The Space Density of Redshift 5.7 Lyα Emitters: First Constraints from a Multislit Windows Search

We present results from a blind, spectroscopic search for redshift 5.7 Lyα emission-line galaxies at Keck I. Using a band-limiting filter and custom slit mask, we cover the LRIS detector with low-resolution spectra in the 8100-8250 Å atmospheric window, which contains no bright night-sky emission lines. We find nine objects with line fluxes greater than our flux limit of 6 × 10 -18 ergs s -1 cm -2 in our ~5.1 arcmin 2 field. We rule out a Lyα identification for six of these based on the absence of the continuum break, expected at rest 1215 Å for high- z galaxies and/or the identification of additional emission lines in our follow-up spectra. We find that extremely metal-poor, foreground emission-line galaxies are the most difficult type of interloper to recognize. For the three remaining emission-line objects, we identify a plausible counterpart for each object in a deep V -band image of the field, suggesting that none of them has a continuum break in the i band. Our preliminary conclusion is that our field contains no z = 5.7 Lyα emitters brighter than 0.6 L , where L ≡ 3.26 × 10 42 ergs s -1 . Selecting a field with zero Lyα emitters is marginally consistent with the no-evolution hypothesis—i.e., we expected to recover 2-3 Lyα emitters, assuming that the Lyα luminosity function at redshift 5.7 is the same as it is at redshift 3. Our null result rules out a brightening of L by more than a factor of 1.7 from redshift 3 to 5.7, or, over the same redshift interval, an increase of more than a factor of 2.2 in the number density of Lyα emitters. The paucity of z = 5.7 Lyα emitters raises the question of whether the Lyα-selected population plays a significant role in maintaining the ionization of the intergalactic medium (IGM) at z = 5.7. We find that if the escape fraction of Lyα radiation is less than 0.4 f Lyc , where f Lyc is the escape fraction of Lyman continuum photons, then the star formation rate in the Lyα-emitting population is high enough in the no-evolution model (our upper limit) to maintain the ionization of the IGM at z = 5.7.