J. CHEM. SOC. FARADAY TRANS., 1994, 90(13), 1967-1972 I967 EPR/ENDOR Characterization of Radicals produced in the Photopolymerization of a Dimethacrylate Monomer Elena Selli, Cesare Oliva and Giorgio Termignone Dipartimento di Chimica Fisica ed Elettrochimica e Centro C.N.R.,Universita di Milano, via Golgi 19, I20133 Milano, Italy The formation of trapped radicals under UV irradiation of ethylene glycol dimethacrylate in the presence of different amounts of photoinitiator, as well as radical decay at different temperatures (90-150 "C)after the end of irradiation, have been studied by electron paramagnetic resonance (EPR) spectroscopy. A fitting procedure of the EPR signals revealed that the nine-line EPR spectrum can be attributed to the propagating radical, under- going a fast exchange process between two conformations.This radical also gives an ENDOR spectrum even at room temperature, thus suggesting that the polymer structure is essentially solid. At least one different radiFal species is present in highly photo-cross-linked samples. At temperatures above 100 "C, the hyperfine structure of the nine-line EPR spectrum is partly washed out, owing to a spin-spin exchange phenomenon between radicals. This becomes more evident with increasing degree of polymerization and cross-linking during radical decay kinetic runs and also affects the rate of exchange between the two radical conformations. Radicals produced during the photoinduced polymerization of methacrylate monomers have been thoroughly investigated by EPR spectroscopy.' Sufficiently high concentrations of long-lived radicals were obtained in early studies2 by addi- tion of difunctional monomers to methyl methacrylate.A densely cross-linked polymer network can thus be created. which hosts trapped radicals, exhibiting a nine-line EPR pattern. Variations of the EPR spectrum with temperature and the fact that the same pattern could also be obtained from UV, y-or X-irradiation of poly(methy1 metha~rylate)~-~ caused spe- culation about the radical assignment. The propagating free radical reported in Scheme 1 is now universally accepted to be the sole originator of the nine-line spectrum,*-' which arises from the hyperfine interaction of the unpaired electron with the freely rotating methyl group and with the p-methylene group, which is constrained in a certain conforma- tion of the polymer chain.However, good matches were obtained between the experi- mental EPR patterns and simulations based on different hypotheses involving constraint of the CH, group. The EPR pattern could, in fact, be simulated from a superposition of spectra due to two stable conformations," as a Gaussian dis- tribution of the dihedral angles about the most stable confor- mati~n,'.'~ as the result of exchange broadening due to hindered oscillations between two stable conformations,' ' or, finally, by a composite of the above models.' 371 In order to obtain more information, eventually supporting one of the above hypotheses, a different approach has been carried out in the present work.Following the recent studies y3 -CH,-C' I C ol" '0 I R 0 CH3 II I R = CH,-CH,-O-C-C=CH, (DMA) Scheme 1 reported by us on radicals trapped in photopolymerized and photo-cross-linked multiacrylate systems and on their ther- mally induced decay,16-' the EPR spectra of radicals trapped in a UV-irradiated difunctional methacrylate monomer, ethylene glycol dimethacrylate, have been recorded both under irradiation and during the post-irradiation radical decay at different temperatures (90-1 50 "C). The experimental EPR patterns have been fitted by an automatic non-linear least-squares procedure. Moreover, ENDOR spectra have been recorded for the first time with this photo- polymerized system.Experimental Materials Monomer and photoinitiator were commercial products. Ethylene glycol dimethacrylate (2-methylprop-2-enoic acid, ethane- 1,2-diyl ester), DMA (Aldrich), was washed twice with 2 mol dm-3 NaOH to remove inhibitor and then washed several times with saturated aqueous NaCl solution. It was successively dried over anhydrous sodium sulfate and molec- ular sieves 4A and stored at 4°C in the dark. The photoini- tiator, 2,2-dimethoxy- 1,2-diphenylethanone (BASF), was used as received. Sample Preparation The monomer, either pure or mixed with photoinitiator (0.3- 5.0 wt.