Molcas Forum

Lukhmanul Hakeem k

Member

Registered: 2024-02-05

Posts: 25

SCF orbitals in Cs vs C1 symmetry: active space selection issue for CA

Dear OpenMolcas experts,

I am facing a difficulty regarding active space selection for CASSCF due to differences in SCF orbital outputs when running my calculations in different symmetries.

I am working with a molecule that has Cs symmetry (confirmed geometry). When I perform the SCF calculation using Cs symmetry, the number of occupied orbitals reported is 47. However, when I perform the same calculation in C1 symmetry (no symmetry), the number of occupied orbitals is 57, which correctly corresponds to the number of electrons in my system.

The problem is that:

In C1 symmetry, I can easily identify chemically meaningful orbitals (e.g. the π and π* orbitals) for selecting the active space.

In Cs symmetry, due to the symmetry constraints, the orbitals are separated into A′ and A″ irreps, and the π orbitals are often mixed or distributed differently, making it difficult to recognize and select the correct active space orbitals.

Because of this, I am unsure how to correctly select the active space for my CASSCF calculations when using symmetry.

So I am planning to do :

1. Perform an SCF calculation in C1 symmetry to obtain the full set of orbitals.

2. Analyze and select the correct chemically meaningful orbitals from the C1 calculation (π, π*, lone pairs, etc.).

3. Use these orbitals as the starting guess for a CASSCF calculation in Cs symmetry, by restarting from the orbital file (JobIph or Orbital file) generated from the C1 calculation.

Is this the correct approach?

Thank you very much for your time and support.

With kind regards and sincere respect,
Lukhmanul Hakeem K.

Lukhmanul Hakeem k

Member

Registered: 2024-02-05

Posts: 25

Re: SCF orbitals in Cs vs C1 symmetry: active space selection issue for CA

Dear OpenMolcas experts,

In addition to the issue described above, I have also opened the molcas.scf.molden file in Chemcraft to visualize the orbitals. However, I can only see the A′ orbitals; the A″ orbitals are not displayed.

Why is it that only A′ orbitals appear in the Molden file? Is this a limitation of how the Molden file is written during the SCF calculation in Cs symmetry, or is there a setting I need to adjust to include both A′ and A″ orbitals in the visualization?

Thank you very much for your time and support.

With kind regards and sincere respect,
Lukhmanul Hakeem K.

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,124

Re: SCF orbitals in Cs vs C1 symmetry: active space selection issue for CA

The orbitals are ordered by symmetry. First all A', occupied and virtual, and then all A'', occupied an virtual. Please provide a full reproducible example if you can't find them.

Lukhmanul Hakeem k

Member

Registered: 2024-02-05

Posts: 25

Re: SCF orbitals in Cs vs C1 symmetry: active space selection issue for CA

Dear Sir,

I have now loaded the molcas.scf.molden file into Multiwfn, where I can observe 57 occupied orbitals under the Cs symmetry. Based on this, I have selected the active space as (8,8). The selection includes 4 π electrons, specifically orbitals 54, 55, 56, and 57 (all of which belong to A" symmetry). For the virtual orbitals, I have chosen orbitals 58 and 59 (A' symmetry with σ* character), and 60 and 61 (A" symmetry with π* character).

Following this, I edited the MOLCAS_SCF.ScfOrb file for the next CASSCF calculation as shown below:

#INDEX 
* 1234567890
0 iiiiiiiiii
1 iiiiiiiiii
2 iiiiiiiiii
3 iiiiiiiiii
4 iiiiiii22s
5 ssssssssss
6 ssssssssss
7 ssssssssss
8 ssssssssss
9 ssssssssss
0 ssssssssss
1 ssssssssss
2 ssssssssss
3 ssssssssss
4 ssssssssss
5 ssssssssss
6 ssssssssss
7 ssssssssss
8 ssssssssss
9 ssssssssss
0 ssssss
* 1234567890
0 iiiiii2222
1 22ssssssss
2 ssssssssss
3 ssssssssss
4 ssssssssss
5 ssssssssss
6 ssssssssss
7 ssss

Could you kindly confirm whether this selection and modification are correct for proceeding with the CASSCF calculation? Your guidance and suggestions would be highly appreciated.

Thank you very much for your time and support.

With kind regards and sincere respect,
Lukhmanul Hakeem K.

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,124

Re: SCF orbitals in Cs vs C1 symmetry: active space selection issue for CA

That selection puts orbitals A' 48,49 and A'' 7,8,9,10,11,12 in the active space.

I advice you to use Pegamoid (https://pypi.org/project/Pegamoid/) and .h5 files instead of molden files if possible.

Lukhmanul Hakeem k

Member

Registered: 2024-02-05

Posts: 25

Re: SCF orbitals in Cs vs C1 symmetry: active space selection issue for CA

Dear Sir,

Thank you for your response. I have tried using Pegamoid, and it is very useful for me. Thank you for suggesting it.

With kind regards and sincere respect,
Lukhmanul Hakeem K.