Drivers of Membrane Fouling in the Non-Aqueous Vanadium Acetylacetonate Redox Flow Battery

Abstract

Due to their unique advantages in terms of long cycle life and modular design, redox flow batteries (RFBs) are being developed to load-level and peak shave energy from intermittent sources such as wind and sunlight. RFBs have yet to meet the stringent cost requirements needed to achieve broad market penetration, however. The second largest cost associated with state-of-the art RFBs after the electrolyte itself is the ion-exchange membrane, which accounts for 30% of the total cell cost [1]. Porous separators provide a pathway to substantially decrease the cost of RFBs, but long-term performance has been difficult to assess due to the coupling of electrolyte degradation, self-discharge, and membrane fouling. This study applies an adaptive observer model to estimate crossover from the open-circuit voltage relaxation of RFB cells, providing detailed information that can be used to decouple membrane fouling from electrolyte degradation by side reactions.For this study the nonaqueous V(acac)3 (vanadium acetylacetonate) electrolyte chemistry [2] is used. Because the RFB cell reaction is based on an electrochemical disproportionation of the neutral V(acac)3 complex, crossover of the active species in this system only results in reversible self-discharge.A novel set of 'canary cell' experiments unequivocally point to membrane fouling as the predominant factor that drives permanent cell degradation. An adaptive observer [3] is used to monitor the crossover rate, which is a proxy for membrane porosity, during self-discharge experiments where the cell is fully charged and the voltage is monitored during an open-circuit relaxation. Figure 1 shows that the porous Daramic separator starts off with ideal Fickian behavior because crossover is a linear function of the state of charge. Over time, however, the cell deviates, with active-species crossover becoming non-Fickian as the membrane fouls. The mass transfer coefficient (determined by the slope of the curve) decreases with exposure to current. It is hypothesized that pore clogging is responsible for this decreasing mass transfer coefficient, and in turn the diminution of cell performance. Deconvoluting membrane and electrolyte processes is a crucial step to determine the stability and long-term viability of RFB electrolyte systems.