User:Double sharp/Group 3 sources

Can someone find ? I should try to find this too.

A list of sources that consider group 3 as Sc-Y-Lu-Lr. (Disagreeing sources are in sections below. Maybe it should become a timeline.)

Regarding software: CrystalMaker agrees with this. :)

A few articles have been added because they are cited in the literature as supporting Lu under Y, but I have been unable to track them down. Not that it really matters because even without them the count is overwhelming: we literally crossed 100. As is the evidence, frankly.

To read:

Most of these are primary-source papers (gotta organise it better). I have indented sources I have seen mention of but have not been able to access, as well as those whose reliability is a bit doubtful or who don't give explanations.

Two early authors, Bassett (1892) and Werner (1905), should be mentioned before the list proper: at this point Lu had not been discovered, but they both present a table with no split in the transition metals and with La clearly placed in a different group from Sc and Y. (And yes, it's the coordination chemistry Werner.) A pretty much completely modern form of the table (except for helium, and moving the s-block to the right) appears in Janet (1928).

BTW it seems that most people working on the theoretical-physics basis of the periodic system take Sc-Y-Lu as basically a done deal and given, e.g. (A. I. Fet), Thyssen and Ceulemanns' Shattered Symmetry. (Though helium as an s-element seems to be taken much more seriously by such physicists than by chemists. The two issues tend to be different, in the sense that Sc-Y-Lu has often been argued for on the grounds of empirical chemistry as well as by physics. He-Be is a lot harder to argue by empirical chemistry, though it's not impossible.)
 * 1) Bury (1921) Lutecium, the rare earth of highest atomic weight, will have the structure (2,8,18,32,8,3); the fourth shell is now complete, and it will be a normal element. Lutecium, rather than lanthanum (2,8,18,18,8,3), resembles scandium (2,8,8,3) and yttrium (2,8,18,8,3), just as, in the first long period, zinc rather than calcium resembles glucinum and magnesium.
 * 2) Shemyakin (1932) Zh. Obshch. Khim. 2, 62 (1932) (Russian)
 * 3) Landau and Lifshitz (1948, Russian; 1958, English) In books of chemistry, lutetium is usually placed in the rare-earth elements. This, however, is incorrect, since the 4f shell is complete in lutetium; it must therefore be placed in the platinum group. (Strictly speaking, only argues that Lu is not an f-element, and does not argue that La is one. But used as a springboard by many later sources to continue on, and referred to by Scerri's history of the dispute as significant.)
 * 4) *Finke (1949), . Die Lanthaniden 57–70, sowie die Actiniden 89–102 erhalten endlich das ihnen zustehende Wohnrecht. Im übrigen ist alles ungeändert. Alas he does not deign to give an explanation.
 * 5) Ramirez Torres (1955), (mostly arguing for considering H+He as a  "1s" block split from everything else; however, puts f-block as La-Yb, notes that La lacks a differentiating f-electron, but writes "The position of lanthanum is based on its chemical properties" anyway).
 * 6) *Dockx (1959), Théorie fondamentale du Système périodique des Éléments (p. 82)
 * 7) Villar (1959), (Spanish; where can I get the fulltext? The abstract says Lu already)
 * 8) Hamilton and Jensen (1963), However, for n=3 we have Sc, Y, and Lu not superconducting (nor magnetic) while La is superconducting. Placing La next to Hf in the 5d row was merely a convention. If, following Landau and Lifshitz, we put Lu next to Hf, then all transition metals with n=3 are not superconducting (see Table I).
 * 9) Villar (1964), The study of emmision spectra of the actinide elements has proved that their electronic structures are characterized by the successive additions of electrons to an inner 5f shell. Therefore, the actinide elements constitute an f-transition series to whom correspond a special place in the Periodic Chart, outside of the d-transition series where the actinide elements are often situated. For this reason there would be necessary the following modifications of the Periodic Chart: the places that actually occupies the lanthanide elements and the actinide elements would be occupied only by lutetium and lawrentium that would be not f-transition elements but d-transition elements.
 * 10) Hamilton (1965), In a widely-used version of the periodic table, lanthanum is classed in the same column with scandium and yttrium. However, in a number of respects lutetium resembles scandium and yttrium more closely than lanthanum does. Hence the periodic table should be modified so that scandium, yttrium, and lutetium are in the same column.
 * 11) Luder (1967), Actually, because the energy levels are so close, from a chemical standpoint the two current apparent exceptions [La and Gd] could just as well be considered as normal. Such is the assumption of this paper: none of the atoms from La to Yb has a 5d electron. Thus the differentiating electron for each of them is an f electron, and they are all lanthanons. Consequently, the differentiating electron for Lu is the first 5d electron, and Lu is the first member of the third row of related metals.
