Search for low-mass electron-recoil dark matter using a single-charge sensitive SuperCDMS-HVeV Detector

Using a single-charge sensitive SuperCDMS-HVeV detector operated underground at Fermilab with minimized background luminescence, this study presents new constraints on low-mass dark matter electron scattering and absorption interactions, achieving sensitivity down to 1 MeV/c² for scattering and probing axionlike particles and dark photons with masses greater than 1.2 eV/c².

The Gist

Imagine the universe is filled with invisible "ghosts" called Dark Matter. Scientists have been hunting these ghosts for decades, but they've mostly been looking for big, heavy ghosts (like bowling balls). Recently, however, physicists started wondering: What if the ghosts are actually tiny, like dust motes or even smaller?

This paper is about a team of scientists (the SuperCDMS Collaboration) who built a super-sensitive "ghost trap" to catch these tiny, low-mass dark matter particles. Here is the story of their hunt, explained simply.

1. The Ghost Hunter's Trap: The SuperCDMS Detector

Imagine you have a block of pure silicon (like a super-clean computer chip) that is colder than outer space (almost absolute zero). This block is so sensitive that if a single electron (a tiny particle of electricity) bumps into it, the block shivers.

To make this shiver easier to see, the scientists applied a high voltage (like a strong electrical wind) across the silicon. When a dark matter particle hits an electron, it creates a tiny spark. Because of that electrical wind, that tiny spark creates a cascade of sound waves (called phonons) inside the crystal. It's like a snowball rolling down a hill; a tiny push creates a huge avalanche of sound that the detector can hear.

2. The Problem: The "Hum" of the Machine

In previous experiments, the scientists had a problem. The machine holding the silicon crystals had some plastic circuit boards (PCBs) nearby. These boards were like tiny, glowing fireflies that were giving off faint light. This light was hitting the detector and creating "fake ghosts" (noise), making it hard to tell if a real dark matter particle had arrived.

The Fix: For this new experiment (called Run 4), they redesigned the trap. They built a new holder out of copper and removed almost all the plastic circuit boards. It's like taking a noisy, flickering lightbulb out of a dark room so you can finally hear a pin drop. This made the detector incredibly quiet.

3. The Hunt: Waiting in the Dark

The detector was placed deep underground (in a facility called NEXUS at Fermilab) to shield it from cosmic rays and other background noise. They waited for 6.1 gram-days of data.

• Analogy: Imagine holding a very sensitive microphone in a quiet library for a few days, waiting to hear a specific whisper.

They used a "Blind Analysis" strategy. This means they put a digital curtain over 70% of their data while they were setting up their rules. They couldn't peek at the results until their rules were perfect. This prevents them from accidentally tweaking the rules to make the results look better than they are (a bit like not looking at the scoreboard until the game is over).

4. The Calibrations: Tuning the Radio

Before they could listen for ghosts, they had to tune their radio. They used tiny LEDs (tiny lights) and a radioactive source to send known signals into the detector.

• They checked: "If we send a signal of exactly this size, does the detector hear it correctly?"
• They also checked for "crosstalk," which is like when two people talking on a walkie-talkie accidentally hear each other's static. They figured out how to subtract that static so they only heard the real signals.

5. The Results: No Ghosts Found (Yet)

After unblinding the data, they looked at the spectrum (the list of all the "shivers" they heard).

• The Good News: The background noise (the fake ghosts) dropped by 100 times compared to their previous experiment. The new copper holder worked perfectly!
• The Bad News: They didn't find any dark matter particles. The "shivers" they heard were all just random noise or known background events.

However, in science, "not finding it" is still a huge victory because it tells us where it isn't.

6. What This Means for the Future

Because they didn't find the ghosts, they drew a new "Exclusion Zone" on the map.

• Analogy: Imagine you are looking for a lost coin in a field. You didn't find it, but you checked the whole north side of the field very carefully. Now, you know the coin isn't in the north. You have to look elsewhere.

This paper sets new, stricter limits on how heavy these tiny dark matter particles can be and how strongly they interact with normal matter.

• They ruled out dark matter particles as light as 1 MeV/c² (which is still incredibly light, about 2,000 times lighter than a proton).
• They also set limits on other weird particles like "Dark Photons" and "Axions."

Summary

The SuperCDMS team built a quieter, more sensitive trap to catch the universe's smallest, lightest dark matter particles. They successfully silenced the noise of their own equipment, waited patiently in the dark, and found no ghosts. But by proving exactly where the ghosts aren't, they have narrowed the search for the next generation of scientists, bringing us one step closer to solving the mystery of what the universe is made of.

Technical Summary

1. Problem Statement

Direct detection experiments have strongly constrained the parameter space for dark matter (DM) candidates with masses above the GeV scale. Consequently, there is a growing focus on sub-GeV dark matter, including:

• MeV-scale fermions ($\chi$): Which interact via electron scattering.
• keV-scale bosons: Such as Dark Photons (DPs) and Axion-like Particles (ALPs), which interact via absorption processes.

Detecting these light particles requires detectors capable of measuring extremely low energy depositions (O(10) eV). Previous SuperCDMS High-Voltage eV-resolution (HVeV) searches were limited by background events, specifically luminescence from Printed Circuit Boards (PCBs) in the detector housing, which obscured low-energy signals. This paper addresses the need for improved background rejection and sensitivity to set stringent exclusion limits on low-mass DM interactions.

