Case Study 4.2 — Shell Effects: Where the Liquid Drop Model Breaks Down

"The five-parameter SEMF reproduces 2,500 binding energies to 1%. But if you plot the residuals, you discover the shell model staring back at you."

The Residuals Tell a Story

The semi-empirical mass formula is a smooth function of $Z$ and $A$. Nuclei do not have smooth properties — they have shell structure. The residuals $\Delta B = B_{\text{exp}} - B_{\text{SEMF}}$ are the footprints of the quantum mechanics that the liquid drop ignores.

When we plot $\Delta B$ systematically across the nuclear chart, the pattern is unmistakable: positive residuals (extra binding beyond the SEMF) cluster at the magic numbers $N$ or $Z = 2, 8, 20, 28, 50, 82, 126$. Negative residuals (less binding than the SEMF predicts) occur at mid-shell, between consecutive magic numbers. Doubly magic nuclei — those with both a magic proton number and a magic neutron number — show the largest positive residuals.

This case study examines the evidence quantitatively.

The Two-Neutron Separation Energy

The most sensitive experimental signature of shell closures is not the binding energy itself (which is dominated by the smooth SEMF trend) but the two-neutron separation energy:

$$S_{2n}(Z,N) = B(Z,N) - B(Z,N-2)$$

$S_{2n}$ measures the energy cost of removing the last pair of neutrons. (We use two-neutron rather than one-neutron separation energies to eliminate the odd-even staggering from the pairing term.) For a smooth liquid-drop nucleus, $S_{2n}$ would vary smoothly with $N$. In reality, $S_{2n}$ shows dramatic discontinuities at magic numbers.

Data for the tin isotopes ($Z = 50$), AME2020:

The drop from $S_{2n} = 10.12$ MeV at $N = 82$ to $S_{2n} = 5.14$ MeV at $N = 84$ is a decrease of nearly 5 MeV — a dramatic discontinuity. This is the $N = 82$ shell closure: it costs about 5 MeV more to remove a pair of neutrons from the closed shell ($N = 82$) than from just above it ($N = 84$). The SEMF, being a smooth function of $N$, predicts no such jump.

This same pattern repeats at every magic number. For the lead isotopes ($Z = 82$), there is a dramatic drop in $S_{2n}$ at $N = 126$. For the calcium isotopes ($Z = 20$), there are drops at $N = 20$ and $N = 28$.

Doubly Magic Nuclei: The Showcase

Doubly magic nuclei have closed shells for both protons and neutrons. They are the most "extra-bound" nuclei relative to the SEMF and display the most dramatic shell effects. The known doubly magic nuclei are:

(Note: The SEMF values depend on the exact parameters used. The residuals here are approximate and meant to illustrate the pattern.)

Several features are notable:

1. Most doubly magic nuclei are more bound than the SEMF predicts ($\Delta B > 0$). The exception is $^{40}$Ca, where the SEMF slightly overestimates the binding — this reflects the specific parameter set used and the fact that $^{40}$Ca sits at a point where the shell correction is partially absorbed into the fitted parameters.

2. $^{132}$Sn shows the largest shell effect among the well-measured doubly magic nuclei. This is because $Z = 50$ and $N = 82$ are "strong" magic numbers (the shell gaps are large), and $^{132}$Sn is far from stability ($N/Z = 1.64$), so the SEMF — which is fit primarily to stable nuclei — does not benefit from absorbing the shell correction.

3. $^{100}$Sn ($Z = N = 50$) is particularly interesting: it is the heaviest known $N = Z$ doubly magic nucleus and lies at the proton drip line. Its mass has been measured only recently (at RIKEN and GSI), and the extra binding from the double shell closure is what allows it to exist as a bound nucleus.

