Acetaldehyde by Dehydrogenation

of Ethyl Alcohol
JAMES RI. CHURCH . ~ K DHANAMANT IC. JOSHP
Chemical Engineering D e p a r t m e n t , Columbia University, R'ew Y o r k , IT. Y .

I n recent years, increasing amounts of acetaldehyde copper catalyst, containing ?IC/, cobalt and 27" chromium
have been produced by the air oxidation of ethyl alcohol as promoters, a t 275' to 300' C.,conversions as high as 959'0
and now nearly equal the production from acetylene by with yields of 8870 acetaldehyde and 10% ethyl acetate are
hydration. Both processes result in a highly diluted gas obtained. The hydrogen by-product recovered is 99% pure
reaction mixture involving a difficult recovery of acetalde- and over-all losses are less than %%of the alcohol consumed.
hyde, in one instance from the nitrogen of the air, un- A sustained activity of the catalyst is obtained over long
reacted alcohol, and other oxidation products and in the periods of performance with little loss in efficiency by
other, from large amounts of unreacted acetylene. In periodic reactivations.
contrast, the dehydrogenation of ethyl alcohol would The proposed dehydrogenation method offers several ad-
simplify the problem of recovery, since only acetaldehyde vantages over present processes for the manufacture of
and hydrogen are obtained as the products from the reac- acetaldehyde-namely, a reaction mixture rich in product
tion. A study of suitable dehydrogenation catalysts and for a simple recovery; a by-product hydrogen of high pu-
conditions has demonstrated the practical advantages of rity; ethyl acetate as a concomitant product in low yield;
such a dehydrogenation process. and a combined yield of 98% of valuable products with high
T h e n ethyl alcohol vapor is passed over a snpported conversions of ethyl alcohol per pass.

F OR many years acetaldrhydc has bzen one of the most im-

portant intermediate aliphatic chemicals serving as the rav
material for the production of acetic acid, acetic anhydride, ethyl
and costly operation involving the absorption of the acetaldehyde
in a dilute alcohol solution a t low temperature followed by distil-
lation of the absorption solution for final separation and purifica-
acetate, butyraldehyde, n-butanol, pentaerythritol, and manj tion. In contrast, the dehydrogenation of ethyl alcohol results in
other products. I t s production in this country since 1930 has in- 1:1 mixture with the by-product hydrogen, and separation and
creased nearly 100-fold, with a reported production in 1948 ( 3 2 )of recovery of the product is simplified by low temperature conden-
425,000,000 pounds. The major part of the annual production sation of most of the acetaldehyde and scrubbing the hydrogen gas
of acetaldehyde is derived from acetylene, although in recent in dilute alcohol for remaining traces of the product.
years the air oxidation of ethyl alcohol has produced a consider- Acetaldehyde was first prepared from ethyl alcohol by Davy
able amount of acetaldehyde, n ith nearly 200,000,000 pounds es- ( 7 ) , who employed a platinum black catalyst and passed a mix-
timated from this source during 1949 (99);it is also obtained i n ture of alcohol and air over it a t a low temperature. More thor-
smaller amounts as one of the products from an air oxidation of ough studies of the air oxidation of ethyl alcohol, employing vari-
lower hydrocarbons ( 2 4 ) . The reactions involved in the first two OUR catalysts, were made by Xuhlman ( I S ) , Strecker (SO), and
of these commercial processes are shoxn for comparison with thc Grimax ( 1 2 ) . Not until the work of Trillatt ( S I ) and Orloff ($2)
reaction of the dehydrogenation process under investigation in in 1902, however, were any quantitative data reported. Nore
this study: recently Faith and Keyes ( I O ) carried out extensive investigations
of this reaction and found that silver gauze was the beet catalyst.
1. Hydration of Acetylene (Commercial Process) Employing a flow rate of 0.57 liter of alcohol vapor per minute
80"/15 lb./sq. inch over a heated silver gauze, 12 mm. in diameter and 15 mm. long,
CH=CH + HzO + CHaCHO maintaining the temperature fairly constant at 515' C., they ob-
(as *o% HzS04) Hg '"
~

(ZSS)
(8% in C2n2i tained an 80.6% conversion of alcohol to acetaldehyde with losses
2 . Oxidation of Ethyl AIcohol (Commercial Process) of 13.3y0to carbon dioxide and 3.2y0 t o acetic acid. Shreve (28)
and Faith, Keyes, and Clark (11) report slightly higher yields
500" c.
2CIJaCH20H
(NZS)
+ '/z02 +2K,
(Air)
- +CH3CHO
Ag Catalyst
+ of 85 to 95y0 by air oxidation of ethyl alcohol to acetaldehyde by
use of a silver catalyst a t a temperature of 550 to 570' C. and a
CaHsOH + 2Nz + HzO conversion per pass of 50 t o 55%. However, Groggins ( I $ ) , re-

-
porting results from a similar process, claims only 70 t o 80% of
3. Dehydrogenation of Ethyl Alcohol (Proposed Process) the theoretical yield of acetaldehyde under identical conditions.
300 C. Based on an average of &heresults claimed by the oxidation proo-
CHaCHzOH
(Slight 2's) cu-Co Catalyst
CH3CHO + HB ess-namely, 50% conversion of alcohol per pass and over-all
theoretical yield of 85% of acetaldehyde-this process possesses
These equations show that the present commercial processes certain inherent disadvantages compared with a similar process
involve a recovery of the product from a diluted gas mixture. employing dehydrogenation instead of air oxidation. First, the
I n the hydration of acetylene the acetaldehyde product is mixed presence of the water by-product, formed by removal of hydrogen
with a n excess of unreacted acetylene in a ratio of 1:12, whereas by oxidation, with appreciable quantities of other oxidation prod-
in the oxidation of ethyl alcohol the product is diluted with nitro- ucts such as acetic and formic acids gives rise to serious corrosion
gen, water, and excess of alcohol in a ratio of 1:4. Recovery of problems in condensation and recovery of the products. Sec-
the acetaldehyde product in both these processes entails a difficult ondly, the use of air as an inexpensive oxidizing agent results in
1 The National Chemical Laboratory, Darrtwade Bldg., Poons 4, India. considerable dilution of the gaseous products from the reaction be-
1804
August 1951 INDUSTRIAL AND ENGINEERING CHEMISTRY 1805

