Research Digest — 2026-08-01¶
Reviews & Roadmaps¶
1. Machine Learning Interatomic Potentials for Energy Materials¶
Source: Advanced Energy Materials · 10.1002/aenm.71046 · 📅 2026-07-15 · ↗ Open paper
Comprehensive review by Park et al. covering MLIP architectures and applications across solid-state electrolytes, battery electrodes, electrocatalysts, photovoltaics, and high-entropy alloys. Discusses data generation strategies, active learning workflows, and the transition from proof-of-concept to routine use in materials discovery pipelines.
Relevance to DENG.Group
Core reference for Yanhao Deng's ML interatomic potential work and for the group's broader ML strategy. Provides a landscape view that can inform tool selection and benchmark comparisons for the group's MTP/nequIP workflows.
ML Interatomic Potentials¶
2. AQVolt26: Dataset and MLIPs for lithium halide electrolyte discovery¶
Source: SandboxAQ (Industry Release) · 📅 2026-07-18 · ↗ Open paper
SandboxAQ released AQVolt26, containing 322,656 high-fidelity DFT calculations of lithium halide electrolytes at the r2SCAN level, along with trained MLIPs. The dataset targets the bottleneck of modeling high-temperature dynamics required for battery performance simulation. Trained on GCP and NVIDIA DGX H100 hardware.
Relevance to DENG.Group
Directly relevant to Yan Li and Mengke Li's halide electrolyte work. The dataset covers the same compositional space (Li-Y-Cl, Li-In-Cl systems) and could serve as a benchmark or supplementary training data for the group's own MLP development.
Halide Solid Electrolytes¶
3. Tuning collective anion motion enables superionic conductivity in halide solid electrolytes (world record 11 mS/cm)¶
Source: Nature Chemistry · 10.1038/s41557-024-01634-6 · 📅 2026-07-10 · ↗ Open paper
Mo group (UMD/GaTech/ORNL) achieved room-temperature ionic conductivity up to 11 mS/cm in mixed-anion halide electrolytes by strategically tuning collective anion motions. Synchrotron X-ray/neutron scattering combined with ab initio MD revealed that anion sublattice dynamics trigger the superionic transition. Lowering this transition temperature to room temperature is the key innovation.
Relevance to DENG.Group
Highly relevant to Yan Li and Mengke Li's halide electrolyte research. The anion-motion-tuning strategy offers a concrete design principle for improving conductivity in the group's halide systems. The combined experimental/computational approach also validates ab initio MD methodology the group uses.
Defects & Grain Boundaries¶
4. Lithiation-dependent solid electrolyte interphase formation in silicon/sulfide solid-state batteries¶
Source: Journal of Energy Storage · 10.1016/j.est.2025.116013 · 📅 2026-06-20 · ↗ Open paper
Zheng et al. use machine-learning-accelerated simulations to reveal atomic-scale mechanisms of SEI formation at Si/sulfide interfaces, showing that SEI composition and morphology depend strongly on the lithiation state of the silicon anode. The work bridges electrochemical cycling with interfacial decomposition chemistry.
Relevance to DENG.Group
Relevant to Umang Agarwal's heterogeneous interface work. The ML-accelerated interface characterization approach and the focus on composition-dependent interface stability are both applicable to the group's interface studies across different electrode/electrolyte combinations.
Phase Field / Dendrites¶
5. Numerical simulation of key factors affecting dendrite growth in solid-state electrolyte batteries¶
Source: Journal of Energy Storage · 10.1016/j.est.2026.1119364 · 📅 2026-07-18 · ↗ Open paper
Zhang et al. present a coupled multi-physics model for dendrite growth in solid-state electrolytes, systematically varying key parameters (critical current density, stack pressure, interfacial defects, electronic conductivity of the SE) to identify which factors dominate under different operating conditions.
Relevance to DENG.Group
Directly relevant to Shoutong Jin's phase-field dendrite work. The parametric study approach and the identified critical factors provide validation targets and design guidelines for the group's own dendrite simulations.