Scaling analysis of cotton cellulose/LiCl·DMAc solution using light scattering and rheological measurements

Semidilute solution of cotton lint (CC1) in 8 wt % LiCl/N,N‐dimethylacetamide was investigated using static light scattering (SLS) and rheological measurements. The reduced osmotic modulus estimated by SLS measurements for CC1 solutions are proportional to c^1.16 in the semidilute region. From the exponent of 1.16, de Gennes' scaling theory derives the relationship between radius of gyration, R_g, and molecular weight, M_w, of CC1 as R_g ∝ M^0.62 This corresponds to the Mark‐Houwink‐Sakurada exponent of 0.86. This exponent is very close to that estimated from scaling analysis of zero shear rate viscosity, that is 0.85. Apparent radius of gyration, R_g,app, estimated by SLS measurements for CC1 solutions are proportional to c^?0.5 in the semidilute region. R_g,app indicates the mesh size of polymer entanglement in the semidilute region. On the assumption of the Gaussian behavior of CC1 molecule in the semidilute region, the exponent of ?0.5 gives the relationship between the molar mass between entanglements, M_e, and c as following relationship: M_e ∝ c^?1. This agrees with the concentration dependence on plateau modulus estimated from the dynamic viscoelastic measurements. © 2006 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 44: 2155–2160, 2006     

Molecular dimensions and melt rheology of an all‐aromatic liquid crystalline polymer

The molecular dimensions and melt rheology of a thermotropic all‐aromatic liquid crystalline polyester (TLCP) composed of p‐hydroxy benzoic acid, hydroquinone, terephthalic acid, and 2,4‐naphthalenedicarboxylic acid is examined. The Mark–Houwink exponent (α) of 0.95 is estimated for the TLCP. The persistence length estimated from molecular weight (M) and intrinsic viscosity ([η]) data using the Bohdanecky–Bushin equation is about 95 Å, whereas that estimated from light scattering data is 117 Å. These persistence lengths and the observed α value, both higher than those for flexible polymers, suggest that the present TLCP is a semirigid polymer. The zero shear melt viscosity (η₀) varies with approximately M⁶ for molecular weight M > 3 × 10⁴ g/mol; below this molecular weight, η₀ varies almost linearly with M. Widely different entanglement molecular weights (M_e) are predicted, depending on the method used; the plateau modulus estimates M_e of about 8 × 10⁵ g/mol, whereas the ratio of mean square end‐to‐end distance and molecular weight (〈R²〉₀/M) predicts M_e's either too small (0.33 g/mol) or too large (2.5 × 10⁶ g/mol), depending on the theory used. Although the change in the molecular weight dependency of melt viscosity appears to be associated with the onset of entanglement coupling of the semirigid molecules, its origin needs further investigation. © 2001 John Wiley & Sons, Inc. J Polym Sci Part B: Polym Phys 39: 2378–2389, 2001     

Explanation for the 3.4-power law for viscosity of polymeric liquids on the basis of the tube model

The viscosity η₀(M) of polymeric liquids of molecular weight M is calculated on the basis of the tube model formulated by Doi and Edwards (ref. 3). The contour length fluctuation of polymers along the tube, which was neglected in ref. 3, is now explicitly taken into account. The result is where M_c = 2M_e, and M_e is the molecular weight between the entanglement points. This result is numerically close to the empirical 3.4-power law, η₀(M) = η₀(M_c)(M/M_c)^3.4, for 10M_c ? M ? 100M_c but approaches the result in ref. 3 for very high molecular weight. We thus conclude that the 3.4-power law is actually an approximate expression for the real curve which slowly approaches the asymptotic form calculated in ref. 3.     

Diffuse interface between polymers: Structure and kinetics

The interfacial structure and diffusion kinetics of two compatible polymers, poly(methyl methacrylate) and poly(vinylidene fluoride) are studied in the melt. The interdiffusion rates of the two components are found to be unequal, giving unequal diffusion coefficients, a net mass flow across the interface, and an asymmetric interfacial composition profile. The structure and kinetics confirm the predictions of the reptation theory. The interfacial thickness d grows with t^1/2, and the interdiffusion coefficient is proportional to M^?2, where t is the time and M is the molecular weight. The scaling law for the interfacial thickness is therefore d ∝ M^?1t^1/2. The number of chains per unit area crossing the original interface reaches a constant value independent of diffusion time after a short induction time on the order of the tube disengagement time (about 0.1–10 s in the present cases depending on the molecular weights). The adhesive bond strength σ is scaled by σ ∝ t^1/4M^?1/2 and σ/σ∞ ∝ t^1/4M^?1/2 [1- (M_c/M)]^?1, where σ _∞is the σ at infinite molecular weight and M_c is the entanglement molecular weight.     

