A previously published simulation and data fitting procedure for the reversible addition fragmentation chain transfer (RAFT) process using the PREDICI simulation program has been extended to cumyl phenyldithioacetate mediated styrene and methyl methacrylate (MMA) bulk homopolymerizations. The experimentally obtained molecular weight distributions (MWDs) for the styrene system are narrow and unimodal and shift linearly with monomer conversion to higher molecular weights. The MMA system displays a hybrid of conventional chain transfer and living behavior, leading to bimodal MWDs. The styrene system has been subjected to a combined experimental and modeling study at 60 °C, yielding a rate coefficient for the addition reaction of free macroradicals to polymeric RAFT agent, k β, of approximately 5.6 × 105 L mol-1 s-1 and a decomposition rate coefficient for macroradical RAFT species, k - β, of about 2.7 × 10-1 s-1. The transfer rate coefficient to cumyl phenyldithioacetate is found to be close to 2.2 × 105 L mol-1 s-1. The MMA system has been studied over the temperature range 25−60 °C. The hybrid behavior observed in the MMA polymerizations has been exploited (at low monomer conversions) to perform a Mayo analysis allowing the determination of the temperature dependence of the transfer to cumyl phenyldithioacetate reaction. The activation energy of this process is close to 26 kJ mol-1. In contrast to the styrene system, the PREDICI simulation procedure cannot be successfully applied to cumyl phenyldithioacetate mediated MMA polymerizations for the deduction of k β and k - β. This inability is due to the hybrid nature of the cumyl phenyldithioacetate−MMA system, leading to a significantly reduced sensitivity toward k β and k - β.

This highlight describes recent developments in reversible addition-fragmentation transfer (RAFT) polymerization. Succinct coverage of the RAFT mechanism is supplemented by details of synthetic methodologies for making a wide range of architectures ranging from stars to combs, microgels, and blocks. In addition, RAFT reactions in different media such as emulsion and ionic liquids receive attention. Finally, a specific example of a novel material design is briefly introduced, whereas polymers prepared via RAFT are adopted for microporous/honeycomb membrane design.

A method is derived for calculating Arrhenius parameters for propagation reactions in freeradical polymerizations from first principles. Ah initio molecular orbital calculations are carried out initially to determine the geometries, vibrational frequencies, and energies of the reactants and the transition state. Transition state theory then yields the Arrhenius parameters. The lowest frequencies are replaced by appropriate (hindered or unhindered) internal rotors, to better model these modes in the calculation of frequency factors. It is found that a high level of molecular orbital theory (e.g., QCISD-(TV6-311G**) is required to produce reasonable activation energies, whereas satisfactory frequency factors can be obtained at a relatively simple level of theory (e.g., HF/3-21G), because the frequency factor is largely determined by molecular geometries which can be reliably predicted at such a level. Obtaining reliable frequency factors for quite large systems is thus possible. The overall procedure is illustrated by calculations on the propagation of ethylene, and the results are in accord with literature experimental data. Means are also derived for extending the results from propagation of monomeric radicals to propagation of polymeric radicals, without additional computational requirements. The method is expected to be generally applicable to those propagation reactions that are not significantly influenced by the presence of solvent (i.e., relatively nonpolar monomers in nonpolar solvents). The calculations show that the magnitude of the frequency factor is largely governed by the degree to which the internal rotations of the transition state are hindered. They also suggest that there can be a significant penultimate unit effect in free-radical copolymerization. Furthermore, the calculations explain the rate-enhancing effect found upon deuteration of the monomers and explain why the rate coefficient for the first propagation step is larger than that for the long-chain propagation step.

Literature data are summarized for the chain‐length‐dependence of the termination rate coefficient in dilute solution free‐radical polymerizations. In essence such experiments have yielded two parameter values: the rate coefficient for termination between monomeric free radicals, k, and a power‐law exponent e quantifying how kt values decrease with increasing chain length. All indications are that the value e ≈ 0.16 in good solvent is accurate, however the values of k which have been deduced are considerably lower than well‐established values for small molecule radicals. This seeming impasse is resolved by putting forward a ‘composite’ model of termination: it is proposed that the value e ≈ 0.16 holds only for long chains, with e being higher for small chains – the value 0.5 is used in this paper, although it is not held to dogmatically. It is then investigated whether this model is consistent with experimental data. This is a non‐trivial task, because although the experiments themselves and the ways in which they are analyzed are elegant and not too complicated, the underlying theory is sophisticated, as is outlined. Simulations of steady‐state polymerization experiments are first of all carried out, and it is shown that the composite model of termination both recovers the e values which have been found and beautifully explains why these experiments considerably underestimate the true value of k. Simulations of pulsed‐laser polymerizations find the same, although not quite so strikingly. It is therefore concluded that our new termination model, which retains the virtue of simplicity and in which all parameter values are physically reasonable, is consistent with experimental data. Taking a wider view, it seems likely that the situation of the exponent e varying with chain length will not just be the case in dilute solution, but will be the norm for all conditions, which would give our model and our work a general relevance.Normalized chain length distributions from PLP simulations.imageNormalized chain length distributions from PLP simulations.