cause of large amounts of nitrogen involved. Over half the prod-

ucts from the oxidation reaction are noncondensable (mainly ni- TABLEI. EQUILIBRIUM
\‘ALUES FOR DEHYDRATION
O F ETHYL
trogen with some carbon dioxide), making separation of the acet- ALCOHOL
Parks and Beyer,
Anderson,
and
aldehyde product fairly difficult. Lastly, conversion of alcohol t o
Temp., Rideal (86) Huffman Watson
acetaldehyde by oxidation cannot be expected to proceed without OC. Exptl. Calod. Calod. (8;) Calod. (i)
appreciable losses by further oxidation, resulting in a lower yield yo Conversion
of aldehyde. Therefore, an investigation of the possibilities of 150 3.9 89 5.0 23.5
200 10.6 98.3 11.5 45.5
employing dehydrogenation for converting ethyl alcohol to acet- 250 38 0 99.2 21.0 70.0
275 60.2 99 7 25 2 78.5
aldehyde wasbelieved justified, as in theory a t least, it was free of 300 .. 33.5 85 5
the disadvantages of the oxidation process. In addition, ad-
vantages such as the valuable hydrogen and ethyl acetate by-
products favor the dehydrogenation reaction for commercial pro-
duction of acetaldehyde from ethyl alcohol. ficiency. Many of these, particularly in the form of their basic
salts, caused decomposition and condensation of the aldehyde
THEORY OF DEHYDROGENATING ALCOHOLS product; the increasing order of this tendency was indicated as
manganese, iron, cobalt, and nickel. Therefore from the stand-
Dehydrogenation of ethyl alcohol to form acetaldehyde is an
point of alcohol conversion for a high yield of aldehyde product,
equilibrium reaction which is complicated, under certain condi-
these catalysts would not be desirable.
tions, to permit further condensation of the acetaldehyde product,
Extensive work has been done on the catalytic activity of cop-
forming such by-products as hydroxybutyraldehyde (aldol),
per in the form of pure reduced metal, mixtures of reduced metal
ethyl acetate, and subsequent hydrogenated products due to the
with other materials, alloys with other metals, and reduced salts
presence of the by-product hydrogen. Under favorable condi-
supported on inert materials. The first of these experiments with
tions for the dehydrogenation reaction, however, these side reac-
reduced copper was carried out by Sabatier and Senderens ( W ) ,
tions may be minimized and a high yield of the acetaldehyde ob-
who found that in a finely reduced state, copper was outstanding
tained. The theory of the dehydrogenation of alcohols has been
among the various metal catalysts studied. Palmer ( 2 3 ) like-
reported ( 4 , Z d ) as proceeding through the formation of a complex
wise, from a study of the effect of reducing conditions, found cop-
between the alcohol and the catalyst:
per metal particularly effective, with a maximum activity when
H H prepared by reduction a t low temperature. Reactivation of cop-
I
R-C-H + 2M + R-~-H--LI per catalysts through successive oxidation and reduction also was
reported ( 2 4 ) as showing some increase in catalytic activity, be-
I I coming constant after the third reactivation. The preparation
0-H Ob--H-11
of refined copper metal from its various mlts as reported by Neish
From various studies involving the catalytic dehydrogenation (2f) indicated that the catalyst obtained by reduction of copper
of alcohols, the following statements will suffice to summarize the oxide, previously heated to 1000’ C., had the highest activity.
generally accepted theories concerning this reaction: In mixtures of reduced copper with other materials, such as
1. The reaction follows a meudo first-order reaction rate oxides or carbonates of sodium, iron, zinc, manganese, and mag-
equation (21 ). nesium, Palmer (2325)reported a decrease in catalytic activity in
2. The energy of activation for the dehydrogenation is con- most cases, even with small additions of 0.5% of these ma-
stant for a given catalyst but varies with the type and method of terials. The addition of oxides of manganese and magnesium
preparation-of the same (df,23). improved the catalytic activity of the copper initially, but the life
3. The reaction rate is dependent on the temperature as
stipulated by the hypothesis of Arrhenius and is consistent of these mixed catalysts was short. No references could be found
with the constancy of the energy of activation (%f,25). in the literature concerning the use of mixed copper catalysts in
4. The alcshol molecules are first adsorbed on the surface of the form of copper alloys for the dehydrogenation of ethyl alcohol.
the catalyst, with the hydroxyl group parallel to the surface and However, copper alloys are frequently mentioned as those used in
the rest of the molecule perpendicular to it (24).
5. The reaction velocity is independent of the pressure on the commercial processes for the production of ketones by the dehy-
system (24), drogenation of secondary alcohols, such asisopropyl, sec-buty1,and
sec-hexyl alcohols. Various types of brasses and bronzes, con-
The maximum conversion of alcohol to aldehyde is of course de-
taining small amounts of tin, cadmium, chromium, nickel, and
pendent on the equilibrium of the reaction. Various workers
magnesium, are claimed as efficient sturdy commercial catalysts
(I, 25, 86) have studied the equilibrium as represented by the
for dehydrogenation purposes (33). One report (22) did include
simple dehydrogenation equation:
the use of a codeposit of copper with a small amount of nickel for
CzHi.OH(g) CH3CHOk) + Hz(g) the dehydrogenation of ethyl alcohol, but the yields were low,
supposedly because of decomposition of the acetaldehyde formed.
They report values for the per cent conversion of the alcohol Supporting materials increase both the life and activity of the
from experimental data and calculated from heats of combustion catalyst owing to the increase in active surface area and minimi-
and specific heat data. These results of the various equilibrium zation of local overheating by increase in thermal conductance
values, expressed in per cent conversion of ethyl alcphol to acetal- of the support. Lorang (19) used both pumice and dolomite as
dehyde at different temperatures, show little agreement between the supports for a copper catalyst prepared by a reduction of the
the experimental values and those calculated from fundamental decomposition product from a deposition of the nitrate salt. He
data, as shown in Table I. claimed a constant activity over a long period of dehydrogena-
Of the various catalysts suggested for dehydrogenating alco- tion, especially in a combination using traces of oxides of zinc,
hols, copper appears to be the most common, by itself and in com- cerium, thorium, and zirconium as promoter catalysts. Dunbar
bination with other catalytic materials. It appears to be fairly (9) also used pumice and diatomaceous earth as supports for cop-
ideal for this reaction in that it is specific and of high activity, but per using chromium as the promoter with considerable success in
it does lack stability, losing its activity quite readily (8). Many the dehydrogenation of higher alcohols. I n contrast with un-
investigators have also reported the use of catalysts other than supported copper catalysts for the dehydrogenation of ethyl alco-
copper ( 2 , 5, 15, 17, 20, 27). Among these catalytic materials hol, which Constable ( 6 )reported as of high activity but of short
were compounds of iron, cobalt, nickel, manganese, and zinc, life, Young (34) claimed a long life high activity catalyst using a
somewhat in the order of their increasing dehydrogenation ef- silica catalyst support with a copper chromium combination,
1806 INDUSTRIAL AND E N G I N E E R I N G CHEMISTRY v0i. 43, N ~ a.

