T H E J O U R N A L OF I N D U S T R I A L A N D E N G I i V E E R I N G C H E X I S T R Y
Feb., 1919
MANUFACTURE OF METHYLDICHLORARSINE’ B y R. H. UHLINGER
AND
R. V. Coon
Received January 4, 1919
INTRODUCTION
Methyldichlorarsine promised t o be a very valuable “gas” for warfare purposes. The’ manufacture of 1000 t o 2 0 0 0 lbs. was attempted so t h a t tests of its properties could be made. During this experimental manufacture, d a t a were gathered looking toward t h e erection and operation of a large-scale plant, should t h e nature of t h e material warrant such an expenditure. This report is a detailed description of t h e manufacturing process used. CHEMICAL DISCUSSION
Sodium arsenite is prepared b y dissolving arsenic trioxide in caustic soda solution as indicated in t h e following reaction: 6 NaOH AS203 + zNa3AsOa 3HzO The reaction proceeds readily, considerable heat being evolved The solution of sodium arsenite is methylated by adding dimethyl sulfate at a temperature of 8 j o C.: Na3AsO3 (CH3)2S04 --t NarCHaAsO3 NaCH3S04 (I) I n t h e above reaction a obnsiderable portion of t h e dimethyl sulfate is wasted, due t o hydrolysis: (c&)2so4 NaOH --f NaCH3S04 CH30H (2) (CIla)2S04 HzO CH3HSO.i CHaOH (3) The products formed by hydrolysis, as indicated in the preceding reactions, are of no value as methylating agents for sodium arsenite. Some dimethyl sulfate is also lost. by t h e formation of methyl ether. In t h e preliminary runs, dimethyl sulfate was added a t various rates, taking from I t o j hrs., b u t no difference was noticed in t h e yield of disodium methyl arsenite as shown b y analysis. The disodium methyl arsenite is converted to methyl arsenine oxide b y sulfur dioxide as indicated: Na~CH3As03 SO2 CH3AsO NazS04 (4) Sodium bisulfite is also formed with excess of sulfur dioxide acting on t h e excess of sodium hydroxide prese n t ; this is decomposed after t h e reduction, by adding sulfuric acid. zNaHSO8 HzS04 -+- Na2S04 z H ~ 0 zSO2 Methyl arsine oxide is finally converted t o methyldichlorargine b y pas&ng hydrogen chloride gas through t h e mixture, when t h e following reaction takes place: C H ~ A S O zHCl + CHsAsClz H20 (5) T h e bisulfite is decomposed before conducting the final step in t h e process, a s otherwise t h e sulfur dioxide liberated would carry off a large p a r t of t h e final product. The methyldichlorarsine is obtained by distilling from t h e mixture and condensing. Hydrogen chloride is added until a concentration of about 20 per cent hydrochloric acid in excess is obtained; this gives a constant boiling mixture t h a t distils without change
+
+
+
+
+
+
+
+
+
1
----f
*
+ +
+
+
+
+
Published by permission of the Director of Chemical Warfare Service.
10 5
in composition, and tends to prevent reversal of Reaction (5). In adding hydrogen chloride i t is important t o keep t h e temperature above 8 j o C. a t t h e end, in order t o prevent formation of arsenic trichlmide. . A two-layer mixture is obtained in distillation consisting of a lower layer of methyldichlorarsine and a n upper layer of constant boiling hydrochloric acid. This is separated, and any product dissolved in t h e acid layer is salted out by t h e addition of a saturated solution of calcium chloride. The methyldichlorarsine is finally freed from any water, methyl alcohol, and hydrogen chloride dissolved in it by distillation. APP A R A T U S
The reaction vessel consisted of a Ioo-gal. jacketed Pfaudler kettle with enamelled agitator. The cover was fitted with a hand-hole closed by a flange in which a thermometer well was inserted. There was also a sampling device in this opening which permitted sampling while t h e process was going on. The inlet for sulfur dioxide and hydrogen chloride was a n open lead pipe reaching almost t o t h e bottom of t h e kettle and connected t o a Duriron T, t h e openings of which were closed b y Duriron cocks; t h e side inlet of t h e T was used for introducing sulfur dioxide and hydrogen chloride, and t h e top inlet for t h e addition of sodium arsenite solution, dimethyl sulfate, and sulfuric acid. The vapor outlet was similarly fitted with a Duriron T and cocks: t h e upper opening led t o a reflux condenser, t h e side opening t o t h e