Artificial Neural Network Modeling of Fungus-Mediated Extracellular Biosynthesis of Zirconium Nanoparticles Using Standard Penicillium spp.

Different growth stages of Penicillium notatum, P. purpurogenum and P. aculeatum were evaluated for their potential for zirconium nanoparticle (Zr-NP) biosynthesis by total protein content assay. Fungal secreted protein concentrations for all three species were higher at deceleration phase. For optimizing biosynthesis conditions with respect to Zr-NP size, different pH values and ZrCl4 concentrations were tested and analyzed. The artificial neural network (ANN) modeling was then applied to find the best conditions. A feed-forward ANN model was used to make a connection between the output as the size of Zr-NPs, with three input variables, including pH, ZrCl4 concentration and fungal species. An optimal multi-layer perceptron neural network was finally designed to model the experimental data for which the correlation coefficients R2 at each phase (training, validation and test sets) were calculated to be 0.9946, 0.9952 and 0.9997, respectively. The accuracy of the model was confirmed through validation experiments.