2013ApJ...768..178F

The stacked spectral energy distribution (SED) 24 µm Lyman break galaxies (MIPS-LBGs) detected by the Multiband Imaging Photometer for Spitzer (MIPS) is fitted by means of the spectrophotometric model GRASIL with an "educated" fitting approach which benefits from the results of chemical evolution models. The star formation rate-age-metallicity degeneracies of SED modeling are broken by using star formation history (SFH) and chemical enrichment history suggested by chemical models. The dust mass, dust abundance, and chemical pattern of elements locked in the dust component are also directly provided by chemical models. Using our new "fitting" approach, we derive the total mass M_tot, stellar mass M_*, gas mass M_g, dust mass M_d , age, and star formation rate (SFR) of the stacked MIPS-LBG in a self-consistent way. Our estimate of M_*= 8x10¹⁰ of the stacked MIPS-LBG agrees with other works based on UV-optical SED fitting. We suggest that the MIPS-LBGs at z ∼ 3 are young (0.3-0.6 Gyr), massive (M_tot∼ 10¹¹ M_☉), dusty (M_d∼ 10⁸ M_☉), and metal-rich (Z ∼ Z_☉) progenitors of elliptical galaxies undergoing a strong burst of star formation (SFR ∼ 200 M_☉/yr). Our estimate of M_d= 7x10⁷ M_☉of the stacked MIPS-LBG is about a factor of eight lower than the estimated value based on single temperature graybody fitting, suggesting that self-consistent SED models are needed to estimate dust mass. By comparing with Milky Way molecular cloud and dust properties, we suggest that denser and dustier environments and flatter dust size distribution are likely in high-redshift massive star-forming galaxies. These dust properties, as well as the different types of SFHs, can cause different SED shapes between high-redshift star-forming ellipticals and local starburst templates. This discrepancy of SED shapes could in turn explain the non-detection at submillimeter wavelengths of IR luminous (L _IR{sce}10¹² L_☉) MIPS-LBGs.