crude
(Mid-continent).
When a large yield of lubricating oil of good quality is desired from crude petroleum,
the distillation is begun in the ordinary way. Steam is let into the oil when the tempera¬
ture in the still is about 230°F. and the crude naphtha is distilled off when the still
temperature is about 280°F. Without steam the temperature would be about 375°F.
Likewise the heavy crude naphtha is over with steam at a still temperature of about
330°F. Without steam the temperature would be about 475°F. The lamp oil is
over at the still temperature of 500°F. compared with 630°F. when no steam is used.
The amount of steam is gradually increased and the distillation is conducted as rapidly
as possible in order to avoid cracking until the wax distillate has distilled over. The
still temperature is now 620°F. The residue consists of cylinder oil stock.
Steam is always used when cracking is to be avoided. It passes up through the
oil as a gas and is condensed with the oil vapors. The maximum temperature attained
with steam in oil distillation is about 600°F. Without its use the still would have to
be heated to a temperature of about 750°F. Such temperature would result in obtain¬
ing an oil below the desired standard. The use of steam, therefore, avoids destructive
distillation.
When distillations are carried on at increased pressures the charge usually con¬
sists of heavy oils or petroleum residues like fuel oil and gas oil. The stills employed
when high pressure and temperature is used are of various shapes and construction.
They may consist of the following: (1) heavily constructed retorts; (2) vertical
stills; (3) cracking tubes; (4) combination of stills and cracking tubes; (5) horizontal
cylindrical reservoir in series with multiple connected heating tubes; (6) column stills
in combination with an horizontal still; (7) still internally divided in compartments;
(8) shell stills; (9) rotating stills.
The technical processes for making motor gasoline and motor fuel by destructive
distillation products at high pressure and temperature, employ either stills, tubes,
or special catalytic methods.
It. F. Bacon and C. W. Clark1 have found that the yield and quality of gasoline
begins to decrease at pressures below 100 lb. in the still and that about 300 lb. is the
upper limit for practical purposes. The percentage of saturated hydrocarbons is
greater when the distillation is carried out at pressures about 100 lb. than at pressures
below it.
In the Burton process stills having a capacity of about 250 bbl. are used. Petro¬
leum residues are distilled at a temperature of 370-420°C. at a pressure of about 80
1 R. F. Bacon and C. W. Clark, U. S. Pat. 1101482.
o
UJ
o
3
cc
o
UJ
z
z
o
o
I
£
<4-
o
Z
o
<
£
o
lU
o
3
a:
to
UJ
a
-F
C
O
uj
o
3
o
o
a:
a.
5
<
5
lb. per square inch. This pressure is maintained by means of a valve placed at the
outlet of the condenser. Heavy coke forming fractions settle out at the bottom of the
stills when high boiling hydrocarbons are cracked and if not removed lead to local
over heating of the still bottoms with all the attendant inconveniences arising there¬
from during the course of distillation. To overcome these difficulties many schemes
have been patented employing in general such methods as watercooled jackets, cir¬
culation of the oil from one still to another one through a trap in which the coke forming
substances are deposited; the use of scrapers, etc.
The Hall process embodies the use of tubes. The oil is vaporized and passed
through a system of continuous coils each about 600 ft. long and 1 in. in internal
diameter. It is fed in with a pressure of 50-75 lbs. per square inch at the rate of over 70
gal. per hour. It enters at the cooler end of the retort and is gradually heated and
vaporized. These vapors attain a speed of 500-6,000 ft. per second in the last portion
of the coil where the temperature is the highest. It is claimed that any material
deposit of carbon is prevented by incomplete cracking and the high rate of flow.
When working for gasoline the temperature of the exit tube is kept at 500-600°C.
The Valpy Lucas and the McAfee processes both make use of special catalysts.
In the former1 a mixture of oil and steam is brought in contact with a heated catalyst
such as nickel or other metal or metallic oxide. The oil to be cracked is vaporized by
super-heated steam in a separate still and the vapors are run through a cracking tube
packed with the catalyst and heated to 550-600°C.
In the McAfee process2 the crude is first heated to free it from water. Anhydrous
aluminum chloride is added and the mixture boiled between 500-550°F. for 24-48
hr. Between the still and final condenser are placed two aerial condensers con¬
nected in series which separate the low boiling oils from the high boiling fractions
returning the latter to the still together with any volatilized aluminum chloride. The
distillate obtained contains a mixture of gasoline, solvent oil and kerosene. The prod¬
ucts are claimed to be all water-white, sweet smelling and saturated. Later3 McAfee
combined the aluminum chloride process with the action of an ordinary pressure still
in such a way that the yield of light naphtha from the latter is materially increased.
Bittman and Egloff show that starting with an oil consisting almost entirely of
paraffin hydrocarbons, it is possible to produce by cracking various aromatic hydro¬
carbons. The method pursued is as follows: High molecular weight paraffins when
subjected to destructive distillation after being first converted to paraffins of lower
molecular weight and olefines, are further converted into acetylenes, naphthalenes and
polycyclic compounds. The final products are benzene and higher and lower homo-
logues of benzene.
Destructive Distillation of Asphalt.—Asphalt is a black non-oxidized bitumin¬

ous hydrocarbon occurring in the heavy residuum from petroleum or in natural
deposits. The residues from petroleum are known as oil asphalts and come
largely from California, Mexican, Texas and Mid-continent petroleums. The
most commonly used natural asphalts are Trinidad, Bermudez, Cuban and
Gilsonite.
Asphalts on destructive distillation first melt; at about 300°C. vapors arise, which
when condensed yield a colorless liquid with concentrated petroleum odor; at about
300°C. decomposition begins, H2S is evolved and odor of the distillate changes. The
residue consists of porous coke.
Asphalt hydrocarbons begin to decompose at a temperature of 600°F. or slightly
below it. The destructive distillation yields distillates which are highly unsaturated.
The proportion of unsaturated hydrocarbons in distillates from asphalts and pyro-
asphalts is generally much greater than is to be found in crude-oil distillates,
particularly of the paraffin and semi-paraffin type.
The destructive distillation of Gilsonite4 yields 25-30 per cent coke, 10 per cent
1 English patent, 20470 (1913).
2 A. M. McAfee, Trans. Amer. Inst. Chem. Eng. 1915, Vol. 8, pp. 177-95.
aU. S. Patent 1326073 (1919).
,
4 C. F. Mabery, Jour. Amer. Chem. Soc., Vol. 39 1917, pp. 2015-2027.
gas, largely methane; and 60-65 per cent liquid distillate which is highly unsaturated.
It contains hydrocarbons of the series CnH2n~2 and a considerable amount of ammonia.
Assyrian asphalt1 containing no sulphur on destructive distillation gave 4.8 per cent
water; 45.9 per cent liquid distillate resembling that of petroleum; 43.5 per cent
coke and 6.3 per cent gas and loss.
The destructive distillation of asphalt from Val de Travers2 yielded a brown crude
oil which on fractional distillation gave the following results:
Table 1
1 2 3 4 5 6
Boiling pythnt, degrees. . 90-120 120-150 150-180 180-200 200-220 220-250
Specific gravity... 0.784 0.790 0.802 0.817 0.845 0.867
Carbon, per cent.. 87.56 87.56 87.31 87.34 87.48 87.40
Hydrogen, per cent-. 12.34 12.50 12.59 12.69 12.60 12.40
The results indicate that the oil was composed of hydrocarbons of the CnH2«_2
series (acetylene). Similar results were obtained from the destructive distillation
of an asphalt from the Dead Sea.3
The following results were obtained by distillation of the various asphalts at
atmospheric pressure, the percentages.
Table 21
Trinidad, Bermudez, Gilsonite, Tablyite, Wurtzilite, Ozokerite,

Temperature per cent per cent per cent per cent per cent per cent
off off off off off off
1
0-150° C. 14.93 9.89 9.34 3.12 16.15 0.21
150-200°C. 10.42 7.99 5.34 11.93 21.70 8.91
200-250°C. 2.26 16.08 12.84 24.87 22.82 8.38
250-300°C. 21.12 28.99 13.21 0.91 17.69
300-350°C. 4.77 25.89
.
350-400°C. . 26.85
Total volatile... 27.61 55.08 56.51 57.90 61.58 87.93

Fixed carbon .. 36.69 39.60 43.13 37.45 36.92 10.03
Ash. 35.70 5.32 0.36 4.65 1.50 0.04
i C. Bardwell, B. A. Berryman, T. B. Brighton and K. D. Ktjhrb, Jour. Jnd. Eng. Chem., Vol. 5,
1913, pp. 973-976.
Destructive Distillation of Wood.—The crude products obtained in the

destructive distillation of hard and resinous woods are charcoal, vapors and gases.
The charcoal is a finished commercial product containing as high as 25 per cent
volatile matter. On condensation the vapors yield a pyroligneous acid and tar.
The acid is composed of an aqueous solution of acetic acid, acetone, methyl
alcohol, methyl acetone, aldehydes, dissolved tar and oils. The tars on treat¬
ment yield flotation oils, solvent oils, heavy tar oils for paints, preservatives
1 K. Kharichkov, Chem. Abs., 1910, p. 1235.
2 C. Volceel, Ann. der Chem., 1853, Vol. 87, pp. 139-148.
3 B. Delachanal, Compt. rend., 1883, Vol. 97, pp. 491-494.
and insulating pitch. The oils are valuable raw materials for medicinal beech-
wood creosote.
The gas contains large quantities of C02 and its main combustible constituents
are CO and CH4. Its calorific value is low and it is ordinarily used for fuel.
From resinous woods, such as yellow pine and fir, the chief commercial products
are charcoal, turpentine and tar. Hard woods, such as oak, beech and maple, yield
larger percentages of alcohol and acetic acid and the resinous woods more tar and
oils.
For a long time the distillation of wood was conducted without the recovery of
most of the byproducts; charcoal was the main product obtained. For this purpose
charcoal kilns or pits were used. The latter varied in shape and size, and the kilns
were either beehive in shape or rectangular. Air was admitted directly to the kiln,
and the heat necessary for the distillation of the wood was generated by the partial
combustion of the charge and its distillation product.
In the modern method for dry distilling wood, whereby the byproducts are
recovered, a standard type of a distillation apparatus, or retort is used. These
are either horizontal or vertical. Kilns of the horizontal type may be either rectangu¬
lar or cylindrical with an average charging capacity of about two cords.
Ovens of the horizontal type are much used. They are 6 ft. in diameter, 7 ft.
high and vary in length. They may have a charging capacity of 4-8 cords of wood,
depending on the length of the oven. They are usually set in furnaces with one fire¬
box for two retorts. Vertical retorts generally yield a charcoal of high carbon content.
The wood is first air dried for 12-18 months; or artificially dried in much less time
by the flue gases from the retort stacks. The actual operation of the distillation
largely depends on the nature of the wood treated. The products obtained and their
quality as well as the yield are largely influenced by the temperature employed.
The distillation of wood is an exothermic reaction. A certain amount of the wood
decomposes below 270°C., largely into C02 and water. The exothermic reaction
begins about 270°C., and the complete decomposition of wood into charcoal, vapor and
gas takes place without any further addition of heat.
Two series of reactions, primary and secondary, take place in the course of the
distillation. The methyl alcohol and acetic acid are among the primary products
during the first decomposition of the wood. These primary products react further
yielding secondary compounds. The velocity of the reaction depends on the tem¬
perature, proceeding very rapidly from 270 to 300°C. Slow distillation produces
more secondary reactions because the vapors remain in the retort for a greater length
of time.
The liquid resulting from the condensation of the vapors separate into two layers,
the crude pyroligneous acid on the top and tar underneath. The acid layer con¬
sisting mainly of crude acetic acid is subjected to various treatments, depending on
the quality of the products to be made, i.e., whether brown acetate of lime or gray
acetate of lime is desired. To make brown acetate of lime, the acid is neutralized
with lime, and the insoluble tarry products removed. The solution is then distilled,
until the wood alcohol is collected and the rest of the liquid is evaporated to dryness.
To make gray acetate of lime, the crude pyroligneous acid is distilled to separate it
from the dissolved tar, the distillate is neutralized with lime, allowed to settle and
the clear liquid containing the wood alcohol and the oils is distilled. The acetate
liquor is taken from the still and evaporated to dryness. The tar is steam distilled
to remove any acid and methyl alcohol contained in it.
Destructive Distillation of Coal.—When coal is heated in the absence of air

or in the presence of a limited supply, chemical reactions occur that are highly
/ 649
uj £ uj
xc cc
S-g S
CL. <_>
complex and the new substances that app >f the coal are gas, liquid
products (water and tars) and solid residue he yield and character of
these products from any coal may be va: y changing the external
conditions of temperatures, pressure, time surface.
The retorts used in the destructive distillation of coal vary considerably in shape,
cross-section and in length. They are set in groups or “benches” of from six to nine
retorts. These may be heated by a direct fire of coke or coal, or by producer gas,
which is the practice in larger and more modern plants. The retorts are arranged
either horizontally or vertically and charged with from 3 to 8 tons of coal each.
In commercial practice the coking of coal is done in beehive and byproduct ovens
and in gas retorts. The beehive and most byproduct ovens are used to produce
metallurgical coke, whereas gas retorts are used for obtaining illuminating gas.
The coke from gas retorts is not so good for smelting as from coking ovens and is
usually sold for domestic fuel.
Coke is ordinarily the only product obtained when the beehive oven is used.
There a semi-combustion process takes place resulting in the loss of the whole of the
gas and byproducts together with part of the coke. The oven is simply a fireclay
chamber, beehive shape, provided with a loose brick charging door and an exit flue.
It is heated from the inside by the combustion of gases driven from the coal and of
part of the fixed carbon, a limited supply of air entering through an opening in the
side of the oven. Much of the volatile matter is burned inside the oven, and
the remainder, which passes out through the top burns when it comes in contact with
the air. In actual practice from 15-20 per cent of the coke is burned.
In byproduct ovens, coke is usually the principal product; gas, tar, ammonia,
cyanides, benzol and toluol are recovered as byproducts. The byproduct retort
is a rectangular chamber, measuring from 30-33 ft. in length, from 5 ft. 9 in. to 7 ft.
6 in. in height, from 17 to 22 in. in width. No oxygen comes in contact with the
volatile matter to burn it and destroy the byproducts, and the latter are separated
from one another in a suitable recovery system.
In gas retorts illuminating gas is the principal product, coke and ammonia being
byproducts. The light oil containing benzol and toluol is left in the gas to increase its
candlepower. The charges are of much smaller size than those of coking ovens.
The gases and products are drawn from the oven through cooling and washing
plants for the recovery of tars, ammonia and benzol. Only one-half to two-thirds
of the permanent gas is burned after the above products have been extracted. The
remainder is used for heating or power purposes.
The gases are drawn from the main by the exhauster, through the cooler, either
air or water cooled. In the hydraulic main the heavier tars or pitches are deposited.
From the condensers the bulk of the remaining tar together with the greater proportion
of the ammonia liquor is recovered.
From the condensers the gases pass into a fog extractor which removes the last
traces of tar, then into the ammonia scrubber, where the last of the ammonia not
recovered by the condensers is removed. From here they pass through the benzol
scrubber, which is similar to that from which the ammonia is recovered. There
heavy oil is used for the solution and recovery of the benzol vapors. The gas is then
passed to the ovens, where it is fed into the heating flues.
The tar and ammonia liquors made in the hydraulic main condensers, tar catchers
and ammonia scrubber, are collected and run into tanks. The tar settles to the
bottom, the liquor remaining above the tar is run off to the ammonium-sulphate
plant. There, with the aid of steam and milk of lime, mixed with the liquor, the
whole of the ammonia, together with steam, hydrogen sulphide, carbonic and hydro¬
cyanic acids, pass into a lead lined vessel, called the saturator, containing sulphuric acid.
Here the ammonia is recovered as ammonium sulphate. The hydrogen sulphide and
other gases pass from the saturator to the preheater, thence to a condenser, and finally
into purifiers where the hydrogen sulphide is removed by oxides of iron. The crude
benzol is recovered from its solution of oil by distillation.
In the Koppers byproduct-oven process, the ammonia scrubber is replaced by the

sulphate saturator, thereby forming the sulphate directly in the gas main.
The processes involved in the carbonization of coal may be divided into its various
steps as follows: first, liberation of primary volatile matter at low temperatures with
alteration of the coal into a partly carbonized solid; second, liberation of volatile
products from the altered residue by subjecting it to higher temperatures; and finally,
thermal reactions among the primary volatile products. The tars and gases from the
distillation of the coal are mixtures of all these different substances, most of them
being secondary products.
Experimental evidence indicates that the bulk of the volatile matter of coal under
proper conditions may be liberated at temperatures below 500°C. The rate at which
the volatile matter distills is dependent only on the rate at which heat is supplied.
More than two-thirds of the organic substance of coal is decomposed at temperatures
below 500°C. (932°F.). Different kinds of coal vary in ease of decomposition.
When coal is heated it begins to evolve gaseous and liquid matter at a temperature
which may be as low as 250°C. The rate of evolution gradually increases with the
temperature. Above 550°C. it begins to decline. Experiments have shown that
in the distillation of bituminous coal the reactions that occur in the coal at tempera¬
tures below 750°C. absorb heat, but those at temperatures above 750°C. evolve heat.1
It is possible to control the thermal decomposition of hydrocarbons in such a way
as to give maximum yields of certain products of decomposition, such as constituents
of gases, aromatic hydrocarbons, etc. The factors that control these yields are
temperature, pressure, length of time in the reaction zone, the starting material and
catalysts.
Thermal decomposition of the volatile matter takes place very easily at tem¬
peratures above 750°C. (1,382°F.). It increases the yield of the permanent gases,
hydrogen and hydrocarbons. Distillation at temperatures above 750°C. yields
gases in which hydrogen predominates, whether secondary reactions are prevented
or not; but secondary reactions of the volatile products will increase the total gas yield
at the expense of the tar.
Paraffin hydrocarbon gases of the series CnH2n+2 higher than methane predomi¬
nate at temperatures below 400°C. Coals rich in oxygen (10-11 per cent) and
hydrogen (above 5 per cent) and which, therefore, have large percentages of resinic
bodies, on being carbonized at low temperatures yield tars rich in phenol and cresol.
Recently the nature of the liquid primary products has been determined.2 On
distilling coal in vacuo at low temperature, 430°C., 6.5 per cent the weight of a bitum¬
inous coal was obtained as tar. Of this tar 50 per cent was a pitch residue and 50 per
cent, an oil separated by distillation to 300°C. The oil consisted of 40-45 per cent
ethylene hydrocarbons; 40 per cent naphthene and liquid paraffins; 12-15 per cent
phenols, chiefly cresols and xylenols; 7 per cent aromatic hydrocarbons apparently
'homologues of naphthalene; and solid paraffin and pyridine bases in small quantity.
Benzene and its homologues as well as anthracene, carbon bisulphide and solid aromatics
were absent. Sulphurous gases are readily produced in the decomposition of coal.
They begin to be formed at temperatures as low as 250°C. (482°F.).
The higher the heat of the distillation the larger the volume of the gas produced
and the lower its quality as an illuminating agent, while the condensible tar obtained
decreases in quantity, but increases in specific gravity owing to the presence of a larger
proportion of heavy and high boiling constituents.
When the temperature is only slightly above that which decomposition commences,
the volatile products formed consist chiefly of liquids of somewhat similar constitution
1H. Hollings and J. W. Cobb, Jour. Chem. Soc., Vol. 107, 1915, pp..1106-1115.
2D. T. Jones and R. V. Wheeler, Jour. Chem. Soc., Vol. 105, 1914, pp. 140—151.
(552 CHEMICAL ENGINEERING
to those contained in petroleum oil, the gas formed being small in volume and of high
illuminating power.
When the temperature is raised, the volatile products first produced undergo
further changes in contact with the red-hot coke and surface of the retort. They
are therefore, converted into hydrogen and simpler hydrocarbons as methane, ethy¬
lene, etc., and more complex substances belonging to the aromatic series, such as
benzene, toluene, naphthalene and anthracene. As the temperature is further raised
the quantity of hydrogen formed increases with simultaneous production of more and
more complex hydrocarbons containing smaller percentages of hydrogen.
Low temperature favors the formation of the total light and heavy oil and tar
acids and correspondingly an increase in temperature, causes a falling off in the
percentage of those compounds with an increase in the percentage yields of heavy
oil; a decrease in the percentage yield of paraffins; and an increase in that of the
aromatics.
Diminishing the duration of the carbonization period tends to increase the actual
yields of total, light and heavy oils as well as the yields of the total benzene, toluene,
and xylene fractions. Slow carbonization tends to increase the yield of coke.
The factors to be considered in any coal carbonization process are as follows
(1) Temperature at which carbonizing reactions take place; (2) pressure in the
carbonizing retort; (3) concentration of the various constituents involved in the
carbonizing reactions; (4) duration of the reactions taking place in the retort
during carbonization; (5) size and character of the surfaces in contact with the con¬
stituents undergoing decomposition, or the catalytic effect; (6) nature of the coal.
TAR
Ordinary coal tar is formed from coal at high temperature, chiefly as the
result of the decomposition of a tar previously formed at low temperature. The
mechanism of the process consists essentially in the decomposition of the naph¬
thenes, paraffins and unsaturated hydrocarbons present in the.low temperature
tar to form olefines of various carbon content which condense to form aromatic
compounds.1
There is a variation in the composition of the tar according to the kind of retort
and carbonizing temperature employed. If the coal is carbonized in chamber or
vertical retorts or at low temperature, a lower gravity tar is obtained. High heat
in horizontal retorts produce tar of high specific gravity. On subjecting the tar to
destructive distillation about 190 definite chemical compounds can be separated from
the distillates and the pitch.
Tars produced at low temperatures yield on distillation phenols of the carbolic
acid series and also of a series less acid in behavior; a smaller amount of naphthalene
and of the benzene hydrocarbons, a large percentage of hydrocarbons of the paraffin
and olefine series; a small amount of free carbon, and most of the nitrogen in the form
of aniline and its homologues.
Tars produced at high temperatures yield traces of paraffinoid hydrocarbons on
distillation. The predominating hydrocarbons are those of the benzene, naphthalene
and anthracene series; the nitrogen occurs principally in the form of pyridine bases;
the phenols consist chiefly of carbolic acid and its homologues; the percentage of free
carbon is generally high.
Tar stills are built in various sizes with a working capacity of 10, 15, 20 or 30 tons.
D. T. Jones, Jour. Soc. Chem. Ind.j Vol. 36, 1917, pp. 3-7.
The vertical still is most satisfactory. The amount of tar usually obtained from gas
coal is equivalent to about 5 per cent of the amount of the coal.
The products of high-temperature carbonization from coal of the Pittsburgh type
on an industrial scale at about 800-900°F. (427-482°C.) consist of a rich gas amount¬
ing to 0.6-0.7 cu. ft. per pound of coal, and a large yield of oil tar comprising 10-12 per
cent of the coal. This tar consists chiefly of paraffin hydrocarbons, a very small
amount of benzene and naphthalene derivatives and practically no carbon. The gas
contains 6-7 per cent of unsaturated hydrocarbons and 20-25 per cent of ethane and
its higher homologues. It will, therefore, have a high calorific and illuminating value.
The tar obtained from the distillation of paraffin base oils is in general much like
a coal tar except that it contains no oxygen, nitrogen or sulphur compounds. The
aromatic tars differ from coal tar in that they contain no straight chain compounds.
BIBLIOGRAPHY
G. B. Taylor and H. C. Porter, “The Primary Volatile Products of the Carbonization

of Coal/7 Bureau of Mines Tech. Paper 140.
J. R. Suydam, “A Comparative Study of the Thermal Decomposition of Coal and some
of the Products of its Carbonization,” 1917.
L. F. Hawley, “Hardwood Distillation Industry,” Chem. and Met. Eng., July, 1921,
p. 137; Aug. 3, 1921, p. 195.
A. R. Warnes, “Coal Tar Distillation.”
E. L. Lomax, A. E. Dunstan, F. B. Thole, “Pyrogenesis of Hydrocarbons,” Jour.
Inst. Pet. Tech., Vol. 3, 1916, pp. 36-120.
T. C. Alderson, “The Oil Shale Industry.”
R. F. Bacon and W. A. Hamor, “American Petroleum Industry.”
W. R. Crowell, “Effect of Temperature and Duration of Reaction on the Formation of
Light Oils in Coal Carbonization at Atmospheric Pressure,” 1916.
W. F. Rittman, C. B. Dutton and E. W. Dean, Bureau of Mines, Bull. 114. “Man¬
ufacture of Gasoline and Benzene-Toluene from Petroleum and other Hydrocar¬
bons.”
Gavin Hill and Perdes, “Notes on Oil Shale Industry,” Bureau of Mines, Serial 2256,
June, 1921.
C. Ellis and J. V. Meigs, “Gasoline and Motor Fuels.”
Thorpes, “Applied Chemistry.”
R. W. Ells, “Report on the Oil Shale Industry in Scotland,” Dept, of Mines, Canada,
1909, No. 55, No. 1107.
Rogers, “Industrial Chemistry.”
SECTION XIX
REFRIGERATION
By Van Rensselaer H. Greene1
Theory of Refrigeration.—Refrigeration is the art of producing temperatures

below the normal for useful purposes. The theory of refrigeration is based upon
the flow of heat from a warm body to one that is cooler. Commercially, there are
two methods possible for producing a refrigerating effect: The heat may be
transferred to a cold body which is afterwards discarded, as when rooms, liquids,
etc. are cooled by means of ice or a mixture of ice and salt; or a continuous
mechanical refrigerating effect may be produced by means of a circulating agent
which absorbs the heat, and is afterwards made to change its physical state and
transfer that heat to some ordinary cooling medium such as water or the atmos¬
phere ; following which the heat-absorbing vehicle reverts to its original state again,
and the cycle begins a second time; and so on ad infinitum.
In the field of refrigeration in this country, the unit of heat measurement is the
“British thermal unit,” written “B.t.u.” One B.t.u. is the quantity of heat required
to raise the temperature of 1 lb. of water, at or near the point of maximum density
(39°F.), 1°F. One pound of ice, in melting, absorbs 144 B.t.u., and this is known as
the latent heat of fusion of ice. The quantity of heat which becomes latent, due to
the melting of 1 ton (2,000 lb.) of ice, is therefore 288,000 B.t.u.; and the absorption of
this quantity of heat is known as “One Ton of Refrigeration.”
When two solid bodies, such as ice and common salt (NaCl), mix to form a liquid,
a certain amount of heat becomes latent, and as soon as the mixture takes place, the
temperature of the resulting liquid falls correspondingly. With mixtures of ice and
salt, this latent heat is equal to 144 B.t.u. for each pound of ice present, minus a small
quantity of heat, known as the heat of solution of salt, which varies with the proportion
of salt used. The temperature to be obtained from a salt and ice mixture depends
principally upon the relative proportions of the mixture, but also, to some extent, upon
the rate at which outside heat is supplied, and the size of the ice lumps and salt parti¬
cles. On this account it is more or less impracticable to fix the exact temperature
which will result from a given ratio of salt and ice. In general, however, it may be
stated that with a 10 per cent mixture, by weight, of ice and salt (NaCl), a temperature
of 20°F. may be obtained; with 15 per cent of salt, a temperature of 12°; with a 20 per
cent mixture, 0°F.; and with 25 per cent of salt; 10° below zero. The properties of
other salts, such as calcium chloride, magnesium chloride, etc., differ somewhat from
those of the ordinary sodium compound. When such salts are present in quantity,
therefore, these figures will not hold true.
If a mixture is made of a given amount of ice and salt, both at a temperature of
32°F., and sufficient heat is supplied to melt the ice, absorb the salt, and raise the
temperature of the resulting brine to the original 32°F., the total quantity of heat
absorbed in carrying out this operation will be the algebraic sum of the latent heat of
the ice and the heat of solution of the salt. This heat of solution will vary with the
1 Refrigerating engineer, 35 Warren St., New York City.
655
relative quantity of salt used; and the heat absorbing value of a brine of any given
density, corresponding to the particular ratio of ice and salt entering into it, will be a
constant quantity.
To illustrate: A 5 per cent salt mixture (NaCl) will absorb 140 B.t.u. per pound;
a 10 per cent solution, 133 B.t.u.; a 16 per cent solution, 125 B.t.u.; a 20 per cent
mixture, 120 B.t.u.; and a 26 per cent solution, 110 B.t.u.
When the temperature of the salt used is much above 32°F., a slight difference in
these figures will result, ranging from zero to 5 or 6 per cent as the temperature of the
salt is higher.
The use of salt with ice in a refrigerator does not increase the cooling effect to be
obtained from a given amount of ice, but it does cause the ice to melt more rapidly, and
so absorb heat more quickly. In other words, when salt is used, the melting point
of the ice is lowered. But whether salt is added or not, 1 lb. of ice, in melting, only
absorbs 144 B.t.u. and no more. With the addition of salt, the heat absorption is
quickened; hence a lower temperature for a shorter period. There is no gain in
efficiency in employing such mixtures. In fact there is a slight loss, due to the heat of
solution of the salt.
Mechanical Refrigeration.—For the purposes of mechanical refrigeration, in

order to produce a continuous cycle, it is necessary to provide a circulating cooling
medium of such a nature, that with the aid of mechanical force, without excessive
pressure, it can be made so to alter its physical state, and in so doing, to increase
its temperature, that the greater part of its heat content may be discharged to
the atmosphere or to ordinary cooling water.
The usual mechanical refrigerating cycle consists of two stages, as follows: (1) A
vapor at low tension is mechanically compressed to a point sufficiently high to permit of
its liquefaction at ordinary temperatures, and the sensible heat of the hot compressed
gas is removed by flowing water in a surface condenser. This causes the gas to liquefy,
and it then flows to a container adapted to store it under pressure. (2) The liquefied
vapor—still a liquid—is brought into contact, under reduced pressure, with the body
to be cooled. The lowered pressure and proximity of a warmer body at once cause the
liquid to vaporize rapidly, and this is necessarily accompanied by the absorption of
heat, which is taken from the most available source, the body which it is desired to
refrigerate.
The heat absorbed, as previously explained, is called the “latent heat” of the body
which absorbs it. It is defined as follows. The latent heat of a substance is the num¬
ber of B.t.u. required to change the state of 1 lb. of the substance without changing
its temperature, as for instance, the latent heat of vaporization of a liquid; the latent
heat of melting of ice.
The circulating refrigerating medium above spoken of is known as the “refriger¬
ant.” The theoretical requirements of a good refrigerant are: (1) A low boiling point
at ordinary pressures; (2) a high latent heat value; (3) a small specific volume.
A low boiling point is desirable because it makes operation possible with low pres¬
sures throughout the major part of the cycle. The latent heat of vaporization is, to a
large extent, a practical measure of the cooling effect; so it is obvious that the greater
the latent heat, the better the refrigerant. The volume of vapor per pound of the
refrigerating medium determines the size of the machine parts.
The mediums most exclusively used in the refrigerating industry are ammonia,
carbon dioxide, sulphur dioxide, ethyl chloride, and methyl chloride. Accompanying
will be found a table giving the absolute pressures of these mediums at 0°F. and 95°F.,
and also their latent heats of vaporization.
REFRIGERATION 657
Absolute pres¬ Absolute pres¬

Latent heat of
sure at 0°F., sure at 95 °F.,
evaporation at
pounds per pounds per
0°F.
square inch square inch
Ammonia. 29.95 197.30 572.2

Carbon dioxide. 314.00 1211.00 124.4
Sulphur dioxide. 9.93 77.64 172.0
Ethyl chloride. 4.04 31.80 190.4
Methyl chloride. 18.73 110.10 176.0
Since ammonia has a much greater heat of vaporization, and reasonably low work¬
ing pressures, this agent is the one most generally used in the refrigerating industry.
Further than this, ammonia is soluble in water, which makes possible the ammonia
absorption process, as well as the ammonia compression process.
Accompanying (jpp. 658 and 659) will be found a table giving the properties of
ammonia at various temperatures and pressures.
Critical Data for Gases.—For each gas there is a temperature, above which
it cannot be liquefied, no matter how much pressure is employed. This point
is called the “critical temperature77 for that gas. At any point below its critical
temperature, any gas or vapor may be liquefied, providing it is subjected to
sufficient pressure.
A gas having been cooled to a point just below its critical temperature, the pres¬
sure necessary to produce liquefaction is called its “ critical pressure.,} As, however,
under these conditions, the exact point of transition from one state to the other is
most difficult to recognize, such determinations must always be regarded as merely
approximate.
The “critical volume77 of a gas is its volume at the critical temperature and
pressure, as compared with its volume a 32°F., under a pressure of one atmos¬
phere taken as a unit (normal pressure and temperature, N.P.T.).
Below will be found a table giving the accepted critical temperatures, pressures,
and volumes of the gases employed as refrigerating agents.
Critical tempera¬
Critical pressure, ! Critical volume, .
ture, degrees
atmospheres calculated
Fahrenheit
Ammonia. 115 266. 0.00482

Carbon dioxide. 77 88. 0.0066
Sulphur dioxide. 79 312.
Ethyl chloride. 54 360.5
Methyl chloride. 73 286.9
Rating of Refrigerating Machines.—A refrigerating machine is rated accord¬

ing to the number of “tons77 of refrigerating effect it is capable of producing
in 24 hr. at a given pressure range, and with a given temperature of condensing
42
Properties op Anhydrous Ammonia
Pressure pounds per Saturated vapor Liquid

Temperature, degrees square inch
Specific Density Density Latent

volume pounds pounds Heat of
Fahrenheit Absolute Absolute Gage per cubic per cubic evapo¬
cubic feet
per pound foot foot ration
-10 450 23.30 8.60 11.63 0.0860 41.46 579.9

- 9 451 23.90 9.20 11.35 0.0881 579.1
- 8 452 24.52 9.82 11.08 0.0903 578.4
- 7 453 25.15 10.45 10.82 0.0924 577.6
- 6 454 25.80 I 11.10 10.57 0.0946 576.8
- 5 455 26.46 1 11.76 10.32 0.0969 41.25 576.1
- 4 456 27.13 12.43 10.08 0.0992 575.3
- 3 457 27.82 13.12 9.85 0.1015 574.6
- 2 458 28.52 13.82 9.62 0.1039 573.8
- 1 459 29.23 14.53 9.40 0.1064 573.0
0 460 29.95 15.25 9.19 0.1089 41.04 572.2
1 461 30.69 15.99 8.98 0.1114 571.4
2 462 31.44 16.74 8.78 0.1139 570.7
3 463 32.21 17.51 8.58 0.1165 569.9
4 464 32.99 18.29 8.39 0.1192 569.1
5 465 33.79 19.09 8.20 0.1219 40.83 568.3
6 466 34.60 19.90 8.02 0.1247 1 567.5
7 467 35.43 20.73 7.84 0.1275 566.7
8 468 36. 28 21.58 7.67 0.1304 565.9
9 469 37.14 22.44 7.50 0.1333 565.2
10 470 38.02 23.32 7.34 0.1363 | 40.61 564.4
11 471 38.93 24.23 7.18 0.1393 563.6
12 472 39.84 25.14 7.02 0.1424 562.8
13 473 40.77 26.07 6.87 0.1455 562.0
14 474 41.71 27.01 6.72 0.1487 561.2
15 475 42.67 27.97 6.583 0.1519 40.39 560.4
16 476 43.65 28.95 6.444 0.1552 559.6
17 477 44.65 29.95 6.308 0.1585 558.8
18 478 45.67 30.97 6.176 0.1619 558.0
19 479 46.70 32.00 6.047 0.1654 557.1
20 480 47.75 33.05 5.920 0.1689 40.17 556.3
21 481 48.82 34.12 5.796 0.1725 . 555.5
22 482 49.91 35.21 5.676 0.1762 554.7
23 483 51.02 36.32 5.560 0.1799 553.9
24 484 52.15 37.45 5.447 0.1836 553.1
25 485 53.30 38.60 5.336 0.1874 39.95 552.2
26 486 54.47 39.77 5.228 0.1913 551.4
27 487 55.66 40.96 5.122 0.1953 550.6
28 488 56.87 42.17 5.019 0.1993 549.8
29 489 58.10 43.40 4.918 0.2034 549.0
30 490 59.35 44.65 4.820 0.2075 39.72 548.1
31 491 60.62 45.92 4.724 0.2117 547.3
32 492 61.91 47.20 4.631 0.2159 546.5
33 493 63.22 48.92 4.540 0.2203 545.6
34 494 64.55 49.85 4.451 0.2247 544.8
35 495 65.91 51.21 4.364 0.2292 39.50 543.9
36 496 67.29 52.59 4.279 0.2337 543.1
37 497 68.69 53.99 4.196 0.2384 542.2
38 498 70.11 55.41 4.115 0.2431 541.4
39 499 71.56 56.86 4.036 0.2478 540.5
40 500 73.03 58.33 3.959 0.2526 39.27 539.7
41 501 74.53 59.83 3.884 0.2575 538.8
42 502 76.05 61.35 3.810 0.2625 537.9
43 503 77.59 62.89 3.738 0.2675 537.1
44 504 79.16 64.46 3.668 0.2727 536.2
45 505 80.75 66.05 3.599 0.2779 39.04 535.3
46 506 82.37 67.67 3.532 0.2832 534.5
47 507 84.01 69.31 3.466 0.2885 533.6
48 508 85.68 70.98 3.402 0.2940 532.7
49 509 87.37 72.67 3.339 0.2995 531.8
50 510 89.09 74.39 3.278 0.3051 38.81 531.0
water. The usual rating is based on a suction pressure of 15.65 lb. above atmos¬
phere, a discharge pressure of 185 lb. above atmosphere, and sufficient con¬
densing water at 60°F.
Refrigerating machines are sometimes listed in terms of their ice-making capacity.
This rating, however, should be used with great caution, as it is apt to vary widely—
sometimes as much as 15 per cent—depending upon the temperature of the water to be
frozen, and other variables.
REFRIGERATION 659
Temperature, degrees Pressure pounds per

square inch Saturated vapor 1 Liquid
Specific Latent
volume Density Density
Fahrenheit Absolute Absolute Gage heat of
cubic feet pounds per pounds per evapora¬
per pound cubic foot cubic foot tion
51 511 90.83 76.13 3.219 0.3107 530.1

52 512 92.59 77.89 3.161 0.3164 529.2
53 513 94.38 79.68 3.104 0.3222 528.3
54 514 96.19 81.49 3.048 0.3281 527.4
55 515 98.03 83.33 2.992 0.3342 38.57 526.5
56 516 99.90 85.20 2.938 0.3404 525.6
57 517 101.8 87.10 2.885 0.3467 524.7
58 518 103.7 89.00 2.833 0.3530 523.8
59 519 105.7 91.00 2.783 0.3594 522.9
60 520 107.7 93.00 2.734 0.3658 38.33 522.0
61 521 109.7 95.00 2.686 0.3723 521.1
62 522 111.7 97.00 2.639 0.3790 520.2
63 523 113.8 99.10 2.592 0.3858 519.3
64 524 115.9 101.20 2.547 0.3927 518.4
65 525 118.1 103.40 2.503 0.3996 38.09 517.5
66 526 120.3 105.60 2.460 0.4066 516.5
67 527 122.5 107.80 2.418 0.4136 515.6
68 528 124.7 110.00 2.377 0.4207 514.7
69 529 126.9 112.20 2.336 0.4280 513.7
70 530 129.2 114.50 2.296 0.4354 37.85 512.8
71 531 131.5 116.80 2.257 0.4430 511.9
72 532 133.9 119.20 2.219 0.4506 511.0
73 533 136.3 121.60 2.182 0.4583 510.0
74 534 138.7 124.00 2.145 0.4662 509.1
75 535 141.1 126.40 2.109 0.4742 37.60 508.1
76 536 143.6 128.90 2.074 0.4823 507.2
77 537 146.1 131.40 2.039 0.4905 506.2
78 538 148.7 134.00 2.005 0.4988 505.3
79 539 151.3 136.60 1.972 0.5071 504.3
80 540 153.9 139.20 1.940 0.5155 37! 35 503.4
81 541 156.5 141.80 1.908 0.5241 502.4
82 542 159.2 144.50 1.877 0.5328 501.4
83 543 161.9 147.20 1.847 0.5416 500.5
84 544 164.6 149.90 1.817 0.5504 499.5
85 545 167.4 152.70 1.788 0.5594 37! io 498.5
86 546 170.2 155.50 1.759 0.5685 497.5
87 547 173.0 158.30 1.731 0.5777 496.5
88 548 175.9 161.20 1.704 0.5869 495.5
89 549 178.8 164.10 1.677 0.5964 494.5
90 550 181.8 167.10 1.650 0.6060 36 '.84 493.5
91 551 184.8 170.10 1.624 • 0.6158 492.5
92 552 187.8 173.10 1.598 0.6258 491.5
93 553 190.9 176.20 1.573 0.6358 490.5
94 554 194.1 179.40 1.548 0.6460 489.5
95 555 197.3 182.60 1.524 0.656 36! 58 488.5
96 556 200.5 185.80 1.500 0.667 487.5
97 557 203.8 189.10 1.477 0.677 486.5
98 558 207.1 192.40 1.454 0.688 485.4
99 559 210.4 195.70 1.431 0.699 484.4
100 560 213.8 199.10 1.408 0.710 36! 32 483.4
Measurement of Refrigerating Effect.—The commercial unit of refrigeration

is the “ton of refrigeration,” before spoken of. This term is used to designate
the number of B.t.u. required to melt (not manufacture) one ton of 2,000 lb.
of pure solid ice. Since the latent heat of ice is 144 B.t.u., 1 ton of refriger¬
ation is equal to 144 B.t.u. multiplied by 2,000, or 288,000 B.t.u. All refriger¬
ating machinery calculations are based on a 24-hr. day, and refrigerating
capacities are figured at the rate of 12,000 B.t.u. per hour, or 200 B.t.u. per
minute.
Refrigerating Systems.—There are two commercial systems of refrigeration
in use. They are known as the “compression system” and the “absorption
system.” The compression system is suitable for use with any of the refrigera¬
ting mediums listed in this section, and doubtless could be used with others not
yet discovered. The absorption system, on the other hand, is distinctly an
ammonia system. Due to its greater flexibility, the majority of the refrigera-
ting plants in use operate by the compression system. Both of these systems
require water to extract the heat from the refrigerating medium.
Compression System.—The ammonia-compression system of refrigeration
depends upon the mechanical compression of ammonia vapor, so that the result¬
ing gas may be liquefied, and its heat content removed, after which it is vapor-
Fig. 1.—Atmospheric condenser.
ized again in contact with the objects to be cooled. The principal pieces of
apparatus required are as follows: The compressor; the condenser; the liquid
receiver; and the expansion coils. These, with the addition of the various pipe
connections and lesser appurtenances, make up the entire plant.
The compressor is a machine much like an air compressor, and is driven in the same
manner. With regard to speed, compressors are of two general types: low-speed
machines, and high-speed machines. The former run at about 75 r.p.m., and the
latter at about 225 r.p.m. The function of the compressor is to compress the ammonia
gas. . ,
The condenser is a coil of pipes for condensing the compressed ammonia gas, the gas
being usually passed through the inside of the pipes, while water is applied externally.
REFRIGERATION 661
Condensers are of two principal types: The “atmospheric” condenser, and the
“double pipe” condenser. The purpose of the condenser is to remove the heat from
the compressed vapor, and so cause it to change its state from a gas to a liquid.
Figure 1 shows the “atmospheric" type of condenser. It is usually constructed
of 2-in. wrought-iron pipe, 12 pipes high, and 20 ft. long. This type of condenser is
best located in the open air, on the roof of a building, or in some other exposed place.
When in operation, the compressed ammonia passes through the pipes, while the cool¬
ing water trickles down over the outside. There is no economy in making these
condensers more than 12 pipes high.
The double-pipe condenser, shown in Fig. 2, consists of two concentric pipe coils
put together with suitable composite end fittings. The ammonia gas is passed
through the outer annular space, while the cooling water circulates through the
central pipe.
The liquid receiver, Fig. 3, is a cylindrical drum, which is used as a reservoir
for the storage of the liquid refrigerating medium, in this case, ammonia. The
Condenser
Fig. 3.—Compression side ammonia refrigeration system.
expansion coils are pipe coils, which are located in the rooms or other places where a
cooling effect is desired. The refrigerated medium is circulated through these coils,
which are usually constructed of 1J4“ or 2-in. pipe.
The cycle of operation of an ammonia compression system, such as that shown in
Fig. 3, is as follows: The ammonia gas enters the compressor, or ice machine, as it is
called, at a low pressure, say 15 lb., and is there compressed to, say 185 lb. As the gas
is compressed, its temperature also rises. It is then passed through an oil separator,
which removes the oil absorbed while going through the compressor, after which it
flows to the condenser. Here, the cooling water, which is sprayed over the outside of
the condenser pipes, removes enough heat from the compressed vapor to cause it to
condense to a liquid. The anhydrous liquid ammonia thus formed then flows to the
liquid receiver. This much of the system is known as the high-pressure side.
The liquid ammonia is then released, through an expansion valve, into the expan¬
sion coils; ancLas it passes through this valve, the pressure upon it automatically drops
from 185 lb.—the condenser pressure—to 15 lb.—the suction pressure maintained by
the ice machine, with which the expansion coils are in direct connection. This drop
in pressure carries with it a drop in temperature from 95 to 0°F., and the cold liquid in
the expansion coils, being wholly unconfined, boils away to a gas, absorbing heat from
everything in its neighborhood. The gas is then returned to the compressor, com¬
pleting the cycle of operation.
In the practical operation of a system of this kind, the discharge pressure is neces¬
sarily governed by the temperature and amount of the available condensing water, while
the suction pressure on the machine must be maintained by the operator at the point
required to produce the desired temperature in the rooms. Assuming a 15° difference
in temperature between the cooling medium and the air in a room, if it is desired to
keep this room at 15°F., the temperature of the ammonia must be 0°F. By referring
to a table of the properties of saturated ammonia, it will be found that 0°F. corresponds
to a pressure of 15.25 lb. The machine, therefore, should be operated at this back
pressure. When the expansion valve is once properly set to pass the required quantity
of ammonia to the expansion coils, the entire system may be controlled by operating
the compressor so as to maintain this suction pressure.
The chief advantages of the system are two in number: (1) Steam is not necessary.
An ammonia compressor and all needed auxiliaries may be operated by gas engine, oil
engine, or electric motor, none of which require a boiler. (2) All the apparatus is
extremely efficient and economical to operate. Therefore, low production costs are
always possible, wherever power is to be had at reasonable cost.
The principal disadvantage of the system lies in the fact that temperatures below
about 0°F. cannot usually be maintained with a single-stage compressor without an
excessive expenditure of power. On this account, in the past, recourse has been had
to the absorption machine, whenever excessively low temperatures were needed. At
the present time, however, high-class two-stage compressors are available, which are
very efficient and economical. A machine of this type, recently installed by the writer
in Philadelphia, keeps a tier of hardening rooms at the factory of the Colonial Ice Cream
Co. at — 16°F., day and night without difficulty.
Two-stage Compression.—In refrigerating plants, the value of two-stage

compression lies in the possibility which it presents of carrying very low back
pressures. This most attractive feature would have led, no doubt, to the general
adoption of two-stage compression long ago, but for the difficulties attending
its use, due to the want of a suitable method of inter-cooling between stages. In
the ammonia machine of this type, operating at a very low back pressure, the
compressed ammonia coming from the first stage is still quite cool, and the tem¬
perature of the gas in the inter-cooler cannot be successfully controlled by ordi¬
nary water, usually warmer than the gas itself. It therefore becomes necessary
where cool water is not available to find some other means of heat regulation,
and the method which immediately presents itself is by the injection of liquid
ammonia into the partially compressed gas.
This is the method of inter-cooling used in the modern two-stage compressor.

Special precautions, however, must be adopted to obviate the great danger which
attends this practice, due to the fact that the liquid ammonia is apt to carry over—
still a liquid—into the second-stage cylinder, and blow out the head of the machine.
Two distinct methods are in use for avoiding this danger. In one—the most
successful—a supplementary container, filled with helical pipe coils nested together,
is placed just below the inter-cooler; and the surplus ammonia injected into the inter¬
cooler, in excess of that needed for the inter-cooling of the partially compressed gas,
flows into it by gravity. Through these helical pipe coils, the entire anhydrous
ammonia supply of the plant is made to pass, its temperature being comparatively
high—about that of the ammonia condenser. So the surplus ammonia from the
REFRIGERATION 663
inter-cooler, being brought in contact with this great mass of warm liquid, expends
itself in cooling it, at the same time boiling away to a gas. In this way it is made to do
useful work in the refrigerating cycle, and the danger above described is entirely
eliminated.
The second method used is to allow the surplus ammonia to drain out of the inter¬
cooler to the suction of an ammonia pump, which discharges this liquid directly back
into the high-pressure feed lines. This, however, is not so reliable as the method first
described, as there is a possibility of the ammonia pump becoming gas bound, and
more or less difficulty in keeping the pump suitably packed to handle the cold liquid.
Tonnage of Ammonia Compression Machines.—As refrigerating effect is

produced by the vaporization of a liquid into a gas, the capacity of a refrigera¬
ting machine is measured by its ability to handle a given number of cubic feet,
per minute, of the gaseous refrigerant. From the ice-machine-builder’s stand¬
ard, approximately 5 cu. ft. of ammonia gas per minute, compressed by a given
machine under standard conditions, is the practical equivalent of one ton of
refrigerating effect per day of 24 hr.
Under normal conditions, the water used to condense the compressed ammonia is
obtained from the city mains; and in the summer months, its temperature is in the
neighborhood of 75°F. With water at 75°F. showered over the ammonia condensers,
the head pressure on the machine, set by the temperature of liquefaction of the gas, will
be somewhere between 155 and 185 lb., depending upon the design and efficiency of
the condenser employed.
Manufacturers of refrigerating machines have adopted as their standard for rating
purposes a condenser or head pressure of 185 lb., and a suction pressure of 15 lb.;
so that the compression of 5 cu. ft. of gas per minute, from a suction pressure of 15 lb. ;
to a condenser pressure of 185 lb., represents an ice-machine capacity of 1 ton of
refrigeration per 24 hr.
Horsepower of Ammonia Compressors.—Under these conditions, the com¬

pression of 5 cu. ft. of ammonia gas per minute will require, theoretically,
horsepower, expended in the cylinder of the compressor, exclusive of the friction
losses of the compressor itself, the friction losses of the steam engine which
drives it (if any), and the belt friction, if a belt is used. These various losses, in
the aggregate, will amount to about 15 per cent, giving a total expenditure of
power, per ton of refrigeration, of 1% hp.
Ice-making Capacity of Compression Machines.—When ammonia-compres¬
sion machines are used for making ice, they are called upon not only to remove the
latent heat of freezing from the water, but also to cool the water itself from its
initial temperature down to 32°F., and the ice, after it has been frozen, from 32°F.
down to the temperature of the freezing tank; and besides all this, to replace all
heat leakage from the various piping, and from the brine tank itself, with its
great charge of ice still being frozen in the cans. These various losses, taken
together, amount to fully 40 per cent of the latent heat of freezing, so that an
ammonia compressor must compress at least 7 cu. ft. of gas per minute for every
ton of ice to be manufactured by it per day of 24 hr. The total expenditure of
of power, per ton of ice manufactured per 24 hr., will be 2.45 hp. for 24 hr.,
or 58.8 hp.-hr.
Absorption System.—The ammonia-absorption system is based upon the
great solubility of ammonia gas in water, which, at ordinary temperatures,
is about 700 volumes of the gas to one volume of water. The apparatus required
consists of the generator, the rectifier, the exchanger, the condenser, the liquid
receiver, the expansion coils, and the absorber; together with the necessary-
pump, some minor apparatus, and the pipe connections. Ammonia is the only
refrigerant adapted for commercial use with an absorption system. No other
cheap medium is sufficiently soluble.
The expansion coils, condenser, and liquid receiver used in this system are identical
with the apparatus similarly named for use in compression plants. They also perform
the same functions, and have the same relation to each other. In the compression
system, the compressor is the main energizing unit or heart of the plant. In the same
way, the absorber is the heart of the absorption system.
The ammonia absorber is a cylindrical drum, much like a horizontal tubular boiler,
with interior tubes parallel to its longitudinal axis, terminating in tube sheets located
near each end. Cooling water is passed through the tubes to cool the mixture of water
Fig. 4.—Ammonia absorption plant.1
and ammonia in the shell surrounding them. A very weak mixture of water and
ammonia, termed weak liquor, continually flows into the absorber from the exchanger
—a piece of apparatus to be described later on. Spent ammonia gas from the expan¬
sion coils is led into the bottom of the absorber through a perforated pipe, and caused
to bubble up through the liquor contained in it. This weak liquor rapidly absorbs it,
and thus enriched by the addition of ammonia gas, now becomes what is called strong
liquor. A stream of this strong liquor is continually pumped out of the absorber by
a horizontal steam-driven piston pump called the ammonia pump, which is the only
moving piece of apparatus required.
From the pump, the strong liquor flows through a set of double pipe coils, spoken of
above, called the exchanger. Here, the cold strong liquor in one pipe, which has just
left the absorber, picks up some of the heat from the hot weak liquor in the other pipe,
coming back from the generator to the absorber. In this way the strong liquor is
warmed to some extent, and then passes on to the generator.
The generator is another cylindrical drum, which has a set of steam coils suspended
in it. Steam, at a pressure of from 10 to 20 lb., furnished from the exhaust of the
ammonia pump or from other sources, is made to pass through these coils, to heat
the strong liquor with which the shell of the generator is filled. The greater part of
the ammonia in this strong liquor, and some of its water, is in this way vaporized,
forming a hot wet gas, which flows away from the generator, leaving a hot weak liquor
behind.
1 Courtesy Carbondjile Machine Co., Carbondale, Pa,
REFRIGERATION 665
The ammonia gas driven out of the liquor produces a head pressure in the interior
of the generator; and, impelled by this, a continual stream of the hot weak liquor
flows away, passing first through the exchanger, as before mentioned, and then
through a weak-liquor cooler, which is another set of double pipe coils, cooled by
running water. From this point, the weak liquor flows back to the absorber, as de¬
scribed in the first paragraph, completing that cycle of operation.
The hot wet ammonia gas, escaping from the generator, is first led to the rectifier,
a series of pipe coils arranged with drip connections on several of the pipes. Cold
water is passed over the coils of the rectifier, to condense the water vapor in the
ammonia gas, and the condensed water, thrown down in this way, is drawn off through
the various drips, and led back to the generator. The warm ammonia gas, now freed
from moisture, passes on to the condensers, and from there to the liquid receiver, where
it is stored in the form of anhydrous liquid ammonia, ready for use in the expansion
system as it is needed. Later on, the spent expanded gas coming from the refrigerat¬
ing coils, is led back to the absorber, completing that cycle of operation.
Figure 4 is a diagram showing the various pieces of apparatus, and the two
cycles of operation, in the ammonia absorption system. The control of the entire
process is in the speed of the ammonia pump, which must be regulated by the operator
to take care of the demands made upon the plant from time to time.
The advantages of this system are three in number: Simplicity of control; the fact
that exhaust steam can be utilized; and the fact that very low temperatures are
possible. Little attention, as a rule, is required on the part of the operator; and the
possibility of using exhaust steam is often an important factor. With plants of this
character, room temperatures as low as 20° below zero can be obtained.
The principal disadvantages of plants of this kind are: (1) That steam is an absolute
necessity; (2) that the thermal efficiency of the process is low; (3) that the system is
rather inflexible, requiring to be operated at a fairly constant rate.
Ammonia Not the Only Practical Refrigerant.—While ammonia, on account

of its high heat of vaporization and low working pressure, is the refrigerant
generally employed in a commercial way, a number of other agents are in use to
a greater or less extent. The mode of operation with these other liquids is
practically similar to that of the ammonia-compression system; and the machines
and other appliances are, in a general way, the same. In this class will be found
compression plants utilizing carbon dioxide, sulphur dioxide, Pictet liquid, ethyl
ether, ethyl chloride, and methyl chloride.
Refrigeration may also be accomplished by means of the compressed-air refrigerat¬
ing machine; vacuum machines, both simple and compound, and by the evaporation of
liquefied air and other gases.
The Carbon-dioxide Machine.—Refrigerating machines employing carbon
dioxide as a refrigerant have been used more or less in the past for particular
cases, where it was feared that leaks might occur, and the contents of the storage
rooms damaged, if ammonia machines were installed. Among the advantages
claimed for this process are the fact that the agent cannot be decomposed by
compression, it is absolutely non-corrosive, and but slightly poisonous. On the
other hand, it must be remembered that carbon dioxide is inodorous, and leaks
might exist for long periods without being discovered.
Below will be found a table giving the properties of carbon dioxide. In making
calculations, however, it should be remembered that the gas is in a superheated
condition during an appreciable part of the refrigerating process, and due allowance
made for the variations which will occur on this account.
Properties of Carbon Dioxide
Absolute
Boiling Total Heat of Latent Heat Weight
pressure,
point, heat of liquid heat of equivalent of 1 cu. ft.
pounds
degrees gas above above evapora¬ of external vapor,
per square
Fahrenheit 32 °F. 32 °F. tion work pounds
inch
—22 210 98.34 -37.80 136.15 16.20 2.321

-13 249 99.14 -32.50 131.64 16.04 2.759
- 4 292 i 99.87 -26.90 126.78 15.80 3.265
5 342 100.57 -20.92 121.50 15.50 3.852

14 396 101.20 -14.50 115.70 15.07 4.535
23 457 101.80 - 7.56 109.38 14.58 5.330
32 525 102.35 0.00 102.35 13.94 6.265
41 599 102.85 8.32 94.52 13.14 7.374
50 680 103.24 17.60 85.64 12.15 8.708

59 767 | 103.58 28.22 75.37 10.91 10.355
68 864 103.83 40.86 1 62.99 9.29 12.480
77 968 103.95 57.05 46.88 7.06 15.474
86 1080 103.72 84.45 19.28 2.96 21.520
Construction of the Carbon-dioxide Plant.—Plants of this type are built much

the same as the ammonia-compression plant, consisting mainly of a compres¬
sor, a condenser, and an evaporator or refrigerator, where the compressed and
liquefied refrigerant is evaporated to absorb heat. The compressor may be either
horizontal or vertical, but must be built extremely strong. It is generally pro¬
vided with an external jacket through which the suction gas passes, this arrange¬
ment giving additional strength, while, at the same time, it tends to keep the
compressor cool. These compressors generally operate at a piston speed of
about 2 ft. per second; and the suction valve has an area somewhat less than
half the area of. the piston, while the discharge valve is about one-fourth of that.
The piston rod is usually packed with leather cups, glycerin being continually
forced into the spaces between the leathers, so that any tendency to leakage
simply results in some of the glycerin being drawn into the cylinder or forced
out into the air, and no escape of gas is ever possible. A belt-driven pump is
generally used to keep up the glycerin pressure.
Any surplus glycerin which passes into the machine, beyond that needed to fill up
the clearances, is forced out with the gas through the discharge valves; and a trap is
provided to separate it from the hot gas on its way to the condenser, and return it to
the glycerin pump, so that it may be used again and again.
The condenser is generally constructed of extra-heavy wrought-iron pipe, much like
an ammonia condenser, but stronger. It may be submerged in a cooling tank, or
arranged for the showering of water over it, as is done with the atmospheric ammonia
condenser. In the best practice, the coils are welded together; and in some cases, it
has been found beneficial to have them constructed of copper; which is possible
with this gas, on account of its non-corrosive character.
REFRIGERATION 667
The evaporating system (or expansion coils) used with carbon-dioxide plants, is
usually constructed of extra-heavy wrought-iron pipe with welded joints. Either
direct expansion or the brine system may be used, but, for most purposes, the former
is preferred. All joints should be made up with the highest class of brass-bushed
flange unions, with gaskets of vulcanized fibre or other material, which, while suffi¬
ciently elastic to insure tightness, will not deteriorate with the constant heat.
The working pressures vary between 50 and 75 atmospheres, but as these com¬
pressors—even in the largest sizes—have very small diameters, no difficulty is experi-
. enced in providing an ample factor of safety in all parts. In designing such machines, a
factor of safety of at least three times the highest working pressure which will come
upon it should be provided for each piece.
Another advantage of this mode of refrigeration besides those given lies in the
cheapness of the agent used. Cheap liquefied carbon dioxide is now manufactured as
a by-product of the brewing industry in Europe. The gas has no effect on any of the
metals, and is inert to nearly all substances.
Among the disadvantages may be mentioned the fact that the gas is inodorous,
and negatively poisonous; for, being heavier than air, it will settle in confined spaces,
and produce death by suffocation, due to the exclusion of oxygen. Further than that,
the working pressures are extremely high, and the system is rather inefficient, due to
the high specific heat of the liquid.
The Sulphur Dioxide Refrigerating Plant.—Machines of this type are also in use
to some extent. For a given amount of refrigeration, however, they must be practi¬
cally three times as large as an ammonia compressor to do the same work. For low
refrigerating temperatures, their efficiency, also, is low. The compressor is built
much like an ammonia compressor, except that there is no need of lubricating devices.
Properties of Sulphur Dioxide
Absolute
Boiling Total heat Heat of Heat Weight of
pressure, Latent heat
point, of gas liquid equivalent 1 cu. ft.
pounds of evapo¬
degrees above above of external vapor,
per square ration
Fahrenheit 32°F. 32°F. work pounds
inch
-22 5.56 157.43 -19.56 176.99 13.59 0.076

-13 7.23 158.64 -16.30 174.95 13.83 0.097
- 4 9.27 159.84 -13.05 172.89 14.05 0.123
|
5 11.76 161.03 - 9.79 170.82 14.26 0.153
14 14.74 162.20 - 6.53 168.73 14.46 0.190
23 18.31 163.36 - 3.27 166.63 14.66 0.232
22.53 164.51 0.00 164.51 14.84 0.282
32
41 27.48 165.65 3.27 162.38 15.01 0.340
50 33.25 166.78 6.55 160.23 15.17 0.407
59 39.93 167190 9.83 158.07 15.32 0.483
68 47.61 168.99 13.11 155.89 15.46 0.570
77 56.39 170.09 16.39 153.70 15.59 0.669

86 66.36 171.17 19.69 151.49 15.71 0.708
95 77.64 172.24 22.98 149.26 15.82 0.906
104 90.31 173.30 26.28 147.02 15.91 1.046
The stroke is usually at least twice the diameter of the piston, which is arranged to
move with a velocity of from 4 to 5 ft. per second, while the compressor valves have
an area of about one-twelfth of that of the piston. In designing machines of this type,
great care must be taken to insure tight joints, as any leakage of the gas is liable to
result in the production of sulphuric acid, very destructive to nearly all metals.
The principal advantage in the use of this refrigerant is, that, in its liquid state,
it is an extremely good lubricant, for which reason no cylinder lubricators, oil separa¬
tors, or rectifiers are required. The condensing pressure, also, is quite low. The
disadvantages are the necessarily large compressor, a lower efficiency, and the extreme
precautions needed to prevent leakage.
Pictet Liquid.—This compound, which has also been used to some extent in
compression refrigerating machines, is either a mixture or a chemical combina¬
tion of carbon dioxide and sulphur dioxide, or both. Pictet, who is responsible
for its introduction, claims that it is a chemical compound, corresponding to the
formula CO4S. As its behavior and temperature upon being compressed are
quite different from either of the gases entering into its formation, there is some
reason for believing that this is true. No exact and reliable information, however,
is available with regard to the characteristics of this gas, and reports from experi¬
ments made by different investigators are, to say the least, rather conflicting.
Before any definite decision can be reached as to the efficiencies to be expected
from any of these compound refrigerants, further investigation should be made of
their physical and chemical characteristics, and particularly, of their thermo¬
chemical behavior.
Pressure and Temperature of Pictet Liquid
Temperature, Absolute pressure Temperature, Absolute pressure

degrees Fahrenheit in atmospheres degrees Fahrenheit in atmospheres
-22 0.77 50 2.55

-13 0.89 59 2.89
- 4 0.98 68 3.04
- 2.2 1.00 77 3.29
5 1.18 86 4.54
14 1.34 95 5.50
23 1.60 104 5.27
32 1.83 113 6.03
41 2.20 122 6.68
Sulphuric-ether Machines.—In former years, many compression machines

were built with sulphuric ether (C2H5)20, di-ethyl ether, as the refrigerant.
The great size of the compressor required, however, and the high inflammability
of the ether itself, have caused this form of plant to be practically wholly
abandoned. The size of a compressor to handle sulphuric ether is necessarily
about 17 times as large as an ammonia machine of the same capacity, and the two
agents are governed by the same formula, with different values for the constants,
of course. One exception to this exists, however. Ether vapor superheats
REFRIGERATION 669
during expansion, and cools during compression; just the opposite of the behavior
of ammonia, steam, carbon dioxide, and sulphur dioxide. Ether compressors,
therefore, are never troubled by superheat, and always handle dry vapor.
Properties op Saturated Ether Vapor
Heat Heat
Temper¬
Pressure Heat of equiva¬ equiva¬
ature, Heat of Total Weight
in pounds •vaporiza¬ lent of lent of Specific
degrees the liquid heat in pounds
per square tion, internal external volume
Fahren¬ B.t.u. B.t.u. of 1 cu. ft.
inch j B.t.u. work work
heit
B.t.u. B.t.u.
32 3.54 0.00 376.00 376.00 345.80 30. 20 1.278 0.048

50 5.51 21.28 393.76 372.48 341.48 31.00 0.844 0.073
68 8.31 42.80 411.12 368. 32 336.52 31.80 0. 574 0. 107
86 12.20 64.56 428.00 363.44 330.88 32. 56 0.401 0. 154
104 17.46 86.42 444.44 357.92 324.60 33.32 0.287 0.232

122 24. 32 88.76 460.44 351.68 317.64 34.04 0.210 0.294
140 33.17 131.20 476.00 344.80 310.12 34.68 0. 158 0.392
158 44.32 153.92 491.12 337.20 301.96 35.24 0.120 0.515
176 58.13 176.84 505.76 328.92 293.28 35.64 0.093 0.705

194 74.96 200.00 520.00 320.00 284.12 35.68 0.073 0.848
212 95.25 223.44 532.76 310.32 274.48 35.84 0.057 1.074
230 119.51 247.08 547.12 300.04 264.52 35. 32 0. 045 1.350
248 148.44 270.96 560.00 289.04 254.28 34.76 0.036 1.703
Ethyl Chloride.—A machine utilizing ethyl chloride as a refrigerant has been

on the market for some years. As the boiling point of this medium is high, the
apparatus works under a vacuum, and the maximum condensing pressure is
only about 15 lb. gage pressure. In a plant of this kind in Chicago, the expan¬
sion coils were made of copper, ovoid in cross-section, the section being about
4 or 5 in. wide and about 1 in. high. This type of machine has been often used in
small plants for butcher shops, restaurants, etc., but is hardly of any utility for
larger establishments. Its use is indicated for small experimental installations.
Methyl Chloride.—To avoid some of the disadvantages of ethyl-chloride
refrigeration, the use of methyl chloride has been proposed, and for some time
past, information has been published in regard to a plant of this type installed in
an ice manufactory in Algiers. A 30-hp. engine is said to be used, and the plant
is declared to work efficiently and smoothly with a brine temperature of 4°F.
Details as to the amount of ice produced, however, and the cost of manufacture,
have not been furnished.
Air Refrigerating Machine.—This form of refrigerating machine is a practical
exemplification of the art of cooling by converting the superfluous heat into work.
By this method, atmospheric air is compressed into a strong cylinder, provided
with cooling coils, through which water is circulated to carry off the heat of
compression. The compressed air is then used to drive an engine, whereupon
its temperature necessarily falls in proportion to the work done by it in the engine.
The air refrigerated in this way is then discharged into the rooms to be cooled,
which are also ventilated by it at the same time. The cycle is a continuous one,
but a certain amount of new power must, of course, be continually expended to
keep up the supply of compressed air, which is discarded as soon as it has taken up
its quota of heat. Plants of this kind were once largely used; but, on account
of their lack of efficiency, and the availability of better appliances, are now
seldom seen.
The Simple Vacuum Machine.—Vacuum machines constitute a class of
refrigerating apparatus in which water is the refrigerating agent. In simple
form, the principle under which they operate is much like that of a compression
machine. A vacuum pump is provided to exhaust the air from a closed chamber,
in which a vessel containing the water to be frozen is placed. The lowered pres¬
sure brings down the boiling point of the liquid, a violent ebullition ensues, and
the refrigerative effect produced by the vaporization of a part of the water turns
the rest into ice. The latent heat of ice being 144, and that of steam at 32°F.,
1,091; theoretically, the evaporation of 1 lb. of water should produce something
over seven pounds of ice.
The latent heat of steam being great in comparison with the specific heat of water,
the theoretical efficiency of a vacuum machine will be found to be much greater than
that of any other machine of the compression type. This however is entirely offset
by the tremendous size of the compressor required, which is 200 times as large as an
ammonia machine to do the same work.
If the temperature wanted is lower than that of freezing water, brine must be
placed in the freezing chamber instead of pure water, to prevent it from freezing and so
stopping the operation.
Compound Vacuum Machine.—To make possible a vacuum machine of

reasonable size, a process has been devised whereby the moisture evaporated
from the freezing chamber is absorbed by some chemical agent, usually sulphuric
acid, and the work on the air pump, in this way, much decreased, Of course,
the concentrated acid which is used, soon takes up so much water that its effi¬
ciency is destroyed, when it must be re-concentrated by heating; and the fuel
used to do this must be considered as part of the operating expense. Vacuum
machines working on this principle are relatively small in size, but, if the opera¬
tion is to be continuous, there must also be provided a pump to handle the dilute
and also the reconcentrated acid, and some sort of exchanger to effect a transfer
of heat from the strong acid leaving the reconcentrator, to the cold weak acid on
its way from the absorber to the reconcentrator. One of the great disadvan¬
tages attending the use of sulphuric acid is the fact that all apparatus and piping
with which the acid comes in contact must be lined with lead. This form of
machine is rather troublesome to operate, but, for ordinary temperatures, its
efficiency may be considered as fairly good.
Southby & Blyth Vacuum Machine.—A machine has been devised by these
gentlemen, in England, which comprises the usual vacuum box, a large
vacuum pump, a small splash air pump, and a water cooled condenser. In
beginning operation, the small pump is used to remove the air from the refrigerat¬
ing chamber, after which the larger pump is set to work to withdraw the water
vapor, discharging the same into the condenser. In order to make this pump
operative, so that the compressed vapors may be discharged without condensing
in the pump, the pump cylinder is heated, and the small air pump is kept in
operation to free the condenser from any occluded air which may come over with
REFRIGERATION 671
the vapors. Owing to the fact that the piston pressure is less than one-fifth of a
pound per square inch the compressor operates smoothly and with little friction.
The apparatus is best adapted for use on a small scale in places where the use
of ammonia or other noxious gases could not be risked.
Refrigeration by Liquefied Gases.—This method of refrigeration has no
practical value from a commercial standpoint at the present time, on account of
the great cost of producing refrigerants of this type. For the production,
however, of extremely low temperatures, chiefly for medical purposes, or for
physical, or chemical experiments, the use of liquefied air, and possibly other
liquefied gases cannot be over-looked, or passed by without a brief exposition.
Many of the so-called permanent gases were liquefied on a small scale long ago by
such scientists as Faraday, and later, on a larger scale, by Pictet and others, who used,
not only pressure, but also a series of other liquefied gases, the first cooling the
next, and so on, until finally a temperature obtained, low enough to liquefy the gas
experimented with.
A few years ago, however, these costly laboratory methods, were superseded by a
practical process devised by Professor Linde, who combined enormous pressures with
a cumulative self-intensifying system of refrigeration, by which the liquefaction of air,
nitrogen, oxygen, etc. can.be produced on a large scale, at a slight expense.
The Linde Liquid Gas Process.—By this method, the gas to be liquefied is
compressed by a powerful compressor, then sent through water cooling-coils
to remove the heat of compression. From there it is forced through a coil
several hundred feet long, which terminates in an expansion valve connecting
with a liquid receiver. Leaving this receiver is another pipe, which concentri¬
cally surrounds the first pipe throughout its entire length, and then leads back
to the suction side of the compressor.
The compressed gas, expanding into the liquid receiver, generates a certain amount
of refrigeration, or, in other words, lowers its own temperature. Then, as it flows back
to the compressor through the annular space between the concentric pipes, it absorbs
a certain amount of heat from the gas in the central pipe, moving toward the liquid
receiver, so that when the first gas reaches the compressor, ready to be compressed
again, it has lost the exact equivalent of the B.t.u. removed from it on the first round
in the water cooling-coils. It is then compressed again, and leaves the compressor
and the water cooling-coils with the loss of another modicum of thermal units, so that
when it reaches the receiver a second time, its temperature is lower than before, and its
pressure higher.
So it goes round and round, on each trip exchanging heat with the gas in the other
circuit of the pair of long concentric pipes, and at each passage through the water
cooling-coils losing a certain number of thermal units, part of which are expressed in a
lower temperature, and part in a higher pressure, until at last it reaches the critical
point, and condenses into a liquid in the liquid receiver.
The process is continuous, so that when the liquid has once commenced to collect
in the receiver, an additional amount is made with each stroke of the compressor.
New gas to take the place of that liquefied, is continually supplied to the compressor at
high pressure, from an auxiliary compressor and a second water cooling-coil. The
entire apparatus, of course, must be heavily insulated, so that practically no exchange
of heat can take place with the outside air or surrounding objects.
Practical Uses of Liquid Air.—The production of liquid air is probably the
crowning achievement of the Linde gas-liquefying apparatus. Its availability
as a refrigerant for the production of extremely low temperatures for medical

and experimental purposes has already been referred to. Among its other
important uses are the following.
Liquid air can be used for the production of liquid oxygen, which is, in reality, when
produced by the Linde process, a mixture of oxygen and nitrogen, consisting principally
of oxygen. The process is based on the fact that when liquid air is vaporized under
certain conditions, the nitrogen vaporizes first, leaving most of the oxygen behind.
Linde devised a special apparatus for the manufacture of this so-called “Linde oxy¬
gen,” which is a valuable agent in connection with certain manufactures.
Another valuable field for liquid air is the cheap and quick production of high
vacua. The vessel to be put under vacuum is filled with some gas more easily liquefied
than air, a neck being provided which is capable of being sealed off by the blow pipe.
The neck is then immersed in liquid air, and when the gas has all condensed into it, it
is isolated and removed, leaving a fine vacuum behind.
Expansion Side of Compression System.—The refrigerating medium, having

been passed through the compressor, and condensed into a liquid by the con¬
denser, is conducted to what is known as the liquid receiver, where it is stored
up ready to be used when needed to produce cold, or, in other words, to absorb
heat from the material which it is required to cool. In order to produce this
refrigerating effect, the liquefied refrigerant must be released, through what is
known as an expansion valve, into the interior of a system of pipes, coils, or
other confined vessels, through the walls of which the heat of the surrounding
liquid or air is absorbed by the refrigerant, as it vaporizes again to a gas.
The so-called expansion valve is a species of needle valve, arranged so that it may
be opened very gradually, and so regulate with great nicety the amount of the refriger¬
ant passing through it in any given time, and the cooling pipes or conduits, in the
interior of which the refrigerant is allowed to vaporize again to a gas, are known as the
expansion coils.
The temperature of the expanding refrigerant in these coils, and, as a consequence,
the cooling effect which will be produced by it, will depend upon its pressure;
and this may be regulated by opening or closing the expansion valve, or altering the
speed of the compressor.
Cold Storage Rooms.—The simplest method of cooling any substance is to

place it in a room which is under refrigeration, or, in other words, which is kept
at a low temperature. The first requisite of a cold-storage room is that it must
be well insulated, that is, constructed in such a manner that heat from the out¬
side cannot readily penetrate into it through its walls, floor, or ceiling. This is
accomplished by lining the room on all sides with what is known as insulating
material, and the value of any such material is based on the number of B.t.u.
which will pass through each square foot of it, per day of 24 hr. for each degree
Fahrenheit by which the temperatures inside and outside may differ.
This number of B.t.u. is known as the coefficient n of the material in question.

Below will be found a table giving the value of n for layers of various materials 1 ft.
thick.
Confined air, in itself, is an excellent insulator, but it must be rigidly confined in
small cells, and the whole kept perfectly dry. In fact, with most of the commonly
used insulating materials, which are usually employed in small aggregates packed
REFRIGERATION 673
tightly together, it is believed that the air

confined in the interstices plays no small Value of Insulating
part in producing a high insulation. Mois¬ Materials 1 Ft. Thick
ture, on the other hand, is most injurious;
and no insulation of any kind can be expected Value
to remain efficacious, unless dampness is Material of n in
wholly excluded. B.t.u.
As a general proposition, it may be stated
that a wall, ceiling, or floor, with first-class
Dry pine wood. 2.0
insulation, will pass from 1 to 2 B.t.u. per
Mineral wool. 1.6
degree Fahrenheit difference, per square foot,
Granulated cork. 1.3
per 24 hr.; while a good insulation will pass
Dry wood ashes. 1.0
from 2 to 3 B.t.u.; and a poor insulation, 3
Dry sawdust. 1.1
B.t.u. and over.
Dry powdered charcoal. 1.3
Among the insulating materials most
Dry cotton. 0.7
used, a material known as cork-board stands
Dry soft paper felt. 0.5
pre-eminent. It is manfuactured out of
natural cork, which is, for the most part,
CEILING 8c FLOOR INSULATION
WALL INSULATION
TEMP.
0°-10°
BELOW
TEMP.
ao-ao*
Fig. 5.—Cork-board construction.

imported from Spain. The cork is brought to this country in its natural condition,
ground up into small fragments, and then dumped into a metal box and pressed
tightly together by a powerful hydraulic press to form a thin slab, after which the
slab, with the box still confining it, is run into a hot oven, where the natural gum in the
cork is melted, and cements the whole together into a permanent sheet, from 2 to 4 in.
in thickness, as a rule.
The finished material is usually to be had in slabs, approximately 3 ft. long, 2 ft.
wide, and of any desired thickness, according to the work required of it. A sheet 2 in.
thick will have a heat leakage value of 5 B.t.u., per square foot, per degree difference,
per 24 hr. The walls of an ordinary ice box are usually lined with from 2 to 6 in. of
cork-board, depending upon the temperature of the outside air, and the degree of cold
to be maintained inside the box. A layer of cork-board 6 in. thick has a heat leakage
factor of only 1.6 B.t.u. per square foot.
In Fig. 5 will be found some details of the usual method of insulating ceilings,
walls, and floors with this material. For these refrigerated rooms, regular cold-
storage doors are to be had upon the market. They are wooden doors, insulated with
granulated cork, and built with a tapered or rabbetted edge, where the door proper
meshes with the heavy framework. Their heat leakage factor is in the neighborhood
of 2 B.t.u.
Cooling of Cold Storage Rooms.—The refrigeration of these cold-storage

rooms is accomplished by means of pipe coils, hung up, either on the side walls
or on the ceiling of the room, and the pipes themselves are cooled either by the
liquid ammonia itself, expanded directly into the pipe, or by brine at a low
temperature circulated through them by what is known as a brine pump.
The first of these methods of cooling the pipes is known as direct expansion, and
the second, as brine circulation. Whichever system of cooling the pipes is employed
however, the rate of heat transfer between the pipes and the still air of the room
is 1.7 B.t.u. per square foot of outside pipe surface, per degree Fahrenheit difference
between the temperature of the pipes and that of the room, per day of 24 hr. By
grouping the cooling coils of such a room in a battery, and circulating the air of the
room through them, by means of a fan, at a velocity of from 900 to 1,200 ft. per
minute, this rate of heat transfer may be increased to 5, or even 6, B.t.u., as a
maximum.
Practically, in order to obtain effective cooling without the installation of an exces¬
sive quantity of pipe, it is usual to assume that the temperature of the cooling medium
in the inside of the pipe will be at least 20°F. lower than the temperature required for
the air of the room, and all pipe surfaces are figured on that basis.
The Brine Cooler.—Where cold brine is to be used as the circulating medium,

it is most efficiently cooled by means of a multi-pass, shell-and-tube type, brine
cooler. A brine cooler of this type consists of a steel shell, cylindrical in form,
with steel tube sheets near each end, into which a large number of longitudinal
tubes are expanded. Cast-iron heads, with baffles or partitions forming an
integral part of them, are fitted over the tube sheets at the ends of the cooler;
and these are so designed as to cause the brine to pass successively through the
tubes six or eight times before it finally leaves the cooler. In this way the brine
is caused to pass through the tubes at a high velocity, resulting in a high rate of
heat transfer between the ammonia and the brine, and this is further increased
to a maximum by the presence of a liquid on each side of the tube walls.
Brine coolers are generally operated on the flooded system, or, in other words, an
REFRIGERATION 675
excess of ammonia is continually fed into the brine cooler, so that it will absorb
a maximum amount of heat from the brine passing through the cooler. The rate of
heat transfer between the ammonia and the brine is, under these conditions, 80 B.t.u.
per square foot of outside pipe surface, per degree Fahrenheit difference between the
ammonia and the brine, per hour. The difference in temperature between the
ammonia and the brine is generally figured at 10°F.; but for extreme conditions, as,
for instance, where a low brine temperature is necessary, notwithstanding the fact
that the back pressure on the compressor is high, by making the brine cooler very
large this difference in temperature can be reduced to 5 or 6°F.
Throughout this entire section, the rate of heat transfer has been spoken of as
being based on the degree of temperature difference existing on the two opposite
sides of the cooling surface. In arriving, however, at this temperature difference, it
should be borne in mind that it is the logarithmic mean temperature difference, and
not the arithmetical mean temperature difference, which must be used. The reason
for this is that when a warm body which is to be cooled, and the cooling medium which
is to be heated, first come into contact with each other, the rate of heat transfer is
greater than at the end of the cooling period when the temperatures are nearer together
and for this reason the mean temperature follows a curve instead of a straight line.
Cooling of Liquids.—Any liquid, of course, may be cooled in an appliance simi¬
lar to the brine cooler already described, and such a cooler may be operated
either on the flooded system or with closely regulated expansion valve, which will
admit only sufficient ammonia to effect the desired amount of cooling. When
so operated, however, the medium on one side of the tube walls will be ammonia
gas, and the transfer of heat, under these conditions, will be 25 B.t.u. per square
foot of outside tube surface, per degree Fahrenheit difference, per hour, in place
of 80 B.t.u. with the flooded system.
Another common method of cooling a liquid is to cause the liquid to flow down by
gravity over a vertical bank of coils, through the interior of which the refrigerating
medium is circulated. With this method, extremely good results are possible, as
the close contact existing between the surfaces of the pipes, and the thin layer of
the liquid flowing over them at a relatively high velocity, causes a rate of heat transfer
of at least 70 B.t u. per square foot of outside pipe surface, per degree Fahrenheit of
temperature difference, per hour.
Another method of cooling a liquid, which is much used in places where the danger
of freezing and bursting the pipes can be ignored, is by passing it through a double-pipe
cooler. This consists of two continuous concentric pipes or pipe circuits, usually built
in the form of a flat coil like that already described and illustrated for the double-pipe
ammonia condenser (see Fig. 2). The refrigerating medium is circulated through the
inner of the two pipes, while the liquid to be cooled is made to flow through the annular
space between the inner and outer pipes. As the area of the annular space is quite
small, the liquid may be made to flow through it at a very high velocity, thereby
greatly increasing the rate of heat transfer. With a speed of from 250 to 300 ft. per
minute for the liquid to be cooled, a rate of heat transfer can often be attained as high
as 125 B.t.u. per square foot per degree Fahrenheit difference, per hour.
Still another method of cooling a liquid is to place it in a brine-jacketed tank.
This, however, is probably the least efficient method of all, on account of the usual lack
of conductivity of the liquid itself. The efficiency of cooling tanks of this kind may be
greatly increased if they are provided with moving paddles to keep the liquid in
motion; but nothing like a maximum efficiency can be obtained, even in this way.
This method is seldom used, except where it is desired to hold a large body of liquid
under refrigeration while in storage.
In general, it may be said that cold-storage room piping should be figured on the
basis of a 20°F. difference between the medium in the pipes and the desired tempera¬
ture in the room; and, on this basis, 280 sq. ft. of external pipe surface per ton of
refrigerating effect will be required. Where double-pipe brine coolers are used, there
should be allowed approximately 15 sq. ft. of cooling surface per ton of refrigerating
effect; and with shell- and- tube coolers, 10 sq. ft. per ton.
Ice Making.—The manufacture of artificial ice is carried on, for the most
part, by one or two systems, which are known as the can system and the
plate system. The can system is the one most generally used.
Can System.—A standard ice plant of this type consists of a steel tank,
usually constructed of K in- metal, strongly riveted together, and filled with
brine, in which are placed cans filled with the water which is to be frozen into
ice. Running longitudinally in the interior of the tank, along its sides and
between each row of cans, below the surface of the bath of brine, are placed a succes¬
sion of cooling coils, which keep the temperature of the brine down to between
15 and 18°F. The tank itself is heavily insulated, both sides and bottom, with
several inches of cork-board, and its top is covered with a heavy wooden frame¬
work, with long rows of carefully fitted covers, 2 or 3 in. thick, one of which is
just above the top of each can, so that it may be removed and the can lifted
out when the water which it contains has been frozen into ice.
The ice cans are made of No. 16 galvanized iron, of a size to conform to the weight
of the ice block which is to be manufactured. They are filled with water, and are then
kept submerged in the cold brine until they are frozen solid. A standard commercial
cake is supposed to weigh 300 lb., and the can designed to produce an ice block of this
weight is 11M by 22K in. at the top, and 10K by 21K in. at the bottom, and
44 in. high. When it is immersed in a brine bath of from 15 to 18°F., it requires two
full days to freeze it solid. This means that when a tank of this kind is designed to
produce a given amount of ice per day, twice as many cans must be provided for as
are needed for a single day’s output.
Cooling Surface in Ice Tanks.—The cooling coils in the ice tank are operated
by direct expansion, or, in other words, they are fed direct with liquid ammonia
through an expansion valve. In regard to the proper quantity of cooling surface
to be installed in ice tanks, it is generally conceded that about 130 sq. ft.
per ton of ice per day is the most efficient amount. The cooling coils are usually
made of 1 in. pipe, and it requires 300 lin. ft. of this pipe to make up 130
sq. ft. of surface.
Hygienic Ice.—In making ice by the can system, the water to be frozen must
be practically free from impurities. Even occluded air is objectionable, as it is
almost sure to make the ice milky. Other impurities, such as earthy salts, metal¬
lic oxides, etc., produce unsightly streaks as the water is gradually frozen, begin¬
ning at the sides of the can, and finally concentrate in the center of the block,
to form a dark-colored, and often evil-smelling, core.
The general practice, in the manufacture of the (so called) hygienic ice, is to freeze
the ice from water obtained by condensing the steam exhausted from the steam engine
which drives the plant. Water so obtained is free from earthy and metallic impurities,
but must be re-boiled to remove all air, and subjected to an elaborate filtering, cleans¬
ing, and de-odorizing process to remove all traces of cylinder oil and make it sweet.
All this adds to the cost of manufacture.
REFRIGERATION 677
The greatest objection, however, to this plan is the fact that it limits the efficiency
of the plant, for even with the most wasteful and uneconomical type of steam engine
fco drive the machinery, the exhaust steam is hardly enough to furnish the necessary
condensed water from which to make the ice, and more must be furnished direct from
the boiler.
Plate-ice Process.—The disadvantages of manufacturing ice by the ordinary

can system, which requires that the water, to be frozen into ice, should be sub¬
jected to a preliminary treatment for the removal of its impurities, has led to the
development of what is known as the plate system. In making ice by this
method, the liquid refrigerant, or cold brine, is made to circulate through a system
of flat coils clamped between iron plates. These are suspended in a tank filled
with water, and the ice gradually forms on both sides of the plates, building
itself up, little by little, into a solid cake of the desired thickness, and leaving
behind the occluded air and the impurities which the water may contain.
With this process, no preliminary preparation whatever is required for the water
which is to be frozen. But, on the other hand, a large degree of skill on the part
of the operators must be had to carry it through successfully. The time needed to
freeze a plate of ice 12 in. thick is about 7 days; and when the freezing is complete, the
two huge sheets of ice, weighing many tons, must be detached from the plates by the
aid of warm water, and sawed up into blocks of the desired size. The original ice
sheets are usually about 12 by 20 ft. in area.
The advantage of this system, as has been stated, is the possibility of manufactur¬
ing clear and colorless ice from water which contains a large percentage of impurities.
The disadvantages are the large first cost of the plant, the difficulties attending its
successful operation, and the long time required to freeze a 12-in. cake, which is, of
course, due to the fact that the freezing is done wholly from one side, and the last
layer of ice on the outside must be frozen through nearly 12 in. of ice already in place.
For these reasons, comparatively few plants of this type have been built.
Raw-water Ice Plants.—Some years ago, it was discovered that clear color¬
less ice could be frozen in cans, from natural water of reasonable purity, if, dur¬
ing the entire process of congelation, the water was continuously aerated and
agitated by a stream of air bubbles passed up through it. A number of plants
were built to operate on this principle, but always they failed from the same cause.
An ice plug invariably formed in the air pipe, and stopped the process before
the freezing was complete. Many plans were tried out, in the effort to obviate
this difficulty, but without success; after a long series of experiments, conducted
by the writer at the plant of the Baltimore Independent Ice Co., in 1912, a feasible
scheme, was developed by which the continuity of the aerating process was made
certain and accident proof, so that every cake of ice produced in a large tank
could be depended upon to be clear and perfect.
By this method, a system of air headers is run along the top of the freezing tank,
past each can, just underneath the wooden covers. The headers are supplied with
compressed air at 20 lb. pressure and, through a series of air cocks, communicate with
the interior of each ice can at the bottom, by means of a short, heavy, flexible rubber-
pipe link, terminating in a metal drop pipe. Between the air cock and the flexible
connection—and right here is the crux of the whole process—is a pin-hole orifice,
through which the air must pass on its way to the can. So long as everything goes
along normally, the tiny stream of air, forced through the pin-hole, passes on through
the rubber hose and drop pipe, and bubbles up through the can of water quietly; but
let a plug of ice form to obstruct the flow of air, the pressure behind it will build up at
once, until the ice is forced out of the drop pipe and normal conditions re-established.
By this plan, a minimum of attention is required on the part of the attendants, quite
a necessary condition, as it would be manifestly out of the question to have men lifting
the covers from can after can, all day, to see if the air pipes were open. Every
successful raw-water ice system in use today utilizes this pin-hole orifice, or a sub¬
stantial modification of it.
The air for aerating the freezing water is furnished from an air compressor, and
after being compressed is passed through a suitable cooler, which reduces its tempera¬
ture approximately to that of the freezing tank and at the same time removes most of
its moisture. By the use of this process, ice may be obtained which is clear and
colorless throughout the whole cake, only a faint whitish trace marking where the air
pipe passed down through the can. In some cases, however, where there are highly
colored impurities, it is necessary, when the cake of ice is almost frozen, to remove the
remaining contaminated water from the core, replacing it with some that is clean, in
this way avoiding the presence of a dark-colored lump in one end of the cake.
The modem ice plant of today is necessarily of the raw-water type, as either of the
two older processes is far too expensive in fuel. With the hygienic ice plants of 10
years ago, a production of six tons of ice, per ton of coal burned, was considered
satisfactory. With the best class of raw-water plants now being built, the ratio is
approximately twenty to one. Beyond this, what is more important, in these days of
scarcity of help and other uncertain conditions, is the fact that raw-water plants
require no steam boilers necessarily. So electric power may be used whenever it is
available.
The Operation of Ammonia-compression Plants.—Since the efficiency of a
refrigerating machine is proportional to the number of pounds of gas it handles,
as expressed in cubic feet, in acordance with the particular conditions under which
it operates, care should be taken that the suction gas, as it enters the compressor,
has a temperature corresponding, as nearly as possible, to the suction pressure
under which the machine is working. To ensure this condition, the expansion
valve, which controls the flow of the liquid into the cooling coils, should be
opened widely enough to permit the coils to fill themselves up with the refrigerant.
This being done, the suction gas leaving the coils is not superheated above its
proper temperature.
The presence or absence of superheat in the suction gas can best be determined by
placing a thermometer in the main suction pipe, close to the machine. By closely
watching this thermometer, and regulating the plant accordingly, a maximum of
operating efficiency can be obtained; for not only will the compressor itself be working
under the most economical conditions, but the cooling coils, completely filled with the
liquid refrigerant, will present a pipe surface every square inch of which is thoroughly
wetted. In no other way can maximum efficiency be obtained from cooling surfaces.
Miniature Ice Plants.—During the last few years, a number of small compact
ice plants have been devised for use in private homes, small shops, or isolated
places where a regular supply of commercial ice cannot be had. They are gener¬
ally driven by some form of oil engine, or by electric motors, and have all parts of
the “high” side integral on one base.
Figure 6 shows a plant of this type driven foy an electric motor, a description of
which will serve for them all. Externally, the assembled plant takes the form of an
REFRIGERATION 679
electric motor and two cylindrical bodies, mounted side by side on an iron base. One
of the cylinders contains the rotary gear-type compressor, with its iron housing, all
being submerged in a bath of oil which fills the lower part of the cylinder. The
compressed gas is discharged into the upper part of the same cylinder, above the oil,
Fig. 6.—Miniature ice-plant.
and from there passes through a bent pipe to the top of the other cylinder, which is a
combined condenser and liquid receiver.
The cooling water enters at the top of the second cylinder, and first passes through
a helical coil, condensing the compressed gas into a liquid which collects in the bottom
of the chamber. From there the water passes on to a second coil, submerged in the
oil bath of the compressor cylinder, where it serves to cool the oil and, incidentally, the
compressor immersed in it, after which it flows away to waste. The lubrication
of the compressor is automatic, from the bath of oil with which it is surrounded, small
pin-holes being provided at the proper locations for the passage of the lubricant.
The lower part of the illustration is the expansion side of plant. This consists of a
tank filled with brine, in which are immersed the expansion coils, in which the refriger¬
ant is expanded back to a gas. The liquid refrigerant is fed from the receiver in the
bottom part of the second cylinder, through an expansion valve, to these coils; and
the expanded gas passes on to a small horizontal container, as shown, from which it
is drawn out by the compressor, in this way completing the cycle. The brine tank
is designed to be located in the interior of the ice box which the plant is supposed
to serve, and the supply of brine contained in it acts as a reservoir of refrigeration, so
to speak. A chamber is provided in the center of the brine tank designed to contain
cans filled with water to be frozen into blocks of ice.
The Conditioning of Air.—Some years ago it was noticed that, by eliminating
moisture from the air which is blown into a blast furnace, the quantity of coke
required to melt the charge was materially reduced and a patent based on this idea
was granted to James Gayley of the United States Steel Corporation. Accord¬
ing to his process, the air was passed through a large chamber, filled with heat¬
absorbing pipes, through which either cold brine or liquid ammonia was fed. For
this purpose brine is most generally used, and the temperature of the air is
lowered to a point where its remaining moisture content is less than 1 gr. per
cubic foot, after which the air is heated up again, before it is sent to the blast
furnace.
With the exception of the work done in connection with the Gayley process, the art
of air conditioning made very little progress until the beginning of the World War, when
it was discovered that the use of conditioned air in certain processes connected with
the manufacture of explosives had a great effect upon the time required for the com¬
plete burning of the powdered charge. For some time past, the U. S. Department of
Agriculture has been making exhaustive studies on the effect of conditioned air in
connection with the cold storage of food stuffs, etc., and it now looks as though, at no
distant date, it might come to be one of the great uses of mechanical refrigeration.
From the standpoint of the amount of power consumed, the cooling of air is much
more difficult than its heating. To heat air, it is simply necessary to supply enough
B.t.u. to raise the volume in question a pre-determined number of degrees, and the
heat required to raise the temperature of a given volume of air is relatively small, on
account of the low specific heat of air.
When air is to be cooled, however, its moisture content must be taken into account,
for in order to extract a pound of water from the air, it must be first condensed, and
this requires a quantity of heat substantially equivalent to the latent heat of steam, or
approximately 1050 B.t.u.
The process of cooling air, therefore, resolves itself into three parts: (1) The cooling
of the air with its original load of moisture from the original temperature down to the
temperature of saturation. (2) The condensation of sufficient moisture to produce the
desired relative humidity at the pre-determined final temperature. (3) The cooling
of the air with its remaining moisture content, and also the water which has been
condensed out of it, from the point of saturation down to the final desired temperature,
unless this water be previously drawn off.
In practice, the conditioning of air is usually carried on by first passing the air
through refrigerated chambers, until its temperature has fallen to a point where its
moisture content, in pounds at saturation, is equal to the number of pounds of
moisture which will obtain at the desired ultimate temperature and relative humidity.
All other moisture then condenses and is precipitated, after which the air, carrying
only the proper amount of moisture, is heated to the required temperature by passing
it over steam coils.
Below will be found a table giving the moisture content of air at various tempera¬
tures and percentages of saturation, by the aid of which any required calculations
may be made.
REFRIGERATION 681
Pounds Vapor in 1,000 Cu. Ft. of Air at Different Percentages and

Temperatures
Percentage saturation
Temperature,
degrees
Fahrenheit 50 55 60 65 70 75 80 85 90 95 100
100 1.411 1.552 1.694 1.835 1.976 2.117 2. 259 2.400 2.541 2.682 2.823
95 1.223 1.345 1.467 1.589 1.712 1.844 1.977 2.089 2.201 2.323 2.446
90 1.056 1.161 1.267 1.373 1.479 1.584 1. 690 1.795 1.901 2.005 2.110
85 0.909 1.000 1.091 1.182 1.273 1.364 1.455 1.546 1.637 1.728 1.819
80 0.781 0.859 0.937 1.015 1.093 1.166 1.249 1.327 1.405 1.483 1.562
75 0.668 0.735 0.802 0.868 0.935 1.002 1. 069 1.135: 1.202 1.269 1.336
70 0.570 0.627 0.684 0.741 0.798 0.855 0.912 0.964 1.026 1.083 1.140
65 0.484 0.532 0.581 0.629 0. 678 0.726 0. 775 0.823 0.872 0.920 0.968
60 0.410 0.451 0.492 0.533 0.574 0.615 0. 656 0.697 0.738 0.779 0.820
55 0.346 0.380 0.415 0.449 0.484 0.519 0.554 0.588 0.623 0.657 0.692
50 0.291 0.320 0.349 0.378 0.407 0.436 0. 465 0.494 0.524 0.553 0.582
45 j 0. 243 0.267 0.292 0.316 0.341 0.365 0.390 0.414 0.439 0.463 0.487
40 0.203 0.223 0.244 0.264 0.284 0.304 0. 325 0.345 0.366 0.386 0.407
Let us assume that a 60 per cent saturation, at a temperature of 65°F. is required

for a certain service. The moisture content under these conditions, as shown by the
above table, is 0.581 lb. of vapor per 1,000 cu. ft. Let us now take the natural air,
whatever it may be, and cool it to 50°F. at 100 per cent saturation. The moisture
content, whatever it may have been previously, will, in this way, be reduced to 0.582
lb. of vapor per 1,000 cu. ft., and when we have heated this air again, from 50°F. to
65°F., the air conditioning problem will be met for this particular case.
In conditioning air commercially, its temperature, as it passes through the brine-
cooled chambers to deposit its surplus moisture, and also its final temperature, when
it has afterward been passed over steam coils, is automatically controlled by thermo¬
static regulation, so that air conditioning, as it is practiced today, is quite automatic
as well as extremely simple.
Note.—In the conditioning of air a combination of chemical and physical means, or of chemical
alone, may be adopted. That is, the air may be passed over calcium or magnesium chloride to remove
some or all of its moisture. In event such chemical means are adopted, chemical control of the process
is necessary for its greatest economy, for some of the hydrates of deliquescent compounds give up their
water much more readily than do others, that is, with a much smaller consumption of heat unitB, and
the process should be run so as to attain that minimum. Even if an entirely chemical or an entirely
physical (refrigeration) method be adopted, it should usually be a two-stage-process.—Editor.
SECTION XX
OXIDATION AND REDUCTION*
By E. P. Schoch1
PART I
Origin of Terms.—Oxidation was first used to denote the combining of

oxygen with other elements or with their compounds—e.g., S + 02 = S02, 2S02
+02 = 2 S03. Next came the extension of the term to such changes as the
following:
2H2S + 02 = 2HaO + S2
on account of which the above definition was extended to: oxidation is the com¬
bining of oxygen with, or the abstraction of hydrogen from, other elements
or their compounds.
It was recognized long ago that some elements can exist in more than two oxidation
stages: thus, the sulphur in H2S is in the lowest oxidation stage; that in free sulphur
in the next higher stage; that in S02 in a still higher stage; and the sulphur in SO3 in its
highest oxidation stage. Since, furthermore, hydration {e.g. SO3 + H20 = H2S04),
and the neutralization of an acid by a base do not involve the action of free oxygen,
either directly or indirectly, it soon became evident that an element may be a com¬
ponent of entirely different compounds and yet be in the same oxidation stage in all of
them. Thus, the iron in FeCl2 and in FeS04 is in the same oxidation stage because
both can be obtained from FeO by reaction with HC1 or with H2S04 respectively,
i.e. by means of changes which do not involve oxidation and hence do not change the
oxidation stage of the iron from what it is in FeO. The same considerations show also
that in FeCl3 the iron is in a higher oxidation stage than in FeCl2—because FeCl3 can
be derived from Fe203 by reaction with HC1, and Fe203 would evidently be obtained
from FeO by oxidation.
Naturally, the term oxidation was then extended so as to include any changes of an
element from a lower to a higher oxidation stage irrespective of the manner in which
the change is actually brought about, and hence it is also applied to changes in
which oxygen takes no part e.g. 2 FeCl2 + Cl2 = 2FeCl3.
With the la,st definition, oxidation will designate widely different changes in the
chemical compounds of the elements affected. Hence, it is difficult—if not impossible
—to define oxidation by referring directly to the changes in chemical composition
produced thereby, and another means is commonly employed. This will be presented
farther on.
Reduction was first used by metallurgists to denote the extraction of metals from
their ores. The latter usually contain oxides of the metals to be extracted, and the
reduction involves the abstraction of oxygen from these compounds: thus reduction is
merely the reverse of oxidation, and that is its significance in chemistry today.
Oxidation of One Substance Involves Reduction of Another.—When a

substance, e.g., sulphur,— is oxidized by the oxygen furnished through the decom-
*In preparing this section the author has taken many of the tables from hiB "Introductory Course
of Lessons and Exercises in Chemistry” and has drawn extensively from the presentation of this
subject in the latter.
1 Professor of Chemistry, University of Texas, Austin,. Tex.
683
position of another substance which yields free oxygen readily—e.g., KClOs—

then it is evident that the latter is reduced, and that the amount of oxidation of the
former (i.e., sulphur) is dependent upon the amount of reduction of the latter (i.e,
chlorate). This is necessarily always true—even when the oxidation consists of
actions other than the direct transfer of oxygen.1
Oxidation-Reduction Reactions in Non-Aqueous Mixtures.—In non-aqueous
mixtures, particularly those which react at high temperatures, the actions
frequently consist of the simple transfer of oxygen, and such actions require no
elaborate theory for their consideration. Thus the action of lead on fused KN03,
which is used to prepare KN02, evidently is expressed by the equation:
Pb + KNOs = PbO + KN02
Similarly, the action of litharge upon metallic sulphides (e.g., PbS) as it occurs in
assaying, is easily understood from the equation
PbS + 2PbO = 3Pb + S02
In more complex examples, we avail ourselves of the fact that many ternary com¬
pounds containing oxygen are numerically equal to the sum of two oxides which can
exist separately. Thus Na2SC>4 is numerically equal to Na20.S03. Formulas
written thus are called “dualistic” formulas: they serve to show the relations be¬
tween the oxidizer and the reducer in complex reactions in non-aqueous mixtures.
Thus, when in the fusion of MnO + K2C03 + KN03, the green manganate K2Mn04
is produced at the expense of oxygen from KN03, the relations are easily obtained
from the following dualistic formulas of these substances:
KNO3 written dualistically is K20.N206

K2C03 written dualistically is K20.C02
K2Mn04 written dualistically is K20.Mn03
N206 gives up oxygen and leaves 2NO which escapes as a gas. MnO takes up oxygen
and becomes Mn03. Hence, 2N205 furnish enough oxygen for 3MnO. Since 3Mn03
combine with 3K20, one molecule of K2CO3 will be required to furnish K20, while its
C02 will escape as a gas. Hence the equation is
2(K2O.N2Ofi) + 3 MnO + K20.C02 = 3(K20.Mn03) + 4NO + C02
Another example of the value of writing formulas dualistically is the following:

potassium cyanate (KCNO) is made from potassium cyanide (KCN) by heating the
latter with K2Cr207. The oxygen is obtained from a part of the chromic acid oxide
in K2Cr207 (K20.2Cr03). The K20 retains one Cr03, and the other is reduced to the
form Cr203. These relations are satisfactorily expressed by the following equation:
3KCN + 2(K20.2Cr03) - 3KCNO + 2(K20.Cr03) + Cr203
For convenient reference, the,dualistic formulas of some common substances are

given here*, sodium carbonate*Na20.C02; sodium silicate, Na20.Si02; sodium
sulphate, Na20.S03; sodium sulphite, Na20.S02; sodium nitrate, Na20.N205;
sodium nitrite, Na20.N203; calcium phosphate, 3CaO.P2Os; chromium sulphate,
Cr203.3S03; potassium chromate, K20.Cr03; potassium dichromate, K20.2Cr08;
1 Logical procedure requires that free oxygen be considered as being in the oxidized state, and combined
oxygen as being in the reduced state, so that the change of oxygen from the free element to any com¬
pound amounts to its reduction. Similarly, free hydrogen must be considered as being in the reduced
state and combined hydrogen as being in the oxidized state. Thus the union of oxygen with hydrogen
appears as the reduction of the oxygen and the oxidation of the hydrogen.
OXIDATION AND REDUCTION 685
ferrous sulphate, FeO.SOs; ferric sulphate, Fe203.3S03,* manganous sulphate, MnO.

SO3; potassium manganate, K20.Mn03,* calcium permangante, CaO.M^Cb; potassium
chlorate, K2O.CI2O5; potassium perchlorate, K2O.CI2O7.
Reactions in Aqueous Mixtures.—The above procedure is quite satisfactory

for reactions in non-aqueous solutions, but it is not satisfactory for reactions in
aqueous solutions because in many of them oxygen is not transferred directly
from the oxidizer to the reducer, or does not appear to take any part whatever.
Hence, a more generally applicable theory is here to be introduced. It is applica¬
ble to all oxidation-reduction reactions without exception, and it represents the
experimental facts directly—that is, it does not require the assumption of any
hypothetical intermediate steps.
Valence Changes Due to the Liberation of Electrons Indicate Oxidation or
Reduction.—It is desirable to have a simple means for recognizing oxidation
changes Such a means should be something that is physically measurable,
and applicable to all changes. For the treatment to be presented here we obtain
a general indicator by considering the changes at the anodes of electrolytic or of
battery cells, and this will’be shown in the following examples.
Example 1.—In the electrolytic refining of silver, the silver anode dissolves in a
nitrate electrolyte and forms silver ion (Ag+), which together with the nitrate ion
present corresponds to silver nitrate. This change from metal to a constituent of
of this solution corresponds to the oxidation of the silver, because the same change
could be accomplished through the following steps:
4Ag + 02 = 2Ag20
Ag20 + 2HN03 = 2AgN03 + H20
Since the second step does not change the oxidation stages of the elements involved, it
follows that the first step—oxidation—is essential to the whole change. Hence, the
same change in the state of silver, however produced, involves oxidation, and the
anodic dissolution of silver considered in this example must involve oxidation.
The change at the silver anode is accompanied by the flowing of negative electric
charges from the anode through the connecting wire to the remainder of the circuit.
The amount of these electric charges is one faraday per 1-g. atom of silver, which
corresponds to one electron (symbol, ©) per Ag.1 Hence the equation for this change
is:
Ag° = Ag+ + 1©
This means that an atom of silver metal forms a univalent silver ion and one free
electron.
From this example, we may formulate the following definition of oxidation:
Oxidation is a change in the valence of an element produced by its giving up electrons.
Example 2.—When a solution of a chloride (e.g. HC1) is electrolyzed with a graphite
or a platinum anode, the change at the anode consists of the discharge of chloride ions
which produces free chlorine and free electrons according to the equation—
2C1- = Cl°2 + 2©
the free electrons constitute the electric current flowing from the anode through
the connecting wires to the rest of the circuit.
1 The electron is the unit charge of negative electricity: it is the amount corresponding to one
negative valence and the amount required to neutralize one positive valence.
The above change in the state of the chlorine—from the compound HC1 to free
chlorine—can be brought about by direct oxidation, as in the Deacon process:
4HC1 + 02 = 2H20 + 2C12
Hence any other procedure for changing chloride ion to free chlorine is considered to be
oxidation, and the anodic change above must involve oxidation.
The symbol for the chloride ion is Cl" and for the electrically neutral element is
Cll . These symbols indicate that the change undergone by chloride ion at the anode
amounts to the giving up of negative electric charges (electrons) as expressed by the
equation:
2C1“ = Cl°2 + 2©
From this we derive the same definition for oxidation as before: it is a change in the
valence of an element produced by its giving up electrons.
The changes at the cathodes of electrolytic cells consist of the reverse of the changes
at the anodes—and hence they are reductions. In these actions elements in various
forms of combination combine with the free electrons coming over the wire to the
cathode. Thus, in a cell with a silver salt solution for an electrolyte, the silver ions
at the cathode combine with the free electrons coming to the cathode through the
connecting wire, and form silver metal—according to the equation
Ag+ + 10 - Ag°
The silver atoms form a deposit on the cathode surface.
Again, with an electrolyte containing free chlorine in solution, the cathode change
consists of the formation of chloride ions, according to the equation
Cl°2 + 2© = 2C1"
The free charges for this change come through the connecting wire to the cathode.
Identity of Battery-cell Changes and Oxidation-Reduction Reactions.—A

battery cell differs from an electrolytic cell only in the fact that the poles of a
battery cell are composed of materials which, collectively, have a tendency to
react so as to send electrons out from the anode through the connecting wire
to the rest of the circuit, and when the battery is in action the current is allowed
to flow in accordance with the battery’s impulse; while the poles of an electrolytic
cell are composed of materials which usually have no tendency to react in any
sense, and in any case are impelled to change by an external force which draws
the electrons out at a predetermined anode through the connecting wire. Other¬
wise there is no essential difference between these two classes of cells and all that
has been said concerning electrolytic cells holds also for battery cells, that is,
while the current flows, oxidation takes place at the anode and reduction at the
cathode.1
Attention is now to be called to the following fact: the sum total of the changes in a
,
battery cell is an oxidation-reduction reaction and every mixture in which an oxidation-
reduction reaction takes place contains also the essential elements for constructing a battery
cell.
1 In assigning algebraic signs to cells, we must consider whether we are interested in the current
going in or coming out. With electrolytic cells, we are interested in the current sent in—hence the pole
at which we send the (positive) current in is designated as the positive pole, or the pole at which the
UPTlt in ift (•‘Q llarl + h a narra fitra "Rivf nri+h
To demonstrate the truth of the above statement, let us consider the gravity or the
Daniell cell, the negative pole of which consists of zinc dipping into zinc-sulphate
solution, while the positive pole consists of copper surrounded by a saturated solution
of copper sulphate. Different means are employed to prevent the two pole solutions
from mixing mechanically—in the gravity cell, the positive copper pole with its heavy
solution is placed in the lower part of the jar, with the negative zinc in the upper part
and the lighter electrolyte of the latter floating on top of the copper sulphate solution.
In the Daniell cell, one pole and its solution is placed in a porous earthenware cup
which is placed in a larger glass container, and the negative with its solution is placed
in the space between the glass jar and the porous cup.
When the poles are connected through a metallic connector, the cell will send the
(positive) current from the copper pole (positive) through the wires to the other
pole; or the electrons from the zinc pole (negative) through the wires to the other
polo. The chemical change at the zinc pole consists of the dissolution of zinc, and
is expressed by the equation:
Zn° - Zn++ + 2©
By definition, this change is oxidation: the zinc is oxidized. The change at the
copper pole consists of the deposition of copper on the pole, and is expressed by the
equation:
Cu++ + 2© = Cu°
By definition, this change is reduction: the copper ion is reduced. Furthermore,

it is evident that the rate at which the copper is deposited depends upon the rate at
which the electrons are supplied by the change of the zinc at the other pole, or in other
words, when estimated in number of electrons, the amount of oxidation taking place in
any period of time is equal to the amount of reduction produced in the same time.
The total change in the cell is the sum of the above two pole changes. Adding
them algebraically, and simplifying by dropping the electrons from both sides of the
equation, we obtain:
Zn° + Cu ,‘h - Cu° + Zn*H‘
The accompanying anion requires a little consideration at this point. Since under
equilibrium conditions, every portion of a solution contains equivalent amounts of
cation and anion, the discharge of the copper ion at the positive pole will leave the
liquid next to the pole momentarily with an excess of anion over cation. At the zinc
pole the formation of zinc ion will leave the liquid next to that pole momentarily with
an excess of cation over anion. This electrically unbalanced condition at the poles
causes a slight shift of all the cations toward, and of all the anions away from, the
zinc pole to an extent sufficient to “balance up” the excess of cations at the zinc pole
and of anions at the copper pole (electric transfer of ions). As a result of this, the
new zinc ions at the zinc pole are furnished sulphate ions in amount equal to those
given up by the copper ions discharged at the copper pole.
Hence, we may add these “inactive” sulphate ions to our equation above, and this
completes it to the usual form:
Zn° + (GuH' + 80r~) * (Zn++ + S04--) + Cu°
The foregoing chemical action in the battery cell takes place also whenever metallic
zinc is dipped into a solution of copper sulphate. A vessel containing such a mixture
really contains all of the essential elements of the above battery cell and wire connec¬
tion, as may be seen by imagining the two metal poles brought together until they
merge into one, and the connecting wire attains zero length, while the two liquids are
thoroughly mixed. Since the copper metal and the zinc sulphate are merely results of
the actions, they need not be used to begin with, hence a rod of zinc dipped into
copper sulphate solution contains all the essential parts of the “condensed” battery
cell.
There is no reason why we should assume that the materials in this mixture should
react in a manner different from the manner in which they react in the battery
cell, hence the details of the action presented by the battery cell shows the details
of the action in this mixture, and thus the battery cell enabled us to study oxidation-
reduction actions much more in detail than we could otherwise.
It is evident that we can reverse the above process, namely, take the constituents
of the oxidation-reduction mixture, and construct a battery cell: for this purpose we
place the substance to be oxidized (i.e. zinc) at one pole, and the substance to be
reduced (i.e. copper ion) at the other. To complete the mechanical arrangement, we
must put a salt solution around the zinc, any salt solution will do provided it does not
produce any change when these two are put together apart from the rest. For the
copper sulphate solution, the completion of the pole evidently requires the supplying
of an electric conductor—and this may be chosen at random from substances which do
not change when these two are put together apart from the rest. Graphite, platinum,
copper, etc., will evidently do for this purpose.
Since every mixture undergoing an oxidation-reduction reaction can generally
supply the oxidizer and the reducer separately, and since each one of these can be
supplied with such extra (inactive) parts necessary to complete the poles as shown
above, it follows that every oxidation-reduction reaction can be studied as a battery
cell.
Oxidation-reduction Reactions are Mostly Reversible.—If an outer electro¬

motive force with a voltage greater than that of the above battery is applied to
its poles in such a manner as to send a current in the direction opposite to that
considered above, then the actions at the poles will, in most cases, take place
in exactly the reverse sense: that is, the products of the above reactions will
change back exactly to the original substances, as shown by these equations:
Zn++ + 2© = Zn°
Cu° = Cu++ + 2©
This perfect reversibility is true of most oxidation-reduction reactions.

