4.4.3 Monopole Harmonics#

Prompts

In a monopole background, neither the canonical nor kinetic angular momentum is conserved alone. What is the conserved total angular momentum \(\hat{\boldsymbol{J}}\), and why does it satisfy the standard SU(2) algebra \([\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{J}_k\)?
The minimum angular momentum in a monopole field is \(j_{\min} = \vert s\vert\). Explain physically why this minimum is nonzero, and why no state with \(j < \vert s\vert\) can exist.
For a Dirac monopole (\(s = 1/2\)), the lowest monopole harmonics form a doublet with \(j = 1/2\) and \(m = \pm 1/2\). How does this relate to spin-1/2? What does this tell you about the origin of intrinsic angular momentum?
Connect the Berry curvature of spin-\(s\) to the monopole field in parameter space. What is the Chern number, and how does it relate to \(s\)?

Lecture Notes#

Overview#

When a charged particle moves on a sphere with a magnetic monopole at the center, a key subtlety arises: neither the canonical angular momentum \(\hat{\boldsymbol{L}}_{\mathrm{can}} = \boldsymbol{r} \times \hat{\boldsymbol{p}}\) nor the kinetic angular momentum \(\hat{\boldsymbol{\Lambda}} = \boldsymbol{r} \times \hat{\boldsymbol{\pi}}\) (where \(\hat{\boldsymbol{\pi}} = \hat{\boldsymbol{p}} - q\boldsymbol{A}\)) is separately conserved. The conserved quantity is the total angular momentum \(\hat{\boldsymbol{J}} = \hat{\boldsymbol{\Lambda}} + s\hat{\boldsymbol{r}}\), including the electromagnetic field’s angular momentum \(s\hat{\boldsymbol{r}}\). This \(\hat{\boldsymbol{J}}\) satisfies the standard SU(2) algebra \([\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{J}_k\), but the minimum eigenvalue is shifted: \(j \geq \vert s\vert\) instead of \(j \geq 0\). The eigenfunctions are monopole harmonics \(Y^s_{j,m}(\theta,\varphi)\). The punchline: for \(s = 1/2\) (a Dirac monopole), the lowest harmonics form a doublet with \(j = 1/2\) — they are spin-1/2 spinors. Half-integer angular momentum is not an independent axiom; it emerges from gauge topology.

Monopole Quantum Number#

Monopole Quantum Number

The monopole quantum number characterizes the coupling strength:

(176)#\[ s = \frac{qg}{4\pi\hbar} \]

The Dirac quantization condition \(qg = 2\pi n\hbar\) (from the two-patch construction of §4.4.2) restricts \(s\) to integer or half-integer values because \(s=n/2\):

\[ s = \frac{n}{2} \in \left\{0,\, \pm\tfrac{1}{2},\, \pm 1,\, \pm\tfrac{3}{2},\, \ldots\right\}. \]

For an electron in the field of a minimal Dirac monopole: \(\vert s\vert = 1/2\).

Convention: magnetic charge vs. monopole quantum number

In this section we adopt the flux-defined (Dirac-monopole) convention: the magnetic charge \(g\) is identified with the total flux of the monopole, so \(\boldsymbol{B} = g\hat{\boldsymbol{r}}/(4\pi r^2)\) and \(\oint\boldsymbol{B}\cdot\mathrm{d}\boldsymbol{S} = g\). The dimensionless quantum number seen by a particle of electric charge \(q\) is \(s\):

\[ s = \frac{qg}{4\pi\hbar}. \]

Dirac quantization says \(2s\in\mathbb{Z}\). Therefore the minimal Dirac monopole is not defined by setting \(g=1\) numerically; it is defined by the smallest nonzero allowed coupling, \(\vert s\vert=1/2\). Equivalently, a loop enclosing the full solid angle gives Berry/Aharonov-Bohm phase

\[\begin{split} \begin{split} \Phi_{\mathrm{Berry}} &= -s\Omega, \\ \Omega = 4\pi &\Rightarrow \Phi_{\mathrm{Berry}} = -4\pi s = -2\pi n. \end{split} \end{split}\]

This is why a minimal monopole has Chern number of magnitude \(\vert c_1\vert = 2\vert s\vert = 1\), while a monopole with \(s=1\) carries twice the minimal charge.

