Central Potentials & Orbital Angular Momentum

4 hours ~6 min read

Central Potentials & Orbital Angular Momentum

One dimension was training. Real atoms are three-dimensional, and the force an electron feels from a nucleus depends only on distance — a central potential. That symmetry splits the Schrödinger equation into a radial problem and a universal angular one, solved by the same ladder trick that cracked the harmonic oscillator — now climbing angular momentum instead of energy. A cat circling a warm laptop conserves L\mathbf L; so does an electron circling a proton — but only the electron's comes in steps of \hbar.

Learning Objectives

After this lesson you will be able to:

  1. Separate variables in the 3D Schrödinger equation for a central potential V(r)V(r), identifying the separation constant (+1)\ell(\ell+1).
  2. Derive the commutator algebra [L^x,L^y]=iL^z[\hat L_x,\hat L_y] = i\hbar\hat L_z from the canonical commutators and recognize its classical shadow, the Poisson bracket {Lx,Ly}=Lz\{L_x,L_y\} = L_z.
  3. Construct the ladder operators L^±\hat L_\pm and derive the eigenvalues of L^2\hat L^2 and L^z\hat L_z purely algebraically.
  4. Explain why orbital angular momentum requires integer \ell, while the algebra alone also permits half-integer values.
  5. Use the spherical harmonics YmY_\ell^m — eigenvalue equations, orthonormality, the 2+12\ell+1 degeneracy — and critique the vector model via Lx2+Ly2>0\langle L_x^2 + L_y^2\rangle > 0.

Intuition

Classically a central force exerts no torque, so L=r×p\mathbf L = \mathbf r\times\mathbf p is conserved and orbits flatten into planes (P.1.3 Hamiltonian Mechanics). Quantum mechanics inherits the structure: rotational symmetry peels the angular dependence off the wavefunction once and for all — the angular factors are the same for hydrogen, a 3D oscillator, any V(r)V(r). But there is a twist: classically Lx,Ly,LzL_x, L_y, L_z are three numbers and you may know them all; quantum mechanically they are noncommuting operators — you may know the length L|\mathbf L| and one component, and the other two are then irreducibly indeterminate. The entire spectrum follows from the commutators alone, by the ladder logic of P.5.3 The Harmonic Oscillator.


Theory

Separation of variables in a central potential

The 3D time-independent Schrödinger equation is 22m2ψ+Vψ=Eψ-\frac{\hbar^2}{2m}\nabla^2\psi + V\psi = E\psi. For V=V(r)V = V(r) use spherical coordinates, with the Laplacian (a standard chain-rule exercise, quoted):

2=1r2r ⁣(r2r)+1r2sinθθ ⁣(sinθθ)+1r2sin2θ2φ2. \nabla^2 = \frac{1}{r^2}\frac{\partial}{\partial r}\!\left(r^2\frac{\partial}{\partial r}\right) + \frac{1}{r^2\sin\theta}\frac{\partial}{\partial\theta}\!\left(\sin\theta\,\frac{\partial}{\partial\theta}\right) + \frac{1}{r^2\sin^2\theta}\frac{\partial^2}{\partial\varphi^2}.

Try ψ(r,θ,φ)=R(r)Y(θ,φ)\psi(r,\theta,\varphi) = R(r)\,Y(\theta,\varphi); substitute and multiply by 2mr221RY-\frac{2mr^2}{\hbar^2}\frac{1}{RY}:

{1Rddr ⁣(r2dRdr)2mr22[V(r)E]}depends only on r+1Y{1sinθθ ⁣(sinθYθ)+1sin2θ2Yφ2}depends only on θ,φ=0. \underbrace{\left\{\frac{1}{R}\frac{d}{dr}\!\left(r^2\frac{dR}{dr}\right) - \frac{2mr^2}{\hbar^2}\bigl[V(r)-E\bigr]\right\}}_{\text{depends only on } r} + \underbrace{\frac{1}{Y}\left\{\frac{1}{\sin\theta}\frac{\partial}{\partial\theta}\!\left(\sin\theta\,\frac{\partial Y}{\partial\theta}\right) + \frac{1}{\sin^2\theta}\frac{\partial^2 Y}{\partial\varphi^2}\right\}}_{\text{depends only on } \theta,\varphi} = 0 .

