Quantifying the error of the core-valence separation approximation

Implementation of the Rayleigh-Quotient-based CVS relaxation discussed in

Michael F. Herbst, Thomas Fransson
Quantifying the error of the core-valence separation approximation
Preprint on arxiv

This repository contains the relaxation code, data and example scripts for reproducing the results discussed in the paper. The implementation makes use of adcc version 0.14.2. For more details see also this blog article.

Table of contents

• relaxation.py: Script containing the QR-based CVS relaxation procedure
• ammonia.py, fluoro_ethene.py, methane.py, water.py, water_singles.py: A few examples how to run CVS relaxations. See below for details how to install requirements and run these scripts.
• basis: Basis set definitions for the modified 6-311++G** basis sets used in the study (p6-311++G**, u6-311++G**, up6-311++G**)
• geometries: xyz-files containing the MP2-optimised geometries, which were used as starting points. Length unit in the files is Ångström.
• data: Various files containing raw data for figures and tables.

Running the code

Required is at least python 3.6. The installation process further requires pipenv (Installing pipenv).

To install all dependencies and run the example scripts just type:

pipenv install  # Install dependencies into separate python environment
pipenv shell    # Get a shell in this environment

./water.py      # Run example scripts