Membrane Functionalization Approaches toward Per- and Polyfluoroalkyl Substances and Selected Metal Ion Separations

Abstract

Adsorption and ion exchange technologies are two of the most widely used approaches to separate pollutants from water; however, their intrinsic diffusion limitations continue to be a challenge. Pore functionalized membranes are a promising technology that can help overcome these challenges, but the extents of their competitive benefits and broad applicability have not been systematically evaluated. Herein, three types of adsorptive/ion exchange (IX) polymers containing strong/weak acid, strong base, and iron-chitosan complex groups were synthesized in the pores and partially on the surface of microfiltration (MF) membranes and tested for the removal of organic and inorganic cations and anions from water, including arsenic, per- and polyfluoroalkyl substances (PFAS), and calcium (hardness). When directly compared with beads (0.5–6 mm) and crushed resins (0.05 mm), adsorptive/IX pore-functionalized membranes demonstrated an increased relative sorption capacity, up to 2 orders of magnitude faster kinetics and the ability to regenerate up to 70–100% of their capacity while concentrating the initial solution concentration up to 12 times. The simple and versatile synthesis approach used to functionalize membranes, notably independent of the polymer type of the MF membrane, utilized pores throughout the entire cross section of the membrane to immobilize the polymers that contain the functional groups. Utilizing the pore volume of commercial membranes (6–112 mL/m²), the scientific weight capacity of the polymer (3.1–11.5 mequiv/g), and the synthesis conditions (e.g., monomer concentration), the theoretical adsorption/IX capacities per area of the membranes were calculated to be as high as 550 mequiv/m², substantially higher than the 175 mequiv/m² value needed to compete with commercially available IX resins. This work therefore shows that pore functionalized membranes are a promising path to tackle water contamination challenges, lowering separation diffusion limitations.

Graphical Abstract

1. INTRODUCTION

Water resources throughout the globe are becoming increasingly more complex and contaminated,^1-4 and combined with global warming,⁵ the accessibility of clean water is becoming less predictable. Whether for the production of safe drinkable water, clean water for reuse or repurpose, or pure water for industrial manufacturing, separation technologies for purifying water are critical to ensure a solution to these challenges. Particularly, the separation of ions and small molecules requires proper understanding of the nanoscale interactions and transport mechanisms,⁶ as opposed to larger solutes that can be removed by size screening or other simple approaches. Most of the common environmentally relevant ions can be classified as organic or inorganic and as anions or cations. Some examples of environmentally relevant ions that demand removal from water sources are arsenic (inorganic oxo-anion)^7-9 with a set limit for drinkable water of 50 μg/L,¹⁰ per- and polyfluoroalkyl substances (PFAS, organic ions)^11,12 with a proposed drinkable water level as low as 4 ng/L,¹³ and calcium (hardness, inorganic cations), which is safe but at levels close to 100 mg/L or higher (low-medium hardness), it increases the chances of scaling in households and industrial processes.

Some of the current and most common technologies to separate ions from water include adsorption, ion exchange (IX), and membrane processes.^14-19 The first two group includes carbon-based, biobased, metal organic frameworks (MOFs), polymeric, and inorganic adsorbents.²⁰ Two popularly applied technologies are activated carbon (AC)²¹ and ion exchangers.²² These two materials can be used for a broad number of separation applications, do not demand highly trained operators to operate, their processes involve low capital cost and moderate maintenance requirements,²³ and these can be regenerated to increase the lifespan of the adsorbents. However, these materials suffer from diffusion limitations,²⁴ complex regeneration matrices^25,26 or byproducts,^27-29 and affinity for certain ions (e.g., short chain PFAS).³⁰ Most common AC/IX beads (>600 μm) are porous solids that allow for water to diffuse into the porous network, enabling solid/liquid separation (i.e., adsorption/IX of ions). This means that ions need to travel and diffuse from the bulk fluid to the inner pores to adsorb, reducing the kinetics of the process and potentially the useful capacity.

Recently adsorptive membranes have gained significant research attention with publications increasing by four folds in the past decade,^15,31-33 offering promising capabilities to help overcome current challenges that adsorptive polymers present in a bead format.^34-37 A good example is shown in the chromatography adsorption processes. In a study by Boi et al.,³⁸ breakthrough curves for adsorptive membranes were shown to be independent of the flow rate, as opposed to IX packed columns which were dependent. The use of convective flow through the membrane pores may ensure better access of the adsorptive active sites, a major benefit of adsorptive membranes in removing contaminants of concern. There are several approaches to manufacture adsorptive membranes,³¹ such as modifying MF membranes with surface grafting from chemistries or adding a nanolayer in the surface, or creating homogeneous membranes from blending or copolymerization techniques.

In this study, the pores and the surface of the microfiltration membranes are functionalized with polymers that contain specific functional groups (e.g., weak and strong acids, strong bases, and iron chitosan complex). These polymers can be designed to absorb/IX a diverse range of environmentally relevant ions from water. Some of the polymers prepared in this study have been previously reported in the literature;^{32,34,35,39-43} however, the scope of these previous studies is limited to the absolute separation values obtained as resins or membranes. This is the first time that new and previously studied functionalization approaches are implemented and holistically compared to showcase membranes as platforms that can help overcome intrinsic diffusion limitations that conventional adsorption technologies present. The goal of this work, as summarized in Figure 1, is (1) to study the benefits of using functionalized adsorptive/IX membranes as compared to traditional technologies (e.g., IX resins), such as improved kinetics, better utilization of the adsorption/IX sites, and potential regeneration benefits, while considering scale-up aspects, and (2) to present simple and versatile methods to add adsorptive/IX polymers into microfiltration (MF) membranes, independent of the MF membrane polymer type, to support the applicability of these membranes for the removal of a diverse range of environmentally relevant ions.

Figure 1.

Summary diagram. Adsorptive/IX membranes in this study: overall scopes, applications for water remediation, membrane functionalization approach, and the rise of benefits from the membrane platforms and their comparison against commercial and synthesized polymers. MF1 and MF2 are the two microfiltration membranes used in this study.

