ENGINEERING, DESIGN, AND PROCESS DEVELOPMENT

ties n-ere calculated from the data obtained in runs 89 and DO (13) K o b e , K. A,, a n d Dournam, T. F., U. 9 Patent 2,353,312 (July
11, 1944).
using an expected yield of 95.4%. The quantity of wash toluene
(14) K o b e , IC. A , , a n d Lueth, P.F., T m . ENG.CHEX, 34, 309 -13
was not increased proportionately, because the amount used in (1942).
the laboratory preparation was excessive. About 60% of the (15) Kobe, K. A,, a n d lIcCutchan, K. T., U. S. P a t e n t 2,628,241
diphenylmercury in the toluene can be obtained with each ( F e b . 10, 1953).
crystallization by cooling the solution t o 27’ 6. The quantity (16) Koton, hi. AI., J . Gen. C h n . (L‘.S.$.E.), 9 , 912-16 (193‘3).
(17) L a d c n b u r a , r l . , Ann., 173, 151 (1874).
of toluene in the mixture in each case is just enough so that the (18) L e i c e s t e r , H . hl., J . Am. Chem. Soc., 5 7 , 1901-2 (1936).
crystal slurry can be handled. (19) I h i d . , 60, 619-20 (1938).
(20) hlakarova, L. G., a n d N e s m e j a n o v , A. IT., J . Gen. Chcin.
(U.S.S.R.), 9 , 77-9 (1939).
References (21) M a y n a r d , J. L., J . Am. Chem. SOC.,46, 1510 (1924).
(22) Michaelis, A , Ann., 181, 290 (1876).
(1) Ironheim. B., Ann., 1 9 4 , 148 (1878). (23) Nad, 51. M., a n d Kochcshkov, K. A,, ,J. Gen. Chem. ( Lr.S.$.Ii.,,
(2) B a c h m a n n , W. E., J . Am. Chem. Soc., 55, 2830 (1933). 1 2 , 109-13 (in E n g l i s h 413-14) (3942).
(3) Blicke, F. F., a n d S m i t h , F. D., Ibzd., 51, 3479-83 (1929). (24) N e s m e j a n o v , A. H., a n d Koaesclikov, K. A., Be?., 63B, 2496
(4) B o r g s t r o m , P., a n d D e w a r , Margaret AI., Ibid., 51, 3387-9
(1930).
(1929).
( 2 5 ) Otto, Robert, J . pial;l. Choii., [ 2 ] ,1, 184 (1870).
(5) Calvery, H.0.. Ibid., 48, 1009-12 (1926), “Organic S y n t h e -
ses,’’ Collected Vol. I, pp. 228-9, S e w York, J o h n Wiley 6- (26) Pesci, L . , Gam. chim,.ital, 291, 394 (1899).
Sons, Inc., 1946. ( 2 7 ) Pfeiffer, P., a n d Ti.uskicr, P., Ber.. 3 7 , 1125 (1904).
(6) D i m r o t h , O., Ber., 3 5 , 2853 (1902). (28) Pyman, 1”. L.,a n d S t e v e n s o n , TI. A., Pharm. J., 133, 269
( 7 ) D i n i r o t h , O., Chem. Zenlr., 1901 I, 449. (1934).
(S) D r e h e r , E., a n d O t t o , R., $nn.. 154, 114-15 (1870). (29) S t e i n k o p f , W., et al., ,4niz., 430, 41 (1923).
(9) Ibid., p . 122. (30) T a b e r n , D. L.. a n d Shelberg, E. F., ISD. ENG.C m x . , ANAL.
(10) Dreher, E.. a n d Otto, R., Be?., 2, 542 (1869). E n . , 4 , 401 (1932).
(11) Freidliiia, R. K h . , a n d S e s i n e j a n o v , 8. N., Compt. w i d . m a d . *(31) K h i t m o r e , F. C., “Organic C o m p o u n d s of Nercury,” pp. 165-
sci. U.R.S.S., 29, 567-70 (1940). 70, S e w York, Chemical C a t a l o g Co., 1921.
(12) G i l m a n , H., a n d Brown, R. E., J . Am. Chem. SOC.,52, 3314-17
(1930). RECEIVED
for review August 10,19j3. :iccaPrED Febrriary l:, 1034.

ynthesis o
WILLIAM G. DOMASK’ AND KENNETH A. KOBE
Universify o f Texas, Aosfin, Tex.

Ethylene chlorohydrin has been made for many years, but the secretiveness that has sur-
rounded the process has allowed little specific information to appear in the literature. A
study has been made of the effect of several of the process variables on the yields and a com-
parison made with results from recycle and single column operation. Low temperature opera-
tion is undesirable because of extensive formation of by-products. High mole ratios of ethylene
to chlorine favor high yields of ethylene chlorohydrin, but the yield decreases as charge rates
increase for a given reactor. Physical observations and data indicate that the recycle-type
reactor has advantages over the single column reactor in terms of both yield and smoothness of
operation. Process variables which must be considered in reactor design are stressed, as weli
as conditions and method of operation.

UJIEROUS processes ham? been dcveloped for the synt,hesis chlorohydrin is particularly remarkable for t,he serretivcncs?
of ethylene chlorohydrin. A fen, of the methods border t’hat has surrounded it. operation. There is a considerablc
upon the bizarre; many morc are of academic interest only. volume of patent’ information but only a rather limited numhcr
The most important methods, from an industrial viewpoint’, of reports of laboratory investigations, which deal for the meet
involve the interaction of ethylene m-ith hypochlorous acid formed part \Tit11 studies of small scale batch preparations.
by ‘the solution of chlorinc in water. One of the principal uses K u r t z ( 1 6 ) first prepared et’hylene chlorohydrin from cthyleiic
of ethylene chlorohydrin is in the production of et’hylene oxide glycol in 1859 by reacting glycol and hydrochloric acid. Pour
which in turn is an intermediate in the synthesis of ethylene years later Carius ( 2 ) reported the more direct synthesis h y thc
glycol and a host of other compounds, including acrylonitrile, addition of hypochlorous acid t o ethylene, and this has become
ethanolamines, et,hers for hydraulic fluids, and polysulfide rubber the basic react’ion upon which most industrial processes arc
polymers. founded. In 1919 Gomberg (6) described the synthesis of
While coileiderable interest has recently been directed to the ethylene chlorohydrin from ethylene and chlorine passed into
catalytic oxidation of ethylene t o ethylene oxide, ethylene chloro- water. Rhern-ood ( 1 1 ) , AIcClellan ( 7 ) ,and Curmc a n d 3‘oiirisl;ori
hydrin will remain an important chemical intermediate for many (3) have described the chlorohydrin process for the production of
years. During the industrial expansion after World ’War 11, ct,hylene glycol, and Y~Iurray(8) and Shilov et nl. (19-1,$) h a w
many firma turned to the production of ethylene glycol through a described some of the more recent invcstigat,ions of thc eynthcsip
process that has been practiced in the United States in one form of ethylene chlorohydrin.
01‘ another since 1922. As pointed out by Sherwood ( 1 1 ) , the When chlorine dissolves in water it forms hydrochloric acid ant1
process for the manufacture of ethylene glycol via ethylene hypochloroup acid
1 Present address, Humble Oil and Refining Po., Baytown, Tex. ‘21, -1- HOH e HCI + HOC1 (1 1

