Wikipedia talk:WikiProject Elements/Archive 55

Duplicate article
Introduction to superheavy elements is a duplicate of Introduction to the heaviest elements; only the latter is transcluded in element articles. –LaundryPizza03 ( d c̄ ) 16:47, 22 October 2020 (UTC)
 * Introduction to superheavy elements is up for deletion, see Articles for deletion/Introduction to superheavy elements. -DePiep (talk) 22:29, 10 December 2020 (UTC)

Picture help


How does one go about getting a picture rotated? This picture needs to be rotated 180&deg;. YBG (talk) 02:41, 19 December 2020 (UTC)
 * It seems this can be done using Transform-rotate, using the syntax from the documentation page. I haven't had a chance to test it myself, but it looks right to me. ComplexRational (talk) 04:11, 19 December 2020 (UTC)
 * Or you can just do it in MS Paint or probably something similar, if memory serves (sorry, on my mobile right now). Double sharp (talk) 04:14, 19 December 2020 (UTC)
 * ✅ Double sharp (talk) 06:27, 19 December 2020 (UTC)
 * Thank you!! YBG (talk) 06:45, 19 December 2020 (UTC)
 * Now added to c:Category:Art based on the periodic table :-) -DePiep (talk) 23:44, 19 December 2020 (UTC)

Isotopes: values from trends
With isotopes, some unmeasured values are estimated from trends. For example, see Isotopes of uranium, footnotes 3 and 8 (TMS, TNN).

These methods are mentioned in NUBASE 2016: TNN and AME2016 II: TMS.

Currently, pages TNN (nuclear physics) and TMS (nuclear physics) redirect to non-existent sections. The topic is not present in the target article nuclear physics. (Actually, therefor the redirects are to be deleted, correctly by itself).

My question is: is it possible to add these two methods usefully to an article, described in content? If not, the two pages can be deleted and their links in the footnotes are made unlinked. -DePiep (talk) 20:16, 17 December 2020 (UTC)
 * I searched these terms, and they appear to be extremely niche topics for which little more than a definition could be provided considering WP's target audience. Even finding sources explaining them clearly and substantially seems difficult, and if they were to be explained in another article, nuclear physics seems too broad (I'd have to give it some more thought and see if any feasible solution exists). That said, I wouldn't oppose the deletion of these redirects and unlinking them from the footnotes, so feel free to send them to RfD. ComplexRational (talk) 16:04, 18 December 2020 (UTC)

Conclusions to recent threads, as at 15 Nov
For those who are overwhelmed by the threads. Hopefully to be updated happily as news to ArbCom members that everything is settled without need for their involvement. Double sharp (talk) 23:17, 15 November 2020 (UTC)
 * Interpersonal conflict DS vs Sandbh is totally resolved and over. Not sure about the other one, but seems to be developing promisingly.
 * Argument over La vs Lu placed on freezer until IUPAC says something. DS is happy to have it at La for now, recognising he has made an attempt to change it already and that it has not resulted in a consensus in favour of his preferred Lu. He plans to revisit the situation when IUPAC says something (which based on info from Sandbh will probably start soon, albeit coming in various stages until the final IUPAC decision or lack thereof is made clear), as that will impact the source situation in a way that nothing else probably will . So that argument is settled.
 * Remaining categorisation questions raised civilly here and seem very close to being resolved.
 * Everyone is working together fine with no need to pass everything to external parties, although occasional questions where outside expertise might be helpful are of course still being brought up.
 * Plans have already been set in motion to improve periodic table the article, with EdChem's excellent proposals being kept in mind as a guide.

WP:ANI
Disruptive behaviour by DePiep, here. --- Sandbh (talk) 06:12, 26 December 2020 (UTC)
 * Commented. I don't think DePiep's suggestion to look for consensus first is unreasonable given the amount of discussion here where the consensus of opinion does not appear to support your position, Sandbh.  EdChem (talk) 07:19, 26 December 2020 (UTC)

One. I feel that EdChem’s statement has several misrepresentations and errors. I’ve asked him, at his talk page, to please address my concerns as he sees fit. Sandbh (talk) 11:29, 26 December 2020 (UTC)

Two. The edits to the PT article did not require consensus since a reversion of deleted content does not require consensus.

In any event I had the courtesy to notify of my intention to revert, and I waited for discussion, which ensued. I did this in order to maintain harmonious relations with Double sharp.

DePiep and YBG asked me to discuss before reverting. YBG agreed with me regarding the restoration of content addressing the wow factor. Double sharp offered his views. R8R offered to do the editing myself. DePiep objected to R8R’s solution. I commented. EdChem offered his thoughts.

I posted a list of items to consider in going forward. Double sharp and I discussed this.

I posted about inconsistent editing.

Double sharp responded, adding that I could edit the periodic table article as I saw fit, and that others could comment.

On that basis I went ahead and reverted Double sharp’s edits in which he removed content in the period table article. Some of this content was longstanding. Double sharp raised concerns about other content and took no action at the time.

Subsequently, DePiep reverted my revert without any previous discussion, and in breach of BRD. Sandbh (talk) 11:29, 26 December 2020 (UTC)
 * Please, can we keep the ANI discussion at ANI? Thanks.  EdChem (talk) 01:26, 27 December 2020 (UTC)

I've withdrawn my complaint. Sandbh (talk) 03:51, 27 December 2020 (UTC)
 * It's not withdrawn. -DePiep (talk) 12:15, 27 December 2020 (UTC)

IUPAC Group 3 project update
From Eric Scerri: "Before delivering a report I will be publishing an article in Chemistry International which discusses our preliminary conclusions and will be asking for comments and feedback. I'm still writing the article." "By all means share this update." --- Sandbh (talk) 09:07, 8 November 2020 (UTC)
 * Thank you, this is useful information to keep in the back of our heads. We can start to consider it for content discussions when his article appears – which I hope happens soon. Double sharp (talk) 19:11, 9 November 2020 (UTC)

If I may ask: was an approximate timeframe for the appearance of this article given? (Totally fine if it wasn't, just curious.) Double sharp (talk) 22:03, 30 November 2020 (UTC)


 * No. I’ll ask Eric how he’s going. I could understand his reluctance to be pinned down to a specific time, however. Sandbh (talk) 00:34, 1 December 2020 (UTC)
 * I'd definitely understand it too; I'm just curious. Double sharp (talk) 01:31, 1 December 2020 (UTC)

Group 3 discussion: formal conclusion (Dec 2020)
(Reacting to Double sharp's ping). We write for a generalist readership. Therefore, our primary goal is to describe the current consensus, as set out (for example) in modern textbooks. The La/Lu debate is only of interest to specialists, and is unresolved. Its proper place is in a section, or perhaps an article, and nowhere else; so as not to confuse readers who just about know what the PT is. Narky Blert (talk) 10:42, 20 December 2020 (UTC)
 * We have a section Periodic table and a main article Group 3 element, the latter of which discusses the situation in more detail; is that what you're thinking of?
 * Thanks for your opinion. IUPAC was planning to resolve it and I got some information from here (who's in touch with the IUPAC project chair) that something is going to be published about it by them soon. That's why I decided that the issue might as well be dropped until we hear from them, since whatever happens we have to look at it then anyway. But, more recent events made me worried that some others who had opinions here might not be happy with this "temporary drop". So my asking was intended to just gauge if that was so or not. Double sharp (talk) 10:50, 20 December 2020 (UTC)
 * I haven't read them - but yes, exactly that
 * I haven't followed this whole debate; but would suggest closure with some headnote or another until IUPAC decides something. There's no point in dragging this discussion out still further; and what IUPAC says may make parts of it moot. Narky Blert (talk) 11:04, 20 December 2020 (UTC)
 * I agree totally with you. I also like your idea of putting some headnote up to say "look, let's wait for IUPAC", because this issue seems to be amazingly effective at driving people who care about it to passionate argumentation (myself included, I've been trying to cut it out). Well, let's just see if everyone else is happy with such a closure, since I never actually asked when I requested a withdrawal there. ^_^ Double sharp (talk) 11:13, 20 December 2020 (UTC)

Yes, there should be some sort of note in the archive to indicate our consensus to let things stand at least until we can look at the IUPAC group's report and the reaction to it. But I am not sure whether it is kosher to add a headnote to an already archived thread. I think the best thing would be to start a new "==" thread, invite all participants in the old discussion to ascent to our implicit closure, and then archive it at the top of archive 52. YBG (talk) 17:06, 20 December 2020 (UTC)
 * Yes, as confirming the status quo. Cannot be a new discussion. -DePiep (talk) 19:49, 20 December 2020 (UTC)
 * I agree with Double sharp. I think we should just wait for the IUPAC. -322UbnBr2 (Talk &#124; Contributions &#124; Actions) 05:13, 21 December 2020 (UTC)

Proposal
As a more-or-less neutral observer, I propose that this and any similar live thread be closed with a headnote along the lines of:
 * This discussion is closed without WP:CONSENSUS, and should not be reopened until IUPAC makes a firm recommendation about the La/Lu question.

I don't care about the exact wording. Anything to the same effect would do. This debate has in the past spilled over into at least two dramaboards. Let's draw a line under it.

Support

 * Support, as proposer. Narky Blert (talk) 20:25, 20 December 2020 (UTC)
 * Support. Good enough for me. It doesn't make sense to fight now for something this controversial, when a "no consensus" result is the most likely outcome, when it seems likely that the IUPAC taskforce will decisively put the matter to rest rather shortly. Double sharp (talk) 03:26, 21 December 2020 (UTC)
 * Support. Sandbh (talk) 03:46, 21 December 2020 (UTC)
 * Support. YBG (talk) 10:11, 21 December 2020 (UTC)
 * Support. -DePiep (talk) 10:23, 21 December 2020 (UTC)

Comments
(signed by proposer) Narky Blert (talk) 20:25, 20 December 2020 (UTC)

IUPAC textbook survey
Released one year ago, the results are here.

The ratios for the three periodic table options are as follows:

Sc-Y-La Sc-Y-Lu  Sc-Y-*-**   Z ====================================================== 1970s to 2010s 4.33     1.0      1.1        193 2000s to 2010s 2.7      1.0      1.0         94 2010s          2.66     1.0      1.83        33

A single decade is probably too small a sample to base any conclusion on.

Anyway that is old news.

From a block perspective, it's interesting to note Lu appears in the f-block about 85% of the time (1970s to 2010s). That is because it shows in the f-block in the traditional Sc-Y-La-Ac form, and in the *-** form. Conversely, La is in the f-block about 33.3% of the time. That is because it shows in the f-block in the *-** form and in the Sc-Y-Lu-Lr form. --- Sandbh (talk) 05:37, 23 December 2020 (UTC)

FYI: Campero and Ponce (2020)

 * Disclaimer: No version of Group 3 is better than another; each version has it uses, depending on the properties of interest.

I flagged C&P's article in the interesting publications section.

There were two items in their article that I noted:


 * "We conclude then that the natural sequence of atomic number Z in the large periodic table, as Scerri has explained,(52c) implies that La is in group G3."


 * "Also, the electronegativity values determined in this work are correlated to similar chemical behavior of the ions through [1] Knight’s, [2] isodiagonal, [3] vertical, and [4] horizontal periodicity."

In my recent article, "The location and composition of Group 3 of the periodic table" I argued for the placement of La in group 3, including on the grounds of [2] isodiagonal and [4] horizontal relationships.

I've posted elsewhere about [3] the vertical trends going down (a) B-Al-Ga-In-Tl, (b) B-Al-Sc-Y-La, and (c) B-Al-Sc-Y-Lu. Based on 23 physical and chemical properties, it turns out the trendlines are the smoothest for (b) B-Al-Sc-Y-La. The next best option is either (a) or (c).

On [1], there is a knight's move relationship between Ca and La: --- Sandbh (talk) 00:26, 24 December 2020 (UTC)
 * The ionic radius of Ca2+ is 114 pm; that of La3+ is 117 pm (cf. Lu3+ 100).
 * The similarity in sizes means La3+ will compete with Ca2+ in the human body, and usually win on account of having a higher valence for roughly the same hydrated radius.
 * The basicity of La2O3 is almost on par with CaO2 whereas Lu2O3 is the least basic of the Ln oxides.
 * Freshly prepared La2O3 added to water reacts with such vigour that it can be quenched like burnt lime (CaO) whereas Lu2O3 is insoluble in water.
 * The electronegativity of Ca is 1.0; that of La is 1.1 (cf. Lu 1.27).

Advice from Eric Scerri
As flagged, I said I would ask him about his report setting out progress wrt to the Group 3 issue. It can be expected to appear in the Jan-Feb, or thereabouts, issue of Chemistry International:


 * "There hasn’t been a huge response, let’s just put it that way. I’m hoping my new article in Chem Int.…might change things."

I hope I'm still safe to have posted this here; haven't looked at ANI yet. --- Sandbh (talk) 07:27, 3 January 2021 (UTC)

Element articles, untangling
I came upon Wikiproject elements after doing some cleanup on the lithium page, and being struck by the combination of two topics in one. It seems this is quite general to many articles, at least according to a quick scan of the first three rows of the PT.

I recommend that each article be divided into two.

1. New articles entitled, say, Elemental XXX. This article will be in exactly the same form as that recommended for all other chemical compounds in the Manual of Style/Chemistry. After all, elemental forms are only unique chemical compounds in that they only contain one element. No need to treat them differently IMHO.

2. The existing article will have information specific to the chemistry of the elemental form migrated out except where it illustrates the general chemistry of the element.

Looking forward to your reaction to the concept. Please don't bother wordsmithing the two sections below, they serve only to illustrate the principle.

Sections to be migrated out of the article "Lithium"
Infobox lithium:

- Move from Infobox 'Physical Properties' and 'Other Properties'

- Copy from 'History' In 1855, larger quantities of lithium were produced through the electrolysis of lithium chloride by Robert Bunsen and Augustus Matthiessen.[58][72] The discovery of this procedure led to commercial production of lithium in 1923 by the German company Metallgesellschaft AG, which performed an electrolysis of a liquid mixture of lithium chloride and potassium chloride.[58][73][74] to History

- Copy from Chemistry ''Lithium reacts with water easily, but with noticeably less vigor than other alkali metals. The reaction forms hydrogen gas and lithium hydroxide in aqueous solution.[4] Because of its reactivity with water, lithium is usually stored in a hydrocarbon sealant, often petroleum jelly. Though the heavier alkali metals can be stored in denser substances such as mineral oil, lithium is not dense enough to fully submerge itself in these liquids.[17] In moist air, lithium rapidly tarnishes to form a black coating of lithium hydroxide (LiOH and LiOH·H2O), lithium nitride (Li3N) and lithium carbonate (Li2CO3, the result of a secondary reaction between LiOH and CO2).[40]''

''When the metal burns strongly, the flame becomes a brilliant silver. Lithium will ignite and burn in oxygen when exposed to water or water vapors.[83] Lithium is flammable, and it is potentially explosive when exposed to air and especially to water, though less so than the other alkali metals. The lithium-water reaction at normal temperatures is brisk but nonviolent because the hydrogen produced does not ignite on its own. As with all alkali metals, lithium fires are difficult to extinguish, requiring dry powder fire extinguishers (Class D type). Lithium is one of the few metals that react with nitrogen under normal conditions.[84][85]''

Copy from Military applications Metallic lithium and its complex hydrides, such as Li[AlH4], are used as high-energy additives to rocket propellants.[17]

Move section 'Precautions'

Example New article ("Elemental Lithium")
ElementalLithium is a metal with the formula Li. It is a soft, silvery-white alkali metal. Under standard conditions, it is the lightest metal and the lightest solid element. Like all alkali metals, lithium is highly reactive and flammable, and must be stored in mineral oil. When cut, it exhibits a metallic luster, but moist air corrodes it quickly to a dull silvery gray, then black tarnish.

Production
Lithium metal is produced by electrolytic reduction of lithium chloride

Applications
Elemental lithium has no known industrial uses but it is used a precursor in academic synthetic chemistry. It can appear as an unwanted layer in lithium batteries, and there is much research to develop solid state batteries which could use lithium as cathode.

Article distinction

 * What, you think, is the essential separation ground (say, the test question: why does subtopic X belong in article Li/1 or Li/2?). Following, could the titles be like, disambiguated: [Lithium (elemental)], [Lithium (chemistry)], [Lithium in real life], [Lithium (element concept)], [Lithium]. -DePiep (talk) 00:58, 17 November 2020 (UTC)
 * pending any other responses as requested by Double sharp, How about: information specific to the elemental form(s) of the element, in particular history, properties, uses (primarily industrial, secondarily academic), and above all hazard data. Otherwise, into the main article as it exists today. Not sure I follow your second question.  There would be a lithium article, therein a section 'Compounds and Chemistry', therein a succinct description of the chemistry of its compounds including theelemental form, and a link to main article 'Elemental Lithium'? --Oldboltonian (talk) 14:06, 17 November 2020 (UTC)
 * With the reader in mind, I ask the question: what is the separation ground? If you have property X of Li, which check do you perform to decide: should go in article Li-such or in Li-so? Then my 2nd question is: This same separation ground should be understandable from the title (to me, "elemental" is not). In short: If you have two lithium articles, what's their difference? -DePiep (talk) 17:24, 17 November 2020 (UTC)
 * Good question. For elements that are metals or gases one could add the word 'metal' (I note that you already gave your opinion on the use of the word 'metal' to indicate 'elemental' or 'gas'.  For the common allotropes of C and O, these articles aleady exist.  But I take your points, DePiep and double sharp, let's let nature take its course here, and where there is enough meat (like the C&O allotropes) there will be a separate article). Still I'd like to see the content around the chemistry and properties of the elemental state grouped together within the articles--Oldboltonian (talk) 11:59, 21 November 2020 (UTC)

Uniform structure of articles
In spite of there being a recommended structure for articles on the project page, adherence to the proposed structure is quite patchy (see table below, analysis just the first 3 rows of the PT. Any energy for editing towards some uniformity?

