Wikipedia talk:WikiProject Elements/Archive 27

Archive 20

←

Archive 25

Archive 26

Archive 27

Archive 28

Archive 29

Archive 30

Reclassifying the nonmetals

This is not a formal proposal from me. I'm only requesting views from you on the alternative classification scheme set out below. There is no rush as I have some more research to do. It's therefore a slow time project. As well, R8R is engaged on getting lead to FA status; and YBG is on a wikibreak until mid-April, so I'm happy to let this sit on the bench top for quite a while.

Background

We currently colour code non-metals as polyatomic, diatomic, or noble gas.

Ever since we adopted these categories I've wondered (partly prompted by R8R) if that was the right decision, and if there was a better alternative.

This has been hard. Categorisation of nonmetals in the literature—aside from the halogens and the noble gases—is shabby. Metalloids complicate the situation. Some authors recognise such a category; others don't. The one that don't have to divvy them up between the metals or the nonmetals.

Authors usually throw up their hands and simply look at the leftover nonmetals—the ones other than the halogens and the noble gases—on a group-by-group basis. So you might have separate sections in a book on, say, hydrogen; carbon; nitrogen and phosphorus; and oxygen, sulfur and selenium. And some or all of the metalloids might get added to the applicable sections.

Otherwise there is the question of what to call these leftover nonmetals. The category names I've seen in the literature are "biogen", "CHONPS", "organogen" or "other". The first three of these categories tend to get tripped up by what to do with selenium. The last category—other nonmetals—is the "I-give-up-it's-too hard-I-need-to-get-published-so-I'll-treat-them-as-leftovers" category.

For our own classification scheme the "halogen" category is unavailable since we count astatine as a metalloid (as we should—it's either that or a post-transition metal).

Alternative scheme, Mk. 1

In this context, an alternative scheme I've been considering has the following nonmetal categories:

Noble gas – He, Ne, Ar, Kr, Xe, Rn
Corrosive – O, F, Cl, Br, I
Intermediate – H, C, N, P, S, Se
Weak nonmetal (metalloid) – B, Si, Ge, As, Sb, Te, At

The corresponding legend looks like this (Mk. 1):

Alkali metal

Alkaline earth metal

Lan­thanide

Actinide

Transition metal

Post-​transition metal

Weak nonmetal (metalloid)

Intermediate nonmetal

Corrosive nonmetal

Noble gas

There is no change to the colours we use; the only change is to the three category names between post-transition metal and noble gas.

Corrosive nonmetal. These are are corrosive, and highly electronegative (> 2.6) and are, or their species are, capable of acting as relatively strong oxidising agents. Here are some examples from the literature as to other similarities between oxygen and fluorine, oxygen and the halogens, and oxygen and chlorine:

• "Fluorine tends to bring out the highest valence of the element with which it combines. In this its shows a strong resemblance to oxygen. In combination with metals, oxygen appears to be the best for the highest valences, e.g. OsO₄ and KMnO₄, but fluorine appears best if the highest valence is relatively low, e.g. for CoF₃, CuF₃, AgF₂, TbF₄, and BrF₅. With non-metals the difference between oxygen and fluorine is less apparent." (Phillips & Williams 1965, p. 446)
• "…oxygen, like fluorine, forms strong covalent bonds, and there are a number of similarities between covalent oxides and fluorides." (Emeléus & Sharpe 1973, p. 318)
• "Simple anionic chemistry is limited to oxygen and the halogens, although polyanions and polycations can be formed by many [nonmetals]." (Cox 2004, p. 145)
• "Chlorination reactions have many similarities to oxidation reactions. They tend not to be limited to thermodynamic equilibrium and often go to complete chlorination. The reactions are often highly exothermic. Chlorine, like oxygen, forms flammable mixtures with organic compounds." (Kent 2010, p. 104)

Intermediate nonmetal. The more temperate nature of the intermediate nonmetals is relatively self-evident, situated as they are between the corrosive nonmetals and the weak nonmetals. This is so-called "other nonmetal" territory. There is a magic thread (my name for it) that binds the intermediate nonmetals, and it goes like this: H → C → P → N → S → Se:

• Chemical similarities between H and C were discussed by Cronyn (2003) in the Journal of Chemical Education. They include proximity in ionization energies, electron affinities and electronegativity values; half-filled valence shells; and correlations between the chemistry of H–H and C–H bonds.
• C and P represent an example of a less-well known diagonal relationship, especially in organic chemistry. Spectacular evidence of this relationship was provided in 1987 with the synthesis of a ferrocene-like molecule in which six of the C atoms were replaced by P atoms (Rayner-Canham 2011, p. 126). Further illustrating the theme is the extraordinary similarity between low coordinate P compounds and unsaturated C compounds, and related research into organophosphorus chemistry (Dillon, Mathey & Nixon 1998).
• P and N are in the same group. Although N and its oxides are gases whereas P and its oxides are solids, the two elements "show many similarities in their compounds" (Malati 1999, p. 83). Despite these similarities "the chemistries of nitrogen and phosphorus are very different" (Wiberg 2001, p. 686). However P and N form an extensive series of phosphorus-nitrogen compounds having chain, ring and cage structures; and the P-N repeat unit in these structures bears a strong resemblance to the S-N repeat unit found in the wide range of sulfur-nitrogen compounds (Roy et al. 1994, p. 345) discussed next.
• N and S have a less-well known diagonal relationship, manifested in like charge densities and electronegativities (the latter are identical if only the p electrons are counted; see Hinze and Jaffe 1962) especially when S is bonded to an electron-withdrawing group. They are able to form an extensive series of seemingly interchangeable sulfur nitrides, the most famous of which, polymeric sulfur nitride, is metallic, and a superconductor below 0.26 K. The aromatic nature of the S₃N₂²⁺ ion, in particular, serves as an exemplar of the similarity of electronic energies between the two nonmetals (Rayner-Canham 2011, p. 126).
• S and Se are in the same group: "As in the case of the halogens, the chemical similarities, at least for sulfur and selenium, are abundantly obvious" (Scerri 2007, p. 49).

Weak nonmetal (metalloid). Some authors count metalloids as nonmetals with weakly nonmetallic properties rather than having a discrete metalloid category, and that's the approach I've taken here. For a recent example see Cox (2004, pp. 26–27), who treats B, Si, Ge, As, Sb, Te and At as nonmetals, but notes that Si, Ge, As, Sb, Se and Te are sometimes called metalloids.

Precedents, approach, and features

In terms of precedents, the literature certainly refers to (a) O, F, Cl, Br, and I as having corrosive qualities; and (b) to the weakly nonmetallic chemistry of the metalloids. As far as I can see nobody has ever referred to an intermediate nonmetals category but then the literature is a terminological wasteland when it comes to a collective name for this part of the periodic table. Just about anything would be better than "other nonmetal".

I had to apply some violence and abstraction of detail to the alternative scheme in order to keep it simple. So there may arguably be some discontinuities and boundary overlaps. For example, counting iodine in the same league as O, F, Cl and Br may raise an eyebrow. Then again, iodine is corrosive, has a pretty decent electronegativity (2.66), and its periodate ion is a formidable oxidising agent (stronger than the perchlorate ion, for example); even the iodate ion is a stronger oxidant than elemental bromine. And pragmatically speaking it makes more sense to keep iodine with its lighter halogen congeners. As another example, nitrogen has a high electronegativity of 3.04 but all of its chemistry is essentially covalent, and the average oxidising power of nitrogen and its species, in aqueous solution, is less than that of both iodine and of sulfur.

On the question of discontinuities and boundary overlaps I turn to Jones (2010, pp. 170–171): "Though classification is an essential feature of all branches of science, there are always hard cases at the boundaries. The boundary of a class is rarely sharp…Scientists should not lose sleep over the hard cases. As long as a classification system is beneficial to economy of description, to structuring knowledge and to our understanding, and hard cases constitute a small minority, then keep it. If the system becomes less than useful, then scrap it and replace it with a system based on different shared characteristics." I feel that the similarities within each of the categories of weak nonmetal (metalloid), intermediate nonmetal, and corrosive nonmetal outweigh their differences, and blurry edges, sufficiently to establish them as discrete divisions.

The alternative scheme is better than what we have now because it's easier to get your head around, the new category names are more natural, and the new categories themselves fall into place quite naturally.

It maintains a good job of showing the progression in metallic to nonmetallic character as you go from left to right across the periodic table. And it facilitates a symmetry that is more rewarding than the contrast between metals and nonmetals, or between the alkali metals and the group 17 nonmetals, as shown in the following side-by-side match-up:

"Reactive metals" ^ Groups 1–3, Ln, An			Corrosive nonmetals O, F, Cl, Br, I
Transition metals (the "mundane" ones) Most of 'em			Intermediate nonmetals H, C, N, P, S, Se
Post-transition metals Ga, Bi etc			Weak nonmetals (metalloids) B, Si, Ge, As, Sb, Te, At
Noble metals ^ Ru, Rh, Pd, Ag, Os, Ir, Pt, Au			Noble gases He, Ne, Ar, Kr, Xe, Rn

^  I am not proposing that we have colour categories for reactive metals, and noble metals. The current colour categories for metals (alkali, alkaline earth, transition, lanthanide, actinide, post-transition) are fine.

There are references in the literature to this pattern. For example:

• "Between Groups I and VII there are gradations from active metals (Col. I) to less active metals to moderately active nonmetals to volatile nonmetals (halogens Col. VII)." (Perlman 1970, p. 439)
• "A period represents a stepwise change from elements strongly metallic to weakly metallic to weakly nonmetallic to strongly nonmetallic, and then, at the end, to an abrupt cessation of almost all chemical properties." (Booth & Bloom 1972, p. 426)
• "By the end of 8th grade, students should know that…there are groups of elements that have similar properties, including highly reactive metals, less reactive metals, highly reactive nonmetals (such as chlorine, fluorine and oxygen) and some almost completely unreactive gases (such as helium and neon)." (AAAS 1994, p. 78)
• Between the "virulent and violent" metals on the left of the periodic table, and the "calm and contented" metals to the right are the transition metals, which form "a transitional bridge between the two" extremes. (Atkins 2001, pp. 24–25)
• "Describe how groups of elements can be classified…including…highly reactive nonmetals, less reactive nonmetals, and some almost completely nonreactive gases." (Padilla, Cyr & Miaoulis 2005, p. 27)

Overall, I find the combination of:

• symmetry (as in the four complimentary metal-nonmetal categories)^† and asymmetry (many metals/few nonmetals);
• the natural fit of the intermediate nonmetals between the weak nonmetals and the corrosive nonmetals;
• the thread that links the nonmetals in this category (i.e. the intermediate nonmetals); and
• the balanced 5-6-6-6 distribution of the nonmetals across their four categories

to be especially pleasing.

† The fact that there are 4 + 4 = 8 symmetry components is cool, too.

The wisdom of YBG

YBG suggested that any new categorisation scheme should be:

• clear—"The criterion for division should be easily explained";
• unambiguous—"It should be relatively obvious which category each element fits into"; and
• meaningful—"The categories should have significance more than just dividing for the sake of dividing. There should be enough within-group similarity and enough between-group dissimilarity so that each group could be the subject of a separate encyclopaedia article."

Clear. In this case, the criteria for division go something like the following:

• The noble gas category is self-evident (so to speak).
• The corrosive nonmetals are, well, corrosive.
• The weak metal (metalloid) category corresponds to the elements commonly recognised as metalloids. Astatine, which is included here, is irregularly recognised as a metalloid but we decided quite a while go to recognise it a metalloid. It has since been predicted to have the band structure of a full-blown metal.
• The remaining nonmetals, which are neither corrosive nor weak (metalloid), are the intermediate ones.

This is so easy it almost explains itself.

Unambiguous. It is relatively obvious which category each element fits into.

Meaningful. As discussed earlier in this section, the resulting categories have enough internal similarity, and between category dissimilarity, to make them meaningful. We already have separate articles for noble gases and metalloids; I'd be inclined to have separate sections for intermediate nonmetals and corrosive nonmetals in the nonmetal article, as we do now for the polyatomic nonmetals and diatomic nonmetals.

References

• AAAS (American Association for the Advancement of Science) 1994, Benchmarks for science literacy, Oxford University Press, New York
• Atkins PA 2001, The periodic kingdom: A journey into the land of the chemical elements, Phoenix, London
• Booth VH & Bloom ML 1972, Physical science: a study of matter and energy, Macmillan, New York
• Cox PA 2004, Inorganic chemistry, 2nd ed., Bios Scientific Publishers, London
• Cronyn MW 2003, "The proper place for hydrogen in the periodic table", Journal of Chemical Education, vol. 80, no. 8, pp. 947–950, doi:10.1021/ed080p947
• Dillon KB, Mathey F & Nixon JF 1998, Phosphorus: The carbon copy: From organophosphorus to phospha-organic chemistry, John Wiley & Sons, Chichester
• Emeléus HJ & Sharpe AG 1973, Modern aspects of inorganic chemistry, 4th ed., Routledge & Kegan Paul, London
• Jones BW 2010, Pluto: Sentinel of the outer Solar System, Cambridge University Press, Cambridge
• Kent JA 2007, Kent and Riegel's Handbook of industrial chemistry and biotechnology, 11th ed., vol. 1, Spring Science + Business Media, New York
• Malati MA 1999, Experimental inorganic/physical chemistry: An investigative, integrated approach to practical project work, Woodhead Publishing, Oxford
• Padilla MJ, Cyr M, Miaoulis I 2005, Science explorer (Indiana Grade 6), teachers's edition, Prentice Hall, Upper Saddle River, New Jersey
• Perlman JS 1970, The atom and the universe, Wadsworth Publishing, Belmont, California
• Phillips CSG & Williams RJP 1965, Inorganic chemistry, vol. 1, Principles and non-metals, Clarendon Press, Oxford
• Rayner-Canham G 2011, "Isodiagonality", Foundations of Chemistry, vol. 13, pp. 121–129, doi:10.1007/s10698-011-9108-y
• Roy AK, Burns GT, Grigora S & Lie GC 1994, "Poly(alkyl/aryloxothiazenes), [N=S(O)R]_n : New direction in inorganic polymers", in P Wisian-Neilson, HR Alcock & KJ Wynne KJ (eds), Inorganic and organometallic polymers II: advanced materials and intermediates, American Chemical Society, Washington DC, pp. 344–357, doi:10.1021/bk-1994-0572.ch026
• Scerri E 2007, The Periodic Table: Its story and its significance, Oxford University Press, Oxford
• Wiberg N 2001, Inorganic chemistry, Academic Press, Berlin
• Wisian-Neilson P, Alcock HR, Wynne KJ (eds) 1994, Inorganic and organometallic polymers II: advanced materials and intermediates, American Chemical Society, Washington DC

User:Sandbh 00:58, 12 March 2017 (sign added late, -DePiep (talk) 12:58, 13 August 2017 (UTC))

Comments Mk. 1

Some first responses. Added bullets to allow for subthreading. -DePiep (talk) 12:07, 12 March 2017 (UTC)

• Interwiki RfC? Is it possible to enlarge support for our enwiki categorisation by making this an interwiki RfC? Our 2013 changes did not pick up well even in large wikis like de:, zh:, ru:, fr: (halfway only); ja: did. Note: checking the At category is enough, I can't blame any wiki for not using the poly-/di-atomic category names. -DePiep (talk) 12:07, 12 March 2017 (UTC)
  • And guess who made it stick for ja:! ^_^
  • BTW, there is some regional variation as well: in Japanese アルカリ土類金属 is a direct calque of "alkaline earth metal", but it only refers to Ca, Sr, Ba, and Ra (not Be or Mg). Double sharp (talk) 13:20, 12 March 2017 (UTC)

I'd like to contain the discussion to our own project, for now. Kudos to Double sharp san. Sandbh (talk) 23:42, 12 March 2017 (UTC)

Changing name from "metalloids" into "weak nonmetals"

• This is a surprise, and I'm not yet convinced. I didn't know there is a problem with 'metalloids'. First, the new name introduces a relative term in "weak": weak compared to others I must understand. And these others are: nonmetals (a de-classifying name by itself). So I must understand that there are elements that are stronger nonmetals, that are opposite of metals altogether. Then, seeing their position next to metals in the periodic table, if they are 'weak nonmetals', are they almost 'strong metals'? Or almost 'weak metals'? This is not playing, this is what words do. -DePiep (talk) 12:07, 12 March 2017 (UTC)

This is truly excellent and thoughtful feedback, thank you DiPiep.

To clarify, I am proposing to change the category name from "metalloid" into "weak nonmetal (metalloid)". I will discuss a problem with the term "metalloid" a little later in my response but for now it has its uses, which is partly why I am proposing to retain this word in the new category name. Yes, I agree, if you see the expression "weak nonmetal (metalloid)" it suggests there are stronger nonmetals, which is indeed the case.

If you see that they are positioned next to metals in the periodic table, and this prompts you to wonder if they are almost weak metals or almost strong metals then this is a good thing. The name piques curiosity, rather than bewilderment. I did not know what a metalloid was the first time I saw that word but I sure knew what metals and nonmetals were. The proposed category name is more informative than our other category names, which rely on some level of familiarity to be able to work out what they mean.

The scenario of wondering if a weak nonmetal would almost be a strong metal seems unlikely. Things in general get named for their dominant character rather than any subsidiary character, don't they? Some 'strong' transition metals are capable of forming oxyanions, which is nonmetal-like behaviour, but nobody calls these elements weak nonmetals. Crudely put, if an element was one-third nonmetallic (i.e. weakly non metallic) and two-thirds metallic (i.e. almost strongly metallic) it would be classified as a metal rather than a nonmetal. An analogy can be drawn with water that is weakly acidic (pH 6). This does not mean the water is almost strongly alkaline (pH 12). It means the water is almost weakly alkaline (pH 8).

I am sorry to say that one cannot work this out from the name "post-transition metal". If we had used something more descriptive like "poor metal" that would likely have answered the general reader's question. They would see that weak nonmetal was next to poor metal. But we decided to use "post transition metal", which is fine for the more technical reader. Alas, I feel that I should only work with the metal category names that we have. Sandbh (talk) 04:32, 20 March 2017 (UTC)

Descriptive terms

You raise very good points here, as always.

Yet I have a bit of an objection to this. Is it not the case that all of these descriptive terms need some knowledge of chemistry to understand? You can't understand "alkali metal" or "halogen" without knowing what an alkali or a salt is (in the latter case you also need to have a little knowledge of Ancient Greek, namely ἅλς "salt, sea"). You need some of this knowledge to get anywhere, and often you only get it retrospectively. If the name of the category does not suffice to tell us how Au or Te behaves, then maybe we should look more at the chemistry of their groups as a whole. Double sharp (talk) 04:16, 21 March 2017 (UTC)

I may have misunderstood the basis for your concern but will plough on anyway.

Largely I would say yes you do need some knowledge of chemistry to know what these terms mean. Alkali metals and alkaline earth metals may be somewhat of an exception. I can recall a high school science demonstration of calcium skipping about on the surface of water so that may have given me some idea of the nature of an alkali or alkaline earth metal. I can't remember when I learnt what "alkaline" meant but it may have been when I was introduced to the idea of acids and bases and phenolphthalein—and that was only in general science. And I'm pretty sure I knew by them what quicklime was and its causticity, although I can't remember if I would have recognised this as being due to its alkaline nature. I can't remember if I saw what happened when a few drops of iodine were added to powdered aluminium. Nor can I remember when I was taught about the contrast between sodium and chlorine. And I don't know if high school chemistry these days includes real experiments designed to show the properties of such reactive metals and nonmetals.

There are some category names that I would recognise without needing a background in chemistry. Base metals, precious metals or coinage metals. Maybe even refractory metals. Ferromagnetic metals. Gaseous nonmetals. Crystallogens perhaps.

It's good thing to have simpler descriptive category names, is it not? Sandbh (talk) 10:12, 21 March 2017 (UTC)

Prerequisites

It is a good thing, but there are always some prerequisites. And we shouldn't forget that some of the names are misleading, even though they are intuitively obvious. Pb may be a crystallogen, but it does not even have a diamond cubic allotrope, and while Sn does have one it's not really that attractive to look at. Some of these have artificial breaks: Re is a precious metal (of sorts) but Tc is not, because no one wants to have radioactive things (never mind that Tc has a long half-life and decays to clean stability, eliminating most of the problems with things like Pu and Cm which have comparable half-lives).

Well, yes, there are always some prerequisites; equally, category names are labels of convenience not labels of absolute truth, are they not?

There are several other category names that fall short under pressure. The alkaline earth metals category is a pretty good label for most of the group II metals but wobbles a bit in accommodating beryllium and its amphoteric oxide, and its capacity to form beryllates. (What is an amphoteric element doing hiding among a bunch of supposedly alkaline metals? Hoping no one will notice? ^_^). Pnictogens is an IUAPC approved name for group 15 but has zero intrinsic relevance to anything after nitrogen. Much of the chemistry of the early actinides is quite different from that of actinium. The crystallogen example is interesting. C, Si and Ge form diamond cubic allotropes at room conditions; tin only does so at low temperatures and the result is an amorphous grey power unless special care is taken to prepare it in its crystalline form (which looks metallic). A diamond cubic form of lead has not so far been prepared but there has been speculation that this might be possible under the right conditions.

Labels of convenience, as such, don't necessarily care or need to worry about artificial breaks. Sandbh (talk) 04:47, 22 March 2017 (UTC)

Polyatomic and diatomic

And it seems to me that "polyatomic" and "diatomic" have fewer prerequisites, as they do not require us to get an idea about what could be intermediate about nonmetals (which means learning the chemistries of many of these elements, which are fantastically complicated, so we don't tend to go past the vital C and H in high school): they just require us to learn a few simple terms. Later we can marvel about how these two are correlated with other important properties as well. Double sharp (talk) 14:37, 21 March 2017 (UTC)

Maybe "polyatomic" and "diatomic" require as much pre-knowledge as intermediate nonmetals i.e. you need to at least know what polyatomic means, and for intermediate you need to know they are neither as extreme as corrosive nonmetals (like Cl) nor as weak as weak nonmetals (like Si). If you happen to have a wikipedia table in front of you with the proposed new colour categories you will at least be able to appreciate straight away the relevance of the intermediate label (just like the transition metals are transitional). I don't think there is much between the two schemes in this sense, nor is there is much need to, at first, go past C and H in either of them. Sandbh (talk) 04:47, 22 March 2017 (UTC)

Grouping criteria

re you do need some knowledge of chemistry to know what these terms mean. Alkali metals .... Is true, but is not the primary issue in categorising. Category name and meaning is secondary, and that's what links are for. Primary is: group together what belongs together, and do not include stuff that does not include there. Just give the group a unique color and name, that's enough. The name can be as fancy or invented, no intrinsic meaning required, no prior knowledge for the reader required. Exactly what the grouping criteria is, should be described in the namelink or in an overview article. OTOH, a name should not be misleading by ignoring existing prior knowledge (is why we cannot reuse and redefine 'rare earth metals' for lanthanides). In this case, I think that 'metalliods' (and 'de:halbmetalle') are fine, unloaded words. Once we would use '.. nonmetal' somehow, it can not be made to mean 'but not really a nonmetal' any more. -DePiep (talk) 08:57, 22 March 2017 (UTC)

Hmm, I don't like the term 'halbmetalle' or its English relation 'semi-metal' given confusion with the physics-based terms half-metal, and 'semimetal'. In the latter case only carbon, arsenic, antimony and bismuth are semi-metals yet few would regard carbon and bismuth as metalloids.

I no longer like the term 'metalloid' (but I acknowledge it is still used in the literature). It means 'resembling a metal' but since metalloids generally behave at least chemically like nonmetals, the term 'metalloid' tells less than half the story. Even physically, metalloids mostly act like semiconductors and are brittle, which are properties associated with nonmetals, so although they have the appearance of faux-metals, the 'metalloid' nomenclature is rather misleading.

I prefer the term 'weak nonmetal (metalloid)' as it captures the best of both worlds, without misleading anyone, and it does not involve a change of meaning. Sandbh (talk) 12:28, 22 March 2017 (UTC)

Not-misguiding words

I used 'halbmetalle' only as an illustration of a not-misguiding word, as is 'metalloid' (or so I thought). If there is no descriptive available spot on, and if 'halfmetal' or 'metalloid' etc. are mainly wrong (I doubt), I'd prefer a new, unloaded wording. Definitely not 'semi-conductor' for that is only one property. I don't think the combination wording 'weak nonmetal (metalloid)' will stick; others will pick one ~randomly, so we'll be introducing a future confusion, also since we (enwiki) cannot control their definitions when used off-site. "Category Q", "MNM", "HM"? (kept as unspecified lettercodes). -DePiep (talk) 16:59, 22 March 2017 (UTC)

That's all good. I intentionally suggested the term "weak nonmetal (metalloid)" to avoid the controversy of coming up with something more novel. What is likely to happen is what happens now. Some people will like our categories; some won't. Some will keep calling boron a nonmetal; others will call it a metalloid, or a semimetal, or a halbmetalle. Others will ignore our polyatomic and diatomic categories and use e.g. halogens (counting At as a nonmetal), and refer to the rest of nonmetals as "the rest of the nonmetals" or "the other nonmetals" or the "less reactive nonmetals" or "less electronegative nonmetals" etc. instead. And that is OK: "weak nonmetal (metalloid)" is unlikely to get anyone's nose out of joint, as it accommodates nearly all preferences.

On using an unloaded category name I strongly fear that the result will be too much of neologism. "Chemical phenomena are highly complex" and hard enough to "treat rigorously from a fundamental theory" (Scerri 1996, p. 171) as is, let alone trying to derive unloaded wording from the existing literature. Since the language of qualitative and descriptive chemistry is reasonably replete with the adjectives "weak", "strong" and e.g. "low" and "high" I figure that "weak nonmetal" as a corollary to "corrosive nonmetal" should be nothing to write home about. (I thought it was bold enough already to claim and implement polyatomic and diatomic as discrete nonmetal categories yet there were no screams. The new scheme is not half as scary; the contours of each category are already present in the literature). [Ref: Scerri ER 1996, "Stephen Brush, the Periodic Table and the nature of chemistry," Die Sprache der Chemie, P Jannich, N Psarros (eds), Könighausen & Neumann, Würzburg, pp. 169–176 (171)] Sandbh (talk) 03:44, 23 March 2017 (UTC)

Polyatomic and diatomic screams?

I think the reason why there were no screams for "polyatomic" and "diatomic" is that those are pretty standard terms. Everyone knows what a nonmetal is; everyone knows what a diatomic molecule is; so no one would bat an eyelid if you said "hydrogen is a diatomic nonmetal". It is a true statement using two standard terms: "diatomic" and "nonmetal". (It is also nice that it extends naturally to the noble gases as "monatomic nonmetals", but that's not the main reason.)

The trouble with "weak nonmetals", "intermediate nonmetals", and "corrosive nonmetals", on the other hand, is that while the contours are certainly there, the terms are not. For example, what exactly does "weak" mean when it modifies a nonmetal? As a counterweight to corrosiveness, it sounds like the noble gases belong there too. And this problem comes from the fact that "weak" is not a standard chemical term for elements (it of course is one for acids and bases, but that definition doesn't carry over). You would have to define what it means here newly, and that is more like original research than "polyatomic" and "diatomic" as standard modifiers.

(I'd also argue that while the contours are there, the exact boundaries are continuous, so it's not clear what should be in what category. We currently have an amicable disagreement on astatine, for example. But this is not the main point that I am making here.) Double sharp (talk) 15:28, 24 March 2017 (UTC)

I thought the screams might have arisen for any of the following reasons.

1. Sure, H is referred to as a diatomic nonmetal but there no concept in the literature of diatomic nonmetals having other shared properties. There are a few references to some nonmetals being "polyatomic" but there is no concept of other shared properties among them. Now, I thought we had good reasons for getting rid of other nonmetals. And diatomic nonmetal was certainly a recognised term, and "polyatomic" by itself was reasonably objective, and found in the literature, and the distinction between diatomic and polyatomic nonmetals happened to bring out a reasonable difference in properties.
2. The categorisation of S as a polyatomic nonmetal, while structurally correct, always worried me from the point of view of attempting to bring out its commonalities with C, P and Se. There is a very large and interesting note about this in the nonmetal article. Truly, sulfur is the Pluto of the polyatomic nonmetals.
3. The existence of ozone, a polyatomic nonmetal, bugged me. Sure oxygen is a diatomic nonmetal but you would hardly call it a borderline diatomic nonmetal, so why does ozone exist? I'd expect this sort of thing for an iodine-like nonmetal, not oxygen.
4. The division was based on a structural distinction, unlike any of the other categories. Chemists don't think of the elements in these terms, not immediately, anyway.
5. I worried a little about calling iodine a diatomic nonmetal. It is as a gas, but in its standard condensed state it has some polyatomic character. There is evidence for significant intermolecular coupling between the individual iodine molecules (each iodine atom forming weak bond with its two next nearest neighbours, as well as a stronger bond with its molecular partner) implying a bulk coordination number of 1+2, rather than 1. There is a note along these lines in the metalloid article. The funkiness of the crystalline structure of iodine is further borne out by its similarity to that of gallium.
6. Only S is a polyatomic nonmetal in the sense that one can point to an S₈ molecule. There are no discrete molecule-equivalents for C, P and Se in their most thermodynamically stable forms.

It feels odd to critique a proposal that I put a lot of effort into at the time. However I thought it was the best of the alternatives then available to us.

I agree "diatomic nonmetal" is a plausibly standard term. As noted, that's about all one could say about it i.e. that some of the chemical elements are diatomic nonmetals. End of conversation. No further discussion in the literature, direct or indirect on their other shared properties.

Except for the noble gases, the literature on classifying the nonmetals is quite challenging, as is the case for metals in, or in the vicinity of, the p block.

This doesn't mean we have to come up with newly defined terms for categorising the nonmetals. The literature has already done our work for us. Unlike the polyatomic and diatomic nonmetal scheme, which took nine unsuccessful proposals to surface and is not readily evident in the literature, this one almost fell into place by itself once I thought about the contrast between the so-called "other nonmetals" and their neighbours to either side. (I have to say it was a mental challenge to have to once again contemplate the construct that is the other nonmetals).

We can look at how the nonmetals of concern are described in the literature and seek to apply the same, or equivalent, terminology.

Before I continue:

• The terms "metal" and "nonmetal" are composites. There is no single measure of metallicity or non-metallicity. Given this, when "weak" modifies a nonmetal or a metal, it means that while the nonmetal or metal involved suffers from a deficiency of the attributes that we associate with being a nonmetal or metal, it still flops over the line sufficiently to make the grade as a nonmetal or a metal. In the literature, the terms "strong" and "weak" are relatively commonly encountered when describing metals. This is much less the case with the nonmetals.
• I don't believe I've advanced or suggested the position that "weak" is a counterweight to "corrosive", so I see no issue here.

• I'm puzzled by your observation that it's not clear what should be in what category. The corrosive nonmetals are clear. The weak nonmetals (metalloids) are clear presuming, as we have done, that it is acceptable to include here only the elements most commonly classified as metalloids, plus astatine as per a prior decision of this project. The intermediate nonmetals are what remain. These allocations are consistent with the way the elements involved are described in the literature.

• I agree there is some overlap between the categories. For example, the line between carbon and, say, boron and silicon; and a little bit between iodine and, say, selenium or tellurium; or between selenium and, say, arsenic and antimony, or tellurium. However, I observed the same kinds of overlaps between the polyatomic and diatomic nonmetal categories. Indeed, our nonmetal article says, "The distinction between the three categories of nonmetals, in terms of receding metallicity is not absolute. Boundary overlaps occur as outlying elements in each category show (or begin to show) less-distinct, hybrid-like or atypical properties." More generally, we have a section in our periodic table article addressing instances of this kind occurring in the rest of the periodic table. In our categorisation decisions it is perhaps preferable to minimise such overlaps but, then again, I'm not so sure that to do so would accurately reflect periodic table reality, which tends to be more messy. So, on the question of overlaps, I cheerily say, yes, that's an interesting feature of the periodic table: there are always hard cases at the boundaries.

Getting back to looking at how the nonmetals of concern are described in the literature and seeking to apply the same, or equivalent, terminology:

• We know that the literature talks about strong metals, weak metals, weak nonmetals and strong nonmetals. Sometimes the language used is more specific and it gets couched in terms of e.g. "less reactive nonmetal", "moderately active nonmetal", or "more strongly electronegative nonmetal".

• We know that oxygen and the halogens regularly attract superlatives in terms of e.g. their electronegativity, ionisation energies, and oxidising power. Indeed, the word "oxidation" arose from oxygen's formidable combining power. We know they are all described as being corrosive.

• We know that metalloids generally behave chemically as nonmetals; that they show less tendency to anionic behaviour than other nonmetals; that they always give compounds less acidic in character than the corresponding compounds of (regular) nonmetals; that they have the lowest electronegativities and ionisation energies of the nonmetallic elements; and that they are wimpy enough—nonmetallically speaking—that they can be cajoled into forming alloys and organometallic compounds.

• We know from the literature that the remaining nonmetals are generally not described in terms of these extremes and that they are most commonly termed as "other nonmetals". We know that nobody really likes this term, and that the meaning of "other" includes, "existing besides, or distinct from, that already mentioned or implied".

• Given the moderate nonmetallic character of these other nonmetals, and since they are found on the periodic table besides the weakly nonmetallic metalloids and the extreme or corrosive nonmetals, calling them "intermediate nonmetals" seems like the most literature-consistent and descriptively detached label available to us, in light of the meaning of "intermediate". Here is that meaning: "coming or occurring between two things, places, etc.; holding the middle place or degree between two extremes; interposed, intervening…in spatial position: situated in the middle place, or between two things or places".

In summary, the proposed division is more natural than that of the current scheme, in terms of the way the nonmetals involved, and their properties, are described and discussed in the literature (anomalies and all). Sandbh (talk) 01:35, 27 March 2017 (UTC)

re "no screams about 'polyatomic' and 'diatomic'", the first part of this section. Indeed there were no screams back in 2013 or afterwards (although Nergaal did make some provocative moves). However, it was not adopted either in other lang-wikis or in talkpage's common parlance. Also telling is that the two categories even today don't have an article. With this, it's quite obvious that readers and iw-translators stick to the 'other nonmetals' name variants, possibly after a 'don't understand' moment or after sensing that 'can't point it, but something is wrong with these names' (as Sandbh decribed above: not a true chemistry categorisation here).

There came a stronger negative effect with this: when not adopting the poly-/di-atomic categories, people also don't adopt the claim 'astatine is a metalloid' (we introduced at the same time). So I can fully support the search for better categories in this part of the PT. -DePiep (talk) 06:33, 3 April 2017 (UTC)

I think that comes about mostly because astatine is a halogen and a metalloid. If you put "halogen" as a subcategory of nonmetals, you inextricably run into the issue that astatine is almost certainly not a nonmetal, and tennessine is predicted to be a metal. The fundamental problem is that those wikis are trying to use a group "halogen, group 17" as a category. Japanese Wikipedia does it OK at ja:Template:周期表 only because it puts metallicity as colour, while it puts groups as borders, so that astatine can simultaneously have the halogen border and the metalloid colour. If you don't break this drastically from the en.wikipedia colour scheme, astatine becomes a big problem. Double sharp (talk) 06:55, 3 April 2017 (UTC)

I quite like this line, "As of 2017, Mendeleev's chart has been elevated to a polychromatic icon [emphasis added] emblematic of the successes of modern science" (Shattered symmetry, Group theory from the eightfold way to the periodic table, 2017, p. 323). While the authors were talking about the periodic table in general, rather than our table, this description struck a chord with me. Sandbh (talk) 21:40, 3 April 2017 (UTC)

Sure, and now we started getting into polychromatic iconoclasm. Agree with the Double sharp description, but my point mainly was: our names did not catch up. btw, fr:wiki did. -DePiep (talk) 09:28, 6 April 2017 (UTC)

"metalloids" as "nonmetals"

• Another issue with the proposed name is, that it claims the metalloids as 'nonmetals'. Sure there will be some arguments for this in literature, but to me it is a surprise major change. Our 192 sources on metalloids did not float what must be such an obvious characteristic. And, seeing the property comparing overviews on multiple characteristics, also does not push them into nonmetals (in some properties they are plain metals even). So naming them 'nonmetal' is a change of concept difficult to get. In other languages they are still called de:halbmetalle. Some naming history is nicely described in metalloid (the lede for starters). All in all, I gather that it is bad practice to use a single characteristic description for the overall category name. -DePiep (talk) 12:07, 12 March 2017 (UTC)

This is why I am proposing to change the name to "weak nonmetal (metalloid)". Our lists of metalloids article was only a list. It's purpose was to show which elements were classified as metalloids in the literature. This explains why it does not say anything about metalloid chemistry.

At this point I feel it would indeed be helpful to summarise more of the state of the literature.

• There is consensus that some elements, in the vicinity of a line running roughly through boron to astatine, can be tricky to classify as either metals or nonmetals.

• Some authors classify these elements as metalloids. Others make a call and classify each of these elements, on a case-by-case basis, as either a nonmetal or a metal. In this scenario, there is consensus that boron, silicon and tellurium are nonmetals. Germanium is sometimes classified as a metal and sometimes as a nonmetal; the same goes for antimony and polonium. In light of astatine's status as a halogen, its unimportance, and seeming uncertainty as to its properties, the lowest common denominator consensus is to assume that it's a nonmetal.

• Each of these classification decisions are open to challenge. This rarely occurs, as long as the rest of what the author writes is not completely inconsistent with their original classification decision (and presuming this decision was not blatantly erroneous on the first place).

• There is consensus in the literature that metalloids have properties that are intermediate between those of metals and nonmetals or properties or a mix of metallic and nonmetallic properties. There is consensus in the literature that metalloids look like metals and are semiconductors. (The fact that arsenic and antimony are not semiconductors in their most stable forms is a common oversight).

