Case Study 1 — Fermi's Golden Rule: The Formula That Governs Nuclear Decay and Reactions

One Formula, Four Phenomena

Enrico Fermi derived his "golden rule" in the context of atomic transitions in the 1930s. It has since become perhaps the single most applied result in all of quantum physics. In nuclear physics, virtually every measurable rate — alpha decay, beta decay, gamma emission, nuclear reactions — is computed from the same master formula:

$$\Gamma = \frac{2\pi}{\hbar}|\langle f|\hat{V}|i\rangle|^2 \rho(E_f)$$

This case study traces how this one formula connects to four major categories of nuclear processes, each of which receives its own detailed treatment later in the book. The purpose here is to see the unifying structure: every nuclear transition is the product of a matrix element (the dynamics) and a density of states (the kinematics).

The Physical Picture

Before examining each application, it is worth reflecting on what Fermi's golden rule tells us at the most basic level. A quantum system in state $|i\rangle$ can transition to state $|f\rangle$ if three conditions are met: (1) an interaction $\hat{V}$ exists that connects the two states (the matrix element is nonzero); (2) the transition conserves energy (the delta function, which becomes a density of states when summed over a continuum); and (3) the perturbation is weak enough that first-order theory applies. The rate of the transition is set by the product of the dynamical coupling strength and the kinematic availability of final states. This is a principle of extraordinary generality.

Application 1: Gamma Decay Rates

When a nucleus in an excited state $|J_i^\pi\rangle$ emits a photon and transitions to a lower state $|J_f^\pi\rangle$, the perturbation $\hat{V}$ is the electromagnetic interaction between the nuclear charge/current distribution and the radiation field. For a transition of electric multipolarity $\lambda$ (an E$\lambda$ transition), the matrix element takes the form:

$$V_{fi} = \langle J_f || \hat{O}(\text{E}\lambda) || J_i \rangle$$

where $\hat{O}(\text{E}\lambda)$ is the electric multipole operator and the double bars indicate a reduced matrix element (independent of $M_J$ projections — the angular dependence has been factored out using the Wigner-Eckart theorem).

The density of states for the emitted photon is $\rho_\gamma = E_\gamma^2 V/(\pi^2\hbar^3 c^3)$. After summing and averaging over magnetic substates, the result (derived in Chapter 9) is:

$$\Gamma(\text{E}\lambda) = \frac{8\pi(\lambda+1)}{\lambda[(2\lambda+1)!!]^2}\left(\frac{E_\gamma}{\hbar c}\right)^{2\lambda+1} B(\text{E}\lambda; J_i \to J_f)$$

where $B(\text{E}\lambda)$ is the reduced transition probability — the square of the reduced matrix element. Every factor in this expression has a clear origin:

• The $(E_\gamma/\hbar c)^{2\lambda+1}$ factor comes from the photon density of states ($E_\gamma^2$) and the matrix element ($\propto E_\gamma^{\lambda-1}$ from the multipole expansion)
• The $[(2\lambda+1)!!]^2$ in the denominator reflects the geometric suppression of higher multipoles
• $B(\text{E}\lambda)$ encodes all the nuclear structure information

The physics: Nuclei with large $B(\text{E}2)$ values — collective excitations involving many nucleons moving coherently — have fast E2 transitions. Single-particle transitions are slower. The ratio $B(\text{E}2)_\text{measured} / B(\text{E}2)_\text{single-particle}$ (the Weisskopf ratio) is the standard measure of collectivity.

Real example: The first $2^+ \to 0^+$ transition in $^{152}$Sm has $B(\text{E}2) = 5610$ e$^2$fm$^4$, about 200 times the single-particle estimate — a strongly collective rotational transition. By contrast, the $2^+ \to 0^+$ transition in $^{18}$O has $B(\text{E}2) = 26$ e$^2$fm$^4$, close to the single-particle value — a non-collective, shell-model transition. Fermi's golden rule, through $B(\text{E}2)$, distinguishes these two very different nuclear structures.

