To Martin Rees, the United Kingdom’s Astronomer Royal, AGN feedback offered a natural way to connect the relatively tiny black hole to the galaxy at large. Two decades earlier, in the 1970s, Rees correctly hypothesized that supermassive black holes power the luminous jets observed in some far-off, brightly glowing galaxies called quasars. He even proposed, along with Donald Lynden-Bell, that a black hole would explain why the Milky Way’s center glows. Could these be signs of a general phenomenon that governs the size of supermassive black holes everywhere?
The idea was that the more matter a black hole swallows, the brighter it gets, and the increased energy and momentum blows gas outward. Eventually, the outward pressure stops gas from falling into the black hole. “That will terminate the growth. In a hand-wavy way, that was the reasoning,” said Rees. Or, in Di Matteo’s words, “the black hole eats and then swallows.” A very big galaxy puts more weight on the central black hole, making it harder to blow gas outward, and so the black hole grows bigger before it swallows.
Yet few astrophysicists were convinced that the energy of infalling matter could be ejected in such a dramatic way. “When I was doing my thesis, we were all obsessed with black holes as a point of no return — just gas going in,” said Natarajan, who helped develop the first AGN feedback models as Rees’ graduate student. “Everyone had to do it very cautiously and gingerly as it was so radical.”
Confirmation of the feedback idea came a few years later, from computer simulations developed by Di Matteo and the astrophysicists Volker Springel and Lars Hernquist. “We wanted to reproduce the amazing zoo of galaxies that we see in the real universe,” Di Matteo said. They knew the basic picture: Galaxies start out small and dense in the early universe. Wind the clock forward and gravity smashes these dwarfs together in a blaze of spectacular mergers, forming rings, whirlpools, cigars and every shape in between. Galaxies grow in size and variety until, after enough collisions, they become big and smooth. “It ends up in a blob,” said Di Matteo. In the simulations, she and her colleagues could re-create these large featureless blobs, called elliptical galaxies, by merging spiral galaxies many times. But there was a problem.
While spiral galaxies like the Milky Way have many young stars that glow blue, giant elliptical galaxies only contain very old stars that glow red. “They are red and dead,” said Springel, of the Max Planck Institute for Astrophysics in Garching, Germany. But every time the team ran their simulation, it spat out ellipticals that glowed blue. Whatever was switching off star formation hadn’t been captured in their computer model.
Then, Springel said, “we had the idea to augment our galaxy mergers with supermassive black holes in the center. We let these black holes swallow gas and release energy until the whole thing flew apart, like a pressure cooker pot. Suddenly, the elliptical galaxy would stop star formation and would become red and dead.”
“My jaw dropped,” he added. “We did not expect [the effect] to be so extreme.”
By reproducing red-and-dead ellipticals, the simulation bolstered the black hole feedback theories of Rees and Natarajan. A black hole, despite its relatively tiny size, can talk to the galaxy as a whole through feedback. Over the last two decades, the computer models have been refined and expanded to simulate large swaths of the cosmos, and they broadly match the eclectic galaxy zoo we see around us. These simulations also show that ejected energy from black holes fills the space between galaxies with hot gas that otherwise should have already cooled and turned into stars. “People are convinced by now that supermassive black holes are very plausible engines,” said Springel. “No one has come up with a successful model without black holes.”
Mysteries of Feedback
Yet the computer simulations are still surprisingly blunt.
As matter creeps inward to the accretion disk around a black hole, friction causes energy to be pushed back out; the amount of energy lost this way is something the coders put into their simulations by hand through trial and error. It’s a sign that the details are still elusive. “There’s a possibility that in some instances we’re getting the right answer for the wrong reason,” said Quataert. “Maybe we’re not capturing what is actually the most important thing about how black holes grow and how they dump energy into their surroundings.”
The truth is that astrophysicists don’t really know how AGN feedback works. “We know how important it is. But it’s escaping us exactly what causes this feedback,” said Di Matteo. “The key, key problem is that we don’t understand feedback deeply, physically.”
They know that some energy is emitted as radiation, which gives the centers of active galaxies their characteristic bright glow. Strong magnetic fields cause matter to fly out from the accretion disk too, either as diffuse galactic winds or in powerful narrow jets. The mechanism by which black holes are thought to launch jets, called the Blandford-Znajek process, was identified in the 1970s, but what determines the beam’s power, and how much of its energy gets absorbed by the galaxy, is “still an open unsolved problem,” said Narayan. The galactic wind, which emanates spherically from the accretion disk and so tends to interact more directly with the galaxy than the narrow jets, is even more mysterious. “The billion-dollar question is: How is the energy coupling to the gas?” said Springel.
