Modeling of In2O3-10 wt% ZnO thin film properties for transparent conductive oxide using neural networks

Effects of deposition process parameters on the deposition rate and the electrical properties of In₂O₃–10 wt% ZnO (IZO) thin films were modeled and analyzed by using the error back-propagation neural networks (BPNN). Output models were represented by response surface plots and the fitness of models was estimated by calculating the root mean square error (RMSE). The deposition rate of IZO thin films is affected by the RF power and the substrate temperature. The electrical properties of the IZO thin films are mainly controlled by O₂ ratio and the substrate temperature. The predicted output characteristics by BPNN can sufficiently explain the mechanism of IZO deposition process. Thus, neural network models can provide the reliable explanation of IZO film deposition.