Experimental and computational intelligent techniques for predicting the thickness-dependent optical behavior of PbI2 thin films

This study presents a comprehensive investigation of thermally evaporated lead iodide (PbI₂) thin films, focusing on the influence of film thickness (50–500 nm) on morphological, optical, and nonlinear properties. Scanning electron microscopy (SEM) and 3D surface topography reveal a distinct evolution in microstructure, transitioning from fine-grained uniformity in 50 nm films to coarse, aggregated crystallites in 300 nm films due to enhanced grain growth and stress accumulation. Optical analyses demonstrate tunable transmittance, reflectance, and absorption characteristics, with the direct bandgap decreasing from 2.89 eV (50 nm) to 2.50 eV (300 nm), attributed to reduced quantum confinement effects in thicker films. A robust artificial neural network (ANN) model is developed to predict key optical parameters—including refractive index (n), extinction coefficient (k), dielectric functions (ε₁, ε₂), and nonlinear susceptibility (χ⁽³⁾), with exceptional accuracy (MSE <10⁻⁷), validating its capability to simulate unmeasured thicknesses. Furthermore, nonlinear optical studies highlight PbI₂'s potential for photonic applications, with thickness-dependent variations in two-photon absorption (β_c) and skin depth. This work establishes a synergistic approach that combines experimental characterization and machine learning to optimize PbI₂ thin films for use in optoelectronic and nonlinear optical devices, providing a scalable framework for material design.