This open source python package MDIOSVCOR (Molecular Dynamics IOn SolVation CORrections) provides routines for calculating the correction terms for solvation parameters and partial molar variables of ion solvation calculated with molecular dynamics using the lattice summation treatment of electrostatic interactions.

For installation instructions see below

The program is used for calculating the correction terms applied in the publication:

Björn Dahlgren, Maria M. Reif, Philippe H. Hünenberger & Niels Hansen "Calculation of derivative thermodynamic hydration and aqueous partial molar properties of ions based on atomistic simulations", J. Chem. Theory Comput., Just Accepted Manuscript DOI: 10.1021/ct300260q Publication Date (Web): June 17, 2012

The program was implemented for use in a research project in the IGC group at ETH Zürich. Unfortunately this program which uses the very useful sympy project is a late rewrite, i.e. the numbers presented in the paper are the same but they were calculated using non analytical derivatives, for this reason and that this version was finished after having submitted the manuscript means that SymPy is not cited in this work. Any future developments in this area (which will use this codebase) will naturally contain the proper citations. The theory behind the correction terms were developed by several persons in the IGC group (see references in the article above)

If you make use the code in your own research, please cite the article referenced above and appropriate references therein.

This package is written in python, i.e. no compilation is required. However, in order to speed up calculations the default behavior of the script is to compile functions via FORTRAN to binaries. This is done on the fly the first time the program needs the binaries.

Therefore you just need to execute "calc_correction_terms.py" (or "calc_water_eps.py"/"calc_water_rho.py") with appropriate flags. E.g:

 ./calc_correction_terms.py -T 298.15 -P 101.3e3 -N 1024 -Y G -c all -i sod

(for further information on invocation see help by executing e.g. ./calc_correction_terms.py --help)

This package relies on:

• Python 2.7.x
• Sympy 0.7.1 to treat algebraic expressions symbolically (Ubuntu package name: python-sympy)

Furthermore, for full functionality, additional requirements apply:

• f2py (provided by NumPy)
• a FORTRAN compiler (if compiled versions are wanted)

To calculate a large number of correction terms takes a rather long while due to the symbolic treatment of the expressions and use of explicit units on the numerical parameters. Therefore it might be of interest to run a batch job and store the output in an easily parsed format.

An example on how to run this kind of batch calculation for large number of different parameters is given below:

 python -c "import calc_correction_terms as cct; cct.batch_calc(['G','H'], [101.3e3], [273.15, 298.15, 323.15], [1024, 512], ['sod','cls'],['B','C1','C2','D'])"

To generate correction terms corresponding to conditions used in the article one can (under Bash shell) execute:

 MEMOIZE_CACHE_DIR=cache/ PYTHONPATH=$PYTHONPATH:/home/dahlbjo/sympy python -c "import calc_correction_terms as cct; cct.calc_all_for_project()" > all.txt

Beyond the core equations used in calculating the correction terms specified in the article above, this project re-implements the IAPWS95 formulation of thermodynamic properties of ordinary water substance. It also uses a sophisticated method the retrieve the density of water, and its temperature and/or pressure derivative of arbitrary order with known numerical precision. This is however due to the large number of non-linear terms computationally expensive for higher order derivatives since the manipulations are done symbolically up to the last step of numerical solution of the resulting implicit relation.

Cached compiled binaries are stored in cache/ directory. When moving between hosts of different architecture/OS one must purge this directory of "*.so" files.

Feel free to write them and make a pull request at github:

• Implement expressions not only for LS but also for CT
• Optimize performance

Any comments and/or improvements of the code are greatly appreciated. Improvements are best done by making a pull request at github It is also the place to raise issues and filing bug reports.

Latest version is available at github.com/bjodah/mdiosvcor

• Author of python package: Björn Ingvar Dahlgren
• Email (@gmail.com): bjodah

This work is open source and is released under the 2-clause BSD license (see LICENSE.txt for further information)

Copyright (c) 2011, 2012 Björn Ingvar Dahlgren