LALA Looks at the Early Universe

Based on a Solicited Contribution from James Rhoads

Sangeeta Malhotra and James Rhoads (both at NOAO) have initiated a Large Area Lyman Alpha ("LALA") survey using the KPNO 4-m Mosaic CCD camera to search for Ly- emitting objects in the early universe. Keck spectra obtained of some of their first candidate objects demonstrate that the survey is indeed detecting galaxies at high redshift.

Ly- emission has long been sought as a signpost of galaxies in formation. The history of the field began with a theoretical prediction by Partridge & Peebles (1967) that the first burst of star formation in giant galaxies should produce bright, easily observable Ly- line emission. This prediction prompted many emission line protogalaxy searches (see Pritchet 1994 for a review). However, no search up to 1997 detected the hoped-for field population, demonstrating that the putative emitters are either faint or rare. Subsequent studies have detected the first such objects, but the total sample remains very small.

The primary challenge in searching for Ly- emitting objects is to achieve a suitable combination of sensitivity and survey area. The wide field of the NOAO CCD Mosaic camera allows narrowband searches to be conducted with a much higher efficiency than previous instruments (measured by the quantity A, the product of telescope collecting area and detector solid angle). Malhotra & Rhoads began the LALA survey in the spring of 1998 to exploit this new capability.

To date, Malhotra & Rhoads have completed imaging observations of one Mosaic field in five overlapping H- filters, with a total integration time of 6 hours per filter. Additional observations of other fields in a subset of the filters nearly double this data set. In all, the sampled volume is 1.3×10^-6 comoving Mpc^-3 at redshifts from 4.37 to 4.56 (assuming a cosmology with H₀=70, _m=0.2, =0). The 5 sensitivity limit in each 80Å filter is 2.6×10^-17 erg/cm²/s. Because the Mosaic H- filters overlap, fainter sources (to 1.8×10^-17 erg/cm²/s) can be detected at wavelengths sampled by two filters, or over 40% of the volume in the current data set. As these limits include both line and continuum fluxes, the faintest detected emission lines can be appreciably weaker (~40 % of the line+continuum limit) for souces with equivalent widths smaller than the 80Å bandpass.

The first spectroscopic follow-up of LALA survey sources was obtained at the Keck 10-m telescope by Arjun Dey (NOAO), Daniel Stern, Hy Spinrad, and collaborators (UC Berkeley). These spectra yielded a confirmed z=4.5 Ly- emitter, a comparatively weak source with a line flux of 1.7×10^-17 erg/cm²/s and an 80Å equivalent width. They also elucidated the nature of the most dramatic emission line yet found in the survey (with equivalent width ~1400Å and flux 7×10^-16 erg/cm²/s), which turned out to be the O[III] 5007Å line from an unusual, high-excitation galaxy at z=0.33. On- and off-band images of both sources are shown in Figure 1, and corresponding spectra are shown in Figure 2. By obtaining spectra of sources with a variety of fluxes, equivalent widths, and broadband colors, Malhotra & Rhoads optimized these first spectra to help refine selection criteria for later follow-up. In so doing, they demonstrated that they could achieve high reliability for sources with equivalent widths as small as 50Å in the observer frame.

0.4% of the LALA project's first field. Each image is 2.2' on a side. Left: The H-+16 filter (=6721 Å, =80 Å). Two emission line sources are marked. Middle: The H-+0 filter (=6561 Å, =80 Å). Right: The NOAO Deep Wide-Field Survey R band data for this field (courtesy Jannuzi, Dey, and the NDWFS Team).

Spectra of the two sources marked in Figure 1, obtained at the Keck by Dey, Stern, Spinrad, and collaborators. Top: The brighter source, an extraordinary O[III] emitter at z=0.34. Bottom: A smoothed spectrum of the fainter source, a z=4.5 Ly- emitter.

To maximize scientific return on the narrowband observations, the primary LALA survey fields are placed in fields studied by the NOAO Deep Wide-Field Survey. This allows LALA to combine narrowband emission line strength estimates with broadband colors to better discriminate among different classes of candidate emission line objects.

Typical z=4.5 Ly- emitters are expected to have faint flux levels (< 5.2×10^-17 erg/cm²/s) and large equivalent widths (>80 Å). The number density of such sources in the first field is about 225 per field per filter brighter than 2.6×10^-17 erg/cm²/s line+continuum flux (the single filter 5 detection threshold). This would correspond to a source density of 11000 per square degree per unit z if all of these sources were bona fide Ly- emitters. Early spectroscopic follow-up suggests that in fact some 30 to 50% of these sources are Ly- emitters, leaving 4000 per square degree per unit z in this narrow range of flux. Further planned spectroscopic follow-up will improve these statistics. They expect a final sample of ~10³ objects.

Given this large sample, Malhotra & Rhoads expect to determine the distribution of line fluxes, equivalent widths, and colors for z=4.5 Ly- emitters. By studying the spatial distribution of emitters, they will also be able to examine large-scale structure issues for this population. Ultimately, by comparing this data set with present day and intermediate redshift galaxy populations, it will be possible to study the evolution of Ly- emitters and to understand their role in the galaxy formation process.