I DONALD F.

OTHMER and MAHESH

Polytechnic Institute, Brooklyn 1, N. Y.
S. THAKAR'

Glycol Production
I
- Hydration of Ethylene Oxide
1 function, but in dilute solutions it ap-
proaches that given here.
A simple and continuous proc has been developed for
Acid Concentration. Straight lines
making glycol from ethylene ox using an ion exchange result (Figure 2) when reaction velocity
resin in a fluidized solid-liquid b constants a t constant temperature are
About 1 billion pounds of glycol is made annually from plotted against normality of the reacting
ethylene oxide, but usual processes have disadvantages- solution or against the hydrogen ion con-
at low temperatures and pressures sulfuric acid is used centration; in a dilute solution these two
are proportional, because of complete
as a catalyst, but it contaminates the product, compli- dissociation.
cates the purification process, and s corrosion. At high Temperature. Rate increased greatly
temperatures and pressures, 'no c t is needed, but the with temperature. Table I1 gives a
reaction is hazardous and carbonization lowers yield. constant ko a t each temperature for a
hydrogen ion concentration of unity
I I (pH = 0). Values of unit velocity con-
stants ko are the slopes of lines of Fig-
OVER HALF O F the ethylene oxide pro-
duced in this country is used for making
increases reaction velocity. Also errors
in measuring the time of sample analysis
ure 2. This assumes that velocity con-
stants in neutral solution are 4.legligible.
glycol (26)' usually from a 10 to 15% are larger near the start. Reaction velocity constants, plotted
aqueous solution, using sulfuric acid as a The first order reaction equation (Fig- (Figure 3,A) against vapor pressures of
catalyst. Yet, although glycol has been ure 1) is dc/dt = -kC, or on integration, water a t the same temperatures give a
known' since 1860 (@), and many pat- 2.303 log C = -kt 2.303 log C,where C straight line (37) whose slope is the ra-
ents have been issued, data are meager. is concentration of ethylene oxide in tio of the molal energy of activation to
Only one study reports on the kinetics of gram-moles per liter; t, time in min- the molal latent heat of vaporization of
ethylene hydration in acids ( 4 ) . utes; k, reaction velocity constant; and water a t the same temperature. In these
A solid, ion exchange resin as a cata- C, initial ethylene oxide. The slopes acid media, this gives 18,000 calories per
lyst can be easily filtered from the reactor equal -k/2.303, and the reaction veloc- gram-mole, compared with 18,100 in
liquid, corrosion is minimized, the same ity constants are in Table I. alkaline solutions (23).
catalyst can be used repeatedly, and the A recent study (35) considers reaction Velocity constants a t unit normality
operation can be continuous. Although velocity and mechanism of the molecu- (Table 11) are plotted in Figure 3,B, to
used as a catalyst elsewhere (9, 70, 79, lar reaction at 0' C., a temperature much give a straight line. Data of others (3,
27, 47, 48), ion exchange resins were un- too low for industrial interest. The rate 45) agree closely. The energy of activa-
reported when this work was done. equation given is a more complicated tion from the slope is also 18,000 calories

Sulfuric Acid Catalyst

T o study the kinetics of a homogeneous Table 1. Hydration of Ethylene 'Oxide Using as catalyst Sulfuric Acid
reaction with sulfuric acid as a catalyst,
Hydrogen Reaction
stoppered 500-ml. glass bottles were im-
mersed in a water bath with temperature
Temp.,
O c. Normality PH
Ion concn.
x 108
'
Veloc. Con-
stant X 10s
controlled to 0.01' C., always kept be- 30 0.00165 2.90 1.26 1.381
low boiling points of the charge. Aque- 0.0105 2.16 6.93 7.600
ous ethylene oxide solution (200 ml.) 0.0225 1.86 13.82 16.200
0.055 1'50 31.65 35.17
was added, and then sulfuric acid in
0.140 1.12 76.00 92.00
small amounts. Samples were frequently
50 0.055 1.50 31.65 367.0
analyzed, and p H a t room temperature 16.20 141.5
0.0275 1.79
was measured with a Beckman meter. 0.0115 2.13 7.42 58.0
Ethylene oxide was determined in dilu- 0.0024 2.77 1.70 14.05
tions of 2 to 10% by Lubatti's method, 70 0.0052 2.45 3.55 129.0
using hydrochloric acid to form ethylene 0.00185 2.85 1.41 54.4
chlorhydrin. Only 97.6% conversion to 0.00125 3.00 1.00 34.1
glycol has been reported (75, 23). This 0.0001 1 3.97 0.107 4.6
was confirmed experimentally and by 20" 0.00272 0.886
kinetic calculations; hence, all values of 0.00757 2.461
0.01015 3.300
exthylene oxide have been corrected. 0.01496 4.900
The straight lines in Figure 1 indicate 0.02017 6.600
that the reaction is first order because of 0.01015 3.415
the large excess of water, even though 0.01493 5.260
moles of two reactants must meet. Ini- 25' 0.0121 6.45
0.0222 11.75
tially, however, the line tends to curve,
0.0421 17.75
probably because the rapid initial reac- 0.0195 9.67
tion gives a large amount of heat which 0.0391 18.90
1 Present address, Indian Aluminum
Co.;Ltd., Calcutta 16, India.

