[molpro-user] CASSCF gradients with RS2

[Alex Brown]

Dear All,

     I am trying to use the RS2 gradient program to computed state- 
averaged CASSCF gradients using the NOEXC option. In particular, I am  
trying to compute the gradient of the first excited state in a state  
averaged calculation - note that the states are "relatively" close in  
energy. The CASSCF calculation converges but when I call the RS2  
module, I get the following error (I attach part of the input file  
below):

wrong ci coefficient:        1101  6.787851797470805E-002
   6.807053294665716E-002

ERROR EXIT
CURRENT STACK:      CIPRO  MAIN

So there appears to be some sort of root flipping (maybe ?). It is  
unclear to me why this will occur as I thought the RS2,NOEXC would  
provide/use the CASSCF results. Any insights into the cause of the  
problem or options/suggestions to clear up the problem, would be  
greatly appreciated.

Thanks,
Alex

Here is a snippet from my input file (I can send the whole input file  
if needed):

basis=avdz
{hf;wf,34,1,0}
{multi;occ,20;closed,10;maxiter,100;wf,34,1,0;state,4;start, 
2100.2;canonical}
{rs2;noexc;maxiter,100,300;wf,34,1,0;state,1,2}
forces

Yes, there are reasons why it is C1 symmetry and I'd be happy to  
explain in case anyone really wants to know.

Alex Brown
Assistant Professor
Department of Chemistry
University of Alberta
Edmonton, AB
T6G 2G2
http://www.chem.ualberta.ca/~abrown/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20070704/eebe9612/attachment.html>