Trinity PeV-Neutrino Observatory

• Trinity PeV-Neutrino Observatory is a ground-based facility that detects Earth-skimming tau neutrinos using horizon-pointed Cherenkov imaging.
• It bridges the energy gap from ~1 PeV to 10 EeV, overlapping IceCube’s flux range while extending into the ultra-high-energy domain.
• Its staged development—from the Demonstrator to Trinity One and a multi-site array—enhances sensitivity, background rejection, and point-source capabilities.

The Trinity PeV-Neutrino Observatory is a proposed ground-based observatory for very-high-energy and ultra-high-energy tau-neutrino astronomy that uses wide-field imaging atmospheric Cherenkov telescopes pointed at the horizon to detect upward-going extensive air showers from Earth-skimming $\nu_\tau$ interactions. Across its design studies and status reports, Trinity is defined as covering the $\sim 1$ PeV to $10$ EeV regime, with the specific role of bridging the energy interval between the PeV astrophysical-neutrino domain observed by IceCube and the higher-energy reach of radio-based UHE neutrino detectors; its development is staged through the Trinity Demonstrator, Trinity One, and a multi-site full observatory (Brown et al., 2021, Stepanoff et al., 22 Sep 2025).

1. Origins, scientific scope, and observational niche

Trinity emerged from 2019 case studies of a modular air-shower imaging system optimized for Earth-skimming tau neutrinos above $10^7$ GeV, using techniques derived from imaging atmospheric Cherenkov telescopes and UHECR fluorescence detectors (Otte et al., 2019, Otte et al., 2019). Subsequent design work lowered the threshold to $\sim10^6$ GeV, or $\sim 1$ PeV, so that Trinity would overlap directly with IceCube’s astrophysical flux band while extending measurements through the PeV–EeV window (Wang et al., 2021). In later status documents the project is explicitly framed as a PeV–EeV tau-neutrino observatory built around arrays of wide-field IACTs aimed slightly below the horizon (Stepanoff et al., 22 Sep 2025).

The scientific program has remained consistent across these stages. Trinity is intended to constrain the high-energy tail of the astrophysical neutrino spectrum, distinguish cutoff scenarios in IceCube-motivated flux models, probe cosmogenic neutrinos, inform UHECR origin and composition, and test neutrino interactions at energies beyond accelerator reach (Wang et al., 2021, Stepanoff et al., 22 Sep 2025). The observatory is also presented as a point-source and transient instrument: sensitivity estimates are given for flaring blazars, short gamma-ray bursts, and high-luminosity GRBs, with the Earth-skimming $\nu_\tau$ channel providing both flavor selectivity and strong geometric background suppression (O. et al., 22 Sep 2025).

A recurrent misconception is that horizon-pointed tau-neutrino instruments are intrinsically limited to the EeV scale. Trinity was instead designed to begin at approximately $1$ PeV, specifically to secure overlap with IceCube and to access the energy decade where a continuation or cutoff of the astrophysical spectrum becomes diagnostically important (Brown et al., 2021, Wang et al., 2021). That overlap is central to Trinity’s comparative role: it is not only an extension beyond IceCube, but also an instrument for cross-calibration, flavor-composition studies, and multimessenger continuity across several decades in energy.

2. Earth-skimming $\nu_\tau$ detection principle

Trinity’s detection channel is the Earth-skimming tau neutrino. A $\nu_\tau$ enters the Earth on a shallow trajectory, undergoes a charged-current interaction in rock, and produces a $\sim 1$0 lepton sufficiently near the surface to emerge from the ground. The emerging $\sim 1$1 then decays in the lower atmosphere and initiates an upward-going extensive air shower whose Cherenkov emission is imaged from near-horizon directions (Brown et al., 2021, Wang et al., 2021). The observational baselines are long: design studies quote shower distances from roughly $\sim 1$2 km to $\sim 1$3 km, and later status reports continue to describe imaging distances up to approximately $\sim 1$4 km (Brown et al., 2021, Stepanoff et al., 22 Sep 2025).

