Orbital Angular Momentum Operators

A new chapter opens with an old friend. Classical mechanics crowned $\mathbf{L} = \mathbf{r}\times\mathbf{p}$ its most durable conserved vector — the quantity that keeps planets in their orbital planes and flywheels stubborn. Quantum mechanics inherits the formula intact: replace position and momentum by their operators, and the cross product $\hat{\mathbf{L}} = \hat{\mathbf{r}}\times\hat{\mathbf{p}}$ is a legitimate trio of observables, no ordering headaches, hermitian out of the box. What changes is not the formula but the job description. In quantum mechanics $\hat{\mathbf{L}}$ turns out to be the generator of rotations — the operator that cranks a wavefunction around an axis, exactly as Canonical Commutation Relation showed momentum translating wavefunctions along a line. That reading organizes the whole chapter. Written in spherical coordinates, $\hat{\mathbf{L}}$ loses every trace of the radius: $\hat L_z = -i\hbar\,\partial_\phi$ and $\hat L^2$ touch only the angles. Angular momentum is the physics of orientation on the sphere, blind to how far out the particle sits — the fact that will crack every central-force problem from hydrogen onward. This article builds the operators and earns three results with full derivations: the azimuthal form $\hat L_z = -i\hbar\,\partial_\phi$ , the operator identity $\hat L^2 = \hat r^2\hat p^2 - (\hat{\mathbf{r}}\cdot\hat{\mathbf{p}})^2 + i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ with its telltale quantum ordering term, and the rotation law $e^{-i\alpha\hat L_z/\hbar}f(\phi) = f(\phi-\alpha)$ . The famous non-commuting algebra these operators obey is stated here in one line and owned, in full, by the next article.

Key takeaways

Orbital angular momentum is $\hat{\mathbf{L}} = \hat{\mathbf{r}}\times\hat{\mathbf{p}}$ — the classical formula promoted verbatim; each component pairs only commuting factors, so there is no ordering ambiguity and hermiticity is automatic.
In spherical coordinates $\hat{\mathbf{L}}$ contains no radial part: $\hat L_z = -i\hbar\,\partial_\phi$ , and $\hat L^2$ is a pure angular differential operator — angular momentum probes orientation, never distance.
$\hat L_z$ generates rotations about $z$ : $e^{-i\alpha\hat L_z/\hbar}f(\phi) = f(\phi - \alpha)$ , the rotational twin of momentum's translation law.
The operator identity $\hat L^2 = \hat r^2\hat p^2 - (\hat{\mathbf{r}}\cdot\hat{\mathbf{p}})^2 + i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ carries a genuine quantum correction term — machine-checked to fail without it.
Components do not commute — $[\hat L_x,\hat L_y] = i\hbar\hat L_z$ — a one-line preview here; the algebra and all its consequences are owned by Angular Momentum Commutation Relations.
Key idea: $\hat{\mathbf{L}} = \hat{\mathbf{r}}\times\hat{\mathbf{p}}$ is the generator of spatial rotations — an angles-only operator whose eigenvalue structure, built in the coming articles, is the universal grammar of atoms.

The core insight

Orbital angular momentum $\hat{\mathbf{L}} = \hat{\mathbf{r}}\times\hat{\mathbf{p}}$ is the generator of spatial rotations: a hermitian vector operator that acts only on angles, cranks wavefunctions around axes the way momentum slides them along lines, and whose component algebra — previewed here, derived next — forces the quantization the rest of the chapter harvests.

Where this lives today

The purest modern laboratory for orbital angular momentum is light. A laser mode with azimuthal phase $e^{im\phi}$ — a "vortex beam", dark on axis — carries exactly $m\hbar$ of orbital angular momentum per photon: this article's winding-phase eigenstates written in a beam profile. Such beams twist particles in optical tweezers (an absorbed 1 mW, $m=2$ beam at 633 nm delivers $3.2\times10^{15}$ photons and $6.7\times10^{-19}$ N·m of torque per second — tiny, and routinely measured), and the unbounded $m$ -ladder now multiplexes communication channels: each winding number is an independent lane in one fiber. Electron microscopy caught up in 2010 with electron vortex beams probing magnetic materials at atomic resolution; cold-atom rings realize Example 1's states literally, reading persistent currents as winding numbers. Wherever a wave circulates, $-i\hbar\,\partial_\phi$ is the meter that counts it.

