5.2
Impact Factor
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages
Filter by Categories
Corrigendum
Current Issue
Editorial
Erratum
Full Length Article
Full lenth article
Letter to Editor
Original Article
Research article
Retraction notice
Review
Review Article
SPECIAL ISSUE: ENVIRONMENTAL CHEMISTRY
5.3
Impact Factor
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages
Filter by Categories
Corrigendum
Current Issue
Editorial
Erratum
Full Length Article
Full lenth article
Letter to Editor
Original Article
Research article
Retraction notice
Review
Review Article
SPECIAL ISSUE: ENVIRONMENTAL CHEMISTRY
View/Download PDF

Translate this page into:

Original Article
ARTICLE IN PRESS
doi:
10.25259/AJC_668_2025

In-situ growth of polyimide (PI)/BiOCl mixed matrix membrane with excellent permeability and photocatalytic performance

, , ,
aSchool of Chemistry and Chemical Engineering, Jinggangshan University, Ji’an, China
bKey Laboratory of Jiangxi Province for Special Optoelectronic Artificial Crystal Materials, Jinggangshan University, Ji’an, China

*Corresponding author: E-mail address: hanrunlin@163.com (R. Han)

Received: , Accepted: ,
© 2026 Arabian Journal of Chemistry - Published by Scientific Scholar
Licence
This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-Share Alike 4.0 License, which allows others to remix, transform, and build upon the work non-commercially, as long as the author is credited and the new creations are licensed under the identical terms.

Abstract

Inorganic nanophotocatalysts have high efficiency in the deep treatment of organic wastewater. However, nanocatalysts are often prone to loss and require additional recycling, which limits their wide application. In this work, the introduction of nano BiOCl into a polyimide (PI) membrane matrix was achieved by in-situ growth and immersion precipitation phase inversion method. This resulted in high-performance PI/BiOCl photocatalytic membranes which efficiently combine the photocatalytic properties of BiOCl and the separation performance of PI membrane. Detailed investigation was conducted into the effect of BiCl3 contents on the photocatalytic activity of the mixed matrix membrane. The membranes exhibited high flux in the separation of organic compounds and demonstrated effective performance in the photocatalytic degradation of organics. The experimental results demonstrated that the PI/BiOCl photocatalytic membranes exhibited rejection of 82.8% for ciprofloxacin (CIP) with fluxes as high as 3269.2 L·m–2·h–1·bar–1. And photocatalytic degradations of tetracycline hydrochloride (TC) and CIP were 99.8% and 96.4%, respectively. It is also demonstrated that the membrane exhibits excellent TC degradation efficiency during five cycles within the same time frame. The mechanism of photocatalytic degradation was explored in detail. The prepared PI/BiOCl photocatalytic membrane in this work shows excellent performance and the study provides new insights for efficient separation and degradation of antibiotics in wastewater.

Keywords

Antibiotics treatment
Mixed matrix membrane
Photocatalytic membrane
Polyimide
Tetracycline hydrochloride

1. Introduction

With large quantities of industrial or farm wastewater being discharged into the ecosystem, harmful small-molecule organics such as antibiotics, dyes, etc. pose a serious threat to the environment [1,2]. In order to solve these problems, various researches have been carried out for the removal of organic matter. Common wastewater treatment technologies for organics include membrane separation, catalytic oxidation, and adsorption [3-5]. Among them, photocatalytic technology has received great attention due to its great potential in removing antibiotics from aqueous solutions through a green, economical and efficient process [6]. Titanium dioxide (TiO2) is one of the most important photocatalysts widely used for degrading dyes, antibiotics, etc., but it has a large band gap that somewhat limits its large-scale application [7]. A large number of studies have been devoted to adjusting the catalyst band gap and thus improving the photocatalytic degradation efficiency [8]. There are also some novel photocatalysts that exhibit high activity and have gained extensive research such as g-C3N4 [9,10], cadmium sulfide (CdS) [11], BiOX (X=Cl, Br, I) [12,13]. Among them, BiOCl has been widely studied in photocatalytic applications due to its ease of preparation, high stability, tunable band gap, and low toxicity [14].

Generally, a photocatalytic suspension system is more popular because most of the semiconductor photocatalysts involved in the catalytic process are in the nanometer or micrometer scale, which is conducive to mass transfer and photocatalytic reactions [15]. However, recovering the catalysts from treated wastewater requires an additional, time-consuming, and costly separation process. Therefore, to improve stability and lifetime, the catalyst is usually immobilized on various carrier materials [16]. Membrane separation technology has many advantages, such as environmental friendliness, simple operation, high processing efficiency, and easy to integrate. It has been widely used in seawater desalination, biomedicine, gas separation, energy, and environment [17-19]. However, membrane separation processes often cause surface contamination problems due to the adsorption and deposition of waste liquid pollutants on the membrane surface. The resulting pollutants will occupy the active sites of the membrane pores and change the surface properties of the material, thus leading to the degradation of the separation efficiency and severely restricting the recycling performance of the membrane material [20,21].

To overcome the issue of catalyst recovering and membrane contamination, there has been an increasing focus on immobilizing catalysts on membrane substrates to form catalytic membranes in the fields of materials and catalysis research [22]. By coupling photocatalytic technology and membrane separation technology to construct a photocatalytic self-cleaning membrane, the organic pollutants adsorbed and deposited on the membrane surface can be effectively degraded, and the formation of a fouling layer on the surface of the membrane can be inhibited, thus improving the membrane separation efficiency [23,24]. Precisely, the photocatalytic membrane is the new system combining photocatalysis and membrane separation technology, which has received widespread attention in the field of water treatment and environmental protection in recent years [25].

Existing photocatalytic membranes are generally prepared by first preparing inorganic nanoparticles with high catalytic activity and then embedding them into a polymer matrix to obtain mixed matrix membranes with photocatalytic activity [22]. However, the poor compatibility of organic and inorganic materials often leads to a generally severe agglomeration of photocatalysts, low catalyst additions and severe catalyst encapsulation, which limits the widespread use of photocatalytic membranes [26].

