Elementary approach on the prediction of next material composition using AI technology

This study aims to identify the factors affecting the characteristics of samples, such as photoluminescence intensities, and identify the relationship between performance improvement and the search parameters for material composition. Subsequently, we optimize the experimental conditions to provide the maximum characteristic value. First, the process parameters are introduced as input values to the artificial intelligence (AI)-based model; then, we obtain a generalized equation to establish relationship between the characteristics of the samples and the process parameters. Subsequently, the new samples suitable for determining an accurate model and optimizing the process parameters are calculated and recommended to the user. Finally, the obtained formula is optimized, and the optimum values for achieving maximum characteristic are determined. Experimental validation using the AI program developed in this study found that the two components (x, y) that provide the strongest PL intensity in the Srx(La10?x?yEuy)(SiO4)6O3?x/2 (x=2?6, y=0.6?1.2) red-emitting phosphors can be easily estimated from approximately 10 initial data points.