680 I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY Vel. 46, No. 4

 UNIT PROCESSES

TEMPERATURE
CONTROL JACKETS

T H E R M O M E T E R WELL

250 M L
FLASK

Figure 1A. Reactor for Synthesis of Ethylene

Figure 1. Reactor for Synthesis of Ethylene Chlorohydrin
Chlorohydrin

Ethylene bubbled into this aqueous solution reacts with the hypo- was borosilicate glass. The various components were joined
chlorous acid t o form ethylene chlorohydrin together by ground-glass ball-and-socket joints or XEL-F tubing
(a polymer oT trifluorochloroethglene, a product of M. W. Kellogg
CzH4 + HOCl + Cl.C2H4.OH (2) Co., Barclay St., N. Y.). -4silicone grease was used on the ball-
and-socket joints to provide a seal, and the joints were held
The by-product formed in greatest amount is ethylene dichloride, together tightly by clamps. The main vertical column was of
and it may be considered as resulting from the direct addition of 1-inch internal diameter, and the smaller return column was of
G//8-inchinternal diameter.
chlorine to ethylene Operation. Chlorine was introduced into the aqueous reaction
CZHa + Cln -+ Cl.CnH4.Cl (3) liquid near the bottom of the larger column. The chlorine vapor
was dispersed as fine bubbles by a fritted glass diffuser and was
A second by-product is bis(2-chloroethyl) ether (Chlorex). dissolved rapidly. Ethylene was introduced a t a point 17
inches above the chlorine through a similar diffuser, and the
While there are several postulations as to the mechanism for ethylene bubbles contacted the aqueous chlorine solution where
the ether, it may be represented as being formed a t the expense ethylene and hypochlorous acid reacted to form ethylene chloro-
of reactants along with the desired product hydrin in the aqueous medium. Reaction continued as the bub-
bles rose throu h the liquid, and unreacted gases disengaged from
CZHd + HOCl + C1.CnHa.OI-I the liquid in t i e overhead flask and passed out the top of the
-+

Cl.CzH4.0.CzHn.CI + HOB (4) reactor to an indicator solution and to traps.

In continuous operation, water entered the main column a t
I t was the object of this investigation to study the influence of a point just below the chlorine inlet. The water inlet rate
certain reaction process variables for the synthesis of ethylene determined the rate a t which the product was removed from the
chlorohydrin from ethylene, chlorine, and water in a recycle top of the column through the constant-level arrangement shown
in Figure 1. The aqueous chlorohydrin product solution was
reactor and compare the results from a recycle reactor with those collected in the product receiver. The normal capacity of the
from a single column reactor. Plant (IO)has recently described reactor with the recycle arrangement was approximately 3 liters.
a new installation in England that is an industrial application Heating or cooling liquid was passed through the temperature
of the recycle principle in ethylene chlorohydrin production. control jackets that surrounded the upper portions of the two
columns. The rapid circulation imparted to the reaction liquid
by the gas-lift principle facilitated close temperature control
Experimental of the reaction liquid within 1" C. of the desired temperature.
Reactor. The reactor shown in Figure 1 was designed to For operation of the reactor as a single column, without
incorporate some of the features of equipment described by recycle of the reaction liquid, the bottom flask and the smaller
Ferrero and Vandendries ( 5 ) ,Britton et al. ( I ) , and Murray (8). column were removed, and ball-and-socket ground-glass fittings
I t had sufficient flexibility to permit studies of both batch and equipped with stopcocks were used to seal the respective unused
continuous operations, and it could be used either as a single- ball-and-socket joints a t the base and top of the column.
column reactor or as a double-column recycle or circulating-type A flow diagram of the equipment is shown in Figure 2. The
reactor. Because of the highly corrosive nature of the reaction ethylene was 95% pure and was derived from alcohol. Com-
mixture, acidproof materials were used. The body of the reactor mercial grade chlorine was used from a standard 100-pound

April 1954 I N D U S T R I A L AND E N G I N E E R I N G CHEMISTRY 681

ENGINEERING, DESIGN, AND PROCESS DEVELOPMENT

Figure 2. Flow Diagram of Equipment for Synthesis of Ethylene Chlorohydrin

cylinder. The chlorine was passed through a filter that con- these deterininations arc concerned are indicated in the
sisted of glass wool and an alunduni thimble. The chlorine t'ables to show that the ethylene dichloride values are low.
control valve was heated slightly by an electric light bulb to
ensure steady flow of the chlorine. Water traps were provided Several other methods were tried for the determination of
t o prevent the aqueous reaction solution from backing into the the water-soluble organic chlorides, and distillatioii was the
rest of the system in the event of a sudden loss of gas pressure. procedure finally adopted. Ethylene dichloride forms an azco-
The reactor was operated with an excess of ethylene. It was trope with water a t 72" C., and ethylene chlorohydrin, bis(2-
desirable that all the chlorine be used before it reached the top
of the column so that there would be no possibilit,y of an explosion chloroethyl) ether, and water form a ternary nxeot,rope a t 97.3 C.
O

