Case Study 2: Rotation Matrices in Practice — NMR and Molecular Spectroscopy

Overview

The Wigner $D$-matrices and rotation operators derived in this chapter are not merely elegant mathematical constructions — they are the working tools of several major experimental fields. This case study examines two domains where rotation matrices are indispensable: Nuclear Magnetic Resonance (NMR) spectroscopy and molecular rotational spectroscopy. In both cases, the abstract angular momentum algebra of Chapter 12 becomes the concrete language in which experimental results are expressed, analyzed, and predicted.

Part 1: Nuclear Magnetic Resonance — Rotating Spin-1/2

The Physical Setup

In NMR spectroscopy, the system of interest is the nuclear spin of atoms (most commonly hydrogen-1, with spin $I = 1/2$). The nuclei are placed in a strong static magnetic field $\mathbf{B}_0 = B_0 \hat{z}$, which splits the spin-up and spin-down energy levels by:

$$\Delta E = \gamma \hbar B_0 = \hbar \omega_0$$

where $\gamma$ is the gyromagnetic ratio (for protons, $\gamma/(2\pi) = 42.577$ MHz/T) and $\omega_0$ is the Larmor frequency. In a typical clinical MRI scanner at $B_0 = 1.5$ T, this gives $\omega_0/(2\pi) \approx 63.87$ MHz — a radiofrequency.

Thermal Equilibrium and the Density Matrix

At thermal equilibrium in a strong field, the nuclear spins are described by the density matrix (previewing Chapter 23):

$$\hat{\rho}_{\text{eq}} \approx \frac{1}{2}\hat{I} + \frac{\gamma\hbar B_0}{4k_BT}\hat{\sigma}_z$$

The first term is the maximally mixed state (no information); the second is a tiny polarization along $z$. For protons at room temperature in a 1.5 T field, the polarization is approximately $5 \times 10^{-6}$ — only about 5 parts per million of the spins are aligned with the field rather than against it. The entire NMR signal comes from this minuscule excess, which is why NMR requires sensitive detection and typically billions of nuclei.

Radiofrequency Pulses as Rotations

The key experimental tool in NMR is the radiofrequency (RF) pulse: a short burst of oscillating magnetic field $\mathbf{B}_1(t)$ perpendicular to $\mathbf{B}_0$, oscillating at or near the Larmor frequency $\omega_0$. In the rotating frame (a reference frame that rotates at $\omega_0$ about $z$), the static field effectively vanishes, and the RF pulse acts as a constant transverse field.

The effect of an RF pulse of duration $\tau$ is a rotation of the spin state:

$$|\psi\rangle \to \hat{R}(\hat{n}, \theta) |\psi\rangle = e^{-i\theta \hat{n} \cdot \hat{\boldsymbol{\sigma}}/2} |\psi\rangle$$

where $\theta = \gamma B_1 \tau$ is the flip angle and $\hat{n}$ is the direction of $B_1$ in the rotating frame. The rotation matrix is exactly the $D^{(1/2)}$ matrix from Section 12.7.

The $90°$ Pulse

The most common pulse in NMR is the $90°$ pulse (also called a $\pi/2$ pulse), which rotates the magnetization from $z$ to the transverse ($xy$) plane. If the pulse is applied along the $x$-axis in the rotating frame:

$$\hat{R}_x(\pi/2) = e^{-i(\pi/2)\hat{\sigma}_x/2} = \frac{1}{\sqrt{2}}\begin{pmatrix} 1 & -i \\ -i & 1 \end{pmatrix}$$

Acting on $|\uparrow\rangle$:

$$\hat{R}_x(\pi/2)|\uparrow\rangle = \frac{1}{\sqrt{2}}(|\uparrow\rangle - i|\downarrow\rangle)$$

This is an eigenstate of $\hat{\sigma}_y$ — the magnetization now points along $-y$ in the rotating frame. It precesses freely in the transverse plane, inducing a voltage in a receiver coil that constitutes the NMR signal (the free induction decay, or FID).

The $180°$ Pulse and Spin Echo

A $180°$ pulse ($\pi$ pulse) about the $x$-axis has the matrix:

$$\hat{R}_x(\pi) = e^{-i(\pi)\hat{\sigma}_x/2} = -i\hat{\sigma}_x = -i\begin{pmatrix} 0 & 1 \\ 1 & 0 \end{pmatrix}$$

This flips $|\uparrow\rangle \to -i|\downarrow\rangle$ and $|\downarrow\rangle \to -i|\uparrow\rangle$. In the rotating frame, it reflects the magnetization about the $x$-axis.