Total Angular Momentum#

In the monopole background, the kinetic angular momentum \(\hat{\boldsymbol{\Lambda}}\) does not commute with itself in the standard way because the gauge field \(\boldsymbol{A}\) depends on position. The physically conserved quantity is:

Total Angular Momentum in a Monopole Field

\[ \hat{\boldsymbol{J}} = \hat{\boldsymbol{\Lambda}} + s\hat{\boldsymbol{r}} \]

where \(\hat{\boldsymbol{\Lambda}} = \boldsymbol{r} \times \hat{\boldsymbol{\pi}}\) is the kinetic angular momentum and \(s\hat{\boldsymbol{r}}\) is the angular momentum carried by the electromagnetic field.

\(\hat{\boldsymbol{J}}\) satisfies the standard SU(2) commutation relations:

\[ [\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\,\epsilon_{ijk}\hat{J}_k \]

with eigenvalues:

\[\begin{split} \begin{split} \hat{J}^2 \vert s, j, m\rangle &= \hbar^2 j(j+1) \vert s, j, m\rangle, \\ \hat{J}_z \vert s, j, m\rangle &= \hbar m \vert s, j, m\rangle. \end{split} \end{split}\]

The monopole modifies only the allowed values of \(j\) (not \(j = 0, 1, 2, \ldots\)):

\[ j = \vert s\vert,\; \vert s\vert+1,\; \vert s\vert+2,\; \ldots \]

The shift \(j_{\min} = \vert s\vert\) has a direct physical interpretation: the electromagnetic field already carries angular momentum \(s\hat{\boldsymbol{r}}\). The total angular momentum cannot be less than this field contribution.

Derivation: \([\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{J}_k\)

Building-block identities. Two commutators are needed.

Ingredient 1. The kinetic angular momentum generates rotations of position, so it acts on the position components \(\hat{r}_j\) exactly as ordinary angular momentum does:

\[ [\hat{\Lambda}_i, \hat{r}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{r}_k. \]

This follows from \(\hat{\boldsymbol{\Lambda}} = \boldsymbol{r}\times\hat{\boldsymbol{\pi}}\) together with \([\hat{r}_a, \hat{\pi}_b] = \mathrm{i}\hbar\delta_{ab}\); the gauge potential drops out because \([\hat{r}_a, \hat{r}_b] = 0\) and \(\hat{r}_a\) commutes with \(\boldsymbol{A}(\hat{\boldsymbol{r}})\).

Ingredient 2. Unlike the field-free case, the kinetic angular momentum does not close on itself. In a magnetic field the kinetic momenta fail to commute (§4.1.2):

\[ [\hat{\pi}_i, \hat{\pi}_j] = \mathrm{i}\hbar q\,\epsilon_{ijk}B_k. \]

For the monopole field \(\boldsymbol{B} = g\hat{\boldsymbol{r}}/(4\pi r^2)\) (§4.4.2), restricted to the unit sphere \(r = 1\), this radial non-commutativity is set by

\[ q B_k \;\longrightarrow\; \frac{qg}{4\pi}\,\hat{r}_k \;=\; \hbar s\,\hat{r}_k, \]

using the monopole quantum number \(s = qg/(4\pi\hbar)\) from (176). Propagating this term through \(\hat{\boldsymbol{\Lambda}} = \boldsymbol{r}\times\hat{\boldsymbol{\pi}}\) gives the standard charge-monopole result: the kinetic angular momentum acquires an anomalous term proportional to \(\hat{\boldsymbol{r}}\),

\[ [\hat{\Lambda}_i, \hat{\Lambda}_j] = \mathrm{i}\hbar\epsilon_{ijk}(\hat{\Lambda}_k - s\,\hat{r}_k) = \mathrm{i}\hbar\epsilon_{ijk}\hat{\Lambda}_k - \mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k. \]

Bilinear assembly. With \(\hat{\boldsymbol{J}} = \hat{\boldsymbol{\Lambda}} + s\hat{\boldsymbol{r}}\), expand the commutator term by term:

\[\begin{split} \begin{split} [\hat{J}_i, \hat{J}_j] &= [\hat{\Lambda}_i + s\hat{r}_i,\; \hat{\Lambda}_j + s\hat{r}_j] \\ &= [\hat{\Lambda}_i, \hat{\Lambda}_j] + s[\hat{\Lambda}_i, \hat{r}_j] + s[\hat{r}_i, \hat{\Lambda}_j] + s^2[\hat{r}_i, \hat{r}_j]. \end{split} \end{split}\]

