PhD Defence Tjerk Watt | Advancing Hollow Fiber Nanofiltration Membranes: Materials, Structure, and Performance

Advancing Hollow Fiber Nanofiltration Membranes: Materials, Structure, and Performance

Tjerk Watt is a PhD student in the department Membrane Science & Technology. (Co)Promotors are prof.dr.ir. W.M. de Vos & dr.ir. J.de Grooth & dr. E. te Brinke from the faculty of Science & Technology (TNW), University of Twente.

Many regions of the world are dealing with an increasing scarcity of fresh water. This increase in scarcity is caused by climate change, a growing population, and increased industrialization, resulting in more droughts, higher demand, and more contaminants in our supplies. One method of combating this water scarcity would be to increase the amount of wastewater reuse. To enable such re-use, adequate and affordable wastewater treatment facilities will need to be installed, which are preferably fabricated using sustainable materials. One of the technologies that can be used for the treatment of wastewaters is polyelectrolyte multilayer (PEM) nanofiltration membranes. PEM membranes excel at the removal of small organic pollutants while only retaining small amounts of monovalent ions, making them ideal for wastewater treatment applications. However, the current method of fabrication of PEMs leaves room for improvement with regard to the active membrane surface area that can be packed into a single membrane module, the permeability of the hollow fiber (HF) PEM membranes, and the sustainability of the coated PEM. Therefore, in this thesis, we investigated methods of improving these aspects of PEM membranes to develop the next generation of affordable and sustainable HF PEM nanofiltration membranes for wastewater reuse.

In Chapter 2 of this thesis, we focused on enhancing the active membrane surface area per volume of HF. Currently, PEMs are made with the selective layer on the inside of the HF. However, to reduce the physical footprint of PEMM modules, it would be beneficial to have the selective layer on the outside. This will lead to an increased surface area per fiber and enable the use of smaller fibers, which further increases the surface area per membrane module. To prove that the concept of an outer-skinned PEM membrane is feasible, we used the dry-jet wet-spinning process to fabricate charged outer-skinned HF supports and coated them through the layer-by-layer self-assembly process to form a PEM membrane. Scanning electron microscopy images of the HFs confirmed the existence of an asymmetric structure with an outer skin, while fluorescence imaging confirmed that the PEM is located on the outside of the HF. Filtration experiments showed that the PEM membranes exhibited nanofiltration properties similar to the conventional inner-skinned PEM membranes. Overall, these membranes show a 3.8x increase in active membrane surface area to volume ratio compared to commercial PEM membranes, clearly highlighting the benefits in terms of footprint.

Due to these newly fabricated HF membranes having a selective layer on the outside of the membrane, they can now more effectively be operated in an outside-in filtration configuration where the feed is located on the outside of the HF. In such a configuration, it is important to know the limits of the external pressure that can be applied to the membrane. Therefore, in Chapter 3 we investigated whether a model for predicting the collapse pressure of isotropic thin-walled cylinders was able to accurately predict the collapse pressure of outer-skinned HF membranes. Theoretical derivations showed that collapse can occur due to plastic or elastic failure, where, in the case of plastic failure, the collapse pressure should equal that of the burst pressure. Combining experimental results with our model revealed that for our membranes, plastic failure was the dominant failure mechanism. The model was able to accurately predict the influence of the membrane's porosity on the collapse pressure. However, the model seemed to only partially predict the influence of different geometric dimensions of the HF on its collapse pressure. Interestingly, though, the results showed that in most cases, the burst pressure was indeed similar to the collapse pressure, highlighting that the burst pressure can be used as an indicator for the collapse pressure of the outer-skinned HF membranes. Additionally, the model provided a clear direction to further improve the collapse pressure.

In Chapter 4, we focused on lowering the operational costs of PEM membranes by enhancing the permeability of the PEM membrane. A continuous challenge with enhancing the permeability of membranes is the trade-off between the membrane's permeability and selectivity. One method of avoiding this trade-off is by reducing the thickness of the selective layer. However, PEM membranes have a limit to this reduction, as they first have to fill the pores of their support membrane before they can form a defect-free selective layer. To overcome this limitation, we propose the use of nanoparticles as sacrificial pore fillers. Therefore, we coated silica (SiO₂) nanoparticles in the initial bilayer of a polydiallyldimethylammonium chloride/poly(sodium styrene sulfonate) (PDADMAC/PSS) PEM to fill the pores of the support membrane. This makes sure that subsequent PDADMAC/PSS bilayers are uniformly deposited on the membrane surface without entering the pores. Subsequently, we dissolved the SiO₂ nanoparticles in a sodium hydroxide solution to open the space inside the pores. The results showed that nanoparticles small enough to fit into the pores of the support membrane were successfully able to act as sacrificial pore fillers and enhance the permeability up to 2x, while the 90% molecular weight cut-off (MWCO) even slightly improved. Membranes showed a slight change in ion retentions as the negatively charged membranes lost some of their overall charge due to the dissolution of the negatively charged SiO₂ nanoparticles. However, the changes in separation properties were very small, and Donnan exclusion remained the dominant separation mechanism. These results show that nanoparticles can act as sacrificial pore fillers, enabling PEM membranes to achieve superior performance compared to commercial nanofiltration membranes by reducing their energy requirements in separation processes.

While Chapters 2, 3, and 4 focused more on the affordability and energy use aspect of PEM membranes, Chapter 5 of this thesis looks more into the sustainability of PEM membranes. A problem regarding this aspect is that they are made of nonrenewable and nonbiodegradable fossil resources. Therefore, to improve the sustainability of PEM membranes, we investigated lignin as a renewable and biodegradable alternative for the fabrication of the selective layer. We used lignosulfonate as a polyanion in combination with modified Kraft lignin as a polycation in a layer-by-layer self-assembly process to coat HF support membranes to obtain so-called all-lignin PEM membranes. The PEM membranes showed loose nanofiltration properties (MWCO > 1 kDa, magnesium sulfate retention 20%) that could easily be fine-tuned by changing the ionic strength of the coating solutions. Furthermore, the lignin PEM membranes exhibited excellent stability in saline solutions of up to 5 M sodium chloride and were stable in a pH range from 1 to 11. Additionally, the lignin retained its biodegradable properties in the presence of laccase enzymes after forming a PEM membrane. Our results indicate that lignins are a suitable candidate for replacing fossil-based polyelectrolytes for the fabrication of chemically stable, renewable, and biodegradable PEM membranes.

In Chapter 6 of this thesis, a summary of the key insights that were obtained in this work is provided, together with an outlook for future research.