Annu. Rev. Astron. Astrophys. 2001. 39: 249-307
Copyright © 2001 by . All rights reserved

Next Contents Previous

5.1. The Contribution of AGN and Other Nonnuclear Sources to the CIB

The energy output from AGN represents the major nonnuclear contribution to the radiative energy budget of the universe exclusive of the CMB. The energy of an AGN is derived from the release of gravitational energy associated with the accretion of matter onto a central black hole (BH) located in the nucleus of a host galaxy.

Simple estimates suggest that AGN are not the dominant contributors to the integrated EBL intensity. Accretion onto a BH releases about 10% of the rest mass energy of the accreted matter, which is large compared with the 0.7% conversion efficiency of nuclear fusion. However, the total amount of matter in central BHs comprises only ~ 0.6% of that in stellar objects (Magorrian et al. 1998, Fabian & Iwasawa 1999). The total energy released from AGN should therefore be about 10%-20% of that released from nuclear processes. However, AGN could still contribute a significant fraction of the CIB intensity in specific wavelength regions.

The energy output of AGN spans a wide spectrum, ranging from radio frequencies to gamma-rays (Sanders & Mirabel 1996, Grupe et al. 1998). Many AGN exhibit a thermal infrared excess over the nonthermal continuum, associated with emission from hot (~ 60-100 K) dust (Haas et al. 1998). At far-infrared wavelengths, such sources manifest themselves as luminous or ultraluminous infrared galaxies, first recognized in the IRAS data (Soifer et al. 1986). The infrared emission can be either powered by an AGN or by starbursts, and the relative contributions of these sources to the infrared emission in most objects is not well understood. Phenomenological criteria for distinguishing the energy sources are summarized by Genzel & Cesarsky (2000). However, the interpretation of the observations is difficult, since source geometry, dust extinction, and the competition of dust for UV photons can affect diagnostics such as line ratios and line-to-continuum ratios (Fischer 2000). For example, NGC 6420, a galaxy first believed on the basis of such diagnostics to be primarily powered by starbursts, may in fact be predominantly powered by an AGN (Vignati et al. 1999).

While the relative contributions of starbursts or AGN to the energy budget of individual infrared galaxies is still unclear, recent developments in our understanding of the origin of the cosmic X-ray background (CXB) may clarify the global contribution of AGN to the spectrum of the CIB. Since the discovery of the CXB (Giacconi et al. 1962), understanding its origin has been a challenge. The strong limits set by the COBE / FIRAS instrument on any y-distortion in the spectrum of the CMB (Fixsen et al. 1996) essentially ruled out diffuse thermal emission from a hot intergalactic gas as a significant source of the hard CXB. Type 1 Seyfert galaxies, an unobscured (NH < 1022 cm-2) subclass of AGN, could account for the soft (~ 0.2-1 keV) X-ray background (Hasinger et al. 1998, Schmidt et al. 1998), but the same sources failed to account for the harder component of the CXB (~ 1-40 keV), which is flatter.

This inability to account for the hard CXB with any superposition of known discrete extragalactic sources was a problem referred to as the "spectral paradox" (for reviews, see Boldt 1987, Fabian & Barcons 1992). A new population of objects seemed to be needed to account for the hard X-ray background. Setti & Woltjer (1989) suggested that a superposition of strongly absorbed sources with an imposed low-energy cutoff in the ~ 10 keV range could explain the hard CXB. They argued that the existence of such sources is a natural prediction of the unified theory of AGN (Antonucci 1993). Under the AGN unification scheme, the X-ray spectra of Seyfert 2 galaxies, which are viewed through large H-column densities, will be characterized by a sharp low-energy cutoff between 0.5 and 10 keV. The CXB could arise from a combination of unobscured Seyfert 1 galaxies and Seyfert 2 galaxies with different degrees of obscuration (Madau et al. 1994, Comastri et al. 1995, and references therein). A recent deep X-ray survey with Chandra confirmed that about 75% of the hard CXB originates from obscured AGN, lending support to the CXB models constructed within the unified theory of AGN (Mushotzky et al. 2000).

