Preparation and Manufacture
The analysis of the first German shell indicated that the mustard gas contained therein had been prepared by the method published by Victor Meyer (1886) and later used by Clark (1912) in England. It was natural, therefore, that attention should be turned to the large scale operation of this method.
The following operations are involved: Ethylene is prepared by the dehydration of ethyl alcohol. The interaction of hypochlorous acid (HClO) and ethylene yields ethylene chlorhydrin, ClCH₂CH₂OH. When this is treated with sodium sulfide, dihydroxyethyl sulfide forms, which, heated with hydrochloric acid, yields dichloroethyl sulfide. Chemically, the reactions may be written as follows:
CH₃CH₂OH = CH₂ : CH₂ + H₂O
CH₂ : CH₂ + HClO = HOCH₂CH₂Cl
2HOCH₂CH₂Cl + Na₂S = (HOCH₂CH₂)₂S + 2NaCl
(HOCH₂CH₂)₂S + 2HCl = (ClCH₂CH₂)₂S + 2H₂O
Without going into the chemistry of this reaction, which is thoroughly discussed by Gomberg[18] ([see also German Manufacture]), it may be said that this “procedure proved to be unsuitable for large scale production” (Dorsey). As Pope remarks, “That he (the German) should have been able to produce three hundred tons of mustard gas per month by the large scale installation of the purely academic method (of Meyer) constitutes indeed ‘a significant tribute to the potentialities represented by the large German fine chemical factories.’” It is true that a great deal of experimental work was carried out by the Allies on this method, but further study was dropped as soon as the Pope method was discovered.
The first step in advance in the manufacture of mustard gas was the discovery that ethylene would react with sulfur dichloride. While American chemists were not very successful in their application of this reaction, either in the laboratory or the plant, it was apparently, according to Zanetti, the only method used by the French (the only one of the Allies that manufactured and fired mustard gas). The plant was that of the Société Chimique des Usines du Rhone and was started early in March, 1918, with a production of two to three tons a day. In July it was producing close to twenty tons a day. The plant was being duplicated at the time of the Armistice, so that probably in December, 1918, the production of mustard gas by the dichloride process would have reached about 40 tons. Zanetti points out, however, that the process involved complicated and costly apparatus and required considerable quantities of carbon tetrachloride as a solvent. It is for this reason that the Levinstein process would have been a tremendous gain, had the war continued.
About the end of January, 1918, Pope and Gibson, in a study of the reaction originally used by Guthrie, found that the action of ethylene upon sulfur chloride (S₂Cl₂) at 60° yielded mustard gas and sulfur:
2CH₂ : CH₂ + S₂Cl₂ = (CH₂ClCH₂)₂S + S
The reaction at this temperature caused the separation of sulfur; this occurred after the product stood for some time or immediately if it was treated with moist ammonia gas. While this process was put into commercial operation, both in England and America, it offered considerable difficulty from an operating standpoint. The sulfur would often separate out and block the inlet tubes (ethylene). While it is comparatively easy to remove the mustard gas from the separated sulfur by decantation, a certain amount always remains with the sulfur. It is almost impossible to economically remove this, and its presence adds to the difficulty of removing the sulfur from the reactors; the men engaged in this operation almost always become casualties.
Fig. 29.—The Levinstein Reactor
as Installed at Edgewood Arsenal.
It was especially important, therefore, when Green discovered that, if the reaction was carried out at 30°, the sulfur did not settle out but remained in “pseudo solution” in the mustard gas (Pope) or as a loose chemical combination of the monosulfide (mustard gas) with an atom of sulfur (Green). This material has all the physiological activity of the pure monosulfide, while the enormous technical difficulties of handling separated sulfur are entirely obviated by this method of manufacture. To carry out the reaction Levinstein, Ltd., devised the Levinstein “reactor.” The apparatus is shown in [Fig. 29]. The process consists essentially in bringing together sulfur chloride and very pure ethylene gas in the presence of crude mustard gas as a solvent at a temperature ranging between 30-35° C. A supply of unchanged monochloride is constantly maintained in the reacting liquid until a sufficiently large batch is built up. Then the sulfur monochloride feed is discontinued and the ethylene feed continued until further absorption ceases. By controlling the ratio of mustard gas to uncombined monochloride, the reaction velocity is so increased that the lower temperature may be used.
The product thus obtained is a pale yellow liquid which deposits no sulfur and requires no further treatment. It is ready for the shell filling plant at once. The obvious advantage of this method led to its adoption in all American plants started for the manufacture of mustard gas (Edgewood, Cleveland and Buffalo).