Like Ben Johnson’s explosive jump out of the starting blocks for 100 m, the growth of supermassive black holes had a “jump start” in the early universe. What triggered this fast build-up has long been a mystery, but now a detailed numerical simulation shows that a major collision between two galaxies rapidly drives huge amounts of gas towards the centre, where it collapses into a supermassive black hole.
How is it possible that fully mature quasars are already observed at a redshift, z, of around 6, corresponding to a time when the universe was only about 1000 million years old? How could their central engine – a supermassive black hole of about 1000 million times the mass of the Sun – have grown so quickly? This rapid spurt has puzzled theorists for many years (CERN Courier July/August 2005 p10).
Whether the starting point is a black hole of about 100 solar masses, which could be the remnants from the first generation of stars, or a larger one of 100,000 solar masses resulting from the gas that can accumulate and collapse in the centre of an isolated protogalaxy, the problem is basically the same. The black hole must be continuously fed at – or close – to its maximum rate, which is controlled by the balance between gravitational attraction and radiation pressure. The surrounding matter is either of too low density or undergoes strong stellar formation, both of which prevent the black hole from effectively capturing the gas.
Numerical simulations have shown that the collision and merging of two galaxies can rapidly drive most of their gas content within about 100 light-years from the centre, but it remained unclear whether this gas could be channelled towards the very centre of the galaxy and collapse into a black hole. This issue has now been addressed by Lucio Mayer from the University of Zurich and colleagues. The trick they use to go to smaller spatial scales with current supercomputer facilities is to split the fluid particles describing the gas into eight lighter particles. They do this in a limited volume, only slightly before the final merger of the two galaxies. This allows them to study the infall of the gas on scales 100 times smaller than previously achieved.
The new simulation, published in Nature, starts with two identical, and relatively large, disc galaxies. The results reveal the formation of a central disc of turbulent gas with a strong spiral pattern that further channels the matter towards the central light-year. The dense, central gas cloud with a mass of about 260 million solar masses suddenly becomes unstable towards gravitational collapse and forms a supermassive black hole in only about 100,000 years after the completion of the merger. Additional simulations show that this direct-collapse scenario would also work for mergers of galaxies with a 10 times lower mass, but for still lighter ones the central gas cloud remains stable, which could explain the absence of supermassive black holes in most dwarf galaxies.
The study has several cosmological implications. The common idea that galaxies grow in parallel with their supermassive black hole needs to be revised. The simulations suggest that the heaviest black holes form first and that galaxy growth is modulated by the size of the black hole, rather than the opposite. Furthermore, the rapid growth of big galaxies seems to be in contradiction with hierarchical structure formation (CERN Courier September 2007 p11). Stelios Kazantzidis, a co-author of the study, solves the paradox by explaining that only dark matter builds up slowly from smaller to larger structures, whereas ordinary, baryonic matter collapses more efficiently.