The Shell Correction Energy: Strutinsky's Method

In 1967, V. M. Strutinsky developed a systematic method for separating the smooth (liquid-drop) and oscillating (shell) contributions to the nuclear binding energy. The idea is elegant:

1. Calculate the single-particle energies from a mean-field potential (e.g., Woods-Saxon).
2. Sum these energies to get the total single-particle energy $E_{\text{s.p.}}$.
3. Compute a smoothed single-particle energy $\tilde{E}_{\text{s.p.}}$ by averaging over a range of energies near the Fermi surface (this removes the shell oscillations and produces a smooth function analogous to the liquid drop).
4. The shell correction is the difference: $\delta E_{\text{shell}} = E_{\text{s.p.}} - \tilde{E}_{\text{s.p.}}$.

This shell correction is negative (extra binding) at magic numbers (where the Fermi level sits in a large gap between shells) and positive (reduced binding) at mid-shell (where the Fermi level sits in the middle of a dense cluster of levels).

The total binding energy in the Strutinsky method is:

$$B = B_{\text{LDM}} + \delta E_{\text{shell}}$$

where $B_{\text{LDM}}$ is the liquid drop (SEMF) energy. This "macroscopic-microscopic" approach — using the SEMF for the smooth part and shell model single-particle energies for the oscillating part — is the foundation of modern nuclear mass models such as the Finite Range Droplet Model (FRDM) of Moller and Nix, which achieves RMS mass residuals of $\sim 0.6$ MeV over 2,000+ nuclei.

Visualizing Shell Effects: The $\Delta B$ Map

If we plot the residual $\Delta B = B_{\text{exp}} - B_{\text{SEMF}}$ as a color map on the $(N, Z)$ plane, the following features emerge:

• Ridges of positive $\Delta B$ (extra binding) along the magic numbers $N = 28, 50, 82, 126$ and $Z = 28, 50, 82$, appearing as horizontal and vertical stripes of enhanced binding.
• Peaks at intersections of magic proton and neutron numbers — the doubly magic nuclei.
• Valleys of negative $\Delta B$ (reduced binding relative to SEMF) at mid-shell: $N \approx 40, 66, 104$ and $Z \approx 40, 66$.
• Deformation regions in the rare earths ($150 \lesssim A \lesssim 190$) and actinides ($220 \lesssim A \lesssim 260$), where the residuals show a characteristic pattern of modest negative values (the deformed nuclei gain some Coulomb energy from their elongated shape, partially compensated by increased surface energy).

This map is one of the most informative visualizations in nuclear physics. It demonstrates conclusively that nuclear binding cannot be described by a smooth function of $Z$ and $A$ alone — there is quantum structure that manifests as systematic oscillations around the liquid-drop baseline.

Quantitative Patterns in the Residuals

The residuals reveal several quantitative regularities:

1. Shell gap energies. The magnitude of the discontinuity in $S_{2n}$ at a magic number measures the shell gap energy — the energy difference between the last occupied shell and the first empty shell. These gap energies are:

These are among the most important experimental quantities for constraining the nuclear mean-field potential (Chapter 6).

2. Subshell closures. In addition to the major magic numbers, the residuals show smaller discontinuities at $N$ or $Z = 14, 16, 32, 34, 40, 56, 64$ in certain regions of the chart. These "subshell closures" or "semimagic" numbers correspond to gaps between subshells within a major shell and are sensitive to the details of the single-particle potential (particularly the spin-orbit interaction and tensor force).

3. Evolution of magic numbers far from stability. One of the most exciting discoveries in modern nuclear physics is that the traditional magic numbers can change for nuclei far from stability. The $N = 20$ shell closure weakens for very neutron-rich nuclei around $^{32}$Mg (the "island of inversion"), and new magic numbers may appear at $N = 16$ (in oxygen isotopes) and $N = 34$ (in calcium isotopes). These effects are completely invisible to the SEMF, which uses the same five parameters for all nuclei.

A Concrete Example: The Calcium Isotopic Chain

The calcium isotopes ($Z = 20$) provide an exceptionally clean demonstration of shell effects in SEMF residuals. Calcium has a magic proton number, so the proton shell is closed throughout the isotopic chain. This isolates the neutron shell effects.

