Temperature dependence of viscosity–shear rate behavior in undiluted polystyrene

The time—temperature superposition principle is well-established for linear viscoelastic properties of polymer systems. It is generally supposed that the same principle carries over into nonlinear phenomena, such as the relationship between viscosity η and shear rate \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document}. Guided by this principle and the forms of various molecular theories, one would expect that η—\documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document} data on the same polymer at different temperatures would superimpose when plotted as η/η0 versus \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document}η0/ρT, η0 being the limiting viscosity at low shear rates, ρ the polymer density, and T the absolute temperature. Data on polystyrene melts, obtained in a plate-cone viscometer, appear systematically to violate this principle in the range 140–190°. Such anomalies are absent in concentrated solutions of polystyrene. The trends are similar to those reported by Plazek in the steady-state compliance of polystyrene melts near T_g, but they appear to persist to higher temperatures than the compliance anomaly.     

Temperature and shear dependencies of rheology of poly(acrylonitrile‐co‐itaconic acid) dope in DMF

Poly(acrylonitrile‐co‐itaconic acid) (poly(AN‐co‐IA)) precursor required for carbon fiber production is made into a dope and spun into fibers using a suitable spinning technique. The viscosity of the resin dope is decided by the polymer concentration, polymer molecular weight, temperature, and shear force. The shear rheology of concentrated poly(AN‐co‐IA) polymer solutions in N,N‐dimethylformamide (DMF), in the range of 1 × 10⁵–1 × 10⁶ g mol^?1, has been investigated in the shear rate (γ′) range of 1 × 10¹–5 × 10⁴ min^?1. The zero shear viscosity (η₀) has been evaluated at different temperatures. The temperature dependence of zero shear viscosity conformed to the Arrhenius–Frenkel–Eyring model. The free energy of activation of viscous flow (ΔG_V) values were in the range 5–32 kJ mol^?1 and this value increased with increase in polymer concentration and molecular weight. A master equation for the ΔG_V value of the polymer solution of any and concentration (c) is suggested. The power law fitted well for the shear dependency of viscosity of these polymer solutions. The pseudoplasticity index (n) diminished with increase in polymer concentration and molecular weight. An empirical relation between viscosity (η) and was found to exist at constant shear rate, concentration and temperature. For each , the equation relating n, c, and T was established. Copyright © 2008 John Wiley & Sons, Ltd.     

Characterization and Miscibility Dynamics of Dextran-Poly(vinylpyrrolidone)-Water System

Summary: The miscibility behavior and intermolecular interactions among Dextran (Dx) with different molecular weight and Polyvinylpyrrolidone (PVP) blends were studied as dilute aqueous solutions at 25 °C by viscosity method. The intrinsic viscosity and the interaction coefficient were experimentally measured for each polymer-water as well as for Dx-PVP-water systems. These results served for the prediction of miscibility of the Dx/PVP blends with various blend compositions by using , , , , and parameters. Except Dx4/PVP with its all compositions (Dx4 with nominal molecular weight of 110 000), other blend systems are found to be almost miscible. The density measurements of these polymer solutions and their blends were conducted in order to compare with the viscosity findings. Lastly, all Dx with different molecular weight, PVP and their blends were characterized by infrared spectroscopy (FT-IR), and differential scanning calorimetry (DSC).     

Rheology of poly(vinyl chloride) melts. II. Shear rate-dependent properties

Four samples containing 40, 60, 80, and 97 wt-% of poly(vinyl chloride), the rest being plasticizer and stabilizer, were tested by using the Weissenberg Rheogoniometer in the steady-shearing mode at temperatures between 155 and 235°C and rates of shear \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma = 0.01 - 400 $\end{document} sec^?1. The viscosity η versus \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document} follows Graessley's theoretical dependence for infinitely entangled system. The primary normal-stress difference coefficient ψ versus \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document} is well described by the same theoretical function, used with the square of its argument. The temperature dependence of η₀ and ψ₀ shows discontinuities at T = T_b. The numerical values of T_b can be calculated from the theory of the melting point depression due to diluent. The activation energy of viscous flow E_η below T_b is 5–9 times as large as above this temperature. The activation energy of normal stress is found to be Eψ ≈ 5E_η. The characteristic relaxation times τ_o, ψ_p, calculated from superposition of η versus \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document} and ψ versus \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \gamma $\end{document} data, respectively, onto Graessley's master curves, and τ_N, computed from zero shear parameters η₀ and ψ₀, differ in their sensitivity to the melting of microcrystalline regions. It is postulated that in the systems investigated, aggregates with long lifetimes are being formed, increasing the effective molecular weight and introducing changes in the effective polydispersity.     

