Dear Computational Chemists: Recently, our research group has determined one X-ray crystal structure with two different crystallographic units (dinuclear). One of the units with the disorder atoms and the other unit is accurate. We have also done their DFT studies, but after optimization, we are not getting the same geometry as in experimental data. So kindly enlighten us about this type of problem in metal complexes.
Thanks in advance
Dr. Rakesh Kumar
Dear Rakesh Kumar ,
As you might know, DFT studies for metal complexes are highly sensitive. Could you give us more information about the functionnals you used and the basis-sets ? What kind of pseudo-potentiel did you used also (if you took this into account) ? The problem can also comes from the spin multiplicity of your whole system, a non-proper spin multiplicity is known to "break" bond between two metals. Another problem can come from dispersion interactions: if your system is very crowded, the non-inclusion of a D3 method (developped by Grimme) could result in the separation of the two metals during optimization.
I hope I could have given you some lines of thought!
Dear Rakesh,
in addition to Corentin, I need to ask about the type of DFT geometry optimisation. Have you done this in the gas phase, or using the COSMO solvent model, or do you talk about calculations in the solid?
For the first two approaches (gas phase or solvent), DFT will never yield anything like disorder or two slightly different independent molecules. The DFT procedures "optimise" the geometry literally.
Only for distinct isomers, you can optimise different structures.
For shure molecular geometry\conformation calculations are wery sensitive to basis set and so on. From other hand in a crystal compounds under the influence of the "forces of the crystal field" like weak van der Waals, electrostatic, H-bonds and other interactions between molecules tend to pack as tightly as possible so that there are no voids in the crystal.
This often leads to some non-optimal energy of the molecules packed into a crystal in terms of its geometrical and conformational characteristics, but it gives a gain on the whole for the crystal! Because weak, but multiple intermolecular interactions easily cover small energy losses due to the non-optimal geometry or conformation of the molecule. And disordering in a crystal often appears due to a deficiency or, on the contrary, a certain excess of free volume during packing of molecules, which leads to several alternative positions of a substituent, groups, counterions, or a solvent(some time partial occupancies). For this reason, the quantum-chem.calculations do not always coincide with the experiment, and this must be clearly understood! :) Nevertheless, the calculations make it possible to clearly show the 'pure' influence of one or another type of substituent in one or another place of the molecule on the distribution of bond lengths and angles, as well as the conformation etc. So it is powerful tool but should be applied wisely!
Dear Rakesh Kumar
You must be optimized the hydrogen atoms and kept all other atoms invariant in the optimization process.
Dear Dr. Rakesh Kumar
I suppose you should consider solvent as a key factor in optimization process, and probably the cell you consider could be effective.
The method of optimization and solvent model are effective in optimization.
Regards
- Badges
- Science topic
- Similar topics
- Indexes