Reference energy in ground-state MRCI is higher than CASSCF

Irikura, Karl K. Dr. (Fed)

unread,

May 10, 2024, 3:13:13 PMMay 10

to molpro-user

Dear Molpro experts,

I have a situation where the reference energy in a ground-state (within irrep) MRCI is higher than the CASSCF energy. I know that when the reference energy is too low, I should request a higher state in the MRCI. But when the reference energy is too low, I don’t know how to help the MRCI find the correct reference state.

Here is an input file

***,PtO, 1-Delta

! Reference energy in ground-state MRCI is higher than CASSCF

memory,900,M;

!gprint,orbitals,civector

symmetry,X,Y

angstrom

geometry={

Pt

O 1 R

}

! Problem for R=2.66 and R=2.70 for TZ, QZ and 5Z basis sets

R=2.66

basis=aug-cc-pwcvtz-pp

! Only O(1s) frozen in CI

core,1,0,0,0;

! triplet Sigma- HF orbitals

{rhf;wf,elec=26,sym=4,spin=2}

! same results with WMK and WMK_OLD

{multi,wmk_old

! O(1s) closed but not frozen

frozen,0,0,0,0

closed,1,0,0,0

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

wf,sym=4,spin=0; state,1; lquant,2

expec2,lzz

}

lsq1 = lzlz(1)

lsq2 = lzlz(2)

ecas1 = energy(1)

ecas2 = energy(2)

table, energy, lzlz

title, Energies and LzLz values

{ci,maxit=60,maxiti=150; wf,sym=1,spin=0; state,1}

eref1 = energr

eci1 = energy

edav1 = energd

ediff1 = eref1 - ecas1

{ci,maxit=60,maxiti=150; wf,sym=4,spin=0; state,1}

eref2 = energr

eci2 = energy

edav2 = energd

ediff2 = eref2 - ecas2

table,ecas1,ecas2,eref1,eref2,ediff1,ediff2,lsq1,lsq2,eci1,eci2,edav1,edav2

!table,ecas2,eref2,ediff2,lsq2,eci2,edav2

The result is:

ECAS1 ECAS2 EREF1 EREF2 EDIFF1 EDIFF2 LSQ1 LSQ2 ECI1 ECI2 EDAV1 EDAV2

-193.3242933 -193.3242933 -193.3242933 -193.3227024 0.0 0.00159094 4.0 4.0 -194.0593274 -194.0573976 -194.1477497 -194.1457518

Grateful for any help,

Karl

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

Dr. Karl K. Irikura

National Institute of Standards and Technology

100 Bureau Drive, mail stop 8320

Gaithersburg, MD 20899-8320 USA

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

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

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

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

Peterson, Kirk

unread,

May 10, 2024, 5:48:14 PMMay 10

to Irikura, Karl K. Dr. (Fed), molpro-user

Hi Karl,

usually that means you need to increase the pspace threshold a bit, e.g., pspace,0.3 . I'm not sure if 0.3 is right or not or since I can't remember exactly what the default is.

best,

-Kirk

From: 'Irikura, Karl K. Dr. (Fed)' via molpro-user <molpr...@googlegroups.com>
Date: Friday, May 10, 2024 at 12:13 PM
To: molpro-user <molpr...@googlegroups.com>
Subject: [molpro-user] Reference energy in ground-state MRCI is higher than CASSCF

[EXTERNAL EMAIL]

--
You received this message because you are subscribed to the Google Groups "molpro-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/SA1PR09MB868804633A39E5BC8D7B93A0EFE72%40SA1PR09MB8688.namprd09.prod.outlook.com.

Irikura, Karl K. Dr. (Fed)

unread,

May 14, 2024, 5:23:24 PMMay 14

to Peterson, Kirk, molpro-user

Thanks, Kirk!

That worked (pspace,1.) at some geometries but not all. What worked well was to increase NSTATI.