Oct. 20. 1959
DECOMPOSITION OF DI-~-BUTYL PEROXIDE
phism is suggested by the shape of the curves in Figs. 1 and 3. As can be seen from the portions of the curves a t the right of the eutectics, the initial addition of aliphatic comonomer to the aryl homopolymer did not immediately produce a linear depression of the melting point. This seems to indicate that, within the range of composition covered by the curvilinear portion of the curves, isomorphism could have occurred. To use suberic acid and 9benzenediacetic acid as an example, partial isomorphism will signify that, a t low molar ratios, suberic acid can enter the crystalline lattice of p benzenediacetic acid. This is not an unreasonable explanation, since i t is conceivable to rotate the bonds of a flexible molecule in order to fit the lattice of a rigid one, but not conversely. However, rotating a comonomer molecule out of its most stable configuration will introduce a strain to the lattice, which increases with increased replacement. When the strain surpasses the lattice stability, nonisomorphism and a linear depression of melting point will ensue. Acknowledgment.-The authors are indebted to Drs. J. A. Howsmon and M . R. Lytton for their interest and encouragement in this work. Thanks are due to Dr. J. T. Massengale for organic syn-
5365
Fig. 9.-Schematic chain alignment of copolyaniide of suberic acid and 9-benzenediacetic acid. AB represents suberic acid and CD p-benzenediacetic acid.
thesis, and to Mr. F. F. Morehead for the determination of polymer melting points. MARCUSHOOK,PENNA.
[CONTRIBUTION FROM THE DEPARTMENT O F CHEMISTRY O F COLUMBIA UNIVERSITY]
Organic Reactions under High Pressure. V. The Decomposition of Di-f-butyl Peroxide BY CHEVES WALLINGAND GERSHON METZGER RECEIVED MARCH23, 1959 The effect of pressures up t o 7300 kg./cm.2 on the decomposition of di-t-butyl peroxide a t 120" has been studied in four solvents. I n all, the rate is depressed by pressure, and values of AV* are calculated in cc./mole as 5.4 (toluene), 6.7 (cyclohexene), 12.6 (benzene) and 13.3 (CCL). The variation has been interpreted as arising from the competition between recombination of t-butoxy radicals, attack on solvent (both within the solvent cage) and diffusion out of the cage. AV* for diffusion is apparently ka, so that kobs = k2. Accordingly AV*'s measured (5-7 cc./mole) are essentially AVz*. On the other hand, in benzene and CCld, k d is negligible, and under pressure k5 < ka, so that
merization processes3s6and the difference may arise because here we are dealing with a very rapid lowactivation energy process in which little van der Waals' compression is required in the transition state. I n cyclohexene (Table IV) essentially no acetone was detected, presumably because the solvent reacts with the t-butoxy radicals before they can decompose. However, yields of t-butyl alcohol become as low as 30% a t 5840 kg./cm.2. Although no search was made for any resulting high molecular weight products, we believe that the loss represents t-butoxy radical addition to the cyclohexene, favored by the relatively small negative A V* for hydrogen abstraction indicated in the toluene system. Turning next to the unreactive solvents, the small amount of t-butyl alcohol noted in CCL has been attributed to some induced decomposition of peroxide, and the rate constants corrected accordingly (Table 11). Methyl radicals produced via reaction 7 in this solvent should yield methyl chloride which was detected in amounts approximately equal to the acetone formed. The other expected product, C2Cl6, was not investigated. I n benzene (Table I) some t-butyl alcohol is also formed, the amount increasing with pressure, although there is considerable scatter in the alcohol/acetone ratios observed. This might arise from direct hydrogen abstraction, induced decomkabn = k%kb/(ka kb) (9) position or via interaction between t-butoxy radiand A V o b s * is a composite quantity. Since the ra- cals and radical intermediates formed by addition to tio kj/(ka kb) decreases with pressure as diffusion the aromatic system. The induced decomposition seems unlikely, since there is no dependence of alout of the cage is retarded the larger values of A I/&* are accounted for. Further, if we assume that cohol yield on initial peroxide concentration. AV2* = G cc./mole, and that the reaction in CC14 However, the correction for any such induced reaction, if i t did occur, would yield an even larger represents the limiting case of k o b s = kS/kj/ka, value for AVObsi than that reported in Table I. d In kobs - A v 3 * - A v 2 * - A v h & (10) Finally, analysis of our reaction products indicated dP RT a t all pressures an amount of toluene equal to about Since A V3* is presumably negative, we conclude 10% of the acetone present. Alkylation of simple that A Vj* < 8 cc./mole. aromatic molecules by methyl radicals (presumably Effect of Pressure on Decomposition Products.by addition followed by some sort of disproportionAcetone and t-butyl alcohol arising from reactions 7 ation of the intermediate radical produced) has and 6 or 4, respectively, are the major products of frequently been postulated, and recently observed f-butyl peroxide decomposition. I n our experi- in the decomposition of acetyl peroxide in benzene.]& ments yields correspond closely to peroxide con- Our results are a further confirmation of the reality sumed (with the exception of reactions in cyclo- of this reaction. hexene discussed further below) and the effect of Other Data on Homolytic Dissociations.-IVe pressure is in good accord with the formulation of may now discuss published data on the effect of the reaction given in the preceding section. pressure on other homolytic dissociations (Table V) I n toluene (Table 111) t-butyl alcohol is the ma- in the light of our own results. The most interestjor product a t atmospheric pressure, and the yield ing data are those of Ewaldj on azobisisobutyroniof acetone formed decreases continuously with trile (AIBK) in which different values of A V* were pressure. X plot of log [t-butyl alcohol]/ [acetone] obtained when the reaction was followed photovs. pressure is linear, corresponding to A V,* - A V6* metrically (which determines the actual disappear(or AV4*, which is presumably the same) = 8.9 ance of AIBN) and by iodine disappearance (which cc./mole. Since AV7* is almost certainly positive measures the production of free radicals capable of to the extent of 3-6 cc./mole, AVs* is apparently reacting with iodine). Ewald has suggested that -3-6 cc./mole. This is smaller than values for the discrepancy arises from a "cage" effect, and we radical displacements reported previously in poly- may interpret it in more detail in the same manner (17) Viscosities of organic liquids increase markedly with pressure. as we have discussed solvent effects previously. References and a summary of d a t a are given by S. D. Hamman Where again substances in parentheses represent "Physico-Chemical Effects of Pressure," Academic Press, Inc., S e w fragments in the solvent cage, the pertinent reacY o r k , X . Y., 1957, pp. 81-84. Theory predicts a n inverse relation betlr-ern viscosity and diffusion, which is obeyed only qualitatively in tions are
+
+
+
+
t h e self diffusion of CSg, t h e only organic system €or which d a t a are available; R . C . Koeller and H . G. Drickarner, .I. Chrm. P h y s . , 21, 267
(1953).
(18) E L Eliel K Rabindran 'ind 5 H XViIen I Ory ( h u m 22 889 (1957).
ALKYLATION OF AROMATICS BY BENZYL ALCOHOLS
Oct. 20. 19.59 kn
AIBK
+ Kz
--+ (2R.)
(11)
kiz
(2R.)
+ R-R
(12)
ku
R.
( 2 R . ) + 2R. Iz ----f R I 1.
+
+
(13) (14)
Here the photometric experiment measures k11, hence AV11* = 3.8 cc./mole. kobs for the iodine experiment is actually knk13/(kn k n ) , or, under pressure where k13