Polyimide (PI) has a high glass transition temperature, excellent resistance to a wide range of organic solvents such as hydrocarbons, toluene, ketones and alcohols. It shows good resistance over a wide pH range, especially when crosslinking is performed [27,28]. Therefore, a lot of research has been conducted on the preparation of PI membranes for water treatment, gas separation and other applications [29]. Due to the excellent membrane-forming properties and chemical stability of PI, it can provide the basis for rapid preparation and long-term stability of photocatalytic membranes. However, PI membrane with both excellent photocatalytic and separation properties has not been reported yet.

Based on the above, we successfully utilized the in-situ growth principle of BiCl₃ to BiOCl by hydrolysis in a gel bath, enabling the preparation of PI/BiOCl membranes with high separation selectivity and photocatalytic activity using the immersion phase inversion method. This method not only avoids the agglomeration of nanocatalysts but also significantly increases the catalyst loading. The resulting photocatalytic membrane combines the advantages of a high-performance PI membrane and BiOCl photocatalysts, with the catalysts being uniformly loaded in the polymer matrix. This avoids the issues of catalyst separation, recycling and loss that are associated with the traditional suspended photocatalytic wastewater treatment process. This provides a new approach to the advanced treatment of salt-containing antibiotic or dye wastewater.

2. Materials and Methods

2.1. Materials

PI (P84, Mw: 25,000) was provided by HP Polymer Inc. (Lenzing, Austria). NaCl, MgCl2, Na2SO4, N,N-dimethylacetamide (DMAc), BiCl3, isopropanol (IPA), ethylenediaminetetraacetic acid disodium salt (EDTA-2Na), p-benzoquinone (PBQ), ciprofloxacin (CIP, C17H18FN3O3) and tetracycline hydrochloride (TC, C22H25ClN2O8) are all analytically pure and obtained from Aladdin (Shanghai, China) without further purification. Rhodamine B (RhB, C28H31ClN2O3), methyl orange (MO, C14H14N3SO3Na) and methylene blue (MB, C16H18ClN3S) are provided by Tianjin Damao Chemical Reagent Partnership Enterprise (Tianjin, China).

2.2. Preparation of photocatalytic membrane

The PI membrane was cast on polyester nonwovens with 14 wt.% of PI in DMAc solution, which was designated as M1. The immersion phase inversion method was used to prepare the membrane, with deionized water employed as the coagulation bath. The M1 membrane was selected as the control sample for this study. The preparation of PI/BiOCl photocatalytic membranes followed the same method except the composition of the casting solution. As shown in Table 1, the solute concentration of the casting solution was fixed at 14 wt.%, and the mass ratio of PI and BiCl₃ was 4:1 and 5:5 for M2 and M3, respectively. As an example, M2 was configured with 30 g of membrane solution. Firstly, 25.8 g of DMAc was added, 3.36 g of PI was added under rapid magnetic stirring, and 0.84 g of BiCl3 was added after dissolution of PI. When the casting solution is dissolved, it was scraped on clean glass plate and immersed into deionized water to form a membrane.

Table 1. Formulation of casting solution.
Name PI (wt.%) BiCl3(wt.%) DMAc(wt.%)
M1 14 0 86
M2 11.2 2.8 86
M3 7 7 86

2.3. Membrane characterization

The infrared absorption spectra of the membrane samples were obtained by transmission method to analyze the chemical composition of the samples with FTIR (Nicolet iS10, Thermo Scientific, USA). Contact angle meter (CA100A, Shanghai Innuo Precision) was used to test the hydrophilicity of the membrane surface. The membranes were coated with a thin-layer of gold to capture SEM (ZEISS Sigma 300, Germany) image at an acceleration voltage of 12 KV. Meantime, the energy dispersive spectroscopy (EDS) was used for elemental content analysis. The elemental composition of the material was analyzed using an X-ray photoelectron spectrometer (XPS, K-Alpha, ThermoFisher). An X-ray diffractometer (XRD, D8ADVANCE, Bruker) was applied to analyze the diffraction patterns of membrane samples with a scanning range from 5° to 60° at a scanning rate of 5°∙min–1. A universal testing machine QT-6203S (Jiangsu Qiantong Instrument Equipment Co., Ltd., China) was used to perform tensile analysis on the membrane samples. The membrane width was selected as 10 mm, and the tensile length was 50 mm. To verify the stability of the catalytic membrane, we simultaneously measured the concentration of Bi ions in the solution during a four-hour continuous catalytic degradation of TC with inductively coupled plasma optical emission spectrometer (ICP-OES, ICP5000, Focused Photonics Inc. China).

Membrane flux and rejection were measured at 0.1 MPa and room temperature. The solution concentrations of organics such as RhB, MB, TC and CIP are tested with a UV-Vis spectrophotometer (752N, Shanghai Youke Instrumentation Co., Ltd.). The concentrations of salts such as NaCl, MgCl2, and Na2SO4 were tested with a conductivity meter (DDS-11A, Lei Magnetic, Shanghai, China). And the membrane rejection (R) was calculated as follow Eq. (1) [18]:

(1)
R= 1− ct c 0

The membrane flux (F) was characterized by the setup of a solvent filter under a vacuum pump at room temperature. It was calculated as follows Eq. (2) [30]:

(2)
F=WAt

where W was the total volume of water or solution permeated during filtration process, A was the effective membrane area, and t was the operating time.

