Cryogenic Detectors

• Cryogenic detectors are ultra-sensitive solid-state sensors operating below 100 mK that convert minimal energy depositions into measurable thermal, ionization, or scintillation signals.
• They employ advanced sensor technologies such as TES, NTD-Ge, MKID, and MMC to achieve eV-scale energy resolutions and sub-keV detection thresholds critical for dark matter and neutrino experiments.
• Hybrid readout modes combined with meticulous calibration and background rejection techniques enable near-noise-free detection of rare events in demanding experimental environments.

Cryogenic detectors are solid-state sensors operated at temperatures typically below 100 mK, designed to register extremely faint signals by transducing the energy depositions of rare particles or photons into measurable phonon, ionization, or scintillation signals. These detectors exploit low heat capacities and the temperature sensitivity of certain sensor materials to achieve eV to sub-eV energy thresholds, high background rejection, and energy resolutions rivaling or surpassing those of semiconductor detectors. Dominant experimental applications include direct searches for dark matter (WIMPs, sub-GeV DM, dark photons), coherent elastic neutrino-nucleus scattering (CEνNS), neutrinoless double-beta decay (0νββ), solar and supernova neutrino detection, and explorations of new physics such as axions or hidden-sector particles (Das et al., 2 Dec 2025).

1. Fundamental Detection Principles

Cryogenic detectors function by converting particle-induced energy into one or more measurable channels: thermal phonons, electron–hole pairs (ionization), and/or scintillation photons. The foundational calorimetric principle is that an energy deposition $E$ in an absorber at temperature $T$ causes a temperature rise $\Delta T = E / C(T)$, with heat capacity $C(T)\propto T^3$ at mK for dielectric crystals. Ultralow $C$ enables efficient detection of O(eV–keV) depositions. Three canonical readout modes are distinguished (Armengaud, 2010, Das et al., 2 Dec 2025):

• Phonon (Heat) Detection: All deposited energy is ultimately thermalized as phonons. A phonon sensor (NTD Ge, TES, MKID, MMC) transduces $\Delta T$ to a voltage or current pulse. The resolution is fundamentally bounded by thermal fluctuation noise,

$\sigma_E \propto \sqrt{k_B T^2 C}\, ,$

where $k_B$ is Boltzmann's constant. TES and MKID architectures enable sub-10 eV thresholds (Armengaud, 2010, Das et al., 2 Dec 2025).

• Ionization Detection: Part of the recoil energy creates $N_{e\text{–}h}=E_\text{ion}/\epsilon_\text{pair}$ electron–hole pairs ($\epsilon_\text{pair}\approx3$ eV in Ge at low $T$). Cryogenic voltages drift the charges to readout electrodes, where the current is amplified. Ionization yield ($Q=E_\text{ion}/E_\text{recoil}$) provides particle ID (nuclear recoils $Q\sim0.3$, electron recoils $Q\sim1$) (Armengaud, 2010, Mast et al., 2018).
• Scintillation Detection: Electronic excitation in the absorber emits optical/UV photons. At cryogenic $T$, light is detected by a secondary phonon calorimeter or photon sensor (TES, NTD, PMT, MCP), directly measuring individual photons and enabling event-by-event discrimination (Angloher et al., 2016, Zhang et al., 2020).

In hybrid mode, multiple channels are measured simultaneously, allowing powerful background rejection. For instance, EDELWEISS-II uses simultaneous phonon and ionization channels, enhancing nuclear/electron-recoil discrimination with $99.99\%$ efficacy above threshold (Armengaud, 2010). CRESST-class detectors use heat–light dual measurement to reach $>10^4$ rejection of $\alpha$ and nuclear recoils (Angloher et al., 2016).