Rheological properties of poly(lactides). Effect of molecular weight and temperature on the viscoelasticity of poly(l‐lactic acid)

The dynamic viscoelastic behavior of Poly(l‐lactic acid) (PLLA), with molecular weights ranging from 2,000 to 360,000, have been studied over a broad range of reduced frequencies (approximately 1 × 10⁻³ s⁻¹ to 1 × 10³ s⁻¹), using time–temperature superposition principle. Melts are shown to have a critical molecular weight, M_c, of approximately 16,000 g/mol, and an entanglement density of 0.16 mmol/cm³ (at 25°C). PLLA polymers are noted to require substantially larger molecular weights in order to display similar melt viscoelastic behavior, at a given temperature, as that for conventional non‐biodegradable polymers such as polystyrene. The reason for this deviation is suspected to be due to steric hindrance, resulting from excessive coil expansion or other tertiary chain interactions. PLLA melts show a dependence of η₀ on chain length to the 4.0 power (M), whilst J is independent of M_W in the terminal region. Low molecular weight PLLA (∼ 40,000) shows Newtonian‐like behavior at shear rates typical of those achieved during film extrusion. © 1999 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 37: 1803–1814, 1999     

Effect of the directionality of the ester group on the formation of hairpins in polyesters containing naphthalene and a flexible spacer

Steady-state fluorescence measurements and molecular dynamics simulations have been used to study the intramolecular formation of excimers in five model compounds for polyesters containing naphthalene groups separated by flexible spacers. The model compounds are derived from 2-hydroxynaphthalene and HOOC (CH₂)_n COOH, n = 2–6. The ratio of the intensity of excimer and monomer emissions, I_D/I_M, is nearly independent of the viscosity of the medium, η, over the range covered in dilute solution. Although I_D/I_M is always very small, it shows an odd–even effect for the first four members of the series, with maxima when n is odd. Molecular dynamics simulations provide an explanation for the small values of I_D/I_M, their weak dependence on η, and the trend of I_D/I_M with n. The results for the present series of model compounds are compared with previous work, which reported larger values of I_D/I_M, and a stronger dependence of I_D/I_M on η, for bichromophoric compounds derived from 2-naphthoic acid and aliphatic glycols, where the direction of the ester groups is reversed. The origin of the difference in the behavior of I_D/I_M in the two series is identified. © 1997 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 35: 1127–1133, 1997     

Rheological properties of poly(2,6-dimethylphenylene oxide)—polystyrene blends

The viscoelastic (VE) response of freeze-dried blends of polystyrene (PS) and poly-(2,6-dimethyl phenylene oxide) (PPO) has been studied as a function of composition, frequency, and temperature to examine the degree of rheological compatibility. When blended together, the relaxation processes of both molecular species exhibit the same temperature dependence. However, the temperature dependence of the VE response is a function of composition. It is shown that this behavior can be predicted from the measured glass transition temperatures by assuming the additivity of the free volumes of the components. The properties of the blends are compared at equal free volumes. The effective segmental friction factor is found to be independent of composition while the modulus of the rubbery plateau increases with PPO concentration. This result is interpreted as a change in the entanglement molecular weight M_e of the blends. When the changes in M_e are considered, the relationship between the zero-shear viscosity η0 and the 3.4 power of the weight-average molecular weight, commonly found for high molecular weight homopolymers, predicts the compositional dependence of η0 for the PPO–PS blends. It is concluded that the PPO–PS system forms a rheologically compatible blend.     