VOL. 50, NO. 9 SEPTEMBER 1958 1235

 2

I
.8
.6
A

.2

.I
,Ot
OE
04
T: 70'C

\s \
02

.01
0 100 200 300 IO 20 30 40 IO 20
Time in Minutes
Figure 1. Straight lines on plot of logarithm of concentration of ethylene oxide remaining vs. time indicate that, with a sulfuric
acid catalyst, the reaction is first order
per gram-mole. Figure 3,C was con- dissociated but insoluble acid. React- (dry) was 200 grams. This is one of the
structed iimilarly for the velocity con- ants must diffuse and react in the pres- strongest acid ion exchangers commer-
stant us. temperature at different pH. ence of hydrogen ions on the surface and cially available and was converted to the
The dotted lines assume that the direct inside the pores of the resin. The prod- acid (hydrogen) form before each use by
proportionality between the velocity ucts then diffuse back to the body of the adding five times the resin weight of 10%
constant and p H holds in almost neutral solution. Ion exchange resins compared sulfuric acid, or twice that theoretically
solutions, and at a p H of zero, used math- to a soluble acid thus have obvious ad- required. The mixture was stirred for
ematically for determining the intercept vantages-the catalyst is easily removed 30 minutes, the resin settled, and the
of the lines. by filtration and can be used repeatedly; acid layer decanted. The same amount
I n neutral solution, the hydration re- the products are not contaminated since of acid was added again, and the batch
action is extremely slow (23), and the the anion of the resin is inert to the reac- was stirred again for 30 minutes. The
data on a plot such as Figure 3 form a tion; little corrosion of the equipment treated catalyst was filtered and washed
straight line, the slope of which gives an occurs; and continuous operation of a with distilled water till the effluent
energy of activation of 19,000 calories reactor may be possible. showed no precipitation with barium
pergram-mole. This is higher than for Resin Used. The resin, Amberlite chloride solution. I t was used immedi-
acid or alkaline media (18,000) although IR-120 (Rohm & Haas), a brown, ately.
possibly within experimental error. granular, beadlike material when washed Procedure. A 500-ml. glass flask with
and drained, had a bulk density of 48 agitator was used in the constant tem-
Ion Exchange Resin Catalyst: pounds per cubic foot, moisture content perature bath. Aqueous ethylene oxide
Heterogeneous Reaction of 35 to 45%, screen grading (wet) of (250 cc.) was stirred for about 10 min-
16 to 50 U. S. standard, and voids to an utes to bring it to the bath temperature.
Ion exchange resins as acid catalysts extent of 45 to 50%. The maximum A weighed quantity of resin in acid form
have a number of practical advantages. operating temperature was 250 F., and of proper particle distribution was
Their catalytic hydrogen ions are bound the operating pH range was from 1 to 14. charged; and the mean time of addition
to the highly insoluble anion part of the The exchange capacity (dry) was 5 meq. was recorded as zero time. Agitation
resin. The resin thus behaves as a highly per gram, and the equivalent weight kept the resin particles in suspension.
Samples were withdrawn a t frequent
intervals by means of a pipet, covered
with a cloth to prevent withdrawing resin
particles. Ethylene oxide was deter-
mined by Lubatti's method. The resin
concentration varied from 20 to 200
grams per liter of the reaction solution,
and average particle size varied from 0.05
to 0.0035 inch.

Table II. Hydration of Ethylene Oxide

4 Using as Catalyst Sulfuric Acid
Figure 2. Straight lines
Vapor ko
result when reaction Press. of for Unit ko for
velocity constants at Temp., Water, H Ion Unit Nor-
constant temperature O C. Mm. Hg Concn. mality
are plotted against 20 17.53 0.32
normality or hydrogen 25 23.76 0.51
30 31.82 1.19 0.71
ion concentration
50 92.51 8.45 4.95
~.01 .02 03 04 05 06 70 233.70 38.00 22.8
Data at 20' C. 131; at 25' C.
Normality of Solution (45)

1236 INDUSTRIAL AND ENGINEERING CHEMISTRY

 GLYCOL FROM ETHYL OXIDE

Temp. "C
10 20 20 40 60 80 100 01
R
/

1
.--
/

I
I: I-
$/
/ 4 .I
I
I
I
/
I
I

I
100 100 100
Vapor Pressure of Water mm
Figure 3. Logarithmic plot of reaction velocity constants against vapor pressures of water at the same temperature gives
straight lines (3 7 1
A. For unit hydrogen ion concentration C. At constant pH
B. Unit normality of acid; solid dot at 20' C. (3); at 25% C. ( 4 5 ) --- Outside experimental range
Results. Straight lines of Figures 4 was steeped in an ethylene oxide solution 70°, and 76' C.
(concentration varying) and Figure 5 for several hours before its conversion to Uncrushed, dried resin was used a t
(particle size varying) indicate that this acid form; but Figure 4 shows this to 76' C.; hence, the velocity constants
heterogeneous reaction is also of the give the same degree of curvature. obtained should check with those ob-
first order. A slight curvature in the At 30' C., straight lines may not quite tained with raw resin (Table 111). The
initial periods, similar to that obtained represent the data. area for the raw resin was calculated
using sulfuric acid, can again be ex- Velocity constants, calculated from from its screen analysis; and the result-
plained by an initial temperature rise. the slopes of the lines of Figures 4 and 5 ing value with the interpolated velocity
At first, the variation from a straight (Tables I11 and IV), vary with catalyst constant a t 76' C. is presented in Table
line, especially at the lowest tempera- concentration (grams raw resin per liter) I V to give a fairly good check. In runs
ture, was thought caused either by ad- and are proportional to catalyst concen- made at 30' and 70' C., the resin had
sorption of ethylene oxide on the surface tration except a t low concentrations, been crushed and oven dried. This
of the resin with a consequent rapid de- where there may be poor distribution may explain the lack of exact agreement
crease in ethylene oxide concentration, or (79). with raw resin.
to deterioration of resin activity by ad- The average particle size was assumed Temperature. With raw resin, the
sorption of ethylene oxide molecules to be the arithmetic mean of the minus velocity constant increases exponentially
which block many of the hydrogen ions. and the plus screen size openings, in a with temperature. The reaction veloc-
T o determine these possibilities, the resin standard screen analysis. A logarithmic ity constants times 1000 for 100 grams of
plot of Table I V gives straight lines of raw resin per liter at 30 ' C. is calculated
negative slope showing the hydration as 29.5; at 50', 114.0; at 70°, 259.0;
Table 111. Hydration of Ethylene Oxide rate to go down as the particle diameter and at 90°, 627.0. These values are
Using Resin Catalyst goes up. The trend a t 30' C. is not plotted logarithmically against the vapor
(Varying weights of resin catalyst) definite, because of extremely slow rates; pressure of water a t the same tempera-
Reaction but the correlation otherwise is good. ture in Figure 7 by the method of Oth-
Temp., Veloc. Con- Interfacial Area, If the particles are mer and Luley (3.7). The slope of the
c. G. Resin/L. stant X 101
assumed to be spheres, the surface area straight lines times the molal latent heat
30 19.3 3.34
76.8 6.67 may be Calculated from the particle of water a t the same temperature equals
112.8 15.55 diameter (Table IV). Because the the ratio for energy of activation of the
150.0 41.00 reaction velocity constant varies directly reaction, or 10,000 calories per gram-
40.0 14.90 with mass of the catalyst (diameter to
80.0 34.50
mole. This is substantially lower than
50 18.00 11.30
the third power) and inversely as the 18,000 for the homogeneous reaction us-
40.0 31.0 particle size (diameter), it is directly ing sulfuric acid catalyst. Athough a
60.0 61.0 proportional to the surface of the par- IOo C. temperature change doubles the
120.0 138.2 ticles (diameter to the second power) rate of homogeneous reaction, about a
70 16.0 28.6 which is the interfacial area of contact 20' C. change doubles the heterogeneous
40.0 94.2
80.0 208.0
with the liquid. This is shown in Figure rate. The point a t 30' C. is again
90 12 66.7
6, using Lichtenstein and Twigg data slightly off the line.
20 123.0 (23)for zero area. Correlation is poor Catalyst Activity-Resin vs. Acid. The
40 253.7 a t 30' C. for reasons previously ex- weight of resin added per liter and its
plained, but good a t temperatures of No, acid equivalence gave normality. Rates

