On the Gas Content and Efficiency of AGN Feedback in Low-redshift Quasars

Abstract

The interstellar medium is crucial to understanding the physics of active galaxies and the coevolution between supermassive black holes and their host galaxies. However, direct gas measurements are limited by sensitivity and other uncertainties. Dust provides an efficient indirect probe of the total gas. We apply this technique to a large sample of quasars, whose total gas content would be prohibitively expensive to measure. We present a comprehensive study of the full (1 to 500 μ m) infrared spectral energy distributions of 87 redshift <0.5 quasars selected from the Palomar-Green sample, using photometric measurements from 2MASS, WISE , and Herschel , combined with Spitzer mid-infrared (5–40 μ m) spectra. With a newly developed Bayesian Markov Chain Monte Carlo fitting method, we decompose various overlapping contributions to the integrated spectral energy distribution, including starlight, warm dust from the torus, and cooler dust on galaxy scales. This procedure yields a robust dust mass, which we use to infer the gas mass, using a gas-to-dust ratio constrained by the host galaxy stellar mass. Most (90%) quasar hosts have gas fractions similar to those of massive, star-forming galaxies, although a minority (10%) seem genuinely gas-deficient, resembling present-day massive early-type galaxies. This result indicates that “quasar mode” feedback does not occur or is ineffective in the host galaxies of low-redshift quasars. We also find that quasars can boost the interstellar radiation field and heat dust on galactic scales. This cautions against the common practice of using the far-infrared luminosity to estimate the host galaxy star formation rate.