Excited state Ehrenfest Dynamics

[flo]

Hi,

getting the proper excited state wavefunction in CP2K is currently not
possible. As a workaround, the option wfn_mix was introduced. This
allows you to replace any occupied orbital by a linear combination of
unoccupied orbitals in the restart file, which than can be used as
starting wfn in the Ehrenfest dynamics. If you have a look in the
regtest, there should be an example. Thsi approach is only good, if
you have an idea about the nature of your excited state. Most of the
time, the first excited state is well approximated by a HOMO to LUMO
transition. In principle it should be possible to represent any other
excited state by a linear combination of LUMO's but obviously, that
soon becomes non trivial.
I don't think that anybody is working on the TDDFT bit at the moment
to allow for a TDDFT wfn to be used as initial state for EMD.

Cheers
Flo