Low-Energy Physics Reach of Xenon Detectors for Nuclear-Recoil-Based Dark Matter and Neutrino Experiments

Dual-phase xenon detectors lead the search for keV-scale nuclear recoil signals expected from the scattering of weakly interacting massive particle (WIMP) dark matter, and can potentially be used to study the coherent nuclear scattering of MeV-scale neutrinos. New capabilities of such experiments can be enabled by extending their nuclear recoil searches down to the lowest measurable energy. The response of the liquid xenon target medium to nuclear recoils, however, is not well characterized below a few keV, leading to large uncertainties in projected sensitivities. In this work, we report a new measurement of ionization signals from nuclear recoils in liquid xenon down to the lowest energy reported to date. At 0.3 keV, we find that the average recoil produces approximately one ionization electron; this is the first measurement of nuclear recoil signals at the single-ionization-electron level, approaching the physical limit of liquid xenon ionization detectors. We discuss the implications of these measurements on the physics reach of xenon detectors for nuclear-recoil-based WIMP dark matter searches and the detection of coherent elastic neutrino-nucleus scattering.

Figure 1

Diagram of experimental setup (not to scale). The collimator and shielding are color coded to show the borated polyethylene (blue), high-density polyethlyene (light gray), and lead (dark gray). The xenon TPC is shown in green, and the LS backing detectors are shown in yellow. We utilized two sizes for the backing detectors (2 and 4 in. diameter), with relative placements shown here.

Figure 2

Background-subtracted and efficiency-corrected nuclear recoil ionization spectra measured at a drift field of $220\mathrm{V}/\mathrm{cm}$ (black), at the three lowest energies explored in this work. We clearly observe quantized single-, double-, and triple-electron signals. The spectra in each channel are fitted with simulated spectra (blue), which include the expected neutron multiple-scattering background (gray), allowing us to extract the average number of ionization electrons as a function of recoil energy. We show the scattering angle, the mean and central 68 percentile of the simulated nuclear recoil energy distribution, and the average number of ionized electrons obtained from the fit.

Figure 3

Average number of ionization electrons produced in liquid xenon as a function of nuclear recoil energy (using electron extraction efficiency values measured in Ref. [30]). For comparison, we show measurements from the LUX dark matter experiment [28], as well as the prediction from the nest simulation package [34], which is based on evaluations of published data [23, 24, 25, 26, 27, 35, 36] and is driven by the LUX measurements at low energies. The inset shows the expected recoil spectra from reactor antineutrinos $\mathrm{CE}\nu \mathrm{NS}$ and dark matter WIMP models with different masses for comparison. Here we assume the standard halo model (Ref. [41]) and a spin-independent interaction cross section (denoted ${\sigma }_0$) of ${10}^{-41}{\mathrm{cm}}^2$ for the WIMP spectra, and use the parametrizations in Ref. [42] for the reactor antineutrino emission spectrum.

Figure 4

Left—Projected ionization spectra from a $3\mathrm{GeV}/c^2$ WIMP (assuming ${\sigma }_0={10}^{-41}{\mathrm{cm}}^2$) and reactor antineutrino coherent scattering. Projections calibrated by the measurements in this work (solid lines) deviate significantly from those using nestv2.0 (dashed), which is tuned to fit existing data. Right—Projected sensitivity of a background-free, ionization-only search for low-mass WIMP dark matter (dashed red). The assumed exposure is 10 kg yr. Solid lines show existing limits from CDMSLite [44], Darkside50 [45], CRESSTII [46], XENON1T [2], and LUX [1] for comparison. Uncertainty bands in both the spectra and projected limits are calculated by propagating the uncertainties in the second-order polynomial fit parameters.