Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

May 27, 2024

Lanthanides and the Exceptional Lie Group G2

Posted by John Baez

The lanthanides are the 14 elements defined by the fact that their electrons fill up, one by one, the 14 orbitals in the so-called f subshell. Here they are:

lanthanum, cerium, praseodymium, neodymium, promethium, samarium, europium, gadolinium, terbium, dysprosium, holmium, erbium, thulium, ytterbium.

They are also called ‘rare earths’, but that term is often also applied to 3 more elements. Why? That’s a fascinating puzzle in its own right. But what matters to me now is something else: an apparent connection between the lanthanides and the exceptional Lie group G2!

Alas, this connection remains profoundly mysterious to me, so I’m pleading for your help.

Why are there 14 lanthanides? It’s because

  • the electrons in the f subshell have orbital angular momentum 33,
  • the irreducible representation of SO(3)SO(3) corresponding to angular momentum j=3j = 3 has dimension 2j+1=72j+1 = 7, and
  • each electron can also come in 22 spin states, for a total of 2×7=142 \times 7 = 14 states.

What does this have to do with the exceptional Lie group G 2\mathrm{G}_2? The aforementioned 77-dimensional representation of SO(3)SO(3) can also be thought of as the space of imaginary octonions, since — rather amazingly — the key algebraic structure on the imaginary octonions, their cross product, is invariant under this representation of SO(3)SO(3). Indeed, the 7-dimensional representation of G 2\mathrm{G}_2 on the imaginary octonions remains irreducible when restricted to a certain SO(3)SO(3) subgroup of G 2\mathrm{G}_2, sometimes called SO(3) irrSO(3)_{irr} — and this gives our friend the j=3j = 3 representation of SO(3)SO(3).

All these facts were noticed and apparently put to some use by the mathematician and physicist Giulio Racah, famous for his work on the quantum mechanics of angular momentum. This was recently brought to my attention by Paul Schwahn, who is working to better understand the underlying math.

But Racah’s thoughts remain deeply mysterious, because Schwahn found them in a fragmentary second-hand account, and we haven’t been able to find more details!

Schwahn writes:

I thought me mentioning the f-orbital was just a crackpot idea.

But in the AMS volume Selected Papers of E. B. Dynkin with Commentary (which also contains Dynkin’s original discovery of SO(3)ᵢᵣᵣ) one finds a short review by Yuval Ne’eman, titled “Dynkin Diagrams in the Physics of Particles, Fields and Strings”. The whole thing is a delight to read, but he writes something particularly interesting about an idea of physicist Giulio Racah:

“Racah found ways of applying various simple algebras in classifying higher spectra. His methods, later developed and extended by such as L. Biedenharn and M. Moshinsky, exploited higher rank Lie algebras applied to the representation spaces of SO(3)SO(3). I recall Racah enjoying (anecdotically) the fact that he had found an application for Cartan’s exceptional G(2)\mathrm{G}(2), in studying the f-subshell in atomic spectra. One defines an SO(7)SO(7) algebra acting on some constructs involving the 7-dimensional f-subshell representation of SO(3)SO(3) - and the inclusion G(2)SO(7)G(2) \subset SO(7) does it. In these very complicated atomic spectra of the lanthanides, it provides some physical insights.”

I really wonder that these physical insights are…

It’s possible that even if Racah’s thoughts are lost in the dark mist of time, later researchers on group representation theory and the quantum mechanics of atoms have used the Lie group G 2\mathrm{G}_2 to understand something about f subshell electrons. For example, Biedenharn’s book may contain some clues. But I haven’t yet turned up any clues yet.

Posted at May 27, 2024 10:15 PM UTC

TrackBack URL for this Entry:   https://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/3537

9 Comments & 0 Trackbacks

Re: Lanthanides and the Exceptional Lie Group G2

Here is a hint of G 2\mathrm{G}_2 showing up in studies of the f shell:

  • R. Bruce King, The group-theoretical structure of the atomic g shell: connection with the alternating group A 6\mathrm{A}_6 as L 2(9)\mathrm{L}_2(9), Journal of Mathematical Chemistry 44(1) (2008), 5–19. (paywalled link.)

For the atomic g shell the group L 2(9)\mathrm{L}_2(9) is isomorphic with the alternating group A 6\mathrm{A}_6 on six objects of order 360 or the symmetry group of the 5-dimensional simplex, a 5-dimensional analogue of the tetrahedron with 6 vertices and 15 edges. This leads to the subgroup chain SO(9)SO(5)L 2(9)SO(9) \supset SO(5) \supset \mathrm{L}_2(9) for the atomic g shell analogous to the subgroup chain SO(7)G 2L 2(7)\mathrm{SO}(7) \supset \mathrm{G}_2 \supset \mathrm{L}_2(7) for the atomic f shell.

L n(q)=PSL(n,𝔽 q)\mathrm{L}_n(q) = PSL(n,\mathbb{F}_q) is the group of invertible n×nn \times n matrices of determinant 11 over the field with qq elements, modulo multiples of the identity matrix. L 2(7)\mathrm{L}_2(7) has 168 elements and it’s famous for lovers of the octonions as the symmetry group of the Fano plane.

I haven’t gotten the actual paper yet. The study of the g shell is deeply unpopular in chemistry since it only applies to elements of atomic number 124 and higher!

