Molcas Forum

philippt

Member

Registered: 2015-11-27

Posts: 2

GASSCF

Hi,

as there are not much information about GASSCF available, I have problems understanding the input.

In your input example you show the following:

NGAS
5
GAS1
2  0  0  0  2  0  0  0
GAS2
0  2  0  0  0  2  0  0
GAS3
0  0  2  0  0  0  2  0
GAS4
0  0  0  1  0  0  0  1
GAS5
4  2  2  1  4  2  2  1
GSOC
4  8  12  14  20
4  8  12  14  20 

Five GAS rooms with different amounts of orbitals inside.

As Obviously GAS1 to GAS4 only have orbitals with different symmetries, I would assume MOLCAS to choose always the highest occupied and the lowest unoccupied orbitals. But GAS5 has orbitals from all the symmetry groups - are those then the next best occupied/unoccupied orbitals below/above all previous orbitals from lower GAS rooms ... ?

Also is there the possibility to specify the GAS rooms as CAS/RAS room/s in the *.*Orb files via 1,2,3,... ?

Thank you in advance
Philipp

limanni

Member

Registered: 2015-11-09

Posts: 5

Re: GASSCF

In terms of starting orbitals (say from a previous RasOrb file) the built of GAS spaces follow the normal CAS rules... you exclude the Inactive and then start putting active orbitals distributed in your GAS spaces per symmetry and with the ordering as given in the RasOrb file itself. So what you get in GAS5 it is really up to your starting set of orbitals... if you want to have a say on that you better change orbital ordering inside RasOrb or you use alter keyword.

That said what you will have after the GASSCF has run is not really in your control as the variationality will mix orbitals. If you are not satisfied with the output orbitals you can add extra constraints by using the Supsym keyword.

Hope that help. Sorry for the delay ... I m not a forum person!