RAFT polymerization kinetics: How long are the cross‐terminating oligomers?

We extend a new model for the kinetics of reversible addition‐fragmentation chain transfer (RAFT) polymerization. The essence of this model is that the termination of the radical intermediate formed by the RAFT process occurs only with very short oligomeric radicals. In this work, we consider cross‐termination of oligomers up to two monomers and an initiator fragment. This model accounts for the absence of three‐armed stars in the molecular weight distribution, which are predicted by other cross‐termination models, since the short third arm makes a negligible difference to the polymer's molecular weight. The model is tested against experiments on styrene mediated by cyano‐isopropyl dithiobenzoate, and ESR experiments of the intermediate radical concentration. By comparing our model to experiments, we may determine the significance of cross‐termination in RAFT kinetics. Our model suggests that to agree with the known data on RAFT kinetics, the majority of cross‐terminating chains are dimeric or shorter. If longer chains are considered in cross‐termination reactions, then significant discrepancies with the experiments (distinguishable star polymers in the molecular weight distribution) and quantum calculations will result. © 2009 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 47: 3455–3466, 2009     

A difference of six orders of magnitude: A reply to “the magnitude of the fragmentation rate coefficient”

There has been an ongoing debate regarding the mechanism that causes rate retardation phenomena observed in some reversible addition‐fragmentation transfer (RAFT) polymerization systems. Some attribute the retardation to slow fragmentation of adduct radicals, others attribute it to fast fragmentation coupled with cross‐termination between propagating and adduct radicals. There exists a difference of six orders of magnitude (10^?2 versus 10⁴/s) in the reported values of the fragmentation rate constant (k_f0) for virtually similar RAFT systems of PSt? S? C · (Ph)? S? PSt. In this communication, we explain the estimates of k_f ～ 10⁴/s and the choices of the rate constant in modeling based on experimental polymerization rate and radical concentration data. The use of k_f ～ 10^?2/s in the model results in a calculated adduct radical concentration level of 10^?4 to 10^?3 mol/L, which appears to directly contradict the reported electron spin resonance (ESR) data in the range of <10^?6 mol/L. We hope that this open discussion can stimulate more effort to resolve this outstanding difference. © 2003 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 2833–2839, 2003     

Substituted propenyl end groups as reactive intermediates in radical polymerization

The addition of propagating radicals of methyl acrylate (MA) and styrene (St) to CH₂?C(CO₂CH₃)CH₂? and CH₂?C(C₆H₅)CH₂? ω‐end groups of poly(methyl methacrylate) (PMMA) and polystyrene (PSt) was investigated. The end groups were as reactive as MA and St toward the poly(methyl acrylate) (PMA) and PSt radicals, respectively. The adduct radical derived from the two types of PMMA end groups and PMA radicals underwent β fragmentation exclusively to yield PMMA radicals and end groups bound to PMA chains. The addition of PSt radicals to PMMA with CH₂?C(CO₂Me)CH₂? end groups resulted in adduct radicals that underwent β fragmentation and addition to St or coupling with PSt radicals. Adduct radicals formed by the addition of PMA radicals to both types of end groups of PSt exclusively formed C? C bond by coupling with PMA radicals to form branched structures or by addition to MA monomer to give a copolymer. The fate of the adduct radicals was highly dependent on the type of polymer chain and the substituent bound to the end group. Steric congestion of the adduct radical arising from the α‐methyl group of the PMMA chain was considered to be crucial for fragmentation to expel the PMMA radical. © 2003 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 645–654, 2003     

Studying the mechanism of thioketone‐mediated polymerization via electrospray ionization mass spectrometry

