ZnS-embedded porous carbon for peroxydisulfate activation: Enhanced electron transfer for bisphenol A degradation

Abstract

Transition metal sulfides have garnered increasing attention for their role in persulfate activation, a crucial process in environmental remediation. However, the function of metal sulfides without reversible valence changes, such as ZnS, remains largely unexplored in this context. Here we report ZnS-embedded porous carbon (ZnS-C), synthesized through the pyrolysis of Zn-MOF-74 and dibenzyl disulfide. ZnS-C demonstrates remarkable activity in activating peroxydisulfate (PDS) across a wide pH range, enabling the efficient mineralization removal of bisphenol A (BPA). Through electrochemical investigation and theoretical simulations of charge density distributions, we unveil that the electron transfer from BPA to PDS mediated by the ZnS-C catalyst governs the reaction. This study, both in theory and experiment, demonstrates metal sulfide as electron pump that enhances electron transfer efficiency in PDS activation. These findings redefine the role of metal sulfide catalysts, shedding new light on their potential for regulating reaction pathways in PDS activation processes.

Graphical abstract

Highlights

• •ZnS/carbon heterojunction is reported for persulfate activation reaction.
• •The electron transfer process is induced by ZnS sites.
• •Deep mineralization removal (78.9% in 2 h) of organic pollutants is achieved.
• •The degradation system has good anti-interference property for natural water.

1. Introduction

Persulfate (PS)-based advanced oxidation processes (AOPs) have shown promise in refractory organics removal [1,2]. Since the first report of persulfate-AOPs for environmental remediation in the late 1990s, extensive efforts have been devoted to improving the generation of hydroxyl radicals (^•OH) and sulfate radicals (SO₄^•−) during the PS activation process [3,4]. For instance, transition metal ions, metal oxides, and metal/carbon hybrids have been adopted to activate peroxymonosulfate (PMS) and peroxydisulfate (PDS) [[5], [6], [7], [8], [9]]. It has been reported that the reaction efficiency largely depended on the redox circulation of multi-valence metal species [5,10,11].

Recently, nonradical processes have also been recognized as an important reaction pathway in PS activation [12,13]. Specifically, singlet oxygen (¹O₂), one of the active species in a nonradical pathway, can be produced by catalysts even under acidic and neutral conditions [14,15]. More recently, electron transfer, another important non-radical reaction pathway, has drawn increasing attention due to its unique feature of high anti-interference property to background substances [13,16]. During the electron transfer process, the catalyst is hypothesized to serve as a conductive channel for the electron migration from the electron donor (pollutant) to the acceptor (reactive complex intermediate or catalyst-PS). This way, the organic substance is oxidized, and PS gains electrons to form SO₄²⁻ [12,13,17].

Carbon nanotube-based catalysts were initially found to proceed through the electron-transfer pathway in the PS activation process, and the oxygen groups and double vacancy defects were reported as the active sites [18,19]. However, the relationship between the intrinsic properties of the catalyst and the selective electron transfer pathway remains controversial, and regulating the reaction pathway to electron transfer is still a big challenge for the PS activation system [20].

Compared with conventional metal oxide catalysts, metal sulfides have drawn increasing attention for their high photosensitivity, large specific capacity, lower redox potential, and longer lifetime [21]. The large specific capacity results in high activity and efficiency in the adsorption and catalytic reaction. Metal sulfide catalysts typically have lower redox potentials, which means they can participate in redox reactions at relatively lower energy levels. Moreover, the S²⁻ can facilitate the redox circulation of multi-valence metals in the redox reaction. Metal sulfides, including FeS and CuS, have been explored for their potential in PS activation [11,22]. Their activation mechanism is similar to the transition metallic compounds with ^•OH and SO₄^•− as main radicals. In these systems, S²⁻ aids in the redox circulation of multi-valence metals, i.e., Fe(II)/Fe(III) and Cu(II)/Cu(III) [11,22]. As another typical metal sulfide, ZnS without reversible valence change of Zn(II) has yet to be studied in PDS activation. ZnS has high conductivity and catalytic activity in various catalysis processes, e.g., photocatalytic hydrogen production [23]. Thus, utilizing ZnS in PS activation may lead to different reaction pathways and new working mechanisms not typically observed with commonly reported metal sulfides.

