Development of a polyzwitterionic hydrogel coating for RO membranes by concentration polarization-enhanced in situ click reaction which can be applied in modules
Although vast research efforts have been contributed to control fouling, it remains a problem in membrane filtration processes. While the most reports on surface modifications are focusing on flat-sheet membranes conducted in laboratory scales, only a few focalize the scale-up of the modification to membrane modules in industrial scale due to their complex configuration and the associated limited accessibility. A promising method overcoming this challenge, is the concentration polarization-enhanced in situ coating of membranes during filtration where the barrier properties of the membrane and the arising concentration polarization is exploited allowing the selective modification of membranes.
The main focus of this work was therefore to develop a polyzwitterionic anti-fouling hydrogel coating for RO TFC membranes by CP-enhanced in situ thiol-based click reaction. The gradual transfer of this modification method from dead-end to modules in cross-flow was to be demonstrated. This was done with the aim of integrating modified RO modules into a cooling circuit of the steel industry to purify and allow the reuse of cooling water, as in Germany approx. 77 % of water is used for product cooling. Therefore, a critical evaluation of the separation performance of established membranes was done first by analyzing fouling with real cooling water. The cross-linking of the zwitterionic copolymer with the cross-linker 1,4-Dithiothreitol (DTT) via thiol-ene Michael addition click reaction was then studied in the bulk via in situ rheology. In a further step, the in situ cross-linking reaction was examined during dead-end filtration. The modification was then transferred to flat-sheet membranes in cross-flow. In the last step, an up-scaling to spiral-wound RO modules was accomplished and were then implemented in a cooling circuit of the steel industry where pilot scale tests were done.
During the fouling tests (dead-end short-term and cross-flow long-term experiments), the membrane BW30 exhibited a very high fouling resistance and showed no irreversible fouling towards real cooling water from the steel industry. This membrane was therefore chosen for the membrane modification tests. By means of rheology, the influence of different parameters (polymer concentration, pH value, acrylate/crosslinker ratio) on the bulk hydrogel formation could be investigated. By this, optimal parameters for the composition of reaction solution for the modification experiments in dead-end could be established. A pH value of 9 and an acrylate/crosslinker ratio of 1:1 turned out to be optimal conditions for the formation of a hydrogel with elastic properties in a short time. For a gelation time of 10 min, a critical concentration of 12 wt% was determined under these conditions which was to be correlated with the membrane surface concentration during modification. During dead-end modification, influencing factors, such as polymer concentration, flux and stirring rate on gel formation were identified. A basic understanding of hydrogel formation was gained. It was found that the hydrogel formation undergoes two processes. At polymer concentrations up to 0.03 wt%, the reactants are filtered below the critical concentration in a gel-particulate state towards the membrane so that only thin hydrogel layers are formed. From a bulk concentration of 0.04 wt%, the reaction solution is concentrated above the required critical concentration, resulting in the formation of a three-dimensional hydrogel layer.
It was also shown that the filtration conditions, especially the flux, are determining for the initiation of a crosslinking reaction via CP and the formation of the hydrogel on the membrane surface. Under the selected conditions (0.06 wt%, pH 9, acrylate/cross-linker ratio of 1:1), the flux must be > 18 LHM for a crosslinking reaction to take place. If these conditions are fulfilled, the hydrogel formation will occur within the first few minutes. The membrane surface concentration cm for successful modifications calculated via film model turned out to be significantly lower than the critical concentration ccrit of 12 wt% calculated via rheology. This is primary caused by the fact that the gel point as an evaluation criterion is not representative for the reaction taking place at the membrane surface. Hence, rheology cannot be used to predict a priori conditions that lead to a certain coating effect, but it can be used to support the understanding of hydrogel formation. For the antifouling experiments, the cooling water from the steel industry was not a suitable test system due to low level of contaminants so that synthetic oil/water emulsions were used instead. However, it was found that membranes modified at even subcritical conditions (0.02 and 0.03 wt%), possess superior antifouling properties compared to unmodified membranes.
For the modification of membranes in spacer-filled channels, operated in cross-flow, conditions were applied from dead-end experiments which led to thick and homogeneous hydrogel layers. Different from dead-end, the selected conditions led only to thin hydrogel layers at subcritical conditions which is attributable to changed boundary conditions caused by the spacer. However, it could be shown that such membranes have outstanding antifouling properties towards crude oil emulsions compared to unmodified membranes. By enhancing the flux for the modification, a clearly pronounced hydrogel layer could be formed on the membrane surface. Even though, the formation of the hydrogel layer was not homogeneously formed over the entire membrane area, the overall charge neutrality could be improved significantly. Thus, it could be demonstrated that also under cross-flow conditions, the crosslinking reaction can be controlled and adjusted mainly via flux even though a spacer is used. Finally, the general feasibility of coating SWMs and the application of such modules for the treatment of cooling water of the steel industry could be proven. The antifouling effect of the modified modules was shown for a number of different experiments in the initial phase of the field tests. However, the modification effect of the coated modules could not be demonstrated in the later stages due to severe irreversible iron oxide particle, organic and oil fouling caused by fluctuating cooling water composition, insufficient feed pre-treatment, production conversion and defects at the plant.
