Talk:Chemical bond

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Chemistry (Rated B-class, Top-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.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 Top  This article has been rated as Top-importance on the project's importance scale.
Wikipedia Version 1.0 Editorial Team / Vital
WikiProject icon This article has been reviewed by the Version 1.0 Editorial Team.
B-Class article B  This article has been rated as B-Class on the quality scale.
Taskforce icon
This article is a vital article.

'03 comments[edit]

Some stuff that could be added to this article includes:

  1. Naming bonds based on their symmetry: sigma, pi, delta.
  2. Bond order and associated nomenclature: single bond, double bond, triple bonds, bonding, non-bonding, and anti-bonding orbitals.
  3. The depiction of bonding: Lewis electron dot structures, line drawings with explicit atom lables, organic bond line drawings in which carbon atom positions can be inferred to exist at the vertices, 3-D wedge diagrams with dashed lines for bonds pointing "back" and solid wedges indicating bonds that point "forward". Stereochemical uncertainty indicated by a bond shown as a wavy line.
  4. Ways of characterizing bonds--bond dissociation energies, atomic radii and bond lengths, vibrational spectroscopy, conjugation and UV-visible spectroscopy.

Much more to add than I can do justice to at the moment, but if I put this here it will help get me started more quickly next time, or maybe prompt someone else to contribute.

Other points. Valency fails with coordination compounds. Octets have no meaning when the bonding happens between (s,p) electrons and d electrons. So why is valency pointed to with the second sentence when valence bonding per se is not general? Dwmyers 18:48 Feb 18, 2003 (UTC)

'05 comments[edit]

Weak chemical bonds are sometimes major players in complex stereochemistry. But for simpler molecules they only effect in the compounds' physical properties. Modern science handles these bonds differently from the other types of bonds (covalent, polar, ionic, metallic). F.e. there is no LCAO (or molecular orbitals) for these. Melianis 13:50, 7 December 2005 (UTC) Expanded the article, tried to make it more understandable by classical terms. This article is in serious need of pictures (differences between various bonding types). I have a few ideas but don't know if i'm up to the task. Anyway, i'm not going to bother if they're accurate or not, basic idea should be enough. Melianis 15:17, 9 December 2005 (UTC)

'06 comments[edit]

I have just reviewed the discussion of bonding, and am concerned that the seminal concept of electronegativity - which drives polar properties of covalent bonds, giving rise to issue such as the ability of HCl(gas) to dissociate in water to form H30+ and Cl- ions, or to cause localised +ve or -ve charges responsible for tertairy structure in proteins) is not here. Have I simply missed seeing the relevant entry? Any feedback would be appreciated. If there is not mention within Wiki of the concept I'll start on it. Many thanks. JohnT 09:26, 8 January 2006 (UTC)

It would also be helpful if there would be something explaining bonding in intercalation compounds, such as those appearing in rechargeable Li-ion batteries (explanation of the process of bond creation and destruction). I have completely no idea about the mechanism leading to the bond formation, and even less on how to classify this type of bond. I hope someone could help in this direction.(Maybe a seperate article is necessary?)

Proposal for new version[edit]

The current version has a strong Pauling-POV, and so I have tried to present the same facts in a more neutral light here. All comments are welcome. Physchim62 (talk) 11:13, 19 July 2006 (UTC)

I like your re-write but would encourage a few changes. One is that the really basic description of a bond as an electrostatically driven thing, should not be held back for the H2 covalent section, but should be discussed up front in the article. The person who wants to know what a chemical bond IS and why it FORMS, should not have to wait to find out. The two electron clouds should repel, but at a certain distance they sometimes don't, but actually atract with some vigor. Why? Because electrostatically a pair of outer electrons would like to spend time between two nuclei any time they can get away with it, where they can feel both sets of (approximately +1) screened nuclear charges, and the number of pairs (or half-pairs) that do this (the bonding electrons) depends on how big this screened charge, and this intra-atom or intra-ionic space can get, which is set by how many electrons have "room" to get into that intranuclear space. According to QM rules for orbitals and the space allowed (which may a lot larger for bigger atoms and bigger ionic atoms and atom-assemblies (anion clumps) in crystals where the electrons have been nearly completely transferred from one assembly to another). But the basic electrostatic reason for being, is more or less true for bonds of all polarities from ionic to covalent. Equal numbers of heavy and light ping pong balls with pos and neg charges should come together with two negatives between each set of positives (in 2 and 3 dimensions). The WHY of bonding is electrostatics, but modified by the limitations of the exclusion principle for how densely you can pack electrons, and also other QM momentum/space constraints on holding down light electrons to small volumes.
Also, I would suggest remembering that primary audience for this article may be very wide, including high school and grade school students. So all math, with exception of bond strength numbers, perhaps, should be left in sub-articles. [Example: when I see two Greek letters added and subtracted from each other, with a factor root2, I'm used to that from basic QM, but anybody without a few years of QM will suffer the MEGO (My Eyes Glaze Over) from that point on.] All new concepts (even linked ones, which most of these are) should be introduced in the very brutally simplest of metaphor (think 10th grader), before you go on.
You've expanded the history here a lot, and there's a running debate as to where this should go. It's okay with me as you have it, but in the past it's been ofloaded to other articles.
And yes, we need a lot more pictures.SBHarris 15:19, 19 July 2006 (UTC)


I have added the disputed tag, because this article puts forward a point of view (POV) that molecular orbital theory is clearly superior to valence bond theory. This POV is false. We need to look at both qualitative and quantitative approaches. In quantitative approaches, methods that go beyond molecular orbital theory (Post-Hartree-Fock methods) can indeed give very accurate results. Molecular orbital theory on its own (Hartree-Fock theory) can not give good results and is particularly bad for bond breaking processes. Here multi-reference wave functions are needed. Modern valence methods are essentially equivalent to those MO methods and give an excellent description of bond breaking. As a simple example MO theory dissociated the molecule F2 into a mixture of 2F atoms and F- + F+, with an energy so much higher than 2F atoms that the energy at the miniumum of the curve for F2 is still higher than that of 2F atoms. So why is MO theory so widely used? It is easier to compute because MOs are orthogonal. VB stuctures are not. It is more black box. However many Valence bond codes for modern valence bond theory can now compete with MO programs although they are still not as black-box (easy to use). Qualitatively, MO ideas have become very popular, but the examples given in this article of qualitative VB ideas are very old ideas that have been revised in recent years. I recommend that people read "Valence Bond Theory, its History, Fundamental and Applications. A Primer" by Sason Shaik and Phillippe C. Hiberty in Reviews of Computational Chemistry, Vol. 20, Page 1-, 2004.

