THE CHEMICAL NEWS,

VOL. CXXXI. No. 3415.

MERTON HOUSE, SALISBURY SQUARE, LONDON, E.C.4.

TELEPHONES:

Administrative Central 6521 Printing Works, Hop 2404.

DECOMPOSITION OF TRINITROTOLUENE BY THE ACTION OF SUNLIGHT.

By PROFESSORS C. KRAUZ And O. TUREK. Till the world-war, trinitrotoluene, i.e., 2, 4, 6-trinitrotoluene was generally considered as a material of an almost unlimited resistance to shock and friction which can be pressed, cut, bored, planed, etc.' Its manufacture had not, up to that time, presented any obstacles, and the reason of the explosions which occurred here and there were not attributed to trinitrotoluene alone. Consequently, TNT (mainly in warfare), has been generally favoured, so that, where possible, it has been employed instead of picric acid.

The great faith in the perfect stability of TNT as far as sentitiveness is concerned was shaken, however, during the war. Already in 1915 the gun-barrel burstings were caused by shells filled with TNT, most of which could only be explained as the "selfexplosion" of the TNT. Such accidents often happened during the filling of shells. and hand-grenades with pressed or molten TNT, (e.g., in Bolevec, 1917). Further, TNT can explode without any apparent reason during manufacture, as was the case in the unexpected explosion in the factory of the Nobel Dynamite Company Bratislava, in 1918, when a whole TNT department was literally blown up.

at

Such unpleasant accidents were usually attributed to the impurities which might be found in technical TNT. That this view was, perhaps, in many cases correct, can be admitted. In such a speedy and sometimes, at least, accelerated production as occurred in the last war-when besides other obstacles, there were difficulties not only with the quality but also with the supply of raw materials the product was not always the cleanest. Further, substances are formed in the manufacture of TNT which can be

SPECIAL NOTICE.

All MSS., Books for Review, Business Letters and Communications of every description should be sent to the offices as under, and not to any other address.

CHEMICAL NEWS, MERTON HOUSE, SALISBURY SQUARE, LONDON, E.C.4.

γ

For

the cause of unexpected explosions. instance, J. M. Weiss points out that technical TNT may contain high nitrated phenols, which form, with metals, unstable salts, and can be the cause of an unexpected explosion. H. Ryan and W. M. O'Riordan2 perceived that crude TNT usually contains small quantities of B- and y-isomers, mainly the derivative, and is accompanied by some kind of dark amorphous substance which easily explodes when heated; the authors do not suppose it for excluded that this substance may be the cause of unexpected explosions of the technical TNT. Further, for instance, M. Kostevich" supposes that the causes of these unexpected phenomena may lie in the mixture of substances which he calls "TNT tarry matter, ' and which usually accompany the technical TNT if it was not sufficiently purified.

All above-named authors may certainly be right about it, and then in the inference of their presuppositions, the possibility of unexpected explosions of TNT should fall off if this product were always sufficiently purified, (by Kostevich), if the purest materials only were used for its production, and if the intermediate mononitrotoluene were freed in practice by distillation of those tarry matters" before further nitration. The question arises if the TNT, carefully freed of all impurities, is capable of changes which might lower its stability respecting its sensitiveness in a disadvantageous practice. If we review the investigations appertaining up to this time to the TNT made there, we find that it does not mean a substance which would especially resist the effects of different chemical agents. Concentrated alkaline solutions cause deep decomposition of TNT which, as Dupré pointed out, easily lead to ignition, and even to explosion. According to Copisarow, even dilute alkalies have some effect. Strong oxidising agents, for instance, the nitrating mixture, on heating,

a

easily changes TNT into trinitrobenzoic acid and eventually into trinitromethane, as was shown by Giua, Will' and Langenscheidt. Even sunlight also affects entirely pure TNT, because changes which take place here are already apparent. That TNT varies in colour between yellow, yellowishred and dark brown, has been known for a long time. H. Kast pointed out that the point of solidification of TNT after fourteen days' action of sunlight, was lowered from 80.0° C., to 79.5° C. F. M. Vasquezi reports on the browning of TNT under the influence of daylight, but he states that the other qualities do not change. (?) E. Molinari and F. Quartieri describe in their book, "Notices sur les Explosifs en Italie" (Milan, 1913), on page 167, experiments which they undertook with the illumination of TNT and at which the original-at 80.4° C. melting and 0.03% insoluble residue in acetone substance shewed after a longer exposure to sunlight, a melting point of only 76.6° C., while its insoluble residue rose to 0.1%; in the same book, Molinari and Giua state that the m.p. of pure TNT after three months' exposure to sunlight in air, was lowered to 74° C., while another sample of the same material illuminated under the same conditions, but in an evacuated glass tube, shewed a much less intense colouration, and a m.p. only 0.6° C. lower than the original substance.

H. Custis10 states that sunlight causes oxidation of TNT and affords proofs that the Urays act quicker. Finally, Oddo12 mentions that TNT on exposure to light becomes acid, and capable of oxidising alcohol.

Independently of these observations, as early as the beginning of the year 1916, one of us began to study (during his time at the former military establishments for the manufacturing of explosives at Blaumau, near Vienna), the changes in TNT which arise by the action of sunlight, and came to the opinion that among the photoreaction products is very probably picric acid. Picric acid is indicated in the alkaline exeracts of the light brown TNT by the colour reactions, and with dye tests on wool, etc. More thorough and systematic experiments in this direction were postponed owing to circumstances.