Semi‐perfluoroalkylated perylene diimides for conjugated polymers with high molecular weight and high electron mobility

Perylene diimides (PDIs) and their derivatives are excellent semiconductors, while conjugated polymers based on PDIs have limited applications because of their low electron mobility (μ_e) derived from low molecular weight. The reported maximum number‐average molecular weight (M_n) of related polymers is only 21 kDa because PDIs have very poor solubility due to strong π–π stacking of their big planar conjugated cores. Herein, it is found that suitable semi‐perfluoroalkyl groups could enhance the solubility of PDIs significantly, and a series of semi‐perfluoroalkyl modified conjugated polymers with high molecular weight and electron mobility were synthesized. The maximum M_n reaches 94.8 kDa [P(4C_F8C_H‐PDI‐T2)_HW]. In their space‐charge‐limited current (SCLC) devices, all polymers exhibit typical characters of electron transporting semiconductors, and the highest μ_e is up to 8.40 × 10⁻³ cm² V⁻¹ s⁻¹ [P(4C_F8C_H‐PDI‐T2)_HW], which is similar as that of widely used electron transporting semiconductor PC₆₁BM (6.41 × 10⁻³ cm² V⁻¹ s⁻¹). © 2017 Wiley Periodicals, Inc. J. Polym. Sci., Part A: Polym. Chem. 2018 , 56, 116–124     

Metallocene-methylaluminoxane catalysts for olefin polymerization. V. Comparison of Cp2ZrCl2 and CpZrCl3

The ethylene polymerization by Cp₂ZrCl₂/MAO (Cp = η⁵: cyclopentadienyl; MAO = methyl aluminoxane) and CpZrCl₃/MAO have been studied. The MW and PD (= M _w/M _w) of polymers obtained after 2.5-60 min are the same, which indicate short chain lifetime. The values of rate constants for Cp₂ZrCl₂ at 70°C are: k_p = 168?1670 (M s)^?1 and k_tr^A1 = 0.012-0.81 s^?1 depending upon [Zr] and [MAO,] k_tr^β = 0.28 s^?1, and k_tr^H = 0.2 M^?1 torr^?1/2 s^?1. These chain transfer rate constant values are two to three orders of magnitude greater than the corresponding values found for MgCl₂ supported titanium catalysts. One significant difference between the heterogeneous and homogeneous catalysts is that the former decays according to an apparent second order kinetics, whereas the latter decay is simple first order at 0°C and biphasic first order at higher temperatures. The productivity of the catalysts depends weekly on temperature while the MW decreases strongly with increase of temperature above 30°C. All the active species were formed upon mixing Cp₂ZrCl₂ with MAO while it took up to 20 min for the CpZnCl₃/MAO system. The productivity of the former increase more strongly with the decrease of [Zr] than the latter. Otherwise, the two catalyst systems have all their kinetic parameters differing less than a factor of two.     

Prediction of microstructures for polydisperse block copolymers,using continuous thermodynamics

A generalization of an earlier theory (Leary–Henderson–Williams) developed for microphase separation in monodisperse block copolymers is made for copolymers having moderate degrees of polydispersity and illustrated for the Schultz molecular weight distribution (MWD). First, an explicit study is made of molecular weight (M) effects for monodisperse poly (styrene–butadiene) diblock (SB) and triblock (SBS) copolymers. For a fixed temperature, it is shown how the critical molecular weight (M_c)—above which the copolymer is phase-separated at equilibrium —varies with molecular composition (?_S, volume fraction of S component) for both molecular architectures. Also predicted are the microstructural parameters ΔT(M) and f(M)—interphase thickness and volume fraction, respectively—and the high-M limiting functions ΔT ∝? M^α2, f ∝? M^α3, D ∝? M^α4 (D is domain repeat distance) and T_s ∝? M^α5 (T_s is separation temperature). Then, for polydisperse systems in the range 1 ? p ? 3 ( where \[ P = \bar M_w /\bar M_n \] ) corresponding predictions at constant \[ \bar M_n \] are made after identifying the mixture free-energy-minimum state with a weight average of the free energy minima of each fraction of the MWD. Calculations are made specifically for ?_S = 0.50 and T_s = 298 K. It is shown that, even when \[ \bar M_n < M_c \] , polydispersity can induce microphase separation if p is sufficiently large. Good success is obtained in comparisons of D predictions with data on blends of two polydisperse diblock samples.     