• Much of the literature stops at this point. Any further discussion of elements that are considered to be metalloids generally occurs only in the parts of texts giving the general descriptive chemistry of the elements in each p block group.

• There is consensus in the more considered literature that metalloids or nonmetals in the vicinity of the dividing line between metals and nonmetals generally behave chemically like weak nonmetals (and that metals in the same vicinity generally behave chemically as weak metals e.g. Tl, Pb, Bi).

In summary, the proposed term "weak nonmetal (metalloid)" captures the complexity of the situation with a high degree of consensus—much higher than we have now with "metalloid". By itself the term "metalloid" is misleading, since it means 'resembling a metal', which is less than half the story. But the term weak nonmetal (metalloid) expresses the nature of these elements much more comprehensively, and accurately.

As to our own content, the metalloid article, the properties of metals, metalloids and nonmetals article, and the origin and use of the term metalloid article each note that the overall chemical behaviour of metalloids is nonmetallic. Yes, some of their properties are halfway between metals and nonmetals, but none of these properties are beyond the scope of what you could reasonably expect to find in nonmetals on the borderline between metals and nonmetals. And most of the metallic properties of metalloids are physical properties, and physical properties are trumped by overall chemical behaviour. Sandbh (talk) 04:32, 20 March 2017 (UTC)

Metallic character and the meaning of "metalloid"

I am not sure that they should be trumped so completely though, especially in such a borderline region when no one's arguing over which elements are in the group. If you look at any p-block group going down, you will start with mostly nonmetallic character (notwithstanding that boron is an "honorary metal atom" and vice versa), and you will gain more and more further down. If we look at group 16, for instance, O and S are insulators, Se and Te are semiconductors, and Po is a metal; and while chemistry lags behind a little, we do see some emerge, as Se is not attacked by dilute HCl, while Te dissolves slowly and Po dissolves enthusiastically to give pink Po^II and self-oxidises to yellow Po^IV. And tellurium even looks like a metal: there's a reason why Müller, who discovered Te, called it metallum problematicum.

I don't think it's a far stretch to say that the metalloids look like metals but act like nonmetals, although of course we have to take this with a pinch of salt (I suppose that needs to be Na₂Te for this post), but in contrast to humans where beauty is no guarantee of moral character, the shiny lustre does say something about an element's physical properties, which we shouldn't neglect. There's even fuzziness on the other side; Ge dissolves slowly in hot concentrated H₂SO₄ and HNO₃, proving that some of the metallic character that we had expected is there, if muted by the d-block contraction. But even Sn is amphoteric, as is shown by the following reaction:

Sn + 2 KOH + 4 H₂O → K₂[Sn(OH)₆] + 2H₂

It also has a clearly nonmetallic allotrope to boot, while Si and Ge transform to metallic tetragonal β allotropes like Sn if you just subject them to pressure along the c-axis (~200 kbar for Si, ~120 kbar for Ge). Even the SnX₄ are tetrahedral volatile solids and liquids and the obvious comparison is with the tetrahalides of Si and Ge. So tin appears to be fair game for being a "weak nonmetal" as well in this sense, which is funny because if you go one step to the right(!) you obtain a more clearly ionic compound in SbF₃ and even a clearly intermediate SbI₃! Furthermore, looking at group 15, As, Sb, and Bi have very similar electronegativities, so that cannot be the only guideline for grouping them as metals, nonmetals, or metalloids; the difference is structures is more reflective of how the octahedral gaps in the hcp iodine lattice are filled by each of these elements, and only Bi is large enough to fill them symmetrically.

Of course, we can wave away the question of why such elements are separated by appealing to the literature, but this is science: if the literature says something, we should critically examine it to figure out why something is being said.

And even with clear if intermediate nonmetals like phosphorus, even the halides are getting close to the ionic-covalent divide: PCl₅ is ionic [PCl₄]⁺[PCl₆]⁻ in the crystalline phase and covalent in the gas phase, and PBr₅ (which is [PBr₄]⁺[Br]⁻) outright decomposes in the gas phase. Indeed, phosphorus is so common an element that I remember being told this one in high school. ^_^ The solvent itself also influences the choice between ionicity and covalence, and even tweaking the substituents can lead to massive changes: PhPCl₄ is molecular while MePCl₄ is ionic. There is metallic character even here, and focusing on the nonmetallicity fails to show it clearly.

But the main point is that "metalloids" is way more common in the literature than "weak nonmetals". The "metal" tells you that at face value, that is what they look like: the "-oid" suggests that appearances can be deceiving. What is there more to ask for a term? Double sharp (talk) 14:59, 22 March 2017 (UTC)

After a long stint at responding to Double sharp, Sandbh reaches for his tequila and lime, when he suddenly sees the label on the salt shaker...

Right then, I'll give this a go.

I recall reading in an old chemistry reference that metals and nonmetals used to be distinguished according to their physical properties but that "these days" (back in the 1800s, I believe) chemical properties are more important in making the distinction. So when I said chemical properties "trump" physical properties I may have gotten a bit carried away. It's probably better to say that chemical properties are more important, and that this does not mean physical properties have no relevance.

Metallic character is present even in the extreme nonmetals i.e. the halogens, never mind the intermediate nonmetals, in the right (extraordinary) circumstances. Same goes for nonmetallic character in the alkali metals. Yes, focussing on the nonmetallic character, of e.g. the intermediate nonmetals, will fail to show their metallic character. Same goes for the metals generally and their nonmetallic character, but these are matters of detail added to first order generalisations and category assignments, as you have noted.

My recollection is that cationic behaviour or the acid/base nature of oxides are important considerations in distinguishing between metals and nonmetals. Tin forms a cation in aqueous solution ergo it is near universally regarded as a metal, albeit a weak one. If the case was any stronger for calling tin a weak nonmetal, I'd expect to see it be predominately called as such in the literature yet no one does.

Germanium is curious. Mendeleev predicted it would be a metal but I don't know what he meant. Was he referring to its appearance or its chemistry? Why would he predict it would have a metallic chemistry if he was presumably familiar with the weakly metallic chemistry of tin? As it turns out, when Winkler announced the discovery of germanium in 1886 he called it a new non-metallic element, on the basis of its chemistry!

None of the other metalloids or nonmetals are capable of forming a simple cation in aqueous solution, as far as I know. Yes, astatine can, but then I reckon we should classify it as a post-transition metal. Neither do any of them, as far as I can discern, form oxides that are predominately basic i.e. metallic.

The term metalloid is certainly more common in the literature than weak nonmetal. That is why I am proposing to keep it. I suppose that the "metal" prefix tells you at face value what they look like (but it doesn't tell you that they mostly don't conduct electricity like metals). As to the suffix "-oid" this means "resemble" or "like" doesn't it? When I read "metalloid" I think of something resembling a metal. So I'd expect it to be almost but not quite a metal. What I get instead is something that is brittle, with no structural applications, and that has the general chemical behaviour of a nonmetal. Ripped off! Sold a pup! ^_^ The term "metalloid" tells less than half the story.

We know that IUPAC tried unsuccessfully to get rid of the term metalloid in 1959, and again in 1971, and to replace it with the term "semimetal". These were poor efforts given the physicists had already appropriated the term, "semimetal". The Encyclopedia of Chemistry (Hampel & Hawley 1973, p. 727) had another go: "The term "metalloid" is becoming obsolete in that the term nonmetal is more precise. The term semiconductor now is also applied to such nonmetal elements as silicon, germanium and selenium...". The Condensed Chemical Dictionary, which has been around since 1919 and is currently in its 15th (2016) edition, is closer to the mark. In the entry for "nonmetal" (p. 988) they count metalloids as nonmetals that more nearly resemble metals than then rest of the nonmetals, but say of the term "metalloid" that "it is no longer used by chemists", which is not quite right. (I'm not sure what is happening here since their nonmetal entry has read this way for at least their last six editions, since 1977, under different sets of editors).

Whatever the goings on with the term metalloid, it still has its place in the literature hence I'm proposing to keep it but to use the much more representative term "weak nonmetal (metalloid)".

PS: I expect I'd politely decline having Na₂Te metalloid salt with my tequila and lime, given the toxicology of Te. Sandbh (talk) 02:41, 24 March 2017 (UTC)

Meaning of "metalloid" • cations; interconnectedness

Well, the suffix of the term "metalloid" is originally -ειδής in the Ancient Greek that it comes from, which comes from εἶδος "form, likeness". So when I see "metalloid", I expect something that has the form and likeness of a metal, but is not actually one; for if it were one in all of its behaviour, then why not just call it a metal? So to me it does not really feel like the term rips us off; rather, it gives us a warning that while these elements might look like metals, they don't act like metals.

I wouldn't call PCl₅ particularly extraordinary. I mean, it is in the sense that it is right next to the ionic–covalent divide, but it is not in the sense that it really is quite a common compound. If not, we should have to call water extraordinary, as you do not see its characteristic properties in H₂S, Li₂O, or F₂O. Looking at things that way seems to be along the lines of Novalis talking about making the familiar strange and the strange familiar, because it is only on account on the singular properties of the elements we use the most that they can support something as complicated as life. So although they do not really show the most common behaviour among the elements, they show the most common behaviour if you factor in how often we see certain compounds.

What this means, of course, is that there are certain elements for which the first-order rationalisation of their properties is so far off the mark that it is essentially worthless. I reckon that this happens to a great extent with the first-row anomaly, which is why H, B, C, N, and O are so different from the rest of their groups, and even Li, Be, and F, while visibly allied, have some qualitative differences that can't be overlooked. If we tried to predict the properties of H₂O in a first-order sort of way from those of H₂S, H₂Se, and H₂Te, I suspect that our first prediction would be a rather foul-smelling gas (though it would presumably be relatively tolerable compared to its congeners) that condensed at around −100 °C. Clearly, the first-order prediction overlooks so many things that it's not very useful, and to make valid predictions for a case like this you need to know the answer already by knowing the singular chemistries of H and O.

Now, the trouble I find with the idea that forming a cation in aqueous solution is that important as a factor (instead of being merely suggestive) is that many of the transition metals do not form such a cation in their most common oxidation state. Indeed, when one considers how anion formation tends to stabilise high oxidation states, one sees many similarities between the middle of the transition metal group and the nonmetals that they are placed with in the short form of the periodic table. Consider silicon, the most electropositive of the metalloids (EN 1.90). This value is equalled by Cu, Tc, and Re! Consider hydrogen at EN 2.20: this is equalled by Ru, Pd, Ir, and Os (with Mo not far behind at 2.16), and even surpassed by Rh and Pt (2.28), W (2.36), and the champion Au (2.54) which almost equals carbon (2.55), and shows a suggestive proclivity towards catenation and even forms a simple anion in Rb⁺Au⁻.

If we consider tungsten, for example, its most characteristic oxidation state is +6, a very high one characteristic of clear nonmetals. And we get molecular polytungstate anions with oxygen in this oxidation state, and the most well-characterised halides are the hexahalides, which are practically covalent compounds (WF₆ is famously a gas, and the other two are pretty volatile). And the only electrode potential listed by NIST for a W cation (W³⁺/W⁰) has a footnote to (a) saying "This half-reaction contains at least one doubtful chemical species"; since no one doubts the existence of W metal, it must be the putative W³⁺ that is being referred to. Finally, what of the tungsten oxides? They are mildly acidic, dissolving in aqueous alkali to form tungstates!

Is any of this what we would have expected for the chemistry of a metal? Absolutely not! The arguments for calling W a metal are mostly physical; it is hard, strong, dense, and a good conductor of electricity. Except that even then we can't say that for sure, because W is also brittle! If we are agreed that tungsten should count as a metal, then I would imagine that these criteria need serious revision or supplementation. Between the elements that define what the public considers to be standard metals – the transition metals – and the criteria that most people think of as chemically defining metals – there is a great conflict.

Admittedly there is not much family resemblance here

Note that I do not believe that this detracts from the power and effectiveness of the periodic table. In all of its supreme principles and rules, there are always exceptions, and the rules and the exceptions strengthen each other in an unceasing spiral until they unite and become indistinguishable from each other. A good example of this is how Greenwood and Earnshaw, on p. 27, include "anomalies" and "anomalous" properties in their list of "periodic trends which occur in the chemical properties of the elements", because it is impossible to get a good understanding of period 2, the lanthanides, and the heavy p-block elements without taking these into account so consistently that they harden into a new set of rules! These necessary insights for reading the periodic table are a concealed, self-destructive assault on the principles that it is itself founded on, and it paradoxically gives it the power to cast its net wide and apply the periodic law to every element we know of, from the beautiful simplicity and singularity of hydrogen to the scintillating complexity and radioactivity of oganesson. Its imposing castle-like structure, radiating the confidence of a mathematician, masks this self-destructiveness, and yet the two are simply two sides of the same coin. I realise that this makes Mendeleev sound like Mozart (and I got this idea from reading Charles Rosen's The Classical Style, which describes Mozart's art as subversive and uncompromising in similar terms to what I have done for Mendeleev's, but of course in an even better way), but the more I think about this comparison the more valid it seems to me.

As for germanium and tin, I would imagine that if Mendeleev saw no reason to deny Sn metallic status on account of its amphoteric aqueous chemistry, then he probably saw no reason to deny it to what he predicted for Ge as well. It should also be remarked that Si was originally thought of as a metal, because its oxide was analogous to aluminium and beryllium oxide, so much so that Sir Humphry Davy originally named it silicium; this could also have been a major factor in considering germanium to be metallic. In any case, Si, Ge, Sn, and Pb all have amphoteric oxides: yes, even the clearly metallic (if soft and low-melting) Pb. But I am not surprised that the proposed criteria runs into this problem in the p-block, if it is already straining in the d-block with tungsten. I have no doubt that they came from a great deal of thought, and might summarise how people mentally think for some portions of the table, but I believe that we need to press on and find something even better.

I am looking forward to your reply after you enjoy your tequila and lime: after all, I did plan this reply for over an hour, and then proceed to edit and expand it another nine times. ^_^ I must say that this conversation has been wonderful so far at making us both think and reexamine the amorphous blob of principles and connections between them that constitute our understanding of chemistry. Double sharp (talk) 04:16, 24 March 2017 (UTC)

Broadly, we seem to agree about the suitability of the term "metalloid". Regardless of our differences in interpreting the term, a metalloid is not a metal. I presume a metalloid can therefore be regarded as a "non"-metal—a distinctive kind of such, but a nonmetal nevertheless.

On the "cation in aqueous solution" consideration, my impression of the literature is that this only tends to become important when attempting to sort out what happens when the metals meet the non-metals in the p block.

I remember reading in C&W that some of the transition metals had no simple cationic chemistry but the fact that tungsten has no basic oxides was something that hadn't occurred to me. It happens that tungsten is listed as forming a "green-yellow" trivalent monoatomic cation in The Aqueous Chemistry of the Elements (Schweitzer & Pesterfield 2010, p. 306) but the same authors show germanium as forming a colourless monatomic divalent cation (p. 190)! No supporting citations are given apart from a general reference list at the back of the book. Now, in the germanium section of our metalloid article there is a note on whether Ge can form such a cation and I stand by my assessment that the evidence is unclear. I haven't done the same degree of research on tungsten but I suspect the same may apply to the notion of a tungsten cation. Having said all that I don't think the fact that a few d block metals may or may not be able to form cations matters. Everyone acknowledges they are metals, wrinkles and all.

(Looking back at your comments it seems to me that your reference to the "cation in aqueous solution" criterion as being "suggestive" was prescient, in the context of what I've written above).

So, no, I think there is no need to come up with a more rigorous way of distinguishing between metals and nonmetals. For our purposes, the literature has already done the work for us in distinguishing between metals and nonmetals (including metalloids).

Continuing in this theme, I'm not so sure that I would describe the irregularities and anomalies in the periodic table as a concealed self-destructive assault on the principles that the PT was founded on, since these were loose in the first place. I like this quote attributed by Scerri (1996, p. 174)^ to Peter Nelson, "who often writes on matters of conceptual chemistry": "The application of periodic law is not simply a case of making logical deductions from basic principles as in the case of thermodynamics. It involves rather a fairly thorough knowledge of how the principles work out in practice - of exceptions, trends and patterns [e.g. the first row anomaly, anionic gold, rogue transition metals etc] - so that a particular deduction can be appraised and if necessary adjusted. In the limit the process becomes virtually intuitive". I think however, that you arrive at essentially the same conclusion!

^ details as previously noted

Not that it matters much but your thoughtful speculation as to why in 1869 Mendeleev predicted germanium would be a metal doesn't seem to mesh with what was known about silicon at the time. Borrowing from our article on silicon, Thomas Thomson, in 1817, renamed silicium as "silicon" since, in his words (and with my best Scottish brogue engaged): "there is not the smallest evidence for its metallic nature, and as it bears a close resemblance to boron and carbon, it is better to class it along with these bodies, and to give it the name of silicon." I presume Mendeleev would've been aware of the nonmetallic nature of silicon by the time he made his prediction. On silicon having an amphoteric oxide, I gather from the literature that silicon is normally regarded as having a weakly acidic oxide. There is a note about this in the silicon section of the metalloid article. I had a look at the Davy paper on silicon but could not make sense of what the similarity was with oxides of aluminium and beryllium.

I did not have any tequila at hand so I settled for some refreshing whites, three walks, watching the AFL Women's grand final on TV, and some study time with musical accompaniment. And yes, this is a most absorbing and rewarding thread. Sandbh (talk) 10:19, 25 March 2017 (UTC)

Key properties; periodic law; W

Yes, I think that would be a fair portrayal. The fact that the metalloids need the "-oid" suffix suggests immediately that they're not metals. The need for an intermediate classification comes more, I suspect, from the fact that both the metals and nonmetals have their own characteristic properties, and if you use those characteristic properties instead of defining "nonmetals" simply as the complement of the set of "metals", then it's difficult to group cases like the standard nonmetals in one group or the other. If we consider the listing at Properties of metals, metalloids and nonmetals, for example, we see that the bulk of the properties of metalloids are reasonably distinct from the characteristic ones of the metals and the nonmetals. This would seem to suggest the validity of regarding them as a class apart, instead of taking the term "nonmetal" at face value.

On the "self-destructiveness" of the amendments to the periodic law, I tried to go for the most colourful and eye-catching language possible, and apparently went too far. ^_^ But this is what I was going for here: the periodic law is loosely constructed as applying to every element from H to Og, but then you have localised exceptions to it, and then the localised exceptions at some points harden into another loosely constructed law, except that these new laws also have exceptions. So every level of the "law" has an exception until the laws and the exceptions become indistinguishable. I did not mean to say that this reduces its efficacy, and indeed I do not believe it does. Like Mozart's art, Mendeleev's magnum opus turns into a conception of pure intuition, as one cannot consider so many factors at once: one merely feels it. To continue that comparison, surely Mozart did not learn the musical language of his time that he mastered so well by reading simplified accounts and principles and building them on each other; he must have learned it through listening and developing an intuitive sense of balance. Much the same sort of intuition seems to hold when "power users" use the periodic table as a predictive tool, exactly as you say. ^_^

This intuition does have a few consequences, because some of those amendments to the periodic law are pretty localised. It seems that when chemists think of criteria like cation formation to distinguish metalloids, they generally only think of the contentious region in the middle of the p-block. The fact that some of these criteria would literally exclude some of the heavier d-block metals from metallicity then gets overlooked.

So, what I would then ask is: if we are very sure that tungsten is a metal, which of its properties give us this confidence, given that its chemistry does not inspire us with this confidence? ^_^ We can then consider these other properties and use them to test the well-known behaviour of the metalloids (ignoring At, which we colour as one mostly because of the absence of experimental evidence of clearly metallic behaviour, instead of evidence of its absence).

What I was thinking of was the fact that Davy considers the oxides of Al, Si, Be (and, looking it up for real now, also Zr) in the same breath. He writes "Had I been so fortunate as to have obtained more certain evidences on this subject, and to have procured the metallic substances I was in search of, I should have proposed for them the names of silicium, alumium, zirconium, and glucium." I suppose Mendeleev may have noticed the somewhat borderline behaviour of Si (which does have a metallic lustre, after all), and expected metallicity to become more evident at Ge as an interpolation between Si and Sn. In fact, since there were no A and B groups yet, he may also have been swayed towards a metallic prediction for Ge by looking at Ti and Zr, also in group IV of his table. But ultimately, I suppose we can't know what he thought unless he wrote it down somewhere. Double sharp (talk) 13:44, 25 March 2017 (UTC)

Cool.

I was under the impression that nonmetals would be thought of in terms of having their own characteristic properties until I started writing the first iteration of the current version of our nonmetal article. Upon looking at the literature I was surprised to see that nonmetals were, with some exceptions, more often thought of as elements that lacked metallic attributes, rather than as elements that possessed a preponderance of nonmetallic attributes. They were thus being characterised in terms of what they weren't rather than what they were! That made writing the article a bit challenging.

I agree metals and nonmetals tend to have their own characteristic properties.

Even so, we know it is well acknowledged in the literature that there is no rigorous definition of a metal. Curiously, there is no such acknowledgement, AFAIK, that there is no rigorous definition of a nonmetal. It seems the fate of the nonmetals still rests upon the nature of the metals.

We know that some authors make the call on an element-by-element basis as to which are metals and which nonmetals, and that some authors decide this is too hard in some cases and call the elements concerned metalloids. We know that some authors who make a call on an element-by-element basis nevertheless acknowledge, in some fashion, that some of the elements in the vicinity of the dividing line between metals and nonmetals are sometimes called metalloids.

We know from the literature that the elements called metalloids generally behave chemically as nonmetals. We suppose that if any of the individual elements involved had been judged to generally or significantly behave chemically (even physically) as metals that they would presumably have been called metals—potentially weak metals, but metals nonetheless.

I agree that the bulk of the properties of metalloids are reasonably distinct, which is a good thing as it makes them easier to characterise. I note that they generally behave chemically as nonmetals. I further note that the bulk of their reasonably distinct properties are consistent with what I would expect to see in nonmetals that more nearly resemble metals than the rest of the nonmetals.

We know that we have to make a call as to how to divide the elements between metals and nonmetals on our own periodic table.

While eliminating the metalloid category name and changing it to "weak nonmetal" would be a step too far, changing the name to "weak metal (metalloid)" would be more consistent with the literature. Adding the two other categories of "intermediate nonmetal", and "corrosive nonmetal" (for the superlative-laden nonmetals) would be a more literature consistent way of completing the annexation of this part of the period table.

We know that people expecting to see metalloids on our periodic table will still be able to find them. We know that people expecting to allocate all of the elements between metals and nonmetals will be able to do so. Winners all round.

On the "self-destructiveness" of the amendments to the periodic law, you have excelled your own writings about this and captured what is going on here beautifully.

About tungsten, I've never thought much about it. It seems to have been regarded as metal from the outset purely because of, in addition to its shininess, its incredible toughness and I suppose thermal conductivity, and subsequently electrical conductivity. Yes, it is brittle, but not tap-shatter brittle. From our article I see that pure single-crystalline tungsten is more ductile, and can be cut with a hard-steel hacksaw. As to its cationically-challenged chemistry, I presume—not having looked at specialised tungsten literature—it otherwise behaves like transition metal (variable oxidation states; complex formation; coloured compounds; magnetic properties; catalytic activity)? I had a quick look at The chemistry of the transition elements (Earnshaw & Harrington 1973, p. 59) and was surprised to see electrode potentials for W → W³⁺ → WO₂ of, respectively, –0.11V and –0.15V in acid solution, when on the previous page they say that Mo and W don't form monoatomic cations. The W³⁺ must be a polycation, or dubious. There are also figures for the Mo equivalent.

If you are interested in something more quantitative, I have previously defined a metal as an element that meets at least two of the following criteria:

a. low ionisation energy (< 750 kJ)

b. low electronegativity (< 1.9)

c. metallic band structure

d. high packing efficiency (> 41%)

e. high Goldhammer-Herzfeld metallicity ratio (> 1.1)

Anything else is nonmetallic. Tungsten gets up on c., d. end e.; it only just misses out on a. (770 kJ) which surprised me. Bismuth meets three of these (a., d., and e.). I see Bailar et al. (1984, p. 951)^ refer to bismuth as being, "the least 'metallic' metal in its physical properties…brittle rather than malleable, and [with]…the lowest electrical conductivity of all metals." Which element has the lowest electrical conductivity is debatable but bismuth is certainly in the lowest cohort.

^ Bailar JC, Moeller T, Kleinberg J, Guss CO, Castellion ME & Metz C 1984, Chemistry, 2nd ed., Academic Press, Orlando

Polonium is the only metal among the first 100 elements that meets only two of the criteria (c. and d.).

RL commitments not considered, tungsten would look like FAC development-bait, not to mention bismuth. Hrmmph. Talk about a muscle-bound colossus of a metal and an uber-feeble metal! Would like to see a side-by-side table of the properties of these two. But this is idle speculation given all my other interests. Sandbh (talk) 10:09, 27 March 2017 (UTC)

How chemists informally use the words

Okay, but I would think that in the absence of a rigorous definition of whatever "metals", "metalloids", and "nonmetals" happen to be, we ought to look at how chemists already informally use the words, instead of how we should redefine them for the better, at least for Wikipedia. In this case, Greenwood and Earnshaw in particular seem to treat the categories as a trichotomy instead of a dichotomy. When comparing Ge, Sn, and Pb on p. 373, they call Ge a metalloid, while they call Sn and Pb metals. They also seem to think of metalloids as being equally far from metals and nonmetals, calling Ge instead a metal when introducing Si and comparing it with the "heavier metals of the group" on p. 328: now this makes no sense when we consider only chemical properties (when nonmetals are much closer), or only physical properties (when metals are much closer), but it makes quite a bit of sense if we consider the union of these sets of properties. (IIRC Greenwood and Earnshaw has a similar list of metalloids as ours, but considers boron a nonmetal and declines to classify astatine for obvious reasons.)

I suspect the metallic elements of the metalloids (e.g. the lustre and all that entails for their physical structure) get a little short shrift here because it seems more obvious: it's how they look, and looking underneath reveals the nonmetallic properties. But the elements are not like people where the personality has little to do with the appearance, and surely both must be taken into account to emphasise both physical and chemical properties. We're not drawing the table here, which is founded more on chemical properties; we're colouring sets of elements by their common properties, and surely a significant number of the weaknesses inherent in the post-transition metals are physical in nature, just as those inherent in the transition metals are chemical in nature (looking for instance at W and Au; the properties you list to classify W as a metal are all physical). So there are actually two groups of "weak metals" which are weak for two different reasons, and I agree with you that I want a side-by-side comparison of W and Bi!

It's only when we get to the metalloids that both sets of properties start getting invaded,^{[involved? -- Sandbh (talk) 03:19, 29 March 2017 (UTC)]} and it's not until we get past the metalloid line that decidedly metallic properties are in the minority, I think. Put like that, I think the trichotomy that we suggest in Properties of metals, metalloids and nonmetals makes a significant amount of sense. We may quibble over some boundaries, and we may wish to annex some territory to clearly metallic or clearly nonmetallic elements to iron out the discrepancies, but I think that alerting ourselves to the fact that there are such "intermediate" cases is pretty important.

P.S. It also occurs to me that if we take chemical properties as primary, hydrogen feels like it should be a "weak nonmetal" as well. After all, didn't you tell me some time ago that earlier chemists thought that solid hydrogen would be metallic? ^_^ Double sharp (talk) 15:55, 28 March 2017 (UTC)

By the way, I was trying to be evocative with "invaded", along the lines of the nonmetallic properties going into both the physical and the chemical realm, but reading it as "involved" will not do any violence to my train of thought and may actually be clearer. ^_^ Double sharp (talk) 06:42, 29 March 2017 (UTC)

LOL! Sandbh (talk) 09:51, 29 March 2017 (UTC)

Yes, certainly, when some chemists describe the elements making up groups 14 to 16 they use the terms nonmetal, metalloid and metal. If you are lucky they may mention the semiconducting status of e.g. germanium. But when they go on to discuss the chemical properties of the elements involved the metalloid status of whichever elements they happened to refer to as metalloids fades away, and the actual descriptive chemistry reads like that of a nonmetal.

I'm doubtful that G & E call germanium a metal. When they introduce silicon on p. 328 they say, "there are notable differences from carbon, on the one hand, and the heavier metals of the group on the other." It isn't clear if they are referring to germanium as one of these heavier metals. I'd be quite surprised if this was the case, given what they would've known about germanium. And later, as you note, they call germanium a metalloid with a band gap smaller than that of silicon, and say that Sn and Pb are very soft, low melting metals (p. 371).

Their use and "non-use" of the term metalloid, and related concepts, is curious:

• As well as not referring to boron as a metalloid, neither do they refer to silicon and tellurium as metalloids.

• They say that, "At higher temperatures, boron reacts directly with all of the non-metals expect H, Ge, Te and the noble gases." (p. 145). So, Ge is a nonmetal?

• They say many of the chalcogenides of Ga, In and Tl are semiconductors, semi-metals, photoconductors, or light emitters (p. 253). The relevance of this citation is some later inconsistency in their use of the term semimetal.

• They say graphite is a semimetal (p. 277) but later refer to arsenic and antimony as "metalloids or semimetals" (p. 548). Does this mean germanium is also a semimetal?

• Silicon is a semiconductor (p. 331).

• On germanium, they say Winkler thought he had discovered a metalloid like arsenic or antimony (p. 367). Winkler actually announced he had discovered a nonmetallic–nicht-metallisches–element.

• Later in the chapter on germanium, tin and lead they refer to SnTe and CsSnI₃, CsSnBr₅, Cs₄SnBr₇ and compositions in the system CsSn₂X₅ (X = Cl, Br) as transforming to black metalloids, on being warmed (pp. 380–381). I presume they are here referring to substances that look like metals.

• Black P is semiconducting (p. 482).

• On arsenic they say its lack of ductility, high electrical resistivity, amphoterism and chemical nature between metals and non-metals have led to it being classified as a metalloid rather than a metal (p. 552). However, in my next citation but one, they refer to arsenic as nonmetal.

• Se and Te are semiconductors (p. 754).

• Cobalt reacts on heating with the halogens and other non-metals such as B, C, P, As and S (p. 1116).

• In their chapter on the actinides they refer to the semi-metallic monocarbides and mono nitrides (p. 1267). I'm not completely sure what they mean here. Is this the same as the black metalloids referred to earlier?

In summary, they regard B as a nonmetal; C as a semimetal or nonmetal; Si as semiconducting nonmetal; Ge as a nonmetal or metalloid; P as a semiconductor or nonmetal; As as a nonmetal or metalloid; Sb as a metalloid; Se as a semiconductor; and Te as a nonmetal or semiconductor.

G & E's work is superb and the terminology they use in referring to the nonmetallic elements is not inconsistent with the literature but it does suffer from some confusing incongruities. That was always my impression of them, just in this regard. If they thought much about the idea of a trichotomy then they only expressed it superficially (at best) in writing.

Not that this matters. While we lack rigorous definitions of metals and nonmetals we have some pretty good literature-based descriptions.

Yes, we're colouring sets of elements by their common properties. Mostly, we do this according to chemical properties, since (as you note) that was what the periodic table was based on.

Yes, a significant number of the weaknesses inherent to the post-transition metals are physical in nature. As we've discussed, the fall off from group 11 to group 12 is dramatic, and things get no better after that. The post-transition metals have their share of chemical weaknesses, too.

The properties I referred to with regard to W's status as a metal included most of the distinctive chemistry-based characteristics of the transition metals, in addition to its singularly formidable physical ones. I wouldn't regard W as a particularly weak metal. Bismuth on the other hand, sure.

Feel free to elaborate or refine what you said about there being different bunches of "weak metals" as I'm not sure what you were trying to say here, or of its relevance.

I regret to advise that the "potent-puny" poster portrait of the W and Bi "titanic tag team twins" is currently backed up behind a heap of other "to do" items.

When you pass from the post-transition metals into metalloid territory the metallic physical properties, with some irregularities, recede still further. When you look at the chemistry, you see that this has now become generally nonmetallic (or rather, this is what the literature tells us). When you get past metalloid territory, but before you run into the teflon-coated corrosive part of the table, the nonmetallic properties become more pronounced and the residual metallic properties become patchy.

The trichotomy in "Properties of metals, metalloids and nonmetals" is artificial. I always wanted to expand that table to show pre-transition metals, transition metals, post-transition metals, metalloids, and the rest of the nonmetal categories. I wanted to see the complete progression in metallic to nonmetallic character that the literature always talks about.^ But I was scared off by the enormity of the task and a little bit of apprehension, perhaps unfounded, that the integrity of the polyatomic and diatomic categories, in some aspects, would look a little weak. It'd be fab enough to have a four-column table in the nonmetal article. We can still note the fact that in some interpretations of the metal-nonmetal spectrum, the elements covered by the weak nonmetal (metalloid) column are thought to occupy the middle ground, and we can add a wikilink to the metalloid article for further information. We do something similar(ish) now, with our articles on rare earth elements, and the lanthanides.

^ In contrast, the great majority of chemistry text books limit themselves to one column for metals and one for nonmetals.

The old chemists thought if they could ever freeze hydrogen that it would be metallic. But we know more about the nature of hydrogen these days. It has the physical properties we normally associate with being a nonmetal, being a colourless, odourless, highly volatile and insulating gas. Frozen hydrogen is transparent, soft, and presumably easily crushed (given it's held together by Van der Waals forces). The only semi-plausible report we have so far of the supposed synthesis of metallic hydrogen said that to do so required a pressure near 500 GPa. "Puny" oxygen, in comparison, only needs 95 GPa.

The chemistry of hydrogen is peculiar, being influenced by its unique atomic structure. Speaking personally, it's hard to get your head around.

• It has quite a high ionisation energy (it's easier to form Xe⁺ than it is to form H⁺) but a modest electronegativity of 2.2, which is at the top limit of the metalloids.
• The great bulk of its compounds are covalent. Some of its compounds with metals have alloy-like qualities (as do some phosphides, nitrides, and sulfides).
• It and the halogens often replace one another in organic compounds, often with no more effect on the general properties of the compound than is produced by the substitution of one halogen for another.
• It does not form a simple anion in aqueous solution although the H^– anion is known in the solid hydrides of the alkali metals, and those of Mg, Ca, Sr and Ba, and Eu and Yb.
• It does not form a simple cation in aqueous solution, there being no such thing as a free H⁺ cation in any condensed medium.
• Most of its chemistry can be explained in terms of its tendency to acquire a 1s² configuration.
• It is clearly not a corrosive nonmetal.
• It's capacity to lose its single electron and form a proton is central to the formation of acids (the latter being associated with nonmetallic behaviour), said proton subsequently and rapidly becoming attached by a covalent bond to another atom.

I think especially the last three dot point considerations, combined with hydrogen's physical nonmetallic properties, are sufficient to assign it to the intermediate nonmetal category.

I expect hydrogen will always be a hard case unless we place it in a category of its own, which some authors have been known to do but has never gotten much of a guernsey around here.

The proposed scheme moves H one category further to the left, which is probably a good thing. Yes, you could argue there is some overlapping of chemical properties with the weak nonmetals (P shows some of this too) but not fatally so, I think. Cue standard response about there always being hard cases at the boundaries etc ^_^

PS. A category called "metalloid (weak nonmetal)" rather than "weak nonmetal (metalloid)" might work too. Sandbh (talk) 07:33, 30 March 2017 (UTC)

Hydrogen is indeed a curious case. However, I would give H⁺ quite a higher position. Certainly, no one has ever seen a naked proton, nor is anyone likely to see one; and yet proton transfer as the basis for acids and bases is very well-established, implying that the protons still have a sort of independent existence as they move from molecule to molecule, and that they're not attached very much to their parent molecule (if they were attached more strongly, then it wouldn't be that strong an acid). So I would say that the acids are more defined by their other constituent elements, instead of H which leaves them as they exhibit their acidity.

I tend to think of hydrogen as a spectacular case of a "zeroth-row anomaly" as applied to what could otherwise have been a rather good metal. Beryllium is a more benign case: even so the Be²⁺ cation would be so small and polarising that it doesn't really exist in more than a notional sense even in BeF₂. In hydrogen's case, H⁺ is another few orders of magnitude smaller, and of course the effect driving it towards covalency is greater, but in its preference for cationic over anionic chemistry I think you could make a very convincing case for analysing it as a metal whose metallicity has been severely compromised by the huge "zeroth-row anomaly", but still surfaces in its predominantly cationic chemistry. For example, if one tries desperately to search for anything in the periodic table resembling H bonds, one finds it in group 1 with Li and Na.

Of course, this is not a particularly strong argument for calling it a metalloid at all, but it certainly suggests that if we are going to put it in a group at all, then the alkali metals are at least visibly distant cousins to it.

As for the trichotomy, I think this is where the "fuzziness" comes in. It seems that most people think of there as being three super-categories (metals, metalloids, nonmetals), but their idea of where a specific element goes can be quite fuzzy. But I think that it would look a lot odder to most of them to consider metalloids as a subset of nonmetals. For Wikipedia especially, the fact that the trichotomy is more common than the proposed dichotomy and appears in the vast majority of reliable sources means that the latter still does not sit well with me. Even for -La-Ac, while I argue for it based on chemical and physical reasons off Wikipedia, I argued for it on Wikipedia on the grounds that it was the most common form in the literature.

I do agree that it would be nice to have more detail as one goes from a metal like Ca to a metal like Cr to a metal like Zn, a metalloid like Ge, and then a nonmetal like Se, a nonmetal like Br, and a nonmetal like Kr. But I would think that the super-categories have merit, because if one were to put those six elements I just listed in very coarse boxes, {Ca, Cr, Zn} would easily fall in one box, {Se, Br, Kr} would more easily fall in another, and Ge would seem to be a toss-up.

