846
INDUSTRIAL AND ENGINEERING CHEMISTRY
the normal curing range incrmsing thc prossure of vulcanization from 1000 to 100,000 pounds per square inch has the same effect on t,he percentage of combined sulfu an increase in curing tern. perature of approximately 10” F.
Vol. 41, No. 4
Straka for assistance with the bibliography. This investigatioii was carried out undcr the iponsorship of the Office of Rubber Reserve, Reconstruction Finance Corporation, in connection w i t h the Government Synthetic Rubber Program.
CONCLUSIONS
LXTER.4J’UKE CITED
One of the most interesting aspects of the rejults is the coiitrast in the efl’ects on GR-S and Hevea of high pressure Vulcanization. There is a possibility that the results might be influenced by secondary factors such as differerices iii the solubilities, inelt,iiig point,s, or diffusion rates of the vulcanizing ingredients under high pressure. It seems much mow probable, in vievi of the essential similarity of the two systems, t,hat the contrast in results is relatcd to sonic differencc in the vulcanization reactions with GR-S and Hevca. The apparerit retardation of cure by high pressure with Ilevea suggests the controlling influence of a reaction with a negative pressure coefficient of velocity, possibly a unimolecular type of reaction. IIauser and Sze (18) have considered that a molecular change in the sulfur from Q g to 82 or SIis a necessary preliminary of vulcanization. 11 is difficult t o understand why such a reaction should occur diflcrently in Hevea than it does in GR-S. For a long time there has been a hypothesis that a thern i d deI)ol~iiit.rizationof Hevea takes place during cure which increases its reactivity. Lewis, Squires, and Nutting ( 1 9 ) discount this theory but ncvcrtheless it remains the most attractive explanatioii of the results obtained here. From this viewpoint, it n-ould be considered t)hat the high pressure rethrds this depolymcrization of Hevea to a sufficient extent to make it the (nuntrollingreaction for the rate of vulcanization. This is consistent with the improved resilience values secured with Hevea by the use of high vulcanizing pressure. The concept of two such conflicting reactions during vulcanization has also been forinulated by Dogadkin, Karmin, and Goldberg ( 1 1 ) . For GR-S, it is then to be assumed that such a depolyinerization during cure does not occur to a controlling extent’. A study of the pressurc coefficients of vulcanization react,iona seems to offer an additional means of investigat,ing thes,.: coinplicated effect,s.
Adam*, L. €and I., Gibson, R . E., J . W a s h . Acod. Sei., 20, 218 (1930). Bridgman, P. W., J . Colloid Sei., 2, 7 (1947). Bridgman, P. W., “Physics of High Pressure,” Sew York, Mac. millan Go., 1932. Bridgmsn, P. W., Proc. Am. A c a d . Arts Sci., 7 6 , 9 (1945). Bridgman, P. W., Revs. M o d e r n Phvs., 18,1 (1946). Bulgin, D., Trans. Inst. R u b b e r l n d . , 21, 188 (1945). Conant, J. B., and Peterson, W. R.,Ibid., 54, 629 (1932). Conant, J B., and Tongberg, C. C., J . Am. Chem. Soc., 52, 1659 (1930). Dillon, J. IT., and Gelman, 8 . D., I n d i a Rubber W o r l d , 115, 61 (1946). Dintses, A. I., Korndoif, B. A, Lachinov, S. S.,L’spekhi Khim., 7, 1173 (1938). Dogadkin, R., Karmin, B., and Golberg, I., Hubber Chern. and Technol., 20, 933 (1947); Compt. rend. ncad sci. U.R.S.S., 53, 327 (1946). , DOW, R. B.,J. Chem. ~ h v s . 7,201 , (1939). (13) Evans, M. G., T r a n s . Faraday Soc., 34, 49 (1938). (14) Evans, M.G., and Polanyi, M . , I b i d . , 31,875 (1935). (15) Fawcett, E. W., and Gibson. R O., J . Chem. Soc., 1934,386, 396. (16) Gehman, S. D., Woodford, D. E ~and > Stambaugh, It. B., IND. E N G . C H E M . , 33, 1032 (1941). (17) Glasstone, S., Laidler, K. J., and Eyring, H., “Theory of Hate Processes,” p. 470, New York, McGraw-Hill Book Co., 1941 (18) Hauser, E. A . , aiidSze, M. C., J . Phys. Chem., 46, 118 (1942). (19) Lewis, W. IC,SquiIea, L., and Nutting, R . D., IND. CNG.CHEW., 29, 1135 (1937). (20) Sapiro, R. H., Linstead, It. P., and Newitt, D. M., J . Chem. Snc., 1937, 1784. (21) Scots, A. H., J . Reseai.ch Xatl. Bur. StandaTds, 14,99 (1935). (22) Ibid., 15, 13 (1935). (23) Starkweather, 11.W., J . Am. Chem. Soc., 56, 1870 (1934)(24) Tammann, G., and Pape, A,, 2. anorg. u allgem. Chcm., 200, 113 (1931). (25) Thiessen, P. A , , and Kirsch, W . , ,~~~turwissenschoften, 26, 387 (1938) ; Rubber Chem. and Technol., 12,12 (1939). (26) JVood, L. A., Bekkedahl, N., and Gibson, R. E., J . Research Natl. BILI..Standards, 35, 376 (1945). (27) Zelinskii, N. D., and Vereshchagin, L. F ~ Bull. , acad. sci. C.IZ.8.AS.,Classe sci. chim., 1945, 44.
