J . Chem. SOC., Faraday Trans. I , 1984,80, 2911-2928 Effects of Hydrocarbon Diluents on the Kinetics of the Seeded Emulsion Polymerization of Styrene BY GOTTFRIED LICHTI-~ AND DAVID F. SANGSTER AINSE and CSIRO Division of Chemical Physics, Lucas Heights Research Laboratories, Private Mail Bag, Sutherland, New South Wales 2232, Australia AND BARRY C. Y. WHANG, DONALD H. NAPPER* AND ROBERT G. GILBERT Departments of Physical and Theoretical Chemistry, The University of Sydney, New South Wales 2006, Australia Received 22nd November, 1983 Kinetic relaxation studies (using y-rays from a 6oCo source as initiator) of the seeded emulsion polymerization of styrene have shown that the presence of hydrocarbon diluents, such as ethylbenzene, toluene and cyclohexane, can significantly increase the rate coefficient for the exit of free radicals from the latex particles.This increase was evident even when the chain-transfer constant for the diluent was less than that for styrene. Such an increase may arise from the low reactivity of the diluent free radicals with the monomer molecules. This allows the diluent free radicals additional time, compared with the monomer free radical, to undergo diffusional exit from the particles before reaction with the monomer confines the species to the latex particles. The free-radical exit rate coefficient is thus determined inter uliu by both the chain-transfer constant for the diluent and the reactivity of the diluent free radicals, as demanded theoretically. Large increases in the exit rate coefficient are predicted to occur with certain diluents at high replacements of the styrene by diluent. Seeded emulsion polymerizations of styrene initiated by potassium peroxydisulphate con- firmed that certain diluents may increase the exit rate coefficient significantly.Surprisingly, it was found that the entry rate coefficient may also be increased significantly. This was attributed to the presence of the diluent promoting the stabilization of those free-radical species that are capable of entering the latex particles at lower degrees of aggregation. This reduces the rate of bimolecular termination of the free-radical species located outside the latex particles. At higher monomer replacement levels, however, it is possible that the rate of entry of free radicals into the latex particles may be reduced as a consequence of the depletion of the monomer concentration in the aqueous phase.Thus the entry rate coefficient in the presence of diluents is determined inter alia by two processes whose effects tend to be opposed: on the one hand, enhanced stabilization of the coagulating colloidal free-radical species and, on the other, depletion of the monomer concentration in the aqueous phase. Seymour and coworkers' were the first to show that certain hydrocarbon diluents (e.g. benzene, cyclohexane and octane) retard the ab initio emulsion polymerization of styrene. They proposed that a decrease in viscosity within the latex particles was responsible for the rate reduction. Subsequently, Blackley and Haynes2 confirmed that four hydrocarbon diluents (benzene, toluene, ethylbenzene and cyclohexane) reduced the rate of emulsion polymerization in ab initio systems more than could be accounted for by the dilution of monomer.They showed experimentally that the decrease in rate was accompanied by a reduction in both the size of the latex particles produced and t Present address: ICI Australia, Central Research Laboratories, Ascot Vale, Victoria, Australia. 291 12912 SEEDED EMULSION POLYMERIZATION OF STYRENE the molecular weight of the polymer formed. These results were also interpreted as indicating that the hydrocarbon diluents diminished the Trommsdorff gel effect at the reaction sites within the latex particles. Azad et al.3 subsequently pointed out that any explanation for the retardation observed by Blackley and Haynes that invokes the Trommsdorff effect must be invalid because the average number of free radicals per particle (a) was always very small (in the range 0.002-0.02) in their experiments.Multiple occupancy of the particles by the radicals (i.e. values of n comparable to, say, one-half) is required for a significant Trommsdorff effect to manifest itself. These authors proposed instead that the reduction in rate observed by Blackley and Haynes2 was a consequence, at least in part, of the smaller size of the particles generated in the presence of hydrocarbon diluents. This increased the rate of exit of free radicals from the particles, leading to a smaller value of n and a slower rate of polymerization. In order to obviate the effects arising from a change in particle size, Azad et al.3 studied experimentally the effects of hydrocarbon diluents on the seeded emulsion polymerization of styrene.However, even under conditions of constant particle size all of the hydrocarbons studied decreased the polymerization rate more than could be accounted for by the dilution of monomer. These authors distinguished between two classes of diluents, which were postulated to reduce the rate of polymerization by quite different mechanisms. First, water-insoluble alkanes (e.g. n-octadecane and n-tetracosane), which are presumably not transported through the aqueous phase, appeared to act by lowering the thermodynamic activity of the monomer in the emulsion droplets and thus reducing the monomer concentration in the swollen latex particles.Secondly, lower-molecular-weight hydrocarbons (e.g. n-pentane and ethyl benzene), which by virtue of their solubility in water are capable of being transported through the aqueous phase, were considered to decrease the rate of polymerization by a chain-transfer mechanism. Chain transfer to the diluent was postulated to result in a low-molecular-weight free radical that was capable of diffusing from the particle. This reduced fi to below the value of one-half that was claimed to be operative in these experiments in the absence of diluents. There are several unresolved difficulties associated with the explanation proposed by Azad et aL3 First, if