1 Introduction
Various organic dyes are widely used in textile, painting, printing, pharmaceutical, rubber, cosmetics, and food industries [1–3]. Along with the development of these industries, the treatment of dyes wastewater is becoming increasingly difficult. Due to the toxicity, carcinogenicity, and mutagenicity of dyes, even a small amount can cause severe environmental damages and jeopardize human health [3,4]. Many methods, such as ion exchange [5], chemical coagulation/flocculation [6], chemical oxidation [7], membrane separation [8,9], adsorption [10], photocatalytic degradation [3,11] and electrochemical method [12], have been proposed and developed to remove dyes in wastewater. Among these methods, membrane separation, one of the most applicable and promising techniques to treat dye wastewater, is attractive due to its convenient operation, high efficiency and low energy requirement [13,14].
However, membrane fouling, resulting from the nonspecific interaction between organic dyes and the membrane surface [15], could cause decreased separation efficiency, cutting membrane life and increasing energy consumption [16,17]. Usually, hydrophilic modification may be an effective way to improve the antifouling property of membranes [18]. In recent years, photocatalytic membrane has received increasing attention, which combines photocatalyst and membrane for dealing with organic contaminants in wastewater [14,19–25]. Hydrophilic photocatalysts can mitigate the inherent membrane fouling owing to the hydrophilicity and the efficient photocatalytic decomposition of pollutants [16].
In the past decades, TiO2 has been widely used in the fabrication of photocatalytic membranes [13,16,22–24,26,27]. However, the wide band gap and low sunlight utilization efficiency have confined its practical applications [28–30]. Since 2008, Ag-AgX (X= Cl, Br, I) has been proven to photodegrade the organic pollutants of wastewater outstandingly owing to its superior visible-light activity [30,31]. However, it is very difficult to form a stable catalytic layer on the membrane surface. In order to obtain the best performance of catalysts, the nanoparticles anchored on the membrane surfaces should be exposed to the environment. Several strategies have been reported, including self-assembly [32], epitaxial growth [33], entrapment [34], and chemical binding [35]. Chitosan (CS), linear polymer, possesses outstanding hydrophilicity due to its hydroxyl and amino groups [36]. It can fix TiO2 nanoparticles on the membrane surface to take advantage of the large specific surface area of TiO2, which can facilitate catalysts to immobilize on the membrane surface.
Herein, for dye removal and photodegradation, we report a method to bind Ag-AgBr photocatalyst on the CS-TiO2 layer of PAN-ETA membranes (Scheme 1). First, CS modified TiO2 was firmly immobilized onto the PAN-ETA membrane surface by a simple vacuum filtration method. The CS-TiO2 layer is not only used as the dye separation layer, but also provides adhesion site for silver nitrate (AgNO3) [13,33]. Then, Ag-AgBr nanoparticles were formed, via in-situ partial reduction onto the CS-TiO2 layer, to improve the separation efficiency of various dyes. Furthermore, the PAN-ETA/CS-TiO2/Ag-AgBr membranes showed excellent degradation performance in dye solutions under visible light irradiation, which shows potential in the performances of anti-fouling, self-cleaning, and recyclable stability.
2 Experimental
2.1 Materials and chemicals
Polyacrylonitrile powder (PAN, molecular weight: ~150000) was obtained from Shanghai Jinshan Petroleum. Dimethyl sulfoxide (DMSO, AR, 99%) and absolute ethanol (AR, 99.7%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Ethanolamine (ETA, AR, 99%) was acquired from Shanghai Lingfeng Chemical Reagent Co., Ltd. Silver nitrate (AR, 99.8%), sodium bromide (AR, 99%), glutaraldehyde (GA, RG, 50 wt-% in water), CS and acetic acid (AR, 99.5%) were provided by Adamas Reagent Co., Ltd. Methyl blue (MB, AR), methyl orange (MO, 85%), congo red (CR,≥35%), rose bengal (RB, 95%), methylene blue (MEB,≥70%) and rhodamine B (RhB,≥97%) were purchased from Aladdin Chemical Co., Ltd., China. Titanium dioxide (5 nm) was purchased from Jing Rui New Material Co., Ltd. Deionized water was prepared using a two-stage reverse osmosis process. All chemicals were used as received without further purification.
