2006AJ....132.1890B

We present an atlas of Spitzer Space Telescope Infrared Spectrograph (IRS) spectra of highly luminous, compact, mid-IR sources in the Large Magellanic Cloud. Sources were selected on the basis of IR colors and 8 µm (Midcourse Space Experiment) fluxes indicative of highly evolved, intermediate-to-high-mass stars with current or recent mass loss at large rates. We determine the chemistry of the circumstellar envelope from the mid-IR continuum and spectral features and classify the spectral types of the stars. In the sample of 60 sources, we find 21 red supergiants (RSGs), 16 C-rich asymptotic giant branch (AGB) stars, 11 H II regions, 4 likely O-rich AGB stars, 4 Galactic O-rich AGB stars, 2 OH/IR stars, and 2 B[e] supergiants with peculiar IR spectra. We find that the overwhelming majority of the sample AGB stars (with typical IR luminosities of ∼10⁴ L_☉) have C-rich envelopes, while the O-rich objects are predominantly luminous RSGs with L_IR∼10⁵ L_☉. For both classes of evolved star (C-rich AGB stars and RSGs), we use the near-to-mid-IR spectral energy distributions to determine mean bolometric corrections to the stellar K-band flux densities. For carbon stars, the bolometric corrections depend on the IR color, whereas for RSGs, the bolometric correction is independent of IR color. Our results reveal that objects previously classified as planetary nebulae on the basis of IR colors are, in fact, compact H II regions with very red IRS spectra that include strong atomic recombination lines and polycyclic aromatic hydrocarbon emission features. We demonstrate that the IRS spectral classes in our sample separate clearly in IR color-color diagrams that use combinations of Two Micron All Sky Survey data and synthetic Spitzer Infrared Array Camera and Multiband Imaging Photometer for Spitzer fluxes derived from the IRS spectra. On this basis we suggest diagnostics to identify and classify, with high confidence levels, IR-luminous evolved stars and compact H II regions in nearby galaxies using Spitzer and near-IR photometry.