--Oldboltonian (talk) 12:18, 21 November 2020 (UTC)
 * The guidelines are mostly historical and nobody really seems to be taking them very seriously. They were probably good for a phase-I standardisation, but at some point we moved into a more of a phase-II article improvement of getting as many to GA and FA as we could, and then the structures naturally diverge per element. So I'd say there's not so much reason to change. Double sharp (talk) 10:58, 23 November 2020 (UTC)
 * If they are merely historical artifacts, perhaps we should note that? YBG (talk) 03:22, 24 November 2020 (UTC)
 * Sounds cool to me, . I've marked it as historical and taken it off the tabs list. That being said, I do think an updated version of these guidelines just for a "standard WP layout decision + colouring style" might be a good idea if/when one is decided on via RFC. (Former is frozen till we hear from IUPAC, latter is to come.) Or maybe that is better for the MOS like WP:ALUM. Anyway, we have time before we'll have such a thing. Double sharp (talk) 21:26, 24 November 2020 (UTC)

WP:NOTAFORUM
Are there any thoughts about this WP policy, given has mentioned it a few times?

Everything I post here I do so with the aim of sharing information and enriching perspectives, in support of improving articles. Good writers are well-informed writers, as I see it. --- Sandbh (talk) 01:13, 8 December 2020 (UTC)
 * Please be specific. -DePiep (talk) 01:45, 8 December 2020 (UTC)

The context is these posts by EdChem:


 * "As a Wikipedian, I think the discussion is wandering into WP:NOTAFORUM territory."
 * "Double sharp, if you are going to leave aside any action on group 3 until there is movement from IUPAC, then why put a lot of time into discussion of the topic, especially in light of WP:NOTAFORUM?"
 * "Double sharp, I mentioned WP:NOTAFORUM to both you and Sandbh as there are some pretty long posts appearing and they are not clearly directed to article-space issues. It was meant as a reminder and an invitation for reflection, so please don't take it as more than was intended."

--- Sandbh (talk) 02:20, 8 December 2020 (UTC)


 * I find that the number and volume of posts makes it difficult for me to concentrate on the forest rather than the trees, and so it distracts me from the main purpose: building article space. I also note that when I see substantial references to primary sources and non-mainstream secondary and tertiary sources, I lose confidence in my ability to distinguish between due and undue WP:WEIGHT. So while I don't think we have violated the letter of WP:NOTAFORUM, I think the project would better served by having fewer and smaller posts. YBG (talk) 16:50, 8 December 2020 (UTC)
 * Well, how about this as a halfway house: we put here whatever is directly describing and discussing what we're planning in the articlespace, and write the sort of thing Sandbh describes to each others' user talk pages? There aren't that many of us, so it shouldn't be hard to find (since mostly the long conversations happen between me and Sandbh), and this talk page becomes easier to follow. Double sharp (talk) 17:08, 8 December 2020 (UTC)
 * This seems to be a very workable solution. And when a topic starts to gel into something that might directly affect article space in the immediate future, a new thread can be started on this page, with a summary of the user talk thread and a link to the complete discussion. Yes, this might work out very well indeed. YBG (talk) 17:42, 8 December 2020 (UTC)
 * in you post you mention two issues: use of primary/secondary sources and long posts. While reducing post and thread lengths by moving preliminary work to userspace by itself is an improvement indeed, that does not solve the sources-item you mentioned. And if I am correct, that is what EdChem is pointing to by mentioning NOTAFORUM. This issue then will pop up again when a usertalk thread arrives on this page, and so introduces the issue for article space. (IOW, in userspace talk is more free, but when a conclusion/proposal is posted here to go live somehow, the sources-problem you mention is activated). ping -DePiep (talk) 21:10, 8 December 2020 (UTC)

My experience of our talk page comes in four flavours:
 * 1) Someone posts something here that I don't have much interest in so I don't follow it, so much.
 * 2) Someone posts something here that I find interesting, so I join in the discussion.
 * 3) I post something I find interesting to see if anybody else is interested. Nobody else is interested. End of thread. Collapse into a hat note after a short while.
 * 4) I post something I find interesting to see if anybody else is interested. Somebody else is interested. I join in the discussion.

Any of these may impact the article space. Personally, this has worked fine for me. I pay as much attention as I feel I need to.

Primary, secondary, tertiary sources Here's what our policy says:


 * "Articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources. Secondary or tertiary sources are needed to establish the topic's notability and to avoid novel interpretations of primary sources. All analyses and interpretive or synthetic claims about primary sources must be referenced to a secondary or tertiary source, and must not be an original analysis of the primary-source material by Wikipedia editors."

For primary sources, these do sometimes appear in the article space, as e.g. "In 2020 it was reported that".

For secondary sources, there seems to be two kinds: mainstream, and higher-level. Mainstream = e.g. textbooks e.g. those that show polonium as a metalloid. Polonium is normally excluded from metalloids by higher-level texts, and for an element that is hard to study those will be more reliable. This example comes from Double sharp.

Tertiary sources (of which WP is one) seem unremarkable to me, provided you keep your wits about you. They can provide a global summary of what secondary sources say.

The beauty of posting things here is the convenience of exposing it to all of us, and the ease of searching. Since we've started moving some of our discussions to each other's talk pages it's notably harder to find things i.e. I may now need to search my talk page, YBG's talk page, Double sharp's talk page, and this talk page, before I can find whatever it was I was looking for.

The things we've talked about here recently are: our periodic table article; categories; noble metals; group 3; colouring. There are plenty of reliable secondary sources addressing these. All of our discussions are related to associated content appearing in the article space. Where we have got to now seems to provide fair coverage. --- Sandbh (talk) 23:38, 8 December 2020 (UTC)

An informal poll
It seems to me that the general WP philosophy about what types of sources we should pay the greatest attention to is:
 * WP:TERTIARY > WP:SECONDARY > WP:PRIMARY > WP:OR
 * Q1 This general assessment is admittedly overly simplified. But is it completely wrong-headed? Why or why not?

However, my general impression of the volume of material discussed on this page is:
 * WP:OR > WP:PRIMARY > WP:SECONDARY > WP:TERTIARY
 * Q2 This general impression is admittedly overly subjective. But is it completely wrong-headed? Why or why not?

I would appreciate my fellow project editors answering these questions and one of the following two:
 * Q3 If you agree with me on these points, what is your overall reaction to this unfortunate contrast?
 * Q4 If you disagree on one or both points, do you have any idea why I might have gotten such wrong-headed ideas?

YBG (talk) 09:14, 14 December 2020 (UTC)


 * from Double sharp
 * My answers.
 * Q1. I think you are correct, though I'd rather nuance it to TERTIARY > SECONDARY >> PRIMARY >>> OR. And I think that when a basic topic is being discussed like chemical periodicity, it becomes even more important to follow this order. In this case I would say that tertiary sources (i.e. standard textbooks) are the way to go, with secondary sources only called in when the tertiary sources don't agree or where secondary sources generally agree that the tertiary ones are wrong. Primary sources, being generally reports of new research, are hardly the way to go here, when we're supposed to provide a standard view on the topic: if something is really standard, it should not be hard to find tertiary sources giving that view. And there is always the danger of interpreting primary sources to say something that they do not really say or cherrypicking primary sources, when actually an analysis of tertiary and secondary sources would better indicate to us how common a view expressed in primary sources really is and whether and how we should describe it per WP:DUE. Finally, OR ought to be kept to a minimum, because its only conceivable purposes are (1) fun diversions or (2) announcements that something is soon to appear in a primary source.
 * Q2. I also think that you are correct.
 * Q3. I think it is not a good situation. I have contributed to it in the past, and I regret having done so. I would also like it to stop, and have since tried to do so. Double sharp (talk) 10:26, 14 December 2020 (UTC)
 * Yes, I agree that my first ordering should be more nuanced, and I generally agree with your reordering. YBG (talk) 05:45, 15 December 2020 (UTC)


 * from DePiep
 * We should also consider related policies like WP:UNDUE and WP:FRINGE. -DePiep (talk) 22:08, 14 December 2020 (UTC)


 * from Sandbh
 * Some more answers.
 * Q1. The general WP philosophy is "Articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources."
 * Q2. An example would be helpful.
 * --- Sandbh (talk) 00:34, 15 December 2020 (UTC)
 * re Q1: Yes, it appears my general impression needs to be a bit more nuanced. See 's better arrangement listed above.
 * re Q2: What brought this subject to the forefront was the extended discussions of correlation coefficients elsewhere on this page, which appeared to me to have huge swaths of OR. But this only crystalized my general impression of our discussions here over the years. acknowledges above that he has contributed to this trend in the past and I is now seeking to take a different tack.
 * --- YBG (talk) 05:45, 15 December 2020 (UTC)
 * --- YBG (talk) 05:45, 15 December 2020 (UTC)

, 1. I hope my What's important response to ComplexRational will overtake all of this. 2. A little more to follow. Sandbh (talk) 06:50, 15 December 2020 (UTC)

, A little more: A simple XY plot of Z v. the value of property for, say, five elements, showing the R2 value is not what I would call OR. Thousands of authors comment about trends going down groups. G&E discuss this and draw the XY plots, including for groups 3 and 13 in the same chart, and comment on the difference in the curves. In any event WP:OR doesn't apply in the talk space. Sandbh (talk) 07:01, 15 December 2020 (UTC)
 * You say is not OR. Perhaps not. But there were many, many tables of such correlations with very little reference to sources. I could not tell whether this is the same sort of analysis present in the sources, and if so, whether they be 1RY/2RY/3RY. I could not tell if this was duplicating the analysis or if it was adding something new. I could not tell whether the sources showed R2, and if so, whether the source was using a linear, quadratic, or another sort of fit. In short, without any of these sorts of details mentioning sources and describing similarities and differences, I am left assuming that the thread was OR without any RS context.
 * Yes, there is no absolute prohibition of OR in talk space. But IMO it should be used in moderation: when a talk thread has a volume of OR or 1RY that vastly exceeds mentions of 2RY/3RY sources, IMO we have lost sight of something very important.
 * In the past, our project used 1RY/OR quite broadly in our talk page discussions. I clearly recall thinking that we were absolutely justified in using OR in our talk that resulted in mono-/di-/poly-atomic nonmetals. But my opinion has now changed, and I now think I would cringe at the the sheer volume of OR that we contributed in deciding highly significant and highly visible mainspace and template space changes.
 * It seems to me that we are at a different place now. IMO, we now have a big difference of opinion about the appropriate usage of 1RY/OR. IMO this disagreement is an important one and we as a project need to come to grips with it.
 * Some time back, Sandbh attempted to start a discussion aimed at coming up with project norms. I opposed it because I saw no compelling reason to restate or supplement the general WP norms. But as far as 1RY/2RY/3RY/OR, IMO we need better agreement than we now have. Maybe we need to develop a common understaning of what the global WP policies mean in our project's context. Maybe we need to supplement those policies with our local agreements. I don't know. I don't know what we need. I don't know how to get it. But IMO we absolutely need something.
 * ---YBG (talk) 08:20, 15 December 2020 (UTC)
 * Totally agree with the above post by . Double sharp (talk) 09:15, 15 December 2020 (UTC)

, I don’t understand what the problem is. —- Sandbh (talk) 08:32, 15 December 2020 (UTC)
 * I see several problems
 * Lost collegiality. In our recent trip to ANI-hell, the presenting symptom was rudeness (aspersions,non-AGF,ad hominen,...), but IMO a contributing (if not underlying) cause was that we hold different views about sourcing. If we continue to have wildly differing ideas about OR/1RY/2RY/3RY, we may still be able to avoid ANI-hell, but only because some (or all) of us expend huge amounts of energy to avoid such rudeness.
 * This argues for our need to reach consensus the appropriate role of OR/1RY/2RY/3RY in talk space. I would hope that this need would be apparent both to (1) those who prefer few limits on OR/1RY in talk space and to (2) those who prefer sparing use of OR/1RY in talk space.
 * In addition to the above reason (which I hope all could agree with), I see specific problems with extensive use of OR/1RY in talk space, which I expect agreement only from those who like me prefer sparing use of OR/1RY in talk space:
 * Lost opportunity. Our 880K discussion consumed 16 months of this project's life and untold hours of our personal lives to come up with poly-di-/mon-atomic nonmetals. A large portion of that discussion was OR/1RY. Some time later, we settled on reactive nonmetals / noble gases. We have since reverted to something very nearly like what we had before poly-/di-/mon-atomic. I believe that one of the big reasons was that we based our discussion on OR/1RY not 2RY/3RY as we should have. Think of how much bluer our WP:ELEM/PTQ could be if we had not expended that effort.
 * Lost focus. Because WP article content needs to reflect the state of 2RY/3RY sources, we need to keep those sources in focus. OR/1RY in moderation can help us better understand the context of 2RY/3RY sources. But if we allow unbridled use of OR/1RY in talk space, we run the risk of unwarrented (conscious or unconscious) use of those sources in deciding main space content issues, so that WP does not accurately reflect the state of 2RY/3RY sources.
 * I hope that explains the problem as I see it. Much as I value consensus, of even greater importance is that we understand each other. YBG (talk) 09:32, 15 December 2020 (UTC)
 * Double sharp (talk) 10:01, 15 December 2020 (UTC)

I see no wildly differing views on sourcing. WP sourcing on policy is clear enough. I agree with DS about the reliability of textbooks v. higher level sources. The age of a source does not automatically disqualify its relevance.

On poly-di-/mon-atomic nonmetals, I recall we were proud of that achievement. What has happened to that pride? There was no OR. The terms involved were already in the literature. The hours involved were voluntary. What OR/1RY did we use? WP:ELEM/PTQ is not our sole focus. Our project is small and this retards blue progress. Having worked on At I know how hard it is, and the standards expected these days for FA I find to be off-putting. The sense of achievement is not worth the effort involved unless I am hugely motivated. That is why e.g. I haven't done more work on metal.

The current nonmetal taxonomy was reached in a relatively short time. I benefited from that experience in terms of increasing my knowledge of the associated contextual factors.

On the risk of unwarranted bias, risk is made up of likelihood of an occurence and its consequences. The level of the consequence informs the risk mitigation strategy. What's the level of risk? I'd say possible. What's the consequence? An article shows some bias, as semi-regularly occurs judging by what I see at ANI. What's the consequence? The article concerned detracts from the calibre of Wikipedia. What risk management strategy is in place? Someone will notice the bias and correct it or raise it here. That's the continuous improvement nature of WP.

As to differing views and expectations for how we achieve project goals, that is the nature of WP projects. Sandbh (talk) 23:56, 15 December 2020 (UTC)


 * You seem to prefer an after-the-fact reactive approach to quality control; I would prefer a proactive before-the-fact approach. I wonder if we can find some clarity on the issue of whether there are wildly differing views. YBG (talk) 01:35, 16 December 2020 (UTC)

, both approaches are valid, are they not? The first is supported by WP:BOLD. Our discussions on nonmetal taxonomy are an example of the second. Sandbh (talk) 04:47, 16 December 2020 (UTC)


 * Since it has been referred to: for the record, my view is that we should only be able to supersede textbooks with higher-level sources if the higher-level sources as an overwhelming majority agree that the textbooks are wrong. And then the higher-level sources must have been agreeing on that for quite some time, extending up to the present, and show a wide spread of authors adhering to that view. And it should at least be possible to find some textbooks or other standard works that take the correction seriously. Just one article by itself, or a bunch that seem to live in a vacuum with no one noticing them, is not enough. Of course, this is only for standard topics covered in textbooks.
 * Basically, it's meant for astatine, not for odd PT's someone suggested and nobody else ever adopted. Double sharp (talk) 02:19, 16 December 2020 (UTC)

Which odd PT's are you referring to? Al over Sc? Sandbh (talk) 04:26, 16 December 2020 (UTC)

Informal poll 2
To help determine if we have wildly differing views on the use of source material on talk pages within our project, I would like to pose ask the following question of project members.
 * Your input would be greatly appreciated in order to gauge the range of opinion in project participants. Your input would be especially appreciated if it is outside the range of those who have already !!voted. (This ping went to all editors found in the last 500 edits on this page who have not yet !!voted. The original ping attempted to go to every editor mentioned in a ping or other template.) Thank you! YBG (talk) 21:14, 17 December 2020 (UTC)
 * Apologies for misspelling YBG (talk) 21:25, 17 December 2020 (UTC)

I am not trying to gain consensus here, merely trying to gauge the range of opinions present at WP:ELEM. YBG (talk) 01:35, 16 December 2020 (UTC)
 * Which statement best represents your opinion about the appropriate use of primary source material (1RY) and original research (OR) on talk pages?
 * 1) Editors should use 1RY/OR on talk pages as they see fit. Efforts to limit such are forbidden by WP policy.
 * 2) Editors should use 1RY/OR on talk pages as they see fit. Efforts to limit such are discouraged by WP policy.
 * 3) Editors should use 1RY/OR on talk pages as they see fit. Efforts to limit such are harmful to our project.
 * 4) I would prefer that 1RY/OR on talk pages have no limits, but our project can still thrive with such limits.
 * 5) As to the use of 1RY/OR on talk pages, I have no opinion.
 * 6) I would prefer that 1RY/OR on talk pages be used sparingly, but our project can still thrive if it is used extensively.
 * 7) Editors should use 1RY/OR on talk pages only sparingly. Extensive use of such is harmful to our project.
 * 8) Editors should use 1RY/OR on talk pages only sparingly. Extensive use of such is discouraged by WP policy.
 * 9) Editors should use 1RY/OR on talk pages only sparingly. Extensive use of such is forbidden by WP policy.