2.2 Fabrication of PAN-ETA/CS-TiO2/Ag-AgBr membrane
2.2.1 Preparation of PAN-ETA membrane
PAN-ETA membrane was prepared by a modified in-situ polymerization/NIPS method according to our previous work [37]. In detail, 3.6 g of ETA (containing 1% water) and 10.8 g of PAN were added into 45.6 g of DMSO solution and the mixture was then stirred at 80 °C for 8 h. After cooling to room temperature and degassing, the solution was cast on a glass plate and immersed into deionized water to obtain PAN-ETA membranes.
2.2.2 Preparation of PAN-ETA/CS-TiO2 membrane
TiO2 (0.1 g) was dispersed in 2 wt-% acetic acid aqueous solution (100 mL) and the solution was ultrasonicated for 1 h, followed by the addition of CS (0.2 g) to form a CS-TiO2 dispersion. The CS-TiO2 coating layer was obtained through simple vacuum filtration of the above-dispersed solution and subsequently crosslinked in 1 wt-% GA aqueous solution to fabricate PAN-ETA/CS-TiO2 membrane.
2.2.3 Preparation of PAN-ETA/CS-TiO2/Ag-AgBr membrane
CS-TiO2 modified membrane was immersed into 50 mL of 1 wt-% AgNO3 solution under dark conditions for a designated time to ensure desorption/sorption equilibrium. After rinsing with deionized deionized water, the membrane was immersed into 0.5 wt-% NaBr solution to form AgBr nanoparticles. Then, the above membrane was submerged in deionized water/EtOH solution and irradiated by a 300 W Hg lamp for 10 min to partly decompose AgBr to Ag [30,38]. Finally, the PAN-ETA/CS-TiO2/Ag-AgBr membrane was stored in deionized water prior to use.
2.3 Characterization
The morphology of the prepared membranes was characterized by field emission scanning electron microscopy (FE-SEM, Nova Nano SEM 450, Japan). To obtain the cross-section, a dry membrane was immersed in liquid nitrogen and fractured, and the fractured surface was sputtered with a thin layer of gold prior to SEM analysis. The chemical components of the modified membranes were analysed by X-ray photoelectron spectroscopy (XPS, VG Microlab II, UK). Attenuated total internal reflectance Fourier transform infrared spectra (ATR-FTIR) of the membrane were recorded on Perkin-Elmer 16 PC FTIR. The contact angle measurement of the dry membrane was done by the sessile drop method using a drop shape analysis system (JC2000A, Shanghai Zhongchen Digital Equipment Co. Ltd., China). The dye concentration was measured by UV-vis spectrophotometer (Shimadzu, UV-2600).
2.4 Dye filtration performance
A self-made cross-flow cell (effective area ~28.26 cm2) was used to evaluate the pure water flux (PWF) (J) and dye rejection (R) of the as-prepared membranes. Four types of dyes, i.e., MB, MO, VB, and CR, were tested. The dye concentration was fixed at 100 mg·L−1, if it is not specially indicated. The membrane was pressurized with deionized water at 0.1 MPa for 30 min before the measurement. The permeate flux (J) was calculated by Eq. (1):
where Q is the volume of permeated water (L), A is the membrane area (m2), and t is the permeation time (h). Dye rejection (R) was calculated based on Eq. (2):
where Cf and Cp are the dye concentrations (mg·L−1) in the feed and permeate solution, respectively.
2.5 Photocatalysis performance
MEB and RhB were chosen to evaluate the photocatalytic activity of the as-prepared membrane due to its negligible photosensitizing effect under visible light. Typically, the PAN-ETA/CS-TiO2/Ag-AgBr membrane was cut into small sectors. In order to ensure the desorption/sorption equilibrium, the membrane was immersed into MEB or RhB solution (10 mg·L−1) for 30 min under a dark environment. The photocatalytic degradation of MEB or RhB was carried out under continuous stirring in a UV-visible irradiator equipped with a 500 W Xe lamp and a UV cut-off filter (l>420 nm). During photodegradation, 1 mL of the dye solution was gathered at 5 min intervals, followed by UV-vis spectrophotometry to measure the absorbance.
