| 11:51:09 AM PST - Fri, Nov 22nd 2019 |
|
You only need to run the calculation without a field once.
Running an optimization first would be fine. My example input assumed you already had the geometry of interest. The reason it was set as a call to tddft gradient is that the excited state density matrix is not computed if you are only calculating the excitation energies; it is only calculated if you need an excited state gradient. If you are doing a geometry optimization on the excited state, then you will be calculating excited state gradients and will be generating the excited state density matrix. The caveat to this is that the XC functional must have analytical 2nd derivatives implemented; otherwise, the code is calculating gradients through numerical differentiation and an excited state density matrix is not calculated. For those functionals (e.g. M06), we showed that one possible workaround is to use a different functional to calculated the excited state density matrix and then use the functional of interest for the RT-TDDFT propagation (see, https://pubs.acs.org/doi/abs/10.1021/acs.jpclett.6b00282)
You need a visualization block for each propagation, including the one without a field. Also, don't include the treference keyword or the dplot keyword.
Use whatever FFT utility works for you. In the past I wrote analysis code in Fortran around FFTPACK from netlib, but most programing languages have FFT libraries that can be linked.
|