2.4. Photocatalytic performance examination

The photocatalytic performance of PI/BiOCl membranes was evaluated under UV light irradiation in the 20 mg∙L–1 of MB, RhB, TC and CIP solution at room temperature without adding acids or bases. The photodegradation experiments of target pollutants were carried out in a beaker with 60 mm diameter of membrane sample placed in 200 mL of target pollutant solution. A 30 W UV lamp (254 nm) was used as the light source. After the degradation system was left in the dark for 30 min to equilibrate the adsorption, the UV lamp was switched on for irradiation, and the same volume of sample solution (2 mL) was taken out at every half an hour for test. The absorbance was then measured by using a UV-Vis spectrophotometer, and the photocatalytic degradation efficiency of the PI/BiOCl membrane was calculated with Eq. (3) [31]:

(3)
D= C0 −Ct C0 ×100%

where C0 and Ct are the initial absorbance and absorbance at any t time for organics.

To investigate the degradation performance of PI/BiOCl membrane under visible light, M2 membrane was tested in degrading TC under a xenon lamp light source system (CEL-PF300-T10, 300W, Beijing Zhongjiao Jinyuan Technology Co., Ltd., China). After equilibrating adsorption by placing the degradation system in darkness for 1 min, the xenon lamp was activated for irradiation. Equivalent volumes of sample solution (2 mL) were collected for testing at hourly intervals.

The selectivity values of salt and organic compound, using SNaCl/MB for example, were calculated using Eqs. (4) [32]:

(4)
S NaCl/MB = 1−RNaCl 1−RMB ×100%

2.5. Photocatalytic stability of membrane

To investigate the stability and reusability of the photocatalytic membrane, a five-cycle photocatalytic performance experiment was conducted on M2 and M3. Under UV irradiation, the membranes were used to treat 100 mL of 10 mg∙L–1 TC solution. The used membranes were then recovered and cleaned using an ethanol soak, after that, it was added into the same TC solution for another photocatalytic degradation test. This process was repeated five times to evaluate the actual performance of photocatalytic membrane.

2.6. Photocatalytic mechanism study of the PI/BiOCl membrane

To detect the active species involved in the photocatalytic process, the hydroxyl radical (‧OH), the superoxide radical (‧O₂−) and the hole (h⁺) were investigated using the trapping agents IPA, PBQ and EDTA-2Na, respectively. This method was similar to the previous photocatalytic activity test, except that 1 mM of a trapping agent was added to 200 mL of TC or CIP solution (20 mg L-1) containing the photocatalytic membrane. The absorbance of the TC and CIP solutions was detected after 12 h. Degradation solution of TC were analyzed using an Ultimate 3000 UHPLC-Q Exactive liquid chromatography-mass spectrometry (LC-MS) system (Thermo Scientific, USA) with a Zorbax Eclipse Plus C18 column (100 mm × 2.1 mm, 1.8 µm) at a column temperature of 30°C and an injection volume of 20 µL. Aqueous solution of 0.1% formic acid and acetonitrile were used as the mobile phases.

3. Results and Discussion

3.1. Membrane characterization

To investigate the relationship between the structures and their properties of the M1, M2, and M3 membrane materials, SEM was used in this experiment to characterize the external appearance of the membrane materials, and the magnifications of the samples were 5,000 and 50,000 times. As can be seen in Figures 1(a-f), the M2 and M3 membranes containing BiOCl exhibit rough surface structure while the M1 membrane is relatively smooth. The M2 membrane shows obvious nanosheets in the membrane surface as seen in Figures 1(d).

SEM maps (a&b) of M1, (c&d) of M2 and (e&f) of M3; (g) EDS zero-loss image (the inset is atomic weight composition of the elements) and elemental mapping images (h-l) of C, O, N, Bi and Cl elements in the selected regions of M2.
Figure 1.
SEM maps (a&b) of M1, (c&d) of M2 and (e&f) of M3; (g) EDS zero-loss image (the inset is atomic weight composition of the elements) and elemental mapping images (h-l) of C, O, N, Bi and Cl elements in the selected regions of M2.

Figures 1(g-l) show the SEM image and corresponding EDS-mapping diagrams of each element for the M2 membrane, which proved that the five elements, C, O, N, Bi, and Cl, existed in the samples and were uniformly dispersed. As shown in the inset of Figure 1(g), the elements Bi and Cl are from BiOCl with atomic percentages of 1.6% and 1.14%, respectively, while the element N is from the PI matrix with an atomic percentage of 3.78%. The higher atomic percentage of C of 80.44% is due to the ability of the X-rays to penetrate the membrane and the fact that some of the carbon is from the conductive adhesive. The atomic percentage of O is 13.08%, both from BiOCl and PI. Therefore PI/BiOCl mixed matrix membrane was successfully prepared with the simple immersion precipitation phase inversion method and in-situ growth of BiCl3 uniformly in the polymer matrix.

In Figure 2(a), the wide-scan XPS spectrum provides the chemical composition of M2 sample. From the figure, it can be noticed that the sample contains five elements, C, O, N, Bi and Cl, as BiOCl is formed in the PI matrix. Figure 2(b) shows the deconvoluted high-resolution XPS N 1s spectrum with the only binding energy peak at 399.6 eV, which can be ascribed to bond of N-C. As shown in Figure 2(c), in the high-resolution XPS O 1s spectrum, there are two peaks located at 531.5 and 529.16 eV. These peaks were attributed to the C-O groups of PI and Bi-O bond of BiOCl, respectively [33]. Figure 2(d) shows two binding energy peaks at 163.8 and 158.5 eV, which are representing Bi4f5/2 and Bi4f7/2, respectively [34]. As shown in Figure 2(e), in the high-resolution XPS C 1s spectrum, two peaks at 287.5 and 284.8 eV corresponding to C-O and C=O, respectively. Figure 2(f) shows the FTIR spectra of M1, M2 and M3 samples. As shown in the figure, a characteristic C=O peak appears at 1722 cm–1 which belongs to PI. The absorption peak at 3440 cm–1 corresponds to the O-H stretching vibration of H2O, which may be the result of incomplete drying of the membrane samples during testing. Compared to the pure membrane M1, new peak at 1420 cm–1 appear in M2 and M3, which can be ascribed to Bi-Cl bond of BiOCl. From the XPS and FTIR spectra, it can be learned that BiOCl was successfully introduced into the PI matrix, resulting in a homogeneously dispersed organic-inorganic hybrid membrane, and the results are consistent with the EDS mapping results.

