Kinetics of the Polymerization of Methyl Methacrylate initiated by Butylmagnesium Bromides and Dibutylmagnesium in Tetrahydrofuran+Toluene BY PETER E. M. ALLEN* AND BRETT 0. BATEUP? Department of Physical and Inorganic Chemistry, University of Adelaide, Box 498, G.P.O., Adelaide, South Australia 5001 Receiued 4th February, 19’75 The rates, up, and degrees of polymerization, .Dp,, are governed by relationships different from those of a typical anionic polymerization (e.g. those initiated by organic compounds of the alkali metals). At low temperatures (<230 K) termination of macrochains is insignificant. At fractional conversions a < 0.25 : - DP, = constant a/[BunMgBr],, up = -d[MMAj/dt = upo = ~ [ B U ~ M ~ B ~ ] ~ [ M M A ] ~ and at CI < 0.4 : where the subscript 0 implies an initial rate or concentration.The kinetics of polymerization when either ButMg or BusrvlgBr is used are similar. These relationships can be explained on the assumption that the rate and efficiency of initiation are governed by the reactions of an initiator-monomer complex and that the propagation reaction proceeds through a complex between monomer and the propagating chain-end. There have been several investigations of the polymerization of methyl metha- crylate (MMA) initiated by Grignard reagents in the presence of diethyl This is a solvent in which associated forms of the reagents are known to be p r e ~ e n t . ~ In this work we have used tetrahydrofuran (THF), a solvent in which Grignard reagents, other than the fluorides, and dialkylmagnesium compounds are unas- ~ociated.~ The molecular composition of the reagent is thus governed only by the Schlenk equilibrium (R,Mg + MgBr, + 2RMgBr).EXPERlMENTAL REAGENTS Grignard reagents were prepared from B.D.H. magnesium turnings, specified suitable for this purpose, in tetrahydrofuran distilled from calcium hydride under a purged atmosphere of 99.9 ”/, nitrogen gas. Iodine crystals were not used as activators. Didkylmagnesium compounds were prepared under vacuum by breaking a capsule of a solution of a appropriate Grignard reagent into dioxan. After two days and periodic shaking, the solution of dialkylmagnesium was filtered from the precipitated magnesium bromide- bisdioxan complex, diluted to the required concentration with THF and dispensed into break-seal capsules.were negative. MMA and solvents were purified by techniques known to be suitable for use in kinetic measurements on anionic polymerizations.6 -f present address C.S.I.R.O., Division of Textile Industry, P.O. Box 21, Belmont, Victoria 3216. Tests for residual bromide by Volhard‘s method 22032204 POLYMERIZATION OF METHYL METHACRYLATE KINETIC MEASUREMENTS Conventional dilatometry and adiabatic reaction calorimetry were used. Details are recorded Initial concentrations of initiator were estimated from an assay of the active alkyl- magnesium groups in the initiator solution. A sample of the solution was added to distilled water. Standard HCl was added to neutralize the MgOH groups formed. The excess acid was back titrated with standard NaOH using phenolphthalein as indicator. Each series of kinetic measurements was carried out with samples from a single batch of initiator and monomer.Rates of polymerization under comparable conditions were reproducible within 5 % when initiator from the same batch was used. However repro- ducibility between batches was not so good, ranging over a factor of 2-3. ISOLATION OF POLYMER Methanol-insoluble polymer was precipitated when a polymerizing mixture, ter- minated by opening the tube, was poured into a tenfold excess of methanol. HCl was added and the mixture stirred to dissolve inorganic material. The polymer was dissolved in benzene and recovered by freeze-drying. The filtrate was evaporated to dryness. Low molecular-weight methanol-soluble polymer