Evaluation of the nanofiltration of brines from seawater desalination plants as pre-treatment in a multimineral brine extraction process

Abstract

Many materials essential to the European Union's (EU) economy lack primary resources, and the major suppliers of these materials are non-EU nations. It has been determined that seawater reverse osmosis (SWRO) desalination brines are a circular answer for the supply of critical elements. SWRO desalination brines, however, are distinguished by high concentrations of elements with negligible economic worth. (e.g., Na, Cl) compared to critical elements present in very low concentrations (e.g. Mg, B, Sc, V, In, Ga, Li, Mo, Rb). Therefore, an efficient separation pre-treatment is necessary to maximize the recovery efficiency in a brine extraction process. Nanofiltration (NF) was proposed as pre-treatment stage to separate monovalent from divalent ions and it was investigated if a Ca(II) removal stage was necessary before the NF to avoid scaling. The performances of three polyamide commercial membranes (NF270, Fortilife XC-N and PRO-XS2) treating SWRO brine (brine 1) and Ca(II)-free SWRO brine (after Ca(II) removal with sodium bicarbonate - brine 2) were evaluated in a closed circuit configuration as a function of the applied pressure from 6 to 30 bar.Results showed that NF was an efficient process to separate target elements present in seawater desalination brine into two different streams. Moreover, removal of Ca(II) before the NF step (brine 2) increased the selectivity between multivalent and monovalent species (higher rejection of multivalent and lower rejection of monovalent species compared to brine 1) for all the membranes tested, reaching the highest selectivity at 20 bar. At 20 bar, the polyamide membrane PRO-XS2 presented rejections of 71% for Ca(II) and 89% for Mg(II) and the highest selectivity in experiments with brine 2, being recommended for this application. In addition, the transport of species was characterized by membrane permeances by using the Solution-Electro-Diffusion-Film model. The scaling assessment at the membrane surface showed a lower scaling potential if Ca(II) was removed before the NF stage.