The Runge-Lenz vector, SO(4) symmetry, and the accidental degeneracy of hydrogen

Intuition Beginner

The hydrogen atom has $n^2$ states at each energy level $E_n$. For $n=2$, there are four states: one with $\ell =0$ (the 2s orbital) and three with $\ell =1$ (the 2p orbitals). All four have the same energy. A generic central potential — say $V(r)=-1/r^3$ — would give different energies for different $\ell$ values. The Coulomb potential $V(r)=-1/r$ is special: it has an extra conserved quantity that forces all those states to share one energy.

In classical mechanics this extra quantity is the Laplace-Runge-Lenz vector $\mathbf{A}$. For a planet orbiting the sun under gravity, $\mathbf{A}$ points along the long axis of the ellipse. Its magnitude tells you the eccentricity. Its direction tells you the orientation of the orbit. The vector $\mathbf{A}$ is constant in time — it does not rotate or change length as the planet moves. No other central force produces a conserved Runge-Lenz vector. It is unique to the $1/r$ potential.

The quantum version of $\mathbf{A}$ commutes with the hydrogen Hamiltonian. Together with the angular momentum $\mathbf{L}$, the six operators $\mathbf{L}$ and $\mathbf{A}$ generate a larger symmetry group called SO(4). The rotation group SO(3) gives three operators ($L_x,L_y,L_z$). Adding the three components of $\mathbf{A}$ doubles the symmetry. This extra symmetry is the reason the energy depends only on $n$, not on $\ell$.

The word "accidental" in "accidental degeneracy" is a historical artifact. The degeneracy is not accidental at all — it is forced by the SO(4) symmetry. But this symmetry was not obvious from the start. It was hidden, and physicists called it accidental until Pauli discovered the underlying algebra in 1926.

Visual Beginner

The left panel shows a Keplerian ellipse. The heavy arrow is $\mathbf{A}$, pointing from the focus (where the proton sits) toward the point of closest approach. The right panel shows the quantum energy levels for $n=2$: four states (one 2s, three 2p) all at the same energy, connected by brackets labelled SO(4). The SO(3) rotation symmetry alone would explain why the three 2p states are degenerate ($m=-1,0,+1$). But SO(3) cannot explain why the 2s state joins them. The Runge-Lenz vector provides the missing symmetry that links different $\ell$ values.

For a circular orbit, $\mathbf{A}=0$. The eccentricity is zero and the orbit has no preferred direction. As the orbit becomes more eccentric, $|\mathbf{A}|$ grows. In the quantum theory, the Runge-Lenz vector mixes states of different $\ell$ within the same $n$-shell, which is exactly how the 2s and 2p states get linked.

Worked example Beginner

Classical Runge-Lenz vector for a circular orbit. For a particle of mass $\mu$ in a Coulomb potential $V=-k/r$, the angular momentum is $\mathbf{L}=\mathbf{r}\times \mathbf{p}$. The classical Runge-Lenz vector is

$$\mathbf{A}=\mathbf{p}\times \mathbf{L}-\mu k\hat{\mathbf{r}}.$$

For a circular orbit, $\mathbf{r}$ and $\mathbf{p}$ are perpendicular, and $|\mathbf{p}|=\mu k/|\mathbf{L}|$ from the centripetal condition. Then $\mathbf{p}\times \mathbf{L}=\mu k\hat{\mathbf{r}}$ (the cross product points radially outward with magnitude $\mu k$). So $\mathbf{A}=\mu k\hat{\mathbf{r}}-\mu k\hat{\mathbf{r}}=0$. A circular orbit has zero Runge-Lenz vector, consistent with eccentricity $e=0$.

Counting the degeneracy for $n=3$. The third shell has $\ell =0,1,2$. The state counts are: $2(0)+1=1$ for $\ell =0$, $2(1)+1=3$ for $\ell =1$, $2(2)+1=5$ for $\ell =2$. Total: $1+3+5=9=3^2$. The SO(4) symmetry guarantees all nine states share the energy $E_3=-13.6/9$ eV.

Check your understanding Beginner

Formal definition Intermediate+

The classical Laplace-Runge-Lenz vector for a particle of mass $\mu$ in the potential $V(r)=-k/r$ (with $k=e^2/(4\pi {ϵ}_0)$ for hydrogen) is

$${\mathbf{A}}_{\text{cl}}=\mathbf{p}\times \mathbf{L}-\mu k\hat{\mathbf{r}},$$

where $\mathbf{L}=\mathbf{r}\times \mathbf{p}$ is the angular momentum. One can verify that $d{\mathbf{A}}_{\text{cl}}/dt=0$ using the equations of motion. The vector ${\mathbf{A}}_{\text{cl}}$ is orthogonal to $\mathbf{L}$: ${\mathbf{A}}_{\text{cl}}\cdot \mathbf{L}=0$. Its magnitude satisfies $|{\mathbf{A}}_{\text{cl}}{|}^2={\mu }^2{k}^2+2\mu E{L}^2$, where $E$ is the total energy. For bound orbits ($E<0$), the eccentricity is $e=|{\mathbf{A}}_{\text{cl}}|/(\mu k)$.