(a) Wide-scan XPS spectrum; High-resolution XPS spectra of (b) N 1s, (c) O 1s, (d) Bi 4f and (e) C 1s of M2 sample; (f) FTIR spectra of M1, M2 and M3 samples.
Figure 2.
(a) Wide-scan XPS spectrum; High-resolution XPS spectra of (b) N 1s, (c) O 1s, (d) Bi 4f and (e) C 1s of M2 sample; (f) FTIR spectra of M1, M2 and M3 samples.

For comparison, BiOCl sample was prepared by dissolving BiCl3 in DMAc and then pouring the solution into water for hydrolysis, the insoluble material was filtered and dried for XRD analysis. Figure 3 shows XRD patterns of pure BiOCl and M2 samples. The BiOCl and M2 samples have several characteristic peaks, where 2θ of 11.99°, 25.86°, 32.53°, 46.64°, 54.09° and 58.60° for (001), (101), (110), (200), (211) and (212), respectively. All the characteristic peaks align with the BiOCl reference (PDF#06–0249). These diffraction peaks indicate a successful synthesis of BiOCl in the membrane.

XRD patterns of BiOCl standard card (PDF#06-0249), BiOCl and M2 samples.
Figure 3.
XRD patterns of BiOCl standard card (PDF#06-0249), BiOCl and M2 samples.

Since all the PI-based membranes are coated onto the polyester nonwoven fabric, the mechanical properties of the nonwoven fabric was also tested as a blank comparison. As shown in the Figure 4, the polyester nonwoven fabric exhibits a fracture elongation of 14%. After being coated as a composite membrane layer, the fracture elongation of the membranes increased. Among them, the M1 membrane, with the highest polymer content, achieved a fracture elongation as high as 24%, indicating excellent elasticity. Regarding the tensile strength of the membranes, the figure reveals that the introduction of nanomaterial BiOCl in the M2 membrane enhances its mechanical strength to a certain extent, resulting in higher tensile strength. The stress of the membrane containing BiOCl increased from 35.7 MPa to 43.2 MPa, exhibiting greater tensile strength than the membrane without BiOCl.

Strain-stress diagrams for the PI membrane (polyester nonwoven fabric for blank, M1, M2, and M3).
Figure 4.
Strain-stress diagrams for the PI membrane (polyester nonwoven fabric for blank, M1, M2, and M3).

3.2. Separation performance of the membranes

Figure 5(a) shows the water flux of the membranes prepared with different contents of BiCl3. For comparison, the M1 membrane without BiOCl shows a water flux of 1597.0 L‧m–1‧h–1‧bar–1 with an initial contact angle of 71.2°, which is generally hydrophilic because of a small amount of carboxyl groups remaining in its molecular main chain. When BiOCl was introduced in the membrane matrix, the water fluxes for M2 and M3 were obviously enhanced to 3269.2 and 3943.9 L‧m–1‧h–1‧bar–1 with initial contact angles of 29.8° and 4.2°, respectively. The hybrid membrane shows improved water flux because of high hydrophilicity of the membrane surface with BiOCl. Figure 5(b) displays the change of contact angle during 12 s of different membranes. It can be concluded that the contact angle of M1 membrane did not change in 12 s. However, the contact angles of M2 and M3 decreased rapidly during 12 s and the water droplets were completely spread out after 12 s. Especially, the contact angle of M3 membrane was smaller and decreased to 0° in 6 s. So, the PI/BiOCl membranes show excellent hydrophilicity and water flux.

(a) Pure water flux of the membranes (b) contact angle of the membrane at different time.
Figure 5.
(a) Pure water flux of the membranes (b) contact angle of the membrane at different time.

There are various dyes and antibiotics in industrial wastewater. To test the practical application of membranes, this experiment tested the rejection and permeate flux of M1, M2 and M3 membranes for inorganic salts, dyes and antibiotics, and the results have been presented in Figure 6. All three membranes exhibit high permeation flux (above 445 L‧m–1‧h–1‧bar–1), which is consistent with the pure water flux. And the introduction of BiOCl evidently improves the permeate ability of the membrane. At the same time, the rejection of inorganic salts for all three membranes is almost zero, which indicates that the membrane pores are large. Specifically, the M1 membrane had a flux of more than 594 L‧m–1‧h–1‧bar–1 and a rejection of 79% for CIP. The fluxes of M2 membrane nearly doubled for the CIP separation, with a rejection of 83% for CIP. The M3 membrane shows higher flux but lower rejection to antibiotics because of loose structure compared with M2 membrane. The membranes showed higher rejection of organic dyes RhB and MB. The rejection of the membranes for MB was close to 100%, which may be due to the possible chemical reaction between the amino group in the molecule of MB and the carboxyl group in the chain structure of PI. Therefore, the dye molecules were almost undetectable in the filtrate.

Flux and rejection of the membranes (a)M1, (b) M2, (c), M3.
Figure 6.
Flux and rejection of the membranes (a)M1, (b) M2, (c), M3.

Compared with traditional drug or dye desalination technology, membrane separation technology has significant advantages in terms of process energy consumption, process flow, and green environmental protection. The development of separation membrane with high selectivity has important application value. The membrane in this study not only has high permeate flux, but also has almost zero rejection of inorganic salts and high rejection of antibiotics and dyes, which are very suitable for the desalination process. The M2 membrane shows high selectivity of SNaCl/MB (510.1) and SNaCl/CIP (7). According to the data, it can effectively separate the inorganic salts in the mixture solution and allow them to permeate through the membrane. In the production process of antibiotics or dyes, a certain amount of inorganic salt must be added. Therefore, the prepared membrane can be used efficiently in desalination technology.