and other products were separated from the inorganic material by dissolving in benzene.The solution was filtered and freeze-dried. MOLECULAR WEIGHTS Molecular weight distributions were determined for the methanol-insoluble polymer by gel-permeation chromatography in chloroform solution through a Waters Associates lo5 A linear column calibrated using Waters' standard samples and Dawkins' universal calibration method based on unperturbed dimensions.? Number- and weight-average molecular weights (Mn, Mw) determined by conventional direct methods would be of little value with the samples obtained because of their polymodal distribution and the accompanying low molecular weight material. The values cited are estimated from the g.p.c. traces and refer to the whole of the methanol-insoluble polymer where this had a unirnodal distribution.Where the distribution was polymodal, Mn and Mw of completely resolved peaks are cited. In the case of incompletely resolved, polymodal distributions, the approximate positions of the peak maxima are given. RESULTS EFFECT OF GRIGNARD ALKYL-GROUP STRUCTURE Polymers produced by t-butylmagnesium bromide were highly isotactic and those by n-butylmagnesium bromide predominantly syndiotactic. In neither case did the proportions of iso-, syndio- and hetero-tactic steric triads correspond to a Bernouillian distribution. Polymers produced by s-butylmagnesium bromide were of intermediate character. The overall rate of reaction increased with extent of alkyl-group branching in the initiator. At 223 K in toluene containing 2.9 mol dm-3 of THF, 2.8 mol dm-3 of MMA and 0.1 mol dm-3 of initiator the initial rate of polymerization was 8 x mol dm-3 s-l with BunMgBr, with But MgBr .with BusMgBr and an estimated 4 x EFFECT OF INITIATOR, INITIAL CONCENTRATIONS AND CONVERSIONS ON POLYMER MOLECULAR WEIGHT The size and size-distributions of the methanol-insoluble polymers depended on the initiator. Those obtained using n-butylmagnesium bromide had molecular weights in the 104-105 range and a fairly narrow distribution: M,/M, % 1.2. ThoseP . E. M. ALLEN AND B . 0. BATEUP 2205 produced with t-butylmagnesium bromide had higher molecular weights (- 3 x lo6) and narrower distribution. Initiation by s-butylmagnesium bromide produced very broad, polymodal molecular-weight distributions in which peaks corresponding to polymer similar in size to that produced by each of the other initiators could be discerned.Systematic identification of the parameters controlling the molecular weight distribution was possible only when n-butyl magnesium bromide was used as the initiator. At 223 K, the number average degree of polymerization En was proportional to conversion, a (a Q 0.25), and inversely proportional to the initial concentration of the initiator [BunMgBrIo, The most important feature of eqn (1) is the missing term. The degree of polymeriza- tion was found to be independent of the initial concentration of monomer, [MMA],. The molecular weight of samples taken at the same conversion (0.08) from solutions of the same concentrations of initiator (0.036 mol dm-3) and tetrahydrofuran (1.0 mol dm-3) remained constant within experimental error [an = (6.4k0.3) x lo4] while the initial monomer concentration was varied between 0.96 and 2.9 mol dm-3.Eqn (1) was tested with five samples prepared at 223 K in toluene solution with [THF] = 2.9 mol drn-3 and [MMAIo = 2.8 mol dm-3. [Bu"MgBr], varied from 0.05 to 0.19 mol dm-3 and conversion, a, from 0.06 to 0.16. The following relation- ships were tested by regression analysis with the results [and their 90 % confidence limits (c.l.)] indicated : Bn = constant x a/[Bu"MgBr],. (1) (i) log(~,[Bu"MgBr],) against log a was linear: (ii) M,[Bu"MgBr], against a was linear through the origin: slopek90 % c.1. = 1.1 kO.28 slope _+ 90 % c.1. = 37 500 9800 mol dm-3 intercept & 90 % c.1.