View all hidden authors and organizations

Edited by Manish Kumar of the University of Texas at Austin, accepted by Pablo G. Debenedetti, member of the editorial board, on February 27, 2021 (reviewed on November 17, 2020)

Inspired by biological models, artificial water channels can be used to overcome the permeability/selectivity trade-offs of traditional desalination membranes. We prove that the I-quartet artificial water channel is reasonably combined in the composite polyamide membrane synthesized by interfacial polymerization to provide a defect-free biomimetic membrane. It has an inherent water pair when operating brackish water reverse osmosis (BWRO)/tap water reverse. Salt Permeability Osmosis (TWRO) Desalination pressure and medium/low salinity conditions.

Membrane-based technology plays a huge role in water purification and desalination. Inspired by biological proteins, artificial water channels (AWC) have been proposed to overcome the permeability/selectivity trade-off of the desalination process. A promising strategy using AWC with angstrom-scale selectivity shows their impressive performance when embedded in a bilayer membrane. Here, we proved that the self-assembled imidazole quartet (I-quartet) AWCs are macroscopically integrated in the industry-related reverse osmosis membrane. In particular, we explored the best combination of I-tetragonal AWC and meta-phenylenediamine (MPD) monomers to achieve seamless integration of AWC in defect-free polyamide membranes. The performance of the membrane was evaluated by cross-flow filtration by filtering the brackish feed stream under actual reverse osmosis conditions (15 to 20 bar applied pressure). The optimized biomimetic membrane has achieved unprecedented improvements, resulting in a similar commercial membrane with more than twice the water permeability (up to 6.9 L⋅m−2⋅h−1⋅bar−1) while maintaining an excellent NaCl rejection rate (>99.5%) ). They also show excellent performance in purifying low-salinity water under low-pressure conditions (6 bar applied pressure), with fluxes as high as 35 L⋅m-2⋅h-1 and 97.5 to 99.3% of the observed rejection.

Over the past few decades, competition for increasingly polluted freshwater resources has made the availability of safe drinking water a serious global challenge in the 21st century. Together with environmental remediation technologies and sustainable protection policies, seawater desalination is one of the ways to solve the problem of freshwater scarcity (1, 2). At present, the production of desalinated water is about 100 million cubic meters per day, with an average annual growth rate of more than 10%. The traditional reverse osmosis (RO) membrane is a thin film composite (TFC), which relies on a supporting polyamide (PA) layer for water/solute separation (3, 4).

The role of the synthetic membrane is to dissolve and diffuse proportionally selective solutes through its polymer matrix: a material property that leads to permeability-selectivity trade-off behavior (5). Many methods have been proposed to overcome this limitation, including membrane grafting, chemical functionalization and process optimization (6⇓ ⇓ ⇓-10). Despite several elegant design strategies, membrane-based technologies always increase water permeability at the expense of lower selectivity, and vice versa (11). Although membrane-based separation provides energy efficiency, excellent water quality, and the ability to be implemented on an industrial scale, there is an urgent need for membranes with higher selectivity and higher permeability, especially to improve the efficiency of water purification technologies (2, 4) .

Recent developments have made materials prepared under molecular control more and more attractive to researchers, providing the possibility of stable and efficient biomimetic membranes (1, 2). Due to their unique functional design, biological aquaporins and artificial water channels (AWC) have proven high performance on the angstrom scale. When embedded in lipid bilayers, they stimulate their combination with seawater desalination membranes (12, 13 ). In the past ten years, people have developed self-assembly (14, 15) and single molecule (16) AWC, hoping to mimic complex biological mechanisms (17⇓ –19). These methods show that when embedded in lipid bilayers, AWC is mainly active, but it can be upgraded according to industry standards for the preparation of biomimetic RO polymer membranes (2).

The defects generated at the interface between the active AWC and the surrounding matrix are critical to the desalination performance of the RO membrane. When manufacturing high-performance reverse osmosis membranes for desalination, many necessary characteristics need to be achieved to maximize productivity, including suitable mechanical resistance, high density of functional channels per unit area, and perhaps the most ambitious, without defects ( 1). We hypothesize that one of the creative strategies to solve the amplification challenges and improve performance is to combine established PA materials prepared by interfacial polymerization (IP), and is known for its integration in typical roll-to-roll processing systems, with permeability and options Sexual AWC (2).

