Development of a diafiltration-nanofiltration-reverse osmosis (DiaNF-RO) process for ion fractionation towards resource recovery in seawater desalination

• Main Author: Truong, Vinh Hien
• Other Authors: Chong Tzyy Haur
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Desalination; Ion fractionation; Optimisation; Membrane process; Energy efficient
• Online Access: https://hdl.handle.net/10356/174821
• Institution: Nanyang Technological University

Seawater reverse osmosis (SWRO) desalination has become one of the most viable options to mitigate the global water shortage issues. However, one of the problems of the SWRO is that it cannot reach high recovery due to the pressure limit and fouling risk. As such, a very concentrated by-product stream, known as brine, are produced. If the brine discharge is not carefully monitored, it may disturb the marine ecology and increase the disposal cost. Hence, many brine management strategies have been proposed worldwide to mitigate this problem. This research introduces the novel diafiltration-nanofiltration-reverse osmosis (nDiaNF-RO) process that aims to achieve both divalent-monovalent ion fractionation and desalination. Firstly, using n numbers of NF stages in series, the divalent-monovalent ion fractionation is initiated at the pre-treatment stage instead of at the conventional post-treatment step to alleviate the high osmotic pressure constraint. Diafiltration is also employed to enhance the divalent-monovalent separation property of NF membrane. Then, an RO stage is integrated to desalinate the NF permeate, which provides both the diluent and product water. Lastly, all these stages are operated in a continuous mode, so that it can be more easily incorporated into new or existing desalination plant. Overall, the process produces 3 streams: Brine 1 (Br1) or retentate of the last NF stage, which is rich in divalent ions; Brine 2 (Br2) or retentate of RO stage, which is rich in monovalent ions; and Product Water (PW) or the net RO permeate. In this research, Mg and Na are the representative divalent and monovalent ions, respectively, due to their seawater abundance and potentially profitable extraction. The Mg-rich Br1 stream could facilitate the subsequent Mg(OH)2 precipitation, which is important for magnesia production and carbon sequestration. The Na-rich Br2 stream could be used to recover the valuable NaOH from the chlor-alkali process. With many potential resource recovery applications, the ion recovery and fractionation properties of nDiaNF-RO should be thoroughly studied. In the first study, optimal performance of nDiaNF-RO was determined via simulations, using a commercial NF membrane with a Mg-Na separation factor of 2.44 at 16 bar and 35 g/L seawater. Firstly, sensitivity analysis was conducted to show some favourable operating conditions, such as dilution should only occur in the last NF stage and no retentate recycling is needed unless fouling potential is significant. Next, genetic algorithm was implemented to find the optimal NF pressures and number of elements to simultaneously maximize the Mg-Na separation factor in Br1 (SF1Mg-Na) and minimize the net specific energy consumption (SECnet). More stages often achieved better optimal results but with a trade-off in Mg recovery. At the highest maximum selectivity condition, the 2DiaNF-RO design amplified the SF1Mg-Na of NF by 3 times to 7.32, with Na dilution (CF1Na = 0.54) and Mg concentration (CF1Mg = 3.95), at SECnet at 5.97 kWh/m3. However, 2DiaNF-RO with commercial NF cannot achieve simultaneous targets of SF1Mg-Na > 8 and SECnet < 6 kWh/m3. In the second study, the novelty and practicality of the nDiaNF-RO was justified. Firstly, the pre-treatment nDiaNF-RO (i.e., treating high volume low concentration seawater) was compared to the corresponding post-treatment RO-nDiaNF (i.e., treating low volume high concentration SWRO brine) in various aspects via simulations. It was shown that the pre-treatment outperformed the post-treatment in Mg recovery, Mg-Na ion fractionation, membrane area requirement, gypsum scaling potential, and more efficient cost utilization at the expense of slightly higher SECnet. This result emphasizes the advantages of the pre-treatment feature in nDiaNF-RO over post-treatment in brine management because NF membrane performance is a strong function of total dissolved solids (TDS). Secondly, the 2DiaNF-RO was shown to become desirable if an improved NF membrane was used or a 2nd RO stage is employed. If there was a more suitable NF membrane (in the order of higher water permeability, lower Na rejection, and higher Mg rejection), the 2DiaNF-RO could achieve the desirable fractionation target (i.e., SF1Mg-Na = 13.7 > 8.0) that was previously impossible for the original 2DiaNF-RO. If the 2nd RO stage was used to just increase the water recovery, the SECnet could be further reduced by 26-30% to 4.01 kWh/m3, which makes it competitive to other RO-based processes in brine management (SECnet < 6 kWh/m3). Overall, given some appropriate modifications, nDiaNF-RO could achieve desirable performance at low stage number, which further justified its practical feasibility. In the last study, the equivalent lab-scaled semi-batch DiaNF-RO experiment was conducted, which showed that the process could indeed achieve Mg-Na and Na-Mg ion fractionation in brine 1 and 2, respectively, while desalinate seawater. Moreover, upon validation, the 2DiaNF-RO model was shown to achieve at least 77.0% and 99.07% prediction accuracy in Br1 and PW, respectively. Due to the simulation’s underestimation in NF rejection and overestimation in NF and RO water permeability, the experimental processes often have higher ion fractionation and require longer operation time than predicted. Finally, by comparing the experimental semi-batch 2DiaNF-RO to the corresponding RO-2DiaNF, the performance benefits of pre-treatment were reiterated. The findings from this research provide a better understanding on the technical feasibility and economic possibility of the DiaNF-RO design. However, more data from the future pilot test are required to further improve the practicality of the process. If the implementation of DiaNF-RO is realized, it will become an important step in the integrated treatment scheme that aims to realize the resource recovery in seawater desalination and brine management.