I need some practical advise. What would be the easiest approach to compute strengthening or weakening of π-interaction in a bond with multipole-bond character. I'm not really fond of going through coefficients of Kohn–Sham orbitals. I did it on occasion, but it's a lot of work and I just do no have time. Unfortunately the molecules I need to compare have low symmetry ($\ce{C1}$) so I can't make use of symmetry to divide $\sigma$ and $\pi$ contributions. I'm not quite sure but I remember to read that the energy decomposition analysis in ADF could be used to divide in $\sigma$ and $\pi$ components. Let me make it clear, I'm not interested in comparing bond lengths, I need quantitative measure for the strength of $\pi$ interaction: either energy component or relative contribution of $\sigma$ and $\pi$ components orbitals.

I welcome any suggestions concerning methodology and/or software that can help.


First off, definitely check out ETS-NOCV, which I believe is a fairly recent addition to ADF which seems to do exactly what you're interested in. Other NBO-esque software may also suit your needs - I seem to recall NBO5 in Gaussian can generate sigma and pi natural bonds.

Regarding doing things the energy decomposition way:

If you have not already seen it, this paper goes about doing (in ADF) what you want to do:


(disclosure - I am a member of the Stranger group)

I did this for a related system a while back but have forgotten the exact details, however fragment bonding analysis is a nifty technique that allows estimation of the energy of single bonds in a molecule as well as decomposition by bond type.

Broadly, you cut your molecule into chunks on either side of the bond you're interested in and generate fragment orbitals from them. (that is, you use the TAPE21 files for these molecular fragments as a basis for the calculation of the full molecule). These fragment orbitals subsume all of the interactions within the fragments so that the total bonding energy only consists of interactions between the fragments.

From here, you can actually remove atomic orbitals from the fragment orbitals. If you prune the orbitals relevant to pi bonding, you can estimate the pi and sigma components by relative energy.

Here's where you should start: ADF manual - Analysis

  • $\begingroup$ Thanks. The ETS-NOCV sound interesting. I'll check the paper you cited. I was using the fragment approach but I was not aware that you can also make a decomposition by bond type. The part I do not understand is "you can actually remove atomic orbitals from the fragment orbitals". I think I need to do my homework (e.g. rtfm). $\endgroup$ – Kris_R Jul 4 '13 at 13:04
  • $\begingroup$ @Kris_R - Sorry, I believe what I said was incorrect - I'm pretty sure you remove molecular orbitals from the fragment orbitals. When I did this, I was looking at d to pi* donation to N2. This involved removing the relevant antibonding orbitals on the N2 fragment to preclude the interaction. I don't recall the specifics. $\endgroup$ – Richard Terrett Jul 5 '13 at 1:59
  • $\begingroup$ For clarity, NBO is NOT a quantitative analysis. It is merely qualitative. In fact, it would be tough to convince me that most of these energy decomposition schemes are quantitative. The best thing you can do is use a variety of methods in tandem but don't take NBO E2 values or whatever and treat them quantitatively. $\endgroup$ – LordStryker Oct 23 '13 at 12:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.