[molpro-user] MRCI records of SOC from different MCSCF calculations

[Peterson, Kirk]

Dear Wenli,

yes, currently in the SO calculations the wavefunctions used in hlsmat must use a common set of orbitals.

best,

-Kirk

On Jan 24, 2014, at 5:03 PM, ZOUWL <qcband at gmail.com> wrote:

> Hello,
> 
> I'm wondering whether it is possible to do SOC calculation with MRCI records from more than one MCSCF calculations. In the attached input file, the reference wavefunctions of singlet and triplet states of BF are calculated separately which is necessary in the case of RASSCF or the general MCSCF since the states with different spin multiplicities cannot be averaged. In the manual, it states "the OCC and CLOSED cards must be the same for all states used in a HLSMAT calculation". I think this is true in my input, but an error was encountered:
> 
>  ?ORBITALS OF BOTH WAVE FUNCTIONS FOR SPIN-ORBIT CALCULATION NOT IDENTICAL
>  GLOBAL ERROR fehler on processor   0
> 
> Does this mean that all the MRCI records must be based on the same state-averaged MCSCF calculation?
> 
> Best regards,
> Wenli
> 
> ***,BF
> memory,800,m
> 
> basis=vdz
> 
> geometry={
> B;
> F, 1, r1;
> }
> 
> r1=1.1 angstrom;
> 
> hf;
> 
> {multi;
> frozen,2,0,0,0;closed,2,0,0,0;occ,6,2,2,0;
> wf,14,1,0;      ! X1Sigma+
> wf,14,2,0;      ! 1Pi
> wf,14,3,0;}
> 
> {ci;noexc;wf,14,1,0;save,2001.3;}
> {ci;noexc;wf,14,2,0;save,2002.3;}
> {ci;noexc;wf,14,3,0;save,2003.3;}
> 
> {multi;
> frozen,2,0,0,0;closed,2,0,0,0;occ,6,2,2,0;
> wf,14,2,2;      ! 3Pi
> wf,14,3,2;}
> 
> {ci;noexc;wf,14,2,2;save,2004.3;}
> {ci;noexc;wf,14,3,2;save,2005.3;}
> 
> {ci;hlsmat,ls,2001.3,2002.3,2003.3,2004.3,2005.3;}
> ---
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user