of unreacted chlorine and ethylene in the disengaging zone above Con.i-enientiy, ethylene dichloride and bis(2-chloroethyl) ether
the liquid in the reactor. I n order to have a rapid indication are only slight'ly soluble in wat,er, so it is possible to recovrr
of the appearance of chlorine in the unreacted gases, t,he gases these materials as an immiscible layer by distillation. The
were bubbled through a starch and potassium iodide indicator
solution that gave an immediate color change when chlorine was distillation column x a s operated a t total reflux, and the immis-
present. Two traps were placed in series in the exit gas line t o cible layer was collected in a side arm a t the top of the column.
recover ethylene dichloride stripped from t,he reaction solution. Yields were based on analyzed chlorine. Inasmuch as t,hrb
The traps consisted of graduated vials partly filled with water and dichloride determinations were low for the batch runs, the ethyl-
immersed in an ice bath. The gases were bubbled through the
cold water, which was at a temperature of 2' to 4 ' C., and ene chlorohydrin yields are slight'ly high. However, t'he data are
ethylene dichloride condensed as an immiscible layer in the bot- on a comparable basis, so the qualitative influence of the operat-
tom of the traps. There was seldom any trace of immiscible ing variable8 remains clearly evident'. The yield data for t,'ric
dichloride in the second trap. continuous runs are considered to reprepent actual coiiciitions.
The procedure used t o put the equipment into operation
involved first charging the reactor wit,h water and then charging
ethylene for about 20 minutes before admitting chlorine. The
initial chlorine rates were kept quite low to prevent t,he appear- Discussion of Results
ance of unreacted chlorine in the exit gases. A s the concentra-
tion of reaction products increased, the gas rates were gradually I n t,hc synthesis of ethylene chlorohydrin, the amount of by-
increased until the desired rates were attained. The appear- product's present in the reaction medium affects t'he rate of by-
ance of reaction products facilitated the reaction in t,hat gas dis- product formation. The amount of by-products present at
persion in the aqueous medium improved, and gas charge rates
could be increased without, loss of unreacted chlorine. Rteady state conditions during continuous operation is greater
than \Then the reaction begins a t zero concentration of ethylene
Analyses. The analyses included determination of inorganic chlorohydrin and increases to a given concentration level. Con-
chlorides (hydrochloric acid), ethylene chlorohydrin, and othrr sequently: data from batch operations do not reflect yields
organic chlorides. The latter included primarily ethylene di- a t any given concentration; hoiwver, batch data do provide in-
chloride and a small amount' of bis(2-chloroethyl) ether and formation concerning the qualitative influence of certain variable*.
was reported as ethylene dichloride. The hydrochloric acid Accordingly, a number of variables were studied with batch
and ethylene chlorohydrin were determined mercurimetrically operation of the recycle arrangement, and based on information
( 4 ) . The hydrochloric acid was determined on one portion of a gained from these runs, conditions were selected for a number
sample; a second portion was treatcd with sodium bicarbonate, of continuous runs with both a single column and the recycle
which selectively hydrolyzed the ethylene chlorohydrin, and t>he reacttor arrangements. The final concentration of the product.
combined chloride determination after hydrolysis permitted the in the batch runs M-as generally about 12% et'hylene chlorohydrin.
chlorohydrin content to be obtained by difference. During the Influence of Reaction Temperature. The influence of the
batch runs, the soluble organic chlorides, other than ethylene heat of reaction for the formation of ethylene chlorohydrin iu
chlorohydrin, were obt*ained by a chloride determination aftcr limited largely by the fact that only a dilute solution of chloro-
hydrolysis of a third portion of a sample with sodium hydroxide. hydrin is formed. Some indicat'ion of the temperature rise in
The hydrolysis with sodium hydroxide was later found t o be un- the reaction medium can be obtained from Table I, which shows
reliable in that. it effected only partial hydrolysis of the water- only a small temperature rise for the particular rates of forma-
soluble portion of ethylene dichloride. The data with which tion represented by the data.

682 I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY Vol. 46, No, 4

 UNIT PROCESSES

The data, combined .Ivith recorded physical observations,

Table I. Data from Preliminary Batch Run with Recycle indicate that nothing is to be gained by resorting to low operating
Reactor temperature was not controlled temperatures that require refrigeration. A disadvantage of low
Room temperature = 26' t o 28" C. temperature operation is that ethylene dichloride accumulates in
Final chlorine rate = 36 g r a m d h o u r
Mole ratio, ethy1ene:chlorine = 1.73 Ihe reaction medium as a separate phase. The presence of the
Products, Wt. 7% dichloride contributes to the formation of more dichloride,
Elapsed Reactor C1C1":HC1 because ethylene and chlorine dissolve in the immiscible phase
Time, Temp., chloro- chloric dichlo- Mole R't.
Hours O C. hydrin acid ride@ ratio ratio and react directly.
1 28.8 0.83 0.38 0.987 2.178 It is possible that the yield of ethylene chlorohydrin is slightly
2 30.9 1.82 0.83 0:OOQ 0.990 2.187
3 32.4 2.85 1.31 0.01 0.985 2.174 favored by a reaction temperature approximating 35" C. The
4 33.8 3.91 1.80 0.01 0.984 2.173 results of this work are not conclusive in this respect, but Murray
5 34.4 4.89 2.27 0.02 0.979 2.159 ( 8 ) also obtained slightly higher yields at about 30' C. than a t
6 34.3 5.88 2.72 0.03 0.978 2,159
7 34.3 6.83 3.15 0.03 0.980 2.164 other temperatures.
8 34.0 7.71 3.59 0.04 0.972 2.147
In large scale production, the selection of an operating tempera-
9 33.3 8.66 4.00 0.04 0.980 2.163
10 33.1 9.47 4.42 0.05 0.970 2.142 ture can probably be left to the dictates of convenience. The
11 32.7 10.31 4.82 0.07 0.969 2.139 heat of reaction promotes an increase in temperature, but the
12 32.6 11.12 5.22 0.09 0.966 2.132
final temperature can be controlled to a great extent by the tem-
a Values shown are the low values of ethylene dichloride determined by perature of the feed water. As the temperature of the reaction
hydrolysis with sodium hydroxide.
is raised, increased amounts of ethylene dichloride are stripped
from the reaction medium; this facilitates the recovery of the
Table II. Data from Batch Run, with Recycle, at 13" C. dichloride which can be condensed from the exit gas.
Reactor temperature = 13' & 1' C. Influence of Mole Ratio of Ethylene to Chlorine. A compari-
Final chlorine rate = 36 2 g r a m d h o u r
Mole ratio, ethy1ene:chlorine = 1.73 son TTas made of the influence of mole ratios of ethylene to chlorine
Products, Wt. 7% of 1.21, 1.48, and 2.85, with other variables heid constant. It
Elapsed Ethylene Hydro- Ethylene C1C2H40H:HC1, is readily apparent from the data of Tables 111, V, and VI that
Time, chloro- chloric dichlo- Mole "t.
Hours hydrin acid ride" ratio ratio much more ethylene dichloride is formed a t the low ratio than
1 0.74 0.34 0.003 0.982 2.168
2 1.28 0.59 0.005 0.985 2.175
3 2 07 0.95 0.007 0.986 2,176
4 2.91 1.33 0.01 0.988 2.182
6 4.57 2.12 0.03 0.977 2.158
35" C.
7 5.57 2.58 0.03 0.977 2.157 Table 111. Data from Batch Run, with Recycle, at
8.5 6.89 3.20 0.06 0.975 2.153 Reactor temperature = 35' 1. 1' C.
10 8.22 3.83 0.07 0.971 2.144 Final chlorine rate = 42 grams/hour
11 8.99 4.23 0.09 0.964 2.127 Mole ratio, etliy1ene:chlorine = 1.48
12 9.79 4.59 0.09 0.966 2.132 Products, Wt. "r,
13 10.55 5.00 0.12 0.956 2 110 Ethylene ~1CIHmi:HCI
14 11.39 5.36 0.14 0.962 2.123 Elapsed Ethylene Hydro-
Time, chloro- chloric dichlo- Mole wt.
Yields for Hours hydrin acid ridea ratio ratio
entire run 94.4 98,O 3.8
1.5 1.03 0.48 0.007 0.980 2 161
a No trap was used in gas-outlet line during this run; however, 15 grains 3 2.01 0.92 0.01 0.988 2.181
of ethylene dichloride was recovered as a separate phase from ,the reactor. 4 2.92 1.34 0 02 0.985 2.176
Values shown are t h e low values of ethylene dichloride determined by hy- 5 3.98 1.84 0.02 0.982 2.168
drolysis with sodium hydroxide. 6 5.06 2.35 0.03 0.976 2.155
7 6.16 2.87 0.04 0.973 2.149
8 7.21 3.33 0.04 0.981 2.16,s
4 8.23 3 82 0.03 0 975 2.162
10 9,13 4.27 0.06 0.968 2.138
11 10.07 4.72 0.06 0.966 2.132
Tables 11, 111, and IV give the data for three runs over the 12 10.98 5.16 0.07 0.963 2.127
temperature range 13 O to 50' C. I n the reaction of ethylene, 13 11.83 5.57 0.08 0.961 2.122
chlorine, and water to form ethylene chlorohydrin, Yields for
entirerun 95.2 98.9 2.9