3 Results and discussion
3.1 Characterization of PAN-ETA and composite membranes
The surface morphologies of the as-prepared membranes are shown in Fig. 1. A relatively smooth surface could be clearly found on the PAN-ETA membrane (Fig. 1(a)) and a homogeneous distribution of nanoparticles on the surface demonstrated the successful loading of CS-TiO2 on the PAN-ETA membrane (Fig. 1(b)). The stable attachment of TiO2 is mainly attributed to the adhesion effect of CS intertwining networks on TiO2 nanoparticles. Furthermore, CS with abundant positively charged amine groups could be firmly immobilized onto the negatively charged surface (–COOH and –OH) of hydrolysed PAN-ETA membrane through electrostatic interaction. Figure 1(c) shows the surface morphology of a large number of nanoparticles decorating membrane. We found that small Ag nanoparticles were epitaxially grown on the surface of AgBr nanoparticles.
The XRD patterns and ATR-FTIR spectra of the as-prepared membranes are demonstrated in Fig. 2. In Fig. 2(a), the sharp peaks at 2q values of 17.3° confirmed the crystalline nature of PAN-ETA. Compared to PAN-ETA membrane, no characteristic peak for crystalline phase was found in PAN-ETA/CS-TiO2 due to the low exposure of TiO2 nanoparticles wrapped by CS. For the PAN-ETA/CS-TiO2/Ag-AgBr membrane, the diffraction peaks of Ag and AgBr simultaneously appears, indicating the formation of AgBr and Ag composite, which was in accordance with the SEM of PAN-ETA/CS-TiO2/Ag-AgBr [30,39]. Specifically, two peaks appear at 2q = 38.1° and 64.6°, ascribed to the face-centered cubic Ag (111) and (220) planes, respectively (JCPDS Card, No. 65-2871). Five strong diffraction peaks are also presented at 2q = 27.0°, 31.0°, 44.3°, 55.0° and 73.3°, which can be indexed to AgBr (111), (200), (220), (222), and (420) plane reflections (JCPDS card, No. 06-0438), respectively.
As shown in Fig. 2(b), two typical peaks, 2243 and 1733 cm–1, appeared for the PAN-ETA membrane, which correspond to the stretching vibration of cyano groups (C≡N) and carbonyl groups (C=O), respectively [40]. The absorption peaks at 2924 and 1452 cm–1 correspond to the stretching vibration and bending vibration of –CH2 groups. After filtration of CS-TiO2 suspension, the peak at 3353 cm–1, which is attributed to the stretching vibration of –NH2 and –OH, was seen for PAN-ETA/CS-TiO2. The stretching vibration of –NH2 occurs at 1405 cm–1, and the peak at 1565 cm–1 corresponding to the C=N group indicates that CS has been crosslinked with GA through the Schiff-Base reaction [41–43]. That would mean crosslinking networks can fix TiO2 tightly. The absorption peaks of C=O shifted to 1670 cm–1 and resulted in a significant enhancement, which suggests the –OH groups reacted with the aldehyde groups. Moreover, it was reported that the interaction between TiO2 and C=O groups would also give rise to a red shift for the C=O groups [13,44]. Compared to the CS-TiO2 decorated membrane, the absorption spectrum of the Ag-AgBr modified membrane showed no distinct difference.
XPS analysis was performed as shown in Fig. 3. Characteristic peaks of C 1s, N 1s and O 1s were detected in the XPS survey spectra of all membranes (Fig. 3(a)). After vacuum filtration of the CS-TiO2 suspension, an obvious signal of Ti 2p was found, which demonstrated the successful introduction of TiO2 on the membrane surface. For the PAN-ETA/CS-TiO2/Ag-AgBr membrane, the Ag 3d and Br 3d XPS spectra appeared, which illustrates that AgBr and Ag nanoparticles were spatially distributed on the membrane surface. To further determine the states of Ag, the Ag 3d XPS spectrum was measured as shown in Fig. 3(b). The Ag 3d5/2 and Ag 3d3/2 peaks can be fitted to two sets of peaks, in which the set of peaks at 367.7 and 373.7 eV is assigned to Ag+ in AgBr while that at 368.4 and 374.4 eV belongs to metallic Ag [30,45–47]. According to the peak areas in XPS, the relative contents of metallic Ag and Ag+ in the total Ag were calculated to be 9.72 at-% and 90.28 at-%, respectively, further illustrating the coexistence of Ag and AgBr in the PAN-ETA/CS-TiO2/Ag-AgBr membrane.
