Trinity Neutrino Observatory

• Trinity Neutrino Observatory is a proposed optical array that detects Earth-skimming tau neutrinos using Cherenkov and fluorescence light from upward air showers.
• It employs a staged development approach—from the Trinity Demonstrator to a multi-site observatory—to bridge the energy range overlapping with IceCube and radio-based detectors.
• Advanced design features include wide field-of-view SiPM cameras, optimized horizon-pointing strategies, and effective background suppression techniques.

Trinity Neutrino Observatory is a proposed mountain-top, horizon-pointing optical neutrino observatory designed to detect Earth-skimming tau neutrinos by imaging the Cherenkov light, and in the original case studies also the fluorescence light, from upward-going air showers initiated by tau decays in the atmosphere. Across its design studies, sensitivity calculations, and status reports, Trinity is presented as a staged program comprising the Trinity Demonstrator, Trinity One, and a final multi-site observatory of 18 wide-angle Cherenkov telescopes, targeting approximately 1 PeV to 10 EeV while overlapping at the low end with IceCube and at the high end with radio-based ultra-high-energy neutrino detectors (Otte et al., 2019, Wang et al., 2021, Stepanoff et al., 22 Sep 2025, O. et al., 22 Sep 2025).

1. Scientific rationale and observational niche

Trinity was proposed to address the largely unexplored neutrino window above the energies already probed by IceCube. Early concept papers framed the relevant regime as approximately $10^7~\mathrm{GeV}$ to $10^{10}~\mathrm{GeV}$, or more generally the ultrahigh-energy band above $10^7~\mathrm{GeV}$, with three main objectives: narrowing in on source classes responsible for the astrophysical neutrino flux measured by IceCube, helping to find the sources and composition of ultrahigh-energy cosmic rays, and testing fundamental neutrino physics at the highest energies (Otte et al., 2019). Those same papers also situated Trinity in the context of cosmogenic neutrinos, ANITA-like anomalous upward-going events, and the use of mature air-shower imaging methods from CTA, H.E.S.S., MAGIC, VERITAS, Pierre Auger, and Telescope Array (Otte et al., 2019).

A later development in the concept was the extension of the low-energy reach. Updated sensitivity calculations argued that Trinity is sensitive to Earth-skimming neutrinos from $10^6~\mathrm{GeV}$ to $10^{10}~\mathrm{GeV}$, and emphasized that this lower threshold enables direct overlap with the upper end of the IceCube astrophysical neutrino band rather than only with the classical ultrahigh-energy regime (Wang et al., 2021). That overlap is central to Trinity’s scientific identity: the observatory is intended not only to search for cosmogenic neutrinos and other extreme-energy components, but also to determine whether the diffuse astrophysical neutrino spectrum continues as a power law, softens, or cuts off above the IceCube range (Wang et al., 2021).

The later observatory papers recast the same niche in PeV–EeV language. In that framing, Trinity targets the 1 PeV to 10 EeV band because it bridges conventional high-energy neutrino astronomy and ultrahigh-energy techniques, probes the origin of ultrahigh-energy cosmic rays, extends neutrino astronomy toward cosmogenic neutrinos, and provides source-discovery and source-characterization capability for point sources and transients in a regime where current measurements remain sparse (Stepanoff et al., 22 Sep 2025, O. et al., 22 Sep 2025).

2. Earth-skimming tau-neutrino detection principle

The defining Trinity channel is the Earth-skimming $\nu_\tau$ channel. A tau neutrino enters the Earth at a shallow angle, typically less than $10^\circ$ below the horizon, undergoes a charged-current interaction in rock, produces a tau lepton, and the tau may emerge from the ground before decaying in the atmosphere. Its decay initiates an upgoing or near-horizontal extensive air shower, and the shower emits Cherenkov light that can be imaged from distances up to about 200 km (Otte et al., 2019, O. et al., 22 Sep 2025).

This geometry is central for two reasons. First, the Earth provides a very large target mass for neutrino interactions while also suppressing many ordinary downward-going backgrounds. Second, the tau’s finite range makes the $\nu_\tau \rightarrow \tau \rightarrow$ air-shower chain uniquely favorable for optical detection in this energy regime. Trinity therefore differs from volumetric neutrino telescopes and from most radio-based concepts in that it observes an atmospheric secondary rather than the primary interaction site itself (Brown et al., 2021).