2. METHODS

2.1. Materials.

The following chemicals were used for synthesis of the adsorptive/IX polymers, water filtration tests, and material characterization: iron(III) nitrate nanohydrate (ACS reagent grade, SigmaMillipore), chitosan (200–600 mPa s, 0.5% in acetic acid, TCI), nitric acid (trace metal analysis grade, Aristar Plus BDH), sodium hydroxide (10N reagent grade, VWR life science), ASTM type 1 water (ACS reagent grade, RICCA), ammonium persulfate (98%, Acros Organics), methacrylic acid stabilized (99.5%, Acros Organics), 2-acrylamido-2-methylpropanesulfonic acid (99%, Sigma-Aldrich), (3-acrylamidopropyl) trimethylammonium chloride stabilized (74–76% in water, TCI), N-isopropylacrylamide stabilized (>98%, TCI), N,N′-methylenebis(acrylamide) (99+%, Thermoscientific), sodium hydrogen arsenate heptahydrate (ACS grade, Alfa Aesar), sodium perfluorooctanoate (97%, Alfa Aesar), potassium perfluorooctanesulfonate (98%, Matrix Scientific), undecafluoro-2-methyl-3-oxahexanoic acid (GenX) (97%, Matrix Scientific), hepta-fluorobutyric acid (98%, Aldridch), nonafluoro-1-butanesulfonic acid (>98%, TCI), calcium chloride (ACS grade, BDG), and neodymium-(III) chloride anhydrous (99.9%, Alfa Aesar). Amberlite IRC120 and IRC83 ion exchange resins were obtained from Thermo Scientific and Dupont, respectively. Millipore Express PLUS Membrane Filter polyether sulfone (PES) 0.45 μm pore size membranes and PV650 polyvinylidene fluoride (PVDF) membranes from Solecta, Inc., were used for the membrane functionalization process. Additional PVDF 0.45 μm and polycarbonate membranes from Millipore were used for pore volume measurements.

2.2. Adsorption/IX Tests.

Adsorption/IX studies were conducted in two different modes. For the polymer resins, batch studies in a centrifuge tube (as illustrated in the Supporting Information (SI)) assisted by an orbital shaker at 250 rpm was used. Samples (aliquots of 500 μL) were collected over time for kinetic studies, and initial and final samples were collected for isotherm studies. More details can be found in the SI. For the membrane experiments, dead-end filtration cells were used. For calcium and PFAS experiments, a stainless-steel Sterlitech P/N 4750, with a membrane active area of 14.6 cm² (2.26 in²) was preferred, while for the arsenic studies a polysulfone-body 50 mL Amicon stirred cell from Millipore, with 13.4 cm² active membrane area, was preferred to avoid potential arsenic adsorption on the cell. Polymers and membranes were initially primed to their sodium or chloride form unless they came in this form from the supplier. Regeneration of the polymer and membranes was achieved using benign solutions; 90000 mg/L sodium chloride for the acid anion exchange polymers was used and 0.7 M sodium hydroxide solution for Fe-Chi materials.

2.3. Analytical Methods and Materials Characterization.

The depth profile of the membranes was studied using Kα X-ray photoelectron spectroscopy (XPS) with an Al Kα X-ray monochromator source, from Thermo Scientific. The elemental composition was measured at five different levels starting from the surface and after intervals of 300 s of etching using an argon ion gun with assistance of the flood gun to minimize surface charging. Error bars in the plots represent the standard deviations from the analysis of two different areas.

A focused ion beam/scanning electron microscope (FIB/SEM) dual beam (Helios Nanolab 660, FEI/ThermoFisher Scientific, Hillsboro, OR, USA) was used for sample preparation and imaging of the surface and cross section. Elemental analyses were conducted immediately after FIB sectioning without exposure to air in the same dual beam system using Energy Dispersive Spectroscopy (EDS, Oxford Instruments X-Max^N 80 mm² detector). Additional sample preparation details are presented in the SI.

High resolution images of the polymers were acquired using a VHX-7000 digital microscope (Keyence, Osaka, Japan), assisted by reconstruction of fully focused images captured at different focus distances.

Arsenic quantification was performed using a NexION 5000 PerkinElmer inductively coupled plasma-mass-spectrometer (ICP-MS). Argon plasma and Ge internal standards were used for the analysis using the O2 DRC mode. Calcium quantification was performed using a 7800 Agilent ICP-MS. Argon plasma, assisted by helium to remove interferences, and scandium as internal standard were used. For samples containing calcium in a high salinity matrix, the calibration curve was prepared with similar sodium chloride content. Samples were prepared in 1% nitric acid, and analyses were run in triplicate with repeating standards/blanks every 15 samples to ensure instrument stability. Error is reported as the standard propagated error.

PFAS analysis was performed using a Thermo Scientific TSQ Altis Plus triple quadrupole mass spectrometer equipped with a HESI ionization probe coupled to a Thermo Scientific Vanquish Horizon UHPLC system fitted with a Thermo Scientific PFC-free kit (P/N 80100–62142). Detailed information is presented in the supporting material.

3. RESULTS AND DISCUSSION

3.1. Preparation and Properties of the Adsorptive/IX Polymers and Membranes.

The methods used for the synthesis of the different polymers, whether as pore and surface coating of the microfiltration membranes or as the polymers alone (i.e., hydrogels or iron chitosan), referred to in this study as “beads”, were similar. The simple functionalization approach consists of passing a polymer or monomer containing solution through the membranes, then applying high pH or heat, resulting in randomly and stably coating pores and partially surfaces, as illustrated in Figure 1. For the case of hydrogels in situ temperature initiated free radical polymerization was used.^44-46 Poly methacrylic acid (PMAA),⁴⁷ poly 2-acrylamido-2-methylpropanoesulfonic acid (PAMPS),⁴¹ and 50/50% poly(3-acrylamidopropyl) trimethylammonium chloride and N-isopropylacrylamide (DMAPA-Q/PNIPAm), the solutions containing the monomers, cross-linker, and initiator were prepared and passed through the membranes. To the best of our knowledge the use of DMAPA-Q/PNIPAm has not been previously reported in literature, and it consists of a random polymerization of quaternary amine groups and NIPAm thermoresponsive amide groups creating a stable polymer that now has additional adsorption capabilities and has potential benefits toward adsorption regeneration using the lower critical solution temperature of the synthesized copolymer.³⁴ The leftover solutions with the reactants were placed on Petri dishes and then placed into the oven at the same time as the membranes. Following heat treatment (85 °C for 2 h), membranes were rinsed in deionized water, dried, and stored; hydrogels were rinsed with deionized water several times, freeze-dried, and stored. Hydrogels were then primed with 23500 mg/L NaCl solution, rinsed, and freeze-dried again for later use in IX experiments. For the case of Fe-Chi,³⁹ chitosan and iron nitrate were dissolved into 0.1 M HNO₃. A syringe with an 18G needle was loaded into a syringe pump to pour drops of the Fe-Chi solution into a 0.4 M NaOH solution for the fabrication of the beads, while membranes soaked (assisted by vacuum filtration) with the Fe-Chi solution were placed directly into the 0.4 M NaOH solution. A detailed description of the synthesis conditions of each of the polymers used in this study is shown in Table S1.

