MaterialsPebax 2533 was supplied by Arkema as pellets. Polyethersulfone was supplied by Goodfellow Cambridge Limited as pellets. The 2,3,5,6-tetrafluoroterephthalonitrile (TFTPN), poly(tetramethylene glycol) (PTMG, MW: 2000 g·mol−1) and octanediol chain extender and dibutyltindilaurate (DBTDL) was purchased from Sigma-Aldrich. 3,3,3′,3′-tetramethyl−1,1′- spirobisindane-5,5′,6,6′-tetrol (TTSBI) was provided from TCI Co., Ltd. Isophorone diisocyanate (IPDI), formic acid (98%), and NaOH were purchased from Wako Pure Chemical Industries. HCl (35%) was provided by Nacalai Tesque, Inc. Solvents were purchased from Wako Pure Chemical Industries except for those at high-performance liquid-chromatography (HPLC) grade that were purchased from Fischer Chemicals. Distilled and Milli-Q water (pH = 6.8 ± 0.1, n = 3) were obtained from E-POD and Q-POD Millipore systems.A total of 13 PPCPs were selected as model pollutants. Analytical standards of methylparaben (MePb), propylparaben (PPb), and benzophenone (BP) were purchased from Wako Pure Chemical Industries. While naproxen (NPX, analytical standard) and methyl anthranilate (MATN, ≥97%) were supplied by Sigma-Aldrich, analytical standards of 2-hydroxy-4-methoxybenzophenone (BP3), octocrylene (OCR), ehtylhexyl methoxycinnamate (EHMC) and homosalate (HMS) were supplied by Supelco®. Triclosan (TRC, ≥99.9%) and N,N-diethyl-m-toluamide (DEET, analytical standard) were acquired from Dr. Ehrenstorfer. Diclofenac (DCF, ≥95%) and losartan (LOS, ≥98%) were purchased from Angene Chemicals and Ambeed, respectively. The DCF analogue 2-(2-(phenylamino)phenyl)acetic acid (PaPAA, ≥99%) was obtained from BLD Pharmatech Ltd. Standard solutions of the analytes were prepared in acetonitrile (HPLC grade) at 1000 mg·L−1, except for LOS, which was prepared in methanol (HPLC grade). Mix standard solutions containing all the PPCPs and selected PPCPs were prepared in acetonitrile at 100 mg·L−1 and 0.5 mg·L−1. These stock solutions were stored in the fridge at 4 °C and used to prepare working solutions in the different water matrices. The main physicochemical characteristics of these compounds are shown in Supplementary Table 1.Tap water sample (pH = 7.18 ± 0.02, n = 3) was collected in the laboratory (Kyoto, Japan) just before the experiments. River water samples were collected in Kamo River (Kyoto, Japan) in three different spots (north, center, and south) using glass brown bottles. The samples were filtered using hydrophilic PTFE filters (OmniporeTM, 0.2 µm), mixed in 1:1:1 proportion (pH = 7.7 ± 0.1, n = 3) and stored in the fridge at 4 °C until analysis.SynthesisSynthesis of HRhMOPHRhMOP was synthesized according to our previous reports23,37. Rhodium acetate Rh2(AcO)4·(MeOH)2 (200 mg, 0.4 mmol) was reacted with benzene-1,3-dicarboxylic acid (H2BDC) (328 mg, 1.97 mM) in 14 ml of Dynamic mechanical analysis (DMA) in the presence of Na2CO3 (210 mg, 2 mmol) at 100 °C for 48 h. The resulting precipitating solid was separated from the reaction mixture, followed by successive washing with DMA, EtOH and MeOH. Finally, the solids were collected and dried overnight at 120 °C under vacuum to obtain the MOP powders.Synthesis of C12RhMOPC12RhMOP was synthesized following our previously published protocol38. Rhodium acetate Rh2(AcO)4·(MeOH)2 (100 mg, 0.2 mmol) was reacted with 5-dodecoxybenzene−1,3- dicarboxylic acid (H2BDC-C12) (174 mg, 0.5 mmol) in 10 ml of DMA in the presence of Na2CO3 (52 mg, 0.5 mmol) at 100 °C for 48 h. The resulting green solution was centrifuged, followed by precipitating the supernatant with MeOH to obtain C12RhMOP solids. The