Molcas Forum

lijingbai2009

Member

Registered: 2019-08-28

Posts: 29

CASPT2 energies are different with using "multistate" and "xmultistat"

Hi,

I just encountered a weird problem with CASPT2 energy.

When I used "Multistate":

&CASPT2
SHIFT
0.1
Multistate
all

I got state-spacific CAPST2:

::    CASPT2 Root  1     Total energy:   -308.51816796
::    CASPT2 Root  2     Total energy:   -308.27718471
::    CASPT2 Root  3     Total energy:   -308.26876705
::    CASPT2 Root  4     Total energy:   -308.26563793
::    CASPT2 Root  5     Total energy:   -308.14596830

and multistate CASPT2:

::    MS-CASPT2 Root  1     Total energy:   -308.51817576
::    MS-CASPT2 Root  2     Total energy:   -308.28300616
::    MS-CASPT2 Root  3     Total energy:   -308.27718476
::    MS-CASPT2 Root  4     Total energy:   -308.25139877
::    MS-CASPT2 Root  5     Total energy:   -308.14596049

But when I used "xMultistate":

&CASPT2
SHIFT
0.1
xMultistate
all

I got state-specific CASPT2:

::    CASPT2 Root  1     Total energy:   -308.53488080
::    CASPT2 Root  2     Total energy:   -308.28328209
::    CASPT2 Root  3     Total energy:   -308.25983207
::    CASPT2 Root  4     Total energy:   -308.30166598
::    CASPT2 Root  5     Total energy:   -308.38389292

and extended multistate CASPT2:

::    XMS-CASPT2 Root  1     Total energy:   -308.53491245
::    XMS-CASPT2 Root  2     Total energy:   -308.38386127
::    XMS-CASPT2 Root  3     Total energy:   -308.30428097
::    XMS-CASPT2 Root  4     Total energy:   -308.28328203
::    XMS-CASPT2 Root  5     Total energy:   -308.25721713

I am wondering why the "xMultistate" changed the state-specific CASPT2 energy?

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: CASPT2 energies are different with using "multistate" and "xmultistat"

That is not a problem, it is expected. Have you seen the note below the energies?

  Note that these CASPT2 energies are obtained using
  the XMS Fock operator and thus do not correspond
  to the true single-state CASPT2 ones.

lijingbai2009

Member

Registered: 2019-08-28

Posts: 29

Re: CASPT2 energies are different with using "multistate" and "xmultistat"

Sorry I haven't seen these notes in the output. Does it mean if I only want to compute single-state CASPT2 I shouldn't use "Multistate" after it?

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: CASPT2 energies are different with using "multistate" and "xmultistat"

Exactly. Or use "MULT" to select the states you want and "NOMULT" to disable the actual mixing and stop after the state-specific energies.