distillate condenser. The kettle was discharged through a bottom outlet, closed by a Duriron cock. During t h e last month of manufacture, t h e Duriron cocks of this whole system were done away with and acid-proof valves used. They were made of antimony lead by t h e Chemical Equipment Co., of Chicago Heights, Ill. This substitution was necessary because t h e Duriron cocks stuck so badly they were not serviceable. The sodium arsenite container was a roo-gal. iron drum with bottom outlet, mounted above the level of the kettle. S d i u r dioxide cylinders were mounted on a platform scale, and connected b y a manifold t o t h e inlet tube. Dimethyl sulfate was received in carboys, from which i t was siphoned into t h e kettle. The carboy was mounted on a platform scale, placed above t h e level of t h e kettle. Sulfuric acid was received in iron drums containing some 1500 lbs. The acid was pumped, b y means of air pressure, into an iron tank mounted on a platform above t h e level of t h e kettle. From this storage tank it was run by gravity into the kettle. The reflux condenser consisted of a coiled lead pipe, set in a galvanized iron can, and cooled by water or ice. Vapors entered t h e bottom coil of t h e reflux, a sight glass being inserted in t h e line between kettle and condenser. The top coil of t h e condenser was connected t o two sight bottles t o observe any escape of gas from t h e apparatus. The distillate condenser was a coiled I1/t-in. lead pipe set in a galvanized iron can, and cooled by ice.
106
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
Vol.
11,
No.
2
&p
w FIG. APPARATUS
FOR
MANUPACTURE OF
A glass was placed between condenser and carboy, by means of which t h e lower layer of product was separated continuously from t h e acid layer. The hydrogen chloride generator was a so-gal. Pfaudler enamelled kettle with stirrer. The acids were added from receivers, connected t o openings in t h e kettle through lead pipes, containing traps and sight glasses. An iron drum was used for t h e sulfuric acid container and a stoneware crock for hydrochloric acid; Duriron valves were used on each container. The containers were charged from carboys, from which t h e acid was discharged b y air pressure. Spent acid was removed from t h e kettle through a bottom outlet, closed by a Duriron cock. This process worked splendidly until t h e wet gas ate holes around t h e top of t h e so-gal. Pfaudler kettle. We then connected on a n additional generating system composed of a 70-gal. fusion kettle made from 11/2-in. cast iron and arranged t o be heated with an oil burner. These two generation systems were so arranged t h a t either could be used while t h e other was being repaired. The fusion kettle was charged with I O O lbs. of fine rock salt, closed, and heated t o about 3 soo C. Sulfuric acid was then dripped in on t h e hot salt until all available hydrogen chloride had been obtained. Because of our lack of control of
R'IETHYLDICHLORARSINE
temperature, a good deal of foaming took place. This blocked t h e lead pipe line leading t o t h e reaction kettle until a lead t r a p made very similar t o a 60-gal. drum was inserted in t h e line. The gas entered t h e t o p of t h e drum on one side and went out a t t h e t o p on t h e other. Thus any solids carried over b y foaming were left in t h e t r a p . This second system of hydrogen chloride generation worked very nicely until t h e outlet of t h e kettle sprung a leak. T h e trouble encountered was due not t o t h e difficulty of t h e problem b u t t o t h e inadequacy of t h e apparatus provided. The still used for redistillation of methyldichlorarsine was a jo-gal. jacketed Pfaudler kettle. Heating was effected b y circulating oil through t h e jacket; t h e oil was heated in a n open iron drum b y means of a gas burner, and was handled b y a small belt-driven gear pump. The still cover was provided with a charging inlet and thermometer well; t h e vapor outlet was provided with a thermometer and was piped t o t h e condenser. T h e condenser consi,sted of two coiled lead pipes, connected in series, and mounted in galvanized iron cans. The condensate drained from t h e condenser into a receiver, consisting of a carboy.
