Formation Rate Constants of the Mayo Dimer in the Autopolymerization of Styrene

Transcription

1 RAPID COMMUNICATION Formation Rate Constants of the Mayo Dimer in the Autopolymerization of Styrene THOMAS KOTHE, HANNS FISCHER Physikalisch-Chemisches Institut der Universitaet Zuerich, Winterthurerstrasse 190, CH 8057 Zuerich, Switzerland Received 16 July 2001; accepted 5September 2001 Published online 00 Month 2001; DOI /pola Keywords: styrene; autopolymerization; Mayo-dimer formation; 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO); activation energy; ESR/EPR; kinetics Since the beginning of polymer chemistry, the mechanism and kinetics of the reactions that initiate the spontaneous polymerization of styrene have been investigated intensively. 1 The major steps are displayed in Scheme 1, and they are well established by product analyses, kinetic evidence, and direct observations of intermediates. 1 Asfirst formulated by Mayo, 2 adominant Diels Alder reaction between two styrene molecules leads to the semi-benzene dimer 1with the rate constant k 1. It is spectroscopically observable, and there are actually two distinguishable isomers that differ in configuration at the chiral centers. 3 They are formed rather similarly but decay at quite different rates. 3 Inpure monomer, the dimers presumably undergo the retro-diels Alder reaction (k 1 ), and they react with styrene into the radicals 2and 3. 4 These radicals start the styrene propagation, couple to trimers, 5 and disproportionate to the tetraline derivative 4 and styrene. By acid-catalyzed reactions, 4 is also formed directly from 1. 6,7 In addition, the diphenyl cyclobutanes 6result from aconcurring minor reaction pathwayinvolvingthe1,4-diradical5that,inprinciple, could also lead to 1. Nitroxideradicalsgenerallyinhibitbutalsomediate living styrene polymerizations. 8 Inthe latter case, the styrene self-initiation affects the conversion rates and the control over the molecular weight and polydispersity of the resulting polymer. 9 Therefore, the reactions Correspondence to: H. Fischer ( hfischer@pci. unizh.ch) JournalofPolymerScience:PartA:PolymerChemistry,Vol.39, (2001) 2001 John Wiley &Sons, Inc. of Scheme 1have also found attention in styrene polymerizations mediated by nitroxides. Moad et al. 10 reported that nitroxide radicals such as 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO; T ) disappear in styrene much faster at 100 C than expected from the thermal polymerization rate in the absence of the radical scavenger. For TEMPO, aquantitative product analysis led to the formulation of the reactions shown in Scheme 2. T abstracts the weakly bonded hydrogen atom from 1to yield the radical 2and the hydroxylamine 7. The coupling of 2with asecond TEMPO molecule provides 8. This mechanism agrees with other known abstractions of hydrogen atoms by TEMPO, 11,12 and it is supported by the observation of practically equal amounts oftempoincorporatedinto7(44%)and8(45%).however, these equal yields can also be explained by the addition of TEMPO to the exo-methylene group of 1, which is followed by hydrogen abstraction from the resulting cyclohexadienyl-type radical. The addition of TEMPO to 1is not unlikely because the bis-tempo adduct 10 to styrene was observed by Moad et al. 10 ina minor yield of 10%, and its formation was confirmed later on by other authors. 12,13 However, at this stage the two possible routes to 7and 8cannot be distinguished. Other products containing TEMPO groups were not detected. The biphenyl cyclobutanes 6 (Scheme 1) accounted for about 7% of the total mixture of products. 10 The 1,4-diradical 5 must have a very short lifetime because it was not trapped by TEMPO, although the nitroxide concentration was 0.05 M, and TEMPO reacts with carbon-centered radical species with rate constants that are larger than 10 7 M 1 s

2 4010 J. POLYM. SCI. PART A: POLYM. CHEM.: VOL. 39 (2001) Scheme 1 Moad et al. 10 also reported that TEMPO decays in styrene with zero-order kinetics and the rate r Ms 1 at 100 C. They did not discuss this observation, but it is obvious that it may be related to the zero-order formation rate of the Mayo dimers 1. In this work, we show that it can be used to determine the total rate constant k 1 for the formation of the Mayo dimers and the activation parameters. Simple steady-state considerations reveal that the decay rate of TEMPO should obey d[t ] 2k dt 1 [S] 2 (1) where [S] is the styrene concentration and k 1 is the sum of the rate constants of formation of the two dimer isomers. The stoichiometric factor of 2 is obvious from Scheme 2. Equation 1 will hold if, first, TEMPO is converted only into 7 and 8. This is approximately valid because the yield of 10 is small. 10 However, the occurrence of the side reaction of TEMPO by addition to alkenes means that rate constants derived from eq 1 are upper limiting values and may deviate from the true values by up to 10%. Second, the concentrations of the carboncentered radicals and 1 should be approximately constant during the TEMPO decay. Although the steadystate assumption will certainly hold for the transient radicals, the dimer reaches a steady state in pure styrene only after about 50 h at 64 C and only after about 10hat90 C. 3(a) However, Hawker et al., 13 Boutevin and Bertin, 15 and Schmidt-Naake et al. 16 showed that styrene polymerizations are completely inhibited by M TEMPO at 125 C for about 4 h, and at 120 C, the induction period lasts about 14 h. Moad et al. 10 analyzed the products formed in 16 h at 100 C, that is, within the induction period. There, the dimer concentration must be very small because of nearly complete scavenging by TEMPO. This implies a stationary state of the dimers, and so the second condition of eq 1 is also justified. During the induction, the styrene concentration is also practically constant, of course. To obtain k 1, solutions of TEMPO (sublimed; Aldrich) in styrene (stabilizer-free and freshly distilled) were freed from oxygen at room temperature by several freeze pump thaw cycles on a vacuum line and sealed in glass tubes 5 mm in inner diameter. After insertion into the preheated probe head of an electron spin resonance spectrometer (Bruker EMX), the TEMPO signal was followed in time. Nearly perfect zero-order decays were observed for about 12 h at 91 C and for about 3.5 h at 121 C if the initial TEMPO concentrations were about 0.05 M or larger. No increase in the viscos-