!!Votes
Please answer below by picking the statement that best represents your viewpoint. Please limit your entry to a simple !!vote indicating the statement and (if necessary) a more nuanced statement that better matches your point of view. YBG (talk) 01:35, 16 December 2020 (UTC)
 * 7. YBG (talk) 01:35, 16 December 2020 (UTC)
 * 7. Double sharp (talk) 01:56, 16 December 2020 (UTC)
 * 7. Droog Andrey (talk) 22:59, 16 December 2020 (UTC)
 * 7. DePiep (talk) 13:37, 17 December 2020 (UTC)
 * 7. -322UbnBr2 (Talk &#124; Contributions &#124; Actions) 21:26, 17 December 2020 (UTC)
 * 7. ComplexRational (talk) 15:46, 18 December 2020 (UTC)
 * 7. Just like everyone else. –LaundryPizza03 ( d c̄ ) 01:13, 19 December 2020 (UTC)
 * Between 7 and 8: As I explain below, I am uncomfortable with the definitive nature of (7) and its placement of OR and 1RY source discussions in the same boat, but there has been way too much of both and that has caused harm. It is also discouraged by policy.  EdChem (talk) 02:32, 19 December 2020 (UTC)

Discussion, reasoning, and proposals
To keep the above section limited to !!votes with optional nuanced restatements, please include discussion here, including your reasoning and justification and any suggestions for what we should do or not do. YBG (talk) 01:35, 16 December 2020 (UTC)
 * Editors may refer to 1RY, noting WP policy provides that, "Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources." Editors may refer to OR, noting WP:OR does not apply to talk pages, and that, "articles must not contain original research…OR is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published sources exist." Sandbh (talk) 04:42, 16 December 2020 (UTC)
 * I think there is no dispute that (a) pan-WP policies allow 1RY in both main space and talk space, and (b) pan-WP policies forbid OR in main space but not in talk space. The question in this poll is not about overturning these pan-WP policies, but whether, in our project talk space, the best volume of usage of 1RY/OR would be, on the one hand, only sparingly, or, on the other, as extensively as an editor sees fit. Your input on this question would be appreciated, especially if it differs from the range of opinions already shared. Thank you. YBG (talk) 02:23, 19 December 2020 (UTC)
 * My only focus is on improving the article space. Sometimes I'll do that on a bold basis, at other times I'll raise it here first. If I raise OR here I do so in the context of seeking to improve the article space. My interest here is in the merits of ideas or items posted here. That's the first thing I look for. If it makes it into the article space well and good. If it doesn't, for whatever reason, then I hope to have learned something, before moving on to something else. The volume of usage here is contingent on its users. Sandbh (talk) 06:09, 19 December 2020 (UTC)
 * I understand that your focus is on improving article space. What I am asking for is your opinion with respect to the extent to which 1RY/OR on talk pages should be used to achieve that goal.
 * Would the use of 1RY/OR as editors see fit improve our project? Does it harm our project? Or perhaps merely hinder it without necessarily preventing it from thriving?
 * Would the limiting 1RY/OR to sparing use improve our project? Does it harm our project? Or perhaps merely hinder it without necessarily preventing it from thriving?
 * I hope that these yes/no questions might help you (or anyone else) answer the poll question. YBG (talk) 07:09, 19 December 2020 (UTC)
 * PS, it would also be helpful to clarify whether different editors see things differently (a) because of different responses to this poll question, or (b) because of different ideas about what constitutes OR. If we have disagreements, it can only help if we understand exactly what we disagree about. YBG (talk) 07:27, 19 December 2020 (UTC)


 * My answers to your two questions are:
 * The use of 1RY/OR as editors see fit may or may not improve our project.
 * Limiting 1RY/OR to sparing use may or may not improve our project.
 * --- Sandbh (talk) 22:24, 19 December 2020 (UTC)
 * Thank you. A couple of follow-on questions:
 * If talk page use of 1RY/OR as editors see fit may or may not improve our project, what do you think would be the difference between use "as editors see fit" that would improve our project and such use that would not improve our project?
 * If lalk page limiting 1RY/OR to sparing use may or may not improve our project, what do you think would be the difference between such limits "to sparing use" that would improve our project and such limits that would not improve our project?
 * --- YBG (talk) 10:34, 21 December 2020 (UTC)

, I feel you should add something about DUE. Since it's perfectly possible to find reliable published sources for things which are basically WP:FRINGE views in modern chemistry. For example the JChemEd article about putting hydrogen over carbon. Double sharp (talk) 05:20, 16 December 2020 (UTC)
 * Yes, I agree. said much the same in the earlier poll. DUE, UNDUE, & FRINGE all relate to how 1RY sources should be used. But what I'm hoping to figure out first relates to the volume of 1RY/OR we use at WT:ELEM. Specifically, I'm trying to gauge whether we have wildly different ideas about it. YBG (talk) 00:03, 17 December 2020 (UTC)
 * Clear to me, good description. Especially fine as you mention the wider issue/consequence (volume). -DePiep (talk) 00:19, 20 December 2020 (UTC)

As I see it, this should be a project talk page, not a forum. I'm all for presenting different sources offering different points of view and having discussions, even if very long, to decide how best to represent this information in articles. However, I don't think we should be making our own arguments, especially when they are OR or otherwise considerably deviate from their cited sources. Such discussions can go off on a tangent, and in the end, won't help the project because OR and our personal viewpoints are not permitted in articles, so let's focus on improving content. ComplexRational (talk) 15:51, 18 December 2020 (UTC)
 * Double sharp (talk) 15:57, 18 December 2020 (UTC)

I agree with the sentiment here but perhaps see this as more nuanced. For example, I see the issues around 1RY sources and OR as overlapping but distinct. Since OR cannot end up in article space, its usefulness in article talk space is limited to situations relevant to content for article space. Its purpose should have a clear connection to article space... but that doesn't mean there aren't legitimate places for OR-related discussion. A simple example might be two facts about a compound published in different journal articles and whether bringing them together is a problematic SYNTHesis resulting in impermissible OR, or how to present them without ending up with policy difficulties. Though SYNTHesis is prohibited and for very good reasons, there are also times in science where a technical SYNTH violation might also result in an utterly uncontroversial statement. Let's suppose that an article included the statement that (4R, 5S)-7-chloro-4,5-diethyl-2,2-dimethylheptan-1-ol reacts with hydrobromic acid to form (4R, 5S)-1-bromo-7-chloro-4,5-diethyl-2,2-dimethylheptane, and that it was challenged. I could find hundreds of textbooks that assert that alcohols react with HBr to form bromoalkanes, that the reaction is tolerant of alkyl branches and other halogen substituents, and that stereochemistry in unrelated parts of the molecule is retained. However, I likely couldn't find anyone reporting this specific reaction unless it was part of a larger synthesis. Put aside the question of why this was in article space and just assume it was justified to be there, what happens next? Policy supports removal of challenged materials if a source can't be provided, a specific source is not available and yet the statement should not be contentious and a mundane SYNTHesis of a multitude of sources would offer a wealth of evidence (though in technical violation of OR). I think a talk page discussion would be appropriate and it would likely include OR etc, and it would not be harmful... but it would also be an exception as much (most?) OR discussion is problematic / harmful and it is certainly discouraged by policy.

Now, I know that much of the recent issues are not like the above example, and they have certainly been harmful given the ANIs and near-ArbCom case, etc. So, I do agree with the idea of 7 but see it as too definitive. Extensive use of OR is likely to be harmful, but I am uncomfortable with the possibility of exceptions being excluded. My misgivings become stronger when it comes to 1RY literature. I am aware that there are plenty of editors who view this more restrictively than I do, and I respect those differences of opinion. There are places where I would include primary literature as a matter of routine. If I was writing the ferrocene, I would mention the discoverers and those who contributed to understanding its structure. There are plenty of textbooks and review articles covering this history and I could easily source it without any 1RY sources, and I would use them... but I'd also include the key primary papers as references. I do this because, as an educator, I tell students that they can look to WP for information but not reference it and that its references can be a great start for writing about a topic. I also do it because, as a reader, those are some sources that I might want to find. And, I do it as a Scientist Wikipedian because when I write that "X did Y", X reporting that they did Y is the direct source for the statement. Now, for WP, we need the significance established by the 2RY and 3RY sources and having the 1RY source is not a reason for not providing the needed support – in other words, I see 1RY sources as often providing a desirable bonus / supplement to the necessary 2RY and 3RY literature. I can't agree that extensive discussion of 1RY material is necessarily harmful.

This again, however, is not the sort of situation that prompted the above poll. In the WT:ELEM context, I very much agree that there has been too much OR and 1RY discussion (for which I am also partly responsible, having fallen into the trap of discussing some OR). It has been harmful. I am !voting differently from other !voters to date, but hope that I am seen as very much in the same ballpark. EdChem (talk) 02:32, 19 December 2020 (UTC)


 * FWIW, I agree and do think some nuance would make it even better, and considered answering with "something between 7 and 8" too. But in the end, I felt that was doing an excellent thing in trying to get at the underlying issues in a simple question, and I did not want to go straight for too many details for fear of being too long-winded and causing his efforts to all end up for naught. ;) Double sharp (talk) 07:35, 19 December 2020 (UTC)


 * Thank you for your nuanced answer. Shoe-horning something that is not only continuous but multi-dimensionally continuous into a discrete set of nine choices is difficult if not problematic. YBG (talk) 08:29, 19 December 2020 (UTC)


 * I get and appreciate the nuanced answer: differentiating between OR and 1RY, and noting that at WT:ELEM this is urgent. -DePiep (talk) 22:27, 19 December 2020 (UTC)

Wisdom: EdChem, there is much wisdom in what you wrote. As you say, the issues around 1RY sources and OR are overlapping but distinct.

Your take on the use 1RY sources in the article space is excellent.

I differ from you re:


 * "much (most?) OR discussion is problematic / harmful and it is certainly discouraged by policy."

There is no basis for such a sweeping generalisation. OR discussion may or may not be problematic / harmful. It depends on the context. There is no WP policy discouraging such discussion in the talk space. --- Sandbh (talk) 22:48, 19 December 2020 (UTC)
 * I disagree with Sandbh. For example, many discussions here at ELEM went astray partially for being about OR; waste of time and no improvement. Worse, I cannot read something else than repeating their editing matra: anything goes, no discussion needed. No reflection on the problematic discussion and editing history in say 2020 here at WT:ELEM, no response regarding the mayfold content issues and their causes/solutions. If I am the only one having this perception: others can correct me. -DePiep (talk) 23:04, 19 December 2020 (UTC)
 * , in you have not !!voted yet. Could you do so, or else describe why not? -DePiep (talk) 23:31, 19 December 2020 (UTC)

I haven't voted due to concerns with the wording of the question:
 * 1) The issues around 1RY and OR are distinct, albeit overlapping.
 * 2) OR policy does not apply in the talk space. References to what editors "should" or should not do with regard to discussing OR in the talk space are wishful thinking. OR may or may not be harmful to our project; it depends on the context.
 * 3) EdChem has provided an excellent summary as to the many ways in which 1RY can be used in the article space, as a matter of routine.

WP:COMMON captures a lot of how I feel about this:


 * "Wikipedia has many policies or what many consider "rules". Instead of following every rule, it is acceptable to use common sense as you go about editing. Being too wrapped up in rules can cause loss of perspective, so there are times when it is better to ignore a rule. Even if a contribution "violates" the precise wording of a rule, it might still be a good contribution. Similarly, just because something is not forbidden in a written document, or is even explicitly permitted, doesn't mean it's a good idea in the given situation. Our goal is to improve Wikipedia so that it better informs readers. Being able to articulate "common sense" reasons why a change helps the encyclopedia is good, and editors should not ignore those reasons because they don't reference a bunch of shortcut links to official policies. The principle of the rules—to make Wikipedia and its sister projects thrive—is more important than the letter. Editors must use their best judgment.


 * Why isn't "use common sense" an official policy? It doesn't need to be; as a fundamental principle, it is above any policy."

--- Sandbh (talk) 00:12, 20 December 2020 (UTC)
 * I find it very worrying and troublesome that now it has come this far . -DePiep (talk) 09:59, 22 December 2020 (UTC)

About thread lengths
A scan of history and archives of this page WT:ELEM shows that since 1 July, 2020 (arbitrary date), that's 23 weeks:
 * 2800 posts were made (65/week)
 * 1950k was added (85k/week).

That's too much to handle, no quality of discussion or results can be established. -DePiep (talk) 17:56, 10 December 2020 (UTC)


 * I have removed all Collapse top etc collapsings from this page, as they disrupted the talk understanding (for example, an F3-search for keywords did not highlight right).
 * Collapsing might be useful for minor subthreads, for obviously closed dead ends, but not to introduce a limbo status. Closed threads go to archives, but only so after obvious closure. Collapsing is no mean to reduce thread length. -DePiep (talk) 18:30, 10 December 2020 (UTC)
 * Please do not use collapse to fold texts. It breaks text searching, also in Archives, and is not a clean way to finish a talk. Instead, you could consider keeping thread flow clean, even afterwards (by agreed reduction somehow); arvhice obviously closed threads, etc. -DePiep (talk) 19:15, 10 December 2020 (UTC)

, that's OK by me. Is there a collapse function that allows text searching? Sandbh (talk) 23:41, 10 December 2020 (UTC)
 * Don't khow. Better be solved by other means (i.e., thread and discussion self-managing). -DePiep (talk) 23:48, 10 December 2020 (UTC)

Use of sources at WP:ELEM
I just finished reading WP:PSTS, which discusses WP:PRIMARY, WP:SECONDARY, and WP:TERTIARY sources. IMHO, I think understanding these policies and how they apply to our domain of knowledge would be of great benefit to our project. It would help us (a) create good content, (b) following WP policy, and (c) avoiding conflict. What our project needs is to interpret these rules and specify what are (1) primary WP:ELEM sources, secondary WP:ELEM sources, and tertiary WP:ELEM sources. I am not saying I have the answers here, just saying that we need to collaboratively reach a consensus on these three question. Thoughts? YBG (talk) 04:11, 25 November 2020 (UTC)




 * 1. Primary sources are those in the peer-reviewed academic literature reporting OR.


 * 2. Secondary source are those primarily relying on secondary sources. Very crudely there are three tiers:


 * C. less RS;
 * CB. lies to children (ltc) RS
 * B. RS; and
 * A. "top-tier" RS, with reputation e.g. Wiberg; G&E; C&W; Oxford University Press, Pergamon Press


 * Wikipedia articles usually rely on material from reliable secondary sources.


 * 3. Tertiary sources are dictionaries, encyclopaedias and compendium. Many introductory undergraduate-level textbooks are regarded as tertiary sources because they sum up multiple secondary sources. Reliable tertiary sources can be helpful in providing broad summaries of topics that involve many primary and secondary sources, and may be helpful in evaluating due weight, especially when primary or secondary sources contradict each other. Some tertiary sources are more reliable than others, and within any given tertiary source, some entries may be more reliable than others.


 * Principles
 * Deciding whether primary, secondary, or tertiary sources are appropriate in any given instance is a matter of good editorial judgment and common sense, and should be discussed on article talk pages.
 * There's no need to necessarily rely on ltc RS, since plenty of RS and top-tier RS are available. Indeed, a theme of distinguishing RS based on e.g. publisher reputation, author reputation, age of publication, etc pervades WP:RS.
 * Higher-level texts, for elements that are hard to study, will be more reliable.
 * The age of a source does not necessarily reduce its reliability.
 * Introductory undergraduate-level textbooks have a reputation for textbook errors,, ,
 * —- Sandbh (talk) 04:08, 29 November 2020 (UTC)


 * I find it very worrying and troublesome that now it has come this far . -DePiep (talk) 09:59, 22 December 2020 (UTC)

Am I going mad?
For group 13 as the traditional B-Al-Ga-In-Tl, the trend line for the sum of the 1st three ionisation energies is disturbed by the appearance of the 3d10 shell in Ga. Thus, the trend line has a goodness of fit value (gf) of "only" 0.54

For group 3 as B-Al-Sc-Y-La, gf is 0.95; as B-Al-Sc-Y-Lu it's 0.94

This prompted me to check the gf values for other group 3 properties as either B-Al-Sc-Y-La or B-Al-Sc-Y-Lu.

The following table includes all the values involved, as well as the values for group 3 as Sc-Y-La-Ac and Sc-Y-Lu-Lr:

B-La B-Lu  B-Tl  | Sc-Ac  Sc-Lr --- Density           0.98  0.94  0.98  | 0.99   0.99 1st IE            0.62  0.62  0.65  | 0.97   0.97 Ionic radius      0.98  0.61  0.84  | 0.99   0.61 3rd IE            0.96  0.95  0.42  | 0.98   0.83 Melting point     0.57  0.42  0.81  | 0.15   0.11 Electron affinity 0.62  0.10  0.08  | 0.85   0.99 Sum of 1st 3 IE   0.95  0.94  0.54  | 0.96   0.96 Electronegativity 0.96  0.94  0.47  | 0.58   0.99 --- Average           0.77  0.65  0.60  | 0.83   0.69


 * The best one is obviously B-Tl with the smallest average. Droog Andrey (talk) 16:22, 6 December 2020 (UTC)
 * Why? Sandbh (talk) 23:07, 6 December 2020 (UTC)


 * Because of secondary periodicity (odd vs. even periods). Droog Andrey (talk) 08:47, 9 December 2020 (UTC)

Secondary periodicity

 * Secondary periodicity doesn't show in group 4: C−Pb = 0.69; C−Hf = 0.71. Chemically speaking, groups 1–3 are better shown as Li-Fr; Be-Ra; B-Ac. That's not going to happen; I expect group B-Al will remain over Ga. And He will stay over Ne. So an electronically consistent (EC) table is He over Be, and Lu in group 3. A chemically consistent (CC) table is He over Ne, and B-Al over Sc. Where one goes from there depends on the context. The EC table can make a concession to chemistry and move He over Ne. The CC table can make a concession to electronics and move B-Al- over Ga, in which case one has the popular form. Sandbh (talk) 00:52, 10 December 2020 (UTC)


 * Secondary periodicity does present in Group 4: Ge is more electronegative than Si, atomic radius increases more from C to Si and from Ge to Sn than from Si to Ge and from Sn to Pb, and so on.


 * B-Al-Sc-Y-La-Ac is neither chemically, nor electronically consistent. It is just comfortable for those who don't understand anything but straight similarities and linear trends. Droog Andrey (talk) 14:49, 12 December 2020 (UTC)

Can someone please provide a definition of secondary periodicity? The term is not used in article space. YBG (talk) 08:23, 14 December 2020 (UTC)
 * Secondary periodicity is the typical difference in properties of elements in even vs odd-numbered periods. It holds in all blocks except s. Even-period elements generally have smaller atomic radius, are more electronegative, prefer lower oxidation states. Odd-period elements differ the opposite way. Not always perfect of course. Therefore, whenever you graph a periodic trend outside the s block, it will rarely show an exact straight line, but rather show a definite zigzag instead.
 * The term is not already in article space because apparently mostly Russian sources know this generalisation and English ones don't. Perhaps because it seems to be a Russian discovery (Biron for the p block, Chistyakov for the d block). Double sharp (talk) 09:06, 14 December 2020 (UTC)

Chemical *and* physical properties
The trends are based on second degree polynomial curves of best fit, using the property value in question cross referenced against Z.

