Prediction of RSM and ANN in the decolorization of Reactive Orange 16 using biochar derived from Ulva lactuca
M. Kumar, S. Sujatha, R. Gokulan, Arun Vijayakumar, S. Praveen, S. Elayaraja
Abstract
ABSTRACT The present research compares the prediction of the response surface methodology (RSM) and artificial neural network (ANN) on the decolorization of Reactive Orange 16 (RO16) using a novel adsorbent produced from Ulva lactuca (seaweed). These mathematical models were designed based on four process conditions biochar dose, pH, temperature, and initial concentration. The experimental trials concluded that the dye removal of 93.10% was achieved at an optimum biochar dosage of 2 g/L, pH of 2, initial concentration of 0.5 mmol/L, and temperature of 40°C. The biochar characterization confirmed the presence of functional groups that are responsible for the adsorption of dye. The mathematical predictive model of RSM and ANN was compared with the experimental trials and a correlation coefficient (R 2 ) of 0.95 is obtained for RSM, whereas a correlation coefficient (R 2 ) of 0.99 was obtained for ANN. ANN prediction model was far better than RSM in the prediction of decolorization of Reactive Orange 16 (RO16) using U. lactuca as a novel adsorbent. The adsorption isotherm studies concluded that four parameter model Fritz–Schlunder – IV and Marczewski– Jaroniec were found to best fit with a correlation coefficient of 0.9999. Pseudo-second-order kinetic model was found to best fit the experimental data.