Talk:Resonance (chemistry)

Distribution of electrons

I just corrected a bunch of statements in the lead which imply that resonance structures have different distributions of electrons. That is extremely poorly worded if not flatout wrong. Only our *depiction* of where the electrons reside (in Lewis structures) changes, not the location (density) of the electrons themselves. Alsosaid1987 (talk) 00:55, 11 May 2018 (UTC)

I have very carefully rewritten the second and third paragraphs of the lead to fix these issues. Also, there are some delicate issues with respect to logic and terminology. On the one hand, we rationalize the structure of a resonance hybrid based on the expected geometries of the individual Lewis structures and taking the "average". On the other hand, we later need to assert that contributing forms of a resonance hybrid do not differ in the geometry or overall electron density but are simply different representations of the real molecule.

Basically, we need to distinguish between standalone Lewis structures and Lewis structures that are part of resonance hybrids (i.e., contributing forms). I welcome changes that clarify this point. Alsosaid1987 (talk) 06:26, 11 May 2018 (UTC)

Thanks for working on this! Some of the key terms were defined several times. I also tightened up the idea that "resonance structures don't actually exist" vs "actuality is an average of them" (especially it felt like the "don't exist" detail was too buried). Hopefully I didn't change your meanings. But I do have a concern about the discussion of geometry. I removed the mention of it (other than bond order/length) where I was working because it was too tentative. Later (where I'm out of time now to consider editing), we talk about geometry...need to sort out when geometry is an average and when geometry is some sort of least-common-denominator. DMacks (talk) 08:26, 11 May 2018 (UTC)

Thank you for the changes. I will continue to tweak this (these changes should be considered preliminary). It is highly nontrivial to introduce resonance in a way that is correct and meaningful to beginners. Personally I like the purple mule = [red horse <-> blue ass] analogy as the clearest elementary explanation by analogy. Describing the true structure as being "intermediate" or "the average" of resonance forms feels somewhat less than satisfactory, but I was unable to think of a way to rigorize this notion.

What do you think about the question of whether singlet biradicals can be resonance forms of formally non-radical species? (I left a note in the appropriate section above.) I don't understand radical chemistry well enough, but it is my impression that as long as the spins pair up, it could be okay. Thus, I'm not sure that it's technically correct to say that all resonance structures need to have the same number of unpaired electrons. Alsosaid1987 (talk) 22:49, 11 May 2018 (UTC)

Example in lead

I think using NO2 as the example is problematic for several reasons. Like nitric oxide or triplet O2, nitrogen dioxide has a 2c3e bond due to its unpaired electron. However, Lewis structures are unable to correctly show the extra half bond, and averaging the structures gives an incorrectly low estimate of the bond order. Also, there are at least two other important Lewis structures that one can draw. Though charge separation is less favorable, they cannot be neglected when determining the bond order or structure. The bond angle is 134 degrees, which reflects the formation of the extra half-pi-bond, and the hybridization is somewhere between sp2 and sp, while the bond lengths are also shorter than expected. There is no simple way of estimating the bond order in this case (which is somewhere between 1.5 and 2). For these reasons, NO2 is really a pathological example that shows the limits of the Lewis representation. On the other hand, NO2-, nitrite is much more straightforward, and I tentatively chose this example to illustrate the concept. Unfortunately, it's hard to find a neutral example of resonance that is simple enough to give. Alsosaid1987 (talk) 03:17, 12 May 2018 (UTC)

Unwieldy struggle in explanation

@Dirac66:@DMacks:@Alsosaid1987: It has come to my attention that this article seems to be struggling a lot to explain that resonance structures do not exist but the resonance hybrid does (the very long introduction and repeated assertions in the article is testament to this). It has also occurred to me that nowhere in this article, except the valence bond (ie quantum mechanics) part explains what resonance truly is. Could I suggest an overhaul of this article in that we introduce in the outset that resonance is basically the equivalent of linear combination of atomic orbitals in Valence Bond theory, where the actual molecular wavefunction is a weighted sum of individual resonance structures just like molecular orbitals are a weighted sum of individual atomic orbitals (the analogy goes quite far in fact. We can even take antisymmetric, ie. antibonding, combinations of resonance structures to attain excited states etc).

