2012ApJ...757..120G

This paper studies the properties of kiloparsec-scale clumps in star-forming galaxies at z ∼ 2 through multi-wavelength broadband photometry. A sample of 40 clumps is identified from Hubble Space Telescope (HST)/Advanced Camera for Surveys (ACS) z-band images through auto-detection and visual inspection from 10 galaxies with 1.5 < z < 2.5 in the Hubble Ultra Deep Field, where deep and high-resolution HST/WFC3 and ACS images enable us to resolve structures of z ∼ 2 galaxies down to the kiloparsec scale in the rest-frame UV and optical bands and to detect clumps toward the faint end. The physical properties of clumps are measured through fitting spatially resolved seven-band (BVizYJH) spectral energy distribution to models. On average, the clumps are blue and have similar median rest-frame UV-optical color as the diffuse components of their host galaxies, but the clumps have large scatter in their colors. Although the star formation rate (SFR)-stellar mass relation of galaxies is dominated by the diffuse components, clumps emerge as regions with enhanced specific star formation rates, contributing individually ∼10% and together ∼50% of the SFR of the host galaxies. However, the contributions of clumps to the rest-frame UV/optical luminosity and stellar mass are smaller, typically a few percent individually and ∼20% together. On average, clumps are younger by 0.2 dex and denser by a factor of eight than diffuse components. Clump properties have obvious radial variations in the sense that central clumps are redder, older, more extincted, denser, and less active on forming stars than outskirt clumps. Our results are broadly consistent with a widely held view that clumps are formed through gravitational instability in gas-rich turbulent disks and would eventually migrate toward galactic centers and coalesce into bulges. Roughly 40% of the galaxies in our sample contain a massive clump that could be identified as a proto-bulge, which seems qualitatively consistent with such a bulge-formation scenario.