The physical picture

Picture the wavefunction as a pattern painted on a sphere around the origin. Angular momentum measures how the pattern circulates — and $\hat{\mathbf{L}}$ is, literally, the crank that rotates it. A state with angular momentum about $z$ has a phase winding around the axis like a screw thread; $\hat L_z$ reads off the winding rate, and exponentiating it turns the whole pattern bodily.

Figure

The cross product r×p and the circulation it measures — the crank that rotates wavefunction patterns

Two features of the picture do real work. First, the radius never enters: rotating a pattern slides it along circles of constant $r$ , so the generator cannot contain $\partial_r$ — and the derivations below confirm $\hat L_z$ and $\hat L^2$ are pure angle operators. That is why central-potential problems separate: the angular machinery is universal, and Separation of Variables in Spherical Coordinates will reuse it wholesale. Second, rotations about different axes interfere — rotate a book $90°$ about $x$ then $y$ , then reverse the order, and you hold two differently oriented books. The generators inherit that friction: $[\hat L_x, \hat L_y] = i\hbar \hat L_z$ . This article verifies the statement on a test function and stops; the derivation and everything it forbids and forces belong to Angular Momentum Commutation Relations, next.

Figure

L acts only on angles — the radial direction is a spectator

Building the operator

Start from the classical components and promote. The cross product in components reads $L_z = xp_y - yp_x$ and cyclic permutations, so the quantum definition is the triple

\hat L_x = \hat y\hat p_z - \hat z\hat p_y, \qquad \hat L_y = \hat z\hat p_x - \hat x\hat p_z, \qquad \hat L_z = \hat x\hat p_y - \hat y\hat p_x \qquad (1)

where: each term multiplies a position by a momentum of a different axis — the whole reason the promotion is painless. Quantizing classical products usually stumbles on ordering ( $xp_x$ vs $p_x x$ ; Canonical Commutation Relation owns that story), but the cross product's structure guarantees no coordinate ever meets its own conjugate momentum. Since $[\hat x, \hat p_y] = 0$ , every ordering is the same operator; the verifier confirms symmetrizing changes nothing. Hermiticity comes free: a product of commuting hermitian operators is hermitian — $(\hat x\hat p_y)^\dagger = \hat p_y\hat x = \hat x\hat p_y$ — so each $\hat L_i$ passes the reality test of Hermitian Operators and Observables with no repair work (contrast the fixes that article performed on $d/dx$ ). Three legitimate observables with real spectra guaranteed; what the spectra are is the payoff of Ladder Operators and the Angular Momentum Spectrum, two serials ahead. One line of algebra deserves its preview now. The components fail to commute with each other:

[\hat L_x, \hat L_y] = i\hbar\,\hat L_z \quad (\text{and cyclic}) \qquad (2)

where: the right-hand side is another component of the same vector — the algebra closes on itself. The verifier spot-checks (2) on a test function, and this article uses it for nothing further: the derivation from $[\hat x_i,\hat p_j] = i\hbar\delta_{ij}$ , the $\varepsilon_{ijk}$ bookkeeping, the impossibility of sharpening two components at once, and the promotion of (2) to the definition of angular momentum are all owned by Angular Momentum Commutation Relations.

The spherical forms: angles only

Angular momentum earns its keep in spherical coordinates $x = r\sin\theta\cos\phi$ , $y = r\sin\theta\sin\phi$ , $z = r\cos\theta$ . The component along the polar axis collapses to a single derivative:

\hat L_z = -i\hbar\,\frac{\partial}{\partial\phi} \qquad (3)

where: $\phi$ is the azimuthal angle winding around $z$ . Note the family resemblance to $\hat p_x = -i\hbar\,\partial_x$ : $\hat L_z$ is the momentum conjugate to the angle — the hint, made precise below, that it does for rotations what $\hat p$ does for translations. On a winding phase $e^{im\phi}$ it reads off $m\hbar$ .

Derivation AThe azimuthal form of Lz from the chain rule

We derive $\hat L_z = -i\hbar\,\partial_\phi$ from the Cartesian definition $-i\hbar(x\,\partial_y - y\,\partial_x)$ . The strategy is one honest chain rule: express the $\phi$ -derivative of any function in Cartesian derivatives and recognize the combination that appears. The care needed is remembering that $\phi$ moves $x$ and $y$ while freezing $r$ , $\theta$ , $z$ — precisely why no radial term can survive. The computation sets the template for the family: $\hat L_x,\hat L_y$ work the same way with messier trigonometry, and $\hat L^2$ assembles from all three.

