It is of great importance to quantitatively characterize feed fouling potential for the effective and efficient prevention and control of reverse osmosis membrane fouling. A gradient filtration method with microfiltration (MF 0.45 μm) → ultrafiltration (UF 100 kDa) → nanofiltration (NF 300 Da) was proposed to extract the cake layer fouling index, I, of different feed foulants in this study. MF, UF, and NF showed high rejection of model suspended solids (kaolin), colloids (sodium alginate and bovine serum albumin), and dissolved organic matters (humic acid) during constant-pressure individual filtration tests, where the cake layer was the dominant fouling mechanism, with I showing a good linear positive correlation with the foulant concentration. MF → UF → NF gradient filtration tests of synthetic wastewater (i.e., model mixture) showed that combined models were more effective than single models to analyze membrane fouling mechanisms. For each membrane of gradient filtration, I showed a positive correlation with the targeted foulant concentration. Therefore, a quantitative assessment method based on MF → UF → NF gradient filtration, the correlation of combined fouling models, and the calculation of I would be useful for characterizing the fouling potentials of different foulants. This method was further successfully applied for characterizing the fouling potential of real wastewater (i.e., sludge supernatant from a membrane bioreactor treating dyeing and finishing wastewater).
Keywords: cake layer fouling index; fouling potential; gradient filtration; model foulants; real wastewater; reverse osmosis; single and combined membrane fouling models.