Accidental Degeneracy and the SO(4) Symmetry

Rydberg-atom laboratories work inside a single $n$-shell the way this article does. At $n \sim 50$ the 2500 nearly degenerate states form a playground steered with small fields, in exactly this language: circular-state preparation (the $\ell = n-1$ rungs of cavity-QED) is adiabatic steering through the $(j, j)$ multiplet, the Runge–Lenz operator generating the $\ell$-transfers, and the linear Stark maps used to field-ionize Rydberg atoms read off the $SO(4)$ quantum numbers. The idea of a spectrum-generating symmetry — solve a system by closing the algebra of its conserved and ladder operators — became a working method across physics: the $SO(4,2)$ treatment of all hydrogen states at once, the interacting boson model's dynamical symmetries, supersymmetric factorization chains, the integrable-model toolbox — all descend from Pauli's 1926 template. And "accidental degeneracy" survives as the standard alarm phrase: when a numerical spectrum shows more coincidence than the visible symmetry explains, the modern reflex is to hunt the hidden conserved operator — in cold-atom traps, photonic lattices, and band structures alike.

The classical tell has been in the sky all along. A planet in a pure $1/r$ potential retraces the same ellipse forever, its long axis pointing at a fixed star. That frozen direction is a conserved vector — the Laplace–Runge–Lenz vector $\mathbf{A}$, focus to perihelion, magnitude proportional to eccentricity. Perturb the potential even slightly and the ellipse becomes a turning rosette: $\mathbf{A}$ rotates, conservation gone (Mercury's perihelion precession is this diagnostic in reverse). Conservation of $\mathbf{A}$ fingerprints the exact Coulomb form. Quantum mechanics inherits the fingerprint with upgraded consequences. Rotations cannot change $\ell$ — the $\hat L$'s only reshuffle a multiplet internally. The components of $\hat{\mathbf{A}}$ are different animals: they commute with $\hat H$ yet carry vector ($k=1$) character, so they connect $\ell$ to $\ell \pm 1$. A conserved operator that changes $\ell$ without changing energy forces degeneracy: apply $\hat A_z$ to 2$s$ and out comes (a multiple of) 2$p_0$ at the same energy, by conservation. The "accident" is this machine at work.

Six conserved generators — three $\hat L$'s, three $\hat A$'s — are too many for the rotation group: $SO(3)$ has three; six is the count for $SO(4)$, rotations of four-dimensional Euclidean space. In 1935 Fock showed the four-dimensionality is literal: in momentum space, bound hydrogen is free motion on a 4-sphere (the portrait tells the story). Below is the operator version — close the algebra, split it into two su(2)s, let the Casimir arithmetic hand over the spectrum.

Everything rests on the quantum Laplace–Runge–Lenz operator. Its construction is delicate — the naive $\hat{\mathbf p}\times\hat{\mathbf L}$ is not even hermitian — and the symmetrized operator, with proofs of all credentials, is the business of The Laplace–Runge–Lenz Vector in Quantum Mechanics. We quote what we need:

where: $m$ — the (reduced) electron mass; the symmetrization makes $\hat{\mathbf{A}}$ hermitian. Its credentials, all cited from that article: $[\hat H, \hat A_i] = 0$ (conserved for the Coulomb $\hat H = \hat p^2/2m - k/r$, and only for it), $\hat{\mathbf{A}}\cdot\hat{\mathbf{L}} = \hat{\mathbf{L}}\cdot\hat{\mathbf{A}} = 0$ (the arrow lies in the orbital plane), and the squared-length identity

where: the lone $+\hbar^2$, absent from the classical $A^2 = 2EL^2/m + k^2$, is the residue of operator ordering — and precisely the term that will make $n$ an integer. A classical physicist would delete it as clutter; deleting it misquantizes hydrogen. The commutators among the six generators are the whole story, so they get a derivation card.

The algebra $\mathfrak{so}(4)$ has a lovely secret: unlike its three-dimensional cousin it is not simple — it falls apart into two independent copies of the angular-momentum algebra. The right change of variables makes the split visible.