n were truly close to one-half, its value could only be reduced to the smaller values (calculated to be in the range 0.25-0.38) in the presence of diluents if the exit rate coefficient k were increased dramatically (e.g.fifty-fold). This follows directly from the steady state (ss) nss = 1 /(2 + k / p ) , where p is the rate coefficient for the entry of free radicals into the latex particles. Such a massive increase in k is not expected since the propagating polystyryl free radical has similar rate constants for chain transfer to styrene and ethylbenzene; moreover, both of the hydrocarbons have comparable solubilities in water. Secondly, although the chain- transfer constant for ethylbenzene is probably greater than that for styrene, those for benzene, toluene and cyclohexane are all significantly smaller than that for styrene. Yet Blackley and Haynes found that all four hydrocarbon diluents reduced the rate of polymerization per particle.In this paper we describe the results of additional experiments on the seeded emulsion polymerization of styrene in the presence of certain hydrocarbon diluents. These experiments were initiated by both y-radiolysis and a chemical initiator. The former allowed the exit rate coefficient to be measured directly by relaxation studies after removal of the system from the 6oCo initiating s o ~ r c e ; ~ the latter permitted the value of n to be varied fi~e-fold.~ The combination of these two types of experiment allowed the effects of diluents to be more clearly discerned. To do this we exploited recently developed technique^^-^ that enable the observed overall polymerizationLICHTI, SANGSTER, WHANG, NAPPER AND GILBERT 2913 kinetics to be resolved into the rate coefficients for individual microscopic events : specifically, those for free-radical entry into the latex particles and those for free-radical exit (desorption) therefrom.As these two events are mechanistically different, the presence of hydrocarbon diluents might be expected to affect them differently. Mechanistic deductions can be made more reliably from the effects of diluents on these microscopic rate coefficients than from the effects on the overall rate. EXPERIMENTAL All studies were performed using seeded latex systems in Interval I1 at a temperature of 50 "C. The methods and recipes used for both the relaxation studies (in which y-rays from a 6oCo source constituted the readily removed initiating source) and the chemically initiated studies have been described in detail previou~ly.~.The sizes of the seed latex particles, which were prepared by the methods described el~ewhere,~ were determined by both ultracentrifugation and calibrated electron microscopy. The radius of the particles used in the relaxation studies was 44 nm; those of the latices used in the chemically initiated studies were 29 and 51 nm for the systems with low and high initiator concentrations, respectively. Electron microscopy established the absence of nucleation in all of systems reported here. Note that the total weight of monomer plus hydrocarbon diluent (if present) at the commencement of each type of experiment was held constant so that the effects of the substitution of the monomer by hydrocarbon diluent could be explored.The precision of the data for the runs containing diluent was estimated by error analysis using the slope and intercept method. That for the runs in the absence of diluent was found from the reproducibility of some 6 runs. The relatively poor precision of the kinetic parameters obtained in the presence of diluents arose from the rapid attainment of the steady state. RESULTS AND DISCUSSION RELAXATION STUDIES Table 1 lists the values of the exit rate coefficient k determined by relaxation studies for the seeded emulsion polymerization of styrene, both when the monomer was pure and when the weight ratio of monomer to diluent present in the dilatometer at the commencement of polymerization was 70 : 30. These results were all calculated from the experimental data by assuming a value for the exited-free-radical rate parameter of a = + 1.This was the value of a found to be appropriate in previous relaxation studiess using the chain-transfer agents carbon tetrabromide and carbon tetrachloride and it corresponds to complete re-entry of the exited free radicals.6 Note, however, that the values calculated for k were relatively insensitive to the value of a adopted and they were also insensitive to the value assumed for A (provided A < 0.5) in the radiation cavity. Under these conditions, A was set equal to one-half, as demanded by previous kinetic5? and particle-size-distribution studiesg It is apparent from table 1 that the substitution of 30 wt% of the monomer by either ethylbenzene or toluene caused an approximately four-fold increase in the exit rate coefficient.Cyclohexane at the same level of substitution only doubled the value of k. Benzene, in contrast to the other hydrocarbons, induced no apparent change in k within the limits of experimental error. The simplest explanation for the effects of hydrocarbon diluents on the exit rate coefficient would assume that the diluents function as chain-transfer agents, thereby increasing the rate of production of small free-radical species. These small free radicals could then diffuse out of the latex particles, leading to an increase in the value of k. This simple model would imply that the value of k in the presence of diluents, whose 95 FAR 12914 SEEDED EMULSION POLYMERIZATION OF STYRENE Table 1.Measured exit rate coefficients for various diluents (particle radius, 44 nm; particle number, 5 x 10l6 dm-3) diluent k / 10-3 s-1 ~ k / 10-3 s-1 1.8 & 0.4 0 - et hy 1 benzene 6.6 1 .O +4.8& 1.1 benzene 1.8 0.4 0.0 f. 0.6 cyclohexane 3.6 l- 0.7 + 1.8k0.8 toluene 6.7 & 1 .O +4.9& 1.1 solubility in water is comparable to that of the monomer, could be expressed approximately as where ktr, refer to the chain-transfer constants for the monomer (M) and diluent (D), respectively, CM and CD are their respective concentrations inside the latex particles