Likewise Johnson ( 1 6 )found that copper carbonate with traces of tor itself consisted of a U-tube with an enlarged leg, 20 mni. in di-
zinc and magnesium oxides supported on alumina gel catalyst ameter, housing the 6-inch catalyst bed supported by plugs of
gave high conversions over a long period of time, although it glass wool. A traveling thermocouple was placed in a tube run-
resulted in the formation of higher condensation products. ning through the center of the catalyst bed and supported by a
stopper in the reactor head. Larger 1000-ml. gas absorption bot-
APPAR4TUS
tles were employed for complete removal of the condensable prod-
ucts and a large calibrated gas holder collected noncondensable
Figure 1 shows the apparatus designed for determining the per- products.
formance characteristics of the more active catalysts. A slightly
PREPARATION O F CATALYSTS
different type of apparatus was employed in preliminary evalua-
tion tests for determining the relative efficiencies of various cata- Previous studies on dehydrogenating alcohols had shovvn cop-
lysts; i t consisted of a constant feed buret for charging alcohol to per to be the best of the catalyst materials and it was rhosen as
the primary catalyst material, alone and in com-
bination with other elements. Various types of
copper catalysts were prepared by a numbcr of
methods from different salts of copper with and
without cocatalysts and support materials. The
catalysts were divided into two groups-series A
Gar Holder without catalyst support and series B with
catalyst support. The catalyst compositions are
given in Tables I1 and 111.
One method of preparation, whirh was ustd
Scrubbers mainly for the A series catalysts, consisted of
suspending 25 grams of cupric oxide in a solution
of 50 grams of cupric nitrate [ C U ( N O ~3Hz0]
)~
dissolved in 1 liter of distilled water. A molar
equivalent of 28% ammonia was added for pre-
cipitation of the cupric hydroxide. The combined
precipitates were then washed thoroughly, mixed
with a small amount (2 to 3%) of animal hide
glue while wet, slowly dried on a steam bath,
and granulated by stirring. Finally the granules
were decomposed by heating over a gas flame for
a. few minutes and Wen were sieved for proper
size. In the mixed catalysts the other materials,
as nitrates or oxides. were added along with the
cupric salts prior to precipitation. A d g h t modi-
Figure 1. Dehydrogenation Apparatus fication included the use of ammonium carbonate
in place of ammonium hydroxide for the precipi-
tation of the mixed carbonates.
a vaporizer and a preheater, made of a horizontal 15-mm. diame- 111 another method, employed chiefly for the preparation of
ter heat-resistant glass tube filled with glass wool and heated in an some of the supported catalysts, a melt or strongly concentrated
electric furnace to the desired temperature. The preheated alco-
hol vapors were then passed to a vertical 20-mm. diameter heat-
resistant glass tube containing a &inch bed of catalyst and heated
by means of an electric furnace. The temperatures of the incom- TARLE11. COMPOSlTION O F UNSUPPORTED COPPER CATALYSTS
ing vapors and the catalyst were measured by suitable thermo- Catalyst Promoter
XO. Copper Salt Catalysta Mesh Siaeb
couples placed in the vapor lines and the catalyst bed. The reac-
None 8-20
tion products leaving the top of the reactor were chilled in an ice Oxide from pptd. hydroxide 5 % CrzOs 8-20
trap where the major part of the alcohol and aldehyde were con- A-3 5 % COO 8-20
-1-4 Metallic screen None No. 100 gauze
densed; then any uncondensed vapors were absorbed in two wash ZnO '1/16-a/ie
7% BIgO
Oxide from PPtd. carbonate 5% /IWa/iS inch spheres
inch spheres
bottles containing water, and the residual gases were finally col-
lected in a n aspirator bottle for measurement. a Based on weight of original copper oxide content.
b Tyler standard mesh.
I n the performance tests the dehydrogenation apparatus waA
changed in order to determine the activity of the catalysts over a
longer period of operation. Automatic controls were required in
order t o maintain constant temperatures, uniform flow of react- TABLE 111. COMPOSITION O F SCPPORTED COPPER CATALYSTS
ant, and collection of products without interruption. In this ap- Catalyst Support
NO. Copper Salt Materiala Promoter*
paratus a larger graduated 5-liter bottle for charging larger quan-
Long fiber asbestos None
tities of alcohol replaced the buret. I n place of electric furnace g:;]
8-31
Nitrate hydrate melt Pumice
Alumina gel
1 . 6 2 % CnOa
%-one
heaters the reactor was heated in a mixed salt bath composed of B-4 Bmmonia solution of
equal weight quantities of sodium and potassium nitrates con- nitrate Pumice 5 7 COO
Pumice 5d M g O
tained in an electric well heater. The salt bath temperature was '-5) >'ormate solution Pumice 5% c o o
controlled b y use of a bimetallic thermoregulator placed in series Short fiber asbestos 5 . 2 y COO
Alundum 5 , 2 9 COO
with the heating element of the well heater. An electrical con-
Kitrate solution
Short fiber asbestos 5 . 2 9 COO
1 . 8 % CrzO:
+
denser in parallel with the bimetallic regulator prevented sparking B-10 Short fiber asbestos 5 . 2 % COO
4 4 % cr90.
+
for better operation of the thermostat in maintaining a constant B-11 Short fiber asbestos Kana '"
temperature within 13' C . A large cold trap, chilled in a dry-
ice bath, condensed most of the products; the head of the reactor a Mesh size: Pumice 1/8 to l/a inch spheres; alumina 8-14 mesh; asbestos
8-20 mesh.
was designed to prevent any products partially condensed in the b Amount based on weight of original copper oxide content of catalyst.
upper part of the tube from returning to the reactor. The reac-
 August 1951 INDUSTRIAL AND ENGINEERING CHEMISTRY 1807

