2018MNRAS.473...80T

The time delays of gravitationally lensed quasars are generally believed to be unique numbers whose measurement is limited only by the quality of the light curves and the models for the contaminating contribution of gravitational microlensing to the light curves. This belief is incorrect - gravitational microlensing also produces changes in the actual time delays on the ∼day(s) light-crossing time-scale of the emission region. This is due to a combination of the inclination of the disc relative to the line of sight and the differential magnification of the temperature fluctuations producing the variability. We demonstrate this both mathematically and with direct calculations using microlensing magnification patterns. Measuring these delay fluctuations can provide a physical scale for microlensing observations, removing the need for priors on either the microlens masses or the component velocities. That time delays in lensed quasars are themselves time variable likely explains why repeated delay measurements of individual lensed quasars appear to vary by more than their estimated uncertainties. This effect is also a new important systematic problem for attempts to use time delays in lensed quasars for cosmology or to detect substructures (satellites) in lens galaxies.