View PDF Version
 
DOI: 10.1039/D4NJ02840F (Paper) New J. Chem., 2024, Advance Article

Novel robust hierarchical porous membrane for uranium enrichment: fabrication, degradation behavior, and uranium sorption performance†

Zhixiao Liu, Lingjun Meng, Haoye Xiong, Lintao Liao, Yuhang Zhao, Yiping Zhong, Tongtong Xie, Yuhang Yan, Gao Hu and Zhiming Mi*
State Key Laboratory of Nuclear Resources and Environment, East China University of Technology, Nanchang, 330013, P. R. China. E-mail: mzm@ecut.edu.cn

Received 21st June 2024 , Accepted 6th August 2024

First published on 7th August 2024

Abstract

Extracting uranium from water bodies is urgently needed whether from a waste nuclear resource utilization or a new energy exploration perspective. Herein, a novel robust hierarchical porous CAP membrane was fabricated from the nucleophilic condensation of phenolphthalein and 2,6-difluorobenzonitrile, followed by classical amidoximation and nonsolvent-induced phase separation (NIPS) methods. Fast uranium uptake kinetics with sorption equilibrium at a mere 12 h was demonstrated, and the uranium uptake capacity was found to be 499 mg g−1. The significant improvement in uranium enrichment performance was probably endowed by the outstanding hydrophilic surface and the hierarchical pores throughout the cross-sectional CAP membrane. Meanwhile, it was confirmed that the CAP molecular backbone may endure polymer degradation during amidoximation, and the amount of hydroxylamine used should be strictly controlled. The sorption mechanism was explored, and practical testing experimentally indicated the excellent desorption, reusability and selectivity properties of the CAP membrane. Overall, the easy processability, robustness and outstanding uranium sorption performance made the CAP membrane an ideal candidate for uranium enrichment from water bodies.

1. Introduction

As a clean energy source, nuclear power is in high demand and is expected to double by 2040.1,2 Uranium is the main raw material for nuclear energy. Nevertheless, the development and utilization of uranium resources inevitably produce massive amounts of radioactive wastewater.3,4 Wherein hexavalent uranium, with good water solubility and strong mobility, can migrate into the ecosystem along with water circulation, posing a serious threat to human health and the living environment.5–7 Therefore, the enrichment of uranium from water bodies is extremely urgent. Popular materials under development for extracting uranium possess a porous structure and large surface area, e.g. MOFs,8,9 COFs,10,11 and PAFs.12,13 However, most of these adsorbents are in powder form, making them unsuitable for large-scale practical use. Therefore, an increasing number of attempts have been made to develop self-standing, easily prepared and economical polymer adsorbents such as membranes14–17 and fibers.18,19

Polyacrylonitrile (PAN)-derived polyacrylonitrile oxime (PAO), has been the most sought-after adsorbent since its inception due to its selectivity towards uranium.20 Nonetheless, there have been some drawbacks which have considerably obstructed its uranium extraction performance. For instance, PAO based adsorbents can hardly be self-supported whether processed into porous membrane or fine fibers. Hence in most cases, a high-strength and flexible matrix has to be chosen as reinforcement to ensure the practicality of PAO. The introduction of matrix supports such as polypropylene,21 polyethylene,22 chitosan,23 and polyvinyl alcohol24 inevitably sacrifices the adsorption capacity. On the other hand, the insufficient hydrophilicity of the PAO surface and the single oxime functional group restricts the diffusion and adsorption of uranyl at the phase interface.25 Consequently, it was not hard to deduce that the integration of amidoxime and some hydrophilic functional groups into the backbone of the polymer chains was entirely justified. Meanwhile, reasonable processing of the adsorbent, e.g. the introduction of a hierarchical porous structure inside the membrane surface, can achieve significant exposure of functional sites, which is beneficial for boosting uranium adsorption properties.26,27

In contrast to PAO with aliphatic molecular chains, aromatic polymer membranes showed better mechanical properties ascribed to their intra- and intermolecular π–π electronic stacking.28,29 For instance, a membrane of amidoximated polymers of intrinsic microporosity (AO-PIM-1) developed by Yang et al.27 showed flexible characteristics with hierarchical pores throughout the cross-sectional membrane, and the uranium uptake was found to be far higher than that of a membrane with only intrinsic microporosity. Regrettably, the adsorption equilibrium took over 50 h due to a lack of sufficient hydrophilicity of the AO-PIM-1 surface. Recently, a flexible aromatic polyamide acid membrane (GPAA) was gently fabricated in our lab.30 Due to the presence of hydrophilic carboxyl, amino and secondary amino on the porous membrane surface, the GPAA membrane demonstrated excellent uranium uptake kinetics with uranium adsorption equilibration at 16 h and capacity of 413 mg g−1.

In this work, high molecular weight CNP was first synthesized via nucleophilic polycondensation between 2,6-difluorobenzonitrile and phenolphthalein, and CAP was subsequently obtained by amidoximation in hydroxylamine solution. Thereafter, a flexible hierarchical porous CAP membrane could be readily fabricated via classical NIPS (Fig. 1) according to our previously reported method.31,32 The selection of aromatic monomers was made in view of the ideal mechanical properties resulting from the strong π–π electronic interaction of high molecular chains, while the contorted structure (Fig. S1, ESI†) of phenolphthalein was expected to create an irregular molecular arrangement and thus improve the processability of CAP in common organic solvents. Additionally, the introduction of carboxyl into the backbone of CAP was utilized to increase the hydrophilicity of the adsorbent surface, as well as synergistically capturing uranyl from aqueous solutions. As a result, the robust hierarchical porous CAP membrane exhibited excellent solubility and ideal hydrophilicity with an electronegative surface at pH > 3.45. Meanwhile, the degradation behavior of CAP in the amidoximation process was studied, and the uranium sorption mechanism is also presented. Furthermore, practical application demonstrated that the obtained CAP membrane showed outstanding desorption, reusability and selectivity performance. To our knowledge, the nucleophilic polycondensation of 2,6-difluorobenzonitrile and phenolphthalein, and the study of a CAP membrane for uranium extraction from water bodies, have not yet been reported.


image file: d4nj02840f-f1.tif
Fig. 1 Fabrication of a flexible hierarchical porous CAP membrane. Synthesis scheme of CNP and CAP polymers (I). NIPS process (II). Insert: hierarchical porous structure, and schematic diagram of adsorption site and uranium interaction.

