First of all, I'd like to thank all of the participants for an exceptionally high quality discussion. Based on my reading of the discussion, there is consensus to use the Sc-Y-La-Ac periodic table as a default. Both periodic tables should be presented when the controversy in group 3 is relevant, but it need not be mentioned otherwise. For example, there is no need to link an alternative periodic table in the article about chlorine.
The primary reason this conclusion was reached by the participants of this discussion is because the Sc-Y-La-Ac periodic table has been shown to be the more common form, both in general texts and among researchers who study the dispute academically. The science-based discussion also seemed to favor the Sc-Y-La-Ac periodic table, but at minimum the Sc-Y-La-Ac periodic table did not appear to be outclassed by the Sc-Y-Lu-Lr table.
The Sc-Y-*-** periodic table was discontinued by IUPAC during this debate, and was rendered historical. IUPAC is currently discussing this issue, but a decision is not expected to be finalized soon. If IUPAC decides on a Sc-Y-Lu-Lr periodic table, that will almost certainly require this discussion to be rehashed.
It should be noted that I am not a chemistry expert; I have had some college level chemistry, but I may have easily missed a subtlety in this discussion. To point out my blunders, or if you have any questions or concerns about the closure, please contact me on my talk page. Tazerdadog
) 05:48, 8 January 2017 (UTC)
- The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
I have noticed this has been an infected discussion in the past, and no resolution seems to have materialized sadly.
Currently in this template, Lu and Lr are directly inserted in group 3 and the additional rows have them excluded; but the IUPAC version shows them at the end their additional rows below.
Should we keep the current configuration or apply the IUPAC version instead?
Regard, →AzaToth 07:20, 30 October 2016 (UTC)
DS's proposed form: Sc-Y-La-Ac
- IMHO it ought to follow the periodic table as presented by prominent textbooks such as Greenwood & Earnshaw, Holleman & Wiberg, Cotton & Wilkinson, Shriver & Atkins, and Housecroft & Sharpe, putting La and Ac in group 3 instead of Lu and Lr. There are a few reasons for this.
- Firstly, this is the way the vast majority of chemists would have learned it, judging by how it is the dominant form in textbooks (I've already listed five of the most commonly used inorganic chemistry textbooks in my opening paragraph, and not a single one uses anything other than Sc-Y-La-Ac). (You can read a chemist's incredulous reaction to the -Lu-Lr form in Talk:Yttrium#Periodic group to judge how uncommon it is in academic circles that are not specifically debating the composition of the periodic table: Petergans seems to never have been aware of the -Lu-Lr option before this.) When common usage conflicts with IUPAC, the former generally wins on Wikipedia: hence we have articles on phosphine, arsine, stibine, and bismuthine, instead of the IUPAC-approved phosphane, arsane, stibane, and bismuthane.
- Secondly, if we look at electron configurations in the solid phase (more relevant for chemistry than the usually provided gas-phase electron configurations), most of the lanthanides and actinides (with only seven exceptions out of thirty: Eu, Yb; Th, Es, Fm, Md, No) have an fnd1s2 configurations. If you go across the lanthanide row, what appears to happen is that a 5d electron is added at La, and it stays there throughout as the 4f shell fills from Ce to Lu (except at Eu and Yb where it drops down to bring the 4f shell up to being half-filled or fully filled, but this is not a problem given that much the same thing happens with Cr and Cu as the 3d shell fills). Much the same happens for the actinide row. This implies that one column of the d-block fills, the f-block intervenes, and then the rest of the d-block fills. This is something you see in the Sc-Y-La-Ac table only, not the IUPAC Sc-Y-*-** table or the Sc-Y-Lu-Lr table currently given.
- Lastly, the IUPAC Sc-Y-*-** table is somewhat unsatisfactory as it does not provide a good answer to the composition of group 3. It seems to imply that there are 32 elements in group 3: Sc, Y, and all 30 lanthanides and actinides. This view has been held in the past, but it is emphatically not the view you will find in most textbooks today in English or German (don't know about other languages), which is that group 3 (like any other transition group) has only four elements: Sc, Y, La, and Ac. Even Cotton & Wilkinson, while covering these elements with the other 14 lanthanides (Ce–Lu) on the grounds of chemical similarities, specifically entitles the chapter "The Group 3 Elements and the Lanthanides", making it quite clear that these are separate categories. Furthermore, when expanded out into the 32-column format as in the footer, it seems to imply that the Sc and Y boxes be stretched to cover fifteen columns each, which is not something I have seen anybody at all seriously suggest.
- In light of this, I support that a Sc-Y-La-Ac table be used instead of the current Sc-Y-Lu-Lr table or the ambiguous IUPAC Sc-Y-*-** table. I get the feeling that there is a tendency to overestimate the scale of the debate on the composition of group 3. It seems to stem largely from a vocal minority, including for example Jensen and Scerri, that has repeatedly argued for the Sc-Y-Lu-Lr table in many articles. But if you look at chemistry education circles, and what professional chemists think, the debate looks like it was concluded long ago in favour of Sc-Y-La-Ac. It does not hurt that there are also sound chemical arguments on this side. The navboxes and infoboxes would then be edited accordingly to move Sc and Y. Double sharp (talk) 08:47, 30 October 2016 (UTC)
A couple of questions from me: [R8R 10:35; sign added to more easily follow threads. -DePiep (talk) 13:56, 10 November 2016 (UTC)]:
- Is the IUPAC choice, whatever it is, going to actually affect the decision? Because if so, why not postpone the discussion given that they are working on this issue now as well? [R8R 10:35]
- If it is going to affect our decision, then would it be a straight following their line, whatever it is, or will it have only some partial influence? If the latter is the case, will their reasoning have effect? [R8R 10:35]
- If these books come in disagreement with IUPAC, which source would be more prestigious? [R8R 10:35]
- Well, you did convince me with group 12 that just because IUPAC says something (e.g. group 12 isn't in the TMs) isn't a reason to do it if people don't always follow it. ^_^ This is even enshrined in our MOS with phosphine and the rest of the pnictogen hydrides. So, to my mind, there are no grounds to use -Lu-Lr until and unless it becomes the most popular form among chemists. Double sharp (talk) 12:19, 30 October 2016 (UTC)
- I see your point. I just believe it is important to determine what our priorities are first. You and I don't make all of Wiki and we intend to write a universal rule.--R8R (talk) 13:23, 30 October 2016 (UTC)
- We already kind of have one in Wikipedia:Naming conventions (chemistry), which says "Some articles on Wikipedia have non-standard titles through consensus that this is the most commonly used name (in scientific circumstances) for the compound concerned, whatever IUPAC or the other rules suggest". Hence we have "phosphine" instead of "phosphane" and "ethylene oxide" instead of "oxirane". So, to my mind, it would not be much of a stretch to say "regarding periodic table layouts and classifications, such as group 3 and group 12, some articles may not follow IUPAC conventions provided that the alternative used is in fact at least as common in the scientific literature". Double sharp (talk) 16:34, 30 October 2016 (UTC)
- I am skeptical given we have WP:ALUM. Rules have been used somewhat deliberately. I sincerely don't see the point to force everyone to write "alumunium" when they use AmE: Americans don't use that! Likewise, the British use, to my knowledge, only "sulphur" and I don't see why they are forced to use the unusual for them spelling. We may stick to the IUPAC (though I wouldn't want to), but then we have no phosphane and that's inconsistent. So apparently we're on our own.--R8R (talk) 11:41, 11 November 2016 (UTC)
- I think the case of sulfur in the UK is different from the case of aluminium in the US, given that for the former the Oxford Dictionary says 'The traditional British spelling is sulphur and the US spelling is sulfur. In chemistry and other technical uses, however, the -f- spelling is now the standard form for this and related words in British as well as US contexts, and is increasingly used in general contexts as well.' We are not yet at the point where the spelling has been unified across the Atlantic, but we are getting close (which agrees with my experience). Besides, IUPAC does not allow "sulphur" for S.
- As for Al and Cs, I would argue that the concept of a primary spelling is very useful. Is the category going to be "Aluminium compounds" or "Aluminum compounds" when you have dozens of articles, some using "-ium" and others "-um"? IUPAC can only print one name for them in the periodic table, after all. I think it is clear here that "aluminium" and "caesium" are the primary names with the American spellings here being minor alternatives. I know "phosphine" ignores IUPAC entirely, but it is a different kind of issue, since nobody says "phosphane". Many people say "aluminium" and "aluminum". Double sharp (talk) 11:57, 11 November 2016 (UTC)
- I remember having seen a Canadian teacher complaining, "who uses 'sulfur'? Nobody does!"
- As for primary spelling, I don't mind having one. But there's a difference between a "primary (to be used in broad contexts such as categorizing as a sort of agreement) spelling" and "the only one you may use." Read WP:ALUM yourself.--R8R (talk) 12:10, 11 November 2016 (UTC)
- Well, if you're asking me what I think for Al and Cs, I do think that there could be more allowance given to "aluminum" and "cesium", although probably not "sulphur" today. We survive with Category:Color having a mix of "color" and "colour", after all. I suspect the real reason why the rule is the way it is is because the consensus at Wikipedia talk:Naming conventions (chemistry) is firmly in favour of using "aluminium" and "caesium" only. (BTW, looking at WP:ENGVAR and the article histories, this wouldn't actually change anything for the element articles, because they were all started at the IUPAC spelling; but perhaps some of the compounds might be different.)
