This post is the fourth in a series that accompanies the publication of my book ‘Gravity’s Engines: How Bubble-Blowing Black Holes Rule Galaxies, Stars, and Life in the Cosmos’ (Scientific American/FSG).

Ten years ago the universe was in trouble.

Or rather, our puny human theories about the nature of all the stars and galaxies in the universe were in trouble. At this time we had managed to discover a number of remarkable things about our cosmic surroundings. We had detected minute non-uniformities in the remains of the primordial radiation field known as the cosmic microwave background, and we had ever more convincing evidence that normal matter - the stuff we are made of - amounts to barely a fifth of the total raw mass of the universe. The rest is dark matter, a shadowy stuff that seldom interacts particle-to-particle but is massive enough that its gravity dictates much of the motion of stars in galaxies and the colossal structures of galaxy clusters and superclusters.

A virtual cosmos. The predicted distribution of stars and galaxies across tens of millions of light years (Credit: Millenium Simulation)

Armed with these observations, and with our models of how these features would translate into the seeding and gravitational growth of structure in the universe, we could build virtual realities in which galaxies and stars would arise and populate an imaginary cosmos. These simulations could then be compared to our great cosmological maps and our statistical counts of all the bright and dark places around us. It was a way to confirm our ideas, compare virtual universes to the real one. And this was the problem, because most of these models did not look quite like the real thing.

The fact was that the simulations overproduced stars and they overproduced big bright galaxies. They didn't produce a cosmos that looked like ours. Debate raged over the possible cause. It was possible that the computer algorithms used to simulate the production of stars in the virtual galaxies were deeply flawed. It was also possible that we were missing a critical component, a source of additional energy that would dampen down the process in the real universe, stunting galaxies and regulating stellar births from virtual nebula. But what could it be? It might be the energy of the stars themselves - massive suns burning through their nuclear fuel quickly and exploding as supernova could impose a speed-limit on the formation of subsequent generations. It might also be something else that was lurking in the dark alleys and back rooms of the cosmos.

Supermassive black holes had long been suspected of being a disruptive element. As I wrote in the last piece in this series, when they consume matter they can be incredibly noisy, casting energy back out into the surrounding universe. Inside clusters of galaxies we see them blowing bubbles and generating waves in intergalactic gas, phenomena that slow down the condensation of that gas into cool nebula and new stars. In other places we see the wash of radiation and particles flooding galaxies, a sporadic but potent force. It looked like a promising solution, but was still incomplete.

I was personally convinced of the role that these giant holes play in 'pruning' galaxies and stars when my colleagues and I stumbled across a great clue in the deep cosmic past. For black holes to provide the missing link, they needed to have helped stunt the growth of the very largest galaxies. And that meant that they needed to have been involved from the very outset, something that was unconfirmed. Luckily, we had discovered just such a situation in a place some 12 billion light years away, in the early morning of galactic evolution.

The birth of a galaxy. An overlay of the X-ray light (blue) and ultraviolet light (red) coming from a system 12 billion years ago (with thanks to Wil van Breugel and Ian Smail)

It has a dreadfully boring name of 4C 41.17. A source of intense radio waves so far away that its light is stretched, redshifted, by a factor of 3.8 due to the expansion of the universe. Using the giant Keck telescope in Hawaii, a seven hour exposure allowed us to make an image of what was originally the ultraviolet light flooding from this system - revealing a tumultuous scene of gas ebbing and flowing. Deep infrared images also showed a colossal mass of star dust, 600 million times the mass of our Sun in carbon-rich grains, a fingerprint of an orgy of stellar birth and death. And then we used NASA's orbiting Chandra X-ray Observatory for 40 hours to gather the barest whisper of X-ray light, a total of only 150 twelve billion year-old photons. But this was all we needed, because they revealed the whole story.

Deep inside this teenage galaxy was a supermassive black hole. As it consumed matter it was spewing forth electrons at close to the speed of light. As these particles sped away they were meeting the cosmic background radiation, primordial photons. But 12 billion years ago the universe was more compact, the background radiation was warmer than it is today, and so when these encounters happened the exchange of energy was far more efficient than it is today. Black hole accelerated electrons boosted primordial photons into X-rays, and the X-rays were penetrating through the murk of this youthful galaxy, modifying its production of new stars.

It was an exciting discovery. A vivid demonstration that we did indeed need black holes, that they were not mere bystanders in the game of cosmic evolution but central players, today, yesterday, and all the yesterdays before that.

The number of stars in the universe, and the very nature of the galaxies, owes an enormous debt to the behavior of these most extreme and fantastical of astrophysical phenomena. Ten years ago they helped solve part of a mystery, who knows what we'll learn about them next.

The whole story of our modern view of black holes is told in Gravity's Engines, available today!