Dec., 1953
KINETICSOF REACTIONS OF Fe(I1)
sure. This method provides quick and reasonably precise surface tension values. It was used in these investigations because the method is independent of the angle of contact and density of the liquid being measured. This is particularly important when measuring interfacial tensions between two liquids.
AND
HYDROPEROXIDES BASEDON CUMENE
925
Because spreading is a function not only of the air-liquid interfacial energy but also of the solid-air and solid-liquid interfaces, i t is too early tjo generalize as to the wetting and penetrating effects of fluorochemicals. However, it is safe to say that fluorochemicals will contribute to the pure and applied science of surface chemistry.
KINETICS OF THE REACTIONS BETWEEN IRON(I1) AND SOME HYDROPEROXIDES BASED UPON CUMENE AND CYCLOHEXANE BY R. J. ORRAND H. LEVERNE WILLIAMS Research and Development Divison, Polymer Corporation Liniited, .Samia, Ontario, Canada Received April 0. 2865
An investigation of the rates of reduction of phenylcyclohexane, p-menthane and p-nitrocumene hydroperoxides by iron( 11) has been made. The Arrhenius equation was: k = 8 X 101Oe-laJm'Er 1. mole-' sec.+ for p-nitrocumene hydroperoxide ; k = 6.3 X 109e-11pIm/RT 1. mole-' set.-' for p-menthane hydroperoxide; and k = 2.4 X 109e-109m/RT1. mole-' set.-' for phenylcyclohexane hydroperoxide. From the change of frequency factor and activation energy with struct,ure of the hydroperoxide for these and other reactions investigated earlier, it was concluded that complex formation must take place between the iron(I1) and hydroperoxide, with subsequent decomposition to the final products. There is a relationship between the frequency factor and energy of activation. The mechanism of the electron transfer is explained in terms of Rollefson's concept of frequency factors. The behavior of the different hydroperoxides in emulsion polymerization is described and an attempt is made to correlate this behavior with the characteristics of the reactions studied.
Introduction hydroperoxide and cumene hydroperoxide. The General purpose chemical rubbers are made in an data showed that less reducer was required and the aqueous emulsion of butadiene and styrene or reaction mixture could tolerate inhibitors introduced acrylonitrile. For many years these copolymers with the monomers, the emulsifier, or other recipe have been prepared at 30 to 50'. It was found that ingredients. Smaller amounts of catalyst with better copolymers could be prepared at. lower poly- other recipe ingredients constant, resulted in little merization temperatures. A large number of pub- change in the initial rate of polymerization but sublications have described methods of preparing poly- sequent cessation of the polymerization reaction at mers at low temperatures, particularly 5 O , and the lower conversion of monomers to polymer. This fault could be overcome by bringing the amount of properties of the copolymers so prepared. Of great importance in the development of proc- reducer more nearly into molar balance with the esses for use at such temperatures was the discov- catalyst. Embree in this Laboratory has studied peroxides ery of cumene hydroperoxidel as catalyst. This hydroperoxide has been used with various reducing as catalysts in the butadiene-acrylonitrile r e ~ i p e . ~ substances, in particular iron(II),l ethylene dini- Similar results were obtained except that an optitrilotetraacetic acid salts2 and polyethylene poly- mal concentration of hydroperoxide for maximal amines. 