DAMIC-M Collaboration: Light Dark Matter Detection

• DAMIC-M Collaboration is an international scientific consortium that employs high-purity silicon CCDs to search for light dark matter with eV-scale energy thresholds.
• The experiment uses thick, fully depleted skipper CCDs capable of single-electron resolution, enabling the detection of rare ionization events from sub-GeV dark matter interactions.
• It implements innovative background suppression and calibration techniques, significantly advancing constraints on both thermal and hidden-sector dark matter models.

The DAMIC-M collaboration is an international scientific consortium focused on the direct detection of light dark matter through the application of advanced, high-purity silicon charge-coupled devices (CCDs) operated in ultra-low background environments. The experiment, based at the Laboratoire Souterrain de Modane (LSM) in France, is the next-generation extension of the DAMIC (Dark Matter In CCDs) program, intensifying efforts to search for dark matter candidates in the MeV to GeV mass range as well as probes of hidden sector scenarios. Its haLLMark innovation is the deployment of large-mass, thick "skipper" CCD arrays capable of achieving single-electron resolution and eV-scale energy thresholds, thus enabling the probing of previously inaccessible parameter space in both dark matter–electron and dark matter–nucleus scattering.

1. Scientific Motivation and Objectives

DAMIC-M is designed to directly detect the interactions of sub-GeV dark matter particles with electrons or nuclei in silicon CCDs. The primary scientific objective is the exploration of light dark matter candidates—including both thermal relic and hidden sector models—by searching for rare ionization signals with ultra-low energy deposits, down to a few electronvolts. The experiment prioritizes sensitivity to candidates for which the recoil energies are beneath the detection threshold of most established WIMP (weakly interacting massive particle) search experiments. This focus encompasses both standard WIMP–nucleus interactions and generalized dark sector models involving ultralight mediators, dark photons, asymmetric dark matter, and freeze-in/freeze-out cosmological production mechanisms (Smida, 30 Sep 2024, Collaboration et al., 18 Mar 2025, Cheek et al., 21 Jul 2025).

2. Experimental Apparatus and Skipper CCD Technology

DAMIC-M employs an array of thick, fully depleted, high-resistivity n-type silicon skipper CCDs, each typically 670 μm thick with 15×15 μm² pixels (Arnquist et al., 2022, Arnquist et al., 25 Jul 2024). The skipper readout architecture, pioneered by collaborators at Lawrence Berkeley National Laboratory, incorporates a floating gate amplifier that enables multiple non-destructive readouts of each pixel. The charge resolution improves as the number of repeated measurements increases, following

$\sigma = \frac{\sigma_{1}}{\sqrt{N_{\text{skip}}}}$

where $\sigma_{1}$ is the single-sample noise and $N_{\text{skip}}$ is the number of samples per pixel. Operationally, this enables sub-electron charge resolution (down to ≈0.07–0.2 e⁻ RMS) and eV-scale ionization thresholds, allowing the unambiguous detection of individual electron–hole pairs (Lee et al., 2020, Castello-Mor et al., 2020, Arnquist et al., 2022, Arnquist et al., 25 Jul 2024). The apparatus is housed in a cryogenic copper box cooled to approximately 130 K within a multilayer shielding structure composed of ancient lead, OFHC/electro-formed copper, and high-density polyethylene, all inside an ISO 5 cleanroom located 1700 m underground to suppress cosmic muon backgrounds (Arnquist et al., 25 Jul 2024). Radiopurity of materials is assured by rigorous HPGe, alpha counting, and ICP-MS screenings, with detailed modeling using Geant4 guiding background minimization strategies (Arnquist et al., 25 Jul 2024).