2. Sensor Technologies, Readout, and Operation

Cryogenic detectors employ a spectrum of sensor technologies, optimized for specific performance domains (Das et al., 2 Dec 2025, Angloher et al., 2016, Shah et al., 2021):

Sensor Type	$T_\mathrm{op}$ [mK]	Resolution [eV]	Threshold [eV]	Remarks
NTD-Ge	$10-20$	$50-200$	$100$	Used in CUORE, CUORICINO
TES	$10-20$	$1-10$	$10$	CRESST, SuperCDMS, ALPS-II
MKID	$100-1000$	$10-100$	$10$ (R&D)	High multiplexing, R&D
MMC	$10$	$20-50$	$20$	AMoRE, MARE projects

• NTD Ge thermistors: Resistance $R(T)=R_0\exp[(T_0/T)^{1/2}]$ enables low-power, low-noise temperature sensing (Das et al., 2 Dec 2025, Collaboration et al., 2021).
• Transition-Edge Sensors (TES): Voltage-biased thin-film (e.g., W, Ir/Au), operate at the steep part of the superconducting transition. Under negative electrothermal feedback,

$$\Delta E \approx 2.355 \sqrt{\frac{4k_BT^2C}{\alpha}},$$

where $\alpha=(T/R)(dR/dT)$ is the transition sharpness. Typical $\Delta E = 0.1-1$ eV at sub-100 mK (Shah et al., 2021, D'Andrea et al., 19 Jan 2024).

• MKID: Resonant shift due to kinetic inductance changes from Cooper-pair breaking by phonons; mass multiplexing possible up to $10^3$ pixels/channel (Das et al., 2 Dec 2025).
• MMC: Paramagnetic sensor in field; $\Delta T$ read via SQUID; employed for high-mass bolometry (Das et al., 2 Dec 2025).
• Low-noise HEMT preamplifiers: Enable sub-20 eV$_{\rm ee}$ noise for ionization readout in Si/Ge crystals (Juillard et al., 2019).

Hybrid sensors with interleaved electrodes (EDELWEISS) or full coverage (FID, CRESST-III) enhance fiducialization and surface-event rejection (Armengaud, 2010, Das et al., 2 Dec 2025). Contact-free electrode techniques allow kg-scale Si detectors with sub-pA leakage and full charge collection at O(10 V/cm) (Mast et al., 2018). Heat channel signals can be amplified via Neganov–Trofimov–Luke (NTL) effect: conducting charge under bias generates additional phonons,

$$E_\text{heat} = E_\text{recoil} + eV N_{e\text{–}h},$$

enabling further threshold reduction (Defay et al., 2015).

3. Detector Architectures and Operating Modes

Cryogenic detectors are realized in a diverse set of module architectures, tailored for distinct detection channels and physics goals (Angloher et al., 2016, Su et al., 22 Sep 2025, Ding, 29 May 2024):

• Cryogenic Solid-State Calorimeters: Bulk absorbers (Ge, Si, TeO$_2$, Li$_2$MoO$_4$, CaWO$_4$, CsI) at O(10) mK, instrumented with one or more phonon and/or ionization sensors. Masses from $\sim$10 g to tonne-scale (CUORE: 988 TeO$_2$ crystals, total 742 kg) are routine (Collaboration et al., 2021).
• Cryogenic Scintillating Calorimeters: Simultaneous heat and light readout for particle discrimination (e.g., undoped CsI: $8-8.1\%$ of energy converted to light). Modular structure with TES phonon sensor on a radiopure carrier and independent light absorber with TES or NTD-Ge (Angloher et al., 2016, Su et al., 22 Sep 2025).
• Hybrid CPSD with Low-T PMTs: Noncontact PMT readout (at 20 K) for single-photon sensitivity enables sub-10 ns timing and modular scalability (Zhang et al., 2020, Ding, 29 May 2024).
• Micropattern/CRADs: Noble gas/condensed-phase detectors employing hole-type MPGDs (GEM/THGEM) for avalanche gain, coupled with SiPM or GAPD optical readout for sub-e$^-$ threshold; used in two-phase Ar/Xe/Ne detectors (Buzulutskov, 2011).
• Light-Detection Arrays (CSC): Multi-pixel, low-mass SOS wafers with TES, achieving $1~{\rm eV}$ baseline (CRESST-III), proposed for direct dark matter–electron interaction searches (Zema et al., 2 Feb 2024).

Novel large-volume pure CsI calorimeters (3.3 kg each) achieve $\sim$29 photoelectrons/keV$_{ee}$ at 95 K with FWHM 7–8% at 60 keV, stable to better than $0.3\%$ over a month, and with upper bounds $<3$ mBq/kg on $^{134}$Cs/ $^{137}$Cs activity (Su et al., 22 Sep 2025, Ding, 29 May 2024). Geant4 simulations are validated to $<5\%$ with these geometries.

