The authors build a simple particle-physics model that can account for the single nuclear-recoil‑like event reported by the LZ experiment near 248 keV. Their idea is a form of “inelastic” dark matter: the dark particle must jump to a slightly heavier state when it scatters, and the amount of that mass gap controls the recoil energy. By arranging a particular mixture of two hypothetical dark states, the model reduces how often such scatters happen without changing the expected recoil energy.

Concretely, the model pairs a Majorana singlet (a neutral fermion that is its own antiparticle) with a vector-like electroweak doublet (a pair of charged and neutral states that transform like known weak-force particles). At the mass splitting implied by LZ’s 248 keV recoil, a standard-strength interaction through the Z boson would have produced thousands of events. The key move is singlet–doublet mixing: it weakens the Z-mediated transition enough to bring the predicted count down to about one event while leaving the recoil spectrum in place.

Other simple explanations that keep the interaction strength fixed need a larger mass splitting so only the fastest dark-matter particles in the galactic halo can scatter. That approach is partly rescued by the Sun’s gravity, which speeds up incoming particles and reduces the suppression. But a larger splitting also pushes recoils to higher energies, into an energy range where LZ saw no events, so those fixed-coupling scenarios have tension with the observed spectrum in a way the mixing model avoids.

The same mixing that reduces scattering also reduces how efficiently dark matter is captured and annihilated inside the Sun. In addition, the small mass gap required for coannihilation in the early universe weakens Higgs-mediated cooling of captured particles, so the captured population can remain dilute and out of equilibrium rather than form a dense core. Imposing the requirement that the model reproduce the observed dark-matter abundance and normalizing the direct-detection rate to one event leaves a two-dimensional space of allowed dark-matter mass and mass splitting. Under different assumptions about how captured particles cool, the paper quotes illustrative limits for the splitting: a thermalized population gives a conservative IceCube neutrino constraint at δ ≈ 300–301 keV, while a nonthermal cooling assumption with tree-level elastic scattering moves the nominal limit to δ ≈ 331–341 keV near a relic-density endpoint around 730–733 GeV — a mass range that the recoil spectrum also prefers.