2006AJ....131..226Y

We analyze the global structure of 34 late-type, edge-on, undisturbed, disk galaxies spanning a wide range of mass. We measure structural parameters for the galaxies using two-dimensional least-squares fitting to our R-band photometry. The fits require both a thick and a thin disk to adequately fit the data. The thick disks have larger scale heights and longer scale lengths than the embedded thin disks by factors of ∼2 and ∼1.25, respectively. The observed structural parameters agree well with the properties of thick and thin disks derived from star counts in the Milky Way and from resolved stellar populations in nearby galaxies. We find that massive galaxies' luminosities are dominated by the thin disk. However, in low-mass galaxies (V_c≲120 km/s) thick disk stars contribute nearly half the luminosity and dominate the stellar mass. Thus, although low-mass dwarf galaxies appear blue, the majority of their stars are probably quite old.

Our data are most easily explained by a formation scenario in which the thick disk is assembled through direct accretion of stellar material from merging satellites while the thin disk is formed from accreted gas. The baryonic fraction in the thin disk therefore constrains the gas richness of the merging pregalactic fragments. If we include the mass in H I as part of the thin disk, the thick disk contains ≲10% of the baryons in high-mass galaxies and ∼25%-30% of the baryons in low-mass galaxies. Our data, therefore, indicate that the fragments were quite gas rich at the time of merging (f_gas=75%-90%). However, because low-mass galaxies have a smaller fraction of baryons in their thin disks, the pregalactic fragments from which they assembled must have been systematically more gas poor. We believe this trend results from increased outflow due to supernova-driven winds in the lower mass pregalactic fragments. We estimate that ∼60% of the total baryonic mass in these systems was lost due to outflows. Pushing the episode of significant winds to early times allows the mass-metallicity relationship for disks to be established early, before the main disk is assembled, and obviates the difficulty in driving winds from diffuse disks with low star formation efficiencies. We discuss other implications of this scenario for solving the G dwarf problem, for predicting abundance trends in thick disks, and for removing discrepancies between semianalytic galaxy formation models and the observed colors of low-mass galaxies.