Coagulation of Effluent Organic Matter for Water Reuse: Demonstrating and Predicting Carbon Removal

Carbon-based advanced treatment (CBAT) is an important water reuse strategy for inland communities. Since CBAT efficiency is negatively impacted by effluent organic matter (EfOM), coagulation pretreatment is expected to improve downstream processes. Jar tests on 26 wastewater effluents demonstrated the systematic effectiveness of alum coagulation, removing 10–35% of dissolved organic carbon (DOC) at 80 mg/L alum, which is expected to significantly extend the service life of activated carbon in CBAT. Applying a common surface water (SW) model, DOC removal for EfOM is systematically overpredicted. Compared to SW, wastewater has a higher DOC concentration, higher alkalinity, lower specific ultraviolet absorbance (SUVA254), and overall lower variability in these parameters between samples. Importantly, SUVA254 is a poor predictor of coagulation efficiency potentially due to its low variability between wastewaters relative to measurement error, or SUVA254 is not an effective compositional surrogate for wastewater treatability. A new set of models was developed for EfOM by reassessing underlying assumptions. General models using only three inputs (raw water DOC concentration, coagulant dose, and coagulation pH) had low error similar to SW models. If calibrated with at least one coagulation observation, site-specific models further decreased the error. Multiple isotherm models (i.e., Langmuir, Freundlich, and linear) predicted DOC removal with similar accuracy. An error propagation simulation demonstrated that unavoidable random error in real data sets limits the interpretation of fitted isotherms mechanistically. A complementary model predicts absorbance removal at 254 nm (UV254). This work demonstrates that heuristics for SW treatability should not be applied to wastewater without validation. Overall, DOC and UV254 models will enable users to assess the value of coagulation for effective CBAT designs.