We chose the HST/WFPC2 field (see Table 1 for coordinates) at an ecliptic latitude || > 30° (to minimize contributions from ZL) and near the Galactic pole in a region of low Galactic 100 µm emission (to minimize diffuse Galactic light). We also selected the field to avoid bright stars, in an effort to minimize the scattered light from stars near the optical axis of the telescope. The positions of stars relative to our field can be seen in the r-band CCD image shown in HST Figure 4 and in the IRAS map in Figure 3. No stars brighter than 12 AB mag fall within 10 arcmin of the center of our field, and no stars brighter than 7 AB mag fall within 2.5 degrees. Based on measurements of the point spread function (PSF) of HST done with WFPC1, the attenuation factor at 1 arcmin from the center of a point source is 10-8 (STScI Technical Memos RSB-85-03, RSB-85-02, ISR/OTA 06.1). Furthermore, the "large-angle" ( > 3 arcsec) scattering is roughly one order of magnitude lower for WFPC2 than for WFPC1. Based on these results, the total contribution from stars closer than 5 arcmin to the field is at most 10-9 of their total flux, which amounts to one 35 mag star on-axis. The EBL signal, by comparison, is approximately 27.5 mag arcsec-2. The flux from off-axis sources is therefore an insignificant contribution to the background and a negligible source of error.
Figure 4. The HST/WFPC2 and FOS footprints overlaid on a mosaic Gunn-r-band image (0.6° × 0.6°) taken with the 1m Swope Telescope at the Las Campanas Observatory.
The exact scheduling of exposures was also critical to this program, as stray solar and terrestrial light in HST observations are a strong function of the orbital position of the satellite. Sunlight scattered off the limb of the Earth can increase the background level by a factor of 10 when the bright limb is near the optical axis of the telescope. All of our science observations were therefore scheduled to execute exclusively in the shadow of the Earth. Because our field was at a viewing angle angle greater than 135° from the Sun during the months of our observations, the satellite was pointed in a direction greater than 45° away from the satellite's direction of motion when our field was observed from within the Earth's shadow. This guaranteed that no upper-atmosphere glow from the satellite's flight path would affect our observations. Scattered moonlight was also explicitly avoided by exposing only with the Moon greater than 65° from the optical axis of the telescope, which is the angular separation at which models predict that the attenuation function for off-axis scattered light becomes flat (STScI Technical Memo RSB-85-03). Based on those models, confirmed by on-orbit measurements (Burrows 1991, Hasan & Burrows 1993), the light from the moon at 65° from the optical telescope assembly (OTA) is less than 10-5 photons s-1 arcsec2 per 100Å, which is 1000 times smaller than the lower limit expected for the EBL, and is therefore also a negligible source of background during our observations.
Our 18 orbits were divided into three visits, staggered by one month each between October and December, 1995. The ZL flux changes with the Earth's orbital position about the Sun and the line of sight through the interplanetary dust during the three visits. We therefore expected to see changes in the total sky flux between visits, at a level predicted by the models of the interplanetary dust. During all three months, our target field has a ZL surface brightness which is within 20% of the absolute minimum intensity of the ZL at any orientation. By separating the orbits for this program into three visits, we were also able to observe the field at two different roll angles - V3 position angle 222° for the November and December visits, V3 angle 132° for the October visit - thereby changing the position of the off-axis stars with respect to the OTA. This allows us to further rule out significant off-axis scattered light and other possible photometric anomalies.