With the machinery assembled, the spectrum falls out of bookkeeping. This is the calculation Wolfgang Pauli published in January 1926 — solving hydrogen by matrix mechanics while the differential-equation route did not yet exist.

where: $g_n$ — states per shell (spin doubles it to $2n^2$ but adds no $\ell$-physics). Compare the chapter's two derivations of hydrogen: the analytic route needed a differential equation, an asymptotic peel, a series, and a termination demand; the algebraic route needed six commutators and one constraint. The price of elegance is fragility — the algebra knows only the exact $1/r$ problem. One more reading deserves a paragraph. Fock's 1935 construction: stereographically project each bound momentum state, from the momentum sphere of radius $p_0 = \sqrt{-2mE}$, onto the unit sphere $S^3$. There hydrogen's Schrödinger equation becomes a free particle's on $S^3$ — symmetry manifest — and the 4D spherical harmonics come in families of dimension exactly $n^2$: hydrogen's degeneracy is the statement that a hypersphere looks the same from every side. (Stated without derivation; the portrait band tells the story.)

1 · The n = 2 shell, assembled from two spin-halves (confidence builder)

Work in the smallest nontrivial representation, $n = 2$: $j_1 = j_2 = \tfrac12$. Build the generators, verify the algebra, and watch the shell's orbital content — one 2s, three 2p — drop out of angular-momentum addition.

Step 1. Represent the two spins on $\mathbb{C}^2\otimes\mathbb{C}^2$: $\hat{\mathbf{J}}^{(1)} = \tfrac{\hbar}{2}\boldsymbol\sigma\otimes\mathbb{1}$, $\hat{\mathbf{J}}^{(2)} = \mathbb{1}\otimes\tfrac{\hbar}{2}\boldsymbol\sigma$ — $4\times4$ matrices. (Symmetry spins, not electron spin: the four dimensions are the $n=2$ shell's four states.)

Step 2. Machine-check the structure (this article's verify script): all nine $[\hat J^{(1)}_i, \hat J^{(2)}_j] = 0$; each triple closes su(2); $\hat{\mathbf{L}} = \hat{\mathbf{J}}^{(1)} + \hat{\mathbf{J}}^{(2)}$ and $\hat{\mathbf{A}}' = \hat{\mathbf{J}}^{(1)} - \hat{\mathbf{J}}^{(2)}$ reproduce the full $\mathfrak{so}(4)$ table; and $\hat{\mathbf{L}}\cdot\hat{\mathbf{A}}' = 0$ exactly — automatic in a $(j,j)$ representation.

Step 3. Decompose by total angular momentum: diagonalizing the $4\times4$ $\hat{\mathbf{L}}^2$ gives eigenvalues $\{0, 2\hbar^2, 2\hbar^2, 2\hbar^2\}$ — one $\ell = 0$ state and three $\ell = 1$ states, the 2s and the 2p triplet: the $\tfrac12\otimes\tfrac12 = 0 \oplus 1$ Clebsch–Gordan series in the flesh.

Step 4. Confirm the energy label: $n = 2j+1 = 2$, and the Casimir arithmetic gives $E_2 = -13.6057/4 = \mathbf{-3.401\,eV}$ for all four states. The 2s–2p degeneracy is not two integrals that happened to agree; it is the statement that both multiplets live in one irreducible $(\tfrac12, \tfrac12)$ box.

2 · Running the Casimir constraint up the ladder (exam level)

Evaluate the spectrum at $j = 0, \tfrac12, 1$ against the radial-equation ladder; then find which states the Runge–Lenz operator annihilates.

