Chapter 9: Electromagnetic Properties and Transitions

"The gamma ray is the messenger. What it tells us about the nucleus depends entirely on how carefully we listen." — Aage Bohr, Nuclear Structure (1969)

Everything we know about the interior of atomic nuclei comes from the particles and radiation they emit or absorb. Among these messengers, gamma rays hold a privileged position. Unlike alpha or beta particles, which involve the nuclear force or the weak interaction, gamma-ray emission is a purely electromagnetic process. We understand electromagnetism to extraordinary precision. This means that when we measure a gamma-ray transition rate, we are measuring a nuclear matrix element with a probe we trust completely. The electromagnetic field does the asking; the nucleus provides the answer.

This chapter develops the formalism connecting nuclear structure to the electromagnetic radiation nuclei emit. We will derive the selection rules that determine which transitions are allowed, calculate the single-particle estimates that serve as benchmarks, and explore the experimental techniques — from lifetime measurements to Coulomb excitation to modern tracking arrays — that extract the nuclear structure information encoded in gamma rays. The result is a complete framework: given a structural model (shell model, collective model, or anything in between), we can predict electromagnetic observables; given measured observables, we can test and constrain the models.

The punchline is powerful: a single number — the reduced transition probability $B(E\lambda)$ or $B(M\lambda)$, expressed in Weisskopf units — tells you immediately whether a transition involves a single nucleon rearranging itself within the mean field or many nucleons moving collectively. This is how we "see" nuclear structure.

9.1 Multipole Moments: A Review and Extension

9.1.1 The Multipole Expansion of the Electromagnetic Field

In Chapter 2, we introduced the electric quadrupole moment as a measure of nuclear charge deformation. Here we develop the full multipole expansion systematically, because it forms the mathematical backbone of everything that follows.

Any charge distribution $\rho(\mathbf{r})$ produces an electrostatic potential that, at points outside the distribution, can be expanded in spherical harmonics:

$$\Phi(\mathbf{r}) = \frac{1}{4\pi\epsilon_0} \sum_{\lambda=0}^{\infty} \sum_{\mu=-\lambda}^{\lambda} \frac{4\pi}{2\lambda+1} \frac{Q_{\lambda\mu}}{r^{\lambda+1}} Y_{\lambda\mu}(\theta,\phi)$$

where the electric multipole moments are:

$$Q_{\lambda\mu} = \int \rho(\mathbf{r}') r'^{\lambda} Y^*_{\lambda\mu}(\theta',\phi') \, d^3r'$$

The $\lambda = 0$ term is the total charge (monopole). The $\lambda = 1$ terms give the electric dipole moment. The $\lambda = 2$ terms give the quadrupole moment introduced in Chapter 2, and so on.

For a nucleus with $Z$ protons, the charge density operator is:

$$\hat{\rho}(\mathbf{r}) = e \sum_{i=1}^{A} \frac{1+\tau_{z,i}}{2} \delta(\mathbf{r} - \mathbf{r}_i)$$

where the isospin projection operator $(1+\tau_z)/2$ selects protons. The corresponding electric multipole operator becomes:

$$\hat{Q}_{\lambda\mu} = e \sum_{i=1}^{A} \frac{1+\tau_{z,i}}{2} r_i^{\lambda} Y_{\lambda\mu}(\theta_i,\phi_i)$$

Key insight: Electric multipole operators depend on $r^\lambda$, so higher multipoles are increasingly sensitive to the surface of the nuclear charge distribution, where $r$ is largest. This is why $E2$ transitions are such powerful probes of nuclear deformation.

9.1.2 Physical Interpretation of Electric Multipole Moments

Before proceeding to the magnetic case, let us develop physical intuition for the electric multipole moments. Each multipole order probes a specific feature of the nuclear charge distribution:

• Monopole ($\lambda = 0$): $Q_{00} = Ze/\sqrt{4\pi}$ — just the total charge. All nuclei with the same $Z$ have the same monopole moment. The monopole cannot radiate (no $E0$ photon), but $E0$ transitions proceed by internal conversion and probe changes in the mean-square charge radius between states.

• Dipole ($\lambda = 1$): The electric dipole moment $\mathbf{d} = e\sum_i \mathbf{r}_i$ measures the displacement of the charge centroid from the origin. In the center-of-mass frame, $\mathbf{d}_{\text{cm}} = e(Z/A)\sum_i \mathbf{r}_i = 0$ by construction, which is why static nuclear electric dipole moments vanish (to observe a permanent EDM, one would need a violation of parity and time-reversal symmetry — see Chapter 32). Dynamic $E1$ transitions connect to the giant dipole resonance, the collective oscillation of protons against neutrons.

• Quadrupole ($\lambda = 2$): The spectroscopic quadrupole moment $Q = \sqrt{16\pi/5} \, Q_{20}$ measures the deviation of the charge distribution from spherical symmetry. Prolate nuclei ($Q > 0$) are elongated along the symmetry axis; oblate nuclei ($Q < 0$) are flattened. The quadrupole moment is the most commonly measured static moment after the magnetic dipole, and $E2$ transitions are the workhorse of nuclear structure spectroscopy.

• Octupole ($\lambda = 3$): The octupole moment probes pear-shaped (reflection-asymmetric) distributions. Nonzero octupole correlations have been identified in nuclei near $Z = 88$, $N = 134$ (e.g., ${}^{224}$Ra) and near $Z = 56$, $N = 88$, providing evidence for static octupole deformation in certain regions of the nuclear chart.

The general pattern is that the $\lambda$th multipole is sensitive to variations in the charge distribution on an angular scale of $\sim \pi/\lambda$. Higher multipoles probe finer details but contribute less to the external field (they fall off as $r^{-(\lambda+1)}$).

9.1.3 Magnetic Multipole Moments

Nuclear current distributions generate magnetic fields. The magnetic multipole moments arise from two sources: the orbital motion of protons (which constitute a current loop) and the intrinsic magnetic moments of all nucleons (both protons and neutrons have anomalous magnetic moments).

The magnetic dipole moment operator is:

$$\hat{\boldsymbol{\mu}} = \mu_N \sum_{i=1}^{A} \left[ g_{\ell,i} \hat{\mathbf{l}}_i + g_{s,i} \hat{\mathbf{s}}_i \right]$$

where $\mu_N = e\hbar/(2m_p) = 3.152 \times 10^{-8}$ eV/T is the nuclear magneton, and the $g$-factors are:

The neutron's nonzero $g_s$ — despite having no net charge — reflects its internal quark structure and has profound consequences: neutrons participate fully in magnetic transitions.

9.1.4 The Schmidt Lines: A Test of the Shell Model

For a single nucleon with quantum numbers $(\ell, j)$, the magnetic moment of a nuclear state with $I = j$ is:

$$\mu = g_j \cdot j \cdot \mu_N$$

where the effective $g$-factor depends on whether $j = \ell + 1/2$ or $j = \ell - 1/2$:

$$g_j = \begin{cases} g_\ell + \frac{g_s - g_\ell}{2\ell+1} & j = \ell + \frac{1}{2} \\ g_\ell \frac{2j+3}{2(j+1)} - g_s \frac{1}{2(j+1)} & j = \ell - \frac{1}{2} \end{cases}$$

Plotting these predictions against measured nuclear magnetic moments yields the famous Schmidt lines — two curves (one for each coupling, separate for protons and neutrons) that bracket the experimental values. The measured moments cluster between the Schmidt lines, pulled toward the center by the residual nucleon-nucleon interactions and meson exchange currents that the extreme single-particle model ignores. The pattern confirms the shell model as a zeroth-order description while quantifying the corrections needed.

