An alternative parameterisation for binary-lens caustic-crossing events

TL;DR

This paper introduces a new parameterisation method for binary-lens microlensing events that improves the robustness and efficiency of fitting models to observed caustic-crossing light curves, reducing false minima issues.

Contribution

The authors propose an alternative parameterisation tailored to caustic-crossing features, enhancing the stability and accuracy of binary lens light curve modeling.

Findings

Improved robustness in binary lens light curve fitting.

Reduced occurrence of false minima during model optimization.

Extension of the method using the straight line fold caustic approximation.

Abstract

Microlensing events are being discovered and alerted by the two survey teams OGLE and MOA at an increasing rate. Around ten percent of these events involve binary lenses. Such events potentially contain much information on the physical properties of the observed binary systems, which can then be used for e.g. statistical studies on binary objects in the Galactic disk or bulge. However, such events are usually not straightforward to study, because the model equations are strongly non-linear and there are many local minima that can fool the search for the best solution if the parameter space is not inspected with great care. In this work an alternative parameterisation for the binary lens fitting problem is proposed, in which the parameters involved are defined to represent as closely as possible the caustic-crossing features observed in most binary lens light curves. Furthermore, we work out an extension of the method in order to make use of the straight line fold caustic approximation, when the latter applies for both the caustic entry and exit. We introduce an alternative parameterisation in order to confine the exploration of the parameter space to regions where the models only involve caustic crossing at the dates seen in the light curve. We find that the proposed parameterisation provides more robustness to the light curve fitting process, in particular in avoiding a code to get stuck in false minima.