Articles techniques et présentations

Hot carriers generated in plasmonic nanostructures are of particular interest in emerging applications such as photocatalysis, photodetection, nanoscale energy harvesting, and enhanced light–matter interactions [1,2]. Predicting where these carriers are generated, how they propagate, and whether they can reach and cross an active interface is particularly challenging because the problem combines electromagnetic field localization, microscopic carrier scattering and transport and energy- and momentum- dependent interfacial transmission [3-6]. Current theoretical approaches to hot-carrier modeling often rely on electronic-structure calculations or semiclassical Boltzmann transport methods [3,5,7]. While density functional theory and Boltzmann transport calculations can provide detailed microscopic insight, their direct application to experimentally relevant plasmonic nanostructures becomes computationally prohibitive when the relevant geometries span nanometer-scale features to dimensions of hundreds of nanometers or micrometers [7]. As a result, these methods are often difficult to use for rapid design iterations, geometry optimization, or direct support of experimental and industrial workflows. The method developed here addresses this gap by combining full-wave electromagnetic simulations with an efficient Monte Carlo transport backend, enabling fast evaluation of hot-carrier generation and extraction in realistic three-dimensional structures. In this study, we develop a COMSOL Multiphysics–coupled Monte Carlo framework to model hot-carrier generation and transport in metallic nanostructures. The electromagnetic response of the plasmonic structure of interest is first simulated by solving the frequency-domain electromagnetic wave problem in COMSOL Multiphysics. From the simulated optical fields, we extract spatially resolved quantities such as the local electric-field intensity. The full field solution is then incorporated into channel-specific generation probability distributions for phonon-assisted processes, electron–electron scattering, and surface-assisted Landau damping [3,4,8]. The COMSOL model is coupled to a MATLAB-based Monte Carlo code through LiveLink for MATLAB. The backend imports geometry information, mesh data, tetrahedral element volumes, field values at mesh-element centroids, and triangulated boundary surfaces. Generation events are sampled from optical absorption distributions. The subsequent carrier propagation accounts for multiple scattering events and carrier impacts at metal boundaries. The workflow produces generation-specific extraction probabilities, barrier-dependent carrier yields, surface impact maps, and average impact-energy distributions. This approach provides a practical way to identify which electromagnetic hot spots are relevant for hot-carrier observables. In particular, it distinguishes regions of strong optical absorption from regions where hot carriers can be efficiently extracted [7,9]. The framework also enables direct comparison between competing generation mechanisms, including bulk absorption-mediated processes and surface-assisted Landau damping. The current workflow is designed to be reusable within COMSOL Application Builder. Parameters such as geometry, photon energy, Schottky barrier height, carrier-generation settings, and Monte Carlo sampling size can be modified directly from the COMSOL graphical user interface. This method is particularly well suited as a design-support tool for experimental groups and industrial users seeking to optimize plasmonic nanostructures for applications in photocatalysis, photodetection, energy harvesting, and nanoscale light–matter interactions [1,2,10,11].