I could go ahead and rewrite the article, but I hope to first convince the editors here that they are putting forward a false POV. Maybe I will rewrite it in future when I have more time. --Bduke 05:01, 26 November 2006 (UTC)

Hey, you've convinced me that you know more about the subject than I do, and I'm the guy who was just fleshing out the MO section (although I wasn't the one who originally promo'd it).

We are in serious need of expertise on Wikipedia, but REAL experts on the matter you bring up (say somebody who'se actually worked with MO vs VB calculation programs) are few and FAR between. That sounds like you. So contribute! Please! I didn't even know a modern valence bond theory wiki existed, until I went back and saw that (indeed) it's referenced at the end of valence bond theory, and is indeed mathematical enough to warrant a separate Wiki, even though the judgemental opinions in it need to be mined out and stuck in some more general articles, like this one.

Anyway, re-write away. Remove POV comments as you think best, but be sure to note that this is fairly recent perspective you bring (those papers are all 21st cen).

Not really - the points that Bduke raises have been discussed in the literature going back to the 1950's, and the basic complementarity between VB and MO theory has been a standard topic in Physical Chemistry textbooks for decades (just about every P.Chem. book works out, or assigns as a homework problem, the demonstration that VB + ionic terms is exactly equivalent to MO + configuration interaction.) The problem, I think, is that the physical organic and physical inorganic chemists noticed that MO gave a much better picture of the problems that they interested in at the time, and came to the incorrect conclusion that it was better in general. They didn't notice that, for example, MO can't describe something as basic as how a molecule dissociates into atoms. --Rparson 23:00, 26 November 2006 (UTC)
Well, it seems to me that MO theory CAN do that, if you give it the same breaks as "modern" valence bond theory, as BDuke points out in the MCSCF Wiki, and (as I read you) you yourself admit above also. So it seems to me (correct me if wrong) you're not quite being fair here. If they're indeed complementary theories, then they give the same numbers and are equally "true" for those problems in which they do-- but that's not saying they're complimentary for all cases.

We're getting into nomenclature issues, where old MO theory with bad H-F approximations (what you call "MO theory on its own") is getting blamed by you for the poor H-F approximations used to calculate it, when there are better "post-H-F" approximations for the sorts of ionic-dissociation systems you're describing, as you know, and which still assume MO's. So it's not MO's fault, and this is still MO theory, and I don't think it's fair to say these things go "beyond" MO theory, so that you have to call these things post-HF or modern-MO, or post-modern-MO or whatever. MO theory is the basic idea that electrons in molecules go into molecular orbitals. If you pull the thing completely apart, the molecular orbitals need to start looking atomic and you should stop worrying about ionization states-- okay. YOur MO theory will do that if you add all the needed parts. I'm simply starting out with the assumption that the basic MO picture (every electron goes everywhere in the molecule) is the "true" one. By which I mean that if you could write down every last interaction term in the Hamiltonian and calculate them all with the biggest supercomputer and gigantic program, MO's would be what you'd find if you graphed the solutions--wavefunctions for each electron which describe a pattern distributed over the whole molecule (though non-uniformly in most instances, but not all). Those electrons are in molecular orbitals-- what else do we call them? And those orbitals really exist in some kind of objective way, at least in the same way that atomic orbitals exist in atoms in that funny quantum world. And so also with antibonding orbitals, as we find when we add electrons and find ourselves sometimes weakening a bond. And this view encompasses all the rest of the bond theories, even if sometimes it's unusable in practice. If in some cases simplifying assumptions must be made which MAY convert it into something equivalent to "modern" valence theory, that's not to say that valence theory is just as "true" for EVERY case. In physics we assume the more general case is the better picture, do we not? Limiting cases, even if they give the same answers, and even if they give them more easily, are not given the same respect. General relativity (GR) is "truer" than Newtonian gravity, even if in going to the moon we're more or less relegated to approximating GR down to the same equations Newton came up with, more or less. If modern versions of MO theory (adding configuration interactions) can describe open-shell dissociation energies, AND in the meantime describe spectroscopy in great big molecules, then should we not note it as the encompassing paradigm? I mean, how do YOU really think about that 13th bonding electron in benzene? Isn't it closer to "objective truth" to think of it as spread all over the whole smear, totally uniformly in one big pi cloud? SBHarris 01:57, 27 November 2006 (UTC)

