An infrared census of accreting supermassive black holes across a wide range of cosmic times indicates that the canonical understanding of how these luminous objects form and evolve may need to be adjusted.
Ask an astronomer when quasars were at their peak and they will probably tell you it was about 10 billion years ago, when the Universe was about one-third of its current size1,2. Before then, the quasar population was still growing along with other large structures in the young Universe; there has since been a steady decrease in quasar numbers. However, in a paper published in The Astrophysical Journal, Vardanyan et al.3 present results suggesting that this widely accepted picture may not be correct — or at least that it does not tell the whole story.
That story started in 1963 with the discovery4,5 of a new type of astronomical object, referred to variously as quasi-stellar objects or quasars, the name that is generically used today. Their physical nature was initially unknown, but it was gradually deduced6 that a quasar is a glowing disk of hot, dense material that can form around the supermassive black hole at the centre of a large galaxy, often the result of a collision with a second galaxy. Although such accretion disks are 'only' about the size of the Solar System, they can outshine all the stars in the host galaxy by a factor of a thousand or so. Quasars can hence be seen comparatively easily at great distances, which makes it possible to trace their evolution back to the first billion years after the Big Bang.
More than a million quasars have been catalogued in the 50 years since their discovery. Although this is more than enough for most demographic studies of astronomical objects, it is difficult to obtain a representative sample of quasars that spans a wide range of distances from Earth, and hence cosmic look-back times. It is also challenging to properly account for all the energy output of a quasar, because some of the ultraviolet light that is emitted from the accretion disk is absorbed by dust in the host galaxy and re-radiated at much longer, infrared wavelengths. Most surveys of the quasar population have been undertaken using observations made at optical or near-infrared wavelengths (between about 0.2 and 2 micrometres), and it is these types of measurement that have provided the strongest evidence that quasar numbers peaked fairly sharply 10 billion years ago.
Vardanyan and colleagues studied a comparatively small sample of 10,000 quasars that were initially identified using optical data from the Sloan Digital Sky Survey. But, crucially, the authors had access to longer wavelength measurements (at about 8 μm) of the same objects from the Wide-Field Infrared Survey Explorer (WISE) satellite. They were thus able to get a more complete census of the quasars' energy output and, after correcting for the various complicated observational selection effects that inevitably make such studies so difficult, found some striking results. They confirmed the steady decrease in the quasar population over the past 10 billion years but, rather than the expected drop at cosmic times before 3 billion years, they found a 'plateau' in the quasars' energy output back to a little over a billion years after the Big Bang (Fig. 1). The authors were unable to probe any earlier than this, and one of their conclusions was that extending these sorts of measurements to earlier times is the best way to explore this issue further.
These results are not unprecedented — there have been several similar previous claims7,8 that the canonical understanding of the quasar population from optical data was incomplete. However, the scale and quality of the WISE data are superior to any previously available. The findings demand serious attention, both in terms of subjecting them to further scrutiny and exploring their implications for quasar formation if the simplest interpretation — that large numbers of high-luminosity quasars were in place just a billion years after the Big Bang — is indeed correct.
The most exciting potential implication of Vardanyan and colleagues' study is that we need to adjust our understanding of the quasar population, especially how the early quasars formed. Most current models are based on the idea that galaxy collisions trigger quasar activation, so the number of quasars should rise sharply as galaxies form, grow and collide in the early Universe. The authors' results suggest that this link is not so strong, and that the most luminous quasars in particular form more rapidly than astronomers might suspect using simple models of black-hole accretion and galaxy collisions.
The word 'suspect' is appropriate here, because this sort of science really is like detective work, in which indirect clues must be combined with inspired deduction to reach any interesting conclusions. It is remarkable that it is possible to make any kind of inference about black holes that are billions of light years away and have long since ceased to exist as quasars. One ambiguity is that the infrared light being used to assess the quasars' energy output could come from other sources, because any mechanism that heated whatever dust was present in the host galaxy would contribute to this signal. Also problematic is that various corrections to the inferred output of the quasars have to account for the expansion of the Universe: the light seen at any given wavelength here and now has, since its emission, been redshifted by an amount that depends on how distant the source is, and hence how far back in time astronomers are seeing it. Perhaps the most uncertain aspect of all attempts to measure the evolution of the quasar population is deciding how best to account for this effect and how to test whether it has been done correctly. The approach taken by Vardanyan et al. is reasonable, but it is easy to imagine future data that would allow these corrections to be improved.
'More data' is something of a mantra in astronomy. Technological developments such as WISE have been one of the main drivers of discovery for the past century, and probably will continue to be in the future. We already have exciting projects such as the Large Synoptic Survey Telescope and the Square Kilometre Array just a few tantalizing years away, and both should tell us a great deal more about the age of the quasars.
Fan, X. et al. Astron. J. 122, 2833 (2001).
Richards, G. T. et al. Astron. J. 131, 2766 (2006).
Vardanyan, V., Weedman, D. & Sargsyan, L. Astrophys. J. 790, 88 (2014).
Hazard, C., Mackey, M. B. & Shimmins, A. J. Nature 197, 1037–1039 (1963).
Schmidt, M. Nature 197, 1040 (1963).
Rees, M. J. Ann. Rev. Astron. Astrophys. 22, 471–506 (1984).
Casey, C. M. et al. Astrophys. J. 761, 139 (2012).
Carilli, C. L. et al. Astrophys. J. 763, 120 (2013).