Dynamic mechanical properties of moderately concentrated polystyrene solutions

The storage (G′) and loss (G″) shear moduli have been measured in the frequency range from 0.04 to 630 Hz for solutions of narrow distribution polystyrenes with molecular weights (M) 19,800 to 860,000, and a few of poly(vinyl acetate), M = 240,000. The concentration (c) range was 0.014–0.40 g/ml and the viscosities of the solvents (diethyl phthalate and chlorinated diphenyls) ranged from 0.12 to 70 poise. Data at different temperatures (0–40°C) were combined by the method of reduced variables. Two types of behavior departing from the usual frequency dependence describable by the Rouse-Zimm-Tschoegl theories were observed. First, for M ? 20,000, the ratio (G″ ? ωη_s)/G′ in the neighborhood of ωτ₁ = 1 was abnormally large and the steady-state compliance J was abnormally small, especially at the lowest concentrations studied. Here ω is circular frequency, η_s solvent viscosity, and τ₁ terminal relaxation time. Related anomalies have been observed by others in undiluted polymers at still lower molecular weights. Second, at the highest concentrations and molecular weights, a “crossover” region of the logarithmic frequency scale appeared in which G″ ? ωη_s < G′. The width of this region is a linear function of log c; the frequency dependence under these conditions can be represented by a sequence of Rouse relaxation times grafted on to a sequence of Zimm relaxation times. For each molecular weight, the terminal relaxation time τ₁ was approximately a single function of c for different solvents of widely different η_s. At lower concentrations, τ₁ was close to the Rouse prediction of 6ηM/π²cRT, where η is the steady-flow viscosity; but at higher concentrations, τ₁ was proportional to η/c² and corresponded, according to a recent theory of Graessley, to an average molecular weight of 20,000 between entanglement coupling points in the undiluted polymer.     

Simple method for determining the critical molecular weight from the loss modulus

The normal concept is that the critical molecular weight (M_C) is about twice as large as the entanglement molecular weight (M_e). However, experimental data have shown considerable deviations from M_C ≈ 2M_e. Furthermore, a determination of M_C requires samples with a wide range of molecular weights, including weights lower than M_C and higher than M_C. In this article, we suggest a simple method for determining M_C from the loss moduli of nearly monodisperse linear polymers with M ? M_C. We consider two characteristic relaxation times, which correspond to the local maximum and minimum of the loss modulus. M_C is determined from the intersection of two phenomenological relaxation times as a function of the molecular weight. The method precisely agrees with M_C ≈ 2M_e, which is not shown by conventional methods. Moreover, our method provides a determination of relaxation time τ_e, at which chain segments first feel the constraints imposed by the conceptual tube, without the measurement of the tube diameter and the monomeric friction coefficient, which may be determined by complicated procedures with a lot of data. © 2004 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 42: 2724–2729, 2004     

Solution characterization of the novel organometallic polymer poly(ferrocenyldimethylsilane)

A series of four well‐defined poly(ferrocenyldimethylsilane) (PFS) samples spanning a molecular weight range of approximately 10,000–100,000 g mol⁻¹ was synthesized by the living anionic polymerization of dimethyl[1]silaferrocenophane initiated with n‐BuLi. The polymers possessed narrow polydispersities and were used to characterize the solution behavior of PFS in tetrahydrofuran (THF). The weight‐average molecular weights (M_w ) of the polymers were determined by low‐angle laser light scattering (LALLS), conventional gel permeation chromatography (GPC), and GPC equipped with a triple detector (refractive index, light scattering, and viscosity). The molecular weight calculated by conventional GPC, with polystyrene standards, underestimated the true value in comparison with LALLS and GPC with the triple detection system. The Mark–Houwink parameter a for PFS in THF was 0.62 (k = 2.5 × 10⁻⁴), which is indicative of fairly marginal polymer–solvent interactions. The scaling exponent between the radius of gyration and M_w was 0.54, also consistent with marginal polymer–solvent interactions for PFS in THF. © 2000 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 38: 3032–3041, 2000     

Chain dimensions and entanglement spacings in dense macromolecular systems

In this article, we reexamine and extend a relationship proposed earlier between entanglement density and chain dimensions in polymer melts. The power-law equation presented in the earlier work, relating the entanglement molecular weight M_e, melt chain density ρ, and the packing length p is tested with additional polymer species. Now included are additional polydienes and their hydrogenated derivatives, the isotactic forms of polypropylene and polystyrene, the essentially syndiotactic form of poly(methyl methacrylate), along with poly(tetrafluoroethylene), poly(vinylmethyl ether), various poly(methacrylates), and polymeric sulfur. We find that within experimental uncertainties, M_e/ρ and p are related through an equation (M_e/ρ = 218p³) that is insensitive to temperature (25°C ≤ T ≤ 380°C) and which seems to be universal for flexible Gaussian chains in the melt state. © 1999 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 37: 1023–1033, 1999     

