Neutrino fog for dark matter-electron scattering experiments

The search for sub-GeV dark matter via scattering on electrons has ramped up in the last few years. Like in the case of dark matter scattering on nuclei, electron-recoil-based searches also face an ultimate background in the form of neutrinos. The so-called “neutrino fog” refers to the range of open dark matter parameter space where the background of neutrinos can potentially prevent a conclusive discovery claim of a dark matter signal from being made. In this study, we map the neutrino fog for a range of electron recoil experiments based on silicon, germanium, xenon, and argon targets. In analogy to the nuclear recoil case, we also calculate the “edge” to the neutrino fog, which can be used as a visual guide to determine where neutrinos become an important background—this boundary excludes some parts of the key theory milestones used to motivate these experiments.

Figure 1

Summary of the main result obtained in this paper—the neutrino fog “edges” for silicon and liquid xenon (LXe) experiments for the commonly used heavy (left) and light (right) DM-electron interaction models. We define the “edge of the neutrino fog” as the cross section at which the scaling of a DM discovery limit scales as $\sigma \propto (MT{)}^{-1/n}$ with $n>2$—implying slower scaling than the Poisson expectation (full visualizations of the value of $n$ as a function of both $m_{\chi }$ and ${\overline{\sigma }}_e$ are shown in Fig. 2). To provide additional context, we have shown the community’s key theory milestones for each case, i.e., models that explain the DM abundance while having the possibility of DM-electron scattering. These correspond to the window between a simple scalar-DM freeze-out scenario [6, 25, 29, 80, 81] and the ELDER model [30, 32] for the heavy-mediator case—and for the light-mediator case, the DM freeze-in scenario [1, 6]. Constraints from existing experiments are shown in gray: SENSEI [10], DAMIC-M [12, 13], CDMS-HVeV [23, 24], EDELWEISS [22], PandaX [21], DarkSide-50 [19], XENON10 [1], XENON1T-S2 [20], and XENON1T for solar-reflected DM [82, 83, 84]. Since big-bang nucleosynthesis (BBN) and the cosmic microwave background (CMB) observations are often incompatible with very-low-mass DM, we have shown an indicative example of such a bound under the assumption that DM is a Dirac fermion with a mediator 3 times its mass [85].

Figure 2

Left panels: event rates for DM (solid lines) and neutrino-induced (gray dashed line) electrons for two of the target media studied here—silicon and liquid xenon. In each case, we show two example DM masses assuming a heavy mediator, with cross sections chosen to show cases that have a stronger overlap with the neutrino background. Right panels: corresponding neutrino fogs for the two targets and under the heavy-mediator case (light-mediator cases are shown in Fig. 5). The color scale shows the value of $n$ calculated from the scaling of the DM discovery as a function of exposure, i.e., $\sigma \propto (MT{)}^{-1/n}$. We also highlight the $n=2$ contour which we showed in Fig. 1.

Figure 3

Scaling of the DM discovery limit in a silicon target as a function of exposure for three specific DM masses. We see clearly here that there is a hard neutrino floor for masses around 1.5 MeV, which is a result of the signal being contained in a single electron-number bin. Higher masses do not show this behavior as expected because the additional spectral information contained over a range of bins allows for greater background discrimination. In those two cases, instead, the neutrino fog is less severe, though still present as implied by the values of $n>2$. As a visual aid, we show three straight lines that have gradients of $\sigma \propto (\text{Exposure}{)}^{-1/n}$ with $n=1$, 2, 7. We emphasize that this plot is to illustrate the impact of the neutrino background—exposures in excess of the ton-year scale are not currently anticipated for this type of experiment.

Figure 4

Additional neutrino fog visualizations for argon and germanium targets assuming a heavy mediator. For argon, the required exposures to map the lower extent of the plot are in excess of ${10}^{15}\mathrm{kg}\text{−}\mathrm{years}$, leading to numerical instability in deriving an accurate value of $n$. In this case, we simply leave this region unfilled.

Figure 5

Additional neutrino fog visualizations for the light-mediator case. For argon, the required exposures to map the lower extent of the plot are in excess of ${10}^{15}\mathrm{kg}\text{−}\mathrm{years}$, leading to numerical instability in deriving an accurate value of $n$. In this case, we simply leave this region unfilled.

Figure 6

The $n=2$ boundaries of the neutrino fog when DM-electron scattering is calculated using DarkEFL (as utilized throughout the paper) [90], QEdark [3], and EXCEED-DM [91] for the Si target.