4. Background Rejection, Calibration, and Data Analysis

Cryogenic detectors achieve sub-keV thresholds and near-background-free operation through engineering, analysis cuts, and software pipelines (Armengaud, 2010, Wagner et al., 2022, Angloher et al., 2016):

• Event Fiducialization: Interleaved or wraparound electrodes define fiducial bulk. Surface $\alpha/\beta$ events are vetoed using charge-sharing and pulse shape (Armengaud, 2010).
• Dual-Channel Discrimination: Simultaneous heat-light or heat-ionization readout distinguishes NR/ER with $\geq99.99\%$ efficiency above threshold; discrimination leakage between recoil bands can be reduced below $10^{-4}$ in $LY$ (Angloher et al., 2016).
• Pulse Shape and Optimum Filtering: Mutually orthogonal time constants for athermal/thermal components; matched filter kernels $H(\omega)\propto S^*(\omega)/N(\omega)$ maximize SNR for low-energy triggers (Wagner et al., 2022).
• Calibration: Onboard heaters inject known energy; external gamma/neutron sources set channels. Internal cosmogenic lines (e.g., 9–10 keV) calibrate fiducial mass; heater-pulse calibration is accurate to $<3\%$ (D'Andrea et al., 19 Jan 2024).
• Data Processing Pipelines: Tools such as Cait automate triggering, template fitting, background rejection, machine-learning-based artifact classification, and efficiency estimation (Wagner et al., 2022). ML-masked 1D CNN classifiers yield $>99\%$ nuclear-recoil retention above threshold and $>$98% artifact rejection.
• Active Shielding: Multi-layer Cu/Pb enclosures, active muon vetoes (up to 98% coverage), polyethylene for neutron moderation, and radon-suppressed environments (e.g., cleanroom assembly) reduce environmental backgrounds (Collaboration et al., 2021).
• Dynamic Coincidence/Veto: CryoAC in ATHENA X-IFU uses simultaneous detection with two TES microcalorimeters to tag muon vs photon/electron events, providing $>50\times$ background reduction (D'Andrea et al., 19 Jan 2024).

5. Performance Metrics and Key Experimental Results

Recent results from major experiments and R&D demonstrators define the state-of-the-art (Armengaud, 2010, Das et al., 2 Dec 2025, Angloher et al., 2016, Su et al., 22 Sep 2025, Ding, 29 May 2024):

• Energy Resolution & Threshold:
  • Heat (phonon): FWHM $\sim$1 keV at sub-100 mK (EDELWEISS-II), $0.1$–$0.9$ keV at CRESST, $1.4$ keV threshold for ATHENA CryoAC (Armengaud, 2010, D'Andrea et al., 19 Jan 2024).
  • Ionization: $\sim0.9$ keV in Ge at 20 mK (Armengaud, 2010); Si achieves $<500$ eV (FWHM) at kg scale (Mast et al., 2018).
  • Light: $29.3 \pm 1.0$ PE/keV$_{ee}$ and $7.7\%$ FWHM at 59.6 keV in 3.3 kg CsI (Su et al., 22 Sep 2025); $\sim$50 PE/keV$_{ee}$ at 77 K (CryoCsI prototype) (Ding, 29 May 2024).
• Background Rejection:
  • Surface event suppression to $10^{-5}$ (EDELWEISS ID), dual-channel $>10^4$ leakage suppression at keV energies (CsI) (Armengaud, 2010, Angloher et al., 2016).
  • Light yield and quenching factor calibrations: CsI shows QF$\sim$0.1–0.2 for NR; neutron beam measured $QF\sim15\%$ at 10 keV$_{nr}$ (Angloher et al., 2016, Ding, 29 May 2024).
• Stability and Scalability:
  • Month-long runs at $\leq 0.5$ K drift in temperature or gain (CsI). No detectable signal degradation after $10^4\times$ photon flux (TES-based NTL Si) (Su et al., 22 Sep 2025, Defay et al., 2015).
  • Tonne-scale operation at 10 mK (CUORE), $7.8$ keV energy resolution at $2615$ keV for $742$ kg of TeO$_2$ (Collaboration et al., 2021).
• Sensitivity and Limits:
  • EDELWEISS-II: WIMP-nucleon cross-section sensitivity of $5 \times 10^{-8}$ pb at $m_\chi=80$ GeV, 322 kg-days exposure (Armengaud, 2010).
  • CRESST-III: $<30$ eV$_{nr}$ threshold, world-leading DM limits $<$0.5 GeV (Das et al., 2 Dec 2025).
  • CEνNS and LDM sensitivity: CryoCsI projects $<0.5$ keV$_{nr}$ thresholds, with full parameter-space coverage for $m_\chi$ below 1 GeV, and $>10^4$ CEνNS events above threshold in 3 years (Ding, 29 May 2024).
  • Intrinsic radioactivity upper bounds in pure CsI: $^{134}$Cs $<$2 mBq/kg, background rates $<$mHz/kg (20–1000 keV) (Su et al., 22 Sep 2025).

