About 4th order force constants in ShengBTE format

[ybC007]

Dear Dr. Tadano,

I'm planning to convert the 4th order force constants generated by alm to ShengBTE format. I found that the convert for 3rd force constant is implemented. So I tried to imitate the source code void Writer::save_fc3_thirdorderpy_format() to write a new function for 4th force constants. But when I compiled and run it, it reports the ERROR ERROR in save_fcs_alamode_oldformat MESSAGE: This cannot happen. and I can't find the reason.

I would like to ask that the time when you planning to implement the function. Or would you like to read the code I write for the format convert ?

Thank you very much in advance!
Looking forward to your kind advice.

[ttadano]

I have added a new option FC4_SHENGBTE in the feature/shengbte_4ph brach. Also see here.

You can try this version, but please be cautious because I have not tested it.

[ybC007]

Cool! I'll have a try.
Thanks a lot!😁

[ybC007]

Dear Dr. Tadano,
Hello! I have tried the new option FC4_SHENGBTE, but the calculated thermal conductivity with four phonon interaction seems a little abnormal. Then I read the source code and found there maybe a mistake. But I'm not sure about it.

In file writer.cpp from line 1264 to 1275.

                                        vecs[0][m] = (x_image[0][atom_tmp[0]][m]
                                                      - x_image[0][jat0][m]
                                                      + x_image[cell_now[0]][0][m]
                                                      - x_image[0][0][m]) * Bohr_in_Angstrom;
                                        vecs[1][m] = (x_image[0][atom_tmp[1]][m]
                                                      - x_image[0][kat0][m]
                                                      + x_image[cell_now[1]][0][m]
                                                      - x_image[0][0][m]) * Bohr_in_Angstrom;
                                        vecs[2][m] = (x_image[0][lat][m]
                                                      - x_image[0][lat0][m]
                                                      + x_image[cell_now[2]][0][m]
                                                      - x_image[0][0][m]) * Bohr_in_Angstrom;

There may be something with vecs[2][m] = (x_image[0][lat][m]? It should probably be modified like this 'vecs[2][m] = (x_image[0][atom_tmp[0]][m]' ?

Thank you very much in advance!
Looking forward to your kind advice.

[ybC007]

Sorry for the spelling mistake. It should probably be modified like this vecs[2][m] = (x_image[0][atom_tmp[2]][m] ?

[ttadano]

Yes, that is obviously a mistake. I've fixed it.