The original Trinity case studies emphasized that the observed optical signal is not purely Cherenkov. Simulations showed a near-distance peak from fluorescence and a broader far-distance peak from Cherenkov-detected events, with fluorescence contributing about 20% of all triggered events in the acceptance calculation (Otte et al., 2019). At the same time, the observatory is fundamentally optimized as a Cherenkov imaging instrument: later status papers describe the sought signal as a well-defined, elongated elliptical air-shower image in the camera, in contrast to backgrounds such as direct muons crossing the camera, spillover events in neighboring parallel SiPMs, fuzzy compact events, and electronics or module malfunctions (Stepanoff et al., 22 Sep 2025).

The horizon-pointing strategy follows the approach explored by MAGIC rather than the nearby-mountain geometry of ASHRA. Trinity telescopes are pointed slightly below the horizon, so that an upper part of the camera can function as a veto region for downward-going cosmic-ray air showers while the lower part images the atmospheric volume in which emerging taus decay and produce visible showers (Stepanoff et al., 22 Sep 2025). In the early design studies, the relevant air showers were expected to remain low in the atmosphere: even for a $10^{10}~\mathrm{GeV}$ tau neutrino, the shower develops over only about $10~\mathrm{km}$, and for the vast majority of tau-induced showers the shower tip reaches a maximum altitude of only 2 km to 5 km (Otte et al., 2019).

3. Observatory architecture and staged development

Trinity’s architecture evolved from a six-telescope baseline concept into a larger multi-site observatory. The early white paper described a system of six telescopes, each with a $10^{10}~\mathrm{GeV}$0 field of view, together providing $10^{10}~\mathrm{GeV}$1 azimuth coverage from a site at 2–3 km altitude, with operational requirements including a $10^{10}~\mathrm{GeV}$2 duty cycle, $10^{10}~\mathrm{GeV}$3 effective light-collection surface, $10^{10}~\mathrm{GeV}$4 pixel size, sensitive wavelength range $10^{10}~\mathrm{GeV}$5–$10^{10}~\mathrm{GeV}$6 nm, and $10^{10}~\mathrm{GeV}$7 megasamples/s readout (Otte et al., 2019). The 2019 case study used a MACHETE-derived optical concept with a $10^{10}~\mathrm{GeV}$8 collection area, 68 mirrors arranged in four rows of 17 mirrors, focal length $10^{10}~\mathrm{GeV}$9, approximately 3,300 pixels, and a point spread function of $10^7~\mathrm{GeV}$0 across the full $10^7~\mathrm{GeV}$1 field of view (Otte et al., 2019).

A 2021 sensitivity study adopted a larger observatory-level layout: three sites, each at altitude $10^7~\mathrm{GeV}$2 km, with telescopes collectively covering $10^7~\mathrm{GeV}$3 in azimuth and a $10^7~\mathrm{GeV}$4 elevation field of view. In that study the assumed angular resolution was $10^7~\mathrm{GeV}$5, the effective light-collection area was $10^7~\mathrm{GeV}$6 in any direction, and the camera model used the characteristics of blue-sensitive Hamamatsu S14520-6050CN SiPMs (Wang et al., 2021). The same paper noted that sensitivity scales approximately linearly with the monitored atmospheric volume and, because telescope fields of view do not overlap, proportionally with the number of telescopes (Wang et al., 2021).

Later observatory papers formalized a three-stage program.

Stage	Role	Selected published parameters
Trinity Demonstrator	Pathfinder	Deployed in Fall 2023 at Frisco Peak, Utah; $10^7~\mathrm{GeV}$7 mirror area; 256-pixel SiPM camera; $10^7~\mathrm{GeV}$8 field of view; $10^7~\mathrm{GeV}$9 angular resolution
Trinity One	First complete telescope	$10^6~\mathrm{GeV}$0 segmented light-collection surface; $10^6~\mathrm{GeV}$1 field of view; 3,328 SiPM pixels; $10^6~\mathrm{GeV}$2 resolution; $10^6~\mathrm{GeV}$3, 12-bit readout; azimuth rotation
Full observatory	Final array	18 wide-angle Cherenkov telescopes across at least three sites; 1 PeV to 10 EeV target range; $10^6~\mathrm{GeV}$4 duty-cycle assumption