CzH4 + Clz + HOH -+ CI.CnHd.OH + HCI (')

Ethylene dichloride recovered = 1 2 . 3 grams. Values shown are t h e
a
low values of ethylene dichloride determined by hydrolysis with sodium
hydroxide.
there should theoretically be 1 mole of ethylene chlorohydrin for -
each moleof hydrochloric acidformed. However, as the mole ratios Table IV. Data from Batch Run, with Recycle, at 50" C.
indicate, the chlorohydrin is always present in lesser amounts.
Reactor temperature = 30" i: 1' C.
The theoretical weight ratio is 2.208. In explanation of the varia- Final chlorine rate = 41 grame/hour
tion of the mole ratio from unity, there are the possibilities of Mole ratio, ethy1ene:ohlorine = 1 . 4 7
decomposition of hypochlorous acid and of its oxidation to chloric Products, Wt. 7c
Elapsed Ethylene Hydro- Ethylene ClCZH4OH: HC1
acid. There is also the probability of loss of some hypochlorous Time, chloro- chloric dichlo- Mole Wt.
acid and ethylene chlorohydrin through the formation of bis(2- Hours hydrin acid ride4 ratio ratio
chloroethyl) ether, as shown by Equation 4. I n this connection, 1.5 0.92 0.42 0.006 0.983 2.170
3 2.46 1.16 0.01 0.959 2 117
the variation of the mole ratio from unity may be used to pro- 4 3.64 1.72 0.02 0.958 2.117
5 4.76 2.25 0.03 0.959 2.118
vide a rough indication of the amount of the ether formed, but
6 5.87 2.77 0.03 0.981 2.122
because of the influence of othei reactions, it would not provide a 7 6.89 3.28 0.05 0.953 2.105
quantitative measure of the amount of the ether present. 8 7.90 3.76 0.06 0.952 2.102
9 8.88 4.24 0.07 0.949 2.095
The solubility of ethylene and chlorine in water increases as the 10 9,85 4.71 0.09 0.948 2.093
temperature decreases; accordingly, it was anticipated that the 11 10.76 5.15 0.10 0.947 2.091
12 11.03 5.58 0.10 0.944 2.085
reactor would operate more smoothly a t a low temperature than 13 12.50 6.01 0.10 0.942 2.080
a t some higher temperature. However, the opposite to this was Yields for
found to be true. At 13" C., the final chlorine rate was limited entire run 92.2 97.7 5.1

to about 36 grams per hour instead of the desired 42 grams per a Ethylene dichloride recovered = 25 6 grams. Values shown are the
low values of ethylene dichloride determined by hydrolysis with sodium
hour, because unreacted chlorine appeared in the exit gases when- hydroxide.
ever attempts were made to increase the chlorine rate.

April 1954 INDUSTRIAL AND ENGINEERING CHEMISTRY 683

ENGINEERING, DESIGN, AND PROCESS DEVELOPMENT

of a larger amount of inert gas would have reduced somewhat

the amount of dichloride formed by stripping it, from the reaction
mixture. This might have reduced the amount of dichloridtl
sufficiently so that a separate liquid phase would not have formed,
for it is in this organic phase that the formation of more dichloride
is promoted. For a given mole ratio of et,hylene to chlorine
there probably is an optimum concentration of inert gases.
Although some advant,age may be gained by the stripping effwt
I I 1 I I 1 I I2 of the exit gases, the decreased partial pressure of ethylene would
0 25 50 75 100 125 60 175 200 decrease t'he rate of its react'ion with hypochlorous acid. Excess
MOLE PER CENT EXCESS ETHYLENE ethylene itself strips out the dichloride, so that any optimum
Figure 3. Influence of Excess Ethylene on Yields of concentration of inert gases is dependent on such factors as esccss
Ethylene Chlorohydrin and Ethylene Dichloride ethylene, temperature, and others.
A comparison of the data shows only limited advantage for
Reactor temperature, 2 5 ' C.; final chlorine rate, 42 grams per hour an increase froin 48 to 185% excess ethylene, A very large escess
of ethylene n-ould be impractical for large scale production he-
cause of the large volume of gas that n.ould have to be recyclrd.
when ethylene is present in large excess. This is shown graphi- The selection of an optimum ethylene to chlorine ratio must
cally in the qualitative curves of Figure 3. Tables V and VI depend upon an ecomomic balance that talres into consideration
show that a t approximately the Same final ethylene chloro- the purity of the ethylene feed gas.
hydrin concentrations, more than four times as much ethylene
Influence of Gas Flow Rates. The data of Tables 111,VII, arid
dichloride was formed when ethylene was present in 21% excess VIII, and the qualitative curves of Figure 4 indicate t,hat low
than when it mas present in 185% excess. At the high ethylene
gas rates favor t'he formation of ethylene chlorohydrin over
excess, most of the dichloride was formed above the 11% chloro-
ethylene dichloride. 4 t a concent'ration of approsimatcly 100;
hydrin concentration level. ,4comparison of the data in Table3
ethylene chlorohydrin, 4.9 grams of ethylene dichloride was
111 and V indicates that approximately three times as much formed a t the low gas rates while approximately 14 grams >vas
dichloride was formed at the low ratio as was formed with 48y0
formed a t the high rates. At the 1270 chlorohydrin level, 12.3
excess ethylene. grams of dichloride was recovered a t the intermediate rates vhilc
23 grams was recovered a t the high rates. At low gas ratcss,
there are relatively few gas bubbles in any given incremcnt of
volume within the reactor; therefore, a given bubble of ethylcrirl
has a relat,ively large volume of solution in which it can dissolve
to react with hypochlorous acid in the solution. Also, with :I
low bubble density, there is relatively low probability that ethyl-
ene and chlorine bubbles will come into direct contact and rc-
act, to forni ethylene dichloride. In the range of gas flow r:ttm