solid was redissolved in dichloromethane, filtered, and evaporated. Finally, the solids were collected, redispersed and washed in ethanol and dried overnight at 120 °C under vacuum to obtain the products.Synthsis of 1,4-bis(imidazol−1-ylmethyl)benzene (bix)bix was synthesized according to a previously reported procedure38. Imidazole (2.13 g, 16.6 mmol) and NaH (0.88 g of 60% in mineral oil) were mixed at room temperature in 50 mL THF for 30 min. Then, a solution of α,α′-dibromo-p-xylene (2 g, 7.57 mmol) in 20 mL of THF was added. After the reaction at 50 °C for 4 h, ice water was added into the mixture to quench the reaction. The organic phase was extracted with chloroform and treated with anhydrous MgSO4. After drying under vacuum, the collected powders were washed with diethyl ether twice and finally dried under vacuum to obtain the white powder products.Synthesis of 4,4’-Bis(imidazol-l-ylmethyl)biphenyl (bibPh)bibPh was synthesized in the same method as described for bix expected for the use of 4,4′-Bis(bromomethyl)biphenyl38.1-dodecyl-1H-imidazole (diz) was synthesized by adapting a previously reported protocal38. Imidazole (1.63 g, 24 mmol) and KOH (1.4 g, 25 mmol) were mixed in 60 mL acetonitrile and stirred at room temperature for 2 h. Then 1-bromododecane (4.98 g, 20 mmol) was added and the solution was stirred for 24 h at room temperature. Finally, the solvent was evaporated and water was added to the resulting solid. The product was extracted in DCM, treated with anhydrous MgSO4, filtered, and dried by removing DCM solvents. Finally, the product was dried under vacuum.Synthesis of AO-PIMAO-PIM was synthesized according to a previously reported procedure39. Firstly, the PIM-1 polymer was synthesized by reacting TTSBI (10.21 g, 30 mmol) and TFTPN (6.00 g, 30 mmol) in 200 mL DMF in the presence of anhydrous K2CO3 (8.40 g, 60.8 mmol) at 70 °C for 72 h under N2. After successive washing with water, acetone and methanol, the obtained polymer was purified by repeated precipitation from chloroform solution by adding methanol dropwise until the solution became turbid. After drying at 110 °C under vacuum, the obtained PIM-1 (4.8 g) was dissolved in 300 mL THF at 65 °C, followed by dropwise addition of hydroxylamine solution (50 mL, 50 wt% in H2O). After reflux for 20 h under N2, the reaction mixture was poured into 1 L ethanol, followed by collecting the solid through filtration. After washing with ethanol, the product was dried at 100 °C under vacuum to yield an off-white powder AO-PIM.Synthesis of polyurethane (PU)The PU was synthesized by a two-step bulk polymerization method. First, PTMG was reacted with an excess IPDI (PTMG:IPDI = 1:3 molar ratio) under nitrogen atmosphere at 75 °C, followed by the addition of 0.1 mL DBTDL to obtain a macro diisocyanate pre-polymer. After 2 h, an exact amount of the chain extender octanediol was added for equimolar adjustment of the NCO:OH ratio. The synthesized polyurethane was washed and precipitated in methanol: water (50:50 wt%) to remove any unreacted monomers or low molecular weight polymers. Samples were dried at 70 °C under vacuum before use. The molar ratio of PTMG:IPDI: chain extender was 1:3:2.Synthesis of HRhMOP(diz)12