Example: The ground state of ${}^{17}$O has $I^\pi = 5/2^+$, corresponding to a single neutron in the $1d_{5/2}$ orbital. The Schmidt prediction is $\mu = -1.913\,\mu_N$; the measured value is $\mu = -1.894\,\mu_N$. The agreement to 1% is exceptional and confirms the single-particle picture for this doubly-magic-plus-one nucleus.

9.2 Electromagnetic Transition Operators

9.2.1 From Static Moments to Transition Rates

Static multipole moments describe the charge and current distributions of a nucleus in a single state. Electromagnetic transitions connect two different states through the emission (or absorption) of a photon. The formalism requires time-dependent perturbation theory and the quantized radiation field, developed in Chapter 5.

From Fermi's golden rule, the transition rate for emission of a photon of multipolarity $(\sigma\lambda)$ — where $\sigma = E$ (electric) or $M$ (magnetic) and $\lambda$ is the multipole order — is:

$$T(\sigma\lambda; I_i \to I_f) = \frac{8\pi(\lambda+1)}{\lambda[(2\lambda+1)!!]^2} \left(\frac{E_\gamma}{\hbar c}\right)^{2\lambda+1} \frac{B(\sigma\lambda; I_i \to I_f)}{\hbar}$$

Here $E_\gamma$ is the transition energy, $(2\lambda+1)!! = 1 \cdot 3 \cdot 5 \cdots (2\lambda+1)$ is the double factorial, and $B(\sigma\lambda; I_i \to I_f)$ is the reduced transition probability, defined as:

$$B(\sigma\lambda; I_i \to I_f) = \frac{1}{2I_i+1} \left| \langle I_f \| \hat{\mathcal{O}}(\sigma\lambda) \| I_i \rangle \right|^2$$

where the double-bar notation denotes the reduced matrix element (Wigner-Eckart theorem, Chapter 5). The factor $1/(2I_i+1)$ averages over initial magnetic substates.

This is one of the most important equations in nuclear structure physics. It cleanly separates the physics into: - Kinematics: the $E_\gamma^{2\lambda+1}$ energy dependence and numerical prefactors - Nuclear structure: the $B(\sigma\lambda)$ value, which encodes everything about the nuclear wavefunctions

9.2.2 Electric Transition Operators

The electric multipole transition operator of order $\lambda$ is:

$$\hat{\mathcal{O}}(E\lambda,\mu) = e \sum_{i=1}^{A} e_{\text{eff},i} \, r_i^{\lambda} \, Y_{\lambda\mu}(\hat{r}_i)$$

where $e_{\text{eff},i}$ is the effective charge of nucleon $i$. In the simplest approximation, $e_{\text{eff}} = 1$ for protons and $0$ for neutrons. However, because the core nucleons are not truly inert — they polarize when a valence nucleon moves — effective charges are used in shell-model calculations:

$$e_p^{\text{eff}} \approx 1.5e, \quad e_n^{\text{eff}} \approx 0.5e$$

These values, determined empirically, account for core polarization. The fact that neutrons carry a nonzero effective charge is physically crucial: a neutron moving from one shell-model orbital to another does create an oscillating charge distribution because the inert core adjusts in response.

9.2.3 Magnetic Transition Operators

The magnetic multipole transition operator is:

$$\hat{\mathcal{O}}(M\lambda,\mu) = \frac{\mu_N}{\lambda+1} \sum_{i=1}^{A} \left[ g_{\ell,i} \sqrt{\lambda(2\lambda+1)} \left\{ \hat{\mathbf{l}}_i \otimes \mathbf{Y}_{\lambda-1}(\hat{r}_i) \right\}_{\lambda\mu} r_i^{\lambda-1} + g_{s,i} \, c_\lambda \nabla_i \left( r_i^{\lambda} Y_{\lambda\mu}(\hat{r}_i) \right) \cdot \hat{\mathbf{s}}_i \right]$$

For the crucial $M1$ case ($\lambda = 1$), this simplifies to:

$$\hat{\mathcal{O}}(M1,\mu) = \frac{3}{4\pi} \mu_N \sum_{i=1}^{A} \left[ g_{\ell,i} \hat{\ell}_{\mu,i} + g_{s,i} \hat{s}_{\mu,i} \right]$$

which is proportional to the magnetic dipole operator. Thus $M1$ transitions probe the same physics as static magnetic moments — the interplay of orbital and spin angular momentum.

Physical picture: Electric transitions probe the charge distribution — where the protons are. Magnetic transitions probe the current distribution — how the nucleons are moving and spinning. The two types of transition are complementary windows into nuclear structure.

9.3 Selection Rules

9.3.1 Angular Momentum Selection Rules

The emitted photon carries angular momentum $\lambda\hbar$. Conservation of angular momentum requires:

$$\mathbf{I}_i = \mathbf{I}_f + \boldsymbol{\lambda}$$

which gives the triangular condition:

$$|I_i - I_f| \leq \lambda \leq I_i + I_f$$

with $\lambda \geq 1$ (there is no $\lambda = 0$ radiation — a photon must carry at least one unit of angular momentum). This immediately tells us:

• $0^+ \to 0^+$ transitions are strictly forbidden by single-photon emission. (The photon cannot carry zero angular momentum.) These states can only de-excite by internal conversion, internal pair creation, or two-photon emission.
• For $I_i = I_f = I$, the minimum multipole order is $\lambda = 1$.
• For $\Delta I = I_i - I_f$, the lowest allowed multipole order is $\lambda_{\min} = |\Delta I|$ (or $\lambda_{\min} = 1$ if $\Delta I = 0$).

9.3.2 Parity Selection Rules

The parity carried by the radiation depends on both the multipole order $\lambda$ and the type (electric or magnetic):

$$\pi(\text{radiation}) = \begin{cases} (-1)^{\lambda} & \text{for } E\lambda \\ (-1)^{\lambda+1} & \text{for } M\lambda \end{cases}$$

Since parity is conserved in electromagnetic interactions:

$$\pi_i = \pi_f \cdot \pi(\text{radiation})$$

This gives the combined selection rules:

9.3.3 Worked Example: The Level Scheme of ${}^{60}$Ni

Let us apply these selection rules systematically to a real nucleus. The low-lying level scheme of ${}^{60}$Ni includes:

Consider each possible transition:

• $2^+_1 \to 0^+$: $\Delta I = 2$, no parity change. The lowest multipole satisfying both conditions is $E2$ ($\lambda = 2$, parity $(-1)^2 = +1$). This is the dominant transition, with a $B(E2)$ value of 635 $e^2$fm$^4$ (about 6 W.u.), indicating moderate collectivity appropriate for a semi-magic $Z = 28$ nucleus.

• $2^+_2 \to 2^+_1$: $\Delta I = 0$, no parity change. Both $M1$ and $E2$ are allowed. This is a case where $E2/M1$ mixing occurs, and the mixing ratio $\delta$ must be measured experimentally.

• $2^+_2 \to 0^+$: $\Delta I = 2$, no parity change. Pure $E2$.

• $4^+ \to 2^+_1$: $\Delta I = 2$, no parity change. Pure $E2$.

• $4^+ \to 2^+_2$: $\Delta I = 2$, no parity change. Pure $E2$, but this transition competes with the $4^+ \to 2^+_1$ transition. The branching ratio depends on the relative $B(E2)$ values and the $E_\gamma^5$ energy factor.

• $4^+ \to 0^+$: $\Delta I = 4$, no parity change. The lowest allowed multipole is $E4$ — highly suppressed by the $E_\gamma^9$ dependence. This transition is almost never observed.