We previously reported on the manufacture of biomimetic membranes, including the I-quartet AWC embedded in the classic PA (2). We know that the composition diversity of PA components and additives has a huge impact on membrane structure and performance (1). In this report, our strategy has given people a deeper understanding of how the incorporation of AWC affects the performance of PA/AWC membranes. The reported mixed AWC-PA materials can be optimized at the nanometer level with different morphologies and chemical properties by using the optimal mixing of various formulations of different amounts of AWC and PA components. Specifically, we studied valuable insights into the manufacturing-structure-performance relationship, and discussed the influence of the composition of AWC and meta-phenylenediamine (MPD) components on the improvement and discovery of better performance of biomimetic membranes. The strategy here is related to the direct quantification of specific ingredients, which allows a seamless active layer to be obtained in all cases, where AWC is gently and optimally combined without defects. Therefore, a uniform sponge-like superstructure of mixed PA-AWC material is obtained, and it is applied under the medium and low pressure filtration under the conditions of real brackish water RO (BWRO) and low-salinity water RO desalination conditions.

The imidazole quartet (I-quartet) (14, 15) consists of four molecules of HC6 and two water molecules, which form a pore size of 2.6 Å and are filled with oriented single water lines (Figure 1A) (20). The orientation of water (21) is essential for enhancing its translocation in the range of 106 to 107 water molecules·s-1 per channel, and perfect salt rejection is achieved through the single thread (20) or sponge cluster (17) mechanism. The crystal stacking reveals the self-assembled alternative I-quartet parallel pieces in the lamellar phase of the parallel single water channels (Figure 1C and D), and matches well with the ordered rows observed in the transmission electron microscope (TEM) micrographs The crystalline powder obtained from the evaporation of the HC6 ethanol casting aqueous solution (Figure 1B).

Film synthesis: (A) I-quartet AWC; (B) high TEM magnification of layered crystalline nanoparticles, obtained by evaporating HC6 ethanol/water solution; (C) top view of I-tetragonal AWC single crystal layered stacking and (D) Side view. (E) Synthetic procedure for the preparation of hybrid PA/AWC membranes: impregnate PSf carrier 1 with HC6 solution to form the colloidal layered phase of AWC, then impregnate with m-phenylenediamine (MPD) aqueous solution 2, and interact through (F) and AWC The H-bonding of colloidal particles and the impregnation of trimesoyl chloride (TMC) in hexane3 resulted in the formation of a mixed H-bonded PA/AWC layer through the nucleation/IP process.

The RO membrane in this study was prepared using MPD and trimesoyl chloride (TMC) IP, resulting in the formation of a PA layer on a commercial polysulfone (PSf) carrier. For the synthesis of the bionic PA layer, AWC is added before the traditional IP procedure (Figure 1E). In the first step, the ethanol solution of HC6 resulted in the formation of stable colloidal aggregates (Dh = 200 nm). When water was added to the ethanol solution of HC6, it was observed by dynamic light scattering (DLS) that there was no visible precipitation. Please note that compared to the other channels HC4-HC8 studied previously, HC6 with the best solubility can avoid precipitation under the described IP experimental conditions, thus forming a defect-free film. Pour the aqueous solution on the PSf carrier, and then immerse it in the MPD aqueous solution to form a mixed AWC/MPD nanocolloid aggregate. DLS analysis shows that when using MPD aqueous solution instead of pure water, smaller aggregates (Dh = 70 nm) are obtained (SI appendix, Figure S1), which is reminiscent of the hydrogen bond interaction between AWC and MPD ( 2). The integration of AWC into the PA mixed layer is firstly related to the nucleation process of this colloidal structure; then, as shown by DLS, the hydrogen bond interaction between MPD and colloidal AWC nano-aggregates promotes heterogeneous nucleation, while observing To any precipitation process. Based on this information, different HC6/MPD ratios were studied to explore the influence of composition on the material morphology that controls and adjusts the transmission performance of the membrane (SI Appendix, Table S1). We know that during the initial formation of the PA film, the growth of the polymer remains almost unchanged, which depends on the diffusion rate of the MPD monomer in the organic phase (22). An excess of MPD is needed to promote significant diffusion into the organic phase during film formation; therefore, an equal or substoichiometric amount of MPD may result in a lower degree of polymerization (23, 24). After immersing the support layer in the MPD solution, use an air gun to remove the excess solution from the surface, and immerse the support member saturated with HC6-MPD in the TMC solution to form the PA layer. In this step, the AWC/MPD amphiphilic colloidal aggregates may help MPD migrate to the organic phase, which is conducive to the IP reaction. At the same time, they may interact with the growing PA fragments through H bonds. We speculate that this process leads to the formation of cross-linked hybrid AWC-PA materials. The control membrane and the biomimetic membrane were prepared according to the IP procedure, and they were cured in 95 °C deionized (DI) water, immersed in 200 ppm NaOCl aqueous solution, and then immersed in 1,000 ppm NaHSO3 aqueous solution. Before testing or characterization, the membrane is finally rinsed and stored in deionized water at 4 °C.

