Molcas Forum
Support and discussions for Molcas and OpenMolcas users and developers
You are not logged in.
- Topics: Active | Unanswered
Announcement
Please note: The forum's URL has changed. The new URL is: https://molcasforum.univie.ac.at. Please update your bookmarks!
You can choose an avatar and change the default style by going to "Profile" → "Personality" or "Display".#1 2020-03-27 20:07:31
- MaxParadiz
- Member
- Registered: 2019-09-01
- Posts: 22
Large off-diagonal elements of the effective Hamiltonian (MS-CASPT2)
Hello all!
I am attempting to calculate the properties of a palladium complex at the MS-CASPT2(12,12)/cc-pVDZ level. I have a few different structures and each one with slightly different geometries (singlet and triplet energy minima). Most of them have worked fine, but there is a particular one that is causing me problems...
This is a singlet. The issue is that at the CASPT2, root 1 and 2 are at -6049.903 and -6049.795. This makes sense, and is consistent with what I expect from experiments and from the other related structures that I have calculated... However, after diagonalizing the MS-CASPT2 effective hamiltionian, the second root jumps to -6049.883, bringing it much closer to the ground state. This appears to happen because of the large coupling between states 2 and 3 in the effective hamiltonian, probably because root 2 and 3 are close in energy. How can I deal with a case like this? Is this where the IPEA shift comes into play? (I thought it came into play at the CASPT2 level rather than the MS-CASPT2). Do you have any recommendations?
Here is what I think is the relevant part of the output:
Total CASPT2 energies:
:: CASPT2 Root 1 Total energy: -6049.90345168
:: CASPT2 Root 2 Total energy: -6049.79491798
:: CASPT2 Root 3 Total energy: -6049.78288149
:: CASPT2 Root 4 Total energy: -6049.72339139
:: CASPT2 Root 5 Total energy: -6049.81991796
++ Multi-State CASPT2 section:
********************************************************************************
MULTI-STATE CASPT2 SECTION
--------------------------------------------------------------------------------
Output diagonal energies have been shifted. Add -6049.0000000000000
Effective Hamiltonian matrix (Symmetric):
1 2 3 4 5
1 -0.90345168
2 -0.00573095 -0.79491798
3 -0.00442572 -0.01066972 -0.78288149
4 -0.00132014 0.00504053 -0.00823764 -0.72339139
5 -0.01602290 -0.05273806 -0.06778662 -0.00845274 -0.81991796
Total MS-CASPT2 energies:
:: MS-CASPT2 Root 1 Total energy: -6049.91703090
:: MS-CASPT2 Root 2 Total energy: -6049.88267992
:: MS-CASPT2 Root 3 Total energy: -6049.78205273
:: MS-CASPT2 Root 4 Total energy: -6049.72334822
:: MS-CASPT2 Root 5 Total energy: -6049.71944873 Thank you!
Offline
#2 2020-03-27 21:47:48
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,012
Re: Large off-diagonal elements of the effective Hamiltonian (MS-CASPT2)
That's probably both why you need an MS treatment and why the MS treatment fails. It is not related to the IPEA shift. Try the XMS or XDW variants. Just an observation: you seem to have even larger couplings with state 5, you could also try removing it.
Offline
#3 2020-03-28 21:17:27
- MaxParadiz
- Member
- Registered: 2019-09-01
- Posts: 22
Re: Large off-diagonal elements of the effective Hamiltonian (MS-CASPT2)
Thank you Ignacio!
I had missed the larger coupling between states 2 and 5. I ran the same calculation using XMS, and state 5 again showed up as the problematic state, but the coupling between 2 and 3 no longer seems problematic. I am running it again with state 5 excluded, and I will a bit more into what this state looks like. I'll look into the literature to read about MS/XMS.
Final state energy(ies):
------------------------
:: RASSCF root number 1 Total energy: -6047.44911807
:: RASSCF root number 2 Total energy: -6047.31235600
:: RASSCF root number 3 Total energy: -6047.29240664
:: RASSCF root number 4 Total energy: -6047.25520005
:: RASSCF root number 5 Total energy: -6047.24669534
Total CASPT2 energies:
:: CASPT2 Root 1 Total energy: -6049.91885153
:: CASPT2 Root 2 Total energy: -6049.81017712
:: CASPT2 Root 3 Total energy: -6049.78220148
:: CASPT2 Root 4 Total energy: -6049.72778050
:: CASPT2 Root 5 Total energy: -6052.02225279
Note that these CASPT2 energies are obtained using
the XMS Fock operator and thus do not correspond
to the true single-state CASPT2 ones.
++ Multi-State CASPT2 section:
********************************************************************************
MULTI-STATE CASPT2 SECTION
--------------------------------------------------------------------------------
Output diagonal energies have been shifted. Add -6049.0000000000000
Effective Hamiltonian matrix (Symmetric):
1 2 3 4 5
1 -0.91868481
2 0.00508065 -0.80382557
3 -0.00100206 0.00138566 -0.78668172
4 0.00389778 -0.01372229 0.00639918 -0.72842068
5 -0.02997914 0.84666584 -0.45357456 -0.46274087 -3.02365064
Total XMS-CASPT2 energies:
:: XMS-CASPT2 Root 1 Total energy: -6052.44927122
:: XMS-CASPT2 Root 2 Total energy: -6049.91865381
:: XMS-CASPT2 Root 3 Total energy: -6049.78983742
:: XMS-CASPT2 Root 4 Total energy: -6049.75631855
:: XMS-CASPT2 Root 5 Total energy: -6049.34718243 Offline