Designed experiments are used to systematically examine the relationship between response variable and a set of experimental factors. In this study, 23 factorial design with center points was used to verify the selected experimental factors influence the removal efficiency and linearity assumption is not valid. Optimization experiments were then done to determine the best settings of the experimental factors and define the nature of curvature. Central Composite Design was used to fit a second-degree model since it was possible to build the previous 23 factorial design up into a central composite design by adding axial points. According to optimization results, the removal efficiency is maximized when concentration:110 mg l-1, dosage: 22 g l-1 and reaction time: 123 min. The predicted yield at these settings is 99.4623%.