We present our survey for optically faint variable objects using multi-epoch (8-10 epochs over 2-4 years) $i$-band imaging data obtained with Subaru Suprime-Cam over 0.918 deg$^2$ in the Subaru/XMM-Newton Deep Field (SXDF). We found 1040 optically va
riable objects by image subtraction for all the combinations of images at different epochs. This is the first statistical sample of variable objects at depths achieved with 8-10m class telescopes or HST. The detection limit for variable components is $i_{rm{vari}}sim25.5$ mag. These variable objects were classified into variable stars, supernovae (SNe), and active galactic nuclei (AGN), based on the optical morphologies, magnitudes, colors, and optical-mid-infrared colors of the host objects, spatial offsets of variable components from the host objects, and light curves. Detection completeness was examined by simulating light curves for periodic and irregular variability. We detected optical variability for $36pm2%$ ($51pm3%$ for a bright sample with $i<24.4$ mag) of X-ray sources in the field. Number densities of variable obejcts as functions of time intervals $Delta{t}$ and variable component magnitudes $i_{rm{vari}}$ are obtained. Number densities of variable stars, SNe, and AGN are 120, 489, and 579 objects deg$^{-2}$, respectively. Bimodal distributions of variable stars in the color-magnitude diagrams indicate that the variable star sample consists of bright ($Vsim22$ mag) blue variable stars of the halo population and faint ($Vsim23.5$ mag) red variable stars of the disk population. There are a few candidates of RR Lyrae providing a possible number density of $sim10^{-2}$ kpc$^{-3}$ at a distance of $>150$ kpc from the Galactic center.
We present the properties of active galactic nuclei (AGN) selected by optical variability in the Subaru/XMM-Newton Deep Field (SXDF). Based on the locations of variable components and light curves, 211 optically variable AGN were reliably selected. W
e made three AGN samples; X-ray detected optically non-variable AGN (XA), X-ray detected optically variable AGN (XVA), and X-ray undetected optically variable AGN (VA). In the VA sample, we found a bimodal distribution of the ratio between the variable component flux and the host flux. One of these two components in the distribution, a class of AGN with a faint variable component $i_{rm{vari}}sim25$ mag in bright host galaxies $isim21$ mag, is not seen in the XVA sample. These AGN are expected to have low Eddington ratios if we naively consider a correlation between bulge luminosity and black hole mass. These galaxies have photometric redshifts $z_{rm{photo}}sim0.5$ and we infer that they are low-luminosity AGN with radiatively inefficient accretion flows (RIAFs). The properties of the XVA and VA objects and the differences from those of the XA objects can be explained within the unified scheme for AGN. Optical variability selection for AGN is an independent method and could provide a complementary AGN sample which even deep X-ray surveys have not found.
Large samples of high-redshift supernovae (SNe) are potentially powerful probes of cosmic star formation, metal enrichment, and SN physics. We present initial results from a new deep SN survey, based on re-imaging in the R, i, z bands, of the 0.25 de
g2 Subaru Deep Field (SDF), with the 8.2-m Subaru telescope and Suprime-Cam. In a single new epoch consisting of two nights of observations, we have discovered 33 candidate SNe, down to a z-band magnitude of 26.3 (AB). We have measured the photometric redshifts of the SN host galaxies, obtained Keck spectroscopic redshifts for 17 of the host galaxies, and classified the SNe using the Bayesian photometric algorithm of Poznanski et al. (2007) that relies on template matching. After correcting for biases in the classification, 55% of our sample consists of Type Ia supernovae and 45% of core-collapse SNe. The redshift distribution of the SNe Ia reaches z ~ 1.6, with a median of z ~ 1.2. The core-collapse SNe reach z ~ 1.0, with a median of z ~ 0.5. Our SN sample is comparable to the Hubble Space Telescope/GOODS sample both in size and redshift range. The redshift distributions of the SNe in the SDF and in GOODS are consistent, but there is a trend (which requires confirmation using a larger sample) for more high-z SNe Ia in the SDF. This trend is also apparent when comparing the SN Ia rates we derive to those based on GOODS data. Our results suggest a fairly constant rate at high redshift that could be tracking the star-formation rate. Additional epochs on this field, already being obtained, will enlarge our SN sample to the hundreds, and determine whether or not there is a decline in the SN Ia rate at z >~ 1.