Application 2: Alpha Decay Rates

Alpha decay is fundamentally a tunneling problem, but Fermi's golden rule still provides the framework. The transition rate is:

$$\Gamma_\alpha = \frac{2\pi}{\hbar} |V_{fi}|^2 \rho_\alpha(E_\alpha)$$

Here the situation is subtler: the "perturbation" is the nuclear potential that binds the alpha particle inside the nucleus, and the final state is the alpha particle in a continuum Coulomb state outside the barrier. The matrix element $|V_{fi}|^2$ contains the preformation factor (the probability that four nucleons inside the nucleus cluster into an alpha particle) and the penetrability (the WKB tunneling probability through the Coulomb barrier):

$$\Gamma_\alpha \approx \nu \cdot P_\alpha \cdot T_l$$

where $\nu$ is the assault frequency (how often the alpha particle hits the barrier from inside), $P_\alpha$ is the preformation probability, and $T_l$ is the barrier penetrability for angular momentum $l$.

The physics: The tunneling factor $T_l \sim e^{-2\gamma}$ dominates the energy dependence. A factor of 2 change in $E_\alpha$ changes the Gamow factor by $\Delta(2\gamma) \approx 40$--$60$, producing a factor of $e^{40}$--$e^{60} \approx 10^{17}$--$10^{26}$ change in the decay rate. This is why alpha decay half-lives span such an extraordinary range.

Real example: Compare $^{212}$Po ($E_\alpha = 8.78$ MeV, $t_{1/2} = 0.30\ \mu$s) with $^{232}$Th ($E_\alpha = 4.01$ MeV, $t_{1/2} = 1.4 \times 10^{10}$ years). The ratio of half-lives is $\sim 10^{24}$. This factor of $10^{24}$ comes almost entirely from the exponential tunneling factor, and it all traces back to Fermi's golden rule applied to a tunneling problem.

Application 3: Beta Decay Rates

Beta decay ($n \to p + e^- + \bar{\nu}_e$ or $p \to n + e^+ + \nu_e$) involves the weak interaction. The perturbation is the weak Hamiltonian, and the final state includes both the outgoing lepton (electron or positron) and the neutrino. Fermi's golden rule gives:

$$\Gamma_\beta = \frac{2\pi}{\hbar} |V_{fi}|^2 \rho_\beta(E_0)$$

The matrix element decomposes into two types: - Fermi transitions ($\hat{V} \propto \hat{1}$ in spin-isospin space): $\Delta J = 0$, no parity change - Gamow-Teller transitions ($\hat{V} \propto \hat{\boldsymbol{\sigma}}\hat{\boldsymbol{\tau}}$): $\Delta J = 0, \pm 1$, no parity change

The density of states involves both the electron and the neutrino. Since the total energy $E_0$ (the $Q$-value) is shared between them, the phase space integral gives the famous Fermi integral $f(Z, E_0)$:

$$\Gamma_\beta = \frac{G_F^2}{2\pi^3\hbar^7 c^6}\left(g_V^2 |M_F|^2 + g_A^2 |M_{GT}|^2\right) f(Z, E_0)$$

where $G_F$ is the Fermi coupling constant, $g_V$ and $g_A$ are the vector and axial-vector coupling constants, and $|M_F|^2$ and $|M_{GT}|^2$ are the Fermi and Gamow-Teller nuclear matrix elements.

The physics: The matrix elements depend on the nuclear wavefunctions (how similar the initial and final nuclear states are), while the Fermi integral $f(Z, E_0) \propto E_0^5$ (approximately, for allowed transitions) encodes the phase space. The strong $E_0^5$ dependence means that beta decay rates are extremely sensitive to the $Q$-value: doubling the $Q$-value increases the rate by a factor of $\sim 32$.

Real example: The $ft$-value (the product $f \times t_{1/2}$) is a measure of the nuclear matrix element alone, with the phase space factored out. Superallowed $0^+ \to 0^+$ Fermi transitions (pure Fermi, $|M_F|^2 = |N-Z|$ for mirror nuclei) have nearly identical $ft$-values: $ft \approx 3070$ s, providing one of the most precise tests of the Standard Model's electroweak sector. This precision is possible because Fermi's golden rule cleanly separates the dynamics ($|V_{fi}|^2$) from the kinematics ($\rho$).

Application 4: Nuclear Reaction Cross Sections

For a nuclear reaction $a + A \to b + B$, the cross section is related to the transition rate by:

$$\sigma = \frac{\Gamma}{n_A \cdot v_\text{rel}}$$

where $n_A$ is the target density and $v_\text{rel}$ is the relative velocity. Using Fermi's golden rule for $\Gamma$:

$$\sigma_{a+A \to b+B} = \frac{2\pi}{\hbar v_\text{rel}} |T_{fi}|^2 \rho_f(E_f)$$

where $T_{fi} = \langle f|\hat{T}|i\rangle$ is the $T$-matrix element (which reduces to $V_{fi}$ in the Born approximation, but is the exact scattering amplitude in general). The density of states $\rho_f$ counts the available states of the outgoing particle $b$.