solution of the cupric nitrate salt was used. The support mate- between this amount and the equivalent amount of the liquid
rial, either acid-washed asbestos fiber, pumice, or alumina gran- products obtained was regarded as unaccountable loss. Thus
ules, was added to the hot melt or solution and stirred for com-
plete impregnation. The solution mixture was then strongly yields could be calculated by the following expression:
heated over a flame for subsequent evaporation of water and
partial decomposition of the salts t o form a granular solid mass Total % yield =
moles aldehyde +
ester
x 100
consisting of the support material thoroughly coated with the moles of Hz
catalytic materials. I n other catalysts the freshly precipitated
cupric hydroxide, with or without coprecipitated hydroxides of Likewise the percentage conversion of the alcohol was calcu-
the promoter materials, was first dissolved in a sufficient amount lated from the amount of alcohol consumed in terms of equivalent
of 28% ammonia solution and the support material thoroughly
saturated with the solution before evaporation of the mixture to amounts of hydrogen produced. In the runs for determining the
dryness. This resulted in a fairly even coating of the catalyst efficiency of the catalyst, the change in catalyst activity was de-
support with thd oxides. In a modification of this procedure noted in terms of the percentage alcohol reacted during a given pe-
cupric formate solution was employed in place of the cupric riod of time, as measured by the volume of gaseous product col-
ammonia solution in similar fashion, and in another a suspension
of the precipitated carbonates was used in place of the hydroxides. lected. Analysis of samples of gaseous and liquid products a t
Of all the methods tried, the best catalysts were prepared by short intervals during the run was sufficient to detect any change
impregnation of the support solution in hot 25% cupric nitrftte in the catalyst activity. Reactivation of the catalyst was read-
solution, with and without added promoter salts, by a long period ily achieved by stopping the alcohol vapor and passing air over
of soaking followed by drying. The impregnated support mate-
rial was finally heated a t elevated temperatures approaching the heated catalyst while maintaining the catalyst bed tempera-
400' C. in order to decompose the nitrate to form an oxide coat- ture below 400 ' C . for an hour or so to convert the metallic cata-
ing on the support material. Repeated impregnations, totaling lyst t o the corresponding oxide mixture and then reducing again
six to eight, gave a heavy evenly coated catalyst of maximum with alcohol vapor with the temperature maintained below 300"
activity and long life. In all catalyst preparations, reagent grade
chemicals were used. C. as described previously.
PROCEDURES The products collected from the dehydrogenation consisted
The catalysts, prepared in the form of their oxides were charged mainly of unreacted ethyl alcohol, acetaldehyde, ethyl acetate,
acetic acid, and hydrogen. The gaseous hydrogen by-product
into the reactor tube and heated for a period of 2 or more hours in was analyzed in an Orsat apparatus by the usual standard gas
a current of dry filtered air t o ensure complete decomposition of analysis method. The aqueous solution of the liquid products,
the last traces of other salts. The temperature was then lowered prepared from the condensate and absorber solutions was ana-
t o a little below 300' C. and a slow stream of alcohol vapor pro- lyzed for free acid by a standard alkali titration of an ali uot
portion. Another sample was refluxed with an excess of stan8ard
duced from 95% ethyl alcohol m-as passed over the catalyst for alkali for 1 hour during which time the ester product was hydro-
reduction of the oxides. Care was exercised during this reduction lyzed, but subsequent condensation of the aldehyde product,
t o avoid overheating of the catalyst as the reaction is exothermic. which forms a colored solution, prevents a direct back titration
Following reduction of the catalyst, noted by the absence of of the alkali. Instead the solution was treated with excess stand-
ard acid; heating was continued to destroy the color; and the
moisture in the exit gases, the temperature of the catalyst bed excess acid was back-titrated with standard alkali. Thus the
was adjusted to the desired level, and the rate of alcohol vapor amount of ester present could be calculated from the volume of
was increased to a maximum for an optimum conversion and standard alkali consumed.
yield of the dehydrogenated products. During the run, which Acetaldehyde was determined by adding an excess of standard
laatcd for several hours, the temperature was maintained within bisulfite solution t o an aliquot sample of the product solution
and allowing the mixture t o stand, with frequent shakings, a t
+ 2 O C. of the desired temperature, and the products were col- room temperature for 2.5 hours according to the standard proce-
lected and measured for determining the efficiency of the catalyst dure for acetaldeh de (g). The sample was then made up to a
a t any time during the run. At the end of the run, the total vol- known volume wit{ distilled water, and the excess bisulfite was
ume of liquid products in the condensate trap, those absorbed in determined iodimetrically by titration of an aliquot with excess
standard iodine solution in the cold, followed by back titration
, the scrubber solution, and the volume of noncondensable gas in with standard sodium thiosulfate using starch-iodide as the indi-
the gas holder were noted. The condensed and absorbed prod- cator. The aldehyde-sulfite addition compound does not react
ucts were combined and made up to a known volume with dis- with the iodine solution, and therefore the amount of aldehyde
tilled water. Aliquots of this product solution were analyzed for could be com uted from the amount of bisulfite consumed in
forming the afdition compound.
alcohol, aldehyde, acid, and ester content according to the meth-
ods described in the following section. DISCUSSION OF RESULTS
The amount of alcohol reacted was equivalent to the hydrogen
content of the gaseous product, and therefore the percentage con- A total of some fifty runs with the unsupported and supported
versions and yields were computed on this basis. The difference catalysts was made t o determine the efficiency of the various
catalysts under varying condi-
tions of temperature and space
velocity for a maximum con-
TABLE IV. PERFORMANCE OF CATALYSTS IN DEHYDROGENATION OF ETHYL ALCOHOL version of alcohol and yield of
Space %con- Compn. of Liquid Productsb, H~ aldehyde. For results from
Run
SO. Catalyst
Temp Veloc-
0 c." itya
yer-
sion
_I_ %
ACH ~ct-40 HAO xc
Gas,
%
the better runs, representing
the optimum conditions for the
NOXSUPPORTED CATALYSTS catalysts under examination,
A-13 CuO-Cu(OH)z from nitrate 335 8.6 65 90 R 0.2 08 99.2
1-33 CuO-Cu(OH)z.from nitrat? + 5% COO 330 15.7 57 88 7 0 3 4.7 93.6 see Table IV. In addition, a
.L52 CuCOr from nitrate + 7% ZnO number of continuous perform-
A-62 CuCOi from nitrate + 5% MgO
335 13 5 80 78 9 04 13.1 99.3
332 14.1 78 78 8 07 13.3 99.0
ance runs, lasting from 50 to
SUPPORTED CATALYSTS 100 hours, were made with the
B-12 CuNOs melt (asbestos) 330 14 0 85 88 10 0.3 1.7 99
3-42 CuNHs + 5% Co (ppmtce) 325 25 0 68 83 10 09 6 1 95 better catalysts. Changes in
B-62 Cu For 5% Co (pumice) 330 17.7 67 84 10 0.6 5 4 95 activity, in terms of initial
3-73 CuNOl 5% Co (asbestos) 337 17.2 94 89 7 06 3.4 96
B-93 CuN03 5% Co, Z% Cr (asbestos) 330 18.0 93 83 9 1.0 7.0 95 activity, are shown in Table V.
B-113 CuN03 Soln. (asbestos) 328 18.8 79 90 8 09 1 2 99
Lastly, a 100-hour run was
Co. alcohol vapor/cc. catalyst bed/minute. m a d e w i t h o n e of the best
b Theoretical yield based on alcohol reacted.
0 Mainly high boiling by-products: also include6 losses. catalysts with reactivation at
24-hour intervals (Table VI).
1808 INDUSTRIAL AND E N G I N E E R I N G CHEMISTRY Vol. 43, No. 8

