Search

We investigate the clustering properties of 2600 Lyman break galaxies (LBGs) at z = 3.5-5.2 in two large blank fields, the Subaru Deep Field and the Subaru/XMM-Newton Deep Field (600 arcmin2 each). The angular correlation functions of these LBGs show a clear clustering at both z [?] 4 and 5. The correlation lengths are r0 = 4.1img1.gif and 5.9img2.gifhimg3.gif Mpc (r0 = 5.1img4.gif and 5.9img2.gifhimg3.gif Mpc) for all the detected LBGs (for L [?] L* LBGs) at z [?] 4 and 5, respectively. These correlation lengths correspond to galaxy-dark matter biases of bg = 2.9img5.gif and 4.6img6.gif (bg = 3.5img7.gif and 4.6img6.gif) for all the detected LBGs (for L [?] L* LBGs) at z [?] 4 and 5, respectively. These results, combined with estimates for z [?] 3 LBGs in the literature, show that the correlation length of L [?] L* LBGs is almost constant, ~5 himg3.gif Mpc, over z [?] 3-5, while the bias monotonically increases with redshift at z [?] 3. We also find that for LBGs at z [?] 4 the clustering amplitude increases with UV-continuum luminosity and with the amount of dust extinction. We estimate the mass of dark halos hosting various kinds of high-z galaxies, including LBGs, with the analytic model given by Sheth & Tormen. We find that the typical mass of dark halos hosting L [?] L* LBGs is about 1 x 1012 himg8.gif M over z [?] 3-5, which is comparable to that of the Milky Way. A single dark halo with ~1012 himg8.gif M is found to host 0.1-0.3 LBGs on average but about four K-band-selected galaxies.

We investigate the photometric properties of Lyman break galaxies (LBGs) at z = 3.5-5.2 based on large samples of 2600 LBGs detected in deep (i' [?] 27) and wide-field (1200 arcmin2) images taken in the Subaru Deep Field (SDF) and the Subaru/XMM-Newton Deep Field (SXDF) using broadband B, V, R, i', and z' filters. The selection criteria for the LBG samples are examined with 85 spectroscopically identified objects, and the completeness and contamination of the samples are estimated from Monte Carlo simulations based on a photometric-redshift catalog of the Hubble Deep Field-North. We find that these LBG samples are nearly rest-frame UV magnitude-limited samples, missing systematically only 10% of red high-z galaxies (in number), which are a dusty population with E(B - V) [?] 0.4. We calculate luminosity functions (LFs) of the LBGs with the estimated completeness and contamination and find (1) that the number density of bright galaxies (M1700 < -22 ; corresponding to SFR [?] 100 himg1.gif M yr-1 with extinction correction) decreases significantly from z = 4 to 5 and (2) that the faint-end slope of the LFs of LBGs may become steeper toward higher redshifts. We estimate the dust extinction of z [?] 4 LBGs with M < M* ([?]-21) from UV-continuum slopes and obtain E(B - V) = 0.15 +- 0.03 as the mean value. The dust extinction remains constant with apparent luminosity but increases with intrinsic (i.e., extinction-corrected) luminosity. We find no evolution in dust extinction between LBGs at z = 3 and 4. We investigate the evolution of UV-luminosity density by integrating the LFs of LBGs and find that the UV-luminosity density at 1700 A, rUV, does not significantly change from z = 3 to 5, i.e., rUV(z = 4)/rUV(z = 3) = 1.0 +- 0.2 and rUV(z = 5)/rUV(z = 3) = 0.8 +- 0.4; thus, the cosmic star formation rate (SFR) density (with correction for dust extinction) remains constant within the error bars, or possibly has a slight decline, from z = 3 to 5. We estimate the stellar mass density from the cosmic SFR thus obtained and find that this stellar mass density is consistent with those derived directly from the stellar mass function at z = 0-1 but exceeds those at z ~ 3 by a factor of 3. We find that the ratio of the UV-luminosity density of Lya emitters (LAEs) to that of LBGs is [?]60% at z [?] 5, and thus about half of star formation probably occurs in LAEs at z [?] 5. We obtain a constraint on the escape fraction of UV ionizing photons (i.e., UV continuum in 900 A) produced by LBGs, fesc [?] 0.13. This implies that the escape fraction of LBGs may be larger than that of star-forming galaxies at z = 0.