The water contact angle is adopted to evaluate the relative surface hydrophilicity of the membranes. Generally, a smaller contact angle displays better hydrophilicity [26]. As shown in Fig. 4, the pristine PAN-ETA membrane exhibits the highest contact angle of 76.5°, corresponding to the lowest surface hydrophilicity. After filtration of CS-TiO2 solution, the hydrophilicity of the PAN-ETA/CS-TiO2 membrane is significantly improved due to the amount of hydroxyl groups on the membrane surface. This result should also be ascribed to the embedded TiO2 nanoparticles on the membrane surface. The loading of Ag-AgBr nanoparticles had no distinct influence on the hydrophilicity of the PAN-ETA/CS-TiO2/Ag-AgBr membrane, and a slightly higher contact angle was obtained resulting from part of hydrophilic TiO2 being covered by Ag-AgBr nanoparticles.
3.2 Dye rejection performance of membrane
The dye rejections of the as-prepared membrane were tested and the results are summarized in Fig. 5. In view of the PAN-ETA membrane (Fig. 5(a)), the water flux reached ~120 L·m–2·h–1·bar–1 and the rejection of MO was comparatively low, only 29.12%. After filtration of CS solution, the water permeability reduced to 41.64 L·m–2·h–1·bar–1. Both MB and MO rejections increased from 76.25% to 98.75% and from 29.12% to 91.67%, respectively. The CS layer gives great resistance to water and dye molecules, and simultaneously, it causes decreased water flux and increased rejections of dyes. When the TiO2 nanoparticles were introduced, the water permeability was found to be enhanced to 54.50 L·m–2·h–1·bar–1, followed by slightly declined dye rejections. This may be explained by the improved membrane surface hydrophilicity and the loose structure with nano-voids after incorporation of CS and TiO2. In addition, the Ag-AgBr decorated membrane can retain high MB rejection of 97.41% and MO rejection of 88.22% similar to the PAN-ETA/CS-TiO2 membrane, yet a slight loss of water flux was examined due to the increased steric hindrance [13]. Five various dyes, i.e., MB, MO, VB, CR and RB, were tested to further evaluate the performance of PAN-ETA/CS-TiO2/Ag-AgBr membrane. As presented in Fig. 5(b), the rejections of negatively charged dyes (MO, CR, MB and RB) increased from 88.22% to 99.98% with the increase of molecular weight (Table S1, cf. Electronic Supplementary Material, ESM).
The MB (100 mg·L−1)/NaCl (100 mg·L−1) mixture solution was chosen to explore the stability of the PAN-ETA/CS-TiO2/Ag-AgBr membrane. Figure 6 shows the time-dependent water flux and rejection of MB and NaCl, which revealed that MB and NaCl rejections exhibited no obvious fluctuation while the water permeability decreased slightly (from ~ 47 to ~ 45 L·m–2·h–1·bar–1). This reduction was reasonable due to MB contamination. Despite a slight drop of flux, the composite membranes exhibited potential long-term stability for dye/salt separation application (dye rejection: ~97%; salt rejection: ~6.5%), benefiting from their excellent structural stability and anti-fouling ability. This further proved that the method to bind Ag-AgBr photocatalyst via the CS-TiO2 layer on PAN-ETA membrane is feasible.
3.3 Photocatalysis properties of PAN-ETA/CS-TiO2/Ag-AgBr
As presented in Fig. 7, RhB and MEB were chosen to evaluate the photodegradation performance of PAN-ETA/CS-TiO2/Ag-AgBr photocatalytic membrane under visible light irradiation. After irradiation with visible light for about 30 min, the MEB and RhB solutions with the membrane were degraded almost completely (Fig. 8), while the color of the dye solution without the membrane showed no significant change (Fig. S1, cf. ESM).