%), was sealed under vacuum in quartz EPR tubes. In radical formation kinetic runs the EPR tubes were irra- diated directly in situ in the EPR cavity.The previously described apparatus and experimental procedure were employed.'* Photopolymerized DMA samples for radical decay and ENDOR studies were prepared by pre-irradiating the EPR tubes outside the spectrometer cavity, under the already reported' 6,1 experimental conditions. A double-bond con-version degree of around 63% was reached in the presence of 1-5% of photoinitiator, as revealed by analysis of residual unsaturations, carried out by Raman spectroscopy on a Perkin-Elmer Model 1720 FTIR spectrometer. The band at 1641 cm-' corresponding to the C=C stretching mode was monitored,20 using the band at 1729 cm-' due to the car- bony1 group as an internal standard. ENDOR analysis was carried out with the samples employed for radical decay kinetic studies, which had been stored at -18"C in the dark.EPRFNDOR Spectroscopy and Fitting Procedure EPR spectra under irradiation were recorded at 25°C by means of a Varian E-line Century Series EPR spectrometer, while EPR and ENDOR spectra of pre-irradiated samples were recorded at different temperatures on a Bruker ESP 300 EPR/ENDOR spectrometer, equipped with an EN1 3100LA RF (200 W) Power Amplifier. In both instruments, the tem- perature of the sample was kept constant at the desired value to within & 10C.16-18 EPR digitized spectra were fitted according to the pro- cedure already reported,' 9*21yielding width, hyperfine split- ting, g factor, spectral area of each overlapping EPR pattern and also parameters of linewidth variation resulting from dynamic processes involving radicals.The radical concentration, assumed to be proportional to the numerical integrated area of the corresponding EPR pattern, was in the range 10-3-10-2 mol dm-3, according to calibration of the two instruments by means of a Varian strong pitch. The best ENDOR signal was obtained with a microwave power of 6.3 mW and 10 dB (200 W) RF power attenuation. The RF modulation depth was kept to 100 kHz to achieve a good resolution. Results EPR Spectra at 25°C under Irradiation The well known nine-line spectrum reported in Fig. 1, typical of propagating methacrylate radicals (see Scheme l), was obtained at 25°C under UV irradiation in situ of DMA samples containing different amounts of photoinitiator. The shape of the spectrum did not change significantly with either irradiation time or photoinitiator content in the photoreac- tive mixture.Moreover, spectra identical both in shape and in intensity were recorded immediately after the light was turned off at the end of each radical formation kinetic run. Very good fittings of the experimental EPR spectra were obtained (see Fig. 1) by assuming that they result from the interaction of the unpaired electron with the three completely equivalent hydrogen atoms [with hyperfine coupling constant (hfcc), a3H = 22.7 f0.1 G] and, at the same time, with the Fig. 1 EPR spectrum at 25°C recorded under irradiation in situ of a DMA sample containing 0.3 wt.% of photoinitiator.Irradiation time: 450 min. (a) Experimental spectrum; (b) least-squares computer-synthesized spectrum. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 two P-methylene hydrogen atoms, characterized by a mean value of the hfcc, u~~ = 10.6 & 0.1 G, undergoing a fast chemical exchange. The central line of the latter multiplet is broadened by a contribution AWexchx 5.3 G to which the intrinsic linewidth AWo x 4.1 G is added. AWo increased slightly with irradiation time, i.e. as the polymer structure became more rigid and entangled, owing to further poly- merization and cross-linking. The concentration of radicals, R', responsible for the nine- line EPR spectrum increased continuously with irradiation time, as shown in Fig.2, being more than one order of mag- nitude lower in samples containing no photoinitiator. Photo- polymerization of DMA was quite slow in this case. Fig. 2 also shows that greater amounts of photoinitiator produced an increase in radical concentration. However, a maximum value was reached in the presence of 