A comprehensive methodology for analysis of coagulation performance: Dosing approach, isotherm modelling, FTIR spectroscopy and floc characterization

Isolated studies for the evaluation of coagulation considering either microscopic or macroscopic dimensions have largely been discussed. However, comprehensive analysis of coagulation including various dimensions is imperative to design an optimal coagulation system. In this work, a comprehensive methodology has been described which includes dosing approach, mathematical modelling, chemical analysis by FTIR (Fourier Transform Infrared) spectroscopy, image analysis for floc properties and various analytical procedures. Jar tests were performed on simulated waters containing humic acid and kaolin using high basicity PACl (Poly Aluminium Chloride). A PODR (point of diminishing return) approach is used for the determination of coagulant dose corresponding to TOC (Total Organic Carbon), DOC (Dissolved Organic Carbon) and UV254 (Ultraviolet Absorbance at 254 nm). On average, PODR for UV254 was slightly greater than that of TOC followed by that of DOC. The correlation between the input TOC and PODR values was in the order: TOC (R2 = 0.84) > UV254 (R2 = 0.83) > DOC (R2 = 0.65). The isotherms are analyzed by non-linear error functions: RMSE (Root mean square error), HYBRID (Hybrid fractional error function), ARE (Average relative error), MPSD (Marquardt's percent standard deviation) and Chi-Square. It was observed that error functions RMSE, HYBRID and Chi-Square can be successfully used for non-linear analysis. However, ARE and MPSD cannot be used at low concentrations as they showed high residual values. Further, image processing is used to analyze floc size and fractal dimension. The interactions of species and the physical structure of flocs derived based on adsorption isotherms are corroborated by using the FTIR data.

Automated summary

Isolated studies for the evaluation of coagulation at different dimensions have been discussed. In this work, a comprehensive methodology was described, including dosing approach, mathematical modeling, chemical analysis, image analysis, and various analytical procedures. The results showed that RMSE, HYBRID, and Chi-Square could be successfully used for non-linear analysis, while ARE and MPSD could not be used at low concentrations. Additionally, image processing was used for analysis of floc size and fractal dimension.