Yung Ying, Chen (ASIAA)Hiroyuki Hirashita (ASIAA)
Dust grains emit infrared (IR) radiation, which is a vital tracer of star formation activities in galaxies. Dust temperatures are the key to estimate the total IR flux. Some studies suggest that dust temperatures in normal high-redshift distant galaxies, Lyman-Break galaxies (LBGs), are systematically higher than in the local galaxies, but the possible bias in the dust temperatures for the detected objects, including the dependence on the chosen bands, is yet to be understood. To understand the dust temperature bias for high-redshift (z≿5) galaxies at various (sub)millimeter wavelengths, we perform a simple simulation based on random realizations of LBGs with various stellar masses, dust temperatures, and dust-to-stellar mass ratios. We find that, if we select galaxies by the total IR dust luminosity IR, the 850 μm and 1200 μm samples are biased to low dust temperatures (≾45 K at LIR=1011 L☉) but that the 450 μm samples are less biased. However, if we select galaxies by stellar mass, there is no significant bias for the dust temperature, and the detection in the stellar-mass-selected sample is predominantly driven by the dust abundance (dust-to-stellar mass ratio). Therefore, if we select LBGs by stellar mass, the tendency of rising temperatures at high redshift reflects the true evolution. For an IR-selected sample, we need to determine the dust temperature by using two bands or constructing a 450 μm-selected sample to avoid the dust temperature.