A function of rr plus a function of (θ,φ)(\theta,\varphi) vanishes identically only if each is constant; write the constant — with foresight — as (+1)\ell(\ell+1). The radial piece keeps V(r)V(r) and is deferred to P.6.2; the angular piece,

1sinθθ ⁣(sinθYθ)+1sin2θ2Yφ2=(+1)Y, \frac{1}{\sin\theta}\frac{\partial}{\partial\theta}\!\left(\sin\theta\,\frac{\partial Y}{\partial\theta}\right) + \frac{1}{\sin^2\theta}\frac{\partial^2 Y}{\partial\varphi^2} = -\ell(\ell+1)\,Y ,

is universal — and, as we now show, it is the eigenvalue problem of angular momentum.

The angular momentum operators and their algebra

Quantizing L=r×p\mathbf L = \mathbf r\times\mathbf p gives L^=r^×p^\hat{\mathbf L} = \hat{\mathbf r}\times\hat{\mathbf p}, with components L^x=y^p^zz^p^y\hat L_x = \hat y\hat p_z - \hat z\hat p_y, L^y=z^p^xx^p^z\hat L_y = \hat z\hat p_x - \hat x\hat p_z, L^z=x^p^yy^p^x\hat L_z = \hat x\hat p_y - \hat y\hat p_x (no ordering ambiguity: each product pairs a coordinate with a different momentum component). Using [x^i,p^j]=iδij[\hat x_i,\hat p_j] = i\hbar\delta_{ij}, expand [L^x,L^y][\hat L_x,\hat L_y] into four brackets; [y^p^z,x^p^z][\hat y\hat p_z,\hat x\hat p_z] and [z^p^y,z^p^x][\hat z\hat p_y,\hat z\hat p_x] contain only mutually commuting factors and vanish, leaving

[L^x,L^y]=[y^p^z,z^p^x]+[z^p^y,x^p^z]=y^[p^z,z^]p^x+x^[z^,p^z]p^y=i(x^p^yy^p^x)=iL^z, [\hat L_x,\hat L_y] = [\hat y\hat p_z,\hat z\hat p_x] + [\hat z\hat p_y,\hat x\hat p_z] = \hat y\,[\hat p_z,\hat z]\,\hat p_x + \hat x\,[\hat z,\hat p_z]\,\hat p_y = i\hbar\,(\hat x\hat p_y - \hat y\hat p_x) = i\hbar\,\hat L_z ,

and cyclically [L^y,L^z]=iL^x[\hat L_y,\hat L_z] = i\hbar\hat L_x, [L^z,L^x]=iL^y[\hat L_z,\hat L_x] = i\hbar\hat L_y. This mirrors the classical Poisson bracket {Lx,Ly}=Lz\{L_x,L_y\} = L_z computed in P.1.3 Hamiltonian Mechanics via Dirac's rule {,}[,]/i\{\cdot,\cdot\}\to[\cdot,\cdot]/i\hbar — the algebra of rotations was there all along.

No two components commute, so no common eigenbasis exists: LxL_x and LzL_z cannot both be sharp (the uncertainty relation behind this is developed in 1.3.2 Expectation & Uncertainty). But L^2=L^x2+L^y2+L^z2\hat L^2 = \hat L_x^2 + \hat L_y^2 + \hat L_z^2 commutes with every component: using [A^2,B^]=A^[A^,B^]+[A^,B^]A^[\hat A^2,\hat B] = \hat A[\hat A,\hat B] + [\hat A,\hat B]\hat A,

[L^2,L^z]=i(L^xL^y+L^yL^x)+i(L^yL^x+L^xL^y)+0=0, [\hat L^2,\hat L_z] = -i\hbar(\hat L_x\hat L_y + \hat L_y\hat L_x) + i\hbar(\hat L_y\hat L_x + \hat L_x\hat L_y) + 0 = 0 ,

so L^2\hat L^2 and L^z\hat L_z do share simultaneous eigenfunctions — the maximal sharp angular data. In spherical coordinates (a rotation about zz changes only φ\varphi, so φ=xyyx\partial_\varphi = x\partial_y - y\partial_x):

L^z=iφ,L^2=2[1sinθθ ⁣(sinθθ)+1sin2θ2φ2], \hat L_z = -i\hbar\frac{\partial}{\partial\varphi}, \qquad \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\varphi^2}\right],

so the angular equation above is L^2Y=2(+1)Y\hat L^2 Y = \hbar^2\ell(\ell+1)\,Y.