- The main argument I see and am somewhat convinced by is that in the case of PH3, everyone calls it "phosphine", nobody calls it "phosphane", so we call it phosphine. That is straightforward. The case of Al is less straightforward. Most of the geographic world calls it "aluminium", but most papers you will see call it "aluminum", and given the percentage of Wikipedia viewers from the USA I suspect many will expect "aluminum", although perhaps not the majority. In such an unclear situation it is quite useful to have the IUPAC guideline. Double sharp (talk) 14:20, 11 November 2016 (UTC)
- Are you both making an analogy between preferred spelling and group 3 composition? Methinks science should be decisive, not (like choosing) a preferred language. And: if "popularity" indeed is to decide which group 3 set we show in top, that should be "popularity" as in "by relevant sources", i.e. sources that deal with group 3 issues and that make claims about this. For this, even these PT's (in the comment section) by Scerri (August 2015) does not count as a 'One more vote for Sc/Y/*/** (by Scerry even!)': while discussing TM elements, Scerri uses a Sc/Y/*/** graph without actually addressing or correcting that group. He just used a PT that was lying around, to mark some elements of topic. The problem is: there are just way too many Sc/Y/*/** PTs out there. Same for the IUPAC PT this RfC opens with: that's not a vote, let alone a science based claim. They just copied something from from a wallsize classroom PT, that was copied from a five year older one, ... etc, from Seaborgs 1940 version. -DePiep (talk) 06:20, 17 November 2016 (UTC)
- If we go for popularity, among sources for which group 3 is an important point of discussion, Sc-Y-La is quite clearly a winner. Every single inorganic chemistry textbook that covers group 3 in depth that I've checked uses this classification. (There are, of course, quite a few texts that use Sc-Y-* when group 3 is not important to what they are trying to say: I have ignored those as irrelevant per what you say.) Even Wulfsberg, which uses Sc-Y-Lu, does not really treat group 3 to a whole chapter the way, say Greenwood & Earnshaw or Holleman & Wiberg do. Even if we choose monographs, the only one I found covering group 3 apart from the lanthanides is R C Vickery's Scandium, Yttrium, and Lanthanum (published by Pergamon Press), whose support is evidently clear from his chosen title. Double sharp (talk) 08:21, 18 November 2016 (UTC)
Also a couple of comments: [R8R 10:35]
- "The books do it, and thus so should we" and "look at the properties!" are quite contradictory lines of reasoning. If you are going to present the idea that -La-Ac suits better property-wise, then look at both Sc-Y-La-Ac and Sc-Y-Lu-Lr. The current line on reasoning is one-sided. [R8R 10:35]
- I did look at -Lu-Lr, didn't I? I looked at the solid-state electron configurations and noted that "This [the split d-block the configurations imply] is something you see in the Sc-Y-La-Ac table only, not the IUPAC Sc-Y-*-** table or the Sc-Y-Lu-Lr table currently given." Even Jensen (a noted -Lu-Lr proponent) writes that when positioning elements, you look first at electron configurations and only second at trends, and the first trumps the latter. Since both -La-Ac and -Lu-Lr give reasonable trends, but only -La-Ac fits the chemically useful electron configurations (and the idea of using solid-state ones as more chemically relevant is Scerri's), -La-Ac seems to win here. I could certainly make arguments for why the -La-Ac trend is better to show as primary than the -Lu-Lr trend, but why should I when I think the electron configuration argument settles it already? Double sharp (talk) 12:19, 30 October 2016 (UTC)
- On a brief read, this didn't look like all of Jensen's arguments have been taken in consideration. Besides, you decide that solid phase is more important than the gas state, which also cannot be taken for granted and only agreed on. Also, if (sure, if) I remember correctly, electron energies are close in almost all of these elements for the d and f configs, which (if I am right) diminishes the importance of this: Cr falls out of line, and so do Pd, Cu, Ag, etc.--R8R (talk) 13:23, 30 October 2016 (UTC)
- They really are not close in the cases of La and Ac. You have to go all the way up to 15196.83 cm−1 before La deigns to occupy the 4f orbital as [Xe]4f16s2, past many more stable configurations that do not involve the 4f orbitals at all. In the case of Ac, the energy of the [Rn]5f17s2 configuration would be so high that ionisation actually takes place first. So, while Ce–Lu and Th–Lr might be able to survive being stuck in the f-block, La and Ac seem to be very unhappy with it. At least for the transition metal anomalies like Cr, Cu, Pd, and Ag, the Aufbau-expected configuration is usually the first excited state even if it isn't the ground state. The Aufbau-expected configuration for La and Ac is nowhere near that. Double sharp (talk) 13:51, 30 October 2016 (UTC)
- P.S. In many cases the Jensen arguments are lame. Some of them are disputed (4f involvement in La is highly suspect), some have been superseded by newer data (both La and Lu are superconducting at low temperatures), and even the chemical arguments stemming from the differences between Y and La (which aren't present for Lu) are simply a result of the larger size of La3+ as compared to Y3+. I could use the exact same arguments to "prove" that Be and Mg do not belong in the same group as Ca. (For instance, Ca dissolves in liquid ammonia but Be and Mg will not; Ca metal is fcc while Be, Mg, and Zn are hcp; the mp and bp trends show a discontinuous jump from Mg to Ca but decrease smoothly from Mg to Zn; Ca has low-lying d-orbitals that can be used for bonding while Be and Mg do not; CaX2 and MgX2 for X = F, Cl, Br do not share the same structure; Be and Mg have a rich organometallic chemistry like Zn while Ca does not.) Yet no one would agree with that because these are all trumped by the electron configuration argument. If they can be used to prove something like this that no one would agree with today, why do they suddenly become useful and important when deciding on the composition of group IIIA? I don't think they do. Double sharp (talk) 01:45, 31 October 2016 (UTC)
- Looks inconsistent to say that La is "unhappy" but Lu "might be able to survive being stuck" citing the data because when you look at Lu, they don't even show information for the 4f6s2 config! How can you judge then? "In the case of Ac, the energy of the [Rn]5f17s2 configuration would be so high that ionisation actually takes place first." that sounds very unlikely. I'd rather suggest they couldn't do much research because of a number of factors: a) actinium is rare, b) the research is difficult, c) actinium keeps decaying away, d) research costs money, e) radiation interferes with energies of electrons. They measured 6s235f for Lu. Surely the same couldn't be impossible for Ac. Just more difficult.
- Of course Lu doesn't show any information for a hypothetical 4f156s2 (what?!?) configuration! The 4f orbitals can only hold fourteen electrons! If you take a Sc-Y-La-Ac table, the f-block starts at Ce and ends at Lu, you are saying that 5d1 enters at La and is conserved throughout the next fourteen elements, so Yb is predicted as [Xe]4f135d16s2 and Lu is predicted as [Xe]4f145d16s2. You will notice that this is the same predicted electron configuration for Lu as with the Sc-Y-Lu-Lr table, so looking at Lu here is a little pointless. What we should be looking at is how much energy it takes to promote one 4f electron for Pr–Yb (except Gd) to 5d (in favour of -La-Ac), compared to how much energy it takes to promote one 5d electron to 4f for La, Ce, and Gd (in favour of -Lu-Lr). The former always requires much less energy than the latter.
- Radium is also a strong radioactive, but somehow we have data up to 7s52p. Besides, Ac+ has been shown to show a [Rn]7s5f configuration at 28201 cm−1. Given that all this has been measured, unlike the really spotty promethium data (I wonder why also), I am quite confident that it is true that "ionisation actually takes place first". Double sharp (talk) 14:20, 11 November 2016 (UTC)
- "Of course Lu doesn't show any information for a hypothetical 4f156s2 (what?!?) configuration!" -- oh yeah. Apparently I was lost for a second there. Please never mind.
- "What we should be looking at" really don't follow why and even if so, that actually weakens your position, doesn't it? Some more explanation, please.
