The Dust Attenuation Law in Distant Galaxies: Evidence for Variation with Spectral Type

TL;DR

This study investigates how dust attenuation laws vary among distant galaxies, revealing that the dust properties depend on spectral type, star formation activity, and geometry, challenging the assumption of a universal dust law.

Contribution

It provides the first comprehensive analysis of dust attenuation curve variations across different spectral types of galaxies at 0.5<z<2.0 using composite SEDs.

Findings

Attenuation law varies significantly with spectral type.

Steeper laws correlate with stronger UV bumps.

Active galaxies have shallower dust curves and weaker bumps.

Abstract

This letter utilizes composite spectral energy distributions (SEDs) constructed from NEWFIRM Medium-Band Survey photometry to constrain the dust attenuation curve in 0.5<z<2.0 galaxies. Based on similarities between the full SED shapes (0.3-8 micron), we have divided galaxies in 32 different spectral classes and stacked their photometry. As each class contains galaxies over a range in redshift, the resulting rest-frame SEDs are well-sampled in wavelength and show various spectral features including Halpha and the UV dust bump at 2175 Angstrom. We fit all composite SEDs with flexible stellar population synthesis models, while exploring attenuation curves with varying slopes and UV bump strengths. The Milky Way and Calzetti law provide poor fits at UV wavelengths for nearly all SEDs. Consistent with previous studies, we find that the best-fit attenuation law varies with spectral type. There is a strong correlation between the best-fit dust slope and UV bump strength, with steeper laws having stronger bumps. Moreover, the attenuation curve correlates with specific star formation rate (SFR), with more active galaxies having shallower dust curves and weaker bumps. There is also a weak correlation with inclination. The observed trends can be explained by differences in the dust-to-star geometry, a varying grain size distribution, or a combination of both. Our results have several implications for galaxy evolution studies. First, the assumption of a universal dust model leads to biases in derived galaxy properties. Second, the presence of a dust bump may result in underestimated values for the UV slope, used to correct SFRs of distant galaxies.