Regarding "weak" metals, my idea was that the only totally strong metals are to be found in the first three groups of the periodic table, together with the lanthanides and actinides, and even then you have to take beryllium out. Then you have some metals that are physically strong but chemically weak: Be and Al fit here, but so do W and Au. I'm not too sure if this works as I was originally thinking though because things like Cd, In, and Sn are both physically and chemically weak (though not as much as Sb and Te of course), so I may have to go back to the drawing board on this point. ^_^ Double sharp (talk) 13:35, 31 March 2017 (UTC)

re the general lines metallic elements "get a little short shrift here" (Double sharp), and earlier "And most of the metallic properties of metalloids are physical properties, and physical properties are trumped by overall chemical behaviour" (Sandbh, with added enthousiasm). I can understand that physical properties are less important for these elements in this topic. And so 'semiconducting' should not be the namesake category. But to brush all physical properties aside and then call then 'nonmetals' is too much pars pro toto. After all, for one other category out of three being a metal fully accepts physical properties. Fixing it to 'nonmetals' somehow definitely forbids even considering metallic properties. It might even miss the target (of getting a good general category name), because of estranging people who by good intuition keep seeing some metal characteristics in there.

Maybe we'd be doing better by keeping 'metalloid' for this category, and add everywhere "metalloid (as specified by en:Metalloid)". Just narrowing and channeling its definition for general and promotional use, no scientific fields get hurt or abandoned. I also can support this one by Sandbh: "A category called "metalloid (weak nonmetal)" rather than "weak nonmetal (metalloid)" might work too". -DePiep (talk) 07:03, 3 April 2017 (UTC)

That last suggestion would look like this:

Alkali metal

Alkaline earth metal

Lan­thanide

Actinide

Transition metal

Post-​transition metal

Metalloid (weak nonmetal)

Intermediate nonmetal

Corrosive nonmetal

Noble gas

I think I prefer this one. Sandbh (talk) 21:23, 3 April 2017 (UTC)

The curious case of hydrogen

Continuing on from Double sharp's comments, I want to put my Chemistry for Dummies hat on. My question is: Does hydrogen actually lose an electron in any of its chemical reactions? When HCl is added to water, the positive charge on the H in HCl means its proton may get extracted by the O atom in a H₂O molelcule, to form H₃O⁺, with the result that the H atom's electron gets left behind with the Cl atom to give Cl^–. (Note my use of the word "extracted" rather than "transferred"). I don't count this as H losing its electron. In Introduction to advanced inorganic chemistry, Durrant and Durrant (1970, p. 404) say that, "H is a characteristically non-metallic element which shows no electropositive properties.[!] The H atom does not lose its electron in any chemical process…When the atom reacts it does so by acquiring an electron, either to form the anion H^–, or to form an electron pair which it shares with another atom." When hydrogen is used to reduce CuO to Cu the usual story is that H loses its electrons to Cu⁺⁺ to give Cu; the O^2– then combines with "H⁺" to give H₂O. Whereas according to Durrant and Durrant, and Ramírez-Torres (J. Chem. Ed. 1955, p. 452) what actually happens is that the O^2– loses its electrons to Cu⁺⁺ to give Cu, and the O atoms then combine with H₂, to give H₂O.

Most peculiar. Durrant and Durrant (p. 406) also observe that, "…the hydroxonium ion…[is] a complex of oxygen; it is not a hydrated proton, just as formaldehyde [CH₂O] is not hydrated carbon." I see our hydroxonium article in fact refers to it as an oxonium ion i.e. an oxygen cation.

The seemingly curious case of H is indeed giving me something to think and read more about. Sandbh (talk) 14:33, 5 April 2017 (UTC)

Here are a couple more quotes:

• "…even in combination with the most powerfully electronegative elements (i.e. F, O and N) hydrogen is covalently bound." (Wilson & Newall 1966, General and inorganic chemistry, p. 267)
• "Hydrogen atoms…do not react by loss of electrons to form simple H⁺ ions." (Sisler et al. 1967, College chemistry, p. 234)

-- Sandbh (talk) 07:31, 6 April 2017 (UTC)

Yes, but the fact that H⁺ is such a useful fiction that proton transfer is a good simplification for so many things makes me wonder if it might not be useful ground to stand on. Certainly the protons are never naked, but hydrogen seems to be behaving here more like a wannabe alkali metal than a wannabe halogen. The 1s shell is not very good at holding a negative charge, as can be seen from how squishy the hydride anion is.

If you look down group 1, you would almost expect hydrogen to be weird in it and be a super anomaly. Consider: for Li, the 1s shell gives great shielding for the last electron (hence the extremely high standard reduction potential), while for Na and below the p-shells give less good but still adequate shielding. But uniquely in the whole table, the sole electron of H in its tiny 1s shell has no shielding at all from the nucleus, and truly removing it would leave a proton, a "naked" charge with tremendous polarising power. So one would expect H to act like a "wannabe" alkali metal, acting as if it was losing its 1s electron and actually not doing so for real (but trying). And that is exactly whay we see, and that's why we put H on the far left of the table instead of the far right. It may be a nonmetal in appearance, but it seems to quite clearly want to be a metal, even if its natural makeup means it can't actually be one. Double sharp (talk) 06:54, 15 May 2017 (UTC)

P.S. To some extent this also follows the principle of "how high an oxidation state can you support as a true cation, without unloading your positive charge on others?" As we proceed across period 4, what starts on the left as basic hydroxides with KOH and Ca(OH)₂ ends with acids like Se(OH)₆, but the formula remains evocative. We are still forming hydroxides or oxyacids, but at some point they cross a border and turn into each other. We start by forming nice aquo complexes in low oxidation states for K⁺, Ca²⁺, and Sc³⁺; then hydrolysis brings itself to the fore, so that we get instead oxocations in TiO²⁺ and VO₂⁺. Then we lose cation formation entirely in VO₃⁻, CrO₄²⁻, and MnO₄⁻. Even though we are nowhere near the metalloid line this time, the high ionic charges needed cause hydrolysis for Ti^IV, V^V, Cr^VI, and Mn^VII just as surely as they do for Si^IV, P^V, S^VI, and Cl^VII.

Where is the difference between {Ti, V, Cr, Mn} and {Si, P, S, Cl} here in terms of chemical behaviour of their compounds? The one clear one I can see is not actually about metallicity; it is the difference between transition chemistry in the first set and main-group chemistry in the second set, as is shown by their common oxidation states in solution.

Perhaps we may then consider H^I as a special case where the naked proton is so small that even the +1 charge is not supportable, thus creating nonmetallic behaviour. This happens because the 1s electrons are in H and He the outermost electrons. Starting from Li they provide very good shielding indeed, so that it takes until B^III in the next row for the charge to become too high. After that the "first unsupportable charge" becomes +4 in periods 3 and 4, and rises to +5 in periods 5 and later (I need to check out group 14 chemistry better; I will probably cover Sn as well as Si and Ge in that planned exposé here). Double sharp (talk) 04:27, 12 June 2017 (UTC)

Hmm, apparently only hydrolysed species exist even for Zr and Hf, not to mention Sn and Pb, so that only for Ce and the actinides are true +4 ions achieveable – and even then they are very readily hydrolysed. Double sharp (talk) 10:23, 12 June 2017 (UTC)

Sampling the literature

Out of curiosity I visited a university library and looked up the indexes of 100 chemistry textbooks to see how many had entries for the terms metalloid/s or semimetal/s. The count was 48 did and 52 didn't. But there are some more parts to this story. I didn't check for the term semiconductors/s which some authors use to erroneously refer to metalloids. And while some indexes did not list the key terms, I know that at least a few of the books in question nevertheless used the terms at least early on in the books when they talk about metals and nonmetals. I looked again at a some of these books to see how they described the group 15 elements and their chemistry. The count here was no better. It was pot luck as to whether any of these elements were described as metalloids, even when the book might've referred to metalloids early on in the introductory writings.

My general impression is that while some authors use the terms metalloid or semimetal, some don't. For the authors that do use the terms metalloid of semimetal, they rarely use them in any meaningful way when describing the chemistry of the elements involved. In other words, the metal-nonmetal dichotomy, at least when it comes to descriptive chemistry, appears to be rather deeply entrenched.

I'm not sure, but I suspect the terms metalloid or semimetal may be more popular in contemporary texts. This may be because fuzzy concepts like metalloids, however ill-conceived, seem to be more fashionable now than more black and white thinking. Sandbh (talk) 08:56, 3 April 2017 (UTC)

I think your mentioning of "the chemistry" is the reason why it is not stated, given that it is rather common knowledge (even if only intuitively so) that metalloids have metallic physical properties and nonmetallic chemical properties.

On the other hand, I think we also give some "short shrift" here to the more metallic chemical properties. As might be similar to P in many aspects, but even Na₃As is an intermetallic compound. Actually As and Sb react like metals with stronger nonmetals like O and the halogens, while Sb reacts with concentrated oxidising acids to give salts (unlike As) like Sb₂O₅, SbCl₅, and Sb₂(SO₄)₃. Going a column further, Te dissolves in dilute HCl and Po does so quite enthusiastically; though Te is admittedly much more on the "nonmetallic" side than Sb, tellurides are intermetallic compounds for the most part, so it kind of feels like As. I do think antimony might serve as a reasonable case study for this. There is a tendency towards more and more nonmetallic behaviour as the oxidation state climbs higher: even Sn(IV) feels "nonmetallic" in a way that Sn(II) doesn't, and bringing Ge(II) into the picture of germanium makes its classification as a metalloid seem much sounder. Double sharp (talk) 09:36, 3 April 2017 (UTC)

Hmm. I wonder if it is really common knowledge that metalloids generally have nonmetallic chemical properties. On metalloid "salts" I will always remember Axiosaurus' views about this, here. "Salt-like" compounds might be safer terminology. But certainly, metalloids can form alloys, and fall within the scope of organometallic chemistry. Sandbh (talk) 11:28, 3 April 2017 (UTC)

And see antimony sulfate. Sandbh (talk) 11:41, 3 April 2017 (UTC)

They are not really "salts", but you could perhaps call them "generalised salts", because they form by the same sorts of reactions (Sb does react with sulfuric acid to give this, it's just that it's not really a sulfate). I think the alloying thing as well as distinct organometallic chemistry might be the best argument for metallic chemical properties in the metalloids, but then tellurium and astatine are shaky because they seem to act quite close to selenium and iodine in their organic chemistry. But tellurium fits the alloying criterion, and as for astatine, I was surprised to see Ag–At and Au–At listed in Phase Equilibria, Crystallographic and Thermodynamic Data of Binary Alloys – that actually gives a wonderful explanation for why silver precipitates astatine only partially (unlike for the lighter halogens), if the astatine alloys with the silver! ^_^ Double sharp (talk) 12:06, 3 April 2017 (UTC)

Tellurium and astatine should be OK from the organometallic perspective. Rochow talks about the organometallic chemistry of Te in The Metalloids (1966, pp. 86–87). Or see this: Tellurium: Organic and Organometallic Compounds. There is a reference to organometallic At compounds in Gmelin. And I have seen an article exploring the potential medical applications of cationic At, which might also count, at least in part. The remarkable references to At alloys with Ag and Au, sure point in this direction. Sandbh (talk) 22:07, 3 April 2017 (UTC)

It occurs to me though that both criteria seem to strike out for C, as it famously forms alloys like steel, and "organocarbon chemistry" is obviously a silly and very ill-defined term. Double sharp (talk) 03:13, 4 April 2017 (UTC)

What if we start from category definition (not from element properties)

• The difficult part is to fomulate the question right, so I'll build it up. Sandbh proposeds to name this category "weak nonmetal (metalloid)", because it 'captures the complexity of the situation with a high degree of consensus'. Now if we accept this new better, more descriptive(!) name, the question is: Would this name change the membership of the category?

Note that this is a different approach, namely from the category definition, not from individual element's properties. Most if not all of above arguments depart from an individual element's properties (set). While, when we have a better category definition, a membership test can and shall be performed from that definition.

Examples. Looking at the arbitrary dividing line between metals and nonmetals (very arbitrary of course, if only because it reduces the number of categories from three to two; not unlike the current proposal btw), it is likely that Ge and Sb are considered metals, and so will not be nonmetals (of whichever subspecification). So, if the new name is better and descriptive, the membership could change? (A curiosity then to be explained is, that these two elements score safely high on the 192 sources list). A contrary example I do not have, or maybe At while likely being an current-name 'metalloid' does not qualify as a nonmetal? -DePiep (talk) 15:50, 26 March 2017 (UTC)

These are insightful questions, thank you DiPiep. I hadn't previously considered the subject matter they raise.

My response to the first question, "Would this name change the membership of the category?" is a firm, but calm, "No." I say this because in developing the proposal I focussed on the elements and their individual and shared properties. And I did not finalise the categories until I was sure they would stand up without the need for membership changes.

A related consideration is that I don't know of a comprehensive, literature-based objective way it could be done. Any attempt to do so would surely fail because there are no other goal posts as credible as the metalloid ones.

The dividing line between metals and nonmetals is only OK as a rough guide.

If applied literally it results in germanium and antimony being regarded as metals.

Germanium is a brittle semiconductor with a mostly nonmetallic chemistry. All other semiconducting elements (B, C perpendicular to its planes, Si, black P, Se, Te, and I in the direction of its planes) are metalloids or nonmetals.

Antimony is a brittle semimetal with a mostly nonmetallic chemistry. Here, "semimetal" means it lacks the electronic band structure of genuine metal. You can get more of an idea of how "metallically-challenged" antimony is by comparing it to bismuth, its heavier, more metallic congener. Bismuth has been described as physically the weakest metal in the periodic table, and chemically as a "very weak" metal. Whereas bismuth is scraping the bottom of the metal envelope, it too being a "semimetal" in the physics-based sense, antimony has fallen though the envelope. Information sources in metallic materials (Patten 1989, p. 192) says antimony is usually classified as a nonmetal or metalloid. (The name of this source is ironic; double pun not intended. ^_^)

Lithium, beryllium, aluminium and polonium, which are the other four elements below the dividing line, are metals, although the nonmetallic properties of the last three sometimes results in one or more of them being regarded as metalloids.

On the other side of the line, we currently classify astatine as a metalloid on the basis of its known trace chemistry (which may not be representative of its bulk chemistry). It may in fact turn out to be a ductile post-transition metal, with the electronic band structure of a genuine metal.

Individual authors nevertheless need to make their own decisions, on a case-by-case basis, as to the metallic or nonmetallic status of elements in the vicinity of the arbitrary dividing line, as we know they do. Sandbh (talk) 06:47, 28 March 2017 (UTC)

By now, I get this impression. You are proposing a better name for the ubiquitous (omnipresent) word 'metalloids', thereby being more specific on properties in a certain way. The categories characteristics and membership should not change. One main point is that the property 'semiconducting' should be given less importance, as (a) it is well more narrowly defined within physical uses (and therefrom has quite a different membership list, of course very useful & meaningful in that field); and (b) it disturbes the category ('metalloids would be so much more homogeneous if we'd leave that one out completely'). -DePiep (talk) 09:51, 28 March 2017 (UTC)

Re the name change and membership, effectively yes. The semiconducting aspect is currently given less prominence in the metalloid article. The properties section explains that metalloids are either semiconductors or semimetals, and that in the case of As and Sb (which are semimetals) that these also exist in less stable semiconducting forms. We cannot leave this out completely as the semiconducting properties of silicon and germanium enabled the establishment of the semiconductor industry in the 1950s, the development of solid-state electronics from the early 1960s, and the associated exploration of the semiconducting properties of the other metalloids and nonmetals. I guess it is understandable that some people reckon metalloids are semiconductors, period, without appreciating that this is not quite right. Sandbh (talk) 11:56, 28 March 2017 (UTC)

Leftover nonmetals categorising

• Leftover nonmetals categorising. This is more problematic, in that it does not succeed in making a (YBG-)convincing subclassification. It could be an improvement in that at least one class (into corrosion) has a wiki article compared to zero for both of todays categories. The magic thread Sandbh points out (I'd call a magic chain) links the elements by pairs only. This breaks an other ground rule for good classification: class elements should all be strongly tied within the category, and badly tied to outside the category. The chain does not do this (no. 3 is not tied to no. 1).

Maybe this is happening: if sub-categorisation of the leftover nonmetals is that difficult, absent or sought after, there probably is no subcategorisation. Keep them one category? Of course, all descriptions and notes made in the reasoning still are valid, and should have a place in the article leftover nonmetals (working title). -DePiep (talk) 12:07, 12 March 2017 (UTC)

Looking at all the elements that are not metals, there are three categories that almost populate themselves; noble gas; corrosive nonmetal; and weak nonmetal (metalloid). We know that the corrosive nonmetals are strong nonmetals and that the weak nonmetals (metalloids) are weak nonmetals. The remaining nonmetals are neither as strongly nonmetallic as the corrosive nonmetals nor as weakly nonmetallic as the weak nonmetals. They are in the nonmetal goldilocks zone—not too chemically hot, not too chemically weak. Effectively, they are intermediate nonmetals.

The magic chain of the intermediate nonmetals is not the primary thing that links them. What links them is their temperate or moderate nature, compared to the corrosive nonmetals and the weak nonmetals. Sure, rather than reading the chain as H → C → P → N → S → Se, you could read it as H → C; C → P; P → N; N → S; S → Se. I think this is cutting off your nose to spite your face. Even across groups 1 to 18, only the elements in groups 1, 3, 17 and 18 are strongly tied together, and group 1 becomes a little wobbly if it includes H. There are more groups that are not tied together strongly than there are groups that are tied together strongly. But I don't think this matters since organising the groups according to their valence states is what revealed the deeper relationships among the elements in the first place, including the diagonal relationships and patterns which become particularly prominent in the p block.

Yes, it is hard to subcategorise the leftover nonmetals i.e. those that are not noble, corrosive, nor weak (metalloid). There are only a few names that have been given to them in the literature and none of these are widely used. That does not mean we necessarily need to adopt an unfortunate category name like other nonmetal, when—consistent with the literature—I like to think that there are more generic, helpful, descriptive and preferably plain English words that we could try. I may be deluded though given how hard it seems to be to come up with something better than "other".

The proposed scheme fine-tunes our nonmetal categories in way that is more consistent with the literature, and retains and enhances the work of the founders of the Wikipedia periodic table colour categories. (The first scheme, from February 2002, was based on the LANL table, and the Environmental Chemistry table, here.) Sandbh (talk) 04:32, 20 March 2017 (UTC)

Threading intermediate nonmetals

The trouble is that I do not see why the thread should stop there. If it can go from S to Se, why not from Se to Te? They're not all that different chemically. Simply put, there does not seem to be a very clear demarcation as to why the thread should contain some elements and not the others, that isn't rationalised from already having selected which elements to exclude from the other categories and put in the thread. Double sharp (talk) 04:16, 21 March 2017 (UTC)

The thread can be traced back as far as H → Li → Mg → Be → Al → B → Si → C → etc. It alternates between diagonal links (which are often forgotten) and vertical links, and since going onto Te will result in two vertical links, S is a natural place to stop. (It did not occur to me at the time to keep going onto Te, since S was an intermediate nonmetal and Te was a weak nonmetal). The fact that there is this thread that runs through all of the intermediate nonmetals wasn't something I realised until later. The natural part—like falling off a log when compared to the poly/di-atomic marathon—was dividing the nonmetals into the corrosive nonmetals (they speak for themselves) and the weakly nonmetallic metalloids (they speak for themselves, too). All that was left after that were the goldilocks nonmetals most of which have been labeled with the "it's-too-hard-to-work-out-what-to-call-them" category name of "other nonmetals". I didn't realise that they happened to be linked by diagonal and vertical links until I thought about what else they had in common, aside from being intermediate nonmetals. If I have made too much of a song and dance about the magic thread, I can only reiterate that the main game is about the contrast between the weak nonmetals (metalloids) and the corrosive nonmetals, and the remaining bystander or sandwich nonmetals (relatively speaking). Sandbh (talk) 13:13, 21 March 2017 (UTC)

Diagonal links

Okay, but surely not all the diagonal links are relevant: surely no one thinks that fluorine has anything to do with argon. I don't think the diagonal links are often forgotten: the ones between Li and Mg, as well as Be and Al, are really quite well-known. What is forgotten is that the analogy may be pushed further (though there are limits; comparing Ce and Pa is rather nonsensical).

More to the point, the trouble I find is that the distinction is very fuzzy, and there is nothing like the precipitous drop in metallic behaviour that occurs when you cross past group 11 (there is another drop between very strong metals and milldy strong metals past group 3 and the Ln and An). There is a reason I raised selenium and tellurium: we split them into two separate categories, even though they are very similar chemically and are in fact very often found together: we shouldn't forget that selenium was named after the moon because it resembled tellurium, which was discovered earlier and was named after the Earth! They even form a continuous range of solid-state solutions in which Se and Te atoms alternate in the helical chains (since hexagonal Se resembles the main modification of Te in its structure).

So why do we put selenium apart from tellurium in the first place? Because Se has a number of properties that are somewhat uncharacteristic of a nonmetal which are only brought to maturity in the next member of the group. But already Se shows an advance on S in incipient metallic properties, so that if one is willing to call S–Se a valid link, then disallowing Se–Te immediately after that seems to show symmetry being taken too far (after all, S forms S₈ molecules while Se and Te form long chains). S–Se–Te is quite a good triad for this "no man's land" between clearly metallic and clearly nonmetallic territory and it does not sit well with me to get rid of it for reasons like this. It is fine if we are applying some artificial sharpness in our categorisation by going for a hard line based on how many atoms happen to be in a molecule of the pure element (never mind that S and I seem a little out of place), but with fuzzier categories like this I am slightly more ill at ease. Double sharp (talk) 14:09, 21 March 2017 (UTC)

Fuzzy v sharp; diagonals

Indeed, there is no diagonal link between fluorine and argon and I did not allude to this. In my experience, the diagonal links are ignored---certainly in high school where the focus is on vertical similarities. Maybe things have changed since I went to HS; I tend to doubt it.

On the fuzziness of the proposed categories, and I hope I am not belabouring things, I did not make the distinction between (a) corrosive nonmetals and (b) metalloids, and in consequence of those extremes, the leftover nonmetals. The first two "categories" are the way they elements involved have been described in the literature whereas the leftover nonmetals don't attract such language. In this case I tend to think that the wisdom of the masses and its delineation of the nonmetals into "naturally sharp" categories is as at least as good as the artificial sharpness in our hard line categorisation based on how many atoms happen to be in a molecule of the pure element.

As noted, extending the magic thread to Te, breaks the alternating diagonal-vertical sequence. I'm not sure what joining a diagonal-vertical-repeat thread (an orange) onto a group-based vertical-repeat thread (an apple) would achieve (apart from causing me to raise an eyebrow).

We currently call Se a polyatomic nonmetal and Te a metalloid even though Te is a polyatomic "not-metal", because the literature tells us that Te is counted as a metalloid 4 times more frequently than is the case with Se (98% v 24% according to lists of metalloids). The fact that Se happened to be polyatomic was a fortuitous bonus that enabled us to keep it grouped with its other polyatomic colleagues. There is no distinction between Se and Te in an atomic structure sense, since both are polyatomic.

Nothing essential changes with the proposed scheme. Te becomes a weak nonmetal (metalloid) and Se becomes intermediate, consistent with distinctions made in the literature. Sure S forms rings whereas Se forms chains but S does form long chains in plastic sulfur. The S-Se-Te triad is still there, and its natural category division is in the same place. The fact that there is a cross-cutting thread fortuitously and curiously running through the intermediate nonmetals is something I happened to notice afterwards. (And the presence of the thread does not mean there is no vertical link between Se and Te), just like the polyatomic-diatomic line between the same two elements doesn't.) Sandbh (talk) 04:47, 22 March 2017 (UTC)

Postscript: A possible alternative way of mentioning or considering the other relationships among the intermediate nonmetals is to merely observe H → C; C → P; and N → S, and not say anything about P → N and S → Se, since the latter two relationships are already part of the periodic table furniture. Sandbh (talk) 06:51, 22 March 2017 (UTC)

Diagonals; threading; H & N

Okay, but surely most of the diagonal relationships come with the periodic table furniture as well: they are one of those things that you learn as you read the thing, sometimes to the point that you intentionally break some things to show some points. I have sometimes naturally wanted to talk about aluminium as if it were above scandium instead of above gallium, for example (and when R8R Gtrs writes the Al article I am looking forward to reading about that similarity). We also know of the linkages between groups n and groups n + 10 that used to be considered A and B subgroups; the "knight's move" relationships stretching across groups 11 to 13; the way the early actinides "pretend" to be transition metals; and so forth. Essentially, you start with a first-order rationalisation, and then keep adding adjustments to it, like some chemical analogy for a Taylor series. Just because not all of this is taught in high school (although I remember that Li–Mg and Be–Al were emphasised, and B–Si was mentioned but not treated in so much detail) does not mean that they are ignored totally: it just means that we're not ready for it then, but might be later.

Certainly the extremes exist, and similarities between the elements that are not clearly in the extremes also exist, but because there is a commonality leading through all of them as you pass through adjacent cells on the periodic table I am not so sure that the "magic thread" is particularly convincing as an argument. It actually feels more convincing to me to say that these elements are just somewhere in the middle, without appealing to the thread. So I see the only difference between the current and the proposed classifications is where we put hydrogen and nitrogen, right? I would note though that both are actually vigorously reactive, except that you have to heat them up first, whereas sulfur does not even need the heating: whereas diamond is extremely unreactive and graphite needs rather violent temperatures to react even with fluorine. Meanwhile black phosphorus is not very reactive at all (the element's reputation comes mostly from the white allotrope) and is even a semiconductor. The difference is mostly whether you can find simple anions as a matter of course, but even then, At would form At⁻, and with the reactive metals even Po forms "quasi-salts" like K₂Po. Double sharp (talk) 09:50, 22 March 2017 (UTC)

I think we are in agreement on diagonal relationships, and adjusting from first order rationalisations as one goes. I do not claim the magic thread as a supporting argument, only as a nice to observe. You refer to the same thing that occurred to me which is that the extremes in nonmetal behaviour are already evident in the literature, and that naturally leaves the rest in the middle. (I don't know why this didn't occur to me when we were doing option 10 etc---in retrospect it was staring me in the face all along).

Yep, H and N are the only movers except that, by default and without doing anything, they end up in the middle. The decision outcome as to which nonmetallic element goes into which category is a composite of the many factors that contribute to, or are aspects of, a nonmetallic element's nonmetallic character, including some that you mention such as reactivity, and anion forming capacity. In the case of the corrosive nonmetals, it happens that their corrosive nature correlates quite well with their overall nonmetallic character. The work on assessing all of the factors that contribute to the nonmetallic-fu of the rest of the nonmetallic elements has already been done via a wisdom of the masses consensus in distinguishing between the metalloids and the remaining nonmetals. (I have a reservation about At as a metalloid, as mentioned, but there you go). Sandbh (talk) 04:14, 23 March 2017 (UTC)

Astatine

The reason why I still think At deserves to be called a metalloid, even if it might form cationic At⁺ in solution, is because it is also homologous with iodine, which is important for its medical use. I should think that this kind of paradoxical, mixed chemistry of a combination of the halogen and the metal should deserve the name "metalloid". After all, we get halogen-like bonds to organic molecules like PhAt, that oxidises like iodine compounds to PhAtCl₂, and the usual species like HAtO₃ and H₅AtO₆ that remind me very much of iodine. I don't think cationic chemistry is enough to prove it as a metal; look at tungsten, for example. Double sharp (talk) 06:57, 24 March 2017 (UTC)

(Please excuse me by the way for not mentioning anything about H and N in this reply. There are two main reasons: the less important one is that I am still pondering what you have said here, and the more important one is that I do not wish to overwhelm you with so many long replies at once. ^_^) Double sharp (talk) 15:30, 24 March 2017 (UTC)

Thank you!

Regarding astatine, I think that's somewhat of a sideshow at least for now. I guess I can only blame myself since if I'm going to raise it I ought to do it well. But then I though it worth mentioning in passing.

Perhaps the key things to say about it might be (1) the recent prediction that the condensed form would be expected to be a (ductile) full blown close-packed metal; (2) the At^– ion is relatively easy to oxidise back to At (0) and there is some speculation that +1 might be the most stable state; (3) the extremely low concentrations at which its chemistry has been observed, and the possibility of reactions with all manner of contaminants (impurities, walls and filters, radioactivity by-products) and other unwanted nano-scale interactions; (4) “since the trace chemistry of I sometimes differs significantly from its own macroscopic chemistry, analogies drawn between At and I are likely to be questionable, at best" (I recall that's attributed to Kirby, writing in the Gmelin handbook on astatine); and (5) as a putative p block metal it could be expected to show significant nonmetallic character. (I caveat all of this lot by saying I haven't looked at the more recent literature on astatine.) Sandbh (talk) 11:26, 25 March 2017 (UTC)

The way I would see it is that while our picture of the chemistry of astatine is definitely fuzzy, it is the best we have at the moment: we shouldn't treat the theoretical predictions as being on a par with experimental results until they are verified (we're not giving astatine a "predicted" colour). So far, the properties known (admittedly on trace quantities) tally reasonably with that of a weird cross between a metal and a halogen, so calling it a metalloid seems reasonable given the current state of limited knowledge. When better data comes in, we can of course revisit this, like we did for polonium. All right, now off to formulate a reply to the bigger block of text. ^_^ Double sharp (talk) 12:08, 25 March 2017 (UTC)

Yes, no worries about astatine. I suspect it needn't impact on the merits of the proposal. It'll be ironic if the 1940 decision of Corson and his colleagues (the discovers of astatine) to classify it as a metal on the basis of its analytical chemistry, turns out to be true. Some 77 years later and we still don't know! That doesn't seem right to me but I have too many other things to do right now. Sandbh (talk) 01:49, 27 March 2017 (UTC)

Radon

Standard reduction potentials of cation-forming elements (Z ≤ 103)

1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
H																	He
Li	Be											B	C	N	O	F	Ne
Na	Mg											Al	Si	P	S	Cl	Ar
K	Ca	Sc	Ti	V	Cr	Mn	Fe	Co	Ni	Cu	Zn	Ga	Ge	As	Se	Br	Kr
Rb	Sr	Y	Zr	Nb	Mo	Tc	Ru	Rh	Pd	Ag	Cd	In	Sn	Sb	Te	I	Xe
Cs	Ba	La	Hf	Ta	W	Re	Os	Ir	Pt	Au	Hg	Tl	Pb	Bi	Po	At	Rn
Fr	Ra	Ac
			Ce	Pr	Nd	Pm	Sm	Eu	Gd	Tb	Dy	Ho	Er	Tm	Yb	Lu
			Th	Pa	U	Np	Pu	Am	Cm	Bk	Cf	Es	Fm	Md	No	Lr
Highly reactive: E0 ≤ –2.0 V   Less reactive: –2.0 < E0 ≤ 0   Noble (unreactive): E0 > 0   No simple cationic chemistry

I was reading a nice 1983 article on radon chemistry (10.1524/ract.1983.32.13.163) and it seems to put a finger on what's not so ideal about the criterion of cationic formation:

"Stable solutions of radon fluoride [RnF₂] have been prepared [through reactions of Rn gas with some famously oxidising interhalogens]...Radon has very unusual properties in these solutions. When elemental radon dissolves in an aqueous or organic solvent, it has a measurable partial pressure above the solution; it is distributed between the gas phase and liquid phase and can be easily removed by pumping on the solution. However, oxidized radon in halogen fluoride solutions is completely involatile. The solutions can be exposed to the atmosphere, poured from one container to another and vacuum distilled without loss of radon. When they are distilled to dryness, residues of radon fluorides are obtained. These residues may consist of RnF₂ or complexes of RnF₂ with halogen fluorides.

Electromigration studies, carried out with the cell shown in Fig. 3, have shown that radon is present in many of these solutions as a cation [63, 73]. In a pure conducting solvent, such as bromine trifluoride, radon fluoride is believed to ionize as follows:

RnF₂ ⇌ RnF⁺ + F⁻

RnF⁺ ⇌ Rn²⁺ + F⁻

In a solution containing added KF or CsF electrolyte, however, dissociation may be suppressed, and radon may even form anionic complexes:

RnF₂ + F⁻ ⇌ RnF⁻
₃

RnF⁻
₃ + F⁻ ⇌ RnF²⁻
₄"

All of this reminds me and the author uncannily of aluminium and especially beryllium, and would indeed create a nonmetal with a markedly metallic chemistry! Another interesting fact is that Rn appears to be reluctant to form oxyanions like Xe does, but this may be more due to an absence of evidence than evidence of absence. In light of this I have added At and Rn in colour to the pretty picture (reproduced at right), and I humbly submit that cation formation should be only one of many criteria, since no one wants to call radon a metal. It doesn't look like one as a solid, does it? Double sharp (talk) 16:33, 4 April 2017 (UTC)

BTW there are also some claims (see the radon article) that a higher radon fluoride (RnF₄ or RnF₆??) exists and coprecipitates with CsXeO₃F as what is presumably CsRnO₃F, strongly suggesting that hydrolysis to RnO₃ has occurred. Some 1989 studies suggest an emerging picture of Rn(VI) along the lines of Xe(VI), forming [HRnO₃]⁺ and the radate anion [HRnO₄]⁻. So Rn(II) appears to be like Be(II) and Rn(VI) like Xe(VI). One Soviet popular science book spoke of astatine in these terms: "The heaviest halogen, iodine's elder brother. It could be a great surprise to chemists, for it was not impossible that its properties would be weakly metallic. Halogen and metal what a splendid example of a two-faced element!" I think I shall bow to them and write the same of radon: noble gas and metal! And it makes me wonder what the chemistries of Ts and Og would look like; I think that for these we not only need longer-lived isotopes, but also a better understanding of At and Rn as the sixth period runs into the void. Double sharp (talk) 16:41, 4 April 2017 (UTC)

Just a little reminder to myself to include a mention of this cationic chemistry in the radon article. (I should also go look for predictions for cationic oganesson!) Double sharp (talk) 08:09, 9 May 2017 (UTC)

 Done Double sharp (talk) 15:23, 10 May 2017 (UTC)

Classification criteria for metals etc

It is with some hesitation that I post this since, in my view, the metalloid category is not necessarily worth elevating to the same level as metals and nonmetals; consequently I feel like I'm undermining my own proposal. Be that as it may, as Double sharp has expressed an interest in the topic of classifying the elements, here it is (I posted the first part, in a slightly different form, earlier in this thread):

The first 100 elements can be satisfactorily classified as metals, metalloids, or nonmetals, using a combination of five quantitative and semi-quantitative properties.

1. A metal is an element that meets at least two of the following criteria:

    a. low ionization energy (< 750 kJ)
    b. low electronegativity (< 1.9)
    c. metallic or semimetallic band structure
    d. high packing efficiency (> 41%)
    e. high Goldhammer-Herzfeld metallicity ratio (> 1.1) Most metals (65 out of 76) meet at least four of these criteria.

2. A metalloid is an element with the following properties:

    a. medium ionization energy (750–1,000 kJ)
    b. medium electronegativity (1.9–2.2)
    c. semimetal or semiconductor band structure
    d. medium packing efficiency (34–41%)
    e. medium metallicity ratio (~0.85 to 1.1)

The six elements that meet all of these criteria (B, Si, Ge, As, Sb, Te) correspond to the six elements commonly classified in the literature as metalloids.

3. A nonmetal is an element that meets at least two of the following criteria:

    a. high ionization energy (> 1,000 kJ)
    b. high electronegativity (> 2.2)
    c. semiconductor or insulator band structure
    d. low packing efficiency (< 34%)
    e. low metallicity ratio (< ~0.85)

Most nonmetals (14 out of 17) meet at least four of these criteria.

Notes

• There is no widely agreed or even contested definition that sets out what is a metal, what is a metalloid, and what is a nonmetal.
• In the literature the elements tend to be collectively characterized in terms of generalities or a few broadly indicative physical or chemical properties, with a single quantitative criterion or attribute (such as electrical conductivity, or the acid−base character of group oxides) being mentioned only occasionally.
• Astatine remains a mystery since we don't know its band structure, nor its crystalline structure, which would determine its packing efficiency and shed light on its metallicity ratio. The most rigorous theoretical (and relativistically-based) calculations to date suggest it will have the band structure of a metal, and have a close-packed crystalline structure; by analogy with high-pressure iodine it might even be a superconductor.

--- Sandbh (talk) 01:22, 5 April 2017 (UTC) @YBG: for when you get back. The above definitions are slightly different from what I sent you (see 1c. and 2c.). Sandbh (talk) 14:40, 5 April 2017 (UTC)

Here is a simpler definition, which comes in successively streamlined versions:

I. A metal is an element that has a lustrous appearance when freshly prepared or fractured and (a) has a densely packed crystalline structure;¹ or (b) forms a simple cation in aqueous solution;² or (c) has a basic oxide.

II. A metal is an element that has a densely packed crystalline structure;¹ and (a) forms a simple cation in aqueous solution;² or (b) has a basic oxide.

III. A metal is an element that has a densely packed crystalline structure¹ or forms a simple cation in aqueous solution.²

All other elements are nonmetals.

Footnotes:

¹ Hexagonal-close packed, face-centred cubic, α-lanthanum, α-samarium, body-centred tetragonal, or body-centred cubic

² Including aqua-cations such as [Bi(OH₂)₈]³⁺

The potential "problem" elements are those with a lustrous or semi-lustrous appearance and that have more open-packed crystalline structures: Mn, Ga, In, Sn, Hg, Bi, Po, U, Np, and Pu. However all of these form one or more (stable) simple cations in aqueous solution, so the problem goes away i.e. they are all counted as metals. The other elements having a metallic or semi-metallic appearance are B, C, Si, P, Ge, As, Se, Sb, Te, and iodine, however they all have more open-packed crystalline structures, and none form a simple cation in aqueous solution, nor do any of them have basic (as opposed to amphoteric or acidic) oxides i.e. they are non-metals.