This example illustrates a general principle: in even-even nuclei, the low-lying level scheme is dominated by $E2$ transitions because the ground-state band consists of even-parity states connected by $\Delta I = 2$ steps.

9.3.4 Dominance of the Lowest Multipole

A transition between two states may satisfy the selection rules for several multipolarities simultaneously. For example, a $2^+ \to 2^+$ transition allows $M1$, $E2$, $M3$, and $E4$. In practice, the lowest permitted multipole almost always dominates because of the strong energy dependence $T \propto E_\gamma^{2\lambda+1}$. Each unit increase in $\lambda$ suppresses the rate by roughly a factor of $(E_\gamma R / \hbar c)^2 \sim 10^{-5}$ to $10^{-7}$ for typical nuclear transition energies ($E_\gamma \sim 1$ MeV) and nuclear radii ($R \sim 5$ fm).

The one systematic exception is the competition between $E2$ and $M1$ for transitions with $\Delta I = 0, 1$ and no parity change. While $M1$ is lower multipole order, $E2$ can be enhanced by collectivity (many nucleons contributing coherently), making the $E2/M1$ mixing ratio $\delta$ an important measurable quantity:

$$\delta = \frac{\langle f \| E2 \| i \rangle}{\langle f \| M1 \| i \rangle}$$

Measured $\delta$ values provide direct information about the interplay between single-particle and collective degrees of freedom.

Example: In the ground-state rotational band of ${}^{166}$Er, the $4^+ \to 2^+$ transition at 265 keV is predominantly $E2$ with $\delta(E2/M1) = -14.4$. The enormous $E2$ strength reflects collective rotation of the entire deformed nucleus — a classic signature of the rotational model (Chapter 8).

9.4 Transition Rates and Weisskopf Estimates

9.4.1 The Single-Particle Estimate

Victor Weisskopf (1951) introduced a benchmark for transition rates by calculating the $B(\sigma\lambda)$ value expected for a single proton making a transition between shell-model orbitals. These Weisskopf estimates (also called single-particle estimates) provide a yardstick against which all measured transition rates are compared.

For an electric transition, Weisskopf assumed: 1. A single proton changes its orbital state 2. The radial wavefunctions are constant inside the nuclear radius $R$ and zero outside 3. $R = r_0 A^{1/3}$ with $r_0 = 1.2$ fm

The resulting estimate is:

$$B_W(E\lambda) = \frac{1}{4\pi} \left(\frac{3}{\lambda+3}\right)^2 R^{2\lambda} \, e^2$$

Substituting $R = 1.2 A^{1/3}$ fm and expressing in units of $e^2 \text{fm}^{2\lambda}$:

$$B_W(E\lambda) = \frac{1}{4\pi} \left(\frac{3}{\lambda+3}\right)^2 (1.2)^{2\lambda} A^{2\lambda/3} \quad [e^2 \text{fm}^{2\lambda}]$$

The first few values are:

9.4.2 Derivation of the Weisskopf Estimate

Let us derive the $E\lambda$ estimate in detail. We need:

$$B_W(E\lambda) = \frac{1}{2I_i+1} |\langle I_f \| \hat{\mathcal{O}}(E\lambda) \| I_i \rangle|^2$$

For a single proton, the reduced matrix element involves the radial integral $\langle n_f \ell_f | r^\lambda | n_i \ell_i \rangle$ and an angular/spin coupling factor. Weisskopf's approximation replaces the radial integral with its uniform-density estimate:

$$\int_0^R r^{\lambda+2} dr = \frac{R^{\lambda+3}}{\lambda+3}$$

normalized by the volume integral $\int_0^R r^2 dr = R^3/3$, giving:

$$\langle r^\lambda \rangle_{\text{sp}} = \frac{3}{\lambda+3} R^\lambda$$

For the angular coupling, Weisskopf took the statistical average over possible angular momentum couplings, yielding a factor of $1/(4\pi)$ when squared. Combining these:

$$B_W(E\lambda) = \frac{1}{4\pi}\left(\frac{3}{\lambda+3}\right)^2 R^{2\lambda} e^2$$

The corresponding transition rate, in inverse seconds, is:

$$T_W(E\lambda) = \frac{8\pi(\lambda+1)}{\lambda[(2\lambda+1)!!]^2} \left(\frac{E_\gamma}{\hbar c}\right)^{2\lambda+1} \frac{c}{\hbar} \cdot \frac{1}{4\pi}\left(\frac{3}{\lambda+3}\right)^2 (1.2 A^{1/3})^{2\lambda} e^2$$

Evaluating numerically with $\hbar c = 197.3$ MeV fm and $e^2 = 1.440$ MeV fm:

$$T_W(E1) = 1.023 \times 10^{14} A^{2/3} E_\gamma^3 \quad \text{s}^{-1}$$

$$T_W(E2) = 7.28 \times 10^{7} A^{4/3} E_\gamma^5 \quad \text{s}^{-1}$$

$$T_W(E3) = 3.39 \times 10^{1} A^{2} E_\gamma^7 \quad \text{s}^{-1}$$

where $E_\gamma$ is in MeV.

9.4.3 Worked Example: The $2^+ \to 0^+$ Transition in ${}^{152}$Sm

Let us work through a complete numerical example. The first excited state of ${}^{152}$Sm is at 121.8 keV with $I^\pi = 2^+$. We want to calculate the Weisskopf estimate for the $E2$ transition to the $0^+$ ground state and compare to the measured value.

Step 1: Weisskopf $B$ value.

$$B_W(E2) = \frac{1}{4\pi}\left(\frac{3}{5}\right)^2 R^4, \quad R = 1.2 \times 152^{1/3} = 1.2 \times 5.337 = 6.40 \text{ fm}$$

$$B_W(E2) = \frac{1}{4\pi} \times 0.36 \times 6.40^4 = 0.02865 \times 0.36 \times 1678 = 17.3 \text{ } e^2\text{fm}^4$$

Step 2: Weisskopf transition rate.

$$T_W(E2) = 7.28 \times 10^7 \times 152^{4/3} \times (0.1218)^5$$

$152^{4/3} = 152 \times 152^{1/3} = 152 \times 5.337 = 811$

$(0.1218)^5 = 2.68 \times 10^{-5}$

$$T_W(E2) = 7.28 \times 10^7 \times 811 \times 2.68 \times 10^{-5} = 1.58 \times 10^6 \text{ s}^{-1}$$

This corresponds to a Weisskopf half-life of $t_{1/2}^W = \ln 2 / T_W = 4.4 \times 10^{-7}$ s $= 440$ ns.

Step 3: Comparison to experiment.

The measured lifetime of the $2^+_1$ state is $\tau = 1.01$ ns, giving $T_{\exp} = 1/\tau = 9.9 \times 10^8$ s$^{-1}$.

The ratio of measured to Weisskopf rate is:

$$\frac{T_{\exp}}{T_W} = \frac{9.9 \times 10^8}{1.58 \times 10^6} = 626$$

So the measured transition is 626 times faster than the single-particle estimate, or equivalently, $B(E2) = 626 \times B_W(E2) = 626 \times 17.3 = 10,830$ $e^2$fm$^4$. (The evaluated value from ENSDF is 8,650 $e^2$fm$^4$; the discrepancy arises partly from the internal conversion correction — the total rate includes both $\gamma$ and conversion electron channels, with $\alpha_{tot} \approx 1.9$ for this transition, so the $\gamma$-ray rate alone is only about 35% of the total rate.)

