I've added such a file (replace -molden.txt by .molden):
It was generated via Orca/NBO7 -> *.47, then acting gennbo on the *.47 file with NBO's molden=nbo option. Mentioning "orca" in the title section of the molden file does not solve the problem.
Indeed, this is an issue. I cannot easily find solution not only because the .molden format was not rigorously defined, but also the order and normalization condition of basis functions employed by ORCA are very strange. So, please always use NBO plot file when using NBO in combination with ORCA.
]]>It was generated via Orca/NBO7 -> *.47, then acting gennbo on the *.47 file with NBO's molden=nbo option. Mentioning "orca" in the title section of the molden file does not solve the problem.
]]>A final remark:
Multiwfn now acts fine on NBO's 31-40 files. However, for users that prefer to use the molden option in NBO's keylist, for example molden=nbo as a substitute for the *.37 file, the problem persists. Now, both Chemcraft and Multiwfn draw those wrong pictures described above.
This issue should also be caused by the weird convention of recording basis functions in ORCA. Currently I don't have such a file, you can try to add "orca" in any place of [Title] section of the molden, then Multiwfn will recognize that this file was generated by ORCA, special consideration will be employed during reading, in this case the problem may be solved.
]]>Thank you very much for solving the problem that quickly.
Best wishes,
Peter
This problem comes from the very special rule of recording basis functions of p shell in ORCA.
]]>