[Octopus-users] TD projections

[Robertson Burgess]

Dear Octopus Users,

I'm trying to analyse the source of various peaks within the optical absorption spectrum, which I am calculating using the explicit time-propagation method. I understand that it would be more straight forward to do this using the Casida method, as each peak can be directly attributed to an occupied/unoccupied pair of orbitals. However since I am calculating for large systems I am using the time-propagation method since it scales better. When I spoke to some people about this in Benasque last January I was told I could try using the TDOuput= td_occup to project the excitation peaks onto the ground-state wavefunctions, which could include some unoccupied orbitals. However when I did this I saw that the projections were given for each time-stamp. Is there some way to convert these time-dependant projections into frequency-dependant projections? That way they could be easily associated with individual peaks.

Sincerely,
Robertson Burgess
University of Newcastle, Australia