Shear thinning and polymer deformation in large flow fields

We theoretically investigate polymer deformation and shear thinning, i.e., a decrease of intrinsic viscosity, in a dilute polymer solution as a function of the applied shear rate $ \dot \gamma $. We use a bead-and-spring model with hydrodynamic interaction in the Rouse-Zimm framework, approximately accounting also for excluded-volume effects, and impose a constraint on the average mean-square spring length to prevent its stretching at large $ \dot \gamma $. When suitably normalized, both the intrinsic viscosity [η] and the components of the mean gyration tensor 〈SS〉 depend on the single variable $ \xi = {{\dot \gamma \tau _1^{\left( 0 \right)} } \mathord{\left/ {\vphantom {{\dot \gamma \tau _1^{\left( 0 \right)} } {N^{1 - v} }}} \right. \kern-\nulldelimiterspace} {N^{1 - v} }} $ where τ is the longest relaxation time for $ \dot \gamma = 0 $, N is the number of chain springs and v is the Flory exponent. The full shear-rate dependence is obtained numerically, and compared with analytical results obtained under free-draining conditions both for low and for very large shear rates. The shortcomings of the theory are also discussed, in particular a substantial stretching under shear of the central springs, where the intramolecular tension is largest, with a corresponding strong contraction of the end springs due to the average character of the constraint.     

Nonlinear viscoelasticity of polystyrene solutions. II. Non-Newtonian viscosity

The steady shear viscosity η(k) and the stress decay function \documentclass{article}\pagestyle{empty}\begin{document}$ \tilde \eta \left({t,k} \right)$\end{document} (the shear stress divided by the rate of shear k after cessation of steady shear flow) were measured for concentrated solutions of polystyrene in diethyl phthalate. Ranges of molecular weight M and concentration c were 7.10 × 10⁵ to 7.62 × 10⁶ and 0.112–0.329 g/cm³, respectively. Measurements were performed with a rheometer of the cone-and-plate type in the range 10^?4 < k < 1 sec^?1. The Cox–Merz relation η(k) = |η^*(ω)|_ω=k was tested with the experimental result (|^*(ω)| is the magnitude of the complex viscosity). It was found to be applicable to solutions of relatively low M or c but not to those of high M and c. For the latter η(k) began to decrease at a lower rate of shear than |η^*(ω)|_ω=k did; the Cox–Merz law underestimated the effect of rate of shear. The stress decay function was assumed to have a functional form \documentclass{article}\pagestyle{empty}\begin{document}$\tilde \eta \left( {t,k} \right) = \sum {\eta _p \left( k \right)e^{ - t/\tau p\left( k \right)} } $\end{document} where τ₁ > τ₂ > …, and the values of τ₁, τ₂ η₁ and η₂ were determined for some solutions. The relaxation times τ₁ and τ₂ were found to be independent of k and equal to the relaxation times of linear viscoelasticity. At the limit of k → 0, η₁ and η₂ were approximately 60 and 20–30%, respectively, of η and the non-Newtonian behavior was due to large decreases of η₁ and η₂ with increasing k. It was shown that η₁(k) may be evaluated from the relaxation strength G₁(s) for the longest relaxation time of the strain-dependent relaxation modulus with a constitutive model for relatively high c–M systems as well as for low c–M systems.     

Tensile stress and recoverable strain measurements on polystyrene melts. Interpretation of results in terms of rubberlike elasticity

Extensional tests at constant strain rate \documentclass{article}\pagestyle{empty}\begin{document}$ \dot \varepsilon $\end{document} have been carried out on polystyrene melts with different molecular weight distributions at various temperatures and strain rates. The true tensile stress is found to be well approximated by the sum of two contributions: (1) a neo-Hookean expression involving the recoverable strain and (2) a contribution rapidly reaching a steady-state value. Two experimental parameters can be defined: an elasticity modulus \documentclass{article}\pagestyle{empty}\begin{document}$ G(\dot \varepsilon ) $\end{document} from (1) and a viscosity \documentclass{article}\pagestyle{empty}\begin{document}$ \eta _{\rm v} (\dot \varepsilon ) $\end{document} from (2). It is further shown that time-temperature equivalence applies not only to the stress but also to the recoverable strain and to G and η_v. The dependence of G and η_v on strain rate is then discussed. For high strain rates, G is close to the linear viscoelastic plateau modulus of PS melts and decreases with decreasing strain rate. The value of η_v is found to a good approximation to be equal to three times the shear viscosity taken at a shear rate equivalent to the elongational strain rate.     