2. Methodology

Experimental Setup

• Location: Data was collected underground at the NEXUS facility at Fermilab, with an overburden of 225 meters water equivalent.
• Detectors: Four high-purity silicon HVeV detectors (NFC1, NFC2, NFE, NFH), each weighing 0.93 g.
  • Design: The detectors utilize Quasiparticle-trap-assisted Electrothermal-feedback Transition Edge Sensors (QETs) patterned on the silicon crystal.
  • Readout: Two independent channels (inner and outer) allow for position reconstruction and background rejection (vetoing multi-scatter events).
  • Bias: A bias voltage ($V_{bias} = 100$ V) is applied to induce the Neganov-Trofimov-Luke (NTL) effect, amplifying the phonon signal generated by electron-hole (eh) pairs.
• Upgrades:
  • Detector Housing: A new copper holder was designed to minimize exposed PCB area, effectively eliminating the luminescence background observed in previous runs.
  • Shielding: An external 10 cm lead castle was added to suppress ambient gamma flux from $^{40}$K and U/Th chains by nearly two orders of magnitude.
  • Environment: The system operates at 11 mK in a cryogen-free dilution refrigerator with a two-layer magnetic shield.

Data Collection & Calibration

• Exposure: A blind analysis was performed on 6.1 g $\cdot$ days of exposure.
• Blinding Strategy: 70% of the data was blinded during analysis development. Only the first 30% was used to finalize cuts and select the optimal energy peak for the final limit.
• Calibration:
  • LEDs: 630 nm LEDs were used for precise energy calibration (1.97 eV photons).
  • Gamma Source: A $^{137}$Cs source provided continuous spectrum calibration.
  • Modeling: The analysis models Charge Trapping (CT) and Impact Ionization (II) effects, which distort the energy spectrum. A two-step fitting process was used to extract calibration parameters and CT/II fractions.
  • Systematics: Uncertainties from cross-talk subtraction, calibration function assumptions, and gain corrections were quantified (total energy scale uncertainty $\lesssim$ 1% up to the 4th eh-pair peak).

Data Selection

• Live-time Cuts: Applied to remove periods of unstable temperature, baseline drifts, elevated trigger rates (likely electronics noise), and coincident events. These removed ~19% of raw exposure.
• Event Reconstruction: An Optimal Filter (OF) amplitude was used as the primary energy estimator. A reduced-$\chi^2$ cut ensured pulse shape consistency, yielding a signal efficiency of $\epsilon_{\chi^2} = 0.95 \pm 0.02$.
• Region of Interest (ROI): 85 eV to 500 eV, covering the first four eh-pair peaks.

Signal Models

The analysis tested four hypotheses assuming a local DM halo density of 0.3 GeV/cm$^3$:

1. DM-electron scattering ($\chi$): With heavy or light mediators (setting limits on cross-section $\sigma_e$).
2. Dark Photon Absorption: Setting limits on kinetic mixing parameter $\epsilon$.
3. ALP Absorption: Setting limits on axioelectric coupling $g_{ae}$.

3. Key Contributions

• Background Suppression: Successfully eliminated the PCB luminescence background that plagued previous HVeV runs, reducing the background rate in the Region of Interest by at least two orders of magnitude.
• Improved Calibration: Developed a robust calibration framework using LED data to model charge trapping and impact ionization, allowing for precise energy reconstruction down to the single electron-hole pair level.
• Blind Analysis: Executed a rigorous blind analysis strategy to prevent bias in the selection of the optimal energy peak and the definition of exclusion limits.
• Earth Shielding Correction: Incorporated Monte Carlo simulations to account for the attenuation of low-mass DM by the Earth's overburden, providing more accurate exclusion limits for strongly interacting particles.

4. Results

• Observation: The data showed a dominant number of events in the first eh-pair peak, with very few events in higher-order peaks. No distinct peak features were observed above the second peak (~200 eV).
• Exclusion Limits (90% Confidence Level):
  • DM-Electron Scattering: Set world-leading constraints on the scattering cross-section $\sigma_e$ for DM masses as low as 1 MeV/c$^2$.
  • Dark Photons & ALPs: Set limits on the photon-dark photon mixing parameter ($\epsilon$) and the ALP-electron coupling constant ($g_{ae}$) for particle masses as low as 1.2 eV/c$^2$.
• Comparison: The new limits significantly improve upon previous SuperCDMS HVeV results (Run 3) and are competitive with or superior to other leading experiments (SENSEI, EDELWEISS, PandaX-II, XENONnT) in the sub-GeV mass range.

5. Significance

This work represents a major milestone in the search for low-mass dark matter. By successfully mitigating the PCB luminescence background, the SuperCDMS HVeV collaboration has demonstrated the capability of cryogenic silicon detectors to probe the MeV and sub-MeV mass regimes with unprecedented sensitivity. The results rule out significant portions of the parameter space for light dark matter candidates, guiding future theoretical models and experimental designs. The paper also highlights the ongoing challenge of understanding low-energy excess (LEE) backgrounds, which remain a focus for future runs (e.g., at SNOLAB) to further push sensitivity limits.