Important. A coating of as-

TABLE
V. PERFORMANCE
CONTINUOUS OF DEHYDROGENATION
CATALYSTS bestos fibers from a mclt of
Catalyst No. 8-51 B-41 B-51 B-71 B-91 B-92 B-111 copper salt, such as the cata-
CHARACTERISTICS lyst used in run B-12, u a s less
Original Cu salt Carbonate KHr Formate Nitrate Nitrate Sitrate Nitiate effective than the catalyst of
Comples
Promoter, % 7% ZnO 5% COO 3% bIgO 5 % COO 5 7 COO 5 7 COO None run 113 prepared from a solu-
2 9 crzda 2 4 c r d a
Support None Pumice Pumice Asbestos Asgestos Asbestos Asbestoq tion of the salt. Little differ-
Reaction temp., C. 275 275 275 2 75 275 300 275 ence, however, was noted in the
Initial conversion EtOH, % 35 44 68 55 50 72 35 activity of catalysts prepared
Yield AcHa, % 81 85 74 90 85 84 89
Yield EtOAc," % 8.7 10.0 11.3 8.7 9.6 9.3 0.2 from various soluble salts; this
CONTISUOUS BCTIYITY, % O F I N I T I A L ACTIVITYb \+as probably because all were
Hours
finally converted to correspond-
2 99 97 96 100 98 96 95
4 97 93 89 100 96 95 92 ing oxides and therefore were
6 96 85 80 98 94 93 89
8 93 79 69 97 92 91 85 fairly identical in chemical
10 84 73 61 95 91 90 80 composition in the active state.
12 77 64 52 94 90 89 79
24 41 43 36 85 88 87 73 The mixed catalysts, which in-
30 34 39 32 84 87 86 72
36 30 36 29 82 87 85 71 cluded a promoter material,
48 26 32 23 80 86 84 70
54 24 32 End 79 86 83 68 behaved quite differently from
60 22 End 78 85 82 66 those containing only copper,
72 End 75 84 80 65
84 73 84 78 63 Additions of cobalt as a pro-
96 70 83 72 61 moter greatly increased ac-
100 70 83 71 60
tivity and resulted in higher
Theoretical yield, based on alcohol reacted, remained unchanged with decreasing catalyst activity.
b Time required t o collect 800 c c . Hz gas. (*onversions of alcohol, as IS
qhown by a comparison of run
B-73 and run B-113, both made
TABLE VI. CONTINUOUS P E R F O R N A N C E WITH INTERMITTENT under almost identical condi-
REACTIVATION O F SPECIAL DEHYDROGENATION CATllLYST
tions. Other promoter materials did not seem to affect catalyst
Catalyst compositjon (B-Q): Cu from cupric nitrate solution + 5.2% COO
+ 1.8% CrzOa, impregnated on short fiber asbestos as oxides
Conditions of run:
activity much but did alter the ra,t'ioof aldehyde yield to other by-
Alcohol feed, 40 cc./hour; catalyst 16.2 grams (38.9 cc.) products. The catalysts of runs .4-52 and A-62, which contained
Temperature, 275' C * initial conversion EtOH, 49.6% alkaline oxides of zinc and magnesium, caused considerable by-
Space velocity (cc. alcbhol vapor/cc. catalyst bedlminute), 16.2
Initial activity: product formation. Undoubtedly this was due to their basic
Conversion of E t O H , 49.6% character, which favored the conversion of the normal aldehyde
Yield aldehyde, 85.1%; yield ester, 9.670p
Continuous activity: % initial activity, time required t o collect 800 cc. product to higher condensat,ion product's. Similar result8 were
H2 gas
noted in runs wit'h catalyst R-3, empIoying alumina as the catalyst
Hours % Hours % support; these indicated that B-3 also behaved as a condensation
0 100 55 catalyst. As it was desirable in thip study to minimize the forma-
3 96 58
6 93 61 tion of such by-products in order to increase the yield of aldehyde,
9 92 64
13 89 67 such alkaline materials were not, considered suitable constituents
22 87 77 for dehydrogenation catalysts. The pure copper catalysts of
28 86 (Reactivated 1 hour) 80
29 105 81 105 runs A-13 and B-113 gave the highest yields of aldehydes with a
31 101 83 102 minimum of by-products. !?%en cobalt was added to the copper,
34 98 85 99
38 96 88 97 an increase in dehydrogenation occurred with only slight louses
40 93 92 94 in yield of acetaldehyde due to decomposition to gaseous by-
51 90 (Reactivated 1 hour) 102 91
54 88 105 88 (End of run) products. However, additions of chromic oxide to the copper
a 7" yield, based on alcohol reacted, remained unchanged throughout run.
catalyst caused some dehydration of' ethyl alcohol thus producing
considerable amount's of ethylene in the gaseous by-products.
The main by-product, ethyl acetate, was Eormed in about the
Catalyst Activity. The results obtained with the various same amounts with the various catalyst's, and therefore its forma-
catalysts show considerable variation in activity in terms of thc tion is probably not' a function of t'he cat'alyst but rather a vapor-
percentage conversion of alcohol by dehydrogenation as well as phase reaction of the Tischenko type.
in theoretical yield of aldehyde and other by-products. Table IV Stability of Catalysts. I n the performance tests with catalysts
gives the results for the supported copper catalyst (run R-113), under continuous operation, activity varied greatly as the de-
which was many times more effective than the copper catalyst hydrogenation reaction was continuod. Furthermore, catalyst
used in run A-52. I n general those catalysts prepared directly life, as determined in these runs, showed different results, in most
from soluble nitrate salts were more effective than insoluble oxidw inst,a,nces, than given in previous reports on the performance of
or carbonates; thus, the method of preparation contributes much similar type catalysts. Constable (6) found that copper cata-
to catalytic activity. Apparently catalyst activity is direct]?- lysts, prepared by various methods, all lost their activity rapidly
proportional to total active surface area which is much greater in a t temperatures above 280' C. with the best of the group losing as
freshly prepared catalysts. . 4 comparison of activities of sup- much as 70% of initial activity a t 300" C. within 2 hours. On
ported and nonsupported catalysts indicated generally that sup- the other hand, Johnson (16) claimed that copper catalysts con-
ported catalysts contained greater active surface area per volume taining magnesium and zinc oxides retained their activity for
of catalyst, although the weight of actual catalyst was consider- long times. However, the reeu1t)s from this study d e k i t e l y
ably less. The nature and type of support material likewise con- demonstrated that a sustained activity could be obtained from
tributes a major effect on catalytic activity-for example, the several of the copper catalysts at temperatures in the neighbor-
asbestos-supported catalysts which proved superior to the pumice- hood of 300" C. (Table V). In runs B-71, B-91, B-92, and 13-111,
supported catalysts, although the catalyst content was quite ail catalysts retained a high percentage of their initial activity
similar. The method of impregnation of the catalyst support over the 100 hours of the test, The reverse was true of the cata-
material for a uniform and even coating of the catalyst is highly l y s t , ~of runs A-51 and R-51 containing zinc and magnesium ox-
August 1951 INDUSTRIAL AND ENGINEERING CHEMISTRY 1809

