geometry-dependent error at end of MULTI

62 views

Skip to first unread message

Irikura, Karl K. Dr. (Fed)

unread,

Dec 9, 2021, 6:38:25 PM12/9/21

to molpro-user

Dear MOLPRO experts,

I got the following fatal error. It occurs after the expectation values are computed.

A longer and shorter bond length ran successfully.

ORDCP: ERROR IN ASSIGNING SYMMETRY FOR STATE 1.2

Here is the relevant part of the input file:

***,PtH 5D7Q SO-CI in C2v point group, full active space

! 5 doublets + 7 quartets

memory,1500,M;

gprint,orbitals,civector,basis;

gthresh,energy=1.d-8,coeff=1.d-8;

core;

basis,default=aug-cc-pV5Z,Pt=aug-cc-pwCV5Z-PP

angstrom

geometry={

H 1 R

}

R=1.9899

! PtH+ for closed-shell orbitals

{rhf;wf,elec=18,sym=1,spin=0}

! PtH, doublet

!{rhf;wf,elec=19,sym=1,spin=1}

! first with (11/7) active space

{multi,wmk_old;maxiter,30

! 5 doublets: Sigma + Delta + Pi

frozen;

wf,elec=19,sym=4,spin=1; state,1; lquant,2;

wf,elec=19,sym=1,spin=1; state,2; lquant,2,0;

wf,elec=19,sym=2,spin=1; state,1; lquant,1;

wf,elec=19,sym=3,spin=1; state,1; lquant,1;

! 7 quartets: Sigma + Pi + Delta + Phi

wf,elec=19,sym=4,spin=3; state,1; lquant,2;

wf,elec=19,sym=1,spin=3; state,2; lquant,2,0;

wf,elec=19,sym=2,spin=3; state,2; lquant,1,3;

wf,elec=19,sym=3,spin=3; state,2; lquant,1,3;

expec2,lzz;

}

table, energy, lzlz

title, Energies and Lz**2 values (C2v)

(Following steps are MULTI(19/11), MRCI, SO-CI)

I’m using wmk_old for compatibility with lquant.

I get the same error if I skip the (11/7) calculation.

I get the same error with the smaller basis sets aug-cc-pwCVQZ-PP, aug-cc-pwCVTZ-PP, cc-pwCVTZ-PP, cc-pVTZ-PP.

I would be grateful for any suggestions!

Thanks,

Karl

-------------------------------------------------------------------

Dr. Karl K. Irikura

National Institute of Standards and Technology

100 Bureau Drive, mail stop 8320

Gaithersburg, MD 20899-8320

voice: 301-975-2510 fax: 301-975-3670

email: karl.i...@nist.gov

http://www.nist.gov/mml/csd/informatics_research/index.cfm

-------------------------------------------------------------------

da...@home-kreplin.de

unread,

Dec 13, 2021, 7:23:48 AM12/13/21

to molpro-user

Dear Karl,

thank you for your input file!

The old program is unfortunately not very robust when it comes to two different lquant numbers in a single state symmetry.

The error occurs because the lz=2 state could not be found in the reduced Hamiltonian in the final CI optimization for the printout.

In principle, you can avoid this by removing the printout (gprint,civector).

Nevertheless, I'm happy to tell you that in the meantime we fixed all lquant issues in the new multi program.

Your input runs now perfectly with the most recent Molpro 2021.3.1 version without the wmk_old.

The new multi version also fixes lquant issues, when states become degenerated for different lquant numbers which is another problem with the cold code.

Therefore, I would strongly recommend to update to the recent Molpro 2021.3.1 version when running lquant calculations.