Electrolyte System Modeling and Classical MD Workflow

This tutorial provides a step-by-step guide for modeling electrolyte solutions and running classical molecular dynamics (MD) simulations. It is aimed at beginners and covers environment setup, required software installation, molecular modeling, simulation script preparation, job submission, and data analysis. Please follow each step closely and refer to documentation or community resources when needed.

Environment Setup and Software Installation
Detailed Modeling Workflow
Classical Molecular Dynamics (MD) Simulations
Machine Learning Molecular Dynamics (MLMD) Workflow
Data Analysis and Processing
FAQ and Useful Resources

Environment Setup and Software Installation

1. Set Up Conda Environment

Module Anaconda/Miniconda:
```
module load miniconda/3
```
- see detail from the official site and follow our group wiki.

Create and activate a Python environment:

conda create --name py311 python=3.11
conda activate py311

Install required packages:

conda install -c rdkit rdkit
conda install -c conda-forge openbabel
pip install ai2-kit
# If network issues occur, change sources in ~/.condarc to Tsinghua mirrors, e.g.:
# channels:
# - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
# - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
# - conda-forge
# - defaults
# ssl_verify: true

2. Other Essential Software

Download and install: packmol, gnuplot, travis
Tip: Always read each software's README before installing.

Detailed Modeling Workflow

1. Prepare Script Tools

Copy auxiliary scripts to your working directory:
- convertLigParGen.py You can read this blog from Xu longkun which gives a detailed example on how to obtain the OPLS-AA force field parameters and the LAMMPS settings for MD simulations.
- fftool.py For the details about using fftool to generate LAMMPS input file and data file, you can check this fftool page of Prof. Agilio Padua, which contains very detailed information.
```
cp /path/convertLigParGen.py ./
cp /path/fftool.py ./
```
For better organization, create directories for scripts and define aliases in your shell:
```
alias convertLigParGen='python /path/to/convertLigParGen/convertLigParGen.py'
alias fftool='python /path/to/fftool/fftool.py'
```
Remember to add these to ~/.bashrc and run source ~/.bashrc to apply.

2. Molecular Modeling and Job Submission

2.1 Module load singularity for using LigParGen

Module load singularity
```
module load singularity/3.9.2
```

check out the output file path

singularity exec /data/share/images/apps/ligpargen.sif pwd

2.2 Generate Molecular Force Field

Example for water molecule:
```
singularity exec /data/share/images/apps/ligpargen.sif ligpargen -s 'O' -n H2O -p H2O -c 0 -o 3 -cgen CM1A
```
- -s: SMILES input
- -n: Output file prefix
- -p: Output folder
- -c: charge
- -o: optimization
- -cgen: charge type
SMILES resources:
- OPLS-AA for LAMMPS & SMILES explanation see detail from the official doc.
- Draw structure and convert to SMILES (Marvin JS) try to use this tool from the official site.
  
  Note: Marvin JS requires JavaScript enabled in your browser.

2.2 Generate Required Files

Check script parameters:
```
convertLigParGen -h
```

Generate .xyz and .ff files:

convertLigParGen -g H2O.gmx.itp -l H2O.lammps.lmp -o H2O.ff

2.3 Build a Box Filled with Water Molecules

Check fftool help:
```
fftool -h
```

Fill box:

fftool 128 H2O.xyz -b 15

This generates pack.inp for packmol:
```
packmol < pack.inp
```

Generate LAMMPS input files:
```
fftool 128 H2O.xyz -b 15 -l
```
- Produces input.lammps and data.lmp

2.4 Submit Simulation Job

Copy job script template and ensure the -in argument matches your input script:
```
cp /data/share/base/scripts/lammps.slurm ./
sbatch lammps.slurm
```

available script

#!/bin/bash
#SBATCH -J lammps
#SBATCH -p c51-medium
#SBATCH -N 1
#SBATCH --ntasks-per-node=24
#SBATCH --exclusive

# add modulefiles
module load lammps/23Jun2022_update1_cpu
module load intel/17.5.239 mpi/intel/2017.5.239 gcc/7.4.0
module load netcdf/4.9.0_intel17
module load plumed/2.8.0

mpiexec.hydra  -env I_MPI_EXTRA_FILESYSTEM on -env I_MPI_EXTRA_FILESYSTEM_LIST gpfs \
  lmp_mpi -in input.lammps

Useful commands:

squeue -u username     # View jobs
scancel JOBID          # Cancel job
tail -f log.lammps     # Monitor log output

Classical Molecular Dynamics (MD) Simulations

Refer to the LAMMPS Manual for parameter settings and input file formats.
LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) is an open-source, highly extensible classical MD engine.
- Learn about NVT/NPT ensembles and other relevant commands.
- If you encounter errors, check common LAMMPS errors and troubleshooting.

Machine Learning Molecular Dynamics (MLMD) Workflow

1. MLMD Script Overview

Upload and extract the mlmd package to your root directory.
Key scripts:
- run.sh: Automates several data extraction and conversion steps
- dump2xyz.py: Converts .lammpstrj to .xyz
- dump2avedata.py: Selects the best energy-converged frame and generates a LAMMPS input
- sub.deepmd: Job submission script

2. Operation Steps

Edit run.sh to match your file names
Copy mlmd to your classical MD folder
Execute:
```
./run.sh
./dump2xyz.py
```
Obtain input_npt.lammps, conf.lmp, simbox.xyz, etc.
Submit the NPT job:
```
sbatch sub.deepmd
```

3. NVT Calculation

After NPT, process the data:
```
./dump2avedata.py
```
Create a new nvt folder in mlmd, copy the average density output file (preferably rename as nvt.lmp), and ensure all necessary files (sub.deepmd, nvt.lmp, input_nvt.lammps, force field files) are present and consistently named.
Submit the NVT job:
```
sbatch sub.deepmd
```
You can terminate the simulation early if energies are converged (check with plot tool). Use VMD to visualize trajectory files.
- About VMD:
  VMD (Visual Molecular Dynamics) is a free, cross-platform program for displaying, animating, and analyzing large biomolecular systems using 3D graphics and scripting. It supports MacOS, Unix, and Windows and is commonly used for trajectory visualization and analysis.

Data Analysis and Processing

1. Install and Use Travis

After installing TravisCI, load the gcc module:

module load gcc
ln -s name.lammpstrj name.lmp
travis -p -name.lmp

Choose central atom and calculation type per your research needs (e.g., RDF). Use plot to visualize and check the quality of your data.

FAQ and Useful Resources

Encountering Errors:
Use Google, Computational Chemistry forums, and official documentation for troubleshooting.
For LAMMPS-specific errors, consult the LAMMPS Error Guide.
More Resources:

Notes

Make good use of bash aliases and scripts to accelerate batch operations.
Always maintain organized file/script naming and archiving habits for reproducibility and collaboration.
If in doubt at any step, consult the relevant README or Wiki, or ask more experienced colleagues.

This tutorial is subject to updates. Contributions and suggestions are welcome!

Siriusitzy/Battery Material Group Tutorial 1.md