Salt-ion crossover through bipolar membranes (BPMs) is a major source of capacity fade and efficiency loss in acid-base flow batteries (ABFBs), yet its mechanisms remain poorly understood. Here, we quantify Na+ and Cl− crossover in a five-cell ABFB stack under varying electrolyte composition, state of charge (SoC), current density, and temperature. The results show that crossover occurs predominantly through BPMs and is governed by diffusion assisted by H+/OH− neutralization. At ambient conditions, Cl− crossover exceeds Na+ by roughly twofold, with apparent activation energies of 15 kJ mol−1 (Cl−) and 33 kJ mol−1 (Na+), reflecting asymmetric co-ion diffusion within the cation- and anion-exchange layers. Increasing current density reduces the relative contribution of crossover, whereas higher electrolyte concentration, SoC, and temperature increase it. Based on the combined voltage efficiency and crossover analysis, we recommend operating ABFBs under conditions corresponding to BPM efficiencies above 80% - that is, with dilute electrolytes (≤0.25 M), near-equimolar acid and base (SoC ≈ 50%), and cycling current densities exceeding 5 mA cm−2. These insights clarify the interplay between salt-ion transport and BPM operation, and provide design guidelines for next-generation ABFBs and other BPM-based electrochemical systems.

The use of gas diffusion electrodes that supply gaseous CO₂ directly to the catalyst layer has greatly improved the performance of electrochemical CO₂ conversion. However, reports of high current densities and Faradaic efficiencies primarily come from small lab scale electrolysers. Such electrolysers typically have a geometric area of 5 cm², while an industrial electrolyser would require an area closer to 1 m². The difference in scales means that many limitations that manifest only for larger electrolysers are not captured in lab scale setups. We develop a 2D computational model of both a lab scale and upscaled CO₂ electrolyser to determine performance limitations at larger scales and how they compare to the performance limitations observed at the lab scale. We find that for the same current density larger electrolysers exhibit much greater reaction and local environment inhomogeneity. Increasing catalyst layer pH and widening concentration boundary layers of the KHCO₃ buffer in the electrolyte channel lead to higher activation overpotential and increased parasitic loss of reactant CO₂ to the electrolyte solution. We show that a variable catalyst loading along the direction of the flow channel may improve the economics of a large scale CO₂ electrolyser.