The celebrated spin echo experiment (Erwin Hahn, 1950) uses a $90°$-$\tau$-$180°$-$\tau$ pulse sequence:

1. A $90°$ pulse tips the magnetization into the $xy$ plane
2. Over time $\tau$, different spins precess at slightly different rates (due to field inhomogeneity), causing the transverse magnetization to dephase
3. A $180°$ pulse flips the magnetization, reversing the relative phases
4. After another interval $\tau$, the spins rephase, producing an echo — a spontaneous reappearance of the signal

The mathematics of the spin echo is entirely a rotation-matrix calculation:

$$\hat{R}_x(\pi) \cdot e^{-i\omega\tau\hat{\sigma}_z/2} \cdot \hat{R}_x(\pi/2)$$

for each spin with its own frequency $\omega$. The rotation matrices from Chapter 12 predict the exact shape and timing of the echo.

📊 By the Numbers: Modern MRI scanners use thousands of RF pulses per image, with carefully designed sequences (FLASH, EPI, spin echo, gradient echo, etc.) that exploit rotation matrix algebra to control contrast, speed, and resolution. A single brain MRI scan involves on the order of $10^6$ individual rotation operations on $\sim 10^{24}$ nuclear spins.

Higher-Spin NMR: Deuterium and Sodium

Not all NMR-active nuclei have spin-1/2. Deuterium ($^2$H) has spin $I = 1$, and sodium-23 ($^{23}$Na) has spin $I = 3/2$. For these nuclei, the rotation operators are the larger $D$-matrices:

• Deuterium ($I = 1$): The $3 \times 3$ rotation matrices $D^{(1)}$ describe pulse effects. The quadrupolar interaction (coupling between the nuclear electric quadrupole moment and local electric field gradients) introduces complications absent in spin-1/2 NMR, but the underlying framework remains the angular momentum algebra.

• Sodium ($I = 3/2$): The $4 \times 4$ $D^{(3/2)}$ matrices are needed. Sodium MRI is used clinically to image tissue viability after stroke, because intracellular sodium concentration changes rapidly in damaged tissue.

The general-$j$ matrix representations from Section 12.6 are essential for analyzing these higher-spin systems.

Part 2: Molecular Rotational Spectroscopy

Rigid Rotor Model

A diatomic molecule (like CO, HCl, or N$_2$) can rotate about its center of mass. In the simplest model — the rigid rotor — the rotational energy levels are:

$$E_J = \frac{\hbar^2}{2\mathcal{I}} J(J+1) = BJ(J+1)$$

where $\mathcal{I}$ is the moment of inertia, $B = \hbar^2/(2\mathcal{I})$ is the rotational constant, and $J$ is the rotational angular momentum quantum number ($J = 0, 1, 2, \ldots$).

This is precisely the eigenvalue spectrum of $\hat{J}^2$ from Section 12.5, with $j \to J$ (integer values only, since molecular rotation is orbital angular momentum). The degeneracy of each level is $2J + 1$ (the number of $M$-values, where $M = -J, \ldots, J$).

Pure Rotational Transitions

A molecule interacting with electromagnetic radiation can undergo transitions between rotational levels. The selection rules follow from the angular momentum algebra:

$$\Delta J = \pm 1, \quad \Delta M = 0, \pm 1$$

These selection rules are a direct consequence of the Wigner-Eckart theorem applied to the electric dipole operator (a rank-1 spherical tensor). The transition frequencies are:

$$\nu_{J \to J+1} = 2B(J+1)/h$$

The rotational spectrum is a series of equally spaced lines (in frequency) separated by $2B/h$. For CO, $B/h = 57.636$ GHz, giving spectral lines at 115.27 GHz, 230.54 GHz, 345.80 GHz, etc. — in the microwave and far-infrared regions.

📊 By the Numbers: The rotational constant $B$ directly gives the bond length of the molecule: $B = \hbar^2/(2\mu r_0^2)$, where $\mu$ is the reduced mass and $r_0$ is the equilibrium bond length. For CO, measuring the rotational spectrum gives $r_0 = 1.131$ \AA with extraordinary precision (better than 0.001 \AA). Rotational spectroscopy is one of the most precise structure-determination methods in all of chemistry.

Symmetric and Asymmetric Tops

For polyatomic molecules, the rotational problem becomes richer:

Symmetric top (like NH$_3$ or CH$_3$Cl): The molecule has two equal moments of inertia. The energy levels depend on two quantum numbers:

$$E_{JK} = BJ(J+1) + (A - B)K^2$$

where $J$ is the total rotational angular momentum quantum number, $K$ is the projection of $\hat{J}$ onto the molecular symmetry axis ($K = -J, \ldots, J$), and $A$, $B$ are rotational constants related to the two distinct moments of inertia.

