Application of hybrid membrane technology to recover resources from concentrates

According to the sustainable development goals of the United Nations, water shortage may cause poor health conditions and affect people's livelihood and food security. Therefore, alternative water sources are required to supply freshwater, and seawater is one of them. Application of membranes is one of the most used and efficient methods to produce freshwater from seawater. Nanofiltration (NF) and reverse osmosis (RO) are primarily employed in desalination. Usually, NF and RO are thin film composites, but NF has lower separation efficiency for smaller and less charged ions (such as sodium and chloride ions) than RO membranes. This study focuses on the application of NF in seawater desalination and investigating its useful applications by modifying its properties. NF membranes with high selectivity for monovalent and divalent ions (and salts) will produce concentrates that would have divalent salts. The permeate from NF that passes through the RO will produce concentrates that is rich in monovalent salts. Such a system will be efficient in producing pure salts from both the NF and RO concentrates. The following research questions are answered in this study through experiments and modelling: (i) What is the current rejection of monovalent and divalent ions by NF membranes, and how to increase the rejection of multivalent ions and decrease the rejection of monovalent ions by a NF membrane? (ii) How do the existing mathematical models need to be modified to predict the performance of the NF membranes? (iii) What are the suitable substrates and monomers to prepare such NF membranes? What is the optimum composition of the above materials to achieve the rejection goals with respect to monovalent and divalent ions? (iv) What are the mechanisms involved in the rejection of monovalent and multivalent ions, and what is the performance of NF membrane with respect to permeate flux? (v) How the synthesized membrane will reject the monovalent and divalent ions present in seawater, and what additional modifications to the NF membrane are required to achieve the goals of separating monovalent and multivalent ions present in seawater? The latest research progress of NF is reviewed and compared to the rejection of monovalent and divalent ions by NF membranes to answer the first research question. The processes of NF membrane preparation using phase inversion, layer by layer, and interfacial polymerization have been described. The rejection mechanism of NF including steric hindrance effect, Donnan effect, and dielectric exclusion are analysed, and some common characterization methods have been summarized. To answer the second research question, the traditional Donnan steric pore model and dielectric exclusion (DSPM-DE) was modified to fit the rejection of monovalent and divalent salts under different operating pressures. The study analysed the effect of membrane thickness on the removal rate of neutral solutes; investigated the effect of two different methods of calculating the Peclet number on the rejection rate of monovalent and divalent salts by fitting the membrane surface charge density and introducing the actual thickness of the active layer. When the operating pressures are above 10 bar, the effective active layer thickness, is more applicable in calculating the Peclet number than using the membrane pore size. Utilizing the effective active layer at low pressures causes the model to overestimate the rejection of monovalent salts by 11.76% and to underestimate the rejection of divalent salts by 8.85%. This study demonstrates that the membrane thickness significantly impacts the separation performance of membranes, and therefore, an appropriate method to calculate the Peclet number should be chosen based on the applied pressure when modelling. In addition to the thickness of membranes, the effect of the substrate on the flux and separation performance of NF is a blank. Traditional monomers piperazine (PIP) and trimesoyl chloride (TMC) were used for interfacial polymerization (IP) on different substrates to investigate the impacts of substrates on the performance of the thin-membrane composite (TFC) nanofiltration membranes. This part of the study answered research questions three and four. It is found that when the molecular weight cut-offs (MWCO) of the substrate increased: (i) the effective thickness of the substrate decreased; (ii) the average pore size of the substrate enlarged in general; and (iii) the surface roughness of the substrate increased. When the MWCO of substrates was increased, the TFC membranes synthesized from them showed the following characteristics: (i) the thickness of the polyamide layer expanded; (ii) average size of the nodules on the membrane surface enlarged, (iii) permeate flux enhanced in general, (iv) the rejection of NaCl increased, and the rejection of MgSO4 did not change significantly; (v) flux recovery rates decreased. The study showed that the MWCO of the substrate had a similar influence on the permeate fluxes produced by the substrate and the TFC membrane. Lower MWCO substrates induced higher selectivity for divalent and monovalent ions. To answer the fifth question, new materials β-cyclodextrin (β-CD), tannic acid (TA), and ZIF-8 were applied in modifying NF membranes for enhancing the separation performance and the flux of NF membranes. The study investigated the effects of monomer concentration, substrate morphology, and reaction temperature on membrane performance. The results showed that when the concentrations of both β-CD and TA were 1 wt%, the membranes had the highest selectivity (2.99) for monovalent and divalent salts. At 10-fold increase in TMC concentration (0.05 to 0.5 wt%), membrane selectivity and flux had optimum values. When both spongy and cavernous structures were present in the substrate, the membranes could achieve more than 85% removal of both monovalent and divalent salts. The surface of a synthesized membrane was smooth with root mean square roughness at 0.396 nm. After the introduction of ZIF-8, the roughness of the membrane increased, and the removal rate of monovalent salts decreased significantly. The growth of ZIF-8 on the membrane surface was assisted by an increase in reaction temperature. This study demonstrated the potential of the synthetic membrane to have high selectivity for monovalent and divalent salts.