A modeling study by artificial neural network on process parameter optimization for silver nanoparticle production

• Main Authors: Chowdhury, Silvia, Yusof, Faridah, Sulaiman, Nadzril, Sidek, Shahrul Na'im, Faruck, Mohammad Omer
• Format: Article
• Language: English; English
• Published: Asian Research Publishing Network (ARPN) 2016
• Subjects: QD Chemistry; T Technology (General)
• Online Access: http://irep.iium.edu.my/52645/1/jeas_1016_5222.pdf; http://irep.iium.edu.my/52645/7/52645_A%20modeling%20study%20by%20artificial%20neural%20network_SCOPUS.pdf; http://irep.iium.edu.my/52645/; http://www.arpnjournals.org/jeas/research_papers/rp_2016/jeas_1016_5222.pd

Artificial neural network (ANN) is the most accepted method for non-parametric modelling and process optimization of chemical engineering. The paper focuses on using ANN to analyse the yield production rate of silver nanoparticles (AgNPs). The study examines the effect of AgNO3 concentration, stirring time and tri-sodium citrate concentration on the production of AgNPs yield. The yield of AgNPs was modelled and optimized as a function of three independent variables. Furthermore, assessment of the model through the coefficient of determination (R2 = 0.9778) and mean square error (MSE) showed that the optimized production conditions were found at 1mM AgNO3 concentration,15 min of stirring time and 1% tri-sodium citrate. Optimal and maximal AgNPs production were 20.62 (Area*) of yield experimentally, which was calculated using area under the curve from UV-vis analysis in the wave length range of 350 nm to 420 nm. Meanwhile, under the same conditions, the ANN predicted value is 19.84 (Area*) of AgNPs yield with 3.95% error. Besides that, the ANN model was employed to construct an output surface plot to reveal the impact of input variable as well as figure out the interaction effect and clear representation of optimized condition. Synthesized AgNPs at optimized condition (absorbance 0.93AU at 420 nm wavelength) were then characterized using Field Emission Scanning Electron Microscopy (FESEM) and UV-vis analysis.