Posted by Lukas Pasteka on January 02, 2009 at 00:57:24:
Hello,
I am trying to compute the geometry optimalization of an excited state with CASPT2 method but I get this error:
"Alaska requests the Numerical_Gradient module to be executed!
...
ERROR !!!
It was selected to run numerical gradients using
energies for root/state number 2
but only 1 exists."
CASSCF optimalization works great but CASPT2 fails as described. I want to optimalize only the second root and not to use SA orbitals (CiRoot 2 2 1). So I use RASSCF keyword
CiRoot
1 2
2
then again CASPT2 keyword
Multistate
1 2
So I do compute only 1 state, and it is the 2nd one, am I right? ALASKA tries to use the 2nd state but there is only one available (which in fact is the 2nd, but the 1st is not available). So should I use keyword
Root
1
in ALASKA? Would it optimize the excited state for me or am I getting this all wrong?
Input I was using (part of interest) is as following:
"...
&RASSCF &END
Inactive
8 5
Ras2
4 2
nActEl
6 0 0
Symmetry
1
Spin
3
CiRoot
1 2
2
LumOrb
End of Input
&CASPT2 &END
MaxIter
50
Density matrix
Multistate
1 2
End of input
&ALASKA &END
End of input
&SLAPAF &END
Iterations
70
End of input
>>> EndDo <<<"
Thanks for any advice.
Lukas