Chandra Observatory Identifies 84 New 'Hypersoft X-ray Sources' in Six Galaxies
NASA's Chandra X-ray Observatory has detected a new class of compact binary systems that emit unusually low-energy X-rays and intense ultraviolet radiation, potentially resolving two long-standing astrophysical puzzles.
Astronomers using NASA’s Chandra X-ray Observatory have identified 84 previously unknown objects scattered across six galaxies that defy easy classification. These systems, described in a Nature Astronomy paper published September 9, 2026, emit X-rays so soft—meaning at such low energies—that they barely register against the background, yet simultaneously pour out intense ultraviolet radiation. The finding has drawn immediate attention because it may resolve two distinct problems that have persisted in X-ray astronomy for years.
The objects were spotted in galaxies including Andromeda (M31) and the Pinwheel Galaxy (M101), among others. Their spectral signature is peculiar: most X-ray binaries produce harder, higher-energy radiation as matter from a companion star crashes onto a compact object, typically a neutron star or black hole. These new sources do the opposite. The gas transfer is happening, but the resulting X-ray emission is unusually weak, while the ultraviolet output is disproportionately strong. Space.com’s reporting on the discovery emphasizes that astronomers have not seen this particular balance of emissions before in compact binary systems.
The mechanism appears to involve what researchers are calling “hypersoft” X-ray sources—compact objects where the accretion disk never fully heats to the temperatures typical of brighter X-ray binaries. Something is stripping or diverting energy that would normally emerge as X-rays, converting it instead to ultraviolet. The exact physics remain under investigation, but the objects’ very existence suggests that the population of accreting compact binaries in nearby galaxies has been systematically undercounted. Standard X-ray surveys would simply miss sources this soft.
This has direct implications for two standing questions. One is the so-called “missing X-ray binary” problem—the apparent shortage of observed systems compared to what population models predict should exist in typical galaxies. If a substantial fraction of binaries operate in this hypersoft regime, they would have escaped detection by instruments tuned for harder X-rays. The other is the “ultraviolet excess” observed in some galactic spectra, where the ultraviolet component is stronger than expected from stellar populations alone. These 84 sources, distributed across six galaxies, could contribute meaningfully to that excess without having been recognized as the origin.
The six galaxies examined were not selected randomly; they represent a range of types and distances that allowed the team to test whether the phenomenon is localized or widespread. Finding 84 such objects across this sample suggests the population is common rather than exotic. The NASA announcement notes that follow-up observations with other wavelengths will be needed to pin down the nature of the compact objects themselves—whether they are neutron stars or black holes—and to understand why their accretion disks behave so differently from the better-studied population.
What makes the finding practically significant is that it points toward a calibration issue in how astronomers inventory the energetic output of galaxies. X-ray observatories have historically prioritized sensitivity to harder radiation, both because it penetrates intervening gas more effectively and because the sources themselves are more dramatic. The Chandra detection of these hypersoft sources required careful analysis of data that might otherwise have been dismissed as noise or background. If similar populations exist in more distant galaxies, the total energy budget from accreting compact objects could be substantially higher than current estimates suggest—shifted toward ultraviolet rather than X-ray, but present nonetheless. The paper’s authors are explicit that this is not merely a cataloging exercise: it changes what questions can be asked about how mass transfer proceeds in close binary systems, and what the end states of stellar evolution actually look like in practice rather than in models.