Multiwfn forum

wham09

Member

Registered: 2025-09-19

Posts: 57

Judgement on DFT vs MR method selection

Dear Prof. Lu,

I'm studying spin-delocalized states. Specifically, I'm looking at a transition state of radical transfer (A-B + C_dot -> A_dot + B-C), 3-center-3-electron bonding situation.

For this, I have the following questions:

1. Is it correct that this situation inherently requires multireference methods, and normal DFT is absolutely not suitable?

2. Is it reasonable (or common) to optimize the geometries of the reactants and the TS by DFT, obtain thermal correction and dissolution free energies by DFT, evaluate the single point gas-phase energies by MRPT, and then sum everything to obtain free energies?

3. Do the answers to 1 and 2 change if the radical donor "A" is 3rd-row (prone to degeneracy) or 4/5th-row (less prone to degeneracy, but relativistic effect emerges) transition metal?

4. Do the answers change if what I want to study is transition state of excited-state homolysis of carbon-carbon, carbon-heteroatom, or main group atom - transition metal bond (as a part of a much bigger molecule)?

Thank you very much.

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,204

Website

Re: Judgement on DFT vs MR method selection

1 In most case DFT works reasonably. For example, H2+H• -> H• + H2, DFT can nicely represent the whole PES.

2 MRPT is not absolutely necessary, depending on the specific reaction.

3 If transition metal is directly involved in the process, multireference treatment may be necessary, but still not always.

4 When excited state is involved, the situation is significantly more complex; whenever possible, using a proper multireference method is recommended, especially when transition metal participates in the process.