We present evidence for ultraviolet/optical microlensing in the gravitationally lensed quasar Q0957+561. We combine new measurements from our optical monitoring campaign at the United States Naval Observatory, Flagstaff (USNO) with measurements from
the literature and find that the time-delay-corrected r-band flux ratio m_A - m_B has increased by ~0.1 magnitudes over a period of five years beginning in the fall of 2005. We apply our Monte Carlo microlensing analysis procedure to the composite light curves, obtaining a measurement of the optical accretion disk size, log {(r_s/cm)[cos(i)/0.5]^{1/2}} = 16.2^{+0.5}_{-0.6}, that is consistent with the quasar accretion disk size - black hole mass relation.
We analyzed the microlensing of the X-ray and optical emission of the lensed quasar PG 1115+080. We find that the effective radius of the X-ray emission is 1.3(+1.1 -0.5) dex smaller than that of the optical emission. Viewed as a thin disk observed a
t inclination angle i, the optical accretion disk has a scale length, defined by the point where the disk temperature matches the rest frame energy of the monitoring band (kT=hc/lambda_rest with lambda_rest=0.3 micron), of log[(r_{s,opt}/cm)(cos(i) / 0.5)^{1/2}] = 16.6 pm 0.4. The X-ray emission region (1.4-21.8 keV in the rest frame) has an effective half-light radius of log[r_{1/2,X}/cm] = 15.6 (+0.6-0.9}. Given an estimated black hole mass of 1.2 * 10^9 M_sun, corresponding to a gravitational radius of log[r_g/cm] = 14.3, the X-ray emission is generated near the inner edge of the disk while the optical emission comes from scales slightly larger than those expected for an Eddington-limited thin disk. We find a weak trend supporting models with low stellar mass fractions near the lensed images, in mild contradiction to inferences from the stellar velocity dispersion and the time delays.
We expand our Bayesian Monte Carlo method for analyzing the light curves of gravitationally lensed quasars to simultaneously estimate time delays and quasar structure including their mutual uncertainties. We apply the method to HE1104-1805 and QJ0158
-4325, two doubly-imaged quasars with microlensing and intrinsic variability on comparable time scales. For HE1104-1805 the resulting time delay of (Delta t_AB) = t_A - t_B = 162.2 -5.9/+6.3 days and accretion disk size estimate of log(r_s/cm) = 15.7 -0.5/+0.4 at 0.2 micron in the rest frame are consistent with earlier estimates but suggest that existing methods for estimating time delays in the presence of microlensing underestimate the uncertainties. We are unable to measure a time delay for QJ0158-4325, but the accretion disk size is log(r_s/cm) = 14.9 +/- 0.3 at 0.3 micron in the rest frame.
We use the microlensing variability observed for nine gravitationally lensed quasars to show that the accretion disk size at 2500 Angstroms is related to the black hole mass by log(R_2500/cm) = (15.6+-0.2) + (0.54+-0.28)log(M_BH/10^9M_sun). This scal
ing is consistent with the expectation from thin disk theory (R ~ M_BH^(2/3)), but it implies that black holes radiate with relatively low efficiency, log(eta) = -1.29+-0.44 + log(L/L_E) where eta=L/(Mdot c^2). These sizes are also larger, by a factor of ~3, than the size needed to produce the observed 0.8 micron quasar flux by thermal radiation from a thin disk with the same T ~ R^(-3/4) temperature profile. More sophisticated disk models are clearly required, particularly as our continuing observations improve the precision of the measurements and yield estimates of the scaling with wavelength and accretion rate.
We analyze V, I and H band HST images and two seasons of R-band monitoring data for the gravitationally lensed quasar SDSS0924+0219. We clearly see that image D is a point-source image of the quasar at the center of its host galaxy. We can easily tra
ck the host galaxy of the quasar close to image D because microlensing has provided a natural coronograph that suppresses the flux of the quasar image by roughly an order of magnitude. We observe low amplitude, uncorrelated variability between the four quasar images due to microlensing, but no correlated variations that could be used to measure a time delay. Monte Carlo models of the microlensing variability provide estimates of the mean stellar mass in the lens galaxy (0.02 Msun < M < 1.0 Msun), the accretion disk size (the disk temperature is 5 x 10^4 K at 3.0 x 10^14 cm < rs < 1.4 x 10^15 cm), and the black hole mass (2.0 x 10^7 Msun < MBH eta_{0.1}^{-1/2} (L/LE)^{1/2} < 3.3 x 10^8 Msun), all at 68% confidence. The black hole mass estimate based on microlensing is consistent with an estimate of MBH = 7.3 +- 2.4 x 10^7 Msun from the MgII emission line width. If we extrapolate the best-fitting light curve models into the future, we expect the the flux of images A and B to remain relatively stable and images C and D to brighten. In particular, we estimate that image D has a roughly 12% probability of brightening by a factor of two during the next year and a 45% probability of brightening by an order of magnitude over the next decade.
In searches for planetary transits in the field, well over half of the survey stars are typically giants or other stars that are too large to permit straightforward detection of planets. For all-sky searches of bright V<~11 stars, the fraction is ~90
%. We show that the great majority of these contaminants can be removed from the sample by analyzing their reduced proper motions (RPMs): giants have much lower RPMs than dwarfs of the same color. We use Hipparcos data to design a RPM selection function that eliminates most evolved stars, while rejecting only 9% of viable transit targets. Our method can be applied using existing or soon-to-be-released all-sky data to stars V<12.5 in the northern hemisphere and V<12 in the south. The method degrades at fainter magnitudes, but does so gracefully. For example, at V=14 it can still be used to eliminate giants redward of V-I~0.95, that is, the blue edge of the red giant clump.
