2.1. WFPC2 Mid-UV Survey Strategy
2.1.1. Mid-UV Filters
With the HST Wide Field and Planetary Camera 2 ("WFPC2"), we have obtained images of 37 nearby galaxies through one or, whenever possible, two wide-band mid-UV filters below the atmospheric cutoff. These filters are F300W (cent 2930Å; 740Å FWHM) and F255W (cent 2550Å; 395Å FWHM), which provide reasonable red-leak suppression. The F255W, F300W, and the Johnson U and B filters are approximately equally spaced in energy (i.e., in the logarithm of the wavelength), and so add significantly to the existing ground-based optical-near-IR color baseline.
Since the HST/WFPC2 system throughput is ~ 2.0% in F300W and ~ 0.5% in F255W (Biretta et al. 2001; Appendix 1), we can only detect the highest SB, bluest objects in F255W in a single HST orbit and so have selected our sample accordingly. The mid-UV is the longest wavelength where younger stars can dominate the integrated galaxy light, and therefore the regime of choice to measure the SFR averaged over 1 Gyr. We have observed all selected galaxies through the F300W filter, spending no more than one full orbit per galaxy. In that same orbit a short exposure through a red filter (F814W) is taken for adequate red-leak correction (see Section 2.4.3). For galaxies in the HST Continuous Viewing Zone (CVZ), we also took exposures in the F255W filter (see Section 2.1.4).
2.1.2. Predicted Mid-UV Surface Brightness
We predict the average mid-UV SB, µF300W, for a given galaxy from its total B magnitude, BT, its (U - B) color, its half-light radius, re, and its ellipticity, b/a, (as tabulated in or derived from the RC3 catalogue [de Vaucouleurs et al. 1991], or the NASA/IPAC Extragalactic Database, NED) as following:
|µF300W = F300WT + 0.75 + 2.5 log ( re2 × b/a) ,||(1)|
i.e., half the total predicted F300W magnitude, F300WT, within the effective area. We used the updated Bruzual & Charlot (1993) models to transform the (U - B) color for each galaxy type to a predicted (F300W - B) color, from which F300WT follows. To a reasonable approximation we find (F300W - B) 2 × (U - B). In the absence of a (U - B) color, a prediction for (U - B) was made from the measured (B - V) color and the known (U - B) vs. (B - V) relation for RC3 galaxies as a function of galaxy type (de Vaucouleurs et al. 1991). Our sample was selected to have 18 µF300W 22.5-23.0 mag arcsec - 2. For this range in SB, a galaxy can be detected out to r 2-3 re with WFPC2 in one orbit with sufficiently high S/N to allow morphological features to be recognized.
The bias toward selecting higher SB galaxies can be addressed as in Driver et al. (1995b). In short, selecting high SB galaxies as the nearby template objects is not an overriding concern, since the high-redshift samples are similarly biased (or more so) in favor of high SB galaxies due to the severe cosmological SB dimming. For monochromatic light, SB-dimming is proportional to (1 + z)(4+), with the spectral index if the object spectrum were to be represented by a power-law SED.
The resolution of HST's Optical Telescope Assembly (OTA) in F300W is ~ 0".04. This is somewhat larger than HST's formal diffraction limit at 2930Å (1.22 · / D 0".03), which doesn't set in until longward of 4000Å due to mirror micro-roughness. The WFPC2 WFC pixel size is 0".0996/pixel. Hence, because the UV images are already severely undersampled, on-chip rebinning to gain SB-sensitivity is not an option. Instead, where needed, we can rebin the images in the post-processing stage to measure the outskirts to fainter SB-levels. This improves the SB-sensitivity (see Section 2.4.2) in the outskirts to ~ 25.8-26.3 mag arcsec - 2 in F300W and to ~ 23.8-24.2 mag arcsec - 2 in F255W, sufficient to get good light-profiles for r 2-3 re.
2.1.3. Target Size and Placement Inside WFPC2
The WFPC2 FOV measures ~ 2.5 (along the WFC CCDs). We selected the sample to fit within the FOV, and preferably within a single 75" × 75" WFC CCD, allowing us to derive reliable SB-profiles without having to mosaic multiple WFPC2 fields. For galaxies with a B-band half-light radius in the range 0.'1 re 1.'0 (as derived from the RC3 catalogue, de Vaucouleurs et al. 1991), about ~ 3-5 scale-lengths fit in a single WFPC2 field.
For most of our sample galaxies, the nucleus has been placed on WFPC2's WF3 CCD, near pixel (X, Y) = (300, 300), so that both the WF2 and WF4 CCDs maximally sample the galaxies' outskirts, allowing optimal subtraction of any sky-background when mosaicing the four WFPC2 CCDs. For some of the larger galaxies and for galaxies in pairs or small groups, we center the object(s) in another part of the WFPC2 FOV, or constrain the HST roll angle ("ORIENT") to assure that the largest possible portion of the galaxy or galaxy group is observed.
2.1.4. The HST Continuous Viewing Zone
Part of the galaxy sample is located in the HST Continuous Viewing Zone (CVZ), i.e., at 53° |DEC| 72°, where objects are observable for an entire HST orbit, typically doubling the available integration time. For any such galaxies, we were able to obtain F255W as well as F300W images without the cost of an extra HST orbit. We select the sample to maximize the fraction of galaxies in the CVZ.
Since the Zodiacal background reflects the color of the Sun, the sky-background will be darker in the F255W filter (~ 24.5-25.0 mag arcsec - 2) than in the F300W filter (~ 24.0 mag arcsec - 2), partly compensating for the lower sensitivity in F255W. Observations in the CVZ may suffer from higher sky-background levels due to the Earth's limb (Williams et al. 1996). We minimize the probability of excessive sky-levels by interspersing the exposures in the different filters using the sequence: F814W, F255W, F300W, F255W, F300W, F814W, F255W, and F300W. This sequence ensures that never more than one F255W or F300W exposure is taken close to the Earth's limb in a full CVZ orbit, and also minimizes the number of Fine Guidance Sensor (FGS) motions needed to create a pointing dither-pattern. For non-CVZ targets we use the sequence: F814W, F300W, F300W, and F814W, to push the F300W exposures farthest from the Earth's limb.