and Q is simply a constant of proportionality. It is readily evaluated because in the absence of diluents eqn (1) reduces to k M Q(ktr, M CM + ktr, n CD) (1) and ktr, k" = Qktr, M C M Q = k"/ktr, M C M (2) where the superscript " denotes the relevant values in the absence of diluents.Accordingly so that a revised version of eqn (1) is (3) Eqn (4) allows the values of k predicted by this simple theory to be calculated from the literature valueslO for ktr, M, ktr, and CMy5 together with the values calculated for CM and C, assuming ideal mixing. The latter assumption has been confirmed experimentally for ethylbenzene by Azad et aZ.3 The above analysis permits the theoretical change in the exit rate coefficient Ak arising from the presence of diluent to be calculated according to this simple model. The results so obtained are compared with those measured experimentally in table 2. It is apparent from table 2 that according to this model of the diluents studied only ethylbenzene is expected to increase the exit rate coefficient.This follows from ethylbenzene being the only diluent studied which had a chain-transfer constant that is apparently greater than that for styrene. The other diluents would be expected to decrease k. With the notable exception of benzene, however, all the diluents were found experimentally to give rise to significant increases in k . It is apparent that the simple theory outlined above is inadequate to explain the observed results. The failure of the simple concepts embodied in eqn (4) has been noted previously8 in relation to strong chain-transfer agents. There it was proposed that the reactivity with monomer of the small free-radical species generated by chain transfer was important in determining the magnitude of k.Reaction of these free radicals with monomer was postulated to reduce their solubility in water. This rendered them less able to diffuse from the latex particles. The importance of this mechanistic step in the process of exit even in the absence of diluents has been recognized previously byLICHTI , SANGSTER, WHANG, NAPPER AND GILBERT 2915 Table 2. Comparison of the predictions of the simple transfer theory with the results of experiment Ak/ lop3 s-l k r , D C D diluent /lo+ dm3 mol-l s-l /mol dm-3 simple theory experiment 1 .oa et hy 1 benzene 1.5 toluene 0.25 benzene 0.03 cyclohexane 0.05 - - - - 1.8 + 0.3 +4.8+ 1.1 2.0 - 0.4 +4.9+ 1 . 1 2.2 - 0.6 0.0 f 0.6 2.1 - 0.6 + 1.8k0.8 Ugelstad and Hansen,ll Nomura and Haradal29 13 and Hawkett et aL5 The value of k would be expected to be proportional to the ratio ktr, CM/kb CM, where k;, is the propagation rate constant for the reaction between the monomeric styryl free radical and ~ t y r e n e .~ ' ~ ~ . ~ ~ Note that, as pointed out by Ugelstad and Hansen,ll k; is not necessarily equal to the conventional propagation rate constant k, relevant to the polymeric free radical. The physical insight implicit in the above ratio is that the rate of free-radical exit depends not only on the frequency (= ktr, C,) with which new free radicals are generated by chain transfer but also on the average time spent by these free radicals before adding monomer units, which results in their confinement to the latex particles. It is possible to formalize the above concepts for diluents whose solubility in water is not too different from that of the monomer using the relationship where R is the constant of proportionality and kpD is the rate constant for the reaction of the diluent free radical with monomer.Eqn (5) is obtained readily from consideration of the total exit rate of the separate free-radical species that arise from transfer to monomer and diluent. It thus embodies the notion that exit of free-radical activity from the latex particles can occur by the diffusion of both the monomeric styryl free radicals (M') and the diluent free radical (DO) from the latex particles. Further, the free radicals M' and D' are assumed to be of comparable solubility in water. In the absence of diluents, the exit rate constant can be written as k" = Rktr,,/kb so that eqn ( 5 ) can be recast as It follows that according to this approach Eqn (8) implies that the change in k is directly dependent upon the ratio of the rates of transfer to diluent and monomer, and inversely related to the relative propagation 95-22916 SEEDED EMULSION POLYMERIZATION OF STYRENE Table 3.Calculated values for the relative reactivities of the diluent and monomeric free radicals diluent 1 .o - ethylbenzene 0.25 toluene 0.05 cyclohexane 0.03 benzene - rates of D' and M'. Note that, in contrast to eqn (I), one essential prediction of eqn (8) is that the value of k must always increase or remain constant, as was indeed observed experimentally. Eqn (8) can be used to calculate the values of the ratio k,,/kL from the measured values of Ak and the literature values for the chain-transfer constants.The values so calculated are presented in table 3. It is apparent that, according to this approach, the free radicals produced by chain transfer to the diluents ethylbenzene, toluene and cyclohexane are all less reactive towards styrene monomer than is the free radical (presumably CH&H Ph and/or CH,=tPh,l* where Ph = phenyl) generated by chain transfer to styrene. The reduced reactivity of the diluent free radicals with monomer means that such free radicals can undergo diffusion within the latex particles for a longer period of time than styryl free radicals before monomer addition occurs. The latter results in the free radicals being confined to the domains of the latex particles.The lengthened period of diffusional motion by the unreacted free radical increases the probability of escape of the free radical from the particle and so leads to the observed increase in k . Note that the data presented in table 2 suggest that the more difficult it is to generate a free radical by chain transfer, the slower is its reaction with monomer. This conforms with the general rule that governs the reactivity of free radicals with a given monomer in copolymerizations. One possible explanation for this rule lies in the stabilization of the transition states involved in the two proces~es.