Snapshot of a collection of scattered trajectories inside a turbid medium upon gaussian beam illumination 

Light transport in multiple scattering materials is notoriously difficult to describe, with most computational methods relying on heavy simplifications, periodic boundary conditions, or statistical methods.

In our group, we push the boundaries of computational methods by developing highly parallelized implementations of software solving light transport problems. In collaboration with Dr. Giacomo Mazzamuto (CNR-INO), we have realized an implementation of the Monte Carlo method to solve the Radiative Transfer Equation which is particularly suited for the study of multiple scattering in bounded media and thin films, where light propagation occurs at ultra-short scales, requiring a huge statistics to collect enough signal. Large simulations statistics comprising up to 10¹⁴ trajectories have been demonstrated, showing excellent agreement with our experimental time-domain results and highlighting exotic transport regimes arising in these confined geometries.

In certain cases, however, the vector wave nature of light and near-field effects arising from the microscopic distribution of scatterers must be taken into account to understand the interplay between light and disordered media with tailored spatial correlations. Also in this case, by leveraging on the massively parallel computing capabilities of modern GPU hardware, we contributed to the development of open-source implementations of the superposition T-matrix method in collaboration with Dr. Amos Egel (now at Hembach Photonik GmbH). Using these software tools, we were able to run rigorous light scattering calculations for aggregates of up to 3 million nanoparticles using an iterative solver on a regular consumer GPU card.

In addition to Monte Carlo and superposition T-matrix software, the group has expertise in the theory of radiative transfer and solution of the diffusion equation, as well as on the simulation of photonic nanostructures and metasurfaces using the Finite-difference time-domain (FDTD), Finite-difference Frequency Domain (FDFD) or Rigorous Coupled Wave Analysis (RCWA) methods.