Contents

SECTION III

EXPERIMENTAL RESEARCHES


CHAPTER IX

THE CHEMISTRY OF WAR GASES ^a

In ordinary life one distinguishes sharply between gases, liquids, and solids; in gas warfare this distinction does not hold, the word gas being used broadly to denote any substance, solid, liquid, or gas, which is dispersed in the air and which irritates the lungs, eyes, or skin.

From the beginning of the great war there was a steady development in gases as the means of defense against them were improved. Since chlorine (the gas used earliest in the war) attacks most substances readily, it can be stopped easily. It was soon found that a cloth steeped in sodium hyposulphite solution and mopped over the nose and mouth gave fairly satisfactory protection against the chlorine, this gas being a respiratory irritant, but it did not protect the eyes. In order to strike at this weak point, the Germans then made use of lacrymators or " tear gases," such as bromacetone and xylyl bromide, which were sent over in shell because they were not very volatile. This form of attack was met first by the use of a hood with eyepieces; later, by the introduction of regular gas masks. The hoods could be impregnated with sodium hyposulphite or other substances. The next move on the part of the Germans was to find a toxic gas which was less readily stopped than chlorine, and they made use of phosgene. This was not volatile enough to be used by itself in cylinders, and consequently it was mixed with chlorine. To stop this the British hoods and the first French masks were steeped in a solution of sodium phenolate and urotropin (hexamethylene tetramine). The British before long changed from the hood, or helmet, as it was called, to an impervious mask with a box respirator or canister attached. The air came through the canister, which contained chiefly soda-lime granules and charcoal, both of which stop gases much more effectively than the solutions in the fabric of the helmets.

There being no other gases suitable for use in cylinders, the Germans were now forced to use substances which were fired in shell and which were scattered by the explosion of the booster charge in the shell. The use of chloropicrin was the next step in advance. This substance may be described as an all-round gas, having associated advantages and disadvantages from the offensive point of view. It is fairly toxic and moderately lacrymatory. It causes vomiting and therefore makes it difficult for a man to keep on his mask, It is not stopped by soda-lime, and it had the great advantage of not being stopped well by the charcoal in use in the early part of 1917. An improvement in the quality of the charcoal was necessary in order to stop chloropicrin, and this was accomplished by the Allies. This improvement removed all danger from chloropicrin and made the mask so good that it stopped practically all gaseous substances fairly well.

^a The data in this chapter are based, in the main, on " History of the Chemical Warfare Service in the United States," May 31,1919, Part I, by Lieut. Col. W. D. Bancroft, C. W. S., Research Division, Chemical Warfare Service, American University Experiment Station.⁰²¹/₂ On file, Chemical Warfare Service, Munitions Building,

296

The introduction of the so-called mustard gas in July, 1917, changed the whole state of affairs again. This is a liquid boiling at about 217^o; it attacks the skin, causing bad burns, which may incapacitate a soldier for a number of weeks. Special ointments and special clothing were devised as a protection against mustard gas; but these were not really satisfactory at the time the armistice was signed, and the best protection had been found to be not to keep troops long in gassed areas. That the gas-mask canister did not stop tobacco smoke was well known, and it had been found necessary to put in cotton wads in order to keep stannic chloride smokes from getting through it. The Germans took advantage of this fact and developed their so-called sneezing gas. This is a high melting solid, diphenylchlorarsine, which is dispersed, by means of high explosives, as a very fine smoke. Protection against it is provided by supplying the canister with suitable filtering pads.

A brief account will now be given of the more important gases used by the Germans and this will be followed by a discussion of some of the war gases developed by the Allies.

GASES USED BY THE GERMANS

CHLORINE

Chlorine, Cl ₂, was first used in April, 1915. It is a greenish-yellow gas with a suffocating and irritating smell. It boils at 33.6°; the vapor pressure of liquid chlorine is 3.66 atmospheres at 0^o and 11.5 at 40^o. The molecular weight is 71, so that the vapor is nearly 2.5 times as dense as air (71: 28.8). The density of liquid chlorine is 1.4685 at 0^o and 1.4108 at 20^o It is easily prepared and easily liquefied. It is so volatile that it can be used in a cylinder or cloud attack. To prevent the cooling of the cylinders the discharge tube runs to the bottom as in a soda siphon and evaporation takes place outside the cylinder.