~~ Any mechanism (e.g. electron delocalization) that stabilizes the transition state of the hydrogen-abstraction reaction between the free radical and diluent necessarily promotes the rate of chain transfer.A very similar mechanism is likely to be operative in stabilizing the transition state of the addition reaction between the diluent free radical and monomer, thus promoting that reaction as well. CHEMICALLY INITIATED STUDIES In the experiments to be described below, the seeded emulsion polymerization of styrene, in both the presence and absence of diluents, was initiated by potassium peroxydisulphate. It was possible in favourable circumstances to measure both the steady-state rate and the approach to the steady state. This allowed the slope and intercept method5* to be used to give values for the rate coefficients for both the entry of free radicals into the latex particles and the exit of free radicals therefrom.Previous studies6T8 have shown that the value of the exited-free-radical rate parameter a that is appropriate in the presence of chemical initiators lies in the range - 1 < a < 0. This corresponds physically to significant heterotermination of the exited free radicals in the aqueous phase.6 The values obtained for the entry and exit rate coefficients wereLICHTI, SANGSTER, WHANG, NAPPER AND GILBERT 2917 relatively insensitive to the precise value chosen for a in this range and so a was set equal to - 1. The extent of substitution of the styrene by the hydrocarbon diluent will be expressed by the initial diluent weight fraction, d,", defined such that d,"m, is the weight of diluent in the dilatometer at the beginning of the experiment.(m, is the corresponding total weight of monomer plus diluent in the dilatometer.) Note that as polymerization proceeds the value of the overall diluent weight fraction increases; its actual value is given by &[l -x(l -&)I, where x is the fraction of monomer converted to polymer. In the quantitative determinations of the rate coefficients to be presented below the numerical data were all derived for x < 0.2. Accordingly, changes in the diluent weight fraction were relatively small and have been neglected. The kinetic results at different initial diluent weight fractions will be presented as plots of the percentage conversion of monomer to polymer as a function of time. This form of presentation was adopted because if simple dilution effects alone were operative, the rate curves would all be identical with that observed in the absence of diluent. Moreover, at constant particle number, the gradient of such curves at any point is proportional to the rate of polymerization per particle expressed relative to the initial amount of monomer present in the system.It is convenient to discuss the results in terms of this relative rate of polymerization per particle rather than the absolute rate of polymerization per particle. The latter will necessarily vary with the extent of the substitution of monomer by diluent. Note that given sufficient time (usually between 24 and 72 h), all polymerizations were found to proceed to complete conversion. Note also that the conversion was measured at intervals of 1 min. This resulted in the data points being too close together to be presented individually in the kinetic curves.A smooth curve has therefore been drawn through the observed points. The results obtained with ethylbenzene as the hydrocarbon diluent were both qualitatively and quantitatively unambiguous. These will therefore be discussed first and in some detail. The results obtained for toluene, benzene, cyclohexane and 11-heptane, whilst being qualitatively unambiguous, proved to be less clear cut quantitatively. Therefore only differences from the effects observed with ethyibenzene will be discussed in these cases. E:THYLBENZENE The effects of ethylbenzene at various initial diluent weight fractions on the kinetics of the seeded emulsion polymerization of styrene at two different initiator concentrations (1.2 x lop2 and 1.2 x mol dmP3) are displayed in fig.1 and 2. It is apparent that the influence of the diluent was different at the higher and the lower initiator concentrations. At the higher concentration, the curves do not coincide, as would be expected if the only factor to be taken into account was dilution of monomer. The results imply that an additional mechanism which reduces the rate of polymerization to below that expected for dilution alone was operative. The magnitude of this additional reduction in rate increased monotonically with increasing mono- mer dilution. At the lower initiator concentration, however, the relative rate of polymerization first increased and then decreased as the monomer became progressively more dilute (see fig.2). Stated differently, this meant that at relatively low levels of replacement of styrene by ethylbenzene, the rate of polymerization was not reduced to the extent expected by simple dilution considerations. This was despite the fact that Azad et al.3 have shown experimentally that the swelling of polystyrene latex particles by styrene + ethylbenzene mixtures closely follows the behaviour expected for the ideal mixing of diluent and monomer. Fortunately, at these low monomer-replacementSEEDED EMULSION POLYMERIZATION OF STYRENE 2918 100 80 8 60 E: .