3.3. Photocatalytic performance of the membranes

Figure 7(a) shows the degradation performance of the membrane when treating 200 mL of MO at a concentration of 20 mg∙L–1. In contrast, the decrease in dye concentration after the addition of M1 and M2 catalytic membranes was very small under dark conditions. In contrast, the dye concentration decreased by 19.6% after 26 h of UV irradiation without the addition of the membrane. With the addition of the M1 membrane, dye degradation was accelerated by 47.0%. With the addition of the M2 catalytic membrane, dye degradation was significantly accelerated, with 99.0% degradation achieved.

Photocatalytic degradation of (a) MO, (b) RhB, (c) TC and (d) CIP; Kinetics plot of M2 and M3 against (e) TC and (f) CIP.
Figure 7.
Photocatalytic degradation of (a) MO, (b) RhB, (c) TC and (d) CIP; Kinetics plot of M2 and M3 against (e) TC and (f) CIP.

Figure 7(b) demonstrates the degradation efficiency of the membranes for RhB. Similar with the treatment of MO, a decrease in dye concentration occurs with the addition of M1 and M2 catalytic membranes under dark conditions, but the adsorption of RhB by the membranes is larger. Whereas, under UV irradiation conditions, after 6 h of treatment without the addition of membranes, RhB dye degraded 53.7% due to UV, the degradation efficiency of the dye was accelerated to 83.1% with the addition of M1 membranes, and the rate of degradation of the dye was significantly accelerated to 95.9% with the addition of catalytic membrane M2. The results showed that the membranes without the introduction of nanocatalysts also promoted the degradation of the dyes to a certain extent, which could be attributed to the adsorption of the dyes by the membranes, which were more susceptible to UV degradation on the surface of the membranes.

As shown in Figure 7(c), after the introduction of M1, M2 and M3 membranes under dark conditions, the concentration of the TC decreased by 38.3%, 36.6%, and 32.8%, respectively, due to adsorption of TC by the membranes. In the absence of a membrane under UV irradiation, the concentration of TC decreased by 54.2% for 14 h. The degradation rates of TC were 76.8%, 98.2% and 99.8%, for M1, M2 and M3 membrane used in the system, respectively, after 14 h UV irradiation. It means the photocatalysts in the membrane matrix play an important role in the degradation of TC.

As shown in Figure 7(d), after the introduction of M1, M2 and M3 membranes under dark conditions, the CIP concentration decreased by 39.5%, 35.1% and 42.7%, respectively, due to the adsorption of CIP by the membranes during 16 h. The greater adsorption capacity of PI and PI/BiOCl photocatalytic membranes for CIP under dark conditions may be attributed to the fact that CIP has a lively piperazine ring in its molecular structure, which is able to produce strong covalent interactions with the carboxyl groups in the PI molecule, resulting in chemisorption. Under the UV irradiation condition without membrane, the concentration of CIP decreased by 37.7% after 16 h. After the introduction of M1, M2 and M3 membranes, the degradation rates of CIP under UV conditions after 16 h were 70.6%, 96.4% and 96.0%, respectively. This also indicates that the photodegradation efficiency of the membranes for CIP was significantly improved after the introduction of photocatalysts in the PI membrane matrix.

The fitting results for the kinetic plot have been shown in Figure 7(e) for TC. The -ln[Ct/C0]∼t curves of M2 and M3 for TC degradation are well fitted linearly to the pseudo-first-order kinetic model with R2 of 0.9861 and 0.9731, respectively. The degradation rate constant k was calculated from the intercept (1/k) of the -ln[Ct/C0]∼t linear curve, which is 0.1343 and 0.1060 h–1 for M2 and M3 membranes, respectively. This means the M2 membrane has a higher activity due to a relatively more nanocrystals in its membrane matrix as shown in the SEM images. Moreover, the fitting results of kinetic plot of CIP degradation was also shown in Figure 7(f). It can be found that the degradation rate constant k is 0.07461 and 0.07823 h–1 for M2 and M3 membranes, respectively, which means that the M2 and M3 membranes have the similar performance on the degradation of CIP.

The photocatalytic performance of M2 membrane was also tested with a visible light (300 W, xenon lamp). As shown in Figure 8, the photocatalytic performance of the M2 membrane under UV light (30 W UV lamp) and visible light conditions are compared. It was found that, visible light has a minimal effect on the degradation of TC. However, after introducing the catalytic membrane, TC degradation reached 84% within 12 h, indicating that visible light also possesses excellent photocatalytic activity. Nevertheless, its catalytic activity is slightly lower than that of UV light. Additionally, the UV lamp used had a low power output, making UV photocatalysis more advantageous for this reaction.

The degradation effect of M2 on TC under UV and visible light conditions.
Figure 8.
The degradation effect of M2 on TC under UV and visible light conditions.

Table 2 [35-39] lists the comparisons between this study and other photocatalytic membrane studies. It can be seen that the PI/BiOCl photocatalytic membranes developed in this study not only have high separation selectivity and permeation flux, but also have high photocatalytic efficiency. The membrane has considerable advantages and application value in the field of antibiotic wastewater treatment.

Table 2. Comparison of photocatalytic membrane performance with other literature.
Membrane Separation performance
 Photocatalytic performance
 Ref
Pollutants F (LMH) R (%) Pollutants Degradation (%) Light
ZnO/PEN Oil/water emulsion 2053.4 99.4 MB 91.7 Xenon lamp (300W) [35]
MXene/PVDF Oil/water emulsion 1022.7 99.3 MB 90 UV [36]
MXene/COF MB 217.6 99.8 n.a. n.a. n.a. [37]
TiO₂/MXene/PES BSA 756.8 95.0 n.a. n.a. n.a. [38]
NCQDs/BiOBr/TiO₂/PVDF n.a. n.a. n.a. TC 77 Xenon lamp (300W) [39]
PI/BiOCl MB 3913.9 99.9 TC, CIP 98, 99 UV (30W) This work
PEN: poly(arylene ether nitrile), PVDF: polyvinylidene fluoride, PES: polyethersulfone, NCQDs: modification of CQDs (Carbon quantum dots) with N atoms.