Independence of Kind of Action at Poles of Battery Cell.—The kind of an
action which an oxidizer (or a reducer) undergoes, and the force with which it
tends to undergo this action are not in any manner affected by, or dependent,
upon, the reducer (or the oxidizer) with which it is “coupled” in a particular
mixture or battery cell. Thus the zinc pole used above may be coupled with
a “silver nitrate-silver” pole in place of the “copper sulphate-copper” pole
used above, and it will be found that the change at the zinc pole is then in no
wise different from what it was before.
Measurement of Oxidizing Tendency.—If a voltmeter is applied to the poles
of the Daniell cell above, it will register 1.1 volts. If a “silver nitrate-silver”
pole is used in place of the copper pole, the cell will exert 1.56 volts. A chlorine
pole may be made up of a solution of chlorine in hydrochloric acid, with a rod of
graphite dipped into the solution to make the electrical connection. When this
is coupled with the zinc pole, and the voltmeter is applied to this cell, it will
register 2.11 volts. The largest voltage with this zine anode would be obtained
if a fluorine pole could be coupled with it: the voltmeter would then register 2.60
volts.1 The question arises: how much of those voltages is due to the zinc pole
in each case? We do not know how much, but if, as stated in the preceding
paragraph, the action of the zinc pole is not affected by the action in the other
pole, then it exerts the same amount of electromotive force in each case.
In order to formulate a definite notion concerning the origin of electric charges
(electrons) from different substances, we assume in accordance with the present day
view of electricity, that metals differ in their tendencies to dissociate* into free elec¬
trons and positive ions.- Since such resulting ions also react with electrons in the
reverse sense, it follows that then* must he* an equilibrium be»fwe*en these two oppositely
reacting ,substantia v,g,
%n >< ZnM >} 20
in the maintainanee of which equilibrium each active* eomponemt must he present in a
certain concent ration or exerting a certain “pressure.” This must he true of the
electrons as well as of the* other components. Hene*e ouch single pole, when at
equilibrium e.g. when no current is passing through it exhibits a particular electron
picssure eharae*terisfic of its essential chemical ingredients. The* two elifferent polexs
of a eudi (*\ert different electron pressures, and when the* wire* circuit, het-w<*cn them is
dosed, the* electrons will tlmv from the* pole* of greater pressure thremgh the* wire* to the
pole* of lessen* pressure: h«*n<*<* the pole* e*\e*rting naturally a gn*nter electron pressure
will have* its suhstane*e*s reacting in such a sense* as to mpluee the* edeetrems se*nt out,
while* the* othe*r pole* must re*net see as to use* up the* chad runs re*e*e*iv(*d or “forced in”
upem it.
The* difference he*twe*<*n the* (*h'etron pre*ssun*s of the* two poh\s at eepiilihnum is
commeud.v measur«*<l in volts.
Tabulation of Single Poles in Order of Forces, In order to connect the

voltage's e>f all the* <*le*edromotive*ly active* substitutes, we* profit'd as follows:
Since* the above* theoretical zinedluoride cell shows the* largest voltage* of any cell
made* up with this zinc anode, the* fluoride cathoele* evidently <*xerts a lesse'r opposing
voltage than any ot 1ht cathoele that could be* eoum*ete*d with this zinc anode*, and we*
may assume the volt age* of the* fluoride* pole to la* any small value*- say 0.1 volts.®
Then we* have* “ voltage of zinc pole* minim 0.1 equals 2.66,” or ”force* of zinc pole* ^
2.76 vedts.”
With this value* feu* the zinc pede*, the* vedtages of the* other ptde*s may bet obtained
from the* cell measurements above In the following simple* calculation, because, as
pointch! out above*, the electromotive force* of e*neh pede is the* same* in all combinations.
1 Thb value* in ohtttiurd by mlrutntinn ftom either dMn, be*<tupee a fluorine** peilo b experimentally
impotisiblt* on arremut of the* fiirf, that flaw its** ncO* upon watur &e*e'e»rclin<( te» thn euiuation:
ilV, ! ;?JlsO - WIliKj ! Oi
8 Or, ge*ne*riilly nprnkinia nil vulrfirr rlmndnu atofnw differ in their tendemdort tee tlbaeudatei into
('lecitrenia uuel tUonw vlmtrn-d \vm nrMnf iwlv to an outwit e’eimepouehutf to tho number of nWtron*
given up,
1 Tim value btt*» br*oi trd in oolrt to nuito* the* voltage* of the hyelroKen edeetrexle (Net. 17 in the
in'rompmiyinK Table*, p tUKU exaetly 2 \ olt«,
Oxidation Changes Arranged in Descending Order of the Electric Pressures

Produced Through their Reaction Tendencies
Relation of ten¬
Elements in reduced state (in order of Elements in oxidized state (in order of
dency to change increasing powers of acting as
decreasing powers of acting as
reducing agents) from left to right oxidizing agents)
expressed in volts
1. Potassium metal. 5.20 K+ salt solution -J- 1©

2. Sodium metal. 4.80 Na+ salt solution + 1©
3. Calcium metal. 4.50 Ca++ salt solution + 2©
4. Magnesium metal. 3.50 Mg++ salt solution + 2©
5. Aluminum metal. 3.00 Al+++ salt solution-j- 3©
6. Hydrogen gas. 2.80 H+ ion in NaOH solution (normal in
OH") + 1©
7. Zinc metal. 2.76 Zn++ salt solution -J- 2©
8. Sulphide ion (sodium sulphide solu¬
tion normal in S““). 2.55 S° element + 2©
9. Iron metal. 2.43 Fe++ salt solution + 2©
10. Cadmium metal. 2.40 Cd++ salt solution + 2©
11. Sulphide ion (saturated solution of
hydrogen sulphide in pure water). 2.35 S° element + 2©
12. Lead metal. 2.34 Pb++ in dil. H2SO4 + 2 © (saturated solu-
tion of PbSC>4)
13. Nickel metal. 2.17 Ni++ salt solution + 2©
14. Sulphide ion (saturated solution of
HaS in normal HC1). 2.15 S° element + 2©
15. Lead metal. 2.12 Pb++ salt solution + 2© (normal in Pb++)
16. Tin metal. 2.10 Sn++ salt solution + 2©
17. Hydrogen gas. 2.00 H+ ion in acids + 1© (normal solution
of H+ ion)
18. Bismuth metal (same for Sb). 1.80 Bi+++ salt solution *f 3©
19. Stannous ion (stannous chloride 1.76 gn++++ sait solution + 2© (stannic chlo-
solution). ride solution)
20. Mercury metal. 1.72 Hg+ salt solution + 1© (IlgaCL in normal
1 Cl- solution)
21. Copper metal. 1.66 Cu++ salt solution + 2©
22. Iodide ion (normal in I“). 1.46 1° element + 1 © (saturated solution of
iodine)
23. Ferrous ion (solution normal in Fe++) 1.25 Fe+++saltsolution + l©(normal inFe+++)
24. Silver metal. 1.20 Ag+ salt solution + 1©
25. Mercury metal. 1.20 Hg+ salt solution + 1©
26. Oxygen ion (in neutral solution)_ 1.18 Oxygen gas (0°2) (saturated solution)
+ 2©
27. Bromide ion (solution normal in Br“) 0.92 Br° element (saturated solution) + 1©
28. Sulphur in compounds, with valence 0.80 S++++++ compounds + free © (concen¬
less than 6 (+). trated H2SO4)
29. Nitrogen compounds, with valence 1.00 to 0.72 Compounds of NOa“ ion + free © plus
less than 5+. H+ (HNOa)
30. Cr3+compounds. 0.70 Cr6+ compounds + free © + H+ (solu¬
tion of chromic acid)
31. Chloride ion (solution normal in Cl~ 0.65 CP element (saturated solution) + 1 ©
ion).
32. Mn++ compounds. 0.48 Mn7+ compounds -f free © + H+ (solu¬
tion of permanganic acid)
33. Cl° element. 0.40 Cl in compounds where it has positive val¬
ences + free © + H+ (HCIO, HClOs)
34. Oxygen ions (in any acid solution Oxygen gas liberated from a platinum
having very few O” ions). 0.30 pole by electrolysis of solutions of oxy-
acids such as nitrates, sulphates, phos¬
phates + free ©
35. Pb++ ion (in saturated solution of Pb++++ (from solid PbOa in diluted
PbS04 in diluted H2SO4). 0.34 H2SO4) -I- free ©
36. Fluoride ion solution. ).10 F° element +1©
37. Theoretical 0 pole. 0.00 Theoretical 0 pole
(а) Force of zinc pole minus force of copper pole = 1.1 volts; or 2.7G volts - force of copper pole
=* 1.1 volts; hence, force of copper pole = 2.76 — 1.1 = 1.66 volts.
(б) Force zinc pole minus force of silver pole = 1.56 volts; or 2.76 volts - force of silver pole
= 1.56 volts; hence, force of silver pole — 2.76 — 1.56 = 1.2 volts.
(c) Force of zinc pole minus force of chloride pole = 2.11 volts; or 2.76 volts — force of chloride
pole = 2.11 volts; hence, force of chloride pole = 2.76 — 2.11 — 0.65 volts.
In the accompanying table, these forces of the single poles are tabulated in descending order of
their values.
Oxidation Changes.—For the changes shown in the accompanying table all

equations in the table were written so as to show the formation of free electrons—
that is, in each example the material on the left side is in the lower oxidation
stage, and on the right side in the higher oxidation stage: they all appear as
oxidation changes.
The voltages in this table have been obtained by using poles in which all solutions
have normal concentrations of the ingredients mentioned except where other concen¬
trations are stated specifically in the table. Since the reacting tendencies of sub¬
stances vary with their concentrations, it follows that the tendencies to reaction will be
different when the concentrations of the substances are different from those given
here: the voltages in the central column will be larger if the substances in the left
column are employed in a more concentrated form than that here mentioned, the vol¬
tages will be less if the substances on the right are present in a more concentrated form
than that here mentioned.
The relative tendencies to reaction indicated by the position of these substances in
the table is not changed extensively by ordinary diluting or concentrating of the
solutions; but with large differences in concentrations, such as the difference between
the hydrogen ion concentration in an acid on one hand and in pure water on the other
(the latter contains one-ten millionth as much H ion as the former!), the voltage of the
same material has largely different values. Only on this account do some materials
appear at several different places in the table. Compare No. 6 with No. 17, and No. 8
with 11 and 14.
In order that a particular reducer and a particular oxidizer may actually react
when mixed, their positions in the table must have the following diagrammatic
relation:
Diagrammatic Relation (in Table) op Reducers to Oxidizers which Actually

React on Mixing
Reducer (chosen from left changes to Products on same

side)-► line
Products on same line <-Oxidizer (chosen from
changes to right side)
This diagram should be “read” as expressing the following: the ‘ 'reducer,”

found in the left-hand column, changes to the right to products on the same line, while
an “oxidizer” found below these products changes to the left to products on the same
line with it;
A reaction between a reducer and an oxidizer the potentials of which arc close
together does not continue until one or the other of these substances is exhausted
because the lessening of the original materials brings the potentials closer together.
Thus, a mixture of KI and FeClg reacts as follows:
KI reacts thus: 2K+I~ = I°2 + 2K+ + 2©

FeCl3 reacts thus: 2 Fe+++Cl3~ + 2© = 2 Fe++Cl2"" + 2C1~
Together, thus: 2KI -f 2FeCl3 = I°2 + 2KC1 + 2FeCl2
With N/l (I-) solution saturated with 1% the potential of the first reaction is 1.46
volts and with a solution normal in both Fe++ and Fe+++, the potential of the second
reaction is 1.25 volts. Hence, with these concentrations, their combined action is
impelled by 1.46 — 1.25 = 0.21 volts. But as the iodide is used up, the first potential
becomes less, and as the Fe+++ is used up, the second potential becomes greater, so
that they will finally become equal. Then their combined reaction has no voltage to
impel it, and it ceases.
It is evident that with still smaller concentrations of iodide and Fe','++ the poten¬
tials will even have the reverse relations, and such a mixture will react in the opposite
sense, i.e. so as to produce iodide and Fe+++ from I°2 and Fe++. This reversal of
potential relations with simple changes in concentrations will occur in all cases in
which the potentials are close together, e.g., with silver metal and mercurous salt
(Nos. 24 and 25); copper and iodine (Nos. 21 and 22); etc.
Selection of Materials for Battery Cells or Oxidation-reduction Mixtures.—

To select the materials for any battery cell possible with the common substances,
proceed as follows. For the negative pole select any substance in the left-
hand column of the table on p. 690, it will change to the corresponding substance
in the right-hand column if any substance in the right-hand column on a lower
line is placed at the positive pole or if these substances are put directly together.
The second substance is forced to change to its corresponding substance in the
left-hand column (on the same line!) because the change of the first substance
produces a greater electron pressure than the second can withstand. The
voltage of the cell is equal to the voltage of the first change minus the voltage of
the second change.
The Daniell cell is made up by placing zinc (i.e., No. 7 left) at the negative pole, and
copper sulphate {i.e., No. 21 right) at the positive pole. Its voltage is 2.66 — 1.56 =
1.10 volts. The reactions which take place are
Zn° - Zn++ + 2©
Cu++ + 2 © = Cu°
Adding we obtain
Zn° + Cu-1"1- = Zn++ + Cu°
To obtain the usual form, we add the “inactive” anions thus
Zn° + (Cu++ + S04—) = (Zn++ + S04”) + Cu°
Other cells in which the same kinds of changes take place—of a free metal at the
negative pole and of a metal ion at the positive pole—are easily made up from the
table in the same way.
Another well known kind of oxidation-reduction reaction—namely the- displace¬
ment of bromine or iodine by chlorine—is obtained by coupling No. 22 left with No. 31
right The reactions which take place are
21- = 1°2 + 2©
Cl°2 + 2© = 201-
Adding we obtain:
21- + Cl°2 = 1°2 + 2C1“
To obtain the usual form of the equation, we must add the inactive cation—
(2K+ + 21 “) + Cl°2 - (2K+ + 2C1") + 1%
A third kind of oxidation-reduction reaction is obtained by coupling No. 19 left

with No. 23 right: since the total number of electrons formed by the oxidized substance
must be taken up by the reduced substance, the equation for No. 23 must be multi¬
plied by two throughout. Thus we obtain:
Sn++ = Sn4+ + 2©
2Fe+++ + 2© = 2 Fe++
Adding and inserting the inactive anions present, we obtain:
SnCl2 + 2FeCl3 = SnCl4 + 2FeCl2
Any other oxidation-reduction reaction that will actually take place may thus be
ascertained from the table, and the equation derived as shown here.
Special properties and other data concerning the substances which take part in
oxidation-reduction reactions are given below under sections devoted to the different
elements. However, before proceeding to those sections, we must consider the single
reactions of ternary and other polyatomic oxidizing (or reducing) agents.
Calculation of Polar Valences cf Atoms in Compounds of More than Two

Different Elements.—In oxidation-reduction actions, many ternary compounds—
such as HN03—or ions composed of radicals—such as N03“, undergo changes
which indicate that they can ionize in a manner different from the ordinary.
Thus the fact that NO2, HN02, NO, N20, or N2 may be obtained from HN03 by
different reducing actions indicates that the latter may form any one of the
following sets of products:
(a) H+, 0— and N02;
(b) O— and HN02;
(c) H+, 20““ and NO;
(d) 2H+, 50““ and N20 (from 2HN03);
(0 2H+, 60““ and N2 (from 2HN03);
All of these sets of products have an excess of negative charges over positive
charges, and since HN03 as a whole is electrically neutral, it follows that this
excess of negative charges must have been added from without—as expressed by
the following equations:
(a) HN03 + 1© = H+ + 0" + N02
(b) HN03 + 2© = 0— + IINO2
(c) HNOa + 3© = H+ + 20““ + NO
(d) 2HN03 + 8© = 2H+ + 50— + N20
(e) 2HN03 + 10© = 2H+ + 60” + N2
This suggests that these actions may take place as follows:
In (e) the hydrogen and the oxygen atoms separate in the form of ions (as H+ and
0 respectively) and leave the nitrogen atom as an ion charged with as many charges
as will make the sum of the negative charges equal to the sum of positive charges on all
these products: since the three O together are charged with six negative charges,
and the H+ with one positive charge, the N must be charged with 6 — 1=5 positive
charges. Then the free electrons act upon the N51* ion (5© upon 1 N5+) to form
the free element, N 2 while the H+ and O ions remain temporarily or permanently
as free ions.
,
In the other reactions above, a, b c, and d—we find it convenient to assume that the
reactions take place as in (e),—namely that HNO3 is ionized into elemental ions—
and to consider that the free electrons react only with the N5+ ion while the other ions
are not affected by electrons. According to the number of electrons in the equations
above, the N54 ion will be changed to N+, N++, N+++, or N++++ in equations (a) to
0d) respectively. Vice versa, the number of electrons required in a change is equal
to the algebraic difference between the valence of N in the original and in the resulting
compound.
Finally, the ions temporarily formed will recombine as far as possible to form those
combinations which have little or no tendency to dissociate. Thus, H+ and 0 ions
will combine to H20 until one or the other of these kinds of ions is exhausted; N4+
will combine with 20 to form N02; N++ will combine with O to form NO; etc.
Hence, the complete changes (a) to (e) above are presented by the following equations:
(а) (H+, N6+, 30—) + 1© + H+ = N02 + H20

(б) (H+, N6+, 30—) + 2© + 2H+ = HN02 + H20
(c) (H+, N6+, 30—) + 3© + 3H+ = NO + 2H20
(d) 2(H+, Ns+, 30“) +8© +8H+ = N20 + 5H20

(e) 2(H+, N5+, 30—) + 10© + 10H+ = N2 + 6H20
We find in the same way, that every compound which is the essential “ actor ”
in an oxidation or in a reduction action contains one particular element which—like
N in HNO3—changes its polar1 valence, and which is the particular element oxidized
or reduced. Hencd, it is important to know how to “figure out” the elemental ions
of such compounds, and for this purpose the following suggestions are given here.
1. In the ionization of compounds, hydrogen and the basic metals of salts always
form ions with positive charges corresponding to their valences.
2. Oxygen generally forms the bivalent anion 0 , or together with hydrogen it
forms the monovalent hydroxyl ion, OH”. The chief exception to this rule is met
with in peroxides—e.g., H2O2 and Na202—in which oxygen is present as 0~ because
there can be no doubt that the hydrogen would form 2H+, or the sodium 2Na+.
If we know what ions the cations and the oxygen in a ternary electrolyte form,
then we can calculate the polar valence of the third element by using only the simple
rule that in ionization positive and negative charges are always formed in equal num¬
bers. This has been shown with HNO3 above, and may be shown again with the
following substances:
(a) KMn04: the oxygen will form four bivalent anions, 40 . Thus eight nega¬
tive charges will have been formed with these 40 and eight posi tive charges formed
simultaneously: the potassium atom takes one of these to form K+, and seven positive
charges must be taken by the only remaining atom, which is Mn. Hence the polar
valence of the latter is Mn7+, and the elemental ions of KMn04 are, K+, Mn7+, 40 .
(1b) K2Cr207: 70—ions will evidently be formed, hence 14 positive charges must be
accounted for. The 2K+ take two of them, and hence the 2Cr atoms take 12, or each
Cr takes 6 positive charges. The elemental ions of K2Cr207 are: 2K+, 2Cr6+, 70 .
In ternary compounds containing a halogen or sulphur in place of oxygen, it is
logical to assume that the halogen forms the well known anions—e.g. Cl~, Br“, I",
and sulphur the ion S . The following examples illustrate this.
NH4C1: on complete ionization of this compound, four H+ will be formed, and
since Cl takes only one of the four negative charges formed simultaneously, the N
must take the remaining three. Thus the elemental ions of NH4C1 are: N3” 4.H+,
Cl”
Na3AsS4: the sodium atoms form 3Na+; the sulphur atoms form 4S ; hence the
arsenic atom will be charged with 8 — 3—5 positive charges.2
The Forming of Any Compound from Elemental Ions in a Mixture.—Just as

any polyatomic compound may be resolved—or considered to be resolved—into
its elemental ions, so any compound may be formed by the assembling of the
necessary elemental ions. Thus when aluminum is formed in or introduced
into a solution containing NaOH, the aluminum ion, Al+++, will appear as
NaA102: this is evidently formed from Na++ Al++++ 20” ”. Similarly,
pentavalent arsenic will appear in a solution of NaOH in the form of Na3As04:
this is formed by the union of 3Na+ + As5+ + 40—. Again, in the reduction
of nitrate ion to ammonia in acid solutions (HC1), the nitrogen will finally be
present as NH4Cl: this is formed by the union of N3~ + 4 H+ + Cl”.
1 Since we must distinguish between positive and negative valences, these valences have been
designated as polar valences.
2 It is not Intended to convey the idea that any large fraction of a ternary substance exists in the
form of separate elemental ions: it is merely intended to show that it must ionize in a manner equiva¬
lent to what w \s shown above. The fraction of a substance present in the form of elemental ions is
It should be realized that all the necessary atoms for such compounds are available
in the form of elemental ions, and need merely to be added together. They may not be
present as free ions in any large amount, but they are obtainable by the ionization of
their compounds which are present. Thus, the needed for NaA102 is not present
as such, but is obtained from OH” or from H20; when thus obtained, the whole complex
(OH“ or H20) may be added to the left side of the equation, and the unused ions (H+)
left temporarily as free ions on the right side.
General directions for writing equations with polyatomic oxidizing or reducing
agents are as follows: (1) Write—in the form of elemental ions—the formula for the
compound of the element to be oxidized, putting it on the left-hand side of the equal
sign, and put on the other side, the “elemental ion’7 formula of the resulting compound
of the oxidized element. (2) Add to the right side, as many electrons as correspond
to the valence-change of the oxidized element. For a “check” method of estimating
the number of electrons, see the next paragraph. (3) Ascertain what other ions will
be required to form final stable compounds of all the elemental ions in the original
compound of the oxidized element, and add these ions—or the compounds from which
they are obtainable by ionization—to both sides of the equation, placing together the
ions of all compounds which naturally belong together.
Thus, if the original compound of the oxidized element is AsCl3, and the As is
changed to H3As04, then we put (As3+, 3C1”) on the left side and (3H+, AsB+, 40—)
on the right side, as per (1) above. Then we add 5 — 3=2 electrons on the right
side, as per (2) above, and finally we secure 3H+ for H3ASO4 and 3H+ more for the
3C1”—total 6H+, and 40 : or it might be better to secure 4H20, and to leave 2H+
on hand. In the latter case, the equation obtained would be—
AsC13 + 4H20 = H3As04 + 2H+ + 2© + 3HC1
For elements that are to be reduced, the corresponding equations are derived in the
same manner except that the electrons are to be placed on the left side.
Thus for the reduction of K2Cr207 to Cr2(S04)3, we obtain by this procedure:
(2K+, 2Cr6+, 70””) + (14H+, 7S04~) + 6© = (2Cr3+, 3S04-~) + 7H20

+K2S04 + 3S04
The equation for the reaction between any such polyatomic oxidizing (or reducing)
agent with any particular reducing (or oxidizing) agent is obtained by the general
procedure illustrated in the last example on p. 692, and it consists of the follow¬
ing: multiply either one or both of the equations for the single actions of the
oxidizer and reducer by factors which will make the number of electrons in one equal
to those in the other; add the resulting equations, and insert such other ions to both
sides as may be present and necessary to “couple up” all ions. For illustrations of
this procedure, see particularly the paragraphs below headed “Nitrogen.”
Simplified Method for Writing Equations for a Single Oxidizer or

Reducer. (Second Method for Estimating Free Electrons Involved).—(1)
Express by their common formulas the compounds, free elements, and ordinary
ions used or produced in a reaction, writing ions with their proper electric
charges, but whole compounds and free elements without any charges.
(2) See that every atom which is introduced on one side appears also on the other.
(3) Take the algebraic sum of the ionic charges shown on all the ions on each side of
the equation and add on the proper side as many electrons as are necessary to
balance the two sums.
Condition (3) above presents a second method for estimating the number of
electrons in the action of an oxidizer (or reducer) and this method enables us to do
without the elemental ionizing of compounds. We consider the compounds or ion

radicals simply with regard to their electrical condition as a whole which is zero for dis¬
tinct compounds, such as H20, HN03, etc., or the usual ionic charges for the ordinary
ions such as N03~, MnO^f, etc.—and we make the difference between the algebraic
sum of all apparent electric charges on the left of the equation equal to that on the
right by inserting the necessary free electrons. If these are to be added to the
right, the change consists of oxidation; if on the left, the change is reduction. For
an illustration, see the equations below on this page.
This method is to be used particularly when the elemental ionizing of compounds
is bothersome e.g., with carbon compounds. For illustration see below, under
“Carbon” (p. 705 et seq.).
Writing of Separate Equations for Oxidizer and Reducer is both Important

and Simple.—The proof that the substance oxidized and the substance reduced
react independently of each other except with respect to the free electrons passed
from the former to the latter has made it possible to study and describe their
properties and relations much more definitely than was possible otherwise:
this separation of the two actions is the key to the whole modem study and treat¬
ment of this subject of oxidation and reduction. Hence, throughout this chapter,
the reactions of the substances oxidized and of those reduced are presented
separately, and it is shown that reactions between any particular individuals are
merely the algebraic sums of their separate reactions. The reader should realize
the importance of studying the individual reactions of oxidizers or reducers, and
to practice obtaining their combined reactions by adding, algebraically, the
individual reactions previously obtained.
The writing of the individual reaction is after all a very simple matter because it is
almost immaterial in what form the symbols of the substances are written. For
illustration let us try to write the equation for the oxidation of SnCl2 to SnCl4. We
may consider that this is simply a change of Sn++ to Sn4+, and hence write the equation
thus
Sn++ = Sn4+ + 2©
Or we may write the equation with both chlorides, adding the 2 Cl" required in the
form of free ions, thus
SnCl2 + 2C1- = SnCl4 + 2©
Or we may add the 2C1“ in the form of 2HC1, thus
SnCl2 + 2HC1 = SnCl4 + 2H+ + 2©

All three equations are equally correct and useful.
An application of the “Simplified Method” above may be made here: note that
all substances are written so as to express electric conditions correctly. Note also
that the number of free electrons necessary to balance the sum of the electric charges
shown on the free ions on the two sides of each equation is 2© in all three equations.
To learn how the latter is obtained, see the last equation above: here the algebraic
sum of the charges on the substances on the left is zero plus zero ( — 0); that for the
substances on the right is zero phis two positive ( = 2+). Hence, we must add 2©
to the right side.
Oxidizing Powers of Different Oxidation Stages of an Element—At first

sight it might appear to us that the different oxidation stages of an element with
three or more oxidation stages e.g., nitrogen—should have oxidizing powers in the
order of their oxidation stages, that is, NO should be a stronger oxidizing

agent that N20, HN02 stronger than NO, etc. In some cases this is true,
but it is not true in general although the formation of the higher oxidation
stages usually takes place by oxidation of lower stages. The above erroneous
notion is probably due to the analogy which most of us draw between the
“raising” of an element to different oxidation stages and the lifting of an object
to higher levels: this analogy may be followed out so as to correspond to the
facts observed.
For this purpose, we need merely consider that the object instead of remaining at
the highest point to which it is raised (and at which it would have the correspondingly
higher potential energy than at the lower oxidation stage), may have dropped to
a lower level on the other side of an obstacle over which it had to be raised: in its
final position it may then have a potential of position even lower than that from which
it was last raised. Thus completed, this analogy may serve to explain why higher
oxidation stages are not necessarily possessed of a greater potential to return to
lower stages—or in other words, are not necessarily stronger oxidizing agents than
lower oxidation stages; and correspondingly, lower oxidation stages are not necessarily
stronger reducing agents than higher oxidation stages. Thus we find that ammonia
(in which the nitrogen is reduced to the N3~ stage) is not os strong a reducing agent
as liydroxylamine (in which nitrogen is reduced only to the N“ stage), and that per¬
chloric acid (in which chlorine is in the Cl7+ stage) is nofcos strong an oxidizing agent
as chloric acid (in which chlorine is in the Cl6‘t' stage).
The chemical fact which in these examples corresponds to the dropping of the
raised object to a lower level on the other side of an obstacle over which it had to be
raised is the formation of a relatively stable combination of atoms after the electronic
oxidation—i.e.f the valence change in the N or in the Cl respectively in the examples
quoted above—has taken place. In these examples, the stable combinations of
atoms are NH3 and HC104 respectively. The word stable designates compounds
which have a less tendency to dissociate into the substances from which they have been
formed than corresponds to the energy expended in forming them, and the formations
of such compounds always involves a loss of free energy in the reacting mixture. The
latter corresponds to the “dropping of the object to a lower level” in the above analogy.
Potentials of Changes Involving Hydrogen and Oxygen Ions Vary with

Acidity of Solutions.—The potentials of polyatomic oxidizing or reducing agents
which in ionizing form 0 or H+, or combine with them to form their final
stable compounds, are affected by the concentration of H+ (or OH~) in the solu¬
tion because either 0— or H+ is formed or used up in excess of the other, and
this “excess” of one or the other must be given to, or obtained from, the rest of
the solution. The ease or difficulty of obtaining 0— or H+ from the rest of the
solution (or giving them to the rest of the solution) naturally depends directly
upon the concentrations of the free O— or H+ in the solution, and the relation is
the following: the potential with which such a substance tends to be oxidized
(a reducing agent!) is greater in alkaline solutions than in neutral, and is least
in acid solutions. This is due to the fact that all such reducing agents use
up an excess of 0— over H+, or produce in the change, an excess of H+ over 0 .
For example, the tendency of arsenic trichloride to be oxidized to the next higher
oxidation stage of arsenic is least in acid solutions, greater in neutral solutions, and
greatest in alkaline solutions. The equation for the change is
(As+++, 3C1-) + 4H20 = (3H+, As5+, 40“) + 2© + 5II+ + 3C1~

This equation shows that the action produces free H+, and these must be added to the
rest of the solution: the disposition of these H+ will naturally be easier in an alkaline
solution in which they combine with OH—or in other words, in which the H+ con¬
centration is low—than in an acid solution in which the H+ concentration is high.
This effect of the H“ concentration is expressed more definitely by the following mass
law relation between the concentrations of all substances present when these materials
are in a solution by themselves and the above reaction has attained equilibrium: the
concentration of the electrons then is proportional to the potential. To get this mass-
law relation at equilibrium, we must take the product of the concentrations of the
substances on the left of the equation above—raising each factor to the power equal
to the coefficient of its substance in the equation above; then we secure in the same
manner the corresponding product of the concentrations of the substances on the
right, and we multiply the latter by a constant, and put the two products equal to
each other. Thus we obtain:
[As+++] X [Cl? X [H20? = [HaAsOJ X [H+]& X [Cl-]* X [©]2 X (Const.)!
Here the square braces around the substances designate their concentrations.
Since the electron concentration is proportional to the potential” or “voltage”
with which this reaction tends to take place, and which this mixture would produce
if used with a platinum pole as one pole of a battery cell, we will divide the whole
expression by all factors on the right except the electron concentration, and thus
obtain:
[As-*"*-+] X [Cl]3 X [H2Q]4_ = rc,12
[H3ASO4] X [H++P X [Cl-]* X (const.) LUJ
It is evident from this expression that any decrease in the value of H+ inceases the
value of the electron concentration, and thus raises the voltage (or position) which
this “pole” would have in our table: in other words, it would increase its tendency
to change to the oxidized state.
A familiar illustration of the corresponding effect upon the potential with which a
substance tends to be reduced (an oxidizing agent) is found in the action of the nitrate
ion in acid and alkaline solution respectively: in both, the nitrate ion may be reduced
to the nitrite ion, according to the equation
(N6+, 30™) + 2© = (N*+ 20-") + 0—

or
(NO,)" + 2© = (N0,)“ + 0—
In acid solutions, the concentration of free 0“ - is extremely small, but in alkaline

solutions, the O concentration is relatively large. This greater concentration of
0 ions in the alkaline solution produces a greater opposing effect than its lesser
concentration in the acid solution, and hence a solution of HNOs (&.e., N03“ in acid
solutions) is a much stronger oxidizing agent than a solution of NaNC>3 (i.e., N03“
in neutral or alkaline solution).
Hence, the general rule that oxidizing agents are strongest in acid solutions, and
reducing ajents strongest in alkaline solutions. However, this applies only to reactions
of oxidizing or reducing agents in which H+ or OH- ions are involved; while reactions
such as the following do not change their “potentials” when the H+ or OH” concen¬
trations of their solutions are changed:
Cl°2 + 2© = 2C1“
Mn04~ + 1© = Mn04~“
For a more extensive consideration of these effects, see the section on “Halogens”
below.
Substances which Act Simultaneously as Oxidizers and as Reducers in the

Same Mixtures.—A very striking example of oxidation-reduction reactions is
found in the case of a number of substances such as NO2 in cold water, free
elements (halogens, sulphur, phosphorus) in alkaline solutions, hypochlorites in
either acid or alkaline solutions, etc., in which some molecules of a substance
change to a more oxidized state while other molecules of the same substance
change to a more reduced state. Thus N02 dissolves in cold water and changes
according to the equation
2N4+ 02 + H20 = HN3+02 + HN5403
The valences of the nitrogen in the different compounds are marked in this equa¬
tion, and these marks show that one N02 is reduced to HN02 while the other is
oxidized to HNO3, the amount of reduction (addition of 1© to N02)
being equal to the amount of oxidation (liberation of 1© from 1N02).
Again, chlorine dissolves in NaOH solution and reacts according to the equation
Cl°2 + 2NaOH = NaCl“ + NaOCl ++ H20.
The “marked” valence on the chlorine atoms shows that 1C1° is reduced to Cl- (with
addition of 1©) while 1C1° is oxidized to Cl+ (with liberation of 1©).
In order to understand how these reactions are related to the ordinary oxidation-
reduction reaction, we must realize, first of all, that the force, potential, or voltage
with which a substance tends to change its valence depends upon the change undergone.
i.e., the whole transformation expressed by the equation—and not merely upon the
original substance. Thus the change from 01° to Cl“ has a potential of 0.65 volt,
while that from Cl° to Cl+ (or HO Cl) in acid solution is only 0.37 volt, or in other
words, as an oxidizing agent the potential of chlorine is 0.65 volts, but as a reducing
agent in acid solution, its potential is only 0.37 volts.- Similarly, any element in an
intermediate stage of oxidation will exert one potential to pass to a higher stage and
another potential to pass to a lower stage.
Furthermore, we must recall the fact that, in order that a reducer may actually
react with a particular oxidizer, the potential of the reducer must be greater than
that of the oxidizer (according to the method of reckoning potentials employed in
this chapter). In this connection, see the diagram given in the “Note on Table of
Oxidation Changes.”
It may now be seen why the same substance may act simultaneously as oxidizer
and as reducer: to do this, it need only have the potential of its change to a higher
oxidation stage greater than the potential of its change to a lower oxidation stage.
Thus, in alkaline solutions, chlorine changes from Cl° to Cl+ (i.e., C10~ ion) according
to the following equation
Cl°2 + 4(OH)“ - 2(Cl+0 ) + 2© + 2H20 (potential 1.57 V.)
while its change from Cl° to Cl”, which takes place according to the equation
Cl°2 + 2© - 2C1-
has the potential of 0.65 volts. Hence these two reactions have their potentials
properly related to act together (see diagram—“Notes on Table”) and their combined
reaction, as always, is the sum of the separate changes of the oxidizer and reducer
(with the inactive 2Na+ added to both sides).
Cl°2 + 2NaOH = NaOCl + NaCl + H20.

General considerations of this kind of reactions show that the general relations
between the potentials of the oxidizing and of the reducing changes should be as
shown: for this particular example, the potentials obtained or calculated from measure¬
ments confirmed our general conclusions. Corresponding measurements have not
been made, so far, on all other examples, but it is quite certain that the same general
relations must exist in all cases. Thus for the reaction between N02 and cold water, the
oxidation change
N02 -f OH- = HNOs + 1©
must have a greater potential than the reduction change

N02 + H+ + 1© = HN02
Since it was shown in the preceding paragraph that reactions which involve the
ions of water (H+ or OH") change their potentials when the solutions are changed from
acid (great H+ and small OH“ concentrations) to alkaline reaction (small H+ and
great OH~ concentration), it follows that the above relation between the potentials
of the oxidizer and the reducer may be reversed if solutions with a different H+ con¬
centration are used. Such is actually the case with Cl2 :in acid solutions, the oxi¬
dation product of chlorine is un-ionized HCIO, and the formation of this substance
from Cl2 in acid solution (few OH“) takes place according to the equation—
Cl°2 + 20H- = 2HOC1 + 2©
The potential of this change is 0.37 volts, while the potential of the corresponding
change in alkaline solution (see above) is 1.57 volts. But the potential of the re¬
duction of chlorine to chloride—
Cl°2 + 2© = 2C1-
is evidently not dependent on the H+ or OH” concentration, and is still 0.65 volts.
The former equation here would have a place in our Table below that of the latter,
and hence these two reactions could not take together (See diagram-note on Table,
p. 691, and correspondingly, in acid solutions, free chlorine remains practically
unchanged. For further consideration of these particular relations, see section
headed below “Halogens.”
Definition of Normal Solution of an Oxidizer.—A normal solution of an oxi¬

dizer (or reducer) contains in one liter volume as much of this substance as
amounts to its formula-weight, in grams, divided by the number of free electrons
which jappear in the equation for its separate action. Thus, a normal solution
of KMnO* contains one-fifth of 158.04 g. per liter, because it reacts according to
the equation
(K+, Mn7+, 40—) + 8H+ + 5© - K+ + Mn++ + 4HaO
A normal solution of oxalic acid contains one-half of 126.06 g. per liter, because
it reacts according to the equation
(COOH)2 + O— = 2C02 + li20 + 2©
Differential Oxidation.—Since no reaction will take place unless the potential

of the change of the reducer is greater than the potential of the change of the
oxidizer, it follows that we may so choose the latter as to affect only one of several
reducers in a mixture. Thus, if it is desired to oxidize iodides only when bromides
also are present, we can accomplish this by using HNO2 because the potential of
n NIIa An onmno* fA Iawav avi ofo/vnci Zct okAn+ 1 O Tr/'O’fa wniln AT fhA
2Br~ = Br°2 + 20
change is 0.92 volt, and with this relation between the potentials there can be no
reaction. However, the potential of the change:
2I~ = I°2 + 2©
which is 1.46 volts, is so related to the potential of the above action of HNO2 as
to bring about a reaction. Hence, HN02 will change the iodide to free iodine,
but will not affect the bromide.
Another way to accomplish the same result is to use KMn04 with differently
acidified solutions, because the potential of the change
(K+, Mn7+, 40—) + 8H+ + 5© = K+ + Mn++ + 4H20
decreases with increase of H+ ion concentration: with such an H+ ion concentration

as is obtained in acetic acid plus sodium acetate, the potential of the permanganate is
great enough to react with iodides only, i.e., its potential is less than the potential of the
iodide-iodine change, 1.46 volts, but greater than the potential of the bromide-bromine
change, 0.92 volts; if the solution is moderately acidified with H2S04or HNO3 bromides
are oxidized, but chlorides only to a negligible extent (potential between 0.92 volts for
bromides and 0.65 volts for chlorides). But with large additions of H2SO4 or HNO3,
the oxidation of chlorides also is appreciable.
PART II
Details of Oxidation-Reduction Behavior of the Common Elements.—The

remainder of this section will present under each element the special facts shown
by them in oxidation-reduction reactions. The elements will be considered
in the order of their positions in the periodic system.
Sodium, Potassium, Magnesium, Calcium, Strontium, Barium, and
Aluminum.—These metals exhibit a great tendency to pass to the oxidized
state: they are strong reducing agents, usable at ordinary or room temperatures
(see Table, Nos. 1 to 5). They undergo only the one valence change illustrated
by the following equations:
Na° = Na+ + 1©
Mg° = Mg++ + 2©
Al° = A1+++ + 3©
Since the alkali metals—sodium and potassium—react violently with water,

they can be employed for the reduction of substances in aqueous solutions only
in the form of amalgams. Calcium, strontium, and barium also are best employed
in the form of amalgams, but magnesium and aluminum are employed direct—
usually in a finely divided form such as filings, etc. For the reduction of sub¬
stances in non-aqueous mixtures—such as solutions in alcohol, ether, benzene,
etc., the alkali metals also are employed directly, i.e., as such.
Aluminum is frequently used as a reducing agent in sodium hydroxide solutions
because aluminum ion (Al+++) in sodium hydroxide solutions unites with Na+ and
20—to form the compound NaAKh which is soluble. For an illustration of the use
of this reducing agent, see “Nitrogen” below.
Copper, Silver and Gold.—These metals have a low tendency to change to

their simple ions, i.e., the ions formed from solutions of their simple salts such as
chlorides, nitrates, etc. (see Table, Nos. 21 and 24: the corresponding place for
gold would be between Nos. 28 and 29). Hence they are very feeble reducing
agents, and are practically never used as such.
In solutions in which the ions of these metals form complex ions, i.e., in ammonia,
or potassium cyanide solutions, most of the simple ions of these metals are combined
with other ions and hence the reverse tendency of these metals to change to the
metallic form is lessened so much as to give these metals a much greater reducing
power (or a higher position in the table). However, no special use is made of this
fact.
In practice, a few examples of monovalent copper ion, Cu+, are met with. Thus
cuprous chloride is used in gas analysis as an absorbent for CO. It is produced by
boiling cupric chloride with metallic copper, and in this reaction lCu++is reduced to
Cu+ while 1 atom of copper metal is oxidized to Cu+, according to the equations:
Cu++ + 1© =Cu+
Cu° = Cu+ + 1©
On adding these equations and inserting the inactive Cl'", we obtain:
CuCl2 + Cu° = 2CuCl
Since cuprous chloride is but slightly soluble in water, it is made and employed in a
solution containing much hydrochloric acid. It is also soluble in ammonia. In
both cases, the increased solubility is due to the extensive union of Cu+ with other
ions to form complex ions (e.g., NH3Cu+ and CuCl2” respectively).
In cyanide solutions, copper is also present in the cuprous form. Its reduction
from cupric to cuprous ion takes place by the oxidation of the cyanide ion, and since
the whole action is met in the treatment of a cupric salt with KCN, the equations for
the whole change are here given:
(1) CuS04 + 2KCN = Cu(CN)2 + K2S04

(2) 2 Cu(CN)2 = 2CuCN + (CN)2
(3) KCN + CuCN = KCu(CN)2
Only (2) is an oxidation-reduction action. To understand it we need to consider

that, on ionizing, Cu(CN)2 gives Cu++ -f- 2(CN)” and that 1(CN)“gives up its nega¬
tive charge to Cu++ reducing the latter to Cu+.
Zinc, Cadmium, Mercury.—This group includes the most commonly used

reducing agent—zinc. It is used with either dilute sulphuric acid or sodium-
hydroxide solution, but the latter are not for the purpose of reacting with it to
evolve hydrogen: they serve to prevent the formation of a layer of Zn(0H)2
upon the surface of the metal if the action of the reduced substance uses up hydro¬
gen ion. For illustration, see under “Nitrogen,” “the reduction of nitrates to
ammonia.” When hydrogen is not used up by the reduced substance, then an
acid or sodium hydroxide is not needed—as shown in the following equations for
the reduction of ferric compounds to ferrous. In this action we have the zinc
changing as follows:
Zn° = Zn++ + 2©
while the ferric ion changes thus:
Fe+++ + I© = Fe++
Evidently, one Zn produces enough electrons for two ferric ions, hence they change
in the ratio Zn:2 Fe+++. Then the sum of the two simultaneous changes is expressed
by:
Zn° + 2Fe+++ = ZN++ + 2Fe++
Adding the inactive anions (e.g., S04 ) we obtain
Zn° + Fe2(S04)s = ZnS04 + 2 FeS04
Cadmium is employed in place of zinc if a less energetic reducing agent is desired.

Its action is just like that of zinc, except that it cannot be used in alkaline solutions
, because Cd(OH)2 docs not react to form a soluble compound with NaOH as Zn(OH)2
does.
Mercury exerts a very small tendency to be oxidized (see Table, No. 23) hence
it is never used as a reducing agent. Since it is easily reduced from its ion form
(Hg+ or Hg++) its compounds are used occasionally as oxidizing agents. Thus in
purifying mercury, metals such as zinc, iron, etc., are removed by treating the impure
mercury with a solution of mercurous nitrate, which converts the zinc or iron to ions
while an equivalent amount of Hg+ is reduced from the solution, as shown by the
following equations:
Zn° = Zn++ + 2Q
2Hg+ + 2© = 2Hg
Adding these equations, and inserting the inactive N03~ ions, we obtain
Zn° +2IIgNOs =Zn(N03)2 + 2Hg
Tin and Lead.—These metals have practically the same electromotive tend¬
ency for the change from the metal form to the bivalent ion (see Table, Nos. 15 and
16). They are milder reducing agents than cadmium. On account of the prox¬
imity of the potentials of lead and tin, lead is used when antimony is to be reduced
to the metal form out of solutions containing both tin and antimony: the potential
of antimony is much less than that of lead (see Table, No. 18), while that of tin is
practically equal to that of lead—hence antimony is entirely reduced by the
action of lead while practically no tin is reduced. The equations for the action
are:
3Pb° = 3Pb++ + 6©
2Sb+++ + 6© = 2Sb°
adding these equations, and inserting the inactive anions (Gl“), we obtain:
3Pb° + 2SbCl3 - SPbCh + 2Sb°
Both tin and lead exist also in the tetravalent form, and the potentials of the
changes from bivalent to tetravalent ions are much less than those of the changes
from metals to bivalent ions (see Table, Nos. 19 and 35). Hence the action of even
mild reducing agents such as metallic lead or tin serves to reduce tetravalent tin to
the bivalent form, as shown by the equation
Sn++++ + Sn° - 2 Sn++
Here tin furnishes both the oxidizing and the reducing agent.
Stannous chloride solution acidified with hydrochloric acid is frequently employed
as a reducing agent. It changes in accordance with the equation
Sn++ - Sn4+ + 2©
or SnClj + 2 Cl" - SnCl4 + 2©
and the potential of this change is given in the Table, line 19: from the latter it is
evident that any of the substances in the right hand column and below line 19 will
react with it. Thus it is seen that it will reduce mercurous compounds to mercury
and hence it will reduce mercuric compounds all the more readily—either to the
mercurous stage or to free mercury. The corresponding equations follow:
SnCl2 + 2 Cl” = SnCl4 + 20

2HgCl2 + 20-2 HgCl + 2 Cl”
Adding, we obtain
SnCl2 + 2HgCl2 - SnCl4 + 2HgCl
If the Hg++ is changed to Hg°, the equation for the change is evidently
HgCl2 + 2© = Hg° + 2C1”
and this combined with the above equation for the change of Sn++ gives
SnCl2 + HgCl2 = SnCl4 + Hg°
The reduction of ferric salts to ferrous salts is frequently accomplished by means

of stannous chloride. Since the mixture will contain a large amount of Cl” we may
consider the iron salts to be chlorides, and hence the reaction for the change from fer¬
rous to ferric is
(Fe+++, 3C1”) + 1© = (Fe+++, 2C1”) + Cl”
To combine this with the equation for the change of stannous chloride above, we must
multiply the last equation by two; then on adding the resulting equation to the one
for the stannous chloride, we obtain
2FeCl3 + SnCl2 = 2FeCl2 + SnCl4
Stannous oxide is soluble in caustic soda with the formation of the compound
Na2Sn02, sodium stannite. The corresponding stannic compound, Na2Sn03, also
exists, and is a fairly stable compound. The change from stannite to stannate natur¬
ally takes place according to the equation
Na2Sn02 + 2 OH” — Na2Sn03 + H20 + 2©
This solution of sodium stannite is a fairly strong reducing agent, as shown by

the fact that it reduces copper hydroxide, bismuth hydroxide, etc., to the free metals.
On boiling the solution reacts to form tin and Na2Sn03.
In other words, one molecule out of two of the stannite acts as the reducing agent
while the other acts as the oxidizing agent. The general relation of the potentials
which makes this possible was given in Part I, last article. The equations for the
reduction of the tin are:
Na2Sn02 + 2H20 + 2© = 2Na+ + (2H20 + 20””) + Sn°

= 2NaOH + 20H” + Sn°
The equation for the reduction of Bi(OH)s is
Bi(OH)s + 3© = Bi° + 30H”
The combining of either one of the two last equations with the preceding equation
for the oxidation of Na2Sn02 to Na2Sn03 is a simple matter and hence need not be
given here.
The most important oxidation stages of lead and their behavior may be observed
in the storage battery.
The lead storage battery has for its chemically active substance spongy lead at
the negative pole and a brownish-black powder, Pb02, at the other pole. These
materials are held by a “skeleton” of lead which remains unchanged except on the
surface. The liquid in which the poles are immersed is dilute sulphuric acid. The
change at the negative pole, when the cell is furnishing current is
Pb° = Pb++ + 2©
The Pb++ immediately combines with S04 because PbS04 is practically insoluble;
hence the whole change is
Pb° + SO 4“ = PbS04 + 2©
The change at the positive pole is
(Pb4+, 20-") + 2© = Pb++ + 20
The free oxygen ions, as soon as formed, will combine with H+ present in the elec¬
trolyte; and the Pb++ will unite with S04 ; hence the whole change at the positive
pole is
(Pb4+, 20—) + 4H+ + SO4— -f 2© = PbS04 + 2H20
The change for the cell as a whole is obtained by adding these equations for the
two pole changes:
Pb° + Pb02 + 4H+ + 2S04 - 2PbS04 + 2H20
It is evident that when the cell is furnishing current, lead at the negative pole is
oxidized, and lead peroxide at the positive pole is reduced.
When the cell is recharged, the above changes take place in the reverse sense.
Lead peroxide in dilute sulphuric acid is a strong oxidizing agent (see its position
in the Table, No. 35, and note that the substances in the lower part of the right hand
column are all strong oxidizing agents). The oxidizing tendency of Pb02 is shown in
the fact that it gives up oxygen gas on heating and changes to PbO.
Carbon.—It is not intended to discuss here any representative set of the

many compounds of carbon which take part in oxidation-reduction reactions,
nor even a partial list of those which are reducing agents—although the latter
are very important, particularly in photography, but to consider merely the
equations of the oxidation of a few compounds in order to show how such equa¬
tions are to be arrived at.
In order to estimate the number of free electrons liberated when a carbon compound
changes from one to another state of oxidation, we use the second method, in Part I,
for the estimation of the number of electrons required for polyatomic substances,
that is, we note
(а) That the original and the final compounds are usually not ionized—that is, they
are electrically neutral.
(б) That ionic hydrogen (H+) and ionic oxygen (0 or 0H“) are used in the
reaction—i.e., when oxidizing agents other than free oxygen gas are used—and that
the algebraic sums of their ionic charges are obtained as free electrons because all
other substances are electrically neutral.
Hence, to figure out the equation for the oxidation of a carbon compound by an
oxidizing agent other than free oxygen, add as many O and H+ as may be necessary
to form the final products, and add to the right side as many free © as is necessary
to balance the algebraic sums of the ionic charges on the two sides of the equations.
Thus, when ethyl alcohol is oxidized to acetaldehyde, the equation for the change is
45
CH3CH2OH + 0- - CH3CHO + H20 + 2©

or
CH3CH2OH - CH3CHO + 2H+ + 2©
For the oxidation of aldehyde to acetic acid, the equation thus obtained is
CH3CHO + O— = CH3COOH + 2©
or
CHaCHO + (2H+, 0—) = CHaCOOH + 2H+ + 20
For the oxidation of oxalic acid, the equation is
(COOH)2 + 0— = 2C02 + H20 + 2©
or
(COOH)2 = 2 C02 + 2H+ + 2©
For the oxidation of formic acid, the equation is
HCOOH + O = C02 + H20 + 2©
or
HCOOH = C02 + 2H+ + 2©
For the oxidation of formalin to C02 and H20, which occurs in the disinfection
of rooms in which some formalin is vaporized by the heat of reaction between another
portion of formalin and permanganate, the equation is
H2CO + 20 = C02 + H20 + 2©
The permanganate used for this purpose is reduced to Mn02, hence, the equation
for its reduction is
(K+, Mn7+, 40—) + 3© = K+ -f (Mn4+, 20~) + 20“
To combine the last two equations, we must multiply the last by 2 and preceding by 3:
adding these equations, and noting that KHC03 will be the most stable combination
in which K+ will remain in this mixture, and that this is formed by the simple union
of K+ + H+ + O -f C02, we obtain
KMn04 + H2CO = KHCOa + Mn02
Since the oxidation of ethyl alcohol to aldehyde is usually accomplished by means

of acidified bichromate, we will derive the complete equation for this change. The
equation for the change of the bichromate is derived under chromium, and is here
merely stated
(2Na+, 2Gr6+, 70“) + 6© + 14H+ = 2Cr+++ + 2Na+ + 7H20

or
(2Na+ 2Cr6+, 70~) + 6© + 7 H2S04 = Cr2(S04)3 + Na2S04 + 7H20 + 3S04™
To combine the last with the equation for the change of ethyl alcohol to aldehyde
given above, we multiply the latter by three in order to make the number of electrons
formed equal to the number used up by one molecule of bichromate. Then adding
the resulting equation to that for the bichromate, and eliminating 3H2S04 from
both sides, we obtain
3CH3CH2OH + Na2Cr207 + 4H2S04 = 3CH3CHO + Na2S04 + Cr2(S04)3 + 7H20
These examples may suffice to show how such equations are derived. It might be
added that, if the oxidizing is done by gaseous oxygen—as in the burning of methane—
no special device is needed to derive the equation, and the amount of oxidizing done
is expressed directly by the number of oxygen atoms used up.
Another type of reaction which may be of special interest is the formation of

chloroform from ethyl alcohol and hypochlorite of lime. Concerning the action of
the latter, see below under “Halogens.” The change of the ethyl alcohol results in
the formation of the formic acid ion and of chloroform, which takes place according
to the following equation:
CH3CH2OH + 3C1- + 30— = CHC13 + HCOO“ 4- 2H20 + 8©

The change in the hypochlorite is
(Ca++, 20— 201+) + 40 = Ca++ Cl2“ + 20““
Multiplying the latter by 4, the former by 2, adding, eliminating 60 and 6C1”

from both sides, we obtain
2CH3CH2OH + 4Ca(0Cl)2 = 2CHC1, + Ca(COOH)2 + 4H20 + 2CaO + CaCl2.
The 2CaO will naturally combine with 2H20 to form 2Ca(OH)2. ’
Nitrogen.—This element furnishes at once the two most important and

generally used oxidizing agents, HNO3 and HNO2, and also two of the strongest
reducing agents—hydroxylamine and hydrazine. We shall consider the action
of nitric acid first.
The nitrogen in nitric acid and in nitrates is in the highest state of oxidation in
which it occurs; and the nitrogen in ammonia or ammonium compounds is in the
lowest state of oxidation, it cannot be reduced further. Between these two limits
there are a number of oxidation stages, the order and relation of which is shown in the
following table. Note that the free element occupies an intermediate position.
Table of Nitrogen Compounds in The Order of their Extent of Reduction

from Nitric Acid
Amount of re¬
Elemental ions
duction from
Compounds from theoretical
HNO3; electrons
ionization
per 1 N
i
1. HN03, nitric acid (and nitrates). H+,N5+, 30— 0

2. N02, nitric peroxide. N4+, 20— 1
3. HN02, nitrous acid (and nitrites). H+N*+20— 2
4. NO, nitric oxide. N++ 0— 3
5. N20, nitrous oxide. 2N+ 0— 4
6. N, nitrogen. N° 5
7. NH2OH, hydroxylamine (or its salts) N" 6
8. NH3(or NH4OH), ammonia (or its salts)— N-, 3H+ 8
From the position of nitric acid in the Table of Oxidation Changes, it is seen that
nitric acid reacts with all substances in the left column above line 29. Technically, we
say: all these substances in the left column from 29 up are oxidized by nitric acid. This
includes all metals except gold, platinum, rhodium, and iridium.
The potential of the change of nitric acid fco lower oxidation stages varies with the
free H+ concentration—from 1.0 volt for normal H+ concentration to a value near
0.72 volt for large H+ concentrations. The lowest potential obtained by measure¬
ment with concentrated HN03 is 0.72 volt.
Nitric acid is a slowly acting substance—it acts as though it had to overcome a
reaction-hindrance. This coupled with the fact that its potential changes almost
directly with its H+ concentration, irrespective of the reduction product formed,

accounts for the fact that very dilute nitric acid frequently does not produce any
oxidizing effects whatever.
When acting as an oxidizing agent, nitric acid is reduced to its various products
named in the table above. It is seldom that, in any case, only one particular reduc¬
tion product is formed; however, under particular conditions some particular product
is mainly formed. The following general rule connects these conditions with the.
chief product formed in each case: The extent to which the nitrogen alone is reduced
increases with the degree of dilution of the nitric acid and with the force of the reducing
agent. For example:
cone. HN03 + copper gives NO*

dil. HN03 + copper gives NO
dil. HNO3 4- zinc gives N
very dil. HNO3 + zinc gives Nil3
The above rule and these examples serve as a means of predicting the extent of
reduction of nitric acid in other mixtures.
The reactions for the formation of seven different reduction products of nitric acid
are expressed by the following equations:
(a) (H+, N5+, 30 ) + 1© + 1 extra H+ = NO* + 11*0

(b) (H+, N*+, 30 ) +2© + 2 extra H+ - UNO* + H*0
(c) (H+, N5+, 30 ) +30 + 3 extra H+ = NO + 211*0
(d) 2(H+, N*+, 30—) + 2X4©+ 2X4 extra H+ - N*0 + 511*0
(fi) 2(H+, N6+, 30—) + 2X5©+ 2X5 extra 11+ = N°a + 511*0
(/) (H+, N5+, 30—) +60 + 6 extra H+ « (N~, 3H+, 0“ 1 + 2H*0
(g) (H+, N5+, 30—) +8©+8 extra H+ - NIT* + 311*0
The equations for the reactions between nitric acid and various reducing agents
will now be presented together with experiments in which these reactions take place.
(a) When a glowing piece of charcoal is dropped into some warm fuming nitric acid,
the charcoal burns vigorously in the acid.
This shows that, at high temperatures, nitric acid dissociates so as to form free
oxygen. This it can do only by forming N-compounds in which N has a lesser valence
than 5, as shown for instance, by the equation
2HNOs = Ha0 + 2N0a +0

n—valence N — valence
Since the N in HN03 actually does this valence changing at high temperatures, it is
to be expected that it exerts this tendency even at ordinary temperatures, its assumed
throughout this discussion.
The Formation of Nitric Peroxide from Nitric Acid.-—When copper reacts

with concentrated nitric acid, reddish-brown fumes of NO* will be evolved.
This reaction between copper and concentrated nitric acid is evidently a combina¬
tion of equation (a), above—
(H+ N6+, 30—) + 10 + 1 extra II+ = NO* + 11*0

and of the electromotive change of copper,
Cu° = Cu++ + 2©
In order that the electrons formed from Cu° will be taken up wholly by the
HNO3, 2HNO3 must change while 1 Cu° changes; hence we multiply the first
equation by 2 and obtain,
2IINO3 + 2© + 2H+ = 2N02 + 2H20
Then, taking the second equation,

Cu° = Cu++ + 2©
adding the left sides and the right sides, cancelling the electrons, and adding the
extra NO3 ions, we obtain the common form of the equation:
2HNO3 + lCu° + 2TINO3 = 2N02 + 2H20 + 1Cu(N03)2
The Formation of Nitric Oxide from Nitric Acid. When copper reacts with
ordinary dilute nitric, a colorless gas, nitric oxide, NO—will be evolved. The
reaction between dilute nitric acid and copper evidently consists of equation
(c), and the equation for the change of copper. Here, HN03 and Cu must
change in the ratio
2HNO3 : 3Cu°
in order that the electrons formed from the Cu° may all be used up by the HN03.
Multiplying the corresponding equations by 2 and 3, we obtain
2IIN03 + G© + 6li = 2NO + 4H20

3Cu° = 3Cu++ + 6©
Aiding, cancelling the electrons, and adding the “inactive” N03 ions, we
obtain
2IINO3 + 3Cu° + 6TINO3 - 2NO + 4H20 + 3Cu(N03)2
The Reduction of Nitrates to Ammonia.—Very strong reducing agents (zinc

or aluminum) reduce very dilute nitric acid or nitrates to the extreme limit—
ammonia. This is conveniently accomplished as follows: secure in a test-tube
some finely divided aluminum, cover it with concentrated solution of caustic
soda, and add a few drops of dilute nitric acid: it will take a few minutes until
vigorous action sets in, then the presence of ammonia will be noticeable.
The equation for this reaction is the sum of equation (g) above, and of the equation
for the electromotive change of aluminum. From the numbers of electrons involved
in these changes, it is evident that 3HN03 change while 8A1 change. Hence we
multiply the first equation by 3,
3IINO3 + 3X8©+ 3X8 extra H+ = 3NHa + 9H20
and the second by 8, thus:

8A1 = 8A1+++ + 24©
On adding these, we obtain
3HNOj + 8A1 + 24 extra H4* - 3NHa + 9H20 + 8A1+++
To get the final equation, we must add extra ions to change 8A1+++ to 8NaA102.
This requires 8Na+ ions, and 1GO— ions. These and also the 24 extra H+ ions in
the equation above will be supplied if we add 8NaOH and 8H20. Hence adding
8NaOH and 8H20 in place of the 24 extra H+, and putting 8N*aA102 in place of
8A1+++, we obtain the final form
3HNOa + 8A1 + 8NaOH + 8H20 = 3NHa + 8NaA102 + 9H20
In its simplest form, this equation is evidently,

3NaNOs -1- 8A1 + 5 NaOH + 2H20 = 3NH3 + 8NaA102
Other Examples of the Oxidizing Actions of Nitric Acid.—(a) Action upon

insoluble sulphides—e.g., CuS, Bi2S3j Ag2S. These sulphides generally change
to the ordinary cations and free sulphur, hence their oxidation is expressed by
the following equations:
CuS = Cu++ + 2© + S°
Bi2S3= 2Bi+++ + 6© + 3S°
Ag2S = 2Ag+ 2© S°
It is useless to try to decide whether both components of the sulphide have a

zero valence, and hence the metal is oxidized, or that the metals have their posi¬
tive valences, and the sulphur its negative valence (e.g., Cu,+ f, S ) and that
the latter is changed from 2“ to 0: either idea may be used.
The reduction product of nitric acid acting upon metal sulphides depends upon
the concentration of the acid: with ordinary dilute acid, NO is formed; hence the
change of the nitric acid is
(H+, N+6, 30—) + 3© + 3H+ .= NO + 2H20
Hence, for CuS, 2HN03 react with 3CuS because 3CuS produce 6© and this is
enough for 2HN03. Adding the corresponding equations—multiplied by 3 and 2
respectively—and inserting the inactive N03~ ions, we obtain
3 CuS + 2HN03 + 6HNO3 = 3Cu(N03)2 + 3S + 2NO + 4H20

With Bi2S3, we obtain in the same manner
Bi2S3 + 2HN03 + 6HN03 = 2Bi(N03)3 + 3S + 2NO + 4H20
With Ag2S, we obtain
3Ag2S + 2HN0S + 6HN03 - 6AgN03 + 3S + 2NO + 4H20.
(b) Action of “Aqua Begia.”—A mixture of one volume concentrated nitric

with three volumes of concentrated hydrochloric acid is employed to dissolve metals
(gold, platinum) and to act upon metal sulphides which are not acted upon by nitric
acid (HgS). In this mixture, the nitric acid is relatively dilute—diluted by the hydro¬
chloric acid—but the high H+ concentration in the mixture helps the nitric acid to
exert a high oxidation potential. A part of the oxidizer is always wasted in the
formation of free chlorine, but this action is not rapid because the chlorine is held for
a while in the form of several unstable compounds such as nitrosyl chloride (NOC1),
hypochlorous acid (HOC1). However, in the end all the chlorine produced by the
“wasted” oxidizing action appears as CI21 and practically all the nitrogen escapes as
NO. Hence chlorine is formed according to the following equations:
(H+, N6+, 30—) + 3© + 3H+ = NO + 2H20
2(H+, Cl") = Cl° + 2H+ + 2©
To make the number of electrons formed by the change of HC1 equal to the number
used up by HN03, we must multiply the first equation by 2 and the last by 3. Adding
the resulting equations, and cancelling the 6H+ and 6© from both sides, we obtain
2HN03 + 6HC1 = 3C1* + 2N0 + 4H20

1 HNOa can oxidize chloride ion only when the latter is sufficiently concentrated to raise the potential
of its change above that of HNOs in our Table, q.v., Nos. 29 and 31.
The portion of the oxidizer (HN03) that spends itself upon another substance
put into the aqua regia (HgS, Au, Pt, etc.) naturally does not form free chlorine as an
end result, and hence is best considered as acting directly with the substance treated.
Since the oxidizing power of the nitric acid is the only one acting in aqua regia, it
follows that the oxidizing power of this mixture cannot be (is not) greater than that
of the nitric acid; and hence the question arises: why is nitric acid alone unable to act
upon such substances as HgS, Au, or Pt? The answer is: the chlorides of these metals
are only slightly ionized, and very few free Hg++ or Au+++, or pt++++ ions remain to
oppose the formation of more of the latter; whereas, in the absence of Cl“ion, the first
(even small) portions formed, of Hg++, etc., remain as free ions, and stop further
action.
Hence the portion of aqua regia which acts direct upon HgS acts in accordance
with the following equations (derived by means of method 2 for writing such equations
—see Part I):
HgS + 2HC1 = HgCl2 + S° 4- 2H+ 4- 2©
HNOa + 3H+ -f- 3© = NO + 2H20
Multiplying the first by 3 and the second by 2, adding the resulting equations, and
simplifying, we obtain
3HgS + 6HC1 + 2HN03 - 3HgCl2 + 3S° + 2NO + 4H20
The action upon gold takes place according to the equations
,Au° + 3 Cl” = AuC13 + 3©
PINO8 + 3H+ -f- 3© = NO + 2H20
Adding, we obtain
Au + 3HC1 + HN03 = AuC13 + NO + 2H20
(c) The action of concentrated nitric acid upon some non-metals such as sulphur,
phosphorus, etc., results in the formation of the highest oxidation product of these
non-metals and of the reduction of HN03 to N02. Thus, with sulphur the equations
are
(H+, N«+, 30”") + 1© + H+ = N02 + PI20
S° + 4(0”” + 2H+) = (2H+, S6+, 40””) + 6© + 6H+
Multiplying the first by 6, and adding the result to the last equation, we obtain—
after cancelling the 6©, 4H20 and 6H+ from both sides
6HN0S + S° - 6N02 + H2S04 -f 2H20
Nitrous Acid and Nitrites.—Although the nitrogen in nitrites is' in a lower

state of oxidation than that in nitrates, and their potential as oxidizing agents
less than that of nitrates, yet they react more readily than nitrates, and hence
they appear to be stronger oxidizing agents. However, they are really weaker.
The potential of the change from HN02 to lower reduction products is 1.2 to
1.3 volts or HN02 is 0.5 to 0.3 volt weaker as an oxidizing agent than HN03,.
Reduction products formed from nitrous acid or nitrites are practically identical
with those of nitric acid, and hence the equations for the reduction of nitrous acid
need not be presented.
Strong oxidizing agents (acidified permanganate), oxidize nitrous acid to nitric
acid. The action is quantitative and hence is employed to estimate the amount of
nitrites in a sample. The equation for this change is obtained as follows: nitrogen in
HN02 has 3+ charges,—that in HN03, 5+, and the latter also contains one O more,
which is obtained from the solution in the form of an ion, O , from water. Hence this
change is expressed as follows:
(H+, N3+, 20—) + (2H+, 0—) = (H+ + N6+, 30“) + 2© + 2H+
The change undergone by the permanganate is explained under the corresponding

heading below, and given here without comment:
2(K+, Mn7+, 40—) + 10© + 8(2H+, 804~) = K2S04 + 2MnS04 + 8H20 + 5SQ —
To make the number of electrons produced by the change of the nitrous acid equal to
that taken up by the permanganate, we must multiply the first equation by 5; then
adding them we obtain:
5HN02 + 2KMn04 + 5H20 + 8H2S04 = 5HN03

+ K2S04 + 2MnS04 + 8H20 + 5H2S04
This equation can be simplified by taking out 5H20 and 5H2S04 from both sides.
Hydroxylamine.—This is a very active reducing agent of the composition

NHaOH. It forms salts with acids just as ammonia does, but it is a much
weaker base than ammonia. Hydroxylamine is produced by the reduction of
nitric acid: when the reduction is not properly regulated, other reduction prod¬
ucts—notably NH3—are formed. The reduction is most easily effected electro-
lytieally: with proper current densities and a mercury cathode, nitric acid in
dilute H2S04 is reduced almost quantitatively to hydroxylamine sulphate.
The equation for this reduction is:
(H+, N5+, 30—) + 6© + 6H+ = (N-, 3H+, 0—) + 2Ha0
On evaporation, the cathode liquid will give crystals of hydroxylamine sulphate

2NH2OH.H2SO4.
Hydroxylamine is a splendid reducing agent. According to the vigor of the

oxidizing agent with which it reacts, its nitrogen will be oxidized to any of the following
stages: Ng, N+, N+++, etc. Thus with Fehling’s solution, it will change quantita¬
tively to N20. Fehling’s solution contains cupric ion, Cu++, and this is reduced to
cuprous ion, G\i+(i.e., Cu20), according to the equation:
2Cu++(0H)r + 2© = Cu20 + H20 + 20H"
The elemental ions of hydroxylamine are, (N“, 3H+, 0—),hence the nitrogen changes
from 1” to 1+ according to the equation:
2(N“, 3H+, O ) == N20 + 4© + H20 +4H+
To make the electrons produced equal to those used up, we must multiply the former
equation by 2. Then adding the result to the last equation above, we obtain:
4Cu(OH)2 + 2(NH2OH) - 2Cu20 + 7H20 + N20

When hydroxylamine salt is put into excess of acidified permanganate solution, the
nitrogen is oxidized almost quantitatively to N02, the action taking place according
to the following equations:
(N- 3H+, 0”) + (2H+, O—) = (N4+, 20“") + 5© + 5H+

(K+, Mn7+, 40—) + 5© + 8H+ = K+ + Mn++ + 4H20
Adding these equations, dropping 4H+ and one H20 from both sides, and inserting
the inactive anions, 2S04 on both sides, we obtain
(NH2OH) + KMn04 + 2H2S04 - N02 + KHS04 + MnS04 + 3H20

OXIDATION AND DEDUCTION 713
The extent to which hydroxylamine may be oxidized by other oxidizing agents

may be inferred by noting the position in the Table of such oxidizing agents, and recall¬
ing the above actions of acidified permanganate (line 32) and of cupric salts (line 21—
the potential of Cu++ changing to Cu+ being practically identical with that of Cu++
changing to Cu°).
Hydrazine.—This substance has the formula N2H4, and it forms salts of the
formula N2H4.HCI. Its elemental ions are evidently (2N”“, 4H+). It is a
strong reducing agent: a solution of the free base is oxidized by the oxygen of
the air even. It reduces silver oxide to silver, cupric to cuprous, etc., and even
free sulphur to H2S. Its oxidation product is almost always N2; and this action
naturally takes place according to the equation:
(2N—, 4H+) = No + 4H+ + 4©
Since the nitrogen in hydrazine is in a higher oxidation stage than that in ammonia
hydrazine can also be reduced to NH3, and even when it is oxidized by such a
strong oxidizing agent as acidified permanganate, a part of the hydrazine is
reduced to ammonia.
The fact that most—or all—of the “hydrazine nitrogen” is oxidized to N2 is
undoubtedly due to the fact that the two N’s in hydrazine are directly connected.
It is on account of the presence of the 2N nucleus that hydrazine is not formed from
nitric acid or other IN compounds by reduction, but is obtained by reducing hyponi-
trous acid, (HON)2.
Other compounds of nitrogen, such as hydrazoic acid, diazo compounds, etc.,
are not used ordinarily in oxidation-reduction reactions and hence need not be con-
considered here.
Ammonia and Its Salts.—Although these compounds contain nitrogen in the

lowest state of oxidation, they are not correspondingly active reducing agents:1
they are oxidized only by strong oxidizing agents, or under special conditions.
With free chlorine, the nitrogen in ammonium salts will change to free nitrogen
or N+++ (i.e., NCI3) according to the following equations:
2(N3", 4H+, Cl”) = N°2 + 8H+ + 2C1” + 6©

(N3“, 4H+, Cl") = N3+ + 4H+ + Cl” + 6©
The equation for the reduction of the chlorine is
Cl°2 + 2© = 2C1-
Multiplying the latter by 3, and adding the result to each of the preceding
equations in turn, we obtain
2NH4C1 + 3C12 = N2 + 8HC1

NH4CI + 3C12 - NC13 + 4HC1
With bromine, which is a weaker oxidizing agent than chlorine, the oxidation
of the nitrogen stops mainly at the zero valence—that is, free nitrogen is produced.
With iodine, the action on acidified solutions of ammonium salts is negligible. In
neutral or alkaline solutions, the tendency of ammonia to be oxidized is much greater
than in acid solutions because these reactions involve the formation of free H+ ion (see
equations above), and a lesser H+ ion concentration will increase the potential
1 Attention has been called to this in the general discussion in Part I and a reason given for this
anomaly.
of this reaction of the reducer. This more than offsets the corresponding change in
the potential of the oxidizing halogen. The latter changes from free halogen to hypo-
halite, and these are weaker oxidizing agents in alkaline solutions than the free
halogens from which they are formed (see under “ Halogens ”). However, the amount
by which the potential of the oxidizing agent is thus lessened is evidently more than
offset by the before-mentioned increase in the reaction tendency of the ammonia,
so that we find even iodine oxidizing free ammonia vigorously, to NI3. The latter
crystallizes with one molecule of NH3, which gives the compound the formula NH3'NI3.
Nitric acid does not oxidize ammonium salts appreciably at ordinary temperatures,
but the rapidity of the action increases with rise of temperature, and is appreciable
in boiling solutions. At slightly higher temperatures, mixtures of ammonium salts
with nitrates react rapidly. The nitrogen from both is changed to the N+ form,
i.eN2O is formed—according to the equation
(N3-, 4H+, Cl") + (Na+, N*+, 30—) = N20 + 2H20 + NaCL
Nitrites—which react more readily than nitrates—oxidize ammonium salts

rapidly even at temperatures slightly below 100°. Thus on boiling a solution con¬
taining NaN02 and NH4C1, the following reaction takes place:
(Na+, N3-, 20—) + (N3“, 4H+, Cl") = NaCl + N°2 + 2H20
At ordinary temperatures, acidified permanganate and bichromate solutions

do not oxidize ammonium salts extensively, but at higher temperatures, e.g., in hot
solutions, the action takes place quite rapidly and results in the formation of nitrogen.
This also takes place in alkaline solutions, although here these oxidizing agents have
necessarily a much smaLer tendency to be reduced: this is more than offset by the
larger tendency of ammonia to be oxidized.
At ordinary temperatures, ammonia gas and oxygen do not react; but at higher
temperatures, i.e., when they are heated from an external source of heat, ammonia
may be ignited and will burn to form nitrogen and water. When the hot gases come
in contact with platinum, nitric oxide is formed (principle of commercial production of
nitric acid from ammonia).
Phosphorus, Arsenic, Antimony, and Bismuth.—The total range of oxidation

stages of these elements is the same as that for nitrogen: it extends from the state
corresponding to nitric acid (E5+) to the stage corresponding to ammonia (E3”)
except in the case of bismuth in which the free element is the lowest oxidation
stage known.
In its highest oxidation stage, phosphorus exerts absolutely no oxidizing tendency,
in other words, phosphoric acid is not reducible in aqueous solutions. Concerning the
other metals of this group, it may be said that the oxidizing tendency—or reducibility
—of their highest oxidation stages increases with increasing atomic weight, so that
the highest oxide of bismuth is a fairly strong oxidizing agent.1 Correspondingly the
lower oxidation stages show less tendencies to pass to the higher stages—or, in other
words, the lower oxidation stages of phosphorus are stronger reducing agents than
those of arsenic, and those of arsenic less than those of antimony. There is no appre¬
ciable difference between bismuth and antimony, so that the- electromotive force
required for the electrolytic deposition of these metals out of solutions of their chlorides
are practically the same, and near that of copper (see Table, Nos. 18 and 21). None
of the compounds of these elements are used as oxidizing agents except the so-called
“bismuth peroxide.”1
1 The formula of such a compound is not definitely known beyond the fact that it contains bismuth
with a valence greater than three.
However, the reduction of Bi(OH)3 to black BiO by means of an alkaline solution

of Na2Sn02 serves as a delicate test for bismuth, hence its reaction is given here.
2Bi(OH)3 + 20 = 2BiO + 3H20 + 0~ Na2Sn02 + 0“ = Na2Sn02 + 20

or, together:
2Bi(OH)3 + Na2Sn02 = 2BiO + 3H20 + Na2Sn03.
Hypophosphites, phosphites, and arsenites are used occasionally as reducing

agents. Hypophosphites are strong reducing agents. They reduce concentrated
sulphuric acid to S02 or to free sulphur, cupric ion (e.g.} from CuS04) to copper hy¬
dride, silver ion (from AgN03) to silver, etc. They are derived from the free acid,
H3P02, the elemental ions of which are (3H+, P+, 20 ). Its salts are readily pre¬
pared by the action of yellow phosphorus on warm aqueous solutions of strong bases.
In this case a part of the phosphorus changes to phosphine (reduction!) while another
part changes to hypophosphite. The latter is a mono-basic acid ion of the formula
H POOH”. It contains phosphorus with a polar valence of 1+ because its elemental
ions are (2H+, P+, 20 ) the electric charges of which give an algebraic sum of 1”
and this belongs to the ion as a whole. Its formation evidently takes place according
to the equation
p -t- h2o o ~ (h2po2)- -f-
while the reduction to PH8 takes place according to the equation
P + 3H0H + 3© - PH3 + 30H-
The relation of the forces which brings about this reaction were discussed in the
latter part of Part I under the heading “ Substances Which Act Simultaneously as
Oxidizers and as Reducers.”
Phosphites are determined quantitatively by titration with iodine in neutral or in
acid solutions. Alkaline solutions cannot be used because in such solutions iodine
will change to iodide and hypoiodite, and the latter will rapidly change further—to
iodate (see below under “Halogens”) and thus the quantitative relation between
iodine and the other reacting substance would be disturbed. Iodine reacts with H3POi
as follows:
(3H+, P+++, 30 ) + H20 = H3PO4 + 2H+ + 2©
I°2 + 2© = 21-
Hence,
H3P03 + H20 + I2 = H3PO4 + 2HI
All oxidizing actions lower in our Table than the iodide-iodine reaction will
naturally also oxidize phosphites.
Arsenites and other trivalent arsenic compounds (AsC13, As203, etc.) also may be
determined quantitatively by oxidation with iodine.
However, the titration of arsenic trioxide with iodine is quantitative only when
carried out in faintly acid solutions. The reason why the solution must be faintly
acidified during the change is this: the oxidation of the arsenic involves H+ ion as is
shown by the following equation:
Na3As03 + HjO = Na8As04 + 2H+ + 2©
and hence its potential decreases with increase in H+ ion concentration. But the
change of the iodine does not involve—is independent of—the H+ ion concentration.
Hence with increasing H+ ion concentration, the difference between the two potentials
decreases and soon becomes zero, and then there is no force to impel the reaction.
With H8POs—which is a stronger reducing agent than arsenious acid—a moderate
concentration of H+ ion still leaves the potential of “phosphorous acid changing to
phosphoric acid ” large enough to act with iodine.
Arsenic trioxide is used frequently to reduce nitric acid to nitric peroxide, N02—
i.eto reduce HN03 to the least extent possible: hence this extremely weak reducing
agent is chosen. This reaction takes place according to the following equations:
(2As+++, 30“) + 6H++ 50— = 2(3H+, As6+, 40“) + 40

(4H+ N5+, 30—) + 4© + 4H+ = 4N02 + 4H20
On addition, we obtain
As203 + 4HN03 + 5H20 = 2H3As04 + 4N02 + 4H20.
The production of arsine (AsH3) and of stibine (SbH3) by reduction—from higher

oxidation stages of arsenic or antimony respectively—can be accomplished in acid
solutions by means of zinc, iron, etc. The equations for the change with zinc are
given here
Zn° = Zn++ + 2©
As+++Cl3“ + 6© + 3H+ = As3"H3+ + 3C1-
Multiplying the first by 3, inserting on both sides the SCI- to accompany the 3+H
in the second, and adding the resulting equations, we obtain
3Zn + As Cl 3 + 3HC1 = AsH3 + 3ZnCl2
Oxygen.—This element can exist in three distinct oxidation stages, which in

their order from lowest to highest stage, are: (1) oxygen as a constituent of all
ordinary oxides, 0—; (2) as a constituent of true peroxides, e.g., hydrogen per¬
oxide, H202, barium peroxide, Ba02, etc., in which the polar valence of oxygen is
1”; and (3) as the free element, 0°2 or 03 (ozone). Thus peroxides appear as
containing oxygen which has been reduced from the free element half-way to the
lowest possible state of oxidation, and this view of the state of the oxygen in
peroxides is in accord with its properties: it can act in either capacity—as an
oxidizer, changing itself to 0—, or as a reducer, changing itself to 02.
Mathematically we could consider peroxides to be compounds formed by the

addition of free oxygen to ordinary oxides, which at the time of action dissociate
into these constituents. But this view does not suggest the behavior of peroxides
as reducing agents, and hence is inferior to the other view.
In aqueous solution, the oxygen ion, 0 , exists as such only in extremely small
amounts; but it is always in equilibrium with its hydrated form, OH”, through the
equation
O— + H20 = 20H”
and the 0H~ concentration is known through the acidity or alkalinity of the solution.
Since the concentration of H20 in all ordinary aqueous solutions is practically the
same, the mass-law gives us the following relation between the materials in the above
equation “at equilibrium.”
0—] X [H20 Const.] = (Const.) X [OH*"]2

or
[0H“]2 =
that is, the ratio of the concentration of O to the square of the 0H~ concentration
is constant—or the variations in 0 concentration are proportional to the square
root of the variations in the OH” concentration. The OH” concentrations—in turn
—are known to vary as the reciprocal of the H+ concentrations, because H+ 4- OH”
= H20, and according to the mass law—[H+] X [0H~] = (Const.). With these
facts and the further fact that elementary oxygen has a slight, but roughly constant
solubility in all aqueous solutions, we can consider the tendency of the various forms
of oxygen to change to free oxygen, O2.
From the foregoing, it is evident that the tendency of O to change to 02 is great¬
est with a high concentration of OH", intermediate in neutral solutions, and lowest
in strongly acidified solutions. Number 26 in the table corresponds to the second,
i.e., neutral solutions. Or conversely, the tendency of 02 to change to O— (or OH”)
is greatest in acid solutions, and least in alkaline solutions. Hence, we may say
that oxygen gas will react with (oxidize) any of the substances in the left column of
the table above line 26, and that this will be particularly the case if the solution is
acidified (small OH+ concentration). With silver and mercury, the forces impelling
the combined changes (differences between the numbers in the central column)
are small, and hence the actions take place slowly; but with substances farther up in
the table, the actions are correspondingly more rapid. In all such cases, the actions
may be inferred from the table because the substances in the left-hand column change
to the substances on the same line in the right-hand column, while oxygen—which
is in the right-hand column on a line below—No. 26—changes to OH”—which either
remains as such or combines with H+ to form H2O. Some examples follow.
(а) An aqueous solution of hydrogen sulphide exposed to the air undergoes the
following changes:
2S— = 2S° + 4© (see table, line 11)
02 + 4© = 20
Adding and inserting the four H+ coupled with 2S , we obtain
2H2S + 02 = 2H20 + 2S
(б) An acidified solution of FeS04 exposed to the air undergoes the following
reaction:
4Fe++ = Fe+++ + 4©
4H+ + 0°2 + 4© = 2H20
Adding and inserting the inactive 6SO4 , we obtain

4FeS0 4 + 2H2SO4 + O2 = 2Fe2(S04)3 + 2H20
(c) A piece of copper projecting into a dilute acid will undergo the following
reactions, particularly at the points where it projects into the air:
2Cu° = 2Cu++ + 4©
4H+ + 02 + 4© = 2H20
Adding, and inserting the inactive anions of the acid, we obtain

2Cu + 4HC1 + 0°2 = 2CuCl2 + 2H20
Since the oxygen in peroxides is in a state intermediate between that of the free
element and of the bivalent anion 0 , it has two distinct tendencies to change,
namely, either to be oxidized to 02 or to be reduced to 0 . Either one of these can
take place in alkaline, or in acid solutions—the corresponding equations and poten¬
tials are here given for iV/lNaOH and AVIH2SO4 respectively.
Oxidation of the peroxide in alkaline solutions—potential, 2.02 volts
(2Na+, 20") + 20H” = 0°2 + 2© + 2NaOH
Under these conditions it is a reducing agent equal in effect to tin in neutral or acid
solutions.
Oxidation of peroxide in acid solutions—potential, 1.20 volts
(2H+, 20") = 0°2 + 2© + 2H+
Under these conditions it is only a feeble reducing agent, equal to ferrous salts changing
to ferric.
Reduction of the peroxide in alkaline solutions—potential, 1.16 volts
(2Na+, 20”) + 2© + 2H20 = 4(0H)” + 2Na+
Under these conditions it is feeble oxidizing agent.

Reduction of peroxide in acid solutions—potential 0.34 volt
(2H+, 20”) + 2© + 2H+ = 2H20
Under these conditions it is a fairly strong oxidizing agent.

Peroxide in alkaline solutions reduces silver oxide (Ag+) to silver metal—a
reaction which is to be expected from the potential relations. The equations for this
change are
(2Na+, 20") + 2(OH)” = 0°2 + 2© + 2NaOH
2Ag+, 0”” + 2© + H20 = 2Ag02 + 2(OH)"
Adding these and cancelling the 2© and the 2(0H)" common to both sides, we obtain:
Na202 + Ag20 + H20 = 02 + 2NaOH.-f 2Ag
In acidified solutions, H202 is oxidizable by strong oxidizing agents, such as acidified

permanganate, acidified bichromate, free chlorine, free bromine, and concentrated
nitric acid, or in other words, by the substances in the right column of the table
below line 24. Since its reaction with acidified permanganate is frequently employed
for the quantitative determination of H202, the equations for it will be given here.
The equation for the peroxide is as derived above, and that for the acidified perman¬
ganate is derived under “Manganese.”
5(2H+, 20”) = 50°2 + 10© + 10H+

2(K+, Mn7+, 40””) + 16H+ + 10© == 2K+ + 2Mn++ + 8H20
Adding, cancelling the 10H+ and 10© common to both sides, and adding 3SO4 to
both sides, we obtain
5H202 + 2KMn04 + 3H2S04 = 50°2 + K2S04 + 2MnS04 + 8H20
I11 concluding this section on oxygen, a few words should be said about ozone. This
is a very strong oxidizing agent, or very active form of oxygen, but in its action only
10 out of the molecule Os is reduced to O —the other two oxygen atoms are obtained
as 02. Ozone reacts quantitatively with H202 in accordance with the following equa¬
tions:
020° + 2© + H20 - 02 + 2(OH)"

(2H+, 20”) = 0°2 + 2© + 2H+
Adding these equations we obtain
O3 + H20 -1- H202 = 202 2H20
Sulphur.—This element, similar to nitrogen, exists in a number of oxidation

stages, the lowest of which is the sulphide stage, and the highest, the persulphate
stage, H2S208 or Na2S20g. These different stages and the relations between
them are shown in the following table.
Compounds of Sulphur in Ascending Order of Oxidation Stages
Theoretical Oxidation elec¬

Order Common form of each oxidation stage
elemental ions trons per S
1 Hydrogen Sulphide (and other Sulphides). 2H+, S 0

2 Sulphur. S° 2
3 Hyposulphurous Acid (H2S204). 2H+, 2S3+, 40 5
(Not Thio-sulphuric Acid.)
4 Sulphur Dioxide (Sulphurous Acid, Sul¬ S4+, 20— 6
phites) .'.
5 Sulphuric Acid (Sulphates). 2H+, S6+ 40— 8
6 Persulphuric Acid (Persulphates). 2H+, 2S7+, 80— 9
In addition to the oxidizing stages in this table there are a number of oxidation
stages such as thiosulphate and the various thionates, which are not formed in the
ordinary oxidation-reduction changes of sulphur compounds.
Persulphates are strong oxidizing agents, which change to sulphates on reduction.
Concentrated sulphuric is a strong oxidizing agent which gives different reduction
products according to the strength of the reducing agent. Sulphur dioxide acts as
an oxidizing agent to strong reducing agents, but it acts as a mild reducing agent to¬
ward oxidizing agents. All lower oxidation stages of sulphur act as reducing agents
only. The strongest of these is the hyposulphite. Sulphur and sulphides are only
feeble reducing agents—partly because they are slightly soluble in water and hence
can be in solution only in small concentrations.
Persulphuric acid is prepared by anodic oxidation of concentrated sulphuric acid—
the change at the anode takes place according to the equation
2(HS04)- = H2S208 + 2©
The action consists of the discharge of electrons from the HS04“ion, and the union
of two such radicals. It is a powerful oxidizing agent even in alkaline solutions.
Sulphuric acid is an oxidizing agent only in the concentrated form. Its oxidizing
power is then approximately the same as nitric acid, and like the latter is used to
dissolve silver, copper, etc. It does not attack gold or platinum. Weak reducing
agents, such as silver or copper, reduce it to the next lower oxidation stage—SO2.
Strong reducing agents such as zinc, iodides, etc., reduce it to a much lower stage
i.e.j sulphur, or H2S even.
The following equations express the reduction of H2S04to its reduction products:
(a) H2S04 + 2© + 2H+ = S02 + 2H20

(ibj H2S04 + 6© + 6H+ = S + 4H20
(c) H2S04 + 8© H- 8H+ « H2S + 4H20
The equation for the reaction between copper and sulphuric acid is obtained by
adding equation (a) above to the equation for the change of copper because 1 Cu
produces as many electrons, i.e., 2—as 1H2S04 uses up to form S02. The addition
of these equations gives
Cu + H2S04 + 2H+ = Cu++ + SO2 + 2H20

Adding the inactive S04 ions, we obtain
Cu 4- 2H2S04 = CuS04 + S02 + 2H20
46
To obtain the equation for the reaction between Zn and II2S04, add together
equation (c) above and
4Zn = 4Zn++ + 80
The fact that stronger reducing agents reduce such oxidizing agents as concen¬
trated sulphuric acid to a greater extent than weaker reducing agents is shown well
by the action between concentrated sulphuric acid and chlorides, bromides, and iodides
respectively It is evident from the position of the latter in the table of Oxidation
Changes that their tendency to form the free halogens increases in the order in which
they are here named—chlorides, bromides, iodides—and in accordance with this we
find that in the action of concentrated sulphuric acid with chlorides there is no
oxidation or reduction, with bromides there is formation of bromine and slight
reduction of the sulphuric acid (to SO2), while with iodides there is formation of iodine
and very extensive reduction of the sulphuric acid (to sulphur, or to hydrogen sul¬
phide even). With bromides, the reactions are
2NaBr - 2Na+ + Br°2 + 2©
Adding this to No. (a) above, and inserting ISO4 ion on both sides, we obtain
2NaBr + 2H2S04 = Na2S04 + S02 + 2H20 + Br2
With iodides, the reactions are

2NaI = 2Na+ + I2 + 2©
Multiplying this by 3 (or by 4) and adding No. (b) (or No. (c)) above, wc obtain
the final equation.
A mixture of sulphur and concentrated sulphuric acid will react to form SOa—
the S° being oxidized and the H2SO4 reduced correspondingly.
Sulphur dioxide and sulphites are oxidized quantitatively to sulphuric acid
(or sulphates) by means of iodine in neutral or acidified solutions. Alkaline solutions
cannot be employed because in these iodine reacts in another manner (to form iodide
and hypoiodite) and this reaction destroys the quantitative relation. In neutral
or acid solutions, the reactions are
Na2S03 + H20 = Na2S04 + 2H+ + 2©

I°2 + 2© = 21“
On addition we obtain
Na2S03 + H20 + I2° = Na2S04 + 2HI
Since the formation of HI would result in the liberation and loss of SO2, some
NaHC03 is added to the solution. Since, furthermore, the action between iodine
and sulphites induces a direct action of air (oxygen) on the sulphite, a trace of alcohol,
glycerin, or sugar is added to the mixture; this acts as a negative catalyzer upon the
latter reaction. All oxidizing changes which appear lower in our Table than the
“iodide-iodine reaction’7 will naturally oxidize S02 and its compounds.
Sulphur dioxide oxidizes hydrogen sulphide both in the gaseous phase, and in
aqueous solution. The sulphur from both forms will be changed to free sulphur.
This reaction has produced many sulphur deposits in the earth. In this action,
H2S is oxidized according to the reaction
H2S = S° + 2H+ + 2©
while S02 is reduced thus,
S02 + 4© = S° + 20““
and the combined action is,
2H2S + S02 = 3S° + 2H20
When this reaction takes place in aqueous solution, some pentathionic acid, HAOe,
is also formed, according to the equation
IOH2SO3 + 4H2S = 3H2S506 + 12H20
Hyposulphites (not thiosulphates!) are very strong reducing agents. The most
familiar of these is the zinc salt which is obtained by dissolving zinc in aqueous
sulphurous acid—it dissolves without the evolution of hydrogen, and forms a yellow
solution of ZnS204. The reactions are the following:
Zn° = Zn++ + 2©
2(2H+, S4+, 30“) + 2© = (S2O4)— + 2H20
Zn + 2H2SO3 = ZnS204 + 2H20
According to the strength of the oxidizing agent reacting with them, hyposulphites
change to sulphites or sulphates.
Free sulphur acts as an oxidizing agent in a number of examples in which its
addition to a compound corresponds to the addition of an atom of oxygen. Thus in
the preparation of “yellow” ammonium sulphide, sulphur is merely added to, and
readily taken up by, ammonium sulphide solution, according to the equation
(NH4)2S + rrS = (NH4)2SS*
The highest value that x can have is 4—which corresponds to the 4 oxygen atoms
in (NH4)2S04. It is logical to assume that the sulphur thus taken up is bivalent—
just as the corresponding oxygen is assumed to be bivalent. However, this sulphur
is so readily separated again from this compound that the latter acts as though it
contained free sulphur merely. Hence when stannous sulphide, antimony trisul¬
phide, or arsenious sulphide are treated with yellow ammonium sulphide, they are
changed to stannic sulphide, antimony pentasulphide, or arsenic sulphide respectively
—i.e., they are changed to a higher valence. Sulphur also combines directly with
sulphites in neutral solutions to form thiosulphate, according to the equation
Na2S03 + S = Na2S203
The sulphur seems to add itself in the same manner as the fourth atom of oxygen
required for the sulphate.
Thiosulphates have an appreciable tendency to dissociate into sulphite and sulphur,
i.e., the above reaction is reversible. Hence, when its solution is boiled with copper
filings, we obtain CuS and Na2S03. On adding a great deal of a strong acid to the
solution of a thiosulphate, we obtain free sulphur and sulphurous acid. In other
words, the free thiosulphurous acid dissociates entirely into these constituents. The
difference in the behavior of the salt and the acid is due to the fact that the salt
solution contains the thiosulphate mainly in the form of the free ion, S203 , which
contains less free energy than the un-ionized form. The free acid, H2S203, however,
is present mainly in the un-ionized form, and this is much less stable—more reactive—
contains more free energy.
Thiosulphates are moderately strong reducing agents. When oxidized by weak
oxidizing agents, such as iodine in neutral or moderately acidified solutions, they
react as follows:
2S203— = 2S40e + 2©
That is, they form tetrathionates (e.g., Na2S406). The potential of this change
is evidently not dependent on the concentration of H+ or 0H“ in the solution. Very
strong oxidizing agents, such as chlorine in neutral or acidified solutions oxidize both
the sulphite sulphur and the “free” sulphur to the sulphate stage, according to the
equations
S2O3— + 5H20 = 2S04 + 80 + 10H+
4Clo + 8© = 8C1"
Inserting the inactive sodium ions, and adding these two equations, we obtain
Na2S203 + 5H20 + 4C12 = Na2S04 + H2S04 + 8HC1
The above reaction shows that the change of thiosulphate to sulphate involves H*
ion, and hence the (reducing) potential of this change is least in acid solutions and
greatest in alkaline solutions. Hence only very strong oxidizing agents will oxidize
thiosulphate to sulphate in acid solutions; but in alkaline solutions even a weak oxi¬
dizing agent such as iodine (more especially hypoiodite) is able to produce this change,
according to the equations
S2O3— + IOOH- - 2S04 + 5H20 + 8©

4I°2 + 8© - 81-
Inserting 2Na+, and adding, we obtain
Na2S203 + lONaOH + 4I2 = 2Na2S04 + 5H20 + 8NaI
A comparison of these equations with the preceding set above will reveal that they
are identical except for the fact that in the acid solutions H20 must be used to secure
O— while in the alkaline solution OH- is used.
Thus it is plain that only weak oxidizing agents, such as iodine, acting upon
thiosulphate in neutral or acid solutions, will change the latter to tetrathionate;
this observation is of great importance in iodometric titrations.
Concerning the reducing actions of hydrogen sulphide, it is merely necessary to
say that the tendency of the change of the sulphide ion (S ) to free sulphur (S°)
varies with the concentration of the former. Accordingly, it appears several times
(at Nos. 8, 11, and 14) in our table of Oxidation Changes. Hence, sulphides will
react with an oxidizing agent in the right-hand column and below lines, 8, 11, or 14.
Furthermore, the sulphur may not only be changed to free sulphur, but to higher
oxidation stages in accordance with the strength of the oxidizing agent.
Sulphur reacts in alkaline solutions to form sulphides and sulphites essentially,
that is a part of the sulphur is reduced from S° to S , while another part is oxidized
from S° to S4+. The general relations involved in such reactions have been discussed
in Part I under the heading “Substances which Act Simultaneously as Oxidizers and
as Reducers.” This behavior of sulphur is not important and hence will not be con¬
sidered further. The sulphide and the sulphite react further with each other,
and they also take up free sulphur—to form a mixture of “yellow” sulphides and
thionates, etc.
Chromium.—This element occurs in the following oxidation stages:

Chromium metal, Cr° Chromic acid (chromates and bichro-
Chromous ion, Cr++ mates) Cr6+, 30
Chromic ion, Cr++i*
Only the last two stages, and the changes from one to the other need to be dis¬
cussed here because the other substances are not commonly used. The oxida¬
tion of the trivalent to the hexavalent form takes place in alkaline mixtures
according to the equation:
(2Cr+++, 30—) + 4Na+ + 10OH" = 2(2Na+, Cr6+, 40“-) + 5H20 + 60
It is evident that this change involves OH" ion: an increase in OH" ion helps
this change. The potential of this change is undoubtedly well above 1.6 volts
(as potentials are calculated in our Table). The oxidation is brought about
either in fused NaOH (or Na2C03), or in an aqueous alkaline solution (e.g., of
NaOH), and by means of mild oxidizing agents even. Thus even iodine (hypo-
iodite) with a potential of 1.41 volts will act. However for fused mixtures KN03
and KC103 and for aqueous alkaline solutions Na2C>2 are the most convenient
oxidizing agents because they leave no troublesome by-product. The reaction
between Na202 and Cr203 takes place according to the equations
Na202 + 2H20 + 2© = 2Na+ + 40H"

Cr203 + 4Na+ -j- 10OH" = 2Na2Cr04 5H20 -f- 6©
Multiplying the first by 3, adding, and simplifying the resulting equation, we

obtain
3Na202 + H20 + Cr203 = 2Na2Cr04 + 2NaOH
Strong oxidizing agents such as KMn04 and Pb02 oxidize trivalent chromium to
the chromic acid stage even in acid solutions, although the potential of this change in
acid solutions is 0.70 volt or less (see table). Chromic salts can also be oxidized
electrolytic ally (at the anode) in acid solutions, the equation for the change being
2Cr+++ + 7H20 = H2Cr207 + 12H+ + 6©
On account of this potential of the change between trivalent and hexavalent

chromium in acid solutions, chromates in acid solutions act as strong oxidizing agents
toward most common reducing agents, i.e., all in the left-hand column of our Table
above line 30. Note that this does not include chlorides. The equation for the
change is the same as that above for alkaline solutions except that H+ and H20 are
employed in place of H20 and OH" respectively (change from alkaline to
acid solution!), and that, furthermore, the bichromic acid H2Cr207 is present instead
of the chromic acid H2Cr04 (the latter dehydrates to form H2Cr207). With these
modifications, the equation appears thus:
(2Na+, 2Cr6+, 70"") + 14H+ + 7S04"“ + 6© =

7H20 + 2Cr2(S04)3 + Na2S04 + 3S04"~
We will give the equations for the action of this oxidizer with several reducers.
(a) KI: its reaction is
2KI = I°2 + 2K+ + 2©
Hence we must multiply the latter by 3 to make the number of electrons formed
equal to the number used up in the above equation of the oxidizer. Adding the two
equations, we obtain
6KI + Na2Cr207 + 7H2S04 = Cr2(S04)3 + 3K2S04 + Na2S04 + 7H20 + 3I2

(b) FeS04: its reaction is
Fe++; SO 4 = Fe+++ + SO4"" + 1©
To obtain 6©, we must multiply this equation by 6. Adding the result to the above
equation of the oxidizer, we obtain
6FeS04 + Na2Cr207 + 7H2S03 - 3Fe2(S04)3 + Cr2(S04)3 + Na2S04 + 7H,0
Other equations may be readily arrived at in the same manner.

Halogens.—In considering the ordinary oxidizing or reducing changes of the

halogens, we can exclude any consideration of fluorine because the only valence
change it undergoes is from 0 to 1" and since free fluorine is not capable of
existence in aqueous solutions (because it reacts thus: 2F2 + 2H20 = 2H2F2 +
02), it is not employed in ordinary operations.
The other halogens—chlorine, bromine, and iodine—in general exist in the follow¬
ing oxidation stages.
The Common Halogen Compounds in Descending Order of Oxidation Stages

(X stands for Cl, Br, or I)
Theoretical Reduction
Order Common form of each oxidation stage
elemental ion electrons per X
1 HXO4 Perhalate (e.g., perchlorate, KC104). H+, X7+, 40” 0

2 HXO3 Halate (e.g., chlorate, KC103). H+, X5+, 30— 2
3 HXO2 Halite (e.g., chlorite, KC102). H+, X3+ 20— 4
4 HXO, Hypohalite (e.g., hypochlorite, IvCl
0). H+, X+, 0— 6
5 Free halogen (e.g., Cl). X° 7
6 HX, Halide (e.g., chloride, KC1). H+, X“
8
A few exceptions should be mentioned: Chlorine exists in all the stages shown in
the table above, and furthermore in the form of the oxide, CIO2, chlorine dioxide, in
which chlorine is evidently tetravalent. Bromine does not exist in the halite stage,
nor in the perhalate stage (and no oxides of bromine appear to be capable of separate
existence). Iodine exists in all stages in the table above except the halite stage.
The most important, and the most puzzling, behaviors of the different halogen
compounds are those of the free halogens and of the hypohalites—both respectively
in acid and in alkaline solution. Hence we will consider these first.
When a halogen is dissolved in pure water, a small amount reacts with water—
hydrolyzes—in accordance with the following equation, using Cl for illustration:
Cl°2 + H20 H+ + Cl~ + HOC1
This is a reversible reaction and since the reaction involves H+ ion, the equilibrium
necessarily is dependent, among other factors, upon the H+ ion concentration. That
is, the reaction stops when the II+ ion concentration has reached a value which
together with the corresponding concentrations of the other substances stops the
reaction.
Hence, when a base is added to this solution and thus the H+ ion concentration
decreased, more of the free halogen will be able to react to form chloride ion—and
HOC1 (or a salt of the latter), and accordingly we find that, in alkaline solutions, free
halogen changes almost completely to chloride and hypochlorite according to the
equation
Cl°2 + 2NaOH = NaCl + NaOCl + H20
We note that 1C1 has changed from Cl° to Cl-, i.e.} has been reduced—while the
other Cl has changed from Cl° to Cl+, i.e., has been oxidized. This phenomenon of
the same element acting both as oxidizing and as reducing agent has been discussed
in Part I under the heading ‘‘Substances which Act Simultaneously as Oxidizers
and. as Reducers,” and it has been shown there that in acid solutions the potential of
the change
2C1- -♦<- CC + 2Q
is greater numerically than the potential of the change
Cl°2 + 2H20 2HOC1 + 2H+ + 2©
The following table presents the corresponding values for all the halogens. The
numerical values are on the basis of the hydrogen pole potential being 2 volts and the
values decreasing as the reactions written to express oxidation—from left to right—
tend to take place with less force (as in the table of “ Oxidation Changes. ”)
Potential in acid solutions (N/1H+)

Nature of change
Cl Br I
Halide —»<— free halogen. 0.65 0.92 1.37

2X~ -+*— X2 + 20
Free halogen —»<—hypohalite. 0.37 0.41 0.63

X, + 2EUO -><- 2HOX + 2H+ + 2©
It is evident from these values that the material on the left of the upper equation will
change to the right while the material on the right side of the lower line will change to
the left. In other words, in acid solutions, halides react with hypohalites1 to produce
free halogen. The sum of these changes is expressed by the equation (with chlorine):
NaCl + NaOCl + H2S04 = Na2S04 + Cl2 + H20
But in alkaline solutions, the numerical values of the potentials of these changes
have the reverse relations—as is shown in the following table :
Potential in alkaline solution (N/10H“)

i
Nature of change
, Cl Br I
Free halogen —><— hypohalite. 1.57 1.61 1.46

X, + 4(OH) —><■- 2(XO)~ + 2H20 + 2©
Halide —m— free halogen. 0.65 0.92 1.37

2X--*<-X°2 +2©
With this arrangement according to the potentials, the material on the left of the
upper equation will react with that on the right of the lower equation, or, in alkaline
solutions, free halogen will react with free halogen to form hypohalite1 and halide.
1 Hypohalous acids are all extremely weak (practically not ionized!) and hence changed to salts only
in strongly alkaline solutions. The gradation in strength is from I to Cl, hence HOI is neutralized but
slightly in alkaline solutions. The salts are naturally extensively ionized.
The difference between the potentials, in acid and alkaline solutions, of the
change of free halogen to hypohalite is due to the fact that this reaction involves
H+ or 011“ ion and these are obtained from, or added to these different solutions
with different ease or difficulty. For further discussion of this topic, see Part I:
“Potentials of Changes Involving Hydrogen and Oxygen Ions.”
The equations for the reaction of free halogens in alkaline solution is the sum of
the equations in the preceding table—the first reading from left to right, and the
second from right to left. Adding them accordingly, and inserting the inactive Na+
ion, we obtain (with chlorine)
2C12 + NaOH = 2NaOCl + 2NaCl + 2H20
This striking difference in the behavior of the free halogens in acid and alkaline
solutions shows itself again in the behavior of the hypohalites themselves. These
substances also can change so as to be reduced (to free halogen or to halide ion) or
so as to be oxidized to the halate stage (the intermediate halite stage does not appear
to be formed by direct oxidation of hypohalite!). In other words, they will act
simultaneously as oxidizer and as reducer provided the potential relations are such
as to impel such a change. On account of the fact that the potentials of all the changes
possible would have to be considered, a presentation of the potentials will be omitted,
and a general statement of their relations only will be made.
In alkaline solutions, we have the following relations:
(a) For each halogen, the potential of the change of its hypohalite salt (i.e., free
X0“ ion!) to halate is numerically the greatest of all the different possible oxidation
changes from hypohalite to halate stage.
(b) For each halogen the potential of the change of its free hypohalous acid to
halide is numerically the least of all the different possible reduction changes of the
hypohalite stage.
(c) The potentials for the change from hypohalite to halate are numerically
greater than those for the change from hypohalite to halide. Hence these two
changes have the proper potential relation for taking place together;1 and since
the greatest potential for their combined action is obtained when the free (un-ionized!)
hypohalous add is reduced to halide while the hypohalite salt (or ion!) is oxidized to
halate, it follows that these are the changes that actually take place when these sub¬
stances are present, and that the equations for these reactions are
HOX + H+ + 2© = X- + HoO
NaOX + 40H“ == NaX03 + 2H20 + 4©
Multiplying the first by 2, adding, and simplifying, we obtain
2H0X +■ NaOX + 2NaOH = 2NaX + NaX03 + 2H20
The proper conditions for the reaction—i.e., the simultaneous presence of free
hypohalous acid and hypohalite salt—are obtained with chlorine by using a solution
of sodium bicarbonate or a “suspension” of lime or of magnesia, and bubbling chlorine
gas into this. To hasten the reaction the solution is kept warm.
The above-mentioned solutions are very nearly neutral to begin with, and as the
reaction uses up 0H~ ion, they will maintain their state of near-neutrality—the bicar¬
bonate by evolving C02—the others by dissolving more oxide. Since the whole
change begins with the free halogen, and ends with halide and halate respectively,
we may write the reactions for it without considering the intermediate hypohalite
stage. Thus the equations are
Cl°2 + 2© » 201“
Cl°2 + 12(0H)“ - 2(C10*)“ + 6H20 + 10©
1 See Part I: Notes on Table of Oxidation-changes.
Multiplying the first by 5, adding, and inserting the inactive cations (Na+), we obtain
6 Cl 2 + 12NaOH = 2NaC103 + lONaCl + 6H20
Since, in acid solutions, all halates are stronger oxidizing agents than in alkaline
solutions, the above resulting mixture of a halide and a halate will, when acidified,
react to form the free halogen again according to the following equations:
2 Cl" = Cl°2 + 2©
KC103 + 6H+ + 6© = KC1 + 3H20
Multiplying the first by 3, and adding the result to the last, we obtain
6HC1 + KC103 = 3Cl2 + KC1 -f 3H20
For bromides and iodides the corresponding equations will express the same changes.
This reaction between iodate and iodide in acid solutions is made use of in iodimetry
because it will stop as soon as the solution has become neutral. Thus it is used to
determine the normality of an acid.
Chloric acid (or a chlorate plus an acid) is used frequently as an oxidizing agent,
and according to its position in our table (No. 33), it will oxidize all common sub¬
stances capable of oxidation. It is a much stronger oxidizing agent than nitric
acid. In analytical work it is used to oxidize As2S3 to arsenic acid—a change which
takes place according to the following equations (using method (2) for writing
these equations—see Part I)
As2S3 + 8H20 = 2H3As04 + 10H+ + 10© + 3S

KCIO3 + 6H+ + 6© = KC1 + 3H20
Multiplying the last equation by 5 and the first by 3, and adding the results, we obtain
on simplifying
3As2S3 + 9H20 + 5 K CIO 3 = 6H3As04 + 9S + 5KC1
Perchlorates are formed by anodic oxidation (electrolysis) of chlorates in neutral

or acid solutions—according to the equation
C103- + H20 = C104- + 2H+ + 2©
They are also obtained by heating chlorates moderately: a portion will react thus,
4KC103 = 3KC104 + KC1
while another portion—the relative amount of which increases with the temperature—
reacts to evolve oxygen gas.
Perchlorates and perchloric acid are very stable and do not act as oxidizing
agents except to extremely strong reducing agents. Hence, they are not employed
as oxidizing agents in solutions, but merely as dry salts in explosives.
Periodates are formed by the action of moderately strong oxidizing agents (e.g.f
chlorine) on iodates in alkaline solutions; also by heating dry iodates, and by anodic
oxidation of iodates. Periodates and periodic acid are not so stable as perchlorates or
perchloric acid, and they act as weak oxidizing agents. Thus zinc reduces periodic
acid to free iodine, but mercury scarce reacts with it. This phenomenon of higher
oxidation stages having lesser oxidizing powers than lower stages has been discussed
in Part I.
In conclusion, attention should be called to the fact that for the changes from
halide to free halogen
2X- - X°2 + 2©
the potentials are the same in acid and alkaline solutions, because they do not involve
H+ or OH ~ ion. But the potentials of all oxidation-reduction changes of oxy-halogen
compounds change with the H+ or OH- concentration of the solutions in which they
occur, and they have values which place them in our Table with reference to their
corresponding “halide-halogen” line, as follows: above in alkaline solutions; below
in acid solutions.
Furthermore, corresponding changes of all compounds of chlorine, bromine, and
iodine, except perhalates, would occupy, in our Table, the same order from top to
bottom as the “halide” halogen-potentials—that is, ranging downward in the order—
iodine, bromine, chlorine. For instance, iodic acid is a weaker oxidizing agent than
chloric—just as iodine is a weaker oxidizing agent than chlorine.
Manganese.—This element exists in the following oxidation stages:

Manganous compounds. MnO, MnS04 — Mn++
Manganic compounds. Mn203, MnCl3 — Mn+++
Manganites. Mn02, CaMn03 — Mn4+
Manganates. Mn02, K2Mn04 — Mn6+
Permanganates. Mn207, KMn04 — Mn7+
The salts of the first and last stage are stable in acid solutions, and unstable in
alkaline solutions, while the reverse is true of the other oxidation stages.
Freshly precipitated manganous hydroxide is flesh-colored, but turns dark from
the formation of Mn(OH)3 by the oxygen of the air. With strong oxidizing agents,
it will naturally change to a higher oxidation stage. Thus with permanganate in
neutral solutions, it will change to Mn02 according to the equations (using method
(2), Part I for writing these equations)
Mn++ + 2H20 = Mn02 + 4H+ + 2©

KMn04 + 4H+ + 3© = Mn02 + 2H20 + K+
Multiplying the first by 3, the second by 2, adding the results, inserting the inactive
anions (e.g., S04 ) and simplifying, we obtain
3MnS04 + 2H20 + 2KMn04 = 5Mn02 + K2S04 + 2H2S04
In order to prevent the solution from becoming acid, we add solid zinc oxide to it:
this serves also to convert the Mn02 to ZnMn03—zinc manganite—the hydrated
Mn02 functioning as an acid of the formula H2Mn03. The above reaction is used for
the quantitative determination of Mn++ ion.
Manganates are formed by fusing together a compound of one of the lower oxi¬
dation stages of Mn together with sodium carbonate, and an oxidizing agent. The
latter can be either air, or KN03, etc. With air and Mn02, the reaction takes place
in accordance with the equation (dualistic formulas to express the relations—see
beginning of Part I)
Mn02 -j- O **{- Na20*C02 — Na20'Mn03 -1- C02
The green manganate will dissolve, and remain intact, in alkaline solutions; but
in acid solutions it changes by the oxidation of some of its molecules and the simul¬
taneous reduction of others—according to the equations
(2Na+, Mn6+, 40“-) + 4H+ + 2© = Mn4+02“ + 2H20 + 2Na+

Mn04““ = Mn04- + 1©
Multiplying the last by adding the inactive anions (bicarbonate, HC03 ), we

obtain
3Na2Mn04 4- 4H2C03 = Mn02 + 2NaMn04 + 4NaHC03
It is advisable to use C02 to acidify the mixture because the excess is easily disposed of.
Otherwise, any acid will do for this purpose.
In a solution acidified so as to be normal in H+ ion, a solution of permanganate
containing 1 gram-molecule per liter has the potential given in No. 32 of our table.
It is evidently one of the most powerful oxidizing agents at our disposal, and it will
oxidize everything in the left column of the table above line 32.
Line 32 presents the potential of the change
Mn++ + 4H20 = Mn04" + 8H+ + 5©

01 MnS04 + 4H20 + K+ = KMn04 + S04 " + 8H+ + 5©
The reaction is quantitative and exact, and hence forms the basis of a great many
titrometric processes. A few examples will be given.
{a) Oxidation of Oxalic Acid—the equation for its oxidation is (see under “ Carbon”
above)
(COOH)2 = 2C02 + 2H+ 4- 2©
The equation for the reduction of permanganate in acid solution is
KMn04 4- 8H+ + S04“- 4- 5© = MnS04 4- 4H20 4- It+
Multiplying the first by 5, the last by 2, adding the results, inserting 1S04— to both
sides, and dropping 10H+ from both sides, we obtain
5(COOH)2 + 3H2S04 4- 2KMn04 = 10CO2 + K2S04 4- 2MnS04 4- 8H20

(b) Oxidation of Nitrites—see under “Nitrites.”
(c) Oxidation of ferrous salts—see under “Iron.”
(d) Oxidation of hydrogen peroxide—see under “Oxygen.”
In alkaline solutions, permanganates exert a much feebler oxidizing tendency—

so that the change
Mn04~ 4- 1© MnO 4
on the basis employed in our Table, has a potential of 1.39 volts, or a position between
the iodide-iodine and the ferrous-ferric potentials (Nos. 22 and 23 in the Table).
If the alkaline or neutral solution contains any manganous salt, the permanganate
will react with the latter to form Mn02 (see above).
Iron.—The only oxidation-reduction change of iron which is of impor¬

tance is the change
Fe++.>< - Fe+++ 4- 1©
the potential of which is in our Table (No. 23). From this it is seen that all
the substances above 23 in the left column will reduce ferric salts, while those
below 23 in the right column will oxidize it. Simple oxidation reduction
equations with iron salts have been given in Part 1.
It will suffice here to give the equations for the oxidation of ferrous salts by HN03
and by KMn04.
(a) With HN03, the equations arc

Fe++ Fe+++ + 1©
HNOa 4* 3H+ 4- 3© = NO + 2HaO
Multiplying the first by 3, adding, and inserting the inactive anions (3S04 and
3 NOr) we obtain
3FeS04 4 4HN03 = 3Fe / N03 + NO + 2HsO

\so4
or 6Fe2S04 + 8HN03 = 2Fe2(S04)3 + 2Fe(N03)3 + 2NO + 4H20.
(b)With KMn04, the equations are
KMn04 + 8H+ + 5© = K+ 4- Mn++ + 4H20
Fe++ = Fe+++ + 1©
Multiplying the first by 2 and the last by 10, adding the result and inserting the inactive
anion (S04 ), we obtain
2KMn04 + 8H2S04 + 10FeSO4 = K2S04 + 2MnS04 4- 8H20 + 5Fe2(S04),
The valence of iron in its double cyanides—e.g.} potassium ferrocyanide,

K4Fe(CN) e, and potassium ferricyanide, K3Fe(CN) 6—is sometimes puzzling and hence
may be considered here.
The simplest way to recognize the valences of the iron in these compounds is to
consider them to be split into their “constituent ” simple cyanides—thus K4Fe(CN)6 ==
4KCN + Fe(CN)2. It is evident that we have a bivalent iron cyanide in this case.
Similarly,
K3Fe(CN)6 = 3KCN + Fe(CN)3
and this shows that we have a trivalent iron cyanide in the last case. Hence the
names—ferrocyanide for the former and ferricyanide for the latter.
The cyanides can be easily changed from one to the other. By anodic oxidation,
ferrocyanides can be changed to ferricyanide according to the equation
Fe(CN)6-= Fe(CN)6-+ 1©
The potential of this change—on the scale in our Table—is 1.60 volts (between
Nos. 21 and 22 in the table).
The oxidation of ferrocyanide can also be effected by means of chlorine—according
to the equations
Cl°2 + 2© = 2C1-
Fe(CN)fi-= Fe(CN)6-+ 1©
Multiplying the second by 2, adding and inserting the inactive cations (e.g., K+),
we obtain
Cl* + 2K4Fe(CN)6 = K3Fe(CN)6 + 2KC1
The reduction of the ferricyanide is also easily accomplished—either by cathodic

reduction, or by the action of mild reducing agents such as iodides, sulphides, etc.—
the equation for the reduction by iodides being
21” = Io 4-2©
Fe(CN)6— 4- 1© = Fe(CN)6-
These two can be combined just like the preceding.

*K<'TIUN XXI
ELECTROLYSIS
iW I h.\ w.n M, Luum.m,1
Whi'n « direct current p:c; .e. through a chemical compound, either in .solution
or (used. 1! the '"huge exceeds a certain minimum, decomposition will take place,
and the eompo. ini* elements will he set free at. the two poles, or new combina¬
tions will he formed with or at the electrodes.
it is poss.il*le .dso tleii slight chemical I’liangca take place when electric currents
are pass, d thioHF.h cei t un snhds, hut these changes, if they occur, at, least are not
«»! mwh umgnitml*- im In be o| enmmetrial uiltltn
l4*|5iitl*i i4«* tliriinrli, m mSu fhiee rlunara; thune uhieh rnnduet <’un*ontH without
iIrrtf!fi|i»:tii!Mii . ;i:i h«|tud mrf.ilro; th*«ne tt hieh tio m »f rumiurt (herurrent; thow» which
thnmmprm* under thr iiiifijnirr nf fhe rurreiit trier!rolyfea). Water probably (tomOH
III ihr rht'id, iiiifi romlueforu, hut uUuhifely imiT water in ho diliieult of
UeO f lit' |«*lht V-* olmnilr {hut tiled Will it in, at hunt, ahliOHt H non-
I’MlItllir till
Electrical Units. \eglerfmg flu* rg.M. defmitioiw, the practical electrical

tint!h uoiv be defined nn ; The ampere hmucha current that panned through
it alv*’t iiitiitt*' .x*iltifi*in hi \%Hti*t, tf mil dopo.nit (I 00111K27 g, of silver per second.
U»r |m|- i *;rr m defined un a coulomb, The ohm is the unit of
fr^mtniire. If m flu" imin.1 iiuro utTered |*y ii column of mercury at 0°( 14,4521 g.
in in y •, *4 *n»n h** * ertinu and 100,,2 em. in length (also defined an the
re i tanee *4 n rolninn 10**20 em long, 1 i4t| em, in enwa-seution at 0* (h). The
\i4f i > fii*4 it!.if Hi pir 't$tr and 1 udt will produce a steady current of 1 amp.
tliiMUfdi i nhiji fr-rtifiiv Tie- wiiti m the unit of energy; 1 amp. flowing for
1 ’re ttfidii ,» pi* -tnr ,*i I \»4t irquiren the expenditure of 1 watt of energy.
Elrrtforfirilikal Kffiiivttfrstlt*. The iiuimitif of any given substance fhnf m
*et fie?\ ptMvalsep ,4 * m * am minimum voltage in exceeded, in a function only of
tin* time and do amp* j* - run* n* that art1 pieced through the electrolyte, and
tin i tjf|„ia?it \ no! uOrjrd b\ the t'tiftiliiiiiifiuii in which the given nulmtaneo
r%r*t >„ ponimg ?hr * .mme* e» nut altered, Thun, 0.01070 g. of eopper are
d^|«r4ted if i mop p,r 'T-, fl,n»ngli a ruprif wilt for l min., whether it he euprie
ftiilpledr, napf^ h}U’,i*ir *«i nipiie rhlonde. In tin* wime way, 0,00700 g. of
mIvit i%i!1 1«* 4ti«n?i'0 h\ I 'yap doming for t lain, and 00 X 0.015700 or 4.025K g,
hy 1 map tloming 0 % I hi 2+he*e eleetroehemieal ecpiivnlentH are proportional
to fhe ntoiisr weigh!- On ofed h\ the \iilriirte A nlmrt table of thene equivalentn
h gn oi
‘d.JI 4 i.s.t,.. i t. »'si r, ,f,-,Nrw \ M.k, N V.
741
Electro¬
Atomic chemical Grams per
Chemical
Element weight Valence equivalent, ampere-
equivalent
O = 16 grams per hour
second
Electropositive
Aluminum. 27.00 3 9.00 0.00009327 0.3358
Antimony. 120.20 3 40.07 0.00041525 1.4949
Bismuth. 209.00 3 69.67 0.00072202 2.5992
Cobalt. 58.97 2 29.48 0.00030558 1.1000
Copper (ic). 63.57 2 31.78 0.00032942 1.1858
Copper (ous). 63.57 1 63.57 0.00065885 2.3718
Gold. 197.20 3 65.73 0.00068122 2.4524
Hydrogen. 1.008 1 1.008 0.00001045 0.03764
Iron (ic). 55.84 3 18.61 0.00019296 0.6946
Iron (ous). 55.84 2 27.92 0.00028934 1.0416
Lead. 207.20 2 103.6 0.00107371 3.8654
Mercury (ic). 200.60 2 100.30 0.00103944 7.4840
Mercury (ous). 200.60 1 200.60 0.00207888 14.9680
Nickel. 58.68 2 29.34 0.00030079 1.0946
Potassium. 39.10 1 39.10 0.00040512 1.4584
Silver. 107.88 1 107.88 0.00111827 4.0258
Sodium. 23.00 1 23.00 0.00023838 0.8582
Tin (ous). 118.70 2 59.35 0.00061510 2.2144
Zinc. 65.37 2 32.68 0.00033869 1.2192
Electronegative
Bromine. 79.92 1 79.92 0.00082824 2.9814
Chlorine. 35.46 1 35.46 0.00036749 1.3230
Iodine. 126.92 1 126.92 0.00131539 4.7353
Nitrogen. 14.008 3 4.67 0.00004840 0.1742
Oxygen. 16.00 2 8.00 0.00008291 0.2985
1 amp. per square decimeter = 9.29 amp. per square foot = 0.0645 amp. per
square inch.
1 amp. per square foot = 0.006944 amp. per square inch = 0.1076 amp. per
square decimeter.
1 amp. per square inch = 144 amp. per square foot = 15.50 amp. per square
decimeter.
In case an equivalent is not given it can easily be found by proportion. Thus,
magnesium has an atomic weight of 24.32 and a valence of 2. Its chemical equivalent
is therefore 24.32 r2 = 12.16. The amount of magnesium deposited per ampere-
hour is therefore Ag ; Mg :: 107.88 :12.16 :: 4.0257 : x.
The general theory of electrolysis is pointed out by one of the calculations

noted above, that the atomic weight must be divided by the valence. It is
believed that all electrolytes dissociate in solution into parts called ions, which
carry electric charges proportioned to the valence. All bivalent ions carry
identical charges; all trivalent ions carry identical charges, which are one-and-a-
half as great as those carried by a bivalent ion. These ions may be simple or
ELECTROLYSIS 733
complex. In copper sulphate they are Cu++ and S04~~, in silver sulphate,
Ag+, Ag+, SO4-” and in ammonium sulphate, NH4+, NH4+, S04—, both positive
and negative ions in this last case being complex.
This dissociation into ions takes place without the aid of the electric current
but is only complete if the solution be very dilute. When electrodes are immersed
in an electrolyte the positively charged ions are attracted to the negatively charged
electrode and there lose their charge, similarly the negatively charged ions are
attracted to the positive electrode. At the electrodes the ion is either deposited, or
sets up secondary reactions. In the electrolysis of copper sulphate, the copper is
deposited as a plating on the negative electrode (cathode) while the S04 ion, if the
positive electrode (anode) be soluble, dissolves it, as Cu++ + S04— = CuS04. If
the anode be insoluble (platinum, graphite), the sulphate ion reacts with the water of
the solution H20 + S04— = H2S04 + O™.
According to this theory there is no real conduction through an electrolyte, as
there is through a metallic conductor, but the charged ions arriving at the electrodes
partially neutralize and thus diminish their charges and permit a further flow of
electricity from the generator to the electrode.
Energy Required for Electrolysis.—As is well known, the formation of chem¬

ical compounds from the elements usually is accompanied by the evolution of
heat, the so-called heat of formation. Some of the heats of formation are shown
in an accompanying table. These are “molecular heats,” i.e., those evolved by
reactions involving gram quantities .equal to the atomic weights (or multiples
thereof), the final mass in grams of the compound produced by the reaction
giving out these quantities of heat being equal to the compound’s molecula
weight.
The small figures under the heats of combination show roughly the combining
quantities in grams. They are the atomic weights rounded off, and are probably at
least as accurate as the values given for the heats of formation.
Beferring again to the table, in general it will be found that these heats divided
by the valence of the base stand roughly in the same order in each column, and it will
also be found that the basicity of the substances decreases from top to bottom of the
table. It is also true that in each list made by dividing combining heats by valences
and arranging the metals in order of the figures thus obtained, a metal will displace in
chemical compounds any metal following it, but will not affect any that precedes it.
Thus iron or lead will throw out copper from solution but copper will not throw down
either lead or iron. In general, such a series is spoken of as an electrochemical series
and the elements in it arc said to be electropositive to those that follow and electro¬
negative to those that precede. The usual order given is: + K, Na, Mg, Mn, Zn, Cd,
Fe, Co, Ni, Pb, Sn, Bi, Cu, lig, Ag, Pt, Au, PI, Sb-. But it can be seen from the
various columns of the accompanying table that the order is dependent somewhat
upon the acid radical and with metals very close the order is not an invariable; thus,
cadmium may be electropositive to cobalt in one electrolyte and electronegative to it
in another. It is also dependent to some extent upon concentration.
Since, as has already been stated, the amount of metal deposited per hour is a
function of the amperes only, and since it requires the same energy to break down
a compound that it gives out when formed, it is obvious that if the heat of formation
be known, the minimum voltage at which the compound will break down with insoluble
electrodes can be calculated.1
1This voltage is equal to the counter e.m.f. of a battery whose electrodes are whatever the anode
and cathode would be if the designated reaction went on and where the activating solution is the
Molecular Heats Evolved in Forming
Element Oxides Chlorides Nitrates Sulphates
Potassium, K'. . 98,200 105,700 119,000 344,300

(78 + 16) (39 + 35.4) (39 + 14 + 48) (78 + 32 + 64)
Sodium, Na'.... 100,900 97,900 110,700 328,100
(46 + 16) (23 +35.4) (23 + 14 + 48) (46 + 32 + 64)
Barium, Ba" . .. 133,400 197,100
(137 + 16) (137 + 71)
Calcium, Ca" .. 131,500 169,900 202,000 317,400
(40 + 16) (40 + 71) (40 -F 28 + 96) (40 + 32 + 64)
Strontium, Sr" . 131,200 184,700 330,200
(87 + 16) (87 + 71) (87 + 32 + 64)
Manganese,
TVTn" 90,900 112,000 249,400
(56 + 16) (55 + 71) (55 + 32 + 96)
Zinc, Zn". 84,800 97,400 131,700 229,600
(65 + 16) (65 + 71) (65 + 28 + 96) (65 + 32 + 96)
Cadmium, Cd". 66,300 93,700 219,900
(112 + 16) (112 + 71) (112 + 32 + 96)
Iron, Fe". 65,700 82,200 234,900
(56 + 16) (56 +35.4) (56 + 32 + 96)
Cobalt, Co".. .. 64,100 76,700 228,900
(59 + 16) (59 + 76) (59 + 32 + 96)
Nickel, Ni".... 61,500 74,700 228,700
(58.5 + 16) (58.5 + 71) (58.5 + 32 + 96)
Lead, Pb". 50,800 83,900 105,400 215,700
(207 + 16) (207 + 71) (207 + 28 4- 96) (207 + 32 + 96)
Tin, Sn". 70,700 80,900
(118 + 16) (118 + 71)
Copper, Cu". . . 37,700 51,400 81,300 181,700
(63.6 + 16) (63.6 + 71) (63.5 4-28 4- 96) (63.5 + 32 + 96)
Mercury, Hg".. 21,500 53,300 165,100
(200 + 16) (200 + 71) (200 + 32 + 90)
Silver, Ag'. 7,000 29,000 28,700 167,100
(216 + 16) (108 + 35.4) (108 + 14 + 48) (216 + 32 + 96)
Gold, Au"'. -11,500 22,800
(394 + 48) (197 + 106.5)
Hydrogen, H' . . 69,000 22,000 48,800 192,200
(2 + 16) (1 + 35.4) (2 + 14 + 48) (2 + 32 + 96)
It requires 96,494 coulombs to deposit 1 g. = equivalent of any metal. (This

quantity is usually known as a Faraday and denoted by I<\)
Then if n be the valence of the compound, the energy to break up the compound
= nEF watts and Q calories per second. Since 1 cal. = 4.20 watts
nEF = 4.2. Since F = 90,494
electrolyte under consideration. What is meant by “ whatever the anode and cathode would be" is
this. In the simple electrolytic analysis of copper (a copper nitrate solution with platinum anode and
cathode) the instant deposition starts the counter e.m.f. is not platinum against platinum in a solution
ELECTROLYSIS 735
4 2Q Q
E = 967i94» = 227975n (Th°mpSOn’S m,e)‘
This is not quite true, a more accurate statement being given by the Gibbs-Helmholz
equation
nEF = Q X 4.2 +
dE
in which T — absolute temperature and ^ is the temperature coefficient of the
electromotive force. This coefficient is usually small and Thompson’s rule suffices
for practical purposes.
Conversely it follows that if the minimum decomposition voltage of a compound
using insoluble anodes is known, the heat of formation can be calculated. It also
follows that the theoretical deposit of metal per killowatt-hour may be calculated
in advance, though 100 per cent efficiency is never reached in practical work. It is,
however, a most valuable guide, for since many of the heats of combination are
published, it is possible to get an approximate idea of the power that is going to be
required per ton of material produced.
It is also possible to calculate exactly how much of any product should be given
by the number of amperes at one’s disposal, and to check up in this way both the
geneal power efficiency and the ampere efficiency of a plant.
With soluble anodes the case is entirely changed. If copper is being dissolved
from an anode as fast as it is being precipitated on the cathode, the energy of solution
is equal to the energy required for decomposition, so that the current expended is
only that needed in overcoming the resistance of the circuit and the internal resistance
on the electrolytic cell.1 It can be seen why it is preferable to carry on electrolytic
oprocesses using soluble anodes to carrying on the same process by dissolving the metal
by purely chemical means and then depositing it using insoluble anodes.
If the potential difference between two insoluble electrodes is less than the decom¬
position voltage of the electrolyte, a “diffusion” current will still flow. This is a
small current carried by actual migration of ions and tends to produce differences in
concentration of the solution until such time that the tendency to produce further
changes is exactly balanced by the tendency to diffuse. The migration velocities of
various ions has been measured. Some of them are as follows:
Ion Migration Velocities, Centimeters per Second

(Potential gradient 1 volt per centimeter, 18 °C.)
H. . 0.00326 OH. . 0.00180
NH4. . 0.00066 I. . 0.00069
K. . 0.00067 Cl. . 0.00068
Ag.. 0.00056 N03. . 0.00064
Na.. 0.00045 CIO,. . 0.00057
Li.. 0.00035 c2h3o2. . 0.00042
Non-ionized colloidal particles may apparently also carry charges (see p. 288) by
actual physical migration.
Osmotic and Solution Pressures.—There is, however, an absolute measure
for an electrochemical series, independent of the acid radical. When any sub¬
stance dissolves in any solvent where there is not infinite miscibility, after a
1 Including in this “internal resistance of the cell’' the effect of osmotic and electrolytic solution
pressures, to be spoken of later.
736' CHEMICAL ENGINEERING
time there is a certain back pressure against further solution, known as osmotic
pressure. In most respects this is strictly an analogous phenomenon to vapor
pressure, the osmotic pressures (in dilute solutions) being proportional to the
number of molecules dissolved, both in the case of a single un-ionized salt in
varying quantities, or in the case of two un-ionized salts of different molecular
weights. Moreover, as in gases, total pressure is the sum of partial pressures.
The reason for specifying “un-ionized substances5* is that with metallic salts in
aqueous solution the pressures are always greater than should be expected from the
number of molecules involved. The accepted explanation, the only one we can see
that fits all the known facts, is that in extremely dilute solution all of the salt disso¬
ciates into its metallic and non-metallic radicals, and that these parts, so far as osmotic
pressure goes, exercise the same effect as the molecules. This is only true in very dilute
solution, and evidently only a small portion of the molecules present are dissociated
in ordinary solutions. These are the ions already mentioned, and while many of them
are elementary substances, it must never be lost to sight that their physical properties
are not those of the elements, beginning with the most important fact that each ion
carries an electric charge.
Electrolytic Solution Pressure.—In a solution after ionization the positive

charges (on the metallic ions) and the negative charges (on the acid radicals)
are in equal numbers. The condition of the solution is unchanged. But
if a plate of iron be immersed in water and a portion of it passes into solution,
the ions passing into solution would be positively charged, and an equal amount
of negative electricity would appear on the plate. The tendency of a metal
to ionize and to pass into solution is known as its electrolytic solution pressure
and when elements are ranged in order of their decreasing electrolytic solution
pressures we have an absolute electrochemical series, independent of the radicals
present.
Some of these pressures are, in atmospheres:
Zinc . .. 9.9 X 101* Cobalt.. . .. 1.9 Copper.. . 4.8 X 10-'°

Cadumim .. . .. 2.7 X 10« Nickel .. . .. 1.3 Mercury. .. 1.1 X 10~18
Iron. .. 1.2 X 10< Lead. . . 1.1 x io-3 Silver.... . 2.3 X 10~17
Hydrogen .. 9.9 X 10”4
This ionizing tendency is opposed by the osmotic pressure; by the difference in

electric potential generated; and by the fact that in aqueous solution the ions usually
form an insoluble compound by reacting with the water, which coats the metal and
stops further action.
Zn+ + 2H20 = Zn(OH)o + 2H+
If the hydroxide is not insoluble, action will go merrily on, as in the well-known
reaction,
2K + H20 - 2KOH + H2
Therefore we may postulate that the solvent action of an acid on a metal is mainly
not a direct one, but consists in dissolving the insoluble hydroxides formed by most
metals.
From these considerations we can see that the requisites for depositing a pure
metal from a mixed electrolyte are that we have a current sufficiently high to exceed
its solution pressure, and that no metal of lower solution pressure be present and that
ELECTROLYSIS 737
we do not have a high enough current to deposit those metals of higher solution pres¬
sures. On the other hand, we can deduce at this point that to deposit an alloy from a
mixed solution it is only necessary to have a current of great enough potential to
overcome the higher solution pressure and of sufficient dilution in the element of
lower solution pressure that the amperes used cannot be carried by the number of
ions present, together with an anode of the alloy desired, and an electrolyte in which
all its components are soluble.
The voltage difference caused by solution pressure and osmotic pressure is
RT
E =
nF
where R is the gas constant in joules per degree, T the absolute temperature, v the
valence, F the Farady, P the solution pressure, and C the osmotic pressure.
In an electrolytic cell the voltage of the cell is
RT Pi RT P2
E = loge loge
niF Cl n%F C2
where the subscripts indicate the different metals.
And for a uniform metal immersed in a segregated solution, the e.m.f. between
RT C
two points on the metal will be E = log ^ where Ci and C2 are the ionic concen¬
trations at the two points.
Overvoltage.—In those processes where hydrogen is liberated at the cathode

it has been found that the potential difference necessary for electrolysis is
greater than the theoretical amount for the decomposition of the given electrolyte
by varying quantities dependent on the cathode material. That is, the elec¬
trolyzing potential for two portions of the same electrolyte will be greatly different
if a platinum cathode is used in one case and a lead cathode in the other.
Taking a platinum-black cathode as 0.0 (as the voltage is less with this cath¬
ode than with any other), other values (according to Caspari) are as follows:
Values of q in volts
Hg. . 1.3 Sn. . 1. 15 Cu.

