English

• 1. Introduction

  With the rapid rate of urbanization and evolution, it is imperative to promptly solve the pressing issues of water and energy shortages to ensure human survival and foster societal advancement [1-5]. Faced with these issues, there is a growing interest in recovering non-traditional water sources [6], such as municipal wastewater [7-9]. Municipal wastewater is characterized by its low strength and high potential for water resource and chemical energy recovery [10, 11]. For the treatment of municipal wastewater, conventional aerobic biological technology requires substantial energy input and emits greenhouse gases [12-14]. Alternatively, anaerobic processes, especially Anaerobic Membrane Bioreactor (AnMBR) embedded with an ultrafiltration (UF) unit, can transform organics into bioenergy and reduce sludge production while maintaining high effluent quality [15, 16]. Furthermore, nutrients in municipal wastewater, such as ammonia nitrogen and phosphorus, are almost fully retained in anaerobic effluent for subsequent recovery and utilization [17], rather than being degraded via nitrification-denitrification process [18] and resynthesized via Haber-Bosch Process [19] at high energy consumption.

  For high-grade water reuse, advanced treatment of anaerobic effluent is essential to remove residual harmful substances [20, 21]. Recently, osmotic membrane technology, including Reverse Osmosis (RO) and Forward Osmosis (FO), has been standing out as a promising advanced treatment for the UF pre-filtered anaerobic effluent (i.e., AnMBR) due to its superior rejection and stable performance [22]. In addition, while drawing freshwater through the osmotic membranes, the ammonia nitrogen and phosphorus become highly concentrated in retentate [23], significantly enhancing subsequent recovery efficiency and reducing capital expenses. Under this circumstance, integrating AnMBR with osmotic membrane processes demonstrates significant potential for simultaneously recovering resources and energy from municipal wastewater in an efficient manner [24-28]. However, osmotic membranes show unsatisfactory rejection for ammonia nitrogen due to the similarity in polarity and hydraulic radius between ammonia nitrogen and water molecules [29], harming for the produced water safety. Numerous approaches have been proposed from different perspectives to maximize the removal of ammonia nitrogen in osmotic membrane systems [30, 31]. Post-treatment units such as ion exchange [25] or chlorination procedures [27] were applied to further remove ammonia nitrogen. Conversely, pre-treatment of biochar adsorption was used to reduce the concentration of ammonia nitrogen in the feed solution (FS) prior to membrane filtration, thus reducing the transmembrane diffusion driving force of ammonia nitrogen [31]. However, these additional units inevitably led to an increase in energy consumption and investment expenses [32]. Therefore, enhancing the ammonia nitrogen rejection of osmotic membranes becomes a fundamental approach to addressing this issue at its source.

  As is well known, the membrane performance is determined by its structure [33]. Therefore, enhancing the ammonia nitrogen rejection of osmotic membranes naturally lies in constructing functional structures that specifically hinder the passage of ammonia nitrogen [30, 34]. Recently, a significant amount of work has reported that surface modification to regulate the active layer of osmotic membranes can effectively enhance the ammonia nitrogen selectivity [30, 34-38]. For instance, our previous work found that grafting polyethyleneimine (PEI) and polyamidoamine (PAMAM) dendrimers onto FO membrane surface can effectively improve the membrane selectivity towards ammonia nitrogen through the combined effect of diffusion resistance and electrostatic repulsion [30, 34, 35]. Unfortunately, there existed a trade-off between ammonia nitrogen selectivity and anti-fouling capability in the surface grafted functional layer [35]. Besides, existing research has predominantly focused on the FO platform, with a lack of exploration in the RO field. Theoretically, reconstructing the fine structure of the active layer and support layer of the osmotic membranes could also achieve the goal of hindering ammonia nitrogen diffusion across the membrane. However, due to the lack of systematically understanding of the transmembrane diffusion mechanism of ammonia nitrogen and its mapping relationship with membrane structure, it has been challenging to propose effective strategies for targeted optimization of membrane fine structure.

  Therefore, in-depth investigation of the transmembrane diffusion behavior of ammonia nitrogen under both FO and RO platform is a crucial precursor step to enhancing the ammonia nitrogen rejection of osmotic membranes. To fill this gap, this review analyzed the transmembrane diffusion mechanism of ammonia nitrogen at the microscopic level and correlated it with membrane fine structure. Firstly, by discussing existing and potential osmotic membrane-based applications in municipal wastewater reclamation processes, the need for ammonia/ammonium-rejecting membranes was underlined. Then, the systematic transmembrane diffusion process was thoroughly investigated using the Donnan-Steric Pore Model with Dielectric Exclusion (DSPM-DE) model to examine how membrane properties affect ammonia nitrogen diffusion. Lastly, specific strategies for improving ammonia nitrogen rejection by membrane modification are suggested. This work offers theoretical guidance and potential methods for the effective design of osmotic membranes with high ammonia nitrogen selectivity for the recovery of energy and resources from municipal wastewater.

  2. Osmotic membrane-based applications in municipal wastewater reclamation process

  In a world facing persistent water and energy shortages, reclaiming unconventional water sources, such as municipal wastewater, offers a viable technical solution for achieving water sustainability [39]. Among the available methods for freshwater reclamation and energy recovery from municipal wastewater, the combination of biological treatment with osmotic membranes has garnered increasing attention due to its high recovery efficiency [40, 41]. A comprehensive description of the applications that integrate biological treatment with osmotic membranes for municipal wastewater recovery are provided below.

  2.1 Current applications in municipal wastewater reclamation

  2.1.1   Water reclamation

  Membranes play a crucial role in effectively removing organics and nutrients from municipal wastewater, featuring high separation efficiency, environmental friendliness, and relatively low cost [42, 43]. Therefore, membranes for advanced treatment of municipal wastewater can successfully retain residual nutrients from bioreactor effluent [44]. Among these membranes, osmotic membranes, including RO and FO membranes, are particularly vital for producing high-quality reclaimed water with significant removal of refractory organics, pathogenic microorganisms, nitrate nitrogen, and other contaminants compared to nanofiltration (NF) membranes [45, 46]. The applications integrating osmotic membranes with biological treatment improve water recovery efficiency compared to traditional municipal wastewater treatment processes. Table S1 (Supporting information) summarizes these applications with detailed removal ratio.

  RO is an emerging membrane separation technology that demonstrates superior rejection of ammonium, nitrate, and chloride [47] compared to similar osmotic membranes like NF. As an advanced treatment process, RO is utilized for water reclamation from aerobic or anaerobic effluents of municipal wastewater (Table S1). Applications integrating RO with Membrane Bioreactor (MBR) (Fig. 1a), which combines the conventional Activated Sludge Process (ASP) with membrane separation technology, offer a promising approach to municipal wastewater reclamation [27, 48]. For instance, Qin et al. demonstrated the feasibility of producing high-grade water directly from municipal wastewater using the MBR-RO process, achieving similar or better product quality compared to the conventional ASP based process [48]. RO can also be used as advanced treatment of AnMBR effluent (Fig. 1b), achieving significant pollution rejection [25, 26, 28]. To further improve the quality of reclaimed water, some research had introduced post-processing equipment (Fig. 1c) [25, 28] with inevitably increasing energy consumption and costs. These studies underscored the feasibility of RO in recovering resources from aerobic or anaerobic effluents.

  Figure 1

  

  Figure 1.  Applications that integrate RO with biological membrane treatment. (a) MBR-RO. (b) AnMBR-RO. (c) AnMBR-RO with post-processing equipment.

  DownLoad: Full-Size Img PowerPoint

  When addressing municipal wastewater treatment, high rejection membrane processes like RO, propelled by hydraulic pressure, effectively remove non-aqueous components [49]. FO, as an osmotically driven process, is characterized by low fouling tendency and minimal energy consumption [50], exhibiting promising potential for municipal wastewater recovery (Table S1) [51]. Systems that couple FO and MBR (Fig. 2a) have been reported to perform excellently in wastewater treatment [4, 52-54]. Additionally, the integration of FO with AnMBR, in the form of the Anaerobic-Osmotic Membrane Bioreactor (An-OMBR) (Fig. 2b), is feasible for recovering energy, water, and enriching nutrients for subsequent recovery simultaneously when treating municipal wastewater [24, 55-57]. This approach achieves high treatment efficiency by concentrating organic substrates and promotes anaerobic degradation in the reactor. However, severe membrane fouling occurs due to the high solids content of the mixed liquor and digested sludge, leading to blockages of membrane pores. Alternatively, a similar configuration with FO pre-concentrated prior to anaerobic digestion has emerged platform for municipal wastewater reclamation [58]. The FO pre-concentration application allows for the simultaneous extraction of clean water for beneficial reuse whilst pre-concentrating wastewater to a higher strength suitable for anaerobic treatment. The key benefit of this configuration is that the FO membrane is in contact with concentrated wastewater, which has lower fouling propensity and fouling reversibility compared with the mixed liquor inside an An-OMBR. Besides, the volume of wastewater feeding the anaerobic bioreactor is greatly reduced, which can greatly reduce the thermal energy and physical footprint required for anaerobic treatment. Recent studies have increasingly focused on exploring the feasibility of utilizing FO processes for the direct concentration of municipal wastewater to recover resources [59, 60]. However, these studies primarily focus on wastewater concentration via FO without subsequent recovery processes. The application of the FO process in directly concentrating wastewater can yield foulant-free permeate, which could serve as pretreatment for RO or AnMBR hybrid systems [24, 45]. Although FO membranes are subjected to fouling during pre-concentration, the fouling is reversible and can be easily removed by simply flushing the membrane surface with water. FO pre-concentrate can be divided into submerged (Fig. 2c) and external configurations (Fig. 2d). To reduce the costs associated with circulating the FS through an external membrane module, the submerged configuration is more suitable for wastewater pre-concentration [61].