Evaluate each term using ingredients 1 and 2, with \([\hat{r}_i, \hat{\Lambda}_j] = -[\hat{\Lambda}_j, \hat{r}_i] = \mathrm{i}\hbar\epsilon_{ijk}\hat{r}_k\) and \([\hat{r}_i, \hat{r}_j] = 0\):

\[\begin{split} \begin{split} [\hat{\Lambda}_i, \hat{\Lambda}_j] &= \mathrm{i}\hbar\epsilon_{ijk}\hat{\Lambda}_k - \mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k, \\ s\,[\hat{\Lambda}_i, \hat{r}_j] &= +\,\mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k, \\ s\,[\hat{r}_i, \hat{\Lambda}_j] &= +\,\mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k, \\ s^2\,[\hat{r}_i, \hat{r}_j] &= 0. \end{split} \end{split}\]

Cancellation. The anomalous \(-\mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k\) from ingredient 2 and the two cross terms \(+\mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k\) sum to \(+\mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k\). Adding this to \(\mathrm{i}\hbar\epsilon_{ijk}\hat{\Lambda}_k\) reassembles \(\hat{J}_k = \hat{\Lambda}_k + s\hat{r}_k\):

\[ [\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{\Lambda}_k + \mathrm{i}\hbar s\,\epsilon_{ijk}\hat{r}_k = \mathrm{i}\hbar\epsilon_{ijk}(\hat{\Lambda}_k + s\hat{r}_k) = \mathrm{i}\hbar\epsilon_{ijk}\hat{J}_k. \]

The field contribution \(s\hat{\boldsymbol{r}}\) is exactly what is needed to absorb the anomalous term and restore the standard SU(2) algebra.

Monopole Harmonics#

The eigenfunctions of \(\hat{J}^2\) and \(\hat{J}_z\) in the monopole background are monopole harmonics:

Monopole Harmonics

\(Y^s_{j,m}(\theta,\varphi)\) are simultaneous eigenfunctions of \(\hat{J}^2\) and \(\hat{J}_z\):

\[\begin{split} \begin{split} \hat{J}^2 Y^s_{j,m} &= \hbar^2 j(j+1) Y^s_{j,m}, \\ \hat{J}_z Y^s_{j,m} &= \hbar m Y^s_{j,m}. \end{split} \end{split}\]

with \(j = \vert s\vert, \vert s\vert+1, \ldots\) and \(m = -j, \ldots, j\).

Orthonormality: \(\displaystyle\int \sin\theta\,\mathrm{d}\theta\,\mathrm{d}\varphi\,(Y^s_{j,m})^* Y^s_{j',m'} = \delta_{jj'}\delta_{mm'}\)

Reduction: For \(s = 0\), \(Y^0_{l,m} = Y_{lm}(\theta,\varphi)\) (ordinary spherical harmonics).

The explicit form of \(Y^s_{j,m}\) requires the two-patch construction (§4.4.2): the monopole gauge field cannot be described by a single vector potential on the whole sphere, so northern and southern hemispheres use separate gauges, related by a transition function on the equator.

Summary of how the monopole modifies the spectrum:

\(s\)	\(j_{\min}\)	Degeneracy at \(j_{\min}\)
\(0\) (no monopole)	\(0\)	\(1\) (singlet)
\(1/2\) (Dirac)	\(1/2\)	\(2\) (doublet = spin-1/2)
\(1\)	\(1\)	\(3\) (triplet)
\(3/2\)	\(3/2\)	\(4\) (quartet)

Punchline: Spin-1/2 from Monopole Topology#

Spin-1/2 Emerges from \(s = 1/2\)

For a Dirac monopole (\(\vert s\vert = 1/2\)), the minimum angular momentum is \(j_{\min} = 1/2\). The lowest states are:

\[ \left\{\, \vert \tfrac{1}{2}, +\tfrac{1}{2}\rangle,\ \vert \tfrac{1}{2}, -\tfrac{1}{2}\rangle \,\right\} \]

These form a doublet transforming exactly as spin-1/2: under a \(2\pi\) rotation, each state acquires a factor \(\mathrm{e}^{2\pi\mathrm{i}s} = \mathrm{e}^{\pi\mathrm{i}} = -1\) (the spinor sign flip).

Conclusion: Half-integer angular momentum is not an independent axiom. It is the minimum angular momentum enforced by the topology of the monopole gauge bundle.

Discussion: spinors from monopole quantization

For a Dirac monopole (\(s = 1/2\)), the minimum angular momentum level is \(j = 1/2\). In what sense are the states at this level “spinors” rather than “vectors”? What would change physically if the minimum were \(j = 0\) instead of \(j = 1/2\)? Could you design an experiment to observe the \(2\pi\) sign flip of monopole harmonics with \(s = 1/2\)? And more broadly: does the topological origin of half-integer angular momentum here suggest that spin itself might be “geometric” rather than an independent axiom?

