td

Parser for excited state calculations done with Gaussian 09 (td keyword), TURBOMOLE (escf and ricc2 module) and ORCA calculations.

The script tries to determine the program type from the supplied log file. For TURBOMOLE calculations it expects escf.out or ricc2.out file names.

The script uses a slightly modified version of Sergey Astanin's tabulate module (https://bitbucket.org/astanin/python-tabulate) and peakdetect from Sixten Bergman. Thanks to them.

Installation

Clone this repository with git clone https://github.com/eljost/td.git and run python setup.py install in the cloned directory.

The argcomplete module (https://pypi.python.org/pypi/argcomplete) has to be configured separately. When using argcomplete td looks for files with .out and .log extensions.

Usage

Display the help message with all available commands:

td -h

Common usage examples are the generation of broadened spectra from calculated excitation energies and oscillator strenghts according to http://dev.gaussian.com/uvvisplot/ and this paper https://dx.doi.org/10.1002/chir.20733. Spectra can be generated in two ways: Not normalized (ε in l mol⁻¹ cm⁻¹) or normalized with the brightest peak set to 1 (ε/ε_max). The spectrum is printed to STDOUT.

Easy plotting of UV absorption spectra

td.py [log] --plot [nm or ev] [--peaks] [--plotalso [additional logs to plot]]

Verbose MO names

Verbose MO names can optionally be loaded from a mos.json file with the following format:

{
	"[MO number] [MO irrep] : "[verbose name]",
	e.g.,
	"61 a'" : "dπ₁",
	"39 a\"" : "dπ₂"
	
	! The last entry must NOT have a comma at the end
}

The mos.json file must reside next to the parsed log file.

Spectrum generation

td exports the spectrum ε(x) for x in two different units: nm and eV. The first two blocks hold the spectrum and the oscillator strength impulses in nm, the third and fourth blocks hold the same data in eV. Within gnuplot the data blocks can be easily accessed by with the index [id] command.

Normalized spectrum with (ε/ε~max~) on the ordinate:

./td [fn] --spectrum [from in nm] [to in nm]  > [outfn]

Non-normalized spectrum with ε on the ordinate:

./td [fn] --spectrum [from in nm] [to in nm] --nnorm > [outfn]

Oscillator strength scale on the ordiante

When used with the argument --e2f the molecular extinction coefficients on the ordinate will be converted to an oscillator strength scale.

./td [fn] --spectrum [from in nm] [to in nm] --e2f --nnorm > [outfn]

Dealing with different multiplicities

A constant shift in a.u. can be added to the excitation energies with --enoffset. This may be useful when one has a calculation with triplet-triplet excitation energies and wants to relate these energies to the corresponding singlet groundstate energy. Additionally all oscillator strengths can be zeroed with --zeroosc.

Filtering

To investigate a Gaussian 09 excited state optimization it may be useful to split the output in chunks, where chunks should correspond to the number of calculated roots at every step of the optimization:

./td [fn] --chunks [roots]

Only show transitions with an oscillator strength greater than or equal to a supplied threshold and sort by oscillator strength:

./td [fn] --fthresh [thresh] --sf

Exporting

Several export-formats are available:

| Format | Argument | Comment | | --------- | ------------- | ------------- | | booktabs | --booktabs | To be used in .tex-documents (doesn't export transitions and weights yet).| | raw | --raw | Export without any formatting. | | docx | --docx | Export to an .docx-document. Needs python-docx-module. | | tiddlywiki | --tiddly | Export as a tiddlywiki-table. | | csv | --csv | Export as CSV. |