Fourier transform infrared (FTIR) spectroscopy demonstrated the formation of a hybrid PA-AWC network (SI appendix, Figure S2 A and B). Vibration bands related to the amide bond of PA are visible in all infrared spectra, namely 1,502 cm-1 (amide-II, NH in-plane bending) and 1,666 cm-1 (amide-I, C=O stretching vibration) (22 ). Evidence of AWC incorporation in the hybrid PA film was obtained from the vibration displacements at 1,585 and 1,609 cm-1, assigned to C=C stretching of imidazole and C=O stretching of urea groups, and -CH2 ,as and -CH2 , The symmetric stretching modes of the HC6 alkyl chain are visible in the range of 2,750 and 2,957 cm-1, respectively. The sharp band centered at 3,400 cm-1 is attributed to the OH stretching vibration of the strong H-bonded water in the I-quartet channel. Due to the presence of hydrogen-bonded water with stronger fluidity and lower density in the matrix of the biomimetic layer, the band becomes very wide for the TFC-HC6 membrane. This PA-AWC hybrid material produced by the nucleation/IP mechanism is reminiscent of the self-assembled I-quadruplex that appears in the porous spongy membrane, and theoretically shows enhanced water permeability, as our previous molecular model Research shows. 17). They imply that dynamic superstructures can provide interconnected transmembrane pathways through water networks.

The SI appendix, Figure S2B shows the average elemental mass fraction X% obtained from energy dispersive X-ray spectroscopy, used for biomimetic films prepared by maintaining a constant amount of HC6 and different MPD concentrations. Compared with control TFC membranes with C% of 61 to 67 wt% of traditional PA membranes (23, 24) reported in the literature, the C% of 80 to 85 wt% of all bionic PAs is significantly higher. AWC layer. This result is expected because the C/O mass ratio of HC6 (~10) is significantly higher than that of PA (~3.5 to 4), thus providing additional evidence of the incorporation of AWCs into the mixed AWC-PA layer. By reducing the molar ratio of MPD, thereby reducing the density of the amine part and the amide bond, the N% is reduced, and the relative C% is increased by about 4%.

