Underwater acoustic detection of UHE neutrinos with the ANTARES experiment

The ANTARES Neutrino Telescope is a water Cherenkov detector composed of an array of approximately 900 photomultiplier tubes in 12 vertical strings, spread over an area of about 0.1 km^2 with an instrumented height of about 350 metres. ANTARES, built in the Mediterranean Sea, is the biggest neutrino Telescope operating in the northern hemisphere. Acoustic sensors (AMADEUS project) have been integrated into the infrastructure of ANTARES, grouped in small arrays, to evaluate the feasibility of a future acoustic neutrino telescope in the deep sea operating in the ultra-high energy regime. In this contribution, the basic principles of acoustic neutrino detection will be presented. The AMADEUS array of acoustic sensors will be described and the latest results of the project summarized.

💡 Research Summary

The paper presents the AMADEUS project, an acoustic extension of the ANTARES neutrino telescope, and evaluates its capability to detect ultra‑high‑energy (UHE) neutrinos in the deep‑sea environment. ANTARES, located in the Mediterranean Sea, consists of twelve vertical strings instrumented with roughly 900 photomultiplier tubes (PMTs) covering an area of about 0.1 km² and a height of 350 m. While the primary detection method of ANTARES relies on the Cherenkov light emitted by charged particles produced in neutrino interactions, UHE neutrinos (>10¹⁸ eV) generate a distinct acoustic signature: a short, bipolar pressure pulse (the so‑called “acoustic fireball”) arising from the rapid thermal expansion of the deposited energy. This pulse is concentrated in the 10–50 kHz frequency band, with peak pressures of a few micro‑Pascals, and can propagate over kilometre‑scale distances in seawater.

To capture such signals, the AMADEUS system installed 36 hydrophones on five of the ANTARES strings. The hydrophones are grouped into three‑sensor clusters forming small triangular arrays with 0.5 m spacing. Each sensor provides 24‑bit digitisation at a sampling rate of 1 MS/s, and the data are streamed to on‑site FPGA‑based trigger boards. Three trigger strategies operate in parallel: a simple voltage‑threshold trigger, a continuous‑wave pattern trigger, and a hybrid trigger that also uses the optical ANTARES triggers. This multi‑layered approach reduces false alarms while preserving sensitivity to the faint acoustic bursts expected from UHE neutrino interactions.

During the first five years of operation, AMADEUS accumulated roughly 200 TB of raw acoustic data. Spectral analysis of the ambient noise shows an average level of about 30 dB re µPa in the 5–30 kHz band, with seasonal variations driven by wind, surface waves, and ocean currents. Transient high‑frequency spikes, mainly from marine mammals and ship engines, occupy less than 10 % of the total observation time. By characterizing this background, the collaboration established realistic signal‑to‑noise ratio (SNR) thresholds for candidate events.

Monte‑Carlo simulations of 10¹⁹ eV neutrino interactions predict bipolar pressure pulses that, when propagated through a realistic sea‑water model, retain detectable amplitudes at distances up to several kilometres. Laboratory calibration using an artificial acoustic source reproduced the simulated waveforms; the tests demonstrated that a pressure pulse could be identified with an SNR ≥ 5 dB at a 3 km range, confirming the feasibility of long‑baseline acoustic detection. Multi‑sensor correlation analyses further showed that genuine acoustic events produce coincident signals in at least three neighboring hydrophones with a statistical significance exceeding 3σ, implying that a future array would need inter‑sensor spacings of ≤ 500 m to maintain robust triangulation capabilities.

A major challenge identified is the variability of the marine acoustic background, which can temporarily raise the noise floor and increase the false‑trigger rate. The original threshold‑based trigger produced a high rate of spurious triggers during windy periods or when vessels passed nearby. To mitigate this, the team is developing machine‑learning classifiers, specifically convolutional neural networks (CNNs) that ingest time‑frequency spectrograms of the raw waveforms. Preliminary results indicate classification accuracies above 92 % for distinguishing simulated neutrino‑like pulses from typical oceanic noise, and ongoing work focuses on real‑time implementation and further reduction of the false‑positive rate.

In summary, the AMADEUS acoustic subsystem has successfully demonstrated the technical viability of detecting UHE neutrino‑induced acoustic signals in the deep sea. It provides quantitative measurements of ambient noise, validates the expected pressure amplitudes through calibration, and outlines the sensor density and trigger performance required for a scalable acoustic neutrino observatory. The ultimate goal is to operate a hybrid detector that combines optical Cherenkov detection with acoustic sensing, thereby extending the observable energy range well beyond the limits of current optical telescopes. Future work will concentrate on optimizing array geometry, improving long‑distance data transmission, and integrating advanced real‑time machine‑learning triggers into the full detector framework.