The Trinity Demonstrator is a proof-of-concept precursor rather than a scaled-down copy of the final telescope, but it shares key performance parameters with the intended observatory, especially angular resolution, readout sampling speed, and nearly identical vertical field of view (Bagheri et al., 14 Mar 2025). Trinity One is the first observatory-grade telescope: it is to be located on Frisco Peak, Utah, can rotate in azimuth over $10^6~\mathrm{GeV}$5, and alone can observe 64% of the sky (O. et al., 22 Sep 2025). The full Trinity Neutrino Observatory is described as an array of 18 wide-angle Cherenkov telescopes distributed across at least three sites, with the observatory after ten years expected to have 36 times the sensitivity of a single telescope after five years (Stepanoff et al., 22 Sep 2025).

The design uses SiPM-based cameras and switched-capacitor-array readout throughout the program. Early Trinity studies employed red-sensitive SiPMs because long-path atmospheric transmission suppresses much of the blue Cherenkov light and leaves a red-peaked spectrum at large distances (Otte et al., 2019). Later papers continued to stress SiPM advantages, particularly their stronger red sensitivity relative to MAGIC’s photomultiplier tubes and the resulting gain in effective area for far-away tau-induced showers (Stepanoff et al., 22 Sep 2025). The readout lineage also remained stable: MUSIC or eMUSIC front-end ASICs and AGET digitizers appear throughout the design and prototype papers (Otte et al., 2019, Bagheri et al., 14 Mar 2025).

4. Sensitivity formalism and projected astrophysical reach

Trinity’s sensitivity calculations are based on a full chain that folds together neutrino interaction in the Earth, tau emergence, tau decay, air-shower development, optical propagation, detector triggering, and image selection. For diffuse neutrinos, the 2021 sensitivity paper defined the acceptance as

$10^6~\mathrm{GeV}$6

where the probability term includes interaction in the Earth, tau emergence, shower development within the field of view, sufficient light for imaging and later reconstruction, and an air-shower image length exceeding $10^6~\mathrm{GeV}$7 (Wang et al., 2021). The same paper defined sensitivity as the flux yielding one detected neutrino in effective observation time $10^6~\mathrm{GeV}$8, with a factor of 3 converting from tau-neutrino-only sensitivity to an assumed $10^6~\mathrm{GeV}$9 flavor-mixed astrophysical flux at Earth (Wang et al., 2021).

For Trinity One, the direction- and energy-dependent effective area is written as $10^{10}~\mathrm{GeV}$0, and the paper introduces an $10^{10}~\mathrm{GeV}$1-weighted integral acceptance,

$10^{10}~\mathrm{GeV}$2

together with the source event-yield relation

$10^{10}~\mathrm{GeV}$3

For short transients with durations below 1000 s, the paper uses a fluence sensitivity based on the Feldman–Cousins 90% C.L. upper limit of 2.44 per energy decade for a background-free observation (O. et al., 22 Sep 2025).

The principal observatory-level result of the updated diffuse study is that a single Trinity telescope can detect the IceCube diffuse astrophysical neutrino flux in five years if the spectrum continues without a turnover, and that the full observatory can distinguish between an unbroken power law and an exponential cutoff scenario constrained by IceCube upper limits (Wang et al., 2021). For the two benchmark scenarios over ten years, the published event counts are 70 detected neutrinos for the no-cutoff case, corresponding to about 7 neutrinos per year, and 6 detected neutrinos in ten years for the cutoff case (Wang et al., 2021). The same paper emphasizes that this separation is large enough to diagnose the high-energy spectral shape even without event-by-event energy reconstruction (Wang et al., 2021).

Point-source and transient reach is strongly declination dependent. For one site at Frisco Peak, Utah, the 2021 exposure study obtained an annual observation time of 1765 hours, corresponding to a 20% duty cycle, an average observability of 4.8 hours per night, and a maximum nightly observation duration of 12 hours (Wang et al., 2021). With full azimuthal coverage, the instantaneous sky acceptance forms a band with $10^{10}~\mathrm{GeV}$4 full width at half maximum, and sources with declinations $10^{10}~\mathrm{GeV}$5 are observable from that site (Wang et al., 2021).