This study reports a ZnS-embedded porous carbon composite (ZnS-C) by pyrolysis of Zn-based metal-organic framework (MOF-74) and dibenzyl disulfide. Interestingly, dibenzyl disulfide can stabilize the Zn element and form ZnS on the porous carbon rather than experiencing Zn evaporation at high pyrolysis temperatures. This ZnS-C catalyst shows highly improved PDS activation activity compared with S-doped or porous carbon. Moreover, this catalyst is effective across a wide pH range (3–9) and achieves a good mineralization efficiency (78.9% in 2 h) for the removal of bisphenol A (BPA). The mechanism study reveals that ZnS can act as an electron pump and tune the reaction pathway of PDS activation to electron transfer due to the optimization of the interface electronic structure of the catalyst. The density functional theory (DFT) calculation on the differential distributions of charge density on ZnS-C/PDS/BPA further reveals the electron transfer process from BPA to PDS through the ZnS-C catalyst. In line with the reported non-radical pathway, this ZnS-C/PDS system has been demonstrated to have high anti-interference properties for real water samples with rich background substances. This study, for the first time, reveals the role of a metal sulfide catalyst with no reversible valence change of its metal element in the reaction pathway regulation for PS activation. This discovery holds the potential to inspire the development of efficient and selective AOP systems for organic pollutant removal.

2. Experimental section

2.1. Pollutant degradation

The degradation experiments were performed in a 150 mL beaker containing 100 mL ultrapure water at ambient temperature (25 ± 5 °C). Before the reaction, a certain amount of pollutant was added, and the initial solution pH was adjusted by H₂SO₄ and NaOH solutions. Subsequently, the catalyst was introduced into the solution, and the reaction was started by adding a certain amount of PDS. During the reaction, samples were withdrawn at different intervals and filtered through a 0.45 μm polytetrafluoroethylene (PTFE) membrane filter to remove the catalyst. To completely stop the degradation reaction, 0.5 mL methanol was injected into the 0.5 mL filtered solution to quench the radical oxygen species (ROS) generated by PDS itself. The degradation experiments of pond water, river water, and tap water were conducted with the same experimental procedure and condition, except that real water solution was used instead of ultrapure water. All experiments were conducted in duplicate. Detailed information on the chemicals, preparation of catalysts, characterizations, and electrochemical tests were supplied in the Supplementary Materials, Text S1–S4.

2.2. Galvanic oxidation process (GOP) test

A GOP experiment was conducted in a two-cell electrochemical reactor to study the electro-transfer mechanism. To prepare the electrode, 5 mg ZnS-C-900 powder was added to a mixture solution (2.5 μL 5 wt% Nafion, 15 μL isopropanol, and 12.5 μL ultrapure water) and sonicated for 30 min. The electrode was obtained by depositing 100 μL of the sonicated suspension on one side of the conductive carbon paper electrode (2 × 2 cm) followed by 1 h drying. In the two-cell reactor, one cell was filled with BPA solution (50 mL, 10 mg L⁻¹ BPA, 0.2 M Na₂SO₄, marked as A), while the other cell was filled with electrolyte (50 mL, 0.2 M Na₂SO₄, marked as B). The reaction solutions in both cells were tuned to pH 7 by phosphate solution (5 mM). The proton exchange membrane connected the two cells during the GOP. The proton exchange membrane selectively allows protons (H⁺) to move while blocking the passage of electrons. This ensures that the electrochemical reactions in the anode and cathode compartments remain separate, preventing short-circuiting and ensuring controlled electron flow through an external circuit. The experiment was conducted by adding 0.2 mL PDS (500 mM) into cell B. For comparison, blank carbon paper electrodes were used and tested under the same experimental conditions.

3. Results and discussion

3.1. Physicochemical properties of ZnS-C

The catalyst synthesis process is schematically shows in Fig. 1. The Zn-MOF-74 was first synthesized according to the literature [24]. After thermal treatment at 900 °C, Zn-MOF-74 was transformed into pure carbon material (marked as C-900) due to the complete Zn evaporation [12]. Unexpectedly, ZnS could be formed on the carbon network (ZnS-C-900) through high-temperature annealing due to the sulphuration effect when dibenzyl disulfide (as the S source) was mixed with Zn-MOF-74. It is worth noting that ZnS has only one valence state of Zn(II) compared with multiple valence states of Fe, Cu, Mn, and Co. Recently, the electron transfer pathway was reported to be an important reaction mechanism in PDS activation [1]. As shown in Fig. 1, it is proposed that coupling conductive ZnS with a carbon catalyst may facilitate the electron transfer from organic pollutants to PDS absorbed on the catalyst, which can regulate the reaction pathway in the whole reaction process.

Fig. 1.

Schematic illustration of the ZnS-C synthesis process and the reaction pathway regulation in the PDS activation process.