Posted by: John Baez on May 28, 2024 7:09 AM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

This paper gives a link to Racah’s work on G 2\mathrm{G}_2 and the f shell:

  • R. Bruce King, The group-theoretical structure of the atomic g shell: connection with the alternating group A 6\mathrm{A}_6 as L 2(9)\mathrm{L}_2(9), Journal of Mathematical Chemistry 44(1) (2008), 5–19. (paywalled link.)

King writes:

The detailed study of the atomic d shell was initiated by Condon and Shortley [1] in 1935 following earlier work by Slater [2] in 1929. In 1949 Racah [3] developed group-theoretical methods for study of both the atomic d and f shells…”

Reference [3] is

  • G. Racah, Theory of complex spectra IV, Physical Review 76(9) (1949), 1352–1365. (paywalled link.)

and it indeed does quite a bit with G 2\mathrm{G}_2!

But King’s paper also has its own interesting stuff on f orbitals, and it refers to this for more:

  • R. Bruce King, Group-theoretical structure of the atomic f shell: connection with the non-Euclidean heptakisoctahedral (didodecahedral) group, Molecular Physics 104(20-21) (2007), 3261–3268. (paywalled link.)

The ‘heptakisoctahedral group’ is our friend the 168-element group PSL(2,𝔽 7)PSL(2,\mathbb{F}_7).

Posted by: John Baez on May 28, 2024 10:04 AM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

Now that you’ve mentioned A 6A_6, I’m suddenly hoping that there’s some amazing chemistry that comes from the existence of its weird outer automorphism.

Posted by: Scott McKuen on May 31, 2024 5:09 AM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

I still don’t know King is embedding A 6PSL(2,𝔽 9A_6 \cong PSL(2,\mathbb{F}_9 in SO(9)SO(9) or PSL(2,𝔽 7)PSL(2,\mathbb{F}_7) in SO(7)SO(7) — he doesn’t say, and I haven’t succeeded in guessing!

But in pondering this, I ran into something which seems to say the outer automorphism of A 6A_6 can be understood as the outer automorphism of PSL(2,𝔽 9)PSL(2,\mathbb{F}_9) that arises from the nontrivial automorphism of the field 𝔽 9\mathbb{F}_9, so that’s nice. The field 𝔽 9\mathbb{F}_9 can be thought of as 𝔽 3\mathbb{F}_3 with a square root of 1-1 thrown in, say “ii”, and the automorphism sends ii to i-i. So it’s just like complex conjugation.

I still don’t understand the isomomorphism A 6PSL(2,𝔽 9)A_6 \cong PSL(2,\mathbb{F}_9), so if anyone knows how that works, I’d be interested.

Posted by: John Baez on June 2, 2024 10:13 AM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

This construction on math.SE might help. The idea is to construct an index-6 subgroup G<PSL(2,𝔽 9)G \lt PSL(2,\mathbb{F}_9), which I assume is so that one gets a certain permutation representation of PSL(2,𝔽 9)PSL(2,\mathbb{F}_9) on PSL(2,𝔽 9)/GPSL(2,\mathbb{F}_9)/G, and then the sizes work out so that the image of the permutation rep inside the symmetric group on this 6-element set has to be the corresponding alternating group.

Posted by: David Roberts on June 4, 2024 1:20 PM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

Hermann Weyl wrote a book called `Group theory and quantum mechanics’ around 1930 which was reprinted as a cheap Dover paperback which I encoountered as a kid and no longer have. I remember it as incomprehensible at the time, but if memory serves it included an account of

….the symmetric permutation group; and algebra of symmetric transformation (invariant sub-spaces in group and tensor space, sub-groups, Young’s symmetry operators, spin and valence, [* group theoretic classification of atomic spectra *]…

(from the Amazon blurb). It’s still in print but was apparently never Mathematically Reviewed; IIRC he talked about about electron shells around Ch 7. I believe you [JB] have had a shot at explaining shell theory elsewhere and I have been waiting a long time to understand this story.

I don’t think Herman knew about G_2; I sent him an email but may not get a quick response.

Posted by: jack morava on May 29, 2024 1:00 AM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

@Jack Morava : the book is available on the Internet Archive !

And it has been reviewed (the 1931 german edition) by none other than Bartel van der Waerden for the Zentralblatt.

Posted by: bruno on May 29, 2024 11:22 AM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

@ Bruno,

Many thanks! It’s still kind of incomprehensible after all these years, but the part I recall is in Ch V \S 15 p 372-6.

[Apparently my father-in-law knew Racah …]

Posted by: jack morava on May 29, 2024 4:53 PM | Permalink | Reply to this

Re: Lanthanides and the Exceptional Lie Group G2

Jack wrote:

I believe you [JB] have had a shot at explaining shell theory elsewhere and I have been waiting a long time to understand this story.

Yes, I summarized the key results of shell theory here:

The biggest thing I left out is the quite difficult calculation starting from the hydrogen atom Hamiltonian and concluding with the decomposition of its Hilbert space of bound states (which I called H\mathbf{H}) into eigenspaces of the Hamiltonian (which I called H n\mathbf{H}_n, and which are called ‘shells’) and then the further decomposition of each eigenspace into irreducible representations of SO(3)SO(3) (which I called H n,\mathbf{H}_{n,\ell}, and which are called ‘subshells’.) When you understand this stuff, you hold the magic key to understanding the periodic table!

I also have a ‘pop’ version of this story, which should be very easy and fun for you:

Keep hitting the “next” button at upper left. (It’s hard to see on the very first page.)

Posted by: John Baez on May 30, 2024 10:23 PM | Permalink | Reply to this

Post a New Comment