Revealing Short- and Long-range Li-ion diffusion in Li$_2$MnO$_3$ from finite-temperature dynamical mean field theory

TL;DR

This work tackles the discrepancy between short-range and long-range Li diffusion barriers in the Li-excess cathode Li$_2$MnO$_3$ by introducing a finite-temperature DMFT framework that treats Mn $d$-electron correlations in the paramagnetic phase. By combining DFT+$U$, DMFT with a CTQMC impurity solver, and NEB calculations for a single Li vacancy, the study evaluates DMFT total energies along six diffusion paths. It finds that dynamical correlations reduce the two lowest barriers to $E_a=0.18$ eV (path (ii)) and $E_a=0.50$ eV (path (iv)); the 0.18 eV barrier matches local $bmu^+$SR measurements, while the 0.50 eV barrier aligns with macroscopic ac-impedance data, suggesting a percolating network that requires both intra- and interlayer hops. The results demonstrate the importance of dynamical electronic correlations for diffusion in correlated oxides and provide a framework for incorporating strong Coulomb interactions in future Li-rich transition-metal oxide battery materials.

Abstract

Li$_2$MnO$_3$ remains a crucial component of the Li-excess layered cathode family, $(1-x)\,\mathrm{LiMO_2} + x\,\mathrm{Li_2MnO_3}$ ($M$ = Mn, Ni, Co, \dots), but its role in limiting Li-ion mobility remains under debate. Here we combine DFT+$U$, finite-temperature DMFT with a continuous-time quantum Monte Carlo impurity solver, and nudged-elastic-band (NEB) calculations to investigate Li$^+$ migration for a single Li vacancy in paramagnetic Li$_2$MnO$_3$. Dynamical electronic correlations within DMFT substantially reduce the activation energies of the lowest-barrier pathways, yielding $E_a = 0.18$ eV for the shortest-range Li jump and $E_a = 0.50$ eV for the next-lowest pathway. The 0.18 eV barrier quantitatively reproduces the short-range activation energy extracted from $μ^+$SR measurements, whereas the 0.50 eV barrier is consistent with the long-range transport activation energy obtained from ac-impedance measurements. This single-vacancy, paramagnetic DMFT description therefore provides a coherent explanation of both local and macroscopic probes without requiring highly clustered vacancy configurations or strong extrinsic disorder, an assumption compatible with nearly stoichiometric Li$_2$MnO$_3$ powders. Our results highlight the importance of finite-temperature dynamical correlations for Li-ion migration in correlated oxides and provide a framework for incorporating strong Coulomb interactions in future studies of transition-metal oxide battery materials.

Figures (6)

• Figure 1: Atomic structure of Li$_2$MnO$_3$ with positions of Li ions.
• Figure 2: Electronic structure of Li$_2$MnO$_3$ with a single Li vacancy from DFT+$U$ and DMFT. (a,b) DFT+$U$ projected density of states (PDOS) onto Mn1 and Mn2 $d$ states. (c,d) DMFT PDOS projected onto the Mn-$d$--like correlated subspace on Mn1 and Mn2. Mn1 and Mn2 denote inequivalent Mn sites. We caution that in the downfolded DMFT setup the correlated orbitals are Mn-$d$--like Wannier functions that implicitly include Mn--O hybridization; therefore changes in the projected Mn-$d$ occupancy reflect the effective low-energy description and should not be interpreted as a direct oxidation-state assignment in the full $p$--$d$ manifold.
• Figure 3: Diffusion path of single Li vacancy in Li$_2$MnO$_3$. (a)-(d) intra layer difffusion. (e)-(f) inter layer diffusion.
• Figure 4: (a) Diffusion path (ii) of single Li vacancy in Li$_2$MnO$_3$, and (b) LiO$_4$ tetrahedral structure at the saddle point.
• Figure 5: (a)–(f) Migration barriers for a single Li vacancy in Li$_2$MnO$_3$ along pathways (i)–(vi), respectively.