Chapter 6 — The Nuclear Shell Model: Magic Numbers and Independent Particles in a Mean Field

"I hesitated a long time before I dared to publish this result, because I felt that physicists would say, 'She is crazy — you can't have spin-orbit coupling that strong.'" — Maria Goeppert Mayer, interview (1963)

The semi-empirical mass formula of Chapter 4 captures the gross features of nuclear binding. Its five terms — volume, surface, Coulomb, asymmetry, and pairing — reproduce experimental binding energies to within about 1% across the chart of nuclides. That is a remarkable achievement. But 1% errors in nuclear physics are not small. A 1% error in the binding energy of a heavy nucleus corresponds to several MeV — enough to completely mispredict whether a nucleus is stable or radioactive, whether it decays by alpha emission or spontaneous fission, whether it lies inside or outside the neutron drip line.

More importantly, the SEMF is smooth. It treats the nucleus as a charged liquid drop, and liquid drops have no internal structure. But nuclei do have internal structure, and the evidence for that structure is dramatic. Certain nuclei — those with specific "magic numbers" of protons or neutrons — are anomalously stable, anomalously spherical, and anomalously resistant to excitation. The SEMF cannot explain this. The liquid drop model cannot explain this. To understand magic numbers, we need a fundamentally different picture of the nucleus: not a featureless droplet, but an ordered quantum system in which nucleons occupy discrete energy levels, much as electrons occupy orbitals in an atom.

This chapter develops that picture — the nuclear shell model. It is, alongside the collective models of Chapter 8, one of the two great pillars of nuclear structure physics. Its discovery by Maria Goeppert Mayer and J. Hans D. Jensen, working independently in 1949, earned them the Nobel Prize in 1963 and transformed our understanding of the atomic nucleus.

6.1 Evidence for Magic Numbers

Before we build any model, we must examine the evidence that demands one. The magic numbers are

$$2, \quad 8, \quad 20, \quad 28, \quad 50, \quad 82, \quad 126.$$

These numbers appear as anomalies in multiple independent observables. No single piece of evidence would be conclusive; it is the convergence of many lines of evidence that makes the case overwhelming.

6.1.1 Binding Energy Anomalies

Recall from Chapter 4 that the SEMF predicts binding energies $B(Z,N)$ as a smooth function of $A$, $Z$, and $N$. If we compute the residuals — the difference between experimental binding energies and SEMF predictions — we find systematic deviations that correlate with the magic numbers.

$$\delta B(Z,N) = B_{\text{exp}}(Z,N) - B_{\text{SEMF}}(Z,N)$$

Nuclei with magic $Z$ or magic $N$ are systematically more bound than the SEMF predicts. Doubly magic nuclei — those with both $Z$ and $N$ magic — show the largest positive residuals. The following doubly magic nuclei are among the most tightly bound for their mass region:

Spaced review (Ch 4): The SEMF residuals we plotted in Chapter 4 already hinted at this pattern. Those were not random scatter — they were the shell model trying to tell us something the liquid drop could not.

6.1.2 Nucleon Separation Energies

A more sensitive probe is the nucleon separation energy — the energy required to remove one neutron or one proton from a nucleus. The one-neutron separation energy is

$$S_n(Z,N) = B(Z,N) - B(Z, N-1)$$

and similarly for one-proton separation energy $S_p$. If we plot $S_n$ as a function of neutron number $N$ for a given element, we observe a striking pattern: $S_n$ decreases roughly smoothly with $N$ (as expected from the asymmetry term), except at the magic numbers, where there is a sharp drop.

Consider the tin isotopes ($Z = 50$). As we add neutrons from $N = 50$ to $N = 82$, $S_n$ decreases gradually from about 11 MeV to about 9 MeV. Then at $N = 83$ — one neutron past the magic number 82 — $S_n$ drops abruptly to about 5.5 MeV. The 82nd neutron is bound by 9 MeV; the 83rd is bound by only 5.5 MeV. This 3.5 MeV discontinuity is the hallmark of a closed shell.

This is precisely analogous to the ionization energy pattern in atomic physics. The first ionization energy of neon (a closed-shell atom) is 21.6 eV; the first ionization energy of sodium (one electron past the closed shell) drops to 5.1 eV. The physics is the same: filling a new shell requires occupying a higher-energy orbital.

6.1.3 First Excited State Energies

Magic nuclei are not only hard to disassemble — they are also hard to excite. The energy of the first excited state, $E(2^+_1)$, is a sensitive probe of shell closure.

For even-even nuclei, the ground state is always $0^+$ (we will understand why shortly). The first excited state is typically $2^+$, and its energy reflects how easy it is to promote a nucleon pair across the shell gap. The pattern is unmistakable:

• $^{208}$Pb: $E(2^+_1) = 4.09$ MeV
• $^{210}$Po ($Z=84$, two protons past the $Z=82$ closure): $E(2^+_1) = 1.18$ MeV
• $^{206}$Hg ($Z=80$, two protons before the closure): $E(2^+_1) = 1.07$ MeV

The doubly magic $^{208}$Pb has a first excited state nearly four times higher than its neighbors. Similar spikes appear at every magic number. The first excited state of $^{48}$Ca ($N = 28$) is at 3.83 MeV; for $^{46}$Ca ($N = 26$) it drops to 1.35 MeV.

6.1.4 Number of Stable Isotopes and Isotones

Nature itself votes on which nuclei are magic. The number of stable isotopes (nuclei with the same $Z$) and isotones (nuclei with the same $N$) peaks at the magic numbers:

• $Z = 50$ (tin): 10 stable isotopes — more than any other element
• $Z = 20$ (calcium): 5 stable isotopes, plus the doubly magic $^{48}$Ca with a half-life of $6.4 \times 10^{19}$ years
• $N = 82$: 7 stable isotones
• $N = 50$: 6 stable isotones
• $N = 20$: 5 stable isotones

Tin, with its 10 stable isotopes, is the champion of the periodic table. This is not a coincidence — it is the magic number $Z = 50$ conferring extraordinary stability across a wide range of neutron numbers.

6.1.5 Nuclear Charge Radii

Laser spectroscopy and electron scattering provide precision measurements of nuclear charge radii across isotopic chains. When these radii are plotted as a function of neutron number, the general trend follows $R \propto A^{1/3}$ with smooth, gradual increases. But at the magic numbers, the charge radii show distinctive kinks — sudden changes in the slope of $\langle r^2 \rangle$ versus $N$.

The physical origin of these kinks is subtle. When a major shell is being filled, each additional neutron contributes to the mean field in a way that slightly expands the proton distribution (through the proton-neutron interaction). When a shell closes, the next neutron begins filling a new, typically higher-$\ell$ orbit, which has a different spatial distribution. The resulting change in the mean field produces a detectable discontinuity in the charge radius.

The lead isotopes provide a particularly clean example. Measurements across the chain from $^{183}$Pb to $^{214}$Pb show a clear kink at $N = 126$, with the mean-square charge radius increasing more steeply for $N > 126$ than for $N < 126$. Similar kinks are observed at $N = 82$ in the barium, cerium, and neodymium isotopes, and at $N = 28$ in the calcium isotopes.

6.1.6 Electric Quadrupole Moments and Nuclear Shapes

The electric quadrupole moment $Q$ measures the deviation of the nuclear charge distribution from spherical symmetry. A positive $Q$ indicates a prolate (elongated) nucleus; a negative $Q$ indicates an oblate (flattened) nucleus; and $Q = 0$ indicates a spherical nucleus (or a nucleus with $J = 0$ or $J = 1/2$, for which $Q$ vanishes by angular momentum selection rules regardless of shape).

The pattern of quadrupole moments across the nuclear chart is striking. Nuclei at or near magic numbers have small or zero quadrupole moments — they are spherical. As we move away from magic numbers toward mid-shell, $|Q|$ grows rapidly, reaching values of several barns in the rare-earth and actinide regions. The largest known quadrupole moments occur in the lanthanide region ($Z \sim 66$, $N \sim 100$), where nuclei can be deformed by 30% or more from spherical shape.