Step 1. Differentiate the coordinate map at fixed $r,\theta$ . From $x = r\sin\theta\cos\phi$ and $y = r\sin\theta\sin\phi$ :

\frac{\partial x}{\partial\phi} = -r\sin\theta\sin\phi = -y, \qquad \frac{\partial y}{\partial\phi} = r\sin\theta\cos\phi = x, \qquad \frac{\partial z}{\partial\phi} = 0

Why these three: a $\phi$ -turn carries the point along a horizontal circle — $x$ and $y$ chase each other while $z$ and the radius stand still.

Step 2. Chain rule on an arbitrary function $f(x,y,z)$ :

\frac{\partial f}{\partial\phi} = \frac{\partial x}{\partial\phi}\frac{\partial f}{\partial x} + \frac{\partial y}{\partial\phi}\frac{\partial f}{\partial y} + \frac{\partial z}{\partial\phi}\frac{\partial f}{\partial z} = \left(x\frac{\partial}{\partial y} - y\frac{\partial}{\partial x}\right) f

Step 3. Multiply by $-i\hbar$ and recognize the definition: $-i\hbar\,\partial_\phi f = -i\hbar(x\,\partial_y - y\,\partial_x)f = \hat L_z f$ for every $f$ — operator equality. The generator of the $\phi$ -circle is the angular momentum about its axis, with no radial contamination possible: the circle never changes $r$ . Verified: symbolically on four test functions, including complex combinations $(x+iy)^2 z$ — Cartesian $\hat L_z$ and $-i\hbar\,\partial_\phi$ agree identically after substitution of the spherical map (verify script, section 1).

The squared magnitude $\hat L^2 = \hat L_x^2 + \hat L_y^2 + \hat L_z^2$ assembles, after the same chain-rule work on all three components, into the pure-angle operator

\hat L^2 = -\hbar^2\left[\frac{1}{\sin\theta}\frac{\partial}{\partial\theta}\left(\sin\theta\,\frac{\partial}{\partial\theta}\right) + \frac{1}{\sin^2\theta}\frac{\partial^2}{\partial\phi^2}\right] \qquad (4)

where: the bracket is exactly the angular part of the Laplacian. Not a $\partial_r$ in sight: $\hat L^2$ measures how sharply a function wiggles across the sphere, nothing else. The verifier checks (4) against the brute-force sum $\sum_i \hat L_i^2$ on four test functions. Its eigenfunctions — the spherical harmonics, eigenvalues $\ell(\ell+1)\hbar^2$ — are constructed in Ladder Operators and the Angular Momentum Spectrum and Spherical Harmonics; we touch no eigenvalues here. How does $\hat L^2$ relate to the full kinetic energy? Through an operator identity worth deriving carefully, because it contains this article's one genuinely quantum surprise:

\hat L^2 = \hat r^2\hat p^2 - (\hat{\mathbf{r}}\cdot\hat{\mathbf{p}})^2 + i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}} \qquad (5)

where: the first two terms are the classical Lagrange identity $|\mathbf{r}\times\mathbf{p}|^2 = r^2p^2 - (\mathbf{r}\cdot\mathbf{p})^2$ , and the last term — pure $i\hbar$ , no classical ancestor — is the toll operator ordering charges for moving $\hat{\mathbf{r}}$ 's past $\hat{\mathbf{p}}$ 's. Solved for $\hat p^2$ , equation (5) splits kinetic energy into radial and centrifugal parts; that split, and the effective potential it feeds, belong to Effective Potential and the Centrifugal Barrier in the hydrogen chapter.

Go deeper · Derivation B · 1 minThe Lagrange identity with its quantum toll

Derivation BThe Lagrange identity with its quantum toll

We derive equation (5) and locate the quantum correction exactly. Classically $|\mathbf{r}\times\mathbf{p}|^2 = r^2p^2 - (\mathbf{r}\cdot\mathbf{p})^2$ — pure vector algebra. Quantum mechanically every reordering of an $\hat x$ past a same-axis $\hat p$ costs $i\hbar$ ; the strategy is to do the index algebra honestly, collect the tolls, and watch a single $+i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ survive. The watch-out: borrowing the classical identity and skipping the bookkeeping gives an operator statement the verifier shows is simply wrong.