I'm putting the above as an idea, not sure how to better write this article to be both concise and succinct. From reading this talk page it appears the writing of this article being problematic goes back some time. Opinions?--Officer781 (talk) 14:19, 19 January 2019 (UTC)

In principle, I agree. Essentially, resonance is taking linear combos of canonical structures, rather than taking linear combos of AO's in MO theory. But this may be tricky. I think this should be done if a relatively simple concrete example can be given. Ideally, this example should make clear that the true wavefunction \Psi is equal to \sum c_i\Psi_i, in which each \Psi_i corresponds to a canonical form. There needs to be some explanation of how that correspondence is made, as well as rules or at least some intuition for how to handle the contributors. (To give a completely rigorous introduction, one might need to start with the Heitler-London model for the bonding of H2, and rules for dealing with spin will also be introduced, but that seems impractical for this article.) The problem with VB theory is that, done correctly and rigorously, the rules are actually more complicated than the rules of MO theory. That's why, from a pedagogical point of view, VB is taught to undergrads in terms of qualitative heuristics for Lewis structures (and resulting in some phenomena that aren't adequately explained, like triplet O2 or aromaticity), while MO theory is relatively easy to teach in a semi-quantitative manner as a "more correct" theory than VB. Shaik and Hiberty wrote a very interesting review that both theories when fully implemented are correct and actually quite similar, but that they handle ionic contributions to bonding differently. The theories converge to each other when all approximations are taken away. Alsosaid1987 (talk) 15:52, 19 January 2019 (UTC)

If it's not already clear, I'm an experimentalist by training (though with a decent undergrad to beginning grad level pure math background), and I don't consider my background in theory to be strong enough to write an intro that is simultaneously clear, concise, rigorous, and introductory. Most of the intro is currently my writing, but I have no objections to others revamping it on a more rigorous theoretical basis. Alsosaid1987 (talk) 00:13, 20 January 2019 (UTC)

I have long been puzzled and unhappy about this article. It seems to be very much like the Irishman who when asked how to get to Dublin, responded by "If I was going to Dublin, I would not start from here". It needs a complete rewrite. However let me deal with "resonance is basically the equivalent of linear combination of atomic orbitals in Valence Bond theory" in Officer781's introduction to this section. I know of no source that suggests this. It is correct that valence bond theory deals with a linear combination of n-electron functions each built with little or no variation from atomic orbitals, while molecular orbitals theory introduces the variation at the atomic orbital level, but I do not think we should start the article off with this point, or indeed even have it in the article. Even what I wrote, "with little or no variation from atomic orbitals", is not strictly correct. In traditional VB theory we just build the total function from atomic orbitals, but in the spin coupled form of modern VB theory, the VB orbitals are built from a linear combination of atomic orbitals, just like in MO theory. However they turn to to be rather localized in contrast to the molecular orbitals which are fully delocalised. We should avoid questions of whether resonance structures exist. They no more exist than molecular orbitals, or even atomic orbitals, exist. However I am perhaps, as an active researcher in VB theory, too close to it all to be helpful. Perhaps the article should start with saying that Resonance is a simple way of describing the bonding in molecules, that it is based on valence bond theory rather than the molecular orbital theory, and that it is just one way or describing and understand bonding in molecules. It could go on to describe the Heitler-London picture of H2 and that move to adding ionic structures and then describe the Kekule structures (2 structures) of benzene, move on to add the Dewar structures (leading to 5 structures) and then mention that if we add ionic structures we add another 170 structures. I might have a go at an introduction later. --Bduke (talk) 00:55, 20 January 2019 (UTC)

You should give this a try. Expertise and accuracy is important on a commonly referenced article like this one. Wikipedia should be authoritative and not just repeat what textbooks say. Most people (myself included) reason using resonance structures, but only have a fuzzy understanding of what is going on physically behind the formalism. (Hiding behind the formalism and jargon just exposes ignorance and lack of understanding, to paraphrase Feynman....) I look forward to a clear, physically intuitive introduction! Alsosaid1987 (talk) 01:58, 20 January 2019 (UTC)

I'm going to try to shorten this article by first merging sections into a unified "overview" section instead (Alsosaid1987's overview is actually quite good in my opinion). I would probably need your help (Bduke, Alsosaid1987, etc) in the revamp. Do please edit if you have any ideas for revamping the article.--Officer781 (talk) 02:56, 20 January 2019 (UTC)

Three types of contributing structures

Just posting an idea here. This article currently talks about major and minor contributors but does not go into the types of resonance structures that can be drawn. As far as I know there are three types (doesn't just refer to aromatic molecules but to any molecule where contributing structures can be drawn):

• Kekule-type (standard)
• Dewar-type (long bond)
• Ionic-type (lone pair/octet deficient)

The Kekule-type structures are the ones currently covered in the article and in elementary discussions of resonance. I am busy currently and if anybody else would like to have a go at giving a brief mention of these in the major and minor contributors section can go ahead.--Officer781 (talk) 15:25, 23 January 2019 (UTC)