Rheology of polystyrene solutions around the coil overlapping concentration: A phenomenological description of stresses in simple shear flow

Linear viscoelasticity behavior is described with the sum of two terms for polystyrene solutions in tricresyl phosphate around the coil overlapping concentration (K. Osaki, T. Inoue, & T. Uematsu, J Polym Sci Part B: Polym Phys 2001, 39, 211). One is a Rouse–Zimm (RZ) term represented by the Zimm theory with arbitrarily chosen values of the hydrodynamic interaction parameter and the longest relaxation time (τ_RZ). The other (the L term) consists of a relaxation mode with a single relaxation time (τ_L > τ_RZ) and a high‐frequency limiting modulus proportional to the square of the concentration. In this study, we describe the viscosity (η) and first normal stress coefficient (Ψ₁) in steady shear with simple formulas. The stress due to the L term is assumed to be given by a Kaye, Bernstein, Kearsley, and Zapas (K‐BKZ) equation with the damping function h(γ) = (1 + 0.2γ²)^?1/2, where γ is the magnitude of shear. Contributions to η and Ψ₁ from the RZ term are derived from the RZ model, in which the relaxation time in steady flow is given by τ_st = τ + (τ_RZ ? τ)/(1 + 0.35τ_RZ γ˙) instead of τ_RZ. Here, γ˙ is the rate of shear, and τ is the τ_RZ value at the infinite dilution limit. η and Ψ₁ at various concentrations for two polystyrene samples (with molecular weights of 2890 and 8420 kg mol^?1) are well described with parameters derived from dynamic viscoelasticity. © 2002 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 40: 1038–1045, 2002     

Entanglement friction and dynamics in blends of starlike and linear polymer molecules

We investigate relaxation dynamics in a series of six‐arm star/linear 1,4‐polybutadiene blends with mechanical rheometry measurements. Blend systems are formulated to systematically probe constraint release and arm relaxation dynamics. Zero shear viscosity and terminal relaxation times of star/linear polymer blends with fixed star arm molecular weights (M_a) and compositions (?_S) are found to follow nonmonotonic dependencies on the linear polymer molecular weight (M_L). At low values of ?_S, at least two scaling regimes are apparent from the data (ξ₀ ～ M and ξ₀ ～ M), where ξ₀ refers to the zero shear viscosity or terminal relaxation time of the blend. The two regimes are separated by a critical linear polymer molecular weight M^* that is more than 20 times larger than the critical molecular weight for entanglements. When the linear polymer contribution to blend properties is removed, a clear transition from dilution dynamics, ξ₀ ～ M, to Rouse‐like constraint‐release dynamics, ξ₀ ～ M, is apparent at low values of ?_S. At higher ?_S values, a new activated constraint‐release dynamic regime is evident in which ξ₀ ～ M and ξ₀ ～ ?, where α changes continuously from approximately 2 to 0.5 as ?_S increases and β varies from 2.0 to 1.0 as M_L increases. The experimental results are compared with theoretical predictions based on a drag coupling model for entangled polymer liquids. All features observed experimentally are captured by this model, including the value of M^* for the transition from dilution to Rouse constraint‐release dynamics. Predictions of the drag coupling model are also compared with published data for the zero shear viscosity and terminal relaxation time in bidisperse linear polymer blends and pure entangled starlike molecules. © 2001 John Wiley & Sons, Inc. J Polym Sci Part B: Polym Phys 39: 2501–2518, 2001     

Nonlinear rheology of highly entangled polymer solutions in start‐up and steady shear flow