Step 1. Write the square in index form with the epsilon contraction $\varepsilon_{ijk}\varepsilon_{ilm} = \delta_{jl}\delta_{km} - \delta_{jm}\delta_{kl}$ :

\hat L^2 = \varepsilon_{ijk}\,\hat x_j\hat p_k\,\varepsilon_{ilm}\,\hat x_l\hat p_m = \hat x_j\hat p_k\hat x_j\hat p_k - \hat x_j\hat p_k\hat x_k\hat p_j

Why the contraction is safe: the epsilons are numbers, not operators — all ordering risk lives in the $\hat x$ 's and $\hat p$ 's, whose order the expression preserves exactly.

Step 2. Normal-order the first term. Move the middle $\hat x_j$ left past $\hat p_k$ using $\hat p_k\hat x_j = \hat x_j\hat p_k - i\hbar\delta_{kj}$ :

\hat x_j\hat p_k\hat x_j\hat p_k = \hat x_j\hat x_j\,\hat p_k\hat p_k - i\hbar\,\delta_{kj}\hat x_j\hat p_k = \hat r^2\hat p^2 - i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}

Step 3. Normal-order the second term the same way — two swaps this time; the honest sum (run in the verifier's symbolic engine) collapses to:

\hat x_j\hat p_k\hat x_k\hat p_j = (\hat{\mathbf{r}}\cdot\hat{\mathbf{p}})^2 - 2i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}

Step 4. Subtract: $\hat L^2 = \hat r^2\hat p^2 - (\hat{\mathbf{r}}\cdot\hat{\mathbf{p}})^2 + i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ . The classical identity survives with exactly one quantum rider. In spherical coordinates $\hat{\mathbf{r}}\cdot\hat{\mathbf{p}} = -i\hbar\,r\,\partial_r$ : the rider is radial, belonging to the kinetic-energy split — which is why $\hat L^2$ itself still contains no $\partial_r$ . Verified: symbolically on three test functions — equation (5) holds identically with the $+i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ term and fails without it (verify script, section 5: the classical borrow is machine-refuted).

The generator of rotations

The reading that makes $\hat{\mathbf{L}}$ more than a promoted formula: Canonical Commutation Relation proved that $e^{-ia\hat p/\hbar}$ translates a wavefunction by $a$ . The identical argument with $\hat L_z$ and $\phi$ gives

e^{-i\alpha\hat L_z/\hbar}\,f(\phi) = f(\phi - \alpha) \qquad (6)

where: $\alpha$ is a rotation angle about $z$ . The exponential of angular momentum is the rotation — the quantum home of Noether's pairing of rotations with angular momentum conservation — and the doorway to symmetry: rotation-invariant Hamiltonians commute with the generator, making $\hat{\mathbf{L}}$ conserved (theory owned by Symmetries and Conservation in Quantum Dynamics and Constants of the Motion and Conservation Laws).

Go deeper · Derivation C · 1 minExponentiated Lz rotates the wavefunction

Derivation CExponentiated Lz rotates the wavefunction

We prove equation (6): the family $e^{-i\alpha\hat L_z/\hbar}$ implements rigid rotation about $z$ . The strategy is the translation argument of Canonical Commutation Relation transplanted from the line to the circle — expand the exponential, recognize the Taylor series of a shifted function, resum. It works verbatim because $\hat L_z = -i\hbar\,\partial_\phi$ has exactly momentum's form in the angle variable; the one novelty is that the circle closes on itself, a fact with consequences (single-valuedness) we flag and hand forward.

Step 1. Insert the azimuthal form into the exponential and expand the power series:

e^{-i\alpha\hat L_z/\hbar} f(\phi) = e^{-\alpha\,\partial_\phi} f(\phi) = \sum_{k=0}^{\infty}\frac{(-\alpha)^k}{k!}\,\frac{\partial^k f}{\partial\phi^k}

Why the $\hbar$ 's vanish: the generator's $-i\hbar$ cancels against the exponent's $-i/\hbar$ — rotation is geometry, and geometry is $\hbar$ -free.

Step 2. Recognize Taylor's theorem: the sum is precisely the expansion of $f(\phi - \alpha)$ about $\phi$ . Resumming,

e^{-i\alpha\hat L_z/\hbar} f(\phi) = f(\phi - \alpha)

— the pattern rotated bodily by $+\alpha$ (its features formerly at $\phi-\alpha$ now sit at $\phi$ ).