Transient sensitivity is summarized operationally in the same paper: for hard-spectrum transients, fluxes of a few $10^{10}~\mathrm{GeV}$6 are detectable within one hour, and fluxes of order $10^{10}~\mathrm{GeV}$7 are detectable within one year (Wang et al., 2021). Among specific benchmark sources, the paper states that the full Trinity Observatory would be sensitive to TXS 0506+056-like ultrahigh-energy emission if the source is in its field of view, and that the extrapolated ultrahigh-energy flux from NGC 1068 would be detectable within one week irrespective of source declination within the accessible range (Wang et al., 2021).

The Trinity One paper sharpened this source-oriented program. For a TXS 0506+056-like 158-day flare, a source at the sky position of TXS 0506+056 would yield $10^{10}~\mathrm{GeV}$8 hours of Trinity One exposure and an expected $10^{10}~\mathrm{GeV}$9 neutrinos, corresponding to a 15% probability of detecting at least one event (O. et al., 22 Sep 2025). The same paper states that a similar flare from a source at $\nu_\tau$0 would yield approximately one detected neutrino within one year of observation, that a GRB 221009A-like event could yield at least one neutrino if it occurs within 10 Gpc and in the field of view, and that a short GRB within 100 Mpc would be detectable (O. et al., 22 Sep 2025). For short GRBs, the cumulative probability of at least one detection reaches 60% after 20 years for a single telescope and approaches 90% after 20 years in an extrapolated configuration of six telescopes at three sites with 80% daily sky coverage (O. et al., 22 Sep 2025).

5. Trinity Demonstrator: prototype, commissioning, and first analyzed data

The Trinity Demonstrator is the first deployed instrument of the program. It was installed on Frisco Peak, Utah, at an elevation of $\nu_\tau$1, with first light on October 3, 2023, and commissioning completed in June 2024 (Bagheri et al., 14 Mar 2025). The telescope uses a Davies–Cotton optical design with a focal sphere radius and focal length of $\nu_\tau$2, 77 round mirror facets of 15 cm diameter, and a combined light-collection area of $\nu_\tau$3 (Bagheri et al., 14 Mar 2025). Its 256-pixel camera is based on Hamamatsu S14161-6050HS-04 SiPM matrices, has a $\nu_\tau$4 field of view and $\nu_\tau$5 angular resolution, and is read out by eMUSIC front-end ASICs and a 256-channel AGET-based switched-capacitor-array system at $\nu_\tau$6 with 12-bit resolution (Bagheri et al., 14 Mar 2025).

The Demonstrator permanently points toward azimuth $\nu_\tau$7, with the camera center $\nu_\tau$8 below the horizon; in later status reporting this geometry is summarized as $\nu_\tau$9 of the vertical field of view above the horizon and $10^\circ$0 below it, with the upper part acting as a veto region for downward-going cosmic-ray air showers (Bagheri et al., 14 Mar 2025, Stepanoff et al., 22 Sep 2025). Triggering is deliberately simple: the camera is read out whenever one SiPM exceeds threshold, and the observing threshold used in practice is about 20 photoelectrons (Bagheri et al., 14 Mar 2025, Stepanoff et al., 22 Sep 2025). Calibration is based on an onboard LED flasher, flat-fielding of relative gains across the camera, absolute conversion to photoelectrons, and corrections for temperature drifts in SiPM breakdown voltage (Bagheri et al., 14 Mar 2025, Stepanoff et al., 22 Sep 2025).

Commissioning established several low-level performance characteristics relevant to the full observatory. The optical point-spread function corresponds to an 80% containment radius of $10^\circ$1 in the focal plane, well within one SiPM pixel (Bagheri et al., 14 Mar 2025). The camera response could be flat-fielded to within 5% standard deviation, and the prototype achieved readout dead time below 1.4% at average trigger rates below 10 Hz (Bagheri et al., 14 Mar 2025). The dominant noise source was the night-sky background rather than the electronics, and the prototype also demonstrated remote autonomous operation with multi-layer safety interlocks and daily data transfer to Georgia Tech (Bagheri et al., 14 Mar 2025).