= - 100 & 1320 mol dm-3. When the tetrahydrofuran concentration was reduced, the molecular weight of the polymer obtained at a given conversion was increased. At low concentrations ([THF] = 0.5 mol dm-3) the distributions were slightly narrower. The polymodal distribution of polymer produced by s-butylmagnesiuni bromide was sensitive to tetrahydrofuran concentration. The gel permeation chromatogram of methanol-insoluble polymer prepared at [THF] < 1.3 in01 dm-3 showed three incompletely resolved peaks at molecular weights ca. lo6, lo5 and lo4. Changing experimental conditions altered the heights of these peaks, but not their position. In polymer prepared at 223 K, with [MMA], = 0.96 and [BusMgBr], = 0.02 mol dm-3, the high molecular weight peak was prominent when [THF] = 0.5 mol dm-3, diminished at 1.3 mol dm-3 and just detectable at 4.6 mol dm-3.The low molecular weight peak increased with increase in tetrahydrofuran concentration. Narrow-distribution, high molecular weight polymer ( M , % 3 x lo6) was obtained if polymerizations initiated by t-butylmagnesium bromide were terminated after 2 min. Increase in reaction time did not influence the position of this peak, but some lower molecular weight polymers appeared which were absent at short reaction time. THE RATE EQUATION AT 223 K The rate equation was established by dilatometry. The final conversion on termina- tion was checked gravimetrically and found to agree within l % in all cases with that estimated from the total contraction.2206 POLYMERIZATION OF METHYL METHACRYLATE At fixed tetrahydrofuran concentration and temperature (223 K) the initial rate of polymerization obeyed the equation where [BunMgBr],, the initial concentration of initiator formally estimated as BunMgBr, corresponds to the initial concentration of BunMg groups present in the equilibrium mixture of Bu;Mg and BunMgBr.The external order of one with respect to [Bu"MgBr], was tested at 223 K, with [MMAIo = 2.8 mol dnr3 and [THF] = 2.9 mol dm-3, for eleven kinetic runs over a range [Bu"MgBr], = 0.05-0.20 rnol dm-3. By regression analysis: vpo = k[BunMgBr],[MMAfo (2) (i) log upo against log[Bu"MgBr], was linear : (ii) up, against [Bunh4gBr], was linear and through the origin : slopek90 % c.1. = 1.0520.14 slopef90 % c.1. = (7.51 f0.88) x intercept+90 % c.1.= (7.2+ 10.5) x s-l rnol dm-3 s-l. The external order of one with respect to monomer was tested at 223 K, with [THF] = 1.0 mol dm-3 and [BunMgBrIo = 0.036 mol dm-3 for five kinetic runs over the range [MMA],, = 1.0-4.1 mol dm-3. By regression analysis : (i) log up,, against lodMMAJ, was linear: (ii) upo against [MMAlo was linear through the origin: sl0pef90 c.1. = 0.90k0.24 slope490 % c.1. = (2.86k0.37) x s-l intercept+90 % c.1. = (9.22 10.8) x rnol dm-3 s-l. / P 2 4 6 8 t/103 s FXG. 1.-The dependence of fractional conversion : a = 1 -[MM~/[MMA]o (-) and hn(lMMAJo/ [MMA] (- - -) against time of reaction t at 233 K when MMAlo = 2.8 rnol dm-3, VHF] = 2.9 rnol dm-3 and [BunMgBrIo = 0.130(0) and 0.103(0) mol dm-3.P. E. M. ALLEN AND B . 0. BATEUP 2207 At 223 K, with [MMAIo -3 2 8 mol dm-3 and [THF] = 2.9 mol ~ I r n - ~ the rate of polymerization remained constant up to fractional conversions of 0.39 (fig.1). This implies that the internal order of reaction with respect to monomer is close to zero. In experiments carried to conversions a > 0.10 the deviation of the data from the integrated first order equation was obvious (fig. 1). Polymerizations initiated by s-butylmagnesium bromide, though faster, obeyed similar kinetic relationships. The initial rate was proportional to [BuSMgBrIo. The rate of polymerization remained constant until high conversion when the solutions became viscous and the rate decreased. ln([MMAJ,/[MMA]) = k't (3) SOLVENT EFFECTS The initial rate of polymerization declined as the concentration of tetrahydrofuran was increased after the manner shown in fig.3 for toluene solutions initiated by n-butylmagnesium bromide ([BunMgBr], = 0.1 mol dm-3, [MMA] = 2.8 mol dm-3, 223 K). A similar decline was observed when s-butylmagnesium bromide was used but it was not so pronounced. FIG. 2.