correct settings for T vs rho LJ equation of state

As a run-up to the work I’d like to do, I want to be able to reproduce some of the results for basic L-J fluid properties (T vs rho, etc.) at a NIST site:

https://www.nist.gov/mml/csd/chemical-informatics-research-group/lennard-jones-fluid-properties

Can someone point me to some sample scripts both for NVE and NVT Molecular Dynamics for this system? I’ve looked at many papers that address this topic, but have not found one that provides enough information for me to create the relevant scripts.

this is a two part problem:

  1. you need to learn how to use the different features in LAMMPS, see what they do, monitor their progress, and analyze the results
  2. you need to set up calculations to meet the descriptions in publications

for 1) there are the basic instructions and explanations in the LAMMPS manual (starting from the “Running LAMMPS” section) and there are the various tutorials given on running LAMMPS. Some are posted on the lammps.sandia.gov website. The material from the introductory (self-study) tutorial done by the LAMMPS developers at the last LAMMPS workshop in 2019 can be found at tutorial.lammps.org.

Please keep in mind that LAMMPS inputs are a bit like a program for running modeling calculations and processed in a stepwise fashion. That is often a major stumbling block for people new to LAMMPS.

  1. should not be a problem, after you have spent enough time with 1) and do not skip over important steps. remember, like for every tool that allows you to do complex tasks, it pays in the long run to get the basics right instead of skipping ahead.

axel.

Issue resolved. All of the scripts used to generate the NIST NVT simulations with LAMMPS can be found here:

https://github.com/dwsideriusNIST/LAMMPS_Examples

It’s been helpful to inspect those line-by-line as a launching point for further work.

If that is of interest to anyone here, I did the "MD NVE" example for the
Lennard-Jones fluid some time ago, without the need for an initial configuration
and with analysis code in Python. I wrote the steps in a makefile for
reproducibility.

blog article: http://pdebuyl.be/blog/2014/reproducible-lennard-jones-simulations.html

code: https://github.com/pdebuyl/ljrr

(some changes are needed due to the inclusion of H5MD in lammps).

Axel's comments remain 100% valid anyway.

Regards,

Pierre