Molecular Statistics (MST) Source Codes
Description
These files contain the source codes for the Molecular Statistics Framework. Abstract: This study presents Molecular Statistics (MST): a novel computational methodology designed to simulate thermodynamic properties of solid materials from room temperature to melting point with high efficiency. MST combines principles of statistical mechanics with molecular potentials, employing position distribution functions to generate atomic oscillations. Unlike conventional methods, MST enables fully independent and parallel calculations on small atomic sub-assemblies, reducing the computational load by orders of magnitude relative to molecular dynamics (MD) and Monte Carlo (MC) simulations. This approach bypasses the need for thermodynamic integration, directly sampling the Helmholtz free energy, which further minimizes computational expense. To validate MST, we benchmark its performance against state-of-the-art methods, including nonequilibrium molecular dynamics (NEMD), and the local harmonic approximation (LHA), a rapid yet less precise method. The evaluation demonstrates MST’s accuracy in predicting key thermodynamic properties, such as Gibbs free energy, entropy, and heat capacity, for several transition metals, aligning closely with experimental data. Furthermore, MST achieves these calculations within seconds, competitive with NEMD in both runtime and accuracy, underscoring its potential as an efficient alternative for high-temperature thermodynamic simulations. The MST framework thus emerges as a promising tool for materials science applications, particularly for complex systems where traditional methods become computationally prohibitive. Please cite the following article, when using the codes: Vincent Feyen, Nele Moelans, Molecular statistics: A novel computationally efficient modeling approach to simulate thermodynamic properties of solid materials up to the melting point, Computational Materials Science, Volume 248, 2025, 113558, ISSN 0927-0256, https://doi.org/10.1016/j.commatsci.2024.113558. (https://www.sciencedirect.com/science/article/pii/S0927025624007791)
Files
Steps to reproduce
- Add all folders to the MATLAB path - Open MST_script - Choose potential type (MEAM or FS) (InputStruct.potentialType = "FS"; %options: "MEAM", "FS") - Choose which atoms to simulate (nrAtomsStable = 1:1:12; %options: 1:1:12, these correspond to the different atoms: Fe,V,Mo,Nb,Ta,W,Cu,Ag,Au,Ni,Pd,Pt) - Run script (saving and plotting will occur automatically, just plotting will also work since the simulated data for all graphs of the paper is available in DATA_FILES) NOTE: see the requirements in the README file
Institutions
Categories
Funding
Research Foundation - Flanders
1SA6319N