- L W 40 V 20 I I I 0 50 150 200 Fig. 1. Effect of ethylbenzene on the rate of polymerization of styrene in seeded systems at a high initator concentration (1.2 x mol dm-9. Diluent weight fraction 4: curve 1, 0.00; 2, 0.15; 3, 0.26; 4, 0.50; 5, 0.75.loo t/min 80 - h rf 60 - 0 150 200 loo t/min 50 Fig. 2. Effect of ethylbenzene on the rate of emulsion polymerization of styrene at a low initiator concentration (1.2 x lop4 mol dm-9. Diluent weight fraction 4: curve 1,O.OO; 2,0.15; 3,0.26; 4, 0.50; 5 , 0.75. levels the slope and intercept method allows the free-radical entry and exit rate coefficients to be determined. The results obtained are listed in table 4. The values of k determined from these chemically initiated experiments show that the presence of ethylbenzene increased the exit rate coefficient. This is in accord with the relaxation results presented above. What is surprising, however, is that the entry rate coefficient pA( = p + kfi, when a = - l), relating to the production from the initiator of free-radical species that could enter the latex particles, was also increasedLICHTI, SANGSTER, WHANG, NAPPER AND GILBERT 2919 Table 4.Effects of diluents on the entry and exit rate coefficients (particle concentration, 3 x 10'' dmP3) particle [I1 diluent radius/nm d; mol dm-3 nss s-la k/lOP3 s-' - - ethylbenzene 5 1 0.00 120 0.43 51 0.26 119 0.38 51 0.50 120 0.33 51 0.75 119 0.24 29 0.00 1.2 0.08 2.3f0.3 1.4 f 0.5 29 0.15 1.2 0.10 6.4f1.5 3.0 + 1 .O 29 0.26 1.2 0.10 8.1 f 2 . 0 3.6f 1.5 - - - - - - toluene - - 51 0.00 120 0.43 51 0.26 118 0.38 - - 51 0.50 118 0.37 51 0.75 116 0.34 29 0.00 1.2 0.08 2.3f0.3 1.4 f 0.5 29 0.15 1.2 0.08 3.9k 1.0 2.1 f 1.0 29 0.26 1.2 0.08 2.5+ 1.5 1.5+ 1.5 - - - - benzene 29 0.00 1.2 0.08 2.3k0.3 1.4 f 0.5 29 0.15 1.2 0.10 3.2f 1.0 l.5f 1.0 29 0.26 1.2 0.07 3.4+ 1.5 2.3f 1.5 cyclohexane 29 0.00 1.2 0.08 2.3k0.3 1.4 f 0.5 29 0.15 1.2 0.09 4.5f1.0 2.2 f 1 .o 29 0.26 1.2 0.08 3.2+ 1.5 1.9+ 1.5 n-heptane 29 0.00 1.2 0.08 2.3f0.3 1.4 f 0.5 29 0.15 1.2 0.08 2.4f 1.0 1.4+ 1.0 29 0.26 1.2 0.07 2.3k 1.5 1.6f 1.5 significantly (e.g.by a factor of ca. 3). The reality of this increase was supported by a significant decrease (from ca. 30 to ca. 10 min) in the induction period in the presence of ethylbenzene. This increase in the rate of production of free-radical species capable of entering the latex particles is responsible for the rate of polymerization not decreasing in accordance with simple dilution requirements. The explanation for the dramatic increase in the entry rate coefficient in the presence of ethylbenzene is not immediately apparent.It appears to be unlikely that a sparingly water-soluble substance, such as ethylbenzene, could promote molecule-induced homolysis of the initiator and so cause a large increase in the rate of production of primary free radicals. It seems more likely that ethylbenzene reduces the rate of bimolecular termination of primary and oligomeric free radicals located outside of the latex particles. Such bimolecular termination reduces the radical capture efficiency to < 100% . There are several possible mechanisms by which bimolecular termination could be reduced but, whatever the mechanism, it is presumably closely related to the mechanism by which ethylbenzene, and other diluent hydrocarbons, reduced the particle size in the ab initio emulsion polymerizations of styrene reported by Blackley and Hayne.s2 Our studies16 of the nucleation mechanism in the emulsion polymerization of2920 SEEDED EMULSION POLYMERIZATION OF STYRENE styrene suggest that there is a range of oligomeric species that can enter the seed latex particles.These include colloidal precursor (or primary) particles, which are aggregates of ‘insoluble’ oligomeric species that are characterized by a very slow growth rate. This probably arises from the poor swelling of such aggregates by monomer as a consequence of their residual hydrophilic character and/or their relatively small size. Free radicals associated with such precursor particles propagate more slowly than those in mature latex particles.They would also be expected, on diffusional considera- tions, to enter the seed particles more slowly than surfactant-like oligomeric species. In the context of the preceding initiation mechanism, one possible way by which ethylbenzene could increase the rate of free-radical entry into the seed particles is by promoting the stabilization of the precursor particles by the surfactant at relatively low degrees of aggregation. This appears to be a synergistic phenomenon in that the effect of surfactant plus ethylbenzene seems to be greater than the sum of the effects of surfactant and diluent acting separately. This synergism may arise from the ethylbenzene rendering the precursor particles more lipophilic because, unlike the monomer, the diluent molecules do not polymerize.Futhermore, ethylbenzene is a good solvent for polystyrene. In these circumstances the surfactant could be adsorbed more readily onto the precursor particles than in the absence of diluent. Such a mechanism would explain the increase in the number of particles produced in the presence of diluents in the ab initio emulsion polymerizations reported by Blackley and Haynes.2 In addition, the above mechanism is consistent with the large increase in the free-radical entry rate observed in seeded systems in the presence of ethylbenzene, since bimolecular termination, which accompanies the coagulation of precursor particles, would be significantly reduced. It might at first sight be thought that the explanation proposed above for the mechanism whereby ethylbenzene increased the free-radical entry rate is in conflict with the results of Piirma and Chen.18 These authors found that, at equilibrium, the saturation amounts of anionic alkyl surfactants adsorbed onto unit surface area of polystyrene latex particles decreased when the particles were swollen by hydrocarbons.It should be recalled, however, that precursor particles are postulated to be different from mature polystyrene latex particles inasmuch as they are more hydrophilic in character. Their behaviour might be expected to resemble more closely that of more polar polymers, such as poly(methy1 methacrylate). Piirma and Chen indeed showed that, in contrast to the results for polystryene, the adsorption of sodium dodecyl sulphate increases significantly when poly(methy1 methacrylate) latex particles are swollen by a hydrocarbon (benzene).This observation is in accord with the concepts postulated above for precursor particles. Note, however, that kinetic factors may also contribute to any increase in surfactant adsorption by the precursor particles in the presence of