I'll now spend the rest of the day wondering if there are flaws in any of these definitions. Usually this happens five minutes after I press the "save changes" button. Sandbh (talk) 04:29, 9 April 2017 (UTC)

I'll note that radon is a problem with III, being fcc and forming Rn²⁺ (aq). ^_^ Presumably this is why the appearance has to come into it. Double sharp (talk) 07:18, 9 April 2017 (UTC)

I had intended for the definitions to apply at room conditions. In specifying crystalline structure I was thinking more about packing efficiency rather than crystalline structure per se. So Rn should fail as it has a minute packing efficiency but, of course, the way I've worded the definitions means it won't work that way. Copernicium might cause some difficulties too, if it turns out to be gaseous. Hmm, where's my thinking hat? Sandbh (talk) 11:57, 9 April 2017 (UTC)

I added the word "solidified" to definition I, so that should work now:

"I. A metal is an element that has a lustrous appearance when freshly prepared, solidified, or fractured and (a) has a densely packed crystalline structure;¹ or (b) forms a simple cation in aqueous solution;² or (c) has a basic oxide." -- Sandbh (talk) 13:34, 9 April 2017 (UTC)

So this would mean that the "chemically strong" pre-transition metals in the s- and f-blocks would meet all the criteria; the not-so-strong ones in the d- and p-blocks (together with "pseudo-d" elements in the early actinides) meet the first two; and the metalloids only meet the first one.

The status of the elements At and Ts where the metalloid line dives into a "sea of instability" is tricky, since astatine is so poorly known and tennessine is not characterised at all. (Oganesson, at least, should not be able to metallise, though it would be a spectacularly bad noble gas.) The crux for both of them would be whether or not they are metallic in bulk quantities; At might be (but the precise result depends a lot on how you do the calculations), so Ts almost certainly is from group trends (relativistic effects seem to push the "metalloid line" further to the right). So tennessine is probably safe to call a "post-transition metal". Whether astatine qualifies as a metal or not seems to rest entirely on whether or not it metallises at standard conditions. The "monatomic and metallic paper" notes that scalar-relativistic DFT calculations suggest astatine to have a structure like iodine, remaining nonmetallic and diatomic to 9 GPa (compare iodine at 16 GPa). But once spin-orbit effects are included the metallic state becomes the ground state at atmospheric pressure! While this is suggestive, the sudden change suggests to me that we should exercise caution and conservatively mark At as a metalloid (and Ts as a predicted post-transition metal) until better data comes in.

I am a bit worried about copernicium too. Not only is it expected to be very noble, Cn is even expected to have a low packing efficiency like the noble gases. The same sort of thing might happen for the quasi-closed shell at Fl. Copernicium might be a semiconductor or insulator (ref); if it is the latter, it might not even be lustrous as a solid. How problematic! Double sharp (talk) 15:34, 9 April 2017 (UTC)

Then again, perhaps this is not a failure of your criteria; if we made a bulk sample of copernicium, and it turned out to be an insulating gas, condensing to a noble-gas-like monatomic solid, I would start questioning if this element had better not be moved out of the group of metals. But then again, it would seem to act like a normal eka-mercury once you forced it to start reacting, so I am not sure about this either. Since we're not likely to get bulk quantities of Cn without exploding multiple nuclear bombs to exploit the r-process to its fullest, we can probably sweep this problem under the rug for now. Double sharp (talk) 15:46, 9 April 2017 (UTC)

• I love this approach. Allow me to note that our categorisation (into colors) is always a summing of those judgements. That is: per criterion there is a ~classification, and after that we do summarise (accumulate) those classifications. That's a lot of grey/uncertain areas to explain. -DePiep (talk) 21:26, 9 April 2017 (UTC)

For DePiep. Thank you, your observation is astute. The chemist inside you is doing you proud! Sandbh (talk) 13:11, 10 April 2017 (UTC)

This time for such a compliment I wont just click 'thank'. I type: Thanks! This enwiki elements cloud is very inspiring.

I tried to say: our category colors are a stack of two categorisations. By single criterium first: eg "semimetal or semiconductor band structure" (with its own grey area). Next step: gathering those classes, and their disputed areas. Great, and also says why the periodic table is not that definitive. Also I keep supporting the YBG/R8R approach, that we want complete categorisation (and coloring). That is: unless one goes into detail, the periodic table should give each element a single position. Or: precious metals is a nice grouping, but not periodic-table-complete.

About my "chemist inside": actually, there is none AFAIK. I am winning those bonus points here because (trick revealed) my mind does classification, while you do individual elements & their patterns. Md did this by himself. -DePiep (talk) 20:36, 10 April 2017 (UTC)

For Double sharp. Yes, the match up is reasonably good. All metals meet criterion 1. The s- and f-block metals mostly meet 2, 3 and 4. The d-block metals generally meet 2 or 3. The p-block metals tend to meet only 3 or 4 (the ones that also meet 2, like Pb, do so abnormally). As DePiep implied, it is the small number of anomalies (Mn is another) that need explaining. The rest of the elements having a lustrous or semi-lustrous appearance are B, C, Si, P, Ge, As, Se, Sb, Te, and I. These can be divided into those having amphoteric or weakly acidic oxides i.e. B, Si, Ge, As, Sb, Te (sometimes called metalloids), and those with more strongly acidic oxides i.e. C, P, Se, and I. And then all that's left are H, N, O, F, S, Cl, Br, the noble gases (and At).

The definition can also serve to establish electrical conductivity benchmarks. Metals have conductivities of ≥ 10² S·cm⁻¹ (based on grey tin's figure of 2–5 × 10² S·cm⁻¹); the semiconductor range starts with Se at around ~10⁻⁹ to 10⁻¹² S•cm⁻¹; and anything less than that of Se is an insulator.

I worried too much about Cn. I don't need to add the word "solidification" since when we refer to the crystalline form of an element, we mean in its solid state. So the definition reverts back to:

A metal is an element that has a lustrous appearance when freshly prepared or fractured and (a) has a densely-packed crystalline structure;1 or (b) forms a simple cation in aqueous solution;2 or (c) has a basic oxide.

   1 Hexagonal-close packed, face-centred cubic, α-lanthanum, α-samarium, body-centred tetragonal, or body-centred cubic
   2 Including aqua-cations such as [Bi(OH2)8]3+

If Cn turns out to be a semiconductor or insulator I somewhat doubt it would behave chemically like a metal, since its electrons would presumably be localised rather than delocalised, and not available for transfer. Sandbh (talk) 13:11, 10 April 2017 (UTC)

Well, given that Cn is the relativistic maximum in period 7 just like Au is in period 6, I'm prepared for almost anything to show up! ^_^ I don't see why a nonmetallic Cn could not form simple cations though: At and Rn seem to be able to do this, and CnSe preliminarily seems to be a fairly normal group 12 selenide. The band gap is also expected to be small (0.2 eV). Double sharp (talk) 23:01, 10 April 2017 (UTC)

I forgot about Rn (noting research as to its cationic tendencies is quite limited); At I'm not surprised about given the prediction it will be a metal, and having regard to the incipient metallic qualities of iodine. I hereby sign up to the "Cn be prepared for anything club" and the "we can probably sweep this problem under the rug for now" club. Rn reminds me of forgetting about H whenever one is talking about the nonmetals on the right hand side of the periodic table. Sandbh (talk) 02:14, 11 April 2017 (UTC)

If you ask me off the record (so with nothing to do with reliable sources and "no original research" ^_-☆), I would agree: I'd bet on the side that astatine is a metal. This is not purely a votum separatum for the sake of having one, but because the majority opinion is mostly focusing on F, Cl, Br, and I, and assumes that At will follow its "younger sisters". It isn't necessarily so. Astatine is such a backwater in chemistry that I can easily believe that Corson et al. were right and most other chemistry texts (most of which predict that astatine is going to be a black solid) are wrong (speaking in a sort of Platonic sense for now, in which the these statements have truth values independent of whether we can or will eventually find out one way on the other), because when the chemist is in a pinch at his frontiers of knowledge, his first recourse is to put on a false beard and pretend to be Mendeleev. ^_^

Now if one knows a lot about F, Cl, Br, and I, and nothing at all about At, then channelling our inner Mendeleev tells us that astatine might look metallic as a solid like iodine, and perhaps not even have a band gap and be monatomic due to the need to correct for relativistic effects, but vaporise to form a black gas of At₂ molecules. How cool that would be, especially if tennessine behaved the same way! I wonder if there have been any calculations on gaseous astatine? What a pity that there seems to be no more practical way to answer most of these questions on At and most of the superheavies than selling one's soul for its weight in francium, wishing for the power to control radioactivity, dying for the sake of the universe, and turning into the witch of instability with a metallic nature. (Sorry for the completely incongruous references, but the joke was simply irresistible.)

Incidentally I think this also accounts for the lack of studies on Fr, Ra, and Ac. They just behave too much like Cs, Ba, and La for chemists to consider them as interesting, as pretending to be Mendeleev in those cases gets you the right answer! At least Po, At, and Rn are more interesting and would deserve more chemical studies. I reckon the same thing has gone on with Mt: the hotbeds of study in period 7 have moved to the Cn–Og region, with scarcely a look at Mt, Ds, and Rg, normal noble metals as they are expected to be. (You can time the lack of interest in Rg almost exactly from when Rg's electron affinity stopped being predicted as high, precluding "roentgenides" along the lines of aurides.) The current interest in elements 119 and 120 likely stems from the fact that they are the next elements to be synthesised; after that I suspect the interest will go back to the almost-zero it was before the synthesis of tennessine completing the 7th row of the table. Double sharp (talk) 15:09, 11 April 2017 (UTC)

I wonder about At as a black gas. The "Mendeleev for a day" literature regularly refers to the halogens becoming "darker" as you descend group 17. But what is really happening? Fluorine is yellow; chlorine is yellow-green; bromine vapour is amber; iodine vapour is violet. There does not seem to darkening here, at least not in terms of the frequency shift of the colours involved. It goes yellow (580 nm?) to yellow-green (555 nm?) to amber (595 nm?) to violet (415 nm?) in other words ↓ ↑ ↓. On this basis I'd be tempted to say that astatine gas would go ↑ and be blue (470 nm), realising that I have no firm basis to make such an extrapolation. I recall that some of the true metals in gaseous forms come in colours (potassium = green) so I'd expect that if astatine was a metal it could presumably be coloured too. But I don't know if I've lost the plot here. Sandbh (talk) 00:18, 16 April 2017 (UTC)

Potassium is green and sodium is blue in the gaseous state, and the vapours of course are K₂ and Na₂ respectively. But the colours are pretty intense, so this may not be Rayleigh scattering like for F₂, Cl₂, Br₂, and I₂, but true molecular rotational-vibrational band absorption.

What we would need to speculate on this is the polarisability of the At₂ molecule, as that's important for Rayleigh scattering. This is not something that is easy to find! WebElements gives the At–At bond energy as ~80 kJ/mol. The values for the lighter halogens are I–I, 151 kJ/mol; Br–Br, 193 kJ/mol; Cl–Cl, 243 kJ/mol; and F–F, 159 kJ/mol. Well, this might be useful for making predictions. ^_^ Extra credit would be obtained for trying to extend this to Ts₂ vapour! Double sharp (talk) 04:45, 16 April 2017 (UTC)

If we have the polarisability values for F₂, Cl₂, Br₂ and I₂, could we extrapolate a value for At₂? And would that enable us to take a stab at its colour? Oh, and might At vapour be colourless, in the same way that Hg vapour is colourless? Sandbh (talk) 06:26, 16 April 2017 (UTC)

Yeah, it would help. If the result is a value outside the visible spectrum, then yes, At vapour would be colourless. Double sharp (talk) 06:50, 16 April 2017 (UTC)

Mean polarisability for F₂ = 1.38 × 10^–40 C² M² J^–1; Cl₂ = 5.01; Br₂ = 7.42; I₂ = <13.0, which would give a figure of about 20 for At₂. Are these the kind of polarisability values you had in mind? Sandbh (talk) 11:20, 16 April 2017 (UTC)

Actually, after reading enough physics to look at this, it occurs to me that for strongly light-absorbing gases like these, it's not Rayleigh scattering that is important, as they significantly absorb all wavelengths; if you had a planet with a chlorine atmosphere, you wouldn't be able to see anything, because while blue and red are indeed preferentially absorbed, green is also absorbed. At 1 atm pressure, a metre of chlorine gas would be enough to cut visibility to zero. So I most humbly apologise for making you go through the work that was necessary to find these! <(_ _)>

Late last year, I rewrote the articles on the three halogens that R8R didn't get to: Cl, Br, and I. I reread them just now and it occurs to me that I explained it there: the colour for F₂, Cl₂, Br₂, and I₂ comes from an electron transition between the highest occupied antibonding π_g molecular orbital and the lowest unoccupied antibonding σ_u molecular orbital. We would then have to calculate this for astatine, where it would involve the 6p electrons, and then see what wavelength of light would be absorbed preferentially. (Or we could hope that someone has done this already, which would suggest the answer for gaseous astatine.) Double sharp (talk) 15:11, 16 April 2017 (UTC)

"Not-so-strong" metals

On the d-block metals that neither form simple cations nor have basic oxides, these seem to be Zr, Hf, Nb, Ta, W, Tc, Os, and maybe Pa. The last one does form Pa^IV in aqueous solution but this is rapidly oxidised by air to Pa^V. In any event, Pa has a body-centred tetragonal crystalline structure so counts as a metal.

(Sorry to cut into your post, but I like how you call Pa a d-block metal! Sometimes I wonder if one shouldn't annex the f-block to the s-block entirely, along with Sc and Y in the d-block, while the elements Th–Am and maybe Cm are moved to the d-block; meanwhile we could also annex group 12 and helium to the p-block. Sure, that would have nothing to do with the electron configurations, but it would make perfect sense for the chemistry.) Double sharp (talk) 15:12, 11 April 2017 (UTC)

That was an occupational hazard of posting too late at night. It would be an odd-looking table, if the f-block was annexed to the s-block. You would have to position the d- and p-blocks below and separate from the s- and f-blocks in order to still be able to fit the thing into portrait orientation. But there would be a fair bit of room for a legend in the space formerly occupied by the period 4 to 5 transition metals in groups 4 to 11. It would be bizarre to see the f-block metals have their time in the sun at the top of the table, after having languished for so long as a forgotten annex at the foot of the table. "What do we want? F-block metals up front! When do want it? Now!" Oh, I'd be inclined to lower the positions of Th to Am slightly, to indicate their resemblances to groups 4 to 9, rather than annex them completely to the d-block. Sandbh (talk) 05:27, 15 April 2017 (UTC)

For the remaining seven less strong d-block metals, Steele (1966, pp. 101–102) says, "Because the energy differences between adjacent electron shells remote from the nucleus is small, the electrons in these shells can more easily pass from one level to the next than can those in shells nearer the nucleus. When these more mobile electrons constitute the valence electrons of an atom, the element will be more stable in the higher oxidation states. This is admirably illustrated by the later transition metals…RARITY OF SIMPLE IONS This feature of the later transition metals is also accounted for by their electronic structures. The mobility of the valence electrons referred to above enables the ions of these elements to gain stability by forming complexes other than the relatively unstable aquo-complexes. Such ions that do exist are to be found chiefly in the noble metals, particularly ruthenium and rhodium." [The chemistry of the metallic elements, Pergamon Press]

Parish (1966, p. 133) says, "Study of the aquo-, hydroxo-, and oxo-complexes of the 4d- and 5d-metals is complicated by two factors. Firstly, when the metals are in relatively low oxidation states a great many ligands, including mostly simple anions, are able to displace water from the coordination sphere. In order to avoid hydrolysis it is usually necessary to work with acidic solutions, so that an excess of anions is present. The metal is usually found in anionic complexes which may or may not contain coordinated water, or hydroxo- or oxo-groups. Consequently, most redox-potential data refer to complexes rather than to aquated species, and the values vary with the anions present. In a few cases special care has been taken to use acids whose anions coordinate very weakly (ClO⁴⁻, NO³⁻), and simple aquated cations have been observed for Mo³⁺, Ru²⁺, Ru³⁺, Rh³⁺, Pd²⁺, Ag⁺. Au³⁺, Cd²⁺ and Hg²⁺. The second complicating factor is that, for the higher oxidation states in alkaline solution, polymeric species are formed which coexist over wide pH-ranges in complex series of slowly-attained equilibria. It is difficult to establish the nature of the species present and, as discussed…the species which can be isolated in solid salts may bear no relation to those present in solution [The metallic elements, Longman]

Schweitzer & Pesterfield (2010, pp. 294) say that a few salts of Zr⁴⁺ can be prepared but that these are stable only in acid solution; and they list a Tc⁺² ion but say it decomposes, and a green-yellow W³⁺ ion, the existence of which we have previously queried. [The aqueous chemistry of the elements, Oxford University Press].

Back to Parish again, he writes (p. 179) that, "a satisfactory chemical definition of a metal (as opposed to one based on physical properties such as electrical conductivity or lustre) is that a metal is an element which, in aqueous solution, displays cationic behaviour or which has an oxide which is soluble in acids." On this basis he counts Ge, Sb and Bi as metals, and Po as a nonmetal, "although those on the boundary [B, Al, Si, Ge, As, Sb, Te, Bi, Po] are probably better classed as metalloids." His definition is actually less than satisfactory since it would preclude W, for example, from being counted as a metal, as noted by Double sharp. He later mentions germanium and grey tin in a passage that says, "these metals have low electrical conductivities (both are semiconductors)." The notion of a semiconducting metal is bizarre; and we now know that grey tin is a semimetal, in the physics-based sense. Writing further on the p-block metals (pp. 190–191) he says, "…few simple aquated cations exist[!], and all are subject to hydrolysis and must be studied in strongly acidic solutions. Uni-, di-, and tri-positive cations are known, e.g. Tl⁺, Sn²⁺, Pb²⁺, and Al³⁺ to Tl³⁺...tripositve antimony and bismuth seem to occur in acidic solution only as oxo-cations, SbO⁺ or BiO⁺…only In⁺…and Ge2⁺ are unstable to disproportionation." We now know that SbO⁺ is in fact [Sb(H₂O)₄(OH)₂]⁺ and that BiO⁺ is in fact Bi₆(OH)₁₂⁶⁺.

In summary, and riding roughshod over some of the nuances and subtleties noted by Double sharp and DePiep, it appears that the s- and f- metals are chemically strong but physically weaker; the d- metals are chemically strong to weaker (or even noble) and physically the strongest; and that the p-block metals are chemically weak and physically weak. I need to look closer at all of this but so far the voyage of exploration has been illuminating and rewarding. Sandbh (talk) 02:14, 11 April 2017 (UTC)

Yes, that's a good approximation. To add a second layer of precision I might note that the behaviour of the 3d metals is quite different from that of the 4d and 5d metals, and that groups 3 and 12 are "honorary f-block and p-block" groups. What little we know about the 6d metals suggests that they will be quite different again: Rf forms an aquated tetrapositive cation, but for Db and Sg cations are unknown, and no aqueous-phase studies have been done on the elements from Bh onwards.

I am not sure if the idea of a semiconducting metal is all that bizarre, given what Parish appears to be trying to do: give a purely chemical definition of a metal, divorced from all physical properties. Such a philosophy seems laudable for elements like copernicium, for which we cannot know the physical properties until we start exploding more nuclear bombs (viz. Es and Fm), but I am not sure if it is really ideal for the first 100 elements. Incidentally I fail to see why he classifies Po as a nonmetal, given that IIRC its cationic chemistry was already known in 1966.

I wonder though about the Sb cation. It looks so much like a hydrolysed Al or Fe tripositive cation that I wonder if we could profitably call Sb a metalloid that has real chemical metallicity; it even copies metallic reactions with acids, even if the results are not really "salts". That would seem to caution us against too much of a nonmetallic view of the metalloids, even if we bias it that way by only considering chemistry. Double sharp (talk) 03:27, 11 April 2017 (UTC)

P.S. I remain after two months slightly sceptical of that Rf cation. The actinide contraction is mostly relativistic, but apparently it goes so far that whenever Nb and Ta behave differently, Db follows the behaviour of Nb rather than that of Ta. That Rf cation must be heavily hydrolysed if that is the case, just like Zr: it may have a "notional" existence but it will surely not be of the same order as Ce^IV, Th^IV, or U^IV. Double sharp (talk) 11:01, 15 June 2017 (UTC)

Viewing metals and nonmetals too purely

Your caution about taking too much of a nonmetallic view of metalloids applies more generally to nonmetals, and it has a corollary in taking too much of a metallic view of metals. I had never thought about it this way but it now seems to me that the term "metalloid" inadvertently tends to encourage the kind of thinking you are cautioning against. By calling some elements metalloids and thereby drawing attention to their supposed funkiness—not to mention elevating them to a categorical status on par with metals and nonmetals—it becomes easy to be left with the mistaken impression that metals behave like metals and that nonmetals behave like nonmetals. Whereas in fact the situation is more permeable.

The following metallic characteristics are seen among nonmetals:

• lustrous appearance: [B], C, [Si], P, [Ge], [As], Se, [Sb], [Te], I
• deformability: C, white P, S, Se
• metallic conductivity: C, [As], [Sb]
• capacity to form alloys or alloy-like compounds: H, [B], C, N, O, [Si], P, S, [Ge], [As], Se, [Sb], [Te]
• organometallic chemistry: [B], [Si], P*, [Ge], [As], Se*, [Sb], [Te]
• cation formation: Rn (of all things)
• oxocation or hydroxocation formation: N, [Sb]
• polycation formation: C, N, O, P, S, Cl, [Ge], [As], Se, Br, [Sb], [Te], I, Xe
• salt-like compounds: H, [B], C, N, [Si?], [Ge], [As], Se, [Sb], [Te], I

* = sometimes   [ ] = metalloid

Whereas the following nonmetallic characteristics are seen among metals:

• open packed crystalline structures: Mn, Ga, In, Sn, Hg, Bi, Po, U, Np, and Pu
• negative temperature coefficient of resistance: Pu
• markedly brittle comportment: Ga, Bi
• oxyanion formation: e.g. Be, Al, V, Nb, Cr, Mo, W, Mn, Tc, Re, Fe, Ru, Os, Ag, Zn, Ga, In, Sn, Tl, Pb, Bi, Po
• exclusively acidic oxide formation:^ Zr, Hf, Nb, Ta, W, Tc, Os, Pa
• covalent bonding tendencies: Be, Al, and especially among the post-transition metals, and to a lesser extent the transition metals
• anion formation: Na, K, Rb, Cs, Ba, Ag, Pt, Au, Po, At

^ this is a bit tough to call as some of the lower oxides of the transition metals and actinides are poorly characterised

In general, the closer an element is to the metal-nonmetal dividing line, regardless of which side it's on, the more likely it is to display funkiness. This is another way of saying, as is often written in the literature, that a period represents a progressive change from strongly metallic to weakly metallic, and weakly nonmetallic to strongly nonmetallic, accompanied by irregularities along the way (and with bells and whistles in the form of the noble metals and the noble gases).

In the context of this more widespread permeability, the term metalloid has a profile that is disproportionate to its significance, consistent with the fact that so few textbooks have anything meaningful to say about metalloid chemistry. The situation immediately on the other side of fence, in terms of categorising the B-subgroup metals and describing their chemistry, is not much better.

Of course, we should retain the term metalloid, but the category name metalloid (weak nonmetal) would represent a significantly more balanced view of the elements in question, having regard to the literature.

Getting back to the aquated Sb hydroxocation [Sb(H₂O)₄(OH)₂]⁺, this has a patchy history. The older texts describe it as the SbO⁺ antimonyl  ion. It forms only in very acidic media. Pourbaix's Atlas of electrochemical equilibria in aqueous solutions (1974) sets the limit of its presence at pH 0.87. Greenwood and Earnshaw (1997) don't mention it as far as I can see. Cotton et al. (1999) say that Sb has some definite cationic chemistry but provide no more insight than saying, "cationic compounds of Sb^III are mostly of the "antimonyl" ion, SbO⁺, although some of the "Sb³⁺ " ion, such as Sb₂(SO₄)₃, are known." (It seems likely that the latter is better described as an Sb₂O₃.3SO₃ mixed oxide.) They go on to mention the presence of "SbO⁺ and/or Sb(OH)₂⁺ " in < 1.5 M H₂SO₄. Wiberg (2001) briefly mentions Sb(OH)₂⁺, and salts such as (SbO)₂SO₄, Sb₂(SO₄)₃, SbONO₃, Sb(NO₃)₃, and SbCl₃ but whether these are true salts or salts only in name is not explained. If they are referring to SbCl₃ as a salt then they are applying the term very liberally—chemists like Axiosaurus would cringe. The Aqueous Chemistry of the Elements (2010) lists the Sb cation as Sb(OH)₂⁺ and records it at a pH of < 0.

It is interesting to consider that antimony has some metallic chemistry, even if this is not that well-characterised, and that antimony may be chemically the most metallic of the elements that aren't metals. Nevertheless its chemistry is generally nonmetallic and even its oxide, while amphoteric, is predominately acidic (unlike e.g. ZnO, which is predominately basic). Physically, I think antimony might be outgunned by carbon's rather impressive electrical conductivity of 3 × 10⁴ S•cm⁻¹.

In the land of the metals I suppose gold might be a candidate for being chemically the most nonmetallic, given it has an electron affinity greater than that of O and S; the absence of a simple aqueous cation; the want of a basic oxide; its capacity to bond to itself with a bond energy greater than that of Br₂; and its halogen-like capacity to form an Au^– auride ion.

More broadly, it is relevant to consider that nearly all metals and nonmetals have contrary characteristics and that, in this regard, there is nothing so unique about the metalloids that warrants any particularly special treatment, or their distinction as a third major category of elements. (I don't intend for any of this to come over as a lecture; it's just the way I tend to write). -- Sandbh (talk) 05:38, 14 April 2017 (UTC)

Of course, all metals are in touch with their nonmetallic sides to some degree, and all nonmetals are in touch with their metallic sides to some degree – with the possible exceptions of fluorine and the light noble gases He, Ne, Ar, and Kr. (Mind you, Kr forms fluorocations if not oxocations, and Ar is predicted to do the same, so even that is a little shaky.) But I would think that the fact that many of the nonmetals who are most in touch with their metallic sides fall into a familiar narrow band, cordoning off the rest of the nonmetals from the metals, suggests that there is a real buffer zone here. We keep seeing the same elements over and over again in your first sets of lists, and they do not travel very far from the metalloid line. Sb is the most metallic, appearing seven times; C, As, and Se appear six times; P, Ge, and Te appear five times; B, N, and Si appear four times; S and I appear thrice; H and O appear twice; and Cl, Br, Xe, and Rn each appear but once. But the concentration is in that narrow strip close to the approximate location of the "metalloid line", even though it seems to be skewed to the left of conventional wisdom in the "non-relativistic" periods and to the right in the "relativistic" periods.

There is nothing like this sort of consistency in the examples you give of metals in touch with their nonmetallic sides. With every nonmetallic property, there seems to be a rather different set of metals that display those, since most elements appear only once there. The closest thing we have to this is the brief dalliance with nonmetallic chemistry in the transition metals, but the physical properties there remain steadfastly metallic – even in gold, which is still the king of metals, with no word in the literature to its chemical weakness.

And perhaps some of the "nonmetallic" properties you cite are a little too strong; anion formation is a very tough measure, since only the really small N lets you meaningfully speak about an N³⁻ anion; even the phosphides and arsenides of sodium are alloy-like. For even more negative charges than −3, one can forget about the possibility even in methanides: although it is formally possible to talk about Be₂C and Al₄C₃ as having C⁴⁻ anions, that is a very unlikely total description given the interatomic distances and the great polarising power such compact cations would have! While these do hydrolyse to give methane, as you would expect for real ionic methanides, I think we have another example here of chemistry along the lines of metallicity in Sb; the reactions are those of a metal, but the appearance of the compounds is not. None of the "big six" metalloids, except Te, form simple anions, if I am not mistaken. Such a measure of metallicity would weirdly imply that polonium is less metallic than phosphorus!

So what distinguishes the metalloids is more that they have a bunch of metallic properties and a bunch of nonmetallic properties – and are not, say, metals with one or two nonmetallic tendencies, or nonmetals with one or two metallic tendencies. There is some special pleading needed in the divisions, because it is rather odd to call carbon, nitrogen, and phosphorus metalloids; but the vague outline of the reasoning for an intermediate buffer zone is there. Double sharp (talk) 17:05, 14 April 2017 (UTC)

I'm still pondering this rather demanding classification science puzzle.

In the interim, I can say something about gold chemistry in the literature. Wiberg mentions it at p. 1279. A summary is given in the metalloid article, here. Chemical principles (Zumdahl and DeCoste 2013, p. 575) discusses what causes gold to emulate the "many" properties of the non-metallic halogens, but only specifically mentions aurophilicity, gold's electron affinity, and CsAu.

Could you clarify what you meant when you described anion formation as being a very tough measure? Are you saying this is not a good criterion for gauging nonmetallic behaviour? Sandbh (talk) 01:57, 17 April 2017 (UTC)

Anion formation

Yes, I would not consider formation of simple anions a good criterion. One reason is that the noble gases (except probably Og) have nonpositive electron affinities, so they cannot bind electrons to form anions at all; yet they are excellent nonmetals from the physical perspective, and from the chemistry of Kr and Xe. The other reason is that nonmetals in groups 14 and 15 do not form simple tetranegative and trinegative anions, with the exception of N; C⁴⁻ and P³⁻ have but a formal existence at best. Double sharp (talk) 04:25, 17 April 2017 (UTC)

Anion formation as a stand alone criterion would likely not be a valid measure of nonmetallic character, as per your noble gas example. The same could be said of (simple) cation formation as a measure of metallic character in light of the absence of this property among some of the heavier transition metals.

On the other hand, since metallic or non-metallic character is composed of different aspects, cation formation or anion formation can be regarded as being at least as good as any other applicable indicators, provided they're not relied on in isolation but rather, form part of composite assessment i.e., one using multiple criteria. That would be why the noble gases make the grade as nonmetals.

I had a look at what Wiberg says about anions of the group 14 and 15 elements. I was able to find explicit mentions of isolated methanide ions C^4– in the lattices of Al₄C₃ and Be₂C (p. 798); alkali and alkaline earth (A&AE) silicides containing Si^4– anions (p. 835); Ge^4–, Ge^2–, or Ge^– anions found in A&AE germanides (p. 896); discrete monophosphide P^3– anions in A&AE phosphides (pp. 687–688); As^3– anions in A&AE arsenides, and in Zn₃As₂ (p. 744); and A&AE antimonides containing Sb^3– anions (p. 759). I couldn't find any unambiguous mention in Wiberg of B anions. However Inorganic chemistry (Weller et al. 2014, p. 367) says, "The simplest metal borides are metal-rich compounds that contain isolated B^3– ions".

Greeenwood & Earnshaw are either skeptical (C, P, As, Sb), unclear (B, Si) or silent (Ge) on the question of such anions.

Wiberg has copious examples of polyanions formed by all these group 14 and 15 elements, e.g. Ge₉^4– and As₁₁^3–. Greenwwood & Earnshaw also make mention of such polyanions.

I find that what Wiberg and G&E say about anions of the group 14 and 15 elements is about what I would expect for these nonmetals, given their proximity to the dividing line between metals and nonmetals. That is to say, some difference of opinion about the existence of genuine anions but no doubts about the existence of numerous polyanions. Boron may be in its own category here given it's such a peculiar element. For the rest of the nonmetals i.e., H, N, O, F, S, Cl, Se, Br, Te, I, and the noble gases, these also behave more or less as I'd expect, anionically speaking. (Am still thinking about the broader classification science question.) Sandbh (talk) 12:39, 17 April 2017 (UTC)

Okay, but again, if the simple-anion-forming nonmetals are all clustered in the "diatomic nonmetals" area of the periodic table plus the rest of groups 16 and 17, then it seems to suggest a fragmentation in the nonmetals between those that do form simple anions; those that don't, but not because of negative electron affinity; and then those that don't, because of negative electron affinity. The first seems to be the "typical nonmetal" area of the periodic table with H, N, O, S, Se, Te, and the halogens; the second seems to be the more metallic area up to group 15; and then the last is group 18. So we've mostly just called out a specific and clear subset of the nonmetals, but it's not a property that is characteristic of all of them. Furthermore you also have all those metal anions – and if we're going to polyanions you'll have to add basically all the group 13–16 metals to your list as Zintl phases. So either this criterion seems to cast too small a net as a subset of the nonmetals, or it seems to subsume pretty much the whole p-block until we reach group 18 and fully fill the shell. Either way, this does not seem to be a property that uniquely characterises the elements we like to think of as nonmetals, since it not only misses out a significant number of nonmetals but also includes a significant number of metals as well, most of which also come from this borderline territory. Since this property is shared by elements on both sides of the metalloid borderline I am not sure if it is very useful in deciding what belongs where. Double sharp (talk) 15:28, 17 April 2017 (UTC)

As for gold, while its nonmetallic properties are certainly noted in the literature, does anyone use them to classify it as a weak metal? Double sharp (talk) 15:28, 17 April 2017 (UTC)

I would not focus too much on the incidence of simple anion formation by itself, but when you combine this with oxyanion formation, the pattern becomes clearer. In general:

• O and the halogens are noted for their simple anion chemistry;
• H, C, N, P, S and Se are more noted for their capacity to form oxyanions than simple anions;
• the metalloids behave similarly with the exception that they show less tendency to anionic behaviour generally

And the noble gases are in a league of their own, just like some of the cationically challenged TMs are in class of their own.

As you note, polyanion formation is a bit of neither here nor there since we would have to add in the group 13--16 metals given their capacity to form Zintl phases. I was reluctant to mention Zintl phases in my initial comparison of funkiness among the metals and nonmetals partly for this reason. And I only listed some details of the rather amazing range of nonmetal polyanions as, until we had this chat, I didn't even know there were so many them.

I'm not sure you can imply that since anion formation is encountered amongst metals as well as nonmetals that it is not very useful in deciding what belongs where. By itself this is maybe right but if you drill down a bit and combine it with other more metallic or more nonmetallic attributes, it has its place.

With the exception of a few authors who note the capacity of gold to act as a main group metal (and hence it can be regarded as post-transition metal in this sense) nobody I know of classifies it as a particularly chemically weak metal since (a) everybody seems to be overwhelmed by its physical superlatives; and (b) general textbook treatment of gold chemistry is rather superficial. I suppose its nobility might be the closest thing to it being counted as chemically weak. It should be classified as a transition metal (or noble metal) noting the transition metals span a range of chemical liveliness. Sandbh (talk) 06:24, 18 April 2017 (UTC)

Metalloids either lacking equally or skewed

Well of course Au is not usually well-studied even up to the high school level; the members of group 11 that are studied there are just Cu and Ag, and of course Au is not treated in detail. But I do not think specialised books on general inorganic chemistry, or even group 11 or Au, would think of it as a weak metal. It is physically a great metal and it that respect is in a similar situation to Be, Al, and Po: they are next to the metalloid line and show some chemical weaknesses, but don't have physical weaknesses.

I am not sure if nobility is really a sign of chemical weakness: copper is pretty noble, but there are even fewer problems with it as a metal. Many metals are not quite as reactive as the most electropositive ones in groups 1 to 3, and even then the lanthanides get less active as they shrink. (The actinides seem to show the opposite trend, but I have to wonder how much of this is due to radioactive self-heating: not only do the actinides shrink from Th to Lr, but their stability drops precipitously too!) The questions start coming with oxyanions; but don't many of the B-subgroup metals do that too, according to post-transition metal?

By the way, doesn't this mean that the metalloids are lacking in several significant nonmetallic properties (e.g. anion formation), just as they are lacking in several significant metallic properties? If the situation is approximately balanced that way, I think we would have a serious justification for retaining the "borderline territory".

A lot of the classification of these borderline elements is contextual anyway: if I was comparing Ge with Sn and Pb, I might take a more metallic view of it than if I was comparing it with Si, and in both cases the viewpoint can be supported by some properties and opposed by some others. The same is, I think, true of the other metalloids: you can draw fruitful analogies of them with metals just as you can with nonmetals. So perhaps they should really be in both categories and are taken to be a class apart purely to avoid creating a seeming contradiction in terms! Double sharp (talk) 06:49, 18 April 2017 (UTC)

I agree gold is physically a pretty good metal but for lacking structural strength. I did not suggest it was a weak metal; I suggested it may be a candidate for being chemically the weakest metal.

Be, Al and Po, being closer to the dividing line between metals and nonmetals, show varying physical shortcomings. Be has a directional (covalent) component in its bonding that lowers its packing density and limits its ductility. The physical weaknesses of aluminium are discussed in the post-transition metal article, here. Po has a crystalline structure characterised by partially covalent bonding; the literature is contradictory as to whether Po's simple cubic structure would be brittle or ductile. We know Po apparently has a softness similar to lead so it is likely to be lacking in structural strength.

On nobility as a marker of weak metallic character I had in mind gold's rather impressive standard electrode potential of +1.52 eV (for Au³⁺) which, in this case, is an indicator of its "buffed up for a metal" capacity to accept electrons. In this regard A companion to physical and inorganic chemistry (Stott 1956, p. 100) notes that a high positive electrode potential indicates either a weak metal or a nonmetal. It seems to me that copper, at +0.337 (for Cu²⁺), is not in the same league.

Rather than saying that the metalloids are lacking in several significant nonmetallic properties and lacking in several significant metallic properties, I would say that they are physically strong for nonmetals but chemically weak. Here, "physically strong" means being solid, lustrous, and of relatively low volatility and, for nonmetals, having pretty good electrical conductivity and relatively close-packed crystalline structures.

