Due to the lack of precise membrane structure control, the fabrication of nanofiltration membranes with excellent acid resistance and high separation performance remains a major challenge. Covalent organic frameworks (COFs), with their abundant porosity and highly ordered structure, are used to synthesize highly permeable nanofiltration membranes. In this paper, an acid-stable COF layer and a polysulfonamide (PSA) layer were sequentially prepared on a polyethersulfone (PES) ultrafiltration substrate by in-situ interfacial polymerization (IP). Due to the staggered stacking of COF and PSA layers and the interfacial polymerization (IP) process regulated by COF layers, the obtained COF-based composite membranes have sub-nanometer pore size and excellent rare earth ion separation performance. In addition, the composite membrane showed high rejection of more than 92.2% for trivalent rare earth ions (RE 3+) and high water permeability greater than 43.3 L h-1m-2bar-1 at both pH=6.8 and pH=1, which can be attributed to the high porosity and abundant transport paths provided by the COF layer and the staggered stacking structure between the COF layer and the PSA layer.
Scheme 1: Preparation of PEI-BDSC/TpPa/PES composite membranes.
Fig. 1 (a) ATR-FTIR spectra of PES substrates, TpPa/PES, and PEI-BDSC/TpPa/PES membranes. (b) WAXD spectra of TpPa powder (prepared at Pa and Tp concentrations of 4 and 0.1, 8 and 0.2, 12 and 0.3, 20 and 0.5, 28 and 0.7 gL-1) and PEI-BDSC/TpPa/PES membrane (Pa of 20 gL-1, Tp of 0.5 gL-1, BDSC of 0.2 gL-1). (c) Pore size distribution of TpPa powder prepared at different monomer concentrations when the Pa/Tp mass concentration ratio is 40. (d) Surface zeta potential of PES substrates, TpPa/PES, and PEI-BDSC/TpPa/PES membranes.
The composite nanofiltration membrane PEI-BDSC/TpPa/PES (scheme 1) was prepared by a two-step reaction, and p-phenylenediamine (Pa) and 2,4,6-trialdehyde-resorcinol (Tp) were first reacted on a PES substrate to obtain a TpPa film, and then polyethylenimine (PEI) and 1,3-benzenodisulfonyl chloride (BDSC) were further stacked through interfacial polymerization to form the final composite membrane. The chemical structure of TpPa/PES and PEI-BDSC/TpPa/PES membranes was confirmed by ATR-FTIR. As shown in Fig. 1a, the characteristic C=N telescopic vibration at 1519 cm−1 indicates that the TpPa COF formed by enols has been successfully synthesized. FTIR spectra also indicate that some of them may have been converted to ketone form, as evidenced by the characteristic C=C telescopic vibration of the ketone form at 1578 cm−1. In addition, the S-N stretching vibrations in the ATR-FTIR spectra of the PEI-BDSC/TpPa/PES composite film were observed at 937 and 914 cm−1, confirming the formation of the PEI-BDSC polyamide layer. The crystal structure of the PEI-BDSC/TpPa/PES composite film was evaluated by wide-field X-ray diffraction analysis (WAXD), and a representative WAXD plot is shown in Figure 1b to illustrate how the crystal structure of the TpPa powder is affected by the concentration of the monomer. As shown in Figure 1c, the predominantly sub-nanopore width of TpPa powder is about 0.7 nm, and the average pore width is slightly higher at a Pa concentration of 28 gL-1, which can be attributed to its higher crystallinity. For nanofiltration membranes, the surface zeta potential is critical for ion separation. As shown in Figure 1d, the zeta potential of TpPa/PES, PEI-BDSC/PES, and PEI-BDSC/TpPa/PES membranes decreased from 10 mV to -15 mV as the pH increased from 3 to 11, indicating that the electrostatic interaction on the membrane surface was weak. In particular, due to the introduction of the –NH2 group by the polyetherimide (PEI) monomer, the PEI-BDSC/PES membrane exhibits a zeta potential of up to 10 mV at pH=3.
Fig.2 Water contact angle (a-c), surface (a-c), and cross-section (d-f) SEM images of PES, TpPa/PES, and PEI-BDSC/TpPa/PES membranes.
As shown in Figure 2a-c, the water contact angles of PES, TpPa/PES, and PEI-BDSC/TpPa/PES are 39.8°, 63.7°, and 70.5°, respectively. Due to the abundant hydroxyl groups present on the surface, the TpPa layer exhibits higher hydrophilicity than the PEI-BDSC layer. According to cross-sectional topography measurements, the thickness of the TpPa and PEI-BDSC/TpPa films is 170 nm and 270 nm, respectively (Fig. 2d-f). In addition, COF nanoparticles were observed within the basal finger pores and remained stable during the filtration experiment, demonstrating that the permeate total organic carbon did not change over time.
Fig.3 AFM surface morphology of polyethersulfone membrane, TpPa/PES membrane and PEI-BDSC/TpPa/PES membrane.
Atomic force microscopy (AFM) is used to characterize membrane surface roughness, as shown in Figure 3. The roughness (Ra) of PES, TpPa/PES, and PEI-BDSC/TpPa/PES were 2.29, 8.47, and 3.80 nm, respectively. The TpPa/PES composite film has the highest surface roughness, which is due to the formation of TpPa particles (about 80 nm). After the formation of a thin layer of PEI-BDSC polysulfonamide on the surface of the composite film, the surface of the composite film becomes smoother, which may be the result of the process of interlayer regulation of IP by COF.