Step 3. Check on the winding phases, exact in one line: $e^{im\phi} \mapsto e^{-im\alpha}e^{im\phi}$ . A state of definite $\hat L_z$ only acquires a phase under rotation — rotationally covariant, the angular twin of a momentum eigenstate under translation. Because $\phi$ and $\phi+2\pi$ are the same point of space, the allowed $m$ are constrained — the single-valuedness argument is owned by Eigenfunctions of L_z and the Azimuthal Quantization, and we use nothing of it here. Verified: the operator series matches the shifted function to $O(\alpha^4)$ for a generic $f$ , and exactly on $e^{im\phi}$ (verify script, section 4). Finite rotations about arbitrary axes, their composition, and the group structure $SU(2)$ they generate are owned by Rotation Operators and SU(2). ***

Limiting cases

case	statement	meaning
isotropic pattern, $f(\theta)$ only	$\hat L_z f = 0$	nothing winds around $z$ — zero angular momentum about that axis
winding phase $e^{im\phi}$	$\hat L_z = m\hbar$ sharp; rotation ⇒ phase $e^{-im\alpha}$	rotationally covariant states; the angular twin of plane waves
$\hbar \to 0$ at fixed pattern	$[\hat L_x,\hat L_y] = i\hbar\hat L_z \to 0$	components compatible again — the classical arrow returns
large quantum numbers	$\hbar\sqrt{\ell(\ell+1)} \approx \hbar(\ell + \tfrac12)$ ; 1 J·s flywheel ⇒ $\ell \sim 9.5\times10^{33}$	discreteness invisible: 2.1% gap at $\ell=2$ , one part in $10^{34}$ for the flywheel
central potential $V(r)$	$[\hat H, \hat{\mathbf{L}}] = 0$	rotation symmetry ⇒ conservation — theory owned by the dynamics chapter
radial functions	$\hat L^2\,g(r) = 0$	angles-only operator: radial structure is invisible to it

Common trapYou do not need to symmetrize r×p — and "fixing" it anyway teaches the wrong lesson

Students drilled on ordering often "repair" the definition to $\tfrac12(\hat{\mathbf{r}}\times\hat{\mathbf{p}} - \hat{\mathbf{p}}\times\hat{\mathbf{r}})$ first. Harmless — but unnecessary. Each cross-product component pairs a position with a different axis's momentum ( $\hat x\hat p_y$ , never $\hat x\hat p_x$ ); the factors commute, so every ordering is the same operator (machine-checked) and hermiticity is automatic. Where ordering does bite is one floor up: squaring $\hat{\mathbf{L}}$ forces same-axis encounters, and the Lagrange identity picks up its genuine $+i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ toll. Ordering vigilance is right; aim it at the identity, not the definition.

Worked examples

Example 1 — a ring current of one particle (confidence builder). A particle on a ring is prepared in the state $\psi(\phi) = \tfrac{1}{\sqrt{2\pi}}\,\sqrt2\,\cos\phi$ (normalized). What values can a measurement of $\hat L_z$ return, with what probabilities, and what are $\langle\hat L_z\rangle$ and $\sigma_{L_z}$ ?

Step 1. Decompose into winding phases: $\cos\phi = \tfrac12(e^{i\phi} + e^{-i\phi})$ , so $\psi = \tfrac{1}{\sqrt2}\bigl(\chi_{+1} + \chi_{-1}\bigr)$ with $\chi_m = e^{im\phi}/\sqrt{2\pi}$ the normalized $\hat L_z$ eigenstates.

Step 2. Measurement menu: outcomes $+\hbar$ and $-\hbar$ , each with probability $\boxed{1/2}$ — a standing wave is an equal mix of counter-circulating currents.

Step 3. Moments: $\langle\hat L_z\rangle = \tfrac12(+\hbar) + \tfrac12(-\hbar) = \boxed{0}$ , while $\langle\hat L_z^2\rangle = \hbar^2$ , so $\sigma_{L_z} = \boxed{\hbar}$ . Zero net circulation, maximal fluctuation. (All three numbers machine-checked by exact integration.)

Example 2 — the orbital magnet, before any new physics (exam level). An electron in an atom occupies a state with definite $\hat L_z = m\hbar$ . Classically a circulating charge is a magnet with moment $\mu_z = -(e/2m_e)L_z$ ; the quantum version inherits the formula with $\hat L_z$ inserted. Find the natural unit of atomic magnetism and the energy splitting of adjacent $m$ -levels in a $B = 1.0$ T field along $z$ .

