[molpro-user] FCI converges to wrong excited state?

Peter Knowles KnowlesPJ at cardiff.ac.uk
Wed Jan 21 09:09:28 GMT 2015

Dear Lorenzo,

I'm afraid the eigensolver in FCI is not as sophisticated as those in MCSCF and MRCI. Once it has generated a trial vector - which in this case happens to be singlet Delta - it then works in a Krylov space generated from that vector, and you will never reach vectors of different symmetry except perhaps by numerical accident following round-off.

Best regards,

> On 20 Jan 2015, at 13:58, Lorenzo Lodi <l.lodi at ucl.ac.uk> wrote:
> Scandium hydride ScH has a ground 1Sigma+ term  and (among other terms)
> a low-lying 1Delta term.
> Consider the attached MCSCF, MRCI and FCI calculations of the 1Sigma+
> and the two degenerate components of the 1Delta terms (takes ~2 minutes
> to run).
> MCSCF gives the values (which are correct, as far as I can tell):
> -760.334208045295 (1Sigma+, symmetry=1)
> -760.319940164809 (1Delta, symmetry=1)
> -760.319940164815 (1Delta, symmetry=4)
> MRCI also gives correct, reasonable values:
> -760.353785720431 (1Sigma+; symmetry=1)
> -760.337101857865 (1Delta; symmetry=1, state=2)
> -760.337101848033 (1Delta; symmetry=4)
> However FCI gives:
> -760.337267414097 (1Delta; symmetry=1)
> -760.269557607316 (????; symmetry=1)
> -760.337267433348 (1Delta; symmetry=4)
> In other words in the two-state FCI run the 1Sigma+ state has 
> disappeared.  Any idea as to what is going on?
> Thanks.
> Lorenzo
> <ScH.out>_______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user

Prof. Peter J. Knowles             
School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF10 3AT, UK
Telephone: +44 29 208 79182
Email: KnowlesPJ at Cardiff.ac.uk 

Yr Athro Peter J. Knowles
Yr Ysgol Cemeg, Prifysgol Caerdydd,
Prif Adeilad, Plas y Parc
Caerdydd CF10 3AT, DU
Ffôn: +44 2920 879182
Ebost: KnowlesPJ at Caerdydd.ac.uk
Ysgrifennwch ataf yn Gymraeg neu Saesneg

More information about the Molpro-user mailing list