egrad does not write ed.plt

[antti_karttunen]

Hello all,

I'm investigating the nature of the excited states of a certain system (TDDFT, closed shell system, both singlet and triplet states). I try to produce the differential density for a excited state by adding $pointval to control file and running egrad -proper. For some reason egrad only produces td.plt and esd.plt (differential spin density for excited state). Although the data in esd.plt seems to make sense when I visualize it and compare to orbital plots, I was expecting an ed.plt file (manual discusses such file). Am I missing something very fundamental here? I'm not too familiar with excited state calculations yet.

[uwe]

Hi Antti,

this is a bug. The file td.plt is the excited state density, the file esd.plt is the one of the differential total density, and not of the differential spin density (which does not exist for a closed shell system anyway).

So for closed shell systems esd.plt = ed.plt. You can check that by taking td.plt from dscf/ridft minus td.plt from egrad (or better calculate a small molecule and print the grid in a readable format). For open shell systems this bug is a problem since egrad first writes out the differential total density and then the differential spin density to the same file, so the ed.plt file gets lost.

Uwe

[antti_karttunen]

Hi Uwe,

Thank you for the prompt reply, this really clarified the issue!