Scanning electron microscope (SEM) micrographs (Figure 2 and SI appendix, Figure S3) show that the layer morphology and HC6 incorporation level can be adjusted by adjusting the MPD/HC6 ratio during the film manufacturing process. All bionic layers are continuous and free of microscopic defects. The nano- and micro-structure of the PA layer that controls mass transmission strongly depends on the degree of cross-linking of the PA, thereby affecting the density and distribution of AWC in the PA matrix. Generally, under a constant HC6 load, as the concentration of MPD in the manufacturing solution decreases, the top surface evolves from a "ridge-valley" morphology (the effect of the rapid diffusion of concentrated MPD into the organic phase), a typical traditional PA layer, to The "flaky" morphology (SI appendix, Figures S4 and S5) observed lower MPD concentrations. In addition to the changes in surface roughness, biomimetic membranes with high HC6 loading and medium to high MPD concentrations show a unique and highly porous core structure with a relatively large thickness ranging from 150 to 300 nm. More interestingly, all layers prepared in the presence of HC6 also contain a sponge-like filling superstructure distributed over the entire active layer (Figure 2A and B and SI Appendix, Figure S6). These innovative internal nanostructures and overall morphology seem obvious to the optimally loaded membranes called 0.8-HC6 [0.8% (wt/wt) MPD and 1.5% (wt/wt) HC6], as shown in Figure 2C and D. Obviously, the cross-section of this representative biomimetic membrane has the following characteristics: 1) a smoother upper surface than the control TFC membrane, 2) a highly porous multilayer structure containing large voids, and 3) a larger The overall thickness, and 4) There are high-density and uniformly distributed nanoparticles embedded in the PA matrix. In an experimental BWRO test, the best combination of productivity and selectivity was obtained using this membrane (see below). When the biomimetic PA-AWC layer is prepared with a substoichiometric MPD molar ratio [≤0.4% (wt/wt)], the concentration of the monomer becomes the main factor controlling the final film structure: these layers usually have no large voids below and It is characterized by a much smaller average thickness of 50 nm; see the representative samples in Figure 2 E and F. The sponge-like nanostructures in the embedded layer are still observed, although the density is significantly lower, which may be due to the harder incorporation of films with a lower degree of polymerization. This result shows that the complex core structure of the selective layer embedded in the self-aggregating AWC-PA sponge-like superstructure is a strong function of the HC6/MPD ratio and the absolute monomer concentration used in the manufacturing process, and the layer structure can ultimately be controlled by these parameters To adjust and improve RO performance, this will be described in detail below.

(A and B) Representative cross-sectional SEM micrographs of H-0.9; (C and D) 0.8-HC6; (E and F) 0.2-HC6. The MPD concentration in the casting solution is (A and B) 3.4%; (C and D) 0.8%; (E and F) 0.2% (wt/wt). The concentration of HC6 in the casting solution is (A and B) 0.9% and (C and F) 1.5% (wt/wt).

The surface morphology of the biomimetic PA layer was further studied by atomic force microscopy (AFM) (SI appendix, Figure S7). Consistent with the SEM analysis, it revealed the disappearance of rough ridge and valley features, and it was found that the RMS of the film made with MPD concentration lower than 2% (wt/wt) was <100-nm roughness value. These RMS values ​​are lower than the typical values ​​of 100 to 300 nm often observed for classic PA films (25). When 1.5% (wt/wt) HC6 and 0.8 to 1.2% (wt/wt) MPD are used to make the bionic film, the average roughness of the total thickness of the indicator layer increases (Figure 3 AE). Consistent with the larger internal porosity and thickness observed in the SEM micrograph.

(A) 0.2-HC6, (B) 0.4-HC6, (C) 0.8-HC6, (D) 1.2-HC6 and (E) 1.8-HC6 membrane TEM cross-sectional images,% (wt/wt) MPD, 0.1 % (wt/wt) TMC and 1.5% (wt/wt) HC6 are dyed with Zn(NO3)2, showing details of the top mixed layer, which contains AWC/PA embedded in a darker PA matrix with high precision White nanoparticles.