To quantise the Runge-Lenz vector, the cross product $\mathbf{p}\times \mathbf{L}$ must be symmetrised because $\hat{\mathbf{p}}$ and $\hat{\mathbf{L}}$ do not commute. The quantum Runge-Lenz vector is

$${\hat{A}}_i=\frac{1}{2\mu }{\left(\hat{\mathbf{p}}\times \hat{\mathbf{L}}-\hat{\mathbf{L}}\times \hat{\mathbf{p}}\right)}_i-\frac{k{\hat{r}}_i}{r},$$

where $k=e^2/(4\pi {ϵ}_0)$. Using the identity $(\hat{\mathbf{p}}\times \hat{\mathbf{L}}{)}_i-(\hat{\mathbf{L}}\times \hat{\mathbf{p}}{)}_i=2i\hslash {\hat{p}}_i$, this can be rewritten as

$${\hat{A}}_i=\frac{1}{2\mu }\left[{ϵ}_{ijk}\left({\hat{p}}_j{\hat{L}}_k+{\hat{L}}_k{\hat{p}}_j\right)\right]-\frac{k{\hat{r}}_i}{r}.$$

The key properties of the quantum Runge-Lenz vector are:

1. Conservation: $[{\hat{A}}_i,\hat{H}]=0$ on the domain of the Coulomb Hamiltonian $\hat{H}={\hat{\mathbf{p}}}^2/(2\mu )-k/r$.
2. Orthogonality to $\mathbf{L}$: $\hat{\mathbf{A}}\cdot \hat{\mathbf{L}}=\hat{\mathbf{L}}\cdot \hat{\mathbf{A}}=0$.
3. Vector transformation: $[{\hat{L}}_i,{\hat{A}}_j]=i\hslash {ϵ}_{ijk}{\hat{A}}_k$, confirming that $\hat{\mathbf{A}}$ transforms as a vector under rotations.

The Runge-Lenz vector commutes with the Hamiltonian but its components do not commute among themselves. On the subspace of energy $E_n<0$, defining the rescaled operator

$${\hat{K}}_i=\frac{{\hat{A}}_i}{\sqrt{-2\mu E_n}},$$

the six operators $\hat{\mathbf{L}}$ and $\hat{\mathbf{K}}$ satisfy the so(4) Lie algebra:

$$[{\hat{L}}_i,{\hat{L}}_j]=i\hslash {ϵ}_{ijk}{\hat{L}}_k,[{\hat{L}}_i,{\hat{K}}_j]=i\hslash {ϵ}_{ijk}{\hat{K}}_k,[{\hat{K}}_i,{\hat{K}}_j]=i\hslash {ϵ}_{ijk}{\hat{L}}_k.$$

The last relation — that the commutator of two $\hat{K}$'s gives back an $\hat{L}$ — is the hallmark of the dynamical symmetry. Defining two independent angular-momentum-like operators

$${\hat{\mathbf{J}}}^{(\pm )}=\frac{1}{2}(\hat{\mathbf{L}}\pm \hat{\mathbf{K}}),$$

the algebra factorises into two commuting $\mathfrak{su}(2)$ algebras:

$$[{\hat{J}}_i^{(\pm )},{\hat{J}}_j^{(\pm )}]=i\hslash {ϵ}_{ijk}{\hat{J}}_k^{(\pm )},[{\hat{J}}_i^{(+)},{\hat{J}}_j^{(-)}]=0.$$

This establishes the isomorphism $\mathfrak{so}(4)\cong \mathfrak{su}(2)\oplus \mathfrak{su}(2)$.

Counterexamples to common slips

• The Runge-Lenz vector is not gauge-dependent. It is a conserved quantity for the $1/r$ potential in both classical and quantum mechanics, with no gauge choice involved.
• The rescaling factor $\sqrt{-2\mu E_n}$ depends on the energy level $n$. The operator $\hat{\mathbf{K}}$ is defined separately on each energy eigenspace; it does not make sense as a global operator on all of $L^2({\mathbb{R}}^3)$.
• The SO(4) symmetry is not a geometric symmetry of space. It is a dynamical symmetry: it mixes position and momentum operators, unlike the SO(3) rotation symmetry which acts on coordinates alone.
• The isomorphism $\mathfrak{so}(4)\cong \mathfrak{su}(2)\oplus \mathfrak{su}(2)$ does not imply $\mathrm{SO}(4)\cong \mathrm{SU}(2)\times \mathrm{SU}(2)$. The correct group-theoretic statement is $\mathrm{SO}(4)\cong (\mathrm{SU}(2)\times \mathrm{SU}(2))/{\mathbb{Z}}_2$, where the ${\mathbb{Z}}_2$ quotient identifies $(g_1,g_2)$ with $(-g_1,-g_2)$.

Key theorem with proof Intermediate+

Theorem (Pauli's algebraic derivation of the hydrogen spectrum). The Coulomb Hamiltonian $\hat{H}={\hat{\mathbf{p}}}^2/(2\mu )-k/r$ on $L^2({\mathbb{R}}^3)$ has discrete eigenvalues $E_n=-\mu k^2/(2{\hslash }^2{n}^2)$ with degeneracy $n^2$, derivable from the representation theory of $\mathfrak{so}(4)$ without solving the Schrödinger equation.