3.4. Cycle stability of photocatalytic membranes

Photocatalytic membranes require a high degree of cyclic stability during use, which can reduce the cost of utilization. The performance of TC cycle test has been shown in Figure 9. The photocatalytic degradation efficiencies of M2 membrane are higher than 91.1% and those of M3 membrane are higher than 91.7% in five-cycle tests. After five cycles, M2 decreased by 2.9% and M3 decreased by 3.1%, demonstrating high stability and consistent degradation efficiency of the photocatalytic membrane. In conclusion, the PI/BiOCl membrane exhibits excellent photocatalytic reproducibility and outstanding stability in the cyclic degradation of TC, making it a promising candidate material for the treatment of antibiotic degradation. No Bi ions were detected in the solution after photocatalytic reaction, indicating that the BiOCl nanocrystals exhibit excellent stability in this catalytic reaction without leaching of Bi ions.

Cycle stability of (a) M2 and (b) M3 in degradation of TC.
Figure 9.
Cycle stability of (a) M2 and (b) M3 in degradation of TC.

3.5. Self-cleaning ability of membranes

PI membranes are commonly used to separate dyes and micropollutants. However, accumulated contaminants on the membrane surface or within pores can lead to fouling and performance degradation. Incorporating photocatalytic materials into the membrane separation layer can effectively remove these contaminants without the need for chemical cleaning. As illustrated in Figures 10(a and b), the TC rejection rate of the composite membrane decreased from 37.0% to 8.0% and its flux decreased by 29% under dark conditions after five cycles due to membrane fouling. However, after UV irradiation, the dye rejection rate of the composite membrane decreased by only 9.0%, while its flux decreased by 9.2%. As shown in Figures 10(c and d), the composite membrane’s CIP rejection rate decreased from 82.1% to 15.7% after five cycles, while its flux decreased by 66.4% due to membrane fouling. Following UV irradiation, the composite membrane’s dye rejection decreased by 32.1%, while its flux decreased by only 11.2%. These results confirm that the PI/BiOCl composite membrane exhibits excellent self-cleaning capabilities under UV irradiation. The incorporation of photocatalytic BiOCl nanocrystals enables the effective degradation of organic compounds that adhere to its surface.

Effect of UV light on the self-cleaning performance of M2 membrane on (a & b) TC, and (c & d) CIP.
Figure 10.
Effect of UV light on the self-cleaning performance of M2 membrane on (a & b) TC, and (c & d) CIP.

3.6. Degradation mechanism of photocatalytic membrane

The effects of different radical trapping agents on the degradation performance of TC and CIP with M2 membrane were shown in Figures 11(a and b). Figure 11(a) shows the degree to which the photocatalytic reaction activity of TC was inhibited by different scavengers. Adding IPA did not affect TC degradation, whereas adding PBQ and EDTA-2Na inhibited TC degradation efficiency by 73.6% and 24.5%, respectively. This demonstrates that ‧O₂− and h⁺ play a major role in the photocatalytic degradation of TC. As shown in Figure 11(b), radical trapping agents have similar effect on the inhibition of photocatalytic reaction by 43.9% and 45.6%, respectively. It confirms that the PI/BiOCl membrane produces ‧O₂− and h⁺, which degrade antibiotics efficiently. However, IPA can also inhibit 4.1% of the reaction efficiency, as ·OH is produced during the photocatalytic reaction.

Photodegradation efficiency of (a) TC and (b) CIP with different radical trapping agents.
Figure 11.
Photodegradation efficiency of (a) TC and (b) CIP with different radical trapping agents.

We also analyzed the degradation products generated during the photocatalytic degradation of TC by the M2 membrane using LCMS. As shown in Figure 12, the figure presents the possible degradation products of TC. T2 is generated through hydroxylation, and product T3 is formed via decarboxylation of T2. At the same time, oxidation and ring-opening processes also take place, resulting in products such as T5 and T7. These products further undergo oxidation reactions of the cyclic structure to generate intermediate products with smaller molecular weights and then decomposed into CO2, H₂O and NH₄+.​

Possible intermediates of TC generated during photocatalytic degradation as analyzed by LCMS.
Figure 12.
Possible intermediates of TC generated during photocatalytic degradation as analyzed by LCMS.

4. Conclusions

In this study, in-situ generated PI/BiOCl photocatalytic membranes were successfully prepared via a simple immersion phase inversion method. SEM images show that a large number of nanocrystals are generated on the membrane surface, while EDS results confirm the presence of Bi and Cl elements in the membrane. XRD analysis reveals that the nanocrystals formed in the membrane structure are BiOCl. XPS analysis further confirms the successful preparation of the PI/BiOCl mixed matrix membrane. With the introduction of BiOCl, the contact angle of the membrane rapidly decreases from 71.2° to 4.2°, and the water droplets spread quickly on the membrane surface, indicating that the hydrophilicity of the membrane is significantly improved. Compared with the PI membrane, the pure water flux of the PI/BiOCl mixed matrix membrane is increased by 2 times, reaching 3269.2 L·m⁻2·h⁻1·bar⁻1. The mixed matrix membrane exhibits typical nanofiltration membrane characteristics with high rejections for CIP and dyes, but low rejection for different salts. Among them, the selective separation coefficient SNaCl/MB and SNaCl/CIP are 510.5 and 7, respectively, which demonstrates high application potential in the field of separation between antibiotics and salts. The prepared membranes exhibit high photocatalytic degradation efficiency for organic compounds such as dyes and antibiotics. The degradation rate of TC reaches 99.8%, and that of CIP reaches 96.4%. After 5 cycles of use, the membranes can still maintain high degradation efficiency, indicating their excellent cyclic stability with minimal catalyst loss. It was demonstrated that ‧O₂− and h⁺ play a major role and oxidize the intermediates during the photocatalytic degradation of TC. The study provides valuable reference ideas for the design and preparation of photocatalytic membranes with excellent separation efficiency to achieve the separation and degradation of antibiotics.

