Prediction of the size of green synthesized silver nanoparticles using RSM-ANN-LM hybrid modeling approach

Abstract

Green synthesized Silver nanoparticles (AgNP) have attracted the most attention among others as it is an efficient antibacterial agent with low toxicity, environmentally beneficial and commercially viable. Since the attributes of AgNP are size-dependent, optimization of size becomes crucial. In this research, the size of AgNP is optimized and evaluated using an Artificial Neural Network (ANN) model. Response Surface Methodology (RSM) was also employed to conduct statistical analysis to determine the influence of the process parameters on the size of AgNP. coleus aromaticus plant's leaf extract was used to synthesize AgNP. The experiments were conducted based on four influential process parameters: Temperature, the volume ratio of the precursor to the extract, revolutions per minute (RPM) of the stirrer, and pH. The scrutinized experimental data were then used for training of ANN model using the Levenberg-Marquardt(LM) backpropagation algorithm for a more acceptable prediction of a minimum size of Ag NPs. The optimized ANN is a multilayer perceptron (MLP), which is a feed-forward (4–10–1) network: (i.e.) 4 nodes in the input layer, ten neurons in the hidden layers, and 1 node in the output layer. The ANN predictions capitulate R2 as 0.98536 and MSE in the range of 0.0371 for the given experimental data set. The regression analysis bespeaks that the ANN model can triumphantly predict the size of AgNP with a high degree of exactness.