Prediction of plasmons in silver nanorods using artificial neural networks with back propagation algorithm

This work presents the prediction of plasmon positions in the optical absorption spectra of silver nanorods using artificial neural networks. Silver nanorods were prepared using a modified seed mediated strategy. At first, seed particles were synthesized by reducing silver ions with sodium borohydride in the presence of a stabilizer viz., cetyltrimethylammonium bromide (CTAB). In order to produce silver nanorods, the seeds were added to a growth solution containing a metal salt (AgNO3), a weak reducing agent (ascorbic acid) and a structure directing agent (CTAB). Four parameters viz., volume of metal seed, concentrations of silver nitrate, ascorbic acid and CTAB, which are found to be highly sensitive to plasmon characteristics are varied and grouped into 45 different sets of experiments. Herein, we report a model to predict the transverse and longitudinal plasmon band wavelengths, using a 4-input artificial neural network, having five neurons in the hidden layer and two neurons in the output layer. Levenberg-Marquardt back-propagation algorithm was used for the design. The network architecture has been trained with the 45 sets of experimental data collected and the regression plots with good correlation coefficient values were obtained for training, testing and validation stages.