Proton exchange membrane water electrolysis (PEMWE) holds promise as an efficient and responsive technique for green hydrogen generation, vital for de-carbonization in diverse applications spanning from transportation, energy storage, to steel production, etc. Electrochemical impedance spectroscopy (EIS), as a non-invasive diagnostic tool, is increasingly being employed to differentiate processes and optimize performance of PEMWE.[1] This talk gives an overview of the status, features, problems and potentials of EIS in the study of PEMWE.Currently, the application of EIS in PEMWE remains in its early stages both in quantity and quality. There are less than 200 related journal papers, dwarfed by that in lithium-ion batteries (~700) and proton exchange membrane fuel cell (PEMFC, ~2300). Interpretations of the impedance spectra largely depend on visual inspection or fitting with empirical equivalent electric circuit models (EECM). Investigations using numerical and analytical frequency-domain model based on the physical processes in PEMWE are rather rare.Typical features in a spectrum for the anode and cathode catalyst in a 3-electrode setup generally include a depressed semi-circle, which is attributed to the charge transfer at high frequency end and surface adsorption at low frequency end in HER/OER reaction.[2-4] Occasionally, a small semi-circle at high frequency range may be observed, yet the origin is unclear.[5] In contrast, typical features in a spectrum for the catalyst layer in a 2-electrode setup include a straight line segment at high frequency range, a depressed semi-circle at medium frequency range and an arc at low frequency range. The straight line segment is a signature of the porous electrode. The depressed semi-circle corresponds to kinetic process. The arc at low frequency is related to mass transfer process. The characteristic frequency corresponding to surface adsorption may overlap with that of mass transfer in catalyst layer, and the arc at low frequency range is generally attributed to mass transfer rather than surface adsorption. [9-10] At high current density, the dynamic generation of bubbles may impair the quality of the EIS data or even invalidate the technique altogether.[11] Sometimes a small arc attributed to contact resistance/capacitance may be detected at the medium to high frequency range.Despite its easiness of use, EIS still grapple with problems or controversies in its application in PEMWE. Firstly, different from PEMFC, where electronic resistance in the catalyst layer is negligible, the electron resistance in the catalyst layer of PEMWE is comparable to the proton resistance. [12-13] Therefore, the intercept of the 45° line segment at high frequency range with the real axis under blocking or non-blocking conditions is affected by the electron resistance of the catalyst layer. Accordingly, analytical EIS model taking into account of both ionic and electronic resistivity should be used in the analysis.[14] Secondly, the combination of high electronic resistivity, the small thickness of the catalyst layer, the pore size jump between the catalyst layer and the porous transport layer (PTL), and the large deformation of the CL, causes significant 2D or 3D effects. Such effects are not yet considered in either analytical models or EECM.[15] Thirdly, while some groups reported inductive loops at low frequency range (0.1~10 Hz) with good repeatability for multiple types of catalyst and measurement setup[16-17], the majority of the impedance spectra in literature didn’t show such behavior. The conditions for the emergence and the underlying mechanism for such low frequency inductive loop remain largely unexplored.Lastly, there is a lacking of guideline for reporting EIS data and presenting the spectrum. Some normalization of the raw data in Ohm with apparent or effective area, catalyst loadings, flow rate per active area, etc., is needed to facilitate the understanding and comparison among the EIS data in literature. [9,10,17-23] The potential of EIS application in PEMWE can be further realized by upgrading its practice into the next stages. In experiment, measures to suppress the disturbance of bubbles are needed to extend the frequency range compliant with Kramer-Kronig transformation. Replacing PTL with gas diffusion layer (GDL) used in fuel cell may help to validate the power and limitation of the 1D analytical model that has considered the electronic resistance. In post-processing, the number and values of characteristic frequencies may be identified with the further development of distribution of relaxation time (DRT) method. In modeling, analytical and numerical EIS models taking into account the low electronic resistivity and 2D as well as 3D catalyst layer structure is needed to reveal the mechanism for the subtle features in the spectra.
Read full abstract