The two-neutron separation energies for calcium isotopes reveal multiple shell closures:

The drop from $S_{2n} = 13.69$ MeV at $N = 28$ to $S_{2n} = 8.18$ MeV at $N = 30$ — a decrease of 5.5 MeV — is the unmistakable signature of the $N = 28$ shell closure. The SEMF predicts no such discontinuity. Equally remarkable is the sharp decrease at $N = 34$, measured at RIKEN in 2013, which suggests a new subshell closure far from stability — a discovery that no macroscopic model could have anticipated.

$^{48}$Ca ($Z = 20$, $N = 28$) is a particularly important doubly magic nucleus. Despite having 8 more neutrons than protons (a large asymmetry for such a light nucleus), it is exceptionally stable — its binding energy exceeds the SEMF prediction by about 6 MeV. This extra binding from the double shell closure is the reason $^{48}$Ca exists as a stable nucleus despite its seemingly unfavorable neutron-to-proton ratio. The SEMF predicts that the most stable $A = 48$ isobar should be $^{48}$Ti ($Z = 22$), which is closer to $N = Z$. The shell model overrules this prediction.

What This Tells Us

The systematic study of SEMF residuals teaches several foundational lessons:

The SEMF is the right starting point. The smooth liquid-drop mass formula captures the dominant physics of nuclear binding. The residuals are perturbations on this smooth background — they are typically $\pm 5$ MeV on a total binding energy of hundreds to thousands of MeV.

Shell effects are real and systematic. The magic numbers are not quirks of a few specific nuclei — they are global features of nuclear structure that affect every nucleus on the chart. The regularity and reproducibility of the shell effects across different proton and neutron numbers is powerful evidence for the independent-particle (shell) model.

The failures of the SEMF define the research program. The entire field of nuclear structure theory — from the shell model (Chapter 6) through collective models (Chapter 8) to modern density functional theory and ab initio methods (Chapter 11) — can be understood as the effort to explain and predict the systematic deviations from the liquid drop mass formula. The SEMF does not solve nuclear physics; it defines the questions.

Mass models that combine macroscopic and microscopic physics are the current state of the art. The Strutinsky method, the FRDM, the Hartree-Fock-Bogoliubov approach with Skyrme or Gogny interactions, and relativistic mean-field models all share a common structure: a smooth macroscopic baseline (akin to the SEMF) plus microscopic corrections from shell structure and pairing. The best of these models achieve RMS residuals of 0.3–0.6 MeV, roughly an order of magnitude better than the SEMF alone.

The SEMF residuals are predictive, not just descriptive. Once we understand that shell closures produce extra binding, we can use the residual pattern to predict where unknown nuclei might be more stable than the SEMF suggests. This is exactly the reasoning behind the predicted "island of stability" at $Z = 114$ (or 120) and $N = 184$: these numbers are predicted to be magic by the shell model, and the extra binding from shell closure should stabilize superheavy elements against the instantaneous fission predicted by the liquid drop model. The experimental discovery of elements up to $Z = 118$ (oganesson) with lifetimes of milliseconds to seconds — far longer than the $\sim 10^{-21}$ second timescale expected for liquid-drop fission — is a dramatic confirmation that shell effects can extend nuclear existence well beyond the liquid drop limit.

Analysis Exercise

Using the Python code from semf_fit.py (this chapter's progressive project), generate the residual plot $\Delta B = B_{\text{exp}} - B_{\text{SEMF}}$ vs. neutron number $N$ for all nuclei in the AME2020 dataset. Annotate the magic numbers $N = 8, 20, 28, 50, 82, 126$. Then answer:

1. At which magic number is the shell effect (peak in $\Delta B$) largest?
2. Is the shell effect symmetric — that is, are the positive residuals at magic numbers comparable in magnitude to the negative residuals at mid-shell?
3. Can you identify the rare-earth deformation region ($60 \lesssim Z \lesssim 70$, $90 \lesssim N \lesssim 110$) in the residual pattern?

The liquid drop model is the forest. The shell effects are the trees. Both are real, and you need to see both to understand the landscape of nuclear binding.