{SPAM}? {SPAM}? Re: [molpro-user] Dipole moment from mp2/rmp2 for (restricted) open shell system

[ishii at tcl.t.u-tokyo.ac.jp]

But, isn't it OK to write that the dipole moments are
calculated by the single-reference CASPT2?
I think the matter is effective digit of absolute values
of dipole momennts.

Ishii

> On Wednesday 05 March 2008 03:14, ishii at tcl.t.u-tokyo.ac.jp wrote:
> > You can use 'rs2' command (CASPT2) instead of rmp2.
> > In rs2 calculations, dipole moments are calculated.
> > Using ROHF as reference in rs2, rs2 is equivalent as RMP2.
> 
> This is actually not the entire truth. RS2 and RMP2 both give MP2 resu
lts when 
> applied to closed shell systems. But in open-shell systems, they diffe
r due 
> to using H_0 Hamiltonians for the perturbation theory. For ROHF wave 
> functions, RMP2 tends to give better results compared with UCCSD(T) 
> calculations.
> 
> That being said, i'm not aware of any way to calculate RMP2 dipole mom
ents 
> with the current version of molpro, except using finite fields and fin
ite 
> differences (i.e., getting dE/d(dmx)...dE/d(dmz) by adding small finit
e 
> fields to the core Hamiltonian and building difference quotients. See 
> "properties and expectation values"/"finite field calculations" in the
 
> manual)
> -- 
> Gerald Knizia
>