Pb. . 1.3 Ni. . 1..00 Pt (bright)
Cd.. 1.22 Zn. . 0..80
No satisfactory explanation is known for this phenomenon.

Passivity and Anode Overvoltage.—If oxygen is evolved at the anode there
are also overvoltage phenomena. These are usually inconsiderable, yet where
millions of pounds of metal are deposited, as at the New Cornelia Copper Co.’s
plant at Ajo, Ariz., or the Chile Copper Co.’s works at Chuquicamata, Chile,
the anode overvoltage may be a considerable monetary item, and can by no
means be neglected. Some of these anode overvoltages are: Nickel, 0.05; iron,
0.24; platinum black, 0.24; lead peroxide, 0.28; graphite, 0.40; bright platinum,
0.44 volts.
There is sometimes an even more serious anode disturbance, known as
passivity. That is, in certain electrolytes certain metals show no tendency to
dissolve completely when used as anodes, even though one would expect them to.
Those exhibiting this quality in the highest degree are: aluminum, chromium,
cobalt, iron, molybdenum, and tungsten.
47
Various hypotheses have been advanced for this. As summarized in 1916 by the
Faraday Society the chief are: (1) The formation of a surface gas film; (2) the forma-
tion of an oxide film; (3) conversion of the surface metal into an allotropi^modifica¬
tion of which the electrolytic solution pressure is low (4) the velocity of ionization
or hydration of the ion is retarded.
Polarization.—Somewhat akin to passivity, if the first hypothesis be correct, is

the phenomenon known as polarization, when minute gas bubbles gather on
either anode or cathode or both and greatly raise the cell resistance. The
phenomenon was of the utmost importance when batteries were a large source
of power and many batteries employed depolarizers, for instance, potassium
bichromate in the Poggendorff cell, which reacted chemically with the hydrogen
formed.
In general polarization disappears when the cell is allowed to stand, partly through
slow coalescence of the small gas bubbles into those large enough to escape, partly by
solution of the gases in the electrolyte. But there is this important point to remember,
that polarization often develops slowly and cell resistances are therefore often lower
during the first few hours of running than they are later, and one important commercial
installation proved of little value because of increased costs due to polarization resis¬
tance. In all experimental electrolytic work, be sure the cells are run long enough for
polarization and passivity effects to develop before drawing final conclusions.
Systems of Refining.—All refining operations may be carried on either by

“multiple” or “series” arrangements. In the first all anodes (and all cathodes)
in any individual tank are in parallel, that is, they all have a direct connection
to the busbars. In the series system only the two headplates in each tank are
connected to the source of current supply, one being the cathode, the other the
anode. Between them are arranged as many parallel plates as one pleases (up to
the point where the resistance of a single tank becomes too high). The side of
each plate toward the tank anode then becomes a cathode and receives a deposit,
while the side of each plate toward the tank cathode becomes an anode and is
eaten away, each plate (theoretically) moving toward the anode end of the tank
by its own thickness as the refining progresses.
Suspending these intermediate plates has always been a great deal of a problem.
There must be a certain amount of tank space below them into which may settle the
insoluble impurities of the metal being refined. The Nichols Copper Co. suspends its
cast series copper “flap jacks” by hooks; the Baltimore Copper Smelting & Rolling
Co. holds its rolled plates by grooved sticks that grip the sides. In each of these
systems if the plate moved over the theoretical distance (its thickness) during refining,
the original support would be detached from the new plate and it would fall. Con¬
sequently a certain proportion of metal even though it be only a strip along the edge
must be left unrefined as a support to the newly deposited metal and there is some¬
times difficulty in parting the two.
There is always a little current lost by its going underneath the plate and traveling
through the slimes in the series tank. The plates must be rolled or cast of very even
thickness, or else some will finish before others, and thereafter pure copper will be
eaten away and redeposited.
While the horsepower efficiency1 of the series tank is greater than that of the mul¬
tiple, the process requires much closer attention to detail and more managerial
1 The ampere efficiency is lower due to lost current through the slimes and other causes, but the
resistance between plates is less, and circuit resistances are smaller, so more copper is deposited per
horsepower-hour in series than in multiple relining, i.e., the horsepower efficiency is greater.
ELECTROLYSIS 739
ability and only two copper refiners use it. The series process has been tried for lead
refining but was discarded, and I am told was tried for electrolytic iron production and
found wanting in simplicity.
It is obvious that a single series tank will have a much higher resistance than a
single multiple tank, though between individual plates the resistance is less. But
whichever is used, these internal resistances cannot be cut down below a certain
limit. They are part of the essential cost of the process. But the circuit resistance
is a dead overhead. So, apart from using large cross-sections and good mechanical
contacts to cut down circuit losses, one will throw many tanks into series to cause
the tank resistances to bear a high ratio to the circuit resistance, and also to get an
economical type of generator. Tor this reason, small electrolytic units are often
built with several cells in series where it would be mechanically possible to install a
single tank.
Tank Arrangement.—In metal-refining (copper, lead, etc.) it is now very
common to build tanks in long rows with a common side to each two tanks and
on this to place a triangular copper conductor. All of the anode lugs of one tank
rest on it and all the cathodes of the next tank. Each small section of the bar
(except for the end anode or cathode) will then carry half the current of one
electrode. It will be objected to at first sight that the triangular rod will not give
a good contact with the electrode, but as a matter of fact it does. Furthermore,
the rods should be kept somewhat oily, as most tank-room atmospheres are
corrosive. The thin skin of oil will not affect the contact. Figure 1 shows
the arrangement of tanks and electrodes referred to.
As said above, the tank arrangement in an installation will be figured to give an

economic ratio of tank to circuit resistance. The amperes to be used is a matter of
dividing the ampere-work to be done by the number of tanks. (One ampere will
deposit 0.00111827 g. of silver per second in one cell. If put through 20 cells in series
deposits 20 times as much.) The outside circuit will be made as short as possible, to
cut down line resistance, and will be calculated to that size where the increased
interest on the cost of the extra copper becomes less than the cost of the power saved.1
i The Institution of Electrical Engineers (British) uses a formula for calculating size of conductors
which is independent of the price of copper. It is C — 2.QA°‘Si or
log C = 0.82 log A + 0.415
where C = current in amperes; A » area in thousandths of a square inch. While undoubtedly sizes
given by this formula are safe, when copper is cheap it pays far to exceed mere safety, and I believe that
the price of copper must enter into the calculation where large currents are being carried.
Electrodes.—Both anodes and cathodes for metal refining are shown in an ac¬
companying illustration. Anodes are usually supported by lugs cast on the anode
(not, however, in series processes). Cathodes are supported by loops or hooks.
The cathode is originally a starting sheet, which may be of the metal to be de¬
posited, in which case the starting sheet is usually melted with the deposit, or it
may be of other material, and the deposited cathode metal is later stripped off.
Special cases of this last are those processes in which tubes are made by depositing
metal on a rotating cylindrical mandrel, or where a wire is made by stripping from a
spirally grooved mandrel. A similar special case where the deposit is not stripped is
where a wire is plated as it is drawn through a bath.
Fig. 2.—Electrode shapes.
The entire bottom of a tank may be made a cathode, as in the Thum silver refin¬
ing process, or the cathode may be a puddle of mercury, as in some variations of elec¬
trolytic sodium-hydrate manufacture.
Anode Materials.—The materials for insoluble anodes must be chosen first

with regard to insolubility, second, with regard to their overvoltage, already
spoken of. To show that it is no small problem it may be mentioned that the
Chile- Copper Co. tried over 3,600 different compositions for insoluble anodes.
For large scale work the usual choices lie between graphite, hard lead, one of the
ferrosilicon alloys (tantiron, duriron, corrosiron, or the like) fused magnetite,
platinum gauze or the Fink cuprosilicide anode.
If hard lead is used, casting appears to. be preferable to rolling, as the cast skin
appears to be somewhat more resistant than the rolled. There are marked differences
in graphite anodes, probably due to the binder material used, and because one
make has proved unsatisfactory for certain work is no reason for abandoning the trial
of graphite anodes. Several makes should be tested. Magnetite anodes, which were
given a most extensive trial at Chuquicamata, eventually were discarded as being
uncommercial. They broke too easily when being handled, and under the particular
conditions existing at Chuquicamata, developed a tendency to crack and split.
ELECTROLYSIS 741
When using anodes of low conductivity it is helpful to cast them hollow and fill
the interior with copper. This insures a much better distribution of current over the
surface of the anode and cuts down the circuit resistance at the same time.
From the electrolytic standpoint it is of advantage to have insoluble anodes in
motion. This decreases the polarization because it sweeps away the gas bubbles
mechanically, and in addition it keeps a layer of more constant concentration at the
anode surface. This again is a mere balancing of commercial factors, the cost of the
power to move the anodes and the mechanical complexities necessary being weighed
against the electrolytic advantages.
Some ordinary soluble anode shapes are shown in Fig. 2. There is not much to be
said about soluble anodes except that uniformity of surface is desirable to prevent
excessive corrosion at projecting points, uniformity of thickness of all the anodes is
essential that all the anodes in a tank may finish at the same time, uniformity of
composition is important that the rate of solution may be the same over the entire
anode and the lugs or other supporting members should have an ample factor of
safety.
Cathode Materials.—Ordinarily, in metal refining, the cathode will be the

same material as the metal deposited. This is not true in zinc refining simply
because there is decided tendency to corrosion of the anode at the solution line
(where the oxygen of the air helps the electrolyte in dissolving zinc) and if zinc
cathodes were used they would often cut through and drop down into the tank
causing short-circuits and broken tanks.
Nor are silver cathodes used in the Thum-Balbach silver refining process, as it
would tie up too much silver and the lower anode resistance would nowhere nearly
offset the interest charge thus created. In the case of zinc electrolysis it would be
more convenient to have an all zinc cathode which could be melted entire, rather
than to have to strip the zinc, but the labor cost of stripping is less a disadvantage
than those evils ensuing from cut cathodes. It is again a balancing of economic
factors, which after all, is what differentiates the chemical engineer from the chemist.
In copper refining, the metal is deposited on “starting sheets,” which are thin
copper sheets which have been deposited on greased copper plates and then stripped
off. Ordinarily grooved sticks are run around the edges of the starting sheet blank,
preventing deposition there and making the sheets easier to strip .from the blank.
As a matter of practical interest it should be noted that the current should be on as
the blanks are placed in the electrolyte so that deposition may begin at once, otherwise
the grease comes off slowly into the electrolyte and when the current is turned on the
new deposit becomes firmly adherent and the starting-sheet blank is spoiled.
What is said of copper starting sheets can be considered typical. Lead and tin
sheets can be made in similar manner, but it is cheaper to cast them. Zinc sheets
may be made, but as already said, it is better not to use them, but to use aluminum
instead.
In plating, the article to be plated is the cathode. As this is a quality process,
and imperfections are as a rule exaggerated in plating, all mechanical imperfections
should be removed before plating, after which all grease must be taken off with soap
or some alkali. After the article has been cleaned it should preferably not be handled
again with bare hands. If articles to be plated can be kept in motion results are
usually superior to those obtained where they are motionless
This is in part due to the keeping the electrolyte stirred so that a solution of uni¬
form composition is presented to the electrodes; in part to the irregular shapes of
articles to be plated, which makes the distance of various parts from the anode
unequal, which effect is minimized by moving the cathode (of course an article will be
suspended in such manner as itself to minimize this effect); and probably to the fact
that mechanical work is performed in this way on the deposited material, thereby
improving its texture.
In this matter of plating, two recent interesting novelties are to remove the cathode
repeatedly from solution, and then reimmerse it,1 and to deposit thin films of nickel
from time to time in a copper plate, thereby greatly increasing its tensile strength.2
The results of this latter innovation seem due to retarding the growth of crystals
through the presence of the interdeposited metal.
Another kink of various applications is the use of a mercury cathode. Because of
the tendency to form amalgams mercury can be used as a cathode to attain some
results otherwise impossible. For instance, mercury can be used to take up sodium
and by having the cathode stirred constantly and project under a partition into a
separate compartment containing water, the water in the second compartment
gradually becomes changed to caustic soda, while if sodium chloride solution is the
electrolyte used in the other cell, chlorine is given off at the anode.
The use of rotating cathodes has already been referred to. Much higher current
densities can be used with rotating than with stationary cathodes, and the deposit
is of better quality. In part this improvement can be attributed to keeping the
electrolyte about the cathode thoroughly mixed and probably in part as in plating to
the mechanical work performed on the deposit, the effect being analogous to cold
rolling or drawing. The effect of the mechanical work done by the electrolyte againts
the cathode is often reinforced by the use of buffing pads against the deposited metal.
Seamless tubes are formed of copper, of iron and of nickel on rotating cathodes.
Usually annealing to remove hydrogen is necessary in order to make them less brittle
particularly if the tube is to be reduced by drawing.
Cell Resistance.—This is composed chiefly of the resistance of the electrolyte.

To reduce this the electrodes will be placed as close together as the mechanics
of the process will permit, i.e., in metal refining the danger of bridging the gap
either by an accidental excrescence on the anode, or the growth of a “tree”
on the cathode due to some local condition. In the electrolytic production of
gases the exigencies of the collecting apparatus will dictate the space. Ths
electrolyte will also be heated to cut down its resistance to the point where the
cost of heating outweighs the saving in power or until adverse chemical reactions
enter.
The saving in current by heating the electrolyte may be entirely counterbalanced

by undesirable effects along other lines. Thus, in silver and lead refining too high
temperatures increase the acid-radical losses beyond tolerance, and if the tanks are an
asphalt or sulphur compound, the tanks will not stand up, while with zinc deposition
with insoluble anodes the electrolyte must be refrigerated to avoid re-solution losses.
There is also a counter e.m.f., due to the electrolytic cell acting as a battery.
This resolves itself into a question of the anode and cathode materials and the con¬
centrations of the electrolyte around them, or the differences of electrolyte if it is a
two-fluid cell.
That is, to take a specific case, anode copper, because of its impurities, has not
the same electrolytic solution pressure as has the refined cathode copper, while the
electrolyte about the anode tends constantly to become different from that about
the cathode, and as a result there is a counter e.m.f. set up.
There is also unquestionably an adsorbed gas film on most anodes and also on the
cathodes if gas is being evolved there. Even where a metal is being deposited on the
1C. P. Marsden, Brit. pat. 142,432 of 1919.
* Chem. and Met. Eng., p. 961, Nov. 23, 1921.
ELECTROLYSIS 743
cathode some hydrogen is usually occluded, which renders it necessary to anneal

many electrolytic deposits that are to be used without remelting. Such are the
seamless steel tubes already referred to. I have been told by one experimenter who
annealed electrically in vacuo that he obtained hydrogen equivalent at normal pres¬
sure and temperature to seven times the volume of the electrolytic nickel from which
it was derived. The so-called protective colloids also form a layer about the cathode
that sends up the tank resistance enormously (see p. 744).
Acid Radical.—In the selection of a salt for any refining process the following
considerations will weigh. The acid radical must be fairly cheap and stable.
The salt must be soluble. If impurities are to be taken out as slime they should
be insoluble in the acid used. The tank lining should not be affected by it.
Poisonous fumes should not be given off, and the cathode should not readily
redissolve.
Plating stands on a little different ground in that it is a quality process. There¬
fore, the cost of the salt used is not vital and in small operations poisonous fumes may
be controlled.
This, however, brings up one important matter. It should always be considered
in commercial operation whether any impurity may exist in the electrolyte that may
give rise to poisonous fumes either by anodic oxidation or cathodic reduction, and if
there is such a possibility, the operation must be conducted where good ventilation
exists and without any close attendance. A typical case is the generation of arsine in
electrolyzing spent copper electrolytes, this reaction taking place at the cathode when
the copper is almost all deposited
Diaphragms.—Where it is desirable to retard diffusion, for instance in a cell

where iron is being deposited on a cathode from a ferrous chloride solution which
was being oxidized to ferric chloride at the anode, or where an organic addition
agent is being used at the cathode a porous diaphragm is placed in the cell.
The classic material for this purpose is porous earthenware.
Diaphragms are also made of “filtros,” apparently a silicon hydrate product,

which permits rapid ionic diffusion, but does not allow rapid intermixture of the
anolyte and catolyte. Diaphragms can also be made of asbestos board for temporary
experimental use, but if they are to be in use long they require strengthening with a
wooden grid or other support. If metal supports are used the experimenter may find
he is working with two or three series cells (see p. 738) instead of a single diaphragm
cell, the side of the metal support toward the anode becoming a series cathode, etc.
Tank Materials.—The usual tank materials are hard or soft lead (extensively
used in multiple copper refining); wood heavily asphalted (lead, zinc and both
series and multiple copper refining); slate (series copper refining); glass (gold
refining and electrochemical industries); porcelain (silver refining and electro¬
chemical industries); Portland cement acidproofed with some bituminous
material or plain for neutral and faintly alkaline solutions; or a sand-sulphur
mixture.
If hard-lead is used 10 per cent of antimony appears to give the best results.
Whether hard or soft lead is used, lead of the same composition as the tank lining
should be used for burning the joints.
Asphalt-lined tanks are subject to the disadvantage that high temperatures cause
the tank lining to run. To reduce the viscosity of the lining finely powdered silica
may be added to the asphalt. Slate, glass and porcelain all have the disadvantage
of being breakable. Slate tanks have the additional one of having to be built up,
so that there are joints that must be carefully sealed, usually with some bituminous
composition. Portland-cement tanks can be made to stand either acids or alkalies,
as explained on page 842, but not the two alternately.
Of the sulphur-sand mixes, the one introduced by Charles S. Bradley may be
considered typical. His mix is composed of 1 part sulphur and 1.4 parts sand ground
to pass at least a 60-mesh screen. The mixture is then heated to about 150°C., where
it flows well and is far enough above the melting point of sulphur (114°C.) to avoid
sudden chilling. But the mixture must not be heated as high as 156°C. for at that
temperature sulphur begins to thicken, and at 180°C. becomes so viscid that it will no
longer flow.
Commercial Electrolytes.—Copper is refined commercially from a sulphate
electrolyte containing about 3 per cent copper and 12 per cent of free acid.
Theoretically the power would be cut in half by using cuprous chloride rather than
copper (ic) sulphate, but other considerations render the sulphate the cheaper.
The same electrolyte is usually used for copper plating (IK lb. copper sulphate
crystals and K lb. of sulphuric acid per gallon of water) but copper cyanide is also
used. Silver is both plated and refined with a nitrate electrolyte (5 per cent Ag,
0.1 per cent free HN03 and some copper). Potassium-silver cyanide is also
used for plating. Gold is refined using gold chloride electrolyte containing a
high percentage of hydrochloric acid (Wohlwill process) about 25 to 30 oz. gold
and the same amount of free hydrochloric acid per cubic foot. There is an
interesting feature in this process that an alternating current is passed simul¬
taneously with the direct current used for deposition, the object of the alternating
current being to shake off the silver chloride formed on the anodes. Potassium-
auric cyanide is usually used for gold plating baths.
Lead refining is done with lead fluosilicate solutions acidulated with hydro-
fluosilicic acid and containing some addition agent such as glue or gelatin (Betts
process). The electrolyte contains about 8.50 per cent Pb and 11.5 per cent of total
H2SiFo, both free and combined, and about 0.18 per cent of free HF. Siemens &
Halske have patented the use of lead perchlorate in free perchloric acid as an elec¬
trolyte, using peptone as the most desirable addition agent. The process is said to be
in use at the Hagener accumulator works,1 but as a true refining bath from impure
anodes it appears inferior to the fluosilicate. Nickel is refined using nickel-sulphate
solutions slightly acidulated with boric acid, but nickel-ammonium sulphate (5 to
8 per cent solution) is usually used fpr plating. Cobalt, so far as my knowledge goes,
is not refined electrolytically. Kalmus recommends the following plating solution:
Cobalt sulphate (CoS04) 312 g.; sodium chloride, 19.6 g.; water, 1 liter; boric acid
almost to saturation. Use current up to 100 amp. per square foot. Iron is com¬
mercially electrodeposited from a ferric-chloride solution. Theoretically ferrous-
sulphate or ferrous-ammonium sulphate would be preferable, but so far as my own
experiments go, it is impossible to keep the deposited iron low enough in sulphur
to make the process commercial. The only use of iron known to me as an electro¬
plate is its deposition on engraved copper plates as a hard protective coat that can
easily be removed, when it begins to wear. Ferrous sulphate, ferrous-ammonium
sulphate, iron alum, and double cyanides may all be used for this work.
Platinum is electroplated from platinic-chloride and ammonium-platinic chloride
(acidulated with citric acid and carrying an excess of ammonium chloride) solutions.
But it is not satisfactory, as the anodes do not dissolve readily, and the bath is soon
1 Rideal, “Electrometallurgy," D. Van Nostrand, p. 88.
ELECTROLYSIS 745
too depleted to work well. Palladium is plated on mirrors and other scientific
apparatus from palladic-chloride solutions, using a moving carbon anode.
Zinc is deposited on a large scale from zinc-sulphate solutions using insoluble
anodes and usually aluminum cathodes (Anaconda, Mont., Park City, Utah). Dur¬
ing the World War it was also electrolytically refined using anodes of Prime Western
spelter. The process is not commercial ordinarily as there is not a sufficient margin of
price between Prime Western and High Grade, and there is no precious-metal content
to bear the refining toll as there is (ordinarily) in impure copper. “Electrogalvaniz-
ing” may be done with zinc-chloride, zinc sulphate or double-cyanide solutions.
As a rule zinc solutions must be carefully freed from cadmium, copper, nickel and
cobalt by agitation with atomized zinc. At" Anaconda the electrolyte contains
about 5 per cent zinc and 2J4 per cent free sulphuric acid, and the current density is
about 23 amp. per square foot.1 At Trail, B. C., the electrolyte carries about 7.5
per cent of zinc and 2 per cent of acid on entering the tanks and discharges at 3 to 5
per cent of zinc and 5 to 7 per cent of free acid. Current density is about 24 amp. per
square foot and tank temperature about 30 to 45°C.2 At Anaconda about 1 to 1oz.
of glue per ton of zinc produced is added to the electrolyte. This is typical of the
small amount of these addition agents necessary to give a dense cathode with small
crystals.
Copper is recovered on a large scale, i.e., millions of pounds per year, by leaching
with sulphuric acid and precipitating the metal by electrolysis with insoluble anodes.
At the Chile Copper Co.’s installation at Chuquicamata, the ore is brochantite, an
oxysulphate of copper, so that there is a constant surplus of acid produced by the
leaching operation, which permits discarding a portion of the electrolyte from time to
time in order to prevent undue contamination (see p. 746). The electrolyte normally
carries: Cu, 50.44 g. per liter; S03, 122.75; free H2S04, 28.0; Fe, 3.71; Mn, 0.07;
P, 0.06; As, Sb, 0.0; CaO, 0.8; MgO, 3.32; A1203, 1.61; Na20, 21.6; K20, 5.0; Cl,
11.52 g. per liter. It is freed from chlorine before electrolysis by agitating with cement
copper in revolving drums, which forms cuprous chloride that is then filtered out and
the copper precipitated on scrap iron thus reforming cement copper.
The work at Ajo has been described in great detail by Tobelmann, Potter and
Gross.3 The process consists essentially in leaching with sulphuric acid (and such
ferric sulphate as is present); reduction of the solution with sulphur-dioxide gas or
cement copper and electrolysis. The cement copper referred to is produced by passing
the discarded electrolyte over scrap iron. Acid is purchased from the C. & A. smelting
plant.
The copper leaching and electro-deposition processes of Hopfner and of Siemens &
Halske, usually described at great length, are now only interesting chemical curiosities.
The former is a cupric-chloride leaching process, the latter a ferric-sulphate leaching
process, but it is time some of the metallurgical dead were decently interred, and
ancient history may be consulted for further details.
. Cadmium is recovered from zinc-solutions and electrolytes by chemical means.
It may be plated from an ammonium-cadmium sulphate solution, although the
process has no commercial application to my knowledge. Antimony is refined using
antimony tri-fluoride solution with antimony anodes.
A curious phenomenon in connection with antimony deposition is that if amor¬
phous antimony is deposited from a solution of 1 part of antimony ter-chloride in 5 or
6 parts of hydrochloric acid, of specific gravity 1.12, the amorphous mass changes to
crystalline antimony under the influence of sudden heating, or of a sudden blow
1 F. Laist, Trans. A. I. M. E., February, 1921 meeting.

,
2 E. H. Hamilton, Trans. Am. Electrochem. Soc., Vol. 32 p. 317, 1918.
3 Trans. A. X. M. E., Vol. 55, p. 830; Vol. 60, p. 20.
with the evolution of sufficient heat to scorch paper or wood, and of a cloud of vapor,
apparently from occluded electrolyte.
Bismuth is difficultly refined, as its great tendency to form complex salts in solution
usually results in evolution of hydrogen, and spongy deposits. While Foerster and
Schwabe claim to have obtained good deposits from a fluosilicate solution it is be¬
lieved that the only commerical installations now operating in the United States are
using bismuth chloride containing enough free hydrochloric acid to prevent precipita¬
tion as oxichloride.
Tin is being refined electrolytically at Perth Amboy using a tin-chloride-sulphate
solution with an organic addition agent but the greatest use of electrolytic processes
for tin is in its recovery from tin scrap.' The tin scrap is used as the anode and polished
steel plates as the cathodes. The electrolyte is a strong solution of sodium hydroxide
(about 8 per cent) which is regenerated, as it carbonates, by calcium hydroxide.
The solution is kept hot, preferably over 150°F., and current is passed until the last
of the tin is removed from the iron. The process has not proved adapted to tin cans
from garbage as it has not proved possible to clean them economically and the elec¬
trolyte soon becomes unworkable while the slime containing the tin is foul and often
will not melt, though much time and money have been spent on efforts to treat “ swill
cans.” It is believed that detinning with dry chlorine offers some advantage in cost
over electrolytic detinning for all applications. For electrolytic tin plating a bath
containing 234 to 3 per cent of stannous chloride, 534 to 6X4 per cent of ammonium
oxalate and 0.35 per cent of free acetic acid is probably as good as any, but it is difficult
to get bright dense adherent deposits. Iron and steel should be given a copper strike
before plating with tin.
Deposition of Alloys.—Usually when a current is passed through a mixed

solution, only the least electropositive metal is precipitated. This may be due
to not reaching the higher decomposition voltage, or it may be that the more
electropositive metal is deposited and then redissolves in the solution, precipitat¬
ing some of the less electropositive metal as it dissolves. That this separation
can be made, is of course the basis for all electrolytic refining processes.
Nevertheless it is sometines desirable to be able to deposit alloys. The conditions
for doing so are: (1) If the current be so strong that the more electropositive metal
has not the time to dissolve before a further deposit over it of both metals occurs;
(2) if the solution used has a nearly identical decomposition voltage for both metals.
(3) If the proportion of the more electronegative metal be so small that it is exhausted
by electrolysis as fast as it can be replaced by diffusion, so that the more electro¬
positive metal, as deposited, has nothing to act upon.1
Purification of Electrolyte.—Possibly the greatest problem of the electrolytic

refiner is the contamination of his electrolyte. Purification may be effected in
various ways, some examples of which have been given. Thus, the electrolyte
used for electrolytic detinning is regenerated chemically by calcium hydroxide.
Copper solutions (in refining) can be partially regenerated by plating out all the
copper with insoluble anodes, crystallizing out the nickel and sending the acid
back to the tank house. Nevertheless, there comes a time in commercial opera¬
tions when all that can be done is to recover if possible what valuable metals are
in a portion of the electrolyte either by chemical or electrochemical means,
and send the treated portion to the sewer.
1 For plating formulas, moulding compositions, stripping baths, etc., see Watt’s “Electrometal¬
lurgy,” D. Van Nostrand, or McMillan and Cooper’s “Electrometallurgy,” C. Griffin & Co.
ELECTROLYSIS 747
This problem of electrolyte contamination is not alone a concern of the hydro¬

electrometallurgist. The thermo-electrical worker likewise has his troubles. Though
chemically purified alumina is put into the aluminum bath, the impurities build up
until part of it must be thrown away. Magnesium chloride baths finally become
unusable because of oxide produced by secondary reactions.
In general it may be said that what is the greatest concern of the commercial
operator is usually the smallest worry to the amateur experimenter, and much work
has been done and published that is of no great immediate value because the elec¬
trolytes were not run continuously for a long enough period for all the troubles due to
impure baths to develop and the problems of electrolyte regeneration are unsolved,
and indeed unknown.
Electrolysis with Fused Electrolytes.—While the conductivity of fused

electrolytes is usually higher than that of aqueous solutions, the horsepower
efficiency is much lower. This is largely due to the following causes: (1) Vola¬
tilization of the deposited metal; (2) chemical side-reactions involving reactions
usually impossible in aqueous solution—for instance the formation of the sub¬
chloride, CaCl, in the electrolysis of calcium chloride; (3) cloud formation, or the
dissipation of the precipitated metal through the fused electrolyte as a cloud.
The use of the lowest feasible temperatures minimizes this. (4) The re-solution
of the metal in the electrolyte.
Aluminum is produced by the use of a bath consisting of cryolite (sodium-alumi¬

num fluoride) in which alumina is dissolved as needed, about 20 per cent should ordi¬
narily be kept in the mixture. Because of the expense of replacing the bath when
contaminated, purified alumina is used. The bath is contained by a carbon or
alumina lining, preferably the latter and run at 750 to 850°C. The potential drop is
usually 5 to 7 volts, and it requires about 10 hp.-hr. per pound of aluminum.
Magnesium is usually recovered by electrolysis of the fused anhydrous chloride.
If a pure magnesium-chloride electrolyte is used the metal will float to the top of the
bath and burn there to oxide, unless it is trapped by inverted steps on the cathode,
or some similar form of construction. A mixture of magnesium and potassium
chlorides (with preferably a little ammonium chloride avoids this). The natural
potassium-magnesium chloride (carnallite) is therefore in favor where obtainable for
magnesium production. Too high a voltage deposits potassium at the same time,
while even small amounts (0.05 per cent) of iron in the electrolyte cut down the cell
efficiency tremendously through alternate reduction and oxidation.
Sodium and potassium may both be recovered from the electrolysis of the molten
hydroxides. In both the metal is lighter than the electrolyte and is caught by having
a tube suspended over the cathode and closed with a movable cover in which the
metals collect. Sodium has a high surface tension and an interesting use is made
of this fact. The metal is ladled out in iron ladles with small holes through which the
metal cannot pass (because of its surface tension). Any retained electrolyte, how¬
ever, flows away freely through the holes. In sodium recovery electrolysis is said to be
conducted at about 320°C. using a current of about 170 amp. per square decimeter.
A commercial cell for the recovery of sodium from direct electrolysis of sodium
chloride is shown in the illustration. (This shows one of the devices for trapping a
metal lighter than the electrolyte.) A circular furnace is employed, lined with
firebrick, containing an annular graphite and a hollow iron cathode, surrounded
at its upper extremity by a watercooled hood.
On electrolysis the liberated sodium flows up through the watercooled hood and
down through the hollow space into a collecting vessel.
Sodium chloride is also electrolyzed using a molten lead cathode, the cathode sub¬
sequently being treated with steam for the production of caustic soda.
Calcium is prepared from fused calcium chloride or from a mixture of calcium
chloride and calcium fluoride (the latter used because it lowers the melting point
of the mix). Since the melting point of the metal is higher than that of the electrolyte
it is possible by close control of the temperature of the electrolyte to produce calcium
as a solid, depositing it on a so-called “contact electrode.” This is an electrode
(usually iron and usually a circular plate) which can be raised as the metal is deposited,
thereby keeping the separation distance between the electrodes constant. The metal
is quenched in petroleum and the porous mass thus obtained is remelted to free it
from chloride. It is said that about 10,000 amp. per square decimeter are used with a
resulting e.m.f. of 20 to 30 volts. Strontium and barium can also be produced by
electrolysis of the fused chlorides.
Electrodes.—In Heroult furnaces 30 amp. per square inch may be allowed

for carbon electrodes, 50 amp. for other types of furnace, and 150 amp. per square
inch for graphite electrodes.
Electrothermal Processes.—Iron and zinc smelting, the preparation of various
ferro alloys, the manufacture of graphite, the reduction of phosphorus and silicon,
and the formation of carbon-disulphide, carborundum, the oxysilicides of carbon
and many carbides and borides is carried on in the electric furnace, but these
seem to be merely thermal processes requiring extremely high temperatures,
and not electrolytic processes, and so are not treated here. “The Electric
Furnace” by J. N. Pring, Longmans, Green & Co., gives a good review of these
processes.
A thermal use of the electric current, but one which must be mentioned, is the
trick of using an alternating current as well as a direct in work with fused elec¬
trolytes. The electrolysis is performed by the direct current, while the alter¬
nating exerts a heating effect only.
SECTION XXII
CATALYSIS
By Hugh S. Taylor1
Introduction.—Temperature, pressure, solubility and the electric current

are the normal agencies whereby the chemical engineer seeks to promote the
chemical activity of the substances with which he is concerned. In constantly
increasing measure,2 modern industrial chemistry is making use of the catalytic
agent as an auxiliary for the more ready promotion of chemical reaction. Thus,
nitrogen and hydrogen manifest but little reactivity toward one another under
varying conditions of temperature or pressure when brought into contact with
one another without a catalytic agent. In the presence of reduced iron, or, better
in the presence of reduced iron intimately admixed with one or more reduced
metals as, for example, molybdenum, nitrogen and hydrogen, over a wide range of
temperature, react to form ammonia, pressure increasing the equilibrium con¬
centration of ammonia obtainable. Pressure, temperature or solubility are in¬
sufficient to bring about the saturation of an unsaturated liquid fat by means of
hydrogen gas. In the presence of finely divided reduced nickel, however, ready
saturation of the most diverse oils occurs at temperatures below 250°C. with the
production of hard fats suitable for edible purposes or for the manufacture of
hard soaps. The operation of the electric current, in, for example, the prepara¬
tion of per-salts such as persulphates or perborates is modified by the addition
of a small quantity of a fluoride, a greatly enchanced yield resulting. Other
electrolytic preparations may be similarly catalyzed.3
The characteristics of the catalytic agent may be briefly summarized and
illustrated as follows:
1. A catalytic agent accelerates (or retards) the velocity of a chemical reaction,
itself remaining unchanged in chemical composition when the cycle of operations
is completed. Normally, the concern of the chemical engineer is with catalysts
which speed up the reaction process, with the so-called positive catalysts. Occa¬
sionally, as in the addition of such substances as sodium pyrophosphate, hydro-
fluosilicic acid and various organic compounds to hydrogen peroxide solutions,
to diminish the velocity of decomposition of the peroxide, negative'catalysts or
retarding agents are utilized. The participation of the catalyst as an inter¬
mediate compound in the cycle of operations is not excluded, but a true catalytic
agent must be regenerated in its original chemical form on completion of the re¬
action of cycle. Thus, sulphuric acid is a catalyst in the dehydration of alcohol
to yield ether, although, as shown by Williamson, an intermediate stage in the
1 Assistant Professor of Physical Chemistry, Princeton University, Princeton, N. J.

2 For the history of catalysis consult Rideal and Taylor: "Catalysis in Theory and Practice,”
Chap. 1., Macmillan, London, 1919.
3 See Rideal and Taylor, loc. cit., Chap. XII.
749
process may be realized in which ethyl-sulphuric acid may be isolated. The

final production of ether is accompanied by the simultaneous regeneration of
sulphuric acid.
C2H5OH + H2SO4 = C2H5-HS04 + H20. (1)
C2H5 HSO4 + C2H5OH = (C2H5)20 + H2S04 (2)
The catalytic substance may change its physical state in the course of the process.
Thus, a smooth stranded platinum gauze employed as an accelerator of the oxidation
of ammonia by atmospheric oxygen to oxides of nitrogen changes its physical con¬
dition; the strands becoming pitted with minute craters in which, under the micro¬
scope, deposits of finely divided platinum may be observed.1
2. It follows as a corollary of the previous characteristic that small quantities of
the catalytic agent are adequate for the promotion of reaction between large quantities
of the interacting substances. A square foot of platinum gauze is sufficient for the
oxidation of sufficient ammonia to supply oxides of nitrogen to a lead-chamber
process sulphuric-acid plant having a capacity of 500 tons of acid per day. The
intervention of side reactions frequently results however, in a diminished activity of
the catalyst material requiring therefore, its renewal (see, also, p. 754, under “Catalyst
Poisons”).
3. The catalyst, normally, promotes the establishment of the same equilibrium
position that is attained by the manifestation of the chemical forces in absence of the
catalyst; the position of equilibrium is governed by the same factors of temperature
and pressure as hold in absence of the catalyst. Thus, in the interaction of steam and
carbon monoxide
CO + H20 = H2 + C02
the reaction is reversible and the reaction constant
__ [H2O][C0]
[H2][C02]
at varying temperatures in presence or absence of a catalyst is the same and has the
values
T°C = 400 500 600 700 800 1200
K » 0.05 0.1 0.3 0.6 0.9 2.13
High temperatures favor the formation of carbon monoxide as manifested in the

water-gas generator; low temperatures favor the production of hydrogen and carbon
dioxide; but in the latter case the reaction velocity is slow. In presence of catalytic
agents such as oxide of iron, the velocity is accelerated enormously. Use is made of
this in the manufacture of hydrogen from water gas. Let us assume that 200
vols. of steam and 100 vols. of water gas (50 vols. H2 + 50 vols. CO) react at 500°C.
In accordance with the equilibrium equation
ni _(200 -s)(50-s)
~ (50+»)(*)
where x is the number of volumes of carbon monoxide or steam which have reacted
to produce hydrogen and carbon dioxide, the gases issuing from the catalyst at equili¬
brium will be composed approximately of H20 = 153 vols.; CO = 3 vols.; C02 = 47
vols.; and H2 = 97 vols. a composition identical with that which would have been
obtained had the gases proceeded slowly to equilibrium in absence of the catalyst.
Similarly with regard to the influence of pressure upon the equilibrium position.
The following table gives the concentration of ammonia in true equilibrium with
nitrogen and hydrogen (N2 4* 3H2) at different temperatures under pressures of 1,
100 and 200 atmospheres in presence or absence of a catalyst.
1 See Rideal and Taylor, p. 96. Parsons, Jour. 2nd. Eng. Chem., 1919, Vol. 2, p. 550.
CATALYSIS 751
Equilibrium amounts of NH3

Pressure in atmospheres
550°C. 650° C. 750°C. 850°C.
1 0.077 0.032 0.016 0.009

100 6.70 3.020 1.540 0.874
200 11.90 5.710 2.990 1.680
Apparent exceptions to the unchangeability of the equilibrium position by the

presence of the catalyst medium have frequently been observed. Thus, in the case
just cited, Jellinek1 from a study of the role of iron as catalyst, concluded that if the
gases were removed rapidl}' from the catalyst, concentrations of ammonia in con¬
siderable excess of those demanded by the equilibrium could be obtained. It was
shown, however, that iron can absorb or dissolve considerable quantities of ammonia,
which would be given up on modification of the reaction conditions, for example, by
sudden release of the pressure. It is obvious, therefore, that under such circum¬
stances the concentration of ammonia in the solid catalyst phase will be quite definite
and may be distinct from that in the surrounding gas phase. Under such circum¬
stances, the laws of equilibrium in a two phase system will apply.
In catalytic reactions in homogeneous systems, also, apparent exceptions are
known. Thus, in the interaction of an acid and an alcohol to yield an ester it was
shown by Berthelot and Paen de St. Giles, that the reaction constant, in absence of
catalysing acid was
__ (H20)(CH3C00C2H6) __
(CH6OH)(CH3COOH) *
Lapworth and Jones2 showed that this value was by no means constant in presence
of varying quantities of hydrogen chloride acting as a catalyst. With increasing
strength of acid the value rose steadily to nearly 9 in a system in which the ratio
H20 (total water ): HC1 was 4.65. In such cases the apparent variation in K is
explainable on the assumption that the active masses of the reacting species do not
correspond to the total concentrations present, or, alternatively, it may be assumed
that the participation of the reaction constituents in secondary reactions modifies
the active masses taking part in the reaction studied.
For true equilibrium, in a single reaction system, it is demanded by the laws

of thermodynamics that the equilibrium position remain unchanged. For,
otherwise, it would be possible to bring about perpetuum mobile by working
alternately with and without the contact substance; the conversion might thus
be carried backwards and forwards indefinitely and so, work might be accom¬
plished in a way without fall of temperature, in contradiction to the laws of ther¬
modynamics.”3
4. Granted that the equilibrium constant, K, remains unchanged in presence or
absence of catalyst, it follows, since K is composite of hi and ft2, the reaction constants
of the forward and reverse reactions,
K = h/kt
1 Zeit. anorg. Chem. 1911, Yol. 71, p. 121.

that the catalyst must influence the velocity of the two opposed reactions in an
equilibrium process to the same degree. This observation is of importance in indus¬
trial operations. Thus, a catalyst which accelerates the combination of sulphuric
dioxide and oxygen will also accelerate the decomposition of sulphuric trioxide if
brought into contact with the catalyst at a concentration in excess of the equilibrium
concentration of the substance at the given temperature of operation. Similarly,
acids which accelerate esterification processes similarly catalyse the reverse process
of hydrolysis.
5. It is maintained by some that a catalyst cannot initiate a reaction, but may
only modify the velocity with which a reaction already occurring may take place.
By others, this view is warmly contested and the opinion is held that the catalyst may
play a profound part in bringing substances to reaction. From the technical stand¬
point the question is relatively unimportant, and owing to the wide range of catalytic
action it would be difficult to devise experimental tests in which all possibilities of
catalytic agency were excluded.
Types of Catalytic Action and of Catalysts.—In the following table an attempt

has been made to classify the various types of catalytic reaction, to summarize
the various types of catalysts which are active in each class and to illustrate
their application in operations of technical importance.
Reaction process Type of catalyst Exemplification in Industrial Practice
Oxidation. Metals, especially (a) Contact sulphuric-acid process using finely divided
the platinum platinum on a suitable support.
metals (b) The oxidation of ammonia using a platinum gauze.
(c) The preparation of formaldehyde, using copper or silver
(see dehydrogenation).
Oxides (a) Lead chamber process, employing a gaseous catalyst,

oxides of nitrogen.
(b) Contact sulphuric acid manufacture with iron oxide
catalyst.
(c) Chance-Claus process of sulphur recovery. Oxide of
iron as catalyst.
(d) Purification of illuminating gas. Hydrated oxide of iron
as catalyst.
(e) Surface combustion processes. Refractory oxides as
catalysts.
Salts (а) Deacon chlorine process. Copper chloride as catalyst.
(б) Oxidation processes in the dye industry. Mercury and
copper salts as catalysts.
(c) Drying of oils. Metallic soaps as accelerators.
Hydrogenation Metals (a) Hydrogenation of Oils. Nickel and platinum as contact
agents.
(b) Fine organic chemical industry e.g., synthesis of cyclo¬
hexane, reduction of organic compounds, nickel as catalyst.
Oxides (a) Hydrogenation of oils. Nickel oxide as catalyst (especi¬

ally at elevated pressures; see Ipatiev).
Dehydrogenation.... Metals (а) Cracking of oils. Nickel, copper, iron, aluminum as
catalysts.
(б) Dehydrogenation of alcohols. Copper, silver and nickel as
catalysts.
Oxides (а) Cracking of oils, AI2O3 and Ti02 as catalyst.
(б) Dehydrogenation of alcohols, ZnO, SnO, CdO and MgO
as catalysts.
Nitrogen fixation.... Metals and (a) Ammonia synthesis. Reduced iron, iron-molybdenum,
metal-oxide iron-potash, nickel-sodium as contact agents.
mixtures (b) Cyanide formation. Reduced iron as catalyst.
(c) Arc process of oxide of nitrogen formation. Metal
electrodes as negative catalysts.
(d) Nitride formation (Serpek process). Iron, copper,
chromium, molybdenum, as catalysts.
CATALYSIS 753
Reaction process Type of catalyst Exemplification in Industrial Practice
Hydration and hy Acids (a) Hydrolysis of glycerides, soap and candle industry.
Sulphuric, sulphonic (Twitchell reagent) acids as catalyst.
(b) Hydrolysis of starches and wood cellulose to yield sugars,
Sulphuric, hydrochloric, sulphurous acids as catalysts.
(c) Hydrolysis of esters using acids.
(d) Hydration of acetylene to yield acetaldehyde. Various
acids as catalysts.
Alkalies (a) Hydrolysis as in (a) and (c) above. NaOH, KOH,
Ca(OH)2, MgO, ZnO as catalysts.
Dehydration. Acids (а) Manufacture of ether. Sulphuric, sulphonic and phos¬
phoric acids as catalysts.
(б) Esterification processes. Hydrochloric and sulphuric
acids as catalysts.
Oxides (a) Manufacture of ethylene and unsaturated hydrocarbons
Clays, AI2O3, SiOa, TiC>2 as catalysts.
Halogenation. Non-metallic (a) Chlorination of carbon disulphide, carbon monoxide and
elements hydrocarbons. Carbon, iodine, sulphur as contact agents,
Chlorides (a) In synthetic organic chemistry. AlCla, FeCls, SbCl(,
SnCU , ZnCla, HgCh, as contact agents.
Of catalytic agents other than those recorded in the above tabulation especial
mefition must be made of the ferments and enzymes (see Rideal and Taylor, Chap.
XI) which chiefly promote hydrolytic, oxidation and reduction processes and various
processes of coagulation. In the industries these become of importance in the fermen¬
tation process of alcohol manufacture and to a lesser degree in acetic acid and acetone
manufacture. Enzyme action is paramount in the syntheses of nature.
Radiant energy and, more especially, light are important catalytic agencies not
classified in the above. Light plays an important role in processes of halogenation,
and in certain oxidation, reduction and condensation reactions; the drying of oils and
the various photographic processes may be mentioned in illustration. Cathodic
reduction and anodic oxidation are susceptible of catalytic influence and the phenome¬
non of catalysis intrudes in the electro-chemical phenomena of passivity and over
potential.
Promoters.—The admixture of small quantities of other substances with

single catalysts, such as are recorded in the preceding section, frequently results
in the abnormal increase in catalytic activity of the mixture. Normally, it would
be anticipated, the catalytic activity of a mixture should be composite of the
several activities of the components and such is of very frequent occurrence.
Nevertheless, exceptions are well known and have been freely employed in indus¬
trial practice. Thus, in British Pat. No. 27963 of 1913, filed by the Badische
Anilin und Soda Fabrik, it is claimed that by admixture of small quantities of
chromium and cerium oxides with iron oxide the velocity of the reaction
H*0 + CO = C02 + H2
can be considerably accelerated. A mixture containing 2.5 per cent Cr203

and 0.5 per cent Ce02 operates 700 times more rapidly at 500°C. than does iron
oxide alone. To such admixtures the name of catalyst promoters has been given.
The explanation of promoter action is thus far not forthcoming. Indeed, it is
probable that no one single theory would be generally applicable. In certain cases,
as for example, in the use of molybdenum as a promoter to iron in ammonia synthesis
the phenomenon may be associated with the nitride forming or nitrogen adsorbing
properties of molybdenum. In cases of carrier action, as, for example, in oxidation
48
processes, where catalytic action is associated with the tendency of an element to

oscillate between two stages of oxidation the function of the promoter may be to
increase the temperature range over which the catalytic agent is labile. It would
seem, in other cases, that promoters may act by decreasing the sensibility of the
catalyst to substances which diminish its activity. Since it is not possible to give any
general theory of promoters, nor to lay down any principles to guide the investigator
in the choice of promoters, the following tabulation of some of the more important
industrial examples has been compiled.
Promoters employed or
Reaction Catalyst
suggested
Contact sulphuric acid proc- Burnt pyrites Oxides of Cu, V, U, Cr,

ess. • Ni, Co, Al, Be, Zr, Zn,
Ce, Th, Di.
Ammonia oxidation. Oxide of iron Oxides of Cu, Ca, Bi, Ce
Cr,
Platinum Oxides of Pb, Te,
Deacon chlorine process.... Cuprous chloride Chlorides of several ele¬
ments
Removal of sulphur com¬ Iron oxides Oxides of Cr, Ce, Th.
pounds from gases.
Production, of formaldehyde. Various metals Metallic couples
Oxidation of aniline to ani¬ Copper sulphate or vana¬ p.-diamine or p.-amido-
line black. dium chloride phenol
Oxidation of naphthalene.... Mercuric sulphate Copper sulphate
CO + h2o = co2 + h2.... Iron oxide Oxides of Cr, Ni, AI, Ce,
Th, Zn, Pb, U, Na, Cu,
Ag, and Cu ' metallic
couples.
2CO + H2 excess + 02 = Iron oxide Oxides of Cr, Ce, Th, La,
2C02 + H2 excess. Al.
Hydrogenation of oils. Reduced nickel Reduced copper, iron.
Ammonia synthesis. Reduced iron Reduced Mo, W, Co, Ni,
U, Traces of alkalis.
Catalyst Poisons.—Paralleling the abnormal increase in catalytic activity,

due to the admixture of small quantities of substances other than the essential
catalytic agent, is the abnormal depreciation in activity produced by the pres¬
ence, in the reacting species, of minute quantities of foreign materials. Such
substances are termed catalyst poisons from their analogy with the effect of
poisons on the natural processes. It is to the presence of catalyst poisons that
much of the earlier ill-success of catalytic processes in industry may be attributed.
Thus, the contact process of sulphuric acid manufacture, in presence of finely
divided platinum, was patented as early as 1831 (Brit. Pat. No. 6069 of 1831)
by Phillips, but the industrial enterprise was not achieved until the beginning
of this century, when success had been obtained in the technical production
of pure gases. The principal poisons of this catalytic reaction were found to
be arsenic, antimony, phosphorus and lead together with suspended solid par-
CATALYSIS 755
tides. Only when such materials were quantitatively removed was technical
synthesis possible. Similar observations hold for the various other industrial
catalytic processes. Volatile silicon compounds poison the platinum gauze
catalyst in ammonia oxidation. Sulphur and chlorine must be rigidly ex¬
cluded in the catalytic hydrogenation of oil in presence of nickel. Oxides
of sulphur and arsenic impair the activity of cuprous chloride in the Deacon
chlorine process. Carbon monoxide in the nitrogen-hydrogen mixture for
ammonia synthesis, even in concentrations of 0.05 per cent severely reduces
the efficiency of the synthetic operation in presence of iron catalysts admixed
with promoters while oxygen and water vapor exercise a similar effect. Minute
concentrations of sulphur in the reacting gases also destroy the catalytic activity
of the medium. It is evident, therefore, that, from the industrial standpoint,
the problem of catalyst poisons is of paramount importance.
The most recent and comprehensive discussion of the problem of poisoning in

catalysis is due to Bancroft.1 The literature of the subject is there assembled and in
the critical analysis it is concluded that poisons act by decreasing the adsorption of the
reacting substances on the catalyst. The selective action of this latter on the materials
with which it is brought into contact determines the behavior of the catalyst. Experi¬
ence shows that the action of a poison varies widely in different cases. Thus, the
action may be physical as appears to be the case in ammonia oxidation. The volatile
silicon compounds are oxidized on reaching the platinum gauze and are deposited in
the form of silica on the strands of the material, thus diminishing the active surface.
This is true also of suspended matter in the sulphur gases led to the platinum catalyst
in the contact sulphuric acid process. In other cases, as, for example, in the sulphur
poisoning of iron catalysts in ammonia synthesis and in processes of preferential com¬
bustion, where the poisoning is progressive and cumulative the action appears to be
due to chemical reaction between the catalyst and the sulphur gas to form sulphide
which is catalytically inert. If the sulphide formed is sufficiently stable to resist
reduction the poisoning produced is permanent and can only be removed by oxidation
of the sulphide followed by subsequent reduction to the metallic form. A large number
of cases are doubtless due to selective adsorption of the poison. This would seem
to afford the best explanation of the poisoning influence of carbon monoxide,
oxygen and water vapor on iron catalysts for ammonia synthesis. Appar¬
ently, even at temperatures of 500 to 600°C. water vapor, resulting from the reduction
of carbon monoxide to methane, or from the interaction of hydrogen and oxygen or
introduced as such into the reaction gases, is tenaciously held by the reduced metals
acting as catalysts and prohibits the exercise of their normal activity. In such cases
the poisoning effect is temporary, the original efficiency of the catalyst being gradually
restored on passage of the pure gases.
Mechanism of Catalytic Action.—The problem of promoter action is closely

allied to the mechanism of the catalytic activity itself. Theories of catalysis
may, in the main, be grouped under the two classifications:
(a) The intermediate compound theory.
(b) The theory of adsorption.
For catalytic reactions in homogeneous system the intermediate compound
theory appears to be generally applicable. Thus the data, first put forward
by Mrs. Fulhame in 1794 and extended by the researches of Dixon (1880), on
* Jour. Phya. Chem., 1917, Vol. *1, p. 767.
the necessity for the presence of water vapor in the reactions of many gases, for
example, in the combination of carbon monoxide and oxygen, can be formulated
on the theory that the reactions occur in two stages:
2CO + 2H20 = 2C02 + 4H

4H + 02 = 2H20
The fact that sulphuretted hydrogen, ethylene, ammonia, hydrogen chloride

and pentane will also facilitate combination of the same two gases seems to
support the viewpoint. In the lead chamber process of sulphuric-acid manu¬
facture a similar cycle of reactions is postulated for the reaction mechanism of
the catalytic action.
It must be observed, however, that the possibility of heterogeneity even in such

apparently homogeneous media as these under discussion, cannot be totally excluded.
Under such circumstances the possibility of the second mode of explanation must also
be envisaged. In homogeneous liquid systems the intermediate compound theory has
been widely accepted since the explanation by Williamson of the function of sulphuric
acid in the etherification process previously given. The intermediate compound is
employed to explain the large group of catalytic reactions in homogeneous liquid
systems in which the active agent is a dilute acid. Thus, in the hydrolysis of esters
a recent formulation of the reactions occurring1 may be represented in the following
equilibrium equations:
CH3 COOC2H6 + HC1 = CH3.COOC2H5.HCl (a binary oxonium compound)

CH3COOC2H5.HCI 4- H20= CH3.COO.C2H5.HCl.H2O (a ternary oxonium compound)
CH3 COOC2H5.HCl.H2O = C2H5.OH.HCl + CH3 COOH
C2H5.OH.HCl - C2H5 OH + HC1.
Such a representation would account for the observation that the acid accelerates
both the hydrolysis and the reverse process of esterification. The proof of the
existence of such binary and ternary compounds is given in the systematic researches
of Kendall and his colleagues.2 From these studies the general rule has been deduced
that the tendency towards compound formation is the more marked, the greater the
chemical contrast between the basic nature of the ester and the acidity of the catalytic
agent. The concordance of this conclusion with the observation that the catalytic
activity in ester hydrolysis is greatest with the strong acids and diminishes with
decreasing strength of acid forms a striking piece of evidence in favor of the interme¬
diate compound theory in such systems.
In heterogeneous catalytic reactions the role of the intermediate compound is less
certain. It is apposite, however, to record that catalytic activity is common in the
case of substances which may exist in more than one state of oxidation. Cuprous
chloride as catalyst in the Deacon chlorine process is an example in point, the reaction
mechanism being formulated by the following cycle of equations:
2CuC12 = CuaCla + Cl2

C112CI2 + O = CuO.CuCl2
CuO.CuCh + 2HC1 = 2CuC12 4- H20
In support of this is the observation that the chlorides of most substances capable of
forming oxychlorides are similarly active in the oxidation of hydrogen chloride to
chlorine. Other similar cycles might be cited from industrial catalytic processes,
especially in the case of the numerous reactions promoted by the presence of iron'
1 Fa.lk and Nelson,Jour. Am. Chem. Soc., 1915, Vol. 37, p. 1732.
iJour. Am. Chem. Soc., 1914, et seq.
CATALYSIS 757
oxide. The adsorption theory is however, also a possibility as regards mechanism in

the Deacon chlorine process.
Faraday in his researches on the power of metals and other solids to induce the
combination of gaseous bodies introduced the physical explanation of catalytic action.1
According to this theory the phenomena of heterogeneous catalysis are dependent on
the exertion of that attractive force possessed by many bodies, ... by which they
are drawn into association more or less close, without at the same time undergoing
chemical combination though often assuming the condition of adhesion and which
occasionally leads, under favorable circumstances, to the combination of solid
bodies simultaneously subjected to this attraction. Bodenstein and Fink’s researches2
form strong experimental evidence in support of such a viewpoint, the reaction kinetics
being explainable by the assumption that there is an adsorption layer on the surface
of the catalyst through which the reactants must pass by diffusion in order to come
into contact with the catalyst, where they react practically instantaneously. Ban¬
croft3 has developed the viewpoint in a comprehensive discussion of contact catalysis,
concluding that catalysts tend to produce the system which they adsorb the most
strongly. That the catalytic agent accelerates both of the reactions in the equilibrium
process must be borne in mind, however, in connection with this expression of view¬
point. The striking researches of Langmuir, in support of his chemical theory of
the mechanism of adsorption,4 supply additional ideas on the nature of the surface
layer and consequently on the mechanism of the heterogeneous catalytic reaction.
The conclusions reached in investigations of the arrangements of oil films on the sur¬
face of water are identical with the conception of contact action which led Twitchell
to the development of naphthalene stearo-sulphonic acids as agents for the hydro¬
lytic splitting of fats.5 Langmuir’s views applied to the phenomena of heterogeneous
catalysis harmonize to a measured degree the chemical and physical theories of
catalytic action. They show that beyond the simple process of adsorption a definite
orientation of the molecules may be achieved resulting in an enhanced catalytic
activity. Extension of the theory to the explanation of promoter action should doubt¬
less soon be possible.
The Industrial Development of the Catalytic Reaction.—The translation of a

given catalytic reaction from the stage of laboratory success to that of the indus¬
trial enterprise involves a series of undertakings which, from the nature of the
catalytic reaction, are distinctive and peculiar to this type of process. Thus,
since catalysis is concerned largely with the velocity with which a process occurs,
this property calls for careful consideration. Again, since the catalyst may
accelerate reaction over a wide temperature range the equilibrium obtaining in
the given process is also of fundamental import since only by a successful combi¬
nation of the equilibrium factor with the velocity factor may the optimum con¬
ditions of reaction be achieved. In a reaction, also, in which the change is
localized on the catalytic agent it will be obvious, on consideration, that the
thermal magnitudes of the process are of primary importance. The preceding
sections which have emphasized the functions of promoter and of poison in
modifying the essential catalytic operation serve to indicate further peculiarities
inherent in the technical development of the catalytic process.
1 Experimental Researches in Electricity, 1849, Vol. 1, p. 165, Gth series, Nos. 564-659; "Everyman’s
Library,” No. 576, pp. 84-111; Phil. Trans., 1834, Vol. 114, p. 55.
2 ZeiU Phys. Chem., 1903, 1904, 1905, 1907.
3 Jour. Phys. Chem., 1917, Vol. 21, pp. 676, 767.
In exemplification of these factors the historical development of the contact sul¬

phuric acid process may be noted. Discovered by Phillips in 1831 it did not reach
large industrial application before the beginning of the present century. In the inter¬
vening time, the problems which required solution included the purification of the
reacting gases, the preparation of the catalytic medium, the necessary distribution
of the catalyst on a suitable support, the determination of the reaction equilibrium
to locate the optimum time-temperature conditions of operation and the design of
converter units to ensure proper distribution of the reacting substances through
the catalytic mass with adequate control of the temperature of the medium. The
necessity for the investigation of this last factor arose from the exothermic nature of
the reaction between sulphur dioxide and oxygen. The general questions involved
in such investigations are set forth in the following pages.
The Equilibrium Data.—For a reversible reaction in which, over the possible

range of reaction temperatures, the direction of change may be more or less
completely reversed, a prior knowledge of the equilibrium conditions is of the
greatest possible advantage. Thus from the equilibrium data of the water-gas
reaction previously cited (page 750) it is evident that for the production of a gas
with maximum hydrogen or minimum carbon monoxide content at the lowest
possible temperature, from water gas and steam, a catalytic agent working at
the lowest possible temperature to ensure equilibrium conditions is required.
Further, given the minimum working temperature for a suitable catalytic agent
at which equilibrium can be attained, a simple calculation from the equilibrium
constant determines the ratio of the reactants required to produce a gas of definite
concentration of the several reacting species. Thus, in the same example, it has
been shown that a steam water-gas ratio of 2 :1 is adequate for the production of
an equilibrium mixture at 500°C. with a carbon monoxide concentration below
2 per cent.
In gas reactions where, from economic causes, there are present in the reaction
mixture inert gases operating as diluents, knowledge of the equilibrium data is of
especial value. Thus in the contact sulphuric acid process
SO2 4" MO2 = SO3

the oxygen supplied is diluted with nitrogen from the air. The fractional conversion
x — .can ke associated with the equilibrium constant KP at any
I0U3J -+- [0U2J
temperature and the concentrations a and b are the percentages of sulphur dioxide and
oxygen respectively in the initial mixture by means of the equation
100x = 100 -^2-
b — 0.5ax
Kp + 100 — 0.5arc
From this equation it may be deduced that, although an increase in the partial
pressure of oxygen is beneficial in that it tends to raise the value of x, a limit is set
to the quantity of air admitted by the diluent effect of the nitrogen simultaneously
introduced. Hence, assuming optimum conditions, namely, a trace of sulphur dioxide
and large excess of air, it follows (since b = 20.9 per cent) that
A approximates = 0.457
\100-0.5ax \100
or correspondingly
CATALYSIS 759
KP
100s - 100
Kp + 2.2
Hence, it is only for large values of Kp that the yield approaches theoretical, and Kp
is large in this reaction when the temperature is comparatively low; for example, at
450°C., Kp = 188. At 700°C. where Kp is only 4.8 it follows that the yield is
approximately 60 per cent of the theoretical in spite of the large excess of air assumed
to be employed. The economic effect of employing enriched air may also be studied
from the same data, since at 700°C. with pure oxygen the value of 100s approximates
to or ^ Per ceirfc* Actually in practice it has been decided that the opti¬
mum results are obtained with excess air to raise the oxygen content to the ratio SO2:
02 = 2:3.1
In absence of determinations of the equilibrium data or of facilities for obtain¬
ing the same, orientation may be secured into the trend of these by experimental
test. Thus, in the hydrogenation of benzene to yield cyclohexane, experimental work
of Sabatier showed that hydrogenation occurs freely at 180°C., whereas at 300°C.
the reverse process of dehydrogenation of cyclohexane readily occurs in contact with
reduced nickel as a catalyst. Obviously within this interval the sign of log Kp
must change. The Nernst approximation formula for equilibrium conditions may
often serve as a check on such observations. Thus, in the case cited, the formula
has recently been utilized in the author's laboratory with results given below con¬
firming the experimental observations of Sabatier set forth above.
C6H6 + 3H2 = C6H12 + 39,300 cal.
According to Nernst's approximation formula

i .3 pC6H6
pH23 -nO.H* - - 39,300
- log Kp = + 3(1.75 log T) +4.8
log^rc„H12“ 4.571T
whence may be obtained the following values:
T° C. = 0 27 200 227 527
logi^= -16.91 -13.85 -1.21 -0.03 8.17
Obviously the hydrogenation of benzene must be conducted at as low a tempera¬
ture as possible below 227°C. in order to attain maximum hydrogenation in a single
passage through the catalyst mass.
The Reaction Velocity.—Given the equilibrium data it is next necessary to

obtain comparative data of reaction velocities with different catalysts under
fixed conditions of experiment in order to ascertain the most suitable agent to
employ. Ordinarily, it is of advantage to make the preliminary measurements
with purified reacting species, so that by restriction of the variables the necessary
orientation as to relative activities of catalyst may be obtained. In this manner,
a norm may be established in terms of which the efficiency of a given catalyst
substance may be established. Thus, as the researches of Haber have shown,
uranium carbide has been employed in comparing the relative efficiencies of
catalysts for ammonia synthesis. Actually, such a catalyst is not employed in
technical synthesis owing to its great sensitivity to catalyst poisons. Catalytic
agents composed of iron admixed with suitable promoters are preferred to the
uranium carbide, although catalytically less efficient, owing to their greater
resistivity to traces of foreign gases such as carbon monoxide or sulphur
compounds.
1 (Similar observations on the equilibrium factors in other catalytic reactions may be obtained in
Rideal and Taylok, loc dt. pp. 83, 91, 101, 162, 237, 302.)
The measurement of the reaction velocity is relatively simple in the case of

homogeneous catalytic reactions, since the concentrations of the several constituents
and the course of the reaction can generally be readily ascertained. With hetero¬
geneous reactions the problem is more complex since the several magnitudes required
are not capable of precise statement. In illustration of the problem let it be assumed
that a given volume of gas or liquid, V, be forced through a bed of granular contact
material, v, per unit of time. The ratio V/v may be termed the “space-velocity”
(&.F.) of the process and is an important magnitude technically since its size deter¬
mines, in part, the dimensions of the plant required for a given output. In passage
through the catalyst space the reaction substances will attain more or less to the
Fig. 1.—Reaction-velocity and yield curve.
equilibrium concentrations according as the rate of passage be slow or rapid. If the

fractional conversion of the reacting substances to their equilibrium state be measured
it is obvious that this value will decrease as the speed of passage through the mass
increases. The product of the fractional conversion R and the space velocity, S. V.
will give the space-time yield or the amount of reaction effected in unit time per pass¬
age through unit volume of the catalyst space. The variation of the space-time yield
with the space velocity will, in general, give a curve of the type shown in the accom¬
panying diagram.
The first branch of the curve ab indicates the region in which the velocity is suffi¬
ciently slow for equilibrium conditions to be attained, along the second branch of the
curve, be, the equilibrium is no longer attained, but the deviation from equilibrium
concentrations is more than compensated, as regards space-time yield, by the increase
in space velocity. On the tailing oranch of the curve, cd, the space velocity is rising
less rapidly than corresponds to the diminution of reaction occurring and so the
space-time yield falls away. The curve shows that the optimum conditions of opera¬
tion do not necessarily coincide with the conditions under which equilibrium is
attained at the highest velocity of passage. Higher values of the space velocity and
a lower conversion per unit passage may give a greater yield of products per unit of
time. For circulatory systems it follows that the point c represents the best con¬
ditions of operation, other factors being equal, while with non-circulatory processes
the best utilization of the reacting species is attained at b.
Comparison of different catalysts can therefore be expressed in terms of the maxi¬
mum space-time yield for the different materials for the particular working conditions
employed, whether circulatory or non-circulatory. It is to this magnitude that all
the variable factors in the catalyst mass, such as influence of temperature, concen¬
tration of reactants, purity of materials and life of the catalyst will be referred, the
choice of operating material being then decided by adjustment of these data to the
economic aspects of the problem.
CATALYSIS 761
The Form of the Contact Mass: Catalyst Support Materials.—It is evident

that the velocity factor is determined in part by the form of the catalyst material,
which also is a factor of considerable importance in the technical catalyst unit.
The space velocity has been referred to the volume occupied by the catalyst.
The actual velocity of the gas or liquid is determined by the shape, size and dis¬
tribution of the catalyst mass, which factors also determine the back pressure
prevailing in the unit. It has been frequently shown that the catalytic effect
is associated with the surface area of the catalyst medium so that, apart from the
mechanical difficulties involved, a catalyst bed of finely divided material is prefer¬
able. Generally a compromise between the several conflicting items is adopted.
The surface factor and the activity of the catalyst are most often greatest when
the materials efnployed are porous and amorphous. If it is impossible to bring
the material into such a condition and such a size that the resistance to the
passage of the reacting constituents shall not be prohibitively great, the device
is adopted of incorporating the catalytic material upon a suitable support.
Thus, in the contact sulphuric acid process the finely divided platinum is employed
in several different forms. The original Badische type of plant employed platinized
asbestos in long narrow tubular converters. The Tenteleff process exposes the plati¬
num on large coarse asbestos string mats impregnated with platinum black. In the
more recent and most successful Grillo-Schroeder contact material, a very active form
of platinum black distributed over a large surface is attained by soaking partly
dehydrated magnesium sulphate heptahydrate with a spray of a platinum solution
the mixture being then ignited in an atmosphere of sulphur dioxide. In this way the
platinum is reduced and the salt undergoes further dehydration, puffing up and becom¬
ing covered with a thin layer of platinum. The recovery of the catalytic material
from such a support is readily accomplished by solution of the magnesium sulphate
and flocculation of the metal particles.
Among other supports which have found employment in technical practice mention
may be made of pumice, various porous siliceous materials such as fireclay, kiesel-
guhr, alundum and unglazed porcelain, charcoal in its various forms, salts and oxides
obtained in a porous condition by ignition and decomposition of suitable raw materials.
Granular metallic aluminum is now being used as a thermally conducting support
material. As with the catalytic agents, it is impossible to anticipate the particular
support which will be found most suitable. In the main, trial and error must be
employed to determine this. It follows, however, from considerations set forth in the
preceding that care must be exercised to insure that with the support no materials
negatively catalytic in nature are introduced. Frequently, in spite of the most patient
researches, a suitable catalytic support material cannot be obtained without sacrifice
of the maximum efficiency of the catalyst. An adjustment between the several factors
must then be made. The ideal support will be a porous.material of low density, free
from impurities negatively catalytic to the reaction process, itself, if possible, positively
catalytic, even though feebly, to the reaction, and as far as possible incapable of re¬
action with the catalyst mass. With suitable gradation of such a material, the
problem of form in the contact material may readily be solved for all sizes of catalyst
unit (see Rideal and Taylor, Chap. 3).
Heat of Reaction.—The heat change accompanying a catalytic action deter¬

mines in large measure the manner of conduct of the process and the design of
unit in which to carry out the reaction. Thus, in the catalytic conversion of
carbon monoxide and steam to hydrogen and carbon dioxide the reaction is
accompanied by the evolution of 10,000 cal. per gram molecule of reacting carbon
monoxide.
CO + H20 = C02 + H2 + lOKg. cal.
It may be calculated from specific heat data, and has been demonstrated on a
technical unit, that, with a system of heat interchange having a 75 per cent
exchange efficiency, this heat evolution is adequate for the maintenance of the
reaction temperature, 500°C., without the application of external heat. The
steam water-gas mixture in the ratio of 2% volumes of steam to 1 volume of water
gas enters a series of heat interchangers at 100°C., is preheated by the sensible
heat of the effluent gases to 450°C., reacts in the catalyst chamber and is thus
raised to a temperature somewhat higher than 500°C. and leaves by the heat
interchangers in the reverse direction to that of the entering gases, finally passing
out at 100°C. With such favorable thermal conditions the process is self sus¬
taining when the heat losses by radiation and conduction do not exceed 25 per
cent of the available heat.
With a more energetically exothermic reaction, temperature control becomes a
matter of difficulty. Thus, for example, in the contact sulphuric-acid process, the
optimum conversion temperature is in the region 400 to 450°C. Should the gases
reach the catalyst at that temperature, it follows that with a heat evolution of 21,700
cal. per gram molecule of sulphur dioxide converted, the gases would leave the catalyst
at considerably higher temperatures, with correspondingly less efficiency of conversion,
due to the equilibrium data. Such complications are avoided in several different
ways. In the case under consideration, the Badische converter consisted of narrow
reaction tubes permitting high radiation to the incoming gas which passed externally
over the tubes in a counter-current direction, the cold gas thus serving to cool the
catalyst medium at its hottest part. In the Mannheim process the reaction was
conducted in two distinct stages. Energetic reaction was permitted in the first
catalyst unit employing burnt pyrites as catalyst at a temperature of 500 to 600°C.
In this manner, a 45 per cent conversion was obtained. The gases, after suitable
diminution of the temperatures, were led to a platinum-catalyst converter operating
at the temperature of maximum efficiency, in which, owing to the lower percentage
of sulphur dioxide requiring oxidation, the temperature control was comparatively
simple. In other types of contact converters, the sulphur gases are passed over trays
of catalyst separated from each other by intervening gas spaces in which the necessary
cooling of the gas temperatures may be achieved.
- Another device of frequent application to overcome the excessive localization of
temperature in gas reactions is the addition of an inert constituent to the gas mixture
or the addition in a circulatory process of excess of one of the reacting constituents,
or, in a process where the reaction is practically exclusively in one direction, the
addition of one of the reaction products to the incoming gases. These devices operate
by diminishing the heat evolution per unit volume of total gas, thus lessening the
temperature increase of the gas mixture per unit of reaction. Thus, in the removal of
2 per cent of carbon monoxide from hydrogen by processes of preferential combustion
with added air or oxygen, the effective temperature range for preference in the
combustion process is some 50°C. The thermal effect of the reaction in the simple
hydrogen-carbon-monoxide air mixture is adequate in a well-insulated converter
to raise the gas temperature at least 150°C., thus destroying the preferential nature
of the process. This difficulty has been overcome by admixing steam as an inert
constituent, thereby raising the specific heat of the gas mixture per molecule of car¬
bon monoxide requiring oxidation. Similarly in the hydrogenation of acetaldehyde
CATALYSIS 763
to yield synthetic alcohol, rigorous temperature control is necessary in order to obviate

catalytically accelerated side reactions resulting in the degradation of the aldehyde to
methane, carbon monoxide, carbon dioxide and carbon. The exothermic reaction
is controlled by employing the hydrogen in large excess in a circulatory process, this
having the further advantage that it tends to inhibit the reverse process of dehydro¬
genation. Finally, the problem of chlorination of hydrocarbons may be cited as an
example of reaction in presence of charcoal as catalyst, in which the intensity of the
reaction may be minimized by admixing, with the incoming gas, carbon tetrachloride
or other saturated chloro-compound. In this way the ready tendency to explosive
violence, which this reaction shows, may be restrained.
With reactions which take place with absorption of heat the problems arising in
connection with control of catalyst temperature are relatively simple. One or other
of the many methods of communicating heat to any reaction system is generally to
be found suitable.
SECTION XXIII
COLLOID CHEMISTRY
By Jerome Alexander1
Colloid chemistry deals with the properties and behavior of matter in a

certain very fine state of subdivision or dispersion, known as the colloidal condi¬
tion. There are no sharp limitations to the size of colloidal particles; they cover
a zone which begins with dimensions somewhat smaller than a wave length of
light, and extends downward into the dimensions commonly ascribed to molecules.
The laws of “universal gravitation5’ do not hold for particles of matter in extremely
close proximity. We have but to mention cohesion, adhesion, capillarity, adsorp¬
tion, and chemical attraction, to call to mind the wide range of practical every-day
variations from Newton’s law. Molecular attraction necessitated the introduction
by van der Waals of the variable a in his well-known gas equation2:
H)
With the increasing subdivision of matter into smaller and still smaller particles
we begin to enter the zone' wherein “molecular” or “physical” forces begin to play
an increasingly important part, until they finally become dominant. If the sub¬
division is profound enough, chemical combination may result. In fact before
substances can react chemically, their particles must first be brought into close prox¬
imity by solution, fusion, ionization, or even pressure, as was shown by W. Spring,
who caused dry powders to unite chemically by extremely high pressures.
No sharp line can, therefore, be drawn between physical and chemical forces;
for they seem to blend into each other over the colloidal zone, and to exhibit differ¬
ences in degree and modification, rather than in lack of common origin. In fact it
is readily conceivable that such forces may result from the same ultimate cause as
gravitation itself, modified when the structure and motion of the particles cease to
be negligible infinitesimals with reference to the distance between the particles. Sir
William Thompson has indeed expressed the opinion3 that it is possible that the phe¬
nomena of cohesion and others which are ordinarily ascribed to a departure, at small
distances, from the law of gravitation, may not be inconsistent with it.
Coarse powders or suspensions, on the other hand, upon increasing subdivision,
pass by insensible gradations into the colloidal zone. About 1827, the English biolo¬
gist Robert Brown observed that particles approaching in size the limit of microscopic
resolvability, 0.1/x, exhibit an oscillatory movement about a mean position, known
as the Brownian motion. With the aid of the ultramicroscope, which brings into
visibility particles as small as 5ju/x, it can be seen that as the particles become still
smaller than O.lju, their motion increases enormously in both speed and amplitude,
until they exhibit a rapid free-path motion which calculation shows to be a visual,
1 Pros., Uniform Adhesive Co., Brooklyn, N. Y.

2 This equation is intended to care for the departures from Boyle’s law for gas. The constants a
and b for some of the more common gases are given on p. 190.
3 Proc. Roy. Inst., Vol. 11, part III, p. 483.
7G5
tration of the kinetic theory.1 By observing an extremely diluted aqueous

1 of milk, it can be seen that the Brownian movement of the smallest fat
s seems to be mainly due to swarms of actively moving colloidally dispersed
s of casein, etc., which keep up a ceaseless bombardment from all sides,
n what has just been said, it is obvious that no substance is a colloid, per se,
t any substance may be reduced to the colloidal state or condition, providing
fficiently subdivided or dispersed. Thus many soaps dissolve in alcohol to
ystalloidally dispersed solutions, whereas their aqueous solutions are colloidal.
chloride, which in water naturally forms a crystalloidal solution, may be
lly dispersed in benzol.2
Lough modern investigation has modified some of the notions originally enter-
it is interesting to give a few extracts from the work of Thomas Graham,
, Master of the Mint. In two basic papers, the first entitled “Liquid Diffusion
. to Analysis,” read before the Royal Society of London, June 13, 1861, the
entitled “On the Properties of Colloidal Silicic Acid and other Analogous
il Substances,” published in the Proceedings of the Royal Society, June 16,
Iraham pointed out the essential facts regarding colloids and the colloidal
>n, and established much of the nomenclature now in use. It is remarkable
ent research has confirmed the correctness of most of Graham’s views.
juid Diffusion Applied to Analysis.—The property of volatility, possessed

>us degrees by so many substances, affords invaluable means of separation,
een in the ever-recurring processes of evaporation and distillation. So
in character to volatility is the diffusive power possessed by all liquid
Lees, that we may fairly reckon upon a class of analogous analytical re-
to arise from it. The range also in the degree of diffusive mobility exhib-
different substances appears to be as wide as the scale of vapor tensions,
ydrate of potash may be said to possess double the velocity of diffusion of
e of potash, and sulphate of potash again double the velocity of sugar,
, and sulphate of magnesia. But the substances named, belong all, as
diffusion, to the more “volatile” class. The comparatively “fixed”
s regards diffusion, is represented by a different order of chemical sub-
, marked out by the absence of the power to crystallize, which are slow in
:eme. Among the latter are hydrated silicic acid, hydrated alumina, and
Letallic peroxides of the aluminous class, when they exist in the soluble form;
r with starch, dextrin, and the gums, caramel, tannin, albumen, gelatin,
>le, and animal extractive matters. Low diffusibility is not the only
y which the bodies last enumerated possess in common.. They are dis-
led by the gelatinous character of their hydrates. Although often largely
in water, they are held in solution by a most feeble force. They appear
rly inert in the capacity of acids and bases, and all in the ordinary chemical
ls. But, on the other hand, their peculiar physical aggregation with the
. Perrin, and The Svedberg. This was clearly forseen by Sir William Ramsay, who in a
;itled, “Pedetic Motion in Relation to Colloidal Solution” (Chem. News, Vol. 65, p. 90, 1892),
follows: “lam disposed to conclude that solution is nothing but subdivision and admixture,
attractions between solvent and dissolved substance accompanied by pedetic motion, that the
>tic pressure has, probably, never been measured, and that a continuous passage can be traced
visible particles in suspension; and matter in solution; that in the words of the old adage,
ihil fit per saltum.”
elementary definition of a colloid was a substance which would not pass a porous membrane
jarently in solution, while a crystalloid would. This, is of course untrue, and as said again
raham himself knew that it was not so.
COLLOID CHEMISTRY 767
chemical indifference, referred to, appears to be required in substances that can

intervene in the organic processes of life. The plastic elements of the animal
body are found in this class. As gelatin appears to be its type, it is proposed to
designate substances of the class as colloids, and to speak of their peculiar form of
aggregation as the colloidal condition of matter. Opposed to the colloidal is the
crystalline condition. Substances affecting the latter form will be classed as
crystalloids. The distinction is no doubt one of intimate molecular constitution.
“Although chemically inert in the ordinary sense, colloids possess a compensating

activity of their own, arising out of their physical properties. While the rigidity of
the crystalline structure shuts out external impressions, the softness of the gelatinous
colloid partakes ot fluidity, and enables the colloid to become a medium for liquid
diffusion, like water itself. The saint' penetrability appears to take the form of
cementation in such colloids as can exist at a high temperature, hence a wide sensi¬
bility on the part of colloids to external agents. Another and eminently character¬
istic quality of colloids is their mutability. Their existence is a continued metastasis.
A colloid may be compared in this respect to water, while existing liquid at a tempera-
tun' under its usual freezing-point, or to a supersaturated saline solution. Fluid
colloids appear to have always a pcctous modification and they often pass under the
slightest influences from the first into the second condition. The solution of hydrated
silicic acid, for instance, is easily obtained in a state of purify, but it cannot be pre¬
served. It may remain fluid for days or weeks in a sealed tube, but is sure to gelatinize
and become insoluble at last. Nor does the change of this colloid appear to stop at
that point. For the mineral forms of silicic acid, deposited from water, such as flint,
an' often found to have passed, during the geological ages of their existence, from the
vitreous or colloidal into the crystalline condition (II. Hose). The colloidal, is, in
fact, a dynamical state' of matter; the. crystalloidal being the statical condition. The
colloid possesses Knergia. It may he looked upon as the probable primary source
of the force appearing in the phenomena of vitality. To the gradual manner in
which colloidal change's take* place (for they always demand time* as an element),
may the* characteristic protraction of chemico-organic changes also be referred.”
(Tp. 183 184.)
“It. may perhaps he* allowed to me to apply the convenient term dialysis to the
method of separatum by diffusion through a septum of gelatinous matter. The
most suitable* of all substance's for the dialytic st'ptum appears to be the commercial
material known us vegetable parchment, or parchment-paper, whic.h was first pro-
duceel by M, (Jaine, ami is now successfully manufactured by Messrs. Do la Rue.”
(P. 186.)
“Preparation of Colloid Substances by Dialysis.—The purification of many

colloid substances may be* effected with great advantage by placing them on the
dialyzer. Accompanying crystalloids are eliminated, and the colloid is left
behind in a state of purity. Tne purification of soluble colloids can rarely be
effected by any other known means, and dialysis is evidently the appropriate
mode of preparing such substances free from crystalloids.
li Soluble Silicic Acid. -A solution of silieia is obtained by pouring silicate of soda

into diluted hydrochloric acid, the acid being maintained in large excess. But in
addition to the hydrochloric acid such a solution contains chloride of sodium, a salt
which causes the nilicia to gelatinize when the solution is heated, and otherwise
modifies its properties. Now such soluble mlieia, placed for 24 hr. in a dialyzer of
parchment-paper, to the usual depth of 10 mm., was found to lose in that time 5
per cent of itn silicic acid and 86 per cent of its hydrochloric acid. After 4 days on the
dialyzer, the liquid ceased to be disturbed by nitrate of silver. All the chlorides were
gone, with no further loss of silicia. In another experiment 112 g. of silicate of soda,
67.2 g. of dry hydrochloric acid, and 1,000 c.c. of water were brought together, and
the solution placed upon a hoop dialyzer, 10 in. in diameter. After 4 days the solution
had increased to 1,235 c.c., by the action of osmose, colloid bodies being generally
highly osmotic. The solution now gave no precipitate with nitrate of silver, and
contained 60.5 g. of silicia, 6.7 g. of that substance having been lost. The solution
contained 4.9 per cent of silicic acid.
“The pure solution of silicic acid so obtained may be boiled in a flask and con¬
siderably concentrated, without change; but when heated in an open vessel a ring of
insoluble silica is apt to form round the margin of the liquid, and soon causes the
whole to gelatinize. The pure solution of hydrated silicic acid is limpid and colorless,
and not in the least degree viscous, even with 14 per cent of silicic acid. The solution
is the more durable the longer it has been dialyzed and the purer it is. But this solu¬
tion is not easily preserved beyond a few days, unless considerably diluted. It soon
appears slightly opalescent, and after a time the whole becomes pectous somewhat
rapidly, forming a solid jelly, transparent and colorless, or slightly opalescent and no
longer soluble in water. This jelly undergoes a contraction after a few days, even in
a close vessel, and pure water separates from it. The coagulation of the silicic acid
is effected in a few minutes by a solution containing Mojooo part of any alkaline
or earthy carbonate, but not by caustic ammonia, nor by neutral or acid salts. Sul¬
phuric, nitric and acetic acids do not coagulate silicic acid, but a few bubbles of car¬
bonic acid passed through the solution produce that effect after the lapse of a certain
time. Alcohol and sugar, in large quantity even, do not act as precipitants; but
neither do they protect silicic acid from the action of alkaline carbonates, nor from the
effect of time in pectinizing the fluid colloid. Hydrochloric acid gives stability
to the solution; so does a small addition of caustic potash or soda.
“This pure water-glass is precipitated on the surface of a calcareous stone without
penetrating, apparently from the coagulating action of soluble lime-salts. The hy¬
drated silicic acid then forms a varnish, which is apt to scale off on drying. The
solution of hydrated silicic acid has an acid reaction somewhat greater than that
of carbonic acid. It appears to be really tasteless (like most colloids) although it
occasions a disagreeable persistent sensation in the mouth after a time, probably from
precipitation.
“Soluble hydrated silicic acid, when dried in the air-pump receiver, at 15°, formed a
transparent glassy mass of great luster, which was no longer soluble in water. It
retained 21.99 per cent of water after being kept 2 days over sulphuric acid.
“The colloidal solution of silicic acid is precipitated by certain other soluble col¬
loids, such as gelatin, alumina, and peroxid of iron, but not by gum nor caramel.
As hydrated silicic acid, after once gelatinizing, cannot be made soluble again by
either water or acids, it appears necessary to admit the existence of two allotropic
modifications of that substance, namely, soluble hydrated silicic acid, and insoluble
hydrated silicic acid, the fluid and pectous forms of this colloid.” (Pp. 204-205.)
“The possession of a soluble and an insoluble (fluid and pectous) modification is
not confined to hydrated silicic acid and the aluminous oxids, but appears to be very
general, if not universal, among colloid substances. The double form is typified in
the fibrin of blood.” (Pp. 212, 213.)
“The separation of colloids from crystalloids by dialysis is, in consequence, gener¬
ally more complete than might be expected from the relative diffusibility of the two
classes of substances.” (P, 214.)
“Colloidal Condition of Matter.—I may be allowed to advert again to the radi¬

cal distinction assumed in this paper to exist between colloids and crystalloids in
their intimate molecular constitution. Every physical and chemical property

is characteristically modified in each class. They appear like different worlds
of matter, and give occasion to a corresponding division of chemical science. The
distinction between these kinds of matter is that subsisting between the material
of a mineral and the material of an organized mass.
“The colloidal character is not obliterated by liquefaction, and is therefore more

than a modification of the physical condition of solids. Some colloids are soluble in
water, as gelatin and gum arabic; and some are insoluble like gum tragacanth. Some
colloids, again, form solid compounds with water, as gelatin and gum tragacanth,
while others, like tannin, do not. In such points the colloids exhibit as great a diver¬
sity of property as the crystalloids. A certain parallelism is maintained between the
two classes, notwithstanding their differences.
“The phenomena of the solution of a salt or crystalloid probably all appear in the
solution of a colloid, but greatly reduced in degree. The process becomes slow;
time, indeed, appearing essential to all colloidal changes. The change of temperature,
usually occurring in the act of solution, becomes barely perceptible. The liquid
is always sensibly gummy or viscous when concentrated. The colloid, although
often dissolved in a large proportion by its solvent, is held in solution by a singularly
feeble force. Hence colloids are generally displaced and precipitated by the addition
to their solution of any substance from the other class. Of all the properties of liquid
colloids, their slow diffusion in water, and their arrest by colloidal septa, are the most
serviceable in distinguishing them from crystalloids. Colloids have feeble chemical
reactions, but they exhibit at the same time a very general sensibility to liquid reagents,
as has already been explained.
“ While soluble crystalloids are always highly sapid, soluble colloids are singularly
insipid. It may be questioned whether a colloid, when tasted, ever reaches the sen¬
tient extremities of the nerves of the palate, as the latter are probably protected by
a colloidal membrane, impermeable to soluble substances of the same physical con¬
stitution.” (P. 220.)
aA tendency to spontaneous change, which is observed occasionally in crystal¬
loids, appears to be general in the other class. The fluid colloid becomes pectous and
insoluble by contact with certain other substances, without combining with these
substances, and often under the influence of time alone. The pectizing substance
appears to hasten merely an impending change. Even while fluid a colloid may alter
sensibly, from colorless becoming opalescent; and while pectous the degree of hydration
may become reduced from internal change. The gradual progress of alteration in the
colloid effected by the agency of time, is an investigation yet to be entered upon.
“The equivalent of a colloid appears to be always high, although the ratio between
the elements of the substances may be simple. Gummic acid, for instance, may be
represented by C^HnOn, but judging from the small proportions of lime and potash
which suffice to neutralize this acid, the true numbers of its formula must be several
times greater. It is difficult to avoid associating the inertness of colloids with* their
high equivalents, particularly where the high number appears to be attained by the
repetition of a smaller number. The inquiry suggests itself whether the colloid
molecule may not be constituted by the grouping together of a number of smaller
crystalloid molecules, and whether the basis of colloidality may not really be this
composite character of the molecule.
“With silicic acid, which can exist in combination both as a crystalloid and
colloid, we have two series of compounds, silicates and cosilicates, the acid of the
latter appearing to have an equivalent much greater (36 times greater in one salt)
than the acid in a variety of metallic salts, such as certain red salts of iron, is accounted
for by the high colloidal equivalent of their bases. The effect of such an insoluble
colloid as Prussian blue in carrying down small proportions of the precipitating
salts, may admit of a similar explanation. (P. 221.)
“ The*hardness of the crystalloid, with its crystalline planes and angles, is replaced
in the colloid by a degree of softness, with a more or less rounded outline. The water
of crystallization is represented by the water of gelatination. The water in gelatinous
hydrates is aptly described by M. Chevreul as retained by 'capillary affinity/ that is,
by an attraction partaking both of the physical and chemical character. J (P. 222.)
“On the Properties of Silicic Acid and Other Analogous Colloidal Substances.—
The prevalent notions respecting solubility have been derived chiefly from obser¬
vations on crystalline salts, and are very imperfectly applicable to the class of
colloidal substances. Hydrated silicic acid, for instance, when in the soluble
condition, is, properly speaking a liquid body, like alcohol, miscible with water
in all proportions. We have no degrees of solubility to speak of with respect to
silicic acid, like the degrees of solubility of a salt, unless it be with reference to
silicic acid in the gelatinous condition, which is usually looked upon as destitute
of solubility. The jelly of silicic acid may be more or less rich in combined water,
as it is first prepared, and it appears to be soluble in proportion to the extent of
its hydration. A jelly containing 1 per cent of silicic acid, gives .with cold water
a solution containing about 1 of silicic acid in 5,000 water; a jelly containing
about 5 per cent of silicic acid gives a solution containing about 1 part of acid in
10,000 water. A less hydrated jelly than the last mentioned is still less soluble;
and finally, when the jelly is rendered anhydrous, it gives gummy-looking, white
masses, which appear to be absolutely insoluble, like the light dusty silicic acid
obtained by drying a jelly charged with salts, in the ordinary analysis of a
silicate.
“The liquidity of silicic acid is only affected by a change, which is permanent
(namely, coagulation or pectization), by which the acid is converted into the gelatinous
or pectous form, and loses its miscibility with water. The liquidity is permanent in
proportion to the degree of dilution of silicic acid, and appears to be favored by a
low temperature. It is opposed, on the contrary, by concentration, and by elevation
of temperature. A liquid silicic acid of 10 or 12 per cent pectizes spontaneously in a
few hours at the ordinary temperature, and immediately when heated. A liquid of
5 per cent may be preserved for 5 or 6 days; a liquid of 2 per cent for 2 or 3 months;
and a liquid of 1 per cent has not pectized after 2 years. Dilute solutions of 0.1 per
cent or less are no doubt practically unalterable by time and hence the possibility
of soluble silicic acid existing in nature. I may add, however, that no solution,
weak or strong, of silicic acid in water has shown any disposition to deposit crystalsy
but always appears on drying as a colloidal glassy hyalite. The formation of quartz
crystals at a low temperature, of so frequent occurrence in nature, remains still a
mystery. I can only imagine that such crystals are formed at an inconceivably slow
rate and from solutions of silicic acid which are extremely dilute. Dilution no doubt
weakens the colloidal character of substances, and may therefore allow their crystal¬
lizing tendency to gain ground and develop itself, particularly where the crystal once
formed is completely insoluble, as with quartz.
“ The pectization of liquid silicic acid is expedited by contact with solid matter in.
the form of powder. By contact with pounded graphite, which is chemically inactive,
the pectization of a 5 per cent silicic acid is brought about in an hour or two, and that
of a 2 per cent silicic acid in 2 days. A rise of temperature of 1.1 °C. was observed
during the formation of the 5 per cent jelly.
“The ultimate pectization of silicic acid is preceded by a gradual thickening in

the liquid itself. The flow of liquid colloids through a capillary tube is always slow
compared with the flow of crystalloid solutions, so that a liquid-transpiration-tube
may be employed as a colloidoscope. With a colloid liquid alterable in viscosity,
such as silicic acid, the increased resistance to passage through the colloidoscope is
obvious from day to day. Just before gelatinizing, silicic acid flows like an oil.
“A dominating quality of colloids is the tendency of their particles to adhere,
aggregate, and contract. This idio-attraction is obvious in the gradual thickening
of the liquid, and when it advances leads to pectization. In the jelly itself, the
specific contraction in question, or synaeresis, still proceeds, causing separation of
water, with the division into a clot and serum; and ending in the production of a hard
stony mass, of vitreous structure, which may be anhydrous, or nearly so, when the
water is allowed to escape by evaporation. The intense synaeresis of isinglass
dried in a glass dish over sulphuric acid in vacuo enables the contracting gelatin to tear
up the surface of the glass. Glass itself is a colloid, and the adhesion of colloid to
colloid appears to be more powerful than that of colloid to crystalloid. The gelatin,
when dried in the manner described upon plates of calcspar and mica, did not adhere
to the crystalline surface, but detached itself on drying. Polished plates of glass
must not be left in contact, as is well known, owing to the risk of permanent adhesion
between their surfaces. The adhesion of broken masses of glacial phosphoric acid to
each other is an old illustration of colloidal synaeresis.
“Bearing in mind that the colloidal phasis of matter is the result of a peculiar
attraction and aggregation of molecules, properties never entirely absent from matter,
but more greatly developed in some substances than in others, it is not surprising
that colloidal characters spread on both sides into the liquid and solid conditions.”
(Pp. 335, 336.)
Graham goes on to say that certain fluid substances which exercise no pectizing
influence on colloids, can totally or partially displace the combined water of colloidal
silicic acid, without causing the coagulation of the colloid. Hydrochloric, nitric,
acetic, and tartaric acids, sugar syrup, glycerin, and alcohol are in this class. He then
describes the production of such a solution of silicic acid in alcohol and the alcogel.
“The alcogel, or solid compound, is readily prepared by placing masses of gelatinous
silicic acid, containing 8 or 10 per cent of the dry acid, in absolute alcohol, and changing
the latter repeatedly till the water of the hydrogel is fully replaced by alcohol. The
alcogel is generally slightly opalescent, and is similar in aspect to the hydrogel,
preserving very nearly its original bulk. The following is the composition of an
alcogel carefully prepared from a hydrogel which contained 9.33 per cent of silicic
acid.
Alcohol. 88.13
Water. 0.23
Silicic acid. 11.64
100.00
“Placed in water, the alcogel is gradually decomposed—alcohol diffusing out and

water entering instead, so that a hydrogel is reproduced. (Pp. 337, 338.)
“The compound of sulphuric acid, sulphagel, is also interesting from the facility
of its formation, and the complete manner in which the water of the original hydrogel
is removed. A mass of hydrated silicic acid may be preserved unbroken if it is first
placed in sulphuric acid diluted with two or three volumes of water, and then trans¬
ferred gradually to stronger acids, till at last it is placed in concentrated oil of
vitriol. The sulphagel sinks in the latter fluid, and may be distilled with an excess
of it for hours without losing its transparency or gelatinous character. It is always

somewhat less in bulk than the primary hydrogel, but not more, to the eye, than one-
fifth or one-sixth part of the original volume. This sulphagel is transparent
and colorless. When a sulphagel is heated strongly in an open vessel, the last portions
of the monohydrated sulphuric acid in combination are found to require a higher
temperature for their expulsion than the boiling-point of the acid. The whole silicic
acid remains behind, forming a white, opaque, porous mass, like pumice. A sulphagel
placed in water is soon decomposed, and the original hydrogel reproduced. No
permanent compound of sulphuric and silicic acids, of the nature of a salt, appears to
be formed in any circumstances. A sulphagel placed in alcohol gives ultimately a
pure alcogel. Similar jellies of silicic acid may readily be formed with the mono¬
hydrates of nitric, acetic, and formic acids, and are all perfectly transparent.
“The production of the compounds of silicic acid now described indicates the
possession of a wider range of affinity by a colloid than could well be anticipated.
The organic colloids are no doubt invested with similar wide powers of combination,
which may become of interest to the physiologist. The capacity of a mass of gelatin¬
ous silicic acid to assume alcohol, or even olein, in the place of water of combination,
without disintegration or alteration of form, may perhaps afford a clue to the pene¬
tration of the albuminous matter of membrance by fatty and other insoluble bodies,
which seems to occur in the digestion of food. Still more remarkable and suggestive
are the fluid compounds of silicic acid. The fluid alcohol or olein compound favors
the possibility of the existence of a compound of the colloid albumen with olein,
soluble also and capable of circulating with the blood.
“The feebleness of the force which holds together two substances belonging
to different physical classes, one being a colloid and the other a crystalloid, is a
subject deserving notice. When such a compound is placed in a fluid the superior
diffusive energy of the crystalloid may cause its separation from the colloid. Thus,
of hydrated silicic acid, the combined water (a crystalloid) leaves the acid (a colloid)
to diffuse into alcohol; and if the alcohol be repeatedly changed, the entire water is
thus removed alcohol (another crystalloid) at the same time taking the place of water
in the combi,nation with the silicic acid. The liquid in excess (here the alcohol)
gains entire possession of the silicic acid. The process is reversed if an alcogel be
placed in a considerable volume of water. Then alcohol separates from combination,
in consequence of the opportunity it possesses to diffuse into water; and water, which is
now the liquid present in excess, recovers possession of the silicic acid. Such changes
illustrate the predominating influence of mass.
“Even the compounds of silicic acid with alkalies yield to the decomposing force
of diffusion. The compound of silicic acid with 1 or 2 per cent of soda is a
colloidal solution, and, when placed in a dialyzer over water in vacuo to exclude car¬
bonic acid, suffers gradual decomposition. The soda diffuses off slowly in the caustic
state, and gives the usual brown oxide of silver when tested with the nitrate of that
base.”
Many other colloids are dealt with by Graham in these two papers, including gum,
dextrin, caramel, albumen, alumina; stannic, tungstic, titanic and molybdic acids,
etc. Contrary to what is commonly believed, Graham was fully aware that colloids
dialyze to some extent. Thus he expressly states that sodium chloride dialyzes
about 200 times faster than tannic acid, about 400 times faster than gum arabic, and
about 1,000 times faster than pure albumen prepared according to the method of
Wurtz.
Classification of Colloids.—Since a colloid results from the subdivision, within
limits, of a substance in a medium, the broadest classification of colloids is that
proposed by Wolfgang Ostwald1 who grouped them according to the physical
i Kolloid Zeil.. Vol. 1. d. 291.
state (gaseous, liquid or solid) of the subdivided substance (dispersed phase), and
of the medium in which the particles are distributed (dispersion medium).
Table 1 shows the nine groups, and gives some instances of each.
Table 1
Dispersion
Disperse phase Example
medium
Gas. Gas. No example since gases are miscible in all

proportions.
Gas. Liquid. Fine foam, gas in beer.
Gas. Solid. Gaseous inclusions in minerals (meerschaum,
pumice), hydrogen in iron, oxygen in silver.
Liquid. Gas. Atmospheric fog, clouds, gases at critical
state.
Liquid. Liquid. Emulsions of oil in water, cream; colloidal
water in chloroform.
Liquid. Solid. Mercury in ointments, water in paraffin
wax. Liquid inclusions in minerals.
Solid. Gas. Cosmic dust, smoke, condensing vapors
(ammonium chloride).
Solid. Liquid. Colloidal gold, colloidal sodium chloride,
colloidal ice in chloroform.
Solid. Solid. Solid solutions, colloidal gold in ruby glass,
coloring matter in gems.
This classification is more theoretical than practical, for the properties of colloids
are dependent mainly upon the specific nature of the dispersed substance, its degree
of subdivision, and the presence of other substances, which though present in minimal
quantities may produce potent stabilizing or coagulative effects. It is not possible
to classify colloids according to the size of the dispersed phase, because most colloidal
solutions contain particles of different sizes.
The most acceptable classification is that suggested by Hardy1 and further
developed by Zsigmondy, whereby colloids are divided into two broad classes, depend¬
ing upon their behavior upon dessication at ordinary temperatures—the reversible
colloids which redissolve, and the irreversible colloids which do not. This classifi¬
cation is illustrated by Table 2, taken from Zsigmondy,2 which gives some instances of
each group. The varying stretch of the names over the horizontal lines, which
indicate dimensions, corresponds to the approximate size-limits of the dispersed
particles in each substance.
It is interesting to note that the addition of a little reversible colloid, may render
an irreversible colloid reversible (protective action). Furthermore no sharp line is
to be drawn between reversible and irreversible colloids, for besides intermediate or
transition cases between the two classes, there may be recognized two groups of
irreversible colloids, roughly defined by their behavior upon concentration: (1) The
completely irreversible, which coagulate while still quite dilute and separate sharply
from the solvent with the formation of a pulverulent precipitate, rather than a gel
(i.e., pure colloidal metals). Chemical or electrical energy is needed to bring them
back again into colloidal solution. (2) The incompletely reversible, which when quite
1 Proc. Roy, Soc.r Vol. 66, p. 95, (1900).
2 “Coloids and the Ultramicroscope,” J. Wiley and Sons, p. 26.
Table 2
O.ljUjU 1/XJtl 10/I/X lOO/^/X lii 10m 100m 1mm.
I 1 n 1 1 1 l
Microscopic field,
13 ltramicroscopic field i
Real forms of particles may be seen
Quartz in solution does not form a | QuarU golution fomis a sediment

sediment ;
Oil emulsions in water do not separate into i Oil emulsions in water separate into two
two layers 1 layers
Particles pass through filter paper ! Particles may be filtered out
Very rapid Brownian movement Slow None
i Hydrosols j
Crystalloidal icolloidal j
solutions isolutions Turbidities Suspensions
Table 2A
Reversible Hydrosols Irreversible Hydrosols
Lea’s colloidal silver Colloidal metals (usually)
Colloidal indigo Bredig’s colloidal gold
PaaPs blue colloidal gold Colloidal sulphides
Glycogen solution Many colloidal sulphides
Soluble starch Purple of Cassius
Crystallized albumen
Dextrin
Molybdic acid
concentrated, form a gel which may be easily redissolved or peptisized by compara¬

tively small amounts of reagents, unless the evaporation has proceeded too far (i.e.,
colloidal stannic acid).
Size of Colloidal Particles.—In order to give some idea as to the size of colloidal
particles as compared with other dimensions of the same order, the table on
p. 775 has been prepared.
The size of colloidal particles may be determined with the aid of the ultramicro¬
scope, by counting the number of particles in a definite volume of a highly diluted
colloidal solution of known concentration, or by measuring the mean distance between
the particles.1
Let d — average diameter of particles,
s = specific gravity of dispersed substances,
a = mass of dispersed substance,
n = number of particles, and
r — mean distance between particles.
1 Siedentopf and Zsigmondy, “Drude’s Annual,” Vol. 10, p. 30-39, 1903. Zsigmondy, 4 ‘Colloids
and the Ultramicroscope,” p. 117.
Then in the first case
And in the second case
These formulas are based on the assumption that all the particles of the dispersed
substance are actually counted, a condition which is practically never realized. Some
particles are amicroscopic (i.e., too small to be seen in the ultramicroscope); some
may be reduced to molecular dispersion; and even at the high illumination at which
the counts should be made, there is the possibility of the superior brilliancy of large
particles obscuring the presence of smaller ones. The error thus introduced does not
exceed 20 per cent, and is usually considerably less.
Table 3
0. 1mm Theoretical size of hydrogen molecule.

0. 5mm Theoretical size of methyl alcohol molecule.
0.8mm Theoretical size of chloroform molecule.
1—0.02mm Thickness of electrical double layer.
1.7mm Size of small particles in colloidal gold solutions.
3— 1mm Thickness of metal films which polarize platinum.
4— 3mm Mean distance between centers of nearest molecules in gases at
760 mm. and 0°C.
5mm Approximate size of molecule of soluble starch.
Approximate limit of ultramicroscopic visibility of individual par¬
ticles in colloids.
10—5mm Thickness of permanent water film on glass at 23 °C.
12mm Thickness of black soap films.
Action of silver plate on phase of reflected light alters.
Range of unstable thickness ends.
40mm Wave length of mean violet light.
50mm Wave length of mean green light.
60mm Wave length of mean red light.
50mm Approximate value of p, the radius of molecular attraction.
60mm Gold hydrosols having this average particle size are unstable.
75mm Gold particles settle out.
96—45mm Range of unstable thickness begins.
90mm Thickness of gold leaf.
100mm Approximate lower limit of size of particles in suspensions.
150mm Limit of resolving power of microscope objective of 1.4 numerical
aperture.
500—2,000mm Diameter of various spherical bacteria (cocci).
(0.5—2m)
4,000—15,000mm Length of anthrax bacillus (width about 1,000mm)-
(4—15m)
The size of colloidal particles may also be estimated by means of ultrafiltration.
An ultrafilter is essentially a layer of a jelly properly supported, through which a
colloidal solution may be forced by appropriate pressure. By using jellies of different
kinds and concentrations, and layers of varying thickness, ultrafilters may be pre¬
pared having varying degrees of permeability to colloidal particles. Four methods
of gauging ultrafilters are given by Bechhold,1 who also prepared the following table
1 “ Colloids in Biology and Medicine” (trans. by Bullowa) D. Van Nostrand Co., p. 90.
(Table 4) showing the relative sizes in decreasing order of size of the particles of
various colloids, as determined by using ultrafilters of different degrees of porosity.
Table 4
Suspensions
Prussian blue Hemoglobin solution, 1 per cent molecular
weight about 16,000
Platinum hydrosol (Bredig) Serum albumin, molecular weight 5,000
to 15,000
Ferric oxide hydrosol Diphtheria toxin
Casein, in milk Protalbumoses
Arsenic sulphide hydrosol Colloidal silicic acid
Gold hydrosol No. 4 (Zsigmondy) about Lysalbinic acid
40mm
Bismon, colloidal bismuth oxide (Paal) Deuteroalbumose A
Deuteroalbumose B, molecular weight
about 2,400
Lysargin, colloidal silver (Paal) Deuteroalbumose C
Collargol, colloidal silver von Heyden, Litmus
20mm
Gold hydrosol No. 0 (Zsigmondy) about Dextrin, molecular weight about 965.
1 tO 4 fJLfX
Gelatin solution, 1 per cent Crystalloids
The XJltramicroscope.—Since this instrument revolutionized colloid research,

a brief description of it is essential. A discussion of the scientific principles
involved in ultramicroscopy together with the prior history of microscopy leading
to its development are given by E. F. Burton.1
The ultramicroscope consists essentially of a compound microscope arranged so
as to examine against a black background (dark field), objects powerfully illuminated
by an intense beam of light at right angles (or approximately at right angles) to the
optical axis of the microscope. If a particle diffracts sufficient light to affect the
retina, it becomes visible against the dark background, just as the motes in the air
become visible when a sun beam enters a dark room. The efficiency of the instru¬
ment depends upon the intensity of the illuminating beam and the elimination of
extraneous light from the dark field.
Just as in the cosmic field even the most powerful telescope fails to resolve the
fixed stars, which are nevertheless visible as points of light, so too in the ultramicro-
scopic field tiny colloidal particles are visible, although we may never hope to resolve
them, since they are smaller than a wave length of light itself. The limits of visibility
of particles in the ultramiscroscope depend upon (1) the specific light-reflecting capac¬
ity of the particles, (2) their difference in refractive index from that of the medium
surrounding them, (3) the intensity of the illuminating beam, (4) the darkness of
the field. Large particles in the field, because of their superior luminosity, obscure
the presence of smaller ultramicrons, just as the full moon renders the smaller stars
invisible to the naked eye. If too large a number of small particles is present there
results only an illuminated field—a miniature Milky Way. Particles below micro¬
scopic range, but visible in the ultramicroscope are called ultramicrons; particles
invisible even in the ultramicroscope are called amicrons. Dyes like uranin, fluorescin
and eosin, even in extremely dilute solution, show an illuminated field, which may
1 “The Physical Properties of Colloidal Solutions.” Longmans, Green and Co., 1916. See also
“Colloids and the UltramicroscoDe.”
COLLOID CHEMISTRY Ui
be apparently due to a large number of amicroscopic particles, although it Ls< ommonly

ascribed to fluorescence.
Preparation of Colloidal Solutions.—Some substances are inherently colloidal

(gelatin, gumarabic, egg albumen) and readily yield hydrosols on being dissolved
in water. As Svedberg has observed the other methods of prepanng colloidal
solutions fall naturally into two classes: (1) condensation methods, vhereby the
colloidal state is reached by the aggregation or building up of molerularly dis¬
persed particles to colloidal dimensions; (2) dispersion methods, whereby larger
particles are reduced down to colloidal state. A few instances ol each class will
be given.
Condensation Methods.—(a) Colloidal Gold Hydrosol (Faraday, Zsigmondy). Dis¬

solve 15 mg. of gold hydrochloride (AUHCI4.3H2O) and 37 mg. of the purest potassium
carbonate in 120 c.c. of conductivity water (re-distilled through a silver worm). "Hum
add 0.5 c.c. of an ethereal solution of yellow phosphorus, made by diluting a concen¬
trated solution with five times its volume of ether. The hydrosol forms slowly,
passing through a yellowish-brown to a bright ruby-red. Traces of impurilies found
in ordinary distilled water, especially alkaline earth phosphates, silicates, etc., are
apt to mar the results, and lead to the formation of bluish or violet suspensions of
gold, which settle out. Ordinary tap water will yield fair results, however, if there
first be added to it a few drops of a good protector, such as warm gelatin solution or
gum arabic. In this case the potassium carbonate may be dispensed with.
([b) Colloidal Arsenious Sulphide.—This is readily prepared by running a solution
of arsenious acid into sulphuretted hydrogen water, kept saturated by a stream of
the gas.
(c) Colloidal Resins, Oils, etc.—These may be first reduced to crystalloids! dis¬
persion by solution in alcohol, and the alcoholic solution diluted with water sufficiently
to cause aggregation to colloidal dimensions. Thus alcoholic mastic solution when
diluted yields a colloidal dispersion. In preparing perfumed alcohol for shaving,
it is usual to make a solution of essential oils in alcohol, which upon proper dilution
with water, gives a cloudy but stable colloidal emulsion.
Dispersion Methods.—(a) By prolonged grinding or agitation, especially in the
presence of protective or deflocculating substances, many substances become reduced,
in part at least, to the colloidal state. The grinding of ultramarine and of ores, the
preparation of Chinese or india ink, the homogenizing of milk, and the “freeing out”
of clay, may be mentioned as examples.
(b) Electrical Atomization.—Bredig1 first applied this method to the manufacture
of stable colloidal solutions of metals. He caused an electric are (currant 4 to 10
amp., potential difference at arc 30 to 300 volts) to pass between electrodes of chemic¬
ally pure metals, below the surface of distilled water contained in n clean vessel of
very insoluble glass or porcelain. It is essential that there be no mixture of electro¬
positive and electronegative metals in the electrodes, and an intermittent sparking
of the arc produced by a mechanically operated circuit breaker m desirable, by the
use of oscillating discharges at from 30 to 200 volts from an induction coil, Hvedberg
reduced the current density to a fraction of that required by Bredig (from 1 or 2
amp. down to about 10 mil-amp.); and he was thus enabled to produce a large
number of hydrosols and organosols.
(c) Chemical Dispersion.—Many precipitates may be brought into colloidal dis¬
persion by the addition of small quantities of proper reagents which may then he
1 Bmdig—“Anorganische Fermente” Faraday had previously xmml the electric arc to nr<«t»c<i
finely dispersed gold. *
dialyzed out; ferric hydroxide by small quantities of ferric chloride or hydrochloric

acid, copper ferrocyanideby ammonium oxalate.
I cannot do better than to quote here the pregnant words of Graham: “The
pectization of liquid silicic acid and many other liquid colloids is effected by contact
with minute quantities of salts in a way which is not understood. On the other hand,
the gelatinous acid may again be liquefied, and have its energy restored by contact
with a very moderate amount of alkali. The latter change is gradual, 1 part of caustic
soda, dissolved in 10,000 water, liquefying 200 parts of silicic acid (estimated dry)
in 60 min. at 100°C. Gelatinous stannic acid is also easily liquefied by a small pro¬
portion of alkali, even at ordinary temperature. The alkali, too, after liquefying
the gelatinous colloid, may be separated again from it by diffusion into water upon
a dialyzer. The solution of these colloids, in such circumstances, may be looked
upon as analogous to the solution of insoluble organic colloids witnessed in animal
digestion, with the difference that the solvent fluid here is not acid, but alkaline.
Liquid silicic acid may be represented as the ‘peptone’ of gelatinous silicic acid; and
the liquefaction of the latter by a trace of alkali may be spoken of as the peptization
of the jelly. The pure jellies of alumina, peroxide of iron, and titanic acid, prepared
by dialysis, are assimilated more closely to albumen, being peptized by minute quan¬
tities of hydrochloric acid.”
Consequences of Subdivision.—The most striking results of the subdivision
of a substance to colloidal dimensions, follow from the enormous development of
free surface. Table 5, adapted from Wo. Ostwald, shows the interfacial develop¬
ment resulting from the progressive decimal subdivision of a cube.
Table 5.—Results of the Progressive Decimal Subdivision of a Cube
Length of edge Number of cubes | Total surface

1
1.0 cm. 1 6 sq. cm. or 0.93 sq. in.
1.0 mm. 103 60 sq. cm. or 9.3 sq.in.
0.1 mm. i 106 600 sq. cm. or 93.0 sq. in.
0.01 mm. 109 6,000 sq. cm. or 6.46 sq. ft.
1.0 M ! 1012 6 sq. m. or 64.6 sq. ft.
0.1 M 1015 60 sq. m. or 645.97 sq. ft.
0.01 M 1018 600 sq. m. or 6,459.72 sq. ft.
1.0 MM 1021 6,000 sq. m. or 1.48 acres
0.1 MM 1024 6 hectares or 14.83 acres
0.01 MM
1Q27 60 hectares or 148.3 acres
o.oow 1030 6 sq. km. or 2.32 sq. mi.
From this it is obvious that surface forces such as surface tension, adsorption,
capillarity, and even chemical attraction become enormously magnified and of trans¬
cendent importance. Thus colloidal Sulphur
reduces silver salts energetically, whereas even fine
precipitated sulphur does not form silver sulphide
in the cold, and does so only partially upon boiling.
Surface Tension.—Figure 1 represents a

/////////////^sy?//////////' droplet of water on an oiled surface. Omitting
from consideration such disturbing factors as
■Fig* lm gravitation, evaporation, etc., the following
main forces are at work: (1) the attraction of the water molecules for each other;
(2) the attraction between the air and the water; (3) the attraction between
the air and the oiled surface; (4) the attraction between the water and the oiled
surface.
Regarding for the moment only the water and the air, the attraction of the water
molecules for each other tends to make the droplet assume the shape exposing a
minimum surface, i.e., a sphere. In considering this sphere, it is evident that the
molecules on its surface occupy an anomalous position; they are attracted inwardly
by the relatively strong attraction of the interior water molecules, and outwardly
by the relatively weak attraction of the air. The residual attraction in the surface
layer results in the formation of a surface skin, and the strength of this surface film
is called the surface tension of the water against air.1 The surface tension is desig¬
nated by <i and may be determined in a variety of ways, i.e., by the force required
to produce distorsion of the fluid surface, by the capillary rise of the fluid, or by meas¬
uring the maximum weight attained by a drop of the fluid falling from a tube.2
The work needed to increase the surface of the water drop is the product of
the surface tension by the surface area (o>) i.e. surface energy = a.co, or <r =
surface energy dynes
expressed in The effect of gravitation on larger droplets of
CO
water is to increase their area by flattening them into oblate spheroids.