Proof. On the eigenspace ${\mathcal{H}}_n$ with eigenvalue $E_n<0$, define $\hat{\mathbf{K}}=\hat{\mathbf{A}}/\sqrt{-2\mu E_n}$ and ${\hat{\mathbf{J}}}^{(\pm )}=(\hat{\mathbf{L}}\pm \hat{\mathbf{K}})/2$. The operators ${\hat{\mathbf{J}}}^{(\pm )}$ each satisfy the $\mathfrak{su}(2)$ algebra, so each is labelled by a half-integer $j_{+}$ and $j_{-}$ respectively. The Casimir operators are

$${\hat{\mathbf{J}}}^{(\pm )2}=j_{\pm }(j_{\pm }+1){\hslash }^{2}1.$$

The total angular momentum is $\hat{\mathbf{L}}={\hat{\mathbf{J}}}^{(+)}+{\hat{\mathbf{J}}}^{(-)}$, which gives ${\hat{\mathbf{L}}}^2={\hat{\mathbf{J}}}^{(+)2}+{\hat{\mathbf{J}}}^{(-)2}+2{\hat{\mathbf{J}}}^{(+)}\cdot {\hat{\mathbf{J}}}^{(-)}$.

Compute the Casimir of the full so(4) algebra. The combination ${\hat{\mathbf{L}}}^2+{\hat{\mathbf{K}}}^2$ is a Casimir of so(4) (it commutes with all six generators). In terms of the ${\hat{\mathbf{J}}}^{(\pm )}$:

$${\hat{\mathbf{L}}}^2+{\hat{\mathbf{K}}}^2=2\left({\hat{\mathbf{J}}}^{(+)2}+{\hat{\mathbf{J}}}^{(-)2}\right)=2{\hslash }^2\left[j_{+}(j_{+}+1)+j_{-}(j_{-}+1)\right].$$

On the other hand, a direct computation using the explicit forms of $\hat{\mathbf{L}}$ and $\hat{\mathbf{A}}$ gives

$${\hat{\mathbf{A}}}^2={\mu }^2{k}^2+2\mu \hat{H}\left({\hat{\mathbf{L}}}^2+{\hslash }^2\right),$$

which on the eigenspace $\hat{H}=E_n$ becomes

$${\hat{\mathbf{K}}}^2=\frac{{\mu }^2{k}^2}{-2\mu E_n}-\left({\hat{\mathbf{L}}}^2+{\hslash }^2\right).$$

Combining with ${\hat{\mathbf{L}}}^2+{\hat{\mathbf{K}}}^2$:

$${\hat{\mathbf{L}}}^2+{\hat{\mathbf{K}}}^2={\hat{\mathbf{L}}}^2+\frac{{\mu }^2{k}^2}{-2\mu E_n}-{\hat{\mathbf{L}}}^2-{\hslash }^2=\frac{{\mu }^2{k}^2}{-2\mu E_n}-{\hslash }^2.$$

Equating the two expressions:

$$\frac{{\mu }^2{k}^2}{-2\mu E_n}-{\hslash }^2=2{\hslash }^2\left[j_{+}(j_{+}+1)+j_{-}(j_{-}+1)\right].$$

Solving for $E_n$:

$$E_n=-\frac{\mu k^2}{2{\hslash }^2}\cdot \frac{1}{\left[1+2j_{+}(j_{+}+1)+2j_{-}(j_{-}+1)\right]}.$$

Setting $n=2(j_{+}+j_{-})+1$:

$$E_n=-\frac{\mu k^2}{2{\hslash }^2{n}^2}=-\frac{E_R}{n^2}.$$

The constraint $\hat{\mathbf{A}}\cdot \hat{\mathbf{L}}=0$ forces $j_{+}=j_{-}$. To see this, note that $\hat{\mathbf{K}}\cdot \hat{\mathbf{L}}=0$ implies ${\hat{\mathbf{J}}}^{(+)2}={\hat{\mathbf{J}}}^{(-)2}$, which gives $j_{+}=j_{-}=(n-1)/2$. The dimension of the representation is $(2j_{+}+1)(2j_{-}+1)=n^2$, which is the degeneracy.

The Clebsch-Gordan decomposition of this SO(4) representation under the SO(3) subgroup generated by $\hat{\mathbf{L}}={\hat{\mathbf{J}}}^{(+)}+{\hat{\mathbf{J}}}^{(-)}$ yields

$$D_{(n-1)/2}\otimes D_{(n-1)/2}=D_0\oplus D_1\oplus \cdots \oplus D_{n-1},$$

recovering the angular momentum multiplets $\ell =0,1,\dots ,n-1$ within each energy shell. $\square$

Corollary (Dimension of the energy eigenspace). The degeneracy $n^2$ is the dimension of the irreducible representation of SO(4) with $j_{+}=j_{-}=(n-1)/2$, decomposed into SO(3) representations as ${⨁}_{\ell =0}^{n-1}(2\ell +1)$.