A series of ZnS-C-T (annealing temperature T = 600, 700, 800, 900 °C) catalysts were prepared with different annealing temperatures. To study the crystalline structure of the prepared ZnS-C catalysts, the X-ray diffraction (XRD) test was first conducted. As shown in Fig. S1, the Zn-MOF was successfully synthesized, which matches well with the typical diffraction peaks of Zn-MOF-74 [25]. After thermal annealing of Zn-MOF-74 and benzyl disulfide, as shown in Fig. S2a, the prepared ZnS-C-900 shows typical diffraction peaks of ZnS (JCPDS No. 65–9585 and 36–1450) and no significant carbon peak is found. For the C-900 catalyst, the XRD patterns only show the peaks of (001), (002), and (100) crystal planes of carbon, indicating that no Zn-based composites were generated without dibenzyl disulfide [12]. For comparison, the S-C-900 sample was prepared by thermal treatment of C-900 mixed with dibenzyl disulfide. As revealed in Fig. S2a, the XRD patterns of S-C-900 are similar to those of C-900, with three peaks belonging to (001), (002), and (100) crystal planes of carbon.

Raman spectroscopy was used to study the structural characteristics of ZnS-C-T and C-900. As shown in Fig. S2b, two typical bands of carbon, i.e., D band and G band, are found in the Raman spectra of all catalysts. It is also found that the I_D/I_G ratio of ZnS-C-T increases with the increase of the calcination temperature. Since the D peak indicates the degree of disorder in the carbon structure, while the G band indicates the degree of graphitization of the carbon, the I_D/I_G value can be used to assess the defective degree. The results indicate that more defects, e.g., edge and topologic, may be generated in the ZnS-C-T through high-temperature pyrolysis. The I_D/I_G ratio of C-900 is lower than that of ZnS-C-900, which suggests that the sulphuration effect causes more defects in the carbonaceous catalyst [26].

The morphology and porous structure of Zn-MOF-74 and ZnS-C-900 were studied by scanning electron microscope (SEM) and transmission electron microscope (TEM) images. As shown in Fig. S3, the synthesized Zn-MOF-74 has a regular nanorod structure, consistent with the reported literature [24]. After sulphuration and thermal treatment, the ZnS-C-900 retains the nanorod structure (Fig. 2a). At the same time, in contrast to the smooth surface of Zn-MOF-74, many cracks appear in the nanorod structure of ZnS-C-900 (Fig. 2b and c). We speculate that these cracks are generated by releasing gases (e.g., CO₂ and Zn) during the high-temperature treatment. The corresponding energy-dispersive X-ray spectroscopy (EDS) mapping (Fig. 2d, e, g, h) shows that the nanorod structure is mainly composed of C, O, S, and Zn, and the four elements are uniformly distributed in the catalyst. This further confirms that Zn is fixed by the sulphuration effect to form ZnS in the ZnS-C-900 [[27], [28], [29]]. The high-resolution TEM (HRTEM) image in Fig. 2f shows that the lattice spacing (d) of 2.705 Å matches with the (200) diffraction plane of ZnS. A clear interface between ZnS and amorphous carbon reveals the presence of ZnS/C heterojunction. The selected-area electron diffraction (SAED) patterns (Fig. 2i) reveal the single-crystal ZnS structure [30].

Fig. 2.

a, SEM image of ZnS-C-900. b–i, TEM images (b–c), EDS mapping images (d, e, g, h), HRTEM image (f), and SAED patterns (i) of ZnS-C-900.

The structure characterization results confirm that ZnS-C-900 has a nanorod structure similar to that of Zn-MOF-74, and ZnS is formed and uniformly loaded on the rod-like carbon structure. The oxygen content in Zn-MOF-74 would promote gaseous products such as CO₂ formation. These reactions lead to cracks and porous structures forming in ZnS-C-900, which will benefit the mass transfer during the pollutant degradation reaction. The N₂ adsorption-desorption isotherms were obtained to study the porous structure to measure the specific surface area. As shown in Fig. S4, the Brunauer-Emmett-Teller (BET) specific surface area of ZnS-C-900 is 884.51 m² g⁻¹, and the BET specific surface area of C-900 is as high as 1530.11 m² g⁻¹. The high specific surface area and porous structure of ZnS-C-900 provide many accessible active sites for pollutant and oxidant adsorption and subsequent degradation removal.

