Turbidity and COD removal from leather processing effluents using $TiO_2$ –assisted photocatalytic-ozonation by response surface methodology

Abstract

In the present study, concurrently removal of COD and turbidity from leather processing effluents (LPE) using $TiO_2$ –assisted photocatalytic-ozonation were investigated by utilization of Box-Behnken design (BBD) in planningexperiments. Effects of ozone dose (OD, mg $L^{-1}$), catalyst dose (CD, g $L^{-1}$), and aeration (A, mL $min^{-1}$) were performedas explanatory variables. An increase both in doses of ozone and catalyst and a decrease in aeration leaded increasesboth in removals of COD and turbidity. Values of 96.77% and 95.37% were obtained as the highest COD and turbidityremoval efficiencies, respectively. This showed that $TiO_2$ -assisted photocatalytic-ozonation process was significantlyeffective for the treatment of LPE. By using BBD, 2.95 g $L^{-1}$ of CD, 19.99 mg $L^{-1}$ of OD, and 1.63 mL $min^{-1}$ of A weredetermined as BBD-optimized operating conditions. BBD suggested removals of 96.77% and 94.93% for COD andturbidity, respectively at these optimized conditions. Validation experiments at BBD-optimized conditions wereresulted as 95.52%±1.28 and 94.36%±2.52 for COD removal and turbidity removal, respectively. Good agreementbetween predicted values and experimental results demonstrated the accuracy of BBD in optimization of explanatoryvariables of $TiO_2$ -assisted photocatalytic-ozonation process. Finally, multiple non-linear regression (MNLR) studieswere performed to state the variation in responses and also to predict the responses. The proposed models predictedCOD and turbidity removals with regression coefficients of 99.99% and 99.97%, respectively. These findings alsoshowed that MNLR was an efficient way to model and to predict the response variables of photocatalytic-ozonationprocess.