3. Data Acquisition, Subsystem Performance, and Background Characterization

The Low Background Chamber (LBC), a 20–40 g skipper CCD prototype module, has validated the detector concept and demonstrated long-term stability of noise and gain, yielding background rates of ≲7 dru in the 1–6 keV range with further declines upon upgrades (dru = 1 event kg⁻¹ day⁻¹ keV⁻¹) (Arnquist et al., 25 Jul 2024). Advanced electronics and readout boards—tested first with commercial ARC controllers and later custom DAMIC-M boards—achieved single-sample noise as low as 3 e⁻, with full system readout noise $\sim$0.1–0.2 e⁻ after multiple non-destructive measurements. The slow controls and DAQ software provide continuous, automated monitoring of environmental (pressure, temperature ±0.2 K), electronics, and vacuum system variables, aided by robust database and web interfaces (Arnquist et al., 25 Jul 2024).

Material selection, cleaning (chemical etching), and shielding have reduced detector backgrounds to levels predicted by Monte Carlo. Environmental effects, notably dark current and pixel leakage, are actively studied, and pixel clusters associated with known surface or instrumental artifacts are excluded from the fiducial data set. The ultra-low dark current (∼20 e⁻ mm⁻² d⁻¹) and operational stability are crucial for identifying genuine rare ionization events from background (Arnquist et al., 2022, Arnquist et al., 2023).

4. Analysis Methods, Signal Modeling, and Constraints on Dark Matter

The DAMIC-M analysis framework targets both event counting (single pixel, few-pixel patterns) and time-dependent signatures. Signal modeling for DM–electron scattering uses the theoretical rate:

$$\frac{dR}{dE_e} \propto \bar{\sigma}_e \int \frac{dq}{q^2} \left[ \int \frac{f(\mathbf{v})}{v} d^3v \right] |F_{\rm DM}(q)|^2 |F_c(q,E_e)|^2,$$

where $\bar{\sigma}_e$ is the DM–electron reference cross section (at $q = \alpha m_e$), $F_{\rm DM}(q)$ is the DM form factor (e.g., $(\alpha m_e/q)^n$ with $n=0$ for heavy, $n=2$ for ultralight mediator), and $F_c$ is the silicon crystal form factor (Arnquist et al., 2023, Collaboration et al., 18 Mar 2025). Background estimation is performed via detailed Monte Carlo for random coincidences and radioactive decays, and pattern recognition is optimized for various charge deposition configurations (e.g., consecutive pixel charge patterns {11}, {21}, {111}, {31}) (Collaboration et al., 18 Mar 2025). Fiducialization in z (depth) using lateral charge diffusion, likelihood clustering, and masking of instrumental defects further suppress residual contamination (Aguilar-Arevalo et al., 2023).

Daily modulation searches for MeV-mass DM exploit the time-varying DM flux caused by Earth scattering and use a binned likelihood function on the daily time series of 1e⁻ cluster rates, fitting for cosine modulations and establishing limits in the absence of a signal (Arnquist et al., 2023).

5. Experimental Results: Constraints, Excesses, and Model Exclusion

Initial scientific runs with the LBC and subsequent prototypes (total exposures of ≳1 kg·d) have excluded previously untested regions of parameter space in both heavy and ultralight mediator DM–electron interactions. Notable results include:

• Exclusion of DM–electron interactions in the mass ranges 1.6–1000 MeV/c² (ultralight mediator) and 1.5–15.1 MeV/c² (heavy mediator) at leading cross-section limits (Arnquist et al., 2023).
• Two orders of magnitude improvement in limits for DM-electron scattering via daily modulation for masses [0.53, 2.7] MeV/c² (Arnquist et al., 2023).
• Exclusion of benchmark models in which hidden-sector DM is produced by freeze-in (ultralight mediator, DM masses 3.5–490 MeV/c²) or freeze-out (heavy mediator, DM masses 2.9–21.5 MeV/c²) at the level needed to obtain the observed cosmic abundance, for both complex scalar and asymmetric fermion DM (Collaboration et al., 18 Mar 2025, Cheek et al., 21 Jul 2025).
• No observed statistically significant signal in excess of background in searches for 2e⁻, 3e⁻, or 4e⁻ clusters with all observed rates consistent with background expectations (Collaboration et al., 18 Mar 2025).
• Confirmation of a bulk, low-energy excess ("exponential spectral excess") in the CCD data at SNOLAB—characterized by decay constant $\varepsilon \sim{80}–90~\mathrm{eV}_{\rm ee}$ and rate $\sim{7}$ events/kg-day, with origins still undetermined—seen with improved depth-fiducialization and surface background rejection (Aguilar-Arevalo et al., 2023). A plausible implication is that further statistics and improved background modeling from DAMIC-M and the SENSEI experiment will be crucial for elucidating this excess.