200 mg of HRhMOP were dispersed in 40 mL of DCM, then 132.5 mg of diz (18 mol. eq.) were added. After sonication for 5 min, the solution was centrifuged to remove the precipitate. The upper purple solution was collected and evaporated in vacuum. The obtained solid residue was washed twice with EtOH to remove any remaining diz. Finally, the purple solid product was dried under vacuum.Synthesis of the pure linked MOP gel and its corresponding aerogel1 mL of DMF solution of HRhMOP(diz)12 at 2.8 mM was added to 1 mL of DMF solution of bix at 33.6 mM (12 mol. eq. related to MOPs) under vigorous stirring. The purple solution was transferred to a sealed syringe and heated at 80 °C for 8 h to complete the reaction and obtain the gel. To get the corresponding aerogel sample, the as-synthesized gel was solvent-exchanged by soaking in acetone for three days, while replacing the fresh acetone every day. The sample was then dried by supercritical CO2 at 14 MPa and 50 °C.Synthesis of PNM/PEBA n wt%The PEBA membrane and its composites with linked MOP networks at different loadings were prepared in DMF by using the solution casting method. In a typical synthesis process, HRhMOP(diz)12 and bix (12 mol. eq. related to MOPs) were separately dissolved in the DMF solution of PEBA (3 wt%), respectively. Then the solution of HRhMOP(diz)12 was added to the solution of bix under vigorous stirring at 40 °C. The resultant transparent purple solution was transferred into the Teflon Petri dish and heated at 80 °C overnight. After the evaporation of DMF solvent upon heating, the films were obtained and removed from the Petri dishes, followed by drying under vacuum at 80 °C overnight. The resulting composite films were referred to as PNM/PEBA n wt% where n indicates the weight percentage of HRhMOP(diz)12 related to the PEBA matrix. For the PNM sample made from C12RhMOP or bibPh, the same synthetic procedure was used except for the use of desired MOPs or linkers.Synthesis of MMM/PEBA n wt%The PEBA membrane and its composites with isolated MOPs at different loadings were prepared in DMF by using the solution casting method. In a typical synthesis process, HRhMOP(diz)12 was dissolved in the DMF solution of PEBA (3 wt%) at 40 °C. Then the resultant transparent pink solution was transferred into the Teflon Petri-dish and heated at 80 °C overnight. After the evaporation of DMF solvent upon heating, the films were obtained and removed from the Petri-dish, followed by drying under vacuum at 80 °C overnight. The resulting composite films were named MMM/PEBA n wt% where n indicates the weight fraction of HRhMOP(diz)12 related to the PEBA matrix.Synthesis of PNMs and MMMs from other polymer matricesFor the membrane samples made from PU, PES and AO-PIM, the same synthetic procedure to PNM/PEBA n wt% and MMM/PEBA n wt% was used except for the use of DMF solutions of the different polymers.Instrumentation and characterizationSEM of cross-sectional morphologies of the membranes were observed using a field-emission scanning electron microscope with a JEOL Model JSM-75FTC system operating at 10 kV and 5 mA current. The surface morphologies of the membrane samples were observed using a scanning electron microscope (FESEM, Hitachi SU-5000) equipped with EDS. The samples were coated with Osmium (14 nm layer thickness) before measurement.Optical microscopy of the membranes was measured using a VHX-5000 series digital microscope (Keyence).UV-visible spectroscopy of the membranes and solution samples was performed in a V-670 spectrophotometer (JASCO).The rheological measurements of the membrane samples were made using a solid analyzer RSA-G2 (TA Instruments, New Castle, DE, USA) rheometer. DMA measurements of the membranes were conducted by frequency sweeping in a tensile mode with a 0.3% strain amplitude that was well inside the linear regime (initial strain is fixed to 0.5%). The tensile tests of the membranes were performed by stretching the samples in an initial strain rate of ε̇0 = 1.5 