Sean McGrath's data and resources for a variety of quantum chemistry simulations performed with Dr. Scott Auerbach.

The files are organized in the following way:

jaguar contains data associated with simulations performed with Schrodinger's Jaguar simulation package. Use of this package is not recommended for future work due to Jaguar's poor support for higher-order basis sets and general instability.

gaussian contains input and output files for a number of simulations conducted in Gaussian, as well as files generated from analysis of the output.

scripts contains several bash tools for parsing Gaussian log files and generating plots from them. These functions are to be integrated into a cohesive command-line tool in the near future.

.mae: A Project for the Maestro simulation manager. Not recommended for use.

.vib: IR vibrational analysis output of Jaguar.

.spm: Raman vibrational analysis ouput of Jaguar.

.in: Input file for Jaguar simulations.

.out: Output file for Jaguar simulations.

.job: A computational job to be submitted to Volta or another queue-managed system via the qsub command.

.com: A gaussian input file.

.log: For Jaguar, a relatively useless log of simulation detailes. For Gaussian, the main output file, which is fed into various parsing scripts to extract vibrational data and distance matrices.

.csv: The output of an analysis script; if it has "distance" in the title, it is probably an atomic distance matrix; otherwise, it is probably raman spectral data.

.gnumeric/.xls: a spreadsheet, probably used to calculate the RMS deviation of simulated Raman frequencies from experimental values.

Gaussian simulation jobs are represented by .com files, and can be submitted directly to Gaussian via the gaussian command or equivalent wherever the software is installed.

However, Gaussian jobs are best executed on the Volta supercomputer, and this is slightly more complicated. A .job file must be created for submission to the Volta' batch server; this can be done be copying gaussian/dft.job and replacing the .com file it references with the one you would like to submit. Ensure that both of these files are copied to volta and placed in a proper working directory (scp is useful for this) then simply execute qsub <your_file>.com.

If everything goes smoothly, Gaussian will generate a .log file in the working directory and no .e file, which documents errors.

Gaussian simulations should be carried out at the B3LYP level of density-functional theory, using a 6-31+G(2DF,P) basis set. If a dissolved system is being simulated, the SMD solvation model seems to work nicely. Jaguar does NOT support this basis set for Raman analysis.

Until a comprehensive tool is implemented, the included scripts are the best way to extract information from Gaussian output.

matrixparse takes a single Gaussian .log file as its argument and outputs a comma-separated table representing the final atomic distance matrix of the simulated molecule. Table entries at position (i,j) represent the distance in angstroms between atom i and atom j. Consult the logfile itself for the atomic number scheme.

matrixRMS takes two matrices created by matrixparse and calculates the RMS deviation between them as a way of quantifying the degree to which two molecular structures differ.

ramanparse takes in a single Gaussian .log file which contains Raman data and outputs a comma-separated list of frequency-intensity pairs, with column headers.

lorPlot takes in the outout of ramanparse and constructs a Lorentzian fit to the frequency data. It outputs a plot of this fit as well. To run this, you will need the Python packages matplotlib, seaborn and their associated dependencies.