Obviously if for air we substitute another surrounding medium, i.e., oil, or if for
water we substitute another droplet, i.e. mercury, the value of o- will change. Table
VI gives a few values of <r.3
Opposed to surface tension is what Wo. Ostwald4 terms “ expansive surface energy”
whose intensity is the so-called ‘‘internal pressure.” This apparently depends upon
the balance between the various forces acting on the molecules within the droplet,
for it is increased by temperature and electrification. It is in effect the net internal
energy of the droplet, and is one of the factors involved in solution. In fact solution is
a condition of equilibrium reached between the internal forces of the solute and those
of the solvent, and the mutual attraction between the solvent and solute. This
equilibrium may be reached before true crystalloidal solution is effected; i.e., tannic
acid goes into crystalloidal dispersion in alcohol, but into colloidal dispersion in
water.
Table G.—Various Values of Surface Tension (<r)

Dynes per square centimeter
Mercury/air. 436.0 Benzol/water. 32.6
Water/air. 75.0 Chloroform/water. 27.7
Glycerin/air. 65.0 Olive oil/water. 22.9
Glue solution/air. 48.3 Rosin/water. 19.9
Olive oil/air. 32.7 Oil of turpentine/water. 12 4
Benzol/air. 28.8 Ethyl ether/water. 9.69
Chlorof orm /air. 26.0 Isobutyl alcohol/water. 1.76
Ethyl alcohol/air. 22.0 Ethyl alcohol /olive oil. 2.26
Ethyl ether/air. 16.5 Olive oil/castile soap (1 to 4,000)... 3.65
Returning now to our water drop on the oiled surface, Fig. 1, it may be seen there
is a three-phase system involved—water/air/oil; and the following three interfacial
tensions are involved: <r water/air, <r water/oil, and a oil/air. Since: <r water/air >
75.
■’•Atmospheric constituents are also condensed on the surface.
2 J. L. R. Morgan’s drop-weight instrument, or Taube’s staglamometer.
8 The figures are approximate only, as cr varies with the temperature.
cr water/oil + cr oil/air, the water droplet maintains its spherical shape. If, however,
22.9 32.7
a drop of oil be placed on water, it spreads itself out over the surface of the water,
because: cr oil/air < cr water/air — cr water/oil. Thus the interplay of surface
32.7 75. 22.9
tension determines the distribution of fluids over interfaces, a fact of technical and
especially biological importance. Glue or gum arabic dissolved in water, enable it
to emulsify oil readily. In fact the lowering of the surface tension between phases
is a forerunner of mixing or even solution; and as Willard Gibbs has demonstrated,
substances which lower surface tension between two phases, tend to collect at their
interface. This introduces a new idea—adsorption. Adsorption is essentially the
concentration1 of a dispersed substance at the interface between two phases. The
power possessed by bone black to decolorize sugar solutions, of fullers earth to bleach
oils, of platinum or palladium black to condense gases (i.e., in automatic gas lighters),
of charcoal to purify air (in gas masks during the recent war), are all examples of
practical uses of adsorption; and a consideration of the adsorbents shows that they
all have greatly extended surface.
The underlying cause of adsorption is found in the anomalous position of the sur¬
face atoms, previously referred to, they are attracted in three directions by atoms of
like kind, whereas in the fourth direction they are in contact with the molecules of
another substance, towards which they exhibit a residual surface attraction. Where
one phase is solid and the other is fluid, the molecules of the fluid, and especially of
substances dispersed in it, are condensed at the surface of the solid; and the most
obvious examples of adsorption are found in this category. It is, however, an im¬
portant factor in emulsions (fluid/fluid), resulting in the formation of emulsostatic
surface films.
While adsorption has been commonly classed as a “physical” force, its depend¬
ence on chemical composition is marked, as is evident from the experiments of
Bechhold,2 who tested the adsorptive action of a series of chemically related com¬
pounds on acid and basic aniline dyes. The results were as follows:
Table 7
Substance Acid dyes Basic dyes
Naphthalin (CiJEIs) neutral.• • Faintly stained Faintly stained

/SNaphthol (C10H7.OH) acid. Faintly stained Strongly stained
/SNaphthylamin (C10H7.NH2) basic. Strongly stained Very faintly stained
Amidonaphthol (C10H7OH.NH2) amphoteric. Stained Strongly stained
Adsorption is frequently the precursor of chemical action. The adsorbed sub¬

stance may diffuse into the adsorbent and there undergo irreversible changes, or it
may by its mere concentration at the surface initiate chemical reactions. Catalyzers
like platinum black and enzymes generally, are colloidal, and adsorption plays an
important part in their activities. Frequently the adsorbed substance is coagulated
at the interface, with the formation of a surface skin or pellicle. This is evident in
solutions of dyes like fuchsin and methyl violet; it is responsible for the inactivation
of enzymes by shaking, and the separation of albumin from aqueous solution in the
foam produced by shaking with benzol or chloroform.
1 In negative adsorption the reverse occurs, probably because the solvent is more strongly attracted
than the solute.
2 “Colloids in Biology and Medicine,” p. 29. The work of Langmuir and others indicates that
adsorption is the result of residual chemical forces.
Swelling is another common property of colloids, which is intimately connected

with their enormous development of free surface, and with surface tension. Colloids
are swollen only by substances which penetrate them; thus gelatin swells in water
but not in benzol, whereas vulcanized rubber swells in benzol but not in water. The
vapor of a solvent, being condensed by adsorption at the surface, acts like the liquid
itself—gelatin swells in water vapor, and rubber in benzol vapor. The force with
which fluids are absorbed by dry colloids is enormous. An old-fashioned method
of splitting rocks is to drive in and then moisten baked wooden wedges. According to
H. Rodewald starch exerts a swelling pressure of 2,523 atmospheres. The swelling
is consequent upon the attraction of the solvent to the interfaces between the particles.
The swelling of gelatine, fibrin, etc., is markedly influenced by electrolytes, especially
by dilute acids and alkalis which strongly increase it. Apparently the hydrogen or
hydroxyl ions are selectively adsorbed, and their electric charge tends to force the
particles further apart. If the concentration is too great, however, the coagulative
or “salting out” effect of electrolytes predominates, and shrinking occurs.1 On the
other hand, if gelatine is super dried at 130°C., it practically loses its swelling capacity,
apparently because its particles are brought into such close proximity by the loss of
their surrounding films of water, that they cannot be readily separated again.
Optical Properties of Colloids.2—The most striking optical characteristic

of colloidal solutions is turbidity or cloudiness, which is due to the scattering of
fight by their small dispersed particles. If a powerful convergent beam of fight
be passed through a vessel containing water carefully redistilled through a silver
worm, nothing is seen in the water on looking against a dark background at right
angles to the beam. Upon mixing with the water a few drops of colloidal gold
solution, the path of the light through the water at once becomes visible as a
luminous cone. The same effect is seen on a large scale when a search fight plays
at night through a smoky or foggy atmosphere. This method was first used by
Faraday (1857) and later by Tyndall, to demonstrate the optical inhomogeneity
of colloidal solutions, and is known as the Faraday-Tyndall effect or Tyndall
phenomenon. It might be considered a criterion of colloidal solution were it
not for the fact that traces of colloidal impurities produce it, and it is to some
extent shown by concentrated solutions of some crystalloid sugars (glucose,
raffinose).3
The fight scattered by colloidal particles is polarized, as may be seen by examining
it with a Nicol prism or a selenite plate. Furthermore colloidal particles absorb,
reflect, and transmit various fractions of white fight, so that they frequently appear
strongly colored. The color of colloidal gold solutions to transmitted fight runs from
yellowish and ruby red for the smaller particles, to purplish red, blue and violet or
even brown, for the larger ones. In fact with colloids generally, as the size of the
particles increases, the fight absorption maxima shift towards the yellow end of the
spectrum; or in other words the smaller particles show a brighter or more yellow
shade. Rayleigh has shown that the blue color of the sky is due to the faint scattering
of fight by the molecules of the air, which is evident to us only when we look through
a great depth of the atmosphere. Similarly distant mountains appear blue, especially
on a hazy day when the atmosphere is full of colloidal particles of water, dust or smoke.4
It is interesting to note that the apparent colors of many organized structures, are
largely influenced by or entirely due to their colloidal nature (i.e., feathers, mother of
pearl, skins of fish.5 Blue feathers usually owe their color, not to a blue pigment,
1 Proctor and Wilson and J. Loeb believe that gelatin forms hydrolyzable salts and that its
behavior is explainable on the basis of Downan’s theory of membrane equilibrium.
2 For fuller discussion, see E. F. Burton, “The Physical Properties of Colloidal Solutions,” Chap. 5.
3 Lobby db Brtjyn and Wolff, Rec. trav. chim. des Pays Bas, Vol. 23, p. 155, 1904.
but to the reflection of the blue light waves with the concomitant absorption of the
other waves. If the reflected blue light passes through a thin layer of yellow pigment,
the feather looks green. Certain green tree toads turn blue when preserved in alcohol,
which dissolves out the droplets of yellow oil from their skins.
Electric Properties of Colloids.—The particles of practically all colloidal

solutions possess an electric charge, and under the influence of an electric current
(difference of potential) move toward the electrode having the opposite charge.
(Electrophoresis.) In general, when two substances are brought into contact,
the one having the higher dielectric constant becomes positively charged, whereas
the one with the lower dielectric constant becomes negatively charged (Coehn’s
Law). Since water has a high dielectric constant (80), most substances sus¬
pended in pure water become negatively charged and wander to the anode.
On the other hand if suspended in oil of turpentine which has a low dielectric
constant (2.23) they become positively charged and wander to the cathode.
If, however, electrolytes are present, Coehn’s law is superseded by other controlling
factors such as the adsorption of ions which may give their charge to the suspended
particles. In fact Hardy found that in pure water albumen was amphoteric; in the
presence of a trace of alkali it acquired a negative charge and migrated to the anode;
but a trace of acid gave it a positive charge and it then migrated to the cathode. The
following table shows the usual charge and migration tendency of a number of aqueous
colloidal solutions.
Table 8
Charged +• Charged —
Migrate to Cathode (— Pole) Migrate to Anode ( + Pole)
1. Hydrates of Fe, Cu, Cd, Al, Zr, Ce, Th. 1. Sulphides of As, Sb, Cu, Pb, Cd
Halides of Ag.
2. Titanic Acid. 2. Stannic Acid, Silicic Acid.
3. Colloidal Bi, Pb, Fe and Cu (Bredig’s 3. Colloidal Pt, Au, Ag, and Hg, I, S, Se.
Method).
4. Albumen, Hemoglobin, Agar. 4. Gum arabic, soluble starch, gamboge,
mastic oil emulsion.
5. Basic Dyes: Methyl Violet, Bismarck 5. Acid Dyes: eosin, fuchsin, aniline
Brown, Methylene Blue, Hofman blue, indigo, soluble Prussian blue.
violet.
Pectization and Peptization.—Briefly stated pectization means the coagula¬

tion of a colloidal gel and peptization its re-solution. If a small quantity of an
electrolyte be added to a pure ruby-red colloidal gold solution, it changes to a
blue or violet color, and deposits its gold as a fine blackish coagulum or precipi¬
tate.1 By watching the coagulation of very dilute milk by acid in the ultrami¬
croscope, the individual particles of the casein may be seen to gather gradually
together into groups, whose motion becomes progressively less as their size in¬
creases, until they are no longer able to stay afloat, and finally coagulate
in large grapelike clusters. Hardy believes that the particles of colloids adsorb
the oppositely charged ions of the electrolyte; at the isoelectric point (that is when
the charges on the particles become zero or are balanced) coagulation occurs.
1 The amount in milligrams of protective colloid just sufficient to prevent the change to violet of
10 c.c. of bright red colloidal gold solution, by the addition of 1 c.c. of a 10 per cent solution of NaCl,
is called the “gold figure” or ‘ gold number” of the protector.
If, however, an excess of electrolyte be added all at once, the isoelectric point may
be passed before coagulation occurs, and the particles acquire a charge opposite to
the one they had originally. Under such conditions, no coagulation may result.
Burton epitomizes the difference in action of various electrolytes as follows:
“Two remarkable results are evident on comparing the coagulative powers of various
electrolytes on colloids of different kinds; first, the coagulation depends entirely on
the ion bearing a charge of sign opposite to that of the colloidal particle; and, second
with solutions of salts, trivalent ions have, in general, immensely greater coagulative
power than divalent ions, and the later, in turn, much greater than univalent. Acids
and alkalis in particular cases act more strongly than the corresponding salts.
Very small quantities of electrolytes may actually stabilize colloidal solutions,
their adsorption by the particles resulting in an electric charge that causes mutual
repulsion. Frequently there are zones of stability representing concentration of
electrolytes which cause no precipitation; but larger or smaller quantities produce
coagulation. Thus a 1:32 mastic solution is flocculated by 0.10 and by 0.00025
molar aluminum sulphate, but not by a 0.0025 molar solution. Furthermore the
coagulation may be reversible or irreversible, depending on the amount of the pre¬
cipitating electrolyte. Thus albumen is irreversibly flocculated by 0.05 to 1.0
normal zinc sulphate; from IN to 2N there is no flocculation; beyond 2N reversible
precipitation begins, reaching a maximum at 4N.
Oppositely charged colloids mutually coagulate each other. If more of one colloid
is added than is necessary to satisfy the electric charge of the other, such excess acts
as a protector, inhibits coagulation, and gives the mixture its charge. For this
reason very minute quantities of protectors may act as sensitizers1 or coagulants.
Thus 0.0003 to 0.0001 pe rcent of gelatin flocculates mastic sols, and in the presence
of one part of gelatin per million colloidal gold is flocculated by quantities of hydro¬
chloric acid which otherwise would not affect it.
The coagulation of a colloid involves a liberation of energy, which may in some
cases be determined calorimetrically. Thus Prange found that 1 g. of colloidal
silver on precipitation by ammonium citrate developed 250.98 to 12G.73 cal., according
to the concentration of the silver solution. A still more profound coagulation of the
resulting hydrogel by sulphuric acid, liberated an additional GO cal. The liberated
energy may appear in other forms than heat, i.e., the so-called “vital forces,” nervous
impulses, and the mechanical energy of contracting muscle, are consequent upon
changes in the degree of dispersion of the body colloids. The electrical energy
released by the gymnotus and similar “electric” fish seems to be, and even self¬
luminosity may be, due to the same cause. In peizo-electrical effects and tribolumi-
nescence we have an analogous phenomenon, due to the aggregation of the constituent
particles of crystals.
Peptization is the reverse of flocculation, a dispersion of groups into separate
particles which once more acquire active colloidal motion which keeps them afloat
and in solution. In fact the term peptization was coined by Graham to express the
liquefaction of a gel, because of the strong analogy between digestion and colloidal
disintegration.2
Colloidal Protection.—A most interesting and important fact regarding rever¬

sible colloids is that they can communicate their reversible property to irreversi¬
ble colloids. The addition of gelatin (as little as 0.0001 per cent) to a solution of
colloidal gold protects the latter against coagulation upon the addition of elec¬
trolytes, and permits it to redissolve after desiccation. Ultramicroscopic exami¬
nation has shown that the gelatin does not affect the motility of the gold particles,
1 Anaphylaxis is apparently a sensitization of this character.
See page 772. Chemical changes may accompany peptization.
thus disposing of the suggestion of Lobry de Bruyn that it acts by decreasing

their motion. The idea advanced by Muller1 that gelatin acts by increasing the
viscosity and thus preventing the deposition of suspended particles, is disproved by
the fact that quince-kernel gum, notwithstanding its viscosity, exercises no pro¬
tective action,2 whereas the small quantities of gelatin necessary to produce this
effect cannot appreciably increase the viscosity, and actually permit gold particles
to settle if they are large enough to do so.
The action of reversible colloids in opposing group-formation, is of great technical

importance, for in many cases it permits them to hinder, modify and even prevent
coagulation, precipitation and crystallization. Thus, in the absence of the colloid
a dilute solution of nitrate of lead deposits beautiful glistening crystals of metallic
lead upon a strip of zinc immersed in it, whereas in the presence of some glue the
deposit appears dull and amorphous.
Dialysis.—Colloid solutions possess a small but definite diffusibility through

colloidal septa (parchment paper, bladder) as was recognized by Graham (see
p. 772). Graham’s original form of dialyzer may be made from a wide-mouthed
bottle whose bottom has been removed.3 The mouth is closed by a piece of
bladder or parchment paper tightly bound on, the solution to be dialyzed is
poured in, and the bottle immersed about half way in water contained in a larger
Fig. 2.—Dialyzer. Fig. 3.
vessel (see Fig. 2). Most of the crystalloids diffuse through the membrane
into the outer water, which should be frequently renewed; while most of the
colloid remains in the interior bottle, and may be thus obtained in a purified
condition. Improved modern dialyzers consist of parchment or collodion sacs,
discs or thimbles, or even of whole bladders, which have the advantage of a
larger dialyzing surface.
SOME PRACTICAL APPLICATIONS
Shower-proofing of Fabrics.—Ordinary textile fabrics are readily wet by

rain; but if coated with a surface film of wax or the like (aluminum stearate),
the falling droplets roll off—“like water off a duck’s back.” This waterproofing
effect is readily understood if we consider the interfacial surface tensions involved
(see Fig. 3). These are:
1. Surface tension water/air, denoted by crWA

2. Surface tension water/fabric, denoted by <rWF.
3. Surface tension air/fabric, denoted by crAF.
Practically speaking, the surface attraction of the untreated fabric for water
1 Ber.. 1904. Vol. 37, p. 11.
* according to Zsigmondy, quince-kernel gum acts as a protector with mose substances.
is so great (i.e., crWF is so small1), that both the surface film or skin of the rain¬
drop and the air film on the fabric, are burst; the rain-drop spreads itself out on
the fabric and is absorbed into its pores. That is wetting occurs if
erWF<<rWA + crAF
or aWA><rWF — crAF
In the case of the treated fabric, crWF is very much increased; becoming in
fact the surface tension water/wax, for surface tension is only skin deep.
Therefore,
crWF>crWA + o-AF
or crWA<o-WF — crAF
and no. wetting takes place.
Flotation.—The understanding of what happens in flotation will be much
simplified, if we bear in mind that, notwithstanding the presence of many dis¬
turbing and variable factors, there are two main triphasic systems involved.
These are:
(V
AIR )
JWATER
—
Fig. 4. Fig. 5.
1. Ore/water/oil from which the true ore (usually a sulphide) must emerge
with a film of oil which enables it to be taken up by the air bubbles and thus
floated with the froth, while the gangue is wetted and flowed off at a lower level.
2. Air/water/oil which yields the bubbles to float the oiled sulphide. Refer¬
ring to the diagram (Fig. 4), and denoting
1. Surface tension water/oil by crWO
2. Surface tension oil/sulphide by <rOS
3. Surface tension water sulphide by <rWS
then the oil will distribute itself over the interface water/sulphide (i.e., the sul¬
phide will be oiled)
if crOSCcrWO + <rWS
or o-WO>crOS-<rWS
The second system is illustrated by diagram (Fig. 5). Denote

1. Surface tension water/air by crWA
2. Surface tension water/oil by <rWO
3. Surface tension air/oil by oAO
The oil or some of its constituents distribute themselves at the interface water/air
because
crWA>crWO + crAO
75 23 33 (approximate values)
1 It should be borne in mind that the surface tension decreases, as the attraction between the phases
increases. In fact when the positive surface tension in a two phase system becomes zero, solution may
take place.
This yields air bubbles with a more or less stable surface film, i.en a foam or froth.
. This film is miscible with the oil film on the sulphide, although not necessarily identical
with it, because there may be a separation of the constituents of the oil by differential
absorption. The oiled sulphide particles therefore act like oil, and, distributing
themselves at the interface water/air, are attached to the air-bubbles and lifted or
floated to the upper froth layer, if they are not too heavy or are not knocked off.
The gangue is not oiled, and flows off at a lower level.
In the foregoing no account has been taken of. complicating factors which are
always present; and while space will not permit an exhaustive discussion of them and
their results, the most important will be considered.1
1. Water.—Here are found dissolved air, crystalloids and colloids. The dissolved
air aids in bubble-formation by reducing the internal pressure of the water. For the
most part the crystalloids, especially acid electrolytes, act beneficially, being adsorbed
by coagulating the fine slimes, clay, etc., which then tend to repel the flotation oil,
recalling the old-fashioned household expedient of moistening the mouth and throat
with orange or lemon juice before taking castor oil, and thus avoiding the taste.
Some electrolytes, especially alkalis (i.e., Na2C03) in small percentages are powerful
deflocculators, and by reducing the surface tension water/oil and water /sulphide tend
to emulsify the oil, wet the sulphide, and deflocculate the slimes. Humic substances
especially when converted into alkali salts, and protective colloids generally, act in
like manner—the injurious effects of glue are well known. Colloidal clay or very
fine gangue act similarly—“fat” clays will emulsify oils, coal tar and asphalt.
2. Ore.—Outside of soluble impurities which dissolve in the water, the fineness of
grinding is most important. The gangue should be fine enough to flow off, but not
so fine as to cause the undesirable effects
just referred to above. If the sulphide
particles are too large, the air bubbles can¬
not lift them, or they are easily knocked off;
if they are too small, there is a great loss in
efficiency, for the surface of the bubbles may
be satisfied by a layer of small particles which
Good Result Poor Result
represents a small weight of sulphide. The
Fig. G.
subjoined diagram (Fig. 6), makes this
obvious.2 The results may also be adversely affected by the surface oxidation of
the sulphide.
3. Oil.—This is deserving of the most intensive study, since it is the factor most
readily controlled; and besides the great and sudden development of the art of flotation
has turned the various oil markets topsy-turvy, necessitating a constant search for
new oils or combinations not only from financial considerations, but also because of
actual shortage in the supply. Since conditions are different at each plant, and since
different lots of ore even from the same mine may vary, a careful study is necessary
to determine what oil or mixture is best, and also to know how to meet changing
conditions. Particularly useful will be a determination of the foam-producing factor
as opposed to the sulphide-oiling factor. Some so-called soluble oils give even better
results than coal and wood tar distillates. According to recent patents, a wide
variety of substances may replace oils.
Some idea of the importance of flotation in the mining industry may be formed
from the subjoined table published by the U. S. Bureau of Mines.
1 Mechanical agitation and temperature exercise considerable influence. Callow estimates that with
four different oils, three oil percentages, two pulp densities, and two temperature changes, there are
about 60,000 possible different combinations of conditions.
2 The layer of tiny adhering ore particles form an armadillo-like armor about the air bubbles, which
to a large extent protects them from coalescence and destruction,
Monthly Consumption of Flotation Oils in United States
Monthly tonange of Monthly consumption of flotation oils, be¬

ore ginning of 1916, pounds
Type of ore Wood products

End of
Beginning 1916, esti¬
1916 (tons) mated Pine Pine Euca¬ Creo¬ Turpen¬
(tons) oil tar lyptus sote tine
i oil i
Copper. 1,248,000 1,942,000 59,300 750 417,000 1,500

1
Zinc and com¬ I
plex. 248,000 350,000 60,750 667 262,500 | 3,330
Lead. 115,000 136,000 3,900 216 121,000
Gold and silver 45,700 123,000 9,830 750 40,250
1,656,700 2,551,000 133,780 2,167 | 216 840,750 4,830
Monthly Consumption of Flotation Oils, Beginning of 1916, Pounds
Coal products Petroleum

Oleic
Type of ore
acid
Tar Creosote Cresol Crude Fractions
i
Copper. 677,000 403,000 8,340 79,000 1,702,000
Zinc and complex.. 5,830 10,670 46,500 157,000 41,000
Lead. 9,250 660
Gold and silver.... 27,450 4,920 7,090 6,250
5,830 715,120 463,670 8,340 243,090 1,749,910
Lubrication.—The consumption of petroleum is increasing so rapidly that a

material decrease in the supply, if not actual exhaustion, lurks in the not far
distant future. W. B. Hardy in a trenchant paper (Journ. Soc. Chem. Ind.,
,
Vol. 38 No. 2, p. 7t, 1919), has shown that we may with confidence look to colloid
chemistry to aid us in finding the lubricants of the future.
Glass surfaces selectively adsorb from the air rather more of the atmospheric
impurities than of the elementary gases and water vapor, yielding a film about W
(1 X 10~7cm.) thick, which Lord Rayleigh termed “grease” because it has the general
properties of an oil.1 For this reason a new or raw glass surface has different mechan¬
ical properties than a satisfied or neutral surface.
“It is not possible to get a raw surface of glass.2 One cannot cleave glass; but
the 'grease' film may be removed by rubbing the surface under water. Soaking
is insufficient; actual vigorous rubbing is necessary, preferably with the finger tips
1 This adsorbed layer of ‘'grease" tends to make pipettes and burettes deliver inaccurately, so
analysts remove it by oxidation with bichromate.
2 Raw surfaces are readily made by cleaving mica, and if immediately put together again, they
adhere so strongly that they withstand a powerful lateral pull, and usually tear the surface when finally
separated.
until a peculiar clinging feeling is produced; the same which is felt when the edge
of a finger bowl is rubbed so as to give out a musical note. A surface cleaned in this
way has some remarkable mechanical properties, the most striking being that two
cleaned surfaces seize when pressed together with a relatively small force. This may
be demonstrated very simply by measuring the tangential force needed to produce
slipping. That true seizing occurs is proved by the tearing of the surfaces, which
takes place in the act of slipping.
“The function of lubricants is to keep the applied surfaces in the neutral con¬
dition by maintaining a ‘grease’ film on each. Not all fluids, however, can act in
this sense as lubricants for any particular surface such as that of glass. Water,
ether, alcohol, benzene, and strong ammonia are apparently entirely incapable of
maintaining a lubricating film on glass. Seizing occurs just as readily when they are
present as it does with cleaned surfaces. Glycerin differs from the fluids mentioned
above in the fact that though it will not maintain a lubricating film it does prevent
seizing when present in excess. For instance, the maximal tangential force which a
pair of cleaned surfaces would support without slipping was measured in grams, 55.
Flooding the surfaces with water, benzene, alcohol, etc., left this value unchanged.
When a film of glycerin was deposited on the surfaces the force still stood at 55, but it
fell to nine when the surfaces were fully flooded with glycerin.
“The expression ‘film’ used above denotes a layer of fluid on the solid surface of the
order 10-7 cm. in thickness. With a true lubricant the facility for slipping is maximal
when a layer of this excesive tenuity separates the solid faces and nothing is gained
by increasing the thickness of the layer.
“Thus with castor oil the weight required just to start one face of glass slipping
over .another was 10 g. when only the invisible film of fluid mentioned above was
present, and it was still 10 g. when the surfaces were flooded with oil.
“Some fluids indeed seem to lubricate better in thin than in thick layers, to act
that is to say in the contrary way to glycerin. Acids as a class behave in this way,
the solid faces again being of glass.
Pull in Geams
Film . Flooded
Acetic acid. 40 47
Sulphuric acid. 37 47
Oleic acid. 10 13
“If these results can be fully substantiated it will be an important and striking
physical fact likely to throw much light upon the process of lubrication. One broad
conclusion emerges from these facts, namely, that lubrication depends wholly upon the
chemical constitution of a fluid, and the fact that the true lubricant is able to render
slipping easy when a film of only about one molecule deep is present on the solid
faces, suggests that the true lubricant is always a fluid which is adsorbed by the solid
face. If this be so then the problem of lubrication is merely a special problem of
colloid physics.”
Some solids, notably graphite, are lubricants themselves. But the chief function
of graphite seems to be as an aid to lubrication by oil. The graphite is disintegrated
in the bearing and besides filling up relatively large imperfections, its colloidal particles
are adsorbed by the iron, forming in reality a graphite surface which has a lower sur¬
face tension against oil than iron has. Practically this means that the oil film adheres
more strongly to the graphitized surface and therefore needs greater force or pressure
to break it down; i.e., the bearing will stand greater speed and pressure. Acheson’s
“oildag” consists of a paste of colloidal graphite dispersed in oil; a small percentage
is mixed with the lubricating oil and feeds with it through the finest orifices. “Aqua-
dag” is the corresponding water-soluble product, stabilized with tannin.
Colloidal Fuel.—What promises to be one of the most far-reaching advances
made under the stress of the recent war, when necessity literally was the mother of
invention, is the discovery that by means of a suitable fixateur or peptizing agent
and suitable treatment, very large percentages of cheap tars and finely powdered
coal waste may be held dispersed in fuel oil with a sufficient degree of permanence
to enable the mixture to be stored, piped, atomized and burned practically like
fuel oil itself. Since it will at one stroke relieve the drain on the earth’s rapidly
diminishing stores of petroleum, as well as lead to the efficient utilization of all
kinds of coal waste (culm, screenings, dust), inferior fuels (peat, lignite), and even
cellulose waste (slabs, saw-dust), this new composite fuel may be hailed as a
powerful factor in the conservation of our natural resources, and a lasting benefit
to mankind.
Realizing the vital importance of the oil supply to the Allies in the conduct of
naval, military and manufacturing operations, the German submarines bent every
effort to destroy tankers, and Marshal Foch is said to have cabled America: “If
you don’t keep up your petroleum service, we shall lose the war.” While the Allied
navies were dealing with this peril in a most decisive fashion, Lindon W. Bates, head
of the Engineering Committee of the Submarine Defence Association, with the assist¬
ance in laboratory matters of Dr. S. E. Shepperd and other chemists of the Eastman
Kodak Laboratory, courteously opened to him, developed “colloidal” fuel, which by
practically doubling the usefulness of every oil cargo, would of itself have materially
assisted the defeat of the enemy efforts.
Coal or other combustible solid is prepared for dispersion by being pulverized so
that about 95 per cent passes through a 100 mesh, and 85 per cent through a 200-mesh
screen. This, of course, means that by far the greatest weight is in particles hundreds
and thousands of times larger than colloidal dimensions. But we should remember
that violent motion of the colloidal particles causes the Brownian movement of much
larger particles, as is the case in milk (see p. 765) where the fat globules can be seen
oscillating about under the ceaseless bombardment of the casein ultramicrons. It
has been found that a fluid fuel may be made containing as much as 40 per cent by
weight of powdered coal, and mobile pastes with up to about 75 per cent; and mobile
gels may be produced both from the. liquid and the pastes.
The exact nature of the fixateur, peptizing agent, and treatment is withheld
pending the issuance of patents here and abroad, but it is stated that ordinarily
between 0.5 per cent and 1.5 per cent of the essential component of the fixateur is
used, whereas about 0J per cent exercises a noticeable influence. The amount of
protecting agent is determined by the nature of the mixture desired, the components,
and the degree of permanence desired. The bulk of the particles does not begin to
settle until the period of “life” has passed. This period is subject to control within
reasonable limits and may be days or months in duration and may be relative or nearly
absolute in degree. Settling thereafter begins gradually but proceeds progressively
unless a gel has formed. Heat and agitation revivify the fluid fuel, and the paste
may be kept for years. The period of adequate stability may be regulated to meet
requirements, days for power plants, weeks and months for ocean going vessels and for
central storage stations.
Being heavier than oil or even water, colloidal fuel compresses in a unit volume
the maximum thermal value, thus economizing storage space; and it may be stored
under and extinguished by water, thus avoiding fire risk. Its operative efficiency
is high, for when sprayed into the hot firebox, the oil-permeated particles of carbonace¬
ous matter are still further atomized by the sudden gasification of their imbibed oil.
It possesses the advantages of fuel oil over coal, absence of smoke, dust and ash,
practical elimination of labor in loading into storage space and in firing, with resultant
saving in time for “coaling” and for raising steam. As in the case of oil, a protective
smoke screen may readily be produced by over-firing, and the fire is subject to instant
control.
Smokes.—While it is an ancient principle of strategy to attack an enemy

under cover of natural fog or mist, the recent war led to an enormous develop¬
ment of artificial smokes as aids to naval and military offensive and defensive
operations. Like the cuttlefish, steamers strove to escape pursuing submarines
in colloidal clouds of their own making, and our soldiers frequently crept to the
attack under a protective smoke barrage. Ammonium chloride, stannic acid,
sulphur trixoide, phosphorus, coal, and oil smokes were among the cloud-forming
substances used, and some of the interesting work of the Dispersoid Section,
Research Division, Chemical Warfare Service, U. S. A., has been published.
Tolman and Vliet1 describe an easily constructed Tyndallometer for determining
the strength of the Tyndall beam in suspensions, colloidal solutions, smokes, and
mists, and find it proportional to the concentration, unless too high a concentration
allows opacity to affect the results. In general for smokes, increased subdivision
results in increased Tyndallometer readings. Gaseous dispersoids have a relatively
short life, for the coagulation and precipitation of the smoke particles begins immedi¬
ately, and is favored by concentration (meaning closer proximity of the particles)
and by fineness of subdivision (meaning greater speed of and more collisions between
the particles). It is therefore impossible to raise the optical density of a smoke be¬
yond a certain point, by the introduction of further smoke material.
Wells and Gerke2 describe an oscillation method for determining the size of ultra-
microscopic particles, by measuring photographically their oscillation amplitudes in
an electric field which is rapidly reversed by a rotating commutator. The formula
used, based on Stokes7 Law, is
Xe = SirTjdv
where X = electric field in volts per centimeter
e = electric charge (assumed to be 1 electron) = 1.59 X 10-20 e.m.f.
7] = viscosity of the dispersion medium (air).
d — diameter of the particle in centimeters
v = velocity of the particle (== oscillation amplitude X frequency of field
reversal)
which gives d = 9.3 X 10-10X/v
Particles in one specimen of tobacco smoke were estimated according to this

method to be about 2.75m; in another specimen, from 0.8 to 1.0m.3
As an instance of the commercial utilization of smoke, may be mentioned the prac¬
tise of burning smudge pots in orchards especially in spring or early summer, when
weather condition indicate that the tender blossoms or fruit are apt to be damaged
by frost. The heat emitted by the pots is a factor, but the efficacy is partly due
to the dense clouds of smoke given off by the burning smudge oil. This remedy is
1Jour. Am. Chem. Soc., Vol. 41, p. 297 et seq.

2 Jour. Am. Chem. Soc., Vol. 41, p. 312 et seq.
3 See also Zsigmondy’s observations on tobacco smoke. “Colloids and the Ultramicroscope,” p. 196.
of course useless in very cold weather, and serves to prevent only what agriculturists
call a light frost, which has for its precursors a still clear atmosphere and a tempera¬
ture approximating the freezing point. No such frost occurs on cloudy or windy
nights.
In the absence of smoke, moisture deposits upon leaves, fruit and other surfaces
furnishing nuclei, and congeals into ice crystals or “frost” when the slight heat thus
generated by condensation, is dissipated into the surrounding cold atmosphere, the
freezing being also facilitated by the evaporation of part of the deposited water.
The smoke, however, furnishes throughout the entire lower atmosphere, innumerable
nuclei for the deposition of moisture and the widely distributed and relatively larger
amount of heat thus set free is sufficient to help carry the temperature above the
freezing point. The smoke fog furthermore forms a kind of protective blanket, which
reflects the terrestrial heat and prevents its dissipation.
On the other hand fine sulphuric acid mist, cement and smelter fume and the like,
are types of injurious industrial smokes which are now electrically coagulated by the
well-known Cottrell process. Although this process was originally worked out to
abate the dust and fume nuisance, cement and blast-furnace dust are now yielding
considerable potash, while in smelter fumes are found large quantities of arsenic,
selenium, tellurium and other substances. The use of the high tension discharge to
dispel fog and thus prevent collision at sea, is a possibility previously suggested by
Sir Oliver Lodge.
Ultracentrifugation.—The non-homogeneity of ordinary crystalloid solutions

was demonstrated by van Calcar and Lobry de Bruyn, who caused about three-
eighths of the sodium sulphate in a saturated solution to crystallize at the per-
phery of a rapidly rotating centrifuge. To do this required several hours at
about 3,000 r.p.m. If such results can be accomplished with crystalloid molecules,
it is to be expected that larger colloidal groups can be separated from solution by
centrifugal force, the main factors being the viscosity of the dispersion medium,
diameter of the centrifuge, speed of rotation, and length of time the centrifugal
force acts.
Certain commercial sized centrifuges rotate between 15,000 and 20,000 r.p.m.,
and develop a separating force over 15,000 times as great as gravity. Such machines
naturally have a much wider range of usefulness than their prototype, the turbine-
driven cream separator, which years ago marked such a big improvement in both
speed and efficiency, over the old-fashioned “pan” or gravitation method of separat¬
ing emulsified butter fat from milk. Thus for example, with modem ultracentrifuges,
varnishes, nitrocellulose lacquers, vegetable oils, gelatin liquors may be clarified;
blood corpuscles may be separated from the serum in making blood albumen and sera;
and there may be recovered from waste wool scouring liquor, wool grease of light color,
suitable for refining into lanolin, instead of the dark degras yielded by the cumbersome
acid coagulation method. The use of the laboratory ultracentrifuge capable of mak¬
ing 40,000 r.p.m. opens a wide field for analytical and physiological research, for with
it can be avoided a factor that often causes trouble in filtration or ultrafiltration,
namely adsorption by the filter medium.
Miscellaneous Practical Applications.—The use of protective colloids in

electrodeposition is well known, use being made of them on a large scale in the Betts’
process for electrolytic lead. Colloids such as glue, dextrin, tannin, are used to
prevent the formation of hard scale in boilers. The successful treatment of
sewage largely depends on the successful precipitation of colloids, as does also
water filtration.
Many other practical applications are mentioned in the reference given below.
These include rubber, brewing, chemical analysis, colored glass, metallurgy, gems,
medicine, biology, paper, dyeing, nitrocellulose, petroleum, asphalt, mineralogy,
geology, tanning, etc. Most journal articles of value are abstracted in “Chemical
Abstracts” or “Kolloid Zeitschrift.”
BIBLIOGRAPHY
Alexander, Jerome: “Colloid Chemistry.”

Bechhold, H.: “ Colloids in Biology and Medicine.”
Burton, E. F.: “ Physical Properties of Colloidal Solutions.”
Cotton and Mouton: “Les Ultramicroscopes et les Ob jets Ultramicroscopiques.”
Gaston, Paul: “LrUltramicroscope dans le Diagnostique Clinique et les Recherches
de Laboratoire.”
Muller, Arthur: “Allgemeine Chemie der Kolloide.”
Ostwald, Wo.: “Grundriss der Kolloidchemie.”
Ostwald, Wo.: “Ail Introduction to Theoretical and Applied Chemistry.”
Ostwald, Wo.: “Theoretical and Applied Colloid Chemistry.”
Svedberg, The: “Herstellung Kolloider Losungen.”
Zsigmondy, R.: “Chemistry of Colloids.”
Zsigmondy, R.: “Colloids and the Ultramicroscope.”
SECTION XXIV
METALLURGY OF ZINC
By Walter Renton Ingalls1
The art of zinc smelting is simple in principle, but complicated in practice.

It depends upon the reducibility of zinc oxide by carbon, or carbon monoxide, or
both. Consequently zinc sulphide—blende—must first be roasted to obtain the
oxide. Zinc silicate (ZnO. Si02) is reducibledirectly by carbon.
Zinc oxide begins to be reduced by carbon at 1,125°C., but at that temperature

zinc is reoxidized by carbon dioxide if the latter be present in more than 0.2 per cent.
With increase of temperature the permissible percentage of carbon dioxide increases
also, until at 1,500°C. there may be 0.76 per cent 2 This fact is of supreme impor¬
tance in the metallurgy of zinc. It explains why metallic zinc may not be distilled in
a blast furnace and why the product is blue powder in one-stage smelting in the elec¬
tric furnace. Zinc oxide can be reduced in the blast furnace just as it is in the
Wetherill furnace, which is a species of blast furnace, but in the Wetherill the reduced
zinc is then intentionally burned to oxide. By limiting the presence of carbon dioxide,
combustion of the zinc may be avoided, but the oxidation that results in blue powder
cannot be avoided. This is why the reduction-distillation of zinc is limited to rela¬
tively small retorts, in which the necessary and peculiar conditions can be maintained,
and why the process must be so arranged as to be divisible into stages. Even so,
although it is possible to establish the condition of very low carbon dioxide in the
gas issuing from the retort during the period of active distillation, it is seldom possible
to attain the minimum theoretically required and consequently there is always some
production of blue powder.
As practically conducted, zinc distillation is an intermittent process, divided

into four stages, viz: (1) The maneuver, during which the spent residues from
the previous charge are withdrawn, broken retorts are replaced, all retorts are
recharged, etc. This requires 3 to 5 hr. (2) Preliminary reduction, during which
water is expelled, hydrocarbons are driven off, and the oxides of metals other than
zinc, such as iron and lead, are reduced. During this stage the temperature
of the retorts is raised gradually. The issuing gas is first pure coal gas and burns
with a luminous flame at the noses of the condensers. This flame becomes less
bright as the reduction of the metallic oxides begins and the percentage of carbon
monoxide increases, gradually becoming purple. When the flame becomes
tinged with bluish-green the beginning of zinc distillation is indicated. This
marks the end of the second stage of the process, although the dividing line is
not sharply drawn. When zinc begins to distil and condense the percentage of
carbon dioxide in the gas is still high and consequently there is much formation
of blue powder at this time. (3) Active distillation, during which the zinc is
reduced and distilled rapidly. (4) Waning distillation, during which the last
1 Consulting engineer, 115 Broadway, New York, N. Y.
2 Bodlander, Zeit. filr Elektrcchemie, VIII, xliv, 833-842.
793
part of the zinc is expelled, the average rate of hourly condensation during this
period being only about one-half that of the previous period. A high tempera¬
ture is maintained and the resistance of the retorts, relieved of the endothermic
reaction within them, is tested most severely. Finally, all of the zinc is expelled
that can be at the temperature used, the condenser flames become purple again,
the cycle has been completed and preparations are made for the maneuver of the
next day.
From the foregoing will be seen one condition that stands in the way of making
zinc reduction-distillation a continuous process, viz., the necessity for performing the
second stage separate from the third. In practice this is done as one of the steps in a
24-hr. cycle in the distilling furnace, but it may be done in a separate furnace as a
separate act and then zinc distillation may be conducted continuously. This has
been proved by electric furnaces operated on preheated, or prereduced, charges.
The same idea of prereduced charges is actually, though not generally, used in prac¬
tical pyrometallurgy, not to the extent of enabling continuous distillation, but
enough to relieve the retorts of some of their work and extend the time available for
active zinc distillation.
However, there is another important condition that stands in the way of continuous
distillation, though it is nob insuperable, viz: Intimate contact of zinc ore and carbon
is necessary, which implies fineness of grains. Indeed, zinc ore being so largely the
product of concentrating mills, including the excessively fine flotation slimes, it comes
to the distiller fine anyhow.. Now, when a retort is charged with fine ore a rod must
be run through the top (this is called “spiessing”) in order to afford a channel for
egress of the gas and vapor during the early stage of distillation. However, the gas
and vapor must seep up through the charge itself in order to get into the channel.
Under certain conditions, e.g., too great depth, excessive fineness, etc., it may be
unable to do that easily. The result then may be the accumulation of such internal
pressure that the entire charge will be blown out of the retort. This is why the dis¬
tillation of ore ordinarily fine is practically limited to the depth that it has in the
Rhenish retort, i.e., about 12 in. For the distillation of coarser charges in Upper
Silesia deeper muffles were formerly used, but when they began to treat richer and
finer ores the Silesian distillers had to adopt the Rhenish retort.
The same condition applies to vertical retorts designed for continuous operation,
although the earlier designers did not think of it. The gas and vapor could not get
out. One or more inventors thought subsequently to meet this condition by arranging
a row of outlets down the shaft. It is possible, also, that the condition may be met in
other ways. However, with the exception of the Roitzheim-Remy furnace that was
introduced in Germany just before the war, about which no thoroughly satisfactory
information was communicated, and since then none at all, continuous distillation has
not become established in zinc pyrometallurgy.
It will help in the consideration of the whole subject of zinc pyrometallurgy to
fix these fundamental principles of practice well in the mind. There is another one,
namely, the matter of temperature of distillation, on which depends largely the
extraction of the zinc, the quality of the primary product and the cost of the process.
Although zinc oxide begins to reduce, under certain conditions, at 1,125°C.

the speed of the reaction is greatly increased by increase of temperature, and since
it is for technical considerations important in practice to complete the working
of a charge in 24 hr., it is essential in practical smelting to conduct the operation
at a considerably higher temperature. Given a certain time for the distillation,
which is commonly about 18 hr., the remaining 6 hrs. being occupied in
METALLURGY OF ZINC 795
discharging and recharging the retorts, the higher the temperature, the more
complete will be the extraction of the zinc. But here enter other complications.
The singulosilicates of iron, lime, etc., begin to form at about 1,200°C., and as the
temperature rises further more infusible slags can form, and the gangue of the ore
containing these elements may give rise to slags which quickly corrode and
destroy the retorts in which the distillation of the zinc is being conducted, to such
an extent indeed as to make the process impracticable. Moreover, with increase
of temperature other metals, especially lead, distil over with the zinc, necessitate
‘ng a further refining of the latter in order to produce a marketable quality of
metal. An ordinary temperature of distillation is 1,250 to 1,350°C. The per¬
formance of the process at 1,200°C. is known as “slow driving/’ while 1,400 to
1,450°C. is “hard driving.”
American smelters lean to “slow driving” while Europeans adhere to hard driving.
In either way there is some variation in the management of the furnace. Having
arrived at the third stage of the process one smelter may raise the temperature
quickly and maintain a high temperature right through the third and fourth stages
while another may raise it more gradually and attain a maximum during the fourth
stage when he is “pinching” the charge. As already pointed out, with increase of
temperature other metals, especially lead, distil over with the zinc. Given the lead
contents of charge and spelter, an experienced metallurgist can tell pretty nearly
hat was the temperature of distillation. American smelters make a practice of
keeping separate their first draw of metal, which will be higher in cadmium and lower
in lead than the second draw. The third draw will be highest in lead. European
3melters, distilling uniformly at high temperature, are apt to make a Ieady spelter,
which has to be refined by melting in a reverberatory furnace and allowing the excess
of lead to separate by gravity. This has the incidental advantage of equalizing the
spelter and producing slabs of uniform composition. Spelter molded directly from
the drawing kettles is apt to be irregular and there is more or less complaint against
American brands on this account.
The matter of temperature introduces consideration of the behavior of the

gangue of the ore, for the higher the temperature the more dangerous become the
lags. The historic method of taking care of the slag is to have present a great
xcess of reduction material to absorb and hold it like a sponge, but in modern
ractice care is taken not to permit the formation of anything like an easily
fusible singulo-silicate, but so to compound the charge that the slag will be a
bi$ilicate or trisilicate. The retorts must be as strong, dense, and non-corrodible
as it is possible to make them, and the lining of the furnace, in fact the general
structure, and especially the middle wall, or middle ledges, must be of the maxi-
part of the zinc is expelled, the average rate of hourly condensation during this
period being only about one-half that of the previous period. A high tempera¬
ture is maintained and the resistance of the retorts, relieved of the endothermic
reaction within them, is tested most severely. Finally, all of the zinc is expelled
that can be at the temperature used, the condenser flames become purple again,
the cycle has been completed and preparations are made for the maneuver of the
next day.
From the foregoing will be seen one condition that stands in the way of making
zinc reduction-distillation a continuous process, viz., the necessity for performing the
second stage separate from the third. In practice this is done as one of the steps in a
24-hr. cycle in the distilling furnace, but it may be done in a separate furnace as a
separate act and then zinc distillation may be conducted continuously. This has
been proved by electric furnaces operated on preheated, or prereduced, charges.
The same idea of prereduced charges is actually, though not generally, used in prac¬
tical pyrometallurgy, not to the extent of enabling continuous distillation, but
enough to relieve the retorts of some of their work and extend the time available for
active zinc distillation.
However, there is another important condition that stands in the way of continuous
distillation, though it is not insuperable, viz: Intimate contact of zinc ore and carbon
is necessary, which implies fineness of grains. Indeed, zinc ore being so largely the
product of concentrating mills, including the excessively fine flotation slimes, it comes
to the distiller fine anyhow.. Now, when a retort is charged with fine ore a rod must
be run through the top (this is called “spiessing”) in order to afford a channel for
egress of the gas and vapor during the early stage of distillation. However, the gas
and vapor must seep up through the charge itself in order to get into the channel.
Under certain conditions, e.g., too great depth, excessive fineness, etc., it may be
unable to do that easily. The result then may be the accumulation of such internal
pressure that the entire charge will be blown out of the retort. This is why the dis¬
tillation of ore ordinarily fine is practically limited to the depth that it has in the
Rhenish retort, i.e., about 12 in. For the distillation of coarser charges in Upper
Silesia deeper muffles were formerly used, but when they began to treat richer and
finer ores the Silesian distillers had to adopt the Rhenish retort.
The same condition applies to vertical retorts designed for continuous operation,
although the earlier designers did not think of it. The gas and vapor could not get
out. One or more inventors thought subsequently to meet this condition by arranging
a row of outlets down the shaft. It is possible, also, that the condition may be met in
other ways. However, with the exception of the Roitzheim-Remy furnace that was
introduced in Germany just before the war, about which no thoroughly satisfactory
information was communicated, and since then none at all, continuous distillation has
not become established in zinc pyrometallurgy.
It will help in the consideration of the whole subject of zinc pyrometallurgy to
fix these fundamental principles of practice well in the mind. There is another one,
namely, the matter of temperature of distillation, on which depends largely the
extraction of the zinc, the quality of the primary product and the cost of the process.
Although zinc oxide begins to reduce, under certain conditions, at 1,125°C.

the speed of the reaction is greatly increased by increase of temperature, and since
it is for technical considerations important in practice to complete the working
of a charge in 24 hr., it is essential in practical smelting to conduct the operation
at a considerably higher temperature. Given a certain time for the distillation,
which is commonly about 18 hr., the remaining 6 hrs. being occupied in
discharging and recharging the retorts, the higher the temperature, the more
complete will be the extraction of the zinc. But here enter other complications.
The singulosilicates of iron, lime, etc., begin to form at about 1,200°C., and as the
temperature rises further more infusible slags can form, and the gangue of the ore
containing these elements may give rise to slags which quickly corrode and
destroy the retorts in which the distillation of the zinc is being conducted, to such
an extent indeed as to make the process impracticable. Moreover, with increase
of temperature other metals, especially lead, distil over with the zinc, necessitat¬
ing a further refining of the latter in order to produce a marketable quality of
metal. An ordinary temperature of distillation is 1,250 to 1,350°C. The per¬
formance of the process at 1,200°C. is known as “slow driving,” while 1,400 to
1,450°C. is “hard driving.”
American smelters lean to “slow driving” while Europeans adhere to hard driving.
In either way there is some variation in the management of the furnace. Having
arrived at the third stage of the process one smelter may raise the temperature
quickly and maintain a high temperature right through the third and fourth stages
while another may raise it more gradually and attain a maximum during the fourth
stage when he is “pinching” the charge. As already pointed out, with increase of
temperature other metals, especially lead, distil over with the zinc. Given the lead
contents of charge and spelter, an experienced metallurgist can tell pretty nearly
what was the temperature of distillation. American smelters make a practice of
keeping separate their first draw of metal, which will be higher in cadmium and lower
in lead than the second draw. The third draw will be highest in lead. European
smelters, distilling uniformly at high temperature, are apt to make a leady spelter,
which has to be refined by melting in a reverberatory furnace and allowing the excess
of lead to separate by gravity. This has the incidental advantage of equalizing the
spelter and producing slabs of uniform composition. Spelter molded directly from
the drawing kettles is apt to be irregular and there is more or less complaint against
American brands on this account.
The matter of temperature introduces consideration of the behavior of the

gangue of the ore, for the higher the temperature the more dangerous become the
slags. The historic method of taking care of the slag is to have present a great
excess of reduction material to absorb and hold it like a sponge, but in modern
practice care is taken not to permit the formation of anything like an easily
fusible singulo-silicate, but so to compound the charge that the slag will be a
bi$ilicate or trisilicate. The retorts must be as strong, dense, and non-corrodible
as it is possible to make them, and the lining of the furnace, in fact the general
structure, and especially the middle wall, or middle ledges, must be of the maxi-
mum durability. The successful withstanding of a temperature of 1,550°C. for a

campaign of upward of 3 years requires refractory material, mason’s work, and
general furnace design of a high order of excellence.
Complying with these conditions, good practice tends strongly to hard driving,
for the extraction of zinc is greatly increased without corresponding increase in the
consumption of coal as compared with slow driving. The European smelters have
been hard drivers because their coal has been dear, even in pre-war times. Note
must here be made that a furnace designed for the distillation of a certain ore at a
relatively low temperature cannot necessarily be speeded up by simply raising the
temperature. In so doing the capacity of the condensers might be overtaxed and
zinc distilled too rapidly might be forced to burn uncondensed at their noses. Refer¬
ence may here be made to the subject of prolongs, sheet iron extensions of the con¬
densers. European distillers use them generally; American but rarely. They save
considerable zinc escaping from the condensers, but in American practice this is
deemed insufficient to compensate for the extra labor of taking them down and
putting them up at each draw. With the Rhenish and Silesian furnaces drawing is
less frequent, nor do the prolongs have necessarily to be taken down during drawing.
This statement of general principles outlines the complexity of the art of zinc
distilling in arriving at the most economical balance among many conflicting con¬
ditions. Within the present limits of space it is impossible even to hint at many
details of the art. The manufacture of retorts and condensers, the design of distilling
furnaces, the organization of labor for their operation, the matter of density of charge,
these and many other things are subjects for voluminous analysis and discussion.
Bulky treatises are devoted to them without exhausting the subjects. The reader is
referred to Ingalls on “Metallurgy of Zinc and Cadmium,” Lodin on “Metallurgie
du Zinc,” Liebig on “Zink und Cadmium,” Hofman on “Metallurgy of Zinc and
Cadmium,” and Prost on “Cours de Metallurgie des Metaux autres que le Fer.”
Coming to practical application, the first step is to roast the ore, blende now
being the common ore. Out of many furnaces that have been designed and used,
but few need here be mentioned. Where labor conditions permit hand-roasting,
as in Europe they have done heretofore, the Delplace kiln is the favorite. This is
a superior form of shelf-burner, so designed as to be very economical of fuel.
American smelters use commonly the Hegeler furnace, which is clumsy and costly.
European smelters and one American have been successful with the Spirlet,
which is distinctly a mechanical furnace. Other mechanical furnaces, such as the
Merton and the Ridge, have been tried here and there but have not established
themselves generally. The metallurgy of zinc still lacks a thoroughly good
mechanical muffle-furnace.
Furnace Design.—Respecting distilling furnaces the designs that are in
successful use are of great variety. It is even difficult to classify them. My
present effort to do so will be different from anything attempted heretofore.
In the first place I shall distinguish between the method of heating and the
metallurgical type, which is roughly equivalent to distinguishing between the inferior
structure and the superior. Confining our attention for the moment to regenerative
furnaces, and the modern furnace is almost necessarily confined to this class, we may
have either the counter-current regenerative or the reversing regenerative. The
latter may have regeneration for both gas and air or for air alone. The checker-
works may be arranged longitudinally, or transversely. In the laboratory, or com¬
bustion chamber, of the furnace the flames may rise vertically, leap over the middle
wall, and descend on the other side, or they may reverse from end to end of the
surface traveling longitudinally and horizontally. They may issue from ports in the
middle wall. This is mainly a problem of efficient combustion and equable heating.
The desiderata are a temperature around the retorts as nearly even as possible and a
combustion product escaping to the chimney at only 200 to 300°G. as it leaves the
furnace. There are designs, such as the Welzer (Overpelt) andNeureuther-Siemens,
that permit a ton of ore to be distilled with as little as 0.9 ton of coal. The best of the
counter-current furnaces and the Hegeler furnace, so extensively used in the United
States, in which the heat recuperation is effected through the medium of a waste-heat
boiler cannot equal that figure.
Metallurgically, distilling furnaces may be classified as Belgian, Silesian and
Rhenish. The distinguishing features are comprised within the superior structure,
which may be combined with different forms of the inferior. Thus, a Rhenish top
was developed a compromise, known first as the Belgo-Silesian and later as the
Rhenish. Being a compromise, in which features were adopted from both the proto¬
types, the characteristic features of each were not everywhere borrowed to the same
degree, and consequently there has always been some uncertainty as to just what
constitutes the pure Rhenish type. I will offer the following generalization.
The Rhenish distilling furnace is one possessing a combustion chamber in which are
suspended three double rows of retorts so that the flames can circulate around ail of
them. The furnace has two facades, or fronts, and is roofed by a single arch springing
from the top of one front to the other. The support of the inner ends of the retorts is
characteristically a vertical lattice work of refractory blocks, but may be a middle
wall, with ledges as in the Belgian furnace. The retorts have the cross-section like
a chain link, are about 12 in. high inside, and hold about 34 kg. of charge. The
condensers are larger and longer than in the Belgian furnace. The front of the
furnace is divided into closets completely enclosing the condensers, one closet com¬
prising six condensers. The closets are subdivided by three horizontal shelves, which
have holes through which the residues raked out of the retorts are dropped into pockets
below the working floor.
As thus defined, the Rhenish furnace has the main design of the Silesian, and ob¬
tains much of the economy of the latter in labor and fuel, but it has the suspended
retorts of the Belgian, giving the advantage of better heating. The dimensions of
the retorts are a compromise between those of the Belgian and the Silesian and are as
large as is consistent with the successful treating of high-grade ore, for reasons pre¬
viously stated, and as long as will permit the load safely to be sustained. However,
the retorts themselves are not a distinguishing feature of the Rhenish furnace, for
the same retorts may be used in a Belgian furnace without destroying the distinct
characteristics of the latter.
The Belgian furnace itself has undergone great modifications which have resulted
in some confusion of description. The American zinc smelter in speaking of the
Belgian furnace means the small, direct-fired blocks installed at Pittsburg, Kan.,
and vicinity about 50 years ago and still used there within a year or two. The gas-
fired furnaces used in the more modern plants he calls “Hegelers” or “Siemens.”
A correct description would be “Old Belgian,” “Hegeler-Belgian” and “Siemens-
Belgian.” The Hegeler-Belgian is simply a very high, gas-fired Belgian, turned over
on its side. The Siemens-Belgian is an elongated Belgian, which was rendered
possible by its method of gas firing. Besides such modifications the Belgian furnace
may be equipped with either cylindrical or elliptical retorts, or even the large retorts
employed with the Rhenish furnaces. Thus, the last were installed in the original
Siemens-Belgian furnaces in this country, but their use did not endure.
The distinguishing features of the Belgian furnace are the use of rather small
retorts, arranged in four or five rows, each retort bridging from front wall to back wall
(or middle wall in the case of a double furnace); but especially the design of the
front walls. These are thin, being scarcely more than sufficient to support the front
ends of the retorts. Outside of the front walls the structure is supported by an iron
devanture of only small depth, the main function of which is to iron the furnace, while
the horizontal shelves merely support the condensers and aid in clearing out the
retort residues. Thus the front of the furnace is latticed off in a series of shallow
pigeon holes instead of the deep closets of the pure Rhenish furnace.
Differences in Practice.—The differences in fronts create a rather important

difference in metallurgical operation. "We used to deplore the necessity for
making the front wall of a distilling furnace so thin, which we thought resulted
in a terrific waste of heat by radiation. We learned, however, from thermal com¬
putations that this radiated heat was necessary in order to keep the condensers
at the proper temperature, which computations received ail illuminating con¬

firmation in some electric furnace smelting when it was found necessary -to arrange
an independent fire under the condenser. In the Belgian furnace the condensers
stick right out from the front wall of the furnace and it is more difficult to regulate
their temperature than it is in the Rhenish furnace wherein they are enclosed in
closets, which often have adjustable, sliding doors in their fronts.
However, as I have previously indicated there are all kinds of departures from
typical Rhenish design and practice. Thus, at Blende, Colo., the furnaces built
originally as Welzers (Overpelt-Rhenish) have been greatly modified. While these
furnaces have the deep closets they are now divided by iron plates instead of fire¬
brick tiling. The condensers fit fully into the mouths of the retorts and of course have
to be taken down during the maneuver. The discharging at Blende is now done by
means of a machine, which demands free access to the retort, anyhow.
European zinc-distilling practice has concentrated rather emphatically upon

the Rhenish type of furnace, to which the Silesian smelters have been going over
in recent years. This type of furnace gives as good zinc extraction as the Belgian,
is more durable both as to retorts and as to furnace itself, and lends itself to
infrastructures that aid in producing rather better economy of fuel than is realized
with the Belgian furnace of Europe. They do not, however, excel the best
regenerative Belgians of the United States, although they equal them. In re¬
placing broken retorts it is more trying on the men to handle and insert the
heavier Rhenish retorts than it is the lighter Belgians, but on the whole the labor
in operating the Rhenish furnace is rather less and is distinctly easier than the
Belgian. Nevertheless the Rhenish furnace, which has had only one prolonged
trial in the United States, has found no favor here, and American smelters in
adhering to the Belgian type stick to the old cylindrical retorts, although the Bel¬
gian smelters themselves use generally the elliptical. There are reasons for this
difference in ideas but present limitations do not permit a discussion of them.
Zinc-lead Smelting.—Until about 20 years ago zinc smelting was conducted
purely as a method of extracting zinc. The zinc smelter did not want to have
lead in his ore and penalized it. With the advent of the great supplies of Austral¬
ian ore, however, and other ores of similar character the zinc smelters of Europe,
especially of the Continent, overcame their aversion to lead and indeed began to
demand its presence to an extent of about 6 per cent as a minimum. They had
learned that lead did not prejudice the distillation of zinc except in so far as it
contaminated the spelter, which was no very serious matter. Zinc-smelting in
Europe was therefore put practically upon a zinc-lead basis. The same thing
was done by numerous smelters in the United States, but not to the same extent
as in Europe.
In distilling an argentiferous and plumbiferous zinc ore some of the silver and lead
go over into the spelter, but the major parts remain in the retort residue. The latter
may be passed to the lead-smelting blast furnace, or it may be subjected to mechanical
concentration and the concentrate alone be smelted for argentiferous lead. When
concentration is done some of the surplus reduction coal is recovered. The choice
of the method is determined by costs. Europeans generally concentrate; Americans
smelt directly. The ultimate extraction of lead and silver from a zinc ore is about 50
to 75 per cent of the original contents, 62.5 per cent being perhaps a fair average.
Another method of treating zinc-retort residues is to burn them in Wetherill
furnaces for the production of zinc oxide, either as a pigment or as an impure product
to be returned to the distilling furnaces. The surplus coal in the residues is usually
sufficient for the execution of this process, which serves to extract about five-eighths of
the zinc remaining in the residues, when the latter assay about 7 per cent Zn. This
process is used extensively by American smelters, who owing to their low-temperature
distillation have richer residues than occur in good European practice. If the residues
contain much silver this process would be inadvisable for the silver would largely be
lost.
The electrothermic smelting of zinc ore need be referred to but briefly, for
only in Scandinavia has it become a commerical art. The metallurgical principles
of zinc distillation obtain in electrothermic smelting just as they do in the ordinary
pyrometallurgy. There are two large plants in Scandinavia, viz. at Trollhattan
in Sweden and at Sarpsborg in Norway. Besides these, there are one or two
small plants. A new plant is being built at Glomfjord. This is to have capacity
for smelting 60,000 metric tons of ore per annum. At the end of 1921 it had not
been completed, but construction was well advanced. The electric furnaces used
for ore smelting in Scandinavia are resistance furnaces, and are designed and
run as combined zinc and lead smelting furnaces, both zinc and lead being
obtained in the same operation, the gangue of the ore being fluxed and drawn
off as slag.
In the smelting at Trollhattan until within a year or two the idea was to
condense spelter directly from the distillation furnace. In so doing there was
experienced the common difficulty in the matter of condensation, an unduly
large proportion of the zinc being condensed as blue powder. We know now
that that is inherent to any zinc ore reduction and distillation process that m
carried on continuously. The common process is conducted intermittently and
in stages. In the first stage there is an expulsion of oxidizing gases. After
they have been driven off, and when the effluent gas becomes essentially carbon
monoxide the temperature of the furnace is raised to the point of zinc reduction
and distillation, and in such an atmosphere condensation of the zinc mainly m
spelter can be effected. In a furnace operated continuously there can be no such
control of gaseous conditions, and as fresh ore is charged into the furnace there
is a constant development of oxidizing gases, and coincidentally there is more or
less dusting. The entrainment of mechanical dust with the stream of gas and
vapor constitutes an additional condition adverse to condensation as spelter.
The continuous reduction, distillation, and condensation of zinc (as spelter) t
requires therefore the conditions of a two-stage operation, just as there is in the
common pyrometallurgy. With a continuously operating furnace there must be
either a preliminary stage of pre-reduction in another furnace, or there must be
a subsequent stage of blue powder retreatment, also in another furnace. Which**
ever be the choice, there are two stages or two steps.
The Trollhattan metallurgists adopted the method of (1) reduction and
distillation and (2) blue powder retreatment. The latter was done by redistilla¬
tion in an electric furnace, but this was found to be unduly expensive. The
discovery was then made that the blue powder could be melted down in a com¬
paratively inexpensive way if subjected to heating to the melting point of zinc
and at the same time to a rubbing action.
In the new process recently developed at Trollhattan, which is also to be
applied at Glomfjord, the resistance furnaces arc run at high temperatures, the
lead being completely volatilized and no attempt being made to condense the
zinc as spelter. The furnace is operated on the principle of producing a highly
siliceous slag, which commonly contains about 50 per cent of silica, and creates
conditions favorable to a high degree of zinc expulsion, i.e. the percentage of
zinc scorified and retained by the slag is relatively small. The vapor is rapidty
cooled, the metals condensing as powder, which is subsequently melted down,
affording a very leady spelter, which is finally refined by redistillation in an
electric arc furnace, but of course an ordinary grade of spelter can be produced
by simple gravity refining. The ore smelting furnaces have now been built
and successfully operated for large capacity, as much as 12,000 kg. of ore per
24 hours being smelted in a simple furnace. This in itself is an achievement
beyond what anyone else has accomplished in this field and is to the credit of
the Swedish metallurgists.
Hydrometallurgy.—The extraction of zinc by purely hydro metallurgical meth¬
ods has never attained any great importance, indeed scarcely anything but occa¬
sional application. Owing to the formation of insoluble zinc ferrite during the
roasting the extraction of zinc is apt to be disappointing. Given a sulphate
solution, white vitriol (zinc sulphate) may be obtained as a commercial product
by evaporation and crystallization as used to be done in the Lower Harz, Germany,
but the market for that product is limited. However, such sulphate solutions
may be, and are, used for the production of lithopone by precipitation with
barium sulphide. To obtain zinc oxide from a zinc sulphate solution is difficult,
but it was once done on a commercial scale by the Parnell process, which was
based on the reaction between zinc sulphate and zinc sulphide when heated in a
muffle furnace.
Hydro-electrometallurgy.—The electrolytic production of zinc, which has now
become a great commercial art, is based upon the ordinary hydrometallurgical
steps up to the point of obtaining a purified solution of zinc sulphate. The same
difficulties of obtaining a satisfactory dissolving of the zinc obtain here, but need
not be insuperable for the lixiviation residues may be burned in Wetherill fur¬
naces, or reverberatory furnaces, for the expulsion of the zinc that they retain.
There are the other difficulties, such as result from gelatinization of silica, re-
precipitation of zinc, etc., that are to be expected in a hydrometallurgical process.
They are minimized by starting with a relatively pure ore. One of the factors
contributing to the success of electrolytic zinc extraction during the last 5
years, whereas all attempts for 25 years previously had been failures, is that the
flotation process gave us a relatively clean, high-grade ore, whereas previous work
had been done on low-grade, impure run-of-mine ore.
The essential conditions for successful electrolytic deposition of zinc are purity of
anodes (lead,) cathodes (aluminum) and electrolyte (solution of zinc sulphate), and
by purity is meant not merely relative purity but a high degree of it. Some foreign
elements, the alkalis for example, may be present with impunity, others such as
manganese may even be desirable, but others, such as cobalt, are poisonous even if
present only to the amount of a few hundredths of a per cent. Consequently, after
the solution of zinc has been obtained it must be purified by the precipitation of ob¬
jectionable elements. This is an essential and often a troublesome part of the proc¬
ess. Iron is converted to the ferric state by the manganic acid, or acid salts, that are
present and is precipitated by lime. Copper and cadmium are thrown down by zinc.
These operations are relatively simple, but not so the removal of chlorine and cobalt.
However, the electrolysis of a solution that is not refined quite so much as it ought to
be may be aided by the addition of glue.
The vats are commonly arranged in cascades. The electrolyte is introduced into
the head cell made up to about 60 g. zinc and 25 g. free acid per liter. A difference
of 3.6 volts in potential and a current density of 25 amp. per square foot are common
practice. In its passage through the cascade the solution heats. Its temperature
is advisedly controlled by means of cooling pipes interposed between the cells. With
the gradual peroxidation of the manganese the electrolyte assumes a deeper and deeper
pinkish hue. Finally it issues from the last cell with a zinc content of about 30 g.
and an acidity of about 75 g. per liter, and is returned for the lixiviation of a fresh
charge of ore. The leaching is commonly done in Pachuca tanks; the separation of
residues and the filtration of the solution by means of Dorr classifiers and Oliver filters.
The mechanical arrangements, the difficulties and the limitations of the hydrometal-
lurgical part of the process are similar to what obtain in any process of that character.
The arrangement of the electrolytic tank room is similar to that of a copper refinery.
In zinc electrolysis, however, the liberation of an acid mist is considerably more
objectionable.
Rotating cathodes have certain theoretical advantages. They were tried in the
Judge plant at Park City, Utah, but they did not develop satisfactorily. However,
such cathodes have been used in the electrolysis of zinc-chloride solution according
to the Hoepfner process. In that process, which has been applied on a small scale at
Fiihrfort in Germany and at Wilmington. in England, chlorine is liberated at the anode
and is conducted away for the manufacture of bleaching powder.
In good practice in zinc electrolysis there is a deposition of about 10 lb. of zinc
per horsepower per day. At $30 per horsepower year this is about 0.8 ct. per pound
of zinc. The other costs—for leaching, purifying, melting, etc.—raise the total
per pound of zinc to a high figure, which must be offset by superior advantages in
other respects, such as superior extraction of zinc, lead and silver. The conditions
under which cheaper power and offsetting advantages exist are rather rare and conse¬
quently electrolytic zinc extraction is hardly to be expected to become a process of
universal application, or likely to displace distilling, although the production of
electrolytic zinc has had and will continue to have a profound effect upon commercial
conditions. Already it has destroyed the historic large premium for high-grade zinc.
METALLURGY OF LEAD
By Walter Renton Ingalls1
The art of extracting lead from its ores is based on these principles: (1) the
reduction of lead oxide by carbon, or carbon monoxide; (2) the reaction between
lead sulphide and lead sulphate or oxide, resulting in a double decomposition
with the formation of lead and sulphur dioxide; and (3) the decomposition of
lead sulphide by metallic iron. All of these reactions are endothermic, i.e., they
absorb heat, which must be supplied extraneously. The first principle is the
basis of what is known as the roast-reduction method of smelting, in which lead
oxide (or lead silicate, or some other oxidized compound) must be prepared from
sulphide ore by a preliminary roasting. The second is what is known as the
roast-reaction method. And the third is the basis of the precipitation method.
METALLURGY OF LEAD 803
All of these methods are employed practically in the United States at the present
time, either alone or in combination, the latter being most commonly the case;
also, the principles are not practically so sharp as theoretically stated, the
reactions fundamental to one process invariably playing a certain part in the
other processes.
If a lead ore were absolutely pure, there would be no other consideration in the
smelting processes than those reactions which relate to lead and its chemical combina¬
tions; but practically lead ores are never pure, the valuable minerals being mixed with
a certain proportion of foreign matter, which must be separated by making a slag
of it. In*the roast-reaction method of smelting, it is true, no slag properly speaking
need be made, the reduced lead being liquated out from the worthless impurities,
but the latter will still retain a high percentage of lead, and in order to effect a high
degree of extraction from the ore it must be subjected to a further smelting process
in which a true slag is made. In making a slag, the object is to combine the impurities
into a fusible silicate, which when molten will be of comparatively low specific gravity,
so that the heavier lead will settle to the bottom of the crucible of the furnace, whence
it may be drawn off separately, while the lighter slag will float on top. Besides the
slag and metal, there is formed practically another substance, called matte, which is
lighter than the metal but heavier than slag. Matte is an artificial sulphide, consisting
in lead smelting usually of the sulphides of iron, lead, and copper (if copper be present
in the ore), and owes its origin to the incomplete elimination of sulphur from the ore.
Similarly, speiss is an artificial arsenide of iron.
The difference between lead smelting in principle and in practice is chiefly due to
the incompleteness with which the basic reactions are carried out, and the qualifying
effect of the impurities that are commonly met with in ores. Thus certain metallic
impurities are reduced with the lead, contaminating it and necessitating a subsequent
refining process. Other impurities affect the composition of suitable slags. Others
affect the running of the furnace in various ways. It is the reconciliation and neutrali¬
zation of these various difficulties that constitute the duty of the metallurgist, whose
purpose is to treat a ton of ore in such way that the difference between the value of the
products extracted and the expense of treatment will be the maximum, the value
of the products depending largely upon the percentage extracted, and the expense
of treatment including interest and amortization charges on the capital invested in
plant, etc.
The most primitive form of lead smelting in the United States was practiced
with the log and ash furnaces in Missouri prior to 1850. They employed the
roast-reaction system of smelting and'were applicable only to non-argentiferous
galenas, of very high grade in lead. They were for the most part displaced about
1840 by the Scotch1 hearth furnace. Later on reverberatory furnaces of the
Flintshire type came into use, but never extensively, nor with such success as
to develop a permanent metallurgical practice. The Scotch hearth, however,
has survived, and in its modern development is the basis of a highly efficient
process in the treatment of certain classes of ore, especially high-grade non-
argentiferous galena.
Neither the Scotch hearth nor the reverberatory furnace is well adapted to the
treatment of argentiferous ore, or to the treatment of ore containing less than 60
per cent of lead, while the smelting of carbonate ores alone in these ways is of course
1 This term is not strictly accurate, since the furnaces of this class that were introduced into America
were early modified materially from their prototype as employed in Great Britain. They are des¬
cribed more precisely as the American-Scotch hearth, which term is to be found in many metallurgical
treatises, but is never heard in practice, the term Scotch hearth having been adopted for brevity.
infeasible. For these reasons, by far the major part of the world’s lead is produced by
smelting in the blast furnace.
In the early history of blast-furnace smelting of silver-lead ore in the United

States, the work was done chiefly for the reduction of the ore oi a, single mine,
or single group of mines, as at Eureka, Nev., Cerro Gordo, Gal., and at several
places in Utah. The charges were generally rich in lead. Their silver and gold
contents went of course chiefly into the lead, but although they wore commonly
the most important elements of value in the ore, the process was essentially one
of lead smelting.
Later on, when many of the rich lead deposits became exhausted, when the tonnage
of rebellious1 gold and silver ores greatly increased, when transportation faeilitieH
had been largely improved, and when the smelting industry had become concen¬
trated in a comparatively few centers, drawing ore supplies from numerous districts,
the average percentage of lead in the ore smelted fell to a low figure, and although
in the aggregate the production of lead increased immensely the process of reduction
became essentially one of gold and silver smelting, the lead in the ore functioning an
collector of the precious metals and being viewed mctallurgically from that standpoint.
Subdivisions of Lead Smelting.—It is useful, therefore, to consider the art

of lead smelting in its present development as being divided under three* heads,
viz.:
1. Reduction of non-argentiferous galena, in which all the methods of smelting are

applicable, if the ore be sufficiently rich in lead.
2. Reduction of special classes of argentiferous ore, the product of one mint* or a
few mines, in which the blast furnace only is advisable. The character of the charge
and the bullion are chiefly dependent upon the composition of the ore, while the slag
is governed by the composition of the ore and the availability of the necessary fluxes
and may vary within rather wide limits.
3. Reduction of general mixtures of argentiferous ores, the product, of many mines.
The most important part of the present lead-smelting industry falls under this head.
The ores are mixed to make a slag of definite composition and a bullion of nearly
uniform grade, which vary only within narrow limits. The percentage of haul in the
composite ore, i.e.f the furnace charge, is comparatively low, and is designed to be
only what is necessary to collect the gold and silver most efficiently. To improve the
collection of gold and silver a proportion of copper is generally added to the charge,
which results in the production of a leady cupriferous matte* as a hetween-produet.
The third system is essentially gold and silver smelting on a haul basis, and is
commercially analogous precisely to smelting on a copper basis. These two kinds
of smelting are indeed competitive, the same ores being to a large extent amenable to
treatment in either way. However, the copper smelter is not desirous of receiving
lead-bearing ores, while the lead smelter needs a certain proportion of copper-bearing
ore. In the treatment of mixtures properly compounded, smelting on a copper Imsw
is more profitable than on a lead basis, and the tendency of metallurgical practice
during the last 10 years has been toward its adoption wherever feasible. The two
kinds of smelting are now to be witnessed in operation at the same places, and even
side by side in the same works (vide Salt Lake City, Utah; Aguas Calientes, Mexico,
and elsewhere). The successful collection of lead fume from the copper converter
blowing leady matte has made some copper smelters less averse to haul in their on*
than they used to be.
1 Thus word when applied to gold and silver ores indicates those of a character not readily ummmbld
to the milling processes.
In smelting any kind of lead ore in any way, the sulphur must be burned off,
partly at least, and the impurities must be combined in a slag, fusible at approxi¬
mately 1,100 to 1,200°C., and of specific gravity not to exceed 3.6, or but little
more. The slag must be of a composition which will form at the right point
in the smelting process, will be thoroughly liquid, in order to insure a satis¬
factory separation from the matte, and will require the minimum consumption
of fuel, the chief part of which in blast-furnace smelting is always consumed in
effecting the formation of the slag. A good slag will cause the charge of ore
and flux to descend evenly and regularly in the furnace, avoiding the formation
of accretions in the hearth or on the walls, keeping the lead in the crucible red-
hot and preventing any creeping up of the fire and the ensuing danger of loss of
lead and silver by volatilization.
The sulphur is eliminated in various ways, which may be enumerated as
follows:
1. Roasting.—Performed usually in hand-raked reverberatory furnaces with a
single, long hearth. Bruckner cylinders are also employed, but most of the other
forms of mechanical furnaces are inefficient because of the stickiness of lead ore in
roasting, wherefore the attempts to apply them have been generally abandoned. Lead
ores, being usually low in sulphur, are not self-burning in the ordinary roasting furnace,
and decomposition of lead sulphate being effected only by reaction with silica, forming
lead silicate, the necessary temperature must be chiefly supplied by the combustion
of carbonaceous fuel. In good practice the consumption of coal (Colorado bituminous
and other western grades) is about 30 per cent of the weight of the ore roasted. The
roasting of lead ore may be done in three ways, viz.:
(а) Ordinary roasting, in which the ore is simply desulphurized, at the same time
being more or less sintered. Often the ore is partially fused, so that upon withdrawal
from the furnace it crusts or may be pounded into cakes, which may be called “sinter-
roasting.”
(б) Slag roasting, in which the ore is completely fused.
(c) Blast roasting, a recent improvement, which will be discussed more fully further
on.1
Slag roasting is performed in modifications of the ordinary reverberatory furnace,
designed so as to permit the maintenance of a suitably high temperature at the
discharge end.
It is objective in roasting to reduce the sulphur as low as possible, without
entailing undue losses in other directions, since it is reckoned roughly that each
unit of sulphur remaining in the ore increases the cost of smelting 25 cts. per
ton (under pre-war conditions) The sulphur is more completely eliminated by
sinter roasting than by ordinary roasting, and more completely by slag roasting
than by sinter roasting, but at the same time the loss of lead by volatilization is
heavily increased and in slag roasting is so high that the process has been aban¬
doned in the United States, save in one or two instances, while sinter roasting
has had no wide practice, and ordinary roasting, by which the sulphur is reduced
to about 4 per cent, had become the generally adopted method until recently
displaced by the “blast-roasting” process.
2. Roast Reaction.—Lead sulphide burned partially to sulphate reacts with unde¬
composed sulphide, setting free metallic lead and sulphur dioxide. This process is
effected in the reverberatory smelting furnace (Flintshire, Tarnowitz, etc.) wherein the
1 See page 820.
charge is first partially roasted and the reaction is then effected under increase of
temperature; also in the Scotch hearth, wherein the roasting and reacting go on
contemporaneously. In the modern blast furnace, which has lines and is operated
under conditions promoting oxidation, roasting and reacting play an important part.
3. Precipitation.—In this method the ore is charged raw into the blast furnace and
the lead sulphide is decomposed by iron, precipitating metallic lead, while the sulphur
combines with the iron, forming a matte, from which the sulphur is subsequently
eliminated by roasting. The quantity of matte to be roasted is apt to be as much as
the quantity of ore smelted, but the loss of lead is less than if the ore were roasted
originally. There are other drawbacks, however, and this method which has never
had any wide application in the United States is now employed by itself in only one
or two instances, if at all. The reaction takes place to some extent in any blast fur¬
nace, however, and it together with the reaction between the sulphide and sulphate
described under the previous caption is largely relied on in modem practice, wherein
galenas of high grade in lead are commonly charged raw into the blast furnace to
save the comparatively high loss of lead which is inevitably suffered even in ordinary
roasting. Precipitation smelting may also be done in the reverberatory furnace, but
that method is quite inefficient and has seldom been practised either in America or
Europe.
Fig. 3.—Blast furnace, International Smelting Co., Tooele, Utah.
After elimination of sulphur by ordinary roasting or by blast roasting, the ore

is ready for smelting in the blast-furnace. Naturally oxidized ores go of course to the
blast furnace without any preliminary treatment, save breaking to the proper size.
The function of the blast furnace is to reduce the metal and slag the impurities
so that a separation can be made. This separation is effected by the difference
in specific gravity of the various products. The specific gravity of liquid (molten)
lead is about 10.5; that of matte ordinarily ranges from 4.5 to 5. In order to
insure a satisfactory separation of matte and slag, therefore, the specific gravity
of the latter should be at least one degree lower than the lightest matte. Both
the matte and the slag must be thoroughly liquid at the temperature of issue
from the furnace. Consequently the matte must be free from components which
make it mushy, like zinc, and the slag must be of a composition which is easily
fusible at 1,100 to 1,200°C. The cleanness of the slag, i.e., its freedom from gold,
silver, lead, and copper, is dependent chiefly on these physical conditions. The
lead content of a good slag should not exceed 1 per cent; the silver content should
not exceed 0.8 oz.; the gold content should not be more than a trace. In order
to achieve such a result the specific gravity of the slag should not be more than
3.6, and it should be of approved chemical composition.
In making a slag, the metallurgist is practically limited by the elements which
commonly occur in ores, and to cheap fluxes like limestone. The commercial slag is
a silicate of two or more bases. The fundamental constituents are silica, iron, and
lime. The iron may be replaced to some extent by manganese, and the lime to some
extent by magnesia, zinc, and baryta. These elements ordinarily constitute 90 per
cent of the slag. The remainder is chiefly alumina, barium sulphate, soda, and
potash, etc. In preparing the furnace charge, a mixture of ores and fluxes is made
in such proportions as to produce a slag of the desired composition,. In lead smelting
the permissible range in silica content is rather narrow, varying from 28 to 36 per
cent, and ordinarily also the percentages of ferrous oxide and lime range only within
narrow limits, the percentages stated nominally as ferrous oxide and lime including
their equivalents respectively of manganese oxide and magnesia or zinc oxide.
The efficiency of the lead-smelting process depends chiefly upon (1) the method
of sulphur elimination, affecting the cost of the operation, the loss of gold, silver, and
lead, and the physical condition of the product delivered to the blast furnace; (2)
the design and construction of the blast furnace and its accessories, the method of
operation, the physical condition of the material charged, the correct composition of
the slag, the character of the bullion produced and the means to collect dust and fume;
(3) the means for handling products throughout the process.
The better the character of the slag, the faster will be the running of the furnace,
i.e., the greater the quantity of charge smelted per square foot of hearth area,1 and
the lower the loss of lead and silver in the slag and in the fume. The lower the
proportion of very fine ore in the charge, the faster will be the running of the furnace
and the smaller the percentage of flue dust to be collected and rehandled. The
better the character of the bullion, the less irregularity in the management of the
furnace and the less the cost of refining. A detailed consideration of all of these
governing and largely interdependent factors is impossible in a brief statement of the
principles of the art as developed in modem practice, but such will be found in the
standard treatises on the subject. It is aimed here to emphasize the predominant
effect of the slag character, i.e., the make-up of the furnace charge, in determining the
cost and efficiency of the smelting process. It may also be said that the successful
development of the silver-lead smelting industry hinged upon this matter. As the
histories of the various lead-smelting districts are examined, it will appear strikingly
how the process was quickly made successful in those which had the advantage of a
self-fluxing ore, like Eureka, Nev., and how it was struggled with in many other dis¬
tricts, and how, as lead smelting became a well understood art, it came to be recog¬
nized that an assured and adequate supply of the right classes of ore was the funda¬
mental prerequisite for the inauguration of a successful business.
Given supplies of silica, iron, and lime, there is no difficulty in producing a slag
of any desired composition, but the smelting might be so costly as to be unprofitable.
The cost of smelting is properly based on the charge, i.e., all of the material except
fuel which is put through the furnace. It is the aim of the metallurgist to have the
ore in the charge at the maximum possible percentage, i.e., it is sought to obtain all of
the necessary fluxing elements in a proper proportion in the ores purchased for
smelting, since the time, fuel, and labor required to smelt a pound of flux are the same
as for a pound of ore, and for each pound of flux put into the charge there is one pound
less, of ore. It is consequently of the utmost importance to the smelter to be able
i The hearth area is the horizontal section o f th e furnace on the line of the tuyeres.
to secure ores that upon mixture will contain the proper proportion of slag-forming
substances and will obviate the necessity of making up the proper proportion by the
addition of barren fluxes. It is seldom possible to attain the acme of that condition,
a small proportion of lime-rock having usually to be added to the charge while there
is always a certain quantity of foul slag, flue dust, roasted matte, and other between-
products that must be added to the charge; but in the modern American practice of
silver-lead smelting the ore in the charge amounts usually to 80 per cent or more.1
The price which a smelter will give for any particular ore is consequently
determined by his fluxing requirements and the supply of various kinds of ore on
hand or offered. The smelter refers to the difference between the cost of a ton of
ore and the proceeds which he obtains from it as the “ margin/J out of which he
has to pay the cost of smelting, the remainder of the margin being his profit.
The margin varies greatly according to the class of ore treated and the profit is
dependent upon the average; consequently it is sought to smelt as much ahigh-
margin5 ’ ore as possible.
As a simple illustration, suppose there were available two kinds of ore, viz., a lead-
carbonate ore, containing 22.5 per cent lead, 20 per cent ferrous oxide, 4 per cent lime,
and 27 per cent silica; and an argentiferous iron ores, containing 60 per cent ferrous
oxide, 4 per cent lime, and 25 per cent silica; and lime-rock containing 1 per cent ferrous
oxide, 50.5 per cent lime, and 8 per cent silica. Suppose it were desired to mix these
ores so that the slag would have the composition 34 per cent silica, 34 per cent ferrous
oxide, and 17 per cent lime, it would be necessary to add to each ton of the lead ore
0.25 ton of iron ore and 0.25 ton of lime-rock. The mixture would contain 15 per
cent lead. In the daily charge of a furnace, say 150 tons, there would be 100 tons of
lead ore, 25 tons of iron ore, and 25 tons of lime-rock. Assume the cost of smelting
to be $3.75 X 6 -f- 5 = $4.50. In order to realize a profit of $1 per ton, assuming
that the full value is paid for gold, silver, and lead, the smelter must have a margin of
$5.50 per ton. But suppose the iron ore is so low in silver that it will not stand a
treatment charge of more than $2 per ton, the smelter obliged to have it as a flux
must make up the deficit by an increased charge on the lead ore, which then would
have to be $6.37}^.
This illustration shows also what advantage it would be to the smelter if he were
able to substitute for the lime-rock an ore of similar content in lime, on which he could
make a treatment charge. It shows also the advantage of making a slag as siliceous
as possible, which enables the maximum percentage of the usually profitable siliceous
ore to be carried in the charge.
The foregoing analysis of the commercial principles of lead smelting relates espe¬
cially to the conditions involved in the operation of a custom smeltery, supplied by ore
from numerous mines, in which the percentage of lead in the aggregate is compara¬
tively small, the business being conducted not with the particular view of producing
lead, but rather with that of smelting a large tonnage of ore. Nevertheless the same
fundamental principles govern the smelter of the product of a single mine, or group of
mines, although he may be more constrained in the matter of ore mixtures; and also
the smelter of high-grade galenas, who may extract directly a large proportion of the
lead in Scotch hearth or Flintshire furnaces, but is still obliged to effect the final
separation of the gangue by smelting the gray slag in a blast furnace.
1 This means 80 per cent or more of the charge, exclusive of the slag which is used in larger or smaller
quantity, often rather large, to facilitate the running of the furnace. Great efforts are made to eliminate
this “side-charge” but it does not appear possible to do so.
The practice of the metallurgy of lead has resolved itself at the present time
into smelting in the blast furnace for argentiferous ores and smelting in the
Scotch hearth for non-argentiferous ores. In both kinds of smelting the smoke
and fume are filtered by means of baghouses, the use of which has greatly reduced
the loss of lead. In silver-lead smelting the loss of lead has been further reduced
by the introduction of the blast-roasting process, which has eliminated the large
loss that was formerly incurred in roasting in reverberatories. The result of these
improvements has been to make it possible to extract 95 per cent of the lead in
smelting silver-lead ores in the blast furnace and 98 per cent in smelting high-
grade galena in the Scotch hearth. The two methods of smelting are applied
under radically different conditions. In silver-lead smelting a prime function
of the lead is to act as collector for the precious metals. The percentage of lead
in the charge is ordinarily low, say about 10 per cent, but there are regions where
the main supply of ore is galena and in those regions the furnace charges are
commonly much higher in lead. In any case the proportion of slag to ore that is
made is much higher than in the treatment of high-grade galena ore in the
Scotch hearth and consequently loss of metal in the slag is higher. The Scotch
hearth method of smelting has been confined to the treatment of non-argentif¬
erous galena containing 60 per cent of lead, or upward. This method of smelting
has been improved by the introduction of the Newnam mechanical rabbler,
which has reduced and ameliorated labor conditions, an important thing in
view of the severity of the latter, especially in hot weather.
Except at Port Pirie, South Australia, the metallurgy of lead is technically less
advanced than in the United States and Mexico. Spain, which is one of the four great
lead-producing countries of the world, is distinctly backward. It is probable that the
Spaniards, in revising their metallurgy, will adopt American methods.
ROASTING
By F. R. Pyne1
Definition.—Roasting is the heating to an elevated temperature, without

fusion, or at most to incipient fusion, of ores or metallic compounds in contact with
oxidizing, reducing or chloridizing materials, in order to produce a chemical
change or to eliminate a component by volatilization. It is a preliminary step
to many metallurgical operations, the usual object being partially or completely to
eliminate sulphur, arsenic, antimony, or other readily volatile constituents, such
as carbon dioxide, which may interfere with subsequent processes.
Types of Roasting.—Roasting may be divided into three principal classes:
oxidizing, reducing, and chloridizing. In the oxidizing roast the operation is
performed by heating with excess of air to such a temperature that the affinity of
the undesirable element for oxygen becomes stronger than for the metal with
which it is associated. In the reducing roast an oxidized metallic compound
is heated in the presence of a strong reducing agent, such as carbon, to a tempera¬
ture sufficient to cause the deoxidization. In the chloridizing roast the object
is to convert the metallic compounds into chlorides by treatment with chlorine
gas or chlorides.
The most common and important type of roast is the oxidizing roast of metallic
sulphides. When a metallic-sulphide is heated to a sufficiently high temperature with
access of air, there is formed sulphur dioxide and a metallic oxide:
2MS + 302 = 2MO + 2S02
The temperature, in degrees Centigrade, at which the leading metallic sulphides
ignite is shown by the following table.2
j
Size of grain, millimeters
Material
1 1-2 Over 2
Pyrite. 325 405 472

Pyrrhotite. 430 525 590
FeS. 535
Ni 73.3 S 26.7. 700 802 886
Co 66.37 S 33.63. 574 684 859
Co 70.20 S 29.80. 514 751 1,019
Stibnite.1 290 340
Molybdenite. 240 508
Cinnabar. 338 420
Chalcocite:. 430 679
Bi. 83.3 S 16.7. 500 626
Mu 61.01 S 3?.98 Fe 2.02. 355 700
Argentite. 605 875
Blende.1 647 810
Galena (in oxygen). 554 847
Millerite. 573 616
1 Works superintendent, U. S. Metals Refining Co., Chrome, N. J.

2 Hofman, “General Metallurgy,” p. 404.
M KTALLVRG Y.ROASTING 811
Proper oxidation requires a temperature suflici ly high to produce the necessary

afhnity between the oxygen and the sulphur, but it must not be so high as to cause
the surface o( the particles ol ore to melt and form a protecting layer which retards the
oxidizing action and may even cause it to cease. This temperature may be obtained
by the combustion of extraneous fuel or by the heat of oxidation of the constituents
of the charge.
A second requirement for proper oxidation is a constant and abundant supply of
air in immediate contact with the surface to be oxidized. The necessity for an
abundant supply of air is duo to the fact that the oxidizing power of the air is lowered
out of all proportion when diluted. In other words, if the air for oxidization is diluted
with its own volume of gaseous furnace products, the oxidizing effect of the resulting
mixture will ho much less than half the oxidizing power of pure air. Therefore, the
more rapidly the gaseous furnace products are removed and replaced by pure air
the more rapid will be the roast.
The speed and thoroughness of the roast are also governed by the size of the
particles, ft is obvious that the finer the ore is crushed the greater will he the
surface exposed to the oxidizing influences. As fine crushing always results in
losses by handling, flue dust, etc., it will he seen that there is an economical point
when the advantages gained by the increased speed of roasting are balanced by
the dust losses which necessitate expensive recovery systems. The usual charge
of ore to the, roasters consists of coarse and fine particles, and the latter will be
thoroughly rousted long before the former and to obtain a dead roast requires
extensive stirring and heating. But should only a partial elimination of sulphur
be desired, as in copper smelting, the roast is carried to a point where the over
roasting of the fines balances tilt's under roasting of the coarser particles, so that the
final mixtures give the desired sulphur content.
Sulphatizing Roasting. The gases produced are more or less rich in 80s»
depending upon the. amount of air furnished per unit of material, and in the
presence of contact substances the SO* may be converted into 80a
80a + o?=*a()i
The 80a formed may combine with metallic oxides to form sulphates
MO + 80a^M8()4
In some types of rousting the aim is to produce a maximum amount of sulphates, in

which cane the ojmration is known as a sulphurizing roasting. The amount of
SO* present in the gums and the existing temperature determines the formation
of the sulphates. The reaction is-a reversible one depending upon the temjwmte
tore and pressure of the 80a,1 The following table taken from Fulton’s “ General
Metallurgy,” page 218, gives the equilibrium conditions for temperature and
pressure of 804 for several sulphates.
* I haw »mtm doubt an to whofchtr thin matter of temporatura rango hi th« formation anti dmotn*
imtdiion of nulphate# hi roantmg m tin limpk a® It i« usually rwprtwmted. Thorn in much morn oxporh
mentation ntwdod to rWr thr* withjnot, hut I am i noli nod to hollow, from mnm oxporimontal ovidnnoo hi
my tummmlmt, that varying ration of water vapor to tho sulphur trioxido prnient profoundly alter th«t
rate of formation of rndphatca and iho quantity formnd. Air Gonditiontng may y«t bo eonsidwod an
o»»ontt&l for tbif propnr control of aulphatixing rcm»th*g,~Knmm,
Fe2S04 A12(S04)3 2CuS04 2Cu0S03 Z11SO4

Fe203 + so3 Ahos + S03 2CuOS03+S03 2CuO + SO 3 ZnO + S03
Temp- Milli- Temp- Milli- Temp- Milli- Temp- Milli- Temp- Milli-
erature meters erature meters erature meters erature m eters erature meters
553 23 572 28 546 43 600 62 675 5

570 33 621 51 588 55 653 98 690 6
592 45 681 120 615 70 686 123 720 24
614 70 702 180 642 98 705 139 750 61
634 113 720 261 665 130 728 173 775 112
650 149 731 356 700 233 745 209 800 189
660 182 742 480 714 324 775 298
680 286 74S 692 725 460 805 542
690 401 731 647
699 560
707 715
i
From the above table it is seen that for sulphatizing roasting a very careful tem¬
perature regulation is necessary. For example, if we take copper sulphate it is seen
that at a temperature of 615°C. the dissociation pressure is 70 mm. If the partial
pressure in the furnace above the copper sulphate is less than 70 mm., the sulphate
will dissociate into CuO and SO3; if the partial pressure is greater than 70 mm.,
dissociation will not take place. However,’ should the temperature of the furnace
be raised while the partial pressure of the S03 remains constant, the dissociation
will be rapid. The gases must, therefore, be removed from the furnace very slowly
in order that the partial pressure of the S03 be kept above the critical point.
If it is not desired to roast to sulphate but only for the elimination of sulphur, the
furnace gases should be removed as rapidly as possible and the temperature of the
furnace limited only by the sintering of the ore as this prevents complete roasting of
the particles.
Blast roasting is the name given to the process of forcing air through finely
divided ignited sulphides, the object being to roast and sinter in a single operation.
This type of roasting is a rapid form of roast in which the SO2 gases are with¬
drawn very rapidly and the oxidation proceeds with such a speed that the heat
generated is sufficient to form a sintered mass of the resulting oxides and the
gangue or flux , in the ore. A careful temperature control is necessary for if too
much heat is generated the charge fuses before roasting has progressed to the
desired degree, while if the temperature is too low the resulting product is not
properly sintered. Excess heat is taken Care of by the following methods:
(a) Wetting the charge.
(b) Preliminary or rough roasting to reduce the amount of available sulphur.
(c) Diluting the ore mixture with inactive fluxes thus reducing the calorific
power of the mixture. This method is particularly beneficial in that it keeps
the mixture open and permits of a faster roast.
Roasting Arsenides and Antimonides.—When these materials are roasted
with free access of air at a low temperature, a part of the arsenic or antimony
is oxidized to AS2O3 orSb203 which is readily volatile. Further oxidation changes
METALLURGY-ROASTING 813
these oxides into AS2O5 or Sb205 which combine with metallic oxides to form very
stable arsenates and antimonates. However, if sulphides are present the sulphur
vapor and SO2 gas reduce the arsenates and antimonates to the sulphide and to
the lower oxides which are readily volatile. In general arsenic and antimony
are most readily removed by alternate oxidation and reduction.
The reducing roast is used to effect a deoxidation below the fusion point, the
material being heated in contact with carbon, sulphur, or reducing gases. An example
of the reduction of metallic oxides is the heating of hematite, Fe203, with coal to form
the magnetic oxide Fe304.
3Fe2Os + C = 2Fe304 + CO
The chloridizing roast is used to convert the metals into chlorides prepara¬
tory to treatment with solvents. The usual source of chlorine is common salt
and it is essential to the success of the process to have sulphides present. A
very careful temperature regulation is necessary.
The objects of chloridizing roasting are (Greenwalt, “Hydromctallurgy of
Copper”)^
1. In copper ores, or in gold or silver ores containing copper, to convert the copper
into chlorides which are directly soluble in water or in chloride solutions.
2. In silver ores, or in gold and copper ores containing silver, to convert insoluble
metallic silver or its insoluble compounds into the more soluble silver chloride.
3. In any ore, to convert the harmful elements into less harmful compounds.
4. To assist in a more efficient oxidizing action than is possible under the same
conditions in ordinary oxidizing roasting.
Ottakar Hofman1 classified the adaptability of various ores to chloridizing roasting
as follows:
1. Those like iron and copper pyrites, gray copper ore, and silver copper glance
which in roasting form sulphates and decompose salt, liberating chlorine.
2. Those like galena and blende, which form sulphates remaining indifferent to
salt.
3. Antimonial and arsenical silver minerals, which form antimonates and arsenates
of silver.
4. The gangue remains indifferent like quartz or porphry, or it takes an active
part like limestone, talc, spar, manganese, or minerals containing magnesia.
If the ore consists of minerals of the first group together with an indifferent gangue,
chloridizing roasting offers no difficulty and a high chloridization can be obtained with¬
out much loss of silver by volatilization and no special skill is required in the roasting;
neither does it matter if the salt is added to the charge before entering the furnace
or after it has been subjected to a partial oxidizing roasting.
If one or both of the minerals of the second class are present in any quantity the
process of chloridizing roasting becomes more difficult.* In such a case the time
of addition of the salt is very important. If added before the charge enters the furnace
a very inferior chloridization is obtained, as is also the case if the salt is added before
the oxidizing period has sufficiently advanced.
If all of the above classes of ore are represented the roasting is still more difficult,
particularly with a gangue like limestone which takes an active and injurious part in
the operation.
Chemistry of Chloridizing Roasting.—The sulphides of iron and copper are
mostly relied upon for chloridization. The sulphates formed during the roast¬
ing react with the salt to form sodium sulphate and the chlorides of the metals.
1 ** The Mineral Industry,” 1896.
Fe2S04 A12(S04)3 2CuS04 2Cu0S03 ZnS04

Feo03 + so3 A1203 + S03 2Cu0-S03+S03 2CuO + SO 3 ZnO -j- S03
Temp- i Milli- Temp- Milli- Temp- Milli- * ' Temp- Milli- Temp- Milli-
erature ' meters erature meters erature meters eraturei meters erature meters
553 23 572 28 546 43 600 62 675 5

570 33 621 51 588 55 653 98 690 6
592 45 681 120 615 70 686 123 720 24
614 70 702 180 642 98 705 139 750 61
634 113 720 261 665 130 728 173 775 112
650 149 731 356 700 233 745 209 800 189
660 182 742 480 714 324 775 298
680 286 748 692 725 460 805 542
690 401 731 647
699 560
707 715
i
From the above table it is seen that for sulphatizing roasting a very careful tem¬
perature regulation is necessary. For example, if we take copper sulphate it is seen
that at a temperature of 615°C. the dissociation pressure is 70 mm. If the partial
pressure in the furnace above the copper sulphate is less than 70 mm., the sulphate
will dissociate into CuO and S03; if the partial pressure is greater than 70 mm.,
dissociation will not take place. However, should the temperature of the furnace
be raised while the partial pressure of the S03 remains constant, the dissociation
will be rapid. The gases must, therefore, be removed from the furnace very slowly
in order that the partial pressure of the S03 be kept above the critical point.
If it is not desired to roast to sulphate but only for the elimination of sulphur, the
furnace gases should be removed as rapidly as possible and the temperature of the
furnace limited only by the sintering of the ore as this prevents complete roasting of
the particles.
Blast roasting is the name given to the process of forcing air through finely
divided ignited sulphides, the object being to roast and sinter in a single operation.
This type of roasting is a rapid form of roast in which the SO2 gases are with¬
drawn very rapidly and the oxidation proceeds with such a speed that the heat
generated is sufficient to form a sintered mass of the resulting oxides and the
gangue or flux in the ore. A careful temperature control is necessary for if too
much heat is generated the charge fuses before roasting has progressed to the
desired degree, while if the temperature is too low the resulting product is not
properly sintered. Excess heat is taken Care of by the following methods:
(a) Wetting the charge.
(b) Preliminary or rough roasting to reduce the amount of available sulphur.
(c) Diluting the ore mixture -with inactive fluxes thus reducing the calorific
power of the mixture. This method is particularly beneficial in that it keeps
the mixture open and permits of a faster roast.
Roasting Arsenides and Antimonides.—When these materials are roasted
with free access of air. at a low temperature, a part of the arsenic or antimony
is oxidized to AS2O3 orSb203 which is readily volatile. Further oxidation changes
these oxides into AS2O5 or Sb205 which combine with metallic oxides to form very
stable arsenates and antimonates. However, if sulphides are present the sulphur
vapor and SO2 gas reduce the arsenates and antimonates to the sulphide and to
the lower oxides which are readily volatile. In general arsenic and antimony
are most readily removed by alternate oxidation and reduction.
The reducing roast is used to effect a deoxidation below the fusion point, the
material being heated in contact with carbon, sulphur, or reducing gases. An example
of the reduction of metallic oxides is the heating of hematite, Fe203, with coal to form
the magnetic oxide Fe304.
3Fe203 + C = 2Fe304 + CO
The chloridizing roast is used to convert the metals into chlorides prepara¬
tory to treatment with solvents. The usual source of chlorine is common salt
and it is essential to the success of the process to have sulphides present. A
very careful temperature regulation is necessary.
The objects of chloridizing roasting are (Grecnwalt, “Hydrometallurgy of
Copper”):
1. In copper ores, or in gold or silver ores containing copper, to convert the copper
into chlorides which are directly soluble in water or in chloride solutions.
2. In silver ores, or in gold and copper ores containing silver, to convert insoluble
metallic silver or its insoluble compounds into the more soluble silver chloride.
3. In any ore, to convert the harmful elements into less harmful compounds.
4. To assist in a more efficient oxidizing action than is possible under the same
conditions in ordinary oxidizing roasting.
Ottakar Hofman1 classified the adaptability of various ores to chloridizing roasting
as follows:
1. Those like iron and copper pyrites, gray copper ore, and silver copper glance
which in roasting form sulphates and decompose salt, liberating chlorine.
2. Those like galena and blende, which form sulphates remaining indifferent to
salt.
3. Antimonial and arsenical silver minerals, which form antimonates and arsenates
of silver.
4. The gangue remains indifferent like quartz or porphry, or it takes an active
part like limestone, talc, spar, manganese, or minerals containing magnesia.
If the ore consists of minerals of the first group together with an indifferent gangue,
chloridizing roasting offers no difficulty and a high chloridization can be obtained with¬
out much loss of silver by volatilization and no special skill is required in the roasting;
neither does it matter if the salt is added to the charge before entering the furnace
or after it has been subjected to a partial oxidizing roasting.
If one or both of the minerals of the second class are present in any quantity the
process of chloridizing roasting becomes more difficult.* In such a case the time
of addition of the salt is very important. If added before the charge enters the furnace
a very inferior chloridization is obtained, as is also bhe case if the salt is added before
the oxidizing period has sufficiently advanced.
If all of the above classes of ore are represented the roasting is still more difficult,
particularly with a gangue like limestone which takes an active and injurious part in
the operation.
Chemistry of Chloridizing Roasting.—The sulphides of iron and copper are
mostly relied upon for chloridization. The sulphates formed during the roast¬
ing react with the salt to form sodium sulphate and the chlorides of the metals.
l" The Mineral Industry,’1 1896.
At the same time some hydrochloric acid and chlorine may be formed due to
the action of sulphur trioxide and sulphuric acid:
2NaCl + FeSC>4 = Na2S04 + FeCl2

2NaCl + CuS04 = Na2S04 + CuCl2
2NaCl + 2S0s - Na2S04 + 2C1 + S02
2NaCl + H2SO4 = Na2S04 + 2HC1
The chlorides and chlorine thus formed react with the silver and silver sulphates
to form silver chloride:
FeCl2 + AgS04 = 2AgCl + FeS04

CuCl2 + AgS04 = 2AgCl + CuS04
2NaCl + AgS04 = 2AgCl + Na2S04
2HC1 + 2Ag +0 = 2AgCl + H20
Ag + Cl = AgCl
Arsenic and antimony form chlorides which are easily volatile. If the ore con¬
tains arsenic and antimony in large amounts, much of the silver may be converted
in the early stages of the roast into arsenates and antimonates.
Blende, zinc sulphide, in the presence of salt remains indifferent and does not
decompose the salt. Zinc sulphate is not decomposed. Zinc oxide may be converted
at a red heat into the chloride, which is very volatile.
Galena is converted into lead sulphate which does not react with the salt, and lead
oxide which may be converted into the chloride. Both lead oxide and chloride are
volatile.
Limestone has a deleterious effect as when roasted with the metallic sulphides it
changes in part to calcium sulphate and in part to calcium oxide. Should the ore
contain limestone in excess, the sulphides of iron and copper which are needed to
decompose the salt will be converted directly to the oxides and a low degree of chlori-
dization will result.
Percentage of Salt.—Ordinarily the amount of salt for silver ores will vary
between 1 and 5 per cent, but it is largely proportional to the amount of lime
or magnesia in the ore. Copper ores require a larger precentage of salt, 5 to
10 per cent being a fair average for an ore containing several per cent of copper.
As the copper content increases the percentage of salt must be increased propor¬
tionally. The time of adding the salt depends upon the composition of the ore.
An ore low in sulphur may have salt added before charging into the furnace, pre¬
ferably before the ore is crushed so as to get an intimate mixture. If the ore con¬
tains considerable sulphur, combined with iron and copper, it may be given a good
oxidizing roast before adding the salt and still have enough sulphur to chloridize
the silver. If copper is to be chloridized the ore should contain at least as much
sulphur as copper before the salt is added. Should the ore be thoroughly oxi¬
dized so that it does not contain sufficient sulphur, salt and pyrites may be added.
If the salt is added while there is a large excess of sulphur remaining in the ore,
it will be largely volatilized, but if added at the proper time the chloridization
takes place very rapidly.
Temperature is the all-important factor in chloridizing roasting. Any ore
chloridized at an excessive heat will volatilize much of the metals, but if the
temperature be kept at the lowest possible point at which the metals can be
M ETALL URGY—ROASTING 815
chloridized, the time of roasting and amount of air are more or less
immaterial.1
Roasting Apparatus.—The various types of apparatus used in roasting may
be roughly classified as follows:
1. Heaps
2. Stalls
3. Shaft Furnaces
4. Muffle Furnaces
5. Reverberatory Furnaces
6. Cylindrical furnaces with superimposed hearths
7. Blast-roasting apparatus
Roasting in Heaps.—This method of roasting, formerly of common occur¬
rence, has lost much of its former importance, It can be used only for coarse
ores, and where the ore is suitable for pyritic or partial pyritic smelting heap
roasting is being abandoned. The operation consists in piling the ore in the
form of a truncated pyramid on a bed of wood which serves to start the roast.
If the ore contains sufficient sulphur to maintain combustion, the roast will
proceed of its own accord, a shortage of sulphur will necessitate the admixture
of extraneous fuel.
The roast yard must be so situated that the heap is protected from strong winds
and so that the prevailing winds carry the sulphurous fumes away from the works.
Proper drainage is of importance as the ore after roasting is in a condition to facilitate
leaching, and heavy metal losses may occur through this source.2
The ore is crushed to about 2-in. size and the resulting fines screened out. A layer
of fines 4 to 6 in. in depth is spread on the ground and covered with a layer of cord-
wood, which is arranged so as to form a series of flues leading to chimneys in the cen¬
tral portions of the heap, the chimneys being constructed of boards of sufficient length
to reach above the finished heap. The area of the heap is a matter of convenience,
the important dimension, being the height which varies with the percentage of sul¬
phur in the ore. An ore carrying 15 per cent of sulphur can have a height of 8 ft.
above the bed of wood, while one with 35 per cent of sulphur cannot be over 5 ft.
The average height of heaps is about 6 ft.
After the cordwood has been placed the coarse ore is piled thereon, and the top
and sides of the pile covered with a layer of a somewhat finer ore. After igniting
the wood the heap is carefully watched and as the ore becomes ignited and the roast
makes progress, a thin layer of fines is spread on the sides and the top of the heap.
Should the temperature at any time become too high, the thickness of the layer of fines
is increased, while if too cool, some of the fines are removed. The flow of air for the
oxidation is regulated at the base of the heap by the addition or removal of the layer
of fines at that point. The time required for heap roasting is considerable, 60 to 70
days being required for a heap containing 250 to 300 tons. Upon the completion of
the roast the heap is stripped of its covering of fines and bhe thoroughly roasted coarse
ore sent to the blast furnaces, while any partially roasted or fused ore is sent to another
heap for further treatment.
The advantages of heap roasting are cheapness of plant and low cost of the roast.
The disadvantages are the length of time required, loss of metal by leaching, and, most
serious, the killing of the vegetation of the surrounding region.
j This forms the basis of such processes as the Swinburne-Ashcroft and Pohle-Croasdale for the
recovery of all the commercial metals by volatilizing as chlorides.
2 Or advantage may be taken of this fact for recovering metal as formerly done at some of the
Spanish mines.
Roasting in Stalls.—This is an improved modification of heap roasting i.

which the heap is enclosed on three sides by permanent brick walls, the fourt]
being added when the roast starts and being removed to facilitate tearing dowi
the heap at the completion of the roast. The chimneys and flues are also per
manent and the utilization of heat is more efficient while the draft and collectio]
of the gases are under better control.
Roasting in Shaft Furnaces.—In this type of furnace, which is generally usei
for coarse ores only, the roast is continuous, the ore and fuel being charged at th
top of the furnace and the roasted product withdrawn at the bottom. The amoun
of fuel required may vary from nothing with rich sulphide ores to 10 per cent witj
low sulphur ores. The air for the oxidation is furnished by the natural draf
of the furnace. The height of the furnace depends upon the size and characte
of the ore to be roasted; if the material is to move rapidly through the furnac
the latter is tapered toward the throat, while if it is desired to have the or
exposed for a considerable time to oxidizing influences the furnace is tapere
toward the bottom. When ores are roasted in shaft furnaces and it is not desire
to have the ore and fuel in intimate contact, a separate firebox is used. Thi
type of furnace may be used for roasting fine ores by allowing the fine materh
to slide over inclined shelves retarding its progress sufficiently to effect the desire
amount of oxidation. When used in this manner, however, a large amount c
dust is produced.1
Roasting in Muffle Furnaces.—This type of furnace is used where it is cssentu
that the products of combustion of the fuel be kept separate from the ore. Th
ores treated are usually fine, and are unsuitable if they are of such a nature tin
they become sticky during the roast. The air for roasting can be carcfull
regulated and the composition of the resulting gases kept under careful contr<
and any volatile metals readily condensed and recovered. As the heat from th
fuel has to pass through the walls and bottom of the muffle, the fuel cost is hig
and the progress of the roast slow. The chief commercial application is to blend*
Roasting in Reverberatory Furnaces.—The reverberatory furnace treai
fine ores, and the heat for the operation is furnished by burning fuel in a separal
compartment, the hot gaseous products passing over the ore which is spread in
thin layer on the hearth, the depth varying with the character of the ore and tl
ease of roasting. The air for the oxidation may be supplied as excess air in tl
products of combustion of the fuel, or, more generally, be admitted throng
the working doors. The ore is fed to the furnace at the cold end and is gradual]
worked toward the fire end, being stirred at longer or shorter intervals dependir
on the character of the ore and the thoroughness of the roast desired. Tl
stirring may be done by hand or by mechanical means. The former is expensn
and only applicable to small units and is rapidly being replaced by mechanic
stirring.
The ratio of the hearth area to the grate area depends upon the nature of the or
one requiring little heat for oxidation would be treated in a furnace having a hip
ratio of hearth to grate area and vice versa. The height of the roof above the hearl
depends to a great extent upon the character of the ore. For fuel economy the low*
the roof the more intimate is the contact between the ore and the hot gases. Wii
1 If for no other reason than its use for the production of quick lime, and calcined magnesite ai
dolomite, shaft roasting is of great importance commercially.
MET ALL URGY—ROASTING 817
ores low in sulphur and a highly oxidizing atmosphere, the height of the roof is de¬
pendent upon convenience in stirring the ore. On the other hand, should the ore be
high in sulphur, the sulphur dioxide formed must be speedily removed in order to
maintain the proper oxidizing atmosphere, and this necessitates the roof being suffi¬
ciently high to permit of large volumes of air passing over the hearth.
The process may be intermittent or continuous. Where small quantities of ore
are to be treated or the installation of an expensive furnace is not warranted, a small
hand-worked furnace would be used. Those are cheap in construction but expensive
to operate. The furnace is charged with the ore which is rabbled by hand until the
roast is complete when it is withdrawn from the furnace and a new charge introduced.
Should there be a considerable supply of ore, the process would be made continuous
by lengthening the hearth to a greater or less degree depending upon the time the ore
has to remain in the furnace. If the ore is to be hand stirred the width of the furnace
is controlled by the ability of the workman to reach and stir the ore with convenience,
a furnace worked on one side only being limited to a hearth width of about 8 ft. while
if worked from both sides the width may be 14 to 16 ft. For hand work the working
doors must be spaced sufficiently close to permit of easy rabbling, and if they are
spaced too far apart, the ore in the intermediate spaces will seldom be stirred.
Hand worked reverberatories are rapidly being replaced by the more effi¬
cient mechanically worked furnaces, due to the high cost of labor, which is the
chief item of expense, and also due to the fact that the quality of the labor is
variable and has a great effect on the quality of the roasted product. The
mechanically worked furnaces are under better control, the operation is more
regular, and the output per square foot of hearth area is greater. They have
the disadvantage that more dust is produced and difficulty is experienced with
ores that become sticky during the roast as this type of ore clogs up the rabble
arms and often causes serious difficulties. However, these disadvantages are
more than offset by the gains. Some types of mechanically operated reverbera¬
tory furnaces are:
Edwards Furnace.—This is a straight-line furnace 76 ft. long by 6 ft. 6 in. in width;
the hearth is sloped from the feed end to the discharge end in. to the foot. It is
heated by two fireboxes, one at the discharge end, the other being placed about 30 ft.
distant. The ore is fed at the cool end from which it is worked down the inclined
hearth to the discharge end where it is discharged through an opening in the hearth.
The stirring is done by 21 horizontally revolving arms fastened to a vertical water-
cooled spindle passing through the roof of the furnace and being driven by bevel
gears. Each spindle moves in the direction opposite to the adjacent one, the circle
of rotation of the end of the rabble arm overlapping that of the adjacent rabbles, so
that the particles of ore move through the furnace in a zig-zag path. In the Duplex
type of furnace the width of the hearth is 13 ft. and there are two rows of rabbles.
The following table taken from Hofman’s “ Metallurgy of Copper” shows the work
of the Edwards furnace at the Yampa plant when roasting copper ores:
Simplex Duplex
Hearth, feet. 6X 57 11 X 92
Inclination of hearth, inches per foot.... %
Spindles, revolutions per minute. 2 1
Power, horsepower. 1 4
Capacity in 24 hr., tons. 25 72
Sulphur in raw ore, per cent. 26-39
Sulphur in roasted ore, per cent. 2.2-8.6
S02 in flue gases, per cent. 4.3
818
Merton Furnace.—This is a multiple-hearth furnace having three superimposed
hearths. The ore is fed at the cool end of the upper hearth and moved by revolving
rabble arms in a manner similar to the Edwards furnace. Upon reaching the end of
the first hearth the ore falls through a drop hole to the second hearth and similarly
to the third hearth from which it passes to a finishing hearth having a water-cooled
spindle and rabble arm. The roasting hearth is 24 ft. long and 8 ft. wide and each
hearth has four rabble arms. The power required is 2 hp. and the capacity per 24
hr. is claimed to be as follows: Sulpho-tellurides 18 to 25 tons; pyrites 6 to 15 tons;
sulphide copper ores 10 tons; galena 8 to 20 tons.
The Ropp furnace is a straight-line single-hearth furnace with an area of 150 by
by 14 ft. and has four fireplaces. Through the center of the hearth there is a slot
through which passes a heavy steel plate to which is attached a rake extending the
width of the hearth and having teeth set at an angle of 45 deg. The steel plate sup¬
porting the rake is attached to a small truck which is moved along a track by a steel
cable. The rakes move slowly through the furnace from the feed end to the discharge
end, continually stirring the ore, and at the discharge end return to the feed end
alongside the furnace during which time they are thoroughly cooled by the air. Each
truck makes a complete circuit in about 5 min. The furnace requires about 5 hp.
and will roast 50 to 75 tons of sulphide ores per 24 hr.
The operation of the Wethey furnace is similar to the above except that the trucks
supporting the rake carriers move on tracks outside the furnace and return to the
feed end over the top instead of along the side. The Wethey multiple-hearth furnace
is a double furnace with four superimposed hearths having drop holes alternately at
opposite ends so that the ore may pass from one hearth to the next. The single¬
hearth type with 1,000 sq. ft. of hearth area will handle 70 tons per 24 hr. and requires
15 hp.
The Pearce turret furnace has an annular hearth open at one point to permit
ore being fed and discharged. It is made with one or two hearths, but as many as
six hearths have been tried. The furnace is heated from three exterior fireplaces and
the ore is stirred by two air-cooled rabble arms passing through a slot in the furnace
wall, the slot being kept closed by a steel band which moves with the rabble arms.
The mechanism for operating the rabble arms is located on the inside of the annu lar
ring. The teeth of the rabble arms are so arranged that the ore on the outer edge of
the hearth, which has the greater distance to travel maintains its relative position to
the ore on the inner edge. A single-hearth furnace having an area of 505 sq. ft. will
handle 14 tons of copper sulphide concentrates per 24 hr.; a double-hearth furnace
with an area of 1,010 sq. ft. will treat twice this amount.
The Bruckner Furnace.—At one time this was an extremely popular furnace but
due to its small capacity it is fast being discarded by the larger plants. It is composed
of a brick-lined horizontal cylinder with truncated ends, revolving on two friction
rollers between a firebox at one end and a flue at the other. The firebox runs on tracks
so that when roasting high sulphur ores one firebox can serve several furnaces. There
are four discharge openings in the horizontal portion of the cylinder. The furnace
requires 2 hp. and makes eight revolutions per hour. Working on Butte concentrates
it will rough roast 15 tons per 24 hr. The operation requires 2 hr. time in drying
the ore, 4 hr. in bringing it to the roasting temperature, and the remaining 18 hr. in
completing the roast and discharging the roasted product. Due to the furnace having
such small dimensions (16 ft. long by 8 ft. in diameter) fche fuel consumption is high,
and the stirring of the ore by the rotation of the cylinder causes a very heavy produc¬
tion of dust.
The cement kiln is at present being used to some extent for the roasting and sinter¬
ing of flue dust, fine concentrate, etc. The kiln may be fired with powdered coal or
oil, and the size of the sintered product depends upon the amount of fuel used per
Fig. 4.—Wedge furnace for lead ore and lead matte.
ton of material. A 60-ft. kiln will treat 50 to 75 tons of material per day depending
upon its nature.
Cylindrical Furnaces with Superimposed Hearths.—This type of furnace is
rapidly becoming universal for the roasting of fine material. It consists of vxitical
brick-lined iron or steel cylinders having, in some cases, as many as seven superim¬
posed hearths. When used for the roasting of material high in sulphur, sufficient
heat is generated to carry on the roast without the use of extraneous fuel. When
used for dead roasting, sulphatizing roasting, or chloridizing roasting a firebox is
used and can be so arranged that the hot gases enter the lowest hearth or any or all
of the upper hearths. There is a central rotating shaft which carries radial stirring
arms having blades set at the proper angle for stirring and moving the material. The
ore is fed to the top hearth where it is stirred and moved by the rabbles from the per¬
iphery to a drop hole in the center where it falls to the next hearth on which it is
moved from the center to the periphery, and so on until it is discharged from the lowest
hearth. The necessary air for the roasting is supplied through the working doom
and travels in a direction opposite to the movement of the ore. The leading types
of this furnace are the Herreshoff, Wedge, and Evans-Klepetko. The capacity of
this type of furnace depends largely upon the size and the character of the material
treated, and varies from 20 to 80 lb. per square foot per 24 hr.
Blast-roasting Apparatus.—Blast roasting is conducted in both up-draft and

down-draft apparatus; the latter appears to be gradually replacing the former.
Typical examples of the two methods are the Huntington-Hcberlein pots for
up-draft work, and the Dwight-Lloyd straight-line machine for down-draft
roasting.
The Huntmgton-Heberlein pots are made of east iron and are about 9 ft. in diam¬
eter with a depth of about 4K ft. and will hold 8 to 10 tons of ore. Owing to the
severe strains undergone during the operation the pots are usually cast in flanged
sections of four sides and a bottom which are bolted together. About 18 in. above
the bottom is a circular cast-iron grate 6 ft. in diameter which is perforated with
numerous *k£-in. holes and which serves to distribute the blast which enters the pot
beneath the grate through a blast pipe 5 to 6 in. in diameter. During the blowing
the pot is covered with a movable hood connected with a flue which carries off
the waste gases. The method of operating the pots is to spread a layer of hot coals
on the grate and start a light blast. About 2 tons of ore is charged on the hot coals
and as the roast gains headway, the balance of the charge is added and the blast
pressure gradually increased to its maximum of about 12 oz. The blow is complete
in from 8 to 10 hr. after which the blast is turned off and the pot emptied by means
of a crane. The cake is then broken up and sent to the blast furnaces.
The Dwight-Lloyd straight-line sintering machine consists of a structural steel
frame which supports a feed hopper, an ignition furnace, a suction box, and a pair of
endless tracks or guides upon which move a train of grates on wheels, known as pallets,
forming a continuous conveyor. At the discharge end of the machine the guides
curve downward and pass under the machine back to the feed end. At this point
there is a break in the continuity of the train of pallets in order to facilitate the dis¬
charge of the sintered material. The suction box is connected with an exhaust fan
maintaining a vacuum of 5 to 10 in. of water and the top is accurately machined so
that when the pallets pass over they can slide over this smooth surface making a con¬
tact that is as airtight as possible. The pallets are moved by sprocket wheels which
lift them from the guides underneath the machine to the upper guides and push them
under the feed hopper. As each pallet is pushed by the one behind it, the joints are
kept closed and the air leakage reduced to a minimum. The speed of travel varies
from 18 to 36 in. per minute. The ignition furnace, which may be fired with coal,
gas, or oil, is placed across the line of pallets and is so arranged that the flame is
deflected downwards upon the layer of ore carried by the pallets. As the roasting
occurs only when the ore is over the suction box, the area of the latter is the effective
hearth area of the machine.
The operation of the machine is as follows: The ore and flux are conveyed to a
receiving hopper having a gate to regulate the discharge to an endless belt which
conveys the material to a horizontal mixer where the ingredients are thoroughly
mixed and any necessary amount of water added, the mixer discharging to the feed
hopper of the machine. As the pallets are pushed by the sprocket wheels to their
position under the feed hopper, they receive an even layer of ore about 4 in. in thick¬
ness and then pass under the ignition furnace where the top surface of the ore is
ignited. At the same time the pallet engaged the suction box and the downward
flow of air causes the ignition and sintering to progress from the surface toward the
grates. The speed of the pallets is so adjusted that as a pallet leaves the suction box
the sinter is complete. The pallets now move towards the discharge end and, if too
hot, are sprayed with water to reduce the temperature. On reaching the discharge
point, where there is a break in the continuity of the train of pallets, the pallet drops
into the guides leading beneath the machine and at the same time strikes the pallet
which has immediately preceded it, with sufficient force to dislodge its load of sinter
which is discharged into a chute leading to a hopper or car. The pallets now return,
by gravity, beneath the machine to the sprocket wheels. The machine has a capacity
of from 50 to 200 tons per 24 hr. depending upon the nature of the material being
treated.
A continuously rotating wheel on which the charge was spread was formerly used
in the same way, the charge being ignited by a fire box occupying part of the per¬
iphery (if the axis was horizontal) or a sector of the circle (if it was a horizontal cir¬
cular surface revolving about a vertical axis), the suction box coming next, then a
cooling portion, and finally the material being scraped off.
THE SMELTING OF COPPER ORES

By F. R. Pyne1
The World’s Copper Supply.—The great bulk of the world’s production of

metallic copper is derived from sulphide ores. Oxide ores, which occur to a
small extent, are usually mixed with sulphides for smelting, or are subjected
to hydrometallurgical treatment. Native copper ores, such as occur in the
Michigan peninsula, consisting as they do of particles of metallic copper dis¬
seminated throughout the accompanying gangue, require simply a fusion (usually
preceded by concentration) with sufficient flux properly to slag the gangue, but
such ores are entirely insufficient to supply the world’s consumption.
Fig. 5.—Standard matting furnace—vertical section.