Commercial ion exchangers used in this work and membranes functionalized with IX polymers were previously primed with NaCl or NaOH solutions, except for strong acid and strong base Amberlite resins that were received in their sodium or chloride form directly from the supplier.

In order to compare adsorption/IX capacities among the different materials, applications, and the format (beads or membrane) of the polymers, this study puts an emphasis on the use of the same or comparable measurement units and terminologies (as further described in the SI). A summary of all of the polymers used in this work and their specifications are presented in Table 1. The scientific weight capacities (mequiv/g) represent the number of ionogenic groups (sulfonic, carboxylic, or amine) per specific amount of ion exchanger. For complex groups (iron complex) this value was calculated as the absolute maximum, therefore the number of amine groups available to complex iron. These values set the maximum ion exchanging capacity per unit of mass by these polymers, assuming no multilayer adsorption or other contributions. The values for the commercial Amberlite ion exchangers were calculated using their datasheet information (Table S2), and more details of the calculations are presented in the SI. For synthesized polymers, the main monomer and its active sites were considered. The scientific weight capacities were calculated considering the molecular weight of the most abundant repeat unit (Table S2) and one active site (functional group) per repeat unit, and the values are summarized in Table 1. The lower the molecular weight of the repeat unit, the higher the scientific weight capacity. Commercial ion exchangers are presented for comparison for the case of calcium removal, while DMAPA-Q/PNIPAm and Fe-Chi materials are compared as polymer beads against pore and surface functionalized membranes.

Table 1.

Summary of Adsorbent/IX Polymers Used in This Study

resin	commercial/ synthesized	type	functional group	polymer	scientific weight capacity (max) (mequiv/g)	application in this study
IRC120	commercial	strong acid	sulfonic acid	styrene- divinylbenzene	6.4	hardness removal (calcium)
IRC83	commercial	weak acid	carboxylic acid	cross-linked acrylic	7.3
PAMPS	synthesized	strong acid	sulfonic acid	cross-linked acrylic	4.8
PMAA	synthesized	weak acid	carboxylic acid	cross-linked acrylic	11.6
Fe-Chi	synthesized	complex	iron-amine	chitosan	3.2–6.3a	arsenic removal
DMAPA-Q/ PNIPAm	synthesized	strong base	quaternary amine	cross-linked acrylic	3.1b	pfas removal

^aThis value is calculated assuming iron forms a bidentate or monodentate complex with ALL nitrogen available from the chitosan. Since a hundred percent complexation is not likely, the true scientific capacity should be lower and proportional to the occupancy of nitrogen complexed with iron.

^bThis value just accounts for the quaternary amine group and does not account for the hydrophilic adsorption from the PNIPAm moieties. Therefore, the true scientific weight capacity should be higher.

The separation mechanisms for the different functional groups in each polymer (Table 1) are well studied in ion exchange and adsorption literature.^34,39,48,49 The mechanism for sulfonic and carboxylic acid groups is based on strongly acidic and weakly acidic cation exchange (such as calcium ions), respectively. Quaternary amine groups are well-known as strong bases and therefore ion exchange of anions (such as several types of anionic PFAS) and PNIPAm has been demonstrated to have adsorptive properties toward PFAS,³⁴ making this a dual IX/adsorptive mechanism. For the case of Fe-Chi, as a hard acid, Fe(III) interacts favorably with partially protonated amine groups at neutral pH chitosan,⁵⁰ and it has previously been shown to form primarily bidentate complexes, with Fe(III) likely bridging two amine groups.³⁹ The ability of Fe-chitosan to selectively adsorb arsenic (As(V) and As(III)) in the presence of a phosphate competitor likely comes from its ability to form inner-sphere complexes with the arsenic oxoanion, supported by Pearson hard soft acid base principle (HSAB) and electrostatic interactions.^39,49

3.2. Lab-Synthesized and Commercial Adsorbents.

As previously mentioned, this study compares the polymers alone referred to as beads in this work against the polymers functionalized in the membrane domain and, in one case, against commercially available IX resins (beads). Figure S1 presents micrographs of the polymer beads by optical microscopy. While commercial ion exchangers (IRC120 and 83) present uniform smaller sizes in the range 0.5–0.75 mm, the synthesized beads were larger (in the range 3–6 mm, Table S4). To prove, quantify, and compare the existence of diffusion limitations on these materials, the strong and weak acid IX beads were also crushed down into a powder form (Figure S2), with sizes in the range of 40–57 μm. The swelling ratios of these polymers are summarized in Figure S3. The highest swelling values were for PAMPS and DMAPA-Q/PNIPAm hydrogels, with average values of 8300% and 5100%, and 1100% and 43% for PMAA and Fe-Chi, respectively. These swelling ratios are affected by high concentration solutions, as observed in Figure S4. Commercial IX beads are designed to experience low swelling, with values of around 20%.

3.3. Functionalized Adsorptive/IX Membranes.

The PV650 MF membranes were functionalized with PMAA, PAMPS and DMAPA-Q/PNIPAm, while the PES MF membranes were functionalized with Fe-Chi, as illustrated in Figure 1. All four types of membranes are shown in Figure S5. Both the surfaces and cross sections of the membranes were investigated to explore the location and semiquantitation of the functionalized polymer.

First, the location of the functionalized polymer was studied on the membrane surface (Figure 2(A.1-6)). Reduction of the pore size and surface porosity was observed for all membranes. The surface of the functionalized membranes maintained the pore shape, becoming surface roughness, meaning that partial coverage was obtained and that at the most a thin layer coated the top surface. Figures S7 and S8 support these findings and present some pore openings, differences in surface morphology between rich and poor functionalized areas, and the cracks found on some PMAA and PAMPS membranes. These cracks may be produced by the swelling behavior of these hydrogels, whereas the Fe-Chi presented no cracks.

Figure 2.

Surface and cross-sectional characterization of the four functionalized (2, 3, 4, and 6) and two blank membranes (1 and 5). Organized as follows: 1: PV650 (blank), 2: DMAPA-Q/PNIPAm, 3: PMAA, 4: PAMPS, 5: PES (blank), 6: Fe-chi. (A) Scanning electron microscopy (SEM) images from the top surface of the membranes at the same 25,000X magnification. PV650 Membranes were ion exchanged with heavy metals (Nd³⁺ or HAsO₄^2-) for better identification of the functional groups from the polymer added into the membrane domain. Panel (A) also includes the atomic percent of the characteristic metal ions from X-ray photoelectron spectroscopy data obtained after etching 1250 s to ensure proper reading of nonpotentially contaminated surface (error indicates standard deviation from 2 samples). Figures in the second column (B) show the overlap of the cross section SEM image and the EDS map from the specified region. Purple corresponds to the nonwoven fabrics from the backing support of the PV650 on 1–4 identified by oxygen as the characteristic element. PES membranes are unsupported. Light blue corresponds to the polymeric MF membrane represented by fluorine or oxygen for PV650 and PES membranes, respectively. Yellow corresponds to the active adsorption and IX sites from the functionalized polymers. Membranes 2–4 were ion exchanged with heavy easy-to-detect elements (the same ones used in XPS analysis). The elements for yellow areas are arsenic for (B.3), neodymium for (B.3) and (B.4) (not detected), and iron for (B.6).