This pattern is exactly what the shell model predicts. Closed-shell nuclei have no available single-particle degrees of freedom to support deformation; all orbits are filled, and the filled shell is inherently spherical. Mid-shell nuclei have many valence nucleons that can collectively polarize the shape through their mutual interactions.

6.1.7 Neutron Capture Cross Sections

Thermal neutron capture cross sections $\sigma_{n,\gamma}$ show dramatic minima at magic neutron numbers. When a slow neutron approaches a nucleus with $N$ equal to a magic number, the capture cross section can be orders of magnitude smaller than for neighboring isotopes. For example:

• $^{88}$Sr ($N = 50$): $\sigma_{n,\gamma} \approx 0.006$ b
• $^{89}$Y ($N = 50$): $\sigma_{n,\gamma} \approx 1.3$ b (but this has $Z$ not magic, so it's larger)
• $^{138}$Ba ($N = 82$): $\sigma_{n,\gamma} \approx 0.4$ b
• $^{208}$Pb ($N = 126$): $\sigma_{n,\gamma} \approx 0.0005$ b (one of the smallest known)

The small capture cross sections at magic $N$ reflect two factors: the large energy gap to the first available neutron state above the shell closure means there are few low-lying resonances to facilitate capture, and the high binding energy of the closed shell means the compound nucleus formed after capture has a relatively low excitation energy with a low level density.

These small cross sections have profound astrophysical consequences. In the slow neutron capture process (s-process), which builds elements in AGB stars, the nuclear reaction flow encounters "bottlenecks" at magic nuclei — the neutron capture is slow, so these nuclei accumulate. The s-process abundance peaks at $A \approx 88$ ($N = 50$), $A \approx 138$ ($N = 82$), and $A \approx 208$ ($N = 126$) are direct observational evidence for the nuclear magic numbers, written in the elemental abundances of the cosmos. We will return to this connection in Chapter 23 (Explosive Nucleosynthesis).

6.1.8 Summary of Evidence

The convergence of all these independent lines of evidence is overwhelming. Seven distinct observables — binding energy residuals, separation energy discontinuities, first-excited-state energies, stable isotope/isotone counts, charge radius kinks, quadrupole moments, and neutron capture cross sections — all single out the same special numbers: 2, 8, 20, 28, 50, 82, and 126. This is not a pattern that can be dismissed as coincidence. Something fundamental happens at these nucleon numbers. Our task is to explain what.

6.2 The Mean-Field Concept

How can nucleons — which interact via the strong nuclear force, as we developed in Chapter 3 — exhibit shell structure? Shell structure requires independent-particle motion: each nucleon moving in a well-defined orbit, largely unperturbed by the other nucleons. But nucleons are packed tightly inside the nucleus, separated by only about 1.8 fm on average. How can they avoid constant collisions?

The answer lies in the Pauli exclusion principle. When two nucleons inside the nucleus attempt to scatter, the final states must be unoccupied (since nucleons are fermions). But in a cold nucleus, all low-lying states are already filled. The only available final states lie above the Fermi surface, and for typical nucleon kinetic energies (~25 MeV), there is not enough energy to scatter into those states. The mean free path of a nucleon inside the nucleus is therefore much longer than the nuclear diameter — on the order of 5-7 fm compared to $R \approx r_0 A^{1/3} \approx 5$ fm for medium nuclei, and even longer for nucleons well below the Fermi surface.

This remarkable fact — that nucleons effectively pass through each other inside the nucleus, blocked from scattering by the Pauli principle — is the physical foundation of the shell model. It means we can replace the complicated two-body interactions between all $A$ nucleons with a single one-body mean-field potential $U(r)$ in which each nucleon moves independently.

Spaced review (Ch 3): Recall that the nuclear force is short-ranged (~1 fm) and strongly attractive. The mean-field potential is the average effect of all the other nucleons' individual pair interactions on a single nucleon. It is the nuclear analogue of the Hartree potential in atomic physics.

Formally, we start from the full $A$-body Hamiltonian:

$$H = \sum_{i=1}^{A} \frac{p_i^2}{2m} + \sum_{i < j} V_{ij}$$

where $V_{ij}$ is the nucleon-nucleon interaction. We introduce the mean-field potential $U(r_i)$ by writing:

$$H = \sum_{i=1}^{A} \left[ \frac{p_i^2}{2m} + U(r_i) \right] + \left[ \sum_{i

The first term $H_0$ is the independent-particle Hamiltonian — $A$ nucleons each moving in the same potential $U(r)$. The second term $H_{\text{res}}$ is the residual interaction — whatever the mean field did not capture. The art of the shell model is choosing $U(r)$ so that $H_{\text{res}}$ is as small as possible. We will treat $H_{\text{res}}$ as a perturbation in Chapter 7; for now, we set it to zero and solve $H_0$.

The single-particle Schrödinger equation is:

$$\left[ -\frac{\hbar^2}{2m} \nabla^2 + U(r) \right] \psi_{n \ell j m}(\mathbf{r}) = \varepsilon_{n \ell j} \, \psi_{n \ell j m}(\mathbf{r})$$

Since the potential is central (spherically symmetric), the angular part of the wave function is a spherical harmonic (or spin-angular function — see Section 6.5), and the quantum numbers are: - $n$: radial quantum number (number of radial nodes + 1, in the convention used in nuclear physics) - $\ell$: orbital angular momentum quantum number - $j$: total angular momentum ($j = \ell \pm 1/2$ after including spin) - $m$: projection of $j$ along the quantization axis

6.2.1 The Fermi Energy and the Nuclear Potential Well

We can estimate the depth of the mean-field potential from the nuclear Fermi energy. Inside the nucleus, nucleons are fermions filling states up to the Fermi energy $\varepsilon_F$. For a degenerate Fermi gas of $A$ nucleons in a volume $V = \frac{4}{3}\pi R^3 = \frac{4}{3}\pi r_0^3 A$, the Fermi momentum is:

$$p_F = \hbar \left(\frac{3\pi^2 \rho}{2}\right)^{1/3}$$

where $\rho = A/V$ is the nucleon number density and the factor of 2 in the denominator accounts for the two isospin states (protons and neutrons are counted separately in a more careful treatment). Using $\rho = 3/(4\pi r_0^3) \approx 0.16$ fm$^{-3}$ and $r_0 = 1.25$ fm:

$$p_F \approx \hbar \left(\frac{9\pi}{8 r_0^3}\right)^{1/3} \approx 250 \text{ MeV}/c$$

The corresponding Fermi kinetic energy is:

$$T_F = \frac{p_F^2}{2m} \approx \frac{(250)^2}{2 \times 939} \approx 33 \text{ MeV}$$

The average binding energy per nucleon is about 8 MeV, which means the separation energy is $S \approx 8$ MeV. The potential depth must be at least $V_0 \approx T_F + S \approx 33 + 8 = 41$ MeV. More careful treatments, accounting for the nuclear surface and the density dependence of the interaction, give $V_0 \approx 50$ MeV, consistent with the Woods-Saxon parametrization.

The average kinetic energy of nucleons in the Fermi gas is $\langle T \rangle = \frac{3}{5} T_F \approx 20$ MeV. This establishes the energy scale: the typical nucleon kinetic energy inside the nucleus is 20 MeV, which is small compared to the available excitation energy needed to scatter above the Fermi surface. This is why the Pauli blocking is so effective and the mean-field picture works.

The question is: what form should $U(r)$ take?

