CFOUR | Main / Plotting Stick Spectra For Vibrational Frequency Calculations

After a calculation of the vibrational frequencies, you will find the file irspectrum_harmonic.txt in the working directory; if you have done a VPT2 calculation, you will also find a file called irspectrum.txt. Both of these offer a visual representation of the calculated infrared spectrum, and can be plotted with gnuplot as follows:

[stanton@hbar cubic]$ gnuplot

(various garbage printed here by gnuplot)

Terminal type set to 'x11'
gnuplot> plot 'irspectrum_harmonic.txt' with lines

which of course can be greatly elaborated upon (and combined with laboratory spectra) if you have a working knowledge of the gnuplot plotting package that is available with essentially all installations of unix/linux/etc.

The filenames are self-explanatory: the irspectrum_harmonic.txt contains the harmonic frequencies for the fundamental vibrations and the associated intensities (in wavenumbers and km/mol, respectively), while irspectrum.txt contains the corresponding numbers from VPT2 calculations. In the latter case, when strong resonances (closely-lying levels differing in an odd number of excitation (Fermi resonances), quite peculiar-looking spectra can be generated; use of the xguinea module and doing a deperturbation-diagonalization analysis is usually required to obtain sensible results.

For irspectrum.txt, the fundamental and two-quantum transition intensities are written to the second and third records on each line, respectively. Hence, to differentiate these, you can use the gnuplot command

gnuplot> plot 'irspectrum.txt' using 1:2 with lines, 'irspectrum.txt' using 1:3 with lines

which will produce the spectrum with color-coded fundamentals and combinations/overtones.

The VPT2 analysis in CFOUR is limited to one- and two-quantum levels for intensities; hence using this graphical display of data will reveal nothing about higher-quantum levels. Again, xguinea can be used for three-quantum intensities, but VPT2 cannot be considered to be reliable for such things.

Finally, for the VPT2 output, and for gnuplot aficionados, there is a set of "labels" for each of the peaks in the stick spectrum. To see these (they are stored in the file irspectrum.labels), just invoke the command

gnuplot> load 'irspectrum.labels'
gnuplot> replot

and you should see labels associated with each stick peak. You can use the "set xrange" command to look at just a narrow region of the spectrum, where this is most useful. Looking at too broad a range of spectrum will cause many/most of the labels to overlap. To see what a particular stick is, zoom in on it with "set xrange". Enjoy!