Removal of arsenic from wastewater by co-precipitation with ferric ions and alum using experimental design approach

Abstract

Full Factorial Design (FFD) and Central Composite Design (CCD) with Response Surface Methodology (RSM) are applied to evaluate optimized conditions for As(III) and As(V) removal from natural water by co-precipitation with either ferric chloride or aluminum sulfate. The independent and interactive effects of factors including pH and co-precipitant concentration were investigated though the linear regression model obtained by Multiple Linear Regression (MLR) technique. For FFD, predicted removal efficiencies were calculated to plot the response surface. The optimized conditions at pH 7 and 225 mg/L of ferric ions were used to remove 90 mg/L of As(III), while alum sulfate was a poor co-precipitant for As(III) removal. In the case of As(V), 225 mg/L of ferric ions and 135 mg/L of aluminum ions were suitable co-precipitant doses to remove 90 mg/L of As(V) at pH 7. In addition, CCD was used to determine the main effects of pH, ferric ions and initial arsenic concentrations. Linear, quadratic and interaction parameters for the major factors were constructed in order to build the regression function, with coefficients calculated by MLR. The function was calibrated and validated using external experimental runs. The correlation coefficients (R2) of the actual vs. predicted arsenic removal percentages were 0.9871 and 0.9478 for As(III) and As(V), respectively. Analysis of variance (ANOVA) was used to determine the significance level of these parameters. All major factors were determined to be significant, with p-values <0.01. Multi-layer response surfaces were developed to determine the best conditions, having highest removal efficiency. The maximum removal efficiencies for arsenic species were approximately 100%, achieved by model prediction with a ferric ion concentration of 225 mg/L at pH 7. These optimized conditions were then applied to remove arsenic from two industrial wastewater samples, giving efficiencies of 93.98, and 91.48%. The results reveal that the chosen conditions from the RSM approach are applicable for arsenic removal from real water samples.