Turn down a lamp and the room should get darker. Astronomers watching a black hole binary called MAXI J0637-430 fade after its 2019 outburst saw the opposite: as the X-ray source dimmed, gas in the outer part of its black hole accretion disc looked more strongly lit, not less. A study published in Nature Astronomy on 22 October 2026 argues that the disc had been casting a shadow over itself, and that this happens at far lower brightness than standard theory allows.
The system lies in our own galaxy. A black hole draws gas from a small companion star, which earlier work estimated circles it roughly every two hours. Japan’s MAXI X-ray monitor on the International Space Station picked it up as a new, uncatalogued source on 2 November 2019. A team led by Montserrat Armas Padilla of the Instituto de Astrofísica de Canarias (IAC) followed its decline with the X-shooter spectrograph on ESO’s Very Large Telescope (VLT) in Chile, alongside X-ray data from NASA’s NICER, Swift and NuSTAR missions. The paper involves researchers in eight countries, and the IAC announced the result on 23 October.
Gas spiralling towards a black hole heats up as it falls, and the innermost disc glows fiercely in X-rays. Some of that radiation travels outwards and strikes the cooler outer disc, heating it in a process astronomers call irradiation. Nobody can photograph a disc this small at this distance. Its temperature has to be read from its spectrum instead, and the outer disc turns out to be a sensitive light meter for whatever the centre is doing.
At the outburst’s peak, when the system shone at about a fifth of its Eddington luminosity (the brightness at which the outward push of radiation on infalling gas balances the pull of gravity), X-shooter recorded a cool spectrum full of strong hydrogen and helium emission lines. That is the signature of a weakly irradiated outer disc. About two weeks later the X-ray output had roughly halved. Yet the lines had almost vanished, leaving the nearly featureless spectrum of a hot, heavily irradiated disc. “It was a surprise, exactly the opposite of what we expected,” Armas Padilla said.
The team’s explanation is geometric. At peak brightness the innermost disc swelled vertically into a thick structure that blocked much of the high-energy radiation from reaching the outer regions. As the flow of gas eased, the inner disc thinned again, the shade came off and the outer disc took the full glare. X-ray spectra taken at the same times fit this picture of a changing inner flow. Co-author Teo Muñoz-Darias likens it to dimming a lamp while removing its shade.
What makes the result awkward for theory is the threshold. Classical models expect the inner disc to thicken only as a source approaches the Eddington limit; here it appears to have done so at roughly 10% to 20% of it. That matters well beyond one binary. Estimates of black hole spin often rely on fitting X-ray spectra with models of a thin, flat inner disc, and a thicker disc at moderate brightness could shift those values. The authors note knock-on effects for X-ray polarisation studies, which probe the shape of the hot gas near a black hole, and for the winds that irradiation drives off the outer disc. They also point to similar behaviour in other systems, suggesting self-shadowing may be common.
The catch lies in the bookkeeping. The evidence comes from one outburst of one system, observed at a handful of epochs. Converting a measured brightness into a fraction of the Eddington limit requires the black hole’s mass and distance, and neither is well pinned down here; an earlier analysis estimated the peak at under a tenth of the limit and the distance at probably less than 7 kiloparsecs (about 23,000 light years). A dimmer peak would make the early thickening more surprising still, but it leaves the exact threshold unsettled. The claim that the effect is widespread rests on comparison with other systems rather than equally detailed monitoring, and testing it means catching new outbursts early enough to watch them in X-rays, visible and infrared light at once.
For now, MAXI J0637-430 offers a tidy lesson in reading indirect evidence. The outer disc never sees the black hole. It only records how much light gets past whatever stands in between. In late 2019 that reading went up as the lamp went down, and the culprit was a shade that the textbooks did not expect to be there.