3-6 of alternative hydroperoxides yield of polymer was observed which varied with have now been described, some of which are consid- the amount of other recipe ingredients present. ered superior to cumene hydroperoxide. I t is of From the conversion of monomers to polymer importance to show in what ways they are superior achieved in three hours, the relative effectiveness of and to understand the reason for the improved the hydroperoxides in promoting polymerization was, in decreasing order of effectiveness, t-butylcuperformance. Davidson in this Laboratory compared hydro- mene hydroperoxide, oxidized dipentene hydroperperoxides as catalysts in a low temperature recipe oxide, nitrocumene hydroperoxide, p-cymene hysimilar to that described earlier.2 The order of droperoxide, cumene hydroperoxide and s-butyldecreasing effectiveness, measured by yield of poly- benzene hydroperoxide. Ineffective were t-butyl mer and rate of polymer formation, was found to be hydroperoxide, benzoyl peroxide, dicumyl peroxide t-butylcumene hydroperoxide, isopropylcumene hy- and tetralin hydroperoxide. The rate of converdroperoxide, p-menthane hydroperoxide, p-cymene sion was found to be more dependent upon the concentration of hydroperoxide Qhanfor the buta(1) E. J. Vandenberg and G . E. Hulae, Ind. E n y . Chem., 40, 932 diene-styrene system. (1948). (2) J. IM. Mitchell, R. Spolsky and H. L. Williams, ibid., 4i, 1592 The relative effectiveness of peroxides as initia(1949). tors of polymerization has been studied by others. (3) W. H. Embree, R. Spolsky and H. L. Williams. i b i d . , 43, 2553 The most recent and comprehensive works are (1951). those of Coopergs10 and Swain, et al.," who studied (4) R. Spolsky and H. L. Williams, ibid., 42, 1847 (1950). (5) G. 8. Whitby, N. Wellman, V. W. Flouts and H. L. Stephens, substituted benzoyl peroxides and concluded that ibid., 49, 445 (1950). the substituent groups affocted the rate of polymer(6) G. 6 . Fisher, L. A. Goldblatt, I. Kniel and A. D. Snyder, i b i d . , 43, 671 (1951). (7) C. F. Fryling and A. E. Follet, J . Polymer Sci., 6, 59 (1951). (8) J. E. Wioklatz, T. J. Kennedy and W. B. Reynolds. ibid., 6, 45 (19511.
(9) W. Cooper, Nature, 162, 897, 927 (1948). (10) W. Cooper, J . Chem. Soc., 3106 (1951). (11) C. G. Swain, W. H. Stockmayerand J. J. Clarke, J . A m . Chem. Soc., 72, 5426 (1950).
926
R. J . ORBAND H.LEVERNEWILLIAMS
Vol. 57
ization through steric and resonance factors asso- ence of impurities which also reacted with iron(I1). ciated with the individual structures. It was sus- The over-all reactions would correspond to two pected that resonance and conjugation would also simultaneous primary reactions be factors in the effectiveness of hydroperoxides as Fe(I1) + PMHP --+ Products (1) initiators OI cold rubber polymerization through af+ Impurities ----f Products Fe(I1) (2) fecting the rate of decomposition of the hydroperThe initial assumption that - Jd[Fe(II)]/dtIl >> oxide. In addition, the rate of initiation of polymerization might be affected by the reactivity of - Id[Fe(II)]/dt\ 2 was made. The errors introthe free radical formed, a property which might be duced into a measurement of k~ by assuming - Idexpressed in a manner described by Herbst.12 It [Fe(II)]/dtl2 = 0, are due to the erroneous was also suggested7that there was an optimal water [Fe(II)]m value. It is necessary to have a meassolubility of the hydroperoxide but this did not hold ure of the Fe(I1) consumed by reaction 1 going to rigidly for all hydroperoxide^.'^ Earlier papers14-ls completion from which could be obtained an and the following results are part of a series intended [Fe(II)]m value valid for reaction 1. After the elapse of a certain time interval, the to elucidate the various factors. The work of [Fe(II)] time plot became nearly linear, as would be Kolthoff is also relevant.19 A study of the kinetics of reduction of these expected if the previous assumption as to the rate of hydroperoxides by iron(I1) must be made before it - Id [Fe(II) ]/dtl were true. Thus extrapolation is possible to find the reason for the varying ef- of this linear portion of the [Fe(II)]-time curve to fectiveness of these hydroperoxides in emulsion zero time at least partially compensates for the polymerization recipes. Such a study should Fe(I1) consumed in reaction 2 and yields a true also indicate the mechanism of electron transfer value of the [Fe(II)Im for reaction 