A resolved map of the infrared excess in a Lyman Break Galaxy at z=3

M. P. Koprowski, K. E. K. Coppin, J. E. Geach, N. K. Hine, M. Bremer, S. C. Chapman, L. J. M. Davies, T. Hayashino, K. K. Knudsen, M. Kubo, B. D. Lehmer, Y. Matsuda, D. J. B. Smith, P. P. van der Werf, G. Violino, T. Yamada

Published 2016-08-17Version 1

We have observed the dust continuum of ten z=3.1 Lyman Break Galaxies with the Atacama Large Millimeter/Submillimeter Array at ~450 mas resolution in Band 7. We detect and resolve the 870um emission in one of the targets with an integrated flux density of S(870)=(192+/-57) uJy, and measure a stacked 3-sigma signal of S(870)=(67+/-23) uJy for the remaining nine. The total infrared luminosities estimated from full spectral energy distribution fits are L(8-1000um)=(8.4+/-2.3)x10^10 Lsun for the detection and L(8-1000um)=(2.9+/-0.9)x10^10 Lsun for the stack. With HST ACS I-band imaging we map the rest-frame UV emission on the same scale as the dust, effectively resolving the 'infrared excess' (IRX=L_FIR/L_UV) in a normal galaxy at z=3. Integrated over the galaxy we measure IRX=0.56+/-0.15, and the galaxy-averaged UV slope is beta=-1.25+/-0.03. This puts the galaxy a factor of ~10 below the IRX-beta relation for local starburst nuclei of Meurer et al. (1999). However, IRX varies by more than a factor of 3 across the galaxy, and we conclude that the complex relative morphology of the dust relative to UV emission is largely responsible for the scatter in the IRX-beta relation at high-z. A naive application of a Meurer-like dust correction based on the UV slope would dramatically over-estimate the total star formation rate, and our results support growing evidence that when integrated over the galaxy, the typical conditions in high-z star-forming galaxies are not analogous to those in the local starburst nuclei used to establish the Meurer relation.