2. Experimental

2.1. Materials

Phenolphthalein (98 + %) and 2,6-difluorobenzonitrile (98 + %) were provided by Adamas and used as received. N,N-Dimethylformamide (DMF, 99.8%) and N,N-dimethylacetamide (DMAc, 99%) were bought from Aldrich Chemical Co. (Shanghai, China), and further dehydrated using reduced pressure distillation and stored in single-ended bottles containing 4 Å molecular sieves. Potassium carbonate (K2CO3, 99.99%) was provided by Aladdin Reagent Co. Ltd and used directedly. Arsenazo III (C22H18As2N4O14S2, 98%), UO2(NO3)2·6H2O (98%) and toluene (99%) were bought from Xilong Chemical Reagent Guangdong Co. Ltd and used as received. Sodium hydroxide (NaOH), hydrochloric acid (HCl, 12 mol L−1), polyvinylpyrrolidone (PVP, 99%) and NH2OH (50 wt% in water) were obtained from Sinopharm Chemical Reagent Beijing Co. Ltd. Deionized water (R = 18.2 mΩ cm) was obtained from a hollow fiber reverse osmosis device and used throughout the experiments. The other commercially available solvents and reagents were provided by Adamas (Shanghai Titan Co. Ltd) and used without further purification.

2.2. Characterization

Fourier transform infrared (FTIR) spectra were investigated with a Thermal Scientific Nicolet 380 spectrometer at a resolution of 4 cm−1 in the range of 500–4000 cm−1. X-ray photoelectron spectra (XPS) were carried out on an ESCALAB 250 (Thermo Fisher Scientific). Differential scanning calorimetric (DSC) analysis was performed on a TA instrument DSC Q2000 at a scanning rate of 10 °C min−1 from 50 °C to 350 °C under nitrogen. Thermal gravimetric analysis (TGA) was recorded on a Netzsch STA2500 under nitrogen atmosphere at a heating rate of 10 °C min−1 from 100 °C to 820 °C. The 1H NMR, 13C NMR, and correlation (COSY) spectra were conducted on a Bruker-AVANCEIII500 (500 MHz) with deuterated dimethyl sulfoxide (DMSO-d6) as the solvent and tetramethyl silane (TMS) as the reference. The pH values were determined with a pH meter (PHS-25, Shanghai Yueping, China), and the pH meter was calibrated using pH = 4.00, pH = 6.86, and pH = 9.18 standard buffer solutions prior to use. Studies of the morphologies of the porous membrane were conducted on a field emission electron microscope (FESEM, NOVA NANOSEM 450). The uranium concentrations were measured on an ultraviolet visible (UV-Vis) spectrophotometer (UV1800PC, Shanghai, China) with arsenazo III as the chromogenic agent. Nitrogen absorption and desorption measurements were carried out using a Quantachrome Autosorb IQ at 77 K. The surface charge of the resultant membrane was obtained by measuring the streaming potential with a SurPASS electrokinetic analyzer (Anton Paar GmbH, Austria). The hydrophilicity was evaluated by water contact angle (CA) on a drop shape analysis system DSA10-MK2 at 25 °C. Weight-average molecular weights (Mw) and number-average molecular weights (Mn) were achieved from gel permeation chromatography (GPC) on the basis of polystyrene calibration on a PL-GPC 220 instrument equipped with a refractive index detector using DMF as an eluent at a flow rate of 1.0 mL min−1. Mechanical properties were studied on a Shimadzu AG-I universal testing apparatus with a crosshead speed of 5 mm min−1, and the tensile strength (TS) and elongation at break (EB) were calculated as the average of five strips.

2.3. Synthesis of carboxylated aromatic cyano polymer (CNP)

Phenolphthalein (3.2034 g, 10 mmol), 2,6-difluorobenzonitrile (1.3910 g, 10 mmol), K2CO3 (3.4551 g, 25 mmol), and anhydrous DMF (30 mL) were transferred to a 100-mL three-necked flask equipped with a mechanical stirring paddle, a nitrogen inlet with a temperature detector, and a 50-mL water separator with a reflux condenser. Thereafter, an extra 10 mL of toluene as the water carrier was added. The mixture was first maintained in a 20 mL min−1 nitrogen atmosphere for 15 min, and then heated to 130 °C for approximately 4 h until the water level in the water separator no longer increased. Then, the reaction was continuously refluxed at 160 °C to obtain a highly viscous gray solution. Afterwards, the mixture was directly poured into dilute acid solution to produce flexible polymer filaments, the crude polymer was washed throughout with ultrapure water and dried overnight in a vacuum at 120 °C to obtain white CNP. Yield: 99%, Tg: 240 °C, 1H NMR (500 MHz, DMSO-d6) δ 13.04 (–COOH), 7.84–7.82 (H1), 7.55–7.50 (H2), 7.38–7.35 (H3), 7.19–7.06 (H4), 6.71 (H5), 6.62–6.60 (H6). 13C NMR (126 MHz, DMSO) δ 169.35 (C1), 160.91 (C2), 153.29 (C3), 144.02 (C44), 140.98 (C5), 136.32 (C6), 131.98 (C7), 131.86 (C8), 131.60 (C9), 130.79 (C10), 127.09 (C11), 120.35 (C12), 113.51 (C13), 111.21 (C14), 100.00 (C15), 94.66 (C16), 50.53 (C17).