I've added nine physical properties, and one more chemical property, for a total of 17 properties. The combined goodness of fit values for all 17 properties are: B-Al-Sc-Y-La = 0.71; B-Al-Y-Sc-Lu = 0.615; B-Al-Ga-In-Tl = 0.685

For the relevance of chemical and physical properties please see the quote by Jensen, here.

So there's a case for keeping B-Al- over Ga, if you want to emphasis blocks i.e. p elements in the p block.

The better fit for B-Al-Sc-Y-La v B-Al-Sc-Y-Lu prompted me to check the historical record for periodic tables showing group 3 as B-Al-Sc-Y-La. There are 17 examples listed here.

What was it that these authors, including luminaries such as Rydberg, Seaborg, and Pauling, knew about B-Al-Sc-Y-La-Ac?

Deming (1947, p. 617) categorised Al as a pre-transition metal, along with the alkali and alkaline earth metals. Pauling (1988, pp. 635, 697) refers to the congeners of B being Al-Sc-Y-La. He discusses Ga-In-Tl separately in his chapter on Cu, Zn, Ga and their congeners. He writes, "The metals Zn, Cd, and Hg (group IIb) are also much different from the alkaline earth metals (group II), as are Ga and its congeners (group IIIb) from the elements of group III. Cox (2004, p. 185) refers to the pre-transition metals as those in groups 1 and 2, and Al. Further suggesting B and Al are better placed over Sc, here is a 2019 periodic table (by me) recording oxidation number trends,.

In this vein, Rayner-Canham (2020, pp. 178–181) writes:


 * “It was Rang in 1893 who seems to have been the first, on the basis of chemical similarity, to place boron and aluminum in Group 3.


 * Such an assignment seems to have been forgotten until more recent times. Greenwood and Earnshaw have discussed the way in which aluminum can be considered as belonging to Group 3 as much as to Group 13 particularly in its physical properties. Habashi has suggested that there are so many similarities between aluminum and scandium that aluminum’s place in the Periodic Table should actually be shifted to Group 3.


 * In terms of the electron configuration of the tripositive ions, one would indeed expect that Al3+ (electron configuration, [Ne]) would resemble Sc3+ (electron configuration, [Ar]) more than Ga3+ (electron configuration, [Ar]3d10). Also of note, the standard reduction potential for aluminum fits better with those of the Group 3 elements than the Group 13 elements (Table 9.2) — as does its melting point.


 * In terms of their comparative solution behaviour, aluminum resembles both scandium(III) and gallium(III). For each ion, the free hydrated cation exists only in acidic solution. On addition of hydroxide ion to the respective cation, the hydroxides are produced as gelatinous precipitates. Each of the hydroxides redissolve in excess base to give an anionic hydroxo-complex, [M(OH)4]−…There does seem to be a triangular relationship between these three elements. However, aluminum does more closely resemble scandium rather than gallium in its chemistry. If hydrogen sulfide is bubbled through a solution of the respective cation, scandium ion gives a precipitate of scandium hydroxide, and aluminum ion gives a corresponding precipitate of aluminum hydroxide. By contrast, gallium ion gives a precipitate of gallium(III) sulfide. Also, scandium and aluminum both form carbides, while gallium does not.”

B-Al-Sc-Y-La-Ac = (sp)8 cores; Lu has another core type (spf)22; Ga, In & Tl have other core types (spd)18 and (fspd)32.”

Based on the preceding evidence, B-Al-Sc-Y-La-Ac is chemically and electronically consistent.

The popular La table represents a case where B-Al- were moved over Ga, on electronic grounds. That said, the n +10 relationship (Rayner-Canham 2020, pp. 173−194) for groups 3 and 13 is still seen. Sandbh (talk) 04:12, 13 December 2020 (UTC) Courtesy ping
 * , I am not sure why I am being pinged even as a courtesy here when the conversation is between you and . I watch this page, so there's no need to ping me unless I am being directly addressed But since I am pinged, I confess I do not see how B-Al-Sc-Y-La-Ac can be seen as electronically consistent considering their electron configurations. Boron and aluminium are s2p1, that doesn't match the s2d1 configuration of the other four. Neither do B nor Al have low-lying d orbitals in the way Lr with its famous anomalous configuration does, that would save the idea with chemically relevant excited states. Neither does B have an (sp)8 core, as the noble gas it is over is helium which lacks a filled p shell. As I see the logic presented, one could just as well argue that Ar with an (sp)8 cores does not belong with Ne with an s2 core, nor with Kr and Xe with (spd)18 cores, and hence split apart the noble gases. Considering that that is literally one of the most homogeneous groups in the periodic table, I find myself unconvinced. Double sharp (talk) 07:21, 13 December 2020 (UTC)


 * , that was not well written be me. Sorry. B-Al-Sc-Y-La-Ac are electronically consistent in that each has an underlying noble gas core. As discussed by G&E (p. 222), for Ga and In it is NG + d10. For Tl it is NG + 4f145d10. "This variation has a substantial influence on the trends in the chemical properties of the group." Sandbh (talk) 05:43, 15 December 2020 (UTC)
 * Yes, but, what other groups in the periodic table have a consistent noble gas core? Look at every p block group from C-Si-Ge-Sn-Pb onwards. At the third element, d10 is added to the core. At the fifth element, f14 is added on top of that. Look at every d block group from Ti-Zr-Hf onwards. At the third element, f14 is added to the core. Why then should the B-Al- and the Sc-Y- groups be exceptions? It simply breaks the global trend across the periodic table and breaks electronic consistency. Not my OR argument, Jensen used it for the d block. Double sharp (talk) 05:48, 15 December 2020 (UTC)

I'll let DA answer for himself. But since I agree with him, I'll give also my reason: because we expect a contraction to wreck a linear trend everywhere else in the table. Just look at the "kinked" p-block EN trends for example. Therefore a linear trend is "wrong" here and we should expect a kinked one: so a lower goodness-of-fit value means we are closer to what the trend should look like. We describe to students how the alkali metals form a nice sequential series. Likewise, we describe to students how the halogens form a nice sequential series. At this point, we have “proved” the validity of periodicity and move on to other topics. The students are happy. Better to avoid the complexities of the middle main groups which such questions as: why is nitrogen dimeric and unreactive, yet one common allotrope of phosphorus is tetrameric and very reactive. Yet, here is some of the richness. This is actually an example of the Uniqueness Principle of these mid-second period elements (8). ... And with the fourth period for the p-block elements, comes the Fourth Period Anomaly. Discussed in detail by Sanderson (9), it can be stated as, “the 4th Period Anomaly for the p-block elements is where the properties of the Group member of the 4th Period do not fit the trend for the other members of the Group.” As exemplars of the Fourth Period Anomaly, Sanderson cited the difficulty in preparing AsCl5 (yet PCl5 and SbCl5 are easily synthesized) and HBrO4 (yet HClO4 and H5IO6 are easily synthesized). Pekka Pyykkö has referred to the phenomenon as “secondary periodicity,” and he has explained it as a relativistic effect (10).

...

Additionally, smooth trends down groups are more of a convenient fiction and that irregularities can be expected acccording to the Uniqueness Principle and the Fourth Period Anomaly. The "kinked" trend of Al-Ga-In is just like that for Si-Ge-Sn, P-As-Sb, S-Se-Te, Cl-Br-I. All is clear. Linear trends, as measured by best-fit lines, are a fiction that frankly only exist in groups 1 and 2 due to weird preemptive filling of s orbitals as explained by Ostrovsky. I think the more global understanding of the RS (e.g. Jensen) is better. Double sharp (talk) 02:59, 7 December 2020 (UTC) They're properties v Z. Sandbh (talk) 00:51, 8 December 2020 (UTC)
 * Question: Are these correlation coefficients of properties vs. period number?, , ?  EdChem (talk) 12:15, 7 December 2020 (UTC)
 * , I think is best placed to answer that question. I assumed it was either properties vs. period number or properties vs. atomic number, myself, as it'd have made sense that way. Double sharp (talk) 12:18, 7 December 2020 (UTC)
 * If they are against period number (as in the diagram above) then either I am missing something important or the analysis is statistically flawed. EdChem (talk) 12:26, 7 December 2020 (UTC)


 * , I have some concerns that I would like to discuss.
 * Firstly, these "goodness of fit" values are correlation coefficients for you fitting a degree 2 polynomial to the data for property vs. Z, correct? If so, I wonder why is the use of a model for continuous data being used for categorical data?  In no way is Z meaningful for any non-integer value, yet you are modelling with continuous curves against Z?  Are the goodness of fit values statistically justified?  Would other statistical measures be more appropriate?
 * On wat basis is the choice of a degree 2 polynomial justified as the model chosen? Why is a linear model or an alternative degree polynomial rejected / discarded / whatever?  Note, obviously there must be an integer N such that a polynomial of degree N or higher will have an R2 = 1 as the polynomial will pass through every data point... only polynomials of degree less than N are relevant.
 * On what basis is the choice of model a polynomial rather than some other function (exponential, logarithmic, etc)?
 * At times, in looking at your comments, I have the impression that you are presenting your view / thoughts / opinions / conclusions that come from your own analysis and then looking for historical work / publications, etc, that support that position. Is this a reasonable impression to form and does it reflect your approach?  In considering material to include in article space, what are you doing to avoid confirmation bias in looking for supporting materials and to ensure that materials selected to support a position do not need to be synthesised together to provide that support?
 * I ask because there are some strategies that are totally appropriate for the primary research literature that don't suit an encyclopaedia. For example, the changes here to our periodic table article have Al shown in group 3 and the caption of the illustration presents an argument that supports La over Lu in group 3, but neither of these is, for me, really supported by the references shown.  My impression is that it more reflects what you believe / want to say than what the sources presented support... am I being fair?  The introductory comment on teaching reflects exactly what you have written in your referenced article, though it is unsupported by references here or there.  Now, it is reasonable to say that contrasting the alkali metals and halogens is an approach used in teaching.  However, the context in which it is made is to begin a section on analogous categories that suggests expanding that approach to contrast pairs like most non-metals with transition metals or Sn and Bi with metalloids... and I don't see the literature support to justify not only including what is stated, but also for WP's voice to be used to imply what is included here.  My own experience as a researcher, for whatever it might be worth, involved mostly organometallic compounds of one of the noble metals and I don't see an analogy between it and the noble gases is apt.  My experience is OR for deciding what to include, of course, but speaking as an educator I am not inclined to adopt the approach that you suggest.
 * I am pinging all regular and / or recent contributors to this page –, , , , , – in search of a breadth of perspectives and thoughts.  Please, can we focus comments on editing rather than on editors, and bear in mind that discussion of differences of opinion / perspective can be valuable if mutually respectful, open-minded, grounded in the extant knowledge of chemistry, and collegial / professional?  Thank you.  EdChem (talk) 06:18, 13 December 2020 (UTC)
 * , I think a polynomial is difficult to justify in the first place, because of the secondary periodicity of Evgeny Biron that mentions. I also do share Droog Andrey's impression that 's analysis here is not paying enough attention to patterns and general trends. But I also don't think that that's the main issue for WP. WP is intended to reflect what the literature is rather than what we feel it ought to be; if we all agree to work on reflecting the former rather than the latter, then there would be no issue. Of course, sometimes it can be difficult when the literature is not united; solving this problem of how to describe the situation for an encyclopaedia like WP is a delicate matter that requires some discussion. But it's not the same as determining which view one personally supports and considers scientifically correct and writing in favour of it in an academic article. What EdChem remarks in 4 and 5 is what makes me worried that the two have been conflated. Double sharp (talk) 06:57, 13 December 2020 (UTC)

Thanks for your interest.

1. Yes, the goodness of fit value is the coefficient of determination for a second order polynomial curve of Z v the value of the property in question. I don't understand your reference to continuous data. For example, if I have five data points for the values of 1st IE for B-Al-Sc-Y-La, I'm looking to see how smooth the trend line is running down the group. Greenwood & Earnshaw (p. 223) plot trend lines for the 1st, 2nd, 3rd, and sum of 1st to 3rd ionization energies for groups 3 and group 13 v Z. The smoothness of the curve for the sum of the 1st to 3rd ionization energies for B-Al-Sc-Y-La is remarkable compared to the choppy trend line for B-Al-Ga-In-Tl. I used the same approach for six charts in my article on the location and composition of group 3. As noted in that article, a comparison of ionic data by Atkins et al. concluded that Sc-Y-La is preferred over Sc-Y-Lu. Their comparison is expressed in the context that ionic radii generally increase down a group. See Fig. 4 in my article, including the 0.99 v 0.61 goodness of fit values.

2. Polynomial curves consistently gave the best goodness of fit values. There was no requirement to go to a higher degree.

3. As per above.

4. I doubt there is that much new about the periodic table that has not already been documented somewhere in RS. Yes, I can see you may have formed the impression that I form an opinion then look for evidence in the literature to support it. In this case, prompted by G&E's plot, I found some evidence supporting B-Al over Sc, and it turned out this has a long history in the literature. I had already read in Wikipedia that Al was more similar to Sc, than Ga, but had not thought more of it at that time.

When I edit, my practice is to provided citations for anything controversial. I don't consciously worry about confirmation bias or WP:SYNTH since so much of descriptive chemistry is just a collection of facts.

5. I posted those changes to our periodic table article since I saw how clean the article structure looked, subsequent to our trial of continuous cooperative improvement editing, and noticed the periodic trends and patterns section could be improved with some discussion about isodiagonality and oxidation number trends. In this context I don't understand 's comment about not paying enough attention to patterns and general trends. Those changes had nothing to do with my advocacy for La in group 3, in a chemistry based table. I also advocate that Lu in group 3, and He over Be, is an excellent electronic based table. More important than either of these is that the context for the choice of particular periodic table is explained. It so happens that the La table happens to be currently more popular in the corpus of chemistry.

Aluminium in group 3 is convenient way to show the diagonal relationship between Be and Al and between Al and Ti, as has been reported in the literature. The Inorganic chemist's periodic table, for example, shows Al over Sc, with a diagonal relationship. Rayner Canham's corresponding article on isodiagonality has a PT with Sc over Al, showing the same kinds of relationships, with a diagonal line running through Li-Mg-Al-Ti-Nb-W. I see no controversy here.

Pairing. The remainder of the caption is non-controversial; as you say, there is no controversy with the comment that, "A traditional aspect of teaching the periodic table is to contrast the alkali metals with the halogens." I discussed the pairing approach, including its didactic implications, at length in my article in the Journal of Chemical Education, on organising the metals and metalloids, as cited.

There is nothing new in pairing. It is seen at the highest level in the metals and nonmetals. Literally thousands of authors have written about the associated L-R progression in metallic to nonmetallic character going across the PT e.g., "The elements change from active metals to less active metals, to metalloids, to moderately active nonmetals, to very active nonmetals, and to a noble gas." Welcher (2001). Or consider this quote [square brackets added]:


 * "The horizontal rows are called series. A series begins with an [0] inert gas, the next member of it is [1] strongly electropositive, the [2] next less strongly so, [3] the next is but weakly electropositive, and [4] the next quite indifferent in this respect. Then follow [5] a very weak electronegative, [6] a more strongly electronegative, and finally [7] an element with intense electronegative qualities. The next element [8] is another inert gas and so begins the next series." (Lemon 1946, p. 308)

The pairing approach article (2020) has so far been downloaded 6,500+ times, and cited twice, including by Rayner-Canham, in his article, "Why don't we really teach about the periodic table?".


 * PS: For the post-transition metals and metalloids there is Dingle (2017, p. 101): "…with 'no-doubt' metals on the far left of the table, and 'no-doubt' non-metals on the far right…the gap between the two extremes is bridged first by the poor metals, and then by the metalloids – which, perhaps by the same token, might collectively be renamed the 'poor non-metals'." Sandbh (talk) 09:32, 14 December 2020 (UTC)

Noble gases and metals. For the analogy between the two I gave a citation from Wiberg. The synthesis in 1962 of the first noble gas compound, of approximate composition XePtF6, involved a noble metal. For another example, xenon, in the +8 oxidation state, forms a pale yellow explosive oxide, XeO4, while osmium forms a yellow strong oxidizing oxide OsO4. There are parallels in the formulas of the oxyfluorides: XeO2F4 and OsO2F4, and XeO3F2 and OsO3F2 (Rayner-Canham and Overton 2010, p. 205). Noble metals and the noble gases each have biological applications: --- Sandbh (talk) 04:37, 14 December 2020 (UTC)
 * Brooks R. R. Noble metals and biological systems: Their role in medicine, mineral exploration, and the environment. CRC Press, Boca Raton, 1992
 * Medici S.; Peana M.F.; Zoroddu M.A. Noble metals in pharmaceuticals: Applications and limitations. In: M. Rai; A. Ingle; S. Medici (eds). Biomedical Applications of Metals. Springer: Cham, Switzerland
 * Winkler, D.A.; Thornton, A; Farjot G.; Katz, I. The diverse biological properties of the chemically inert noble gases. Pharmacology & Therapeutics. 2016, 160, 44-64
 * Winkler D. A.; Katz, I.; Farjot, G; Warden, A. C.; Thornton, A. W. Decoding the rich biological properties of noble Gases: How well can we predict noble gas binding to diverse proteins? ChemMedChem. 2018, 13, 1931.