The TEM cross-sectional image in Figure 3 also proves that by increasing the MPD concentration from 0.2% (wt/wt) to 2.0% (wt/wt) while keeping the amount of HC6 [1.5% (wt/wt)] constant, This leads to an increase in film thickness; the maximum height of the PA layer is related to the number of MPDs, with alternating thinner/thicker areas on top of the PSf support. For 0.2% and 0.4% (wt/wt) MPD (Figure 3A and B), the PA layer has a more irregular appearance, including larger parts (up to 1 μm, as in 0.2-HC6 film) with 20 to 40 nanometers The thickness of the ridge, the part with more regular ridges and valleys, and the area with higher thickness (0.2 and 0.4-HC6 up to 240 and 170 nanometers, respectively). Starting from 0.8% to 1.8% (wt/wt) MPD (Figure 3 CE), the IP layer has a more uniform ridge-valley morphology. In addition, more developed areas are being observed, with branches growing more evenly in all dimensions (with flower/rose morphology). For these films, the PA layer has a more tortuous appearance and therefore a higher surface area. It is also clear that when AWC is added, the IP layer has a complex internal structure, with specific nanoparticle domains present throughout its thickness. AWC/PA nanoparticles correspond to brighter spots, with a diameter of 15 to 20 nm in the middle of the PA layer, for 0.2% and 0.4% (wt/wt) MPD, and higher observed in the case of 0.8 to 1.8 Nanoparticle density% (wt/wt) MPD. In addition, they are more evenly distributed throughout the film thickness, from the top to the bottom of the layer. During PA formation, these low-density white areas are filled with porous nanoparticles of the AWC/PA mixture. This apparently lower layer density may result in lower transmission resistance, which partly explains the larger water permeability values ​​observed for TFC-HC6 membranes. In conclusion, compared with the reference TFC counterpart without HC6, the low concentration of MPD promoted the formation of a smoother surface, while the presence of HC6 during the IP process produced a thicker and more porous selective layer. In addition, the proper MPD loading is higher than 0.4% (wt/wt) and HC6 is lower than 1.8% (wt/wt), and the proper HC6/MPD ratio during the manufacturing process is 0.7 to 1.9 (wt/wt), allowing It is appropriate to embed self-assembled AWC/PA nanoparticles in the PA layer, assuming that it promotes faster and selective water passage during membrane filtration.

The filtration performance of biomimetic membranes was measured in BWRO and tap water reverse osmosis (TWRO) desalination (26).

First, by changing the HC6 loading from 0% (control TFC) to 1.8% (wt/wt) while maintaining the MPD concentration constant [3.4% (wt/wt)] (Figure 4A and B). By increasing the HC6 load, the membrane exhibits a gradually higher water permeability, increasing from 1.5 L⋅m−2⋅h−1⋅bar−1 to 185%, while maintaining a very low and nearly uniform salt flux up to the initial load 1.3% (wt/wt) of HC6. Above this concentration, performance degradation is observed. This result is rationalized when the HC6 concentration is up to 1.8% (wt/wt), which inhibits the formation of a defect-free selection layer during the IP process. The large error bars related to the unstable filtration performance of different samples also indicate the same use Program production. This set of filtering results is converted into A/B value, and the peak concentration is 0.9% (wt/wt) HC6. Then, we observed that by reducing the MPD concentration to 2% (wt/wt), a higher A was obtained with an average A of 4.5 L⋅m−2⋅h−1⋅bar−1 and a low salt pass rate The water penetration value of water: up to 1.5% (wt/wt) HC6 loading (empty symbols in Figure 4A and B), a true average removal rate of 99.8% of NaCl was observed.

The desalination performance of BWRO is a function of (A and B) HC6 and (C and D) MPD concentrations at 15.5 bar. Experimental conditions: the pure water of A is measured; the feed solution contains 100 mM NaCl, which is used to measure the desalination performance of B. TWRO at 6 bar. (E) Permeation flux as a function of MPD concentration and rejection observed at steady state. (F) The repulsion of different cations and anions in the steady state of 1.6-HC6 membrane was observed under TWRO conditions. The lines connecting the data points are only used as a guide for the eyes.

