Talk:Resonance (chemistry)

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Chemistry (Rated C-class, High-importance)
WikiProject icon This article is within the scope of WikiProject Chemistry, a collaborative effort to improve the coverage of chemistry on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.

Random comments at the top[edit]

At the end of the article, it says that resonance diagrams are soon to come. That was 6 months ago. Is anybody going to be adding resonance diagrams soon here? H Padleckas 16:17, 25 Jan 2005 (UTC)

You can steal some from Benzene_ring. - Omegatron 16:47, Jan 25, 2005 (UTC)

They don't really switch between the two possible states constantly, do they? Then the electrons would be net going in a circle and would create a magnetic field. Rather, they exist in a less-than particle state or a "cloud" or something, right? - Omegatron 16:47, Jan 25, 2005 (UTC)

No, resonance structures are not compounds that switch between the two (or more) possible forms constantly. Those two kinds of forms of a compound which readily switch back and forth are called tautomers. Nuclear magnetic resonance studies have provided a lot of information on various tautomer pairs, including how fast they switch back and forth depending on temperature. In a compound whose chemical structure can be represented by two or more resonance structures, the actual "constant" structure or condition of the molecule can be thought of as an "average" or hybrid of the its resonance structures, although this average is not always evenly weighted among its resonance structures. I have never heard and I have no reason to believe that the electrons in aromatic molecules move around in a circle creating a magnetic field. Most likely, whatever motions the electrons in an aromatic molecule engage in are more likely to be random, or random to the extent that we do not understand their order. Maybe these aromatic electrons do exist in a less-than particle state or a "cloud" or something as perhaps suggested by the Heisenberg uncertainty principle. However, if you're a scientific guy, maybe you design an experiment to look for these kinds of magnetic fields in aromatic molecules or compounds.  :-)
H Padleckas 10:42, 26 Jan 2005 (UTC)

Well, on a large scale they would swamp each other out, as they are all facing different random directions. I didn't think they really circulated around like that, but I wonder what effect it does have on the magnetic properties of these materials... I know that carbon nanofoam exhibits weak ferromagnetism. I wonder if they're related. I wish there was enough time to learn every subject in depth... - Omegatron 15:00, Jan 26, 2005 (UTC)
Actually if you apply magnetic field (such as in the aforementioned NMR), you can experience Aromatic ring current. Cubbi 22:09, 20 April 2007 (UTC)
The word I've always used for electrons in a resonant system is delocalised. ben.c.roberts

Resonant bonds[edit]

Has anyone actually heard of the term "resonant bonds"? I've look in texts before, and while I've definately heard of resonance structures, "resonant bonds" just doesn't seem standard.

The use of verb forms of "resonance" died out with Pauling; even his student Wheland eschewed them. "Resonant" is like a participial form that is not used because it implies that resonance is a process. Rather, resonance is a (rather clunky) description of a variety of chemical bonding involving electrons dispersed over more than two atoms.Rectifico (talk) 19:27, 18 June 2010 (UTC)
Electrons are continuously whizzing around, so you could say a compound with resonance has an infinite number of contributing structures, with the major ones most often occurring. In that sense, resonance may be considered as a process and the term (defined broader than merely the presence of different Lewis formulas) is not so bad as many think.--Wickey-nl (talk) 10:02, 19 June 2010 (UTC)


There's a merge tag over at benzene. That section on resonance would look mighty good here. Isopropyl 21:08, 11 April 2006 (UTC)

I agree that some of that material should be added here, but it should also be retained in the benzene article. Resonance is a crucial part of the benzene story and readers should not have to come over here to read about it. --Bduke 21:27, 11 April 2006 (UTC)

Maintain separate articles, sharing whichever sections would improve the other. They are sufficiently separate entities. -Ayeroxor 21:41, 1 May 2006 (UTC)

Sigma bonds[edit]

The article states that sigma bonds are meaningless in MO theory for benzene. This is strictly accurate but goes against useage. Pi orbitals in benzene are also meaningless strictly. The terms sigms and pi strictly apply only to linear molecules giving the spectroscopic terms for angular momentum about the molecular axis. We can talk therefore about sigma and pi molecular orbitals even in long linear molecules where the MOs are delocalised. For planar congugated systems such as benzene the useage has come to mean how the MOs change on reflection in the molecular plane. If it changes sign, it is pi. If it does not change sign it is sigma. The term "sigma-pi separation" is very common in both MO and VB discussions of benzene, so the current wording could serious confuse people. I'm not clear how to rewrite the section. The section is:-

"Often when describing benzene the VB picture and the MO picture are intermixed, talking both about sigma 'bonds' (a meaningless concept in MO) and 'delocalized' pi electrons (a meaningless concept in VB). This is not a good practice, because mathematically the models are incompatible."

Any body got any ideas? --Bduke 22:24, 18 December 2006 (UTC)

MO has sigma orbitals, but no sigma bonds. A bond implies that an electron pair belongs to two and only two nucleus, whereas in MO the electron state vectors are calculated with the )approximate) potential of the entire molecule and not any two particular atoms. That is why in MO it does not make sense to say that every carbon has 1 sigma bond with each of its neighbours. At best we could say that these electrons are in highly localised orbitals. As is indicated in the article, sigma-pi in VB is somewhat different from sigma-pi in MO. Loom91 07:07, 19 December 2006 (UTC)

The problem is that sigma orbitals and sigma bonds do not appear that different and indeed they are not as the delocalised MOs can be transformed into localised orbitals. This means that sigma bonds are not a meaningless concept in MO although I agree that sigma in MO is not exactly the same as sigma in VB. I note also that most quantitative VB treatments of benzene such as the famous spin coupled calculations of Gerratt, Cooper and Raimondi use VB only for the pi electrons and treat the sigma as MOs. It is true that MOs are calculated with the potential of the entire molecule, but so are the VB wave functions. It is the orbitals that are delocalised or localised. The potential is always the whole molecule included the repulsion of all other electrons. We are not too far apart but I think the current wording will lead to misunderstandings and there is already too much misunderstanding in the MO/VB debate. --Bduke 07:46, 19 December 2006 (UTC)

The point made in the article is that VB and MO are two different mathematical models with different approaches to solving a problem. The sigma-orbital approach in MO and the sigma-bond approach in VB give the same predicted expectation of observables (otherwise one theory would be more 'wrong' than the other), but that doesn't mean that the theories are the same and you can go around taking a slice from VB and a scoop from MO. Also, MO is a more 'natural' theory than VB. The postulates or methodology of MO are closer to the actual quantum mechanical picture, which exists exactly only for the simplest molecules like H2+. Loom91 09:51, 20 December 2006 (UTC)

There are two major problems with what you say. (1) MO and VB do NOT give the same predicted expectation values of observables and indeed one is more wrong than the other. I never said they are the same. They are different. (2) MO is not more natural than VB. Just because the exact solution of H2+ (a one electron problem) is a MO does not mean that MO treats electron repulsion in cases with more than one electron correctly. Both are approximations. MO is much easier to do and is very popular. Computationally a lot of methods build on it or add to it. However in fact simple VB energies almost always lie below simple MO energies (and hence are better) and MO does not dissociate homonulcear diatomic molecules correctly. VB is less wrong. I would also add that it may seems odd to take "a slice from VB and a scoop from MO" but that is exactly what a lot of people do, including as I said above, some of the best VB calculations on benzene. Have you studied quantum chemistry to any depth? Of course this article is not the place for advanced quantum chemistry, but we must try to avoid too many "lies for children" as we simplify the quantum chemistry. --Bduke 10:16, 20 December 2006 (UTC)

I think the confusion is resulting from talking about VB and MO as theories in the ordinary sense. I haven't studied VB in detail, but as far as I know they are two different approaches to approximating the solution, not methods of approximation as such. As you say, computational methods build on them. There are different levels of approximation in VB and different levels of approximation in MO (you could start with Huckel and proceed to DFT). If a certain level of approximation in VB gives a result, MO will also give similar or better results at some level of approximation. You say simple VB energies lie below simple MO energies, but we can't decide superiority between the two simply by looking at the 'simple' limit.

The framework provided by MO (or VB) can not be identified with any particular method or level of approximation in either. The MO road is closer to what a physicist will do if asked to solve a system, but that does not mean the VB road will not give the same results at sufficient approximation. The converse is also true. You claim "MO does not dissociate homonuclear diatomic molecules correctly"-I've never read this, but VB fails to predict that hybridisation in CH4 does not produce equivalent orbitals. VB is not more correct than MO.

As for mixing MO and VB, you are talking about mathematical calculations where we have quantitative ways of verifying whether we are mixing in a valid manner. The article however talks about mixing the two in qualitative descriptions ("lies for children"), where there is no way to know (without having actually done the calculations) whether the pictorial descriptions reflect the actual mathematics. It is such situations that trying to mix the two models is dangerous. Also, I don't think that this article is not a place for advanced quantum chemistry. Wikipedia is a specialised as well as general purpose encyclopedia, so feel free to add a section on the detailed mathematics behind the two-headed arrows. It will be best if you cite references. Loom91 07:48, 21 December 2006 (UTC)

You state that you have not studied VB in detail. Sorry, but I'm afraid that is pretty clear. Let us take the dissociation of MO theory for homonuclear diatomics first. The simple MO wavefunction is an equal mixture at all values of the bond length of the simple VB covalent and ionic terms, so it dissociates into an equal mixture of 2 H atoms and (H+ + H-). The energy is the mean at infinite bond length of these two and thus very much higher in energy than 2 H atoms which H2 experimentally dissociates into. VB dissociates correctly. The predicted MO dissociation energy is too large. The predicted VB dissocaiation energy is closer to experiment but two small because the two atoms at large distance are exact and the VB energy lies above the exact energy at the predicted bond length. The MO energy lies even higher. These facts matter because both methods are variational - they lie above the exact result. The simple Heitler-London VB lies below the simple MO energy curve and is thus better. The similar case of F2 is so bad that the MO energy lies higher than the correct dissociation of 2 F atoms. This is in many texts. By MO I mean for H2 a single molecular orbital built as a linear combination of atomic orbitals (with simple MO using just the two 1s orbitals). Of course one can add configuration interaction and do as well as VB. Adding all possible excited configurations to MO and all possible VB structures to the simple VB gives identical results but such calculations, called full configuration interaction, are only possible for fairly small molecules with fairly small basis sets. "VB fails to predict that hybridisation in CH4 does not produce equivalent orbitals" - please explain more carefully or give a source. Hybridisation is not a physical think. It is an artifact of VB theory and the hybrids can and are closen to be equivalent. DFT is not MO although I grant you it looks like it. VB can easily do better than the best MO wavefunction - called the Hartree-Fock limit. It does exactly because it get bond breaking better.