CHLORINE RATE, GRAMS PER HOUR

Figure 4. influence of Gas Flow Rutes

on Yields of Ethylene Chlorohydrin and
Ethylene Dichloride
Reactor temperature, 35 C.; approximately 50%
excess ethylene 51 I I I
5 6 7 8 9
WEIGHT PER CENT ETHYLENE CHLOROHYDRIN
The ratio of ethylene chlorohydrin to hydrochloric acid de-
Figure 5. Yields from Continuous Operations of
creases relatively rapidly as the chlorohydrin concentrafion
Single-column and Recycle Reactors at Various
increases a t the low ethylene to chloritic ratio. Table T' also
Concentrations of Ethylene Chlorohydrin
s h a m an appreciable amount of ethylene dichloride accumulated
in the reactor as an immiscible phase a t the Ion- ratio; this con- Reactor temperature, 35' C.; approximately 50% excess ethylene
dition is unfavorable since it accelerates the formation of the tli-
chloride by-product.
The large amount of ethylene dichloride as probably formed investigated, after the reactor had attained the desired operatlng
as a result of the retarded reaction of chlorine xith ivater to form conditions, the chlorine dissolved rapidly and there was no notire-
hypochlorous acid, as s h o m by Equat,ion 1. The increase in able difference in the rates of solution of chlorine a t high and l o a
concentration of hydrochloric acid a? the reaction proceeds causes rates of flow. It would be expected, however, that as gas fccti
a natural shift to the left in Equation 1. The added effect of the rates increased, consideration would be given t o the spacing of
presence of a smaller amount of ethylene t o react with hppo- chlorine and ethylene injection points to ensure solution of tlw
chlorous acid resulted in more favorable conditions for direct chlorine.
reaction between ethylene arid dissolved hut unhydrolyzed chlo- An inherent feature of the reactor used in this work provided
rine to form ethylene dichloride. for increased circulationrate as gas flox rates increased. Since the
K i t h 95% ethylene, there \vas only a sniall amount of un- height of the reaction zone was fixed, this resulted in a decieascd
reacted and diluent gas a t the low ratio. Possibly the presence time of contact of ethylene bubblcs mith the rrartion solution.

684 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 46,No. 4

 UNIT PROCESSES

Table V. Data from Batch Run, with Recycle, and with Low Table VI. Data from Batch Run, with Recycle, a n d with
Ethylene to Chlorine Ratio High Ethylene to Chlorine Ratio
Reactor temperature = 35O =t l o C. Reactor temperature = 35' =t l o C.
Final chlorine rate = 42 grams/hour Final chlorine rate = 42 crams/hour
Mole ratio, ethy1ene:chlorine = 1 . 2 1 Mole ratio, ethylene: chlorine = 2 . 8 5
^.--- _"", I , ".2"- Products, Wt. ?' &
Elapsed Hydro- Ethylene .?jCsHIOH:HC1 Elapsed Ethylene Hydro- E t h lene C1CsHdOH:H%
Time,
Hours
Chloro-
hydrin
chlorio
acid
dichlo-
ridea
Mole
ratio
Wt.
ratio
Time, chloro- chlorio dicxlo- Mole y.
Hours hydrin acid rideQ ratio ratio
1.5 0.97 0.44 0,004 0.988 2.183 1.5 1.04 0.48 0.007 0.989 2.181
3 2.51 1 20 0.02 0.946 2.089 3 2.08 0.96 0.02 0.987 2.180
4 3.68 1.79 0.03 0.932 2.058 4 2.75 1.27 0.02 0.984 2.172
5 4.78 2.37 0 04 0.915 2 020 5 3.86 1.78 0.02 0.981 2.166
6 5.82 2.94 0.05 0.897 1.980 6 5.04 2.34 0.02 0.977 2.157
7 6.79 3.51 0.05 0.975 1.931 7 6.18 2.87 0.03 0,977 2.157
8 7.72 4.08 0.06 0.856 1.890 8 7.30 3.39 0.04 0.976 2.156
9 8.58 4 63 0.07 0.839 1.852 9 8.35 3.89 0.04 0.972 2.146
10 9.59 5.11 0 09 0.849 1.875 10 9.39 4.42 0.04 0.963 2.127
11 10.60 5.58 0.10 0.852 1.881 11 10.39 4.89 0.04 0.963 2.126
12 11.36 6.02 0.10 0.854 1.886 12 11.33 5.34 0.05 0.961 2.121
13 12.29 6.45 0.12 0.863 1.906 13 12.17 5.75 0.06 0.958 2.116
Yields for entire run Yields for entire run
85.4 98 8 7.9 95.6 99.6 2.4
a Ethylene dichloride recovered = 1 0 . 0 grams. Values shown are low
(IEthylene dichloride recovered = 42 6 grams, including 29 grams ,with- values of ethylene dichloride determined b y hydrolysis with sodium hy-
drawn from the reactor. Values shown are low values of ethylene dichlo- droxide.
ride determined by hydrolysis with sodium hydroxide.

