In the present photoelectrochemical (PEC) system utilizing MoS2 nanoflakes as a wide spectrum absorptive photoanode, simultaneous PEC degradation of different organic pollutants was achieved by employing in situ generated oxysulfur radicals, superoxide and hydroxyl radicals as strong oxidants. In order to better understand the cooperative PEC degradation of representative organic pollutants including rhodamine B dye (RhB), chlorpyrifos (CPF) and ciprofloxacin (CIP), the influences by bias potential, solution pH, radical scavenger, dissolved oxygen concentration and electrolyte concentration have been investigated. The selective PEC degradation efficiency follows the order of CPF > RhB > CIP in mixed substrates condition. In addition, the degradation rate for the single substrate degradation was about two times higher compared to that in mixed substrates degradation. The experimental results verified that reactive oxidation species (ROS) including oxysulfur radicals, superoxide and hydroxyl radicals can be efficiently produced on both anode and cathode under visible light irradiation, and they work together for simultaneous degradation of different pollutants, but the contribution of each ROS for pollutant degradation is substrate dependent. These results indicate that cooperative oxidation of multiple pollutants by miscellaneous oxygen-based radicals should be further considered as a promising advanced oxidation technique.
MoS2 nanoflakes on Ti sheet were prepared as a wide spectrum absorption anode.
Sulfite ions served as electrolyte, precursor of oxysulfur radicals and hole scavenger.
Multiple radical species can be produced by PEC activation of sulfite under visible light.
Cooperative degradation of organic pollutants in wastewater was achieved.
The contribution of each radical species for pollutant degradation was substrate-dependent.
Advanced oxidation processes (AOPs), as pretreatment approaches to increase biodegradability of wastewater or secondary treatment approaches after biological water treatment to further reduce the amount of organic pollutants, are increasingly employed for treatment of different types of organic wastewater. In addition, recalcitrant, high toxic, persistent and new emerging pollutants with low concentrations in wastewater, such as endocrine disrupting chemicals (EDCs), pharmaceuticals and personal care products (PPCPs) and persistent organic pollutants (POPs), are hardly removed by traditional physiochemical or biochemical approaches (Zhang et al. 2015; Huang et al. 2017). The AOPs utilize radical species (mainly hydroxyl radical, •OH) as strong oxidants for mineralization of different types of pollutants. Among the different types of AOPs (ozone oxidation, electro-catalytic oxidation, photo(electro)catalytic oxidation, Fenton oxidation, wet oxidation etc.), and photoelectrocatalytic (PEC) degradation of water pollutants has been broadly studied as one of the promising oxidation techniques (Pawar et al. 2015; Wu et al. 2015; Xia et al. 2017a, 2017b). However, whether in the photocatalytic or photoelectrocatalytic process, the search for abundant, stable and efficient photocatalytic materials is especially crucial from the perspective of practical application (Zhao et al. 2013; Ji et al. 2017; Liu et al. 2017a, 2017b).
On the other hand, even though hydroxyl radical has strong oxidation potential (2.8 V vs. SHE), it is not omnipotent since some kinds of water pollutants are resistant to its oxidation. For example, ammonia nitrogen is hardly completely removed by hydroxyl radicals (Deng & Ezyske 2011). Some research groups have reported conversion of ammonia nitrogen by photocatalysis, but most of the processes are less efficient (Wang et al. 2014; Bian et al. 2015; Dozzi et al. 2017; Liu et al. 2017a, 2017b). Our previous work also found that the molecular structures and properties of pollutants have a strong influence on the •OH mediated PEC degradation efficiency with the degradation rate order of chlorpyrifos > rhodamine B > 4–chlorophenol > ciprofloxacin (Zhou et al. 2019). Varanasi et al. (2018) also proposed that •OH selectively attack the aliphatic components, whereas Cl• induces transformation of olefinic species, and transformation of aromatic and olefinic moieties by SO4•− are the predominant pathway for removal of dissolved organic matter in UV–AOPs. Thus, the AOPs relying on hydroxyl radical alone are not sufficient for the handling of multiple recalcitrant pollutants. Besides the conventional •OH based AOPs, other radical species including superoxide radical (O2•−), chlorine radical (Cl•−), oxysulfur radicals (SO3•−, SO4•− and SO5•− radicals), persulfate radical (S2O8•−) and their combination mediated newly emerging AOPs have been paid more attention recently (He et al. 2014; Lutze et al. 2015; Remucal & Manley 2016; Jasper et al. 2017; Xia et al. 2017a, 2017b; Zhu et al. 2020; Ji et al. 2021). For instance, electrochemical and photoelectrochemical oxidation of ammonia by Cl•− have been reported as efficient approaches for complete removal of ammonia (Yang et al. 2016; Shin et al. 2017). SO4•− radical in company with Cl•− and •OH radicals-based water treatment has been studied (Lutze et al. 2015). In addition, an expanding range of sulfite, sulfate or persulfates activation methods have been established to generate oxysulfur radicals for degradation of organic pollutants (Farhat et al. 2015; Hu & Long 2016; Chen et al. 2017; Zhi et al. 2017).