hydrocarbons. It is also possible that the increase in the free-radical exit rate coefficient induced by the presence ofethylbenzene may contribute to the increase in the overall free-radical entry rate coefficient. It was mentioned above that precursor particles would be expected to enter the seed latex particles more slowly than oligomeric surfactant-like species. If the free radicals trapped inside precursor particles were released by chain transfer and exit to form additional aligomeric species in the aqueous phase, the rate of entry of free radicals into the seed particles could be enhanced.As noted previously, inspection of the rate curves shown in fig. 2 reveals that the rate of polymerization, expressed relative to the amount of monomer originally present in the system, first increased on replacement of monomer by ethylbenzene. However, as the extent of monomer substitution increased, the relative rate of polymerization dropped significantly so that the relative rates at 50% and 75%LICHTI, SANGSTER, WHANG, NAPPER AND GILBERT 292 1 substitution levels were both considerably smaller than that in the absence of diluent. Note that the same pattern of behaviour was observed at all degrees of monomer replacement at the higher initiator concentration (fig.1). Consideration must therefore be given to possible effects, other than dilution, that may retard the polymerization rate. The most obvious reason why the rate of polymerization is reduced to below that expected from dilution is an increase in the exit rate coefficient induced by the presence of ethylbenzene (see tables 1 and 4). It should be recognized that high levels of substitution of the monomer by ethylbenzene can result in very large increases in k according to eqn (8). This can be rewritten as Ak/ko = A(CD/CM) (9) where A is a constant. From the relaxation data recorded in table 1 for 30% replacement of styrene by ethylbenzene, A is found to have the value 6.2.Alternatively, the data for the chemically initiated reaction with 15% substitution of styrene by ethylbenzene gives A = 6.5, in good agreement with the relaxation value. Using the former value for A , it is readily predicted from eqn (9) that the exit rate coefficient at 75% replacement level should be ca. 20 times larger than that in the absence of diluent. This is a sufficiently large increase in the value of k to explain the observed decrease in the relative rate of polymerization at high dilutions, irrespective of the initiator concentration. There is, however, an additional mechanism which might be implicated in the reduction in the rate of polymerization at high levels of substitution of monomer by ethylbenzene. This is a significant reduction in the rate of production in the aqueous phase of the oligomeric free-radical species that are capable of entering the latex particles.It would be expected that at very high dilutions of monomer the reduction in the chemical potential of the monomer in the emulsion droplets must be accompanied by a significant decrease in the equilibrium concentration of monomer in the aqueous phase. This could reduce the rate of addition of monomer to the primary free radicals in the aqueous phase to such an extent that the rate of production of oligomeric species is markedly reduced. This decrease in the rate of production of oligomeric free radicals could occur at high degrees of replacement of the monomer by the hydrocarbon, despite the fact that the entry rate coefficient is increased at lower levels of replacement.It is not possible from the data presented in fig. 2 to discriminate between the two possible causes of the reduction in rate at high levels of substitution: a large increase in k and/or a reduction in p. Note, however, that at high replacement levels the kinetic curves display virtually no approach to the steady state. The time to achieve the steady state is of the order of (2p+k)-l. The rapid attainment of the steady state, which precludes the determination of the entry and exit rate coefficients in these cases, suggests that either k and/or p must have been relatively large in these systems. The above results might seem at first sight to be in conflict with those previously reported by Azad et aL2 and discussed in the introduction.It is apparent that since the value of k at low replacement levels is not dramatically increased by the presence of ethylbenzene, it is unlikely that the value of iiss could have been reduced from 0.5 in the absence of ethylbenzene to ca. 0.38 in the presence of ethylbenzene (4 = 0.23), as was claimed by those authors. This apparent discrepancy may be readily resolved, however, by recognizing that the calculation of A from rate data involves an assumption as to the value of the propagation rate constant for styrene, kp. Azad et aL2 assumed that k, = 206 dm3 mol-1 s-l, whereas our previous studies5y l7 have shown that for the emulsion polymerization of styrene at 50 "C the appropriate value2922 SEEDED EMULSION POLYMERIZATION OF STYRENE loo 1 80 60 c 0 .- E $ 40 > 20 0 50 100 150 200 t/min Fig.3. Effect of toluene on the rate of emulsion polymerization of styrene at a high initiator concentration (1.2 x lop2 mol dm-3). Diluent weight fraction 4: curve 1,O.OO; 2,0.15; 3,0.26; 4, 0.50; 5, 0.75. is significantly greater (kp = 255 10 dm3 mol-l s-l). It follows that the correct value for A,, in the absence of diluent in the experiments reported by Azad et al. was ca. 0.40, not 0.50 as claimed. The corresponding revised value of A,, calculated from their rate data in the presence of ethylbenzene at 4 = 0.23 is 0.30. This measured change in A,, can be compared with that predicted from the data presented in table 1 for the increase in k on adding ethylbenzene to this level of replacement, provided it is assumed that the rate coefficient for the entry of free radicals into the particles remains constant.This latter assumption appears reasonable at the relatively high initiator concentration (1.2 x mol dm-3) used in these studies. It would be predicted from the measured increase in k that