Electrochemical porosimetry: Deconvolution of distribution functions

Abstract

Discrete Fourier transform (DFT) was used to deconvolute the distribution of penetrability coefficient (ao-distribution, equivalent to pore size distribution) from electrochemical impedance data of porous electrodes. The working equation is the Fredholm integral equation of the first kind to correlate macroscopic impedance data to a theoretical model describing microscopic signal with the ao-distribution. Simulated and experimental impedance data were tested. Noise observed at high frequencies in Fourier space was removed before inversely Fourier-transforming the ao-distribution from Fourier space to real space. The accuracy of ao-distributions deconvoluted by DFT depended on the number, frequency range and quality of impedance data. The examples in this work showed that fairly accurate ao-distributions could be obtained by DFT deconvolution. Most promising method was to use the ao-distribution obtained from DFT deconvolution as the first guess to shape the true ao-distribution. Then, accurate ao-distributions could be obtained by estimating parameters of the pre-assumed distributions by using complex nonlinear least square fitting.