Physical evidence
I compared the trend lines for B-Al-Sc-Y-La; B-Al-Sc-Y-Lu; and B-Al-Ga-In-Tl, for the following properties: 1. density; 2. electrical resistivity; 3. enthalpy of atomisation; 4. enthalpy of fusion; 5. enthalpy of vaporisation; 6. standard molar entropy; 7. heat conductivity; 8. specific heat capacity; and 9. surface tension at the melting point. These properties were used by Horovitz and Sârbu C 2005, "Characterisation and classification of lanthanides by multivariate-analysis methods", Journal of Chemical Education, vol. 82 no. 3, pp. 473–483,. They also used the Gibbs energy of formation of the chloride LnCl3 per mol of Cl, as a characteristic of chemical reactivity. I added this to my chemical property data set and removed density and MP, which explains the updated chemical values below.

The average goodness of fit values, alongside the chemical values were:

Physical    Chemical  Combined B-La    0.63         0.88      0.755 B-Lu    0.49         0.81      0.65 B-Tl    0.80         0.62      0.71


 * *1st update to accommodate nine physical and nine chemical properties; and double checking of property values, and charts, and corrections thereto. More work still to be done here. Sandbh (talk) 05:16, 16 December 2020 (UTC)

Physically, it seems like the d electrons in Sc-Y-La etc exert sufficient influence, compared to the sp electrons in B-Al, to upset the smoothness of physical trends. So the B-Tl trend has a better fit.

Chemically, it seems like the consistency of the underling (sp)8 core in B-La makes for better trends lines, compared to the irregular underlying cores in Lu (spf)22 and Ga-Tl (spd)18 and (fspd)32 Sandbh (talk) 02:30, 10 December 2020 (UTC)


 * The reason this argument doesn't convince me is because most groups in the PT don't have consistent underlying cores. The 3d and 4f contractions are well-known effects in the literature (cf. Greenwood & Earnshaw), and the secondary-periodicity effects that DA mentions are well-known even if not under that name (why Br is so much harder to get in the +7 state than Cl or I, for example). When you proceed Si-Ge (p2), P-As (p3), S-Se (p4), Cl-Br (p5), Ar-Kr (p6), you always add an underlying d10 to the core, so it seems natural to ask that Al-Ga (p1) do the same. (Basically Jensen's argument applied to the p instead of the d block.) Therefore, just because B-La by itself gives a better trend line locally, does not clearly dictate that it should be the standard format, because B-Tl + Sc-Lu gives trend lines that are globally more consistent with other groups. And that is likely the reason why B-Tl is standard. Of course, to rigorise this argument requires clearly defining what exactly a block is, in order to make it clear why we should expect the p1 trend to follow the others. (Without defining what a block is, you cannot answer why they shouldn't be split.) But since most natural definitions that have been thrown around for blocks are based on physics and electronic structure rather than chemistry, they are surely not going to favour B-La. Double sharp (talk) 02:38, 10 December 2020 (UTC)

Yes, B-Al over Sc is not going to happen. Neither can I see He over Be happening. I suppose the situation is at the broad contour level, where the broad contours are the four idealised blocks. Where you go from there depends on the context, and as long as that is explained, you'll be good. For example, moving He over Ne is one chemistry-based step. Moving La-Ac into group 3 is another. Moving B-Al over Sc, old school style, is another. The current popularity of the La form has many historical vectors. Will Lu topple La? I don't know. I tend to doubt it, unless IUPAC adopts this form. Sandbh (talk) 03:11, 10 December 2020 (UTC)

BTW check out this old version of our scandium article, position in periodic table section. Sandbh (talk) 03:19, 10 December 2020 (UTC)


 * I tend to think that Lu will eventually topple La, the reason being the quote from Jones you like to give.

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.
 * From the perspective of those primarily interested in classifying by the final chemistry observed, the Sc-Y-La system is not less than useful, so they won't bother with reexamining it. However, if Sc-Y-Lu had been standard already, it would also not be less than useful. All the elements here are fairly chemically similar, and Y-Lu is not chemically worse than Y-La. It just prioritises different chemical relationships. This rather neatly explains why the general feeling among chemists wrt this issue seems to be not so much supporting one side or the other but total indifference. So if Sc-Y-Lu become standard, chemists of this view would probably just shrug their shoulders and use it anyway, typical of the way in which Be-Mg-Ca and B-Al-Ga got acquiesced to. Because to them the issue is not of any importance.
 * However, from the perspective of those primarily interested in going from physics, Sc-Y-La is quite definitely less than useful, and Sc-Y-Lu is needed for consistency. For them, Sc-Y-La must be scrapped and replaced by Sc-Y-Lu.
 * So you have a situation where physicists and physically minded chemists will push for Lu endlessly, and less physically minded chemists won't care. Ergo, the push for will be greater than the push back, which suggests an eventual Lu victory. That seems borne out by the steady fall in the percentage of textbooks that show a La table. The fact that the IUPAC project got established after fifty years of the fight being around is, in my view, a manifestation of this: fifty years were not enough for Lu supporters to win, but they gave them enough representation and made the dispute visible to the public. It is easy to find different books and different posters with different representations of group 3. And since most people who care about this are going to be Lu supporters for the above reasons, plus Scerri's comments, I find it highly likely that Lu toppling La is something that will eventually happen, because more will be pushing that way than the other. Extrapolation of the figures in textbooks suggests that it should happen around midcentury: of course, if the IUPAC project comes to a Lu conclusion, then that would speed it up quite a bit. (The interesting question might be whether Lu topples La before element 121 is discovered. I suspect that might happen. If not, it would be interesting to see if anyone actually applies the block-start criterion for the La table that I see as flawed in order to delay the g block to start at element 125. I highly doubt it to be honest.)
 * Once that happens, helium becomes the only exception from the four blocks that everyone starts from. (They are "idealised" only from a chemical perspective, I guess; from a physical perspective they are quite real.) So if we look forward in time to a period when He-Ne + Sc-Y-Lu is standard, I can actually see arguments for He-Be starting to become more mainstream. They are already appearing now in significant journals from noble gas chemists (Grochala, Grandinetti), which is already more recognition than the self-published book of Henry Bent. Extrapolation naturally suggests that, especially as helium chemistry takes off, suddenly He-Ne will seem more and more like an anomaly to more chemists. But that is a far more fraught extrapolation. Double sharp (talk) 03:32, 10 December 2020 (UTC)


 * This perfectly ties into one of the main arguments behind my support for Lu-Lr. We see with known elements that ground state electron configurations (mostly) shape the blocks, yet there is a single exception to the pattern in the case of La and Ac. Scerri himself says that we don't make an exception for Th to delay the onset of the actinides, and starting the g-block at 125 seems very imprudent. Not to mention relativistic effects and the fact that if/when 125 is discovered, it will certainly be of more interests to physicists. In other words, if every exception to the Aufbau rule or deviation due to relativistic effects were accommodated, the periodic table would be a jumbled mess, and elements would no longer be in a meaningful order of atomic number or grouped by similar chemical properties and valence electron configurations. (And this would likely confuse everyone ranging from high school chemistry students to theoretical physicists.) As I see it, splitting group 13 not only disregards the issues at the root of the group 3 debate, but justifies breaking trends and whimsically making exceptions. I know this is not a proposal for change on WP, but I very strongly oppose it on these grounds. ComplexRational (talk) 16:47, 12 December 2020 (UTC)
 * Thank you! Double sharp (talk) 07:25, 13 December 2020 (UTC)

What's important
Regardless of the form of the periodic table, the most important consideration is to explain the context for the chosen form.
 * Splitting the s-block, so that He is over Ne; and splitting the p-block so that B-Al is over Sc (as was done historically, pre the quantum revolution, and even after by e.g. Pauling); and splitting the d-block, is chemically congruous.
 * From there, an electronic concession can be made by placing B-Al over Sc, as is the case with the currently popular form.
 * From there:
 * One can make a concession to chemistry, and show a 15-column wide f-block as per the IUPAC form; or
 * Another concession can be made to electronics by showing Lu in group 3, as is the case with the "popular" Lu form.
 * Finally, one can go the whole electronic hog and move He over Be.

As Scerri says, there is no optimal form of table:. Each of the above options have their place. None is worse or better than any other one. The only confusion that arises is caused by not explaining the context for the different forms.

The popular form is an accident of history combined with acceptance by the chemistry community whereby the noble gases where all placed into group 0 because of their seeming inertness; La made it into group 3 because it was discovered before Lu, and when it was discovered that the f-shell closed at Yb rather than Lu, nothing changed about the chemistry involved, so La stayed where it was; B-Al were moved out of group 3 and into group 13, on electronic grounds (never mind the damned chemistry). This is all easily explained, as are the four less popular variations.

Here at WP we can't do anything per se, about the fact the La form is currently the most popular. We can however write about the contextual relevance of this form, and at least the other four variations. Then we could stop "arguing", so to speak, and focus yet more on improving articles. --- Sandbh (talk)

Conclusions
Based on these eight properties:
 * Group 3 is best shown as B-Al-Sc-Y-La.
 * If B-Al remain in group 13, then group 3 is more cohesive as Sc-Y-La-Ac.
 * Seemingly the focus on lining up groups into pure blocks detracts from the regularity of periodic trends.

Pauling, in his book, General Chemistry (1988) has group 3 as B-Al-Sc-Y-La in his PT showing EN values (p. 182). Elsewhere he refers to the congeners of B as being Al-Sc-Y-La. He discusses Ga-In-Tl separately in his chapter on Cu, Zn, Ga and their congeners. He writes, "The metals Zn, Cd, and Hg (group IIb) are also much different from the AEM (group II), as are Ga and its congeners (group IIIb) from the elements of group III." (p. 697)

Goodness! Not only do we have a split s-block, and a split d-block, now we have a split p-block! --- Sandbh (talk) 06:49, 4 December 2020 (UTC)
 * It's nice to see this being discussed. ^_^
 * What you have shown is exactly why, as you know, I am so against La in the d block. Because as you have just demonstrated, the arguments that lead to cleaving group 3 away from group 4 also succeed marvellously at cleaving group 13 away from group 14. Of course I was also saying that in the previous discussion, so it pleases me to see that you're seeing this now. Indeed, see the quote you used at Wikipedia talk:WikiProject Elements/Archive 49 from R. J. P. Williams:

2. THE CHEMICAL CLASSIFICATION OF ELEMENTS

"The classiﬁcation of elements in the Periodic Table is now known to be a reﬂection of restrictions imposed by quantization of energy states of electrons in atoms. However, without recourse to other than empiricism in the study of chemistry the same classiﬁcation had been observed for over 100 years. In fact it has long been a standard educational practice to separate elements into Groups IA, IIA, and IIIA; transition metals; Groups IB, IIB and IIIB; and the non-metals of Groups IVB to VIIB of the Periodic Table to simplify discussion of their chemistry. Although the distinctive properties in aqueous solution of each of the four classes does not provide sharp divisions it is very useful to treat separately three types of metal: Groups IA, IIA and IIIA metals are associated with equilibrium ionic-model chemistry; transition metals with one-electron redox chemistry and, across each such series, increasingly covalent chemistry concommitant with increasing Lewis-acid strengths of ions, usually at equilibrium with their surroundings; and Groups (IB), IIB and IIIB metal ions with a compromise ion chemistry involving strong Lewis-acid properties while maintaining fast equilibration but little redox activity…Finally, there is the further chemistry of non-metals…"
 * So a split between groups 13 and 14 is just as present as one between groups 3 and 4. Behaviour of Al is also closer mostly to Sc than to Ga, according to Rayner-Canham. Metallicity drops clearly going Al to Si, Ga to Ge, even In to &alpha;-Sn. The +3 state of Al-Ga-In has significant ionic tendencies that are not present for the +4 state of Si-Ge-Sn. And in fact, because Cn-Nh-Fl breaks the 234 pattern in all likelihood, whereas Ra-Ac-Th does not, in fact the 234 pattern also points to B-Al-Sc-Y-La-Ac as group 3. Isodiagonality also works better with Al atop Sc, as shown in your own article on group 3.
 * I also point out that at Names for sets of chemical elements there is "earth metals" that unites groups 3 and 13, and that preserves left-to-right order better with B-Al-Sc: B, Al, Sc and Ga, Y and In, etc. A +3 d-block contraction is highlighted better with B-Al-Sc, because Ti3+ is the first tripositive cation with a d electron and Ga3+ is the last. Shchukarev's imitator trend, if we believe in Sc-Y-La, looks more consistent with a Ti-Ga d block: then we start each half with +4 (Ti, [Co not so good]), dip down to +2 (Mn, Zn) before restoring +3 (Fe, Ga) at the last element; just like respectively Ce/Tb (and indeed again Tb is much less good at this than Ce!), Eu/Yb, Gd/Lu.
 * I hope you can now see why I was always saying that Sc-Y-La alone, without something like B-Al-Sc, is a double standard. Maybe I would be less harsh about it now, but I'd still say you can't have one without the other: the same sort of arguments imply both. (I took some of your ten arguments above and applied them to B-Al-Sc!)
 * Well, you seem to be seeing it as "since I believe Sc-Y-La, I have to believe B-Al-Sc". And that's consistent! Only, I think you'll find some more problems once you look at C-Si-Ti-Zr-Ce-Th. That trend is linear and obeys a 345 rule, whereas C-Si-Ge-Sn-Pb-Fl shows kinks at every block insertion and fails 345 at Nh-Fl-Mc (likely 3-2-3).
 * But I see it a little bit differently. I say "I cannot believe B-Al-Sc". After all, the vast majority of chemists show B-Al-Ga. Contradiction. Therefore, the logic goes: "since Sc-Y-La implies B-Al-Sc, and that's wrong, Sc-Y-La must also be wrong". Because I have no confidence in an argument that fails to give the expected answer in the non-controversial cases. If it can fail for group 13, then why should I believe it will not fail for group 3 too? I am not so sure of my own chemical knowledge as to think I have the right answer here when almost no chemist sees group 13 as an unsettled question anymore. Does anyone think so other than Habashi, really? Granted, Sc-Y-La is the more common form there, so either way a significant number of chemists are wrong. But Sc-Y-Lu is gaining ground, it has already a significant minority: there's hardly any minority arguing for B-Al-Sc anymore. Even for He-Be there is at least a minority of multiple people, a significant number of whom are noble gas chemists. And I can far more readily believe that some but not all chemists are wrong for the not-so-common rare earth elements, than that practically all chemists are wrong for ultra-common aluminium; especially when a majority of chemists who closely examines the group 3 question comes out in favour of Lu.
 * And then everything is clear: it's simply that after a contraction, things change. B and Al are normal; then Ga has a kink in the trend, because 3d10 has been inserted. We should not expect it to continue the trend. We know Ga comes after the 3d contraction, because it cannot use 3d anymore: Zn is the last element that can do that. That's exactly the same as the kink that happens going from C to Si to Ge; from N to P to As; and so on. In the exact same way: Sc and Y are normal; then Lu has a kink in the trend, because 4f14 has been inserted. We should not expect it to continue the trend. And we know Lu comes after the 4f contraction, because it cannot use 4f anymore: Yb is the last element that can do that. Just like the kink that happens going from Ti to Zr to Hf; from V to Nb to Ta; and so on. As Rayner-Canham noted in the article you emailed me, in fact the nice linear trends you see in the alkali and alkaline earth metals are not the norm at all in the periodic table. So we shouldn't ask for a goodness of fit to a linear trend, when we're not expecting a linear trend. It'd be like taking points perfectly spaced on the parabola y = x2, and saying that there's no relationship between the variables because you can't fit a line. Of course you can't, the trend is clearly quadratic in that case.
 * If you look only at chemistry, you'll never decide between B-Al-Sc and B-Al-Ga; they both look plausible. If you like linear trends too much, you may even go for B-Al-Sc. It's just like Sc-Y-La vs Sc-Y-Lu. That's why chemistry alone is inconclusive and you have to drill down to the electronic structure. As Scerri noted, chemical and physical properties will never provide a decisive answer, and you've got to go for something more fundamental. And that is electronic structure which has the bonus of explaining chemical and physical properties. That's why I go for strict physics and strict blocks alone. And even in your article on group 3, you say that "From a Platonic symmetry perspective or perhaps that of physics ... it can be argued that lutetium is better placed under yttrium" in the conclusion (my emphasis). That should explain to you briefly why I am a strong partisan for the Lu side: I have rejected chemistry as a criterion as inconclusive and decided on physics, following the philosophy of Scerri: 'Scerri thinks a focus on chemical or physical properties is misguided. He compares it to early botanists’ classification of flowers by their color or petal number. “You’ve got to go for something fundamental,” Scerri says, like electronic configuration. “Just to amass properties is never going to give you a definitive answer.”' And that's why the table I am convinced is theoretically correct has not a single split block: even helium takes its place in the s block above beryllium. Of course, just because I'm convinced by it does not necessarily mean it's the right answer, just that I find it the most convincing. But we'll see what the IUPAC project says.
 * It is alright if you are not convinced right now. After all, it would be very hypocritical of me to criticise you for not being immediately convinced, when I was vacillating about the group 3 issue for quite a while after DA talked to us in the middle of 2018, and only finally decisively rejected Sc-Y-La in late 2019 / early 2020 and never looked back. And of course there's no reason this should impact what we show on WP for now when we haven't heard from the IUPAC project yet. After all: I think He-Be is definitely the correct form, but I don't advocate it for the mainspace because it's so obviously not the majority view. So in this thread we're just having a friendly chat, bringing to the table interesting points. But it really pleases me to see that you're now applying your argument outside the bounds of group 3, both because it means I feel you're no longer artificially limiting your arguments, and also because it means you're now examining a case where the standard classification doesn't seem less than useful to any significant number of people rather than deciding it's not worth losing sleep about. That is exactly what took me on the first step to decisively rejecting Sc-Y-La: using B-Al-Sc as a test case. And the beauty of this is that this is science: if the right evidence comes in, I will reevaluate my stance. As you may know, if some evidence came in suggesting that actually Lu had some categorical 4f valence involvement and it wasn't just buried and giving incomplete shielding like it is for Hf, I'd certainly move to thinking that Sc-Y-La isn't that bad after all. It's a contest of ideas only, not personalities, and a nice diversion while we take a break from describing what is in the literature, and dream of what perhaps should be – and in my case, what I hope will soon be. Before returning to earth of course once we saunter out into the mainspace. ^_^
 * I'll probably not have time to reply at this length after today, so I'll let you think about it. I also feel that this is the sort of thing that works better if you think about it by yourself for a while: that's how it worked for me. And so I think it's best if this is my only long reply about this topic to not let it get out of hand. However you choose to see it, it will for you hopefully be a journey of scientific inquiry and examination. And I feel it is always good to go down such a journey away from the WP mainspace sometimes. Whatever your conclusions are, I have no doubt they will be interesting, even if it's away from the WP mainspace. Double sharp (talk) 17:44, 4 December 2020 (UTC)
 * Just to get it: what would the split be, like? -DePiep (talk) 00:13, 6 December 2020 (UTC)
 * I don't recall having seen a split p-block in any of my studies or work in chemistry, so I would need a LOT of RS to be persuaded that any such split being shown widely on WP would be justifiable on DUE grounds. Given the reference to Pauling, there may be a place for a mention somewhere.  Double sharp's comments on how the idea led to a firm view on La v. Lu etc, if they reflect discussion in RS, could be of relevance for main space somewhere as well.  This thread includes much OR, it appears to me, but there are parts that could be article-space relevant if sufficient sources exist.  EdChem (talk) 00:26, 6 December 2020 (UTC)
 * Yes, that's why I ended off by saying that it was interesting to talk about but not for the mainspace. Only source I'm aware of that mentions analogy of situations of B-Al-Sc-Y vs B-Al-Ga-In to Sc-Y-La-Ac vs Sc-Y-Lu-Lr and uses it to argue for the latter situation is a 2004 Russian article by : I think it is cited under Periodic table for something else already. Might be worth a tiny mention, but not a do-or-die thing for me as it is a minor thing. Double sharp (talk) 08:08, 6 December 2020 (UTC)

What it looks like


There it is, a PT with split s-, p- and d-blocks.

