[ C O N T R I B U T I O N FROM T H E N O Y E S C H E M I C A L
LABORATORY, UNIVERSITY
OF ILLINOIS]
LEVIXSTEIN MUSTARD GAS. 111. T H E STRUCTURE O F T H E MONOCHLORINATION PRODUCT OF MUSTARD GAS1 REYNOLD C. FUSON
AND
WILLIAM E. PARHAM
Received March 15, 1946
When mustard gas (I) is made by the condensation of ethylene with sulfur moncchloride according to the Levinstein process, the distilled prcduct contains small amounts of 2-chloroethyl 2-chlorovinyl sulfide (IV) as a n impurity (1). For this reason it has been assumed that chlorination of the bis(2-chloroethyl) sulfide occurs in the process. This assumption is based upon the observations of several investigators. When Lawson and Dawson (2) treated bis(2-chloroethyl) sulfide 11ith one equivalent of chlorine at 0", they succeeded in isolating bis(2-chloroethyl) sulfor ium chloride (11). This compound proved to be unstable; it decomposed spontaneously into 2-chl~cethyl1,2-dichlorcethyl sulfide (111), which slowly underwent further deccmpcsition a t room temperature with the evolution of hydrogen chloride and the formation of 2-chloroethyl 2-chlorovinyl sulfide (IV). (ClCHzCH2)ZS -*+ I
C1
(ClCH&H2)2SClg + I1 ClCH2CH&ZHClCH2Cl+ ClCH&HzSCH=CHCl I11 IV
dartori (3) has cited Libermann as authority for saying that sulfur monochloride also reacts with the dichloro sulfide (I)to give 11,while Mann and Pope (4) have reported the isolation of the trichloro compound (111) as a prcduct of the action of sulfur monochloride on bis(2-chlorcethyl) sulfide. Another possikility, which hitherto has not been considered, is that the unsaturated sulfide (IV) may result from the action of a n intermediate in the Levinstein procers. The discovery of a methcd for the synthesis ( 5 ) of 2-chloroethylsulfenyl chloride (V), a probable intern-ediate in the preparation of bis(2chloroethyl) sulfide, opened the way for the investigation of this possibility. Experiment has shown that the sulfenyl chloride and pure mustard gas react with esch other, and that 2-chlorcethyl 2-chlorovinyl sulfide (IV) actually is prcduced along with bis(Schlorcethy1) disulfide (VI). [The disulfide also has been identified as oce of the impurities in the crude prcduct of the con6ensation of ethylene with sulfur monochloride (6)]. Presumably, the intermediate products are 2-chloroethyl 1,2-dichlorcethyl sulfide (111) and 2-chlorcethyl x e r captan. The latter would be expected (7) to react with another molecule of 2-chloroethylsulfenyl chloride to produce the disulfide (VI).
* This paper is based on work done for the Office of Scientific Research and Development under Contracts h'os. OEMsr-300 and OEMsr-48 with the Board of Trustees of the University of Illinois. 482
LEVIR'STEIN MUSTARD GAS.
ClCH:!CHzSCI Ti
483
I11
+ ClCH2CH2SCH2CB2CI + I
+
ClCH2CH2SCHCICHzCI ClCKzCHzSH 111 HC1 ClCH2CHzSC HClC HzCl -+ ClCHzCH2SCH=@ HCl IV ClCHzCHzSCl ClCHzCHzSH + (CICHzCHz)zS2 HCl VI Although it has been accepted geEerally and without proof that the urstable precursor of 2-chlorcethyl 2-chlorovinyl sulfide has the designated structure (111),it became a matter of special interest to establish its identity bey0r.d any reasonable doubt. The only other structure which was considered for the trichloro ccmpound was 2-chlorcethyl 2,2-dichloroethj-l sulfide (VIII). By use of the method of Salzberg and Lazier (8) it has been found possible to produce the corresponding hydroxy compound (VII) from monothioglycol and vinylidere chloride. It was converted to the trichloro compound by the action of thionyl chloride. HOCISzCH2SH CH*=CC12 + HOCH2CH2SCH2CHC12 VI1 socl?+ CICH2CHzSCH2CHCla VI11 The structure of the trichloro sulfide was established by hydrolysis, which converted the compound to the expected (2-hydroxyethy1thio)acetaldehyde(IX).