-The effect of temperature (T/K) on the rate of polymerization, up (at conversions where up = upo), when UHF] = 2.9, [BunMgBrIo = 0.1 and [MMAIo = 2.8 rnol dm-3. The upper curve shows the rates observed as the temperature was increased stepwise in two separate experiments initiated below the inversion temperature, The lower curve shows the results for an experiment initiated at 263 K. TEMPERATURE EFFECTS There was an optimum temperature for polymerization at which the temperature dependence of the rate inverted.Below this temperature, Ti, an Arrhenius de- pendence, with a positive, apparent energy-of-activation (EBpp) prevailed. Above Ti, the rate of formation of macromolecular products decreased rapidly with increasing temperature. Fig. 2 shows the behaviour observed at concentrations [BunMgBr], = 0.10 mol dm-3, [MMA], = 2.8 mol dnr3, [THF) = 2.9 rnol d ~ n - ~ , as the tempera- ture was successively raised. Provided the initiation was carried out below Ti, the 1-702208 POLYMERIZATION OF METHYL METHACRYLATE rate of polymerization in successive sweeps of the temperature range using different samples having comparable concentrations, traced out the same curve : the upper one in fig. 2. If the initiation was carried out above the optimum temperature, the temperature-dependence followed a similar pattern but the rates were very sub- stantially reduced (fig.2, lower curve). The optimum temperature occurred at lower temperatures when the concentration of tetrahydrofuran was reduced (224 K when [THF] = 0.72 mol dm-3). The apparent overall activation energy, EaDp, below Ti, was 17+(90 % c.1.) 3 kJ mol-1 when [THF] = 2.9 and 19-t(90 % c.1.) 5 kJ mol-1 when [THF} = 0.72. *t I 2 3 4 [THF]/mol dm-3 FIG. 3.-The dependence of rate of polymerization, vpo, on the concentration of tetrahydrofuran [THFJ, in toluene solution at 223 K when [BunMgBrIo = 0.1 and [MMAIo = 2.8 mol d111-~. SPECTROSCOPIC OBSERVATIONS At temperatures above Ti an absorption peak at 296-300 nm developed during the early stages of the polymerization.This was absent if the polymerization tenipera- ture was kept below Ti. However, once formed, the peak remained even if the temperature was reduced below Ti. The peak was destroyed when the polymerization was terminated by methanol, but persisted when the polymerization was terminated by admitting oxygen. No trace could be found of peaks in the 330 nm region observed in the anionic polymerization of methyl methacrylate in similar solvent mixtures when Li+ and Naf6 are the counterions. However, a transient greenish-yellow colour appeared instantly the monomer and initiator were mixed at 223 K, but it was too short-lived (ca. 100 s) for the spectrum to be determined. The duration of this transient colour was greater at lower temperatures and tetrahydrofuran concentration.At 200 K, when [THF] = 0.76 mol dm-3, it persisted for almost 20 min, but no facilities were available for maintaining this temperature in a spectrophotometer. A similar, but less intense, transient colour was observed on initiation by Bu’MgBr, Bu”,g, BuiMg, but not by Bu‘MgBr or Bu‘,Mg. REACTION CALORIMETRY The temperature rise, corrected for a slight deviation from adiabatic conditions, did not settle down to a linear dependence on time until some 20 min after initiationP. E . M. ALLEN AND B . 