2.4. Synthesis of carboxylated aromatic amidoxime polymer (CAP)

CNP (2.0972 g, 5 mmol) was first dissolved in anhydrous DMAc (40 mL) in a three-neck bottle equipped with a magnetic stirrer, a temperature detector, and a condenser. Subsequently, NH2OH (12.5 mmol) aqueous solution was dropped in stepwise, and the mixture was continuously reacted at 45 °C for nearly 48 h until no apparent cyan groups could be observed in the FTIR spectrum. Finally, a goose yellow floccule could be achieved by pouring the reaction mixture into ultrapure water, followed by filtration and oven drying overnight at 80 °C. This was denoted as CAP. Yield: 87%, 1H NMR (500 MHz, DMSO-d6) δ 12.88 (–COOH), 9.32 (–OH), 7.81–7.79 (H1), 7.48–7.49 (H2), 7.32–7.35 (H3), 7.01–7.04 (H4), 6.62 (H5), 6.56–6.59 (H6).

2.5. Fabrication of porous CAP membrane

In a 10-mL sample tube, CAP (0.5000 g) and PVP (0.1000 g) were dissolved in DMF (2.8333 g) under magnetic stirring to obtain a transparent solution. Then, the fully degassed-polymer solution was poured linearly onto a horizontal dry and clean glass plate, and subsequently a scraping knife was used to scrape the casting solution into a square pattern with dimensions of 100 mm (length) × 100 mm (width) × 250 μm (thickness). After a few seconds, the glass plate was fully immersed into ultrapure water for the NIPS process, and a white flexible membrane (Fig. 5a) could be achieved after the thorough exchange of organic solvent DMF and non-solvent water for 24 h.

For comparison, polyacrylonitrile (Mw = 150[thin space (1/6-em)]000) was also amidoximized in a similar manner to that of CNP to achieve classical polyamidoxime (PAO), and a PAO membrane could be readily obtained by casting the polymer solution onto a glass plate, followed by NIPS treatment.

3. Results and discussion

3.1. Synthesis and characterization of CAP

Based on our reported method,33 CNP with high molecular weight can be readily achieved via nucleophilic polycondensation between phenolphthalein and 2,6-difluorobenzonitrile. Nevertheless, neither 3,5-difluorobenzonitrile nor 2,5-difluorobenzonitrile was able to routinely react with phenolphthalein to produce high molecular weight polymers, attributed to the passivation effect of meta-CN on fluorine in the condensation reaction. Various characterizations, including FTIR, NMR, DSC, and TGA, were utilized to detect the chemical structural and physical properties of CNP and CAP. As displayed in the FTIR spectra (Fig. 2a) of CNP, characteristic absorption peaks around 3500–2500 cm−1 (associated with the hydrogen bonding of –OH) suggested the existence of carboxyl, and peaks near 2972 cm−1, 2233 cm−1 and 1720 cm−1 indicated methylene C–H in phenolphthalein, CN and carbonyl in the backbone of CNP polymers, respectively. After oximation, the absence of the characteristic absorption of cyan, and the appearance of N–H (3484 cm−1, 3360 cm−1), C[double bond, length as m-dash]N (1656 cm−1) and C–N (1376 cm−1) proved that CNP had been fully converted to CAP. Meanwhile, characteristic absorptions of an aromatic ether bond (1225 cm−1 and 1010 cm−1) both appeared in CNP and CAP, providing solid evidence for the successful nucleophilic polycondensation of the monomers. Additionally, NMR spectra of CNP and CAP were carried out. As depicted in Fig. 2b and Fig. S2 (ESI†), each hydrogen or carbon atom of CNP can be unambiguously quantitatively assigned. Interestingly, for CAP, apart from the various well-assigned hydrogens (Fig. 2c), small splitting peaks emerged around 6.8 ppm, 7.15 ppm, and 7.7 ppm (Fig. S3, ESI†). This aroused our interest and will be discussed in the following section. Furthermore, high-resolution C 1s (Fig. 2d) and N 1s (Fig. 2e) XPS of CAP was conducted. The appearance of binding energies around 291.49 eV (C[double bond, length as m-dash]O), 285.26 eV (C[double bond, length as m-dash]N) in C 1s, and 400.36 eV (N–H), 399.82 eV (C[double bond, length as m-dash]N) in N 1s indicated the presence of carboxyl and amidoxime groups.
image file: d4nj02840f-f2.tif
Fig. 2 Characterizations of CNP and CAP. FTIR spectra (a). 1H NMR spectra of CNP (b) and CAP (c). High-resolution C 1s (d) and N 1s (e) XPS spectra of CAP. Thermal properties evaluated by DSC (f).

The thermal stability of adsorbents is an important indicator, especially when facing high-temperature nuclear wastewater. Fig. 2f shows that CNP exhibited a Tg as high as 240 °C, demonstrating outstanding thermal resistance. This was mainly ascribed to the tough aromatic molecular chains. Whereas for CAP, the high molecular chains began to degrade at 215 °C, as can be confirmed by the TGA curve (Fig. S4, ESI†). This resulted from the thermal instability of amidoxime, as reported in our previous work.25 Interestingly, CAP exhibited nearly 10% thermal weight loss before thermal decomposition, probably attributable to the removal of surface moisture adsorbed by the hydrophilic polymer. The fine thermal stability of the obtained polymers demonstrated the serviceability of the porous CAP membrane in high-temperature circumstances.

