6.2. Conditions for exploiting submm lensing by galaxies
The key advantage of observing background galaxies that are gravitationally lensed by foreground mass concentrations in the submm waveband is that the K correction (Fig. 4) acts to brighten the distant background lensed galaxy as compared with the lens. This is already very familiar from surveys of lensed radio AGN (Rusin, 2001), and is illustrated clearly in Fig. 1, in which only the central cD galaxy in the lensing cluster shows any significant submm emission.
In SCUBA cluster lens surveys, both the image separations, and the extent of the high-magnification regions are of order 1 arcmin, a scale which is well matched both to the 15-arcsec resolution of the JCMT and to the 2.5-arcmin field of view of SCUBA. The magnification ensures that a significantly greater fraction of the submm-wave background radiation intensity is thus resolved into detectable galaxies in surveys in the fields of gravitational lensing clusters than in even the deepest blank-field surveys (Blain et al., 1999a). However, for background sources lensed by galaxies rather than clusters, the relevant image separations and the extent of the high-magnification region are only of order 1 arcsec, and so cannot be resolved using any single-antenna telescope. High-resolution submm observations are required to disentangle lensed and unlensed galaxies; this capability will be provided by ALMA (Blain, 2002), while pilot studies of should be possible using the CARMA, SMA and IRAM PdBI interferometers. The most luminous lensed sources can already be resolved into multiple images using the IRAM mm-wave interferometer (Alloin et al., 1997).
The only caveat for exploiting galaxy-scale lensing is that the source size must be small as compared with the area of sky behind the lens that is strongly magnified. The intense far-IR and submm emission from low-redshift ULIRGs is typically very compact (several hundred pc across; Downes and Solomon, 1998), and would easily meet this condition; however, there are indications that the dust emission from at least some luminous high-redshift submm galaxies could extend over scales greater than 10-kpc (Papadopoulos et al., 2001; Chapman et al., 2001a; Lutz et al., 2001; Isaak et al., 2002; Ivison et al., 2001). The whole area of sky covered by these galaxies would not then be lensed efficiently by an intervening galaxy, although bright knots of emission within them could still be magnified by large factors. This concern about lensing efficiency and the angular size of distant submm galaxies does not apply to lensing by much larger clusters of galaxies, which will always be effective.