other 25 hours showed another loss of

activity t o 88% of the initial activity.
Two further reactivations increased the

.-
activity to 107 and 105% of the initial
activity, with decreases to 8870 after 25
2
hours continued dehydrogenation in
, both cases. Thus a sustained activity
I1
v1
L
W
- "--- over an indefinite period, averaging
z 96% of the initial activity, can be ex-
0 pected by intermittent reactivation of
-I
the catalyst at 24-hour intervals.
5
t PROPOSED COMMERCIAL PROCESS
z
LL
The results of this investigation indi-
0
cate that a process for the manufacture
ae of acetaldehyde b y dehydrogenating
95% ethyl alcohol is not only commer-
cially feasible but that the dehydro-
genation reaction offers several ad-
25 50 76 100 vantages over existing commercial
TIME - HOURS processes employing air oxidation of
Figure 2. Continuous Performance of Catalysts ethyl aIcohol or hydration of acetylene.
Under the optimum conditions deter-
mined by this study, using a copper
ides, respectively; these were the poorest of the group, and lost catalyst prepared from copper nitrate solution impregnated on an
over 50% of their initial activity within 24 hours of continued de- asbestos support and containing 5% cobalt oxide and 2% chro-
hydrogenation. Again, the type of catalyst support affected the mium oxide, based on the weight of copper oxide equivalent, a
life of the catalyst, as was noted in initial catalytic activity in sustained conversion of about 90% of the alcohol a t 275" C. is
these performance tests. Those catalysts containing asbestos as obtained, resulting in 85% of the theoretical yield of acetalde-
the support proved to have the greatest stability in terms of con- hyde and 10% of the theoretical yield of ethyl acetate, thus rep-
tinued sustained activity, whereas those supported on pumice, as resenting less than 5% losses of alcohol in other by-products. I n
in runs B-41 and B-51, showed low catalyst life. The greatest terms of pounds of products this means 89 pounds of acetal-
contrast in this regard is shown in the results obtained in runs B- dehyde and 10pounds of ethyl acetate per 100 pounds of catalyst
41 and B-71; the catalyst was almost identical but where pumice each hour, an acceptable production capacity for a commercial
was used as the support in the former, asbestos was used in the unit.
latter. The changes in activity, in terms of percentage of initial
activity, were 43 and 8570, respectively, after 24 hours. At the
end of 48 hours the activities were 32% for the pumice as com-
pared to 80% for the asbestos support. The uniformity obtained
in impregnation and coating of the asbestos with the catalytic ma-
terial not only provides a larger active surface area but one less
susceptible to thermal deterioration on sustained heating a t higher
catalytic temperatures. The greatest sustained activity was ob-
tained with catalyst B-9 containing both cobalt and chromium
oxides. It retained over 80% of its initial activity after 100 hours
of dehydrogenation a t 275" C., as obtained in run B-91. With
cobalt alone as the promoter ingredient with copper, the activity
of catalyst B-7 after 100 hours (run B-71) at 275' C. was con- 70 I I I I
siderably le~s-707~as compared to 83% for the combined pro- 25 50 75 100