 * 12) Merz and Ulmer (1967), This X-ray spectroscopic result shows clearly, that Lu, but not La, has the typical behavior of a transition metal in the structure of its conduction band. Arrangements of the periodic table, which ascribe Lu to the transition metals and not to the lanthanides, are therefore favored by this isochromat spectroscopic investigation.
 * 13) Chadwick et al. (1968) - Kurchatovium [rutherfordium] has been heralded as the first member of the fourth transition series: if, however, an essential feature of an nth-row transition element is that the (n + 1)f-shell (if there is one) is full throughout its chemistry, then lutetium is really the first member of the third transition series, and lawrencium (if the 5g-shell does not intervene) is the first member of the fourth.
 * 14) Chistyakov (1968) Zh. Obshch. Khim. 38, 209 (1968) (Russian) Mentions lanthanide contraction as the reason (absent in La, but affecting Lu and subsequent d-elements). Also cites five other articles for the Sc-Y-Lu conclusion, which I haven't found yet: Chistyakov 1964 and 1966 (same author), Shemyakin 1932, El'yashevich (Spectra of the Rare Earths 1953, Atomic and Molecular Spectroscopy 1962). I didn't find such a statement in El'yashevich, but he does note that Lu does not have a rare-earth-like spectrum (presumably meaning that 4f is never ionised), whereas La "is not yet fully typical" of rare earths but does involve 4f levels.
 * 15) *Seel (1969), Bild der Wissenschaft 6, 45
 * 16) Matthias (1969)
 * 17) *Chistyakov (1970)
 * 18) Gol'danskii (1970), . In Figure 4 the whole Mendeleev table is given with allocation of eighteen g elements of the eighth and ninth periods (octadecanides), fourteen f elements of the sixth, seventh, eighth, and ninth periods, and the "inserted decades" of d elements in the 4-9 periods. The (n + l) sequence rule does not reproduce individual irregularities in the sequence of filling of 5d and 4f shells among lanthanides or 6d and 5f shells among actinides (especially essential from thorium to neptunium). However, from what is put in the table in Figure 4 the fact clearly stands out that the f elements of the sixth and seventh periods adjoin the electronic analogs of scandium and yttrium—lutecium and lawrencium—(EK = 17) with the configuration of outer shells of ns2 (n − 1)d and the main valency equal to three.
 * 19) Luder (1970). I can't possibly give a quote, most of this article is treating the situation at length!
 * 20) Wittig (1973), We are convinced that La is a 4f-band metal. It belongs, therefore, to the rare earth group. The pressure induced superconductivity of Lu [35], revealing a Tc which fits very well into the general pattern of the transition temperatures of the d-period metals, now strengthens appreciably our confidence that Lu is the proper element below Y in the periodic table. There is apparently no interference by the closed 4f shell which proves to be without influence on the metallic properties Consequently, Lu should now be removed from the rare earth group and accommodated below Y in the 5d transition period in just the same way as one is used placing the next elements to the right (Hf, Ta ...) below the corresponding 4d-period metals.
 * 21) Jørgensen (1973), – "lutetium (which can most conveniently be considered as the first member of the 5d series")
 * 22) Byakov, Kulakov, Rumer, Fet (1977). Argues not only for Sc-Y-Lu but also for H-Li and He-Be. Note that in the table II hydrogen and helium are included, correspondingly, into subgroups of alkali metals and alkaline-earths metals, but not into these of halogens and rare gases, as accepted in most modern tables. Then, in contrast to the subgroup (IIIb), Sc - Y - La - Ac, of the classical Mendeleev table, according to table II immediate chemical analogues of Sc and Y are Lu and Lr, and not La and Ac. Therefore, the first and the second groups of rare-earths metals begin corr. with lanthanum and actinium, and not with cerium and thorium, whereas lutetium and laurentium fall out of these families.
 * 23) Desclaux and Fricke (1980)
 * 24) *This is one of the old predictions of Lr's anomalous configuration. The authors write On the other hand, the increase in the weight of the s2p configuration is only continuous for the elements Y, Lu and Lr on one side and for La and Ac on the other. This difference of behaviour between these two groups of atoms is also reflected in the expectation  values of the analogous wave-functions both for the ground state [2] and the excited ones considered here. The physical reason for this is the SCF influence of the underlying shells. While 4f and 5f shells are just filled at Lu and Lr, La and Ac are at the beginning of the f-series. Y is very comparable with Lu and Lr because the 3d shell, which has a very similar influence, is just filled. This, of course, is not valid for Sc, as there is no underlying d shell. Consequently, only the elements Y and Lu can directly be compared with Lr.
 * 25) Cowan (1981)
 * 26) *Seel (1981), Chemie Labor Betr. 32, 152 [written as 1961 by Fluck]
 * 27) Jensen (1982; important for collecting many arguments)
 * 28) Cherkesov (1984) Calculations, accomplished using the method of hyperbola, turned out to be possible only in the case of introduction of nuclear-proton criteria for separation of analog elements and putting them down into the third group of short dyad form of the Periodic system.