I agree that classification of the elements to either side of the dividing line between metals and nonmetals is contextual. By this I mean comparing post-transition metals to the transition metals, and to the pre-transition metals, lanthanides, and actinides; and comparing metalloids to the rest of the nonmetals, particularly the corrosive nonmetals. Sandbh (talk) 00:09, 19 April 2017 (UTC)

I am a bit confused about your penultimate paragraph: surely your physically strong properties listed imply weak nonmetallicity? Surely physically the best nonmetals are the noble gases, as almost the complete opposites of the metals in their properties. Their chemistry is indeed significantly nonmetallic, but not only is there a significant nonmetallic component in the properties of the metals of group 12 onwards, but there are some metallic properties too there that become clear upon comparing these elements with their more metallic congeners. From C to Pb, there is a clear trend between the nonmetallic C and the metallic Pb. All of us can say that C is not a metal and Pb is, but where does one colour of the rainbow turn into another? That is why I think the idea of metalloids as a separate category has merit.

If the value for Cu isn't high enough to be conclusive, then perhaps Po (+0.6 V) would be more so, considering that it equals the values for Ru and Rh.

I'm also a bit unconvinced the nonmetallic nature of Be from covalent bonding, given that many of the early 5d metals have that too (accounting for their extremely high melting points). And Pb isn't all that close to the metalloid line, suggesting to me that it's simply a symptom of the abrupt reduction in metallic character that happens between groups 11 and 12. Those things seem to follow vertical lines between groups instead of the diagonal stretch of metalloids.

And I think some of the later metals (Bi, Po?) might be even weaker than Au chemically. Double sharp (talk) 08:43, 19 April 2017 (UTC)

Oh dear, I hope we aren't confusing one another too much.

Yes, the "physically-strong-for-nonmetals" properties I listed for the metalloids imply weak nonmetallicity. Physically, the "best" (i.e. the most representative in this context) nonmetals would be the most volatile and most insulating i.e., the gaseous nonmetals including the noble gases. The metals from group 12 onwards, even allowing for their nonmetallic properties, are more metallic than the metalloids. Yes, C is a nonmetal and Pb is a metal. The colour of the rainbow changes as informed by the literature. Thus Si and Ge are relatively commonly recognised as a metalloids)^ and Sn is a metal. The simplified four-part definition of metals I discussed previously is consistent with this. Yes, keep metalloids as a separate category, with a parenthetical qualifier "(weak nonmetal)" tacked on. I recall we agree they are nonmetals: more like metals than other nonmetals but nonmetals nonetheless.

^ Noting that descriptions of the overall chemistry of individual metalloids are invariably nonmetallic in nature, and that the general chemistry of metalloids is described as being nonmetallic.

Po does looks more noble than Cu, judging by electrode potential.

The reference to directional bonding and reduced packing efficiency etc comes from Structure-property relations in nonferrous metals (Russell and Lee 2005) who discuss the increasing incidence of distorted or odd crystalline structures in the metals, the closer you get to the nonmetals, with some surprises along the way like Mn and Pa–Pu. According to them, all other transition metals and the Ln and An display normal metal crystal structures with the exception of Hg (and they note that "HCP" Zn and Cd have packing efficiencies of only 65% due to contributions from covalency).

I'm not sure about your particular context for mentioning Pb. It has a close-packed structure but an abnormally large inter-atomic distance that has been attributed to partial ionisation of the lead atoms.

Bi and Po might also be candidates for chemically the weakest metal. Inorganic chemistry (Phillips & Williams 1966, vol. 2, p. 459) says, "Because of the increase of nuclear charge across each of the transition series, the B metals are distinguished from the early A metals by their much weaker tendency to form ions or to form compounds with the non-metals. The ions themselves have high electron affinities and tend to seek out, wherever possible, polarisable anions or ligands. This feature is particularly marked in the final row of the B metals, Au, Hg, Tl, Pb, Bi, and Po, where the nuclear charge has been built up across the lanthanide as well as the third transition series. In some respects these elements might almost be classed as super-B or C metals." Sandbh (talk) 04:59, 20 April 2017 (UTC)

Preponderance principle

I think I would only agree about metalloids being nonmetals if we defined the latter word purely etymologically. The way I see it, an element is a metal if it has a preponderance of metallic properties, a nonmetal if it has a preponderance of their negation, and then the term "metalloid" is useful for the intermediate things. Of course group 12 are more metallic than B, Si, Ge, As, Sb, and Te. But they have nonmetallic properties too and the important thing is that more of their properties are metallic than not. Every metal has some nonmetallic tendency and every nonmetal (well, except He, Ne, and Ar) has some metallic tendency. The question is simply which properties dominate, and there must be a region in the middle. Double sharp (talk) 06:53, 20 April 2017 (UTC)

(Placeholder: when I get back I'll look at the p-block groups and discuss where we can see the seeds of the metallicity of Sn and Pb already sown in Si and Ge.) Double sharp (talk) 08:45, 20 April 2017 (UTC)

I believe you may be onto something with the preponderance principle. Physically, if this is applied to the metalloids I think the result is that, at best, they are a mix of metallic and nonmetallic. The academic in me would, in some cases, move the bar to the left in terms of where the cut off is between what is a metallic property and what is a nonmetallic property. But I won't do this for the purpose of this edit. Chemically, the metalloids are generally nonmetallic. Some aspects of their chemistry are metallic but overall it's nonmetallic, as per the literature. The preponderance of properties combined then seems to me to be nonmetallic. Thus, the "region in the middle" is then represented by the post-transition metals and the junior nonmetals. In this region are found metals like "paradoxical" Al, "covalent molecular" Ga, "not too cold please" Sn, and "simple cubic" Po, collectively showing quite a few nonmetallic properties; and junior nonmetals like B ("Pick me! Metals are in fact honorary boron atoms!"),¹ Ge ("No, pick me! The boxes I'm shipped in are labeled in large bold letters, GERMANIUM METAL"!),² and Sb ("Forget those chumps! Pick me! I can conduct electricity like a poor metal!").³ Wimpy post-transition metals and braggart junior nonmetals aside, I feel this gives a more accurate picture of the lay of the land in this part of periodic table, where the metals meet the nonmetals. (One could even refer to the junior nonmetals as metalloid nonmetals, since they are the nonmetals that most resemble metals.) Sandbh (talk) 12:02, 21 April 2017 (UTC)

1. Greenwood NN 2001, "Main Group element chemistry at the Millennium", Journal of the Chemical Society, Dalton Transactions, issue 14, pp. 2055–66 (2057)
2. Haller EE 2006, Germanium: From its discovery to SiGe devices, p. 3, https://www.osti.gov/scitech/servlets/purl/922705
3. Masterton WL & Slowinski EJ 1977, Chemical principles, 4th ed., W. B. Saunders, Philadelphia, p. 160

But then why not just call them "metalloids"? Doesn't the "-oid" suffix already adequately imply that they are like metals but in reality don't qualify as such? As I previously mentioned, the suffix comes from Greek -ειδής, related to εἶδος "form, likeness", so it seems to me that the name "metalloid" already tells you that these elements are similar to metals but are not metals. (In fact, Wiktionary defines the English suffix "-oid" in exactly that way!)

The distinguishing feature between the metalloids and the nonmetals would be that while neither set passes the bar to be called metals, the former are much closer than the latter. I'll save more detail for the coming text. Double sharp (talk) 14:20, 22 April 2017 (UTC)

Oh, I was using "metalloid" here as an adjective rather than a noun, and to better maintain consistency of nomenclature. Thus: metalloid nonmetal, intermediate nonmetal, and corrosive nonmetal. (And I believe the adjectival form of metalloid predates the introduction of the nominal form). I was also conscious of our earlier discussion on loading too much into the term "metalloid" just by itself. Sandbh (talk) 23:47, 22 April 2017 (UTC)

"Orphan" nonmetals

There is a juvenile-level book called, Science of Everyday Things: Real-life chemistry (Knight 2002, Gale Group), which divides the nonmetals into noble gases, halogens, and "orphan" nonmetals. I laughed when I saw this. In a funny way it is marginally better than other nonmetals. Five chemistry credits for novelty value. Sandbh (talk) 22:24, 3 April 2017 (UTC)

The future of periodic table categorisation

(Paragraphs were moved here from [1] and [2]). -DePiep (talk) 19:31, 22 March 2017 (UTC)

• I'm sorry to say what I'm about to say next. Considering how important classes are in chemistry, one would think that IUPAC would have gotten its terminological and conceptual house in order but no, it hasn't, so we're stuck with the mess that is metalloids and semimetals, while the physicists happily apply these words to their own, much better defined, physics-based uses. Poor show, IUPAC. Sandbh (talk) 12:28, 22 March 2017 (UTC)

Sure this is a task waiting for that IUPAC initiative. After groups, periods and blocks I think categories (we needed to invent this word too) are the 4th graphic structural visible in the periodic table (the rest is text in element cells). BTW, we could filter & funnel our threads into some Categorisation of elements by metallic-nonmetallic characteristics. -DePiep (talk) 16:59, 22 March 2017 (UTC)

• Closer at home, there is this same classification issue in Wikidata. In short, an element hydrogen (Q556) does not have wikilinks at Wikidata, instead it has eg a property called "PT group". Then Wikidata structures this all by formal claims like "is a ..", "is part of ..", etc. This is called classification, or orthology, or categorisation. These class relations are being defined right now at Wikidata ("is the H atom: also element H, or a manifestation of element H?"). AFAIsee, our infoboxes are way more precise & correct. -DePiep (talk) 21:30, 22 March 2017 (UTC)
  • What a philosophical issue we have run into here! I tend to see the elements more as abstract concepts that exist whenever we have atoms of that particular element. That explains the term "pure element", which otherwise seems fairly redundant; so you can have a pure sample of Li, having only Li atoms, but you can equally well have LiCl or BuLi and that would fall under the chemistry of the element Li as well. That is also why you can talk about elements in nonzero oxidation states. Double sharp (talk) 10:08, 13 April 2017 (UTC)

'Symmetry' comments

• Symmetry, and " 4 + 4 = 8 symmetry components is cool". I get the feeling, but please take care not to use an outcome as an input argument. "It is symmetric so it must be something good" is not sound reasoning. The periodic table is full of exceptions, we'd not want to miss one this way (by being misguided seeing a non-existent regularity). -DePiep (talk) 12:07, 12 March 2017 (UTC)

I partly agree. I guess that is why I put it in as a footnote. I suspect Mendeleev would have approved :)

While there are irregularities, the overall pattern in the periodic table of strong metals to weak metals and weak nonmetals to strong nonmetals is very well recognised (not forgetting the noble metals and noble gases). Sandbh (talk) 06:09, 15 March 2017 (UTC)

Of course and how could I forget. Mendeleev's approach should not be disqualified, because a broad brush painting is what helped him create the table in the first place (stepping over say disruptive details like atomic weight issues). -DePiep (talk) 08:44, 28 March 2017 (UTC)

Other comments

... split from previous section

Descriptive v structural nomenclature

DePiep's comments are brilliantly insightful. I would add one more thing, though: I find our current use of "polyatomic" and "diatomic" quite felicitous, because it relates the molecular structure of the pure elements to their chemical behaviour (even in compounds), thus linking together two slightly different things. Double sharp (talk) 13:22, 12 March 2017 (UTC)

Hmm. I like the polyatomic, diatomic and monatomic division too. This nomenclature is a bit different from the other categories, which tend to be more descriptive. That is, alkali, alkaline = caustic. Ln/An = like lanthanum/like actinium (these two are a bit more obscure); transition = going from one thing to another; post-transition = coming after the transition metals (these metals used to be called B metals, which at least conveyed the impression that they weren't as buff as the rest of the metals); metalloid = metal-like (mind you, the first time I saw this word I didn't know what it meant); halogen = salt former; noble gas = relatively inert. Polyatomic and diatomic describe the structures involved but don't tell you anything about the character of these nonmetals. And showing the progression in metallic to non-metallic character going across the periodic table is a bit harder to grasp with these two categories than it is with the alternative proposal. It certainly is easier to find references in the literature to this progression based on a first order observation of the nature of the elements as strong, transitional, moderate or weak metals or nonmetals, than it is to have to drill down and find explanations of the same progression based on crystalline structures.

It might come down to comparing the two options…

• polyatomic nonmetal ✦ diatomic nonmetal; or
• weak nonmetal (metalloid) ✦ intermediate nonmetal ✦ corrosive nonmetal,

…and working out which one is more information-friendly?

If our role is to summarise topics with an eye to toward broad-based, literature-supported consensus on subjects, then I think the alternative proposal has more consensus credibility. There is consensus that:

• relatively weak nonmetals are sometimes called metalloids
• oxygen, fluorine and chlorine are the most buff nonmetals; bromine and iodine are not quite so bad-ass but still in the same ballpark
• the remaining nonmetals (setting aside the noble gases) are described, more or less, using more moderate language.

There is consensus that diatomic and polyatomic nonmetals are diatomic and polyatomic but it is quite hard to find discussion in the literature on their other shared properties.

And if the way chemists tend to write is a reflection of the way they think, then they will think about the chemistry of different nonmetals in terms of where each nonmetal approximately lies on a composite scale (or spectrum) of nonmetallic "liveliness" or "intensity" rather than in terms of polyatomic or diatomic nonmetals. At one end of such a scale lie the corrosive nonmetals; at the other end are found the weakly nonmetallic metalloids. Sandbh (talk) 04:32, 20 March 2017 (UTC)

• Let's not forget, before we get used to these appearances: this is once again a wonderful great overview of element property & behaviour by Sandbh. Well researched & sourced, good to read, scholarly level. -DePiep (talk) 18:51, 12 March 2017 (UTC)
  • Hear, hear! Double sharp (talk) 04:12, 13 March 2017 (UTC)

Thank you!, DePiep and Double sharp. Sandbh (talk) 04:32, 20 March 2017 (UTC)

Why metal-normative?

• As a non-specialist in this field, but as someone who has thoroughly enjoyed reading this thread, I do have a question that I would like to bring up: Why is the entire periodic table framed through the context of "metals"? As mentioned above, "metalloids" are almost non-metals in metallic sheep's clothing. 'Fake metals', if you will. And the group of things not metals at all are just called "Non-metals". I always thought of the periodic table as having two equal groups of elements with a fuzzy bit in between. But the more I see the word "non-metal" in this conversation I think to myself "...hey, wait a minute... surely chemists throughout history observed non metallic elements from a very early time". Why didn't they give it a more fabulous name than simply "not-that-other-thing"? Like if memory serves me well, I seem to remember that even things like heat and light were thought to be elements at one point. So it's just strange to me that in all these discussions about how to categorise the periodic table (indeed formed by the literature, but still,) that it seems to be, in a sense, metalnormative, with the others either relegated to "kinda metals" or "not metals at all". The categorisation is basically how much like a metal or how little like a metal something is. Like it's one big continuum and the futher right you get the less strong the element's metallic qualities become. Strange. Any thoughts on this observation?--Coin945 (talk) 17:01, 17 April 2017 (UTC)
• (As a random side note, if I had my way I'd choose a categorisation something like this: horny metal, ehh metal, pretend metal (metalloid), ehh loid, horny loid, buddhist loid) :P --Coin945 (talk) 17:13, 17 April 2017 (UTC)

metalnormative ;-) that's only a PT-subtopic these months, this talkpage.

Thanks Coin945. This is a fine contribution, nice to read your view; it makes me think again. Spoiler: I don't have an answer. Just this: the 'metallishness categorisation' issue (and so main color keys, here at enwiki) is only one dimension of the PT. While it occupies six weeks and this whole talkpage: just one dimension, and not even a top-3 one (How important then? -- I say: after atomic number [absolutely], groups, periods, blocks: quite relevant). So we want to show that property as good as possible. Disclosure: I'm from the PT graphics department, not the chem/phys side. I want the PT to look great. -DePiep (talk) 19:14, 17 April 2017 (UTC)

I suspect that the reason is that not only are there far more metals than metalloids or nonmetals (H, He, B, C, N, O, F, Ne, Si, P, S, Cl, Ar, Ge, As, Se, Br, Kr, Sb, Te, I, Xe, At, Rn, Og is a reasonable list; 25 such elements versus 93 metals!), but they are also a far more diverse bunch. I love the d-block metals, and I'll be among the first to argue fot their individualities (and goodness knows the "dark d-block" in the first few groups needs more love in organic synthesis!), but you won't go too far wrong thinking of the transition metals as hard, strong, with high melting and boiling points, variable oxidation states, resistance to corrosion (well, for the most part), and high thermal and electrical conductivity. Until you get to group 12 they really are a quite homogeneous bunch, with the borders and trends repeating from period to period.

Then you go into the p-block, and it's a frightful mess, because of the huge differences in chemical properties. To take one example, things gradually become more active the further you go to the right of the table, giving the stereotypical middle-school nonmetallic properties – and then you reach the far right, with the noble gases, and suddenly instead of delight in setting various things on fire to be tamed for their great power (N, O, F – the former if you give it a push, since the triple bond of the elementary state is too stable for such fun), we reach Ne and the element stops doing any of this work and metaphorically just sits down and sips tea by itself. And yet physically the noble gases are the best nonmetals, while their chemistry is really weird from the point of view of the other elements (even though you can rationalise it in hindsight: Br and I in various oxidation states tend to correspond to Kr and Xe in the oxidation state one above). But even then the properties are mostly clear by negation; carbon, phosphorus, selenium, and iodine are semiconductors; most of the metalloids are that too, and arsenic and antimony are even semimetals. The trouble with such a small sample is that deviations like that count more significantly than they do with the overwhelming majority of metals.

I love your categorisation and find it hilarious! Truth be told, I tend to think of the groups past 13 (except boron) as a spectrum towards metallicity, with no really clearly drawn "metalloid line" in my head. But I will note that the metalloid line seems to produce significantly better results if you draw it one more step to the right. After all, the left side of the current one (Be, Al, Ge, Sb, Po, Ts) is more metallic than not: Sb is "almost" a metal and Ge is the only weird one here. Maybe we should do a "loop" there to replace Ge with As! ^_-☆ Double sharp (talk) 23:03, 17 April 2017 (UTC)

• That's an interesting assessment! And thank you for your compliment for my cute categorisation. :P Another way I like to think about it is that (excluding the noble gases), the left side is male and the right side is female, and the further out you go the more desperate they are to 'have chemistry' with the other gender to become a couple. I see the elements on both edges as sexually repressed conservatives. They can only be with one mate and are unable to be content by themselves due to horniness/anti-masturbation rules, so they're willing to mate with anybody of the opposite gender. Doesn't really matter who. First in, best undressed. But once they've found someone their particles combine like nobody's business, the male gives their sexy electron to the female and they become monogamous. (In this universe, Na+ doesn't get a divorce and bond with extremely horny older man Caesium ten years down the track).

• The further inward you go, you get gender fluidity and liberalism. I would consider metalloids and those nearby to be closer to gender flexible or even transgender. But certainly liberals, as they're cool to go with either gender, and they're also into orgies (metal) and threesomes (water) and other cool stuff. And noble gases are kinda asexual. Not super into sex at all. Just chillin' and playing GTA5. Oh, there's also #IncestuousCarbon and #IncestuousSilicon, and molecular gases could be considered 'cousins with benefits'?

• I could go on... :P But that's just the ramblings of a madman... ^_^ Would make chemistry class more fun though. Was shown a video once where Sodium was anthropomorphised as a playaaa. (As another random side note, why is "noble gas" not called "noble non-metal", so like, I dunno, consistency? Because you can have gaseous metals. So the term "Noble gas" doesn't really tell you much by itself).--Coin945 (talk) 02:55, 18 April 2017 (UTC)

• Well, even the edges of the table bond to themselves a lot: a piece of sodium is just a regular arrangement of Na⁺ ions in a sea of delocalised electrons, and elemental chlorine is Cl₂. I do find that when I "personify" elements for explanations I follow a scheme like that: Pb is clearly male and Cl is clearly female. Clearly Au is very much in touch with his feminine side but I don't think anyone thinks it's anything more than that, to paraphrase my misgivings about the nonmetallic properties of gold in this great popularisation, regardless of what Rb and Cs might say (and even they form anions). So "gender flexibility" seems to permeate the whole table; the metalloids tie into your proposed analogy fairly well, looking like metals but behaving in many ways more like nonmetals (though the metallic proprties are there too). I don't think I've ever used the analogy for such borderline cases as Ge, As, and Sb, but it is cute!
• As for "noble gas", high melting and boiling points are quite the normal state of affairs for metals, so the gaseous state suggests nonmetallicity until you get to the region of huge relativistic effects (copernicium and flerovium are probably gaseous metals; oganesson is probably neither noble nor a gas, but surely no one will change the name of the group for it, creating a funny opportunity for retelling Voltaire's joke about the Holy Roman Empire in an entirely different context). Double sharp (talk) 05:50, 18 April 2017 (UTC)

Hi Coin945; nice to hear from you. I guess the metal-normative thing might be able to be traced to the seven or so metals of antiquity. The Oxford English Dictionary gives this etymology for the word metal: "[a. OF. metal, metail (mod.F. métal), ad. L. metallum mine, quarry, substance obtained by mining, metal, ad. Gr. µέταλλον mine; app. related in some way to µεταλλᾶν to seek after, explore. The word has passed (directly or indirectly) from Latin into all the Rom. and Teut. langs.: cf. Pr. metalh, Sp., Pg. metal, It. metallo; G. metall, Du. metaal, Sw. metall, Da. metal.]" and has this to say about the meaning of the word, "Any member of the class of substances represented by gold, silver, copper, iron, lead, and tin. Originally this class was regarded as including only these bodies together with certain alloys (as brass and bronze), and hence as definable by their common properties, viz. high specific gravity and density, fusibility, malleability, opacity, and a peculiar lustre (known specifically as ‘metallic’). In process of time other substances were discovered to have most but not all of these properties; the class was thus gradually extended, the properties viewed as essential to its definition becoming fewer. From the point of view of modern Chemistry, the ‘metals’ are a division (including by far the greater number) of the ‘elements’ or simple substances. Among them are all the original (simple) ‘metals’; of the later additions to the list some possess all the properties formerly viewed as characteristic of a metal, while others possess hardly any of them; the ‘metallic lustre’ is perhaps the most constant."

The only other elemental substances that the ancients were aware of were C and S, but I suspect they viewed these things no differently from other substances, like salt, water, earth, wood, oil, and other minerals. I guess the metals and alloys of antiquity distinguished themselves by their shared characteristics and usefulness. Whereas others substances were more diverse and not so amenable to division into widely recognised classes. Over the years more and more metals were discovered with more or less similar shared characteristics thereby reinforcing the identity of the class called "metals" whereas when new nonmetallic substances or nonmetals were discovered there did not seem to be so much of a rhyme and rhythm to their character.

I once tried to trace the origin of the word non-metal. The furtherest I could go back unambiguously was to the prodigious Antoine Lavoisier in 1789, when he used the term "non metalliques" in his revolutionary work, Elementary Treatise on Chemistry. In its first 17 years this work was republished in twenty-three editions and six languages, and carried Lavoisier's 'new chemistry' across Europe and America. I do recall from looking up the works of the ancients that it was relatively easy to find references to things called metals, but that there was no conception of a single class of non-metals.

You and Double sharp have given me more to think about, which I'll see if I can address as part of the broader "whither metalloids" question. Sandbh (talk) 00:15, 18 April 2017 (UTC)

@Coin945: Given your observation about Na, it occurs to me that Goethe wrote the motto for the most electropositive metals way back in 1776 in his Claudine von Villa Bella:

Mit Mädeln sich vertragen,
Mit Männern 'rumgeschlagen,
Und mehr Kredit als Geld;
So kommt man durch die Welt.

Getting along with women,
Knocking around with men,
Having more credit than money,
Thus one goes through the world.

One should of course read "women" as "nonmetals", "men" as "metals", and "money" as "electrons". ^_^ Double sharp (talk) 14:04, 22 April 2017 (UTC)

Category names with attitude

I was inspired by Coin945's side-note about more edgy category names, and Double sharp's crack-me-up image of neon quietly sipping tea by itself, to compile the following list. It strikes me that these (for fun) category names actually explain what's going on better than the mundane names.

• Hyper metals: Groups 1–3, Ln, An
• Working-class metals: Goups 4–11 (excluding the noble metals)
• High society metals: Ru, Rh, Pd, Ag, Os, Ir, Pt, Au
• Poor metals: Zn, Cd, Hg, Al, Ga, In, Tl, Sn, Pb, Bi, Po, At

 ----- Dividing line between metals and nometals ----- 

• Junior nonmetals: B, Si, Ge, As, Sb, Te
• Respectable nonmetals: H, C, N, P, S, Se
• Psycho nonmetals: O, F, Cl, Br, I
• Cup of tea nonmetals ("Tea should be take in solitude"― C.S. Lewis): He, Ne, Ar, Kr, Xe, Rn

--- Sandbh (talk) 02:19, 19 April 2017 (UTC)

I see that Se was the last discovered of the respectable nonmetals and is therefore the most junior ^_^ Sandbh (talk) 09:56, 19 April 2017 (UTC)

I corrected a few things that seemed to be typos; if they were intended, please do not hesitate to revert me! The An are "hyper" in a slightly stranger way than the Ln; the actinide contraction ought to make them get less reactive from Ac to Lr, but actually they get more reactive, because they get more and more unstable and the increasingly violent α- and β-radiation provides the necessary activation energy.

This state of affairs makes me wonder if in bulk, the 6d and 7p elements from Hs to Lv inclusive might not be as noble as they are expected to be! Not that it is a bad fantasy. We know perfectly well that our knowledge of the chemistries of many elements are hampered by radioactivity and the resulting severing of bonds. But we just want to peek into the rose garden, where the strong force blooms with splendour, and see what would happen if the terrible mistress of the elements beyond Bi were to be distracted momentarily from her job acting as a chemical Lachesis.

I'm not sure any name here works well for the transactinides, BTW. Double sharp (talk) 09:56, 20 April 2017 (UTC)

When I read the Hypermetal group, I read Ln and An as "Alan" and "Anne. I was impressed by how method you were going this the whole personification angle. Yep now I've re-read it. :P--Coin945 (talk) 11:17, 20 April 2017 (UTC)

N and S can be pretty psycho as well, if given the opportunity with high temperatures. The high electronegativity of N also makes parallels with O and F evident. Double sharp (talk) 09:56, 20 April 2017 (UTC)

I'm still unsuccessfully searching for a good name for the transactinides in this scheme. "YOLO elements"? "Suicidal elements"? It depends on whether you think of them as wanting to decay. Their lifetime is so short that their chemistry is not well-known at all; so paradoxically the best way of identifying them is by their alpha decay. Chemistry can certainly isolate long-lived copernicium isotopes (²⁹¹Cn and ²⁹³Cn are expected to live for millennia); but how do you prove that they are copernicium without the alpha decay? Double sharp (talk) 14:00, 22 April 2017 (UTC)

Cheshire metals? Nah, YOLO is better. YBG (talk) 18:44, 22 April 2017 (UTC)

Well, oganesson is probably not a metal. Then again it is probably not a "cup of tea" nonmetal either; already radon is starting to leave the tea room. Double sharp (talk) 05:06, 23 April 2017 (UTC)

Nitrogen

Thank you Double sharp. The high electronegativity of N bothered me for quite a while until I read that it is nevertheless a poor oxidising agent. Only when it is in a positive oxidation state (i.e. in combination with oxygen or fluorine) are its compounds good oxidising agents but even then their reactivity is often limited by kinetic factors. When I was reading about the descriptive chemistry of nitrogen I found that it was somewhat hard to get a bead on its nature, whereas was not an issue with O (by far the most abundant oxidizing agent) and the halogens (those who speak for themselves, including iodine the gold-slayer). I note that, like O and F, N is still a good hydrogen bond former. Sandbh (talk) 10:38, 20 April 2017 (UTC)

I wrote about this last year at the nitrogen article: "Nitrogen resembles oxygen far more than it does carbon with its high electronegativity and concomitant capability for hydrogen bonding and the ability to form coordination complexes by donating its lone pairs of electrons. It does not share carbon's proclivity for catenation, with the longest chain of nitrogen yet discovered being composed of only eight nitrogen atoms (PhN=N–N(Ph)–N=N–N(Ph)–N=NPh)." I think we have a case where the personality of nitrogen is more like that of a "psycho" nonmetal and it is simply remaining functional by keeping itself on a tight leash in the elemental state with the extremely strong N≡N triple bond; this also makes its compounds tend to be rather "psycho", as they are usually endothermic for the same reason. The only property I can think of that is shared by N with C is the proclivity towards multiple bonding, but that's not confined to them; it's a thing for the entire second row by the double bond rule. So keeping nitrogen away from oxygen and fluorine still really doesn't feel quite right to me: it doesn't feel different enough. Double sharp (talk) 15:01, 20 April 2017 (UTC)

Yes, the N thing is quite interesting. Writing about the chemistry of N in Comprehensive inorganic chemistry (vol 2., Bailar et al. 1973, p. 162) Jones says: "The traditional exercise of vertical relationships between nitrogen and the other elements of Group V is probably not very meaningful except to emphasise their great differences. So few nitrogen compounds have isostructural phosphorous analogues that only isolated comparisons can be made. It is more profitable to relate the chemistry of nitrogen to that of its neighbouring elements in the first short period, carbon and oxygen, the so-called horizontal relationship. For example, nitrogen is strongly electronegative, forms hydrogen bonds, is capable of catenation to a minor extent (the longest nitrogen chain so far reported being N₈ in the organic derivative [as per above], and is limited to a maximum covalency of four, all properties which would correctly position nitrogen between carbon and oxygen in the first row of the Periodic Table." In a case like this where it translates to 1–1.5–2, I tend to round down, so that it becomes 1–1–2. I have a harder time conceiving of N being on par with the psycho nonmetals. It strikes me as being more of a respectable nonmetal in nature, albeit with some idiosyncrasies. Sandbh (talk) 22:37, 20 April 2017 (UTC)

But it's exactly the high electronegativity and hydrogen bonding that make me want to round up instead of down, given that you do not see this consistently and strongly anywhere else other than in N, O, and F. Even the "minor extent" of catenation reminds me far more of oxygen than it does carbon, as for O it is also minor while for C it is obviously anything but. Certainly there are similarities between N and both of C and O, and there are even occasional if very dim ones between N and P (e.g. orthonitrates correspond to orthophosphates). But in general, I find the similarities to O more convincing. Double sharp (talk) 22:54, 20 April 2017 (UTC)

Well, I'd be reluctant to round up on the basis of two properties absent of other considerations. The high electronegativity of N does not appear to play out as much as would be expected compared to O and the halogens, at least that is my impression—so far—of the descriptive chemistry involved. Nitrogen's ability to catenate strikes me as rather impressive, in comparison to oxygen, following the tour de force synthesis of the N₅⁺ pentazenium cation in bulk quantities, and the related existence of the cyclo-N₅^– pentazolide anion. I see the N catenation count is up to at least eleven (I was impressed by even the eight mentioned by Jones), whereas I thought the effective limit for O was three?

Earlier you referred to the personality of nitrogen being more like that of a "psycho" nonmetal and "it is simply remaining functional by keeping itself on a tight leash in the elemental state with the extremely strong N≡N triple bond; this also makes its compounds tend to be rather "psycho", as they are usually endothermic for the same reason." I'm not sure we're talking about the same kinds of "psycho". When I refer to psycho nonmetals I mean the electron greedy, corrosive, thoroughly oxidising nonmetals that are antithetical to the electropositive metals represented by groups 1, 2, 3, and the Ln and An. Nitrogen is largely not that kind of psycho. Sandbh (talk) 04:27, 21 April 2017 (UTC)

Well, by the same token, I'm not sure I would think of Sc and Y as "psycho" on the metallic side either; and if Be is "psycho" then surely so is Al, and I would actually strongly consider moving it there. There are always trends; short of some of the lanthanides we can hardly expect elements in the same category to behave exactly like each other. But as one goes towards the left and down the psychoness of the metals goes up, and it is simply that Sc and Y have more in common with those "psycho" metals than any others even if they are on a tighter leash thanks to their smaller size.

I reckon the same is going on for N, as psychoness increases up and to the right, except for the farthest right in the tea room. So it is clearly not as psycho as O, F, Cl, Br, and I, but the psychoness is already mostly there. The synthesis of all those chains of nitrogen atoms is impressive, but given how much of a tour de force it is I don't think it's anywhere near C, or even P or S. Even oxygen forms chains of five, admittedly in unstable intermediates! This is not characteristic like dinitrogen and dioxygen are, though. Double sharp (talk) 04:51, 21 April 2017 (UTC)

There is a reasonable case for considering Sc and Y as hyper in light of their low electronegativity values of 1.36 and 1.22, low standard electrode potentials of –2.03 and –2.37, and similarities in their chemistry with the group 1 and 2 metals, and Ln. I'm OK with treating Be and Al as hyper metals, partly because they have reasonably low standard electrode potentials of –2.03 and –1.65, partly because of similarities in their chemistry with Sc and Y, and partly because of their similarity to Mg in that beryllium, magnesium and aluminium make up the only pre-transition metals with structural applications. (Off the record, one could consider the hyper metals as encompassing the pre-transition metals, rare earths, and actinides).

I struggle with N as a psycho nonmetal. It has no peers among the regular psycho nonmetals. Its electronegativity is misleadingly high (Phillips and Williams 1965, p. 609). It has no electron affinity. In the oxidation stakes it is outclassed by sulfur. Mellor's nonmetal displacement series (Parkes & Mellor 1939, p. 205) puts S ahead of N, as does that of Ashford (1967, p. 312). For stable species in aqueous solution,^ while nitrogen as the NO₃^– nitrate anion is a reasonably effective oxidizing agent i.e., 2NO₃^– → N₂ = 1.25 V, it is surpassed by sulfur as the S₂O₈^–2 peroxydisulfate anion i.e., S₂O₈^–2 → 2HSO₄^– = 2.06 (Wulfsberg 2000, p. 247; Schweitzer & Pesterfield 2010, pp. 200–202, 228).

^ pH 0, –3.0 to 3.0 V

N has more peers among the reputable nonmetals. The phosphides, for example, "resemble in many ways the metal borides, carbides, and nitrides" (Greenwood & Earnshaw 1998, p. 490). Fluorides, on the other hand, "frequently adopt the 3D 'ionic' structures typical of oxides" (op. cit., p. 819). Consistent with periodic trends, the chlorides, bromides and iodides more often show sulfide-like layer-lattices or chain structures, presumably in the order Cl < Br < I, i.e. the heavier the corrosive psycho nonmetal the more respectable it becomes. Sandbh (talk) 10:56, 22 April 2017 (UTC)

Ashford TA 1967, The physical sciences, 2nd ed., Holt, Reinhart and Winston, New York

Greenwood NN & Earnshaw A 1998, Chemistry of the elements, 2nd ed., Butterworth-Heinemann, Oxford

Parkes GD & Mellor JW 1939, Mellor's modern inorganic chemistry, Longmans, Green and Co., London

Phillips CSG & Williams RJP 1965, Inorganic chemistry, vol. 1, Principles and non-metals, Clarendon Press, Oxford

Schweitzer GK & Pesterfield LL 2010, The aqueous chemistry of the elements, Oxford University Press, Oxford

Wulfsberg G 2000, Inorganic chemistry, University Science Books, Sausalito, California

Indeed electronegativity isn't everything here, but I must confess that I am unsure as to why you're treating it as more important for classifying Sc and Y than for classifying N. (I advance this of course with the greatest caution, given your generally excellent level of scholarship!) For one thing, if we adopt the electronegativity of Al as a sort of "upper limit", then it looks as though we have to admit not only group 3 and the f-block to the "mental hospital", but also group 4, Nb and Ta, and Mn(!), and V, Cr, Zn, Cd(!), and Tl(!) are really not too far away from the magic figure of 1.61 either. Furthermore, I am not sure if I would dare to call the early actinides "psycho". They only resemble the lanthanides when they are in low oxidation states; in their more characteristic higher oxidation states they are more like the d-block "reputable metals". In fact, the more I look at all the conflicting trends that the periodic table puts togeter not cumulatively but synthetically, the more I think that the chemically most defensible categorisation of elements would be to throw up one's hands in defeat and cover them group by group. Of course, this would be utterly boring, and there is wisdom in the verdict of the masses; and I shall proceed to discuss one of the sources of this wisdom.

Just like Sc and Y have similarities chemically with groups 1 and 2 and the Ln and An, but are significantly less active (along the lines of Be and Al), I would argue that N is probably in the same boat when compared to O and the halogens. The similarities then would outweigh the differences, since every category needs to have an epitome (e.g. Ca or Cl), an overenthusiastic member (e.g. Cs or F), and a runt (e.g. Be or N) which nevertheless shows its belonging by being more similar to its chosen category than any other. Incidentally, the refractory transition metals tend to form non-stoichiometric cluster halides, like MX_~3 (X = Cl, Br, I; M = Nb, Ta, W), where the now small metal cations simply take up positions within the hcp lattice of the halide anions, in a sort of inverse of the situation of interstitial borides, carbides, and nitrides; so this seems to be not so much based on the elements themselves which are involved, but rather their differences in size and electronegativity. It doesn't seem to be an even trend of F < Cl < Br < I so much as F <<< Cl, Br, I. The famous psychoness of compounds like ClF₃ ("The list of elements it sets on fire is diverse" – myself on the chlorine article), BrF₅, and IF₇ has more to do with them being fluorinating agents than anything else, so I'd ascribe it to F rather than Cl, Br, or I.

In fact, sometimes I wonder if the only nonmetals who have clearly cast off all links to sanity are oxygen and fluorine, the ones who are willing to commit the mortal sin of jumping into the noble gases' tea ceremony without being invited. ^_^ There must be a reason why it is so illuminating to classify elements by the acidity or basicity of their oxides: the two exceptions which cannot be treated this way are oxygen itself and fluorine, which oxidises oxygen.

I'll also jump in to note that peroxodisulfate is beaten as an oxidising agent by hyponitrous acid, a compound of nitrogen: the standard reduction potential of the H₂N₂O₂/N₂ couple is +2.65 V. Double sharp (talk) 13:54, 22 April 2017 (UTC)

I didn't think I treated the electronegativity of Sc and Y, for classification purposes, as more important than that of N. In both cases I attempted to take account of their electronegativity values, electrode potentials, and general chemistry.