Fig. 6.—Standard matting furnace—vertical section.
The sulphide ores of copper may be considered as mixtures of copper and iron sul¬
phides accompanied by siliceous or basic gangue and the object of smelting is to
cause by fusion the conversion of the gangue into a valueless slag by the addition of
proper fluxes, and at the same time to concentrate the copper and other valuable
constituents of the ore, such as silver and gold, into a small amount of high-grade
material, this high-grade material then being submitted to oxidation to eliminate all
but the valuable portions which it is desired to recover. In order to accomplish this
1 Works superintendent, U. S. Metals Refining Co., Chrome, N. J.
METALLURGY OF COPPER 823
result advantage is taken of the strong affinity of copper for sulphur and its weak
affinity for oxygen in comparison with iron and the other base metals in the ore.
Slag and Matte Formation.—There are two important materials formed dur¬
ing the smelting of copper-sulphide ores; the slag produced by the combinations
of the gangue of the ore and the flux, and the matte which is the product of the
fusion of the metallic sulphides: of these two products, the slag receives the first
consideration of the metallurgist, as the sulphides melt readily under almost any
circumstances while unless the slag-forming constituents are properly propor¬
tioned, serious difficulties will result. A satisfactory slag must possess the follow¬
ing qualifications:
1. It must be as economical as possible and to this end the other qualifications
are subordinated, as a scientifically perfect slag may in the end be so expensive that
any margin of profit is wiped out by its cost.
2. It must be sufficiently liquid to flow freely, but should not require such a
temperature that an excessive amount of coke is necessary to produce it.
3. The specific gravity should be sufficiently below that of the matte to permit
the latter to separate from it thoroughly.
The following slags are illustrative of general copper practice:
Per cent
1 2 3 4 5 6
Si02. 32.5 34.0 44.5 39.3 42.2 42.0

FeO. 52.2 46.0 39.7 24.8 39.4 22.0
CaO. 4.8 11.0 9.9 25.2 7.0 26.0
AI2O3. 7.2 5.0 6.2 6.2 4.0 4.0
Cu. 0.39 0.69 0.21 0.3 0.29 0.23
Matte is the product of fusion of metallic sulphides. Upon fusion of a mix¬

ture of copper-iron sulphides, a certain part of the sulphur is volatilized while of
the remainder sufficient is seized by the copper to form cuprous sulphide, Cu2S,
the balance being taken up by iron and other base metals in the order of their
affinities to form sulphides.
For all practical purposes, copper matte may be considered to be a mixture
of Cu2S.and FeS in varying proportions. Matte is also an excellent collector
or gold, silver, and other precious metals.
The amount of matte produced in smelting is dependent upon the sulphur in the
ore. An ore high in sulphur and iron but low in copper will, upon fusion, produce a
matte containing so low a percentage of copper that its treatment would, under
ordinary circumstances, cost too much, and as it is the desire of the metallurgist to
concentrate the copper in his ores up to the most economical point, steps mustbe taken
to reduce the amount of sulphur available for the formation of matte and 1 ower the
amount of the latter produced per ton of ore smelted.
Such reduction in the sulphur content may be accomplished in two ways; by
subjecting the ore to an oxidizing roasting or by subjecting it to an oxidizing fusion.
The latter is used where the ore occurs as massive sulphides and will be considered
later under “Pyritic Smelting.” If the ores are not massive sulphides, oxidizing
roasting is the predominating method. In roasting the object is to oxidize the excess
sulphur to SO2, at fche same time oxidizing the excess iron which will then be available
for use as flux for the silica in the ore and thereby save an equivalent amount of
expensive foreign flux. An ore low in copper will require a greater amount of roasting
than one rich in copper in order to produce a matte with the desired copper content.
The following table is illustrative of mattes1 formed in copper smelting:
Per cent
1 2 3 | 4 5 6
Cu. 49.2 55.0 54.9 60.7 48.2 44.6

s. 19.5 24.0 23.4 23.3 23.7 23.5
Fe.j 22.8 13.9 20.3 11.4 24.7 27.0
There are two kinds of apparatus in which the smelting is conducted; the blast
furnace and the reverberatory furnace. The blast furnace is used for the treatment
of coarse ores and the ore and fuel are in intimate contact, the fuel being burned by
a blast of air forced through the body of the charge. The reverberatory furnace is
used for the treatment of fine ores and the fuel and ore are kept separate, the former
being burned in a separate compartment from which the flame and hot gases pass
over the ore being guided by a more or less horizontal roof. The heating of the ore
is accomplished by the radiation from the roof and walls rather than by contact with
the hot gases.
The modem copper blast furnace consists of a water jacketed shaft having an
oblong cross-section and varying in height from 15 to 20 ft. The water jackets
throughout the shaft are necessitated by the furnace running with a hot top due
to the large amount of air forced through the charge and also by the low charge
column, both of which are an aid to the obtaining of a pyritic effect so much
desired by copper metallurgists. The oblong cross-section of the shaft is due to
the fact that the width at the tuyere level is limited by the ability of the blast
to penetrate the charge. This dimension varies from 44 to 56 in. and this being
fixed, the only manner in which an increased capacity is obtained is by lengthen¬
ing out the furnace and modern furnaces have lengths varying from 180 in. to as
much as 1,044 in. in the case of the huge furnaces at Anaconda. The ends of the
furnace are generally vertical while the sides are boshed or sloping. The amount
of bosh for the sides is governed by the amount of reducing action to which it is
desired to subject the charge; the greater the bosh, the stronger will be the reduc¬
ing action. In furnaces where there are two tiers of water jackets, the usual
practice is to have only the lower tier of jackets boshed, the upper tier being
vertical. The amount of bosh per foot of height varies from 1 to 2 in.
The blast enters the furnace through tuyeres passing through the water jackets.
The pressure varies with the width of the furnace, the number of tuyeres, the amount
of fines in the charge and the height of the ore column, a fair average figure being
about 36 oz.
The lowest point of the furnace consists of a shallow hearth or crucible in which
the molten slag and matte are collected. The crucible may be built up solid from the
1 In general mattes carry about 23 per cent of sulphur, a fact, I believe, first pointed out by Dr.
Edward Keller, and in any calculations concerning matte-smelting this figure is a safe one to use in
calculating the charge, correcting the charge later by actual furnace results.
foundations but the general practice is to build it upon ribbed cast-iron plates sup¬
ported by jack screws. The refractory lining of the crucible is generally made of a
good grade of fireclay brick though special circumstances may require the use of
silica or chrome brick. The outside of the crucible is encased in heavy cast-iron
plates bolted securely together, these plates also serving as a support for fche lower
tier of water jackets."
The upper part of the furnace consists of a brick or metal flue through which the
waste gases are led off into a dust chamber. Feed openings located just above the
upper tier of water jackets and occupying the entire length of the furnace, are used,
the charges of fuel and ore being shoveled through by hand or else dumped from a
car. Between charges the feed openings are kept closed by sliding doors.
The slag and matte are continuously discharged from the crucible through a
water cooled breast jacket into a spout, which conveys the mixture to a fore¬
hearth or settler.1 The overflow end of the spout is higher than the opening in
the breast jacket so that when the spout is filled with molten material the blast
is trapped. The spout is usually constructed of water-cooled iron, copper, or
bronze, though the presence of a highly corrosive matte may necessitate the use
of chrome or magnesite brick.
The settler serves not only as a means of separating the matte and slag by
reason of the differences in their specific gravities but also as a reservoir for the
accumulation of matte until such time as it is required. A plant which treats its
own matte, therefore, requires a larger settler than does one which casts its
matte direct into molds for shipment to some other plant for treatment. The
settler is circular or oval and consists of an iron shell lined with refractory
material. There are usually two matte taps, located in the lower part of the
settler, which'are kept closed between taps by clay plugs. The slag overflows
from the top of the settler into a pot car or else into a launder where it is granu¬
lated by water.
The following table gives data concerning some modern copper blastfurnaces.
Tenn.
C. &C. Granby Anaconda Garfield
Co.
Section at throat.... 72 by 270 64H by 72 by 612 72 by 1,044 66 by 240

266K
Section at tuyeres... 56 by 270 44 by 266K 56 by 612 56 by 1,044 42 by 240
Height, tuyeres to
throat. 18 ft. 3 in. 12 ft. 10 in. 13 ft. 10 in. 13 ft. 4 in. 13 ft.
Number of tuyeres... 25 30 88 150 44
Diameter of tuyeres.. 4in. 5 in. 4 in. 4 in. 4 in.
Tons per 24 hr. 441 515 1,460 2,500 560 to 750
Tons per square foot
hearth area per 24
hr. 4.30 6.32 5.88 6.17 8.00 to 10.71
Blast pressure, ounces 45 26.30 40 40 38 to 41
1 There are a few small old-fashioned furnaces in which the slag and matte are tapped intermittently.
These are known as ‘‘internal settlers." The slag tap is on a higher level than the matte tap. With
these furnaces the slag may be run continuously or almost so, and the matteônly tapped intermittently.
Sub-divisions of Blast-furnace Smelting.—In blast furnace smelting there

are three distinct processes:
1. The reduction process in which a considerable percentage, 12 to 15, of
coke is used. The blast oxidizes the carbon of the coke but not the sulphur of the
ore. This process is the one most generally used.
2. The pyritic process in which raw massive sulphides are smelted in a highly
oxidizing atmosphere without the addition of carbonaceous fuel. The heat
generated by the oxidation of the sulphur and iron is sufficient to maintain a con¬
tinuous operation.
3. The partial (or semi-) pyritic process in which sufficient heat is not gen¬
erated by the oxidization of the sulphide ores and carbonaceous fuel is added to
the charge in barely sufficient quantity to overcome the deficiency.
The Reduction Process.—This process is characterized by the use of car¬
bonaceous fuel as the principal source of heat. The .sulphur content of the
charge is sufficient to form the desired matte and any oxidation is generally
undesirable. The percentage of coke, which is the chief carbonaceous fuel,
varies from 12 to 15 per cent. An excessive amount of coke is likely to cause a
reduction of iron from the slag, while with a shortage the furnace becomes chilled
and suffers a loss of tonnage.
The use of such an amount of coke means that the furnace always contains a
large mass of glowing coke for a considerable distance above the tuyeres and conse¬
quently the oxygen of the blast is consumed almost immediately resulting in a strong
reducing atmosphere. Such being the case, it is readily seen that any sulphides
present will melt down unchanged and form matte. An ore high in sulphur would,
therefore, yield a large amount of matte which is expensive to treat and it is clear
that little would be gained by the reducing smelting of such an ore and that a portion
of the sulphur must be eliminated by roasting previous to the treatment in the blast
furnace, or the sulphides may be mixed with oxide copper ores to increase the grade
of the matte. A certain oxidation of sulphur may be accomplished in the blast
furnace by keeping the ore column low and increasing the blast but such procedure
results in a very hot top and high temperatures of the waste gases and in general is
uneconomical.
Regarding the chemistry of the process, the blast entering through the tuyeres
oxidizes the carbon of the coke to CO which is partly oxidized to CO2, the resulting
gases being a mixture of CO, C02, and the nitrogen of the air.1 As these gases rise in
the furnace and reduce the oxidized materials, the percentage of C02 increases and will
predominate in the waste gases. Any unchanged CO is usually burned to C02 at
the top of the ore columns. As the ore charge descends in the furnace, its moisture
is driven off almost immediately, followed by the decomposition of such carbonates
as may be present. At a lower point the oxides and silicates of copper will be reduced
and on coming into contact with metallic sulphides will be converted to sulphides
which, together with any existing copper sulphides, melt and flow downwards to the
crucible collecting the silver and gold in the ore as they descend. Ferric oxide, Fe2Oa,
is reduced to the ferrous condition, FeO, and when the temperature becomes suffi¬
ciently high, which occurs a short distance above the tuyeres, unites with silica to
form slag, as does also any lime or other flux added to the charge.
The molten materials having collected in the crucible below the tuyeres the sul¬
phides adjust themselves to form the proper matte, and likewise the various slag
components to form the proper slag.
1 But in only one or two plants in the world does the proportion of CO become high enough to
use the gas as a fuel, as may be done with the gases from iron smelting.—Editor.
The Pyritic Process.—The feature of this type of smelting is that the heat
necessary to conduct the operation is furnished by the oxidization of the constitu¬
ents of the ore, no extraneous fuel being used for this purpose. Up to 2 or 3
per cent of coke may be added but its function is probably to preheat the charge
and it never reaches the actual smelting zone.1 The essential requirements for
pyritic smelting are siliceous material high in free silica, and heavy pyrite ore.
It is necessary for as much of the silica as possible to be in the free state in order
that it may combine with the ferrous oxide at the instant of its formation,
combined silica is unsuitable for the pyritic process as it is already united with
one or more bases and consequently requires a large amount of heat to break up
the existing combinations and form new ones with the ferrous oxide. The pyrite
ore not only furnishes the heat for the operation but also the matte required and
it must largely consist of iron pyrites, as any oxidation of the copper sulphide
would result in the slagging and loss of the resulting copper oxide.
The blast entering the furnace through the tuyeres strikes the hot fused sulphides
and oxidizes the FeS to FeO and SO2, the FeO simultaneously uniting with silica to
form a ferrous-silicate slag while the SO2 rises through the charge and preheats it,
assisted by the oxidation of such coke as may be present.
C + S02 * S + CO2
As the charge sinks in the furnace the moisture is driven off quickly followed by the
dissociation of the limestone added as flux. When the smelting zone is reached the
sulphides fuse and run down over the gangue and flux and are oxidized by the blast
as described above. The matte collects below the tuyeres as does also the ferrous
silicate slag after uniting with the lime to form the final slag. The slag and matte
mixture then flows through the spout to the settler.
The grade of matte produced and the degree of acidity of the slag are controlled
by the volume of air blown into the furnace. Any reduction in the air will cause
less of the sulphide to be oxidized and the grade of the matte will be lowered, and as
less iron is oxidized the slag will become excessively siliceous and interfere with the
process unless the amount of siliceous material has been correspondingly reduced.
On the other hand, an increase in the amount of air oxidizes additional iron, raising
the grade of the matte and increasing the iron in the slag unless additional silica is
provided.
A peculiarity of the pyritic process is the artificial bosh formed on the side and
end walls of the furnace. This contracts the smelting area to a long narrow slit.
' The bosh composed of fragments of quartz or gangue rock stuck together by slag or
by superficial softening, seldom contains any matte. Its position in the furnace is
not fixed but varies under different conditions.
In reducing smelting the tuyeres are bright, in pyritic smelting they are dark and
are bridged across. It has been stated that a bar can be passed through the furnace
from one tuyere to the opposite tuyere and be cool when withdrawn. This indicates
that the zone of fusion is well above the tuyere line and that the molten material
passes through channels between the tuyeres.
Partial pyritic smelting is used when the available sulphide ores are not mas¬
sive but consist of pyritic material disseminated throughout the gangue which
is generally low in free silica and usually contains alumina. Such ores are not
capable of furnishing by themselves sufficient heat for the continuance of the proc-
t Jt probably also aids by securing a certain looseness of charge.—Editor.

2 V' 1 -v
■ 12 -31J£ Lb. Is
> 3^1
-Reverberatory for smelting copper

“TC.I. Tap Hole
L--J
1* ?
M*—f—15~ 2j . 1
Bi 11
mwg* -tiiil lift
ess and it is, therefore, necessary to add carbonaceous fuel up to about 8 per
cent. Preheating the blast has been found advantageous in some instances.
In the partial pyritic process the slags are high in silica and lime but low in iron
and frequently contain considerable alumina. Due to the presence of carbon
a large volume of air is necessary in order that the atmosphere shall be sufficiently
oxidizing to obtain any considerable pyritic effect. This results in the furnace
having a very hot top. The smelting zone is nearer the tuyere level and is not
so contracted as in pyritic smelting, but higher and more contracted than in
reducing smelting. The tuyeres are frequently dark and require a great deal of
punching to keep them open.
In reverberatory smelting the ore supply is generally fine sulphide ores which
have been given a preliminary rough roasting. This is smelted on a hearth with
the addition of proper fluxes, to a matte containing 35 to 45 per cent copper, and a
slag containing 35 to 40 per cent of silica. The reducing agent is the sulphur
remaining in the ore, the fuel used not being in contact with the ore serves only
to obtain the desired temperature.
The hearth of the reverberatory furnace is generally made of silica sand sintered
into place though silica and magnesite brick have been used. It is long and narrow,
the width being limited to about 18 ft. by the difficulty and cost of maintaining the
roof. The capacity therefore depends upon the length and this is limited only by the
ability to maintain the proper temperature at the flue end. In the most recent fur¬
naces the length is 135 ft. though present practice is to limit it to about 112 ft. The
former practice of having a vault under the hearth has been abandoned in favor of a
solid bottom in order to economize on heat, as any radiation through the hearth
seriously curtails the rate of smelting. The hearth is enclosed in silica- or clay-
brick walls which are about 24 in. thick to reduce radiation to a minimum. They are
heavily ironed and buckstayed for resisting the thrust of the hearth and its contents.
The roof is invariably constructed of silica brick varying in bhickness from 15 to 20 in.
Modern reverberatories have no side doors, the charges being introduced through
holes in the roof along the side walls, the practice being small charges at frequent
intervals. The charge consequently lies against the side walls, protecting them from
the hot gases, and as it gradually melts more charge fis introduced. A deep bath of
matte is carried on the hearth and serves to equalize and distribute the heat, thus
preventing the overheating of any part of the hearth with possible damage to the
furnace. It also is of great assistance in rapidly heating up the fresh charges as they
are dropped into the furnace as the matte being highly superheated readily gives up
its heat to the cold charge above it, while the hot furnace gases heat the charge from
above. The matte is tapped as desired from the end of the furnace while the slag
may flow off continuously or be skimmed periodically; the former method is coming
more into favor as it gives a cleaner slag.
The fuel used may be soft coal fired by hand or in stokers, pulverized coal, oil,
or producer gas. The largest plants are rapidly changing from hand firing to pul¬
verized coal or oil on account of the more economical handling and the greater thermal
efficiency obtained. The ratio of charge to fuel will average about as follows:
Soft coal. 4:1

Pulverized coal. 7:1
Oil. 0.6 to 0.7 bbl. per ton charge.
Flue-dust is an important though undesirable by-product from the roaster,

blast furnaces and reverberatories and the recovery and working up of this
material is a considerable item in the cost of operation. The recovery may be
830
Fig, 10.—Basic-lined converter. At left, side sectional elevation. At right, individual tuyere pipes.
accomplished by one or more of the following methods; passing the gases through
large dust chambers in which the velocity is so reduced that all but the very
finest particles will settle out; by filtering the gases through woolen bags after
previously cooling to a proper temperature; by electrostatic precipitation as in
the Cottrell process; or by causing the suspended particles to impinge upon wires
or baffles hung in a dust chamber as in the Roesing system. The treatment of
recovered flue-dust consists of briquetting and smelting in the blast furnace, a
method formerly widely practiced but now rapidly falling into disuse; agglomerat¬
ing in rotary kilns, Dwight-Lloyd sintering machines or Huntington-Heberlein
pots; incorporating with converter slag and smelting in the blast furnace; or by
direct smelting in the reverberatory furnaces, the last method being the most
satisfactory.
Metal loss is a very important element in smelting costs and one -which is
watched and studied carefully in order to reduce it to a minimum. The three
chief sources are dust loss in handling, losses in slags, and losses in flue dust and
fume. The loss in handling increases with the fineness of the ore and may
amount to a considerable amount; the loss in slags is the largest source owing to
the relatively enormous amount of slag produced though the assay may be low;
and the dust and fume loss will depend upon the fineness of the ore, the amount of
volatile constituents, and the facilities for recovering the flue-dust.
Comparing blast furnace and reverberatory smelting, the former requires coarse
ore in order to work satisfactorily, little space is required and the investment is low
for any given tonnage, the fuel though small in amount is relatively expensive, and
considerable power and cooling water are required. Blast-furnace slags can be made
between widely varying limits but the addition of a large amount of flux causes a
large amount of slag with a corresponding metal loss. The reverberatory furnace
is the most satisfactory and advantageous apparatus in which to treat fine ores but
as it requires an extensive roasting plant and in itself occupies a large amount of space
and locks up a large amount of valuable metal, the investment is large for any given
tonnage. While large amounts of low-priced fuel are consumed, a large amount of
the heat in the waste gases is recoverable in waste heat boilers which greatly reduce
the power costs. The amount of slag and consequently the metal loss is less than the
blast furnace. With the ever increasing amount of fine ores, the reverberatory is
fast displacing the blast furnace and it has been found advantageous in certain locali¬
ties to use reverberatory smelting for coarse material that under ordinary conditions
would go to the blast furnace.
Converting.—The copper and other valuable ingredients of the ore having
been concentrated into matte, and the valueless material disposed of as slag,
the next step in the process is the removal by oxidation of the iron and the sul¬
phur. This is accomplished by transferring the molten matte to a refractory
lined vessel, known as a converter, and forcing thin streams of air through
the liquid mass. The iron in the matte is oxidized to FeO and immediately
combines with silica to form slag while the sulphur is oxidized to S02 and passes
off in the waste gases. The copper together with any gold or silver is reduced
to the metallic form and is cast into suitable shapes for transportation to the
refinery.
The converter is a cylindrical iron or steel vessel lined with refractory material.
A conical mouth is provided through which the matte is introduced, the slag and
blister copper withdrawn, and the gases pass to the stack. The air for the oxidation
is introduced through the side of the converter by means of tuyeres leading from an
air box attached to the outside and having a connection to the high pressure air main.
Converters are classified as upright or horizontal, depending upon the direction
of the long axis of the cylinder and are so mounted that they can swing in a vertical
plane for the purpose of receiving or withdrawing the molten contents.
The refractory lining is composed of either acid or basic material, the acid
lining though formerly universal is rapidly being supplanted by the basic lining.
In the acid-lined converter the material used should contain the largest possible
amount of free silica consistent with proper mechanical strength. The material
generally used is low-grade siliceous ore, for while this may not be so desirable
metallurgically as quartz, the ore is practically smelted for nothing and the values
recovered from an otherwise profitless material make its substitution more
economical. The lining is rapidly destroyed by the union of the FeO, formed
during the blowing, with the silica of the lining and the life of an acid lining is
limited to about five charges, depending upon the grade of matte converted, a
low-grade matte being more destructive than one of higher grade due to the
larger amount of iron slagged. The great expense of frequently renewing the
acid fining has led to its replacement by a fining of magnesite brick which is not
attacked chemically by the process. When using such a fining the silica neces¬
sary to flux the iron in the matte is added to the charge before starting the
blow. The basic fining is rapidly displacing the acid fining unless special con¬
siderations make it more profitable to retain the latter. The main advantage of
the basic fining are: the decreased cost of lining, a basic fining producing 6,000
tons of copper as against 25 to 40 tons for an acid fining, longer converters can
be used resulting in lower operating costs, and mattes of lower grade may be
converted.
The operation of converting is conducted by first turning down the converter so
that the mouth is in proper position to receive the molten matte from a ladle. The
charge having been poured in, the blast at about 16-lb. pressure is turned on and the
converter turned up until the tuyeres are below the surface of the matte and the
mouth of the converter is under the hood conveying the waste gases to the stack.
The oxidation commences and is divided into two stages, the first or slagging period,
and the second or blister period. The reactions occuring during the slagging period are
FeS + 30 - FeO + S02

Cu2S -b 30 = Cu20 ~b S02
Cu20 + FeS - Cu2S + FeO
2FeO -1- Si02 = 2Fe0-Si02
With the acid fining the FeO formed unites directly with the silica of the fining and
on this account, as previously stated, it is necessary to have the matte as high grade
as economically possible. A 30 per cent copper matte contains approximately 1.7
times as much iron as a 50 per cent matte and the lining will be consumed that much
more rapidly. As an offset to the economy of converting high grade matte, there is
the question of increased slag loss in the blast furnace or reverberatory when making
high-grade matte and the question of the proper grade of matte to be made and con¬
verted depends upon local conditions but will average about 45 per cent copper. This
condition does not exist to the same extent when using the basic fining as the siliceous
flux is added to the converter before the blast is turned on.
As the oxidation proceeds and the iron is slagged the matte remaining in the con¬
verter gradually approaches white metal (CU2S). This change may be followed by
the character of the flame issuing from the mouth of the converter. At the start
of the operation it has a reddish color and, as the white metal stage approaches,
changes to a pale blue. However, if an impure matte is being converted, the flame
indications are unreliable and dependence is placed upon the appearance of the
layer of matte on the punching rod which is inserted in the tuyeres to keep them
open for the passage of the blast.
The iron having been slagged, the converter is turned down, the blast shut off and
the slag poured or skimmed into a ladle. A fresh charge of matte, or matte and flux,
is introduced and the operation repeated until sufficient white metal has accumulated
in the converter for finishing. The converter slag contains several per cent of copper
and is too valuable to waste and is, therefore, returned to the blast furnace or rever¬
beratory for retreatment.
The second or blister period consists in oxidizing the sulphur of the white metal,
the copper together with silver and gold remaining in the metallic state. The
reactions during this period are
Cu2S + 30 = Cu20 + S02
2Cu20 + Cu2S = 6Cu + SO2
The blast oxidizes the white metal to cuprous oxide which reacting with the sulphide
liberates copper and forms S02. During this period the flame of the converter changes
from a pale-blue to a reddish-brown but the appearance of the metal on the punching
rod is a surer indicator and is invariably used. Upon the completion of the blow the
converter is turned down, the blast shut off and the copper poured into a ladle for
casting into molds.1
Copper produced in the converter is known as blister copper due to the appearance
of the upper surface of the cast pig. It contains approximately 99 per cent copper,
the balance consisting of sulphur and small amounts of impurities together with the
silver and gold originally in the matte. It is too impure for commercial use and must
be submitted to a refining process before it can be marketed.
THE SMELTING OF IRON
By Bbadley Stoughton2
General Description.—Iron is found in the earth in the form of ore, or oxide,

of which the commonest varieties are Fe203 and Fe304. The former is occa¬
sionally combined with more or less water. The smelting of these ores consists
in heating them to a high temperature, first in contact with reducing gases
(chiefly CO—which removes the greater part of the oxygen) and, finally, in con¬
tact with white-hot carbon, which completes the reduction. The smelting takes
place in blast furnaces the form of which is shown in Figs. 11 and 12. The blast
furnace itself is shown in Fig. 12, in section, and on the left of the picture in Fig. 11;
while the four other furnaces in Fig. 11 are the so-called hot-blast stoves, which are
used for the purpose of preheating the blast that is driven into the furnace for the
1 The first copper formed carries down the great bulk of the gold and silver, and with matte carrying
very little of these metals, so that when the bullion does not carry enough precious metals to make refin¬
ing by electrolysis commercially advisable, selective converting is sometimes resorted to. The old Davids
aelecteur converter had a pocket to catch the first copper formed. The reaction is one worth remember¬
ing.—Editor.
53
purpose of burning the coke and thereby producing the temperature and hot
gases required for the chemical reactions and the melting.
In Fig. 13 we have an outline section of a furnace showing the temperatures and the
typical chemical reactions predominant at each zone. A column of coke fills the body
of the furnace from the bottom of the hearth to the widest portion, which is known as
Fig. 11.—Iron blast furnace and stoves.
the top of the bosh. Above this “bed” are alternate layers of iron ore and coke,
together with an appropriate flux, which is generally limestone, and which is intro¬
duced for the purpose of uniting with the impurities of the coke and the gangue of the
iron ore, so as to produce a slag which will be fusible at the temperature of the smelting
zone.
The preheated air, at a temperature of usually 800 to 1,200°F. (425 to 650°C.) and
at a pressure of 15 to 30 lb. per square inch, enters through the tuyere pipes at the top
of the hearth, combines with the fuel and creates a volume of intensely hot reducing
METALLURGY OF IRON 835
gases, which pass up through the interstices of the charge, heating, melting, and reduc¬
ing the ore which it meets, and finally passing out the throat of the furnace.
The solid materials, consisting of ore, coke and limestone, are charged by mechan¬
ical means at the top of the furnace, and descend the 80 ft. to the hearth in approxi-
:e
Fig. 12.—Cross-section of blast furnace and skip hoist.
mately 20 hr. The moderate heat met near the throat first removes any moisture
present in the raw materials; then the CO gas begins to rob the ore of oxygen, until,
about midway of the furnace height, the iron is all in condition of a metallic sponge,
the interstices of which are largely filled with carbon in the form of a sort of soot,
produced according to the reactious given in Fig. 3. The heat separates from
the limestone its content of CO2, which passes out through the off-take at the
throat, with the other gases. When the descending solids reach the top of the
smelting zone, which, is coincident with the greatest diameter of the stack, the tempera¬
ture is sufficient to reduce them all to the liquid state, with the exception of the
coke, which is infusible and which extends as a “bed” to the bottom of the furnace.
This bed, plus the friction of the up-rushing gases, plus the friction against the
furnace walls, is the support for the weight of the charge in the furnace.
Stock 75-rO 450° F

.Line (1) 2Fe20s4- 8CO = 7C02+ 4Fe4- C (begins)
(2) 2Fe203+CO = 2FeO+CO2+F62O3 (begins)
65+-10 770° F 575° (3) Fe 4- C02«Fe0+C0 (begins)

750° (6) Fe203-h 3C=2Fe+ 3CO
65+20 1090°F 1025° (4) C 4- C02=2C0 (rapid)

1100° Deposition of carbon ceases
1300° (7) FeO 4- C-Fe+CO (begins)

45+30 1410° F
id7no (D FeO 4- C=Fe+CO (complete)
1475 (8) CaCo3 = Ca04- CO2
35+40 1780°F
1830° (4) C 4* CO 2 » 2CO (prevails)

CO2 cannot exist below this level
29+50 2050*F
■60 2870° F
Smelt- (9) Si02+2C«Si4-2C0

ing (10) FeS-hCaO+C=CaS4-Fe+CO
■70 Feet
2750*F Zone (11) Mn02+2C = Mn+2CO
(12) P2Ofi4-5C=2P + 5CO
Tuyeres =1
Fig. 13.—Diagram showing chemical action m blast furnace.
The Liquid Furnace Products.—The iron sponge melts above the top limit
of the coke bed and trickles down, absorbing carbon, silicon, sulphur, phos¬
phorus and manganese as described in the next paragraph. It collects in the
bottom of the hearth and is drawn off at intervals by drilling a hole through the
tapping notch. It is then either used in the fluid state at nearby steel works,
or else is allowed to solidify in the form of “pigs.” The slag also begins to form
at the top of the smelting zone, by the union of lime with gangue of the ore.
This too trickles down and completes its composition by dissolving the ash of the
coke. This slag, or “cinder,” floats on top of the iron and is drawn off through
the “cinder notch,” or “monkey,” as often as is necessary to prevent its flooding
the tuyeres. When very basic it may he used for the manufacture of portland
cement.
Controlling the Grade of the Product.—The reactions in the smelting zone
determine the chemical composition of the iron product, which is called “pig
iron,” because of the form in which it is cast at old-fashioned types of furnace.

The exact smelting zone reactions depend in turn upon the temperature of the
smelting zone, and upon the constitution of the slag. Thus, the higher the tem¬
perature, the greater amount of carbon will dissolve in the metal; also the greater
will be the amount of silicon reduced by carbon and the greater the amount of
FeS1 transformed into CaS. These reactions are important, because all the
silicon reduced will go into the iron, and so will all the FeS, whereas CaS will
dissolve in the slag. Therefore, a high temperature will mean—other things
being equal—more carbon and more silicon in the iron, and less sulphur. On the
other hand, a low temperature will tend to produce iron with low carbon, low
silicon and high sulphur. The constitution of the slag is regulated by increasing
or decreasing the percentage of lime put into it. Thus, a basic slag, high in
lime, will tend to retain silica more tenaciously and resist its reduction to silicon.
This action is important without being as powerful as the increased reduction of
silicon with increased temperature of hearth. A basic slag will also facilitate
the formation of CaS, and will yield more readily any manganese present in the
charge, thus tending to decrease the sulphur and increase the manganese in the
pig iron. These slag influences become most important in connection with the
production of special grades of iron containing large proportions of silicon (10
per cent) and of manganese (up to 80 per cent). In general it is the temperature
of the smelting zone which is the chief controlling influence, but it is also to be
noted that a basic slag will increase the temperature of the smelting zone, because
it requires a higher heat to bring it to that fluid condition necessary for the opera¬
tion of the furnace.
Some analyses of typical slags, within wide limits of operation, are given below:
Table 1.—Typical Iron Blast-furnace Slag Limits

Per Cent
Si<V. 20 to 70
CaO & MgO: 25 to 55
A12034. 5 to 20
CaS. Up to 4J£
FeO. Up to 1
The constitution of typical grades of pig iron will vary somewhat as follows:
Table 2.—Limits in Analysis of Ordinary Grades of Pig Iron
Per Cent
Carbon.... 2.75 to 4.50

Silicon. 0.75 to 3.00
Sulphur 0.30 to 0.02
Phosphorus 0.03 to 3.00
Manganese. 0.20 to 1.00
Practically all the phosphorus that goes into the furnace with the iron ore (and,
in less extent, with other raw materials) will be reduced and absorbed by the iron;
1 The FeS gets into the furnace chiefly with the coke. Rarely iron ores contain it.
2 The silica comes from the ash of the fuel and the gangue of the ore, as well as that of the limestone.
3 Lime and magnesia are chemically equivalent approximately, except that magnesia is about 40
per cent more powerful weight for weight.
4 Alumina comes from the gangue of the ore and from the coke ash.
the same is true with manganese, except that an acid slag will carry away manganese
oxide to partially satisfy its affinity for bases.
Controlling the Temperature of the Smelting Zone.—The all-important

matter of the temperature of the smelting zone is determined by: the proportion
of fuel used with the charge, known as the “burden” which the coke has to carry;
the temperature to which the blast is preheated; and the basicity of the
slag.
Preheating the Blast.—Modern furnaces have four “hot-blast stoves,” and
some very recent ones have even five. These stoves are heated by the waste gas
from the blast furnace, which contains a little sensible heat, and which is a fuel by
virtue of containing about 22 to 27 per cent of CO. It is drawn off at the off¬
takes; some of the dust is separated from it, and it may even be scrubbed clean.
Then about one-third of it is burned in the stoves, and the remaining two-thirds
is used to create power, by burning under boilers or in gas engines. One type of
stove is shown in Fig. 4, the important feature being the large surface of brick ex¬
posed by the “ checkerwork” construction, in order that it may absorb and impound
heat from the burning gas. In a plant consisting of four stoves, three will be
heating up in this way and one will be giving up its impounded heat to the air
blast that is being driven into the furnace. A new stove is put on blast heating
about once an hour. This is all regulated by the manager in proportion as he
wishes to raise or lower the temperature of the air blast.
Dry Blast.—The most important and spectacular improvement in the iron
blast-furnace process of the past few decades is James Gayley’s use of dessicated
blast. The moisture usually driven into the furnace with the blast cools the
smelting zone by the endothermic reaction of dissociating the steam. Elimi¬
nating this reaction produces an effect in fuel economy greatly in excess of the
theoretical amount of heat saved. This has been explained by J. E. Johnson,
Jr., upon the basis that the operative heat of the furnace is the heat in excess of
the minimum temperature at which the process can proceed. This operative
heat is only about 100 to 200°; therefore an increase of 100° of hearth temperature
might mean an increase of 50 or more per cent in heat efficiency.
Fuels Used.—Coke is the predominant blast-furnace fuel, and wood charcoal
is the next commoner. Sometimes this is used because ore can be smelted
economically where charcoal is cheap and coke is not; but more often it is because
charcoal contains less sulphur than coke, and therefore pig iron with lower sulphur
may be produced. The charcoal furnaces must be low in height, however,
because the structural weakness of charcoal makes it incapable of supporting the
weight of material that would prevail in the ordinary coke furnace having 100 ft.
height. The use of anthracite coal mixed with coke is a commercial matter
dependent only slightly upon the lower sulphur involved.
Other Processes of Reduction.—Iron ores occur in great quantity in many
localities where coke is not available for smelting. Therefore electric smelting
has been tried, in which charcoal is used in amount only sufficient for chemical
reduction, and the heat has been furnished by electricity generated from water¬
power. In most cases the cost of the electricity and the small output in relation
to the overhead cost has rendered these attempts commercial failures, although
the technical operation of the processes has been successful.
Iron may be separated from its oxygen by the action of reducing gases below the
melting point. This results in the production of an unmelted metal, usually in the
form of sponge. The process is so easy and so economical of fuel that it has tempted
many inventors, especially since it requires only liquid or gaseous fuel. The com¬
mercial stumbling block is that the product is only equivalent to steel scrap, which is
usually abundant and cheap, even in many localities where coke is not available.
Millions of dollars have been lost by men who have been influenced by the technical
success of the processes, without understanding this simple commercial factor. How¬
ever, the process may have a certain commercial value through the use of the iron
sponge in isolated regions as a precipitant for copper in the hydrometallurgy of that
metal.
k
SECTION XXV
CEMENTS AND GLUES
By S. S. Sadtlee1
Classification.—There is a nearly infinite number of cements and they may

be classified in various ways, but for convenience at this time they might well be
subdivided into three classes: (1) Building cements, including portland or, so
called, ordinary cement, Puzzolana, Rosendale mortar and hydraulic limes; (2)
miscellaneous cements, or cements for other than building purposes, and lutes;
(3) glues and casein preparations. The building cements I have considered under
the section on “Materials of Construction” and the present section will treat
only of classes (2) and (3).
The word cement connotes a more or less permanent bonding material while lute
refers to rather temporary bonds. For instance, plaster of paris with water forms a
cement, and glue and glycerine form a lute. For convenience in reference, cements
and lutes will be classified by the uses to which they are to be put, though what is
given here can only be a skeleton or outline of the innumerable number of possible
combinations that will adhere or cause other substances to adhere after first mixing,
by melting, evaporation or chemical union. In many cases better proportions of the
substances given may be found and other fillers and binding agents will suggest them¬
selves to the chemist or engineer.
Lutes are generally used to confine gases in or out and fresh portions are added as
needed. It is not serious if there is a leak for more of the preparation can be applied.
A luting compostition is chiefly selected to resist chemically the action of the gases or
liquids, while cements are generally selected for a certain strength and durability.
Cements are therefore more carefully applied. The surfaces to which they are to
adhere are often roughened or scraped so as to promote adhesion. Sometimes they
are painted with silicate of soda, neat cement, glue, shellac, etc., so as to give good
contacts. Solder is rightfully a cement. A good operator tins the surfaces of the
work after applying the flux, before putting the solder on in quantity. It might be
mentioned here, for the few, perhaps, who do not already know, that solder in paste
form is very convenient to use. It consists of finely divided solder mixed with a non¬
corroding flux and a lighted match often supplies enough heat to melt the solder and
heat the flux sufficiently to cause it to clean the tarnish off the work so the solder
can unite the surfaces.
Waterproof Cements.—As the chief waterproofing substances available we

have the bituminous compounds, asphalts, pitches, gilsonite and blown petroleum
residues, mixtures of these substances and mixtures of them with inert material.
Candle pitch is an artificial bitumen that is tough and very sticky and much used
for waterproofing. For instance, it is used as a waterproof coating between layers
of cardboard.
1 Chemical engineer, S. P. Sadtler & Son, Philadelphia, Pa.
841
Of the asphalts, refined Bermudez is to be preferred as it is a nearly pure bitumen

and forms homogeneous solutions with suitable solvents. A minor percentage of
boiled linseed oil or blown petroleum oil is useful sometimes for tempering and fluxing.
For solvent heavy naphtha is generally used. It does not dissolve all the asphalt, such
as the so-called asphaltene portion, but it thins the asphalt very well. Coal-tar
naphtha is better, however, at present it is slightly more costly. As voids are apt to
form in a painted asphalt coating, some form of filler is generally employed, either by
coating paper so that it becomes more or less impregnated or by the addition of a filler,
such as silex or infusorial "earth, to the asphalt.
For a soft, watertight coating in cement work, blown asphalt and infusorial earth,
thinned with a solvent such as heavy'naphtha, is employed.
The use of blown asphalts with natural asphalts and gilsonite is in some cases
desirable, as the former are ideal fluxes for hard asphalts and tend to prevent brittle¬
ness or separation of the films when the solvent evaporates. The harder or higher
melting point grades of these blown residuums are capable of making good cements
alone or with filler, but are rather hard to dissolve. Such artificial asphalts have
melting points as high as 150°C., while the softer ones that are probably best for
fluxing have melting points about 80°C.
Testing' of Bituminous Cements.—As to the choice of a suitable water¬

proofing substance for reservoirs, etc., we are chiefly guided by the experience
of those who have constructed such works in the past and we seek from them
in formation as to the materials used and are concerned their specifications for those
materials. But in addition to this, certain small-scale waterproofing tests can
be made by coating porous bricks with bituminous cements with and without
fillers and noting the difference in weights, before and in given intervals after
soaking in water.1
Lutes of boiled linseed oil, thickened with clay, asbestos, red or white lead, etc.,
are waterproof, but must be made thick enough from the filler added to resist mechan¬
ical action at the start. Linseed oil and whiting also make a good waterproof cement.
Flaxseed meal made into a stiff paste with water is useful as a lute for low-pressure
steam connections and is easily applied. Portland cement only serves as a waterproof
cement when given time for the preliminary setting to take place. It is not gener¬
ally entirely impervious to water.
For all practical purposes the use of trade preparations, such as those containing
metallic soaps or oil emulsions, serves to render concrete approximately impervious
when the proper amount of water is used and the mass thoroughly puddled so as to
eliminate voids.
For making small electrolytic cells it is desirable to use white iron-free portland
cement and white plasterers* sand that has been sieved. If a 1:2 mixture is made
with this sand and the prescribed quantity of an approved waterproofing substance,
such as the product of the Newberry patent, aluminum oleate, Aquabar, Toxement,
or the General Fireproofing Co.’s compounds, is added and care is taken to workout
voids and the article in the green is kept moist, a waterproof container can be made
which will stand dilute acids or aklalies, but not the two alternately. Of course
acids act on the free lime, but silica is formed from the calcium silicate which tends to
protect the preparation. When acids are to be used it is best to treat the finished
1 The time of absorption of a given number of drops of water by a given amount of surface is also
a fair guide as to the waterproof qualities of the solid.—Editor.
CEMENTS 843
cell with hot paraffin unless the melting point of paraffin is below the temperature at
which the cell is to be used.
Making Lutes with Portland Cement and Silicate of Soda.—A little silicate of
soda in the water used with Portland cements will help toward a quick initial set
for under-water work. Water containing 2 to 40 per cent of silicate by volume
causes a prel minary set in 2 hr. that prevents washing away by water. It
injuriously affects the natural gradual hydration, however, unless the sodium
hydroxide, carbonate or sulphate formed is washed out by excess of water, which
is difficult to do. Mixtures containing above 4 per cent of silicate of soda show
efflorescence; this is particularly marked when made with 20 per cent aqueous
solutions of silicate. The silicate-of-soda-setting is at first due purely to double
decomposition of sodium silicate and calcium hydroxide. If very strong silicate
is used the mass stiffens so fast that it cannot be worked. This is true of all
concentrations of 50 per cent or over.
Therefore for immediate immersion in running water silicate of soda acts as a

binder, while the particles of the cement form gelatinous colloids instead of being
washed away. The soluble sodium salts are then gradually more or less eliminated
and apparently normal concrete results. Glue solution (3 to 5 per cent strength)
with white portland cement insures very dense compositions, but the tensile strength
is likely to be a little deficient. It is possible that the glue aids colloid formation, and
protects the mixture against too rapid crystallization.
Oilproofing is often necessary to lute apparatus containing oil vapors or vapors

of liquids immiscible with water. Glue or silicate of soda compositions are
probably most used. A typical composition is: Good glue 2 parts by weight,
glycerine 1 part, water 7 parts. The glue is first softened by the water, then
liquefied by heat and the glycerine is incorporated. This is a good lute to render
corks vacuum tight and for stopping small leaks of combustible gases and oil
vapors.
A putty of molasses and flour has been found useful in laboratories and even in the
works for oil vapors. A lute made according to the following proportions is very
satisfactory: Glycerine 90 parts by volume, water 10 parts, made into a stiff putty
with: Litharge 90 parts by weight, red lead 10 parts. It takes several hours to stiffen
and about a day to set. This mixture was formerly used for luting in gage-glasses
on low-pressure boilers.
Several cements of which silicate of soda is an active principle or the chief binding
substance are as follows: Silicate of soda (about 30°Bd.) and whiting, made into a
stiff putty. This slowly sets by drying if not by chemical action. If precipitated
magnesium carbonate be used the setting is so quick that it is hard to use the mixture.
If it is used, however, the silicate should be diluted to one-third to one-half ordinary
strength.
Barium sulphate is often used with silicate of soda. The silicate should be about
30°B6. If the mixture is heated the cement dries faster and thereby sets. There is
no chemical action. A very strong and ultimately hard cement for a variety of
purposes is made by incorporating hot glue solution with plaster of paris.
Acid Proofing.—The asphaltic preparations referred to under waterproofing

are acidproof if other than basic fillers are used. There are also some good pro-
prietary acidproofing preparations on the market, but their compositions are

not disclosed.
Black putty is made by intimately mixing equal weights of china clay, linseed oil and
melted pitch. The ingredients must be anhydrous.
Rubber cements may have varying compositions, but the following are prac¬
tically the extremes. Equal weights of rubber and boiled oil are taken; the
rubber is first dissolved in carbon disulphide in the proportion of 4 c.c. CS2 to
1 g. of cut-up rubber. Hot boiled linseed oil is then added. The solvent is not
removed by evaporation until the paste is applied to the apparatus to be luted.
The other formula, to which reference has just been made, differs in having four
times as much linseed oil and then fireclay or other filler, such as silex, is used: Crude,
finely cut rubber 1 part, boiled linseed oil 4 parts, fireclay 6 parts by weight. If equal
parts of fresh unvulcanized rubber are used the masses are so stiff that they should
probably be used without filler. If as much as 4 parts of linseed oil are used consider¬
able filler can be incorporated and made into a workable putty.
Melted sulphur with fillers of stone powder, cement, sand, etc., are used, such a
mixture even being recommended for small electrolytic tanks. The following is used
for hydrochloric acid vapors: Rosin 1 part, sulphur 1 part, fireclay 2 parts, by weight.
Linseed oil (boiled) and fireclay stand most acid vapors.
Inert fillers in boiled oil stand acid vapors (even nitric acid): Silex 20 to
30 lb., flock asbestos 10 lb. are fed into a mixer, a little at a time, with 6 qt. of
boiled linseed oil. There are numerous acid-resisting mixtures possible in which
silicate of soda is used. This is because in spite of the strong basic character of
silicate of soda, the silicate is superficially changed to colloidal silica, which
continues the cementing work, at first effected by the silicate of soda. Barium
sulphate, powdered glass, china clay, etc., are used with silicate of soda slightly
diluted with water. A strength of about 30°B6. is best.
For dilute hydrochloric acid one may use white china clay 1 part by volume, fine
white sand or powdered quartz and sand 2 parts by volume. Mix thoroughly and
work up with just enough silicate of soda, dilute with an equal volume of water to
make a paste. This can be rendered more impervious to water by the judicious incor¬
poration of organic colloids. If a little fine casein be incorporated with the silicate
of soda in a mixer so that the mixture is quite smooth the mass is better. About
5 per cent of fine, dry casein powder is added, based on the weight of silicate. If fresh
milk curd can be used, corresponding to the same dry weight of casein and allowance
is made for the water contained, a mechanical mixer will not be needed.
Chlorine Resistant.—The most reliable is made with portland cement as the

chief ingredient: Powdered glass 1 part, portland cement 1 part, silicate of soda
1 part. The last mentioned should be diluted considerably so as not to set too
rapidly. Portland cement quickly reacts with silicate of soda, while the pow¬
dered glass and clay react more slowly but finally render the cement insoluble as
regards its mass.
General Mixtures with Plaster of Paris.—This series of lutes and cements is
highly important, and individual formulas are adapted to prevent the escape of
hydrocarbon and other gases in furnace work, as cements for mechanical purposes
and for wall coatings such as plaster. When mixed with asbestos, straw, plush
CEMENTS 845
trimmings, hair, etc., the plaster of paris mixtures are physically strengthened or
reinforced.
Some soluble sulphates form double crystallizable sulphates with calcium sulphate
and water. They set harder and are more impervious than calcium sulphate (plaster
of paris) alone. It is desirable not to take equal molecular quantities of the soluble
sulphates to form the full double sulphates, but about half of these quantities.
Sodium, potassium and aluminum sulphates are used for this purpose. The two
latter are preferable.
According to Sigmund Lehner, a little borax in the water used makes hard plaster
cements and regulates the setting: 12 volumes of water to 1 volume saturated borax
solution sets with plaster in 15 to 20 min. at 10°C.; 8 volumes water to 1 volume satu¬
rated borax solution sets in 1 hr. In another recipe, 1,500 g. of borax and 150 g. of
magnesium oxide are melted together and powdered. This powder is then mixed with
75 g. of plaster of paris. Borate of magnesium thus predominates and protects the
plaster from being washed away by water.
Marine Glue.—A standard preparation of this class of lutes (which are

applied to crevices in a heated and fluid state, and become firm but not brittle
when cold) is composed as follows: Crude rubber 1 part, shellac 2 parts, pitch
3 parts, by weight. The rubber is first dissolved in carbon disulphide or turpen¬
tine before mixing with the heated (not superheated) mixture of the other two.
The advent of blown petroleum residuums has made it possible to make up hard
but flexible compounds without rubber. Grahamite is a good base to which
fluxes such as these just mentioned, or soft asphalts, are added.1
Gasket Compositions.—In the laboratory one can generally make out for
low temperatures and pressures by saturating heavy “kraft” wrapping paper
with soft pitch, such as wood pitch for steam or with gelatine and glue (hecto¬
graph composition) for oils. For high pressures, slots filled with lead rings and a
V-shaped rim to the lid are most satisfactory. For large retorts asbestos rope
and graphite are used for moderate pressures and lead for high pressures. Rub¬
ber compositions are most used for low temperatures where oils or solvents are
not present.
Machinists’ Cements.—These are the well-known red and white leads.
The red lead is often diluted with an equal bulk of silica or other inert substance
so as to make the mixture with oil less powdery on drying. One way to accom¬
plish this is to add rubber or gutta-percha to linseed oil in the following way:
Linseed oil 6 parts, rubber or gutta-percha 1 part, by weight. The rubber or
gutta-perchd is dissolved in sufficient carbon disulphide to give it the consistency
of molasses. This solution is then mixed with the oil and left exposed to the air
for about 24 hr. The red lead is then mixed to a putty. Oxide of iron makes less
brittle cements than red lead. A portion of the red lead may be substituted by
the cheaper red oxide of iron.
Leather Cements.—Equal parts of good hide glue and American isinglass, are
softened in water for 10 hr. and then boiled with pure tannin until the whole mass
is sticky. The surface of the joint should be roughened and the cement applied
hot. Another is: One pound of finely shredded gutta-percha is digested over a
water-bath with 10 lb. of benzol, until dissolved, and 12 lb. of linseed oil varnish
1 Blown fish oil and blown cotton seed oil may also be used where quality is desired and expense
a subsidiary consideration.—Editor.
stirred in. Still another cement is formed by melting together gutta-percha

8 oz., pitch 1 oz., shellac 1 oz., olive oil 1 oz.
Iron Cements.—When iron in a fine state of division, as in fresh oil-free
filings or cast-iron borings that have been powdered, is mixed with an oxidizing
agent, such as manganese dioxide or a substance electronegative to iron, such as
sulphur, in a good conducting solution such as one of salt or salammoniac,
galvanic action sets in very rapidly, ammonia is given off (if salammoniac be
used) and the iron swells, by forming iron oxide, and cements the mass together.
It is best diluted with portland cement. One formula is: Iron filings 40 parts,
manganese dioxide or flowers of sulphur 10 parts, salammoniac 1 part, portland
cement 20 to 40 parts, water to form a paste.1 These cements are used exten¬
sively in foundries, etc. An odd application of the use of this cement is made in
works that have quantities of iron in a more or less fine state, such as scrap from a
magnetic separator. The iron is spread in layers, several inches in thickness, and
is then moistened with salammoniac solution. Nature does the rest. The
expansion of the iron, in forming oxide, knits the whole into a mass that forms
good paths where it might otherwise be muddy in the yards of factories.
Crucible and Furnace Cements.—These cements for furnace linings are gen¬
erally applied by hand. Sometimes air guns are used, however. For acid
slags silica is most used and for basic slags magnesia, chromite, etc.
Silicate of soda and powdered glass or sand are suitable compounds for cementing
lids on crucibles, etc. Sometimes the “iron cements” are used for such purposes, or
iron filings and a little manganese dioxide are added to the above compositions for
crucible cements. The first of these is the most practical for most purposes.
The best known cement for graphite compositions is fireclay, which with water
binds the graphite fairly well and is used in making crucibles, muffles, etc. Carborun¬
dum powder is sometimes used with fireclay for furnace or crucible cement.
In some cases it is desirable to have an all-carbon binder, and for this purpose tars2
or soft pitches are taken. Very little binder can be used, however, or the material
will crack when heated, as the carbonizing of the pitch shrinks the binder somewhat.
Starch paste has been recommended for this purpose, but the shrinkage is greater and
the binding is not so good.
A strong, waterproof cement that will stand high temperatures may be made by
mixing powdered silica or fine sand and powdered silica with a solution of magnesium
chloride of about 10 per cent strength. This composition is applied as a putty
and then painted or soaked in a solution of silicate of soda of about 30 per
cent strength. This forms magnesium silicate as a binding material for the silica.
Magnesia composition for furnaces, etc., may be made by mixing magnesia burned
at incandescent heat or hard-burned, 80 per cent and magnesia light burned (just
sufficient to drive off CO2 (560°C.), 20 per cent. The composition is made into a stiff
putty with water and shaped as desired (or better still with magnesium-chloride
solution, 18°B<$., if the cement is not to be heated to over 500°F.). The water must be
driven off very slowly. A small proportion (3 to 6 per cent) of good asbestos fiber may
be worked in to keep from cracking.
Oxy-chloride (Sorel) Cements.—These are sometimes called stone cements

and the only one of practical importance outside of dental work is that made with
magnesium chloride. If the magnesium chloride is quite pure, i.e., free from
1 The well-known “Smooth On” is such a paste.—Editor.

CEMENTS 847
potassium, sodium or calcium chloride, the solution used need not be more than
19°BA, but if commercial German magnesium chloride be used it requires a
solution of 20 to 22°B6.1 The magnesium oxide must be freshly burned at a
dull-red heat as the light oxide that is requisite for this cement readily absorbs
carbon dioxide from the air and becomes unsatisfactory or of no use for making
cement compositions. A diluent for the magnesium oxide is generally used, such
as silica or wood pulp, ground wood, fine cork, infusorial earth, etc. Colors such
as iron oxide and ochres are used, or anilines on a fuller’s earth base.
The preparation is applied as a paste after mixing solids and the liquid. A good
diluent is moistened portland cement as it tends to harden the mixture. This cement
is said to have been used by the Germans in France and Belgium for emplacements for
big guns instead of concrete as it sets much quicker. It is not entirely waterproof in
itself, but can be rendered so. The biggest use is for floors, in stores, railway cars, etc.
When wood meal or cork is used with, say, infusorial earth as a filler with the Sorel
stone it makes a tough resilient surface that wears well and is less tiring to the feet
than, cement, concrete, etc.
Fig. 1.—Cement-testing machine.
Testing in General.—Many of the cements treated on pages 940 to 943, and

949 to 951, may be tested as to strength by being made into test forms or bri¬
quettes and, after setting, pulled with a test machine such as shown above.
Resistance to acids, alkalies, heat, oil vapors, steam, etc., are all tested in obvious
ways by the use of the same test pieces. In all important work testing should
be done.
1 The figures, 27 to 30°B6., frequently given, are wasteful of chloride.—Editor.
Glue.—There are several varieties of glue.1 The most important are bone and
hide glues, obtained from sources disclosed by their names. A fish glue is made
from fish skins and offal, but is a cheaper grade and inferior for must purposes.
It is generally met with as a rather badly smelling liquid. It is used by cabinet
makers for certain purposes, where strength is not the great desideratum. Of
course fish gelatin made from bladders of the sturgeon and other fish is a distinct
and superior product. Liquid glue is made from high-grade hide glue by treating
with strong acids such as nitric or hydrochloric. It has lost its power of gelatiniz¬
ing but not its adhesive power.
Glue Testing.—Glue and gelatin occur in commerce in a great variety of
forms (sheet, flake, shred, ground, powdered, etc.). While gelatin is usually
colorless or else clear pale-yellow or brown, glue varies in color from milk-white
to dark brown, and all possible shades yellow, transparent or opaque.
Although appearance2 is no certain criterion of quality, glue is often erroneously

judged by its color, clearness, fracture, etc. Since glue is used for a great variety of
widely differing purposes, the use for which the glue is intended should always be borne
in mind when subjecting it to test or technical examination. A number of tests2 will,
therefore, be given, which may he conveniently run in consecutive series, and their
practical significance will be pointed out in a few instances.
Thin blown glasses about 3Yi in. high and 2% in. in diameter are convenient for
making tests. Twenty-five grams of each glue to be tested is broken into small pieces
and soaked in 100 c.c. of cold water until thoroughly softened. Thick sheet or flake
glues must be soaked over night and should be allowed to stand in a cool place. With
the glues under examination, there are at the same time soaked up a number of glues
of known strengths (standards), for tests of glue should always be comparative. It
is desirable and convenient to use the standards hereafter described (see page 850).
With high-test glues, or in cold weather, less glue may be used, providing the unknown
glues and the standards are treated exactly alike. In warm weather low-test glues
must sometimes be tested 30 g. to 100 c.c. Gelatins are usually tested 10 to 100, or
even 3 to 100.
When the glues are thoroughly softened, the glasses are immersed in a water-bath,
and their temperature raised to 80° with constant stirring to insure complete solution.
With thick-cut glues, care must be taken to see that no undissolved pieces stick to the
bottom of the glass. The following tests are then made in the order given:
Reaction.—This is determined with strips of litmus paper. Where the degree of
acidity or alkalinity is desired, a seaparate titration is made.
Viscosity or Running Test.2—The viscosity is taken by running the hot glue solu¬
tion at 80° from a pipette, and noting the time of efflux by a stop-watch. The relative
viscosities are thus fixed in seconds.
i Glue in small oblong cakes (3 by 6 in.) is known as Cologne glue. Scotch glue is dark-reddish-
brown, in large oblong sheets. French glue is generally about 9 in. square, and is frequently stamped
with a trade-mark (medal glue). Glue opaqued with oxide of zinc is often called Rvssian glue, although
in America thin-cut opaques are commonly known as white-shell glues. Ribbon glue comes in
thin strips about 2 by 6 in., and noodle glues in thicker and narrower pieces. The dark, heavy noodle
glue popular in Eastern countries is called Bazaar glue. The form in which a glue is dried or cut does
not of course affect its quality.
« E. G. Clayton (Journ. Soc. Chem. Ind., 1902, Vol. 21, p. 670) at the conclusion of a paper on the
technical examination of glue says “while it would be rash to form a judgment on glue from a single
test, the evidence afforded by a number may be irresistible. The expert’s wisest system appears to be,
not to rely upon single short-cut tests of general quality, but to employ a number of methods, including
any having especial bearing on the prospective or present uses of the glue, and then base his conclusions
on a consideration of all the results together."
CEMENTS 849
The pipette adopted as standard has the following dimensions:

Capacity. 45 c.c. of water at 80°C.
Internal diameter of effluent tube. 6 mm.
External diameter of effluent tube. 9 mm.
Length over all of effluent tube. 7 cm.
Smallest diameter of outlet (about). 1.5 mm.
Outside diameter of bulb. 3 cm.
Length of bulb. 9.5 cm.
Length of upper tube. 22 cm.
The pipette should permit the efflux of 45 c.c. of hot water at 80° from the glue
bath in exactly 15 sec. The viscosities of glues vary widely, as may be seen from the
table of standards on page 850. Great care must be taken to make two pipettes
that will give concordant results. The size and shape of the outlet hole, and the length
and diameter of the effluent tube are the chief factors controlling the time delivery.
The efflux hole is made by cutting the effluent tube square across, and holding it
vertically in a bunsen flame with constant rotation. As the glass softens, the hole
gradually draws together, and after a few trials can be brought to the desired size.
It is desirable to have the lower graduation joint just where the effluent tube joins the
bulb, for otherwise in glues of high viscosity there is much uncertainty caused by
dribbling of the last few drops.
While running the pipette may be kept in a simple thermostat, consisting of a
water-bath specially made for the purpose. The projecting effluent tube is protected
by a mica cylinder through which the end point is observed. When the thermostat is
used a small piece of rubber tubing controlled by a pinch cock is slipped over the upper
end of the pipette, or a glass stopcock may be fused on.
After each determination, the pipette is washed out with hot water from the glue
bath, and care must be taken that no undissolved glue, glue “skin,” slime, or other
obstruction clogs the outlet, even momentarily during use.
More complicated viscosimeters, as Engler’s1 or the Rideal-Slotte,2 may be used,
but they are cumbersome and slow of operation, and therefore impractical in routine
work, where many determinations are made.
Odor.—To an experienced nose, the odor of the hot glue solution usually gives some
indication of the raw material from which it is made. Decomposition is readily
detected, although it is frequently masked by antiseptics or ethereal oils. Glues are
rated as “sweet ” or “off ” as the case may be. With food gelatins, freedom from odor
is essential. •
Grease.—A brush-full of the glue solution is mixed with little aniline or other color,
and painted out on a piece of white paper, when spots or “eyes” appear roughly pro¬
portionate to the amount of grease present. The grease is recorded arbitrarily on a
comparative basis. For an exact estimation of fat in glue, etc., Kissling dissolves 20 g.
of the sample in 150 c.c. of water containing 10 c.c. of hydrochloric acid of 1.19 sp. gr.
The liquid is heated for 3 or 4 hr. on a water-bath under a reflux condenser. The
solution is cooled, 50 c.c. of petroleum-ether added, the liquid well shaken, and, after
standing until clear, a known measure of the solvent is drawn off, evaporated, and the
residue weighed.
Foam.—Foam is determined by agitating the glue solution with a rod or
mechanical agitator and, like grease, is recorded comparatively. Some of the condi¬
tions affecting the foaming of glues are discussed by Trotman and Hackford3 and H. J.
Watson.4
1 Journ. Soc. Chem. Ind., 1890, Vol. 9, p. 654.
2 Journ. Soc. Chem. Ind. 1891, Vol. 10, p. 615.
3 Jour. Soc. Chem. Ind., 1906, Vol. 26, p. 104.
4 Ibid., p. 209.
Comparative Set.—The glasses are taken from the water-bath, and note is made
of the comparative set or order in which the solutions gelatinize. In warm weather,
especially with glues of low jelly strength, the glasses must be allowed to stand in ice
water. As a rule, glues and gelatin gelatinize in order of their jelly strength.
Jelly Strength or “Finger” Test.—When the glue solutions have gelatinized, the
glasses are arranged in order of the strength or resiliency of the jellies, as determined
by pressure with the finger tips. Notwithstanding the personal equation, expert
operators obtain much more uniform results than are given by the various mechanical
devices.
Standards.—The selection of standards is of great importance, for once they

are taken, all unknown glues are measured by them. Most published results of
glue tests cannot be compared with each other, because of the great diversity of
methods and glues used, and the absence of standards. That definite standards
of jelly strength will simplify and harmonize the grading of glue is self-evident.
Sixteen arbitrarily established, nearly equidistant grades cover the usual range of
jelly strengths. The grades are given values from 10 to 160, thus allowing 10 points
between each grade. The following table gives the viscosities (as determined by the
standard pipette, see p. 849) and the jelly strengths (as determined by Alexander’s
jelly tester) of these 16 standards. The determinations were made on solutions of the
glues containing 25 g. of glue to 100 c.c. of water. For the standard viscosities there
are taken the average of laboratory tests extending over many years, and inasmuch as
standards of the same jelly strength are apt to vary in viscosity, there is indicated in the
table the reasonable limit of such variation. Glues often greatly exceed these limits,
but should not be chosen as standards.
Viscosities (in Allowable

Jelly strength Jelly strength
seconds) at 80°, variation of
Standards (in ounces) at (in grams) at
solution 25 to viscosities
10°G. 10°C.
100 (in seconds)
10 15 Yi ± 1 to 4
20 16 ± 1 to 4
30 wx ± 1 to 4
40 17 ± 1 to 4 60 1,701
50 18 ± 1 to 2 82 2,324
60 19 ± 1 to 2 104 2,948
70 20 ± 1 to 2 126 3,572
80 21 ± 1 to 2 148 4,196
90 22 ± 3 to 4 170 4,820
100 23 ± 3 to 4 1*92 5,443
110 24 ± 3 to 4 214 6,067
120 25 ± 1 236 6,691
130 26 ± 3 258 7,314
140 28 ± 5
150 34 ± 8
160 40 ± 12
The highest standards from 60 and upward should be neutral hide glues, clear,
clean, well made, free from any odor of decomposition, and practically free from foam
CEMENTS 851
a-nd grease. For the lower standards, bone glues should be chosen, because most
tow-test glues are bone glues, and besides low-test hide glues usually have abnormally
aigh viscosities.1
Water.—From 2 to 3 g. of glue are roughly granulated and dried at 110 to 115°
until constant in weight. The estimation of water is without much practical signifi¬
cance if the glue is commercially dry; for the more moisture a glue contains, the weaker
i-t will show on the viscosity and jelly tests.
Ash.—The ash of glue varies considerably, both in amount and composition. In
the case of some bone glues the ash consists largely of phosphates, while in hide glues
considerable calcium sulphate or other calcium salts (formed upon neutralization of
the lime used in preparing the stock) are apt to be present. Chalk, clay, zinc oxide,
and lead carbonate and sulphate, which are sometimes added as whiteners, and zinc
sulphate, which is used as a preservative, are often found.2
For the estimation of ash, a weighed quantity (2 or 3 g.) of glue is incinerated in a
large covered platinum crucible, using a few drops of nitric acid if necessary, to insure
the oxidation of the carbon. The heating must be conducted with care, as the glue
at first intumesces violently. The ash of bone glue is supposed to be fused by the
heat of the bunsen burner, to be neutral, and to contain phosphoric acid and chlorides;
whereas the ash from hide glue is supposed to be infusible owing to the presence of
lime, to be alkaline, and generally free from phosphoric acid and chlorides.
The composition of the ash in reality depends more upon the nature of the process
than upon the raw material used in making the glue, and from a practical standpoint
its estimation is generally unnecessary, unless to detect some added adulterant or
for the purpose of comparison. Hide and bone glues are frequently mixed, both in
the liquor and in the finished form, and it is unsafe to attempt to judge the raw mate¬
rial of a glue on the basis of its ash.
Water-absorption Test (Schattenm aim’s Test).—A known weight of glue is im¬
mersed in water at ordinary temperature for 24 hr. In this time, the high-test glues
absorb from 10 to 15 times their weight of water, whereas weaker glues absorb only
3 to 5 parts.3 Very low-test glues form a slime rather than a jelly.
1 Many American manufacturers and dealers use for comparative testing the so-called “Cooper
grades” which were established many years ago by Peter Cooper, a pioneer American manufacturer,
as grades under which his glues were sold.
A Extra.. Standard 130.
1 Extra. Standard 120.
1. Standard 110.
1 X (called one cross) Standard 100.
1J4. Standard 90.
m. Standard 80.
m. Standard 70.
Standard 60.
m.
m. Standard 50.
Standard 40.
i% .
2 . Standard 30.
In all probability, these were originally based upon the distance a certain weighted rule would compress
a certain dish or bowl of jelly. The “ No. 2” allowed it to sink 2 in., the “No. 1,” only 1 in., while the
intermediate grades were measured in eighths of an inch.
2 It is probable that any crystalloid mineral compounds will affect the glue somewhat injuriously.
3 S.Rideal soaks 10 g. of the coarsely powdered sample of glue in water at 15° in a weighed beaker for
48 hr. At the end of this time the water is carefully decanted, and the increase of weight ascertained,
the character and odor of the jelly being also, noted Good glues give a firm jelly, and absorb from 5
to 9 times their weight of water. In other cases, the product is not jelly but a slimy liquid. According
to Kissling the amount of water absorbed by glue affords no indication of its cohesive power, and Fels
describes it as an uncertain test. If the sample is finely ground or powdered, this test cannot be made.
H. R. Proctor (Kolloidchem. Beihefte, 1911, Vol. 2, p. 243) found experimentally that the swell¬
ing maximum of gelatin is influenced by the volume of the original jelly from which it was dried.
The presence of highly ionized acids enormously increases the water absorption of gelatin.
Although the best glues take up more water than inferior qualities when immersed
in the liquid, well-made and well-dried glues are much less hygroscopic than badly
made specimens, or than those prepared from inferior materials. The latter are also
liable to undergo putrefaction on exposure to dampness. .
The water absorption is roughly proportionate to the jelly strength, but it so inac¬
curate that it is hardly worth the trouble involved, unless as a rough approximation.
Tensile Strength of Glue.1—Many methods have been suggested such as

Bauschinger’s, Kissling’s, and Bideal’s.1 Bideal has devised the following
method: Two plane surfaces of biscuit porcelain, ground so as to fit, 1 sq. in. in
area, are soaked with a solution made up of 1 part of glue in 2 parts of water at 70°
for 30 min.; they are placed together, weighted with 5 lb. and kept for 5 days
in a cool room. They are then fixed in a testing machine, and the weight neces¬
sary to tear them asunder is determined.
The writer has used with success a slightly modified form of a method originally
suggested by Millar.2 Ten per cent solutions of the glues to be tested are made up.
Strips of filter or other paper are cut, 1 in. in width and about 18 in. in length. These
are dipped in the solution at about 50°, and then hung up to dry. When the first
coat has dried, they are again immersed in the solution and allowed to dry, the por¬
tion that was lowest being placed at the top, so that as far as possible a uniform layer
of glue is produced on their surfaces. After air drying, the strips are heated for 1 hr.
in the water-oven. Two or three lengths of 3 in. are cut out of the central portion of
the strips. These are then separately tested by fixing between two indiarubber-
covered clamps, one of which is fixed to the table and the other attached to the
beam of a strong balance, the balance with the clamp, of course, being first counter¬
poised. Weights are added to the other pan until the paper breaks. A blank test is
then made with the unsized paper, and this is deducted. Several experiments should
be made, and the mean taken. A good glue should require some 9 or 10 lb. when
tested in this way to break the sized paper.
Interpretation of the Results of Tests.—Glue is used for a multitude of pur¬

poses, and each line of work has its special requirements. Experience has
taught that glue of certain characteristics made from certain kinds of stock will,
in certain cases, do the best work at the lowest cost. It is, therefore, best to test
unknown glues against those which are actually doing the work, and give the
ones which show up well a trial under actual working conditions. If this is
impossible, a laboratory test must be devised, in which the actual working condi¬
tions are exactly simulated; the more nearly this is done, the more valuable the
test will be.
Practical Work with Glue.—The following precautions should be observed
in practical work with glue in order to get the best results. The glue should be
weighed and the water measured according to proportions already established,
either by laboratory test or in practical work. The amounts should not merely
be guessed at. The water and glue should then be mixed and stirred and allowed
to stand in a covered vessel for at least 2 hr. It should then be heated carefully.
About 140°F. is the best working temperature and it should never exceed 150°F.
The glue should not be heated longer than is necessary and for the highest type
of work the glue should not be used after being heated for more than 8 hr. The
1 “Dictionarv of Applied Chemistry”—Thohpe, 1912. Vol. 2, p. 750.
2 Joum. Soc. Chem. Ind,. 1899, Vol. 16.
CEMENTS 853
surfaces to be united should be dry and at least as warm as the air in the shop, which
should be as warm and humid as the health of the operatives will permit.
Casein Preparations.—Casein is a protein substance obtained by drying the
nearly fat-free curd of separator skimmed milk. It is a mild acid in its properties
and combines with bases such as soda and ammonia and also with alkaline
earths and heavy metal oxides. The usual substances employed to form soluble
compounds with casein are soda ash, borax and ammonia.
Calcium, magnesium, aluminum, zinc and lead salts are used to precipitate casein
as the corresponding insoluble salts. They, are hard to form into colloidal masses,
however, and it is more usual to make insoluble masses of casein with formaldehyde.
Casein, besides being acid in its properties, because of its amino-groups acts like a
weak base with acids and combines with certain quantities of sulphuric and hydro¬
chloric and other strong acids.
Casein is usually precipitated with sulphuric or hydrochloric acid but “self
soured” and rennet-precipitated casein aie on the market and are better grades for
some purposes. It is much used as a sizing or finish for paper, as a glue for veneers,
etc., and for making horn or ivory-like substances for billard balls, cigar holders,
insulating parts, etc.
Casein is often used in the form of sodium caseinate as it can be kept indefinitely
as a dry powder, and when added to water makes a strong glue or a sizing preparation.
Instead of using casein in this way, if sodium silicate is mixed to form a thick paste
with finely powdered casein an especially strong glue is obtained. Doubtless sodium
caseinate is formed in large part setting free colloidal silica. When the glue dries
the composition becomes hard and strong.
The strongest and most insoluble glue made from casein is that developed for
glueing the laminations of plane propellors and now used for veneering wood. It is
insoluble when set and is not even loosened by steam. The principal requirements
of casein to make it desirable for glue making are (1) that the percentage of impurities
such as acid, moisture (not over 8 per cent) and fat (not over 1.5 per cent) must be
small; (2) freedom from sour odors; (3) clearness and uniformity of color; and, (4)
fineness of particles (less than 50 mesh). By grinding a casein to fine particles a
minimum of time is required in the mixing and a more uniform glue is obtained.
The ash should not exceed 3 per cent for natural sour, or 4.5 per cent for acid
precipitated casein.
Manufacture of Casein Glue.1—To produce glue, casein is mixed with other
ingredients, the chief of which are lime and water. These three constituents
alone will give a glue of good water-resistant properties but of short life. Other
ingredients are, therefore, added to increase the working life and the water resist¬
ance and improve the other qualities of the glue. Sodium hydroxide, sodium
fluoride (patented) and sodium silicate are used in various formulas. They all
lengthen the life of the glue and sodium fluoride probably gives the glue antiseptic
properties. Still other ingredients are added to give the glue some particularly
desired property. Oils are usually added to the dry-mix glue to prevent dusting
in handling the glue. Formulas vary therefore, not only with reference to
materials used but as to the proportions of different ingredients.
Several formulas for mixing casein glues have been developed at the Forest
Products Laboratory. One of the best of these, with a discussion of the method of
mixing and a description of materials required, is as follows:
1 Taken from formulas devised by Forest Products Laboratory, U. S. Dept, of Agriculture.

Glue No. 4-B1
Formula
100 parts casein
soak 15 min.
150 to 250 parts water
mix
15 to 22 parts hydrated powdered (high calcium) lime
mix mix
90 parts water
70 parts silicate of soda (40°B6.)
Method of Mixing.—The proper quantity of water is introduced into the glue pot
and the mixing blade is brought into action at a speed corresponding to about 50 or
60 r.p.m. The stirring is allowed to continue during the addition of the casein to the
water and for a few minutes thereafter until the mixture becomes mush-like in con¬
sistency, through the absorption of the free water by the casein; the blade is then
stopped and the mixture allowed to soak.
After a period of 15 min. the soaking is considered complete and the mixing blade
is again brought into action. The lime-water mixture is now added and 2 or 3 min.
later the liquid silicate of soda is introduced. The mixing is allowed to continue for
20 min. to ^ hr. after the addition of the silicate of soda, whereupon a smooth freely
flowing mixture of uniform texture and free from lumps should be obtained.2
Usually actual experience in the mixing of this glue is necessary before satisfac¬
tory results can be expected, and it is advisable to have a new operator witness an
actual demonstration. This is due to the fact that no precise quantity of water can
be prescribed because of the variation in the water-absorbing qualities of different
caseins. The criterion of whether or not the proper quantity of soaking water has
been added is the viscosity of the finished (mixed) glue. If its consistency is too thin
an excess of water beyond that required has been used, and it is best to reject the
batch and try again. Similarly, if the consistency is too thick and heavy an insuffi¬
cient quantity of water has been used. The water required for various types of
casein lies in the following ranges:
Lactic-acid casein. 130 to 170 parts water per 100 parts casein,
Sulphuric acid casein j
to 220 parts water per 100 parts casein,
Hydrochloric acid casein / * ’' *
Hennet casein. 280 parts water per 100 parts casein.
Galilith is the best-known horn or celluloid-like substance obtained by the

action of formaldehyde on casein (from milk). The casein is dissolved in alkalis
(such as caustic soda or aqueous ammonia), and the clarified solution is precipi¬
tated with acids, filtered, pressed, and dried slowly during some weeks or months.
The plates so obtained, when soaked in formaldehyde and dried, form a yellowish-
white translucent product, very tough like horn or ivory.
Casein Compositions Direct from Milk.—For many purposes, it is preferable
to make casein as wanted direct from separator skimmed milk, as the homogeneity
of the resulting colloid is greater. One hundred pounds of skimmed milk pro¬
duces about 334 lb. of dried casein. The milk may be curdled with muriatic acid
or rennet. The latter method is preferable for making casein for Galilith.
1 U. S. Patent No. 1,291,396 on this process has been granted to Samuel Butterman of the Forest
Products Laboratory and assigned to the United States Government
2 The old household method of mending broken crockery by using the white of an egg mixed with
lime is akin to this process. The cement produced in this way is a good one, but sets quickly. Good
casein glues are now placed on the market by the Casein Manufacturing Co. and Certus Cold Glue Co.—
Editor.
CEMENTS 855
The coagulation with rennet is effected at about 100 °F. but with acid at 110 to
120°F. In both cases the coagulum after preliminary straining is heated to about
160°F. and strained and re-strained, or pressed. If diluents to make the goods opaque
are put in they are best added before the coagulation as the resulting colloid in such
cases becomes more homogeneous.
If transparent goods are aimed at, or very strong compositions are requisite, the
following method, described at greater length in Scherer,1 is to be recommended, as
it eliminates the fat. Dissolve the casein (rennet) in sodium hydroxide, and then
keep adding quite strong sodium hydroxide until all the casein is precipitated as
sodium caseinate; it is then filtered off from the excess alkali containing the fat in a
thin emulsified form. The precipitated sodium casemate is dissolved in water, pre¬
cipitated with the exact amount of hydrochloric acid necessary, and well washed.
It is dried to the desired content of water (say 10 to 15 per cent) at low heat, pressed
and treated with formaldehyde vapors for a considerable time to render insoluble.
The relatively large amount of alkali in the filtrate can be recovered by causticising
with lime, to be used over again.
The following recipes are often very servicable to technical men: For cementing
glass use 10 parts of gelatine in water mixed with a saturated solution of 2 parts of
potassium bichromate. These two solutions should be mixed together and used in
a subdued light. The compound forms an insoluble mixture quickly in the light.
The so-called “cap cement, ” used by electrical instrument makers to fasten metal
fittings to glass, is made as follows: Put 11 oz. of glue to soak over night in enough
water to cover it. Next day add it and 1 oz. of white curd soap to enough water to
make the total water 2 qt., the water being previously heated to boiling. To the,
take some of the above mixture and add enough plaster-of-paris to make a smooth
paste. Use quickly, for it soon hardens.
The well-known “universal wax,” so useful in physical laboratories, is made by
melting together beeswax, 4 parts; turpentine, 2 parts; then adding about 1 part
Venetian red and working the latter in. The Venetian red may be varied somewhat
to give differing consistencies. Cloth may be stuck to iron rolls using glue to which
enough tannin has been added to give a ropy consistency. This must be used
quickly. For sticking paper to metal use a solution of 2 parts gelatine and 1 of starch
in the least quantity of water necessary to dissolve them.
1 “Das Kaaein” by Robert Scherer, Vienna and Liepsig, 1905.
SECTION XXVI
RADIUM AND RADIOACTIVITY
, By Richard B. Moore1
Discovery of Radioactivity.—The property of radioactivity was discovered

partly by accident, by Henri Becquerel, the French physicist, in 1896. He was
experimenting with certain fluorescent substances in order to find, if possible, a
connection between fluorescence and the recently discovered 3-rays. Among
other chemicals which possess the property of fluorescence, he used salts of uranium.
His custom was to expose the fluorescent substance to the action of the sunlight,
and then register the effect of possible penetrating radiation on a photographic
plate protected from ordinary light. BecquereFs experiments gave positive
results at once, and he at first believed that he could discover a relation
between fluorescence and 3-rays. Later, he exposed a plate to uranium salts
which had not been previously exposed to sunlight. To his surprise, on develop¬
ing this plate he found that he had obtained the same effect as he had previously
secured when the uranium salts had been exposed. The following up of this
partly accidental discovery has given us our whole science of radioactivity.
It was found that radium and its salts had the property of ionizing gases, or con¬
verting the molecules of gases into charged particles. It is this property which is
used almost exclusively in identifying radioactive substances, and in making
quantitative determinations.
Madame Curie’s Work.—Madame Curie, wife of Pierre Curie, after¬

ward Professor of Physics in the Sorbonne in Paris, became interested in the
work of Professor Becquerel, and examined all the known elements to see whether
any of them possessed properties similar to those of uranium. She found that
thorium and its salts would also affect a photographic plate without previous
exposure to sunlight, and would also ionize gases. Thorium as well as uranium
is therefore, radioactive.
Madame Curie found that radioactivity was an atomic property. A given weight
of uranium metal had the same activity, no matter whether it was combined with
chlorine, bromine, the (S04) radical or the (N03) radical. The other elements in
combination with the uranium did not affect the activity in any way. She then made
a study of radioactive minerals, paying especial attention to pitchblende, which is a
natural uranium oxide containing traces of lead, arsenic, bismuth, and other impurities.
This was furnished her by the Austrian Government from its mines at St. Joachimsthal.
To her surprise, she found that a piece of pitchblende carrying a given weight of
uranium had approximately four times greater activity than any pure uranium
salt containing the same weight of uranium. This indicated that either her original
conclusion that radioactivity is an atomic property was wrong, or the pitchblende
contained another element, or elements, which were also radioactive. She proceeded
to test these conclusions and was assisted by the Austrian Government, which sent
1 Chief Chemist, Bureau of Mines, Washington, D. C.
857
her a considerable amount of pitchblende for this purpose. The mineral was dissolved
and the different groups of elements were successively precipitated, each precipitate
in turn being tested for its radioactivity. The lead group was found to be slightly
active, and we know now that this was due to the presence of radium D, or radioactive
lead. The copper group was also active, due to the presence of polonium, which in
many of its properties, is allied to bismuth. The activity associated with the iron
group was due to actinium, which is allied to some of the rare earths. The majority
of the radioactivity, however, was found to be concentrated in the barium, strontium,
and calcium group. The separation of the small amount of highly active material
found associated with these elements was difli ult and tedious. On separating the
calcium and strontium, the activity still remained with the barium,* and the element,
radium, was finally separated by fractional crystallization of its salts, cither chloride
or bromide, from the corresponding barium salts.
Other Workers in Radioactivity.—Madame Curie deserves great credit for

the discovery of radium, and for a great deal of other scientific work she has
carried out with marked success in connection with this element. But those who
are not versed in the subject are likely to forget to some extent what has been
accomplished by those who entered the field of radioactivity after the discovery
of radium. Among those, Sir Ernest Rutherford, Sir J. J. Thomson and W. II.
Bragg, stand out prominently on the physical side, while Sir William Ramsay
and Professor Soddy probably lead on the chemical side of the subject.
Properties of the Rays.—Radioactive substances will affect a photographic
plate and will ionize gases. This is due to the fact that radium and its salts give
off three types of rays, called the alpha, beta, and gamma. The alpha rays travel
with a velocity of about 20,000 miles per second and are positively charged. It
was early found by Rutherford that their mass was comparable to that of a
helium atom; and he definitely made the statement, before the proof was actually
obtained, that the alpha particle was a helium atom with two positive charges
on it. This was afterward proved by Sir William Ramsay and Profonsor Soddy,
who dissolved some radium chloride in water and allowed the occluded gases
to run into a spectrum tube which had previously been evacuated. On allowing
these gases to stand for a day or two, the spectrum of helium gradually appeared.
The beta rays consist of negatively charged electrons, with a mass of about
Heoo of a hydrogen atom. During radioactive changes, they arc ejected with a
velocity of from 100,000 to 186,000 miles per second. While the alpha particle in
stopped by an ordinary sheet of note-paper, the beta particle will penetrate a thin
piece of glass, but is completely stopped by a millimeter thickness of lead. All
evidence points to the fact that the beta particle is similar in its properties to the
electron found in a Crookes tube, and gives rise to cathode rays.
The gamma rays are not material in character but are vibrations of very short
wave length in the ether. Just as the arrays are formed in an x-ray tube by the
stoppage of the cathode rays by impinging on the target, so the gamma rays are
formed during radioactive changes when such changes give rise to beta rays, and
these rays are expelled from the atom with the velocity of light. It is evident, there¬
fore, that the gamma rays are practically identical with the x-rays, except that, they
are of shorter wave length and penetrate matter to a much greater extent.
Anything, therefore, is radioactive which is undergoing atomic disintegration and
giving off alpha, beta, or gamma rays, or all of them. All radioactive changes are
accompanied by at least one of these rays. Their emission is* due to the explosion of
RADIUM AND RADIOACTIVITY 859
the radioactive atom, such explosion taking place at a definite rate, so that in the
case of radium itself one-half is completely transformed in 1,580 years. In the second
1,580 years, half of what is left will have been changed. In the third period of 1,580
years, half of what is left at the end of the second period will have been changed—and
so on. The 1580-year period is called the half-life or half-value period of radium; and
it can be readily seen that in 10 times the half-life period, less than 0.1 percent of the
element will remain unchanged.
Disintegration Series.—The manner in which radioactive elements change is
shown in Table 1, giving the uranium series, and Table 2, the thorium series.
Uranium 1, changes into uranium Xi with the elimination of alpha rays; uranium
Xi changes into uranium X2 with the elimination of beta rays; uranium X2
changes into uranium 2 with the elimination of both beta and gamma rays-r-
and so on down the list. It is thus plainly seen that the metal uranium is the
parent of radium and, indeed, of all the radioactive elements which are found in
any uranium mineral, and are shown in Table 1. Indeed, all of the elements
of Table 1 are found in any uranium ore, and most of them can actually be sepa¬
rated chemically, and their physical and chemical properties identified.
It has already been stated that an alpha particle is a helium atom which has an
atomic weight 4. Theoretically, therefore, whenever a radioactive atom explodes,
with the elimination of an alpha particle, the resulting atom, left behind after the
expulsion of the alpha particle, should have the atomic weight of the original atom
minus 4, the atomic weight of the expelled helium atom. The atomic weight of
radium has been determined experimentally as 226. The radium atom, during its
change, loses an alpha particle with atomic weight 4 and, therefore, the residual
radium emanation atom will have an atomic weight 222.
By examining the fourth column of Table 1, it is seen that whenever a change
occurs involving an alpha particle, the atomic weight of the resulting element if
reduced by 4. As the beta particle is an electron, it has not sufficient mass to affect
the resulting atomic weight.
Table 1.—The Uranium Radioactive Series1
Uranium series Half-value period Rays Atomic weight
Uranium 1. 4.5 by 10° years alpha 238

Uranium Xi. 23.8 days beta 234
Uranium X2. 1.15 min. beta and gamma 234
Uranium 2. 2 by 106 years alpha 234
Ionium. 9 X 104 years alpha 230
Radium. 1,580 years alpha and slow beta 226
Radium emanation.... 3.86 days alpha 222
Radium A. 3.0 min. alpha 218
Radium B. 26.8 min. beta and gamma 214
Radium C. 19.5 min. alpha, beta and
gamma 214
Radium D. 16.5 years beta and gamma 210
Radium E. 5.0 days beta 210
Radium F. 140 days alpha 210
Radium G (lead'.... stable 206
1 The branches of Tables 1 and 2 are omitted for the sake of simplicity.
Table 2.—The Thorium Radioactive Series
Thorium series | Half-value period Rays Atomic weight
Thorium.. 2.2 by 1010 years alpha 232

Mesothorium 1. 5.5 years beta 228
Mesothorium 2. 6.2 hr. beta and gamma 228
Radiothorium. 2 years alpha 228
Thorium X. 3.65 days alpha 224
Thorium emanation.. 54 sec. alpha 220
Thorium A. 0.14 sec. alpha 216
Thorium B. 10.6 hr. beta and gamma 212
Thorium C. 60.8 min. alpha and beta 212
Thorium Di. 3.1 min. beta and gamma 208
Thorium D2 (lead)... stable 208
Experimental Proof.—There is some definite experimental proof that the

above statements are correct. Sir William Ramsay and Professor Soddy actually
determined the density of the radium emanation, and the figure obtained as a
mean of five determinations was 223. In addition, lead is always found in uran¬
ium minerals, and the atomic weight of radium G, or the final disintegration
product, according to theory, is 206. This does not coincide with the atomic
weight of ordinary lead, which is 207; but some experimental work on the atomic
weight of lead obtained from uranium and thorium minerals, by Prof. T. W.
Richards, O. Honigschmid and Professor Soddy, has shown that the atomic
weight of uranium lead is 206, while that of thorium lead is 208—an exceedingly
interesting experimental confirmation of the theory. We have, therefore,
actually three forms of lead, the only difference among them being their atomic
weights all having exactly the same chemical and physical properties. If they
were mixed by fusion, no known method could separate them; and the only way
of telling one from another would be by making an actual atomic-weight
determination.
The Actinium Series.—There is another radioactive series called the actinium
series, which is very similar to the uranium series, but of less importance.
Table 3.—The Actinium Radioactive Series
Actinium series Half-value period Rays J Atomic weight
Protactinium 1,200 to 180,000 alpha Not known

years
Actinium. 20 years (?) Not known
Radio-actinium. 18.8 days alpha and beta Not known
Actinium X. 11.2 days alpha Not known
Actinium emanation. 5.6 sec. alpha Not known
Actinium A. 0.0029 sec. alpha Not known
Actinium B. 52.1 min. beta Not known
Actinium C. 2.1 min. alpha Not known
Actinium C2(?). (?) alpha Not known
Actinium D. 4.76 min. beta and gamma Not known
Actinium E.
A'mVn/-vnm\
The origin of actinium is not absolutely certain, but it and its disintegration
products are supposed to be formed from uranium as a branch chain probably
starting with Uranium 2. It is always found in uranium minerals.
RADIUM ORE DEPOSITS
Commercial Ores.™The two principal commercial ores of radium are pitch¬

blende and carnot ite. The former mineral has no definite composition, consisting
of uranium oxides (UOa, U02) with oxides of lead, calcium, iron, bismuth,
manganese, copper, silicon, aluminum, and rare earths. Carnotite has a more
definite composition, being a potassium uranyl vanadate containing small
quantities of barium and calcium. The formula K A) • 2 IT ()a. V 20 6.3112()
expresses its composition fairly well, although not exactly. Of lesser importance
are aufunite, a hydrated calcium uranium phosphate, Ca(UOt>) 2(1*002.8HA), and
torbernite, a hydrated copper uranium phosphate, Cu(U(>2)2 (POO 2-811-A).
St. Joachimsthal. -The pitchblende deposit at St. Joachimsthal, Austria1—*
is in mica schist interbedded with lime schist and crystalline limestone. Toward
the east and northeast the formation is gneiss. The gneiss was intruded
by quartz porphyry subsequent to the deposition of the vein material. In
the mica schist are fissures filled with volcanic material which cut the mineral¬
ized zone at various points and depths. The veins are usually 0 in. to 2 ft.
wide, in rare cases widening out to II ft. The mode of mineralization varies
greatly, the ores occurring in both stringers and pockets. They contain the
following metals: Silver, metallic and as argentite, polybasite, tetrahedrite, etc.;
nickel, as nickelin, ehlonnthite, etc.; cobalt as smaltite, bismutosmaltite, etc.;
bismuth, as metallic bismuth, bismite, etc.; arsenic, as metallic arsenic and
arsenopyrite; and uranium, as pitchblende and other alteration products. (lal-
enite, zinc blende, pyrite, marcasite, and copper occur in minor quantities.
The mines at Joachimsthal have been worked since 1517. In 15*15 the
production of silver ores declined considerably, but since then the deposits
have been mined for bismuth and cobalt. During the last 25 years the mines
have been worked for uranium.
Saxony. In the vicinity of Annaberg, on the Saxon side of the Erzge¬
birge, the silver-cobalt veins resemble those at Joachimsthal. At Johanngeorgen-
stadt, the veins contain tin and silver-cobalt ores. Where dolomite spar is
found, the silver-cobalt ores contain pitchblende, as at Annaberg. In the
(loltessegen mine t he pitchblende occurs in the spar in pieces 2 to 3 in. in diameter.
These mines are worked principally for bismuth ocher, but also for cobalt and
nickel
In the cobalt-bismuth mines of Helmeeberg are found bismutite and various
minerals of nickel, silver, and arsenic. There is also some pitchblende, urano-
ehaleite, uranospinite, galenite, zinc blende etc.
Cornwall. - Pitchblende has also been found in Cornwall, England in
the tin region. As at Joachimsthal and Johunngeorgenstadt, the mineral is found
associated with nickel-cobalt veins, although only part of the veins are highly
1 Ku’Hako Bk<*k: “Lidsro von dim Krxlagmtaotton," 3d, Kd.( Vol. 1, pp. 4OH 410. BrrHn, Burnt”
rnogor, itaox: “Thr Nuturo of Oro I)<«ponit«." Translation by W. II. Wood. lit. Kd.»
Vol.1, pp. 2H4 2H7. Now York and London, Eng. and Min. Jour., 1005.
argentiferous. Even though these veins are closely connected with the tin
veins, they apparently are not of the same age as the latter, but belong to the
same general period of mineralization According to IJssher, Barrow and Mac-
Allis ter, “The Geology of the Country Around Bodmin and St. Austell,n Memoirs
of the Geological Survey of England and Wales (1909), p. 157, the most important
uranium producers are the South Terras mine, the Carharrack, Dolcoath, Wheal
Unity, Wheal Gorland, Wheal Lovell, and Trenwith. The South Terras mine is
situated in the valley of the Fal, southwest of St. Austell. The country rock
is slate, intruded by greenstone and granite porphyry dikes.
Pitchblende in the United States.—Pitchblende has been found in the
following localities in the United States: feldspar quarry, at Middletown,
Conn., in large octahedrons; in Hall’s quarry, at Glastonbury, Branchville, Conn.,
in a pegmatite vein and usually embedded in albite; at Marietta, S.C.; in the
Baringer Hill District, Llano County, Texas; in the Bald Mountain district,
Black Hills, S. Dak.; in Mitchell County, N. C.; and in Gilpin County, Colorado,
The latter district is the only one of commercial importance.
All of the Gilpin County mines, with one exception, are found on or near Quartz
Hill, a few miles from Central City. There are five that have produced pitchblende
in quantity; the Kirk, Wood, German, Belcher, and Calhoun. The Kirk, Belcher
and German mines are close together on Quartz Hill, the Wood and the Calhoun
being in the valley below.
These mines, until recently, have been worked mainly for gold. In this district,
gneiss and crystalline schist predominate, with intrusive andesitic dikes and occasion¬
ally acid granitic dikes. The rock containing the pitchblende, galena, sphalerite,
etc., is a fine grained aplitic granite which probably once contained an appreciable
amount of biotite. The ore deposits are of two general types, one containing pitch¬
blende with pyrite, sphalerite and galena, and sometimes marcasite; the other type
contains pyrite, chalcopyrite, sphalerite and galena, with some gold and silver.
Generally speaking, the two types are not associated, so that the miner has a choice of
mining either for pitchblende, or for gold.
The Kirk mine has probably been the most important producer of the five men¬
tioned although reliable data on the output of pitchblende from this mine, up to a
few years ago, has been difficult to obtain. During the last 12 years, about 20
tons of ore, with an average content of 35 per cent UaOa, and over 100 tons with a
content of 3 to 4 per cent U3Os, have been mined. The mine has been shut down
for some time. More recent operations of the German and Belcher mines produced
120 tons of low-grade ore, averaging about 1 per cent U3Os, and 6 tons of high
grade. Smaller quantities of ore have been produced at various times from these
mines and from the Wood and Calhoun, also the Joe Reynolds mine 5 miles west of
Idaho Springs, Colo.
Australia.—Uranium ores are found in certain localities in Australia. One

of these deposits is 80 miles east of Farina, a railroad station on the Great
Northern line in South Australia, and lies between Mount Painter and Mount
Pitt. H. L. Y. Brown1 states that the rocks of the district consist of coarse and fi ne
feldspathic, siliceous, and micaceous granite, gneiss, quartzite and mica schist.
The rocks are contorted in places and penetrated by dikes of coarse, pink-colored
eruptive granite.
1 Occurence of Uranium Ore and other Rare Minerals near Mount Painter, in the Flinders Range
of South Australia.” South Australia, Mines Dept. 1911.
‘Two of the prospect pits are on outcrops of iron oxide with cellular quartz and
gossan, the whole having the appearance of an irregular lode. The uranium minerals,
torbernite and autunite, are disseminated through the ore, and are also crystallized
on the walls of the fissures and cavities in it. Uranophane and gummite occur
sparingly; fergusonite and some monazite are also present. Another uranium deposit
lies southeast of the one just described, about 20 miles southeast of Olary, on the
railroad line from Petersburg to Broken Hill, South Australia. The ore occurs as a
yellow and greenish-yellow incrustation and powder on the faces, joints and cavities of
a lode, which consists of magnetite, titaniferous magnetite, etc., and quartz in asso¬
ciation with black mica.
Portugal.—Autunite is found in commercial quantities in Portugal in the

district between the towns of Guarda and Sabugal. An excellent description is
given by Segaud and Humery.1 Apart from the uranium, the rocks of the region
are much mineralized, showing deposits of galena, arsenopyrite, chalcopyrite,
tungsten and cassiterite.
Camotite Deposits of Colorado and Utah.—All of the deposits referred
to above are important and have been developed as commercial source
of radium; in fact, until about 7 years ago, they were the only sources from which
radium ores were obtained. About 1910, the carnotite deposits of southwestern
Colorado and eastern Utah began to receive attention. They were known as far
back as 1881, but the composition of the ore was unknown until 1897. In 1899,
an analysis showed that the ore not only contained uranium, but was a potassium-
uranyl vanadate.
In 1900, a small plant was erected in the McIntyre district, south of the Paradox
Valley, Colo., for the extraction and recovery of uranium oxide from carnotite ore.
Only moderate success was achieved, and the plant was shortly shut down. Opera¬
tions were also started by other concerns, notably the Western Refining Co. and the
Dolores Refining Co.; these plants extracted uranium and vanadium only. None of
these operations was of importance, and it was not until 1909 to 1910 that any interest
was shown in the carnotite deposits as a source of radium. At that time, the ore was
almost exclusively shipped abroad.
In the fall of 1912, representatives of the U. S. Bureau of Mines made a thorough
survey of the carnotite fields and announced the fact2 that the camotite deposits of
Colorado and Utah constituted by far the largest source of radium-bearing ores in the
world. Developments since that time have proved this statement to be correct, as
the larger part of the radium that has been produced in the world has been derived
from American carnotite ore.
The deposits are found mainly in Dolores, San Miguel, and Montrose Counties,
Colorado, and extend over a belt about 60 miles long by 20 miles wide. The ore is also
found to the west of the La Sal Mountains, in Utah, and along the San Rafael Swell,
southwest of Green River, Utah. Small patches of ore are found scattered between
these points and extend as far north as Meeker, Colorado.
The deposits are invariably pockets, many of which however, are of considerable
size. A few individual claims have produced as high as 500 tons of shipping ore,
which, however, is exceptional. The ore is found in a light-colored sandstone overlain
1 Segaud et Humery: “Les Gisements d’Uranium du Portugal.” Annales des Mines, Memoires, Ser.
11 (1913), 3, 111-118.
2 Richard B. Moorb and Karl L. Kithil: “A Preliminary Report on Uranium, Radium and Vana-
in places with, shale and conglomerate; according to Hillebrand and Ransome, this is
the McElmo formation.11
Amount of Radium in Ore.—Most of the radium obtained at the present time, as
has already been stated, is obtained from carnotite ore. The average commercial
grade of this ore contains from 5 to 10 milligrams of radium element per ton of ore.
Some high-grade pitchblende ores may carry as much as 50 to 60 milligrams of radium
per ton, but this is unusual. Allowing for losses in extraction and recovery, it takes
about 5,000 to 6,000 tons of ore to give an ounce of radium element. It can be seen
therefore, that the metallurgical processes involved in the extraction of radium are
entirely different from those connected with any other element. There are two general
steps which must be carried out; first, to obtain a radium concentrate from the ore;
and second, the retreatment of this concentrate in order to extract a high-grade
product. The first step is necessarily carried out on a large scale, while the major
part of the second step is done in the laboratory.
Metallurgical Treatment.—Pitchblende was the first ore treated for the
extraction of radium. This was done by Madame Curie who obtained her
ore through the Austrian Government, from the St. Joachimsthal deposits. The
original method used was to fuse the ore with sodium sulphate. This converted
the uranium into sodium uranate. On washing with water the excess sodium
sulphate and other soluble products were separated leaving the sodium uranate in
the residue with most of the other heavy metals, sillica etc. The residue was
then treated with dilute sulphuric acid which dissolved the sodium uranate and
some heavy metals, leaving a residue which consisted very largely of silica,
alumina, etc. The radium remained in this residue as radium sulphate. The
acid solution carrying the uranium was treated for the recovery of this element
while the residue carrying the radium was boiled with a solution of sodium
carbonate. This treatment converted a portion of the radium sulphate into
radium carbonate and sodium sulphate. The insoluble product was thoroughly
washed free from sodium sulphate, and the residue treated with dilute hydro¬
chloric acid. This dissolved the radium carbonate. The process of the boiling
with sodium carbonate was repeated several times until all the radium sulphate
was converted into radium carbonate and finally into radium chloride by solution
in the hydrochloric acid. The mixed liquors were then added together and a
small amount of barium chloride and sodium sulphate added in order to precipi¬
tate in the solution a definite amount of barium sulphate. The liquor was kept
thoroughly agitated during the precipitation and practically all of the radium was
precipitated with the barium sulphate. This product constituted a high-grade
radium concentrate, and was refined by methods which mil be described later on.
In many ways still better results can be obtained by fusing the original ore with
sodium carbonate instead of with sodium sulphate. After thoroughly washing,
the residue, consisting of metallic carbonates and hydroxides, is dissolved in dilute
chemically pure hydrochloric acid, and from this a high-grade radium-barium
sulphate can be directly precipitated. If the ore is high-grade and largely
free from pyrites it can be treated directly with boiling concentrated sulphuric
acid. Radium sulphate is moderately soluble in hot concentrated sulphuric
acid, and on diluting to a concentration of less than 10 per cent free acid the
radium is precipitated as sulphate along with any barium and traces of calcium,
1W. F. Hillebrand and F. L. Ransome: “On Carnotite and Associated Vanadiferous Minerals
in Western Colorado.” U. S. Geological Survey Bull. No. 262 (1905), 11.
etc., that may be present. This represents a crude concentrate which can be
further refined. The American pitchblende ores cannot be handled in this manner
because they contain a considerable amount of pyrites. The pyrite crystals have
small particles of pitchblende embedded in them, and as the pyrite is not readily
soluble in concentrated sulphuric acid, the radium recovery is poor. For the
same reason fusion with sodium sulphate or acid sodium sulphate gives also a
poor recovery. The American pitchblende ore is best treated by fusion with
sodium carbonate or by the method used by the U. S. Bureau of Mines, viz.—
boiling the ore with strong nitric acid until all the pyrite is decomposed, diluting,
filtering and then treating the insoluble residue containing the radium with
boiling concentrated sulphuric acid in the manner above described.
Camotite.—The different methods of treating carnotite ore to obtain a
concentrate may be classed under three general heads; (1) an alkaline leach
followed by an acid leach; (2) fusing the ore with some material that will dis¬
integrate it and make the extraction of the valuable contents possible; (3) an
acid leach.
Alkaline Leach Methods.—It is probable that some of the early experi¬
mental work on extracting radium from camotite ore involved boiling the
ore with a solution of sodium carbonate, thus getting rid of the uranium and
vanadium, which go into solution. Since radium has properties similar to those
of barium, any radium in the ore would be converted into radium carbonate, and
on treating the residue with dilute chemically pure hydrochloric acid, the radium
would be dissolved with any other acid-soluble products.
The radium concentrate always obtained is radium-barium sulphate. By adding
barium chloride and sulphuric acid, or sodium sulphate, to the slightly acid solution
carrying the radium, barium sulphate is formed in the solution and drags down
radium sulphate with it. The radium is almost always precipitated in this manner
in a liquor sufficient in volume to hold it actually in solution. Undoubtedly adsorp¬
tion has something to do with the precipitation of the radium along with the barium
sulphate, but this does not fully explain the small losses that accompany such precipi¬
tation. The term ‘‘pseudo-isotopy” has been given to this property by Dr. S. C.
Lind.1
The general principles outlined above are included in the Haynes-Engle process,
which involves boiling the ore with an alkaline carbonate solution (see p. 864). The
object of this process was to recover uranium and vanadium only, and did not attempt
to obtain the radium in any form. A patent taken out by Warren F. Bleecker involved
the extra step which the Haynes-Engle process did not cover, namely, the leaching of
the residues with hydrochloric acid in order to obtain the radium in solution, after the
ore had been boiled with an alkaline-carbonate solution.
The alkaline-leach method has many advantages and some disadvantages. It
separates the uranium and vanadium from the ore during the first stage of the process.
It eliminates sulphates by converting the metallic sulphates in the ore into metallic
carbonates and soluble sulphates, which go into the filtrate with the uranium and
vanadium. The radium, therefore, is left behind in the residue as carbonate, prac¬
tically free from sulphates. This prevents the re-precipitation of the radium as sul¬
phate, on treating with acid, until after the acid solution is filtered from the tailings.
On the other hand, it has the following disadvantages: It converts some of the silica
in the ore into sodium silicate, which makes filtration very difficult—in fact, most of
1 S. C. Lind, J. E. Underwood, and C. F. Whittbmore: “The Solubility of Pure Radium Sulphate.*'
,
Jour. Am. Chem. Soc., (March, 1918) Vol. 40 pp. 465-472.
55
the filtering and washing has to be done by decantation. It is difficult to treat con¬
centrates by this process, as these are almost invariably of very fine mesh,which
adds to the filtration difficulties. The treatment with alkaline carbonate converts
most of the iron and a good part of the aluminum in the ore into an acid-soluble form,
so that the acid consumption is high. The method, however, can be used with success
for the treatment of certain uranium ores, particularly carnotite, autunite, and
torbernite.
Fusion Methods.—A treatment somewhat similar to the Austrian method
for treating pitchblende, is given to the ores from Olary, South Australia, by
the Radcliffe process. The main uranium mineral is carnotite, but this is asso¬
ciated with considerable quantities of ilsemannite and other rare-earth minerals
that are not found in American carnotite. The concentrates are mixed with
three times their weight of salt-cake, and fused in a reverberatory furnace;
the fused product is then crushed and agitated in wooden vats, with water.
By suitable adjustments it is possible to separate on the bottom of the vats a
considerable amount of comparatively coarse material that is almost free from
radium and uranium. The turbid overflow carries in suspension the radium, lead
and barium, as sulphates, together with a considerable amount of finely divided
silica. The overflow is pumped to large lead-lined tanks and allowed to stand all
night. This is nothing but a sliming process and has the advantage that the
radium in the form of sulphate always remains with the fine material. The
slimes settle completely in 12 hr. and are collected periodically and treated for
the recovery of radium.
The process of Schlundt, which is being used by one company in the United States
in connection with carnotite, is very similar to the Radcliffe process. The ore is
fused with acid sodium sulphate, leached and washed with water to extract the uran¬
ium, vanadium, and other soluble products, and the residue is slimed in order to obtain
a crude concentration of the radium, which stays with the fine material.
• The U. S. Bureau of Mines has found that if a radium-barium sulphate high in
silica is fused with caustic soda containing a small quantity of sodium carbonate,
the silica can be easily washed out as sodium silicate, while the barium and radium
remain behind as radium-barium carbonates, which can be readily dissolved in hydro¬
chloric acid. Usually the commercial caustic soda contains enough carbonate with¬
out the addition of any extra amount. This method has been applied commercially
to the treatment of crude concentrates, such as are obtained by the Schlundt process.
One firm has used a sodium-carbonate fusion. The ore is fused with soda ash,
usually about two and one-half times the weight of the ore being required. This is
done in a reverberatory furnace, lined with magnesite brick, and the fused mass is
run directly into vats, in which it is leached. The silica is thus converted into sodium
silicate and so passes into solution, together with the uranium and vanadium. The
iron, calcium, barium, radium, etc., remain as the insoluble residue, which is washed
in filter presses. The material is then treated with dilute sulphate-free hydrochloric
acid, which dissolves the carbonates, and the radium and barium are precipitated by
the addition of the requisite amount of sulphuric acid, or sodium sulphate. The
whole is allowed to stand in settling tanks and the clear liquid is drawn off, leaving
the barium-radium sulphates mixed with a considerable amount of silica and other
impurities, as a sludge at the bottom of the tank. This is taken out without previous
filtration, and dried, forming a crude radium-barium sulphate which is then refined.
The main advantages of this method are that it will treat an ore containing con¬
siderable quantities of sulphates, as the process removes the sulphates as soluble
sodium sulphate in a manner similar to the alkaline leach method. In addition it is
applicable to the treatment of concentrates, as fineness of material is really an advan¬

tage instead of a disadvantage, as it does not involve any tlsliming.” The main
disadvantage is the cost, both for the chemicals and the labor required. The con¬
centrate obtained is also low-grade, which involves additional refining costs.
Acid Leach Methods.—Leaching with hydrochloric acid gives an excellent

extraction, provided the ore is comparatively free from sulphates. As a consid¬
erable proportion of such ores contain traces of gypsum, it is a method which must
be used on selected ores. A hydrochloric acid leach has been used successfully
on Cornish pitchblende, which is practically free from pyrite, but is not applicable
in any way to American pitchblende, which contains considerable quantities
of pyrite. It has also been used on Portugese autunite. In general, the hydro¬
chloric method has a limited use.
Leaching with nitric acid has been used more successfully. This is the method
originated by the U. S. Bureau of Mines1 and used in the plant of the National Radium
Institute. As is generally known, barium sulphate is much more soluble in nitric
acid than in hydrochloric acid, and this is especially true when the nitric acid is con¬
centrated and hot. For this reason, ores containing sulphates can be treated by a
nitric acid leach, and the method has been successfully applied to ores carrying as
much as 1 per cent gypsum. On ores carrying small quantities of sulphates, the ex¬
traction obtained is very high, frequently going, on a commercial scale, to 93 and even
to 95 per cent. Boiling, 40-per cent nitric acid is used, and filtration is obtained on
either a vacuum or a pressure filter. Filtration must be done rapidly, because as the
acid cools, the radium has a tendency to precipitate out, especially in the presence of
sulphates. During the process of recovering uranium and vanadium by this method,
practically nothing but sodium hydroxide and sodium carbonate are added to the
nitric acid. Consequently, sodium nitrate is obtained as the final product after the
precipitation of the tradum, uranium and vanadium, and this sodium nitrate is
recovered by evaporation and crystallization and used over again for the manufacture
of nitric acid in the ordinary way. The average losses of nitric acid were about 15
per cent, so that 85 per cent of the acid used was recovered as sodium nitrate. This
actually reduced the cost of the nitric acid below that of hydrochloric, and this, to¬
gether with the high extraction of the radium obtained, was largely responsible for
the low cosl of the recovered radium and the success of the method.
The process cannot treat successfully ores carrying as large quantities of gypsum
as canbe treated by some of the other methods, such as the sodium carbonate fusion,
but it is applicable to a very large percentage of the carnotite ore produced. Diffi¬
culty was experienced in treating concentrates, most of which are below 150 mesh.
As already stated, it has been necessary to filter rapidly in order to get a good extrac¬
tion; and the problem of filtering a 40-per cent boiling nitric solution, containing fine
material in suspension, was found to be difficult. It was finally overcome by the
development and use of a pressure filter instead of a suction filter; this consisted of a
steel shell containing an earthenware filter set in concrete. The earthenware filter
had at its bottom, as filtering medium, a plate of either filtros or alundum. The steel
top was made tight with swing bolts, and a pressure of 100 lb. was applied.
Refilling.—After the radium-barium sulphate is obtained, it is necessary

to change this insoluble product into soluble form so that the radium can be
fractionated from the barium. The old method of converting the insoluble
sulphates into soluble chlorides consisted in boiling the radium-barium sulphate
1 Charles L. Parsons, R. B. Moore, S. C. Lind, and O. C. Schaefer: “Extraction and Recovery
of Radium, Uranium and Vanadium from Carnotite,” U. S. Bureau of Mines Bull. No. 104 (1915L
g6g CHEMICAL ENGINEERING
with a strong solution of soda ash so that the sulphates were converted into
carbonates. After the first treatment, the insoluble residue which consisted
of a mixture of sulphates and carbonates, was thoroughly washed so as to ice
it from soluble sulphates, and was then leached with dilute chemically puu,
hydrochloric acid. This dissolved out the radium and barium carbonates wine i
went into solution as chlorides, leaving the insoluble sulphates which had not
been converted into carbonates by the first treatment. The process was then
repeated; the sulphates being boiled a second time with soda ash, washed and
leached with hydrochloric acid. Three or four treatments usually served to eon-
vert all of the sulphates into carbonates, which were ultimately obtained m the
form of soluble chlorides by adding hydrochloric acid. AH the acid liquors were
then added together for fractionation purposes. It can readily be seen that this
method involved a considerable amount of time and trouble and added consid¬
erably to the cost of radium production.
Reduction with Coal Gas.—It is a well-known fact that barium sulphate
can be reduced by heating in a current of coal gas. A temperature control
within reasonably narrow limits is necessary in order to get proper reduction,
and at the same time to obtain the material in a non-clinkering condition so
that it can be readily removed from the tube in which it is heated. This tempera¬
ture control is the main objection to this method, as it is difficult to obtain {airly
satisfactory results by heating with coal, oil or gas, and It is not easy to build
an electric tube furnace which is suitable for the reduction. With coal gas,
reduction begins at about 575°C.; the actual working temperature being between
725 and 740°C. The gas consumption averages from 200 to 300 eu. ft. for 25
to 28 lb. of sulphate. The heat of reduction is positive at high temperature,
producing reaction clinkers, some of which are friable and some glassy and hard
to break. With a high-grade sulphate, reasonably free from silica, the reduction
varies from 75 to 90 per cent. If a considerable amount of silica is present,
the reduction is decreased. The barium-radium sulphides obtained are dissolved
in dilute hydrochloric acid and the solution filtered from unreduced sulphate.
During the reaction with the acid, there is a copious evolution of hydrogen
sulphide so that the work must be done outdoors or in a strong draft of air.
The solution of barium-radium chlorides thus obtained is ready for
fractionation.
A more satisfactory reduction of the sulphates to sulphides can be obtained by
using charcoal instead of coal gas. The sulphate is mixed with about one-fifth its
weight of powdered charcoal and heated for 7 or 8 hr. in a graphite crucible in an oil
furnace at about 800°C. Reduction takes place readily, according to the equation:
BaS04 4- 4C =* BaS + 4CO
The crucible is covered with a lid, but a small opening is left at the spout at which
the carbon monoxide bums. When the evolution of gas is complete, the crucible is
allowed to cool with the lid on, and the radium-barium sulphide is removed and broken
up. Under such conditions, with a sulphate reasonably free from silica, it is usual
to get a 90 per cent reduction of the sulphate to sulphide; the ratio of the reduction
of the radium being about the same. The sulphide is then dissolved in dilute hydro-
chloric acid; the whole is thrown on an earthenware suction filter, and the unreduced
sulphate thoroughly washed with distilled water. This unreduced sulphate is either
reduced a second time separately, or mixed with a fresh batch of the sulphate. The
RADIUM AND RADIOACTIVITY' 869
ratio of radium to barium in the unreduced sulphate is about the same as in the orig¬
inal sulphate, but there is a greater concentration of impurities, such as lead. After
these impurities have accumulated, it is more convenient to fuse the final product
obtained with sodium carbonate. During this fusion, the lead accumulates at the
bottom of the melt as metallic lead, and in this way is eliminated. The carbonates
are dissolved in hydrochloric acid.
Plant Crystallization.—The liquor obtained from the above procedure will

probably carry in solution radium and barium chlorides in a ratio of about 1 part
of radium to 1,000,000 of barium. The only efficient method of separation of
these two elements so far discovered is fractional crystallization. Radium
chloride is less soluble than barium chloride and, therefore, will tend to concen¬
trate in the crystals, whereas the barium will concentrate in the mother liquor.
The crystallization factor in hydrochloric acid solution is about 1.5 to 1.6; that
is to say, if 50 per cent of the barium chloride is removed, there will be 50 per
cent more radium in the crystals removed than in the liquor left behind.
The crystallization factor in neutral solution is considerably less, therefore, it is
advantageous to crystallize in an acid solution. Acidproof ware of some kind must
be used; and, as the volumes of liquor obtained from the plant are at first large, it is
necessary to have vessels of considerable size. The most satisfactory vessels to use
are steel or cast-iron pots with acidproof fused silica lining. Steam-jacketed vessels
of 50 to 80 gal. capacity can be obtained of this material, and, if used carefully, will
last a considerable time. The method of crystallization is as follows:
About six evaporators are required. The radium-barium chloride solution in the
first evaporator is concentrated so that on cooling one-half of the barium chloride
crystallizes out. This is richer in radium than the barium chloride in the original
solution. The liquor is siphoned into evaporator No. 2, and the process repeated;
one-half of the remaining barium chloride being allowed to crystallize. This pro¬
cedure can be continued all the way down the series, or a fresh batch of liquor may
be introduced into the first pot after the crystals have been removed. It is advisable
to establish a permanent crystallizing system so that fresh liquor can be introduced
each day into evaporator No. 1 and a batch of crystals richer in radium can be taken
out each day for further refining. The system may be established in several ways:
Either by adding fresh material to the first pot each day and gradually working the
mother liquor down the evaporators to the bottom of the series; or by crystallizing
the first batch of liquor introduced into the first pot all the way down the series before
the introduction of fresh material. Whichever method is used, the procedure after¬
wards adopted is the same. Half of the barium chloride in No. 6 evaporator is crys¬
tallized out and removed from the mother liquor. This liquor should be very low
in its radium content; carrying not more than 25 or 30 parts of radium per billion of
barium. It is discarded. After the crystals have been removed from the evaporator,
the liquor obtained from a similar crystallization in pot No. 5 is siphoned into No. 6,
and the crystals from No. 6 are dissolved in this liquor. In other words, the crystals
from each evaporator ascend and the mother liquor at the same time descends; the
crystals from a given evaporator being dissolved in the liquor from the second evapo¬
rator above. The crystals taken from pot No. 1 can be further concentrated by giving
two or three additional crystallizations before being sent to the laboratory for final
refining.
Laboratory Refining.—The crystals obtained from a plant refining should
contain from 5 to 10 parts of radium per million parts of barium. The salts of
most metals, such as iron, aluminum and vanadium, that may occur with the
870 CUE'S!K'M* ESElXEEEi\E
radium-barium salt as impurities, pus> into the mother liquor>, ami only .small
traces remain with the crystals. Lead is an exception. however. and requires
special treatment. Where barium-radium sulphide, a!'t«*r n-durtiou from sul¬
phate, is dissolved in hydrochloric acid solution with eopmu . «*\i»lution nt* hydm-
gen sulphide, one might expect the almost complete ivuuaai m the lead content
This does not occur, liowever, and some lead always remain . with the crystals
obtained from the plant crystallization system and mud be returned later cm.
The crude barium-radium chloride4 received in»m the plant L devolved in water
in large porcelain dishes, and hydrochloric acid b added to precipitate as much
lead as possible. After the solution has Mood over night, it L tittered to remove
the lead chloride and whatever curiam and barium Milphate m;i\ haw têajaui
previous filtration, All filter papers used mud. of emir-e, be retained and after¬
wards ignited for recovery of tin* radium retained by the paper .. The laboratory
crystallizing system is very similar to that te-ed in the plant. except the volumes
of the dishes are much smaller and there are at least tour wv*eb> »»n t In* minus side
and whatever number is required on tin* pirn side to bring the rehurd eatf tip to
the purity desired. If the crystals from the plant contain nut more than b or t>
parts of radium per million of barium, four eryMalli/atious on the utiuu * ddc are
usually suflieieut to reduce tin* radium low enough so that the tad * ran be put
back into the plant crystallizing system. No. b wêt, therefore, eminting from
the minus end, is the one into which all fresh charges of liquor are placed; and
from this point the radium travels up the plus aide and become» more anti more
concentrated, and the barium travels to the last wml on the minim ride, from
which point it goes back with whatever radium it ‘dill retains to tin4 plant crys¬
tallizing system.
The vessels in which the actual cryntalligation is curried out mav cither be of
silica, porcelain, or steel, lined with acid-proof material. The vr^rb in which the
higher grade material is crystallized should uIwhvh be of qmutz. The procedure is
the same as in the plant crystallizing system; for example, the erv*tub* fioin pot No. 0
arc dissolved in pot No. 7 in the mother liquor obtained from pot No, K, As the fac¬
tor of enrichment of radium chloride from barium chloride and a bo of radium bromide
from barium bromide is more favorable in arid than in neutral aotuiiomi, a fair arid
concentration is maintained throughout the chloride and nlno th bromide system
which is afterwards obtained.
Removal of Lead.~~~The chloride crystals from vessel Xo, *J on the plus side are
dissolved in water without acid, and placed in a large glu*s precipitating jar III or
18 in. high. Make ammoniacal and pass in hydrogen sulphide until the pre¬
cipitation of lead as lead sulphide4 is complete, lids sulphide t« filtered and col¬
lected and may be analyzed fur its radium content It nlioulil In* Mured for
recovery of the radium by fusion with sodium carbonate. Attempts to precipi¬
tate the lead as sulphide even from slightly acid solution timidly result in obtain¬
ing a brick-red precipitate of the formula (PbS), * PbCl3 which is far more soluble
in acid than lead sulphide; hence, complete preeipitation of flu* lead U possible
only in ammoniacal solution. If the solution is umad alkaline with ammonia
before passing in the hydrogen sulphide, only a small amount of radium is
precipitated with the lead, usually not more than Cl/i per rent
Conversion to Bromide.-—-Into the filtrate front the lead precipitate, fiowdered
ammonium carbonate is introduced gradually with vigorous stirring until all
he barium has been precipitated as carbonate. After standing over night for
he settling of the barium carbonate and the thorough precipitation of the radium
arbonate, the supernatant solution is siphoned off as far as possible and the
st is thrown on a Buchner funnel where it is filtered and washed several times
nth distilled water. The filtrate which carries only small traces of radium is
tored and may be returned to the plant at some convenient point before the pre-
'pitation of the radium. The amount of radium in the liquor is surprisingly
)w, usually 0.001 to 0.003 mg. per liter. The barium-radium carbonate is
smoved from the Buchner funnel and is dissolved in chemically pure hydrobromic
cid of 20 to 35 per cent strength in a large glass precipitating jar. Frac-
[onal crystallization is now continued in silica vessels in fairly strong hydro-
romic acid solution in which the separation factor is now 2 instead of 1 % as is the
ase for the chlorides. Crystallization, therefore, takes place more rapidly, but
reater care in handling the vessels is, of course, necessary on account of the
igher radium concentration. Heating may be carried on on tripods with bare
as flames until the richer fractions are reached, when the evaporation is carried
ut on an electrically heated water bath in which only distilled water is used,
a case of an accidental loss of radium solution into the bath, all the water can be
rawn off and returned into the system just before the treatment with hydrogen
llphide.
The evaporation required to obtain a suitable batch of crystals may be generally
sgulated by concentrating the solution until vigorous fanning just begins to cause
le formation of crystals on the surface of the hot solution. Of course, the higher the
fid concentration, the more generous the crystal batch will be, and it is usually
mvenient to have the acid concentration such that about half of the barium in solu-
on will crystallize out. Owing to the high factor of enrichment as bromide, the
idium content of the mother liquor from the “minus two” bromide vessel is extremely
•w. In general, the amount of radium returned to the plant in this bromide mother
luor is only about 0.2 per cent of the total amount going through the system. If
ch a result is actually obtained, it is more convenient and economical to use the
romide mother liquor for the recovery of hydrobromic acid than to return the mother
ûor to the plant to recover its small radium content.
Number of Fractions Employed.—The number of fractions employed in the

Lus direction in the bromide system varies with conditions; 10 to 12 being the
sual number. The crystallization is conducted in such a way that the barium
romide collected in the final fraction should not fall below 1 per cent of radium
romide, and sometimes is as high as 3 or 4 per cent. The total weight of the
action should be 1 to 2 g. This fraction, after thorough drying in a hot-air
pen, is sealed in a glass tube and kept for final purification, when a sufficient
umber of tubes have been accumulated. The amount of radium in the tube
m be easily determined by means of the gamma-ray method, to be described
ter on. After 20 or 30 tubes have accumulated, they are opened and the cou¬
nts dissolved in hydrobromic acid in a small silica dish.
If there is any considerable difference in the activity of the salt in the different
ibes, it is wiser to put those together which have approximately the same activity
r the initial solution, and to put those of higher activity in plus 1, plus 2 or plus
dish, depending upon the amount of radium present. Crystallization is then con-
aued with a general tendency to push up the radium from the lower fractions into
the last silica dish in which the majority of the radium is finally accumulated. After
thorough drying in a hot-air oven to free the salt from moisture and from water of
crystallization, it is placed in a glass tube and hermetically sealed. In this tube a
small platinum wire is sealed through one end in order to conduct away the unipolar
charge that may collect in the interior, attaining voltages that could cause destructive
sparking. Reports are on record of serious radium losses having resulted through
neglect of this precaution.
RADIUM MEASUREMENTS
Measuring Instruments.—The methods used for determining the amount of

radium in either an ore, or in a product such as a concentrate containing radium,
depend entirely upon the fact that radium and other radioactive substances
ionize gases. Such ionization is due to the alpha, beta and gamma rays which
are emanated by radioactive substances. The methods are, therefore, entirely
physical in character and involve any means of recognizing qualitatively or
quantitatively the ionization in air or other gases produced by the alpha, beta
and gamma rays. There are two instruments which are usually used for such a
purpose, namely, an electrometer or an electroscope.
The first instrument is adapted for use in chemical and physical laboratories, and
is especially useful where a large number of readings are desired on a radioactive
material in a short space of time. For example, where a decay curve is required and
the points on the curve involve very short intervals of time, an electrometer is very
useful, owing to the fact that the length of time in making a reading on even a not
very sensitive product is short, due to the senstiveness of the instrument that may be
used. For ordinary practical purposes, however, an electroscope is much more
satisfactory, and the use of the electroscope only will be described in detail in this
article. A suitable instrument usually consists of two compartments, one above
containing a suspended gold or aluminum leaf in front of which is attached a reading
microscope, and one below in which the ore, radioactive solid or radioactive gas, to be
tested is placed.
If the material is a solid, there is a suitable door to the lower compartment which
can be opened and closed for the introduction of the material. If radioactive gas is to
be tested, the lower compartment is airtight and has two stopcocks, one for exhausting
the chamber, and the other for the introduction of the gas after partial or complete
exhaustion. Usually the leaf is electrically charged by means of a piece of vulcanite
rubbed on the sleeve of a coat, or by means of a battery of small dry cells, or any
other suitable means of getting a sufficiently high voltage; the charge causing the
leaf to rise. Then the natural leak of electricity from the leaf is noted on the scale,
using a stop-watch to determine the time the leaf drops between two different points,
and calculated as a certain number of divisions per minute.
Approximate Method for Solids.—If an approximate determination of the

activity of a solid such as an ore is desired, the material is placed in the compart¬
ment below and the leak of the leaf noted as before. If the ore contains uranium,
radium, or any other radioactive element, the rate at which the leaf falls will
always be faster than the natural leak of the instrument itself, owing to the
ionization of the air in the chamber by the rays given off from the radioactive
material. There are a number of precautions, however, to be taken in making
such measurements. First, the illumination during the taking of the readings
should be constant, and, therefore, it is better to have the electroscope in a room

artifically illuminated rather than using ordinary daylight which will vary from
time to time. Second, readings should always be taken between the same points
on the scale. Third, in comparing two ores, their physical condition should be
as nearly as possible the same. This may be roughly done by passing
them through the same mesh sieve, preferably 40 or 60 mesh. Of course, every
particle of the ore must be ground until it finally passes this sieve. The same
weight should be taken and the same surface should be exposed in the
electroscope.
In order to get these conditions conveniently, it is advisable to use a brass plate

about one-quarter inch thick, of a size to fit into the bottom compartment of the
electroscope; and in this should be cut by means of a lathe a circular depression one-
eighth inch deep and about 3 in. in diameter. This can be done by any brass worker.
The bottom and sides of the depression should be perfectly smooth. The ore to be
tested is poured into the depression, the plate tapped gently so as to settle the ore,
and then by passing the edge of a flat piece of metal across the surface of the plate the
extra ore is wiped off and the depression left exactly filled with ore, having a flat
surface. In this manner a fairly uniform weight of material is obtained for comparison
and the surface exposed in the electroscope is approximately constant. Of course,
the density of the ores tested varies, but the method is accurate enough to give
approximate results. The plate with the ore is introduced into the bottom com¬
partment of the electroscope and a reading taken. The ore is removed and replaced
by a sample of carnotite of known uranium content, which serves as a standard.
This sample, of course, is passed through the same mesh sieve as the sample being
tested. The relative radioactivities: that is, the rates at which the leaf falls, are
roughly proportional to the amount of radioactive elements present.
In the case of an ore, the total activity will be due to the uranium, radium and
other disintegration products of the series. In the case of a concentrate, such as
radium-barium sulphate, the activity will be due to radium alone. Too much
emphasis can not be placed upon the fact that this method is only applicable to
comparison between similar radioactive products, and results are only approximate.
For example, a carnotite ore must be checked against an analyzed carnotite ore;
pitchblende against an analyzed pitchblende ore, and a concentrate against a similar
type concentrate. As has already been pointed out, the physical condition of the
material affects the results very markedly.
The following example will show how to make a calculation:
Method of Calculating Percentage of UaOa
Natural leak of instrument — 5 divisions in 10 min.