min−1.Differential scanning calorimetry (DSC) was carried out in the temperature range from 30 °C to 150 °C at a heating rate of 10 °C·min−1 using Hitachi DSC 7020 and under a nitrogen atmosphere. Supplementary Fig. 37 shows the DSC plot for pristine PEBA, PNM/PEBA n wt% and MMM/PEBA n wt% samples.Fourier transform infrared spectra of the MOPs and the membrane samples were conducted on a Jasco FT/IR-6100 in the range of 4000–500 cm−1 with a resolution of 4 cm−1 and 128 scans. Supplementary Fig. 38 shows the spectra for pristine PEBA, PNM/PEBA n wt% and MMM/PEBA n wt% samples.Thermogravimetric analyses (TGA) were performed in the temperature range from room temperature to 500 °C at a heating rate of 10 °C·min−1 with a Rigaku Thermo plus EVO2 and under a nitrogen atmosphere. Supplementary Fig. 39 shows the TGA plot for HRhMOP, pristine PEBA PNM/PEBA n wt% and MMM/PEBA n wt% samples.The super-critical CO2 drying process was carried out on SCLEAD-2BD autoclave (KISCO) using supercritical CO2 at 14 MPa and 50 °C.The stability of the membranes was evaluated by immersing them in 2 mL of the tested solutions for 24 h. Then, the membranes were air-dried and characterized by different techniques. The tested solutions were distilled water (pH = 5.8 ± 0.1, n = 3), HCl solutions at 0.1 M (pH = 0.96 ± 0.04, n = 3) and 10−3 M (pH = 2.90 ± 0.02, n = 3), NaOH solution at 10−4 M (pH = 9.64 ± 0.04, n = 3) and acetonitrile.Uptake and release of PPCPsA Prominence-20A modular HPLC system from Shimadzu (Japan) consisting of two LC-20AT pumps, a DGU-20A3R degassing unit, a CTO-20AC column oven, and an SPD-20A UV-visible detector was used for the quantification of the PPCPs in the solutions. The separation was carried out on a Shim-pack Velox C18 column (150 mm × 4.6 mm, 5 µm) (Shimadzu), which was protected with a Velox EXP C18 guard cartridge (5 mm × 4.6 mm, 5 µm) and kept at 30 °C. Methanol and water acidified with formic acid (0.02% v/v, pH = 3) were used as mobile phases at a flow rate of 1 mL·min−1. The separation of the 13 PPCPs was achieved using a starting mobile composition of 50% (v/v) of methanol, held for 10 min, followed by an increase to 60% in 2 min, held for 2 min, increased to 90% in 2 min, held for 4 min, and finally increased up to 100% in 2 min, and held for 3 additional min to ensure elution of all the compounds. The injection volume was 20 µL and the detection wavelengths were set at 220, 254, 280, and 312 nm at different times depending on the PPCP (see Supplementary Table 1). The data were acquired and analyzed using Shimadzu LCsolution software.For the determination of LOS during adsorption isotherm experiments, the mobile phase composition started at 70% (v/v) of methanol, increased to 100% in 3 min, and kept for 0.5 additional min. In the case of DCF, the initial mobile phase was set at 80% (v/v) of methanol, which was increased to 100% in 3 min and held for 0.5 min. The same elution gradient was used for the separation and determination of DCF and PaPAA and using 280 nm as detection wavelength. During the competitive adsorption isotherm experiments with LOS and DCF, the separation was achieved using an initial mobile phase composition of 70% (v/v) of methanol, which was increased to 80% in 2 min, followed by another increase up to 100% in 2 min, which was kept for 1 additional min.Calibration curves for the PPCPs were obtained using standard solutions of the analytes prepared in Milli-Q water by dilution of the mix standard solution (100 mg·L−1). Several quality analytical parameters of the calibration curves are included in Table S5. The limits of