This is the simple level. Another way to distinguish different types of structure is to use what is called seniority (S) or ionicity (I). I will take benzene as an example. The Kekule and Dewar structures all have 6 electrons in 6 orbitals, so S=6 and I=0. The electrons are paired either just by neighbours (Kekule) or one pair across the molecule and two by neighbours (Dewar). An ionic structure with one +ve atom and one -ve atom would have S=5 (only 5 of the 6 orbitals are used) and I=1. If there are two -ve atoms and two +ve atoms, S=4 and I=2. If there are three -ve atoms and three +ve atoms, S=3 and I=3. Thus S = N - I, where N is the number of electrons, in this case 6. On this basis the full set of 175 resonance structures for benzene can be classified into 4 different seniorities or ionicities. I would add, although this is very advanced, that this assumes the Rumer pairing schemes. There are other pairing schemes that that do not distingish Kekule and Dewar structures, just as seniorities or ionicities do not. I do not think this is for the article, but it might be in a few years time. --Bduke (talk) 21:38, 23 January 2019 (UTC)

As usual we must confine ourselves to facts from external sources. The Kekule-ionic classification conforms to many organic chemistry texts, and the Dewar structures are often mentioned too although their contribution is minor in most molecules. I don't think though that we can say these are the only three possibilities unless someone can find a proof.

Seniority I have never heard of as a chemical bonding concept. This article [1] defines "seniority number" as the number of singly occupied orbitals in an orbital configuration, but I don't see that this definition corresponds to your values. As for ionicity, I would have assumed it means partial ionic character, going back to Pauling, but that would have fractional values. So if we want to put these two words into an article, we need clear definitions and sources. Dirac66 (talk) 00:16, 25 January 2019 (UTC)

You are quite correct. I must have been asleep when I wrote the above. Let me have another go at it. This is the simple level. Another way to distinguish different types of structure is to use what is called seniority (S) or ionicity (I). Seniority number is the number of singly occupied orbitals in an orbital configuration as the paper you reference from Wu's group indicates. Ionicity is the number of doubly occupied orbitals (yes, it does have other meanings). It refers to a resonance structure so it does not have fractional values. I will take benzene as an example. The Kekule and Dewar structures all have 6 electrons in 6 orbitals, so S=6 and I=0. This is correct. There are no doubly occupied orbitals so no ions. The electrons are paired either just by neighbours (Kekule) or one pair across the molecule and two by neighbours (Dewar). An ionic structure with one +ve atom and one -ve atom would have I=1. There are 4 singly occupied orbitals in 2 pairs so S=4. If there are two -ve atoms and two +ve atoms, S=2 and I=2. If there are three -ve atoms and three +ve atoms, S=0 and I=3. Thus S = N - 2I, where N is the number of electrons, in this case 6. On this basis the full set of 175 resonance structures for benzene can be classified into 4 different seniorities or ionicities. I do not think this is for the article, but it might be in a few years time. It is discussed in several papers by several independent authors but I do not yet know of a review article that discusses it. The fact that benzene has 175 resonance structures is well known. The use of S or I provides a pathway to go from the simple 5 resonance structures for benzene in stages up the full set of 175 structures. It also suggests that it is best to consider the Kekule and Dewar structures together which is the basis of the now well known spin coupled (SC) method [also known as the full generalized valence bond method (full GVB) or a recent compromise as spin coupled generalized valence bond (SCGVB)]. The 5 structures of SCGVB for benzene obtain over 90% of the energy difference between a molecular orbital calculation and the energy of the full 175 structures. These calculations are not new. They go back to the 1990s and have been discussed often in review articles. There are dozens, probably hundreds, of published papers using SCGVB and many book chapters and review articles. It should be mentioned somewhere in wikipedia but not in this article. --Bduke (talk) 02:36, 25 January 2019 (UTC)

We now have a short article on Generalized valence bond (GVB) but it contains nothing on spin coupling. And the redirect spin coupling points to J-coupling in NMR which is not what you want. So I think you could either add a Spin coupling section to the GVB article, or start a new article. Dirac66 (talk) 19:38, 26 January 2019 (UTC)

Ok. I will see what I can do. I will add it to Generalized valence bond. Note the method is called "spin coupled..", not "spin coupling...". --Bduke (talk) 22:10, 26 January 2019 (UTC)

Hyperconjugation

What is difference between resonance and hyperconjugation? Umer ilyas shaaheen (talk) 15:55, 26 January 2020 (UTC)

In organic chemistry, resonance involves mixing (conjugation) of structures which differ in the placement of pi-bonds. Hyperconjugation also involves structures with different sigma bonds or lone pairs. See the article on hyperconjugation. Dirac66 (talk) 22:13, 5 February 2020 (UTC)