VELOCITY OF DECOMPOSlTION OF HIGH EXPLOSIVES ETC. 1603 CLXXX-The Velocity of Decomposition of High Explosives in a Vucuzcm. Part II. Yi*initro-p hen ylmetlzylniti.oc~mine ( Il’etryl). By ROBERT CROSBIE FARMER. TETRYL is widely used for explosive purposes and the control of its stability is of considerable practical importance. On account of the instability of the nitroarninegroup the compound decom-poses much more rapidly than compounds of the triiiitrobenzene type and may undergo detmkration on storage if not completely purified. It was mainly for the control of the manufacture of tetryl that the vacuum stability test described by the author (this vd. p. 1432) was devised and this test has been used as the standard method for a number of years. A temperature of 120° has been found best for the measure-ments.‘Tetryl is solid a t this temperature (m. p. 1 2 9 O ) and gives convenient volumes of gas in aboat two days. The gases consist of carbon dioxide carbon monoxide nitrogen and oxygen and the residue is a complex mixture containing picric acid etc. The 3 ~ 1604 FARMER THE VELOCITY OF DECOMPOSITION OF evolution of gas proceeds with an acceleration and an arbitrary period of forty hours has been taken for the standard test. Samples of well-purified tetryl give evolutions of 1.5 to 3 C.C. from 5 grams of explosive. The reaction is very sensitive to catalytic influences and products containing minute quantities of residual impurities decompose much more rapidly. It appears probable that one of the main causes of instability is the presence of analogues of tetryl containing a nitro-group in the meta-position.Traces of picric acid may also be present; this compound has been found to1 decrease the stability greatly. The temperature-coefficient corresponds with a factor of 1.9 for each 5O that is approximately the same as for moet other explosives which undergo gradual decomposition. The logarithms of the velocities give approximately a straight line when plotted against the temperatures. By extrapolation to ordinary temperature it is found that forty hours a t 120° corresponds with about 1700 years a t 20". Such extrapolations must naturally be taken with reserve as the decomposition is of a complex nature and may include reactions the temperature-coefficients of which vary but the result shows that the purified substance is almost free from any tendency to1 decompose when stored a t ordinary temperatures, unless decomposition is induced by contact with any material of a reactive nature.A point of interest in the study of the stability of tetryl is that, measurements can be made both above and below the melting point. I n the case of most nitro-compounds the velocity of decom-position becomes sol low at the melting point that measurements on tho solid compound are scarcely possible. In view of the behaviour of many substances of melting with simultaneous decomposition it is instructive to make quantitative tests in the solid and the molten condition. When measurements are made on the solid substance at different temperatures a uniform tempera-ture-coeEcient is obtained.A t the melting point, an abrupt change takes place in the velocity of decomposition. The effect of the molten condition is to increase the velocity approximately fifty-fold. Beyond this point the velocity again increases uniformly with the temperature as before. When the logarithms of the velocities are plotted against tho temperature 'the measurements abo've the melting point give a straight line which is approxim-ately parallel to the line representing the velocities below the melting point (Fig. 5). This great increase of velocity on melting is the main cause of the acceleration in the decomposition at 1 2 0 O . When a slight decomposition has occurred the products of the decompositio HIGH EXPLOSIVES IN A VACUUM.PART 11. 