Earth-Scattering Induced Modulation in Low-Threshold Dark Matter Experiments

TL;DR

The paper investigates how strong DM interactions with ordinary matter can cause DM particles to scatter in the Earth, imprinting a daily sidereal modulation in low-threshold detectors targeting sub-GeV DM. It develops a dark-photon mediator benchmark and uses two transport tools, DaMaSCUS for full 3D Monte Carlo simulations and Verne as a faster analytic approach, to predict modulation patterns in silicon, xenon, and argon detectors located at SNOLAB (Northern Hemisphere) and SUPL (Southern Hemisphere). The results show that Earth-scattering can significantly modulate the flux and hence the event rate, with substantial modulation amplitudes near current experimental limits, especially for light mediators in the 1e− electron-ionization channel. These modulation signals can outperform rate-only searches when backgrounds dominate, and the authors provide open-source code and data to enable broader exploration of the MeV–GeV DM parameter space.

Abstract

Dark matter particles with sufficiently large interactions with ordinary matter can scatter in the Earth before reaching and scattering in a detector. This induces a modulation in the signal rate with a period of one sidereal day. We calculate this modulation for sub-GeV dark matter particles that interact either with a heavy or an ultralight dark-photon mediator and investigate the resulting signal in low-threshold detectors consisting of silicon, xenon, or argon targets. The scattering in the Earth is dominated by dark matter scatters off nuclei, while the signal in the detector is easiest to observe from dark matter scattering off electrons. We investigate the properties of the modulation signal and provide projections of the sensitivity of future experiments. We find that a search for a modulation signal can probe new regions of parameter space near the energy thresholds of current experiments, where the data are typically dominated by backgrounds.

Figures (33)

• Figure 1: Contours of the mean free path through the mantle $\lambda$ in units of Earth radius ($R_\oplus$) in the DM mass $m_\chi$ and DM-electron scattering cross-section $\overline\sigma_e$ parameter space, for a heavy mediator ( top) and light mediator ( bottom). For the light mediator, the halo constraints ( solid black) are taken from SENSEI:2024yytdamicmcollaboration2025probingbenchmarkmodelshiddensectorDamicModArnquist_2024 while for the heavy mediator, constraints are combined using damicmcollaboration2025probingbenchmarkmodelshiddensectorDamicModArnquist_2024SENSEI:2024yytPandaXTLi_2023XENON:2024zncDarkSide:2022knj. Constraints from the Migdal effect from PandaXTLi_2023 are shown in dotted black. The dashed black line shows the excluded region which arises from the solar reflected constraint from DM XENON:2024znc. Along the dash-dotted line, the mean free path equals $R_\oplus$.
• Figure 2: An example of an isodetection ring with a finite $\Delta \Theta$. The Earth's velocity, $\bm{v}_\oplus$, points in the direction of the Cygnus constellation.
• Figure 3: Examples of the shift in $\eta$ (defined in Eq. (\ref{['eq:eta']})) as a function of isoangle $\Theta$ for DM interacting with a heavy mediator ( left) or a light mediator ( right), for a DM mass of $m_\chi=1$ MeV, and DM-electron scattering cross-section of $\sigma_e=10^{-32}$ cm$^2$. The unshifted $\eta$ for the SHM is shown as a black, dashed line in both panels. These lines are calculated using DaMaSCUS.
• Figure 4: Left: Isoangle as a function of time for a detector at SNOLAB (blue) and SUPL ( orange) on March 8th, 2025). Right: Differential exposure as a function of isoangle for the same sites.
• Figure 5: Scattering rate for $m_\chi=1$ MeV ( left), 10 MeV ( middle), and 100 MeV ( right) in silicon for the $n_e=1$ bin for the indicated scattering cross-sections, which are near the current constraints. The blue and orange regions indicate the isoangles spanned by our two representative locations, SNOLAB and SUPL, respectively. Here the data are from DaMaSCUS simulations, while the (red line) is the fit to these data using Eq. \ref{['eq: hyptan fit']}. Results from Verne are shown with a grey dashed line.