Research Digest — 2026-09-09¶
Halide & Sulfide Electrolytes¶
1. Non-blocking grain boundaries for Li-ion transport in Li3YCl6 halide electrolyte¶
Source: Applied Physics Letters · 📅 2026-09-07 · ↗ Open paper
Machine-learning-driven deep potential molecular dynamics is used to quantify grain-boundary contributions to Li-ion transport in Li3YCl6. All GBs examined show low formation energies (an inherent tendency toward polycrystallinity) and reduced conductivity relative to bulk, but the penalties are modest: the worst GB activation energy is 0.246 eV and the lowest-GB-energy {3032}<1013> boundary reaches 13.9 mS/cm. GBs hinder transport yet do not act as the dominant barrier in this halide.
Relevance to DENG.Group
Cheng Peng — this is almost exactly his thesis topic (GBs in solid electrolytes) done with the group's own toolchain (deep potential MD); a direct benchmark for his GB conductivity calculations and a design datapoint for Yan Li and Mengke Li's halide transport/degradation work. Yanhao Deng can compare their DP-gen training strategy against the group's MLIPs.
2. Electrochemical and Spectroscopic Evaluation of Li7Si2S7I Solid Electrolyte in All-Solid-State Batteries¶
Source: ACS Applied Energy Materials · 📅 2026-08-28 · ↗ Open paper
Three-electrode ASSB cells with impedance plus ex-situ Raman and XPS decouple cathode-side and anode-side degradation of the silicon-based sulfide Li7Si2S7I in NMC82 | SE | Li-In cells. The measurements resolve concurrent interfacial decomposition at both electrodes under practical operation, clarifying which interface limits the cell.
Relevance to DENG.Group
Yan Li and Mengke Li — provides exactly the kind of interface-resolved degradation phenomenology (decomposition products, which electrode drives failure) that their simulation of halide/sulfide electrolyte degradation should reproduce and rationalize; the Zeier-group characterization protocols are a useful reference for designing validation experiments.
3. Particle Size and Composition Coupling Determine the Interfacial Stability in Sulfide All-Solid-State Battery Composite Cathodes¶
Source: ACS Applied Energy Materials · 📅 2026-09-05 · ↗ Open paper
Three-dimensional LAMMPS simulations plus multiscale characterization probe how Li6PS5Cl particle size and NCM90/SSE ratio control interfacial transport in composite cathodes. Refining LPSC improves void filling, lowers Li-ion tortuosity and boosts ionic continuity, but simultaneously increases electronic-path tortuosity — a 50:50 CAM:SSE volume ratio balances the two percolation networks.
Relevance to DENG.Group
Umang Agarwal — a concrete example of microstructure-resolved simulation of electrolyte/electrode contacts, complementing his interface work with composite-cathode-scale thinking; the ionic-vs-electronic tortuosity trade-off is also relevant context for the group's dendrite/transport simulations.
Composite Polymer Electrolytes & Interphases¶
4. Residual alkali effect in LLZO-modified composite solid polymer electrolyte¶
Source: Energy Materials · 📅 2026-09-03 · ↗ Open paper
Ga-doped LLZO nanoparticles prepared by wet vs dry routes — carrying abundant vs minimal surface residual alkali (e.g. Li2CO3) — are incorporated into PEO-based composite solid polymer electrolytes. Surface residual alkali is shown to disrupt the ordered arrangement of PEO chains, and this restructuring, rather than LLZO conductivity alone, governs the conductivity gains of the composite.
Relevance to DENG.Group
Naibing Wu — directly on his polymer/composite simulation theme: the finding that filler surface chemistry controls polymer-chain ordering (and hence transport) is a concrete target for his PEO/oxide-filler interface simulations and for interpreting filler-selection rules.
5. Mapping Solid Electrolyte Interphase Evolution and Lithium Deposition at Solid-State Interfaces by Operando Nano-focus WAXS¶
Source: ACS Energy Letters · 📅 2026-09-03 · ↗ Open paper
An anion-trapping layered double hydroxide (LDH) additive is embedded in PEO-based composite polymer electrolytes, yielding 950 h stable Li plating and 90.1% retention over 250 cycles in LiFePO4||Li cells. Operando nano-focus wide-angle X-ray scattering spatially maps SEI species and Li dendrite distributions at the Li/electrolyte interface across micron-scale lateral and vertical dimensions, linking regulated Li+ transport to suppressed dendrite penetration.
Relevance to DENG.Group
Naibing Wu (composite-polymer interphase chemistry) and Shoutong Jin — the operando dendrite/SEI maps are the kind of quantitative growth data his phase-field dendrite model can be calibrated and validated against; Umang Agarwal can mine the Li/electrolyte interphase evolution for interface-stability context.
Grain Boundary Science¶
6. Complexity of disconnection dynamics during grain boundary migration¶
Source: npj Computational Materials · 📅 2026-08-29 · ↗ Open paper
Using Interfacial Line Defect Analysis on MD of a Sigma5(210)[001] tilt boundary in Al, migration is resolved into a reaction-dense network of transient disconnection activity rather than smooth glide of straight steps. A 'flickering' mechanism — small loops repeatedly nucleating, fluctuating and annihilating — drives boundary translation, and shear coupling emerges from instantaneous defect densities.
Relevance to DENG.Group
Cheng Peng — a state-of-the-art template for analyzing GB migration and defect networks atomistically; the ILDA methodology and disconnection picture transfer directly to studying GB structure/motion in solid electrolytes, where GB-mediated transport and mechanical response are his core questions.
ML Interatomic Potentials¶
7. Hessian-based molecular conformation augmentation for a scalable and efficient strategy of machine learning interatomic potentials¶
Source: arXiv · 📅 2026-09-04 · ↗ Open paper
Introduces a data-augmentation strategy for MLIP training that uses Hessian information to generate enhanced molecular conformations, improving the model's fidelity to the potential energy surface beyond energy/force fitting alone. The approach targets applications such as vibrational analysis and transition-state search where curvature quality matters, with better sample efficiency than brute-force expansion of training sets.
Relevance to DENG.Group
Yanhao Deng — a practical training-data strategy for his MLIP development: curvature-aware augmentation could improve transferability of the group's potentials for solid electrolytes while cutting DFT labeling cost, which matters for the large configuration spaces his electrode/electrolyte models must cover.
8. AdaptNTK: Adaptive Uncertainty Quantification and Active Learning for Neural Network Potentials¶
Source: arXiv · 📅 2026-08-31 · ↗ Open paper
Proposes an adaptive uncertainty-quantification framework with active learning for neural network potentials, iteratively expanding training sets by identifying configurations where the model is least reliable. The method aims to make first-principles-accuracy MD more dependable in extrapolative regimes without exhaustive DFT sampling.
Relevance to DENG.Group
Yanhao Deng — directly applicable to his active-learning loops for electrolyte/electrode potentials: better UQ means fewer wasted DFT single-points and safer sampling of reactive interfacial events (decomposition, charge transfer) where the group's potentials are most likely to extrapolate.