[CP2K-user] MOLog coefficients are not normalized

[msha...@gmail.com]

Hello everyone,

I wanted to compute the MO overlaps for a geometry using the AO overlap and 
the MO coefficients in MOLog files. First I obtained the AO overlap matrix 
from DFT%PRINT%AO_MATRICES%OVERLAP and MOLogs from DFT%PRINT%MO. When I 
computed the MO overlaps using C^T * S * C (C is the eigenvector and S is 
the AO orbital matrix) I found that the eigenvectors coefficients are not 
normalized (the AO overlap matrix has 1 on its diagonal). Here is a 
response in the group which states that the Gaussian basis sets are not 
normalized (https://groups.google.com/g/cp2k/c/qCtT8eWENu0). Now, my 
question is that can we normalize the coefficients and then compute the MO 
overlaps or we should perform some other post-processing steps to correctly 
compute the MO overlaps?

Thanks very much in advance.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210503/83ff936a/attachment.htm>