No abstract

A positron annihilation lifetime spectroscopic (PALS) study was carried out on a typical representative of simple hydrogen-bonded glass formers-glycerolover a wide temperature region from 100 K up to room temperature. Several crossover temperatures, T bi , were identified in the temperature dependences of the ortho-positronium (o-Ps) lifetime and the relative o-Ps intensity. The onset temperature of free-volume change, 137 K, was compared with the value of 135 ± 3 K obtained from extrapolating thermodynamic data and with the temperatures from relaxation data. In order to determine the influence of the free-volume evolution on the relaxation dynamics above the glass transition temperature T g , the free-volume fraction from the PALS data was related to the behaviour of the dielectric loss of glycerol. A new relationship between the temperature parameters of the onset of free-volume change and of the highfrequency wing of the dielectric loss is presented. Taking the PALS free-volume data as input for a free-volume model (Doolittle ansatz), it was found that the relaxation times τ α can be reproduced above the bend temperature T b2 = 241 K; the latter lies in the vicinity of crossover temperatures reported from mode coupling theory analyses of dynamic data. In the lower-temperature region T b2 > T > T b1 ∼ = T g , a contribution from thermally activated mobility has to be included to reproduce the dielectric relaxation time data.

We have reexamined the free-volume concept presented by Cohen and Turnbull on the basis of two microscopic quantities: the excess mean-square displacement (f) and the total free volume V(PA,t), of poly- butadiene evaluated from the quasielastic neutron scattering and the positron annihilation lifetime spectroscopy (PALS) data, respectively. Comparing with the viscosity eta we found two relations, eta=eta(0) exp u( 2)(0)/(f) and eta=eta(0) exp V(*)(PA,0)/V(PA,t)=eta(0) exp V(*)(PA,0)/v(PA,f), where u(2)(0), V(*)(PA,0) and v(*)(PA,0) are the critical values for the mean-square displacement, the total PALS free volume, and the PALS free volume per molecule, respectively, and further v(*)(PA,0)=V(*)(PA,0)/N, N being the total number of molecules or segments. On the basis of these relations, we discuss the microscopic basis of the free-volume theory. The experimentally evaluated critical values u(2)(0) and v(*)(PA,0) are much larger than the average values of (f) and v(PA,f) calculated from the distributions. This has been explained from the low probability of escaping motions from a molecular cage. The free volume per monomer and the free-volume fraction were calculated from the excess mean-square displacement (f). The former was compared with the free-volume hole obtained by PALS, suggesting that 22 monomers are required for one PALS free-volume hole. The free-volume fraction obtained from the excess mean-square displacement was found to be 6.4% at 250 K, which is in reasonable agreement with that evaluated from the rheological data (9.0%).

A free-volume theory is developed based on the defect diffusion model (DDM). In addition, positronium annihilation lifetime spectroscopy (PALS) ortho-positronium free-volume and intensity data are presented for poly(propylene glycol) with a molecular weight of 4000 (PPG 4000) in both the glassy and liquid states and dielectric relaxation and electrical conductivity data are reported for PPG 4000 in the liquid state. The DDM is used to interpret all of the data for PPG 4000 and previously reported PALS and dielectric relaxation data for glycerol. It is shown that while the PPG 4000 data exhibit a preference for the three-halves power law, the data for glycerol favor the first power (standard Vogel-Fulcher-Tammann) law. Good agreement between the DDM and the experimental results is found for all of the electrical data and the PALS free-volume data. While reasonable agreement is also found for the PALS intensity data for PPG 4000, a discrepancy exists between the experimental PALS intensity data and theory for glycerol. For the electrical conductivity for PPG 4000, a transition is observed at the same temperature (about 1.4 T(g) where T(g) is the glass transition temperature) where the PALS free volume changes from steeply rising with temperature to approximately independent of temperature. The same behavior is observed at about 1.5 T(g) for previously reported dielectric relaxation and PALS data for glycerol. Model parameters are presented that show the dominance of mobile single defects above (1.4-1.5) T(g) and the dominance of immobile clustered single defects below T(g) . Finally, a coherent picture of glasses and glass-forming liquids is presented based on the theory and results of the experiments.