Second, a high-sensitivity subsurface XPS analysis was used to quantify the presence of the polymers on the surface of functionalized membranes. Fe-Chi membranes already contain iron as a heavy metal, facilitating detection. PMAA and PAMPS membranes were ion-exchanged with neodymium (NdCl₃ solution), and the DMAPA-Q/PNIPAm membranes were ion exchanged with arsenic (HAsO₄²⁻). Therefore, iron, neodymium, and arsenic represent the active adsorption/IX sites of the functionalized polymers. The atomic percent values of neodymium, arsenic, and iron were estimated based on the weight gain of the membranes and the atomic percent of the carboxylic acid or amine groups within the respective polymers, assuming the repeat units to be the same as the main monomers. Concentrations were estimated to be 0.5–1.4% for Nd in PMAA (depending on mono-, di-, and trivalent bonding nature), 0.1–0.4% for Nd in PAMPS, 0.5–0.9% for As in DMAPA-Q/PNIPAm (mono- or divalent bonding), and 2–7% Fe in Fe-Chi (based on XPS N% and mono- or bidentate complex). After enough etching of the membrane surfaces to prevent contamination (depth profiles shown in Figure S6), the calculated concentrations of metals (shown in Figure 2(A)) for the PMAA, PAMPS, and DMAPA-Q/PNIPAm functionalized membranes were 0.8%, 0.7%, and 0.2%, respectively. These values fell within the estimated concentrations presented previously. For the Fe-Chi membrane, concentrations of iron were above the anticipated values, with Fe/N ratios between 2 and 4. Steady concentrations of the polymers on the top surface (and bottom surface as shown in Figure S15) were also observed, presented as evolution of the concentration over etching time (potentially the top few micrometers or less).

Finally, the cross sections of the membranes were characterized. SEM images of the cross sections for all studied membranes, at various magnifications, are shown on Figures S9 and S10. Figure 2(B.1-6) presents the energy dispersive spectroscopy (EDS) maps, overlaid on the SEM images. Blue and purple correspond to the MF membranes, represented by fluorine and oxygen, respectively. For the case of PV650 membranes, fluorine corresponds to the polyvinylidene fluoride polymer, and oxygen corresponds to the polyester nonwoven support fabric. For the unsupported PES membranes, oxygen corresponds to the poly(ether sulfone) polymer. Yellow corresponds to the metals of interest, as explained in the analysis of the XPS results. Figures S11 and S12 show the EDS peak location and intensity of the elements of interest. It can be observed that the adsorptive/IX polymer coated the entire cross section of all functionalized membranes, except for PAMPS–PV650, in which at concentrations below 0.2 at. % of neodymium (from XPS data) EDS cannot detect the signal of the metal. It can be inferred that this simple membrane functionalization approach utilizes the entire cross section of the MF membranes, and therefore, the maximum adsorption/IX capacity of these membranes can be estimated simply knowing the total pore volume per membrane area. Figure S14 shows the line scan of the Fe-Chi-PES membrane, where a higher atomic percent was found in the top surface and the large-pore zone of the cross section. The top surface was exposed first on the infiltration process of the polymeric solution (Figure 1). It is suspected that the pores were able to capture more solution, while the solution may not have reached all the membrane pores in the middle section. Also, large pores sections were likely more easily filled with polymeric solution and therefore offer an adsorptive/IX-polymer rich zone.

The addition of a polymer in the membrane pores is expected to affect the water permeance of these membranes because of the reduction of the water flow path. To evaluate the operational impact, we measured the water permeability coefficients of the functionalized membranes. The water permeance of the modified membranes dropped as little as three times and as much as a hundred times, with values between 100 and 1000 LMH/bar (Figures S16 and S17). These results are promising as they allow for low-pressure operations. The amount of polymer loaded into the pores as well as the random nature of this functionalization affect the water permeance values, as it has been shown in a previous study.⁴⁴ Hereby, the values from the water permeance (with a range within an order of magnitude) and the SEM (with surface coverage variability) prove that regardless of the heterogeneities of this synthesis approach, these membranes preserve good conditions for water transport and great separation performances (as presented later in this work).

3.4. Removal and Regeneration of Inorganic Cations (Calcium).

This section has as its objective the application of the proposed simple, broad, and versatile approach to create IX membranes that work for separating inorganic cations. Additionally, an in-depth analysis of the diffusion limitations that commercially available ion exchangers face was conducted to highlight some of the key benefits of these functionalized membranes. To this end, all of the formats of the IX polymers were studied. This includes the commercial IX resins (IRC120 and 83) and synthesized IX polymers (PAMPS and PMAA) in both bead and powder form as well as functionalized membranes with the IX polymers.

3.4.1. Comparing Synthesized and Commercial Strong and Weak Acid Cation Exchange Polymers.

The bead and powder polymers (Figure S1) were tested for an ion exchange isotherm experiment in a batch study for about a hundred h for different concentrations of CaCl₂. Results are summarized in Figure 3(A). Strong acid polymers (IRC120 and PAMPS) presented higher equilibrium capacity (q_e) for calcium ions than weak acid polymers, and the synthesized polymers closely followed values from commercial IX resins. The Misak isotherm model (eq 1)^51,52 for heterovalent exchange was used to fit the experimental data and calculate the maximum capacities:

$$\frac{C_A}{q_A}=\frac{\left(\frac{x}{y}\right)C_A}{Q}+\frac{{(C_0-x{C}_A)}^{x∕y}}{y^{x∕y}Q{K}^{1∕y}}$$ (1)

where C_A (C_e) is the equilibrium concentration (mmol/cm³), q_A is the mequiv/g of sorbate A sorbed by the solid (sorbent), x and y are the charges of ions A and B, Q is the maximum sorption capacity (mequiv/g) of the solid, C₀ is the total concentration (mequiv/cm³) of A and B in solution, and K is the selectivity coefficient.

Figure 3.

Ion exchange capacity for calcium and kinetics from weak and strong acid commercial (IRC120 and IRC83) and synthesized (PAMPS and PMAA) ion exchange polymers. (A) Isotherms from “bead” and “powder” ion exchangers. Fitted lines represent the model isotherm values obtained by using the Misak isotherm for heterovalent ions. Each batch study contained approximately 40 mg of polymer in 40 mL of CaCl₂ solution made in ASTM type I water, pH = 6, with their respective initial concentrations. (A.2) Summary of the maximum calcium exchange capacities (Q) obtained from the Misak model. (B) Cumulative calcium exchange over time (q_t). Dotted and segmented lines do not represent the model, and they are used just for visual help. Values from powder are represented by thin- outlined and filled markers, while those from beads are represented by thick-outlined markers. Calcium ions were measured by ICP-MS.

Q values (Figure 3(A.2)) indicated that strong acid IX polymers in their bead form followed closely the maximum weight capacities (Table 1), however, weak acid ion exchange polymers performed Q values below their maximum scientific weight capacity (Table 1), which can be partially attributed to their pK_a values being closer to the tested solution pH solution (pH = 6).

These polymers were also tested for their kinetics properties (Figure 3(B)) for the case with [Ca²+]₀ = 25 mg/L. Two main outcomes can be observed. First, the powder form of the polymers achieved faster and higher steady-state IX capacity values (q_t). This is an indication of overcoming the diffusion limitations offered in larger beads, achieving faster and more access to the active sites in the powder form. Second, synthesized polymers present more favorable kinetics than the commercial IX resins, which may be attributed to their swelling ratio (Figure S3). These outcomes directly relate to the intrinsic diffusion limitations that commercial IX resins experience. Both swelling properties and convective flow through the membrane domain are expected to help overcome the diffusion challenges when the adsorptive/IX membranes are used. It is worth noting that these experiments included high swelling ratio polymers and IX in a powder form, neither of which would be operationally viable due to the drastic pressure drop these would cause, and therefore, these results represent theoretical maximum comparisons to the membrane experiments.

3.4.2. Membrane Performance.

Results of the cumulative ions exchanged per mass of polymer within the membrane domain are presented in Figure 4, and the volume filtered was normalized per square meter of the membrane tested. The ion exchanging process between sodium co-ions and calcium counterions was well captured by the quantitation of these, with equivalents of sodium being released into solution following closely the ones from calcium being exchanged out from solution, in both PMAA and PAMPS PV650 membrane experiments, and under residence times (i.e., the time that the solution is exposed to the adsorptive or IX polymer inside the pores) as low as 2—7 s. Differences between calcium and sodium may be attributed to the H+ form of ion exchange sites (groups in an acid form). Great reproducibility was found on PMAA-PV650 membranes with the ion-exchanging results from two membranes falling on top of each other (Figure 4(A)). PAMPS—PV650 membranes were tested using the same method but passing the solution two times, one initial pass and a second postregeneration (Figure 4(B)). The calcium exchange after regeneration was even greater, which can be attributed to full regeneration and complete priming of the sulfonate groups with sodium. The IX capacity observed for PMAA—PV650 was higher than the capacity obtained from PAMPS—PV650 membranes, with values up to 5 mequiv/g compared to 2.5 mequiv/g, respectively, opposed to the results found with the bead and powder experiments. These results suggest that weak acid IX polymers in the membrane domain can make greater use of their IX capacity (Table 1) as compared to their bead form (Figure 3(A)). Three cycles of Ca²+ exchange and regeneration with Na+ of a PMAA—PV650 membrane are shown in Figure 4(C). Operating ion exchange capacity of about 3.5 mequiv/g, higher than the bead and powder values obtained from the PMAA polymer, and stable regeneration cycles were observed. These results suggest that these membrane platforms can be easily regenerated with traditional IX regenerating methods with minimum decrease of the operating capacity (70% of the useful capacity) while offering concentration factors from 3 to 6 times.

Figure 4.

Ion exchange capacity of calcium on weak and a strong acid cation exchanging membranes as mequiv of Ca²⁺ per mass of IX polymer on the membrane. (A) Exchange between sodium and calcium on the sodium-primed PMAA membranes during calcium IX, two membranes are presented as m-1 and m-2 (4% weight gain and 8% weight gain respectively). (B) Exchange between sodium and calcium on the sodium primed PAMPS membrane (6% weight gain), first pass and second pass after regeneration are illustrated. Calcium solution made in ASTM type I water, with [Ca²⁺]₀ = 25 mg/L made with CaCl₂, pH = 6, T = 22 °C. (C) Ca²⁺ exchange and regeneration with Na+ cycles for the PMAA membrane. Regeneration was done with 50 mL of a 90,000 mg/L NaCl solution. Cumulative ions exchanged are normalized to mequiv to easily compare H+ with Ca²⁺ and divided by the mass of polymer in the membrane sample tested. Volumes were normalized by the membrane area. A dead-end cell P/ N 4750 from Sterlitech was used. The active surface membrane area tested was 0.00146 m² and the volume filtered was ~210 mL. Water flux was kept between 200 and 460 L/(m² h). Ca⁺ and Na+ were measured by ICP-MS.

3.5. Removal of Inorganic Oxo-anions (Arsenic).

This section has the objective to prove the application of the proposed simple, broad, and versatile approach to create adsorptive membranes that enable the separation of inorganic anions. Additionally, we aim to show a different functionaliza-tion approach (transition metal chitosan polymer) and a different porous membrane (poly(ether sulfone)) than used previously to highlight the wide applicability of the proposed adsorptive membrane synthesis approach. Notably, the concentration of ions dealt in this section is between 50 and 100 μg/L (in the upper range of high concentrations found in groundwater throughout the globe),⁷ saturating the membrane active sites more gradually, and consequently, the break-through curve can be captured more easily. The effects of some operating parameters, regeneration aspects, and application of this adsorptive membrane in more challenging solutions are also highlighted here.

Fe-Chi beads have previously been studied for their separation of arsenic.³⁹ The maximum adsorption capacity for arsenic (III and V) was found to be around 20 mg/g, equivalent to 0.27 mmol/g (reported in mmol since arsenic is a multivalent ion). This value falls about ten times below the maximum possible scientific weight capacity presented in Table 1, meaning that partial utilization of the total number of nitrogen groups available from the chitosan are complexing to iron. In comparison to these results, Figure 5(A) shows the cumulative arsenic(V) adsorbed (^mol/g) by the Fe-Chi- functionalized and the blank membranes. Clear adsorption capability by the Fe-Chi-PV650 membrane is observed at low- pressure requirements (0.14 bar). It is worth noting that when removing arsenic, where lower concentrations of solutions are found, the membrane is under a longer transient state before reaching full saturation, therefore, the cumulative adsorption values for As(V) are lower than the ones for calcium removal.

