The MOSDEF Survey: Implications of the Lack of Evolution in the Dust Attenuation-Mass Relation to z~2

TL;DR

This study examines the dust attenuation-stellar mass relation in star-forming galaxies from redshift 0 to 2.3, finding it remains constant despite significant evolution in dust mass and gas properties, challenging existing models of dust evolution.

Contribution

It provides evidence that the dust attenuation-stellar mass relation does not evolve from z~0 to z~2.3, highlighting the need to understand dust properties and geometry over cosmic time.

Findings

No significant evolution in dust attenuation vs. stellar mass relation.

Dust mass and surface density are higher at z~2.3 than at z~0.

Constancy of attenuation relation despite dust mass increase.

Abstract

We investigate the relationship between dust attenuation and stellar mass ($M_*$) in star-forming galaxies over cosmic time. For this analysis, we compare measurements from the MOSFIRE Deep Evolution Field (MOSDEF) survey at $z\sim2.3$ and the Sloan Digital Sky Survey (SDSS) at $z\sim0$, augmenting the latter optical dataset with both UV Galaxy Evolution Explorer (GALEX) and mid-infrared Wide-field Infrared Survey Explorer (WISE) photometry from the GALEX-SDSS-WISE Catalog. We quantify dust attenuation using both spectroscopic measurements of H$\alpha$ and H$\beta$ emission lines, and photometric measurements of the rest-UV stellar continuum. The H$\alpha$/H$\beta$ ratio is used to determine the magnitude of attenuation at the wavelength of H$\alpha$, $A_{{\rm H}\alpha}$. Rest-UV colors and spectral-energy-distribution fitting are used to estimate $A_{1600}$, the magnitude of attenuation at a rest wavelength of 1600\AA. As in previous work, we find a lack of significant evolution in the relation between dust attenuation and $M_*$ over the redshift range $z\sim0$ to $z\sim2.3$. Folding in the latest estimates of the evolution of $M_{{\rm dust}}$, $({M_{{\rm dust}}}/{M_{{\rm gas}}})$, and gas surface density at fixed $M_*$, we find that the expected $M_{{\rm dust}}$ and dust mass surface density are both significantly higher at $z\sim2.3$ than at $z\sim0$. These differences appear at odds with the lack of evolution in dust attenuation. To explain the striking constancy in attenuation vs. $M_*$, it is essential to determine the relationship between metallicity and $({M_{{\rm dust}}}/{M_{{\rm gas}}})$, the dust mass absorption coefficient, and dust geometry, and the evolution of these relations and quantities from $z\sim0$ to $z\sim2.3$.