Orientation angle and stress‐relaxation dynamics of entangled polystyrene (PS)/diethyl phthalate solutions were investigated in steady and step shear flows. Concentrated (19 vol %) solutions of 0.995, 1.81, and 3.84 million molecular weight (MW) PS and a semidilute (6.4 vol %) solution of 20.6 million MW PS were used to study the effects of entanglement loss on dynamics. A phase‐modulated flow birefringence apparatus was developed to facilitate measurements of time‐dependent changes in optical equivalents of shear stress (n₁₂ ≈ Cσ) and first normal stress differences (n₁ = n₁₁ ? n₂₂ ≈ CN₁) in a planar‐Couette shear‐flow geometry. Flow birefringence results were supplemented with cone‐and‐plate mechanical rheometry measurements to extend the range of shear rates over which entangled polymer dynamics are studied. In slow (τ > ) steady shear‐flow experiments using the ultrahigh MW polymer sample (20.6 × 10⁶ MW PS), steady‐state n₁₂ and n₁ results manifest unusual power‐law dependencies on shear rate [n_12,ss ～ ^0.4 and n_1,ss ～ ^0.8]. At shear rates in the range τ < < τ, steady‐state orientation angles χ_SS are found to be nearly independent of shear rate for all but the most weakly entangled materials investigated. For solutions containing the highest MW PS, an approximate plateau orientation angle χ_p in the range 20–24° is observed; χ_p values ranging from 14 to 16° are found for the other materials. In the start‐up of fast steady shear flow (γ˙ ≥ τ), transient undershoots in orientation angle are also reported. The molecular origins of these observations were examined with the help of a tube model theory that accommodates changes in polymer entanglement density during flow. © 2001 John Wiley & Sons, Inc. J Polym Sci Part B: Polym Phys 39: 2275–2289, 2001     

Electron density and its derivatives at the nucleus for spherically confined hydrogen atom

It is shown that the energy of a hydrogen‐like atom confined inside a spherical cavity of radius, R, and potential barrier, V₀, is quantitatively defined by the ratio . Here, the conventional spherical density (r) is scaled as η_l(r) = and the ratio of the second derivative η(r) to η_l(r) is evaluated at the nucleus. Numerical results of the ratios are presented for 1s, 2s, 2p, and 3d states at several values of V₀. For such states, the characteristic radii of confinement leading to the well‐defined values of energy are identified. © 2008 Wiley Periodicals, Inc. Int J Quantum Chem, 2009     

Steady flow measurements on moderately concentrated solutions of poly(sodium acrylate)

Shear stress P₂₁ and normal stress P₁₁ –P₂₂ functions of moderately concentrated aqueous solutions of sodium poly (acrylate) were measured under steady shear flow at various shear rates \documentclass{article}\pagestyle{empty}\begin{document}$\dot \gamma$\end{document} with a Weissenberg rheogoniometer. The compliance function J_s = (P₁₁ – P₂₂)/2P₂₁² is found to vary with shear rate. The double logarithmic plot of J_s versus \documentclass{article}\pagestyle{empty}\begin{document}$\dot \gamma$\end{document} can be superposed with respect to molecular weight at constant concentrations of added neutral salt and polymer. These reduced plots differ markedly as the concentration of added salt is varied but converge to a single line at high shear rates.     

Anisotropy analysis of the surface stress in Ag stepped surfaces with the modified embedded atom method

The surface stresses in Ag stepped surfaces (910), (710), (510), (410), (310), (210), (320), (430) and (540) have been calculated by using the modified embedded atom method (MEAM). The surface stresses in the surface plane τ_xx (along the step edge) and τ_yy (normal to the step edge) have similar orders of magnitude as the surface energy. For surfaces having the (100) and (110) terraces, the change of τ_xx and τ_yy is very small and the variation of the surface energy is smooth. The stress τ_zz (normal to the surface plane) is always tensile in the unrelaxed state. The linear variation of the change in surface energy per unit change in elastic strain , and (i.e. τ_xx ? γ, τ_yy ? γ and τ_zz) with the angle α between the (hk0) and (100) planes has a turning point corresponding to the (210) surface. The anisotropic ratio in the stepped surface having the (110) terrace is larger than that having the (100) terrace, and the wider the (110) terrace or the narrower the (100) terrace, the larger the ratio τ_yy/τ_xx. Copyright © 2006 John Wiley & Sons, Ltd.     

