locking the desired MO configuration
Clicked A Few Times
8:25:15 AM PDT - Thu, Aug 22nd 2013
Dear developers;

I would like to perform an optimization on the
locked orbital configuration and I am not sure
whether I can do it or not. The principle story is
simple; imagine there is a three fold degenerate
state of Tg (No.;50,51,52) and we have two
situations (by adding 3 electrons to this state);

Low spin === 50(up-dn), 51(up)
High spin == 50(up),51(up),52(up)

The question is a low spin state with the higher
energy than previous one. This spin state, can be
called spin-flip state;

LS(spin flip)===50(up),51(dn),52(up)

I would like to do it within DFT, by fixing
configuration of orbital. In principle it should
be possible.

I would be grateful, if somebody give a hint to me.


Best regards,

Shahab
-- 
ººººººººººººººººººººººººººººººººººººººººººººººººººººººººº

                  S. Shahab Naghavi      

                        (SISSA)   

  Scuola Internazionale Superiore di Studi Avanzati

          Via Bonomea 265, Tireste, Italy                 

Phone:             +39-403787448          

Email(1):        snaghavi@sissa.it

ººººººººººººººººººººººººººººººººººººººººººººººººººººººººº

Forum Regular
2:31:14 PM PDT - Thu, Aug 22nd 2013
Hi Shahab,

Take a look at the description of the occup block in the DFT documentation for NWChem 6.3.
http://nwchemgit.github.io/index.php/Release62:Density_Functional_Theory_for_Molecules#OCCUP...
This can be only be used for single point energy calculations (i.e. you cannot perform geometry optimizations with this turned on because MO ordering can change if the geometry changes)

Let us know if you have further questions.

Best regards,
-Niri
niri.govind@pnl.gov

Clicked A Few Times
10:41:21 AM PDT - Mon, Aug 26th 2013
Dear Niri,

Thanks a lot for your complete answer. I compiled the
newest version of nwchem and tried what you said, and
it is perfect. My question is, if I knew the
structural change is so small to change the ordering of
orbitals, then optimization should be possible with the
"occup" keyboard. Because I noticed that "occup"
keyboard keeps the ordering as I defined in
input. An infinitesimal change shouldn't perturb the
correct ordering.

Moreover, I did a trick to get my desired structure, in
GAMESS, but I would like to replicate it with
nwchem. In that code, I used D5d symmetry, and within
this symmetry the only possibility is T==E+A
(E_up,E_dn,A_up). The code follow this symmetry, and I
can get it converged for Sz=1/2, which is not the
ground state (L=2). In the NWCHEM, the code break the
symmetry---which is absolutely true--- to obtain the
ground state but I want to avoid it by follow the
symmetry blindly. ADAPT doesn't change the outcome.

However, I think "occup" will do the same.

Thanks in advance, for your time and concern.

Best regards,
Shahab


Forum >> NWChem's corner >> General Topics