This Week in Astronomy: a Stellar Merger Reconstructed, and Hera Opens Its Eyes
This week delivered an unusually satisfying mix of results. A stellar collision was reconstructed from beginning to end, a European spacecraft settled in alongside the asteroid we deliberately crashed into, a dozen specks of dust were traced to an ancient reservoir of ice beyond Neptune, and some clever mathematics turned one of astronomy's most stubborn searches into something a laptop could manage. All four arrive with measured numbers attached rather than promises, which is always the mark of a good week.
Webb clinches the first full reconstruction of a stellar merger
Astronomers have long believed that stars in close binary systems sometimes spiral together and merge. Watching it happen, and then proving it, is another matter entirely. A team led by Viraj Karambelkar, a NASA Hubble Fellow at Columbia University, has now done exactly that for an event in the outskirts of the Andromeda galaxy, our nearest large spiral neighbour at roughly two and a half million light years.
The object, formally AT 2025abao and nicknamed Bao, was caught erupting by the Zwicky Transient Facility on 17 October 2025. It brightened over about a day and peaked within a week. What makes the result so complete is the archive the team then assembled around it. The retired NEOWISE mission had recorded the star growing restless as early as February 2023. WINTER had it in September 2024, SPHEREx happened across it in July 2025, and Swift and Gemini North followed up in November. The decisive piece was a single James Webb Space Telescope observation taken on 24 December 2025, just 47 days after the peak, granted as Director's Discretionary Time.
Webb's mid infrared view revealed dust that the system had been shedding for centuries, and that record is what allowed the team to reconstruct the whole sequence. Simulations led by Karel Temmink at Radboud University matched a star of roughly seven solar masses, swollen to around 350 times the Sun's radius and some 18,000 times brighter, with a companion of about 0.7 solar masses. Once the two made contact, the smaller star completed only about seven orbits, roughly seven years, before being dragged into its partner's core.
Why this matters: common envelope evolution, the phase in which one star is engulfed by another, sits upstream of a great deal of astrophysics, from the binaries that become gravitational wave sources to certain classes of supernova. It has been modelled for decades with an efficiency factor nobody could measure, and here at last is a system where the bookkeeping can be checked against reality. The team is more cautious about the surviving star; the suggestion that it has been rejuvenated awaits later observations. Andromeda is a favourite target among Telescope Live users, and those wide field frames contain a laboratory as well as a portrait.
Reference: NASA Data Clinches First-of-Its-Kind Stellar Merger, NASA, 18 November 2026.
Hera opens its eyes at Didymos
In September 2022 NASA's DART spacecraft, all 580 kg of it, struck the small moonlet Dimorphos at 6.1 km per second, threw a plume of debris across some 10,000 km of space and shortened the moonlet's orbit by about 33 minutes. That was a far larger deflection than the impact simulations predicted, and nobody has been back to find out why. ESA's Hera mission has now arrived to do precisely that.
ESA released Hera's first survey images on 17 November. The Asteroid Framing Camera, shooting from 28.5 km out, resolves the 780 metre primary Didymos at about 2.7 metres per pixel as it spins once every 2.2 hours. Patrick Michel, the mission's Principal Investigator at CNRS and the Observatoire de la Côte d'Azur, picked out boulder strewn ground, clearer patches and "traces of what might be landslides towards its central ridge". The 151 metre Dimorphos is almost entirely in shadow in the visible frames, though TIRI, the thermal imager supplied by JAXA, caught the whole moonlet by its heat. HyperScout H is building up a 25 band view from the visible into the near infrared for comparison against laboratory samples.
The real questions stay open for now. Did DART leave a crater, or did it reshape Dimorphos entirely? What is the moonlet's mass and internal structure, and which of its properties produced that oversized deflection? Hera will spend the coming months on a close range investigation, assisted by two CubeSats, and ESA has said the detailed Dimorphos images will be unveiled on 25 November. If planetary defence is ever going to be engineering rather than hope, this is the dataset it rests on.
Reference: Hera surveying asteroids with superhuman vision, ESA, 17 November 2026.
Twelve grains of dust point to ancient ice beyond Neptune
Some of the week's best science came from a sample you could lose on a fingertip. Researchers at NASA's Johnson Space Center, working with colleagues at UCLA, analysed twelve interplanetary dust particles collected high in the stratosphere by NASA's ER-2 research aircraft. The density of solar energetic particle tracks in the grains let the team argue for an origin in the Kuiper Belt, out beyond Neptune.
Electron microscopy at Johnson identified minerals that can only have formed where water met dust, and an ion microprobe at UCLA measured the oxygen isotopes locked inside them, calibrated against the standard set by the Genesis mission. That water was enriched in the heavier isotopes of oxygen, which the team describes as the strongest evidence yet for a distinct and very ancient reservoir of ice in the outer Solar System. Equally interesting is what is missing: such signatures are rare in meteorites, hinting that the giant planets acted as a barrier and kept the outer and inner Solar System from mixing freely. That is a long reach from twelve specks of dust, and the claim will now be tested against every cometary sample available.
Reference: NASA Scientists Find Clues to Ancient Ice Beyond Neptune, NASA, 18 November 2026.
A pulsar search that got a million times easier
Finding pulsars in gamma ray data is a brutal computational problem. The signal is a faint periodicity buried in sparse photon arrival times, and the number of trial combinations is astronomical in the least helpful sense of the word. A team led by Daniel Pook-Kolb at the Max Planck Institute for Gravitational Physics, with Dotan Gazith and Barak Zackay at the Weizmann Institute and Colin Clark at Max Planck, has reformulated the search using lattice methods, and the improvement is remarkable.
Applied to 15 years of data from NASA's Fermi Gamma-ray Space Telescope, the new approach turned up PSR J1622-7204, an isolated pulsar spinning 261 times every second. The older method had failed to find it, and would have demanded a computation equivalent to 600 years on a standard processor. The new one did the job in about ten hours of equivalent single processor time. NASA's summary of about a million times faster is a generous rounding of the team's own figures, but the gain is genuine and very large.
The attraction is that all 18 years of Fermi observations are public, so the haystack is already in hand. Clark suggests there are "probably tens or maybe even hundreds" of these objects still sitting unrecognised in the archive, and the team's next target is the harder case of pulsars in binary systems, where orbital motion smears the signal further. There are plans to pair the technique with the Einstein@Home volunteer computing network.
Reference: NASA Science Data Powers New Method to Discover Hard-to-Find Pulsars, NASA, 19 November 2026.
Clear skies, and we shall see you next week.