Far infrared constraints on the contamination by dust obscured galaxies of high-z dropout searches

TL;DR

This study demonstrates how far infrared observations can distinguish low-redshift dusty galaxies from true high-redshift candidates, reducing contamination in high-z galaxy searches.

Contribution

It introduces a method using FIR SED analysis to identify and exclude dust-obscured low-z interlopers in high-redshift galaxy surveys.

Findings

FIR data suggests the candidates are at z~2, not >7.

Standard templates fail to fit the IR part of the SEDs.

Sources resemble dust-obscured galaxies at z~2.

Abstract

The spectral energy distributions (SED) of dusty galaxies at intermediate redshift may look similar to very high redshift galaxies in the optical/near infrared (NIR) domain. This can lead to the contamination of high redshift galaxy searches based on broad band optical/NIR photometry by lower redshift dusty galaxies as both kind of galaxies cannot be distinguished. The contamination rate could be as high as 50%. {This work shows how the far infrared (FIR) domain can help to recognize likely low-z interlopers in an optical/NIR search for high-z galaxies.} We analyse the FIR SEDs of two galaxies proposed as very high redshift ($z>7$) dropout candidates based on deep Hawk-I/VLT observations. The FIR SEDs are sampled with PACS/Herschel at 100 and 160\,$\mu$m, with SPIRE/Herschel at 250, 350 and 500\,$\mu$m and with LABOCA/APEX at 870\,$\mu$m. We find that redshifts $>7$ would imply extreme FIR SEDs (with dust temperatures $>100$\,K and FIR luminosities $>10^{13}$\,$L_{\odot}$). At z$\sim$2, instead, the SEDs of both sources would be compatible with that of typical ULIRGs/SMGs. Considering all the data available for these sources from visible to FIR we re-estimate the redshifts and we find $z\sim$1.6--2.5. Due to the strong spectral breaks observed in these galaxies, standard templates from the literature fail to reproduce the visible-near IR part of the SEDs even when additional extinction is included. These sources resemble strongly dust obscured galaxies selected in Spitzer observations with extreme visible-to-FIR colors, and the galaxy GN10 at $z=4$. Galaxies with similar SEDs could contaminate other high redshift surveys.