[molpro-user] conical optimisation question

Jayashree yfpjaya at gmail.com
Wed Jan 6 10:20:39 GMT 2010


Hi all,
I am working on minimising the energy of methoxy molecule (CH3O) with
respect to symmetric displacements that will preserve the C3v symmetry of
the molecule. In other words, I am using the active variables- CO bond
length, 3CH bonds displacements (all in same direction) and 3 HCO bending
displacements (all in same direction as in umbrella motion). As you will see
in the output of the following input, the final geometry after optimisation
is CO ~ 1.427 A, and corresponding parameters for CH and HCO angle.

I notice that when I scan the potential with respect to only CO bond
compression, the energy of the molecule when CO ~ 1.38 A for example is
lower than CO ~ 1.427 A. The CO bond compression preserves the C3v symmetry
of the molecule, and so I would expect the Conical optimisation procedure to
arrive at a minimum with respect to symmetric motions in the molecule. The
difference in energy between 2 above such geometries is about 100cm-1. How
can I force MOLPRO to search for  the correct minimum with respect to the
symmetric motions? (I know that the asymmetric motions do not have a minimum
at C3v symmetry, hence the Jahn-Teller effect in the molecule.) I notice the
same trend with the other 2 symmetric motions as well. Any comments or
suggestions would be great.

*Input-*
basis=6-311+g(2d,2p)
geomtype={zmat}
geometry={Angstrom;
          c;
          o,c,oc;
          h1,c,h1c,o,h1co;
          h2,c,h1c,o,h1co,h1,120.0;
          h3,c,h1c,o,h1co,h1,240.0}
oc=1.30,h1c=1.1,h1co=109.1

{uhf;occ,7,2;open,7.1;wf,17,1,1}
{casscf;maxiter,40;occ,9,3;closed,5,1;wf,17,1,1;state,1;weight,0.5;wf,17,2,1;state,1;weight,0.5;
 CPMCSCF,GRAD,1.1,spin=0.5,accu=1.0d-8,record=5101.1
 CPMCSCF,GRAD,1.2,spin=0.5,accu=1.0d-8,record=5102.1}
 {Force,
  SAMC,5101.1
  CONICAL,6100.1}
 {Force
  SAMC,5102.1
  CONICAL,6100.1}
{optg,startcmd=casscf}

-- 
Jayashree
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20100106/535294d9/attachment.html>


More information about the Molpro-user mailing list