1605 form a eutectic mixture with a part of the tetryl and this liquid portion at once decomposes with the higher velocity. It is always observed after testing that the crystals show signs of having been partly melted. This acceleration tends t o accentuate the differ-ence between tetryls of different initial stability. I n the molten condition the acceleration is much less marked and is in this case due to auto-catalysis. Pure and impure samples also show much less difference in their velocity curves in the molten conditim. The rapid decomposition of molten tetryl makes it impossible to obtain accurate measurements of its freezing pointl (" setting point' ") but the melting point can be determined by the capillary method without any appreciable error due to decomposition.The purest samples gave a melting point of 1 2 9 . 1 O . The above behaviour on melting formed an explanation of the results obtained on mixing tetryl with trinitrobenzene trinitro-toluene etc. These mixtures decomposed much more rapidly than tetryl itself a t looo to 1 2 0 O . This appeared a t first to indicate a chemical interaction but was simply due t o the loweiring of the melting point of the tetryl. The rates of decomposition of the mixtures at' temperatures above the melting point of tetryl agreed with those of tetryl whilst at lower temperatures they agreed with the extrapolated velocities from the temperature-velocity curve of molten tetryl (Fig. 5 ) . It may therefore be concluded t h a t ' a t temperatures belo'w the mellting point of the eutectic mixtures the stability of these mixtures will not differ from that of tetryl alone.It was not possible to test this as the rate of decomposition of tetryl can scarcely be measured a t temperatures much bebw 100'. I n the case of s-trinitrocm-xylem however it was possible to obtain measuremynts below the melting pclint of the eutectic mixture and these agreed approximately with those made on tetryl alone. These deductions have been confirmed by prolonged climatic trials of such mixturw att 500. Trinitrophenol caused a great increase in the rate of decomposi-tion of tetryl after making allowance for the loiwering 09 the melting point. As trinitrophenol is a decomposition product of tetryl this probably forms the main explanation of the auto-catalysis.Mixtures of small proportions of nitric and acetic acid did not accelerate the decomposition. Probably the acids were removed by distillation in the vacuum before they were able to exert any decomposing action. Sulphuric acid on the other hand gave rise to' a very rapid evolution of gas. Since the completion of the above work a few measurements have been given by Knowles (J. I n d . Eng. Chem. 1920 12 246) 1606 FARMER THE VELOCITY OF DECOMPOSITION OF using a modification of Obermuller's test. He gives the gas evolu-tion as 20 to 30 mm. per hour for commercial tetryls but the temperature and the volume of the apparatus are not stated. E x P E R I M E N T A L. Method of 7Vorking.-This was essentially the same as that described in Part I (Zoc.cit.). The standard procedure was as follows The tetryl was dried for four hours in a steam-oven and 5 grams were weighed into the heating tube which was then con-nected with the manometer. A mixture of heavy petroleum and ceresine wax was used as lubricant; these ingredients were care-fully examined to ensure that they were inert and the smallest possible quantity of the lubricant was used. After exhaustion, FIG. 1. 