6.3 The Harmonic Oscillator Potential

The simplest choice for the mean-field potential is the three-dimensional isotropic harmonic oscillator:

$$U_{\text{HO}}(r) = \frac{1}{2} m \omega^2 r^2$$

This has the enormous advantage of being exactly solvable. The energy eigenvalues are:

$$\varepsilon_N = \left( N + \frac{3}{2} \right) \hbar \omega, \qquad N = 0, 1, 2, 3, \ldots$$

where $N = 2(n-1) + \ell$ is the principal oscillator quantum number. The parameter $\hbar\omega$ sets the level spacing and is empirically fitted as:

$$\hbar\omega \approx 41 \, A^{-1/3} \text{ MeV}$$

This parametrization comes from requiring that the oscillator potential reproduce the correct nuclear radius. The oscillator length parameter is $b = \sqrt{\hbar / m\omega}$, and the root-mean-square radius of the $A$-nucleon system in the ground state is $\langle r^2 \rangle^{1/2} \approx \sqrt{3/5} \, r_0 A^{1/3}$. Matching these gives $\hbar\omega \propto A^{-1/3}$, with the proportionality constant determined empirically.

For $^{16}$O: $\hbar\omega \approx 41 \times 16^{-1/3} = 16.3$ MeV. For $^{208}$Pb: $\hbar\omega \approx 41 \times 208^{-1/3} = 6.9$ MeV.

The level spacing decreases for heavier nuclei, which makes physical sense: heavier nuclei are larger, and a particle in a larger box has more closely spaced energy levels.

6.3.1 The Three-Dimensional Harmonic Oscillator Solution

The radial Schrodinger equation for the three-dimensional isotropic harmonic oscillator, after separating out the angular part, is:

$$-\frac{\hbar^2}{2m} \left[ \frac{d^2 u}{dr^2} - \frac{\ell(\ell+1)}{r^2} u \right] + \frac{1}{2} m \omega^2 r^2 u = \varepsilon \, u$$

where $u(r) = r R(r)$ is the reduced radial wave function. The solutions are associated Laguerre polynomials:

$$R_{n\ell}(r) = N_{n\ell} \left(\frac{r}{b}\right)^{\ell} L_{n-1}^{\ell+1/2}\left(\frac{r^2}{b^2}\right) \exp\left(-\frac{r^2}{2b^2}\right)$$

where $N_{n\ell}$ is a normalization constant and $b = \sqrt{\hbar/m\omega}$ is the oscillator length. The key result is that the energy depends only on $N = 2(n-1) + \ell$, not on $n$ and $\ell$ separately. This means that for a given $N$, all combinations of $n$ and $\ell$ satisfying $2(n-1) + \ell = N$ are degenerate. This $\ell$-degeneracy is an accidental symmetry of the harmonic oscillator — it arises from a hidden SU(3) symmetry of the $r^2$ potential and is broken by any deviation from the $r^2$ form.

6.3.2 Degeneracies and the First Magic Numbers

For each value of $N$, the allowed $\ell$ values are $N, N-2, N-4, \ldots, 1 \text{ or } 0$. Each $\ell$ sublevel has a degeneracy of $2(2\ell + 1)$ — the factor of 2 from the two spin orientations and $2\ell + 1$ from the magnetic substates.

The total degeneracy of each $N$-shell is:

$$g_N = (N+1)(N+2)$$

Let us enumerate:

The cumulative particle numbers at each shell closure are: 2, 8, 20, 40, 70, 112, ...

The harmonic oscillator gives us the first three magic numbers correctly! But 40, 70, and 112 are not magic numbers. The observed magic numbers above 20 are 28, 50, 82, and 126 — completely different from the oscillator predictions.

6.3.2 Why the Harmonic Oscillator Fails

The harmonic oscillator potential has two fundamental deficiencies:

1. Wrong shape at large $r$: The HO potential rises as $r^2$ to infinity, confining the nucleon at all distances. A real nuclear potential has a finite depth and goes to zero beyond the nuclear surface. This means the HO over-confines nucleons and overestimates the splitting between high-$\ell$ and low-$\ell$ orbits within each $N$-shell.

2. Missing spin-orbit interaction: The HO has no mechanism for breaking the degeneracy between states of different $j$ within a given $\ell$ sublevel. As we will see in Section 6.5, the spin-orbit interaction is the essential missing ingredient.

Let us address the first deficiency before tackling the second.

6.4 The Woods-Saxon Potential

A more realistic mean-field potential should have a flat bottom (reflecting nuclear density saturation — the interior of a large nucleus has nearly constant density) and a smooth fall-off at the nuclear surface. The Woods-Saxon potential satisfies both requirements:

$$U_{\text{WS}}(r) = \frac{-V_0}{1 + \exp\left(\frac{r - R}{a}\right)}$$

where: - $V_0 \approx 50$ MeV is the well depth - $R = r_0 A^{1/3}$ with $r_0 \approx 1.25$ fm is the half-density radius - $a \approx 0.65$ fm is the surface diffuseness

Spaced review (Ch 1): The Woods-Saxon form is just the Fermi distribution for the nuclear density $\rho(r)$ that we encountered in Chapter 1, now used as a potential shape. This connection is self-consistent: if the potential seen by each nucleon is proportional to the density of the other nucleons, then $U(r) \propto \rho(r)$, and the potential should have the same shape as the density.

For protons, we must add the Coulomb potential:

$$U_C(r) = \begin{cases} \frac{Ze^2}{4\pi\epsilon_0} \frac{1}{2R} \left(3 - \frac{r^2}{R^2}\right) & r \leq R \\ \frac{Ze^2}{4\pi\epsilon_0} \frac{1}{r} & r > R \end{cases}$$

6.4.1 Effects of the Realistic Shape

The Woods-Saxon potential cannot be solved analytically — it requires numerical methods (or, as a pedagogical approximation, the "intermediate" approach of a finite square well). However, the qualitative effects of the more realistic shape are clear:

1. Level ordering within each $N$-shell changes. In the harmonic oscillator, all sublevels within a given $N$ are degenerate. In the Woods-Saxon potential, higher-$\ell$ orbits are pushed down in energy relative to lower-$\ell$ orbits. This is because high-$\ell$ nucleons have their wave function concentrated at larger radii (due to the centrifugal barrier $\ell(\ell+1)\hbar^2/2mr^2$), where the Woods-Saxon potential is less deep than the harmonic oscillator. Wait — that argument would push high-$\ell$ up. The actual effect is the opposite: the flat bottom of the Woods-Saxon means that the centrifugal barrier is the dominant energy contributor at small $r$, and high-$\ell$ orbits effectively sample a deeper potential at the surface. More precisely, the $\ell$-degeneracy of the harmonic oscillator is an accidental symmetry; any deviation from the $r^2$ shape breaks it, and for a flat-bottomed well, the high-$\ell$ states are indeed lowered.

2. Partial lifting of the HO degeneracy. The $N = 3$ shell, for example, splits into $1f$ and $2p$, with $1f$ lower. But the splitting is not large enough to close the shell gaps at 28, 50, 82, and 126. The Woods-Saxon potential alone gives magic numbers approximately at 2, 8, 20, 34, 58, 92, 138 — closer to reality than the HO, but still wrong.

6.4.2 A Numerical Example: $^{208}$Pb

To make the Woods-Saxon potential concrete, let us evaluate it for $^{208}$Pb. The parameters are:

• $V_0 = 50$ MeV
• $R = 1.25 \times 208^{1/3} = 1.25 \times 5.925 = 7.41$ fm
• $a = 0.65$ fm

At the center ($r = 0$): $U(0) = -50 / (1 + e^{-7.41/0.65}) = -50 / (1 + e^{-11.4}) \approx -50$ MeV.

At the half-radius ($r = R$): $U(R) = -50 / (1 + 1) = -25$ MeV.

At $r = R + 2a = 8.71$ fm: $U = -50 / (1 + e^{2}) \approx -50/8.4 \approx -5.9$ MeV.

At $r = R + 4a = 10.01$ fm: $U \approx -50/55.6 \approx -0.9$ MeV.

The potential is essentially flat at $-50$ MeV throughout the nuclear interior, then drops to half its depth at the nuclear surface $r = R$, and effectively vanishes for $r > R + 3a \approx 9.4$ fm. This is dramatically different from the harmonic oscillator, which at $r = R$ would have $U_{\text{HO}}(R) = \frac{1}{2} m \omega^2 R^2 \approx 35$ MeV (positive — already well above zero and climbing).