- Data for radium comes from a source focused on radium. Data for actinium comes from a source that gives a shot on many elements. No wonder the former has more data, right? (I still don't follow: do you actually think that a [Ac+]+5f system is less favorable energetically than a [Ac+]+an infinitely distant electron?)--R8R (talk) 17:44, 11 November 2016 (UTC)
- Is it that impossible? The Moore source is quite good enough to give 5f and 12d levels for radon. Maybe in actinium the 5f level is really very high up? (Further, since 5f would penetrate the core quite deeply, maybe the electron–electron repulsion is enough to push it away?) Double sharp (talk) 02:13, 12 November 2016 (UTC)
- It feels so to me. If Rn does it, why wouldn't Ac? A not too detailed calculation on Ac seems more plausible. Anyway, we're doing some crystal ball guessing here, so I can't defend this position too strongly and I won't pretend I can; but apparently, neither can you, so let's leave it there.--R8R (talk) 18:35, 13 November 2016 (UTC)
- I cannot, of course, explain it; but I think the fact that 5f is listed for Ac+ suggests that they would have listed it for Ac0, if only they could find it. However, I think my point that the 5f level is too high to contribute in Ac is still valid. Double sharp (talk) 14:53, 14 November 2016 (UTC)
- Re P.S. I'll take your word for "disputed." I'll say superconductivity never mattered to me anyway. However, "are simply a result of the larger size" is a huge deal and you put "simply" to possibly to underestimate the importance of this, while I immediately think, "if it's so simple, could it be it's a structural reason?" As for comparison with group 2, I think you're misinterpreting the argument in a way you wouldn't convince a -Lu-Lr supporter. Group 2 would be expected to show the change, as does its s-block mate, group 1. Group 3, on the other hand, would be expected to do what other early d-block groups do, especially groups 4 and 5, and -Lu-Lr fits that trend better.--R8R (talk) 12:04, 11 November 2016 (UTC)
- The reason I don't think it's important is that there are numerous examples throughout the periodic table where the universally-agreed placement does not result from Jensen-style arguments. I can even cite examples within the same group. It is well-known that Hg is very different from Zn and Cd. ZnF2 and CdF2 are like the alkaline earth fluorides in having high lattice energies and being sparingly soluble in water, but HgF2 instead gets hydrolysed to HgO and HgF. Zn and Cd favour 4-coordinate tetrahedral complexes, but Hg actually favours 2-coordinate linear complexes. Now we ask: why does this happen? Due to the much higher effective nuclear charge on the outer electrons at Hg, HgII compounds are singular in group 12 because of the intense polarising power of Hg2+ and great covalent character of its compounds. The 2-coordination thing stems precisely from the ease of distortion of its electronic cloud due to its filled d10 configuration poorly shielded from the nucleus by the f14 configuration: if you have two ligands approaching a Hg2+ cation from the z-axis, the electron density gets concentrated in the xy plane, so much so that other ligands find it difficult to approach. So we have the lanthanide contraction working both ways; at Hf, Ta, W, and Re, it makes the 6th-period element similar to the 5th-period-element, but when we get to Au and Hg it makes them very different. What all of this tells me is that we should consider such effects to simply be second-order course corrections. The periodic table is meant to show the most basic relationships, not things like the Al–Fe link, the group n–(n+10) relationships, and all that. Therefore we should not be looking at this sort of thing to place elements, but rather simply place them according to what subshells are active in their chemistry. This leads to a clear answer: 4f becomes active for the first time only in Ce, so that is the start of the f-block.
- As for group 3 being in the d-block: I think there is a fundamental difference between group 3 and group 4. The 4f electron level does not fall below the 6s level and become chemically active until Ce (and similarly the 5f electron level does not fall below the 7s level until Th). Thus the f-block cannot be considered to begin until Ce and Th, and that is where the lanthanide and actinide contractions begin, because it makes no sense to me to consider La an f-block element when 4f is not doing anything for its chemistry. From that basis, of course group 3 should not follow group 4; the latter comes after the lanthanide contraction, but the former comes before it! Therefore the irregularity of Hf being almost exactly like Zr should not be expected to occur in group 3. Double sharp (talk) 14:20, 11 November 2016 (UTC)
In the first para, you make a good point, a very good one. And it will be great if you show how this can be applied to gold, since gold mainly shows 3+ rather than +, but nonetheless I take your point on Hg. However, in the list of your arguments, some of which are very good, I find this one not nearly so. First of all, this one seems to apply to mercury only. I can't see how you could apply this to lanthanum, but I'll take it the periodic table is a bizarre place because when you want to draw some simple trends, it turns out not everything so so simple and the oversimplicity breaks down. Mostly because I wouldn't want to argue about that, as this moves from scientific to philosophical, and I just feel we're moving from the issue. So will also declare this to be not too relevant unless you show me it is.
This is a La-centered argument, and a something-centered argument isn't the best. This one, I really don't like: you can replace "lanthanide contraction" with "d-block contraction" and see that Al should be located above Sc. See why you don't like that statement and how you would disprove that claim: this is exactly why how I would disprove yours.
I also want to specify that I am, and have throughout this discussion, arguing with your arguments rather than the idea they are intended to support. I'd want to ask you: do you see the point (not if you agree with it; do you see it) behind me saying this all is a matter of preference rather than actual science? Don't even want to convince you this is the case; just genuinely interested to know the answer.--R8R (talk) 15:18, 11 November 2016 (UTC)
- Yes, I think it is partially a preference thing. I think that the choice of -La-Ac and -Lu-Lr depends a great deal on the philosophy of what you think the periodic table is meant to show first. However, that basic philosophy also makes a difference on the conclusions you would draw on many other fraught elements in the table, like H and He, Be and Mg, etc. Thus my argument for -La-Ac is mostly based on the fact that the philosophy of the standard table, in its placement of those other elements, seems to point to a -La-Ac sort of philosophy for me.
- For example, to me the philosophy leading to -Lu-Lr is what Jensen and Scerri have said: we think of the table as being based on abstractions, like ideal electron configurations based on a simple rule like Madelung's. The table thus splits neatly into rectangular s-, p-, d- and f-blocks, and we judge things that way. Thus we look at group 3, and think 'oh, it should act like group 4, and not so much like group 2'.
- However, I think the fact that we almost always shift helium to the noble gases illustrates to me that there are really grounds to break away from such a Platonic table when the chemistry does not support it. Instead of starting from the theory, and placing helium with the alkaline earth metals (complete with caveats as to its actual chemistry), here we have a profound statement in favour of scrapping the theory in a case like this in favour of what we experimentally see in chemical properties.
- I argue therefore that similarly, although the theory predicts that lanthanum and actinium should show f-orbital involvement, experiment says they do not. I see that every block of the standard periodic table only begins when its first subshell becomes chemically active, and the rest of the periods below it follow. This results in defining the theoretical blocks from the practical results, not forcing the practical results into a theory. The s-block starts at Li (ignoring the weird first row as an outlier), the p-block starts at B, and the d-block starts at Sc.
- I look at the Ca–Sc divide, and note that Ca can occasionally be forced to use the 3d orbitals for supervalent hybridisation, but that 3d is an important part of Sc chemistry. Then I am reminded of La–Ce, where La can occasionally be forced to use the 4f orbitals similarly, but 4f doesn't become important until Ce when it falls down in energy. (Yes, really: the source is 10.1021/ed8001286. This is what I was trying to say in the "lanthanide contraction" argument; I must not have expressed it well enough.) Then I think that if Ca isn't a d-block element even though it can use its d-orbitals in hybridisation, neither should La be an f-block element. And I look at the metallic lanthanides and I see that the d-electron added at La is mostly retained (only two exceptions, Eu and Yb); I look at the metallic transition metals, and the s-electrons added in groups 1 and 2 are similarly retained (only two exceptions, Rh and Pd). This is why I want to start the d-block at La and interrupt it for the f-block lanthanides, to more closely reflect when the f-orbitals first become active.
- And this results in a table that is consistent with our choice to move helium, because while the theory is perfectly happy with it above beryllium, almost every chemist who has worked with both elements would consider such a placement ridiculous. May I also add that it feels more comfortable to my preferred philosophy of science if our theories are based on the facts, instead of us forcing the facts to fit our theories?
- P.S. For gold, I would note that while CuI and AgI form aquo ions, AuI does not, and while CuI and AgI form complexes with unsaturated hydrocarbons, AuI does not. The +1 oxidation state for gold is known and is reasonably stable; it is indeed important in its extraction by formation of [Au(CN)2]−, although it is more susceptible to oxidation than for silver and so you have to keep it away from water (but once you do, there is no problem). You are right of course that it is easier to make this kind of point where all three elements in the group like to show the same oxidation state; but even here I can make the point about size again, because CuCl has the zincblende structure while AgCl has the rocksalt structure similar to NaCl and KCl. The issue is simply: are the ions roughly the same size (in which case you get the CsCl structure), is the cation a little smaller (in which case you get the NaCl structure), or is the cation much smaller (in which case you get the CuCl structure)? And the crucial thing is, the size is the only factor here, and it really has not much to do with how the elements actually behave as long as the stoichiometry works out. That's why, despite AgCl and NaCl having the same structure, Ag is very unreactive and Na is famously reactive. So to me, just saying that La is bigger than Y while Lu is about the same size by itself is not saying much. K is bigger than Na while Cu is closer to being the same size. Double sharp (talk) 15:56, 11 November 2016 (UTC)
- This line of reasoning in incomplete in a way that I've noted before: When you have an argument for -La-Ac, you are left to show the same couldn't be said about lutetium. If a comparison is impossible, then there is no point. You've defeated my counter-argument to your argument above, which I made by mistake. But nonetheless, that doesn't make that argument of yours any better, and you had to replace it with another one, which I don't quite understand. See?
- I was unable to meaningfully apply your argument to lutetium and period 4. I think this disqualifies it.--R8R (talk) 17:44, 11 November 2016 (UTC)
- Let me try to make the comparison very clear. If I look at the first row of the d-block, I notice that we start it at Sc where the 3d orbitals first become active and end it at Zn where they finally sink into the core and become chemically inactive. This is true for every row of every block except the s-block (we can take the activity of Kr, Xe, and Rn as a minor exception). A case could be made for starting it at Ca and ending it at Cu, because Ca shows some 3d involvement under extreme conditions, but that is not what we do. So this acts as the precedent I am citing.
