A log-quadratic relation for predicting supermassive black hole masses from the host bulge Sersic index

Abstract

We reinvestigate the correlation between black hole mass and bulge concentration. With an increased galaxy sample, updated estimates of galaxy distances, black hole masses, and Sersic indices `n' - a measure of concentration - we perform a least-squares regression analysis to obtain a relation suitable for the purpose of predicting black hole masses in other galaxies. In addition to the linear relation, log(M_bh) = 7.81(+/-0.08) + 2.69(+/-0.28)[log(n/3)] with epsilon_(intrin)=0.31 dex, we investigated the possibility of a higher order M_bh-n relation, finding the second order term in the best-fitting quadratic relation to be inconsistent with a value of zero at greater than the 99.99% confidence level. The optimal relation is given by log(M_bh) = 7.98(+/-0.09) + 3.70(+/-0.46)[log(n/3)] - 3.10(+/-0.84)[log(n/3)]^2, with epsilon_(intrin)=0.18 dex and a total absolute scatter of 0.31 dex. Extrapolating the quadratic relation, it predicts black holes with masses of ~10^3 M_sun in n=0.5 dwarf elliptical galaxies, compared to ~10^5 M_sun from the linear relation, and an upper bound on the largest black hole masses in the local universe, equal to 1.2^{+2.6}_{-0.4}x10^9 M_sun}. In addition, we show that the nuclear star clusters at the centers of low-luminosity elliptical galaxies follow an extrapolation of the same quadratic relation. Moreover, we speculate that the merger of two such nucleated galaxies, accompanied by the merger and runaway collision of their central star clusters, may result in the late-time formation of some supermassive black holes. Finally, we predict the existence of, and provide equations for, a relation between M_bh and the central surface brightness of the host bulge.