The polymeric product spectrum generated in thioketone‐mediated free radical polymerization (TKMP) was analyzed via electrospray ionization mass spectrometry. Poly(n‐butyl acrylate) samples were synthesized in the presence of the (commercially available) thioketone 4,4‐bis(dimethylamino)thiobenzophenone under variable reaction conditions in toluene solution at 80 °C. To unambiguously assign the mass spectra, the samples are prepared under variation of the monomer (going from n‐butyl acrylate to ethyl acrylate) as well as by employing variable thermally decomposing initiators [i.e., 2,2′‐azoisobutyronitrile and azobis(cyclohexanecarbonitrile)]. In all mass spectra, significant amounts of the expected cross‐termination product, formed via bimolecular termination of propagating macroradicals with the dormant thioketone radical adduct (consisting of a propagating chain and the mediating thioketone) alongside conventional termination products can be identified. As the study was carried out on acrylate polymers, acrylate‐specific reaction products arising from intramolecular transfer reactions followed by β‐scission of the generated mid‐chain radicals are also identified in the mass spectra. In addition, a species congruent with the dormant thioketone radical adduct itself (oxidized to its cationic state) was identified. Products that could potentially be formed via a chain transfer mechanism cannot be identified. The results presented here thus support the earlier suggested TKMP mechanism involving a highly stabilized adduct radical which undergoes significant cross‐termination reactions. © 2009 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 47: 1864–1876, 2009     

Thiocarbonylthio end group removal from RAFT‐synthesized polymers by a radical‐induced process

The removal of thiocarbonylthio end groups by radical‐addition‐fragmentation‐ coupling from polymers synthesized by RAFT polymerization has been studied. We found that a method, which involves heating the polymer with a large excess (20 molar equivalents) of azobis(isobutyronitrile) (AIBN), while successful with methacrylic polymers, is less effective with styrenic or acrylic polymers and provides only partial end group removal. This is attributed to the propagating radicals generated from the latter polymers being poor radical leaving groups relative to the cyanoisopropyl radical. Similar use of lauroyl peroxide (LPO) completely removes the thiocarbonylthio groups from styrenic or acrylic polymers but, even with LPO in large excess, produces a polymer with a bimodal molecular weight distribution. The formation of a peak of double molecular weight is indicative of the occurrence of self‐termination and ineffective radical trapping. We now report that by use of a combination of LPO (2 molar equivalents) and AIBN (20 molar equivalents) we are able to completely remove thiocarbonylthio end groups of styrenic or acrylic polymers and minimize the occurrence of self termination. © 2009 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 47: 6704–6714, 2009     

Living/controlled copolymerization of acrylates with nonactivated alkenes

The living/controlled copolymerization of methyl acrylate with 1‐alkenes and norbornene derivatives through several radical polymerization techniques has been achieved. These techniques include atom transfer radical polymerization, reversible addition–fragmentation transfer polymerization, nitroxide‐mediated polymerization, and degenerative transfer polymerization. These systems display many of the characteristics of a living polymerization process: the molecular weight increases linearly with the overall conversion, but the polydispersity remains low. Novel block copolymers have been synthesized through the sequential addition of monomers or chain extension. © 2004 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 42: 6175–6192, 2004     

Living free‐radical polymerization (reversible addition–fragmentation chain transfer) of 6‐[4‐(4′‐methoxyphenyl)phenoxy]hexyl methacrylate: A route to architectural control of side‐chain liquid‐crystalline polymers

Side‐chain liquid‐crystalline polymers of 6‐[4‐(4′‐methoxyphenyl)phenoxy]hexyl methacrylate with controlled molecular weights and narrow polydispersities were prepared via reversible addition–fragmentation chain transfer (RAFT) polymerization with 2‐(2‐cyanopropyl) dithiobenzoate as the RAFT agent. Differential scanning calorimetry studies showed that the polymers produced via the RAFT process had a narrower thermal stability range of the liquid‐crystalline mesophase than the polymers formed via conventional free‐radical polymerization. In addition, a chain length dependence of this stability range was found. The generated RAFT polymers displayed optical textures similar to those of polymers produced via conventional free‐radical polymerization. © 2003 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 2949–2963, 2003     

Living radical polymerization of diisopropyl fumarate to obtain block copolymers containing rigid poly(substituted methylene) and flexible polyacrylate segments