I'm sorry but the above is largely wrong, although I grant that there are nomenclature problems. However it simply is not true that if one calculated the exact wavefunction it would look like MOs. No one-electron functions, which is what MOs are, would be seen. MO theory is when we take a set of one-electron functions and "occupy' them, usually 2 at a time for closed shell molecules. The result is what is called a single determinant function. The orbitals do not necessarily spread over the whole molecule. We can convert the delocalised orbitals that you think are reality and tranform them to localised orbitals by a simple transformation that leaves the total wavefunction the same.
If we take several determinants where that MO determinant is the first and the others are created by removing an electron from an orbital in the first (i.e from a 'bonding' orbital) and put it into a virtual orbital (one created by the calculation of the first determinant but not used, i.e. not occupied) then we get configuration interaction. All post-Hartree-Fock methods are essentially this, but are formulated differently. The individual MO occupancies are lost.
The important theorem that gives the ultimate equivalence of MO and VB based approaches is this:-
  1. we take a set of N atomic orbitals (AO), where N is equal or larger than the number of electrons.
  2. we form a set of N MOs from these N AOs and create many determinants by assigning the electrons in all possible ways to these MOs. This is called full configuration interaction.
  3. we form all possible covalent and ionic resonance structures from these N atomic orbitals, each represented by a determinant. This is full valence bond.
then in 2 and 3 we have the same number of deteminants and an optimised linear of combination in each case gives exactly the same wavefunction, energy, properties etc. They are equivalent. Example, H2 using 2 2s atomic orbitals. We form two MOs. We can put both electrons in one or both in the other (one in each does not interact with these due to different symmetry). We have two determinants. That is full configuration interaction. We now create the covalent structure (Heitler-London) and the ionic structure (H+ H- <--> H- H+) and mix these two functions. The two results are equal.
However note that the CI function, while based on molecular orbitals is not the simple MO function with 2 electrons for H2 in a bonding orbital. It is a mixture of 2 electrons in the bonding orbital and 2 electrons in the antibonding orbital. Note also there are an infinite number of other ways of getting the equivalent of full configuration or full ionic-covalent resonance. All give the same result. You can not say that molecular orbitals are ultimate reality.
In practice we have a large number of atomic orbitals and the resulting number of determinats (configurations or structures) is even larger. A 100 atomic orbitals can give millions of configurations. If we take all configurations and all structures the two approaches is equivalent. In practice we take less than the full number and the MO-based approaches and the VB-based appoaches are not equivalent but compliment each other.
MO theory has won out in programs like Gaussians because it is computationally simple, not because it is closer to reality or whatever. Good calculations however are really not MO, just based on the orbitals found in a MO calculation. Qualitatively both MO and VB ideas are used and do complement each other. Increasingly good VB calculations are being published, not just in the last 6 years - some go back to the 1980s.
I see some good changes to the article have been done. I will try to add something myself when it is clear that the editor who made these changes has finished for a while. The detail of above is not needed of course at this level but we should not make statements that are false. --Bduke 03:12, 27 November 2006 (UTC)
You say: I'm sorry but the above is largely wrong, although I grant that there are nomenclature problems. However it simply is not true that if one calculated the exact wavefunction it would look like MOs. No one-electron functions, which is what MOs are, would be seen. COMMENT: Huh? News to me. Then what would be seen? IOW, let me back up a step, and ask whether, if you calculated the wavefunction for a multielectron atom, if you'd see electrons in ye olde atomic orbitals? No one-electron wave-functions there, either, and the analogy to MO is exact (really it is exact-- you cannot give me any argument for the unreality of MOs that I can't return for multi-electon AOs). When I look at a periodic table and the electron structure of (say) lithium is written 1s22s, is this a matter of chemists handing me a convenient fiction because they're too lazy to do a full computer program? Is this an approximation which is put up there by analogy with the Schroedinger hydrogen atom, and which works well enough to explain the periodic table, but actually is false? Or no more true than many another analogy, which is to say it won out because it was computationally simple, but is no closer to reality than any other view?? So what should I think of it? These spectroscopists seem to believe in these atomic orbitals as though they were objective. SBHarris 08:26, 1 December 2006 (UTC)

Let me try to be reasonably clear. First, I do not want to use terms like false or analogy. Let us start with lithium as (1s)2(2s). This is an approximation. It is pretty good, but it assumes, for example, that the repulsion between the two 1s electrons is averaged. It does not take into account that if one electron is close to the nucleus, the other will be further away, or if one is to one side of the atom, the other will tend to be at the other side. It is simplifying the repulsion between electrons. The calculation of He as (1s)2 is an approximation. It gives a pretty poor value of the total energy. Accurate calculations of He give agreement with experiment, but there are no atomic orbitals in them. Do spectroscopists think orbitals are real? No, they do not. They think atomic and molecular states are real. The difference? Orbitals are about one electron at a time. States are about all the electrons. So, good calculations do not show electrons in 'ye old atomic orbitals'. They are a usefull, but approximate idea. Chemists do , sort of, give you a 'convenient fiction'. The problem is always that chemists want to use simple ideas, but this always means making approximations. The exact solutions are just mathematics and they lose the simple pictures. As chemists we have to play one side against the other - get good numbers but have good pictures to think about. It is always difficult. --Bduke 10:08, 1 December 2006 (UTC)

  • Okay, we are making progress-- good. And when you get this explained to me, you will be able to explain it to those who know no chemistry at all. Yes, specroscopists think states are real, but for many cases these states are close to linear combinations of AO's. Examples: the Ritz-Rydberg laws would not work otherwise, and Moseley's law basically treats the K-alpha line in all elements as a hydrogenic Lyman-alpha, corrected by a (Z-1)2 term, so you see screening by one electron (the remaining 1s, we presume) and no others. A simple AO picture here works for the heaviest atoms, even. Is that the full quantum picture? No. But here it gives correct answers for K-alpha, even for energy, for any element. And to some extent it's rather unfair to focus on energy itself as an issue, because this is where AO combination fails worst, due to simple screening. On the other hand, the screening factors tend to be simple, and they affect orbital energies and sizes, not so much shapes and behaviors (which is a lot of chemistry). Thus elements get denser down columns but tend to behave chemically and physically in similar ways. The periodic table itself results from the approximation to one-electron AO combinations, and where it begins to fail, it's the result of relativistic effects and differential screening, and the diagonalization of chem properties on the table is still explained by an AO picture.

    For example, I have a modern periodic table on my wall which informs me that the electron structure of rhodium is [Kr]4d85s1. Those are atomic orbitals, as "real" as molecular orbitals. But of course this notation presumes that there is such a thing as a "5s orbital" in this atom, and that this thing holds 1 electron, not 2. And that somebody has performed some kind of objective test to decide that the atom does NOT have its outer 9 electrons in (say) the d7s2 configuration of iridium, the element below it. And that is why these notations are up on my wall, as the modern chemist's view of "reality." If there wasn't something special in this picture, I presume that instead, there'd just be a note telling me that (hey, dummy!) for multielectron atoms there are no atomic orbitals in any sort of real or objective sense, so therefore they can't be differentially populated and notated, and that any approximation to the real picture for atoms is no better than any other, except computationally. And furthermore that the aufbau picture of electrons building up in orbitals/shells (whatever you want to call these things) which had been noticed as early as 1920, and which explains the periodicity of the chemical properties of the elements so well, and which also happens to come out in terms of combinations of one-electron states for Schrödinger hydrogenic atoms, is (as a picture) not worth any special comment, and all this experience does not make atomic orbitals "real" at all, at all, so forget it.

    But since, strangely, that is not what's up on my wall, I remain unconvinced by your argument, and obviously so also the ACS and the folks who made my poster, convinced that this approximation to the truth was more useful than any other. The full calculations, I have read, show electron clouds that look very much like (though of course never exactly) combinations of AO's. Example, for lithium, full computation of probability density will give you a total electron cloud which looks like two nested spheres, even if they are the "wrong" size for combinations of hydrogen 1s and 2s AOs.

