problems of spin contamination in MCSCF of molpro2021

[Joonghan Kim]

Dear Molpro users

 I faced the problems of spin contamination in MCSCF calculation when I use Molpro2021.2 and Molpro2021.3.

I performed the state-average CASSCF calculations for singlets and triplets simultaneously using the Molpro2021.3.

The result of calculation is as follows:

 ITER  MIC  NCI  NEG     ENERGY(VAR)    ENERGY(PROJ) ENERGY CHANGE       GRAD(0)  GRAD(ORB)   GRAD(CI)      STEP      TIME

 Spin contamination in state 1.3: <S**2> =  0.3E-01 (+2.0)
 Annihilate S =  2.0 component
 Spin contamination in state 2.3: <S**2> =  0.2E-02 (+2.0)
 Annihilate S =  2.0 component
 Spin contamination in state 1.3: <S**2> =  0.5E-02 (+2.0)
 Annihilate S =  3.0 component
 Spin contamination in state 2.3: <S**2> =  0.9E-04 (+2.0)
 Annihilate S =  3.0 component
 Spin contamination in state 1.3: <S**2> =  0.5E-04 (+2.0)
 Annihilate S =  4.0 component
 Spin contamination in state 2.3: <S**2> =  0.3E-07 (+2.0)
 Annihilate S =  4.0 component
 Spin contamination in state 1.3: <S**2> =  0.3E-05 (+2.0)
 Annihilate S =  5.0 component
 Spin contamination in state 2.3: <S**2> =  0.1E-08 (+2.0)
 Annihilate S =  5.0 component
 Spin contamination in state 1.3: <S**2> =  0.3E-07 (+2.0)
 Annihilate S =  6.0 component
 Spin contamination in state 1.4: <S**2> =  0.3E-02 (+2.0)
 Annihilate S =  2.0 component
 Spin contamination in state 1.4: <S**2> =  0.5E-03 (+2.0)
 Annihilate S =  3.0 component
 Spin contamination in state 1.4: <S**2> =  0.5E-05 (+2.0)
 Annihilate S =  4.0 component
 Spin contamination in state 1.4: <S**2> =  0.2E-07 (+2.0)
 Annihilate S =  5.0 component
   1   77  825    0   -123.24579816    -123.50160923   -0.25581107    0.20599686 0.01754770 0.09428132  0.19E+01    127.90
 Spin contamination in state 1.1: <S**2> =  0.3E-07 (+0.0)
 Annihilate S =  1.0 component
 Spin contamination in state 2.1: <S**2> =  0.2E-03 (+0.0)
 Annihilate S =  1.0 component
 Spin contamination in state 2.1: <S**2> =  0.1E-06 (+0.0)
 Annihilate S =  2.0 component
 Spin contamination in state 2.1: <S**2> =  0.2E-08 (+0.0)
 Annihilate S =  3.0 component
 Spin contamination in state 1.2: <S**2> =  0.3E-06 (+0.0)
 Annihilate S =  1.0 component
 Spin contamination in state 1.3: <S**2> =  0.2E-01 (+2.0)
 Annihilate S =  2.0 component

Eventually, it has not been converged.

However, it has been converged with the same input file when I used the Molpro2020.2.

The result is as follows:

 ITE  MIC  NCI  NEG      ENERGY(VAR)     ENERGY(PROJ) ENERGY CHANGE       GRAD(0)  GRAD(ORB)   GRAD(CI)      STEP      TIME
   1   54  506    0    -123.24579733    -123.50251803   -0.25672070    0.20600085 0.00131431 0.00181733  0.19E+01     90.43
Spin contamination in state  1.2: <S**2> =  0.2E-07 (+0.0)
Annihilate S =  1.0 component
   2   19  269    0    -123.42192082    -123.47021921   -0.04829839    0.26853821 0.00172090 0.00216772  0.57E+00    139.67
Spin contamination in state  2.1: <S**2> =  0.2E-08 (+0.0)
Annihilate S =  1.0 component
   3   43  586    0    -123.47377575    -123.47928515   -0.00550940    0.07356155 0.00029067 0.00041992  0.21E+00    237.44
   4   19  296    0    -123.47978305    -123.47987018   -0.00008712    0.00692155 0.00003007 0.00003137  0.22E-01    287.63
   5   27  406    0    -123.47987093    -123.47987095   -0.00000001    0.00009474 0.00000000 0.00000002  0.30E-03    355.58
   6    4   17    0    -123.47987095    -123.47987095   -0.00000000    0.00000002 0.00000001 0.00000002  0.43E-07    359.36
CONVERGENCE REACHED!   Final gradient:  0.00000000 ( 0.21E-08)
                       Final energy:    -123.47987095

How can I solve this problem?

Do I have to use the Molpro2020.2 ?

Thank you in advance.

Joonghan Kim

[da...@home-kreplin.de]

Dear Joonghan,

we have changed the spin-projection Method from Molpro 2020 to Molpro 2021, so there might be a small chance that there is something going wrong.
It would be great if you could share an input file so that we can reproduce and investigate the problem.

Best wishes,

David

[Joonghan Kim]

Dear David

I'm sending you the input and output files.

As I mentioned previously, the input file is the same.

The output using the Molpro2021.3 shows oscillation in the MCSCF calculation. 

I hope to resolve this problem soon.

Thank you in advance.

Joonghan Kim

2021년 12월 18일 (토) 오전 2:45, da...@home-kreplin.de <da...@home-kreplin.de>님이 작성:

--
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/914ba3d3-c64f-4567-a1c3-4be75fd4fc14n%40googlegroups.com.

--

Joonghan Kim, Ph. D.

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

Associate Professor

Department of Chemistry

The Catholic University of Korea

Bucheon, Gyunggi-do 14662, Korea

(Tel)+82-2-2164-4338

(Fax)+82-2-2164-4764

(E-mail) joong...@catholic.ac.kr, quan...@gmail.com

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

[Joonghan Kim]

Dear David

Is there any improvement on this problem?

Do I have to use the Molpro2020?

Thank you in advance.

Joonghan Kim

2021년 12월 18일 (토) 오후 2:09, Joonghan Kim <quan...@gmail.com>님이 작성:

[iri...@verizon.net]

Hi Joonghan,

Sometimes using WMK_OLD works for me in these cases ("MULTI,WMK_OLD;"). 

Good luck,

Karl