The main focus of this work was therefore to develop a polyzwitterionic anti-fouling hydrogel coating for RO TFC membranes by CP-enhanced in situ thiol-based click reaction. The gradual transfer of this modification method from dead-end to modules in cross-flow was to be demonstrated. This was done with the aim of integrating modified RO modules into a cooling circuit of the steel industry to purify and allow the reuse of cooling water, as in Germany approx. 77 % of water is used for product cooling. Therefore, a critical evaluation of the separation performance of established membranes was done first by analyzing fouling with real cooling water. The cross-linking of the zwitterionic copolymer with the cross-linker 1,4-Dithiothreitol (DTT) via thiol-ene Michael addition click reaction was then studied in the bulk via in situ rheology. In a further step, the in situ cross-linking reaction was examined during dead-end filtration. The modification was then transferred to flat-sheet membranes in cross-flow. In the last step, an up-scaling to spiral-wound RO modules was accomplished and were then implemented in a cooling circuit of the steel industry where pilot scale tests were done.
During the fouling tests (dead-end short-term and cross-flow long-term experiments), the membrane BW30 exhibited a very high fouling resistance and showed no irreversible fouling towards real cooling water from the steel industry. This membrane was therefore chosen for the membrane modification tests. By means of rheology, the influence of different parameters (polymer concentration, pH value, acrylate/crosslinker ratio) on the bulk hydrogel formation could be investigated. By this, optimal parameters for the composition of reaction solution for the modification experiments in dead-end could be established. A pH value of 9 and an acrylate/crosslinker ratio of 1:1 turned out to be optimal conditions for the formation of a hydrogel with elastic properties in a short time. For a gelation time of 10 min, a critical concentration of 12 wt% was determined under these conditions which was to be correlated with the membrane surface concentration during modification. During dead-end modification, influencing factors, such as polymer concentration, flux and stirring rate on gel formation were identified. A basic understanding of hydrogel formation was gained. It was found that the hydrogel formation undergoes two processes. At polymer concentrations up to 0.03 wt%, the reactants are filtered below the critical concentration in a gel-particulate state towards the membrane so that only thin hydrogel layers are formed. From a bulk concentration of 0.04 wt%, the reaction solution is concentrated above the required critical concentration, resulting in the formation of a three-dimensional hydrogel layer.
It was also shown that the filtration conditions, especially the flux, are determining for the initiation of a crosslinking reaction via CP and the formation of the hydrogel on the membrane surface. Under the selected conditions (0.06 wt%, pH 9, acrylate/cross-linker ratio of 1:1), the flux must be > 18 LHM for a crosslinking reaction to take place. If these conditions are fulfilled, the hydrogel formation will occur within the first few minutes. The membrane surface concentration cm for successful modifications calculated via film model turned out to be significantly lower than the critical concentration ccrit of 12 wt% calculated via rheology. This is primary caused by the fact that the gel point as an evaluation criterion is not representative for the reaction taking place at the membrane surface. Hence, rheology cannot be used to predict a priori conditions that lead to a certain coating effect, but it can be used to support the understanding of hydrogel formation. For the antifouling experiments, the cooling water from the steel industry was not a suitable test system due to low level of contaminants so that synthetic oil/water emulsions were used instead. However, it was found that membranes modified at even subcritical conditions (0.02 and 0.03 wt%), possess superior antifouling properties compared to unmodified membranes.
For the modification of membranes in spacer-filled channels, operated in cross-flow, conditions were applied from dead-end experiments which led to thick and homogeneous hydrogel layers. Different from dead-end, the selected conditions led only to thin hydrogel layers at subcritical conditions which is attributable to changed boundary conditions caused by the spacer. However, it could be shown that such membranes have outstanding antifouling properties towards crude oil emulsions compared to unmodified membranes. By enhancing the flux for the modification, a clearly pronounced hydrogel layer could be formed on the membrane surface. Even though, the formation of the hydrogel layer was not homogeneously formed over the entire membrane area, the overall charge neutrality could be improved significantly. Thus, it could be demonstrated that also under cross-flow conditions, the crosslinking reaction can be controlled and adjusted mainly via flux even though a spacer is used. Finally, the general feasibility of coating SWMs and the application of such modules for the treatment of cooling water of the steel industry could be proven. The antifouling effect of the modified modules was shown for a number of different experiments in the initial phase of the field tests. However, the modification effect of the coated modules could not be demonstrated in the later stages due to severe irreversible iron oxide particle, organic and oil fouling caused by fluctuating cooling water composition, insufficient feed pre-treatment, production conversion and defects at the plant.