Chlorine is not very toxic, the lethal concentration being 2.5 mg. per liter (770 p. p. m.) for dogs on 30 minutes exposure. It is very corrosive but reacts so readily with most things that it is easily stopped. In the canister it reacts direct with soda-lime. It is also absorbed by charcoal and reacts with moisture, according to the equation 2Cl₂+2H₂O=2HCl+2HClO=4HCl=O₂, the hydrochloric acid being taken up by the soda-lime. Dry chlorine does not react with iron and can therefore be kept in steel cylinders. It is soluble to about 10 percent in carbon tetrachloride. One volume of water absorbs about 2.6 volumes of chlorine at 760 mm. (reduced to 0^o). In aqueous solution there is a reversible hydrolysis represented by the equation Cl₂+ H₂O=HCl=HClO. Light, charcoal, and certain catalytic agents cause the decomposition of hypochlorous acid, 2HClO=2HCl+O₂.

Chlorine is prepared by electrolysis of an aqueous solution of sodium chloride.

PHOSGENE

Phosgene, COCl₂, was first used in December, 1915. It is a colorless gas with a smell like musty hay. It boils at 8.2^o, and the vapor pressure of liquid phosgene is 1.6 atmospheres at 20^o and 3.1 at 40°. It is over three times as dense as air. The density of the liquid is 1.432 at 0^o. It is not

297

sufficiently volatile to be used by itself in cloud attacks and is always mixed with chlorine in such cases, the mixture usually containing 20 to 25 percent phosgene. It is about seven times as toxic as chlorine, the lethal concentration for dogs on 30-minute exposure being 0.35 mg. per liter (80 p. p. m.).

Phosgene reacts readily with water according to the equation COCl₂+ H₂O=CO₂+2HCl. Although this reaction is not reversible, phosgene is very stable when in contact with concentrated hydrochloric acid. The English physiologist, Barcroft, has found that under the conditions of his experiments, phosgene vapor is hydrolyzed only to about 10 percent in the presence of an excess of water vapor, even though the reverse reaction does not take place to any measurable extent. It seems probable that phosgene and water react chiefly, and perhaps solely, at the surface of the containing vessel and that the reaction comes practically to a standstill when the surface becomes covered with a film of hydrochloric acid of sufficient concentration. Phosgene reacts readily with ammonia, aniline, hexamethylene tetramine, pyridine, and many other organic compounds. When heated to 300° or so, it dissociates to some extent into carbon monoxide and chlorine. It does not react with cyanogen chloride, and the two substances can be separated by fractional distillation. The data on the corrosion of metals are contradictory, probably owing to differences in the phosgene used. Steel and Monel metal stand up well in most experiments; aluminum is resistant to phosgene containing traces of chlorine, but does not make a good showing when there is 25 percent chlorine. Lead is usually attacked readily. On the other hand, one set of experiments showed that shell steel lost over 10 times as much as lead when submerged for 30 days at room temperature. There are apparently no data on the corrosion of metals by the mixtures used in cylinder attacks.

Phosgene can be detected by the color change in filter paper treated with dimethylaminobenzaldehyde and diphenylamine. The paper changes to yellow and then to orange with increasing concentration. When used according to directions, it will detect 1 part of phosgene per 1,000,000 of air.

In the canister, phosgene is absorbed by charcoal and reacts with the moisture in the latter to form carbon dioxide and hydrochloric acid, which are taken up by the soda-lime. Soda-lime does not absorb or decompose phosgene sufficiently rapidly to give adequate protection. The charcoal gives the activity and the soda-lime the capacity. Increased moisture in the charcoal increases its efficiency toward phosgene.

Phosgene is made by the combination of carbon monoxide and chlorine in the presence of charcoal as a catalyzer, CO+Cl₂=COCl₂.

CHLOROPICRIN
Chloropicrin, CCl₃NO₂, is a colorless liquid, boiling at 112°, and having a vapor pressure of 5.8 mm. at 0^o, 14.0 mm. at 15°, and 23.8 mm. at 25^o. The vapor is nearly six times as dense as air. The density of the liquid is 1.6924 at 4^o and 1.6539 at 20^o, the two determinations not being made by the same man. The melting point is 69.2^o. Chloropicrin is not sufficiently volatile for use by itself in cloud attacks. While it has been used mixed with 75 percent chlorine, it was usually fired in shell. It is moderately toxic, 0.8 mg. per liter (110 p. p. m.); somewhat lacrymatory, 0.016 mg. per liter, and liable to cause vomiting, thus forcing removal of the mask. It was not stopped

298

satisfactorily by the charcoal first used in the masks. The laboratory charcoal eventually employed was about one thousand times as effective as the earlier material. Chloropicrin is practically nonmiscible with water, and a mixture of the two boils at about 84^o. It is miscible in all proportions with many organic solvents. There is a marked evolution of heat when it is mixed with methyl alcohol, ether, or acetophenone; a slight evolution of heat when mixed with isobutyl alcohol, isoamyl alcohol, or carbon bisulphide.