For protons in $^{208}$Pb, the Coulomb potential adds approximately $V_C(0) = 3Ze^2/(8\pi\epsilon_0 R) \approx 3 \times 82 \times 1.44 / (2 \times 7.41) \approx 24$ MeV to the nuclear potential at the center, reducing the effective well depth to about 26 MeV. This significant Coulomb correction is why the proton single-particle energies differ from the neutron single-particle energies, and why the proton and neutron shell gaps are not identical.

The punchline: a central potential — no matter how carefully chosen — cannot reproduce the magic numbers above 20. Something beyond a central force is required. That something is the spin-orbit interaction.

6.5 Spin-Orbit Coupling: The Key to Magic Numbers

6.5.1 The Spin-Orbit Interaction in Nuclear Physics

Every nucleon has orbital angular momentum $\boldsymbol{\ell}$ and intrinsic spin $\mathbf{s}$ (with $s = 1/2$). These couple to form the total angular momentum:

$$\mathbf{j} = \boldsymbol{\ell} + \mathbf{s}$$

with $j = \ell + 1/2$ or $j = \ell - 1/2$.

The crucial spin-orbit potential is:

$$U_{ls}(r) = V_{ls}(r) \, \boldsymbol{\ell} \cdot \mathbf{s}$$

where $V_{ls}(r)$ is a radial function peaked at the nuclear surface (typically the derivative of the Woods-Saxon potential):

$$V_{ls}(r) = -V_{ls}^0 \frac{1}{r} \frac{d}{dr} \left[ \frac{1}{1 + \exp\left(\frac{r - R}{a}\right)} \right]$$

with $V_{ls}^0 \approx 20$-$35$ MeV, depending on the parametrization.

The expectation value of $\boldsymbol{\ell} \cdot \mathbf{s}$ is:

$$\langle \boldsymbol{\ell} \cdot \mathbf{s} \rangle = \frac{1}{2} \left[ j(j+1) - \ell(\ell+1) - s(s+1) \right]$$

For $j = \ell + 1/2$ (spin and orbital angular momentum parallel, the "stretched" state):

$$\langle \boldsymbol{\ell} \cdot \mathbf{s} \rangle_{j = \ell + 1/2} = \frac{\ell}{2}$$

For $j = \ell - 1/2$ (spin and orbital angular momentum antiparallel):

$$\langle \boldsymbol{\ell} \cdot \mathbf{s} \rangle_{j = \ell - 1/2} = -\frac{(\ell + 1)}{2}$$

The energy splitting between the two $j$-partners is:

$$\Delta E_{ls} = \varepsilon_{j=\ell-1/2} - \varepsilon_{j=\ell+1/2} = \frac{2\ell + 1}{2} \langle V_{ls}(r) \rangle$$

This splitting grows linearly with $\ell$. For small $\ell$, the splitting is modest; for large $\ell$, it is enormous — large enough to push the $j = \ell + 1/2$ member of a high-$\ell$ orbit down into the shell below.

6.5.2 The Critical Sign Convention

The nuclear spin-orbit force is attractive for $j = \ell + 1/2$ and repulsive for $j = \ell - 1/2$. In our convention, $V_{ls}^0 > 0$, and the $j = \ell + 1/2$ level is pushed down in energy while the $j = \ell - 1/2$ level is pushed up.

Warning: atomic vs. nuclear. In atomic physics, the spin-orbit interaction is a relativistic correction ($\sim Z^4 \alpha^4 m_e c^2$) and is typically a small perturbation. In nuclear physics, the spin-orbit splitting is enormous — comparable to the spacing between major shells. This was the astonishing realization of Mayer and Jensen. The nuclear spin-orbit force is not a small correction. It is a dominant feature of nuclear structure.

The nuclear spin-orbit force has a different origin from the atomic one. The atomic $\boldsymbol{\ell} \cdot \mathbf{s}$ coupling arises from the interaction between the electron's magnetic moment and the magnetic field in its rest frame (Thomas precession). The nuclear $\boldsymbol{\ell} \cdot \mathbf{s}$ coupling arises from the spin-orbit component of the nucleon-nucleon interaction itself — specifically, from the tensor and spin-orbit components of the NN force that we discussed in Chapter 3. Its magnitude (20-35 MeV) is about $10^3$ times larger, relative to the level spacing, than the atomic spin-orbit coupling.

6.5.3 How Spin-Orbit Coupling Produces the Correct Magic Numbers

Let us trace through the shell filling with spin-orbit coupling, starting from the Woods-Saxon + spin-orbit potential. The key events:

$N = 3$ shell ($1f$, $2p$): Without spin-orbit, this shell has 20 states and fills at cumulative particle number 40. With spin-orbit: - $1f_{7/2}$: $j = 7/2$, degeneracy $2j+1 = 8$, pushed down (large $\ell = 3$, big splitting) - $1f_{5/2}$: $j = 5/2$, degeneracy 6, pushed up - $2p_{3/2}$: degeneracy 4 - $2p_{1/2}$: degeneracy 2

The $1f_{7/2}$ orbital is pushed down so far that it joins the $N = 2$ shell, creating a gap after 28 particles (= 20 from the $N = 0,1,2$ shells + 8 from $1f_{7/2}$). This gives us magic number 28.

$N = 4$ shell ($1g$, $2d$, $3s$): The $1g_{9/2}$ ($j = 9/2$, degeneracy 10) is pushed down to join the remaining $N = 3$ orbits. The shell closes at $28 + 22 = 50$. Magic number 50.

$N = 5$ shell ($1h$, $2f$, $3p$): The $1h_{11/2}$ ($j = 11/2$, degeneracy 12) is pushed down to join the remaining $N = 4$ orbits. Shell closure at $50 + 32 = 82$. Magic number 82.

$N = 6$ shell ($1i$, $2g$, $3d$, $4s$): The $1i_{13/2}$ ($j = 13/2$, degeneracy 14) is pushed down to join the remaining $N = 5$ orbits. Shell closure at $82 + 44 = 126$. Magic number 126.

Let us verify the arithmetic for each shell closure, since getting this right is essential:

Magic 28: Shells 1-3 hold $2 + 6 + 12 = 20$ nucleons. The $1f_{7/2}$ intruder from the $N = 3$ shell holds 8 more. Total: $20 + 8 = 28$.

Magic 50: After 28, the remaining $N = 3$ orbits ($1f_{5/2}$: 6, $2p_{3/2}$: 4, $2p_{1/2}$: 2) hold 12 nucleons. The $1g_{9/2}$ intruder from $N = 4$ holds 10 more. Total: $28 + 12 + 10 = 50$.

Magic 82: After 50, the remaining $N = 4$ orbits ($1g_{7/2}$: 8, $2d_{5/2}$: 6, $2d_{3/2}$: 4, $3s_{1/2}$: 2) hold 20 nucleons. The $1h_{11/2}$ intruder from $N = 5$ holds 12 more. Total: $50 + 20 + 12 = 82$.

Magic 126: After 82, the remaining $N = 5$ orbits ($1h_{9/2}$: 10, $2f_{7/2}$: 8, $2f_{5/2}$: 6, $3p_{3/2}$: 4, $3p_{1/2}$: 2) hold 30 nucleons. The $1i_{13/2}$ intruder from $N = 6$ holds 14 more. Total: $82 + 30 + 14 = 126$.

The pattern is clear: at each shell closure above 20, the $j = \ell + 1/2$ member of the highest-$\ell$ orbit in the next oscillator shell is pulled down into the current shell by the spin-orbit force. The intruder orbit has the highest $j$ value in the shell, and because $\ell$ increases with each successive shell, the intruder always has a large enough degeneracy ($2j + 1$) to create a significant gap. The mechanism is elegant and universal — a single physical effect (spin-orbit coupling) with a single sign convention (attractive for $j = \ell + 1/2$) produces all four magic numbers above 20.

This is the shell model. The spin-orbit force — that single $\boldsymbol{\ell} \cdot \mathbf{s}$ term — is the key to the entire magic number sequence. Without it, nuclear physics would be a different subject.