Living radical polymerizations of diisopropyl fumarate (DiPF) are carried out to synthesize poly(diisopropyl fumarate) (PDiPF) as a rigid poly(substituted methylene) and its block copolymers combined with a flexible polyacrylate segment. Reversible addition‐fragmentation chain transfer (RAFT) polymerization is suitable to obtain a high‐molecular‐weight PDiPF with well‐controlled molecular weight, molecular weight distribution, and chain‐end structures, while organotellurium‐mediated living radical polymerization (TERP) and reversible chain transfer catalyzed polymerization (RTCP) give PDiPF with controlled chain structures under limited polymerization conditions. In contrast, controlled polymerization for the production of high‐molecular‐weight and well‐defined PDiPF is not achieved by atom transfer radical polymerization (ATRP) and nitroxide‐mediated radical polymerization (NMP). The block copolymers consisting of rigid poly(substituted methylene) and flexible polyacrylate segments are synthesized by the RAFT polymerization. © 2016 Wiley Periodicals, Inc. J. Polym. Sci., Part A: Polym. Chem. 2016 , 54, 2136–2147     

Genesis of the CSIRO polymer group and the discovery and significance of nitroxide‐mediated living radical polymerization

The background to the formation of the Commonwealth Scientific and Industrial Research Organization (CSIRO) polymer group is discussed. In particular, the challenges of working with high‐conversion polymerization, as found in commercial systems, and the need to explain variations in polymer properties led to important advances in the theory of radical polymerization and control over both the initiation and termination steps. Studies on the fate of the macromonomer, formed in termination by disproportionation, led to an early form of addition/fragmentation now known as reversible addition–fragmentation chain transfer, whereas detailed studies on initiation pathways using nitroxide trapping led to nitroxide‐mediated living radical polymerization. These studies contributed to the renaissance in free‐radical polymerization studies. © 2005 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 43: 5748–5764, 2005     

Design strategies for controlling the molecular weight and rate using reversible addition–fragmentation chain transfer mediated living radical polymerization

Living radical polymerization has allowed complex polymer architectures to be synthesized in bulk, solution, and water. The most versatile of these techniques is reversible addition–fragmentation chain transfer (RAFT), which allows a wide range of functional and nonfunctional polymers to be made with predictable molecular weight distributions (MWDs), ranging from very narrow to quite broad. The great complexity of the RAFT mechanism and how the kinetic parameters affect the rate of polymerization and MWD are not obvious. Therefore, the aim of this article is to provide useful insights into the important kinetic parameters that control the rate of polymerization and the evolution of the MWD with conversion. We discuss how a change in the chain‐transfer constant can affect the evolution of the MWD. It is shown how we can, in principle, use only one RAFT agent to obtain a polymer with any MWD. Retardation and inhibition are discussed in terms of (1) the leaving R group reactivity and (2) the intermediate radical termination model versus the slow fragmentation model. © 2005 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 43: 3189–3204, 2005     

Reversible addition‐fragmentation chain transfer polymerization of diisopropyl fumarate using various dithiobenzoates as chain transfer agents

Dialkyl fumarates as 1,2‐disubstituted ethylenes exhibit unique features of radical polymerization kinetics due to their significant steric hindrance in propagation and termination processes and provide polymers with a rigid chain structure different from conventional vinyl polymers. In this study, we carried out reversible addition‐fragmentation chain transfer polymerization of diisopropyl fumarate (DiPF) in bulk at 80 °C using various dithiobenzoates with different leaving R groups as chain transfer agents to reveal their performance for control of molecular weight, molecular weight distribution, and chain end functionality of the resulting poly(DiPF) (PDiPF). 2‐(Ethoxycarbonyl)‐2‐propyl dithiobenzoate ( DB1 ) and 2,4,4‐trimethyl‐2‐pentyl dithiobenzoate ( DB2 ) underwent fragmentation and reinitiation at a moderate rate and consequently led to the formation of PDiPF with well‐controlled chain structures. It was confirmed that molecular weight of PDiPF produced by controlled polymerization with DB1 and DB2 agreed with theoretical one and molecular weight distribution was narrow. Dithiobenzoate and R fragments were introduced into the polymer chain ends with high functionality as 95% by the use of DB1 . In contrast, polymerizations using 1‐(ethoxycarbonyl)benzyl dithiobenzoate ( DB3 ), 1‐phenylethyl dithiobenzoate ( DB4 ), and 2‐phenyl‐2‐propyl dithiobenzoate ( DB5 ) resulted in poor control of molecular weight, molecular weight distribution, and chain end structures of PDiPF. Fragmentation and reinitiation rates of the used benzoates as chain transfer agents significantly varied depending on the R structures in an opposite fashion; that is, introduction of bulky and conjugating substituents accelerated fragmentation, but it retarded initiation of DiPF polymerization. It was revealed that balance of fragmentation and reinitiation was important for controlled polymerization of DiPF. © 2017 Wiley Periodicals, Inc. J. Polym. Sci., Part A: Polym. Chem. 2017 , 55, 3266–3275     