0. BATEUP 2209 by BunMgBr, even though the conversion, as measured dilatometrically, was pro- portional to reaction time from the first measurement. An initial rapid temperature rise, decelerating to a steady state, was also observed with both s- and t-butyl- magnesium bromides. The heat capacities of the calorimeter and contents were determined using an inbuilt calibrated heating coil.When a solution of composition [MMAIo = 0.96 mol dm-3, [Bu"MgBrJo = 0.37, [THF] = 0.99, initially at 228 K, was polymerized to a fractional conversion, estimated by gravimetric determination of both methanol- soluble and insoluble polymer, of a = 0.135, the corrected, total temperature-rise immediately prior to termination was 4.1 K. This corresponded to an enthalpy loss of 52+ 5 kJ (mol of monomer polymerized)-l. [The enthalpy of polymerization lies in the range -(56+2) kJ mol-1.12a] In a similar experiment with s-butylmagnesium bromide (at a lower concentration, 0.02 mol dm-3, because of the faster reaction) the enthalpy change prior to termina- tion was 50 & 5 kJ (mol monomer polymerized)-'.DIBUTYLMAGNESIUM-INITIATED SYSTEMS Polymerizations initiated by di-n-butylmagnesium obeyed the same rate equations as those initiated by the corresponding Grignard reagent. An upward curve of the plot of eqn (3) was discernible in polymerizations taken beyond a conversion of 0.1 while the direct plot of conversion against time remained linear. The fit of the data to eqn (2) was confirmed by regression analysis. At 223 K, [BuzMg], = 0.05, [THF] = 2.6 mol dm-3: (i) log upo against log[MMA], was linear: (ii) up, against [MMA], was linear through the origin: slope+90 Xc.1. = 1.05-t.0.16 slope+90 % c.1. = (3.92k0.84) x s-I interceptk90 "/o c.1. = (0.8k6.7) x At 223 K, [MMA], = 0.50, [THF] = 2.6 mol dm-3: (i) log vp0 against log[Bu;Mg], was linear: slope$90 Xc.1.= 1.06k0.16 (ii) cp0 against 2[Bu;Mg], was linear through the origin: slope+90 % c.1. = (2.14f0.32) x mol d n ~ - ~ s-I. s-l intercept-(-90 % c.1. = (- 1.19k3.7) x niol dn? s-l. The rates of polymerizations initiated by B$Mg were slightly faster than those initiated by Bu"MgBr at a similar concentration of Bun-Mg bonds. At 223 K, [MMAIo = 0.96, [THF] = 1.1, [Bu"MgBr], = 2[BuZMg] = 0.036 mol the initial rates were 3.86 x for BuZMg. Very similar colours and U.V. absorptions were observed with the two initiators. The polymers produced had identical proportions of steric triads and very similar molecular weight distributions. Polymerizations initiated by BuiMg and Bu',Mg were respectively slightly faster and very much slower than those initiated by the corresponding Grignard reagents.Tn both cases the steric triad content and molecular weight distributions of polymers produced using dialkylmagnesium and the corresponding Grignard reagents were entirely different. mol dm-3 s-l for Bu"MgBr and 4.36 x2210 POLYMERIZATION OF METHYL METHACRYLATE DISCUSSION The temperature dependences of the kinetic behaviour in tetrahydrofuran +toluene and diethyl ether + toluene solution are very similar. In the latter the phenomena were attributed to a decrease in initiator efficiency with increasing temperature and the presence of reversible and irreversible termination of propagating chains above Ti. The substance absorbing at 290-3OOnm is probably a product of irreversible termination.A similar absorption has been seen as side products develop when the anionic polymerization of MMA (counter ion Na+) was carried out at near ambient temperatures.6 In both cases the peaks were unaffected by oxygen, but were destroyed by methanol, which suggests that they arose from alkoxide products of attack of initiator or propagating chains on the monomer-carbonyl. Our kinetic studies were therefore carried out at 223 K, well below Ti. to conversion, at least up to a = 0.25, is evidence that termination of propagating chains is insignificant. Initiation is very rapid. For such “ living ” systems the degree of polymerization is given by 