VOL. 50, NO. 9 SEPlEhABER 1958 1237

 0 20 40 60 80 100 0 10 20 30 40 50 60 0 IO 20 30 40 50 0 IO 20 30
Time in Minutes
Figure 4. The heterogeneous reaction using an ion exchange resin catalyst is first order as shown b y straight lines on plot of
logarithm of concentration of ethylene oxide remaining vs. time
0 100 200 -__ Catalyst first steeped in ethylene oxide solution

using the acid catalyst are from 10 to 100 compared to that of sulfuric acid indi-
times those using an equivalent normality cates that hydrogen ions in the interior
of resin (Figure 8); or a temperature of of the resin are not readily available,
about 30" C. with the acid gives a rate because of diffusion resistance.
equal to that at about 70" C. with the I n the surface mechanism: dissolved
same normality of resin. This lower ethylene oxide is adsorbed on the sur-
activity of a resin, noted previously, is face and reacts with water in the presence
probably caused by the phenomenon at of hydrogen ions. Glycol is desorbed
the interface where physical diffusion and diffuses to the bulk of the liquid.
may be required before chemical cata- Absorption, chemical reaction, and de-
lytic effect can be appreciated. sorption are all probably slower than
Probable Mechanism of Resin Catal- diffusion in the liquid of product or re-
ysis. Besides the chemical hydration re- actant.
action of ethylene oxide with water In the interior mechanism, ethylene
catalyzed by hydrogen ions on the sur- oxide which has been adsorbed on the
face of the resin, there are at least four surface will also diffuse within the resin
physical steps. Two are of diffusion: mass, before reacting with water to form
of ethylene oxide to the resin surface glycol. This is desorbed and diffuses
and through its pores, and of glycol out back through the pores into the liquid.
the pores to the solution; one is of ad- The slowest step in this process seems to
=
-1
0 IO 20 30 40 50 sorption of ethylene oxide on the sur- be the diffusion through the resin pores.
face; and one is of desorption of glycol In the actual reaction, the combined
L
from the surface. effect of these two types of mechanism
E 2
The slowest step controls the over-all may give the median result of the effec-
-
0, disappearance of ethylene oxide; this is tive energy of activation of 10,000 calories
2 0:
probably diffxion through the pores of per gram-mole as shown experimentally.
06
the resin, because:
E 04 1. The energies of activation of ad- Yields of Glycol Using
zL

sorption, reaction, and desorption are Resin Catalyst

about 20,000 calories per gram-mole, Ethylene oxide reacts with glycol to
3 02 while that for diffusion is about 5000.
6 give diethylene glycol, which may react
The energy of activation of the over-all
.01 disappearance of ethylene oxide as deter- with ethylene oxide to give triethylene
008 mined previously is 10,000 calories per glycol, and thus to glycols (7) having
.OM gram-mole, indicating that the control- molecular weights of 1000 (viscous liq-
E 004 ling step lies somewhere in between the uids) to 7000 (waxy solids). These
.c diffusion and the chemical reaction. straight chain compounds, with two hy-
0 10 20 30 40 50 2. The reaction rate is proportional droxyl groups at either end (32) are dis-
to the interfacial area between the resin tributed (77) according to Poisson's dis-
.--C0 4 and the reacting solution, as must also
e tribution law. As the ratio of ethylene
c be the diffusion rate. Decreasing the oxide to water increases, yield of higher
W particle size reduces the length of the
E .2 path for diffusion. This facilitates the glycols increases and that of glycol de-
u over-all reaction. Also, increasing sur- creases (8,27). With l mole of ethylene
.I face area of contact increases the re- oxide to 1 mole water, only 30% of re-
.OB action occurring on the surface. acting ethylene oxide goes to glycol with
.06 3. Lower activity of the resin as 70yo to higher glycols. The maximum
.04 amount of ethylene glycol present de-
creases as its rate of formation becomes
0 IO 20 30 40 50 less than its reaction to diethylene gly-
Time in Minutes 4 Figure 5. Ion exchange resin cata- col. The effect of the resin on the yield
Dry Resin, G. per Liter lysts at vatious temperatures with resin of glycol and of higher glycols on the
A. 40 particle size varied gives straight lines life of the catalyst is important.
B. 60
C. 20 of plot of logarithm of concentration of Procedure. I n a constant tempera-
D. 20 ethylene oxide remaining vs. time ture bath, a three-necked 1-liter flask