The governing propagation scales are those of neutrino interaction, tau energy loss, and tau decay. One design study writes the neutrino interaction length in rock as

$\sim 1$5

with typical crustal density $\sim 1$6 (Brown et al., 2021). The tau decay length is written as

$\sim 1$7

with numerical examples of approximately $\sim 1$8 m at $\sim 1$9 PeV, $10$0 km at $10$1 PeV, $10$2 km at $10$3 PeV, and $10$4 km at $10$5 EeV (Brown et al., 2021). Tau energy loss in rock is parameterized as

$10$6

with the $10$7 term dominating at ultra-high energy and shaping the emergence probability (Brown et al., 2021).

The optical signature is set by the Cherenkov geometry. For air,

$10$8

with $10$9 and $10^7$0, giving $10^7$1 in one description and a degree-scale cone in later status work (Brown et al., 2021, Stepanoff et al., 22 Sep 2025). Near the horizon, the resulting image is thin and elongated across the camera, and timing information at nanosecond scale is used to distinguish upward-going fronts from downward-going or scattered-light backgrounds (Brown et al., 2021).

The treatment of emitted light evolved across the literature. The 2019 case studies described Trinity as using both Cherenkov and air-fluorescence light, and a 2021 sensitivity study still notes that fluorescence light can be used for showers developing within tens of kilometers (Otte et al., 2019, Wang et al., 2021). By contrast, the 2025 status report states that Trinity observes the Cherenkov light, not fluorescence, produced by these upward showers (Stepanoff et al., 22 Sep 2025). This suggests a narrowing of the operative design emphasis toward Cherenkov-only reconstruction in the implemented program.

3. Observatory architecture and instrument design

The core Trinity concept is a horizon-pointed wide-field IACT optimized for a narrow elevation band and a broad azimuthal span. A principal 2021 design paper specifies a field of view per telescope of $10^7$2, a spherical primary mirror composed of $10^7$3 square facets of $10^7$4 each, a total mirror area of $10^7$5, and a curved camera focal plane matched to the spherical optics (Brown et al., 2021). The optical requirement is to keep $10^7$6 of the light within one $10^7$7 pixel, and the per-pixel angular scale is given as approximately $10^7$8 (Brown et al., 2021).

The camera architecture is SiPM-based and modular. In that design, the camera follows the EUSO-SPB2 design philosophy and consists of $10^7$9 camera units with $\sim10^6$0 pixels each, for a total of $\sim10^6$1 SiPM pixels; each $\sim10^6$2-pixel unit is built from $\sim10^6$3 modules, and each module hosts $\sim10^6$4 SiPMs in a $\sim10^6$5 matrix (Brown et al., 2021). Solid PMMA light guides are placed in front of the SiPMs, MUSIC ASICs provide low-power amplification and shaping with per-channel bias setting and current monitoring, and AGET boards digitize at $\sim10^6$6 MS/s with $\sim10^6$7-bit resolution and implement the event trigger (Brown et al., 2021). The 2019 design studies likewise emphasized red-sensitive SiPMs, non-imaging light concentrators, MUSIC front-end electronics, and AGET switched-capacitor-array digitization at $\sim10^6$8 MS/s (Otte et al., 2019, Otte et al., 2019).

Trinity One is the first full-scale telescope in the staged program. Its 2025 design paper gives a $\sim10^6$9 light-collection surface, a $\sim 1$0 field of view, a $\sim 1$1-pixel SiPM camera with $\sim 1$2 resolution, and readout at $\sim 1$3 MS/s with $\sim 1$4-bit ADC resolution (O. et al., 22 Sep 2025). It can rotate in azimuth between $\sim 1$5 and $\sim 1$6, making it a point-source-capable horizon imager rather than a fixed-sector prototype (O. et al., 22 Sep 2025).