The bottom line is that the simple pictures grew out of calculations. Pauling would have done nothing without the Heitler-London calculation on H2 and some extensions of it. We forget this at our peril. Qualitative ideas did not grow out of the air. "Also, I don't think that this article is not a place for advanced quantum chemistry". I think you meant "Also, I do think that this article is not a place for advanced quantum chemistry". I agree. It is about the simple pictures but we must not mislead. I'll think about it more after the holiday period which is going to be very busy for me. --Bduke 08:42, 21 December 2006 (UTC)

What you are saying is not that VB is a more accurate theory than MO, but that VB is computationally less intensive. As for hybridisation, the 4 CH bonds are not exactly equivalent as calculated in VB. As predicted by MO, one of the bonding pairs have a different energy from the other three. The photoelectron spectra shows two characteristic bands [1]. Also, why are you excluding post-Hartree-Fock methods from the umbrella of MO? And I meant exactly what I said: "Also, I don't think that this article is not a place for advanced quantum chemistry." That is, this article IS a place for advanced quantum chemistry. An accurate section on what exactly is mathematically meant by resonance will add greatly to the article. It may even become a GA. So feel free to add such a section. Loom91 07:12, 22 December 2006 (UTC)

No, I am saying that VB is more accurate at equivalent levels than MO, but that MO is computationally less intensive. The orbitals in MO are orthogonal and that simplifies things. In VB they are not orthogonal and it has taken a long time to get code that competes with MO. Why am I excluding post-Hartree-Fock (HF) methods? Because while based on a MO reference function they are not MO. Configuration interaction at the full level is entirely equivalent to full VB, so no comparision is fruitfull. Bond breaking is still badly handled by post-HF that uses a single determinant reference. To handle bond breaking correctly the MO guys use multi-configuration SCF where you can no longer say that 2n electrons are in n MOs for a closed shell singlet. Also it can be shown that these methods are very similar and in some cases identical to some spin-coupled VB methods, so again comparision is not fruitfull. The only meaningfull comparision is between the methods Pauling and Mulliken and their respective supporters fought about in the 1930s, for example simple MO for H2 - (a + b)(1)(a + b)(2) and Heitler London for H2 - a(1)b(2) + b(1)a(2) where a and b are the 2 is orbitals on the 2 H atoms. In passing note that expanding out the MO function, you get a(1)b(2) + b(1)a(2) + a(1)a(2) + b(1)b(2). The first 2 terms are the Heitler London terms and the last 2 are ionic terms - H- H+ and H+ H- so as I said earlier MO is a mixture of the VB covalent term that dissociates in 2 H atoms and the VB ionic terms that dissociate into two ions at a higher energy.

I have no idea where you have got the idea that the 4 CH bonds are not equivalent in VB. They are. The photoelectron spectrum with 2 peaks is best explained by the fact that there are only 2 energy-distinct MOs - 1 triply degenerate group and a single degenerate one for the 4 pairs of valence electrons. Ionisation is certainly best explained by MO theory because the electron does not leave one bond but the whole molecule. A VB description of CH4+ would have to include resonance between the 4 structures each with 3 two electron bonds and 1 one electron bond. In this way the ion would come out symmetric and there are indeed 2 solutions just as in MO theory. Yes, MO theory is simpler to describe ionisation and spectroscopy. VB can be simpler to describe bonding and generally gives better numbers. Getting numbers to agree with your PE spectra from MO theory is not easy, but the simple picture is. The orbital energies, for example, will only predict the position of your peaks well, using Koopman's approximation, if the massive correlation energy corrections and relaxation energy corrections are of opposite sign and similar magnitude which they often are for organic molecules but rarely are for metal complexes. --Bduke 08:12, 22 December 2006 (UTC)

What do you mean by equivalent levels? How would you say a particular VB method is equivalent to a particular MO method? Loom91 18:07, 23 December 2006 (UTC)

I assume you are asking about my statement that, for example, full CI is equivalent to full VB. OK, let us take the classic example of H2. The simple MO is (a + b)(1)(a + b)(2) which expands to:-

(a + b)(1)(a + b)(2) = a(1)b(2) + b(1)a(2) + a(1)a(2) + b(1)b(2) as above.

The excited state with both electrons in the antibonding orbital is:-

(a - b)(1)(a - b)(2) = a(1)b(2) + b(1)a(2) - a(1)a(2) - b(1)b(2)

Now mix these and collect terms (K is the mixing weight):-

(a + b)(1)(a + b)(2) + K {(a - b)(1)(a - b)(2)} =
(1+K){a(1)b(2) + b(1)a(2)} + {1-K){a(1)a(2) + b(1)b(2)}

The above is the full CI result for this small basis set of 2 1s orbitals. The full VB is ionic - covalent resonance, which is (C is the mixing coefficient):-

{a(1)b(2) + b(1)a(2)} + C{a(1)a(2) + b(1)b(2)}

Neither of these are normalised. In both cases the mixing coefficent is determined by finding the value that minimises the energy. Since both allow any proportion of the covalent - {a(1)b(2) + b(1)a(2)} - and ionic terms - {a(1)a(2) + b(1)b(2)}, the final results will be the same. This result is general. If we take a simple MO and mix in all possible excitatations that mix with the ground state, and then take all possible VB structures from the same set of atomic orbitals, the results are equivalent. The general result is perhaps surprising - approximations that look very different and start from different ideas, can actually be completely identical.

To our other readers, I apologise. This is getting over complicated and technical. Loom91, if you want to continue this, please move it to e-mail. I have e-mail set from my user page. I am happy to continue helping you to learn about VB theory, but I think the discussion is getting beyond relevence to this article. --Bduke 21:42, 23 December 2006 (UTC)

You misunderstand me. I know VB == MO in the high accuracy limit. I was asking in lower accuracy levels how you say that VB is more accurate than 'equivalent' MO. As for the article, what changes do you propose? Loom91 07:02, 27 December 2006 (UTC)

Let us take H2. The simplest MO approach just using the two hydrogen atom like 1s orbitals is as above. The simplest VB using the same orbitals is the Heitler-London. These are at an equivalent level, yet give different results. The latter lies lower in energy than the former at all interatomic distances and particularly at large distance and so is better. We can then optimise the orbital exponent of the 2s orbital in both cases. These are at equivalent levels. Again VB is better. That is what I mean by equivalent - same basis set and simplest possible MO or VB approach or comparable improvements to simplest approach.

I have made the changes to the article that I think should be made. My reasons are many. First, it is quite common to mix MO and VB ideas. Coulson in both "Valence" and in McWeeny's "Coulson's Valence" says this about the sigma bonds in benzene, "These bonds can be described either in MO or VB language; their essential character is the same in either case". I know of no book that criticises this statement. He goes on to give the VB and MO approaches. This mixing of language is commonly done in simple qualitative explanations and a mixing of methods is commonly done in quantitative calculations as I mention above. I do not think sigma is "meaningless in MO" or delocalised is "meaningless in VB". I do agree it is best to use "delocalisation energy" in MO descriptions, but note that somewhere on WP is a reference to a Journal of Chemical Education article that recommends delocalisation rather than resonance for VB descriptions. I also suggest it is stretching it to say about the two methods that "mathematically the models are incompatible". Different, yes, but not incompatible. To say, for example that MO for H2 is entirely identical to VB resonance between the covalent and ionic structures, but with equal weights, demonstrates this lack of incompatability. The article is best made simpler at this point. The wording was confusing and not clarifying matters, so is best removed. --Bduke 02:10, 28 December 2006 (UTC)

Reasonance diagrams for heteroaromtic compounds?[edit]

Reasonance diagrams for heteroaromtic compounds would be nice -- Quantockgoblin 23:47, 20 March 2007 (UTC)

{{Chemical drawing needed}}

It does not seem useful to clutter the article with more examples. Heteroaromatic compounds present no particular challenge in drawing resonance diagrams. For example in furan, you simply move the lone pair on the O atom through the other 4 carbon atoms. Resonance diagrams for homoaromatic compounds would be more interesting. Loom91 08:35, 21 March 2007 (UTC)
The set of resonance structures of Furan is illustrated at this time, so I'm calling this request Yes check.svg Done. DMacks (talk) 04:34, 4 June 2012 (UTC)

Strange definition in introduction[edit]

Just out of interest, why is resonance referred to as "a tool used (predominantly in organic chemistry) to represent certain types of molecular structures."? , it's a bit of a LARGE generalisation; aside from being referred to as canonical forms, Miessler refers to it as when there is "more than one possible way in which valence electrons can be placed in a lewis[-based] structure.", Chambers refers to it as "when [a] true structure of .. [a molecule or compound] cannot be accurately represented by a single structure, ... several resonance structures are suggested." Gaim.svg ♥♥ ΜÏΠЄSΓRΘΠ€ ♥♥ slurp me! 10:40, 6 April 2007 (UTC)

The sentence you refer to simply identifies what sort of thing we propose to discuss. It does not try to give a definition of resonance. Loom91 19:11, 8 April 2007 (UTC)
Well, with that said -- wiki articles aren't for proposing anything, but rather reporting it impartially. For instance;
"It is also not right to say that resonance occurs because electrons "flow" or change their place within the molecules. Such a thing would produce a magnetic field that is not observed in reality."
Although correct, it has a more "teaching" than "telling" approach. Gaim.svg ♥♥ ΜÏΠЄSΓRΘΠ€ ♥♥ slurp me! 13:15, 9 April 2007 (UTC)
Referencing that statement is rather trivial, any chemistry textbook will do. An actual definition of resonance is neither useful nor easy to give. Would you like a sentence in the introduction saying "Resonance in the context of chemical structure may be defined as the method of approximating an actual state vector of a molecule as a linear combination of (not necessarily complete) basis state vectors representing states where the molecule only contains localised bonds, where bond is taken in the meaning of Valence Bond Theory"? Loom91 06:53, 10 April 2007 (UTC)
No, but i'll surmise what i mean. I'll withdraw my objection to the introduction so long as a generic definition of resonance structures is there. Gaim.svg ♥♥ ΜÏΠЄSΓRΘΠ€ ♥♥ slurp me! 10:23, 11 April 2007 (UTC)

Resonance hybrid stability[edit]

To clarify this point "Resonance hybrids are always more stable than any of the canonical structures", the wave function Ψ is given by:

Ψ = a1C1 + a2C2 + a3C3 + ..

where Ψ is the resonance hybrid function and C1, C2, .. are the canonical structure functions. a1, a2 , a3, .. are coefficients chosen to minimise the energy. It follows from the variation theorem that the energy of Ψ is less than or equal to the energy of all of C1, C2, .. taken separately. It would be equal if one of the a1, a2, .. was 1 and all the others zero, and lower otherwise. Loom91 is correct and he gives a good simple reference. In antiaromaticity, the geometry changes to a more stable form. A good discussion is chapter 4 of "Facts and Theories of Aromaticity" by David Lewis and David Peters, Macmillan, 1975. --Bduke 13:37, 12 April 2007 (UTC)

Convoluted statement in section "Writing resonance structures" should be rewritten[edit]

When separating charge (giving rise to ions), usually structures where negative charges are on less electronegative elements have little contribution, but this may not be true if additional bonds are gained.