Figure 5.

Adsorption of arsenic(V) by Fe-Chi-PES membranes. (A) Cumulative adsorption of arsenic by blank PES and modified Fe-Chi- PES membranes. Adsorption is normalized by the amount of Fe-Chi polymerized on the PES membrane (4.7 mg), and the volume is normalized by the surface area of the membrane tested. The error bars represent the standard deviation of 2 samples tested with similar mass of polymerized Fe-Chi. (A.2) Absolute mass of arsenic adsorbed vs absolute mass of arsenic passed on a 1:1 proportion. Three different Fe-Chi-PES membranes were tested, and a PES blank is also included. (B) Adsorption/desorption cycles. Adsorption was performed at 3 different pressures (indicated), while desorption was at 0.34 bar for all cycles. Adsorption is normalized by the amount of Fe-Chi polymerized on the PES membrane (2.7 mg), and volume is normalized by the surface area of the membrane. A 50 mL Amicon stirred cell from Millipore was used for these experiments. The surface area of the membrane tested is (0.00134 m²). Arsenic was analyzed by ICP-MS. Adsorption solutions were at pH = 6.9. Desorption was performed by using a 0.7 M NaOH solution.

The adsorption breakthrough (i.e., when the As(V) adsorbed starts deviating from the 1:1 straight line in Figure 5(A.2)) for the 150 μm thick Fe-Chi-PES membrane, which represents the height of an adsorbent packed column, started closely after passing 9–12 μg of arsenic (equivalent to 6.7–9.0 mg/m²). After the breakthrough point, the quality of the filtered solution starts decreasing gradually to a point where it does not meet the recommended safe drinking water level (10 μg/L for arsenic). Figure S19 summarizes the filtered water quality obtained under different testing conditions and shows how stacking membranes (or increasing membrane capacity) may help overcome losses in the adsorption capacity when competing with other anions in solution. Remarkably, concentrations below the maximum contaminant level (MCL) were obtained when mean residence times were as low as 3.2 s, as compared to commercial technologies that operate at values close to 1.5 min, when filtering as much as 125 L/m² of a solution containing As(V) alone.

Figure 5(B) shows four adsorption/desorption cycles using 20 mL of a 0.7 M NaOH regenerating solution, suggesting that the adsorptive lifetime of these membranes could be extended through facile regeneration procedures with near zero losses of the operating capacity while obtaining up to 12 times concentration factors for arsenic (Figure S20). Longer mean residence times ensured higher adsorption capacities, as the observed adsorption of As(V) was higher in Figure 5(B) when values were 1.4 s compared to 0.08 s (i.e., 0.14 and 3.1 bar, respectively).

3.6. Removal of Organic Anions (PFAS).

This section has as objective to prove the synthesis and application of these IX/adsorptive membranes for applications with organic anions. An array of five different PFAS with different lengths and functional groups were tested. Figure 6 summarizes the different PFAS exchanged and adsorbed on a filtration experiment using DMAPA-Q/PNIPAm-PV650 and blank PV650 membranes. High exchange/adsorption of all PFAS passed through the functionalized membranes was observed (at the concentration tested), suggesting this platform could bring PFAS concentration to required low levels (not evaluated in this study). PV650 showed some minimum adsorption, which can be contributed to its fluorinated structure, but closely after passing a few ng of PFAS, it lost its adsorption capability. Figure S18 shows the percent removal of each of these PFAS. PFBA experienced a decreased performance down to 86% removal after 240 L/m², followed by GenX with a decreased removal down to 96% after filtering the same normalized volume, while PFOS, PFOA, and PFBS maintained removals above 99.6% under the same conditions. Decreased hydrophobicity and favorable partitioning in water of the short chain PFAS make these challenging to separate out from water.^25,53 The relative decreased in the removal for PFBA, PFBS, and GenX was not as significant as it has been in other adsorptive materials,^30,54 providing promising results to overcome the challenges of removing short chain PFAS.

Figure 6.

Captured PFAS as IX/adsorption. Ratio of mass (μg) adsorbed for five different PFAS versus the mass of PFAS passed through the membrane during the filtration experiment. Results in color (non-PV650) correspond to DMAPA-Q/PNIPAm-PV650 membrane filters, while the gray symbols correspond to unfunction-alized PV650 membrane filters. The active surface membrane area tested was 0.00146 m², and the volume filtered was ~350 mL. Water flux was kept between 80 L/(m² h) and P = 0.14 bar, and the feed concentration of PFAS was ~500 μg/L each.

3.7. Benefits of Adsorptive/IX Membrane Platforms.

The goal of this section is to highlight the benefits obtained from better accessibility to the active sites, such as faster adsorption/IX kinetics and higher useful capacities, by comparing the separation performances between the poly-merized membranes and the polymer beads at similar conditions. Comparing membrane dead-end cell filtration tests against polymer beads batch adsorption experiments is challenging. A detailed analysis of the ion exchanging variables and factors to be considered when comparing membranes to polymer beads is presented in the supporting material. Assuming that the properties of the IX polymers do not change between its pure form (beads) and as integrated on the IX membrane, and that the limiting step is internal diffusion, the two main variables of the system are the number of co-ions (i.e., active sites) and the number of counterions (i.e., ions in solution). Based on this we define the ratio R = #^c°^unterlons to facilitate a fair comparison between the two systems (Table S5). This ratio has two main implications. First, R values under unity mean that the system is demanding partial use of the adsorption/IX capacity, while values above unity mean that there are not enough active sites to exchange all ions from solution. Second, the higher the R value, the faster the IX kinetics on the system (i.e., the higher the fractional attainment of equilibrium U(t)). Figure S21 compares the cumulative calcium exchange of IRC120 in powder and bead forms for two different R values, and both higher useful capacity and kinetics were obtained for the higher R value experiment.

Taking this into account, we considered a conservative approach, in which adsorptive/IX membranes are fairly or less favorably compared to the polymer resins (e.g., lower R values when comparing adsorption/IX capacities or less mass of polymer when comparing removal rates); thus, the actual benefits could be greater. Figure 7(A and B) compares the weak and strong acid calcium exchange experiments for the beads, powder, and membranes. It can be observed that for the strong acid polymers, membranes outperformed the kinetics obtained by the powder polymer by more than eight times and three times for low and high R values, respectively, while exchanging similar calcium ions. For the weak acid calcium exchangers (Figure 7(B1-2)) PMAA membranes outper-formed the best polymer powder results with seven times higher calcium exchanged within 75% of the time (at low R values) and almost nine times higher calcium exchanged within an order of magnitude less time compared with both commercial and synthesized polymers (at high R values). The full data for the 25 mg/L calcium concentration, where R values are not equalized, are compared in Figure S22. Not only did membranes outperform the polymer beads, but they also outperformed the polymer powder (smaller size) form. Polymer powders are expected to outperform beads but these cannot be used in practical industrial processes.

