The history of star formation in nearby dwarf galaxies.

Abstract

We present detailed analysis of color-magnitude diagrams (CMDs) of resolved stellar populations in nearby dwarf galaxies based on observations taken with the Hubble Space Telescope (HST). From the positions of individual stars on a CMD, we are able to derive the star formation histories (SFHs), i.e., the star formation rate (SFR) as a function of time and metallicity, of the observed stellar populations. Specifically, we apply this technique to a number of nearby dwarf galaxies to better understand the mechanisms driving their evolution. The ACS Nearby Galaxy Survey Treasury program (ANGST) provides multi-color photometry of resolved stars in ∼ 60 nearby dwarf galaxies from images taken with HST, This sample contains 12 dSph, 5 dwarf spiral, 28 dIrr, 12 dSph/dIrr (transition), and 3 tidal dwarf galaxies. The sample spans a range of ∼ 10 in MB and covers a wide range of environments, from highly interacting to truly isolated. From the best fit lifetime SFHs we find three significant results: (1) the average dwarf galaxy formed ∼ 60% of its stars by z ∼ 2 and 70% of its stars by z ∼ 1, regardless of morphological type, (2) the only statistically significant difference between the SFHs of different morphological types is within the most recent 1 Gyr (excluding tidal dwarf galaxies), and (3) the SFHs are complex and the mean values are inconsistent with simple SFH models, e.g., single epoch SF or constant SFH. We find that an exponentially declining star formation model with τ ∼ 6.4 Gyr provides a reasonable representation of the mean SFHs. The dominance of the older stellar populations implies that the typical dwarf galaxy formed most of its stars at times similar to more massive galaxies. The sample shows a strong density-morphology relationship, i.e., the dSphs in the sample are less isolated than dIs. We find that the transition from a gas-rich to gas-poor galaxy cannot be solely due to internal mechanisms, e.g., stellar feedback, and is likely the result of external mechanisms, e.g., ram pressure and tidal stripping and tidal forces. The average transition dwarf galaxy is slightly less isolated and less gas-rich than the typical dwarf irregular. From a comparison between the ANGST and Local Group dwarf galaxy SFHs, we find consistency between the two samples, suggesting that the Local Group dwarf galaxies are a good representation of the broader universe. Both samples show a similarly strong morphology-density relationship, further emphasizing the importance of environment in dwarf galaxy evolution. We then present the recent (. 1 Gyr) SFHs of nine M81 Group Dwarf Galaxies. Comparing the SFHs, birthrate parameters, fraction of stars formed per time interval, and spatial distribution of stellar components as a function of luminosity, we find only minor differences in SF characteristics among the M81 Group dIs despite a wide range of physical properties. We extend our comparison to select dIs in the Local Group (LG), with similar quality photometry, and again find only minor differences in SF parameters. The lack of a clear trend in SF parameters over a wide range of diverse environments suggests that SF in low mass systems may be dominated by stochastic processes. The fraction of stars formed per time interval for an average M81 Group and LG dI is consistent with a constant SFH. However, individual galaxies can show significant departures from a constant SFH. Thus, we find this result underlines the importance of stochastic SF in dIs. In addition to a statistical comparison, we examine possible formation scenarios of the less luminous and candidate tidal dwarfs in the M81 Group. The SFHs and the lack of an overdensity of associated red stars suggest that the Garland and Ho IX are not dIs and are potentially tidal dwarf galaxies. Interestingly, a noteworthy difference between the LG and the M81 Group is the lack of tidal dwarf candidates in the LG. Comparing the recent SFHs and spatial locations of young stars with observations of the neutral interstellar medium (HI), we are able to gain new insight into the physics of stellar ‘feedback’. We first make this type of comparison in IC 2754, a luminous dwarf irregular galaxy in the M81 Group with a ∼ 1 kpc supergiant HI shell. We find two significant episodes of SF inside the SGS from 200 − 300 Myr and ∼ 25 Myr ago. Comparing the timing of the SF events to the dynamic age of the SGS and the energetics from the HI and SF, we find compelling evidence that stellar feedback is responsible for creating the SGS and triggering secondary SF around its rim. We then conduct an extensive analysis of HI holes in M81 Group dwarf irregular galaxy, Holmberg II. From the deep photometry, we construct the CMDs and measure the SFHs for stars contained in HI holes from two independent holes catalogs, as well as select control fields, i.e., similar sized regions that span a range of HI column densities. The CMDs reveal young (< 200 Myr) stellar populations inside all HI holes, which contain very few bright OB stars with ages less than 10 Myr, indicating they are not reliable tracers of HI hole locations while the recent SFHs confirm multiple episodes of star formation within most holes. Converting the recent SFHs into stellar feedback energies, we find that enough energy has been generated to have created all holes. However, the required energy is not always produced over a time scale that is less than the estimated kinematic age of the hole. A similar analysis of stars in the control fields finds that the stellar populations of the control fields and HI holes are statistically indistinguishable. However, because we are only sensitive to holes ∼ 100 pc in diameter, we cannot tell if there are smaller holes inside the control fields. The combination of the CMDs, recent SFHs, and locations of young stars shows that the stellar populations inside HI holes are not coherent, single-aged, stellar clusters, as previously suggested, but rather multi-age populations distributed across each hole. From a comparison of the modeled and observed integrated magnitudes, and the locations and energetics of stars inside of HI holes, we propose a potential new model: a viable mechanism for creating the observed HI holes in Ho II is stellar feedback from multiple generations of SF spread out over tens or hundreds of Myr, and thus, the concept of an age for an HI hole is intrinsically ambiguous. For HI holes in the outer parts of Ho II, located beyond the HST/ACS coverage, we use Monte Carlo simulations of expected stellar populations to show that low level SF could provide the energy necessary to form these holes. Applying the same method to the SMC, we find that the holes that appear to be void of stars could have formed via stellar feedback from low level SF. We further find that Hα and 24μm emission, tracers of the most recent star formation, do not correlate well with the positions of the HI holes. However, UV emission, which traces star formation over roughly the last 100 Myr, shows a much better correlation with the locations of the HI holes.