TDDFT Long Range Corrections
RNAlo... at gmail.com
Tue May 19 02:37:21 UTC 2009
Dear CP2K Users,
I am hoping someone can comment/correct anything that I am saying
I am attempting to see if I can use the TDDFT modules within CP2K to
calculated some excitation energies of a charge transfer state (CT).
One of the big troubles with TDDFT and charge transfer states, is that
it underestimates the excitation energies and often leads to spurious
low lying CT states being predicted. Several schemes have been
designed to correct for this so that the long range behavior is
treated more accurately (for example by including non-local Hartree
Fock exchange in the asymptotic limit).
I think the the SIC (self interaction error correction) tries to do
something similar specifically for ground state but I wasn't sure if
there were any explicit methods programmed into the TDDFT modules to
deal with this problem. Based on some preliminary calculations on a
model compound, I find several spurious CT states in both Gaussian and
CP2K that shouldn't be there.
Any help/ideas on this issue will be appreciated.
More information about the CP2K-user