I believe this statement should be rewritten. Not being a native english speaker, my opinion may be misguided; anyway, to me it looks convoluted and is nearly uncomprehensible. A statement like this, being a list item, should speak for itself. It remains unclear however, to what phelomenon contribution should contribute.
Bertus van Heusden 10:53, 4 September 2007 (UTC)

kcal vs kJ[edit]

I see someone changed the units under "resonance energy" to kcal/mol without changing the numeric values. These should definitely stay as SI units, but someone should now check the correct values. Unfortunately I don't have time right now. --Slashme 06:53, 10 October 2007 (UTC)

The numbers are correct in kcal/mol (the resonance energy of benzene from this type of analysis is closer to 36 kcal/mol, not to 36 kJ/mol, which would be less than 9 kcal/mol, way too low). I don't think that the units should be converted to SI exclusively, although I would recommend giving the numbers in both systems. In my experience, kcal/mol are used more often in physical organic chemistry than kJ/mol, although both systems are commonly seen, and there is some regional variation. --Itub 10:02, 10 October 2007 (UTC)

You have a point, it's probably a good idea to keep the kcals, but a quick browse through the literature shows that kJ is gaining ground. I can't yet find a good ref. for the values quoted, because my chem. books are at the lab. If I get around to it, I'll sort it out, but I might not... --Slashme 13:21, 10 October 2007 (UTC)

The original estimates were in kcal, probably in the era of Pauling, but kJ are now preferred. So yes, we need both units, so I have now inserted kJ values. I just multiplied the kcal values by 4.184 and rounded off to two figures. Dirac66 13:33, 10 October 2007 (UTC)

terminology, physical reason[edit]

To me, as a physicist, there are a couple of problems with this article. One is that the article never really explains the reason for the term "resonance," and there is no obvious (to me) physical phenomenon going on that is in any way (that's obvious to me) analogous to resonance. The section near the end about Pauling's introduction of the term doesn't really explain anything very clearly: why the quantum-mechanical treatment of H2+ was relevant to Pauling, or why he used the word "resonance." The other problem IMO is that the article never gives any very transparent physical explanation of what's going on. Although I understand the general argument made above on the talk page that a superposition of trial wavefunctions can be optimized variationally to lower its energy, that argument is so generic that it really has nothing in particular to do with chemical bonds, or even chemistry. If I had to take a stab at it, I would guess that the general physical mechanism is that, compared to a structure made of single and double bonds, the actual structure delocalizes the electrons, which means that they have a larger wavelength, thus a lower momentum and kinetic energy. In the case of an ion like CH3CO2-, I can also imagine that the delocalization would lead to a lower Coulomb energy.-- (talk) 23:56, 11 December 2007 (UTC)

I think your puzzlement is very understandable, and your inferences are perceptive and correct, as far as I can judge. The term "resonance" was picked up by Pauling from Heisenberg, who noted that the two electrons in an excited helium atom had to be treated by the mathematics used for coupled pendulums, said to be in resonance. This was then used by Pauling to describe all chemical bonding, although by now it is applied only to cases like those in this article. Some of us believe that this would be a good case for the old New Yorker column, "Block that Metaphor".

Rectifico (talk) 19:27, 18 June 2010 (UTC)

Hmm...okay, I think I understand the reason for the term now. See . The paragraph beginning "Resonance: In attempting to explain ..." seems to be saying clearly what the WP article is saying unclearly. I'd take a whack at it myself, but I'm not a chemist, so I don't want to get this wrong.-- (talk) 00:08, 12 December 2007 (UTC)

Problems with this article[edit]

I know I'm coming in late into this discussion, but ...

There are several small problems:

  • Use of 'pi' instead of π.
  • I am no expert on the history of the concept, but I doubt very much that Kekulé came up with resonance, as is implied in the 'Examples' section.
  • Whereas ozone is correctly presented as bent and benzene as a hexagon, the allylic cation is linear (?) and the middle H changes position, contrary to rule 1 governing the correct drawing of resonance structures.

The biggest problem I find with this article is that it is not well targeted to the intended audience.

After all, who will look up 'Resonance' on Wikipedia? The average Joe wouldn't ever. Practitioners don't need to. Only students would -- students who have encountered resonance in their studies and in their textbook, and who need help in understanding it. Their first encounter would be with main group examples (e.g. sulfate ion), but this article barely acknowledges this reality. My evaluation is that this article would confuse more than enlighten.

Let me be clear: I can live with this article, because I see no substantial problem with the content, but I would never refer any of my students to it. It's simply not written for them and assumes knowledge.

Specific problems:

  • Resonance as a theory and key component of VB theory vs. a diagrammatic tool (I prefer the term 'representational device') vs. its 'True Nature'. Students struggle enough with chemical bonding, and resonance is especially problematic because it is a total misnomer. The presentation of resonance in this article is, I believe, a recipe for confusion.
  • Students want to know what's what. Mixing VB theory, MO theory and resonance in the way that this article does serves poorly to answer their questions "What is the real structure?", "What is going on?", "Why?".
  • Using the undefined phrase 'non-classical' in the very first sentence and the undefined 'conventional model' in the second is counter-productive.
  • Defining resonance, in part, as a 'mathematical description of a molecule using valence bond theory' is unclear and useless. Students are not interested in mathematical descriptions, and want to know how to understand and generate resonance structures.
  • The idea that resonance is a way to explain properties that cannot be explained in terms of canonical Lewis structures is not incorrect. This is an empirical approach to discussing resonance, answering the question "How did the idea emerge?", but not the question "What is the real structure?".
  • Defining (a hybrid) structure as an 'average' of 'several simpler but incorrect structures' is misleading. What is really incorrect is the representation of resonant structures with a non-resonant Lewis one, not the resonance extremes ('contributing structures') themselves. Whether these contributing structures are simpler or not is a semantic issue. Delocalized hybrids are more correctly defined as an energy-weighted average of localized structures, not a simple average.
  • "Resonance hybrids are always more stable than any of the canonical structures would be, if they existed." Why?
  • "A canonical structure that is lower in energy makes a relatively greater contribution to the resonance hybrid..." Lower than what? In the context where this sentence appears, one could easily be led to think that there are canonical structures that are lower in energy than resonance hybrids.
  • Returning repeatedly to benzene is repetitive.
  • "This observation of greater delocalisation in less stable molecules is quite general." Greater than what?
  • "The excited states of conjugated dienes are stabilised more by conjugation than their ground states,..." Surely this requires some literature reference.
  • "... causing them to become organic dyes." I know of no simple diene that can ever be called a dye, while this actually says that the excited states are dyes and implies that the ground states are not. The idea of their 'becoming' dyes is disturbing.
  • The "allylic rearrangement" is not real. It is so named because the starting material appears to undergo a rearrangement in forming the product, but this is ignoring the non-rearranging nature of the cationic intermediate. "Isotope labelling experiments have shown that what happens here is that the additional double bond shifts from 1,2 position to 2,3 position in some of the product." What 'additional' double bond?
  • You cannot dissociate a discussion of resonance without talking about orbital overlap, conjugation and energy.
  • Set aside the historical perspective, which is interesting but not to beginning students.
  • The mention of ring currents (and NMR) is inappropriate.
  • The mention of superacids is unneeded.
  • The vector analogy is not wrong but unnecessary and does not merit a sub-heading. The point is that the true structures can be represented as a linear combination of structures. Students understand linear combinations (from LCAO).
  • "In the mathematical discipline of graph theory, a Kekulé structure is a matching or edge-independent set in a graph." Whoopee-do. Unnecessary and unilluminating.
  • "In benzene both Kekulé structures have equal weight." Why? (Don't answer that; I know the answer. My point is that this needs to be substantiated, but nowhere is this explicited.)
  • I disagree that there is a theory called 'resonance'.

Pgpotvin (talk) 00:23, 11 January 2009 (UTC)

I largely agree and will look at this, if others do not, when I have more time. It will not be easy.

"In the mathematical discipline of graph theory, a Kekulé structure is a matching or edge-independent set in a graph." I added this to get the links in. In particular, Kekulé structure rather oddly redirects to matching. I could not see a better way of drawing attention to this. Can you think of a better way of getting the links right to all the related articles? --Bduke (Discussion) 23:12, 10 January 2009 (UTC)

The term "Kekulé structure" is used elsewhere. The redirection can be changed, or the link can be to Kekulé structure.

Pgpotvin (talk) 00:30, 11 January 2009 (UTC)

This link leads to the article on Kekulé which only had a verbal description of the structure, so I have added the diagram from Commons. Dirac66 (talk) 01:32, 11 January 2009 (UTC)

On the other hand, maybe I'm off. I followed Bduke's links to Wikiversity. I'd never looked at that before. There is there an article on Resonance with many of the same elements as here (and which I presume was also written by Bduke). Maybe the didactic approach for which I argued earlier rightly belongs there, whereas the article here can speak to other, more general audiences. Perhaps a link to the Wikiversity article can be added to direct students? Then this article can be greatly simplified.Pgpotvin (talk) 01:56, 11 January 2009 (UTC)

You are indeed being less than clear. You need to look at the history. I presume you mean links on User:Bduke to wikiversity in general. I did not write the material on wikiversity that you mention (look at the history) and I did not write this article, although I have done a few minor edits. Wikepedia and wikiversity serve a different purpose. --Bduke (Discussion) 07:38, 11 January 2009 (UTC)
Sorry, but where exactly are "Bduke's links to Wikiversity"?? I tried searching Wikiversity for resonance + chemistry, and all I found was this fragment (not by Bduke).

In any case I don't really think it is useful to have two sets of articles named Wikipedia and Wikiversity. There is enough work to do without having two articles on each subject. Sans compter les articles en d'autres langues. Dirac66 (talk) 04:47, 11 January 2009 (UTC)

  • Agreed, this article has some issues that need to be addressed. Already relinked Kekulé structure to aromaticity V8rik (talk) 21:48, 11 January 2009 (UTC)

Sorry, I only assumed without evidence that Bduke had written both articles. The other article is indeed a Wikibooks entry, the link to which I got from Wikiversity. Here it is: Resonance (Wikibooks). I agree that two sets of articles is much to create and maintain, but I can see value in both. Pgpotvin (talk) 22:06, 11 January 2009 (UTC)

New version[edit]

Because there have been no further comments, I've gone ahead and re-written this article. See my sandbox draft here. Depending on your comments, this will replace the existing article in a few days. The changes include removal of material that is not germane, correction of some falsehoods, and extensive examples of various resonance situations. The tone is different, as well, with not a single mention of Valence Bond theory. Pgpotvin (talk) 21:06, 18 January 2009 (UTC)