detection (LODs) were determined by decreasing the concentration of the PPCP until a signal-to-noise ratio of 3 was obtained. The limits of quantification (LOQs) were estimated as 10/3 times the LODs and experimentally verified. The calibration ranges (and calibration levels) were different depending on the analyte and ranged from the LOQ to 12.5 mg·L−1.The uptake and release of the membranes towards PPCPs was evaluated by determining their concentration before and after immersing the membranes in aqueous solutions of the analytes. In a typical experiment, a membrane was pierced using a syringe needle and hanged inside 10 mL aqueous solutions (Milli-Q, tap or river water) containing the PPCP(s) at a specific concentration depending on the study. The solution was stirred at 500 rpm using a magnetic stir bar for a fixed time depending on the study. Then, the membrane was separated from the aqueous solution, which was analyzed by HPLC-UV-visible to determine the remaining concentration of PPCPs. The membrane was dried with a tissue and stirred in acetonitrile for 5 min using vortex to release the compounds. Acetonitrile was selected to release these compounds after uptake because of its capability to coordinate with HRhMOP for replacing possible coordinating PPCPs and its compatibility with HPLC analysis. This extract was diluted (1:1) with Milli-Q water containing formic acid 0.02% v/v, and then analyzed by HPLC-UV to determine the released concentration of PPCPs. Before experiments, the membranes were cut in square shapes of 1.5 × 1.5 cm or 0.5 × 0.5 cm and dried in the vacuum oven at 80 °C for 2 h to remove any remaining solvent.For the uptake and release screening experiments using membranes, the concentration of the PPCPs was 0.1 or 1 mg·L−1 and the uptake time was fixed at 24 h. When using 0.5 cm membranes for uptake of 0.1 mg·L−1 PPCPs, the UV filter compounds were excluded because of the negligible contribution of MOP fillers in their uptake. The aqueous solutions were prepared using Milli-Q water, tap water or river water depending on the experiment. The volume of acetonitrile for the release was 0.5 mL using 1.5 cm membranes and 0.15 mL using 0.5 cm membranes. Each experiment was carried out by triplicate using 3 different membranes. For the reusability study, after each uptake/release cycle, the membranes were washed using 1 mL of acetonitrile and 5 min of vortex stirring. Then, they were air-dried and used again for the uptake experiments.The uptake (in percentage) was calculated by comparing the concentration of PPCPs found in the aqueous solution before and after experiment. The release (in percentage) was calculated by comparing the mg of PPCPs found in the acetonitrile extract and the initial mg in the aqueous solution.$${{\rm{Uptake}}}\,=\,\frac{{{{\rm{C}}}}_{{{\rm{initial}}}}{{-}{{\rm{C}}}}_{{{\rm{final}}}}}{{{{\rm{C}}}}_{{{\rm{initial}}}}}\cdot 100,{{\rm{Release}}}\,=\,\frac{{{{\rm{C}}}}_{{{\rm{extract}}}}\cdot {{{\rm{V}}}}_{{{\rm{extract}}}}}{{{{\rm{C}}}}_{{{\rm{initial}}}}\cdot {{{\rm{V}}}}_{{{\rm{initial}}}}}\cdot 100$$For the uptake and release screening experiments using the aerogel sample (composition: HRhMOP(bix)9.4(diz)1.1)24, 0.25 mg of aerogel (mimicking PNM/PEBA 11 wt% membranes of 0.5 × 0.5 cm) were dispersed in 10 mL of aqueous solution containing the PPCPs at 0.1 mg·L−1 and stirring at 500 rpm with a magnetic stir bar. After 24 h, the solution was transferred to a plastic centrifuge tube and centrifuged at 3500 rpm for 5 min. The supernatant was analyzed by HPLC-UV-visible and discarded. 