We knew something heavy orbited four nearby stars back in 1997, so why did it take until 2026 to see one?
Welcome to FreeAstroScience. We went through the MNRAS paper’s own caveats rather than the summary, so the numbers below arrive with the limits the authors put on them.
By the team’s count, twenty-seven years separate the first solid clue from the ultraviolet spectrum Hubble recorded on 25 March 2026. That star sits 24.8 light-years away, ninth on the list of nearest white dwarfs, and its coordinates were on file for most of that time. Nobody could see it, because at the wavelengths our eyes and most telescopes work in, the star next to it is simply brighter.
We updated this piece on 28 July 2026 with our own check of the local space density, the detection floor the authors report, and a correction to the way the result is being described elsewhere.
Four dead stars, suspected since the 1990s
A white dwarf is what a star like the Sun leaves behind: a core roughly the size of Earth carrying half a solar mass or more, cooling for billions of years with no fusion left to run. We covered that whole sequence in our guide to how stars form, live and end, so this piece picks up after the funeral.
Mairi O’Brien at the University of Warwick led the new work, with David Wilson at the University of Colorado Boulder. Monthly Notices of the Royal Astronomical Society accepted it on 19 June 2026. Their paper reports the first direct spectroscopic detection of the white dwarf inside four binary systems within 20 parsecs of the Sun, which is 65 light-years. Each of those objects orbits a red dwarf bright enough at visible wavelengths to erase it.
None of the four arrived out of nowhere. Neil Reid and John Gizis flagged G 203-47 as a spectroscopic binary in 1997, and Xavier Delfosse’s team showed two years later that the unseen partner carried at least half a solar mass, too heavy for a planet and too dark for an ordinary star. Gizis had already suspected a degenerate companion around Wolf 1130 in 1998. Jennifer Winters’ group caught the motion of LHS 1817 in 2020, and the CARMENES survey did the same for GJ 207.1 in 2021. One detail worth pinning down: the twenty-seven-year clock starts at the 1999 mass limit, not at the 1997 sighting of the binary.
Table 1: the four systems, with distances converted from Gaia parallaxes (O’Brien et al., June 2026).
| System | Distance | Orbital period | Temperature |
|---|---|---|---|
| G 203-47 | 24.8 light-years | 14.9 days | about 5,300 K |
| GJ 207.1 | 51.5 light-years | 14.5 hours | about 6,100 K |
| LHS 1817 | 53.1 light-years | 7.4 hours | about 6,300 K |
| Wolf 1130 | 54.1 light-years | 11.9 hours | about 6,100 K |
Wolf 1130 is the strange one, and it was not on our radar before this paper landed on the desk. Its primary is a metal-poor subdwarf, the burnt-out core beside it carries between 1.1 and 1.4 solar masses, close to the ceiling for a white dwarf, and a cold brown-dwarf-like third body orbits far outside the inner pair. We are leaving that one alone here. It does not affect the mass estimate, and it deserves an article of its own.
Objects buried inside what looks like a single point of light are an old story in this field. Alnitak concealed an extra star until 1998, as we set out in our look at the triple system in Orion’s Belt, and Alnitak is one of the brightest things in the winter sky.
Ultraviolet was the only way in
Each of these white dwarfs runs roughly twice as hot as the red dwarf it orbits. At visible wavelengths that buys nothing, because the larger star wins on total output. Move to the ultraviolet and the ranking inverts.
So the team pointed Hubble’s STIS spectrograph at all four systems, hunting for the smooth ultraviolet glow of something compact sitting beneath a brighter partner. For the quiet G 203-47 they used the well-calibrated G230L grating. GJ 207.1 and LHS 1817 flare often enough that the same optics risked damage, so a sturdier grating went in, and that one scatters stray red light into the measurement. To get around it, the researchers observed G 203-47 through both settings, compared the two results, and built their correction out of the difference.
Flares were the second trap, since a strong outburst on a red dwarf lifts the ultraviolet in a way that can impersonate a white dwarf. Swift watched two of the targets at the same moment Hubble did, and its light curves stayed flat. The team then masked the emission lines that stellar activity produces, which left the clean continuum of each object behind, and four temperatures landed between roughly 5,300 and 6,300 kelvin.