- Now if I look at the first row of the f-block, I notice (citing Greenwood, Schwarz, Johanssen and many others) that the 4f orbital first becomes active only at Ce. This is the main point I am trying to make. Despite a flurry of speculations in the early 1980s literature, Sandbh and I were unable to find any further references which actually firmly substantiated 4f character in La, and even the "evidence" in the early 1980s literature for it could be more easily used to support the absence of 4f character. It is because of this that I think it is better to begin the f-block at Ce and create Sc-Y-La table, when the 4f orbitals first become active, and end it at Lu where they finally sink into the core and become chemically inactive. Whereas if you start it at La to create a Sc-Y-Lu table, you're creating a Ca–Cu d-block situation, because you're starting the f-block with La with no 4f involvement except under extreme conditions and ending it at Yb where the 4f orbitals are still active. Double sharp (talk) 02:04, 12 November 2016 (UTC)
@R8R Gtrs: P.S. This argument cannot be used to "prove" that Al must be placed above Sc. In period 4, 3d first drops below 4s at Sc, so the d-block has to begin there. Therefore, if we are building up an 8-column table in the first three periods, we have to expand a new block and rip Mg apart from Al, so that Ca goes below Mg, Ga goes below Al, and the ten columns Sc–Zn of the d-block go between them.
Now, if we apply this to period 6, 5d first drops below 6s at La, so the d-block has to begin there. 4f does not drop below 5d and 6s at La; it does so at Ce. So the f-block has to begin there. Therefore, if we are building up an 18-column table in the first five periods, we have to expand a new block and rip La apart from Hf, so that La goes below Y, Hf goes below Zr, and the fourteen columns Ce–Lu of the f-block go between them. The reason this argument is not applicable to Lu is that the facts just do not support it. This is not a matter of subjective opinion here: if you accept the principles of blocks as starting only when their subshell first becomes active, then we are drawn inexorably to the conclusion that La cannot be considered the first 4f-element as the 4f subshell is not yet active there. Double sharp (talk) 08:35, 13 November 2016 (UTC)
- Okay, point taken. I'll rephrase my original saying to, I think, more closely match my feeling about this: this is kind of question that relies on a) La being in group 3 and b) Lu not being in group 3. Both are relative; personally, I readily take a position that there's no such thing as an easy way out (as we're having this discussion in first place, as Sandbh mentioned he'd changed his side three times, etc.). Therefore, I believe, to more or less convincingly point that group 3 is -La-Ac one must show that -La-Ac is a good fit in a way that -Lu-Lr isn't. In this argument, you do say -La-Ac is a fine fit but then you don't say how this excludes the possibility that -Lu-Lr is even better.
- All this said, I don't reject the whole idea; only this argument.--R8R (talk) 18:29, 13 November 2016 (UTC)
- As I've said many times, if you use -La-Ac here, the f-block corresponds precisely to the filling of the f-orbitals, from their initial transition to activity at Ce and Th to their completion at Lu and Lr. If you use -Lu-Lr, then you start the f-block when the f-orbitals are not yet involved at La and Ac, and you end it early at Yb and No when it can still participate in chemistry and be ionised to a stable f13 configuration. The former certainly looks better, especially since it looks closer to Sc–Zn and Y–Cd in the d-block (where as a -Lu-Lr group 3 gives an f-block that would be similar to a "d-block" consisting of Ca–Cu and Sr–Ag). So, I think that not only does this show that -La-Ac is "a fine fit" as you put it, but that -Lu-Lr is worse.
- (I must give credit to Lavelle's "pair out of place" argument here. He notes that since La is [Xe]5d16s2 and Ac is [Rn]6d17s2, it would be the only example of a column in the f-block in which neither element had any f-electrons at all outside the previous noble gas core. What pushes it over for me is the total lack of any substantiation for 4f and 5f involvement in La and Ac respectively. I think he must have been thinking of that in how he insists that Th as [Rn]6d27s2 can still be placed in the f-block while Ac does not; while he looks inconsistent and was rightly scolded by Jensen for that, he is actually completely right if you consider condensed-phase configurations, where Th has fds hybridisations amounting to half a 5f electron on average, and consider the configuration of the Th3+ cation, which is [Rn]5f1 in the ground state. Contrast that to Ac whose valence band is purely (6d7s)3 and for which Ac3+ has the simple noble-gas configuration [Rn].)
- And yes, I agree that Sandbh and I have changed our sides on it way too many times (I think I started as -La-Ac, flip-flopped to -Lu-Lr after reading Jensen, back to -La-Ac after reading Greenwood, back to -Lu-Lr after reading Sandbh's summary, thought about it again when writing the La article, and went back to -La-Ac a few months ago). I do think this sort of argumentation based on what happens elsewhere in the periodic table can give a reasonably straightforward solution in favour of -La-Ac, of course if we agree that He remains above Ne and Be and Mg remain above Ca. The reason I didn't bring this up is that until the middle of this year, I do not think I actually ever went to think that deeply about what the principles of the periodic table as it is drawn actually are. (This is not an idle question, because I don't think they are actually ever explicitly officially stated.) Double sharp (talk) 15:01, 14 November 2016 (UTC)
- "The books do it, and thus so should we" and "few people even care" are quite contradictory as the latter diminishes the intensity of the former. If we presume the books don't care and a commission from a very important organization makes a detailed decision on the issue, it will probably overrule the books that don't even care, right? [R8R 10:35]
- It's not that people don't care. That's not what I'm trying to say. I'm trying to say that most people in the field seem to act as though the debate has already been long concluded in favour of Sc-Y-La-Ac. It's not as if they're saying today "oh, the precise composition of group 3 doesn't matter because all those elements are similar". No, they're saying "group 3 is Sc, Y, La, Ac", without acknowledging that there is any debate at all, acting as though the decision has already been made. (Maybe part of this is because the RSC and the ACS both use -La-Ac in their publications exclusively, unless of course the composition of group 3 is the subject under debate, which would explain the near-unanimity behind -La-Ac among English-speaking chemists. So you'll find that, while in general chemists may differ on why exactly they think -La-Ac is a superior presentation to -Lu-Lr, they will not differ for the most part in that conclusion.) Double sharp (talk) 12:19, 30 October 2016 (UTC)
- I unintendedly exaggerated, of course. Still, your last sentence, or at least its wording make me unsure and that makes me suspicious. "So you'll find that, while in general chemists may differ on why exactly they think -La-Ac is a superior presentation to -Lu-Lr, they will not differ for the most part in that conclusion." -- what you wrote above does not read like they are all sure this is superior; that's, as I see it, is what they're used to. Which is a strong point, too, noted, but I don't see how this "widely believed superiority" takes place.
- Petergans and Axiosaurus on WP seemed representative, and their reactions to -Lu-Lr at WT:ELEM were both to raise arguments about why it should not be preferred (and the former referred to -Lu-Lr in WebElements as "nonsense on the Web"). One other thing I note about Lavelle's article is that although some of his points are lame, he appears to feel very strongly about his conviction in favour of Sc-Y-La-Ac. You can really feel the "someone-is-wrong" vibe in his article. Even if this stems from the fact that he learned it that way, he certainly tries to prove it is superior (after all, how else do you convince people in science?). Double sharp (talk) 13:51, 30 October 2016 (UTC)
- It also contains statements on matters that a different person could take differently. [R8R 10:35]
- So, if Jensen puts on this "my-opponent-is-wrong" mode, will his arguments be any better? Why does it work with Lavelle?
- I've had enough math in university to say two people can't be representative. Speaking sociologically, these are two people who bothered to say something rather than silently observe. Speaking mathematically, this is just bad. Two people. Come on.--R8R (talk) 17:44, 11 November 2016 (UTC)
- I don't think his "my-opponent-is-wrong" mode improves his arguments. In fact, I think it makes him forget himself and open holes in them (like the Th vs Ac thing, where his base is right, but he messes up the wording). It is just an illustration of how -La-Ac proponents usually think -Lu-Lr is not as good. Whether it is because they learnt -La-Ac (like Lavelle, and somewhat me), or because they thought about the thing in detail (I hope to be able to include myself in this category), that's the general response. Sc-Y-La-Ac is definitely the mainstream viewpoint and it is -Lu-Lr that faces an uphill struggle in getting accepted. Whether I think it should be accepted (I obviously think it shouldn't now) is not relevant to this particular point. Double sharp (talk) 08:42, 18 November 2016 (UTC)
- "the IUPAC Sc-Y-*-** table is somewhat unsatisfactory as it does not provide a good answer to the composition of group 3" -- I've seen people argue it is good exactly because it does so. [R8R 10:35]
- Yes, but is that really the majority view today? I'm not convinced that it is, especially when the first five inorganic chemistry textbooks I've checked (and they're all very common) unanimously support that the composition of group 3 is clear and that it includes lanthanum and actinium. Double sharp (talk) 12:28, 30 October 2016 (UTC)
- This sounds better than "it's unsatisfactory." I'll only note that many people elsewhere use -*-** and are just fine.
