Toward a neutrino-limited dark matter search with crystalline xenon

Experiments searching for weakly interacting massive particle dark matter are now detecting background events from solar neutrino-electron scattering. However, the dominant radioactive background in state-of-the-art experiments such as LZ and XENONnT is beta decays from radon contamination. In spite of careful detector material screening, radon progenitor atoms are ubiquitous and long-lived, and radon is extremely soluble in liquid xenon. We propose a change of phase and demonstrate that crystalline xenon offers more than a factor $\times 500$ exclusion against radon ingress, compared with the liquid state. This level of radon exclusion would allow crystallized versions of existing experiments to probe spin-independent cross sections near ${10}^{-47}{\mathrm{cm}}^2$ in roughly 11 years, as opposed to the 35 years required otherwise.

Figure 1

Schematic of the experimental apparatus, approximately to scale. Radon atoms (indicated) were introduced via the sealed xenon gas circulation loop.

Figure 2

Top: distribution of scintillation (S1) and ionization (S2) signals from single scatter events in the measurement region, after circulating ${}^{222}\mathrm{Rn}$ in liquid/vapor mode for 210 hours. Bottom: the same distribution in crystal/vapor mode, about 45 hours after crystallizing. Inset plots show the reconstructed energy spectra of the bulk radon events. The population of events with $\mathrm{S}2\sim 200$ phd are due to ${}^{210}\mathrm{Po}$ source decays at the edges of the disk.

Figure 3

Count rate of alpha decays in the measurement region during continuous introduction of ${}^{222}\mathrm{Rn}$. The count rate initially increased in proportion to $1-e^{-\lambda t}$, with $\lambda =\log (2)/t_{1/2}$ and $t_{1/2}=92$ hours. In crystal/vapor mode the count rate decreased as $e^{-\lambda t}$, indicating that no new radon could enter the crystal. Arrows indicate the two data sets shown in Fig. 2.

Figure 4

Left: distribution of scintillation (S1) and ionization (S2) signals from single scatter events in the measurement region during four hours of continuously circulating ${}^{220}\mathrm{Rn}$ in liquid/vapor mode. Center: the same distributions obtained during four hours of continuously circulating ${}^{220}\mathrm{Rn}$ in crystal/vapor mode. Right: same as center, but with cathode events removed for clarity. Three candidate alpha events in the crystal are observed. The population with $\mathrm{S}2\sim 1000$ is due to radon decays in the vapor above the anode (short electron drift times) and to the tail of ${}^{210}\mathrm{Po}$ source decays (long electron drift times).

Figure 5

The projected sensitivity of an LZ-like detector to spin-independent scattering of 1 TeV dark matter particles, in the case of either 5.5 tonne active liquid xenon mass (filled circles) or 6.6 tonne active crystal xenon mass (open circles). Also shown are the recent LZ first results [7], LZ projection [20], and projected sensitivity of a 200 tonne-year exposure of a next generation experiment [26] scaled to a 5.5 tonne target.