Furthermore, the solubility of CAP was examined in common organic solvents, and the results are listed in Table S1 (ESI†). It was discovered that the resultant CAP not only exhibited excellent solubility performance in aprotic solvents, including DMF, DMAc and DMSO, but in low boiling point solvents, such as CHCl3 and THF. This was mainly attributed to the twisted and flexible triphenylmethane moieties in phenolphthalein fragments, which markedly reduced the regularity of the polymer chains. It is worth noting that the ease of processability using common organic solvents, especially low-boiling solvents, made CAP suitable for solution processing techniques.34

3.2. Degradation and stability behavior of CAP in oximation process

Theoretically, polymer-based adsorbents revealed good mechanical properties, e.g. flexible self-supporting fibers, thin films and porous membranes, if these amidoxime adsorbents were transformed from high molecular weight cyano-containing polymers. Unexpectedly, Mw decreased from 18.4 w for CNP to 15.9 w for CAP in the course of amidoximation. The solution temperature, the amount of NH2OH, and the reaction time have been recognized as three key factors during the amidoximation process. In most cases, the temperature was kept at no more than 60 °C to prevent severe shrinkage or damage to the oximation products, and the reaction time was usually more than 48 h to ensure the complete conversion of cyan into amidoxime.35 Therefore, the dosage of hydroxylamine was considered the key factor affecting the molecular structure. In this work, 2.5, 5, 7.5, and 10.0 molar equivalents of hydroxylamine were employed (Table S2, ESI†). All the reactions were kept at 45 °C for 48 h to achieve fully amidoximized CAP. Astonishingly, the resultant CAPs exhibited sharply decreased molecular weights (Mw) from 18.4 w for CNP to 15.9 w, 11.8 w, 9.5 w, 7.0 w for CAP with the use of NH2OH at 2.5 eq., 5.0 eq., 7.5 eq., and 10.0 eq., respectively (Fig. S5 and Table S3, ESI†).

To explore the subtle changes in molecular structure, both 1H NMR and 2D 1H–1H COSY NMR spectra were carried out to identify and track proton variations. As displayed in Fig. 3a, cross-peaks including H1/H2, H2/H4 and H3/H6 found in the COSY spectrum at 7.80/7.48, 7.48/7.07, and 7.35/6.57 ppm illustrated the existence of H1, H2, H3, H4 and H6, respectively (Fig. 3a). This illustrated the unambiguous assignment of protons in CAP. Moreover, as described in Fig. 3b, the hydrogen resonance around 6.6–6.7 ppm of methylene in the phenolphthalein moiety showed significant attenuation with an increase in the dosage of hydroxylamine from 2.5 eq. to 10 eq. This suggested the degradation of the molecular backbone of CAP, causing breakage of the molecular chain and hence reduced molecular weights. Further, the aromatic fragment connected to amidoxime at ∼7.05 ppm may also be dramatically attacked when the dosage of hydroxylamine is as high as 10 eq. Therefore, the dosage of hydroxylamine was strictly controlled within 2.5 eq. in this work.


image file: d4nj02840f-f3.tif
Fig. 3 1H–1H COSY (a), and 1H NMR (b) spectra of CAP with hydroxylamine concentrations of 2.5 eq., 5.0 eq., 7.5 eq., and 10.0 eq.

Similarly, it was not hard to explain that, for the most popular PAN-derived PAO, the non-self-standing or fragile PAO membrane was caused mainly by the inevitable methylene fragment destruction during the amidoximation process. Up to now, the degradation mechanism of methylene attacked by hydroxylamine has still been unclear and further research is ongoing in our lab. But at least it is clear that the amount of hydroxylamine should be strictly limited during the oximation reaction process to prevent severe molecular structural destruction.

3.3. Characterization of CAP membrane

A tough and flexible porous CAP membrane can be readily achieved via a classical NIPS process.30 As displayed in the morphology diagrams in Fig. 4a and b, both the upper and sub-surfaces presented large pores with average pore diameters of 0.25 μm and 2 μm, respectively. In fact, these large pores on the membranes were undoubtedly recognized as surface defects either for a nanofiltration or an ultrafiltration membrane. Whereas for the sorption situation, the large pore characteristics on the membrane surface facilitated the rapid entry of uranyl ions into the membrane body. In addition, the cross-section of the CAP membrane (Fig. 4c and d) demonstrated that there were finger-shaped pore channels with large pores running through the whole membrane, and smaller pores were distributed on the channel walls. Moreover, the pore size gradually increased from the top to the bottom surface of the membrane, consistent with the trend in pore size changes of the upper and sub-surfaces of the membrane. This unique hierarchical pore structure can significantly reduce the transport resistance of uranyl ions within the adsorbent, improve diffusion efficiency, and facilitate an improvement in uranium adsorption kinetics.27
image file: d4nj02840f-f4.tif
Fig. 4 Morphologies and surface properties of the CAP membrane. The upper surface (a), sub-surface (b), cross-sectional overview (c) and partial enlarged image (d) of hierarchical porous structure. Dynamic water contact angle results (e).

In addition, the mesoporous properties of the membrane were also examined via BET analysis (Fig. 4e), and the surface area of the porous CAP membrane was calculated to be 8.224 m2 g−1. A relatively large specific surface area was conducive to the exposure of active sites for uranium adsorption. Furthermore, the isothermal curve exhibited a typical IV isotherm with a sorption hysteresis loop, indicating the existence of mesopores. As demonstrated in the pore diameter distribution (inserted picture in Fig. S6, ESI†), mesopores of nearly ∼16 nm and ∼30 nm can be clearly observed. The hierarchical porous structure was beneficial for reducing the mass transfer resistance in the sorption process, thus significantly improving the uranium sorption efficiency.

The hydrophilicity of the membrane surface affects the diffusion rate of uranyl ions at the phase interface. The water contact angle results (Fig. 4e) of the membrane surface indicated that water droplets can quickly spread out on the membrane surface within 5.2 seconds. This indicated the superhydrophilicity of the membrane, mainly ascribed to two aspects. Firstly, the abundant polar hydrophilic carboxyl groups in the backbone of the CAP molecular chains enhanced the surface hydrophilicity. Secondly, the rough and porous surface promoted the wettability of the CAP membrane.32 These above factors were conducive to the rapid spreading of aqueous solutions on the membrane surface.