- "it seems to imply that the Sc and Y boxes be stretched to cover fifteen columns each" -- not necessarily. This is not necessarily meant to ever be expanded. [R8R 10:35]
- Except that we keep using the 32-column table in our navboxes, so it is a problem for us. We cannot really logically use anything other than a Sc-Y-La-Ac or Sc-Y-Lu-Lr table if we keep using the 32-column table. I suppose one could argue for changing that too (for the record, I would support that), but that is not the purpose of this RFC. The purpose of this RFC, as I see it, is simply to decide on whether we put lanthanum or lutetium under yttrium, and then enforce that as default throughout WP. Double sharp (talk) 12:19, 30 October 2016 (UTC)
- I must intervene, re "it seems to imply that [Sc and Y stretch over 15 columns]": No, it is actually so. Not 'seems to imply'. The graph states that Sc and Y stretch those columns. Both in 18-column form, and in 32-column form (check: just resolve the placeholders for the lanthanides and actinides). On top of this, those columns are all under number "3": group 3. You can not haggle with this graphic fact. -DePiep (talk) 23:33, 3 November 2016 (UTC)
- "I get the feeling that there is a tendency to overestimate the scale of the debate on the composition of group 3." -- please take it with a note of irony, but it was you who wrote such a lengthy treatise on this issue. [R8R 10:35]
- Yes, I know. I recognise that I used to be one of those "periodic table fans" I criticise today. (That's why I still have patience with those IPs at extended periodic table: I was once like that.) But I find that once you study enough and read enough on inorganic chemistry in general you tend to realise that this is a very minor thing, to be maybe covered in one page at the very most in a thousand-page textbook. Double sharp (talk) 12:19, 30 October 2016 (UTC)
- Double sharp, you're changing your point of view so fast. Not so long ago you said we should just follow IUPAC with whatever they decide? [R8R 10:35]
- Like I said, you convinced me away from that idea with group 12. Can't I change my mind? ^_^ I was once a -Lu-Lr supporter, you know. It's just that I did more and more reading and convinced myself that -La-Ac was better. Double sharp (talk) 12:19, 30 October 2016 (UTC)
- Sure, you can, this just came out unexpectedly.--R8R (talk) 13:23, 30 October 2016 (UTC)
I suggest we take our time to wait what IUPAC has to say.--R8R (talk) 10:35, 30 October 2016 (UTC) end of post [R8R 10:35]
- I would indeed not have opened an RFC this early. However, it has been opened, so I felt bound to comment. Double sharp (talk) 12:20, 30 October 2016 (UTC)
- What I would probably do is to wait for IUPAC. Don't see the need for changes that go back-and-forth. A long-term solution is on its way because we will apparently take theirs as our own. Collecting comments is fine, I think. It's just that it doesn't look like a good time for action now. I think I've made your point stronger and I am unwilling to make mine, so I'll leave it here.--R8R (talk) 13:23, 30 October 2016 (UTC)
- OK, except that I'm not entirely sure if we will follow IUPAC's decision or defer to common usage. We already do have precedents of doing the latter (e.g. phosphine), and after all you proposed deferring to common usage for "element 113" and the like a while ago (I agreed then; would disagree now because I'm not sure that is common usage among all chemists). So I think that even if IUPAC decides to change it to -Lu-Lr, it is not a given that we convert too if no one adopts it. Double sharp (talk) 14:21, 30 October 2016 (UTC)
- I had no idea IUPAC was actually deliberating this question at this time and a decision is soon to be declared. Do you have any up to date pointers towards such discussion? →AzaToth 15:53, 30 October 2016 (UTC)
- Here's the latest article from them about the creation of the task force deliberating this question. Note however that it completely misrepresents the Sc-Y-La-Ac position by equating it with a stupid one that doesn't keep the atomic number sequence that nobody supports; hopefully their final report will be better. Double sharp (talk) 01:36, 31 October 2016 (UTC)
- I see; As with most decision processes in science, I assume a decision is many years in the future... →AzaToth 14:54, 31 October 2016 (UTC)
- The IUPAC page for the project says it started on 18 Dec 2015, and that the task group will next meet during the first week of April 2017. Once the task group complete their deliberations and arrive at a recommendation this will require a review by, "a large group of outside experts (~15) and a public comment period." So the process could take some time yet. Sandbh (talk) 10:32, 1 November 2016 (UTC)
Something else I have found from querying the NIST database: even when the prediction by the Sc-Y-La-Ac table for the ground state is wrong for the gas phase, it is always the first excited state (except for Pm, where there just is no data, and Yb where it is the second). For instance, Pr may be [Xe]4f36s2 in the gas phase, but its first excited state is [Xe]4f25d16s2, and that is its configuration in the solid phase. Whereas, when the prediction by the Sc-Y-Lu-Lr table is wrong (La, Ce, Gd), it is very far from being correct, always being a very high-energy configuration. This is also true for Ac and Th, but after that I ran into a lack of data. (In the case of Th, the expected [Rn]5f16d17s2 configuration from Sc-Y-La-Ac is the second excited state, while the [Rn]5f27s2 configuration from Sc-Y-Lu-Lr – well, let's just say you'll have to scroll down a long distance to find it. For Ac, the energy of [Rn]5f17s2 is so high that you'll end up ionising the atom if you try to get it to show that configuration.) Double sharp (talk) 14:05, 30 October 2016 (UTC)
(I've studied chemistry and worked in biochemistry, but I've never seriously thought about rare elements.) I like the version shown at the top of this section, with La and Ac used as placeholders. I like even more the IUPAC version, which makes it very clear that the placeholders are placeholders. I do not like the version currently at the top of Periodic table: the things are called "lanthanides" and "actinides", so the obvious placeholder elements to use are La and Ac, not Lu and Lr. Maproom (talk) 09:10, 1 November 2016 (UTC)
- In the top image "La" and "Ac" are not placeholders. "La" does not 'represent' the lanthanides, etc for Ac. They are simple tables cells containing exactly one element each (elements lanthanium and actinium). The placeholders in the top imagine are "*" and "†", and they hold the place for the two 14-element rows below (being the lanthanides and actinides except La and Ac). Would the placeholding be resolved, the bottom rectangle is inserted and we have a changed the 18-column graph into an equivalent 32-column periodic table (18 respectively 32 columns with elements). Both graphs state that group 3 (the column numbered 3) contains the four elements Sc, Y, La, Ac.
- The IUPAC table you link to, on the other hand, states that group 3 (column) contains 32 elements, either before resolving (as shown) or after resolving into 32 columns. Because: the "3" is spreading those elements.
- On the third hand, group 3 can consist of the four elements Sc, Y, Lu, Lr with scientific base. Again, it does not matter whether this statement is shown in 18 or 32 column graph.
- So: there are three scientific statements about group 3 constitution in view, each of them can be shown equally in 18- and in 32-column graph (six graphs). We need two independent decisions. -DePiep (talk) 19:45, 1 November 2016 (UTC)
- Thank you. I understand that now. I still prefer the IUPAC version because it makes it clear that there are 15 lanthanides. Worst is the table now at Periodic table, which might make a reader think that there are only 14 lanthanides, before we get back to a standard group-3 metal. Maproom (talk) 20:13, 1 November 2016 (UTC)
- You're free to hold the preference of course. But I note that this topic is not about showing which are the lanthanides. Core issue is the constitution of group 3 (and people here reject the IUPAC pt because it has the wrong group 3 content). Marking lanthanides etc. in this wiki is solely done by coloring them (category colors). Just as with the, say, metalloids. -DePiep (talk) 23:19, 1 November 2016 (UTC)
- I should add that in many of the inorganic-chem textbooks I mentioned at the start of the section, La is not considered a lanthanide (and Ac is not considered an actinide), leaving 14 of each only. You will never find Lu or Lr excluded on the other side, however. Double sharp (talk) 04:55, 2 November 2016 (UTC)
- Oops, I now see this sections title is RFC: ... the IUPAC version for lanthanides, and actinides?. So Maproom is not off-topic as I thought.
- Still, above I already answered that part of the RFC: we mark lanthanides and actinides by category color, not by table structure (i.e., not by putting them together in one column). Columns are groups. Also in the table structure: rows (periods) and apparently blocks (steps). But just think of how we would have to draw (show) the metalloids category this way? Same-colored cells (same-category elements) may be scattered across the periodic table, and coloring handles that nicely. Of course, for lanthanides and actinides 32-column table happens to solve the question nicely, irrespective of group-3 composition. -DePiep (talk) 08:09, 2 November 2016 (UTC)
For the La and Ac and the Lu and Lr periodic tables, the following two tables (courtesy of Sandbh) compare the idealized numbers of f electrons for period 6 and 7 f-block elements with the actual numbers of f electrons in their solid states (rather than their gaseous states). There are 6½ irregularities in the first table compared to 21½ in the second out of a total of 28 elements.