To address the problems of AOPs as mentioned above, we attempt to explore a new PEC oxidation system based on multiple oxygen based radical species. Specifically, oxysulfur radicals are generated in situ at the photoanode, and O2 reduction generated radicals like O2•− and •OH are produced at the cathode. These oxidative radicals can realize simultaneous degradation of different types of wastewater pollutants. Herein, MoS2 nanoflakes grown on Ti sheet as a wide spectrum absorption photoanode was employed since it has a strong visible light absorption capability (band gap ∼1.5–1.8 eV) and suitable valence band position (1.3–1.5 V vs SHE) for oxidation of sulfite to generate sulfite radical (Liu et al. 2016). Upon the analytical, microscopic and spectroscopic characterization of prepared MoS2 electrodes, the degradation of three organic pollutants (chlorpyrifos as a representative of pesticide, rhodamine B as a representative of dye and ciprofloxacin as a representative of pharmaceutical) either in single substrate condition or mixed substrates condition was tested. These organic pollutants may coexist in some kinds of wastewater or sewage streams in different concentration levels, therefore the PEC degradation of three pollutants in mixed systems was studied. Furthermore, another goal of this work is to explore the molecular structure related degradation behavior, as the three pollutants have different molecular structures. Importantly, we have demonstrated that the in situ generated oxysulfur radicals, superoxide and hydroxyl radicals are key oxidants for the simultaneous degradation of different organic pollutants.
Preparation and characterization of MoS2 photoanode
The as-prepared MoS2 nanoflakes on Ti film (MoS2) and further annealed at 300 °C in reducing atmosphere (MoS2(300 °C)) were obtained by a hydrothermal method, and detailed information can be found in our previous work (Zhou et al. 2019). The morphology, microstructure and crystal phases of MoS2/Ti electrodes were characterized using a field emission scanning electron microscope (FESEM, SU8010), high resolution transmission electron microscope (HRTEM, Tecnai G2 F30), and X-ray diffraction (XRD, Bruker D8 Advance). Raman spectra of MoS2 samples was measured with 532 nm excitation (LabRAMHR–800). Surface chemical composition and valence state of the samples were analysed by X-ray photoelectron spectroscopy (XPS, PHI 5000 VersaProbe II) with a monochromatic Al Kα ray source. The band-gap energies of different MoS2/Ti films were measured by a UV-Vis spectrophotometer (JASCO V–7000).
Photoelectrochemical (PEC) performance evaluation
The PEC performance of different MoS2 samples was tested by an electrochemical workstation (CHI660E, Shanghai Chenhua). Typically, the photocurrents of the MoS2 photoanodes were recorded in 0.02 M Na2SO3 electrolyte and 0.5 V bias under visible light irradiation (≥420 nm) by a 300 W Xe arc lamp with a 420 nm filter (PLS–SXE300, light intensity 95.5 mW/cm2) according to our previous optimization study (Zhou et al. 2019). The PEC degradation of rhodamine B (RhB) dye (2.0 μmol/ L) and chlorpyrifos (CPF) and ciprofloxacin (CIP) (10.0 μmol/L) was carried out with applied bias of 0.3–0.5 V in 0.01–0.05 M Na2SO3 electrolyte in the pH range of 3–11 utilizing an MoS2(300 °C) electrode (electrode area = 1.0 cm2). Continuous and slow air bubbling into the reactor aimed to provide higher amounts of oxygen for enhancing generation of radicals on the cathode and degradation efficiency. The concentration of RhB, CPF and CIP at different intervals was analysed by a UV–Visible spectrophotometer (Shimadu UV–3600), gas chromatograph (GC, Fuli 9790) and high performance liquid chromatography (HPLC, Waters E2695). Detailed analytical information can be found in our previous work (Zhou et al. 2019). A series of radical trapping experiments were carried out by adding 5 vol.% methanol, tert-butyl alcohol (TBA) and 2,2,6,6-tetramethylpiperidine (TEMPO, 1 mM) as scavengers, respectively, in order to determine the role of each radical species.
