The VLA-COSMOS Perspective on the IR-Radio Relation. I. New Constraints on Selection Biases and the Non-Evolution of the IR/Radio Properties of Star Forming and AGN Galaxies at Intermediate and High Redshift

TL;DR

This study uses extensive multi-wavelength data from the COSMOS field to examine the IR-radio relation in galaxies, finding it remains consistent up to high redshifts and addressing previous discrepancies with robust statistical methods.

Contribution

It provides new constraints on the IR-radio relation's evolution, demonstrating its stability up to z~1.4 and possibly beyond, using a large, bias-reduced sample and survival analysis techniques.

Findings

IR-radio relation remains unchanged up to z~1.4

Relation likely holds at 2.5<z<5

Radio-quiet AGN follow the IR-radio relation similarly to star-forming galaxies

Abstract

VLA 1.4 GHz (rms noise ~0.012 mJy) and MIPS 24 and 70 micron (rms noise ~0.02 and ~1.7 mJy, respectively) observations covering the 2 square degree COSMOS field are combined with an extensive multi-wavelength data set to study the evolution of the IR-radio relation at intermediate and high redshift. With ~4500 sources -- of which ~30% have spectroscopic redshifts -- the current sample is significantly larger than previous ones used for the same purpose. Both monochromatic IR/radio flux ratios (q24 & q70), as well as the ratio of the total IR and the 1.4 GHz luminosity (qTIR) are used as indicators for the IR/radio properties of star forming galaxies and AGN. Using a sample jointly selected at IR and radio wavelengths in order to reduce selection biases, we provide firm support for previous findings that the IR-radio relation remains unchanged out to at least z~1.4. Moreover, based on data from ~150 objects we also find that the local relation likely still holds at 2.5<z<5. At redshift z<1.4 we observe that radio-quiet AGN populate the locus of the IR-radio relation in similar numbers as star forming sources. In our analysis we employ the methods of survival analysis in order to ensure a statistically sound treatment of flux limits arising from non-detections. We determine the observed shift in average IR/radio properties of IR- and radio- selected populations and show that it can reconcile apparently discrepant measurements presented in the literature. Finally, we also investigate variations of the IR/radio ratio with IR and radio luminosity and find that it hardly varies with IR luminosity but is a decreasing function of radio luminosity.