3% photoini- tiator, while a higher photoinitiator content was less effective in this respect. EPR Spectra at Different Temperatures The shape of the EPR spectrum of DMA irradiated in situ at 25°C did not change significantly in the temperature range from -120 to + 100°C. The fittings of the EPR spectra reveal that AWexchincreased only by ca. 7% if the tem-perature was lowered from + 100"C to -120 "C. In contrast, at higher temperatures (120-150 "C) some modifications were observed in the EPR spectral shapes.In fact, much better fittings of the experimental EPR spectra could be obtained by assuming that they are the super- position of the usual nine-line hyperfine structured pattern and a single-line pattern with the same g-value, having a line- width of around 10 G at 150°C. The mole fraction of rad- icals, S', giving this latter EPR pattern increased with temperature, being cu. 0.2 at 150 "C. In the framework of this fitting model, the exchange broadening, A Wexch,decreased, being about 30% less than in the absence of the single-line EPR pattern. Note that two new lines appeared on both sides of the EPR spectra with samples irradiated outside the spectrom- eter cavity, which are more homogeneously polymerized to a relatively high degree of conversion.At temperatures above 100"C, the two new lines also appeared in the EPR spectra of 250 200 h v)4-.-C :150 v m2? U$ 100 L rn 4-.-t 50 0 fi -ma 200 400 600 irradiation tirne/rnin Fig. 2 EPR nine-line pattern integrated area (arb. units) as a func- tion of irradiation time for DMA (A) pure or photopolymerizing in the presence of (a)0.3, (0) 3.0, (A)5.0 wt.% of photoinitia- 1.0, (0) tor J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Fig. 3 EPR spectrum at 120°C of DMA pre-irradiated in situ at 25°C in the presence of 0.3% photoinitiator (same sample as in Fig. 1). New EPR lines in the experimental spectrum (a) are indicated by arrows.(b) Least-squares computer-synthesized spectrum. samples pre-irradiated in situ, as shown in Fig. 3. They did not disappear if samples were cooled again to room tem-perature. Radical Decay Radical decay, investigated at 90-150°C in DMA samples behaviour was also reported recently by other authors.22 The results of the fitting procedure indicate that the mole fraction of S' radicals, giving the single-line EPR pattern, increased continuously during radical decay kinetics. This also occurred at 90°C, as shown in Fig. 5, although no single-line pattern could be detected in the EPR spectrum at the beginning of radical decay. The single-line spectrum rep- resented up to 90% of the whole integrated area at the end of radical decay at 150 "C.Fig. 5 also shows that, while the hyperfine-structured EPR pattern decreased continuously at 90-1 50 "C, the single-line spectrum increased continuously with decay time at 90 and Q 0.4 .-R J.a-5 0.2 \/ -c--E x.C ...n 0 100 300 400 time/min A) 90"C, (0,0)120kc,(m, 0)150°C. ,. Fig. 4 EPR spectra of DMA pre-irradiated in the presence of 1.0%photoinitiator at various times during radical decay at 150°C (noisy trace) and least-squares fittings of their central parts. Maximum spectral intensity was normalized. (a) 0, (b) 50, (c) 130, (d)310 min. 11 12 13 14 15 16 radiofreq u ency/M Hz Fig. 6 ENDOR spectrum of pre-irradiated DMA (3.0% photoinitiator) at (a) 150 K and (b)room temperature 120"C, while at 150°C it reached a maximum value and then decreased. The fitting of the EPR spectra also revealed that AWexchof the hyperfine-structured spectrum decreased during each radical decay kinetic run.A greater amount of photoinitiator in the photoreactive mixture caused a slight enhancement in the rate of radical termination, as already observed and discussed in previous studies on similar systems." ENDOR Spectra 'H ENDOR of pre-irradiated DMA (see Fig. 6)was detected with the magnetic field set at the centre of the EPR spectrum. Different settings always gave ENDOR patterns character- ized by an intensity roughly proportional to the intensity of the nine-line EPR pattern. At the lowest attained temperature (150 K) the ENDOR spectrum [Fig.qa)]was composed of a narrow singlet, over- lapping a broader feature. This pattern is qualitatively similar to that already rep~rted~~*" at similar temperatures