Data analysis for “Mechanistic Origin of Superionic Lithium Diffusion in Anion-Disordered Li₆PS₅X Argyrodites”

Authors: - Benjamin J. Morgan ORCID 0000-0002-3056-8233

Summary

This repository contains a series of Jupyter notebooks used for the data analysis described in “Mechanistic Origin of Superionic Lithium Diffusion in Anion-Disordered Li₆PS₅X Argyrodites” (10.1021/acs.chemmater.0c03738).

The dataset contains raw simulation data (VASP XDATCAR files) from a series of ab initio molecular dynamics simulations on Li₆PS₅X argyrodites, and Jupyter notebooks containing code to process these data to produce the relevant manuscript figures.

The full DFT dataset, including all DFT inputs and outputs is available at the University of Bath Research Data Archive.

To run the complete data analysis, run the following commands from the top level directory: pip install -r requirements.txt snakemake clean snakemake --cores all

Contents:

• README.md: This file.
• requirements.txt: Python dependencies.
• Snakefile: The top-level snakemake analysis workflow script.
• system_info.txt: Information about the runtime environment (e.g. OS version and installed versions of Python packages) used to generate this dataset before depositing. here also provide code for plotting the relevent figures.
• data: Nested directories containing the raw molecular dynamics trajectory data (as VASP XDATCAR files).
• scripts: Auxiliary Python scripts.
• analysis: Nested directories containing Jupyter notebooks for various analyses of the molecular dynamics data.
• figures: Manuscript figures generated by running the analysis workflow.

Dependencies

The analysis workflow has the following Python module dependencies: - crystal-torture - jupyter - kinisi - matplotlib - numpy - polyhedral-analysis - pymatgen - pyyaml - site-analysis - snakemake - tqdm - version-information

These can be installed using pip via pip install -r requirements.txt

Packages can be installed using the “frozen” versions that were used for the as-published dataset via pip install -r frozen_requirements.txt

Note: Installation of crystal-torture requires a Fortran compiler is installed.

Full system information and Python environment used to generate the as-published analysis dataset is listed in system_info.txt.

Data processing workflow

Data processing is mostly managed using snakemake, and uses the Jupyter notebooks in the analysis directory, and the Python scripts in the scripts directory. The full data analysis workflow can be rerun with: snakemake --cores all clean snakemake --cores all