Astronomers Found 84 Cosmic Beacons in a Part of the Spectrum We Can Barely See

Astronomers found 84 luminous objects by looking in a place where X-ray searches rarely linger: below 0.3 kiloelectronvolts, at the ragged lower edge of NASA’s Chandra X-ray Observatory data.

In one set of images, the objects appear as points of extremely low-energy X-ray light. Move only slightly higher in energy and they nearly all vanish.

That disappearing act is their defining feature. A team led by Mustafa Muhibullah of the University of Alabama has named them hypersoft X-ray sources, or HSSs. The team reported the population in a peer-reviewed study published in Nature Astronomy.

What the researchers have probably found is not the full light from these objects, but a sliver of it. Their models suggest the radiation peaks just below the X-ray range, in the extreme ultraviolet, a part of the spectrum that is notoriously difficult to observe across astronomical distances.

The result is a peculiar combination. These sources are exceptionally cool by X-ray standards, yet some may be among the most energetic compact objects in their galaxies.

And nobody yet knows exactly what most of them are.

The universe’s missing color

Extreme-ultraviolet radiation, or EUV, occupies the territory between familiar ultraviolet light and X-rays. It is useful radiation for studying very hot matter, but terrible radiation for surveying the distant universe.

Neutral hydrogen and helium absorb EUV photons readily. That is helpful if the photons are interacting with gas inside a galaxy, but disastrous if an astronomer is waiting for them to arrive at a telescope. Gas between the source and Earth can erase the signal long before it reaches us.

The new search used an indirect route around that barrier. If a hot source’s spectrum peaks in the EUV, a small high-energy tail may extend into the softest X-rays that Chandra can register. Find that tail, and you may find the otherwise hidden source.

Muhibullah and colleagues searched public observations of six nearby galaxies: Andromeda, also known as M31; the Pinwheel Galaxy, M101; and the early-type galaxies NGC 3115, NGC 3379, NGC 4697 and NGC 4472.

They divided the X-ray photons into narrow energy bands. To qualify as hypersoft, a source had to be detected between 0.15 and 0.3 keV, show no significant emission above 0.3 keV and produce more than eight times as many photons in that lowest band as it did between 0.3 and 1.0 keV.

The search produced 84 candidates: 18 in M31, seven in M101, 18 in NGC 3115, 12 in NGC 3379, eight in NGC 4697 and 21 in NGC 4472.

Those numbers are not a census of which galaxy has the most. The galaxies sit at different distances and were observed for different amounts of time through different columns of intervening gas. M101 and NGC 3115, for example, each accumulated more than one million seconds of usable exposure, while NGC 4697 had less than one-fifth that amount.

The important result is that HSSs appeared in every galaxy the team examined, and in very different stellar environments. They trace the spiral structure of M101, but they also occur among the much older stars of elliptical galaxies. Whatever produces them is not restricted to recently formed stars.

How 84 objects escaped the catalogues

The data were public. Chandra has been operating since 1999. So why did these sources remain largely unrecognized?

The energy interval in which they appear is unusually difficult terrain.

Chandra’s Advanced CCD Imaging Spectrometer has steadily lost sensitivity to soft X-rays as molecular contamination has accumulated on its optical blocking filters. The Chandra X-ray Center’s calibration documentation says the contamination reduces the detector’s low-energy efficiency, with the softest energies particularly affected.

Calibration is also more uncertain below 0.3 keV. Conventional source searches therefore tend to begin at 0.3 keV or higher. The discovery paper reports that none of the 84 HSSs appeared in Chandra Source Catalog 2.1, even though a few nearby objects had been noticed in earlier work under other classifications.

The team compensated for the difficult calibration by using ordinary, brighter X-ray binaries in each galaxy to convert the low-energy photon counts into estimated energy fluxes. A second approach using hot gas in NGC 4472 produced results within about 25 percent of the first.

That agreement helps, but it does not make the measurements simple. Many HSSs supplied only a small number of photons, and the telescope never recorded the part of their spectra where most of their energy may lie.

Astronomers are trying to reconstruct an unseen mountain from its visible slope.

The photon ratios imply characteristic temperatures no higher than roughly 20 electronvolts, or about 230,000 kelvin. That is fiercely hot compared with an ordinary star’s surface, yet cool for a source selected in X-rays.