Connection to Berry Phase#

The monopole structure of angular momentum has a direct physical manifestation through Berry phase (§4.2). For a spin-1/2 particle in a slowly rotating magnetic field \(\boldsymbol{B}(t)\), as \(\boldsymbol{B}\) traces a closed loop on the unit sphere in parameter space, the spin acquires a geometric phase \(\Phi_{\mathrm{Berry}} = -\frac{1}{2}\Omega\) where \(\Omega\) is the solid angle subtended. The corresponding Berry curvature is the radial monopole field

\[ \boldsymbol{F} = -\frac{1}{2}\frac{\hat{\boldsymbol{n}}}{\vert\boldsymbol{n}\vert^2} \]

whose flux through the enclosed cap reproduces \(\Phi_{\mathrm{Berry}} = -\frac{1}{2}\Omega\) by Stokes’ theorem — the same curvature \(\boldsymbol{\Omega} = -\tfrac{1}{2}\hat{\boldsymbol{n}}\) derived for spin-1/2 in §4.2.1.

Spin = Monopole in Parameter Space

The Berry curvature of a spin-\(s\) system is the field of a monopole at the origin of parameter space — of strength \(-s\) for the aligned eigenstate (§4.2.1 finds strength \(-1/2\) for spin-1/2; the sign depends on which eigenstate is transported, the magnitude does not). This is not an analogy — it is the same mathematical structure.

The Chern number — the curvature flux through any closed surface enclosing the degeneracy, divided by \(2\pi\) — is the integer topological invariant

\[ \vert c_1\vert = \frac{1}{2\pi}\left\vert\,\oint \boldsymbol{F}\cdot\mathrm{d}\boldsymbol{S}\,\right\vert = 2\vert s\vert \]

This closes the circle: spin-1/2 emerges from monopole topology (§4.4.2), and the Berry phase of spin-1/2 is itself a monopole field (§4.2.1).

Poll: Monopole harmonics vs. spherical harmonics

On a sphere in the presence of a magnetic monopole at the origin, the allowed quantum states are labeled by monopole harmonics (or monopole spherical harmonics) rather than ordinary spherical harmonics. The key difference is that monopole harmonics are eigenstates of a modified angular momentum operator that includes the effect of the monopole. Which statement correctly describes this modification?

(A) The monopole shifts the origin of the angular momentum algebra, changing the commutation relations.

(B) The covariant angular momentum operator includes a term proportional to the vector potential of the monopole, reflecting the coupling of orbital angular momentum to the gauge field.

(C) The monopole adds a constant \(s\hbar\) to every angular momentum eigenvalue, rigidly shifting the whole spectrum upward.

(D) The angular momentum operator is unchanged; the monopole only alters the radial wavefunction, so the angular eigenfunctions remain ordinary spherical harmonics.

Summary#

The conserved total angular momentum in a monopole field is \(\hat{\boldsymbol{J}} = \hat{\boldsymbol{\Lambda}} + s\hat{\boldsymbol{r}}\), satisfying standard SU(2): \([\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{J}_k\).
Monopole harmonics \(Y^s_{j,m}\) are eigenstates of \(\hat{J}^2\) and \(\hat{J}_z\) with eigenvalues \(\hbar^2 j(j+1)\), but \(j_{\min} = \vert s\vert\) instead of 0.
The shift \(j_{\min} = \vert s\vert\) reflects the angular momentum already carried by the electromagnetic field.
For \(s = 1/2\) (Dirac monopole): the lowest harmonics form a spin-1/2 doublet — half-integer angular momentum emerges from topology, not as an axiom.
Berry connection: the Berry curvature of spin-\(s\) is a monopole of strength \(-s\) in parameter space — spin is topological.

Homework#

1. Zero-monopole limit. Show that for monopole quantum number \(s = 0\), the monopole harmonics \(Y^0_{j,m}(\theta,\varphi)\) reduce to ordinary spherical harmonics \(Y_{lm}(\theta,\varphi)\), and the minimum angular momentum is \(j_{\min} = 0\). Explain why a nonzero \(s\) shifts the minimum \(j\).

2. Completeness and spectrum. The monopole harmonics satisfy orthonormality \(\int \sin\theta\,\mathrm{d}\theta\,\mathrm{d}\varphi\,(Y^s_{j,m})^* Y^s_{j',m'} = \delta_{jj'}\delta_{mm'}\).

