14
$\begingroup$

There is no question that the study of chemistry (the study of interatomic interactions and their transformations) predates the quantum mechanical description of matter. This has given rise to descriptions of bonding including the famous theory of Lewis.

From this vantage point, it seems that many "exotic" chemical motifs (e.g. noble gas molecules) have to be explained as perturbations away from the traditional rules of bonding, e.g. octet rules, hybridization, hyper valency, and d-block participation.

I believe that many of the questions seek answers that would get correct answers scored on exams in lower-division chemistry, and that would be by simply executing these empirical rules until the desired answer is found.

However, it is my opinion that all chemical scenarios can be described from concepts drawn from electronic structure theory (i.e. theory that could be used for quantitative quantum chemical computation.)

In this light, it is very difficult to have to start answers by saying that the rules that are propagated from generation to generation might be bogus, and are simply summaries of general observations by creating "rules" that may or may not be supported by solution of a Schroedinger equation.

It seems that StackExchange should be correct. I suppose I am venting a little bit, but my question would be: given that there are dual explanations (one traditional, the other quantum mechanical), should we make a concerted effort to give answers and votes to the most correct answers, or the answers that best fit the expectation of the person who is asking the question?

| |
$\endgroup$
  • $\begingroup$ This might be somewhat related: meta.chemistry.stackexchange.com/questions/101/… $\endgroup$ – Martin - マーチン Jun 1 '14 at 13:59
  • 1
    $\begingroup$ Eric, thanks for taking the initiative to bring this to everyone's attention. I'd love to see more voting on great content in general (though we're improving on that day by day), but we definitely appreciate answers that are as accurate as possible. By the nature of SE, the responsibility for marking answers as "correct" will likely always lie on the OP, but the votes will ultimately flesh out the highest quality replies. $\endgroup$ – jonsca Jun 1 '14 at 23:20
12
$\begingroup$

In general, I try to remember two things:

  1. All theories/models are imperfect
  2. The best model is the one that adequately describes a phenomenon in the simplest terms

With that in mind:

From this vantage point, it seems that many "exotic" chemical motifs (e.g. noble gas molecules) have to be explained as perturbations away from the traditional rules of bonding, e.g. octet rules, hybridization, hyper valency, and d-block participation.

I don't think of them as perturbations - they are limitations of the simpler models.

I believe that many of the questions seek answers that would get correct answers scored on exams in lower-division chemistry, and that would be by simply executing these empirical rules until the desired answer is found.

What you are describing is using simple models for situations where they adequately describe observations.

However, it is my opinion that all chemical scenarios can be described from concepts drawn from electronic structure theory (i.e. theory that could be used for quantitative quantum chemical computation.)

Sure they can. But why stop there? Why not break out the Feynman diagrams and standard model? It's because we stop at the simplest model that adequately describes phenomena. Remember that even solutions to Schrodinger's equation are approximations for all but the simplest systems - we have to choose the level of approximation that we are comfortable with.

In this light, it is very difficult to have to start answers by saying that the rules that are propagated from generation to generation might be bogus, and are simply summaries of general observations by creating "rules" that may or may not be supported by solution of a Schroedinger equation.

You don't have to do anything. You can give as much or as little detail as you want. However, to be useful, I have found that good answers to scientific questions keep the above two principles in mind. So, for example, if you are answering a question about the stability of carboxylate in solution, resonance structures and hybridization theory should work just fine. On the other hand, if you want to explain why chlorate ions can form even though the chlorine atom would have an expanded octet, then it might make sense to explain that the Lewis model and valence bonding theory is just an empirically (and pedagogically) useful approximation of what is really happening, which is ... (insert quantum mechanical explanation here)

It seems that StackExchange should be correct. I suppose I am venting a little bit, but my question would be: given that there are dual explanations (one traditional, the other quantum mechanical), should we make a concerted effort to give answers and votes to the most correct answers, or the answers that best fit the expectation of the person who is asking the question?

