The SCUBA-2 Cosmology Legacy Survey: the submillimetre properties of Lyman-break galaxies at z = 3-5

We present detections at 850 μm of the Lyman-break galaxy (LBG) population at z ≈ 3, 4, and 5 using data from the Submillimetre Common User Bolometer Array 2 Cosmology Legacy Survey in the United Kingdom Infrared Deep Sky Survey ‘Ultra Deep Survey’ field. We employ stacking to probe beneath the survey limit, measuring the average 850 μm flux density of LBGs at z ≈ 3, 4, and 5 with typical ultraviolet luminosities of L1700 ≈ 1029 erg s−1 Hz−1. We measure 850 μm flux densities of (0.25 ± 0.03), (0.41 ± 0.06), and (0.88 ± 0.23) mJy, respectively, finding that they contribute at most 20 per cent to the cosmic far-infrared (IR) background at 850 μm. Fitting an appropriate range of spectral energy distributions to the z ∼ 3, 4, and 5 LBG stacked 24–850 μm fluxes, we derive IR luminosities of L8-1000 μm ≈ 3.2, 5.5, and 11.0 × 1011 L⊙ [and star formation rates (SFRs) of ≈50–200 M⊙ yr−1], respectively. We find that the evolution in the IR luminosity density of LBGs is broadly consistent with model predictions for the expected contribution of luminous-to-ultraluminous IR galaxies at these epochs. We observe a positive correlation between stellar mass and IR luminosity and confirm that, for a fixed mass, the reddest LBGs (UV slope β → 0) are redder due to dust extinction, with SFR(IR)/SFR(UV) increasing by about an order of magnitude over −2 < β < 0 with SFR(IR)/SFR(UV) ∼ 20 for the reddest LBGs. Furthermore, the most massive LBGs tend to have higher obscured-to-unobscured ratios, hinting at a variation in the obscuration properties across the mass range.

For more details, the full paper can be found in Coppin, K.E.K. et al. 2015, MNRAS, 446,1293

Evolution of the bolometric IR luminosity density (solid line) as a function of redshift from the parametric backward evolution model of Béthermin et al. (2011). Also shown are the individual contributions in the models from normal galaxies (LIR 1013 L⊙). Measurements from Pascale et al. (2009) and Rodighiero et al. (2010) are overplotted, along with our results. To derive the IR luminosity density, we have multiplied the space density of LBGs at each redshift by the average LIR we have measured, which is obviously affected by the choice of luminosity integration limit; we show the effect in integrating down to M* and M* + 1. These results extend observational constraints of the models from z = 3 to 5, showing that the observations of LBGs are broadly consistent with the model predictions for the LIRG class of galaxies for our z ∼ 3 and 4 LBGs and with the ULIRG class for the z ∼ 5 LBGs.

Evolution of the bolometric IR luminosity density (solid line) as a function of redshift from the parametric backward evolution model of Béthermin et al. (2011). Also shown are the individual contributions in the models from normal galaxies (LIR < 1011 L⊙), LIRGs (1011 < LIR < 1012 L⊙), ULIRGs (1012 < LIR < 1013 L⊙), and HyLIRGs (LIR > 1013 L⊙). Measurements from Pascale et al. (2009) and Rodighiero et al. (2010) are overplotted, along with our results. To derive the IR luminosity density, we have multiplied the space density of LBGs at each redshift by the average LIR we have measured, which is obviously affected by the choice of luminosity integration limit; we show the effect in integrating down to M* and M* + 1. These results extend observational constraints of the models from z = 3 to 5, showing that the observations of LBGs are broadly consistent with the model predictions for the LIRG class of galaxies for our z ∼ 3 and 4 LBGs and with the ULIRG class for the z ∼ 5 LBGs.