moter under identical conditions of run B-91. This is in accord TIME- HOURS

with the results reported (9, 19)from the use of chromium as Figure 3. Reactivity of Catalyst
a promoter with copper in the dehydrogenation of higher alco- B-91 = Cu-Co-Cr, 275O C.
hols. The effect of increased temperature on sustained activity
is shown in a comparison of runs B-91 and B-92. As might be
expected the loss of activity was greatest at the higher tempera- A flow sheet of the proposed commercial process is shown in
ture. The results of these continued performance tests are also Figure 4. The alcohol, which is a tax-free special denatured
shown in the curves of Figure 2. type (S.D.29) containing traces of acetaldehyde, is vaporized a t a
Reactivation of Catalyst. Surprising results were obtained in constant rate in an evaporator at 100" C. and preheated in a heat
the reactivation of some of the sturdier catalysts; these regained exchanger by the product gases t o 200" C. before it enters the
not only their initial activity but a slightly improved activity catalytic converters, where dehydrogenation takes place. The
following reactivation and sustained activity a t a higher level converters consist of a series of the conventional vertical tubular
for part of the continued dehydrogenation. This is shown in the type, packed with the supported copper-cobalt-chromium oxide
results obtained from catalyst B-9, the sturdiest and one of the catalyst combination, maintained a t 275" to 280" C. by means of
most active, given in Table VI and plotted in the graph of Figure a suitable heat transfer medium such as superheated steam, hot
3. At the end of 28 hours of continuous performance the cata- air, molten salts, or some high temperature fluid. A uniform
lyst had lost 14% of its initial activity, but after reactivation by a supply of heat is essential since the reaction proceeds with a cal-
slow oxidation of the catalyst with air for 1 hour at 350" to 400" culated absorption of about 600 B.t.u. per pound of alcohol re-
C., the activity was 105% of the initial activity. Continued de- acted, or 540 B.t.u. per pound charged with a 90% conversion as
hydrogenation with the reactivated catalyst over a period of an- expected. The products of reaction with unchanged alcohol
1810 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 43, No. 8

, 1 EtOH

are cooled after passage through the heat exchanger to -10" C., For continued uninterruptrd operation of the dehydrogenution
first with water and then wit,h a refrigerant in two tubular con- converter, where periodic reacthation of the catalyst is esseni8in.l,
densers to condense as much of 'the alcohol, acetaldehyde, ~ i i d devera1 catalyst units are desiralk so that reactivation of any of
ester as possible. The uncondensed gases are then sent to a di- t8hese can be made wit,hout, serious curtailment of production.
lute alcohol scrubber for renioval of the remaining alcohol, acetal- The reactivation can easily be made by shutting off the alcohol
dehyde, and other dissolvable vapors. The hydrogen product' vapor feed to the converter, flushing with steam to remove aloo-
collected from the scrubber may be used for hydrogenation pur- hol vapor, passing hot, air t'hrough t'he catalyst tubes while main-
poses directly since it is about 98% pure containing only traces of taining the temperature below 400" C. for about 1 hour, followed
unobjectionable by-products--methane, carbon monoxide, and by alcohol reduction a t 280" C. Careful control of the c a t a l p t
et,hylene. The alcohol-aldehyde condensate containing inorc. tt:mpera,ture during reactivation and the subsequent reduction
than 90% of the reaction products is sent to a fractionating col- are most essential, since both reactions are exothermic. Regu-
umn for separation of the lower boiling acetaldehyde a t 21' C.; lation of the amount, and ttmperature of the air during the first
the alcohol-ester bottoms are sent to a separate column for re- stage is sufficient for maintaining the desired maximum temp
covery of the alcohol a t 78 o C. Likewise the dilut,ealcohol scruli- ture. During the reduction stage, which is perhaps more c
her solution is first fractionated in a column for recovery of tlie cal since a low temperature of reduction is desirable for a high ac-
acetaldehyde product, the bottoms being sent to the alcohol re- tivit8yand long life of the catalyst, the amount and temperaturc
covery column. The ethyl acetate by-product, separated from of the alcohol vapor ca,n be regulat~rdfor a proper texnperxtiiit
the diluted ternary distillate of the alcohol recovery column, i, control.
purified in a separate column for recovery.
The dehydrogenation process is much simpler than the ethyl
alcohol oxidation process for acetaldehyde. The temperatures o f
the dehydrogenation convert,ers are only 280" C. as comparpd t,o Va,rious copper catalysts, supported arid unsupported, P O ~ oCf
a maximum temperature of 550" C. for the oxidation convclrtcrs: which contained promoter n1a.t Is, were thoroughly evaluatcd
hence fabricat'ion and maintenance of this part of the equipmcnt for their efficiency in the dehydmgenrtt>ionof ethyl alcohol for the
~vouldbe much easier. The endothermic nature of the dehydro- formation of acetaldehyde. Tliese copper catalysts werc pre-
genation process makes it less imperative to maintain 6he derired pared from the metal and several salts of copper, such as the car-
operating temperature, but heating costs are higher. The exo- bonate, formate, hydroxide, nitrate, oxide, and cupramrnoniuin
thermic reaction of t,he oxidat'ion process, requiring heat removed complexes. The added promoter materials included cobalt
for close control of the reaction temperature, is much more sensi- metal and oxides of cobalt, chromium, magnesium, and zinc in
tive to loss in yield by fluctuations in temperature due to over- varying amounts and in combinations. Supporter material con-
oxidation of the alcohol amounting to complete conibustion and sisted of alumina, a,lunduni, pumice, and asbestos; various
loss of starting material. Very small losses in yield and starting methods were used for impregnating thme with the cat,alj-tic
material occur with wide variations in temperature in the dehy- materials.
drogenation method. Under varying conditions of tcmpernt,ure and rate of passagc of
August 1951 INDUSTRIAL AND ENGINEERING CHEMISTRY 1811