12b At this temperature the proportionality of Dp,= 4 3 4 1 0 f [initiator] (4) which is a modification of Szwarc’s equation for “ living ” polymerizations.l Szwarc’s original equation, where the initiator efficiency f = 1 , is obeyed by a number of anionic polymerizations of hydrocarbon molecules, of which styrene is the best studied.13J4 There is good evidence l 5 that the same is true of the anionic poly- merization of MMA at 198 K. At higher temperatures however f < 1.6 Eqn (l), which we found applied to polymerizations initiated by n-butylmagnesium bromide, is only consistent with eqn (4) iffis proportional to the initial monomer concentration The fact that the polymerizations were slower at higher THF concentrations is an indication that they are not simple anionic mechanisms in which the propagating centres are ions, ion-pairs or equilibrium mixtures of these species.In these systems the proportions of more reactive free ions and solvent-separated ion-pairs increase as the solvent is made more polar. This is observed in the anionic polymerization of MMA in THF toluene mixtures where the second order rate coefficient for propa- gation is proportional to [THF].6 The absence of oxygen-sensitive U.V. absorptions at 290 and 330 nm associated with the propagating species in anionic polymerizations in the presence of the counter ions Na+ and Li+ l 1 confirms that the mechanism is different. Under certain conditions when the relaxation between the propagating species of the anionic polymerization of MMA is slow, very rapid polymerization occurs in the first seconds of the reaction and high conversions can be attained before the system settles down to a steady-state rate.6 Such phenomena can be missed in conventional dilatometer experiments as the steady state may be established Sefore the first reading can be taken.However in the present study all dilatometric conversions corres- ponded to those determined gravimetrically at the end of the experiment showing that the rates of contraction must have been steady from the beginning of the reaction. The high rates of evolution of heat in the early stages must have a different explanation and are presumably associated with either exothermic side reactions or complexing between monomer and initiator. The total heat evolved in both the steady-state and pre-steady-state periods is accounted for by the enthalpy of polymerization of mono- mer converted to polymer.This suggests that the pre-steady-state exothermicity “MI* f [MMAIO).P . E. M. ALLEN AND B . 0. BATEUP 221 1 must be primarily due to a reaction step involved in this course of events, i.e. complex formation. This is supported by the fact that although a significant proportion of initiator is involved in side reactions, the amount of monomer so consumed is much less than that incorporated in polymeric products. The transient yellow-green colour present in the initial stages may be due to monomer complexes. Complexes between MMA and triethylaluminium have a similar colour. 39 The kinetics of polymerization can be explained on the basis that both initiation and propagation involve an equilibrium, complex-forming step between initiator (I) and monomer and propagating chain (M,") and monomer.The additional conditions that eqn (1) must also be satisfied puts a considerable constraint on the stoichiometric and kinetic possibilities. The simplest and only really plausible mechanism is as follows. Initiation is a rapid but inefficient process and is complete early in the poly- merization. The mechanism proceeds through a monomer-initiator complex (1,M) which may either rearrange to give non-propagating side products or react with monomer to form the first of the propagating chain-carriers (M,*) Ki I+M+I,M ks I,M -+ side products ki I,M + M-+Ml* + . . . . (7) At low temperatures