RESULTS AND DISCUSSION
Physiochemical characterization of different MoS2 electrodes
Figures 1(a) and S1 display the morphologies and microstructures of as-prepared MoS2/Ti electrode (MoS2) from hydrothermal synthesis and further annealed electrode at 300 °C (MoS2(300 °C)). The vertical grown MoS2 nanoflakes with a thickness of ca. 4–10 nm standing on Ti substrate were obtained by hydrothermal synthesis as shown in the SEM images. However, there are more cracks on the further annealed MoS2 sample at 300 °C (Figure S1), probably because of the shrinkage and further growth of MoS2, whereas the thickness of MoS2 was not markedly varied. The length of vertically aligned MoS2 layer on Ti substrate is about 1.2–2.0 μm for MoS2 and MoS2(300 °C) from the SEM image cross-section view, as shown in the inset of Figure 1(a). In addition, the layered crystalline structure of MoS2 can be identified from the HR–TEM images and SAED patterns (see Figure S2). The XRD patterns of different MoS2 samples are shown in Figure 1(b). In reference to the standard MoS2 diffraction peaks (JCPDS card: 73–1508), broad diffraction peaks at 14.4 and 34.1° with relatively low intensities are clearly exhibited for the MoS2 sample. The MoS2(300 °C) sample presents a higher diffraction peak of the (002) plane at 14.4° due to the enhanced crystallinity of MoS2 nanoflakes after annealing at higher temperature.
Surface elemental XPS analysis of two prepared MoS2 samples along with a commercial bulk MoS2 sample (shown in Figure 1(c) and 1(d)) reflects some structural difference between the two prepared MoS2 samples. The two peaks centred at ca. 230 and 233 eV (Mo 3d5/2 and Mo 3d3/2) for the bulk MoS2 and MoS2(300 °C) samples, and appeared at relatively higher binding energies compared to those of the as-prepared MoS2 sample. Meanwhile, the S 2p doublet peaks for the bulk MoS2 and MoS2(300 °C) samples located at 163–164 eV, which is relatively higher than that of the as-prepared MoS2. It was reported that Mo and S peaks of MoS2 in 2H phase located at higher binding energy compared to 1 T phase MoS2, indicating that structural transformation from 1 T phase to stable 2H phase occurred during the post annealing treatment. The representative Raman peaks of different MoS2 samples are shown in Figure 1(e). The Raman peaks located at 383 and 409 cm−1 correspond to the in-plane (E2 g) and out-of-plane (A1 g) lattice vibrations. The Raman peaks of commercial bulk MoS2 are narrow and sharp, whereas those for the prepared MoS2 samples are broad, and there are slight Raman shifts for the prepared MoS2 nanoflakes samples compared to the bulk MoS2, probably due to edge effect and increased electronic transition energies in few-layered MoS2. The band-gap energies of prepared MoS2 film samples were evaluated by diffused reflection spectroscopy and the KM-plots of DRS spectra are shown in Figure 1(f). The band-gap energy of MoS2 is associated with transitions from p-orbitals of the S atom to d-orbitals of the Mo atom in the visible light region (400–800 nm), and depends on physicochemical characters such as layer thickness and defects etc. The band-gap energies of the MoS2 sample and MoS2(300 °C) sample are about 1.35 and 1.60 eV, respectively. The increased band-gap energy of the MoS2(300 °C) sample was probably due to the removal of intermediate defects, when the MoS2 nanoflakes were annealed at higher temperature. The suitable band-gap energies of MoS2 samples allow them to efficiently absorb solar light in a visible light region.