Step 1. The natural unit is the Bohr magneton: $\mu_B = \dfrac{e\hbar}{2m_e} = \boxed{9.274\times10^{-24}\ \text{J/T}} = 5.788\times10^{-5}$ eV/T.

Step 2. Interaction energy $E = -\mu_z B = +m\,\mu_B B$ (orbital case): adjacent $m$ -levels split by $\Delta E = \mu_B B = \boxed{57.9\ \mu\text{eV}}$ at 1 T. A level with $\ell = 1$ fans into three equally spaced lines spanning $2\mu_B B = 115.8\ \mu$ eV — the Zeeman fan, counted here purely from the $m$ -menu.

Step 3. As a frequency: $\nu = \mu_B B/h = \boxed{13.996\ \text{GHz}}$ per tesla — microwave territory, the working frequency of electron-orbital resonance. (Spin runs at ≈ twice this — The Time-Evolution Operator's 28 GHz/T — because spin carries $g\approx2$ ; anatomy in Magnetic Moment and the Gyromagnetic Ratio.) All numbers recomputed from CODATA constants.

The physicist: the man who quantized rotation itself

Origins · 1 minHendrik Brugt Gerhard Casimir and how this was discovered

Hendrik Brugt Gerhard Casimir

1909–2000 · Leiden, Copenhagen, then Philips

Casimir wrote the book on quantum rotation before turning twenty-two. His 1931 Leiden thesis under Paul Ehrenfest — Rotation of a rigid body in quantum mechanics — took the operators this article builds and asked what they do to a spinning object as a whole: the quantum top, the molecule tumbling in a gas. The problem forced him to isolate the one operator every component of $\hat{\mathbf{L}}$ respects, $\hat L^2$ , and to exploit its invariance systematically — the object algebraists call a Casimir operator carries his name from that thesis, generalized far beyond rotations. He learned the trade at the source, as Bohr's Copenhagen assistant and later beside Pauli; then came the second act, joining Philips Research in 1942 and eventually directing it, arguing all his life that deep theory and industry feed each other — his 1948 prediction of a vacuum force between plates, pointed at elsewhere in this track, is the emblem. For this chapter he is the right patron: the first person to run the $\hat{\mathbf{L}}$ machinery at industrial strength.

History

1913

— Bohr quantizes orbital angular momentum by decree,

L = n\hbar

, to hold the hydrogen atom together

1922

— Stern and Gerlach observe Sommerfeld's space quantization directly in a split silver beam

1925

— Born, Heisenberg and Jordan derive the angular momentum commutators inside matrix mechanics

1926

— Schrödinger's wave mechanics realizes L as the differential operators on the sphere used here

1931

— Casimir's Leiden thesis treats quantum rotation systematically; the invariant L² becomes the prototype "Casimir operator"

1992

— Allen and coworkers show laser beams carry orbital angular momentum mħ per photon, opening OAM optics

Going further

References. Griffiths, Introduction to Quantum Mechanics, 3rd ed., Ch. 4 · Sakurai & Napolitano, Modern Quantum Mechanics, 2nd ed., Ch. 3 · Shankar, Principles of Quantum Mechanics, 2nd ed., Ch. 12 · Cohen-Tannoudji, Diu & Laloë, Quantum Mechanics, Vol. 1, Ch. VI · Townsend, A Modern Approach to Quantum Mechanics, 2nd ed., Ch. 9

Provenance. Verifier should recompute: (1) $[L_x,L_y]=i\hbar L_z$ from canonical commutators (Sakurai §3.1, Griffiths §4.3); (2) the spherical-coordinate forms of $L_z$ and $L^2$ and the Laplacian split (Cohen-Tannoudji Ch. VI, Shankar §12.6); (3) integer quantization of $m$ from single-valuedness; (4) eigenvalues $\hbar^2\ell(\ell+1)$ and $\hbar m$ . Bohr magneton $\mu_B=e\hbar/2m_e\approx9.274\times10^{-24}\,\mathrm{J/T}$ (CODATA) is the only physical constant cited and is standard. No fabricated numbers. For this edition: items (1) and (2) plus every numeric claim were machine-verified (sympy/scipy, 33 checks) — $L_z$ chain rule, spherical $L^2$ vs the Cartesian sum, circle hermiticity, the rotation series, the Lagrange identity with and without its $i\hbar$ term, the ring-state numbers, and the Bohr-magneton set. Items (3) and (4) are deliberately NOT reproduced: per the chapter ownership map, single-valuedness belongs to Eigenfunctions of L_z and the Azimuthal Quantization, the eigenvalue spectrum to Ladder Operators and the Angular Momentum Spectrum, and the algebra derivation to Angular Momentum Commutation Relations.

