Rest-frame UV spectroscopy of extreme [OIII] emitters at $1.3<z<3.7$: Toward a high-redshift UV reference sample for JWST

Abstract

Deep spectroscopy of galaxies in the reionization-era has revealed intense CIII] and CIV line emission (EW $>15-20$ \r{A}). In order to interpret the nebular emission emerging at $z>6$, we have begun targeting rest-frame UV emission lines in galaxies with large specific star formation rates (sSFRs) at $1.3<z<3.7$. We find that CIII] reaches the EWs seen at $z>6$ only in large sSFR galaxies with [OIII]+H$\beta$ EW $>1500$ \r{A}. In contrast to previous studies, we find that many galaxies with intense [OIII] have weak CIII] emission (EW $=5-8$ \r{A}), suggesting that the radiation field associated with young stellar populations is not sufficient to power strong CIII]. Photoionization models demonstrate that the spread in CIII] among systems with large sSFRs ([OIII]+H$\beta$ EW $>1500$ \r{A}) is driven by variations in metallicity, a result of the extreme sensitivity of CIII] to electron temperature. We find that the strong CIII] emission seen at $z>6$ (EW $>15$ \r{A}) requires metal poor gas ($\simeq0.1\ Z_\odot$) whereas the weaker CIII] emission in our sample tends to be found at moderate metallicities ($\simeq0.3\ Z_\odot$). The luminosity distribution of the CIII] emitters in our $z\simeq1-3$ sample presents a consistent picture, with stronger emission generally linked to low luminosity systems ($M_{\rm{UV}}>-19.5$) where low metallicities are more likely. We quantify the fraction of strong CIII] and CIV emitters at $z\simeq1-3$, providing a baseline for comparison against $z>6$ samples. We suggest that the first UV line detections at $z>6$ can be explained if a significant fraction of the early galaxy population is found at large sSFR ($>200$ Gyr$^{-1}$) and low metallicity ($<0.1\ Z_\odot$).