Reversible addition–fragmentation chain‐transfer graft polymerization of styrene: Solid phases for organic and peptide synthesis

The γ‐initiated reversible addition–fragmentation chain‐transfer (RAFT)‐agent‐mediated free‐radical graft polymerization of styrene onto a polypropylene solid phase has been performed with cumyl phenyldithioacetate (CPDA). The initial CPDA concentrations range between 1 × 10^?2 and 2 × 10^?3 mol L^?1 with dose rates of 0.18, 0.08, 0.07, 0.05, and 0.03 kGy h^?1. The RAFT graft polymerization is compared with the conventional free‐radical graft polymerization of styrene onto polypropylene. Both processes show two distinct regimes of grafting: (1) the grafting layer regime, in which the surface is not yet totally covered with polymer chains, and (2) a regime in which a second polymer layer is formed. Here, we hypothesize that the surface is totally covered with polymer chains and that new polymer chains are started by polystyrene radicals from already grafted chains. The grafting ratio of the RAFT‐agent‐mediated process is controlled via the initial CPDA concentration. The molecular weight of the polystyrene from the solution (PS_free) shows a linear behavior with conversion and has a low polydispersity index. Furthermore, the loading of the grafted solid phase shows a linear relationship with the molecular weight of PS_free for both regimes. Regime 2 has a higher loading capacity per molecular weight than regime 1. © 2002 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 40: 4180–4192, 2002     

Macromolecular brushes synthesized by “grafting from” approach based on “click chemistry” and RAFT polymerization

Well‐defined macromolecular brushes with poly(N‐isopropyl acrylamide) (PNIPAM) side chains on random copolymer backbones were synthesized by “grafting from” approach based on click chemistry and reversible addition‐fragmentation chain transfer (RAFT) polymerization. To prepare macromolecular brushes, two linear random copolymers of 2‐(trimethylsilyloxy)ethyl methacrylate (HEMA‐TMS) and methyl methacrylate (MMA) (poly(MMA‐co‐HEMA‐TMS)) were synthesized by atom transfer radical polymerization and were subsequently derivated to azide‐containing polymers. Novel alkyne‐terminated RAFT chain transfer agent (CTA) was grafted to polymer backbones by copper‐catalyzed 1,3‐dipolar cycloaddition (azide‐alkyne click chemistry), and macro‐RAFT CTAs were obtained. PNIPAM side chains were prepared by RAFT polymerization. The macromolecular brushes have well‐defined structures, controlled molecular weights, and molecular weight distributions (M_w/M_n ≦ 1.23). The RAFT polymerization of NIPAM exhibited pseudo‐first‐order kinetics and a linear molecular weight dependence on monomer conversion, and no detectable termination was observed in the polymerization. The macromolecular brushes can self‐assemble into micelles in aqueous solution. © 2009 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 48: 443–453, 2010     

Macromonomer synthesis using α‐(2‐methyl‐2‐phenylpropyl)acrylates as addition–fragmentation chain‐transfer agents expelling the cumyl radical

