Molcas Forum

moabe

Member

Registered: 2018-06-03

Posts: 23

Active Space in CASSCF Calculations

Hi Everybody,

I have a problem in choosing the active space for CASSCF calculation. I'm going to investigate a ring-opening reaction in which a single C-C bond being breaking. So, it is necessary to include the corresponding sigma and sigma* orbitals as well as some pi and pi* orbitals in the active space. Based on the KS/DFT orbitals, I've designed my active space and for the first attempt I've selected 10 electrons in 10 orbitals and set 28 orbitals in inactive part and froze the core orbitals (my closed shell structure has 22 atoms and 94 electrons). But when I checked the orbitals after CASSCF orbital optimization in order to make sure that it has maintained the selected orbitals in DFT, surprisingly those sigma orbitals disappeared. I increased my active space from (10,10) to (12,12) and (14,14) but no sigma orbitals observed in the optimized CASSCF orbitals. Does anyone have any idea to solve this problem?

Thank you so much

Best,
Mostafa

niko

Member

From: Marseille

Registered: 2015-11-08

Posts: 59

Website

Re: Active Space in CASSCF Calculations

Hi, if you want to keep these orbitals in your active space, you need to significantly reduce the population in the sigma one and increase the population in the other one. One way to achieve this is to increase the C-C bond length. Alternatively, you may start from the transition state geometry obtained with a different method.

moabe

Member

Registered: 2018-06-03

Posts: 23

Re: Active Space in CASSCF Calculations

Hi. Thank you for your suggestion. Actually, my purpose is doing relaxed PE scan along the C-C bond which is breaking. To do so, I prefer to do a separate constraint optimization for each point along the PES using CASSCF method. Based on your suggestion, it seems scanning from the outside to the inside is the recommended procedure. Meaning that it would be better to start from the largest C-C bond length (can be TS geometry or even product) and end up with the equilibrium one. Have I understood correctly? If yes, should I check the active space for each point of PES and then start to do constraint optimization? I mean by approaching to the equilibrium geometry the bond is forming and the corresponding sigma orbitals must be included. How about using the converged CASSCF orbitals from the previous geometry step for the next one? 

Best,
Mostafa

Last edited by moabe (2018-06-06 12:34:10)

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Active Space in CASSCF Calculations

Another possible way to stabilize the active space is increasing the number of states in the state-averaging (CIRoot keyword).

niko

Member

From: Marseille

Registered: 2015-11-08

Posts: 59

Website

Re: Active Space in CASSCF Calculations

Yes, you always need to check the active space. Starting your PES cut at the TS geometry is, in my opinion, the best strategy.