Black-hole-binary simulations for gravitational-wave astronomy

Gravitational waves were among the first predictions of Einstein's general theory of relativity, but almost a 100 years later they have not yet been directly observed. Gravitational waves are produced when an object accelerates: space and time are distorted, and the distortion travels away as a wave. Gravity is a weak force and the waves produced by the motion of everyday objects are too weak for us to even think about trying to detect them. Very dense and massive astrophysical objects are a different story. Among the most massive and dense objects that we know of are black holes, and the most extreme source of gravitational waves we can imagine is the collision of two of them. For example, if a black hole has the mass of the sun, its radius (crudely speaking) is only about 1.5km. If two of these black holes are orbiting each other and are 10km apart, they will move at almost a tenth of the speed of light, and would complete their 10km orbits over 800 times every second. That is what we mean by the most extreme gravitational-wave source we can imagine. In fact, such a binary wouldn't be able to complete more than a few orbits: it will lose so much energy from gravitational-wave emission that the black holes will quickly fall together and form a single black hole. The merger process itself produces an intense final burst of gravitational radiation that carries away three percent of the holes' mass, an incredible amount of energy. An international network of detectors (in the USA, Germany and Italy) is trying to measure these signals. Unfortunately, we expect that from most sources the signals will still be so weak as to be almost indistinguishable from background noise. The only hope of finding them is to know precisely what the signals should look like. That is the goal of my work: accurately predicting the signals from black-hole mergers. There are approximate formulas that we can use to calculate the waves from black holes in orbit --- they tell us that many binaries emit gravitational waves at the same frequencies as sound; the binaries are quietly buzzing at us --- but when the black holes are close, and when they merge, these formulas are no longer valid. The only way to predict the merger signal is to solve the full set of Einstein's equations, and the only way to do that is with a computer. This first became possible in 2005. Since then many discoveries have come from black-hole simulations. I was involved in one of the most exciting, the prediction that in some situations the final merged black hole can receive such a 'kick' from the emitted gravitational waves (like a canon recoiling after firing) that it can be ejected from its host galaxy and sent flying, alone and unseen --- it is a black hole, after all --- across the universe. Now that black-hole simulations are possible, my challenge is to provide enough information to experimentalists so that they can finally detect gravitational waves and confirm Einstein's predictions. No-one knows how many black-hole binaries the universe contains, buzzing or whirring or squeaking at us. But when we have sufficiently accurate computer simulations and enough reliable detections, we will have a chance to find out. And that will be only the beginning: a new field of Gravitational Wave Astronomy will be born, which will teach us more about the processes of galaxy formation, the physics of the early universe, and maybe even reveal astrophysical objects we never imagined.