Average Infrared Galaxy Spectra From Spitzer Flux Limited Samples

TL;DR

This study analyzes mid-infrared spectra of flux-limited galaxy samples from Spitzer, revealing how spectral features correlate with luminosity and enabling predictions of galaxy detectability at various redshifts.

Contribution

It provides the first combined analysis of Spitzer spectra for a flux-limited galaxy sample, establishing spectral diagnostics for luminosity and redshift predictions.

Findings

Lower luminosity sources are dominated by PAH features.

Higher luminosity sources show silicate absorption or emission.

Starbursts require luminosity evolution to match high-redshift observations.

Abstract

The mid-infrared spectroscopic analysis of a flux-limited sample of galaxies with fv(24um) > 10 mJy is presented. Sources observed are taken from the Spitzer First Look Survey (FLS) catalog and from the NOAO Deep Wide-Field Survey region in Bootes (NDWFS). The spectroscopic sample includes 60 of the 100 sources in these combined catalogs having fv(24um) > 10 mJy. New spectra from the Spitzer Infrared Spectrograph are presented for 25 FLS sources and for 11 Bootes AGN; these are combined with 24 Bootes starburst galaxies previously published to determine the distribution of mid-infrared spectral characteristics for the total 10 mJy sample. Sources have 0.01 < z < 2.4 and 41.8 < log vLv(15um) < 46.2 (ergs/s). Average spectra are determined as a function of luminosity; lower luminosity sources (log vLv(15um) < 44.0) are dominated by PAH features and higher luminosity sources (log vLv(15um) > 44.0) are dominated by silicate absorption or emission. We find that a rest frame equivalent width of 0.4um for the 6.2um PAH emission feature provides a well defined division between lower luminosity, "pure" starbursts and higher luminosity AGN or composite sources. Using the average spectra, fluxes fv(24um) which would be observed with the Spitzer MIPS are predicted as a function of redshift for sources with luminosities that correspond to the average spectra. AGN identical to those in this 10 mJy sample could be seen to z = 3 with fv(24um) > 1 mJy, but starbursts fall to fv(24um) < 1 mJy by z ~ 0.5. This indicates that substantial luminosity evolution of starbursts is required to explain the numerous starbursts found in other IRS results having fv(24um) > 1 mJy and z ~ 2.