PC GAMESS/Firefly-related discussion club

Learn how to ask questions correctly

Slawomir Janicki
slawomir.janicki@comcast.net

When selecting I like to visualize all non-core orbitals. Use your favorite software. For me back in the day it was Molden, now it is ChemCraft. You will find that most often the orbitals are NOT ordered as you would like to see them. Pay attention to the orbital symmetry - you want to have all orbitals that you think you need with the correct symmetry. Aromatic systems are particularly messy as sigma orbitals mix with pi when symmetry allows. After a few attempts you will notice when the selection works for you. The orbitals in the CAS space will become "cleaner" - the admixing of other parts from outside of the space will diminish. Alternativelly, when you miss something, the calculation will likely pull in a contribution with the missing symmetry from the surrounding orbitals.

For more complex systems the cannonical orbitals are just to messy for me. In these cases I use the NBO orbitals as a starting point. NBO now comes with an orbital viewer, I don't know if any other program can display NBOs.

Slawomir

On Thu Aug 20 '09 1:53am, Igor Polyakov wrote
---------------------------------------------
>Hi,

>i didn't check your files but i wanted to give some general comment- in any MCSCF calculation with "not proper" orbital selection the results should be very obscure. The general rule is to include "important" orbitals in active space, for example if u study s0-s1 homo-lumo transition u must include these right homo and lumo orbitals, but in general these orbitals are not homo and lumo orbitals of RHF printout- you should conduct CIS calculations or inspect your orbitals manually. In general case of some reaction i suppose u could try to include high-lying occupied orbitals and low-lying vacant orbitals which concern the chemical bond to be broken or made during the reaction.

>Best luck. Igor.
>
>
>On Wed Aug 19 '09 4:33pm, Alex Maiorov wrote
>--------------------------------------------
>>Hallo,

>>This i have read in the previous discussion about SA and SS MCSCF
>>>>>>The proper selection of active space is very important.
>>>>>>CAS (2,2) is natural for your problem. You can use larger spaces
>>>>>>but be careful to select proper orbitals!
>>>>>>And sure, you can use $guess norder=1 iorder(1)=... to reorder orbitals.

>>i cant understand how to find out which orbitals are proper or not!
>>How can i see what to reorder? What`s the metter with my results if i didnot select the proper
>>orbitals?

>>For example
>>I made calculations of Transition state of reaction O3+C2H4 in basis 6-311pg**:

>>1)by the means of ss cas(8,8) - name of file           d_c2h4o3_cas.inp
>>2)i calculated the caspt2 energy - name of file        d_c2h4o3_cas_pt2.inp

>>Don`t pay attention than files were made for GAMES US, i made this calculation
>>in pcgamess too, but i could not find them.

>>Thank you very much/