Copyright © 2001 by Annual Reviews. All rights reserved
|
| Annu. Rev. Astron. Astrophys. 2001. 39:
249-307 Copyright © 2001 by Annual Reviews. All rights reserved |
3.10. Summary of Extragalactic Background Observations
As the preceding Sections show, there is now a great deal of direct and
indirect observational evidence providing CIB upper limits, lower
limits, and tentative or probable detections.
Figure 5 summarizes this evidence, together with
the data at UV and optical wavelengths to provide a more complete
picture of the EBL. References for the direct infrared background
measurements and the limits from infrared fluctuation measurements are
in Tables 1 and
2,
respectively. References for the UV-optical measurements are in
Table 4.
Figure 5 also shows the lower limits on the EBL
found from the integrated light of resolved extragalactic
sources. References for these limits are in
Table 3. Missing from this figure
for the sake of
clarity are the CIB upper limits deduced from analysis of TeV
-ray data,
which are shown in
Figure 4. The dotted line in
Figure 5 indicates the nominal measurement where
detections are reported, or a somewhat arbitrary intermediate value
between upper and lower limits where there is no claimed detection.
 |
Figure 5. Summary of extragalactic
background light measurements and limits. Error bars for detections are
1 |
Not surprisingly, the most convincing detections of the diffuse infrared
background are in the spectral windows in the local foregrounds near 3.5
µm and longward of 100 µm. At other wavelengths,
the most certain and constraining results come from integrated galaxy
light. The limits inferred from the less direct measures, including
fluctuations in the infrared sky brightness and attenuation of TeV
-rays, are
somewhat model dependent and therefore less certain. However, they are
particularly useful in the 5 to 100 µm range, where direct
measurements are so difficult.
It is notable that present detections of the EBL and the independently
determined upper and lower limits are not in conflict with each other
and, at some wavelengths, are not very far apart. In particular, the
deepest SCUBA counts at 850 µm
(Blain et al. 1999b)
yield an integrated source light comparable to the CIB brightness
determined from the FIRAS data. Thus, the discovery of the submillimeter
background and resolution of its sources seem to have occurred virtually
simultaneously. In the 7 to 180 µm range, the counts from
the IRAS and ISO instruments are rising too steeply for the light to
have converged, so the integrated light is necessarily below the actual
CIB. However, the lowest TeV
-ray limits
on the CIB are only modestly greater than the integrated light from the
ISO 7 and 15 µm number counts, which suggests that the
counts may be resolving a significant fraction of the CIB at these
wavelengths.
The situation in the optical and near-infrared (0.3-2.2 µm)
is somewhat different. Although the galaxy counts have been extended to
sufficiently faint levels that the integrated light has converged
(Madau & Pozzetti
2000),
there remains an apparent gap between the integrated galaxy light and
the reported EBL measurements. However, there may be systematic
underestimates by a factor of ~ 2 in the integrated galaxy light due to
photometry incompleteness at low surface brightness, as argued by
Bernstein (1999).
Furthermore, the claimed EBL measurements in this range are less than
4
positive and are at
most only slightly more than
2 above the integrated
galaxy light. In view of the potentially large systematic errors in both
the integrated galaxy light and the extragalactic background
determinations, one cannot confidently conclude that there is a
significant difference between these two measures. Such a difference
would imply the existence of unidentified discrete or diffuse sources
contributing to the background light.
The shaded region in Figure 5 indicates conservative upper and lower limits on the spectral energy distribution of the EBL based on all available measurements and their uncertainties. We use these limits in the discussions of implications (Section 4) and models (Section 5).