Electron Transfer Processes at Aryl‐Modified Glassy Carbon Electrode

The effect of a layer of electrochemically grafted 4‐diazo‐N,N‐diethylaniline (DEA) groups on the electron transfer kinetics of redox systems, displaying fast and slow heterogeneous electron transfer rate constants at edge and basal planes of carbon, was investigated. The properties of the modified electrode were characterized by cyclic voltammetry using four different inorganic redox systems (Fe(CN) , Co(phen) , Ru(NH₃) , and IrCl in acidic, neutral, and basic media. Two distinct blocking behaviors and electrostatic effects were observed. More precisely, a strong blocking effect of the grafted layer on Fe(CN) and Co(phen) was found, whereas Ru(NH₃) and IrCl showed to be rather unaffected by the presence of the DEA grafted layer.     

Stepwise Strategy to Cyclometallated PtII Complexes with N‐Heterocyclic Carbene Ligands: A Luminescence Study on New β‐Diketonate Complexes

The imidazolium salt 3‐methyl‐1‐(naphthalen‐2‐yl)‐1H‐imidazolium iodide ( 2 ) has been treated with silver(I) oxide and [{Pt(μ‐Cl)(η³‐2‐Me‐C₃H₄)}₂] (η³‐2‐Me‐C₃H₄=η³‐2‐methylallyl) to give the intermediate N‐heterocyclic carbene complex [PtCl(η³‐2‐Me‐C₃H₄)(H$\widehat{CC}$ *‐κC*)] ( 3 ) (H$\widehat{CC}$ *‐κC*=3‐methyl‐1‐(naphthalen‐2‐yl)‐1H‐imidazol‐2‐ylidene). Compound 3 undergoes regiospecific cyclometallation at the naphthyl ring of the NHC ligand to give the five‐membered platinacycle compound [{Pt(μ‐Cl)($\widehat{CC}$ *)}₂] ( 4 ). Chlorine abstraction from 4 with β‐diketonate Tl derivatives rendered the corresponding neutral compounds [Pt($\widehat{CC}$ *)(L‐O,O′)] {L=acac (HL=acetylacetone) 5 , phacac (HL=1,3‐diphenyl‐1,3‐propanedione) 6 , hfacac (HL=hexafluoroacetylacetone) 7 }. All of the compounds ( 3 – 7 ) were fully characterized by standard spectroscopic and analytical methods. X‐ray diffraction studies were performed on 5 – 7 , revealing short Pt?Pt and π–π interactions in the solid‐state structure. The influence of the R‐substituents of the β‐diketonate ligand on the photophysical properties and the use of the most efficient emitter, 5 , as phosphor converter has also been studied.     

Hole growth as a microrheological probe to measure the viscosity of polymers confined to thin films

We review recent hole growth measurements performed at elevated temperatures in freely-standing polystyrene (PS) films, using optical microscopy and a differential pressure experiment (DPE). In the hole growth experiments, which were performed at temperatures close to the bulk glass-transition temperature of PS, T = 97 °C, we find evidence for nonlinear viscoelastic effects, which markedly affect the growth of holes in freely-standing PS films. The hole radius R initially grew linearly with time t before undergoing a transition to exponential growth characterized by a growth time τ. The time scale τ₁ for the decay of the initial transient behavior prior to reaching steady state was consistent with the convective constraint release mechanism of the tube theory of entangled polymer dynamics, while the characteristic hole growth times τ of the holes were consistent with significant reductions in viscosity of over eight orders of magnitude with increasing shear strain rate due to shear thinning. DPE measurements of hole growth on very thin freely-standing films revealed that hole formation and growth occurs only at temperatures that are comparable to or greater than T, even for films for which the T_g value was reduced by many tens of degrees Celsius below the bulk value. © 2006 Wiley Periodicals, Inc. J Polym Sci Part B:Polym Phys 44: 3011–3021, 2006     

Nitroxide spin probe study of probe size,hydrogen bonding and polymer matrix rigidity effects on poly(acrylic acid)/poly(ethylene oxide) complexes

