Broken symmetry question

[Sibo Lin]

I am attempting to simulate a dichromium compound with two bridging chloride ligands. Currently I am trying the fragment approach from the archived forum ( https://nwchemgit.github.io/Special_AWCforum/st/id1680/DFT_Broken-symmetry_for_molecule....html ). Each fragment converges properly but the combination of all the fragments never finishes one SCF cycle. One option I will try is simulating a smaller system (my current system has 444 atoms). 

I was also wondering, in the link above which discusses a diiron compound with two bridging sulfide ligands, why was the fragment containing those sulfides set as a singlet with 0 charge? I would expect each sulfide to be a dianion for overall -4 charge of the fragment. In my case with the 2 bridging chlorides, should I set that fragment as neutral and singlet as well?

Thank you!

[Edoardo Aprà]

The trick I used on the FeS molecule was to get the spin multiplicity as high as possible on the Fe fragments. This seems often a good strategy when using DFT.

Assigning a 0 charge on the sulfide ligands allows you to set a charge of -1 on the iron fragments (while a charge of -2 on S fragment  would require 0 on Fe fragments),  and the spin on the Fe fragments can be 6 (instead of 5).

Another solution for this tricky system with un-paired electrons could be the use of HF: you first converge using HF ("xc hfexch" in the DFT input section) and then use the molecular orbitals from HF as starting m.o.s for the DFT part.

dft

  xc hfexch

  vectors input .... output mymolecule_hf.mos

end

task dft

dft

 xc pbe0

 vectors input mymolecule_hf.mos output mymolecule_dft.mos

end

task dft

[Sibo Lin]

Thanks Edoardo, that did the trick! Setting the two bridging Cl to 0 charge in order give more electrons to the Cr fragments (4 unpaired spins each) allowed me to converge the wavefunction to the proper broken symmetry state (3 unpaired spins on each Cr atom). 

If anyone knows of a paper that uses such a procedure (minimizing bridging ligand charge to overexaggerate metal fragment spins for broken symmetry calculations), I would happy to cite it if my current research ends up being published.