WINTER on S250206dm: A near-infrared search for an electromagnetic counterpart to a gravitational-wave event

TL;DR

WINTER performed a systematic near-infrared follow-up of the GW event S250206dm to search for a kilonova counterpart. Despite tiling about $43\%$ of the localization probability and real-time data processing, no EM counterpart was detected at a distance of $373\,\mathrm{Mpc}$ with median $J$-band depths near $17.4$–$18.0$ mag (AB). The non-detection is broadly consistent with NSBH and BNS kilonova models under the observed conditions, underscoring that deeper and faster infrared surveys are required to constrain ejecta properties for distant mergers. The study confirms WINTER's role as the first wide-field NIR GW-follow-up instrument and outlines a path forward with next-generation facilities to improve detection prospects and ejecta characterization in multi-messenger astronomy.

Abstract

We present near-infrared follow-up observations of the International Gravitational Wave Network (IGWN) event S250206dm with the Wide-Field Infrared Transient Explorer (WINTER). WINTER is a near-infrared time-domain survey designed for electromagnetic follow-up of gravitational-wave sources localized to $\leq$300 deg$^{2}$. The instrument's wide field of view (1.2 deg$^2$), dedicated 1-m robotic telescope, and near-infrared coverage (0.9-1.7 microns) are optimized for searching for kilonovae, which are expected to exhibit a relatively long-lived near-infrared component. S250206dm is the only neutron star merger in the fourth observing run (to date) localized to $\leq$300 deg$^{2}$ with a False Alarm Rate below one per year. It has a $55\%$ probability of being a neutron star-black hole (NSBH) merger and a $37\%$ probability of being a binary neutron star (BNS) merger, with a $50\%$ credible region spanning 38 deg$^2$, an estimated distance of 373 Mpc, and an overall false alarm rate of approximately one in 25 years. WINTER covered $43\%$ of the probability area at least once and $35\%$ at least three times. Through automated and human candidate vetting, all transient candidates found in WINTER coverage were rejected as kilonova candidates. Unsurprisingly, given the large estimated distance of 373 Mpc, the WINTER upper limits do not constrain kilonova models. This study highlights the promise of systematic infrared searches and the need for future wider and deeper infrared surveys.

Figures (2)

• Figure 1: WINTER J-band coverage of S250206dm. The probability skymap is plotted in red overlaid with WINTER pointings shown as black rectangles. Each rectangle corresponds to a single WINTER detector and six detectors combine to produce a 1 deg x 1.2 deg FOV in a single WINTER pointing. Our observations covered a total of 43% of the skymap probability for this gravitational wave event, of which 40% was observed at least twice. Our observations comprised a total of 39 WINTER fields, a total on-sky time of 130 hours, and achieved median J-band limiting magnitudes of 17.6, 17.6, 17.4, 18.0, 17.5, and 17.4 mag (AB) on the six WINTER detectors respectively.
• Figure 2: WINTER observations compared to kilonova models for NSBH (left) and BNS (right) mergers from Ahumada et al., in prep, computed using the radiative transfer code possisBulla:2023. The kilonova light curve models are marginalized over the full parameter grid and 100 distances sampled from the event probability distribution function for the estimated distance. The median $J$-band light curve across all distance realizations and the full model grid is shown as a red line, while the shaded regions represent the 1$\sigma$ (68$\%$) and 3$\sigma$ (99.7$\%$) credible intervals. For comparison, the $r$-band light curve is shown from the same exercise to demonstrate the extended NIR plateau in kilonova models. The limiting magnitudes of WINTER observations are shown for comparison and do not constrain any of the presented models.