Probing the Coevolution of Supermassive Black Holes and Quasar Host Galaxies

Abstract

At low redshift, there are fundamental correlations between the mass of supermassive black holes (M BH ) and the mass (M bulge ) and luminosity of the host galaxy bulge. We investigate the same relation at z k 1. Using virial mass estimates for 11 quasars at z k 2 to measure their black hole mass, we find that black holes at high z fall nearly on the same M BH versus R-band magnitude (M R ) relation (to $0.3 mag) as low-redshift active and inactive galaxies, without making any correction for luminosity evolution. Using a set of conservative assumptions about the host galaxy stellar population, we show that at z k 2 (10 Gyr ago), the ratio of M BH /M bulge was 3-6 times larger than today. Barring unknown systematic errors on the measurement of M BH , we also rule out scenarios in which moderately luminous quasar hosts at z k 2 were fully formed bulges that passively faded to the present epoch. On the other hand, five quasar hosts at z % 1 are consistent with the current-day M BH -M R relationship after taking into account evolution that is appropriate for E/S0 galaxies. Therefore, z % 1 host galaxies appear to fit the hypothesis that they are fully formed early-type galaxies. We also find that most quasar hosts with absolute magnitudes brighter than M R ¼ À23 cannot fade below L Ã galaxies today, regardless of their stellar population makeup, because their black hole masses are too high and they must arrive at the local M BH -M R relationship by z ¼ 0.

An Upper Limit to the Degree of Evolution between Supermassive Black Holes and Their Host Galaxies

“…The upper limits we derive are competitive with or stronger than those from direct, but difficult, observations of high-redshift BHs and their hosts, and can be applied in any band (including those for which direct estimates of these correlations at z > 0 do not exist). Our upper limits are consistent with passive evolution (with reasonable formation redshifts z f P 6) at all redshifts and in every band, as well as hydrodynamic simulations of BH and spheroid coformation at different redshifts (Robertson et al 2006) and various estimates of these correlations at low redshifts (z P 1; Shields et al 2003;Peng et al 2006). The weak evolution inferred by Merloni et al (2004) cannot be ruled out for sufficiently low z f P 3, but the mass-function data prefer a noevolution case at $2 .…”

“…2, open symbols) by, e.g., Treu et al (2004) (2006) are ruled out at k6 with respect to the upper limits we measure over each corresponding redshift interval (note that this does not necessarily mean their measurements are inconsistent with our upper limits at 6 , as the error bars in the direct observations are large and generally inconsistent with passive evolution at only $2Y3 ). It should be noted, however, that there is a factor $2 systematic normalization uncertainty in the virial BH mass estimators these authors adopt, and lowering, e.g., the estimates of Peng et al (2006) by such a factor (i.e., allowing for a factor of $2 increase in M BH /M Ã by z $2) makes them consistent with our upper limits.…”

“…Note that this notation for À(z) is used by Peng et al (2006) but in their case refers to individual systems, rather than the evolution of the mean relation. The third equality comes from the physical fact that it is not possible to destroy BH mass; the total mass above a given M min BH can never decrease from z to the present.…”

“…We also plot (open black circles) the recent estimates of the evolution in these correlations from Peng et al (2006). The authors directly measure the correlation with R; we use their adopted model to convert to mass, B-band, and K-band estimates.…”

“…Consequently, different groups have reached seemingly contradictory conclusions, with O [iii] velocity dispersion (Shields et al 2003) and BH clustering (Adelberger & Steidel 2005) measurements finding no evolution in these correlations out to redshifts z $ 4, while CO velocity dispersion ( Walter et al 2004;Shields et al 2006) and host R-band luminosity ( Peng et al 2006) observations imply strong evolution (with BHs becoming overmassive relative to their hosts) occurring at z k 1. Spectral template fitting ( Treu et al 2004;Woo et al 2006) suggests strong evolution at much lower redshifts z ¼ 0:37, and comparison of radio quasar and galaxy populations ( McLure et al 2006) indicates evolution of the form /(1 þ z) 2 .…”