3.2. BPA degradation

BPA was used as the model pollutant to study the removal performance of the ZnS-C/PDS system. As shown in Fig. 3a, the use of PDS in isolation and the Zn-MOF-74/PDS system yielded notably low rates of BPA removal. The C-900/PDS and S-C-900/PDS systems can remove 58.1% and 50.0% of BPA within 60 min, respectively. In contrast, the ZnS-C-900/PDS system shows the highest BPA removal efficiency (∼90.4%). Since ZnS-C-900 has a porous structure and a large specific surface area, its adsorption behavior was investigated. The BPA removal by ZnS-C-900 adsorption (without PDS) is around 60.4%, indicating that adsorption plays an important role in the removal process (Fig. S5a). To confirm whether the adsorbed BPA was further degraded on the catalyst surface, we performed desorption experiments on the used catalysts with ethanol. As shown in Fig. S5b, it is found that the extraction amount of BPA from the ZnS-C-900/PDS system is much lower than that from the adsorption reaction (ZnS-C-900 alone). This evidence confirms that the PDS activation-induced degradation rather than adsorption contributes to the complete BPA removal. The PDS utilization efficiency was further studied. As shown in Fig. S6, the PDS concentration was recorded during the reaction process in PDS alone, C-900/PDS, and ZnS-C-900/PDS systems. It is evident that the PDS utilization efficiency matches the trend of BPA removal, further confirming the superior catalytic activity of ZnS-C-900.

Fig. 3.

a, BPA removal performances by PDS alone, Zn-MOF-74/PDS, C-900/PDS, ZnS-C-900/PDS, and S-C-900/PDS systems. b, Comparison of the reaction rate constants in different reaction systems. c, TOC removal rates of BPA in PDS alone and ZnS-C-900/PDS systems. Reaction conditions: [BPA]₀ = 20 mg L⁻¹, [Catalyst]₀ = 100 mg L⁻¹, [PDS]₀ = 2.0 mM, initial pH₀ = 7.

To further investigate the possible origin of BPA adsorption on the catalyst, the Zeta potential and contact angle measurements were conducted. As depicted in Fig. S7a, when the solution pH exceeds 3.4, the catalyst surface becomes negatively charged, while when pH is below 3.4, the catalyst surface becomes positively charged. The pKa of BPA is between 9.78 and 10.39 [31], indicating that when the solution pH is below 9.78, BPA exists as a molecular form. As shown in Fig. S7b, the ZnS-C-900 possesses hydrophobic characteristics based on the contact angle measurement. Therefore, the adsorption of BPA onto the catalyst is not due to electrostatic attraction but rather attributed to hydrophobic adsorption.

Notably, the S-C-900 shows slightly lower BPA removal efficiency than C-900. This result suggests that S-doing in the carbon support does not improve the catalytic activity in PDS activation for BPA removal. Thus, compared with C-900 and S-C-900, the highest BPA removal efficiency of ZnS-C-900 is mainly due to the presence of ZnS. To compare the catalytic efficiency, the time-course data of BPA degradation were fitted by the pseudo-first-order kinetic model (Fig. S8). As shown in Fig. 3b, the reaction rate constant (k_obs) of PDS alone, Zn-MOF-74/PDS, S-C-900/PDS, C-900/PDS, and ZnS-C-900/PDS is 0.0015, 0.0015, 0.0145, 0.0226, and 0.0679 min⁻¹, respectively. The k_obs value of ZnS-C-900/PDS is 4.7 times higher than that of the C-900/PDS system, indicating the obvious enhancement of removal efficiency by ZnS.

The thermal annealing temperature also influences the catalyst activity. As shown in Fig. S9, with the annealing temperature ramping from 600 to 900 °C, the BPA removal rate increases from 36.8% to 90.4%. As shown in Fig. S10, the XRD patterns of the catalysts show that ZnO and ZnS coexist in the samples at an annealing temperature of 600–800 °C. When the temperature reaches 900 °C, the ZnO is completely sulphurized to ZnS. The ZnS-C-900 catalyst was digested, and the Zn content in ZnS-C-900 was measured as 21.7 wt% using inductively coupled plasma-optical emission spectrometry (ICP-OES). We then compared the performance of ZnS-C-900 with commercial ZnO and ZnS with the same Zn content. The results (Fig. S11) show that ZnO- and ZnS-based PDS systems have almost no degradation performance for BPA, indicating that ZnS alone cannot be used as a catalyst for PDS activation. In addition to the apparent removal rate, the mineralization rate of the ZnS-C-900/PDS system was also studied. As shown in Fig. 3c, PDS alone has almost no effect on the TOC removal, while the ZnS-C-900/PDS system shows a high TOC removal rate (78.9%) within 2 h. The TOC analysis indicates that this ZnS-C-900/PDS system has good TOC removal performance for BPA.

