Question regarding the use of different supercell sizes for harmonic and anharmonic parts

[czp520]

Dear Dr. Tadano,
Hello, when obtaining the second-order force constant file (.xml) at a specific temperature using dfc2, I used the 3x3x3 supercell results for the input XML file containing the original FC2 (i.e., the harmonic part was based on the 3x3x3 supercell), while for the FC2 correction file from the SCPH calculation, I used the 2x2x2 supercell results (i.e., the anharmonic part was based on the 2x2x2 supercell). However, I still obtained the second-order force constant file at the specific temperature under the 3x3x3 supercell. I would like to ask whether this approach is reasonable.

Below is my procedure for running dfc2:

DFC2 -- a generator of renormalized harmonic Fcs from ScPH outputs.
XML file containing original Fc2 : ../../../333/har/MgSnSb333.xml
Output xml filename with anharmonic correction : 300-SCPH-333.xml
FC2 correction file fromSCPH calculation : MgSnSb.scph dfc2
Target temperature : 300

New XMl file 300-SCPH-333.xml was created successfully.

Thank you very much for your time and attention.
Kind regards

[ttadano]

That approach is quite normal. The anharmonic interaction is expected to be shorter ranged compared to the harmonic interaction, so adding anharmonic corrections only for pairs within the 2x2x2 supercell range is reasonable. Of course, please check whether the generated renormalized FCSXML yields a reasonable result.

[czp520]

Dear Dr. Tadano, thank you very much for your reply, I have received it.