The single star population synthesis models (Section 3.3) predict a rapid decline in FUV-V with increasing redshift (lookback time), while the binary models suggest no significant change. This stark contrast provides an important test.
GALEX is NASA's UV space telescope mission that can do just this. It has sensitive FUV and NUV detectors and reaches passive (no star formation) old populations (such as many elliptical galaxies) out to z ~ 0.2 (Martin et al. 2005). Its Deep Imaging Survey (DIS) is obtaining the UV images of tens of galaxy clusters using 20,000 seconds of exposure. The UV upturn is found to be the strongest in the brightest cluster elliptical galaxies (BCGs) and hence we have tried to obtain accurate photometric data on the BCGs in our galaxy cluster sample. Besides, a number of lower-redshift (z ~ 0.1) BCGs have been sampled from the shallower Medium-deep Imaging Survey (MIS) as described in Schawinski et al. (2007). The UV photometry turned out to be very tricky because there are many background UV sources that are not easily identifiable in shallow images. The background confusion would easily cause underestimation on the UV brightness. Occasionally, small foreground objects that are invisible in the optical images contaminate the UV flux of our target galaxy as well.
From the up-to-date GALEX database, Ree et al. (2007) obtained the data for seven BCGs from DIS and five from MIS. A small fraction of the BCGs had star formation signatures (Yi et al. 2005) and hence had to be removed from our sample. Figure 3 shows the look-back time evolution of the apparent (not K-corrected) FUV - V colour for the BCGs at z < 0.2. The FUV flux fades rapidly with redshift. The colours are derived from total magnitudes to minimize aperture effect. Model lines are calibrated to the colour range (FUV - V = 5.4 - 6.4) of the giant elliptical galaxies in nearby clusters (open circles), and passively evolved and redshifted with look-back time so that they can be directly compared with the observed data of the BCGs (filled circles) in GALEX DIS (black) and MIS (grey) mode. The size of circle symbols represents the absolute total luminosity in r-band. The solid and dashed lines are from the passively evolving UV-to-optical spectra of the "metal-poor" and "metal-rich" HB models (Section 3). The regions filled with oblique lines denote the predicted colour range from these two extreme models. The dotted line indicates the apparent colour expected when the local UV upturn galaxy NGC 1399 model spectrum is redshifted without the effect of stellar evolution. The binary population models would be similar to the non-evolving model. The higher redshift data points at 0.33 and 0.6 are the HST data from Brown et al. (2000, 2003) The model fits by Ree et al. (2007) and Lee et al. (2005a) suggest that the GALEX data show a UV flux decline with lookback time at the rate (FUV - V) /t = 0.54 mag / Gyr. Although a definite answer requires more data, the current sample seems more consistent with the prediction from the single-star population models. Any population model aiming to explain the UV upturn phenomenon would be obliged to reproduce this unique data.
Figure 3. Look-back time evolution of the apparent (not K-corrected) FUV - V colour for the brightest cluster elliptical galaxies (BCGs) at z 0.2. FUV flux fades rapidly with redshift which is consistent with the prediction from the single-star population models (Section 3.3). See text for details. Excerpted from Ree et al. (2007).