Machine Learning-Driven Alloy Anode Accompanied by Interfacial Kinetic Compensation Design Toward High-Rate Zinc Pouch Cells.

Developing practical aqueous zinc metal batteries is crucial for safe grid energy storage. However, zinc anode imposes severe interfacial mass transfer and kinetic bottlenecks at high rates and high capacities, driven by thermodynamically competitive hydrogen evolution reaction and significant polarization. In this work, we establish a predictive design paradigm that resolves this dilemma by integrating active‑learning materials screening and a composition‑gradient alloy (GL) anode with an interfacial kinetic compensation. Moreover, the gradient grain boundaries preclude structural failure under high capacities arising from stress concentration at high rates. Consequently, GL anode achieves 20,000 cycles at 40 mA cm-2 and the cumulative capacity of 81.36 Ah at 80 mA cm-2. The symmetric cell endures more than 400 h at 80% depth of discharge and 20 mA cm-2, while pouch cells exceed 1000 h at 5 mAh cm-2. Coupled with high-loading vanadium-based cathode (30 mg cm-2, 4.5 mAh cm-2), the full cell retains 91% capacity after 1500 cycles. Notably, a 560 mAh pouch cell stable for 350 cycles over 90% capacity retention at a high rate of 20 mA cm-2. This work offers a design concept of practical zinc anode with high-rate stability for zinc pouch cells.