Mid‐Infrared Spectral Diagnosis of Submillimeter Galaxies

Abstract

We present deep mid-IR spectroscopy with Spitzer of 13 SMGs in the GOODS-N field. We find strong PAH emission in all of our targets, which allows us to measure mid-IR spectroscopic redshifts and place constraints on the contribution from star formation and AGN activity to the mid-IR emission. In the high-S/ N composite spectrum, we find that the hot dust continuum from an AGN contributes at most 30% of the mid-IR luminosity. Individually, only 2/13 SMGs have continuum emission dominating the mid-IR luminosity; one of these SMGs, C1, remains undetected in the deep X-ray images but shows a steeply rising continuum in the mid-IR indicative of a Compton-thick AGN. We find that the mid-IR properties of SMGs are distinct from those of 24 m-selected ULIRGs at z $ 2; the former are predominantly dominated by star formation, while the latter are a more heterogeneous sample with many showing significant AGN activity. We fit the IRS spectrum and the mid-IR to radio photometry of SMGs with template SEDs to determine the best estimate of the total IR luminosity from star formation. While many SMGs contain an AGN as evinced by their X-ray properties, our multiwavelength analysis shows that the total IR luminosity, L IR , in SMGs is dominated by star formation. We find that high-redshift SMGs lie on the relation between L IR and L PAH;6:2 (or L PAH;7:7 or L PAH;11:3 ) that has been established for local starburst galaxies. This suggests that PAH luminosity can be used as a proxy for the SFR in SMGs. SMGs are consistent with being a short-lived cool phase in a massive merger where the AGN does not appear to have become strong enough to heat the dust and dominate the mid-or far-IR emission.

Dusty star-forming galaxies at high redshift

“…This figure highlights the expected change in observed flux density with redshift for a template DSFG SED of fixed luminosity (Pope et al, 2008b) across many observed-frame wavelengths, from the optical (i-band, 0.8µm) through the near-IR, mid-IR, far-IR, millimeter, and the radio (1.4 GHz). The K-correction at 250 < λ < 500 µm is still negative, though is far less dramatic than the very negative K-correction at ∼1 mm and the very positive K-correction at optical/near-infrared and radio wavelengths which causes a dramatic drop in redshift flux density.…”

“…The timing was perfect for investigating the mid-infrared spectral properties of 850µm-selected SMGs, now described in Menéndez-Delmestre et al (2007), Valiante et al (2007), Pope et al (2008b) and Menéndez-Delmestre et al (2009). These works find that SMGs have a majority of broad PAH emission features (80%) with the remaining 20% of SMGs being dominated in the mid-IR by AGN.…”

“…Figure 5 illustrates a characteristic SED of a z = 2 L IR = 3 × 10 12 L galaxy (Pope et al, 2008b) in relation to some well known local ULIRGs (Arp 220, Mrk 231, templates provided by Polletta et al, 2007). Over-plotted on this SED are characteristic detection limits for the deepest mid-IR, submillimeter and millimeter facilities available.…”

“…This illustrates the nearly-unchanged flux densities which DSFGs have in the ∼1 mm bands across a wide range of epochs 1 < ∼ z < ∼ 10. The Pope et al (2008b) composite SMG SED is used to generate the flux evolution, or K − correction, at wavelengths 24µm-2 mm, and at 1.4 GHz. An Arp 220 SED, adjusted to have a luminosity of 10 12.5 L , is used to generate optical and near-IR K−corrections.…”

“…A Figure 5: A schematic spectral energy distribution for a dusty star forming galaxy at z = 2. A representative SED for a ∼500 M yr −1 SMG (black; Pope et al, 2008b) is overplotted on a gray band representing the plausible range of SED types for a galaxy of fixed infrared luminosity 10 12.5 L with peak SED wavelengths ranging from 70-130µm (corresponding to temperatures ∼30-58K). Model SEDs for local galaxies Arp 220 and Mrk 231 are also overplotted in red dot-dashed and dotted lines.…”