T H E J O U R N A L O F I N D U S T R I A L A N D ENGILTTEERING C H E M I S T R Y
Feb., 1919
MASUPACTURING
OPERATION
Caustic soda is dissolved in water in a n iron container and arsenic trioxide added. On warm days t h e reaction proceeds nicely with considerable generation of heat. On cold days t h e temperature is raised by inserting a steam hose. When it reaches about 70" C. t h e reaction starts. The solution of sodium arsenite is run into t h e reaction kettle, stirred 15 min., and sampled. The titration of t h e sample of arsenic is described under "Methods of Analysis." T;CTith agitator running, and the reflux condenser packed in ice, dimethyl sulfate is siphoned into t h e kettle directly from the carboy. The mixture is cooled by circulating water through t h e kettle jacket, keeping t h e temperature of t h e mixture at 8 j ' C. The addition requires about one hour, after which the cooling water is shut off, and stirring continued until t h e reaction is complete. This is indicated by the temperature falling and by constant values on titration for arsenic; it rehrs. t o allow t h e reaction t o go t o quires about completion. Throughout t h e methylation, there is slight refluxing of methyl alcohol, and some methyl ether is lost as gas. fA"
omwe/;
T4e?mem&~
FIG.&-APPARATUS
FOR
DISTILLATION OF
10
7
The cock t o t h e reflux condenser is left open t o allow the escape of dimethyl ether and sulfur dioxide, and t h e distillate condenser charged with cracked ice, Steam is turned into t h e jacket of t h e kettle and distillation started. The greater part of t h e product distils a t a temperature of 95" t o 99" c. in t h e liquid; distillation is continued until no more oil is observed in t h e distillate, a t which point t h e liquid temperature is about 106' C. Complete distillation requires 3 t o 4 hrs. The condensate is a two-layer mixture of methyldichloramine and hydrochloric acid, which is separated. The upper layer of acid is mixed with shturated calcium chloride solution and, after removing t h e separated product, is discarded. The product is finally redistilled in t h e 50-gal. Pfaudler kettle. The kettle is charged with t h e product obtained after salting out and separating from t h e acid layer. After packing the condensers well with ice and salt, heating is started, and oil circulated through t h e kettle jacket. The first fraction is made between 80' and 129' C. (vapor temperature) and consists principally of methyl alcohol, with some hydrochloric acid. The specific gravity a t t h e s t a r t is 0.852 and increases until a final gravity of 1 . 7 1 is reached. The receiver is then changed t o receive t h e pure product; this fraction is collected between 129' and 1 3 2 " C. (sp. gr. 1 . 7 1 t o 1.84). The first runnings contain a slight red suspension which can be removed by filtration, b u t the main portion is almost water white. There is a slight amount of gas liberated a t t h e beginning of t h e distillation; this is probably methyl ether. The residue left in t h e still is almost solid, and consists of arsenic trioxide and lead and iron compounds. The distillation requires about 8 hrs. CHARGG
METHYGDICHLORARSINE
PREPARATION O F SODIUM ARSENITE
The mixture i s now cooled t o jo' t o 55 O C. and sulfur dioxide: added from cylinders kept wfrm with steam. When the mixture is completely reduced, there is no further absorption of gas in t h e kettle and gas is evolved from the mixture; t h e addition requires about z hrs. The temperature is next raised t o 6 5 " C. and sulfuric acid siphoned into t h e kettle while cooling (which takes about 2 hrs.). Care must be observed in this process, for if sulfuric acid is added too rapidly the decompo:,ition of t h e bisulfite is so rapid t h a t the charge in the kettle roams out through t h e reflux. Several "busts" of this nature happened during t h e operation of the plant. Hydrogen chloride is now introduced at 7 0 ' t o 85O C. while cooling. The hydrogen chloride is generated in the jo-gal. Pfaudler kettle by mixing salt and hydrocnloric acid, heating t o I O O O C., and adding concentrated sulfuric acid. The addition of hydrogen chloride is continued until the acidity of t h e solution is constant; this is determined b y titration with standard sodium hydroxide solution, using phenolphthalein as indicator, and averages 2 0 t o 2 1 per cent hydrochloric acid. The acidifying extends over a period of 1 2 hrs.
375 lbs. water 128 lbs. caustic soda 84 lbs. arsenic trioxide METIIYLAT~ON-I 28 lbs. dimethyl sulfate. REDUCTION-^^ to 90 lbs sulfur dioxide ACIDIFICATION-63 lbs. sulfuric acid, 66 B6. PREPARATiON OF HYDROGEN CHLORIDE
1
batches {
IIO
lbs. hydrochloric acid,
20'
Be.
salt 2 0 gal sulfuric acid, 64' Be. SALTING OUT-2 gal. water, saturated with calcium chloride. FRACTIONAL DISTILLATION--23Q lbs crude rnethyldichlorarsine from 4 runs. 3
7 j lbs.
