First-Principles Thermal Transport

Jul 12, 2023 · 1 min read

Overview

Heat conduction at the nanoscale is governed by phonons, and predicting it from first principles remains computationally demanding. This project develops and applies state-of-the-art methods to overcome these bottlenecks.

Key Contributions

Symmetry-adapted approach: Exploiting the line group symmetry of quasi-1D systems (carbon nanotubes, nanowires) to dramatically reduce the computational cost of thermal conductivity calculations via the Green-Kubo formalism.
ML-accelerated MD: Benchmarking and applying machine-learned interatomic potentials (MLIPs) to enable long-timescale molecular dynamics for accurate thermal conductivity prediction.
Defect effects: Systematic study of how intrinsic defects alter heat transport in 1D nanomaterials, bridging theory and experimental observations.

Ab-initio heat transport in defect-laden quasi-1D systems — npj Computational Materials (2026)
Accelerating first-principles MD thermal conductivity calculations — J. Chem. Theory Comput. (2025)
Accelerating Green-Kubo heat transport for quasi-1D systems using ML force fields — PSI-K 2025