α‐(2‐Methyl‐2‐phenylpropyl)acrylate (RS‐2) was examined as a C? C bond‐cleavage type addition–fragmentation chain transfer (AFCT) agent in the benzene solution polymerizations of styrene (St), ethyl methacrylate (EMA), and cyclohexyl acrylate (CHA) with the objective of achieving efficient macromonomer synthesis by radical polymerization. The AFCT efficiency was evaluated in terms of the decrease in the number‐average molecular weight (M_n) upon the addition of the AFCT agent and the number of unsaturated end groups introduced per chain (f). The AFCT efficiency was rationalized by the consideration of the relative importance of AFCT as an end‐forming event and the competition between ‐fragmentation and crosspropagation as adduct radical reaction pathways. In St and EMA polymerizations at 60 °C, RS‐2 resulted in higher f values and lower M_n values than methyl α‐(2‐methyl‐2‐carbomethoxypropyl)acrylate (MMA‐2), and this suggested the facilitation of ‐fragmentation due to the expulsion of the more stable cumyl radical from the RS‐2 adduct radical. Higher f values were observed for MMA‐2 than for RS‐2 in CHA polymerization because of unsaturated end group formation by ‐fragmentation of midchain radicals. However, RS‐2 resulted in lower M_n values for poly(CHA) than MMA‐2 because of a smaller contribution of crosspropagation. Retardation in the presence of the AFCT agents was affected by the balance between b‐fragmentation and crosspropagation and by the addition rate of the propagating radical to the AFCT agent. © 2004 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 42: 6021–6030, 2004     

Calculations of Monomer Conversion and Radical Concentration in Reversible Addition‐Fragmentation Chain Transfer Radical Polymerization

Simple expressions are derived for the development of monomer conversion, as well as propagating radical, adduct radical, dormant chain, and dead chain concentrations in reverse addition‐fragmentation transfer polymerization (RAFT). The relations for the profiles of propagating radical concentration and conversion versus time are derived and depend on group parameters of rate constants and chemical recipe. The analytical equations are verified against numerical solutions of the mass‐balance differential equations. This derivation involves the steady‐state hypothesis for radical and RAFT agent concentrations. The errors introduced by these assumptions are negligible when the fragmentation rate constant, k_f, is higher than 10 s⁻¹ or when the cross‐termination rate constant, k_ct, is higher than 10⁵ L · mol⁻¹ s⁻¹.

Calculated concentration profiles (points: numerical, lines: analytical) of propagating radical R, adduct radical A, dormant T, and dead D (= P + P′) chains.     

Kinetics of surface‐initiated atom transfer radical polymerization

Although atom transfer radical polymerization (ATRP) is often a controlled/living process, the growth rate of polymer films during surface‐initiated ATRP frequently decreases with time. This article investigates the mechanism behind the termination of film growth. Studies of methyl methacrylate and methyl acrylate polymerization with a Cu/tris[2‐(dimethylamino)ethyl]amine catalyst system show a constant but slow growth rate at low catalyst concentrations and rapid growth followed by early termination at higher catalyst concentrations. For a given polymerization time, there is, therefore, an optimum intermediate catalyst concentration for achieving maximum film thickness. Simulations of polymerization that consider activation, deactivation, and termination show trends similar to those of the experimental data, and the addition of Cu(II) to polymerization solutions results in a more constant rate of film growth by decreasing the concentration of radicals on the surface. Taken together, these studies suggest that at high concentrations of radicals, termination of polymerization by radical recombination limits film growth. Interestingly, stirring of polymerization solutions decreases film thickness in some cases, presumably because chain motion facilitates radical recombination. © 2002 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 386–394, 2003     