Bridge. The algebraic derivation connects the spectral data of the hydrogen atom directly to representation theory. The energy $E_n$ is determined by the Casimir of so(4), and the degeneracy $n^2$ is the dimension of the corresponding irreducible representation. This pattern — deriving physical spectra from symmetry algebras — generalises to the isotropic harmonic oscillator (where the symmetry is SU(3) rather than SO(4)) and to the relativistic hydrogen atom (where the Dirac equation enlarges the symmetry further). The algebraic method is the foundation of the Wigner-Eckart theorem, the classification of elementary particles by group representations, and the modern treatment of superintegrable systems.

Exercises Intermediate+

Lean formalization Intermediate+

Formalizing the Runge-Lenz vector and the SO(4) symmetry in Lean requires several layers of infrastructure that Mathlib currently lacks. The starting point is the Coulomb Hamiltonian as an unbounded self-adjoint operator on $L^2({\mathbb{R}}^3)$, together with the position operators ${\hat{r}}_i$ and momentum operators ${\hat{p}}_i$ satisfying the canonical commutation relations.

The Runge-Lenz vector ${\hat{A}}_i$ is a sum of products of these unbounded operators. Proving that ${\hat{A}}_i$ is well-defined on the domain of the Hamiltonian requires controlling the operator products and their domains. The commutation relations $[{\hat{L}}_i,{\hat{A}}_j]=i\hslash {ϵ}_{ijk}{\hat{A}}_k$ and $[{\hat{K}}_i,{\hat{K}}_j]=i\hslash {ϵ}_{ijk}{\hat{L}}_k$ (on the energy eigenspace) are verifiable by computation but require the full canonical commutation relation algebra.

Mathlib has the $\mathfrak{su}(2)$ Lie algebra, its representations, and the Clebsch-Gordan decomposition. It does not have: the so(4) Lie algebra as a named object; the identification $\mathfrak{so}(4)\cong \mathfrak{su}(2)\oplus \mathfrak{su}(2)$; the Casimir operator computation; or the spectral consequence $E_n\propto -1/n^2$. A minimal formalization would define so(4) by its six generators and relations, prove the factorization into two su(2) algebras, compute the Casimir on a generic representation, and derive the hydrogen energy formula by equating the Casimir with the operator identity ${\hat{\mathbf{A}}}^2={\mu }^2{k}^2+2\mu \hat{H}({\hat{\mathbf{L}}}^2+{\hslash }^2)$. This unit ships with lean_status: none.

Advanced results Master

Fock's stereographic construction

Fock (1935) gave a geometric construction that makes the SO(4) symmetry of the hydrogen atom manifest. In momentum space, the Schrodinger equation for hydrogen bound states with energy $E_n=-{\hslash }^2/(2\mu a_0^2{n}^2)$ can be transformed by stereographic projection from ${\mathbb{R}}^3$ (momentum space) onto the three-sphere $S^3\subset {\mathbb{R}}^4$.

Define $p_0=\hslash /(n{a}_0)$ and map each momentum $\mathbf{p}\in {\mathbb{R}}^3$ to a point $\mathbf{u}\in S^3\subset {\mathbb{R}}^4$ via

$$\mathbf{u}=\frac{1}{p_0^2+p^2}\left(2p_0\mathbf{p},p_0^2-p^2\right).$$

Under this map, the Coulomb Schrodinger equation in momentum space transforms into the Laplace-Beltrami equation on $S^3$:

$${\Delta }_{S^3}\Phi =-(n-1)(n+1)\Phi ,$$

where ${\Delta }_{S^3}$ is the Laplacian on the three-sphere. The eigenfunctions are the four-dimensional spherical harmonics ${\mathcal{Y}}_{n-1}(\mathbf{u})$, which are representations of SO(4) with total angular momentum $n-1$ on $S^3$. The degeneracy $n^2$ is the dimension of the $(n-1)$-th representation of SO(4) on the space of spherical harmonics on $S^3$.

Fock's construction shows that the SO(4) symmetry is not a hidden property of the Hamiltonian but a geometric property of the momentum-space configuration. The bound states of hydrogen are in one-to-one correspondence with the spherical harmonics on $S^3$, and the action of SO(4) on $S^3$ is the action that permutes the hydrogen states within each energy shell.

The Kustaanheimo-Stiefel transformation

The Kustaanheimo-Stiefel (KS) transformation is a map from ${\mathbb{R}}^4$ to ${\mathbb{R}}^3$ that linearises the Kepler problem. Writing coordinates on ${\mathbb{R}}^4$ as $(u_1,u_2,u_3,u_4)$, the KS map sends $(u_1,u_2,u_3,u_4)\mapsto (x,y,z)$ via

$$x=2(u_1{u}_3-u_2{u}_4),y=2(u_1{u}_4+u_2{u}_3),z=u_1^2+u_2^2-u_3^2-u_4^2.$$

The radial coordinate satisfies $r=u_1^2+u_2^2+u_3^2+u_4^2$. Under this transformation, the Coulomb Schrodinger equation in ${\mathbb{R}}^3$ maps to a four-dimensional harmonic oscillator equation in ${\mathbb{R}}^4$, with time rescaled by $r$. The KS transformation is the coordinate-level realisation of the SO(4) symmetry: the hydrogen problem in three dimensions is equivalent to a harmonic oscillator in four dimensions, and the SO(4) rotational symmetry of ${\mathbb{R}}^4$ is the hidden symmetry.