1. This method would then be applicable to partially decomposed taking place in reactions of this nature. samples from the sodium salt as well as to unpurified Results hydroperoxides. For each experiment a value of The reaction of a hydroperoxide with iron(I1) [Fe(II)], for reaction 1 was calculated. This was consists of several steps. The primary radical- the value substituted in the rate equation for the producing step is followed by several radical-in- calculation of the rate constant. Hydroperoxide duced reactions. Some of these cause oxidation of samples for which purification had been attempted iron(I1) so that in the study of this reaction by but which gave different results from sample to measuring the rate of iron(I1) disappearance, it is sample, as well as hydroperoxide samples for which necessary to eliminate the side reactions. This is no purification had been attempted, were studied. done by conducting the reaction in a solution of Some typical plots are in Fig. 1 and the rate consome water-soluble monomer such as acrylonitrile. stants determined for this hydroperoxide are in When this is done, iron(I1) disappears according to Table I. a second order reaction for which the integrated TABLEI equation can be shown to be CORRECTED RATE CONSTANTS FOR p-MENTHANE HYDRO-In (1 - [Fe(II)],/[Fe(II)I) = [Fe(II)],kt + C PEROXIDE-FE( 11) REACTION k, 1. kw., T, lFe(%rlm mole-! seo.-1 1. mole-1 sea.-! OC. where [Fe(II) I m is the residual iron(I1) concen0 3.45 12.0 11.3 i0 . 2 tration, and [Fe(II) ] is the iron(I1) concentration 3.50" 11.8 at time t. This equation is such as to eliminate any 3.50" 10.5 effect from residual traces of oxygen which reacts 4.76 10.5 rapidly during the initial part of the experiment. 5,05" 10.0 p-Menthane Hydroperoxide-Iron(I1) Reaction. 6.10" 11.1 -Preliminary experiments were done using samples 6.24 12.4 of hydroperoxide which had been concentrated and 9 2.36 . 23.1 24.1 i 1 . 9 purified in the form of the sodium salt. The 4.74 22.1 results were quite irreproducible and differed when 4.90 26.9 various samples of hydroperoxide were used. This 5.85 22.6 was ascribed to the sodium salt being unstable and 6.00 25.5 decomposing into less active peroxides or hydroper16 1.34 38.3 33.5 f 3 . 8 oxides. Since some of these impurities were be3.15 34.3 lieved to be isomers, physical and chemical meth3.50 40.0 ods of separation did not show promise. Atten4.30 27.5 tion was directed to methods of studying the kinet5.02 32.0 ics of the oxidation-reduction reaction in the pres(12) R. L. Herbst. Jr.. J . Polymer Sei., 7 , 687 (1951). (13) Heroules Powder Company, private communication. (14) J. W. L. Fordham and H. L. Williams, J . Am. Chem. Soc., 7 2 , 4465 (1950). (15) J. W. L. Foidham and H. L. Williams, (bid., 78, 1634 (1951). (16) J. W. L. Fordham and H. L. Williams, Con. J . Reseorch, 278, 943 (1949). (17) J. W. L. Fordham and H. L. Williams, ibid., 288, 551 (1950). (18) R. J. Orr and H. L. Williams, Can. J . Chsm.. 30, 985 (1952). (19) I. M. Kolthoff and A. I. Medalia, J . Am. Chsm. Soo., 71, 3789 ( 1949).
6.10 30.1 7.10 32.6 5 Determined on unpurified p-menthane hydroperoxide samples.
The changes in the observed rate constants brought about by substitution of an [Fe(II)], value determined by extrapolation for a value which was observed directly differed from sample to sample according to the purity of the ,sample.
KINETICS OF REACTIONS OF Fe(I1) AND HYDROPEROXIDES BASEDON CUMENE
Dec., 1953
927
longed periods of time had elapsed; and (4) nonlinearity of the Arrhenius plot of the rate constants. Rate constants were determined a t three temperatures. Typical plots of data are in Fig. 2 and the rate
" 9 " " / I
I
0
8
10
IO
I
LO
LO MINUTES.
30
I
40
4U
I
80
Fig. 1.-Disappearance of iron(I1) by reaction with pmenthane hydroperoxide at 0". Final [Fe(II)] co : curve 1, 6.24 X 10-6 M; curve 2, 4.76 X M ; curve 3, 3.45 X 10-6 M.
The above data are only as reliable as the initial assumption that - Id(Fe(II))/dtl 2