Figure 7.

Comparison of the adsorption/IX rate of ions between the single polymers and the functionalized membranes for equivalent or less favorable conditions for the membrane experiment. Cumulative calcium exchange over time for the synthesized and commercial beads (“B”) and powder (“P”) single polymers, and membranes for (A) strong acid cation exchangers and (B) weak acid cation exchanger. Comparison is done as useful capacity over time at similar R= # counterions (mequiv)/# co-ions (mequiv) values, meaning that the maximum useful capacity for both membrane and the polymer alone is expected to be the same. Standard deviation on the R values represents the difference between the values from the polymers and the membranes. For (1) low R values, [Ca²⁺]₀ = 25 mg/L of CaCl₂ salt. For (2) high R values, [Ca²⁺]pol_ymeralo_ne,0 = 390 mg/L of CaCh salt, while the concentration for the membrane experiments was the same as in (1) but accounts for larger volumes filtered. The active surface membrane area tested was 0.00146 m² and the volume filtered was ~210 mL. Water flux was kept between 200 and 460 L/(m² h). (C) Percent removal of arsenic by membrane and polymer alone (beads) from a 40 mL solution containing 100 μg/L of arsenic. In this case, the least favorable case for the membrane experiment is set as the same arsenic required to adsorb but with less polymeric mass material (2.7 mf Fe-Chi used on the membrane and 35 mg Fe-Chi used on beads experiments). (C.2) Zoom-in to the membrane removal over the first hours. Calcium and arsenic ions were analyzed by ICP-MS and PFAS were analyzed LC-MS/MS. (D) Percent removal of total PFAS (PFOA, PFOS, GenX, PFBA, and PFBS) by DMAPA-Q/PNIPAm-PV650 and DMAPA-Q/PNIPAm beads. Both batch and filtration experiments treated solutions with a concentration of 500 μg/L of each of the PFAS, a fixed 200 mL volume for the batch and up to a 350 mL volume for the membrane filtration. The membrane contained 7.8 mg of adsorptive/IX polymer, and the batch study used 33 mg of polymer beads. (D.2) Zoomed-in view of the membrane filtration data.

Figure 7(C) compares the results obtained for the Fe-Chi polymers in their bead and membrane format. Since the removal of arsenic involves low concentrations of this ion, R values tend to be small (0.02 and 0.002 for membrane and beads, respectively). Instead of comparing same R values, here the removal efficiency (%) was compared, maintaining the same volume and concentrations of arsenic in solution, but having 35 mg of beads in the batch adsorption experiment (over time) and just 3 mg in the membrane experiment (in a recirculation mode). 99.7% removal of 100 μg/L As was achieved by the Fe-Chi-PES (membrane) in 30 min, nearly 2 orders of magnitude less time than the neat beads while producing a filtered solution with 12 times less arsenic.

For comparing the PFAS removal efficiency between polymer beads and functionalized membranes, the same solutions of PFAS were filtered but different volumes were used (350 mL for the membrane and 200 mL for the beads). The separation of PFAS obtained by the DMAPA-Q/ PNIPAm-PV650 membranes showed high separation rates in a short period of time (Figure 7(D)). The initial removal rates were as high as 100% for the sum of all the PFAS in solution, and after 3 h (passing an equivalent of 242 L/m²), the removal experienced a slight decrease to 96% total removal, as observed in Figure 8(D.2). Even though the volume filtered was smaller and the mass of polymer was larger than in the membrane experiments (33 mg and 7.8 mg for beads and membranes, respectively), the removal efficiency values obtained by the DMAPA-Q/PNIPAm beads oscillated between 30 and 40%, significantly lower than the removal rates obtained by the membranes.

Figure 8.

Analysis of the adsorptive/IX capacity of the functionalized membranes and its potential toward large-scale applications. (A) Estimation of the pore volume and volumetric porosity of different microfiltration membranes by membrane wetting with low surface tension solvents (IPA and Silwick). Error bars represent the standard deviation obtained from three different membranes tested. Mem-branes thickness is also indicated. (B) Membrane adsorption/IX capacity in mequiv/m² as a function of pore volume of the microfiltration blank membrane and the monomer concentration in the polymerization solution. A monomer with a fixed scientific weight capacity of 3 mequiv/g was used. PMAA-PV650 experimental useful capacity is presented as reference. (C) Comparison between the bed volumes obtained by IX membranes and ion exchangers as a function of the capacity of ion exchange materials and the concentration of environmentally relevant ions in solution, including PFAS, arsenic, and calcium. A range of modeled and experimentally obtained values, and commercial technical volume capacities (packed bed, fully swollen) were used for membranes and commercial IX, respectively.

Two main experimentally observed benefits of the function-alized membranes were presented above: faster kinetics and higher utilization of the active sites (by at least 1 order of magnitude in some examples). A third benefit arises from the original purpose for which the porous membranes were designed for: size exclusion. Micro- and ultrafiltration membranes, which would be the range for these postfunction- alized membranes, are design to exclude bacteria, viruses, proteins, and even large natural organic matter, depending on the surface pore size.⁵⁵ Therefore, these adsorptive membranes could offer a dual separation mechanism, adsorption/IX and exclusion, under a single driving force. This concept can be brought even a step further, and the membrane can be functionalized with a nanofiltration layer on the top surface, obtaining nanofiltration-like rejections while also offering adsorption capabilities.³⁵

3.8. Scale-Up Considerations.

3.8.1. Adsorptive/IX Membranes Capacity and Comparison against Commercial Ion Exchangers.

Up to this point, the comparison between functionalized membranes and polymer resins has been on the basis of adsorption/IX per unit of mass of the adsorptive/IX polymer. Even though adsorptive materials are sold in units of mass (or volume), membranes are sold in units of surface area. Therefore, to facilitate a reasonable comparison, the adsorptive/IX capacities from the membranes need to be transformed to unit of area (mequiv/m²). To further understand the maximum adsorption capacity of these platforms, the amount of mass of adsorptive/IX polymer that can be functionalized within these microfiltration membranes also needs to be studied. Ultimately, from an industrial process standpoint, it is important to know whether the competing technology will require the same, fewer, or more stages (and potentially capital cost) to achieve the same separation performance as the current technology.