Chloropicrin is not hydrolyzed by water, nor by cold hydrochloric, sulphuric, or nitric acid. When heated with these acids it is said to distil unchanged. Dilute aqueous sodium hydroxide does not attack it; but alcoholic sodium hydroxide decomposes it slowly, and sodium ethylate attacks it fairly readily, forming the orthocarbonic ether, CC1₃NO₂ +4C₂H₅ONa=C(OC₂H₅)₄+3NaCl+NaNO₂. Chloropicrin can be heated for several days with aqueous ammonium hydroxide at 100^o without undergoing any appreciable change. At 150^o, or when heated with alcoholic ammonia, a reaction takes place in a few hours, guanidine being formed, HN: C: (NH₂) ₂. Alcoholic potassium acetate decomposes chloropicrin completely at 100^o and alcoholic potassium cyanide reacts at a lower temperature, the product in this last case having the formula (CN)₂C(NO₂)₂-Cl. Though chloropicrin is attacked very slowly by dilute aqueous sodium hydroxide, it unites readily with neutral potassium sulphite, CC1₃NO,+3K₂SO₃+H2O=CH(NO₂)(S0₃K)₂+3KCl+KHSO₄. This reaction is the basis of a quantitative method for determining the concentration of chloropicrin vapor in air. A definite volume of air is passed through a neutral solution of sodium sulphite and the resulting amount of sodium chloride is determined. The data on the corrosion of metals are conflicting, but dry chloropicrin apparently attacks steel but slightly and copper and lead considerably more.

Chloropicrin may be detected by its giving a pink color with a suitably prepared solution of alpha-napthol or a blue color with a different solution of beta-napthol. A flame test with copper gauze may be used also, the appearance of a green flame showing the presence of chlorine in the flame. This is a general test and not a specific one for chloropicrin. A concentration of one in a million can be detected by passing the air through a sodium ethylate solution and testing for sodium nitrate.

TRICHLORMETHYLCHLOROFORMATE

Trichlormethylchloroformate, ClCO₂CCI₃, is called diphosgene by the British, surpalite by the French, and superpalite by the Americans. It is a colorless, mobile liquid with a fairly pleasant sweet odor. It boils at 128^o and has a vapor pressure of 2 to 4 mm. at 0^o and of 10.3 mm. at 20^o. The vapor is over seven times as dense as air and is twice as dense as phosgene. The density of an impure sample of the liquid is 1.687 at O^o and 1.656 at 20^o. Owing to the low volatility superpalite was used only in shell and only by the Germans. The shell usually contained mixtures of superpalite and phosgene, though some duds have been found containing superpalite and chloropicrin. Diphenylchlorarsine also has been found in some of the green cross shell.

299

The lethal concentration for dogs is 0.35 mg. per liter for exposure of 30 minutes (40 p. p. m.), but is much lower on longer exposures. Superpalite owes its importance to its high vapor density, to its persistency because of its high boiling point, and to the increased toxicity on long exposures.

Superpalite is hydrolyzed slowly by water at room temperature and fairly rapidly at 100^o, the products being HCl and CO₂ presumably according to the equation, ClCO₂CCl₃+2H₂O=4HCl+2CO₂.

Boiling with an aqueous solution of sodium hydroxide for half an hour decomposes it completely. Heated by itself to 300^o, it is said to decompose into phosgene, but this may be the result of a catalytic action. Superpalite reacts with methyl alcohol in the cold to give trichlormethylmethoxyformate: ClCO₂CCl₂+CH₃OH=CH₃OCO₂CCl₃+HCl.

On long boiling with methyl alcohol the methoxyformate reacts according to the equation, CH₃OCO₂CCl₃+CH₃OH=2CH₃OCOCl+HCl.

Ammonia reacts rapidly with superpalite vapor forming ammonium chloride and urea, ClCO₂CCl₃+8NH₃=4NH₄Cl+2CO(NH₂)₂.

Alumina causes superpalite to decompose into carbon tetrachloride and carbon dioxide, while iron oxide and charcoal decompose it to phosgene,

ClCO₂CCl₃=CCl₄+CO₂
ClCO₂CCl₃=2COCl₂

Some preliminary experiments seem to indicate that in a sealed tube at constant temperature, the decomposition of superpalite in the presence of iron oxides does not run to an end, even though the reverse reaction does not take place. This raises the question whether the Germans really put a mixture of superpalite and phosgene into their shell or whether the extremely variable concentration of phosgene may be due to the catalytic decomposition by the steel shell. There are no experiments as yet to show what effect chloropicrin has on this decomposition. In the canister, superpalite is decomposed by the charcoal to phosgene, which is then decomposed by moisture. Superpalite is also decomposed readily by soda-lime.

Superpalite was probably made in Germany by chlorinating methylformate to methylchloroformate and then chlorinating this to superpalite.