    I suppose it all comes down to your idea of ontology. What is real in physics or chemistry, anyway? Is Newton's inverse square law picture of gravity "real"? Or should we say that it's simply false, and that even as an approximation has no more reality than is conferred by ease of computation in some circumstances? I think that's the wrong way to look at it. Or present it, for that matter. If you let yourself get snookered into that game, then nothing is real, and no progress is ever made by approximation and the pictures that come with approximation, since any calculation may fail tomorrow when it's found that it doesn't perfectly fit experimental results to the last significant digit. Do you see the point? Good and useful and simple approximations do have a kind of reality. In fact THAT is the only kind of reality we ever see. SBHarris 03:17, 2 December 2006 (UTC)

I am not going to respond to all of this. The Hartree-Fock method based on electron configurations for atoms gives good approximations to the properties of atoms but it is not exact. For example the best we can do with (1s)2 for helium gives a total energy of -77.9 eV and an ionisation energy of 25.0 eV. The experimental values are -79.0 eV and 24.6 eV. We can do calculations that get that bit extra but we lose the idea of occupied orbitals. Of course in atoms different electron configurations correspond to different spectroscopic states, so the electron configuration on your periodic table is the one that gives the correct state for the ground state of the atom and the basis for this being approximated by a Hartree-Fock calculation. You may think that these differences from experiment for total energies are small, but for molecules and reactions we are looking at small differences between two systems and the small errors can lead to complete nonsense in predicting reaction energies. The exact wavefunction, i.e. one that gives exact agreement with experiment has in it no one-electron orbitals, either AOs or MOs. AOs and MOs remain a usefull approximation. Yes, approximations do have a kind of reality, but the one-electron approximation in chemistry can be seriously misleading. We are not talking about very small differences. I know of lots of cases where molecules that have been found experimentally, although unstable, do not even exist at the Hartree-Fock approximation level. --Bduke 04:23, 2 December 2006 (UTC)

Let me add a further comment. I think Steve's analogy between atomic orbitals and molecular orbitals misses some important points. AO's really are a good first-order approximation for describing atomic properties, at least in ground states, in that there aren't too many places where the AO picture leads you to qualitatively incorrect conclusions. You can get by slapping on a few empirical fixes to account for those properties that depend inherently upon correlation (e.g. Hund's Rule). Molecular orbitals, on the other hand, really aren't a very good first-order description of the way electrons are distributed in a chemical bond. The valence bond wave function is somewhat better, but it isn't all that close either. As a result, both pictures lead right away to qualitatively false predictions - MO gets dissociation products all wrong, VB gets paramagnetism all wrong, and you can't fix these up without changing the wave function in a big way, by adding CI to MO or by adding ionic terms to VB. So when thinking qualitatively abound chemical bonding, you have to keep both pictures in mind, and use whichever is most appropriate. It's a bit like Bohr's "comlementarity" - the physical picture that one adopts depends upon the experiment in which one is interested. In quantitative ab-initio work, MO won out not because it was closer to reality than VB (it isn't), but because it was better adapted to computation - even though MO had farther to go, the road was much easier, for reasons that are mathematical rather than physical. --Rparson 17:55, 2 December 2006 (UTC)


The last half of the 20th century (to first approx) saw MO theory slowly winning out over Pauling, and if the balance has since tipped, when and how and for what problem classes needs to be noted, just as you've done it in the non-math parts of the modern VB article. Try to keep this one non-mathematical, but if you see examples you don't like or that you think are misleading, fix them with replacements. You know enough to do it.

Yeah, I know you don't have time for full rewrite. But please note the tricky way Wikipedia has of drawing out the time of true experts in various fields. Wikipedians write articles which are get better and better until some true expert reads them and says "Gak, that's SORT of right, and I can't STAND the difference between this and true". So you go to the TALK page and complain (see, we suckered you out of that much of your time). And those who did the rough work then cheer you on to do whatever you can to be WP:BOLD and WP:SOFIXIT. Go for it. Start with the simplest metaphors you can, and work your way up. Compare and contrast methods. Remember you're writing for an audience who's mostly never seen the math function for a wavefunction, so pretend you're Steven Weinberg writing The First Three Minutes. No question you know what you're talking about, but you have to explain it anyway, and give real-world examples (benzene, carotene, graphite). SBHarris 05:24, 26 November 2006 (UTC)

I've made a start on the first couple of sections - it may be too jargon-ridden, feel free to simplify. Rparson 23:00, 26 November 2006 (UTC)
  • Well Bonding is an important topic, and one that requires good editing. I would welcome a theorist taking a whack. The article was awful, failing to distinguish "descriptions" and artificial "classifications" from reality. So I plead guilty to shifting to the implication that VB is inferior to MO. I am no theorist but try to find a grad level course in organic or inorganic that pushes hybridization. Or requires d-orbitals to explain PF5. Or views coordination compounds as a special kind of coordinate covalent bond, an article that I find particularly lame. Or explains the PES spectrum of CH4 with VB. So I took the charge of "be bold" seriously, but by the same token, I welcome a healthy balance based on modern analysis written by experts.--Smokefoot 06:02, 26 November 2006 (UTC)

I have rewritten the sections on MO and VB theory and hopefully this is now better balanced and better structured. If you think I have taken important material out, please ask me why. I removed a lot of material that was not quite correct or not clear. It needs referencing properly and I will get around to that in a few days. I have removed the disputed tag now. --Bduke 07:05, 27 November 2006 (UTC) Fascinating discussion guys. This article is much better than when I first read it. 18:26, 29 June 2007 (UTC)

Intermolecular Bonding[edit]

The section on intermolecular bonding is deficient - it omits several types of intermolecular force (e.g. higher order multipole interactions such as dipole-quadrupole or quadrupole quadrupole, and induction forces such as ion-induced dipole.) I could add these in, but I wonder if it's better to just cut thhe entire section, as it's only tangentially related to the subject of the article. The article is about Chemical Bonds, after all, and this section is about interactions that are for the most part not generally referred to as chemical bonds (hydrogen bonds being the exception, perhaps there should just be a section on Hydrogen Bonds and toss out the rest.) Rparson 18:44, 28 November 2006 (UTC)