(a) The completeness relation sums \(j\) from \(\vert s\vert\), not from \(0\). Explain why the states with \(j < \vert s\vert\) are absent from the spectrum.

(b) How many angular momentum states exist for \(j \leq J_{\max}\)? Express the total count in terms of \(J_{\max}\) and \(s\), and compare to the free-space count \((J_{\max}+1)^2\).

3. Restoring SU(2). The kinetic angular momentum \(\hat{\boldsymbol{\Lambda}}\) in a monopole field does not satisfy the standard SU(2) algebra. The total angular momentum \(\hat{\boldsymbol{J}} = \hat{\boldsymbol{\Lambda}} + s\hat{\boldsymbol{r}}\) does.

(a) The commutator \([\hat{\Lambda}_i, \hat{\Lambda}_j]\) has an extra term \(-\mathrm{i}\hbar s\,\epsilon_{ijk}\hat{\boldsymbol r}_k\) compared to the standard algebra. Show that adding \(s\hat{\boldsymbol{r}}\) cancels this anomaly.

(b) Verify that \([\hat{J}_i, \hat{J}_j] = \mathrm{i}\hbar\epsilon_{ijk}\hat{J}_k\).

(c) Explain physically why \(\hat{\boldsymbol{J}}\), not \(\hat{\boldsymbol{\Lambda}}\), gives the good quantum numbers for a particle in a monopole field.

4. Spectrum enumeration. For each monopole quantum number \(s\) below, list the first three allowed values of \(j\) and state the degeneracy \(2j+1\) of the lowest level.

(a) \(s = 0\) (no monopole)

(b) \(s = 1/2\) (minimal Dirac monopole)

(d) \(s = 3/2\)

5. Spin from a monopole. For \(s = 1/2\), the lowest angular momentum level has \(j = 1/2\) with states \(m = \pm 1/2\).

(a) Compare the quantum numbers of this lowest multiplet to those of a spin-1/2 particle. How many states are there?

(b) Under a \(2\pi\) rotation, monopole harmonics with quantum number \(s\) acquire a phase \(\mathrm{e}^{2\pi\mathrm{i}s}\). Compute this phase for \(s = 1/2\). What familiar property of spinors does this reproduce?

(c) Explain why this result supports the claim that half-integer angular momentum can be understood as orbital motion in a monopole background, without invoking intrinsic spin as a separate postulate.

6. Dirac condition. The Dirac quantization condition \(qg = 2\pi n\hbar\) determines the monopole quantum number \(s = qg/(4\pi\hbar)\).

(a) For the minimal monopole (\(n = 1\)) and an electron (\(q = -e\)), compute \(\vert s\vert\) and the minimum angular momentum \(j_{\min}\).

(b) For a Cooper pair (\(q = -2e\)), what are \(\vert s\vert\) and \(j_{\min}\)? Relate this to flux quantization in superconductors (\(\Phi_0 = h/(2e)\)).

(c) If quarks with charge \(q = e/3\) existed as free particles, what monopole charge would be needed for the Dirac condition? How would the angular momentum spectrum change?

7. Berry phase and Chern number. The Berry connection for a spin-\(s\) system on the Bloch sphere has Chern number of magnitude \(\vert c_1\vert = 2\vert s\vert\).

(a) For \(s = 1/2\), verify that integrating the Berry curvature component \(F_{\theta\varphi} = -\frac{1}{2}\sin\theta\) over the unit sphere gives \(\vert c_1\vert = 1\).

(b) For \(s = 1\), what is the Berry curvature and the Chern number magnitude? Compare to the spin-1/2 case.

(c) The Chern number must be an integer. Explain how this requirement, applied to the monopole harmonics, is equivalent to the Dirac quantization condition.

8. Monopole harmonics versus spherical. Compare the monopole harmonics \(Y^s_{j,m}\) with ordinary spherical harmonics \(Y_{lm}\) to understand the effect of the monopole.

(a) For \(s = 0, 1/2, 1\), write down the quantum numbers \((j, m)\) for the lowest two values of \(j\).

(b) The ordinary spherical harmonics have a node structure: \(Y_{lm}\) has \(l - \vert m\vert\) nodes in \(\theta\). How is the node structure of \(Y^s_{j,m}\) modified by a nonzero \(s\)?

(c) In the limit \(s \to \infty\) (strong monopole), the lowest Landau level on the sphere becomes highly degenerate. Relate the number of states in the \(j = s\) multiplet to the number of flux quanta through the sphere, and connect this to the quantum Hall effect on a sphere (Haldane sphere).