[molpro-user] Artificial stabilization in MRCI calculation for HBO molecule.

[Gang]

Dear Molpro users,

I've been doing the PES calculation for HBO molecule. I wrote a short 
script with DO loop to stretch the H-B bond continuously.  Please check 
the  attached  output file and potential energy curve.   I also 
recalculate all the points individually from scratch.  The two different 
approaches gave me different final energies which I don't understand.  
The only difference is that the continuous MCSCF calculation will use 
the MCSCF orbital (2140.2)from last point calculation as an initial 
guess, when the individually calculation will use HF wavefunction 
(2100.2)as a initial guess for MCSCF calculation.   I can understand a 
bad initial guess of electronic wavefunction may cause some convergence 
difficulties.  But why the continuous calculation gave us an artificial 
stabilization? 
How should I modify the input file in order to force the program to use 
the HF wavefunction as initial guess without losing the convenience of 
using DO loop?
Thank you very much for your time!

Gang
-----------------------
University of York

-------------- next part --------------
A non-text attachment was scrubbed...
Name: HBO PES.JPG
Type: image/jpeg
Size: 52460 bytes
Desc: not available
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20090909/82dddacd/attachment.jpe>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: HBO-acvtz210a.out
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20090909/82dddacd/attachment.ksh>