Bimolecular radical termination: New perspectives and insights

The reversible addition‐fragmentation chain transfer‐chain length dependent termination (RAFT‐CLD‐T) method has allowed us to answer a number of fundamental questions regarding the mechanism of diffusion‐controlled bimolecular termination in free‐radical polymerization (FRP). We carried out RAFT‐mediated polymerizations of methyl acrylate (MA) in the presence of a star matrix to develop an understanding of the effect of polymer matrix architecture on the termination of linear polyMA radicals and compared this to polystyrene, polymethyl methacrylate, and polyvinyl acetate systems. It was found that the matrix architecture had little or no influence on termination in the dilute regime. However, due to the smaller hydrodynamic volumes of the stars in solution compared to linear polymer of the same molecular weight, the gel onset point occurred at greater conversions, and supported the postulate that chain overlap (or c*) is the main cause for the observed autoacceleration observed in FRP. Other theories based on “short–long” termination or free‐volume should be disregarded. Additionally, since our systems are well below the entanglement molecular weight, entanglements should also be disregarded as the cause of the gel onset. The semidilute regime occurs over a small conversion range and is difficult to quantify. However, we obtain accurate dependencies for termination in the concentrated regime, and observed that the star polymers (through the tethering of the arms) provided constriction points in the matrix that significantly slow the diffusion of linear polymeric radicals. Although, this could at first sight be postulated to be due to reptation, the dependencies showed that reptation could be considered only at very high conversions (close to the glass transition regime). In general, we find from our data that the polymer matrix is much more mobile than what is expected if reptation were to dominate. © 2008 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 46: 3155–3173, 2008     

Gel effect in the bulk reversible addition–fragmentation chain transfer polymerization of methyl methacrylate: Modeling and experiments

A model for the evaluation of the kinetics and the chain length distribution in living/controlled radical polymerization mediated by reversible addition–fragmentation chain transfer (RAFT) in bulk is presented. Using the free volume theory, the model accounts for the diffusion limitations over both termination and RAFT exchange reactions. Model predictions are compared to experimental results of methyl methacrylate polymerization with cumyl dithiobenzoate as a RAFT agent. It is shown that the polymerization retardation observed in living systems at large conversions is well predicted. © 2005 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 44: 1071–1085, 2006     

2‐Cyanoprop‐2‐yl dithiobenzoate mediated reversible addition–fragmentation chain transfer polymerization of acrylonitrile targeting a polymer with a higher molecular weight

The reversible addition–fragmentation chain transfer (RAFT) polymerization of acrylonitrile (AN) mediated by 2‐cyanoprop‐2‐yl dithiobenzoate was first applied to synthesize polyacrylonitrile (PAN) with a high molecular weight up to 32,800 and a polydispersity index as low as 1.29. The key to success was ascribed to the optimization of the experimental conditions to increase the fragmentation reaction efficiency of the intermediate radical. In accordance with the atom transfer radical polymerization of AN, ethylene carbonate was also a better solvent candidate for providing higher controlled/living RAFT polymerization behaviors than dimethylformamide and dimethyl sulfoxide. The various experimental parameters, including the temperature, the molar ratio of dithiobenzoate to the initiator, the molar ratio of the monomer to dithiobenzoate, the monomer concentration, and the addition of the comonomer, were varied to improve the control of the molecular weight and polydispersity index. The molecular weights of PANs were validated by gel permeation chromatography along with a universal calibration procedure and intrinsic viscosity measurements. ¹H NMR analysis confirmed the high chain‐end functionality of the resultant polymers. © 2007 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 45: 1272–1281, 2007     

Addition–fragmentation chain transfer: Molecular weight distributions and retardation in the system methyl methacrylate/methyl α‐(bromomethyl)acrylate

A detailed investigation of addition–fragmentation chain transfer (AFCT) in the free‐radical polymerization of methyl methacrylate (MMA) in the presence of methyl α‐(bromomethyl)acrylate (MBMA) was carried out to elucidate mechanistic details with efficient macromonomer synthesis as an underlying goal. Advanced modeling techniques were used in connection with the experimental work. Curve fitting of simulated and experimental molecular weight distributions with respect to the rate coefficient for addition of propagating radicals to MBMA (k_add) over 60–120 °C resulted in E_add = 21.7 kJ mol^?1 and A_add = 2.18 × 10⁶ M^?1 s^?1 and a very weak temperature dependence of the chain‐transfer constant (E_add ≈ E_p). The rate coefficient for fragmentation of adduct radicals at 60 °C was estimated as k_f ≈ 39 s^?1 on the basis of experimental data of the MMA conversion and the concentration of 2‐carbomethoxy‐2‐propenyl end groups. The approach developed is generic and can be applied to any AFCT system in which copolymerization does not occur and in which the resulting unsaturated end groups do not undergo further reactions. © 2004 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 42: 2640–2650, 2004