The role of the Havriliak-Negami relaxation in the description of local structure of Kohlrausch's function in the frequency domain. Part I

Abstract

An improved approximation via Havriliak-Negami (HN) functions to the Fourier Transform (FT) of certain Weibull distributions, -\psi_{\beta}, (the time derivative of the Kohlrausch-Williams-Watts function), is given for a large interval of frequencies: \omega/2\pi\in[0,10^{12}] if 0<\beta\leq1 and \omega/2\pi\in[0,10^{7}] if 1<\beta\leq2. The model is free from the usual numerical distortions, or restrictions associated to sampling step and finite size, present in similar adjustments to complex relaxation functions. Further indicates that the identification of (FT) Weibull data with a double HN approximant, \psi_{\beta}\simeq\mathcal{A}p_{2}HN, is quite exact locally even though the parameters involved should vary adiabatically with the frequency, i.e. \{\alpha_{1,2},\gamma_{1,2},\tau_{1,2},\lambda\}(\omega). This fact is the base for the high sensibility of the parameter models, as functions of \beta, to the available data and sampling associated errors. Presumably is also the root of the different proposals for asymptotic traits of \alpha\cdot\gamma(\beta)_{1,2}|_{\omega\rightarrow\infty} functions found in scientific literature.