Constraints on the VHE counterpart of two binary black hole mergers observed by the MAGIC and CTAO LST-1 telescopes

Technical Summary: Constraints on the VHE Counterpart of Two Binary Black Hole Mergers

Problem and Context
While the detection of electromagnetic (EM) counterparts to binary neutron star (BNS) mergers has become established (e.g., GW170817), no EM counterpart has been observationally confirmed for binary black hole (BBH) mergers. Theoretical models suggest that BBH mergers embedded in the dense accretion disks of active galactic nuclei (AGN) could produce detectable EM emission, including very-high-energy (VHE) gamma rays via synchrotron and inverse Compton processes. However, the lack of confirmed detections leaves the physical conditions required for such emission unconstrained. This paper addresses the challenge of searching for VHE counterparts to BBH mergers, specifically focusing on two candidates from the fourth observing run (O4b) of the LIGO-Virgo-KAGRA (LVK) collaboration: GW240615_113620 (GW240615) and GW241125_010116 (GW241125).

Observational Campaign and Data
The authors conducted follow-up observations using the Major Atmospheric Gamma Imaging Cherenkov (MAGIC) telescopes and the first Large-Sized Telescope (LST-1) of the Cherenkov Telescope Array Observatory (CTAO). Both telescopes are located at the Roque de los Muchachos Observatory in La Palma.

• GW240615: This event was the best-localized BBH merger of the LVK runs, with a 90% confidence area of only 5 deg². Observations began ~14.9 hours post-trigger. Due to the small localization area, the entire region of interest fit within the telescopes' fields of view (FoV), allowing for a coordinated "wobble" observation strategy.
• GW241125: This event had a larger localization area (76 deg²) but was temporally and spatially coincident with a sub-threshold short-duration burst detected by Swift-BAT, Swift-XRT, and the Einstein Probe FXT. Observations began ~19 hours post-trigger, targeting the Swift-BAT candidate coordinates.

Methodology
The paper details a joint analysis pipeline and specific adaptations required by the observing conditions:

1. Joint Stereo Analysis: The authors combined data from MAGIC and LST-1 using the magic-cta-pipe package. MAGIC data were converted to the ctapipe format to enable a unified stereoscopic reconstruction, improving energy and direction resolution compared to single-telescope observations.
2. Handling Uncertain Localization (GW240615): Since the source localization was not a single point but a probability map, the authors employed a frequentist approach (following Hussain et al. 2020). They modified the standard test statistic ($TS$) by weighting the likelihood of each sky pixel with the LVK 2D probability map ($P^{GW}_i$). This created a modified statistic ($TS'$) to derive a global significance and flux upper limit (U.L.) conditional on the source being within the observed region.
3. Handling Suboptimal Weather (GW241125): Observations of GW241125 were severely affected by cloud cover, reducing atmospheric transmission. Standard analysis chains were insufficient. The authors implemented a tailored analysis using Monte Carlo (MC) simulations adapted to the specific atmospheric conditions measured by a LIDAR system. These simulations incorporated cloud transmission profiles (derived from LIDAR and keogram data) into the sim_telarray and CORSIKA frameworks to correct for energy bias and effective area reduction. Only data from windows with higher transmission were retained, and a conservative energy cut of 0.5 TeV was applied.
4. Background Modeling: For both events, background estimation utilized a 3D model (pybkgmodel) and the ring background method, normalized over the dataset.

Key Results

• No Detection: No significant gamma-ray emission was detected in the GeV-TeV band for either event.
• GW240615 Upper Limits:
  • In the low-energy range (0.15–0.6 TeV, LST-1 only), the 95% confidence level (CL) flux upper limit is $5.17 \times 10^{-11}$ cm$^{-2}$ s$^{-1}$.
  • In the high-energy range ($E > 0.6$ TeV, MAGIC+LST-1), the 95% CL flux upper limit is $4.47 \times 10^{-12}$ cm$^{-2}$ s$^{-1}$.
  • The global significance of the signal was low (1.01$\sigma$ and 1.43$\sigma$ for the respective ranges).
• GW241125 Upper Limits: Due to the poor weather conditions and the resulting reduction in effective area, the analysis yielded non-detections with significantly higher flux upper limits compared to standard conditions. The paper notes that the data quality precluded a standard analysis, necessitating the custom approach described.
• Theoretical Context: The authors discuss these upper limits in the context of theoretical models predicting VHE emission from BBH mergers in AGN disks. They note that while synchrotron emission alone cannot produce GeV-TeV photons, inverse Compton scattering could. The non-detection places constraints on the luminosity of such potential counterparts.

Significance and Claims
The paper claims several technical and observational milestones:

1. First Coordinated IACT GW Follow-up: This represents the first time observations were scheduled in a coordinated manner between different Imaging Atmospheric Cherenkov Telescopes (IACTs) for a GW event, leveraging the joint MAGIC+LST-1 stereoscopic configuration.
2. Methodological Advances: The study demonstrates the feasibility of analyzing GW counterparts with uncertain localization using a weighted frequentist approach that integrates GW probability maps directly into the gamma-ray significance calculation. Furthermore, it showcases a robust methodology for analyzing VHE data taken under suboptimal, cloudy conditions by adapting MC simulations to real-time atmospheric monitoring data.
3. Constraints on BBH Models: By failing to detect VHE counterparts for these two specific events (one with excellent localization, one associated with a GRB-like burst), the paper provides empirical constraints on theoretical models that predict bright VHE emission from BBH mergers, particularly those occurring in AGN disks. The authors emphasize that these events, due to their specific characteristics (small localization area and GRB-like association), offered an unprecedented opportunity to test these models, even though the results were null detections.

The paper concludes modestly, stating that while no detection was achieved, the successful execution of the joint observation strategy and the development of analysis techniques for challenging conditions lay the groundwork for future, more sensitive searches as the CTAO array expands.