Two caveats in that paper deserve more attention than they have been getting. When the authors fitted broadband photometry alone, without Hubble spectra, temperatures came out 5 to 8 percent too high, inflated by the red dwarfs’ own chromospheric glow. Detection has a floor too: anything cooler than about 4,700 to 5,500 kelvin, depending on the system, would have slipped past unseen. So this census is a lower bound, and the paper says as much in its own discussion instead of burying it.
G 203-47 refuses to spin on schedule
The red dwarf circles its white dwarf every 14.9 days. Rotation takes 100 days or more, with a tentative photometric signal near 126 days. Divide one into the other and the star completes about eight and a half orbits for every single turn on its axis. In a pair this tight, tides should have pulled those two numbers together ages ago, the way the Moon keeps one face turned toward us.
Wilson and his colleagues read the sluggish spin as a fossil. Binaries like this form when the future white dwarf swells into a giant and swallows its neighbor inside a shared envelope of gas. Friction within that envelope drags the two stars closer before the gas is thrown clear. Systems that survived a long, violent envelope phase emerged tightly bound and tidally locked, while G 203-47, the authors argue, got away with something gentler. Its core has been cooling for roughly four billion years.
Plenty of time to lock, if locking were ever going to happen.
Our own Sun takes this same road, and we walked through every stage of the journey in our stage-by-stage account of the Sun’s red giant phase. What the O’Brien paper adds is precisely the part we could not cover there, since our star has nothing nearby to swallow.
How many are still missing?
Nine or ten more, by the team’s own estimate, inside the same 65 light-years.
With the four newcomers counted, confirmed white dwarfs within 20 parsecs reach 153. Only about 30 percent of the more than 2,000 red dwarfs in that volume have been watched with instruments precise enough to register a wobble. Pier-Emmanuel Tremblay of Warwick puts the likely yield of a dedicated search of the rest at nine or ten further pairs.
One figure here does not reconcile, and we spent a while trying before we gave up. A sphere of radius 20 parsecs encloses about 33,500 cubic parsecs, so 153 of them work out to 4.6 per thousand cubic parsecs, or one in every 219. The density quoted in the coverage is 5.2 per thousand, one in 190. Those two published values agree with each other and not with a count of 153, which points to a completeness correction we cannot see from the material in front of us. Printing a number we had not checked would have been easier.
What does hold up is more interesting. Population models predicted about four or five close white dwarf and red dwarf pairs in this bubble. Observers found four. A prediction that specific about our own neighborhood is a harder test than it sounds, and we think that match, rather than any single detection, is the durable result of this work.
Gaia’s fourth data release should shorten the wait for whatever is left. Its positional precision will expose the remaining astrometric wobblers without anyone tracking radial velocities for a quarter of a century.
Where we think this is being oversold
The line doing the rounds is that four dead stars were hiding in our backyard and have now turned up. That is not quite what happened. Three of these four had an unseen companion of known minimum mass, in G 203-47’s case since the 1990s, and no one seriously expected them to be anything but white dwarfs. What landed in June 2026 is confirmation, plus something more useful: temperatures, and the cooling ages those imply. Calling it a discovery gives the old velocity measurements less credit than they earned and makes the ultraviolet work sound easier than it was.
- Four systems, all within 65 light-years.
- G 203-47 is the ninth closest known white dwarf, at 24.8 light-years.
- Temperatures from photometry alone run 5 to 8 percent hot.
- Anything below roughly 4,700 kelvin stays invisible to this technique.
- Gaia’s fourth release should find the rest without another decades-long wait.
Twenty parsecs is the best-catalogued volume of space in existence, and it still held four objects nobody had directly seen. Our census of the neighborhood is incomplete, and after this paper we know roughly by how much, which is not a bad place to be standing. If another headline about a nearby star crosses your feed this year, the useful question is what the catalogue behind it is leaving out.
Gerd Dani, President of FreeAstroScience
Sources
- O’Brien, M. W., Wilson, D. J., Tremblay, P.-E., Gänsicke, B. T., Byrne, C. M., Lagos-Vilches, F., Pineda, J. S., Barraza-Jorquera, J. A. (2026). Direct detections of white dwarfs in four WD+dM post-common envelope binaries within 20 pc. Monthly Notices of the Royal Astronomical Society, 550, pages 1 to 13. Accepted 19 June 2026. https://doi.org/10.1093/mnras/stag1195
- Focus.it (2026). Scoperte quattro “stelle morte” nascoste nel nostro cortile cosmico. Italian coverage of the MNRAS study, translated in our newsroom.
