Quantitative Analysis And Systematic Elucidation Of The Interfacial Interactions Involved In Microfiltration Membrane Fouling By Humic Acid

Membrane fouling is the major factor inhibiting the development of microfiltration (MF) technology. Although MF membrane fouling is affected by multiple factors and appeared to be an intricate process, the essence of membrane fouling lies in the interfacial interactions between foulant and membrane surface. In this study, the xDLVO approach was utilized to quantitatively analyze the interfacial interactions involved in MF membrane fouling by humic acid (HA) at various pHs. The contribution of different interactions (i.e. van de Waals interaction, double layer electronic interaction and acid-base interaction) towards fouling was assessed and the main mechanism through which pH influenced HA fouling of MF membrane was examined. In order to obtain a more comprehensive and systematic evaluation of the validity of the xDLVO approach in describing MF membrane fouling by HA, both qualitative and quantitative studies were conducted on correlations between fouling potential and interfacial interaction energy UTOTmif (or UTOTflf) at various fouling conditions.It was found that acid-base (AB) interaction, which was a short-range interfacial interaction, contributed the most to MF membrane fouling by HA and turned out to be the major factor through which pH influenced fouling. As pH decreased, MF membrane-HA and HA-HA interfacial interaction energies reduced. At the same pH, hydrophobic membrane-HA interaction energy was smaller than HA-HA interaction energy and HA-HA interaction energy was smaller than hydrophilic membrane-HA interaction energy.For both the hydrophobic and hydrophilic membranes, permeate flux during the entire filtration process decreased with the decreasing pH. Physical cleaning efficiency also decreased at lower pH, whereas chemical cleaning efficiency increased at lower pH. In initial filtration stage, permeate flux of hydrophobic membrane with smaller interaction energy was lower than that of hydrophilic membrane; however, in final filtration stage where fouling behaviors of both hydrophobic and hydrophilic membranes were controlled by the same HA-HA interfacial interaction energy, permeate flux of hydrophobic membrane was still smaller than that of hydrophilic membrane. For hydrophobic membrane, flux reduction in initial filtration stage was greater than that in final filtration stage because of the smaller membrane-HA interaction energy compared with HA-HA interaction energy; for hydrophilic membrane, flux reduction in initial filtration stage was also greater than that in final filtration stage despite the greater membrane-HA interaction energy in initial filtration stage. Therefore, the xDLVO approach could qualitatively predict MF membrane fouling by HA at various pHs but a consideration of other factors was also needed when the xDLVO approach was used to predict fouling behaviors of different membranes and different filtration stages.For both hydrophobic and hydrophilic membranes, fouling potentials in initial and final filtration stages correlated well with membrane-HA interaction energy (UTOTmlf) and HA-HA interaction energy (UTOTflf), respectively. This suggests that the xDLVO approach could successfully describe MF membrane fouling by HA at different pHs. Slope of linear regression between fouling potential and interaction energy was higher in initial filtration stage compared to that in final filtration stage. For hydrophobic membrane, slope of linear regression was higher than that of hydrophilic membrane. Nevertheless, discrepancy of slopes for different membranes was remarkably smaller than that for different filtration stages. The different slopes of linear regression indicate that fouling potential in initial HA microfiltration stage is more susceptible to interaction energy and hydrophobic membrane fouling potential is more susceptible to interaction energy compared to hydrophilic membrane fouling potential. Since fouling potential was more susceptible to interaction energy when fouling occurred inside the membrane pores and the permeate flux was great, the inside membrane fouling and the greater permeate flux in initial filtration stage accounted for the higher slope of linear regression in this stage.