The whole model of stack exchange is one of populism. The idea is that people, given free reign, will choose the answers that are best for them, and that will eventually result in the "best" and "most correct" answers. It is true that the usefulness of the site in general will depend on the correctness of the answers on it. It is also true that scientific truth doesn't depend at all on popular opinion. However, on stack exchange, the two collide. We would like for that collision to result in answers that are both correct and useful, and so I think that people should:

  1. Give the most useful and correct answer that they can when answering questions
  2. Accept the most useful and correct answer when asking questions
  3. Vote for the most useful and correct answer when voting for questions.

I think everyone can (and should) use their own judgement on what counts as "most useful," but for myself, I look for:

  1. Answers that are as simple as possible - but no simpler!
  2. Answers that are correct within the realm of the model being used
  3. Answers that cite references

Correctness has to be hashed out in the comments, competing answers, and discussions - and as a reader, I often have found the discussions to be at least as valuable as the answers themselves.

| |
$\endgroup$
  • $\begingroup$ I agree with the gist of the above answers - that there are varying shades of correctness. And above all, scientific truth is independent of what we might think, or draw on a piece of paper, or write in a 200 page dissertation. On the other hand, we can get a lot of mileage from basic models, at least at the undergrad level. It is nice that many people here are obviously way beyond me in understanding chemistry and I appreciate the additional insight these people have to offer, and thanks to them, I am aware of many of the limitations of models. $\endgroup$ – Dissenter Jun 4 '14 at 20:35
  • $\begingroup$ Nice response. I think the concept of "model" is a very important one to bring up, so much so that this Q&A could very well be part of the main site: "Why do historical bonding theories persist in the chemical sciences?" If the purpose of answering questions is to allow the "asker" to develop his/her scientific knowledge, then answering with an historical model (including the limitations of that model, when necessary) is a good way to meet that goal. $\endgroup$ – bobthechemist Jun 12 '14 at 12:27
  • 2
    $\begingroup$ I wish I could recommend this answer 100 times. If every answer to a mechanics question started with "Newtonian mechanics is (incorrect/inadequate/obsolete)" we'd never get anywhere. $\endgroup$ – Silvio Levy Jul 25 '14 at 18:36
  • $\begingroup$ Also - I come from a math background and find a lot of explanations involving MOs and stuff very unrigorous; sometimes statements that are outright false are widely believed because they're given in textbooks (perhaps with caveats that are easy to forget). My question at chemistry.stackexchange.com/questions/14578/… deals with one such example. $\endgroup$ – Silvio Levy Jul 25 '14 at 18:40
10
$\begingroup$

First of all, there is nothing stopping votes from being given to both the empirical hand-waving explanations and the deeper theoretically calculated ones with their accompanying framework. Of course, which one gets picked as correct is up to the asker, though one might expect they'll have a tendency to go for the easier one, given the site's current audience, which is an unhelpful disincentive towards more elaborate responses.

I've contributed to a few chemical bonding questions with the textbook answers, not because I feel it's the most adequate in the situation or some other deeper reason, but simply because it's the only one I know. If I did know better, I would happily delve into detail, using the classical explanation as a stepping stone into a more precise investigation. I would love to know more, but due to several factors in my academic formation, I have regrettably not yet acquired much depth in the nature of chemical bonding. Even though I am fascinated at the possibilities of computational chemistry, it's a world I have limited access to. I suspect this isn't the case just for me.

Therefore, while I don't see any reason to recommend against explaining phenomena with an undergraduate-level understanding of bond theory, I would very kindly ask that anyone capable of contributing with information at a higher level do so whenever possible. For example, it's been a while since I'd seen you, Eric, but I still remember how interesting your answers were several months ago when you attacked multiple questions with AIM calculations and arguments against the traditional picture of hybridization and hypervalency. So please don't lose heart, contributions like yours are very useful, even if possibly underappreciated by most.

| |
$\endgroup$
  • $\begingroup$ I'd like to second that. For chemical bonding in general I usually try to provide a more canonical approach. The empirically found observations are usually in some way approximations to the truth and should therefore be treated as these and then there are concepts like hypervalency, which resembles an incomplete (ancient) point of view - one should always state the flaws of these concepts. $\endgroup$ – Martin - マーチン Jun 1 '14 at 13:52
  • $\begingroup$ Good questions and comments. I'm afraid a lot of handwaving does go on at the undergrad level when describing bonding. For example, hybridization has been described as "a meat grinder; throw in 3/4th of a pound of ham and 1/4th of a pound of pork and what do you get? An sp3 hybridized orbital with 75% ham character and 25% pork character." I'm not kidding, I've heard this meat analogy several times. $\endgroup$ – Dissenter Jun 2 '14 at 19:16
  • $\begingroup$ The best I think most undergrad students of chemistry can do is understand: 1) Qualitative aspects of valence bond theory 2) Hybridization and its limitations 3) Molecular orbital theory (qualitative aspects only) $\endgroup$ – Dissenter Jun 2 '14 at 19:16
4
$\begingroup$

This is a question I have pondered too. Being a primarily synthetic chemist and one for whom a lot of physical chemistry was never properly placed in a context that made it relevant, I naturally err to the lower rigour, more 'hand-wavy' end of the explanation spectrum, because, as others have pointed out of themselves, I am simply unqualified most of the time to provide much more rigorous explanations.

I think the 'best' format of answer, would be one that is understandable on multiple levels of complexity or abstraction, and doesn't require a firm grasp of the underlying mathematics to follow, but provides them for digestion. This makes the answers more useful to a broader range of readers beyond the level the initial question is pitched at.

This idealised answer I think should touch upon three sections, reflecting the three main levels of comprehension as I see them:

1) Metaphor(s) tying the behaviour to a very well understood or basic principle, for example, in one answer I pointed out that a way of grasping that 2s is lower energy than 2p because of penetration might be to consider Coulombic attraction of point charges. The relevant equation is simple if unfamiliar and crops up in most secondary/high school curricula, and provides an initial handle for the reader to understand the behaviour. Think about trying to explain a phenomenon to a parent or non STEM friend, you aren't aiming for detail, you are aiming for broad strokes, which in this case can be built upon in 2).

Obviously this isn't an explanation and certainly not a rigorous one, but it is supposed to be an 'easy way in'. I gave the particular answer I mention after a number of very technically proficient and rigorous answers that I found very hard to follow even as someone completely familiar and comfortable with the concept of radial distribution functions and most of the relevant QM, where I felt the question was more about being unable to wrap their head around the behaviour intuitively, rather than that not believing the maths.

2) The classical model. Variously hand-wavy, this is the next level of rigour, giving for example details of relevant orbital characteristics in qualitative terms. Most technical diagrams or pictures (particularly orbital diagrams or mechanisms) would come under this section. Here the goal is to explain the answer in a functional and chemical framework which allows common application of the knowledge.

This section is aimed at giving a functional understanding of the behaviour of something, ideally in terms of generally applicable principles. e.g. Orbital hybridisation, mechanism classes. This level of explanation might not be particularly precise, or always a good representation of exact behaviour, but most of the time it serves you well. HOMO/LUMO/SOMO type explanations I would probably consider part of this section as well.

3) The QM perspective. Here is where all the physically minded respondents shine, stepping through logical and formula-based explanations of phenomena, resulting in as quantitative a picture as can be gathered of a given bonding behaviour. This is for the reader who really wants to get to the bottom of what is understood about the phenomena in question.

The one point I will note personally about this section is that often notation choices are made for efficiency or elegance from the perspective of someone knowledgable in the art, not for someone less fluent or only just dipping their feet in. Sometimes it can be easier to follow in plain, bog-standard algebraic notation (and/or with accompanying text detailing the parts of the equation) rather than fancy short-hands with lots of field-specific notation (Bra-Ket and group theory notation spring to mind as obvious barriers to comprehension for the otherwise mathematically competent).

To achieve this total may require multiple distinct answers, but ideally each would at least be considered.

| |
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .