The evolution of the size-mass relation at $z$=1-3 derived from the complete Hubble Frontier Fields data set

TL;DR

This study measures the size-mass relation of galaxies at redshifts 1-3 using gravitational lensing from Hubble Frontier Fields, revealing size evolution and demonstrating the effectiveness of lensing for high-redshift galaxy studies.

Contribution

It provides new empirical size-mass relations at high redshift using lensing data and assesses lens model uncertainties, advancing methods for studying galaxy evolution.

Findings

Galaxy sizes evolve as R_eff ∝ (1+z)^{-1.05±0.37}

Intrinsic scatter increases from <0.1 dex at z<1.5 to ~0.3 dex at higher z

Lens model choice introduces ~3.7% median uncertainty and ~25% scatter for individual galaxies.

Abstract

We measure the size-mass relation and its evolution between redshifts 1$<z<$3, using galaxies lensed by six foreground Hubble Frontier Fields clusters. The power afforded by strong gravitation lensing allows us to observe galaxies with higher angular resolution beyond current facilities. We select a stellar mass limited sample and divide them into star-forming or quiescent classes based on their rest-frame UVJ colors from the ASTRODEEP catalogs. Source reconstruction is carried out with the recently-released $lenstruction$ software, which is built on the multi-purpose gravitational lensing software $lenstronomy$. We derive the empirical relation between size and mass for the late-type galaxies with $M_{*}>3\times10^{9}M_{\odot}$ at 1$<z<$2.5 and $M_* >5 \times 10^{9} M_{\odot}$ at 2.5$<z<$3, and at a fixed stellar mass, we find galaxy sizes evolve as $R_{eff}\propto (1+z)^{-1.05\pm0.37}$. The intrinsic scatter is $<0.1$ dex at $z<1.5$ but increases to $\sim0.3$ dex at higher redshift. The results are in good agreement with those obtained in blank fields. We evaluate the uncertainties associated with the choice of lens model by comparing size measurements using five different and publicly available models, finding the choice of lens model leads to a 3.7 % uncertainty of the median value, and $\sim 25$ % scatter for individual galaxies. Our work demonstrates the use of strong lensing magnification to boost resolution does not introduce significant uncertainties in this kind of work, and paves the way for wholesale applications of the sophisticated lens reconstruction technique to higher redshifts and larger samples.