After correcting for internal conversion: $T_\gamma = T_{total}/(1+\alpha) = 9.9 \times 10^8 / 2.9 = 3.4 \times 10^8$ s$^{-1}$, giving $B(E2) \approx 215 \times B_W = 3,700$ $e^2$fm$^4$... but this is the emission (downward) $B$ value, $B(E2; 2^+ \to 0^+)$. The excitation (upward) $B$ value used in data tables is:

$$B(E2\uparrow) = \frac{2I_i+1}{2I_f+1} B(E2\downarrow) = \frac{1}{5} B(E2\downarrow)$$

Wait — the convention matters. The tabulated value $B(E2; 0^+ \to 2^+) = 8650$ $e^2$fm$^4$ is the upward $B$ value. The downward value is $B(E2; 2^+ \to 0^+) = (2 \times 0 + 1)/(2 \times 2 + 1) \times 8650 = 1730$ $e^2$fm$^4$. This illustrates a crucial point: always check the direction convention when comparing $B$ values from different sources.

The bottom line: $B(E2\uparrow) = 8650$ $e^2$fm$^4$ $\approx 500$ W.u. — a massively collective transition. ${}^{152}$Sm is a well-deformed rotor, and this enormous $B(E2)$ value was one of the key pieces of evidence for collective rotation in nuclei.

9.4.4 Magnetic Weisskopf Estimates

For magnetic transitions, the single-particle estimate gives:

$$B_W(M\lambda) = \frac{10}{\pi} \left(\frac{3}{\lambda+3}\right)^2 R^{2\lambda-2} \mu_N^2$$

The corresponding transition rates are:

$$T_W(M1) = 3.15 \times 10^{13} E_\gamma^3 \quad \text{s}^{-1}$$

$$T_W(M2) = 2.24 \times 10^{7} A^{2/3} E_\gamma^5 \quad \text{s}^{-1}$$

$$T_W(M3) = 1.04 \times 10^{1} A^{4/3} E_\gamma^7 \quad \text{s}^{-1}$$

Note that $M\lambda$ rates are roughly $10 \times (R/\hbar c)^{-2} \approx 0.3$ times the corresponding $E\lambda$ rates — magnetic transitions are systematically slower than electric transitions of the same multipole order.

9.4.4 The Weisskopf Unit as a Diagnostic Tool

Expressing measured $B$ values in Weisskopf units (W.u.) — that is, as multiples of $B_W$ — immediately reveals the character of the transition:

This diagnostic is extraordinarily powerful. Some examples from real nuclei:

Single-particle transitions ($\sim$ 1 W.u.): - $B(E2; 2^+_1 \to 0^+_1)$ in ${}^{208}$Pb: $0.028$ $e^2$b$^2$ $= 4.8$ W.u. Near doubly-magic ${}^{208}$Pb, the first excited state involves moving a single nucleon — the shell model works.

Collective transitions ($\gg$ 1 W.u.): - $B(E2; 2^+_1 \to 0^+_1)$ in ${}^{168}$Er: $3.46$ $e^2$b$^2$ $= 204$ W.u. This is the signature of a well-deformed rotor: the entire nucleus — all 168 nucleons — participates coherently in the rotational motion.

Hindered transitions ($\ll$ 1 W.u.): - Isospin-forbidden $E1$ transitions: typically $10^{-4}$ to $10^{-6}$ W.u. The $E1$ operator cannot change isospin ($\Delta T = 0$) in the long-wavelength limit, so $E1$ transitions between states of the same isospin in $N \approx Z$ nuclei are strongly suppressed. This was a major puzzle in early nuclear physics: $E1$ transitions should be fast, but they are not.

Why $E1$ transitions are slow: The $E1$ operator, to leading order, is proportional to $\sum_i (e_i - Ze/A) r_i Y_{1\mu}$, which is the center-of-mass displacement. Since the center of mass does not move in internal excitations, the leading $E1$ matrix element vanishes. The surviving $E1$ strength comes from higher-order corrections (isovector contributions), which are small. This is why $E1$ transitions are typically $10^{-3}$ to $10^{-5}$ W.u. despite being the "lowest" electric multipole.

9.4.5 Connecting to Models

The power of the $B$ value framework is its model independence. We can calculate $B(\sigma\lambda)$ values within:

• The shell model: Single-particle matrix elements, effective charges, configuration mixing. Good for nuclei near closed shells. Predicts $B$ values of order $\sim 1$ W.u. for single-particle transitions and moderate enhancements for collective transitions if the model space is large enough.

• The rotational model (Chapter 8): For a symmetric rotor,

$$B(E2; I_i \to I_f) = \frac{5}{16\pi} e^2 Q_0^2 \langle I_i K 2 0 | I_f K \rangle^2$$

where $Q_0$ is the intrinsic quadrupole moment and $\langle \, | \, \rangle$ is a Clebsch-Gordan coefficient. This gives the characteristic ratio $B(E2; 4^+ \to 2^+) / B(E2; 2^+ \to 0^+) = 10/7 = 1.43$ for a $K = 0$ band — a ratio that has been verified in hundreds of deformed nuclei.

• The vibrational model (Chapter 8): Transitions within a vibrational band follow $B(E2; n+1 \to n) \propto (n+1)$, the boson number enhancement factor.

9.5 Internal Conversion

9.5.1 The Process

Electromagnetic de-excitation of an excited nuclear state need not produce a gamma ray. An alternative process is internal conversion, in which the nuclear transition energy is transferred directly to a bound atomic electron, which is then ejected from the atom. The ejected electron has kinetic energy:

$$T_e = E_\gamma - B_e$$

where $E_\gamma$ is the nuclear transition energy and $B_e$ is the binding energy of the atomic shell from which the electron was ejected ($K$, $L_I$, $L_{II}$, etc.).

This is not a two-step process (gamma emission followed by photoelectric absorption). The electron interacts directly with the electromagnetic field of the transitioning nucleus. The conversion electron spectrum shows discrete lines — one for each atomic shell — superimposed on any continuum background, providing a spectroscopic fingerprint.

9.5.2 Internal Conversion Coefficients

The internal conversion coefficient $\alpha$ is defined as the ratio of the conversion electron rate to the gamma-ray rate:

$$\alpha = \frac{T_e}{T_\gamma}$$

The total transition rate is:

$$T_{\text{total}} = T_\gamma + T_e = T_\gamma (1 + \alpha)$$

Partial conversion coefficients are defined for each atomic shell:

$$\alpha = \alpha_K + \alpha_L + \alpha_M + \cdots$$

The $K$-shell coefficient dominates because the $K$-electron wavefunction has the largest overlap with the nuclear volume. For a transition of multipolarity $(\sigma\lambda)$:

$$\alpha(\sigma\lambda) \propto Z^3 \lambda \left(\frac{E_\gamma}{m_e c^2}\right)^{-(\lambda+5/2)} \quad \text{(approximate)}$$

Key features of this dependence: - $Z^3$ dependence: Internal conversion is much more important in heavy nuclei. In ${}^{238}$U, low-energy transitions can have $\alpha_K > 100$, meaning almost all de-excitations go through conversion. - $E_\gamma^{-(\lambda+5/2)}$: Conversion is favored for low-energy transitions. As $E_\gamma \to 0$, $\alpha \to \infty$. - $\lambda$ dependence: Higher multipoles favor conversion more strongly. - $M\lambda$ transitions have larger $\alpha$ values than $E\lambda$ transitions of the same $\lambda$ because the magnetic interaction with the electron has an extra derivative.

9.5.3 Diagnostic Power of Conversion Coefficients

Measuring the conversion coefficient — or, more powerfully, the $K/L$ conversion ratio — determines the multipolarity of the transition independent of any model. This is because $\alpha_K/\alpha_L$ depends strongly on $\lambda$ and on whether the transition is $E$ or $M$ type.

