Catalytic Membrane based Advanced Oxidation Processes for Micropollutant Removal
Tao Wang is a PhD student in the department Membrane Science & Technology. (Co-)Supervisors are prof.dr.ir. W.M. de Vos and dr.ir. J. de Grooth from the faculty of Science & Technology.
In this thesis, catalytic membranes, which can successfully integrate membrane separation and sulfate radicals-based advanced oxidation processes (SR-AOPs) are developed. These membranes are then immediately applied to treat water streams containing organic micropollutants (MPs). Given the wide spread of MPs in our surface waters and their toxicities, MPs pose a significant threat, not only to aquatic organisms but also to human beings. Since wastewater treatment plants only provide limited removal to MPs, different technologies, including adsorption, membrane separation, and AOPs have been applied to remove MPs from wastewater. However, the diverse physicochemical properties of MPs make them very challenging when just a single technology is used. The integration of membrane separation and AOPs has the potential to provide sufficient removal efficiently. Indeed, the membrane is very suitable as a substrate for required small-sized catalysts, avoiding the difficulties of reusing and recycling them. Meanwhile, through activating the oxidants by the catalysts embedded within the membrane structure, the reactive species generated can degrade MPs when the solution passes through the membrane structure. In this case, not only the MP concentrations in the permeate are reduced but also the rejected MPs can be degraded, preventing the formation of the highly concentrated retentate. This thesis not only focuses on the fabrication of SR-AOPs-based catalytic membranes but also studies the influence of operational parameters on MPs removal.
In Chapter 2, catalytic support membranes were developed by blending CoFe2O4 nanoparticles with polyethersulfone (PES), leading to CoFe2O4 catalysts that were embedded homogeneously within the polymeric membrane structure. Batch experiments show that 70% naproxen can be degraded in the presence of peroxymonosulfate (PMS) and membranes with 2.0% CoFe2O4 in the casting solution, demonstrating that the CoFe2O4 catalysts embedded within the polymeric membranes can activate PMS to generate reactive species. Since the traditional polyamide selective layer cannot withstand oxidation from hydroxyl radicals, strong polyelectrolytes, poly (diallyldimethylammonium chloride) (PDADMAC) and poly(styrene sulfonate) (PSS) were used to build selective polyelectrolyte multilayers on top of the catalytic support. To explore the effect of residence time, the catalytic ultrafiltration (UF) and nanofiltration (NF) membranes were measured under various fluxes in both dead-end and cross-flow set-ups. The significant role of residence time is revealed, where the removal of MPs increases with a prolonged residence time in both UF and NF processes. Moreover, the MPs removal of NF and UF catalytic membranes are comparable under the same flux (residence time). However, since the pressure required to reach a certain flux is much lower for UF membranes than for NF membranes, the application of UF catalytic membranes for MPs removal is more energy friendly than the application of catalytic NF membranes.
It needs to be mentioned that before the addition of PMS, 24 hours recycling of naproxen is always performed in all the measurements in Chapter 2 to make sure that the adsorption and desorption of MPs on the membranes reach equilibrium under the set conditions. However, it was found that due to the acidity of PMS, the pH of the naproxen solution decreased once PMS was added, resulting in increased adsorption of naproxen on the membrane structure. To rule out the effect of this decreased pH on naproxen adsorption after adding PMS and to systematically explore the effects of pH on the degradation efficiency of MPs, two ways of controlling the pH of the naproxen solution were compared in Chapter 3: adjusting the pH before and after the addition of PMS. Surprisingly, when the pH of the naproxen solution was adjusted after adding PMS, the kinetic constant of the naproxen oxidation (0.42/min, pH 6) in the batch experiment is 23 times higher than the kinetic constant obtained when pH was adjusted before adding PMS (0.018/min, pH 6). This enhancement reveals that the increased naproxen adsorption induced by the addition of PMS inhibits the degradation of MPs, the effects of which are clearly significant but normally they are overlooked in catalytic polymeric membranes. Moreover, when the pH was readjusted to 8 after adding PMS, the highest degradation kinetic constant was obtained (0.54/min). At this pH, the UF catalytic membranes with 2.0 % of CoFe2O4 exhibited 98 %, 97 %, and 74 % removal of naproxen, bisphenol A, and atrazine, respectively in a dead-end experiment. The inhibition caused by the adsorption of naproxen or its byproducts was also observed in a reusability study, where the naproxen removal of the catalytic membranes with 2.0 % of CoFe2O4 decreased by 81 % and 45 % after 5 rounds in the batch experiment and dead-end cell, respectively. However, chemical cleaning between each round is shown to be an effective method to mitigate the negative effects of the adsorption of MPs or their intermediates during the SR-AOPs.