The mechanical properties of the CAP membrane were evaluated with a universal tensile testing machine. The tensile strength was found to be 6.2 MPa with elongation at break of 11.5% (Fig. S7 and Table S4, ESI†), similar to the polysulfone ultrafiltration membrane in our previous report.33 In comparison with the mechanical performance of a PAO-based membrane, the tensile strength of the resultant CAP membrane was far higher than that of PAO, ascribed to the π–π interaction of the aromatic molecular chains of CAP. Whereas for PAO, there were no π electric interactions due to its aliphatic chain attributes, as in the case for thermal performance.

3.4. Uranium adsorption performance of CAP membrane

3.4.1 Adsorption performance in uranium aqueous solution. The uranium adsorption behavior of the hierarchical porous CAP membrane was initially explored in 20 ppm uranium-spiked aqueous solution at pH = 6. The uranium concentrations before and after adsorption were measured via UV-vis with arsenazo III as a chromogenic agent, and the detailed measurement process, e.g. the drawing of the standard curve (Fig. S8, ESI†), is covered in detail in the ESI.† As shown, the appearance of the membrane changed from white to yellow after a mere 12 h of uranium sorption (Fig. 5a), illustrating the successful uranyl enrichment on the CAP membrane surface. Additionally, the whole XPS before and after uranium uptake (Fig. 5b) demonstrated distinctive U 4f peaks near 393.14 eV (U 4f5/2) and 382.33 eV (U 4f7/2) with a spin–orbit splitting of 10.8 eV (Fig. S9, ESI†), suggesting that the chemical valence of uranium did not change. Furthermore, the SEM-EDS images in Fig. 5d depict that a considerable amount of uranium had already been enriched inside the membrane surface, and this was consistent with the color change and XPS results of the hierarchical porous CAP membrane.
image file: d4nj02840f-f5.tif
Fig. 5 Uranium adsorption performance of the CAP membrane in 20 ppm uranium-spiked aqueous solution. Physical images before and after uranium sorption, and CAP-U membrane bending (a). XPS of the CAP membrane before and after uranium adsorption (b). pH influence and surface charge variations at different pH values (c). SEM-EDS results of local cross-sectional SEM image (d). High-resolution N 1s XPS of CAP-U (e). High-resolution O 1s XPS spectra before (f) and after (g) uranium sorption.

pH is an essential parameter that affects adsorption performance. As indicated in Fig. 5c, no significant uranium adsorption amounts were observed at pH < 4. This was mainly due to the protonation of amidoxime and carboxyl under acid conditions. Meanwhile, the zeta-potential results demonstrated that the CAP membrane surface had an isoelectric point of pH = 3.45 (Fig. 5c). Therefore, it can be speculated that the positively charged membrane surface (pH < 4) increased the electrostatic repulsion between uranyl and the functional sites. However, the uranium uptakes sharply increased from 167 mg g−1 for pH = 4 to 270 mg g−1 for pH = 5, and 454 mg g−1 for pH = 6. This was because the deprotonation of the chelating active sites and the electronegativity of the CAP membrane surface towards the positive uranyl, boosted the enrichment of uranium on the membrane surface. To increase pH (pH > 8) one step further, the uranium uptake began to decrease, probably due to the hydrolysis or even precipitation of uranyl in alkaline aqueous solution. Notably, the resultant CAP membrane exhibited high amounts of uranium adsorption over a wide pH range, implying the practicality of the CAP membrane both in wastewater and seawater uranium enrichment scenarios.

To gain a deeper insight into the uranium capture mechanism, high-resolution XPS spectra, including N 1s and O 1s before and after uranium uptake, were carried out. It was identified that the binding energies of amidoxime N–H shifted from 400.36 eV (Fig. 2e) for CAP to 400.59 eV (Fig. 5e) for CAP-U, whereas no apparent binding energy differences were observed for C[double bond, length as m-dash]N before (399.82 eV, Fig. 2e) and after (399.81 eV, Fig. 5e) uranium uptake. This suggested the complexation of amidoxime N–H with uranyl. Moreover, high-resolution O 1s XPS (Fig. 5g) demonstrated that the binding energies of COOH and C[double bond, length as m-dash]N–OH were located at 533.84 eV and 533.10 eV, respectively, after uranium uptake, which were higher than for the pristine CAP membrane with binding energies at 533.71 eV for COOH and 532.97 eV for C[double bond, length as m-dash]N–OH (Fig. 5f). This illustrated that the oxygens in carboxyl and oxime were responsible for capturing uranium during the sorption process. Therefore, it can be understood that the nitrogen in amidoxime, as well as the oxygens in carboxyl and oxime played dominant roles in collecting uranium from aqueous solution.

3.4.2 Uranium adsorption kinetics and isothermal adsorption characteristics. The adsorption kinetics were investigated by soaking CAP membranes (10 mg) in 250 mL of uranium-spiked aqueous solutions with concentrations of 10 ppm and 20 ppm, respectively. Testing samples were separately taken out at intervals of 0, 0.5, 1, 2, 3, 4, 6, 8, 10 and 12 h, and all concentrations were determined via UV-vis at 652 nm. It was found that the CAP membrane exhibited fast sorption kinetics within 6 h, and reached adsorption equilibrium in a mere 12 h (Fig. 6a). Whereas for the PAO-derived membrane, no adsorption equilibrium could be observed, even when the sorption time was extended to 48 h.36 This was mainly attributed to the carboxyl in the backbone of the molecular chains not only increasing the surface hydrophilicity of the CAP membrane, but synergistically capturing uranium via the formation of stable complexes between uranyl and the empty orbitals in oxygen.37
image file: d4nj02840f-f6.tif
Fig. 6 Uranium enrichment performance of the CAP membrane. Uranium sorption kinetics in 10 ppm and 20 ppm uranium-spiked aqueous solutions for the CAP membrane, and 20 ppm for the PAO membrane (a). Linear regressions of pseudo-first-order (b) and pseudo-second-order (c) kinetics. Langmuir and Freundlich fitting curves of the sorption isotherm (d). Weber and Morris fitting curves of sorption kinetics in 10 ppm uranium-spiked aqueous solution (e). Desorption (f) and reusability performance (g). Cross-sectional image (h) in the fifth sorption–desorption cycle (h). Selectivity in the face of common interfering ions (i).

