2009ApJ...697..610M

We present mid-infrared imaging at 11.7 µm for the quadruple lens systems, MG0414+0534 and Q2237+030, using the cooled mid-infrared camera and spectrometer attached on the Subaru telescope. MG0414+0534 is characterized by a bright pair of lensed images (A1, A2) and their optical flux ratio A2/A1 deviates significantly from the prediction of a smooth-lens model. Q2237+030 is "the Einstein Cross" being comprised of four lensed images, which are significantly affected by microlensing in a foreground lensing galaxy. Our mid-infrared observations of these lensed images have revealed that the mid-infrared flux ratio for A2/A1 of MG0414+0534 is nearly unity (0.90±0.04). We find that this flux ratio is systematically small, at 4-5σ level, compared with the prediction of a best smooth-lens model (1.09) represented by a singular isothermal ellipsoid and external shear. The smooth-lens model, which also considers the additional lensing effect of the possible faint satellite, object X, still provides a large flux ratio of A2/A1=1.06, thereby suggesting the presence of more substructures that can explain our observational result. In contrast, for Q2237+030, our high signal-to-noise observation indicates that the mid-infrared flux ratios between all the four images of Q2237+030 are virtually consistent with the prediction of a smooth-lens model. Based on the size estimate of the dust torus surrounding the nuclei of these QSOs, we set limits on the mass of a substructure in these lens systems, which can cause anomalies in the flux ratios. For MG0414+0534, since the required mass of a substructure inside its Einstein radius is ≳360 M_☉, millilensing by a cold dark matter substructure is most likely. If it is modeled as a singular isothermal sphere, the mass inside a radius of 100 pc is given as ≳1.0x10⁵ M_☉. For Q2237+030, there is no significant evidence of millilensing, so the reported anomalous flux ratios in shorter wavelengths are entirely caused due to microlensing by stars.