Help with understanding results of Shubin Liu's EDA

wham09 · 2025-10-19 07:38:34

Dear Prof. Lu,

I have two questions about interpreting EDA-SBL results.

1. Since my background is mostly synthetic organic, I have a difficulty understanding the chemical meaning of Weiszacker energy (E_steric). From what I read in the manual, E_quantum is basically a Pauli repulsion term, so to me, E_steric and E_quantum seem to be both steric hindrance terms. What is the difference of E_steric and E_quantum in terms of chemical meaning (structure and/or reactivity)?

2. From ETS-NOCV analysis, we can get stabilization energy of each NOCV pair. Can this stabilization energy be categorized as one of the EDA-SBL terms (E_steric, E_electrostatic, or E_quantum)?

sobereva · 2025-10-19 18:39:51

Hello,

1 Neither the steric term nor quantum term in EDA-SBL fully correlate to steric effect in common sense.
To characterize steric effect between fragments, I would like to suggest calculating Pauli repulsion term in sobEDA energy decomposition method, see J. Phys. Chem. A 2023, 127, 7023−7035. sobEDA is very easy to use if you have Gaussian 16 and Multiwfn.

2 EDA-SBL and ETS-NOCV have very different theoretical framework, the former treats the system as a whole, while the latter focus on analyzing interfragment interactions, there is no any direct correlation between them.

wham09 · 2025-10-19 20:01:22

Thank you for the reply.

1. I actually really wanted to try sobEDA, but I only have access to g16 in a remote server, where I'm not allowed to install Multiwfn. Is there a way to run sobEDA in this situation? If not, can I get any insight on how to interpret E_steric and E_quantum in the EDA-SBL framework?

2. I understand that EDA-SBL and ETS-NOCV are very different, but I thought that E_T, E_J, E_Nuc, E_V, E_xc, etc. are general terms in DFT. Then, shouldn't the NOCV stabilization energy have some correlation with a certain general energy component of the whole system?

sobereva · 2025-10-20 23:28:48

1 You can install Multiwfn in your private user directory, you must have privilege. Also you can consider to run commands in sobEDA.sh manually, for all commands involving g16, run them on remote server, while run all other commands in Linux environment of your local machine.

2 The most important value of ETS-NOCV is decomposing E_orb term between fragments. However, EDA-SBL decomposes total energy of a system, there is no correlation between them.

wham09 · 2025-10-21 10:50:47

1. I looked into the sobEDA script file and concluded that the best option for me is using Multiwfn only to combine wavefunctions of fragments and generate promolecular & frozen input files, then run g16 jobs, and then tabulate the results in the l608 section myself. Could you confirm that this will work?

2. I now understand what you mean. Thank you very much!

sobereva · 2025-10-21 13:24:02

1 It should work

wham09 · 2025-10-22 07:53:36

Just an additional technical question about sobEDA:
I noticed that among the sobEDA tutorials, the template gjf files for metal-containing systems (chromium alkylidene and copper/H2O) contain an additional keyword IOp(3/174=1000000,3/175=2238200,3/177=452900,3/178=4655000), which I believe is additional dispersion adjustment. Should this keyword (with these exact numbers) be applied to any organometallic species?

sobereva · 2025-10-23 03:08:09

Yes, it should be applied to all systems involved in sobEDA calculation.

wham09 · 2025-10-23 03:49:36

I'm a little confused because in the 5 template.gjf files,
The templates for b3lyp, blyp, TPSSTPSS: Do not contain the iop keyword.
The template for BHandHLYP contains: IOp(3/174=1000000,3/175=1035400,3/177=279300,3/178=4961500)
The template for TPSSh contains: IOp(3/174=1000000,3/175=2238200,3/177=452900,3/178=4655000)

So, are these options specific to each functional? or a combination of functional/basis set?

sobereva · 2025-10-23 04:53:36

b3lyp, blyp, TPSSTPSS: Their D3 parameters are built-in in Gaussian, so do not need to customize
For BHandHLYP and TPSSh, their D3 parameters must be customized via respective IOps.

wham09 · 2025-10-25 16:36:31

Thank you for the clarification.

1. For sobEDAw, is it okay to use SDD (in combination with 6-31+G** or 6-311+G**)? I couldn't find any sobEDAw example in your paper and tutorial that deals with transition metals.
2. For both sobEDA and sobEDAw (actually, for any calculation), if I decided to use def2- family, would you recommend using a mixed basis set with SDD for TM-complex? Or use the def2- basis set solely?
3. If I want to do sobEDA or sobEDAw analysis using TPSSh/def2-TZVP, should I do the initial geometry optimization also using TPSSh/def2-family? I guess this question is also relevant to any calculation.

Last edited by wham09 (2025-10-25 16:53:33)

sobereva · 2025-10-26 04:07:36

1 sobEDAw was not parameterized for transition metals
2 Solely using def2 is good, note that accuracy of SDD pseudopotential basis set is worse than def2-TZVP
3 You can use any appropriate level to perform geometry optimization.

Multiwfn forum

#1 2025-10-19 07:38:34

Help with understanding results of Shubin Liu's EDA

#2 2025-10-19 18:39:51

Re: Help with understanding results of Shubin Liu's EDA

#3 2025-10-19 20:01:22

Re: Help with understanding results of Shubin Liu's EDA

#4 2025-10-20 23:28:48

Re: Help with understanding results of Shubin Liu's EDA

#5 2025-10-21 10:50:47

Re: Help with understanding results of Shubin Liu's EDA

#6 2025-10-21 13:24:02

Re: Help with understanding results of Shubin Liu's EDA

#7 2025-10-22 07:53:36

Re: Help with understanding results of Shubin Liu's EDA

#8 2025-10-23 03:08:09

Re: Help with understanding results of Shubin Liu's EDA

#9 2025-10-23 03:49:36

Re: Help with understanding results of Shubin Liu's EDA

#10 2025-10-23 04:53:36

Re: Help with understanding results of Shubin Liu's EDA

#11 2025-10-25 16:36:31

Re: Help with understanding results of Shubin Liu's EDA

#12 2025-10-26 04:07:36

Re: Help with understanding results of Shubin Liu's EDA

Board footer