Table of Contents
Fetching ...

Ly$α$ profile, dust, and prediction of Ly$α$ escape fraction in Green Pea Galaxies

Huan Yang, Sangeeta Malhotra, Max Gronke, James E. Rhoads, Claus Leitherer, Aida Wofford, Tianxing Jiang, Mark Dijkstra, V. Tilvi, Junxian Wang

TL;DR

This study addresses how Ly$\alpha$ photons escape star-forming galaxies and how to predict the Ly$\alpha$ escape fraction $f^{Ly\alpha}_{esc}$ to disentangle ISM/CGM effects from IGM attenuation in reionization contexts. Utilizing a large, representative sample of 43 Green Pea galaxies with HST/COS Ly$\alpha$ spectra and rich SDSS optical data, the authors measure Ly$\alpha$ flux, equivalent width, and line profiles, and fit Ly$\alpha$ profiles with an HI shell radiative-transfer model. They reveal strong anti-correlations between $f^{Ly\alpha}_{esc}$ and HI-related kinematic quantities and HI column density $N_{HI}$, and show that lower dust, metallicity, and stellar mass favor higher Ly$\alpha$ escape, with [OIII]/[OII] acting as a proxy for LyC leakage potential. Importantly, they derive an empirical linear relation linking $f^{Ly\alpha}_{esc}$ to dust extinction $E(B-V)$ and the Ly$\alpha$ red peak velocity, with a scatter of about 0.3 dex, enabling estimates of IGM transmission along lines of sight to high-$z$ LAEs observed by JWST. These results support Green Peas as local analogs of high-$z$ LAEs and provide a practical tool to isolate IGM effects in reionization-era Ly$\alpha$ observations.

Abstract

We studied Lyman-$α$ (Ly$α$) escape in a statistical sample of 43 Green Peas with HST/COS Ly$α$ spectra. Green Peas are nearby star-forming galaxies with strong [OIII]$λ$5007 emission lines. Our sample is four times larger than the previous sample and covers a much more complete range of Green Pea properties. We found that about 2/3 of Green Peas are strong Ly$α$ line emitters with rest-frame Ly$α$ equivalent width $>20$ Å. The Ly$α$ profiles of Green Peas are diverse. The Ly$α$ escape fraction, defined as the ratio of observed Ly$α$ flux to intrinsic Ly$α$ flux, shows anti-correlations with a few Ly$α$ kinematic features -- both the blue peak and red peak velocities, the peak separations, and FWHM of the red portion of the Ly$α$ profile. Using properties measured from SDSS optical spectra, we found many correlations -- Ly$α$ escape fraction generally increases at lower dust reddening, lower metallicity, lower stellar mass, and higher [OIII]/[OII] ratio. We fit their Ly$α$ profiles with the HI shell radiative transfer model and found Ly$α$ escape fraction anti-correlates with the best-fit $N_{HI}$. Finally, we fit an empirical linear relation to predict Ly$α$ escape fraction from the dust extinction and Ly$α$ red peak velocity. The standard deviation of this relation is about 0.3 dex. This relation can be used to isolate the effect of IGM scatterings from Ly$α$ escape and to probe the IGM optical depth along the line of sight of each $z>7$ Ly$α$ emission line galaxy in the JWST era.

Ly$α$ profile, dust, and prediction of Ly$α$ escape fraction in Green Pea Galaxies

TL;DR

This study addresses how Ly photons escape star-forming galaxies and how to predict the Ly escape fraction to disentangle ISM/CGM effects from IGM attenuation in reionization contexts. Utilizing a large, representative sample of 43 Green Pea galaxies with HST/COS Ly spectra and rich SDSS optical data, the authors measure Ly flux, equivalent width, and line profiles, and fit Ly profiles with an HI shell radiative-transfer model. They reveal strong anti-correlations between and HI-related kinematic quantities and HI column density , and show that lower dust, metallicity, and stellar mass favor higher Ly escape, with [OIII]/[OII] acting as a proxy for LyC leakage potential. Importantly, they derive an empirical linear relation linking to dust extinction and the Ly red peak velocity, with a scatter of about 0.3 dex, enabling estimates of IGM transmission along lines of sight to high- LAEs observed by JWST. These results support Green Peas as local analogs of high- LAEs and provide a practical tool to isolate IGM effects in reionization-era Ly observations.

