Enis Tuncer, M. Furlani, B. E. Mellander

• In this article, a recently developed numerical technique [E. Tuncer and S. M. Gubanski, IEEE Trans. Dielectr. Electr. Insul. 8, 310 (2001)] is applied to poly(propylene glycol) (PPG) complex dielectric data to extract more information about the molecular relaxation processes. The method is based on a constrained-least-squares (C-LSQ) data fitting procedure together with the Monte Carlo method. We preselect the number of relaxation times with no a priori physical assumption, and use the Debye single relaxation as "kernel," then the obtained weighting factors at each MC step from the C-LSQ method builds up a relaxation time spectrum. When the analysis is repeated for data at different temperatures a relaxation image is created. The obtained relaxation are analyzed using the Lorentz (Cauchy) distribution, which is a special form of the Levy statistics. In the present report the beta and alpha relaxations are resolved for the PPG. A comparison of the relaxations to those earlier reported in the literature indicate that the presentedIn this article, a recently developed numerical technique [E. Tuncer and S. M. Gubanski, IEEE Trans. Dielectr. Electr. Insul. 8, 310 (2001)] is applied to poly(propylene glycol) (PPG) complex dielectric data to extract more information about the molecular relaxation processes. The method is based on a constrained-least-squares (C-LSQ) data fitting procedure together with the Monte Carlo method. We preselect the number of relaxation times with no a priori physical assumption, and use the Debye single relaxation as "kernel," then the obtained weighting factors at each MC step from the C-LSQ method builds up a relaxation time spectrum. When the analysis is repeated for data at different temperatures a relaxation image is created. The obtained relaxation are analyzed using the Lorentz (Cauchy) distribution, which is a special form of the Levy statistics. In the present report the beta and alpha relaxations are resolved for the PPG. A comparison of the relaxations to those earlier reported in the literature indicate that the presented method provides additional information compared to methods based on empirical formulas. The distribution of relaxation times analysis is especially useful to probe the crossover region where the alpha and beta relaxations merge and the results show that the relaxation after the crossover region at higher temperatures is Arrhenius-type as the beta relaxation. Moreover, this relaxation is more likely to be the continuation of the beta relaxation, but with a different activation energy. (C) 2004 American Institute of Physics…