Quasiannealed Monte Carlo method for light transport in strongly heterogeneous media.

Abstract

Random-walk Monte Carlo simulations are widely used to predict the optical properties of complex, disordered materials. In the presence of large heterogeneities (e.g., spatially extended nonscattering regions in a turbid environment), an explicit description of the microstructure and the macrostructure and of the light propagation therein is generally required, in addition to a statistical average over a representative set of microstructures, thereby making simulations in the so-called "quenched" disorder particularly time consuming. Here, we explore the possibility of modeling light transport in finite-size strongly heterogeneous media without an explicit description of the underlying microstructure but from the knowledge of typical random-walk trajectories in infinite-size media that take correlations between successive interaction events into account. Simulations may thus be performed for media of any macroscopic shape and size more efficiently. We illustrate this approach, coined "quasiannealed," with the case of a two-phase emulsion consisting of transparent spherical droplets dispersed in a turbid medium. Good agreement with predictions from simulations in quenched disorder on the reflectance of a finite-thickness slab is found for a large set of microstructure properties and thicknesses with typical errors on the reflectance on the order of a percent.