The quantum number $K$ is the molecular analogue of $m$: it quantifies the angular momentum about the molecular axis. The matrix representations from Section 12.6 directly apply.

Asymmetric top (like H$_2$O): All three moments of inertia are different, and there is no "good" quantum number analogous to $K$. The energy levels must be found by diagonalizing the rotational Hamiltonian in the symmetric top basis — a problem that requires the full matrix representation machinery of Section 12.6.

Rotation Matrices in Spectroscopy

The Wigner $D$-matrices appear explicitly in molecular spectroscopy in several ways:

1. Wavefunctions of the symmetric top: The rotational wavefunctions are proportional to $D$-matrix elements:

$$\Psi_{JMK}(\alpha, \beta, \gamma) = \sqrt{\frac{2J+1}{8\pi^2}} D^{(J)*}_{MK}(\alpha, \beta, \gamma)$$

where $(\alpha, \beta, \gamma)$ are Euler angles describing the orientation of the molecule-fixed frame relative to the lab frame.

2. Transition intensities: The line strengths for rotational transitions involve integrals over products of three $D$-matrices, which are evaluated using the Clebsch-Gordan series (Chapter 14):

$$\int D^{(J_1)*}_{M_1 K_1} D^{(1)}_{q_1 q_2} D^{(J_2)}_{M_2 K_2} \, d\Omega = \frac{8\pi^2}{2J_1+1} \langle J_2, M_2; 1, q_1 | J_1, M_1 \rangle \langle J_2, K_2; 1, q_2 | J_1, K_1 \rangle$$

3. Orientation averaging: In a gas-phase sample, molecules are randomly oriented. Computing the average signal requires integrating over all orientations using the orthogonality properties of the $D$-matrices.

Part 3: Magnetic Resonance Imaging — From Algebra to Clinical Diagnosis

How MRI Works

MRI (Magnetic Resonance Imaging) is the most sophisticated practical application of spin-1/2 angular momentum algebra. The essential steps are:

1. Polarization: Place the patient in a strong static field ($B_0 = 1.5$ or $3$ T). Proton spins partially align with the field.

2. Excitation: Apply an RF pulse (a rotation) to tip the magnetization into the transverse plane.

3. Spatial encoding: Apply spatially varying gradient fields $\mathbf{G} = (G_x x, G_y y, G_z z)$ so that the Larmor frequency varies with position: $\omega(\mathbf{r}) = \gamma(B_0 + \mathbf{G} \cdot \mathbf{r})$.

4. Detection: Record the time-varying transverse magnetization (the FID or echo).

5. Reconstruction: Fourier-transform the signal to recover the spatial distribution of protons (the image).

Every step of this process is described by the angular momentum algebra:

• Excitation is a rotation: $D^{(1/2)}$ matrix
• Precession in a gradient is a position-dependent rotation about $z$: $e^{-i\omega(\mathbf{r})t\hat{\sigma}_z/2}$
• The spin echo refocusing is a $\pi$ rotation: $D^{(1/2)}$ with $\beta = \pi$
• Relaxation (T$_1$ and T$_2$ processes) requires the density matrix formalism (Chapter 23) but the underlying unitary evolution is still described by rotation operators

Quantitative Example: The Spin Echo in MRI

Consider a collection of proton spins at different positions, each experiencing a slightly different field due to the gradient. In the rotating frame, spin $k$ at position $z_k$ precesses at frequency $\delta\omega_k = \gamma G_z z_k$.

After the $90°_x$ pulse at $t = 0$, each spin is in the state:

$$|\psi_k(0^+)\rangle = \frac{1}{\sqrt{2}}(|\uparrow\rangle - i|\downarrow\rangle)$$

At time $\tau$ (just before the $180°$ pulse):

$$|\psi_k(\tau^-)\rangle = \frac{1}{\sqrt{2}}(e^{-i\delta\omega_k\tau/2}|\uparrow\rangle - ie^{+i\delta\omega_k\tau/2}|\downarrow\rangle)$$

The $180°_x$ pulse transforms this to:

$$|\psi_k(\tau^+)\rangle = \frac{1}{\sqrt{2}}(-ie^{+i\delta\omega_k\tau/2}|\uparrow\rangle + e^{-i\delta\omega_k\tau/2}|\downarrow\rangle)$$