Tabulated conversion coefficients (the BrIcc database, maintained at ANU Canberra) are calculated using relativistic Dirac-Fock atomic wavefunctions and are accurate to better than 1% for most cases. Comparing measured $\alpha_K$ values to these tables is a standard tool for multipolarity assignment.

Practical note: The $0^+ \to 0^+$ transition, forbidden for single-photon emission, proceeds entirely by internal conversion (or, for $E_\gamma > 2m_e c^2 = 1.022$ MeV, by internal pair creation). The classic example is the $0^+_2 \to 0^+_1$ transition in ${}^{16}$O at 6.05 MeV ($E0$).

9.5.4 Worked Example: Internal Conversion in ${}^{137}$Ba

The 661.66 keV transition in ${}^{137}$Ba (the daughter of ${}^{137}$Cs, one of the most important calibration sources in gamma-ray spectroscopy) provides an instructive example. The transition connects the $11/2^-$ isomeric state to the $3/2^+$ ground state.

From the selection rules: $\Delta I = 4$, parity changes ($- \to +$). The lowest allowed electric multipole is $E4$ ($\lambda = 4$, parity $(-1)^4 = +1$, yes). The lowest magnetic multipole is $M4$ ($\lambda = 4$, parity $(-1)^5 = -1$, no parity change — wrong). So we need $E4$ or $M5$... actually, let us be careful: $(-1)^{\lambda+1} = (-1)^5 = -1$ for $M4$, which means parity does change. So $M4$ requires $\pi_i \neq \pi_f$, which is satisfied. Both $E4$ and $M4$ are allowed, with $M4$ being lower order for the magnetic type. In practice, the dominant multipolarity is $M4$ (the lowest multipole overall when both $E$ and $M$ are checked is $M4$, since $|I_i - I_f| = 4$ and $M4$ satisfies the parity rule).

The measured half-life is 2.552 minutes — extraordinarily long for a 662 keV transition. The Weisskopf estimate for $M4$ at 662 keV and $A = 137$ gives $t_{1/2}^W \approx 26$ s, so the transition is hindered by about a factor of 6 relative to the single-particle estimate.

The internal conversion coefficient is $\alpha_K = 0.0916$, $\alpha_{\text{total}} = 0.110$. This means: - Fraction as $\gamma$ rays: $1/(1+\alpha) = 1/1.110 = 90.1\%$ - Fraction as $K$-shell conversion electrons: $\alpha_K/(1+\alpha) = 0.0916/1.110 = 8.3\%$ - Fraction as higher-shell conversion electrons: $(\alpha - \alpha_K)/(1+\alpha) = 1.7\%$

This transition is predominantly a gamma ray because $\alpha < 1$ — the conversion coefficient is modest at this relatively high energy. Compare this to the 13.3 keV $M4$ isomeric transition in ${}^{73}$Ge, where $\alpha_K \approx 2920$ and virtually all decays produce conversion electrons rather than gamma rays.

9.5.5 $E0$ Transitions: A Special Case

Electric monopole ($E0$) transitions deserve special attention. The $E0$ operator is:

$$\hat{\mathcal{O}}(E0) = e \sum_i r_i^2$$

which measures the mean-square radius of the charge distribution. An $E0$ transition between two $0^+$ states requires a change in the nuclear charge radius — a direct probe of shape coexistence. Nuclei with two $0^+$ states at similar energies (e.g., ${}^{186}$Pb, with spherical and oblate $0^+$ states within 650 keV) show strong $E0$ transitions between them, providing unambiguous evidence for coexisting nuclear shapes.

The $E0$ strength is parameterized by $\rho^2(E0)$, defined as:

$$\rho^2(E0) = \left|\frac{\langle 0^+_f | \sum_i e_i r_i^2 | 0^+_i \rangle}{e R^2}\right|^2$$

Values of $\rho^2(E0) \sim 10^{-3}$ to $10^{-2}$ are typical for $0^+$ states with similar structure, while $\rho^2(E0) \sim 0.1$ indicates states with very different charge distributions — strong evidence for shape coexistence. The mercury-lead region ($Z = 78$-$84$) is particularly rich in shape coexistence, with $\rho^2(E0)$ values as large as 0.1 observed in ${}^{184}$Hg and ${}^{186}$Pb.

9.6 Angular Correlations

9.6.1 The Method

When a nucleus de-excites through a cascade of two (or more) successive gamma rays, the angular distribution of the second gamma ray relative to the first is not isotropic. This angular correlation encodes information about the spins of the nuclear levels and the multipolarities of the transitions.

Consider a cascade $I_i \xrightarrow{\gamma_1} I \xrightarrow{\gamma_2} I_f$. If both gamma rays are observed, the angular correlation function is:

$$W(\theta) = \sum_{k} A_{kk} P_k(\cos\theta)$$

where $\theta$ is the angle between the two gamma-ray directions, $P_k$ are Legendre polynomials, and the sum runs over even values of $k$ from 0 to $\min(2I, 2\lambda_1, 2\lambda_2)$. The coefficients $A_{kk}$ depend on the spins $I_i$, $I$, $I_f$, and the multipolarities $\lambda_1$, $\lambda_2$ through products of Racah coefficients.

9.6.2 The $4^+ \to 2^+ \to 0^+$ Cascade

The classic example is the $4^+ \to 2^+ \to 0^+$ cascade of $E2$ transitions in an even-even nucleus (e.g., in the ground-state rotational band of a deformed nucleus). Both transitions are pure $E2$, so $\lambda_1 = \lambda_2 = 2$. The angular correlation is:

$$W(\theta) = 1 + a_{22} P_2(\cos\theta) + a_{44} P_4(\cos\theta)$$

with theoretical coefficients $a_{22} = 0.1020$ and $a_{44} = 0.0091$ for the $4$-$2$-$0$ cascade. The small but nonzero $a_{44}$ coefficient provides a distinctive signature.

Deviations from the theoretical values indicate either: - Mixed multipolarities (e.g., $E2/M1$ mixing in one of the transitions) - Perturbation of the intermediate state's alignment by extranuclear fields before the second gamma ray is emitted (e.g., in condensed matter, hyperfine interactions can cause decoherence — this is the basis of perturbed angular correlations, or PAC, used in materials science)

9.6.3 Physical Origin of Angular Correlations

The anisotropy of the angular correlation has a simple physical origin. The first gamma ray is detected in a specific direction, which selects a subset of the magnetic substates $m$ of the intermediate state. This selection creates an aligned intermediate state — a non-uniform population of $m$ substates. The second gamma ray, emitted from this aligned state, has an angular distribution that reflects the alignment.

If the intermediate state lives long enough for the alignment to be disturbed — for example, by the hyperfine interaction between the nuclear magnetic moment and the local magnetic field in a solid — the angular correlation is attenuated. This is the basis of perturbed angular correlation (PAC) spectroscopy, a technique used in condensed-matter physics and materials science. By measuring the attenuation of the angular correlation as a function of time, one can extract the strength of the hyperfine field at the nuclear site, which provides information about the local electronic environment of the probe nucleus.

PAC has been used to study defects in semiconductors, phase transitions in metals, and the electronic structure of biological molecules. The nuclear physics formalism of this chapter thus has applications far beyond nuclear structure.

9.6.4 Directional Correlations from Oriented States (DCO)

In modern nuclear spectroscopy, angular correlation measurements are extended to DCO ratios, where gamma-ray coincidences are measured between detectors at different angles relative to the beam axis. The DCO ratio:

$$R_{\text{DCO}} = \frac{I_\gamma(\theta_1, \text{gated on } \gamma_2 \text{ at } \theta_2)}{I_\gamma(\theta_2, \text{gated on } \gamma_2 \text{ at } \theta_1)}$$

is sensitive to the multipole character of $\gamma_1$. For an array like Gammasphere, with detectors at specific angles, $R_{\text{DCO}} \approx 1$ for stretched $E2$ transitions and $R_{\text{DCO}} \approx 0.5$-$0.6$ for dipole transitions. This provides a fast, reliable method for assigning multipolarities to newly discovered transitions.