 * 29) Johnson (1984), the entirety of pages 54 and 55
 * 30) Holden (1985; it was supposed to be presented at the 33rd IUPAC General Assembly in Lyon) The strongest argument in my estimation is that of Landau. In both the case of lutetium and lawrencium, the group elements have 4f14 and 5f14 configurations, respectively. Neither lanthanum nor actinium would be appropriate group members as a result.
 * 31) Eggers et al., (1986), (German)
 * 32) Eggers et al., (1986), (German, a little bit ambiguous; calls Lu a pseudo-5d-element, but also says Yb is the "last" 4f element with quotes, and cites Jensen)
 * 33) Fluck and Rumpf (1986), (German)
 * 34) Jensen (1986)
 * 35) Brodersen and Seel (1986), (German)
 * 36) *Seel (1987), MNU 40, 304
 * 37) Fluck (1988, a IUPAC report)
 * 38) *"According to the electron configurations of the elements, the scandium group consists of the elements Sc, Y, Lu, Lr. ... Based upon their electronic  configurations and their chemical and physical properties, the elements La to Yb and Ac to No should be inserted between barium and lutetium and between radium and lawrencium or  for practical reasons be listed at the bottom of the table."
 * 39) *Although also mentions that a compromise was chosen for the next Red Book. This is the Sc-Y-*-** form which "invents" the weird idea of a 15 element f block and one element (yttrium) sitting atop fifteen others. This is true in a "yes and no" way though: the next Red Book had a Sc-Y-*-** 18-column table, but a Sc-Y-Lu-Lr 32-column table. So, any way you look at it, Sc-Y-La-Ac has no IUPAC imprimatur, despite what Lavelle claimed in 2008.
 * 40) Bokii (1989),
 * 41) Eriksson et al. (1990), Suggestions that lawrencium is a simple sp metal are refuted. Instead it is found that this element is the first member of the 6d transition series, in analogy with the position of lutetium relative to the 5d series. This conclusion is confirmed by calculations of the cohesive energy for kurchatovium [rutherfordium].
 * 42) Scerri (1991),
 * 43) Scerri (1994),
 * 44) Heyes (1997–8; alternate link), a page from Oxford University's 1998 second-year inorganic chemistry course
 * 45) Kulsha (1999): available on the author's website, pages 1, 2, 3, 4 (Russian)
 * 46) Leach (1999-present), Chemogenesis (well, a bit arguable because it's self-published, but the author has non-self-published works in this area)
 * 47) Fang et al. (2000),
 * 48) Wulfsberg (2000), Inorganic Chemistry (this is a textbook, but quite unusually among textbooks, actually explains why it chooses to show it this way)
 * 49) Katz (2001, in the context of supporting Janet table)
 * 50) Lombardi and Davis (2002),
 * 51) Sastri et al. (2003)
 * 52) Dommelen (2004, 2007, 2008, 2010, 2011, etc.; hosted by FAMU-FSU College of Engineering, written by a retired prof)
 * 53) *Also, both lutetium and lawren­cium are ac­cord­ing to IU­PAC in­cluded in the f-block. That makes the f-block 15 columns wide in­stead of the 14 col­umn block shown at the bot­tom of fig­ure 5.8. Of course, that does not make any sense at all. The name f-block sup­pos­edly in­di­cates that an f-shell is be­ing filled. An f-shell holds 14 elec­trons, not 15. For lutetium, the f-shell is full and other shells have be­gun to fill. The same is, at the time of writ­ing, be­lieved to be true for lawren­cium. And while the first f-shell elec­trons for lan­thanum and ac­tinium get tem­porar­ily bumped to the d-shell, that is ob­vi­ously a mi­nor er­ror in the over­all logic of fill­ing the f-shell. (Ap­par­ently, there is a long-stand­ing con­tro­versy whether lan­thanum and ac­tinium or lutetium and lawren­cium should be in­cluded in the f-block. By com­pro­mis­ing and putting both in the f-block of their 2007 pe­ri­odic ta­ble, the IU­PAC got the worst of both worlds.)
 * 54) Kulsha (2004) (Russian – technically this is student work, but he later became a published computational chemist, so I think it's reliable enough; in any case the section on group 3 cites other work)
 * 55) Horovitz and Sârbu (2005),
 * 56) Wulfsberg (2006),
 * 57) Bent and Weinhold (2007) Such a Dobereiner-type argument can also be used to infer that element 71 (not 57) should appear below Sc and Y as the [5d1] element (18), in order that it forms a proper even-starting triad with Y (Z = 39) and Lr (Z = 103). Also argues for He-Be.
 * 58) Jensen (2008)
 * 59) Ouyang et al. (2008),
 * 60) Stewart (2008) The rationale for dividing the continuous sequence of elements into blocks is that they correspond with the subshells of electrons. Thus, by definition, the s-, p-, d-, and f-blocks should have respectively 2, 6, 10, and 14 elements. If the definition of membership of the f-block is that the differentiating electron of an element could be and is in an f-orbital, only La, Gd, Ac, and Th are exceptions, since neither Lu nor Lr could be distinguished by a 15th f-electron. A slightly weaker generalization is that the nth element in each series has n f-electrons, in which case Yb and No have to be the 14th and last. To this, Pa, U, Np, and Cm are exceptions, but even a generalization that applies to three quarters of the block is worth having. Lu fits into the d-block, as La does not, by sharing with Hf and so forth the lanthanide contraction and its ensuing chemical likeness with its homologue Y.