The first analyzed science-quality data set covered October 2024 to May 2025 and comprised 346 hours of observations with 979,957 hardware triggers (Stepanoff et al., 22 Sep 2025). After image cleaning and a requirement of at least three cleaned pixels, 497 events survived, corresponding to about 1.4 events per hour (Stepanoff et al., 22 Sep 2025). Visual inspection identified four classes: 461 clean muon tracks, 15 muon tracks with signals in parallel SiPMs, 2 broad fuzzy events, and 19 events from a misbehaving camera module (Stepanoff et al., 22 Sep 2025). No neutrino candidate event was found (Stepanoff et al., 22 Sep 2025).

The significance of this null result is methodological rather than astrophysical. The collaboration explicitly states that, given the Demonstrator’s sensitivity, no neutrino-induced air-shower signals were expected in the analyzed data set (Stepanoff et al., 22 Sep 2025). What the prototype established instead was that the dominant background is direct charged-particle hits in the SiPM camera, mainly muons, and that these backgrounds are morphologically distinct from the expected “gold-plated” neutrino-induced shower image (Stepanoff et al., 22 Sep 2025). The same work notes that a plastic-scintillator-based veto detector surrounding the focal plane could reject the muon background during data taking in future configurations (Stepanoff et al., 22 Sep 2025).

6. Physics implications, limitations, and future trajectory

Trinity is not designed as an all-flavor, all-sky, all-time neutrino survey instrument. It is optimized for the Earth-skimming $10^\circ$2 channel, and its published exposures depend strongly on horizon geometry, declination, accessible azimuth, dark-sky observing constraints, and duty cycle (Wang et al., 2021, O. et al., 22 Sep 2025). This specialization is also its physics strength: because Trinity is specifically sensitive to tau neutrinos, the observatory papers argue that joint analyses with IceCube can probe the astrophysical flavor ratio, source-environment diagnostics such as magnetic fields affecting production and cooling, and beyond-standard-model scenarios including neutrino decay and non-standard oscillations (Wang et al., 2021).

A broader methodological context for that claim comes from solar-neutrino decay studies. The Sudbury Neutrino Observatory showed that searches for invisible neutrino decay require flavor sensitivity and a direct calculation of survival and appearance probabilities, because invisible disappearance violates the usual active-flavor flux-conservation relation $10^\circ$3 (Collaboration et al., 2018). That result concerns solar $10^\circ$4B neutrinos rather than Trinity’s PeV–EeV tau-neutrino channel, but it suggests that Trinity’s tau-neutrino specificity is most powerful in a multi-detector framework with complementary flavor response and energy-resolved spectral analysis.

Published Trinity sensitivity estimates also carry explicit caveats. The 2021 conference paper did not provide a standalone budget for atmospheric attenuation parameters, residual background rates, trigger thresholds in photoelectrons, or systematic uncertainties from neutrino cross sections or Earth density profiles; these were folded into the simulation framework rather than unpacked individually (Wang et al., 2021). The same work noted that the strong performance claims still need validation with prototype detectors (Wang et al., 2021). Earlier design papers were more direct still: whether the quoted sensitivity can be achieved depends on sufficiently suppressing background events near the horizon, and that had to be tested with a prototype telescope station (Otte et al., 2019).

The program’s development trajectory is therefore inseparable from prototype validation. The Demonstrator has already validated remote operation, calibration strategy, and a first background taxonomy, and later status papers conclude that these results support proceeding to Trinity One (Stepanoff et al., 22 Sep 2025). Trinity One, in turn, is presented as the bridge between proof-of-concept and full observatory deployment: a $10^\circ$5, azimuth-rotatable, $10^\circ$6 telescope with 3,328 SiPM pixels and astrophysically useful point-source sensitivity (O. et al., 22 Sep 2025). In that sense, the Trinity Neutrino Observatory remains both a detector concept and an evolving experimental program whose defining claim is that wide-field optical imaging of Earth-skimming tau-induced air showers can make the PeV–EeV neutrino sky observationally accessible.