HCO₂CH₃+Cl₂=ClCO₂CH₃+HCl.
ClCO₂CH₃+3C1₂ = ClCO₂CCl₃+3HCl.

While superpalite was not used by the Allies as a war gas, it has been prepared in this country for testing purposes by the action of phosgene on methyl alcohol, giving methyl chloroformate, which is then chlorinated to superpalite.

COCl₂+CH₃OH=ClCO₂CH₃+HCl.
ClCO₂CH₃+3C1₂=ClCO₂CCl₃+3HCl.

In the first stage, a possible side reaction is COCl₂+2CH₃OH=(CH₃)₂CO₃+3HCl, which has no toxic value and which has been thrown away in the past, although it can be decomposed by prolonged heating into superpalite and phosgene, C(Cl₃)₂CO₃=ClCO₂ CC1₃ + COC1₂. The chlorination of methyl chloroformate to superpalite takes place when the heated liquids exposed to intense light while the chlorine is passed in. Nitrogen-filled lamps may be used as the source of light.

300

BROMACETONE

Bromacetone, CH₂ BrCOCH₃, is a water-clear liquid which turns brown or black on standing, owing to charring. It boils with decomposition at about 126°, and the vapor pressure of the crude product is 1 mm. at 10^o and 9 mm. at 20^o. It is not quite five times as dense as air. The density of the liquid is given as 1.631 at 0^o and 1.603 at 20^o. It causes lacrymation at 0.0013 mg. per liter (0.22 p. p.m.) when pure, and at 0.0011 mg. per liter when containing 20 percent chloracetone (Martonite), although the chloracetone is a poorer lacrymator than bromacetone. Being fairly volatile and readily decomposed, it is classed as a nonpersistent lacrymator. Troops can advance a few hours after the shelling. Bromacetone attacks steel and most other metals and must be used in shell lined with lead, glass, or enamel.

Bromacetone is only slightly soluble in water, but readily miscible with alcohol and acetone. Traces of water stabilize the product somewhat, and addition of chloracetone seems to have the same effect. There is some reason to believe that the instability is due to the presence of some impurity, but it is not known what impurity has this effect. Both the charcoal and the soda- lime in the canisters stop bromacetone.

Bromacetone can be made by passing bromine into acetone to which small pieces of marble have been added and then shaking with water and separating the bromacetone layer, which is afterwards distilled with steam. It can be made also by adding bromine dissolved in a saturated solution of potassium bromide to the acetone, or by adding bromine to a solution of acetone in 15 percent sulphuric acid. The product usually contains some dibromacetone. These methods have not gone beyond the laboratory stage in this country. The French have manufactured a mixture of about 80 percent bromacetone, and 20 percent chloracetone, which they call Martonite. In order to prevent the loss of half the bromine as hydrobromic acid, they add a mixture of sodium chlorate and sulphuric acid to oxidize the hydrobromic acid. The reaction is as follows: 5CH₃,COCH₃+4Br+H₂SO₄+NaC10₃=4CH₂BrCOCH₃+CH₂ClCOCH₃+ NaHSO₄ +3H₂O.

BROMMETHYLETHYLKETONE

Brommethylethylketone is a mixture of CH₂BrCOC₂H,, boiling at 145^o, and of CH₃COCHBrCH₃, boiling at 133^o. It is made by brominating methylethylketone, CH₃COCH₂CH₃. It lacrymates at 0.009 mg. per liter (1.3 p. p.m.) and is substituted for bromacetone solely on account of shortage of acetone. Shell must be lined to prevent corrosion.

XYLYL BROMIDE

Xylyl bromide, CH₃C₆H,CH₂Br, is a mixture of the ortho-, meta-, and para-compounds, and boils at about 212 ^o. It lacrymates at 0.002 mg. per liter (0.25 p. p. m.) and is classed as a persistent lacrymator. Lined shell must be used. Both the charcoal and the soda-lime stop xylyl bromide. The mixture of the three xylenes, which is ordinarily called xylene, is heated, exposed to sunlight, and treated with bromine. Under these conditions the bromine substitutes in the methyl side chain and not in the benzene ring. Care must be taken not to carry the bromination too far, as the dibromide is of no value. Xylyl bromide is sometimes called toluyl bromide, because the bromine substitution compound of toluene is called benzyl bromide.

301

DICHLORETHYLSULPHIDE (MUSTARD GAS)

Mustard gas, so-called, (CH₂ClCH₂)₂S, melts at 14.2^o, when very pure, to a colorless, oily liquid which boils at 217^o at 750 mm. The name "mustard gas" was given to it by-the British soldiers, and is an unfortunate one because the compound has nothing to do with what the chemist calls mustard oil. The vapor is a little less than six times as dense as air. The vapor pressure of a sample melting at 13.8° was about 40 mm. at 140^o, 30 mm. at 120^o to 125^o, 20 mm. at 111^o, and 12 mm. at 97°. Some British data are 44 mm. at 128° to 132^o and 10 mm. at 109^o. The density of the liquid referred to water at 0^o is 1.2790 at 15^o, 1.2686 at 25^o, and 1.2584 at 35^o. Owing to the low vapor pressure the substance can be used only in shell and is very persistent. The Germans marked their mustard gas shell with a yellow cross.

