Programme

P.75 -- Posters, VCTP-51

Date: Thursday, 6 August 2026

Time: 09:30 - 10:30

Structural and thermal stability of pentagonal PtN$_2$/PdN$_2$ heterostructures: Insights from machine learning interatomic potentials trained on first-principles data

Mai Ngoc Qui (1, 2), Tran Thi Anh Thu (1), Nguyen Hai Dang (1, 3), Pham Thi Bich Thao (1), Le Huu Nghia (1), and Nguyen Thanh Tien (1, ∗)

(1) College of Natural Sciences, Can Tho University, 3-2 Road, Can Tho City 94000, Vietnam (2) Vinh Long University of Technology Education, Vinh Long City, Viet Nam (3) Faculty of basic, Nam Can Tho University, Can Tho City 900000, Vietnam

The integration of first-principles calculations (DFT), machine-learning interatomic potentials (MLIPs), and large-scale molecular dynamics (MD) simulations has emerged as a powerful approach for investigating the properties of advanced 2D materials. In this work, the monolayer of PdN$_2$ and PtN$_2$ are first based on DFT by electronic structures, phonon dispersions, and thermal stability. Based on the DFT-generated dataset, the MLIP models were assessed using the root mean square error (RMSE) of atomic energies and forces. The energy/force RMSE values were $8.71 \times 10^{-4}$ eV atom$^{-1}$ and $8.14 \times 10^{-2}$ eV \AA$^{-1}$ for PdN$_2$, $6.11 \times 10^{-4}$ eV atom$^{-1}$ and $4.35 \times 10^{-2}$ eV \AA$^{-1}$ for PtN$_2$, and $4.14 \times 10^{-4}$ eV atom$^{-1}$ and $4.02 \times 10^{-2}$ eV \AA$^{-1}$ for the PdN$_2$/PtN$_2$ heterostructure, respectively. The reliability of the trained potentials is further validated through a comparative analysis of phonon spectra obtained from DFT and PhonoLAMMPS calculations. Large-scale MD simulations are performed using NPT and NVE ensembles to optimize the structures and investigate their thermal behavior. The resulting radial distribution functions (RDFs) reveal a strong consistency with the equilibrium atomic configurations predicted by DFT. The proposed multiscale DFT–MLIP–MD approach opens new opportunities for the accelerated design and large-scale simulation of next-generation functional materials.

Presenter: Mai Ngoc Qui