 * 61) Jensen (2009)
 * 62) *Marshall and Marshall (2009, historical series; not actually part of the list :D)
 * 63) Scerri (2009)
 * 64) Scerri (2009)
 * 65) Scerri (2009)
 * 66) Scerri (2009)
 * 67) Thyssen and Binnemans (2010)
 * 68) Emsley (2011; a popular book, but citing Jensen 1982 as the reason; correctly remarks following Jensen that the Sc-Y-La table mostly exists for "historical reasons"; and John Emsley was a chemist before turning to writing popularisations)
 * 69) Imyanitov (2011)
 * 70) *Saleh et al. (2011) As expected,13 all of the metal-ligand bond distances follow the periodic trend Sc-X < Y-X > Lu-X. This doesn't really argue for the placement, but takes it as a given.
 * 71) Scerri (2012; nicer pdf link)
 * 72) Scerri (2012)
 * 73) Scerri (2012, talk by him)
 * 74) Bhandari (2013)
 * 75) Keeler and Wothers (2013)
 * 76) Leach (2013)
 * 77) Nelson (2013)
 * 78) Scerri (2013)
 * 79) Settouti and Aourag (2014),
 * 80) Imyanitov (2014)
 * 81) Imyanitov (2015)
 * 82) Jensen (2015)
 * 83) Poliakoff (2015; Periodic Videos, a secondary RS, but he's an actual chemist)
 * 84) Scerri (2015)
 * 85) Settouti (2015) (French)
 * 86) de Sousa Filho et al. (2015)
 * 87) Deblonde and Abergel (2016)
 * 88) *"Although the name actinium would unambiguously make it the first member of the actinide family, Ac is still placed below Sc, Y and La in most periodic tables. Recent arguments tend to support its reassignment as an actual f-block member with irregular electronic configuration, whereas the current heaviest actinide, Lr, would be drafted to the d-block below Sc, Y and Lu."
 * 89) Ball (2017)
 * 90) Jensen (2017) 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 (2, 4, 12). This 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.
 * 91) Stewart (2017) There are good reasons for giving the place to Lu: it follows the lanthanoid contraction, so its relation to Y is as close as that of Hf to Zr; indeed it was discovered in yttria, while La was first found in ceria. An argument for making La the first element in the f block is that then, except in Gd, the pth element in the series has p f electrons.
 * 92) Imyanitov (2018)
 * 93) Scerri and Parsons (2018)
 * 94) Tsimmerman (2018)
 * 95) Labarca and González (2019) (Spanish)
 * 96) Pereira et al. (2019)
 * 97) *In terms of the displacement of the elements in the third group of the periodic Table, this work shows a thermodynamic similarity, at least for the fifth and sixth period, between Y and Lu much larger than that observed for Y and La.
 * 98) Scerri (2019)
 * 99) Scerri (2019)
 * 100) Scerri (2019)
 * 101) Scerri (2019)
 * 102) Scerri (2019, another talk)
 * 103) *Thyssen (2019, Christmas lecture; seems to follow the end of 2011 Thyssen and Binnemans)
 * 104) Tsimmerman and Boyce (2019)
 * 105) Alvárez (2020)
 * 106) Jean (2020)
 * 107) Maeno et al. (2020)
 * 108) Scerri (2020)
 * 109) Scerri (2020)
 * 110) Scerri (2020; a workshop)
 * 111) Scerri (2020; briefly mentions and cites Scerri and Parsons 2018 as the stuff that needs to be brought to bear on this)
 * 112) *Zhang (2020 – personal website of a chemist, so not quite as RS as the other ones, but it hardly makes a difference at this point)
 * 113) Ghosh (2021)
 * 114) *Leach (2021, praised by Scerri, but it's just a YouTube presentation.)
 * 115) Scerri (2021, lecture)
 * 116) Scerri (2021 IUPAC provisional report)
 * 117) Amézaga, García-Suárez, Fernández-Martínez (2022, see the footnote)
 * 118) *Baez (2022; this is John Baez's blog, so maybe not quite as RS as the others as he's not a chemist :D)
 * 119) Scerri (2022; lecture, mentions what's happened since the provisional report came out)
 * 120) Teknowijoyo and Gulian (2023) Demonstration of superconductivity in this material [Re-Lu] justifies the point of view that Lu plays a role of group 3 transition metal in period 6 of the Periodic table of elements.
 * 121) Winter (2024) The trends for the corresponding average CN [coordination number] values of groups 4–7 elements are reminiscent of the Sc–Y–Lu trend and on that basis Lu is better placed beneath Y rather than La in a chemical periodic table based upon VN [valence number] and CN.
 * 122) Strub et al. (2024) Finally, the bond lengths of perrhenates and pertechnetates of Sc, Y, La and Lu shall be compared (Fig. 30). These are again very similar for the pertechnetates and the perrhenates. It also reflects that bond lengths as well as crystal radii of the Lu and Y compounds are very similar, as would be expected for any 4d/5d element pair. In this sense, our data might be a humble contribution to the ongoing discussion as to whether La or Lu (and Ac or Lr, respectively) should be placed beneath Sc and Y in the periodic table. Considering our data, we widely agree with the view of [Jensen and Scerri]. The main obstacle of placing Lu underneath Sc and Y is the reluctance to display the periodic table in its 32-column form. Consequently, as there is also evidence that Lr is a Lu homologue, it may be hypothesised that Lr(TcO4)3 will also fit this trend. A 32-column periodic table representing the bond lengths of all known metal pertechnetates is shown in the SI.