I cottoned on to this as a result of reading Greenwood & Earnshaw's discussion of the chemistry of Al-Ga. I recall I was prompted to do so by something DS wrote, which I've now forgotten. Anyway, G&E have a graph (p. 223) showing trends in successive ionisation energies of B-Al-Sc-Y-La v B-Al-Ga-In-Tl. They write:


 * "It is notable that these irregularities for the Group 13 elements do not occur for the Group 3 elements Sc, Y, and La, which shows a steady decrease in IE from B to Al, all 5 elements having the same type of underlying core (noble gas). This has a decisive influence on the comparative chemistry of the two subgroups…the steady decrease in EN in the series B > Al > Sc > Y > La > Ac is reversed in Group 13 and there is a steady increase in EN from Al to Tl." (pp. 222, 224)."

A similar pattern is seen in standard electrode potentials (p. 225).

B-Al-Sc-Y-La tables

 * Rang (1893)
 * Gooch & Walker (1905)
 * Cuthbertson & Metcalfe (1907)
 * Baur (1911)
 * Rydberg (1913)
 * Black & Conant (1920)
 * Lewis (1923)
 * Hubbard (1924)
 * Deming's table (1925), which popularised the medium-long form
 * Antropoff (1926)
 * LeRoy's table (1927)
 * Irwin (1939)
 * Seaborg (1945), with Al in group 3 and in group 13
 * Yost & Russell (1946)
 * Coryell (1952)
 * Pauling's table (1960)
 * Habishi's Metallurgist's Periodic Table (1992), as mentioned by DS (Habishi leaves B in group 13).

I suspect what has happened here is that it was historically known that group 3 was better represented as B-Al-Sc-Y-La-[Ac]. Then, with the advent and rise of modern electronic structure theory, B-Al- got moved to the p-block because, after all, they were p-block elements, never mind the damned chemistry. And La stayed in the d-block since it was the first element to show 5d electron, and 4f did not show until Ce. DS and I have discussed the latter phenomenon at length.

Nowadays, this represents forgotten chemistry. It underscores the fact that electron configurations are not the final arbiters of periodic trends.


 * I found the smoking gun. Here's an extract from Mellor's modern inorganic chemistry (1943 impression of 1939 edition, p. 677):


 * "21 Relationships of the elements of Group III. The exact sub-classification to be adopted in this group has in the past occasioned some difficulty, but it is now generally agreed that B and Al are best associated with Ga, In and Tl. This is supported by the conclusions at present accepted for the electronic configurations of these elements…"


 * So B-Al- over Ga-In-Tl is an electronic decision albeit B-Al- over Sc-Y-La is a better fit in terms of chemical and physical periodic trends going down group 3. Fortunately, these trends can still be seen via the n + 10 relationship, as discussed by Rayner-Canham (2020), for group 3 and group 13. Sandbh (talk) 00:09, 11 December 2020 (UTC)

B-Al-Sc-Y-La-Ac as a linking or bridging group
It's odd to see how group 3 consistently represents a linking group. So B-Al link the s-block with the bulk of the p-block; Sc-Y link to the bulk of the d-block; and La-Ac link to the f-block.

On Sc-Y-La-Ac, Greenwood & Earnshaw write:


 * "All are rather soft-slivery-white metals, and the display the gradation in properties that might be expected for elements immediately following the strongly-electroposive AEM and preceding the transition metals proper." [italics added]

B is between Be (a metal) and C (a nonmetal). G&E note B has many similarities with C (and Si). At the same time, Greenwood (2001) commented:


 * "The extent to which metallic elements mimic boron (in having fewer electrons than orbitals available for bonding) has been a fruitful cohering concept in the development of metalloborane chemistry…Indeed, metals have been referred to as "honorary boron atoms" or even as "flexiboron atoms". The converse of this relationship is clearly also valid…"

Al is between Mg, a metal and Si, a metalloid. It has some properties that are unusual for a metal; its crystalline structure shows some evidence of directional bonding. Aluminium bonds covalently in most compounds. The oxide Al2O3 is amphoteric and a conditional glass-former. Stott (1956) labels aluminium as a weak metal. It has the physical properties of a metal but some of the chemical properties of a nonmetal. Steele (1966) notes the paradoxical chemical behaviour of aluminium: "It resembles a weak metal in its amphoteric oxide and in the covalent character of many of its compounds…Yet it is a highly electropositive metal…[with] a high negative electrode potential". Moody (1991) says, "aluminium is on the 'diagonal borderland' between metals and non-metals in the chemical sense."

--- Sandbh (talk) 06:31, 6 December 2020 (UTC)

EdChem

 * So, if I am following correctly:
 * Habishi argued in 1994 for moving Al but not B to group 3
 * Greenwood and Earnshaw comment on the similarity of metals to B (also noted by Habishi) and note irregularities in IE, while retaining B and Al in group 13 and so not actually presenting a split p-block
 * Moody's comment from 1991 that Al is on the "'diagonal borderland' between metals and non-metals" sounds to me to be consistent with placing B and Al in group 13 alongside all other such elements
 * And, you have no RS literature in support of a split p-block with B and Al in group 3 that was published in the last 60 years?
 * Arguing that much older work had such placements tells us nothing about the present RS consensus, and nor does referring to it as "forgotten chemistry." Do you have any significant body of recent RS that actually uses a split p-block?  If not, then what you are discussing may be useful for our articles on the history of the PT but it doesn't offer any reason to alter the presentation of the PT as it is currently constructed by RS and used in practice.  EdChem (talk) 08:18, 6 December 2020 (UTC)

Yes, you're following correctly. It's interesting to talk about and not for the mainspace other than in a historical sense. I refer to it as forgotten chemistry since modern texts do not discuss the idea of a split p-block. It's peculiar that modern periodic tables seem to be electronic block-based tables, rather than chemistry-based tables in which B-Al are over Sc. Even Seaborg, who arguably contributed to the dominance of the electronic table, showed Al in two places ie group 3 and 13. Looking at the list of tables with B-Al over Sc, there are some other relatively big hitters, too: Rydberg; Lewis; Hubbard; Deming; Antropoff. What was it that these luminaries knew about B-Al-Sc-Y-La-Ac that is deemed to be no longer relevant, and why is that the case?

I recall seeing split-p block tables from time to time but have never paid much attention to them. The most recent one I have at hand is in the corrected 1988 edition of Pauling's General Chemistry.

Deming (1947, Fundamental Chemistry, 2nd ed. p. 617) located Al with the pre-transition metals in groups 1−2. Cox (2004, Inorganic Chemistry, 2nd ed. p. 185) refers to the pre-transition metals as those in groups 1 and 2, and Al. Here's a 2019 periodic table (by me), recording oxidation number trends, further suggesting B and Al are better placed over Sc.

Rayner-Canham (2020, The periodic table: Past, present, and future, pp. 178–181) writes:


 * "It was Rang in 1893 who seems to have been the first, on the basis of chemical similarity, to place boron and aluminum in Group 3.


 * Such an assignment seems to have been forgotten until more recent times. Greenwood and Earnshaw have discussed the way in which aluminum can be considered as belonging to Group 3 as much as to Group 13 particularly in its physical properties. Habashi has suggested that there are so many similarities between aluminum and scandium that aluminum’s place in the Periodic Table should actually be shifted to Group 3.


 * In terms of the electron configuration of the tripositive ions, one would indeed expect that Al3+ (electron configuration, [Ne]) would resemble Sc3+ (electron configuration, [Ar]) more than Ga3+ (electron configuration, [Ar]3d10). Also of note, the standard reduction potential for aluminum fits better with those of the Group 3 elements than the Group 13 elements (Table 9.2) — as does its melting point.


 * In terms of their comparative solution behavior, aluminum resembles both scandium(III) and gallium(III). For each ion, the free hydrated cation exists only in acidic solution. On addition of hydroxide ion to the respective cation, the hydroxides are produced as gelatinous precipitates. Each of the hydroxides redissolve in excess base to give an anionic hydroxo-complex, [M(OH)4]−…There does seem to be a triangular relationship between these three elements. However, aluminum does more closely resemble scandium rather than gallium in its chemistry. If hydrogen sulfide is bubbled through a solution of the respective cation, scandium ion gives a precipitate of scandium hydroxide, and aluminum ion gives a corresponding precipitate of aluminum hydroxide. By contrast, gallium ion gives a precipitate of gallium(III) sulfide. Also, scandium and aluminum both form carbides, while gallium does not."

As a chemist, what is your take on the merits of B-Al-Sc-Y-La-Ac in comparison to B-Al-Ga-In-Tl? Sandbh (talk) 22:46, 6 December 2020 (UTC)
 * As a chemist, I think the periodic table has an unsplit p-block and B and Al are p-block elements. As a Wikipedian, I think the discussion is wandering into WP:NOTAFORUM territory.  As a researcher (and engaging in a little OR), I think that the correlations (if I am not mistaken in how they are constructed) have serious problems with validity in a statistical sense.  EdChem (talk) 12:22, 7 December 2020 (UTC)

How the B-Al-Sc vs. B-Al-Ga situation relates to the Sc-Y-La vs. Sc-Y-Lu situation
As noted above the source I am aware of that mentions this is an old article by (= Andrey Kulsha). Link is here. I grant it is not formally published but seems to be something he did as a student at Belarusian State University in 2004. But he has gotten published outside as an academic, here (some computational chemistry about superacids) and here (in 1999 as a student, but already in the journal Хiмiя: праблемы выкладання 1999, 5, pp. 102–109; this is his article supporting Lu in group 3). Hopefully this is good enough as an RS for WP; if you feel it's not, I will listen to that.

Следует заметить, что до недавнего времени к семействам f-элементов относили не La – Yb и Ac – No, а Ce – Lu и Th – Lr, в то время как d-элементами IIIб группы были La и Ac вместо Lu и Lr из-за прохода первого (n – 2)f-электрона на (n – 1)d-подоболочку [4]. Однако детальное исследование закономерностей изменения химических и физических свойств элементов [5 – 6] в рядах Ba – Hf и Ra – Rf [7], а также сравнительное исследование групп Sc-Y-La-Ac и Sc-Y-Lu-Lr [8] и проведение аналогий с группами B-Al-Sc-Y и B-Al-Ga-In соответственно показало, что целесообразнее, практичнее и закономернее относить к f-элементам именно семейства La – Yb и Ac – No [3]. В настоящее время вариант периодической системы со «сдвинутыми» семействами еще распространен [4], однако он постепенно вытесняется новой формой, более строгой формально и лучше описывающей реальные свойства f-элементов и d-элементов IIIб подгруппы.

Translation courtesy of Google Translate: "It should be noted that until recently, not La - Yb and Ac - No, but Ce - Lu and Th - Lr were attributed to the families of f-elements, while the d-elements of group IIIb were La and Ac instead of Lu and Lr from for the passage of the first (n - 2) f-electron to the (n - 1) d-subshell [4]. However, a detailed study of the regularities of changes in the chemical and physical properties of elements [5 - 6] in the series Ba - Hf and Ra - Rf [7], as well as a comparative study of the Sc-Y-La-Ac and Sc-Y-Lu-Lr groups [8 ] and drawing analogies with the B-Al-Sc-Y and B-Al-Ga-In groups, respectively, showed that it is more expedient, more practical, and more logical to attribute the families La - Yb and Ac - No to f-elements [3]. At present, the version of the periodic system with "shifted" families is still widespread [4], but it is gradually being replaced by a new form, formally more rigorous and better describing the real properties of f-elements and d-elements of IIIb subgroup."

If I used it, it would in any case only be for a sentence in the main group 3 element article, along these lines: "Kulsha pointed out that the changes going down Sc-Y-La-Ac and Sc-Y-Lu-Lr are respectively analogous to those going down B-Al-Sc-Y and B-Al-Ga-In; since B-Al-Ga-In is standard, this constitutes an argument for Sc-Y-Lu-Lr." In any case there are very many other arguments for Lu in the literature there, so even if this is considered not RS enough, it's still OK with me. Again, it's just for a descriptive sentence, in what's basically a subarticle and not the main periodic table article, not about changing the periodic table format.

I'm not arguing that we change to -Lu-Lr right now on WP anymore. Oh, yes, I still think it's the absolutely correct form and all that, but I also know that's not the point for WP. Please correct me if I'm wrong about the outside situation, but it seems the situation there among chemists and on WP is this:
 * The majority of chemists don't care and if anything wonder what all the fuss is about.
 * A minority thinks the issue is of great and paramount importance.
 * When this minority writes articles about the matter, a crushing majority of them argues for -Lu-Lr for various reasons. (Link is to all the sources I've managed to collect, or at least have mentions of, in the literature. If you know somewhere where I can find the ones that don't have links, please write on my talk page!) There are a few who care greatly about this and who think the situation supports -La-Ac (Sandbh's view), but it is very few. So: as a general rule, those who care about this issue think -Lu-Lr is the scientifically correct, or at least a scientifically better, form.
 * However, -La-Ac was the traditional form for a long time. Therefore, among those who wonder what all the fuss is about, that form predominates because it's what they learnt and they don't think there are any strong reasons to change it.
 * Therefore the effect is that most textbooks continue to use the -La-Ac form.
 * However, some textbook writers have been convinced by the -Lu-Lr partisans to change their periodic tables. So thanks to them, the amount of support for -La-Ac has eroded noticeably, but it still has a majority.
 * The debate also seems to be responsible for the -*-** compromise of IUPAC gaining ground, but that is not as common either. See this 1988 report by IUPAC, in which Sc-Y-Lu was admitted to be the correct form according to electronic configurations (with some sources being given to justify this), but Sc-Y-* was chosen as a compromise and appeared in the next Red Book because Sc-Y-La dominated in the literature. Although inconsistently: in the next 1990 Red Book, the periodic table was shown in 18 column format as Sc-Y-*, and then on the next page in 32 column format as Sc-Y-Lu. I don't know if this was just a mistake or a deliberate attempt to please everybody; if so it clearly didn't work. The 32 column form was dropped from the 2005 Red Book anyway, so for now IUPAC shows Sc-Y-*, but does not recommend it. Strange, I know. Judging by the fact that if anything the rate of publications on this argument increased rather than decreased after 1988, it seems that this ended up being the classic sort of compromise that satisfied nobody that was promptly ignored in favour of continuing the fight.
 * Therefore the average reader is likely to be familiar with all forms because it's easy to find textbooks on both sides (La-Ac: Greenwood & Earnshaw; Lu-Lr: Clayden et al.) and even popular-science books on both sides (La-Ac: The Cartoon Guide to Chemistry, Gonick & Criddle; Lu-Lr: The Periodic Table: Its Story and Its Significance, Scerri), but -La-Ac is what he or she will find most of the time when reading off WP at his or her likely level if he or she is looking up basic stuff about the PT anyway.
 * Ergo, the conflict is between which sources to listen to. Should we go for the sources people are more familiar with, in which -La-Ac predominates with -Lu-Lr as just a rising upstart in the minority? Or should we go for the sources that directly take on the matter, in which -Lu-Lr crushingly predominates, except for the small problem that those sources are written at a level quite a few levels up from the basics that the periodic table is at?
 * The issue has been raised a couple of times this year by me, and it is somewhat split; but the viewpoint in favour of going for the sources directly about the thing did not get a clear majority. So I have to admit there is currently no consensus for a change. Both viewpoints seem legitimate to me by the measure of "it's not against WP policy", and I do still think WP is worse off for not showing -Lu-Lr, but I think that pales before the need to respect consensus. Without that, WP can no longer remain WP. So I hold my peace.
 * The issue has been going on since 1965 and apparently enough students and instructors are confused by the differences in group 3 found in the literature that IUPAC started a project to deal with it in 2015. Its chair's words here make me hopeful that -Lu-Lr will be the ruling. If that happens, the chance of getting a consensus change will be far greater, and then I will propose it again. But not before that and not if it doesn't happen. After all: at any time it is possible that they rule something that isn't -Lu-Lr, and then suddenly -Lu-Lr has even less of a position by WP policy than before. So there isn't any point in raising it again now until and unless a -Lu-Lr ruling happens soon: not much will have changed.

So, to make it clear, I'm not looking to change anything about how we present group 3 now. (I mention this here because I'm not sure I've said all this in one place yet.) It has been fairly tried and gotten no consensus, and although I still prefer -Lu-Lr as everyone knows by now, WP has to work by consensus and I respect that. Not only that, a IUPAC ruling is coming that with a bit of luck could end the entire debate: since we usually wait for IUPAC approval in any case when new elements get discovered (the only common way the PT changes at this moment), it makes sense to use that as a precedent here and wait for it too.