TABLE 1: Sc-Y-La-Ac periodic table f-block showing electron configurations
(light grey shading = match with idealized number of f electrons; dark grey shading = irregularity)
TABLE 2: Sc-Y-Lu-Lr periodic table
Double sharp (talk) 14:05, 30 October 2016 (UTC)
To clarify, IUPAC hasn't recommended any particular form of periodic table. They do have a periodic table on their site but it isn't an "offically endorsed by IUPAC" table, even though it's commonly referred to as the IUPAC periodic table.
The template we have now (which shows group 3 as Sc-Y-Lu-Lr) draws on the work of Jensen, and is informed by the electron configurations of the lanthanides and the actinides. This is discussed in our periodic table article. These configurations are for the metals in their gaseous form. It has recently been noted by Scerri, here and in the articles he refers to there, that the electron configurations of the elements in their condensed (solid) phases are more pertinent to their chemistry. That supports an Sc-Y-La-Ac template, as per the Tables section, above.
As mentioned by Double sharp, Sc-Y-La-Ac is the most popular form (and we say this with a citation in the periodic table article), despite Jensen and others advocating Sc-Y-Lu-Lr since 1982. I recall there was little controversy when the template was changed to Sc-Y-Lu-Lr. I'd expect less controversy if the template was changed to the most popular form (together with updated explanatory text in the periodic table article). And I'd support changing the 32-column table in our infobox to the 18-column form, as Double sharp has discussed elsewhere.
I'm inclined to not wait for the IUPAC task group. Articles change more or less continuously on Wikipedia, as new information comes to light. No doubt the same thing will happen when the IUPAC task group publishes their recommendation. If the Sc-Y-La-Ac form is currently the most popular form then that's the one that warrants appearing at the top of the article, unless better evidence suggests otherwise. Sandbh (talk) 11:02, 1 November 2016 (UTC)
- The last para contains a reasonable notion. Why I argue it would be better to wait is that this discussion has become somewhat lengthy. If we agree to do the change now, and then IUPAC makes a different decision, I am more or less confident those who participated in this discussion and especially those among them who supported the change will be thinking that the arguments had been laid down and nothing would need to be changed. I am not confident that the discussion will be the same if we halt until then. That's my what I'm worried about.
- In theory, we could do the change and agree now IUPAC won't matter whatever they decide. I, however, think it would be best to take in account their decision, which, as I have just described above, should be easier if we just wait.--R8R (talk) 00:10, 17 November 2016 (UTC)
- I think common usage is commonly agreed to trump IUPAC on Wikipedia, which is why we have phosphine instead of phosphane. It's only when common usage is unclear or varies geographically (e.g. spelling of Al, S, and Cs; propene vs propylene) that we use IUPAC as a discussion-ender. Given that I've never actually seen any reference work discussing the chemistry of the group 3 elements that includes Lu instead of La, I think the choice in favour of Sc-Y-La-Ac is clear even in the absence of all these arguments. If Sc-Y-Lu-Lr becomes more popular in future, we can revisit it then, but the way group 3 element is written (for example) is explicitly making the less popular form primary. This does not seem to mesh very well with policy. Double sharp (talk) 08:14, 18 November 2016 (UTC)
- Generally I would support this line of reasoning. But, say, spellings of those elements are imposed on Wiki to a far greater extent than you would need (those spellings are not discussion-enders in a regular Wiki way, they are the standard to comply with in every case, even though we're doing fine without that in most situations). Such standardization is not needed and violates the spirit of WP:ENGVAR. This displays some, in my opinion, excessive reverence towards IUPAC in this part of Wiki. This makes me question if this is what they do in here?
- What I don't like about this whole discussion is that it does not try to begin with basics. Those who chose names for superheavy elements, for example, first set up the rules and only then went for it. I'd want us, too, to define rules first and shift the focus of the discussion to that for now. You can be sure there will be people that would change the periodic table to the whatever form IUPAC chooses saying that they did it and we should follow. Popular pressure for something always makes me think if the thing being pushed should actually be approved.
- Maybe we'll decide to ignore them, for the time of publication. Okay then, but let's agree to do so ahead of choosing between the two variants.--R8R (talk) 13:15, 18 November 2016 (UTC)
- The spellings are only imposed in chemistry articles on WP, which is the only place where the IUPAC decision is in anyway relevant. We have Aluminum electrolytic capacitor and Aluminum Christmas tree under the American spellings. Anyway, the reason for our current practice is "there was a consensus to follow IUPAC here, just as there was a consensus to ignore IUPAC for ethylene and phosphine". See Wikipedia talk:Naming conventions (chemistry) for the previous discussions. (Of course, they are rather old, so one could discuss changing them now.)
- As I have said before, I am in favour of ignoring IUPAC at the time of publication regardless of what they choose, and simply going with common usage until that changes, as Sandbh and Silvio73 have suggested here. Double sharp (talk) 14:18, 18 November 2016 (UTC)
- Sure. No objections from me.
- Speaking of old discussions, can you help me find the discussion that resulted in WP:ALUM? I do remember seeing it, but apparently can't find it.--R8R (talk) 15:50, 18 November 2016 (UTC)
First, I need to split this RFC into three independent issues. 'Independent' says: we can decide on each of them separately. For readability, I'll detail on lanthanides when needed, and just say 'and for actinides, mutatis mutandis, the same' or simply 'etc'. IMO, these are the separate issues:
- 1. Show lanthanides and actinides by using the IUPAC graph.
- 2. Show the periodic table in 18- or 32-column? The presentation form.
- 3. Which elements in group 3? Options are: Sc/Y/Lu/Lr; Sc/Y/La/Ac; Sc/Y/*/* (=Sc, Y + all lanthanides and actinides = 32 elements). This is the scientific topic.
RE 1: Wrong approach by the OP. The table structure (columns & rows) does not exist to show the lanthanides etc. A table has columns (=groups) and rows (=periods). The PT also happens to have blocks showing (as an outcome of the physics, not as a setup). But lanthanides etc are a category (as we name that on this enwiki): the metallishness. All those categories are marked by color. So the lanthanides are shown by color. Simple: it would be impossible to adjust the table showing metalloids by column/row. I conclude: we shall not support this request.
RE 2: Whatever the scientific statement made by a PT: the 18-col and the 32-col form must represent that same statement. One can not change group 3 when converting between those two, such a change is unacceptable. I conclude: 18- and 32-col are/must be equivalent in content. Prescribing one form (18 or 32 col) is wrong (and usually indicates a misinforming).
RE 3: Between the scientific claims, this is actually the only real choice to be made in this RFC. I'd say, of the three options Sc/Y/*/**, 32 elements, is outdated (it was popular & common). The IUPAC team is still researching. We at enwiki can choose one as the preferred PT presentation, while describing the other two as wide as needed in relevant articles (as is done in the article periodic table). So far, I have no preference and I'm interested (not an expert) in the discussions. That is, both the discussions via WP:ELEMENTS and by the IUPAC-Scerri group. I conclude: (1) Sc/Y/*/** looks outdated; (2) between the two "group 3 = Sc/Y/?/?" options, we do need a preferred one (the one we'll use everywhere around enwiki), based on science.
- -DePiep (talk) 00:56, 4 November 2016 (UTC)
- I reread this RfC. I repeat my point: showing 18- or 32-column form is independent of the group 3 constitution to be shown. They are two unrelated choices. In other words, whatever group 3 is stated to contain (the scientific statement), both in 18- and 32-column form that statement must be shown correctly.
- - So, when going from 18 to 32 columns (or vice versa) that statement may not be changed. It is a graphic cut & paste exercise only. For each scientific statement on group 3, there is a correct 18/32 pair of PT graphs.
- - Now, wrt group 3, three scientific statements are know: group 3 = Sc/Y/Lu/Lr, Sc/Y/La/Ac, and Sc/Y/*/** (the latter says under "3": group 3 has 32 elements). Those three are unambiguous, and mutually exclusive. Note that Sc/Y/*/** does not leave any ambivalence about group 3: it is a fixed listing too.
- - Together, that's 3 statements × 2 forms = 6 periodic tables (in three pairs).
- - This being about scientific statements, our description & preference must be based on science. It may be unsolvable or 50/50, but hey don't blame the PT for it. If two (three?) claims are equally correct, it's up to us to solve that presentation question. But for sure don't corrupt its presentation for this.
- -DePiep (talk) 14:58, 10 November 2016 (UTC)
||This project will deliver a recommendation in favor of the composition of group 3 of the periodic table as consisting either of
- the elements Sc, Y, Lu and Lr, or
- the elements Sc, Y, La and Ac.
The task group does not intend to recommend the use of a 32-column periodic table or an 18-column. This choice which is a matter of convention, rather than a scientific one, should be left to individual authors and educators. The task group will only concern itselve with the constitution of group 3. Once this is established, one is free to represent the periodic table in an 18 or 32 column format.
- First, IUPAC has abandoned the Sc/Y/*/** option altogether.