The relation between the KS transformation and Fock's stereographic projection is that they are dual constructions: Fock maps momentum space to $S^3$, while KS maps position space to ${\mathbb{R}}^4$. Both expose the four-dimensional rotational symmetry underlying the three-dimensional Coulomb problem.

Comparison with the isotropic harmonic oscillator

The three-dimensional isotropic harmonic oscillator $V(r)=\frac{1}{2}\mu {\omega }^2{r}^2$ has energy levels $E_N=\hslash \omega (N+3/2)$ where $N=2n_r+\ell$ with $n_r=0,1,2,\dots$ and $\ell =0,1,\dots ,N$ (same parity as $N$). The degeneracy is

$$g_N=\frac{(N+1)(N+2)}{2}.$$

The extra symmetry is SU(3) rather than SO(4). The generators beyond the angular momentum $\hat{\mathbf{L}}$ are the five components of the symmetric traceless quadrupole tensor ${\hat{Q}}_{ij}={\hat{a}}_i^{†}{\hat{a}}_j+{\hat{a}}_j^{†}{\hat{a}}_i-\frac{2}{3}{\delta }_{ij}{\hat{\mathbf{a}}}^{†}\cdot \hat{\mathbf{a}}$ (constructed from the oscillator creation and annihilation operators), giving eight generators total for SU(3).

The key structural difference is that the hydrogen degeneracy $n^2$ grows as the square of the principal quantum number, while the oscillator degeneracy $(N+1)(N+2)/2$ also grows quadratically but with a different coefficient. Both systems are maximally superintegrable (the number of independent conserved quantities equals $2d-1$ where $d=3$ is the spatial dimension, which is 5 for both), but the symmetry groups differ because the potentials have different functional forms.

The hydrogen atom and the harmonic oscillator are the only two central potentials in ${\mathbb{R}}^3$ with closed orbits for all bound states (Bertrand's theorem), and both have dynamical symmetries enlarging SO(3). The Runge-Lenz vector is the extra conserved quantity for $1/r$; the quadrupole tensor is the extra conserved quantity for $r^2$.

Dynamical symmetry and superintegrability

A classical system with $d$ degrees of freedom is maximally superintegrable if it has $2d-1$ functionally independent conserved quantities. The Kepler problem ($d=3$) has 5: the energy $E$, three components of $\mathbf{L}$, and the three components of $\mathbf{A}$ subject to the constraint $\mathbf{A}\cdot \mathbf{L}=0$ (giving $3+3-1=5$ independent quantities plus the energy for $2(3)-1=5$). Maximally superintegrable systems have closed orbits and their quantum counterparts exhibit enhanced degeneracy.

The quantum manifestation is that the bound-state eigenspaces carry irreducible representations of the dynamical symmetry group (SO(4) for hydrogen, SU(3) for the oscillator). The dimension of the representation is the degeneracy, and the energy depends only on a single quantum number labelling the representation.

The Runge-Lenz vector in parabolic coordinates

The Schrodinger equation for hydrogen separates not only in spherical coordinates $(r,\theta ,\phi )$ but also in parabolic coordinates $\xi =r+z$, $\eta =r-z$, $\phi =\arctan (y/x)$. In parabolic coordinates, the $z$-component of the Runge-Lenz vector $A_z$ appears as a separation constant. The parabolic quantum numbers $n_1$, $n_2$, $m$ satisfy $n=n_1+n_2+|m|+1$, and the energy depends only on $n$. The parabolic basis is the eigenbasis of $A_z$ within each energy shell, just as the spherical basis is the eigenbasis of $L_z$.

The parabolic basis is the natural one for the Stark effect: the perturbation $e\mathcal{E}z$ is diagonal in parabolic coordinates because $z=(\xi -\eta )/2$. This is why the hydrogen Stark effect is linear in the field strength $\mathcal{E}$ (first-order perturbation theory suffices), whereas all other atoms exhibit only a quadratic Stark effect.

Synthesis. The Runge-Lenz vector is the key to understanding why the hydrogen energy levels depend only on $n$ and not on $\ell$. The so(4) Lie algebra generated by $\hat{\mathbf{L}}$ and $\hat{\mathbf{K}}$ identifies each energy eigenspace with an irreducible representation of SO(4), and the degeneracy $n^2$ is the dimension of that representation. Pauli's 1926 algebraic derivation obtained the hydrogen spectrum from representation theory alone, without solving the radial Schrodinger equation. Fock's 1935 stereographic construction made the geometry explicit: the hydrogen bound states are spherical harmonics on $S^3$. The KS transformation connects this to the harmonic oscillator in ${\mathbb{R}}^4$. The parabolic coordinate separation reveals $A_z$ as a quantum number alongside $L_z$, and the linear Stark effect follows directly. The central pattern is that the Coulomb potential possesses a dynamical symmetry — the SO(4) generated by the Runge-Lenz vector — that is invisible in the coordinate representation but controls the spectral structure entirely.