A,, should change from 0.40 to 0.26, in fair agreement with the revised experimental range (from 0.40 to 0.30). It is clear that the data reported by Azad et al. are in reasonable accord, both qualitatively and quantitatively, with those reported here provided that the value for kp that we previously determined from rates of emulsion polymerization is adopted. TOLUENE The data obtained with toluene as diluent at both high and low initiator concen- trations are presented in fig. 3 and 4. The results for the higher initiator concentration closely parallel those obtained with ethylbenzene as diluent and so will not be discussed further.At the lower initiator concentration, however, a different pattern of behaviour to that obtained with ethylbenzene was observed. At low diluent weight fractions the entry rate coefficient first increased and then decreased (see table 4). This pattern of behaviour is confirmed qualitatively by the sequence of the curves displayed in fig. 4 with increasing degrees of monomer substitution. This observation of a maximum in the entry rate coefficient can be ascribed to the competing effects discussed in relation to ethylbenzene as diluent: on the one hand, the toluene may allow the surfactant to stabilize the precursor particles at a smaller particle size, thus increasing the entry rate coefficient, and on the other, the presence of toluene mayLICHTI, SANGSTER, WHANG, NAPPER AND GILBERT loo 4 80 60 c .- v1 e, 40 8 20 t 2923 0 50 100 150 200 Fig.4. Effects of toluene on the rate of emulsion polymerization of styrene at a low initiator concentration (1.2 x mol dm-3). Diluent weight fraction 4: curve 1,O.OO; 2,0.15; 3,0.26; 4, 0.50; 5, 0.75. t/min decrease the styrene concentration in the aqueous phase, thus reducing the rate of production of oligomeric free-radical species that can enter the latex particles. Futhermore, as styrene is consumed by the polymerization reaction, so the rate of production of free-radical species that can enter the latex particles should decline. BENZENE The effects of benzene on the seeded emulsion polymerization of styrene at a relatively low initiator concentration (1.2 x mol dm-3) are displayed in fig.5. The entry and exit rate coefficients determined at low initial diluent weight fractions (see table 4) show that the exit rate coefficient was not significantly changed by the presence of benzene at these dilution levels. This is in conformity with the results of the relaxation studies presented above. It is apparent from the sequence of the curves shown in fig. 5 that the relative rate of polymerization per particle increased at low levels of substitution of the monomer by benzene. The quantitative data presented in table 4 suggest that this increase resulted from an increase in the entry rate coefficient.CYCLOHEXANE The effects of cyclohexane as a diluent at a low initiator concentration are shown in fig. 6. Again at the lowest initial diluent weight fraction studied the value of the entry rate coefficient was significantly increased (see table 4), but as the level of substitution increased, the rate coefficient again declined. Note, however, that cyclohexane did not exert as large an effect on the free-radical entry rate parameter as did ethylbenzene. This may be a consequence of the fact that cyclohexane at 50 "C is only slightly better than a0-solvent for polystyrene (0 = 34 "C), whereas ethylbenzene is a significantly better solvent for polystyrene.2924 SEEDED EMULSION POLYMERIZATION OF STYRENE - 8o t 50 100 150 200 t/min 0 Fig. 6. Effect of cyclohexane on the rate of emulsion polymerization of styrene at a low initiator concentration (1.2 x mol dmP3).Diluent weight fraction 4: curve 1,O.OO; 2,0.15; 3,0.26; 4, 0.50; 5, 0.75. n-HEPTANE The kinetic curves obtained at a low initiator concentration with n-heptane as a diluent are displayed in fig. 7. The entry and exit rate coefficients determined from these curves at low initial diluent weight fractions are listed in table 4. These results show that for 4 < 0.28 the rate coefficients were not dramatically changed by the replacement of styrene by n-heptane.100 80 h 60 0 .- E 0, 40 > LICHTI, SANGSTER, WHANG, NAPPER AND GILBERT I I I 50 100 150 200 t/min 0 2925 Fig. 7. Effects of n-heptane on the rate of emulsion polymerization of styrene at a low initiator concentration (1.2 x mol dm-3).Diluent weight fraction d;: curve 1,O.OO; 2,O. 15; 3,0.26; 4, 0.50; 5, 0.75. Note that n-heptane is a non-solvent for polystyrene at 50 "C. As a result, it might be expected on free-energy considerations that n-heptane would be located preferentially in the emulsion droplets rather than in the particles. This could result in more of the monomer being present in the droplets at equilibrium than predicted by dilution considerations. As a consequence, the monomer concentration in the particles could well be reduced to below that expected for ideal mixing. Any such depletion of the monomer concentration in the latex particles would necessarily lead to a reduction in the rate of polymerization. CONCLUSIONS The kinetic relaxation studies described above imply that the effect of hydrocarbon diluents on the exit rate coefficient for styrene is determined not only by the chain-transfer constant of the diluent but also by the reactivity of the diluent free radicals with monomer.The latter appears to follow the rule that the more difficult it is to form a free radical, the slower is the reaction of that free radical with monomer. If a diluent free radical reacts only slowly with monomer, there is increased probability of that free radical escaping by diffusion from the latex particle before it adds on monomer and is rendered incapable of undergoing exit. This can result in an increase in the observed exit rate coefficient, even if the constant for chain transfer to the diluent is less than that for chain transfer to monomer.Very large increases in the exit rate coefficient are predicted theoretically to be operative with certain diluents at high monomer replacement levels. The effects of two diluents (ethylbenzene and toluene) at various monomer- replacement levels on the relative rate of polymerization at the higher concentration of potassium peroxydisulphate are summarized in fig. 8. The rate of polymerization was reduced in both cases more than expected on the basis of dilution of monomer alone. This decrease may arise from at least two possible causes: an increase in the2926 SEEDED EMULSION POLYMERIZATION OF STYRENE I 1 c .