40
Part One · Quantum mechanics

Spin-1/2 and Stern–Gerlach

In Frankfurt in February 1922, Otto Stern and Walther Gerlach sent a beam of silver atoms between the poles of a specially shaped magnet. Each atom is a tiny magnet, and classically its orientation should be random, so the beam should have smeared into a continuous band. On the glass plate where the atoms landed they found two separate marks. Not a smear. Two.

On the plate: the experiment. Silver atoms enter from the left; the magnet sorts each one up or down.

Two spots

The deposit was so thin that it was invisible until Stern breathed on it. He smoked cheap cigars, and the sulphur in his breath turned the silver into black silver sulphide, making the two marks appear. That, at least, is how Stern told it, and Bretislav Friedrich and Dudley Herschbach later reproduced the effect.

Toggle the expectation: or . At the time, the split was read as Bohr’s orbits lining up. The right explanation came in 1925: the electron itself carries an intrinsic angular momentum, spin, with s=12s = \tfrac12, and the outer electron of silver has no orbital angular momentum to confuse things. Spin-½ has two states, so there are two spots. Established

Spinors and the Pauli matrices

A spin-½ state is a two-component vector, a spinor: ∣ψ⟩=α∣↑⟩+β∣↓⟩|\psi\rangle = \alpha|{\uparrow}\rangle + \beta|{\downarrow}\rangle. Its spin operators are S=ℏ2σ\mathbf S = \tfrac\hbar2\boldsymbol\sigma, with the Pauli matrices of Chapter 8:

σx=(0110), σy=(0−ii0), σz=(100−1)\sigma_x = \begin{pmatrix}0&1\\1&0\end{pmatrix},\ \sigma_y = \begin{pmatrix}0&-i\\i&0\end{pmatrix},\ \sigma_z = \begin{pmatrix}1&0\\0&-1\end{pmatrix}

A magnet tilted at angle θ\theta from zz measures n^⋅σ\hat n\cdot\boldsymbol\sigma, whose “up” state is cos⁡θ2∣↑⟩+sin⁡θ2∣↓⟩\cos\tfrac\theta2|{\uparrow}\rangle + \sin\tfrac\theta2|{\downarrow}\rangle. Prepare atoms spin-up along zz, then send them through a second magnet tilted by 60°. The fraction going up is cos⁡2(θ/2)\cos^2(\theta/2): all of them at 0°, half at 90°, none at 180°.

A chain of magnets

Build a sequence: , , or . After each magnet the lower beam is blocked.

Two z magnets in a row: everything that passed the first passes the second. Insert an x magnet between them and half of what comes out of the last one is spin-down along z again, although every one of those atoms was spin-up along z before. Measuring SxS_x erased the SzS_z information, because the two do not commute. This three-magnet experiment is the shortest demonstration there is that quantum measurement changes the state.

The Bloch sphere

Any spin-½ state, up to an overall phase, can be written cos⁡θ2∣↑⟩+eiϕsin⁡θ2∣↓⟩\cos\tfrac\theta2|{\uparrow}\rangle + e^{i\phi}\sin\tfrac\theta2|{\downarrow}\rangle, so it is a point on a sphere with polar angle θ\theta and azimuth ϕ\phi. Set θ=\theta = 60° and ϕ=\phi = 45°.

The arrow points along the direction in which a magnet would find spin-up with certainty, and ⟨σ⟩\langle\boldsymbol\sigma\rangle is exactly that arrow. Opposite points on the sphere are orthogonal states: up and down are 180° apart here, though they are perpendicular in Hilbert space. That halving of angles is the same one you met in Chapter 39. Part Three uses this sphere for every qubit.

Precession, and how MRI works

In a magnetic field B\mathbf B along zz, a magnetic moment μ=γS\boldsymbol\mu = \gamma\mathbf S has energy H=−γBSzH = -\gamma B S_z. The up and down parts pick up opposite phases, so the Bloch arrow turns around zz at the Larmor frequency ω=γB\omega = \gamma B, without changing its tilt. Ehrenfest’s theorem (Chapter 31) says the average spin obeys the classical gyroscope equation.

For protons, γ/2π=42.577\gamma/2\pi = 42.577 MHz per tesla. Set the scanner field to 3.0 T. An MRI scanner listens for exactly this radio-frequency signal from the hydrogen in water and fat, and uses field gradients to tell which part of the body it came from. Established

Rabi oscillations

To flip a spin, add a small field rotating at frequency ω\omega near the Larmor frequency ω0\omega_0. Seen from a frame rotating with it, the spin precesses about a tilted axis, and the probability of finding it flipped oscillates:

P↓(t)=Ω2Ω2+Δ2 sin⁡2 ⁣(Ω2+Δ22 t)P_\downarrow(t) = \frac{\Omega^2}{\Omega^2 + \Delta^2}\,\sin^2\!\left(\frac{\sqrt{\Omega^2+\Delta^2}}{2}\,t\right)

Here Ω\Omega is the strength of the drive and Δ=ω−ω0\Delta = \omega - \omega_0 the detuning. Set Δ/Ω=\Delta/\Omega = 0.00. On resonance the spin flips completely; off resonance it flips partly and faster. Isidor Rabi used this in 1938 to measure nuclear magnetic moments, and won the 1944 Nobel Prize. MRI pulses, atomic clocks and every gate on a qubit are Rabi rotations.

What you now have

The Stern–Gerlach experiment, spinors and Pauli matrices, the cos⁡2(θ/2)\cos^2(\theta/2) rule, sequential measurements, the Bloch sphere, Larmor precession and MRI, and Rabi oscillations.

· · ·

Three questions to carry with you.

Why did z, x, z magnets give spin-down atoms again?tap
Measuring S_x leaves the atom in an S_x eigenstate, which is an equal superposition of z-up and z-down.
How far apart are orthogonal states on the Bloch sphere?tap
180°: diametrically opposite.
What sets the MRI signal frequency?tap
The Larmor frequency γB: 42.577 MHz per tesla for protons.

And one to work out. What is the proton Larmor frequency in a 1.5 T scanner, in MHz?

f =MHz

Sources: W. Gerlach and O. Stern, Z. Phys. 9, 349 (1922). B. Friedrich and D. Herschbach, Physics Today 56(12), 53 (2003). G. E. Uhlenbeck and S. Goudsmit, Naturwissenschaften 13, 953 (1925). I. I. Rabi et al., Phys. Rev. 53, 318 (1938). CODATA 2022: γ_p/2π = 42.577 MHz T⁻¹. Checked by code: P(up) = cos²(θ/2); z–x–z output 12.5% each way of the input; Larmor 127.7 MHz at 3 T, 63.87 MHz at 1.5 T; Rabi formula against direct integration.

Next · Chapter 41
Addition of angular momenta

Two spins make four states, but nature groups them as 3 + 1. That split drives magnetism and chemical bonding.

Turn the page