1238 INDUSTRIAL AND ENGINEERING CHEMISTRY

 GLYCOL FROM ETHYLENE OXIDE

10
-
50 -
?O -

10 -

T. 76' C

0 1 1 ' 1 ( I I I I l I

0 100 200 300 400 500 600 0 500 1000 1500

lnlerfaciol Area of Contact (crn2)
Figure 6. The reaction velocity constant is shown b y the straight lines to be directly proportional to the surface area of the
particles which is the interfacial area of contact
0 Raw resin
was fitted with agitator, a gas feed tube Periodic acid oxidizes only compounds glycols increase with increasing ratio
connected from a supply of ethylene containing hydroxyl groups on adjacent (Figure 10). The same yields of glycol
oxide to the bottom, and a reflux con- carbon atoms, not di- and polyglycols were obtained a t 20' and 50' C. while '
denser using water chilled to 5' C. containing hydroxyl groups on nonadja- an average yield at 80' C. was 5 to 8%
Water vapor could condense but un- cent carbon atoms. Check analysis of lower. Yields of glycol later obtained
reacted ethylene oxide escaped. Ethyl- synthetic mixtures of glycol with diglycol at about 100' C. in a continuous reactor
ene oxide feed was determined from dif- proved this. are still lower.
ferences in weight of the supply bottle. Glycol was fractionally distilled from Glycol yields at 80' C. from chemical
Raw Amberlite IR-120 containing the reactor liquid in a 30-sieve plate analysis checked very well with those
40% moisture after conversion to the hy- Oldershaw column to check the chemi- from distillation (Figure IO). Material
drogen form, was charged to the reactor, cal agalyses and to estimate the amounts balances are good, assuming formation
to which was added the desired weight of of higher glycols. of only mono-, di-, and triglycols.
water. The agitator was started. Eth- Results. Data of 15 runs are in Losses may be caused mainly by carboni-
ylene oxide (40 to 50 grams per hour) Table V. Figure 9 compares glycol zation in the distillation pot and to some
was fed to maintain a saturated solu- yields with those from reactions using liquid holdup in the column-about 15
tion of ethylene oxide for maximum reac- sulfuric acid catalyst and with data of ml. This was assumed to be an equi-
tion rate. From kinetic data thus ob- others. Yields decrease rapidly with molar mixture to di- and triglycol with-
tained, a weight of catalyst was calcu- increasing ratio of ethylene oxide to out large error.
lated to maintain an absorption rate of water. This confirms previous investi- Time and Temperature. Excess
70 to SO%, with molal ratio of ethylene gators (8, 27). Production of higher ethylene oxide kept the liquid saturated;
oxide to water of 0.038 to 0.78.
At 80' C., 50 grams of raw resin were
used; a t 50' and 20' C., 100 grams and Table IV. Hydration of Ethylene Oxide Using Resin Catalyst
200 grams, respectively. The loss of (Varying sizes of particles)
water vapor (0.1 gram per hour) through
the reflux condenser was negligible; and Interfacial
D Veloc. Area of
the longest run (20 hours) lost only 2 Temp., G . Catalyst Av. Part. Constant Contact,
grams. The feed was stopped at the O c. 250 CC. Size, In. 1/D x 108 Sq. Cm.
desired weight of ethylene oxide. T h e 30 0 ... 0.063 0
bath was heated to 80' C. for half an 10 o.Oi27 78.7 0.78 1330
hour to allow all ethylene oxide to react. 0.00925 108.1 1.29 1830
0.00635 157.3 1.22 2665
T h e reactor was cooled to room tem-
perature and weighed. The increase in 50 0
15 0.0&55
...
20.12
0.46
76.0
0
507
weight gave the total amount of ethyl- 0.02815 35.52 106.2 900
ene oxide reacted. The liquid, con- 0.01985 50.40 \ 177.5 1276
taining glycol and higher glycols, was 0.01 175 85.10 208.0 2155
filtered and measured. A correction 70 0 ... ... 2.6 0
was made for the liquid adhering to the 5 0.0127 78.7 36.3 666
catalyst. This volume was estimated by 0.00925 108.1 49.9 916
0,00495 201.7 86.3 1710
weighing the wet catalyst, subtracting
the weight of the dry catalyst, and divid- 76 0 ... ... 4.1 0
ing by the specific gravity of the reactor 5 0,04955 20.17 27.1 169
0.02815 35.52 40.3 300
liquid. 0.01985 50.40 57.1 426
Glycol was determined by the periodic a
... 85.0 574
method (34, 47) after boiling gently for a Raw dry resin.
15 minutes to drive off ethylene oxide.

VOL. 50, NO. 9 SEPTEMBER 1958 1239

 Temperature "G 4 1000
IO 20 40 60 80 100 120 140 ~i~~~~ 7, A
v

l o g a r i t h m i c r)
plot of reac- 0
tion velocity x
n
constant
vapor pressure
vs. -
5
IOC

0 of water at
x same temper- 2
ature gives
straight lines. $'
The slopes of 9 IC
these lines times >
the molal lot-
ent heat of 8 0 AMBERLITE 18-420
0 SULFURIC AClO
water at the
same temper- 12: I

I
ature equals
VI I
the energy of
activation ( 3 I ) A Normality of Reacting Solution
the same magnitude. This also explains
Figure 8. The logarithmic why distribution of products does not
plot of reaction velocity change greatly with temperature.
constants vs. normality
From the above derivation, it follows
shows rates using the sul- that the velocity constant k for unit cata-
IO 100 1000 furic acid catolyst are
lyst concentration varies directly with
Vapor Pressure of Water mm HQ from 10 to O0 times those ( l / M ) ( d W / d t ) ( l / C ) . Values of C at
using an equivalent nor- various temperatures and atmospheric
thus. the rate of reaction was equal to mality of resin pressure are available (6). Thus, a log-
that of absorption. At least for initial arithmic plot of absorption rate ex-
period, where glycol is dominant prod- Rate of reaction dC/dt equals ( d W / d t ) pressed as ( d W / d t ) ( l / M C ) us. the tem-
uct, absorption rate will depend only on ( l / V M )where d W / d t is rate of absorption perature derived from the vapor pres-
weight of catalyst, and will be inde- of ethylene oxide and M is molecular sure of a reference substance is a straighc
pendent of amount of water present. weight of ethylene oxide; hence: line, as in Figure 11. Its slope times the
The first order reaction, -dc/dt = kC, - d W / d t = k" mM. Thus, the rate latent heat of water is the energy of ac-
holds a t any temperature. Velocity varies directly with the weight of catalyst. tivation, 11,500 calories per gram-mole
constant k varies as catalyst concentra- The rate of reaction remained almost which is an average value for reactions
tion: k = k' m/V, where m is weight of constant regardless of reaction time or forming both glycol and higher glycols.
catalyst; and V , volume of liquid. Also, ratio of ethylene oxide to water. Thus, I t is a little higher for glycol alone
for the saturated solution, ethylene oxide ethylene oxide reacts with glycols a t (10,000). Thus, for higher glycols the
concentration C a t a given temperature about the same rate as with water. This energy of activation must be greatet
is constant; therefore, holds true for every temperature; thus, than that for glycol, and shows that the
-dC/dt = k" m 1V. the energy of activation for all runs is of comparative yield of glycol must de-

Table V. Hydration of Ethylene Oxide in Batch Agitated Reactor

800 c. 60' C. 200 c.
Time of operations, hr. 6.0 1.83 8.0 8.5 6.0 9.0 2.0 3.0 5.5 14.17 3.17 16.75 5.33 7.0 19.8
Water, g. 400 400 300 200 100 100 300 300 200 200 ... ,.. ... ... ,. ,
Dry resin, g. 30 30 30 30 30 30 30 60 60 60 60 60 120 120 120
Washed resin, g.
Total water charged, g.
72
442
73
443
74
344
73
243
74
144
72.3 73.3
142.3 343.3
143.2
383.2
135.3
273.5
156
296
... ...
454 203.0
...
381.3 233
...
393.9
. . #