0.15 mL of acetonitrile was added to the recovered aerogel and the tube was stirred for 5 min using vortex to release the compounds. After centrifugation at 3500 rpm for 5 min, the supernatant was diluted (1:1) with Milli-Q water containing formic acid 0.02% v/v and then analyzed by HPLC-UV-visible. Filtration was not possible because of the adsorption of the compounds by the syringe-filters evaluated. The aerogel was activated at 120 °C under vacuum overnight before experiments.For the kinetic experiments using PNM/PEBA 11 wt% and MMM/PEBA 11 wt% membranes (0.5 cm), the concentration of the PPCPs was 0.5 mg·L−1, and the uptake experiments were carried out for 36 h. Aliquots of 25 µL of the aqueous solution were taken at different times and analyzed by HPLC-UV-visible. Each experiment was carried out by triplicate using three different membranes to calculate the standard deviation.The kinetic data was fitted to different models using Python (v. 3.9.7) and the following non-linear regressions31:Pseudo-first order model. Empirical model based on diffusion. Rate is directly proportional to the difference in concentration and the amount of uptake with the time.$${{{\rm{q}}}}_{{{\rm{t}}}}={{{\rm{q}}}}_{{{\rm{e}}}}\,\cdot \,(1\,{-}{{{\rm{e}}}}^{{{-}{{\rm{K}}}}_{1}\cdot {{\rm{t}}}})$$Pseudo-second order model. Empirical model based on chemisorption. Rate depends on the adsorption capacity of the material and not on the solute concentration. Fitting is better when adsorption equilibrium is reached.$${{{\rm{q}}}}_{{{\rm{t}}}}\,=\,\frac{{{{\rm{K}}}}_{2}{\cdot \,{{\rm{q}}}}_{{{\rm{e}}}}^{2}\,\cdot \,{{\rm{t}}}}{{1+{{\rm{K}}}}_{2}{\cdot {{\rm{q}}}}_{{{\rm{e}}}}\,\cdot \,{{\rm{t}}}}$$Elovich model. Empirical model based on chemisorption. Rate depends on the adsorption capacity of the material and decreases exponentially as the amount of adsorbed solute increases. It fits better when adsorption equilibrium is not reached.$${{{\rm{q}}}}_{{{\rm{t}}}}\,=\,\frac{1}{{{\rm{b}}}}\,{\mathrm{ln}}\,(1\,+\,{{\rm{a}}}\,\cdot \,{{\rm{b}}}\,\cdot \,{{\rm{t}}})$$The adsorbed amount at each time (qt, mg·g−1) and the adsorbed amount at equilibrium (qe, mg·g−1) were expressed as mg of analyte per g of HRhMOP in the membrane. The amount of HRhMOP in each membrane was calculated considering the weight of the membrane, the reagent ratios used during the synthesis, and the MW of HRhMOP(diz)12. In this case, the average and standard deviation for three membranes were 0.14 ± 0.01 mg of HRhMOP for PNM/PEBA 11 wt% and 0.22 ± 0.05 mg for MMM/PEBA 11 wt%. In the previous equations, K1 is the pseudo-first-order kinetic parameter (min−1), K2 is the pseudo-second-order kinetic parameter (min−1), b is the desorption rate constant of the Elovich model (g·mg−1), a is the initial adsorption rate constant of the Elovich model (mg·g−1·min−1), and t is the time at which qt was measured (min). The chi-square value was used to evaluate the goodness of the fitting.For the adsorption isotherm experiments, pure PEBA, PNM/PEBA 11 wt%, and MMM/PEBA 11 wt% membranes (0.5 cm) were immersed in aqueous solutions containing the selected PPCPs at different concentrations for 24 h. Individual adsorption isotherms were obtained using aqueous solutions containing one compound, and competitive adsorption isotherms were obtained using aqueous solutions containing two compounds. The isotherms were obtained in PPCP concentration ranges within the water solubility domain and the HPLC limitation of quantification. In the case of LOS, the concentration ranged from 0.08 to 8 mg·L−1. For DCF and PaPAA, the range was from 0.02 to 2 mg·L−1. In all the experiments, the acetonitrile content in the aqueous solution was adjusted to 1% (v/v). The concentration of the PPCP(s) solutions before and after the uptake was used to determine the adsorbed amount of analyte (mg or mol). The error for each point of the isotherms was estimated by error propagation starting from the standard error of the concentrations determined by the calibration curves, which was calculated using the following equation:$${{{\rm{s}}}}_{ {{\rm{x0}}}}=\frac{ { {{\rm{s}}}}_{{{\rm{y}}}/{{\rm{x}}}}} {{{\rm{b}}}}\sqrt{1+\frac{1} {{{\rm{n}}}}+\frac{{({{{\rm{y}}}}_{0}{-}\overline{{{\rm{y}}}})}^{2}} {{{{\rm{b}}}}^{2}{\sum }_{{{\rm{i}}}} {({{{\rm{x}}}}_{{{\rm{i}}}}{-}\overline{{{\rm{x}}}})}^{2}}}$$where sy/x is the residual standard deviation from the calibration, b is the calibration slope, n is the number of calibration levels, y0 is the measured peak area, ȳ is the mean of the peak area for the calibration standards, xi is a concentration value from the calibration curve and x̄ is the mean of the concentrations for calibration standards.The experimental data was fitted to different models using Python (v. 3.9.7) and the following non-linear regressions13,30:Langmuir model. It assumes the adsorption occurs at specific homogeneous sites in the sorbent and there is not significant interaction between the adsorbed molecules and the sorbent. The sorbent is saturated after one layer or adsorbed molecules; thus, a plateau is reached.$${{{\rm{q}}}}_{{{\rm{e}}}}=\frac{{{{\rm{q}}}}_{\max }\cdot {{{\rm{K}}}}_{{{\rm{L}}}}\cdot {{{\rm{C}}}}_{{{\rm{e}}}}}{1+{{{\rm{K}}}}_{{{\rm{L}}}}\cdot {{{\rm{C}}}}_{{{\rm{e}}}}}$$Freundlich model. Empirical model based on multilayer adsorption over heterogeneous surfaces; thus, a plateau is not reached.$${{{\rm{q}}}}_{{{\rm{e}}}}{={{\rm{K}}}}_{{{\rm{F}}}}{\,\cdot \,{{\rm{C}}}}_{{{\rm{e}}}}^{\,{{\rm{n}}}}$$Sips model (Langmuir-Freundlich model). Semiempirical model that combines Langmuir and Freundlich models. Langmuir model that includes a constant associated to the adsorption capacity since it predicts there is a full coverage of the surface; thus, a plateau is reached.$${{{\rm{q}}}}_{{{\rm{e}}}}=\frac{{{{\rm{q}}}}_{\max }\cdot ({{{{\rm{K}}}}_{{{\rm{s}}}}\cdot {{{\rm{C}}}}_{{{\rm{e}}}}})^{{{\rm{n}}}}}{1+({{{{\rm{K}}}}_{{{\rm{s}}}}\cdot {{{\rm{C}}}}_{{{\rm{e}}}}})^{{{\rm{n}}}}}$$Redlich-Peterson model. Semiempirical model that combines Langmuir and Freundlich models. It facilitates the validation of the data close to the equilibrium conditions.$${{{\rm{q}}}}_{{{\rm{e}}}}\,=\,\frac{{{{\rm{K}}}}_{{{\rm{RP}}}}\cdot {{{\rm{C}}}}_{{{\rm{e}}}}}{{1\,+\,{{\rm{\alpha }}}\,\cdot \,{{\rm{C}}}}_{{{\rm{e}}}}^{{{\rm{\beta }}}}}$$Czinkota multi-step model (2 steps). It assumes multi-step adsorptions over a heterogeneous surface and validates sequential filling of different pores in the sorbent. Mathematically, it is expressed as the sum of Langmuir isotherms.$${{{\rm{q}}}}_{{{\rm{e}}}}\,=\frac{{{{\rm{a}}}}_{1}{\cdot {{\rm{K}}}}_{1}{\cdot {{\rm{C}}}}_{{{\rm{e}}}}}{{1+{{\rm{K}}}}_{1}{\cdot {{\rm{C}}}}_{{{\rm{e}}}}}+\frac{{{{\rm{a}}}}_{2}{\cdot {{\rm{K}}}}_{2}\,\cdot \,[({{{\rm{C}}}}_{{{\rm{e}}}}\,-\,{{\rm{b}}})\,+\,|{{{\rm{C}}}}_{{{\rm{e}}}}-\,{{\rm{b}}}|]}{{2+{{\rm{K}}}}_{2}\,\cdot \,[({{{\rm{C}}}}_{{{\rm{e}}}}\,-\,{{\rm{b}}})\,+\,|{{{\rm{C}}}}_{{{\rm{e}}}}\,-\,{{\rm{b}}}|]}$$The adsorbed amount at equilibrium (qe, mg·g−1 or mol·mol−1) and the maximum adsorption capacity (qmax, mg·g−1 or mol·mol−1) were expressed as mg or mol of analyte per g or mol of HRhMOP in the membrane. Ce is the analyte concentration in the solution at equilibrium (mg·L−1), KL is