Atmospheric and instrumental calibration are treated as fundamental design elements because the line of sight runs through long, low-altitude slant paths. One design description lists three atmospheric calibration modes: continuous star photometry near the horizon, multi-wavelength light beacons at known distances, and periodic UAV-based calibration flights (Brown et al., 2021). The same paper notes that measured Cherenkov spectra redden with distance, using examples at $\sim 1$7, $\sim 1$8, $\sim 1$9, $\nu_\tau$0, and $\nu_\tau$1 km, and therefore require real-time atmospheric transmission corrections (Brown et al., 2021). Later status reports for future stages also anticipate expanded calibration and atmospheric monitoring suites typical of IACT-class systems (Stepanoff et al., 22 Sep 2025).

4. Acceptance, sensitivity, and astrophysical reach

Trinity sensitivity calculations are based on Earth-skimming propagation, detector geometry, and detector response. A 2021 sensitivity study writes the acceptance for diffuse flux as

$\nu_\tau$2

with the propagation implemented numerically using NuTauSim, including $\nu_\tau$3 energy losses and $\nu_\tau$4 regeneration (Wang et al., 2021). Event yields are then computed through

$\nu_\tau$5

A complementary design description gives the conceptually equivalent expression

$\nu_\tau$6

for benchmark flux models (Brown et al., 2021).

For diffuse astrophysical fluxes, Trinity’s low threshold of $\nu_\tau$7 GeV is intended to overlap IceCube’s band and extend measurements into the multi-PeV to EeV regime (Wang et al., 2021). The published benchmark scenarios are notable. One study reports that, for a power-law extrapolation of IceCube’s astrophysical flux without a cutoff, Trinity expects approximately $\nu_\tau$8 events over $\nu_\tau$9 years; for an astrophysical spectrum with an exponential cutoff consistent with IceCube’s upper limit at approximately $1$0 PeV, the same paper reports approximately $1$1 events over $1$2 years (Brown et al., 2021). The updated sensitivity study quotes approximately $1$3 events in the no-cutoff case and approximately $1$4 events over $1$5 years for a strong exponential cutoff consistent with IceCube upper limits around $1$6 GeV (Wang et al., 2021). The significance of these bracketing cases is that Trinity could statistically discriminate spectral cutoffs even without per-event energy reconstruction, by combining high-energy event counts with IceCube’s measured overlapping spectrum (Wang et al., 2021).

Point-source and transient performance is also quantified. For one Trinity site, the instantaneous acceptance on the sky is a ring below the local horizon with a full width at half maximum of about $1$7, peaking for elevations greater than $1$8 below the horizon (Wang et al., 2021). Under dark-sky constraints the assumed duty cycle is about $1$9, or approximately $\nu_\tau$0 h per site per year; for a source, the average observable duration per night is approximately $\nu_\tau$1 h, with a maximum of approximately $\nu_\tau$2 h (Wang et al., 2021). At Frisco Peak, Utah, a single site can observe declinations between about $\nu_\tau$3 and $\nu_\tau$4, with maximal exposure near $\nu_\tau$5 and $\nu_\tau$6 (Wang et al., 2021). For hard spectra with $\nu_\tau$7, the integral point-source sensitivity between $\nu_\tau$8 and $\nu_\tau$9 GeV reaches few $\nu_\tau$0 within approximately one hour at favorable declinations, and $\nu_\tau$1 within approximately one year (Wang et al., 2021).

Trinity One refines this source-oriented picture. Its 2025 design paper states that, from Frisco Peak, Trinity One can observe $\nu_\tau$2 of the sky and declinations approximately between $\nu_\tau$3 and $\nu_\tau$4 when visibility and operations are included (O. et al., 22 Sep 2025). For the 158-day TXS 0506+056 flare, the expected Trinity One yield is reported as $\nu_\tau$5, corresponding to a Poisson probability of about $\nu_\tau$6 to detect at least one neutrino (O. et al., 22 Sep 2025). The same paper states that a BOAT-like high-luminosity GRB would yield at least one detected $\nu_\tau$7 if it occurred within about $\nu_\tau$8 Gpc in the field of view, and that the cumulative probability to detect at least one neutrino from a short GRB within the field of view is about $\nu_\tau$9 over $\sim 1$00 years for a single Trinity One and about $\sim 1$01 for a partial multi-site array of six telescopes over the same interval (O. et al., 22 Sep 2025).