Natural leak of instrument = 0.5 division per min.
Rate of fall of leaf with standard ore (3 per cent UsOs) = 48.5 divisions per min.
Rate of fall of leaf with ore to be tested = 36.5 divisions per min.
Subtracting from each of these figures the natural leak, 0.5 divisions per min.,
36 X 3
we have 48 and 36. The percentage of UsOs in the ore will then be — = 2.2.
If the natural leak is as low as 0.5 division per minute and the radioactivities
of the samples are as high as those indicated in the above experiment, the natural leak
can be neglected, as the error from it is less than the probable experimental error.
Exact Determinations by Emanation Method.—The exact determination of
radium is by means of the emanation method which involves separating radium
emanation as a gas from its parent radium, and measuring its quantity in a gas-
tight electroscope previously standardized with a known amount of radium
emanation. Analyzed pitchblende has been employed to furnish known quan¬
tities of emanation for purposes of standardization. Pitchblende is selected
because it is a primary uranium mineral and contains the equilibrium amount of
radium to uranium. Therefore, by analyzing the mineral carefully foi uinnium
it is possible to calculate exactly the amount of radium present; and a satisfactory
standard is thus readily obtained. Secondary uranium minerals, such as carnotitc,
autenite, torbernite, etc., do not always contain the equilibrium amount of
radium and, therefore, can not be used as a standard.
Three general methods of procedure may be used, as follows: (1) Release and
measure the emanation from a substance in which it is in equilibrium with the radium
content. This condition will usually not be fulfilled unless the substance has been
retained for a month or more in a closed container. In exceptional instances, however,
the radium might be contained in a solid of very compact structure, or with a glazed
surface, so that no spontaneous loss of emanation could take place. But even with a
dense mineral like pitchblende, the leak of emanation, called “emanating power,”
amounts at ordinary temperature to several per cent.
This circumstance suggests the second procedure. (2) Liberate and measure
the emanation retained in the solid and apply as correction the “emanating power,”
which must be determined separately and preferably after the solid has been in a
closed retainer for 1 month. Both of the above procedures, applicable in general
to solids only, involve in practice long delays; and, although they are adapted to
scientific investigation, they are not suited to radium measurements for the purpose
of plant control when quick results are desired. The following procedure is shorter
and preferable when its use is possible. (3) Remove the emanation completely from
sample of the substance to be analyzed for radium; close it at once in a gastight vessel
and allow the emanation to accumulate for a convenient period, such as from 1 to 10
days. Then remove it and measure it, making a time correction to find the maximum
amount that would have been formed on the attainment of equilibrium. For removal
of emanation the radium must be contained either in solution, or in a state of fusion.
When radium and barium are in a solution together, and there is a tendency for
partial precipitation, either as a sulphate or as a silicate, the two elements will usually
precipitate in the proportion in which they exist in the solution. The presence, there¬
fore, of a moderate amount of barium in solution has a tendency to hold the radium
in solution and give more exact results by the emanation method.
Treatment for Solution Containing Barium in Large Excess over Radium.—

For a solution containing barium in large excess over radium, the treatment is as
follows: Place a suitable portion of the solution—such as will contain about 1 X
10 # radium in a small Jena flask and add to it a suitable quantity of 1:1
nitric acid. Add a few glass beads and boil 5 to 10 min. to remove all emanation.
Allow slight cooling and then close the flask tightly with a one-hole rubber
stopper provided with a glass tube drawn out above to a capillary tip. Seal the
t!p some steam is stiU in the flask, in order to provide a partial vacuum which
should be maintained until the flask is again opened, thus affording a proof that
no outward leak of gas has taken place. Note the exact time and date of sealing.
-°r a.flultion containing little or no barium is to add a suitable

^Tn XL♦ aCldy,hl? is satur^ted ™th barium nitrate, and then proceed
as m the treatment desenbed above.
Procedure for Liquids Containing Excess of Sulphate or Carbonate.—The

•detailed procedure for treating a liquid containing an excess of sulphate or carbon¬
ate, but no barium, is as follows: An excess of barium salt is added to the liquid,
and the precipitate is filtered off. The filtrate containing an excess of barium is
made acid with nitric acid to the point of precipitation, and is given the treat¬
ment outlined above for a solution containing barium in large excess over radium.
The precipitate, if barium sulphate, is fused with four or five times its weight of
fusion mixture of sodium and potassium carbonates and is treated as described
later for fusions. If the precipitate is barium carbonate, it is dissolved in nitric
acid containing sufficient sulphuric acid to precipitate an amount of barium sul¬
phate convenient for fusion, which is filtered off. The filtrate that is obtained
may be combined with the original filtrate and given the treatment as described
for a solution containing barium in large excess over radium. All radium is then
contained either in the filtrate, with excess of barium, or in the barium sulphate
precipitate. The latter is fused with sodium carbonate and treated as described
in the next section. Both of the liquid fractions are closed simultaneously
(within 15 min.) so that the time of accumulation will be the same for both lots of
emanation which can be later introduced into one electroscope to determine the
total radium.
Fusion Method for Radium Determinations.—If the radium is contained in a
substance not readily soluble, such as a radium-barium sulphate, fuse a suitable
quantity in a small platinum or porcelain boat with four to five times the weight
of sodium or potassium carbonate and note the exact time at which the material
becomes solid. Close this boat in a glass tube as shown in Fig. 1. Allow the
emanation to accumulate 2 or 3 days. Connect the glass tube at one end to a
highly exhausted electroscope and at the other to a stopcock. Break the glass
tips inside the rubber connections and exhaust the air from the glass tube into
the electroscope several times, leaving enough vacuum in the electroscope cham¬
ber to accommodate the gas to be introduced later. Break the glass tube,
remove the boat and its contents, wrap in a filter paper and place in the neck of a
flask as shown in Fig. 2; it is then ready for treatment with 1:1 nitric acid after
the flask has been connected with the gas burette as shown in Fig. 3. In this
treatment the flask is tipped until the acid comes in contact with the carbonate
fusion, thus beginning a gas evolution. The stopcock is immediately opened
to the gas burette above and the boat and contents are then thoroughly wet with
acid and jarred down from the neck of the flask to the body of the acid.
As regards larger fusions, the evolution of carbon dioxide may become rapid and
care should be taken in handling them; but in small fusions not exceeding 1 g., the
boat may be shaken down directly into the acid, which should be heated to boiling
as soon as the gas evolution begins to slacken. All of the carbon dioxide is, of course,
absorbed by the strong sodium hydroxide solution which is contained in the gas burette.
The boiling off from this point is performed as with solutions discussed below.
Boiling off Emanation from Solutions.—For boiling off emanation from solu¬
tions, the procedure is as follows: Set up an apparatus as shown in Fig. 3, wiring
rubber connections at a and b to insure tightness. Put into the leveling bulb
c a stick of sodium hydroxide 2 to 3 in. long, or more if a large quantity of carbon
dioxide is to be absorbed; make sure that stopcock d is closed and stopcock e
open; pour boiling distilled water into the leveling bulb and allow the alkali to
go into solution. If the boiling is too violent, put a one-hole stopper lightly into
the mouth of the leveling bulb. After the alkali has gone into solution raise the
leveling bulb until the gas burette is filled to the stopcock e. If the quantity of
air to be boiled off is small, some air may at first be left in the gas burette. Close
stopcock e and lower bulb c to its original position. Break the glass tip / inside
SZ.~ZZL
Fig. 1.—Apparatus for fusing Ra-Ba sulphate.
the rubber tubing at a, and slowly open d to ascertain whether there is vacuum
in the flask g. If so, close d again and begin to heat flask g over wire gauze.
Test the vacuum every few seconds and as soon as the pressure is outward open
Continue boili*g ™til live steam has

less thsffl 5 anci 6 âs ûreê h. This boding should never be
glass°,-0,1,5 ^ bom^ is desirable' After the
in someinstances has lodged in thT t ** ^ ^t0 be carried uPward by steam and
a precaution a roll of thin nl r st?P®ock d and caused serious explosions. As
as indicated'at i or the stnn ^ 01 °an ^ “ôduced into the glass tubing,
tea t, or the stopcock d may have a wide bore, which also obviates
the danger mentioned. After the boiling off has been completed, remove the
tlîne, and as soon as the liquid begins to draw back through the stopcock d close
the stopcock and remove the flask entirely. Evacuate the electroscope chamber
to a suitable vacuum, either by means of an aspirator, or more conveniently,
^ hand pump, and connect the sulphuric acid micro drying bulb l to the electro¬
scope and to the gas burette as indicated in Fig. 3. Be sure that stopcock j is
dosed; open first the cock of the electroscope for a moment and reclose it; then
slowly open stopcock e to full width and then gradually open the stopcock to
the electroscope, allowing the gas to bubble through the microdrying bulb at a
fairly rapid rate.
When the liquid in the gas burette has risen exactly to the point h close stopcock
e and open stopcock j, allowing dry, dust-free air which should preferably be taken
from outside the laboratory to sweep out the connections for a few minutes; then close
the stopcock to the electroscope, reopen stopcock e and allow the liquid in the gas
bnrette to fall back 3 or 4 in. below the shoulder; close e, and then pour off all excess
licjuid out of c; close j and again open e to the electroscope, allowing air to bubble
from the bottom of the gas burette h through its entire length to insure the removal
of any emanation that may have remained dissolved in the liquid. Air should be
allowed to bubble into the electroscope chamber until normal pressure has been almost
restored. The above procedure for boiling off radium emanation is used for carbonate
fnsions introduced into acid, and also in handling any solids that are to be dissolved
directly. For example, ground pitchblende and carnotite ore may be wrapped in
filter paper in the way in which such a fusion is wrapped in Fig. 3, or sealed in small
glass bulbs which are opened by being crushed against the bottom of the flask by
tapping on the glass stem projecting through a second hole in the rubber stopper.
Construction of an Electroscope.—Figure 4 shows an electroscope which is
very suitable for radium determinations by the emanation method. Figure 5
represents a section of this electroscope, showing the gastight chamber at the
bottom with openings at o for connections with stopcock. The electrode e is a
fc>rass cylinder M in. in diameter. It is suspended by a small brass rod H in. in
diameter, which screws into the top of the electrode, passing upward through the
Insulating material d, and terminating in a small conical top, c, serving to make
metallic contact with the leaf system above. The leaf system, /, is supported
from the top of the cylinder where it is held in place by the sealing-wax insula¬
tion set in a milled-head cap, g, which screws into a vertical collar on the cylinder
in. in height. The cap is hollowed out inside to contain the insulating wax,
from which a flat brass rod, /, J4 in. broad, about }{e hi. thick and 234 in. long
projects downward, terminating below in a light brass spring, $, to make contact
with the conical top of the electrode of the ionization chamber. The leaf is of
aJLiaminum, about 2 in. in length, and is attached to a small offset at the top of the
brass rod by a moisture contact. The electroscope can either be charged by
means of a piece of hard rubber or, better, by means of a battery of flashlight
storage cells giving 500 or 600 volts. The reading microscope in front of the
Instrument carries a micrometer scale serving to measure the rate of discharge of
tire leaf. The eye-piece fits firmly into the case so that its rotation is difficult
adFter the micrometer scale has been set parallel to the leaf.
Order of Procedure in Using Electroscope.—(1) Set up electroscope and
charge for 15 min. from a battery with just sufficient voltage to hold the leaf on
the part of the scale to be used later. (2) Observe the natural leak during 15 or
more minutes. (3) Carry out the calibration control by means of penetrating
rays if radium is available for this purpose. (4) Detach the top and evacuate tilt?
lower chamber to the desired vacuum. (5) Pass the emanation-air mixture
through a sulphuric acid drying tube into the evacuated chamber and restore
Fig. 5.-—CrosH-Bcctt ion of elec¬

troscope (one-fourth sisse.)
for 3 hrs.^fDemanaôn to stand in the discharge chamber

meats are jmod n !f mm* aS before* Take three readings if agree-
out the erJLti’n j!° t deiVlatl0ns.are greater than 1 per cent. (9) Clean
time (over nwht if ° ^ ^ ^rawblg dust-free air through it for some
natural leak ^ Calculate the discharge and subtract the
^ (H) Compare th« eor-
tity of radiunfunder ™ Cahbratlon of the instrument to determine the quan-
CaliSaZ TmlZmemen^takin^ time corrîons into consideration,
out in exactly the samp °SC0I>e#. be cakbration of the electroscope is carried
two ways as follows: (Ii Fm™

rs r~ts' iL *
+ • This known quantity may be obtained in
ah through it until its Pm« *• sta.nda;^d solution of some radium salt by passing
method is not satisfactory aTiiTis diffiU 1t,rans1ferred into the electroscope. This
in solution, owing to the t cu^ know always the amount of radium
solution as sulphate or silicate ^ ~?0rtl0a of *he radium to precipitate out of
*mdyw*d pitchblende a sni^wl Iile Preferable practice is to use high-grade
“rdiiatiaa, and the quantity nf ^“antlt^ bemS dissolved for each stand¬
by®. The quantiWradif ^ being calculated from the uranium
7 radium emanation obtained on dissolving the pitch
blende will not correspond exactly to the radium content because a small frac¬
tion of from 2 to 10 per cent of the gas diffuses from the ore; this fraction, termed
the “emanating power” of the ore in the cold, must be determined by sealing a
quantity of the ore in a tube for a month or more, and drawing off the emanation
into an electroscope by the passage of air. The “emanating power” thus deter¬
mined in the standard sample is used as a subtraction correction. Convenient
quantities of radium emanation are those that will produce a discharge of the
order of one to two scale divisions per second. The use of pitchblende as a
standard is based upon the fact that in any unaltered uranium mineral the ratio
of the radium present to the radium is constant. One gram of uranium is in
radioactive equilibrium with 3.3 X 10~7 g. of radium.
Sample Determination of Radium Content of an Ore.—A sample of pitchblende
loses 10 per cent of its emanation at room temperatures. It contains 50 per
cent metallic uranium. Therefore 22 mg. of the ore will, on dissolving in acid,
liberate emanation in equilibrium with 10 mg. (0.01 g.) or metallic uranium.
This emanation, 3 hr. after introduction into the electroscope, causes the leaf to
fall at the rate of 40.5 divisions per minute. The natural leak (0.5 divisions
per minute) subtracted from this leaves 40 divisions per minute due to the emana¬
tion. Therefore the fall of one division per minute represents the total emana¬
tion associated with --- = 2.5 X 10“4 g. of uranium in the mineral. This is the
40
“constant” for the electroscope. One gram of ore is fused with fusion mixture
as already described. At the end of a month the emanation obtained from the
two solutions is introduced into the electroscope. After 3 hr. the rate of fall of
the leaf is 18.5 divisions per minute. Subtracting the natural leak (0.5) leaves
18 divisions per minute. Therefore 1 g. of the ore contains 18 X 2.5 X 10~4 =
45 X 10""4 g. of uranium.1 As 1 g. of uranium is in radioactive equilibrium
with 3.3 X 10~7 g. of radium, 1 g. of the mineral will contain (45 X 10“4)X
(3.3 X 10-7) = 1.48 X 10~9 g. of radium.
OTHER RADIOACTIVE ELEMENTS
Me so thorium.—Of the other radioactive elements which have commercial

use mesothorium is the most important. As is seen from Table 2, it is a disinte¬
gration product of thorium and is, therefore, associated with this metal in all
thorium minerals. Its half-life period is very short, compared to that of radium;
being only 5.5 years. Therefore, commercially it has not the same value as
radium, especially for cancer treatment or for purposes where it would naturally
be carefully preserved and a long life would be of advantage. When its com¬
mercial use is likely to last over a limited period, such as in luminous paint for
cheap watch dials, electric push buttons, etc., it is just as useful as radium and
can be substituted for the latter element in luminous paint for such purposes.
Owing to the fact that mesothorium 1 gives off beta rays only, and the alpha rays
1 This is true only when the uranium and radium are in equilibrium. In other cases it represents
the theoretical amount of uranium in equilibrium with the radium actually present. In pitchblende,
since it is a primary mineral, the ratio of uranium to radium is constant, 1 g. uranium = 3.3 X IQ"7 g. of
radium, or 2,000 lb. UaOii = 331 mg. RaCh, and if the percentage of uranium is known by analysis,
the ftmnnnt, of mHinm ■nrân.Tvt- r>«-n enlenleterl Hireetlv In oarnntite and other recent u anium
are the main source of luminosity in paints used for watch dials, etc., it is neces¬
sary for the mesothorium 1 after preparation to be allowed to “ripen” for a
year or even 2 years so that the alpha rays, due to the gradual accumulation of
radiothorium, can be used. The general effect, therefore, is for the luminosity
of such paints gradually to increase for 2 or 3 years and then, after coming to
a maximum, gradually to decrease. Mesothorium 1 chemically is allied to barium
and radium and, therefore, can be precipitated with barium just as radium is.
The usual procedure in manufacture is to add 2 or 3 lbs. of barium chloride
per ton of monazite treated. When the monazite is heated with sulphuric
acid in order to extract the thorium and other rare earths, the barium sulphate
and mesothorium sulphate are left behind in the residue and can be recovered
from the coarse silica, etc., by sliming. The crude concentrates so obtained can
be still further purified by fusion with a mixture of caustic soda and sodium
carbonate, by which means the silica is converted into sodium silicate and can
be washed away from the barium-mesothorium carbonates, or the silica can be
eliminated by the use of hydrofluoric acid. The refined sulphate so obtained
is purified by the same methods described under the refining and purification
of radium.
Actinium.—Actinium was discovered by Debierne in the iron group separated
from pitchblende, very shortly after the discovery of radium and polonium. Act¬
inium itself is probably rayless, but its first product, radioactinium, has a half-life
period of 18.8 days. Actinium preparations when first made increase enorm¬
ously in activity over a period of several months. It is, therefore, comparatively
easy to overlook its presence, unless the preparations are kept and their activity
tested periodically. Actinium is easily obtained with iron and rare earths by
precipitating barium as sulphate in an acid solution; and this accounts for its
presence in pitchblende residues. In working these up, the acid solutions, after
removal of polonium by means of hydrogen sulphide, are oxidized and precipi¬
tated with ammonia; the actinium being precipitated and the radium and
barium, of course, remaining in solution. The precipitate may be extracted with
dilute hydrofluoric acid; the insoluble part consisting of La, Di, Ce and Th retain¬
ing most of the actinium. The rare-earth element most closely associated with
actinium is lanthanum. In the presence of ammonium salts the precipitation
of actinium is far from complete, but it is completely precipitated in the pres¬
ence of manganese from basic solutions.
Polonium.—(Radium F). Polonium is one of the short-lived radioelements,
having a half-life of 140 days and giving off alpha rays only. It was discovered
owing to the fact that it is closely allied to bismuth and precipitates with bis¬
muth in the second group. For this reason it is easily separated from radioactive
minerals in crude form by digesting radium containing residues of pitchblende
or carnotite ore itself with hydrochloric acid. A part of the polonium is dis¬
solved and may be precipitated with hydrogen sulphide. The polonium may be
purified: (1) By fractional precipitation from solutions made very acid with
hydrochloric acid, the polonium being enriched in the precipitate. (2) By
fractional precipitation of the basic nitrate with water, the precipitate being
enriched. (3) By sublimation in vacuo, the polonium being more volatile.
By immersing a plate of bismuth, silver, copper, etc. in a hydrochloric acid solu¬
tion, the polonium is practically completely precipitated. The theoretical quantity
of polonium in minerals is 1 g. for 14 tons of uranium element, assuming that

no radium emanation escapes from the mineral. As such a condition never
exists in nature, the actual amount is less then the theoretical.
Radiolead.—(Radium D). As has been previously stated, there are three
kinds of lead: One representing the final disintegration product of the uranium
series; one the final disintegration product of the thorium series; and ordinary
lead which so far as we know does not owe its origin to radioactive changes.
The first has an atomic weight of 206, the second 208, and ordinary lead 207.
Since radiolead, or radium D, resembles lead perfectly in all of its chemical
reactions, in the treatment of uranium minerals it can not be separated from the
lead which is found in such minerals. It is, therefore, obtained in association
with this lead when the latter is precipitated during the refining of the radium-
barium sulphate. Common commercial lead contains traces of radium D and
is more distinctly radioactive than most other metals. For this reason, in making
instruments, a very old lead should if possible be employed, as in this the radio¬
active constituents will largely have decayed. Another source of radium D
is old radium, in which the radiolead has to a certain extent accumulated. If,
however, a sample of this element is required in concentrated form, the best
method is to remove the emanation periodically from a radium solution kept in
an airtight vessel and introduce the emanation into a closed vessel. As the
emanation decays the radiolead will be formed and will be deposited on the
sides of the vessel. This can then be dissolved by acid and precipitated by
ammonia or other lead precipitants, provided there is a small amount of lead
salt introduced in order to give a precipitate sufficiently large to be handled.
SECTION XXVII
RARE METALS
By Richard B. Moore1
Introduction.—It is not possible within the scope of the present volume to give
even an outline of the chemistry and metallurgy of all the rare metals. A selec¬
tion, therefore, has been made of those which are of most importance in the indus¬
tries, and the present article will deal only with the following: Vanadium, uranium,
tungsten, molybdenum and thorium.
VANADIUM
(Atomic Weight = 51.0)
Occurrence.—Vanadium is never found free in nature. Its ores are fairly

widely distributed, but seldom occur in quantity in any one locality. The prin¬
cipal ores are patronite, roscoelite, carnotite, vanadinite, cuprodescloizite, zinc
descloizite, mottramite and pucherite. Vanadium is found in small quantities
in certain iron ores, and in the ash from lignites, coals and other carbonaceous
materials.
Patronite is an impure vanadium sulphide found in Minasragra, Peru. The

composition is approximately V2S9, and it is associated to a certain extent with car¬
bonaceous material, pyrites and free sulphur. The area in which the ore lies2 is along
the western limit of a broad anticline in1 ‘ Jura-trias ’ ’ and Cretaceous rocks. A section
shows the series in this locality to be composed of green shales, thin beds of limestone
and red shales. Vanadium is found only in the red shales. The deposit proper ap¬
pears to be a lens-shaped mass, 28 ft. wide and 350 ft. long. The mineral that con¬
stitutes the larger portion of the deposit has been called “Quisqueite.” It is a black
carbonaceous substance containing sulphur, with a hardness of 2.5 and a specific
gravity of 1.75. There is also a lesser quantity of a coke-like material with a hardness
of 4.5 and a specific gravity of 2.2. Neither of these contain vanadium. The vana¬
dium is mostly in the southern end of the ore body, and to a depth of 20 ft. is largely
in the form of a red calcium vanadate, and carries as much as 50 per cent vanadic
oxide. It occurs in pockets and fills the cracks and fissures in a fine shale.
Below this shale is the mother lode. It is from 9 to 30 ft. thick and extends along
the greater length of the deposit. It carries as high as 10 per cent vanadic oxide and
nearly as much sulphur. On the east and south sides below the mother lode is found
a hard blue-black vanadium shale carrying as much as 13 per cent vanadic oxide and
4 to 5 per cent sulphur. Patronite, the main vanadium mineral, is greenish-black
and has a hardness of 2.5 and a specific gravity.of 2.71. It contains from 19 to 24.8
per cent vanadic oxide and sometimes 50 per cent of combined sulphur. This deposit
in Peru constitutes the largest known source of vanadium at the present time and
1 Chief chemist, Bureau of Mines, Washington, D. C.
,
* D. F. Hewett, Eng. and Min. Jour., Vol. 82 Sept. 1, 1906, p. 385; W. F. Hillebrand, Jour.
,
Am. Chem. Soc., Vol. 29 1907, p. 1019.
furnishes the American Vanadium Company of Pittsburgh, Pa., with practically its
only supply of ore. As this company has been the largest producer of ferrovanadium
and vanadium compounds in the world, patronite is by far the most important ore
of vanadium from the commercial standpoint.
Boscoelite, or vanadium mica, is the second most important vanadium mineral
commercially. It has a somewhat indefinite composition; the formula HsK2(MgFe)-
(AlV)4(Si03)i2 representing its composition fairly well. It exists in minute scales of
a brown to greenish-brown color. Sometimes the color is actually green. The
specific gravity is 2.92 to 2.94. It exists rather widely distributed in nature, but at
only one place in the United States is it worked commercially, namely, near Newmire,
San Miguel County, Colo.
Carnotite is mainly of importance on account of its uranium and radium content,
but also carries usually from 3 to 5 per cent V2Os in ordinary commercial ores. It is
found mainly in Southwestern Colorado and Eastern Utah, but also exists in smaller
quantities in other localities, such as South Australia, Portugal, etc. It is a potassium
uranyl vanadate of the following approximate composition K2O.2UO3.V2O5.3H20.
The crystals are canary-yellow in color and usually exist as incrustations on other
minerals, or are disseminated through a sandstone in which they are usually found.
It also contains small traces of calcium and barium.
Vanadinite, or lead vanadate, is widely distributed, especially in certain of the
western states of the United States, particularly Arizona, New Mexico, Nevada and
California. The largest deposits are in Arizona, where it is frequently associated
with wulfenite, or lead molybdate. The deposits are almost always low-grade and
for this reason this mineral has not been used to any great extent as a source of vana¬
dium in the United States. In Mexico the vanadinite occurs at Villarosales,
Chihuahua. In Spain important deposits of vanadinite occur near Santa Marta, Estra-
maduras. The ore is found in a sandstone which contains on an average 3 per cent
vanadic oxide. This can be concentrated, and these deposits supplied most of the
vanadium used in the world up to the opening of the Peruvian mines. The mineral
occurs in prismatic hexagonal crystals, and has a specific gravity of 6.6 to 7.1. It
varies in color from ruby-red to yellowish and even brown.
Cuprodescloizite is a hydrated lead-copper vanadate of the probable compositino
(PbCu)2(0H)V04. Commercial samples of the ore may carry as much as 23 per
cent PbO, 7 per cent CuO and 8 to 10 per cent V205. The largest deposit of this
mineral known at the present time is in the Shattuck mine at Bisbee, Arizona, although
it is found in a few other localities in the United States.
Zinc descloizite is a mineral very similar to cuprodescloizite, where the zinc re¬
places the copper. It is found in Nevada, especially around Goodsprings, in New
Mexico, Arizona and in Argentina. Mottramite, a vanadate of lead and copper, and
pucherite, a bismuth vanadate, are of lesser importance to the minerals already
described.
Extraction from the Ores.—The method of extraction which can be used for
vanadium ores varies materially with the ore and its grade. In general, a
different metallurgical process must be used for every mineral treated, although
there are, of course, some general similarities between the methods used.
Patronite is treated by first roasting in order to eliminate the sulphur as far as
possible. It can then be leached with either sulphuric acid, filtered and the filtrate
evaporated to a crude vanadic oxide; or the roasted ore can be leached with alkali—
so urn. carbonate or caustic soda, which will dissolve the vanadium as sodium vana-
ate. From this a high-grade vanadic oxide can be obtained by acidifying with
s p uric acid, or after neutralization the vanadium can be precipitated as vanadate
of iron by means of ferrous sulphate.
RARE METALS 885
Roscoelite is treated commercially at only one place in the United States, namely,
at Newmire, Colo., where the Primos Chemical Co. has a mill. The method they use
is to roast the ground ore with salt and a small percentage of ground pyrites. The
ore contains almost no uranium, and no attempt is made to recover the minute quan¬
tity found in some of it. The mixed ore, salt and pyrites are coarsely ground and
run into a dryer which reduces the moisture to 1 per cent. The material which
comes out somewhat caked is ground to 20 mesh and roasted in a furnace for about
3 hr. The roasted material is delivered at the bottom of the furnace to conveyors
which carry it to lixiviation tanks where it is treated with water and filtered. The
vanadium is now in solution as sodium vanadate. A solution of ferrous sulphate is
added, and the vanadium is precipitated as vanadate of iron. This precipitate is
filtered out of the solution and partially dried and shipped to the company’s reduction
works at Primos, Pa. Roscoelite is not very amenable to acid-leaching methods as
the mineral is quite insoluble. G. A. Koenig (U. S. Pat. 986,180) claims, however,
that the mineral can be decomposed by either dilute sulphuric acid or hydrochloric
acid under proper treatment as regards both heat and pressure. In practice, he
proposes to use a solution containing 20 per cent sulphuric, hydrochloric or other
acid at a temperature of about 200°C. and under a pressure of 225 lb. per square inch;
and he claims that this process will completely decompose and dissolve the roscoelite
within a few hours.
Carnotite is almost exclusively treated for its radium content, and practically all
metallurgical processes for the treatment of carnotite have in view the extraction of
the radium primarily, while the recovery of the uranium and vanadium is of second¬
ary importance. Some processes, therefore, do not obtain a high recovery of the
vanadium. Metallurgical methods for the treatment of carnotite are discussed more
fully in the article on “Radium” in this volume, and under “Uranium” in this
section. In order to extract vanadium from carnotite, all that is necessary is to
follow the process originally patented by Haynes and Engle (U. S. Pat. 808,839).
The ore is first crushed to 12 mesh and is then boiled with a solution of an alkaline
carbonate, preferably sodium carbonate, until the uranium and vanadium are dis¬
solved. The strength of the sodium-carbonate solution and the time necessary to
boil are determined by the proportion of uranium and vanadium in the ore, and will
vary considerably. After these materials are dissolved, the clear solution is treated
in a separate tank, where the uranium is precipitated as sodium uranate by the addi¬
tion of sodium hydroxide to the solution. This precipitate is then filtered from the
solution which contains the vanadium. The vanadium can be precipitated on neu¬
tralization as iron vanadate or calcium vanadate.
Vanadinite was originally treated by leaching with sulphuric acid. This dissolved
the vanadium and left the lead as insoluble lead sulphate, from which the lead could
be recovered by smelting. The process was not a success, because the action stopped
after a short time, owing to the crystals of the vanadinite being coated with the lead
sulphate, which prevented further action. J. E. Conley has thoroughly studied the
treatment of vanadinite.1 The method recommended by him is to fuse the vanadinite
concentrate with a mixture of soda ash and caustic soda, which gives a better melt
and a better recovery than soda ash alone. The lead is recovered as metallic lead
and the vanadium is precipitated by means of slaked lime, giving calcium vanadate
as a product. If any trace of molybdenum is present, the latter metal does not come
down with the vanadium and is thus eliminated. The calcium vanadate can be
treated with sulphuric acid, the precipitated calcium sulphate filtered off, and the
vanadium recovered as high-grade vanadic oxide.
Cuprodescloizite has never been treated commercially but, as it constitutes an
important and potentially fairly large source of vanadium, the method of Conley2 is
,
1 Chem. and Met. Eng., Vol. 20 No. 10, p 514.
of interest. The object is to recover the lead, copper and vanadium. Conley found
that a sulphuric-acid leach gives a fairly satisfactory extraction, but his preferable
method is to subject the ore to a preliminary fusion with nitre cake, followed by a
hot-water leach, and then treating the residue with sulphuric acid. By heating the
ore with an equal weight of nitre cake, approximately two-thirds of the ore can be
decomposed and the corresponding values extracted. The remaining values are then
extracted by leaching the residue with about one-half as much acid as is required in a
straight-acid treatment.
Zinc descloizite up to the present time has not been treated commercially, and no
methods have been published or suggested for its treatment. The principal deposits
in Nevada carry considerable quantities of calcite, which make an acid treatment
expensive and, therefore, undesirable.
Concentration Methods.—Most vanadium ores are not readily amenable to

concentration. Patronite is partially concentrated by roasting the sulphide to
the oxide. Cuprodescloizite has not been concentrated commercially, and the
same can be said of zinc descloizite. It is probable that a partial concentration
could be made in both cases by screening or by sliming, but the recoveries are not
satisfactory. Roscoehte is not concentrated by the Primos Chemical Co., the
ore being treated direct. The writer, however, knows of one gold ore in Califor¬
nia, which contains roscoelite, and from which a high-grade vanadium concen¬
trate is obtained during the regular concentration of the ore for gold. This,
however, is exceptional. Vanadinite is the only vanadium ore which can readily
be concentrated, and this is done by the usual methods involving tables and
slimers. It is, however, difficult to separate vanadinite from wulfenite, a mineral
with which it is frequently associated, owing to the fact that both minerals have
almost the same specific gravity.
Vanadium Compounds.—The compounds of vanadium which are met with
commercially are the oxides and the vanadates. The principal oxides are vana¬
dium trioxide (V2O3), vanadium dioxide (V2O4) and vanadium pentoxide or
vanadic anhydride (V2O5). The latter is the one which is mainly of interest
commercially. It is usually prepared from a solution of sodium vanadate by
adding either sulphuric or nitric acid to the solution until slightly acid. In a
concentration containing 10 g. of V2O5, as sodium vanadate, per liter, if sulphuric
acid is added until the acidity is 1/20AT, 92 per cent of the vanadium present can
be recovered as pentoxide under favorable conditions. Some vanadium always
remains in the filtrate, and if the concentration and acid strength are not favor¬
able, a very considerable amount may remain unprecipitated. There are five
sodium vanadates, namely, sodium metavanadate (NaVCh), sodium ortho¬
vanadate (NaaVO-O, sodium pyrovanadate (Na^C^), sodium tetravanadate
(Na3HV60i7) and sodium hexavanadate (Na2H2V60i7). The first three are the
important ones commercially. In solution the metavanadates are the most
stable, and the orthovanadates the least stable. At a high temperature the
reverse is true, the orthovanadates being the most stable, and are formed when
vanadium pentoxide is fused with an alkali carbonate, while the meta salt is pro¬
duced when a solution of alkali carbonate is boiled with vanadium pentoxide.
When ferrous sulphate is added to a solution of soluble vanadate, an iron vanadate
is precipitated. In commercial practice, the color of the iron vanadate produced
varies very materially from dark-brown to yellow, depending upon the conditions
RARE METALS 887
of the precipitation, and may contain from 20 to 45 per cent V205. The precipi¬
tate may consist of mixtures of the iron vanadates in different proportions, and
is frequently contaminated by ferric hydroxide.
Metallic Vanadium.—Metallic vanadium has been produced by Gin1 by the
electrolysis of a solution of vanadium trioxide in fused calcium vanadate. The
material may also be prepared by the alumino-thermic method, and by reducing
the dichloride in hydrogen, in which case the metal is a light whitish-gray colored
powder. The metal has a hardness greater than that of steel or quartz. It takes
a good polish and is not affected by air. If the powdered metal is thrown into a
flame, or rapidly heated in oxygen, it burns brilliantly. Its specific gravity at
15° is 5.5.
Ferrovanadium.—As the principal use of vanadium is in steel and as the
vanadium is always added to the steel in the form of ferrovanadium, the manu¬
facture of the latter product is of great importance. Probably 75 per cent of the
ferrovanadium produced in this country is made in the open-hearth or crucible
furnace by a modification of the thermit process, using aluminum as reducing
agent.2 The remainder is produced by the electric furnace, using 90 per cent
silicon as the reducing agent. A satisfactory ferrovanadium for commercial
purposes contains from 30 to 40 per cent vanadium and not more than 0.5 per
cent carbon, 1 per cent silicon, 2 per cent aluminum, 0.1 per cent sulphur and
0.1 per cent phosphorus. It has a good fracture, is not crystalline and is bright
gray in color. Too much carbon is injurious, owing to the fact that it makes a
carbide with the vanadium, which is not satisfactory in steel making. In the
aluminothermic or Goldschmidt method, vanadium oxide or iron vanadate is
reduced with aluminum shot in a gas-fired open-hearth furnace, slagging off the
alumina by addition of soda ash or fluorspar. When the oxide is used, iron turn¬
ings must also be added. The process is partly exothermic, but not enough heat
is generated to carry the reduction to completeness without the application of
external heat. The use of iron vanadate requires more aluminum than the oxide,
as the iron must be reduced also. In the electric-furnace method, carbon is not a
satisfactory reducing agent, and in addition it has a tendency to form carbides
which are not acceptable to steel manulacturers. Silicon works excellently and
is usually used. The raw materials required are steel turnings, vanadium oxide,
silicon, lime and fluorspar. If iron vanadate is used, the steel turnings can be
eliminated. While 90 per cent silicon material ground to 60 mesh is preferable,
ferrosilicon can be used if necessary. All material should be low in phosphorus,
but the presence of sulphur is not so objectionable, as the sulphur content can
readily be reduced by re-melting the ferro; the sulphur going into the slag. The
melting point of ferrovanadium, practically free from other elements and contain¬
ing 40 per cent vanadium, is about 1,480°C. The melting point becomes gradu¬
ally lower as the amount of vanadium is decreased until 35 per cent is reached,
when it melts at 1,425°C., and remains stationary until the alloy contains 30
per cent. At this point the melting point gradually rises to about 1,450°C.
Uses of Vanadium.—The chief use of vanadium is for making vanadium steel.
The general effect of this rare element on steel is to increase the elastic limit and
tensile strength without reducing the ductility. It is usually alloyed with
1 Blectrochem. and Met. lnd., 1909, No. 7, p. 2G4.
2 Bull A. I. M, E., August, 1910, p. 1342.
chromium or manganese to give chromium-vanadium or manganese-vanadium

steel, and is also used in open-hearth and high-speed tool steels. Vanadium is
found in much of the steel used in motor cars, locomotive axles, rock drills, etc.
The metal causes oxides and nitrides present to pass into the slag, and a certain
portion of it also goes into solid solution, rendering the metal more coherent
and less liable to disintegration. The quantity of vanadium added to steel is
small; the amount usually being between 0.1 and 0.4 per cent.
Other uses for vanadium are as alloys with copper and aluminum. These alloys
make excellent castings. It has been used in photography for toning silver-bromide
plates to a green color. Vanadium salts have also been used in pottery and glass,
and as mordants.
Analytical Methods for Vanadium.—Vanadium in an ore can be recognized

qualitatively in the following manner: The ore is ground and boiled with hydro¬
chloric acid, filtered, and to the cold acid solution hydrogen peroxide is added.
A deep-red color indicates the presence of vanadium. In the case of a mineral
such as roscoehte which is with difficulty soluble in hydrochloric acid, the ore
should be fused with sodium carbonate, leached with water, filtered, and the fil¬
trate made acid with hydrochloric acid. On adding hydrogen peroxide, the red
color will then be obtained. The choice of a quantitative method will depend
upon the ore, or vanadium compound, to be analyzed. Full analytical details
are given in U. S. Bureau of Mines Bulletin No. 70. The following method is
quick and applicable to a large variety of vanadium ores, especially those which
contain little insoluble material and are decomposed by acids. Weigh out 1 g.
sample of the pulverized ore and add 25 c.c. of concentrated hydrochloric acid.
Heat on a hot plate to a small volume, then add 10 c.c. of concentrated nitric acid
and continue heating until decomposition is complete. Remove from the hot
plate, cool and dilute to about 25 c.c. Now add 10 c.c. of concentrated sulphuric
acid and heat to copious fumes. Cool and dilute to 25 to 30 c.c., warm sufficiently
to dissolve any ferric sulphate which may have settled out, and then add either
sufficient powdered potassium permanganate, or a solution of potassium perman¬
ganate (10 g. per liter) to produce a strong permanent pink. Now add 50 c.c. of
concentrated hydrochloric and evaporate on a hot plate to fumes of sulphur triox¬
ide. Cool and dilute to about 150 c.c., warm to dissolve the ferric sulphate, and
then titrate at 70 or 80° with iV/20 KMn04 to a pink which persists for half a
minute. It is advisable to run a blank determination, which, usually requires
from 0.4 to 0.5 c.c. of N/20 KMn04. Fe factor X 1.63 = V2O5. In the case of
an insoluble mineral, such as roscoelite, it is advisable to filter off the insoluble
material and to treat with hydrofluoric and sulphuric acids in the usual manner to
eliminate the silica, and then to redissolve in a mixture of hydrochloric and nitric
acids. The method is applicable in the presence of large amounts of iron;"
molybdenum does not interfere; and arsenic, if present only in small amount, may
be overlooked. Large amounts of arsenic, however, should be removed by first
passing sulphur dioxide through the solution after evaporating the first time to
sulphur trioxide fumes, boiling off excess and passing in hydrogen sulphide.
Chromium interferes, but is seldom met in vanadium ores.
BARE METALS 889
URANIUM
(Atomic weight—238.2)
Occurrence.—The principal ores carrying uranium are pitchblende, carnotite,

autunite and torbernite. Pitchblende is an impure uranium oxide with rare
earths, carrying traces of lead, calcium, iron, bismuth, manganese, copper, silica
and aluminum. The specific gravity varies from 6.4 to 9.7, and the hardness is
usually about 5.5. Carnotite is described under “Vanadium.” Autunite is a
hydrated uranium-calcium phosphate, with the formula Ca(U02)2 (P04)2.8H20.
It is a bright-yellow mineral usually crystallizing in orthorhombic plates, and
has a specific gravity of from 3.5 to 3.9. Torbernite is a copper-uranium phos¬
phate with the formula Cu(U02)2.P208.8H20. The crystals are tetragonal with
a pearly lustre and of an emerald-green color. The specific gravity is 3.4 to
3.6. A further description of uranium minerals and their distribution can be
found under the chapter on “Radium.”
Extraction from the Ores.—All uranium minerals carry radium. The amount
of radium is usually proportional to the amount of uranium present, although
this is not always the case. In primary minerals the equilibrium amount of
radium is present, but in secondary minerals such as autunite and torbernite the
proportion of radium to uranium varies and is often as low as 60 or even 50 per
cent of the equilibrium amount. Notwithstanding this fact, uranium minerals
are now always treated for their radium content and not for the uranium; the
recovery of the latter being of secondary importance. Consequently, the metal¬
lurgy of uranium is really the metallurgy of radium, and full details concerning
the treatment of uranium ores will be found under the chapter on “Radium.”
In order to obtain the uranium alone, practically the same methods that are used
for the extraction of radium can be used, the only difference lying in the fact that
the radium need not be precipitated as a radium-barium sulphate, this step being
omitted. Any method that gives a good extraction of the radium may be used
to recover uranium, as it is necessary, of course, to decompose the mineral thor¬
oughly in order to get the radium.
For analytical purposes uranium can be extracted from one of its ores by fusing
with excess of sodium carbonate. The melt is broken up, boiled with water and
filtered. The uranium goes into solution as the double carbonate of sodium and
uranium, and will contain only small traces of aluminum and iron. On the addition
of caustic soda in excess, followed by boiling, the uranium is precipitated as sodium
uranate. If an ore is treated which also carries vanadium, the latter element will also
be found in solution with the uranium. On the addition of caustic soda, the uranium
is precipitated, as already stated, and the vanadium remains in the filtrate. It is
practically impossible, however, in such a case to precipitate the uranium free from
vanadium. Commercially, the precipitate nearly always carries from 5 to 8 per cent
V2O5. The formula of sodium uranate is Na2U207, and it usually possesses 6
molecules of water of crystallization.
Concentration Methods.—Most uranium minerals can be concentrated

mechanically. This is largely due to the fact that the specific gravity of these
minerals is high, particularly pitchblende. Pitchblende can be concentrated
by the use of jigs, tables and slimers. On the table it appears as a black streak
representing the heaviest mineral present. Next to this is pyrite, and usually
a portion of the pitchblende is found in the pyrite, which constitutes a middling
which must either be reground and retreated, or handled chemically. As the
crystals of carnotite are extremely small, this mineral is not amenable to ordinary
concentration methods. It can, however, be concentrated in two ways: (1)
By dry methods; and, (2) by sliming. The most satisfactory dry method involves
the use of the ordinary Raymond pulverizing machine with tubular dust collector.
This has a beater chamber containing two sets of rapidly revolving beaters.
Over the chamber is a cone of galvanized iron, within which is an inner cone pro¬
vided at its lower end with a swinging discharge gate. The two cones are con¬
nected at their upper ends by a number of small gate shutters and have a common
top or cover plate. A large pipe from the center of the top leads to an ex¬
haust fan, mounted on the same shaft as the beaters. This fan discharges into a
cyclone dust and air separator which is connected by a return pipe with the
beater chamber of the pulverizer. A small pipe leads from this return pipe to a
tubular dust collector. The Raymond mill is used extensively to pulverize
cement, paints, etc. In the sliming method the ore is ground and then stirred
in water, getting as much attrition as possible in order to remove the carnotite
crystals from the silica grains. The latter settle quickly, and the carnotite with
other fine material is slimed off into another vat and allowed to settle. Undoubt¬
edly torbernite and autunite could be concentrated in the same way as carnotite,
but up to the present this has not been commercially carried out.
Uranium Compounds.—There are two well defined oxides, namely, UO2 and
UO3, and these combine to form intermediate oxides, such as U3O8. The dioxide
is basic and gives rise to uranous salts, the metal being tetravalent. The tri-
oxide is an acid-forming oxide yielding salts known as the uranates, correspond¬
ing to tungstates, chromates, etc. Uranium trioxide also gives a number of
derivatives in which only one of the three oxygen atoms is replaced by negative
groups, and these salts may be regarded as derivatives of the divalent compound
radical, uranyl (U02). U02 is formed by the reduction of UjOs by means of
hydrogen. It can be made directly from sodium uranate (Na2U207) by the
method of Parsons, which consists in fusing the oxide with salt containing a little
carbon. The fused product is leached and washed; everything being in solution
except the U02. U03 can be obtained by heating uranyl nitrate to 250°C., out
of contact with air, the oxide being left behind as a brownish-yellow powder.
U3O8 may be formed by Wohler’s method. The solution of the uranium ore
in dilute sulphuric acid is treated with hydrogen sulphide to remove arsenic,
antimony, lead, bismuth, etc.; filtered, and after the addition of nitric acid to
oxidize the iron, excess of ammonia is added. The precipitate consists of iron
hydroxide and ammonium uranate, and is digested with a strong solution of
ammonium carbonate with excess of ammonia in order to dissolve the ammonium
uranate. After filtering the liquid, it is allowed to cool, and ammonium uranyl
carbonate crystallized out. On igniting this, U308 is left behind. The alkali
uranates are obtained by precipitating a uranic salt with an alkali. The uranates
generally have the composition M20.2U02, and correspond to the dichromates.
They are yellow, insoluble in water, but soluble in acids. The two principal
uranates from the commercial standpoint are sodium diuranate, Na2U207, and
ammonium uranate (NH4)2U207. The sodium salt is known as uranium yellow,
RARE METALS 891
or yellow oxide. The hydrated salt contains six molecules of wafer of crystalliza¬
tion. The color of the sodium salt varies from a dark-yellow to a very bright-
canary-yellow. The color is dependent upon the conditions of precipitation, and
also upon the presence of excess of alkali. Ammonium uranafo is also yellow and
has the formula (NII.iAUA)?. On heating to redness, if yields pure HaOs-
Metallic Uranium.- -Metallic uranium was first prepared in the electric
furnace by Moissan.1 lie used 300 parts of U30« and 40 parts of sugar charcoal
and reduced the mixture in a carbon tube in Ike electric furnace. The crude
product contained carbon and was purilied by heating in a crucible, brasqued with
UaOg. This crucible was imbedded in another brasqued with titanium, the object
being to prevent action of nitrogen on the reduced uranium. The metal may also
be prepared by the electrolysis of the fused double chloride of uranium and sodium
in an inert atmosphere, using carbon electrodes.’2 Keeney3 has prepared the metal
by reducing U30h with carbon in a Siemen’s furnace with magnesite wails and
carbon bottom. The metal was allowed to cool in the furnace. He got a product
containing SS per cent uranium, 3.07 per cent carbon, and 2.47 per cent silicon,
which really consisted of uranium metal and uranium carbide.
Ferrouranium. Many unsuccessful attempts were made to produce
ferrouranium before success was finally achieved. The main difficulty met with
was that a low-carl>on alloy could not be made by variation of the carbon in the
charge because there was suck a low recovery of uranium, both when the amount
of carbon charged was about the theoretical amount required, and when the
carbon in the ferrouranium was low. Carbon could not be removed from an
alloy containing more than tin* allowable amount by using either iron oxide or
uranium oxide as can be done in the case of other ferro-alloys high in carbon.
When this was attempted the uranium in the alloy oxidized and went into the
slag leaving pig iron in the furnace. Keeney4 was unable to get a ferrouranium
containing as much as 30 per cent uranium, and in addition the carbon content
was too high. lie was, however, able to get a commercially high-grade uranium
metal as described above. Gillett and Mack6 succeeded in making a satisfactory
ferrouranium by using a pure U02, a low ash coke and a pure iron as raw materials,
with calcium fluoride as slag former. They stated that by using a tilting direct-arc
tyi>e furnace with water-cooled magnesite hearth and sides, it should he possible
to produce commercially without a second refining operation ferrouranium of any
desired uranium content, say 40 to 70 per cent, with carbon averaging below 2 per
cent, silicon below 0.75 per cent, vanadium below 0.5 per cent, and with alumi¬
num, sulphur, phosphorus and manganese all so low as to be negligible.
Uses of Uranium.'I insides its use in steel, uranium has been used for
many years as a coloring agent for glass. Sodium uranate is usually desired,
although uranium oxide is sometimes used as a substitute. The color produced is
opalescent yellow, turning green by reflected light. Usually about 20 per cent of
uranium calculated as oxide is required to produce the color; this making the
glass quite expensive. Uranium salts are also used in the ceramic industry.
It produces yellow, orange and black glazes. The coloring power is great and
1 Compt rrnd , I KM, V<»1. 116, p, 347.
3 Mwhiiav, Compt, rmuL Vol 122, p. 10BH.
8 Bull. A. I. M. IS., August,. HUH, 13f>4.
4 Bull. A. I. M. K„ August, HUH, j>. 1333,
only small quantities are required. Uranium salts have also been used to some
extent as mordants for silk and wool, and also a catalyst in the synthesis of
ammonia from nitrogen and hydrogen.1 In photography, uranium nitrate has
been used as a sensitizing agent for paper. With potassium sulphocyanide it can
be used for toning bromide prints.
Since Keeney and Gillett made ferrouranium and commercial uranium metal, a
considerable amount of interest has developed in connection with uranium steel.
One firm in this country has had ferrouranium on sale for some time and has been
cooperating with one or two steel manufacturers in the manufacture of uranium steel.
One part of uranium is supposed to replace from 2 to 3 parts of tungsten in high-speed
steels. Keeney2 states that ferrouranium can be added to steel in quantities up to 4
per cent with a uranium recovery of at least 50 per cent; and a recovery of 70 per
cent can be made in steel containing less than 2 per cent uranium. A considerable
portion of the carbon and silicon in the ferrouranium seems to enter the steel. For a
steel containing less than 2 per cent uranium, the ferrouranium can be added in the
ladle, but a higher percentage of uranium in the steel requires addition in the furnace,
or chilling will occur. These results were obtained in the electric furnace, but uranium
can also be made in the crucible furnace. Uranium steel has not come into general
use, although some of the steel makers claim that uranium is a satisfactory substitute
for tungsten in certain cases.
Analytical Methods for Uranium.—There are a number of meth¬

ods suitable for the analysis of uranium ores, and these are discussed at
length in Bulletin 70 of the U. S. Bureau of Mines. One of the best methods which
is particularly applicable to ores such as carnotite is as follows: Treat from 2 to
5 g. of ore, according to the proportion of vanadium, iron, and uranium present, in
a covered beaker, with 10 c.c. of HC1 and let it stand 15 min., shaking it occasion¬
ally. Add 5 c.c. of HNO3 and heat on a steam bath. When the solution is
quiet, remove the cover and evaporate to dryness. Add 3 c.c. of HC1 and 5 c.c.
of water to the residue and let it stand on the steam bath for a few minutes,
stirring occasionally. Dilute with 25 c.c. of hot water, filter into a small beaker,
and wash the residue with warm water. Some ores do not yield all the vanadium
to this treatment, a little remaining with the insoluble residue.
To make sure that all vanadium is in solution, ignite the residue in a platinum
dish, treat it with 5 c.c. of HF, and evaporate to dryness on a steam bath. Do not
bake the residue, for it is not necessary to expel all Si02. Add 3 c.c. of HC1 to the
residue from the HF treatment and evaporate to dryness. Repeat this treatment
to insure expulsion of HF. Treat the residue with 2 c.c. of HC1 and 2 c.c. of water
and stir with a glass rod until any red crust is dissolved, then dilute the solution with
water and filter it into the main liquid. Pass H2S into the liquid to separate copper,
lead, and other metals of this group, filter and boil the liquid to expel the H2S. Con¬
centrate the liquid to 100 c.c. if necessary, oxidize it with an excess of H202, and
then neutralize with dry Na2C03, adding 2 or 3 g. in excess. Boil the liquid for
about 15 min. until the yellowish uranium precipitate dissolves, leaving a brown
precipitate which is principally iron. Filter and wash the iron precipitate with
water, reserving the filtrate. Dissolve the iron precipitate in the least possible amount
of HNO3 (1:1), and add 10 c.c. of H202, neutralize with Na2C03, add an excess of 2
g. of Na2C03, and boil as before. Filter into the beaker containing the first filtrate.
1 Haber and Lb Rossignol, Zeit. Elektrochem., 1913, Vol. 19, p. 53.
2 Bull. A. I. M. E., August, 1918, p. 1366.
RARE METALS 893
The iron precipitate may contain a little vanadium—reserve it for further treat¬
ment. Evaporate the united filtrates from the iron precipitation to a volume of
about 200 c.c., add 10 c.c. of strong HN03 and boil until all C02 is expelled. Neutral¬
ize the free acid with ammonia (until a slight permanent precipitate appears), then
add 4 c.c. of HN03 for each 100 c.c. of liquid. Now add 10 c.c. of a 20 per cent lead
acetate solution, and enough of a strong solution of ammonium acetate to neutralize
the nitric acid present and substitute acetic acid for it. The object is to precipitate
the vanadium as lead vanadate in an acetic-acid solution. The ammonium-acetate
solution may be made by mixing 80 c.c. of strong ammonia, 100 c.c. of water, and
70 c.c. of acetic acid 99 per cent pure. Heat the liquid containing the lead vanadate,
precipitate on the steam bath for 1 hr. or more, filter on a tight filter, and wash with
warm water. Dissolve the precipitate in the least possible quantity of hot dilute
nitric acid, neutralize as before, add 3 c.c. of nitric acid in excess, add 2 c.c. of lead-
acetate solution and repeat the precipitation of lead vanadate by adding ammonium
acetate in excess, filter and add the filtrate to the one from the first precipitation of
lead vanadate.
Reserve the precipitate of lead vanadate for treatment as described below. Evapo¬
rate the united filtrates from the lead vanadate to about 400 c.c. Add 10 c.c. of
strong H2SO4 to separate the bulk of the lead (derived from the excess of lead acetate)
as PbS04, filter and wash the precipitate with cold water. Neutralize the filtrate
from the PbSC>4 with ammonia and add freshly prepared (NBU)HS until the solution
is yellow and the uranium and what little lead is present are precipitated as sulphides.
Warm the mixture on a steam bath until the sulphides settle well. Filter and wash
slightly with warm water. Dissolve the precipitate in a No. 2 beaker with hot
dilute (1:2) HN03, add 5 c.c. of H2SO4 and evaporate till fumes of H2SO4 appear, cool
and take up with water, boil, and let the small precipitate of PbS04 settle until
the solution is cold, filter the precipitate and wash it with very dilute H2SO4. Nearly
neutralize the filtrate with ammonia, have the solutions cool (not warmer than 30°C.),
and add powdered carbonate of ammonia in about 2 g. excess to precipitate the alumi¬
num. Let the precipitate settle over night, filter, and wash it with warm water.
If the precipitate is bulky or is at all yellow, dissolve it in a little dilute H2SO4 and
reprecipitate with carbonate of ammonia as described. Acidulate the filtrate from
the alumina with H2S04, and boil thoroughly to expel C02. Active boiling for M hr.
is required. Make the liquid slightly alkaline with NH4OH while it is hot, and heat
on the water bath until the ammonium uranate collects and settles. Filter and wash
with a very dilute (2 per cent) solution of (NH4)N03. Do not allow the precipitate
to run dry on the filter after the first washing. Dry the precipitate, ignite it in a
porcelain crucible, and weigh as U308. Dissolve the precipitate in HN03 and test
it with H202 for vanadium and with (NH4)2C03 for aluminum. Dissolve the lead
vanadate in dilute HN03, add 10 c.c. of H2S04, and evaporate the mixture to fumes.
Cool, take up with water, add 10 c.c. of a concentrated solution of S02 to the mixture,
boil until the excess of S02 is expelled, and titrate the hot solution with a standard
solution of potassium permanganate. The vanadium compound is reduced by S02
from V2Ob to V204- The iron equivalent of the permanganate solution multiplied by
1.6264 = the V206 equivalent.
It is not necessary to filter out the lead sulphate before boiling the mixture to expel
the S02. The boiling is best done in a large flask. The iron precipitate that was
produced by the addition of Na2C03 and H202 to the original acid solution may contain
vanadium. Ignite the precipitate in a platinum crucible and fuse the residue with
Na2C03. Leach the fused mass with water, filter, and acidulate the filtrate with
H2SO4. The filtrate may be added to the main solution before reducing with S02,
or reduced and titrated separately, as preferred. In expelling the excess of S02, it is
necessary to boil the liquid for at least 10 minutes after the smell of SO2 can no longer
be detected, or until the fumes do not decolorize dilute KMn04 solution. Some of
the factors used in calculations are: Fe value of permanganate times 0.9133 = V. Fe
value of permanganate times 1.6294 = V205. 1.78431V = V206. U = 0.84808U308
TUNGSTEN
(Atomic weight = 184)
Occurrence.—There are four reasonably common tungsten minerals found in

the United States. Ferberite, wolframite and hubnerite are closely associated as
regards their composition. Ferberite is usually classed as an iron tungstate,
although it may and often does carry moderate quantities of manganese. Wolf¬
ramite is an iron-manganese tungstate carrying usually, however, more iron than
manganese. Hubnerite is a manganese-iron tungstate containing more manga¬
nese than wolframite. These minerals form a complete series with almost an
infinite number of members between a pure iron tungstate and a pure manganese
tungstate.
Hess1 has suggested the following definitions: “Ferberite should be considered

as an iron tungstate (FeW04) contaminated by not more than 20 per cent MnWOi,
a proportion equivalent to 4.69 per cent MnO, or 3.63 per cent Mn in the pure tungsten
mineral. Hubnerite should be considered as managnese tungstate (MnW04) con¬
taminated by not more than 20 per cent FeW04, a proportion equivalent to 4.74
per cent FeO, or 3.69 per cent Fe. Wolframite should cover the ground between the
limits above indicated. That is, wolframite should be considered a mixture of iron
and manganese tungstates containing not less than 20 per cent nor more than 80
per cent of either. Except the light-colored htibnerites, most of these minerals can
not be detected by the eye or by simple tests, and in the absence of analyses it is,
therefore, convenient to refer to the dark minerals of the series as wolframites.”
Ferberite and wolframite when pure are black. When partly oxidized they may
appear brownish from the presence of iron oxide. Hubnerite is characteristically
brown, some specimens however, being yellowish, others reddish and some nearly
black. These minerals are all monoclinic in form. Ferberite has more of a tendency
to form well defined crystals than do the others. In hardness, all of these minerals are
a little over 5. Their specific gravities range from 7.2 or 7.3 in hubnerite to 7.5 in
ferberite, while the specific gravity of wolframite is intermediate between these
figures. Scheelite is a calcium tungstate (CaW04). In color it is usually white,
light-gray, or light-yellow, although its color may occasionally vary from greenish-
yellow to brown. Its luster is slightly glassy. It is found in granular masses, or
irregular lumps of a more or less coarse texture. Its hardness is a little less than 5,
and its specific gravity is approximately 6. It possesses four directions of good
cleavage. Other tungsten minerals of less importance are stolzite or lead tungstate,
cuprotungstite or hydrated copper tungstate, and tungstite or tungstic ocher, a
hydrated tungstic oxide.
Ferberite is found in several western states, but mainly in Colorado in the Boulder
district. This district has a length of 12 miles in a northeast and southwest direction,
and a width of about 7 miles, and begins about 6 miles due west of the town of Boulder.
The tungsten veins are in many ways similar to the gold veins of the district, and seem
to have a close connection with them in genesis. Hubernit-e and wolframite are
mainly found in New Mexico, Arizona, Nevada and South Dakota, although the
1 TJ. S. Geol. Surv. Bull, No. 652, p. 22.
RARE METALS 895
minerals are frequently met with in small quantities in other states. They do not
form as important a commercial source of tungsten as ferberite and scheelite. The
latter mineral is mainly found in California, principally near Atolia in northwest
San Bernardino County. A second area of considerable importance lies in Inyo
County, near Bishop. The Atolia-Randsburg district occupies an area of about 2H
by 10 miles in extent. Tungsten minerals are found to a very considerable extent in
foreign countries, namely, in Burma, Siam, the Malay States, Portugal, Boliva,
Argentina, Peru, England, China, Australia and New Zealand. During the last
few years, ores from these countries have competed to a very considerable extent with
United States ores.
Extraction from the Ores.—Up to recent years, most methods of treatment of
tungsten ores were small modifications of the original Oxland process which was
patented in 1847. This involves a fusion with sodium carbonate of the finely
ground ore previously leached with hydrochloric acid to get rid of traces of bis¬
muth if this metal is present. Sometimes a small amount of sodium nitrate is
added to the sodium carbonate. The mixture is placed in a reverberatory furnace
and sintered, care being taken not to fuse it. It is then removed and ground.
Frequently, this treatment does not give a sufficiently high extraction of the
tungsten, and involves a second sintering process. The sintered mass is leached
with hot water in tanks. Tungsten dissolves as sodium tungstate, while the
iron, calcium and most of the manganese remain insoluble. Some of the phos¬
phoric acid and silicic acid also dissolve as silico- and phospho-tungstates.
Traces of manganese may also dissolve, owing to the sodium nitrate present, as
sodium manganate. The solution of sodium tungstate may be either evaporated
to dryness, or evaporated to the point at which crystallization takes place. The
chief impurities are sulphates, silicates and arsenates of sodium, with traces of
iron and manganese. Iron, manganese and arsenic are precipitated by a small
quantity of caustic soda, and the greater portion of the sodium sulphate will
crystallize out before sodium tungstate. The silica remains in the mother liquor,
after the crystallization of the sodium tungstate, and causes some loss of tungsten
in the form of soluble silico-tungstates, which, however, can be used as
by-products for fireproofing purposes. The fairly pure sodium tungstate is redis¬
solved in boiling water and added to a boiling solution of hydrochloric acid made
up of 1 part of concentrated acid and 7 of water. Sometimes about 5 per cent of
nitric acid is added. The result is the precipitation of the tungsten as hydrated
tungstic oxide, which is filtered off and washed in filter presses. It is important
that the oxide should be washed free from sodium salts, as otherwise on drying
it has a greenish tinge which is not attractive to purchasers.
One of the best methods, if not the best, for producing a high-grade tungstic
acid is that of Ekeley and Stoddard (U. S. Pat. No. 1,255,144). This process is
used at one mill in Boulder County for the treatment of medium grade ferberite ores
and concentrates. Usually middlings from the mills, or ore running from 15 to 25
per cent WO3 are used. The product is a high-grade tungstic acid said to contain
less than 0.03 per cent phosphorus and 0.03 per cent sulphur, with only a small per
cent of silica. After drying the oxide runs from 99 to 99.7 WO3. The ore is mixed
with sodium carbonate and salt and the charge is placed in sheet-iron pans 2 by 4 ft.
and 6 in. in depth. This is placed in a firebrick furnace heated by oil burners. The
charge is so proportioned that a glass slag is obtained. The heat treatment requires
about an hour, at the end of which the pans are pulled out of the furnace onto an iron
rack, clamped and inverted, and the fused charge dumped into an iron receptacle
where it is allowed to cool. It is broken into pieces and ground in a crusher to approxi¬
mately J^-in. size. It is then mixed in a cement mixer with a hot-water wash liquor
from a previous run. The contents are dumped into a screen-bottom trough which
retains the coarse residue. This is shoveled back into the mixer and washed with hot
water; the wash water being used in the next run. The concentrated solution is
allowed to settle and the clear solution is given a secret treatment which removes the
last traces of the phosphorus, arsenic, silica and other deleterious substances. This
probably involves the use of magnesium chloride as a precipitant. After filtering,
the clear solutions are run into a large wooden tank, heated and calcium tungstate
is precipitated by the addition of a solution of calcium chloride. The precipitate is
allowed to settle, scraped out and treated with commercial hydrochloric acid in a
wooden revolving drum. After settling, the solution is drawn off through suction
filters. The residue is treated twice with hydrochloric acid in this manner and finally
the whole is dumped onto the suction filters. This treatment produces tungstic
acid as a precipitate, and the acid solution contains calcium chloride and calcium
sulphate. This is allowed to run into stone vessels where the calcium sulphate
crystallizes out, and the clear solution is drawn off into barrels where lime is added,
which produces calcium chloride for further use.
Acid-leaching methods are also used in connection with all the commercial tung¬
sten ores. They are especially applicable to getting rid of manganese and give good
results with scheelite. Ferberite, wolframite and hubnerite are soluble in hot con¬
centrated hydrochloric acid, but it is necessary to digest the ores in the acid for a
considerable time. Under such conditions the ores are decomposed and the manga¬
nese goes into solution while the tungsten is precipitated as tungstic acid and remains
with the insoluble residue. After filtering, the residue is extracted with ammonia
which dissolves the tungstic acid, giving ammonium tungstate. On ignition this is
decomposed giving tungstic acid and ammonia which may be recovered for further
use. Phosphorus is gotten rid of at some stage in the operation by precipitating
with magnesium chloride, or if tungsten powder is made from the tungstic acid, a
considerable amount of phosphorus can be leached out of the powder with hydro¬
chloric acid. Scheelite is soluble in hydrochloric acid, but one firm adds to the ease
of solubility by using hydrochloric acid and sodium chlorate. . The decomposition
of the ore takes place in acidproof earthenware pots in a hood with a steam draft.
After complete decomposition the tungstic acid and insoluble silica are filtered off.
By using high-grade scheelite the amount of silica present is not large, and the refin¬
ing of the product is not carried any further, but the acid without further removal
of the silica is used for the manufacture of tungsten powder.
Gin1 suggests a bisulphate fusion for the ore whenever tin is present. This
method, therefore,, does not apply specially to American ores, but is useful for some
of the foreign ores. The tin can be largely separated from the tungsten ore by means
of magnetic separation, but there is usually at least 1 per cent of tin oxide left behind
with the tungsten. The ore is decomposed by means of acid potassium sulphate in
a muffle furnace. After the sulphate is fused, the ore is thrown in and the mass is
stirred continually and the temperature increased until the whole mass is fluid enough
to run out of the furnace. After solidification the fused mass is ground and treated
with water, which dissolves the soluble sulphates and phosphoric acid and leaves
insoluble potassium acid tungstate as a white amorphous precipitate. About 50
per cent excess of bisulphate over that theoretically required is actually necessary.
The insoluble residue is dried and treated with a warm solution of ammonium car¬
bonate, or cold ammonia water in which carbon dioxide is passed. The potassium
1 Trans. Am. Electrochem. Soc., Vol. 13, p. 481.
RARE METALS 897
acid tungstate dissolves leaving the silica, the cassiterite and the insoluble sulphates.
The solution is evaporated to crystallization, which gives ammonium tungstate from
which tungstic oxide is made by heating.
Concentration Methods.—The tungsten minerals are all very amenable to

concentration methods, the methods used involving jigs, tables and slimers.
A good many mills also have “rag plants/7 instead of slimers, or work the two
together. The average grade of concentrate that was demanded up to 2 or 3
years ago was around 60 per cent. Recently (1918) concentrates of 55 per
cent were classed as high-grade. One of the difficulties in connection with milling
operations is the handling of second-jig and low-grade slime products. These
to a certain extent have been taken care of by a chemical treatment plant.
Where this is not used, the product, of course, is often reground and reconcen¬
trated, but this always involves considerable loss. Mill operations have been
improved very much during the last 5 years, and an average recovery of 85 to
87 per cent is not now uncommon.
Tungsten Compounds.—Tungsten forms two definite oxides, namely, tung¬
sten dioxide (WO2) and tungsten trioxide (WO3). These combine together to
form more complex oxides. The dioxide is formed when a current of hydrogen
is passed over the trioxide at dull-red heat. Under these conditions, different
colored oxides are obtained, including the dioxide. It may also be prepared
in the wet way by reducing the trioxide by means of zinc and hydrochloric acid.
The dioxide is a brown powder of specific gravity of 12.1. It is strongly pyro¬
phoric and must be cooled in hydrogen before exposure to air. It is slightly
soluble in hydrochloric acid and sulphuric acid, giving purple solutions. It is
soluble in sodium and notassium hydroxides giving the alkaline tungstates with
the evolution of hydrogen. Tungsten trioxide is prepared by heating ferberite
or scheelite with concentrated hydrochloric acid for a considerable time, as already
described under “extraction from the ores.77 It can also be prepared by adding
hydrochloric or nitric acids to a solution of sodium or potassium tungstate,
the hydrated oxide (tungstic acid) being precipitated. On ignition this gives the
trioxide. It is a bright canary-yellow colored amorphous powder which becomes
dark-orange on heating, but regains its color on cooling. A small amount of
sodium salt gives a greenish tint to the oxide, which is difficult to remove. It
slowly becomes greenish on exposure to light. It can be obtained in the crystal¬
line state by igniting a mixture of sodium tungstate and carbonate in a current
of hydrogen chloride when the trioxide is obtained in olive-green prisms. The
specific gravity of the oxide is 6.34. The trioxide is an acid-forming oxide and
yields two tungstic acids, namely, the normal acid (H2WO4) and metatungstic
acid (H2W4O13). In addition to the salts corresponding to these acids, a large
number of other tungstates have also been prepared, of which the following are
a few:
M2WO4 = M2O.WO3
M2W2O7 = M2O.2WO3
M^WÔio =: M2O.3WO3
M2W4O13 - M2O.4WO3
M2WfiOi6 = M2O.5WO3
M2W6019 = M2O.6WO3
57
Tungstic acid (H2WO4) is formed as a white precipitate when a soluble tungstate

is treated with an acid in the cold. If, however, an excess of hot acid be used, H2W0*
separates as a yellow powder, insoluble in water and all acids except hydrofluoric.
Metatungstic acid (H2W4O13.7H2O) is prepared by decomposing the barium salt with
dilute sulphuric acid. It crystallizes in small yellow octohedra which lose their water
of crystallization at 100°C., and on ignition are converted into the trioxide. The
crystals are soluble in water. Normal sodium tungstate (Na2W04.2H20) is prepared
by fusing the tungsten minerals with soda ash and treating with water. It crystallizes
as thin rhombic prisms, soluble in four parts of cold and in two parts of boiling water.
Heated to 200°C., sodium tungstate becomes opaque and loses its water. It fuses
at a red heat. Sodium metatungstate (Na2W4013.1OH2O) is formed by boiling the
normal salt with tungstic acid for a considerable period. It crystallizes in octahedra
belonging to the regular system. It is soluble in about 10 times its weight of cold
water.
Metallic Tungsten.—Metallic tungsten can be made in a number of ways.

The purest product is obtained by heating a very high-grade tungstic oxide in a
current of hydrogen. The oxide can also be reduced by means of carbon or
metallic zinc. Metallic aluminum and magnesium also can be used instead of
zinc. Most metallic tungsten prepared commercially is in the form of powder
for making tungsten steel. Reduction with carbon is the method usually used
as the reaction goes readily and presents few difficulties. The reduction can
take place either in a steel tube or in a covered steel crucible. Davis1 has care¬
fully studied the conditions and temperatures necessary for the reduction of
tungstic oxide to the metal. He finds that at 650 to 850° the oxide goes to a
black or purple on heating with carbon; at 950 to 1050°C. a chocolate-colored prod¬
uct is the result (probably W02); and at temperatures above 1050°C. the gray
powdered metallic tungsten results. The ratio of tungstic oxide to carbon
varies from 10:1 to 10:1.6 depending upon the process used, the temperature of
reduction and the time involved. A small excess of carbon can be partially
removed by washing. Both fireclay crucibles and iron tubes give satisfactory
results, and a product of over 98 per cent metallic tungsten can be produced.
Whereas the reduction to metal is complete at 1050°C., it is better to keep the
temperature around 1,100°C.
Davis also found that the temperature of reduction with hydrogen at ordinary
pressures is much the same as with carbon. The properties of metallic tungsten are
so unique that when it was first prepared it led to a search for methods for producing
the metal in a ductile form. It was found that this could be done by repeated heat-
mg, rolling, hammering, drawing and swaging. Cast tungsten, or tungsten powder,
era widely in its properties from the pure raw material, such as is used in the manu-
aeture of incandescent lamp filaments. The tungsten powder is brittle, crystalline
and hard and has a specific gravity of 16 to 17. The pure metal is softer and tougher,
?al* ® w ded a yellow heat, and even drawn into fine wire. Its specific gravity
is 18 8, and the melting-point close to 3,270°C. A wire of 5 mm. in diameter has a
e s ength of more than 450,000 lb. to the square inch. At a red heat, the metal
K^1 i tf 6 a^r* -^^d sulphur and phosphorus attack it slowly; potassium
T, ? ‘ cau^c alkalis and fused nitrates attack the metal, but solutions of these
salts have practically no action on it. Hot dilute sulphuric acid affects it only slightly,
e concen rated acid dissolves the metal slowly. At ordinary temperatures
lJ<mr. Ind. Eng. Chem.f Vol. 11, No. 3, p. 201. *

RARE METALS 899
hydrochloric acid of any concentration has practically no action; but the strong boiling
acid dissolves it slowly. Concentrated nitric acid and hydrofluoric acid have very
little action on tungsten, but the two acids when mixed attack it fairly rapidly.
Ferrotungsten.—Tungsten is the principal rare metal used for alloy purposes
in connection with the manufacture of high-speed steels. In making this steel,
either ferrotungsten or tungsten powder may be used. The annual production of
the two is probably about equal in quantity. When ferberite is used, practically
all ferrotungsten is made directly from the high-grade concentrate without previous
chemical treatment of the ore. The product obtained in such a case contains
usually from 70 to 85 per cent tungsten, 0.5 per cent carbon, 0.4 per cent silicon,
0.5 per cent manganese, 0.01 per cent sulphur, and 0.02 per cent phosphorus.
Ferrotungsten has a high density, fine gray fracture and is not crystalline.
Keeney1 gives a detailed description of the manufacture of ferrotungsten. The
reduction with carbon goes readily. The slag may contain as high as 8 per cent FeO,
but less than 1 per cent W03. Theoretically, the reduction of 100 parts of alloy
from 122 parts of ferberite requires 16 parts of carbon, but practically about 25 per
cent excess of carbon is charged. Small amounts of lime and fluorspar are used to
flux the silica. Operating in this manner, with an excess of carbon, a product is
made containing 3 per cent carbon, 70 per cent tungsten, 0.05 per cent phosphorus
and 0.01 per cent sulphur, with slag containing below 1 per cent W03. The analysis
of a typical concentrate for the manufacture of such a product is W03, 60.36 per cent;
Fe, 22.0; Si02, 8.0; Mn, 0.05; S, 0.35 and P, 0.05 per cent.
A typical operation involves charging into the furnace 65 lb. of a mixture composed
of 200 lb. concentrate, 42 lb. coke, 56 lb. lime and 6 lb. fluorspar. Three more 65-lb.
charges are added at intervals of hr., and at 2J-£ hr. from the start the furnace is
tilted and the slag poured. This cycle is repeated until a 1,200-lb. button has been
formed, requiring 24 to 36 hr. The furnace is allowed to cool, is torn down and the
button of metal removed, cleaned and broken up. This crude metal is refined as
follows: A charge of 150 lb. metal and 75 lb. ferberite concentrate are smelted for
K hr., when 12 lb. of fluorspar are added. , After another 3 hr. the slag is poured and
a fresh charge is started. The process is continued for from 36 to 48 hr. until a
button weighing 1,500 lb. has been formed, when the furnace is allowed to cool, is
torn down and the button of metal removed, cleaned and broken. Ferrotungsten
containing less than 1 per cent carbon can be made in a single smelting operation by
careful regulation of the carbon in the charge and the use of an acid slag. The prod¬
uct, however, is less pure and the slag loss is higher. In addition, the metal will
contain higher phosphorus and sulphur.
Uses of Tungsten.—The main use for tungsten is in the manufacture of steel,
especially high-speed steel. Such steel is generally made by the crucible process,
although not always. A steel suitable for such use may contain as high as 15
or 20 per cent tungsten, and as low as 8 or 10 per cent tungsten, and 4 per cent
chromium. Vanadium and molybdenum may also be used in small quantity,
and uranium has in some special steels partially replaced the tungsten. Tung¬
sten is also used in self-hardening steels, or those requiring no tempering after
forging. They may have a composition between the following limits: Tungsten,
2.4 to 3.4 per cent; chromium up to 6 per cent; carbon 0.4 to 2.2 per cent and
silicon 0.2 to 3 per cent.
The chief property which tungsten gives to steels is tensile strength, which, with
the elastic limit, increases within certain limits as the percentage of tungsten increases.
1 Buzz. A. I. M. E., August, 1918, p. 1338.
Elongation and resistance to shock diminish proportionately, and the hardness in'
creases fairly consistently with the per cent of tungsten. Tungsten is also used for
filaments in electric lamps. For making the filaments the “squirting” process was
originally employed. This involved mixing metallic tungsten powder with an organic
binding material, such as gum, and these were squirted into the filament. The car¬
bon was eliminated by placing the filament in an atmosphere containing a volatile
compound of tungsten, such as the oxychloride, and a small quantity of hydrogen.
On heating the filament by an electric current, the carbon was replaced by tungsten.
Another method consists in mixing metallic tungsten powder with an amalgam
containing equal amounts of cadmium and mercury and squirting the mixture through
a die in the usual way. The cadmium and mercury in the filament are volatilized by
heating and the filament retains the usual brittleness, but after continued heating
at higher temperatures the filament becomes pliable and can be bent. Drawn fila¬
ments are now used; the drawing being made possible by hammering, heating, swag¬
ing, etc. Tungsten is also a constituent in the so-called “bronze powders” employed
for decorative purposes, and has also been used to a small extent in the fireproofing
of cloth and other fabrics, as a mordant for silk, and for coloring glass and porcelain.
Analytical Methods for Tungsten.—Tungsten in ores can be determined

quantitatively by the following method: The ore ground as finely as possible in
an agate mortar is fused with from 5 to 10 times its weight of sodium-potassium
carbonates and extracted with hot water. If any gritty particles remain, another
fusion is necessary. An equal volume of concentrated hydrochloric acid is
added to the aqueous solution of the alkaline tungstate, and the whole evaporated
to dryness, after which the silica is dehydrated by heating at 120°C. for an hour.
The residue is moistened with hydrochloric acid, taken up with water and boiled;
then filtered and washed with 5 per cent HC1 or ammonium-nitrate solution.
The filtrate contains a small amount of tungsten, which is determined by heating
the filtrate and the washings to boiling and adding 5 to 6 c.c. of a cinchonine solu¬
tion made by dissolving 25 g. of cinchonine in 200 c.c. of 1:1 HC1. This is allowed
to digest on a hot plate while the other operations are being performed.
The precipitated tungstic oxide is dissolved in hot ammonia solution (Watt's

solution of 200 c.c. strong ammonia, 1,000 c.c. water, 10 c.c. IIC1), filtered into a
platinum dish, the filtrate being washed with the ammonia solution and evaporated
to dryness and ignited. The cinchonine precipitate is filtered and washed with hot
dilute cinchonine solution (100 c.c. water to 5 to 6 c.c. of the above solution), dried
and ignited in the platinum dish with the tungstic oxide. A few drops of sulphuric
acid and a little HF are added, and the SiC>2 driven off at a dull red heat. The tung¬
sten is weighed as WO3. Small quantities of impurities may be removed by treating
the W03 with the ammonia solution, filtering and again igniting.
Ferrotungsten can be analyzed by the following method which is used at the
laboratories of the Tungsten Products Company at Boulder, Colo.
The refined ferrotungsten, crushedând rolled to pass a J^-in. mesh screen, goes
to a machine which automatically mixes, samples and sacks the product for shipment.
The sampler is of type developed by this company, which cuts- one-fifth of the total
mix. This cut passes through the machine again after being rolled to approximately
34-in. mesh. If the original mix is very large the operation is repeated to bring the
weight of the sample down to less than 500 lb. This final sample from the automatic
is passed repeatedly through a riffle sampler until the last cut weighs 15 to 50 lb.
This goes to the laboratory sampling department where it is reduced in a steel mortar
to pass a %q-in. screen. The pounding necessary for this process is done by a 2-in.
RARE METALS 901
Ingersoll Rand piston drill set up vertically in a frame above the mortar and pestle
in such a manner that it is used as an air hammer. The sample is thoroughly mixed
and split twice, and one-quarter is then further reduced to pass a 10-mesh screen.
This procedure of screening, mixing and splitting is continued through a series of
screens until the last portion passes the 200-mesh screen giving the analyst about
100 g. of finely powdered metal. The operation of reducing a 20-lb. sample received
from the smelter to a 100-g. laboratory pulp of 200-mesh powder by this system
requires about 1 hr.
For the determination of tungsten, two 1-g. samples are mixed separately in nickel
crucibles with about 5 g. sodium peroxide and a cover of sodium carbonate. Fusion
is done in an electric muffle at about 800°C. which requires 6 to 7 min. to make a
perfect decomposition. Cool, leach out in water in covered 400 c.c. beakers. Re¬
move crucibles and acidify carefully with hydrochloric acid using a moderate excess.
Add about 2 c.c. nitric acid or H2O2 to assist in dissolving any small fragments of
nickel oxide scale and bring to a boil. Most of the tungsten separates as H2WO4
and all iron and nickel salts are in solution. Add about 40 c.c. cinchonine solution,
stir and let stand in a warm place for about 4 hr. or preferably over night. This
precipitates all the remaining tungsten as cinchonine tungstate, and it has proved to
be just as effective a reagent for this purpose as quinine hydrochloride. Filter off the
settled precipitate through a 15 cm. close weave paper arranged over a platinum cone
and suction flask. Wash thoroughly with warm dilute cinchonine wash water
using gentle suction. No trouble will be experienced due to the precipitate sticking
to the beaker if a few drops of hydrofluoric acid are used.
The washed precipitate is transferred to a shallow gold dish of known weight and
ignited to W03 in the muffie. Cool, moisten with 10 or 12 drops of hydrofluoric
acid to remove any traces of Si02 which might be carried down, dry and ignite again.
Weigh and calculate to W. The factor is 0.793. Duplicate determinations carried
out in this way should check to within 0.10 per cent. The proportions used in the
cinchonine solution are 50 g. of cinchonine alkaloid dissolved in 2 liters of cold water
containing 150 c.c. hydrochloric acid; for the cinchonine wash water, 100 c.c. of above
cinchonine solution and 50 c.c. hydrochloric acid diluted to 11. with hot water and
used from a bulb wash bottle is convenient. For the determination of carbon—
f errotungsten and tungsten metal powder burn completely and readily in oxygen—and
carbon is easily determined in any type of combustion train. In the laboratory
mentioned they prefer to catch the C02 in a Meyer bulb in 2 per cent barium-hydrate
solution, to filter, wash and weigh the BaC03.
The usual fusion-oxidation methods of determining sulphur are all open to objec¬
tions when applied to a ferro alloy or other material whose sulphur content is below
0.05 per cent. The blank is always high, due to traces of sulphates in fluxes and
reagents used and the danger of picking up fumes or traces of sulphuric acid from the
apparatus of a general laboratory. The sulphur content of ferrotungsten occurs as
sulphides which, in a finely ground sample, are completely decomposed yielding H2S.
This suggests at once the application of the method which has been used successfully
for 2 years. Two to five grams of 200-mesh ferrotungsten or metal powder are weighed
into an evolution flask. Cover with 50 c.c. water and add 25 c.c. strong hydrochloric
acid by way of the separatory funnel. Heat to boiling and boil for about 10 min.
catching the evolved gas, air and steam in ammoniacal cadmium-chloride solution.
The usual yellow precipitate of cadmium sulphide shows the presence of sulphides
in the alloy. Wash the cadmium-chloride solution from the bulbsj then cool, acidify
with hydrochloric acid and titrate at once against a weak iodine solution (1 c.c. =
0.0005 g. sulphur) using starch as an indicator. The iodine solution is standardized
by using a Government standard steel pf known sulphur content, or by using a ferro¬
tungsten of known sulphur content. Boiling with dilute hydrochloric acid does not
dissolve the alloy, but does decompose the sulphides present as may be shown by
filtering off the contents of the evolution flask after the operation and examining for
sulphur by a fusion method or by a combustion method.
The removal of phosphorus from the alkaline liquor after a fusion decomposition
of ferrotungsten presents certain difficulties. Precipitation with magnesia mixture
from dilute solutions is incomplete unless much time is allowed for this step, also
precipitation from concentrated solutions either as magnesium-ammonium phosphate
or as aluminum phosphate tends to drag down some tungsten, involving another
step for its removal. To avoid these difficulties the use of uranium acetate has been
introduced at this stage. The precipitation is quantitative from moderately dilute
solutions and no dragging down of tungsten occurs. Weigh 1 g. sample into a nickel
crucible containing 6 to 8 g. chemically pure Na2(>2. Mix thoroughly and cover
with about 1 g. of chemically pure Na2C03- Ignite in a muffle to complete
fusion of the fluxes. Cool and leach out carefully with warm water in 250 c.c. covered
beaker. Remove crucible and partly neutralize the NaOH with 25 c.c. HC1 (1-3) to
permit filtering without dilution. Stir and then allow the dense Fe(OH)3 to settle
out while standing on a warm plate. Decant off through a 12 J^-cm. qualitative paper
into a 600-c.c. beaker. Wash precipitate into the filter, allow it to drain thoroughly
then give it two good washes with hot water stirring up the precipitate well with a
fine jet. Reserve this precipitate of Fe(OH)3 which contains a small amount of
phosphorus, to add to the uranium precipitate obtained from the tungsten solution.
Acidify the filtrate of sodium tungstate and sodium phosphate with acetic acid
using 30 c.c. of 40 per cent acetic. Add 2 c.c. of 5 per cent uranium-acetate solution.
Boil off all CO2. Make just alkaline with NH4OH (1:1). This will require about
20 c.c. The precipitate is ammonium-uranyl phosphate. Add 2 c.c. more uranium
acetate which precipitates at once as uranium hydroxide and serves to drag down the
last traces of phosphate, as well as to assist in filtering the ammonium-uranyl phos¬
phate. Boil off any large excess of NH4OH. Filter through a qualitative paper and
wash twice with hot water to remove tungsten. Place the precipitate in the same
beaker together with the iron precipitate obtained above, add 30 c.c. water and 25
c.c. strong nitric acid. Boil until both precipitates are in solution and the filters
reduced to pulp. Filter into a 500-c.c. Erlenmeyer flask washing the pulp once with
hot water. Add 5 c.c. of 5 per cent KMn04 solution to acid filtrate. Boil to oxidize
all phosphorus to H3PO4. Clear of Mn02 by adding 2 c.c. H2O2 (1:1). Boil to
remove excess H202 and cool under the tap. Add 45 c.c. NH4OH (1:1) which will
nearly neutralize the nitric acid, then 50 c.c. of molybdate solution. Stopper the
flask and shake for 5 min., and let stand at least 15 min. before filtering. Filter through
a 9-cm. filter. Wash with acid ammonium-sulphate wash solution to remove all
nitro-molybdate. Dissolve the yellow phospho-molybdate in hot dilute ammonia
water which must be free from chlorides, into the same Erlenmeyer flask. Wash the
paper well with alternate washes of ammonia water and hot water. When cool,
acidify with 5 c.c. of H2S04 (1:1) and pass through the reductor, following with a
thorough wash of cold water. Titrate against standard KMn04 solution to the same
end point as used in determining the reductor blank. The iron value of the standard
permanganate solution times 0.0163 equals the P value. The ammonium-sulphate
wash solution must be distinctly acid. A good formula is: H2S04 (1:1) 100 c.c.; NH4-
OH (1:1) 60 c.c.; water, 21. If convenient it is well to allow the precipitate of yellow
phospho-molybdate to stand about 1 hr. after shaking for 2 min. A blank on all
reagents including a fusion of Na202 and Na2C03 should be carried through with the
determinations. It is sometimes desirable to add a measured quantity of dilute
phosphate solution of known P content to the blank for close work. Checks should
agree within 0.015 per cent P.
RARE METALS 903
MOLYBDENUM
(Atomic Weight = 96)
Occurrence.—The chief commercial minerals of molybdenum are: Molybde¬

nite (MoS2); wulfenite (PbMoO*) and molybdite (Fe203.3Mo03.7%H20). Of
lesser importance are ilsemannite, possibly (MoO3.SO3.SH2O), belonesite (Mg Mo-
04), and powellite (CaMoO*). The first three are the only ones from which
molybdenum is obtained commercially. Molybdenite contains 59.5 per cent
molybdenum and 40.05 per cent sulphur. It is a soft, opaque, lead-gray mineral
with metallic luster. It frequently occurs in flakes or scales resembling some
micas as regards its cleavage. In hardness it ranges from 1 to 1.5, and readily
leaves a mark on paper of a bluish-gray color. ‘Its specific gravity is from 4.7 to
4.8. The mineral is widely distributed throughout the United States, especially
in the West, and whereas it is found in a large number of localities in small
quantities, there are also several large deposits which are being worked
commercially.
The most important of these is at Climax, Colorado, near Leadville. Here a large
portion of a mountain is impregnated with the mineral, and whereas the grade aver¬
ages less than 1 per cent M0S2, the amount of ore available is very large. Another
large deposit is on Red Mountain, near Empire, Colo., about 50 miles due west of
Denver. There are two mills at Climax and one at Red Mountain. Smaller deposits
of molybdenite are found in several places in the United States, especially in Arizona,
New Mexico, Nevada, California, etc. Wulfenite contains theoretically 26.15 per
cent of molybdenum and 56.42 per cent of lead. It is a brittle, heavy, semi-trans¬
parent mineral with resinous lustre and is generally of a wax or orange-yellow color,
although occasionally it is olive-green, yellowish-gray or even brown. Its hardness
is 2.75 to 3, and its specific gravity is 6.7 to 7. It crystallizes in a tetragonal system,
and the crystals are commonly square and tabular, and are sometimes very thin.
Less frequently, they are octahedral or prismatic. Like molybdenite, wulfenite is
widely distributed in the western states; probably the largest known deposit is at the
Mammoth mine, Mammoth, Arizona, from which a very considerable tonnage of
concentrates has already been obtained. It is, however, found in numerous other
places in Arizona, Nevada and to some extent New and Old Mexico. Molybdite is a
hydrous ferric molybdate, lemon-yellow to pale-yellow in color, and occurs as an earthy
powder, usually as incrustations. Molybdite is an alteration product of molybdenite,
probably formed by the interaction of molybdic acid and limonite; the molybdic acid
being an oxidation product of the molybdenite.
Molybdite, therefore, is frequently found associated with molybdenite, especially
where the latter mineral has had a chance to weather. Molybdenum minerals are
found in notable quantities in several foreign countries. Molybdenite is found
in Queensland, New South Wales, Norway and Canada. It has only been during
the last 3 years (written 1919) that the United States has produced concentrates of
molybdenum minerals in commercial quantities. Up to that time practically all of
the world's production came from Queensland, New South Wales and Norway.
Extraction from the Ores.—As ferromolybdenum, which is the chief use for
molybdenum, can be made directly from molybdenite concentrates, it is not
necessary to treat molybdenite chemically unless molybdic oxide or ammonium
molybdate are required for chemical purposes. In such a case, the molybdenite
may be roasted to oxide, treated with ammonia which dissolves the molybdic
oxide as ammonium molybdate, filtered and evaporated to dryness, or to crys¬

tallization of the ammonium molybdate. On ignition the ammonia is driven off
and may be recovered for further use. Doerner (U. S. Pat. 1,299,560) gives an
oxidizing roast to a mixture of molybdenite and salt. The presence of the sulphur
in the charge, together with the oxidizing atmosphere which is maintained during
the treatment, results in the liberation of free chlorine in two stages. In the first
stage, the sulphur in the charge is converted by the oxidizing roast into sulphur
dioxide, and in the second stage, the sulphur dioxide together with additional
oxygen reacts with the chloride to liberate chlorine. This reacts with the
molybdenum, forming volatile compounds, either the chlorides, oxychlorides or
mixtures. The chlorides and oxychlorides pass off to a condensing chamber and
are decomposed into molybdic oxide and hydrochloric acid by the action of steam
or water.
Wulfenite is a more difficult metallurgical problem than molybdenite. This
mineral nearly always carries small quantities of vanadinite, and the two can not be
separated mechanically. Kissock first used a sodium carbonate fusion method in a
small blast furnace. The wulfenite mixed with sodium carbonate and coke was
strongly heated and the lead was obtained as metallic lead, while the molybdenum
went into the slag in the form of sodium molybdate. This slag was used directly in
the electric furnace for the making of ferromolybdenum, but the excess of sodium
carbonate had very decided and deleterious effects on the furnace linings. Holladay
(U. S. Pat. 1,281,961) prepares a sodium molybdate slag reasonably free from lead and
other impurities, then leaches the slag with water, and adds dilute calcium-chloride
solution in the cold. This precipitates most of the impurities which are still present,
while the molybdenum stays in solution. On filtering and boiling and adding excess
calcium chloride, the molybdenum is precipitated as Calcium molybdate which
can be used for the preparation of molybdic oxide, or can be used directly in the
making of molybdenum steel. Conditions for the precipitation of calcium molybdate
have been studied carefully by Bonardi.1
Concentration Methods.—Molybdenite can be concentrated readily by flota¬
tion methods. An ore carrying less than 1 per cent can be concentrated to a
product carrying 60 or even 70 per cent molybdenite. Usually mixtures of
kerosene and pine oil are used although other oils give satisfactory results on
certain ores. Wulfenite can be concentrated by ordinary milling methods, using
tables and slimers. The mineral crushes readily, but an ore carrying as little as
0.5 per cent M0O3 as wulfenite can be concentrated with a fair recovery. Vana¬
dinite which is frequently associated with it can not be separated to form very
satisfactory products as the specific gravities of the two minerals are almost the
same.
Molybdenum Compounds.—The molybdenum oxides are the sesquioxide
(M02O3), molybdenum dioxide (M0O2) and molybdenum trioxide (M0O3).
The first two oxides are basic and little is known concerning them except their
halogen derivatives. The trioxide is the important oxide and is not only used
commercially but also gives rise to the molybdates. It is a white powder which
on heating becomes yellow. It melts at a red heat to a dark-yellow liquid and
this on cooling solidifies to a yellowish-white crystalline mass with a specific
gravity of 4.39. It volatilizes at high temperatures when heated in closed vessels,
but sublimes readily in the air, giving small transparent rhombic tables. It is
1 Chem, and Met. Eng.t Sept. 15, 1919.
RARE METALS 905
soluble in 500 parts of cold water and 960 parts of hot water; the solution being
slightly acid to litmus.
Molybdic acid (H2M0O4.H2O) may crystallize out when a solution of ammonium

molybdate is treated with nitric acid. The addition of a crystal of the compound
renders the separation more easy. Normal sodium molybdate (Na2Mo04) is formed
by fusing the trioxide with sodium carbonate. It crystallizes from water in acute
rhomb ohedra containing 2 molecules of water. Below 6° it contains 10 molecules of
water of crystallization. The dimolybdate (Na2Mo207) is formed when sodium
carbonate and the trioxide are fused in the correct proportions. The tetramolybdate
(Na2Mo4Oi3.11H20)J the octomolybdate (Na2Mo8025.4H20) and the decamolybdate
(Na2Moio03i.l2H20) are of lesser importance. Normal ammonium molybdate,
(NH4)2Mo04} is produced when molybdenum trioxide is heated with excess of con¬
centrated ammonia. It crystallizes in four-sided prisms which are decomposed
by water. If a solution of the trioxide in ammonia is evaporated the compound
(NH4)6Mo7024.4H20 crystallizes out in monoclinic crystals. This is the salt whichis
usually known as “ammonium molybdate.”
Metallic Molybdenum.—Metallic molybdenum may be prepared by reducing

molybdic oxide with aluminum powder or by heating the trioxide or one of the
chlorides in a current of hydrogen. It may also be prepared by heating a mixture
of molybdenum trioxide with Jio of its weight of sugar charcoal in an electric
furnace in a carbon crucible. The pure metal is malleable and is not hard enough
to scratch glass. It has a specific gravity of about 9 and can be forged when hot.
It oxidizes readily at a dull red heat and is attacked by fused potassium chlorate,
nitrate, etc.
Ferromolybdenum.—The raw materials for ferromolybdenum may be either
molybdenite, molybdic oxide, a sodium-molybdate slag, or calcium molybdate.
The reducing agent is usually carbon although 90 per cent silicon material ground
to 60 mesh has been used. Lime and fluorspar are used as fluxes. The reaction
that takes place with molybdenite is as follows:
2MoS2 + 2CaO + 3C = 2Mo + 2CaS + 2CO + CS2.
Reduction with silicon metal gives the following reaction:
M0S2 + Si = Mo + SiS2.
According to Keeney1 for about 100 parts of molybdenite, 58 parts of lime are
necessary for slagging the sulphur as calcium sulphide. The reaction works
closely to the theoretical, and there is no difficulty in making a product with
about 0.1 per cent sulphur and from to 3 per cent carbon. With sodium-
molybdate slag, the reaction is as follows:
Na2Mo04 + 3C = Mo + 3CO + NasO.

The reduction of sodium molybdate requires considerably more power than the
reduction of the sulphide or oxide. The average power consumption is 7 to 7%
kw. -hr. per pound of molybdenum produced. The recovery varies from 78 to 80
per cent with a loss of 10 per cent in the slag and 10 per cent mechanically by volatili¬
zation. Ferromolybdenum containing 80 per cent molybdenum and under 1 per
cent carbon can not be regularly tapped from the electric furnace.
1 Bull. A. I. M. E., August, 1918, p. 1334.
Uses of Molybdenum.—The main use of molybdenum is in making molyb¬

denum steel. Undoubtedly, the introduction of molybdenum into steel increases
the elastic limit without diminishing its ductility. Molybdenum also may
replace tungsten in high-speed steels. Up to recently, only from 0.5 to 2 per
cent molybdenum was used, but recently the amount has been considerably
increased. High-speed molybdenum steel can be welded with difficulty and
is quite brittle on forging. Magnetic steel contains from 2% to 3 per cent molyb¬
denum and about 1 per cent carbon. A chromium-molybdenum steel has been
used recently quite extensively in connection with certain automobile parts, espe¬
cially where resistance to heavy wear is required, such as in bearings, etc. Such
a steel contains 0.2 to 0.6 per cent Mo. Ammonium molybdate is used exten¬
sively in chemical laboratories. Some salts of molybdenum are also used for
coloring glazes in porcelain and to some extent in coloring silks, leather and
rubber.
Analytical Methods for Molybdenum.—The following method has been

worked out by Messrs. Bonardi and Barrett of the U. S. Bureau of Mines:
Weigh 0.25 to 5.0 g. of the finely pulverized ore into a 250 c.c. Erlenmeyer flask.
Add 15 c.c. of HN03 and heat until the brown fumes are gone. This will be
found to be the most rapid manner of decomposing the sulphides present. Care¬
fully add 5 to 10 c.c. of HC1 and heat about 20 min, or until the ore is completely
decomposed. Wash down the sides of the flask with hot water and heat about
5 min. longer. Cool. Add 5 to 10 c.c. excess NH4OH, heat a few minutes and
filter hot, washing with hot water. If the residue of insoluble, iron and aluminum
hydroxides is large, it is advisable to puncture filter and wash residue into flask
in which decomposition was made, add sufficient HC1 to dissolve the precipitated
hydroxides, make slightly ammoniacal, heat and filter into the same beakers
used for the first filtration. Wash well with hot water. This second treatment
is made in order to prevent an appreciable amount of molybdenum being held
mechanically in the residue. In case the ore should carry lead, as wulfenite for
example, decompose with HNO3 and HC1 as previously stated, but after cooling
add 10 c.c. of H2SO4 and evaporate to dense fumes. Cool, add about 50 to 60 c.c.
of water and heat to dissolve soluble sulphates. Cool. Filter to remove insolu¬
ble lead. Wash well and add to the filtrate 5 to 10 c.c. excess NH4OH, heat
and filter. Repeat this precipitation in order to insure extracting all the molyb¬
denum from the residue.
If considerable phosphorus is present, it should be removed by adding a few cubic
centimeters of magnesia mixture to the ammoniacal solution before filtering off the
iron and aluminum hydroxides. The ammoniacal solution should be warmed about
20 min. to allow the precipitated magnesium phosphate to settle. If not removed,
the phosphorus will precipitate as lead phosphate along with the lead molybdate.
The molybdenum is now in the form of ammonium molybdate in a solution practically
free from silica, iron and alumina, also lead in case the sample was taken to fumes
with H2SO4. The solution should have a volume of about 250 c.c. Make acid with
HC1 using methyl orange as an indicator, adding about 5 c.c. excess. Add 5 to 10 g.
of sodium or ammonium acetate. This amount will be sufficient to destroy any
free mineral acid. It has been found that the use of HC1 to neutralize the NH4OH
and the addition of 5 c.c. excess followed by sodium or ammonium acetate in sufficient
quantity to destroy the mineral acid produces the proper conditions for the precipita-
RARE METALS 907
tion of the molybdenum as PbMo04 so that the precipitate will be granular, will
settle rapidly, and will be easily filtered and washed. The addition of a very small
amount, about 8 drops, of HN03 before precipitating the molybdenum as PbMo04
has been advocated by Weiser.1
Precipitation is best made by titrating this hot solution with a lead acetate solu¬
tion (about 18 g. crystallized lead acetate per liter) until a test drop gives no color
change with a drop of tannic-acid solution. Add about 2 c.c. excess lead acetate,
2 or 3 c.c. of acetic acid, and place the beaker back on the hot plate for about 15 to
20 min. to allow the precipitated PbMo04 to crystallize and settle. Filter hot and
wash well with hot water. It has been suggested by Weiser that the precipitate be
washed twice by decantation with a 2 per cent solution of NH4N03 and then be trans¬
ferred to the filter paper and washed with (NH4)N03. By following the above details
a precipitate of PbMo04 will be obtained which will settle rapidly and is easily
filtered and washed. It will also be a very pure PbMo04. The washed filter is
placed in a fireclay annealing cup and ignited at a dull-red heat in a muffle. The
filter need not be previously dried. After cooling, the precipitates are brushed out
of the annealing cups onto the balance pan and weighed. Weight of PbMo04 times
.2615 gives weight of molybdenum.
THORIUM
(Atomic Weight = 232.4)
Occurrence.—The principal commercial mineral carrying thorium is monazite,

although small quantities of thorianite and thorite have been used commercially.
These latter minerals, however, are limited in quantity. Monazite is composed
mainly of phosphates of the cerium and lanthanum earths together with a vari¬
able percentage of thoria. Its specific gravity is from 4.8 to 5.5, and its color
varies from yellow-gold to reddish-brown and is occasionally even dark-brown or
black. Monazite occurs in certain gneissic and granitic rocks, but the actual
commercial deposits which are worked are alluvial. The principal sources are
Brazil and India, although the mineral has been mined successfully in the United
States, in the Carolinas, Idaho and Florida. It has been found in Switzerland,
Africa and in Australia, and to a limited extent in river rocks and placers in
Ekaterinburg, Russia.
Monazite is usually found in gravels of small streams and in bottom lands. In

Brazil and India, it occurs mainly in the beach sands of sea water. It is found in
small crystals in gneiss, granite and pegmatite rocks. As these rocks become disinte¬
grated the crystals are washed into the creeks and streams and together with other
heavy sands are deposited in the beds of water courses. On the coast of Brazil the
monazite from the crystalline rocks of the coast mountains is concentrated in the sea
sands by the waves of the sea. The bulk of the monazite in Brazil is found in the
states of Espirito Santo and Bahia. In India, the main location is in the Travancore
district in Southern India. The deposits of the Carolinas in the United States cover
an area of several hundred square miles east of the Blue Ridge Mountains. Prac¬
tically all of the monazite mined in the Carolinas is derived from the gravel in the
streams and bottom lands, the miners usually following the course of the streams and
creeks. The gravels vary greatly in thickness; in general they are between 1 y% and
2}i ft. thick. Monazite is also found in Florida.
lJour. Phya. Chem., Vol. 20, 1916, p. 640.
Extraction from the Ores.—The methods of treating monazite for the pro¬
duction of thorium nitrate are more or less secret and are changed from time to
time, depending upon conditions, costs of chemicals, etc. The general principles
of the commoner method used before the war are as follows: The mineral is
heated in cast-iron pans or pots with about twice its weight of concentrated sul¬
phuric acid, until the monazite is completely decomposed, giving a white mass of
sulphates which are largely insoluble in the acid. This mixture is run into cold
water in a lead-lined vat and the whole stirred until solution is complete. The
material is allowed to stand for a considerable period in order that the insoluble
matter, consisting of silica, zircon, rutile and other minerals insoluble in concen¬
trated sulphuric acid, may settle out, and the solution which contains the rare
earths, phosphates, etc., is decanted off. If now the free acid is partially neutral¬
ized so as to reduce the acidity, thorium phosphate is precipitated first owing to
the fact that it is less soluble than the phosphates of the other rare earths. The
thorium phosphate still carrying quantities of other rare-earth phosphates is
filtered, dissolved in a minimum amount of acid, and the fractional precipitation
repeated.
One method of still further purifying the precipitate is to boil it with oxalic acid
which causes the thorium to be precipitated as thorium oxalate, while the phosphoric
acid remains in solution. The precipitated phosphates are digested for a prolonged
period with sodium hydroxide and the hydroxides formed are dissolved in hydro¬
chloric acid. If the acid solution is then carefully treated with sodium hydroxide,
until about one-sixth of the bases have been precipitated, thorium hydroxide will
be precipitated before the other hydroxides. Baskerville has suggested the volatiliza¬
tion of the phosphoric acid by mixing 1 part of monazite with 1.1 parts of petroleum
coke, 0.8 parts of lime and 0.15 parts of fluorspar, and heating in an electric furnace.
This not only removes the phosphorus but also gives the latter in a marketable form.
Further purification of the thorium hydroxide may be carried on in one of several
ways. The oxalate method depends upon the fact that thorium oxalate forms a
double salt with ammonium oxalate which is soluble, while cerium oxalate is almost
insoluble. The carbonate method is based upon the fact that thorium carbonate
forms double salts with the alkali carbonates which are soluble, whereas the double
salts of the cerium earth oxalates are insoluble. The sulphate method depends upon
the fact that certain hydrated sulphates of thorium possess a considerable difference
in solubility from the sulphates of the cerium earths. This fact is used as a basis
for fractional crystallization. Whichever method is used, after the thorium is suffi¬
ciently purified it is converted into thorium carbonate or hydroxide and dissolved in
nitric acid and the solution evaporated until thorium nitrate, Th(N0s)4.4H20, crys¬
tallizes out. After the beginning of the war, owing to the greatly increased cost of
certain chemicals, especially oxalic acid, considerable modifications were necessary in
the process used for treating monazite sand. These changes in practice have been
kept secret, but consist essentially in combinations of the methods described above.
Concentration Methods.—The first stage of concentration in the Carolinas
involves the use of oscillating tables or sluice boxes. The concentrates produced
in the sluice boxes contain 20 to 60 per cent monazite. The crude concentrates
must be further refined, and are best treated by electromagnetic separators, of
which the Wetherill type has proved to be the most successful.1 The separation
of the minerals is dependent upon the difference in their magnetic permeability.
1 Gunther: “Electromagnetic Ore Separation,” 1909.
RARE METALS 909
The magnets are best adjusted so that the first pole of the first magnet removes
from the sand the highly magnetic material, as for instance, the magnetite and
ilsemannite; the second pole of the first magnet extracts the garnets and also the
finer grains of the ilsemannite; the third magnet (being the first pole of the second
magnet) removes all of the coarser grains of the monazite; and the last pole
extracts the finer grains of the monazite. At the end-turn of the rubber belt of
the machine, the residues are dropped into a receptable.1 The grade of the
concentrates obtained from the different large sources of supply varies considerably.
The average concentrate obtained in the Carolinas was about to 4 per cent
thorium oxide, whereas that obtained from Brazil averages around 6 per cent. It
was for this reason very largely that the industry in the Carolinas was ultimately
given up. The monazite from India has even a higher grade than that from Brazil,
averaging around 9 per cent thorium oxide, and some of it going still higher. As
concentrate of this type carries about 27 per cent cerium oxide, 29 per cent lantha¬
num and allied oxides, and 2 to 4 per cent yttria and allied oxides. It also contains
about 26 per cent phosphoric acid.
Thorium Compounds.—Thorium hydroxide, Th(0H)4.£H20, is precipi¬
tated from solutions of thorium salts by means of ammonia or alkalis. It is a
white gelatinous precipitate, soluble in excess of the reagent. A granular product
is obtained if crystallization is carried out below 30°C. Thorium dioxide,
Th02, can be made by ignition of the hydroxide, the nitrate, etc. The oxide is
insoluble in dilute acids. Thorium sulphate, Th(S04) 2, can be obtained by evapor¬
ating to dryness a solution of the dioxide in sulphuric acid. The sulphate forms
double salts with the sulphates of the alkali metals; it crystallizes in transparent
monoclinic crystals with nine molecules of water of crystallization above 43°C.
Below this temperature it has four molecules. Thorium nitrate, Th(N03)4.12H20,
crystallizes at ordinary temperatures in large tablets, soluble in water and alcohol.
It also may be obtained with six and five molecules of water of crystallization by
crystallization from a hot-water solution and from a nitric acid solution. The
nitrate is the most important commercial salt of thorium. It forms double nitrates
of the general type R'2Th (N03)e where R' = NH4, K, Rb, Cs, and R"Th(N03)6*
8H2O, where R" = Mg, Mn, Zn, Ni, Co.
The thorium phosphates which are obtained by the addition of phosphoric acid
or alkali phosphates to solutions of thorium salts are of somewhat doubtful com¬
position. Thorium oxalate, Th(C204)2-6H20, may be precipitated quantitatively by
the addition of oxalic acid to a thorium salt even in the presence of considerable
quantities of mineral acids. It is much less soluble in mineral acids than the oxalates
of the rare earths. The salt dissolves in excess of alkali oxalate, but it is precipitated
from the solutions by mineral acids.
Metallic Thorium.—The metal has probably not yet been prepared in a

perfectly pure state. Reduction of the oxide with magnesium is never complete,
and attempts to reduce with carbon give a mixture of the metal and carbide.
Moissan and Honigschmid heated the purified anhydrous chloride with sodium in
a sealed glass tube and obtained a product which was said to contain only 3 per
cent of the oxide. The amorphous metal is a dark-gray powder of specific
gravity 11.3 and when hammered and strongly heated, it has a density of 12.16.
1 K. L. Kithil: Technical Paper, 110, XJ. S. Bureau of Mines.
910 C HEMIC A L ENGINEERING
It burns readily in air, and melts at about 1,700°C. although, the pure metal would
probably melt at a higher temperature.
Uses of Thorium.—The main use of thorium is in connection with the incan¬
descent gas-mantle industry. The successful use of incandescent gas mantles
began in 1893 when Welsbach patented the use of a mixture containing 98 to 99
per cent thorium oxide with 1 to 2 per cent cerium oxide (English Pat. 124 of 1893).
With this mixture the oxides give a brilliant light and have a maximum
illuminating power. The cerium oxide is probably in solid solution in the thorium
oxide.1 Thorium oxide is a very poor heat conductor and hence can be raised
to a high temperature. The small amount of cerium oxide gives enough color
for efficient light emission. An outline of the methods used for obtaining mantles
is as follows: The mantle itself is made either of ramie fiber or of artificial silk.
The woven mantles are thoroughly washed in a 2 per cent solution of nitric acid,
distilled water, and then in dilute ammonia. . The latter is thoroughly washed
out by distilled water. The mantle is dried in a current of hot air and then dipped
in a solution of thorium and cerium nitrates, consisting of 99 per cent thorium salt
and 1 per cent of cerium.
Small quantities of beryllium or magnesium nitrate are added for strengthening
purposes. • The upright mantles are then fitted with an asbestos loop by which they
are suspended, whereas the inverted mantles are fixed to a supporting ring. Those
portions on which there is especial strain are sometimes treated with a thorium solution
containing much larger proportions of calcium, aluminum, or magnesium salts. The
mantle is then shaped and burnt off from the top downward by applying a Bunsen
flame. During this process of burning off, there is considerable shrinkage. As soon
as the carbonization is finished, the mantle is shaped and heated with low-pressure
burners, after which it is hardened by heating in a high-pressure burner. The
mantles now consisting of nothing but the ash skeleton are immersed in a collodion
solution containing collodion, ether, camphor and castor-oil. The ingredients may
vary to some extent. After drying at a moderate temperature, they are ready for
use.
1 White and Tbavers: Jour. Soc. .
Chem Ind., 1902, Vol. 21, 1012.
SECTION XXVIII
RARE GASES
By Richard B. Moore1
HISTORY
Discovery of Argon.—It is a carious fact that during the 30 or 40 years imme¬

diately preceding the discovery of argon, little work was done by chemists on
the atmosphere. They seemed to feel that here at least an end had been reached,
and the wise man should bend his energies in other directions. No suspicion was
entertained that the air was a veritable gold mine for chemical research, and even
after the evidence seemed to indicate that we had not yet found out all there was
to know about our atmosphere, there was a feeling of doubt and uncertainty that
made the work move slowly. The first clue to the undiscovered elements was
partly accidental. In 1882 and several succeeding years, Lord Rayleigh worked
on the relative densities of hydrogen and oxygen. In 1893 he published his
results on the densities of some other gases, notably those of oxygen, nitrogen and
air.2 He prepared oxygen and nitrogen by different methods; and while the
density of oxygen proved in all cases to be the same, that of nitrogen prepared by
chemical methods differed from that of nitrogen obtained from the air. The latter
was about 0.5 per cent heavier than the former. A difference of this magnitude
could not be accounted for on the basis of experimental error, and the riddle
remained unanswered for some time. Finally, Prof, (afterwards Sir) William
Ramsay asked permission of Lord Rayleigh to take up the matter from the chem¬
ical standpoint. He had already in some previous work found that red-hot
magnesium was a good absorbent for nitrogen. He concluded that if there was
any part of atmospheric nitrogen different from the rest, a separation might be
made by passing the gas over the heated metal. This was tried, the oxygen first
being removed by metallic copper. A slight increase in the density of the prod¬
uct was observed. Arrangements were now made to circulate the gas back¬
wards and forwards over the heated magnesium. After several days of this
treatment the density was found to be 16.1. At this stage Ramsay did not
know that he had to deal with a new element, but thought that the results were
due to a triatomic form of nitrogen, corresponding to ozone. A further absorp¬
tion by magnesium finally gave a gas with density of about 20. It was only when
the spectrum of this product was examined that Ramsay became convinced
that he had found a new element. Meanwhile, Lord Rayleigh was dealing
with the problem in a different manner. He passed sparks through a mixture of
atmospheric nitrogen and oxygen in the presence of a solution of sodium hydrox¬
ide. The hydroxide absorbed the oxides of nitrogen formed, and after removal of
the excess of oxygen, the resulting gas showed an increased density. From this
point, Ramsay and Rayleigh continued the work together and announced the
discovery of the new element “argon” at the British Association Meeting in 1894.
1 Chief chemist, Bureau of Mines, Washington, D. C.
2 Proc. Roy. Soc., 1893, Vol. 83, p. 146; 1894, Vol. 58, p. 340.
911
Discovery of Helium.—After the discovery of argon, Ramsay looked for other

sources of the element. He heard through Sir Henry Miers of a paper by W. F.
Hillebrand of the U. S. Geological Survey, in which it was stated that the mineral
cleveite, on heating, gave off a considerable quantity of inactive gas which was
supposed to be nitrogen. Ramsay suspected that this gas might be argon and so
obtained a sample of the mineral; heated it with sulphuric acid; pumped off the
evolved gas; sparked it with oxygen, and finally ran it into a spectrum tube. The
spectrum was entirely different from that of argon; having as a chief character¬
istic a bright-yellow fine. At first Ramsay’s surprise was so great that he thought
there was something wrong with the spectroscope. In-addition, others suggested
various explanations for the bright-yellow line; the most common being that it
was the sodium line due to dirty electrodes, or other causes. It was only when
the spectrum of the sodium flame was compared through a comparison prism
with the yellow line from the spectrum tube that it became evident beyond ques¬
tion that a new element—namely, terrestrial helium—had been discovered.
Discovery of Krypton.—It now seemed probable that other undiscovered
gases beside argon and helium existed. After a large number of minerals had
been examined, in addition to the gases from mineral waters, as well as sea water,
which were investigated with the object of finding all possible sources of argon,
helium and any new gases, Ramsay, assisted by his pupil Travers, decided on the
fractionation of a considerable quantity of argon. Fifteen liters of the latter gas
had previously been prepared in connection with the work on helium, and it
was planned to use this for the investigation. Before it was fractionated, how¬
ever, a liter of liquid air, presented by Dr. Hampson, was allowed to evaporate
and the last fraction of gas was collected separately. The spectrum of this gas,
after purification from nitrogen and oxygen, showed a bright-green and a bright-
yellow line not belonging to the argon spectrum. The density of the sample was
also greater than that of argon, and the new element was called krypton.
Discovery of Neon.—It was natural with this encouragement that the work
should be carried on with great eagerness. With the assistance of Dr. Travers,
the 151. of argon referred to above was now fractionated. The first fraction which
came off, containing the gases with the lowest boiling points, was separated and
examined. The spectrum showed a large number of brilliant red and yellow
lines, and the density of this particular sample was about 15. The new gas was
called neon. Apparatus was now devised in connection with the Hampson
liquid-air machine, with which the lighter fractions from liquid air could be
reliquefied and refractionated. Finally, a portion was obtained which could not
be liquefied even by air boiling under reduced pressure. This proved to consist
largely of neon, but also contained helium—proving the existence of the latter
gas in the air. All attempts to separate the helium from the neon were futile,
until later on the two were fractionated at the temperature of liquid hydrogen.
At this temperature the neon solidified and the helium could be pumped off.
Discovery of Xenon.—The residues or last fractions from 30 1. of liquid air
which had been allowed to evaporate, were afterwards used to make a larger quan¬
tity of krypton. The majority of the gas consisted of argon which was carried
through a rigorous course of fractionation; the last fraction containing the kryp¬
tons showing some new blue lines, with a density greater than that of krypton.
After careful refractionation, the final gas obtained gave a density of 64, and the
RARE GASES 913
new element was called xenon. The density put this gas in its proper place in the
periodic table. The volume of xenon obtained was only 3 c.c., while the total
volume of krypton recovered was 12 c.c. All of the physical constants of these
gases were afterwards found with the small volumes indicated; and only men
possessed of the manipulative skill of Ramsay and Travers could have done this.
General Properties of the Gases.—The main property common to all the
rare or “noble” gases is the fact that they are absolutely inert to chemical
action. So far, no attempts at getting them to combine with other elements
have been successful. When Ramsay first obtained argon in quantity, he used
rigorous methods to act on it chemically. He sparked it with oxygen, chlorine
and hydrogen without success. Red-hot sodium hydroxide, sodium peroxide, or
fused potassium nitrate had no action. Fluorine, nascent chlorine, and red-hot
magnesium were equally ineffective. Helium, neon and the other gases appeared
to be just as inactive as argon. More recent attempts to get argon to combine
with other elements involved an arc discharge between electrodes of different
metals immersed in pure liquid argon. Whereas the metals usually underwent a
decided change themselves, in no case could it be proved that any argon com¬
pounds had even a momentary existence. And yet there is some evidence
that possibly compounds may be formed under certain conditions. Fergusonite
is a mineral which, on heating, evolves considerable quantities of helium.1 In
doing so, it becomes quite hot; and Ramsay called it an “endothermic mineral.”
Again, when a discharge is passed between the electrodes of ordinary spectrum
tubes containing the rare gases, there is a very decided tendency for the electrodes
to “splash” from a mirror coating) on the sides of the tubes. This coating
adsorbs the gas and shortens the life of the tube. This “splashing” would seem
to indicate that endothermic compounds were momentarily formed. The ratio
of the specific heats indicates that every one of the rare gases is monatomic.
Additional evidence to this effect is furnished by the results of Sir J. J. Thomson’s
“ Rays of Positive Electricity. ”2
Position in the Periodic Table.—The inert gases constitute a group by
themselves. Their atomic weights do not fit anywhere in the periodic table,
unless a place can be found in the eighth group, or by themselves in a new group
preceding group one. The latter position has been preferred and accepted by
all chemists. Usually the chemical properties of an element have a direct con¬
nection with its position in the periodic table. The fact that the rare gases are
inert chemically fully justifies placing them is a separate group by themselves.
Argon is the only one of the rare gases which appears to be out of place with the
arrangement of the table as given above; its atomic weight being higher than that
of potassium, the element which follows it. It is not, however, out of place as
regards its atomic number. Originally it was thought that argon was probably a
mixture, and vigorous efforts were made to separate it into two or more compounds.
One plausible explanation has been advanced—that argon contains a small
percentage of diatomic molecules; 5 per cent of such molecules would reduce its
density to 19 and the atomic weight to 38. Such a condition is not inconceivable,
but if it were a fact the density of the gas would probably vary with the tempera-
1 When, the zirconium silicate from the Buckman & Pritchard property near Jacksonville, Fla.
is ground, it suddenly evolves helium after reaching a certain fineness.—Editor.
2 Sir J. J. Thomson: “Rays of Positive Electricity,” Longmans, Green and Company, 1913.
58
ture, which is not true. The ratio of the specific heats for argon is 1.659, which
closely approximates the theoretical 1.667 for a perfect gas. Five per cent of
diatomic molecules should give a ratio of 1.648, instead of 1.659. No satis¬
factory explanation, therefore, can be advanced at the present time for the
anomalous position of argon in the periodic table.
HELIUM
Homogeneity of the Gas.—The discovery of helium by Ramsay has already
been described. Before long some doubt was cast upon the elementary nature of
the gas and on its identity with solar helium. When the gas was allowed to stream
through a porous plug into a Plticker tube, the light at first was green, the line
X 5,016 of the single-line group being equal in intensity to the D3 line, and then
gradually became yellow as D3 became relatively stronger.1 Runge and Paschen,
who made the observation, concluded that they had separated the gas obtained
by Ramsay from cleveite into two components: Helium of density 2.2, and
a lighter gas for which a name “parhehum” was proposed. Some claims
were also made that helium obtained from different minerals showed con¬
siderable variations in density2 and that by diffusion the gas could be separated
into fractions differing still more in density. The homogeneity of helium was
satisfactorily proved by Travers,3 who showed that on passing an electrical
discharge through helium in a vacuum tube, the pressure fell steadily, owing to
the fact that the gas was adsorbed by the finely divided platinum which was
deposited on the walls of the tube; and that with this fall in pressure, the color of
the glow changed from orange-yellow to bright-yellow, then to yellowish-green,
and finally to green. The tube was then allowed to cool; the remaining gas
was pumped out, and the tube was then heated in order to drive out the gas
which had been occluded by the platinum. This gas showed exactly the same
behavior as the original gas when a discharge was again passed through the
tube: Ramsay and Travers also carried out some elaborate experiments in
fractionating by diffusion the gas obtained from cleveite. They showed con¬
clusively that whereas fractions of different densities could be obtained, this
variation was due to the presence of argon; and the elementary nature of helium
was, therefore, firmly established.
Occurrence.—Kayser4 was the first one to detect the presence of helium in
the atmosphere by spectroscopic methods, and his results were confirmed by
Friedlander,5 Ramsay’s original determination of the quantity in the atmos-
sphere was one volume of helium in 250,000 volumes of air.6 Watson,7 with
better facilities at his command, calculated that the amount is more probably
1 part by volume of helium in 185,000 volumes of air. Helium is occluded in
a large number of minerals; indeed it is contained to a greater or lesser extent
in all minerals and rocks. As will be explained more fully later, Ramsay and
Soddy8 showed that the radium emanation gave rise to helium. The question
1 Runge ,
and Paschen, Phil. Mag., 1895, Vol. 40 p. 297.
,
2 Ramsay, Proc. Roy. Soc.f 1896, Vol. 59 p. 325.
3 Proc. Roy. Soc., 1897, Vol. 60, p. 449.
* Chem. News, 1895, Vol. 72, p. 89.
s Zeit. physik. Chem., 1896, Vol. 10, p. 657.
RARE GASES 915
received still further light when Rutherford and Geiger proved the identity
of the alpha particle given off by radioactive changes with the helium atom.
The helium found in the earth’s crust, therefore, is undoubtedly partly, if not
entirely, due to the disintegration of radioactive types of matter in the earth’s
crust with the elimination of alpha particles. Generally speaking, the amount
of helium in minerals is proportionate to the amount of uranium or thorium pres¬
ent. The Hon. R. J. Strutt has, therefore, sought to determine by this means
the age of certain minerals and rock deposits with fair success. In some cases,
however, the rate of leakage is greater than the rate of formation from radio¬
active changes.
Helium is found in the gases evolved from many mineral springs, notably the
Bath springs in England, the springs at Mazi&res, at Monte Irone (Albano), at
Railliere in the Pyrennes, at Wildbad in the Black Forest, and in a number of springs
in Tuscany. In some cases the gases evolved carry as much as several per cent of
helium. The most notable source of helium in quantity, however, is certain natural
gases in the United States. Cady and MacFarland1 showed that a number of natural
gases in Kansas contained as much as 1 per cent or even more of helium. At Dexter
in Cowley County, Kansas, the content of some wells was as high as 1.8 per cent.
At Eureka, samples were analyzed containing as high as 1.5 per cent, and in the
El Dorado field certain of the wells contained as much as lJi per cent helium. The
Augusta field was almost as rich as the others. In certain localities in Oklahoma
natural gas samples carry as much as per cent helium, while in Vinton County,
Ohio, about 80 miles south of Columbus, there is a very large supply of gas containing
between 0.4 and 0.5 per cent helium. The most satisfactory supply at present in this
country, however, is at Petrolia, Tex., 20 miles northeast of Wichita Falls. The gas
here carries 0.93 per cent helium, and as much as 15,000,000 or 20,000,000 cu. ft. a day
of natural gas can be pumped. There are several small sources of supply in Europe.
One test well near Pechelbronn, in Alsace, carried 0.4 per cent helium. The gas in
Transylvania has less than 0.002 per cent, but it is claimed that the gases from some
of the fumaroles in Italy are reasonably rich in helium.
All radioactive minerals carry helium to a greater or less extent. The amount of
helium in the mineral depends on the amount of uranium or thorium, the age of the
mineral itself, and also on the physical properties of the mineral, which affect its
power of occlusion of helium. Since the radioactive minerals are those which carry
uranium or thorium, all minerals containing these elements also contain helium,
provided the gas is not able to diffuse out so fast as it is formed.
Liberation of Helium from Minerals.—Helium can be obtained from minerals

either by strongly heating, or by solution in acids. A still better method is to
fuse the mineral with acid sodium sulphate. Whichever method is used, the tube
containing the mineral is connected with a Topler pump, and the evolved helium
may be taken through the pump and collected over mercury. A high tempera¬
ture is required to get rid of all of the gas by heating alone. In case of monazitc,
it has been shown that very little gas comes off at 500°C., and at 900°C. it took
24 hr. of heating to liberate all the gas. For thorianite, 0.7 per cent came off at
350°C. after 5 hr. heating, 8.5 per cent at 500°C. after 80 hr., and the whole at
1,000° after 30 hr.
Purification.—Helium can best be purified by means of charcoal at liquid
air temperature. If the gas containing traces of nitrogen, oxygen, carbon
,
* Jour. Am. Chem. Soc., 1907, Vol. 29 p. 1523.
dioxide, or other similar impurity, is run into an evacuated tube containing

cocoanut charcoal, the tube having previously been heated and pumped dry by
means of an efficient air-pump, all other gases but the helium will be adsorbed
and the helium can be pumped out in practically a pure condition. If argon
is present, it will also be adsorbed by the charcoal, but in the case of neon, only
partial adsorption will take place and some neon will probably come over with
the helium unless the quantity is very small. In order to get a complete separa¬
tion it is, therefore, necessary, actually to fractionate the two, using the above
method.
Properties.—Helium is a colorless gas, is odorless and tasteless.1 Its density

is 1.995, according to Watson.2 The compressibility coefficient at zero is —0.0006
between pressures of 400 and 800 mm. of mercury.3 The molecular weight and
therefore, the atomic weight calculated from Heuse’s value for the density is
4.0. Helium does not obey Graham’s law of diffusion of gases, but passes through
a porous diaphragm more slowly than is expressed by the law. The solubility of
helium in water was determined by Antropoff,4 who found for the absorption
coefficient the value 0.0134 at zero, and 0.0226 at 50°, with minimum at 10°. The
thermal conductivity of helium is 0.00003386.5 The refractivity was first deter¬
mined by Hayleigh and afterwards more accurately by Ramsay and Travers, who
found it was 0.1238 times that of air, which gave the value for helium of 1.000361.6
The viscosity was first determined by Rayleigh by measurement of the rate of
flow of the gas through a capillary tube. He found the value 0.96; air being
1.0. Schultz redetermined the viscosity and found it to be 1.086 times that of
air at 15°C.7 The spectrum of helium has been studied by Runge and Paschen,
by Watson and others. The lines fall naturally into groups, in each of which we
have a principal series, and two subordinate series. One of these groups consists
of doublet series and the line (D3) is the first in the principal series. It is the
lines of this group which characterize the solar spectrum of helium, and are seen
when an ordinary vacuum tube containing helium is examined with a spectro¬
scope. Generally speaking, all of the lines are seen in a vacuum tube, but the
relative intensity of the lines varies with the pressure and, therefore, the color
of the tube itself may vary materially with different pressures. The main helium
line, usually called the “helium yellow/’ consists of a double line with the wave
lengths 5876.21 and 5875.88.
Formation of Helium from Radioactive Substances.—The fact that helium
is formed from radium was proven by Ramsey and Soddy.8 By dissolving a
soluble radium salt in water, it was found after purification of the evolved gas
that an appreciable residue of helium was left behind. By allowing the solution
containing the radium salt to stand, the helium was reformed and could be
pumped off again. The origin of helium in the air has always been ascribed
1 Ramsay and Travers, Proc. Roy. Soc., 1898, Vol. 62, p. 316.
2 .
Trans. Chem. Soc., 1910, Vol. 97, p. 810. It is 1.999, according to Hettse (Ber deutsch. physikal.
,
Ges., 1913, Vol. 15 p. 518).
3 Jaouerod and Scheuer, Mem. Soc. Phys. Nat., 1908, p. 35.
4 Proc. Roy. Soc., 1910, Vol. A 83, p. 414.
RARE GASES 917
to radioactive material in the earth’s crust, but such an explanation is not satis¬
factory for the large amounts of helium present in natural gas found in certain
localities in the United States. It would take immense quantities of radium,
uranium, or thorium to give rise to the large volumes of helium contained in these
gases, and some other explanation is necessary. It has been shown that if the
velocity of the alpha particle (helium atom) in radioactive changes were to fall
below a certain value, the alpha particle would not be able to ionize gases and,
therefore, its presence could not be detected by ordinary analytical radioactive
methods.1 In other words, an element might be radioactive in so far as dis¬
integration and the elimination of alpha particles is concerned without our
being able to detect this fact. If this is true, it would give at least a reasonable
explanation of the presence of helium in such quantities in natural gas.
Liquefaction of Helium.—After several attempts by others, helium was success¬
fully liquefied by Onnes. This was accomplished by the use of the Joule-Thomp-
son effect after the gas had been pre-cooled by liquid hydrogen. The gas was
circulated through an apparatus in which it was cooled first to the temperature of
liquid air and then to — 258°C. by means of liquid hydrogen boiling under reduced
pressure. It was then liquefied in an apparatus using the general principles of
the Hampson type of machine. Liquid helium is colorless and very mobile; its
density is 0.154, and its maximum density is at about 2° absolute.2 The boiling-
point is 4.3° absolute; the critical temperature 5.25° absolute3 and the critical
pressure is 2.26 atmospheres. The critical volume (calculated) is 0.00271.
Commercial Manufacture of Helium.—In the fall of 1917, it was decided by
the United States Army and Navy to attempt to extract helium in commercial
quantities from the natural-gas wells at Petrolia, Tex., to be used in dirigibles and
war balloons instead of hydrogen, as helium is non-inflammable and has many
other advantages. A sufficient amount of money for experimental purposes
was allotted from Army and Navy appropriations to the United States Bureau of
Mines, which took charge of the work. Two experimental units were constructed
at Fort Worth, Tex., connecting up with the gas supply at Petrolia by means of a
pipe line. The plants at Fort Worth were an Air Reduction (Claude process)
and a Linde Air Products plant (Linde Process). A third plant, using the Jefferies-
Norton process, was constructed at Petrolia, Tex., at the source of supply of the
gas. The natural gas at Petrolia carries about 36 per cent nitrogen and 60 per
cent methane, in addition to small quantities of heavier hydrocarbons, such as
ethane, propane, butane, etc. The problem calls for the successive liquefaction
and fractionation of the hydrocarbons in the gas, and also of the nitrogen.
Although the helium does not liquefy at the temperature involved, namely, that
of liquid nitrogen, it is to a certain extent soluble in the liquefied gases, and as
it is present in the proportion of only 0.93 per cent., it is necessary thoroughly
to fractionate the liquids so that too much helium will not be lost by solubility.
The plants of the Air Reduction Company and Linde Company were regular
oxygen or nitrogen units modified to meet the particular problem at hand. The
most serious trouble was met with in keeping the tops of the columns sufficiently
cool for the final liquefaction of the nitrogen. The grade of product obtained
1 Cf. p. 879.
2 Onnes: Comm. Phys. Lab. Leyden, No. 119.
3 Onnes, Proc. K. Akad. Wetensch, Amsterdam, 1911, Vol. 14, p. 678.
varied from 20 per cent to 70 per cent helium. This gas was refined by a process
which practically involved squeezing out the nitrogen by passing the gas through
a condenser at a pressure of 1,000 lb. per square inch at the temperature of liquid
nitrogen. In this manner, a final product containing about 92 per cent helium
was obtained, and over 200,000 cu. ft. were manufactured during the experimental
period.
A commercial unit costing $2,000,000, consisting of six large Linde units, was
decided upon by the Government as the result of the experimental work. The
Jefferies-Norton process at Petrolia involves the use of three expansion engines for
refrigeration purposes, and a heat interchanger consisting of seven sections, in which
the hydrocarbons are progressively liquefied. The original pressure of 300 lb. is
held all through the still and inter changer, and the gas is expanded through the
expansion engines only after it has passed through the still and has returned to the
high-pressure outside sections of the interchanger, from whence it is expanded through
the expansion engines back into the low-pressure sections. The same general
principles in the stills are involved as regards the extraction of the helium, but the
still of the Jefferies-Norton Company uses reflux coils, through which the liquids are
throttled, instead of condensers as is the case in the Claude and Linde columns.
ARGON
Occurrence.—Argon is found widely distributed in nature. It exists in the

atmosphere in the proportion of 0.941 per cent by volume, after water vapor, and
carbon dioxide have been removed from the air, and 1.1845 per cent by volume
of atmospheric nitrogen.1 The proportion of argon in the air is constant, as
samples taken from widely different localities show practically no variation in
the argon content. Over the open sea, however, slightly higher proportions of
argon have been found. It is found to a small extent in sea and river water,
owing to its being dissolved with the air. Since argon is slightly more soluble than
air itself, the proportion by volume in natural waters is slightly higher. It has
been found in many mineral springs, among them being those of Bath and Buxton;
the old sulphur well at Harrogate, England; Monte Irone, Albano; the Poretta
baths, Bolognian Apennines; Maizi&res, Cote d’Or, etc. It is also present in
certain natural gas in Kentucky, and in the helium-bearing gas at Petrolia, Tex.
Only a few minerals yield argon on heating; the chief of these are malacone2 and
a mineral from Brazil described by Antropoff .3 Both minerals contain zirconium
in considerable quantities, and both are radioactive. The amount of argon pres¬
ent is probably too large to be derived from atmospheric sources and, therefore,
there is at least a possibility that argon may also be a product of the disintegration
of radioactive elements, although there is no direct proof of this statement.
Argon has also been found in eliastite4 and in a mineral carrying cerium from the
South Caucasus.5
Recovery from the Atmosphere.—Argon is extracted from the atmosphere by
passing air over metallic copper and a strongly heated mixture of metallic magne¬
sium and lime. The absorption of nitrogen by magnesium alone is a very slow
, ,
1 Schloesing, Comp, rend., 1895, Vcl. 121 p. 605; Kellas, Proc. Roy. Soc., 1896, Vol. 59 p. 66.
2 Ramsay and Travers, Proc. Roy. Soc.t 1897, Vol. 60, p. 442.
,
3 Zeit. Elektrochem., 1908, Vol. 14 p. 585.
4 Lockyer, Proc. Roy. Soc.t 1896, Vol. 60, p. 133.
,
6 Tschernik, J. Russ. Phys. Chem. Soc., 1897, Vol. 29 p. 291.
RARE GASES 919
process, but when mixed with lime the absorption is rapid and complete. The
air is first passed over soda lime and phosphorus pentoxide, and then through
a copper tube in which the oxygen is absorbed. From this it goes through the
magnesium-lime mixture, over heated copper oxide, and finally through other
soda lime and pentoxide tubes. In order to have the absorption complete, some
care is required. First the apparatus must be completely exhausted and must
remain airtight. Connections may be made by antimony pressure tubing of
small bore. Black and gray rubber do not stick tightly to the joints, and the
apparatus is sure to leak. It is bettter to make glass joints wherever possible,
and reserve the rubber for connections with the hard-glass tubes. The lime must
be as free from carbonate as possible and, of course, perfectly dry. Before the
air is passed through it, the magnesium-lime tube is strongly heated and soon
turns black, owing to the glass being attacked by the magnesium. During this
process carbon monoxide and hydrogen are formed in considerable quantities;
the first by the reduction of the carbon dioxide obtained from the traces of carbon
in the lime, and the second from the action of water vapor on the magnesium.
These gases should be pumped off as fast as formed, and when the evolution is
practically finished, air is slowly admitted to the apparatus. Traces of carbon
monoxide and hydrogen formed during the actual experiment are oxidized by
the copper oxide, and the products absorbed by the soda lime and pentoxide
tubes.
When the nitrogen reaches the magnesium-lime tube, absorption is indicated by

a sudden rise in temperature and the formation of a glowing ring at the rear end of the
tube. Unless this ring is formed, absorption is not taking place. The glow gradually
goes down the tube as the magnesium is used up, making it easy to know when to stop
the experiment. If the apparatus leaks, small traces of air may be removed from the
argon obtained by sparking with oxygen in presence of a strong solution of caustic
soda. The excess of oxygen is then removed by means of phosphorus. For larger
quantities of argon it is more convenient to use the method of Fisher and Ringe,
according to which air is passed over a mixture of 90 per cent calcium carbide and 10
per cent calcium chloride heated to 800°C. In this case, the temperature is too high
for glass, and iron tubes must be used. With a properly prepared apparatus, consider¬
able quantities of argon may readily be made. The nitrogen and carbide give carbon
and cyanamide. The oxygen and carbide give lime and carbon, while carbon dioxide
and monoxide formed at the same time react with the carbide to give carbon. Argon
can also be extracted from the atmosphere by the original sparking method of Lord
Rayleigh, in which an inverted flask is used containing a strong solution of caustic
soda. Through a two-hole rubber stopper and passing up through the solution are
two glass tubes containing electrodes. The oxides of nitrogen formed by the sparking
are absorbed by the caustic-soda solution. Argon is now made commercially from
the atmosphere by using the ordinary refrigeration method for liquifying air. The
fractionation column—either the Claude process or the Linde process is used—is
adapted so that a fraction of the gas is taken off between the liquefied oxygen and the
liquid nitrogen, as liquid argon has a boiling-point between liquid oxygen and liquid
nitrogen. In this manner, a product containing about 80 per cent of argon is obtained.
It is purified by chemical methods already described, and finally used in incandescent
lamps instead of nitrogen. Liquid argon is colorless and transparent, and boils at
— 186.1°C. Its density at this temperature is 1.4046. If allowed to cool by
means of its own evaporation it can be solidified to a white mass which melts
at — 188.9°C.
Physical Properties.—Argon is an odorless, colorless and tasteless gas. Its

density is 19.95 (0 = 16), and the weight of a normal liter of gas is 1.782 g. Its
atomic and molecular weights are, therefore 39.9. The critical constants of argon
are as follows: Critical density, O.SOOg./c.c;1 critical pressure, 47.996 atmospheres2
and the critical temperature, -122.44°C.3 The vapor pressure of solid argon is
411 mm. at —190.9°. The triple point of argon is —189.3°. The molecular heat
of vaporization of the liquid is 12.92 cal. at —125.49°; 24.01 cal. at —140.8°,
and 35 cal. at -183.06°. The coefficient of expansion of the gas is 0.003668.4
Its viscosity is 1.2 times that of air.5 Its solubility in water is four volumes per
100 volumes of water at ordinary temperatures.6 The refractivity (p — 1) of
argon is:
Wave length X U - 1) X 10
6,439 2,796
5,461 2,823
4,359 2,851
The gas has a low dielectric cohesion, namely, 38, which is about one-fifth that of
air. The sparking distance is about 40 per cent greater in argon than in air
under similar conditions. The gas is diamagnetic. The thermal conductivity
is K = 0.00003894 at 0°C. The specific heats at constant pressure and constant
volume at ordinary temperatures have not been determined, but they have been
determined at higher temperatures, namely, over 1,300°C. Argon gives two
sets of spectrum lines, depending upon the pressure the gas is subjected to. For
the sake of convenience, the two sets are called the red and blue spectrum.
With a spark-gap and Leyden jar the blue spectrum is given. The principal
ines are as follows:
“Red” Spectrum of Argon
Wave-length Intensity Wave-length Intensity Wave-length Intensity
7,066.60 7 4,272.30 8 4,182.00 7

