Silica mitigated calcium mineral scaling in brackish water reverse osmosis

Mixed scaling patterns and mechanisms of high-pressure nanofiltration in hypersaline wastewater desalination

Nanofiltration (NF) will play a crucial role in salt fractionation and recovery, but the complicated and severe mixed scaling is not yet fully understood. In this work, the mixed scaling patterns and mechanisms of high-pressure NF in zero-liquid discharge (ZLD) scenarios were investigated by disclosing the role of key foulants. The bulk crystallization of CaSO4 and Mg-Si complexes and the resultant pore blocking and cake formation under high pressure were the main scaling mechanisms in hypersaline desalination. The incipient scalants were Mg-Si hydrates, CaF2, CaCO3, and CaMg(CO3)2. Si deposited by adsorption and polymerization prior to and impeded Ca scaling when Mg was not added, thus pore blocking was the main mechanism. The amorphous Mg-Si hydrates contribute to dense cake formation under high hydraulic pressure and permeate drag force, causing rapid flux decline as Mg was added. Humic acid has a high affinity to Ca2+by complexation, which enhances incipient scaling by adsorption or lowers the energy barrier of nucleation but improves the interconnectivity of the foulants layer and inhibits bulk crystallization due to the chelation and directional adsorption. Bovine serum albumin promotes cake formation due to the low electrostatic repulsion and acts as a cement to particles by adsorption and bridging in bulk. This work fills the research gaps in mixed scaling of NF, which is believed to support the application of ZLD and shed light on scaling in hypersaline/ultra-hypersaline wastewater desalination applications.

Antiscalants for mitigating silica scaling in membrane desalination: Effects of molecular structure and membrane process

Silica scaling is a major type of mineral scaling that significantly constrains the performance and efficiency of membrane desalination. While antiscalants have been commonly used to control mineral scaling formed via crystallization, there is a lack of antiscalants for silica scaling due to its unique formation mechanism of polymerization. In this study, we performed a systematic study that investigated and compared antiscalants with different functional groups and molecular weights for mitigating silica scaling in membrane distillation (MD) and reverse osmosis (RO). The efficiencies of these antiscalants were tested in both static experiments (for hindering silicic acid polymerization) as well as crossflow, dynamic MD and RO experiments (for reducing water flux decline). Our results show that antiscalants enriched with strong H-accepters and H-donors were both able to hinder silicic acid polymerization efficiently in static experiments, with their antiscaling performance being a function of both molecular functionality and weight. Although poly(ethylene glycol) (PEG) with abundant H-accepters exhibited high antiscaling efficiencies during static experiments, it displayed limited performance of mitigating silica scaling during MD and RO. Poly (ethylene glycol) diamine (PEGD), which has a PEG backbone but is terminated by two amino groups, was efficient to both hinder silicic acid polymerization and reduce water flux decline in MD and RO. Antiscalants enriched with H-donors, such as poly(ethylenimine) (PEI) and poly(amidoamine) (PAMAM), were effective of extending the water recovery of MD but conversely facilitated water flux decline of RO in the presence of supersaturated silica. Further analyses of silica scales formed on the membrane surfaces confirmed that the antiscalants interacted with silica via hydrogen bonding and showed that the presence of antiscalants governed the silica morphology. Our work indicates that discrepancy in antiscalant efficiency exists between static experiments and dynamic membrane filtration as well as between different membrane processes associated with silica scaling, providing valuable insights on the design principle and mechanisms of antiscalants tailored to silica scaling.