An Artificial Intelligence‐Assisted Virtual Laboratory for Optoelectronics Education Using the Transfer Matrix Method

ABSTRACT

Visualizing sub‐wavelength optical phenomena, such as thin‐film interference and parasitic absorption, remains a significant pedagogical challenge in optoelectronics education. Furthermore, traditional physical laboratories require costly equipment, limiting student accessibility. To address these barriers, this study presents an interactive, MATLAB‐based virtual laboratory designed to teach solar cell optics using the transfer matrix method (TMM). The software enables students to construct standard multi‐layer architectures (e.g., ITO/PEDOT:PSS/Active Blend/TiO ₂ /Al) and immediately visualize critical optical outputs, including the electric field distribution | E ( z )| ² , spatial generation rate G ( z ), and macroscopic spectral reflectance. While traditional simulators provide mathematical outputs, they function as unguided platforms, leaving students to interpret dense graphical data alone. To bridge the gap between complex optical data and physical intuition, the defining novelty of this virtual laboratory is the integration of a multimodal, context‐aware Artificial Intelligence (AI) tutor powered by a Large Language Model (LLM). This embedded assistant dynamically processes both numerical parameters and rasterized visual plots to guide students through Socratic questioning and engineering design critiques. Enriched by 2D and 3D parameter sweeps and a comparative case manager to explore the short‐circuit current density ( J _sc ) optimization landscape, this platform merges rigorous computational physics with automated pedagogical guidance to cultivate practical design skills in next‐generation semiconductor engineers.