Comparison with PandaX-4T S2-only searches shows DAMIC-M competing at or exceeding sensitivity in most of the sub-GeV region for light mediators, while PandaX-4T rivals it (or dominates) for heavy mediator scenarios at 20–200 MeV (Cheek et al., 21 Jul 2025).

6. Theoretical Impact, Complementarity, and Future Prospects

DAMIC-M results now robustly test both thermal and non-thermal dark matter production models in the sub-GeV regime, critically shrinking or eliminating large viable parameter spaces for benchmark scenarios previously considered weakly testable. For scalar DM ($J_\mu' = i\chi^*\partial_\mu\chi$) and asymmetric DM ($J_\mu' = \bar{\chi} \gamma_\mu \chi$), nearly all the cosmologically favored regions above a few MeV are excluded aside from narrow mass windows. Complementarity is highlighted with cosmological (BBN, CMB), astrophysical (stellar cooling, cosmic ray signatures), accelerator-based, and alternative direct detection approaches (e.g., nuclear recoils via the Migdal effect), which together create a tightly constrained landscape for sub-GeV DM (Cheek et al., 21 Jul 2025).

A table summarizing the comparative experimental reach:

Experiment	Mass Range (MeV/c²)	Best Probed Scenario
DAMIC-M	0.5–1000	Light mediator (ultralight)
PandaX-4T S2	20–200	Heavy mediator (scalar/asymm.)

Continued advances are anticipated as DAMIC-M completes its scale-up to ≳700 g of target mass (208 CCDs, 1 kg·yr exposure) with projected backgrounds <1 dru and sub-e⁻ readout noise. Targeted further improvements include additional reductions in dark current and backgrounds (via ultra-radiopure material handling), optimization of multi-CCD array operation, and expanded search channels (including dark photon absorption and Migdal effect signals). This suggests DAMIC-M will play a leading role over the next decade in closing the remaining windows for hidden-sector and sub-GeV dark matter candidates (Smida, 30 Sep 2024, Collaboration et al., 18 Mar 2025, Arnquist et al., 25 Jul 2024).

7. Collaboration Structure, Network, and Technical Innovation

The DAMIC-M collaboration comprises multiple international institutes, combining expertise in low-noise electronics, silicon sensor technology, underground operations, and statistical data analysis. Key technical contributions originate from LBNL (skipper amplifiers), SUBATECH/IN2P3, University of Chicago, and the Instituto de Física de Cantabria. Dedicated subgroups pursue radiopurity control, firmware/software infrastructure, calibration campaigns, and new methods for particle identification (e.g., crystalline defect tagging from nuclear recoils (Lee, 2022)). Joint developments, such as the QCDark framework, ensure robust modeling of silicon atomic and crystal effects for theoretical rate calculations (Collaboration et al., 18 Mar 2025).

The collaborative effort is also engaged in extensive cross-comparisons with parallel technologies (e.g., SENSEI, SuperCDMS) and synergistic global dark matter initiatives, directly influencing future projects (OSCURA, XLZD) aimed at closing the remaining sub-GeV parameter space.

---

DAMIC-M exemplifies the state-of-the-art in ultra-low threshold, low-background direct dark matter detection by leveraging highly engineered skipper CCD technology, advanced background control, and comprehensive analysis techniques. Its results have already excluded key hidden-sector dark matter scenarios and set industry-leading limits on the coupling of sub-GeV dark matter to electrons, with future data expected to fully explore much of the remaining viable parameter space for thermal and non-thermal dark matter models (Collaboration et al., 18 Mar 2025, Cheek et al., 21 Jul 2025).