I agree, but first, we need to fix something. I have altered two main article links. Dipole-dipole attraction redirected to Intermolecular force so I altered that. We should have a link there. Instantaneous dipole attraction also directs there. However Van der Waals forces is a quite separate article. I just fixed a small redirect issue with it. Perhaps Van der Waals forces and Intermolecular force need to be merged. I'll look into it but not immediately. --Bduke 21:38, 28 November 2006 (UTC)

Good work![edit]

This article is really improving nicely. I would suggest adding more references (aim for ten) and add an external links section to “good articles” on chemical bonding. Keep up the good work: --Sadi Carnot 18:10, 10 December 2006 (UTC)

Banana bond[edit]

The article currently gives diborane as an example of banana bonds. I think that this usage of the term "banana bond" is nonstandard, and three-center two-electron bond is a more appropriate term. Banana bond is more commonly used to refer to "bent" bonds in cyclopropane, or to a view of the C=C double bond as a combination of two banana bonds instead of a combination of a sigma bond plus a pi bond. Note that the current version of Diborane doesn't mention bananas at all, and the current version of banana bond doesn't mention diborane at all. I do recognize that the bonding in diborane is sometimes informally called "banana" (and deserves a brief mention in the banana bond article), but it would be more precise to simply call it a three-center two-electron bond. Itub 22:46, 18 December 2006 (UTC)

Note: I just expanded banana bond and added a link to three-center two-electron bond, as well as a section about double and triple bonds. Itub 00:47, 19 December 2006 (UTC)
Agree. Looks like you fixed the problem nicely, yourself. Isn't it nice when Wikipedia works like that? SBHarris 01:14, 19 December 2006 (UTC)

Comment moved from article[edit]

((EDIT: I wish I knew what was real on this page, and what isn't. -Lindsey)) (added by User:

It might help if the anon explained this point. --Bduke 01:18, 12 January 2007 (UTC)

hybridization or hybridisation?[edit]

The spelling is now a hybridiz/sation of US and UK--P.wormer 01:01, 24 March 2007 (UTC)

Proposal to create Simple bonding ?[edit]

Is there any possibility that we would be able to create such an article? I think it'd be an important distinction to explain the fundamentals of simplified bonding theories in an article away from a general one; stuff such as lewis theory, VSEPR, Ligand-packing and soforth. Gaim.svg ♥♥ ΜÏΠЄSΓRΘΠ€ ♥♥ slurp me! 21:01, 4 April 2007 (UTC)

Amplitude addition resulting in charge expansion[edit]

There is some kind of confused comment about this now, added by Enormousdude, in the INTRO. It implies that the nature of a bond depends somehow on constructive interference between two electrons. It doesn't. There are plenty of one-electron bonds (eg H2+), and two-electron bonds are NOT 4 times as strong as one-electron bonds. In fact, they are generally less than half as strong, due to electron-electron repulsion. (Example: bond enthalpy in H2+ is about 2.65 eV and in H2 it is 4.0 eV). So I'm going to remove this comment in a few days, unless somebody can convince me otherwise. SBHarris 20:23, 18 April 2007 (UTC)

I noticed that too. It is certainly confusing. The article does need a simple explanation of the covalent bond and it needs to be correct. --Bduke 22:30, 18 April 2007 (UTC)
I don't think he meant the interference of two electrons, but the interference of the functions that you could use to describe the unbound system (as in sp3 orbital + sp3 orbital -> sigma orbital, with an increased electron density between the two nuclei). --Itub 06:20, 26 April 2007 (UTC)
Comment should stay removed, as per Sbharris. I think we have a good-faith edit by a physcics undergrad who hasn't quite undertood all of his or her courses. Physchim62 (talk) 13:22, 26 April 2007 (UTC)

'Bonded' vs 'bound'[edit]

I think that perhaps this is a good place to ask for advice, as many Wikipedia articles suffer from this (or in my ignorance, I think they suffer from this). My question is the following: what is the correct use of bind|bound|bound and bond|bonded|bonded when talking about chemical bonds? I mean, is the expression "atoms A and B are unbound" correct, to mean "atoms A and B are not bonded"? To me the former sounds like "the size of the atoms is unbound (i.e. it is unlimited)", yet I see it in many places. Other uses include "binding energy" (which is correct), or "atoms A and B bind together" (how about "atoms A and B bond together", or "atom A bonds atom B" or "atom A binds (to) atom B"). I must confess that after a Ph.D. in chemistry, this silly thing still haunts me (obviously, I am not a native English speaker). — isilanes (talk|contribs) 11:04, 4 June 2007 (UTC)

Page Focus[edit]

I recently started a page called chemical bonding models a modest page at best which is currently under dispute. I started this page since I considered incorporating segments on Valence shell electron pair repulsion (VSEPR) Theory, Crystal Field Theory (CFT), and Ligand Field Theory (LFT) into chemical bonding along side the discussion of Valence Bond Theory (VBT) and Molecular Orbital (MO) Theory and the well written "Comparison of valence bond and molecular orbital theory". I didn't go this route since chemical bonding seemed bloated and ripe for division into separate pages; especially if I added these segments which don't deserve equal footing with VBT and MO theory.
Another example of this bloat is that all the bonds types are listed on equal terms. The major bonding types or strong chemical bonds as they are referred to on this page are covalent, ionic, and metallic; even coordinate covalent bond is an subcategory of covalent bonding. The minor bonding types should some how be identified as less significant. Does there need to be a page dedicated to "types of chemical bonds" to collectively list things like Bent bonds, Polar covalent bond, multi-center multi-electron bonds and the others minor bonds in a single place? The same idea can be extended to the section on "Intermolecular bonding" while I realize that Cation-pi interaction should be noted as a intermolecular bonding event its not on equal footing with Hydrogen bonding. Should there also be a "List of Intermolecular bonding"?
The question I pose is: Does the page simple increase in size to the point where the significant information is diluted by second tier information or do we give the second tier their own venue? How has this been handled on other pages?--OMCV (talk) 05:05, 19 March 2008 (UTC)
The usual pattern on Wikipedia is indeed that (as information is added from various sources) length increases to the point that it's obvious that some of it must be spun off into a subarticle(s). At that time, somebody with a good overview knowledge has to perform a delicate operation in which less-important and/or more-technical stuff gets spun off, with some semblance of summary of it left behind. Sometimes that summary already exists in the main article. Othertimes not, and it must be written at the time the other stuff goes off.