From group 4 onwards we enter transitional metal chemistry territory proper, so I wouldn't consider admitting any of the other metals you mentioned into the psycho ward. Certainly, there are some reactive metals here but I'm being pragmatic in terms of trying to maintain a reasonable distinction between different kinds of chemistry.

You are right about the early actinides. I'd give them a backstage pass to the psycho ward in recognition of their reactive nature and the similarity of their chemistry in the +3 state to that of the Ln.

I'd regard iodine as the runt of the corrosive nonmetals just as (maybe) selenium could be regarded as the runt of the respectable metals. I appreciate it's possible to drill down and find interesting aspects of any categorisation scheme. But I'm trying to keep things at the broad generalisation level, for want of avoiding a sea of hands up in the air.

I based the comparison of standard reduction potentials on stable species in aqueous solution. I understand hyponitrous acid is unstable in aqueous solution and, going by the sodium hyponitrite article, that the hyponitrite ion in solution decomposes. There is a silver hyponitrite but this is slowly decomposed by light. Solutions of peroxodisulfates, on the other hand, are relatively stable (Wiberg 1995, p. 558), and I see that the sodium, potassium, and ammonium salts (that last of these being a standard ingredient in hair bleach---eek!) are stable or appear to be so.

Please don't jump into the aqua acidum! I prefer Double sharp to Semi blunt. Sandbh (talk) 11:59, 23 April 2017 (UTC)

Transition metal territory certainly, but for Zr and Hf it is fairly weak: Zr^III and Hf^III oxidise water and are mostly known solely as the trihalides. As for the chemistry of the actinides in the +3 state, this is not even very well-defined for Th and Pa, and still for U it is thermodynamically unstable and gets more stable as atomic number increases, until we get to Am and it henceforth becomes the normal oxidation state in aqueous solution for everything but No. So relying on those for a classification seems to be on about the same level as using hyponitrous acid. ^_^

Of course, all of these are slowly moving trends, and one could pick a division a group earlier or later for various reasons. When dividing the pre-transition metals from the transition metals, group 3 is a sensible place to do it, but one could choose an earlier division at group 2, or a later one swerving to Zr, Hf, and Rf, for slightly different reasons. The ideal place to mark a division would be when all characteristic properties of the elements change in sync; that is of course completely impossible outside the noble gases. But we can have the other divisions aspire to that high level of clarity. Choosing to divide at a certain place indicates that you believe that the discontinuities are more important than the continuities there, and that you can back it up. If all else fails, the elements you put on one side should have more in common with the elements on their side than the elements on the other side.

I'm not really convinced about nitrogen being on the less reactive side. It only seems to do that when it is left alone as individual N₂ molecules; once you do actual chemistry with it, instead of leaving it to pretend to be a noble gas and sip tea, the similarities with oxygen and fluorine seem more striking, with hydrogen bonding playing a minor major ^{oops! Double sharp (talk) 15:13, 28 June 2017 (UTC)} role, and a greatly reduced tendency towards catenation compared with C, P, and S. In that respect they are kind of like Sc, Y, and some of the later and more unreactive lanthanides; they are like their fellows when they let their hair down, but it takes some effort for them to do it.

As for nitrogen having no electron affinity, this is purely because of its half-filled shell, without any relativistic corrections allowing for it like in the heavier pnictogens; beryllium, manganese, and mercury do not have an electron affinity either, while carbon, sulfur, and selenium do. Double sharp (talk) 15:17, 23 April 2017 (UTC)

P.S. Might I add that the fact that such a compound is strong enough an oxidising agent to decompose water seems to be evidence for the (perhaps somewhat repressed) psychoness of nitrogen? As another example, it is not as if F₂ is something you can get in water, and the huge aqueous couples (F₂/2F⁻, 2.87 V; F₂,2H⁺/2HF, 3.06 V) bear witness to fluorine's ferocity. Double sharp (talk) 03:31, 1 May 2017 (UTC)

On thinking about this a little more, I absolutely do think that the reason why N is commonly thought of as more akin to O and F than to C is its chemistry. But I might draw the line between the 2nd and 3rd periods, instead of zigzagging like the nearby metalloid line. The high electronegativity of these elements is not misleading at all when you consider how it reflects in their lack of hypervalent compounds (unwillingness to enter ionic bonds when they carry the significantly positive charge). This has a lot to do with the similar size of the 2s and 2p orbitals, so that they easily hybridise with each other; when there are lone pairs around (carbenes, hydrazine, peroxides, fluorine gas), the repulsion between electrons tends to significantly weaken the bonding. So if you look at N₂ alone, it does not look very much like O₂ and F₂; but its compounds show the same sorts of effects. This is not even that uncommon in the table: if you look at Br and I vs Kr and Xe, for instance, there couldn't be a starker contrast in chemical activity. Yet the chemistry of Kr and Xe in the +2, +4, +6, and +8 oxidation states quite clearly parallels that of Br and I in the +1, +3, +5, and +7 states. Just as you have "hydroxides" of tellurium and iodine that are really oxoacids, so you have them of xenon too, completing the series; and you can even get a good idea of tin and antimony by positing similar compounds for them as precursors (in fact Sn seems to have a strong case for being put in the nonmetallic class for chemical reasons; if you focus on 273 K, it even has good physical reasons too)!

This sort of thing, with period and group trends acting and continuing even beyond metal-nonmetal boundaries, is why I think the great comprehensive inorganic chemistry texts all treat the elements by groups (and occasionally periods to separate the first-row anomalies for B, C, N, and O and the zeroth-row anomaly for H), treating metallicity as an afterthought. Given its lack of importance for chemistry rather than physics, perhaps we should leave the distinction mostly to physics, where the metalloids emerge as more clearly defined territory. Double sharp (talk) 06:48, 15 May 2017 (UTC)

Pressing pause

I'd like to pause our exchange, due to RL commitments, and resume proceedings in early to mid-June. Sandbh (talk) 08:48, 24 April 2017 (UTC)

No problem; please take as much time as you like! At the most I might fill up that promised post on metallicity in the metalloids that is now a placeholder, but I won't expect a reply anytime soon of course. Double sharp (talk) 09:21, 24 April 2017 (UTC)

Update. I've returned from my break and intend to resume participation in this thread shortly. I'm delighted to see contributions from R8R, Double sharp, and YBG, none of which I've read yet. I thought a fair bit about the proposal during the break and expect to be able to post a refinement shortly, subject to considering the new posts. Sandbh (talk) 02:05, 9 June 2017 (UTC)

I just read through the most recent posts and am now considering how to respond to the various items raised therein. There are a fair number I agree with or have previously thought about. I just need to gather everything together and respond in a methodical fashion. This will take a little while. Sandbh (talk) 04:55, 10 June 2017 (UTC)

Let's create an article for this

This discussion will be talkpage-archived, ok. Let's also create an article for this topic.

Could someone (Sandbh, Double sharp) make a setup? Title suggestion? -DePiep (talk) 19:33, 17 April 2017 (UTC)

I didn't know articles could be created out of talk page threads. Is there an example of such a thing elsewhere? My title suggestion is Nonmetal (subtypes including metalloids). Or are you saying the talk page is archived and that we create a new encyclopedia-style article on this topic? Sandbh (talk) 02:18, 18 April 2017 (UTC)

Yes, we archive the talk page, but perhaps the idea of including metalloids as nonmetals could be expounded upon further in articles. Given that it seems to make sense from the names of the categories (albeit we now have oganesson as a noble gas, though it is probably neither noble nor a gas), I would be very surprised if nobody else has done this. Isn't it a common practice to divide up the metalloids among the metals and nonmetals anyway, depending on which side of the line they are on? If you swapped Ge and As it would be as perfect as you could get with a dichotomy instead of a trichotomy. Double sharp (talk) 12:08, 18 April 2017 (UTC)

• Now what to do with all this? Someone draw a conclusion? -DePiep (talk) 23:13, 19 May 2017 (UTC)

This talkpage is 333k now. At least, Sandbh and Double sharp, could you make a plan on how to proceed? Hard to see any fun ahead. -DePiep (talk) 23:17, 19 May 2017 (UTC)

The chat has not yet finished. It is on hold due to my RL commitments. Proceedings should resume by mid-June. Sandbh (talk) 01:11, 20 May 2017 (UTC)

Given the natural hiatus in the discussion here, this would strike me as the best possible archive point. Everything before this section is pretty old; everything after it is very recent and is still actively being discussed. Double sharp (talk) 08:21, 13 June 2017 (UTC)

R8R's comment

Sorry for not writing for so long. I've wanted to write a review for this for a long while now. I've set up a draft for it but I clearly don't like it myself because my reaction to the proposal seems to get constantly reduced to one essential point. The attempt is good; clearly, "corrosive nonmetals" is a good name, I like it very much. The problem is that, though, it comes in a bundle with "intermediate nonmetals," which is terrible because it means nothing by itself: you have to know intermediate between what and what. "Mild nonemtals" is not a complete name, either: you need to know in which respect it is "mild." Compare with "alkali metals": metals that form alkalies. As simple as that. Of course, we have "transition metal," one incomplete name, too, but only because this stuck really well. "Transition metals" are a thing: you don't break this term into parts; these two words are one unit, if that's appropriate to say. "Mild nonmetals" are nonmetals that are mild (a combination of two terms rather than one and that's the problem).

By no means I am saying that the idea is bad in general because if there was a good solution, we'd already have it, and this could be it, there is merit. Yet such a change should not start in Wikipedia, which is meant to reflect what others say and not say for itself. If there should be a nonmetal divide, I'd say this was a decent attempt but we'll have to try again. Yet, in our case, I believe that should be sticking to the well-known terms or, when not available, minimize the number of those that are not and make that as understandable as possible. (Again, I'm not saying that the solution is bad or inappropriate at all; it is not great for Wiki only. I could see this in a printed book, for example.)

In that respect, I'd suggest to not look for a nonmetal divide at all. First of all, 10 elements is not too many anyway and any line we draw is sort of arbitrary (it is by no means clear if sulfur and iodine belong together; what about sulfur and selenium?). Also, essentially, any proposal that would come up to my mind divides into "strong nonmetals" and "mild nonmetals" and I can't seem to escape this Wiki-related problem with that latter. However, we can agree that we have a medium-size diverse (scandium, gold, and mercury are quite different, too) group of elements. If we only have "reactive nonmetals," that's one easily describable term: nonmetals that (commonly) react.--R8R (talk) 13:31, 7 June 2017 (UTC)

Let me jump in and analyse one thing I find really important here: why is "transition metals" fine? The name is not too good, yes, but it is a fairly sharp category: one of the most clear-cut distinctions is between main-group element chemistry and transition element chemistry. Yes, some of the actinides are weirdly on the border here (Pa, U, Np, Pu, and Am), and Zr and Hf are not very good transition metals (though Zr^III and Hf^III are undoubtedly well-defined in halides), but for the most part you can point to some very characteristic properties that distinguish one category from the other.

I don't see how you can separate "mild nonmetals" from "strong nonmetals" anywhere near that well; I can't think of a "smoking gun" property that only one category has and not the other (and this, I suspect, is partly related to how small a sample size we are dealing with). Actually, the only really clear-cut line I could draw in the nonmetals would be the line cutting period 2 away from period 3, and that only if you forced me to draw one.

Incidentally, I also don't like "reactive nonmetals". This would seem to exclude nitrogen, even though a common motif in nitrogen chemistry is the great stability of dinitrogen, so that most nitrogen compounds are very reactive in a bid to return to being nitrogen. I think we're focusing too much on elements in the free state rather than elements in compounds, which are surely part of chemistry as well: who would not claim NaCl as the province of sodium and chlorine chemistry, even though obviously it reacts like neither elemental Na nor elemental Cl? As a result we overemphasise fluorine-like reactivity (where the element is unstable and wants to form compounds) and underemphasise nitrogen-like reactivity (where the compounds are unstable and want to return to being the element), even though they create almost the same redox potentials. It is just that nitrogen achieves +2.85 V for reducing hyponitrous acid to elemental nitrogen, while fluorine achieves +2.87 V for reducing elemental fluorine to fluoride. Double sharp (talk) 15:44, 8 June 2017 (UTC)

I don't quite follow that last one. When you said about nitrogen bot being too reactive in its free state, my first response to that was, "we're talking about elements and free substances"; then I saw you recognized that one. So, hmmm, I don't know. You're essentially naming the problem that my proposal seems to be not falling into it and then you dismiss it on this basis. I don't see the logical pattern here.

By the way, I'm glad to know you follow my main point about not dividing the nonmetals. "Reactive nonmetals" is, in my opinion, the best name for this category, because we only have two categories, one being "noble gases" (i.e., nonmetals that don't react) and the antithesis seems obvious here. I'm open for suggestions, though.--R8R (talk) 16:17, 8 June 2017 (UTC)

Yeah, it's a bit of a separate thing. It just makes me a little uncomfortable to say that N is unreactive when its chemistry is so famously explosive, which makes me think that perhaps we should consider how the elements behave in forming compounds as well as in their free states. Aren't the names "alkali metal", "halogen", and "noble gas" predicated on forming compounds, anyway? They're called alkali metals because they, well, form alkalis when they're dissolved in water; but then you don't have M, you have MOH. They're called halogens because they readily form ionic salts, but then you don't have X, you have X⁻ with some counter-ion. They're called noble gases because they don't like forming compounds, but they're not called inert gases because Xe has quite a rich chemistry, and those of Kr and Rn are emerging but still clearly well-defined.

Put like that, I think the polyatomic/diatomic distinction puts N with more similar elements than the proposed one. Mind you, this only means that I think the current scheme is better than the proposed scheme (because it's correct, doesn't require names that mean nothing by themselves, and generally splits the group about as well as could be expected); I am not sure if splitting the nonmetals is a good thing at all, but am willing to accept that the current scheme might be the "least-worst" way of doing it. Double sharp (talk) 16:23, 8 June 2017 (UTC)

A few notes about nitrogen

Incidentally, I think it's pretty unfair to hold nitrogen's negative electron affinity against it. This is purely because of its electron configuration; being 1s²2s²2p³, any added electron would either have to pair itself in one of the 2p orbitals (and then inter-electron repulsion makes this unfavourable), or it would have to go into 3s or higher (and then the high energy makes this unfavourable). You could say the same about species like O⁻; oxygen's second electron affinity is of course negative, as the repulsion between the like negative charges of O⁻ and the incoming electron is high. And yet oxygen quite readily forms O²⁻ – that is, if your metal is electropositive enough. If it isn't, you may end up with situations like NiO (a semiconductor) or ReO₃ (a good metallic conductor). Similarly, if your element is highly electropositive, you will get salt-like nitrides like Li₃N with what can be considered N³⁻ ions. There is not a clear discontinuity here between this and oxygen; the increased difficulty of forming ionic nitrides simply follows what one would expect from the −3 charge instead of the −2 charge. Instead there seems to be a discontinuity between N and P, just like between O and S. Even carbon can just about form C⁴⁻ in Be₂C, but this is pretty much the limit. So we get a smooth trend from C to N to O, but even with oxygen metallic or covalent oxides are still quite possible for many metals; it's not like fluorine where you need really high oxidation states and exceptionally weakly electropositive metals to create covalence like for MoF₅ and PtF₆ (and even then it might not be enough, as seen in the case of PbF₄).

It thus strikes me that if we are looking for nonmetals which are clearly strong and almost always form ionic compounds, only the halogens seem to pass the bar, and then only because of their low −1 charge when they gain their coveted noble-gas configuration.

As for the "misleadingly" high electronegativity of nitrogen, it is not so misleading when one considers that N compounds have very distinct effects caused by hydrogen bonding. Cl has the same Pauling electronegativity, but those effects are hardly noticeable there. By this measure, the electronegativity of N seems to misleadingly low instead of high; and the fact that H bonding is very much a first-row thing further makes me think that the main division in the p-block is between the 2p and 3p elements, and not so much along the metalloid line. Double sharp (talk) 09:58, 11 June 2017 (UTC)

Groups trumping metallicity

Sandbh wrote to me about this by email, and it got me thinking: given how the p-block is pretty much characterised by a head-on collision between horizontal and vertical trends, why is it that pretty much all authorities divide the table by groups?

Everyone tacitly acknowledges the problem. You can't write a homogeneous article about group 16 because the family segments immediately. On the light end, the youngest sister oxygen is very different from sulfur, selenium, and tellurium, a fairly homogeneous set of triplets. On the heavy end, the eldest sister polonium appears to suddenly suffer a discontinuous metallisation catastrophe (skipping over the superheavy elements). But somehow, everyone covers these elements together in consecutive chapters, even if they can't do it in just one.

Let's step back for a second. What does an element's placement tell us about its chemistry? Surely it tells us about the ways in which it can achieve a stable octet or octadecet configuration, or perhaps through hypervalence ignore this (yes, yes, I know d-orbitals are not involved, but I'm going to keep saying they are because it's a fine first approximation even if it's completely wrong). From oxygen's placement in group 16, we know that it is short two electrons; so we know that its main modes of getting the octet it so dearly covets are through forming a dinegative ion, or through forming (2+n) covalent bonds compensated by the loss of n electrons (allowing negative values of n; actually, this encompasses the first case as well). And that's all, to a first approximation. Its position in the second row tells us that hypervalence is not to be expected, and neither are high coordination numbers, but we would expect this for S, Se, Te, and Po.

What is particularly striking is that when forming analogous compounds, it does not seem to matter whether the elements forming them are physically metals or not in the elemental state. Let's consider Si, Ge, Sn, and Pb. How can we summarise their chemistry? First of all, they all have four valence electrons and thus have the same number of valence electrons as available orbitals, hence forming the "covalent divide" together with C. Now what can we say about the trends? Electronegativity decreases down the group as Si→Pb. Fine, and this happens to pass through the metalloid line, but the same thing is going on for Mg→Ba and that doesn't. The important thing here is the shift between ionic and covalent bonding, and that is shown in a decrease of strength in the M–M bonds and hence lessened catenation.

But wait: ionicity vs covalence is not only predicated on how metallic an element is, but also by the value of its main oxidation state. This is why the early transition metals are such bad metals chemically; you cannot expect much cationic chemistry if the states you are looking at are things like Ta^V, W^VI, and Re^VII. So we can just as well explain the increasing metallicity down the group purely via appealing to the inert pair effect, can we not, noting how it changes the main oxidation state involved? We can also appeal, for the lessened stability of highly negative oxidation states, by the need for the nucleus to be really close to handle such an overload of negative charge (and even that is not really enough to save H⁻ and C⁴⁻ for the most part); but that is again simply a gap between the first-row elements and the ones below, which works for Li/Na and Be/Mg just as surely as it does for the other pairs in the p-block. And if one considers hypervalence and stereochemistry, that is certainly characteristic of Si–Pb in a way that emphatically isn't for C, but this is once again a divide between the first-row elements and the ones below. Nowhere does metallicity get involved in it; and if one wants to bring cationic chemistry and the "reality" of metalloid salts, one is then confronted with the fact that Sn^IV is more like Sb and Te, while Sn^II is more like Pb, proving once again that it's the oxidation state that is running the show. And that is predicated on the group.

I would not dare to say that this is the magic key for reading the periodic table, and why everyone seems to organise it that way over the physical resemblances, but I think we are getting closer to the answer. Double sharp (talk) 16:17, 8 June 2017 (UTC)

@Double sharp: makes an interesting point about the conflict between horizontal vs. vertical trends in the nonmetals. Think about it as a battle that fought to a stalemate or an argument badly in need of an arbiter. Without thinking about chemistry or even elements, the simple minded solution is to propose "diagonal" as the compromise or stalemate result. Then think about the PT, and what do we see in running across the right hand side of the PT but a diagonal battle scar, which is sometimes displayed as though it is completely closed up and healed, and other times displayed as though it is still an open and festering wound.

What about the following categorization?

1. Elements He, Ne, Ar, Kr, Xe, Rn

   Proposed name: Unreactive nonmetals (noble gases)

   Other possibilities: Noble gases, Unreactive nonmetals, Inactive nonmetals, Noble gases (unreactive nonmetals)

2. Elements H, C, N, O, F, P, S, Cl, Se, Br, I

   Proposed name: Reactive nonmetals

   Other possibilities: Active nonmetals, Chemically active nonmetals

3. Elements B, Si, Ge, As, Sb, Te, At

   Proposed name: Borderline nonmetals (metalloids)

   Other possibilities: Metalloids, Weak nonmetals, Weak nonmetals (metalloids), Borderline nonmetals, Metalloids (borderline nonmetals)

There are two issues about this that need discussion

• Is this (i.e., the lists of elements) a good categorization?
• What would be the best name to use for each category?

Some of this overlaps with what has been stated elsewhere above, but I thought it easiest to write from scratch rather than respond to specific other parts of this thread. YBG (talk) 23:23, 8 June 2017 (UTC)

The following are the issues I would want to consider carefully:

1. Is metallicity really the best way to categorise the post-transition elements (groups 12 through 18) for our purposes?
2. If so, should we consider the metalloids a subset of the nonmetals?
3. Should we split the current set of nonmetals at all, leaving aside group 18 which is unique?
4. If so, how? (Which elements fall on which side?)

The more I think about it, the more I want to answer only a qualified "yes" to (1). While I would consider metallicity important from the physical perspective, for chemistry it seems to be at best a needless distraction from the group trends. The early B subgroup is by no means alone in having physical and chemical properties work against each other: much the same happens in the early A subgroup as well.

Now here's a really radical proposal: get rid of all the colours. The group numbers tell us everything we need for chemistry. (^_^) I am partially proposing this tongue-in-cheek and partially seriously. At least, I would like to ask, to clarify our thinking: I think most of us would consider this a bad idea, but why?

Also, is there a reason to put physical properties on the same footing as chemical properties? The former only apply to the elemental state; whereas the latter apply to the vast amount of compounds formed by the element. Surely the latter should triumph for the most part. Now I realise that this stand is going to force me to mount a full chemistry-based defence of B, Si, Ge, As, Sb, and Te as showing significant metallic behaviour. Well, one thing I'm planning and writing is a look through each period: perhaps I'll start by doing {Cd, In, Sn, Sb, Te, I, (Xe)}, and then work my way up. Double sharp (talk) 14:13, 9 June 2017 (UTC)

Refined proposal, Mk. 2

During my break I concluded it would be much simpler to retain the category name "metalloid" instead of replacing it with "metalloid (weak nonmetal)". I note DePiep in fact made this suggestion early on. So the informal proposal now looks like this (Mk. 2):

Metal						Nonmetal				Unknown chemical properties
Alkali metal	Alkaline earth metal	Lan­thanide	Actinide	Transition metal	Post-​transition metal	Metalloid	Intermediate nonmetal	Corrosive nonmetal	Noble gas

Hereafter are some notes explaining the refined proposal; my responses to earlier relevant comments by R8R, Double sharp, and YBG; and comments about my own concerns. Thank you DePiep for your help with the basic html code for the legend.

Notes

We categorise the elements based on their overall chemical properties hence the already existing "unknown chemical properties" super-category at the end.

The elements most commonly designated as metalloids are B, Si, Ge, As, Sb, and Te. Per Double sharp's observation, the term metalloid gives us a warning that while these elements might look like metals, they don't act like metals. Indeed, the literature (here, 1894–2016) records that their chemistry is generally nonmetallic. Even so they are the most metallic of the nonmetals.

The intermediate nonmetals are H, C, N, P, S, and Se. They are neither as metallic as the metalloids nor as chemically active as the corrosive nonmetals. This is a better explanation of the intermediate nature of these nonmetals.

The corrosive nonmetals are O, F, Cl, Br and I. Their category name speaks for itself.

Summary of R8R comments

• "intermediate nonmetals" is terrible as it means nothing
• say what others say
• use well-known terms if possible; minimise those that are not; make them as understandable as possible
• suggest no divide; any divide is sort of arbitrary

Response
I mostly agree with DePiep who said

[the name] is not the primary issue in categorising. Category name and meaning is secondary, and that's what links are for. Primary is: group together what belongs together, and do not include stuff that does not include there. Just give the group a unique color and name, that's enough. The name can be as fancy or invented, no intrinsic meaning required, no prior knowledge for the reader required.

Since we write for the general reader not the technical specialist, the name "intermediate nonmetal" is as good as "alkaline earth metal" or "lanthanide". In fact the general reader is likely to be more comfortable with the word "intermediate" than they would be with "alkaline earth" or "lanthanide". "Transition metal" is no better for the general reader. The general reader thinks, "a metal that is transitonal", and then, "transitional between what?". But, as DePiep notes, this largely does not matter.

I agree we should seek to use what others say. The quotes listed here are examples of terms used by authors to refer to nonmetals:

weak very weak electronegative more strongly electronegative intense electronegative weakly electronegative weakly nonmetallic somewhat electronegative (nonmetallic)           less active; active; inert weakly active to highly active

moderately active strongly nonmetallic mild strong weakly nonmetallic to strongly nonmetallic strongly nonmetallic highly reactive less reactive nonmetals

In this context, "intermediate" is a reasonable equivalent for terms such as "somewhat", "moderately" or "mild".

I do not agree that we could use invented names, and I agree with you that we should use what others say, where possible. If this is not possible it is OK to use plain English descriptive phrases. It is good to make these as understandable as possible but this aspiration needs to be balanced against the need to avoid introducing new terminology or having to use cumbersome, too long terminology. Clearly, many of the authors whose terms I listed above where comfortable using quite short, plain English words.

It is not true that any divide is arbitrary. The current divide between polyatomic and diatomic nonmetals is not arbitrary. The proposed divide is based on a natural division between metalloids and corrosive nonmetals.

Double sharp writes: Okay, but those terms used by the authors make sense: they not only say that something is moderate, but also what is moderate, whether it be reactivity, electronegativity, or metallicity. The proposed "intermediate nonmetals" does not tell us what exactly is intermediate about them.

For sure, "alkali metal", "lanthanide", "transition metal" and the rest are opaque to a reader who is clueless about chemistry. But these are standard terms, so that if s/he is interested enough to read on the matter, s/he will have to become familiar with these terms. Not so with intermediate nonmetal. It is certainly an English phrase, instead of a single term by itself. But by putting it there with other phrases, particularly as it never tells us what is intermediate about these nonmetals, the reader might easily think of it as a standard term with Wikipedia's imprimatur, as has happened several times with "poor metals".

Incidentally, this is why I don't criticise transition metals: it is clearly a standard term, and there is a clear divide between transition and non-transition chemistry. I do not see a very clear divide between N and O or between Se and Te that is anything like the clear divide between Cu and Zn. Actually I do not see one between Sb and Bi either, and I am refraining from mentioning Po and At because we have only tracer chemistry for them for the most part. One thing I am curious about is the tracer chemistry of iodine: from what I have heard it is very different from the usual chemistry of the element in bulk, which is of course that of an orthodox halogen. Is it anything like what we know of the tracer chemistry of astatine?

And one other thing I find slightly problematic here is the subsuming of metalloids as a subclass of nonmetals, even in their placement in the legend. Firstly, you can quite readily get the chemistry of H and B by extrapolating upwards from Li and Al and asking what would happen if covalence was forced. And secondly, despite the prominent nonmetallic characteristics of these elements (many of which I would note are shared with the transition metals, such as a rich oxyanion chemistry and a reluctance to form simple cations, and can be put down to a general difference between low and high oxidation states instead), it does not seem particularly common to make this decision in the literature. If no one dares to take that step, then perhaps we should be cautious about it, and perhaps retreat to scouring the descriptive chemistry of these elements and searching for reasons, as I am sure Greenwood & Earnshaw for example must have had. Double sharp (talk) 03:51, 12 June 2017 (UTC)

P.S. Indeed, in trace quantities iodine is only partially extracted from organic solvents by NaOH, just like At! (10.1063/1.1699161) Perhaps it really is significantly less metallic in bulk. Double sharp (talk) 06:48, 12 June 2017 (UTC)

P.P.S. Trace radioiodine appears to form a sort of radiocolloid that does not act very much like diatomic iodine (10.1007/BF02162882); given the sparsity of iodine atoms produced this way with an atom-at-a-time procedure, this also seems fairly reasonable. It seems that trace iodine behaves a lot more like trace astatine than bulk iodine, which perhaps should give us caution that bulk astatine might be more like a halogen and less like a metal than trace astatine. Tennessine is more likely to be a true metal in bulk, as even non-relativistic extrapolations would seem to predict its metallisation, but I would be hesitant to say anything more given the lack of chemical studies. There is a good reason why the chemistry of Cn was compared with that of trace quantities of Hg and Rn, and not with the behaviour of bulk Hg and Rn. Double sharp (talk) 08:53, 12 June 2017 (UTC)

Response by Sandbh to Double sharp's comments of 03:51, 12 June 2017

A careful reading of the above terms used by some of the authors in fact shows they are quite comfortable with referring to nonmetals in the purest of terms i.e. "weak nonmetals", "weakly nonmetallic", "strongly nonmetallic", "mild nonmetal", "strong nonmetal", "weakly to strongly nonmetallic", "strongly nonmetallic".

This is entirely proper and appropriate.

It is preferable or at least satisfactory to NOT combine adjectives with aspect-specific subjects like reactivity, electronegativity, or metallicity (which results in category names such as "moderately reactive") since nonmetallic character is a composite of many factors that contribute to, or are aspects of, a nonmetallic element's nonmetallic character.

When you add subjects it starts to become very tough to establish representative categories in which every member meets just that particular criterion. By "representative" I mean a category in which a single criterion correlates reasonably well with all the other factors that make up non-metallic character.

The other issue is that category names like "highly reactive" or "very electronegative", which I have previously proposed or mentioned, have been criticised for their subjectivity e.g. what is "highly"(?), or what is "very"(?).

So now I let the literature do the talking i.e. it is not me that has decided which elements are commonly recognised as metalloids nor is it me who has decided which elements are corrosive.

With respect to "intermediate nonmetal" being thought of as a standard term this will not happen, just as "other nonmetal" and "polyatomic nonmetal" have not become standard terms. Yes, I do see "poor metal" from time to time but this term has been around since before Wikipedia and hasn't become thought of as standard term. There will be no "intermediate nonmetal" article only a link to the nonmetal article. The nonmetal article will explain the sorry state of nonmetal category names, a bit like it does now, and that "intermediate nonmetal" is a descriptive phrase only.

As noted, the divide among the nonmetals is sourced directly from the literature.

Subsuming the metalloids as a subclass of nonmetals is not my doing, it is a natural outcome of the fact that our colour categories are based on the chemical properties of the elements, as I previously explained. And, as noted, while metalloids are reasonably well recognised as a category of elements, the literature has, since at least 1894(!), noted that metalloids generally behave chemically as nonmetals.

Hawley's Condensed Chemical Dictionary, which has been around since 1919, has been treating metalloids as the most metallic of nonmetals---without any fuss---for at least their last six editions (since 1977). The 2016 educational periodic table placemat published by Painless Learning Placemats in the USA (they only have a pre-2016 version on their web page, which does not have this feature) shows metalloids as a subclass of nonmetals. When I asked the publisher why they did this they said they asked IUPAC (!) for some good examples of periodic tables but that they (the publisher) had forgotten which periodic table it was that showed metalloids as being subsumed under nonmetals.

By all means keep "metalloid" as a category but don't elevate them to a status that is incompatible with the superficial way they are treated in the literature.

My standard disclaimer continues to apply. The refined proposal is not perfect, just like our current categorisation scheme is not perfect. But it is more consistent with the rest of our categorisation scheme, and the literature, and the progression in metallic to nonmetallic character as you go from left to right across the periodic table, and the way chemists think, which is in terms of where each nonmetal approximately lies on a composite scale (or spectrum) of nonmetallic "liveliness" or "intensity" rather than in terms of polyatomic or diatomic nonmetals. Sandbh (talk) 04:23, 14 June 2017 (UTC)

DS comment. I think, IIRC, that you will find "poor metals" as a category on many Internet beginning-chemistry resources, because Wikipedia gave that category an imprimatur for so long; and you will always find it referring there to the same elements {Al, Ga, In, Sn, Tl, Pb, Bi}, maybe sometimes including Po. The trouble is that most such pages do not go very far into chemistry at all; there never seem to be any calculations beyond the basic ones (working with moles, finding empirical formulae, limiting reagents, etc.), and if you see organic chemistry covered beyond nomenclature and perhaps aliphatic alkanes I should like to see it. Again, the problem is that the periodic table is too fundamental. We all know which categories are in serious use in science and which are not, but consider a reader just beginning to search things on Wikipedia: he does not have this knowledge. He does not have the context, by definition, to know what the commonly used terms in science are. So he takes us at face value, and thinks that "corrosive" and "reactive" nonmetals are inscribed in stone, and then has a severe case of cognitive dissonance when he finds otherwise. I daresay a lot of people think from the firm colouring that we are surer about astatine than we really are, but elements are not radioactive, rare Pokémon which each belong to exactly one type.

We talk about getting classifications from the literature, but the fact remains that by the time a reader gets to the level where she can read those chemistry books without her eyes glazing over, she already knows what a nonmetal is, and what properties are characteristic of them. Hence she knows that "nonmetal" is the base term, and that you can apply imprecise modifiers to it. But because we don't have that barrier, we cannot count on readers having this background knowledge. So our needs for a classification are a bit different from those books.

If we were to link "Reactive nonmetal" and "Corrosive nonmetal", that would be an improvement, but the reader not already in the know will be surprised by pathological edge cases like nitrogen (reactive in a slightly different sense) or iodine (corrosive but honestly not that reactive, certainly less so than sulfur; you can smell iodine and come out OK, which is certainly not the case for bromine just above it). If we link "Polyatomic nonmetal" and "Diatomic nonmetal", it would be even better; these names have the naïve certainty of the beginner combined with technical correctness, and the experienced reader will see how this correlates nicely though approximately with many chemical and physical properties. (I am inclined not to worry about ozone; after all, with enough pressure, everything is a metal, so we might as well confine ourselves to the elements in their standard states.)

As for the treatment of the metalloids, you give two sources treating metalloids explicitly as a subclass of the nonmetals; many others do not. Furthermore, it is not really true that metallicity increases monotonically from left to right in the periodic table: I daresay tungsten is less metallic than lead chemically. The generalisation I would substitute is instead: metallic character increases with higher atomic radius and lower oxidation state.

It would be remiss of me to make this generalisation without letting you put it to the test. As an example of the relevance of the oxidation state, I invite you to compare the metallicities of Cr^III and Cr^VI. ^_^ Does the latter really look that much more metallic than Se^VI or Te^VI, even though the former shows quite good metallic behaviour like Al^III? Now Se^IV and Te^IV have a lower oxidation state: do these look any more metallic? Now, to gauge the effects of size, how metallic do U^VI and U^IV look? And when you're done looking at the chemistry this way, how metallic do Cr, Se, Te, and U look in the elemental state, judging by their physical properties?

I think, therefore, that we are rather forced to admit that the existence of a separate "metalloid" band comes from the differing effects of active d and f electrons compared to p electrons, and these largely in physical properties. By all means there are different nonmetals, some sometimes stronger, some sometimes weaker. But you will not find a one-size-fits-all categorisation, and that is why I prefer a clearly defined categorisation that suggests most trends, to a fuzzy categorisation that tries to show a clear divide and neglects that it moves with the point you are trying to make. Double sharp (talk) 05:19, 15 June 2017 (UTC)

Sandbh responds: Alas, alack, I feel like the troubles of the internet are being attributed to Wikipedia.

I regret having to repeat myself: the category name "poor metals" predates its use on Wikipedia. We used it for a while then we got rid of it. So what? The whole of Wikipedia changes on a daily basis. In an interconnected world the relevance of worrying about the calibre of Internet beginning-resources based purely on a non-critical harvesting of Wikipedia content escapes me.

There is no rational nor reasonable basis to think that anything found in Wikipedia is necessarily inscribed in stone, and I accept no responsibility for the views of irrational folk that do. Even in science there are controversies, and developments in progress, and we note the status of these in the articles, and it is up to the reader to make of that content what they will. We explain the multiple category names that have been bandied about for the post-transition metals. We explain the uncertain status of astatine, at some length, in the astatine article. We explain the uncertainty around nonmetal categorisation in the nonmetal article. I have said we would explain the background and status of corrosive nonmetals, and intermediate nonmetals, in the nonmetal article, as we should. I am not sure what else we are expected to do. Maybe I should add more background to the polyatomic nonmetals category name, noting the use of the concept in this manner is a Wikipedia "best-intention" initiative, lest the innocent reader gains the impression that this category name is inscribed in stone (which it self-evidently isn't).

The innocent reader will not be surprised by pathological edge cases such as nitrogen or iodine, since the nonmetal article currently makes clear in its lead and main body that there is some variation and overlapping of properties within and across each category of nonmetal, and that the boundaries between the categories are not absolute. Indeed there is a rather large footnote effectively apologising for the seeming absence of meaningful polyatomic-associated character in sulfur. As I have previously observed about classification science, hard case at the edges are nothing unusual.

I hope the innocent reader is struck with a sense of wonder at the elements that are either anomalous given their category, or otherwise extraordinary, as the nonmetal article will explain, and as comprehensively set out in the anomalous properties section of the Properties of metals, metalloids and nonmetals article. Chemistry would be rather dull but for these rogue elements.

There may be an apparent "naïve certainty for the beginner" in the names "polyatomic nonmetal" and "diatomic normal". However I am not sure that the general reader would gain any more real certainty from these names as they would from "intermediate nonmetal" and "corrosive nonmetal". In my experience, when I explain to other people what I do on Wikipedia, the term "nonmetal" is hard enough for the general reader to understand, let alone the qualifying adjectives. "Intermediate" most will get straight away; ditto "corrosive". "Polyatomic" and "diatomic"? Impressive sounding names but for the general reader I may as well be speaking another language.