An electron spin resonance (ESR) spin probe study was performed on 1 : 1 by weight poly(acrylic acid) (PAA)/poly(ethylene oxide) (PEO) complex over the 100–450 K temperature range with a series of tetramethylpiperidyloxy‐based spin probes. Measurements of the parameters T_5mT, Ta and Td demonstrated the effects of probe size and the strength of hydrogen bonding. The probes in the series Tempone, Tempo, Tempol and Tamine (respectively 4‐oxo‐, unsubstituted, 4‐hydroxy‐ and 4‐amino‐2,2,6,6,‐tetramethylpiperidine ‐1‐oxyl) displayed noticeable increases in the hydrogen‐bonding effect, as indicated by Ta and Td. These increases correlated with increasing hydrogen bond acceptor strength. On the other hand, as the probe size became larger, T_5mT gradually increased due to the free volume decrease. These effects were analyzed using the established theoretical relationship of T_5mT to probe volume expressed by f. Meanwhile, in order to investigate the effect of polymer matrix rigidity, a similar study was performed with a nitroxide spin probe, 2,2,6,6‐tetramethyl‐1‐piperidine‐1‐oxyl (Tempo), on PAA/PEO complexes of different weight compositions. The quantitative fast motion fraction in the composite ESR spectrum was calculated. The influence of changes in the composition of PAA on the molecular mobility was characterized by changes of the spectral parameters and τ_c. The molecular mobility was shown to diminish with increasing content of PAA in PAA/PEO blends duo to the restriction of the polymer matrix rigidity increase. Copyright © 2003 John Wiley & Sons, Ltd.     

Stable γ-distonic oxonium ions: R\documentclass{article}\pagestyle{empty}\begin{document}$ \mathop {\rm O}\limits^ + $\end{document}HCHR′ĊH2CHR″ and their characteristic reaction

The γ-distonic radical ions R$ \mathop {\rm O}\limits^ + $CHR′CH₂?HR″ and their molecular ion counterparts R$ \mathop {\rm O}\limits^{{\rm + } \cdot } $CHR′CH₂CH₂R″ have been studied by isotopic labelling and collision-induced dissociation, applying a potential to the collision cell in order to separate activated from spontaneous decompositions. The stability of CH₃$ \mathop {\rm O}\limits^ + $HCH(CH₃)CH₂?HCH₃, C₂H₅$ \mathop {\rm O}\limits^ + $HCH(CH₃)CH₂?HCH₃, CH₃$ \mathop {\rm O}\limits^ + $HCH(CH₃)CH₂?H₂, CH₃$ \mathop {\rm O}\limits^ + $HCH₂CH₂?HCH₃ and C₂H₅$ \mathop {\rm O}\limits^ + $HCH₂CH₂?HCH₃, has been demonstrated and their characteristic decomposition, alcohol loss, identified. For all these γ-distonic ions, the 1,4-H abstraction leading to their molecular ion counterpart exhibits a primary isotope effect.     

Incorporating glycidyl methacrylate into block copolymers using poly(methacrylate‐ran‐styrene) macroinitiators synthesized by nitroxide‐mediated polymerization

Methyl methacrylate/styrene (MMA/S), ethyl methacrylate/styrene (EMA/S) and butyl methacrylate/styrene (BMA/S) feeds (>90 mol % methacrylate) were copolymerized in 50 wt % p‐xylene at 90 °C with 10 mol % of additional SG1‐free nitroxide mediator relative to unimolecular initiator (BlocBuilder^®) to yield methacrylate rich copolymers with polydispersities _w/ _n = 1.23–1.46. k_pK values (k_p = propagation rate constant, K = equilibrium constant) for MMA/S copolymerizations were comparable with previous literature, whereas EMA/S and BMA/S copolymerizations were characterized by slightly higher k_pK's. Chain extensions with styrene at 110 °C initiated by the methacrylate‐rich macroinitiators (number average molecular weight _n = 12.9–33.5 kg mol^?1) resulted in slightly broader molecular weight distributions with _w/ _n = 1.24–1.86 and were often bimodal. Chain extensions with glycidyl methacrylate/styrene/methacrylate (GMA/S/XMA where XMA = MMA, EMA or BMA) mixtures at 90 °C using the same macroinitiators resulted frequently in bimodal molecular weight distributions with many inactive macroinitiators and higher _w/ _n = 2.01–2.48. P(XMA/S) macroinitiators ( _n = 4.9–8.9 kg mol^?1), polymerized to low conversion and purified to remove “dead” chains, initiated chain extensions with GMA/MMA/S and GMA/EMA/S giving products with _w/ _n ～ 1.5 and much fewer unreacted macroinitiators (<5%), whereas the GMA/BMA/S chain extension was characterized by slightly more unreacted macroinitiators (～20%). © 2009 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 47: 2574–2588, 2009