Kinetics of styrene emulsion polymerization above the critical micelle concentration: Effect of the initial monomer concentration on the molecular weight

The emulsion polymerization of styrene above the critical micelle concentration has been experimentally studied from a low final polymer content up to a high polymer content (～50%). A maximum in the molecular weight (M) evolution has been observed in all cases. The presence or absence of such a maximum depends on the relative values of the rate of free‐radical entry (ρ) and the rate of chain transfer to the monomer (K_trC_Mp, where K_tr is the chain transfer to monomer rate coefficient and C_Mp is the monomer concentration in particles). If ρ ? K_trC_Mp, M is constant and equal to K_p/K_tr (where K_p is the propagation rate coefficient), except at very low particles sizes typical of the early stages of the reaction, in which the chain length is limited by the particle size. On the other hand, if ρ ? K_trC_Mp, M is determined by both C_Mp and ρ. It is proposed that ρ is determined by the sum of the entry of the oligomeric radicals formed in the aqueous phase and those contained in particles that undergo limited coagulation. This coagulative entry can become very significant; therefore, reactor hydrodynamics can play a major role in the kinetic behavior observed. Disagreement between Clay and Gilbert's model and molecular weight distribution data can be ascribed, to a lesser or greater extent, to the degree of correctness of the quasi‐steady‐state and instantaneous‐termination approaches. © 2005 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 43: 1963–1972, 2005     

Preparation and characterization of polypropylene/mesoporous silica nanocomposites with confined polypropylene

Polypropylene (PP)/Ti-MCM-41 nanocomposites were prepared by isospecific propylene polymerization with Ti-MCM-41/Al(i-C₄H₉)₃ catalyst. The cross polarization/magic angle spinning (CP/MAS) ¹³C NMR spectrum of the composite was similar to that of the conventional isotactic PP, and the decrease in the pore volume of Ti-MCM-41 in the nanocomposites, as measured by N₂ adsorption, was consistent with the value calculated from the weight loss in the thermogravimetric analysis (TGA) curve; both these facts attest to propylene polymerization within the mesopores of Ti-MCM-41. Alkali treatment followed by extraction with o-dichlorobenzene allows us to extract the confined PP out of the Ti-MCM-41 mesopores. Although the PP/Ti-MCM-41 nanocomposites do not exhibit a crystalline melting point, the same PP when extracted from the mesopores showed a clear melting point at 154.7 °C; this indicates that the crystallization of PP confined in mesopores is strongly hindered. For the PP polymerized within the confinement, the molecular weight (M_w) and molecular weight distribution (M_w/M_n) were 84,000 and 4.3, respectively; these values were considerably smaller than those of the PP polymerized concurrently outside the Ti-MCM-41 mesopores (M_w = 200,000–450,000, M_w/M_n = 40–75). Therefore, the confinement also has a marked effect on the molecular weight of the PP. © 2003 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 41: 3324–3332, 2003     

Chain dimensions and transport coefficients of sodium poly(isoprenesulfonate) in aqueous sodium chloride

Sodium poly(isoprenesulfonate) (NaPIS) fractions consisting of 1,4‐ and 3,4‐isomeric units (0.44:0.56) and ranging in molecular weight from 4.9 × 10³ to 2.0 × 10⁵ were studied by static and dynamic light scattering, sedimentation equilibrium, and viscometry in aqueous NaCl of a salt concentration (C_s) of 0.5‐M at 25 °C. Viscosity data were also obtained at C_s = 0.05, 0.1, and 1 M. The measured z‐average radii of gyration 〈S²〉_z^1/2, intrinsic viscosities [η], and translational diffusion coefficients D at C_s = 0.5‐M showed that high molecular weight NaPIS in the aqueous salt behaves like a flexible chain in the good solvent limit. On the assumption that the distribution of 1,4‐ and 3,4‐isomeric units in the NaPIS chain is completely random, the [η] data for high molecular weights at C_s = 0.5 and 1 M were analyzed first in the conventional two‐parameter scheme to estimate the unperturbed dimension at infinite molecular weight and the mean binary cluster integral. By further invoking a coarse‐graining of the NaPIS molecule, all the [η] and D data in the entire molecular weight range were then analyzed on the basis of the current theories for the unperturbed wormlike chain combined with the quasi‐two‐parameter theory. It is shown that the experimental 〈S²〉_z, [η], and D are explained by the theories with a degree of accuracy similar to that known for uncharged linear flexible homopolymers. © 2001 John Wiley & Sons, Inc. J Polym Sci Part B: Polym Phys 39: 2071–2080, 2001     