An upper limit of gas rates mould exist, beyond which unreacted

chlorine would appear in the exit gas stream, but this limit Table VII. Data from Batch Run, with Recycle, and with
could be extended by a constriction placed in the system to con- Low G a s Rates
trol the rate of circulation. Reactor temperature = 35' i 1' C.
From an industrial standpoint, extremely low gas rates are Final chlorine rate = 26 grams/hour
Mole ratio, ethylene: chlorine = 1 . 4 7
impractical since the gas rates directly influence rates of produc- Products, Wt. %
tion. An industrial reactor would be operated a t as high a rate Elapsed Ethylene Hydro- Ethylene CIC&dOH:HC1
of production as could be attained without the formation of ethyl- Time, chloro- ohloric dichlo- &Tole wt.
Hours hydrin acid ridea ratio ratio
ene dichloride in amounts that mould detrimentally affect the 1.5 0.80 0.37 0,005 0.980 2.166
economics of the process. 3 1.89 0.88 0.01 0.978 2.160
4.5 3.06 1.42 0.01 0.974 2.152
6 4.05 1.90 0.02 0.964 2.128
7.5 5.12 2.42 0.03 0.960 2.120
Comparison of Continuous Operation of Single-Column 9 6.19 2.92 0.03 0.960 2.118
and Recycle Reactors 10 6.89 3.24 0.04 0.963 2.127
11 7.57 3.54 0.04 0.968 2.137
12 8 24 3.86 0.04 0.966 2.133
The reactor, shown in Figure I, i w s modified to be operated 13 8.88 4.18 0.04 0,963 2.127
as a single-column reactor with concurrent flow of the inlet water 14 9.62 4.49 0.06 0.960 2.120
and feed gases but without recycle of the reaction solution. 15 10.14 4.81 0.06 0.955 2.109
This represents only one of several possible arrangements for a Yields for entire run
single-column reactor. Table I X shows the results of the last 96.2 100.6 1.6
several hours of three runs at levels of 5.9, 6.5, and 8.4% ethylene a Ethylene dichloride recoiered = 4 9 grams. Values shoivn are low
chlorohydrin in solution. The data for ethylene dichloride con- valups of ethylene dichloride determined b y hydrolysis with sodium hy-
centrations for all of the continuous runs are considered reliable. droxide.
Table IX and Figure 5 show that as the concentration of ethyl-
ene chlorohydrin was increased, the yield, based on chlorine, de-
creased, and the dichloride by-product increased both in con- Table Vlll. Data from Batch Run, with Recycle, and with
centration and yield. I n connection with the physical operation High G a s Rates
of the single-column reactor, it was observed that the chlorine
was well dispersed, but relatively little absorption of chlorine Reactor temperatiire = 35O i 1" C.
Final chlorine rate = 71 grarns/hour
appeared to occur until the bubbles reached the region of ethylene Mole ratio, ethy1ene:chlorine = 1 , 5 0
injection. There appeared to be a high probability of direct
Products, W t . yo _ _
contact between ethylene and chlorine bubbles. IIlapsed Ethylene Hydro- Ethylene C1C2H4OH:HCI-
When chlorine and ethylene were injected into water, the gas Time, chloro- chloric dichlo- Mole "t.
bubbles were large initiallv, and the rate of solution appeared Hours hydrin acid ridea ratio ratio
to be slow. As the concentration of products increased, gas dis- 1.5 0.94 0.44 0.009 0.978 2.158
persion improved, and the rate of solution of the gases appeared 3 2.49 1.16 0.01 0.977 2.158
to increase. In the continuous operation of a single-column 4 4.06 1.89 0.04 0.975 2.153
reactor, there is a concentration gradient between the water 5 5.95 2.77 0.05 0.974 2.160
inlet and the product removal regions, and with an arrangement 6 7.54 3 53 0.06 0.968 2.137
such 8 s has just been considered, the gases are injected in a 7 9.15 4.30 0.06 0.964 2.128
region that has a low concentration of reaction products; this 8 10.63 5.03 0.07 0.958 2.115
9 12.07 6.70 0 11 0.959 2.118
would appear to have some disadvantages. In a recycle arrange- 10 13.38 6.37 0.11 0.951 2.110
ment, such as shown in Figure 1 the concentration of reaction
products is nearly constant at ali points in the reactor because Yields for 10% chlorohydrin
94.5 98.5 3.6
of the rapid circulation imparted to the reaction medium. Thus, Yield8 for 12% chlorohydrin
it appeared that there would be some advantage in the production 93.1 97.1 4.9
of ethylene chlorohydrin in a recycle reactor rather than in a Yields for entire run
91.8 96.4 5.9
single-column reactor of the type considered.
Table X shows data from continuous runs with the recycle a Ethylene dichloride recovered = 32 6 grams: approximately 14 grams
arrangement a t 5.2, 6.3, 6 7 , and 8.2% ethylene chlorohydrin. recovered at 10% chlorohydrin concentration. approximately 23 grama
recovered a t 12% chloroh drin co?centratiok Values shown are low
The yield and ratio data a t the 5.9% level in Table IX and a t values of ethylene dioh1oridTe determined b y hydrolysis with sodium hy-
the 6.3% level in Table X are almost identical. A slight advan- droxide.
tage is shown by the recycle method in that it was operated a t

April 1954 INDUSTRIAL AND ENGINEERING CHEMISTRY 685

 ENGINEERING, DESIGN, AND PROCESS DEVELOPMENT

Table IX. Results of Continuous Operation without Recycle

Reactor temperature = 3.j0 =k 3 O C.
Chlorine rate = 42 grarns/hour
31ole ratio, etliylene:cIilorinc = 1 , 4 7

Sainnle chloro- chlorio dichlo- Mole

S O . hsdiin acld ride ratio ratio
.Iverage Chlorohydrin Concn. = 5 . 9 z 0 . 2 %
I 5 69 2 76 0 87 0 936 2.064
7 6 05 2 YO 0 85 0 943 2 083
3 6 12 2 92 0 78 0 948 2 . 0Q4
4 5 Y2 2 83 0 74 0 951 2,100
.krera.ge yieids
87.9 92.3 Y 7
Averwe Chlorohydrin Concn. = 6 , ; I 0.396
1 6.84 3.58 0.97 0,866 1,911
2 6.46 3 36 0.92 0 871 1.923
3 6.32 3.21 1.04 0.891 1.968
? 6 47
6.73
3.25
3.3'4
0 97
0 93
0.903
0 916
1.993
a 022
AT-erage yields
84 5 95 4 10 1
Ai-erape Chlorohydrin Concn. = 8 . 4 t 0.6y0
1 i.83 4.11 1.80 0.864 1.908
7 8.17 4 23 L49 0.874 1 928
3 8.99 4 63 1.6-1 0 879 1.941
.Ii-erage yields
81.2 92 '3 12.9

Initial ethylene dispersion in wofer Ethylene dispersion during normal op-

with no chlorine injection eration with simultaneous injection of
chlorine; product contains 470 ethyl- 3 to 6 mm. in diameter. The size of bubbles was sorneTThat
ene chlorohydrin
dependent upon gas rates. d slight decrease in bubble size was
Figure 6. Ethylene Dispersion in Single-Column Reactor observed as an individual gas continued t o flow into iyater over a
period of time; the magnitude o i decrease \yas greater for chlorine
than for ethylene.
K h e n the reactor vias put into operation, ethylene was bubbled
a higher level of concentration. The numerical difference in into the distilled water charge for 20 to 30 minutes before chlorine
concentrations is small; however, the recycle reactor rvas able
t o operate at approximately a 770 higher chlorohydrin concen- was introduced. Within a few minutes after chlorine flow had
tration with the same yield. A slight advantage is also shown romnimced, t,herr was a noticeable decrease in the bubble size
by the recycle reactor a t the other concentration levcls.
The batch runs indicated that the ethylene chlorohydrin to
hydrochloric acid mole ratio decreased as the product concen-
tration increased. I n the results reported for continuous runs,
there was in each case a slight increase in male ratio as the time
of reaction a t a given concentration level grew longer. The
decrease in mole ratio in the batch runs results from an increase
in by-product formation a$ the ethylene chlorohydrin concentra-
t,ion increased. In the continuous runs, a rapid decrease in mole
ratio occurred when the water inlet rate as suddenly increased
after the reaction medium had reached the desired level of con-
centration. Then. as this accumulation of by-products was
removed by the product stream, the mole ratio slowly increased
as the reactor approached steady state conditions.
Tables X I and XI1 present, data from a run a t relatively high
gas rates and a prolonged continuous run. h comparison 01
these data indicates that, the lover gas rates givr more favorable
results; this supports the conclusion drawn from batch studies
of this variable. The curves of Figure 5 reflect the increased by-
product formation and decrease in yield of ethvlene chlorohydrin
as the concentration level is increased. As k i t h many of the
other variables, tBheselectmionof a maximum concentration level
for industrial production must depend largely upon the maximum
yield of by-products that can bc tolerated.