3 0- 2 t, 1 Hours 10 20 30 40 Normal decomposition of tetryls as manujactured ( 5 grams at 120"). the tube was heated for a few hours a t 80° to remove any volatile matter and was then again exhausted and flushed out twice with dry air. The exhausted tube was then inserted in the bath a t 120° and packed round the top with asbestos wool.A period of one and a-half hours was allowed for the level of the mercury in th0 manometer to become steady and readings were then taken ov0r a period of forty hours. For special purposes such as the measurement of temperature-coefficients the above conditions were modified as required. Precautions against explosion were taken as described in Part I but no explosion of tetryl occurred in the whole of the experiments. (a) Typical Results of Tetryl as Manufactured and Purified by -4cetone and TT'ater.-A number of these is shown in Fig. 1. Th HIGH EXPLOSIVES IN A VACUUM. PART 11. 1 GO? general character of the evolution is seen from the curves. A marked accelleration occurs due in part to auto-catalysis but in a greater degree to the progressive formation of eutectics.Attempts to determine the order of reaction serve therefore no useful purpose. Analysis of the gases from three experiments gave the following results : (i. 1 (ii.) (iii.) CO ........................... 20.0 21-5 23.3 co ........................... 5.5 7-1 5.8 0 ............................ 6-9 6.1 4.9 N? ........................... 67.6 65.3 66.0 These analyses do not take account of the nitric peroxide present. This was determined spectroscopically as indicated in experiments cited below. FIG. 2. Prolonged decomposition ( 5 grams 01 tc tryl). (b) Dlecomposition in Air and in a Current of Carbon Dioxide. -The velocity of evolution of gas at atmospheric pressure was originally used for the comparison of the stability of samples of tetryl but although this formed some guide in the purification of tetryl during the manufacture it was not so trustworthy as the vacuum method and may therefore be passed over.Two experi-ments may however be quoted in which the decomposition a t atmospheric pressure was continued until the gas-evolution almost ceased. The object of these was to ascertain whether the acceler-ation which occurred a t first would continue until i t culminated in an explosion. The thermostat was surrounded by a niound of earth and the readings were taken with a telescope. No explosion occurred however and it was possible to observe the whole course of the gas-evolution. Fig. 2 shows the gas-evolution from 5 grams of tetryl at 120° and 1250 respectively. In the latter case tb 1608 FARMER THE VELOCITY O F DECOMPOSITION OF gas-evolution ultimately almost ceased and the total volume of gas amounted to 1.27 mols.per mol. of tetryl. The residue was a dark resinous mass which did not yield any results on cry stallisation. Will's method (Zeitsch. angew. Chent. 1901 14 743 774) in which the exploaive is heated in a current of carbon dioxide was also applied t o tetryl. Fig. 3 shows the evolution of nitrogen from 2.5 grams of tetryl a t 125O over a period of thirty-two hours. The proportion of nitric peroxide was also measured spectroscopically by Robertson and Napper's method (T. 1907 91 761). On account of the low rate of evolution a very long observation tube was necessary in order to render the spectrum of the nitric peroxide WiZZ test (2.5 grams of tetry4 at 126").distinctly visible. The proportion of nitric peroxide decreased as the evolution proceeded (Fig. 3): It is of interest tot co'mpare the evolution of nitrogen from tetryl and guncotton respectively under the conditions of Will's test. A g o d guncotton gives approximately 1.5 milligrams of nitrogen i n four hours at 125O in a current of carbon dioxide whereas tetryl when well purified gives only 0.15 to 0.21 milligram in four hours. Its rate of decomposition is therefore only about one-tenth that of guncotton. (c) Purification of TetryZ.