The lethal concentration is 0.05 mg. per liter (7 p. p. m.). The liquid produces burns which appear 4 to 12 hours after exposure and heal very slowly. The vapor also causes burns, but to a much less extent. It attacks the eyes, causing conjunctivitis and temporary blindness. The percentage of deaths was rather low in mustard gas cases, only about 5 percent; but this is not due to any low toxicity. It is because the number of casualties due to mustard gas burns was very large. A comparison, from this point of view, with phosgene, which does not burn the skin, is therefore quite improper.

Mustard gas is very slightly soluble in water, less than 0.1 percent. It is freely soluble in alcohol, ether, chloroform, tetrachlorethane, chlorobenzene, and trioxymethylene. It is miscible in all proportions with ligroin above l9^o and with kerosene above 25.6^o. At a pressure of 760 mm., 100 volumes of dichlorethylsulphide dissolve 182 volumes of ethylene at 15^o and 10^ovolumes at 95^o. At 0^o mustard gas dissolves about 3 percent of dry hydrochloric acid. At room temperature it dissolves about 1 per cent of sulphur, the solubility becoming about 6 percent at 100. Dichlorethylsulphide is hydrolyzed very slowly by cold water and quite rapidly by hot water to thiodiglycol, which is harmless, (CH₂ClCH),S+H₂O=(CH₂OHCH₂)₂S+2HCl.

Sodium perborate, hydrogen peroxide, and the dry peroxides of zinc, magnesium and sodium have only a slight effect upon dichilorethylsulphide. Sodium and ammonium sulphides react slowly in the cold, more rapidly upon warming. Calcium, sodium, and potassium hypochlorites, when present in excess, react quickly with evolution of heat. Dry bleaching powder was used by the Germans to destroy mustard gas on the ground. The sulphur is oxidized only partially to sulphate, a water-soluble sulphur compound being formed as well. Potassium permanganate reacts with mustard gas. In acid solutions about four atoms of oxygen are used up by each molecule of the sulphide. Concentrated nitric acid oxidizes the dichlorethylsulphide to the sulphoxide (CH₂ClCH₂)₂SO, melting at 109.50^o while fuming nitric acid carries it to the sulphone (CH₂ClCH₂)₂SO₂, melting at 54^o.

Zinc and acetic acid or aluminum powder and sodium hydroxide destroy dichlorethylsulphide very rapidly, but sulphur dioxide, sodium thiosulphate, and sodium hydrosulphite do not react. Chlorine reacts readily, giving the symmetrical tetrachlorosulphide, which is not irritant. Sulphur diehloride reacts rapidly with mustard gas, forming the tetrachlorosulphide. It is this property which makes sulphur dichloride such a valuable reagent in the laboratory for removing mustard gas. The reaction also takes place in carbon tetra-

302

chloride solution and more slowly the more dilute the solution. Sodium sulphide solution reacts, forming the ring compound S:(CH₂CH₂)₂: S. Chloramine-T(p-toluene sodium sulphochloramine) and dichloramine-T(p-toluene sodium sulphodichloramine) react vigorously with mustard gas, forming white, crystalline compounds which are not irritating.

At ordinary temperature pure mustard gas has practically no action on aluminum, zinc, tin, lead, copper, bronze, or steel. At 100^o aluminum, lead, and brass are not attacked appreciably, while copper, bronze, and steel are corroded slightly, and zinc and tin are attached rapidly. Mustard gas can be detected by smell at about 1 part in 3,000,000; the selenious acid test is sensitive to about 1 part in 1,000,000; while a flame test has been developed which is sensitive to 1 part in 10,000,000, but is not specific, being a test for chlorine.

The Germans made mustard gas by the chlorhydrin method. Chlorine and water react to form hypochlorous acid, which combines with ethylene to give chlorhydrin, C₂H₄+Cl₂+H₂O=CH₂ClCHOH+HCl.

The chlorhydrin reacts with sodium sulphide to form dihydroxyethyl sulphide, 2CH₂ClCH₂OH+Na₂S=(CH₂OHCH₂)₂S+2NaCl. On treating with hydrochloric acid, mustard gas is formed according to the equation, (CH₂OHCH₂),S+2HCl=(CH₂ClCH₂)₂S+2HCl.