Observations of Gas Dispersion

Extensive studies have been reported conccrriing the ioriiicltion
of gas bubbles in liquids. Much of t h c work has: heen concerned
with studies of systems whcre the gases arid liquids have been
mutually inert. Relatively little appears t o have heen done
with systems that involve extensive mass transfer hetween the
gas and liquid phases, and it is this typc of system that is involvrd
in the synthesis of ethylene chlorohydrin.
Initial chlorine dispersion in water Chlorine dispersion during normal
When ethyIene and chlorine indiv-idually were dispersed into operation
distilled water, the et,h>-lcnebubbles initially were in thc range of
4 to 8 mm. in diameter, and the chlorinP bubbles appeared to be Figure 7, Chlorine Dispersion in Single-Column Reactor

686 INDUSTRIAL AND ENGINEERING CHEMISTRY Vel. 46, No. 4

 UNIT PROCESSES

Table X. Results of Continuous Operation with Recycle

Reactor temperature = 35O i l o C.
Chlorine rate = 42 grams/hour
Mole ratio, ethylene :chlorine = 1.47
Products, Wt. %
Hourly Ethylene Hydro- Ethylene . CiC2HI01r: HC1
Sample chloro- ohloric dichio- Mole Wt.
NO. hydrin acid ride ratio ratio
Average Chlorohydrin Concn. = 5 . 2 i 0.1%
1 6 09 2.52 0.65 0.915 2.020
2 5.12 2.53 0.62 0.917 2.026
3 5.21 2.52 0.62 0.936 2.066
4 6.19 2.50 0.68 0,939 2,074
5 5.21 2,50 0.61 0.944 2.086
Average yields
88.2 94.7 8.6
Average Chlorohydrin Concn. = 6 . 3 f 0 . 1 %
1 6.34 3.06 0.82 0.937 2.069
2 6.28 3 03 0.85 0.941 2,077
3 6.28 3.02 0.84 0.943 2.081
4 6.33 3.02 0.84 0.980 2.097
5 6.35 3.01 0.86 , 0.956 2.110
Average yields
87.9 93.0 9.5
Average Chlorohydrin Concn. = 6 . 7 rt 0 . 1 %
1 6.66 3.48 0.98 0.867 1.914
2 6.73 3.45 1.00 0.884 1.952
3 6.68 3.40 0 99 0.891 1,967
4 6.77 3.38 1.09 0.907 2.003
5 6.83 3.39 1.03 0.911 2.012
Average yields
84 5 94 7 10 4 Ethylene dispersion in water ratu- Ethylene dispersion during normal
Average Chlorohydrin Concn. = 8 2 0 1% r a t e d with ethylene but with no operation with simultaneous injection
1.933 chlorine injection of chlorine; solution contains 6%
1.946 ethylene chlorohydrin
1.970
1.991
2.011
Figure 9. Ethylene Dispersion in Recycle Reactor
Average yields
84.0 94.1 11.0 chlorine bubbles appeared to Ise from 0.5 to 1 mm. in diameter.
These bubble dimensions prevailed within 6 to 8 inches above the
points of injection. As a result of mass transfer during reaction,
there was a steady decrease in bubble size as the bubbles rose in
of both gases, and the bubble size continued to diminish as the the column.
concentration of products increased. I n the recycle arrangement, Some coalescence of ethylene bubbles was observed TT hen
at an ethylene chlorohydrin concentration of 6%, the ethylene ethylene was bubbled into pure water, b u t coalescence decreased
bubbles were approximately 1 to 3 mm. in diameter, and the as the concentration of reaction products increased. This con-
forms with observations by Pattle (9) and T'erschoor (15), who
reported that the presence of organic substances or impurities
prevents the coalescence of bubbles through an increase in film
resistance. Tests were conducted to determine whether such a
film resistance might decrease the solubility of reaction gases at
high concentrations of reaction products with a possible contiibu-
tion to the relatively high yields of by-products. The results
indicated that the solubility of ethylene was about the same in
distilled water as in a solution that was saturated with ethl-lene
dichloride and contained 6% ethylene chlorohydrin and 3% hy-
drochloric acid.

Table XI. Results of Continuous Run, with Recycle, at High

Rate and at Concentration of 6.4y0Chlorohydrin
Av. chlorohydrin canon. = 6 . 4 i. 0 . 1 %
Reactor temperature = 35' k l o C.
Chlorine rate = 71 grams/hour
Mole ratio, ethyiene:chlorine = 1.42
Products, Wt. %
Hourly Ethylene Hydro- Ethylene =OH:HCI
Sample chloro- chloric diohlo- Mole TI-!.
h-0. hydrin acid ride ratio ratio

Figure 8. Direct Contact of Chlorine

and Ethylene in Single-Column Reactor

April 1954 INDUSTRIAL AND ENGINEERING CHEMISTRY 687

ENGINEERING. DESIGN. A N D PROCESS DEVELOPMENT

able formation of eddy currents near the wall of the column.

These currents did not flow domn\Tard very far, but they resem-
bled an oscillating motion in relatively localized regions. The
formation of mirls and eddies a t high gas rates WRS undoubtedly
influenced by the relatively m a l l diameter of the column. The
direct contact of riging chlorine liubbles with ethylene in the
single-column reactor is clearlj- evident in Figures 6 and 8.
At the 6% ethylene chlorohydrin level iu the recycle reactor, 95%
of the chlorine was dissolved before chlorine bubbles reached the
ethylene injection point. The rapid rate of chlorine solution
during normal operation is shown a t the right in Figure 10, and
the relatively limited opportunity for direct contact of ethylene
and chlorine bubbles is shown a t the right in Figure 9, where the
region below the ethylene inlet is almost entirely free of rising
chlorine bubbles.

Initial chlorine dispersion in wafer Chlorine dispersion during normal

saturated with ethylene operation; solution contains 670
ethylene chlorohydrin

Figure 10. Chlorine Dispersion in Recycle Reactor

As as indicated by Xurray (8),there is only a limited change

in the surface tension of the liquid as the concentration of re-
action products increases. Tests indicate that the changes
that do occur can be attributed almost entirely to the influence
of ethylene chlorohydrin.
The differences that exist in the dispersion of ethylene and chlo-
rine under different conditions are illustrated in Figures 6 to 11.
Figures 6, 7 , and 8 show the dispersion in a single-column reactor,
and Figures 9, 10, and 11, refer to the recycle reactor. The few
exceptionally large bubbles in Figures 6 and 9 resulted from Figure 11. Effect of Chlorine on
coalescence of bubbles along the top of the side arm-not within Ethylene Dispersion after Complete
the column itself. I n the single-column reactor there wa2 notice- Saturation with Ethylene