 * Textbooks

As shown in the IUPAC survey, this form has made quite a few textbooks (including Clayden et al.'s famous Organic Chemistry), so it's got a significant case. Moreover we have a precedent for listening to journal articles over textbooks when we have a lot of journal articles complaining about what they see as a textbook error: cf. the saga of whether or not d orbitals are involved in hypervalent compounds (most textbooks say yes, but most journal articles say no).

At times there are cases of confusion. Cotton & Wilkinson, as well as Housecroft & Sharpe, at different points give Sc-Y-La-Ac and Sc-Y-*-**. Mingos' Essential Trends in Inorganic Chemistry mostly gives Sc-Y-La-Ac tables, but on p. 381 when he lists longest-lived isotopes of the actinoids, he only lists Ac-No. Same when he lists stable oxidation states on p. 382. And on p. 387 Lr, as well as Rf-Mt, is called a transactinide. All that only makes sense with Sc-Y-Lu-Lr.


 * Reference books
 * 1) Encyklopedia PWN (Polish): "skandowce, pierwiastki chemiczne tworzące 3. grupę układu okresowego pierwiastków chemicznych: skand Sc, itr Y, lutet Lu i lorens Lr (dawniej do skandowców zaliczano skand, itr, lantan La i aktyn Ac, często także tzw. pierwiastki f-elektronowe: lantanowce i aktynowce)" [i.e. the scandium group contains Sc-Y-Lu-Lr, but formerly ran Sc-Y-La-Ac and included the Ln and An]


 * Others


 * Opposition

In support of La there are among research articles (many textbooks show this but don't bother justifying themselves):


 * 1) Smith (1927)
 * 2) Wybourne (1965), Spectroscopic Properties of Rare Earths. Notes that for chemists La is a rare earth (those are supposed to be characterised by progressive f-shell filling) because of its chemistry, whereas for spectroscopists it isn't because it doesn't actually have an f-electron. (Although that doesn't stop him from including thorium, which doesn't have any f-electrons either.)
 * 3) Dash (1967),
 * 4) Trifonov (1970) could it be "Редкоземельные элементы и их место в периодической системе"?
 * 5) Shchukarev (1974)
 * 6) Atkins (2006)
 * 7) Lavelle (2008)
 * 8) Lavelle (2008)
 * 9) Lavelle (2009)
 * 10) *The above three are quite clearly rebutted by Jensen (2015) above in the Lu list.
 * 11) Restrepo (2017), . Concludes that stoichiometrically, La acts more similarly to Y than Lu. A pity that it is common knowledge in rare earth chemistry that this is false: yttrium is well-known to be very similar to heavy lanthanides, see e.g. Greenwood & Earnshaw, Cotton's Lanthanide and Actinide Chemistry.
 * 12) Vernon (2020, quite long; important for collecting many arguments. Even if the first one makes an uncited claim about the periodic law that is not present in the footnote statement of the periodic law.)
 * 13) Campero and Ponce (2020), supposedly anyway
 * 14) *But in fact, if you read the supplementary information, you will see that it does not clearly support La in group 3. In page S103 it says The fail of continuity of La to Hf in the electronegativity scales, Vg., &chi;&alpha; and GH scales, gives more validity to include La as the beginning of the lanthanides and less to include it before to Hf. On the other hand, Lu has a continuity in its electronegativity ratio with those of the lanthanides in 9 of the 11 the electronegativity scales as seen in Table S10B. Then, Lu belongs to the Lanthanides. On the other hand, the electronegativity ratio of Lu has a similar relationship with G3 as La, thus both La and Lu have a relationship with G3 as seen in Table S10B. This is simply fence-sitting, saying that La and Lu both fit well with lanthanides and with group 3 – well obviously yes, they're all called rare earths for a reason, but that doesn't prove they're f-elements (Sc and Y fit well with lanthanides, but they are not f-elements either).
 * 15) Vernon (2023, quite long; important for collecting many arguments, even if once again it goes on about the irrelevant gas-phase configurations of isolated atoms and isolated cations)

In support of * there are:


 * 1) Xu and Pyykkö (2016)
 * 2) *But the logic is strange, it supports Lu and Lr in the f-block on the basis of chemical bonding, after just having concluded that they don't use f-orbitals for that
 * 3) Kaltsoyannis (2017)
 * 4) Chandrasekar, Joshi, and Ghanty (2019)
 * 5) Chemey and Albrecht-Schmitt (2019)
 * 6) *But self-contradictory: first correctly says that Lu and Lr don't have readily accessible f-orbitals, but then claims that putting them in f-block rows correctly models their chemical bonding. Even though the Xu and Pyykkö article cited calls 5f a core orbital for Lr. This is also an issue with Kaltsoyannis (2017): first correctly saying the atomic ground state matters less than the chemical bonding, but not addressing the fact that there is no f-bonding for Lu and Lr.
 * Pyykkö (2019)

Interestingly, this source (2023) claims that Sc-Y-* really means Sc-Y-Lu: The 15th entry of the f-block represents the first slot of the d-block which is left vacant to indicate the place of the f-block inserts. This actually quite agrees with how IUPAC uses it: in 18-column they tend to show Sc-Y-*, but when expanding it into 32-column they show Sc-Y-Lu. Of course, the obvious question is: if you really mean that Lu belongs in group 3, then why not just show that in the 18-column form too, instead of creating all this confusion?