Reaction parameters in the RAFT mediated dispersion polymerization of styrene

Monodisperse polystyrene (PS) particles were prepared by a living radical dispersion polymerization with a reversible addition‐fragmentation chain transfer (RAFT) agent in an ethanol medium. In the presence of RAFT agent, the effects of various reaction parameters on the characteristics of PS particles were systematically investigated. When no RAFT agent was involved, the number‐average molecular weight (M_n) of the PS particles increased from 17,800 to 30,000 g/mol, but the weight‐average diameter (D_w) decreased from 2.54 to 2.06 μm with the increase of poly(N‐vinylpyrrolidone) content from 4.0 to 16.0 wt %. No correlation between the M_n and the coefficient of variation (CV) was observed. However, when the RAFT concentration varied from 0 to 2.0 wt %, all of the conversion, M_n, D_w, CV, and polydispersity index (M_w/M_n) decreased. This indicates that the RAFT agent alters the inverse behavior between the molecular weight (MW) and particle size shown in the conventional dispersion polymerization. © 2007 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 46: 872–885, 2008     

Hyperbranched polymers as scaffolds for multifunctional reversible addition–fragmentation chain‐transfer agents: A route to polystyrene‐core‐polyesters and polystyrene‐block‐poly(butyl acrylate)‐core‐polyesters

Polydisperse hyperbranched polyesters were modified for use as novel multifunctional reversible addition–fragmentation chain‐transfer (RAFT) agents. The polyester‐core‐based RAFT agents were subsequently employed to synthesize star polymers of n‐butyl acrylate and styrene with low polydispersity (polydispersity index < 1.3) in a living free‐radical process. Although the polyester‐core‐based RAFT agent mediated polymerization of n‐butyl acrylate displayed a linear evolution of the number‐average molecular weight (M_n) up to high monomer conversions (>70%) and molecular weights [M_n > 140,000 g mol^?1, linear poly(methyl methacrylate) equivalents)], the corresponding styrene‐based system reached a maximum molecular weight at low conversions (≈30%, M_n = 45,500 g mol^?1, linear polystyrene equivalents). The resulting star polymers were subsequently used as platforms for the preparation of star block copolymers of styrene and n‐butyl acrylate with a polyester core with low polydispersities (polydispersity index < 1.25). The generated polystyrene‐based star polymers were successfully cast into highly regular honeycomb‐structured microarrays. © 2003 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 3847–3861, 2003     

Relaxation dynamics of salt‐free polyelectrolyte solutions using flow birefringence and rheometry

Relaxation dynamics of salt‐free, aqueous solutions of sodium poly(styrene sulfonate) (NaPSS) were investigated by mechanical rheometry and flow birefringence measurements. Two semidilute concentration regimes were studied in detail for a range of polymer molecular weights. At solution concentrations c < 10 mg mL, limiting shear viscosity η₀ was found to scale with molecular weight and concentration as η₀ ∼ c^0.5M_w over nearly two decades in concentration. At higher solution concentrations, c > 10 mg mL, a change in viscosity scaling was observed η₀ ∼ c^1.5M, consistent with a change from simple Rouse dynamics for unentangled polyions to near‐perfect reptation dynamics for entangled chains. Characteristic relaxation times τ deduced from shear stress and birefringence relaxation measurements following start‐up of steady shearing at high rates reveal very different physics. For c < 10 mg mL, both methods yield τ ∼ c^−0.42M and τ ∼ c⁰M for c > 10 mg mL. Curiously, the concentration scalings seen in both regimes are consistent with theoretical expectations for salt‐free polyelectrolyte solutions undergoing Rouse and reptation dynamics, respectively, but the molecular weight scalings are not. Based on earlier light scattering studies using salt‐free NaPSS solutions, we contend that the unusual relaxation behavior is likely due to aggregation and/or coupled polyion diffusion. Simultaneous stress and birefringence measurements suggest that in concentrated solution, NaPSS aggregates are likely well permeated by solvent, supporting a loose collective of aggregated chains rather than the dense polymer aggregates previously supposed. Nonetheless, polyion aggregates of either variety cannot account for the inverse dependence of relaxation time on polymer molecular weight for c < 10 mg mL. © 1999 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 37: 825–835, 1999