Fig.4 Cross-sectional X-ray spectroscopy (EDS) of PEI-BDSC/TpPa/PES films.
Fig. 5 (a) The effect of Pa and Tp concentrations on the permeability of composite membranes was studied when the ratio of Pa to Tp concentrations was 40. (b) The effect of PEI and BDSC concentrations on the permeability of composite membranes was studied when the ratio of PEI and BDSC concentrations was 20. Test conditions: cross-flow filtration, circulating flux > 300 Lh-1, feed solution of 1 gL-1 YCl3, pressure of 10 bar, pH 6.8, room temperature.
As shown in Figure 4, the structure of the composite membrane was further verified by SEM-EDS cross-section analysis. In the 0-100 nm range, the intensity of element S decreases sharply, then remains constant, and finally continues to rise. The EDS analysis results were consistent with the SEM morphology, which confirmed the stacked structure of the composite film.
The effects of synthesis conditions (TpPa and PEI-BDSC layers) on the permeability of the membrane were systematically studied by cross-flow filtration. Given the same Pa/Tp mass concentration ratio of 40, as the Pa concentration increased from 0 to 28 gL-1, the Y3+ rejection increased significantly from 81.2% to 93.1%, and the membrane permeability decreased from 105.7 Lh−1m−2bar−1 to 43.9 Lh−1m−2bar−1 (Fig. 5a). As the PEI concentration increased from 0 to 10 gL-1, the membrane rejection increased continuously from 26.9% to 93.7%, and the permeability decreased from 103.9 Lh-1m-2 bar-1 to 36.9 Lh-1 m-2 bar-1 (Fig. 5b).
Table 1: Comparison of the membranes prepared in this work with other reported acid-resistant polymer membranes or COF membranes.
As shown in Table 1, TpPa/PES membranes and PEI-BDSC membranes exhibited low Y3+ rejection rates of 26.8% and 80.2%, respectively. In contrast, the PEI-BDSC/TpPa/PES composite membrane showed an enhanced rejection rate of 92.3% (Table 1). The improvement of salt rejection rate of PEI-BDSC/TpPa/PES membranes is mainly due to the staggered pore structure between the PEI-BDSC layer and the TpPa layer.
Fig. 6 (a) Performance of PEI-BDSC/TpPa/PES membranes for different RE3+ and methyl blues. (b) YCl3 filtration performance of PEI-BDSC/TpPa/PES membranes at pH=6.8 before acid filtration and pH=1 after acid filtration. (c) Long-term acid stability of PEI-BDSC/TpPa/PES membranes. (d) Comparison of the permeability and rejection of the trivalent ions of the PEI-BDSC/TpPa/PES membranes with the reported data. Test conditions: cross-flow filtration, flux > 300 Lh−1, feed solution of 1 gL-1 YCl3 or 0.5 gL-1 polyethylene glycol (PEG), feed pressure of 10 bar, pH 6.8, room temperature.
Further studies of the nanofiltration performance of other rare earth ions were investigated and compared using methylene blue, as shown in Figure 6a. The permeability of the composite membrane was greater than 52.7 Lh-1m-2bar-1, and the RE3+ rejection rates of LaCl3, NdCl3, GaCl3, YbCl3 and YCl3 were 92.9%, 92.8%, 92.8%, 92.3% and 92.2%, respectively, indicating that the composite membrane could be used for the separation of other rare earth elements. As shown in Fig. 6b, the YCl3 rejection rates of TpPa/PES, PEI-BDSC/PES and PEI-BDSC/TpPa/PES membranes were 29.2%, 81.8% and 92.2%, respectively, which were higher than those at pH=6.8. The slight increase in salt rejection may be due to the higher positive zeta potential under acidic conditions. Long-term acid resistance measurements were made by soaking the membrane in 0.1 molL-1HCl and 1 gL-1YCl3 solutions for a specific period of time. As shown in Figure 6c, the Y3+ rejection rate remains unchanged for 93 days.
Shows high acid resistance. In the first 7 days, the permeability decreased slightly from 47.0 Lh−1m−2bar−1 to 43.3 Lh−1m−2bar−1, which may be attributed to chain rearrangement to form a new ideal acid-resistant structure. As shown in Figure 6d and Table 1, the PEI-BDSC/TpPa/PES membranes show comparable trivalent ion rejection to most reported commercial acid-resistant membranes and polysulfonamide-based nanofiltration membranes, but with a much higher permeability of 53.6 Lh-1m-2bar-1.
The above was published in Chemical Engineering Journal. The first authors of the paper are Wei Lai, Jiangxi University of Science and Technology, Ganjiang Innovation Research Institute, Chinese Academy of Sciences, Institute of Process Engineering, Chinese Academy of Sciences, Key Laboratory of Green Process and Engineering, State Key Laboratory of Multiphase Complex Systems, Beijing Key Laboratory of Ionic Liquid Cleaning Processes, Jiangxi University of Science and Technology, and Linglong Shan, Institute of Process Engineering, Chinese Academy of Sciences, Langfang Green Industrial Technology Center, and the corresponding author is Institute of Process Engineering, Chinese Academy of Sciences. Luo Shuangjiang from Langfang Green Industrial Technology Center and Zhang Suojiang from the Institute of Process Engineering, Chinese Academy of Sciences.
Original link: https://doi.org/10.1016/j.cej.2022.137965