pyArtaios

Introduction

pyArtaios is a simple python interface to control ARTAIOS calculations. Without any knowledge of how to start quantum chemical calculations and to perform the post-processing using Artaios, it yields properties as the transmissions function for a given structure of a junction and a handfull of settings.

# import module
import pyartaios as pa

# define settings
settings = {'basis set': 'lanl2dz', 'functional': 'b3lyp'}

# read structure
natoms, atoms, coords = pa.read_xyz('file.xyz')

# initialize class object
transport = artaios(natoms, atoms, coords, settings)

# perform calculations
transport.calculate()

# get results
transmission = transport.get('transmission')

Requirements

pyArtaios needs ARTAIOS installed on your system, and it needs to be able to find the executables. Since it can start Gaussian and Turbomole calculations by itself, it needs to find those as well. The way these things are hardcoded right now is far from perfect, but will change in the future. Here is a list of executables, pyArtaios tries to execute:

For Calculations involving Turbomole:

| ridft / dscf | TM single point calculations, path to executable can be specified with settings['path tm']. Right now, it tries to execute ridft_smp or dscf_omp. Adjust this to your liking in pyArtaios._singlepoint(). | | tm2molden | part of Turbomole, to extract a molden file for calculations of subsystems |

For Calculations involving Gaussian:

| g09 | Gaussian executable |

This is messy, so it is advised to change the code according to your system. Future developments will include more flexibel and appropriate ways of linking executables

python Libraries

Following python libraries are needed:

numpy
subprocess
cclib (for for analyzing output files from calculations)
os
matplotlib (for the interactive plotting of the local transmissions some fancy stuff is needed, like some imports from matplotlib.widges and mpl_toolkits. Make sure you have an up-to-date version of matplotlib)

Adding the library to the python path

There are two ways to make the pyArtaios library findable for python on your system. The first way is to add it to the $PYTHONPATH variable in the .bashrc by adding the line export PYTHONPATH=/home/<username>/<path to your local git>. But, and this is crucial: We have to be careful with the file and functionnames! If somebody creates a file 'numpy', this would overwrite all numpy functionalities, since now two python modules exist with that name.

The second way is to add the link at the beginning of each script by using

import sys
sys.path.append(<path to pyartaios library>)

Settings

The default settings are set in the _default_settings function of the pyartaios class and are overwritten by the settings supplied by the user. The full list of settings can be found there, too. The keywords which cover most of the calculations are given below.

keyword	default	description
basis set	'lanl2dz'	basis set for the quantum chemical calculation
charge	0	charge of the molecule/system
elow	-8	lower energy limit for the transmission function in eV
estep	400	number of steps of the transmissions function
eupp	-8	lower energy limit for the transmission function in eV
fermi	-5	Fermi energy used for transport calculation in eV
functional	'bp86'	exchange-correlation functional
method	'dft'	method for quantum chemical calculation
multi	1	multiplicity
n cores	1	number of cores for calculation
nosymmetry	False	Use of "nosymmetry" keyword in Gaussian
qcprog	'gaussian'	'gaussian' or 'turbomole'
ri	False	switch for density fitting

The settings are organized as a dictionary, as shown in the example above.

Functions

With the initialization of the pyartaios class object, all necessary checks will be performed and the consistency of the chosen settings will be checked. If problems occur, the class will raise an error, so the user has to adjust the settings accordingly.

If the settings are not consistent, the pyartaios.calculate function will not execute. This function performs all necessary calculations in a subfolder called artaios. Existing folders with the same name will be moved to artaios_backup_<i>.

The available properties and results of the calculations can bet obtained with pyartaios.get(<keyword>). The keywords can be shown with transport.results.keys(), which yields the keys of the internal "results" dictionary.

List of Functions

The function which are needed for all basic calculations are already given in the example above. This is a full list of all function of the pyArtaios class. Under normal circumstances, they should not be used to perform calculations.

function name	description
`calculate`	performs all necessary calcualtions
`get`	getter function for results
`plot_local`	plots the local transmissions
`_artaios_transport`	performs the artaios transport calculation
`_check`	performs consistency check
`_default_settings`	sets default settings
`_get_subsys_energies`	routine to perform the subsystem analysis with artaios
`_get_local_transmission`	get local transmissions from artaios calculation
`_get_mo_energies`	get mo energies from single point calculation
`_singlepoint`	performs single point calculation
`__init__`	initializes pyArtaios class

List of Helperfunctions

pyArtaios contains a bunch of scripts to facilitate the usage of python, see below for a list.

function name	description
`read_xyz`	reads an xyz file

List of accessible variables

Even though for the results a getter function exists (pyartaios.get('keyword')), the necessity of adjusting a variable of the class can arise

variable name	description
natoms	number of atoms
atoms	list of atom symbols
coords	numpy array with cartesian coordinates
left	list of atoms for left electrode
central	list of atoms for central region
right	list of atoms for right electrode
results	dictionary with results

molspintron/pyArtaios