ET0 fed, g. 209.8 56.6 277 275 214 271.8 93.5 96.6 155.8 295 44.2 356.1 132.5 178.5 367.7
Absorption of ETO, % 64.6 ... 65.2 79.5 87.7 82.4 ... 95.7 78.3 82.4 88.4 50.1 94.5 #.. ..#

Rate of absorption (9. 70.0 61.9 69.3 64.7 71.3 60.4 93.5 32.2 28.3 20.8 13.9 33.2 12.4 12.74 9.8
ET0/100 g. resin)
x 100
(Moles ETO/moIes 19.45 5.23 33.0 46.3 60.9 78.1 11.17 10.32 22.95 40.8 3.98 71.8 19.2 31.33 38.2
H20)X 100
Yield Based on Chemical Analysis by Periodic Method
Glycol, % 69.0 79.7 60.0 52.9 44.1 38.8 75.7 75.3 72.2 56.9 92.3 43.7 78.0 64.9 59.0
Results of Fractional Distillation of Products"
Glycol yield, % 68.2 77.3 61.1 53.5 43.2 75.7
Diglycol yield, % 13.55 14.5 15.62 14.15 32.6 25.7
Triglycol yield, % ... ...
... ,.. ... ... 21.8 ...
...
Di- +
triglycol, % 16.5 15.20 24.10
Total wt. in., E T 0 + 651.8 499.6 621.0 518.0 358.0 436.8
HnO, g.
Total wt. out, glycols 643.8 499.3 602.8 509.9 347.4 435.6
+ &O, g*
ET0 balance, % 98.1 91.9 92.0 91.8 98.7 101.3
Total balance, % 98.9 100.0 97.1 98.4 97.1 99.8
a Based on amount of ethylene oxide added.

1 240 INDUSTRIAL AND ENGINEERING CHEMISTRY

 .
GLYCOL FROM ETHYLENE OXIDE

4
Figure 9.
Glycol yield
with all dif-
ferent cata-
lysts d e -
creases rap-
idly as the
ratio of
ethylene ox-
ide to water
increases

Mol Ethylene Oxide Reacted x IOO/Mol Water

crease slightly as the temperature in- Figure 10. Higher

creases. glycol production increases
Higher glycol formation did not inter- as the ratio of ethylene
fere with catalytic activity. After 18 to oxide to water increases,
20 hours there was no decrease in the both with ion exchange 0 IO 20 30 40 50 60 70
rate. Also, the same catalyst was used resin catalysts and without Mol EthO 100
for several successive runs without any catalysts Mol Water
decrease in activity. ---Without catalyst (8)
Glycol yields a t 20' to 50' C. com- A flanged borosilicate glass pipe, 4
ethylene oxide. The bomb was weighed feet long, having an inside diameter of 1
pared favorably with Matignon, Mourau,
to determine any loss of material, usually inch and tested to pressures of 100 p.s.i.
and Dode (27), using sulfuric acid as
less than 1%. The liquid was analyzed was the reactor. A thermocouple, pushed
catalyst, and were about 10 td 1270
for glycol (Table VI). up or down a '/rinch thermowell,
higher than Davis (8) without catalyst.
U p to a molal ratio of 1 to 5, Amberlite
No catalyst gave a higher yield than showed temperatures a t any height. T o
sulfuric acid or Amberlitd IR-120 but confine the resin bed, an 80-mesh screen
gave about 12 to 15% higher yields than
silver oxide, phenol, benzoic acid, zinc, was placed between the flanges a t the
sulfuric acid. At 80" C., Amberlite
and potassium acid phthalate gave yields top and bottom of the glass reactor, with
gave yields about 6 to 870 lower than
of the same magnitude. Disodium hy- glass wool above the top screen. Dis-
sulfuric acid u p to a molal ratio of 0.30,
drogen phosphate, sodium bisulfite, and charge from the reactor passed through a
while a t higher ratios, Amberlite again
sodium chloride gave much lower yields. coil of a/8-inch tubing immersed in cold
gave higher yields.
Sodium hydroxide gave extremely poor water and followed by a rotameter, then
At lower temperatures, with repeated
yields (Figure 9) which indicated that it by a needle valve controlling the flow,
use and vigorous agitation, the resin did
acts more as a polymerizing agent than a and thence to a glass jar for collecting
not break up into fine particles as it did
hydrolyzing agent and thus is good for the product. Nitrogen supplied through
a t higher temperatures.
making higher glycols. Phenol has an a pressure regulator maintained the sys-
advantage-besides giving high yields, tem a t 50 to 70 p s i . , to keep the ethyl-
Other Catalysts it can be volatilized from the liquid dur- ene oxide in solution. For protection
For sodium hydroxide and phenol, the ing purification. against explosions, there was a relief
glass reactor described previously was valve; and the glass reactor was en-
used. For silver oxide, zinc dust, ben- Fluidized Ion-Exchange Resin Bed closed in a cage of wire screen. A small
zoic acid, potassium acid phthalate, and A resin catalyst allows the use of a orifice in the feed line would have pre-
disodium hydrogen phosphate, a strain- continuous reaction in a fluidized ion vented major discharge of feed liquid if
-less steel bomb, of 2-liter capacity, was exchange resin bed, with ethylene oxide the reactor burst.
used and for sodium chloride, a steel in water flowing upward. Several runs Procedure. T o about 40 pounds of
autoclave of 1-liter capacity was em- were made in such a reactor. distilled water in the feed tank, liquid
ployed. All liquid-contacting metal parts of ethylene oxide is fed, and resin, converted
Using sodium hydroxide and phenol, the equipment were of stainless steel. to the hydrogen form, is added to the re-
the same procedure was followed as with The bomb in Figure 12, 1.2 cubic feet actor.
Amberlite IR-120. When using the capacity, was a feed tank for ethylene A constant pressure of nitrogen is ad-
bomb or the autoclave, it was chilled, and oxide solution. The feed was heated justed. The final needle valve controls
to it was added the weighed ethylene to the reaction temperature by an elec- the flow, indicated by the rotameter.
3 C. and the cat-
oxide solution a t 0' to ' trical coil wound around a standard 1/4- Electrical heat raises the feed liquid to the
alyst. I t was then placed in the oil bath inch stainless steel pipe. A central l/8- desired temperature. Because of heat of
a t a desired temperature. Pressure in- inch rod, wound spirally with a wire, in- hydration, the temperature increases
creased, then dropped as ethylene oxide side the '/a-inch pipe reduced the effec- with height in the reactor and is deter-
reacted, finally dropping to the vapor tive pipe cross section and increased tur- mined by the adjustable thermocouple.
pressure of water, with disappearance of bulence, to improve heat transfer. When temperatures are steady, the rotam-