We know that when the IP reaction occurs, the MPD concentration controls its diffusion into the organic phase (23), but this parameter will also interfere with HC6 aggregation through the H bond. Therefore, the HC6/MPD ratio clearly determines the shape and transmission characteristics of the final layer. Taking these assumptions into account, further experiments aimed to study lower MPD concentrations, from 2% to 0.2% (wt/wt), combined with the optimal HC6 loading, fixed at 1.5% (wt/wt) (Figure 4 C and D) ). The best sample prepared with this load showed a consistent A/B value close to 35 1/bar, so it is related to the overall significant permeability-selectivity combination used for BWRO desalination, where the permeability is >3 L ·M−2⋅h−1⋅bar−1 is required. Under BWRO conditions, when the MPD concentration is 0.8% to 1.2% (wt/wt), the water permeability increases significantly and reaches a peak of ∼6.9 L⋅m−2⋅h−1⋅bar−1, while maintaining a very high value. Low relative salinity flux rate. This result shows that compared with the commercial BW30 membrane, the optimal membrane made with 3.4% (wt/wt) MPD concentration and water permeability increased by about 130%, and the water permeability increased at the equivalent A/B ratio About 360% (~3.0 L⋅m−2⋅h−1⋅bar−1 and 99.5% NaCl rejection), the observed NaCl rejection is ≥99.4%. We noticed that by using sub-stoichiometric MPD concentrations below 0.8% (wt/wt), the membrane performance dropped rapidly. The data shows that the optimal porous sponge-like AWC/PA structure is constructed and the formation of defects in the selective layer is prevented. In addition, 0.8 to 1.2% (wt/wt) MPD and 1.5% (wt/wt) HC6 are obviously the best combination, resulting in an HC6/MPD ratio of about 1.2 (wt/wt) to achieve high performance defects-free bionics membrane.

This study also aims to evaluate the performance of biomimetic membranes in TWRO seawater desalination. The water flux follows the same trend as measured under BWRO desalination, by increasing the MPD concentration from 0.8% (wt/wt) to 1.8% (wt/wt) with a constant decrease from 34 L⋅m−2⋅h−1 to 24 The optimal HC6 load of L⋅m−2⋅h−1 1.5% (wt/wt) (applied pressure is 6 bar; SI appendix, tables S5 and S6). At the same time, the observed rejection rate based on global conductivity measurements correspondingly increased from 97.5% to 99.3% (Figure 4E). These values ​​are in the same range as those measured by commercial TWRO membranes under the same experimental conditions, and in some cases are slightly higher (31 ± 1 L⋅m−2⋅h−1; 97.5% observed rejection), such as DuPont The company's TW30. In addition, compared with Na and K cations, multivalent cations including Ca2 were rejected at a higher rate of 99.8%. This is due to the size exclusion mechanism (27), which specifically indicates that these membranes have the potential for softening water applications ( Figure 4F).

In this study, an unexplored mechanical strategy was discovered that allowed the uniform incorporation of AWC from the in-situ colloidal self-assembled superstructure to identify the sponge-like particles present in the mixed PA-AWC material. During IP, MPD interacts with soft self-assembled HC6 colloidal nanoparticles through H bonds. Due to their amphiphilic properties, AWC aggregates may also help enhance the diffusion of MPD into the organic phase to react with TMC, as previously revealed (25). The formation of the mixed layer can be further produced by the interaction of amphiphilic AWC/MPD nanoparticles and nascent PA oligomers through H bonds. These interactions are strongly dependent on the HC6/MPD ratio. Under favorable conditions, they effectively promote the formation of PA and promote the gentle bonding of AWC aggregates, thereby preventing the formation of defects when solid-state nanoparticles are directly bound to PA (1, 2). In addition to the seamless in-situ adaptive integration of the distributed AWC/PA sponge-like nanostructures into the PA layer, the morphological results observed in this study indicate that the uniform formation of the unique highly porous AWC/PA structure at the interface Reduce the water transmission resistance. Selective layer with PSf support. This membrane provides the best performance in terms of water transmission and selectivity, and the porous structure does not affect the mechanical resistance or membrane performance under RO filtration, which is different from the results of PA membranes prepared with additives previously reported (26, 28 ).

It can be concluded that the improvement in transmission performance is due to the combination of higher porosity, defect-free and fast transmission through the HC6 nanostructure of the entire PA base layer. Its excellent supramolecular self-adaptive properties confirm their ability to selectively translocate. Even when combined with a hybrid PA-AWC membrane, it can remove small ions in water. In contrast, for membranes made with too low MPD concentration or too high HC6 loading, a significant loss of permeation selectivity was observed. For example, membranes manufactured with insufficient MPD exhibit a dense, more symmetrical, and extremely thin selective layer. These results indicate that there seems to be an urgent need for excess MPD to synthesize a defect-free biomimetic layer (29).