An immediate consequence of these results is that some fraction of the energy absorbed by the obscuring material must be absorbed by dust and reradiated in the infrared. This provides a link between the sources of the hard CXB and the CIB. Since the SCUBA surveys at 850 µm have resolved most of the CIB, this linkage can be observationally explored by identifying the AGN contribution to the SCUBA sources.

Steps towards that goal were recently taken (Ivison et al. 2000, Severgnini et al. 2000, Barger et al. 2001). Ivison et al., using optical and radio data, found evidence for AGN activity in three of seven SCUBA sources, though it was not clear whether the AGN activity dominated the submillimeter emission of these sources. Severgnini et al. correlated a sample of 850 µm SCUBA sources and 2 to 10 keV X-ray sources from Chandra and BeppoSAX observations at limiting fluxes that resolve more than 75% of the background in each of the two energy bands. They detected only one SCUBA source among the hard X-ray sources, limiting the AGN contribution to the submillimeter background to less than ~ 7%. A significantly larger AGN contribution to the CIB would require a close association of the SCUBA sources with the fainter hard X-ray sources, which contribute at most 25% of the hard CXB.

Barger et al. (2001) conducted deep optical, near-infrared, SCUBA, and radio surveys centered on 20 hard X-ray sources that had been detected in a flux-limited deep Chandra survey. One of the X-ray sources was detected by the SCUBA survey, again implying that only about 10% of the 850 µm CIB arises from such sources.

The CXB-submillimeter connection has been recently explored theoretically in the context of the unified AGN model for the origin of the CXB (Almaini et al. 1999, Gunn & Shanks 2001). The models are essentially backward evolution models (Section 5.2.1), in which the local X-ray luminosity density of AGN is evolved backward in time assuming pure luminosity evolution. Model parameters are constrained so that the AGN produce the hard CXB. With somewhat different assumptions regarding the relation between infrared and X-ray luminosity in individual galaxies, both models conclude that AGN contribute ~ 10%-20% of the measured 850 µm background. Because of the rather simple assumptions about the template infrared SED used in these models, the models are not particularly informative about the AGN contribution at wavelengths shorter than ~ 100 µm.

A more physically oriented approach was used by Granato et al. (1997), who constructed detailed radiative transfer models for calculating the infrared spectral energy distribution (SED) from circumnuclear dusty tori powered by an accreting BH, for a distribution of viewing angles. Adopting parameters that force these objects to produce the hard CXB, Granato et al. found that AGN contribute only about 1% to the CIB intensity at 850 µm, substantially less than the value found with the aid of the recent hard X-ray observations. The differences between their results and those of Almaini et al. (1999), Gunn & Shanks (2001) arise primarily from the different infrared spectra adopted for these objects, and different prescriptions for the evolution of their X-ray luminosity function.

The available evidence therefore strongly suggests that AGN contribute at most only 10%-20% of the CIB. This conclusion is supported by the global energetic argument based on the abundance of BHs, by the observational studies of the relation between AGN, hard X-ray, and submillimeter sources, and by theoretical models for the CIB contributions from the sources of the hard CXB.

Brown dwarfs are other gravitational energy sources that might contribute to the CIB. These are objects whose mass falls below the minimum mass (~ 0.08 Modot) required for stable hydrogen burning (Burrows & Liebert 1993). Karimabadi & Blitz (1984) assumed that these objects radiate as blackbodies and calculated their evolutionary tracks on the H-R diagram. Assuming that all the dark matter required to close the universe is contained in these objects, they found the brown dwarf contribution to the CIB to be at most 3 nW m-2 sr-1 in the 10 to 100 µm wavelength region. Because the assumed amount of mass in such objects was greatly exaggerated, cooling substellar objects clearly make a negligible contribution to the CIB.

An additional nonnuclear contribution to the EBL might arise from the radiative decay of primordial particles. By appropriately selecting the particle number density, particle mass, and decay redshift, one can produce a wide range of background intensities at UV to far-infrared wavelengths (Bond et al. 1986, Wang & Field 1989, Sciama 1998). However, lacking physical evidence to substantiate particular choices, predictions of background contributions from such mechanisms remain highly conjectural.

In summary, it appears reasonable to assume that the energy in the CIB arises largely from nuclear processes. In the following, we review models in which nuclear processes provide the energy for the background radiation.

Next Contents Previous