TRISTAN: KATRIN's Upgrade for keV Sterile Neutrino Dark Matter

KATRIN is built around the tritium beta-decay endpoint at 18.6 keV. Its huge electrostatic spectrometer and intricate windowless source design are optimised to measure the last few tens of electron-volts of the beta spectrum, where the shape encodes the active neutrino mass. Outside that narrow window, the standard KATRIN configuration is essentially blind: the spectrometer voltage filters out everything below 18.5 keV, and the detector cannot handle the much higher rates at lower electron energies.

The same tritium source, however, contains an unrivalled wealth of beta-decay events at all energies below the endpoint. A 1 keV electron has a rate something like $10^{10}$ times higher than an 18.6 keV electron. If a keV-scale sterile neutrino existed and mixed with the active sector, it would imprint a small kink on this huge low-energy spectrum — and KATRIN’s source is already running at the level required to look for it.

TRISTAN is the planned hardware upgrade that turns KATRIN into a keV-sterile-neutrino dark-matter search. The existing pin-diode detector is replaced with a silicon-drift-detector array sized to handle the full-spectrum rate, and the analysis runs across the entire 1-to-18.6-keV range instead of the last 30 eV. Installation is scheduled for after KATRIN’s main active-mass campaign concludes, and the resulting sensitivity reaches into the parameter space favoured by warm-dark-matter cosmology and by some structure-formation arguments. This post is about why keV sterile neutrinos are interesting, what a sterile-kink signature looks like, and what TRISTAN is built to find.

Why keV sterile neutrinos are a serious dark-matter candidate

Dark matter is conventionally pictured as a cold, weakly interacting particle. But “cold” is a property the particle has at decoupling, not a fundamental requirement. A particle with mass in the keV range and a non-thermal velocity distribution that has cooled with the Hubble expansion behaves as warm dark matter: cold enough on galactic scales to produce the observed large-scale structure, but warm enough to suppress small-scale structure below sub-galactic scales. The suppression is potentially attractive because cold-dark-matter simulations overpredict the number of small satellites of galaxies like the Milky Way, and warm dark matter naturally addresses this.

A keV sterile neutrino is the simplest particle-physics realisation. Production proceeds through the Dodelson-Widrow mechanism (later refined by Shi and Fuller): in the early universe, active-neutrino oscillations into the sterile state convert a small fraction of the active population into sterile, with the conversion rate and total abundance set by the mixing angle and the lepton-number asymmetry. For sterile mass in the keV range and mixing angle around ${\sin }^2(2\theta )\sim 10^{-11}$ to $10^{-7}$, the resulting abundance can match the observed dark-matter density.

The same particle can also decay through the mixing back to an active neutrino plus a photon:

${\nu }_s\to {\nu }_a+\gamma \text{with}{E}_{\gamma }=m_s/2$

X-ray telescopes therefore search for narrow keV-scale X-ray emission from galaxy clusters and from dwarf galaxies, where the decay signal would be loudest. The detection of a 3.5 keV X-ray line in galaxy-cluster stacks reported in 2014 by two independent groups led to a long-running debate about whether it represented a 7 keV sterile-neutrino decay or an atomic line. The line has been argued for and against in the years since, and a clear resolution awaits a dedicated mission. But the episode established that the parameter space is real and reachable, and that direct laboratory tests would be valuable.

That is what TRISTAN provides: a clean laboratory test, independent of any astrophysical-modelling uncertainties.

The kink signature

A standard tritium beta decay produces an electron with continuous energy from 0 to the Q-value of 18.6 keV, with the neutrino taking the remaining energy. The exact spectrum shape is given by the Kurie function with a phase-space term depending on the neutrino mass.

If a keV sterile neutrino is mixed into the neutrino mass eigenstates, then a fraction ${\sin }^2(2\theta )$ of decays produces the sterile rather than the active mass eigenstate. The sterile carries a minimum energy equal to its rest mass $m_s$ when produced at rest, so it cannot be produced at all if the electron’s energy exceeds $Q-m_s$. Above $E_e=Q-m_s$ the spectrum follows the active-only shape; below, the spectrum includes the additional sterile contribution.

The result is a kink in the spectrum at electron energy $E_{\mathrm{kink}}\approx Q-m_s$. The kink amplitude is ${\sin }^2(2\theta )$; the kink position is set by $m_s$. Detecting a kink at the $10^{-6}$ level requires very high statistics across the spectrum and very well-controlled detector systematics, since any systematic energy-dependent efficiency variation could mimic the signal.

