2007MNRAS.381.1591B

Several gravitationally lensed quasars are observed with anomalous magnifications in pairs of images that straddle a critical curve. Simple theoretical arguments suggest that the magnification of these images should be approximately equivalent, whereas one image is observed to be significantly demagnified. Microlensing provides a possible explanation for this discrepancy. There are two key parameters when modelling this effect. The first, the fraction of smooth matter in the lens at the image positions, has been explored by Schechter & Wambsganss. They have shown that the anomalous flux ratio observed in the lensed quasar MG 0414+0534 is a priori a factor of 5 more likely if the assumed smooth matter content in the lens model is increased from 0 to 93 per cent. The second parameter, the size of the emission region, is explored in this paper, and shown to be more significant. We find that the broadening of the magnification probability distributions due to smooth matter content is washed out for source sizes that are predicted by standard models for quasars. We apply our model to the anomalous lensed quasar MG 0414+0534, and find a 95 per cent upper limit of 2.62x10¹⁶h^–1/2₇₀(M/M_☉)^1/2 cm on the radius of the I-band emission region. The smooth matter percentage in the lens is unconstrained.