To explore the kinetics mechanism, pseudo-first-order (eqn (S1), ESI†) as well as the pseudo-second-order (eqn (S2), ESI†) kinetics were used to further evaluate the sorption behavior. The resultant fitting curves and detailed data are shown in Fig. 6b and c and Table S4 (ESI†), respectively. It was identified that the R2 of pseudo-second-order were 0.9972 and 0.9956 for 10 ppm and 20 ppm uranium-spiked aqueous solutions, respectively, which were only slightly higher than the pseudo-first-order with R2 at 0.9956 and 0.9968 for 10 ppm and 20 ppm uranium aqueous solutions, respectively. This suggested that both chemical and physical sorption occurred in the process of uranium capture. Excellent physical adsorption behavior is uncommon in most adsorbents, and it might be ascribed to the strong electrostatic interaction between electronegative membrane surfaces and positively charged uranyl.

The distribution coefficient (Kd, mg L−1) is an important indicator of an adsorbent, and a Kd value higher than 104 mL g−1 is considered to show good affinity for uranium.38 Herein, the Kd values were calculated according to eqn (S3) (ESI†) and the results are listed in Table S5 (ESI†). It was discovered that the CAP membrane exhibited high Kd values of 3.84 × 107 mL g−1 and 1.63 × 105 mL g−1 in 10 ppm and 20 ppm uranium-spiked aqueous solutions, respectively, both far higher than 104 mL g−1. This suggested that the obtained hierarchical porous CAP membrane possessed excellent affinity towards uranyl, which could effectively facilitate the rapid enrichment of uranium from aqueous solution.

To estimate the uranium adsorption capacity, isothermal adsorption experiments were carried out at 298.15 K. Specifically, CAP membranes (10 mg) were separately added into 250 mL of aqueous solutions with uranium concentrations ranging from 5 to 45 ppm at intervals of 5 ppm. Both Langmuir (eqn (S4), ESI†) and Freundlich (eqn (S5), ESI†) models were employed to fit the isothermal curve, and the results are presented in Fig. 6d and Table S6 (ESI†). The Langmuir model was a better fit with a higher correlation coefficient (R2 = 0.9940) than that of the Freundlich model (R2 = 0.8527), showing that single molecule layer adsorption dominated the entire adsorption process. Furthermore, the uranium capacity was calculated to be 499 mg g−1, implying that the obtained hierarchical porous membrane was an ideal candidate among uranium adsorbents.

To gain a deeper insight into uranyl diffusion throughout the hierarchical porous CAP membrane, a Weber and Morris intraparticle diffusion model (eqn (S6), ESI†) was employed. As displayed in Fig. 6e, the linear regression curves of qt–t0.5 were divided into three main regions with R2 at 0.9897, 0.9827 and 0.9350, corresponding to 0–120 min, 180–360 min and 360–720 min, respectively. This illustrated that the uranyl transfer process was controlled mainly by external process transfer, surface-to-interior ion migration and intrapore diffusion courses. It was not hard to deduce that gradual diffusion was beneficial for maximizing the utilization of functional groups on the membrane surface, which significantly accelerated the uranium enrichment efficiency.27

3.4.3 Uranium desorption, reusability, and selectivity of CAP membrane. Uranium desorption measurement was conducted by directly immersing the uranium-loaded CAP membrane into an eluent containing 0.1 M HCl aqueous solution. As displayed in Fig. 6f, nearly 90% of uranium was washed out within 8 min, and a desorption rate of 99% could be reached in 20 min. This suggested the favorable regenerative ability of the uranium-sorbed CAP membrane. Notably, in contrast to traditional PAO adsorbents, no complex eluent, such as sodium carbonate or hydrogen peroxide aqueous solution, was needed for uranium elution from the CAP membrane. This was because of the presence of p–π electron conjugation in aromatic amidoxime, facilitating a lower electrostatic potential on the aromatic amidoxime in comparation with that in PAO. This endowed the selective amidoxime in the CAP membrane with excellent environmental tolerance.

In practice, reusability is also a key parameter. For the reusability measurement, a piece of CAP membrane (10 mg) was immersed in a 250-mL, 20 ppm uranium aqueous solution for 12 h, and subsequently eluted with 0.1 M HCl. For each sorption–desorption cycle, no extra treatment was conducted except that the adsorbent membrane was thoroughly flushed with ultrapure water. As demonstrated in Fig. 6g, no more than 5% adsorption capacity loss was observed after five sorption–desorption cycles. Meanwhile, the uranium elution rate still retained a high level of 95% in the fifth cycle. Moreover, the FTIR spectra of the hierarchical porous CAP membrane in the 1st, 3rd, and 5th sorption–desorption cycles were carried out (Fig. S10, ESI†), and no significant degradation was observed in functional groups such as carboxyl or amidoxime. Additionally, the cross-sectional morphology of the membrane in the 5th cycle presented in Fig. 6h indicated that the overall appearance of the membrane was well preserved except for some small local cracks inside the membrane's finger-like holes. These results illustrated the outstanding reusability of the resultant CAP membrane.