It is only that Sandbh's mentioning of the group 13 situation made me remember that this argument exists in the literature, and wondered if it would be permissible by WP standards to give it a one-sentence mention in the article group 3 element that discusses the arguments (of course not in the main periodic table article for which it would be overkill). That's all I'm wondering. Double sharp (talk) 11:09, 6 December 2020 (UTC)


 * DA's paper is less than reliable.


 * The key sentence is:


 * "However, detailed study of the regularities of changes in the chemical and physical properties of elements [5 - 6] in series Ba - Hf and Ra - Rf [7], as well as a comparative study of groups Sc-Y-La-Ac and Sc-Y-Lu-Lr[8] and drawing analogies with the groups B-Al-Sc-Y and B-Al-Ga-In, respectively, showed that it is more expedient, more practical, and more logical to attribute to f-elements precisely the families La - Yb andAc - No [3].


 * No summary is provided as to "the regularities of changes in the chemical and physical properties of elements [5 - 6] in series Ba - Hf and Ra - Rf [7]."


 * References 5−7 are:


 * 5. Ugai Ya.A. General and inorganic chemistry. - M .: Higher school, 1998
 * 6. Physical quantities: Handbook. - M .: Energoatomizdat, 1991.
 * 7. Kulsha A.V. f-Elements in the periodic system of D.I.Mendeleev // Chemistry: problems laid out 1999, No. 5, pp. 102 - 10


 * I expect these references would be hard to check.


 * Reference 3 i.e. Winter M. WebElements Periodic table. URL: http://www.webelements.com, is not a RS.


 * Reference 8 is Jensen. Scerri wrote that Jensen was too selective in the evidence he put forward in order to support his case.


 * The real-life situation seems to be as follows:


 * 1) An electronic (physics-based) table would show (a) He over Be; and (b) group 3 as Sc-Y-Lu-Lr; and (c) group 13 as B-Al-Ga-In-Tl
 * 2) A chemistry-based table would show (d) He over Ne; and (e) B-Al over Sc-Y-La-Ac.


 * What we have instead is a hybrid table with 1(c) and 2(d).


 * Is that a fair summary? --- Sandbh (talk) 23:42, 6 December 2020 (UTC)
 * I am asking, actually. I am not sure why WebElements is not an RS considering that we actually use it in chemical elements data references. Source #7 you mention is exactly where the "regularities of changes" are mentioned: it is his 1999 article that I already linked you. The selectivity of Jensen seems to be only a charge that Scerri levels: it did not seem to stop many people from citing and referring to him as the authority for the Sc-Y-Lu table, and Scerri himself also supports Sc-Y-Lu for different reasons.
 * I also do not think you have a fair summary of the chemistry-based situation. (The physics-based one is correct, with maybe some quibbles about He-Be since He chemistry is not too well-known yet.) The grounds that you accept B-Al-Sc for (mostly historical and chemical) are exactly those that have in the same era been used to support Be-Mg-Zn, according to RS. As Jensen noted, Be and Mg do share more in common with Zn than with Ca from a chemical perspective. Jensen explained the situation in that paper:

As we have seen, from a spectroscopic point of view, the members of the Zn group are analogous to the alkaline earth metals or group II elements. From a chemical point of view, Zn and Cd most resemble Be and Mg, not only in terms of their atomic radii, ionic radii, and electronegativities (Table 4), but also in terms of the structures of their binary compounds and in their ability to form complex ions with a wide variety of oxygen and nitrogen donor ligands (including complex hydrates and amines). Indeed, prior to the introduction of electronic periodic tables, the similarity between Be and Mg and Zn and Cd was often considered to be greater than the similarity between Be and Mg and the rest of the alkaline earth metals (Ca–Ra). Many inorganic texts written before the Second World War placed their discussion of the chemistry of Be and Mg in the chapter dealing with the Zn subgroup rather than in the chapter dealing with the Ca subgroup, and the same is true of many older periodic tables, including those originally proposed by Mendeleev (34, 35). Even as late as 1950, N. V. Sidgwick, in his classic two-volume survey of The Chemical Elements and Their Compounds, felt that it was necessary to justify his departure from this scheme in the case of Mg (36)...
 * And indeed, if you look at modern texts, you will find that indeed those similarities exist and are widely acknowledged. Greenwood & Earnshaw on p. 113 says it: "In these properties [peroxides and dissolving directly in ammonia are mentioned], as in many others, the heavier alkaline earth metals [Ca, Sr, Ba, Ra] resemble the alkali metals rather than Mg (which has many similarities to Zn) or Be (which is analogous to Al)." What's changed is that nobody thinks it has significance for PT placement anymore. Of course, you can't exactly have Be-Mg-Zn and B-Al-Sc in the same periodic table. It would break the order of increasing atomic number because B would have to come to the left of Be. That doesn't look good for your argument: if we insist that chemical similarity is the first criterion, we end up with both Be-Mg-Zn and B-Al-Sc and hence break the periodic law!
 * Jensen has clearly pointed out here that this is simply coming from the change between chemical and physical understandings of the periodic table. B-Al-Sc and Be-Mg-Zn both come from this same era (of course never in the same book because it breaks the atomic number order). And in both cases, we can still find significant relationships to the other one: Al-Ga is well-known, Mg-Ca is also well-known. Look at any chemistry textbook that describes main group elements, some of their similarities will be mentioned. Over 50 or 60 years ago, because physics and electronic periodic tables weren't the norm or widely understood yet, people were using chemistry as the sole guide. But modern periodic tables are based on physics and electronic structure, not chemistry. That's also why, for example, hydrogen is placed where it is over lithium, and not over fluorine. (Helium over neon is arguable because one can think of 1s2 vs 2s22p6 as the core, in which case electronic structure is not against He-Ne: I expect this will change as helium chemistry gets revealed, because that is what proves that those are not just core electrons. Which is fascinatingly consistent with how a significant number of He-Be partisans seem to be noble gas chemists: Grochala, Grandinetti.) As Scerri himself noted, chemical arguments are not conclusive: they are rather the equivalent of classifying flowers by colour or petal number rather than genetics. (Jensen citing Jorgensen also pointed out that the path from electronic structure to final chemical behaviour is not always that straight, especially for the d and f block elements with multiple low-lying excited states that contribute for chemistry. Seems to me just like the winding road from genotype to phenotype. ) And as Jensen also noted, if you just pick and choose elements from different rows and columns arbitrarily, then thanks to the possibility of trends cancelling each other you can prove pretty much anything, which is no good. As he said, you have to get the self-consistent periodic trends from the atomic structure. That's why his latest publicly expressed stand (2017) seems to be that the most important argument for -Lu-Lr is because La and Ac have low-lying f orbitals that they use for chemistry, but Lu and Lr don't: in other words, the physical argument is coming first for him too. That is why B-Al-Ga is standard, that is why Be-Mg-Ca is standard. Elements are placed in blocks according to their electronic structure, even if some other elements may be chemically more similar. And although I still can't find a source making the point I am about to make, it boggles my mind exactly how chemical properties could still be thought of as a valid way to build up the periodic table today, when everyone is familiar with nitrogen and bismuth in the same group. You could hardly come up with two more different elements. (If you object that N and Bi share the maximum oxidation state +5, simply make it nitrogen and moscovium.) In short: what you have, talking about chemistry, is widely agreed by authorities to be a historical view: everything I said above that isn't in small print is from the RS. My impression concurs with EdChem's. Other than a few people on the side, nobody thinks these added chemical similarities actually mean the periodic table should change anymore. Be-Mg-Zn and B-Al-Sc are interesting sidelines in the history of the periodic table, connected with the trouble encountered when you expand Mendeleev's original 8 column format to the modern 18 and 32 column ones; but they are only historical sidelines. As could be expected because Be and Mg are more like Zn than Ca according to RS, and B and Al are more like Sc than Ga according to RS, so if you adopted chemical similarity as the criterion for both you'd end up with the atomic number sequence getting lost because B came before Be!)
 * I might also add that if one really believed in chemistry as the basis, then Th, Pa, and U would not be under Ce, Pr, and Nd, but under Hf, Ta, and W. That is where they were before WWII as well, because the valences match better that way, and the physics was not then well-understood. Nd cannot be hexavalent; U can and often is.

Nevertheless, the actinoids are not a “second rare-earth set.” Instead, the early actinoids are not limited to +3 common oxidation states but have highest oxidation states paralleling their transition metal analogs. This “double-allegiance” was acknowledged by Darlene Hoffman and D. M. Lee in their discussion of the chemistry of the heaviest elements (27). This resemblance of the early actinoids to corresponding transition metals has been named the Actinoid Relationship, which “relates to similarities in chemical formulas and chemical properties between early members of the actinoid series and the corresponding members of the transition metal series” (27). The Actinoid Relationship provides such rich chemistry of the early actinoids. In fact, there has been a continuing interest in whether higher oxidation states can be found farther along the actinoids, even up to plutonium(VIII) (28). Thus, we should not give students the impression that the lanthanoids and actinoids are matching series. Instead, the answer to whether the early actinoids are actinoids or transition metals is: “both”!
 * Of course, nobody believes that anymore: the physical fact that the f orbitals are involved here has trumped all chemical ideas here. No doubt to a traditional chemist of Mendeleev's time, the placement of U under Nd would look deranged: now it has conquered all. Chemistry has lost utterly to physics here. I suspect this paper of Scerri is of relevance here, given its header "The Actinides. A Case Where Spectroscopy Triumphs over Chemistry?" In any case, even if one tried to use chemistry here: one could not. Th, Pa, and U would be jammed in the same spaces as Rf, Db, and Sg which cannot be placed anywhere else but as eka-Hf, eka-Ta, and eka-W. Just another example of how chemistry simply doesn't work as a basis for PT placement: it forces you to break the law that elements appear in order of increasing atomic number. It has been abandoned for a very good reason.
 * I don't think the future is going to be particularly good for Sc-Y-La. We seem to be agreed that physics is not for it; whether chemistry is for it or not is somewhat disputed (Jensen thinks no, Restrepo thinks yes – but only for La, since data for Ac/Lr is thin on the ground), but chemistry has already been supplanted as a basis for the PT by physics long ago. So the dispute has almost become moot because according to the modern criterion of physics, Sc-Y-Lu is a shoo-in. After all, using chemistry as the basis means wrecking the atomic number order (B before Be, Al before Mg) and forcing two elements in the same place (U and Sg jammed together under W). Scerri has outright said in his recent workshop: "My money's on number two [Sc-Y-Lu] and I certainly will be arguing to IUPAC that they should adopt number two. ... The one I'm suggesting and other people on this committee have suggested is lutetium and lawrencium should be the official version." I doubt he'd say that if he didn't think the IUPAC committee he's chairing was on board with it, particularly since he's apparently writing an article summarising the results of their deliberations. So, I grant that I am an outsider to these deliberations, but I would be very, very surprised after hearing that if the results were not an IUPAC ruling in favour of -Lu-Lr. So I'd personally advise once again that you drop that idea and return to Sc-Y-Lu as you supported before 2016, before I to my regret convinced you otherwise based on wrong data (in fact, what I told you were the condensed-phase configurations of the Ln and An, that convinced you to -La-Ac, were not the correct such configurations; my bad). RS indicate that for the wider chemical community, periodic tables based on chemistry rather than electronic structure are a historical thing. We've advanced since then, and the reason why -Lu-Lr which is consistent with this idea hasn't gotten up seems to be just inertia or people unaware of the matter. It seems to be going as DA said: "Eventually Lu will replace La. That's just a matter of time needed for obsolete understandings to die." But it's up to you, of course. You can have whatever view you like, and so can I; the point is just to work together to let WP accurately display what's in the literature, and update things only with it. As I said, I'm totally OK with WP only getting updated to the -Lu-Lr format when an IUPAC ruling comes along. Double sharp (talk) 00:38, 7 December 2020 (UTC)
 * , if you are going to leave aside any action on group 3 until there is movement from IUPAC, then why put a lot of time into discussion of the topic, especially in light of WP:NOTAFORUM? EdChem (talk) 12:31, 7 December 2020 (UTC)
 * Well, I did it mostly because it seemed to me that was using partly the same sort of argumentation as he was for Sc-Y-La as he's now using for B-Al-Sc. In this case the RS situation is a lot clearer (to a first approximation no one seems to take B-Al-Sc seriously), so I thought this might serve well at explaining to Sandbh why I feel RS was not supporting his approach. Maybe it skirts to NOTAFORUM, but my judgement might be a bit compromised by a lingering guilt: before 2016 Sandbh was a Lu supporter and I convinced him away from it based on what I now with better knowledge think was wrong data and some bad arguments. Since in 2018 DA explained to us more about this and I realised that what I was saying wasn't really correct, but Sandbh wasn't convinced, I find it a bit difficult to suppress the lingering guilt and it might impact what I say here more than it really should. Anyway, it seems clear now from his reply that it didn't really work at explaining to Sandbh why I feel this use of correlation to look for a linear trend isn't justified per Jensen. So I decided after my second post today (see the bottom) to stop posting about it here. In other words: I agree with you now, and had just decided to stop. ^_^
 * Anyway, the point of me starting this subthread was just to ask you if you thought DA's article could serve as an RS for including the argument "B-Al-Sc vs B-Al-Ga is analogous to Sc-Y-La vs Sc-Y-Lu, and since B-Al-Ga is standard that makes Sc-Y-Lu seem more reasonable by analogy" in the ancillary article on group 3 element (not the main periodic table article): it was a source reliability for WP question. Unfortunately I guess it is now buried. I'd still like an opinion but it's not really that important; there are lots of other arguments which have already been listed in that article and are cleanly cited to RS, so having or not having this one isn't that important.
 * I am afraid to say I have some doubts now about whether trying to continue the collaboration to improve periodic table to real FA status is really going to be good for us as a project. That's because of recent edits by Sandbh coupled with this discussion. The two I have in mind are this and this. Please correct me if I am wrong but I have serious doubts if these two trends (category analogies and isodiagonality respectively) are that common in the introductory chemistry literature to deserve this kind of high-profile mention. They are things that Sandbh has discussed in his recent journal-published articles of 2020 (this year), and he cites himself for them, but I was not aware of them being common trends to discuss among chemists rather than just appearing in a few places (isodiagonality is common for a few specific pairs Li-Mg, Be-Al, B-Si, I know, but I'm not aware of it being commonly applied to things like Ca-Y-Ce as Sandbh gives in his illustration). Again, please correct me if I'm wrong and these actually are common to discuss in the ways the figures show, but as it stands I'm not really comfortable with these additions. I'm especially not happy with the figure for isodiagonality, which he took from his own article on group 3, has aluminium moved to over scandium (as he's discussing here) against common chemical practice, contains an uncited claim about "predominantly ionic" chemistry that to my understanding is against typical freshman knowledge and not really often used to describe chemistry (surely that always depends on the counteranion; sure, NaCl is predominantly ionic, but one can hardly say it for Na3Bi, say; surely one can't exactly say the chemistry of Ce is "predominantly ionic" without making it clear what elements we're comparing it against, because with most elements Ce probably forms some sort of intermetallic phase anyway; and isn't the more common way to express this electropositivity?), and finally at the end throws in what seems to be a gratuitous Sc-Y-La argument from the atomic number triad (21+57)/2 = 39 that isn't exactly relevant to this topic (isodiagonality) IMHO. I cannot say I'm happy with this, but perhaps my opinion is coloured by the fact that I'm generally against Sandbh's research conclusions on periodicity, so your opinion might help as a third one. But in general I feel that (1) we just have very different visions for how the article should look, and (2) it doesn't look very good if I'm always the one who pounces and points out something I feel is problematic when it's the same editor each time, and (3) it seems likely that in the absence of a categorical statement of Sc-Y-Lu from IUPAC this and other issues will be hard to really provide a ruling on because this La vs Lu issue keeps sneaking in in the form of these sorts of periodic table figures and illustrations. I'd be fine with it if it was just the default PTs being shown as La, without any text trying to "prove" it correct that way when the debate is not of importance; but when we have a PT shown like that, and specially moved to have Al atop Sc in something else that the editor in question is championing, with a justification for Sc-Y-La presented in the middle of a long caption in a section not related to the dispute, it somehow leaves a bad taste in my mouth as if the argument was being given too much weight and it spilled outside its own section. So I suspect it might be better for me to just leave this article be and let Sandbh do as he wants with it, when there are a bunch of much less controversial elements in the middle of the table for which good reliable sources are not hard to come by and for which categorisation is completely undisputed (I admit that probably means just some nondescript transition metals...). After all, if what he's done is really problematic, it's better for someone less involved in the dispute to say it, and if it isn't, then I'd better also hold my peace; and I think writing this neutrally is going to be so so so much easier once IUPAC says something and I don't trust myself to be able to do it perfectly either till then. For instance, I guess the GAs on vanadium and palladium ought to somehow be restored by an article improvement drive. Most of the remaining not-even-GA articles are very meaty (e.g. gold, sulfur) or hard-to-research-because-historical (radium) topics, but taking a GA somewhere and FA-ing it might be fun. All that might very well be far less controversial than trying to achieve a shared vision for the periodic table article as the situation stands... Double sharp (talk) 13:04, 7 December 2020 (UTC)
 * , I mentioned WP:NOTAFORUM to both you and Sandbh as there are some pretty long posts appearing and they are not clearly directed to article-space issues. It was meant as a reminder and an invitation for reflection, so please don't take it as more than was intended.
 * You asked if I thought that DA's article could serve as an RS for including the argument "B-Al-Sc vs B-Al-Ga is analogous to Sc-Y-La vs Sc-Y-Lu, and since B-Al-Ga is standard that makes Sc-Y-Lu seem more reasonable by analogy" in the ancillary article on group 3 element (not the main periodic table article). Leaving aside reliability, a more pertinent question is whether DA's article makes that argument, or if it has been used by other RS to make that argument... because if the argument is built from DA's article but the article does not actually make the argument directly then it would be OR or SYNTH to use it unless another RS has already advanced the argument using DA's work.
 * Sandbh citing his own work does not raise OR or SYNTH concerns so long as the work cited does support what it is used to support. That doesn't mean that its inclusion is appropriate based on DUE, however.  You link to two additions and I can see why you have concerns based on a first look.  Starting a talk page discussion and neutrally inviting comment from the Chemistry WikiProject might be a way forward.  I tried to encourage work on the PT article to look at the non-controversial parts that can be improved, but it does not appear that my approach is developing along the lines that I envisaged.  EdChem (talk) 13:52, 7 December 2020 (UTC)
 * Right, I understand what you intended now. ^_^
 * Well, DA's article says (relevant bit in translation) However, a detailed study of the regularities of changes in the chemical and physical properties of elements [5 - 6] in the series Ba - Hf and Ra - Rf [7], as well as a comparative study of the Sc-Y-La-Ac and Sc-Y-Lu-Lr groups [8 ] and drawing analogies with the B-Al-Sc-Y and B-Al-Ga-In groups, respectively, showed that it is more expedient, more practical, and more logical to attribute the families La - Yb and Ac - No to f-elements [3]. I guess it's pretty clear what he means by it, but it's maybe expressed in a way that requires the leader to draw the line to the final step. Do you think it'd be SYNTH to say what I said? I'm not sure, that's why I'm asking. ^_^
 * Oh yes, I agree that Sandbh citing his own work is totally OK wrt OR and SYNTH. It is cited to an RS and it supports what it claims to. My concern was just about DUE, as mentioned. But I think I'll let you opine first, because as I mentioned I may have some unconscious biases. Double sharp (talk) 15:11, 7 December 2020 (UTC)

Group 2 as Be-Mg-Ca-Sr-Ba-Ra or Be-Mg-Zn-Cd-Hg
I mapped the trend-lines going down these two options using the same eight properties I used for group 3 and 13.