The Allies made mustard gas by the sulphur chloride method. Gaseous ethylene is passed into liquid sulphur monochloride contained in large iron reaction vessels, which are usually lead lined. The reaction occurs spontaneously with evolution of much heat. Sulphur is set free and the temperature must be controlled carefully in order to keep this sulphur in colloidal suspension and thus to prevent its precipitation in the solid form in the reacting vessel and the connecting pipes. The equation for the reaction may be written: 2C₂H₄+S₂Cl₂=(CH₂ClCH₂)₂S+S.

It is probable, however, that sulphur monoehloride dissociates to a very slight extent into sulphur and sulphur dichloride, S₂Cl₂S+SCl₂, and that the dichloride is the substance which reacts with the ethylene. It is certain that the reaction takes place in two stages and it is probable that the intermediate product is CH₂ClCH₂SCl. The colloidal sulphur can be precipitated with ammonia if desired.

DIPHENYLCHLORARSINE

Diphenylchlorarsine, (C₆H₅)₂AsCl, is asolid melting at about 44^o and boiling at about 330^o. The vapor pressure is 25 mm. at 233^o and 7 mm. at 180^o. The density of the vapor is about nine times that of air. The density of the crystals is 1.4223 at 15^o. It was used in shell in the presence of high explosive which scatters it as a very fine powder or smoke in the air. The Germans marked these shells with a blue cross. The lethal concentration is about 0.1 mg. per liter, but the substance is used chieflv to cause sneezing and thus to force removal of the mask, and is often called "sneeze gas." It can be detected at 1 part in 100,000,000, produces nasal irritation at 1 part in 50,000,000, and is intolerable at 1 part in 1,000,000, attacking the eyes as well as the respiratory tract. It was first used by the Germans in July, 1917. It is not soluble in water or ammonia, but is readily soluble in alcohol, ether, or benzene. It is hydrolyzed by water to (C₆H₅)₂AsOH and is oxidized by concentrated nitric acid to diphenyl arsenic acid. Chlorine destroys the irritating effect of diphenylchlorarsine, probably due to formation of (C₆H₅)₂AsCl₃, or (C₆H₅)₂AsOCl.

303

though it is possible that the chlorine either causes the agglomeration of the smoke and causes it to be stopped by the canister. Phosgene has no such effect but phosgene containing 1 percent of chlorine does, and it is stated that a marked increase in the capacity of the cloud is noticed.

The vapor of diphenylchlorarsine is stopped by charcoal; and the suspended substance by special filters or otherwise.

It is not known how the Germans made this compound, but it is probable that the reactions are the same as those made use of in this country and in England; the formation of triphenylarsine from sodium, chlorobenzene, and arsenic trichloride in presence of benzene, and the conversion of triphenylarsine and arsenic trichloride into diphenvlchlorarsine by heating in an autoclave.

6Na+3C₆H₅Cl+AsCl₃=(C₆H₅)₃As+6NaCl
2(C₆H₅)₃As+AsCl₃=3(C₆H₅)₂AsCl

While the Germans used a large number of other gases in small amounts, the list just given includes all the really important ones, and it will now be desirable to discuss a few substances which were used or developed by the Allies.

GASES DEVELOPED BY THE ALLIES

BROMBENZYL CYANIDE

Brombenzyl cyanide, C₆HCHBrCN, is a colorless solid melting at 29^o. The crystals soon turn pink, owing to a slight decomposition, which does not proceed far, however. The commercial product melts at 16^o to 22^o, and the crystals are varying shades of dark brown, often with a marked greenish tint. The vapor pressure is given as 0.025 mm. at 0^o and 0.250 mm. at 40°. The compound decomposes before the boiling point is reached, even in a high vacuum. The density of the solid is about 1.51 at 25^o.

Brombenzyl cyanide is a very effective lacrymator. Most people can detect it at 0.021 parts per 1,000,000 and are lacrymated at 0.04 parts per 1,000,000 (0.00033 mg. per liter).

The compound is insoluble in water, moderately soluble in cold alcohol, freely soluble in hot alcohol, and soluble in ether, glacial acetic acid, carbon bisulphide, and benzene. It is hydrolyzed very slowly by boiling water and by cold solutions of sodium hydroxide. A cold alcohol solution of sodium hydroxide decomposes it rapidly, forming sodium bromide. It is oxidized slowly by potassium permanganate, bleaching powder, chromic acid mixture, etc. Brombenzyl cyanide attacks all metals rapidly except lead, and it corrodes lead. It would probably have to be loaded in enamel-lined or glass-lined shell. The magnesium and kaolin cements are satisfactory in presence of brombenzyl cyanide. It does not react with mustard gas. The charcoal in the American canister stops it very well; but the German charcoal appears not to be so effective.