9.7 Lifetime Measurements

9.7.1 Why Lifetimes Matter

The connection between lifetime and transition rate is direct:

$$\tau = \frac{1}{\sum_f T(I_i \to I_f)}$$

where the sum runs over all final states $f$ that the initial state can decay to. For a state with a single dominant decay branch, $\tau = 1/T$, and measuring the lifetime gives the $B(\sigma\lambda)$ value directly (through the formulas of Section 9.4). Nuclear lifetimes span an enormous range — from $10^{-15}$ s (fast collective $E2$ transitions) to years (highly hindered low-energy transitions in nuclear isomers). Different experimental techniques cover different ranges.

9.7.2 Doppler Shift Attenuation Method (DSAM)

For lifetimes in the range $10^{-15}$ to $10^{-12}$ s (femtoseconds to picoseconds), the Doppler Shift Attenuation Method exploits the fact that the recoiling nucleus slows down in the target/backing material on a timescale comparable to its lifetime. If the nucleus emits a gamma ray while still moving with velocity $v$, the observed energy is Doppler-shifted:

$$E_{\text{obs}} = E_0 \left(1 + \frac{v}{c}\cos\theta\right)$$

The fraction of the full Doppler shift observed — the attenuation factor $F(\tau)$ — depends on the ratio of the nuclear lifetime to the stopping time in the medium:

$$F(\tau) = \frac{1}{v_0} \int_0^\infty v(t) e^{-t/\tau} \frac{dt}{\tau}$$

where $v(t)$ is the velocity history of the recoiling ion. Measuring $F(\tau)$ at multiple detector angles, combined with calculated or measured stopping powers, determines $\tau$.

Practical example: In a ${}^{152}$Sm Coulomb excitation experiment, the $2^+_1$ state at 122 keV has $\tau = 1.01$ ns — too long for DSAM but measurable by other methods. The $4^+_1$ state at 367 keV has $\tau = 31$ ps — ideal for DSAM.

9.7.3 Recoil Distance Method (RDM)

For lifetimes in the range $10^{-12}$ to $10^{-9}$ s (picoseconds to nanoseconds), the Recoil Distance Method (also called the plunger method) uses a thin target and a movable stopper foil separated by a distance $d$. Nuclei produced in the target recoil with velocity $v$ toward the stopper. If the nucleus decays in flight (between target and stopper), the gamma ray is fully Doppler-shifted; if it decays after implanting in the stopper, the gamma ray is unshifted.

The ratio of shifted to unshifted intensities as a function of the target-stopper distance $d$ gives the decay curve directly:

$$R(d) = \frac{I_{\text{shifted}}}{I_{\text{total}}} = e^{-d/(v\tau)}$$

This technique is beautifully direct — it literally watches the exponential decay in space rather than time. Typical plunger distances range from a few micrometers to several centimeters, controlled by a precision piezoelectric actuator.

9.7.4 Electronic Timing

For lifetimes longer than $\sim 1$ ns, electronic timing methods using fast scintillation detectors (BaF$_2$, LaBr$_3$:Ce) can measure the time interval between two gamma rays in a cascade. The time spectrum is an exponential convoluted with the detector response function. Modern LaBr$_3$ detectors achieve timing resolution of $\sim 100$ ps (FWHM), enabling lifetime measurements down to $\sim 10$ ps using the centroid-shift method and down to $\sim 50$ ps with direct fitting.

The Generalized Centroid Difference Method (GCD) has become the standard approach: by comparing the centroid of the time-difference distribution for coincidences with $\gamma_1$ in detector A and $\gamma_2$ in detector B (and vice versa), systematic effects cancel, and the lifetime of the intermediate state can be extracted with precision of $\sim 1$ ps.

9.7.5 Nuclear Isomers: When Lifetimes Become Long

A nuclear isomer is an excited nuclear state with a measurably long lifetime — typically $> 1$ ns, though the definition is somewhat arbitrary. Long lifetimes arise when the transition to lower-lying states requires a high multipole order (large $\Delta I$), proceeds through a highly converted low-energy transition, or both. Isomers are classified by their hindrance mechanism:

• Spin isomers (spin traps): Large angular momentum difference between the isomeric state and the nearest lower state. Example: the $I^\pi = 8^+$ isomer in ${}^{178}$Hf at 2446 keV, which has a half-life of 4 seconds despite sitting 2.4 MeV above the ground state. The transition to the $6^+$ state below it requires only an $E2$ transition, but the total cascade down to the ground state involves many steps.

• K-isomers: In deformed nuclei, the quantum number $K$ (projection of $I$ on the symmetry axis) is approximately conserved. Transitions that violate $K$ conservation by $\Delta K > \lambda$ (where $\lambda$ is the multipole order) are hindered by factors of $10^{-2}$ per unit of "forbidden" $\Delta K$. The famous $I^\pi = 16^+$, $K = 16$ isomer in ${}^{178}$Hf (the "Hf isomer") at 2446 keV has $t_{1/2} = 31$ years because its decay requires breaking $K$ conservation by a large amount.

• Shape isomers: In some nuclei, a second minimum in the potential energy surface (a superdeformed or fission isomer) traps the nucleus in a metastable deformed shape. Fission isomers in the actinides ($t_{1/2} \sim$ ns to $\mu$s) are the classic examples.

Isomers have practical importance. ${}^{99m}$Tc (the "m" denotes metastable/isomeric) with $t_{1/2} = 6.0$ hours is the most widely used radioisotope in medical imaging. The isomeric transition energy of 140.5 keV is ideal for SPECT imaging, and the 6-hour half-life is long enough for preparation and imaging but short enough to minimize patient dose.

9.7.6 The Lifetime Landscape

9.8 Coulomb Excitation

9.8.1 The Principle

Coulomb excitation is the excitation of nuclear states by the time-varying electromagnetic field experienced by a nucleus as a charged projectile passes nearby. The key requirement is that the closest approach distance must be well outside the range of the nuclear force:

$$d_{\min} > R_1 + R_2 + \delta \quad (\delta \approx 5 \text{ fm safety margin})$$

This ensures that the excitation is purely electromagnetic — the nuclear force plays no role. Since we understand electromagnetism exactly, Coulomb excitation provides a model-independent measurement of electromagnetic matrix elements.

9.8.2 The Semiclassical Theory

In the standard semiclassical treatment (Alder and Winther, 1966), the projectile follows a classical Rutherford trajectory while the nuclear excitation is treated quantum mechanically. The excitation probability for a transition from state $|I_i\rangle$ to $|I_f\rangle$ via multipole $E\lambda$ is:

$$P(E\lambda) = \left(\frac{a_\lambda}{a}\right)^{2\lambda} f_{E\lambda}(\xi, \theta)$$

where $a$ is half the distance of closest approach, $a_\lambda$ contains the nuclear matrix element, and $f_{E\lambda}(\xi,\theta)$ is a known function of the adiabaticity parameter:

$$\xi = \frac{a}{\hbar v} E_\gamma$$

The adiabaticity parameter measures whether the nuclear excitation energy $E_\gamma$ is "fast" ($\xi \ll 1$, sudden limit) or "slow" ($\xi \gg 1$, adiabatic limit) compared to the collision time. Coulomb excitation is most efficient when $\xi \sim 1$.