First, the pore volume inside membranes was measured to know how much adsorptive/IX polymer can be possibly polymerized inside of these. A set of commercial micro-filtration membranes with different properties were tested to identify a range of pore volumes available on the market. The volumetric porosity of asymmetric PVDF and PES membranes was in between 60% and 67%, while for the isopore polycarbonate membrane it was just 14%, however, the pore volume per membrane area values were more separate apart, with values of 112, 86, 65, and 6 mL/m² for PV650, Millipore PES 0.45 μm pore size, Millipore PVDF 0.45 μm pore size, and Millipore polycarbonate membranes, respectively (Figure 8(A)). These results indicate that thicker asymmetric membranes (175, 150, and 110 μm in the same order as presented above) offer higher pore volume per membrane area, suggesting that the more mass of functionalized adsorptive/IX polymer can be achieved by using thicker MF membranes. Details on the measurement procedure are described in the SI. Also, the affinity between the solution to polymerize and the membrane (e.g., hydrophilicity) may dictate how easily the pores can be infiltrated by the solution. Figure S23 shows how water did not occupy or remain inside all of the pore volume of the PV650, while it did for the other membranes tested.

Then, the theoretical membrane adsorption/IX capacity (mequiv/m²) as a function of the pore volume, the monomer (or polymer) concentration in solution, and the scientific weight capacity of the most-likely repeat unit was calculated. The results for a polymer with 3 mequiv/g scientific weight capacity are shown in Figure 8(B), and theoretical membrane capacities of up to 150 mequiv/m² were observed. Comparison against the experimental values obtained from PMAA-PV650 membranes (41 mequiv/m²) exposed a gap between theoretical and experimental values, which can be attributed to lower degree of pore volume occupancy or lower monomer concentration polymerizing during the functionalization process. Theoretical membrane capacities as high as 350 and 550 mequiv/m² were obtained for the polymers with 7 and 11 mequiv/g (Figure S24), showcasing the maximum values possible this platform can achieve. Details and assumptions on the calculation for the membrane adsorption/IX capacities are presented in the SI.

Finally, to bring the different units from membranes (mequiv/m²) and IX resins (equiv/L) into a normalized unit, the bed volumes were calculated (as described in the SI). The dependency of the concentration of ions and the identification of the environmental applications shown in this study (PFAS, arsenic, and water hardness) are presented in Figure 8(C). Even though the calculated number of bed volumes for membranes follows those from commercially available ion exchange resins, at membrane IX capacities of 200 mequiv/m², which are within the modeled values obtained, can offer the same number of bed volumes as the commercially available IX resins with low technical volume capacity (eq/L). This highlights that the gap between membrane IX capacities and commercially available ion exchangers can be bridged. Furthermore, experimental values of a favorable membrane performance (weak acid PMAA vs IRC83) and the dimensions/properties of a commercially available membrane module were used to predict the number of stages the process would take to remove the same amount of calcium from water (details in the SI). As a conservative calculation, for the ion exchanger the entire module volume was filled up with resin. Figure S25 shows that the IX membrane would require slightly under 2 times the equivalent number of IX packed columns, highlighting how the actual operating adsorptive/IX capacities may reduce the gap previously mentioned.

3.8.2. Challenges and Opportunities.

Based on the experimental and theoretical results, functionalized membranes possess multiple benefits compared to the adsorptive/IX polymers, including faster kinetics, dual adsorption/IX- exclusion mechanisms, and higher useful capacities per mass of adsorptive/IX polymers. Due to the amount of mass able to polymerize in the membrane pores, the major challenge becomes the maximum capacity functionalized membranes can offer. The calculations made in this study, however, prove that this gap can be theoretically closed by utilizing thicker MF membranes with higher cross-sectional porosity and robustly controlling the polymer coating densities in all the MF membrane pores. The use of different chemistries, such as grafting from, may support better control, with the trade-off of making this membrane modification approach less simple and versatile. There is also room for investigating if the adsorptive/ IX properties of the polymers in the membrane pores change or remain constant while just taking advantage of the convective flow. An important variable to compare in future works is the pressure drop and the processing volume based on the water permeance of the functionalized membranes, how the degree of polymerization in the pores affects it, and how it compares to packed bed columns. Fouling onto the membrane or polymeric resins were not evaluated, but it is a common challenge in membrane technologies^55,56 that should be considered in future work. Finally, there are some challenges associated with scale-up processes, such as high frequency adsorption/desorption cycles along with high dynamic flow (pumps needing to turn on and off more frequently). Finding the right application can either minimize the significance of these challenges or help to outweigh these by the benefits functionalized membranes offer, to a point where these technologies can reach industrial scale production.

4. CONCLUSIONS

Membranes that are pore functionalized with adsorptive/ion exchange polymers offered faster transport kinetics (up to 2 orders of magnitude) and higher useful capacity (up to eight times) in direct comparison to commercially available and laboratory-synthesized polymer beads and crushed-down beads. These results suggest that pore functionalized membranes offer a path to overcome the diffusion limitations that current adsorption technologies possess. The regeneration capabilities of the adsorptive/IX membranes synthesized in this work were found to be between 70 and 100% while offering concentration factors as high as 12 times the initial concentration. Remarkably, these separation performances were achieved at operating conditions with residence times (by varying transmembrane pressure) in the order of 2–7 s. Also, postfunctionalized membranes can still operate in low pressure applications, as their values for water permeance ranged between 100 and 1000 (L/(m² h bar)). Knowing the pore volume of commercial microfiltration membranes, which was found to be between 6 and 112 mL/m², the scientific weight capacity of the polymer (3.1–11.5 mequiv/g), and the synthesis conditions (e.g., monomer concentration) allowed for predictions of theoretical adsorption/IX capacities of ions per unit of membrane area, with values as high as 150–550 mequiv/m² depending on the polymer weight capacity. However, experimental capacities were around an order of magnitude lower. This remains the main challenge for functionalized membranes, and we suggest that future research focus on strategies for increasing these adsorption/IX capacities, such as functionalizing thicker MF membranes and robustly controlling the pore functionalization throughout all the membrane pores. Additionally, a simultaneous separation mechanism (exclusion and adsorption/IX) is an intrinsic benefit from these functionalized membrane plat-forms, which are normally designed for size exclusion separations. All these findings were supported by a simple membrane modification approach, with versatile integration of different desired functional groups, and with application toward a broad variety of environmentally relevant ions (e.g., PFAS, arsenic, and calcium). These advantages make these advanced separation technologies promising for scale-up and can offer new avenues to make safe drinking water more accessible.