-The application of the vacuum test to the control of stability in the purification treatment is illus-trated in the following examples. Fig. 4 shows the effect of fractional precipitation of crude tetryl from acetolne by water afte HIGH EXPLOSIVES IN A VACUUM.PART 11. 1609 removing t4he nitration acids by t*reatment with hot water. The bulk of the tetryl was contained in fraction 2 which showed the greatest stability. Gas-evolution from 5 Grams at 12W. Time required for evolution of 2 C.C. C.C. in 40 Hours. hours. - Undiesolved residue Fig. 4 (c) ......... 4-6 First precipitation Fig. 4 ( e ) ......... 14.5 11.80 Second , Fig. 4 (k) ......... 34.0 2-40 Final ,) Fig. 4 (u) ......... 2.5 -FIG. 4. Purijkation 01 tetryl. This was confirmed by the Will test which gave the follolwing evolutions of nitrogen from 2.5 grams of tetryl in four hours at 125O undissolved residue 3.39; crop 1 0.51; crop 2 0.16; final crop 5-76 milligrams.Re-purification of the tailings by the same method gave the f dowing results : Crop 1 ( 5 grams) Fig. 4 (m) ... 1.90 C.C. in 40 hours. Crop 2 ( 5 grams) Fig. 4 ( h ) ... 2-55 C.C. 99 Residue (5 grams) Fig. 4 (n) ... 1.80 C.C. 9 ) A stable crop of crystals from the acetone-water treatment was 3 N 1610 FARMER THE VELOCITY OF DECOMPOSITION OF further crystallised from alcohol but this did not decrease the rate of decomposition. Before crystallisation from alcohol 2.40 C.C. in 40 hours Fig. 4 (k) After 9 ) ¶ ¶ ). 2-80 C.C. 9 p . ) Fig. 4 ( 9 ) In order to ascertain whether tetryl which had been partly decomposeld could be restored to its original stability by crystal-liaation the residue from a test after forty-four hours a t 120° was recrystallised from alcohol.This gave 6.85 C.C. from 5 grams in forty hours at 120° showing that the harmful impurities had not been removed (Fig. 4 f). Crystallisatio'n from toluene was applied to a very impure sample and to a sample previously purified by acetone and water. Gas-evolution from 5 Grams at 1 2 0 O . Time required for an evolution of 2 C.C. C.C. in 40 HOW& hours. - Unetable sample Fig. 4 (b) . . . 2.5 After crystallisation Fig. 4 ( d ) . . . 10.6 Stable sample Fig. 4 (k) ... 34.0 2.40 After crystallisation Fig. 4 (E) ... 37.0 2.30 -In some cases it was found that. the stability decreased on crystal-lisation especially from solvents in which tetryl was readily soluble. As this could not be due to the introduction of chemical impuri-ties it could only be attributable to the physical condition of the substance.The stability was found to be affected by the size of crystal large crystals decomposing more rapidly than small ones. This was probably due to the retention of volatile decomposition products within the crystals. The slight differences in the size of crystals of the tetryl in ordinary use for explosive purposes did not however affect the stability appreciably. (d) Temperature-coefficient of the Velocity of Becomposition.-As the evolution of gas proceeds with an acceleration it was con-ZnfEuence of Temperature on Rate of Decomposition. Initial velocity. & Tetryl Tem- /->- C.C. per gram Grams. perature. 0.5 C.C. 1.0 C.C. 2.0 C.C. 6.0 C.C. per hour. _Log. 5.0 100.0" 141 - 644 - 0.00071 3-85 0*00061 4.79 6.0 100.0 164 - - -6.0 111.0 - 73 152 255 0.0027 3'43 6.0 111.0 - 69 147 253 0.0029 $46 5.0 120.0 - 26 42 65 0.0077 T89 5.0 120.0 - 23 37 - 0.0087 3-94 0.5 126.8 6.0 7.3 - - 0.10 1.00 0.6 129.9 - 1-36 2.2 4.4 1.5 0.18 0.5 134.6 - 0.56 1.09 2.42 3.6 0.56 0.6 135.3 - 0.42 0.78 1-66 4.8 0.68 0.6 138.6 - 0.27 0.63 1.19 7.4 0.87 Time (hours) for evolution of HIGH EXPLOSIVES IN A VACUUM.PART 11. 