9.8.3 "Safe" Coulomb Excitation

In the safe energy regime, the beam energy is chosen so that the Rutherford orbit remains safely outside the nuclear radius at all scattering angles. This condition, combined with the Rutherford cross section (which is exactly known), means that the measured excitation cross section is directly proportional to $B(E2)$:

$$\frac{d\sigma}{d\Omega}(E2) = \left(\frac{d\sigma}{d\Omega}\right)_{\text{Ruth}} \cdot P(E2)$$

This is why Coulomb excitation is the gold standard for measuring $B(E2; 0^+ \to 2^+_1)$ values in even-even nuclei. The theoretical framework is exact; the only unknown is the nuclear matrix element.

9.8.4 Worked Example: Coulomb Excitation of ${}^{208}$Pb

Let us estimate the conditions for safe Coulomb excitation of ${}^{208}$Pb. Using a ${}^{58}$Ni beam:

Distance of closest approach for a head-on collision at center-of-mass energy $E_{\text{cm}}$:

$$d_{\min} = \frac{Z_1 Z_2 e^2}{2 E_{\text{cm}}} = \frac{28 \times 82 \times 1.44 \text{ MeV fm}}{2 E_{\text{cm}}}$$

The nuclear radii (with safety margin): $R_1 + R_2 + 5 = 1.2(58^{1/3} + 208^{1/3}) + 5 = 1.2(3.87 + 5.93) + 5 = 11.8 + 5 = 16.8$ fm.

For safe Coulomb excitation, we need $d_{\min} > 16.8$ fm, which gives:

$$E_{\text{cm}} < \frac{28 \times 82 \times 1.44}{2 \times 16.8} = \frac{3305}{33.6} = 98 \text{ MeV}$$

At $E_{\text{cm}} = 90$ MeV (safely below the Coulomb barrier), the adiabaticity parameter for the $2^+_1$ state ($E_\gamma = 4.086$ MeV) is:

$$\xi = \frac{a}{v} \cdot \frac{E_\gamma}{\hbar} = \frac{d_{\min}}{2v} \cdot \frac{E_\gamma}{\hbar}$$

where $v = \sqrt{2E_{\text{cm}}/\mu}$ is the relative velocity. With $\mu \approx 45.4$ u $= 42,300$ MeV/$c^2$:

$v/c = \sqrt{2 \times 90/42300} = 0.065$

$\xi = \frac{18.4 \text{ fm}}{2 \times 0.065 \times 197.3 \text{ fm}} \times 4.086 = \frac{18.4}{25.6} \times 4.086 = 2.94$

This is in the adiabatic regime ($\xi > 1$), where the excitation probability is exponentially suppressed: $P \propto e^{-\pi\xi}$. The high excitation energy of the first $2^+$ in ${}^{208}$Pb (4.086 MeV — typical of a doubly magic nucleus with a large shell gap) makes Coulomb excitation inefficient. Deformed nuclei with their low-lying $2^+$ states ($E_{2^+} \sim 50$-$100$ keV) are far easier to Coulomb-excite.

9.8.5 Intermediate-Energy and Relativistic Coulomb Excitation

At radioactive beam facilities like FRIB and GSI/FAIR, beams of exotic short-lived nuclei are produced at intermediate energies ($\sim 100$ MeV/nucleon) or relativistic energies ($\sim 500$ MeV/nucleon). At these energies, Coulomb excitation on high-$Z$ targets (typically ${}^{197}$Au or ${}^{208}$Pb) provides access to $B(E2)$ values for nuclei far from stability that can only be produced in-flight.

The relativistic treatment replaces the Rutherford orbit with a straight-line trajectory, and the virtual photon spectrum is Lorentz-boosted. The technique has been used to measure $B(E2; 0^+ \to 2^+)$ values for nuclei as exotic as ${}^{32}$Mg (in the "island of inversion") and ${}^{78}$Ni (near doubly magic), providing critical tests of shell evolution far from stability.

Milestone measurement: The $B(E2)$ value for ${}^{32}$Mg measured by Motobayashi et al. (1995) at RIKEN was $454 \pm 78$ $e^2$fm$^4$, corresponding to $18 \pm 3$ W.u. This large value proved that $N = 20$ is not magic in this region — the "island of inversion" where neutron-rich nuclei develop unexpected deformation.

9.8.6 Multi-Step Coulomb Excitation (COULEX)

At energies near the Coulomb barrier, the electromagnetic interaction can be strong enough to excite a nucleus through multiple successive steps: $0^+ \to 2^+ \to 4^+$, or $0^+ \to 2^+ \to 0^+_2$. Multi-step Coulomb excitation (analyzed with codes like GOSIA) allows measurement of a complete set of matrix elements connecting many states — both diagonal (static moments) and off-diagonal (transition matrix elements). The Rochester/Warsaw group has used this technique to map the shapes of nuclei with remarkable precision, revealing phenomena like shape coexistence in the mercury-lead region.

9.9 Modern Gamma-Ray Detector Arrays

9.9.1 The Evolution of Gamma-Ray Detection

The history of nuclear structure physics is inseparable from the history of gamma-ray detectors. Each generation of detector technology opened new windows:

1. NaI(Tl) scintillators (1950s): Energy resolution $\sim 7\%$ at 1 MeV. Sufficient for simple level schemes.
2. Ge(Li) and HPGe detectors (1970s-1980s): Resolution $\sim 0.1\%$ (2 keV at 1 MeV). Revealed the rich complexity of nuclear level schemes.
3. Compton-suppressed arrays (1980s): Arrays of HPGe detectors surrounded by BGO anti-Compton shields. Reduced the Compton continuum, dramatically improving peak-to-total ratio.
4. $4\pi$ arrays (1990s): Gammasphere (US, 110 Compton-suppressed Ge detectors) and Euroball (Europe, 239 detectors of various types). Achieved unprecedented sensitivity through high granularity, high efficiency, and fold selection.
5. Tracking arrays (2010s-present): GRETINA/GRETA (US) and AGATA (Europe). The current state of the art.

9.9.2 Gamma-Ray Tracking: How It Works

The breakthrough concept behind tracking arrays is the replacement of passive Compton suppression (which wastes photons) with active tracking of every gamma-ray interaction within the detector volume.

A gamma ray entering a germanium detector undergoes a series of Compton scatterings before a final photoelectric absorption. In a conventional detector, these interactions are summed to give the full energy. In a tracking detector, the individual interaction positions and energies are measured by using a highly segmented HPGe crystal (36-fold segmentation in GRETINA, 36 in AGATA) combined with digital signal processing.

The key algorithm is gamma-ray tracking: given a set of interaction points (positions and energies), reconstruct which interactions belong to which gamma ray and determine the full energy and first interaction point of each. The Compton scattering formula:

$$\cos\theta = 1 - m_e c^2 \left(\frac{1}{E'} - \frac{1}{E}\right)$$

relates the scattering angle $\theta$ to the energies before ($E$) and after ($E'$) each scatter, providing a consistency check that is used to cluster interactions into tracks.

9.9.3 GRETINA and GRETA

GRETINA (Gamma-Ray Energy Tracking In-beam Nuclear Array) consists of 28 highly segmented HPGe crystals covering approximately $1\pi$ of solid angle. It has been operating at NSCL/FRIB (Michigan State University) and ATLAS (Argonne National Laboratory) since 2012. Its successor, GRETA (Gamma-Ray Energy Tracking Array), will complete the $4\pi$ coverage with 120 crystals.

