2002ApJ...571...56B

We present the first detection of the mean flux of the optical extragalactic background light (EBL) at 3000, 5500, and 8000 Å. Diffuse foreground flux at these wavelengths comes from terrestrial airglow, dust-scattered sunlight (zodiacal light), and dust-scattered Galactic starlight (diffuse Galactic light). We have avoided the brightest of these, terrestrial airglow, by measuring the absolute surface brightness of the night sky from above the Earth's atmosphere using the Wide Field Planetary Camera 2 (WFPC2) and Faint Object Spectrograph, both on board the Hubble Space Telescope (HST). On the ground we have used the du Pont 2.5 m Telescope at the Las Campanas Observatory (LCO) to obtain contemporaneous spectrophotometry of ``blank'' sky in the HST field of view to measure and then subtract foreground zodiacal light from the HST observations. We have minimized the diffuse Galactic light in advance by selecting the HST target field along a line of sight with low Galactic dust column density and then estimated the low-level Galactic foreground using a simple scattering model and the observed correlation between thermal, 100 µm emission, and optical scattered flux from the same dust. In this paper, we describe the coordinated LCO/HST program and the HST observations and data reduction and present the resulting measurements of the EBL. Galaxies brighter than V=23 AB mag are not well sampled in an image the size of the WFPC2 field of view. We have therefore measured the EBL from unresolved and resolved galaxies fainter than V=23 AB mag by masking out brighter galaxies. We write this as EBL23 to emphasize this bright magnitude cutoff. From absolute surface photometry using WFPC2 and ground-based spectroscopy, we find mean values for the EBL23 of 4.0 (±2.5), 2.7 (±1.4), and 2.2 (±1.0) in units of 10^–9 ergs.s^–1.cm^–2.sr^–1.Å^–1 in the F300W, F555W, and F814W bandpasses, respectively. The errors quoted are 1 σ combined statistical and systematic uncertainties. The total flux measured in resolved galaxies with V>23 AB mag by standard photometric methods is roughly 15% of the EBL23 flux in each band. We have also developed a new method of source photometry, uniquely suited to these data, with which we can measure the ensemble flux from detectable sources much more accurately than is possible with standard methods for faint galaxy photometry. Using this method, we have quantified systematic biases affecting standard galaxy photometry, which prevent light from being recovered in isophotes within a few percent of the sky level. These biases have a significant effect on faint galaxy counts. The flux from resolved sources as measured by our ensemble photometry method is 3.2 (±0.22), 0.89 (±0.01), and 0.76 (±0.01) in units of 10^–9 ergs.s^–1.cm^–2.sr^–1.Å^–1 in the F300W, F555W, and F814W bandpasses, respectively, with 1 σ combined errors. These values, the total flux from resolved sources, represent absolute minima for the EBL23 in each band and are roughly 30% of the mean flux we measure for the total EBL23.