The real pain happens with the "and/or" above. It's usually the case that some more-technical and lengthy parts that need to be spun off due to length-alone, are actually of higher quality than the more general parts that need to stay. So doing the spin-off subtly decreases the overall quality of the parent article, due to dilution. But there's no help for it. So long as people insist on writing what they know and are interested in, vs. what the encyclopedia NEEDS :), there are going to be problems of unevenness of content. These show up as particularly threadbare patches in the main rug, that become more apparent every time a lovely new Persian throw-rug gets so big that it has to be moved to another room of the house. SBHarris 18:50, 9 October 2008 (UTC)


As I understand it - the overview and introduction are strictly correct. They however will cause anyone who does not understand the subject to be very confused.

Chemical bonds hold atoms together in various ways, with different strengths, to form molecules and compounds. Everyday objects are all comprised of atoms bound together in several ways. The accurate description of why this happens is complex, and described by the laws of quantum electrodynamics. In practice, simpler approximations are often used.

Simplifying, there are two basic types of bonding, ionic, which involve the sharing or transfer of electrons between atoms, and the usually weaker covalent bonding, which does not.(subtly incorrect)

Something of this sort of length and level of complexity perhaps for the first couple of paragraphs, going on to a simple description of the main sorts?

I wonder why quantum electrodynamics needs to be mentioned in the lead at all, particularly with the suggestion that instead, quantum mechanics is used as an approximation. Quantum electrodynamics is a more accurate but more complex version of QM (one that takes into account subtle effects caused by quantization of the EM fields themselves), and so it really doesn't have to be mentioned as an "alternative" to QM. All these things are QM. The question is what approximations we want to make. Are the fields to be approximated as "even" with no vacuum polarization into virtual particles? Are the mechanics to be approximated as non-relativistic? Are multi-electron wavefunctions to be approximated as linear combinations of single electron wavefunctions, plus fudging? And so on. I think the LEAD really only needs to mention that QM explains all (so far as we can tell from the very high agreement (11 sig digit) between theory and measurement for simple systems), but approximations are necessary in nearly all complex systems to get answers, as with all engineering problems. I'll write a little suggestion, and others can restore all or part of the present lead, as they like. SBHarris 19:07, 9 October 2008 (UTC)

Sulfur/pi interactions[edit]

Can someone please write a small section on sulfur/pi interactions? --kupirijo (talk) 17:37, 9 October 2008 (UTC)

Basics are missing[edit]

The intro in the article electromagnetic force states correctly that: It is the electromagnetic force that holds electrons and protons together in atoms, and which hold atoms together to make molecules. This is the basis of the covalent "bond". The electromagnetic force is also the basis for the other chemical bonds. But this central fact is not present in this article on the chemical bond. I found that my chemistry course book is also very unclear on this so cannot give a good reference. I would say something like:

All chemical bonds are examples of coulomb attraction in particular configurations of atoms. In particular this is the case for the covalent bond. When two or more atoms are in the vicinity of each other there is a complex interaction between all the electrons and all the nuclei. There are certain configurations in terms of distances and angles that form local optima (small changes from these configurations require addition of energy from outside the system). Detailed descriptions and motivations of this are possible only by use of quantum mechanical models. For a particular element there is a typical number of directions and angles between directions where other atoms are to be placed for a stable configuration to be formed. The distance depends on relations between the atoms in the pair. When two atoms take part in such a configuration they are said to bond with each other. In this perspective, the bond can be said to be the relation between two of the atoms in such a configuration. The bond is not an entity in itself. But each bond has a relation to a particular electron that "wanders" close to both the atoms that are bonded. The bond is similar to the relation between me and earth. It is the thing that makes me mostly stay in place instead of floating around. The main difference is that may bond to the earth is a way of viewing an interplay between gravitation and electromagnetism, while the chemical bond involves only electromagnetism. --Ettrig (talk) 10:10, 18 October 2008 (UTC)

My bad, this is described well under the headline "Overview". --Ettrig (talk) 12:20, 18 October 2008 (UTC)

OH bond energy[edit]

I was looking at the bond energies table, and it gives the oxygen-hydrogen bond as being equal to 366 kJ/mol. I think this is wrong. I did the calculation for the oxidation of hydrogen to give water, and by my working, assuming this releases 572kJ per mole of oxygen molecules, and using the other bond energies in the table, the OH bond should be 486 kJ/mol. I appreciate that these are approximate values that depend on the entire molecule etc, but I would assume that they wouldn't vary by 30%, and so the 366 value is a typo that should read 466. [1] Mathemancer (talk) 07:25, 27 December 2008 (UTC)

Hmm, yes, there's a problem there, and the figure is correctly taken from the source so the problem is in the source. The dissociation enthalpy of the O–H bond in water is 542.8 kJ/mol, according to NIST data (although I know where you got your slightly lower figure from). A very weak O–H bond would be that in hydrogen peroxide, which has a dissociation enthalpy of 356.2 kJ/mol (from gass phase H2O2, slightly more from the liquid); an in-between example is methanol, with a bond dissociation enthalpy of 474±4 kJ/mol. It looks like that table (and section) needs some work doing on it, not least in checking. Physchim62 (talk) 17:45, 27 December 2008 (UTC)
  • I suggest sticking to the data in the table in bond energy. In general compile this type of data in one article only and not different data in several articles. V8rik (talk) 21:27, 27 December 2008 (UTC)

Improvements for a GA?[edit]

I was just wondering whether anyone could help me with cleaning the article up a tad? I took out the long intro, which repeated a large amount of information in the article and replaced it with a brief (maybe too brief?) summary.

just another couple of points about this article:

  • Does anyone else think the history section looks too long?
  • There is a section for valence bond theory; why is there not one for MO theory? could the valence section be shortened?
  • Could the comparision of these two theories and their explanations be combined into one section on "Bond Theories" for example?
  • Should the theories be moved to after the explanation of bonds?

of course, all this could just be pointless rubbish, if anyone has any input, please say! it would be nice to promote this article eventually Thecurran91 (talk) 20:54, 21 September 2009 (UTC)

Yes, you shortened it too much. The original intro may have been too long, but that was because somebody had the idea of putting all the non-technical "overview" material which originally began the article, up into the LEAD. Some of this is actually NOT repeated anywhere in the article (why DO a bunch of random nuclei and electrons stick together? In small words.). And of course, LEAD material is supposed to repeat, at the same time it stands alone. As it it is, your lead is two paragraphs for a very long article, which is not enough. 4 paragraphs are recommended at minimum. So now we have an article which doesn't have a lead long enough to explain why a chemical bond forms, but you have to slog through the entire rest of the article to get that (and good luck).

