Annu. Rev. Astron. Astrophys. 2005. 43: xxx-xxx Copyright © 2005 by Annual Reviews. All rights reserved

3. THE COSMIC INFRARED BACKGROUND

The CIB is the infrared part of the extragalactic background, the radiation content of the Universe today produced by galaxies at all redshifts and seen as an isotropic extragalactic background radiation. Partridge & Peebles (1967) predicted that observations of such a background could give powerful constraints on the cosmological evolution.

3.1. General Observations and Direct Cosmological Implications

The detection of the infrared part of the extragalactic background (the CIB for Cosmic Infrared Background) was the major objective of the DIRBE experiment aboard COBE. In fact, the CIB was first detected at long wavelengths by using the FIRAS spectrometer: > 200 µm (Puget et al. 1996). The CIB has subsequently been detected by DIRBE at 2.4, 3.5, 100, 140, 240 µm (see Hauser & Dwek 2001 and Kashlinsky 2005 for two reviews). The extragalactic background at 2.4 and 3.5 µm is significantly larger than that predicted by the integrated galaxy counts and their extrapolation. Similarly, the ext
ragalactic background in the optical has been finally evaluated in combining several methods by Bernstein et al. (2002) and found to be larger than the value given by the integrated fluxes of galaxies by a factor larger than 2. In the mid-infrared, the interplanetary zodiacal dust emission is so strong that only upper limits were obtained by DIRBE. The combination of number counts by ISO/ISOCAM at 15 µm (see Elbaz & Cesarsky 2003) and by Spitzer at 24 µm (e.g., Papovich et al. 2004) giving lower limits, with the observations of TeV gamma ray emission from distant AGNs (e.g., Renault et al. 2001; Dwek & Krennrich 2005), gives a good measurement of the background at these wavelengths. The full cosmic background spectrum is shown in Figure 2. Only most recent and strongly constraining measurements have been plotted for clarity.

Figure 2. The extragalactic background over three decades in frequency from the near UV to millimeter wavelengths. Only strongly constraining measurements have been reported. We show for comparison in grey an SED of M82 (Chanial, 2003) - a starburst galaxy at L = 3 × 1010 L - normalized to the peak of the CIB at 140 µm. References for data points are given in Table 1.

Figure 2 clearly shows that the optical and infrared cosmic backgrounds are well separated. The first surprising result is that the power in the infrared is comparable to the power in the optical. In contrast, we know that locally, the infrared output of galaxies is only one third of the optical output. This implies that infrared galaxies grow more luminous with increasing z faster than do optical galaxies. A second important property to note is that the slope of the long wavelength part of the CIB, I (Gispert et al. 2000), is much less steep than the long wavelengths spectrum of galaxies (as illustrated in Figure 2 with the M82 SED). This implies that the millimeter CIB is not due to the millimeter emission of the galaxies that account for the peak of the CIB ( 150 µm). The implications in terms of energy output have been drawn by, e.g. Gispert et al. (2000). The infrared production rate per comoving unit volume (a) evolves faster between redshift zero and 1 than the optical one and (b) has to stay roughly constant at higher redshifts up to redshift 3 at least.