We know from literature data that the sprayed carbon nanotube layer on the polyethersulfone support before IP provides an interface that can generate a highly permeable and selective PA layer with a large effective surface area for water transport ( 30). Some similarities can explain the formation mechanism related to this study, that is, adding HC6 to the PA layer will lead to a comprehensive effect, such as the formation of a drainage layer or higher surface roughness and/or leaf morphology may be different from HC6. The affinity is related (in contrast to the use of additives, such as sodium lauryl sulfate). Nevertheless, if we look at the experimental results in the upper limit graph (Figure 5C), our membrane shows a high degree of permeability and selectivity, which is at the limit that exists between the BWRO and SWRO regions. High recovery and pollution experiments (SI appendix, Figures S8 and S9) confirmed the high quality characteristics of our biomimetic membranes. Therefore, in defining the scope/area of ​​BWRO membranes, our membranes are highly selective, more selective than other membranes prepared on a laboratory scale, including the use of nanofillers (ie zeolite, carbon nanotubes, graphite oxide Ene) thin film nanocomposites, additives (ie proton acceptors, surfactants, acids), or through the use of salts/hydrophilic additives, different solvents or co-solvents, or by changing other experimental conditions in the membrane manufacturing process (Ie pH, T or the concentration in the casting solution) for optimization (11).

Enhance penetration mechanism. (A) An ideal film (blue) with perfectly aligned channels despite the entire thickness of the PA film; (B) AWC nanoparticles are uniformly distributed in the layer. Nano-level enhanced penetration is related to the area of ​​AWCs nanoparticles (see text for details); (C) Commercial seawater reverse osmosis (SWRO) (red), brackish water reverse osmosis (BWRO) (blue) and NF membrane (green) Water permeability and selectivity. The blue dashed lines correspond to 99% and 90% NaCl repulsion, respectively, and the flux is 20 L·m-2·h-1. The yellow star highlights the selective permeability of our optimized BWRO membrane prepared using AWC, while the empty symbol represents all other laboratory-scale membranes (11). Reprinted from reference. 11. Copyright (2019), with permission from Elsevier.

Of particular interest is the potential ability of this PA film to provide a directional pathway for water transport. Here, microscopy studies have shown that this hybrid PA-AWC material is composed of AWC crystalline nanoparticles, randomly dispersed in the PA matrix (Figure 5B). For 1.2 to 1.8% (wt/wt) MPD, these particles are uniformly and densely distributed in all thicknesses of the PA layer, while for lower concentrations of 0.2 to 0.8% (wt/wt) MPD, their density is smaller and Located in the middle part of the PA layer. For sponge-like nanoparticles of 20 to 40 nm (17), theoretically, the water permeability can be estimated as high as PAWC = 131 L·m−2∙h−1∙bar−1 (SI appendix). Independently of whether microscopic examination is feasible, a high-density channel that penetrates from one side of the membrane to the other would be ideal, but we have not done this yet. This has nothing to do with the arrangement, which clearly proves to occur along the nanometer distance of the AWC crystals that show order. What is currently achieved on this nanometer scale is penetration and high-density channels. Although these nanoparticles do not merge through the micron-sized PA film (Figure 5A), they are randomly distributed in the PA layer (PPA = 1 LMH/bar). AWC crystals locally help to enhance water translocation, such as PAWC >> PPA (Figure 5B). We are not going to prove anything here in an explicit or completely mathematical way; this is just an intuitive way to explain our results regarding the experimental and unquestionable permeability of AWC-PA membranes. The hybrid PA-AWC material is reminiscent of the previous hybrid ion channel siloxanes, and provides high ion conduction (31⇓ –33) through the nano self-organization of the binding sites in the hybrid material.

