Quasi-Elastic Spectator Recoils from Atmospheric Neutrinos and the LUX-ZEPLIN 248 keV Event

The LUX-ZEPLIN (LZ) collaboration has reported a single event consistent with a xenon nuclear recoil of 248 ± 23 (stat) ± 23 (sys) keV in a 2.84 tonne-year exposure, in tension with its background-only model at 2.6σ global significance. The published background model treats atmospheric neutrinos only through coherent elastic neutrino–nucleus scattering (CEνNS), which cannot populate this energy: the momentum transfer is q ∼ 246 MeV/c, where the Helm form factor gives |F(q)|² ≈ 1.6×10⁻⁴. We examine the incoherent channel: neutral-current quasi-elastic (NCQE) knockout of a neutron by an atmospheric neutrino of Eν ≳ 123 MeV, in which the residual nucleus carries a spectator recoil set by the struck neutron's intranuclear momentum. Because the event topology requires the residual to be left with no prompt de-excitation, the relevant knockout channels are those from the 2d3/2, 3s1/2, and 1h11/2 valence orbitals — the ground-state and long-lived-isomer channels of the A−1 xenon isotopes. A harmonic-oscillator shell calculation of these exclusive channels gives a recoil spectrum peaking at 189 keV with mean 192 keV; the observed event lies on the populated falling shoulder at 75% of peak height (74th percentile), with 16% of stealthy spectator recoils falling within ±1σ of the observed energy. The rate, however, is small: two independent anchors — a direct flux-times-cross-section estimate and a scaling of the measured Super-Kamiokande atmospheric event rate, agreeing to ~20% — give μ ≈ 5×10⁻⁵ stealthy events (bracket 3×10⁻⁶–6×10⁻⁴) in the WS2024 exposure, a likelihood ratio of roughly 1:190 against this channel relative to the experiment's modeled rare backgrounds in the event's S1c slice. We therefore do not attribute the event to this channel. We present it as an unmodeled background whose exclusive spectral peak coincides with the observed energy region, quantify its discriminators against the favored heavy-WIMP effective-field-theory interpretations, and recommend that generator-level ν–Xe NCQE simulation be added to extended-window analyses at LZ, XENONnT, PandaX-4T, and future XLZD-scale exposures, where this channel enters the zero-background budget. This is a standalone Standard-Model note. The record includes the note (PDF and markdown source), the gate script verify_lz248_ncqe_2026-09-02.py that recomputes every numerical claim (23 gates), the figure and its generator, and the adversarial cross-model review record.