The integrated optical flux from all extragalactic sources is a record of the stellar nucleosynthesis in the universe and the chemical evolution which has resulted from it. In Bernstein, Freedman, & Madore 2001a (henceforth, Paper I), we presented detections of the optical EBL in the HST/WFPC2 wide-band filters F300W (U300, 0 ~ 3000Å), F555W (V555, 0 ~ 5500Å), and F814W (I814, 0 ~ 8000Å) based on simultaneous data sets from Hubble Space Telescope (HST) and Las Campanas Observatory (LCO). In Bernstein, Freedman, & Madore 2001b (henceforth, Paper II), we presented details of a measurement of the diffuse foreground zodiacal light which we use in Paper I. Here we briefly summarize the results of Papers I and II and discuss the cosmological implications of these detections of the EBL.
The majority of the EBL at UV to IR wavelengths is produced by stars at restframe wavelengths of 0.1-10µm. Due to cosmic expansion, the EBL at U300, V555, and I814 potentially includes redshifted light from stellar populations out to z ~ 8 (the redshifted Lyman-limit cut-off of the I814 filter). Although stars themselves do not emit much light at wavelengths longer than 10µm, a complete census of the energy produced by stellar nucleosynthesis in the universe must consider the EBL over the full wavelength range 0.1-1000µm because dust in the emitting galaxies will absorb and re-radiate starlight, redistributing energy from nucleosynthesis into the thermal IR.
With 8m-class telescopes and HST, the limits of resolved-source methods (i.e., number counts, redshift surveys, QSO absorption lines, etc.) for studying star formation in the universe are being extended to ever fainter levels; however, a direct measurement of the EBL remains an invaluable complement to these methods. Galaxies with low apparent surface brightness - both intrinsically low surface brightness galaxies at low redshift and normal surface brightness galaxies at high redshift - are easily missed in surface-brightness-limited galaxy counts and consequentally in follow-up redshift surveys. Identification, not to mention photometry, of faint galaxies becomes very uncertain near the detection limits. Even efforts to understand galaxy evolution, chemical enrichment, and star formation through QSO absorption line studies appear to be biased against chemically enriched, dustier systems, as these systems can obscure QSOs which might lie behind them (Fall & Pei 1989, Pei & Fall 1995, Pettini et al. 1999). In contrast, a direct measurement of the spectral energy distribution (SED) of the EBL from the UV to the far-IR is a complete record of the energy produced by star formation and is immune to surface brightness selection effects.
In addition to energy originating from stellar nucleosynthesis, the EBL includes energy emitted by accreting black holes in quasars and active galactic nuclei. However, at optical wavelengths, the quasar luminosity functions at redshifts z 5 indicate that the optical luminosity density from quasars is a small fraction (~ 2.5%) of the that from galaxies (e.g. Boyle & Terlevich 1998). In addition, our measurement of the EBL excludes any point-like sources (of which there are 3 in our images), under the prior assumption that those sources are Galactic foreground stars. We therefore expect quasars to be a negligible source of flux in our measurements of the optical EBL.
The contribution from active galactic nuclei (AGN) is more difficult to assess, as recent dynamical evidence (Richstone et al. 1998) indicates that massive black holes reside in most galaxies and sensitive optical spectroscopy (Ho et al. 1997a, 1997b) indicates that AGN have at least a weak contribution to more than 50% of nearby galaxies. Nonetheless, simple accretion models, the total X-ray background, and the X-ray to far-IR spectral energy distribution of AGN and quasars all indicate that the total contribution to the bolometric EBL from accretion onto central black holes is 15% (see Section 6.2), and is emitted at thermal IR wavelengths. In principle, measurements of the EBL also constrain possible the total energy output from more exotic sources, such as gravitationally collapsing systems, brown dwarfs, and decaying particles (see Carr, Bond, & Hogan 1986, 1991 and Dwek et al. 1998 for discussions).
The outline of the paper is as follows. In Section 2, we give an overview of the observations and methods used to measure the EBL as discussed in Papers I and II. In Section 3, we summarize the individual measurements and associated errors we have obtained from each data set and the final EBL detections which result from them. In Section 4, we compare the measured EBL with the integrated optical flux from resolvable sources as quantified by number counts and luminosity functions. In Section 5, we quantify the contributions to the optical EBL which one might expected from sources which fall below the detection limits of the HDF based on explicit assumptions regarding the surface brightness, luminosity, and redshift distribution of galaxy populations in the universe. In Section 6, we discuss models of the SED of the EBL based on these and recent results in the far infrared. Finally, in Section 7 we discuss the total star formation and chemical enrichment history of the universe required to produce the bolometric flux of the EBL, and compare the inferred values to other observations of the total baryon fraction in stars and the metal mass density in the local universe. We abbreviate the adopted units ergs s-1 cm-2 sr-1 Å-1 as cgs throughout.