Getting rs2 to find the desired ref state

[Falcetta_Michael]

Dear molpro users,

I am trying to do a rs2 calculation on the 6,7,&8th roots derived from
an mcscf.  The mcscf energies are

Root   MCSCF ENERGY
6   	-114.88831918
7   	-114.86572251
8   	-114.83010692
9   	-114.75248957

When I take the state averaged canonical orbitals fromt the mcscf (and
use the same config list for the ref space of the rs2 as I used in the
mcscf), the rs2 module find the following:

Number of p-space configurations:  13

 Reference wavefunction optimized for reference space (refopt=1)

 State     Reference Energy
   6        -114.88831918
   7        -114.83010692
   8        -114.75248951

Clearly RS2 has missed what was the 7th root in the mcscf!! 

 If I ask the RS2 to extract 10 roots it does correctly find the 7th
root to correspond to the 7th root in the mcscf (i.e., the p-space does
contain the correct root and it is calculating the reference energy
correctly)

My question is: Is there an efficient way to get the RS2 program to get
the correct roots?  I think I will need to extract roots 6,7,&8 because
at certain geometries these roots will be close in energy and may cross,
but how can I ensure that the RS2 calculation gets the correct roots.

Thanks in advance

Mike


Dr. Mike Falcetta

Associate Professor of Chemistry
Roberts Wesleyan College
2301 Westside Drive
Rochester, NY 14624