YIELDS
A general avcrage of 7 0 lbs. crude per run was obtained during the 3 months' operation. This is quite low because very low yields were secured during t h e first weeks of operation. The addition of hydrogen chloride seemed to have t h e most t o do with t h e yield, and when t h e hydrochloric acid concentration was brought up t o 2 1 t o 2 2 per cent a high yield was generally obtained, Thus a n average of 8 7 lbs. crude was obtained during t h e last 5 consecutive runs. The fractionation of t h e 435 lbs. obtained gave t h e following results:
THE JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
108
MANUFACTURE O F METHYL
Vol.
DICHLORARSINE'
t
Temp. Deg. C. First Fraction.. . . . . . . . . 60-127 Second Fraction.. ....... 123-135 Residue (difference) . . . . . . .
Sp. Gr. 0.82-1.71 1.71-1.84
..... TOTAL ........................................
Yield Lbs. 56 329 50 I
435
Since 3 29 lbs. methyldichlorarsine were obtained, t h e product of t h e last five runs was 32g/435 X 100 = 75.6 per cent pure. This means t h a t 8 7 X 0.756 = 65.8 lbs. of pure methyldichlorarsine were obtained per run. Now the theoretical yield is 134 lbs. when calculated on t h e AszOs as 98 per cent pure. This gives a yield of 6 5.8/134 x I O O = 48.8 per cent of t h e theoretical for the last 5 runs. PRODUCTION CAPACITY
Sixty-five t o 70 lbs. of product per 24 hrs. are obtained in t h e apparatus described. This is limited principally b y t h e rate of production of hydrogen chloride and with a more satisfactory apparatus for t h e latter could be increased considerably. METHODS O F AXALYSIS
ARsENIc-Mix a 5 cc. sample with 2 0 0 cc. water, and neutralize with hydrochloric acid, using litmus as an indicator. Add excess of sodium bicarbonate, and starch. Titrate in t h e cold, with N l r o solution of iodine in potassium iodide. TITRATION
1 Numbers
FOR
refer to order in which chemicals are mixed.
XI,
No.
2
a
COSTS
The cost of chemicals per pound of crude methyldichlorarsine is calculated below: 128 lbs. 84 lbs. 128 lbs. 85 Ibs. 677 lbs. 220 lbs. 150 lhs. 10 lbs.
..................... ...................
caustic s'oda a t 6.5 cts.. $ 8.31 arsenic trioxide a t 10 cts.. 8.40 dimethyl sulfate a t 75 cts.. 96.00 sulfur dioxide a t 10 c t s . . .................... 8.50 sulfuric acid, 66O B6.,at $28 per t o n . . . . . . . . . . 9.50 hydrochloric acid, 20' Be.,a t 3 c t s . . 6.60 salt a t $3.75 per ton ......................... 0.28 calcium chloride a t 2l/4 cts.. . . . . . . . . . . . . . . . . . 0.23
..................
..........
TOTAL .......................................... 65-70 lbs. cost $137.82 Minimum cost = $1.97 per lb. crude Maximum cost = 2.12 per lb. crude The cost of pure product (yield 55-59 lbs.) is: Minimum = $2.33 per lb. Maximum = 2.50 per lb.
DETAILS OF DESIGN REACTION
KETTLE
100-gal. jacketed Pfaudler AgitatorSpeed 65 R. P. M. Gas inlet-1 in. REFLUX CONDENSER
Diameter-13 in. turns of I in. lead pipe
51/2
$137.82
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
Feb., 1919
DISTILLATE CONDENSER
Elyria enamelled--g in. diam. X I O f t . Later used I '/z in. lead-coil condenser HYDROCHLORIC ACID CONTAINER
25-gal. stoneware crock HYDROGEN CHLORIDE GENERATOR
so-gal. jacketed Pfaudler kettle Agitator-Speed 60 R. P. M, STILI,
so-gal. jacketed Pfaudler kettle Oil container-so gal. drum Burner--r in. gas line Condenser--1 in series-13 in. diarn.-Sl/a turns of I-in. lead pipe This investigation was started a t the American University Experiment Station of t h e Bureau of Mines and was continued under t h e Research Division of the Chemical Warfare Service. SWALL SCALEMANUPACTURINQ SECTION RESEARCH DIVISION, C. W. S., U. S. A. EXPERIMENT STATION AMECRICAN UNIVERSITY WASHINGTON, D. c.
