Research

My research combines periodic first-principles calculations, finite-temperature atomistic simulations, machine-learned interatomic potentials, and data-driven analysis to understand how atomic-scale structure controls transport, stability, and mechanics in functional materials. While many of my recent applications are in solid-state electrolytes, the computational framework is broader and transferable to other classes of crystalline materials.

I use periodic density functional theory and finite-temperature molecular dynamics to investigate atomic structure, bonding, transport, and stability in crystalline materials. A major part of my work has focused on solid-state ion conductors, where static crystal structures alone are often insufficient and finite-temperature effects play a central role.

My work in this area includes structure optimisation, electronic-structure analysis, NpT Car-Parrinello molecular dynamics, and dynamic bonding analysis using Wannier90. I also use trajectory-based analysis to extract transport and correlation metrics from atomistic simulations.

Representative paper

Cooperative Transport of Lithium in Disordered Li₁₀MP₂S₁₂ (M = Sn, Si) Electrolytes for Li-Ion Batteries

Chemistry of Materials (2024) 

To extend simulations beyond the time and length scales accessible to direct ab initio molecular dynamics, I am developing and deploying machine-learned interatomic potentials trained on first-principles reference data. This work is aimed at preserving close contact with DFT-level physics while enabling larger systems, longer trajectories, and broader statistical sampling.

My current efforts include MLIP development using MACE, Allegro, and MTP, construction of training datasets from DFT reference calculations, validation against energies and forces, and deployment in LAMMPS for extended NpT and NVT simulations.

A central theme of my work is that promising functional materials must be evaluated not only for transport, but also for thermodynamic stability and mechanical response. I therefore combine first-principles thermodynamics with elastic and vibrational analysis to develop a more complete understanding of material performance.

This includes convex-hull and grand-potential phase diagrams, electrochemical stability analysis, elastic tensors, directional anisotropy, and vibrational behaviour linked to anharmonicity and superionic transport.

Representative paper
Beyond LGPS: Superionic Conduction Meets Chemomechanical Anisotropy in Li₁₁AlP₂S₁₂ Solid-State Li-Ion Batteries
Acta Materialia (2026)

I am interested in machine learning not only as a predictive tool, but also as a way to extract physically meaningful trends from materials datasets. In this area, I use interpretable machine-learning methods to identify which structural or defect-related features most strongly influence observed properties.

My work includes predictive modelling using XGBoost and neural-network approaches, SHAP-based interpretation of feature importance, and data-driven analysis of structure-property relationships in simulation-generated datasets.

Representative paper
Machine Learning Predicted Inelasticity in Defective Two-Dimensional Transition Metal Dichalcogenides Using SHAP Analysis
Physical Chemistry Chemical Physics (2024)

Beyond bulk crystalline materials, I have also contributed to DFT studies of interfacial mechanics in layered systems. This work broadens my research into adhesion, shear response, and interfacial property prediction.

Representative paper
Interfacial Shear Strength of MXene Interfaces
Cell Reports Physical Science (2025)

My ongoing work includes the development of ML interatomic potentials for atomistic simulations and a manuscript under review on anharmonicity-correlated liquid-like lithium transport and electrochemical stability in halide solid electrolytes.

Current manuscript
Anharmonicity-Correlated Liquid-Like Lithium Transport and Electrochemical Stability in Halide Solid Electrolytes
Under review at PRX Energy