Coupling Lattice Distortion and Cation Disorder to Control Li-ion Transport in Cation-Disordered Rocksalt Oxides

Cation-disordered solids offer a rich chemical landscape where local coordination, lattice responses, and configurational disorder collectively, yet often implicitly, govern ion transport. In cation-disordered rocksalt oxides, Li+ diffusion has conventionally been rationalized by the static 0-transition-metal (0-TM) percolation rule, which assumes an ideal, passive lattice and thus fails to capture experimentally accessible capacities. Here, we show that lattice distortion is an essential, previously overlooked degree of freedom that actively reshapes Li+ percolation networks. By developing a lattice-responsive framework combining Monte Carlo sampling of cation configurations with machine-learning-accelerated molecular dynamics, we quantitatively predict Li+ percolation and electrochemical capacities within 5% of experiment. Our results reveal a causal coupling between lattice distortion and cation short-range order: enhanced local distortions precede and suppress short-range ordering, activating Li+ migration through nominally inaccessible 1-TM channels, fundamentally extending percolation beyond the 0-TM paradigm. Guided by this, we design and synthesize a high-entropy oxide, Li1.2Mn0.2Ti0.2V0.2Mo0.2O2, which exhibits enhanced distortion and achieves a 71.9% Li+ percolation network, surpassing 65.8% in Li1.2Mn0.4Ti0.4O2, delivering 256.3 mAh/g capacity, closely matching our prediction of 255.1 mAh/g. These findings establish lattice distortion as an active control parameter for ion transport, revising percolation concepts and offering a general design principle beyond metal-ion cathodes.

💡 Research Summary

In this work the authors overturn the long‑standing view that lithium transport in cation‑disordered rocksalt (DRX) oxides is governed solely by the static 0‑transition‑metal (0‑TM) percolation rule. By combining Monte Carlo sampling of cation configurations with machine‑learning‑accelerated molecular dynamics (ML‑MD) based on a fine‑tuned CHGNet potential, they construct a “lattice‑responsive” computational framework that simultaneously captures configurational disorder and local lattice relaxations. The key structural descriptor they introduce is the height of the tetrahedral cluster that forms the transition state for O‑T‑O Li migration; this height quantifies lattice distortion arising from size and electronegativity mismatches among Li⁺ and the various transition‑metal (TM) cations.

Density‑functional calculations of migration barriers reveal that while 0‑TM channels remain low‑energy (≈0.30 eV) and relatively insensitive to distortion, 1‑TM channels exhibit a strong dependence on tetrahedral height. When the height exceeds critical values (2.476 Å for Mn, 2.501 Å for Ti, 2.522 Å for Zr), the barrier drops below 0.4 eV, making these channels effectively conductive. By feeding these height‑dependent barriers into the MC‑MD simulations, the authors map the evolution of the Li‑percolation network in realistic, thermally equilibrated structures.

Applying the method to Li₁.₂Mn₀.₄Zr₀.₄O₂ (LMZO) and Li₁.₂Mn₀.₄Ti₀.₄O₂ (LMTO) demonstrates the failure of the pure 0‑TM rule: LMZO would have zero percolation, LMTO only 43 % Li‑connectivity. When lattice‑distortion‑activated 1‑TM pathways are included, the percolation fractions rise to 41.6 % (LMZO) and 65.2 % (LMTO), and the predicted capacities (135.5 mAh g⁻¹ and 257.5 mAh g⁻¹) match experimental values (141.3 mAh g⁻¹ and 260 mAh g⁻¹) within 5 % error. The analysis also shows that the larger ionic radius mismatch in LMTO (Ti⁴⁺ vs Li⁺) produces greater TM displacement (0.20 Å → 0.27 Å) and thus stronger distortion.

Guided by this mechanistic insight, the authors employ a high‑entropy design strategy, screening 289 special quasi‑random structures of Li₁.₂Mn₀.₂Ti₀.₂M_I₀.₂M_II₀.₂O₂ with various TM pairs. The composition Li₁.₂Mn₀.₂Ti₀.₂V₀.₂Mo₀.₂O₂ (LMTVMO) emerges as optimal: it exhibits the lowest calculated mixing temperature (≈1712 K), the highest average TM displacement (0.27 Å), and activates the largest number of 1‑TM channels, yielding a percolation fraction of 71.6 % in simulation. Experimentally, LMTVMO delivers 71.9 % percolation and a reversible capacity of 256.3 mAh g⁻¹, essentially identical to the predicted 255.1 mAh g⁻¹.

The study therefore establishes lattice distortion as an active chemical degree of freedom that couples to short‑range order, suppresses ordering, and unlocks otherwise blocked diffusion pathways. This expands the percolation landscape beyond the conventional 0‑TM paradigm and provides a general, quantitatively validated design principle for disordered ion conductors. The authors suggest that the same lattice‑responsive MC‑MD framework can be extended to other mobile ions (Na⁺, Mg²⁺, etc.) and to solid‑state electrolytes, opening new avenues for high‑performance, earth‑abundant battery materials.