3 RAPID COMMUNICATION 4011 Scheme 2 ity was detected, so the observations refer to the induction period. As expected from eq 1, two different TEMPO concentrations of and M gave practically identical decay rates of and Ms 1 at 120 C. Moreover, TEMPO solutions in styrene that were diluted with the inert solvent tertbutyl benzene (purified and distilled) in molar ratios of 1:1 and 1:3 provided zero-order decay rates at 120 C that were, by factors of and 0.069, smaller than those found for undiluted styrene. From eq 1, one expects a quadratic dependence of the decay rate on the styrene concentration, that is, factors of 0.25 and for the given dilutions, and the observed values are very close to these. Moreover, the addition of TEMPO to styrene becomes less likely upon dilution, so the agreement of the data supports the validity of the rate constants obtained from measurements with undiluted styrene. After these confirmations of the rate law, the decay rates were measured at various temperatures with M TEMPO solutions in pure styrene. The styrene concentrations were calculated from the temperature dependence of the density given by Hui and Hamielic. 17 Equation 1 was then used to convert the observed zeroorder decay rates into k 1. Figure 1 shows the results as closed symbols in the Arrhenius form. The regression line is represented by log(a) (M 1 s 1 ) and E a kj/mol with R (where E a is the activation energy, A is the frequency factor and R is the correlation coefficient). Also shown (by open squares) is the value of k 1 calculated from the decay rate reported by Moad et al. 10 It agrees well with our data. The other open symbols refer to the early work of Kirchner and Buchholz. 5,18 These authors deduced k 1 at 137 C rather indirectly from the rates of the trimer formation during the spontaneous polymerizations of diluted styrene in combination with the monomer conversion rate. 5 Later, they followed the formation of 1 directly by UV spectroscopy but used the earlier value of k 1 for calibration. 18 Their treatment involved several assumptions, and these appear justified now by the agreement of the data with the more directly measured values. To discuss the results, we first note that the quality of the Arrhenius plot agrees with the earlier observation that the two dimers are formed with rather similar rates. 3 Second, Kirchner and Buchholz 5 found that the formation rate of 6 (Scheme 1) varies quadratically with the styrene concentration as that of the dimer 1. At 137 C, the rate constant is M 1 s 1, and the activation parameters are log(a) (M 1 s 1 ) 5.7 and E a 108 kj mol 1. At the same temperature, 1 is formed about four times faster, and the Arrhenius parameters for the formations of 6 and 1 are different. This suggests that 1 is probably not formed via the diradical 5. However, this cannot yet be strictly ex-

4 4012 J. POLYM. SCI. PART A: POLYM. CHEM.: VOL. 39 (2001) Figure 1. Rate constants for the formation of Mayo dimer 1 at various temperatures: (F) this work, ( ) ref. 10, (ƒ) ref. 5, and ( ) ref. 18. cluded from these data because the rates are not very far apart. Third, Arrhenius parameters are known for various 2 4 Diels Alder additions. 19,20 The frequency factors range from log(a) (M 1 s 1 ) 4.1 to log(a) (M 1 s 1 ) 7.5. This means a large negative activation entropy that is caused by steric constraints on the transition state. The frequency factor of k 1, log(a)(m 1 s 1 ) 4.4, is within the accepted range, and this lends additional support to the dimer formation by a concerted Diels Alder process and not via diradical 6. Reported activation energies of Diels Alder reactions range from a low E a 36 kj/mol for additions that are facilitated by large polar rate enhancements 19 to E a 115 kj/mol for the addition of ethene to butadiene. 20 The value observed here indicates small polar effects, as expected from the absence of strongly electron-donating or electron-accepting substituents. Moreover, the small frequency factor further excludes large contributions of the loss of TEMPO by the addition reaction to 10 to the observed decay of TEMPO because radical addition rate constants to alkenes have considerably larger frequency factors. 21 Finally, in their classic modeling of the conversion rates and the number-average and weight-average molecular weights during the spontaneous polymerization of styrene between 100 and 200 C and up to very large conversions, Hui and Hamielec 17 found very reasonable fits for a conversion-independent third-order rate of initiation, I k i,eff [M] 3, if conversion-dependent ratios k p 2 /k t and k f /k p between the propagation (k p ), termination (k t ), and transfer constants to the monomer or the Mayo dimer (k f ) were taken into account. The initiation rate constant was expressed as k i,eff 2k 1 k i / k 1, and k i is the rate constant for the radical formation from 1 in Scheme 1. The derivation of k i,eff implied a stationary state of the dimer concentration, but this develops only slowly in pure styrene. 3 Moreover, the rate constant k 1 for the cleavage of the two isomers of 1 into styrene differs by 1 order of magnitude. 3 Hence, our total rate constant k 1 of the dimer formation cannot simply be combined with k i,eff to extract additional quantitative information about the elementary steps of the spontaneous styrene polymerization. The authors thank the Swiss National Foundation for Scientific Research for its financial support. REFERENCES AND NOTES 1. Pryor, W. A.; Lasswell, L. D. Adv Free-Radical Chem 1975, 5, (a) Mayo, F. R. J Am Chem Soc 1953, 75, 6133; (b) Mayo, F. R. J Am Chem Soc 1968, 90, 1289 and references therein. 3. (a) Kauffmann, H. F.; Olaj, O. F.; Breitenbach, J. W. Makromol Chem 1976, 177, 939; (b) Olaj, O. F.; Kauffmann, H. F.; Breitenbach, J. W. Makromol Chem 1977, 178, 2707; (c) Kauffmann, H. F. Makromol Chem 1977, 178, 3007.