We present measurements of the cosmic shear correlation in the shapes of galaxies in the Suprime-Cam 2.1 deg2 RC-band imaging data. As an estimator of the shear correlation originating from the gravitational lensing, we adopt the aperture mass variance, which most naturally decomposes the correlation signal into E and B (non-gravitational lensing) modes. We detect a nonzero E mode variance on scales between thap = 2' and 40'. We also detect a small but nonzero B-mode variance on scales larger than thap > 5'. We compare the measured E-mode variance to the model predictions in CDM cosmologies using maximum likelihood analysis. A four-dimensional space is explored, which examines s8, Om, G (the shape parameter of the CDM power spectrum), andimg1.gif s (mean redshift of galaxies). We include three possible sources of error: statistical noise, the cosmic variance estimated using numerical experiments, and a residual systematic effect estimated from the B-mode variance. We derive joint constraints on two parameters by marginalizing over the two remaining parameters. We obtain an upper limit of G < 0.5 forimg1.gif s > 0.9 (68% confidence). For a prior G [?] [0.1, 0.4] andimg1.gif s [?] [0.6, 1.4], we find s8 = (0.50img2.gif)Oimg3.gif for Om + OL = 1 and s8 = (0.51img4.gif)Oimg5.gif for OL = 0 (95% confidence). If we take the currently popular LCDM model (Om = 0.3, Ol = 0.7, G = 0.21), we obtain a one-dimensional confidence interval on s8 for the 95.4% level, 0.62 < s8 < 1.32 forimg1.gif s [?] [0.6, 1.4]. Information on the redshift distribution of galaxies is key to obtaining a correct cosmological constraint. An independent constraint on G from other observations is useful to tighten the constraint.

We report on early results from very deep and wide-field narrowband imaging on a 543 arcmin2 area of the Subaru Deep Field. We find 87 Lya emitters (LAEs) at z = 4.86 +- 0.03 that are photometrically selected by a combination of two broad bands (R and i') and one narrow band (NB711; lc = 7126 A, Dl = 73 A). We derive the luminosity functions (LFs) of the LAEs at Lya luminosity and at UV-continuum (rest frame 1700 A) luminosity. The LFs show little evolution between z = 3.4 and 4.86 either in Lya or UV-continuum emission. The amplitude of the LAE LF tends to decline at the bright magnitudes more rapidly than that of the LBG LF at similar redshifts. We calculate the angular correlation function of our LAEs up to ~15' separations. The angular correlation o(th) is found to increase with decreasing angular separations, showing a clear signal of clustering. It is also found that the distribution of LAEs shows a large-density gradient with a scale of [?]15', which would indicate the existence of a large-scale structure of LAEs on [?]20 h-1 Mpc scales. We fit the observed correlation function by Aoth-0.8 to find Ao = 29 arcsec0.8. The estimated correlation length is r0 = 3.5img1.gifh-1 Mpc in comoving units (Om = 0.3 and OL = 0.7), which is slightly larger than the value for z ~ 4 LBGs with i' < 26. We calculate the angular correlation function for two subsamples of the 87 LAEs divided by Lya luminosity, UV-continuum luminosity, and Lya equivalent width (EW). The Lya-bright subsample shows a larger correlation amplitude than the Lya-faint subsample, while no significant difference is found for the subsamples divided by UV-continuum luminosity or EW. This may indicate that galaxies with bright Lya emission are possibly biased against the underlying dark matter halos more strongly than those with bright UV continuum.