ACKNOWLEDGMENT
The authors wish to express their thanks t o the Goodyear Tire
th Rubber Company and L. B. Sebrell for permission t o publish this work, t o G. H . Gates for supplying the compounds used, t o R.. E. Johnson for combined sulfur determinations, and t o Leora.
RECEIVED J u n e 22, 1948. Present,ed before t h e Division of Rubber Chemistry a t llie 113th hIeeting of the AMIEHICAN CHEMICAL SOCIETY, Chicago, 111. Contribution No. 160 from the Research Laboratory of the Goodyear n r t & Rubber Company.
S Stability in Aqueous Pastes and Solutions RONALD A. HENRY, SOL SICOLNIK, ROBERT RSAKOSICV,
AND G.
B. L. SMITH
C’. S. Naval Ordnance Test Station, Inyokern, Calif.
F
OR safety reasons, lnterstate Commerce Commission regu-
lations (6) require that the cool, initiating explosive, nitrosoguanidine, be shipped as a paste containing a t least 10y0water. However, such a practice is opposed by several qualitative studies (1, 2, 6, 8, 9, 11) which show t h a t this material decomposes in aqueous systems, usually with the formation of gaseous products that could create excessive and dangerous pressures if storage or shipment were made in sealed containers. Therefore, a brief study was undertaken to determine the velocity of this decomposition in aqueous solutions and, more particularly, in pastes, so that the pressure increase under various circumstances could be predicted. X o consideration has been given in this investigation t o the sensitivity or other explosive properties of aqueous pastes of nitrosoguanidine.
Quantitative measurements on the rate of this decomposition in dilute solutions have been made previously (4, 7 , 10, l b ) , using colorimetric methods to follow changes in the nitrosoguanidine concentration. These studies indicate that the reaction obeys the first-order rate law and t h a t the specific rate constants arc only a pronounced function of the hydrogen ion concentration when the pH is less than 3, and hydroxyl ion concentration when the p H is greater than 10. Because the colorimetric procedures involve several variables that are difficult to control (9),the rates of decomposition in initially near-saturated, aqueous solutions have been redetermined a t three temperatures, using a titration method developed by Sahetta, Himmelfarb, and Smith (9) t o measure nitroeoguanidine concentration. This analysis involves titration in strongly acidic medium with standard potassium
April 1949
INDUSTRIAL AND ENGINEERING CHEMISTRY
permanganate using a manganous salt as a catalyst; the equivalent reducing weight of the nitrosoguanidine is one half its molecular weight. Distilled water (750 ml.) was heated to the temperature at which the rate study was to be made. Recrystallized nitrosoguanidine (S), in slight excess over that required to yield a saturated solution (see 9 for solubility data), was added and stirred for 10 minutes. This solution was then rapidly poured through a Whatman 41H filter paper into a 1-liter Erlenmeyer flask and the p H was determined or adjusted t o some desired value. After the flask had been tightly stoppered, the solution was brought to temperature equilibrium in a constant temperature bath capable of holding within * O . l o C. of the preset value. Aliquots (25 t o 50 ml. depending on the concentration) were removed and titrated at convenient intervals according t o the procedure outlined above. The average deviation of the determinations was 2.6%.
