Impact of future submillimeter and millimeter large facilities on the studies of galaxy formation and evolution

We investigate what we can learn about galaxy formation and evolution from the data which will be obtained by the forthcoming large submillimeter/mill imeter facilities, mainly by the Atacama Submillimeter Telescope Experiment (ASTE) and the Atacama Large Millimeter Array/Large Millimeter and Submill imeter Array (ALMA/LMSA). We first calculate the source counts from 350 mum to 3 mm using the empirical infrared galaxy number count model of Takeuchi et al. Based on the number counts, we evaluate the source confusion and de termine the confusion limit at various wave bands as a function of the char acteristic beam size. At submillimeter wavelengths, source confusion with t he 10-15 m class facilities becomes severe at the 0.1-1 mJy level, and astr ometry and flux measurements will be difficult. However, we show that a ver y large area survey of submillimeter sources brighter than 10-50 mJy can pr ovide a unique constraint on infrared galaxy evolution at z = 1-2, and such a survey is suitable for the ASTE. In addition, such a survey enables us t o study the clustering properties of submillimeter sources, which are still highly unknown. We also find that the 5 sigma confusion limit of LMSA is f ainter than 1 mu Jy, which enables us to study the contribution of sources at extremely large redshift. When we discuss such a deep flux limit, the dy namic range of a detector should be taken into account, since extremely bri ght sources make it impossible to detect the faintest sources near the dete ction limit. We evaluate the probability that sources that are 10(3) times brighter than the 5 sigma detection limit of LMSA and ALMA exist in the fie ld of view. We find that the probability is less than or similar to3 x 10(- 4), and therefore we do not have to worry about the dynamic range. The sour ce counts at such faint flux levels give important information about the ep och of galaxy formation. We then show that multiband photometry from the in frared (by ASTRO-F) to the millimeter can be used as a redshift estimator. We examined the performance of this method by Monte Carlo simulations and f ound that it works successfully if we have reasonable measurement accuracy. In addition, we compare the observed 1.4, 5, and 8 GHz source counts with our model counts to examine the contribution of star-forming galaxies to fa int radio galaxies. We find that the faintest radio number counts (similar to1 mu Jy) are dominated by actively star-forming galaxies which lie at int ermediate redshift z similar to 1-2.