Quasars in the Early Universe: Smokestacks of the first Cosmic Cities?

A team of astronomers, including members of the Max-Planck Institute for Astrophysics (MPA), have combined the observational power provided by the Hubble Space Telescope (HST) with the predictive power of the Millennium Run cosmological simulations to investigate an intriguing cosmic puzzle. If luminous quasars in the early Universe mark the regions that were the first to collapse and form massive clusters of galaxies as predicted by theory, then why is the observational evidence for such cosmic ”cities-under-construction“ currently so scarce?

Fig. 1: The distribution of z~6 galaxies on the sky as simulated by the Millennium Run simulations. Contours indicate regions of equal density with over- and underdense regions shown in blue and red, respectively. The mean density is shown in green. Large black circles mark galaxies that end up in galaxy clusters at z=0. These proto-cluster regions stand out as relatively overdense regions already at z~6.

Fig. 2: Close-up views of three of the richest regions at z~6 encountered in the simulations. The green square marks the position of the most massive halo (possibly the quasar). Large and small dots correspond to bright and faint galaxies, respectively. The scale bar at the top left in each panel corresponds to the size of an HST/ACS pointing similar to that used to observe real z~6 quasars. Blue circles identify galaxies that will become part of a massive galaxy cluster in the present-day Universe.

Fig. 3: The number of companion galaxies in cubic regions of (20 Mpc/h) ³ versus the mass of the most massive halo in each region. The panel on the left shows the results for faint galaxies (27.5 mag), while the panel on the right shows the results for brighter galaxies (26.5 mag). Although at z~6 there is a large scatter in the number of companions per halo mass, the most massive halos tend to have the most companions. The scatter reduces when going to fainter magnitudes (left panel). The small squares in the right-hand panel correspond to the three regions shown in Fig. 2.

Theory and simulations both predict that the highly rare, luminous quasars discovered at z~6 (roughly one billion years after the Big Bang) are powered by supermassive black holes in the nuclei of massive galaxies that formed deep inside the gravitational potential wells of the densest regions. Therefore, it has been a longstanding prediction - and widely-used assumption - that the quasars be surrounded by vast numbers of smaller galaxies that trace these dense regions. This prediction is consistent with the fact that the most massive and oldest galaxies and the largest (dormant) black holes in the present-day Universe are found at the centers of massive galaxy clusters. If, however, the prediction is proven to be false, theories of the formation of quasars in the early Universe may have to be revised.

In order to test the theory, attempts have been made to detect faint star-forming galaxies using the Advanced Camera for Surveys (ACS) on-board the HST, with mixed results. While the fields surrounding some quasars show a possible excess of galaxies, most quasar fields do not appear any different from random sightlines observed to a similar depth. Also, some studies have found arbitrary regions in the sky that contain structures of galaxies at z~6 that substantially outnumber regions near quasars. Thus, it appears that crucial observational evidence relating quasars and overdense regions in the early Universe is currently lacking. A recent study led by Roderik Overzier (MPA) gives a possible explanation for this unexpected result.

Using the Millennium Run N-body simulations coupled with semi-analytic models of galaxy formation developed at the MPA in Garching (linkPfeil.gifResearch Highlights August 2004 and linkPfeil.gifJune 2007), the team has simulated a very large region of the early Universe to show what it would look like through the eyes of our largest telescopes. This simulated survey shown in Fig. 1 predicts the locations and magnitudes of galaxies and shows how the Universe obtained its characteristic web-like shape as early as z~6. This pattern on the sky is analogous to a system of long highways (”filaments“) inter-connecting large cities of galaxies between vast open regions (”voids“). Quasars are expected to be hosted by the densest regions seen in Fig. 1. However, several important effects combine to explain why the hunt for faint galaxies surrounding quasars is so challenging. First, although current surveys are quite successful at identifying relatively bright galaxies lying roughly at z~6, we are not yet able to determine the exact redshifts of these galaxies very efficiently. In order to prove a physical association between galaxies and the target quasars (of which the redshifts are known), much more precise redshift information would be required. Second, the simulations suggest that the sensitivity and covered area of the surveys performed to date may not be optimal for finding structures of galaxies associated with the quasars. In Fig. 2 we show some examples of high density regions found in the simulations. The large black circles indicate the faintest galaxies that can currently be detected in quasar fields. The red bar at the top left corresponds to the diameter of the HST/ACS field-of-view (3.4 arcmin). Because the number of bright galaxies is relatively small and they scatter over an area typically two or three times larger, it is easy to miss any structures, if present, in the observations. By observing fainter galaxies (small dots in Fig. 2) the large-scale environments stand out much more clearly making them more easy to detect. Third, the team analysed the number of companion galaxies as a function of the mass of dark matter halos assumed to be hosting the quasars, finding that the two quantities are only mildly correlated (Fig. 3). This implies that even if it can be shown through observations that quasars are preferably located in overdense regions, it will be very difficult to determine the exact mass of the quasar halo from the number of companion galaxies observed.

Thanks to the success of the recent NASA HST Servicing Mission (linkPfeilExtern.gif http://www.nasa.gov/mission_pages/shuttle/shuttlemissions/hst_sm4/index.html), the newly repaired ACS will be able to perform new surveys that could target regions around the z~6 quasars. Alternatively, ground based telescopes should be used in order to determine the exact redshifts of faint galaxies near quasars by targeting the Lya emission line. Such studies are best performed by targeting quasars at z~5.7, corresponding to a good atmospheric transmission window. In the next decade, the James Webb Space Telescope (JWST), with its sensitive near-infrared camera and spectrograph able to detect extremely faint galaxies, should be able to provide a definitive answer to the question of whether the luminous quasars formed inside the densest regions in the early Universe.


Roderik Overzier


For further information see:

Overzier, R.A., Guo, Q., Kauffmann, G., De Lucia, G., Bouwens, R., Lemson,G., "LCDM predictions for galaxy protoclusters - I. The relation between galaxies, protoclusters and quasars at z~6", 2009, MNRAS, 394, 577