6,964.80 8 4,266.43 8 4,164.30 7
5,607.44 8 4,259.49 9 4,158.72 10
4,510.85 7 4,200.80 10 4,044.56 8
4,348.11 8 4,198.16 10 3,949.11 8
4,345.32 7 4.191.84 10 3,834.77 8
4,333.71 8 4.190.84 7 3,567.79 7
4,300.25 8
iCsoMMELiN, Proc. K. Akad. Wetensch, Amsterda.ni, 1910 Vol 13 p 607

2 Cbommelin, itrib., p. 54.
3 Crommelest, ibid., p. 54.
4 Ktjenenand Randall, Proc. Roy. Soc., 1895, Vol. 59, p. 60
6 Rayleigh,Proc. Roy. Soc., 1896, Vol. 59, p. 198.
* Raleigh and Ramsay, Phil Trans., 1895, Vol 186 p 187
RARE GASES 921
“Blue” Spectrum of Argon
Wave-length Intensity j Wave-length Intensity Wave-length Intensity
5,559.02 8 3,781.02 7 3,285.91 7

4,880.00 8 3,729.45 9 2,942.94 7
4,806.17 8 3,638.02 7 2,806.30 I 8
4,609.74 7 3,588.63 9 2,769.70 8
4,426.17 8 3,582.55 7 2,753.90 | 8
4,348.22 9 3,576.81 8 2,744.90 8
4,104.11 7 3,561.21 7 2,708.40 8
4,014.00 7 3,559.70 8 2,647.60 8
3,928.75 7 3,546.01 7 2,516.80 8
3,868.72 7 3,545.79 7 2,515.60 8
3,850.72 8 3,491.72 9
NEON
Occurrence.—Neon is found terrestrially principally in the atmosphere.

Watson's estimate is probably the most accurate, and his figure is one volume of
neon in 55,000 volumes of air. Neon has been recognized in the gases from a
number of thermal springs.1 Recently, neon has also been detected in certain
natural gases from Kansas and Texas, which also carry helium.
Separation.—Neon has usually been prepared by extracting the helium and
neon together from the atmosphere and then separating these two gases either
at the temperature of liquid hydrogen, at which temperature neon solidifies and
the helium can be pumped off, or by fractionation from charcoal at liquid-air
temperatures. A convenient method of getting a volume of nitrogen enriched
with neon and helium is to take the gas escaping from an ordinary Hampson
liquefying machine and return it to the intake of the compression pump. In this
manner, the gas is partially liquefied and the more volatile constituents, namely,
neon and helium, become concentrated in the gas given off. A somewhat similar
result can be obtained by reliquefying all of the gas escaping from the air liquefier,
and if a current of air is now brought through this liquid, the first portion that
evaporates will be extremely rich in helium and neon.
The enriched gas must then be freed from nitrogen and traces of oxygen in the
usual manner of passing over heated metallic copper to remove the oxygen, and over
a heated mixture of magnesium and lime or over calcium carbide in order to remove
the nitrogen. The mixture of neon and helium can be separated by means of liquid
hydrogen, or more cheaply by fractionation from charcoal at the temperature of
liquid air. In this latter process, the neon is more readily adsorbed by the charcoal
than the helium, and it is possible to get a reasonably pure sample of neon in this
manner. By means of a modification of his apparatus for the fractionation, conden¬
sation and distillation of liquid air, Claude has been able to obtain the neon, helium
and hydrogen which exist in the air in considerable quantities. The gas thus obtained
contains about 50 per cent nitrogen.2
1 Noureu , ,
and Biqtjard: Compt. rend., 1906, Vol. 143 p. 180; 1908, Vol. 146 p. 435.
2 Claude, , ,
Compt. rend., 1905, Vol. 141 p. 823; 1908, Vol. 147 p. 624.
Physical Properties.—The absorption coefficient of neon in water is 0.0114 at

0°C., and increases up to 0.0317 at 50°.1 The density of the gas, according to
Watson, is 10.08 (0 = 16) the weight of a normal liter being 0.9002. The atomic
weight is therefore 20.16. The compressibility coefficient at 0° between 0 and 1
atmosphere is —0.00105, and neon, therefore, like hydrogen is an ultra-perfect
gas.2 The refractivity of neon is 1.00006716 at 0° and 760 mm. for the green mer¬
cury line. The dispersion at N. P. T. is about twice that of helium. The
viscosity of the gas at 0°C. is 2.981 X 10~4 in absolute units.3 The dielectric
cohesion of neon at 17° is 5.6; hydrogen being 205. The dielectric cohesion,
therefore, is less than helium, whereas one would expect it to be between that of
helium and argon. Very small traces of impurities raise the dielectric cohesion
of neon considerably.
Liquefaction of Neon.—Neon can be liquefied and even solidified by means of
liquid hydrogen. Its boiling-point is — 245.92°C. Its triple-point pressure
is 35 cm. of mercury, and its critical pressure is 29 atmospheres. According to
Onnes and Crommelin4 the critical temperature of neon is approximately 42° to
45° absolute. The surface tension, density and refractive index of liquid neon
have not yet been measured, but it has been calculated that the refractive index
should be about 1.24.
Spectrum.—Neon has a very characteristic spectrum, the most prominent
visual lines being in the red and yellow. The wave lengths of the more prominent
lines are as follows:
Wave-length Intensity Wave-length | Intensity | Wave-length Intensity
6,402.40 ! 10 6,183.79 10 5,974.73 6

6,383.15 8 6,143.28 10 5,944.91 10
6,328.38 6 6,128.63 8 5,882.04 8
6,304.99 8 6,096.37 10 5,852.65 20
6,266.66 10 6,074.52 10 5,764.54 8
6,217.50 8 6,030.20 10 4,259.53 6
6.182.37 10 5,975.78 8
KRYPTON
Occurrence. Krypton occurs in the atmosphere in the proportion of 1 part

by volume in 20,000,000.5 It has been found spectroscopically to exist in certain
mineral springs, and also in certain natural gases in Kansas and Texas.
Separation. Krypton has only been made in quantity once, and this was
by Moore.6 The oxygen residues obtained in a Claude air liquefying plant were
used after about 100 tons of air had passed through the plant. The 501. of liquid
oxygen left were allowed to evaporate slowly from Dewar flasks; the gas passing
through a glass spiral dipping in liquid air. In this manner the oxygen and traces
1 Antrqpoff: Proc. Roy. Soc., ,
1910, Vol. A 83 p. 474.
,
2 Burt, Trans. Faraday Soc.t 1910, Vol. 6 p. 19.
, ,
3 Ranxine, Proc. Roy. Soc., 1910, Vol. A 83 p. 516; Vol. A 84 p. 181.
, ,
4 Pr°c. A". Akad. Wetensch, Amsterdam, 1915, Vol. 18 p. 15; 1911, Vol. 14 p. 163.
, ,
6 Ramsay, Proc. Roy. Soc., 1903, Vol. 71 p. 424; 1908, Vol. A 80 p. 599.
6 Trans. Chem. Soc., 1908, Vol. 93, p. 2181; Proc. Roy. Soc., 1908, Vol. A 81 195. ,
RARE GASES 923
of nitrogen were not liquefied, but the krypton and xenon contained in the liquid
oxygen were solidified in the spiral. The spiral was then removed from the
liquid air, and the gases thus obtained collected over mercury and fractionated
from each other. This is possible since the vapor pressure of krypton at the
temperature of liquid air (17mm.) is nearly 10 times that of xenon.
Physical Properties.—Krypton is a colorless, odorless and tasteless gas.
The density has been determined by Moore, using the best fractions obtained
in the work described above. The value obtained was 41.506; oxygen being
16. The weight of a normal liter of krypton is, therefore, 3.708 g.1 The com¬
pressibility of krypton was determined by Ramsay and Travers, and they showed
that the value of pv decreases markedly with increase of pressure. Between
0 and 1 atmospheres, the compressibility is +0.00215. The solubility of krypton
in water has been determined, the absorption coefficient is 0.1249 at 0°, 0.0788 at
20°, and 0.0823 at 50°. The viscosity at 10.6°C. is 1.361 times that of air; at 0° in
absolute units, it is 2.344. Krypton is easily liquefied; it boils at — 151.7°C.
The density of the liquid at its boiling-point is 2.155. The critical temperature
of the liquid is — 62.5°C.; and its critical pressure is 54.3 atmospheres.2 Krypton
can easily be solidified, and the solid melts at — 169°C. The specific heat of krypton
has not been determined. The ratio of the specific heats was determined to be
1.689. Krypton undoubtedly has a monatomic molecule, and the atomic weight
is, consequently, identical with the molecular weight and equals 82.92.
This figure is based on Watson’s theoretical correction of Moore’s density
determination.
Spectrum.—The spectrum of krypton depends materially upon the nature of
the discharge. With the direct discharge in a Geissler tube, the number of lines
is not large and is characterized mainly by a bright-green and a yellow line.
When a spark-gap is used with a Leyden jar, a much larger number of lines show
up in the blue. The aurora borealis gives a spectrum with a persistent green line
which is probably the krypton green. The following represent the principal
lines in the two spectra:
First Krypton Spectrum (Direct Discharge)
Wave-length Intensity Wave-length Intensity
5,871.12 10 4,454.12 10
5,570.50* 10 4,400.11 6
5,562.45 6 4,376.33 10
4,671.40 10 4,362.83 9
4,624.48 10 4,319.76 10
4,502.56 9 4,318.74 8
4,501.13 7 4,274.15 10
4,463.88 10
* Probably the green auroral line.
,
1 Trans. Ckem. Soc., 1908, Vol. 98 p. 2181.
, ,
2 Ramsay and Travers, Proc, Roy. Soc., 1901, Vol. 67 p. 329; Phil, Trans. 1901, Vol. 197 p. 47.
Second Krypton Spectrum (Condensed Discharge)
Wave-length ; Intensity Wave-length Intensity
5,633.17 ' 6 4,109.38 6

4,765.90 6 4,098.89 7
4,739.16 7 4,088.48 8
4,619.31 6 4,065.22 8
4,577.40 6 4,057.17 8
4,475.18 7 3,994.98 6
4,355.67 10 3,920.29 8
4,293.10 6 3,917.76 6
4,145.28 C 3,906.37 8
XENON
Occurrence.—Xenon exists in the atmosphere to the extent of one volume in

190,000,000. . It is also found to a very limited extent in the gases evolved from
certain mineral springs1 and also in certain natural gases from Kansas and Texas.
Separation.—Xenon has only been prepared in reasonable quantity on one
occasion by It. B. Moore, as described under krypton.2 From the oxygen residues
obtained from 100 tons of liquid air, about 200 c.c. of pure xenon were finally
obtained. Owing to the fact that the vapor pressure of xenon at the temperature
of liquid air is very low, namely, 0.17 mm., its separation is greatly assisted. In
fractionating xenon from krypton, the krypton can be gradually pumped off from
the mixed solids since its vapor pressure is 10 times that of xenon. Finally a
point is reached when the material seems to be pumped dry and no more krypton
will come off. At this stage, it is necessary to warm up the gases and thus
vaporize them, and then solidify them again by means of liquid air. A fresh
surface is thus exposed, and a further small quantity of krypton can be pumped
off. In this manner, pure xenon can be obtained.
Physical Properties.—It is a colorless, tasteless and odorless gas. Its density
as determined by Moore,3 is 65.35 (0 = 16). Therefore, a normal liter of xenon
weighs 5.815 g. Xenon is reasonably soluble in water; its absorption coefficient
being 0.2189 at 0°C. and 0.0878 at 50°C. A minimum solubility occurs at about
40°C.4 The viscosity of xenon at 0°C. is
Nq = 2.107 X 10-4,
while its refractivity is
/a — 1 = 705.49 X 10~6 for the green mercury line (X = 5461).
Xenon can be easily liquefied and solidified. The liquid is colorless and boils at
-109.1°C. under atmospheric pressure. Its density at the boiling-point is 3.06
and, therefore, the molecular volume is 42.7. The critical temperature of xenon
is +16.6 C., and the gas may be liquefied at this temperature by a pressure of
,
1 Motjbbh and Lepape, Compt. rend., 1909, Vol. 149 p. 1174.
. . ,
1 Proc Roy Soc., 1908, Vol. A 81 p. 195.
, ,
z Tram. Chem. Soc., 1908, Vol. 93 p. 2181; Watson, ibid., 1910, Vol 97 p 833
,
4 Antrapgfe, Proc. Roy. Soc., 1910, Vol. A 83 p. 474.
RARE GASES 925
58.2 atmospheres. The melting-point of solid xenon is — 140°C. Its atomic

weight is the same as its molecular weight, namely, 130.2 (0 = 16).1 The ratio
of specific heats is 1.666.
Spectrum.—A Geissler tube containing xenon emits a blue line when a current
is passed through it, and an examination of the spectrum shows that the principal
lines are in the blue end of the spectrum. If a spark-gap and Leyden jar be used,
the color of the tube changes to green. The following lists give the principal
lines in the two spectra:
First Xenon Spectrum (Uncondensed Discharge)
Wave-length Intensity Wave-length Intensity Wave-length Intensity
4,923.28 6 4,624.46 15 4,116.25 7

4,916.63 6 4,582.89 5 4,109.84 5
4,807.19 6 4,524.83 6 4,078.94 10
4,734.30 8 4,501.13 10 3,967.74 10
4,697.17 7 4,193.70 8 3,951.16 10
4,671.42 10
I
Second Xenon Spectrum (Condensed Discharge)
l
W av e-length Intensity Wave-length Intensity Wave-length Intensity
6,097.80 7 4,921.68 6 4,415.00 7

6,051.36 7 4,883.68 6 4,406.99 5
6,036.40 6 4,876.68 7 4,395.91 10
5,976.67 7 4,862.69 8 4,393.34 10
5,719.83 6 4,844.50 10 4,330.63 15
5,667.85 6 4,823.47 6 4,245.54 10
5,616.99 6 4,652.15 6 4,238.37 10
5,531.33 7 4,603.21 10 4,208.61 6
5,472.90 7 4,592.22 6 4,193.25 8
5,460.63 6 4,585.65 10 4,180.20 10
5,439.19 8 4,577.36 6 4,109.20 6
5,419.40 10 4,545.34 8 4,050.19 6
5,372.62 8 4,541.03 8 3,950.70 8
5,339.56 9 4,481.01 7 3,922.67 10
5,314.15 8 4,462.38 20 3,908.00 7
5,292.40 10 4,448.28 10
5,080.88 7 4,434.35 6
RADIUM EMANATION
Niton.—The gas which is given of! from radium as its first disintegration prod¬
uct has properties similar to those of the rare gases and is, therefore, often classed
with these gases. The name, radium emanation, was given to the gas at the time
of the first observation of its formation from radium, and was an excellent tem¬
porary name, but is not satisfactory at the present time, now that we know more
i Moore, Trans. Chem. Soc., 1908, Vol. 93, p. 2181.
about the gas and recognize it as an element, although a transition one. The
name, niton (Nt), suggested by Sir William Ramsay, is an excellent one and at
least as far as chemical usage goes should be adopted.
Occurrence.—Niton occurs in all minerals containing uranium and, therefore,
in radium. As radium is widely distributed through the earth’s crust and exists in
minute quantities in all rocks and minerals, niton is an extremely widely dis¬
seminated element, although, owing to the fact that the quantities of radium in
the earth’s crust are small and to the fact that it decays rapidly, it exists only
in almost infinitesimal quantities. The gas, therefore, occurs in the atmosphere,
but the amount in a cubic meter of air is that which would be in equilibrium
with 1 X 10~10 g. of radium.
Separation.—The only way by which niton can be obtained in anything like
measurable quantities is from a high-grade radium salt. If such a salt is placed
in water solution, the niton will accumulate, and after about 4 days approximately
half the equilibrium amount will have been produced. At the same time a certain
proportion of the water will be decomposed by the a rays into hydrogen and oxy¬
gen; there always being a slight excess of hydrogen, owing to the simultaneous
formation of a small amount of hydrogen peroxide. If the gases are pumped off,
the hydrogen and oxygen can be caused to recombine to form water, and the
excess of hydrogen can be oxidized by means of heated copper oxide. Traces of
carbon dioxide can be absorbed by solid caustic soda, and in this manner the
niton can be purified and run into a small capillary tube for experiment, or
observation. The volume of gas obtained, even under these favorable conditions,
is small, as the amount that is in equilibrium with 1 g. of radium element is
0.585 cu. mm. at N.P.T.
Physical Properties—Niton obeys all of the gas laws and can be manipulated
like any other gas. Its density was determined by Ramsay and Gray1 by weighing
on a microbalance sensitive to 2X10~6 mg. The amount of emanation used was
not more than 0.005 cu. mm. The mean of several experiments gave the value
111.5 (H = 1), which gives 223 as the atomic weight of niton. The gas is soluble
in water; its absorption coefficient being approximately 0.5 at 0°; 0.3 at 15°; 0.27
at 18 ; 0.23 at 20°; and 0.15 at 40°. No work has been done on the specific heats
of niton, but from analogy with the rare gases, it is assumed that it is monatomic.
The gas is easily liquefied and frozen. Its critical temperature is 104.5°, and its
critical pressure is 62.5 atmospheres. Niton boils at - 62°, and freezes at - 71°C.
Its vapor pressure at — 71°C. is 500 mm. Its heat evolution (including dis¬
integration products up to Ra C) is about 106 g. calories per hour for the emana¬
tion in equilibrium with 1 g. of radium.
Spectrum.—When placed in a capillary tube and an electrical discharge is
passed through the gas, niton gives a spectrum which does not vary with the
character of the discharge. Ramsay and Collie2 first examined the spectrum
visually and noticed a strong and persistent line at X = 5,595. Since that time
the spectrum has been mapped in more detail by other observers.3
1 Proc. Roy, Soc., 1911, VoL 84, p. 536.
1 Proc. Roy. Soc., 1914, Vol. 73, p. 470.
8 ***• *°* Soc-> 1909* yol. A 83, p. 50; Ramsay and Cameron: Proc. Roy. Soc., 1908,
1909 VoL 17 p 202RUTHERPORD R°YD8: Ma°" l908’ VoL 16> P- 313; Royds: Phil. Mag.,
SECTION XXIX
MATERIALS OF CONSTRUCTION
By Percy E. Barbour1
Introduction.—The general design of plants is considered in Section XXX,

so that this Section is merely a discussion of the properties of materials, without
consideration of what is to be done with them. Further information, both as to
the properties of materials in the abstract, and as to their use in structures may
"be found in: Kidder’s “Architects and Builder’s Pocket Book,” M. S. Ketchum’s
**Steel Mill Buildings” H. G. Tyrrell’s “Engineering of Shops and Factories;”
Day’s “Industrial Plants (N. Y. Eng. Mag.)] Dyson & Clarkson’s “Chemical
Works” (Scott, Greenwood & Son, London and Van Nostrand, New York);
Charnock’s “Mechanical Technology” (Constable & Co., London); Knott &
Ntackay’s “Practical Mathematics” (W. & R. Chambers, Edinburgh, Van
ISfostrand, New York).
Iron and Steel.—The following definitions were adopted by the Committee
on the Uniform Nomenclature of Iron and Steel of the International Associa¬
tion for Testing Materials, September 1906.
Cast Iron.—Iron containing so much carbon or its equivalent that it is not malle¬
able at any temperature. The committee recommends drawing the line between
cast iron and steel at 2.20 per cent carbon.
Pig Iron.—Cast iron which has been cast into pigs direct from the blast furnace.
Bessemer Pig Iron.—Iron which contains so little phosphorus and sulphur that
it can be used for conversion into steel by the original or acid bessemer process restricted
to pig iron containing not more th n 0.10 per cent of phosphorus.
Basic Pig Iron.—Pig iron containing so little silicon and sulphur that it is suited
for easy conversion into steel by the basic open-hearth process (restricted to pig iron
containing not more than 1.00 per cent of silicon).
Gray Pig Iron and Gray Cast Iron.—Pig iron and cast iron in the fracture of which
the iron itself is nearly or quite concealed by graphite, so that the fracture has the
gray color of graphite.
White Pig Iron and White Cast Iron.—Pig iron and cast iron in the fracture of which
little or no graphite is visible, so that the fracture is silvery and white.
Malleable Castings.—Castings made from iron which when first made is in the
condition of cast iron, and is made malleable by subsequent treatment without
fusion.
Malleable Pig Iron.—An American trade name for pig iron suitable for converting
into malleable castings through the process- of melting, treating when molten, casting
in a brittle state, and then making malleable without remelting.
Wrought Iron.—Slag-bearing, malleable iron, which does not harden materially
when suddenly cooled.
Steel.—Iron which is malleable at least in some one range of temperature and in
1 Mining engineer, Assistant Secretary, American Institute of Mining and Metallurgical Engineers
927
addition is either (a) cast into an initially malleable mass; or, (b) is capable of harden¬
ing greatly by sudden cooling; or, (c) is both so cast and so capable of hardening.
Openhearth Steel—Steel made by the openhearth process, irrespective of carbon
content.
Bessemer Steel.—Steel made by the bessemer process, irrespective of carbon
content.
Blister Steel.—Steel made by carburizing wrought iron by heating it in contact
with carbonaceous matter.
Crucible Steel.—Steel made by the crucible process, irrespective of carbon content.
Steel Castings.—Unforged and unrolled castings made of bessemer, openhearth,
crucible or any other steel.
AUoy Steels.—Steels which owe their properties chiefly to the presence of an
element other than carbon.
Classification of Iron and Steel.—The limits of carbon, the specific gravity

and properties of iron and steel are as follows:
Per cent of Specific
Carbon Gravity Properties
Cast iron 5 to 1.50 7.2 Not malleable, not temperable

Steel 1.50 to 0.10 7.8 Malleable, and temperable
Wrought iron 0.30 to 0.05 7.7 Malleable, not temperable
It will be seen that the percentage of carbon alone is not sufficient to distinguish
between steel and wrought iron. The softer grades of steel resemble wrought
iron. Very mild openhearth steel is often sold under the trade name of “ ingot
iron/7 and is reputed to have many advantages over structural steel, some of
which are questionable.
Cast Iron.—The product of the blast furnace, where the iron ore is reduced
in the presence of a flux, is called pig iron. The term cast iron is commonly
applied to pig iron after it has been again melted and cast into finished form.
Cast iron contains carbon, silicon, sulphur, phosphorus and manganese in addi¬
tion to pure iron, and occasionally very small quantities of other elements.
Carbon.—The percentage of carbon ordinarily varies between 1 % and 4 per cent,
but in the presence of manganese the carbon may be much higher. Carbon may
occur in the form of combined carbon, giving a white brittle cast iron, or in the form
of graphite, giving a gray cast iron, which is the form used in structural castings. The
proper amount of carbon in cast iron depends upon the amount of other impurities
and upon the use that is to be made of the finished product.
Silicon.—The carbon is controlled by varying the amount of silicon and sulphur.
Silicon acts as a precipitant of carbon, changing it from the combined form to the
graphite form. The silicon in gray cast iron is usually between % and 3 per cent.
Sulphur. Sulphur has the opposite effect of silicon and its presence is considered
objectionable. Sulphur produces ‘ ‘red-shortness” (brittleness when the iron is heated).
The amount of sulphur in gray-iron castings should not exceed 0.12 per cent.
Manganese. Manganese and sulphur both tend to increase the amount of com¬
bined carbon, but they tend to neutralize each other. Manganese gives closeness of
grain and prevents the absorption of sulphur on remelting. The amount of manga¬
nese in gray-iron castings is usually less than per cent; more than 2 per cent makes
cast iron brittle.
Phosphorus. Phosphorus increases the fusibility and fluidity of cast iron but at
the same time makes it brittle. A high phosphorus content is necessary in cast iron
ENGINEERING MATERIALS 929
for light ornamental castings where strength is not required. The phosphorus in
gray-iron castings varies from to 1per cent.
Nickel steel for structural purposes is made by the openhearth process and con¬
forms to the following requirements as to chemical composition in terms of per cent.:
i
Structural steel Rivet steel
Carbon.... not over 0.45 not over 0.30

Manganese. not over 0.70 not over 0.60
/ Acid. not over 0.05 not over 0.04
Phosphorus
\ Basic not over 0.04 not over 0.03
Sulphur.... not over 0.05 ! not over 0.045
Nickel. not under 3.25 not under 3.25
The material has a tensile strength of 70,000 to 80,000 lb. per square inch for
rivet steel and 85,000 to 100,000 lb. per square inch for plates, shapes and bars.
Structural steel may be made by the bessemer or the openhearth process except
that rivet steel and steel for plates or angles over % in. in thickness which are to be
punched shall be made by the openhearth process. Structural steel to meet these
standard specifications must conform to the following requirements as to chemical
composition:
Structural steel, Rivet steel, per

per cent cent
Phosphorus bessemer. not over 0.10

Openhearth. not over 0.06 not over 0.06
Sulphur. not over 0.045
Structural steel has a tensile strength of 55,000 to 65,000 lb. per square inch and
rivet steel 46,000 to 56,000 lb. per square inch.
When steel plates for structural purposes are ordered to gage, the thickness of each
plate shall not vary more than 0.01 in. under that ordered. One cubic inch of rolled
steel is assumed to weigh 0.2833 lb.
Malleable Castings.—Small thin castings made of white cast iron may be decarbon¬
ized by heating the castings in annealing pots containing hematite ore or forge iron
scale. The castings are kept at a cherry red heat for 3 to 4 days, and are then allowed
to cool slowly. The metal in malleable castings should not exceed in. in thickness
in small castings, nor 3^ in. in large castings, and should be of uniform thickness.
Steel for concrete reinforcement bars may be made by bessemer or the openhearth
process. Bars shall be rolled from new billets only and shall conform to the following
chemical compositions:
. /Bessemer. not over 0.10 per cent

osp orus j Openhearth. not over 0.05 per cent
Carbon Steel Forgings.—Carbon steel for welding and case hardening Class A
and for case hardening when subsequently heat treated Class B must conform
to the following requirements as to chemical composition:
59
Class A Class B
0.08 to 0.18 0.15 to 0.25

Manganese, per cent. 0.30 to 0.50 0.30 to 0.50
Phosphorus maximum, per cent. 0.045 0.045
Sulphur maximum, per cent. 0.05 0.05
Refined wrought iron bars shall be made wholly from puddled iron and may consist
either of new muck-bar iron or a mixture of muck-bar iron and scrap but shall be
free from any admixture of steel. The iron shall have a tensile strength of 48,000
lb. per square inch.
Lake Copper.—In order to be classed as Lake, copper must originate on the
northern peninsula of Michigan, and may be furnished in wire bars, cakes, slabs,
billets, ingot bars, or ingots. Lake copper offered for electrical purposes, whether
fire or electrolytically refined shall be known as “low-resistance lake.” Low-
resistance Lake wire bars shall have a resistivity not to exceed 0.15535 interna¬
tional ohms per meter-gram at 20°C. (annealed). All ingots and ingot bars
shall have a resistivity not to exceed 0.15694 international ohms per meter-gram
at 20°C. (annealed).
Lake copper having a resistivity greater than 0.15694 international ohms per
meter-gram at 20°C. (annealed) shall be known as “high-resistance Lake.” Low-
resistance Lake copper shall have a purity of at least 99.880 per cent as determined by
electrolytic assay, silver being counted as copper. High-resistance Lake copper
shall have a purity of at least 99.880 per cent, copper, silver, and arsenic being counted
together. The arsenic content of high-resistance Lake copper for special purposes
shall be subject to agreement at time of purchase.
Electrolytic Copper.—Electrolytic-copper ware bars, cakes, slabs, billets,

ingots, and ingot bars shall have a purity of at least 99.880 per cent as determined
by electrolytic assay, silver being counted as copper. All wire bars shall have a
resistivity not to exceed 0.15535 international ohms per meter-gram at 20°C.
(annealed); all ingot and ingot bars shall have a resistivity not to exceed 0.15694
international ohms per meter-gram at 20°C. (annealed). Cakes, slabs, and
billets shall come under the ingot classification except when specified for elec¬
trical use at time of purchase, in which case wire bar classification will apply.
Spelter.—Under the standard specifications of the A. S. T. M., virgin spelter
is the zinc product resulting from ore or similar raw material by reduction and
distillation and not produced from re-worked metal and is considered in four
grades, as follows:
Contains Not oveb
Pb Fe Cd A1 Pb + Cd 4
A. High grade. 0.07 0.03 0.07 0.0 0.10
B. Intermediate. 0.20 0.03 0.50 0.0 0.50
C. Brass special. 0.60 0.03 0.50 0.0 1.00
D. Prime Western. 1.60 0.08 not specified
Manganese Bronze.—The A. S. T. M. standard specification covers man¬

ganese-bronze ingots having notched flat bottoms, approximately 3 by 2% in.
wide by 12 in. long, properly tapered to strip easily from an iron mold. The
chemical composition shall be as follows:
Copper. 53 to 62 per cent

Zinc. 36 to 54 per cent
Aluminum. 0.05 to 0.5 per cent
Lead. not over 0.15 per cent
The ultimate tensile strength shall not be less than 70,000 lb. per square
inch.
Corrugated steel is used for roofing and siding of many plants where the
climatic conditions permit. It is too hot in hot countries for closed-in structures
and too cold for cold countries. Corrugated steel is made from plain flat besse-
mer steel sheets by rolling in the corrugations one at a time. Iron and open-
hearth steel corrugated sheets may be obtained but only with difficulty and extra
cost. The standard corrugation is 2% in. wide and % in. deep; special corruga¬
tions may be obtained at extra cost but like all special material is inadvisable if
avoidable.
Corrugated steel may be either plain black unpainted, galvanized or covered with
a coating of asbestos material, lead, etc. Plain black must be painted before and
also after erection to prevent rusting and corrosion. Galvanized does not take paint
readily unless washed with weak acid.
The standard width of corrugated sheets is 26 in. after corrugating. It is laid
with either one or two corrugations side lap depending upon climatic and other
conditions incident to the use of the building. When laid with one lap a sheet will
cover a width of 24 in. and with two laps will cover 21in. in width. Endwise the
lap should be 6 in. for roofing and 4 in. for siding. The United States standard gage
is used for specifying thickness and weight. Amounts of corrugated steel are speci¬
fied in squares 10 by 10 ft. or 100 sq. ft. The weights of different gages, in general use,
per square (100 sq. ft.) for standard sheets with 2J^ in. by % in., corrugations are
given in the table below, no allowances being made for laps.
Weight of Flat, and Corrugated Steel Sheets with 23^-in. Corrugations.

Weight per Square (100 Sq. Ft.)
Flat sheets Corrugated sheets

Thickness
Gage No.
in inches Black
Black Galvanized Galvanized
painted
16 0.0625 250 266 275 291

18 0.0500 200 216 220 236
20 0.0375 150 166 165 182
22 0.0313 125 141 138 154
24 0.0250 100 116 111 127
26 i 0.0188 75 91 84 99
28 0.0156 63 79 69 86
Trautwine allows about one seventh more weight per square for side lap of one
corrugation and end lap of 4 in. Standard stock lengths of corrugated vary by 6-in.
differences and run from 5 ft. to 10 ft. long. Odd lengths may be obtained at extra
cost and generally with extra delay.
Corrugated Steel Fastenings.—Corrugated steel is fastened to wooden sheath¬
ing, girts, nailing strips, or purlins with barbed roofing nails; 8 d. nails 2%
in. long being commonly used spaced 6 to 8 in. apart. For estimating, 100 nails
per square of corrugated covering is used.
For fastening corrugated to angle or channel girts or purlins without wooden
nailing strips clinch nails are used made of in. No. 10 soft-iron wire with flat head
and are clinched around the steel member.
Table of Clinch Nails
Z. Purlin leg. 3 inches 4 inches 5 inches 6 inches 7 inches

Length. 5 inches 6 inches 7 inches 8 inches 9 inches
No. per pound.... 32 29 23 21 18
[. Purlin leg. 3 inches 4 inches 5 inches 6 inches 7 inches

Length. 6 inches 7 or 8 inches 9 inches 10 inches 11 inches
No. per pound. ... 29 21 18 16 14
Number of Barbed Roofing Nails in 1 Lb.
Length, Gage Number Length, Gage Number

Size Size
inches No. in 1 lb. inches No. in 1 lb.
4 d. IX 13 339 20 d. 4 6 30
6 d. 2 12 205 30 d. 4X 5 23
8 d. 2M 10 96 40 d. 5 4 17
10 d. 3 9 63 50 d. 5X 3 13
12 d. m 8 52 60 d. 6 2 10
16 d. 3X 7 38
Holes for nails are punched in the top of the corrugations. Side laps may be riveted
with copper rivets to make a tight job. Corrugated may also be fastened by metal
straps around the supporting members riveted to the corrugated.
The safe load for corrugated steel sheets may be obtained from trade catalogues
or computed by Rankine’s formula
15 l
Where, W — safe load in pounds;
S - working stress in pounds;
h = depth of the corrugations in inches;
b = width of the sheet in inches;
t — thickness of the sheet in inches;
l = clear span in inches.
The ‘Carnegie Handbook” gives the following formula for transverse strength
of corrugated.
l = unsupported length of sheet in inches.

49.95 tbd t = thickness of sheet in inches.
l 6 =«= width of sheet in inches.
99,900 tbd d = depth of corrugations in inches.
W — breaking weight distributed in tons.
w — breaking weight distributed in lbs.
Corrugated steel of 24 gage and lighter is seldom used, it being too thin to stand
service and corrosion. Roofing is generally used a gage or two heavier than siding
in the same structure because of the increased severity of service conditions.
Sulphurous gases from coal or smelting operations are particularly corrosive to
galvanized or plain-black corrugated; lead or oxide paints afford little protection.
Asphaltum or graphite paints should be used.
Anti-condensation lining. Where corrugated roofing is used in a heated building
moisture will collect and considerable dripping may result. To obviate this chicken
wire is stretched from eaves to ridge and covered with paper or asbestos and then
felt and the corrugated is then laid on top in the usual way.
Protection of Iron and Steel against Corrosion.—If iron and steel are left
unprotected to the atmosphere or to water they will corrode by rusting. Rust
is ferric hydroxide, approximating Fe03H3. Dry air or water free from oxygen
has no effect on iron or steel. Acids greatly hasten corrosion though not
essential to it. Where iron and steel are in contact with water an electrolytic
action always takes place and where the electrolytic force is considerable the
corrosion is greatly increased. Covering iron or steel with a protective coating
not acted on by air or water will retard or prevent corrosion. Cast iron owing
to its carbon content and the silicon coating from its molding rusts much more
slowly than steel or wrought iron.
The ordinary method of protecting iron and steel is by paints, of which there are
many kinds, the most common being oil, tar and asphaltum paints.
An oil paint consists of a drying oil and a pigment thoroughly mixed that will be
workable and is readily applied, has good covering powers, adheres to the metal and
is durable. Linseed oil is commonly used and is unsurpassed in durability. The
pigments used in making oil paints for protecting metal are lead, zinc, iron and carbon.
Raw linseed oil, made from crushing and pressing flaxseed, dries very slowly and
is not often used in the pure state for structural paints. It is made to dry more
rapidly by boiling or by adding a drier. Driers are made in various ways, the usual
ones being fatty salts of those metals having more than one common valence, such as
iron, manganese, etc., and the more drier is added the quicker the paint will dry and
the poorer it becomes.
White lead (hydrated carbonate of lead) is the best white oil paint for interior
and exterior wood work, has good adhesion and covering quality but is readily acted
on by corrosive gases, turns dark and disintegrates. Is not a good paint to use on
metal except over another paint.
Red lead (lead tetroxide) is the most commonly used paint for the first protective
coating of steel for structural, bridge, ship and other purposes. It is stable and dur¬
able under all weather conditions. One ounce of lampblack added to one pound of
red lead gives a deep chocolate color and lengthens the time of drying.
Zinc white (zinc oxide) is a white powder used for paint which has a good covering
quality; used alone it has a tendency to peel and to deteriorate and wash off. Mixed
with lead, however, and ground in linseed oil it forms a very durable paint for metallic
surfaces.
Iron oxide is used frequently for red paint for wooden buildings but is not a good
paint for metal. It is composed of anhydrous sesquioxide (hematite) and hydrated
sesquioxide of iron which is iron rust. It hastens corrosion when applied to structural
steel.
Carbon in the form of lampblack and graphite is much used for black paints.
Lampblack in oil makes an excellent paint, the oil absorbing the carbon. Graphite
is not absorbed by oil but forms a good paint the particles being held together by
cementing action. The Dixon Graphite Co. of Jersey City uses a flake graphite
combined with silica, while the Detroit Graphite Manufacturing Co. uses a mineral
ore with a large percentage of graphite carbon in granulated form. Graphite paints
have a large covering capacity.
Asphalt paint is a mixture of coal tar or mineral asphalt combined with a mineral
base or oil. Various trade brands are purchasable of various combinations. Black
Bridge paint made by Edward Smith & Co., New York, contains asphaltum, linseed
oil, turpentine and Kauri gum. Asphalt paints are very satisfactory for metal.
Coal tar is often used for painting iron work. It should be free from acid and
should be applied hot. It will run in hot weather, so is not a good paint for many
purposes but where this can be obviated by local conditions finds much use. Oil
paint will not stick to tar paint and the latter must first be removed before an oil
paint is applied.
Ketchum quotes from the “Pencoyd Handbook” the following table for the cover¬
ing capacity of paints:
Average Surface Covered per Gallon of Paint
Square feet
Volume Pounds Volume and
Paint of oil, of weight of
One Two
gallons pigment paint
coat coats
Gallons Pounds
Iron oxide (powdered). 1 8.00 1.2 = 16.00 600 350
Iron oxide (ground in oil)... 1 24.75 2.6 = 32.75 630 375
Red lead (powdered). 1' 22.40 1.4 = 30.40 630 375
White lead (ground in oil).. 1 25.00 1.7 = 33.00 500 300
Graphite (ground in oil).... 1 12.50 2.0 = 20.50 630 350
Black asphalt. 1 17.25 4.0 = 30.00 515 ! 310
(turpentine)
Linseed oil (no pigment).... 1 875
Ketchum gives K gal. for first coat and % gal. for second coat per ton of structural
steel as average paint required under average conditions, light structural steel aver¬
aging 250 sq. ft. and heavy structural steel about 150 sq. ft. per ton of metal.
STRUCTURAL TIMBER
The term “Structural Timber” was defined by the A. S. T. M. Committee as

all such products of wood in which the strength of the timber is the controlling
element in their selection and use. The following is a list of products which
are recommended for consideration as structural timbers: Trestle timbers—
stringers, caps, posts, mud sills, bracing, bridge ties, guardrails; Car timbers—
car framing, including upper framing; car sills; Framing for buildings—posts,
mud sills, girders, framing, joists; Ship timbers—ship timbers, ship decking;
Cross arms for poles.
Standard Defects.—Measurements which refer to the diameter of knots
or holes should be considered as referring to the mean or average diameter
1. Sound Knot.—A sound knot is one which is solid across its face and which is
as hard as the wood surrounding it; it may be either red or black, and is so fixed by
growth or position that it will retain its place in the piece.
2. Loose Knot.—A loose knot is one not firmly held in place by growth or position.
3. Pith Knot.—A pith knot is a sound knot with a pith hole not more than }:£ in.
in diameter in the center.
4. Encased Knot.—An encased knot is one whose growth rings are not intergrown
and homogeneous with the growth rings of the piece it is in. The encasement may
be partial or complete; if intergrown partially or so fixed by growth or position that
it will retain its place in the piece, it shall be considered a sound knot; if completely
intergrown on one face, it is a water tight knot.
5. Rotten Knot.—A rotten knot is one not as hard as the wood it is in.
6. Pin Knot.—A pin knot is a sound knot not over in. in diameter.
7. Standard Knot.—A standard knot is a sound knot not over in. in diameter.
8. Large Knot.—A large knot is a sound knot, more than 1 Yi in. in diameter.
9. Round Knot.—A round knot is one which is oval or circular in form.
10. Spike Knot.—A spike knot is one sawn in a lengthwise direction; the mean or
average width shall be considered in measuring these knots.
11. Pitch Pockets.—Pitch pockets are openings between the grain of the wood
containing more or less pitch or bark. These shall be classified as small, standard,
and large pitch pockets, (a) A small pitch pocket is one not over in. wide. (b)
A standard pitch pocket is one not over % in. wide, or 3 in. in length, (c) A large
pitch pocket is one over % in. wide, or over 3 in. in length.
12. Pitch Streak.—A pitch streak is a well-defined accumulation of pitch at one
point in the piece. When not sufficient to develop a well-defined streak, or where the
fiber between grains, that is, the coarse-grained fiber, usually termed “spring wood,”
is not saturated with pitch, it shall not be considered a defect.
13. Wane.—Wane is bark, or the lack of wood from any cause, on edges of
timbers.
14. Shakes.—Shakes are splits or checks in timbers which usually cause a sepa¬
ration of the wood between annual rings:
15. Rot, Dote and Red Heart.—Any form of decay which may be evident either as
a dark-red discoloration not found in the sound wood, or the presence of white or red
rotten spots, shall be considered as a defect.
16. Ring Shake.—An opening between the annual rings.
17. Through Shake.—A shake which extends between two faces of a timber.
Standard Names for Structural Timbers.—1. Southern Yellow Pine.—Under

this heading two classes of timber are used: (a) Longleaf pine; (b) shortleaf pine.
It is understood that these two terms are descriptive of quality, rather than of
botanical species. Thus, shortleaf pine would cover such species as are now
known as North Carolina pine, loblolly pine, and shortleaf pine. Longleaf pine
is descriptive of quality, and if Cuban, shortleaf, or loblolly pine is grown under
such conditions that it produces a large percentage of hard summer wood, so as
to be equivalent to the wood produced by the true longleaf, it would be covered
by the term longleaf pine.
Average Safe Allowable Working Unit Stresses, in Pounds per Square Inch
ENGINEERING MATERIALS
The above tables are based on those recommended by the committee on “ Strength of Bridge and Trestle Timbers " of the Association of Railway Superintendents
of Bridges and Buildings at their Fifth Annual Convention in October, 1895, but the arrangement and values in many cases are now modified by later data from
9:
2. Douglas Fir.—The term Douglas fir is to cover the timber known likewise as
yellow fir, red fir, western fir, Washington fir, Oregon or Puget Sound fir or pine,
northwest and west coast fir.
3. Norway Pine, to cover what is known also as red pine.
4. Hemlock, to cover Southern or Eastern hemlock; that is, hemlock from all
States east of and including Minnesota.
5. Western hemlock, to cover hemlock from the Pacific coast.
6. Spruce, to cover Eastern spruce; that is, the spruce timber coming from points
east of Minnesota.
7. Western spruce, to cover the spruce timber from the Pacific coast.
8. White pine, to cover the timber which has hitherto been known as white pine,
from Maine, Michigan, Wisconsin and Minnesota.
9. Idaho white pine, the variety of white pine from western Montana, northern
Idaho, and eastern Washington.
10. Western pine, to cover the timber sold as white pine coming from Arizona,
California, New Mexico, Colorado, Oregon and Washington. This is the timber
sometimes known as Western yellow pine, or Ponderosa pine, or California white
pine, or western white pine.
11. Western larch, to cover the species of larch or tamarack from the Rocky
Mountain and Pacific coast regions.
12. Tamarack, to cover the timber known as tamarack, or Eastern tamarack,
from States east of and including Minnesota.
13. Redwood, to include the California wood usually known by that name.
Yellow-pine bridge and trestle timber specifications were defined by the

A. S. T. M. in 1910 and revised in 1915. Except as noted, all timber shall be
cut from sound trees and sawed to standard size, close-grained and solid, free
from defects, such as injurious ring shakes and crooked grain, unsound or loose
knots, knots in groups, decay, large pitch pockets or other defects that will
materially impair its strength.
Standard Size of Sawed Timber.—Rough timbers sawed to standard size
shall mean that they shall not be over % in. scant from the actual size specified.
For instance, a 12 by 12-in. timber shall measure not less than 11% by 11% in..
Standard Dressing of Sawed Timber.—Standard dressing shall mean that
not more than % in. shall be allowed for dressing each surface. For instance,
a 12 by 12-in. timber after being dressed on four sides shall measure not less than
11H by 11 y2 in.
STONE
Stone.—Very few stones will withstand the action of severe heat which neces¬
sitates caution in their use in structures where fireproofing is required. Granite
may explode or entirely disintegrate when subjected to severe heat. Limestones
and marbles have little resistance in an ordinary fire. Sandstone will withstand
more fire effect than any of the others but often after a severe conflagration will
need to be replaced.
Kidder gives the following crushing strengths of stone and brick.
Crushing Resistance of Brick, Stone, and Concretes, (Pressure at Right

Angles to Bed)
Pounds
per Sq. In.
Brick: Common, Massachusetts. 10,000
Common, St. Louis. 6,417
Common, Washington, D. C. 7,370
Paving, Illinois. 6,000 to 13,000
Granites: Blue, Fox Island, Me. 14,875
Gray, Vinal Haven, Me. 13,000 to 18,000
Westerly, R. 1. 15,000
Rockport and Quincy, Mass. 17,750
Milford, Conn. 22,600
Staten Island, N. Y. 22,250
East St. Cloud, Minn. 28,000
Gunnison, Colo. 13,000
Red, Platte Canon, Colo. 14,600
Limestones: Glens Falls, N. Y. 11,475
Joliet, B1. 12,775
Bedford, Ind. 6,000 to 10,000
Salem, Ind. 8,625
Red Wing, Minn. 23,000
Stillwater, Minn. 10,750
Sandstones: Dorchester, N. B. (brown). 9,150
Mary’s Point, N. B. (fine grain, dark brown). 7,700
Connecticut brown stone,* on bed. 7,000 to 13,000
Longmeadow, Mass, (reddish brown). 7,000 to 14,000
Longmeadow, Mass, average, for good quality. 12,000
Little Falls, N. Y. 9,850
Medina, N. Y. 17,000
Potsdam, N. Y. (red). 18,000 to 42, Q00
Cleveland, Ohio. 6,800
North Amherst, Ohio. 6,212
Berea, Ohio. 8,000 to 10,000
Hummelstown, Pa. 12,810
Fond du Lac, Minn. 8,750
Fond du Lac, Wis. 6,237
Manitou, Colo, (light red). 6,000 to 11,000
St. Vrain, Colo, (hard laminated). 11,505
Marbles: Lee, Mass. 22,900
Rutland, Yt. 10,746
Montgomery Co., Pa. 10,000
Colton, Cal. 17,783
Italy. 12,156
Flagging: North River, N. Y. 13,425
* This stone should not be set on edge.
Fireproof Construction.—In modern structures fireproof floors and walls are a

necessity. Such floors are made by using terracotta tile or by use of some form of
corrugated steel filled or covered with concrete, cinder concrete being the
best fireproof material. Walls are made by using terracotta or expanded metal
covered with cement mortar.
Terracotta tiling is made from clay mixed with water, the plastic mass being
pressed through dies to give it the requisite shape and is then baked in high tempera¬
ture kilns.
The hollow blocks are formed with webs inside to give extra strength and to pre¬
serve the proper shape while drying. It cannot be safely made with walls over 1%
in. thick. Greater strength can be secured by filling the hollow spaces with concrete
or brick work.
The lightness of terracotta, its enormous resisting strength, its durability and
absolute indestructibility by fire, water or frost make it particularly valuable in the
construction of large edifices.
Terracotta is used for floor arches; column and girder protection and for light
hollow walls. There are three different grades made by varying the composition.
Dense tiling is made from a mixture of clays including fireclay. Porous tiling is made
by mixing sawdust with the clay which is burned out in the kiln leaving the porous
clay product. Semi-porous tiling is made by mixing finely ground coal with the
t clay. This type of terracotta tile is a better fire resistant than either of the other
types. A solid block of terracotta 1 ft. cube has borne a crushing force of over 500
tons.
Tests at Columbia University developed a crushing strength of 5,820 lb. per square
inch on dense tile and 3,292 lb. per square inch on semiporous tile.
Concrete is now almost universally used for foundations, walks, basement

floors and is coming into wider and wider use for pavements and for walks of
building and interior supports. Concrete is an artificial stone made by cementing
together sand and broken stone or similar materials by means of natural or arti¬
ficial cement. Crushed stone should be about the size of hens' eggs and should
be sharp and hard like granite or dike rock. Clean gravel may be used with
good results for floors where not too much strength is required. Cinder concrete
is cheap and satisfactory. For fireproofing this is probably the best material,
but where this is used in connection with structural steel corrosive action must
be guarded against, apparently due to the sulphur compounds in it.
The sand used should be clean coarse sand free from chemical salts. The best
proportion of cement, sand and rock depends on the qualities of each and the purposes
of the aggregate in the structure. The sand theoretically should be just the required
amount to fill the voids in the aggregate, but in practice a 5 per cent excess over
this amount seems advisable. Concrete will spall under the action of heat and if
water is put on the heated surface the action is increased.
Hydraulic Cement.—There are three principal cements used in construction
which, mixed with water and poured into the necessary forms, set and give great
resistant power to crushing. The three principal hydraulic cements are: natural
cement, portland cement, and puzzolan or slag cement.
Natural cement is the product resulting from calcination to a point below fusion
of certain natural argillaceous limestones having approximately the correct propor¬
tions for a cementing material. The limestone is not pulverized prior to calcination,
but the resulting product is ground to an impalpable powder as in the case of portland
cement. The product is normally of a yellow or brownish color. Natural cement
is far less uniform in quality than portland. It is considerably less strong and develops
its strength (although not its set) more slowly.
Natural cements are not so uniform as to color as portland cement but vary from
light to dark brown. The weight varies from 49 to 56 lb. per cubic foot and from 265
to 300 lb. per barrel. Natural cements set more quickly than portland cements,
generally within 30 min., always within an hour. Natural cement neat has a tensile
strength of at least 80 lb. per square inch when 1 week old and 120 lb. per square inch
at the end of 30 days, or when mixed with 1 part sand 40 lb. and 70 lb. respectively.
The strength of 1 to 2 natural cement mortar is about equal to a 1 to 4 portland cement
mortar.
Natural Cement Mortar and Concrete.—For stone rubble and brickwork

1 part natural cement and 3 parts sand by measure make a good mortar. For
first class brickwork 1 to 2 or 1 to IK should be used. For concretes 1 part
natural cement may be used to 2 parts sand and 4 parts gravel or crushed rock.
Kidder says “natural cement mortar possesses sufficient adhesion and crushing
strength for any ordinary masonry and when mixed in the proportion of 1 to 2 is
probably just as good for masonry above ground and for ordinary foundations as
portland cement.”
Natural cement is more suitable for works where great weight or mass rather than
strength is required, such as sewers, massive foundations, dams, piers, abutments,
etc., and especially for brick or stone work, either alone or mixed with lime. For
this class of work it is a useful structural material which will ordinarily give satis¬
factory results. French natural cements branded as “Ciment Prompt” set very
rapidly, often in 5 to 15 min., and their value is greatly decreased if they are disturbed
after this initial set has commenced. When used therefore, the concrete should not
be rammed nor tamped unless this can be done very soon after placing.
Portland Cement.—Portland cement is the product resulting from calcination

to the point of incipient fusion of an intimate mixture of calcareous and argilla¬
ceous materials. The materials commonly employed are limestone, chalk,
marl, etc., and clay, indurated clay, shale, etc. This cement has a definite
composition varying within narrow limits. The ingredients are ground to an
impalpable powder. The resulting product is normally a dull blue-gray or pearl-
gray color. Portland cement is characterized by marked uniformity of quality
and high strength rapidly acquired.
The chemical composition of a good portland cement will be about: Lime, 60.1
per cent; silica, 23.16; alumina, 8.5; ferric oxide, 5.3; magnesia, alkalis, sulphur tri¬
oxide, etc., 2.94 per cent. Portland cement weighs from 85 to 90 lb. per cubic foot.
A barrel contains 3K cu. ft. and weighs about 376 lb. net. When put up in sacks
each weighs 94 lb. or four sacks to the barrel.
Portland cement should be used for all works which will be subject to artillery,
fire, shock, abrasion, etc., for all reinforced concrete, and generally for all important
works or where great strength rapidly developed is required. French portland
cement is slower in hardening than American cements and precautions should be
taken that the forms are not removed too soon, especially if this brings a strainon the
new concrete.
Calcium sulphate delays the setting of cement. If no gypsum were added cement
would be too quick setting. Amounts up to 2 per cent S03 are desirable but the pro¬
portion should not exceed this figure.
The influence of magnesia in cement is not as yet fully determined. Some years
ago it was not thought safe to permit more than 2 per cent but now amounts up to
5 per cent are allowed. The English permit even more than do we. Magnesia
comes naturally from the rock or marl used in making the cement. The methods of
chemical analysis may be found in various textbooks, but the standard specifications
for testing portland cement and concrete are given on pp. 949 et seq.
Puzzolan or Slag Cement.—Slag cement is the product resulting from finely

pulverizing a mixture of basic furnace slag (granulated by being run into water)
with slaked lime without subsequent calcination. The product is light-blue
to lilac in color.
Slag cement sets very slowly and never attains or maintains any great degree of
hardness unless continuously wet. Accordingly it is absolutely unfitted for use
above ground where it dries out and weakens. If used below ground or wherever it
is continuously wet, it gives good service. It may be employed for reservoir linings
or large masses of masonry submerged in water or below the ground water level.
The use of slag cement is not recommended where it is practicable to obtain portland
or natural cement.
A report of U. S. army engineers on “Tests in Hydraulic Cements,, states that
“ Mortars and concretes made from puzzolan approximate in tensile strength similar
mixtures of portland cement, but their resistance to crushing is considerably less.
On account of its extreme fine grinding, puzzolan often gives nearly as great tensile
strength in 3 ; 1 mixture as neat.” Slag cement weighs 330 lb. per barrel, or 82% lb.
per sack.
Concrete Mixing.—Kidder gives the following table of material required per yard
(27 cu. ft.) of concrete of different mixtures.
Quantities Required for 1 Cu. Yd. of Rammed Concrete. (Compiled by

Edwin Thacher, C. E.)
Mixtures Stone Gravel
Sand, Stone, Sand, Gravel

Cement, Cement,
Cement Sand Stone cubic cubic cubic cubic
barrels barrels
yards yards yards yards
1 1.0 2.0 2.63 0.40 0.80 2.30 0.35 0.74

1 1.0 3.0 2.10 ’ 0.32 0.96 1.89 0.29 0.86
1* 1.5 3.0 1.90 0.43 0.87 1.71 0.39 0.78
1 1.5 4.0 1.61 0.37 0.98 1.46 0.33 0.88
1 2.0 3.0 1.73 0.53 0.79 1.54 0.47 0.73
1 2.0 4.0 1.48 0.45 0.90 1.34 0.41 0.81
1 2.0 5.0 1.29 0.39 0.98 1.17 0.36 0.89
1 2.5 4.0 1.38 0.53 0.84 1.24 0.47 0.75
1 2.5 5.0 1.21 0.46 0.92 1.10 0.42 0.83
1 2.5 6.0 1.07 0.41 0.98 0.98 0.37 0.89
1 3.0 6.0 0.91 0.42 0.97 0.84 0.38 0.89
i
Actual Volume of Rammed Concrete Resulting from Different Proportions

of Ingredients. (As Determined by Messrs. A. W. Dow and W. J.
Douglas)
Ingredients
Quantity of
1
Stone, Gravel, ' Proportions concrete,
Sand,
Cement cubic feet cubic feet cubic feet
cubic feet
1 bbl. 9 20M 0 1:2:5 21.4

cu. ft. , ii K 27 0 1:2K:6 27.66
or im 13K 13X 1:2K:3:3 27.66
378V4, lb. 13.5 45 1:3:10 45.0
For sand and gravel mixed as it comes from the pit 125 yd. will make about 100
yd. of concrete.
Hydraulic Lime.—Hydraulic lime is the product resulting from burning a
siliceous limestone. It is a natural cement containing a sufficient proportion of
free lime to cause the mass to slake. The product is white to light cream in color.
Hydraulic lime is not manufactured nor used in the United States, and should
not be confounded with hydrated lime, a mere factory slaked quicklime, not pos¬
sessing hydraulic qualities.
Hydraulic lime may be used in mortar for brick or stone work, but is unsatisfac¬
tory for concrete except in mass work, on unimportant structures where great or
rapidly acquired strength is not essential. The use of hydraulic lime for any struc¬
tural purposes except as a bond for brick or stone is not recommended where if is
practicable to obtain portland or natural cement. Hydraulic lime is very slow set¬
ting. The time of setting may be reduced by the addition of caustic soda.
STANDARD SPECIFICATIONS FOR CEMENT, AM. SOC.

FOR TEST. MAT., 1913
GENERAL CONDITIONS
1. All cement should be inspected.

2. Cement may be inspected either at the place of manufacture or on the work.
3. In order to allow ample time for inspecting and testing, the cement should bo
stored in a suitable weather-tight building having the floor properly blocked or raised
from the ground.
4. The cement should be stored in such a manner as to permit easy access for proper
inspection and identification of each shipment.
5. Every facility should be provided by the Contractor and a period of at least
12 days allowed for the inspection and necessary tests.
6. Cement should be delivered in suitable packages with the brand and name of
manufacturer plainly marked thereon.
7. A bag of cement shall contain 94 lb. of cement net. Each barrel of portland
cement shall contain 4 bags, and each barrel of natural cement shall contain 3 bags
of the above net weight.
8. Cement failing to meet the seven-day requirements may be hold awaiting the
results of the 28-day tests before rejection.
9. All tests should be made in accordance with the methods proposed by the Com¬
mittee on Uniform Tests of Cement of the American Society of Civil Engineers, pre-
sented to the Society January 21, 1903, and amended January 20, 1904, and January
15, 1908, with all subsequent amendments thereto. (See p. 951.)
10. The acceptance or rejection should be based on the following requirements:
NATURAL CEMENT
11. Definition. This term shall be applied to the finely pulverized product
resulting from the calcination of an argillaceous limestone at a temperature only
sufficient to drive off the carbonic acid gas.
Fineness
12. It shall leave by weight a residue of not more than 10 per cent on the No.
100, and 30 per cent on the No. 200 sieve.
Time of Setting
13. It shall not develop initial set in less than 10 min. and shall not develop
hard set in less than 30 min. or in more than 3 hr.
Tensile Strength
14. The minimum requirements for tensile strength for briquettes 1 sq. in. in
cross section shall be as follows, and the cement shall show no retrogression in strength
within the periods specified:
Neat Cement
Strength,
Age Pounds
24 hours in moist air. 75
7 days (1 day in moist air, 6 days in water). 150
One Part Cement, Three Parts Standard Ottawa Sand
Constancy of Volume
15. Pats of neat cement about 3 in. in diameter, y in. thick at center, tapering
to a thin edge, shall be kept in moist air for a period of 24 hr.
(а) A pat is then kept in air at normal temperature.
(б) Another is kept in water maintained as near 70°F. as practicable.
16. These pats are observed at intervals for at least 28 days, and, to satisfactorily
pass the tests, shall remain firm and hard and show no signs of distortion, checking,
cracking, or disintegrating.
PORTLAND CEMENT
17. Definition. This term is applied to the finely pulverized product resulting
from the calcination to incipient fusion of an intimate mixture of properly proportioned
argillaceous and calcareous materials, and to which no addition greater than 3 per
cent has been made subsequent to calcination.
Specific Gravity
1$. The specific gravity of cement shall not be less than 3.10 (3.07 for white port-
land). Should the test of cement as received fall below this requirement, a second
test may be made upon a sample ignited at a low red heat. The loss in weight of the
ignited cement shall not exceed 4 per cent.
Fineness
19. It shall leave by weight a residue of not more than 8 per cent on the No. 100,
and not more than 22 per cent on the No. 200 sieve.
Time of Setting
20. It shall not develop initial set in less than 30 min., and must develop hard set
in not less than 1 hr., nor more than 10 hr.
Tensile Strength
21. The minimum requirements for tensile strength for briquettes 1 sq. in. in
cross section shall be asf oil ows, and the cement shall show no retrogression in strength
within the periods specified:
Neat Cement
Strength
Age Pounds
24 hr. in moist air. 175

One Part Cement, Three Parts Standard Ottawa Sand
28 days ( 1 day in moist air, 27 days in water). 300
Constancy of Volume
22. Pats of neat cement about 3 in. in diameter, in. thick at the center, and
tapering to a thin edge, shall be kept in moist air for a period of 24 hr.
(a) A pat is then kept in air at normal temperature and observed at intervals for
at least 28 days.
(b) Another pat is kept in water maintained as near 70°F. as practicable, and
observed at intervals for at least 28 days.
(c) A third pat is exposed in any convenient way in an atmosphere of steam,
above boiling water, in a loosely closed vessel for 5 hr.
23. These pats, to satisfactorily pass the requirements, shall remain firm and
hard, and show no signs of distortion, checking, cracking, or disintegrating.
SULPHURIC ACID AND MAGNESIA
24. The cement shall not contain more than 1.75 per cent of anhydrous sulphuric
acid (SO 8), nor more than 4 per cent of magnesia (MgO).
60
REFRACTORY BRICK
Magnesia Brick are made from ground magnesite carefully selected and dead
burned. It contains little silica and lime and has just the proper fluxing qualities
to set at high temperatures. Magnesia bricks are used in basic openhearth
steel furnaces, mechanical puddling furnaces, basic copper converters and to
some extent in copper reverberatory furnaces, bullion furnaces, electric smelt¬
ing, welding furnaces, calcium carbide furnaces, kilns, rotary cement kilns, etc.
Magnesia brick should be laid in magnesite cement made of ground magnesite
and 20 to 25 per cent fireclay. They are good conductors of heat and should be
backed up with other material if used where this heat conductivity would injure
plate or frame work. They expand sufficiently on heating so that when laid as the
lining of circular converters, etc., every four or five courses should have a slip of lath
laid in which on burning out allows the brick to expand without deforming the arch.
They are better conductors of electricity than porcelain at 2,000°F. or over; at
low temperatures this conductivity is less than porcelain. The best results are obtained
from magnesia brick in furnaces where continuous heats are used. Great variations
of temperature, exposure while hot to currents of cold air or to contact while hot with
water or oil will cause them to shatter and spall. They should not be subjected to
excessive weight when hot.
Chrome brick are made from chrome ore (FeCrC^) which is exceedingly
refractory, dense and neutral. It is neither acid, basic, reducing, nor oxidizing
and is practically infusible.
Chrome brick are used in basic openhearth steel furnaces between firebrick and
magnesia brick courses on the bottom, for lining ports, uptakes, etc., along the slag
line in steel-heating furnaces, copper-smelting and refining furnaces, lining settlers,
converters next to the shell, lead softening and refining furnaces, cement kilns, etc.1
Chrome brick should be laid in finely ground chrome ore except when they are
laid next a steel plate shell when they should be laid in magnesite cement which sticks
well to iron work. Chrome bricks are useful in making quick repairs in furnaces
at working heat as they are not affected by sudden changes of temperature. At
furnace heat they can come in contact with water without breaking or spalling, but
they should not be subjected to excessive pressure when hot.
Ganister and dinas brick are English brick made from rock containing about
82 per cent and 98 per cent respectively of silica.
They are no longer much used
in this country since the development of high-grade brick manufacture here.
Glazed or vitrified brick will neither absorb nor give off heat as rapidly as
porous or non-vitrified material.
Firebrick are made by combining certain' proportions of hard or flint clays
and of soft or plastic clays. The former are the heat resisting elements of the
mass. The latter softens at a lower temperature and act as a bond binding the
hard-clay fragments together when the brick is burned giving a homogeneous
mass of requisite physical strength.
This fireclay is a combination of silica and alumina essentially always in com¬
bination with greater or less quantities of impurities such as lime, iron, magnesia,
i However, it should be noted that both magnesite and chrome brick have one great objection when
used in non-ferrous metallurgy. If used in contact with the metal they absorb some of it and it is then
difficult to recover. Particularly is this true with chrome bricks, which are impossible to smelt.—
Editor.
alkalis and titanic acid. The quality of the firebrick depends on the proportions
of the silica and alumina.
A good firebrick must give the maximum of resistance to either heat or abrasion
with the minimum of contraction or expansion.
For laying firebrick a fireclay of the same quality as the brick should be used,
using as little of it as possible, by laying a tight brick to brick joint.
The joints in brickwork are +he weakest points and if laid with inferior clay it
will melt and run before the bricks are affected leaving open points where the flame
and slag easily attack the lining. A low-grade clay will often melt and flux the higher
grade firebrick destroying the arch or walls which they compose.
Tables giving the number of brick required in circles or arches of any diameter
laid up with straight, wedge and key bricks are given in various handbooks and
catalogs of the manufacturers.
Firebrick for boiler settings must fulfill different requirements. “In some cases the
best brick to use depends entirely upon the heat-resisting quality; in others, upon
resistance to the impinging action of flame and spalling; while in others, upon the
ability to resist the action of clinker and poker, together with heat-resisting qualities.”
(Harbison-Walker.)
Silica brick are made of sand or calcined flints (almost pure silica). Con¬
taining no clay this material does not become plastic when wet, therefore, it
has to be mixed with some material to bind the silica particles together. About
1 per cent of lime is used which forms a small quantity of silicate of lime and
acts as a binder. Silica brick are very fragile and have a slight but uniform
known expansion and are better conductors of heat than fireclay brick, but they
are very refractory. They are the most widely used brick for copper smelt¬
ing furnaces, reverberatories, etc., but because of their high heat conductivity
have to be backed with firebrick. They are used in byproduct oven construc¬
tion, as they decrease the time required for coking by reason of this greater
conductivity.
Special shapes are made for roof arches, door jambs, skewbacks, etc. On account
of the severity of the service they are also used for lining the flues leading from smelting
furnaces to waste heat boilers, etc. However, they are acid and will slag in furnaces
in the presence of molten basic charges unless proper amounts of extraneous silica
or highly siliceous ores are added to the molten bath. The side walls of reverberatory
furnaces are protected by charging this siliceous material through side ports or through
holes in its roof arch. This is known as fettling.
For the best results silica brick should be laid in silica cement, composed of silica
sand and about 20 to 25 per cent fine clay. The quantity required to lay 1,000 bricks
is about 1,500 lb. or 1,000 lb. if they are dipped.
Sudden variations of temperature cause silica brick to spall and reduce their
refractoriness. Furnaces in which they are used should be heated or cooled gradually.
Weight of brickwork per cubic foot runs about as follows: Common red brick,
100 lb.; fireclay, 150 lb.; silica, 128 lb.; chrome, 175 lb.; magnesia, 160 lb.
The Harbison-Walker Refractories Co. gives the following general rules regarding
brickwork. All firebrick should be kept in a dry place. Moisture especially in
cold weather will greatly injure any firebrick.
To obtain the best results from brickwork, observe the following precautions:
Use good fireclay to lay. The fireclay should be equal in refractoriness to the brick
itself, mixing with water to a thin soup. Dip brick and rub to make a brick-to-brick
joint. From 250 to 350 lb. of fireclay are sufficient to lay 1,000 brick.
Heat slowly to expel moisture as it is absolutely essential to dry out brickwork

thoroughly. Firing the boiler while the brickwork is green will always result in open¬
ing up cracks in the setting. If possible, as soon as stack connections are made,
damper and ashpit doors should be blocked open to allow circul tion of air.
After filling the boiler with water a light fire of shavings should be kindled, gradu¬
ally increasing the wood fire, and if possible, continuing several days until the walls
are thoroughly dried out.
For estimating on firebrick work, use the following figures:

1 sq. ft. of wall requires 7 brick;
1 sq. ft. of 9-in. wall requires 14 brick;
1 sq. ft. of 13J^-in. wall requires 21 brick;
1 cu. ft. of brickwork requires 17 nine-inch straight brick;
1 cu. ft. of fireclay brickwork weighs 150 lb.;
1 cu. ft. of silica brickwork weighs 130 lb.;
1,000 brick (closely stacked) occupy 56 cu. ft.;
1,000 brick (loosely stacked) occupy 72 cu. ft.
For estimating on red brickwork, figure on 9 cu. ft. of sand and 3 bu. of lime for
laying 1,000 brick.
Common Brick.—Common brick formerly were made from clay pressed into
molds by hand or machinery and baked in kilns. At present bricks made from
sand and lime are extensively used, while they have been used in Europe for 50
years.
According to their position in the kiln, common brick are designated as arch or
hard burned brick, red or well burned brick, and salmon or soft brick. As a rule,
salmon brick are not fit for use in an exterior or bearing wall.
The color of brick depends on the presence of iron, lime, and magnesia in the
clay; iron giving them the characteristic red color when burned. There is no legal
standard size for common brick. In the Eastern states they average 7% by 3% by
2}£ in., and in the Western states 8M by 4>6 by 2% in. The weight varies from 4>£
to 5H lb. each.
Common brick are ordinarily laid in lime mortar made of 1 part lime paste and
2 parts sand, this having been determined from experiments by Government engineers
to be the best proportion. Lime for this purpose is produced by the calcining lime¬
stone or shells in kilns and is a soft white substance which slakes rapidly in open air.
It is usually sold in barrels weighing 220 lb. each or is sold by the bushel of 80 lb.;
2)4, bu. or 200 lb. being considered equivalent to a barrel. Lime paste yields about
2.62 times the bulk of unslaked lime. A barrel of average lime makes about 8 cu.
ft. or 20 buckets of lime paste. Cement mortar is also used.
To lay 1,000 common brick about 2}% bu. or 1 bbl. of lime and % yd. of sand are
required: For lime cement mortar 2 bu. lime, 1 bbl. cement and % yd. of sand: For
1:3 cement mortar, bbl. cement and % yd. of sand are required. Sand from
pits or banks free from clay and earthy material are considered best for mortar
although some river bed sand is used. Sea sand contains alkaline salts which cause
efflorescence when used in brickwork.
The use of sand in mortar is to prevent excessive shrinkage and save lime and
when used in the proportion 1:2 strengthens it. Sand added to cement gives a
weaker mixture than neat cement.
Kidder gives the following safe-working loads for masonry in walls or piers:
Tons per square

feet Eastern Western brick
brick
Red brick in lime mortar. 7 5

Red brick in natural cement mortar 1:3. 10 8
Arch brick in lime mortar. 8 6
Arch brick in natural cement. 12 9
Arch brick in Portland cement. 15 12H
Plain wall are generally figured at 15 bricks per square foot of 8 or 9 in. wall;
22}4 bricks per square foot 12 or 13 in. wall, 30 bricks per square foot of 16 or 17 in.
wall and bricks for each additional 4 or in. in thickness of the wall.
CONCRETE MAKING MATERIALS AND CEMENT TESTING

By S. S. Sadtler1
The quality of the cement to be used is determined by chemical analysis

together with physical tests. Probably the most important test of all is the
soundness or boiling test. The presence of much free lime not well distributed by
initial grinding and sintering tends to cause unsoundness in use with consequent
cracking and expansion. Much trouble may be saved by refusing shipments that
do not meet specifications in'this particular (see p. 958).
Sand.—As all kinds of local sand must be used in making concrete one cannot
be iron clad as to general specifications. Sand should always be levigated. Sea¬
shore sand is naturally free from clay. So are river sands, but bank sands must
be washed to render them suitable. This is best done by the producer who
handles the sand by conveyors whereas the user would probably have to depend
upon hand shoveling into the washing trough which costs more and the sand is not
apt to be so well washed. The product will be wet and allowance will have to be
made for the difference in volume of wet and dry sand. Sand shoveled but not
packed contains about 45 per cent voids, while sand saturated with water has
about 38 per cent. Of course, the presence of water in the voids of the sand is
considered as being the same as air. The great desideratum in concrete is to have
the maximum density (free from air when set). Sand and aggregates are used
jointly so as to secure without sacrifice of strength the greatest amount of natural
unmanufactured material suitable to the concrete. It has been found that the
use of both sand and aggregate gives the best net result. As to the sand itself
it might naturally be supposed that the greatest latitude in sizes would produce
the best result. This is not altogether the case, the best mixture being fine and
coarse, with no sand of intermediate fineness. The finer sand must be small
enough to fit into the voids of the coarse. The coarse sand should pass a 4-mesh
and be held by a 20-mesh sieve. The fine sand should pass a 40-mesh, and of
course contain no appreciable amount of clay.
The voids in fine sand (if the grains were perfect spheres) are in just the same
proportion as if the sand were all coarse, but manifestly fine particles can fill into the
1 Chemical engineer, S. P. Sadtler & Son, Philadelphia, Pa.
voids of the coarse, and what we are aiming to accomplish is this: Fill the voids of
aggregate with coarse sand, the voids of the coarse sand with fine sand and the
voids of the fine sand with Portland cement. This grading of the sand necessitates
discarding a portion of the sand after screening and using the discard for other purposes.
In practice sand is only screened for securing the best results obtainable. For ground
floors it is not necessary to have a concrete as rich in cement as is required for arches,
pillars, etc. A mixture 1:4:8 is sufficient if properly tamped or rolled. If rolled,
very light rollers are used. The thickness of this concrete is generally from 3 to 5
in. depending upon the character of the bottom.
When cinders are used more cement is required and somewhat less sand. Instead
of 1:4:8 it would probably be 1:3:6. This is due to the fines in the cinders and the
relatively large amount of surface to be covered by the cement and sand. Cinders
from power plants using soft coal only should be used and they should be screened.
All that passed a %-in. screen should be discarded.
Surfacing.—For the final coat for floors, etc. a mixture of cement and sand
only is used and this is quite rich, 1:2 being used. This is well troweled so as to
bring the cement to the surface. It is almost superfluous to say that concrete
that is not tamped must be free flowing. For surface work it is used not quite
so wet, as the trowelling is an aid in removing voids, but it must flow easily with
the aid of trowelling.
After floors have had several weeks setting and begin to dry out, wherever there
is much walking or any trucking whatever it pays to paint the surface with oil. Raw
linseed oil is used and to secure greater penetration add 1 per cent each of turpentine
and denatured alcohol. This treatment adds to the first cost but it pays in the end.
It need not be done all over the floor of the average works, where machinery is per¬
manently installed, but only in the aisles and in front of machines and wherever else
the cement would be abraded. One cubic yard of 1:2 cement mortar contains about
3 bbl. or 12 bags of cement.
Aggregates.—Broken stone and screened gravel are the best materials for
this purpose and may be used in the proportions given above, say 1:2:4 for arches
and 1:3:6 for piers. When the mixture with the cements and sand is made in
power mixers it is always more uniform and takes less cement for the same
strength of structure. Cinders are rarely used except for floors.
Small cement tanks, pipes, etc., are made of 1:3 to 1:4 mortar and tamped. The
sand is carefully selected and for nice work is graded, taking only medium coarse
and fine sand. A good formula is 20 to 40 and through 80 for small work and 10 to
20 and through 40 for larger. Waterproofing substances such as sold by the
cement companies (composed of soap mixtures) or about 1 per cent of medium heavy
lubricating oil on the weight of cement may be used, but careful proportioning of sand
and cement, careful grading of the sand, tamping or the use of a cement gun generally
renders other waterproofing unnecessary. Great care must be taken to keep work wet
for 2 weeks after the preliminary set.
For making very small tanks it is desirable to use white Portland cement and
white plasterer’s sand so as to secure a nice looking article.
Steel articles do not rust when well embedded in cement or concrete. Steel
rods are used for reinforcing nearly all concrete work where the maximum strength
with the minimum cement is aimed at. For small objects, chicken wire (large or
small mesh) or even finer screening is used.
Cement and concrete are used for holding neutral and slightly alkaline liquids.
For cases where mild acid mixtures come in contact with it, such as in superphosphate,
dense siliceous aggregates must be used with heavy walls and with moderately lean
cement mixtures that are well tamped. The writer would think 10 to 20 per cent of
ground silica could be used to advantage in the cement but as this would affect its
strength adversely the mixture could not be made so lean as in the case just mentioned.
Where concrete is laid in winter it is generally sufficiently covered with forms
to protect it in most climates. In the case of floors, marsh hay, etc., is covered over
after laying. In anything approaching zero weather salt may be added to quite a
large extent. The writer has been told by a competent engineer who has built much
in cold weather that he used freely 10 per cent of salt to the cement taken, in the water,
without seriously affecting the strength of the work.
Acidproofing.—After the surface coating has been put on concrete for floors
they can be made more or less resistant to acids in various ways. Washing with
a 20 per cent aqueous aluminum sulphate solution and, when dry, painting with
linseed oil is referred to on page 950. For making cisterns to hold acid the con¬
crete should be painted with acidproof bituminous composition and then lined
with acidproof brick laid in hot asphalt mastic.
Rosendale cement is lower in lime than portland and does not develop more than
half the strength of the latter so can not be used with as much sand and aggregate.
It is less basic and may have some application where acids would come in contact with
the work, it is practically off the market, however, in this country.
A quick and very good test as to the chemical suitability of sand or gravel
for concrete construction is to shake with a 3 per cent solution of sodium hydrox¬
ide, about 150 c.c. of solution to 100 g. of sand, allow the mixture to stand over
night and note the color of the supernatant solution. A suitable sand will give
a colorless or very light-brown solution. One containing much humus will give
a dark solution and should be rejected.
The percentage of voids may be quickly tested by taking 100 g. of dry sand
and pouring it into a dry 100 c.c. graduate, shaking or tapping until the sand is
compacted and note the volume, 7, occupied. Put 50 c.c. of water in another
100 c.c. graduate and pour the sand slowly into the water. (It is more difficult
to avoid entrapped air if the water is poured into the sand, although it can be
slowly flowed down the side of the graduate and the air slowly displaced from the
sand.) Note the volume, v, of the combined water and sand. Then:
50-(v - V)
percentage of voids =
V
Portland Cement Testing.—The best tests for portland cement are those
given by the American Society for Testing Materials, and are as follows:
1. Portland cement is the product obtained by finely pulverizing clinker produced

by calcining to incipient fusion, an intimate and properly proportioned mixture of
argillaceous and calcareous materials, with no additions subsequent to calcination
excepting water and calcined or uncalcined gypsum.
I. Chemical Properties
2. The following limits shall not be exceeded:
Loss on ignition, per cent. 4.00
Insoluble residue, per cent. 0.85
Sulphuric anhydride (SOa), per cent. 2.00
Magnesia (MgO), per cent... 5.00
II. Physical Properties
3. The specific gravity of cement shall be not less than 3.10 (3.07 for white port-
land cement). Should the test of cement as received fall below this requirement a
second test may be made upon an ignited sample. The specific gravity test will not
be made unless specifically ordered.
4. The residue on a standard No. 200 sieve shall not exceed 22 per cent by weight.
5. A pat of neat cement shall remain firm and hard, and show no signs of distor¬
tion, cracking, checking, or disintegration in the steam test for soundness.
6. The cement shall not develop initial set in less than 45 min. when the Vicat
needle is used or 60 min, when the Gilmore needle is used. Final set shall be attained
within 10 hr.
7. The average tensile strength in pounds per square inch of not less than three
standard mortar briquettes (see p. 957) composed of 1 part cement and 3 parts
standard sand, by weight, shall be equal to or higher than the following:
Age of Tensile strength,

test, Storage of briquettes pounds per
days square inch
7 1 day in moist air, 6 days in water. 200

28 1 day in moist air, 27 days in water. 300
8. The average tensile strength of standard mortar at 28 days shall be higher

than the strength at 7 days.
HI. Packages; Marking and Storage
9. The cement shall be delivered in suitable bags or barrels with the brand and
name of the manufacturer plainly marked thereon, unless shipped in bulk. A bag
shall contain 94 lb. net. A barrel shall contain 376 lb. net.
10. The cement shall be stored in such a manner as to permit easy access for proper
inspection and identification of each shipment, and in a suitable weather-tight build¬
ing which will protect the cement from dampness.
IV. Inspection
11. Every facility shall be provided the purchaser for careful sampling and inspec¬
tion at either the mill or at the site of the work, as may be specified by the purchaser.
At least 10 days from the time of sampling shall be allowed for the completion of the
7-day test, and at least 31 days shall be allowed for the completion of the 28-day
test. The cement shall be tested in accordance with the methods hereinafter pre¬
scribed. The 28-day test shall be waived only when specifically so ordered.
V. Rejection
12. The cement may be rejected if it fails to meet any of the requirements of
these specifications.
13. Cement shall not be rejected on account of failure to meet the fineness require¬
ment if upon retest after drying at 100°C. for 1 hr. it meets this requirement.
14. Cement failing to meet the test for soundness in steam may be accepted if
it passes a retest using a new sample at any time within 28 days thereafter.
15. Packages varying more than 5 per cent from the specified weight may be
rejected; and if the average weight of packages in any shipment, as shown by weigh¬
ing 50 packages taken at random, is less than that specified, the entire shipment may
be rejected.
TESTS
VI. Sampling
16. Tests may be made on individual or composite samples as may be ordered.

Each test sample should weigh at least 8 lb.
17. (a) Individual Sample.—If sampled in cars one test sample shall be taken
from each 50 bbl. or fraction thereof. If sampled in bins one sample shall be taken
from each 100 bbl.
(6) Composite Sample.—If sampled in cars one sample shall be taken from one.
sack in each 40 sacks (or 1 bbl. in each 10 bbl.) and combined to form one test sample.
If sampled in bins or warehouses one test sample shall represent not more than 200 bbl.
18. Cement may be sampled at the mill by any of the following methods that may
be practicable, as ordered:
(а) From the Conveyor Delivering to the Bin.—At least 8 lb. of cement shall be taken
from approximately each 100 bbl. passing from the conveyor.
(б) From Filled Bins by Means of Proper Sampling Tubes.—Tubes inserted ver¬
tically may be used for sampling cement to a maximum depth of 10 ft. Tubes inserted
horizontally may be used where the construction of the bin permits. Samples shall
be taken from points well distributed over the face of the bin.
(c) From Filled Bins at Points of Discharge.—Sufficient cement shall be drawn
from the discharge openings to obtain samples representative of the cement contained
in the bin, as determined by the appearance at the discharge openings of indicators
placed on the surface of the cement directly above these openings before drawing
of the cement is started.
19. Samples preferably shall be shipped and stored in airtight containers. Sam¬
ples shall be passed through a sieve having 20 meshes per linear inch in order to mix
the sample thoroughly, break up the lumps and remove foreign materials.
VII. Chemical Analysis
Loss on Ignition
20. One gram of cement shall be heated in a weighed covered platinum crucible,
of 20 to 25-c.c. capacity, as follows, using either method (a) or (b) as ordered:
(a) The crucible shall be placed in a hole in an asbestos board, clamped horizon¬
tally so that about three-fifths of the crucible projects below, and blasted at a full
red heat for 15 min. with an inclined flame; the loss in weight shall be checked by a
second blasting for 5 min. Care shall be taken to wipe off particles of asbestos that
may adhere to the crucible when withdrawn from the hole in the board. Greater
neatness and shortening of the time of heating are secured by making a hole to fit
the crucible in a circular disc of sheet platinum and placing this disc over a somewhat
larger hole in an asbestos board.
(b) The crucible shall be placed in a muffle at any temperature between 900 and
1,000°C. for 15 min. and the loss in weight shall be checked by a second heating for
5 min.
21. A permissible variation of 0.25 will be allowed, and all results in excess of the
specified limit but within this permissible variation shall be reported as 4 per cent.
Insoluble Residue
22. To a 1-g. sample of cement shall be added 10 c.c. of water and 5 c.c. of con¬
centrated hydrochloric acid; the liquid shall be warmed until effervescence ceases.
The solution shall be diluted to 50 c.c. and digested on a steam bath or hot plate until
it is evident that decomposition of the cement is complete. The residue shall be
filtered, washed with cold water, and the filter paper and contents digested in about
30 c.c. of a 5 per cent solution of sodium carbonate, the liquid being held at a tempera¬
ture just short of boiling for 15 min. The remaining residue shall be filtered, washed
with cold water, then with a few drops of hot hydrochloric acid, 1:9, and finally
with hot water, and then ignited at a red heat and weighed as the insoluble residue.
specified limit but within the permissible limit shall be reported as 0.85 per cent.
Sulphuric Anhydride
24. One gram of the cement shall be dissolved in 5 c.c. of concentrated hydro¬
chloric acid diluted with 5 c.c. of water, with gentle warming; when solution is coim
plete 40 c.c. of water shall be added, the solution filtered, and the residue washed
thoroughly with water. The solution shall be diluted to 250 c.c., heated to boiling
and 10 c.c. of a hot 10 per cent solution of barium chloride shall be added slowly,
drop by drop, from a pipette and the boiling continued until the precipitate is well
formed. The solution shall be digested on the steam bath until the precipitate has
settled. The precipitate shall be filtered, washed, and the paper and contents placed
in a weighed platinum crucible and the paper slowly charred and consumed without
flaming. The barium sulphate shall then be ignited and weighed. The weight
obtained multiplied by 34.4 gives the percentage of sulphuric anhydride. The
acid filtrate obtained in the determination of the insoluble residue may be used for
the estimation of sulphuric anhydride instead of using a separate sample.
specified limit but within this permissible variation shall be reported as 2.00 per
cent. *
Magnesia
26. To 0.5 g. of the cement in an evaporating dish shall be added 10 c.c. of water
to prevent lumping and then 10 c.c. of concentrated hydrochloric acid. The liquid
shall be gently heated and agitated until attack is complete. The solution shall then
be evaporated to complete dryness on a steam or water bath. To hasten dehydration
the residue may be heated to 150 or even 200°C. for to 1 hr. The residue shall
be treated with 10 c.c. of concentrated hydrochloric acid diluted with an equal
amount of water. The dish shall be covered and the solution digested for 10 min.
on a steam bath or water bath. The diluted solution shall be filtered and the sepa¬
rated silica washed thoroughly with water.1 Five cubic centimeters of concentrated
hydrochloric acid and sufficient bromine water to precipitate any manganese which
may be present, shall be added to the filtrate (about 250 c.c.). This shall be made
alkaline with ammonium hydroxide, boiled until there is but a faint odor of ammonia,
1Since this procedure does not involve the determination of silica, a second evaporation is
unnecessary.
and the precipitated iron and aluminum hydroxides, after settling, shall be washed
with hot water, once by decantation and slightly on the filter. Setting aside the
filtrate, the precipitate shall be transferred by a jet of hot water to the precipitating
vessel and dissolved in 10 c.c. of hot hydrochloric acid. The paper shall be extracted
with acid, the solution and washings being added to the main solution. The alumi¬
num and iron shall then be reprecipitated at boiling heat by ammonium hydroxide
and bromine water in a volume of about 10 c.c., and the second precipitate shall be
collected and washed on the filter used in the first instance if this is still intact. To
the combined filtrates from the hydroxides of iron and aluminum, reduced in volume
if need be, 1 c.c. of ammonium hydroxide shall be added, the solution brought to
boiling, 25 c.c. of a saturated solution of boiling ammonium oxalate added, and the
boiling continued until the precipitated calcium oxalate has assumed a well-defined
granular form. The precipitate after 1 hr. shall be filtered and washed, then with the
filter shall be placed wet in a platinum crucible, and the paper burned off over a small
flame of a bunsen burner; after ignition it shall be redissolved in hydrochloric acid and
the solution diluted to 100 c.c. Ammonia shall be added in slight excess, and the
liquid boiled. The lime shall then be reprecipitated by ammonium oxalate, allowed
to stand until settled, filtered and washed. The combined filtrates from the calcium
precipitates shall be acidified with hydrochloric acid, concentrated on the steam bath
to about 150 c.c., and made slightly alkaline with ammonium hydroxide, boiled and
filtered (to remove a little aluminum and iron and perhaps calcium). When cool,
10 c.c. of saturated solution of sodium-ammonium-hydrogen phosphate shall be added
with constant stirring.1 When the crystalline ammonium-magnesium orthophosphate
has formed, ammonia shall be added in moderate excess. The solution shall be set
aside for several hours in a cool place, filtered and washed with water containing
2.5 per cent of NH3. The precipitate shall be dissolved in a small quantity of hot
hydrochloric acid, the solution diluted to about 100 c.c., 1 c.c. of a saturated solution
of sodium-ammonium-hydrogen phosphate added, and ammonia drop by drop, with
constant stirring, until the precipitate is again formed as described and the ammonia
is in moderate excess. The precipitate shall then be allowed to stand about 2 hr.,
filtered and washed as before. The paper and contents shall be placed in a weighed
platinum crucible, the paper slowly charred, and the resulting carbon carefully
burned off. The precipitate shall then be ignited to constant weight over a Meker
burner, or a blast not strong enough to soften or melt the pyrophosphate. The
weight of magnesium pyrophosphate obtained multiplied by 72.5 gives the per¬
centage of magnesia. The precipitate so obtained always contains some calcium and
usually small quantities of iron, aluminum, and manganese as phosphates.2
specified limit but within this permissable variation shall be reported as 5.00 per cent.
VIII. Determination of Specific Gravity

28. The determination of specific gravity shall be made with a standardized Le
Chatelier apparatus. This apparatus is standardized by the United States Bureau of
Standards. Kerosene free from water, or benzine not fighter than 62°B6, shall be
used in making this determination.
1 In this stirring the sides of the beaker should be hit by the stirring rod as little as possible, for
where the beaker is touched the precipitate sticks tenaciously. Some authorities say complete precipi¬
tation is not obtaining except by “scratching the beaker.” Undoubtedly higher results can be
obtained by scratching off some of the beaker into the precipitate, but the result is not ammonium-
magnesium phosphate.—Editor.
2 If the last washing is given with a 10 per cent solution of ammonium nitrate containing 2.5 per
cent of free ammonia it is much easier to burn the precipitate white. Indeed it can usually be done
without blasting. The use of an asbestos mat on a tared Gooch crucible is infinitely preferable to filter
paper for filtering magnesias.—Editor.
29. The flask shall be filled with either of these liquids to a point on the stem
between zero and 1 c.c., and 64 g. of cement, of the same temperature of the liquid,
shall be slowly introduced, taking care that the cement does not adhere to the inside
of the flask above the liquid and to free the cement from air by rolling the flask in an
inclined position. After all the cement is introduced, the level of the liquid will
rise to some division of the graduated neck; the difference between readings is the
volume displaced by 64 g. of the cement.
The specific gravity shall be obtained from the formula
0 weight of cement grams
peci c gravi y — displaced voiume cubic centimeters
30. The flask, during the operation, shall be kept immersed in water, in order to
avoid variations in the temperature of the liquid in the flask, which shall not exceed
0.5°C. The results of repeated tests shall agree within 0.01.
31. The determination of specific gravity shall be made on the cement as received;
if it falls below 3.10, a second determination shall be made after igniting the sample
as described in Section 20.
IX. Determination of Fineness

32. Wire cloth for standard sieves for cement shall be woven (not twilled) from
brass, bronze, or other suitable wire, and mounted without distortion on frames not
less than 1J4 in., below the top of the frame. The sieve frames shall be circular,
approximately 8 in. in diameter, and may be provided with a pan and cover.
33. A standard No. 200 sieve is one having nominally an 0.0029-in. opening and
200 wires per inch standardized by the U. S. Bureau of Standards, and conforming
to the following requirements:
The No. 200 sieve should have 200 wires per inch, and the number of wires in
any whole inch shall not be outside the limits of 192 to 208. No opening between
adjacent parallel wires shall be more than 0.0050 in. in width. The diameter of the
wire shall be more than 0.0050 in. in width. The diameter of the wire should be
0.0021 in. and the average diameter shall not be outside the limits of 0.0019 to
0.0023 in. The value of the sieve as determined by sieving tests made in conformity
with the standard specification for these tests on a standardized cement which gives a
residue of 25 to 20 per cent on the No. 200 sieve, or on other similarly graded material,
shall not show a variation of more than 1.5 per cent above or below the standards
maintained at the Bureau of Standards.
34. The test shall be made with 50 g. of cement. The sieve shall be thoroughly
clean and dry. The cement shall be placed on the No. 200 sieve, with pan and cover
attached, if desired, and shall be held in one hand in a slightly inclined position so
that the sample will be well distributed over the sieve, at the same time gently
striking the side about 150 times per minute against the palm of the other hand on the
up stroke. The sieve shall be turned every 25 strokes about one-sixth of a revolution
in the same direction. The operation shall continue until not more than 0.05 g.
passes through in 1 min. of continuous sieving. The fineness shall be determined from
the weight of the residue on the sieve expressed as a percentage of the weight of the
original sample.
35. Mechanical sieving devices may be used, but the cement shall not be rejected
if it meets the fineness requirement when tested by the hand method described in
Section 34.
36. A permissible variation of 1 per cent will be allowed, and all results in excess
of the specified limit but within this permissible variation shall be reported as 22 per
cent.
X. Mixing Cement Pastes and Mortars
37. The quality of dry material to be mixed at one time shall not exceed 1,000 g.
nor be less than 500 g. The proportions of cement or cement and sand shall be
stated by weight in grams of the dry materials; the quantity of water shall be expressed
in cubic centimeters (1 c.c. of water = 1 g.). The dry materials shall be weighed,
placed upon a non-absorbent surface, thoroughly mixed dry if sand is used, and a
crater formed in the center, into which the proper percentage of clean water shall
be poured; the material on the outer edge shall be turned into the crater by the aid
of a trowel. After an interval of min. for the absorption of the water the operation
shall be completed by continuous, vigorous mixing, squeezing and kneading with the
hands for at least 1 min.1 During the operation of mixing, the hands should be pro¬
tected by rubber gloves.
38. The temperature of the room and the mixing water shall be maintained as
nearly as practicable at 21 °C. (70°F.).
XI. Normal Consistency
39. The Vicat apparatus consists of a frame bearing a movable rod weighing 300 g.,
one end being 1 cm. in diameter for a distance of 6 cm., the other having a removable
needle, 1 mm. in diameter, 6 cm. long. The rod is reversible, and can be held in any
desired position by a screw, and has midway between the ends a mark which moves
under a scale (graduated to millimeters) attached to the frame. The paste is held
in a conical, hard-rubber ring, 7 cm. in diameter at the base, 4 cm. high, resting on a
glass plate about 10 cm. square.
40. In making the determination 500 g. of cement, with a measured quantity of
water, shall be kneaded into a paste, as described in Section 37, and quickly formed
into a ball with the hands, completing the operation by tossing it six times from one
hand to the other, maintained about 6 in. apart; the ball resting in the palm of one
hand shall be pressed into the larger end of the rubber ring held in the other hand,
completely filling the ring with paste; the excess at the larger end shall be removed by
a single movement on the palm of the hand; the ring shall then be placed on its larger
end on a glass plate and excess paste at the smaller end sliced off at the top of
the ring by a single oblique stroke of a trowel held at a slight angle with the top of
the ring. During these operations care shall be taken not to compress the paste.
The paste confined in the ring, resting on the plate, shall be placed under the rod, the
larger end of which shall be brought in contact with the surface of the paste; the
scale shall be then read, and the rod quickly released. The paste shall be of normal
consistency when the rod settles to a point 10 mm. below the original surface in
min. after being released. The apparatus shall be free from all vibrations during
the test. Trial pastes shall be made with varying percentages of water until the
normal consistency is obtained. The amount of water required shall be expressed in
percentage by weight of the dry cement.
41. The consistency of standard mortar shall depend on the amount of water
required to produce a paste of normal consistency from the same sample of cement.
Having determined the normal consistency of the sample, the consistency of standard
mortar made from the same sample shall be as indicated in Table I, the values being
in percentage of the combined dry weights of the cement and standard sand.
1 In order to secure uniformity in the results of tests for the time of setting and tensile strength the
manner of mixing above described should be carefully followed. At least 1 min. is necessary to
obtain the desired plasticity which is not appreciably affected by continuing the mixing for several
minutes. The exact time necessary is dependent upon the personal equation of the operator. The
Table I.—Percentage of Water foe Standard Mortars
Percentage of water Percentage of water Percentage of water Percentage of water

for neat cement for one cement, for neat cement for one cement,
paste of normal three standard paste of normal three standard
consistency Ottawa sand consistency Ottawa sand
15 9.0 23 10.3
16 9.2 24 10.5
17 9.3 25 10.7
18 9.5 26 10.8
19 9.7 27 11.0
20 9.8 28 11.2
21 10.0 29 11.3
22 5 10.2 30 11.5
XII. Determination of Soundness1
42. A steam apparatus, which can be maintained at a temperature between 98

and 100°C., is to be used. The capacity of this apparatus may be increased by using
a rack for holding the pats in a vertical or inclined position.
43. A pat from cement paste of normal consistency about 3 in. in diameter, in.
thick at the center, and tapering to a thin edge, shall be made on clean glass plates
about 4 in. square, and stored in moist air for 24 hr. In molding the pat, the cement
paste shall be first flattened on the glass and the pat then formed by drawing the
trowel from the outer edge toward the center.
44. The pat shall then be placed in an atmosphere of steam at a temperature
between 98 and 100°C. upon a suitable support 1 in. above boiling water for 5 hr.
45. Should the pat leave the plate, distortion may be detected best with a straight
edge applied to the surface which was in contact with the plate.
XIII. Determination of Time of Setting
46. The following are alternate methods, either of which may be used as ordered:
47. The time of setting shall be determined with the "Vicat apparatus described
in section 39.
48. A paste of normal consistency shall be molded in the hard-rubber ring as
described in Section 40, and placed under the rod, the smaller end of which shall then
be carefully brought in contact with the surface of the paste, and the rod quickly
released. The initial set shall be said to have occurred when the needle ceases
to pass a point 5 mm. above the glass plate in ^ min. after being released; and the
final set, when the needle does not sink visibly into the paste. The test pieces shall
be kept in moist air during the test. This may be accomplished by placing them on a
rack over water contained in a pan and covered by a damp cloth, kept from contact
• with them by means of a wire screen; or they may be stored in a moist closet. Care
shall be taken to keep the needle clean as the collection of cement on the sides of the
needle retards the penetration, while cement on the point may increase the penetra-
1 Unsoundness is usually manifested by change in volume which causes distortion, cracking, checking
or disintegration.
Pats improperly made or exposed to drying may develop what are known as shrink-cracks within
the first 24 hrs. and are not an indication of unsoundness.
tion. The time of setting is affected not only by the percentage and temperature
of the water used and the amount of kneading the paste received, but by the tempera¬
ture and humidity of the air, and its determination is therefore only approximate.
49. The time of setting shall be determined by the Gilmore needles.
50. The time of setting shall be determined as follows: A pat of neat cement paste
about 3 in. in diameter and in. in thickness with a flat top, mixed to a normal con¬
sistency, shall be kept in moist air at a temperature maintained as nearly as prac¬
ticable at 21°C. (70°F.). The cement shall be considered to have acquired its initial
set when the pat will bear, without appreciable indentation, the Gilmore needle
34 2 in. in diameter, loaded to weigh 34 15. The final set has been acquired when the
pat will bear without appreciable indentation, the Gilmore needle 3^4 in* in diam¬
eter, loaded to weigh 1 lb. In making the test, the needles shall be held in a vertical
position, and applied lightly to the surface of the pat.
XIV. Extension Tests
51. The usual form of test piece shall be used. The molds shall be made of non¬
corroding metal and have sufficient material in the sides to prevent spreading during
molding. Molds shall be wiped with an oily cloth before using.
52. The sand to be used shall be natural sand from Ottawa, Ill., screened to pass
a No. 20 sieve and retained on a No. 30 sieve. This sand may be obtained from the
Ottawa Silica Co., at a cost of 2 cts. per pound, f.o.b. cars, Ottawa, Ill.
53. This sand, having passed the No. 20 sieve, shall be considered standard when
not more than 5 g. pass the No. 30 sieve after 1 min. continuous sieving of a 500-g.
sample.
54. The sieves shall conform to the following specifications:
The No. 20 sieve shall have between 19.5 and 20.5 wires per whole inch of the
warp wires and between 19 and 21 wires per whole inch of the woof wires. The
diameter of the wire should be 0.0165 in. and the average diameter shall not be out¬
side the limits of 0.0160 and 0.0170 in.
The No. 30 sieve shall have between 29.5 and 30.5 wires per whole inch of the
warp wires and between 28.5 and 31.5 wires per whole inch of the woof wires. The
diameter of the wire should be 0.0110 in. and the average diameter shall not be out¬
side the limits 0.0105 to 0.0115 in.
55. Immediately after mixing, the standard mortar shall be placed in the molds,
pressed in firmly with the thumbs and smoothed off with a trowel without ramming.
Additional mortar shall be heaped above the mold and smoothed off with a trowel;
the trowel shall be drawn over the mold in such a manner as to exert a moderate
pressure on the material. The mold shall then be turned over and the operation of
heaping, thumbing and smoothing off repeated.
56. Tests shall be made with any standard machine. The briquettes shall be
tested as soon as they are removed from the water. The bearing surfaces of the clips
and briquettes shall be free from grains of sand or dirt. The briquettes shall be care¬
fully centered and the load applied continuously at the rate of 600 lb. per minute.
57. Testing machines should be frequently calibrated in order to determine their
accuracy.
58. Briquettes that are manifestly faulty, or which give strengths differing more
than 15 per cent from the average value of all test pieces made from the same sample
and broken at the same period, shall not be considered in determining the tensile
strength.
XV. Storage of Test Pieces
59. The moist closet may consist of a soapstone, slate or concrete box, or a wooden
box lined with metal. If a wooden box is used, the interior should be covered with
felt or broad wicking kept wet. The bottom of the moist closet should be covered
with water. The interior of the closet should be provided with non-absorbent shelves
on which to place the test pieces, the shelves being so arranged that they may be
withdrawn readily.
60. Unless otherwise specified all test pieces, immediately after molding, shall be
placed in the moist closet for from 20 to 24 hr.
61. The briquettes shall be kept in molds on glass plates in the moist closet for
at least 20 hr. After 24 hr. in moist air the briquettes shall be immersed in clean
water in storage tanks of non-corroding material.
62. The air and water shall be maintained as nearly as practicable at a tempera¬
ture of 21°C. (70°F.).
SECTION XXX
THE THEORY OF PLANT LOCATION AND DESIGN

By Henry H. Buckman1
Introductory.—To select and constantly maintain a definite point of view is

the first duty of the designer. At first blush it might appear that the design
of chemical engineering plants, owing to the multiplicity of the processes involved
and the apparently highly specific nature of almost every case, is a subject impos¬
sible to treat as a coordinated whole. But a closer view of the field will disclose
a remarkable simplicity of the underlying structure, obscured it is true, by an
infinity of details, but everywhere yielding to a common system of analysis. The
plant embodying truly unique features is rare indeed. The exchange of a given
amount of energy between given masses of material with corresponding increase
of entropy of the system involved; the whole covering given changes of time and
space—that is the whole of any art. Our methods of causing, controlling and
measuring such changes are rather less than more than enough to cause embar¬
rassment to the engineer.
In order intelligently to select his point of view, it is absolutely necessary that
the designer give a large amount of study to certain general aspects of the field within
which his design is expected to function. Too much stress cannot be laid upon the
preliminary study. The success or failure of a design will certainly depend upon it.
Luck seldom favors the careful designer and never the careless one. Yet the number
of really unknown conditions surrounding all ordinary materials, insofar as they
affect practice in the arts utilizing those materials, is surprisingly few. The steps
by which the designer may achieve the comprehensive point of view so necessary to
a true design are so plainly indicated as to leave little excuse for a failure to meet
and provide in advance for all conditions likely to arise. It should be remembered
that that design is a failure which condemns a given amount of capital to turn over and
over in a groove without a chance of return, just as much as is that set of blue prints
and specifications which commits so much labor and material to the scrap heap.
Both are graveyards for capital.
Those general considerations which govern the detail of all designing work, and
which the designer must thoroughly study before proceeding to formulate his plans
at all are the following:
1. The nature of the investment.
2. The nature of the industry.
3. The type process. ,
4. The location: geographic, topographic, economic.
When the designer has accumulated all available data under these four general
heads, and has thoroughly digested the significance of each point developed, he will
be in a position to view his enterprise as a whole, and his ability to do this will indicate
his caliber as a designer.
1 President, The Buckman Corporation, Jacksonville, Fla.