- 0 0.25 0.50 0 0.25 0.50 0.75 diluent weight fraction Fig. 8. Effect of increasing diluent weight fraction on the relative rate of emulsion polymerization of styrene at high initiator concentration (1.2 x mol dm-3).Diluent: (a) ethylbenzene and (b) toluene. exit rate coefficient and/or a change in the entry rate coefficient. The latter may either increase or decrease in the presence of diluent. The increase, when it occurs, appears to arise from the ability of the diluent to render the surfactant more effective in stabilizing the oligomeric and colloidal free-radical species that can enter the latex particles. Stabilization at lower degrees of aggregation leads to an increase in the rate of entry of free radicals into the latex particles. Any decrease in the entry rate coefficient may be postulated to arise from the reduction in the monomer concentration in the aqueous phase.This depletion necessarily accompanies the reduction in the chemical potential of the styrene in the emulsion droplets on dilution with the hydrocarbon. The results obtained for the relative rates of polymerization of styrene in the steady state in the presence of five diluents at the lower initiator concentration are summarized in fig. 9. It is apparent that a complex pattern of behaviour is observed, which is scarcely surprising in the light of the subtle interplay of the various competing effects operative in these systems. These include the mechanisms which were outlined above that cause an increase in the exit rate coefficient and may either increase or decrease the entry rate coefficient. Ethylbenzene, benzene and cyclohexane, which are all good solvents for polystyrene at 50 "C, produced significant increases in the relative rate of polymerization at low initial diluent weight fractions.These apparently resulted from an increase in the entry rate coefficient, presumably as a consequence of a synergistic interaction between the diluent and the surfactant. This allowed the surfactant to stabilize the entering species at relatively low degrees of aggregation and so reduced the rate of biomolecular termination of the free radicals located outside the latex particles. More free radicals were thus available for entry. n-Heptane, which is a non-solvent for polystyrene, had little effect on the entry or exit rate parameters. At higher hydrocarbon weight fractions, the rate of polymerization was reduced in the presence of all diluents significantly more than that expected from dilution considerations alone. Several factors may contribute to this reduction in rate.First, at high dilutions some hydrocarbons may increase the exit rate coefficient massively. Secondly depletion of the monomer concentration in the aqueous phase may reduceC 0 m .- Y .g 1.0 9) - E 0 a rr QJ 0 5 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 diluent weight fraction Fig. 9. Effect of increasing diluent weight fraction on the relative rate of emulsion polymerization of styrene at low initiator concentration (1.2 x mol dm-"). Diluent: (a) ethylbenzene, (6) toluene, (c) benzene, ( d ) cyclohexane and (e) n-heptane. E Q2928 SEEDED EMULSION POLYMERIZATION OF STYRENE the rate of production of species capable of entering the latex particles and thus lower the rate coefficient for entry of the free radicals into the particles. Thirdly, the concentration of monomer in the latex particles may be reduced below that expected for ideal mixing. We thank the Australian Research Grants Committee (B. C. Y . W.) and AINSE (G.L.) for financial support. The Electron Microscope Unit of the University of Sydney is also thanked for their generous provision of facilities. D. R. Owen, D. McLemore, Liu Wan-Li, R. B. Seymour and W. N. Tinnerman, in ACS Symp. Ser. no. 24, Emulsion Polymerization, ed. I. Piirma and J. L. Gardon (American Chemical Society, Washington, D.C., 1976), p. 299. D. C. Blackley and A. C. Haynes, Br. Polym. J., 1977, 9, 312. A. R. M. Azad, R. M. Fitch and M. Nomura, in ACS Symp. Ser. no. 165, Emulsion Polymers and Emulsion Polymerization, ed. D. R. Bassett and A. E. Hamielec (American Chemical Society, Wash- ington, D.C., 1981), p. 357. B. S. Hawkett, D. H. Napper and R. G. Gilbert, J. Chem. SOC., Faraday Trans. I , 1975, 71, 2288. B. S. Hawkett, D. H. Napper and R. G. Gilbert, J. Chem. SOC., Faraday Trans. I , 1980,76, 1323. B. C. Y. Whang, D. H. Napper, M. J. Ballard and R. G. Gilbert, J. Chem. SOC., Faraday Trans. I , 1982, 78, 11 17. S. W. Lansdowne, R. G. Gilbert, D. H. Napper and D. F. Sangster, J. Chem. SOC., Faraday Trans. I , 1980, 76, 1344. G. Lichti, D. F. Sangster, B. C. Y. Whang, D. H. Napper and R. G. Gilbert, J. Chem. SOC., Faraday Trans. I , 1982, 78, 2129. G. Lichti, B. S. Hawkett, R. G. Gilbert and D. H. Napper, J. Polym. Sci., Polvm. Chem. Ed., 1981, 19, 925. lo Polymer Handbook, ed. J. Brandrup and E. H. Immergut (Wiley, New York, 2nd edn, 1975). l1 J. Ugelstad and J. K. Hansen, Rubber Chem. Technol., 1976,49, 536. l2 M. Nomura and M. Harada, J. Appl. Polym. Sci., 1981, 26, 17. l3 M. Nomura, in Emulsion Polymerization, ed. I. Piirma (Academic Press, New York, 1982), chap. 5. l4 M. H. George, in Vinyl Polymerization, ed. G. E. Ham (Marcel Dekker, New York, 1969), vol. 1, l5 P. J. Flory, Principles of Polymer Chemistry (Cornell University Press, Ithaca, New York, 1953), l6 G. Lichti, R. G. Gilbert and D. H. Napper, J. Polym. Sci., Polym. Chem. Ed., 1983, 21, 269. l7 B. S. Hawkett, D. H. Napper and R. G. Gilbert, J. Chem. SOC., Faraday Trans. I , 1981,77, 2395. 18 I. Piirma and S-R. Chen, J. Colloid Interface Sci., 1980, 74, 90. chap. 3. pp. 189-195. (PAPER 3/2078)