Posted by Anders Öhrn on June 20, 2006 at 09:25:01:
In Reply to: Re: Some questions about molcas's usage posted by Jellby on June 19, 2006 at 11:40:21:
: :The electronic term is not explicitly listed but is a part of the electronic energy. If you run the calculation without the external field it is again trivial to figure out the electronic term.
: Are you sure? That would be assuming the wavefunction does not change in response to the external field, or if you consider the distortion energy as part of the electronic interaction energy. In other words, I need to partition the energy change due to the external field into two contributions: the energy spent in changing the wavefunction from the unperturbed one, and the pure interaction energy between the molecule and the external field. To calculate this last term, I have to obtain the electric potential and/or field in the positions of the external charges and/or dipoles, and this is what I want to avoid.
Maybe this could be a solution. If you want to figure out the self-energy in the polarization (i.e. the cost in perturbing the charge density from its optimal configuration) you can do like this (assuming HF or single-state CASSCF): (1) run variational calculation with external field present; (2) copy the output orbitals from this calculation and link to INPORB and give as input to a HF calculation without external potential; (3) the variational calculation will minize the energy, and the energy difference between the first iteration and the last should be your self-energy. Then from the procedure Roland suggested you can get the total polarization energy, and then decompose it into self-energy and external interaction.
But I agree that this solution is a bit tedious and an improvment would be to compute the expectation value of the external potential operator and print it along the dipole etc at the end of the HF or CASSCF output. Its not so hard to modify the code to do this.