VOL. 50, NO. 9 SEPTEMBER 1958 1241

 -
Temperature O C Pressure Regulator 5 Pressure

-
+
100
IO 20 40 60 80 100

60 K

-
7560
0

0
c
40

E 20
3

h '8
c
6
C
0
c 4
VI
6
0 2
%
5 1
-O .8
>" .6
c .4
.-
c
0
i

s
V
.2 Figure 12. The fluidized-bed unit for continuous hydration of ethylene
'I oxide has all liquid-contacting parts of stainless steel
01
IO 100 1000
4 Figure 1 1. A logarithmic plot of the reaction velocity constant using
a resin catalyst vs. temperature derived from the vapor pressure of a
Vapor Pressure of Woter (mm HQ) reference substance gives a straight line (37)

eter reading, pH of the clear liquid and the products obtained by distilla- may decompose the catalyst, and high
product (usually about 6.0), tempera- tion are within the expected precision, pressure drop when using small particles
ture, pressure, and the bed height are and show that ethylene oxide reacts to increase the surface. The expansion
recorded. Samples are analyzed by only to form glycols. of the Amberlite bed at different flow
Lubatti's method and by distillation. \\'hen the feed to catalyst weight ratio rates and temperatures (Figure 14) gives
Results. Data are summarized in was increased, at 100' C., the conver- data for design of fluidized reactors.
Table VII. The yield of glycol is shown sions followed those predicted. How-
in Figure 13 as a function of feed ratio ever. partial deactivation of the catalyst Recommendations
of ethylene oxide reacted per mole of (10% loss in 24 hours) resulted in con-
water. Yield declines rapidly as feed versions lower than predictions. Slight Batch hydration of ethylene oxide in
concentration, ratio of ethylene oxide to corrosion of stainless steel (Type 304) water using sulfuric acid catalyst
water reacted, or reaction temperature may be the cause of gradual catalyst has the advantage of low temperatures
increases, as also shown in batch studies. deactivation. and low pressures; however, the acid
This correlates with the higher energy of The usual advantages of fluidized- has to be replaced with each batch, there
activation for reaction to higher glycols over fixed-bed catalysts were noted: less is corrosion. and the purification of the
previously noted. The material bal- danger of fouling, channelling, hot spots glycol product is difficult. Also, sul-
ances between ethylene oxide reacted because of poor heat distribution, which furic acid Lvith ethylene oxide and glycol

Table VI. Hydration of Ethylene Oxide Using Various Solid Catalysts

Re-
Catalyst Moles ET0 actor Mat.
Temp., Max. H20, Approx. ETO, Moles HzO Liquid, Total Total Bal., - Glycol, G.
c. Press. G. Material G. Wt., % G. x 100 VOl. Wt. In Wt. out % Estd. Yield
50 0 398.7 NaOH 8.1 2 162.7 16.7 538 ... ... ... 42.3 18.45
497 NaOH 4.0 0.8 84.7 6.96 565 *.. ... ... 44.3 37.2
838 NaOH 4.0 0.5 57 2.78 873 ... ... ... 46.6 58.1
367.5 NaOH 2.0 0.5 225.5 25.1 558 ... ... 40.2 12.65
530.5 Phenol 25 5 90.0 6.94 633 ... ... 111.0 87.8
75 0 265 Phenol 13.2 5 56.8 8.77 322 ... ... ... 67.1 83.8
50 0 296 Phenol 15 5 67 9.25 368 ... ... ... 71.2 75.5
100 36 418 NaCl 50 12.5 77 7.55 495 545 534 14 98.1 52.1 48.0
53 200 NaCl 55 27 79.4 16.25 273 334.4 313.6 93.8 33.6 30.0
113 ... 430.8 Ag2O 4.3 1 118.7 7.75 ... 553.5 516.3 93.4 98.2 86.4
115 ... 513.8 AgzO 5 1 119.6 6.6 . . I 638.4 610 96.3 112 95.9
110 ... 326.3 Ag2O 5 1.5 118.4 9.64 ... 445.2 408.2 91.7 93.1 85

118
103
77
488.5
407.7
Ag?O
Zn
5
5
1
1
123.5
66.4
10.35
6.67
...
...
617
479.1
612.6
477.1
99.3
99.6
138.2
76.7
79.4
82.0
122 103 348.1 Zn 5 1 68.0 8.0 ... 421.1 420.6 100.0 77.3 80.0
115 73 363.5 Benzoic acid 5 1.3 69.1 7.78 ... 437.6 435.6 99.6 79.7 81.9
... 300 Benzoic acid 5 1 43.2 5.89 ... 348.2 349.5 100.3 48.2 79.2
120 .. 299 KH phthalate 4 I 51 6.98 ... 354 353 99.7 59.2 82.5
... 310.8 KH phthalate 5 1.25 93.5 12.32 ... 409.3 407 99.4 84.1 63.8
120 314.8 NazHP03 10 2.5 105.1 13.67 ... 429.3 427.2 99.5 67 45.2
115 90 310 NazHPOa 10 2.5 72.4 9.55 ... 392.4 391.4 99.7 58.3 57.2
100 ..* 312.1 NaHS03 5 1.0 128.6 15.8 ... 443 * 7 439 99 85 48.7

1242 INDUSTRIAL AND ENGINEERINGCHEMISTRY

 GLYCOL FROM ETHYLENE OXIDE

~ ~
\
-0
al - \ -
c
u -
0 - -
B. -
O\'\ I'o4
'\ \0
\ '%"
*,% % -
0 -
80
C
- \ % -
i
U
al
u)
- \ %$ 1 20
0
0 - 0 '
'0
- 0 1000 2000 3000 4000
00 - \
-
- I
-
Flow Rate Lb/Hr x Ft2
8a 70 - 0 \o -
1 Figure 14. Data of per cent bed expansion vs. flow rates
'5 - Continuous Reactor
- at different temperatures can be used for designing fluidized
- -
L
0 reactors
rral - - 4
F
o\.o
60 i
0
-
''' ''''
1
' I