with hexane- 1,Bdiol diacrylate, tetraethylene glycol diacrylate and butane-1,Cdiol diacrylate. However, in the present case, at variance with the previously investigated systems, an ENDOR broad band was also detectable at room tem-perature [Fig. qb)],the width of which was the mean of the linewidths measured at 150 K (see Table 1). Discussion Radical Formation under Irradiation At 25°C the EPR spectra detected under irradiation (Fig. 1) are identical to those obtained after the light had been switched off. This confirms that in both cases the observed nine-line EPR pattern is due to the propagating radical (see Scheme l), which is already trapped in the polymer matrix Table 1 ENDOR results sample photoinitiator content (%) T/K 6"/MHz rb/A 0.6 150 1 .OO9-0.194 4.3 -7.4 295 0.6 5.1 3 150 0.95-0.2 4.4-7.4 295 0.65 5.0 a 6, ENDOR peak-to-peak linewidth (maximum dipolar constant value).r, unpaired electron-proton distance. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 during irradiation. No reversible phenomena, such as those occurring by switching off and on the light in the presence of methyl methacrylate,' have been observed with the inves- 5923 tigated difunctional monomer. The fact that a maximum radical concentration value is obtained in the presence of 3% photoinitiator (Fig.2) demon-strates, as already discussed,18 that this amount of photoini- tiator represents in this case the optimum balance between a greater number of simultaneously initiated radical chains and the parallel increase of radical-terminating encounters, which are more probable at higher radical concentration. Also a filter effect, increasing with increasing photoinitiator concen- tration, could play a role in this respect, by reducing signifi- cantly the penetration depth of the radiation impinging on the sample. However, this effect should not be predominant, as the same extent of double-bond conversion (ca. 63%) was reached in the presence of different amounts of photoinitia-tor. Exchange between Radical Conformations Good fittings of the experimental EPR spectra have been obtained (see Fig.1 and 3) by assuming that the CH, group of the propagation radical (Scheme 1) undergoes a fast exchange process, causing a line-broadening phenomenon. The most likely explanation is that the two C,-H bonds are subject to hindered oscillations between two possible radical conformations, R,' and R,' , corresponding to two orienta- tions of the CH, group with respect to the C,-C, bond. The exchange time, z/s, between these two conformations can be evaluated from AW,,,,,/G, the exchange contribution to the width of the EPR lines characterized by a total nuclear spin quantum number, rn, = 0. The following equation has been employed :' I ye I = 17.61 x lo6 s-' G-' is the magnetogyric ratio, a, and a,, are the hyperfine coupling constants of the two protons involved in the exchange process. We have estimated a, = 15.1 G and a,, = 8.1 G, according to calculations' based on the assumption that in the two oscillating radical chain con- formations the dihedral angles of the two C,-H bonds rela- tive to the 7~ orbital of the unpaired electron are 60 k 5".This assumption is also strengthened by the fact that a mean value, ii,, = 11.6 G, is obtained with these two values, rather close to mean values of about 11 G obtained from the fitting procedures. The exchange time, z, is around 22 ns in the temperature range between -120 and + 100 "C, with an exchange activa- tion energy practically equal to zero.At higher temperatures, in contrast, the exchange phenomenon becomes faster (z z 14 ns at 150°C). However, z values do not seem to decrease with temperature according to an Arrhenius plot, as the exchange activation energy apparently increases continuously with increasing temperature, being around 20 kJ mol-' at 120- 150°C. This effect, which has already been observed in similar system^,'^ is clearly related to major modifications occurring in the polymeric structure. The glass-transition temperature of this polymer has been evaluated' very recently as ca. 140°C. Spin-Spin Exchange At temperatures higher than 100°C the EPR spectra are the superposition of the hyperfine-structured pattern, due to the exchange between conformations R,'and R2*, and the single- line pattern, due to the s' radicals.The mole fraction x(S') J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 increases during radical decay kinetic runs (Fig. 5), which are accompanied by further polymerization and cross-linking in the dark.16,26 Moreover, S' radicals decay at 150°C, but at a slower rate than R' radicals. Similar experimental evidence was observed by us in our previous EPR studies on pre-irradiated multifunctional acry- lates. A single-line pattern superposed on a hyperfine-structured pattern was attributed to radicals trapped in dense, highly cross-linked polymeric regions. 