6.6 The Complete Shell Model Level Scheme

We can now assemble the complete picture. The Woods-Saxon potential with spin-orbit coupling gives us a unique ordering of single-particle energy levels. This ordering is not derived from first principles in a single closed-form expression — it requires solving the Schrodinger equation numerically for the full potential. But the resulting level scheme is well established, confirmed by decades of experimental measurements of single-particle states in nuclei near doubly magic closures, and it forms the basis for all shell model calculations.

The standard filling order, from lowest to highest energy, is:

$$1s_{1/2} \quad | \quad 1p_{3/2}, \, 1p_{1/2} \quad | \quad 1d_{5/2}, \, 2s_{1/2}, \, 1d_{3/2} \quad | \quad 1f_{7/2} \quad |$$

$$2p_{3/2}, \, 1f_{5/2}, \, 2p_{1/2}, \, 1g_{9/2} \quad | \quad 1g_{7/2}, \, 2d_{5/2}, \, 2d_{3/2}, \, 3s_{1/2}, \, 1h_{11/2} \quad |$$

$$2f_{7/2}, \, 1h_{9/2}, \, 3p_{3/2}, \, 2f_{5/2}, \, 3p_{1/2}, \, 1i_{13/2} \quad |$$

where the vertical bars indicate the magic shell closures.

The cumulative nucleon counts at each gap:

Every magic number is now accounted for. This is not a fit — the magic numbers emerge naturally from the eigenvalues of a Woods-Saxon potential with spin-orbit coupling. The shell model has only a few adjustable parameters (the well depth $V_0$, the radius parameter $r_0$, the diffuseness $a$, and the spin-orbit strength $V_{ls}^0$), and these are constrained by other data (nuclear sizes, separation energies). The magic numbers are predictions, not inputs.

It is worth emphasizing how remarkable this is. The magic numbers 2, 8, 20, 28, 50, 82, 126 are not simple products of any obvious number-theoretic sequence. They are not multiples of 2 or powers of 2. They do not follow any polynomial pattern. They are the specific cumulative occupancies at which the particular eigenvalue spectrum of a three-dimensional Woods-Saxon + spin-orbit potential produces large energy gaps. The magic numbers are, in a deep sense, consequences of the shape of the nuclear force.

6.6.1 Notation Conventions

Nuclear physics uses a specific labeling convention for single-particle orbits: $n\ell_j$, where:

• $n$ is the radial quantum number (counting from 1, not 0 — the lowest $s$-state is $1s$, not $0s$)
• $\ell$ is given in spectroscopic notation: $s, p, d, f, g, h, i, j, k, \ldots$ for $\ell = 0, 1, 2, 3, 4, 5, 6, 7, 8, \ldots$
• $j$ is the total angular momentum quantum number

Each orbit $n\ell_j$ can accommodate $2j + 1$ nucleons (the magnetic substates $m_j = -j, -j+1, \ldots, +j$). The factor of 2 for spin is already incorporated through $j$.

6.6.2 The Parity of Each Level

The parity of a single-particle state is determined solely by $\ell$:

$$\pi = (-1)^{\ell}$$

States with even $\ell$ ($s, d, g, i, \ldots$) have positive parity; states with odd $\ell$ ($p, f, h, j, \ldots$) have negative parity. This will be crucial for ground-state predictions.

6.7 Ground-State Predictions: Spin, Parity, and the Pairing Interaction

The power of the shell model lies in its predictive ability. Given the proton number $Z$ and neutron number $N$, we can predict the ground-state spin and parity $J^{\pi}$ using the following rules.

6.7.1 Even-Even Nuclei

For even-even nuclei (even $Z$, even $N$), the prediction is simple and spectacularly successful:

$$J^{\pi} = 0^+$$

for every even-even nucleus.

This is because of the pairing interaction: nucleons in a given orbit $n\ell_j$ tend to form pairs with $m_j$ and $-m_j$, coupling to total angular momentum zero. When all occupied orbits are completely filled (or filled with an even number of nucleons that all pair up), the total spin is zero and the total parity is positive.

This prediction is confirmed for every known even-even nucleus — over 800 ground states measured, with zero exceptions. No other model in nuclear physics has such a perfect track record.

6.7.2 Odd-$A$ Nuclei: The Single-Particle Prediction

For odd-$A$ nuclei — those with an odd number of either protons or neutrons (but not both) — the ground-state $J^{\pi}$ is determined by the unpaired nucleon. The even-nucleon group pairs to $0^+$ and contributes nothing to the ground-state spin and parity. The single unpaired nucleon determines everything.

The procedure: 1. Identify the odd species (proton or neutron). 2. Count the number of odd-species nucleons. 3. Fill the single-particle levels in order, placing nucleons according to the shell model filling scheme. 4. The last unpaired nucleon occupies some orbit $n\ell_j$. 5. The predicted ground state is $J^{\pi} = j^{(-1)^{\ell}}$.

Example 1: $^{17}$O ($Z = 8$, $N = 9$) — The textbook case

Protons: $Z = 8$ (magic, all shells closed through $1p_{1/2}$) $\rightarrow 0^+$ contribution.

Neutrons: $N = 9$. The first 8 neutrons fill: $1s_{1/2}$ (2), $1p_{3/2}$ (4), $1p_{1/2}$ (2) = 8. The 9th neutron goes into the next available orbit: $1d_{5/2}$.

Prediction: $J^{\pi} = 5/2^+$. Experiment: $J^{\pi} = 5/2^+$. Confirmed.

This is the cleanest possible shell model test: a single neutron outside a doubly magic core ($^{16}$O). The 9th neutron is in the $1d_{5/2}$ orbit, which has $\ell = 2$ (positive parity) and $j = 5/2$. Everything about $^{17}$O's ground state — its spin, its parity, and (as we will see in Section 6.8) approximately its magnetic moment — follows from identifying that single neutron's orbit.

Example 2: $^{41}$Ca ($Z = 20$, $N = 21$) — Crossing the $N = 20$ closure

Protons: $Z = 20$ (magic) $\rightarrow 0^+$.

Neutrons: $N = 21$. The first 20 fill shells through $1d_{3/2}$. The 21st neutron enters $1f_{7/2}$.

Prediction: $J^{\pi} = 7/2^-$. Experiment: $J^{\pi} = 7/2^-$. Confirmed.

Note the parity: $\ell = 3$ gives $\pi = (-1)^3 = -1$, so the parity is negative. Also note that $1f_{7/2}$ is the beginning of a new shell (shell 4, leading to magic number 28), so $^{41}$Ca has one neutron in a completely empty shell — another clean single-particle test case.

Example 3: $^{91}$Zr ($Z = 40$, $N = 51$) — Beyond the $N = 50$ closure

Protons: $Z = 40$ is not magic, but is even, so all protons pair to $0^+$.

Neutrons: $N = 51 = 50 + 1$. The first 50 fill through $1g_{9/2}$. The 51st neutron enters $2d_{5/2}$, the next orbit in the filling order.

Prediction: $J^{\pi} = 5/2^+$ ($\ell = 2$, positive parity). Experiment: $J^{\pi} = 5/2^+$. Confirmed.

Example 4: $^{207}$Pb ($Z = 82$, $N = 125$) — A hole state

Protons: $Z = 82$ (magic) $\rightarrow 0^+$.

Neutrons: $N = 125$ = 126 - 1. This is one neutron short of the magic number 126. The filling order for the $N = 82$-$126$ shell ends with $1i_{13/2}$, but just before that comes $3p_{1/2}$ (the last orbit before $1i_{13/2}$). Actually, we need to be precise about which orbit has the hole. The complete filling of this shell is: $2f_{7/2}$(8) + $1h_{9/2}$(10) + $3p_{3/2}$(4) + $2f_{5/2}$(6) + $3p_{1/2}$(2) + $1i_{13/2}$(14) = 44 nucleons (from 82 to 126). With 125 neutrons, we are missing one from $3p_{1/2}$ — the second-to-last orbit. A hole in $3p_{1/2}$ has the same $J^{\pi}$ as a particle in $3p_{1/2}$.

