Exploring the Correlation between $\rm{H}\alpha$-to-UV Ratio and Burstiness for Typical Star-forming Galaxies at $z\sim2$

TL;DR

This study investigates the relationship between the Hα-to-UV luminosity ratio and star formation burstiness in z~2 galaxies, finding no strong evidence linking high ratios to recent starburst activity.

Contribution

It provides a detailed statistical analysis of resolved star formation and stellar age distributions, validating the Hα-to-UV ratio as a burstiness proxy at high redshift.

Findings

No significant link between Hα-to-UV ratio and bursty star formation.

High ratio galaxies are slightly younger but not conclusively.

Spectral features of massive stars show no variation between subsets.

Abstract

The $\rm{H}\alpha$-to-UV luminosity ratio ($L(\rm H\alpha)/L(\rm UV)$) is often used to probe SFHs of star-forming galaxies and it is important to validate it against other proxies for burstiness. To address this issue, we present a statistical analysis of the resolved distribution of $\Sigma_{\rm{SFR}}$ as well as stellar age and their correlations with the globally measured $L(\rm H\alpha)/L(\rm UV)$ for a sample of 310 star-forming galaxies in two redshift bins of $1.37 < z < 1.70$ and $ 2.09 < z < 2.61$ observed by the MOSDEF survey. We use the multi-waveband CANDELS/3D-HST imaging of MOSDEF galaxies to construct $\Sigma_{\rm{SFR}}$ and stellar age maps. We analyze the composite rest-frame far-UV spectra of a subsample of MOSDEF targets obtained by the Keck/LRIS, which includes 124 star-forming galaxies (MOSDEF-LRIS) at redshifts $1.4 < z < 2.6$, to examine the average stellar population properties, and the strength of age-sensitive FUV spectral features in bins of $L(\rm H\alpha)/L(\rm UV)$. Our results show no significant evidence that individual galaxies with higher $L(\rm H\alpha)/L(\rm UV)$ are undergoing a burst of star formation based on the resolved distribution of $\Sigma_{\rm{SFR}}$ of individual star-forming galaxies. We segregate the sample into subsets with low and high $L(\rm H\alpha)/L(\rm UV)$. The high-$L(\rm H\alpha)/L(\rm UV)$ subset exhibits, on average, an age of $\log[\rm{Age/yr}]$ = 8.0, compared to $\log[\rm{Age/yr}]$ = 8.4 for the low-$L(\rm H\alpha)/L(\rm UV)$ galaxies, though the difference in age is significant at only the $2\sigma$ level. Furthermore, we find no variation in the strengths of Siiv$\lambda\lambda1393, 1402$ and Civ$\lambda\lambda1548, 1550$ P-Cygni features from massive stars between the two subsamples.