Average metallicity and star formation rate of Ly$\alpha$ emitters probed by a triple narrowband survey*

Abstract

We present the average metallicity and star formation rate (SFR) of $Ly\alpha$ emitters (LAEs) measured from our large-area survey with three narrowband (NB) filters covering the $Ly\alpha, \big[O II\big]\lambda3727$, and $\text{H}\alpha+\big[\text{N} II\big]$ lines of LAEs at $z = 2.2$. We select 919 $z = 2.2$ LAEs from Subaru/Suprime-Cam NB data in conjunction with Magellan/IMACS spectroscopy. Of these LAEs, 561 and 105 are observed with KPNO/NEWFIRM near-infrared NB filters whose central wavelengths are matched to redshifted $\big[\text{O} II\big]$ and $\text{H}\alpha$ nebular lines, respectively. By stacking the near-infrared images of the LAEs, we successfully obtain average nebular-line fluxes of LAEs, the majority of which are too faint to be identified individually by NB imaging or deep spectroscopy. The stacked object has an Hα luminosity of $1.7 × 10^{42} erg s^{–1}$ corresponding to an SFR of $14 M_{\odot} yr6{–1}$. We place, for the first time, a firm lower limit to the average metallicity of LAEs of $Z \gtrsim 0.09 Z_{\odot} \big(2\sigma\big)$ based on the $\big[\text{O} II\big]/\big(\text{H}\alpha+\big[\text{N} II\big]\big)$ index together with photoionization models and empirical relations. This lower limit of metallicity rules out the hypothesis that LAEs, so far observed at $z \sim 2$, are extremely metal-poor $\big(Z < 2 × 10–2 Z_{\odot}\big)$ galaxies at the $4\sigma$ level. This limit is higher than a simple extrapolation of the observed mass-metallicity relation of $z \sim 2$ UV-selected galaxies toward lower masses $\big(5 × 108 M_{\odot}\big)$, but roughly consistent with a recently proposed fundamental mass-metallicity relation when the LAEs' relatively low SFR is taken into account. The $\text{H}\alpha$ and $\text{Ly}\alpha$ luminosities of our NB-selected LAEs indicate that the escape fraction of $\text{Ly}\alpha$ photons is $\sim12%-30%$, much higher than the values derived for other galaxy populations at $z \sim 2$.