(up to an overall phase). At time $2\tau$:

$$|\psi_k(2\tau)\rangle = \frac{1}{\sqrt{2}}(-ie^{+i\delta\omega_k\tau/2}e^{-i\delta\omega_k\tau/2}|\uparrow\rangle + e^{-i\delta\omega_k\tau/2}e^{+i\delta\omega_k\tau/2}|\downarrow\rangle) = \frac{1}{\sqrt{2}}(-i|\uparrow\rangle + |\downarrow\rangle)$$

The position-dependent phase $\delta\omega_k\tau$ has cancelled! All spins are back in phase at $t = 2\tau$, regardless of their position. This is the spin echo — and the entire calculation is rotation matrix algebra.

🧪 Experiment: The 2003 Nobel Prize in Physiology or Medicine was awarded to Paul Lauterbur and Peter Mansfield for their contributions to MRI. The underlying physics — nuclear spin precession, rotation by RF pulses, and spatial encoding by gradients — is entirely described by the angular momentum algebra of this chapter, applied to $j = 1/2$.

Part 4: Beyond Spin-1/2 — Rotation Matrices in Particle Physics

Polarization of Photons

Photons are spin-1 particles, but with a crucial restriction: for massless particles, only the helicity states $m = \pm 1$ are physical (not $m = 0$). The rotation matrices for photon polarization are submatrices of $D^{(1)}$, which describe how polarization states transform under rotations.

The Jones calculus used in optics — where polarization states are described by two-component vectors and optical elements by $2 \times 2$ matrices — is a restriction of the spin-1 rotation formalism to the two physical helicity states.

Angular Distributions in Scattering

When particles scatter, the angular distribution of the final-state particles is governed by the Wigner $d$-matrices. For a reaction where a spin-$j$ particle is produced and decays, the angular distribution relative to the production axis is:

$$\frac{dN}{d(\cos\theta)} \propto |d^{(j)}_{m_i, m_f}(\theta)|^2$$

This formula is used routinely in particle physics experiments at CERN, Fermilab, KEK, and other facilities to determine the spin and parity of newly discovered particles. The 2012 discovery of the Higgs boson at the LHC included an analysis of its decay angular distributions using $d^{(0)}(\theta) = 1$ (spin-0) vs. $d^{(1)}(\theta)$ (spin-1) vs. $d^{(2)}(\theta)$ (spin-2) — the rotation matrices from Chapter 12 were directly used to establish that the Higgs is a spin-0 particle.

Discussion Questions

1. In NMR, the RF pulse rotates the spin state by a precise angle. What determines the precision of this rotation in practice? What happens if the flip angle is slightly off from the intended $90°$?

2. The spin echo cancels the effects of static field inhomogeneity but not the effects of T$_2$ relaxation (random, fluctuating fields). Explain qualitatively why static and fluctuating perturbations are treated differently by the echo.

3. Molecular rotational spectroscopy is restricted to integer $J$ because molecular rotation is orbital angular momentum. If a molecule had an unpaired electron spin ($S = 1/2$), the total angular momentum would be $\hat{F} = \hat{J} + \hat{S}$, which can be half-integer. Would this change the rotational spectrum? How?

4. The Higgs boson spin determination used Wigner $d$-matrices. Why is angular distribution analysis more powerful for spin determination than simply counting the number of polarization states?

5. MRI routinely achieves sub-millimeter spatial resolution while manipulating individual nuclear spins with nanosecond precision. In what sense is MRI a "quantum technology"? Is the quantum nature of spin essential, or could a classical model of magnetic dipoles explain MRI equally well?

Key Takeaways from This Case Study

• NMR spectroscopy uses $D^{(1/2)}$ rotation matrices to describe the effect of RF pulses on nuclear spins. The $90°$ and $180°$ pulses are specific rotations whose effects are computed using the $j = 1/2$ matrices from Section 12.6.
• The spin echo is a purely quantum-mechanical phenomenon where a $180°$ rotation (a $D$-matrix operation) reverses dephasing caused by field inhomogeneity.
• Molecular spectroscopy uses the eigenvalue spectrum $E_J = BJ(J+1)$ — directly from the angular momentum algebra — to determine molecular structure with extraordinary precision.
• The Wigner $D$-matrices are the wavefunctions of symmetric-top molecules and appear in transition intensity calculations.
• Particle physics uses angular distributions described by $d^{(j)}(\theta)$ to determine the spin of particles, including the Higgs boson.
• The abstract algebra of Chapter 12 has concrete, quantitative applications across physics, chemistry, medicine, and technology.