Evolution of dust attenuation in star-forming galaxies with UV slope, stellar mass, and redshift out to $z \sim 5$

TL;DR

The paper tackles how dust attenuation in star-forming galaxies, quantified by infrared excess IRX, depends on UV slope $β$, stellar mass $M_*$, and redshift up to $z\sim5$. It uses a $K$-band–selected sample of ~10^5 galaxies from the UDS and COSMOS fields, employing FIR stacking to recover $L_{IR}$ and SED-based $β$ estimates to derive IRX relations, including a mass dependence via the reddening-law slope $dA_{1600}/dβ$. The authors find that IRX–$β$ aligns with a Calzetti-like attenuation for $β\gtrsim -1$, while at bluer $β$ the IRX grows with redshift due to mass-completeness effects; IRX–$β$ in bins of $M_*$ reveals a quadratic dependence of $dA_{1600}/dβ$ on $ ext{log}(M_*/M_⊙)$, indicating grayer attenuation in more massive systems. Expressing IRX as a function of $M_*$ shows a tight increasing trend with a high-mass turnover at $z\lesssim 2-3$, consistent with reduced cold-gas accretion and dust growth in massive galaxies, and the redshift evolution of the $β$–$M_*$ relation helps explain this turnover. Together, these results provide a practical, mass- and redshift-dependent framework to correct UV/optical galaxy samples for dust obscuration and reconcile previous stacking studies across a broad cosmic span.

Abstract

Aims. We derive a dependence of the IRX on UV slope $β$, stellar mass $M_\ast$, and redshift out to $z \simeq 5$, and establish consistent functional relations that can be used for correcting the UV/optical-selected galaxy samples for the effects of dust absorption. Methods. This work is based on a $K$-band selected sample of $\sim 10^5$ star-forming galaxies detected in the UDS and COSMOS fields. Quiescent sources and known starbursts are removed, and the IR luminosities are established through stacking in FIR {\it Herschel} and JCMT maps. UV slopes are found from SED fits and stacked IRX values are derived by taking the median of individual IRX measurements in bins of $β$, $M_\ast$ and redshift. Results. While our best-fit IRX-$β$ relation is consistent with a Calzetti-like attenuation curve at $β\gtrsim -1$, at bluer values the IRX seems to increase with redshift due to different mass-completeness limits imposed. When deriving the IRX-$β$ relation in stellar-mass bins, a systematic trend is found, where the effective slope of the attenuation law becomes progressively shallower with increasing mass. We incorporate this into the IRX-$β$ relation through the slope of the underlying reddening law, $dA_{1600}/dβ$, being a quadratic function of $\log(M_\ast/{\rm M_\odot})$. Expressing IRX as a function of the stellar mass we find a tight correlation, with IRX rising monotonically with mass but exhibiting a clear high-mass turnover at $z\lesssim 2-3$, consistent with suppressed cold-gas accretion and dust growth in massive systems.

Figures (11)

• Figure 1: IRX-$\beta$ relationship for $2.0<z\leq 5.0$ sample studied in this work. The color points with error bars show the stacked values of the IRX in bins of $\beta$, summarized in Table \ref{['tab:irxb']}. The black solid line represents the best-fit functional form (Equation \ref{['eq:sum1']}), with $dA_{1600}/d\beta=1.97$ being consistent with the attenuation curve of Calzetti_2000. The scatter at $\beta\lesssim-1$ is caused by different stellar mass completeness limits imposed at each redshift bin (see Section \ref{['sec:irxb']} for details). For reference, the SMC curve is plotted in dashed line.
• Figure 2: Relationship between the UV slope and the stellar mass for the sample presented in Figure \ref{['fig:irxb']}. Green background image shows a 2D histogram of the $\beta$-$M_\ast$ distribution, with 1, 10, 100 and 1000 sources contours displayed with black curves. It can be seen that due to different mass-completeness limits imposed at each redshift bin (Table \ref{['tab:irxb']}), the median stellar masses at bluest $\beta$ bins increase with redshift, which in turn drives the scatter in Figure \ref{['fig:irxb']} (see Section \ref{['sec:irxb']} for details).
• Figure 3: IRX-$\beta$ relationship in bins of stellar mass. The black circles show the stacked data (Table \ref{['tab:irxbm']}), with a clear correlation between the stellar mass and the slope of the reddening law, $dA_{1600}/d\beta$. Adopting Equation \ref{['eq:sum1']}, with $dA_{1600}/d\beta$ set as free parameter, best-fit values for the reddening slope were found (last column in Table \ref{['tab:irxbm']}), with the corresponding correlation with the stellar mass derived in Section \ref{['sec:mirx']} and plotted in Figure \ref{['fig:dadbm']}. For reference, this work’s best-fit relationship (consistent with the Calzetti curve), together with the SMC curve are plotted in solid and dashed lines, respectively.
• Figure 4: Relationship between the dust reddening slope, $dA_{1600}/d\beta$, and the stellar mass for the sample plotted in Figure \ref{['fig:irxbm']}, where a clear correlation can be seen, with more massive systems being characterized by flatter attenuation curves. The functional form of the relation is derived and discussed in Section \ref{['sec:mirx']} and summarized in Equation \ref{['eq:sum1']}.
• Figure 5: IRX-$\beta$ relationship derived for the $2.0<z \leq 5.0$ sample of this work (gray circles; Table \ref{['tab:irxb1']}). Our median data lying slightly below the best-fit curve at blue UV slopes is driven by variations in the corresponding median stellar masses. We compare our data with the recent literature results of Heinis_2013Reddy_2018McLure_2018Koprowski_2018Alvarez_2019Fudamoto_2020 and Bouwens_2020, with apparent inconsistencies discussed in Section \ref{['sec:irxbcomp']}. For reference, this work's best-fit relationship and the SMC-like curve are plotted in solid and dashed lines, respectively.