Brombenzyl cyanide is prepared in successive steps, starting with toluene, which is converted into benzyl chloride. Benzyl cyanide is made from this by mixing with alcoholic sodium cyanide and distilling. The benzyl cyanide is brominated by treatment with bromine vapor in presence of light. All the apparatus is made of lead or is lead lined.

C₆H₅CH₃+Cl₂= C₆H₅CH₂Cl+HCl
C₆ H₅Cl+NaCN=C₆H₅CH₂ CN+NaCl
C₆H₅CH₂CN+Br₂=C₆H₅CHBrCN+HBr.

304

ETHYL IODOACETATE

Ethyl iodoacetate, CH₂ICO₂C₂ H₅, is a colorless oil, of extremely penetrating odor, which turns brown in the air with liberation of iodine. It decomposes if boiled in the air. The vapor pressure is 250 mm. at 143^o and 0.87 mm. at 30^o. The density of the liquid is 1.8320 at 4^o. The toxic concentration for dogs is about 1.6 mg. per liter; but the substance is primarily a lacrymator. Nine people out of ten tested were lacrymated at 0.14 parts per 1,000,000 (0.0014 mg. per liter). Owing to the scarcity of iodine, this is not a very satisfactory substance for war purposes. It was made by the English at a time when the price of bromine was very high.

PHENYLDICHLORARSINE

Phenyldichlorarsine, C₆H₅AsCl₂, is a highly refractive liquid boiling at 253^o to 255^o. The vapor pressure is 27 mm. at 142^o. The substance blisters the skin much more rapidly than does mustard gas. A burn up to four days old would be judged three to four times as extensive as a mustard gas burn of the same age, and equally as severe. The burns heal more rapidly than do those from mustard gas, so that the usefulness of this liquid is not established. A 60 percent yield can be obtained by heating triphenylarsine and arsenic chloride in suitable proportions in an autoclave at 250^o for 14 hours. (C₆H₅)₃As+2AsCl₃=3(C₆H₅)AsCl₂.

METHYLDICHLORARSINE

Methyldichlorarsine, CH₃AsCl₂ is a colorless liquid witb a powerful burning odor. It boils at 131.5^o, and has a vapor pressure of about 2.2 mm. at 0^o and 19.3 mm. at 35^o. The vapor is between five and six times as dense as air. The density of the pure liquid is given by Richter and Byers as 1.873 at 0^o and 1.81 at 35^o. The toxic concentration for dogs is 0.20 mg. per liter (78 p. p. m.).

Methyldichlorarsine is miscible in all proportions with arsenic chloride, while water dissolves 29 percent by weight and 16 percent by volume. It is insoluble in concentrated hydrochloric acid and very sparingly soluble in the constant-boiling acid. It may be distilled without decomposition alone or with hydrochloric acid stronger than 15 percent. Distillation with water causes a good deal of hydrolysis. Alkalies and alkali carbonates cause hydrolysis in the cold. The liquid has only a very slight action on shell steel even at 54^o. The substance is stopped in the canister both by the charcoal and the soda-lime.

Methyldichlorarsine is made in three stages:

(1) Dimethyl sulphate reacts with sodium arsenite to form disodium metbyl arsenite, Na ₃AsO₃+(CH₃)₂SO₄=NaCH₃AsO₃+NaCH₃SO₄. Possible side reactions are:

(CH₃)₂S0₄+NaOH=NaCH₃SO₄+CH₃OH.
(CH₃)₂SO₄+H₂=CH₃HSO₄+CH₃OH.

(2) Dimethyl sodium arsenite reacts with sulphur dioxide to form methyl arsine oxide,

NaCH₃SO₄+SO₂=CH₃AsO+Na₂SO_4.

The bisulphite formed by the excess of sulphur dioxide must be decomposed before the third stage is carried out, as otherwise the sulphur dioxide liberated would carry off with it a large part of the methyldichlorarsine.

(3) Methyl arsine oxide reacts with hydrochloric acid to form methyldichlorarsine. CH₂AsO+HCl=CH₃AsCl₂+H₂O.

305

CYANOGEN CHLORIDE

Cyanogen chloride, CNCl, is a colorless liquid, boiling at 12.6°, and solidifying at about -7.3^o. The vapor pressure is 444 mm. at O^o and 682 mm. at 10^o.The density of the vapor is only a little more than double that of air. The density of the liquid is 1.22 at 0^o. It is a good lacrymator (0.015 mg. per liter) and is highly toxic, low concentrations causing cramps in the chest and higher concentrations causing symptoms similar to those of hydrocyanic acid. Like hydrocyanic acid, there is no cumulative effect. The toxic concentration for dogs is 0.20 mg. per liter (72 p. p. m); but dogs are the most sensitive to cyanogen chloride of any of the animals.

