Supermassive black hole dims to 5% of its original brightness in two decades

Astronomers watched a supermassive black hole go quiet in real time. A galaxy roughly 10 billion light-years from Earth has dimmed to about one-twentieth of its original brightness over roughly two decades, and the mass feeding rate into the black hole at its center fell to roughly one-fiftieth of what it was in about seven years. That is not supposed to happen on human timescales. The results were published in the Publications of the Astronomical Society of Japan in November 2025.

The galaxy, designated J0218-0036 at redshift 1.8, was flagged when researchers compared images from the Sloan Digital Sky Survey with newer observations from Hyper Suprime-Cam on the Subaru Telescope. The brightness difference was extreme enough that the team, led by Tomoki Morokuma at the Chiba Institute of Technology, immediately pursued follow-up observations with the Gran Telescopio Canarias and the W.M. Keck Observatory, and dug through archival data going back about 70 years. They also pulled X-ray and infrared data spanning multiple decades. The cross-wavelength picture ruled out a simple dust obstruction in front of the accretion disk: dust would not produce the specific pattern of dimming they measured across optical and infrared bands simultaneously. The physical state of the accretion disk itself appears to have changed.

The standard picture of how supermassive black holes feed holds that the mass accretion rate onto the central object varies over timescales of tens of thousands of years or longer. What the J0218-0036 data suggest is that at least some SMBHs can shift their fuel intake dramatically within years. "This object shows rapid variability that cannot be explained by standard models," said co-author Toshihiro Kawaguchi of the University of Toyama, who worked on the theoretical interpretation. "It provides an important test case for developing new theoretical models."

Morokuma put it more directly: "It is fascinating that an active galactic nucleus can change its brightness so dramatically over such a short period of time, and that this fading appears to be caused by a large change in the accretion rate onto the supermassive black hole."

The co-evolution question is why this matters beyond the observatory. Most large galaxies host a supermassive black hole at their center, and the prevailing view has been that the activity of the central engine and the evolution of the host galaxy influence each other over cosmological time. If an AGN can transition from bright to dim within years rather than millennia, the feedback mechanisms that connect black hole activity to star formation, gas supply, and galaxy structure may operate on much faster adaptive timescales than the models assumed. That is a non-trivial recalibration for anyone working on galaxy formation simulations or observational surveys that try to use AGN activity as a proxy for galactic history.

The detection itself is a product of survey methodology. Hyper Suprime-Cam on Subaru is a wide-field instrument, and the researchers specifically used multi-epoch, multi-wavelength archival comparisons to find objects that had change significantly rather than just cataloging what existed at one moment. The team included collaborators from the University of Potsdam, the Instituto de Astrofisica de Canarias in Spain, the National Astronomical Observatory of Japan, and Ritsumeikan University. The paper of record is Morokuma et al., "A possible shutting-down event of mass accretion in an active galactic nucleus at z ~ 1.8," PASJ, November 4, 2025.

What caused the gas supply to drop so sharply is not yet answered. The authors note that further observations and theoretical work are needed. A cloud of intervening material large enough to block the inner disk would show a different dimming signature across wavelengths, and that possibility was evaluated and ruled out. The alternative left is a genuine reduction in the accretion rate itself, meaning less material is reaching the black hole. Whether that reflects a structural change in the surrounding medium, some instabilities in the disk dynamics, or something external to the galaxy remains open.

The broader implication is observational. If SMBH accretion states can flip within years, then single-epoch surveys miss a dimension of variability that matters for understanding what AGN actually do. The researchers are already proposing to use wide-field imaging surveys to find more objects in transition. If they are right that this is not a rare exception but an observable phase that has been systematically missed, the sample of objects with rapid AGN state changes could grow substantially. That would either reinforce or reshape the theoretical picture depending on how common the phenomenon turns out to be.

The paper does not make that claim explicitly. The authors are careful to call this "a possible shutting-down event." But the data are real, the timescales are real, and the gap between what was expected and what was observed is real. What happens next requires more objects and more time on sky. For now, J0218-0036 is the clearest evidence yet that supermassive black holes do not always behave like geological processes.