  Figure 2

  

  Figure 2.  Applications that integrate FO with biological membrane treatment. (a) Ae-OMBR. (b) AnOMBR. (c) Submerged FO pre-concentration. (d) External FO pre-concentration.

  DownLoad: Full-Size Img PowerPoint

  2.1.2   Organics and nutrients enrichment for energy and nutrients recovery

  Rapid global industrialization and population growth have intensified energy consumption, leading to a pressing issue of energy scarcity. One potential solution to this challenge is to recognize municipal wastewater as a valuable renewable resource [62]. Municipal wastewater contains significant chemical energy within its organic pollutants, as well as substantial freshwater resources [63]. This energy can be harnessed in the form of heat and electricity to offset the energy demands of treatment facilities [64]. Additionally, municipal wastewater contains valuable nutrients, including phosphorus and ammonia nitrogen [65]. Phosphorus is a nonrenewable resource predominantly sourced from phosphate rock. Global production of phosphorus is expected to peak around 2033, followed by a gradual decline [66]. Similarly, removing ammonia nitrogen from municipal wastewater is vital to prevent environmental pollution and enhance freshwater production [62]. While various biological nitrogen removal processes have been extensively studied [67, 68], these aerobic processes convert ammonia nitrogen to nitrogen gas. In contrast, directly recovering ammonia nitrogen from municipal wastewater rather than having it degraded [69] and resynthesized [70].

  To recover energy and nutrients from municipal wastewater, biological treatment is employed [10]. Among them, aerobic biological processes cannot effectively generate energy and emit greenhouse gases [71]. As an alternative, anaerobic biological processes are promising for energy output and nutrient reserve. However, direct anaerobic treatment is unsuitable for municipal wastewater. The relatively low organic matter content of municipal wastewater, typically indicated by a chemical oxygen demand (COD) of < 500 mg/L, results in excessive thermal energy and physical space requirements for anaerobic treatment [72]. Additionally, anaerobic treatment necessitates FS with COD levels exceeding 1000 mg/L to ensure system stability and process efficiency [73]. Notably, AnMBR stands out as a superior option for treating municipal wastewater. It achieves high sludge concentration through membrane separation and offers minimal sludge production and a reduced operating volume [74]. Building on the successful application of AnMBR in industrial wastewater treatment [75], its adoption in municipal wastewater treatment has garnered considerable attention. This approach has the potential to recover substantial amounts of energy from large quantities of municipal wastewater. To further enhance the quality of reclaimed water and increase nutrient recovery, AnMBR can be integrated with RO or FO [24-26, 28] and this feasibility has been demonstrated.

  2.2 Energy consumption calculation and comparison

  Based on the above analysis, we can conclude that coupling RO and FO membranes with AnMBR can effectively recover water resources and energy from municipal wastewater. The main reason for choosing AnMBR over MBR is that MBR cannot recover energy and retain nutrients in municipal wastewater. To further demonstrate the advantages of coupling AnMBR with RO and FO membranes, energy consumption calculations will be presented. Three applications represent typical methods for municipal wastewater treatment and detailed calculations are shown in Fig. S1 (Supporting information). As shown in Table 1, aerobic treatment requires more energy input, making it unsuitable for municipal wastewater recovery. In contrast, the AnMBR-RO process and the FO pre-concentration process have lower energy consumption. Quantitative calculations demonstrate the superiority of coupling AnMBR with RO and FO membranes in terms of energy savings.

  Table 1

  

  Table 1.  Energy consumption of applications of the process coupling biological treatment and osmotic membranes in municipal wastewater recovery.

  DownLoad: CSV

  Application	Energy consumption (kWh/m3)	Ref.
  MBR-RO	1.41	[25, 76]
  AnMBR-RO	0.37	[25]
  FO-RO+AnMBR	0.54	[25, 58]

  To summarize, a thorough description of the applications coupling biological treatment and osmotic membranes in municipal wastewater recovery is provided. Two major systems stand out for their capability to simultaneously reclamation water and recovery energy with lower energy input. Firstly, the essential components of the first configuration consist of AnMBR and RO devices. AnMBR demonstrates remarkable COD removal performance and energy output while retaining nutrients for subsequent recovery. RO efficiently extracts nutrients and produces freshwater from the AnMBR effluent due to its strong ability to reject impurities. The second configuration applies FO pre-concentration, typically with draw solution (DS) regeneration through RO. FO reduces the volume of wastewater and concentrates it for subsequent anaerobic treatment. However, AnMBR has limited capacity to retain nutrients like ammonia nitrogen and phosphorus. Their rejection primarily relies on RO and FO membranes to ensure the quality of reused water. Due to the large molecular size of phosphorus, FO and RO membranes can achieve nearly complete rejection. In contrast, the similar polarity and size between ammonia nitrogen and water molecules pose a challenge for RO and FO membranes. Therefore, to improve the efficiency of municipal wastewater recovery and meet sustainable development goals, it is crucial to enhance the rejection capability of RO and FO membranes for ammonia nitrogen.

  3. The transmembrane diffusion mechanisms of ammonia nitrogen in osmotic membranes

  The integration of AnMBR and osmotic membranes in municipal wastewater reclamation offers a promising solution for addressing water scarcity and energy shortages, due to their efficient nutrient removal and energy recovery capabilities. However, the efficiency of freshwater reclamation heavily depends on the ability of RO and FO membranes to effectively reject ammonia nitrogen from municipal wastewater. Previous studies have shown that, under typical municipal wastewater conditions (pH 6–8), standard commercial RO and FO membranes often achieve an ammonia nitrogen rejection rate of < 95% [47, 76]. In municipal wastewater treatment, which typically contains around 40 mg/L of ammonia nitrogen, approximately 20 mg/L of ammonia nitrogen remains in the treated water even when using RO or FO treatment with a 90% rejection rate and an 80% recovery efficiency [46, 77]. This residual concentration exceeds the reuse water standards in many countries, posing a significant risk to public health and environmental safety. Moreover, the concentration of ammonia nitrogen in the treated water increases with the production rate. This suboptimal performance is primarily due to the membrane structure limitations affecting ion selectivity and permeability. To address these challenges and improve the ammonia nitrogen retention capabilities of FO and RO membranes, it is crucial to investigate the mechanisms governing ammonia nitrogen transport through these membranes. Besides, a deeper understanding of relationship between membrane structure and ammonia nitrogen transmembrane diffusion behavior can lead to the development of more efficient membrane designs. For a more comprehensive analysis of this issue, we utilize the DSPM-DE model to explore how membrane surface, active layer, and support layer structures and external conditions influence ammonia nitrogen diffusion across the membrane. This research aims to provide insights for the design of RO and FO membranes with high ammonia nitrogen rejection, ultimately enhancing the efficiency of municipal wastewater reclamation.

  3.1 The theory overview of the DSPM-DE model

  To accurately describe the transmembrane diffusion of ammonia nitrogen, an appropriate mass transfer model is essential. The classical solution-diffusion (SD) model, proposed in 1965 [78], simplifies water and salt transfer processes based on certain assumptions [78, 79]. Consequently, to predict the key factors that enhance ammonia nitrogen rejection, a more robust model is required. The DSPM-DE model currently serves this purpose by describing NF performance, including steric, dielectric, and Donnan effects [80]. Table 2 compares the DSPM-DE model with the classical SD model and the solution-friction (SF) model. Among them, the DSPM-DE model offers a dynamic and comprehensive analysis of membrane properties, providing valuable guidance for designing high-performance membranes. Given that NF membranes share similarities with RO and FO membranes [81], applying the DSPM-DE model to these membranes for ammonia nitrogen diffusion is appropriate. This approach allows for a more comprehensive understanding and optimization of ammonia nitrogen rejection in RO and FO processes.

  Table 2

  

  Table 2.  Comparison between the SD model, SF model, and DSPM-DE model.

  DownLoad: CSV

  Parameter	SD model	SF model	DSPM-DE model
  Salt permeability	Constant	Varies with salt concentration and applied pressure	Varies with salt concentration and applied pressure
  Independence of salt and water transport	Independent of each other	Coupling with each other	Coupling with each other
  Mechanisms of water transport	Diffusion	Hydraulic pressure difference	Hydraulic pressure difference
  Mechanisms of salt transport inside membrane	Diffusion	Concentration difference, potential difference, and convective effect	Concentration difference, potential difference, and convective effect
  Mechanisms of salt transport at the sides of membrane	Diffusion	Seric exclusion, Donnan effect, and dielectric exclusion	Steric exclusion, Donnan effect, and dielectric exclusion
  Membrane characteristics	Static	Static	Dynamic

  The DSPM-DE model, based on the extended Nernst-Planck (ENP) equation, describes water and salt mass transfer by considering solute partitioning at the solution-membrane interfaces and transport through membrane pores. Due to the inherent complexity of these processes, simplifications are necessary for practical calculations. In this model, solute transport occurs in three steps. (1) Partitioning at the solution-membrane interface with the FS, (2) transport through nanoscale pores, and (3) partitioning at the solution-membrane interface with the permeate. These processes involve mechanisms such as steric exclusion, dielectric exclusion, and Donnan exclusion. The ENP equation accounts for diffusion, convection, and electromigration within the membrane pores. However, the DSPM-DE model does not address all aspects, such as membrane-ion interactions. The following analyses will incorporate both the mechanisms outlined in the DSPM-DE model and additional factors not currently covered. Detailed explanations of these mechanisms are available in Section S1 (Supporting information). By expanding the model's scope, we aim to provide a more comprehensive understanding of solute transport.

  3.2 The characteristics of osmotic membranes influence the transmembrane diffusion behavior of ammonia nitrogen

  Membrane characteristics are crucial in the complex transmembrane diffusion process of ammonia nitrogen, significantly influencing its migration, surface partitioning, and internal diffusion. These properties impact the diffusion through mechanisms such as steric exclusion, dielectric exclusion, and Donnan exclusion, as described by the DSPM-DE model. This part explores how the active layer surface, internal structure, and support layer of osmotic membranes directly and indirectly affect the transmembrane behavior of ammonia nitrogen.

  3.2.1   The effect of membrane surface characteristics on the transmembrane diffusion behavior of ammonia nitrogen

  The membrane surface interacts with both the FS and the permeate solution or DS. This analysis focuses on the FS interface, where surface characteristics significantly impact ammonia nitrogen diffusion and partitioning. We explore how surface charge, pore size, chemical properties, hydrophilicity, and surface morphology influence ammonia nitrogen transmembrane transport (Fig. 3a).

  Figure 3

  

  Figure 3.  The transmembrane diffusion mechanism of ammonia nitrogen. (a) The characteristics of the membrane surface. (b) The characteristics of the active layer internal.

  DownLoad: Full-Size Img PowerPoint

  Thin-film composite (TFC) membranes exhibit negative surface charges due to the deprotonation of carboxyl groups from interfacial interactions [82]. The surface charge of the membrane is influenced by the pH of the solution, with higher pH levels leading to increased deprotonation and more negative charges on the membrane surface [83-85]. These negative charges significantly influence the diffusion and partitioning of ammonia nitrogen within the FS (Fig. 3a). According to the DSPM-DE model, Donnan exclusion maintains charge balance by coupling the diffusion of positive and negative ions, enhancing the membrane's selectivity for ammonia nitrogen by restricting ion movement based on their charges [80]. Conversely, the negatively charged membrane surface attracts positively charged ammonium (NH₄⁺) of ammonia nitrogen through electrostatic attraction, promoting its accumulation and diffusion near the membrane. This creates a contrasting effect, where higher negative charge density on the membrane surface enhances ammonia nitrogen partitioning. Both Donnan exclusion and electrostatic attraction affect the transmembrane transport of ammonia nitrogen, with their relative importance needing further experimental validation. In FO processes, the reverse permeation of the DS creates different transport dynamics compared to RO [86, 87]. The uneven ion distribution within the FO membrane generates an electric field, enhancing ammonia nitrogen electromigration (Fig. 3a), resulting in lower selectivity compared to RO [88].

  Although the osmotic membranes, especially RO membranes, are generally considered nearly pore-free, research reveals that these membranes contain a network of pores and aggregated pores [89]. The relative size of solutes to these pores is critical. Larger solutes cannot pass through, whereas smaller solutes are more likely to traverse the pores. This principle underscores the importance of the steric exclusion mechanism described in the DSPM-DE model (Fig. 3a). Steric exclusion involves high energy barriers as solutes reorganize at the molecular level to fit through smaller pores [90]. Thus, membranes with pores smaller than ammonia nitrogen restrict its entry and diffusion due to steric exclusion. However, such membranes also exhibit reduced water flux for the similar size of ammonia nitrogen and water molecules. In terms of FO, the reverse migration of ions from the DS [91] hinders the forward migration of ammonia nitrogen through steric exclusion [92]. Research indicates that the obstruction of DS cations is a key factor in the lower rejection of ammonia nitrogen in FS [77]. Additionally, DS containing multivalent cations and larger hydrated radii enhances ammonia nitrogen rejection, while non-ionic DSs achieve nearly complete rejection [77].

  The extended ENP equation explains that ammonia nitrogen diffusion through the osmotic membrane involves convection, diffusion, and electromigration [93]. Convection is the ion flow driven by transmembrane resistance to water permeation, resulting in ion dissolution and transport [94]. Increased water flux enhances the dissolution and convection-driven transmembrane diffusion of ammonia nitrogen. Furthermore, higher water flux impacts the ammonia nitrogen retention rate. If ammonia nitrogen transmembrane diffusion remains constant (c_{f, i}) while water flux increases, the concentration of ammonia nitrogen on the permeate side (c_{p, i}) declines accordingly. As indicated by Eq. 1, this decreases in c_{p, i} leads to an increase in the rejection rate (R).

  $ R=1-\frac{c_{\mathrm{p}, \mathrm{i}}}{c_{\mathrm{f}, \mathrm{i}}} $

  (1)

  The hydrophilicity and surface features of the membrane directly affect water flux. Hydrophilicity is measured via contact angle assessments and is inversely related to the contact angle magnitude [95]. The TFC membrane features a distinctive ridge-valley structure [96, 97], resembling a vesicle-like form. The effective filtration area of the membrane increases with ridge structure in surface, thereby turn boosts water flux. These characteristics enhance water flux, promoting the dissolution and convection-driven transmembrane diffusion of ammonia nitrogen.

  Besides the effects of membrane surface properties on ammonia nitrogen transport, which align with the DSPM-DE model mechanisms, other properties also impact ammonia nitrogen diffusion through various mechanisms. Notably, the membrane's affinity for ammonia nitrogen significantly impacts its diffusion (Fig. 3a). Detailed descriptions are provided in Section S2 (Supporting information).

  3.2.2   The effect of active layer internal characteristic on the transmembrane diffusion behavior of ammonia nitrogen

  The internal properties of the active layer of RO and FO membranes have unique influences on the transmembrane diffusion of ammonia nitrogen. While both the surface and interior collectively contribute to overall diffusion, specific internal characteristics such as cross-linking and porosity play a significant role.

  The active layer of the osmotic membrane is a cross-linked structure generated by interfacial polymerization (IP) on the support layer [98]. The extent of cross-linking within the osmotic membrane critically determines water flux and salt rejection. The structure of cross-linking creates pores within the active layer, characterized by size and morphology (e.g., tortuosity) [99], and leads to greater steric exclusion of water molecules and ions. Studies [100] indicate that higher porosity and less tortuous pore channels facilitate the diffusion of ions and water through the membrane. Consequently, while increased cross-linking and smaller, more tortuous pores enhance ammonia nitrogen rejection (Fig. 3b), they also reduce water flux. Therefore, achieving optimal efficiency requires strategies to balance the trade-off between water flux and salt rejection [101]. Potential strategies include optimizing cross-linking density and tailoring pore structures to enhance both ammonia nitrogen rejection and water flux.

  Dielectric effects, manifesting as a reduced dielectric constant within the pores compared to the bulk solution, play a crucial role in ion transport phenomena in the DSPM-DE model [102]. This reduction results from a single layer of water molecules tightly bound to the pore wall [103, 104]. Importantly, the dielectric reduction is an inherent characteristic of the membrane, independent of solution chemistry. However, the effect of dielectric exclusion varies among different ions [105] due to differences in their solvation energies, which are influenced by ion charge and size. Higher valency and smaller ions encounter stronger dielectric exclusion effects upon entering the pores [80]. For ammonia nitrogen, dielectric exclusion imposes significant limitations on its permeation through the membrane (Fig. 3b). This is due to the unique solvation properties of ammonia nitrogen interacting with the membrane's dielectric properties. Therefore, understanding and controlling dielectric effects are essential for optimizing membrane performance in separation and filtration processes. Specific strategies may include modifying membrane materials to tailor dielectric properties or engineering pore structures to mitigate dielectric exclusion effects.

  The interaction between ions and membrane pores significantly impacts water and salt transport, a factor not adequately addressed by the DSPM-DE model. This oversight is important because these interactions can substantially affect the efficiency of membrane processes. Detailed discussions on these interactions, including viscosity and friction effects (Fig. 3b), are provided in Section S2 [106]. Specifically, viscosity effects impact ammonia nitrogen diffusion by creating resistance within the pores, while it also facilitates its distribution within the pore environment. This dual role of viscosity effects highlights the complexity of the osmotic membrane and the requirement for comprehensive models that account for such factors.

  3.2.3   The effect of support layer characteristic on the transmembrane diffusion behavior of ammonia nitrogen

  In contrast to the active layer of RO and FO membranes, the support layer is characterized by larger pores and exhibits negligible ammonia nitrogen rejection. However, the properties of the support layer, such as pore size and hydrophilicity, play a significant role in the formation of the active layer, which in turn impacts the transmembrane transport of ammonia nitrogen. Specifically, the pore size of the support layer influences the surface roughness of the active polyamide layer. Additionally, the hydrophilicity of the support layer affects water flux and causes convection. A comprehensive discussion on the support layer's characteristics and their effect on the active layer can be found in Section S3 (Supporting information).

  3.3 The characteristics of the solution and external operating conditions influence the transmembrane diffusion behavior of ammonia nitrogen

  The diffusion of ammonia nitrogen across an osmotic membrane is fundamentally driven by the ratio and concentrations of ammonia species at the membrane-solution interface. This driving force is closely linked to various characteristics of the FS, such as pH, ammonia nitrogen concentration, viscosity, and ionic strength. This section explores how changes in solution pH impact the transmembrane transport of ammonia nitrogen. A detailed discussion on the effects of ionic strength and specific operating conditions are provided in Sections S4 and S5 (Supporting information).

  The ratio of NH⁴⁺ to volatile ammonia (NH₃) in aqueous solution is pH-dependent due to the reversible reaction between this conjugate acid-base pair being sensitive to pH [107]. The predominant species of the conjugate acid-base pair of ammonia nitrogen is dependent on the relationship between pK_a (Eq. 2) and pH values. As the solution pH increases, NH⁴⁺ gradually transforms into volatile ammonia. As shown in Table S2 (Supporting information), once the pH exceeds the pK_a of 9.25, free ammonia predominates in the solution. Otherwise, ammonia nitrogen is the principal component.

  Compared to uncharged ammonia, the positively charged NH⁴⁺, which is larger in size, has more difficulty transporting through the membrane due to Donnan exclusion by a charged membrane. Therefore, the membrane rejection of ammonia species is closely related to solution pH. Furthermore, the solution pH impacts the extent of deprotonation of carboxyl groups on the surface of the TFC membrane [108]. With a higher degree of carboxyl deprotonation, there are more negative charges on the membrane surface, resulting in stronger electrostatic attraction to ammonia nitrogen. To some extent, this phenomenon facilitates the transmembrane transport of ammonia nitrogen (Eq. 2).

  $ \mathrm{NH}_4^{+} \stackrel{K_{\mathrm{a}}}{\Leftrightarrow} \mathrm{H}^{+}+\mathrm{NH}_3(\mathrm{aq})\left(\mathrm{p} K_{\mathrm{a}}=9.25\right) $

  (2)

  To summarize, we provide a comprehensive analysis of ammonia nitrogen transmembrane diffusion mechanisms with well explanation. Initially, the DSPM-DE model was introduced to describe the transport of ammonia nitrogen. Following this, we examined how different membrane layers, including surface, internal active layer, and support layer, affect ammonia nitrogen transport. The membrane surface plays a critical role in ammonia nitrogen diffusion through electrostatic properties and pore characteristics. Specifically, Donnan exclusion, electrostatic attraction, and electromigration influence ammonia nitrogen distribution. Pore size further impacts ammonia nitrogen through steric exclusion. Additionally, convection driven by water flux facilitates ammonia nitrogen transport through the membrane. In the active layer, ammonia nitrogen transport is influenced by cross-linking density and dielectric exclusion. Although the support layer has minimal direct impact on ammonia nitrogen rejection, it indirectly affects this rejection by influencing the formation and characteristics of the active layer. Furthermore, the membrane's affinity for ammonia nitrogen, an aspect not fully addressed by the DSPM-DE model, also influences selectivity. Solution properties, such as pH, affect the ammonia nitrogen form and the deprotonation degree of carboxyl groups on the TFC membrane surface. Additionally, the presence of other ions in the solution can either compete with or assist the transport of ammonia nitrogen. External operating conditions, including temperature, cross-flow velocity, and hydraulic or osmotic pressure, significantly affect ammonia nitrogen mass transfer and its ratio to water entering the membrane. The investigative results offer theoretical insights for optimizing osmotic membranes to enhance ammonia nitrogen rejection.

  4. Strategies to improve ammonia nitrogen rejection of RO and FO membrane

  In view of the unsatisfactory performance of ammonia nitrogen rejecting membranes in concentrating municipal wastewater, improving the ammonia rejection performance is crucial to the development of fresh water and nutrient recovery from wastewater. Following the discussion of ammonia nitrogen transmembrane diffusion, we summarize and propose strategies to improve ammonia nitrogen rejection in RO and FO membranes.

  4.1 Membrane modification strategies to enhance the selectivity of ammonia nitrogen

  Developing membranes with superior ammonia nitrogen selectivity is a crucial and necessary step to fundamentally reduce ammonia nitrogen transmembrane diffusion. Based on the ammonia nitrogen transport mechanism described outlined earlier, improving membrane selectivity for ammonia nitrogen involves three key strategies. These strategies include minimizing diffusion towards the membrane within the FS, reducing dispersion across the membrane surface, and inhibiting internal transport within the membrane. We will review existing membrane modification techniques and propose potential measures to enhance ammonia nitrogen rejection.

  4.1.1   Surface modification of osmotic membrane

  The surface modification strategy aims to enhance ammonia nitrogen rejection by reducing the ammonia nitrogen concentration on the membrane surface. Typically, membranes used in wastewater treatment are negatively charged within the pH range of wastewater [25, 26, 109], particularly polyamide (PA) based TFC membranes with abundant carboxyl groups on the surface [110, 111]. During the concentration process, ammonia nitrogen in the FS is electrostatically attracted to the membrane surface, increasing the surface concentration of ammonium compared to the bulk solution. Therefore, reducing the negative charges on the membrane surface can theoretically alleviate the enrichment of ammonium, thereby reducing ammonium transmembrane diffusion. Lu et al. introduced ethylenediamine (EDC) onto PA-based TFC-FO membranes through surface grafting, forming amide bonds between carboxyl groups and primary amine groups using a coupling reaction [112]. This modification increased the isoelectric point (IEP) from 3.5 to 4.2 by reducing carboxyl group density, significantly reducing forward sodium diffusion and reverse ammonium diffusion. It should be noted that the modified membrane surface did not directly face to ammonium solution and the alleviated ammonium transmembrane diffusion was achieved by Donnan dialysis-induced bidirectional diffusion of cations. Similarly, Yao et al. prepared ethylenediamine (EDA) and 2-aminoethanol (AEA)-grafted TFC-FO membranes. These modifications achieved up to 97.5% ammonia nitrogen rejection with increasing grafting times [38], as shown in Fig. S2 (Supporting information). Table S3 (Supporting information) summarizes the performance of modified membranes.

  Recently, Bao et al. comprehensively investigated the feasibility of adopting a surface amine functionalization strategy to reinforce the ammonium rejection of TFC-FO membrane for domestic wastewater concentration (Fig. 4a) [30, 34, 35]. In-situ grafting of amine-rich molecules on the membrane surface was directly conducted due to the high reactivity between the primary amine groups and the unreacted acyl chloride groups on the PA layer surface after IP of MPD and TMC. Benefiting from the abundant positively charged protonated primary (-NH³⁺) and tertiary (-NR³⁺) amines, the PEI-grafted membranes exhibited significantly enhanced surface potential and hence possessed superior ammonium rejection of up to 95.88% for the synthetic ammonium solution, compared to 70.36% for the pristine membrane (Fig. S2). Similarity, Gonzales et al. used 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (EDC) and n-hydroxysuccinimide (NHS) to activate the TFC-FO membrane surface, converting carboxyl groups of polyamide into amine-reactive NHS esters (Fig. S3 in Supporting information) [36]. After grafting PEI, quaternization of the amine functional groups of PEI was conducted with iodomethane (Fig. S3). This modification achieved ammonia nitrogen rejection of over 98.5% while maintaining similar water permeability compared to the PEI-grafted TFC membrane. The changes on the membrane surface are closely related to the molecular structure of PEI [113, 114]. With increased PEI grafting concentration, the surface potential of the grafted membrane rapidly improves and becomes positively charged. However, the surface hydrophilicity only slightly increases, which helps prevent fouling [115]. The positively charged surface enhances Donnan exclusion and electrostatic repulsion, as described in the DSPM-DE model. However, negatively charged foulants in domestic wastewater tend to encapsulate and incapacitate PEI molecules, weakening the membrane's selectivity towards ammonium.

  Figure 4

  

  Figure 4.  Strategies to improve ammonia nitrogen rejection of RO and FO membrane. (a) Membrane surface grafting modification. (b) Membrane surface secondary IP. (c) Membrane interior modification. (d) Support layer modification.

  DownLoad: Full-Size Img PowerPoint

  In order to coordinate the unsynchronized variation in surface hydrophilicity and potential of the membrane, spherically structured PAMAM dendrimers, terminated with primary amines and embedded with abundant secondary and tertiary amines, were applied for surface amine functionalization as shown in Fig. S3 [35]. Once grafted with PAMAM dendrimer, a spatially distributed amine layer forms on the membrane surface. This layer causes the migration of ammonium through the amine layer to experience electrostatic hindrance from both the terminal and embedded amines of the PAMAM dendrimer. By increasing the concentration of PAMAM dendrimer in the grafting solution, the ammonium rejection of the PAMAM-grafted membranes for synthetic ammonium solution was further enhanced to 98.81% [35]. In addition, the abundant primary amines also greatly improved the hydrophilicity of the membrane surface, which in turn significantly improved the anti-fouling capacity of the PAMAM-grafted membrane. Consequently, a superior ammonium rejection of 91.81% for domestic wastewater (50% volume reduction) was achieved by the PAMAM-grafted membrane [35]. Moreover, The PAMAM dendrimer layer effectively isolates the membrane active layer from the wastewater and provides an ammonium rejecting channel that preferentially transports the foulant-free aqueous solution. The applicability of the PAMAM grafted ammonium rejecting membrane in the long-term concentration of domestic wastewater was also confirmed [35]. In summary, the surface amine functionalization strategy, especially PAMAM dendrimer grafting, proved to be a feasible approach for preparing ammonium rejecting membranes for wastewater concentration.

  The aforementioned studies concentrate on surface grafting modification of TFC-FO membranes through electrostatic repulsion and diffusion resistance [30, 34, 35], while others enhanced selectivity through steric exclusion [37]. One such method is secondary IP, where a tris(2-aminoethyl)amine (TAEA) solution is applied to the TFC-FO membrane after the initial IP process (Fig. 4b). This secondary IP process introduces amine groups that react with unreacted carboxyl groups of TMC, resulting in a modified membrane with reduced carboxyl density and increased cross-linking. As a result, the TAEA-modified membrane shows decreased negative charge and an observed isoelectric point at a higher pH compared to the original TFC membrane. This modification leads to a significant enhancement in ammonia nitrogen retention, rising from 64.92% to 94.74%. Mechanistic studies suggest that the improved rejection is primarily due to size-sieving mechanisms.

  Dedicated research aimed at enhancing the rejection of ammonia nitrogen in RO membranes has not yet been conducted. However, insights can be gleaned from the modification strategies employed for FO membranes. To enhance the rejection of ammonia nitrogen in RO membranes, a surface modification strategy focusing on reducing the concentration of ammonium on the membrane surface can be employed. This reduction is achieved by increasing the diffusion resistance of ammonium from the bulk solution to the membrane surface, primarily through electrostatic repulsion exerted by the membrane surface on the ammonium ions. Consequently, introducing a substantial number of positive charges onto the membrane surface is expected to inhibit the accumulation of ammonium on the membrane surface and thereby increase ammonium rejection by the membrane. According to the works reported by Bao et al. [35, 116], it can be speculated that constructing a highly permeable (e.g., dendritic) and hierarchically distributed charge layer on the membrane surface, in which functional groups with strong charge ability are sealed inside by hydrophilic and weakly charged functional groups, is a very promising method for increasing the membrane rejection of ammonium. Moreover, through secondary IP residual carboxyl groups present on the membrane surface can be masked, while the density of amine groups can be increased. This approach results in a reduction of negative charge on the membrane surface and weakens the diffusion of ammonia nitrogen towards the membrane surface due to the higher amine density. Furthermore, there is a certain degree of enhancement in membrane cross-linking.

  4.1.2   Active layer modification of osmotic membrane

  The size of ammonia nitrogen is extremely small and very similar to water molecules, increasing the density or cross-linking degree of the active layer to increase the selectivity of the membrane to ammonia nitrogen will greatly reduce the water flux. Therefore, increasing the interaction (i.e., repulsion or adsorption) is more appropriate to reduce the diffusion capacity of ammonia within the active layer (Fig. 4c). Modification of the membrane active layer to positively charged can provide a much longer repelling distance for ammonium diffusion through the active layer than the surface modification strategy. Theoretically speaking, preparing an active layer with alternative monomers to introduce positive charges throughout the active layer will influent the interaction between ions and membrane pores and greatly hinder ammonium diffusion. In addition, for membranes that can be used for wastewater concentration, the active layer has a high retention capacity, and foulants cannot enter the active layer and damage the rejection capacity for ammonia nitrogen. However, maintaining salt rejection is challenging due to the weakened cross-linking degree that is derived from the loss of the binding site of the additional charges [117]. Incorporating ammonia nitrogen reject molecules or nanoparticles with water channels within the active layer seems to be a promising solution. Future work should be conducted to investigate the feasibility of this strategy. Conversely, modifying the active layer to possess ammonium adsorption properties, such as doping with zeolite nanoparticles [118], can also increase the retention of ammonium by the membrane. But unlike the adsorptive membrane, the membrane used for wastewater concentration needs a dense active layer to provide high-quality effluent for subsequent water recovery, which makes the ammonia nitrogen desorption process inconvenient. Besides, since the adsorption of ammonium adsorbent is mainly based on the ion-exchange mechanism [119], the substituted cations by ammonium in adsorbent would diffuse into permeate and may contaminate the effluent (e.g., the produced fresh water from RO). However, since there are no related works have been reported, the above-mentioned strategies for the modification of the membrane active layer are only speculated based on theory, and the feasibility needs to be demonstrated by subsequent experiments.

  4.1.3   Support layer modification of osmotic membrane

  The support layer provides essential mechanical support to the active layer. For support layer modification strategy, depending on its application, it can be modified with the ability to repel or absorb ammonia nitrogen (Fig. 4d). There are three primary approaches for modifying the support layer. Firstly, introducing functional groups onto the support layer's surface can form covalent bonds with the active layer, resulting in TFC membranes with enhanced ammonia nitrogen barrier performance and structural stability. Secondly, plasma treatment with hydrophilic monomers can alter the active layer's structure, allowing an exploration of its impact on ammonia nitrogen retention. Thirdly, and most commonly, incorporating materials such as nanoparticles into the support layer can endow it with the capability to adsorb or repel ammonia nitrogen. These three modification strategies collectively enhance the ammonia nitrogen retention effectiveness of the membrane. Otherwise, the support layer acts as a post-treatment for ammonia nitrogen remaining in the permeate. However, given the large pores of the substrate and insufficient contact of the aqueous solution with the pore walls, the adsorption rate, adsorption capacity, and regeneration might be the limiting factors.

  The design of permeation membranes with high ammonia nitrogen retention capacity is addressed as outlined in Table 3. Additionally, the practical feasibility of these modification strategies must be evaluated. Firstly, production costs should be kept within an acceptable range. Secondly, the membrane's operational durability must be assessed to ensure long-term stability and performance without degradation. Finally, it is crucial to verify whether the modification process generates hazardous waste and complies with environmental regulations. Evaluating the feasibility of membrane modification strategies requires consideration of not only technical aspects but also practical factors such as cost, sustainability, and environmental impact. Besides, faced with high-strength wastewater, such as industrial wastewater, the performance of modified osmotic membranes addresses challenges such as server membrane fouling and concentration polarization.

  Table 3

  

  Table 3.  Summary and comparison of modification strategies, mechanisms, and challenges.

  DownLoad: CSV

  Membrane structures	Modification strategies	Mechanisms of ammonia nitrogen rejection	Challenge
  Surface of active layer	Grafting, chemical coupling, etc.	Decrease the partition of ammonia nitrogen	Degradation of the constructed function layer
  Insides of active layer	Co-solvent IP	Repel or absorb ammonia nitrogen	Adsorbate release and adsorbent regeneration
  Support layer	Surface functionalization, plasma treatment, inorganic particles doping	Repel or absorb ammonia nitrogen	Mechanical strength, adsorbate release and adsorbent regeneration

  4.2 Operational conditional optimizing strategies to enhance the selectivity of ammonia nitrogen

  The influence of operational conditions on the transmembrane transport of ammonia nitrogen is highlighted. Specifically, pH impacts the chemical forms of ammonia nitrogen and the deprotonation level of carboxyl groups on the membrane surface. Additionally, other ions in the solution can either facilitate or compete with ammonia nitrogen transport. Factors such as turbulence on the membrane surface and temperature also play a role in affecting the properties of the solution or ammonia nitrogen. Manipulating these operational conditions allows for the regulation of ammonia nitrogen transmembrane transport.

  As mentioned earlier, ammonia nitrogen exists in aqueous solution mainly in neutral free ammonia and negatively charged ammonium forms. Benefiting from the Donnan exclusion by a charged membrane, the membrane rejection for electroneutral ammonia is significantly lower than that for negatively charged ammonium. Pre-treating the FS to lower pH will contribute to not only transforming ammonia into ammonium ions but also leading to a less negatively charged membrane surface by decreasing the deprotonation of carboxylate groups (pK_a = 5.2). Both the variations of the ammonia speciation and membrane surface charge under lower pH would be beneficial for ammonium rejection [38]. However, drastically lowering the pH of wastewater requires the addition of a large amount of chemicals, which will lead to a significant increase in the cost of wastewater treatment. Similarly, adding agents to FS to adjust the ionic composition and inhibit the diffusion of ammonia nitrogen consumes excess cost.

  Although ammonia nitrogen rejection can be enhanced by optimizing operation conditions (Section S6 in Supporting information), the accompanying higher operating costs make it less economical compared to the limited increase in rejection. Besides, the removal of ammonia in the feed side can incipiently mitigate the ammonia transmembrane diffusion, but it would reduce the ammonia enrichment and hence make it less economically and technically feasible for subsequent recovery.

  In summary, enhancing the selectivity of osmotic membranes towards ammonia nitrogen rejection can be achieved through various modifications to the membrane surface, active layer, and support layer, either individually or in combination. Surface modifications typically focus on constructing functional structures that prevent the accumulation of ammonia nitrogen via electrostatic repulsion, Donnan exclusion, and steric exclusion. However, balancing ammonia nitrogen selectivity with anti-fouling capability remains a critical challenge. For the active layer, modifications including nanomaterial doping can be employed to absorb or repel ammonia nitrogen, although increasing cross- linking density might reduce water flux. Similar strategies can be applied to the support layer to enhance its functionality. Additionally, operational strategies like adjusting the pH of the FS can further improve ammonia nitrogen rejection. However, these approaches pose economic and technical challenges for subsequent recovery. Overall, these strategies are essential for advancing water and energy recovery from municipal wastewater, highlighting the need for further research to develop innovative and economically viable strategies.

  5. Conclusion and perspective

  Recovering water resources and energy from municipal wastewater is an effective approach to addressing global water and energy scarcity. Integrating osmotic membrane, including RO and FO membranes, with AnMBR offers the potential for low energy consumption and high resource recovery efficiency. In these integrated applications, RO membranes are used for advanced treatment of AnMBR effluent to improve reused water quality, while FO membranes act as a pre-treatment unit to concentrate the wastewater and reduce the volume on the AnMBR.

  To optimize these applications, enhancing the osmotic membrane's selectivity for ammonia nitrogen, a key component in municipal wastewater, is crucial for improving recovery efficiency. This review analyzes the transmembrane diffusion mechanism of ammonia nitrogen based on the DSPM-DE model, exploring how membrane characteristics affect this process. Initially, the membrane surface's charge and pore size determine the ammonia nitrogen diffusion and distribution on surface through mechanism including Donnan exclusion, electrostatic attraction, and steric exclusion, and its hydrophilicity and morphology influence water flux and convection-driven transmembrane diffusion. Secondly, the cross-linking structure governs pore morphology and size, affecting ammonia nitrogen selectivity through steric exclusion, and the reduced dielectric constant of the water monolayer on the pore walls contributes to dielectric exclusion. Notably, the active layer's affinity for ammonia nitrogen significantly influences its transmembrane behavior, a factor not accounted for in the DSPM-DE model. Subsequently, the support layer has minimal direct impact on ion selectivity but influences ammonia nitrogen diffusion indirectly by affecting the formation of the active layer. Aside from the effect of membrane, external conditions, such as pH, primarily affect the proportion of ammonia and ammonia nitrogen, thereby impacting transmembrane behavior.

  By summarizing the key mechanisms influencing ammonia nitrogen transmembrane behavior, we have developed a comprehensive understanding of the impact of membrane micro/nanostructures. Based on these insights, several strategies are proposed for membrane modification to enhance ammonia nitrogen rejection. Firstly, the most widely recognized approach is constructing functional layers on the membrane surface and reduce ammonia nitrogen diffusion through the mechanisms such as electrostatic repulsion and diffusion resistance. Secondly, modifying active layer, such as doping with nanomaterials, can enhance adsorption or repulsion of ammonia nitrogen and improve rejection efficiency. The third strategy utilize the support layer as a post-treatment process and provide additional adsorption of ammonia nitrogen or resistance to its passage.

  From this review, several prospects are summarized to guide future studies.

  (ⅰ) Currently, with the main products being residual sludge, carbon dioxide, and nitrogen, the efficiency of energy and resource recovery from municipal wastewater is low. To mitigate the environmental consequences of wastewater treatment, it is imperative to transition municipal wastewater treatment from a linear economic model to a circular economic model. Hence, it is crucial to investigate solutions that offer both environmental and economic sustainability for future urban development. In order to tackle the issues, the unique A-B approach provides useful insights for the reuse of municipal wastewater. The A stage utilizes anaerobic units to immediately collect organic matter and greatly decrease sludge generation, whereas the B stage is generally employed for the biological, physical, or chemical retrieval of nutrients. The application discussed in this paper exemplifies the A-B treatment methodology. In the future, we can explore and create more applications that meet the requirements of the A-B model to achieve the circular economy of municipal wastewater treatment.

  (ⅱ) To explain ammonia nitrogen transport through osmotic membranes, we employed the DSPM-DE model, which advances beyond the traditional SD model by incorporating mechanisms like size exclusion, dielectric exclusion, and Donnan exclusion. However, these mechanisms are insufficient to fully describe ion transmembrane diffusion, as factors such as the affinity between ions and membranes are overlooked. Future research should develop more comprehensive models to accurately capture the complexities of ion diffusion across osmotic membranes.

  (ⅲ) Enhancing the performance of permeable membranes through modification is crucial. Chemical surface modifications, such as grafting and chemical coupling, can improve ammonia and nitrogen retention. For the membrane interior, exploring the feasibility of Co-solvent IP and developing new doped materials can advance technical feasibility. Regarding the support layer, new materials should be considered to ensure adequate mechanical strength while achieving effective salt retention and water permeability.

  (ⅳ) This review highlights the need to improve the ability of osmotic membranes to reject ammonia nitrogen. Furthermore, addressing both high ammonia nitrogen rejection and anti-fouling performance is critical to guarantee the sustainable functioning of membranes. To accomplish this, the membrane's surface roughness can be minimized and its hydrophilicity improved, alongside enhancing the capacity to reject ammonia nitrogen.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Yujie Xie: Writing – original draft, Methodology, Data curation, Conceptualization. Kexin Yuan: Writing – review & editing, Methodology, Data curation, Conceptualization. Beiyang Luo: Writing – original draft, Methodology, Data curation, Conceptualization. Haoran Feng: Writing – original draft, Methodology, Data curation, Conceptualization. Xian Bao: Writing – review & editing, Supervision, Project administration, Methodology. Jun Ma: Writing – review & editing, Validation, Supervision.

  Acknowledgments

  This work was supported by National Natural Science Foundation of China (No. 52200051), Harbin Institute of Technology (No. HC202236), and Outstanding Youth Fund of Heilongjiang Natural Science Foundation (No. YQ2023E021),

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110443.

  
  
• 1. [1]

     T. Hülsen, E.M. Barry, Y. Lu, et al., Water Res. 100 (2016) 486–495.

  2. [2]

     K. Kümmerer, D.D. Dionysiou, O. Olsson, D. Fatta-Kassinos, Science 361 (2018) 222–224 (1979).

  3. [3]

     W.H. Luo, H.V. Phan, M. Xie, et al., Water Res. 109 (2017) 122–134.

  4. [4]

     G.L. Qiu, Y.M. Law, S. Das, Y.P. Ting, Environ. Sci. Technol. 49 (2015) 6156–6163. doi: 10.1021/es504554f

  5. [5]

     A. Zirehpour, A. Rahimpour, A.A. Shamsabadi, et al., Environ. Sci. Technol. 51 (2017) 5511–5522. doi: 10.1021/acs.est.7b00782

  6. [6]

     D. Wang, S. Chen, S. Lai, et al., Chin. Chem. Lett. 34 (2023) 107861.

  7. [7]

     G.T. Daigger, Water Environ. Res. 81 (2009) 809–823.

  8. [8]

     H. Furumai, Phys. Chem. Earth 33 (2008) 340–346.

  9. [9]

     L.R. Parsons, B. Sheikh, R. Holden, D.W. York, HortScience 45 (2010) 1626–1629. doi: 10.21273/hortsci.45.11.1626

  10. [10]

      D.J. Batstone, T. Hülsen, C.M. Mehta, J. Keller, Chemosphere 140 (2015) 2–11.

  11. [11]

      P.R. Rout, M.K. Shahid, R.R. Dash, et al., J. Environ. Manage. 296 (2021) 113246.

  12. [12]

      T. Pan, X.D. Zhu, Y.P. Ye, Ecol. Eng. 37 (2011) 248–254.

  13. [13]

      I. Sharawat, R. Dahiya, R.P. Dahiya, Int. J. Environ. Sci. Technol. 18 (2021) 871–884. doi: 10.1007/s13762-020-02893-9

  14. [14]

      J.W. Tang, C.H. Zhang, X.L. Shi, et al., J. Environ. Manage. 234 (2019) 396–403.

  15. [15]

      E. Ranieri, G. D'Onghia, L. Lopopolo, et al., J. Environ. Manage. 337 (2023) 117767.

  16. [16]

      J.D. Vela, L.B. Stadler, K.J. Martin, et al., Environ. Sci. Technol. Lett. 2 (2015) 234–244.

  17. [17]

      C.M. Mehta, W.O. Khunjar, V. Nguyen, et al., Crit. Rev. Environ. Sci. Technol. 45 (2015) 385–427. doi: 10.1080/10643389.2013.866621

  18. [18]

      L. Miao, G.Q. Yang, T. Tao, Y.Z. Peng, J. Environ. Manage. 235 (2019) 178–185.

  19. [19]

      M. Wang, M.A. Khan, I. Mohsin, et al., Energy Environ. Sci. 14 (2021) 2535–2548. doi: 10.1039/d0ee03808c

  20. [20]

      E. Debik, G. Kaykioglu, A. Coban, I. Koyuncu, Desalination 256 (2010) 174–180.

  21. [21]

      V. Stazi, M.C. Tomei, Sci. Total Environ. 635 (2018) 78–91.

  22. [22]

      X. Zhang, Y. Liu, Chem. Eng. J. 421 (2021) 127773.

  23. [23]

      G. Naidu, L. Tijing, M.A. Johir, et al., J. Membr. Sci. 599 (2020) 117832.

  24. [24]

      L. Chen, Y.S. Gu, C.Q. Cao, et al., Water Res. 50 (2014) 114–123. doi: 10.4103/0973-1482.145816

  25. [25]

      J. Gu, H. Liu, S.Y. Wang, et al., J. Clean. Prod. 230 (2019) 1287–1293.

  26. [26]

      H. Liu, J. Gu, S. Wang, et al., J. Membr. Sci. 593 (2020) 117442.

  27. [27]

      S. Wang, H. Liu, J. Gu, et al., Chem. Eng. J. 375 (2019) 122078.

  28. [28]

      S. Wang, H. Liu, J. Gu, et al., Chemosphere 287 (2022) 132060.

  29. [29]

      Q. He, T. Tu, S. Yan, et al., Sep. Purif. Technol. 191 (2018) 182–191.

  30. [30]

      X. Bao, Q. She, W. Long, Q. Wu, Water Res. 190 (2021) 116678.

  31. [31]

      X. Zhang, J. Gu, Y. Liu, Water Res. 211 (2022) 118058.

  32. [32]

      Z. Zhuo, E. Du, N. Zhang, et al., Nat. Commun. 13 (2022) 3172.

  33. [33]

      B.S. Lalia, V. Kochkodan, R. Hashaikeh, N. Hilal, Desalination 326 (2013) 77–95.

  34. [34]

      X. Bao, Q.L. Wu, W.X. Shi, et al., J. Membr. Sci. 573 (2019) 135–144.

  35. [35]

      X. Bao, Q. Wu, W. Shi, et al., J. Membr. Sci. 584 (2019) 9–19.

  36. [36]

      R.R. Gonzales, Y. Sasaki, T. Istirokhatun, et al., Sep. Purif. Technol. 297 (2022) 121534.

  37. [37]

      J. Li, R.R. Gonzales, R. Takagi, et al., Desalination 541 (2022) 116002.

  38. [38]

      X. Yao, R.R. Gonzales, Y. Sasaki, et al., J. Membr. Sci. 650 (2022) 120429.

  39. [39]

      S. Manikandan, R. Subbaiya, M. Saravanan, et al., Chemosphere 289 (2022) 132867.

  40. [40]

      G.S. Arcanjo, C.R. Dos Santos, B.F. Cavalcante, et al., Chemosphere 301 (2022) 134716.

  41. [41]

      Y. Ye, H.H. Ngo, W. Guo, et al., Chemosphere 289 (2022) 133175.

  42. [42]

      D.L. Zhao, S. Japip, Y. Zhang, et al., Water Res. 173 (2020) 115557.

  43. [43]

      F. Wang, Z. Zhang, I. Shakir, et al., Adv. Sci. 9 (2022) 2103814.

  44. [44]

      H. Feng, K. Yuan, Y. Liu, et al., Chem. Eng. J. (2023) 145580.

  45. [45]

      S. Vinardell, S. Astals, J. Mata-Alvarez, J. Dosta, Bioresour. Technol. 297 (2020) 122395.

  46. [46]

      K. Arola, B. Van der Bruggen, M. Mänttäri, M. Kallioinen, Crit. Rev. Environ. Sci. Technol. 49 (2019) 2049–2116. doi: 10.1080/10643389.2019.1594519

  47. [47]

      K. Häyrynen, E. Pongrácz, V. Väisänen, et al., Desalination 240 (2009) 280–289.

  48. [48]

      H.J. Qin, K.A. Kekre, G.H. Tao, et al., J. Membr. Sci. 272 (2006) 70–77.

  49. [49]

      N.A. Khan, S. Singh, E.A. López-Maldonado, et al., Desalination 565 (2023) 116873.

  50. [50]

      M. Xu, P. Zhao, C.Y. Tang, et al., Chin. Chem. Lett. 33 (2022) 3818–3822.

  51. [51]

      R.V. Linares, Z. Li, S. Sarp, et al., Water Res. 66 (2014) 122–139.

  52. [52]

      A. Achilli, T.Y. Cath, E.A. Marchand, A.E. Childress, Desalination 239 (2009) 10–21.

  53. [53]

      G.L. Qiu, Y.P. Ting, Bioresour. Technol. 150 (2013) 287–297.

  54. [54]

      X.H. Wang, Y.X. Zhao, B. Yuan, et al., Bioresour. Technol. 202 (2016) 50–58. doi: 10.1111/1744-7917.12226

  55. [55]

      Y. Gao, Z. Fang, C. Chen, et al., Bioresour. Technol. 307 (2020) 123254.

  56. [56]

      Y.S. Gu, L. Chen, J.W. Ng, et al., J. Membr. Sci. 490 (2015) 197–208.

  57. [57]

      M.K.Y. Tang, H.Y. Ng, Water Sci. Technol. 69 (2014) 2036–2042. doi: 10.2166/wst.2014.116

  58. [58]

      Z.L. Cheng, X. Li, T.S. Chung, J. Membr. Sci. 559 (2018) 63–74.

  59. [59]

      Y. Gao, Z. Fang, P. Liang, X. Huang, Bioresour. Technol. 247 (2018) 730–735.

  60. [60]

      Y. Sun, J.Y. Tian, Z.W. Zhao, et al., Water Res. 104 (2016) 330–339.

  61. [61]

      S.Q. Zou, H.Y. Yuan, A. Childress, Z. He, Environ. Sci. Technol. 50 (2016) 6827–6829. doi: 10.1021/acs.est.6b02849

  62. [62]

      Y.Y. Ye, H.H. Ngo, W.S. Guo, et al., Bioresour. Technol. 268 (2018) 749–758.

  63. [63]

      Y. Sun, Z. Chen, G. Wu, et al., J. Clean. Prod. 131 (2016) 1–9.

  64. [64]

      P.L. McCarty, J. Bae, J. Kim, Environ. Sci. Technol. 45 (2011) 7100–7106. doi: 10.1021/es2014264

  65. [65]

      M.E. Martínez, S. Sánchez, J.M. Jiménez, et al., Bioresour. Technol. 73 (2000) 263–272.

  66. [66]

      K. Arola, M. Mänttäri, M. Kallioinen, Sep. Purif. Technol. 256 (2021) 117255.

  67. [67]

      P.L. McCarty, Environ. Sci. Technol. 52 (2018) 3835–3841. doi: 10.1021/acs.est.7b05832

  68. [68]

      M.K.H. Winkler, L. Straka, Curr. Opin. Biotechnol. 57 (2019) 50–55.

  69. [69]

      J.J. Wang, B.C. Huang, J. Li, R.C. Jin, Chin. Chem. Lett. 31 (2020) 2567–2574.

  70. [70]

      J. Humphreys, R. Lan, S. Tao, Adv. Energy Sustain. Res. 2 (2021) 2000043.

  71. [71]

      H. Liu, S. Qin, A. Li, et al., Sci. Total Environ. 859 (2023) 160183.

  72. [72]

      Z. Lei, S.M. Yang, Y.Y. Li, et al., Bioresour. Technol. 267 (2018) 756–768.

  73. [73]

      A.J. Ansari, F.I. Hai, W.S. Guo, et al., Sci. Total Environ. 566 (2016) 559–566.