Why a new detector is needed

KATRIN’s existing pin-diode detector array is a 148-pixel silicon device positioned downstream of the main spectrometer. It is optimised for the narrow energy window near the 18.6 keV endpoint and for the modest event rate that passes the spectrometer’s high-pass voltage filter — typically a few hundred events per second in standard operation.

For TRISTAN’s full-spectrum search the situation changes drastically. The total event rate from a tritium source observed across the full spectrum is roughly $10^{11}$ Hz. Even after the spectrometer voltage is lowered to admit lower-energy electrons, the resulting count rate at the detector reaches $10^8$ Hz — six orders of magnitude above what the pin-diode array can handle.

The TRISTAN detector is a 21-module silicon drift detector (SDD) array with about 3,000 pixels total. SDDs are sensitive to keV-scale electron energies with energy resolution of about 100 eV FWHM and can handle the required count rate per pixel without pileup. The total active area is roughly 200 cm². Each pixel reads out independently into a multichannel ADC system that records each event’s energy and time.

The detector also needs to handle additional backgrounds that are negligible in the standard KATRIN configuration but become important at higher rates: pile-up of two simultaneous electrons producing a single high-energy hit, background events from beta-induced X-ray fluorescence in the detector chamber, and detector dead-layer effects that affect the lowest-energy events.

Sensitivity and competition

The TRISTAN sensitivity to keV sterile neutrinos depends on the achievable statistics, the systematic control of the detector response, and the modelling of the standard tritium spectrum shape. Current projections target ${\sin }^2(2\theta )<10^{-6}$ for sterile masses from 1 to 18 keV after a multi-year exposure — covering a significant portion of the parameter space favoured by warm-dark-matter cosmology, including most of the region where the 3.5 keV X-ray line could be consistent with a sterile-neutrino origin.

Competing experiments approach the same parameter space differently. Project 8 uses cyclotron radiation emission spectroscopy to measure individual tritium electrons across the full spectrum without a spectrometer, achieving exceptional energy resolution but at substantially lower statistics. HUNTER uses cesium-131 electron capture in a magneto-optical trap to measure the missing momentum from the emitted neutrino, looking for a sterile contribution. PTOLEMY, primarily a cosmic-neutrino-background experiment, also has sensitivity to keV steriles through its tritium analysis. Each technique probes overlapping but distinct parameter regions, and the combination would substantially tighten constraints if no signal appears.

Beyond dark matter

A keV sterile neutrino signal would have implications well beyond the dark-matter question. The same mixing that produces the dark matter abundance would constrain the seesaw mechanism, since keV-scale right-handed neutrinos with the required mixings appear in low-scale seesaw realisations such as the $\nu$MSM (neutrino Minimal Standard Model) proposed by Asaka, Blanchet and Shaposhnikov. In the $\nu$MSM, three right-handed neutrinos at the keV-to-GeV scale simultaneously explain dark matter (the lightest), neutrino masses (via the seesaw), and the baryon asymmetry (through resonant leptogenesis with the two heavier states). A TRISTAN signal would directly probe one of the three required ingredients.

A null result from TRISTAN would tighten the constraints on the $\nu$MSM and on warm-dark-matter cosmology generally, and would force any sterile-neutrino dark-matter scenario into either tighter parameter regions or more elaborate model assumptions.

Summary

TRISTAN is KATRIN’s planned upgrade for searching for keV-scale sterile-neutrino dark matter through the full tritium beta-decay spectrum. The signature is a tiny kink in the electron energy distribution at $E_{\mathrm{kink}}\approx Q-m_{\mathrm{sterile}}$, with amplitude set by the mixing angle ${\sin }^2(2\theta )$. The existing KATRIN detector is replaced with a 3,000-pixel silicon drift detector array capable of handling the $10^8$ Hz count rate that the full-spectrum measurement demands. Projected sensitivity reaches ${\sin }^2(2\theta )<10^{-6}$ for sterile masses from 1 to 18 keV after multi-year operation — covering the parameter space favoured by warm-dark-matter cosmology and by the $\nu$MSM scenario that ties together dark matter, neutrino mass and baryogenesis. Installation is planned for the late 2020s after KATRIN’s active-mass campaign concludes, making TRISTAN one of the leading dedicated probes of keV sterile-neutrino dark matter in the next decade.