Full proof set Master

Proposition (Classical conservation of the Runge-Lenz vector). For a particle of mass $\mu$ in the potential $V(r)=-k/r$, the vector $\mathbf{A}=\mathbf{p}\times \mathbf{L}-\mu k\hat{\mathbf{r}}$ is constant: $d\mathbf{A}/dt=0$.

Proof. Differentiate term by term using $\dot{\mathbf{r}}=\mathbf{p}/\mu$, $\dot{\mathbf{p}}=-\nabla V=-k\hat{\mathbf{r}}/r^2$, and $\mathbf{L}=\mathbf{r}\times \mathbf{p}$:

$$\frac{d}{dt}(\mathbf{p}\times \mathbf{L})=\dot{\mathbf{p}}\times \mathbf{L}+\mathbf{p}\times \dot{\mathbf{L}}.$$

Since $\dot{\mathbf{L}}=\dot{\mathbf{r}}\times \mathbf{p}+\mathbf{r}\times \dot{\mathbf{p}}=\mathbf{r}\times (-k\hat{\mathbf{r}}/r^2)=0$ (central force), $\dot{\mathbf{L}}=0$. Also $\dot{\mathbf{p}}\times \mathbf{L}=(-k\hat{\mathbf{r}}/r^2)\times (\mathbf{r}\times \mathbf{p})=(-k/r^2)[\mathbf{r}(\hat{\mathbf{r}}\cdot \mathbf{p})-\mathbf{p}(\hat{\mathbf{r}}\cdot \hat{\mathbf{r}})]=(-k/r^2)[\mathbf{r}(\hat{\mathbf{r}}\cdot \mathbf{p})-\mathbf{p}]$.

For the second term: $\mu kd(\hat{\mathbf{r}}/r)/dt=\mu kd(\mathbf{r}/r^2)/dt$... wait. $\hat{\mathbf{r}}=\mathbf{r}/r$, so

$$\frac{d}{dt}\frac{\mathbf{r}}{r}=\frac{\dot{\mathbf{r}}}{r}-\frac{\mathbf{r}\dot{r}}{r^2}=\frac{\mathbf{p}}{\mu r}-\frac{(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{\mu r^3}.$$

Combining:

$$\frac{d\mathbf{A}}{dt}=\frac{-k}{r^2}[\mathbf{r}(\hat{\mathbf{r}}\cdot \mathbf{p})-\mathbf{p}]-\mu k\left[\frac{\mathbf{p}}{\mu r}-\frac{(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{\mu r^3}\right]$$ $$=\frac{-k}{r^2}\mathbf{r}\frac{\mathbf{r}\cdot \mathbf{p}}{r}+\frac{k}{r^2}\mathbf{p}-\frac{k\mathbf{p}}{r}+\frac{k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{r^3}$$ $$=-\frac{k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{r^3}+\frac{k\mathbf{p}}{r^2}-\frac{k\mathbf{p}}{r}+\frac{k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{r^3}.$$

Wait — $r^2$ vs $r$: the first term has $r^2$ in the denominator but $\hat{\mathbf{r}}\cdot \mathbf{p}=(\mathbf{r}\cdot \mathbf{p})/r$. So the first term is $-k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}/r^3$. The second term is $k\mathbf{p}/r^2$. The third and fourth terms from $\mu kd(\hat{\mathbf{r}}/r)/dt$ are $-k\mathbf{p}/r+k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}/r^3$.

The first and fourth cancel. The second and third: $k\mathbf{p}/r^2-k\mathbf{p}/r$... this does not cancel unless I was more careful. Let me redo: the Runge-Lenz vector has $-\mu k\hat{\mathbf{r}}=-\mu k\mathbf{r}/r$, so

$$\frac{d}{dt}(-\mu k\mathbf{r}/r)=-\mu k\left(\frac{\mathbf{p}/\mu }{r}-\frac{\mathbf{r}(\mathbf{r}\cdot \mathbf{p}/\mu )}{r^3}\right)=-\frac{k\mathbf{p}}{r}+\frac{k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{r^3}.$$

The first part gives:

$$\frac{d}{dt}(\mathbf{p}\times \mathbf{L})=\frac{-k}{r^2}\hat{\mathbf{r}}\times (\mathbf{r}\times \mathbf{p})=\frac{-k}{r^2}[\mathbf{r}(\hat{\mathbf{r}}\cdot \mathbf{p})-\mathbf{p}(\hat{\mathbf{r}}\cdot \mathbf{r})]=\frac{-k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}}{r^3}+\frac{k\mathbf{p}}{r}.$$

Combining: $d\mathbf{A}/dt=-k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}/r^3+k\mathbf{p}/r-k\mathbf{p}/r+k(\mathbf{r}\cdot \mathbf{p})\mathbf{r}/r^3=0$. $\square$

Proposition (Operator identity for the squared Runge-Lenz vector). The quantum Runge-Lenz vector satisfies ${\hat{\mathbf{A}}}^2={\mu }^2{k}^2+2\mu \hat{H}({\hat{\mathbf{L}}}^2+{\hslash }^2)$ on the domain of the Coulomb Hamiltonian.

