Installation

This page explains how to install and set up SolvHybrid before running it.

Clone the repository:

git clone https://gitlab.com/lch_interfaces/solvhybrid
cd solvhybrid

Add SolvHybrid to your PATH (for example in .bashrc or .zshrc):

export PATH="$PATH:$HOME/solvhybrid/"

Python environment

To run the SolvHybrid workflow, you need Python 3.10+ and the dependencies listed in environment.yml. You can set up the environment using venv, Miniforge, Conda, or Mamba.

Example with Conda, using the provided environment.yml file:

conda env create -f environment.yml
conda activate SolvHybrid

External dependencies

Once the environment is created and activated, you can run SolvHybrid locally with AmberTools (without MPI). Note, however, that such runs may be slow depending on the system size. In addition, some Amber modules (e.g. MUSIC) are not distributed with AmberTools.

For best performance, we recommend installing the full Amber package with MPI support.

  • Amber (sander.MPI) or AmberTools (sander, tleap, cpptraj)

CM5 charges are computed by default using the Python implementation from the charges package. Alternatively, you can use the standalone cm5pac.exe program by specifying its directory in the input file. The executable is typically ~10% faster than the Python implementation.

  • cm5pac (optional, for CM5 charge correction)

To start a run you must also provide:

  • The structures of the isolated molecule, the clean surface, and the adsorbed system

  • Corresponding charge files (Hirshfeld charges), usually extracted from VASP OUTCAR files (CP2K support is planned)

Therefore, you will need access to a DFT engine:

  • DFT engine (VASP supported, CP2K experimental)

Note

For the structure, as it is read by ASE, you can use any format supported by ASE (e.g. POSCAR, CONTCAR, CIF, XYZ, etc.). But for charges, only VASP is supported at the moment.