Iron and alpha element distributions in Milky Way dwarf satellite galaxies from medium-resolution spectroscopy

Multi-element abundance measurements provide a window into the star formation histories of dwarf galaxies. A dearth of stellar spectra suitable for such measurements has previously hampered progress in this field. We present Keck/DEIMOS spectra for nearly 2500 stars in 13 dwarf satellite galaxies of the Milky Way. Through spectral synthesis, we measure their Fe, Mg, Si, Ca, and Ti abundances. The seven dwarfs with L < 2 x 105 L⊙ have 145 synthesis-based abundances of the above elements, compared to six previously published high-resolution spectroscopic measurements. Each of the six more luminous dwarfs contains more multi-element abundance measurements than has been presented before. We assess the accuracy of our measurements and estimate their measurement uncertainties by comparisons to previously published high-resolution spectroscopic abundance measurements of the same stars in globular clusters, the Milky Way field halo, and Milky Way dwarf satellite galaxies. Mg, Si, Ca, and Ti are alpha elements, and the [alpha/Fe] ratio of a stellar system reveals the profile of star formation. From the metallicity distributions and [alpha/Fe] patterns, we qualitatively characterize how the present dwarf galaxy luminosity predicts its star formation profile. For a more luminous dSph, the average metallicity is higher; the metallicity distribution is more consistent with a chemical evolution model that incorporates gas accretion than a closed box; and [alpha/Fe] declines less steeply or not at all with increasing [Fe/H]. From these observations, we conclude that more luminous dwarf galaxies experience more prolonged or multiple episodes of vigorous star formation whereas less luminous dwarf galaxies form stars less vigorously. We also detect extremely metal-poor ([Fe/H] < -3) stars in eight dwarf satellite galaxies. Furthermore, the fractional number of stars with [Fe/H] < -2 exceeds that of the Milky Way halo. We therefore conclude that the observed dwarf galaxy [Fe/H] distributions can no longer be used as an argument against the hierarchical formation of the Milky Way stellar halo from ancient dwarf galaxies.