161 1 sidered best' to take the initial velocities as the basis for the calcu-lation of the temperature-coefficient. Measurements were made both above and below the melting point,. The lower curve in Fig. 5 shows the logarithms of the velocities in relation to the temperature. It is seen that the solid substance has a regular temperature-coefficient (approximately equal to 1.9 for 5O).A t the melting point' there is a break in the curve and above this point it again proceeds in a straight line. The measure-ment a t 126.8 gave a low velocity a t the start but this increased extremely rapidly as the tetryl melted. The difference in velocity between solid and molten tetryl at 120° is in the ratio of 1 to 50. The following table shows the effect of heating mixtures of tetryl with trinitrotoluene and other nitro- compounds. Mixtures of Te tryl with Nitro-compounds. InflueTzce of Temperature. Initial velocity. YAP Admix- Time (hours) for evolution of C.C. per gram of tetryl Tetryl. ture. Tem- c A - Grams. Grams. perature. 0.5 C.C. 1.0 c.c.. 3.0 C.C. 5-0 C.C. Trinitrotoluene mixtures : 5.0 5.0 100.0" -5.0 5.0 100.0 -5.0 5.0 110.5 -5.0 5.0 110.6 -0.6 0.5 120.0 -0.5 0.5 128.6 -0.5 0.5 129.0 -0.5 0.5 134.6 -Trinitrobenzene mixtures : 5.0 2.5 100.0 -5.0 5.0 110.5 -5.0 5.0 110.6 -0.5 0.5 120.0 -0.5 0.5 135.3 -Trinitroxylene mixtures : 6.0 2.5 100.0 124-0 5.0 2.5 110-0 21.5 5.0 2.5 120.0 1-86 0.5 0.5 140.7 -0.5 0-5 140-8 -Dinitrobenzene mixtures : 6.0 5.0 100.0 -6.0 5.0 100.0 -8.5 8.5 2.6 2.5 4.5; 1-23 1.41 0-51 10.4 2.4 2.5 4-05 0.56 -36.7 -0.24 0.20 11.1 12.0 Fig.5 shows the relationship 14.3 28.7 14.7 30.4 4.4 8.5 4-3 8.3 7.8 14.8 2.23 -2.48 4.85 0.90 -17.4 35.0 3.95 -3.92 -6.40 13.2 1.00 2.18 - -61-5 99.0 - _-0.43 0.99 0.38 0.85 16.8 29.9 18.3 31.8 between the per hour.0.023 0.023 0.077 0.080 0-44 1.63 1.42 3-92 0*019 0-083 0.080 0-49 3.57 0.00081 0-0047 0-054 8.3 10.0 0.018 0.017 Log. -- 2.36 2-36 3.89 1-64 0.21 0.16 0.59 2.90 2.28 3-92 5.90 1-69 0.66 -4-01 $67 2-73 0.92 1.00 2-26 3-23 logarithms of the velocit.ies and the temperature. It is seen that above the melt-3 N* 1612 FARMER THE VELOCITY OF DECOMPOSITION OF ing point of tetryl ( 1 2 9 O ) the velocities for tetryl and the mixtures coincide. Below 1 2 9 O the values for solid tetryl are much lower (curve 1). The values for the mixtures with dinitrobenzene tri-nitrobenzene and trinitrotoluene are shown in curve 3. These admixtures lower the melting point of the tetryl and form liquid mixtures ; the velocities of deco'mposition correspond therefore, with those of molten tetryl and their logarithms lie approxim-ately on a straight line forming an extrapolation of the values for molten tetryl.In the case of trinitroxylene the eutectic is almost completely solid a t looo and here the velocity becomes nearly equal to that of somlid tetryl (curve 2). Measurements on the trinitrotoluene and FIG. 6. Influence of temperature on uelocity. trinitrobenzene mixtures were not taken below the eutectic melt-ing points but prolonged climatic trials a t 50° indicated that very little decomposition occurred) in a year a t this temperature. A noticeable feature of the mixtures of tetryl with trinitro-benzene (and to a less extent with trinitrotoluene) was that on cooling t o the ordinary temperature after melting they remained very persistently supercooled forming viscous dark reddish-brown colloids.The trinitrobenzene mixtures crystallised only after several days. Varioas tetryls were mixed with trinitrotoluene to ascertain whether tests at looo would be of value in discriminating between pure and impure samples. This method was however of relatively little value as compared with the direct test on solid tetryl a t 12QO HIGH EXPLOSIVES IN A VACUUM. PART 11. 1613 I n all cases the acceleration of the decomposition is much greater in the solid condition in consequence of the progressive melting clue to the decomposition. This is shown in the following table. Acceleration in the Decomposition of Tetryl.Tetryl. Grams. Solid ............ 