Figure 2(a) exhibits the photocurrent response (i−t curves) of different MoS2 films under visible light irradiation at biased potentials of 0.5 V in 0.02 M Na2SO3 electrolyte. The photocurrent of MoS2 was slightly reduced for the sample after annealing MoS2 (300 °C). It was also found that the generated photocurrents are slowly increased and decreased during the cycled light ‘on-off’ processes. This slow photoresponse may be attributed to several aspects such as weak interfacial connection between Ti substrate and MoS2 nanoflakes, intrinsic defects within MoS2, and fast recombination of photogenerated charge carriers as evidenced from the physiochemical characterization. Furthermore, in the continuous test condition (Figure 2(b)), the as-prepared MoS2 sample is extremely unstable, whereas the MoS2(300 °C) sample is relatively more stable for the generation of photocurrents due to the post thermal treatment. Thus, the MoS2(300 °C) sample was further selected for testing the PEC degradation of pollutants.
Three kinds of organic pollutants including rhodamine B dye (RhB), chlorpyrifos (CPF) and ciprofloxacin (CIP) (as shown in Table 1) were selected as representative pollutants in wastewater. Both a single substrate degradation experiment and mixed substrates degradation experiment were performed to investigate the selective degradation behavior of the MoS2(300 °C) sample. The experimental parameters such as pH, bias potential, and electrolyte concentration were first studied for the degradation of mixed substrates, and the results are shown in Figure 3. The PEC degradation of three substrates was more favored in acidic conditions, as the PEC degradation rates were gradually increased when decreasing the solution pH from 11 to 3. Reducing the bias potential from 0.5 to 0.3 V does not markedly reduce the PEC degradation of the three substrates. However, increasing the electrolyte concentration from 0.01 to 0.05 M gradually improves the PEC degradation rate. From an energy and cost consumption point of view, the optimized PEC degradation condition was screened out with 0.3 V bias, pH 3 and 0.02 M Na2SO3. Under such optimized conditions, substrate-dependent degradation behavior can be clearly observed. The PEC degradation efficiency follows the order of CPF > RhB > CIP, as shown in Figure 4(a). The PEC degradation rate constant of CPF was approximately five times higher than that of CIP (Figure 4(b) and Table 1), implying that CIP is a more chemically stable compound. In addition, a single substrate degradation experiment was performed and its degradation profile was compared with the mixed substrate condition (Figure 5). Generally, the degradation rate for the single substrate degradation was about two times higher compared to that of mixed substrates degradation (shown in Table 1). This is reasonable since there would be more radical species attacking pollutants in the single substrate condition. On the other hand, there would be more pollutants subjected to oxidation in the presence of mixed pollutants compared to that in the single pollutant conditions, because the recombination of generated oxidative radical species could be inhibited. In order to test the stability of the photoanode, multicycle experiments of RhB degradation were performed. As shown in Figure S3, after five cycle tests only slight PEC activity decay was observed, indicating the stability of the MoS2(300 °C) sample. The stable characters of the MoS2(300 °C) photoanode was attributed to the use of an Na2SO3 electrolyte that can consume photo-generated holes on MoS2 to prevent self-corrosion.
|Pollutant .||Structure .||Single pollutant|
|Rate k (min−1) .||R2 .||Rate k (min−1) .||R2 .|
|Rhodamine B (RhB)||0.0259||0.97||0.0147||0.95|
|Pollutant .||Structure .||Single pollutant|
|Rate k (min−1) .||R2 .||Rate k (min−1) .||R2 .|
|Rhodamine B (RhB)||0.0259||0.97||0.0147||0.95|
Mechanism of PEC degradation of pollutants
The generation of reactive oxidative species (ROS) including •OH, O2•−, and H2O2 on cathode and oxysulfur radicals (SO3•−, SO4•− and SO5•−) on the anode has been demonstrated in our previous work (Fan et al. 2019; Zhou et al. 2019). On the anode, the valence band potential (1.3–1.5 V vs. SHE) of MoS2 is not sufficient for oxidizing water to generate •OH radicals, thus the self–oxidation of S2− on the anode would be possible without hole scavengers. However, in the presence of sulfite electrolyte, the valence band holes generated on the nanostructured MoS2 film under visible light irradiation are able to oxidize sulfite to sulfite radicals, because the redox potential of SO3•−/SO32− couple is 1.2 V vs. SHE (Fan et al. 2019). Under such a setting, the self-oxidation of MoS2 electrode was significantly inhibited. On the other hand, the SO32− could also be electrochemically oxidized under biased conditions to produce SO3•−. On the cathode, both the photochemically and electrochemically generated electrons would reduce O2 to generate oxidative species like O2•−, H2O2 and •OH. To inhibit the side reactions of H2 production and enhance the generation of ROS, the reactor was continuously purged with air to feed higher amounts of oxygen. The oxygen effect on the degradation of pollutants can be seen in Figure 6. Without continuous air purging, only slight degradation of pollutants occurred for all the cases, probably because the dissolved oxygen was gradually depleted. This indicates that oxygen reduction-related ROS dominates the organics degradation.
