AT 2025abao: the fourth luminous red nova in M 31

TL;DR

AT 2025abao is a nearby luminous red nova in M31 with a seven-year pre-outburst rise, enabling direct progenitor constraints on an AGB star in a common-envelope phase. The outburst displays a fast blue rise followed by a long, 70-day red plateau, with a peak bolometric luminosity of about $L_{bol}\sim10^{39.5}$ erg s$^{-1}$, making it the brightest LRN in M31; its spectral evolution tracks the canonical LRN sequence, including a blue continuum with narrow Balmer emission, cooling to $T_{BB}\sim6000$ K with metal absorption lines, a high-velocity Ca II H&K forest, and a later counter-P Cygni Hα profile indicating asymmetric outflows and CSM structure. SPHEREx pre-outburst spectroscopy reveals an M-type AGB-like precursor with a dusty environment, while multi-epoch SEDs imply a photosphere of order $T_{BB}\sim4000-6000$ K and radii of several thousand solar radii during the plateau, consistent with hydrogen recombination in an extended envelope. The authors argue that the light-curve dichotomy among LRNe (two peaks versus plateau) reflects differences in common-envelope extent and hydrogen content, with implications for progenitor masses, event rates, and the post-merger remnant evolution. Together, the work provides a comprehensive, multi-wavelength view of a nearby LRN and strengthens the link between progenitor properties, envelope geometry, and observable transients.

Abstract

We present photometric and spectroscopic observations of luminous red nova (LRN) AT 2025abao, the fourth discovered in M 31. The LRN, associated to the AGB star WNTR23bzdiq, was discovered during the fast rise following the minimum phase. It reached the peak at $g=15.1$ mag ($M_g=-9.5\pm0.1$ mag), and then it settled onto a long-duration plateau in the red bands, lasting 70 days, while it was slowly linearly declining in the blue bands. The object showed similarities at peak with the canonical LRNe V838 Monocerotis, V1309 Scorpii, and with the faint and fast-evolving AT 2019zhd, the third LRN in M31, though the later evolution is different. Spectroscopically, AT 2025abao evolved as a canonical LRN: the early spectra present a blue continuum with narrow Balmer lines in emission; at peak, the spectral continuum has cooled to a yellow colour, with a photospheric temperature of 6000 K. Balmer lines have weakened while absorption lines from metals (Fe I, Fe II, Sc II, Ba II, Ti II) have developed, and in particular broad (FWHM$\sim$700 km/s) from the UV Ca II H&K lines. Medium- and high-resolution spectra reveal a counter-P Cygni absorption profile in H$α$. Finally, late time spectra show an orange continuum ($T\sim$4000-5000 K), a return in strength of the Balmer lines and the formation of molecular absorption bands. AT 2025abao is the rare case of a LRN with detailed archival information regarding the progenitor system. For the first time, we obtained the spectral energy distribution in the infrared of the precursor of a LRN, which is consistent with that of an M giant/AGB. We propose that the dichotomy of light curve behaviour in LRNe (two peaks vs. plateau) can be explained by the extent and H-richness of the common envelope.

Figures (17)

• Figure 1: Finding chart of AT 2025abao. The colour image is a composition of $B$, $V$, and $r$ frames obtained by the Asiago 67/92cm Schmidt telescope on 7 November 2025, at the time of the maximum light. The spiral disk of M 31 is visible towards the left edge of the frame. The transient location is highlighted in the zoom-in panel to the right. Scale and orientation are reported.
• Figure 2: Optical light curves of AT 2025abao. Top panel: the entire evolution since October 2024, when the follow-up campaign of WNTR23bzdiq conducted by Karambelkar2025ApJ...993..109K ended. The pre-LRN light curve, its decline to the minimum, and the final rapid rise to the LRN peak are visible. Bottom panel: zoom in on the fast rise towards the maximum light, the 70 days-long plateau, and the fall from it.
• Figure 3: Colour curves of AT 2025abao. Top panel: Evolution of the $g-r$ colour from $-1.1$ yr to $+3$ months. The colour changes prominently between the slow pre-outburst rise and optically thick phase, the fast rise and the first peak. Bottom panel: Evolution of multiple colour curves ($B-V$, $g-r$, $r-i$, $u-g$, $u-i$) around the maximum light and plateau.
• Figure 4: Absolute light curve of AT 2025abao in the $r$ band, compared with those of AT 2019zhd ($r$-band), V838 Mon ($V$-band), and V1309 Sco ($I$-band). The light curve of V1309 Sco is scaled upwards by 2.5 magnitudes to match the peak brightness, and to see the pre-outburst slow rise of both V1309 Sco and AT 2025abao. The error bar on the distance modulus of V1309 Sco (0.724 mag) is not shown.
• Figure 5: Pseudo-bolometric (just $g, c, V, r, o, I$ bands) light curve of AT 2025abao, in black, and UV+Optical+NIR ($UVW2$ to $K$ bands) bolometric curve of the main outburst, in red. The separation is made at phase $-3.5$ d, when our multi-band photometric follow-up campaign started (after this phase, the $gcVroI$ pseudo-bolometric is shown as a dashed line). While the bolometric luminosity of AT 2025abao was similar at the time of the $g$-band maximum to those of the comparison objects, the subsequent evolution is different.