Quantitative interpretation of quasar microlensing light curves

We develop a general method for analyzing the light curves of microlensed quasars and apply it to the OGLE light curves of the four-image lens Q2237+0305. We simultaneously estimate the effective source velocity, the average stellar mass, the stellar mass function, and the size and structure of the quasar accretion disk. The light curves imply an effective source-plane velocity of 10, 200 km s(-1) less than or similar to v(e)h[M/M-.](-1/2) less than or similar to39, 600 km s(-1) (68% confidence). Given an independent estimate for the source velocity, found by combining estimates for the peculiar velocity of the lens galaxy with its measured stellar velocity dispersion, we obtain a mean stellar mass of [M] similar or equal to 0.037 h(2) M-. (0.0059 h(2) M-. less than or similar to [M] less than or similar to 0.20 h(2) M-.). We were unable to distinguish a Salpeter mass function from one in which all stars had the same mass, but we do find a strong lower bound of kappa(*)/kappa greater than or similar to 0.5 on the fraction of the surface mass density represented by the microlenses. Our models favor a standard thin accretion disk model as the source structure over a simple Gaussian source. For a face-on, thin disk radiating as a blackbody with temperature profile T-s proportional to R-3/4, the radius r(s) where the temperature matches the filter pass band [2000 Angstrom or T-s(r(s)) similar or equal to 7 x 10(4) K] is 1.4 x 10(15) h(-1) cm less than or similar to r(s) less than or similar to 4.5 x 10(15) h(-1) cm. The flux predicted by the disk model agrees with the observed flux of the quasar, so nonthermal or optically thin emission processes are not required. From the disk structure we estimate a black hole mass of M-BH similar or equal to 1.1(-0.7)(+1.4) x 10(9) h(-3/2) eta(0.1)(1/2) (L/L-E)(-1/2) M-., consistent with the mass estimated under the assumption that the quasar is radiating at the Eddington luminosity (L/L-E = 1).