IceCube Real-time Searches for High-energy Neutrinos Coincident with LIGO/Virgo/KAGRA Gravitational-Wave Alerts in O4a

Technical Summary: IceCube Real-time Searches for High-energy Neutrinos Coincident with LIGO/Virgo/KAGRA Gravitational-Wave Alerts in O4a

Problem and Motivation
Compact binary coalescences (CBCs), including binary neutron star (BNS), neutron star–black hole (NSBH), and binary black hole (BBH) mergers, are predicted sources of high-energy (TeV–PeV) neutrinos. These events are expected to produce relativistic jets capable of accelerating cosmic rays, which subsequently interact to produce pions and decay into neutrinos. While the detection of GW170817 and its electromagnetic counterparts established the viability of multimessenger astronomy, the sources of the majority of astrophysical neutrinos remain unidentified. IceCube has previously searched for neutrino counterparts to gravitational wave (GW) events during observing runs O1, O2, and O3, finding no statistically significant emission. This paper reports on real-time searches conducted during the first part of the fourth observing run (O4a) of the LIGO-Virgo-KAGRA (LVK) network, aiming to identify coincident high-energy neutrinos and improve the latency of the response pipeline.

Methodology
The analysis utilizes data from the IceCube Neutrino Observatory, specifically the "Gamma-ray Follow-up" (GFU) dataset, which consists of high-energy track-like events reconstructed with low latency. The study covers 85 significant and 945 low-significance CBC candidate events published in real-time by LVK during O4a and the preceding engineering run (ER15). Two distinct analysis methods are employed:

1. Unbinned Maximum Likelihood (UML) Analysis: A generic transient search used for significant GW alerts. This method constructs a likelihood function summing signal and background probability density functions (PDFs) over the observed neutrino events. The signal PDF assumes a point source with a power-law energy spectrum ($E^{-\gamma}$) and incorporates the GW skymap probability ($w(\Omega)$). The test statistic (TS) is maximized over source location, signal count, and spectral index. P-values are derived by comparing the observed TS against a background-only distribution generated via pseudo-experiments where events are scrambled in right ascension and time.
2. Low-Latency Algorithm for Multimessenger Astrophysics (LLAMA): A model-dependent Bayesian approach applied to both significant and low-significance alerts. This pipeline incorporates astrophysical priors regarding the distribution of GW emission energies, the probability of events being astrophysical versus noise, and the isotropic equivalent energy of neutrino emission. The test statistic weights the likelihood of four hypotheses: true coincidence, chance coincidence with an astrophysical GW, chance coincidence with an astrophysical neutrino, and pure background.

The primary search window is set to $\pm 500$ seconds relative to the merger time, consistent with the expected duration of short gamma-ray bursts. Additionally, an extended search window of $[-0.1, +14]$ days is applied to three specific candidates: two NSBH mergers and one low-significance event with a potential subthreshold gamma-ray counterpart.

Key Contributions and Operational Improvements
A significant contribution of this work is the enhancement of IceCube's real-time operational infrastructure. The authors report the implementation of an automated system that ingests results from both the UML and LLAMA pipelines, merges them, and sends machine-readable GCN Notices automatically. This automation reduced the median latency of the IceCube response to GW events by a factor of 2.7 compared to the previous observing run (O3), achieving a median total latency of 21 minutes. The system also automatically releases coincident neutrino information (direction, error, timing) if pre-trial p-values fall below 0.1, and issues GCN Circulars for p-values below 0.01.

Results

• Short-Duration Search ($\pm 500$ s): No statistically significant neutrino emission was detected from any of the 85 significant or 945 low-significance GW candidates analyzed. The distribution of observed p-values for both the UML and LLAMA pipelines is consistent with background-only expectations. The lowest pre-trial p-value observed was 0.0011 for a low-significance candidate (S231025a), which, after accounting for trials across the full dataset, showed no evidence of a true coincidence.
• Long-Duration Search: The extended search for the three specific candidates (two NSBH mergers and one low-significance event) also yielded no significant emission.
• Upper Limits: For all analyzed events, the authors set 90% confidence level upper limits on the time-integrated neutrino flux and the isotropic equivalent energy ($E_{iso}$) emitted in high-energy neutrinos. These limits assume isotropic emission and a power-law spectrum. The $E_{iso}$ limits are calculated over the energy range 500 GeV to 50 PeV.
• Specific Event Follow-up: The event S230904n, a likely BBH merger, was followed up in real-time. While a potential counterpart (AT2023rkw) was identified by the Zwicky Transient Facility, subsequent spectroscopy identified it as a Type Ia supernova, unrelated to the GW event.

Significance and Claims
The paper claims that the addition of the 85 O4a events nearly doubles the number of candidate GW events analyzed with these searches compared to previous runs. The primary significance lies in the demonstration of a robust, low-latency real-time analysis framework capable of processing both significant and low-significance GW alerts. The authors emphasize that while no neutrino counterparts were found, the ability to perform these searches in real-time allows for immediate follow-up by other observatories, potentially increasing the significance of future candidates and refining localization. The results further constrain theoretical models of neutrino production in compact binary mergers by setting stringent upper limits on the energy emitted in high-energy neutrinos. The authors note that future improvements in GW detector sensitivity and the planned expansion to IceCube-Gen2 will increase the prospects for detecting such coincidences.