  • I am not in favor in replacing whole articles by brand new ones. It is simply not in the spirit of Wikipedia. It will discourage people from contributing (next week somebody else will come along to replace your new version). I not even going to check what information gets deleted in the process as that is a lot of work. Why is VB not allowed in the article ? that will create inconsistencies V8rik (talk) 21:55, 18 January 2009 (UTC)
  • I think your new version contains some good material, but I agree with V8rik. You should use your material to make a series of improvements to the article, starting with the most confusing sections. Leave time for other editors to clean up your edits and for changes to be accepted before moving on. It makes absolutely no sense to try to avoid mentioning valence bond theory in this article. Resonance like hybridisation is a child of VB theory. It has no foundation outside VB theory. Note 3 is confusing if not outright wrong. Pauling's book was 1939 not 1928. I will try to find time to make more detailed comments on your new version. --Bduke (Discussion) 22:02, 18 January 2009 (UTC)
OK, OK. I will follow your advice. Piecemeal edits. I'm glad I put it in my Sandbox.
I didn't actually set out to exclude VB. It turned out to be possible to discuss resonance thoroughly without mentioning the words. Pgpotvin (talk) 12:18, 19 January 2009 (UTC)
  • thanks for your consideration. I suggest that at least one reference to BV stays in and that references to MO theory go out. Overall you have raised many interesting points regarding this article so I am looking forward to the improvements. V8rik (talk) 21:21, 19 January 2009 (UTC)
I'm no expert on the subject, but it seems to me that MO theory is invaluable for understanding delocalization, no? Pgpotvin (talk) 01:14, 20 January 2009 (UTC)
Delocalisation is the language we use in MO theory to understand this phenomena. Resonance is the language we use in VB theory. That is why VB theory has to be mentioned, while MO theory can be played down in this article. I still think MO theory should be mentioned but only when it is clear that resonance in chemistry is part of VB theory. --Bduke (Discussion) 01:55, 20 January 2009 (UTC)
Sounds like a turf war to me, and I wouldn't want to be caught up in it. Every practicing chemist I know talks of resonance without meaning to speak the VB language while refering to delocalization, without knowing that to be strictly MO language. To me and to them, these deal with (are) the same thing. Whether VB people want to keep the term 'resonance' for themselves or not, whether they are historically justified in doing so or not, is of little import. We are all talking structure and properties, after all.
The concept of resonance is taught early on to non-savvy people on the heels of Lewis structures and is a source of confusion and struggle when dealing with real structures. My purpose here is to give a palatable and useful presentation to the average non-theorist and to the early chemistry student, without at the same time insulting the theorist. So the references to VB will stay, but I would argue that the relationship to delocalization has to stay as well. There is no intent to put up MO diagrams, so rest easy. Pgpotvin (talk) 03:00, 20 January 2009 (UTC)
I've taken a stab at the opening paragraph. It is simpler, with no limitation to just bonding pairs, no vague mention that "the other [meaning] has to do with the mathematical description of a molecule using valence bond theory" while adding in the connection to properties, and including that what links resonance contributors together is the different electron localizations on the same atomic skeleton. Pgpotvin (talk) 03:06, 20 January 2009 (UTC)

I also think that both VB and MO theories should be mentioned in order to properly show the conceptual development of the subject. First note that some early chemists did occasionally represent a molecule by multiple structures without any reference to quantum theory, including Thiele in 1899 (!) and Arndt more systematically starting from 1924. (For Arndt see references in the Kerber article.)

Pauling of course was the first to relate this practice to quantum mechanics by writing the molecular wave function as a combination of VB functions. The inclusion of VB is justified in this article because of its extensive use in the semi-quantitative theories of the chemical bond by Pauling and Wheland (about 1930-1955), since the concepts of these theories are still used in qualitative bonding theory. [As for the date of “Nature of the Chemical Bond”, Kerber notes that Pauling first used this title for a series of papers starting in 1931, and then for the book, first edition 1939]

As for MO theory, we can note that although it is less clearly related to resonance, it is better suited to systematic quantitative calculations of “delocalization energy” which may be considered another form of “resonance energy”. From the early calculations of Huckel and Coulson to modern ab initio quantum chemistry methods, MO theory has evolved to provide the most reliable way to evaluate “resonance energy”, which justifies its mention in the article as well. In sum: explain that VB is for bonding concepts, MO for quantitative calculations. Dirac66 (talk) 03:12, 20 January 2009 (UTC).

I am not disagreeing in general. “Resonance energy” and “delocalization energy” are about the same thing. The first is from VB and the second from MO. It is no longer a turf war as it was in the 1940s. However to say that resonance is using a combination of Lewis structures without mentioning VB is odd, since here the Lewis structures are VB structures. I would also mention that "ab initio quantum chemistry methods" now include both MO and VB calculations. The latter are no longer just the empirical calculations of Pauling. So "explain that VB is for bonding concepts, MO for quantitative calculations" is just wrong (both VB and MO are for both), but these issues do not need to be mentioned here but in the articles on MO and VB. --Bduke (Discussion) 03:28, 20 January 2009 (UTC)

Yes, please, not here. Let's all agree that this article is about resonance, which has roots in VB and is equivalent to delocalization, a word from MO theory that will not be sacrilegious to mention. Pgpotvin (talk) 03:33, 20 January 2009 (UTC)

Resonance structures must contain the same number of unpaired electrons[edit]

Pauling makes the statement about conditions for resonance that "the two structures must involve the same numbers of unpaired electrons" (Nature of Chemical bond, 1940).Is this (still?) true - and if it is, shouldn't the statement be made in the section "writing resonance structures". --Axiosaurus (talk) 10:04, 16 February 2009 (UTC)

Good point, and still true since it is related to the quantum mechanical statement that total spin is constant. I have now added this point (slightly reworded) to the article. Dirac66 (talk) 14:25, 16 February 2009 (UTC)

I don't know if Pauling's statement is still held to be true or not, but I don't think Dirac66's explanation here isn't consistent with it: as I understand the reaction 3O21O2, the initial diradical undergoes an intersystem crossing to become a singlet diradical, and then the two lone electrons combine to form (as we draw it usually) the pi bond. That second step does not involve a change in spin but it does involve a change in the number of unpaired electrons. Is it resonance? DMacks (talk) 15:02, 16 February 2009 (UTC)
No, this second step is not resonance. As discussed under singlet oxygen, it is the transition from a higher excited singlet state to the lowest singlet state, which is metastable with a lifetime of about 1 hour before returning to the triplet ground state. There is no "return" to the initial higher singlet state.
Resonance, on the other hand, means that one quantum state can be considered as a combination of different contributing structures, which are not quantum states.
I have also found a reference now. Douglas, McDaniel and Alexander, Concepts and Models of Inorganic Chemistry (2nd edn 1983) gives rules for resonance (p.51-53) including "2.All contributing structures of a particular molecule must have the same number of unpaired electrons." Their example of a structure which is ruled out (for benzene) is a diradical formed from Dewar benzene by breaking the long bond. If this is assumed singlet, it is equivalent to your example. Dirac66 (talk) 19:56, 16 February 2009 (UTC)
Thanks for checking up on it! DMacks (talk) 19:59, 16 February 2009 (UTC)

Why has this been labelled a personal essay?[edit]

Yesterday Wickey-nl inserted Template:Essay-like which says that "This article is written like a personal reflection or essay and may require cleanup." The template links to WP:Not#Essay, which I have just read, but it is not at all clear that it refers to this article. In fact the content appears to me quite similar to what is found in chemistry textbooks at various levels (some paragraphs are more advanced than others). Please explain what is personal about this article.

It is true that very few sources are given. Perhaps a more appropriate notice would be Template:Refimprove which asks for more sources. Dirac66 (talk) 15:39, 4 April 2010 (UTC)

  • I am equally puzzled V8rik (talk) 22:14, 4 April 2010 (UTC)
  • As am I, so I have removed it. --Bduke (Discussion) 00:02, 5 April 2010 (UTC)
Sorry for the late reaction. Indeed the wrong template. I meant to say this is not an ecyclopedic article (despite internal links), but a wikibook. The first meaning of essay. It is written like a textbook. Funny that the term is used above. Wikipedia is not a textbook. --Wickey-nl (talk) 10:41, 6 April 2010 (UTC)
We could modify the tone of some sections. I have started with the section Writing resonance structures, where I have inserted "Chemists use ... rules ..." and removed the word "must" which was used three times. Dirac66 (talk) 14:23, 6 April 2010 (UTC)
Ecyclopedic style is principally compact, as much as possible.--Wickey-nl (talk) 15:02, 7 April 2010 (UTC)
  • Compact? where does it say articles have to be compact V8rik (talk) 17:24, 7 April 2010 (UTC)

Writing resonance structures - rule 4 is not useful[edit]

In the section Writing resonance structures, rule 4 does not really belong because it is not helpful at an elementary level. Rules 1,2,3 and 5 can each be illustrated in a first-year lecture with simple examples of acceptable and unacceptable structures, and perhaps such examples should be included in the article.

However rule 4 says "Resonance hybrids can not be made to have lower energy than the actual molecules." This is a special case of the variational principle of quantum mechanics and therefore a true statement. But it is not a useful rule because one cannot quickly determine the energy by inspection, unlike for example the number of unpaired electrons. I propose that this rule be deleted from the article. Dirac66 (talk) 14:14, 4 May 2010 (UTC)

Another viewpoint is that rules 1,2,3 and 5 deal with how to choose contributing structures (even if rule 5 now incorrectly refers to hybrids), whereas rule 4 gives a property of resonance hybrids. So instead of deleting rule 4, I will move it to the section on True nature of resonance, where I think it belongs. Dirac66 (talk) 15:21, 4 May 2010 (UTC)
What does that sentence formerly known as rule 4 even mean? DMacks (talk) 15:30, 4 May 2010 (UTC)

What is the definition of "resonance hybrid"[edit]

The resonance hybrid by definition has the lowest energy. May be the two last rules are added later?--Wickey-nl (talk) 15:48, 4 May 2010 (UTC)
I interpret it as equivalent to the variational principle - the resonance hybrid corresponds to a wavefunction of lower energy than the individual resonance structures, but higher energy than the true wavefunction. Of course this is only meaningful if one has the means (software etc.) to compute the energy, so it did not belong in a set of rules for drawing chemical structures by hand. Dirac66 (talk) 17:55, 4 May 2010 (UTC)
Yup, that's all right. But that's almost opposite how I read the sentence. Should be that individual resonance-structure contributors are never lower in energy than the resonance hybrid, which approximates the electronics of the actual molecule. At least as my textbooks usually use the terms, the "resonance hybrid" is the nonclassical average of the individual simple Lewis-structures and is nearly the "real picture", and the point of the article is that resonance does pretty well reflect reality (vs this being just another pretty far-off approximation from the true wavefunction). DMacks (talk) 19:13, 4 May 2010 (UTC)
Hmm, I had not thought of that interpretation. I presume your "nonclassical average" refers to averaging electron density and not energy. Anyway, at the level of this article we now have two interpretations so the sentence is ambiguous and needs rewording, or ... we can just delete it after all since it doesn't really add much to the (now) preceding paragraph. Dirac66 (talk) 20:55, 4 May 2010 (UTC)
Let's explain, without the variational principle and without mathmatics, what the "resonance hybrid" is.
Can we say: The resonance hybrid has a wavefunction that is the weighted average of the wavefunctions of the contributing structures? This would implicate that the energy of the resonance hybrid also is the weighted average of the energy of all individual contributing structures (?), so even a bit higher than that of the major contributing structures. The term resonance hybrid would be meaningless.
Or would the resonance hybrid be the structure with the lowest possible energy? That would implicate that the resonance hybrid is the structure that reflects the real compound. --Wickey-nl (talk) 10:51, 5 May 2010 (UTC)
"Weighted average" is not totally correct here. The hybrid wavefunction is a linear combination, which can be considered a weighted average IF all the coefficients are positive, but this is not always true. The electron density is a weighted average of densities for the contributing structures, but its coefficients involve the squares of the wavefunction coefficients. And the energy of the resonance hybrid is NOT a weighted average of energies of contributing structures (which would be between the highest and the lowest!); instead it is usually lower than the lowest of the contributing structures due to quantum-mechanical interaction between the structures. Dirac66 (talk) 01:43, 6 May 2010 (UTC)
Definitions can change with time, but Linus Pauling (in 1939!) refers to a resonance hybrid as a system that has to be described by the wave functions of more than one structure, without referring to a certain level of energy (see page 12). I think it is best to keep it similarly simple in this article, unless good references can be given for another meaning.--Wickey-nl (talk) 13:49, 6 May 2010 (UTC)
This discussion has convinced me that the former rule 4 really adds nothing useful. Three of us have each read a different meaning into it, so it must be unclear and ambiguous. Is it worth trying to clarify? The ΔE between resonance hybrid and exact wave function is really outside the scope of this article, and the ΔE between resonance hybrid and contributing structures is already clear in the section on resonance energy. So I have now decided to just delete the former rule 4 - one step toward a simpler article. Dirac66 (talk) 22:41, 6 May 2010 (UTC)

