Lyα luminosity functions at redshift z ≈ 4.5

Abstract

We present a spectroscopically confirmed sample of Lyman α emitting galaxies (LAEs) at z ∼ 4.5 in the Extended Chandra Deep Field-South (ECDFS), which we combine with a sample of z ∼ 4.5 LAEs from previous narrow-band surveys from the Large Area Lyman Alpha (LALA) survey to build a unified Lyα luminosity function. We spectroscopically observed 64 candidate LAEs in the ECDFS, confirming 46 objects as z ∼ 4.5 LAEs based on single-line detections with no continuum emission bluewards of the line, resulting in a Lyα confirmation rate of ∼70 per cent. We did not detect significant flux from neither the C IV λ1549 Å emission line nor the He II λ1640 Å emission line in individual LAE spectra. These lines were also undetected in a co-added spectrum, with the co-added line ratio of He II to Lyα constraining the Population III star formation rate (SFR) to be <0.3 per cent of the total SFR and <1.25 per cent of the observed SFR (both at the 2σ level). We combine the optical spectra with deep X-ray and radio images to constrain the AGN fraction in the sample. Only LAE was detected in both the X-ray and radio, while the other objects remained undetected, even when stacked. The Lyα luminosity functions in our two deepest narrow-band filters in the ECDFS differ at greater than 2σ significance, and the product L*Φ* differs by a factor of >3. Similar luminosity function differences have been used to infer evolution in the neutral gas fraction in the intergalactic medium at z > 6, yet here the difference is likely due to cosmic variance, given that the two samples are from adjoining line-of-sight volumes. Combining our new sample of LAEs with those from previous LALA narrow-band surveys at z = 4.5, we obtain one of the best measured Lyα luminosity functions to date, with our sample of over 200 spectroscopically confirmed Lyα galaxies yielding log₁₀(L*) = 42.83 ± 0.06 (erg s^-1) and log₁₀(Φ*) = -3.48 ± 0.09 (Mpc^-3). We compare our new luminosity function to others from the literature to study the evolution of the Lyα luminosity density at 0 < z < 7. We find tentative evidence for evolution in the product L*Φ*, which approximately tracks the cosmic SFR density, but since field-to-field and survey-to-survey variations are in some cases as large as the possible evolution, some caution is needed in interpreting this trend.