For uranium enrichment whether from wastewater or seawater, common interfering ions such as Ca2+, K+, Na+, Mg2+, or Fe3+ should inevitably be taken into account. Therefore, the most popular interfering cations in their nitrates were added to uranium-spiked aqueous solution, and the concentration of each interfering ion (2000 ppm) was 100 times that of the uranium concentration. As illustrated in Fig. 6i, uranium uptake still maintained a high level of 432 mg g−1 in the presence of K+, Na+, or Mg2+, demonstrating high selectivity towards uranium. Whereas for Ca2+ and Fe3+, the uranium adsorption amount decreased to 91.8% and 87.9% in the fifth cycle, respectively. This can be attributed to the empty orbitals of oxygen or nitrogen being occupied by electrophilic cations, leading to the formation of stable complexes between iron (calcium) ions and functional amidoxime (carboxyl) groups. Overall, the flexible hierarchical porous CAP membrane performed with satisfactory selectivity for uranium enrichment.

4. Conclusions

In this study, a novel flexible porous membrane adsorbent was designed and polycondensed from phenolphthalein and 2,6-difluorobenzonitrile, followed by amidoximation and NIPS processes. The chemical structure and the physical properties were fully characterized via FTIR, XPS, zeta-potential, SEM, AFM, CA, etc. It was demonstrated that the methylene of the phenolphthalein moiety may suffer serious structural degradation when the concentration of hydroxylamine is higher than 2.5 eq. The resultant CAP was rough and flexible with a large surface area, and had hierarchical pores throughout the hydrophilic membrane body. The excellent hydrophilicity and hierarchical porous properties facilitated diffusion kinetics and greatly reduced mass transfer resistance. Therefore, the equilibrium sorption of uranyl could be reached within a mere 12 h in 20 ppm uranium-spiked aqueous solution. In contrast to popular PAO adsorbents, both physical and chemical sorption occurred in the uranium enrichment process, and the maximum uranium uptake was found to be 499 mg g−1. This was mainly ascribed to the low isoelectric point of the membrane surface, resulting in strong electrostatic interaction between the electronegative membrane body and positively charged uranyl. Meanwhile, the carboxyl in the backbone of the molecular chains not only increased the surface hydrophilicity of the CAP membrane, but synergistically captured uranium via the formation of stable complexes between uranyl and the empty orbitals in oxygen. Furthermore, practical testing results indicated that the CAP membrane maintained a high uranium desorption rate with no more than 5% uranium uptake loss after the fifth sorption–desorption cycle. In addition, the obtained CAP membrane demonstrated fine selectivity in the face of common interfering ions, endowed by the natural selectivity of oxime groups towards uranyl. Overall, the flexible hierarchical porous CAP membrane in this work was a promising candidate for uranium enrichment from water bodies.

Author contributions

Zhixiao Liu: writing, original draft, preparation, investigation. Lingjun Meng: data curation, validation, resources. Haoye Xiong: software. Lintao Liao: software, validation. Yuhang Zhao: supervision, methodology. Yiping Zhong: software, methodology. Tongtong Xie: data curation, software. Yuhang Yan: data curation. Gao Hu: data curation. Zhiming Mi: writing – reviewing and editing, visualization, investigation.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The work was financially supported by the Opening Project of Jiangxi Province Key Laboratory of Polymer Micro/Nano Manufacturing and Devices (PMND201904, PMND202109), the Doctoral Scientific Research Starting Foundation of East China University of Technology (DHBK2019115, DHBK2019116), the Foundation of Jiangxi Educational Commission (GJJ210701, GJJ210717), and Jiangxi Province Natural Sciences Fund Financing Project (20232BAB213037).

