J-GEM near-infrared follow-up observations of the gravitational wave event S240422ed

TL;DR

This study reports a targeted near-infrared follow-up of the gravitational-wave event S240422ed using Subaru/MOIRCS, observing 206 nearby galaxies to search for a red kilonova at ~200 Mpc. Despite the initial BH-NS classification and high EM-emission probability, the analysis identified five candidate transients, none of which match the expected kilonova signature, with interpretations ranging from reddened supernovae and red stars to potential Galactic or solar-system contaminants. The work demonstrates how deep $Y$ and $K_{ m s}$-band data can constrain kilonova models, showing that non-detections at 0.4 and 1.4 days limit combinations of ejecta mass $M_{ m ej}$ and opacity $\\kappa$, and that future deep NIR reference data and wide-field surveys will improve GW counterpart searches. The findings offer practical insights for future NIR follow-ups with facilities like WINTER, PRIME, Euclid, and Roman, particularly regarding reference image depth, Galactic-plane contamination, and the need for multi-epoch observations to distinguish transients from moving or glinting objects.

Abstract

We report our near-infrared (NIR) follow-up observations of the gravitational wave (GW) event S240422ed using the Subaru Telescope/MOIRCS. S240422ed was initially classified as a black hole-neutron star merger with $>$ 99% probability of electromagnetic wave emission. We started follow-up observations 7.8 hours after the event. Over two nights, we observed 206 nearby galaxies in $Y$ and $K_{\rm s}$ bands down to about 21.4 and 21.1 AB mag (3$σ$), respectively. The total completeness of our survey based on galaxy $B$-band luminosity is 22%. As a result of our observations, five candidate counterparts were identified. We show that properties of these five objects are not consistent with kilonova such as AT2017gfo. Four objects are consistent with known classes of transients such as supernovae or dwarf nova outbursts. On the other hand, the nature of the remaining one object, which shows a red color and rapid decline, remains unclear. Although later analyses of GW signal reclassified S240422ed as likely terrestrial noise, our NIR observations provide valuable lessons for future NIR surveys for GW sources. We demonstrate that deep NIR follow-up observations as presented in this work would effectively constrain the presence of red kilonova even at 200 Mpc distance. We also discuss the importance of deep and wide NIR reference images and of understanding the properties and frequency of Galactic transients.

Figures (8)

• Figure 1: Sky localization map of the gravitational wave event S240422ed, overlaid with the positions of galaxies targeted in our follow-up observations. The background shows the probability density map derived from the 3D localization of the gravitational wave. The 50% and 90% confidence regions are indicated by labeled solid contours. The positions of the galaxies observed in our follow-up are shown as blue scatter points. Alt text: Sky map of a gravitational wave event showing probability contours and observed galaxy positions. The horizontal and vertical axes represent right ascension and declination in equatorial coordinates.
• Figure 2: Cutout images of the EM counterpart candidates identified in our follow-up observations, including reference images (columns 1--3). Each row corresponds to one candidate (from top to bottom: J-GEM24a, 24b, 24c, 24e, and 24f), while each column (from left to right) shows: (1) a DECaPS2 color-composite image created from $g$, $r$, and $z$ bands; (2) a $Y$-band reference image from DECaPS2; (3) a $K_{\rm s}$-band reference image taken from either VISTA or VLT archival data; and (4–7) MOIRCS follow-up images taken on two nights in the order of Day 1 $K_{\rm s}$, Day 1 $Y$, Day 2 $K_{\rm s}$, and Day 2 $Y$. The white square in the DECaPS2 $grz$-band panel indicates the $16.8{}^{\prime\prime} \times 16.8{}^{\prime\prime}$ square field, corresponding to the cutout images to its right. For J-GEM24a, the $K_{\rm s}$-band reference image is substituted with a follow-up image obtained by VLT six days after our observations. Alt text: Cutout images of five electromagnetic counterpart candidates. Rows represent individual candidates. Columns show archival images in optical and near-infrared bands, followed by MOIRCS follow-up images from two nights.
• Figure 3: Photometric measurements of the EM counterpart candidates (J-GEM24a, 24b, 24c, 24e, and 24f) from our follow-up observations. The orange and dark red points represent detections in the $Y$ and $K_{\rm s}$ bands, respectively, while downward arrows indicate 3$\sigma$ upper limits for non-detection. An open (white-filled) marker indicates a marginal detection with S/N between 3 and 5. Horizontal dashed lines show the 3$\sigma$ limiting magnitudes of the reference images in each band. Alt text: Photometric measurements of five candidate objects, each shown in a separate panel. Magnitudes in Y band and K sub s band are plotted for two nights. Symbols indicate detections, a marginal detection, and upper limits.
• Figure 4: Color evolution of the EM counterpart candidates (J-GEM24a, 24b, 24c, 24e, and 24f from top to bottom), shown as $Y - K_{\rm s}$ versus time since the GW event. Filled circles indicate color measurements, while arrows represent color limits due to non-detections in either band. Alt text: Color evolution of five candidate objects shown in separate panels. Each panel plots Y minus K sub s color over time. Symbols indicate measured colors or limits due to non-detections.
• Figure 5: Comparison of the absolute magnitudes of the EM counterpart candidates with those of kilonova AT2017gfo and a typical Type Ia SN. The upper panel shows J-GEM24a, while the lower panel displays the other candidates (J-GEM24b, 24c, 24e, and 24f). Orange and dark red symbols represent measurements in the $Y$ and $K_{\rm s}$ bands, respectively. Solid lines connect detections, while dashed lines indicate connections involving upper limits. Open circles in the upper panel denote upper limits from Gemini and VLT reported via GCN Circulars; the dark red circle corresponds to the VLT $K_{\rm s}$-band observation, while the other four points are from Gemini, shown in purple ($g$ band), blue ($r$ band), green ($i$ band), and olive ($z$ band) respectively. Alt text: Two-panel figure comparing absolute magnitudes of five candidate objects with those of a kilonova and a Type Ia supernova. The upper panel shows one candidate, and the lower panel shows four others. Time since the gravitational wave event is on the horizontal axis. Symbols indicate detections or upper limits in multiple photometric bands, including our follow-up observations and archival constraints reported from other telescopes.