Ladder operators: the spectrum from pure algebra

In perfect analogy with the a^±\hat a_\pm of P.5.3, define L^±L^x±iL^y\hat L_\pm \equiv \hat L_x \pm i\hat L_y. Then [L^2,L^±]=0[\hat L^2,\hat L_\pm] = 0, and

[L^z,L^±]=[L^z,L^x]±i[L^z,L^y]=iL^y±i(iL^x)=±L^±. [\hat L_z,\hat L_\pm] = [\hat L_z,\hat L_x] \pm i[\hat L_z,\hat L_y] = i\hbar\hat L_y \pm i(-i\hbar\hat L_x) = \pm\hbar\,\hat L_\pm .

Let λ,μ\lvert\lambda,\mu\rangle satisfy L^2λ,μ=λλ,μ\hat L^2\lvert\lambda,\mu\rangle = \lambda\lvert\lambda,\mu\rangle and L^zλ,μ=μλ,μ\hat L_z\lvert\lambda,\mu\rangle = \mu\lvert\lambda,\mu\rangle. Then L^z(L^±λ,μ)=(L^±L^z±L^±)λ,μ=(μ±)L^±λ,μ\hat L_z(\hat L_\pm\lvert\lambda,\mu\rangle) = (\hat L_\pm\hat L_z \pm \hbar\hat L_\pm)\lvert\lambda,\mu\rangle = (\mu\pm\hbar)\,\hat L_\pm\lvert\lambda,\mu\rangle while L^2(L^±λ,μ)=λL^±λ,μ\hat L^2(\hat L_\pm\lvert\lambda,\mu\rangle) = \lambda\,\hat L_\pm\lvert\lambda,\mu\rangle: a ladder in LzL_z, steps of \hbar, fixed total length. It is bounded, since λμ2=L^x2+L^y2=L^xλ,μ2+L^yλ,μ20\lambda - \mu^2 = \langle\hat L_x^2 + \hat L_y^2\rangle = \lVert\hat L_x\lvert\lambda,\mu\rangle\rVert^2 + \lVert\hat L_y\lvert\lambda,\mu\rangle\rVert^2 \ge 0. So there is a top rung, L^+λ,μmax=0\hat L_+\lvert\lambda,\mu_{\max}\rangle = 0 with μmax\mu_{\max} \equiv \hbar\ell. Expanding L^L^±=L^x2+L^y2±i[L^x,L^y]=L^2L^z2L^z\hat L_\mp\hat L_\pm = \hat L_x^2 + \hat L_y^2 \pm i[\hat L_x,\hat L_y] = \hat L^2 - \hat L_z^2 \mp \hbar\hat L_z and applying L^L^+\hat L_-\hat L_+ to the top rung: 0=λ2220 = \lambda - \hbar^2\ell^2 - \hbar^2\ell, i.e. λ=2(+1)\lambda = \hbar^2\ell(\ell+1). Likewise a bottom rung ˉ\hbar\bar\ell gives λ=2ˉ(ˉ1)\lambda = \hbar^2\bar\ell(\bar\ell - 1), so ˉ=\bar\ell = -\ell (the root +1\ell+1 would put the bottom above the top). Descending from \hbar\ell to -\hbar\ell in steps of \hbar requires 22\ell to be a non-negative integer:

L^22(+1),L^zm,m=,,  (integer steps),=0,12,1,32, \hat L^2 \to \hbar^2\ell(\ell+1), \qquad \hat L_z \to \hbar m, \qquad m = -\ell, \dots, \ell \ \ (\text{integer steps}), \qquad \ell = 0, \tfrac12, 1, \tfrac32, \dots

The commutators alone — no wavefunctions, no boundary conditions — fixed the entire spectrum, and they permit both integer and half-integer \ell.

The differential route: why orbital \ell is an integer

For angular momentum realized as ir×-i\hbar\,\mathbf r\times\nabla on wavefunctions, more is true. Solving L^zΦ=mΦ\hat L_z\Phi = \hbar m\Phi gives Φ=eimφ\Phi = e^{im\varphi}; but φ\varphi and φ+2π\varphi + 2\pi label the same point of space, and single-valuedness eim(φ+2π)=eimφe^{im(\varphi + 2\pi)} = e^{im\varphi} forces mZm \in \mathbb Z — hence integer \ell for orbital angular momentum. The half-integer solutions are not garbage: they describe angular momentum that is not motion through space — no φ\varphi, no constraint. That loophole stays open until the Stern–Gerlach experiment forces us through it in P.6.3 Magnetic Moments, Stern–Gerlach & Spin.

Spherical harmonics

The normalized simultaneous eigenfunctions of L^2\hat L^2 and L^z\hat L_z are the spherical harmonics, built from the associated Legendre functions PmP_\ell^m:

Ym(θ,φ)=(1)m2+14π(m)!(+m)!Pm(cosθ)eimφ(m0;  Ym=(1)mYm). Y_\ell^m(\theta,\varphi) = (-1)^m\sqrt{\frac{2\ell+1}{4\pi}\frac{(\ell-m)!}{(\ell+m)!}}\, P_\ell^m(\cos\theta)\,e^{im\varphi} \quad (m\ge0;\; Y_\ell^{-m} = (-1)^mY_\ell^{m*}).
\ell mm Ym(θ,φ)Y_\ell^m(\theta,\varphi)
0 0 1/4π\sqrt{1/4\pi}
1 0 3/4πcosθ\sqrt{3/4\pi}\,\cos\theta
1 ±1\pm1 3/8πsinθe±iφ\mp\sqrt{3/8\pi}\,\sin\theta\,e^{\pm i\varphi}
2 0 5/16π(3cos2θ1)\sqrt{5/16\pi}\,(3\cos^2\theta - 1)
2 ±1\pm1 15/8πsinθcosθe±iφ\mp\sqrt{15/8\pi}\,\sin\theta\cos\theta\,e^{\pm i\varphi}
2 ±2\pm2 15/32πsin2θe±2iφ\sqrt{15/32\pi}\,\sin^2\theta\,e^{\pm2i\varphi}

They are orthonormal on the sphere02π ⁣0πYmYmsinθdθdφ=δδmm\int_0^{2\pi}\!\int_0^{\pi} Y_{\ell'}^{m'*}\,Y_\ell^m\,\sin\theta\,d\theta\,d\varphi = \delta_{\ell\ell'}\delta_{mm'} — and complete. For each \ell there are 2+12\ell+1 values of mm with the same L^2\hat L^2 eigenvalue: a degeneracy every central potential inherits, since energy cannot depend on the orientation of L\mathbf L when the potential has no preferred direction.

The vector model and its limits

A common mnemonic draws L\mathbf L as an arrow of length (+1)\hbar\sqrt{\ell(\ell+1)} on a cone about zz, with fixed projection m\hbar m and smeared azimuth. The cone's opening is real physics:

Lx2+Ly2=L^2L^z2=2[(+1)m2]>0for >0, \langle L_x^2 + L_y^2\rangle = \langle\hat L^2 - \hat L_z^2\rangle = \hbar^2\bigl[\ell(\ell+1) - m^2\bigr] > 0 \quad\text{for } \ell > 0 ,

even at m=m = \ell (where it equals 2\hbar^2\ell): L\mathbf L never points exactly along zz, tilting at best to cosθmin=/(+1)<1\cos\theta_{\min} = \ell/\sqrt{\ell(\ell+1)} < 1. But don't take the cone literally — nothing secretly precesses at a hidden azimuth.

Caution. L=(+1)>m|\mathbf L| = \hbar\sqrt{\ell(\ell+1)} > \hbar\ell \ge \hbar m. The Bohr-era phrase "the angular momentum is \ell\hbar" is shorthand and strictly false — the length exceeds the largest projection, which is exactly why L\mathbf L can never align with an axis. And a state with definite LzL_z has indeterminate LxL_x and LyL_y: not unknown values — no values.


Worked Examples

Example 1 — Sizing up =2\ell = 2

L=23=62.449|\mathbf L| = \hbar\sqrt{2\cdot3} = \sqrt6\,\hbar \approx 2.449\,\hbar, with five projections Lz=mL_z = m\hbar, m{2,,2}m\in\{-2,\dots,2\}. Even at m=2m = 2 the vector tilts off-axis by cosθ=2/6=0.8165\cos\theta = 2/\sqrt6 = 0.8165, i.e. θ35.3\theta \approx 35.3^\circ, with transverse spread Lx2+Ly2=2(64)=22\langle L_x^2 + L_y^2\rangle = \hbar^2(6-4) = 2\hbar^2. The largest projection 22\hbar never reaches the length 2.4492.449\,\hbar.

Example 2 — Y10Y_1^0 through the machinery

Y10=3/4πcosθY_1^0 = \sqrt{3/4\pi}\cos\theta has no φ\varphi-dependence, so L^zY10=0\hat L_zY_1^0 = 0: m=0m = 0. ✓ For L^2\hat L^2 (the φ\varphi term drops):

1sinθddθ ⁣(sinθdcosθdθ)=1sinθddθ(sin2θ)=2cosθ, \frac{1}{\sin\theta}\frac{d}{d\theta}\!\left(\sin\theta\,\frac{d\cos\theta}{d\theta}\right) = \frac{1}{\sin\theta}\frac{d}{d\theta}\bigl(-\sin^2\theta\bigr) = -2\cos\theta ,

so L^2Y10=22Y10=2(+1)Y10\hat L^2Y_1^0 = 2\hbar^2\,Y_1^0 = \hbar^2\ell(\ell+1)Y_1^0 with =1\ell = 1. ✓ Lowering with L^=eiφ(θicotθφ)\hat L_- = -\hbar e^{-i\varphi}(\partial_\theta - i\cot\theta\,\partial_\varphi): L^Y10=3/4πsinθeiφ=2Y11\hat L_-Y_1^0 = \hbar\sqrt{3/4\pi}\,\sin\theta\,e^{-i\varphi} = \hbar\sqrt2\,Y_1^{-1}, with coefficient exactly (+1)m(m1)=2\hbar\sqrt{\ell(\ell+1) - m(m-1)} = \hbar\sqrt2 — the ladder in action.


Hands-on (Python)

Visualize Ym2|Y_\ell^m|^2 for =0,1,2\ell = 0,1,2 as polar surface plots (radius = magnitude):

import numpy as np
import matplotlib.pyplot as plt
from scipy.special import sph_harm

# NOTE scipy's argument order/naming: sph_harm(m, l, azimuthal, polar).
polar = np.linspace(0, np.pi, 101)          # theta in physics texts
azim  = np.linspace(0, 2*np.pi, 201)        # varphi in physics texts
TH, PH = np.meshgrid(polar, azim)

fig = plt.figure(figsize=(13, 8))
for l in range(3):
    for m in range(-l, l + 1):
        rad = np.abs(sph_harm(m, l, PH, TH))**2          # radius = |Y_l^m|^2
        X, Y, Z = (rad*np.sin(TH)*np.cos(PH), rad*np.sin(TH)*np.sin(PH), rad*np.cos(TH))
        ax = fig.add_subplot(3, 5, 5*l + m + 3, projection='3d')  # center each row
        ax.plot_surface(X, Y, Z, cmap='viridis', rstride=2, cstride=2)
        ax.set_title(f"$\\ell={l},\\ m={m}$"); ax.set_axis_off()
plt.tight_layout(); plt.show()
# Expect: l=0 a sphere; l=1 dumbbells (m=0 along z, |m|=1 donuts); l=2 the
# donut/dumbbell family. |Y|^2 is always symmetric about the z-axis.

Verify orthonormality YmYmdΩ=δδmm\int Y_{\ell'}^{m'*}Y_\ell^m\,d\Omega = \delta_{\ell\ell'}\delta_{mm'} by 2D quadrature:

def overlap(l1, m1, l2, m2, n_th=400, n_ph=800):
    th, ph = np.linspace(0, np.pi, n_th), np.linspace(0, 2*np.pi, n_ph)
    TH, PH = np.meshgrid(th, ph, indexing='ij')
    f = np.conj(sph_harm(m1, l1, PH, TH))*sph_harm(m2, l2, PH, TH)*np.sin(TH)
    return np.trapezoid(np.trapezoid(f, ph, axis=1), th)   # np.trapz on older NumPy

labels = [(l, m) for l in range(3) for m in range(-l, l + 1)]   # 9 states
G = np.array([[overlap(*a, *b) for b in labels] for a in labels])
print(np.max(np.abs(G - np.eye(9))))   # ~1e-6: Gram matrix = identity

Exercises

E1 (easy). For =3\ell = 3: how many mm values? Compute L|\mathbf L| and the smallest possible angle between L\mathbf L and the zz-axis.