+
+
+
H20
ClCHzCHzSCHzCHC12 -+ HOCHzCHzSCHzCHO IX The 2,4-dinitrophenylhydrazonederived from this aldehyde was found to be the same as that prepared from (2-hydroxyethylthio)acetal (X), made by the condensation of chloroacetal m-ith the sodium salt of monothicglyccl. CXCHZCH(OC~HE.)Z NaSCHzCHzOH --j H O C H Z C H ~ S C H ~ C H ( O C ~ H ~ ) ~ X The new trichloro compound was characterized further by preparation of the sulfone, the p-toluenesulfilimine, and the morpholyldithiocarbamate derivatives. The stability of this trichloro compound excluded the possibility that it was a precursor of 2-chloroethyl 2-chlorovinyl sulfide; it was possible to dehydrochlorinate it only to the extent of 3% by loag heating with triethylamine. It is interesting that dehydrochlorination of the sulfone occurred with ease; treatment with one mole of triethylamine yielded 2-chloroethyl 2-chlorovinyl sulfone (XI). 0 0
+
r
ClCHzCHzSCH2CHC12
.1
0
-HCl+
-
r
C lCH2 C H2 SC H-CHC 1
1
0 XI
484
R. C. FUSON AND W. E. PARHAM
The sulfone was characterized by comparison with a sample obtained from 2-chloroethyl 2-chlorovinyl sulfide made by the condensation of 2-chloroethylsulfenyl chloride with acetylene (5). EXPERIMENTAL
Reaction of 8-chloroethylsulfeny2 chloride with bis(2-chloroethyl) sulfide.' Sixty-six grams of the sulfenyl chloride was added, during the course of fifteen minutes and at room temperature, to 40 g. of the sulfide. The sulfenyl chloride was decolorized rapidly and hydrogen chloride was evolved. The product was distilled repeatedly at a pressure of 1 mm. until hydrogen chloride ceased to be evolved. Fractionation of this liquid yielded bis(2chloroethyl) disulfide and a compound which is believed t o be 2-chloroethyl2-chlorovinyl sulfide (5). The latter compound was a colorless liquid (n: 1.5458) which gave a reddishbrown color when treated with sulfuric acid. Anal. Calc'd for CdH&Id3: GI, 45.10. Found: C1, 45.15. I-Chloroethyl $,$-dichloroethyl suljidc. In a 500-ml. Vycor flask, fitted with a reflux condenser and a Glas-Col heater were placed 120 g. of redistilled monothioglycol, 318 g. of vinylidene chloride (dried over magnesium sulfate), and 1 g. of benzoyl peroxide. The mixture was irradiated with ultraviolet light for twenty-one hours at a temperature of 40-45". Two additional 0.5-g. portions of the peroxide were added during the irradiation period. The excess vinylidene chloride was allowed to evaporate, and the unchanged monothioglycol (96 9.) was removed at a pressure of 0.3 mm. The residual oil (90 g.), which was not heated above 80', was dissolved in 180 g. of chloroform and the solution washed with water. The solution was dried over anhydrous magnesium sulfate and filtered. The crude 2,2-dichloroethyl 2-hydroxyethyl sulfide was kept in chloroform solution, since in the absence of a solvent i t gradually darkened. The chloroform solution of the hydroxy compound was added, dropwise over a twohour period, t o a well-stirred solution of 73 g. of thionyl chloride in 75 ml. of dry chloroform. Removal of the solvent and fractional distillation of the residue yielded 35.3 g. of 2-chloroethyl 2,2-dichloroethyl sulfide boiling a t 68-69' (0.05 mm.); n: 1.5380. The yield was 11.3% if based on the amount of monothioglycol taken, 47.3% if based on the quantity of monothioglycol actually consumed in the reaction. Anal. Calc'd for CJIiC13S: C, 24.82; H, 3.63; GI, 54.96; S, 16.57. Found: C, 24.74; H, 3.44; C1,54.79; S, 17.22. SuZfilirnine. The trichloro sulfide was treated with Chloramine-T in an acetone-water solution. The product, after three recrystallizations from dilute ethanol, melted, with decomposition, at 157-158'. Anal. Calc'd for CllHl&laNOk&: C, 36.44; H, 3.87. Found: C, 36.58; H, 3.75. Morpholyldithiocarbamate. A solution of 18.8 g. of potassium hydroxide, 29 g. of morpholine, and 60 ml. of absolute ethanol was dropped slowly into a well-stirred solution of 31.5 g. of carbon disulfide in 50 ml. of ether. The reaction vessel was kept in a n ice-bath during the addition. The potassium morpholyldithiocarbamate was isolated by filtration and washed with absolute ethanol; yield 63.6 g., or 94%. Six grams of the trichloro sulfide was dissolved in 100 ml. of ethanol, and 8 g. of the potassium salt in 50 ml. of water was added. The mixture was heated under reflux for sixteen hours, cooled, and kept at 0" for twentyfour hours. The crystals which formed were separated by filtration and recrystallized four times from dilute ethanol; m.p. 77-50'. Anal. Calc'd for C~H&I&'OSJ: C, 33.75; H, 4.77. Found: C, 34.56; H, 4.71. The composition of the product corresponds approximately t o that calculated for the derivative in which one chlorine atom has been replaced.
* This experiment was carried out by Dr. E. W.Maynert.
LEVINSTEIN MUSTARD GAS.
