A Comprehensive Investigation of Chitosan/tripolyphosphate Nanoparticles Using Artificial Neural Networks

Abstract Chitosan/tripolyphosphate (CS/TPP) nanoparticles have been widely investigated in many applications. Many experimental studies evaluated effective parameters on CS/TPP nanoparticles without a comprehensive study to explain the influence of parameters on nanoparticles. The purpose of present work was to build a mathematical model capable of predicting the particle size and zeta potential of CS/TPP nanoparticles using seven significant factors, including CS and TPP concentrations, deacetylation degree of CS (DD), the molecular weight of CS, pH of CS solution and temperature, and their interactions on the size and zeta potential of CS/TPP nanoparticles. A model was built using artificial neural network including properties of nanoparticles as input with particle size and zeta potential as output. Artificial neural networks (ANN) models were used based on 8 experimental works consisting 160 data to estimate the variation tendency of size and zeta potential. The established model successfully predicted particle size and zeta potential of nanoparticles covering a range of 50-1000 nm. All parameters had significant effects on the size, the interaction between parameters changed the relationship pattern between them. In addition, results indicated that the main reason for the unexplained difference in previous works is the interactions between parameters. In addition, there is a relationship between size and zeta potential, which is due to the attractive and repulsive electrostatic charges, ionic interactions, CS chain length and viscosity. The ANN models in this work were valid for other papers.