There are some sources that are either ambiguous (self-contradictory), or just discuss/mention the problem but do not ultimately choose a side. Although they clearly don't count, they at least show that awareness of the problem is rising.
 * 1) Clark and White (2008). Seems a bit confused. First says We recalled the 1982 article in this Journal in which Jensen’s definitive arguments settled this once and for all (1). Jensen chose 14LaAc [Sc-Y-Lu] as the answer. But then says Then we found that in 2005 the IUPAC had decided this question. The chosen standard subscripted-form table is 15LaAc [Sc-Y-*] (2). We prefer the 15LaAc table as the flyleaf table. So, which do they want? Later Clark changed his mind after seeing Jensen's and Lavelle's responses and advocated that the flyleaf form (18-column) vanish in favour of the long form (32-column), but still not saying what group 3 should be.
 * 2) Daintith (2008): The Oxford Dictionary of Chemistry "Group IIIA consisted of scandium (Sc), yttrium (Yt), and lanthanum (La), which are generally considered with the lanthanoids, and actinium (Ac), which is classified with the actinoids. Scandium and yttrium now belong to group 3 (along with lutetium and lawrencium)." But elsewhere it thinks La is a transition metal and that Ce-Lu are the f-block (e.g. entries on "d-block elements" and "transition elements"). So it's just confused.
 * 3) Bünzli (2014). First says it's Sc-Y-La-Ac, then says there is a dispute.
 * 4) Jemmis (2018). Talks about Ghanty et al.'s findings, doesn't come to a decision.
 * 5) Walshe (2018). Both quotes Ghanty et al. recommending *, and Lavelle recommending La.
 * 6) Cao et al. (2019). Calls La/Ac and Lu/Lr both d elements.
 * 7) Geckeis (2019). Says that both La and Lu could have a claim to be d-block elements because they're both d1s2. In aktuellen Kurzversionen, die von der Iupac und der GDCh erhältlich sind, werden sowohl Lanthan und die darauf folgenden Lanthanoide als auch Actinium und die folgenden Actinoide aus dem PSE ausgeklinkt und darunter angeordnet. Kritisieren könnte man daran, dass diese Darstellung eine Zugehörigkeit von Lanthan und Actinium zu den 4f- beziehungsweise den 5f-Elementen suggeriert. Beide sollten vielmehr entsprechend ihrer d1s2-Elektronenkonfiguration im Grundzustand zur Gruppe 3 gehören. In dieser Hinsicht ist die comicartige Darstellung im PSE-Poster der GDCh „korrekter“. Hier fehlt noch ein Hinweis darauf, wo die f-Elemente einzufügen sind. Über die Stellung im PSE sowohl von Lanthan und Actinium einerseits als auch von Lutetium und Lawrencium andererseits wird jedoch derzeit diskutiert. Zumindest für Lutetium ist sicher, dass es neben der voll besetzten 4f-Schale ebenfalls eine d1s2-Konfiguration besitzt und damit auch den Anspruch erheben könnte, dem d-Block zugeschrieben zu werden.
 * 8) Rayner-Canham (2020). Contains whole chapter about this very issue. Writes "Using the long (32-column) form of the Periodic Table, there is no issue: lutetium and lawrencium fall naturally in place". The Luder/Scerri argument is mentioned (Sc/Y/La/Ac breaks the d block); Jensen's arguments appear without criticism, Lavelle's electron-configuration appears but it is also mentioned that ECs are more consistent with Sc/Y/Lu/Lr. However, Rayner-Canham then writes "In this book, ... the lower placements of Group 3 will be left empty" (in order to emphasise continuity over La-Lu and Ac-Lr). Nevertheless on his front page, Lu and Lr go below Y (albeit with a gap so that La-Yb and Ac-No can also be squeezed in).
 * 9) Schwerdtfeger, Smits, and Pyykkö (2020) [Notes that Sc-Y-Lu is more consistent with Madelung and more natural, but says Sc-Y-La keeps La as first member of lanthanoid series. (How does Sc-Y-Lu not do that, though? In fact it's rather Sc-Y-La that makes La fall out of the 4f series Ce-Lu.) Then just stops there and says IUPAC avoids this controversy (which, as you can see, it doesn't).] Still, in Fig. 3 Sc-Y-Lu is assumed and adopted.
 * 10) Cotton (2021). Mentions the problem, notes that Sc-Y-Lu matches group 4 better, Sc-Y-La matches group 2 better. Then doesn't give an opinion. Surely the d-block groups are more relevant though, since Sc and Y are uncontroversially d-block. Unfortunately this is one of those things that are so obvious that it is hard to cite (Jensen and Chistyakov only compare to the other d-elements.)
 * 11) Kwarsick et al. (2021). Talks about measuring IP2 of Lr, says that supports that it's the last actinide (though the actinide series and the 5f series are not necessarily the same thing), but says: However, due to the complexities of this issue, it is unlikely that this issue will be settled on objective means. Rather, the community will likely need to agree upon an accepted convention.
 * 12) *I normally don't comment (other than to mention cases where the articles contradict themselves), but this is a really bizarro-world article where the authors remark that Fig. 1 show the trends going crazy for Lu and Lr that jump far away from the earlier elements, and then proceed to conclude that it suggests that Lu and Lr belong with the f-elements after all. They even say that The measured IP1 and IP2 values for Lr confirm that it is the last actinide element, having a ground-state electronic configuration with a single weakly bound electron outside a [Rn]5f147s2 core ... This fundamental information, in conjunction with state-of-the-art relativistic calculations, will unambiguously show that the Lu and Lr each contain one additional electron outside of fully occupied f orbitals. Following the bizarre logic thus implied, every period must end with the alkali metals, because they contain one additional electron outside of a fully occupied closed shell. Huh?
 * 13) Cotton et al. (2022) [Says "the case for replacing La by Lu in Group 3 has not been made on structural grounds", but says "the Periodic Table as presented currently, not least in the IUPAC form, gives the most appropriate description of the chemistry of Group 3 and of the lanthanides". But the IUPAC table is Sc-Y-* and not Sc-Y-La...]
 * 14) *BTW, in fact their graphical abstract clearly shows that Lu is between Sc and Y, but closer to Y; and that La is much bigger. So only Sc-Y-Lu matches the early transition-metal pattern where the 4d and 5d elements are similar in size, but the 3d one is noticeably smaller, e.g. Ti-Zr-Hf, V-Nb-Ta. Essentially, the secondary periodicity argument of Chistyakov (1970). Thus when Sc and Y give the same structure, Lu tends to match them, while La not doing so: consider the metal, oxides, and chlorides as Jensen (1982) does.
 * 15) Cotton et al. (2022), a related article that mentions the problem but doesn't give an opinion.
 * 16) Labarca et al. (2022), part one, part two
 * 17) Neve (2022). Mentions the IUPAC report, mentions some opposition, but shows (Fig. 1) a Sc-Y-Lu-Lr table, and while it sometimes mentions/shows data for Lr to extend the 6d series backward to group 3, it never does so for Ac. Admittedly, might be because Ac is not a transactinide (in a Sc-Y-La-Ac table, it would appear before the 5f series) and this article is about transactinides.
 * 18) Sato and Nagame (2022), somewhat leaning towards Sc-Y-Lu-Lr (though it does point out that Lr probably has a p-electron). Note that they consider the debate just to be about Lu and Lr (so it becomes a debate between Sc-Y-* and Sc-Y-Lu).
 * 19) Scerri (2022), an interview