The experienced reader will appreciate the logic, fit and consistency of the "intermediate nonmetal" and "corrosive nonmetal" categories compared with the other categories, acknowledging I would still need to sandbox how this plays out in writing. Agree re ozone, it is one of those boundary things, like sulfur and iodine.

On iodine and sulfur I explained earlier that

"counting iodine in the same league as O, F, Cl and Br may raise an eyebrow. Then again, iodine is corrosive, has a pretty decent electronegativity (2.66), and its periodate ion is a formidable oxidising agent (stronger than the perchlorate ion, for example); even the iodate ion is a stronger oxidant than elemental bromine. And pragmatically speaking it makes more sense to keep iodine with its lighter halogen congeners."

Looking quickly at other likely indicators of nonmetallic character (ionisation energy, electron affinity, average electrode potential, enthalpy of dissociation) I see that iodine is ahead of sulfur on each of these. I don't pretend that the numbers are so clear in every other case, nor that they tell the full story, but there they are, in this case. As well, every nonmetal activity series I have seen places iodine ahead of sulfur.

Those two source treating metalloids as nonmetals are particularly notable given one is an encyclopaedia of chemistry in which the editors actively seek feedback from the chemical community, yet no-one has batted an eyelid about the treatment of metalloids as nonmetals during the last six editions, from 1977 to 2016, under different editors. The other source was notable in that it resulted from an IUPAC referral, of all things.

These sources don't tell the full story.

There are the other multiple sources I referred to earlier noting the nonmetallic chemistry of the metalloids.

Yes, quite a few authors divide the elements into metals, metalloids, and nonmetals but the descriptive chemistry provided for the metalloids is inevitably (generally) nonmetallic, acknowledging that that metalloids are nevertheless the most metallic of the nonmetals e.g. they make regular appearances in the organometallic literature (but even hydrogen, phosphorus, and selenium are sometimes seen in the organometallic literature, which is another boundary phenomenon).

Thus, much of the literature is inconsistent. We can note this in the metalloid article but we need to make a decision on how to deal with it. Accuracy, Science, Pragmatism, Educational value and Professionalism (they are watching me as I type this) say we should treat metalloids as a subclass of nonmetals, consistent with the general chemistry at hand.

As I explained, the top row of our periodic table legend divides the elements into super-categories according to their chemical properties, hence our 4th super category of "unknown chemical properties. Metalloid chemistry is generally nonmetallic ergo metalloids are categorised as nonmetals: a distinctive kind of nonmetal, but nonmetals nevertheless.

Yes, metallicity does not increase monotonically from left to right in the periodic table.

It shows a progression at the category level, which is the level of abstraction we are dealing with here.

I agree with you there is no one-size-fits-all categorisation for the nonmetals.

Indeed, I have mentioned that each of our schemes has imperfections.

The polyatomic-diatomic categorisation is well defined, at face value, but that does not imply that the resulting categories are clear cut, as I have mentioned with regard to boundary overlaps, or that they necessarily show the most relevant trends.

The intermediate-corrosive categorisation is as well defined as the literature provides, and more consistent with the logic framing the rest of our categories, and the way chemists think about nonmetals. Sandbh (talk) 05:09, 16 June 2017 (UTC)

I am not sure that chemists think about them as two separate categories, though. It would be more of a continuum if anything of nonmetallic behaviour, largely correlating with electronegativity. Certainly nonmetallicity decreases as Cl > S > P > Si > Al > Mg > Na, but in all seven there is some nonmetallic character. If anything, given how closely allied triads like {S, Se, Te} are, the usual "working classification" seems to be "first-row" for the anomalous ones, and then "pnictogen, chalcogen, halogen, group 14, group 13". You will find tomes about the chemistry of the pnictogens, but largely not for those of the nonmetals. Even if one draws the line as one does currently, you want to continue the story. Se acts a lot like Te and you want to include it, and O does not act very much like S, yet moving from Se to Te leaves the nonmetals and moving from O to S does not. (I think we can extend that generalised Fajans' rule to negative oxidation states to cover this.) And in case this is thought of as corroborating the idea that metalloids should be considered nonmetals, Ge acts reasonably like Sn too.

If one points to the literature to split the nonmetals as corrosive and intermediate, I would note that even more simply split the p-block as "B, group 13, C, group 14, N, pnictogens, O, chalcogens, halogens, noble gases".

If we want to consider only chemical properties, I do not see how you can keep things like molybdenum and tungsten out of the "metalloid" category. Even Th, Pa, and U seem to be almost there, especially uranium in its most stable oxidation state of +6 (and +4 isn't really much better). If the idea of what chemistry is nonmetallic tends to include that of such elements (Th, Pa, and U are even very electropositive and form spalling oxides!), then perhaps we need to narrow down what it means to be a nonmetal.

In my experience the innocent reader, in an effort to put the chemistry of all 118 elements (or even just the 83 ones he can actually legally see) into perspective, would be more inclined if anything to put too much trust in generalisations. Many terms for the metals in groups 12 to 16 have been used in the literature. One wonders why "poor metals" is so common online, as if one looked at the literature, I would've expected something like "B-subgroup metals" to be more common.

If a categorisation is this uncertain, and there is a more certain alternative, why use the uncertain one? Double sharp (talk) 07:59, 16 June 2017 (UTC)

Double sharp, I tend to think that chemists think of halogens, and of metalloids (where "some of the most interesting stuff" happens---not my words I am quoting a chemist) and then, the rest of the nonmetals, which corresponds closely to our proposed scheme. Certainly, at the same time, they will think about the nonmetals on a group-by-group basis. And yes, they probably overlay a continuum over all of that, as well.

I know that the literature examines the p-block as "B, group 13, C, group 14, N, pnictogens, O, chalcogens, halogens, noble gases", but that is after the general survey of the lay of the land into strong metals-less strong metals-intermediate somethings-weaker nonmetals-and strong nonmetals, or categorisation variations of that kind. The second approach does not extinguish the former; the two are complementary.

I do not advocate considering chemical properties only; as I said, the literature does consider chemical and physical properties but it does not strike me as a partnership of equals rather, the chemical informs the physical, at least in chemistry, etc.

We cannot be responsible for the actions of general readers who carelessly glance at Wikipedia and then act on that basis, rather than reading a little further. I suppose poor metals is common on-line due to people preferring plain English, the absence of any IUPAC terminology, and the limited use of the expression "post-transition metals". (I confess to liking poor metal myself. Even Greenwood & Earnshaw refer to Sn and Pb as "poor metals".)

I presume the elements are classed in more than one way in order to help with developing what Peter Nelson wrote about i.e. "a fairly thorough knowledge of how the principles work out in practice - of exceptions, trends and patterns [e.g. the first row anomaly, anionic gold, rogue transition metals etc] - so that a particular deduction can be appraised and if necessary adjusted. In the limit the process becomes virtually intuitive". Sandbh (talk) 06:52, 17 June 2017 (UTC)

R8R writes: Thank you for your overview. I think, however, that I might've failed to clarify some points which, in turn, may lead to misunderstandings.

First, I don't think that "intermediate nonmetals" is clearer to many: those readers who entirely understand the meaning of "intermediate metals" should be able to understand "alkaline earth metals." Those who don't understand the latter will probably not gwt the forner in its entirety, either. If it is clear between what and what this "intermediate" is, then comes the question: why can we not do with a term that just says what it means in its own words? "Alkaline earth metals" is good in its three words. "Intermediate nonmetals" requires clarification; even if it's easy to make, it is still a requirement too many. All other terms don't do that, by the way, they're good by themselves. If, however, it is not clear between what and what that "intermediate" is, then the fact that the word "intermediate" is comprehensive by itself is not -- and I believe you should agree -- much help at all.

As for DePiep's quote: this would be a very good point if we were to write a book by ourselves; not so much in Wikipedia because it will fail the WP:OR policy.

I see, however, that you agree with my point re minimization of non-standard terms and good clear English in them. This pleases me a lot.

As for examples of "intermediate nonmetals": first and foremost, they're not all alike. "Moderately active nonmetals", for instance, is a thousand times better than "intermediate nonmetals"; all because it doesn't fail as obviously the self-descriptiveness criterion. (It is not perfect at its clumsiness but is far better to begin with.) Second, I cannot judge about those terms without context: sentences, chapters, the general expected reader's proficiency, etc. So I can't comment, not because I use all possibilities not to, bur because I don't even feel I have enough information to be the judge.

As for "arbitrary": I certainly could have described this better. It is absolutely clear what nonmetals are di- and polyatomic. It is clear what nonmetals are corrosive and what are not. What is unclear is why either is the divide to go with; why it is better than the other. DS, for instance, claims now that the current divide is better. Any divide so far is clearer: AM and AEM are different because both terms are far better-known than "s-block metals." that's a perfectly clear (and very Wikipedia-y) reason. That's the difference; O believe this is a huge factor.

I hope thos clarifies my thinking; you are very welcome to provide any further comments.--R8R (talk) 17:03, 12 June 2017 (UTC)

• A note about my quote above in this R8R section (Category name and meaning is secondary [to catogisation criteria]). Spoiler: I see no trespassing, so just some clarification here. I mean to say that the scientific categorisation is the main issue. That is: find & set criteria, choose in grey areas, accumulate this in element properties (to get to a stable category definition). Then, the name or description for that category is secondary, and following (not: setting the definition/name beforehand. This is relevant especially since our categories are covering the periodic table 'all and just once': classification requirements R8R or YBG mentioned long time ago). Next, even more irrelevant, is the category color (I'll guard all degrees of freedom to improve those later on ;-) ).

Being an enwiki only, category names can not be OR (but descriptive names can be).

And, as was noted in this thread, let's hope IUPAC picks up this, er, hole-filling hint. -DePiep (talk) 18:59, 12 June 2017 (UTC)

I think it would not be a good thing if IUPAC mandated a particular categorisation, as the ideal categorisation depends greatly on what you are trying to show. There are good arguments for categories like "pre-transition metals" for groups 1, 2, 3, the lanthanides and actinides, and Al; "rare earth metals" for Sc, Y, La, and the lanthanides; "semimetals" for As, Sb, and Bi; or even elevating "icosagens, crystallogens, pnictogens, chalcogens, halogens" into categories as well as groups. All of this depends on what point you are trying to make. But we are an encyclopaedia, trying to sum up the average views of most authors, and trying to use categories that work best for as wide a variety of purposes as possible.

It almost seems like the "ideal categorisation" is to go purely by groups and follow the periodic table, by that measure (and there are very good chemical reasons for it too that have nothing to do with parroting Mendeleev because he was Mendeleev). But that seems to make the colours unnecessary. And that actually seems like a fairly sensible answer: leave the whole thing blank, and consider it a colouring book whenever we are trying to show something. Double sharp (talk) 08:11, 13 June 2017 (UTC)

Response by Sandbh to R8R's comments of 17:03, 12 June 2017

The clarity of intermediate metals is not an issue. As you say, a reader who understands "intermediate nonmetals" will understand "alkaline earth metals"; a reader who doesn't understand the latter won't understand intermediate nonmetals, nor are they likely to understand e.g. "lanthanide" or "metalloid". As DePiep said, that is what links are for. As noted, we write for the general reader so I fully expect that all our category names will be like Greek to them.

Intermediate nonmetals is a good descriptive phrase for several reasons.

1. As I explained to Double sharp, it is preferable or at least satisfactory to NOT add subjects like activity, electronegativity, or metallicity (which results in category names such as "moderately active nonmetals") since nonmetallic character is a composite of many factors that contribute to, or are aspects of, a nonmetallic element's nonmetallic character. When you add subjects it starts to become very tough to establish representative categories in which every member meets just that particular criterion. By "representative" I mean a category in which a single criterion correlates reasonably well with all the other factors that make up non-metallic character.

2. Category names like "highly reactive" or "very electronegative", which I have previously proposed or mentioned, have been criticised for their subjectivity e.g. what is "highly"(?), or what is "very"(?). I believe you may even have criticised some of those suggestions for these reasons and you are now suggesting them to me!

3. "Intermediate nonmetals" reflects the overall intermediate nonmetallic character of these elements and corresponds reasonably well with their geographic position in the period table. Just like the transition metals are positioned between the alkalis and the post-transition metals. Indeed, the Oxford English Dictionary says "intermediate" means, "Coming or occurring between two things, places, etc.; holding the middle place or degree between two extremes".

4. It is a short and plain English, as you noted.

5. It is a thousand times better :) than "other nonmetals".

The revised proposal is better than the current division because it is a purer solution, one that is more consistent with the rest of the category names, how elements are described in the literature, and the way chemists think about them. The polyatomic-diatomic split worked, but it was essentially a "make-do" solution focussed on the "other nonmetals" and the light halogens, rather than blending in with the logic of the rest of the element categories. Sandbh (talk) 04:23, 14 June 2017 (UTC)

There's been one particular situation on my mind that I haven't mentioned yet.

Imagine a reader wants to read an article on a nonmetal like sulfur. They open the article, the first thing they see is the infobox (which should be especially good after we reduce its length). They see the field "category" and its value "intermediate nonmetal." How is this going to be helpful? There are no elements in the infobox that this category could be intermediate between. Almost all views we have are views of articles; not all readers read the infobox (yet perhaps far fewer go for the table in the end), but still it is safe to assume the scenario I've described is quite common.

If "alkali metal" is Greek, too, at very least it is supplied with a wikilink. Of course, this problem is not as terrible when you have the table with a legend, but that only happens in the end of the article (after references, notes, etc.) and it is still collapsed there, so this does not count as a solution. On its own, "reactive nonmetal" is not a perfect name either (not clear if there any gradation etc.), but "intermediate nonmetal" is meaningless.

That's the problem very much.

Here is a counter question, though: is there anything particularly important about the intermediate nonmetals that would give them a good reason for a category other than not being both extreme nonmetals and metals that could be described with a descriptive term? If so, could we go for such a term? If not, is there a strong reason to have a separate category for them? Many people think of "moderate nonmetals" or the like but also many people think of "early transition metals," for instance, and we won't have the latter term.

To be clear: I have not suggested actually using the name "moderately active nonmetals," I only said it was better because it was far better in respect of self-explanatoriness.--R8R (talk) 14:57, 14 June 2017 (UTC)

far fewer go for the table in the end: Thank you for summarising so nicely why I think it's a mistake that the infobox periodic table does not include the symbols.

I like your second-last paragraph. In fact, perhaps I shall elevate it to a principle: a category must be definable without making reference to another category. ("Nonmetals" are at least fine: you can define metallic and nonmetallic properties, and nothing will change if instead of "metals", "metalloids", and "nonmetals" you say "tables", "chairs", and "beer mugs", with apologies to Hilbert.) Double sharp (talk) 04:43, 15 June 2017 (UTC)

With careful wording I suspect this should not be an issue (but I won't know for sure until, if ever, I try sandboxing a revised nonmetal article). I note that the spirit of the principle does not apply to the category names "transition metal", "post-transition metal", and "nonmetal", and I see that the opening definitions of our articles on metalloids, post-transitional metals, and nonmetals do not comply with this principle. I once tried explaining what a nonmetal was to a non-scientist and was met with stupefaction until I explained it in terms of something that lacks metallic attributes, and gave some examples. Indeed, the Oxford English Dictionary defines a nonmetal as, "A non-metallic element".

I tend to suspect that the principle you have suggested is not something fundamental to classification science (sometimes economy of description will be better served by defining a category relative to another category—"other nonmetals", for example) and may represent something more of a useful non-essential aspiration rather than a "categorical" :) principle. Maybe a clunky example of what I am trying say is the colour grey: our article says it is an intermediate colour between black and white; the OED says, "The adjective denoting the colour intermediate between black and white, or composed of a mixture of black and white with little or no positive hue; ash-coloured, lead-coloured. Said of sea, sky, and cloud when not illuminated by the sun." Here grey is defined by reference to other things. So what? It's plain English. That's the thing. It seems to me that the definition is not so important as the description. Actually, I'll take that back: the definition is very important for the general reader. Sometimes defining something by reference to something else will be more powerful than trying to define it purely. Sandbh (talk) 21:44, 15 June 2017 (UTC)

Of course you cannot do it for a reader who doesn't understand chemistry yet, since all the highly characteristic properties of nonmetals are chemical, like most categories. You could hardly explain "alkali metal" to him either; you can't say they are "metals which form alkalis when dissolved in water", because groups 2, 3, and many of the lanthanides and actinides do the same. You have to bring in chemistry. No doubt the average person, not really knowing what substances actually are elements, would think of wood and plastic as nonmetals. And if you give them the physical definition of a metal they will proceed to think of things like lithium as metalloids, because it looks like a metal and doesn't behave like how they think one should. We do not expect the article on, say, quaternions to make any sense to someone not well-versed in mathematics. It is surely within the bounds of common sense that we define things at the lowest level at which the topic can even be understood.

Naturally you can define grey relative to other things. But there is no need to when "having little to no positive hue" already tells you everything you need to know about it; the rest of the definition there is more to give you a visceral confirmation of what that looks like. And what a short description it is! You will tend to be able to write one like that for most useful categorisations in colours, even if it presupposes that you know what technical terms like "hue" or "saturation" mean. If you can't write a short description that both defines the scope and the boundary of a category, then perhaps the category is not very natural.

If we had known more nonmetals in the free state earlier perhaps they would have another name; the same is true for the other categories. But their definitions are fairly obvious in the sense that reasonable possibilities immediately spring to mind. Double sharp (talk) 23:34, 15 June 2017 (UTC)

R8R, for the imaginary general reader reading an article on a nonmetal like sulfur, I imagine that the expression "intermediate nonmetal" would be as meaningless as "transition metal". In both cases this does not matter since each term is wiki-linked: "intermediate nonmetal" points to what would be the appropriate section of the nonmetal article; "transition metal" points to the transition metal article. If the general reader reads the wiki text explaining what an intermediate nonmetal is they straight away can see that intermediate nonmetals are intermediate in nonmetallic character, as well as intermediate in periodic table position between the more metallic metalloids and the more nonmetallic halogens, consistent with the progression in metallic to nonmetallic character going across the periodic table.

For some reason a meaningless IUPAC-endorsed term like "transition metal", or a meaningless non-IUPAC endorsed term like "post-transition metal" is OK but our project seems to feel like it has to knock itself out and come up with superior ("more meaningful") terminology while at the same time striving to avoid the use of neologisms. As I see it, we seem to be killing ourselves for no reason that is worth being particularly concerned about.

With one caveat (set out in the next paragraph) I would say that the intermediate nonmetals are typical or exemplary nonmetals, somewhat like the transition metals are the most representative or typical of the metals. In other words, the intermediate nonmetals are not as metallic as the metalloids, nor are they as nonmetallic as the corrosive nonmetals. The same goes for the transition metals: they are largely neither as metallic as the electro-active metals nor as weak as the post-transition metals. From a terminology view what else could you call the transition metals apart from transition metals? That is their nature: transitional between the extreme electro-active metals on the left of the periodic table and the metallically-challenged poor or post-transition metals further to the right.

The caveat is that the terms "typical" and "exemplary" won't do for the intermediate nonmetals since hydrogen could hardly be called a "typical" nonmetal. However it is neither as blatantly metallic as the metalloids nor as extremely non-metallic as the corrosive nonmetals, so in that regard is an intermediate nonmetal.

When the literature talks about the progression in metallic to nonmetallic character across the periodic table, there is a very strong focus on the alkali metals at the start of the table, and the halogens (and the noble gases) at the other end.

But the rest of the nonmetals, the ones between the halogens and the metalloids, are collectively glossed over---at least until the descriptive chemistry section of the applicable book when they get treated on a group-by-group basis. That being said, some writers, as we have noted, use descriptive terms like "somewhat", "moderately" or "mild" when describing these nonmetals, and we are all familiar with the term "other nonmetal".

Whereas in fact some of the most important nonmetallic chemistry occurs among these glossed over mild-mannered orphan nonmetals (= GOMO nonmetals :) simply because they encompasses the heartland of the non-metallic chemistry spectrum.

Nobody likes "other nonmetals" (I don't since it does a disservice to the elements involved and teaches nothing) yet anything more technical or complicated attracts concerns about being neologistic.

So, I am damned if I do and damned if I don't, so I now stick with generic, plain English descriptive adjectives and I rely on the flavour of the term "transition metal" as a precedent, in the absence of anything better in the literature. Sandbh (talk) 12:17, 15 June 2017 (UTC)

When the literature talks about the progression of character from metallic to nonmetallic, one would suspect that it is considering both chemical and physical properties. But: if we consider physical properties, the metalloids look a lot more metallic than your proposed categorisation would have them (and I would argue that they already show a lot of latent metallicity, even chemically – I'm still writing that exposé) – and if we do not, most of the early transition metals look very nonmetallic. If you plot metallicity in the periodic table by some sort of set of criteria, what you would find is a correlation with oxidation state and atomic radius, not a monotonic decrease to the right. I daresay Mo would appear less metallic than Ag, for example, simply because the group oxidation state is +6 in the former and +1 in the latter; thus the chemistry of Ag^I is mostly cationic and that of Mo^VI is mostly polyoxomolybdates. Uranium would look far more metallic than Mo in the +6 oxidation state: at least it is still cationic in that state, forming the uranyl cation, and the uranates are far more like double oxide salts than actual oxyanions of uranium. But, of course, due to the high oxidation state you will not find any simple cationic chemistry for U^VI; you will find some for U^IV, but this only works because uranium is so heavy, and even so U⁴⁺ (aq) is a very strong acid and is hydrolysed unless [H⁺] exceeds 0.5 M. In fact it quite often hydrolyses all the way to U(OH)₄ (or rather a polymer of that), which reminds me eerily of germanium(!). So I think that ignoring physical properties entirely won't do, and with that goes the idea of classifying metalloids as nonmetals.

Let me try to spell out the problem with splitting the nonmetals this way in detail: "transition metal" has a meaning that is more than the sum of its parts, while "intermediate nonmetal" does not. The word "transition" has stopped being an adjective in its usual English sense in inorganic chemistry; instead it means "that series of elements in the middle of the d-block that commonly exhibit variable oxidation states and like forming coordination complexes with Lewis bases", and if you want to learn chemistry well you have to know that to progress. Whereas pretty much everything in the chemistry of selenium (an "intermediate nonmetal" in this scheme) is also found in tellurium (a metalloid), and there is little in the chemistry of nitrogen (an "intermediate nonmetal") that finds echoes in that of phosphorus (another "intermediate nonmetal"), so clearly there is no hope for a similarly pithy definition that gives "intermediate" a meaning of its own here. The term is undefinable without referencing something else.

At least "polyatomic" and "diatomic" are indisputable, and in their narrowness they allow one to relax and say "oh, it's not meant to refer to all properties, but just vaguely correlate with most of them"; "intermediate" tends to try to master all the properties and always runs into an exception to each one of them. Double sharp (talk) 14:54, 15 June 2017 (UTC)

"for the imaginary general reader reading an article on a nonmetal like sulfur, I imagine that the expression 'intermediate nonmetal' would be as meaningless as 'transition metal'." Sure. Yet you've mentioned classification science on this page; let's also take a look from that perspective because it would be a mistake to miss it. "Transition metals" was probably a contamination of two terms; now, however, it is a term of its own. It wasn't that from the beginning but it grew to be one.

After some thinking, I am not saying that "transition metal" is a perfect term. But it's a widely recognized singular term. In this respect, that's an advantage that the term "intermediate nonmetal" does not have. When I see the phrase "transition metal," I first think of the d-block and not about "transition from what to what." When I see "intermediate nonmetal," I first think about "between what and what." This shows how one imperfect originally term grew to be established in chemical nomenclature in its own right and one has not. That is the difference. This is why the former is not meaningless on its own now while the latter still is.

This well implies that the term "intermediate nonmetals" may once become a term in its own right, and so it is. But not yet.

However, from your response I still don't see the need of having a term for what you call "intermediate nonmetals" despite having asked for it. I genuinely don't see it. Is it there at all?

Just to clarify it: I am not being picky, I am not just saying "I don't like it, suggest something else." I actually think that there are (not may be; are) better options.--R8R (talk) 13:20, 18 June 2017 (UTC)

"Transition metal" is a fine term for the general reader in that it probably doesn't mean any more to them than any of our other terms. I know when you see it that you, as science professional, think of the term d-block, and this is fine. I talk some more about imperfections, chemists, and tidying up the nonmetal category garden in my overview guide, and I hope this makes things a bit clearer. If this does not address your question about the need for having a term called "intermediate nonmetals" please let me know. I hope the overview guide will serve as a divider between what we have discussed so far, and what comes after that. I talk about sandboxing a proposed revised nonmetal article and would be happy to take on any outstanding requests. Sandbh (talk) 03:52, 19 June 2017 (UTC)

Hi @R8R: As per your comment and query above, I believe I've addressed the need for a term such as "intermediate nonmetals" in the "Average view" section, where I wrote that this is attributable to the well-documented progression in metallic to nonmetallic character going across the PT, including fundamental notions of weaker, stronger, and almost inactive nonmetals. Sandbh (talk) 04:06, 21 June 2017 (UTC)

The thing is, though, the current category names are first-year-chemistry material. Even "polyatomic" and "diatomic" are first-year terminology. We know at least from the start where the transition metals are on the table, and even the Ln and An, even though their chemistry is not covered till much later. It is just that important, because of how sharp the main-group vs transition distinction is. Even if I am not sure if the blocks per se are the best way to go about doing it, I see a great difference between the electropositive metals on the left end (with the early actinides being a weird subset), the transition metals in groups 4 through 11 in the middle, and the other main group elements on the right. The break between groups 11 and 12, as well as that between groups 3 and 4 is very sharp in a way that is not the case for splitting the nonmetals. Double sharp (talk) 05:10, 21 June 2017 (UTC)

Our category names are unlikely to mean anything to the general reader. Since there is no suitable formal category name for the weaker and stronger nonmetals, we use plain English and concise, non-novel, descriptive phrases. The general reader will appreciate the visual reinforcement of the positioning of the intermediate nonmetals between the metalloids and the corrosive nonmetals, in the same way that the transition metals bridge the pre-transition metals, and post-transition metals. A first-year chemistry student will understand the concept of an intermediate nonmetal and they will certainly appreciate what a corrosive nonmetal is.

For the general reader, knowing where the transition metals are on the periodic table will not help with the main-group vs transition distinction given transition metal chemistry does not come into its own until group 4; that it starts to flicker at group 11; and becomes vestigial at group 12, even though groups 3 to 12 make up the d block, which gives rise to confusion between the concept of transition metals and the concept of d-block metals.

Yes, one can see a great difference between the electropositive metals on the left end, the transition metals mostly in the middle, and other main group elements on the right. Actually, this is not the first thing mentioned in text books, on average, which is the contra-distinction between metals and nonmetals, and the alkali/alkaline earth metals and the halogens/noble gases. The rest of the metals and nonmetals are distinctive by virtue of the natural focus on alkali/alkaline earth metals and the halogens/noble gases. One cannot make such distinctions in isolation. Whatever is left is distinctive by virtue of its exclusion.

The break between the corrosive nonmetals and the rest of the non-noble nonmetals is self-evidently sharp; the other breaks exist but are less tangible; much less so for the general reader. And what the substantive nature of the break is between the alkali and alkaline earth metals, from the general reader's perspective, pales in comparison. Sandbh (talk) 11:57, 21 June 2017 (UTC)

Progression from metals to nonmetals

Plot of effective ionic radii versus oxidation state for various elements (Greenwood and Earnshaw, p. 52). Double sharp (talk) 06:28, 25 June 2017 (UTC)

On the metal to nonmetal progression, the literature does consider chemical and physical properties but it does not strike me as a partnership of equals rather, the chemical informs the physical, at least in chemistry. I know this is weird since it is the physical, I think, that informs the chemical, not the other way around. But ever since the emphasis on the physical to distinguish between metals and nonmetals was superseded by chemical properties, the physical properties, it seems to me, have taken more of a backseat to the chemical.

If you are of the opinion that the metalloids show a lot of of latent metallicity chemically then that is a good thing and, as the most metallic of the nonmetals, entirely what I would expect.

I think the literature does record a reduction in metallic character among the transition metals, from the quite reactive early transition metals in groups, say, 4 and 5, through to the more well-behaved transition metals, the schizoid coinage metals, and then the pretend transition metals in group 12. The progression has some bumps along the way but the overall trend is there.

As noted, the meaning of transition metals is well understood by the expert but means nothing, apart from its plain English meaning, to the general reader. As discussed, the meanings of "transition" and "intermediate" are equally as meaningful for the expert reader as they are for the general reader.

You asked me once if thought selenium was a metalloid and I said no, its 2.55 electronegativity is too high for a metalloid (Te is 2.1). It is nevertheless what I would consider as being right on the border of metalloid territory. Yes, its chemistry is similar to that of tellurium but the literature has decided by a wide margin that selenium is more properly characterised as a nonmetal and tellurium as a metalloid I guess, in part, because tellurium dioxide is amphoteric whereas selenium is not. That said, I'm not sure of the relevance of your point since, in the current scheme and the proposed scheme selenium and tellurium are in separate categories.

I've previously summarised the shared chemistry of N and P, including resemblances between the phosphides and the nitrides. Analogously, should we abandon the metalloid category on the basis that the chemistry of B has little to do with that of Te? You seem to drilling down to a level of detail that is at odds with broad brush strokes of our categories.

The statements "clearly there is no hope for a similarly pithy definition that gives "intermediate" a meaning of its own here" and "The term is undefinable without referencing something else" are bold claims that remain to be tested. For one thing I don't believe the usefulness of category is necessarily related to its pithiness (although pithiness can be a good thing).

(You reminded me of the IUPAC definition of a planet. Their item (3) strikes me as an example of self-referential definition).

Arguments to do with the meaning and scope of "polyatomic", "diatomic", and "intermediate" are meaningless to the general reader, in the first instance. Superficially, the polyatomic-diatomic division is sharp but, as discussed, the boundaries are by no means as sharp as they may appear. There is nothing that different, in this respect, about the proposals under discussion.

May I add a comment that I feel like I am juggling multiple coloured balls (= your views and those of others) while trying to skate around the cracks in thin ice (= the literature) at the same time as dancing a hula (= attempting to comply with, or not bend beyond breaking point, the Wikipedia pillars) all in an effort to improve the way we categorise nonmetals. Would anybody like me to whistle Dixie at the same time? Fortunately I don't have ants in my pants yet. Sandbh (talk) 10:56, 16 June 2017 (UTC)

Here's the problem I find: if it is the chemical informing the physical (and hence why As and Sb are considered metalloids despite physically being as metallic as their "older sister" Bi), then I am not sure how you can legitimately apply the term "metalloid" to things like Ge and not to things like W. You call the centre of the d-block the more "well-behaved" transition metals, which is true as far as physical properties go, but as far as chemistry goes these seem to be even less metallic than some of the weakest main-group metals in the p-block like bismuth. It tends to be instead the right edge of the d-block (and mostly the 3d row) at that that behaves how you would expect for metals in Mn, Fe, Co, Ni, and Cu (Cr joins them mostly when it is in the +3 state, but is very different in the +6 state); Ru and Rh are similar to their "younger sisters", but they are rather exceptional in the 4d row for being so, and Ag has one foot already in main-group territory. These late 3d elements are the elements you tend to see until high school as representatives of the transition metals, but they are really not typical: they are almost like main-group elements in all but their oxidation states differing by 1 instead of 2 because the 3d shell is much more compact than the 4d or 5d shells. And early on in the series with Ti, V, and Cr, this simple picture is totally blown out of the water.

Suppose I talk to you about an element M which is usually tetravalent. Now I tell you that MCl₄ is a volatile molecular liquid that hydrolyses readily and completely to MO₂ even when in contact with moist air; that oxoacid salts of M^IV do not exist, only hydrolysed species being formed; that supposed "MO²⁺" ions sometimes alluded to in the literature are really polymeric –M–O–M–O– species; and that it has no simple cationic chemistry. Looking at how you are interpreting the literature to make the case for the metalloids as nonmetals, I am sure you would say that M should be classified as a nonmetal, although perhaps a weak nonmetal, analogous to silicon and germanium. Except that M is, of course, titanium. Now are we going to call that a "weak nonmetal"?

This is a problem endemic to the early transition metals, because of their overly high oxidation states preventing the onset of ionicity. ZrCl₄ doesn't fare much better, readily becoming ZrOCl₂ even in concentrated HCl; Hf almost certainly acts the same way. Even if you take Zr^IV and Hf^IV, it is only possible to isolate salts like Zr(NO₃)₄ and Zr(SO₄)₂ at extremely low pH; otherwise what you get are basic salts and anionic complexes. Greenwood and Earnshaw bravely state that "acidic solutions if sufficiently dilute (<10⁻⁴ M) probably contain the Zr⁴⁺(aq) ion" – but anything more forthcoming than that eludes us. I have to wonder if this is real, given that even for the largest tetrapositive ion in the periodic table, Th⁴⁺(aq), the predominant species even in very acidic solutions is [Th₂(OH)₂]⁶⁺. At least the oxoacid salts of Ce^IV and the An^IV are real, once again proving that it's size that counts. Regarding the unwillingness of Ti to react with acids, I am amused to note that Si is about as unwilling to do so (in both cases you need to use hydrofluoric acid), and that Ge is actually more willing to react with acids than Ti is. This is an illustration of what I think is wrong with calling the metalloids "weak nonmetals": their descriptive chemistry is not wholly nonmetallic (and indeed there are some strikingly metallic elements, as you know concentrated in their organometallic chemistry – I think this is because the higher oxidation state tends to be stabilised by a lower difference in electronegativity(!), as can be seen in inorganic vs organic Pb compounds), and if we take them as nonmetals we end up being forced to admit most of the early transition metals as nonmetals, since they are even less metallic in their descriptive chemistry.

Then let's go to group 5. Sure, they react readily with most nonmetals, but the products tend to be interstitial and nonstoichiometric. For vanadium, the main oxidation state is +4, and this at least gives us some real oxocationic chemistry with the vanadyl ion VO²⁺. But let's look at the group oxidation state of +5: V₂O₅ is famously amphoteric. If you dissolve it in NaOH, you get VO³⁻
₄, and if you acidify it to very low pH you get VO⁺
₂. But in between you get all manner of hydrolysis-polymerisation reactions, forming species like V
₂O⁴⁻
₇, V
₃O³⁻
₉, and V
₄O⁴⁻
₁₂ among others. Now, is anyone surprised to hear that phosphorus does essentially the same thing, even if vanadium is even more enthusiastic? Nb and Ta are more respectable; most niobates and tantalates are insoluble and are better thought of as mixed oxides. As an example of how it's oxidation state that matters, vanadium forms simple oxoanion salts as V^II and V^III as sulfates, but not as V^IV and V^V.

Now let's look at group 6, to see if we really get any better behaviour. As one would have expected from the ever-diminishing electropositivity, the answer is a resounding no. I previously half-seriously set you an exercise about the chemistry of chromium in its two main oxidation states: now I'll give you the answer. Chromium reacts readily with dilute HCl to form Cr²⁺(aq), which rapidly reduces water to produce Cr³⁺(aq). But if you put Cr in the presence of KNO₃ or KClO₃, ensuring its oxidation to Cr^VI, the metal is instead attack by alkali melts to produce CrO²⁻
₄. This massive difference between Cr^III and Cr^VI demonstrates that the contrast between stereotypical metallic and nonmetallic behaviour is not a distinction that exists between elements: it can indeed be provoked in a single element by raising the oxidation state (a main-group example would be Ge^II and Ge^IV, even if not as spectacular). The principle I have stated many times is quite well-known to Greenwood and Earnshaw, who write: "CrO₃, as is to be expected with such a small cation, is a strongly acidic and rather covalent oxide with a mp of only 197°C." Whereas MoO₃ and WO₃ have much higher melting points and are insoluble in water – standard ionic properties – yet are acidic and dissolve in aqueous alkali to give salts of the MoO²⁻
₄ and WO²⁻
₄ anions. This two-facedness within the same oxidation state seems to make the idea of a "half-metal, half-nonmetal" appear most reasonable here, perhaps even more so than in the p-block. I do not see metallic character reducing here in the first row; actually I see it decrease from {K, Ca, Sc} abruptly when {Ti, V} are reached, and then go back up with Cr when the common oxidation state drops down to more manageable levels.

I'm pressing onward to group 7, mostly because I'm actually enjoying this little tour: I believe I've made my point fairly well already. The +7 oxidation state, dominant in Tc and Re, is so high that you can forget about seeing anything remotely metallic there. I find it really cool that the original separation procedure for Tc was to oxidise it to pertechnetate and use perchlorate as the carrier. Tc and Re do not dissolve in non-oxidising acids; but oxidising acids quite happily bring them to the +4 state and produce pertechnetic and perrhenic acid. The polymerisation present in groups 5 and 6 has ceased here; surely it is also not a coincidence that P and S form various polyoxoacids and Cl does not. Manganese, of course, is more well-behaved as a metal because it is most stable as Mn²⁺, a simple high-spin cation.

When we reach group 8 we reach the standard-bearer of the d-block, iron. It would not be an exaggeration to say that the chemistry of iron is probably the usual thing most high-school students will think of when imagining a typical transition metal. Well, I'm sorry to say that while this is true for the middle of the d-block, it is assuredly not true for most of the rest of it. Iron is of course a fine, cationic, electropositive metal that is mostly found as Fe²⁺ and Fe³⁺: Fe^VI, like Mn^VII, tends to be an astonishingly powerful oxidising agent. (Cr^VI isn't quite as impressive; you can actually use it selectively in organic oxidations, unlike Mn^VII and Fe^VI which will indiscriminately oxidise everything in sight.) Now we actually have a reasonable, equally-spaced group trend of the sort we know and love from the easy-to-understand groups'; Ru is markedly less happy then Os to reach the +8 oxidation state. Os thus lacks aqueous cationic chemistry while Fe and Ru have them – but Fe has the different problem that the aquo ion [Fe(H₂O)₆]³⁺ is too readily hydrolysed to be really common, whereas the Ru analogue has a much more real existence. So, we've now gone over halfway through the transition metal group, and we've only now arrived at the first element for which aqueous cationic chemistry is real and common at normal acidities.

