Machine learning-assisted kinetic matching model for rational electrode design in aqueous zinc-ion batteries

Aqueous zinc-ion batteries offer inherent safety and low cost, yet performance is limited by unstable zinc metal negative electrodes and dissolution-prone positive electrodes, causing dendrite growth, sluggish ion transport, and rapid capacity decay. Replacing both electrodes with intercalation hosts provides a solution, but progress is slowed by the lack of a universal principle for selecting kinetically compatible pairs. Most existing efforts optimize single components rather than addressing the electrodes’ kinetic mismatch governing full-cell stability. Here we show a machine-learning-assisted kinetic-matching framework that quantitatively evaluates ion-transport compatibility in intercalation-type zinc-ion batteries electrodes. By correlating interlayer spacing with Zn2+ diffusion behavior, the model introduces two descriptors predicting synchronized ion flux for rational electrode pairing. Using this framework, an optimized Zn3V3O8 | |NH4V4O10 system achieves a specific capacity of 310 mAh g-1 and retains over 12,000 cycles at 5 A g-1. The strategy further extends to deformable formats through conductive hydrogel architectures, enabling omnidirectionally stretchable, all-hydrogel zinc-ion batteries with an areal capacity of 1.2 mAh cm-2 and an energy density of 1070 μWh cm-2. These results provide a quantitative design route for next-generation zinc-ion batteries. Aqueous zinc-ion batteries are safe and affordable but limited by incompatible electrode kinetics. Here, authors present a machine learning framework that resolves this mismatch, enabling the rational design of durable and stretchable zinc-ion batteries.