For the special case of compound nucleus reactions (Chapter 18), the cross section factorizes into a formation cross section (getting into the compound nucleus) and a branching ratio (how the compound nucleus decays):

$$\sigma(a + A \to b + B) = \sigma_\text{CN}(a + A) \times \frac{\Gamma_b}{\Gamma_\text{total}}$$

Each partial width $\Gamma_b$ is again given by Fermi's golden rule, with the level density of the compound nucleus playing the role of $\rho$.

Real example: Thermal neutron capture on $^{113}$Cd has a cross section of 20,600 barns — one of the largest known. This enormous cross section arises because the compound nucleus $^{114}$Cd$^*$ has a level very near the neutron separation energy, creating a resonance. The width of this resonance, and hence the cross section, is computed from Fermi's golden rule with the neutron and gamma partial widths.

A second instructive example: the $^{12}$C($\alpha$, $\gamma$)$^{16}$O reaction at stellar energies ($E_{cm} \approx 300$ keV) determines the carbon-to-oxygen ratio produced in helium-burning stars, which in turn affects all subsequent stellar nucleosynthesis and the composition of the universe. The rate of this reaction depends on a delicate cancellation between E1 and E2 contributions to the cross section — both computed from Fermi's golden rule with different multipole matrix elements. Measuring and calculating this rate to the required precision of $\sim$10% has been one of the great challenges of nuclear astrophysics for fifty years.

The Unifying Structure

All four applications share the same structure:

The matrix element carries the nuclear structure information — it tells us how the nucleus makes the transition. The density of states carries the kinematic information — it tells us how many ways the transition can occur. Every chapter from 9 through 21 is, at its core, an exercise in evaluating Fermi's golden rule for a specific process with the appropriate matrix element and density of states.

This is why Fermi's golden rule deserves to be called the master formula of nuclear physics: not because it is the deepest result (it is first-order perturbation theory, after all), but because it provides the universal framework within which all the detailed nuclear physics lives.

Beyond First Order: When the Golden Rule Fails

Fermi's golden rule is first-order perturbation theory. It can fail when:

1. Strong coupling: In nuclear reactions at energies near a narrow resonance, the coupling between the entrance channel and the compound nucleus state is strong enough that higher-order terms matter. The result is the Breit-Wigner resonance formula (Chapter 17), which generalizes the golden rule.

2. Multiple decay channels: When many final states compete, the total width $\Gamma_\text{total} = \sum_i \Gamma_i$ can become comparable to the level spacing, and the compound nucleus picture replaces the perturbative one. This is the regime of overlapping resonances (Ericson fluctuations).

3. Very fast decays: For particle-unstable states with lifetimes $\sim 10^{-22}$ s, the width $\Gamma$ is comparable to the excitation energy, and the concept of a well-defined initial state breaks down. Here, the $R$-matrix or $K$-matrix formalism is needed.

Despite these limitations, the golden rule provides the correct starting point for virtually every calculation in nuclear physics, and the more sophisticated approaches can be understood as extensions of it.

Discussion Questions

1. Why does the Fermi integral for beta decay scale approximately as $E_0^5$ for allowed transitions? Trace each factor of $E_0$ to either the electron phase space, the neutrino phase space, or the matrix element.

2. The Weisskopf single-particle estimate for E1 transitions gives $\Gamma \propto A^{2/3} E_\gamma^3$. Explain the physical origin of each factor using Fermi's golden rule.

3. In alpha decay, the "matrix element" contains a preformation factor — the probability that nucleons cluster into an alpha particle inside the nucleus. This quantity is notoriously difficult to calculate from first principles. Why? What does this tell us about the limitations of the shell model for decay problems?

4. For compound nucleus reactions at high excitation energy, the statistical model replaces the specific matrix element $|V_{fi}|^2$ with an average over many levels: $\overline{|V_{fi}|^2}$. Under what conditions is this averaging justified? (Hint: think about the density of states and the concept of "Ericson fluctuations.")