In the broader methodological context, Earth-traversing $\sim 1$02 regeneration is directly favorable to a PeV-threshold horizon Cherenkov instrument. A separate study on PeV tau neutrinos shows that including Earth-traversing directions can improve sensitivity to transient point sources above $\sim 1$03 GeV by more than an order of magnitude and improve sensitivity to the diffuse UHE flux by up to a factor of $\sim 1$04 (Argüelles et al., 2022). This is closely aligned with Trinity’s near-horizon operating geometry.

5. Trinity Demonstrator: commissioning, operations, and first data

The Trinity Demonstrator is the first deployed hardware stage of the observatory. It is a one-square-meter-class, horizon-pointing Cherenkov telescope located on Frisco Peak, Utah, at approximately $\sim 1$05 m above mean sea level (Bagheri et al., 14 Mar 2025). Its optics are a conventional Davies–Cotton design with $\sim 1$06 round mirror facets of $\sim 1$07 cm diameter and a total light-collection area of $\sim 1$08 (Bagheri et al., 14 Mar 2025, Stepanoff et al., 22 Sep 2025). The camera has $\sim 1$09 SiPM pixels, a field of view of approximately $\sim 1$10, and pixel angular resolution of approximately $\sim 1$11; signals are digitized at $\sim 1$12 MS/s with $\sim 1$13-bit resolution using an AGET switched-capacitor-array readout with eMUSIC signal conditioning (Bagheri et al., 14 Mar 2025, Stepanoff et al., 22 Sep 2025).

The commissioning paper reports remote autonomous operations, environmental monitoring, and a permanently horizon-facing configuration. In standard observing, the optical axis is set at azimuth $\sim 1$14, and the camera center points $\sim 1$15 below the horizon (Bagheri et al., 14 Mar 2025). Measured optical quality is consistent with the design purpose: a tuned simulation of the measured 2f alignment images implies an $\sim 1$16 containment radius of $\sim 1$17 mm at the focal plane for a point source at infinity, comfortably within one camera pixel (Bagheri et al., 14 Mar 2025). Gain and trigger flat-fielding are described quantitatively; the operational threshold is set to $\sim 1$18 DAC, corresponding to $\sim 1$19 photoelectrons, and stable trigger rates below approximately $\sim 1$20 Hz are reported for dark moonless nights (Bagheri et al., 14 Mar 2025).

The same paper provides an engineering and noise characterization of the horizon-pointing mode. The measured RMS noise per pixel is dominated by night-sky background, with totals of $\sim 1$21 photoelectrons for ground-view pixels and $\sim 1$22 photoelectrons for sky-view pixels (Bagheri et al., 14 Mar 2025). Direct muon hits through the camera are a known background, but they are confined to single pixels and can be rejected by requiring at least two adjacent pixels during image cleaning (Bagheri et al., 14 Mar 2025). Between June 1, 2024 and February 28, 2025, the commissioning paper reports $\sim 1$23 h of observing time, including $\sim 1$24 h during which TXS 0506+056 crossed the demonstrator field of view and $\sim 1$25 h for NGC 1068 (Bagheri et al., 14 Mar 2025).

A later status paper reports the first science-quality data sample. Construction was completed in Fall 2023, commissioning ended in Summer 2024, and science-quality operations were taken between October 2024 and May 2025 under a cloud-base requirement of at least $\sim 1$26 m above the site (Stepanoff et al., 22 Sep 2025). In $\sim 1$27 h of observations, the Demonstrator produced $\sim 1$28 total triggers, corresponding to about $\sim 1$29 Hz, and after cleaning and topology-based selection yielded $\sim 1$30 events with at least three pixels (Stepanoff et al., 22 Sep 2025). These surviving events were classified into four categories: $\sim 1$31 clean muon tracks, $\sim 1$32 spillover muon-like events, $\sim 1$33 fuzzy or broad events, and $\sim 1$34 events from a misbehaving module (Stepanoff et al., 22 Sep 2025). No $\sim 1$35 candidate event was identified (Stepanoff et al., 22 Sep 2025).

