We present new mid-IR observations of the quadruply lensed quasar Q2237+0305 taken with CanariCam on the Gran Telescopio Canarias .
Mid-IR emission by hot dust , unlike the optical and near-IR emission from the accretion disk , is unaffected by the interstellar medium ( extinction/scattering ) or stellar microlensing .
We compare these “ true ” ratios to the ( stellar ) microlensed flux ratios observed in the optical/near-IR to constrain the structure of the quasar accretion disk .
We find a half-light radius of R _ { 1 / 2 } = 3.4 _ { -2.1 } ^ { +5.3 } \sqrt { \langle M \rangle / 0.3 { M _ { \sun } } } light-days at \lambda _ { rest } = 1736 Å , and an exponent for the temperature profile R \propto \lambda ^ { p } of p = 0.79 \pm 0.55 , where p = 4 / 3 for a standard thin-disk model .
If we assume that the differences in the mid-IR flux ratios measured over the years are due to microlensing variability , we find a lower limit for the size of the mid-IR-emitting region of R _ { 1 / 2 } \gtrsim 200 \sqrt { \langle M \rangle / 0.3 { M _ { \sun } } } light-days .
We also test for the presence of substructure/satellites by comparing the observed mid-IR flux ratios with those predicted from smooth lens models .
We can explain the differences if the surface density fraction in satellites near the lensed images is \alpha = 0.033 _ { -0.019 } ^ { +0.046 } for a singular isothermal ellipsoid plus external shear mass model or \alpha = 0.013 _ { -0.008 } ^ { +0.019 } for a mass model combining ellipsoidal NFW and de Vaucouleurs profiles in an external shear .