For an interesting example of confusion, see Ullmann. First it is stated that Sc, Y, and La are the first members of d-series, and that all rare earths fit between La and Hf. But then immediately it is stated that all rare earths formally belong to group 3. Then the lanthanoid contraction is stated to be Ce3+ to Lu3+. So, is it La or *? The logic is also strange near the end of that section: it is said that Eu and Yb are exceptional because they are almost at f7 and f14 and try to achieve this. So why aren't Gd and Lu exceptional, if they are supposedly really at those configurations? This is a bit like saying that periods start at group 2 and end at group 1 and that there is always a covering s1 shell. Noble gases as p5s1 are exceptional because they are almost at p6 and try to achieve this. Dunno, seems strange to me.

Not directly related but still gold:

10.1002/anie.197300121

see also: https://www.degruyter.com/view/journals/pac/92/3/article-p515.xml


 * Interesting

Vyatkin (2019)

To read also: https://www.degruyter.com/document/doi/10.1515/pac-2019-graphabs12/html

Why are people so obsessed with the gas-phase configurations of isolated atoms or ions? It is about the same kind of thing as using van der Waals radii rather than covalent radii. See Johnson (1984) for a nice takedown on this. In short, he notes: transition-metal behaviour in the actinides lasts from Ac to U, but transition-metal gas-phase configurations only last from Ac to Th. Meanwhile, lanthanide-like behaviour in the actinides starts at Cm, but lanthanide-like gas-phase configurations start at Pu. What exactly has one got to do with the other, anyway?

