[molpro-user] optg with procedures

[harding]

Dear Molpro Community,

  We are trying to do geometry optimizations with an energy expression 
that combines ccsd(t) and rs2c energies,

  i.e.  energy = e(ccsdt/dz) + e(rs2c/tz) - e(rs2c/dz)

  We have defined a procedure as follows:

PROC mix
basis=cc-pvdz
{rhf;closed,3;occ,4;wf,7,1,1}
{rccsd(t);closed,3;occ,4;wf,7,1,1}
e2 = energy
{multi;closed,1;occ,6;wf,7,1,1}
rs2c,shift=0.2
e1 = energy
basis=cc-pvtz
{multi;closed,1;occ,6;wf,7,1,1}
rs2c,shift=0.2
e3 = energy
e4 = e2+e3-e1
ENDPROC


  The first time it tries to do a ccsd(t) calculation following a rs2c 
calculation it crashes with the error message:

NUMBER OF OCCUPIED ORBITALS NOT CONSISTENT WITH NUMBER OF ELECTRONS FOR 
CCSD. NELEC=   7  NCORE=  1  NCLOS=  0  NOCC=  6  MS2= 1

It appears that numbers of occupied and closed orbitals for the second 
ccsd(t) calculation are set to values appropriate for the multi/rs2c 
step rather than the values on the rccsd(t) command line or the 
immediately preceding rhf step. If we do the same sequence of 
calculations but not part of an optg calcualtion (i.e. rhf, 
rccsd(t),multi,rs2c,rhf,rccsd(t)) it works fine. We get the same error 
message if we do the optg using startcmd=rhf rather than a procedure. 
This was done using version 2010.1.

Any suggestions as to what is wrong would be appreciated.

  Thanks.

    Larry Harding



The complete data set is as follows:

PROC mix
basis=cc-pvdz
{rhf;closed,3;occ,4;wf,7,1,1}
{rccsd(t);closed,3;occ,4;wf,7,1,1}
e2 = energy
{multi;closed,1;occ,6;wf,7,1,1}
rs2c,shift=0.2
e1 = energy
basis=cc-pvtz
{multi;closed,1;occ,6;wf,7,1,1}
rs2c,shift=0.2
e3 = energy
e4 = e2+e3-e1
ENDPROC

memory,20,M
angstrom
nosymmetry
geometry={
  c1
  H1 c1 rch
}
  RCH  =   1.1

basis=cc-pvdz
{rhf;closed,3;occ,4;wf,7,1,1}
{rccsd(t);closed,3;occ,4;wf,7,1,1}
e2 = energy
{multi;closed,1;occ,6;wf,7,1,1}
rs2c,shift=0.2
e1 = energy
basis=cc-pvtz
{multi;closed,1;occ,6;wf,7,1,1}
rs2c,shift=0.2
e3 = energy
e4 = e2+e3-e1

optg,proc=mix,variable=e4
frequencies,proc=mix,variable=e4
---