5 RAPID COMMUNICATION In principle, the slow retrodisproportionation reaction between 1 and styrene to 2 and 3 could also lead to 3 and an isomer of 2 with the radical center at the phenyl-substituted carbon atom. However, because of the gain of aromaticity, the mechanism of Scheme 1 appears much more likely. 5. Kirchner, L.; Buchholz, K. Angew Makromol Chem 1970, 13, Buzanowski, W. C.; Graham, J. D.; Priddy, D. B.; Shero, E. Polymer 1992, 33, Compound 4 is certainly more stable than its isomer 1. However, the direct conversion of 1 into 4 would involve a very unlikely 1,3-shift of a hydrogen atom. 8. (a) Solomon, D. H.; Rizzardo, E.; Cacioli, P. U.S. Patent 4,581,429, March 27, 1985; (b) Georges, M. K.; Veregin, R. P. N.; Kazmaier, P. M.; Hamer, G. K. Macromolecules 1993, 26, (a) Catala, J. M.; Bubel, F.; Oulad Hammouch, S. Macromolecules 1995, 28, 8841; (b) Geortges, M. K.; Kee, R. A.; Veregin, R. P. N.; Hamer, G. K.; Kazmaier, P. J Phys Org Chem 1995, 8, 301; (c) Oulad Hammouch, S.; Catala, J. M. Macromol Rapid Commun 1996, 17, 8841; (d) Greszta, D.; Matyjaszewski, K. Macromolecules 1996, 29, 5239; (e) Greszta, D.; Matyjaszewski, K. Macromolecules 1996, 29, 7661; (f) Greszta, D.; Matyjaszewski, K. J Polym Sci Part A: Polym Chem 1997, 35, 1857; (g) Goto, A.; Fukuda, T. Macromolecules 1997, 30, Moad, G.; Rizzardo, E.; Solomon, D. H. Polym Bull 1982, 6, Opeida, I. A.; Matvienko, A. G.; Ostrovskaja, O. Z. Kinet Catal 1995, 36, For additions of nitroxide radicals to styrene and other alkenes and for abstractions by nitroxides, see (a) Connolly, T. J.; Scaiano, J. C. Tetrahedron Lett 1997, 38, 1133; (b) Aldabbagh, F.; Busfield, W. K.; Jenkins, I. D.; Thang, S. H. Tetrahedron Lett 2000, 41, Devonport, W.; Michalak, L.; Malmström, E.; Mate, M.; Kurdi, B.; Hawker, C. J.; Barclay, G. G.; Sinta, R. Macromolecules 1997, 30, (a) Beckwith, A. L. J.; Bowry, V. W.; Ingold, K. U. J Am Chem Soc 1992, 114, 4983; (b) Sobek, J.; Martschke, R.; Fischer, H. J Am Chem Soc 2001, 123, 2849 and references therein. 15. Boutevin, B.; Bertin, D. Eur Polym J 1999, 35, Han, C. H.; Drache, M.; Schmidt-Naake, G. Angew Makromol Chem 1999, 264, Hui, A. W.; Hamielec, A. E. J Appl Polym Sci 1972, 16, Buchholz, K.; Kirchner, K. Makromol Chem 1976, 177, (a) Sauer, J.; Wiest, H.; Mielert, A. Chem Ber 1964, 97, 3183; (b) Kerr, J. A.; Moss, S. J. Handbook of Bimolecular and Termolecular Gas Reactions; CRC: Boca Raton, FL, Rowley, D.; Steiner, H. Discuss Faraday Soc 1951, 10, Fischer, H.; Radom, L. Angew Chem Int Ed 2001, 40, 1341.