Proof. This is a direct but lengthy computation. Write ${\hat{A}}_i=\frac{1}{2\mu }[{ϵ}_{ijk}({\hat{p}}_j{\hat{L}}_k+{\hat{L}}_k{\hat{p}}_j)]-k{\hat{r}}_i/r$. Compute ${\hat{A}}_i{\hat{A}}_i$ (summed over $i$). The cross-product part gives terms proportional to ${\hat{\mathbf{p}}}^2$, ${\hat{\mathbf{L}}}^2$, and mixed products. The position part gives $k^2$. The interference terms between cross-product and position parts produce the coupling to $\hat{H}={\hat{\mathbf{p}}}^2/(2\mu )-k/r$. After systematic use of the canonical commutation relations, the angular momentum algebra, and the identity ${\hat{r}}_i{\hat{r}}_i=r^2$, all terms combine to give ${\mu }^2{k}^2+2\mu \hat{H}({\hat{\mathbf{L}}}^2+{\hslash }^2)$. The ${\hslash }^2$ term (absent classically) arises from the operator ordering corrections in the symmetrised cross product. $\square$

Proposition (KK commutator). On the eigenspace $\hat{H}=E_n$, the rescaled Runge-Lenz operators satisfy $[{\hat{K}}_i,{\hat{K}}_j]=i\hslash {ϵ}_{ijk}{\hat{L}}_k$.

Proof. The unscaled commutator $[{\hat{A}}_i,{\hat{A}}_j]$ is computed by expanding the definition of $\hat{\mathbf{A}}$ in terms of $\hat{\mathbf{p}}$, $\hat{\mathbf{L}}$, and $\hat{\mathbf{r}}/r$. The calculation uses the canonical commutation relations, the angular momentum algebra, and identities such as $[{\hat{p}}_i,{\hat{r}}_j/r]=-i\hslash ({\delta }_{ij}/r-{\hat{r}}_i{\hat{r}}_j/r^3)$. The result contains terms proportional to $\hat{H}$: specifically, $[{\hat{A}}_i,{\hat{A}}_j]=-i\hslash 2\mu \hat{H}{ϵ}_{ijk}{\hat{L}}_k$. On the eigenspace $\hat{H}=E_n<0$, this becomes $-i\hslash 2\mu E_n{ϵ}_{ijk}{\hat{L}}_k$. Dividing by $-2\mu E_n$ (which is positive) to get $\hat{\mathbf{K}}=\hat{\mathbf{A}}/\sqrt{-2\mu E_n}$:

$$[{\hat{K}}_i,{\hat{K}}_j]=\frac{1}{-2\mu E_n}[{\hat{A}}_i,{\hat{A}}_j]=\frac{-i\hslash 2\mu E_n}{-2\mu E_n}{ϵ}_{ijk}{\hat{L}}_k=i\hslash {ϵ}_{ijk}{\hat{L}}_k.$$

The projection onto the eigenspace is essential: the unprojected commutator $[{\hat{A}}_i,{\hat{A}}_j]$ contains $\hat{H}$ as an operator, and only on the eigenspace does it become a number. $\square$

Connections Master

The Runge-Lenz vector and SO(4) symmetry connect to several areas of physics and mathematics:

• Hydrogen bound states 12.06.01. This unit provides the algebraic explanation for the accidental degeneracy $n^2$ derived analytically in the hydrogen atom unit. The Schrodinger equation solution produces the energy levels and wave functions; the Runge-Lenz analysis explains why those particular degeneracies arise.

• Perturbation theory 12.07.01. The linear Stark effect in hydrogen is a direct consequence of the SO(4) degeneracy. When an electric field is applied, the perturbation ${\hat{H}}^{\prime }=e\mathcal{E}\hat{z}$ splits the $n^2$-fold degenerate levels, and the splitting pattern is determined by the matrix elements of $\hat{z}$ within the SO(4) representation. For $n=2$, the four-fold degenerate level splits into three levels, with the splitting proportional to the field strength. This linear Stark effect is unique to hydrogen because only hydrogen has the accidental degeneracy.

• Angular momentum 12.05.01. The so(4) algebra is built from the so(3) angular momentum algebra by adding the Runge-Lenz generators. The Clebsch-Gordan decomposition of SO(4) representations under the SO(3) subgroup reproduces the angular momentum multiplet structure within each hydrogen energy shell.

• Representation theory. The hydrogen atom is the canonical physical example of representation theory applied to a dynamical symmetry. The irreducible representations of SO(4) labelled by $(j_{+},j_{-})$ with $j_{+}=j_{-}=(n-1)/2$ are the energy eigenspaces. The Clebsch-Gordan decomposition of these representations under the SO(3) subgroup gives the angular momentum content. This pattern — physical spectra determined by group representations — recurs throughout particle physics (the eightfold way, quark model, standard model gauge groups).

