Adaptive learning of prediction and simulation on the influence of Fe thickness on the energy gap of ZnO/Fe/ZnO tri-layer thin films

This work focuses on the investigation of the effect of Iron (Fe) interlayer thickness, on the experimental, simulation and prediction of the optical energy gap of ZnO/Fe/ZnO tri-layer thin films. The optical properties and the modeling were studied by varying the Fe interlayer thickness from 20 to 80 nm. The tri-layer thin films were successfully fabricated using atomic layer deposition for the ZnO layers and direct current magnetron sputtering unit for Fe layers. The optical transmission in the range of 400–2500 nm has been studied to determine the variation of absorption coefficient (α) and optical gap \(({E}_{\text{g}})\) of the investigated thin films. The appearance of ZnO/Fe/ZnO system as a Fabry–Pérot “interference filter” is consistent with the optical transmission curves. Adaptive neuro-fuzzy inference system (ANFIS) was utilized for simulation and prediction based on experimental data. Using ANFIS enabled the prediction of the transmittance of different Fe interlayer thicknesses for the unmeasured values. The E_g, showed a noticeable dependence on the Fe layer thickness, where a significant reduction decreasing from 2.45 to 1.75 eV was observed, which has been attributed to the increase of the metallicity, while the metallization criterion on the basis of E_g showed a decreasing trend from 0.35 to 0.312, which suggests potential applications for nonlinear optics. The evaluating mean squared error of the ANFIS model was indicating the positive influence of the used model. Therefore, modeling approach has computational efficiency and flexibility that provides a rapid and reliable technique to investigate the impact of Fe thickness on the energy gap of ZnO/Fe/ZnO tri-layer thin films.

Graphical abstract