Prediction: $J^{\pi} = 1/2^-$ ($\ell = 1$, negative parity). Experiment: $J^{\pi} = 1/2^-$. Confirmed.

The concept of a hole state is powerful. One neutron missing from a closed shell is mathematically equivalent to one neutron present in that orbit — the hole carries the same quantum numbers as the absent particle. This particle-hole symmetry, familiar from solid-state physics (holes in a semiconductor valence band), is equally important in nuclear physics.

Example 5: $^{209}$Bi ($Z = 83$, $N = 126$) — The first proton beyond $Z = 82$

Neutrons: $N = 126$ (magic) $\rightarrow 0^+$.

Protons: $Z = 83 = 82 + 1$. The 83rd proton enters the first proton orbit above $Z = 82$, which is $1h_{9/2}$.

Prediction: $J^{\pi} = 9/2^-$ ($\ell = 5$, negative parity). Experiment: $J^{\pi} = 9/2^-$. Confirmed.

$^{209}$Bi is notable as the heaviest nucleus with a traditionally "stable" isotope (its half-life against alpha decay is $1.9 \times 10^{19}$ years, far longer than the age of the universe). The single unpaired proton in $1h_{9/2}$ gives it a large spin that makes its alpha decay to $^{205}$Tl highly forbidden — a direct consequence of the shell model.

6.7.3 Odd-Odd Nuclei

For odd-odd nuclei (odd $Z$, odd $N$), we have one unpaired proton and one unpaired neutron. Their angular momenta must be coupled:

$$\mathbf{J} = \mathbf{j}_p + \mathbf{j}_n$$

The allowed values of $J$ range from $|j_p - j_n|$ to $j_p + j_n$. Determining which value is the ground state requires knowledge of the residual interaction between the two unpaired nucleons (the Nordheim rules and their extensions). The parity is unambiguous:

$$\pi = (-1)^{\ell_p + \ell_n}$$

We defer the detailed treatment of odd-odd nuclei to Chapter 7, where we discuss the residual interaction.

6.7.4 Scorecard: How Well Do the Predictions Work?

The shell model predictions for ground-state $J^{\pi}$ are remarkably successful for nuclei near closed shells. A systematic comparison for odd-$A$ nuclei within two particles of a magic number yields correct predictions in approximately 90% of cases. The successes include:

The failures — and there are some — tend to occur for nuclei far from closed shells, where deformation sets in and the spherical shell model breaks down. We will discuss these limitations in Section 6.9.

6.8 Magnetic Moments: Schmidt Values

A more stringent test of the shell model is its prediction of nuclear magnetic moments. If the ground state of an odd-$A$ nucleus is indeed determined by a single unpaired nucleon, then the nuclear magnetic moment should be that of the single nucleon in its orbit.

6.8.1 The Nuclear Magnetic Moment

The magnetic moment of a nucleus is defined as:

$$\mu = g_J \, j \, \mu_N$$

where $\mu_N = e\hbar/2m_p = 3.152 \times 10^{-8}$ eV/T is the nuclear magneton, and $g_J$ is the nuclear $g$-factor.

For a single nucleon in orbit $n\ell_j$, the magnetic moment receives contributions from both the orbital motion and the intrinsic spin:

$$\boldsymbol{\mu} = g_\ell \, \boldsymbol{\ell} \, \mu_N + g_s \, \mathbf{s} \, \mu_N$$

The orbital $g$-factors are: $$g_\ell = \begin{cases} 1 & \text{(proton)} \\ 0 & \text{(neutron)} \end{cases}$$

The spin $g$-factors (measured, not Dirac values): $$g_s = \begin{cases} +5.586 & \text{(proton)} \\ -3.826 & \text{(neutron)} \end{cases}$$

6.8.2 Derivation of the Schmidt Values

The magnetic moment of a nucleus in state $|j, m_j = j\rangle$ is:

$$\mu = \langle j, m_j = j | g_\ell \ell_z + g_s s_z | j, m_j = j \rangle \, \mu_N$$

Using the projection theorem (a consequence of the Wigner-Eckart theorem from Chapter 5):

$$\langle \ell_z \rangle = \frac{\langle \mathbf{j} \cdot \boldsymbol{\ell} \rangle}{j(j+1)} \, j, \qquad \langle s_z \rangle = \frac{\langle \mathbf{j} \cdot \mathbf{s} \rangle}{j(j+1)} \, j$$

Spaced review (Ch 5): The projection theorem states that for any vector operator $\mathbf{V}$ acting within a subspace of definite $j$, the expectation value of $V_z$ in the state $|j, m_j = j\rangle$ is proportional to $j$, with proportionality constant $\langle \mathbf{j} \cdot \mathbf{V} \rangle / j(j+1)$.

Using $\mathbf{j} = \boldsymbol{\ell} + \mathbf{s}$: $$\langle \mathbf{j} \cdot \boldsymbol{\ell} \rangle = \frac{1}{2}\left[ j(j+1) + \ell(\ell+1) - s(s+1) \right]$$ $$\langle \mathbf{j} \cdot \mathbf{s} \rangle = \frac{1}{2}\left[ j(j+1) - \ell(\ell+1) + s(s+1) \right]$$

Substituting with $s = 1/2$:

For $j = \ell + 1/2$:

$$\mu = \left[ g_\ell \left( j - \frac{1}{2} \right) + \frac{1}{2} g_s \right] \mu_N = \left[ j \left( g_\ell - \frac{g_s}{2j} + \frac{g_s}{2j} \right) \right] \mu_N$$

Simplifying: $$\mu = \left[ \left( j - \frac{1}{2} \right) g_\ell + \frac{1}{2} g_s \right] \mu_N$$

For $j = \ell - 1/2$:

$$\mu = \frac{j}{j+1} \left[ \left( j + \frac{3}{2} \right) g_\ell - \frac{1}{2} g_s \right] \mu_N$$

These are the Schmidt values — the magnetic moments predicted by the extreme single-particle shell model.

6.8.3 Worked Examples

Let us compute the Schmidt values for several specific nuclei:

$^{17}$O: odd neutron in $1d_{5/2}$ ($j = 5/2, \ell = 2$, so $j = \ell + 1/2$):

$$\mu = \left(\frac{5}{2} - \frac{1}{2}\right) \times 0 + \frac{1}{2} \times (-3.826) = -1.913 \, \mu_N$$

Experimental value: $\mu = -1.894 \, \mu_N$. The agreement is excellent — within 1%.

$^{209}$Bi: odd proton in $1h_{9/2}$ ($j = 9/2, \ell = 5$, so $j = \ell - 1/2$):

$$\mu = \frac{9/2}{9/2 + 1} \left[\left(\frac{9}{2} + \frac{3}{2}\right) \times 1 - \frac{1}{2} \times 5.586 \right] = \frac{9}{11} \left[6 - 2.793\right] = \frac{9}{11} \times 3.207 = 2.624 \, \mu_N$$

Experimental value: $\mu = +4.111 \, \mu_N$. This is a significant overestimate by the experiment relative to Schmidt — the ratio is 1.57. $^{209}$Bi is only one proton beyond the $Z = 82$ closure, yet the Schmidt value already deviates by 57%. This deviation arises primarily from core polarization: the unpaired proton polarizes the $^{208}$Pb core, inducing small $M1$ admixtures from particle-hole excitations that shift the magnetic moment away from the pure single-particle value.

$^{15}$N: odd proton in $1p_{1/2}$ ($j = 1/2, \ell = 1$, so $j = \ell - 1/2$):

$$\mu = \frac{1/2}{1/2 + 1} \left[\left(\frac{1}{2} + \frac{3}{2}\right) \times 1 - \frac{1}{2} \times 5.586 \right] = \frac{1}{3} \left[2 - 2.793\right] = \frac{1}{3} \times (-0.793) = -0.264 \, \mu_N$$

Experimental value: $\mu = -0.283 \, \mu_N$. Agreement within 7%.