So try again. Don't delete, improve. Don't delete material in a too-long lead, but put it back in an up-front overview, which is where it was before. SBHarris 22:32, 21 September 2009 (UTC)

Hmm. well I reverted my edit as all the stuff was commented. I still don't think the bits about covalent and ionic bonds are short enough for the intro, you agree? and you could also try to sound a bit nicer to newbies like me in future. Thecurran91 (talk) 10:05, 22 September 2009 (UTC)
Agree, we should not BITE each other, newbies or not. The question really comes down to "what do we want in this article?" Straight away, and reading from the top, I wouldn't classify a chemical bond as a "process", I would classify it as a situation. I don't think the reference to QED is useful either, as current theories of chemical bonding only use QED as a perturbation in the calculations, and then only in the most precise of calculations. If someone else with more experience in the area would like to correct me, I would be grateful. My main question is: "what do editors want in this article?" we could ask at WT:CHEM for more opinions. Physchim62 (talk) 13:41, 22 September 2009 (UTC)
Hi, thanks for the good advice! I was wondering, having taken SBHarris' advice, I have lengthened the intro in a sandbox a little more, giving a bit more detail. Maybe there should be some info about the theories in the lead but I don't have the knowledge to do this; Maybe SBHarris would care to lend a hand? Anyway, if you could give me some constructive criticism about my sandbox I would be very grateful! Thecurran91 (talk) 14:53, 22 September 2009 (UTC)
Sorry that sounded gruff. I should have looked to see how new you were. We all assume that somebody who looks like they know what they're doing insofar as the basic moves (like knowing how to pipelink to a policy link as you did below with BITE), has been around enough to have read the WP:MOS and so on. Anyway, welcome. ;)

Yes, the present LEAD is indeed too long, and I (for one) am not happy with it. I didn't make it that long. Again, that happened due to a summary section getting put up there by somebody else. I did help with the summary. We have to do something or other more general before diving into a long history, but the question is: what? I'll be happy to comment on your sandbox stuff. Of course I don't feel I WP:OWN this article (notice I didn't revert you but let you think and others comment); I just feel the need to warn against drastic change which delete text and leave an impoverished lead which isn't long enough. Some compromise is needed.

And (BTW) I also surely think QED is silly to mention here (another issue of somebody else), anymore than relativity in a car crash. Nobody uses QED to calculate atomic bonds-- it's a higher energy thing for far higher order corrections (Lamb shift, etc) which are totally lost in the noise of having to do with messy non-linear multi-body problems. Ab initio calculation of bond strengths are lucky to be acurate to one part in 1000, if I read the literature right. QED at atomic electron binding energies introduces corrections on the order of (what?) 10 parts in a million for Z=1,and maybe something more interesting for heavier elements. SBHarris 02:58, 23 September 2009 (UTC)

Time Taken[edit]

Can we put something about how long bonding takes? The latest research from Canada places bromine molecule breakup at around 200 femtoseconds. Does anybody have times for the converse process -- bonding itself? —Preceding unsigned comment added by (talk) 20:22, 16 September 2010 (UTC)

Huckels Rule - benzene example incorrect[edit]

Huckels rule regarding aromatic bond - gives benzene as an example with n=4, whereas it should be n=1, Huckels rule applies to the pi electrons only. Huckels rule article is correct. — Preceding unsigned comment added by (talk) 14:43, 5 December 2012 (UTC)

Where are electrons most?[edit]

The phrase "the electrons spend more time between nuclei, than anywhere else in space" is incorrect. They spend more time near the nuclei than anywhere else. The electron density between nuclei however increases on bond formation over the situation of two non-interacting atoms. I leave it to others to rewrite that section. --Bduke (Discussion) 21:56, 28 May 2013 (UTC)

I also don't like this statement. At the moment, I can't think of a good way of editing it. Maybe just get rid of it for now. It's not really necessary. I would be careful with the nuclei comment. While electron wavefunctions for atomic systems are typically high at the nucleus, their probability density is not. The probability density will be peaked at some distance away from the nuclei, depending on the number of electrons and the nuclear charge. For example, an electron in a hydrogen atom has its peak electron density at a value of r = a0. Sirsparksalot (talk) 17:36, 4 June 2013 (UTC)
The peak at r = a0 is saying that a0 is the most probable radius - i.e the probability of finding the electron in the volume element 4 pi r**2 delta r. This volume element increases as r increases. However if you look at the probability of finding the electron in equal volumes - (delta x) (delta y) (delta z) the probability is greatest at the nucleus. Several texts plot the electron density in molecules. The peaks are always at the nucleus. The density is the square of the wave function. You are confusing it with the radial density which is density times 4 pi r**2. --Bduke (Discussion) 21:41, 4 June 2013 (UTC)
Yep. Probabily volumetric density dP/dV = psi^2. So the straight differential probability dP is psi^2 dV = psi^2 (4piR^2dR). So RADIAL DENSITY dP/dR = Psi^2 (4piR^2). To see where that function is maximum, you must differentiate yet again WRT R, finding dP^2/dR^2, and set that to zero. The extra differentiation (psi^2 is here a simple exponential in terms of a, the Bohr radius) gives a function of two terms set to zero, and solving that gives an R for P(R)max and that's the Bohr radius.

Going to the original complaint above, since our wavefunctions in molecules are stationary states that are time-independent, it's better not to speak of where the electron spends "most time", as the wavefunction doesn't change with time. Best to speak of electron density in space, like some charged jelly that doesn't move in time but does change in density from place to place in SPACE. That motionless jelly has a radial density (as above) that is maximal on a plane somewhere between two atoms in the bonding molecular orbital (MO) of a chemical bond (in a bond between atoms of the same element in a diatomic molecule, of course it's exactly between them). For an antibonding MO, there's a node in that plane, so the electron is least likely there. The radial increase in density between atoms defines a bonding orbital, and the node an antibonding one.

