Application of finite Gaussian process distribution of relaxation times on SOFC electrodes

Electrochemical impedance spectroscopy (EIS) is a powerful tool in characterisation of processes electrochemical systems, allowing us to elucidate the resistance and characteristic frequency physical properties such as reaction transport rates. The essence EIS relationship between current potential at given frequency. However, it often case that we do not understand system well enough fit meaningful model data. distribution relaxation times (DRT) calculation assumes an infinite series distributed over timescale. DRT may identify number occurring, their respective resistivity timescale, resolve which have relatively similar timescales. Using nonparametric known Gaussian process (GP) regression, showcase method finding unique solution ill-posed problem by optimising kernel hyperparameters opposed ad-hoc regularisation. In this work, use finite GP regression under inequality constraints (fGP) analysed data generated (Ni/CGO|CGO|YSZ|Reference Cathode) solid-oxide fuel cell gas mixture 0.5 bar H2/0.5 H2O temperature 600 °C. By varying density, can characterise current-voltage electrode shed light on mechanism governing charge transfer solid-gas interface. Our findings also show even high densities (±600mAcm-2) limited transfer.