Molcas Forum

abid

Member

Registered: 2015-11-12

Posts: 41

Can we run atomic calculations

Hi all
I have one question that can we run the rasscf calculations on one atom using molcas? I was running them but have a ground state with weight 50% along with 50% excited configurations. Please let me know if any of you have some idea
Thank you very much for your kind response

Last edited by abid (2016-10-18 15:00:36)

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Can we run atomic calculations

I have one question that can we run the rasscf calculations on one atom using molcas?

Sure

I was running them but have a ground state with weight 50% along with 50% excited configurations.

What do you mean with "ground state"? I presume you mean the "Hartree-Fock" configuration: 22...20...00. First, the main configuration depends on the orbitals, and an their order. And then, you could have a system with multiconfigurational nature, which cannot be properly represented by a single configuration, then of course you will have several configurations with important contributions.

abid

Member

Registered: 2015-11-12

Posts: 41

Re: Can we run atomic calculations

Ground state mean no excitations e.g. 2220002 if 8 electrons are active 2 electrons in RAS3 and 6 electrons in RAS2. RAS2=6orbitals RAS1=0;RAS3=1orbital

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Can we run atomic calculations

Well, that's a double excitation from RAS2 to RAS3, isn't it?

Anyway, my point is that you cannot identify a configuration only by its occupation/spin numbers, you need the orbitals too. A particular state may be represented by a single 222000 configuration with a set of orbitals, but if you change the orbitals (and a rotation of active orbitals is always allowed in CASSCF), the representation won't be 222000 any more, but some mixture of "excitations". In your case, I guess it simply means that your orbitals are no longer tailored for the ground state, or that the ground state has become significantly multiconfigurational (it cannot be expressed as a single Slater determinant, or CSF).