- Second: Scerri c.s. supports
mythe idea to make presentation form 18 or 32 independent of group 3 constitution (Added: over the earlier "medium-long version" etc. mistake-pool. -DePiep (talk) 11:25, 11 November 2016 (UTC))
I propose: we should abandon the Sc/Y/*/** (32 element) variant right away. A nice historical version introduced by Seaborg, but now obsolete. -DePiep (talk) 21:04, 10 November 2016 (UTC)
- And yet they still explicitly are using the old 32-element version in the 2016 JWP report for the discoveries of elements 113, 115, and 117. The quotes are "The technique employed was claimed to distinguish between Group 3 elements (lanthanides and actinides) and combined Groups 4 and 5" and "No carrier-free actinide tracers were employed despite the extremely complex oxidation chemistry and adsorption quirks of those Group 3 elements in contrast to lanthanide behavior." Since neither Ac nor Lr desplays complex oxidation chemistry, this must be referring to the other actinides as group 3 elements as well, implying a Sc-Y-*-** table. Double sharp (talk) 02:10, 11 November 2016 (UTC)
- To me it looks like the notion 'group 3' is used too loosely and needlessly here. Eg, 'Group 3 elements (lanthanides and actinides)' says Sc and Y are not group 3 elements. Didn't they meant simply '[the] lanthanides and actinides', without invoking a (any) group 3 notion? Or did they really wanted to refer to 'group 3 elements & properties', not '[the] LN & AN elements & properties'? (I'm just reading, I can not argue the science described).
- Scerri's project remit I quoted was published 18 Dec 2016, the report you link to was written 15 Dec 2015. So back then it was still legal to use Sc/Y/*/** ;-)
- But alas, we'll wait and once the Scerri group has decided, no IUPAC publication can use that Sc/Y/*/** any more. -DePiep (talk) 11:25, 11 November 2016 (UTC)
- Intermediate proposal Given the Scerri's project The constitution of group 3, quoted above, I propose to reduce this RfC to being the choice between those two group 3 sets. The third (much older) option Sc/Y/*/** (=32 elements) is to be treated as a historical form. IMO, this does not prejudice the outcome, having read the discussion so far (which does not hint at that third option either). Meanwhile, the Group 3 article can have both remaining option presented equally, e.g. present two micro-PTs in the infobox, each marking one option. -DePiep (talk) 06:40, 17 November 2016 (UTC)
- It cannot only be historical if it was still being actively used just three days before Scerri's project remit was published. We can certainly say that it is now depreceated and should not be used, but we should at least acknowledge its existence in the lede before ignoring it for the rest of the article, perhaps by saying that it is still sometimes used in practice despite this depreceation. Double sharp (talk) 06:44, 17 November 2016 (UTC)
- Fine too. re "being actively used": yes still used, but that's not the same as "claimed to be a third group 3 version". As I wrote above, about the 2016 JWP report, it more looks like they mixed up "group 3" and "lanthanides and actinides" (without much damage, but hardly a scientific claim). A similar mixup I read in your very first link here, re Petergans in this talk (I rephrase): 'Lu can not be in group 3, because it is a lanthanide'. DePiep (talk) 07:44, 17 November 2016 (UTC)
- I should note that if you use the traditional definitions of the lanthanides and actinides (Z = 58–71 and 90–103) respectively, and define group 3 as Sc-Y-La-Ac (as Greenwood and Earnshaw do), there is no mixup in what he said. Double sharp (talk) 15:49, 17 November 2016 (UTC)
- That's the point: the traditional definitions. Double sharp: the traditional group 3 etc. etc. is dead. Tear down & burn your wallpaper PT, and print a new one from our en:wikipedia. Not yet perfect, but a better periodic table it is. And you can pick your group 3-quartet. -DePiep (talk) 18:38, 17 November 2016 (UTC)
- Wait wait. Did you just say that, by traditional definitions, 57lanthanum and 89actinium are not a lanthanide, actanide? -DePiep (talk) 22:47, 17 November 2016 (UTC)
- Yes, that's right. As Greenwood and Earnshaw put it on p. 1250, "The "actinides" are the fourteen elements from thorium to lawrencium inclusive, which follow actinium in the periodic table. They are analogous to the lanthanides and result from the filling of the 5f orbitals, as the lanthanides result from the filling of the 4f. The position of actinium, like that of lanthanum, is somewhat equivocal and, although not itself an actinide, it is often included with them for comparative purposes." (And yes, my wallpaper PT indeed has a Sc-Y-La-Ac quartet with the lanthanides being labelled as Ce–Lu and the actinides as Th–Lr. ^_^) Double sharp (talk) 02:43, 18 November 2016 (UTC)
- Double sharp, this might be a sidenote, but still worth fleshing out. In general, as I read the GandE quote, this is an example of a triple mixup. We ordinary's classify lanthanides as a category (metallishness properties). [for actinides all similar; skipped for shortness]. But I understand that GandE says "La is not an lanthinide, because of not 4f filling" (that's the "0" in the #Tables no 2., right?). But IMO, "f-orbit filling" is just about blocks not categories. And since La is not filling the 4f shell, this only means La is not in the f-block. I assume La fills a d-shell, and so belongs in the d-block of the PT. Which is, right below Sc and Y (who are undisputed d-block members). IOW, GandE state that the group is the familiar Sc/Y/La[/Ac], into the d-block. Only awkward effect: in the graph, the d-block left-to-right is interrupted by the true f-block. Still, the f-block fits the pattern of a 2-period step (esp in Janets), and d-block a 4-period rectangle. Also, since our categories can freely spread and form in the PT (no group- or block borders to be followed eg metalloids), it is no problem to have the lanthanides color run into the d-block (Sc-column). GandE could have said: "La is not in the f-block". What's left to do after this is the conclusion about group 3. (The triple confusion is: block/category/group). -DePiep (talk) 12:44, 22 November 2016 (UTC)
- Yes, what they are saying does not actually imply that La isn't a lanthanide (same for Ac), but rather that it isn't an f-block element (which, as you note, leads to a Sc-Y-La-Ac conclusion). You are indeed right about the electron configuration, where La is [Xe]5d16s2 and Ac is [Rn]6d17s2, exactly following the smaller Y ([Kr]4d15s2) and Sc ([Ar]3d14s2). One should however note that we categorise the elements primarily based on their properties: elements are categorised in the same category for having common properties. In the case of lanthanum, the lack of f-electrons (f0) makes it qualitatively different from the other lanthanides, so that for example there is no 4f spin–orbit coupling, no magnetic moment, and no fine structure in the bands of the La3+ spectrum, so in that sense La appears as a bit of an outlier or degenerate case. (To be fair the same is true for Lu which is f14, but we have a precedent in the transition metals as we currently categorise them where the empty Ca with d0 is not included but the full Zn with d10 is.) Of course, this is a minority view nowadays; we will simply carry along as we have and use the IUPAC categorisation, where La is a lanthanide and Ac is an actinide. Double sharp (talk) 14:16, 22 November 2016 (UTC)
- OK, clearer now. -DePiep (talk) 18:33, 22 November 2016 (UTC)
- I think the hardest problem to overcome is that old school scientists have learned & lived this old group 3 set of 32 elements without any problem — because it was not relevant. Part of the resistance to change is maintained by its the ubiquitous but unhelpfully difficult to unfold 18-column format. The group 3 statement was tied to the graphic presentation. Good the Scerri project will break this tie in the first place.-DePiep (talk) 07:44, 17 November 2016 (UTC)
@DePiep: I understand that you would like to get rid of Sc-Y-*-** except as a historical form, and have all 18-column and 32-column tables be consistent. I was wondering if you had an opinion on Sc-Y-La-Ac vs Sc-Y-Lu-Lr, since you hadn't expressed one. Or is either okay as long as we're consistent as a default, in your opinion? Double sharp (talk) 08:23, 18 November 2016 (UTC)
- TL;DR: Yes the wrong Sc/Y/*/** will have to disappear; I have no opinion between the other two; hope that Scerri concludes usefully (for our WP); and yes we can choose & change earlier.
- 0. No reply to that weird "La is not a lanthanide" claim by Greenwood etc above, is about category not group 3 (I hope).
- 1. Yes, I expect Sc/Y/*/** will have to disappear, when Scerri concludes. Given Scerri's project remit, and also following this and similar discussions, few people claim group 3 to be Sc/Y/*/** (32 elem). e.g., even the IUPAC table we are talking about does not source it. Current proponents for this group 3 version I check on what they actually mean: easily it is confused with preferring the 18-col table format (which is unrelated to group 3) or about the LN and AN series (=category, not group). For examples, see this RfC lede and your (Ds) "2016 JWP report"above in this thread & my 11:25 reply. So after Scerri reports I expect little reason to draw that 32-elements group 3 in PTs. Sure it should be described in history and background places. Actually it is quite special that the PT structure (a group) is being redefined, stating that the older version (say Seaborg 1940?) was wrong.
- 2. Between Sc-Y-La-Ac vs Sc-Y-Lu-Lr I have no opinion. A preference should be pointed out on scientific base (sources). I'm not that scientific enough, I can mostly follow but not argue with the reasonings. I object to arguments like "This one is as it was learned, the old habit, every scientist uses this one". Counting group 3 forms (in sources, schoolbooks, wallpapers) is not valid either: they picked a version that was not the outcome of a scientific debate (so far), or drew the group 3 sloppy (squeezing in the asterisks somewhere). And anyway, old and current books mostly use the
32-element group 3. The only PTs worth counting as a vote are those where that source bases its group 3 on arguments (ie the publication is stating & sourcing group 3). A bit like the metalloid lists starts.