The hypothesis is that in addition to 1) the adjustable formation of colloidal nanoparticles in PA, 2) the formation of internal pores/voids (due to the influence of HC6 "sandwich"), and 3) the hydrophilicity and selection in the layered structure of the drainage ditch Spongy spongy AWCs-PA (we speculate that it is also in the form of I-quartet). Therefore, we assume that this performance is achieved due to the combined effects of all these aspects. It is also true that AWC is essentially a "nano filler", but in contrast, our membrane is more selective and permeable, which may be due to the compatibility of HC6. Obviously, a high density of channels is needed to promote high permeability. Alignment is not required, but the penetration of AWC particles is important and should be optimized for water transport mainly through the channel. It should preferably have a uniform distribution of particle penetration. High-density particles without agglomeration should be obtained. Achieving these conditions is the main goal of optimization in this work.

In this work, by adding I-quartet AWC to the classic PA layer, a biomimetic membrane for low-salinity BWRO and TWRO seawater desalination was fabricated. Specifically, the performance of the hybrid AWC-PA membrane was adjusted by studying the effects of optimal AWC loading and MPD monomer concentration during IP to optimize the selective layer. It can be concluded that the self-aggregated AWC colloidal nanoparticles are combined through the supramolecular interaction with the MPD monomer, and their presence changes the IP process, thereby changing the final layer characteristics. These dynamic self-assembly processes are equivalent to adaptive colloidal entities with nascent PA oligomers.

This study illustrates the complete interaction of supramolecular aggregation and IP process 1) related to supramolecular aggregation, once the AWC load increases, the selective PA-AWC layer becomes more porous, and the AWC nano aggregates are uniformly distributed in the mixed PA , Resulting in a high-performance film. However, an excessively high concentration of AWC nano-aggregates can lead to the formation of defects, leading to the worst performance. 2) The MPD monomer concentration also has an important influence in the film manufacturing process. By adjusting this parameter, the best HC6/MPD ratio can be achieved between 0.7 and 1.9 (wt/wt) to synthesize a seamless bionic layer. Generally speaking, the lowest concentration of MPD [~0.8% (wt/wt)] is required to obtain high selectivity. However, even this concentration should not be too high to maintain a high water flux. In particular, compared with the membrane using 3.4, the membrane made with 1.2% (wt/wt) MPD and 1.5% (wt/wt) HC6 provides approximately 360% water permeability at an equivalent water-salt flux ratio Increase the% (wt/wt) MPD.

The best membrane shows excellent productivity of 110 L⋅m-2⋅h-1 and 35 L⋅m-2⋅h-1 under an applied pressure of 15.5 bar at an applied pressure of 6 bar, while maintaining high selectivity, excellent For the current BWRO and TWRO commercial membranes. In the end, we proved that the biomimetic membranes containing I-quartet AWC have the potential to improve existing low-pressure RO applications, and their composition can be adjusted to adjust their superstructure and performance for different applications.

Chemicals, film morphology and physicochemical characterization, DLS measurement, attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR), SEM, TEM and film performance evaluation are provided in the SI appendix.

All research data is included in the article and/or SI appendix.

This work was supported by Agence Nationale de la Recherche authorization number ANR-18-CE06-0004-02, WATERCHANNELS and authorization number ERANETMED 2-72-357, IDEA.

Author contributions: MB design research; MDV, AT, MD, FR and SPN research; V.-EM, SC, DC and SPN contributed new reagents/analysis tools; MDV, AT, V.-EM, SC, SPN and MB analyzed the data; MB wrote this paper.

Competitive interest statement: The results reported in this article have been submitted with a provisional patent number. FR1910152, September 2019, MDV, AT and MB

This article is directly contributed by PNAS. MK is a guest editor invited by the editorial board.

This article contains online support information https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2022200118/-/DCSupplemental.

Released under PNAS license.

Thank you for your interest in advertising on PNAS.

Note: We only ask you to provide your email address so that the people you recommend the page to know that you want them to see it, and that it is not spam. We do not capture any email addresses.

Feedback privacy/legal

Copyright © 2021 National Academy of Sciences. Online ISSN 1091-6490. PNAS is a partner of CHORUS, COPE, CrossRef, ORCID and Research4Life.