The results of these experiments are fecorded in Figure 1, where the logarithm of the concentration has been plotted against elapsed time. Specific reaction rate constants, calculated from the slopes of these graphs, according to the equation: k = -2.303 X slope, are summarized in Table I. Although the pH of the neutral and alkaline solutions decreased during the decomposition, the rates of decomposition in these systems did not appear to change with time. No suitable explanation for this observed change in p H can be offered unless the cyanamide (see Equation 1) behaves as a weak acid. Application of the Arrhenius equation to the constants for the initially neutral solution (Figure 2) leads to an energy of ac4ivation of 25,500 calories per mole for the reaction: E = -.2.303 X 1.987 X slope.
847
T h e amount of decomposition occurring in short intervals of time, up to several weeks, and at the temperatures likely to be encountered will not cause an appreciable loss of nitrosoguanidine when shipped or stored as a 10% aqueous paste. However, from a safety fitandpoint, adequate precautions must be taken to prevent the development of pressures in the containers, either by leaving ample dead space or by utilizing fiome pressurereleasing device.
.
reasonable results have been obtained by using a manometric method based on the assumption that the dearrangement in neutral solution (2, 11): NitroEoguanidine +nitrogen f cyanamide
+ water (1)
occurs exclusively and that the pressure increase in a constantvolume system is directly proportional to the amount of decomposition. Four t o 5 grams of once-recrystallized nitrosoguanidine were weighed into a 220-ml. round-bottomed, narrow-necked flask, then a measured weight of water was added. After loosely stoppering, the flask and contents were suspended in the constant temperature bath for 10 to 15 minutes before being attached t o a differential, mercury-filled manometer. Connections involving rubber stoppers were wired tightly and sealed with Fisher's Sealit. A ground-glass stopcock separated the flask from the mercury and was normally kept closed, except when readings were being made,
DECOMPOSITION IN PASTES OR SLURRIES
With pastes containing 10% or more of water, the titration method is not sufficiently precise to reveal consistently the small changes in composition that are experienced. Satisfactory and
TABLE I . SPECIFIC REACTION RATE PH Temp., QC.
After 116 hours
Initial
k, hr.-* X 102
a Titration data similar t o those for A , Figure 1.
TI ME, HOURS
Figure 1. Rate of Decomposition of Nitrosogugnidine in Initially Neutral Aqueous Solutions t,
c.
31.7
40.7 48.3
4
3s
4
3.3
3.2 I/T'K
x 103
Figure 2. Rate of Decomposition of Nitrosoguanidine in Aqueous Solutions as Function of Temperature
848
INDUSTRIAL AND ENGINEERING CHEMISTRY 1.2
I
I
I
IdBl
I
9
% nitrosoguanidine (on total basis) decomposingperhour = 0.00898 X 100 900
Volume of gas evolved per hour
B
0.8
22,414 X
.A
"'-0 Figure 3.
40
1 -I
d
A. E.
1
=
r
z
0
I
Vol. 41, No. 4
= i
.oo x
!O-P
0.00898 -X 88.07
313.9 273.2
= 2.61
cc
Other similarly calculated results, together with the corresponding rates as determined by the manometric procedure, arc summarized in Table 11. Considering the differences in the two analytical methods and the calculations involved, the results are in good agreement and arc taken as coniirming the validity of the assumpt'ions made above pert'aining to the decomposition of the nitrosoguanidine in pastes. This agreement also indicates t h a t the mode of decomposition is probably identical in both solutions and pastes, and that. t,he analytical methods (permanganate titration and manometric procedure) are measuring the same decomposition.
80
TIME, HOURS Decomposition of Nitrosoguanidine in Aqueous Pastes
TABLE 11. DEcoxPosmox !!.QrEoLiS
( R Decomposition/Hr.)
Run at 31.7' C . , paste contains 44.1% nitrosoguanidine Run at 40.7' C . , paste contains 5 0 . 6 % nitrosoguanidine
Lemp.,
Results obtained at 31.7" and 40.7' C. for pastcs containing about 50y0 water are plotted in Figurc 3; the differences here are due not solely to the temperature change but in part to differences of concentration in the aqueous phase. The slopcs of the curves in this graph give directly the per cent decomposition per hour, which for a given paste is essentially constant \vith time during the intervals considcrcd. However, these rates are characteristic of only these particular pastes; pastes with more or less water exhibit higher or lower rates, related as follows a t any given temperature:
% decompositlon per hour in paste A
=
X
(Cr.oi Gas Evolved/&-.)/ 100 C;. of Water e t Indicated
102 In Pastes Containing 10% %Vater
__,
