Carbon-12 plus carbon-12 fusion powers the late stages of massive stars — the carbon-burning shell that produces neon, sodium, and magnesium before the star's core collapses. The reaction cross-section at stellar energies is unmeasurable in the laboratory: the Coulomb barrier suppresses the fusion rate by many orders of magnitude, and experiments can only reach energies well above those relevant to stellar interiors. Theory must bridge the gap.
The gap is conceptual as well as energetic. The standard picture treats certain excited states of magnesium-24 (the compound nucleus) as “molecular states” — configurations where the two carbon-12 nuclei retain their identity inside the compound nucleus, orbiting each other like a dumbbell. The molecular picture is intuitive and has guided decades of nuclear astrophysics.
Descouvemont (arXiv 2602.22763, February 2026) constructs a multichannel microscopic model that simultaneously describes carbon-carbon scattering, fusion into the alpha + neon-20 channel, and the spectroscopy of the magnesium-24 compound nucleus. The resonating group method treats all 24 nucleons explicitly, with antisymmetrization and channel coupling handled without approximation.
The model reproduces the measured cross-sections and predicts both narrow and broad resonances near the Coulomb barrier. At low energies, the cross-section decreases — supporting the fusion hindrance hypothesis, where quantum tunneling is further suppressed relative to simple barrier penetration models.
But the conceptual result is what matters. The magnesium-24 states that were classified as molecular are not molecular. They are highly mixed configurations — superpositions of carbon-carbon and alpha-neon cluster structures with no dominant component. The pure molecular state does not exist in the calculation. What exists is a mixture that was misidentified as a molecule because the experiments that probed it could only see one channel at a time.
The labels preceded the calculation. The calculation dissolved them.