The absence of candidates should not be misread as evidence against the method. The Demonstrator was explicitly a pathfinder rather than a sensitivity-matched discovery instrument, and its main contribution was background characterization. The status paper argues that the upward-shower morphology remains distinctive relative to the dominant local muon background, and that a plastic-scintillator-based muon veto is planned for Trinity One to suppress these local tracks already at the trigger level (Stepanoff et al., 22 Sep 2025).

6. Trinity One, the 18-telescope observatory, and broader significance

The post-Demonstrator roadmap is three-stage. The first stage is the deployed Trinity Demonstrator; the second is Trinity One, the first full Trinity telescope with azimuthal rotation and point-source sensitivity; the third is the full Trinity Neutrino Observatory, envisioned as $\sim 1$36 telescopes distributed over at least three mountain sites (Stepanoff et al., 22 Sep 2025). The full observatory is described as providing wide sky coverage and high diffuse sensitivity at modest cost, stated as approximately $\sim 1$3720$$M <a href="https://www.emergentmind.com/topics/uncertainty-aware-semantic-decoding-usd" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">USD</a> in the 2025 status report (<a href="/papers/2509.18236" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">Stepanoff et al., 22 Sep 2025</a>).</p> <p>The scaling arguments are explicit. A 2021 sensitivity study assumes three sites at an altitude of about$$\sim 1$38 km, with multiple telescopes at each site providing combined $\sim 1$39 azimuthal coverage, and notes that sensitivity scales approximately linearly with the number of telescopes because the telescope fields do not overlap (Wang et al., 2021). The 2025 status report states similarly that diffuse sensitivity scales linearly with combined horizontal field of view and time; after $\sim 1$40 years, the full observatory’s diffuse sensitivity would be approximately $\sim 1$41 that of a single Trinity One after $\sim 1$42 years, from $\sim 1$43 telescopes and a factor of $\sim 1$44 longer operation (Stepanoff et al., 22 Sep 2025).

Trinity’s distinctiveness is architectural as well as scientific. It is a horizon-pointing Cherenkov imaging system, closer in observational logic to MAGIC’s $\sim 1$45 searches than to mountain-target concepts such as ASHRA, and it is intended to bridge the gap between IceCube/KM3NeT at the low-energy end and radio arrays such as BEACON, GRAND, RNO-G, and IceCube-Gen2 radio at the high-energy end (Stepanoff et al., 22 Sep 2025). A common misconception is that “background-free” Earth-skimming searches are literally free of instrumental or local backgrounds. Trinity publications use background-free assumptions for some sensitivity definitions, especially narrow transient windows, but the Demonstrator has already shown real local backgrounds from muon traversals, spillover events, and module malfunctions; what the geometry provides is strong suppression of the dominant downward-going cosmic-ray air-shower background, not a literal absence of non-neutrino events (Wang et al., 2021, Stepanoff et al., 22 Sep 2025).

Taken together, the Trinity literature defines a coherent observatory program rather than a single telescope design frozen in time. The early case studies established the imaging and sensitivity logic, the 2021 papers quantified overlap with IceCube and the ability to test spectral cutoffs, the Demonstrator validated remote operations and morphology-based background control, and the Trinity One design extends the concept into a full-scale point-source instrument with azimuthal slewing and $\sim 1$46 sky accessibility (Otte et al., 2019, Wang et al., 2021, O. et al., 22 Sep 2025). The full Trinity PeV-Neutrino Observatory is therefore best understood as a staged, horizon-pointed $\sim 1$47 observatory whose central scientific claim is not merely access to higher energies, but controlled coverage of the otherwise under-instrumented interval from approximately $\sim 1$48 PeV to $\sim 1$49 EeV.