the alcohol vapor, the maximum activity in terms of conversion D a w . H.. Phil. Trans.. 97. 4 5 (1817).
of the alcohol to acetaldehyde and other valuable by-products, DogOIor, B. N., and Koton, M.‘ M.,‘J. Gen. Chem. (U.S.S.R.),
6, 1444-51 (1935).
was determined. A majority of the catalysts were fairly specific; Dunbar, R. E., Cooper, D., and Cooper, R., J . Am. Chem. Soc.,
copper alone gave the highest yields of the aldehyde. A con- 58, 1053 (1936).
stant small amount of ethyl acetate was formed as the chief by- Faith, W. L., and Keyes, D. B., IND.ENG.CHEM.,23, 1250
product with all the catalysts. For a sturdier type catalyst with (1931).
Faith, Keyes, and Clark, “Industrial Chemicals,” New York,
fairly high activity, added cobalt and chromium as promoter J. Wiley & Sons, 1950.
agents were found beneficial; the former increased the percent- Grimax, E., Bull. SOC.Chim. (II), 45, 481 (1886).
age conversion of the alcohol and the latter contributed to a sus- Groggins, P. H., “Unit Processes in Organic Synthesis,” 3rd
tained high activity for a long continuous performance of the ed., p. 466, New York, McGraw-Hill Book Co., 1947.
Hightower, J. V., Chem. & Met. Eng., 55, 7-105 (1948).
catalyst. Ipatieff, V. N., Ber., 34, 3579, 9596 (1901).
From a practical viewpoint, the best catalyst was the one pre- Johnson, J. Y., Brit. Patent 331,883 (1930).
pared by impregnation of short-fibered asbestos with a solution of Knoevenagel, E., and Heckel, W., Ber., 36, 2816 (1903).
copper nitrate containing 5% cobalt oxide and 2% chromic oxide, Kuhlman, F., Ann., 29, 286 (1839).
Lorang, F. J. H., U. S. Patents 1,956,088 (1934) and 2,039,543
based on the weight of copper oxide equivalent. At a tempera- (1936).
ture of 275“ C. a maximum theoretical yield of 88% of acetalde- Marchand, R. F., J. Proct. Chem., 15, 7 (1838).
hyde and 9.6%.of ethyl acetate was obtained with a 93% conver- Neish, A. C., Can. J. Research, 23B, 4 9 (1945).
sion of the alcohol per pass. After 100 hours under these condi- Orloff, J. E., “Oxidation des Alcohols par L’action de Contact,”
Paris (1901).
tions the catalyst lost only 16% of its initial activity, and reac- Palmer, W. G., PTOC.Rou. Soe. (London). 98A, 13 (1920):
tivation b y a short air oxidation restored the activity to 105% of 99A, 412 (1921); IOIA, 175 (1922).
that initially present. Thus it was possible to obtain sustained Palmer. W. G.. and Constable. F. H.. Ibid.. 106A. 250: 107A
I ,

activity, averaging gay0 of the initial activity, over a long period 255, 270 (1925).
Parks, G. S., and Huffman, H. M., “Free Energies of Some Or-
by reactivations. ganic Compounds,” New York, Reinhold Pub. Corp., 1932
Based on the favorable results of this investigation, a potential Rideal. E. K., Proc. Roy. SOC.(London), 99A, 153 (1921).
process for the manufacture of acetaldehyde by the dehydrogena- Sabatier, P., and Senderens, J. B., Compt. rend., 136, 738, 921,
tion of 95% ethyl alcohol is proposed. 983 (1903).
Shreve, R. Norris, ”Chemical Process Industries,” 1st ed., p.
918, New York, McGraw-Hill Book Co., 1945.
LITERATURE CITED Steverson, H., Chem. Eng. News,27, 1182 (1949).
(1) Anderson, J. W., Beyer, G. H., and Watson, K. M., .Vatl. Strecker, A., Ann., 93, 370 (1855).
Petroleum News,36, R476 (1944). Trillatt, A,, Bull. SOC.Chim. (III),27, 96, 797 (1902).
(2) Armstrong, E. F., and Hilditch, T. P., Proc. 2 2 0 ~ SOC.
. (London), U. S. Tariff Commission, Washington 25, D. C., “Production
97A, 259 (1920). and Sales of Synthetic Organic Chemicals,” 1948.
(3) dssoc. Official Agricultural Chemists, “Tentative and Official Whitmore, F. C., “Organic Chemistry,” p. 243, New York, D.
Methods of Analysis,” p. 195, 1945. Van Nostrand Co., 1937.
(4) Baladin, A. A., 2. physik. Chem., B2, 289 (1929); B3, 167 Young, C. O., U. S. Patent 1,977,750 (1934).
(1929). RECEIVED
September 15, 1950. In partial fulfillment of the requirements for
(6) Berthelot, M., Les Carbures d’hydrogene, 2, 202 (1901). the Ph.D. degree granted in June 1949. Contribution No. 19 from the
(6) Constable, F. H., J . Chem. Soc., 1578, 2795 (1927). Chemical Engineering Labs., Columbia University,. New York, N. Y.

Equilibrium Relations of Two Methane-

Aromatic Binarv Svstems at 150° F. J J
M. ELBISHLAWI’ AND J. R. SPENCER
Texas Petroleum Research Committee, University of Texas, Austin, Tex.

Previous work on mixtures of methane and hydrocarbons

indicated a correlation between the chemical nature of the
sol\ent and the solubility of the methane, the chemical
B ECAUSE of their importance in predicting the phase be-
havior of hydrocarbon systems, vapor equilibrium data for
hydrocarbons have been extensively studied in the petroleum
nature being expressed by the Universal Oil Products char- industry over the past two decades. Such data are usually used
acterization factor (13). The purpose of this work was to in the form of vaporization equilibrium constants as these can be
test the correlation for low values of the U.O.P. charac- applied directly to problems involving the effect of temperature
terization factor and for systems containing aromatic and pressure on the compositions of liquid and vapor in two phase
hydrocarbons. systems. This constant is defined as
Equilibrium constants for the systems methane-benzene
and methane-toluene which were studied bear out the cor-
relation even to extreme ranges of the factors.
If, as these binary systems suggest, the solubility of
methane in crude oil-natural gas systems has a similar where y is the mole fraction of the component in the vapor phase,
dependence on the chemical character of the solvent, the and x, the mole fraction of the same component in the liquid
prediction of phase behavior in petroleum reservoirs would phase.
be greatly facilitated. This prediction is of great economic Simplified systems of two and three components have been
importance in the petroleum industry, especially during studied t o provide a basis for the consideration of complex prob-
the national emergency when oil and gas must be pio- lems of more practical interest and thus enable the production
cluced at maximum rates consistent without sacrifice of
ultiniate recovery. 1 Bureau of Minea, Cairo, Egypt.