A deep AAOmega survey of low-luminosity galaxies in the Shapley supercluster: stellar population trends

We present new optical spectroscopy for 342 R < 18 galaxies in the Shapley supercluster ( and 198 supplementary galaxies), obtained from 8- h integrations with the AAOmega facility at the Anglo- Australian Telescope. We describe the observations and measurements of central velocity dispersion sigma, emission- line equivalent widths and absorption- line indices. The distinguishing characteristic of the survey is its coverage of a very wide baseline in velocity dispersion ( 90 per cent range s = 40 - 230 km s(-1)), while achieving high signal- to- noise ratio throughout ( median 60 angstrom- 1 at 5000 angstrom). The data quality will enable estimates of Balmer- line ages to better than 20 per cent precision even for the faintest galaxies in the sample. Significant emission at Ha was detected in similar to 30 per cent of the supercluster galaxies, including similar to 20 per cent of red- sequence members. Using line- ratio diagnostics, we find that the emission is LINER ( low ionization nuclear emission region) like at high luminosity, but driven by star formation in low- luminosity galaxies. To characterize the absorption lines, we use the classical Lick indices in the spectral range 4000 - 5200 angstrom. We introduce a new method for applying resolution corrections to the line- strength indices. We define a subset of galaxies with very low emission contamination, based on the Ha line, and fit the index - s relations for this subset. The relations show the continuation of the familiar trends for giant galaxies into the low- luminosity regime, with little change in slope for most indices. Comparing the index - s slopes against predictions from single- burst stellar population models, we infer the scaling relations of age, total metallicity, [ Z/ H], and alpha- element abundance ratio, [ a/ Fe]. To reproduce the observed index - s slopes, all three parameters must increase significantly with increasing velocity dispersion. Specifically, we recover Age alpha sigma(0.52 +/- 0.06 +/- 0.10), Z/ H alpha sigma(0.34 +/- 0.04 +/- 0.07) and a/ Fe alpha sigma(0.23 +/- 0.04 +/- 0.06) ( where the second error reflects systematic effects), derived over a decade baseline in velocity dispersion, s = 30 - 300 km s(-1). The equivalent slopes for the subset of galaxies withs > 100 km s(-1) are similar for age and Z/ H. For a/ Fe, a steeper slope is recovered for the high-sigma subset, a/ Fe alpha sigma(0.36 +/- 0.07). The recovered age -sigma relation is shown to be consistent with the observed evolution in the giant- to- dwarf galaxy ratio in clusters at redshifts z = 0.4 - 0.8. In a companion paper, we determine the age, [ Z/ H] and [ a/ Fe] for individual galaxies, and investigate in detail the distribution of galaxy properties at fixed velocity dispersion.