Acknowledgment

This research was funded by National Natural Science Foundation of China (No, 22366021, 22568025, 22268023), Jiangxi Province Double Thousand Talents Plan (No, jxsq2020101049), China Postdoctoral Science Foundation (No, 2021M691964), Ji’an Natural Science Foundation (No, 20255-061493).

CRediT authorship contribution statement

R. Han: Conceptualization, writing—review and editing, funding acquisition; F. Feng, G. Dong, Z. Zhu: Investigation; F. Feng: Writing—original draft preparation. All authors have read and agreed to the publication version of the manuscript.

Declaration of competing interest

There are no conflicts of interest.

Data availability

Data will be made available on request.

Declaration of generative AI and AI-assisted technologies in the writing process

The authors confirm that there was no use of artificial intelligence (AI)-assisted technology for assisting in the writing or editing of the manuscript and no images were manipulated using AI.

References

  1. , , , . Current state and new trends in the use of cellulose nanomaterials for wastewater treatment. Biomacromolecules. 2019;20:573-597. https://doi.org/10.1021/acs.biomac.8b00839
    [Google Scholar]
  2. , , , . Hybrid materials for heterogeneous photocatalytic degradation of antibiotics. Coordination Chemistry Reviews. 2019;395:63-85. https://doi.org/10.1016/j.ccr.2019.05.004
    [Google Scholar]
  3. , , , , , , . Cellulose/aminated multi-walled carbon nanotube nanocomposite aerogels for oil adsorption. Polymers. 2025;17:869. https://doi.org/10.3390/polym17070869
    [Google Scholar]
  4. , , . Easy preparation of porous boron nitride with excellent cycle stability for water purification. Desalination and Water Treatment. 2024;317:100024. https://doi.org/10.1016/j.dwt.2024.100024
    [Google Scholar]
  5. , , , , , . Selective solvent resistant polyimide composite membranes with efficient separation performance in antibiotics. Results in Engineering. 2026;29:108653. https://doi.org/10.1016/j.rineng.2025.108653
    [Google Scholar]
  6. , , , . A critical review on application of photocatalysis for toxicity reduction of real wastewaters. Journal of Cleaner Production. 2020;258:120694. https://doi.org/10.1016/j.jclepro.2020.120694
    [Google Scholar]
  7. , , . Enhancement of photocatalytic activity of TiO2 by immobilization on activated carbon for degradation of pharmaceuticals. Journal of Environmental Chemical Engineering. 2016;4:1929-1937. https://doi.org/10.1016/j.jece.2016.03.023
    [Google Scholar]
  8. , . Novel visible-light-responsive black-TiO2/CoTiO3 Z-scheme heterojunction photocatalyst with efficient photocatalytic performance for the degradation of different organic dyes and tetracycline. Journal of the Taiwan Institute of Chemical Engineers. 2021;121:168-183. https://doi.org/10.1016/j.jtice.2021.04.009
    [Google Scholar]
  9. , , . Impact of doped metals on urea-derived g-C3N4 for photocatalytic degradation of antibiotics: Structure, photoactivity and degradation mechanisms. Applied Catalysis B: Environmental. 2019;244:475-485. https://doi.org/10.1016/j.apcatb.2018.11.069
    [Google Scholar]
  10. , , , , . Enhancing the catalytic activity of g-C3N4 through Me doping (Me = Cu, Co and Fe) for selective sulfathiazole degradation via redox-based advanced oxidation process. Chemical Engineering Journal. 2017;323:260-269. https://doi.org/10.1016/j.cej.2017.04.107
    [Google Scholar]
  11. , , , , , , , , . One-step synthesis of CdS-CeO2/sepiolite and its efficient removal of Rhodamine B by synergistic adsorption and photocatalysis. Applied Clay Science. 2025;273:107831. https://doi.org/10.1016/j.clay.2025.107831
    [Google Scholar]
  12. , , , , . Study on highly efficient BiOCl/ZnO p-n heterojunction: Synthesis, characterization and visible-light-excited photocatalytic activity. Journal of Molecular Structure. 2019;1183:209-216. https://doi.org/10.1016/j.molstruc.2019.01.095
    [Google Scholar]
  13. , , , , , , . Rational synthesis of x-Bi5O7I/BiOBr microspheres for efficacious photo-induced degradation of tetracycline hydrochloride. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2025;723:137296. https://doi.org/10.1016/j.colsurfa.2025.137296
    [Google Scholar]
  14. , , , . Copper-doped BiOCl nanoflowers promote efficient formate production in electrochemical CO2 reduction reaction. Materials Science and Engineering: B. 2025;320:118402. https://doi.org/10.1016/j.mseb.2025.118402
    [Google Scholar]
  15. , , , , , , , , , , , , , , , , . Engineering the defect distribution via boron doping in amorphous TiO2 for robust photocatalytic NO removal. Applied Catalysis B: Environment and Energy. 2024;356:124239. https://doi.org/10.1016/j.apcatb.2024.124239
    [Google Scholar]
  16. , , , , . Two-dimensional Prussian blue analog-based catalytic membrane for effective decontamination of micropollutants. Water Research. 2025;283:123855. https://doi.org/10.1016/j.watres.2025.123855
    [Google Scholar]
  17. , , , , . Structurally stabilized defective ZIF-8 with open “windows” constructed by coordination node occupancy for enhanced CO2 separation. Separation and Purification Technology. 2025;360:131032. https://doi.org/10.1016/j.seppur.2024.131032
    [Google Scholar]
  18. , , . Facile preparation of loose P84 copolyimide/GO composite membrane with excellent selectivity and solvent resistance. Polymers. 2022;14:1353. https://doi.org/10.3390/polym14071353
    [Google Scholar]