GRETINA's capabilities include: - Position resolution: $\sim 2$ mm for each interaction point - Energy resolution: $\sim 2.2$ keV FWHM at 1.33 MeV - Doppler correction: Because the first interaction point is known to $\sim 2$ mm, the Doppler correction for in-flight gamma rays from fast beams is dramatically improved. This is critical at fragmentation facilities where $v/c \sim 0.3$-$0.4$. - Effective photopeak efficiency: $\sim 8\%$ at 1 MeV (for $1\pi$ coverage)

9.9.4 AGATA

The European counterpart is AGATA (Advanced Gamma Tracking Array), which uses 180 HPGe crystals in a $4\pi$ geometry. Like GRETINA, each crystal is 36-fold segmented. The full AGATA will have: - Photopeak efficiency: $\sim 28\%$ at 1 MeV (full $4\pi$) - Position resolution: $\sim 5$ mm (pulse-shape analysis) - Angular resolution: $\sim 1°$

AGATA has been operating in a series of campaigns with increasing numbers of detectors at facilities including GANIL (France), GSI (Germany), and LNL (Italy).

9.9.5 The Doppler Correction Challenge: A Quantitative Example

To appreciate why tracking is transformative, consider a concrete example. A ${}^{32}$Mg nucleus produced by fragmentation at FRIB travels at $v/c = 0.35$ and emits a 885 keV gamma ray ($2^+ \to 0^+$ transition). The observed energy in the lab frame depends on the emission angle:

$$E_{\text{lab}} = E_0 \frac{\sqrt{1-\beta^2}}{1-\beta\cos\theta}$$

At $\theta = 30°$: $E_{\text{lab}} = 885 \times 0.937 / (1 - 0.303) = 885 \times 0.937 / 0.697 = 1190$ keV. At $\theta = 90°$: $E_{\text{lab}} = 885 \times 0.937 / 1.000 = 829$ keV. The full range spans from 1382 keV (forward) to 606 keV (backward).

Now consider the energy resolution. Differentiating the Doppler formula:

$$\frac{\Delta E}{E} \approx \beta \sin\theta \cdot \Delta\theta$$

At $\theta = 30°$ with a conventional detector ($\Delta\theta = 8° = 0.14$ rad):

$\Delta E = 1190 \times 0.35 \times 0.5 \times 0.14 = 29$ keV

With gamma-ray tracking ($\Delta\theta = 1° = 0.017$ rad):

$\Delta E = 1190 \times 0.35 \times 0.5 \times 0.017 = 3.5$ keV

The intrinsic germanium resolution at 1.2 MeV is about 2.5 keV. With tracking, the Doppler broadening is comparable to the intrinsic resolution — the gamma-ray line is essentially resolved. Without tracking, the 29 keV broadening would make it impossible to distinguish closely-spaced transitions or to measure precise energies.

This factor-of-eight improvement in effective resolution is what makes spectroscopy with fast beams practical. Before tracking, such experiments could only identify the strongest transitions; with tracking, detailed level schemes can be constructed for exotic nuclei produced at rates of just a few particles per second.

9.9.6 What Tracking Arrays Discover

The capabilities of tracking arrays have enabled discoveries impossible with earlier detectors:

1. Superdeformed and hyperdeformed bands: The improved sensitivity allows identification of very weak, highly deformed rotational bands at high angular momentum.

2. Gamma-ray spectroscopy with radioactive beams: At fragmentation facilities like FRIB, nuclei are produced at $v/c \sim 0.3$-$0.5$. The resulting Doppler broadening of gamma-ray lines would be debilitating with conventional detectors but is corrected to $\sim 10$ keV resolution with tracking.

3. Lifetime measurements: The position sensitivity enables DSAM-type lifetime measurements with unprecedented precision.

4. Coulomb excitation of exotic nuclei: Combining tracking arrays with Coulomb excitation of radioactive beams at intermediate energies provides $B(E2)$ values for the most exotic nuclei produced at modern facilities.

5. Pair structure via two-photon correlations: Tracking arrays can reconstruct the full kinematics of $e^+e^-$ pairs from internal pair conversion, enabling $E0$ spectroscopy.

Looking ahead: GRETA at FRIB will combine the world's most powerful radioactive beam facility with the world's most capable gamma-ray detector. This combination will open the neutron-rich frontier to precision spectroscopy, testing nuclear structure models in regions of the chart of nuclides that have never been accessible before.

9.10 Synthesis: The Electromagnetic Fingerprint of Nuclear Structure

Let us step back and survey what electromagnetic transitions reveal about the nucleus. The framework of this chapter provides a complete chain from theory to experiment:

$$\text{Nuclear model} \xrightarrow{\text{wavefunctions}} \langle f \| \hat{\mathcal{O}}(\sigma\lambda) \| i \rangle \xrightarrow{\text{Wigner-Eckart}} B(\sigma\lambda) \xrightarrow{E_\gamma^{2\lambda+1}} \tau \xrightarrow{\text{experiment}} \text{measured lifetime or cross section}$$

Running this chain in reverse — from measured quantities back to nuclear structure — is the daily work of nuclear spectroscopists. The key messages:

1. Selection rules determine which transitions are allowed and identify the multipolarities.
2. Weisskopf estimates provide a benchmark: $B$ values in W.u. immediately reveal whether a transition is single-particle or collective.
3. $B(E2; 0^+ \to 2^+_1)$ values across the chart of nuclides map nuclear collectivity: small near magic numbers, large in deformed regions. This single observable encodes the competition between the shell model and collective models.
4. $E2/M1$ mixing ratios probe the interplay of collective and single-particle degrees of freedom.
5. Internal conversion coefficients provide model-independent multipolarity assignments.
6. Coulomb excitation yields $B$ values without any nuclear-model assumptions — pure electromagnetism.
7. Modern tracking arrays push these measurements to the most exotic nuclei, testing structure models at the frontiers.

Gamma-ray spectroscopy is the microscope through which we view nuclear structure. The formalism of this chapter is the instruction manual for that microscope. In Chapter 15, we will return to gamma decay as a decay mode, treating it within the systematic framework of radioactive decay. Here, the emphasis has been on the structural information that electromagnetic transitions encode.

The experimental physicist's creed: A nucleus is known when its level scheme is known — spins, parities, energies, lifetimes, branching ratios, mixing ratios, and the $B$ values derived from them. Everything else is interpretation.

Chapter Summary

• Multipole operators — electric ($E\lambda$) and magnetic ($M\lambda$) — connect nuclear wavefunctions to observable transition rates and moments.
• Selection rules constrain the allowed multipole orders: $|I_i - I_f| \leq \lambda \leq I_i + I_f$; parity change $(-1)^\lambda$ for $E\lambda$, $(-1)^{\lambda+1}$ for $M\lambda$.
• Weisskopf estimates provide single-particle benchmarks. $B$ values expressed in W.u. immediately diagnose single-particle ($\sim 1$ W.u.) versus collective ($\gg 1$ W.u.) character.
• Weisskopf transition rates: $T_W(E1) \approx 10^{14} A^{2/3} E_\gamma^3$ s$^{-1}$; $T_W(E2) \approx 7 \times 10^7 A^{4/3} E_\gamma^5$ s$^{-1}$ ($E_\gamma$ in MeV).
• Internal conversion competes with gamma emission, especially for low-energy, high-multipolarity transitions in heavy nuclei. $E0$ transitions proceed exclusively by conversion.
• Angular correlations between successive gamma rays encode spin and multipolarity information.
• Lifetime methods (DSAM, RDM, electronic timing) cover the range from $10^{-15}$ s to seconds.
• Coulomb excitation provides model-independent $B(E\lambda)$ values using the electromagnetic interaction alone.
• Tracking arrays (GRETINA/GRETA, AGATA) represent the current frontier, enabling precision spectroscopy with exotic beams.