41
Part One · Quantum mechanics

Addition of angular momenta

Two electrons, each spin up or down: four combinations. But nature does not treat them as four equal states. It sorts them into a group of three and a single one left over, and the energy difference between the three and the one decides whether iron is a magnet, whether two hydrogen atoms bond, and how an oxygen molecule sticks to haemoglobin. This chapter shows how angular momenta add, and where the three and the one come from.

On the plate: four two-spin states, regrouping into three plus one.

Two ways to label the same states

Combine angular momenta j1j_1 and j2j_2. You can label the joint states by each part’s own m1,m2m_1, m_2, the uncoupled basis, or by the total JJ and its component M=m1+m2M = m_1 + m_2, the coupled basis. The total runs in whole steps:

J=∣j1−j2∣, ∣j1−j2∣+1, …, j1+j2J = |j_1 - j_2|,\ |j_1 - j_2| + 1,\ \dots,\ j_1 + j_2

Choose j1=j_1 = 0.5 and j2=j_2 = 0.5. The plate lays out the uncoupled states as a grid; states on the same diagonal share MM. Counting the diagonals shows why the totals come out as they do, and the numbers always match: (2j1+1)(2j2+1)=∑J(2J+1)(2j_1+1)(2j_2+1) = \sum_J (2J+1). Established

Building the coupled states

The recipe needs only the lowering operator of Chapter 39. The top state is unambiguous: ∣J=j1+j2,M=J⟩=∣j1,j1⟩∣j2,j2⟩|J{=}j_1{+}j_2, M{=}J\rangle = |j_1, j_1\rangle|j_2, j_2\rangle. Apply J−=J1−+J2−J_- = J_{1-} + J_{2-} to step down the J=j1+j2J = j_1 + j_2 ladder. At each MM, the state orthogonal to it starts the next ladder down, J=j1+j2−1J = j_1 + j_2 - 1, and so on. The numbers that appear are the Clebsch–Gordan coefficients, ⟨j1m1;j2m2∣JM⟩\langle j_1 m_1; j_2 m_2|J M\rangle.

The plate shows the whole table for your j1j_1 and j2j_2: rows are coupled states, columns uncoupled ones, amber positive and red negative. Most entries are zero, because M=m1+m2M = m_1 + m_2. The table is an orthogonal matrix, a change of basis (Chapter 8) and nothing more.

Three and one

For two spin-½ particles, the recipe gives

∣1,1⟩=∣↑↑⟩,∣1,0⟩=12(∣↑↓⟩+∣↓↑⟩),∣1,−1⟩=∣↓↓⟩∣0,0⟩=12(∣↑↓⟩−∣↓↑⟩)\begin{gathered}|1,1\rangle = |{\uparrow\uparrow}\rangle,\quad |1,0\rangle = \tfrac{1}{\sqrt2}\big(|{\uparrow\downarrow}\rangle + |{\downarrow\uparrow}\rangle\big),\quad |1,-1\rangle = |{\downarrow\downarrow}\rangle \\ |0,0\rangle = \tfrac{1}{\sqrt2}\big(|{\uparrow\downarrow}\rangle - |{\downarrow\uparrow}\rangle\big)\end{gathered}

The triplet is symmetric when the particles are swapped; the singlet is antisymmetric. The singlet has total spin zero in every direction, so if you measure the two spins along the same axis, any axis, they always disagree. Measure them along axes 60° apart, and they agree with probability sin⁡2(θ/2)\sin^2(\theta/2). Neither spin has a direction of its own, yet the pair is perfectly correlated. That is entanglement, and Part Two builds Bell’s theorem on exactly this state.

Spin meets orbit

An electron moving through the nucleus’s electric field sees, in its own frame, a magnetic field, which acts on its spin. The energy is proportional to L⋅S\mathbf L\cdot\mathbf S, which does not commute with LzL_z or SzS_z but does commute with J2J^2. So the coupled basis is the right one, and the trick is

L⋅S=12(J2−L2−S2)\mathbf L\cdot\mathbf S = \tfrac12\left(J^2 - L^2 - S^2\right)

A p electron (ℓ=1\ell = 1) can have j=32j = \tfrac32 or 12\tfrac12, at slightly different energies. In sodium this splits the famous yellow line into two, at 589.0 and 589.6 nm, about 17 cm⁻¹ apart. The plate shows the doublet. Established

The Landé g-factor

Orbital and spin angular momentum make magnetic moments with different strengths: the spin counts about twice as much (gs≈2g_s \approx 2). So the total moment is not parallel to J\mathbf J. In the vector model, L\mathbf L and S\mathbf S precess around J\mathbf J, and only the moment’s projection on J\mathbf J survives on average:

gJ=1+J(J+1)+S(S+1)−L(L+1)2J(J+1)g_J = 1 + \frac{J(J+1) + S(S+1) - L(L+1)}{2J(J+1)}

Pick a level: , or . These factors, 2, 2/3 and 4/3, set how far each level shifts in a magnetic field, which the next chapter uses.

What you now have

Coupled and uncoupled bases, the allowed totals, Clebsch–Gordan coefficients from the lowering operator, the singlet and triplet, spin–orbit coupling with the J2−L2−S2J^2 - L^2 - S^2 trick, and the Landé g-factor.

· · ·

Three questions to carry with you.

Why do two spin-½ particles make J = 1 and J = 0?tap
J runs from |½ − ½| = 0 to ½ + ½ = 1, and 2 × 2 = 3 + 1 states.
What makes the singlet special?tap
It has zero total spin along every axis and is antisymmetric under exchange; the two spins always disagree along a common axis.
Why is the coupled basis right for spin–orbit coupling?tap
L·S commutes with J², L² and S², not with L_z or S_z separately.

And one to work out. What is gJg_J for a ²P₃/₂ level (L=1,S=12,J=32L = 1, S = \tfrac12, J = \tfrac32)?

g_J =

Sources: J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, §3.8. A. R. Edmonds, Angular Momentum in Quantum Mechanics. NIST Atomic Spectra Database (Na D lines 588.995 and 589.592 nm). Checked by code (SymPy): triplet and singlet coefficients ±1/√2; ⟨1 0; ½ ½|³⁄₂ ½⟩ = √(2/3); ⟨1 0; 1 0|1 0⟩ = 0; g_J = 2, 2/3, 4/3; Na D splitting 17.19 cm⁻¹.

Next · Chapter 42
Magnetic fields and the Zeeman effect

An electron feels a magnetic field it never touches.

Turn the page