  19. , , , , , . Bioinspired PDA/PVDF composite membrane with excellent antibiotic desalination properties. Emerging Materials Research. 2025;14:381-388. https://doi.org/10.1680/jemmr.25.00087
    [Google Scholar]
  20. , . Fabrication and modification of PVDF membrane by PDA@ZnO for enhancing hydrophilic and antifouling property. Arabian Journal of Chemistry. 2023;16:105206. https://doi.org/10.1016/j.arabjc.2023.105206
    [Google Scholar]
  21. , , , , , . Facile preparation of quaternary ammonium graft-modified PVDF nanofiltration membrane with excellent antibiotic desalination and antibacterial performance. Journal of Water Process Engineering. 2026;81:109284. https://doi.org/10.1016/j.jwpe.2025.109284
    [Google Scholar]
  22. , , , , , , , , . Recent advancement in emerging mxene-based photocatalytic membrane for revolutionizing wastewater treatment. Small (Weinheim an der Bergstrasse, Germany). 2024;20:e2311427. https://doi.org/10.1002/smll.202311427
    [Google Scholar]
  23. , , , , , , . MOF‐based photocatalytic membrane for water purification: A review. Small. 2024;20 https://doi.org/10.1002/smll.202305066
    [Google Scholar]
  24. , , , , , . Photocatalytic degradation of nonylphenol by immobilized TiO2 in dual layer hollow fibre membranes. Chemical Engineering Journal. 2015;269:255-261. https://doi.org/10.1016/j.cej.2015.01.114
    [Google Scholar]
  25. , , , , , , , . Innovative bi-functional cellulose acetate/BiOCl photocatalytic membrane with excellent antibiotics degradation and photoresponsive antimicrobial performance. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2026;730:138959. https://doi.org/10.1016/j.colsurfa.2025.138959
    [Google Scholar]
  26. , , , , , , , , . Construction of MXene-MOF membranes with photocatalytic self-cleaning for enhanced oil-water emulsion separation. Journal of Membrane Science. 2025;718:123685. https://doi.org/10.1016/j.memsci.2024.123685
    [Google Scholar]
  27. , . Thin-film composite membranes for organophilic nanofiltration based on photo-cross-linkable polyimide. Reactive and Functional Polymers. 2015;86:233-242. https://doi.org/10.1016/j.reactfunctpolym.2014.09.027
    [Google Scholar]
  28. , , , , , . Recent advances in polymeric solvent‐resistant nanofiltration membranes. Advances in Polymer Technology. 2014;33:1-24. https://doi.org/10.1002/adv.21455
    [Google Scholar]
  29. , , , , , , , . Cross-linked copolyimide-P84/CAU-10-H/branched polyethyleneimine membranes for organic solvent nanofiltration. Separation and Purification Technology. 2025;355:129663. https://doi.org/10.1016/j.seppur.2024.129663
    [Google Scholar]
  30. , , , . Facile preparation of high performance GO/Mn3O4/PVDF composite membranes with intercalation of manganese oxide nanowires. RSC Advances. 2023;13:19002-19010. https://doi.org/10.1039/d3ra02594b
    [Google Scholar]
  31. , , , , , , , . Design and performance investigation of novel efficient photocatalysts PVP-modified PVDF/BiOBr photocatalytic membranes for wastewater treatment. Chemical Engineering Journal. 2025;507:160781. https://doi.org/10.1016/j.cej.2025.160781
    [Google Scholar]
  32. , , , , , . Tailored nanofiltration membrane for enhanced antibiotic desalination by surface modification using branched quaternary triethanolamine assembly. Journal of Membrane Science. 2025;713:123364. https://doi.org/10.1016/j.memsci.2024.123364
    [Google Scholar]
  33. , , , . A dual strategy for synthesizing crystal plane/defect co-modified BiOCl microsphere and photodegradation mechanism insights. Journal of Colloid and Interface Science. 2022;617:73-83. https://doi.org/10.1016/j.jcis.2022.02.082
    [Google Scholar]
  34. , , , , , , , . Rapid synthesis of BiOCl graded microspheres with highly exposed (110) facets and oxygen vacancies at room temperature to enhance visible light photocatalytic activity. Catalysis Communications. 2019;130:105769. https://doi.org/10.1016/j.catcom.2019.105769
    [Google Scholar]
  35. , , , , , , , . Assembly of MXene/ZnO heterojunction onto electrospun poly(arylene ether nitrile) fibrous membrane for favorable oil/water separation with high permeability and synergetic antifouling performance. Journal of Membrane Science. 2022;663:120933. https://doi.org/10.1016/j.memsci.2022.120933
    [Google Scholar]
  36. , , , , , , , . Durable and super-hydrophilic/underwater super-oleophobic two-dimensional MXene composite lamellar membrane with photocatalytic self-cleaning property for efficient oil/water separation in harsh environments. Journal of Membrane Science. 2021;637:119627. https://doi.org/10.1016/j.memsci.2021.119627
    [Google Scholar]
  37. , , , , , . Self-assembled hierarchical heterogeneous MXene/COF membranes for efficient dye separations. Journal of Membrane Science. 2022;657:120667. https://doi.org/10.1016/j.memsci.2022.120667
    [Google Scholar]
  38. , , , , , , , . Facile synthesis of 2D TiO2@MXene composite membrane with enhanced separation and antifouling performance. Journal of Membrane Science. 2021;640:119854. https://doi.org/10.1016/j.memsci.2021.119854
    [Google Scholar]
  39. , , , , . Facile synthesis of PVDF photocatalytic membrane based on NCQDs/BiOBr/TiO2 heterojunction for effective removal of tetracycline. Materials Science and Engineering: B. 2021;265:114996. https://doi.org/10.1016/j.mseb.2020.114996
    [Google Scholar]

Fulltext Views
605

PDF downloads
865
View/Download PDF
Download Citations
BibTeX
RIS
Show Sections

Suggested read for related articles: