Dark Matter Quantum Information Science
Dark Matter/Quantum Information Science

HVeV Detectors: TES-Based Phonon Sensors

What is an HVeV Detector?

HVeV detectors are gram-scale cryogenic silicon calorimeters capable of detecting individual electron-hole pairs produced by particle interactions in the crystal. The name reflects the detector's operating principle: a high voltage (HV) bias is applied across the substrate, so that each electron-hole pair produced by an interaction drifts across the crystal and produces a characteristic Neganov-Trofimov-Luke phonon signal proportional to the bias voltage times the charge. This amplification mechanism allows the phonon sensor to resolve individual charge quanta even with modest intrinsic phonon resolution.

The phonon signal is read out by arrays of Quasiparticle-trap-assisted Electrothermal-feedback Transition Edge Sensors (QETs) — thin-film superconducting sensors that transduce phonon energy into a measurable current change. Hundreds of QETs are connected in parallel to form a single detector channel, each consisting of aluminum fins that collect phonons from the substrate and funnel their energy into a tungsten TES biased near its superconducting transition.

State-of-the-Art Performance

The latest generation of SuperCDMS HVeV detectors has achieved a baseline phonon energy resolution of 612 meV — sub-eV resolution that enables detection of single electron-hole pairs and individual photons (see here for more details). This improvement over prior generations was achieved primarily by reducing the tungsten TES critical temperature from ~65 mK to ~40 mK, which lowers the fundamental thermal noise floor of the sensor. These detectors have been operated successfully at our SLAC surface-level test facility, the Northwestern EXperimental Underground Site at Fermilab (NEXUS) (Runs 3 and 4), and the Cryogenic Underground TEst (CUTE) facility at SNOLAB, located approximately 2 km underground near Sudbury, Ontario. The have also been operated in a neutron beam at Triangle Universities Nuclear Laboratory (TUNL) to measure nuclear recoil ionization yield down to 100 eV, and implanted with Be7 to study eV-scale nuclear recoils and search for new physics in the neutrino sector.

Phonon collection efficiency — the fraction of event energy ultimately absorbed by the TES — has been measured at approximately 45% at CUTE for our Si detectors, establishing these as the most efficient athermal phonon detectors demonstrated to date and approaching the theoretical limit for phonon-to-quasiparticle conversion. Noise performance is consistent with the standard TES noise model, with no evidence for the anomalous correlated phonon noise reported by other collaborations. We recently used our MEMs scanning system to show that phonon efficiency is position independent to a high degree of precision, validation our phonon collection models. Work is ongoing in our group to port this technology to 4H-SiC and Diamond substrates to enable new paradigms in HVeV sensing for dark matter detection at the single phonon limit. Our group has also been working with Lawrence Livermore National Laboratory to demonstrate phonon sensing in diamond, showing that diamond can achieve better than 99% athermal phonon detection efficiency even in low sensor coverage configurations.

Application: Anti-Coincidence Detectors for X-ray Observatories

The QET sensor design at the heart of HVeV detectors is readily adaptable to new science targets. One such application is the anti-coincidence (anti-co) detector required by future space-based X-ray observatories to veto charged particle backgrounds from galactic cosmic rays. Working from an existing two-channel, 1 cm² HVeV design, the group scaled the W-TES QET unit cell geometry directly into a twelve-channel, 14 cm² prototype — tiling approximately 6,300 QET cells across a 0.5 mm-thick hexagonal silicon wafer. The sensor pattern and QET count per channel were kept fixed from the HVeV design, with the coverage area simply expanded.

This scaling has predictable consequences for detector performance: energy resolution and threshold degrade modestly with increased channel area and phonon collection time, while dynamic range improves. The prototype demonstrated a low-energy threshold well below 1 keV and a live-time fraction above 96% across an energy range spanning four keV to 5.5 MeV — meeting or exceeding the requirements for a LEM-class mission by a wide margin. The multi-channel readout also yielded mm-scale spatial resolution in event position, a capability that emerges naturally from the distributed channel layout and the same athermal phonon diffusion physics that governs HVeV detector response.

G4CMP phonon transport simulations — the same modeling framework developed for HVeV and KIPM detectors — are being used to understand the position- and energy-dependent response of the anti-co and guide future design iterations. The shared simulation infrastructure and fabrication expertise between the dark matter and X-ray observatory programs reflects the versatility of the QET-based athermal phonon detection approach.

For more details, see this article in SLAC news.

HVeV Talks from DMQIS Group