Impedance-Resolved Beginning-of-Life Electrochemical Performance and Durability of Treated Carbon Felt Electrodes in Vanadium Redox Flow Batteries

Abstract

Vanadium redox flow batteries (VRFBs) have been targeted as a potential technology for integration of renewable energy sources onto grids as well as improving grid-level power generation efficiency. The primary reasons VRFBs are well-suited for this role are on account of their modularity, scalability, and relatively high performance [1]. There are still, however, many design parameters that can be more highly optimized to improve and specialize aspects of this technology. The kinetics at the anode of VRFBS utilizing graphite felt electrodes on are generally more limiting than at the cathode [2] as well as showing more rapid signs of performance degradation and are as such an area where sources of losses can be improved significantly.Electrochemical impedance spectroscopy is a commonly used method for characterizing sources of losses in VRFB systems. Impedance spectra are often analyzed in the context of the theory of Randles circuits, where electrochemical phenomena are treated analogously to circuit elements. Data reduction equations can then quantify these circuit elements and subsequently determine the values of physical parameters of interest in the system. The Randles circuit, however, has no geometric framework, nor any term which can capture the physical behavior of electrolyte resistance that is distributed throughout a porous medium. A model which was developed by Paasch and co-workers [3], referred to as the macrohomogeneous porous electrode (MHPE) model, addresses this phenomenon and has been used by others to characterize EIS data in systems with porous electrode where electrolyte resistance is non-negligible. Nguyen [4] et. al furthered this model in the context of inhomogeneous porous electrodes and membranes, Sun and co-workers [5] were the first to demonstrate the single MHPE model in the context of the all vanadium chemistry and Pezeshki and co-workers applied the model to quantify the effects of cell architecture in VRFBs [6].Much work on improving carbon felt-based electrodes in VRFBs does so by attempting to make charge transfer more facile [7]. As the charge transfer becomes smaller the effect of the distributed ohmic (pore) resistance begins to account for a larger portion of the cell losses and neglecting to account for it can misrepresent the series ohmic resistance (primarily membrane) and the charge transfer resistance.This work looks at the impact of using three different EIS models on different electrodes (seven treatments and one control) and the impact on the fitted parameters based on the utilization of a specific model. The first two models are both Randles circuits, which differ in the portions of the frequency spectrum which they are using. The first Randles circuit utilizes the full frequency spectrum referred to as RFS model (Randles Full Spectrum), the second Randles circuit truncates the spectrum, referred to as RTS model (Randles Truncated Spectrum), and excludes the distributed ohmic (pore) resistance, and third model is the aforementioned MHPE model. The electrode treatments include a nitric acid soak, heat treatment in 3 different gaseous environments (Oxygen, Nitrogen and Ammonia), and hydrothermal treatments in a nitric acid | sulfuric acid environment for 3 different durations. Electrodes were tested in-situ and cycled to characterize beginning-of-life stability.The results show that there are non-trivial tradeoffs between utilizing the different models. The serial ohmic resistance cannot be predicted accurately by either of the Randles circuit models (RFS and RTS) for all of the electrode treatments where it is overestimated as seen in Figure 1. Further, if the RFS model is used, it can yield kinetics parameters that differ from the MHPE model by over 50%. However, the RTS model is in relatively high agreement with the MHPE model for kinetics parameters and mass transfer parameters that can be extracted from EIS data. Over the course of cycling the electrodes treated in the ammonia atmosphere demonstrated both the highest stability and the lowest charge transfer resistance of all of the tested electrodes. Figure 1