2
' '
3
' I

4
I'

5
I' ''' I

6
'
Y
Figure 13. Glycol yield
' declines rapidly as feed
7 concentration,ratio of ethy-
A 10 cubic foot bed of resin a t 100' C.
should convert 50,000,000 pounds of
ethylene oxide to glycols per year. Since
higher temperature gives somewhat
Mol Eth Oxide lene oxide to water in feed lower yields, an operating temperature
,oo or reaction temperature, of about 50' C. is preferable, giving
Mol Woter
increases yields of -glycol
. as high as with any other
I

process, and requiring a bed volume of

form a complex compound which causes continuous process with all the advan-
about 60 cubic feet. The operation
a loss in yield. Hydration a t high pres- tages of the direct hydration method. requires deionized water. The resin is
sure and temperature eliminates the The product emerges a t almost neutral
deactivated by ions which exchange with
disadvantages of acid catalyst; but p H and is readily refined. No additive
its hydrogen ions. Condensate water
equipment is costly and yield is lower. compounds are formed; thus a high con- from the purification and distillation
Also, ethylene oxide under high pres- centration of catalyst can be used. I t system can be used.
sures and temperatures has caused vio- has ,a high productivity, and excellent
lent explosions. heat distribution iq the reactor, which re-
The fluidized ion exchange resin bed sults from the usual properties of a fluid- References
achieves the advantages of both proc- ized bed. The catalyst can be regener-
esses without the disadvantages of either. ated by fluidization with circulating sul- (I) Auslerveil, J. V., Brit. Patent 676,891
I t is a low temperature, low pressure, furic acid and then wash water. (Aug. (7, 1952).

Table VII. Fluidized Ion Exchange Resin Bed

(Continuous hydration of ethylene oxide)
Feed rate, lb./hr. 13 13 13 13 13 13 13 11.0 7.5 7.5 5.5 5.5
Feed rate, lb./hr. sq. ft. 2600 2600 2600 2600 2600 2600 2600 2200 1500 1500 1100 1100
E T 0 in feed, % 9,96 9.4 6.20 9.22 9.22 10.5 10.51 3.2 9.1 3.2 10.45 3.2
Reactor temp., O C.
In 88 94.5 86 91.2 92.8 91.6 91.1 79.4 70 70 49.4 71
out 91 125.5 107.3 114.5 104.5 105 103.3 79.4 76.7 70 73.9 71.5
Av. 89.5 110 96.6 102.8 96.1 98.8 97.2 79.4 73.3 70 61.6 71.2
Resin bed
Wt., lb. 0.342 0.455 0.456 0.22 0.22 0.119 0,119 0.44 0.44 0.44 0.44 0.44
Static height, in. 17.5 24 24 11.25 11.25 6 6 22.5 22.5 22.5 22.5 22.5
Fluidized height, in. 21 30 40 14.5 16.5 8 8 25.5 28 23 23 23
Product
ETO, % 0.2 0.47 0.0 0.47 1.43 0.48 3.28 0.42 0.48 0.0 0.0 0.0
PH 4.6 5.6 4.1 ' 6.2 6.2 6.5 4.7 4.7 4.7 4.7
Conversion, %
Actual 98 95 100 95.5 84.5 95.6 68.8 87 95 100 100 100
Predicted (kinetics) 100 100 100 100 100 99.2 99.2 100 100 100 100 100
Productivity, Ib. ETO/lb. 3.65 2.66 1.81 5.15 4.45 11.35 7.90 0.70 1.47 0.545 1.305 0.40
resin/hr.
-
Moles E T 0 reacted 4.48 4.05 2.72 3.98 3.56 4.66 3.37 1.18 3.92 1.36 4.84 1.427
x 100
Moles HzO
Distillation analysis
Glycol yield, % 73.7 80.8 92.0 71.1 74.6 71 84.5 95.4 95.4 91.3 87.1 101.5
Diglycol yield, % 17.6 17.3 7.8 16.55 10.62 13.5 8.72 0 8.1 0 16.9 0
Triglycol yield, % 0 0 0 9.57 7.91 9.54 0 0 5.4 0 0 0
E T 0 reacted recovered 91.4 98.1 99.7 98.5 93.1 91.1 93.3 95.4 97.7 91.3 103.9 101.5
a s products, %