MANUFACTURE OF ARSENIC TRICHLORIDE' B y R. C. SMITH Received January 4, 1919
Arsenic trichloride has occupied an important place
as an intermediate in t h e manufacture of several toxic gases in connection with the Research Division, Chemical Warfare Service, a t t h e American University Experiment Station, Washington, D. C. At first three general methods were considered for the preparation of this compound: I-The action of dry chlorine gas on arsenic metal -1-The action of hydrochloric acid on arsenic trioxide g-The action of sulfur monochloride on arsenic trioxide This latter method was selected as t h e most practical because of t h e simplicity of the process, the small amount of apparatus required, the short time necessary t o complete the operation, a n d the high yields of almost pure product. T h e reaction employed in the manufacture is indicated below: z A s ~ O ~ 6SzC13 -3 4AsC13 +.3S02 gS With t h e proper temperature control the reaction proceeds with no difficulty. It is necessary t h a t the arsenic trioxide be as dry as possible. The commercial supplies of raw materials proved satisfactory. The reaction was carried out in a 7 j-gal. unjacketed cast-iron kettle heated directly by a gas burner. Agitation was effected by means of a stirrer operated a t 60 R . P. M. T h e bottom of t h e kettle was provided with a 2-in. outlet closed by a cap. The top was provided with a hand-hole for charging the solid material. A thermometer well was inserted for obtaining t h e temperature of t h e mixture. The vapor during the reaction and the distillation passed through a 2-in. opening in the cover connected t o a T, the outlets of which led t o the condensers. A sight glass and iron stopcock were inserted between t h e reflux condenser and the kettle. A 2-in. iron gate-valve was inserted be-
+
+
1 Published by permission of the Director of the Chemical Warfare Setvice.
I09
tween the distillate condenser and the kettle. The upper end of the reflux was connected t o the bottom of a 4-in. lead tower filled with stones, over which water was allowed t o run; this served t o absorb the sulfur dioxide formed in the reaction. Forty-six pounds of arsenic trioxide were charged through the hand-hole, which was then closed. The kettle was heated until the temperature inside reached g j ' to 100' C., a t which temperature the sulfur monochloride starts t o reflux. This usually took about 4j min. A t intervals of j t o I O min. during this heating 2 0 0 t o 300 cc. sulfur monochloride were added. The heat of reaction is very great and the addition of t h e sulfur monochloride greatly assisted in bringing up the temperature. The sulfur monochloride was run in by gravity from a 30-gal. iron drum mounted on a platform balance above the reaction kettle. This container held enough for one complete run and was filled from t h e supply drums by means of a small Roc0 pump. The sulfur monochloride could thus be weighed as i t was added. A sight glass inserted in the line just before it entered t h e reaction kettle enabled us t o estimate t h e rate of adding the sulfur chloride. Agitation was started as soon as the mixture was sufficiently fluid. When the temperature of t h e mixture reached g j o t o 100' C., the gas was turned off and the sulfur chloride allowed t o run in a t such rate as t o produce a moderate refluxing. The temperature of the mixture rose t o 120' t o 1 2 5 ' C. The sulfur dioxide formed in the reaction passed up through the reflux condenser and was absorbed in t h e scrubber and carried away with the wash water. T h e sulfur chloride was stopped when 89.5 lbs. had been'added, and as soon as refluxing stopped t h e handhole was opened and 138 lbs. more arsenic trioxide added and t h e pot closed again. Heat was then a p plied t o bring the temperature up t o 95' t o 100' C. and the sulfur chloride started again; 158.5 lbs. more sulfur chloride were added t o complete t h e reaction. Care was taken to keep the arsenic trioxide in excess in order t o prevent the product from being contaminated with sulfur chloride. When the reaction was complete, heat was applied t o the kettle, the reflux condenser cut off, the distillate condenser opened, and t h e distillation begun. The distillate was collected in a Io-liter aspirator bottle and drawn off into bottles or other containers. The first 2 5 t o j o lbs. arsenic trichloride was usually slightly colored and was saved to be redistilled. The remainder of the product was water white and averaged g g per cent pure. The temperature of the mixture was 140' C. when the distillation began and did not rise until the product was nearly all distilled. Then t h e temC., a t which temperperature gradually rose t o 200' ature the distillation was stopped. Above this temperature very little distillate comes over a n d i t is yellow in color. T h e distillation required about 3 hrs. The average yield was 300 lbs. arsenic trichloride. The residue in the kettle was discharged after cooling slightly, by unscrewing the cap on the bottom outlet. This residue consisted of molten sulfur containing a