If you work out where the electron density is (you have to sum it for more than one electron), then simple electrostatics gives the force with which the bond holds the atoms together (Feynman first realized this in an early paper in 1939). It sounds like a classical not quantum answer, but it's merely the classical outcome of a quantum calculation, as force is just the distance-derivative of potential, and the electron-density lines up along potentials, so you can integrate expectation-forces right along with potentials and electron densities. This calculation explains the electrostatic force you have to overcome to start to separate two bonded atoms, but doesn't entirely explain what force keeps them apart in opposition to their wholey electrostatic attraction. Still, the attractive part of a bond is what we are interested in.

It's true that even the radial probability is not much affected for inner electrons of atoms that participate in a bond (so these electrons have about equal probability of being between atoms as not, and do have almost their summed probability associated with just one atom). But we just ignore these electrons as inner electrons that effectively don't participate in the bond. It is the outer electrons with warped orbitals that brings their charge-density more into the space beteen atoms than the space that isn't between atoms, that are responsible for the attractive part of a covalent chemical bond.SBHarris 22:19, 4 June 2013 (UTC)

The increasing volume element as a function of r is precisely what I was referring to. Wouldn't "near" the nucleus imply the electron's proximity to the nucleus or, in other words, the radius? Similarly, what does it mean to be "between" the nuclei as opposed to near the nuclei? Sirsparksalot (talk) 19:52, 25 June 2013 (UTC)

Look at sigma bond. Draw two parallel planes each containing an atom, with the bond axis orthogonal to the planes. An electron is "between" two atoms if it is "between" the two planes, and not if not. The probability sum functions for this do have two peaks at each atom, but also there is a significant density exactly between the atoms, so the sum ends up looking a bit like the scaffolding of a two tower bridge (like the Brooklyn Bridge). The math depends on quantitation, but because electrons between atoms feel a higher total nuclear charge (on average), this region ends up with more electron density "in" it, than if the atoms were separated by a large distance. [2]. Effectively, if you take two H atoms with spherical 1s clouds, and move them closer to each other, each cloud is warped somewhat toward the space between the atoms, and this effectively moves some of each electron "off center", so each cloud is no longer spherical (although each one still has the highest volumetric density at its own nucleus). That overlap-- the amount of electron moved off each nuclear center due to the action of BOTH nuclei, is what makes up the bond. Each nucleus still "wins" in electron density, but those wins are "ties." The important thing is that the space between the nuclei now beats the space to either side of the two nuclei, whereas when the nuclei were separated at large distances, THIS was a tie. Put the two atoms close to each other, and it no longer is. SBHarris 03:16, 26 June 2013 (UTC)

Maybe my inquiry is a moot point since the text has been removed from the article. I understand where the electrons are, but I still have trouble with the semantics of the comparison of "between" nuclei as opposed to "near" the nuclei. An electron can be both between and near. If the between classification is used, then the argument should be that the integrated probability density is higher between the nuclei as opposed to outside of the nuclei (although I don't necessarily think "outside" is the best choice of terms). Also, I don't like the term "near" the nuclei because, in my opinion, it specifically implies the radial distance between a nucleus and an electron. In this case, due to the larger volume element of the radial distribution function, the electron is more likely to be located away from the nucleus as opposed to "near" the nucleus, despite the wavefunction being peaked there and regardless of the fact that a chemical bond is involved. Am I making sense? Sirsparksalot (talk) 15:38, 26 June 2013 (UTC)

Sentence case[edit]

Section headings, per WP manual of style, should be in sentence case, with only the first letter capitalized. Kortoso (talk) 22:51, 16 December 2013 (UTC)

Placing of aromaticity under resonant bonding[edit]

Recently, Leprof 7272 moved the section describing aromaticity to single and multiple bonds. I presume this is because of the bond order. However, all forms of resonant bonding have non-integer bond orders. If it's due to some other reason, please do state here because it doesn't seem to me like it should belong under the single or multiple bond sub-heading. Thanks.--Officer781 (talk) 06:33, 19 March 2015 (UTC)

I agree with you that Aromatic bonding should be Resonant bonding. In fact I think it should be the first subsection of Resonant bonding, as benzene is the classic case of resonance which is usually taught before hypervalence and 3c-2e bonding.
Also I think we have too many levels in the hierarchy, so I am going to promote Resonant to the same level as Ionic and Covalent. Since hypervalent and 3c-2e bonding are both partly ionic, we should not consider Resonant to be a subsection of Covalent. Dirac66 (talk) 20:22, 19 March 2015 (UTC)

Covalent bond list[edit]

I've transferred the covalent bond list over to the covalent bond article. This page is meant to be an introduction to bonding generally, and not meant to go in-depth. Stating here in case there's any issues.--Officer781 (talk) 15:16, 18 April 2015 (UTC)

Speaking of which, this article is too extensive for a general article in my opinion. Sections such as "Valence bond theory", "Comparison of valence bond and molecular orbital theory" and "Bonds in chemical formulas" could really do better under the covalent bond page. Should a content transfer be done?--Officer781 (talk) 15:25, 18 April 2015 (UTC)
My comments:
  • Transferring the covalent bond list is ok, but I would also add a briefer version of the Single and multiple bonds subsection to this article, in order to help beginners (say in high school) find this elementary topic.
  • The resonant bonding list should also be transferred to Covalent bond in my opinion. This is actually a more specialized subject than the types of covalent bonding.
  • In correct English, adjectives without nouns should not be used as (sub)section titles. So Aromatic should be either Aromatic bonding or Aromaticity. Similarly for Hypervalent and Electron-deficient. Dirac66 (talk) 18:58, 18 April 2015 (UTC)
All three points done, although I simply transferred back the "Single and multiple bonds" subsection. As for the three sections I've mentioned, would it be good to transfer or to retain them here?--Officer781 (talk) 07:55, 19 April 2015 (UTC)
I suggest transferring the first 2 of the 3. The VB and VB-MO sections are about using quantum mechanics to describe bonds and will not be accessible to beginners, so they should be in the more specialized article Covalent bond, preferably near the end so that those who have no QM can read the beginning of the article. However Bonds in chemical formulas is a much simpler section with no QM which could help readers trying to understand an organic structure, so I would keep it here. Dirac66 (talk) 12:52, 19 April 2015 (UTC)
Done. Thanks!--Officer781 (talk) 14:46, 19 April 2015 (UTC)