A IUPAC timeline

 * 1985: Norman E. Holden gives a presentation about the group 3 issue at the 33rd IUPAC General Assembly in Lyon, France. He concludes that Sc-Y-Lu-Lr should replace Sc-Y-La-Ac on the basis of electronic and physicochemical data.
 * 1988: IUPAC issues a report on notations for the groups of the periodic table. It shows an f-block as La–Yb and Ac–No (consistent with Sc-Y-Lu-Lr). It says that "According to the electron configurations of the elements, the scandium group consists of the elements Sc, Y, Lu, Lr", and later notes that physicochemical properties justify this as well. Nonetheless, it notes that "most periodic tables in textbooks and classrooms" (well, at the time it was certainly true) showed Sc-Y-La-Ac. The report states that the next Red Book will show "a compromise", which based on the sources seems to be Sc-Y-*-** (but it doesn't explicitly say so).
 * 1990: The next Red Book appears, giving three periodic tables: 8-column (an update of Mendeleev's original arrangement, common in the Soviet Union), 18-column, and 32-column. The first two give Sc-Y-*-**, but the last gives Sc-Y-Lu-Lr.
 * 2003–present: As the 1990 Red Book table requires updates due to the confirmation and naming of elements beyond 103 (Lr), IUPAC publishes a so-called "IUPAC Periodic Table" on its website, extending to element 111 (the highest confirmed at that time). This table is 18-column and in the Sc-Y-*-** format. It has been continually updated to the present year by adding new elements and updating atomic weights to present knowledge.
 * 2005: The latest edition of the Red Book appears. All forms but the 18-column one are dropped, leaving only Sc-Y-*-**.
 * 2009: The IUPAC newsmagazine Chemistry International publishes an article by Jeffrey Leigh (the editor of the 1990 Red Book). It clarifies that IUPAC "has not approved any specific form of the periodic table, and an IUPAC-approved form does not exist, though even members of IUPAC themselves have published diagrams titled “IUPAC Periodic Table of the Elements." Just as in the 1990 Red Book, Leigh's article shows Sc-Y-*-** in 18-column, but Sc-Y-Lu-Lr in 32-column.
 * 2012: Chemistry International publishes an article by Eric Scerri arguing that IUPAC should take a stance on the composition of specific groups. It argues in favour of Sc-Y-Lu-Lr. The reasons given are physicochemical data in which Lu is more similar to Sc and Y than La, and also because Sc-Y-La-Ac would either necessitate breaking the Z order or splitting the d-block.
 * 2015: IUPAC begins a project to decide on the composition of group 3, chaired by Scerri. It lists Sc-Y-Lu-Lr and Sc-Y-La-Ac as possible outcomes, but not the Sc-Y-*-** compromise that had been partially but not fully adopted in 1990.
 * 2016: Chemistry International announces the beginning of the group 3 project, once again confirming that the question is between Sc-Y-Lu-Lr and Sc-Y-La-Ac. The Sc-Y-La-Ac form it displays is not the one with a split d-block, but rather the one that has La-Ac between Lu-Lr and Hf-Rf, thus breaking the order of atomic numbers. Probably since about then, the IUPAC page on the periodic table (giving the "IUPAC periodic table") has mentioned the group 3 problem, the project to resolve it, and says "Stay tune[d]".
 * 2017: The IUPAC project group meets. One member (Philip Ball) publishes an article on the group 3 problem in the magazine Chemistry World, published by the Royal Society of Chemistry. It ends by arguing for Sc-Y-Lu-Lr: Consider what the alternative of putting Lr and Ac [sic] in the d block entails: either the f block has to be placed before the d block (which disrupts the steady progression of atomic number, and so seems indefensible), or the f block must be interposed between group 3 and the rest of the d block, which seems awkward at best. What’s more, this arrangement would then disrupt vertical trends in several chemical properties, such as atomic radius and melting point. Besides, Lu has physical and mechanical properties resembling those of transition metals – although whether it is chemically more akin than La to Sc and Y is less clear. (This is not strictly done by IUPAC, but it is announced on the IUPAC project website.)
 * 2018: Another member of the IUPAC project (Lars Öhrström) publishes an article in Nature Chemistry that briefly mentions the group 3 problem. He does not take a side. Scerri and Parsons publish an book chapter in which Sc-Y-Lu-Lr is argued for, to retain the sequence of increasing Z, avoid splitting the d-block, and stay consistent with the Aufbau principle. (Same considerations as the 2017 entry on the timeline.)
 * 2019: Chemistry International publishes an article by Scerri on the history of the periodic table. Although he accepts that neither chemical and physical evidence on the elements concerned, nor microscopic evidence in the form of electronic configurations, provide an unambiguous resolution of the question, and that there is no absolutely correct version of the periodic table, he still recommends that IUPAC a compromise periodic table that most effectively conveys the largest amount of information to the largest group of users. He argues for Sc-Y-Lu-Lr so that the sequence of increasing Z can be retained, and illustrates the 18-column table that way. Scerri and Wong prepare for IUPAC a survey of university textbooks and publish the data on which form of group 3 they show. It shows that while in the past Sc-Y-La-Ac dominated, it lost its majority in the 2010s (though it still retains a plurality).
 * 2021: Chemistry International publishes a provisional report from the group 3 project, written by Scerri. Though it says that the situation can't be resolved objectively, it then says that that makes it more important for IUPAC to say something as it's a matter of convention. It argues for Sc-Y-Lu-Lr to make the blocks have the expected widths from quantum mechanics (2, 6, 10, 14), retain the sequence of increasing Z, and avoid splitting the d-block. The textbook situation is not referred to at all. The Sc-Y-*-** form that appeared in the 1990 Red Book is stated as having been designed by practitioners of specialized branch of relativistic quantum mechanics concerned with the properties of super-heavy elements, and the report makes the statement that Such interest-dependence should not, in our view, dictate how the periodic table is presented to the general chemical and scientific community. The IUPAC Inorganic Chemistry Division is happy with this conclusion, but considers the update of the PT to be non-urgent until the next element is discovered. The project ends.