the Langmuir constant (L·mg−1 or L·mol−1), KF is the Freundlich constant (L·g−1 or L·mol−1), n is the heterogeneity factor in Freundlich and Sips models, KRP (L·g−1 or L·mol−1) and α (L·g−1 or L·mg−1) are the Redlich-Peterson constants, β is exponential Redlich-Peterson term, KS is the Sips constant (L·mg−1 or L·mol−1). In the case of the Czinkota multi-step isotherm, K1, and K2 refer to the adsorption rate constants at the different steps (L·mg−1 or L·mol−1), a1 and a2 corresponds to the adsorption capacity at the different steps (mg·g−1 or mol·mol−1), and b is the critical concentration limit. The chi-square value was used to evaluate the goodness of the fitting.The detection of PPCPs at low concentrations was evaluated using pure PEBA and PNM/PEBA 11 wt% membranes of 0.5 ×0.5 cm. The membranes were immersed in 10 mL of Milli-Q, tap water or river water spiked with the PPCPs at different concentrations: 0.5, 1.5, 3.5, and 10 µg·L−1. The experiments were carried out under stirring at 500 rpm for 24 h. Then, the membranes were taken out from the solution, dried with a tissue, and immersed in 100 µL of a 1:1 mixture of acetonitrile and Milli-Q water containing formic acid 0.02% v/v. The solution was stirred using vortex for 5 min to release the analytes from the membrane. The final extract was directly injected in the HPLC-UV-visible instrument to determine the PPCPs concentration. The experiments in Milli-Q water at 10 µg·L−1 were carried out by triplicate using three different membranes.The enrichment factor (EF) was calculated by comparing the concentration of the final extract and the initial concentration of the solution. The LODs of the microextraction procedure were estimated as the ratio between the LOD of the instrumental HPLC-UV-visible determination (see Supplementary Table 5) and the calculated EF.$${{\rm{EF}}}\,=\,\frac{{{{\rm{C}}}}_{{{\rm{extract}}}}}{{{{\rm{C}}}}_{{{\rm{initial}}}}},{{{\rm{LOD}}}}_{{{\rm{microextraction}}}}\,=\,\frac{{{{\rm{LOD}}}}_{{{\rm{instrument}}}}}{{{{\rm{E}}}}_{{{\rm{F}}}}}$$Simulation studiesTo prepare a system for analysis, a single unit cell of HRhMOP(bix)6 is produced with ToBaCCo 3.040,41,42. Forcefield parameters are obtained with lammps-interface43, obtaining UFF parameters44.With the crystalline system, an Atomic Biasing Force (ABF) simulation is run using SSAGES45 on the crystalline system with a single inserted molecule (either DCF or PaPAA) to obtain a 2D free energy profile averaged along the x-axis. DCF and PaPAA are modeled using the OPLS force field46. LAMMPS simulations consisted of a timestep of 1 fs, the Verlet integrator, energy minimization using a combination of steepest descent and conjugate gradient to a final criterion of 10−4  kcal mol−1 for energy and 10–4 kcal mol−1 Å−1 for force. Fix ssages is used to run SSAGES, and fix NVT is used for movement of atoms at a temperature of 298 K and a damping time of 100 fs, except for 1 atom in the framework which is held fixed so the unit cell does not shift dramatically. The free energy calculation with ABF used the z and y coordinates as collective variables with 250 bins for each one. The simulation was performed using 20 walkers and ran for 75 ns.Additionally, as a measure of the HRhMOP(bix)6 structure deformation, the end-to-end bix linker lengths (nitrogen to nitrogen distance) from the molecule trajectory file was calculated, and plot it as a function of the distance from the molecule of interest, either DCF or PaPAA.
Pore-networked membrane using linked metal-organic polyhedra for trace-level pollutant removal and detection in environmental water