I VOL. 50, NO. 9 SEPTEMBER 1958 1243

(2) Balcar, Frederick R., U. S. Patent M., U. S. Patent 2,255,411 (Sept. 9, (10) Dierichs, P. B., Rept. 866, U. S.
2,135,271 (Nov. 1, 1939). 1942). Dept. of Commerce, Washington, D. C..
(3) Bronsted, J., Kilpatrick, M., J . Am. (6) Coles, K., Popper, F., IND. ENO. 1945.
Chem. SOC.51, 428 (1929). CHEM.42, 1435 (1950). (11) Dreyfus, Henry, Brit. Patent 387,372
(7) Curme, G., Jr., “Glycols,” Reinhold, (Feb. 6, 1933).
(4) Cartmell, R. R., Galloway, J. R., New York, 1952. (12) Dreyfus, Henry, French Patent 729,-
Olson, R. W., Smith, J. M., IND.ENG. (8) Davis, P., M.Ch.E. thesis, University 957 (March 27, 1931).
CHEM.40, 389 (1948). of Kansas, 1949. (13) Zbid., 737,612 (May 25, 1932).
(5) Cohen, Charles A,, Beames, Clayton (9) Davis, P., fiaature 159, 372 (1947). (14) Dreyfus, Henry, U. S. patent 2,071,-
395 (Feb. 23, 1937).
(15) Eastham, A,, Lalrensville, G., Can. J .
Research B 28, 264 (1950).
(16 ) Ferrere. Paul. Vandendries. Cornille.
’ Berbe, Frincois,’ U. S. Patent’2,108,936
Literature Background (Feb. 22, 1938).
(Chronological sequence, 1860-1957) (17) Flory,
I . ^ -~ P., J . Am. Chem. SOC.62, 1561
(1Y4U).
Subject Conditions Catalyst Ref. Remarks (18) Hamilton, G. E., Metzner, A. B.,
Hydration of ethylene to (49) First hydration reported IND.ENG.CHEM.49, 839 (1957).
glycol (36) Hydration confirmed (19) Haskell, V., Hammet, L., J . Am.
($0) Chem. SOC.71, 1284 (1949).
(20) (20) Henry, Comfit. rend. 144, 1404 (1907).
Hydration of ethylene Nitric-perchloric (46) First use of acid cata- (21) Kreps, Saul I., Nachod, Frederick
oxide in presence of acids lyst C., U. S. Patent 2,477,380 (July 26,
acids 1949).
Kinetics of ethylene 200 c. Perchloric acid (3) First order rate coeffi- (22) Lavesque, C., Craig, A., IND.EXG.
oxide hydration cient directly propor- CHEM.40, 96 (1948).
tional to hydrogen (23’1 Lichtenstein. Twiw. G. H.. Trans.
concentration ‘ Faraday Svc. 905 (194gr’
Direct production of High press. Oxygen-carrying (12) Manufacturing proce- (24) Long, F. A., Pritchard, J. G., J . .4m.
glycol from ethylene (42) dure described Chem. SOC.78, 2663 (1956).
and oxygen in water (18) (25) ?r.lcAteer, James H., U. S. Patent
Minimizing formation Large excess Sulfuric acid (60) Excess water minimizes 2,547,766 (April 3, 1951).
of higher glycols water formation of higher (26) McClellan, P. P., IND.ENG.CHEM.
glycols 42, 2404 (1950).
Effect of reactant com- 50-900 c. Sulfuric acid (27) Ethylene glycol yields (27) Matignon, C., Mourau, H., Dode,
position on yield of decreased with de- M., Bull. svc. chim. (3) 1, 1308 (1924).
glycol creased ratio of H?O/ (28) Metzger, Floyd J., U. S . Patent
oxide 2,409,441 (Oct. 15, 1956).
Direct conversion of Presence of Oxidizing (@) (29) Millas, Nicholas A,, U. S. Patent
ethylene and oxygen water (14) 2,437,648 (March 9, 1948).
to glycol (30) Nef, Ann. 325, 231 (1904).
Hydration of ethylene Oxalic acid (16) Recovery of acid by pre- (31) Othmer, D. F., Luley, A., IND.LG.
oxide cipitation with lime CHEM.38, 408 (1946).
Conversion of reactor In water soh. (2) Absorption of gases in (32) Perry, S., Hibbert, J., J . Am. Chem.
gases water SOC. 62, 2599 (1940).
Ethylene oxide hydra- Under pres- Iron compounds (59) (33) Petrol. Refiner 28, No. 3, 129 (1949).
tion sure (34) Pohle, W., Mehlenbacher, V., Cook,
Conversion of reactor In water s o h Sulfuric acid (6) Modification of ( 3 ) J., Ozl 6 SvaF 22, 115 (1945).
gases (35) Pritchard, J. G., Long, F. A., J . Am.
Hydration of ethylene Trichloroacetic (46) Recovery of acid by
Chem. SOC.78, 2667 (1956).
oxide acid volatilization
Direct conversion of In inert sol- Active oxide, (2s) Use of solid catalyst (36) Reboul, Ann. Suppl. 1, 113, 233
ethylene by hydro- vent soh. OsSOa, RuOi, (1861).
gen peroxide vzos,MOzO8, (37) Reed, C. F., U. S. Patent 2,407,344
CrvOs (Sept. 10, 1946).
Kinetics of ethylene 2Oo-13O0 C. Sodium hy- (93) In alkaline solution, (38) Reed, L. M., Wenzel, L. A., O’Hara,
oxide hydration in droxide reaction catalyzed by J. B., IND.ENG.CHEM.48, 205 (1956).
neutral and alkaline hydroxyl ion; first 139) Revnhart. Arthur F. H., U. S. Patent
solution order reaction with ‘ 2;236,919 (.4pril 1, 1941). ‘
respect to ethylene (40) Riddle, E.. Chem. @ Ind. (London) 67,
oxide ‘ 187 (Aug& i950).
Vapor phase catalytic 15Oo-20O0C.; Silver oxide on (4) This catalyst most ef- (41) Siggia, S., “Quantitative Organic
hydration of ethyl- vapor alumina and fective catalyst; Analysis,” p. 8, John Wiley, New York,
ene oxide phase others aldehyde by-prod- 1949.
ucts noted (42) Skarblom, Karl E., French Patent
Process description In s o h of 1%
’ Sulfuric acid (33) Continuous process
722,000 (Aug. 27, 1931).
sulfuric described
and 20% (43) Skarblom, Karl E., Swed. Patent
glycol 83413 (May 21, 1935).
Product distribution in 10Oo-14O0C.; (8) Data checked with ($7) (44) Skeen, J., Chem. Eng. 57 (7),331
noncatalytic hydra- 90- 120 (1950).
tion p.s.i.g. (45) Smith, L., Wode, G., M’idhe, T.,
Hydration of propylene Vapor phase Aluminum fluo- (26) 2.physik. cham. 130, 154 (1927).
oxide to glycol 15Oo-35O0 ride (46) Solvay & Cie., Belg. Patent 448,518
C.; 1 atm. (Jan. 1943).
Ethylene oxide hydra- Ion exchange (1) Fixed bed (47) Sussman, Sydney, IND.ENG. CHEM.
tion process resin 38, 1228 (1946).
Ion exchange resins as 18Oo-33O0 C. 7 ion exchange (38) Ion exchange materials (48) Underwood, G., Deartherage, - F.,
catalyst for ethylene 7.5-95 materials used effective; temp., . Science 115, 95 (Jan. 25, 1952).
oxide hydration p.s.i.a. press., and the effects (49) Wurtz, A.,Ann. 325, 255 (1863).
of excess water con- 150) Youtz, Merril A., U. S. Patent
tent noted
Perchloric acid ($4) Details of reaction 1,875,312 (Aug. 30, 1932).
Mechanics of hydroly- Oo and 25O
sis of substituted c.; 3 to (35) mechanism
ethylene oxides 14 pH RECEIVED
for review September 24, 1957
Kinetics of vapor phase 1100-2000 c. ; Amberlite IR-20 (18) Rate determining steps ACCEPTED May 1, 19%
hydration of ethylene 1.2 atm. and IR-C-50 evaluated
oxide Research suworted bv the Vulcan Copper
..
and Supply LCo.

1244 INDUSTRIAL AND ENGINEERING CHEMISTRY