Solution

2+1=72\ell+1 = 7 values, m=3,,3m = -3,\dots,3; L=12=233.464|\mathbf L| = \hbar\sqrt{12} = 2\sqrt3\,\hbar \approx 3.464\,\hbar. At m=3m = 3: cosθ=3/12=3/2\cos\theta = 3/\sqrt{12} = \sqrt3/2, so θ=30\theta = 30^\circ exactly.

E2 (easy). Derive [L^z,L^x]=iL^y[\hat L_z,\hat L_x] = i\hbar\hat L_y (a) directly from the canonical commutators, (b) by cyclic relabeling.

Solution

(a) [x^p^yy^p^x, y^p^zz^p^y][\hat x\hat p_y - \hat y\hat p_x,\ \hat y\hat p_z - \hat z\hat p_y] has nonvanishing pieces [x^p^y,y^p^z]=x^[p^y,y^]p^z=ix^p^z[\hat x\hat p_y,\hat y\hat p_z] = \hat x[\hat p_y,\hat y]\hat p_z = -i\hbar\hat x\hat p_z and [y^p^x,z^p^y]=z^p^x[y^,p^y]=iz^p^x[\hat y\hat p_x,\hat z\hat p_y] = \hat z\hat p_x[\hat y,\hat p_y] = i\hbar\hat z\hat p_x; sum =i(z^p^xx^p^z)=iL^y= i\hbar(\hat z\hat p_x - \hat x\hat p_z) = i\hbar\hat L_y. (b) xyzxx\to y\to z\to x preserves [x^i,p^j]=iδij[\hat x_i,\hat p_j] = i\hbar\delta_{ij} and maps L^xL^yL^zL^x\hat L_x\to\hat L_y\to\hat L_z\to\hat L_x, carrying [L^x,L^y]=iL^z[\hat L_x,\hat L_y] = i\hbar\hat L_z into the other two identities.

E3 (medium). Prove L^L^±=L^2L^z2L^z\hat L_\mp\hat L_\pm = \hat L^2 - \hat L_z^2 \mp \hbar\hat L_z and deduce L^±,m=(+1)m(m±1),m±1\hat L_\pm\lvert \ell,m\rangle = \hbar\sqrt{\ell(\ell+1) - m(m\pm1)}\,\lvert \ell,m\pm1\rangle.

Solution

$(\hat L_x \mp i\hat L_y)(\hat L_x \pm i\hat L_y) = \hat L_x^2 + \hat L_y^2 \pm i[\hat L_x,\hat L_y] = \hat L^2 - \hat L_z^2 \mp \hbar\hat L_z$. Then $\lVert\hat L_\pm\lvert \ell,m\rangle\rVert^2 = \langle \ell,m\rvert\hat L_\mp\hat L_\pm\lvert \ell,m\rangle = \hbar^2[\ell(\ell+1) - m(m\pm1)]$; the standard positive phase gives the stated coefficient, which vanishes at m=±m = \pm\ell: the ladder terminates itself.

E4 (medium). By explicit integration, verify that Y11Y_1^1 is normalized and orthogonal to Y10Y_1^0.

Solution

Y112dΩ=38π2π0πsin3θdθ=38π2π43=1\int|Y_1^1|^2d\Omega = \frac{3}{8\pi}\cdot2\pi\int_0^\pi\sin^3\theta\,d\theta = \frac{3}{8\pi}\cdot2\pi\cdot\frac43 = 1 ✓. Orthogonality: Y10Y11dΩ02πeiφdφ=0\int Y_1^{0*}Y_1^1\,d\Omega \propto \int_0^{2\pi}e^{i\varphi}d\varphi = 0 — the azimuthal integral kills any pair with mmm\ne m'. ✓

E5 (hard). In ,m\lvert \ell,m\rangle, show Lx=Ly=0\langle L_x\rangle = \langle L_y\rangle = 0 and Lx2=Ly2=22[(+1)m2]\langle L_x^2\rangle = \langle L_y^2\rangle = \tfrac{\hbar^2}{2}[\ell(\ell+1) - m^2], then verify ΔLxΔLy2Lz\Delta L_x\,\Delta L_y \ge \tfrac\hbar2|\langle L_z\rangle|.