16-' In fact, spin exchange between radicals separated by distances com- parable to molecular diameters could be responsible for the lack of hyperfine splitting in the single-line spectrum. At higher temperatures, higher spin-spin exchange frequencies correspond to narrower EPR singlets.27 The single-line spectra observed in pre-irradiated DMA samples may thus be correlated with the fact that the polymer is not homogeneous.The propagation radical of Scheme 1 can undergo spin-spin exchange through its rigid polymeric surroundings. This effect is more evident at higher temperatures, as in previously investigated systems. l6 On the other hand, the presence of clusters of radicals has been detected in y-irradiated DMA6 and methyl metha~rylate.~ The increase in single-line mole fraction with radical decay, that is, with methacrylate double-bond conversion, and the slower decay of the single-line can be well explained by this model. However, in the case of pre-irradiated DMA the con- centration of s' radicals increases with time at high tem-peratures (Fig.5), while the total radical concentration ([R']+ [S.]) decreases. This implies that the fraction of propaga-ting radicals undergoing spin-spin exchange increases as the polymer becomes more entangled, owing to further poly- merization and cross-linking in the dark. Also the variation in EPR spectral shapes observed28 in the absence or presence of monomer in UV-irradiated poly(methy1 methacrylate) samples can be easily explained according to this model. Interplay between Dynamic Processes During radical decay kinetic runs at 90-150°C the rate of exchange k = 7-l between the two conformations, R,' and R2', of the propagating radical increases, together with an increase in the mole fraction of these radicals giving the single-line EPR pattern.Fig. 7 shows that the exchange time 25 1 I I I I I I 1 r i1 I 1 I I 1 I I 1 J 0 0.2 0.4 0.6 0.8 X(S7 Fig. 7 Exchange time between radical conformations R,' and R2', r/ns, us. mole fraction x(S'). Data have been calculated from the EPR spectra recorded during radical decay kinetic runs at 90°C (open symbols), 120"C (half-solid symbols) and 150"C (solid symbols). Amount of photoinitiator: (0,a)5.0%. a,0)1.0%; (A, A, A) 3.0%; (U, 0, values, z, for pre-irradiated samples containing different amounts of photoinitiator at various temperatures, are clearly correlated with the corresponding x(S') values.This means that the exchange Rl'+R2' does not involve R' radicals alone, but it can also occur through a intermolec- ular pathway involving S', according to the following equa- tion : R,' + S*=R2' + S' (1) Evidence for this reaction was also obtained in a similar system," with $3') values varying over a much smaller range. The present results clearly show that at low x(S') values the rate of exchange, z-l,undergoes a small enhance- ment due to reaction (I), while it increases for @') + 1. Presence of Different Radicals The propagating radical of Scheme 1 is thus responsible for both the hyperfine-structured line pattern and the superposed single-line pattern.However, in some EPR spectra, recorded with highly photo-cross-linked samples, two new lines appear on both sides of the usual pattern (Fig. 3). The presence of these lines, occasionally also reported by other author^,^^.