5.0 ............ 6-0 ............ 5.0 Liquid ............... 0.5 , ............... 0.5 ................. 0.5 , ............... 0.5 9 , 9 , Tern -peraturc. l l l . o o 111.0 120.0 129.9 134.6 135.3 138.6 Initi a1 velocity. 0-0027 0-0029 0.0077 1-5 3.6 4.8 7.4 Mean velocity from 2 C.C. to 6 C.C. 0.0058 0.0057 0.026 2.7 4.5 6.8 9.1 Ratio. 2.1 2.0 3.4 1.8 1.3 1.4 1.2 This was confirmed by the Will test in a current of carbon dioxide. Thus the values at 1 2 5 O were as follows: First Second , , 0.53 ) 9 9 I ?# Third , , 2.25 milligrams , 9 P, Fourth , ) 6.28 , 3 9 1 9 Fifth , , 13.39 , 7 9 ? Y , 4 hours 0.21 milligram of N per 2.5 grams of tetryl.(e) Reactions of Tetryl.-Nitric and acetic acids in small pro-po'rtion did not affect the velocity of evolution materially. In general acids accelerate the decomposition but these acids being volatile were probably removed by the vacuum before they had time to exert any catalytic action. Gas-evolution at 120°'. 20 hours. 40 hours. Tetryl 6 grams ............ 0.96 2.70 2-38 Tetryl 6 grams Nitric acid 0.01 gram } **.*.***. 2.33 Tetryl 6 grams Acetic acid 0.01 gram} **""." 0'85 0'83 Sulphuric acid on the other hand being non-volatile accelerated Even at 80° a marked evolution the decomposition very greatly. of gas occurred. Gas-evolution at 80°. 20 40 60 80 100 hours. hours. hodre. hours hourr. 4.16 4.46 Tetryl 6 grams Sulphuric acid 0.01 gram} le60 2.70 3*55 Picric acid also gave rise to a rapid evolution of gas when mixed with tetryl 1614 VELOCITY OF DECOMPOSITION OF HIGH EXPLOSIVES ETC.Tem- f 20 40 60 80 100 perature. hour. hours. hours. hours. hours. hours. Tetryl 6 grams ............... 120" - 0.50 1.25 3.50 8.50 - ......... 100 - 0.15 0.20 0.25 0.30 0.35 1-60 3.90 6-40 9.20 12.35 ** 9 ) ... pic,,ic 0.5 E;ram ... 120 0.85 - - - - - 4.5 grams I Tetryl * 9 9 9 9 9 ... 100 - , , ) , ... 60 - - - - - 0-02 Some allowance must be made for partial melting of the tetryl containing 10 per cent. of picric acid butl apart from this the picric acid certainly affected the stability adversely. Picric acid was found to be a decomposition product of tetryl and this is doubtless one of the causes of the auto-catalysis which is observed.The influence of some other admixtures is shown in ths following table. These all formed liquid mixtures with tetryl a t looo and, on this account the velocities should be compareld with that of a mixture of tetryl and trinitrotoluene rather than tetryl alone. In the case of p-dichlorobenzene the liquefaction was probably i ncom p 1 et e . Gas-evolution at looo. Hour3 required for evolution of Initial velocity. /-A \ C.C. per gram 1 C.C. 2 C.C. 3 C.C. of tetryl per hour. 9.0 17-1 0.036 :grrffns) i: 10.3 18.6 0.029 " } 17.0 25.0 42.0 0.012 Tetryl Nitrobenzene Tetryl p-Dichlorobenzene 6 ,, Tetryl Diphenyl ether 5 )) ' ') } 2.7 6.1 - 0-074 Summary. The velocity of evolution of gas in a. vacuum a t 1200 forms a useful methosd for the control of the stability of tetryl in the manu-facture. The effect of purification and of various admixtures is shown. The temperature-coefficient of the decomposition of solid tetryl is 1.9 fo'r 5 O . At the melting point an abrupt change in the velocity is observed the molten tetryl decomposing about fifty times as rapidly as the solid. The acceleration in the decornposi-tion of tetryl a t 120" is to a great extent due to progressive melt-ing. Admixtures which lower the melting point also give a rapid evolution apart from any chemical action which they may exert. The thanks of the author are due to the Director of Artillery for permission to publish the above results. RESEARCH DEPARTMENT, ROYAL ARSENAL, WOOLWICB. [Received November 13th. 1920.