- 3. For this encyclopedia, I hope that Scerri c.s. concludes a clear preference between the two. Of course we shall describe both (plus the old wrong one) in some articles. But as the ubiquitous boilerplate PT when discussing say (unrelated to group 3) halogens: we use that preferred version (either in 18 or 32 col, is an unrelated choice). But if Scerri ends with a 50/50 preference, it is up to us to decide and by which WP-grounds (educational ones? RfC voting?). And maybe Scerri could conclude to prefer that crippled version (not-continuous atomic numbers). Well, these are only a problem after Scerri publishes.
- 4. Earlier choice? OK with me. We already have a preferred version (as article periodic table says: Sc/Y/Lu/Lr). I'm fine with a change into Sc/Y/La/Ac as you discuss here. Even by more relaxed criteria as source-counting and -weighing, because that is reflecting the current state in science we must & should follow to publish. And it is only some months until Scerri publishes. For this early version too: better one for all those WP PTs. -DePiep (talk) 11:39, 18 November 2016 (UTC)
In Periodic table, sections #Group 3 constitution variants and #Period 6 and 7 elements in group 3 overlap. The first one, being about graphics only, could be removed. The second one talks science, and conveniently already omits the Sc/Y/*/** version. -DePiep (talk) 17:51, 27 November 2016 (UTC)
- Continued at Talk:Periodic table -DePiep (talk) 12:14, 29 November 2016 (UTC)
regarding: "Should we keep the current configuration or apply the IUPAC version instead?"
Show Both. Per WP:NPOV all important points of view are to be presented and explained. CuriousMind01 (talk) 11:57, 10 November 2016 (UTC)
- The problem with that is that it looks daft to give multiple periodic tables whenever you want to talk about an element's position there, especially if it's something like iodine which is nowhere near the contentious zone. Of course we talk about all three common versions in periodic table, but surely we need to choose one to use as a default? Double sharp (talk) 12:23, 10 November 2016 (UTC)
- In any event, there is no such thing as an IUPAC version since IUPAC does not have a position on a preferred form of table. Sandbh (talk) 20:35, 10 November 2016 (UTC)
- I should also note that as DePiep has pointed out, even IUPAC has now abandoned the Sc-Y-*-** version attributed to them, and is deciding only between Sc-Y-La-Ac (currently more common) and Sc-Y-Lu-Lr (the current configuration on WP, advocated by some for various reasons, although I disagree with them). So the only two possible outcomes of this RFC as I see it are: (1) keep the current configuration and apply it consistently, harmonising 32- and 18-column tables, including marking Lu and Lr as group 3 members in the infoboxes, etc.; (2) switch to Sc-Y-La-Ac and apply that consistently, harmonising 32- and 18-column tables, including marking La and Ac as group 3 members in the infoboxes, etc. Whatever we do, La/Lu (and Ac/Lr) will still be classified as lanthanides (and actinides), so it is only their spatial position that is in question. Double sharp (talk) 08:27, 18 November 2016 (UTC)
- From a Wikipedia's approach IUPAC is just one of the sources and unless it is not dediced that IUPAC is prominent against other sources, I don't see why it should be preferred, although I noticed that in some instances WP has preferred to use IUPAC definitions even when alternative are more common (e.g. propene instead of propylene). However, I just checked on three books at home. All of them (Morrison & Boyd - Organic Chemistry, Perry's Chemical Engineers' Handbook - Solomons - Organic Chemistry) use the Sc-Y-La-Ac template, with Lu-Lr at the end of the lanthanides/actinides block. Said that, my books are 20 years old, I don't now if more recent editions report the same version of the periodic table.
- It looks IUPAC propose a different template to overcome the issue of the Sc-Y-La-Ac or Sc-Y-Lu-Lr choice. Nevertheless, IUPAC template is somehow in discontinuity with the historical arguments from which the periodic table stems, namely but nor limited to the homogeneity of physical and chemical properties of the elements across the same group (please note that La and Ac have very similar properties). Somehow, the IUPAC template suggest the group III has 32 elements and not 4 (as all other d-block elements). My preference would be therefore for the Sc-Y-La-Ac template. Silvio1973 (talk) 09:32, 17 November 2016 (UTC)
- "It looks IUPAC propose a different template ...": I don't understand. AFAIK, this is the "IUPAC PT" (saying group 3 = 32 elements). But this is not a proposal to overcome ...', it is the predecessor from which the more recent pair of 'group 3 = 4 elements' come. IUPAC, and most other older sources, have just copied the original LN/AN extension (all into group 3). That IUPAC today still has it up is not a source, it just never has been evaluated (until Jensen, Scerri, etc., came along). The IUPAC version is not a scientific statement for group 3 any more. -DePiep (talk) 12:48, 17 November 2016 (UTC)
- Well, fair interpretation. It is a fact that IUPAC does not say the group 3 contains 32 elements, but their template for the periodic template strongly suggests it or alternatively should we understand that group 3 contains only 2 elements? The classification of elements in groups is a direct consequence of the periodic table, even if per se a statement is not made. However my point is another. This template  does not seem to be used in secondary sources. Silvio1973 (talk) 12:59, 17 November 2016 (UTC)
- Good point. (As a sidenote: I claim that "IUPAC PT says that group 3 has 32 elements", because the column heading "3" contains all LN and AN too (becomes clear once the placeholding is revolved). -DePiep (talk) 13:19, 17 November 2016 (UTC)
- BTW, regarding propene vs propylene; there is a regional difference on what name is used. (As another example, isopropyl alcohol is common in the US, but in the UK you will hear instead isopropanol. No one calls it 2-propanol! But I think laymen will know it better by the first name, so we use that.) So I think the WP policy is more like "we follow common usage, except when that is unclear, in which case we follow IUPAC". Double sharp (talk) 08:46, 18 November 2016 (UTC)
- Well, I work in the petrochemical and oil business and it is propylene in US, the UK, everywhere. No-one uses propene. Silvio1973 (talk) 19:28, 20 November 2016 (UTC)
- "propene vs propylene" is not part of this RfC. Stop it. -DePiep (talk) 01:14, 21 November 2016 (UTC)
- This one may depend somewhat on field. I've heard chemists using both 'propene' and 'propylene' and it seems about equally split, though maybe part of this comes from the fact that in English the systematic names are taught first (only for fresh-faced chemistry students to walk into their first lecture and get utterly confused by acetaldehyde, acetone, butyric acid, and phosgene...). But as DePiep says, this is off-topic here. Double sharp (talk) 02:05, 21 November 2016 (UTC)
Summoned by bot. Way too lengthy and complex discussion. Why not create two templates, explain the two differences, link each to the other and move on?Timtempleton (talk) 02:38, 25 November 2016 (UTC)
- Because, whatever gets chosen as default ends up having to appear on every element article, every article when the position of an element in the periodic table is important, every element category article, and so on. Certainly, when the composition of group 3 is the main topic, there can be multiple images shown. But that is silly for an element article like chlorine, for which this issue doesn't matter. We need to pick one. Double sharp (talk) 03:01, 25 November 2016 (UTC)
- re Timtempleton. What's wrong with a complex discussion? And lengthy by size? And don't forget: long by time? No problem in there, especially since it reflects the current state of science in this topic. IIRC, this was started in 1982 by Jensen. Since 2015, it is a formal dispute in IUPAC. On top of this, this encyclopedia must decide of how to present all that.
- An example. In 2013, we at enwiki decided to change the main category (background color) scheme of the periodic table (in real life science, it is not a simple fact so it was/is open for options). This archive page reflects the complex, lengthy, long discussion. A discussion about the state of science, and about how to show that in Wikipedia. In the end, we at enwiki have re-structured the periodic table. And the outcome has not been challenged since, so something must have been done well.
- About this RfC: already two intermediate results after the original RfC proposal are apparent from the fleshing-out. One: we will not (do not) show 'lanthanides' by column structure. They are noted by background color. Two: the IUPAC presentation saying "group 3 has 32 elements" is off the table (by IUPAC themselves). Out. Not 'resting', it will not return. As you write yourself: two other options are left to be considered.
- -DePiep (talk) 07:44, 25 November 2016 (UTC)
- My chemistry experience ended at Freshman chemistry in college, so I'm not qualified to chime in from a scientific perspective, but from a quick reading of the arguments it would seem that each camp makes good arguments, and that's why this has gone on for so long. Settling on a compromise will have to be the result. One will be eventually picked, but every opportunity should be given to allow readers to see that there's another popular option - just not as popular as the one chosen.Timtempleton (talk) 23:09, 25 November 2016 (UTC)
- The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Erratum to the above: since writing the above I have been alerted to a source listing a 6d→5f promotion energy for Ac at ~30000 cm−1 (listed at User:Sandbh/Group 3). However since this is even higher and closer to the ionisation limit than for the analogous 5d→4f energy for La I think my point is not seriously affected. Double sharp (talk) 16:59, 24 February 2018 (UTC)
I am seeking comments on a proposal to change the name and composition of two of the colour categories appearing on our periodic table, as follows: