Comparison of Membrane-Based Treatment Methods for the Removal of Micro-Pollutants from Reclaimed Water

: Dual membrane hybrid systems generally produce reclaimed water for non-potable uses by blending microﬁltered biologically treated sewage efﬂuent (BTSE) and reverse osmosis (RO) permeate. This reclaimed water is found to contain a signiﬁcant amount of micro-pollutants, which possibly cause toxicity effects to aquatic organisms and plants when exposed to it. Therefore, removing such pollutants from the reclaimed water before reaching the community is highly emphasized nowadays. The currently used treatment of the RO treatment of microﬁltered BTSE is energy intensive and not cost effective. This paper focuses on less costly and efﬁcient membrane-based hybrid treatment systems such as the microﬁltration-adsorption (MF-GAC) hybrid system, Nano ﬁlter (NF) and RO in the removal of micro-pollutants from the microﬁltered BTSE. Both the MF-GAC hybrid system and NF (with NTR 729HF membrane) removed 70 to 95% of micropollutants from microﬁltered BTSE. The removal depends on the hydrophobicity, charge, and size of the micropollutants. RO was excellent in removing more than 90% of pollutants, while MF was inefﬁcient, as the latter primarily depends on the size exclusion mechanism. Based on the ﬁnding, it is suggested to treat only a portion of microﬁltered BTSE through the MF-GAC or NF membrane before blending with RO permeate to enhance the removal of micro-pollutants from reclaimed water. The development of sustainable hybrid systems for the removal of all micropollutants of different chemical and physical properties is the key for the water reclamation.


Introduction
Biologically treated sewage effluent (BTSE) has progressively been used to produce recycled water as an alternative water resource in many countries (Singapore, Saudi Arabia, Qatar, Namibia, California and Australia) to meet the basic water demands [1]. The dual membrane process is a sustainable treatment solution [2]. In most of the dual membrane treatment plants, microfiltration (MF) or ultrafiltration (UF) is used as the first step to remove particulate and colloids from BTSE, and the reverse osmosis (RO) is used as the second stage [2,3]. RO is used to polish water and to improve the quality of the treated water [3,4].
Regarding the quality of recycled water, the presence of pathogens and chemical contaminants was the main concern in recycled water [5]. In the early 1990s, scientists observed endocrine disruption effects (feminization of male fish) in some fish species when exposed to wastewater effluents, even at trace levels of hormone concentration (a few ng/L) [5]. The micropollutants include pharmaceuticals, personal care products (PCPs), hormones, flame retardants, endocrine disrupting chemicals (EDCs), herbicides, pesticides, and nanomaterials. All of these are widely used by modern society.
Detailed experiments were conducted with the above two systems with microfiltered BTSE (MF/BTSE) in a water reclamation plant in Sydney, Australia, and the removal of representative micro-pollutants were measured.

Feed Water
The WTP treats wastewater through a dual membrane system consisting of continuous flow MF and RO with a total capacity of 7.5 ML/d from MF units and a maximum of 2 ML/d from RO units. The pH, conductivity, and sodium adsorption ratio (SAR) of BTSE were 6.8-7.6, 0.52-1.12 dS/cm and 39, respectively. The concentration of dissolved organic carbon was 3.6 to 7.7 mg/L. BTSE contain high concentration of Na + and Cl − ions in the ranges of 81-120 and 150-300 mg/L, respectively. All the BTSE is firstly passed through MF to remove suspended particles (>0.2 µm). A portion of the water filtered through the MF units was passed through the RO units. The MF unit consists of hollow fiber (Memcor) filters with 0.2 µm pores and arranged in three blocks. One block can produce approximately 2.5 ML/d. Two RO modules, each producing 1 ML/d, were used as a post-treatment to improve the quality of the water by reducing salinity. Following filtration, a portion of MF and RO permeate is blended at an appropriate ratio in the water reclamation/blending tank. The proportion of RO in the effluent usually ranges from 20-50%. After blending, the recycled water is sold back to the households for non-potable uses [4] (Figure 1).  Figure 1 shows the treatment path of BTSE (S1), microfiltered BTSE (S2), and RO permeates (S3) collected from the WTP. The various effluent was transported to the laboratory and maintained at 20 • C for the experiments. The DOC, conductivity, fluoride, chloride, sulfate, nitrate, calcium, and magnesium contents of the feed were 4.5-7.5 mg/L, 680-1120 µS/cm, 0.7-1.1 mg/L, 132-150 mg/L, 49-51 mg/L, 4.5-6.1 mg/L, 25-35 mg/L, and 10-12 mg/L, respectively.

Membranes
The following membranes were used in this study.

a.
Flat sheet membrane module: A flat sheet MF membrane module made up of polyvinylidene fluoride (PVDF) with a nominal pore size of 0.14 µm and surface area 0.2 m 2 was used for the MF-GAC hybrid system. It had 8 vertical membrane sheets with 11 mm gap between any two adjacent membrane sheets. The dimensions of this membrane are: 11.5 cm (width), 10.5 cm (length), and 22.5 cm (height). It was manufactured by the A3 Membrane Company, Gelsenkirchen, Germany. b. Nanofiltration (NF): Two hydrophilic NF membranes, NP030 and NTR729HF, were used to assess the performance of NF in removing micro-pollutants. The NP030 (manufactured by Macrodyn ® Nadir) was made of polyethersulfone with a molecular weight cutoff (MWCO) 400 Da and Zeta potential of −15 mV at pH 7 [28]. The NTR 729HF, manufactured by Nittto Denko, was made of Polyvinylalcoho/polyamides (Heterocyclic aromatic) with a MWCO 700 Da, and the Zeta potential was −100 mV [29]. c.
Reverse Osmosis (RO) membranes: RO membrane (manufactured by Woongjin Chemical) made of polyamides, with MWCO of 100 Da and zeta potential −21 mV [30] was used to investigate the RO removal performance of micro-pollutants from the feed water.

Granular-Activated Carbon (GAC)
Coal-based premium grade (MDW4050CB) GAC supplied by James Cumming and Sons Pty. Ltd., Sydney, Australia was used in the MF-GAC hybrid system. The GAC had an exchange/adsorption capacity of Iodine #1000 mg/g; moisture content 2% max; particle size 300-600 µm; surface area of 1000 ± 50 m 2 /g; and average pore diameter of 30 Å. This particle size was selected as it lowers the Trans Membrane Pressure (TMP) development by reducing the membrane-clogging effect [17].

Methodology
The following treatment methods were employed to compare the removal of micropollutants from the feed water.

MF-GAC Hybrid System
The performance of GAC on the removal of micro-pollutants was studied using a MF-GAC hybrid system ( Figure 2). The experimental procedure is stated in the previous studies [17,31]. An MF membrane (flat sheet) was submerged in the reactor tank containing 10 L of micro-filtered BTSE. Peristaltic pumps were used control the flow of feed water in and out of the tank. The TMP of the membrane filtration was measured using a pressure gauge. Initially, 20 g of GAC was added to the reactor tank to maintain the GAC dose of 2 g/L. The reactor tank was fed with continuous air flow at 1.5 m 3 /m 2 membrane area h (pre-determined), which keeps the GAC particles in suspension. The air flow also produces shear stress across the membrane surface, and this leads to a reduction of fouling of the membrane. The loss of volatile organic compounds (VOCs) due to aeration was neglected as the wastewater used in this study was previously biologically treated and stabilized. The primary purpose of the MF (flat sheet membrane) was to remove particulate GAC [32].

Nanofiltration
As shown in Figure 3, the filtration unit equipped with NF membranes (NP030 and NTR 729HF; membrane area of 68 cm 2 ) was used to filter the micro-filtered BTSE [12,19]. A cooling coil was submerged in the feed tank to maintain the feed water temperature constant at 20 • C. The filtration pressure was maintained at 4 bar. The clean water fluxes of NP030 and NTR 729HF were 12 and 62 L/m 2 ·h, whilst their permeability was 3 and 15.5 L/m 2 ·bar·h, respectively. The reject was re-circulated to the feed tank.

Reverse Osmosis
A similar experimental setup ( Figure 3) was used for the experiment with the RO membrane at the filtration pressure of 40 bar. The clean water flux was maintained to be 23.5 L/m 2 ·h, whilst its permeability was 0.59 L/m 2 ·bar·h.

Analysis of Micro-Pollutants
Concentrations of micro-pollutants were extracted using solid phase extraction (SPE) and examined with high performance liquid chromatograph with tandem mass spectroscopy (HPLC-MSMS) and quantified by isotope dilution. Samples (500 mL) were collected in clean glass bottles, stored in the dark (<4 • C) and extracted within 48 h.
Prior to SPE, samples were spiked with 50 ng (50 µL of a 1 mg/L solution) of isotope labelled analogues of all targets' micro-pollutants. Samples were loaded onto pre-conditioned OASIS 500 mg hydrophilic or lipophilic balance cartridges (Waters, Millford, MA, USA) at a rate of 10 mL/min. After completion, cartridges were rinsed with 5 mL of water and dried with a stream of nitrogen. Then, they were stored at 4 • C if not immediately eluted. The elution of target micro-pollutants was conducted from the cartridges with methanol (3 × 2.5 mL) and 1/9 (v/v) methanol/ethyl-tert-butylether (2.5 mL), and evaporated under a stream of nitrogen to approximately 100 µL. The extract was made up to approximately 1 mL with 60% methanol/water (v/v) and transferred to a 2 mL amber auto-sampler vial for instrumental analysis.
Target micro-pollutants were chromatographically separated and quantified using an Agilent (Santa Clara, CA 95051, USA) 1200 series HPLC equipped with a 150 × 4.6 mm, 5 µm particle size, Luna C18 column (Phenomenex, Torrence, CA, USA). Identification and quantification of micro-pollutants were carried out using an API 4000 triple quadrupole mass spectrometer (Sciex/Applied Biosystems, Forster City, CA, USA), which is equipped with a turbo-V ion source employed in both positive and negative electro-spray modes. A binary gradient consisting of 5 mM ammonium acetate in water (A) and 100% methanol at a flow rate of 800 mL/min was used. For electrospray ionization (ESI) positive analyses, the gradient was as follows: 10% methanol held for 0.50 min, stepped to 50% methanol at 0.51 min and increased linearly to 100% methanol by 8 min, then held at 100% methanol for 2 min. For ESI negative analyses, the gradient was as follows: 10% methanol held for 0.50 min, stepped to 60% methanol at 0.51 min and increased linearly to 100% methanol by 8 min, then held at 100% methanol for 3 min. A 5 min equilibration step at 10% methanol was used at the beginning of each run. An injection volume of 10 mL was used for all methods. Using scheduled multiple reaction monitoring (sMRM), two precursor-product ion transitions were monitored for all analytes, the most abundant used for quantitation.
Calibration was achieved by the construction of a minimum 5-point analyte/ISTD relative response ratio over a concentration range of 0.5-500 ng/mL. Linearity coefficients for all analytes were ≥0.99. Limits of Detection (LODs) were defined as the concentration of an extracted analyte giving a signal to noise (s/n) ratio greater than 3. Limits of Quantification (LOQ) were determined as the concentration that gave a peak of an extracted target analyte with a signal to noise ratio (s/n) of >10. For quality assurance and control, laboratory grade water blanks and fortified blanks at 10 and 100 ng/L were extracted with every batch of samples. No analytes were detected in blank samples above the quantitation limit. Recoveries of all target analytes were within 20% of expected concentrations. Solvents, buffer reagents and analytical standards were purchased from Sigma Aldrich (North Ryde, Australia). Isotope labelled internal standards were purchased from CDN Isotopes (PointeClaire, Canada). Water used in analysis was obtained from a Milli-Q purification system (Merck, Damstadt, Germany).

Estimation of Micro-Pollutants in the Reclaimed Water
Since the MF/BTSE is blended with the RO permeate to produce reclaimed water, the concentration of micro-pollutants in the reclaimed water was estimated using the following, Equation (1): where C 1 is the concentration of micro-pollutant in MF/BTSE; V 1 is the volume of the MF/BTSE sample; and C 2 is the concentration of micro-pollutant in RO permeate; and V 2 is the volume of the RO permeate. The upper limit of the concentration range of the respective MP in MF/BTSE was considered for C 1 , and the C 2 was considered as 0 assuming the RO achieves 100% rejection for all the micro-pollutants.

Removal Efficiency (%) of Micro-Pollutants
The removal efficiency of the micro-pollutants was calculated relative to the influent concentrations using the following, Equation (2).

Micro-Pollutants in the Microfiltered BTSE
The dissolved organic content of the feedwater ranged between 3-7 mg/L. The feedwater was tested for 36 commonly available micro-pollutants in municipal wastewater, out of which 20 micro-pollutants were detected above the LOQ (Table 1). Micro-pollutants that were detected at less than its LOQ (<5 ng/L or <10 ng/L) include Bisphenol A, Simvastatin-hydroxyacid, simvastatin, polyparaben, phenylphenol, paracetamol, dilantin, enalapril, risperidone, atrazine, linuron, omeprazole, clozapine, triamterene, hydroxyzine and diazepam. The micro-pollutant analysis was conducted in triplicate and the linearity coefficients for all analytes were ≥0.99. The micro-pollutants, naproxen, gemfibrozil, diclofenac, caffeine, trimethoprim, and carbamazepine, were detected at more than 100 ng/L in the MF/BTSE. Carbamazepine was detected at a high concentration, as it was less effectively removed by conventional wastewater treatment systems [33]. There is a possibility that this water could cause toxicity in the zebra embryo due to the levels of micro-pollutants such as diclofenac, diuron, carbamazepine, diazinon, and triclosan, as their Predicted No Effect Concentration (PNEC) are 11.64, 9.41, 53.05, 7.97, and 1.14 ng/L, respectively [34]. The algae toxicity might be due to the concentration of diuron [35], diazinon [36], and triclosan, as their PNEC were 8.00, 10.00, and 1.40, respectively [37]. Carbamazepine could possibly cause invertebrate toxicity (PNEC 25.00) [38].
Based on the review by Orias and Perrodin [39], Gemfibrozil reduces the plasmatic testosterone in the species C. auratus (PNEC 300 ng/L), triclosan affects the biomass of biofilms (PNEC 24 ng/L), trimethoprim might reduce the metabolic activity of D. polymorpha (PNEC 5.8 ng/L), Caffein affects the development of the number of viable offspring of X. laevis (PNEC 0.05 ng/L) and fluoxetine delays the development of R. pipiens species (PNEC 2.9 ng/L). The feedwater analyzed in this study detected the aforementioned micropollutants at levels greater than their respective PNEC. In some cases, the LOQs were higher than PNEC of the respective micro-pollutants. Micro-pollutants detected at less than LOQ in the feed water do not necessarily mean that the compound is absent or within safe levels [39].
The reclaimed water supplied to the consumers was detected with a range of micropollutants which are greater than PNEC. This might possibly cause several aquatic and environmental toxicity when using reclaimed water, especially for filling ornamental ponds, irrigation, water fountains, and some other water features.

Removal of Micro-Pollutants by MF
MF bearing a nominal pore size of 0.2 µm was found to only reduce the micropollutants by less than 50% except for fluoxetine and triclosan ( Table 2). Similar results were reported by Gidstedt et al. [23], who observed less than 55% removal of micro-pollutants by MF (0.2 µm) of wastewater. It was further noticed that there was no removal (0%) of Sulfamethoxazole, which agrees with the observation in this study [14]. Rodriguez-Mozaz et al. [14] reported that MF is not capable of rejecting any micro-pollutants as the MWCO of MF (~300,000 g/mol) is far bigger than that of the size of micro-pollutants (<1000 g/mol). The reduction of micro-pollutants observed in this study could be due to the interaction of the membrane surface with the nature of the micro-pollutants. Table 2. Removal (%) of micro-pollutants by MF filtration of 0.2 µm pore size. Amtriptyline  19  13  30  Atenolol  104  84  19  Caffeine  73  68  6  Carbamazepine  393  369  6  Diclofenac  164  152  7  Fluoxetine  19  9  53  Gemfibrozil  334  298  11  Ketoprofen  114  95  17  Naproxen  820  522  36  Primidone  29  26  12  Sulfamethoxazole  61  61  0  Triclocarban  81  41  50  Triclosan  91  5  95  Trimethoprim  236  208  12  Verapamil  35  27  23

Removal of Micro-Pollutants by MF-GAC Hybrid System
The MF-GAC hybrid system was found to reduce the concentration of micro-pollutants by more than 60% for 15 micro-pollutants out of the 18 monitored (Table 3). The micropollutants such as gemfibrozil, atenolol and trimethoprim were efficiently adsorbed into GAC and the removals were more than 95%. Micro-pollutants naproxen, ibuprofen, tricloson and carbamazepine were also efficiently removed (90-95%). The remaining micropollutants also had large removals from the feedwater after treatment with the MF-GAC hybrid system (70-95%), except primidone, which demonstrated the lowest removal (54%). Table 3. Removal (%) of micro-pollutants by MF-GAC hybrid system (GAC's initial dosage of 2 g/L). The literature explains the removals based on the hydrophobicity of the micro-pollutants, which is typically described by the K ow value (octonol-water partition co-efficient) and acid dissociation constant (pKa) [27,40]. The micro-pollutants having high hydrophobicity had more affinity towards GAC than the micro-pollutants with low hydrophobicity [21]. As shown in Table 3, the micro-pollutants, namely naproxen, ibuprofen, gemfibrozil, and tricloson with Log K ow values > 3.0, were effectively removed (>90%) by the GAC. The primidone and sulfamethoxazole with Log K ow values 0.91 and 0.89, respectively, showed smaller removal of 54% and 73% respectively. The reduced removal can be explained by the lower Log K ow values that have a lower affinity to GAC.

Micro-Pollutants
Contrary to the above, micro-pollutants such as atenolol and trimethoprim demonstrated higher removals (97%) despite their Log K ow values (0.16 and 0.89, respectively). A similar observation was made by Liu et al. [41]. They also observed high removal of trimethoprim and the sulfamethoxazone (80-90% and 60-70%, respectively) with the treatment with an artificial composite soil treatment system. Here, the removal is strongly related to sorption. Atenolol having low Log K ow (0.16) showed greater removals by GAC (97%). This could be due to its positive charge. Caffeine with low Log K ow (−0.07) also had a high removal of 78%, as it bears a neutral charge. A similar observation was reported by Jamil et al. [27] The MF/BTSE contains approximately 85% of low molecular weight (LMW) hydrophilic compounds (humics and LMW neutrals) [32], which may possibly compete with micro-pollutants for GAC adsorption sites. However, the micro-pollutants with higher log K ow can become adsorbed into GAC even in the presence of high load of NOM [42]. This has been clearly demonstrated by the results obtained in this study.
The removal of individual trace organics cannot be described by the GAC adsorption based solely on the log K ow , molecular size, and charge. Removal also depends on a variety of physical and chemical interactions among trace organics, dissolved organic matter, and GAC, its dosage, temperature, pH, etc. [43,44].

Removal of Micro-Pollutants by NF
The performance of the two NF membranes, NP030 and NTR729HF, in terms of the removal of selected micro-pollutants, are presented in Table 4. The NP030 and NTR729HF membranes were found to reduce 9 and 11 micro-pollutants, respectively, by more than 50% out of the total number of micro-pollutants detected in the feed water. Comparing NP030 and NTR729 HF, the latter was more efficient at removal. The removal of micro-pollutants by the NF membrane is governed by two mechanisms, namely size exclusion and electrostatic screening [12,27]. The fundamental mechanism behind the rejection of micro-pollutants by membrane filtration is size exclusion, although the removal by electrostatic repulsion, adsorption due to hydrophobic interactions, and adsorption onto the membrane fouling layer also can play a part [45].
The rejections (%) of micro-pollutants such as sulfamethoxazole, naproxene and diclofenac were much higher (>90%) with NTR729HF than the NP 030 despite the higher MWCO (700 Da) of NTR 729HF compared NP 030 (400 Da). Since these micro-pollutants are negatively charged, the electrostatic interaction between the micro-pollutants and the membrane surface exerts electrostatic repulsions and hinders its passage through the membrane [45]. The zeta potential of the membrane surface of NTR729HF (−100) is higher compared to that of NP030 (−15).
The insight mechanism on the removal of micro-pollutants by the NF cannot preciously be explained, as the removal mechanism pertaining to electrostatic repulsion depends on several factors such as the properties of membrane, the micro-pollutant and the chemical properties of the solution [46]. They further reported that the charge of the micro-pollutant and the membrane determines the distribution of ions on the surface of the membrane. If the micro-pollutant charge is opposite to the membrane charge, there is a possibility to transport the micro-pollutant across the membrane. In case of micro-pollutant and membrane being with the same charge, the membrane will repulse the ions. However, the charge of the micro-pollutant and the membrane surface are highly influenced by the presence of the charge of other compounds and the solution pH [46]. Therefore, considering the other physico-chemical characteristics of the membrane and the properties of the micro pollutants are essential [47,48]. For example, the lower removal of pharmaceuticals was observed by NF/RO membranes with the presence of calcium ions [49].
The performance of NP030 is better in the removal of positively charged micropollutants (namely atenolol, amitriptyline, verapamil, fluoxetine) than that of NTR729HF. This can be explained by the size exclusion mechanism of the NF membrane [27,45]. The removal of amitriptyline and verapamil were between 71-76% by the NP030 and 55-57% by the NTR 729HF, as the MWCO of the former is more restrictive. Atenolol was removed 65% by the NP030, and 75% by the NTR729 HF could be due to the size exclusion mechanism. This could be due to the adsorption of the compound on the membrane surface due to several other factors, such as the shape of the molecule, hydrophobic interaction and hydrogen bonding. Fluoxetine showed lesser removal (44%) by NP030, and this may be due to the larger size of the molecule (MW = 309 g/mol). The relation between the rejection and the molecular weight of the compounds agrees with Shirley et al. [50]. Further, the direction of the dipole moment also determines the position of the molecule with respect to the membrane surface in rejecting compounds by NF [51].
The removal mechanism of uncharged compounds (caffeine, trimethoprim, carbamazepine, primidone, diuron, triclosan) cannot be explained by size exclusion nor electrostatic repulsion by the membrane surface. The positively charged or uncharged compounds can reach the membrane surface and become adsorbed onto the polymer matrix more easily compared to negatively charged molecules [52]. Removal could be due to hydrophobic interactions among the compounds, membrane surface or foulants [53]. Caffeine was detected at high concentrations in the feedwater (675 ng/L), and its removal was insignificant by NP030 (6%) and by NTR729HF (35%). The low MW of the compound allowed it to pass through the pores of the membrane (MW = 194 g/mol).

Removal of Micro-Pollutants by RO
RO membranes are nonporous, and the removal of micro-pollutants occurs through the sorption diffusion mechanism, where the micro-pollutants are solubilized and moved by gradient of the chemical potential [54].
RO demonstrated excellent removal of micro-pollutants from feed water and most of the compounds remained below LOQ ( Table 5). The compounds, namely caffeine, carbamazepine, atenolol, and triclosan, detected in permeate, demonstrated more than 92% removal. The micro-pollutant caffeine was left at a higher concentration in permeate (21 ng/L) compared to the rest. The smaller MW of the caffeine (194 g/mol) would have reduced the RO removal rate. This is supported by Dolar et al. [51], as they found the RO rejects the compounds primarily by the size exclusion mechanism.

Conclusions
A total of 20 micro-pollutants out of 36 measured were detected at the above LOQ in the reclaimed water. The micro-pollutants naproxen, gemfibrozil, diclofenac, caffeine, trimethoprim, and carbamazepine were detected at more than 100 ng/L. Out of the 20 detected micro-pollutants, about 50% of them were above PNEC levels in the feed water. Therefore, this reclaimed water may not be ecologically safe if living species including plants are exposed to it.
The MF is not effective in removing micro-pollutants (removed less than 20%). NF is effective in rejecting micro-pollutants through electrostatic interactions and size exclusion and most of the micro-pollutants' removal rates were in the range of 70-95%. NTR729HF was more efficient than NP030, as the former had smaller MWCO and higher zeta potential. GAC also efficiently (75-95%) removed micro-pollutants, and the hydrophobic substances demonstrated higher affinity towards the solid surface of GAC. The removal of the micropollutant by RO membrane was excellent.
Development of sustainable hybrid systems for the removal of all micro-pollutants of different chemical and physical properties is the key for the water reclamation. In this study, the performance of MF-GAC hybrid system and NF were equally good in the removal of micro-pollutants. Therefore, a portion of the MF permeate (presently in blends of between 70-50%) could be sent to either the MF-GAC hybrid system or NF membrane (NTR 729HF) to further reduce the micro-pollutants in the MF permeate before blending with the RO permeate ( Figure 4). Option 1: Blending of MF, MF-GAC, and RO permeate: Since MF-GAC is effective in reducing micro-pollutants, a portion of the MF permeate could be treated through MF-GAC hybrid system to further reduce the micro-pollutants before it is blended with RO permeate at an appropriate ratio.
Option 2: Blending of MF, NF permeate, and RO permeate: Since the NTR729HF is effective in rejecting micro-pollutants, a portion of the MF permeate can be filtered through NTR729HF, and the permeate can be blended with the RO permeate.
The appropriate blending ratio of MF permeate, either MF-GAC or NF permeate, and RO permeate should be decided based on the water quality parameters and the level of micro-pollutants in the reclaimed water. Therefore, a proper monitoring mechanism should be developed to monitor the micro-pollutants levels in the reclaimed water at regular intervals to ensure environmental safety.
Different low-pressure NF membranes and alternative adsorbents need to be studied to remove the broad spectrum of micro-pollutants (of different charges and molecular weights) to achieve sustainable water reuse.