Table XII. Results of Prolonged Continuous Run, with A high concentration of reaction products is not the only factor
Recycle that results in excellent dispersion of ethylene. Where the water
.4v. oblorohydrin concn. = G 3 & 0 . 2 %
in the recycle reactor had been saturated with ethylene over a
Reactor teiiiperature = 35' = 1' C. period of 1.5 hours, the injection of only a small amount of chlo-
Chlorine Tare = 42 g ~ t & i i i s / h o ~ rine resulted in an ethylene dispersion that was nearly the same as
Mole ratio, ethylene :chlorine = 1,57
Prodiicts,-J%L% ~- __ that normally obtained a t a concentration of 6% ethylene chloro-
Hocrly Ethylene Hydro- E tbylene B1i?!?El?lsi- hydrin. This is illustrated in Figure 11 which shows that, under
Sample ohloro- ciiloric diciilo- Mole wt, these conditions, the ethylene dispersion in a solution of 0.1%
KO. hydrin acid ride ratio ratio
I 6.11 2.92 0 . 7 3 0.94'2 2.096 ethylene chlorohydrin is nearly equivalent to that shown on the
2 6.20 2.94 0.82 0.95.5 2.108 right in Figure 9.
3 6.30 2 58 0 84 0 557 2.112
4 6 41 3.02 0 7,; 0 960 2.119 Obervations of physical operations suggest that the recycle
6.46 3.05 0.82 0.955 2.118
reactor has advantages over the single-column reactor used in
this work. The excellent disperiion attained is effectively utilized
through rapid dissolution of the reaction gases.
Average Chlorohydrin Concn. = 6 6 & 0 . 1 %
6 6.44 3.06 0.84 0.953 2 104 Conclusions
7 6.49 3 08 0.90 0.956 2.109
8 6.56 3 11 0 89 0.9.55 2.109
9 6.65 3.14 0.86 0.960 2.120 The rate of formation of by-products, including principally
Ayerage yields ethylene dichloride and bis(2-chloroethyl) ether, increases as the
8s 3 52 4 9 6 ... ... concentration of ethylene chlorohydrin increases.
/Average ratios
... ... ... 0 956 2 111 Temperature studies in the range 13" t o 50" C. indicate that a
diqadvantage of low temperature operation is that by-product

688 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 46, No. 4

 UNIT PROCESSES

ethylene dichloride accumulates in the reaction medium where literature Cited

its presence contributes to the formation of more dichloride.
Britton, E. C., Nutting, H. S., and Huscher, M. E. (to Dow
At temperatures of 35" to 50" C., the extent of dichloride ac- Chemical Co.), U. S. Patent 2,130,226 (Sept. 13, 1938).
cumulation in the reactor is reduced by the stripping action of Carius, von L., Ann., 126, 195 (1863).
unreacted gases. Curme, G. O., and Johnston, F., "Glycols," New York, Rein-
High mole ratios of ethylene to chlorine are favorable t o high hold Publishing Corp., 1952.
Domask, W. G., and Kobe, K. A., Anal. Chem., 24, 989-91
yields of ethylene chlorohydrin and to the formation of relatively (1952).
small amounts of by-products. The large quantities of ethylene Ferrero, P., and Vandendries, C. (to Societe Carbochimique),
that must be recycled a t high ratios limit practical ratios t o about U. S. Patent 2,103,813 (Dec. 28, 1937).
50% excess ethylene. Gomberg, h'l., J. Am. Chem. Soc., 41, 1414-31 (1919).
McClellan, P. P., IND.ENG.CHEM.,42, 2402-7 (1950).
With other conditions constant, yields of ethylene chloro- Murray, K. E., J. Council Sci. Ind. Research, 17, 213-21 (1944).
hydrin decrease, and by-product yields increase as gas feed rates Pattle, R. E., Trans. Inst. Chem. Engrs. (London), 28, 21-7
increase. From an industrial standpoint, the maximum rate of (1950).
production must be limited by the degree of by-product formation Plant, J., Ind. Chemist, 28, 259-64 (1952).
Sherwood, P. W., Petroleum Refher, 28, No. 7, 120-4 (1949).
that can be tolerated. Shilov, E. A., Zhur. Priklad. Khim., 22, 7 3 4 4 6 (1949).
Physical observations and data for continuous operation of a Shilov, E. A., and Kanyaev, N. P., J . Gen. Chem. (U.S.S.R.),
single-column reactor and a recycle-type reactor indicate that the 17. 109-16 (1947).
recycle reactor has certain advantages in terms of both yields and Shilov, E. A., and Solodushenkov, S. N., Zhur. Priklad. Khim.,
20, 1270-82 (1947).
smoothness of operation. The rapid circulation imparted to the Verschoor, H., Trans. Inst. Chem. Engrs. (London), 28, 42-6
reaction medium permits the reaction to be conducted a t essen- (1950).
tially a given concentration of ethylene chlorohydrin rather Wurtz, von A., Ann., 110, 125 (1859).
than with a concentration gradient between the top and bottom
of the column. RECEIVED
for review Auguat 10, 1953. ACCEPTED
November 17, 1953

Pyrolysis of Oil Shale in Gas

Combustion Retort
M. W. PUTMAN AND RUSSELL J. CAMERON
Oil-Shale Mine and Engineering Experimenl Sfofion, Bureou of Mines, Rifle, Colo.

Important features of the gas combustion process to recover oil from Green River oil shale are
discussed, and data are cited from a pilot plant (6 tons per day) to illustrate certain operational
characteristics. Condensation as a stable mist allows the oil to be removed from the shale bed
below its dew point. The ratio of gas to solids in the product cooling zone influences strongly
the oil-condensation phenomenon, exerting an effect on the extent to which the oil condenses
as an aerosol and upon the size and stability of the oil-mist particles. Combustion-zone tem-
perature should be above 1400' F. to ensure a rapid combustion rate but less than 1800" F. to
avoid fusing the inorganic constituents in the shale. To recover a high yield of liquid products
in a gas circulation process, a low pyrolysis temperature, about 800" to 900' F., i s very de-
sirable. Low pyrolysis temperatures are readily attained in the gas combustion process.

P RODUCTIOX of liquid fuels from oil shale is not a new

concept. Shale oil was produced from oil shale long before our
modern petroleum industry was established. Before the dis-
fuels a t coats that will make them a feasible petroleum supple-
ment.
Oil shale is a sedimentary rock that contains a solid organic
covery of petroleum in Pennsylvania in 1859 perhaps as many as material known as kerogen. When oil shale is heated t o a
50 oil-shale plants had been erected in the United States for the temperature of the order of 800' F., the kerogen is converted by
production of heating and lighting oils. These enterprises were pyrolysis to shale oil, gas, and a carbonaceous residue. The
unable to compete with the newly discovered petroleum and process of heating the oil shale is known as retorting. The crude
soon were forced to close. shale oil can be refined to produce fuels similar to those derived
In the interim, interest in the production of liquid fuels from from petroleum. The gases and the carbonaceous residue that
oil shale has been closely related to the estimated extent of our remain on the shale are available to supply the energy require-
petroleum reserves. A t times the subject has been merely of ments of the retorting process.
academic interest. However, since World War I1 the consumption The most promising oil-shale reserves in the United States
rate of liquid fuels has increased to such an extent that the occur in the Green River formation of Colorado, Utah, and
United States is now a net importer of petroleum. This trend Wyoming. The principal oil-shale measure is about 500 feet
has revived interest in the production of fuels from oil shale, so thick and has an average assay value of about 15 gallons of oil
that today we find both Government and industry earnestly per ton. The richest part of the 500-foot section is a series of
working toward the development of processes t o produce oil-shale strata commonly referred to as the Mahogany ledge, which

April 1954 INDUSTRIAL AND ENGINEERING CHEMISTRY 689