Topological Control of Transition Metal Networks for Reversible High-Capacity Li-rich Cathodes

Abstract

Developing high-energy-density batteries is essential for advancing sustainable energy technologies. However, leading cathode materials such as Li-rich oxides, including Li$_2$MnO$_3$, suffer from capacity loss due to irreversible oxygen release and structural degradation, both consequences of the oxygen redox activity that also enables their high capacity. The atomic-scale mechanisms behind this degradation, and whether it can be made reversible, remain open questions. Here, using submicrosecond-scale molecular dynamics simulations with first-principles accuracy, we directly visualize the entire charge-discharge cycle of Li$_2$MnO$_3$, uncovering the full lifecycle of the O$_2$-filled nanovoids responsible for degradation and identifying the critical size limit for voids to remain fully repairable upon discharge. Our results reveal that the topology of the Mn cation network is the key factor governing void growth, coalescence, and reparability. Based on a structural topology-informed design principle, we computationally develop a novel Li$_2$MnO$_3$ structure featuring a Mn lattice with a Kagome-like pattern, demonstrating full electrochemical reversibility even under extreme 80% delithiation. Our work establishes a new paradigm for designing high-energy cathodes, shifting the focus from mitigating damage to engineering inherent stability through atomic-level topological control of transition metal network.

Figures (5)

• Figure 1: Molecular dynamics simulation of the charging process. (a) Schematic illustration of the simulation setup, where Li atoms (Li$^+$ + e$^-$) are sequentially removed from a designated region of the Li layers within the supercell. Electroneutrality is maintained by assuming that Li$^+$ removal occurs instantaneously relative to the slower timescale of structural relaxation. (b) Simulated Li occupancies in the Li and Mn layers (lines) at various states of delithiation show excellent agreement with experimental results (solid circles) from Li et al. Liu16p1502143. (c) Time evolution of 4-fold coordinated Li over a 10 ns MD trajectory. Delithiation induces a transition from 6-fold to 4-fold coordination, forming a dumbbell structural motif (inset).
• Figure 2: Atomistic mechanism of nanovoid nucleation and growth in the 40% residual Li structure. (a) Snapshots from MD simulations showing nanovoids (yellow isosurfaces) with trapped O$_2$ molecules distributed heterogeneously throughout the structure, taken after 10 ns of simulation at 500 K. (b) Time evolution of key structural metrics. The size and number of nanovoids initially increase simultaneously, accompanied by increased Mn disorder (CN$_{\text{Mn–Mn}} > 3$) and rising O$_2$ content. The rapid increase in void size and the simultaneous decrease in void number around 1.4 ns are caused by void coalescence. (c) Schematic illustration of the self-catalytic cycle for O$_2$ formation and the nanovoid nucleation mechanism derived from MD simulations. (d) Mn atomic network showing nanovoid growth driven by interlayer Mn migration. The initial Mn framework adopts a honeycomb lattice, with atoms colored by their out-of-plane displacement ($z$) relative to the layer. Mn atoms with negative $z$ values (colored in red) indicate ongoing interlayer diffusion, which drives void growth.
• Figure 3: Nanovoid size and connectivity at various delithiation levels. (a) O$_2$ content and Mn disorder versus residual Li content. Both increase sharply as Li content decreases, indicating structural destabilization. (b–d) Nanovoids evolution with delithiation. Nanovoids increase in size as Li content decreases. At 30% Li, they coalesce into a continuous diffusion channel spanning the simulation cell with periodic boundary condition (PBC). (e) Root-mean-square displacements (RMSD) of the 20 most mobile oxygen atoms at different residual Li contents.
• Figure 4: MD simulations of full charge-discharge cycles and void repair behavior at different charge states. (a) Charge--discharge simulation protocol. Schematic of the full-cycle simulation using DPMD. Structures with 50%, 40%, and 35% residual Li were equilibrated and used as starting points for discharge simulations, during which Li atoms were progressively reinserted into a designated region of the supercell. (b) Post-discharge void distributions. At 50% residual Li, small voids ($<1$ nm) were fully repaired. At 35% residual Li, larger voids ($>1$ nm) remained partially unrepaired, correlating with irreversible capacity loss in (a). (c) Complete void repair at 50% residual Li. A small $\approx$1 nm void fully heals during discharge. Mn ions near the void wall assist in stabilizing LiO$_2$ intermediates. Those exhibiting significant out-of-plane displacements are highlighted in gold. (d) Incomplete void repair at 35% residual Li. A large $\approx$2.5 nm void contracts to $\approx$1.5 nm but remains unrepaired. The void core falls outside the catalytic range of Mn, preventing further redox activity and trapping O$_2$.
• Figure 5: High reversible capacity and structural stability of Kagome-like Li$_2$MnO$_3$. (a) Mn network transformation via topological engineering. The conventional honeycomb Mn lattice is reconfigured into a Kagome-like framework composed of triangular Mn clusters. This topology introduces a regular array of $\approx$1 nm voids in the fully delithiated state. (b) Near-perfect reversibility of Kagome-like Li$_2$MnO$_3$ under 80% delithiation. The inset shows O$_2$ molecules remain trapped within pre-designed voids. (c) RMSD analysis of the 20 most mobile oxygen atoms in conventional and Kagome-like electrodes at 20% residual Li.