2013ApJ...773..113C

We use the photometric redshift method of Chakrabarti & McKee to infer photometric redshifts of submillimeter galaxies with far-IR (FIR) Herschel data obtained as part of the PACS Evolutionary Probe program. For the sample with spectroscopic redshifts, we demonstrate the validity of this method over a large range of redshifts (4 ≳ z ≳ 0.3) and luminosities, finding an average accuracy in (1 + z_phot)/(1 + z_spec) of 10%. Thus, this method is more accurate than other FIR photometric redshift methods. This method is different from typical FIR photometric methods in deriving redshifts from the light-to-gas mass (L/M) ratio of infrared-bright galaxies inferred from the FIR spectral energy distribution, rather than dust temperatures. To assess the dependence of our photometric redshift method on the data in this sample, we contrast the average accuracy of our method when we use PACS data, versus SPIRE data, versus both PACS and SPIRE data. We also discuss potential selection effects that may affect the Herschel sample. Once the redshift is derived, we can determine physical properties of infrared-bright galaxies, including the temperature variation within the dust envelope, luminosity, mass, and surface density. We use data from the GOODS-S field to calculate the star formation rate density (SFRD) of submillimeter bright sources detected by AzTEC and PACS. The AzTEC-PACS sources, which have a threshold 850 µm flux ≳ 5 mJy, contribute 15% of the SFRD from all ultraluminous infrared galaxies (L_IR ≳ 10¹² L_☉), and 3% of the total SFRD at z ∼ 2.