Revealing Li-ion Stage reaction in Graphite via Machine-Learning Interatomic Potential and Genetic Algorithm

Revealing Li-ion Stage reaction in Graphite via Machine-Learning Interatomic Potential and Genetic AlgorithmFiles are composed of these directories:DFTGA-MLIPDFTThis directory contains the structural models used for DFT calculations, provided in CIF format.Ordering_modelSOC25, SOC50, SOC75 : Contains structural models representing specific ordering types (Off-facing, Alternating, Facing) for each SOC level.SOC100 : Contains structures with distinct space groups designed to vary the facing degree of Li ions.Stacking_modelContains the ground-state structural models for the stacking transition discussed in the manuscript.Global-Lithiation : CIF files representing the global lithiation model.Local-Lithiation : CIF files representing the local lithiation model.Includes sub-directories: soc100_fix , soc75_fix , soc50_fixGA-MLIPThis directory includes the source code, input files, machine learning potential, and environment configurations required to reproduce the GA-MLIP simulation.Root Filesrun_pipeline.py : The main execution script for running the GA-MLIP workflow.input.cif : The initial input structure file used for the simulation.sevenet_model.pth : The fine-tunned machine learning interatomic potential (SevenNet) model file used in this study.GA-MLIP.yaml : The Conda environment file included to ensure reproducibility of the computational environment.Sub-directoriesconfigparams.json : Manages simulation parameters, defining the hyperparameters and settings for the GA process.operatorsContains scripts for genetic operators, including structure generation (random) and variation (mutation) algorithms.srcHouses the main source code for each step of the genetic algorithm pipeline.utilityA collection of utility scripts and helper functions required for code execution.resultsBest Structures : Contains the final optimized structures (best CIFs) for each SOC level derived from the GA-MLIP simulations.