References

  1. OECD, World energy outlook 2023,  DOI:10.1787/827374a6-en.
  2. R. Leng, Y. Sun, C. Wang, Z. Qu, R. Feng, G. Zhao, B. Han, J. Wang, Z. Ji and X. Wang, Environ. Sci. Technol., 2023, 57, 9615–9626 CrossRef PubMed .
  3. C. Meng, M. Du, Z. Zhang, Q. Liu, C. Yan, Z. Li, Z. Dong, J. Luo, J. Ma, Y. Liu and X. Wang, Environ. Sci. Technol., 2024, 58, 9456–9465 CrossRef PubMed .
  4. X. Zhang, X. Yang, Q. Rong, X. Liu, Y. Zhou, H. Yang, G. Wang, Z. Chen and X. Wang, ACS EST Eng., 2024, 4, 250–268 CrossRef .
  5. Z. Qu, R. Leng, S. Wang, Z. Ji and X. Wang, Rev. Environ. Contam. Toxicol., 2024, 262, 12 CrossRef .
  6. Y. Li, S. Wang and X. Wang, EcoEnergy, 2024, 2, 205–219 CrossRef .
  7. B. Vellingiri, Environ. Res., 2023, 233, 116430 CrossRef CAS PubMed .
  8. L. Rani, A. L. Srivastav, J. Kaushal, D. P. Shukla, T. D. Pham and E. D. van Hullebusch, Environ. Res., 2023, 236, 116795 CrossRef CAS .
  9. D. Mei, L. Liu and B. Yan, Coord. Chem. Rev., 2023, 475, 214917 CrossRef CAS .
  10. S. Wang, G. Wei, Y. Xie, H. Shang, Z. Chen, H. Wang, H. Yang, G. I. N. Waterhouse and X. Wang, Sep. Purif. Technol., 2022, 303, 122256 CrossRef CAS .
  11. Y.-G. Wang, W. Jiang, X. Liu, L. Zhang, R.-P. Liang and J.-D. Qiu, Chem. Eng. J., 2023, 477, 146975 CrossRef CAS .
  12. Z. Li, Q. Meng, Y. Yang, X. Zou, Y. Yuan and G. Zhu, Chem. Sci., 2020, 11, 4747–4752 RSC .
  13. Z. Wang, Q. Meng, R. Ma, Z. Wang, Y. Yang, H. Sha, X. Ma, X. Ruan, X. Zou, Y. Yuan and G. Zhu, Chem, 2020, 6, 1683–1691 CAS .
  14. X. Liu, M. Ji, H. Lin, W. Jin, Y. Xue, Q. Wang and F. Ma, Sep. Purif. Technol., 2023, 327, 124890 CrossRef CAS .
  15. Y. Wang, Y. Zhang, Q. Li, Y. Li, L. Cao and W. Li, Carbohyd. Polym., 2020, 245, 116627 CrossRef CAS .
  16. J. Das, A. Rawat, L. Singh, A. Maiti, A. Bhatnagar and P. Mohanty, ACS Appl. Eng. Mater., 2023, 1, 2004–2017 CrossRef CAS .
  17. X. He, M. P. Dugas, J. N. Hodul, B. W. Boudouris and W. A. Phillip, Appl. Surf. Sci., 2024, 643, 158650 CrossRef CAS .
  18. J.-X. Ao, Y.-H. Yuan, X. Xu, L. Xu, Z. Xing, R. Li, G.-Z. Wu, X.-J. Guo, H.-J. Ma and Q.-N. Li, Ind. Eng. Chem. Res., 2019, 58, 8026–8034 CrossRef CAS .
  19. A. Ladshaw, L.-J. Kuo, J. Strivens, J. Wood, N. Schlafer, S. Yiacoumi, C. Tsouris and G. Gill, Ind. Eng. Chem. Res., 2017, 56, 2205–2211 CrossRef CAS .
  20. L. Astheimer, H. J. Schenk, E. G. Witte and K. Schwochau, Sep. Sci. Technol., 1983, 18, 307–339 CrossRef CAS .
  21. M. Fu, J. Ao, L. Ma, D. Kong, S. Qi, P. Zhang, G. Xu, M. Wu and H. Ma, Sep. Purif. Technol., 2022, 287, 120572 CrossRef CAS .
  22. T. Saito, S. Brown, S. Chatterjee, J. Kim, C. Tsouris, R. T. Mayes, L.-J. Kuo, G. Gill, Y. Oyola, C. J. Janke and S. Dai, J. Mater. Chem. A, 2014, 2, 14674–14681 RSC .
  23. J. Yu, H. Zhang, Q. Liu, J. Zhu, J. Liu, R. Chen and J. Wang, Int. J. Biol. Macromol., 2023, 253, 126866 CrossRef CAS .
  24. J. Yu, H. Zhang, Q. Liu, J. Yu, J. Zhu, R. Li, P. Liu, Y. Li and J. Wang, Chem. Eng. J., 2023, 471, 144705 CrossRef .
  25. Z. Mi, D. Zhang, J. Wang, S. Bi, J. Liu, X. Gao, D. Zhang, Y. Jiang, Z. Li, Y. Zhu and Z. Liu, New J. Chem., 2022, 46, 6296–6306 RSC .
  26. Z. Sun, Y. Chen, Y. Liu, B. Na, C. Meng, S. Zhang, S. Zou and H. Liu, J. Mater. Chem. A, 2021, 9, 21402–21409 RSC .
  27. L. Yang, H. Xiao, Y. Qian, X. Zhao, X.-Y. Kong, P. Liu, W. Xin, L. Fu, L. Jiang and L. Wen, Nat. Sustainable, 2022, 5, 71–80 CrossRef .
  28. B. C. Kholkhoev, Z. A. Matveev, K. N. Bardakova, I. A. Farion, P. S. Timashev and V. F. Burdukovskii, Polymer, 2023, 279, 126014 CrossRef .
  29. Z. Mi, Z. Liu, J. Yao, C. Wang, C. Zhou, D. Wang, X. Zhao, H. Zhou, Y. Zhang and C. Chen, Polym. Degrad. and Stabil., 2018, 151, 80–89 CrossRef .
  30. Z. Mi, L. Meng, J. Wang, L. Liao, Y. Huang, K. Zhang, J. Xiao, T. Xie, Y. Yan, Y. Zhong and Z. Liu, New J. Chem., 2023, 47, 14364–14373 RSC .
  31. Z. Liu, L. Wang, Z. Mi, S. Jin, D. Wang, X. Zhao, H. Zhou and C. Chen, Appl. Surf. Sci., 2019, 490, 7–17 CrossRef .
  32. Z. X. Liu, Z. M. Mi, S. Z. Jin, C. B. Wang, D. M. Wang, X. G. Zhao, H. W. Zhou and C. H. Chen, J. Membrane Sci., 2018, 557, 13–23 CrossRef CAS .
  33. Z. X. Liu, Z. M. Mi, C. H. Chen, H. W. Zhou, X. G. Zhao and D. M. Wang, Appl. Surf. Sci., 2017, 401, 69–78 CrossRef CAS .
  34. Z. M. Mi, Z. X. Liu, C. S. Tian, X. G. Zhao, H. W. Zhou, D. M. Wang and C. H. Chen, J. Polym. Sci. Polym. Chem., 2017, 55, 3253–3265 CrossRef CAS .
  35. W. Lin, Y. Lu and H. Zeng, J. Appl. Polym. Sci., 1993, 47, 45–52 CrossRef CAS .
  36. S. Shi, Y. Qian, P. Mei, Y. Yuan, N. Jia, M. Dong, J. Fan, Z. Guo and N. Wang, Nano Energy, 2020, 71, 104629 CrossRef .
  37. S. Vukovic, L. A. Watson, S. O. Kang, R. Custelcean and B. P. Hay, Inorg. Chem., 2012, 51, 3855–3859 CrossRef PubMed .
  38. D. Wang, J. Song, S. Lin, J. Wen, C. Ma, Y. Yuan, M. Lei, X. Wang, N. Wang and H. Wu, Adv. Funct. Mater., 2019, 29, 1901009 CrossRef .

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4nj02840f
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2024