Two experiments measured the neutron skin thickness of nuclei — the excess of neutron distribution radius over proton distribution radius — using parity-violating electron scattering. PREX-II measured lead-208 and found a thick skin (0.283 fm), implying a stiff symmetry energy that rises sharply with density. CREX measured calcium-48 and found a thin skin (0.121 fm), implying a soft symmetry energy. The two measurements, interpreted through standard energy-density functional theory, pointed in opposite directions. The nuclear physics community spent years trying to reconcile them: was one measurement wrong? Were the theoretical models inadequate? Did the contradiction signal missing physics?
Papakonstantinou (arXiv 2602.22722, February 2026) identifies the source: a hidden constraint in the theory itself. Standard energy-density functionals tie the behavior of nuclear matter at the dilute nuclear surface to its behavior near saturation density through a single parameterization. The functional describes both regimes with linked parameters — change the symmetry energy slope at saturation, and you automatically change how the functional behaves at the low densities that characterize the nuclear surface.
This linkage creates the contradiction. A stiff symmetry energy at saturation (needed for PREX-II's thick lead skin) forces a specific low-density behavior that disagrees with CREX's thin calcium skin. But the two nuclei probe different density regimes: lead-208 is large enough that its neutron skin reflects near-saturation properties; calcium-48 is smaller, and its skin is more sensitive to surface (low-density) behavior. The experiments aren't contradictory — they're probing different parts of the density dependence that the standard functional incorrectly yokes together.
When the low-density and near-saturation behaviors are allowed to vary independently — while still requiring reasonable behavior at saturation density, agreement with binding energies, and consistency with neutron-star observations — both measurements can be simultaneously reproduced. The neutron skins, the electric dipole polarizabilities, and the neutron-star mass-radius relations all come into agreement.
The tension was in the model, not the data. The experiments were measuring different things. The theory was assuming they measured the same thing.