NASA's Chandra X ray Observatory tracked 22 supernova remnants in the spiral galaxy and found that roughly half are not dimming as expected, hinting that the collapsed stars at their centers are feeding on companion stars or their own ejected material.
When Andrea Prestwich's team reexamined 14 years of Chandra X-ray observations of the spiral galaxy Messier 83, they expected to watch old supernova remnants fade into the background. Instead, roughly half of the 22 known remnants brightened and dimmed in ways that simple blast-wave physics cannot explain. The dead stars at their centers, the team argues, may not be dead. They are stealing matter from companion stars they were never supposed to have.
Messier 83 (M83) sits about 15 million light-years from Earth and has been a fixture of X-ray astronomy since Chandra first pointed at it in 2000. The galaxy's spiral arms are dense with young, massive stars, the kind that end their lives in violent collapses. Some of those collapses leave behind a neutron star or a black hole. A shock wave from the explosion races outward and heats the surrounding gas to X-ray-emitting temperatures for thousands of years. Once the shock has swept past its immediate neighborhood, the remnant should settle into a slow, predictable fade.
That is not what the data show. The team's catalog of 22 sources in M83, all previously identified as supernova remnants, includes one, SN 1957D, whose variability is straightforward: the blast wave is still plowing into denser surrounding material and producing fresh X-rays. The other roughly ten variable sources have no such ready explanation. Some of them are clearly not young remnants at all. Something is feeding them.
The leading explanation, laid out in a paper led by Andrea Prestwich of the Catholic University of America and presented at the 248th American Astronomical Society meeting in Pasadena, is that each variable source began as a massive binary. The more massive star exploded and collapsed to a black hole or neutron star. The companion survived. The compact object is now pulling gas off that companion, and the resulting accretion generates the X-ray flicker. These are high-mass X-ray binaries, already known to be among the most variable X-ray sources in any galaxy.
The catch is that supernova-remnant-associated high-mass X-ray binaries were rare before this work. Only a handful of candidates were known across all galaxies. Finding more than 20 strong candidates in a single galaxy, M83, is unprecedented, and the variable remnants cluster in the regions of the galaxy with the highest concentrations of massive stars, exactly where such binaries should form. That correlation is what makes the binary explanation more than a guess.
There is also a second possibility, one the authors say may run in parallel. The compact object could be recapturing fallback debris from the original supernova, gas that never quite escaped the explosion. Prestwich has called this "cosmic recycling." The two channels are not mutually exclusive, and the X-ray data alone cannot tell them apart. Distinguishing them will need either a direct image of the companion star, impossible at M83's distance with current instruments, or a multi-wavelength campaign that catches an accretion disk in the act.
The M83 result does not stand alone. A follow-up study of M51, the Whirlpool Galaxy, by Zoe Hoiland at Vassar College and Roy Kilgard at Wesleyan University found a similar population of variable X-ray sources tied to supernova remnants, as documented in the Chandra X-ray Observatory photo album for M83. The pattern is starting to look like a feature of vigorously star-forming galaxies rather than an M83 oddity.
Chandra is uniquely suited to this work because the hottest, most energetic debris from a stellar explosion glows in X-rays for thousands of years, while the visible light fades. Visible-light telescopes see the cooling remnant; an X-ray telescope sees the still-running engine inside it. The M83 survey is archival reanalysis rather than new observation, and the cadence is sparse: single visits in 2000 and 2001, ten visits clustered in 2010 and 2011, and one in 2014. That is enough to catch large-amplitude variability across years, not enough to track a single outburst hour by hour.
The next test is straightforward in principle and demanding in practice. Watch one of the variable remnants continuously with Chandra for a full outburst cycle, then point radio, optical, and ultraviolet instruments at it during a high state. If the binary channel dominates, an accretion disk should appear. If the fallback channel dominates, the spectrum should look more like a cooling blast wave with a twist. Either answer will reshape how astronomers read the rest of Chandra's archive, which contains years of unresolved X-ray sources in other star-forming galaxies that may have been hiding similar behavior.