Builds on: Canonical Commutation Relation · Momentum Operator · Vector calculus and the cross product · Spherical coordinates and the Laplacian

Leads to: Angular Momentum Commutation Relations · Ladder Operators and the Angular Momentum Spectrum · Spherical Harmonics · Eigenfunctions of L_z and the Azimuthal Quantization · Spin Angular Momentum · Rotation Operators and SU(2)

Why is there no ordering problem in r × p when every quantization lecture warns me about ordering?

The warning applies to conjugate pairs — $x$ with $p_x$ . The cross product structurally avoids them: $\hat L_z = \hat x\hat p_y - \hat y\hat p_x$ pairs each coordinate with the other axis's momentum, and $[\hat x, \hat p_y] = 0$ , so every ordering is the same operator. The toll does get paid — one floor up, inside $\hat L^2$ , where the Lagrange identity acquires its $+i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ term.

How can angular momentum contain no radial derivative at all? Surely moving the particle outward changes its angular momentum.

Distinguish the operator from the state. Rotations slide points along circles of constant $r$ , so the generator cannot contain $\partial_r$ — the chain rule confirms it. The state still knows about radius: a packet orbiting far out with winding $m$ has the same $\hat L_z$ eigenvalue $m\hbar$ as one close in, but different energy. Classically $L = mvr$ mixes radius and speed; quantum mechanically the winding number absorbs both, and the radius reappears only in the energetics — the centrifugal term of the hydrogen chapter.

Is L_z = −iħ∂φ just the momentum operator in disguise?

Formally yes, and the analogy is load-bearing: $\hat L_z$ is $-i\hbar$ times the derivative along its coordinate, exactly like $\hat p_x$ , and it generates shifts of that coordinate exactly like $\hat p_x$ generates translations. The one structural difference is global: $\phi$ lives on a circle, so its "plane waves" $e^{im\phi}$ must close on themselves after $2\pi$ . That closure discretizes the spectrum — the argument, owned by Eigenfunctions of L_z and the Azimuthal Quantization, has no analogue on the infinite line, and it is why atoms have discrete $m$ -menus while free particles have continuous momenta.

The article keeps saying the eigenvalues come later. Why can't we just solve L_z ψ = λψ right now?

You can, and the solutions are the winding phases $e^{i\lambda\phi/\hbar}$ — used freely here as test states. What is deferred is the quantization story: which $\lambda$ survive the circle's closure (single-valuedness, owned by Eigenfunctions of L_z and the Azimuthal Quantization) and how the full $\ell(\ell+1)\hbar^2,\ m\hbar$ structure follows from the algebra alone (Ladder Operators and the Angular Momentum Spectrum). The chapter separates machinery from spectrum deliberately — the algebraic route is the one that generalizes to spin, where no $\phi$ exists to differentiate.

Why is the operator identity for L² not just the classical |r×p|² = r²p² − (r·p)²?

Because operators remember their order. Expanding $\hat L^2$ forces same-axis products like $\hat x_j\hat p_j$ to march past each other, and each pass costs $i\hbar$ . The tolls do not cancel; they sum to exactly $+i\hbar\,\hat{\mathbf{r}}\cdot\hat{\mathbf{p}}$ . The verifier makes the point brutally: on explicit test functions, the classical form misses the true $\hat L^2$ by precisely that term. It is a small, clean example of a general lesson — classical identities survive quantization only up to commutator debris, and the debris is physics (here it feeds the radial-momentum bookkeeping of the central-force problem).

Where does spin fit into this construction?

Nowhere — and that is the profound part. Everything here came from $\hat{\mathbf{r}}$ and $\hat{\mathbf{p}}$ : a spatial pattern to rotate, angles to differentiate. Spin has neither. What survives the loss of the scaffolding is the algebra previewed in equation (2) — the next article's pivot is to take that algebra as the definition of angular momentum, formula optional. Spin then enters as the algebra's "extra" representations in Spin Angular Momentum, with matrices in place of derivatives.