A neutrino flare candidate potentially associated with X-ray emission from tidal disruption event ATLAS17jrp

Tidal disruption events (TDEs), in which stars are disrupted by supermassive black holes, have been proposed as potential sources of high-energy neutrinos through hadronic interactions. X-ray-bright TDEs provide dense photon fields conducive to neutrino production via proton-photon ($pγ$) processes. We conducted a time-dependent unbinned likelihood analysis of ten years (2008-2018) of IceCube muon-track data, focusing on ten TDEs with confirmed X-ray detections during this period. We report a neutrino flare candidate spatially and temporally coincident with the TDE ATLAS17jrp, occurring 19 days after the onset of its X-ray activity and lasting for 56 days, with a post-trial $p$-value of 0.01. This significance is modest, representing a hint of an association. We illustrate the neutrino emission using a simple lepton-hadronic model, where X-ray photons serve as target fields. While this model can account for the neutrino data around 100 TeV, the low-energy neutrinos may imply contributions from an additional component. Although constrained by the sample size of X-ray-detected TDEs, these results underscore the need for high-cadence X-ray monitoring and future neutrino observatories to further explore the connection between TDEs and high-energy neutrinos.

💡 Research Summary

The authors present a systematic, time‑dependent search for high‑energy neutrino emission from tidal disruption events (TDEs) that were observed in X‑rays between 2008 and 2018. Using the publicly released IceCube muon‑track dataset spanning April 2008 to July 2018 (1 134 450 events), they focus on a curated sample of ten X‑ray‑bright TDEs, including ATLAS17jrp, ASASSN‑14li, ASASSN‑15oi, OGLE16aaa, PS18kh, SDSS J1201+30, 2MASX 0740‑85, XMMSL2 J1446+68, Swift J1644+57, and Swift J2058+05. For each source they compile redshift, peak bolometric, X‑ray and infrared luminosities, peak X‑ray flux, and black‑hole mass estimates.

The analysis employs an unbinned maximum‑likelihood method that models the data as a mixture of signal and background. The signal probability density function (PDF) is factorised into spatial, energy, and temporal components. Spatially, a two‑dimensional Gaussian centred on the source position with the event‑by‑event angular uncertainty is used. Energetically, a simple power‑law neutrino spectrum (dN/dE ∝ E⁻ᵞ, with 1 ≤ γ ≤ 4) defines the signal energy PDF, mirroring standard IceCube point‑source searches. Temporally, a box‑shaped PDF of width T_W and centre T₀ is adopted, allowing the search to concentrate on transient emission periods and thereby reduce atmospheric background. The background PDFs are taken from IceCube’s published characterisations and are essentially uniform in right ascension and time.

For each TDE the likelihood ratio test statistic (TS) is maximised over four free parameters: the number of signal events n_s, the spectral index γ, the flare centre T₀, and the flare duration T_W (constrained to be shorter than the total electromagnetic activity window). The TS is corrected for the look‑elsewhere effect by the factor T_W/T, where T is the total duration of the X‑ray emission.

Among the ten candidates, only ATLAS17jrp yields a significant excess. The best‑fit parameters are T₀ ≈ MJD 58173 (≈19 days after the onset of X‑ray activity), T_W ≈ 56 days, n_s ≈ 2.5, and γ ≈ 2.3, resulting in TS ≈ 9.99. To assess significance, the authors perform 10 000 pseudo‑experiments by scrambling event times and right ascensions, obtaining a post‑trial p‑value of 0.01 (≈2.6σ). This constitutes a modest hint rather than a definitive detection.

To interpret the neutrino signal, the authors construct a simple lepto‑hadronic model in which the dense X‑ray photon field serves as the target for proton‑photon (pγ) interactions. They approximate the X‑ray spectrum as a blackbody with temperature ∼0.1 keV plus a power‑law tail, and assume a proton spectrum ∝ E_p⁻² extending up to ∼10 PeV. The Δ‑resonance condition then predicts neutrinos peaking around 100 TeV, consistent with the two to three IceCube events that dominate the flare. The model reproduces the high‑energy part of the observed neutrino spectrum but under‑predicts the lower‑energy (∼10–30 TeV) events, suggesting an additional component such as pp collisions in a dense outflow or contributions from multiple emission zones (corona, wind, or jet).

The paper discusses several limitations: the small sample size (only ten X‑ray‑bright TDEs), uncertainties in the exact start and end times of the X‑ray emission, and IceCube’s angular and energy reconstruction errors. ATLAS17jrp’s relatively low redshift (z = 0.066) and black‑hole mass (≈6.5 × 10⁶ M_⊙) make it a favorable case for dense photon fields, yet the temporal alignment between the X‑ray light curve and the neutrino flare remains uncertain.

In conclusion, the study demonstrates that a time‑dependent, unbinned likelihood approach can reveal transient neutrino excesses associated with TDEs and that X‑ray‑bright TDEs are viable sites for pγ‑driven neutrino production. The modest significance of the ATLAS17jrp flare underscores the need for higher‑cadence X‑ray monitoring (e.g., eROSITA, Athena) and next‑generation neutrino observatories (IceCube‑Gen2, KM3NeT) to increase sample sizes, improve localisation, and ultimately confirm or refute the TDE–neutrino connection.