Step 1. The constraint reads $-mk^2/2E = \hbar^2(2j+1)^2$, with $mk^2/2\hbar^2 = 13.6057\,\text{eV}$ (the Rydberg — assembled from CODATA in The Bohr Radius and Characteristic Atomic Scales): $j = 0 \Rightarrow \mathbf{-13.606\,eV}$; $j = \tfrac12 \Rightarrow \mathbf{-3.401\,eV}$; $j = 1 \Rightarrow \mathbf{-1.512\,eV}$. All three match the $-\text{Ry}/n^2$ ladder of The Hydrogen Atom: Coulomb Radial Equation exactly — two derivations, zero daylight.

Step 2. Multiplet sizes: $(2j+1)^2 = 1, 4, 9$ — with electron spin, $2n^2 = 2, 8, 18$: the periodic table's first row lengths emerge as $SO(4)$ representation dimensions (the filling order beyond hydrogen involves screening, which breaks the symmetry being counted).

Step 3. Where does $\hat{\mathbf{A}}'$ run out of work? On an $(n, \ell)$ level, Derivation C's Step 2 gives $\hat{\mathbf{A}}'^2 = \hbar^2[n^2 - \ell(\ell+1) - 1]$: at the top rung $\ell = n-1$ this is $\hbar^2(n-1)$, and it vanishes only at the ground state $(1, 0)$ — the one shell with no other $\ell$ to shuffle to.

Step 4. Positivity sanity check: $n^2 - \ell(\ell+1) - 1 \ge 0$ demands $\ell \le n-1$ — the operator identity knows the allowed quantum numbers. (Verified for $n \le 7$ in the script.)

Texts. Sakurai & Napolitano, 3rd ed., §4.1 + Runge–Lenz/SO(4) supplement · Shankar, 2nd ed., Ch. 13 · Cohen-Tannoudji Vol. I, Runge–Lenz complement · Griffiths & Schroeter §4.2 · Pauli, Z. Phys. 36, 336 (1926) · Fock, Z. Phys. 98, 145 (1935) · Bargmann, Z. Phys. 99, 576 (1936).

Provenance. Spectrum and degeneracy: $E_n=-13.6057\,\text{eV}/n^2$, $g_n=n^2$ — verify against Bohr/Schrödinger solution (Griffiths §4.2, Sakurai §4). Runge–Lenz operator, its Hermitian symmetrization, conservation $[H,\mathbf{A}]=0$, orthogonality $\mathbf{A}\cdot\mathbf{L}=0$, and the identity $\mathbf{A}^2=\tfrac{2H}{\mu}(\mathbf{L}^2+\hbar^2)+k^2$ — recompute by direct commutator algebra (Shankar Ch. 13; Cohen-Tannoudji complement). $SO(4)\cong SU(2)\times SU(2)$ decomposition with $j_1=j_2=(n-1)/2$ and the Casimir derivation of $E_n$ — reproduce Pauli (1926). Clebsch–Gordan series $j\otimes j=\bigoplus_{\ell=0}^{2j}\ell$ gives $\ell=0\dots n-1$. Bertrand's theorem (closed orbits only for $1/r$ and $r^2$) — classical mechanics (Goldstein). No nonstandard numbers used; the only numeric value is the Rydberg energy $13.6057$ eV. Edition note (2026-07-08). The Runge–Lenz operator's construction, hermiticity, conservation, orthogonality, and squared-length identity are quoted with their owner named — The Laplace–Runge–Lenz Vector in Quantum Mechanics proves them; this article owns the closure, the su(2)⊕su(2) decoupling, and the Casimir spectrum. The base's second derivation (a matching-condition sketch that only $V \propto 1/r$ conserves $\mathbf{A}$) is not reproduced: its conclusion is stated via Bertrand's theorem, and operator-level uniqueness stays with the operator's article. All algebra machine-verified (symbolic differential operators + the explicit $n=2$ representation).

Prerequisites. The hydrogen atom: Coulomb radial equation · the Bohr energy spectrum and Rydberg constant · degeneracy and the role of the magnetic quantum number · the Laplace–Runge–Lenz vector in quantum mechanics. Leads to. Hydrogenic wavefunctions and quantum numbers · hydrogenic ions and scaling with nuclear charge.