when termination is not significant the concentration of propa- gating centres, [IlO being the [MI0 % [I30 is once initiation is complete, is initial concentration of initiator.The c = f [ I l o (8) efficiency of initiation when The limiting case assumption is retrospectively justified by eqn (15). The propagation reaction also involves an intermediate complex KP kP M,*+M + M:, M -+ M:+l where c = c ([Mn*I+[M;,MI). n = l Under long chain-length conditions : which corresponds to both the observed rate equation (2) and the internal orders of zero (i.e. up = vpo up to at least a = 0.4). Substitution of eqn (9) (limiting value) into eqn (4) gives2212 POLYMERIZATION OF METHYL METHACRYLATE which is again consistent with the experimental observation [eqn (l)]. Comparison of eqn (14) with the regression coefficient of eqn (1) gives, at 223 K, [THF] = 2.9 rnol dm-3 and comparing this with eqn (1 3) and the regression coefficient of eqn (2) gives kJki = 375&(90 % c.1.) 10 mol dm-3 (1 5) k, f 0.09 s-l.The latter value is statistically reliable, but contains a systematic uncertainty in that the parameters of eqn (1) refer to high molecular-weight polymer only, whereas the rate- equation parameters contain some contribution from low molecular-weight products. It is interesting to note that the kinetics of polymerization of acrylonitrile at low temperature also take an unusual form when initiated by organomagnesium com- pounds. When organolithium compounds are used the system follows the typical kinetic equations of anionic polymerization. However when organomagnesium compounds are used eqn (13) and (14) apply. The mechanism we propose is very similar to Erussalimskii's mechanism l7 for acrylonitrile which in its turn is similar to the Fontana-Kidder' mechanism for the cationic polymerization of propene initiated by AlBr, and HBr.This project was supported by the Australian Research Grants Committee to whom we are grateful. H. Watanabe, J. Chem. SOC. Japan, 1961, 82, 362; J. Chem. SOC. Japan Ind. Chem., 1962, 65, 976, 1104. A Nishioka, H. Watanabe, K. Abe and I. Sono, J. Polynzer Sci., 1960,48, 241. P. E. M. Allen and A. G. Moody, Makromol. Chem., 1965,81,234; 83,220. G. E. Parris and E. C. Ashby, J. Amer. Chem. SOC., 1971,93, 1206. P. E. M. Allen, R. P. Chaplin and D. 0. Jordan, European Polymer J., 1972, 8, 271. A. I. Vogel, Quantitative Inorganic Analysis (Longman, London, 2nd edn., 1970), p. 259. J. V. Dawkins, R. Denger and J. W. Maddock, Polymer, 1969,10,154. 1°Polymer Handbook, ed. J. Brandup and E. H. Immergut (Interscience, New York, 1966), p. IV 48. l 1 D. M. Wiles and S. Bywater, J. Polymer Sci. B, 1964,2,1175. l 2 P. E. M. Allen and C. R. Patrick, Kinetics and Mechanism of Polymerization Reactions (Hor- wood, Chichester and Wiley, New York, 1974), (a) p. 210, (b) p. 162. l 3 R. Waack, A. Rembaum, J. D. Coombes and M. Szwarc, J. Amer. Chem. Soc., 1957,79,2026 ; M. Szwarc, Carbanions, Living Polymers and Electron Transfer Processes (Interscience, New York, 1968). l 4 G. V. Schulz, L. L. Bohm, M. Chmelif, G. Lohr and B. J. Schmitt, Kinetics and Mechanism of Polyreactions: IUPAC Int. Symp. Macromolecular Chemistry, 1969 (AkadCmiai Kiado, Buda- pest, 1971), p. 223, Fortschritte Hochpolym-Forsch, 1972, 9, 1. l 5 G. Lohr and G. V. Schulz, Makromol. Chem., 1973,172, 137 ; European Polymer J., 1974, 10, 121. l6 P. E. M. Allen, B. 0. Bateup and B. A. Casey, J. Organornetallic Chem., 1971, 29, 185. l 7 B. L. Erussalimskii, I. V. Kulevskaya and V. V. Mazurek, J. Polymer Sci. C, 1967, 16, 1355 ; B. L. Yerusalimskii, Polymer Sci. U.S.S.R., 1971, 13, 1452 (Vysokomol. soed. A, 1971, 13, 1293). 4E. C. Ashby, Bull. SOC. chim. France, 1972, 2133. ' B. 0. Bateup, Ph.D. Thesis (University of Adelaide, 1974). l 8 C. M. Fontana and G. A. Kidder, J. Amer. Chem. SOC., 1948,70, 3745.