  74. [74]

      J. Wu, Z. Kong, Z. Luo, et al., Water Res. 207 (2021) 117783.

  75. [75]

      A. Aslam, S.J. Khan, H.M.A. Shahzad, Sci. Total Environ. 802 (2022) 149612.

  76. [76]

      M. Li, Z. Lv, J. Zheng, et al., ACS Sustain. Chem. Eng. 5 (2017) 784–792. doi: 10.1021/acssuschemeng.6b02119

  77. [77]

      K. Almoalimi, Y.Q. Liu, J. Membr. Sci. 648 (2022) 120365.

  78. [78]

      D.R. Paul, J. Membr. Sci. 241 (2004) 371–386.

  79. [79]

      G. Mauviel, J. Berthiaud, U. Vallieres, et al., J. Membr. Sci. 266 (2005) 62–67.

  80. [80]

      R. Wang, S. Lin, J. Membr. Sci. 620 (2021) 118809.

  81. [81]

      Y. Yoon, R.M. Lueptow, J. Membr. Sci. 261 (2005) 76–86.

  82. [82]

      A. Tiraferri, M. Elimelech, J. Membr. Sci. 389 (2012) 499–508.

  83. [83]

      Q. Li, H. Liu, B. He, et al., J. Membr. Sci. 641 (2022) 119880.

  84. [84]

      Q. Li, H. Liu, Y. Ji, et al., Desalination 535 (2022) 115825.

  85. [85]

      M. Mänttäri, A. Pihlajamäki, M. Nyström, J. Membr. Sci. 280 (2006) 311–320.

  86. [86]

      N.T. Hancock, T.Y. Cath, Environ. Sci. Technol. 43 (2009) 6769–6775. doi: 10.1021/es901132x

  87. [87]

      G.J. Irvine, S. Rajesh, M. Georgiadis, W.A. Phillip, Environ. Sci. Technol. 47 (2013) 13745–13753. doi: 10.1021/es403581t

  88. [88]

      F.X. Kong, L.Q. Dong, T. Zhang, et al., Desalination 437 (2018) 144–153.

  89. [89]

      S.H. Kim, S.Y. Kwak, T. Suzuki, Environ. Sci. Technol. 39 (2005) 1764–1770. doi: 10.1021/es049453k

  90. [90]

      R.R. Sharma, R. Agrawal, S. Chellam, J. Membr. Sci. 223 (2003) 69–87.

  91. [91]

      H. Liu, B. Li, P. Zhao, et al., Chin. Chem. Lett. 34 (2023) 108369.

  92. [92]

      J. Xu, T.N. Tran, H. Lin, N. Dai, Water Res. 185 (2020) 116255.

  93. [93]

      L. Wang, T.C. Cao, J.E. Dykstra, et al., Environ. Sci. Technol. 55 (2021) 16665–16675. doi: 10.1021/acs.est.1c05649

  94. [94]

      P.M. Biesheuvel, L. Zhang, P. Gasquet, et al., Environ. Sci. Technol. Lett. 7 (2020) 42–47. doi: 10.1021/acs.estlett.9b00686

  95. [95]

      Y. Baek, J. Kang, P. Theato, J. Yoon, Desalination 303 (2012) 23–28.

  96. [96]

      F. Pacheco, R. Sougrat, M. Reinhard, et al., J. Membr. Sci. 501 (2016) 33–44.

  97. [97]

      F.A. Pacheco, I. Pinnau, M. Reinhard, J.O. Leckie, J. Membr. Sci. 358 (2010) 51–59.

  98. [98]

      L.W. Huang, J.R. McCutcheon, J. Membr. Sci. 483 (2015) 25–33.

  99. [99]

      F.L. Usseglio-Viretta, D.P. Finegan, A. Colclasure, et al., J. Electrochem. Soc. 167 (2020) 100513. doi: 10.1149/1945-7111/ab913b

  100. [100]

       K. Yuan, Y. Liu, H. Feng, et al., Chin. Chem. Lett. 35 (2024) 109022.

  101. [101]

       Z. Zhang, J. Hu, S. Liu, et al., Chin. Chem. Lett. 32 (2021) 2882–2886.

  102. [102]

       T.R. Nickerson, E.N. Antonio, D.P. McNally, et al., Chem. Sci. 14 (2023) 751–770. doi: 10.1039/d2sc04920a

  103. [103]

       W.R. Bowen, J.S. Welfoot, Chem. Eng. Sci. 57 (2002) 1121–1137.

  104. [104]

       H.C. Zhu, F.R. Yang, Y.J. Zhu, et al., RSC Adv. 10 (2020) 8628–8635. doi: 10.1039/c9ra09399k

  105. [105]

       A.E. Yaroshchuk, Adv. Colloid Interface Sci. 85 (2000) 193–230.

  106. [106]

       R. Epsztein, R.M. DuChanois, C.L. Ritt, et al., Nat. Nanotechnol. 15 (2020) 426–436. doi: 10.1038/s41565-020-0713-6

  107. [107]

       J.J. Qin, M.H. Oo, M.N. Wai, F.S. Wong, J. Membr. Sci. 217 (2003) 261–268.

  108. [108]

       H. Zhang, B. Li, J. Pan, et al., J. Membr. Sci. 539 (2017) 128–137.

  109. [109]

       M.F. Tay, C. Liu, E.R. Cornelissen, et al., Water Res. 129 (2018) 180–189.

  110. [110]

       M. Elimelech, W.H. Chen, J.J. Waypa, Desalination 95 (1994) 269–286.

  111. [111]

       R.J. Petersen, J. Membr. Sci. 83 (1993) 81–150.

  112. [112]

       X.L. Lu, C. Boo, J. Ma, M. Elimelech, Environ. Sci. Technol. 48 (2014) 14369–14376. doi: 10.1021/es504162v

  113. [113]

       Y.F. Huang, X.S. Feng, J. Membr. Sci. 586 (2019) 53–83.

  114. [114]

       Y. Song, Y. Wang, N. Zhang, et al., J. Membr. Sci. 630 (2021) 119332.

  115. [115]

       S. Zhao, M. Golestani, A. Penesyan, et al., Chin. Chem. Lett. 31 (2020) 851–854.

  116. [116]

       X. Bao, Q.L. Wu, W.X. Shi, et al., Water Res. 153 (2019) 1–10.

  117. [117]

       K.F. Gu, S.H. Wang, Y.H. Li, et al., J. Membr. Sci. 581 (2019) 214–223.

  118. [118]

       M.R. Adam, T. Matsuura, M.H.D. Othman, et al., Process Saf. Environ. Prot. 122 (2019) 378–385.

  119. [119]

       M.R. Adam, M.H.D. Othman, R. Abu Samah, et al., Sep. Purif. Technol. 213 (2019) 114–132.

• 

  Figure 1  Applications that integrate RO with biological membrane treatment. (a) MBR-RO. (b) AnMBR-RO. (c) AnMBR-RO with post-processing equipment.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Applications that integrate FO with biological membrane treatment. (a) Ae-OMBR. (b) AnOMBR. (c) Submerged FO pre-concentration. (d) External FO pre-concentration.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  The transmembrane diffusion mechanism of ammonia nitrogen. (a) The characteristics of the membrane surface. (b) The characteristics of the active layer internal.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Strategies to improve ammonia nitrogen rejection of RO and FO membrane. (a) Membrane surface grafting modification. (b) Membrane surface secondary IP. (c) Membrane interior modification. (d) Support layer modification.

  下载: 全尺寸图片 幻灯片

  Table 1.  Energy consumption of applications of the process coupling biological treatment and osmotic membranes in municipal wastewater recovery.

  Application	Energy consumption (kWh/m3)	Ref.
  MBR-RO	1.41	[25, 76]
  AnMBR-RO	0.37	[25]
  FO-RO+AnMBR	0.54	[25, 58]

  下载: 导出CSV

  Table 2.  Comparison between the SD model, SF model, and DSPM-DE model.

  Parameter	SD model	SF model	DSPM-DE model
  Salt permeability	Constant	Varies with salt concentration and applied pressure	Varies with salt concentration and applied pressure
  Independence of salt and water transport	Independent of each other	Coupling with each other	Coupling with each other
  Mechanisms of water transport	Diffusion	Hydraulic pressure difference	Hydraulic pressure difference
  Mechanisms of salt transport inside membrane	Diffusion	Concentration difference, potential difference, and convective effect	Concentration difference, potential difference, and convective effect
  Mechanisms of salt transport at the sides of membrane	Diffusion	Seric exclusion, Donnan effect, and dielectric exclusion	Steric exclusion, Donnan effect, and dielectric exclusion
  Membrane characteristics	Static	Static	Dynamic

  下载: 导出CSV

  Table 3.  Summary and comparison of modification strategies, mechanisms, and challenges.

  Membrane structures	Modification strategies	Mechanisms of ammonia nitrogen rejection	Challenge
  Surface of active layer	Grafting, chemical coupling, etc.	Decrease the partition of ammonia nitrogen	Degradation of the constructed function layer
  Insides of active layer	Co-solvent IP	Repel or absorb ammonia nitrogen	Adsorbate release and adsorbent regeneration
  Support layer	Surface functionalization, plasma treatment, inorganic particles doping	Repel or absorb ammonia nitrogen	Mechanical strength, adsorbate release and adsorbent regeneration

  下载: 导出CSV
•