6. SUMMARY AND IMPLICATIONS

To put the results of the COBE CIB measurements in a broader perspective, Figure 2 shows the extragalactic background light from the UV to submillimeter wavelengths. All of the measurements are shown with 2 error bars. Measurements based upon COBE data are as shown in Figure 1 (see Table 1 for references). The recently reported CIB measurement from the IRTS mission (Matsumoto et al. 2000) is shown as a curve from 1.4-4.0 µm. The detections at UV-optical wavelengths are from the recent report by Bernstein (1999) ( I = 12, 15, and 18 nW m^-2 sr^-1 at 0.3, 0.55 and 0.8 µm respectively). The lower limits to the background light from UV to 2.2 µm are obtained from integrated galaxy counts obtained from the ground and the Hubble Space Telescope (summarized by Madau & Pozzetti 2000). Lower limits in the mid and far infrared are based upon galaxy counts with the ISOCAM (7-15 µm) and ISOPHOT (90-175 µm) instruments on the ISO mission (Altieri et al. 1999, 7 and 15 µm; Clements et al. 1999, 12 µm; Juvela, Mattila, & Lemke 2000, 90 and 150 µm; Puget et al. 1999, 175 µm). The lower limit at 850 µm is based on the integrated light from source counts with the SCUBA instrument (Blain et al. 1999). The shaded region in Figure 2 indicates 2 upper and lower limits on the brightness of the extragalactic background light based upon all of these data.

Figure 2. Measurements of the extragalactic background light. See text for references.

Figure 2 shows that the resolved component of the extragalactic light is beginning to approach the detected background at the shortest and longest wavelengths in this range. This reflects a dramatic change in observational knowledge over the past few years. It is reassuring that the resolved component has not exceeded the claimed CIB detections or upper limits at any wavelength. The CIB at 140 µm and in the near infrared at 2.2 and 3.5 µm is relatively well-determined by the direct sky brightness analyses. However, systematic uncertainties in the foregrounds are substantial, with none of the CIB detections being positive by more than several times the systematic uncertainties. Isotropy tests, if done at all, are generally over very limited sky areas. It would therefore seem premature to consider the differences between detections and lower limits as clear evidence for missing components of the background. In terms of energy distribution, the I values in the shortest and longest wavelength regions are rather comparable.

In the range from 4.9 to 100 µm the picture is less certain. Limits from the direct brightness analyses are relatively high, and quite possibly strongly contaminated by residual emission from the interplanetary dust. The upper limits from fluctuation analyses suggested by Kashlinsky et al. (1996b) and Kashlinsky & Odenwald (2000) in this range are certainly the most restrictive, though dependent upon their estimate of the relationship between fluctuations and total light. Other fairly restrictive upper limits in this spectral range come from analysis of attenuation of TeV -rays, as reviewed elsewhere at this conference (Stecker 2001).

Taking the most restrictive limits at face value, the emerging picture is one of substantial background energy in the far infrared and submillimeter, with a comparable level in the near infrared. To make this picture somewhat quantitative, one can integrate the background intensity in the near infrared (1.25-3.5 µm), thermal infrared (3.5-100 µm), and submillimeter (100-2000 µm) ranges. One finds integrated intensities of ~ 20, < 56, and ~ 16 nW m^-2 sr^-1 respectively. Using the measurements reported by Bernstein (1999), the UV-optical background (0.3-1.25 µm) contains ~ 25 nW m^-2 sr^-1 of additional background energy. Thus, 0.4 I_FIR / I_{UV - OPT} 1.6, where I_FIR is the integrated long-wavelength background (3.5-2000 µm) and I_{UV - OPT} is the integrated short-wavelength background (0.3-3.5 µm). The long-wavelength and short-wavelength backgrounds contain comparable total energy.

Assuming that the short-wavelength energy comes primarily from redshifted starlight and that the long wavelength radiation comes from starlight absorbed and reradiated by dust, the large far-infrared background implies that a substantial amount of cosmic star formation has been enshrouded in dust, a conclusion reached by many investigators as soon as evidence for the CIB emerged (e.g., Puget et al. 1996; Schlegel et al. 1998; Hauser et al. 1998). That interpretation has been reinforced by the deep galaxy counts at far infrared wavelengths with the ISO and SCUBA instruments. The total background energy in the spectral range 0.3-2000 µm, I_EBL, is given by I_EBL 60-120 nW m^-2 sr^-1, or, in terms of the critical density,

where h = H₀ / (100 km sec^-1 Mpc^-1) and H₀ is the Hubble constant. For comparison, the integrated EBL contains only about 10% of the integrated energy in the cosmic microwave background, _CMB = 2.5 × 10^-5 h^-2.

With the assumption that the integrated extragalactic background energy primarily arises from nucleosynthesis in stars, one can estimate the mass density consumed in the production of helium and heavier elements, _Z, from the relation (Peebles 1993):

Expressing _Z as a fraction of the critical mass density, _c, this yields

where we have assumed that most of the energy release occurs at redshift z_e 1. Since Big Bang nucleosynthesis arguments give a cosmic baryon mass density (Kolb & Turner 1990; Steigman et al. 1999)

this simple argument suggests that the current estimate of the integrated extragalactic background light implies conversion of 1-8% of the cosmic hydrogen into helium and heavier elements by stars.

The present knowledge of the CIB provides significant constraints on models of star formation and galactic evolution (e.g., Dwek et al. 1998; Pei, Fall, & Hauser 1999). These topics are addressed by others at this conference.