At that temperature, a simple blackbody model says the object’s total radiation could be about 18 times its measured luminosity in the 0.15-0.3 keV band. If the source is cooler, the correction climbs precipitously because even less of its light reaches Chandra’s band.

That is why the paper’s most spectacular energy estimates need care. Several sources approach or exceed 10^38 ergs per second in the narrow X-ray band alone, but their total output depends heavily on temperatures and spectral shapes that have not been measured directly.

The safe conclusion is still remarkable: the sources are extremely luminous at the softest X-ray energies, and their true output is probably higher. Exactly how much higher is not known.

One name, several possible machines

“HSS” tells astronomers what the sources look like. It does not yet tell them what they are.

The best clues come from Andromeda. Several of its lower-luminosity HSSs coincide with known novae, outbursts produced when a white dwarf pulls hydrogen-rich matter from a companion star and ignites it on its surface. Residual nuclear burning after an eruption can leave the white dwarf radiating strongly at very low X-ray and EUV energies.

A white dwarf steadily accreting and burning material offers another possibility. If its hot photosphere expands, its effective temperature falls and its emission moves out of the traditional supersoft X-ray band toward the EUV. The star may not have switched off at all; its light may simply have slid into the observational gap.

That possibility touches one of astronomy’s most persistent problems. Type Ia supernovae involve exploding white dwarfs, but researchers still debate which binary systems and evolutionary paths produce them. Astronomers have searched for populations of steadily burning, accreting white dwarfs and generally found too few obvious supersoft X-ray sources to account for every Type Ia event.

Adding the hypersoft objects roughly doubles the discovery paper’s count of potential white-dwarf accretors in these fields. Some could represent a previously missed phase in the lives of Type Ia progenitor systems.

“Could” is doing essential work there. No HSS in this study has been shown to be a white dwarf destined to explode. A nova association demonstrates that some members of the sample involve white dwarfs; it does not establish the nature or fate of the entire population.

White dwarfs may also struggle to explain the brightest HSSs. If those sources are as cool and luminous as the models suggest, their total output could exceed what a white dwarf can plausibly sustain. Accreting black holes are one alternative, and neutron-star systems remain possible.

A brief theoretical paper in Research Notes of the AAS has even proposed a circumbinary disk feeding two black holes in a hierarchical triple system as an explanation for the most luminous cases. The journal describes its research notes as moderated but not peer-reviewed, and the model requires finely tuned conditions. It is an intriguing possibility, not an identification of any observed source.

Its existence is nevertheless revealing: the measurements leave room for some genuinely exotic machinery.

Light that could reshape a galaxy

The radiation that may be hiding these sources could also make them important.

EUV photons can strip electrons from atoms. Ordinary hot stars produce abundant ultraviolet light capable of ionizing hydrogen, but astronomers also observe strongly ionized helium in some galaxies without active galactic nuclei. Removing the second electron from helium requires a photon carrying at least 54.4 eV, within the suspected output of very soft, very hot sources.

A 2025 photoionization study found that adding populations of supersoft X-ray sources to simulated galactic radiation fields could raise helium emission toward observed levels. Hypersoft sources occupy an even more elusive portion of this general temperature range, making them plausible contributors to the ionizing-photon budget.

But finding 84 candidates does not solve the ionization problem. Researchers still need to know how many HSSs actually exist, how long they remain active, what their spectra look like and how much of their radiation escapes into the surrounding galaxy. The same gas that makes EUV astronomy so difficult also determines how far those photons travel.

Follow-up observations may help separate the suspects. The paper suggests that sufficiently cool HSSs could have detectable ultraviolet tails. Better optical and ultraviolet counterparts could reveal companion stars, nova histories or star-cluster environments. Repeated X-ray observations could show whether individual sources persist, recur or merely shift in and out of Chandra’s narrow window.

For now, the discovery has established something simpler and stranger. Eighty-four compact sources were detected at the softest X-ray energies, and models suggest that much of their radiation may be pouring into the difficult-to-observe extreme ultraviolet.

Sometimes astronomy opens a new window by launching a new telescope. This time, it opened one by looking more carefully through the very bottom of an old window, and noticing the light spilling around its edge.