Abstract

We studied Lyman- (Ly) escape in a statistical sample of 43 Green Peas with HST/COS Ly spectra. Green Peas are nearby star-forming galaxies with strong [OIII]5007 emission lines. Our sample is four times larger than the previous sample and covers a much more complete range of Green Pea properties. We found that about 2/3 of Green Peas are strong Ly line emitters with rest-frame Ly equivalent width Å. The Ly profiles of Green Peas are diverse. The Ly escape fraction, defined as the ratio of observed Ly flux to intrinsic Ly flux, shows anti-correlations with a few Ly kinematic features -- both the blue peak and red peak velocities, the peak separations, and FWHM of the red portion of the Ly profile. Using properties measured from SDSS optical spectra, we found many correlations -- Ly escape fraction generally increases at lower dust reddening, lower metallicity, lower stellar mass, and higher [OIII]/[OII] ratio. We fit their Ly profiles with the HI shell radiative transfer model and found Ly escape fraction anti-correlates with the best-fit . Finally, we fit an empirical linear relation to predict Ly escape fraction from the dust extinction and Ly red peak velocity. The standard deviation of this relation is about 0.3 dex. This relation can be used to isolate the effect of IGM scatterings from Ly escape and to probe the IGM optical depth along the line of sight of each Ly emission line galaxy in the JWST era.

Paper Structure

This paper contains 18 sections, 2 equations, 11 figures, 4 tables.

Figures (11)

  • Figure 1: The metallicity and dust extinction (H$\alpha$/H$\beta$ ratio) diagram of our Green Pea sample. Red squares shows the 20 galaxies with new HST observations (GO 14201, PI S. Malhotra). The other samples include 9 Green Pea galaxies with low dust extinction (cyan circle, Paper I; Henry et al. 2015), 7 Lyman-break analog galaxies (magenta pentagon, Heckman et al. 2011; Alexandroff et al. 2015), 2 Lyman-continuum leaker candidates (blue star, Jaskot et al. 2014), and 5 confirmed Lyman-continuum leakers (blue triangle, two blue triangles overlap; Izotov et al. 2016). The black hollow circles show the other galaxies without HST UV spectra in the sample of 66 Green Peas. Note that a few sources have very small Ha/Hb values. The reasons are not yet well understood, but could be 1) poor flat-field calibration or sky subtraction, 2) different gas conditions from the case-B assumption.
  • Figure 2: The $3\hbox{$^{\prime\prime}$}\times3\hbox{$^{\prime\prime}$}$ NUV images of Green Peas from the COS target acquisitions. In all panels, the colors are in log-scale with the same count-rates limits (from 0 to 0.4). These images are sorted by decreasing $f^{Ly\alpha}_{esc}$ from left to right, and from top to bottom. The label shows the ID of each Green Pea. The five LyC leakers are marked with 'LyC'. The green bar in each panel shows the physical scale of 1 Kpc.
  • Figure 3: Ly$\alpha$ emission line spectra of Green Peas before subtracting continuum. These 42 galaxies are sorted by decreasing $f^{Ly\alpha}_{esc}$ from left to right, and from top to bottom. The ID and $f^{Ly\alpha}_{esc}$ are given in each panel. The five LyC leakers are marked with 'LyC'. The last one galaxy (GP0339$-$0725) shows weak Ly$\alpha$ absorption. One Green Pea (GP0747$+$2336) is not shown here, because its Ly$\alpha$ spectra is very noisy and no Ly$\alpha$ emission or absorption lines are detected.
  • Figure 4: Comparison of the $f^{Ly\alpha}_{esc}$ and EW(Ly$\alpha$) of Green Peas.
  • Figure 5: Here we compare the rest-frame EW(Ly$\alpha$) distribution of Green Peas with different samples. The solid green line shows the sample of 28 Green Peas with EW(Ly$\alpha$) $\mathrel{\hbox{$\sim$$>$}}$ 20Å. The blue dash-dot line shows the GALEX $z\sim0.3$ LAE sample (Cowie et al. 2011; Finkelstein et al. 2009; Scarlata et al. 2009). The magenta dashed line shows the $z=2.8$ LAE sample from Zheng et al. (2016). The red dotted line shows the $z=5.7$ and $6.5$ LAE sample from Kashikawa et al. (2011).
  • ...and 6 more figures