Page 1 of 1

frozen core approx. with soppa

Posted: 09 Jun 2019, 13:56
by leonid
Dear Dalton Community,

I'm calculating the dispersion coefficients using SOPPA(CCSD) and I want to use the frozen core approx. which is not the default one in Dalton.
Here is the example for Ne. It looks like Dalton does not recognize the command
.FREEZE
1 0

like e.g. for regular CCSD calculations.


The inputs and the output are attached. What is the best way to define the FC for such calculations?

Re: frozen core approx. with soppa

Posted: 09 Jun 2019, 15:22
by hjaaj
As you have found out, this has not been implemented: the information about frozen orbitals in the CC module is not transferred to the RESPONS module, and therefore Dalton aborts when trying to read the T2 coefficients as if no orbitals were frozen in the CC.

It might work if you try a calculation in two steps:

1. Just the SCF, retrieve the MO coefficients.
2. Use your .dal file with two additions: read in the converged SCF MOs from step 1 and freeze the first orbital in the new SCF (it will be converged in iteration 1). Then RESPONS should know not to include the first orbital in the linear response equations.

Re: frozen core approx. with soppa

Posted: 09 Jun 2019, 15:23
by hjaaj
PS. I recommend using Dalton2018 instead of Dalton2013.