Black Hole Takes a Seven-Hour Snack, Stuns Scientists

On July 2, 2025, NASA's Fermi Gamma-ray Space Telescope caught something that should not have lasted. Gamma-ray bursts are, by definition, brief — a few seconds to a couple of minutes, generated when a massive star collapses into a black hole and unleashes its remaining fuel in a single cataclysmic flash. GRB 250702B kept going for seven hours. Then it kept going some more. X-ray afterglow persisted for 65 days. It was, in the words of one researcher, unlike anything seen in 50 years.

The duration alone would be notable. But the real signal is in the structure. GRB 250702B did not fade in a smooth arc. It pulsed — three distinct gamma-ray episodes spread across roughly 3.2 hours, each separated by about 2,825 seconds. That interval is not random. It is consistent with an orbital period. Something was feeding the black hole in stages, on a schedule set by gravity.

"A stellar-mass black hole — roughly three times the mass of our Sun — orbiting and gradually merging with a companion star, getting fed from the outside in phases," is how Huei Sears, a postdoctoral researcher in the Department of Physics and Astronomy at Rutgers University, described the leading scenario. Sears is lead author on one of several papers posted to arXiv analyzing the event.

The physics is not subtle. Standard GRBs destroy their own fuel source. The collapse is the explosion; the explosion is the end. What NASA Goddard astronomer Eliza Neights and colleagues modeled instead is a binary system where a compact object — the black hole — is slowly tearing apart a close companion and accreting the debris in repeated episodes. Each pulse corresponds to a fresh delivery of matter onto the black hole's disk. The orbital clock kept time; the feast continued.

Brendan O'Connor, a McWilliams Postdoctoral Fellow at Carnegie Mellon University's McWilliams Center for Cosmology and Astrophysics, led the follow-up campaign. His team directed three space-based X-ray telescopes — Swift, NuSTAR, and Chandra — to monitor the afterglow for 65 days after the initial detection. "The continued accretion of matter by the black hole powered an outflow that produced these flares, but the process continued far longer than is possible in standard GRB models," O'Connor said in a university statement. "The late X-ray flares show us that the blast's power source refused to shut off."

The energy released was enormous: at least 2.2 × 10^54 ergs, making GRB 250702B the most energetic single event in the Fermi dataset by a wide margin. Its host galaxy — an unusually large, dusty, high-mass system — was also unlike any previously catalogued as a GRB host. The burst occurred near the edge of the galaxy's dark dust lane, ruling out the supermassive black hole at the galactic core as the source. This was not a central engine. It was something smaller and closer to the edge.

The event was rare to the point of statistical invisibility. Based on one occurrence in Fermi's operational lifetime, researchers estimate such bursts occur at volumetric rates more than 1,000 times lower than normal high-luminosity long GRBs and more than 100,000 times lower than core-collapse supernovae, after correcting for jet beaming. Of roughly 15,000 gamma-ray bursts observed since their discovery in 1973, none approach this duration.

The dataset is genuinely multi-wavelength. NASA's Fermi caught the gamma rays. The James Webb Space Telescope's NIRCam instrument identified narrow hydrogen emission lines placing the burst at a redshift of z = 1.036 — roughly 8 billion light-years away. The Hubble Space Telescope, Einstein Probe, and ground-based facilities including the Very Large Array contributed observations across infrared, X-ray, and radio bands. Adelle Goodwin, an astronomer at Curtin University in Australia leading the VLA radio follow-up, noted that classical collapsar models cannot account for the emission duration — "more exotic scenarios need to be invoked."

The alternative scenarios are still being argued. The leading candidates are an extreme collapsar jet — a standard GRB with an abnormally energetic and sustained engine — or a tidal disruption event where a middleweight black hole shredded a star. The black-hole-plus-stripped-helium-star model fits the three-pulse periodicity most cleanly, but it requires conditions that models have not fully specified. Eric Burns, an astrophysicist at Louisiana State University, noted that only the combined power of instruments on multiple spacecraft made the event legible at all.

What GRB 250702B does not fit is the textbook. The periodicity in the pulse structure is the clearest signal: this was not one explosion but several, timed by orbital mechanics rather than a single collapse. Whether the fuel arrived via stellar cannibalism, tidal stripping, or some mechanism not yet formalized, the black hole was being repeatedly replenished — which is not supposed to happen in the current GRB framework. Researchers expect to know more once the full multi-wavelength analysis is published. For now, the short answer is that something kept feeding the fire, and nobody had a model that predicted it.