Ensemble Learning for Modeling Luminance in Plasmonic Nanoellipse for Tunable Spaser

in a recent study we provided a theoretical framework for plasmonic coupled elliptical nanostructure capable of producing high-intensity spasers or nanolasers. The nanostructure which consists of a gold film substrate, underwent testing for various light polarizations where we selected a simple elliptical shape and then modified it to explore different geometries. The structure exhibited strong coupled resonance modes, particularly localized surface plasmon (LSP). Through finite-element method (FEM) simulations, we investigated the localized surface plasmon resonance (LSPR) of the system. Results revealed that introducing symmetry-breaking techniques, such as creating a symmetry-broken nano-ellipse (SBE), significantly increased both LSPR peaks and local field intensity (LFI). By selecting appropriate gain media, the nanostructure could support various plasmonic resonance modes across different wavelengths. Moreover, it demonstrated potential for developing a compact nanolaser with a broad tunability range. Adjusting light polarization and making structural modifications led to the observation of Surface Plasmon Amplification by Stimulated Emission of Radiation (SPASER) phenomena. The proposed model, with its substantial LSPR values, shows promise for applications like surface-enhanced Raman spectroscopy and sensing. The application of the data-driven methods in modeling the above phenomena has not been explored yet. Due to the complex nature of the phenomena, it is expected that the data-driven models can provide promising results. Consequently, in this article the primary results of proposing ensemble machine learning in modeling the above phenomenon are presented. In this work random forest (RF) and support vector machine (SVM) were employed for modeling and analysis of outcomes.