Search for gamma-ray bursts with the Antares neutrino telescope

Satellites that are capable of detecting gamma-ray bursts can trigger the Antares neutrino telescope via the real-time gamma-ray bursts coordinates network. Thanks to the “all-data-to-shore” concept that is implemented in the data acquisition system of Antares, the sensitivity to neutrinos from gamma-ray bursts is significantly increased when a gamma-ray burst is detected by these satellites. The performance of the satellite-triggered data taking is shown, as well as the resulting gain in detection efficiency. Different search methods can be applied to the data taken in coincidence with gamma-ray bursts. For gamma-ray bursts above the Antares horizon, for which a neutrino signal is more difficult to find, an analysis method is applied to detect muons induced by the high-energy gamma rays from the source.

💡 Research Summary

The paper presents the implementation and performance of a real‑time gamma‑ray burst (GRB) triggered data acquisition system for the ANTARES neutrino telescope. ANTARES, located at a depth of about 2.5 km in the Mediterranean Sea, detects high‑energy muon neutrinos via the Cherenkov light emitted by secondary muons traversing the instrumented water volume (~50 M m³). Traditionally, ANTARES employs an on‑site trigger that selects events with at least ten coincident photon hits, reducing the raw data flow (≈1 GB s⁻¹) to a manageable trigger rate of 5–10 Hz.

The key innovation described is the “all‑data‑to‑shore” concept, whereby every digitised photomultiplier signal is streamed to a shore‑based computing farm. This architecture enables the storage of the complete raw data stream for a short, predefined interval (currently two minutes) surrounding any GRB alert received from the Gamma‑ray Burst Coordinates Network (GCN). The GCN distributes real‑time alerts from satellites such as Swift and Fermi; upon receipt, ANTARES continues its standard filtering while simultaneously buffering and writing the unfiltered data to disk. Because the on‑shore PCs maintain a buffer of up to about one minute, the saved data can include information from before the satellite detection, yielding negative response times in some cases.

Operational statistics from autumn 2006 to early 2009 show that the satellite‑triggered acquisition was active for roughly 90 % of all GCN alerts, with typical response times centred around zero seconds. Delays larger than a few seconds result in incomplete coverage of the GRB emission window, but the system’s buffering capability often compensates for network latency.

Two complementary analysis strategies are explored. The standard approach uses the real‑time filtered data and performs a five‑parameter muon track fit (including two directional angles). The alternative, GRB‑specific method processes the stored unfiltered data offline, exploiting the known sky position of the burst to constrain the fit. By lowering the photon‑hit requirement from ten to six and applying directional cuts based on the GRB coordinates, the effective detection volume for neutrinos from a given burst increases dramatically. Simulations indicate a gain in detection efficiency of up to an order of magnitude compared with the standard pipeline (Fig. 4). This gain stems from the reduced trigger threshold, which is feasible only because the data are not required to be processed in real time.

For bursts located above the ANTARES horizon (i.e., downgoing events), where atmospheric muon background hampers neutrino searches, the authors propose a dedicated analysis to detect muons produced by high‑energy gamma rays from the GRB itself, following the method outlined in Guillard et al. (in preparation). This technique also benefits from the unfiltered data and is expected to provide a comparable improvement in sensitivity.

The authors conclude that the combination of the GCN real‑time alert system and the all‑data‑to‑shore architecture uniquely positions ANTARES to respond instantly to transient astrophysical phenomena. The ability to buffer large data volumes yields very short, sometimes negative, response times, ensuring maximal overlap between the neutrino detector’s observation window and the GRB emission. Looking ahead, the forthcoming KM3NeT neutrino telescope will adopt a similar trigger concept, promising even larger effective volumes and enhanced discovery potential for GRB‑associated neutrinos. The paper emphasizes that rapid distribution of GRB alerts (within a few tens of seconds) is essential to fully exploit the neutrino telescope’s capabilities.