First measurement of surface nuclear recoil background for argon dark matter searches

One major background in direct searches for weakly interacting massive particles (WIMPs) comes from the deposition of radon progeny on detector surfaces. A dangerous surface background is the ${}^{206}\mathrm{Pb}$ nuclear recoils produced by ${}^{210}\mathrm{Po}$ decays. In this paper, we report the first characterization of this background in liquid argon. The scintillation signal of low energy Pb recoils is measured to be highly quenched in argon, and we estimate that the 103 keV ${}^{206}\mathrm{Pb}$ recoil background will produce a signal equal to that of a $\sim 5\mathrm{keV}$ (30 keV) electron recoil (${}^{40}\mathrm{Ar}$ recoil). In addition, we demonstrate that this dangerous ${}^{210}\mathrm{Po}$ surface background can be suppressed, using pulse shape discrimination methods, by a factor of $\sim 100$ or higher, which can make argon dark matter detectors near background-free and enhance their potential for discovery of medium- and high-mass WIMPs. We also discuss the impact on other low background experiments.

Figure 1

Illustration of the single-phase liquid argon detector used in this study. The lower chamber ($\mathrm{\Phi }72\mathrm{mm}\times 48\mathrm{mm}$) hosts a Spectralon reflector cell that contains the radon sample in purified liquid argon; the upper chamber hosts a Hamamatsu R11065 photmultiplier tube (PMT). The two chambers are hermetically sealed with a quartz window. The reflector cell and the quartz window are coated with a WLS for argon light collection. An external liquid argon bath provides the needed cooling power.

Figure 2

Scintillation spectra of liquid argon excited by ${}^{214}\mathrm{Po}$ decays with (solid black) and without (dotted blue) ${\mathrm{F}}_{\text{prompt}}$ cut. The ${}^{210}\mathrm{Pb}$ recoils are observed below 100 p.e. and the $\alpha$ events are observed around 20 000 p.e. The inset figure shows the ${\mathrm{F}}_{\text{prompt}}$ distribution of ${}^{210}\mathrm{Pb}$ recoils. The ${\mathrm{F}}_{\text{prompt}}$ cut selected events within $2\sigma$ above the most probable ${\mathrm{F}}_{\text{prompt}}$ value, and a Gaussian fit to the spectrum is also shown.

Figure 3

The scintillation energy spectra of ${}^{214}\mathrm{Po}$-induced surface events measured in argon, using TPB ($\sim 0.4\mathrm{mg}/{\mathrm{cm}}^2$) and pTP ($\sim 0.3\mathrm{mg}/{\mathrm{cm}}^2$) as the WLS, respectively. The low energy events ($\sim 100$ p.e.) contain the ${}^{210}\mathrm{Pb}$ signals in argon and the $\alpha$ signals in WLS; the high energy events ($\sim 20,000$ p.e.) are dominated by $\alpha$ signals in argon. The inset shows the ${\mathrm{F}}_{\text{prompt}}$ distribution ($15\mu \mathrm{s}$ maximum integral window) for the low energy surface events. For comparison, the 50% ${\mathrm{F}}_{\text{prompt}}$ value for ${}^{40}\mathrm{Ar}$ is 0.69 at 88 p.e. (the TPB measurement) and 0.72 at 192 p.e. (the pTP measurement) [5].