Gravitational lensing of light from some of the most distant objects known could be more widespread than had been thought. If so, it could be good — and bad — news for cosmologists.
A massive object, such as a galaxy, that happens to lie near our line of sight from Earth to a distant astronomical source can gravitationally deflect the path of light from that source. As a result, the image that reaches us may be displaced, distorted, magnified and multiplied. These cosmic optical effects, known as gravitational lensing, have been the subject of increasingly intense study by astronomers since the discovery of the first multiple images1 of a quasi-stellar object, or quasar, in 1979. Quasars are the brightest objects in the Universe, mostly found at the centres of galaxies and thought to be associated with black holes.
Lensing can reveal valuable, and often otherwise unavailable, information about the source, the lensing object and, in cosmological contexts, even the geometrical properties of the intervening space2. Because a strong lensing effect requires a rather precise chance alignment between the distant source and the foreground lensing object (Fig. 1), such systems are quite rare. For example, a few quasars in every thousand3 and only around one star in a million in nearby galaxies4 are strongly lensed. Gravitational-lens systems have thus been highly prized discoveries. But on page 923 of this issue, Wyithe and Loeb5 argue that a substantial fraction — up to a third — of a recently discovered population of quasars are strongly lensed objects. These quasars are the most distant astronomical objects known and the light from them, emitted billions of years ago, enables us to probe the history of the very early Universe. If Wyithe and Loeb are right, their conclusions will be both good and bad news for cosmologists: some of the problems in understanding the formation and evolution of structure in the early Universe will be made easier, others more difficult.
First, the good news. The quasars, discovered by the Sloan Digital Sky Survey (SDSS)6, are astonishingly bright. They were discovered with exposure times of only a few minutes using a 2.5-m telescope — quite a small instrument by modern standards. A measure of how far the quasars' light has travelled to reach us is given by their redshift, z. For this distant population z ≈ 6, which means that the Universe (and the wavelength of all photons travelling through it) has expanded by a factor of 1 + z, so in this case a factor of seven, since the light was emitted. These high-redshift quasars are roughly as luminous as any that have been discovered at lower redshifts, and which therefore existed much later in the history of the Universe.
This is puzzling because quasars are usually believed to be powered by the rapid infall of streams of gaseous matter into supermassive black holes at the centres of their host galaxies7. Given their brightness, a quasar in this z ≈ 6 population would typically be expected to contain a black hole that has a mass three billion times that of the Sun, swallowing matter at a rate approaching 100 solar masses per year! The daunting problem for theories of structure formation in the Universe is to understand how such huge black holes and the vast reservoirs of gaseous fuel were assembled so soon after the Big Bang8,9 — at the z = 6 epoch the Universe was less than a billion years old, less than a tenth of its present age.
But if strong gravitational lensing is as common in this quasar population as Wyithe and Loeb propose, the problem is eased significantly. They estimate that as many as a third of the observed z ≈ 6 quasars may actually be at least ten times fainter than they appear, and most may be magnified by a smaller but still significant factor. The problem would not be entirely resolved though, as a quasar powered by a black hole ten times less massive and absorbing matter ten times more slowly still poses an interesting theoretical challenge.
But now the bad news. If gravitational lensing of the z ≈ 6 quasars is common, then determining the properties of that population is more complicated and more dependent on the parameters chosen in models of the early Universe. Naive interpretation of the quasar data, ignoring lensing effects, will lead us to overestimate not only their individual luminosities but also their total numbers and cumulative contributions to the mean radiation environment of the Universe in early times. For example, understanding the radiation environment is crucial to understanding the temperature and ionization of the diffuse gas that is found between galaxies and quasars and that is believed to have contained most of the ordinary matter in the Universe at that epoch10. Wyithe and Loeb suggest that we are seeing this early cosmic epoch modified by a complex set of gravitational lensing effects and distortions. We can attempt to correct for them, of course, but only at the cost of introducing an additional layer of systematic uncertainty and model dependence. Studies of the early Universe are difficult enough without a clouded and unreliable view of the proceedings.
But why, if Wyithe and Loeb5 are right, should strong gravitational lensing be 30 to 100 times more common among the z ≈ 6 quasars than among the far more numerous and extensively studied low-redshift quasars (z ≈ 2, say)? Obviously, the z ≈ 6 quasars are more distant, thus increasing the chance of an intervening object being near our sight line and acting as a gravitational lens. But this makes only a small contribution to the difference. The main effect is due to a selection bias — as magnification makes a strongly lensed quasar appear brighter, it also makes it more likely to be discovered. The importance of such selection biases has long been understood11,12 and early worries that they might compromise our ability to study the quasar population and its evolution were voiced12 soon after the discovery of the first lensed system. But Wyithe and Loeb make a plausible case that we may only now have found the first heavily lensed population of astronomical objects.
Fortunately, a direct observational test of Wyithe and Loeb's hypothesis is possible. The strong gravitational lensing they postulate typically produces multiple images of a source, as well as magnifying it. These images are expected to be so close together in the sky (around 10−4 degrees apart) that they may not be resolvable in the SDSS images. But they should be readily resolved by other ground-based telescopes and by the Hubble Space Telescope. Such observational programmes are being initiated. We should soon have a direct measurement of the lensing that Wyithe and Loeb predict.
Walsh, D., Carswell, R. F. & Weymann, R. J. Nature 279, 381–384 (1979).
Wambsganss, J. Living Reviews in Relativity 1, No. 12 (1998); http://www.livingreviews.org/Articles/Volume1/1998-12wamb
King, L. J., Browne, I. W. A., Marlow, D. R., Patnaik, A. R. & Wilkinson, P. N. Mon. Not. R. Astron. Soc. 307, 225–235 (1999).
Paczynski, B. Astrophys. J. 304, 1–5 (1986).
Wyithe, J. S. B. & Loeb, A. Nature 417, 923–925 (2002).
Fan, X. et al. Astron. J. 122, 2833–2849 (2001).
Rees, M. J. Annu. Rev. Astron. Astrophys. 22, 471–506 (1984).
Turner, E. L. Astron. J. 101, 5–17 (1991).
Haiman, Z. & Loeb, A. Astrophys. J. 503, 505–517 (2001).
Madau, P., Haardt, F. & Rees, M. J. Astrophys. J. 514, 648–659 (1999).
Turner, E. L., Ostriker, J. P. & Gott, J. R. Astrophys. J. 284, 1–22 (1984).
Turner, E. L. Astrophys. J. Lett. 242, L135–L139 (1980).