The use of artificial neural network (ANN) for modeling optical properties of hydrothermally synthesized ZnO nanoparticles designed based on Doehlert method

摘要

In the present work, the influences of synthetic parameters on the optical properties of hydrothermally synthesized ZnO nanoparticles were investigated. Multivariate experimental design was applied to study the growth behavior and optical properties of obtained nanoparticles. Doehlert experimental design allowed determining the influence of three parameters (Synthesis temperature; synthesis period; and, initial concentration of precursors) on the different properties of the obtained nanoparticles; including: crystallite size obtained from Debby-Scherer calculation, exciton energy and band-gap energy obtained from optical absorption spectra of synthesized nanoparticles. Experimental data were fitted using artificial neural networks (ANNs). The reproduced experimental data from mathematical model shows a confidence within 90% and allows the simulation of the process for any value of parameters in the experimental range studied. Also, the saliency of the input variables was measured using the connection weights of the neural networks in which the relative relevance of each variable with respect to the others was estimated. The ANN results indicated that the exciton band edge which was observed in UV-Vis spectra of the obtained nanoparticles due to confinement effects, exciton energy increase by increasing the crystallite size while the band gap shows shrinkage.