6. Advanced Architectures and Emerging Directions

Next-generation cryogenic detectors are leveraging new materials, sensors, and assembly methods (Das et al., 2 Dec 2025, Zema et al., 2 Feb 2024, Ding, 29 May 2024, Buzulutskov, 2011):

• Single-Photon/Single-Electron Sensitivity: TES-based detectors (ALPS II) reach $0.1$ eV FWHM, $>90\%$ efficiency for $1.165$ eV photons, with dark rates $<7\times10^{-6}$ cps (Shah et al., 2021). Neganov–Luke-amplified Si reaches $5$ eV threshold and eV-scale energy resolution (Defay et al., 2015).
• Large-Volume/Low-Threshold Arrays: Modular, kg-scale high-purity CsI at 95 K with dual-ended PMT or SiPM readout, demonstrating sub-keV thresholds and best-in-class spatial/energy uniformity (Su et al., 22 Sep 2025, Ding, 29 May 2024).
• Noncontact and Contact-Free Electrodes: Si devices with vacuum-gap contacts suppress leakage to $<1$ pA and enable high biases (100 V) over kg-scale detectors for enhanced NTL gain and threshold reduction (Mast et al., 2018).
• Automated Manufacturing QA: Vision-pipeline analysis for wafer-level defect detection in superconducting sensor arrays achieves $98.6\%$ simulated detection accuracy, critical for scaling to megapixel arrays (Ferguson et al., 4 Jan 2025).
• Cryogenic Avalanche Detectors (CRAD): GEM/THGEM-based noble-gas detectors achieve stable avalanche gain O($10^3$–$10^5$) at $T=2-180$ K with single-e$^-$ counting and mm spatial resolution (Buzulutskov, 2011).

7. Challenges, Optimization, and Outlook

Scaling, reproducibility, and further threshold reduction remain areas of active research:

• Purity and Background Mitigation: Achieving sub-ppb bulk/gas purity in noble targets with optimized recirculation, radiopure materials, fully scintillating housings, and radon suppression (Grauso et al., 2022, Su et al., 22 Sep 2025).
• Low-Energy Backgrounds: CRESST-like excesses below $\sim 100$ eV require suppression by $<10^{-4}$ for electron-scattering searches. Active veto, improved sensor fabrication, and deep underground deployment are principal strategies (Zema et al., 2 Feb 2024).
• Multiplexed Readout: For MKIDs and TES, massive channel counts drive developments in low-noise, scalable SQUID/ASIC readout and firmware-triggered pipeline analysis (Das et al., 2 Dec 2025, Ferguson et al., 4 Jan 2025).
• Sensor Materials and Geometry: Ongoing work to optimize Teflon wrapping, surface polish, absorber thickness, and SiPM design seeks further gains in photon detection, energy resolution, and scalability (Ding, 29 May 2024, Su et al., 22 Sep 2025).
• Quantum-Limited and Directional Readout: Efforts on microwave cavity (SNSPD) and distributed sensor arrays seek to push into the $\mu$eV regime and enable 3D event reconstruction and directional sensitivity (Das et al., 2 Dec 2025, Collaboration et al., 2021).

Cryogenic detectors have set new standards for rare-event physics, enabling world-best sub-keV thresholds, SI/SD interaction sensitivities, and mass scalability up to the tonne scale. Ongoing advances in sensor physics, materials engineering, automated QA, and data analysis are poised to open further discovery potential in dark matter, CEνNS, $0\nu\beta\beta$, solar/geo-neutrino physics, and beyond (Das et al., 2 Dec 2025, Su et al., 22 Sep 2025).