Anharmonic Peaks

In PhononA one makes a section accross the phonon dispersion bands at fixed wavevector k, obtains the intensity spectra at k, compares the intensity with relevant phonon polarization vectors and receives the anharmonic peaks, their shapes, hence the widths and average positions. This is rather complex analysis, but it allows to segregate the anharmonic phonon peaks still within conventional classification performed for the harmonic phonons. Inspite of the fact that anharmonic phonons can be quit wide, and not necessarely similar to Lorenzian, or Gaussian shapes.

In FIGURE.3 an example of anharmonic phonon peak analysis is presented. They are constructed as follows. From data of FIGURE.2 (down) of map of anharmonic phonons we plot intesity along wavevector k = Γ . From group analysis (present in Phonon Software) one finds that at k = Γ sharp phonon modes (including acoustic B_1u,B_2u,B_3u,) can be labelled as: 7A_g + 5B_1g + 5B_2g + 7B_3g + 8A_u + 10B_1u + 10B_2u + 8B_3u). All modes are single degenerate. To carry on more transparent analysis we limit the spectrum to "all A_u modes", which is shown at the bottom raw of FIGURE.3 (left). All eight A_u peaks are shown on first and second raw of the same FIGURE.3. The peak positions and widths are estimated from these plots. It is also shown that all A_u modes fit to the total spectrum, see FIGURE. 3 (down, right).

Similar analysis can be performed from any wavevector k, any collection of anharmonic phonons, and independent of peak degeneracy. Note as well, that this option requires a lot of computer memory. But this we do not consider as a defect of the method, since anyway one never will be able to make physical use of such a vast set of information.

Currently to all anharmonic modes the Gaussian functions are fitted, and tabularized.

FIGURE.3. (Up) Plot of anharmonic phonon density of states for A_u modes of MgSiO₃ at V=const conditions.

Last update: September 20, 2017