^' was interpreted as an improbable superposed doublet with a splitting of 12.8 mT.29 We also think that these lines are not due to the propaga- ting radical, but to a different species, N', which is generated after long irradiation. This is also suggested by the observa- tion that the two extra lines never appear in spectra recorded during monomer irradiation in situ, where the irradiation on the sample is not uniform, owing to the shape of the EPR cavity. Furthermore, they appear and become more evident if the samples are kept at high temperatures for a certain time, thus allowing further polymerization and cross-linking in the dark, as well as chain transfer.However, all attempts to fit the EPR spectra assuming the presence of another superposed pattern, carried out in order to identify the second radical species N', did not give com- pletely reliable results, mainly because the shape of the central part of the experimental EPR spectrum does not change in an understandable manner when the two new lines appear. The most plausible hypothesis on N' is that it consists of the radical -CH2-C'(CH3)-CH,-, i.e. a mid-chain radical also undergoing a chemical exchange phenomenon, involving the two pairs of two equivalent C,-H bonds. The hfcc values, u~~ and a2H, obtained from the fitting procedure are close to those found for the propagating radical of Scheme 1.This mid-chain radical could be easily generated by homolytic chain scission in the a position to the carbonyl group. Its presence in similar systems has already been rep~rted.~"~' Also the delocalized allylic radical CH,"'C(CH3)=CH-, which has been proposed to be present in y-irradiated poly(methy1 rnetha~rylate),'~ is compatible with the spectral fitting. However, it should not be generated easily in the present case by UV irradiation, as this would imply a C-C bond scission in the polymer chain. In contrast, the radical y3 y3 -CH2-CdH----CdH2-I I C0,R C0,R which could easily be obtained by hydrogen abstraction from the polymer chain,14 should be discarded, as it is not compat- ible with the experimental EPR spectra on the basis of our fitting procedure.Radical Decay Kinetics During post-irradiation heat treatments, the decay of the total radical concentration is accompanied by an increase in the fraction of propagating radicals undergoing spin-spin exchange, as shown in Fig. 5. Thus an R' -+ S' step should be included in the mechanism in order to account for the experi- mental picture, together with the mutual termination steps involving the radicals R', S' and N, which should occur according to bimolecular reactions. Unfortunately, in this case we are not able to evaluate the rate constants of the single steps from the integrated areas of the EPR signals, which are supposed to be proportional to the radical concentrations.In fact, the amount of N' radicals cannot be evaluated from the EPR spectra with sufficient accuracy. ENDOR Analysis The trapped radicals, although affected by the dynamical phenomenon [see eqn. (I)], are essentially embedded in a solid phase. The ENDOR spectra (Fig. 6), which are typical of a solid matrix, strongly support this hypothesis. In fact, they can be interpreted, as in the previous case^,'^.'^ as due to matrix nuclear spins (protons), which interact with the unpaired electron spin by hyperfine coupling. A lower limit for the distance between them can be estimated using the equation : 6 = 80/r3 (2) where r is the proton-electron distance in 8, and 6 is the dipolar constant in MHz.6 is less than the experimental ENDOR linewidth. At 150 K the narrow singlet can thus be attributed to an interaction between the unpaired electron and protons at a distance of ~7.48, away. The broader ENDOR feature can be attrib- uted, using the same model, to an interaction between the unpaired electron and more closely placed protons, i.e. situ-ated about 4.3 A from the electron. At room temperature thermal vibrations would destroy the order in the sample, and the unpaired electron would interact with protons at a mean distance of about 5 A. Furthermore, the relationship between the intensities of the ENDOR and nine-line EPR patterns suggests that they are due to the same species, although the isotropic hyperfine couplings measured by EPR are no longer detectable by ENDOR.Conclusions Propagating radicals (Scheme 1) are trapped within the poly- meric structure produced by UV irradiation of difunctional methacrylates. The experimental nine-line EPR spectrum results from the interaction of the unpaired electron with the three completely