The photocatalytic adsorption-degradation efficiency (Fig. 8(a)) of RhB and MEB dyes was obtained with or without the PAN-ETA/CS-TiO2/Ag-AgBr membrane. In Fig. 8(b), the first-order kinetic equation was used to fit the kinetic of RhB and MEB degradation. The photocatalytic reaction rate constant (k) was 0.11924 and 0.08828 min–1 for RhB and MEB, respectively. It can be seen in Fig. 8(c) that the photodegradation of RhB and MEB reached 96.81% and 90.75%, respectively. After going through five cycles, small changes were observed in the photocatalytic performance (Fig. 8(d)), which may be partly attributed to the loss of photocatalyst during the washing process. Compared to other photocatalytic membranes (Table 1), the PAN-ETA/CS-TiO2/Ag-AgBr membrane shows a comparable degradation capacity for various dyes. These results are due to the existence of Ag-AgBr as a photocatalyst on the membranes.
Tab.1 Comparison of the photodegradation performance of different membranes |
| Membrane | Degradation condition a) | Irradiation condition | Degradation efficiency | Ref. | |
|---|---|---|---|---|---|
| Membrane area | Dye solution volume (concentration) | ||||
| PVDF@CuFe2O4 | 10 cm2 | 50 mL MB (100 mg/L) | Vis, 30 min | 99% | [16] |
| PDA/RGO/Ag3PO4/PVDF | 36 cm2 | 100 mL MEB (20 mg/L) | Vis, 480 min | 96.8% | [48] |
| PVDF/GO/ZnO | 9 cm2 | 60 mL MEB (10 mg/L) | Vis, 100 min | 86.8% | [49] |
| ZnO/PAN nanofiber | 50 mg | 10 mL MEB (10 mg/L) | Vis, 480 min | 96% | [50] |
| Ag/TiO2/PVDF | – | 30 mL MEB (10 mg/L) | Vis, 100 min | 51% | [51] |
| Ag/g-C3N4/PES | 20 cm2 | 50 mL MO (10 mg/L) | Vis, 100 min | 77% | [52] |
| PS/CCA/TiO2 | 25 cm2 | 100 mL RhB (10 mg/L) | Vis, 180 min | 82.4% | [53] |
| Pluronic-TiO2 | – | 12 mL RhB (100 μmol/L) | UV, 480 min | 90% | [54] |
| PAN-ETA/CS-TiO2/Ag-AgBr | 4 cm2 | 20 mL RhB (10 mg/L) 20 mL MEB (10 mg/L) | Vis, 30 min | 96.81% 90.75% | This work |
The degradation mechanism could be illustrated in Fig. 9. Under visible light irradiation, the electron-hole pairs are easily produced in Ag and AgBr nanoparticles. Then, the motivated electrons and holes transfer from Ag to the conduction band (CB) and the valence bands (VBs), respectively. Subsequently, the electrons on AgBr CB would be captured by O2 to produce •O2- radical, while the holes on AgBr VB would react with Br- and H2O to produce •Br and •OH radicals. The holes, •O2-, •Br, and •OH radicals degrade MB and RhB dyes into a colorless solution.
4 Conclusions
In this work, PAN-ETA and composite membranes were successfully prepared via facile processes. The introduction of CS and TiO2 improved membrane surface hydrophilicity, which enhanced dye rejection. In order to obtain anti-fouling and self-cleaning performance, PAN-ETA/CS-TiO2/Ag-AgBr photocatalytic membranes were fabricated by an in-situ partial reduction method. These membranes exhibited excellent dye rejection (MO: 88.22%; CR: 95%; MB: 97.41%; RB: 99.98%) and dye/salt mixture separation efficiency (dye rejection: ~97%; salt rejection: ~6.5%). Especially, the photocatalytic membranes could degrade dye solution under visible light irradiation, which shows potential in regeneration and reuse of the designed membranes.