Chromatic Microlensing Time Delays

TL;DR

This paper proposes a method to detect wavelength-dependent microlensing time delays in strongly lensed AGN, enabling direct testing of quasar disk models by measuring time delay differences across different bands.

Contribution

It introduces a novel approach to verify microlensing-induced time delays and test the thin-disk quasar model through multi-band light curve analysis and simulations.

Findings

Time delay differences vary with wavelength due to disk size dependence.

With 2-day measurement precision, the effect can be verified in about 4 observations.

The standard thin-disk model can be rejected with approximately 35 measurements.

Abstract

Due to the finite size of the disk and the temperature fluctuations producing the variability, microlensing changes the actual time delays between images of strongly lensed AGN on the $\sim$day(s) light-crossing time scale of the emission region. This microlensing-induced time delay depends on the disk model, primarily the disk size $R_\mathrm{disk}$ which has been found to be larger than predicted by the thin-disk model. In this work, we propose that light curves measured in different bands will give different time delays since $R_\mathrm{disk}$ is a function of wavelength, and by measuring the time delay differences between bands, one can 1) directly verify such an new effect; 2) test the thin-disk model of quasars. For the second goal, our method can avoid the potential inconsistency between multi-band light curves that may bias the results by continuum reverberation mapping. We conduct a simulation based on a PG 1115+080-like lensed quasar, calculating the theoretical distributions of time delay differences between two bands: u and i centered around 354nm and 780nm, under and beyond the thin-disk model, respectively. Assuming the disk size is twice larger than the standard one, we find that with a precision of 2 days in the time delay difference measurements, the microlensing time delay effect can be verified with $\sim4$ measurements while with $\sim35$ measurements the standard model can be excluded. This approach could be realized in the ongoing and upcoming multi-band wide-field surveys with follow-up observations.