A Paradigm Shift in Anode-Free Lithium Metal Battery: Pressure-Activated Solid-State Interfaces for High-Rate Desolvated Cation Diffusion

Anode-free lithium metal batteries (AF-LMBs) promise ultrahigh gravimetric/volumetric energy densities (> 400 Wh kg⁻¹/1000 Wh L⁻¹) and simplified anode manufacturing as compared to conventional alkali-ion batteries that rely on intercalation chemistry. However, their practical implementation remains plagued by dendritic protrusion from the substrate and rapid lithium inventory depletion, which further exacerbate in Ah-scale pouch cells. Here, this study innovates a paradigm shift through a multiscale interfacial strategy addressing the core limitations of AF-LMBs. Scalable cation-exchange and mechanical exfoliation firstly produce few-layer lithium montmorillonite nanosheets that integrated with polyacrylamide gel, which are functionalized onto the polyethylene separator (FMT-Li/PAM-PE). The composite separator thus reconciles high mechanical strength (204.4 MPa), thermal stability (< 2.5% shrinkage at 180°C), anion screening capability (t₊ = 0.78), and pressure-activated adhesion to the substrate (peel strength > 3.4 N m⁻¹ via hydrogen bonding). Upon the formation cycle at stack pressure of 0.5 MPa, more crucially, the composite separator intimately attaches onto the Cu substrate modified with the recycled spent graphite rich in lithiophilic defects (SGR-Cu), establishing the solid-state Li⁺ diffusion pathway at the separator-anode interface and mitigating solvated Li⁺ interaction. As assembled with a densely-packed LiNi_0.8Co_0.1Mn_0.1O₂ (3.6 mAh cm⁻²) cathode, the 1.0 Ah pouch cell achieves 81.1% capacity retention over 200 cycles, gravimetric/volumetric energy densities of 453.3 Wh kg⁻¹/ 1183.2 Wh L⁻¹ and extreme power output of 1045.0 W kg⁻¹. Beyond insights into multiscale ion regulation, this interfacial strategy also unlocks viability across diverse cell configurations (e.g., LiFePO₄/Ni92||Cu), enabling the high-rate cation diffusion for the commercial AF-LMB prototyping.