Solution

L^x=12(L^++L^)\hat L_x = \tfrac12(\hat L_+ + \hat L_-) maps ,m\lvert \ell,m\rangle to orthogonal states, so Lx=0\langle L_x\rangle = 0 (same for LyL_y). In L^x2=14(L^+2+L^2+L^+L^+L^L^+)\hat L_x^2 = \tfrac14(\hat L_+^2 + \hat L_-^2 + \hat L_+\hat L_- + \hat L_-\hat L_+) the L^±2\hat L_\pm^2 terms shift mm by ±2\pm2 and drop; E3 gives L^+L^+L^L^+=22[(+1)m2]\langle\hat L_+\hat L_- + \hat L_-\hat L_+\rangle = 2\hbar^2[\ell(\ell+1) - m^2], so Lx2=Ly2=22[(+1)m2]\langle L_x^2\rangle = \langle L_y^2\rangle = \tfrac{\hbar^2}{2}[\ell(\ell+1) - m^2]. Then ΔLxΔLy22m    (+1)m(m+1)\Delta L_x\Delta L_y \ge \tfrac{\hbar^2}{2}|m| \iff \ell(\ell+1) \ge |m|(|m|+1), true since m\ell \ge |m| — with equality at m=±m = \pm\ell: the end rungs are minimum-uncertainty states.


Checkpoint

  1. In ψ=RY\psi = R\,Y, what is the separation constant, and what operator equation does YY satisfy?
  2. State [L^x,L^y][\hat L_x,\hat L_y] and its classical counterpart. What rule connects them?
  3. Why do L^2\hat L^2 and L^z\hat L_z share eigenfunctions while L^x\hat L_x and L^z\hat L_z do not?
  4. Which step of which argument restricts orbital \ell to integers — and why doesn't it bind spin?
  5. Why can L\mathbf L never point exactly along the zz-axis for >0\ell > 0?
Answers
  1. (+1)\ell(\ell+1); the angular equation is L^2Y=2(+1)Y\hat L^2Y = \hbar^2\ell(\ell+1)Y.
  2. [L^x,L^y]=iL^z[\hat L_x,\hat L_y] = i\hbar\hat L_z, mirroring {Lx,Ly}=Lz\{L_x,L_y\} = L_z via Dirac's rule {,}[,]/i\{\cdot,\cdot\}\to[\cdot,\cdot]/i\hbar.
  3. [L^2,L^z]=0[\hat L^2,\hat L_z] = 0 guarantees a common eigenbasis; [L^x,L^z]=iL^y0[\hat L_x,\hat L_z] = -i\hbar\hat L_y \ne 0 forbids one.
  4. Single-valuedness of eimφe^{im\varphi} under φφ+2π\varphi\to\varphi+2\pi in the differential realization forces mZm\in\mathbb Z. Spin is not a function of spatial angles, so the constraint never applies — the algebra's half-integers survive.
  5. L=(+1)|\mathbf L| = \hbar\sqrt{\ell(\ell+1)} strictly exceeds the largest projection \hbar\ell; equivalently Lx2+Ly2=2[(+1)m2]>0\langle L_x^2 + L_y^2\rangle = \hbar^2[\ell(\ell+1) - m^2] > 0.

Further Reading

  • [Gri] Griffiths & Schroeter, §4.1 — separation of variables and the angular equation; this lesson's conventions.
  • [Gri] Griffiths & Schroeter, §4.3 — the algebraic theory of angular momentum; our ladder derivation follows it closely.
  • [Sha] Shankar, Ch. 12 — angular momentum from rotational symmetry; the algebra as the group theory of rotations.
  • [ER] Eisberg & Resnick, Ch. 7 — quantum numbers with the experimental context of one-electron atoms.

← Prev: The Harmonic Oscillator · Up: Pre-Term · Next: The Radial Equation & the Hydrogen Atom

Ready to measure your state?

5 exercises · 10 checkpoint questions

Start the quiz