High-redshift quasars lensed by spiral galaxies

Given its extraordinary spatial resolution and sensi- tivity, the projected Next Generation Space Telescope (NGST) is likely to detect a large number of high-redshift QSOs lensed by spiral galaxies. Using realistic models for the QSO and spiral populations, we calculate the expected number density of de- tectable QSOs multiply imaged by spirals, and investigate the influence of various evolution effects on that number density. It is shown that NGST will probably find of order ten lensed QSOs per square degree at 26th magnitude in the V and L bands, and that various observable quantities like the total number density of lensed QSOs in these two bands, the ratio between the num- ber densities of lensed QSOs in the V and L bands, the fraction of QSOs with more than two images and so forth can be used to constrain the evolution of the QSOs, the spirals, the dust in spirals, and the masses of spiral disks. reach a sensitivity of order one nano-Jansky in the visible and near-infrared wave bands. Going to such observational limits, QSOs lensed by spiral galaxies should be discovered in substan- tial numbers. Many of the multiple-image systems produced by spiral galaxies will straddle near-edge-on galactic disks. The statistics of image separations and time delays will therefore allow to determine properties of spiral disks at moderate and high redshifts. The wavelength regime between the visible and the near infrared is particularly interesting for such investiga- tions, since the importance of dust changes dramatically across this wavelength range. While multiple QSO images, in particu- lar near galactic disks, are substantially obscured by dust in the V band, they are almost unaffected in the L band. Comparing observations taken in these two wavebands, it will therefore be possible to quantify the amount and the distribution of dust in distant spiral disks. The very faint flux limits envisaged will bring high-redshift QSOs into view. The QSO population at high redshifts is un- certain. Observations show a quick drop of the QSO number density beyond redshifts of 4 5 (Pei 1995). This may reflect rapid QSO evolution at these redshifts, but also incompleteness of the QSO samples observed. We therefore investigate two al- ternatives, assuming either that the observed QSO population is complete, or that it is to be extrapolated towards high redshifts. It is also important for this study to reliably estimate QSO K-corrections in the visible and the near infrared. For this pur- pose, we model the dominant features of QSO spectra together with the Lyman- forest, and determine K-corrections by con- volving the redshifted model spectrum with the appropriate filter functions.