One volume of water dissolves 25 volumes of cyanogen chloride, 1 volume of ether dissolves 50 volumes, and I volume of alcohol 100 volumes. The alcohol solution decomposes, esters of carbonic and carbamic acids being formed. Sodium ethylate converts cyanogen chloride into cyanic ether, while potassium chloride solution changes it to potassium cyanide and chloride, CNCl+2KOH=KCNO+KCl+H₂O.

An aqueous solution of cyanogen chloride is turned black by alkali, but is not polymerized by hydrochloric acid or chlorine. Nearly pure cyanogen chloride is polymerized to a white solid, cyanuric chloride (CNCl)₃, by small amounts of hydrochloric acid or chlorine. When dry, cyanogen chloride does not attack iron, lead or silver, but does attack copper. If moist it attacks all these metals.

Cyanogen chloride is made by the chlorination of aqueous hydrocyanic acid HCN+Cl₂=CNCl+HCl. Disturbing side reactions are:

CNCl+2H₂O=CO₂+NH₄Cl (in presence of HCl).
3CNCl=(CNCl)₃ (in presence of Cl₂ or HCl).

HYDROCYANIC ACID

Hydrocyanic acid. HCN, is a colorless, mobile liquid, boiling at 26.5^o. The vapor is slightly less dense than air. The toxic concentration for dogs is about 0.08 mg. per liter (90 p. p. m.), but dogs are exceptionally sensitive to this gas. The English physiologist, Barcroft, went into a gas chamber with a dog and stayed there unhurt until the dog had been killed by hydrocyanic acid. There is no cumulative effect. Neither the British nor the Germans used hydrocyanic acid. The French used a mixture called Vincennite; but there seems to be no evidence of its value.

SMOKES

In addition to the toxic gases, several substances have been used as irritant or incendiary smokes. A brief mention of the more important of these may be desirable.

PHOSPHORUS

Phosphorus is prepared by heating phosphate rock with sand and coke in an electric furnace, Ca₃(PO₄)₂+3SiO₂+5C=3CaSiO₃+5CO+2P.

Phosphorus comes on the market as either white (yellow) or red phosphorus. Either form burns to phosphorus pentoxide and is then converted to phosphoric acid, 4P+5O₂+6H2O=2P₂O₅+6H₂O=4H₃PO₄.

306

Since one pound of phosphorus takes up 1.33 pounds of oxygen and 0.9 pounds of water, it is not surprising that phosphorus is the best smoke producer per pound of material. In addition to its use as a smoke producer, it is used in incendiary shell and for coating tracer bullets.

TIN TETRACHLORIDE

The tetrachloride, SnCl₄, is a liquid made by treating tin with chlorine. It boils at 114⁰, and fumes in moist air because it hydrolyzes to stannic hydroxide, SnCl₄+4H₂O=Sn(OH),+4HCl. It makes a better and more irritating smoke for shell and hand grenades than either silicon tetrachloride or titanium tetrachloride. It goes through the charcoal and the soda-lime; but is stopped by the layers of cotton wool in the canister. Since there is practically no tin produced in the United States, silicon tetrachloride and titanium tetrachloride have been developed as substitutes.

SILICON TETRACHLORIDE

Silicon tetrachloride, SiCl₄, is made from silicon or from impure silicon carbide by heating with chlorine in an electric furnace, Si+2Cl₂ =SiCl₄.

It is a colorless liquid, boiling about 58° and fuming in moist air owing to hydrolysis, SiCl₄+4H₂O=Si(OH)₄+4HCl.

It is not of much value in shell, but is better on moist cool days than on warm dry ones. An ammonia cylinder and a silicon tetrachloride cylinder with liquid carbon dioxide as propellant give a first-class smoke when the jets from the two cylinders impinge. SiCl₄+4H₂O+4NH₃=Si(OH)₄+4NHCl. By adding a lacrymator to silicon tetrachloride one gets a mixture which works well in hand grenades for mopping up trenches.

TITANIUM TETRACHLORIDE

Titanium tetrachloride, TiCl₄, is made from rutile, TiO₂, by mixing with carbon and heating in an electric furnace. A carbonitride is formed which is said to have the composition Ti ₅C₄N₄ ; but the actual composition may vary from this to the carbide TiC. When these products are heated with chlorine, titanium tetrachloride is formed. This is a colorless, strongly refracting liquid which boils at about 136^o, is stable in dry air, and fumes in moist, air. It is said that the addition of water to form TiCl₄5H₂O gives a good smoke and that the hydrolysis to Ti(OH)₄ gives a poorer smoke. Titanium tetrachloride is poorer than tin tetrachloride and silicon tetrachloride in hand grenades. In the smoke funnel it is better than tin but not so good as silicon. Since it costs more than silicon tetrachloride, it would really be used only in case of shortage of the former.