The average goodness-of-fit values were:


 * Be-Mg-Ca-Sr-Ba = 0.84
 * Be-Mg-Zn-Cd-Hg = 0.66

This is better than the figures for group 3 and group 13, which I did not find surprising. It is consistent with 16 of the 17 B-Al-Sc-Y-La-Ac tables. Only Hubbard shows Be-Mg-Zn-Cd-Hg. It is consistent with C & W (1999): "The Group 12 elements differ markedly from those in Group 2 in nearly all aspects except having II as their only important oxidation state." (p. 599).

I haven't double checked my figures, so maybe I got something wrong.

The quality of DA's source 7 is too low to be legible.

I agree many people cite and refer to Jensen as the authority for the Sc-Y-Lu table. The only people I know who have looked critically at Jensen rather than superficially, as I did at first, are Scerri, myself and you. And I only did so with the help of Scerri. In our joint IUPAC paper we showed how nearly all (as I recall) of Jensen's arguments don't hold water.

I feel there is too much confidence being placed in physics, as opposed to chemistry, given the nature of IUPAC. I sort of addressed this is in my article on the location and composition of Group 3:


 * "Nor does IUPAC prescribe which elements, for example hydrogen and helium, belong in which group. All they have done is approve collective names for like elements e.g. chalcogens (oxygen, selenium, selenium, tellurium, polonium) and noble gases (helium, neon, argon, krypton, xenon, radon) (Connelly et al. 2005, p. 61). For example, if helium is placed over beryllium then helium still belongs to the noble gases.


 * Why then would IUPAC be interested in prescribing which element goes under yttrium when, depending on the circumstances, either lanthanum or lutetium could be suitable? For example, in electronic structure terms, lanthanum has the advantage of incumbency, since the 5d1 electron appears for the first time in its structure whereas it appears for the third time in lutetium, having already made a brief appearance in gadolinium (Trifonov 1970, pp. 201–202). On the other hand, lutetium may be a better choice if crystallography is the focus, since the crystalline structures of scandium, yttrium, and lutetium are all hexagonal close packed whereas that of lanthanum is double hexagonal close packed. As another example of the flexibility of group assignments, Group 14 in the Earth Scientist’s periodic table is composed of carbon, silicon, titanium, zirconium, and hafnium rather than the standard set of carbon, silicon, germanium, tin, and lead (Railsback 2018).


 * The current confusion for students and teachers as to why the IUPAC table has a 15-element wide f-block, whereas other authors show lanthanum or lutetium under yttrium and an associated 14-element wide f-block, arises from a lack of IUPAC guidance along the lines of the previous paragraph, rather than because it has not formed a view as to the composition of Group 3.


 * Such guidance could read, for example, as follows:


 * Draft IUPAC Red Book guidance


 * ELEMENTS IN THE PERIODIC TABLE


 * The periodic table on the insider cover is the form agreed and used within the IUPAC, rather than being IUPAC recommended or approved. In this instance, the lanthanoids are shown as a 15-element wide series in light of their chemical similarities.


 * Different forms of the periodic table may be more or less appropriate in particular contexts. For example, a 14-element wide lanthanoid series may be more appropriate to better bring out the concept of an f-block. Such a series could start with, for example, lanthanum or cerium depending on the context.


 * IUPAC does not recommend or approve any particular format of periodic table or system, nor does it mandate the composition of Groups."

An IUPAC format table leaves open the implications associated with each group 3 option. [1]

It says the chemistry of La to Lu is the primary item of interest, with group 3 being able to be treated flexibly as Sc-Y-La-Ac [2] or Sc-Y-Lu-Lr. [3]

If consistency is also important, you can move B-Al over Sc-Y-La-Ac if the focus is on chemistry [4]; or, with Sc-Y-Lu-Lr, move He over Be, if the focus is on electronic structure [5].

So the options are: Basis Group 2 Group 3          Group 13  Group 18  Chem  Elec [1] Lu form (consistent) He-Ra    Sc-Y-Lu-Lr       B-Tl      Ne-Rn Elec    Elec             Elec      Elec       0     4 - [2] Traditional Lu form  Be-Ra    Sc-Y-Lu-Lr       B-Tl      He-Rn Chem    Elec             Elec      Chem       2     2 - [3] IUPAC form           Be-Ra    Sc-Y-*-**        B-Tl      He-Rn Chem    ? Elec     Chem       2½    1½ - [4] Traditional La form  Be-Ra    Sc-Y-La-Ac       B-Tl      He-Rn Chem    Chem             Elec      Chem       3     1 - [5] La form (consistent) Be-Ra    B-Al-Sc-Y-La-Ac  Ga-Tl     He-Rn Chem    Chem             Chem      Chem       4     0 -

I note that:
 * the Lu [1] form with He over Be is perfectly unambiguous, and highly symmetrical.
 * the traditional Lu [2] form is a fence-sitter in that it cannot make up its mind if it is Arthur or Martha;
 * the IUPAC form is not really sitting on the fence i.e. it has a chemistry bias: chem. (2½) elec. (1½);
 * the most common, traditional [4] form is inconsistent—not so surprising given the nature of the subject matter i.e. chemistry; and
 * the La [5] form is unambiguous, and requires three split blocks.


 * Above table and notes updated. --- Sandbh (talk) 00:57, 11 December 2020 (UTC)

I know of nothing that would prompt IUPAC to make a decision on the composition of group 3. Recall that the commissioning of the IUPAC project on group 3 does not necessarily mean the recommendation of the project will be accepted by IUPAC.

Of course, I may be be mistaken, and I'm looking forward to what eventually comes out of the IUPAC Group 3 project, and its reception, by the chemistry community, and IUPAC and its world-wide affiliated bodies, as much as you are. --- Sandbh (talk) 06:17, 7 December 2020 (UTC)
 * I think you're still asking for goodness-of-fit trend values comparing against straight lines. But the point is that the trend should not be a straight line. If you do that, then it's like staring at exact parabolas and shouting "there's no trend! there's no trend!". Of course there is, but if one insists on the linear dogma, then it becomes simply impossible to see. Just try and plot the goodness-of-fit value for the following data:


 * The correlation coefficient happens to be exactly zero. But there is a clear relationship between x and y here. An exact one, in fact. I think it's pretty obvious. ^_^
 * There's two kinds of trends in the table: Be-Mg-Ca-Sr-Ba-Ra is the "straight line" kind, Be-Mg-Zn-Cd-Hg-Cn is the "kinked" block-insertion kind. The latter is natural everywhere outside the s block, which is why for those trends a lower goodness-of-fit value is actually better: it indicates we have the right type of trend. But if you insist on those, then I suggest plotting goodness-of-fit values for C-Si-Ti-Zr-Ce-Th (all four electrons above a noble gas) vs C-Si-Ti-Zr-Hf-Rf (Earth scientist's periodic table) vs C-Si-Ge-Sn-Pb-Fl (standard). The former will probably be better. And the same for N-P-V-Nb-Pr-Pa vs N-P-V-Nb-Ta-Db vs N-P-As-Sb-Bi-Mc. Are those forgotten chemistry also? Maybe in a way, they are the A vs B secondary analogies that Jensen agreed were kind of "lost knowledge" today. But that doesn't mean that a true chemical periodic table "should" really contain C-Si-Ti or N-P-V (unless we talk about the 8 column form still common in Russia, in which case yes, C-Si-Ti is there, but so is C-Si-Ge simultaneously); apart from the Earth scientist's periodic table for C-Si-Ti, nobody thinks either of these belong in the 18 or 32 column forms. Also maybe try Ti-Zr-Hf-Th vs Ti-Zr-Hf-Rf, V-Nb-Ta-Pa vs V-Nb-Ta-Db, and Cr-Mo-W-U vs Cr-Mo-W-Sg. Thanks to the actinide contraction probably the former would look better in best-of-fit values. ;)
 * Indeed, commissioning a project does not mean the recommendations will be accepted. But doing so would defeat the entire purpose of it, which is that students and instructors are puzzled by the differences. Many IUPAC projects do indeed result in recommendations, even for such basic issues as those that appear on the PT. Such has happened quite recently for the project on naming new elements and the project on discovery criteria for new elements. Well, they've since both progressed to recommendations and have become the new rules. The first was already in effect for why Ts and Og have -ine and -on suffixes; the second is already also in effect, it just hasn't been applied yet because no serious claims for elements 119 and up have come. Don't see any reason why this one will not progress similarly when Scerri already did write in Chemistry International back in 2012 about how IUPAC should be recommending one form of group 3 and indeed exactly the -Lu-Lr one. That's a IUPAC magazine, so IUPAC knows what his stand is. I also suspect IUPAC must have learned its lesson from the 1988 -*-** compromise which pleased nobody. After that debacle, they probably know that if they don't make a ruling, the issue will just end up being kicked down the can for another few decades before flaring up again just like it did this time (1988 compromise, 2015 opening of IUPAC project). Besides, predicted chemical properties meant absolutely nothing to IUPAC when Ts and Og were recommended to be named with -ine and -on; the symmetry of keeping the whole column the same won the day. So in that case chemistry was shot down in favour of symmetry which came from physics (the predicted electron configurations match F-Cl-Br-I-At and Ne-Ar-Kr-Xe-Rn, even if chemical behaviour probably totally doesn't). So if IUPAC was fine going for physics rather than chemistry then, it doesn't seem that implausible that they'd do it again for -Lu-Lr. Again, I'm not saying I know this is how it will work out (I don't); just saying that the situation and what Scerri has said makes it seem to me that we should not be too surprised if the end result is a IUPAC recommendation of -Lu-Lr. Of course that's not the only thing that might happen; I just think it seems likelier than I had imagined previously.
 * I am indeed in the camp that says that Lu is the only correct form and that the La form is an old historical misunderstanding, as you know. That's a stronger view than your support of La, I see, which seems to account for the possibility of different forms being more or less valid in different contexts. (This I see as secondary relationships, like Jensen.) So let me just state this honestly: I've been trying to fix the consequence of my faulty understanding in 2016. Then I had a faulty understanding, and used what I now know were faulty data and faulty arguments to influence you in User talk:Double sharp/Archive 11, probably you would not be supporting Lu now. If I knew then what I knew now, you'd never have been influenced by me to La, and probably you'd be still supporting Lu unless somebody else made La arguments to you that convinced you. You may recall that I totally withdrew my support for that old IUPAC submission we worked together on and now think that actually Jensen's arguments do hold a great deal of water after all. For me to do that when it went that far says something, probably. If you don't think that's enough for you to reevaluate your position, and you want to continue supporting La, that's your choice and I respect it. So if this post of mine doesn't convince you, then that's alright and my stance returns to "let's wait for IUPAC". Scerri's the chair of that project, he understands quite a bit about these issues, and I am sure he will come up with something interesting and that is worthy of following on WP as a IUPAC recommendation. And then we'll see. Till then, there's no point discussing it further. But you may still go ahead if you like and post interesting things you see here like your B-Al-Sc support, of course. I just probably won't reply, both because it'll go in circles, and because you probably already know what I think and what I see the RS as saying. Basically, my viewpoint is settled enough and tested to my personal satisfaction, and I don't think further discussion of this point is going to change our disagreement: only a IUPAC ruling might. To that end also, I think that if you're not convinced by what I've pointed out here (plotting goodness of fit will probably also support C-Si-Ti which G&E also show), then there's no point in me discussing the PT article until IUPAC comes around, as our visions for it are going to differ too much for a very meaningful collaboration. Maybe I'll just find an element article that is far away from all PT placement questions and also PT categorisation questions. That probably means a nondescript transition metal. Most are GA already, but some could use further improvement pretty easily. So maybe I'll channel whatever free time I have (not too much from tomorrow onwards, but we'll see) into that and wait for IUPAC to say something about group 3 before touching the PT article. There is no deadline after all.
 * And in any case, speaking personally: I'm already happy that you see my He-Be + Sc-Y-Lu personal view as a "perfectly unambiguous" one. I recognise it's pretty much on the opposite end of the spectrum to your preferred one, so it really pleases me that you see something in it. Double sharp (talk) 09:20, 7 December 2020 (UTC)

Basis for goodness of fit

 * There are no straight lines, as understand it. The goodness-of-fit trend values are for polynomial lines.

Oh yes, He-Be + Sc-Y-Lu is wonderful in its clarity! As far as IUPAC is concerned, they may well endorse it, on these grounds! Consider what Jensen wrote in JChemEd (2008):


 * "Above all, such demands should be tempered by the sobering recollection that IUPAC is the organization that brought us density in units of kg/m3, 4πε0 in the denominator of Coulomb's law, and the finger-count labels 1–18 in the periodic table."

If IUPAC introduced the dumbed down 1−18 group numbering system, it would seem to be within the realm of possibility that they could recommended at least Sc-Y-Lu on "simplicity" grounds, and just put up with He over Ne.

It's ironic that Jensen is on the IUPAC Group 3 project and will presumably support Lu on, inter alia, symmetry grounds even though he disparagingly referred to the 1−18 numbering system, and has rallied against the abuse of symmetry considerations when it comes to organising the periodic table! We know he won't support Lu on the grounds of triads per Scerri, since Jensen has disparaged the relevance of triads. Since Scerri referred to Jensen's argument as being too selection, I don't know on what basis Jensen would support Lu. Lavelle is also on the project team, and his support for La and opposition to Lu is well known. And Restrepo most recently supported La.

On students and instructors being puzzled by the differences, that is the fault of IUAPC in not providing sufficient guidance in e.g. the Red Book, along the lines I suggested.

On group 4 v 14, I looked at this, and the average goodness of fit values are about the same. This being so, I'd leave group 14 as C−Pb. Sandbh (talk) 01:39, 8 December 2020 (UTC)
 * Over here I'll just answer the question about Jensen as it is a question that can be answered from publicly available RS without editorialising on my part: presumably Jensen would support Lu for the reason he publicly expressed in 2017, which is also the reason I support it: a consideration of their available excited-state configurations (2, 4, 12)... reveals that both La and Ac have low-lying empty f-orbitals and that these are implicated in several low-lying excited configurations, whereas Lu and Lr have no available empty f-orbitals. This implies, in turn, that both La and Ac – like 25% of the other d- and f-block elements (including their nearest neighbor Th) – are best viewed as having anomalous ground-state valence configurations, which, in their case, take the place of an idealized (n-2)f1ns2 valence configuration. Since such an interpretation is not possible for Lu/Lr, it pretty much determines that this pair should be assigned to the d-block whereas the La/Ac pair should be assigned to the beginning of the f-block. He cites among other sources 10.1364/JOSA.61.001112. I don't think Scerri referred to this argument when he mentioned that Jensen's arguments were too selective, although I may be wrong. Regarding the other points, I'll write on your talk page. Double sharp (talk) 01:49, 8 December 2020 (UTC)

Jensen's 2017 paper
, you've honed in on a key consideration.

As I read it, Jensen's 2017 unpublished and non-peer reviewed paper is a mixed bag:

On the one hand he wrote:


 * "To tip the scales in favor of one pair of elements or the other requires instead a consideration of the additional chemical and physical properties discussed in my original paper of 1982 and, above all, a consideration of their available excited-state configurations.

This is bad. We showed that nearly all of the chemical and physical properties he referred to in his 1982 paper were inconclusive. Not to mention Scerri's view.

To his credit, Jensen brilliantly goes on to write:


 * "Though there are many misconceptions concerning the nature and function of the periodic law and table, perhaps the most prevalent among modern chemists is the belief that the periodic table is nothing more than an electron configuration table. While there is certainly a significant correlation between electron configurations and chemical periodicity, the correlation is, as already noted, far from perfect. The increasing prevalence of irregular configurations among the d- and fblock elements, the increasing lack of correlation between minor irregularities in these configurations and actual chemical behavior, and the ever present empirical question of how to properly divide an atom’s configuration into the chemically relevant categories of valence versus core, all require a careful balancing of both chemical and physical evidence rather than an appeal to authority and a naive, and apparently arbitrary, freshman chemistry application of spectroscopic atomic ground states."

Exactly. The PT is not a simple electron configuration block based table. Electron configuration considerations, in isolation, do not tip the balance. Sandbh (talk) 05:59, 8 December 2020 (UTC)


 * PS: Of course, they can tip the balance if that is the context of interest, in which case you have the beautiful He-Be + Sc-Y-Lu-Lr table. And that is a perfectly fine way to approach it, among others. Sandbh (talk) 06:03, 8 December 2020 (UTC)