To effectively degrade MPs in the feed solution and avoid the formation of a concentrated retentate, a different membrane orientation was explored in Chapter 4. In this approach, the catalytic support faces the feed solution. In this case, the rejection of MPs unavoidably induces severe internal concentration polarization within the porous support. The effect of concentration polarization under different membrane orientations was explored with NF catalytic membranes in Chapter 4. On top of the catalytic support membrane used in Chapter 2 and 3, three types of polyelectrolyte selective layers with different rejection behaviors were produced by dip-coating. Since the concentrations of naproxen and PMS within the catalytic support strongly affect the degradation kinetics of SR-AOPs in the batch experiment, concentration polarization models were used to calculate the concentrations of naproxen and PMS within the porous catalytic support. The results show that a higher removal of naproxen can be obtained with a higher ratio of PMS to naproxen (). On this basis, the calculation of the concentrations of MPs and oxidants within the porous catalytic support can be used as an indicator of the degradation efficiency of MPs. Moreover, a decrease in naproxen concentration within the feed solution was observed when the catalytic support faces the feed solution, exhibiting the potential of eliminating the concentrated retentate. However, it is also found that less naproxen in the feed solution can be degraded when more multilayers were coated on the membrane surface. This can be attributed to the coating of the selective layer on top of the catalytic ultrafiltration membranes, which lowers the accessibility of PMS and naproxen to the catalysts embedded within the polymeric membranes. By coating only one side of the membrane, this negative effect caused by the polyelectrolyte coating can be mitigated. Overall, a 97 % removal of naproxen on the permeate side and a 12% degradation of naproxen on the feed side were observed with the one-side coated membranes under a catalysis-separation sequence.
As it was found in Chapter 4 that the coating of polyelectrolyte multilayers hinders the accessibility of the oxidants and MPs to the catalysts fixed within the polymeric support membranes, the possibility of introducing catalysts in or on top of the selective layer is explored in Chapter 5. When coating an outmost layer of poly (acrylic acid) (PAA) on the hollow fiber membranes, Fe2+ ions can be bonded with PAA on the membrane surface and then reduced into Fe0 by NaBH4. To investigate the effects of Fe2+ binding/reducing cycles, 1, 2, and 3 binding/reducing cycles were performed on membranes with 4.0 bilayers of PDADMAC and PAA. The results show that when 3 binding/reducing cycles were conducted, the Fe0 particles generated are enough to improve the naproxen removal. However, it was also found that the harsh synthesis conditions exhibited negative effects on the MP selectivity of the multilayer. The penetration of Fe2+ ions during the binding/reducing process, together with the increase in salinity and pH from NaBH4, makes the multilayer less dense. To improve the stability of the polyelectrolyte multilayer, the in situ synthesis of Fe0 was performed on top of asymmetric multilayers, which consist of PDADMAC/PAA multilayers on top of 7.0 bilayers of PDADMAC and PSS. It is observed that the above negative effects of the Fe0 on the membrane performances can be mitigated with the asymmetric multilayers. The obtained membranes with asymmetric polyelectrolyte multilayers exhibited an excellent naproxen treatment efficiency, which is over 80 % naproxen rejection on the permeate side and 25 % naproxen removal on the feed solution side (after 1 hour).
This work highlights the possibility of utilizing SR-AOPs-based catalytic membranes to treat MPs. Based on the position where the catalysts are introduced, different fabrication methods are demonstrated. Meanwhile, the significant roles of the residence time and pH in the SR-AOPs-based membrane processes are also revealed. To further push forward the application of the SR-AOPs-based membrane, more relevant work can be done in the future by exploring the effects of complex water matrices, upscaling with hollow fiber geometry, and developing new types of membranes. With the development of catalytic membranes and a good understanding of the AOPs-coupled membrane process, the combination of membrane separation and AOPs can significantly improve the quality of the wastewater, possibly helping to alleviate the global shortage of water resources.