6.8.4 The Schmidt Lines

If we plot the Schmidt predictions for odd-proton nuclei and odd-neutron nuclei separately as a function of $j$, we get two lines for each (corresponding to $j = \ell + 1/2$ and $j = \ell - 1/2$). These are the "Schmidt lines."

For odd-proton nuclei ($g_\ell = 1, g_s = 5.586$):

$j = \ell + 1/2$: $\mu = \left(j - \frac{1}{2} + 2.793\right) \mu_N = (j + 2.293) \mu_N$

$j = \ell - 1/2$: $\mu = \frac{j}{j+1} \left(j + \frac{3}{2} - 2.793\right) \mu_N = \frac{j}{j+1}(j - 1.293) \mu_N$

Note the remarkable feature for odd-proton nuclei with $j = \ell + 1/2$: the magnetic moment increases linearly with $j$. The proton's orbital motion and spin both contribute in the same direction, producing large magnetic moments that grow with the size of the orbit.

For odd-neutron nuclei ($g_\ell = 0, g_s = -3.826$):

$j = \ell + 1/2$: $\mu = -1.913 \, \mu_N$

$j = \ell - 1/2$: $\mu = \frac{j}{j+1} \cdot 1.913 \, \mu_N$

The odd-neutron Schmidt values reveal something striking: for $j = \ell + 1/2$, the magnetic moment is independent of $j$. It is always $-1.913 \, \mu_N$ — exactly half the free neutron magnetic moment — regardless of which orbit the neutron occupies. This is because the neutron has $g_\ell = 0$ (no orbital contribution to the magnetic moment), so the entire magnetic moment comes from the spin, and the projection of the spin along the $j$-axis is always $1/2$ for the stretched state $j = \ell + 1/2$.

6.8.4 Comparison with Experiment

The Schmidt values are only approximately correct. When experimental magnetic moments are plotted against the Schmidt lines, they fall between the two lines — closer to the Schmidt values than to any other simple prediction, but with significant scatter.

Some representative comparisons:

The pattern is clear: nuclei near closed shells ($^{17}$O, $^{207}$Pb, $^{15}$N) agree well with Schmidt values. Nuclei far from closed shells ($^{133}$Cs, $^{209}$Bi) deviate significantly. The deviations arise because the single-particle picture breaks down: the "unpaired" nucleon is actually coupled to the core, which is not inert but is polarized by the valence nucleon. These corrections involve the residual interaction and configuration mixing — topics for Chapter 7.

Nevertheless, the Schmidt lines provide the correct sign and order of magnitude for virtually all measured nuclear magnetic moments. This is strong evidence that the single-particle picture captures the essential physics, even when quantitative agreement requires refinements.

6.9 Where the Shell Model Succeeds and Where It Fails

6.9.1 Spectacular Successes

The shell model is most powerful for:

1. Nuclei near closed shells. The predictions for $J^{\pi}$ are nearly perfect within one or two nucleons of a magic number. The entire low-lying spectroscopy of nuclei like $^{17}$O, $^{41}$Ca, $^{89}$Y, $^{91}$Zr, $^{133}$Sb, $^{207}$Pb, and $^{209}$Bi is quantitatively described.

2. Doubly magic nuclei. The properties of $^{4}$He, $^{16}$O, $^{40}$Ca, $^{48}$Ca, $^{56}$Ni, $^{100}$Sn, $^{132}$Sn, and $^{208}$Pb — including their high first excited states, spherical shapes, and $0^+$ ground states — are natural consequences.

3. Magic number systematics. The entire pattern of enhanced stability, separation energy discontinuities, and first-excited-state spikes at $N$ or $Z = 2, 8, 20, 28, 50, 82, 126$ emerges from a single potential with one crucial ingredient (spin-orbit coupling).

4. Isomeric states. The shell model predicts which nuclei should have long-lived excited states (isomers), based on large spin differences between adjacent single-particle orbits. The "islands of isomerism" in the nuclear chart (near $Z$ or $N = 40, 50, 82, 126$) are exactly where the shell model places large $\Delta j$ transitions.

5. Ground-state spins and parities for odd-$A$ nuclei near magic numbers: ~90% correct.

6.9.2 Where It Begins to Fail

The simple (extreme single-particle) shell model fails or becomes inadequate in several regimes:

1. Mid-shell nuclei. When many nucleons occupy partially filled shells, the assumption that only one nucleon matters breaks down. The rare-earth region ($50 < Z < 82$, $82 < N < 126$) and the actinide region ($Z > 82$, $N > 126$) are particularly problematic for the single-particle shell model.

2. Deformed nuclei. Nuclei far from closed shells develop large quadrupole deformations — they are prolate (cigar-shaped) or oblate (frisbee-shaped) rather than spherical. For these nuclei, the spherical shell model is the wrong starting point. One must use the deformed shell model (Nilsson model, Chapter 7) or the collective models (Chapter 8).

3. Magnetic moments far from closed shells. As we saw in Section 6.8, the Schmidt values become increasingly inaccurate away from magic numbers. Corrections from core polarization, meson exchange currents, and configuration mixing are needed.

4. Transition rates. The single-particle model predicts electromagnetic transition rates (Weisskopf estimates, Chapter 9) to within an order of magnitude, but quantitative agreement requires the full configuration-interaction shell model with effective charges and operators.

5. Exotic nuclei far from stability. Recent experiments at radioactive beam facilities (FRIB, RIKEN, ISOLDE) have revealed that magic numbers are not immutable. The conventional magic numbers 8, 20, and 28 weaken or vanish in very neutron-rich nuclei, while new magic numbers (e.g., $N = 16, 32, 34$) emerge. This "shell evolution" far from stability is one of the frontier topics of modern nuclear structure and is covered in Chapter 10.

6.9.3 Shell Evolution Far from Stability

One of the most exciting developments in nuclear structure physics over the past two decades has been the discovery that magic numbers are not universal constants. In nuclei far from the valley of stability — particularly very neutron-rich nuclei accessible at radioactive beam facilities such as FRIB (Facility for Rare Isotope Beams), RIKEN (Japan), and ISOLDE (CERN) — the conventional magic numbers can weaken or disappear, and new magic numbers can emerge.

The most dramatic example is the "island of inversion" around $^{32}$Mg ($Z = 12$, $N = 20$). Despite having the magic neutron number $N = 20$, $^{32}$Mg does not behave like a closed-shell nucleus. Its first excited $2^+$ state is at only 0.885 MeV (low for a supposed magic nucleus), and its $B(E2)$ transition probability is anomalously large, indicating substantial quadrupole deformation. The explanation is that the neutron $1f_{7/2}$ and $2p_{3/2}$ orbits from the $fp$-shell intrude below the $sd$-shell orbits when $Z$ is small, effectively erasing the $N = 20$ gap. The driving mechanism is the proton-neutron tensor force: when the proton $1d_{5/2}$ orbit is emptied (as $Z$ decreases from 20 toward 10), the attractive proton-neutron tensor interaction that had been pulling the neutron $1d_{3/2}$ orbit down weakens, while the neutron $1f_{7/2}$ orbit drops, closing the gap.

Similarly, the magic number $N = 28$ appears to weaken in very neutron-rich nuclei near $^{42}$Si ($Z = 14$, $N = 28$), and new shell closures at $N = 16$ and $N = 32, 34$ have been identified experimentally. These phenomena — collectively known as "shell evolution" — are a major focus of modern nuclear structure research and are covered in detail in Chapter 10.

The lesson for the student is important: the magic numbers 2, 8, 20, 28, 50, 82, 126 are properties of nuclei near stability. They are not engraved on stone tablets. The single-particle energies are themselves determined by the proton-neutron composition of the nucleus, and when that composition changes dramatically (as in nuclei with extreme neutron-to-proton ratios), the shell structure changes with it.

6.9.4 Beyond the Extreme Single-Particle Model

The limitations listed above are not failures of the shell model concept — they are failures of the extreme single-particle approximation. The full shell model, which includes the residual interaction $H_{\text{res}}$ and allows configuration mixing among multiple nucleons in the valence space, is extraordinarily powerful. Large-scale shell model calculations, performed using sophisticated computer codes (NuShellX, KSHELL, ANTOINE, BigstickL), can accurately describe the spectroscopy of nuclei with up to about 15 valence nucleons, including energy levels, transition rates, moments, and beta-decay rates.

The computational challenge is combinatorial: for $n$ valence nucleons in a set of orbits with total degeneracy $\Omega$, the dimension of the many-body Hilbert space grows as $\binom{\Omega}{n}$, which can reach $10^{10}$ or more for mid-shell nuclei. This is why the single-particle shell model — despite its limitations — remains the essential first step. It provides the basis states and the energy scale; the residual interaction then mixes these states.

The modern interacting shell model — also known as the configuration-interaction (CI) shell model — is the subject of Chapter 7. But its success rests entirely on the foundation we have built here: that nucleons move in a mean-field potential, that the magic numbers arise from spin-orbit coupling, and that the single-particle orbits provide the correct starting point for a more refined treatment.

6.10 The Doubly Magic Showcase: $^{208}$Pb

No discussion of the shell model is complete without a detailed examination of $^{208}$Pb — the heaviest stable doubly magic nucleus and the single most important test case for nuclear structure models.

Shell closure: $Z = 82$ fills through $1h_{11/2}$ at the 6th magic number. $N = 126$ fills through $1i_{13/2}$ at the 7th magic number. Both proton and neutron shells are closed.

Ground state: Predicted $J^{\pi} = 0^+$. Measured: $0^+$. (Mandatory for a doubly magic even-even nucleus.)

First excited state: The lowest excitation requires promoting a nucleon across the shell gap. The proton gap (between $1h_{11/2}$ and $2f_{7/2}$) and the neutron gap (between $1i_{13/2}$ and $2g_{9/2}$) are both large — about 3.5 MeV and 3.4 MeV respectively. The actual first excited state is at 2.61 MeV ($3^-$), which is a collective octupole vibration — but even this collective state is high compared to non-magic nuclei, because the shell gaps suppress single-particle excitations.

Spherical shape: The electric quadrupole moment $Q$ of $^{208}$Pb is zero (as it must be for a $J = 0$ ground state). Even the excited states show minimal quadrupole collectivity compared to mid-shell nuclei. $^{208}$Pb is, to an excellent approximation, a perfect sphere.

Binding energy: The SEMF predicts $B(^{208}\text{Pb}) \approx 1629$ MeV. The experimental value is $B_{\text{exp}} = 1636.4$ MeV. The 7 MeV excess is the combined effect of the proton and neutron shell closures — additional binding that the smooth liquid drop formula misses.

Single-particle states in neighbors: Perhaps the most detailed confirmation of the shell model comes from the energy spectra of nuclei with one nucleon more or less than $^{208}$Pb. The nucleus $^{209}$Pb ($N = 127$) has one neutron beyond the closed shell; its low-lying states at 0 MeV ($9/2^+$), 0.779 MeV ($11/2^+$), 1.423 MeV ($15/2^-$), 1.567 MeV ($5/2^+$), 2.032 MeV ($1/2^+$), and 2.150 MeV ($3/2^+$) are cleanly identified as the single-neutron orbits $2g_{9/2}$, $1i_{11/2}$, $1j_{15/2}$, $3d_{5/2}$, $4s_{1/2}$, and $3d_{3/2}$ — the first six orbits above the $N = 126$ gap. The energy spacings between these states directly map out the single-particle level scheme for the next neutron shell.

Similarly, $^{207}$Pb ($N = 125$) has one neutron fewer than $^{208}$Pb, and its low-lying states map out the hole spectrum: the orbits of the $N = 82$-$126$ shell, seen "from below." The ground state at $1/2^-$ ($3p_{1/2}$), states at 0.570 MeV ($5/2^-$, $2f_{5/2}$), 0.898 MeV ($3/2^-$, $3p_{3/2}$), 1.633 MeV ($13/2^+$, $1i_{13/2}$), and 2.340 MeV ($7/2^-$, $2f_{7/2}$) are the single-hole states, with quantum numbers and ordering exactly as the shell model predicts. Analogous particle and hole spectra exist for $^{209}$Bi ($Z = 83$, proton particle) and $^{207}$Tl ($Z = 81$, proton hole).

Neutron skin: Because $^{208}$Pb has 44 more neutrons than protons, the neutron distribution extends slightly beyond the proton distribution, creating a "neutron skin." The thickness of this skin, measured by the PREX experiment at Jefferson Lab using parity-violating electron scattering, is $R_n - R_p = 0.283 \pm 0.071$ fm. This measurement constrains the equation of state of neutron-rich matter — a direct connection from nuclear structure to neutron stars (Chapter 25).

End of decay chains: $^{208}$Pb is the final product of the thorium decay series ($^{232}$Th $\to$ ... $\to$ $^{208}$Pb). It is also the end point of decay chains starting from $^{236}$U and $^{232}$U. The reason these three natural decay chains all terminate at $^{208}$Pb, rather than continuing to some lighter nucleus, is precisely the double shell closure: the exceptional stability conferred by $Z = 82$ and $N = 126$ makes $^{208}$Pb immune to alpha decay ($Q_\alpha < 0$), so the chain stops. The magic numbers are literally written in the radioactive genealogy of natural uranium and thorium.

$^{208}$Pb is the nuclear structure physicist's best friend. It anchors the shell model, calibrates nuclear interactions, and connects nuclear structure to astrophysics. We will return to it repeatedly throughout this book.

Summary

This chapter has taken us from an empirical puzzle — the magic numbers — to a complete theoretical framework — the nuclear shell model. The key ideas, in logical order:

1. Evidence demands shell structure. Binding energy anomalies, separation energy discontinuities, high first-excited states, stable isotope counts, and spherical shapes at $N$ or $Z = 2, 8, 20, 28, 50, 82, 126$ all point to closed-shell structure.

2. The mean-field concept works because the Pauli principle suppresses nucleon-nucleon scattering inside the nucleus, allowing independent-particle motion.

3. The harmonic oscillator gives magic numbers 2, 8, 20 — and then fails.

4. The Woods-Saxon potential improves the potential shape but still cannot produce the correct magic numbers above 20.

5. Spin-orbit coupling — the $\boldsymbol{\ell} \cdot \mathbf{s}$ interaction — splits each $\ell$-level into $j = \ell + 1/2$ (pushed down) and $j = \ell - 1/2$ (pushed up). For high $\ell$, the splitting is so large that the $j = \ell + 1/2$ state drops into the shell below, creating the observed magic numbers.

6. Ground-state predictions from the shell model — $J^{\pi}$ from the unpaired nucleon, $0^+$ for all even-even nuclei — are confirmed experimentally with remarkable accuracy near closed shells.

7. Magnetic moments (Schmidt values) are approximately correct near closed shells but deviate in mid-shell regions, signaling the need for residual interactions (Chapter 7).

Threshold concept: Shell structure emerges from independent-particle motion in a self-consistent mean-field potential with spin-orbit coupling. This single sentence encodes the essential physics of nuclear structure.

In Chapter 7, we will go beyond the single-particle picture to include the residual interaction, pairing correlations, and configuration mixing — the refinements that extend the shell model's reach from the closed-shell champions to the full breadth of the nuclear chart. We will see that the pairing interaction — which we used here only to justify the $0^+$ ground states of even-even nuclei — is itself a rich and beautiful subject, with deep connections to superconductivity in metals and superfluidity in neutron stars. And the Nilsson model, which adapts the shell model to deformed nuclei by replacing the spherical potential with an axially deformed one, will bridge the gap between single-particle motion and collective rotation, setting the stage for Chapter 8.

Chapter 6 is part of the progressive project: Nuclear Data Analysis Toolkit. See code/shell_model.py for the computational implementation and code/project-checkpoint.md for the current project status.