CCL: Correctly evaluating spin states of a cobalt trimer (using Single Points)?

Dear Professor Neese, thank you for your kind reply and pointing me to your review; that was really helpful.

All the best.

On Sun, Sep 10, 2017 at 1:43 PM, Neese, Frank frank.neese * <owner-chemistry^> wrote:
Dear Henrique,

This is the kind of problem where you have to pay attention to the actual electronic structure and not just run calculations.

You have to pay attention to spin coupling. Co(III) is d6 system. Depending on the coordination environment, the most likely local spin states are Sl=0 or 1. The three local spins - if there are any - could couple to St=3,2,1 or perhaps 0. In a triangle spin frustration occurs and the coupling is most likely antiferromagnetic which would leave St=1 most likely (if Sl=1), but it could well be more complicated if there is any orbital degeneracy.

The point is that the lower spin states are not single determinants and hence just entering a given multiplicity in a DFT program and run with it is incorrect. You obtain broken symmetry determinants that are eigenfunctions to Sz but not S**2. Spin projection techniques must then be applied to estimate spin state energies. 

A multireference method is an alternative, but you have to work hard to overcome the self consistent field bias for high spin states and the underestimation of exchange coupling.

IMHO these problems are most easily solved experimentally. Connecting to the experiment (susceptibility, MCD, EPR ...) is vital in the first place.

Please allow me - My coordination chemistry reviews article from 2009 provides a somewhat detailed discussion. From a more conceptual point of view you may want to check a just published Angewandte Chemie Essay. A whole issue of coordination chemistry rev has recently been devoted to molecular magnetism - there generally  is a rich literature on this type of problem.

Good luck,

Von meinem iPhone gesendet

Am 10.09.2017 um 18:15 schrieb Sergio Emanuel Galembeck segalemb]-[ <owner-chemistry(~)>:

Dear Henrique,

I suggest that for each spin state you optimize the geometry. Some of this states could generate an unstable geometry. 

Best regards,


Prof. Sergio Emanuel Galembeck
Computational Quantum Chemistry Laboratory
Departamento de Química - FFCLRP-USP
14040-901 - Ribeirao Preto-SP

2017-09-10 9:03 GMT-03:00 Henrique C. S. Junior henriquecsj+/ <owner-chemistry~!>:

Dear colleagues, I’m working with a Cobalt(II) trimer whose molecular structure was achieved by Single Crystal X-Ray Diffraction. My task now is to check the spin states of the structure (High or Low spin). Since Co(II) can have 1 or 3 unpaired electrons, I’m approaching this problem by calculating Single Points for every possible multiplicity (10, 8, 6, 4, 2) and assuming that the most stable is the one that represents my structure (and my spin states).

Is this approach correct?

Thank you

Henrique C. S. Junior

Henrique C. S. Junior