How about group 9? Now we are finally getting somewhere, with Co^II, Rh^III, and Ir^III being the most common oxidation states with well-defined cationic chemistries (though Ir readily oxidises to Ir^IV which does not). Oxoanions have finally become rare in this group. In group 10, the +2 oxidation state dominates (with +4 joining in for Pt), although simple Pd²⁺ and Pt²⁺ aquated ions only exist if potential ligands have been rigorously excluded. So we now have a double problem encroaching on 4d and 5d transition metal cations on both sides; if they are stable enough with water to not ditch it for any other passing ligand, their oxidation states are too high and they get hydrolysed, but if their oxidation states are too low to be hydrolysed then they will tend to ditch water for any other passing ligand. With gold in group 11 we reach the limit in strangeness, in which even the +3 state can boast no simple aquated cations, and Cu and even more so Ag in their resolute sticking to one and only one oxidation state in aqueous solution feel like they have one foot out of the strange place that is the d-block already.

If I see any trend in metallic character from Ti to Cu, it seems to increase rather than decrease as simple cations and lower oxidation states become the rule; this is only mildly offset near the end of the row by the rapidly diminishing size of the cations themselves. In their reluctance to form simple aquated cations at all, the general amount of metallic character in the rows of Zr–Ag and Hf–Au seem to be very consistently low, although a mild increase like Ti–Cu seems to be suggested.

All right, that's the end of the d-block exposé; now I'll respond to the remainder of your comments.

I do not consider what the average reader with no knowledge of chemistry makes of our names to be very important. If s/he wants to read up on chemistry and get better at it, the real names like "transition metal" will begin to have actual meaning, while things like "intermediate nonmetal" will not. Names like the former carry on being useful for a lifetime; names like the latter become useful or not depending on what point you are trying to make at one moment or another.

I suppose you can speak of latent cationicity in Te, although it does not seem to change its behaviour all that much from that of Se; the general personalities of both elements are similar. This does not seem to be the case for N and P; yes, you can point to some isolated points of their character and say "oh look, they really are related!", but the similarities between them play out against a background of difference, whereas the differences between Se and Te play out against a background of similarity. If we might permit ourselves some anthropomorphisation, Se and Te are clearly twin sisters, tellurium being perhaps a slightly more tomboyish one, but the similarities are so obvious in how they look and act. N and P on the other hand are so different that no one would believe they were sisters unless they produced their birth certificates for you.

Perhaps I would say that there are two distinguishing features of a good category, at the very least. First, it must have a sizeable number of members in the first place; otherwise every element is no doubt the unparalleled winner at being itself, and there is no generality and hence no power. But second, we must be able to clearly relate the members of the category; they should generally be allied and have something very important and defining in common, to the point that something which does not have those defining properties does not sit well at all with the category (witness groups 3 and 12 in the transition metals). For instance, B and Sb might be very different even when both in the same oxidation state, but in their generally equal proportions of metallic and nonmetallic character they can be considered to act as one.

I do not think of (3) as being intrinsic to the definition of a planet, merely as a bit of legalese. It's simply making it clear that the implied "if" in the definitions (1) and (2) is an "if and only if", so that anything that doesn't meet those criteria is not a planet or a dwarf planet. You could very well rephrase it to say "A Small Solar System Body is a celestial body that is (a) in orbit around the Sun; and (b) does not have enough mass to be in hydrostatic equilibrium". This is not a particularly serious case of self-reference because it is so easily removed; whereas the self-reference seems to be intrinsic to the "intermediate nonmetal" category.

I realise that all of this is coming all at once, and I am sorry for adding on to it. Please take it as a sign that we are all eager to do this right, and are trying to figure out just what that implies. Double sharp (talk) 13:04, 18 June 2017 (UTC)

W is a transition metal since it behaves as a transition metal (variable oxidation states; complex formation; coloured compounds; magnetic properties; catalytic activity). So does Ti (according to Earnshaw and Harrington, The chemistry of the transition elements, 1973). None of the metalloids behave appreciably consistently in this way.

I agree that the descriptive chemistry of the metalloids is not wholly nonmetallic. I never said it was. I believe I've consistently maintained that it is generally nonmetallic. I address the metalloid question further in the overview guide.

Electrochemically, and with some exceptions, the transition metals (the last time I looked) are mostly intermediate in character, and the post-transition metals are mostly weak. I expect it is possible to drill down into the transition metals and find exceptions to my generalisations however I'm working at the category as a whole level and not necessarily at the individual group or element level. I discuss this some more in my over guide which I will post after I respond to R8R.

I promise to consider, at length, your outstanding descriptive chemistry of the d-block when I sandbox the revised nonmetal article.

Certainly for the general reader a name like "transition metal" will begin to have actual meaning should they want to read up on actual chemistry; ditto a name like "intermediate nonmetal" when the reader learns that these relatively pedestrian (with some exceptions) nonmetals are described in the literature using the language of the "hoi polloi"---largely not that of the metalloids and their "disreputable for nonmetals" organometallic tendencies nor that of the corrosive nonmetals and their equally out there "party animal" reputations. I acknowledge that my description of the intermediate character of the intermediate nonmetals perhaps relies to much on contrasts with the metalloids and the corrosive nonmetals, and I'll see what I can do about that when I sandbox the revised nonmetal article.

On the IUPAC definition of a planet I will consider your point. I think the way they have expressed it speaks to my observation about the desirability of pithiness, and it fits better with the overall construction of their definition, but I'll see how this works with the intermediate nonmetals category.

I apologise for the brevity of my response the reason for which I hope will become clearer once you have seen the overview guide, and concur with your sentiments. The last 3.5 days have been exceptionally productive for me. Sandbh (talk) 03:40, 19 June 2017 (UTC)

I agree that tungsten and titanium have all the characteristic properties of transition elements. I am not sure how metallic they come out looking, unless you appeal to their physical properties, because nowhere in the main-group vs transition distinction does metallicity come into the picture. I even still have to say physical even when I talk about the bulk state, because one reason why tungsten and its friends in the early d-block melt at such high temperatures is the formation of covalent bonds within the metal supplementing the metallic bonding. If the chemistry of Ge is generally main-group nonmetallic, then I would think that the chemistry of W could be described as being generally transition-group nonmetallic, and this is where I find a problem with the proposed classification.

Electrochemically, looking at the pretty picture which is even reproduced above, it is not unfair to say that the 3d transition metals are usually intermediate, and the 4d and 5d transition metals are usually as weak as the post-transition metals and even the metalloids. And the reason why the 4d and 5d rows are not completely weak is because we were forcibly considering hypothetical simple cations of the early elements in those rows, and as we all know, metallicity rises with lower oxidation state. In their more common oxidation states they are certainly weak, IIRC: there is a reason why they are so noble.

The brevity is no problem, since your summary is giving me a lot to think about too: thank you very much for your complimentary words, as well!

P.S. The planet definition comes from the IAU, not IUPAC. ^_-☆ Double sharp (talk) 07:06, 19 June 2017 (UTC)

P.P.S. From Greenwood and Earnshaw, p. 347: "The metals which form silicate and aluminosilicate minerals are the more electropositive metals, i.e. those in Groups 1, 2, and the 3d transition series (except Co), together with Y, La and the lanthanoids, Zr, Hf, Th, U, and to a much lesser extent the post-transition elements Sn^II, Pb^II, and Bi^III." This accords pretty well with what I've always been thinking, with a few exceptions. The exclusion of the rest of the actinides must be due to their short half-lives, as this is a geochemical context; the need to specify the oxidation states for Sn, Pb, and Bi is quite reasonable given what I've been saying about the increase of metallicity as the oxidation state lowers. The exclusion of Tl (Ga and In are too happy in the +3 state) may be due to the fact that it is quite widely distributed and tends not to occur in quantity; given that Co is almost invariably associated with Ni, I also understand its exclusion from the club as being for geochemical reasons. Assuredly, it points to the fact that the 4d and 5d transition metals are very weird for metals, and are usually swept under the rug for as long as possible (i.e. until you get to university). Double sharp (talk) 15:27, 19 June 2017 (UTC)

Summary of Double sharp comments

• Don't see how you can separate mild from strong nonmetals
• Focusing too much on the elements in the free state
• Nitrogen chemistry is more reactive than it's intermediate category
• Oxygen does not pass the "strong" nonmetals bar
• Current scheme = least worst
• Why do nearly all authorities divide the table by groups
• Get rid of colours?
• Chemistry trumps physical

Response
On separating "mild" nonmetals from strong nonmetals please see my above response to R8R re any divide being arbitrary.

I don't feel we are focusing too much on the elements in the free state. Much descriptive chemistry is concerned with the nature of the elements in their free states, and how this translates to their chemical reactions. Some of this focus on the free state occurs elsewhere among the category names. For example, and as you alluded to, adding Na to water produces a characteristic alkali solution but dissolving NaCl in water initially yields a moderately acidic solution; much of the importance of the transition metals (e.g. titanium, cobalt, nickel and copper, silver, tungsten, gold) has to do with their properties in the free state (I know this is a bit of a simplification but this how the general reader will think about it); the metalloids have a reputation for semi-conductivity and looking like metals but behaving like nonmetals; and the noble gases are largely about the elements in their free states.

On nitrogen, I have previously noted or alluded to:

• its misleading high electronegativity;
• its nil electron affinity—this does not have anything to do with fairness nor does the electron affinity of O⁻ have any relevance; the only valid comparison is with the electron affinity of atomic oxygen;
• its reluctance to form simple anions;
• its essentially covalent chemistry;
• that the average oxidising power of it and its species, in aqueous solution, is less than that of both iodine and of sulfur;
• that nonmetal displacement series put sulfur ahead of nitrogen; and
• that nitrides resemble the metal borides, carbides and phosphides in many ways.

Here are a couple of other pertinent references:

• Chemistry: Structure and reactions (Synder 1966, pp. 235–242) rates the electronegativity of nitrogen as moderate, based on a consideration of bond strengths in nitrogen compounds;
• Inorganic chemistry (Housecroft & Sharpe 2008, p. 433) write that, "Very little of the chemistry of the group 15 elements in that of simple ions…nearly all the chemistry of the group…involves covalently bonded compounds. The thermochemical basis of the chemistry of such species is much harder to establish than that of ionic compounds. In addition, they are much more likely to be kinetically inert, both to substitution reactions…and to oxidation or reduction when these processes involved making or breaking covalent bonds, as well as the transfer of electrons.
• Introduction to inorganic chemistry (Nelson 2011, pp. 55, 57) describes nitrogen—in electrochemical terms—as weakly electronegative, and comments that practising chemists make use of this classification in thinking about nitrogen chemistry.

Certainly almost any compound of nitrogen is less stable than diatomic nitrogen, so nitrogen atoms in compounds like to recombine if possible and this releases energy and nitrogen gas, which can be leveraged for explosive purposes.

But nitrogen's "nitrophilicity" is a peculiarity of its atomic configuration and is not associated with nonmetallic character.

So, having regard to properties that are associated with nonmetallic character I conclude that nitrogen is appropriately classed as an intermediate nonmetal—albeit an idiosyncratic one.

Incidentally, unless I missed something, neither Cotton & Wilkinson, nor Greenwood & Earnshaw, nor Wiberg have anything systematic to say about the explosive aspects of nitrogen chemistry.

On the question of nonnmetals which are clearly strong and almost always form ionic compounds, I have previously here referred to:

• the corrosive nature of O, F, Cl, Br and I;
• their high electronegativities (> 2.6); and
• the fact that they are, or their species are, capable of acting as relatively strong oxidising agents.

I also gave examples from the literature as to other similarities between oxygen and fluorine, oxygen and the halogens, and oxygen and chlorine.

Here are a few more supporting quotes and citations concerning oxygen:

• "…oxygen is a potent, reactive, and corrosive gas…liable to generate free radicals, highly reactive bleachlike compounds that burn up organic molecules. We are able to tolerate an oxygen-rich environment only because our cells possess complex biochemical mechanisms for suppressing its many harmful influences." (Life's matrix: A biography of water, Ball 2013, p. 232)
• Inorganic chemistry, vol. 1, Principles and non-metals (Phillips & Williams 1965, p. 478) has a nice table showing that metal oxides are often ionic. The only covalent ones are MO₃ for groups 6 and 7; M₂O₇ for group 7; and MO₄ for group 8. The following are shown as being often metallic: MO in groups 4–6 and 8; M₂O₂ in group 1; M₂O₃ in group 4; and MO₂ in groups 1 and 2.

I tend to agree the current approach is the least worst. I view the refined proposal as even less worse. It is not perfect, just as their is no perfect Wikipedia article.

Nearly all authorities divide the table by groups but they do this after discussing the contours of the table in terms of metals and nonmetals, and discussing the gradual transition in metallic to nonmetallic character going across the table, and usually the change in metallic character going down at least the representative groups. Then they divide the table by groups. Our big table incorporates both of these approaches. (I guess one reason why authors take a group-by-group approach is that ths was the way Mendeleev did it, more or less, in his definitive Principles of Chemistry, 1868–1871).

In Comparative inorganic chemistry (Moody 1991, p. 388–389) the author briefly discusses the relationship of O and S to Se and Te. He says while many differences are apparent between O and S, S can replace O in compounds and that many analogous compounds can be formed. S allies itself to Se and Te. While O is characterised by ionic oxides, the others tend to covalency. O and S are clearly nonmetals but some metallic character is present in the others yet is latent until Po. Catenation is prevalent in S, less so in O. There is naught we don't know here but at least the author had a go. Whilst I acknowledge the differences, I am not so sure that a semi-homogenous characterisation of group 16 is quite as challenging as it is made out. I suspect this issue is that authors focus on the differences rather than giving as much consideration to the similarities.

Keep the colours. Our polychromatic icon is fabulous.

Yes, chemistry trumps physical properties, as I noted above. [sign?]

re Double sharp's "Get rid of colours?" I consider this off-topic (no need to change or discuss the category colors). But please point if I mistake. -DePiep (talk) 19:12, 12 June 2017 (UTC)

I wanted to ask the reasons why most of us want to keep the colours, as then we would have a good idea of their intended purpose, and be able to see how splitting the nonmetals would do it. Currently I see at least three possibilities for the final result of this discussion:

1. Keep the status quo division of the nonmetals (polyatomic/diatomic);
2. Use Sandbh's new proposal (intermediate/corrosive, with metalloids wholly as nonmetals);
3. Use something else: for example, I think R8R and I have questioned if it is really necessary to break up the nonmetals at all, and I have even suggested colouring only hydrogen separately.

The last one could result in one colour going missing, and already we need to infer the homologous series like S/Se/Te and the halogens purely from their positions in the periodic table. Since all but the metallicity categories of the p-block can readily be inferred this way, I am genuinely curious to see why removing the colours is not something that people want to do. Double sharp (talk) 03:36, 13 June 2017 (UTC)

Summary of YBG comments

• Divide the nonmetals into unreactive (noble gases); reactive nonmetals; and borderline nonmetals

Response
Including the metalloids, the Wikipedia periodic table has divided the nonmetals into four categories since its inception. When the literature divides the nonmetals into classes it most commonly uses a similar division. Doing so better illustrates the stepwise progression in metallic to nonmetallic character, and the symmetry of same. We previously considered a division along the lines you suggested during the megadiscussion that preceded the current divison. That proposal was one of the unsuccessful ones.

I think one reason for it, if I am not mistaken, is because oganesson is predicted to "act funky" (to use Nergaal's phrase), and so it becomes oddly a reactive nonmetal that is not marked as such. I am not sure whether I think this is a good objection given that the chemistry of At, Ts, and Og is not really the main focus if you are teaching descriptive halogen and noble gas chemistry.

Still, I am rather of the impression that the progression of nonmetallic character is more continuous than proceeding by discrete steps, and while inserting more colours might make it look more continuous along the lines of an integral it neglects the fact that the early p-block region is hardly the only place where chemical metallicity is attentuated among elements that physically look like metals. Indeed, this seems to happen just about every time the predicted oxidation state from the periodic table rises too high for the element's ionic radius to bear: so many of the transition metals show a progression to nonmetallic character, as do the early actinides.

And while the literature may often divide the nonmetals, besides the noble gases it is not at all clear where any given element is going to be placed. You could make a reasonable case for shifting nitrogen, sulfur, and iodine a category up or down; you could also make a reasonable case for considering the metalloids as "nonmetallic metals" instead of "metallic nonmetals", and actually I think doing both at once is even better. Meanwhile hydrogen is so odd that it often gets excluded from the game altogether, if I am not mistaken (so I actually wouldn't mind highlighting H on its own and calling the rest "typical nonmetals"). The absence of a clear division suggests to me that it is better to leave things to trends and consider the whole block one particularly heterogeneous family. (After all, what are we to make of Au and Hg no matter where we put them? It seems to me that there isn't anything quite as strange as that in the clearly nonmetallic category, once you push hydrogen aside.) Double sharp (talk) 04:05, 12 June 2017 (UTC)

Nicely summarized, @Double sharp:. Thank you.

Yes, my proposal was one of the ones rejected in our megadiscussion. We rejected it in favor of the mono/di/poly-atomic classification, which received our joint acceptance, I believe, primarily because of its clarity and lack of ambiguity, that is to say, no fuzzy lines. However, that classification has come into question of late because it overemphasizes physical properties of the elemental form and is largely irrelevant to (a) chemical properties of elemental forms, (b) physical properties of combined forms and (c) chemical properties of combined forms. For all these reasons, the mono/di/poly subdivision seems on its way out.

Now the question is, with what do we replace it? I submit that all of our other options suffer from a lack of clarity and an abundance of ambiguity. Once you hive off the noble gasses, the remaining nonmetals will admit to no clear lines of demarcatiion. They present themselves as a sprawl of suburbs, whose only clear dividing lines are those drawn by postal and electoral authorities.

Hence, I believe that having eliminated the mono/di/poly division we agreed to in the past, the only sensible alternative left is to combine the non-noble non-metals together into a single category.

The remainder of my proposal, moving the metalloids into the nonmetal supercategory, was just continuing the idea already expressed, and can stand or fall on its own. My contribution was to propose the term "borderline nonmetals". But I note that all three adjectives "borderline", "intermediate", and "other" suffer from the same malady, which IMHO is mostly benign in the borderline patient, mildly debilitating to the intermediate patient but has metastasized terminally in the other patient. YBG (talk) 07:32, 14 June 2017 (UTC)

To some extent I think that the polyatomic-diatomic divide might be the best if you insist on dividing up the nonmetals at all; while it is unarguably a physical property, it correlates well with chemistry. So it shows up the strongest nonmetals as the hyperelectronic ones with more valence electrons than valence orbitals, which can simply end off their structures quickly with lone pairs not bonding to anything else when they form their sp³ hybrid orbitals; the noble gases are the extreme case when no bonding at all is necessary. But because pπ–pπ overlap becomes less effective than dπ–pπ overlap once you get past the first row thanks to the nodelessness of the 2p subshells, the multiple bonding necessary for this becomes unfavourable in groups 15 and 16 once you get past the first row, and a large polymeric structure (though still not yet with delocalised electrons moving through the whole of 3D space) is expected before metallisation sets in at As and Po in each group. Thus simple arguments from electronic structure predict where the line is going to be drawn, and because electronic structure is so fundamental it correlates very well with chemistry as well as physics. Predicating this on the polyatomic-diatomic distinction is to a certain extent putting the cart before the horse, but it is no accident that it works this well. And, I might add, for all its flaws, it is correct and does not require any mental supplementation by the reader.

I don't think any other split would work remotely this well, except perhaps the idea of putting H aside and having everything else be a "typical nonmetal". Indeed I am not sure H is treated well by any criterion at all. Double sharp (talk) 08:23, 14 June 2017 (UTC)

@Double sharp: So do you mean we should keep the status quo? I myself am getting headaches trying to understand the long, long proposals and responses here. (And what about the dubnium GA review you reserved for yourself?) Parcly Taxel 09:08, 14 June 2017 (UTC)

The current serious proposals are:

1. Status quo;
2. Sandbh's last one at the bottom of this section (#Alternative proposal, Mk. 3). (third proposal)

The older discussion was helpful for refining Sandbh's proposal to its current state, but it's not really necessary to read all of it to understand what is going on. As for the dubnium GA review, I'll do it later today; I gave the thing a read first. Double sharp (talk) 09:43, 14 June 2017 (UTC)

Sandbh comments

In putting together this proposal I have wondered if I have breached our guidelines on neologisms and original research. The descriptive literature on the p block is dominated by a group-by-group approach. Few authors categorise the p block as we do, which seeks to show the diagonal pattern that arises due to the interaction of horizontal trends across the periodic table and vertical trends going down groups. The inconsistent efforts of the few authors that do take such an approach represent nothing particularly worth quoting given this is where the execrable "other nonmetals" category name came from (I will make an exception for the "post transition metal" category name, which is a reasonable, if not that common a term). These authors are hamstrung by retaining the halogens as a category. They do this because they pay only lip service to the nature of astatine, and treat it as a halogen, and ignore its metallic nature. Whereas astatine is most likely to be a metalloid or post-transition metal but recognising this makes it too hard to categorise the remaining nonmetals. And yet our categorisation approach, which seeks to accommodate horizontal and verticals trends, and does not ignore the astatine challenge, is worth doing since it shows what is actually happening, and we additionally show all the vertical group names in our bigger table. The need to address the astatine challenge becomes even more relevant when we consider that tennessine is predicted to be a metal.

The approach I am proposing is to categorise the nonmetals based on the way they are described. The metalloids as the most metallic of the nonmetals; the corrosive nonmetals as the most chemically active; and the intermediate nonmetals as being neither as metallic as the metalloids nor as active as the corrosive nonmetals.

There are no neologisms here, only plain English descriptive phrases.

There is no original research (OR) here, unless you count identifying descriptive commonalities among similar nonmetals as OR.

As long as our categories accurately reflect the way the elements involved are described, even though the precise terminology may not be settled in the literature, and we seek to avoid neologisms, I believe we will be good.

This is potentially the best categorisation scheme proposal we will have developed and finally sorts out what to do with the nonmetals in a way that is consistent with how the elements involved are described, and the way the metals are categorised. I say "potentially" since the acid test is to see if a sandbox nonmetal article incorporating the proposed new categorisation scheme would stand up.

Coincidentally, I see that green, which is the shading for the intermediate nonmetals category, is the intermediate colour in the ROYGBIV colour spectrum.

Sandbh (talk) 03:00, 12 June 2017 (UTC)

DePiep comments

• When going live, there definitely must be a blue link for Intermediate nonmetal and Corrosive nonmetal. Could be a R to #section of Nonmetal. -DePiep (talk) 19:30, 12 June 2017 (UTC)

• • If this ever goes live then yes, I agree: an R to #section of Nonmetal. Sandbh (talk) 02:29, 13 June 2017 (UTC)

Discussion out of hand by size and by structure

• This talkpage is 400k now. Already inaccessible by mobile. Is there a way to condense earlier parts, or manually move subsections to (a dedicated) Wikipedia talk:WikiProject Elements/Archive 28? -DePiep (talk) 19:30, 12 June 2017 (UTC)
  • I can access the talk page on my mobile. If accessibility is an issue I'd be quite comfortable if everything preceding the Revised proposal section was moved into a dedicated archive. Sandbh (talk) 02:29, 13 June 2017 (UTC)
    • re Sandbh: Sure my mobile does sho0w tha page., However, it only has the top section title "2 Reclassifying the nonmetals", then unfolding that opens the whole. So: tevchnically everything is there, practically it cannot be reached or overviewd. Then, yesterday, in the end on my desktop the page did not even render any more. 400k for a talkpage does not work. -DePiep (talk) 08:14, 13 June 2017 (UTC)
    • I have no objection against the archiving, although I think we should always keep the original chronological order: this is a record of discussions and as such is kind of expected to weave in and out of points, so the table of contents should probably just give signposts of where it might be going at any given time. I am quite amused to be taking part in yet another megadiscussion on categorisation, although I am now slightly worried that I might have to wait very long to get a chance to propose making group 12 post-transition metals... Double sharp (talk) 03:46, 13 June 2017 (UTC)

Double sharp the 'discussion' is incomprehensible (as you reverted my structuring). 45.000 words "weave in and out of points", yes. I am astonished that you don't see the advantage of a topical arrangement by topic and relevance. (REmember that I just reordered complete sections). I don't think people are able to follow the conclusions that some are drawing here. For sure, I am lost. -DePiep (talk) 08:14, 13 June 2017 (UTC)

I think that advantage is outweighed somewhat by the fact that the topical arrangement makes it more difficult to see what has been discussed and what has not; judging from Sandbh's comment above, he seems to have missed my argument from boron (the other five are coming) because it was put with the already-discussed sections on N, At, and Rn. As I and Sandbh have said, I would prefer if the sections before his revised proposal were archived, to make it clear that we have finished dealing with those. The rest should stay here in order. Double sharp (talk) 08:19, 13 June 2017 (UTC)

I think the best point for the archive to stop is the hiatus before the start of R8R's comments. If no one gets to it before then I shall archive the material up to there into Archive 27 later today: while my poor phone can take the 400k talk page, I am not sure it can take the massive copy-pasting this manual archive would need. ^_^ Double sharp (talk) 08:24, 13 June 2017 (UTC)

(ec) The Revised proposal does not state any conclusion from the 3-months long massive chemistry ===Comments===. Nor are there actual statements that stuff has been "dealt with". IOW, the massive word flow is has not been made related to the topic. Then, after the Revised proposal, you start a completely new section that does not relate to anything proposed previous (original nor revised).

• The TOC that could have been. -DePiep (talk) 08:34, 13 June 2017 (UTC)

There are two issues being proposed by Sandbh's classification:

1. Splitting nonmetals as intermediate-corrosive instead of polyatomic-diatomic;
2. Considering metalloids as a subclass of nonmetals, instead of their own separate class.

Both issues have been discussed above. The content below is a continuation going into even more detail.

As I understand it, the revised proposal is exactly the same as the original, except that it does not explicitly call the metalloids weak nonmetals. It nevertheless still considers them in the legend a subclass of the nonmetals. Double sharp (talk) 09:42, 13 June 2017 (UTC)

Metallic behaviour of metalloids

Shall we reposition this section? Does not seem to respond to the #Redefined proposal section. I suggest: make it follow current 2.6.3 #Leftover nonmetals categorising (keep level-3 === depth). After that, current 2.6.4 #The future of periodic table categorisation sort of changes route away from core element chemistry. (when concluded, remove this text allright) -DePiep (talk) 18:37, 12 June 2017 (UTC)

It is responding to this, because as I mentioned above one point I have not critically engaged with is whether it is really reasonable to sweep the metalloids as a subset of nonmetals. To do that I have to examine each of the six and talk about them from the metallic side, thus treating them as chemically very weak metals instead of very weak nonmetals; so since both can be done, it would then make sense to keep them as neither. So far I am only really satisfied with the section for B; I am thinking about the rest currently. Double sharp (talk) 03:28, 13 June 2017 (UTC)

re whether it is really reasonable to sweep the metalloids as a subset of nonmetals: that's exactly what the proposal is about. What actually was discussed then above, or better: how was that on topic? -- rethorically. -DePiep (talk) 08:40, 13 June 2017 (UTC)

The regrouping of the nonmetals as corrosive-intermediate (moving H and N), of course. As the second part of Sandbh's proposal it was most definitely on topic. Double sharp (talk) 05:44, 14 June 2017 (UTC)

Both here and in the earlier #Comments section re properties I am totally missing the connection with the proposals made explicit, and the conclusions that make these elaborate descriptions archiveable. As it is now, it may be exploring but not helping the discussion. Without structural improvements this discussion is bound to be bogged down without any effect. -DePiep (talk) 11:42, 16 June 2017 (UTC)

Boron

In some ways, boron is both the easiest and the hardest of these elements to mount a defence of its metallicity for. On the one hand, boron is a hypoelectronic element (it has three valence electrons but four chemically active orbitals), and all the other hypoelectronic main-group elements (groups 1, 2, 12, and the rest of 13) are quite clearly metals. The problem is that boron is only in period 2. It takes far too much energy to remove electrons entirely from boron; there is no chemically reasonable situation where forming B³⁺ becomes chemically feasible, whether it be in ionic lattices or in aqueous solution. Looking at its congeners, it seems pretty clear that boron would be a metal, if it did not happen to be so small. To some extent it therefore reminds me of hydrogen, which commonly forms what would be H⁺ in the limit (to make analogies with mathematics), but the size difference between H⁺ and Li⁺ means that a purely ionic formulation is inconceivable; but boron is yet weaker, as it doesn't even form "quasi-cations" like this (my bad phrase for things like H⁺ that only sort of exist in the limit, in this case of stronger and stronger acids, for simplification, but for which you can take their existence as a reasonable theory that gives the right answers).

What is characteristic about boron chemistry? Its difficulty in getting a complete octet is well-known; simple BX₃ compounds have tricoordinate boron with only 6 electrons, and are Lewis acidic, forming adducts like R₃N→BX₃. There are certainly analogies to carbon in the resultant sp³ hybridisation; but to be honest, this reminds me if anything of transition metal complexes. Now, I will be the first to admit that the analogy is not quite exact, as for quite a few transition metal complexes, particularly those overshooting the 18-electron count, the bonding is mostly electrostatic: in [Cu(H₂O)₆]²⁺, the central copper atom does not particularly care that it's ended up with 21 electrons. It mostly cares that it has a high positive charge for its size and drags water molecules towards itself to counterbalance it, if I may be permitted some personification. But for the most part, we can largely consider covalence to be the predominant explanation here, just like it is for boron; in fact, in the middle of the transition metal rows, covalence is the rule and true ionic behaviour is quite the exception. This is doubtless because the oxidation states involved are high considering the size of the atoms (e.g. Ti^IV is too high to be ionic, but Zr^IV, Ce^IV, and Hf^IV are intermediate, and Th^IV can be considered to be about as ionic as Fe^III, for instance); much the same is going on when considering the lack of ionicity in B^III compared to Al^III or Sc^III (intermediate), and then Y^III, La^III, and Ac^III (quite clearly ionic).

The hypoelectronicity also brings another analogy to the fore: because B is hypoelectronic, its signature is 3c–2e bonds (prominently in the hydrides). (For example, if you consider B and Si hydrides, there are similarities in outward behaviour as expected from the diagonal relationship – both are volatile and highly flammable compounds – but the Si hydrides look "normal" from the grade-school perspective while the B hydrides look odd, because Si is happy with 2c–2e bonds while B needs 3c–2e bonds.) In the limit of hypoelectronicity, we get electrons delocalised across many, many centres, which sounds like a perfect description of metallic bonding. This is where I would want to reference the heavier group 13 hydrides, except that they get increasingly unstable. Greenwood tells me that AlH₃ has 3c–2e bonding like the boranes, but despite the short Al–Al distance there is no direct metallic bonding. GaH₃ does not appear to be very different, as I expected from the d-block contraction. What would be interesting is knowing the structure of InH₃ and TlH₃, but they seem too unstable for it. ScH₂, YH₂, and LaH₂ are metallic hydrides (showing the intermediate character), culminating in YH₃ and LaH₃ as essentially completely ionic.

I think I've said that thinking of metallicity as a binary thing "does it form cations or not? does it form a basic oxide or not (never mind most of the B-subgroup ones)?" doesn't really do justice to the fact that there are things in between the limits of "normal nonmetallic behaviour" and "normal metallic behaviour", and I think this sort of thing going on with boron is an example. You can think of it not only as a nonmetal that happens to be electron-deficient as a weird exception, but also as a "nonmetallised" version of its heavier congeners, sort of like hydrogen in group 1.

Well, it's late here and it took a great deal of time to organise my thoughts for this; the rest should be easier to write and should come tomorrow. Double sharp (talk) 16:50, 10 June 2017 (UTC)

P.S. When we speak of acidic and basic oxides, we are really dealing with the "hydroxides" that form upon reaction with water. Now it is interesting to note that while B(OH)₃ (boric acid) is certainly acidic, it is not a Brønsted acid like most of these nonmetal acids are! It does not lose its protons to produce salts of BO₃³⁻. Instead, like Al³⁺, the acidity of B(OH)₃ comes from it acting as a Lewis acid via abstraction of OH⁻ from water!

B(OH)₃ + H₂O ${\ce {<<=>}}$ B(OH)₄⁻ + H⁺ (K = 7.3×10⁻¹⁰; pK = 9.14)

This is definitely much more like what happens with the hydrolysis products of [Al(H₂O)₆]³⁺, which similarly abstract OH⁻ from water to form Al(OH)₄⁻. Contrasting this with the formation of oxyanions by the later elements in the 3p row, I think it's fair to say that boron's behaviour here is more akin to a forcibly nonmetallised version of aluminium than a willingly nonmetallic ally of silicon (with which it has a diagonal relationship). To me, it is pretty astonishing that even here, hydrolysis for supposed "H⁺" and "B³⁺" species does not proceed far enough to create oxyanions. Double sharp (talk) 13:34, 12 June 2017 (UTC)

P.P.S. I must be tired. I forgot to mention that the fact that B is less electronegative than H makes a difference for its chemistry; for instance, hydroboration is usually anti-Markovnikov, the reverse of what happens usually with H^δ+–X^δ− reagents. The same is true for silane and germane, which explains why silane is more likely to form complexes with transition metals than methane. But I'm getting ahead of myself; silicon and germanium will be covered in the next section! Double sharp (talk) 14:44, 12 June 2017 (UTC)

Silicon

Placeholder for now; I may have to speed-write the silicon article for this from Greenwood, like I did N and the halogens in a frenzy last year. Double sharp (talk) 15:22, 12 June 2017 (UTC)

(Written after the archive, but I'm keeping this all in one place.) One thing I find intriguing is that the behaviour of Si in its oxoacid compounds and silicate minerals is rather like U; there is no real evidence for anything but traces of simple silicate anions, and polymerism like V, Mo, and W is widespread instead, with large polyhedral clusters of Si^IV which can be readily substituted by Li^I, Be^II, and Al^III. This is how it shows fewer anionic tendencies than an ordinary nonmetal, even though I will of course concede that unlike Ge²⁺, there is probably no Si²⁺. Similarly, I would expect much stronger metallicity in the second of each pair {B, Al}, {Si, Ge}, {As, Sb}, {Te, Po}, and {At, Ts}, with metallicity in the first being weaker and spottier and the same true of nonmetallicity in the second. It is interesting perhaps to note that liquid Si conducts electricity like a metal without the need for pressurisation, and is about as good as it as Hg. Double sharp (talk) 09:51, 10 July 2017 (UTC)

Germanium

One particularly interesting thing about Ge is that it can be made, with difficulty, to dissolve in acids. This is actually more than can be said than for Ti(!), for which HF is necessary to get things going, like Si. In light of this, significant metallicity is to be expected.

Germanium perhaps also offers the best example of the principle that metallicity increases as you lower the oxidation state. It is well-known that the main oxidation state drops from +4 to +2 as group 14 is descended. Now Ge^IV and Sn^IV are quite nonmetallic, while Ge^II and Sn^II are significantly more metallic. This suggests to me that the increased metallic behaviour of Sn compared to Ge is at least partially because Sn^II becomes a more important oxidation state, while Ge chemistry remains mostly that of Ge^IV. Similarly, the reduced metallic behaviour of Mo and W compared to Cr is mostly because Cr^VI shares its status of a common oxidation state with Cr^III, while Mo^VI and W^VI stand pretty much alone; even the large size of U^VI does not succeed in making uranium look very metallic in this oxidation state (it is certainly more so than molybdenum and tungsten, but is still not in the same class as Cr^III).

Another interesting remark that I have stumbled on in the notes to the metalloid article is that Ge²⁺ appears to exist as a simple cation in ionic compounds. For example, GeI₂ is isostructural to CdI₂, and many Ge^II oxoacid salts exist and are isostructural to those of Sn^II. (Of course, there is no hope for the possible existence of "Ge⁴⁺"; while the CdI₂ structure is sometimes intermediate between ionic and covalent bonding, that doesn't seem to stop us speaking of Be²⁺ and Al³⁺, and Ge²⁺ would actually be about the same size as Mg²⁺, Zn²⁺, and the late 3d transition metal dipositive cations.) Yet this simple cation has not been seen in aqueous solution – except perhaps with weakly complexing anions like ClO⁻
₄. Now this sounds very much like the behaviour of many of the late 4d and 5d transition metal cations; in the same oxidation state, we even have Pd²⁺ and Pt²⁺ as refreshing examples, which certainly at least seem to exist in salts, and then are actually only about 10 pm larger than Ge²⁺ is said to be!

I think we thus have another example when calling an element like this a "weak nonmetal" forces us to call pretty much all the transition metals "weak nonmetals". The standard high-school metallic behaviour of the main group metals, as well as a few exceptional transition metals (Mn, Fe, Co, Ni, Cu, Ag), is actually not the norm.

The problem is twofold. In high oxidation states, water is very readily complexed to form a simple solvation shell, but then hydrolysis happens very readily. In low oxidation states, water is only weakly complexed, and just about any other passing ligand does better. (An exception I guess is the s-block metals, but then not only is water weakly complexed, but so is everything else, and the primary solvation shells for things like Rb⁺ and Cs⁺ are not all that well-defined.) You need a particular fine-tuning of size and charge to get things to work the way you would expect them to from high school, and there is a good reason why high-school chemistry sticks to the 3d elements and silver alone from the 4d row. ^_^ Double sharp (talk) 03:27, 19 June 2017 (UTC)

P.S. GeO₂ is soluble in acidic solutions; as everyone should have expected by now, GeO is even more so. It may be more on the nonmetallic side than the metallic side, but metallicity in Ge seems to be fairly real. Double sharp (talk) 14:17, 19 June 2017 (UTC)

Tellurium

Placeholder