Bond lengths[edit]

Added yesterday: "If the bond lengths are measured, for example with NMR spectroscopy, no single and multiple bonds can be distinct. All bonds appear to have the same bond length, somewhere between single bond and multiple bond length."

Two comments: 1. Bond lengths are not usually measured with NMR, but with x-rays in solid state, or by microwave, ir or uv-vis spectroscopy in gas. 2. Can we specify exactly which bonds have the same bond length? Even in benzene, C-C and C-H are not the same length. Do we mean all bonds involved in the resonance? No, because today the thiocyanate ion was added, and the S-C and C-N bonds are not the same length. Do we mean all bonds between similar atoms? No, in ethylbenzene the side-chain C-C is a single bond and longer than the ring bonds. And in naphthalene there are ring bonds which are longer than other ring bonds.

I think what we really mean here as a general statement is that bonds with different bond orders in different contributing structures usually have intermediate bond lengths. Then we can give some examples of equal bond lengths (benzene, carbonate) without claiming that all bond lengths are equal. Dirac66 (talk) 20:20, 8 May 2010 (UTC)

Agree totally, but have no time to do anything. --Bduke (Discussion) 22:34, 8 May 2010 (UTC)
I made this section to break the structure of the old article, which prevented improvements. Of course it has to be rewritten completely.
I am not familiar with measurement technics, but it is called to emphasise that it is not only theoretical guess, but information from practise. NMR can be replaced by x-rays or so. A separate section about measurement of resonance would be interesting.
The characteristic of resonance is that single and multiple bonds are not as could be expected, compared with lengths as measured in simpler (reference) compounds. The reference lengths are mentioned in the current article and in other WP articles e.g the German one. I do not have an overview of all possible bond types. Currently the classical example of benzene is used. Is the differing bond length of adjacent single and multiple bonds (compared with normal single and multiple bonds) better as general characteristic instead of all single and multiple bonds are identical? In the thiocyanate ion S-C versus S=C and C=N versus C≡N.--Wickey-nl (talk) 15:55, 9 May 2010 (UTC) Thanks for your review--Wickey-nl (talk) 09:57, 10 May 2010 (UTC)

Delete the chapter True nature of resonance?[edit]

Can we delete the chapter True nature of resonance? If not, what part is worth to keep?--Wickey-nl (talk) 11:19, 10 May 2010 (UTC)

I think this section is needed to describe the relation of Resonance to the more quantum-mechanically based bonding theories such as VB and MO. I would keep everything except the sentence about graph theory, which is less used by chemists. Perhaps the section title should be changed to "Resonance and bonding theories" or "Resonance and quantum mechanics". Dirac66 (talk) 14:55, 10 May 2010 (UTC)
I think it are many words to tell little, lacking any structure. That relationship is not made clear. No one will read it with pleasure.--Wickey-nl (talk) 16:16, 10 May 2010 (UTC)
Actually it seems quite clear (except for the graph theory sentence) for advanced readers who already know about bonding theories and a little about quantum chemistry. Admittedly it will not help readers at a first-year level. It can always be improved further, but I believe it should be retained until someone has time to improve it. Dirac66 (talk) 20:30, 10 May 2010 (UTC)
For beginners it is worthless, links to VBt and MOt will not help. For advanced readers it is not interesting. I propose to delete all and write (not by me) a new section about the relationship with quantum chemistry, wavefunctions, VBt and MOt. At least I will change the heading.--Wickey-nl (talk) 14:28, 11 May 2010 (UTC)
Do not delete until the rewrite has been discussed here. --Bduke (Discussion) 14:22, 12 May 2010 (UTC)

Resonance energy[edit]

"In fact, resonance energy, and consequently stability, increase with the number of canonical structures possible, especially when these (non-existent) structures are equal in energy."

This can only be true if all canonical structures have comparable energy and have low energy. A canonical structure with higher energy would, by definition, not change the resonance energy and make the compound even less stable.

Furthermore, I think you cannot say that contributing structures are non-existent structures. Although they do not represent the actual compound because they "ignore" the other contributing structures, no one can say this structure never exists at certain moments.--Wickey-nl (talk) 14:06, 13 May 2010 (UTC)

These are good points. I would change "number of canonical structures possible" to "number of canonical structures of low energy" or "of comparable energy". I disagree with your "even less stable" however; according to quantum mechanics structures with higher energy just don't contribute (or contribute minimally) to the ground state, and so leave the stability (almost) unchanged. Their actual effect is very slightly stabilizing, but since the effect is minimal it is probably best not to mention them at all.
As for "non-existent", the correct phrase would be "non-existent in pure form" (or uncombined form) Also I note that this 12-line paragraph says "if they existed" once, "non-existent" twice and "no physical existence" once. I really don't think we have to tell the reader this four times in the same paragraph. Dirac66 (talk) 15:01, 13 May 2010 (UTC)
1. "of comparable low energy" sounds best, but is not unambiguous. Can also mean low/comparable on high-energy level. 2. Non-contributing high-energy structures is not logical. Contributing minimally will be the rule. For outsiders it sounds not logical that structures with higher energy stabilize the compound, although minimally. Apparently that is exactly the paradox and the essence of resonance.--Wickey-nl (talk) 16:00, 13 May 2010 (UTC)

Replace "compound" by "molecule" everywhere[edit]

This article now contains the word "compound" 14 times, and I suggest they all be replaced by "molecule". First, resonance can be a property of elements too; ozone, graphite and fullerenes come to mind. More fundamentally, resonance structures are a property of the microscopic molecular unit and not of a molecular compound (or element). It is not correct to talk about (Lewis) structures of a "compound".

Some may argue that "molecule" excludes ions, but we can add "or ion" a few times, or else specify "molecule (neutral or charged)". In any case, "compound" also does not include ions. Dirac66 (talk) 23:21, 1 June 2010 (UTC)

That resonance also occurs in elements is really a good point. Changing into compound or element in the first sentence would be an improvement.
Compounds include compound ions, but molecules exclude ions. In most cases it regards compounds and in many cases it regards ions, thus replacing compound by molecule would make the article less correct. In most cases only "molecule or ion" will be correct. The number of times a word is used can of course not be an argument. This is not a novel.--Wickey-nl (talk) 12:53, 2 June 2010 (UTC)
I agree that "compound" is definitely not the best choice here (not inclusive enough and suggests macroscopic effect). Is "structure" too confusing a term? Resonance really is a feature of the electronic structure, regardless of how one classifies what that entity is. Just have to be clear when we talk about "individual resonance structures" (the two cyclohexatriene forms of benzene) vs "a structure that has resonance" (benzene). DMacks (talk) 03:12, 4 June 2010 (UTC)
I have already admitted that changes are needed, but "molecules" is too exclusive. "Structure" is a little confusing between contributing structures, the combination should be avoided. A single word will always be a compromise. "Molecule or ion" will in most cases be most exact. Why can a compound not be just a compound with a special electronic structure? B.t.w., the first sentence clearly indicates that the article is about the electronic structure, not about compounds of molecules in general, but I will adapt it also.--Wickey-nl (talk) 12:38, 4 June 2010 (UTC)

Molecule vs. compound vs. structure[edit]

I still think that "molecule" is the best word in most cases (in this article), because electrons are usually delocalized over one molecule and not over a macroscopic sample of a compound or element. (Graphite is an exception and has no molecules). I will however add "or polyatomic ion" and "or ion" a few times to make it clear that ions are included. The idea that "molecules exclude ions" comes from the simple chemistry of small molecules. Organic chemists and biochemists routinely refer to ions as molecules - try telling a biochemist that proteins and DNA are not molecules just because they are charged!

We can however make the point explicitly that the molecules (and ions) can be of both compounds and elements. Since most of the molecules are compounds, we might include an explicit list of elements: I mentioned ozone, graphite and fullerenes which are all neutral molecules, and an example of an ion is the azide ion N3-. As for "structure" we can use it in a few places where the distinction is clear between the hybrid structure and the contributing structures, but I think it would be confusing to use it everywhere.

I will try to make these changes now. Dirac66 (talk) 01:16, 8 June 2010 (UTC)

No one is talking about compounds on macroscopic scale, nor it is suggested, except on this talk page.--Wickey-nl (talk) 13:18, 8 June 2010 (UTC)
OK, instead of "macroscopic", let us say (here on the talk page) that a "compound" refers to an "arbitrary amount" of a (pure) substance containing two or more elements. Since resonance refers to electron delocalization over a single molecule, it is better to use "molecule" in this article. I did leave in the one mention of compounds or elements. Dirac66 (talk) 23:17, 8 June 2010 (UTC)
I don't like this word play. Either you don't want to or you cannot get the point. If you say that biochemists use molecule while it is an ion, I can say in general, compound is used also if it is a molecule. With a few exceptions as pure oxygen or carbon, all structures are compounds. This is not a biochemistry article. Dipolar compounds are never called dipolar molecules and I never use compound to refer to an amount of substance.--Wickey-nl (talk) 15:15, 9 June 2010 (UTC)

Um... DNA isn't a molecule. Perhaps a pair of molecules held together by hydrogen bonds, but not a single molecule. Proteins can be single molecules though. Biologists are terribly sloppy with nomenclature, and can often fail to tell a compound and an ion apart. The IUPAC definition of a molecule is an electrically neutral entity containing more than one atom. Thus zwitterions can be molecules, while atoms like argon or helium are not. This squares exactly with my understanding as a chemist. --Rifleman 82 (talk) 15:34, 9 June 2010 (UTC)

Resonance is not a rapidly-interconverting set of its contributors?[edit]

Is this an axiom?--Wickey-nl (talk) 16:19, 22 July 2010 (UTC)

It's logically provable if that's your concern (but it's also nearly just a restatement of other content). It's the way both undergrad-orgo texts I have available describe the situation. DMacks (talk) 16:48, 22 July 2010 (UTC)
Separate contributing structures cannot be isolated or observed. This does not automatically mean they do not exist. Otherwise they were useless to explain the actual structure. I am curious about the non-mathmatical logical prove.--Wickey-nl (talk) 10:36, 23 July 2010 (UTC)
Since there is no discussion I give my view here. The statement above possibly would be true if there were infinite many particles. However, there are few nuclei and valence electrons. As electrons cannot be on different places at the same moment, there must be interconverting configurations. So the statement is wrong.--Wickey-nl (talk) 15:26, 28 July 2010 (UTC)
In chemistry a structure is only said to "exist" if it corresponds to a local or global potential energy minimum (as a function of nuclear geometry). It is true that any nuclear configuration can occur instantaneously, but it is not considered as a structure which "exists" in the chemical sense if it is not an energy minimum. This is a practical definition of existence, since configurations which are not minima do not last long enough to detect spectroscopically. This is the case for example for the contributing forms of ozone O=O-O and O-O=O. Each has one long and one short bond, which does not correspond to an energy minimum according to spectroscopic evidence.
As for the question of whether this is an axiom, I would say instead that it is the definition of "resonance", a word which is only used when the contributing forms are not minima. In cases where there is really interconversion of two energy minima, we describe them by other names such as isomers or conformers. I hope this helps. Dirac66 (talk) 23:27, 28 July 2010 (UTC)
Agreed. Do we really consider electrons to have instantaneous positions when we discuss valence bonding? Doesn't make much sense to do so, since then we would have to consider (as a logical extreme) both ionic forms and the pair-of-radical forms too--even for a non-resonance case! Sure, electrons are particles, but the analysis for this purpose is as a probability cloud (example: "it's a shared pair, period--it's a sigma-shaped region, or atomic orbital/hybrid overhaps" for methane). So ozone can sure be thought of as an average two traditional-looking valence structures, but it's not really either one. At an electrons-are-particles, or even short-term position average, you might happen to have one side with approximately 2x the electron amount as the other. But is there any type of analysis that supports that? Or are the pi electrons too fast (or too delocalized, or not even localizeable even philosophically?), and it's just a mind game to even ask? Drawing single- and double-bonds also suggests all sorts of other properties that don't make sense: different bond-lengths, perhaps different geometries, etc. But the bonds are the same length (or a resonance-average of not-detectable contributors...there I go again), not changing on the time-scale of electronic motion. The real picture, in any studiable sense, is 1.5 bonds each (still staying with ozone or other simple/symmetric cases).
Let's go to a source. McMurry Organic Chemistry 7th ed: "Acetate doesn't jump back and forth between two resonance forms, spending part of the time looking like one and part of the time looking like the other. Rather, acetate has a single unchanging structure that is a resonance hybrid of the two individual forms and has characteristics of both." ... "Rules for Resonance Forms...Rule 1 Individual resonance forms are imaginary, not real. The real structure is a composite, or resonance hybrid of the different forms... They have a single, unchanging structures, and they do not switch back and forth between resonance forms." (bold in original, typos assumed to be my own). My graduate-level texts are packed for an office-move, so I don't know how exactly March describes the situation. DMacks (talk) 23:52, 28 July 2010 (UTC)
Yes. I was of course referring to the nuclear positions, and you have given a complementary discussion of the electronic charge density with which I also agree. To be fair to the resonance method, however, we should concede that despite the unsatisfactory conceptual basis which we have discussed, it does provide very simple qualitative predictions of organic chemical reactivity which are often correct. Dirac66 (talk) 00:40, 29 July 2010 (UTC)
Resonance structures exist in our minds if we are thinking valence bond language. Resonance structures are made up. The molecule is not rapidly-interconverting between the structures. If we are thinking molecular orbital language, then we do not even think about resonance. Resonance structures are one approximate mathematical way that we can think about bonding. --Bduke (Discussion) 01:14, 29 July 2010 (UTC)

This helps, but I like to make some remarks:

  1. First of all, the article does not contain a word about these fundamental things.
  2. This article regards both the concept of contributing structures and the nature of resonance in chemical species. Contributing structures can be considered as relative energy minima.
  3. If "being not a rapidly-interconverting set of contributors" is the definition of "resonance", ok, but then it should be referenced.
  4. AFAIK, structures cannot be observed directly, so all this is basically theoretical/hypothetical. Molecular/atomic orbitals are also not more and less than models to approximate the reality. Even bond-lengths cannot be observed directly and continuously; observing equal bond-lengths may be just a matter of shortcoming observation technics.
  5. Individual resonance forms are imaginary, not real. This statement can only be true in a certain context. The problem is the word "real". In the concept of resonance the contributing structures are real Lewis structures. You can only say the real structure is not eqal to the contributing structures (which is not provable by direct observation). The bare statement is obscuring, not clarifying

Unfortunately, I have to do with online references.--Wickey-nl (talk) 16:32, 29 July 2010 (UTC)

Comment, point by point:

  1. So add them.
  2. "Contributing structures can be considered as relative energy minima". This means nothing to me. Relative to what? Each contributing structure can be given an energy, but it is not observable and it is not a minimum. It is just <Psi|H|Psi>, where H is the Hamiltonian and Psi is the wavefunction constructed from the structure. The energy for the system comes from the energy of a linear contribution of the various Psis for the various structures.
  3. "being not a rapidly-interconverting set of contributors" is NOT the definition of "resonance". It is merely a comment about a common misunderstanding about resonance.
  4. I agree about orbitals, resonance structures, wave functions. For bond lengths, we need to more careful about what we mean by a bond length and the nature of the experiments.
  5. What do you mean by a "real" Lewis structure? Lewis structures are just diagrams assigning electrons to atoms. For resonance we have to map them to functions - my Psi above. We can do that in different ways. It is not clear cut. --Bduke (Discussion) 23:10, 29 July 2010 (UTC)

Re 2. "Relative minimum" is an accepted synonym for "local minimum"; see Maxima and minima. However the statement is incorrect here since as I have already stated, the contributing structures do not in fact correspond to local minima. Dirac66 (talk) 03:10, 30 July 2010 (UTC)

I think I have the right formulation now:
"The real structure is not a rapid interconversion of contributing structures. Different contributing structures are used, because none of them exactly represent the real structure."
--Wickey-nl (talk) 15:05, 30 July 2010 (UTC)
Yes, your last statement is correct. Note also that I have restored the word "real" which you deleted from Bduke's comment above, because as a rule on talk pages we leave each other's signed comments intact (except in cases of vandalism). Dirac66 (talk) 17:46, 30 July 2010 (UTC)
I assume we're talking about Wickey-nl's "right formulation" statement? The second sentence is logically confusing for the word "used"--if they aren't real, what is the context of "using" them? Could reword to focus on simple diagrams/traditional bonding patterns, or predicting stability/reactivity/etc. DMacks (talk) 19:01, 30 July 2010 (UTC)
The word must be deleted unintentional, while copying it. May be DMacks is reading something I don't. I see no contradiction. Is in the second sentence the use of "A combination of different Lewis structures" instead more clear?
Nevertheless, I still think the statement in the heading of this section is essentially wrong. In my idea every structure has an infinite number of contributing structures; most if them minor, fewer of them major ones. The problem: only extremes (the single and multiple bondings) can be pictured by a Lewis structure and major contributors cannot be weighted in a hybrid picture. But I realize personal views are not appreciated in WP.
The statement is aceptable for me as far as it expresses that the structure cannot be represented by a single Lewis structure. In that respect it can be emphasized that even the use of several Lewis structures is not more than an approximation.--Wickey-nl (talk) 13:59, 1 August 2010 (UTC)

Last edit, gallery and hyperconjugation[edit]

I have reverted the last edit. First, the hyperconjugation heading is for a section that is not about hyperconjugation. It is about 3-centre 2-electron bonding with sigma orbitals. As it says, hyperconjugation is about pi electrons. Second, I suggest the gallery is inappropriate. These diagrams could just as well be describing delocalisation with MOs. They are certainly not resonance structures. --Bduke (Discussion) 02:18, 5 August 2010 (UTC)

I have removed the resonance hybrid images. These are misleading as they could equally be MO diagrams. Indeed the circle in the hexagon is generally accepted as short hand for the MOs. The point is already made clearly in the examples where diagrams for ozone and benzene are shown as the clear consequence of mixing resonance hybrids. That is clear. The images on their own are not. --Bduke (Discussion) 02:26, 5 August 2010 (UTC)
Nothing misleading. MOs may be true, but the hybrid images are unambiguous resonance structures. Molecules with delocalized pi-electrons.--Wickey-nl (talk) 10:24, 7 August 2010 (UTC)
No, you are wrong. Take a look at the diagrams under the "Examples" heading. The diagrams on the left of the equal sign are resonance structures. The diagrams on the right of the equal sign are not resonance structures. They are just representations of delocalization, which can come, as here, from mixing resonance structures, or equally from MO. Just having diagrams of the right hand side labeled as resonance hybrids is misleading. On the other hand these example diagrams make everything clear. We only need those. There is no need for extra diagrams of the right hand side for benzene and ozone. Note, I am not pushing MO or fighting the old VB v MO battle, so your "MOs may be true" is not relevant. I am asking for clarity about the use of terms. --Bduke (Discussion) 23:44, 7 August 2010 (UTC)
The misunderstanding is that I referred to entities that can be pictured as resonance structures, not to the form of the pictures as I indicated. That is why "identical to signs" are used, not "equal signs". And as this article is about resonance, the context is clear. There is nothing wrong with giving some examples of typical, but different types.--Wickey-nl (talk) 11:17, 11 August 2010 (UTC)
I still maintain that they are confusing. I also say that we already have the ozone and benzene diagrams as the right hand side of what you correctly call "identical to signs" (I was not trying to be absolutely correct there - just to point where they are). You were just adding a bunch of unnecessary and misleading images. There is no misunderstanding. What is your background in chemistry? Have you studied bonding theories such as valence bond and molecular orbital to an advanced level? I find it difficult to pitch what I say to you, because I do not know your background. Mine, I think is clear, from my user page. --Bduke (Discussion) 11:48, 11 August 2010 (UTC)
OK, I give you away I haven't studied the subject to an advanced level. I found an article of very poor quality and without appropriate references. Apparently there are no authors willing to write at advanced level. Only to make comments, minor edits and revert edits.--Wickey-nl (talk) 13:57, 11 August 2010 (UTC)
OK. I think you are a little confused about what resonance is all about, but you are not alone. I hesitate to wade in on this article for two reasons. First, I am not sure it needs to be written at an advanced level, but it should at least not be wrong. Second, most of my current research is on valence bond theory, so I am sort of an expert. However VB experts notably all disagree with all the other experts in a fairly profound way. I exaggerate, but not much. Maybe that is why there are no authors willing to write at advanced level. Nevertheless I still state very clearly that I think the image in the examples section is the only place where we have these very vague diagrams on the right that just state that the electrons are delocalised. The diagrams on the left are resonance structures and map directly to the wave functions that are then combined to give a VB total wave function. The diagrams on the right have no direct correspondence to any wave function, but rather to several, including the MO wave function, the VB function from those two structures, and the VB function from many structures (6 for ozone and 5 or a lot more for benzene). So in the context the diagrams on the right are OK. However on their own they are not OK and are also misleading. Part of the problem is the name of these images which include the word "resonance". They should not as it should be replaced by the word "delocalization". Does this explain things and are you now happy to leave these images out? --Bduke (Discussion) 00:18, 12 August 2010 (UTC)
(outdent) Agree with Bduke here. A loose collection of images illustrating delocalized electrons isn't resonance, it's more like the result of it. Resonance is the pieces and an explanation behind it. It's great to use them as part of a LCAO diagram because that's how they are relevant to this article. While we're picking at this though, File:Resonance examples.png is wrong in two ways related to the idea of resonance as a combination of contributors (rather than a "real" MO diagram). First, allyl should +1/2 on each end not + in the middle (formally a + delocalized, but again that's not this article). Second, benzene should have half-bonds (dotted circle) again because we're talking about combining/averaging bond-orders rather than an orbital picture. DMacks (talk) 00:37, 12 August 2010 (UTC)
I agree. We just have too many bad diagrams of bonding done by people who do not entirely know what they doing or what the target article is for them. --Bduke (Discussion) 01:36, 12 August 2010 (UTC)
Best I can discern, that image is only used in this same context (i.e., misused) anywhere in the *pedia world. Should I just alter the original, or place it at a new filename? While we're at it, should allyl+) be redrawn bent? Should it be replaced by something with more than just two resonance contributors (pentadienyl+ perhaps)? DMacks (talk) 14:51, 12 August 2010 (UTC)
@Bduke. "Resonance hybrid diagrams" can be renamed to "hybrid diagrams". In the paragraph about diagrams the pictures were/are appropriate. I fully agree about the level of the article. The subject is not trivial, but must be accessible. A better text about the wave functions would be appropriate in the section about quantum mechanics, instead of focus again on the example of benzene. Can you indicate in a few words what the disputations are about resonance?
@DMacks. If the original is altered it will appear everywhere where used. If you add a new structure I think it may be regarded as a new image.--Wickey-nl (talk) 16:02, 12 August 2010 (UTC)
Sorry, no. I have just broken my arm. --Bduke (Discussion) 08:38, 13 August 2010 (UTC)
  • With respect to File:Resonance examples.png I am more than happy to redo the image. The benzene ring should have a dotted circle (blame my old molecular graphics editor). The allyl cation is typically depicted in the literature with one full and one dotted bond (semi circle) and one full overall positive charge. It should be bent though V8rik (talk) 21:28, 12 August 2010 (UTC)
I am not happy at all with several examples in one image. It limits the possibility to add other examples in an elegant way and no explanation per example is possible. Separated examples are preferable and more flexible.--Wickey-nl (talk) 16:50, 15 August 2010 (UTC)
fair enough. I am having some issues with other images though: the nitrogen atom in the nitrate ion has a positive charge and the oxygen atoms a delta minus charge as in polarity. This should be a -2/3 charge for each oxygen atom. The continuous circle for benzene vs the dashed line is confusing. According to the aromatic hydrocarbon page the full circle representation was invented in 1925 and conveys a different meaning than the one with the dashed lines in the resonance concept of 1928. Also, the single covalent bond image and the double covalent image do not add anyhting to the text and should go. V8rik (talk) 21:26, 15 August 2010 (UTC)
The benzene circle diagram should be removed unless you can find a reliable source that uses it in the context of VB/resonance theory. --Bduke (Discussion) 23:30, 15 August 2010 (UTC)
The image of the nitrate ion may be not completely wrong, but is somewhat ambiguous indeed. For the benzene circle: all versions refer to the same delocalized electrons. Just ignore the most widely used symbol is not the right way. The difference can also be explained in the article and I have not a problem at all to use a version with dashed circle. The covalent bond images are used to illustrate the normal bonds in the Lewis formulas.--Wickey-nl (talk) 15:39, 16 August 2010 (UTC)
The point is that I do not think that I have seen the circle diagram other than in the context of a discussion of the MO theory of benzene. That is why I asked for a source. It is a reasonable request. --Bduke (Discussion) 21:54, 16 August 2010 (UTC)
OK, that was an overreaction. The cyclohexatriene is the standard, but that is not the point. The point is that the circle diagram is generally accepted as symbol of benzene and thus depictures delocalized π electrons and thus represents the resonance character of benzene ( Second, what is the difference between a continuous and a dashed circle? --Wickey-nl (talk) 15:35, 17 August 2010 (UTC)
Per the source you cite, "Delocalization in such species may be represented by partial bonds", which is the dotted-bond notation. A solid-line circle is not a dotted line--it suggests a whole bond rather than the partial. The solid circle is a different representation than the specific idea of the bonding involving resonance. It (some would say more accurately) describes the electronic situation (unified pi orbital around ring, with an broken electron density there) and it's a common notation used in the literature. But it's off-topic here. DMacks (talk) 16:32, 17 August 2010 (UTC)
Of course the solid circle is only used for convenience (easy to draw), but can you say it represents something different? No, you can only say the dashed or dotted version is the version that is more consonant with the usual notations.--Wickey-nl (talk) 14:15, 18 August 2010 (UTC)

I am not sure if this (may be boring) discussion has lead to some agreement. Otherwise it has been a waste of time. My logic is: the essence of resonance is the existence of a pi-system with delocalized electrons. The pictures with dotted bonds represent delocalized electrons, thus they represent resonance structures ("identical to signs" in the image; was not just a simple correction from me). This differs fundamentally from the view of DMacks that delocalized electrons are only the result of resonance. We can also reverse the question: Are there compounds with delocalized electrons, but without resonance?--Wickey-nl (talk) 10:15, 19 August 2010 (UTC)

This continues to be wrong in several ways. First, resonance is not restricted to pi-systems, although it is common there. However the main problem is "the essence of resonance is the existence of a system with delocalized electrons" (I have removed the pi). Resonance is just one way to understand delocalized electrons. MO is another quite different way to understand delocalized electrons. It is important to distinguish between talking about delocalization in general, the resonance interpretation of it, and the MO interpretation of it. This is what you are failing to do. Finally, you can discuss delocalized electrons entirely without using resonance, or you can discuss them by using resonance. It is your choice. It is meaningless to talk of compounds without resonance or with resonance. It is not something like carbon atoms that compounds either have or do not have. It is just an idea that some people find useful and others do not. --Bduke (Discussion) 01:56, 20 August 2010 (UTC)
I have to disagree that "It is meaningless to talk of compounds [molecules] without resonance or with resonance." At the level of simple Lewis structural diagrams, some molecules (such as ethane) can be described reasonably well by a single Lewis structure while others (such as benzene) require a combination of structures which we call resonance. It is true that the difference is harder to define for MO calculations which yield delocalized MOs for all molecules, but much of chemistry (especially organic) still uses simple Lewis structures for which the concept of resonance is useful. Dirac66 (talk) 02:58, 20 August 2010 (UTC)
But you defined the framework, Of course, if you start with "At the level of simple Lewis structural diagrams", then of course some molecules are represented by one Lewis structure and some by several, but you do not have to start there. On the one hand, you could start with ionic-covalent resonance and then both ethane and benzene have many structures, or you could start with MO when Lewis structures do no even come into it. But all this was about images and I am merely saying the images have to match the appropriate framework and to not assume that images from any framework are appropriate in an article on one framework, This article is about resonance. It is not about delocalization or MO. --Bduke (Discussion) 04:08, 20 August 2010 (UTC)
Dirac66 worded pretty well my thoughts. For the "resonance compounds" the pictures with dashed pi bonds are useful, for simple ones not useful and not needed. Do we have such a picture anyway, for the water molecule e.g.? Again, are there compounds with delocalized electrons, but without resonance? For Bduke: are there compounds with delocalized electrons that cannot be explained within the resonance framework? If not, pictures with delocalized electrons are always appropriate within the resonance framework (because delocalized electrons are a condition for the occurrence of resonance and resonance energy). For the article, however, it is may be more useful to explain the difference between the two types of images. --Wickey-nl (talk) 14:33, 20 August 2010 (UTC)
I am going to be pedantic again. Delocalization can be detected from the electron density which is an observable. Resonance can not be so detected. Therefore "are there compounds with delocalized electrons, but without resonance?" is a rather confused question. Delocalisation can be interpreted theoretically in several ways, ONE of which is resonance. "delocalized electrons are a condition for the occurrence of resonance and resonance energy" is false, but one might say that delocalized electrons are a condition for using the explanation of resonance and resonance energy as one possible explanation. "are there compounds with delocalized electrons that cannot be explained within the resonance framework?". No, resonance can always be used as that kind of explanation. However, there are cases where such an explanation is clumsy and seems much less simple than the MO explanation. Two classic examples are the bonding in ferrocene or dibenzene chromium, and large borane structures like B12H14 or the B12H12 dianion. The number of resonance structures is very large, while the MO picture makes use of symmetry and is relatively simple. In the opinion of many, including myself, Pauling lost a lot of credibility in the 1960 edition of his book, "Nature of the Chemical Bond", by only using the resonance approach for those systems with not a mention of the MO approach, which had been used so successfully by Wilkinson, Cotton and Lipscomb. All of these systems are now ones where we naturally use MO, and not resonance, to think about them. --Bduke (Discussion) 00:38, 21 August 2010 (UTC)
To reassure, your point is clear and justified, and I don't want to add a lot of hybrids again (actually it was just an easy and quick way to give a number of examples of compounds that can be explained with resonance). Pictures are seldom fully self-explaining and the accompanying text is decisive. But a religious approach is not useful.--Wickey-nl (talk) 14:26, 21 August 2010 (UTC)

Resonance hybrids[edit]

I just see "resonance hybrid" is actually a synonym of "contributing structure" → That means the intro should be adapted.--Wickey-nl (talk) 16:17, 17 August 2010 (UTC)

That definition doesn't make much sense. A hybrid of any kind is the product of mixing two things, not one of the ingredients.
Ben (talk) 17:32, 17 August 2010 (UTC)
That is my first association too, so the Gold Book definition surprises me. Or do I interpret it wrong? Or is that definition wrong?--Wickey-nl (talk) 14:27, 18 August 2010 (UTC)
I think you do interpret it wrongly, but it is also a very poor, very unclear definition. It is saying that a VB function is an approximation to the true state function and a resonance hybrid approach, i.e. a mixture of resonance structures, is needed when only one Lewis structure is not good enough. Ben has it right. --Bduke (Discussion) 01:38, 19 August 2010 (UTC)