To further prove the relative contributions of ROS taking part in the degradation of pollutants, quenching tests were carried out by adding different radical scavengers (methanol, tert-butyl alcohol (TBA), and TEMPO). TEMPO is a good scavenger of radical species, methanol with an α–H can effectively quench oxysulfur radicals (e.g. SO4•−, second-order rate constant 1.1 × 107 Lmol−1 s−1), •OH (second-order rate constant 9.1 × 108 Lmol−1 s−1) and holes, whereas TBA without an α–H mainly reacts with •OH (second-order rate constant 6 × 108 Lmol−1 s−1) and is not effective for oxysulfur radicals (e.g. SO4•−, 4 × 105 Lmol−1 s−1). According to the different scavenging effects, the key ROS to each pollutant can be inferred. As shown in Figure 7(a), the scavenging effect from high to low in the CPF degradation according to the reaction rate is TEMPO (for •OH, SOx•−, and O2•−) > methanol (for •OH, SOx•−) > TBA (for •OH). This result is consistent with the fact that oxygen reduction-related ROS played a critical role in the degradation of pollutants. The holes may be effectively scavenged by the sulfite electrolyte, thus the difference of methanol and TBA scavenging may be attributed to the SOx•− effect. For the RhB degradation test (Figure 7(b)), the TEMPO scavenging effect was excluded since its yellowish color has interference on the U-vis absorption of RhB solution. It can be seen that both methanol and TBA can effectively inhibit RhB degradation, indicating that •OH and SOx•− play a critical role in the RhB degradation. According to the scavenging effect, •OH could be the main ROS for RhB degradation. In the case of CIP degradation (Figure 7(c)), TEMPO and methanol present more significant scavenging effects compared to TBA, implying that SOx•− and O2•− radicals could be the key ROS for its degradation. Previously, we also found that •OH could not effectively degrade CIP, and Wang et al. (2018) reported that O2•− and photohole (h+) from the C3N4 photocatalyst are the main oxidants for CIP degradation. The scavenging experiments prove that all the ROS, including •OH, O2•− and SOx•−, participate in the organic degradation, as shown in Figure 8, but the contribution of each ROS is substrate-dependent. However, oxygen reduction-related ROS generally contributes most to the pollutant degradation. The different oxidation characteristics of each ROS determine that they will preferentially attack different sites of pollutant molecules, and the degradation pathways and intermediates would be varied, which merits further study.
In summary, an efficient MoS2–SO32− based AOP employing oxysulfur radicals, hydroxyl and superoxide radicals together for the degradation of different structured pollutants has been clearly demonstrated. The coexisting radicals allowed more efficient PEC degradation of multiple organic pollutants with different structures. The PEC degradation rates were affected by many parameters such as applied bias, solution pH, dissolved oxygen concentration, and properties of the pollutants. Generally, acidic pH conditions and higher electrolyte concentrations were more favored for pollutants degradation. The PEC degradation efficiency was also substrate-dependent, with the general degradation rate order of chlorpyrifos > rhodamine B > ciprofloxacin in both single and mixed substrates conditions.
The authors are grateful for the financial support from NSFC (no. 51978197) and the Shenzhen Science and Technology Innovation Commission (Grant No. JCYJ20190813171403664), Basic research program of Guangdong Province (No. 2018A030313851).
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.