Research Digest — 2026-10-31¶
Halide Solid Electrolytes¶
1. Sulfur-Induced Lattice Disordering in Halide Electrolytes Enables Superionic Sodium-Ion Conduction¶
Source: ACS Nano (10.1021/acsnano.6c11189) · 📅 2026-10-29 · ↗ Open paper
A mixed-anion amorphous Na-ion electrolyte, TaCl5-0.5Na2S, made by high-energy ball milling reaches 2.9 mS/cm room-temperature conductivity with a low 0.276 eV activation energy. Sulfur incorporation creates a highly disordered network of Ta-centered bipolyhedra, and AIMD simulations identify continuous low-barrier Na+ diffusion pathways through the amorphous framework.
Relevance to DENG.Group
Directly extends our halide electrolyte work (Yan Li, Mengke Li) from Li to Na chemistry and shows how amorphization plus mixed anions beats crystalline halide transport limits. The AIMD-plus-topology analysis pipeline mirrors what we run for halide degradation and transport, and Na-halides are a natural expansion axis for the group's electrolyte portfolio.
2. Local Bonding and Reconstruction Enhancing Ion Transport in Li3InCl6¶
Source: Small (10.1002/smll.76542) · 📅 2026-10-29 · ↗ Open paper
This study links local bonding environments and surface reconstruction in Li3InCl6 to enhanced Li-ion transport, showing that local structural rearrangements rather than bulk composition alone govern the fast-ion behavior of this humidity-sensitive halide conductor.
Relevance to DENG.Group
Li3InCl6 keeps appearing in our halide transport discussions (Mengke Li's ion-transport mechanisms project); this paper's local-bonding perspective complements our MD analyses of how coordination environments set migration barriers in trivalent-metal chlorides.
3. Virtual electrode XPS experiments on Li3-xSc1-xZrxCl6 (x ~ 0.2): Oxygen-mediated interface passivation on metallic lithium anodes¶
Source: ChemRxiv (10.26434/chemrxiv.15011105/v1) · 📅 2026-10-27 · ↗ Open paper
Using virtual-electrode XPS simulations on Li3-xSc1-xZrxCl6, the authors show that trace oxygen mediates formation of a self-passivating interphase on metallic lithium, stabilizing the halide/Li interface. Zr substitution shifts the decomposition chemistry toward benign oxygen-containing products rather than resistive reduction layers.
Relevance to DENG.Group
Speaks directly to our halide degradation and interface-stability projects (Yan Li): the virtual-electrode XPS approach is a cheap way to predict which decomposition products actually form at halide/metal interfaces, and oxygen-mediated passivation is a hypothesis we can test in our own AIMD interfacial models.
Cluster Expansion & Configurational Energetics¶
4. First-Principles Identification of Higher-Order Configurational Energy Terms and Their Magnetic-State Dependence in Lithium-Rich Oxyfluorides: From QUBO to HUBO¶
Source: arXiv:2610.38321 · 📅 2026-10-28 · ↗ Open paper
Spin-polarized DFT energy maps of disordered-rocksalt Li2MnO2F and Li2Mn0.5Ti0.5O2F reveal substantial third-order (nonadditive) configurational interactions that quadratic cluster expansions (QUBO) cannot capture. Extending to cubic HUBO terms substantially reduces fitting errors, though the third-order contribution depends strongly on the converged magnetic state.
Relevance to DENG.Group
A caution flag for Tzu-Chen Liu's cluster-expansion work: truncating at pair terms can miss real three-body physics, and the demonstrated magnetic-state sensitivity matters for any CE we fit to disordered rocksalts or amorphous electrolytes with magnetic transition metals.
ML Interatomic Potentials¶
5. Distilling Large-Scale Machine Learning Interatomic Potentials to Predict Long-Timescale Electrolyte Transport Properties¶
Source: arXiv:2610.39513 · 📅 2026-10-29 · ↗ Open paper
A knowledge-distillation procedure transfers the accuracy of a large pretrained MLIP 'teacher' into fast system-specific 'student' models with no additional ab initio data, running ~40x faster and delivering stable 20 ns trajectories in days instead of months. Across Na- and Li-ion liquid electrolytes, student-predicted diffusivities, conductivities, and viscosities track experiment and beat classical force fields.
Relevance to DENG.Group
Directly useful for Yanhao Deng's ML-potential work and Naibing Wu's polymer electrolyte transport: distillation from foundation models is exactly the trick we need to converge ionic transport statistics without months of compute, and the no-new-DFT-data recipe makes it cheap to adopt for our electrolyte systems.
6. ML-COHP: predicting bond descriptors using latent space of a universal interatomic potentials¶
Source: arXiv:2610.40445 · 📅 2026-10-29 · ↗ Open paper
A symmetric pair head on frozen MACE foundation-model descriptors predicts ICOHP bond strengths to 1.6-4.1% error for bonds above 0.5 eV, reproducing LOBSTER results on relaxed grain-boundary cells to 0.06 eV in about a second - versus days of DFT plus projection for a hundred-atom cell.
Relevance to DENG.Group
Grain-boundary bond analysis without DFT round-trips is immediately attractive for Cheng Peng's GB models in solid electrolytes: we could screen bonding changes across many GB structures and chemistries at MLIP cost, and the head is architecture-agnostic enough to bolt onto our existing MACE-based potentials.
Polymer Electrolytes¶
7. Preferential Ion Solvation Drives Divergent Miscibility in Polyether Blend Electrolytes¶
Source: ACS Macro Letters (10.1021/acsmacrolett.6c00451) · 📅 2026-10-26 · ↗ Open paper
Coarse-grained MD of PEO/P(2EO-MO) and PEO/P(EO-MO) blends with LiTFSI reproduces experimental miscibility sequences: PEO/P(2EO-MO) phase-separates at low salt but remixes at high salt. RPA-evaluated Flory-Huggins parameters flip from critical to negative as preferential PEO solvation gives way to coordination saturation and ion bridging.
Relevance to DENG.Group
Right on target for Naibing Wu's polymer-electrolyte simulations: the salt-dependent chi parameter mechanism and the re-entrant mixing prediction give us a concrete design rule for blend electrolytes, and the CG-MD-plus-RPA workflow is directly transferable to our PEO-based systems.
MOF Adsorption & Transport¶
8. Elucidating Liquid-Phase Guest Transport Across MOF Single Crystals: Interplay of Adsorption, Desorption, and Intracrystalline Diffusion¶
Source: Angewandte Chemie International Edition (10.1002/anie.9654577) · 📅 2026-10-24 · ↗ Open paper
Tracking molecular transport across MOF single crystals in binary and ternary liquid mixtures, the authors find diffusion rates vary strongly even among guests with identical kinetic diameters. Slow transport is traced to strong guest-framework interactions rather than pore clogging, and diffusion speeds up when a more strongly interacting solvent weakens guest-framework binding.
Relevance to DENG.Group
Clean experimental grounding for Jason Li's MOF adsorption and gas-kinetics modelling: it separates adsorption, desorption, and intracrystalline diffusion contributions - the same decomposition our simulations compute - and suggests solvent-choice levers we could explore for tuning transport in the frameworks we study.