I11
485
Sulfone. Two and two-tenths grams of 30% hydrogen peroxide ww( added slowly t o a solution of 3.8 g. of the trichloro sulfide in 20 ml. of glacial acetic acid. After the introduction of a n additional 5.2 g . of the peroxide solution, the mixture was heated under reflux for thirty minutes, cooled, and poured into ice-water. Two grams of white, crystalline sulfone separated. After three recrystallizations from dilute ethanol, the white needles melted at 73-74'. Anal. Calc'd for C4HrCla02S: C, 21.30;I€, 3.13. Found: C, 21.39; H, 3.36. When a mixture of 4 g. (0.017mole) of the sulfone, 1.8 g. (0.017 mole) of triethylamine, and 50 ml. of absolute ether was heated to boiling, a heavy, white precipitate of triethylamine hydrochloride formed immediately. The precipitate was removed by filtration, the ether evaporated, and the resulting liquid poured into ice-water. The white, platelike crystals which separated were recrystallized from petroleum ether; m.p. 38-39'. A mixture of this compound with an authentic sample of 2-chloroethyE 2-chlorovinyl sulfone (5) (1n.p. 37.5-38.5'.) melted a t 38-39'. Attempted dehydrohalogenation of 2-chloroethyl S,I-dichloroethyl suljide. One-tenth of a mole each of the sulfide and triethylamine were heated in 50 ml. of absolute ether at 37" for nine days. Only 0.003 mole of triethylamine hydrochloride separated. Upon long standing at room temperature the solution turned black, but no more of the hydrochloride separatt:d. Hydrolysis of 2-chloroethyl 2,2-dichloroethyl suljide. A mixture of 23.5 g. of the sulfide and 600 ml. of water was stirred at room temperature for seven days. The insoluble residue (1.9 9.) was shown to be unchanged trichloro sulfide by preparation of the sulfiliniine. Titration of the water solution indicated that 75% of the maximum possible amount of hydrochloric acid had been formed. The (2-hydroxyethy1thio)acetaldehydewas extracted with ether and the resulting solution dried over sodium sulfate. Distillation of the solvent left 7 g. (48%) of the aldehyde as a pale yellow oil; n: 1.5050. It possessed a sweet, esterlike odor, gave a positive fuchsin test, and reduced Tollens' reagent. The 2,d-dinitrophenylhydrazone, formed in the usual way, was recrystallized twice from dilute ethanol; m.p. 74-75". C, 4O.CO; H, 4.03;N,18.66;S,10.67. Anal. Calc'd for CIOHI~N,O&: Found: C,40.12;H, 4.09;N, 18.36;S,10.85. Preparation of (2-hydioxyelhyZthio)acetal. Fifteen and six-tenths grams of redistilled monothioglycol was added to a solution of sodium ethoxide prepared from 4.6 g. of sodium and 1251ml. of absolute ethanol. T o this solution was added slowly, with stirring, 30.4 g. of chloroacetal. The mixture became cloudy but no sodium chloride was precipitated. A small amount of potassium iodide was added, and the mixture was heated overnight at 55', with stirring. After removal of the salt by filtration and evaporation of the ethanol, the residual oil was extracted with dry ether. Evaporation of the ether left the (&hydroxyethylthio)acetaZ as a dark orange oil; yield 15 g. or 38.67' of the theory. A mixture of 3 ml. of the acetal and 5 ml. of dilute hydrochloric acid was warmed, and a aolution of 1.5 g. of 2,4-dinitrophenylhydrazinein 100 ml. of boiling ethanol was added. The solution was shaken for a few minutes, 2 ml. of concentrated hydrochloric acid %'as added, and the mixture was heated under reflux for seven minutes. When the solution was cooled and diluted with water, a large amount of yellow crystals separated; m.p. 6267'. They were purified by repeated recrystallization from ethanol; m.p. 74-75'. A mixture of this compound with the 2,4-dinitrophenylhydrazoneprepared from the hydrolysis product of the trichloro sulfide, showed no lowering of the melting point. SUMMARY
Evidence has been presented in support of the belief that the monochlorination product of mustard gas is 2-chloroethyl 1,2dichloroethyl sulfide. The
486
R. C. FUSON AND W. E. PARHAM
isomeric monochloro derivative, 2-chloroethyl2,2-dichIoroethylsulfide, has been prepared, and the two products have been shown to be different. The 2,2-dichloroethyl compound was n a d e by the addition of monothioglycol to vinylidene chloride. Its structure was established by hydrolysis, which converted it to (2-hydroxyethylthio)acetaldehyde. URBANA,ILL. REFERENCES (1) (2) (3) (4)
(5) (6)
(7) (3)
Chemical Warfare Service and O.S.R.D. Reports. LAWSON AND DAWSON, J . A m . Chem. Soc., 49, 3119 (1927). SARTORI, “The War Gases,” D. Van Nostrand Co., Inc., New York, 1939, p. 231. MANNAND POPE,J. Chem. soc., 121, 594 (1922). FUSON,PRICE,BAUMAN,BULLITT,HATCHARD, AND MAYNERT,J . Org. Chem., Paper I of this series. Reference (1). See also FUSON,PRICE,BURNESS, FOSTER,HATCHARD, AND LIPSCOMB, J. Org. Chem., Paper IV of this series. RHEINBOLDT, Rev. brosil. Chin. Sao Paulo, 4, 169 (1937); Chem. Abrtr., 32, 481 (1938). SALZBERG AND LAZIER, O.S.R.D. Reports.
