Quasiannealed Monte Carlo method for light transport in strongly heterogeneous media

TL;DR

This paper introduces a quasiannealed Monte Carlo method that efficiently models light transport in strongly heterogeneous media without explicit microstructure description, achieving accurate results with reduced computational effort.

Contribution

The paper presents a novel quasiannealed Monte Carlo approach that accounts for correlations in light transport, enabling efficient simulations of complex heterogeneous media without microstructure details.

Findings

Accurately predicts reflectance with ~1% error across various microstructures.

Reduces simulation time compared to traditional quenched disorder methods.

Applicable to media of arbitrary shape and size.

Abstract

Random-walk Monte Carlo simulations are widely used to predict the optical properties of complex, disordered materials. In presence of large heterogeneities (e.g., spatially-extended nonscattering regions in a turbid environment), an explicit description of the micro and macrostructures 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 so-called ``quenched'' disorder particularly time-consuming. We explore here the possibility to model 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 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.