Basis sets for ions with paired/unpaired electrons
-
Hello.
If I want to run a calculation with ATOMSPIN 1 for an ion that has no unpaired electrons in its valence orbitals (as defined in its basis set), should I define a new basis set for it ?
For example, I have a basis set for Oxygen 2- that is closed shell, and I'd like to force crystal to see it as O1- from the very beginning.
Should I write a new basis set for it or can I simply just use ATOMSPIN with the same basis set as for O2-?
Thanks a lot!
Chris
-
Dear Chris,
Ultimately, it shouldn't matter. If the SCF goes well, you should arrive at the correct ground state regardless of the initial shell population (as long as it is neutral). The atomic configuration (neutral atoms or ions, see MgO input example in https://tutorials.crystalsolutions.eu/tutorial.html?td=barebone&tf=basic_tutorials#bs) in the initial SCF guess (default choice: density matrix as superposition of atomic densities) is defined by the atomic shells population.
And yes, with ATOMSPIN you set the initial spin on a specific atom (take care about symmetry and atoms that are translationally related) to set the starting point for the SCF.
If the SCF goes well, when you do the corresponding population analysis (say Mulliken), you should be able to distinguish the different "oxidation" states of different oxygen atoms and locate the unpaired electrons. I have done so with systems containing Cu1+ and Cu2+ in the lattice and this has worked beautifully (within the same basis set).
If you have a specific INPUT file in mind, you can upload it for cross-checking.
Cheers,
Aleks -
Hi Aleks, that's very helpful, thank you! I'll try that on my cells, cheers,
C.
