Comparison of PCA-based neural network models using the screening of X-ray diffraction data for MOMBE-grown HfO2 thin film characteristics

In this paper, the principal component analysis based neural network process models of the HfO2 thin films are investigated. The input process parameters are extracted by analyzing the process conditions and the accumulation capacitance and the hysteresis index are extracted to be the main responses to examine the characteristics of the HfO2 dielectric films. Here, the screened X-ray diffraction data are used to analyze the characteristic variation for the different process conditions and predict the crystallinity-based the response models for the electrical characteristics. For the data screening, principal component analysis was carried out to reduce the dimension of two types of XRD data that are compressed into a small number of principal components. The compressed data are trained using the neural networks. The results show that the physical or material properties can be predicted by the models using the large dimension of the data.