As is seen in the above-mentioned literature, no one hitherto has worked on the chemical characteristics of the products which are formed by the action of sunlight on TNT. Therefore our experiments were carried out in this direction, i.e., we exposed

entirely pure 2, 4, 6 trinitrotoluene to the direct action of sunlight, then we examined the products formed. As will be mentioned below, we at first isolated from the mixture of these photoreaction products, only trinitrobenzoic and picric acids, while the identification of the other substances which also formed was postponed, mainly for the scarcity of material on the one hand, and for the long time which is necessary for the exposition of light on the other. For further tests with TNT the illumination which would offer us sufficient quantities of the material for the study of other, and unknown reaction products, we used the Urays instead of sunlight. The results of these very interesting experiments which are not yet finished will be given later.

Being satisfied with the proved presence of trinitrobenzoic and picric acids in the photoreaction products, we first proceeded to study the influence of these two substances on TNT sensitiveness. For this purpose we compared the sensitiveness to shock of the original entirely pure TNT and of TNT subjected to sunlight, as well as of the pure TNT to which we added different quantities of trinitrobenzoic or picric acids.

In our latest work the experiments are divided into three groups, i.e., the experi ments on the illumination of TNT and the isolation of the products; the preparation of the necessary trinitrobenzoates and picrates; and finally, the experiments on sensitiveness.

EXPERIMENTS ON THE ILLUMINATION OF TNT.

As elementary raw material, we used the technical tritrotoluene made in the Nobel Dynamite Factory in Bratislava. This material was good for all conditions required in practice. The substance was purified by crystallisation from hot alcohol to the constant m.p. of 81.4° C.; for this a double crystallisation was usually sufficient. The pure substance was always filtered by strong suction, and spread on filter paper in a thin layer and left to dry in the air in the dark for about 12 hours (i.e.. during the night), then finished drying (again in the dark), for several weeks in a desiccator above concentrated sulphuric acid.

In this manner, almost snow-white TNT was obtained in the form of fine needles, and this was used for the illumination experiments. For this purpose the substance was equally spread in a thin layer on a crystallising basin, protected from dust, etc., by a glass cover. Basins so arranged

were placed in an open window, directly in a sunny place. Already after several hours it was possible to see the slight turn to yellow which, with the advance of time, became more intense and slowly changed to brown. In order to obtain as much action as possible from sunlight, the TNT layers in each basin were occasionally thoroughly mixed.

(a) EXTRACTION WITH WATER.

After fourteen days of illuminating, we extracted a sample with water, first in cold and then in boiling. in boiling. By this method we gained yellow coloured solutions evidently reacting acid and colouring on adding soda or alkaline solutions, intensely red, and soagulating with acetic-nitrone solution.

The residues from the water extraction, (whether in cold or in warm) remain darkly coloured, from which it is seen that the extraction was not perfect.

The fact that TNT becomes acid by the action of light points to an oxidising reaction which would be in full agreement with the above-mentioned observation of Custis.

(b) EXTRACTION WITH POTASSIUM HYDRATE SOLUTION.

Supposing that the acid reacting products. of oxidation to be more easily soluble in alkalies than in water alone, we tried to extract the exposed TNT with a dilute (about 50%) potassium hydrate solution. This gave an intensely dark red solution, from which dark tarry matters separated after several minutes.. Also, even the residue which was insoluble in the extracting liquid was almost black brown, i.e., disproportionately darker than formerly. These circumstances also point to the origin of some kind of tarry matters which are formed by the influence of the alkali. A similar appearance may also be observed when KOH acts on the purest TNT alone, again in agreement with the above-mentioned observation of Dupré and Copisarov. fore, it is plain that it is not possible in this case to use alkali hydrates as solvents and it is necessary to find a weaker agent.

(e)

There

EXTRACTION WITH SODA SOLUTION.

We even tried to use as extracting liquid. a 10% sodium carbonate solution, with which we shook out three time a sample of exposed TNT. Although we did not get the total colour from the extracted sample, yet we noticed that, contrary to the original brown colour, it became remarkably pale.

The soda solution, which was intensely red, we inade acid with dilute sulphuric acid, by which dark brown flocks separated and the liquid changed to a dark yellow colouration. The dark sediment, here separated, we considered contained no important impurities; therefore we filtered it off and neglected it. The filtrate was then extracted with ether, the dark yellow extract was dried with anhydrous calcium chloride. and freed from ether by vacuum distillation at laboratory temperature. The resulting residue was dark red and greasy, soluble in alcohol, giving an intensely coloured solution and giving with barium, silver and mercury salts, alnost black precipitates. The main part of the ether extract was dissolved in alcohol and precipitated with a 10% alcoholic solution of silver nitrate in excess; the dark sediment, settled on the second day, was suspended in water and decomposed by hydrogen sulphide. After separating the precipitated silver sulphide by filtration, we shook out the filtrate with ether and freed the extract from it by distillation under diminished pressure. The yellowish red residue was so little that it was scarcely sufficient for the test tube reactions. solved in alcohol, it gave with barium and potassium hydroxide solutions, an intense dark red colouration and was precipitated by mercury nitrate and nitrone-acetate as flocky sediments.

Dis

As it seems even the soda solution acts on the illuminated TNT unfavourably, we decided to use sodium bicarbonate for the next extraction experiments. Even the purest TNT is coloured in the shortest time, pink to red, by contact with normal sodium carbonate solution, while the bi-carbonate. solution under the same conditions remains colourless, even after several days.

(d) EXTRACTION WITH SODIUM BICARBONATE.

For further experiments in which we used TNT exposed quite four months (from November to February) and whose m.p. was lowered from the original 81.4° C. to 73.5° C., we proceeded as follows: 20 gr. of the brown substance was dissolved in 50cc of perfectly pure benzene and the dark solution was extracted twice with about a double volume of a 5% sodium bicarbonate solution. The original brownish red benzene solution was at once discoloured almost completely, by which it was possible to judge that the larger quantity of the photoreaction products passed into the aqueous solution. The benzene solution yielded, after evaporation