61 961
Up to this point the work of the designer is assimilative. From this point forward,
the work becomes creative, and necessarily enters the phase of detail. This should
be thoroughly systematized and subdivided. The following system is recommended
as representing the minimum of care and division allowable, and yet one which,
conscientiously followed, should insure a successful design. Each division should be
studied in turn, and each with reference to all the others. In the order recommended
they are:
Type of construction.
Influence of separate related units.
Control.
(a) Administrative.
(b) Labor,
(c) Time and material,
(d) Temperature and moisture, and
(e) Light.
Power distribution.
Waste and by-product disposal.
Sanitation.
Hazards and insurance.
Storage.
Drafting the design.
Each of these subdivisions is treated in more detail in later paragraphs. We
return now to a detailed consideration of the four fundamental general aspects govern¬
ing design, the first of which is:
The Nature of the Investment.—That the nature of the investment itself
should be the fundamental governing factor in all plant design seems a fact too
obvious to be mentioned, but experience has shown that this supposedly first
principle of commercial design is more often violated than is any other, and to
this fact is due a large percentage of failures. The engineer should bear in mind
that the primary object of the plant he is to design is to return earnings on a cer¬
tain investment—the making of a product must always be a secondary object,
no matter how important. When this ceases to be the case, we have almost
invariably failure of both earnings and product. It is only necessary to consider
the course of certain war-time engineering activities to find ample illustration of
this.
The engineer is essentially an economist,1 and he is a poor engineer who loses sight
of the fact that the plant which he is about to design is simply an unavoidable detail
in achieving the main object of his activity—the creation of wealth. He should keep
constantly before him not only the size of the investment proposed, but the entircr-
scheme of financing. He should know and weave into his design the rate of interest
demanded, the number of years allowed for amortization of any proposed debt, the
net earnings expected, and the number and general character of the securities of
which the proposed plant will be the physical basis. Each of these factors should
be carefully considered, and only that design is well conceived which gives due regard
to all. Just as the relation of working capital to plant investment would govern
the design controlling goods in process, so should the amortization period and deprecia¬
tion factor influence the choice of construction type and materials. And so with all
these factors—each will be found to govern some one or more vital detail of the design}
1 An old definition of an engineer was “one who could do with $1 what any fool could do with $2.”—
Editor.
THE THEORY OF PLANT LOCATION AND DESIGN 963
and when he finds that he cannot square the design with their demands, the wise
engineer realizes that he has been attempting the commercially impossible.
A proper appreciation of the nature of the investment is impossible without due
consideration of the scale of projected operations. All industrial processes have an
optimum scale upon which they may be carried on. More important still, all have a
minimum1 scale upon which they can be profitably engaged in. Minimum, optimum
and actual proposed scale of operations should therefore be first of all determined by
the designer, and set up as a guide in all subsequent decisions.
Nature of the Enterprise.—After the nature of the investment has been con¬
sidered, the next most fundamental point which should engage the attention of
the designer is the nature of the enterprise itself. By this is meant the general
type of the industry in which the proposed design will function; whether public
service or private; whether a well established process with an accumulation of
standard practice behind it, and if so, whether it represents a large or small class
of producers, or whether the proposed design represents a launching into a more
or less untried industry, where the unexpected is certain to happen.
Regardless of the nature of the product to be dealt with, a design expected to

function in public service is surrounded by more rigid limitations than is the design
for private service. With public-service designing, permanency of construction and
efficiency of operation invariably take precedence over first cost, while in private
service, the reverse is often properly the case. On the other hand, both topographic
and geographic location of the public-service design very often inject factors into the
designer’s problem which should seldom, if ever, appear with good reason in the case
of private service, where the designer is usually free to exercise his prior judgment as
to location.
The question as to whether the enterprise represents a small or large class of pro¬
ducer is one which the designer should consider most carefully. The smaller the class,
the more important becomes the factor of location with all its attendant conditions.
In any event, and especially if the class be a large one, the designer will do well to
select from all available examples the best type of construction and arrangement as
proved by experience. Too often, this obvious mode of procedure is violated by the
designer.
Whether from unwillingness to recognize the value of the experience of others,
or from sheer infatuation with certain ideas of his own, there are frequently found
incorporated in a design details which other plants have long since proven inefficient
or inoperable; all of which a careful study of the field would have brought to light.
It has been said, and with considerable truth, that whereas the number of new and
good ideas yet to be conceived and tried out is legion, on the other hand, all possible
mistakes have already been made. The difficulty probably lies in the fact that while
successful practice is common knowledge, the man who has made an expensive mis¬
take usually considers it as his personal property. Even so, there is very little excuse
for mistakes in design in standard practice. That mistakes are numerous and that
they are made over and over again is due chiefly to the failure of the designer to in¬
vestigate the mistakes of existing industries of the same kind.
Some latitude must be allowed the designer in a new field, but even here, the only
legitimate excuse for failure of the design is unavoidable ignorance of certain charac¬
teristics of the materials themselves, which develop only after the design is complete.
Modern chemical engineering methods have been so standardized that, given a com¬
plete knowledge of all the chemical and physical constants of the materials to be
1 As interestingly pointed out by J. It. Finlay in his discussion of the Utah Copper Co’s, scale of
operations.
treated, the designer has only to select known processes best fitted to these constants.
A critical examination of mistakes in chemical engineering design will usually show
that failure is the result of the designer’s not allowing for the behavior of a certain
material under certain conditions—behavior which could have been predicted in
nine cases out of ten.
When the contemplated design is in a new field, this study of the physical and
chemical constants and characteristics of the materials and products becomes pre¬
eminently important. If these constants are not known, they should be determined
with the greatest care before the designing work is begun. These should include
specific gravity, viscosity, behavior toward moisture and behavior under varying
temperatures, and varying conditions of light, magnetic and static charges; critical
angle of fall both wet and dry, hardness, friability, heat and electrical conductivity,
solubility, specific heat, latent heat of vaporization, behavior in storage, results of
friction, influence of pressure, etc., etc., of both original materials and intermediate
and final products. It is a vicious mistake, but one made only too often, to assume that
because the new material to be dealt with has superficially the character of, say,
a syrup, that it will behave like a syrup; or because it has apparently a sandy nature,
that it can be handled as such.
Type Process.—There are but five types of process involved in any chemical plant.
These are transporting, energizing, reacting, separating and conditioning. The innum¬
erable details and subdivisions under these five heads constitute proper subjects for
machine design, but they have no place in the general subject of plant design. Plant
design is the physiology of an industry; machine design is its anatomy. Plant design
deals with the correlation of functions, machine design with construction and work.
In order, however, that the designer may keep more clearly in view these five processes
and their government of his design, they may be approximately delimited as follows:
Transporting Separating Conditioning Reacting
Rail, cable and Filters Grinders Vats

other ways Dryers Crushers Tanks
Drags Separators Briquetting machines Smelting furnaces
Elevators Collectors Mixers Cells
Conveyors Stills Shredders Combustion chambers
Pipes Skimmers Digesters Retorts
Pumps Sifters Rolls Roasters
Chutes Settlers Atomizers
Energizing
Feeders Strainers Sprayers Heating equipment
Conduits Extractors Gasifiers Cooling equipment
Ditches Dewaterers Compressors Exchangers
Conductor wires Centrifuges
Loading and unload- Flotation appara-
ing machines tus
Canals Wet tables
Evaporators
Presses
Non-chemical pre¬
cipitators
Sizers
With these five major processes in mind, the designer should weigh carefully the
various phases of the proposed industry as affected by each, and he should clearly
determine and set down in their order of importance the relative position of these
processes. By importance, should be understood not the numerical frequency or

preponderance in the enterprise of one of the five major processes, but that process
involved which represents the greatest percentage of the production cost (all things con¬
sidered). The consideration of a few specific instances will not be without value.
In the manufacture of sulphuric acid, dyestuffs and rubber goods, reaction processes
are the governing factors in design. The design of plants producing cereal products,
glue and glass will have conditioning processes as the limiting factors. The produc¬
tion of steel and fertilizers represents chiefly design in transportation processes.
Carborundum and graphite call primarily for attention to energizing processes.
Thorium nitrate and salt production have separation as their type of process. The
failure to recognize the type process represented by a proposed plant is one of the
most fruitful and common sources of error in design.
Not until the designer has settled clearly in his mind this question of the type
processes involved; not until he can vision a clear picture of the relative places to be
occupied in his design by each type process (using always production cost involved
as the criterion of importance) is he prepared to proceed with the next fundamental
consideration which will govern his design, namely location of the plant.
Location.—Three phases of the location of the proposed plant will influence

the design. These are geographic, topographic, economic. Upon the geographic
location will depend to a considerable extent the materials of construction, the
rain, snow, ice, heat, wind, sunlight and moisture protection and accommodation
features. This refers to physical limitations imposed by climate and other truly
geographic factors.
In all locations reinforced concrete is probably the best of all materials for per¬
manent structures, where the height does not exceed 80 ft., and where the foundations
do not enter salt water. For less permanent construction, the choice lies between
galvanized and black iron for roofing and sheathing, and between steel or wooden
framing. In the case of wood, there is the choice of treated or untreated materials.
The economic factors enter so largely into these selections, that it is better to view the
geographic factors as being of a negative and limiting type. In general, it may be
said that for plants located between 35° N. and 35° S. latitude, except for the most
temporary structures, the less wood used the better. In higher latitudes, where
bacterial action and sun effects are less acute, and where the question of excessive sweat¬
ing makes iron structures objectionable, wood may well be the predominating material.
As between galvanized iron and black iron, the former is always the best prac¬
tice, except for plants located below 35° latitude and on the seacoast. Here, experience
has shown that well-painted black iron will outlast the galvanized material. This
applies to pipes and ^fittings, as well as to sheathing. It should be borne in mind
that wherever great humidity and marked changes of temperature are consistently
present, the designer must expect metal roofs and walls to sweat at intervals, and this
may have very serious results on process or product. Sweating of the roof is par¬
ticularly to be guarded against.
The question of nails and general ironing should be very carefully considered with
regard to geographical location. In low latitudes near the seacoast, galvanized nails
and irons should always be specified. In studying the geography of the proposed
location, particular attention should be paid to rainfall and snowfall data. This is
especially important in the case of roof design.
Below a latitude of 30° and an altitude of 400 ft., the designer need make no special
provision for ice. Above this latitude and altitude, however, this feature becomes
of paramount importance. The designer must allow for protection of all liquids and
gases and associated service apparatus in storage, process and transit, for the contrac-
tion of all parts other than wood, for mechanical failures due to loss of friction from
ice coatings, for mechanical strains due to weight of accumulated ice and for the
increased accident hazard to employees. The design of piping systems is too large
a subject to be more than touched upon here, but it may be well to call attention
to a few of the more important points. One expansion joint for all straight runs of
above %-in. size is necessary. Wide-angle elbows on all steam and other piping are
recommended where intermittent pressures are likely to obtain, or “water hammer”
to develop. All automatic valves need especial protection. Few such will fail to
give trouble if unprotected from low temperatures. Especial care should be taken
with the layout to avoid “traps” which will not drain automatically. Wherever
possible, all gravity return drains should be given a pitch of not less than }{ in. to
the lineal foot. Flexible hangers should be used in all cases. Close return bends
and straight ells should be avoided wherever possible. Whenever there is a choice,
piping should be on interior south walls, or overhead, when the design calls for more
than one story. When only one story is contemplated, placing the great valve of the
permanent supply pipes below the floors and beneath the frost line is often well worth
the additional cost. In any event, pet cocks for bleeders and stop-and-waste cocks
should be plentifully supplied in all lines. Flanged-joint piping and fittings must be
particularly guarded, and especially should the designer avoid curves and offsets in
this variety of piping, unless special flanges are used. Expansion washers should be
allowed on all flange bolts.
Expansion joints in all horizontal concrete surfaces should be provided. Such
joints should pierce the entire thickness of the concrete floor or other surface, and
should be from 34 in. to % in. in width for sections not exceeding 10 ft. in length along
the line of expansion. Asphalt, gilsonite and lead wool are the best fillers for such
joints. Concrete floors, platforms and steps should all be designed with a minimum
slope to drains of J-fe in. to the lineal foot, and “dips” and depressions carefully
avoided. If such surfaces are to be exposed to more than an ordinary amount of
ice, the use of thin finishing coats should be avoided.
In connection with the geographical location of the design, temperature records
of the locality should be carefully studied—daily, monthly and annual means, minima
and maxima. In localities subject to extremely high temperatures, all-metal sheath¬
ing and roofing is usually to be preferred to wood. In such locations storage should
not be provided or contemplated in lofts directly beneath the roof, unless very inert
materials are to be stored. Temperatures of 140°F. are not uncommon in such con¬
ditions, and must be experienced to be appreciated. The overhang width of eaves
is worthy of attention in this connection. Nine to 18 in. will suffice in temperate
latitudes, while 18 to 36 in. is recommended for lower latitudes. It is the duty of
the designer to eliminate as far as possible by the fundamentals of his layout the
necessity for temperature control. ■
Mean and maximum wind velocities should receive careful attention. Likewise,
the direction of prevailing winds should be noted and taken into account when orient¬
ing the plan. Fire doors of boilers and furnaces and doors of all chambers where
hot or finely divided materials are handled should whenever possible be placed to the
leeward. The same holds for drying-room doors, and doors opening to rooms where
gravity feeds of light materials are used. In many geographical locations the
maximum velocity of the prevailing winds is the limiting factor in the height of
structures. Wind pressure should be carefully studied and allowed for.
When studying the geographical location, relative humidity records should be
carefully taken into account. Many processes are vitally affected by moisture in the
air, and without artificial conditioning of the plant atmosphere, it is frequently im¬
possible to operate both night and day shifts owing to the greatly varying humidity
in some localities.
In designing all refrigerating or supercooled systems, it is especially necessary

to study the geographical location in connection with the relative humidity records.
Excessive humidity is often a most troublesome factor, and in some localities may
become prohibitive.
In introducing internal combustion engines into the design, due allowance must
be made for humidity as controlled by geography. This is not only from the point of
view of efficiency, but by actual operability. Locations which experience prolonged
fogs are not adapted to internal combustion engines using low-grade fuels, and if
gasoline is not available, the designer will do well to provide a source of dry air for
combustion purposes.
The economic location of the proposed plant may profoundly influence the
design. The economic location is determined by supply of materials, power and
labor on one hand, and markets and competition on the other. A plant located
far from its source of raw materials must have included in its design much more
ample provision for storage of raw materials than when such are close at hand.
The same is true of fuel. When purchased power is used from only local source
of supply, and continuous operation is essential, auxiliary units should be pro¬
vided, if the size of the operation and the cost of a shutdown will justify it. If
more than one source of purchasable power is available provision should be made
for utilization of an auxiliary source in emergencies. If labor is scarce it is
often good designing policy to increase the ratio of machines to men, and perform
all operations on one shift. With a plentiful supply of labor, it may be cheaper
to substitute men for machines, and design a plant to operate two or three
shifts on a lesser number of machines. The important point is that the designer
should know and decide upon this feature in the beginning, and mould his design
accordingly. Markets and competition often effect the design. It is often not
sufficient to produce goods of a given chemical analysis or character. They may,
owing to existing market demands, or criteria developed by competing products,
have to be sent to market in a given physical condition of form, color or consist¬
ency, or a special form of packing. These factors should be well considered and
allowance made for them in the design.
The most important factor in the economic location of a plant as it affects the
design, is the situation with regard to transportation. Here, the scale of the opera¬
tion is the governing factor. It is too often assumed that a location affording water
transportation is preferable to one affording rail shipments, and in many cases the
conclusion has been an expensive one. If the raw materials to be brought in or the
finished products to be shipped out are handled in quantities equal to cargo shipments,
then facilities for water transportation become of paramount importance, for not
only are cargo shipments of raw materials and finished products certain to be at a
lower rate than car shipments by rail, but the added advantage of bringing in fuel
in cargo lots can be enjoyed. On the other hand, for enterprises where the incoming
and outgoing shipments are in less than cargo lots, it may very well be that rail
shipments even at considerably higher cost are much to be preferred. A small lot of
say, 15 to 30 tons of material can be loaded into a car in bulk, and enjoy all the advan¬
tages of bulk loading and unloading as well as escape the cost of containers. Even
when containers must be used, the material itself usually sets the limitations on the
nature and cost of these.
As opposed to this, less than cargo shipments by water offer conditions which
often prove troublesome; le., the necessity for a regular service of bottoms engaged
in the general carrier trade, the usually more strict and expensive regulations as to con-
tamers, the general Impossibility of making less than cargo shipments in bulk, the
infrequency of sailings as compared with railroads, and the necessity for transhipment
at destination. The tonnage involved in the proposed operations will usually decide
which location, rail or water, is the more advantageous. The ideal location will
afford access to both, and in such case the designer should always make his layout
and orient his design so that either means of transportation can be used, even if he
never intends to use but one. This ability to divert, on short notice and at little
expense, the flow of shipments from rail to water or vice versa, will be the most potent
of all factors in establishing and maintaining favorable commodity freight rates on
the raw materials or finished products under consideration.
The topographic location of a plant is perhaps the most important of all loca¬
tion factors in the design. Upon the skillful utilization of topographical features
will depend the construction cost and the efficiency of the system of transporta¬
tion of material into, through and away from the plant. The reader is here
referred to the paragraph on “Type Process.” If the plant to be designed has
transportation for its type process, then the utilization of topographical features
toward economical transportation becomes the chief duty of the designer, and
all other considerations must be subordinated to this one. If the type process
represented is one of the other four enumerated, the question of transportation
becomes of relative importance, the degree of which must be decided in each
individual case. Here we come upon the familiar question as to the relative
desirability of the “hillside” type of layout as compared to the “broadcast”
design. As regards plants where the type process is transportation, there can be
little doubt that the “hillside” type of design is not only the more desirable but
is vital to success if competition in production costs has to be met.1
As stated above, in other type processes, the question must be answered in each
individual case, the answer depending solely upon the degree of importance of trans¬
portation in the design. In every case, it is well to remember that the cheapest and
easiest way to move any substance from one place to another is to allow it to roll
downhill. Likewise, the cheapest and most convenient way to get large quantities
of any solid material up hill is to haul them up in standard railway equipment. In
other words, it may be assumed that the “hillside” type of design (whether actually
located on a hillside, or having hillside conditions in a measure counterfeited by alti¬
tude of structure) is in general the economical type for chemical-engineering plants,
and that the fairly numerous exceptions to this rule are the results of special features
of the process or location involved which more than offset the transportational
disadvantages of the “ broadcast” type.
Type of Construction.—From the point of view of the chemical-plant designer,

there are but two main types of construction, wooden frame and steel frame.
These may be subdivided as follows:
Wooden Frames Steel Frames

(a) With wooden walls and roof. (e) With wooden walls and metal roof.
(b) With wooden walls and metal roof. (/) With masonry walls and metal roof.
(c) With metal walls and metal roof. (g) With concrete walls and metal roof,
(id) With masonry walls and metal roof.
1 This type of design is especially well exemplified by many of the Western lead- and copper-smelting
plants.
The type to be adopted depends upon a great many factors, chief of which are
the type process, the nature of the investment and the location.
Examples of each type to be found in good practice are given below:

(a) Canning factories, drying and curing sheds, mining mill buildings, and numer¬
ous others which serve chiefly as shelters for machinery or operations in comparatively
removed locations.
(b) Same as class (a) but in the warmer climates where the usual objections to
metal roofs do not obtain.
(c) Smaller foundries and metallurgical mills where overhead traveling cranes
are not used; Naval stores and textile warehouses, usually confined to one story
structures.
(d) This class represents a compromise between the hazardous and the fire¬
proof types, and is often termed “ slow-burning ” construction. It probably more
generally meets the demands of the average designer than does any other type. It
has much to recommend it, and numerous examples of almost every type of industry
will be found utilizing this type of construction. It is not, however, recommended for
plants over two stories in height.
(e) Many foundries, alloy and heavy chemical plants employ this type. The steel
frame gives fire protection and supporting strength for cranes and other heavy fixtures
which cannot be directly supported from the floor, and at the same time affords no
sweating of walls—often a most objectionable feature. This type is slowly gaining in
favor and has much to recommend it for single-story work.
(f and g) These types are the ideal structures for long time investments, and for
type processes requiring more than two-story elevation.
Influence of Separate Associated Units.—Careful consideration of this factor

is necessary throughout the design if the plant be composed of more than one
building or if it is one of a number of plants engaged in working up the same
material. In the former case, if the type process is transportation, the location
of buildings should conform as nearly as possible to the physical flow sheet. This
is not always the best practice in other types however, where control or condi¬
tioning may be the guide to location regardless of transportation. While the
uninterrupted flow of material from process to process is always the ideal condi¬
tion to be obtained, it is often good strategy to arrest or even reverse the direction
of flow under certain conditions where transportation is of secondary importance.
This fact makes necessary a most careful survey of the influence of all units in a
plant upon each other before definite location of each is decided. For example:
in a process which requires natural sunlight for control of the product by the work¬
man, separate units must be located so that in flowing from one to the other, the
material will pass, or lie to the north or west of the workman, regardless of
transportation difficulties.
Again, if the material requires retreatment at any stage, it must not be allowed
to get too far away from the unit constituting the treating equipment in question.
In all cases where re treatment is practiced, intermediate storage units should be
provided to take up the surges in the flow always consequent on interruption of the
process from any cause. The more frequent or lengthy these interruptions are likely
to be, the larger should be these intermediate units. While they condemn perpetually
a certain additional amount of capital to “goods in process,” this is often more than
justified by continuity of output. The designer should beware of a flow sheet too
tightly and rigidly knit together. In such a case interruption at one point causes a
surge and loss of control throughout the entire process. These surges result always
in loss of time and usually in loss of material. Proper attention must be given to the
location of intermediate storage depots between the more important steps of the flow
to serve as elastic units to absorb the surges.
Control.—Provision for successful control is the keynote of success in design¬

ing. Systems of control superimposed on already existing designs are never
entirely successful. The original design must embody the elements which make
efficient control not only possible, but not too expensive. Without proper pro¬
vision for control, proper departmentalization is impossible, and cost accounting
becomes involved and inaccurate. Modern accounting is merely a comprehen¬
sive system of control. With proper designing in the first place it becomes a
real tool which the management may keep constantly in use to lower costs and
' raise production. Otherwise, no available system of accounting is anything more
than a record.
Administrative control should be provided for in the design. This includes
the proper allowance for the activities of all administrative labor from the chief
executive down to the foremen. Due attention must be paid to the actual diffi¬
culties opposed by time and space to close coordination. Other things being
equal, the best design is that which places the administrative labor in closest
proximity to the processes being supervised. Despite the numerous instances
of widely separated plant and office, it can be shown that where one plant only is
under consideration, the most efficient and least expensive administrative control
is possible when the entire office of a plant is located at the plant itself. Further
than this, the office should be centrally located at the plant as regards to flow
sheet. Only where one office is called on to control several widely separated
plants, is it justifiable to locate it at a considerable distance, and this always
entails expense. The most modern tendency is to swing away from the com¬
bination with a “downtown” or city office, to that design which places the entire
administrative control at the plant itself.
Control of labor should be most carefully considered in the layout. The ideal
control will place the laborer at his task immediately after his time is chargeable
to the company, or in other words, as soon as he punches the clock, and will
remove his name from the payroll immediately upon his laying down of his task.
This is never entirely achieved in actual practice, but a good design will provide
for its practical realization. The physical control of labor is best obtained by
proper departmentalization, and then by adequate fences and barriers, with
automatic devices for recording the passage of laborers from one department to
another. A man-tight fence around the entire plant is most desirable. Ingress
and egress of all labor should be restricted so far as possible to one gate. The
ideal design should preclude the admission of a laborer to the precincts until
after he has punched the clock, and should allow of a direct, supervised route
from the clock to the job, and return, and the laborer should be removed from the
precincts immediately his time ceases to be chargeable to the company. If
company dining rooms, rest rooms, etc., are included in the design, they should be
beyond the time clock and outside the precincts. Pay windows, commissaries,
wash rooms and employment offices should be similarly located. The location
of toilets necessarily within departments should be most carefully considered.

They should be close to the job and within sight of the foremen. The amount of
labor lost in toilets is usually 1 to 2 per cent and may greatly exceed this figure
unless proper supervision is possible. Proper provision for drinking water is an
economy in design. Men work better and less time is lost in going to the tap.
Control of time in the design is the next most important feature after control
of labor. It is next to impossible to make this control too elaborate—for it
embodies not only time of labor but time of process and time of capital. Time
control is obtained by two general methods, both of which are necessary. These
are control by record and control by observation. Let the designer remember
that no device, human or otherwise, has ever improved on the clock for telling
time. Estimated time in process work should be forbidden just as much as it is
in labor time. Two complete sets of clocks are advisable. One of these is for
automatic record and the other for observation. The former are for the use of
all labor according to circumstances. The latter are for foremen only. Time
control should center in the plant manager’s office. The use of master clocks
with electric standardization is recommended. The proper location of both
recording and observation clocks is most important. Observation clocks should
be placed in every department, and should be large enough and so located as to be
visible from every part of the department. All should be synchronized and
controlled by a master clock. Wherever human time-keepers are employed,
provision for the control of these automatic clocks should be made. It is recom¬
mended that the designer employ clocks freely. It is difficult to err in this direc¬
tion, if the clocks are standardized and under electric control of a master clock.
The continuous proper speed of flow of material is absolutely dependent on clock-
controlled labor and machines.
Control of material is achieved by control of volume and control of weight.
Ultimately the former is always resolved into the latter. The proper types of
volume- and weight-control devices is a subject belonging properly to detailed
construction, to which the reader is referred. At this point only the general
subject of control is considered. The designer’s guiding principle should be
the fact that in the vast majority of processes (as a practical matter due to limi¬
tations of the art) the materials handled change both in volume and weight as
they proceed through the flow sheet. The amount of material control desired
will depend on the system of cost accounting selected. The simplest case is where
only weights of incoming and outgoing raw material and finished product are
required. From this case to the most elaborate control we have an infinite grada¬
tion of complexity. The designer, with his accounting system in. mind must de¬
cide to what extent it is necessary to carry this control. Here, unlike the case
of time control, it is easily possible to overdo. Generally speaking, it may. be
said that the minimum requirement is sufficient control to permit an accurate
knowledge of raw materials (including fuel) on hand, goods in process, and
finished product. Practically this entails car and platform scales for incoming
and outgoing material, and platform or hopper and automatic scales for goods in
process. The designer’s task here is comparatively simple, though important,
and consists chiefly in, first, determining the number of weighings or measurings
demanded by the control in view, and then the proper location of the weighing or
measuring devices, with due regard to economy of time and labor in getting the
Control of temperature ‘ applies to both plant and process, and is secured

along two general and somewhat different lines. Broadly speaking, control of
temperature in the plant is a control of atmospheric temperature within the plant
buildings. Control of temperature in the process is the control of temperature
within the material itself. Not infrequently the two conflict rather sharply,
and the amount of control necessary and the means to be employed in both cases
should be scrutinized with care. Control of atmospheric temperature within
the plant is best secured by central heating units, when the temperature is to be
maintained above that of the outside air, and by proper ventilating systems when
the temperature within is to be kept down to that of the outside air. No really
practical method of atmospheric-temperature control has yet been devised which
will permit of any considerable, continuous lowering of the atmospheric tempera¬
ture within the plant below that outside. This is because of the inevitable
throwing out of moisture by cooled air, and the consequent dampness of all
things in the supercooled area. But by the proper utilization of artificial and
natural draughts and shading, a degree of comfort for the workmen can be
achieved which is eminently worth while. As this comfort is dependent largely
on increased evaporation of moisture from the skin and not on decreased tempera¬
ture, the designer will do well to content himself with providing for perfect venti¬
lation and then turn his attention to possible moisture control.
The maintenance of higher temperatures within the plant than obtain without is
subject to more positive treatment. Of the two general systems of plant heating;
that of bringing the heat directly to the air of the plant, and that of first introducing
the heat into the air and then bringing the warmed air into the plant; the choice
is largely governed by the type of construction under consideration. Buildings of
wood, and particularly buildings of shed type and all lightly or loosely built structures
require direct heating from radiators, salamanders, or hot air blasts. It is difficult and
expensive to attempt systematic atmospheric conditioning in such structures, and
the designer should use care not to make his temperature control more elaborate than
his type of construction or his climatic conditions will justify. Where buildings are
of steel and concrete and tightly constructed, there can be little doubt of the superior
desirability and efficiency of heating and controlling temperature by properly con¬
ditioned air. In certain climates, a careful weighing of this fact should be allowed
to influence the choice of type of construction within reason, especially if the labor
element in the process is large. The optimum atmospheric temperature at which
human labor should be carried on is that which will throw the least work on the
temperature control system of the human organism itself. Obviously this will
vary for different forms of labor and for different atmospheric conditions.
It is probable that the human laborer is most efficient when there exists the best
balance between the work required of the skin, the kidneys and the lungs, and this
should be the balance sought in control of plant atmospheric temperatures. When
his type of buildings permits the use of conditioned air, the designer can do much to
increase the efficiency of his design by careful consideration of these facts. When
the type of construction necessitates direct heating, he will do well to confine his
efforts to provision for ample heat and rely on properly conducted outside atmosphere
to achieve as much control as is possible under the circumstances. In all cases,
recording thermometers are called for to the extent that the record can serve as a
basis for control. Seldom is the control so slight as to make the use of recording
thermometers unjustifiable.
The control of process temperatures is a local problem and belongs more to
the constructor than the designer. In general, it may be said that recording
thermometers are almost always to be preferred to merely indicating instruments.
Of the three types (neglecting the fusion thermometers or cones used in the
ceramic industries), expansion, thermo-electric and radiation, the first is preferable
wherever it may be employed. Thermometers of this type are more accurate
and more positive and will outlast the other types. At temperatures or distances
where the expansion thermometers cannot be employed, the thermo-electric
types are recommended. Their chief advantage is the feature of recording at
relatively great distances. Radiation thermometers are useful at temperatures
where other types cannot be employed, but the field in which they can be accu¬
rately employed is limited. They have also the drawbacks of the personal equa¬
tion, and the lack of a recording feature.
Control of moisture like control of temperature, is sought both for general
plant atmosphere and for atmosphere of the process chamber. In the former
case, much that has been said relative to atmospheric-temperature control applies
here also. Only in tight, well constructed buildings should this be attempted.
When the type of construction will permit, however, it is usually good design
to provide for the most careful control of this factor. The only feasible method
of drying air on a large scale is to chill it and cause a precipitation of a part of the
contained moisture. Moist air is produced by the direct addition of moisture.
As it is impractical to carry out either the drying or the moistening process
within the rooms occupied by workmen or processes, the usual method is to main¬
tain a balance between moist and dry air by introducing to the plant air already
conditioned both as to temperature and moisture and then providing for a
continuous flow of this atmosphere through the plant. Control is thus secured,
and is indicated by psychrometers. Satisfactory recording psychrometers are
not available, and control must be maintained by indicator readings, taken as
often as necessary. Here, as in the case of temperature, the optimum humidity
is that under which the temperature control of the human body will have the
least work to do. This is also obviously dependent upon the nature of the labor,
and no hard and fast rule can be laid down. The fundamental fact to be kept in
mind by the designer here is the same as in the case of temperature control,
namely, that the normal temperature of the blood is 98.6°F., and that the normal
temperature of the outermost skin surface is considerably below this point. The
difference represents the working margin of safety allowed the organism to con¬
trol the temperature of the blood and exercise the excretory functions. Advan¬
tage may be taken of this within certain narrow limits, and increased moisture
substituted for increased temperature and vice versa—but it should be realized
that optimum temperature is a more or less definite point and that maintainence
of comfort by the body at temperatures away from this point by the aid of
increased or decreased moisture is only achieved by work of the organism itself
through its temperature-control devices. Hence maintainence of comfort need
not necessarily mean maintainence of healthy efficiency.
The control of moisture in the processes themselves is achieved in the same way
as in general plant atmosphere. There is one exception to this, and that is where
moisture is continually contributed to the atmosphere of the process chamber by the
material itself, and where the necessities of the process make it impossible to remove
this by chilling the air or by mechanical removal of the atmosphere itself, if it be other
than air. In such cases, direct drying by lime or other moisture absorbing substances
is called for.
Control of light includes the control of both artificial and natural light. The
control of natural light is brought about by proper coloring of walls and ceilings,
and by properly sized and shaded openings. It is better for the designer to
exercise his talent on these points in the first place than to be obliged to make up
for deficiencies in design by an unnecessarily elaborate artificial lighting system.
Where wall space is not actually required for apparatus, the modern tendency is
more and more toward large windows and many of them. The advent of wire-
filled glass and steel-framed windows has made this possible on an increased scale.
It should be borne in mind, however, that ordinary transparent glass is a poor
reflector, and that rooms with a great window surface of plain glass are correspond¬
ingly difficult to light from within by artificial means. The use of a light frosted
glass serves as a compromise between day and night requirements.
Walls, ceilings and, wherever possible, fixtures should be painted white. It
is safe to assume that provided it is properly diffused, there is no danger of
getting too much light. Wherever the use of white paint is impractical, the next
lightest color should be used. Black is a very inefficient color from a lighting stand¬
point, and almost any other color is preferable. For artificial lighting, the question
as to which is the better, monochromatic or polychromatic light has not been com¬
pletely answered, but the present concensus of opinion leans decidedly toward the
use of ordinary “white light,” either produced by the tungsten filament or the incan¬
descent gas mantles. For special purposes or highly specific processes, the light of
the mercury-vapor arc may prove more satisfactory. Under our modern conditions,
the designer in arranging light control has only diffusion and reflection with which
to work. Generation of light itself has become too much standardized to permit
of any special adaptations for specific designs. Diffusion is merely compound
reflection, so the problem resolves itself into two very simple elements; the selection
of angles, surfaces and distances in the form elements of the design itself and the
selection of materials with the highest practical albedo. There can be little doubt
that the semi-indirect system of lighting is usually preferable. Reflectors should
be translucent, in order evenly to diffuse the light and prevent shadow zones. It is a
mistake to illuminate only the plane of work, except in isolated cases. Individual
illumination of special work is a function of construction, not of design.
Distribution of power is properly regarded as one of the major phases of

designing and usually comes in for its full share of attention from the designer.
Unfortunately the true theory of power distribution is often imperfectly under¬
stood, and elaborate plans do not always bring efficiency. Briefly stated; it is
cheaper and often more practical to transmit power than it is to transfer material.
The tendency of the mechanical engineer, with his mind centered on the most
efficient and economical generation, transmission and application of power, is
too often to subordinate to this the even more important flow sheet or general
process. The distribution of power like all other factors in the design must be
guided primarily by the type process—after that by the layout limitations of
the flow sheet. It is a mistake to attempt to warp either of these to fit any
arbitrary scheme of power distribution, regardless of how efficient.
From the above, it necessarily follows that the ideal system of power distri¬
bution is the most elastic, and there can be no question that this is at present the
electrical system. It does not necessarily follow however, that the electrical system
is in all cases the most expedient, the most economical, or the most efficient. The
nature of the investment, the nature of the industry, the type process and the location
all have to be taken into account. Hence, it becomes necessary for the designer to
consider the several possible systems of power distribution. These are: (1) the
generation of electrical power at a central prime mover, and distribution by electrical
transmission to various motor drives. (2) The generation of mechanical power at a
central prime mover and distribution mechanically by line shafts. (3) Division
and distribution of the prime movers themselves.
Where operations are on a considerable scale covering a considerable period
of time and where the requirements for the immediate future are definitely known,
there can be little doubt that the first system is preferable. Its advantages are the
practically unlimited elasticity of distribution over any operating area, economy of
operation, high efficiency of transmission and low depreciation under normal con¬
ditions. Its disadvantages are high first cost, absolute dependence of the entire
plant upon one unit or group of units, impracticability of gradually increasing the
available power supply to provide for gradual expansion, and the often serious effects
of surges when a considerable proportion of the total load is subject to violent fluc¬
tuations, such as those attendant on heavy hauling and excavating work.
Where fuel is cheap and plant area not too great, the second system may very
well become the most desirable. This is especially the case in certain types of indus¬
try or certain special locations. The chief advantages of this system are the ease
and positiveness of operation, the low first cost of installation when compared with
the first system, and the ability to absorb heavy surges. Its disadvantages are its
greater cost of operation, its decreased efficiency of transmission and greater rigidity
as regards distribution.
The third system represents a compromise class, when the first system is not
justified by the nature of the investment and the location, and the second is made
impossible or undesirable by the area covered by the plant. Such cases are not
uncommon, but they can never show the low production costs of the first or second
systems. The fundamental feature which usually justifies the adoption of this
third system is too small a scale of projected operation. Here the designer is thrown
back on the nature of the investment itself and the fact that this will not permit of the
adoption of the first and second systems may often (although by no means necessarily)
serve as a danger signal to indicate the fundamental weakness of the enterprise
itself—due to the impracticability of the proposed scale of operations.
Waste and By-product Disposal.—All material which is not included in the

regular product of the plant, and which is chemically or physically altered as a
result of the operation of plant processes is either by-product or waste, depending
upon the presence or absence of earning power. If the material thus altered is
subject to profitable utilization or disposal, it is by-product;if not, it is waste. In
either case due provision in the design must be made for disposal, whether profit¬
able or otherwise. If by-product, this provision must take the form of storage,
and transportation to market exactly as in the case of the main product. If waste,
provision must be made for the cheapest possible disposal, otherwise the item of
transportation will add direct cost to this indirect cost item. If the waste is
voluminous, its disposal may often prove as important as any question to be
decided by the designer. Solid waste which cannot be burned or otherwise con¬
sumed, may be either stored on the premises or removed beyond the plant pre¬
cincts. In either case the problem presented is simply one of handling and trans¬
porting solid material. On the other hand, if the waste be liquid or gaseous, it
97 G CHEMICAL ENGINEERING
should be borne in mind that ultimately it is impossible to store it on the premises.

It must eventually trespass on the property of others, both public and private.
If the waste is of such nature and volume as to permit of this trespass without the
objection of the property holders or the general public, no problem is presented.
But if such trespass is objectionable or is likely to become so, then the designer
has his work cut out for him. Here, his earlier study and analysis of the nature
of the investment and the location will stand him in good stead. At best such
problems are difficult. When certain legal restrictions exist, they are sometimes
prohibitive for the industry itself.
Sanitation.—Broadly speaking the design must provide for sanitation from
three points of view: (1) The effect of the plant on the public health; (2) the effect
of the plant on the health of employees; and (3) the effect of employees upon one
another. The first is less personal, less understood, and less likely to have atten¬
tion demanded for it. For this reason it is usually most neglected by designers.
It is, however, from a purely business point of view, well worthy of most serious
consideration. A fundamental and not easily removed objection in the relations
between any plant and the public health is as dangerous to the safety of the
investment as are equally subtle and often ignored flaws in legal title to real
estate or processes. They only need dragging into the public eye to condemn the
particular investment with which they are associated. Increase of population
is rendering the isolated plant more and more the exception, and the tendency of
modern medicine is more and more to seek in great industrial enterprises the roots
of menaces to the public health. Whether such menaces arise from the nature of
the processes carried on, or spring from an unsanitary relation between the
mass of employees and the surrounding public, it is much better to avoid or
provide against these in the original design.
The effect of the plant processes on the health of employees is a subject which
modern practice has reduced to fairly uniform procedure. Confinement and isolation
of the objectionable process as far as possible on the one hand, and protection of the
workman on the other constitute more or less standard practice. The endeavor should
be, of course, to prevent entrance to the organism of the workman of harmful matter;
the maintenance of normal atmosphere, light and temperature conditions and the
avoidance of straining nerve and muscle beyond the elastic limit. The last, no less
than the first two are secured by the design itself, for the policy which achieves the
last is only possible in an adequately designed plant.
The effect of employees on one another as regards the health of each individual
is largely a question of control and handling of excreta; gaseous, liquid and solid,
and should be handled in the same way and with the same care as is the disposal of
other waste. This practice has been reduced to a very exact science and there is
absolutely no excuse for failure of the design in this respect.
Hazards and insurance bear a rather peculiar relation to each other. Both
represent items of cost, one potential and the other actual. The cost of insurance
is directly proportional to the degree of hazard. Aside from the absolute inherent
hazards in the particular industry and type process under consideration and the
hazard arising from the intellectual limitations of employees, the designer can
absolutely control the degree of hazard and hence the cost of insurance. Hazards
may be divided into two main classes: human and natural. Under the former
heading can be classed fire, explosion liability, burglary, mob violence. and war.
Under the latter come storm and earthquake. It is instructive to consider that
probably 95 per cent of all industrial losses come from human hazard and only
5 per cent from natural hazard. Yet this fairly represents the ratio of control of
modern engineering over animate and inanimate nature.
It becomes evident that hazard depends upon control, and chiefly upon control
of labor. There are two ways of attacking this problem, and the designer will do
well to use both to the limit of his ability. The first is to remove as far as possible
the human factor from the hazard by making protection automatic. Fireproof con¬
struction, sprinkler systems, self-acting alarms, safety valves, the housing of danger¬
ous parts, etc., are typical examples of this method, and represent a constantly
increasing tendency in modern engineering. This tendency is founded on the fairly
well understood fact that the last factor to be brought under control of the plant
design and the least reliable in its processes is the mind of the plant employe. The
second method is to surround the activities of the employee with prohibitory equip¬
ment and regulations. Watchmen’s pull boxes, no-smoking zones, etc., are examples
of this method. The reader is again referred to the paragraph on control. Control
is the essence of hazard prevention, and that design wherein control of labor and
material is perfect will automatically exhibit the least degree of all hazards.
Design is Strategy; Construction is Tactics.—The engineer will do well to keep
this difference clearly in mind. The actual production of design drawings from
the first rough sketches to the finished blue prints requires the most careful
method and systematic routine to secure results which are satisfactory. The
difference between ordinary structural design and machine design on the one
hand and plant design on the other is simply this: Ordinary structural design is
in the last analysis a question of shelter from the elements; ordinary machine
design is a question of pure mechanics, but plant design is a problem in coordina¬
tion. There must be no loose ends, no blind alleys. Before the first ground is
broken, the design must be complete to the last detailed drawing; the ground
should further be accurately surveyed and staked to avoid errors in location.
Structural drawings should be on a uniform scale of % in. to the foot.
No tracing should be allowed to be made until the original drawing has been
checked by the designer and one other competent engineer. All calculations involved
should be retained on original sheets, and these should be checked by the designer and
one other competent engineer. The tracing may then be made, and must be checked
and signed by the designer and the tracer. All tracings should bear a stamp carrying
signatures of the designer and one other competent engineer under a statement that
each has personally checked all calculations and drawing involved in that sheet.
Drawings should be divided into three classes, and given a sheet and serial number
accordingly—layout, detail and assembly. The actual work of designing should
proceed in this order. The layout must be absolutely complete before detail and
assembly sheets are begun. The next step after completing the layout is the making
of the survey, and establishing of levels. Fills and cuts can then be calculated.
The designer then locates his means of transporting structural material—usually a
spur of standard-gage railway track—making this conform as far as possible with one
of the tracks to be permanently used for the plant. The layout should contain in
properly proportioned outline every space occupying feature of the plant and process.
Having located his construction track or tracks, the next step for the designer is
to locate on his layout unloading, storage and handling depots for all materials of
construction and mechanical units which will enter into the design. Especial atten¬
tion should be given to allotment of ample room for lumber and steel stacks, with jib
02
cranes or other devices for economical handling. Warehouse room must be provided
and accurately located for cement, plaster, and all materials which cannot be stored
in the open. Temporary fire protection, watchman’s call-box systems and (if deemed
necessary) fencing, should all be provided for in the layout at this time, and not later.
The economy of plant construction is achieved or lost in the drafting room, and is
inseparable from the work of the designer.
One additional point of which every design should take account is the possible
need for enlargement. This can be accommodated in the design by one of two meth¬
ods; the unit group system and the individual unit system. The latter system pro¬
vides for the addition of additional individual units to any given battery, by the
necessary spacial arrangement of the original layout of that group or battery. The
former allows of expansion by alloting proper spare ground space in the original lay¬
out for the construction of entire additional batteries or groups.
It is not to be expected that that design which fails to take in and allow for every
detail of the process of construction itself will provide more perfectly for the process
of operation. One is a good index of the other. The designer should remember
that charges against his process begin to accumulate with the very first time check of the
first man employed in construction, no matter in what preliminary capacity, and there
is no possible way in the future of avoiding or unloading these charges. They are
more permanent than the plant itself, and will always stand as offsetting entries
against the efficiency of the design itself, for they represent truly the load incurred
by the design and the load it must carry throughout its existence. It is not an over¬
statement to say that by creating the detail of his design to serve the economy of
construction as well as the economy of operation, the designer may often reduce by
10 to 20 per cent the capital sunk forever in that most unwelcome of all entries,
“construction labor.”
SECTION XXXI
METHODS OF FINANCING
By Donald M. Liddell1
Introduction.—While there may be some question as to the propriety of this

section in a chemical engineering handbook, the courses in college economics many
times absolutely ignore the various types of security by which enterprises are finan¬
ced, and since after all, the object of most engineering work is to give a return on
these securities, they seem deserving of some notice.
While a privately owned company is in many respects an excellent thing, a success¬
ful enterprise usually outgrows one-man ownership, while there is always the chance
that a failure will involve the owner’s entire fortune. Therefore, both to limit indi¬
vidual liability, and to facilitate raising money, enterprises are usually incorporated.
Those who take the chances of the business, the partners, are the stockholders, those
who lend money at fixed rates (creditors) are the note and bond holders.
In discussing the various forms of corporate issues, illustrations have been given
of many securities falling under the various classes, so that if further information is
desired concerning a particular form of indenture, easy reference may be made to
at least one illustrative type.
Common Stocks.—Theoretically the common stockholder is the ultimate

owner of the business and (except in rare cases, such as the former American
Brake Shoe & Foundry Co.’s organization) the profits of the business after pay¬
ment of interest and guaranteed dividends accrue to him.
For purposes of control by an unchanged management over a period of several

years the holders of the common stock often deposit their stock certificates in a so-
called voting trust and receive instead voting-trust certificates, the trustees of the trust
then having all the voting powers. There has been much debate as to whether voting
trusts are consistent with public welfare, and most states now limit their life to 5
years.
The value of each share of stock may be assigned a definite value, its par value,
the number of shares times this par value being the original theoretical value of the
portion of the business belonging to the common stockholders. Or the shares may
be of “no par value,” each share merely representing a fraction of the ownership of
the business, which is practically what all common stocks are. However, it often
appears that a certain sale value is inherent in the phrase “par value, $100 per share”
even though the entire $100 consists of capitalized aspirations, and the “no par value”
idea does not make much headway.
Occasionally the common stock is split into “founders’ shares” and “ordinary
shares,” the founders’ shares, usually but few in numbed, taking a certain fixed per¬
centage of the profits, or a fixed percentage of the profits2 over a certain amount,
giving them a return entirely incommensurate with that on the “ordinary shares,”
and often out of all proportion to any risk or responsibility assumed by the founders.
As an elementary principle, both of sound business and also of law, dividends may
1Weld & Liddell, Engineers and Economists, New York, N. Y.
2 That is, of the profits accruing to the common stock.
979
not be paid by a company after such payment impairs the value of the com mon stock
but may be paid from a previously earned surplus when there are no current earnings.
Nevertheless, instances are not wanting where dividends have been paid from the
proceeds of stock sales.
Preferred Stocks.—The name of course implies a certain preference over the

common stock. This usually consists in a prior claim to the earnings until a
given dividend is paid before anything may be paid on the common. The pref¬
erence may end here, as it does with the 7 per cent preferred of the American
Sugar Refining Co., or it may also include a prior claim on assets in event of
dissolution, up to a given amount, usually par and accrued dividends, before the
common stock receives anything.
But the preference on assets is limited only by the terms of the indenture, for the
American Zinc, Lead & Smelting Co.’s preferred has a stated par of $25, but must
be paid for at $100 if called. However, this contingency appears most remote.
Preferred stocks may be either cumulative or non-cumulative. In the first case,
if the dividend is not earned and cannot be paid from a previously earned surplus it
must be paid from later earnings before the common stock can receive anything.
Examples of such payments were those of Republic Iron & Steel and Crucible Steel,
which made up previous deficits on their preferred stock during the War. On the
other hand a preferred stock may be non-cumulative and entitled only to fixed divi¬
dends that may properly be paid from earnings or surplus. In this case, if for any
reason a dividend is not paid during any period, the stockholder receives nothing on
this account later. Such preferred stocks are those of the Pressed Steel Car Co. and
the Atlantic, Gulf & West Indies S. S. Co.
A preferred stock may be “participating.” In this case, after payment of the
guaranteed dividend on the preferred, and a stated dividend on the common, the
preferred shares with the common in the further earnings.
Or the preferred stock may be entitled to all the earnings after a given amount
has been paid on the common. For instance, the old preferred stock of the American
Brake Shoe & Foundry Co. was entitled to 7 per cent before the common received any¬
thing, then the common received the benefit of the earnings until 7 per cent had been
distributed on it, after which the preferred received the benefit of all the earnings.
But merely calling a stock “preferred” where no actual preference exists, does not
make it so. Great Northern “preferred” is the only class of stock outstanding,
and the name represents only the personal “preference” of the late J. J. Hill.
Two or more classes of preferred stock may be outstanding simultaneously. In
this case their -rights may be very simply related, as in the case of the Superior Steel
Corporation, where the first preferred has prior claim over the second preferred both
as to earnings applicable to dividends and as to assets in case of liquidation; the second
preferred in turn having similar claims over the common, or the respective rights
may be exceedingly complicated, as in the case of the Bethlehem Steel Corporation,
where it appears that the 8 per cent cumulative preferred has prior claim on earnings,
while the 7 per cent non-cumulative preferred would appear to be a lien prior to the
8 per cent in case of liquidation and distribution of assets.*
The case may be further complicated by the assumption of the liability for the
dividends of the preferred stocks of other companies: Thus it appears that the
preferred A and B stocks of the American Smelters Securities Co. all of whose common
stock is owned by the American Smelting & Refining Co. are prior liens over the
preferred stock of the latter corporation. The preferred A of the former appears to
have first preference as to payment of dividends, but the preferred R, being guaran-
* This refers to the stocks of the former unmerged company.
METHODS OF FINANCING 981
teed by indorsement, probably has preference in regard to assets. Common stocks of

controlled corporations where a given return is assured rank for all practical purposes
with preferred stocks.
In order to facilitate the raising of money, other privileges than that of priority
regarding dividends and assets are occasionally given to the purchasers of preferred
stock. Thus, the preferred stock may be made convertible into common stock. In
this way, if the company attains marked success, the preferred stockholders can
exchange their stock for common and share in the prosperity of the common stock¬
holders. The convertibility may be at any agreed ratio.
In the case of the Superior Steel Corporation’s preferred stocks, already referred
to, the preferred is exchangable for common share for share. In the case of the General
Asphalt Co. one share of preferred is exchangable for 1 shares of common. Further
protection may be thrown about a preferred stock in order to increase its salability
by stipulating that no mortgages (or none except purchase-money mortgages) may
be created without the written consent of some determined percentage of the preferred
stockholders.
Preferred stocks may or may not be retirable, and in the former case may be
retirable in part, either by lot or by purchase by sinking fund. The U. S. Steel
Corporations preferred stock may not be retired. That of the American Zinc, Lead
& Smelting Co. may be retired only as a whole. The preferred A of the American
Smelting Securities Co. is retirable by lot, a part each year. In these cases a sinking
fund is set up by allotting a certain percentage of the earnings or a certain set amount
from the earnings, or a certain set amount based on the preferred stock outstanding,
and the stock either bought in by lot, or purchased in open market, or acquired by
tender. In any case there is usually a premium given the holder of the stock if the
stock is forcibly retired either in part by lot, or as a whole. Similar sinking funds are
used to retire bonds. A case where all of an issue of bonds must be retired at 110
of face is that of the U. S. Steel Corporation’s 10 to 60 year second-mortgage sinking-
fund bonds of 1963.
Preferred stocks may have equal voting power with common, no voting power, or
voting power only in case of default either in dividends or sinking funds. In the case
of preferred stock without voting power, the rights of the stockholder differ but little
from those of a holder of unsecured debentures, although the latter could probably
force payment of interest if earned in a court of equity where the preferred stockholder
might not be able to force the payment of dividends, even if earned.
Debenture has no definite legal meaning, except that it is always a debt.1
It may be applied to any promise or security of the company to pay money. It
may be a mere promise to pay, or a covenant under seal to pay, or a mortgage
under the seal of the company.
In the United States, however, it is usually considered an unsecured credit obli¬
gation only and default in interest does not necessarily serve as a basis for legal
proceedings. However, there may be a provision that if any mortgage be placed
on the property after the creation of the debentures, the debentures shall then be
equally secured with the new bonds. Such a provision appeared in the convertible
debentures of the Baltimore & Ohio R. R. Co., due 1933.
Debentures may or may not be convertible. Those of the Alaska Gold Mines
were convertible into common stock &t $30 per share. Debentures are sometimes
spoken of as “notes” or “bonds,” according to whether the maturity is short or long,
but there seems to be no justification for the term bonds in connection with them un¬
less actually secured by mortgage deed of trust, in which case they cease to be
debentures.
1 Cook on “Corporations,” Vol. 1, Sec. 14.
The word “stock” has a peculiar meaning in England, referring to debentures as

well as to share capital. Shares, bonds and debentures are of fixed amounts, but
stock is issued in any amount or multiple of certain sums, usually £1.
Debenture stock is an English term also. It does not mean shares of stock,
but means an English bond,1 an absolute obligation of the company to pay prin¬
cipal and interest at fixed times. It is doubtful whether the debenture stock of
the General Motors Corporation is correctly named, whereas the preferred stock
of the S. S. Kresge Co. comes close to being such an obligation, except that the
dividends must be earned before they can be paid.
Notes may be unsecured (debentures) or secured either by collateral trust
or mortgage. In the last case the only difference between notes and bonds
seems to lie in the names based on popular usage resting on the very indeterminate
matter of short or long maturity, roughly obligations running 10 years or less being
spoken of as notes. On the other hand, the mortgage deed of trust obligationsôf
the Hanna Furnace Co., running only 5 years, are spoken of as bonds.
The security may be actual pledge of real property (mortgage notes), or the
deposit if stocks or bonds of the issuing company or of other companies (collateral-
trust notes), or they may be secured by indorsement of other individuals or corpora¬
tions. In this latter case, an indorsement on each note or bond by the guaranteeing
corporation is a better security than a guarantee merely stated in the general deed of
trust. Bonds may even have two guarantors, as is the case with the 4J^ per cent
bonds of the Chicago, Lake Shore <fe Eastern Ry. (guaranteed by the U. S. Steel Cor¬
poration and the Elgin, Joliet <fc Eastern Ry.).
To make them attractive at the time of issue, conversion privileges may be given
to notes and bonds, just as with stock. This is particularly done with the issues of
mining and oil companies upon the theory that the investor in these issues takes a
greater risk than does the ordinary industrial or railway-bond buyer and some means
of giving him a speculative chance as well as his fixed return is necessary. For this
reason the bonds or notes are usually made convertible into stock at some higher
figure than the market price of the stock at the time of issue, so that the bond holder
does not have so great a speculative chance as the original stockholder and yet has
some chance for appreciation on his securities. Examples of this are the Chile Copper
Co. 7’s, each $1,000 bond convertible into 40 shares of stock at $25 per share; the
Chile Copper Co. 6’s, a later issue, convertible into stock at $35 per share; the General
Asphalt Co. 8’s, convertible into common stock at $100 per share.
The conversion privilege often is terminated before the maturity of the bond, in
order not to give the holder too great a speculative chance. Such were the convertibles
of the Erie R. R.
Income Bonds.—These are a half-way house between preferred stocks and
true mortgages. The interest is payable only if earned, is usually, but not always
cumulative, as in preferred stocks, but if the bonds are in default at maturity
there is (usually) specific property that can be levied on, and the holder is, of
course, legally a creditor and not a partner (stockholder). The income bond
holder has no vote, but neither do some preferred stockholders. The Hudson
& Manhattan Ry. Co7s. adjustment income bond is a good example of this
class of security.
Collateral-trust Bonds.—These are secured by deposit of stocks, notes or
bonds with a trustee, rather than by pledge of real property. A bond may be
both mortgage and collateral trust, as are the International Agricultural 57s of
1932.
1 Cook on "Corporations."
METHODS OF FINANCING 983
Mortgage Bonds.—These are secured by the pledge of real property, and are
supposedly the strongest form of security that can be given. Corporate mort¬
gages are usually made in the form of a mortgage deed of trust, by which the
bonds are secured, necessarily so if many bonds are to be issued. Of course if
there is to be but one bond holder the mortgage may run to said holder, just as
in individual mortgages.
As said above, the mortgage feature may be joined with collateral trust
provisions, and may also be convertible into stock. They may cover only specific
property owned by the debtor corporation at the time of issue, or they may cover
“all property hereafter acquired/’ the latter clause usually being qualified by,
“subject to purchase money mortgages.”
In case a company has originally placed a first mortgage on its property, it may
also place a second mortgage, which is a lien on the property of the company after
and subject to the satisfaction of the first mortgage. A company may also issue so-
called refunding mortgages, which are usually first liens on property acquired later
than and not subject to the previously placed first mortgage, and which provide for
bonds to be issued later to take the place of the first mortgage at maturity.
Commercial Paper.—This is a method of raising money available only to

companies of recognized standing, and consists in placing notes with various banks
through a broker. These notes run usually 3 months, and while it is the cheap¬
est method of all to raise money in good times, it may leave the issuing firm
stranded in bad ones. Usually commercial paper is not considered a good risk
if the liquid assets (cash accounts and bills receivable, etc.) are not twice the
current liabilities, including the paper issued.
Discounting Accounts Receivable.—Recently there has been a great growth
of the so-called “credit companies.” These advance money on the pledge of
certain specific accounts receivable, backed by the general credit of the pledging
firm. It is an expensive method of raising money, but is often of great service.
APPENDIX
METRIC—ENGLISH EQUIVALENTS
Length
ft. = 0.3048060096 m.
in. = 2.540005 cm.
m. = 3.28083 ft. = 39.370000 in.
Area
sq. in. = 6.452 sq. cm.
sq. ft. = 0.09290 sq. m.
sq. yd. = 0.83613 sq. m.
sq. m. = 1.1960 sq. yd. = 10.764 sq. ft.
sq. cm. — 0.15500 sq. in.
Volume
cu. yd. = 0.764559445 cu. m
cu. ft. = 0.028317 cu. m.
cu. in. = 16.3872 c.c.
cu. ft. = 0.0283161.
cu. in. = 16.3867 ml.
. cu. m. = 1.3079 cu. yd.
c.c. = 0.06102 cu. in.
. 1. = 61.025 cu. in. = 0.035315 cu. ft.
Mass
. lb. (Avoirdupois) = 0.4535924277 kg.
L oz. (Avoirdupois) = 28.3495 g.
L oz. (Troy) = 31.10348 g.
L dram (5 =3 3,apoth.) = 3.887935 g.
[ = 2.204622341b. (Avoirdupois) = 2.67923 1b. (Troy)
[ g# = 0.035274 oz. (Avoirdupois) = 0.032151 oz. (Troy)
= 15.4324 grains
Capacity
1 qt. (liquid) = 0.94633307 1.
1 qt. (dry) = 1.1012 1.
11. = 1.05671 qt. (liquid) = 0.9081 qt. (dry) = 33.8147 fl. oz.
1 fl. oz. = 0.0295729 1. = 1.80469 cu. in.
Note.—1 gal. (liquid) = 231.0 cu. in.
1 bu. (dry) = 2,150.42 cu. in.
1 1. = 1,000.027 c.c.
Temperature
°C. = |(°F. - 32)
°F. = | °C. + 32
o
985
986 APPENDIX
Energy
X hp. ~ 0.746 kw. - 33,000 ft.-lb. per minute
1 kw. = 1.341 hp. = 1,000 joules per second
1 ft.-lb. = 1.383 X 107 ergs = 1.383 joules = 0.1383 kg.-m.
1 poundal = 13,825 dynes.
1 gram’s weight == 980 dynes.
1 pound’s weight = 444,518 dynes.
1 hp. year = 6,535 kw.-hr.
1 kw.-yr. = 11,747 hp.-hr.
Many tables of constants will be found in the sections to which the>

particularly pertain. Among those to which particular attention is directed are
Specific Gravity of Solids. Sec. i, p. 108.
Specific gravity, molecular weights, critical temperature and critical pres¬
sure of gases. Sec. iv, p. 144.
Specific heat of gases. Sec. rv, p. 146.
Specific gravity-Baumd-Brix conversion tables. Sec. x, pp. 397, 398.
Standard thermometric points, Sec. xn, p. 410.
Van der Waals’ constants, Sec. rv, p. 190.
Alcohol Proof
Specific Tralles’
Percentage C2H5OH British U. S. ' Fiscal
gravity hydro¬
by weight ^ by volume proof proof proof
60°F. meter
100.00 100 0.79389 +74.8 200 175.35 100

48.98 | 57 0.91999 j proof 114 | 100.00 57
42.25 1 50 0.93426 J -12.4 100 86.17 50
j i
While the basis of U. S. proof is very readily understood, it has no historic justifica
tion. In former times, the smugglers “ proved” their alcoholic spirits by pouring
some over gunpowder and lighting the mass, if the gunpowder burned after the alcoho
burned off the spirit was “over proof.” If the gunpowder remained a sodden mass
the spirit was “under proof.” Unfortunately for those of centigrade mind, the divid
ing point did not come at exactly 50 per cent of alcohol by volume.
SOLUBILITIES OF SOLIDS IN WATER
S = number of grams of substance of the given formula, including water of crys

tallization, if that be given, which when dissolved in 100 g. of wate
makes a saturated solution at the temperature stated.
p = number of grams of anhydrous substance per 100 g. of saturated solution.
w = number of grams of substance as given by formula in 100 g. of water requiret
to form a saturated solution.
APPENDIX 987
COÔO 00 O CO CNjiO
321.3
57.35
23.83
11.40
3 <N • © lO »C iO t- iO 00
0.^«au«
o »o 00 5
^ CO mO
_ ® M +jr
: IKs
:. .ÔO
9.59
7.60
005
m CO
-cojo
• 1—t ^_q
00iO m
5.65
ci » S
D. Van Nostrand Co., New York.

alum, KA1(SC>4)2'I2H2O, S.
ganic and Organic Substances.”
*'QoS
SjH$2-•
>00 ®a ; : :od«;£ :
;cq : «o &
:7g j0' : :
o?&R
'■^ôSS «
'Q~ ûOO
■ .OOOÛOO ®oa .0 so*
mi*g|g§5f- l-'S^pqÔQ .’C«g*sO - >»~.5~’-5q "V£ d ^ £•§£1' - oP
:Jj*fl oW/§ o5"|3,fl3 c|'3-5“fo ffl-Sa §-§ ©'§_©'§•§''
za o o’c s-g 03^ S'1? sla
a s ota-d531 §5a|l» §.j
||f!||||i|l! dill l-S-S'si eS3.s o-g a«| as las al3g!
i Si-all^ 3 ° lal^
-gaJ ■S'S
O o5 ^c3 I IO
jQ I
d &&
do &
O 3
®
i« mo o o o o ofe o
988 APPENDIX
lOOOiONOOOOOO <M O 0ic>0«300003‘*
0*003
>30 X> to O 00 CO 03O
00 tH cd O O oi oi (N 00 00 o’ CO <N
OOOOOUDNOOOONO^ <N O
rtH rHCM<M 05
! . I . I . I
lOOOOOOOON
th co to oo co co oo co
CO r-l rji CO r—( 05 ^
th cocM'-i
I0i>000000)0 ÔOOOOOOm j
eooHcoooô 3
<M ChONOOCCOO OHCOiOOiONCOONOÔOOTtUM^ 0*010 .
J>. g 03 tO •<*< 00 t>" (M C$ ^ CM JC'- rH rH CO CO tO CO CO CO rH\«H ^ )> O <
rM HH Q3 CO j
0*0*
COOOOCOOOONNOOO
I • *T7*7** i . i . I.S (z(zOmzO)<\d
COO»OOOOOCOOO
pWOÔÔO-^COCOCOOOOOMoOiO oo moos
i0Tt*OC0>O(N»0>C003C0t»03iHe01>N.C0(N00 C9 ■*#< ho CO
COtHOO <N OOO (MOCOH03MH NCO fen
rH rH >. iH
**§:
vqj* 33 •■IS .S'? UJ? 88.9 *°Q ^r<Ul£<N •’
*h10Q
i|s|S»"R|s^gKSs^SS-g.;i
illsilllsli lHa
jsjsfj 3
cfl5^ o-r ° J" e-g-g
?pS° aOmsKS
.g'-'-g I S ® ■§ i ■§'§ W) oT <$£&
5P 4 ^ o rt ^
0-Sfc .£
-JJ dfl a’2'
rt-2* oj-S
co-tf flratzi
-rTSlZI - c
2 ® 3 S3 s| § jp 3 a pis Si 21 l«a 2 111 2 SS^l-g®-.!’ §21 6%*
u ta J3 o U § g o-2 o 2t}13! a aa-l01 § a S-tl P s|3 21
§-d •mu p#!ii| linn mi
O S ftd O'-2'a°'cL0 o § S S'o.® ° 55-3 S
-d.-3
C3 fl 00.2 o d5aM5i§§oo9 m.2 mÔ-Q^Q o c o Oi2i«pd,2 03 a® «
3 233 ^ S 23-2.3 o 8 2 31*3
« 93,0,0,0,0 o o-d.2 d c
P™*!
53 SP § SiSfs II
ass ssgg o£
APPENDIX 989
t-HCO»H i>eouo O
O lO O O0^T-( 00
tH CO CO
t~t'- CM CO CO CO
HOJ 00
CO CM lO
0 000 COCOO 1>

CM* 1> Tfi N* f- 00 CM
00 os CQt-100 00
CM CM
•JOCOrH 00 CO CM Is- O
CO OS CONDON
CMt-h-^
W oi? g «
sv wS*3
o® s! “a
O'?! U n
o oOo d.-gW
-®*-W .2^0 oO
goT-d
SO £ a ^ _0 o s Wo
l-s-l! SÎIh^s ?8
20l°. * 3^0* ^ *
“or-?
oSP^Pr^00ag "MWo^S
<dco 2.9
s*SS«s0||^|j'Sos-s
-«SS!oS>2»'s-§**
5 J* f^ncm fn P 3*5 «.3 C ^ ^
r«SS^‘3nS<2^l| “II
.-ssS^SSwsi^l!
qO cT : •gjliiiia^ s-s igg,
:“ea2605 a®e||BB|os|5g|?®
Is -m-SSCB gojo-Q^rfaua “.3 2is
-2ac,«oaj#'da}'SoDa’S«®
ûT'd'd - o
DQWot
S 68£ g'S ‘o'S's-SrS'S'S a-s S ©J3 <0 "S3
E3 =p.2.fi£ g o o'S^ “ijag^gg
3fl^0,Sm H „OJM(Z}CZJCC<!o2H^PqEHfcâ3CQ
rl 03 a »'2n«®“#M«#SS2!32S
*3sJ
990 APPENDIX
International Atomic Weights, 1922
Electro- Boiling
chem. equi Melting points
Element Symbol Weight Val¬ valents, g. points, deg. C.,
ence1 per amp.- deg. C. visible
hr. ebullition
Aluminum.. A1 27.0 3 0.3354 658.7 1800.0

Antimony.., Sb 120.2 3 1.4949 630.0 1460.0
A 39.9 0 — 189.3 — 186 0
Arsenic. As 74.96 3 0.9324 850.0 450.02
Barium. Ba 137.37 2 2.5619 850.0
Bismuth.... Bi 209.0 3 2.5992 271.0 1440.0
B 10.9 3 2350.0 35002
Bromine.... Br 79.92 1 2.9814 -7.3 58.75
Cadmium... Cd 112.40 2 2.0955 320.9 778.0
Caesium.... Cs 132.81 1 26.0
Calcium.... Ca 40.07 2 0.7477 810.0
Carbon. C 12.05 4 0.1118 >3600.0 3700.0
Cerium. Ce 140.25 4 623.0
Chlorine.... Cl 35.46 1 1.3230 -101.5 -37.6
Chromium.. Cr 52.0 3 0.6476 1520 to>Fe 2200.0
Cobalt. Co 58.97 2 1.1000 16108
Columbium. Cb 93.1 5 1950-2200
Copper. Cu 63.57 2 1.1858 1083.0 2100.0
Dysprosium. Dy 162.5
Erbium. Er 167.7
Europium... Eu 152.0
Fluorine.... F 19.0 1 6.7085 -223.0 -187.6
Gadolinium. Gd 157.3
Gallium. Ga 70.1 30. i
Germanium. Ge 72.5 958.0
Glucinum... G1 9.1 1280.0
Gold. Au 197.2 3 2.4524 1063.0 2100.0
Helium. He 4.002 0 -271.9 -268.8
Holmium... Ho 163.5
Hydrogen... H 1.008 1 0.03764 -259.0 -252.8
Indium. In 114.8 154 5 3000 0
Iodine. I 126.92 1 4.7353 114.0 184!35
Iridium. Ir 193.1 4 2300 0 2850 0
Iron. Fe 55.84 2 1.0416 1530 ±5 2450!o
Krypton.... Kr 82.92 — 169.0 -151.7
Lanthanum. La 139.0 810.0
Lead. Pb 207.20 2 3.8654 327.4 1525.0
Lithium. Li 6.94 1 0.2622 186.0 500.0
Lutecium... Lu 175.0
Magnesium. Mg 24.32 2 0.4531 651.0 1120.0
Manganese. Mn 54.93 2 1.0255 1260 ±20 1900.0
Mercury.... Hg 200.6 2 7.4840 -38.87 357.0
Molybde¬
num. Mo 96.0 2 1.7900 2550.0 3350.0
Neodymium Nd 144.3 . 840.0
Neon. Ne 20.0 0 . -253.0 -245.9
1 In those cases in which a metal has two valences, the valence given corre¬
sponds to the electrochemical equivalent, and may not necessarily be the
commoner one.
2 Sublimes. 8 Commercial metal, about 1480° C.
APPENDIX 991
International Atomic Weights, 1922.—Continued
Electro- Boiling
Val¬ chem. equi¬ Melting points
Element Symbol Weight ence1 valents, g. points, deg. C.
per amp.- deg. C. visible
hr. ebullition
Nickel. Ni 58.68 2 1.0946 1452 ± 3 2450.0

Niton. Nt 222.4 0 — 71 0 — 62.0
Nitrogen... N 1 14.008 3 0.1742 -210. 5 -195.7
Osmium... Os i 190. 9 2700. 0 2950. 0
Oxygen.... O 16.00 2 0. 2985 -218.0 -183.0
Palladium. Pd 106.7 2 1.9951 1550. 0 2540. 0
Phosphorus P 31.04 44. 1 287.0
Platinum.. Pt 195.2 4 1.8206 1755. 0 2650.0
Potassium. K 39.10 1 1.4584 62.3 667.0
Prase ody-
mium.... Pr 140.9 940. 0
Radium. .. Ra 226.0 2 900. 0
Rhodium.. Rh 102.9 1940. 0 2750.0
Rubidium. Rb 85.45 38. 0 696. 0
Ruthenium 101.7 >1950. 0 2780.0
Samarium. Sa 150.4 1350.0
Scandium.. Sc 45.1 1200. 0(?)
Selenium .. Se 79.2 2 1.477 218.5 690.0
Silicon. Si ' 28.1 4 0.2638 1420. 0 3800.0
Silver. Ag 107.88 1 4.0258 961.0 1955.0
Sodium.... Na 23.00 1 0.8582 97.5 742.0
Strontium. Sr 87.63 2 1. 6333 >Ca<Ba
Sulphur.... s 32.06 2 0.5980 112.8-119.2 444.5
Tantalum.. Ta 181.5 2850. 0
Tellurium.. Te 127.5 2 2.379 451.0 1390.0
Terbium... Tb 159.2
Thallium .. T1 204. 0 302.0 1700.02
Thorium... Th 232.15 >1700. 0<Pt
Thulium.. . Tm 169.9
Tin. Sn 118.7 2 2.2144 231.9 2270. 0
Titanium.. Ti 48.1 4 0.4490 1795.0 ±15.0 2700. 0
Tungsten.. W 184.0 6 1. 1437 3267. 0 3700. 0
Uranium... U 238. 2 Near Mo. 3100.0
Vanadium. V 51.0 1720.0 + 20.0
Xenon. Xe 130. 7 0 -140. 0 -109. 1
Ytterbium. Yb 173.5 1800.0(?)
Yttrium... Yt 89. 33 1490. 0
Zinc. Zn 65.37 2 1.2192 419.4 918.0
Zirconium. Zr 90. 6 1700. 0(?)
Note.—In addition to the above elements, there is some reason to believe

in the existence of a gas “ cornium” (so called from its existence in the solar
corona) which would form 0.00058 per cent, of the earth’s atmosphere ac¬
cording to Dr. A. Wegener’s calculations (Science, Oct. 31, 1913).
1 In those cases in which a metal has two valences, the valence given corre¬
sponds to the electrochemical equivalent, and may not necessarily be the
commoner one. 2 Also given as 1280° C.
INDEX
Plain figures represent First Volume. Bold face figures represent Second Volume.
A Ammonia, absorber for, 664

critical constants for, 657
option glasses for pyrometers, 449, 450 formation from nitrate, 709
;ate of lime, evaporation of, 382 as oxidizer, 713
of soda, evaporation of, 382 properties of, 668, 669
;ie acid fermentation, 602 refrigeration by, 661
l, eggs, 120 Ammonium chloride, recovery of, 375
pumps for, 129 nitrate, manufacture of, 408
lity, effect on potential, 697 recovery of, 370
[proofing cements, 843, 844, 951 sulphate, recovery of, 370
.-radicals, in electro-refining, 743 Ampere, definition of, 731
nium, properties of, 880 Analysis by diffusion, 766
radio active series, 860 Anglo of nip, 192
ibatic compression, 147, 148 Anode, insoluble, 740
dnistrative control, problems of, 970 materials, 740
jrbents as weighting agents, 287 over voltage, 737
)rption, definition of, 780 shapes, 740
tion by air, 352 Anthracite, burning, 35
continuous, 347 separation, 207- 2(59
role of, in leaching, 344, 351 sizes of, 35
ator, baffle paddle, 351, 550 Anti monidea, roasting, 812
Dorr, 352 Antimony, electrolytic production of, 746
Pachuca, 353 oxidation-reduction behavior, 714
addle, 351, 550 Appendix, 985
propellor, 352, 550 Aqua regia, as oxidizer, 710
or agitation, 352 Are, temperature of, 524
conditioning of, 680 Arens, table of measures, 985
energy in compressed, 147 Argon, discovery of, 911
grading by, 225, 226 occurrence, 918
horsepower to compress, 170 properties, 920
for melting iron, 183 recovery from atmosphere, 918
necessary for combustion, 37 spectrum of, 920
specific heats of, 145, 14G Arsom furnace, 523
weight at various pressures, 143 Arsenic, oxidation-reduction behavior, 71
temperatures, 143 trioxide uh reduction agent, 716
ifts, 121 Arsenides, roasting, 812
classifier, 271, 272, 354 Asphalts, distillation products of, 846
?el defined, 771 for waterproof cements, 842
lol, oxidation of, 706 Atmospheric pressure at different altitudes, 29(1
Droduction of, 591, 598 Atomic weight table, 990
roof, 986 Atomization, electrical, 777
lyde, oxidation of, 706 Ayro’a coal-separator, 2(57
mder, Jerome, 565
iator (pumps), 127 U
s, deposition of, 746 Badger evaporator, 374
i rays, 858 Baghouses, general considerations, 31(1
nating current in electrolysis, 744 mechanically operated, 317
ide, effect on compression, 172 Baily furnace, 522
ides, atmospheric pressure at various, 290 Ball clays, 492, 493, 503
>arometer heights at various, 29G mill, 203, 200
inothermics, 520 Bank, rotation of, in trommels, 253
inous refractories, 507 Barbour, Percy ,K., 926
inum, electrolytic production of, 747 Barium chloride, recovery of, 370
xidation-reduction behavior, 701 precipitation from salt liquors,
ulphate, recovery of, 370 separation from radium, 874
ican filter, 297
sucratc, properties of, 405
993
994 INDEX
Barometer readings at different altitudes, 296 Bottom crystal, 402, 403

Battery-cell changes identical with oxidation and Boyles law, 144
reduction, 686 Breaking, primary, 211
Baum6 conversion table, 397 Brewery presses, 303, 305
Beater, 215 Brewing, 589
Beck arc, 524 Brick, red, 948
Beef extract, evaporating, 381 refractory, 946
Beet sugar, diffusion process for, 350 stacks, cost of, 186
evaporation factors, 379 Brine cooler, 674
osmosis in, 347 ,
British thermal unit defined, 1 655
Belgian zinc furnace, 798 Bromine, oxidation-reduction behavior, 724
Belting, horsepower transmitted by, 74 Bronze, manganese, specifications, 930
Beta rays, 858 Brownian movement, 765
Bichromates, oxidation by, 723 ,
Bruckner furnace, 805 818
Binary mixtures, distillation of, 625 Brunner-Mond process, 408
Bismuth, electrolytic production of, 746 Buckets, weight of, for cranes, 80
* oxidation-reduction behavior, 714 Buckman, H. H., 961
reduction of hydrate, 704 Buddies, 334
Bituminous cements, 841 Buhr-stones, use in grinding, 191
testing, 842 Building materials, 926
Black-body conditions, 443, 451-454, 461 stone, 936
Black liquor, concentrating, 383 timber, 934
Blake crusher, 194 ultimate strength of, 938
Blast, dry, 839 working stresses, 939
,
furnace, copper, 822 824 Bullion, moisture in, 579
iron, 834 sampling, 577
lead, 806 Burton process, 644
smelting, copper, 824
iron, 833 C
lead, 804
preheating, 839 Cadmium, electrolytes for, 745
,
roasting, 812 820 oxidation-reduction behavior, 702
Bleecker process, 865 Cake, mixers, 561
Bleininger, A. V., 473 moisture content on filtering, 295
Blinding of screens, 256 thickness in filtering, 295
prevention, 257, 263 Calcium acetate, recovery of, 382
Blowers, fan, 148 chloride, recovery of, 376
piston, 167 electrolytic production of, 748
Boots, 183 oxidation-reduction behavior, 701
steam jet, 166 sucrate, properties of, 405
Sturtevant, 183 Caliche, leaching of, 346
Blowing, grading by, 224, 225, 226 Callow cones, 276, 278
Boilers, Babcock & Wilcox, 6, 37 flotation process, 337
efficiency-capacity diagram, 15 screen, 246
fire-tube, 9 Calorie, defined, 1
grates, 12 Canvas tables, 334
Heine, 7 Cap cement, 855
horizontal, 8 Capacity, filter, 294
joints, 10 overload, defined, 1
settings, 12 reserve, defined, 1
steel for, 12 tables of, 985
Stirling, 7, 12 Carbon dioxide, critical constants for, 657
tubes for, 9, 11 properties of, 666
types of, 6 refrigerating by, 665
waste-heat, 13 content, iron and steel, 928
water-tube, 11 oxidation-reduction behavior, 705
Wickes, 8 refractories, 512
Boiling point-composition curves, 625 Carborundum, grading of, 232
points, standard, 410, 990 ,
Carnotite, Colorado deposits, 863 884
Bolters, gyratory, 243
Bonds, 982
,
treatment of, 865 885
Casein glue, 858
Bonecourt combustion, 520 from milk, 854
Booth water-softening apparatus, 278 Cast iron, properties of, 928
practice, 18 Catalysis, 749
IXDEX 995
•sta, in fermentation, 585 Clay, compositions, 497
. oil ('racking, 641 firing behavior, 495, 503
>180118 for, 754 flint, 491
romoter.s, 753 grog, 502
.'tic action, industrial development of, 757 kaolin, 491
mechanism of, 755 melting point of, 490
types of, 752 plastic refractory, 493, 501
horcniN, 288, 782 porcelains, 504
mry screen, 270 purifying by cataphoroHis, 289
dc, materials* for, 741 refractories, 490
icrrury, 742 refractoriness, 494
fluting, 741, 742 shrinkage water in, 494
•nintunct*, 742 siliceous, 490, 501, 50(5
it kiln for roasting, 818 uses, 505
ith, 841 Clay-bonded refractories, 507
UmninmiN, 842 Cleveland flash-point tester correction, 414
tp, 855 Cloth, cementing to metal, 855
■ucible, 841 screen, mesh, 259
irnnee, 846 silk, mesh, 258
lass, 856 Coal, analyses of, 34
vdruulie, 940 burning, 35
on, 848 classification of, 33
other, 848 ,
destructive distillation of, 648 649
uriiml consistency, 987 gas, reduction of Ita-Ha-sulphate by, 868
nrtlnnd, 941 grading of, 207 209
uzaulnn, 942 handling and storage, 37
oscudnlc, 951 heat value of, 33
ibher, 844 powdered, 40
,
»eeiheutmn«, 944 949 sampling, 571
*»ting, 847, 944, 949 separation by sliding friction, 207
uterproofmg, 841 steam sizes of, 35
tftiKul compressors, theory of, 178 Cobalt, electrolyte for, 744
umprt, 111 120 Cockle cylinders, 203
theory of, 117 Coehn’H law, 782
'eh*, 255 ("old junctions for thermocouples, 430, 438
'punihuu of gases, 312 correcting factors, 431
of Ihjuida, 309, 791 storage rooms, 672
'futhtiuc, 355 cooling of, 674
iftige, Hharplen, 309, 319, 321 Collateral trust bonds, 982
m law tit* affecting aerceniug, 289 Colloid chemistry, 765
, Ktingenberg, 4 scope of, 765
ingetnmnn, 88 ;
fuel, 522 789
oi mill, 202, 3*25 state, 768
tey«, draff, 184 Colloids, ehissifiefttion, 772
mm of, 81, 22 electrical properties of, 782
tdfisinjc rotufling, 813 optical properties of, 781
ne, oxidation mluetion behavior, 724 preparation by dialysis, 767
tie nlum, crystallising, 404 of solutions, 777
rick, building with, 946 ,
size of, 774 776
itel resist si ticca, *128 Combustion, fuel oil, 521
site refmetorlw, 810, 946 surface, 520
uium, oxidatiomreducthm Iwdmvior, 722 Bee also “fuels,” “boilers,” “furnaces.”
concent ration of, 383 Commercial paper, 988
nr mil, defined, 80 Compensating leads, 431
sin evaporator, 871, *178 ("ompresaed gases, 187
flout ion, 270 2B0 Compression, adiabatic, 147, 148
tier, Akin*, 271, 272, 884 altitude, effect of, 172
onea, 270 efficiency in, 169
)<mt, 271, 888 gas engines, 67
Wleral- Ksperan *h, 2*10, 271 multistage, 169
mohanieal, 270 of refractories, testing, 485
Ivoea, 284 theory of, 168
hw*C8a|ieyrfiri equation, 611, 612 two stage, in refrigeration, 662
ball, 492, 498, AOS Compressors, centrifugal, 175
996 INDEX
Compressors, hydraulic, 166 Conveyors, screw, 96-98

mechanical details of, 173 suction, 106-107
piston, 167 Cooling, in air compression, 181
Concentration, definition of, 222 Copper bullion sampling, 578
effect on electric pressure, 691 converter, 830
electrostatic, 336 ,
electrolytic production of, 744 745
jigs, 329 matte formation, 823
magnetic, 335 occurrence, 822
principles of, 323 oxidation-reduction behavior, 702
, ,
of rare ores, 886 889 904 pyritic smelting, 827
table, 334 slag formation, 823
testing for, 325 specifications for, 930
Concrete, acid proofing, 951 Copperas, evaporation, 385
aggregates for, 950 Corby dough-mixer, 557
as building material, 940 Cork board, 673
cement required, 942 Corrosion, prevention of, 933
materials for, 949 Corrugated steel, safe load for, 932
mixing, 554 specifications, 931
surfacing floors, 950 Cotton, loss of strength when heated, 316
water proofing, 842 Cottrell process, 320
Condensate, return in fractional condenser, 629 Coulomb, definition of, 731
Condensers, jet, 128 Counter-current washing, 286
for refrigeration, 660 Cracking oil, 639
steam plant, 30 Cranes, locomotive, 81-82
suction, 128 overhead, 83-86
surface, 128 Cristobalite, inversion to, 506
system for evaporators, 374 Critical pressures of gases, 657
Conditioning air, 680 temperatures of gases, 667
Conductivity, thermal, of clays, 500 Crosses, pipe, 45, 46
Conductors, proper size for, 739 Crucible cements, 846
Cones, Callow, 276, 278 Crucibles, electric, 626
Seger, melting points, 479 magnesite, 512
Coning samples. 567 mixtures, 513
Constants, tables of, areas, 985 Crude oil-distillation products, 643
atomic weights, 990 Crusher, Blake, 194
BaumS conversion, 397 disc, 214
boiling points, 410, 990 Dodge, 195
capacity, 985 gyratory, 196
elements, 990 * for sampling, 671
energy, 986 Crushing, general principles, 192
hydrometer equivalents, 397 intermediate, 211
length, 985 power required for, 198, 199
mass, 985 primary, 193, 211, 324
melting points, 990 secondary, 199, 324
solubility, 986 stamps, 213
specific gravity, 108, 144 strength of clays, testing, 486
heat, 144 of stone, 937
temperature, 985 surface diagram, 215
thermometric points, 410 Crystallization, 399
van der Waals’, 190 agitation, 402
volume, 985 concentration in, 400
Constant-temperature stil-healds, 636 cooling and concentration, 401
Construction, steel frame, 968 deep tank, 404
types of, 968 in motion, 403
wooden frame, 968 precipitation prior to, 404
Contact materials, for catalysis, 761 of radium salts, 869
Continuous agitation, 347 on strings, 404
cone classifiers, 276 of thorium salts, 909
mixing machines, 546 Cultures, production of pure, 593
Contraction of clays, testing, 480 typical, 599
Control, administrative, 970 Curb presses, 304
Converter, 830 Curie, Mme., work of, 857
Converting, 831
Conveyors, belt, 87-96
,
Curves, partial pressure, 618 621
vapor-composition, 622
INDEX 9E
, , ,
Curves, vapor pressure, 609 610 618 621 Draft for various coals, 32
Cut-out blades for mixers, 541 Drogin, D., 638
Cyanates, formation of, 684 Drug mill, 213
Cyanides, solution pressure, 702 Drum mixer, 544
Cylinders, cockle, 263 Dry blast, 680
Cytase, 604 Dryers, belt, 391
chamber, 388
D direct heat, 390
drum, 394
, ,
Daniell cell, 687 688 692 pan, 391
Debenture, meaning of, 981 plate, 392, 393
stocks, 982 ribbon, 395
Decantation, 276-287 steam, 392
counter current, 286, 354 tubular, 392, 393
Decortication, 268 tunnel, 388
Decrepitation, separation by, 269 vacuum, 394
Deflection galvanometer, 465 Drying, air, 386
Dehydration, catalysts for, 763 bibliography, 386
Dehydrogenation, catalysts for, 752 in chambers, 388
Deister tables, 335 practice of, 396
Depreciation factors, buildings, 5 steam, 391
conveyors, 95, 98 theory of, 386
formula for, 6 vacuum, 394
Destructive distillation, 638 Dualistic formulas, 684
of asphalt, 646 Duhem-Margules equation, 682
of coal, 648 Dust chambers, 311
of oil, 643 water in, 312
of wood, 647 Dutch oven, 38
Diagrams, crushing surface, 215 Dwight-Lloyd roaster, 821
entrance diagram, 116
ideal entrance, 115 E
settling, 327
Dialysis, 290 Economizers, 29
preparation of colloids by, 767 Edge runners, 205
Dialyzer, 784 Edwards furnace, 817
Diamond recovery on greased tables, 337 Efficiency (pump), 135-138
Diaphragms, 743 Elbows, pipe, 45, 46
Diastase, 587 Electric arc furnaces, 524
Diffusion, as factor in leaching, 341 cranes, 83-86
applied to analysis, 766 crucible, 526
of heat, 436 current, alternating, 50
Dinas brick, 946 heating effect, 50
•Direct current, advantages, 50 discharge, fume settlement by, 319-321
effect of, on compounds, 731 furnaces, 522
Disc mixing machine, 539 generators, 52
Disintegration of radioelements, 859 machinery, 50
Dissociation, energy of, 734 motors, 51, 53
Dissolution, see “Leaching.” pressures, effect of concentration on, 691
Distillation, destructive, 638 of various ions, 690
fractional, 607 Electrical units, 731
theory of, 623 ,
Electrochemical equivalents, 731 732
,
products of oil, 643 645 ,
series, 733 734
Distilled water, production of, 375 Electrodes, materials for, 748
Distillery slop, concentrating, 384 shapes of, 740
Distribution of power, economic, 974 Electrolysis, 731
Dodge crusher, 195 energy required for, 738
Dolomite, as refractory, 510 fused electrolytes, 747
Dorr classifier, 271, 353 in sedimentation, 285
thickener, 280, 286 separation by, 288
Dorr, J. V. N., 341 theory of, 732
998 INDEX
Electrolytes, cadmium, 745 Ether, refrigeration by, 668

calcium, 748 Ethyl chloride, refrigeration by, 669
cobalt, 744 Evaporation, bibliography, 357
copper, 744, 745 direct heat, 358
fused, 747 history, 357
gold, 744 multiple effect, 363
iron, 744 solar, 358, 401
lead, 744 steam, 359
magnesium, 747 theory of, 357
palladium, 745 vacuum, 362
platinum, 744 by waste gases, 358
potassium, 747 Evaporators, arrangement, 368
purification of, 746 film-type, 373
silver, 744 horizontal tube, 370, 372
sodium, 747 jacket-and-coil, 369
tin, 746 materials for, 369
zinc, 745, 801 Expansion of clays, testing, 480
Electrolytic cells, experimental, 842 thermal, 488
Electrolytic-solution pressure, 736
Electron, defined, 685 E
valence changes due to, 685
Electrophoresis, 288, 782 Factor, demand, 2
Electroscope, calibration of, 878 depreciation, 5
construction of, 877 diversity, 1
rise of, 877 load, 1
lectrostatic separation, 336 overload, 1
Eletcotihermic iron smelting processes, 748 use, 1
zinc smelting, 800 Fairchild, C. O., 409
I- lements, table of properties of, 990 Falling bodies, Stokes formula for, 225
El vators, bucket, 99-105 Fans, capacity, 148
Emanation, (See also “alpha," “beta," “gam¬ characteristics, 152
ma rays.’’) constants, 149, 150
boiling off radium, 875 design of, 155
imethod for radium, 873 for forced draft, 33
Emissivity constants, 454 fundamental formulas, 150, 151
corrections, 455 horsepower, 148
Emulsifying, 563 performances of, 151
Energy in compressed air, 147 testing, 155
units of, 990 types of, 163-166
Engineering materials, 926 Faraday, definition of, 734
Engines, Diesel, 68 Faraday-Tyndall effect, 781
economy, 20, 23 Faragher, W. F., 607
expansion ratios, 19 Feathers, grading of, 226
gas, 63 Federal Esperanza classifier, 269, 354
oil, 66 Feed water, heaters, 28
Otto cycle, 64 requirements for, 17
size of, 19, 20 treatment of, 18
steam, 17 Feld scrubber, 314
consumption, 22 Fermentation, 585
Ennis, W. D., 1 acetic acid, 602
Entrainment in evaporators, 368 alcoholic, 591, 698
Entrance diagrams, fans, 176 of grape must, 699
pumps, 116 lactic acid, 600
Enzymes, 586, 753 sewage, 604
Equation, Clausius-Clapeyron, 611, 612 Fermenting cellar, 698
Duhem-Margules, 632 Ferricyanide, oxidation-reduction behavior,
Gibbs-Helmholz, 735 Ferrocyanides, crystallizing, 406
Thomson, 734 oxidation-reduction behavior, 730
van der Waals, 144, 765 Ferromolybdenum, 905
Equilibrium, chemical, 750 Ferrotungsten, 899
data, 758 Ferrouranium, 891
Nernst’s formula, 759 Ferro vanadium, 887
Esperanza classifier, 354 Ferrous sulphate, evaporation, 385
Ether, properties of vapor, 669 Fery optical pyrometer, 445
INDEX 995)
•Fery radiation pyrometer, 458 Furnace, Pearce, 818

Film-type evaporators, 373 Ropp, 818
Filter aids, 303 Wedge, 819
block, 306 zinc, 796
capacity, 294 Furnace cements, 846
continuous, 291 Furnaces, electric, 522
general subject, 290 for testing refractories, 477
life of fabric, 294 Fused electrolytes, 747
pressure, 290, 867
salt, 374 G
vacuum, 291
Filtration, cost of, 297 Galena, settling ratios, 327, 328
general subject, 290-306 Galilith, 864
intermittent, 297 Galvanometer method with thermocouples, 422
pressure, 297, 867 Gamma rays, 858
vacuum, 294 Ganister brick, 946
Filtros blocks, 306 Gas engines, 63, 64, 65
financing, methods of, 979 mantle manufacture, 910
Fink, Colin, G., 619 producers, 59, 60
firebrick, 946 thermometer, 409
firing behavior of clays, 495 Gases, properties of, 144
Sittings, general specifications, 45, 46, 47 specific gravity, 144
rlames, effect in pyrometry, 456 heats of, 146
•’lange fittings, templates, 43, 44 transportation of, 143-191
•'lash-point corrections, 414, 415 Gasket compositions, 848
i’lat screens, capacity of, 247 Gates, A. O., crushing surface* diagram, 215
screening with, 233 Gauze, gritz meshes, 258
’’lint clays, 491, 499 temperature of glowing, 455
dotation, 337-340 Gelatin, evaporating, 381
Callow process, 337 Generators, types of, 52
consumption of oils in, 787 Germination, chemistry of, 688
Minerals Separation, 338 Gibbs-Helmliolz equation, 735
oils for, 340 Given, Arthur, 399
principles of, 786 Glaser, G., 685
'lue-dust, 829 Glass, cementing, 866
collection of, 311 Glass-pot mixtures, 502
luorides, oxidation-reduction behavior, 724 Glowing gauze, temperature of, 455
oaming in evaporators, 368 Glucose, evaporation factors for, 379, 380
oot-pound, definition of, 1 Gluos, 841
oot valve, 127 casein, 853
oote, Paul D., 409 evaporating, 381
oote-Fisher pyrometer, 448 marine, 845
orce, centrifugal, chart for, 308 practical work with, 852
formula for, 310 testing, 848
ormulas, dualistic, 684
water proof, 854
ractional distillation, 607
Glycerin, concentration of, 383
theory of, 623
produced by fermentation, 605
ractionating column, 627
Gold bullion sampling, 577
ractionation of radium solutions, 871 colloidal, preparation of, 777
agments, shape of rock, 234
electrolytic production of, 744
ree settling ratio, 327
number, colloidal chemistry, 782
reeth process, NH4NO» making, 408
oxidation-reduction behavior, 702
reezing points, standard, 410
Goldschmidt method of heat production, 02
action, rolling, 228
Golodetz theory of distillation, 681
separation by sliding, 266
Gooch crucible, works size, 306
lels, analyses of solid, 39
Grading, by blowing, 224
colloidal, 521, 789
definition of, 224
gaseous, 40, 59, 520
general principles, 221
oil, 37, 521
miscellaneous means of, 223
powdered coal, 40
by one dimension, 223
wood, 39, 41
by projection, 226, 227
tme settlement by Cottrell process, 320
by rolling, 227
irnace, blast, 806, 822, 834
by screens, 233
Merton, *818
by volume, 227
1000 INDEX
Grading, by weights, 223 Hydraulic cement, 940

Granulating, 212 Hydraulic compressors, 166
Grape juice, concentration of, 383 efficiency, definition, 119
Graphite crucibles, 513 lime, 943
Greene, Van Rensselaer H., 655 ram, 122
Grinding, general principles, 192, 204 Hydrazine, as reducing agent, 707, 713
machines for, 204 Hydrogen peroxide, oxidation-reduction b<
pans, 205 havior, 717
tube mill, 207 Hydrogenation, catalysts for, 752
wet, 356 Hydrolysis, acid, 590
Gritz gauze meshes, 258 Hydrometer equivalents, 397
Grizzly sections, 256 Hydroxylamine, as reducing agent, 707, 712
Grog, 502, 505 Hypophosphites as reduction agents, 715, 721
Grossalmerode, clay from, 499, 502 Hytor compressor, 184
Gyratory bolters, 243
I
crusher, 196
Ice, latent heat of fusion of, 655
H
making, 676
Ice-plants, miniature, 678
Halides, oxidation-reduction behavior, 724
Income-bonds, 982
Halogenation, catalysts for, 753
Incrustations in evaporators, 368
Halogens, oxidation-reduction behavior, 724
Indorsements on bonds, 982
Hancock jig, 330 Induction furnaces, 626
Hansen's pure-culture apparatus, 594
Industrial Filtration Gorp., 291
operation of, 596
Ingalls, Walter Renton, 793, 802
sterilizing, 596
Injectors, boiler feed, 28
Hardinge mill, 211 Insulating refractories, 517, 673
Harrison, T. R., 409 International magnetic separator, 336
Harrison-Foote compensated indicator, 424
Interstitial action, 243
Harz jig, 330 separation by, 266
Haynes-Engle process, 866 Inversion of starch, 589, 590
Hazards, plant, 976
Invertase, 587
Heat balance, in boiler operation, 14
Investment, nature of plant, 962
gas engine, 65 Iodine, oxidation-reduction behavior, 724
multiple effect evaporators, 364
Ions, formation of compounds from, 694
oil engine, 65 migration velocity, 736
Heat transmission of, 360
theory, 732
units of, 1
Iron cements, 846
of reaction in catalysis, 761
carbon content, 928
Heating buildings, steam for, 6
cast iron, 928
electrical, 50, 54
definitions, 927
surface boilers, 13 electrolytic production of, 744
Helium, commercial manufacture, 917 oxidation-reduction behavior, 729
discovery of, 912 protection against corrosion, 933
formation of, 916
in samples, 575
homogeneity of, 914
sampling of, 679
liberation from minerals, 916
segregation in, 580
liquefaction, 917
Isinglass, fining with, 600
occurrence, 914
properties, 916 J
purification, 915
High-temperature production, 619 Jigging, principles of, 326
Hindered settling ratios, 328 Jigs, air, 329
unsized material, 334 grain-stroke ratio, 332
Hoists, 83 Hancock, 330
Holborn-Kurlbaum optical pyrometer, 449 Harz, 329
Hopjack, 689 types of, 329
Horsepower, defined, 1 Joints, pipe, 41
Horsepower-hour defined, 1 Jones sampler, 668
Howard vacuum pan, 369 Junction box, 436
Huff electrostatic separator, 336
Humus in sand, 951 K
Huntington-Heberlein roaster, 809
Hydration, catalysts for, 753 K, equilibrium constant, 750, 751, 752
INDEX 1001
Kaolin, 491 Locomotive cranes, 79-83
Kataphoresis, 288, 782 Lubrication a colloidal matter, 787
Kelly filter, 304 Lutes, composition of, 841, 842, 843
Kestner evaporator, 373
M
Kick’s law, 216
Kilowatt defined, 1 McKesson-Rice screenless sizer, 229
hour defined, 1 Magnesia-alumina fusion diagram, 511
Kish, formation of, 512 Magnesite, as refractory, 509, 510
Klingenberg chart, 4 Magnesium chloride, recovery of, 376
Kneader, reciprocating, 659 electrolytic production of, 747
Kneading, 529 oxidation-reduction behavior, 701
Koppers byproduct coke, 651 oxychloride, 846
Krypton, discovery of, 912 sulphate, recovery of, 376
occurrence, 922 Magnetic concentration, 335
properties, 923 Maltase, 687
separation, 922 Malting process, 588
spectrum, 923 Maltose, evaporation factors for, 379, 380
Manganates, formation of, 684
L Manganese bronze specifications, 930
Lactic acid, bacteria, 601 Mantius evaporator, 373, 374
evaporation of, 382 Mantius, Otto, 357
fermentation, 600 Mashing process, 589
Lamps, arc, 58 Mass action in colloidal chemistry, 772
incandescent, 57 in leaching, 342
Latent heats of gases, 657 tables of units of, 985
Laterals, pipe, 45, 46 Mechanical draft, 185
Leaching, agitation in, 344, 351 Megraw, H. A., 191
copper, 745 Melting points, Seger cones, 479
counter current, 348 standard, 410, 990
definition of, 341 Mercerizing liquors, concentrating, 384
diffusion in, 341 Mercury, cathode, 742
mass action in, 342 oxidation-reduction behavior, 702
by percolation, 340 Merton roaster, 818
temperature, 345 Mesothorium, properties of, 879
Lead, blast-furnace smelting, 804 series, 860
electrolytic production of, 747 Metallics, sampling of, 681
metallurgy of, 802 Methyl chloride, refrigeration by, 669
oxidation-reduction behavior, 703 Metric equivalents, 986
removal of from radium salts, 870 Middlings purifier, 264
sampling, 578 Milk, evaporating, 381
slags, 806 Minerals Separation process, 338
smelting, 803 Miniature ice plants, 678
Leads, compensating, 431 Miscibility, conditions for, 615
Lead-zinc smelting, 799 Mixing and kneading, 529
Leather cements, 846 disc machines for, 540
Le Chatelier-van’t Hoff law of mobile equilib¬ dough, 541
rium, 609, 617 liquids with liquids, 547
Leeds & Northrup optical pyrometer, 440 with solids, 552
Length, units of, 985 solids with solids, 533
Liddell, Donald M., 143, 565, 731, 979 wheel, 546
Lighting, electric, 56, 57, 58 Mobile equilibrium, law of, 609, 617
Lillie film evaporator, 372 Moisture in filter cake, 295
Lime, hydraulic, 943 samples, 566
Linde liquid gas process, 671 Molasses, treatment of, 403
Lipase, 604 Molybdenum, analytical methods, 906
Liquefied gases, refrigeration by, 671 compounds, 904
Liquid air, uses of, 671 concentration, 994
Liquids, cooling of, 675 extraction, 903
Load, average, 1 ferroalloy, 905
defined, 1 metallic, 905
factor, 1 occurrence, 903
Location, economio, of plant, 967 reduction of, 905
topographic, 968 uses of, 906
1002 INDEX
Monazite, extraction, 908 Osmosis, in beet sugar industry, 347

occurrence, 907 in cane sugar industry, 405
Moore, R. B., 867, 883, 911 Otto cycle, 64
Morrison, A. Cressy, 187 Overload capacity defined, 1
Morse pyrometer, 449 Overvoltage, 737
Mortar and pestle, 644 Ovoca classifier, 354
Mortar, water required for, 968 Oxidation, 683
Mortgages, 982 catalysts for, 752
Motors, alternating current, 53 electric pressures of, 690
induction, 53 equation of single, 695
speed control of, 51 identity with battery-cell changes, 686
types of, 51 powers, 696
Multideck classifier, 271 reversible reaction, 688
Multiple effect evaporation, 363 Oxidizer may be reducer at same time, 699
heat balance in, 364 Oxyacetylene flame, 520
steam consumption, 364 Oxychloride cements, 846
water consumption in, 367 Oxyhydrogen flame, 520
Multiple refining, 738 Oxygen, oxidation-reduction behavior, 717
Oxywater-gas flame, 520
N Ozone, as oxidizer, 718
Neon, discovery of, 912 P

liquefaction of, 922
occurrence, 921 Paddle mixer, 537
properties, 922 Paints, covering power of, 934
separation, 921 for iron and steel, 933
spectrum of, 922 Palladium, electroplating, 745
Nemst’s equilibrium formula, 759 Pans, grinding, 205
Nichrome resistance, 623 Howard’s vacuum, 369
Nickel, F. F., 109 Paper, cementing to metal," 856
Nickel steel specifications, 929 Partial pressures, measurement of, 624
Nip, angle of, 192 Passivity, 737
Niton, 926 Patronite, treatment of, 885
occurrence, 926 Pearce roaster, 818
properties, 926 Pecqueur’s evaporator, 369, 370
separation, 926 Pectization, 771, 782
spectrum, 926 Peltier e.m.f., 419
Nitrates, reduction to ammonia, 709 Pensky-Martin flash point corrections, 415
Nitric acid as oxidizer, 710 Peptization, 782
oxide, formation of, 709 Percolation, leaching by, 349-351
peroxide, formation of, 708 Perfecto filter bags, 317
Nitrites, as oxidizer, 711 Perhalates, oxidation-reduction behavior, 724
Nitrogen fixation, catalysts for, 762 Periclase, formation of, 510
oxidation-reduction behavior, 707 Periodic table, rare gases in, 913
valences of, 708 Permanganates, oxidation-reduction behavior,
Nitrous acid, as oxidizer, 711 729, 730
Normal consistency of cement, 957 Permeable membranes in selective dissolution,
solution, definition of, 700 346
Northrup pyrovolter, 427 Persulphates, oxidation by, 719
Notes, 982 Peters, F. F., 341
Phase-rule, 608
O Phosphorus, oxidation-reduction behavior, 714
Pictet liquid, 668
Ohm, definition of, 731 Pigment mixer, 537
Oil cracking, 639 Pinch effect, 526
distillation products, 643, 645 Pipe, flange fittings, 43, 44
flotation, 340, 787 Joints, 41
fuel, 32 sizes required for steam, 48
Oilproof lutes, 843 specifications for, 42
Olary ore treatment, 866 Pipette for glue testing, 849
Oliver filter, 291 Piping, design of, 49
Optical pyrometry, 443-455 for steam plants, 41
Organic extracts, preparation of, 350 specialties, 48
Osmosis, 290 Pitchblende, American deposits, 862, 863
INDEX 1003
chblende, Australian deposits, 862 Pressure, vapor (see “Vapor pressures").

European deposits, 861, 863 Processes, types of, 964
metallurgy of, 864 Producer gas, 59, 60
t design, 961 composition of, 61, 62
administrative control, 970 Projection, grading by, 227
hazards, 976 Promoters, in catalysis, 753
insurance, 976 Proof of alcohol, 986
location, 961 Propeller fans, capacity of, 148
power costs, 3, 4, 5 Protecting tubes for thermocouples, 440-442
sanitation, 976 Protective colloids, 773, 783, 791
waste disposal at, 975 gold figure, 782
ter of Paris, cements with, 844 Pulsometers, 120
tic clays, 493, 501 Pulverizing, 213
sties, mixing, 563 Pumps, 109-142
tinum, electrolytic production of, 744 auxiliary for steam plants, 27
thermometer, 462 capacity, 133
arization, 738 centrifugal, 114
ar valences, calculation of, 693 compound, 110
es, tabulation in order of forces, 689 deep well, 113
onium, properties of, 880 efficiencies, 135-138
series, 859 high duty, 112
ysaccharides, inversion of, 590 operating, 140
■celain type clays, 504 pistonless, 120
•osity of refractories, 483 power, 112
•tland cement, definition, 941 setting up, 141
lutes with, 843 simple cylinder direct, 109
specifications for, 943 speeds for, 111
•tland filter, 291 testing, 138
covering, 293 triple expansion, 120
;ash recovery by Cottrell process, 320 vacuum, 30
,assium bichromate, recovery of, 377 Pure-culture apparatus, 594
carbonate, recovery of, 377 operation of, 596
chloride, recovery of, 377 Purification of electrolytes, 746
electrolytic production of, 747 Puzzolan cement, 942
hydroxide, recovery of, 377 Pyne, F. R., 812
nitrate, recovery of, 377 Pyod, 422
oxidation-reduction behavior, 701 Pyritic smelting, 827
sulphate, recovery of, 377 Pyroligneous acid, evaporating, 381
,ential change with varying acidity, 697 Pyrometry, 409-472
entiometer method with thermocouples, • optical, 443—455
426-430 radiation, 443-462
ver, cost of gas, 66 recording, 466
economic distribution, 974 thermoelectric, 419-440
purchase of, 57 thermometers, 409-416
transmission, 74 windows in, 456
ver-efficiency curve, centrifugal pumps, 116 Pyrovolter, 427
ver plant, auxiliaries, 27
costs, 3, 4, 5 Q
general discussion, 1
overhead, 4 Quartz, inversion of, 506
size of, 3
ver required for gyratories, 198 R
for jaw crushers, 199
for rolls, 201 Radiation, from evaporators, 368
for tube mills, 210 pyrometry, 443-462
heaters, in evaporation, 375 Radioactive series, actinium, 860
ss, brewery, 303 305 thorium, 860
curb, 304 uranium, 859
filter, 298 Radioactive substances, helium from, 916
leaves for, 299 Radioactivity, discovery of, 857
isure, critical, of gases, 657 (see also “radium”).
in oil cracking, 640 Radiolead, properties of, 881
osmotic. 785 R ium. 857
1004 INDEX
Radium, discovery of, 857 Refrigerating machines, compression, 660

disintegration of, 859 vacuum, 670
emanation, see “Niton.” Refrigerating systems, 659
measurements of, 872 Refrigeration, 655
metallurgy of, 864-872 by air, 669
ores of, 861 by ammonia, 661
Raisin seeder, 261 by carbon dioxide, 665
Ram, hydraulic, 122 by ether, 668
Rare gases, discovery, 911 by ethyl chloride, 669
general properties, 913 by liquefied gases, 671
in periodic table, 913 mechanical, 656
Rare metals, 883 by methyl chloride, 669
Raymond mill, 313 Pictet liquid, 668
Rays, alpha, 858 by sulphur dioxide, 667
beta, 858 Reheaters, for compressors, 174
gamma, 858 Reilly evaporator, 370
X, 858 Resistance, cell, 742
Reaction velocity, 750, 759 thermometry, 462
Reciprocating kneader, 559 Resistances, measuring, 464
Recording pyrometers, 466 Retort, zinc, 795
Red-lead cements, 843, 845 Reverberatory roasting, 817
Reducer may be oxidizer at same time, 699 smelting, 828
Reducers, pipe, 45, 46 Revolving screens, 269
Reduction, 683 Rhenish zinc furnace, 797, 798
electric pressures of, 690 Riehle machine, 847
equation of single, 695 Riffle, assayers, 669
identity with battery cell changes, 686 Rillieux horizontal-tube evaporator, 370, 372
of iron oxide, 835, 839 Ringelmann’s smoke chart, 35
reversible reaction, 688 Ring-roll machines, 205
Reels, centrifugal, 255 Rittinger’s law, 216
Reese, L. C., 529 Roaster, Brilckner, 818
Refining, systems of, 738 cement kiln for, 818
Refractories, 473, 946 Dwight-Lloyd, 821
aluminous, 473 Edwards, 817
bibliography, 517 Huntington-Heberlein, 811
carbon, 512 Merton, 818
chemical composition, 474 Ropp, 818
chromite, 516, 946 Wedge, 822
clay, 490 Roasting, 810
compression, resistance to, 485 blast, 812
crucible mixes, 513 chloridizing, 813
dolomite, 511 heap, 815
expansion of, 480, 488 lead ores, 805
furnace for testing, 477 muffle, 816
high temperature, 519 reverberatory, 816
lime bond, 505 shaft, 816
magnesite, 509, 510 stall, 816
miscellaneous, 516 sulphatizing, 811
molecular composition, 475 temperatures of, 810
porosity, 483 types of, 810
refractoriness, 476 Roast-reaction process, 805
silica, 505, 946 Robert evaporator, 373
sdicon carbide, 514 Rock-breaking machinery, 192
slagging of, 488 Rock candy, crystallizing, 404
special, 515 Rolling friction, theory of, 229
specific gravity of, 484 Rolls, crushing with, 200
spinel, 511 feed for, 201, 202
types of, 490 power consumed by, 201
volume studies on, 482 Roofing nails, weight of, 9321
zirconia, 515 Roots blower, 183
® Refractoriness of clays, testing, 476 Rope drives, 76
Refrigerating effect, measurement of, 659 Ropp roaster, 818
rating of, 657 Roscoelite, occurrence, 884
Refrigerating machines, absorption, 663 treatment of, 885
INDEX 1005
Seamless tubes, production of, 742
Rosendale cement, 951
Rotary filter, Vallez, 303 Sedimentation, 276—287
Seger cones, melting points, 479
Rubber cements, 844
softening temperature diagram, 498
Runners, for water turbines, 71
Segregation in pig iron, 580
S Selective dissolution, 345
permeable membranes in, 346
Sadtler, S. S., 841, 949 Separation, air, 263, 265
decortication, 269
St. Joachimsthal, 861
Salt, in chloridizing roasting, 814 decrepitation, 269
definition of, 222, 261
filter, 374
production. 359, 401 electrolytic, of clays, 288
Salting samples, 583 electrostatic, 336
Samplers, assayers, 569 interstitial action, 266
Brunton, 569 magnetic, 334
Jones, 568 screening, 261
Snyder, 570, 571 sliding friction, 266
Separators, in evaporation, 375
Vezin, 570
Samples, contamination of, 575 Series refining, 738
Setting time of cement, determination of, 958
cutting down, 567
dirt and scale in, 576 Settlement, 276, 282
Sampling, bibliography, 584 effect of electrolytes on, 285
bullion, 577 Settling ratio, free, 327
coal, 571 hindered, 328
crushers for, 571 Sewage, fermentation, 604
definition, 565 liming of, 287
Shafting, horsepower transmitted by, 74
gases, 582
by ladle, 578 Shaft roasting, 816
liquids, 582 Shale-distillation, 641
metallic material, 676 Shape of rock fragments, 234
moisture determination, 566 Sharpies centrifuge, 309, 310, 321
salting, 583 Shore pyroscope, 446
size-weight ratio, 572 Shot, grading of, 227
Shrinkage of clays in firing, 503
theory, 565
tungsten powder, 900 Shrinkage water in clays, 494
Sand, leaching of, 349 Sieves for cement testing, 959
Sand-sulphur mixes for tanks, 744 Silica brick, 505, 507, 947
Sanitation at plants, 976 Siliceous clays, 496, 501
Scale, heat transmission by, 3C0 Silicic acid, soluble, 767, 770
Silicon carbide refractories, 514
Schoch, E. P., 683
Schurecht voluminometer, 482 Silk cloth meshes, 258
Schwerin process for purifying clay, 288 Sillimanite type of refractories, 508
Scimatco pyrometer, 448 Silver bullion sampling, 577
electrolytic production of, 744
Screening, definition, 221
general principles, 221
grading by, 233, 234, 261 Simerenko evaporator, 372
motion, proper, for, 244 Sine-curve block, 224
Sizer, McKesson-Rice Screenless, 229
precise, 249
shape as affecting, 237 Size-weight ratio, 572
tests, 250 Slag cement, 942
‘ theory of, 238 Slags, copper, 823
wet, 245 iron, 837
Screenless Sizer, McKesson-Rice, 229 lead, 806, 809
Screens, blinding of, 256 zinc, 795
capacity of, 239, 247 Smelting, copper, 822, 824
caternary, 270 lead, 806
classification of, 242 iron, 834
cloth for, 258 pyritic, 827
mesh, 258, 259 zinc, 793
Smoke, a colloidal phenomena, 790
revolving, 250
section materials, 257 effect in pyrometry, 456
Ringelmann’s chart for, 35
vibrating, 248
Sodium, acetate, evaporation of, 382
Scrubbers, Feld, 313
benzol sulDhonate evaporation, 382
gas, 311. 313
1006 INDEX
Sodium, bichromate, recovery of, 377 Steffen’s process, 406

carbonate, recovery of, 377 Still-heads, constant-temperature, 638
chloride, recovery of, 378, 401 Stirring rod mixer, 536
chlorate, recovery of, 378 Stocking dust collectors, 318, 319
electrolytic production of, 747 Stocks, debenture, 982
hydroxide, recovery of, 378 common, 979
nitrate, recovery of, 378 preferred, 980
solubility, 346 Stokers, chain grate, 36
nitrite, recovery of, 378 Roney, 36
oxidation-reduction behavior, 701 Stokes formula, 225
peroxide, oxidation by, 718 Stone, building, crushing strength of, 939
phosphate, recovery of, 379 Stove, hot blast, 838
silicate, recovery of, 379 Strontium, oxidation-reduction behavior, 701
silicate lutes, 843 in sugar-refining, 406
sulphate, recovery of, 379 Structural steel specifications, 929
sulphide, recovery of, 379 timber specifications, 934
Softening-temperature diagram (clays), 498 Sturtevant blower, 183
Solar evaporation, 358, 401 Sublimation, 638
Solder, 841 Sucrates, properties of, 405
Solubilities, tables of, 986 Suction condenser, 128
Solution curves, 343 head (pumps), 128
normal, defined, 700 lift (pumps), 127
pressure, 735 Sugar, conversion of starch to, 589
speed of, 344 crystallization of, 403
Sorel stone, 846 diffusion process for, 350
Space-velocity ratio in catalysis, 760 evaporation factors for, 379
Spark-plug composition, 516 osmosis in, 347
Specific gravity, of clays, testing, 484 Steffens house work, 406
tables of, 108, 144 Sulphagel, defined, 771
Specific heats of gases, 144 I'-ib Sulphate liquor, concentrating, 383
Speed of pumps. 111 Sulphatizing roast, 811
of tube mills, 211 Sulphite liquor, concentrating, 384
Spelter, specifications for, 930 Sulphites, reduction by, 720
Spiral anthracite separator, 268 Sulphur dioxide, critical constants of, 657
Split shovel, 568 properties of, 667
Sponge iron, 839 reduction by, 720
Stack, cost of brick, 33 refrigeration by, 667
losses, 16 Sulphur, oxidation-reduction behavior, 718, 722
sizes, for coal, 31 roasting off, 805, 810
for oil, 32 Sulphuric acid, concentrating, 379
Stamps, 213 oxidation by, 719
Standards for glue, 850, 851 Sulphur-sand mixes for tanks, 744
Stannites as reducers, 704 Superheaters, steam plant, 29
Starch, conversion of, to sugar, 589 Surface combustion, 520
Starting sheets, production of, 741 condenser, 128
Steam, consumption in engines, 17 tension, 778
in evaporators, 359, 364 values of, 779
flow of, in pipes, 48 Sutton, Steele <fc Steele table, 265
for heating buildings, 6 Sweetland filter press, 303
in oil cracking, 639 Swing hammer crusher, 199
pipes, sizes required, 48 Symons disc crusher, '214
properties of, 49 Synaeresis, defined, 771
for pumps, 137
T
turbines, 23
Steel, carbon content, 928 Table, Canvas, 334
corrugated, specifications, 931 Deister, 335
definitions, 927 Sutton, Steele & Steele, 265
frame construction, 968 Wilfley, 333
nickel, specifications, 929 Tank, arrangements, 739
protection against corrosion, 933 for experimental electrolysis, 842
structural specifications, 929 materials, 743
weight per cu. in., 929 water, concentrating, 385
Steepwater, evaporation factors for, 380 Tanning extract, evaporating, 381
Stefan-Boltzmann law, 445 Tar, composition of, 652
INDEX 1007
Parian c acid, evaporating, 3S2 Transformation point recorders, 400
Pate process of waterproofing, USD Transmission of heat, 300
Pay lor, Hugh S., 749 Tmut.Hchold, Reginald, 70
Pees, pipe, 45, *1(5 TrayH in fractionating column, 628, 629
IVmperuture, eonversion eonstants, 985 Tridymite, inversion to, 500
control, nutomntie, 470, 471 Trommels, 250, 251
ofTeet. of, on leaching, 345 capacity of, 254
Pemplets, copper sampling, 578 motion in, 252
I'esting cements, 847 range of, 254
faun, loo wot screening in, 254
glue, 848 Tube mill, 203, 200
'etruhedrons, production of, in crushing, 23.r> Tubes for protecting thermocouples, 440
'etmthionntes, formation of, 721 Tumble-box mixers, 540
'hernial conductivity, clays, 500 Tungsten, analytical methods for, 900
'h<‘rmal expansion of clays, test ing, 4S,X compounds, 897
'hermit method of limit, production, 320 concentration of, 897
'hennocouplcH, 410 extraction of, 896
calibration data for, 443 ferroalloy, 899
eoiiHtancy of, 420 metallic, 898
galvanometer with, 422 occurrence, 894
immersion depth, 430 reduction of, 898
i nstullntions of, 432 sampling, 900
junction box, 430 uses of, 899
mounting of, 421 Turbines, economy, 24
potentiometer with, 420 elliciency curve of, 20, 27
protecting tubes for, 440 extraction, 27
reproducibility, 420 low pressure, 25
wiring diagrams, 434, 433 sixes, 24
hertnodynumie scale, 400 steam, 23
hermometer, accuracy of, 412 superheat for, 25
distance-reading, 417 water wheel, 71
emergent-stem correction, 414 Tyler standard screen, 250
lean, 400
indicating, 4 Hi IT
industrial, 413
mercurial, 4 10, 410
Ultracentrifugation, 791
partial immersion, 410, 411, 415
UltramicroHCOpe, 776
platinum, 402
United Kastern, classification at, 274, 278
pressure, 417
Universal wax, 855
recording, 410
Uranium, analytical methods, 892
resistance, 402
compounds of, 890
vapor-pressure, 417
concentration of, 889
hermometrir points, standard, 410
extraction, 889
tiermometry, 410 t<( mq.
ferroalloy, 891
icrmo-opticnl pyrometer, 440
metallic, 891
dekener, Dorr, 2K0
occurrence, 889
deknew* of filter cake, 205
oxide, calculating percentage, 878
lioHulphntea, reduction by, 721
radioactive series, 859
imnson e.m.f., 410
uses of, 891
rule, 735
Utah Copper Co., classification at, 272
torium, analytical methods, 907
compounds, 909
extraction, 908
V
metallic, 909
occurrence, 907 Vacuum augmented 31
radioactive series, 880 chamber (pump®), 127
lines of, 910 evaporation, 302
iwing radiation pyrometer, 457 filtration, 204-208
rnber Butte, classification at, 275 Howard’s pan, 800
tuber, structural, 984 machines for refrigeration, 670
ultimate strength, 98® pan with agitator, 650
working stresses, 987 pump, 30
n, electrolytic production of, 74® Valence, calculation of polar, 698
n, oxidation-reduction behavior, 70S changes due to liberation of electrons, 68®
1008 INDEX
Vallez rotary filter, 303 Water-softening apparatus, Booth, 278

Valve materials, pumps, 125 Water-softening methods, IS
motions, pumps, 123 Wedge furnace, 819
Vanadinite, treatment of, 885 Weighting, separation by, 287
Vanadium, analytical methods, 888 use of adsorbents in, 287
compounds of, 886 Wet screening with trommels, 254
concentration of, 886 Wetherill magnetic separator, 336
extraction of, 884 Wheatstone bridge, use of, 464
ferroalloy, 887 Whisking, 563
metallic, 887 White lead, 933
occurrence, 88 3 Wiard, E. S., 221, 261, 323
uses of, 887 Wien’s law, 444
van der Waals’ constants, 190 Wilfley table, 333
equation, 144, 765 Wilhelm, R. M., 409
Vapor-composition curve, 622 Windows, in pyrometry, 456
Vapor pressure, curves, 609, 610, 618, 621 Wiring, allowable sizes, 55
defined, 607 specifications, 55
dependent on form of liquid surface, 614 thermocouples, 434, 435
measurement of, B12 weights required, 56
of mixed liquids, 614, 617 Wood, destructive distillation of, 647
of non miscible liquids, 616 for fuel, 39, 41
water, in air, 681 ultimate strength, 936
Velocity diagram (pumps), 115 working stresses, safe, 937
Velocity of reaction, 750 Wooden frame buildings, 968
Vezin samplers, 570 Wool scouring wastes, concentrating, 384
size weight ratio for, 574
X
Vibrating screens, 248
Vinegar, production of, 602
Xenon, discovery of, 912
Vinous fermentation, 599
occurrence, 924
Viscosity pipette for glue, 849
separation, 924
Volt definition of, 731
spectrum of, 925
Voltage due to segregation, 737
to solution pressure, 737 Y
Volume changes of clays, testing, 482
Volume, table of measures, 985 Yaryan evaporator, 372, 373
Voluminometer, Schurecht, 482 Yeast, 592
disposition of, 598
W Yelinek evaporator, 370, 372
Waals’, van der, constants, 190 Z

equation 144, 765
Wanner pyrometer, 446-448 Zaremba evaporator, 372
Washing, counter-current, 286, 354 Zeolites, 18
in filter presses, 300 Zinc, chloride, concentrating, 385
Waste, disposal at plants, 975 distillation of, 793
gases, temperature of, 13 electrolytic refining, 745, 802
Water, apparatus for distilling, 375 electrothermic smelting, 800
in dust settlement, 312 furnace design, 796
latent heat of freezing of, 655 hydrometallurgy of, 801
in multiple effect evaporation, 367 metallurgy of, 793
required in mortar, 958 oxidation-reduction behavior, 702
shrinkage, in clays, 494 paints, 933
vapor in air, 681 reduction temperature of, 793, 794
Water-gas manufacture, 750 retort, 795
Waterpower, 69, 70, 71, 72 specifications for, 930
cost of development, 73 Zinc-lead smelting, 799
Waterproofing, by electrolysis, 290 Zirconia refractories, 515
principles of, 784 Zymase, 587

Handbook of Chemical Engineering Vol II

Uploaded by

Copyright:

Available Formats

Handbook of Chemical Engineering Vol II

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handbook of Chemical Engineering Vol II

Uploaded by

Copyright:

Available Formats

HANDBOOK

Electrical World ^ Engineering News-Record

DONALD M. LIDDELL, Editor-in-Chief

McGRAW-HILL BOOK COMPANY, Inc.

PRINTED IN THE UNITED STATES OF AMERICA

THE MAPLE PRESS - YORK PA

HIGH TEMPERATURE PRODUCTION

A Modem Art.—One of the most distinctive features of modern metallurgy

method, which is commercially applied to melting tantalum into buttons or ingots

ignition point; (2) such an accelerated surface combustion is dependent upon an

treated as to form a stable fuel capable of being atomized and burned in a

2,500°C. in an atmosphere absolutely free from carbon monoxide or hydrocarbons.

Fig. 2.—Girod. Fig. 3.—Stassano. Fig. 4.—Rennerfelt. Fig. 5.—Heroult.

fclCCYRlCAL — ■.—CRUCIBLE LINING

IOX-FOR, MECHANICAL RlCIDITV

\ / I SUPPORTS THE ELECTRIC^RUClBLE AND PROVIDES I \ /

Fig. 7.—Crucible induction-furnace.

MIXING AND KNEADING

In order to produce in a machine several mixtures of exactly the same comixnsdion

Figs. 4-5.—Paddle mixer.

of agitator is frequently used, as mixing of different minerals, cereals, chemicals, lumps

Figs. 6-9.—Mixers with scrapers.

Fioh. 10 13. Oblique mixing-dim* machine.

Fig. IK. ~ Single-blade mixer. Fiu. 10.--Two-blado mixer.

Fig. 18.—Single-blade mixer. Fig. 19.—Two-blade mixer.

Figs. 20-21. Fig. 22.

movements caused by the modifications interfere with, derange, counteract or deflect

Km. 'Jit. M ixing \\hiM*t,

For accelerating the mixing action it is sometimes advantageous to supply the

Fig. 25.—Simple agitator with direct fire.

Fig. 26.—Vacuum pan. Fig. 27.—Double-motion agitator.

Fig. 29. Fig. 30.

Fig. 31.—Double-motion kettle.

Fig. 32.—Double mixer.

1 Made by J. H. Day Co., Cincinnati, Ohio.

Figs. 33-34.—Corby dough-mixer.

Figs. 35-36.—High-speed mixer.

Fig. 40.—Typical cake-mixing blades.

Definition.—Sampling is the process of obtaining a small quantity of material

Owing to the difficulties of getting a small sample of a few ounces in weight to

In the sampling of solid materials we come immediately upon two classes—

Cutting Down Samples.—Having obtained the original sample it is then sub¬

Fig. 3.—Assayers riffle.

Fig. 4.—Brunton sampler.

rejected portion leaving an undue proportion of fines in the sample. This is a

Fig. 5.—Vezin. sampler. Fig. 6.—Snyder automatic ore sampler.

Coal Sampling.—Probably coal sampling is subject to greater proportional

Size of Slate Contained in Coal, and Size of Original Sample Required to

Weight of largest piece of Original sample should

Weight of sample Weight of sample

Limit Beyond Which Samples Should not be Divided when Crushed to

Size of coal mesh Should not be divided to less than, grams

l T51. d ■Ra.ilfiv. Jour. Ind. Ena. Chem. . 1909. r>. 161.

In order to obtain a sample which shall correctly represent any material it is

Weights to be Taken in Sampling Ore1

Weights Diameters of largest particle

Very low Low Medium ores Rich Rich and

20,000.0 207.0 114.0 76.2 50.8 31.6 5.4

i Richards: “Ore Dressing,” Vol. 2.

|fii|«#t ao* r«4»gi4iiri| tw n preserving action on l«crf nunh* Umn