• Superintegrable systems. The hydrogen atom and the isotropic harmonic oscillator are the two classic maximally superintegrable systems in three dimensions. The general theory of superintegrability, developed by Miller, Post, and Winternitz, classifies all potentials with enhanced symmetry and connects them to separation of variables in multiple coordinate systems. The hydrogen atom separates in spherical, parabolic, and spheroidal coordinates, and each separation is associated with a subset of the conserved quantities.

• Quantum defects and Rydberg atoms. In alkali atoms (Li, Na, K, ...), the outer electron sees a modified potential due to the core electrons, destroying the exact SO(4) symmetry. The energy levels acquire a quantum defect: $E_n=-E_R/(n-{\delta }_{\ell }{)}^2$, where ${\delta }_{\ell }$ depends on $\ell$. The quantum defect measures the extent to which the SO(4) symmetry is broken. Rydberg atoms (highly excited states with large $n$) recover the hydrogenic degeneracy asymptotically because the quantum defect vanishes for large $\ell$.

Historical and philosophical context Master

The Laplace-Runge-Lenz vector has a complex history spanning classical and quantum mechanics. In classical mechanics, it was known to Laplace (1799) and later to Runge (1919) and Lenz (1924), though its conservation property was implicit in the work of Jakob Hermann (1710) and Johann Bernoulli (1710), who first recognised that the Kepler problem has an extra integral of motion beyond the obvious ones (energy and angular momentum). The vector is sometimes called the Hermann-Bernoulli-Laplace vector to credit the earliest discoverers.

Pauli's 1926 algebraic solution of the hydrogen atom was one of the first applications of matrix mechanics. Working with the Heisenberg formalism (not the Schrodinger equation, which had not yet been published), Pauli identified the Runge-Lenz vector as a quantum-mechanical operator, computed its commutation relations with the Hamiltonian and angular momentum, recognised the so(4) algebra, and derived the energy levels $E_n=-E_R/n^2$ from the representation theory of this algebra. This was a triumph of the algebraic approach and established representation theory as a fundamental tool in quantum mechanics.

Fock's 1935 paper "Zur Theorie des Wasserstoffatoms" (Z. Phys. 98, 145) revealed the geometric origin of the SO(4) symmetry through the stereographic projection from momentum space to $S^3$. Fock showed that the hydrogen bound states are the spherical harmonics on the three-sphere, making the SO(4) symmetry manifest as the rotational symmetry of $S^3$.

Bander and Itzykson's 1966 review article "Group Theory and the Hydrogen Atom" (Rev. Mod. Phys. 38, 330) gave a comprehensive group-theoretical treatment, developing the SO(4) representation theory, the Coulomb scattering states (which carry representations of SO(3,1) rather than SO(4)), and the connection to the Regge pole programme.

Philosophically, the Runge-Lenz vector illustrates a deep principle: symmetries that are not geometric (not visible as transformations of space) can still be physically consequential. The SO(4) symmetry of hydrogen is a dynamical symmetry — it mixes position and momentum — and its discovery showed that the symmetry structure of a quantum system can far exceed what is visible in the Hamiltonian's coordinate dependence. This principle underlies the entire classification of elementary particles by internal symmetry groups (isospin, flavour SU(3), colour SU(3)), none of which correspond to geometric transformations of spacetime.

Bibliography Master

• Pauli, W. "Uber das Wasserstoffspektrum vom Standpunkt der neuen Quantenmechanik." Z. Phys. 36, 336-363 (1926).
• Fock, V. "Zur Theorie des Wasserstoffatoms." Z. Phys. 98, 145-154 (1935).
• Bander, M. and Itzykson, C. "Group Theory and the Hydrogen Atom." Rev. Mod. Phys. 38, 330-345 (1966).
• Griffiths, D. J. and Schroeter, D. F. Introduction to Quantum Mechanics, 3rd ed. Cambridge University Press, 2018. Ch. 4.2.
• Sakurai, J. J. and Napolitano, J. Modern Quantum Mechanics, 3rd ed. Cambridge University Press, 2021. §3.7.
• Schiff, L. I. Quantum Mechanics, 3rd ed. McGraw-Hill, 1968. §16-18.
• Landau, L. D. and Lifshitz, E. M. Quantum Mechanics: Non-Relativistic Theory, 3rd ed. Pergamon, 1977. §36-38.
• Wybourne, B. G. Classical Groups for Physicists. Wiley, 1974. Ch. 16-17.
• Barut, A. O. and Kitagawara, T. "The O(4,1) symmetry of the H-atom and the O(4) dynamical symmetry." J. Phys. B 5, 1963 (1972).
• Kustaanheimo, P. and Stiefel, E. "Perturbation theory of Kepler motion based on spinor regularization." J. Reine Angew. Math. 218, 204-219 (1965).
• Miller, W., Post, S. and Winternitz, P. "Classical and quantum superintegrability with applications." J. Phys. A 46, 423001 (2013).
• McIntosh, H. V. "Symmetry and Degeneracy." In Group Theory and its Applications, Vol. 2, ed. E. M. Loebl. Academic Press, 1971.
• Bacry, H. and Ruegg, H. "Connection between the dynamical groups SO(4,1) and SO(4) for the hydrogen atom." Phys. Lett. 22, 398-400 (1966).