This subject is already discussed above → What is the definition of "resonance hybrid"
I will cite now what Linus Pauling said about the resonance hybrid:

"In this case the best wave funtion ψ would be formed in part from ψI and in part from ψII and the normal state of the system would be described correspondendingly as involving both structure I and structure II. It has become conventional to speak of such a system as resonating between structures I and II, or as being a resonance hybrid of structures I and II."
[2] Linus Pauling, The Nature of the chemical bond - An Introduction to Modern Structural Chemistry . Third Edition 1960, p.12

Pauling is speaking of the normal state, the state with the lowest possible value of energy (page 11). Structures I and II are contributing structures. In the next paragraph (page 12) he says (in my words):
The resonance hybrid is not exactly intermediate in character between structures I and II, because the resonance stabilized hybrid is lower in energy than either of the contributing structures. As we can suppose the real structure will have the state of lowest possible energy (in the normal state), we can say the real structure is the resonance hybrid.--Wickey-nl (talk) 08:52, 19 August 2010 (UTC)

Only single, double and triple bonds?[edit]

The sentence in the second paragraph, "Each contributing structure can be represented by a Lewis structure, with normal single, double or triple covalent bonds between every pair of adjacent atoms within the structure." is wrong. I see it is supported by the reference to the Gold Book, but it is contradicted by a classic example of resonance for the bridge region of diborane, where the resonance structures have alternating "single" and "no bond" bonds giving each bond in the resonance hybrid to be a half bond. Can anyone suggest a better wording? Note also that diborane also contradicts the statements that imply that resonance is only about pi bonds. --Bduke (Discussion) 22:24, 5 September 2010 (UTC)

First I am defining this comment as a new section, since it does not depend on the previous discussion. My suggestion would be to replace "single, double or triple" with "integer number", which can include not only zero as in diborane, but also in principle quadruple bonds and higher. Also I would eliminate the word "adjacent" which does not appear in the IUPAC source document and is not really well defined here - how does one decide which pairs of atoms are "adjacent"? So "Each contributing structure can be represented by a Lewis structure, with only an integer number of bonds between each pair of atoms within the structure." For distant atoms the integer will of course be zero, unless one includes minor structures such as Dewar benzene. Dirac66 (talk) 00:45, 6 September 2010 (UTC)
Exceptions are already mentioned in the last section, but a separate section about sigma bonds involved in resonance is still missing. About the word "adjacent": every bond is connecting adjacent atoms.--Wickey-nl (talk) 16:03, 7 September 2010 (UTC)
Several points. There is no need for separate sections on sigma and pi bonds. Resonance can occur in either. "every bond is connecting adjacent atoms" is not true. The 1-4 bond in Dewar structures of benzene can hardly be called connecting adjacent atoms. Another example is the 1-3 long bond in the singlet diradical structure of ozone where there is a massive literature saying that it is important. I agree with Dirac66 that "integer number" should be used and "adjacent" should be removed. --Bduke (Discussion) 23:48, 7 September 2010 (UTC)
OK, done. Dirac66 (talk) 00:57, 8 September 2010 (UTC)

Adding an external link[edit]

The following link (Resonance Theory) clarifies the topic and enhances the material covered. Material covered in this website is available to anyone and was written by a tenured PhD Organic Chemistry Professor at Utah Valley University. The website is a non-profit website and is intended to advance students understanding of Organic Chemistry. Thanks for your time, Nickcc20 (talk) 14:45, 19 January 2011 (UTC)

Pauling's principle of electroneutrality[edit]

Pauling's principle of electroneutrality is still taught as being the method by which favourable and unfavourable resonanace structures are "selected". A simple statement is that the charge on an atom should be between +/- 1 (formal charge)with the corollary that the negative charges should reside on the most electronegative atom and positive on more electropositive. Is this worth a mention? Axiosaurus (talk) 13:04, 17 March 2013 (UTC)

Yes, I think this principle is important as it is still used to eliminate some structures which students may draw. It chould be included in the section on Major and minor contibutors. Dirac66 (talk) 18:31, 17 March 2013 (UTC)

Resonance can apply to ionic compounds, metals etc .- well at least Pauling thought so[edit]

I have reread this article and see that it is very molecular in its scope- why? Ionic structures with "covalent character", metals, intermetallics and other unusual solid state substances were all tackled by Pauling, was he wrong? Axiosaurus (talk) 16:52, 17 March 2013 (UTC)

Here I think resonance is very rarely used to describe solid-state bonding. Probably because a good resonance description of a nonmolecular solid would require a large number of resonance structures. So perhaps Pauling was wrong about the importance and usefulness of this approach, and I would not include it in the article. Dirac66 (talk) 18:31, 17 March 2013 (UTC)

Recent edits - Misisconceptions heuristic paragraph[edit]

The statement regarding ionic contributions reads as if it applies to both to homonuclear and heteronuclear bonds. I am not familiar with the referenced book- however other books by Shaik discuss in detail the Heitler-London treatment of H2, is this where this quote comes from? . Historically ionic contributions in A-B bonds were the basis of the electronegativity concept.Axiosaurus (talk) 06:24, 12 April 2015 (UTC)

Historically you also have the Weinberg treatment of H2, which combines ionic structures with covalent structures, so ionic contributions do apply to both homonuclear and heteronuclear bonds. --Bduke (Discussion) 08:08, 12 April 2015 (UTC)
I added the statement to mention excited states of a molecule partly to add meaning to resonance as most people are only familiar with the "mental exercise" to quote one website on the net, and also to draw a comparison to LCAO as that has negative combinations (anti-bonding orbitals). Reading the book again only some excited states require the inclusion of ionic structures. I think I'll remove that reference to ionic structures.--Officer781 (talk) 08:28, 12 April 2015 (UTC)

Charge shift bonding[edit]

Charge shift bonding is not the the same as ionic-covalent resonance. The claim is that some molecules are stable only because of ionic-covalent resonance. In F2 for example the calculations using just covalent terms do not show bonding. Ionic terms on their own are not that good. The claim is that it is resonance between them that is responsible for bonding. I want to stress that charge shift bonding is controversial. For example a generalised valence bond (GVB) function function gives a reasonable description of bonding for F2. Some workers argue that this GVB is just a description of covalent bonding. There is no ionic-covalent resonance. Others argue that it disguises the ionic terms and thus the ionic-covalent resonance. There are wide differences of opinion between researchers on valence bond theory. Take care. --Bduke (Discussion) 11:45, 26 April 2015 (UTC)

Dubious energy diagram for benzene[edit]

I am wondering about the exact meaning and validity of the energy diagram added today for benzene in the section Resonance in quantum mechanics. The file description (obtained by clicking on the image) says it is based on the MO diagram for H2, which is of course well known. However that diagram and all the other diatomic MO diagrams are for individual orbitals (one-electron wave functions). This diagram for benzene appears to show the combination of two many-electron (at least six pi-electron) wave functions, one for each Kekulé structure. Is there a source for combining 2 Kekulé structures in an energy-level diagram, as opposed to the elementary diagram which joins the 2 structures with a simple ↔ ? The lower energy level is acceptable, as it is true that the wave function may be written as a (normalized) sum of two functions representing Kekulé structures. But the upper level is quite mysterious - is it antibonding at all 6 C-C bonds? Or bonding and antibonding at alternate positions, so that there are 3 double pi bonds and 3 antibonds? or null bonds? In the absence of a source, a proper answer to this question would require detailed mathematical analysis of the proposed wave function, which of course would be original research. In any case, I think the upper state would be at very high energy and inaccessible by one- or even two-electron transitions, so it is of no experimental interest.

In summary, I have never seen such a diagram for benzene and I think it requires a better explanation with a source. If this is not available, then I recommend the diagram be deleted. Dirac66 (talk) 15:45, 3 May 2015 (UTC)

[3] In fact, the source states that it is the first excited state of benzene. After checking the net it is a single-electron transition to one of the two LUMOs. [4] the book uses molecular orbital-like diagrams for resonance mixing. See also page 201 which discusses the A1g and B2u state from the Kekule structures.--Officer781 (talk) 02:41, 4 May 2015 (UTC)
OK, thanks for the sources. In the article it might be better to add the page(s) where these points are discussed. Dirac66 (talk) 20:36, 4 May 2015 (UTC)

Last section Charge delocalization has many undefined terms[edit]

The last section on Charge delocalization contains mysterious acronyms and other terms. In encyclopedia articles, terms likely to be unfamiliar to many readers should either be explained at first usage or else linked to another article which does provide an explanation. So would someone please provide the answers to the following questions?

  1. What does WAPS stand for? None of the definitions at WAPS is at all relevant.
  2. What does WANS stand for? The article WANS is totally irrelevant.
  3. What is fluoradene? We have no article on this molecule which is not well-known.
  4. What is dma given as a solvent for t-BuP4? The page DMA includes 3 solvents: dimethylacetamide, dimethylamine and dimethylaniline. Which one is meant here?
  5. What is MTBD? I can see by holding my cursor over the formula, but why not write it out?
  6. What are DBU and TBD? OK, these two do have links to articles. But would it not be clearer to write the names out in the table and put the acronyms in parentheses? Dirac66 (talk) 18:28, 24 May 2015 (UTC)
WAPS - weighted average positive σ, introduced in ref 21, a J. Chem. Phys (A) paper. WANS - weighted average negative sigma, introduced in ref 22, a Journal of Solution Chemistry article. They are both calculated using the COSMO program, but I know nothing of them. I'll leave the chemicals to an organic chemist. This section does seem over-detailed. --Bduke (Discussion) 21:45, 24 May 2015 (UTC)
Thanks. I realized that your information on WAPS and WANS comes from the free-access first page of refs. 21 and 22. With that start to encourage me, I also checked refs. 23 and 24. Ref.24 is an entirely free-access paper from CROATICA CHEMICA ACTA (in English) and has a defining equation on the third page for the parameter WANS. Sigma is the polarization charge density - not sure if that is the same as polarization density.
As for over-detailed, the discussion of WAPS/WANS is 3 sentences plus a table of 22 values. Perhaps we should keep the 3 sentences, add a reference to the Croatian paper with the free-access definition (for WANS), and drastically reduce the number of values in the table. I think 4-6 well-known molecules would be sufficient. Dirac66 (talk) 00:45, 25 May 2015 (UTC)
I have now written out WAPS, WANS in full, and also displayed the names of MTBD, DBU and TBD in full. I left most of the items in the table, but deleted fluoradene whose structure is neither in Wikipedia nor in the source article, as well as t-BuP4 because the solvent dma is not identified in the source article. Dirac66 (talk) 22:00, 8 September 2015 (UTC)
"(dma)" is a ligand in the structure—dimethylamino—not the solvent. See doi:0.1021/ja053543n. See Phosphazene#Phosphazene bases for sample structure of "P4", having substitutents on the phosphazine's N and secondary-amino groups on the other phosphazines' P. I generally dislike long tables of primary-sourced data like this, especially when called "common" examples. Demonstrate the range and how they compare to one or two well-known strong and weak acids/bases. DMacks (talk) 01:52, 9 September 2015 (UTC)