BLACK HOLE MASSES OF INTERMEDIATE-REDSHIFT QUASARS: NEAR-INFRARED SPECTROSCOPY

We present near-infrared spectra of 10 luminous, intermediate-redshift quasars (z similar or equal to 2; L-bol similar or equal to 10(47) erg s(-1)), observed with SofI at the NTT of ESO/La Silla. With these rest-frame optical spectra we probe the H beta-[O III] emission line region. Using the standard scaling relation involving the width of the H beta line and the continuum luminosity, we measure black hole (BH) masses in the range of similar to 2 x 10(9) less than or similar to M-bh less than or similar to 10(10) M-circle dot for these sources. We also used Sloan Digital Sky Survey spectra to probe Mg II lambda 2798 and C IV lambda 1549 emission lines and used these for BH mass measurements as well. The BH mass estimates using C IV lambda 1549 are on average smaller by about 60% than those based on II beta. The massive BHs we observe could not have grown by simple radiatively efficient accretion at the observed accretion rate starting from seeds of up to a thousand solar masses. About 10% of the observed BH mass must have been accumulated by earlier merger events and radiatively inefficient accretion. Radiatively efficient accretion would further grow these BHs to masses of several 10(9) M-circle dot in 2-3 e-folding times, i.e., in several 10(8) yr. This scenario is consistent with recent models of BH growth. The H beta-based Eddington luminosity ratios are in the range of similar to 0.2-0.7, with an average of < L-bol/L-edd > = 0.39 +/- 0.05. The Lbol/Ledd ratio distribution follows a lognormal distribution which is consistent with prior studies of quasars with comparable luminosity. We also find that the gas metallicity of the broad-line region is super-solar with similar to 3 Z/Z(circle dot), based on N III]lambda 1750/O III]lambda 1663 and N v lambda 1240/C IV lambda 1549 emission line ratios. We find no correlation of the gas metallicity with the optical Fe II emission line strength in our small sample, contrary to a recent suggestion.