equivalent hydrogen atoms of the methyl group and with the two P-methylene hydrogens, which undergo a fast chemical exchange between two different con- formations. The very minor changes observed in the EPR spectrum in the temperature range -120-+ 100 "C (although its shape is affected by a dynamic process) and the fact that a matrix ENDOR spectrum is detectable even at room temperature lead to the conclusion that radicals are surrounded by an essentially solid polymeric structure.A spin-spin exchange phenomenon between radicals trapped at relatively short distances within a rigid network can be observed at temperatures above 100"C, leading to the partial washing out of the hyperfine structure in the EPR spectrum. The fraction of S' radicals undergoing spin-spin J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 exchange increases with the degree of polymerization and cross-linking during radical decay kinetics, as expected. Moreover, the presence of S' radicals induces an enhance- ment in the rate of chemical exchange between the two con- formations R,' and R,' of the propagating radical, thus suggesting that this latter process could occur through an intermolecular pathway also involving S' radicals. Finally, at least one radical species other than the propaga- ting radical is formed by chain transfer in highly photo-cross- linked samples.We thank Prof. I. R. Bellobono for suggesting this research area, Dr. M. Barzaghi for helpful discussions and Dr. A. Giussani for preparing samples and recording some EPR spectra. This work was financially supported by Minister0 dell'universita e della Ricerca Scientifica e Tecnologica and by Consiglio Nazionale delle Ricerche through Progetto Finalizzato Chimica Fine. References 1 B. RAnby and J. Rabek, ESR Spectroscopy in Polymer Research, Springer-Verlag, Berlin, 1977. 2 N. M. Atherton, H. Melville and D. H.Whiffen, Trans. Faruday SOC.,1958,54, 1300; J. Polym. Sci., 1959,34, 199. 3 I. D. Campbell and F. D. Looney, Austr. J. Chem., 1962,15,642. 4 A. T. Bullock and L. H. Sutcliffe, Trans. Faraday SOC., 1964,60, 625. 5 F. Szocs and K. Ulbert, J. Polym. Sci., Part B, 1967,5, 671. 6 J. Zimbrick, F. Hoecker and L. Kevan, J. Phys. Chem., 1968, 72, 3277. 7 W. Kaul and L. Kevan, J. Phys. Chem., 1971,75,2443. 8 M. C. R. Symons, J. Chem. SOC.,1963,1186. 9 M. Iwasaki and Y. Sakai, J.Polym. Sci., Part Al, 1969, 1537. 10 M. Kamachi, M. Kohno, D. J. Liaw and S. Katsuki, Polym. J., 1978, 10, 69. 11 M. Kamachi, Y. Kuwae, S. Nozakura, K. Hatada and H. Yuki, Polym. J., 1981, 10,919. 12 M. Kamachi, Adv. Polym. Sci., 1987,82, 207. 13 M. E. Best and P. H.Kasai, Macromolecules, 1989,22,2622. 14 J. PlaEek and F. Szocs, Eur. Polym. J., 1989, 11, 11 49. 15 Y. Tian, S. Zhu, A. E. Hamielec, D. B. Fulton and D. R. Eaton, Polymer, 1992,33, 384. 16 I. R. Bellobono, C. Oliva, R. Morelli, E. Selli and A. Ponti, J. Chem. Soc., Faraday Trans., 1990,86,3273. 17 C. Oliva, E. Selli, A. Ponti, I. R. Bellobono and R. Morelli, J. Phys. Org. Chem., 1992,5,55. 18 E. Selli, C. Oliva, M. Galbiati and I. R. Bellobono, J. Chem. SOC., Perkin Trans. 2, 1992, 1391. 19 E. Selli, C. Oliva and A. Giussani, J. Chem. SOC., Faraday Trans., 1993,89,4215. 20 B. Chu and D. Lee, Macromolecules, 1984, 17,926. 21 M. Barzaghi and M. Simonetta, J. Magn. Reson., 1983,51, 175. 22 S. Zhu, Y. Tian, A. E. Hamielec and D. R. Eaton, Polymer, 1990, 31, 1726. 23 J. Shen, Y. Tian, G. Wang and M. Yang, Makromol. Chem., 1991,192,2669. 24 N. M. Atherton, Principles of Electron Spin Resonance, Ellis Horwood, London, 1993, ch. 9. 25 D. Li, S. Zhu and A. E. Hamielec, Polymer, 1993,34, 1383. 26 E. Selli, I. R. Bellobono and C. Oliva, Makromol. Chem., 1994, 195,661. 27 Yu. N. Molin, K. M. Salikhov and K. I. Zamaraev, Spin Exchange, Springer-Verlag, Berlin, 1980, p. 49. 28 R. E. Michel, F. W. Chapman and T. J. Mao, J. Chem. Phys., 1966,12,4604. 29 F. Szocs, J. Appl. Polym. Sci., 1986,32, 5673. 30 K. Hatada, T. Kitayama, E. Masuda and M. Kamachi, Makro-mol. Chem., Rapid Commun., 1990, 11, 101. 31 J. F. Kircher, F. A. Sliemers, R. A. Markle, W. B. Gager and R. I. Leininger, J. Phys. Chem., 1965,69, 189. 32 D. D. Leniart, J. S. Hyde and J. C. Vedrine, J. Phys. Chem., 1972,76,2079. Paper 4/01560F ;Received 15th March, 1994