The optimization of reaction parameters is significant to the practical application of persulfate activation. Within this study, we have explored various reaction parameters, including the initial pH, the PDS dosage, and the catalyst dosage in this system. As shown in Fig. S12a, the initial pH has little effect on the removal rate of BPA. In a wide pH range from 3 to 9, the BPA removal rates are higher than 90%, confirming the good pH tolerance of the system. As shown in Fig. S12b, an increase in PDS dosage from 0.1 to 2 mM leads to a notable elevation in BPA removal, rising from 71.1% to 90.4%. However, when the dosage further increases to 4 mM, the removal rate drops to 85.6%. This phenomenon is attributed to the phenomenon of excess PDS quenching the active species [30]. For the catalyst dosage study, as shown in Fig. S12c, the removal rate of BPA increases with the increase of the catalyst dosage since higher catalyst dosage leads to more active sites for the BPA enrichment and degradation [32].

3.3. Reaction pathway

The inhibition experiments were first conducted to study the radicals generated in the ZnS-C-900/PDS system. As reported, methanol (MeOH) is the inhibitor of ^•OH and SO₄^•−, while tert-butanol (TBA) can efficiently quench ^•OH [33,34]. As shown in Fig. 4a, MeOH has no obvious inhibitory effect on BPA degradation, while TBA also shows a small inhibitory effect (the removal rate maintains at 83.4%). The higher inhibition effect from TBA than MeOH may be caused by the difference in hydrophilicity and hydrophobicity of the quenching agent. Considering that ZnS-C-900 has strong adsorption capacity to hydrophobic BPA, we suppose that the active species are generated on the catalyst surface and can react with pollutants before contacting the hydrophilic inhibitor [5]. To clarify this supposition, pure acetone instead of water was used as the solvent during the reaction since it could quench the direct contact between BPA and catalyst [35,36]. As shown in Fig. 4a and b, acetone can completely inhibit the degradation reaction, indicating that the reaction occurred on the heterogeneous surface of the ZnS-C-900. Therefore, hydrophilic MeOH shows a lower inhibition ratio than hydrophobic TBA. The very limited inhibition efficiency of MeOH and TBA suggests that the radicals, i.e., ^•OH and SO₄^•−, are not the main active species for the BPA degradation.

Fig. 4.

a–b, Quenching effects (a) and corresponding k_obs (b) of BPA degradation by different scavengers in ZnS-C-900/PDS system. c, The k_obs of BPA degradation by different mix orders. d, Current responses of ZnS-C-900 and C-900 electrodes after adding PDS and BPA. e, The open circuit potential of ZnS-C-900 electrode in different electrolytes. f–g, Tafel diagrams (f) and Nyquist diagrams (g) of ZnS-C-900 and C-900 electrodes. h, The current flow from the PDS cell to the BPA cell in the GOP system. i, The two-cell device setup. Reaction conditions: in panel a–c, [Catalysts]₀ = 100 mg L⁻¹, initial pH₀ = 7, [PDS]₀ = 2.0 mM, [BPA]₀ = 20 mg L⁻¹, [MeOH]₀ = 1000 mM, [TBA]₀ = 1000 mM; [Na₂SO₄] = 0.1 M; in panel d–e, [BPA]₀ = 10 mg L⁻¹; in panel h, one cell:10 mg L⁻¹ BPA, 0.2 M Na₂SO₄, the other cell: 2 mM PDS, 0.2 M Na₂SO₄, 5 mM phosphate solution at pH 7.

The electron paramagnetic resonance (EPR) tests were further conducted to identify the reactive species. With 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the spin-trapping agent, the reaction system with PDS alone (no catalyst) shows no signal, suggesting the weak oxidizing capacity of PDS alone (Fig. S13a). In the ZnS-C-900/PDS system, both DMPO-OH and DMPO-SO₄⁻ signals are observed, whose intensities increase over the reaction time, suggesting the existence of ^•OH and SO₄^•−. Although the EPR results indicate the presence of radicals, the non-radical pathways may dominate the oxidation of BPA in the ZnS-C-900/PDS system. In recent years, there have been increasing reports on non-radical pathways in PS activation by carbon-based catalysts [[37], [38], [39]]. This study employed the TEMP as a specific trapping agent to capture ¹O₂. As shown in Fig. S13b, the triplet signals of ¹O₂ with an intensity ratio of 1:1:1 are observed in the ZnS-C-900/PDS system, indicating the existence of ¹O₂ as well.

Since the degradation performance of ZnS-C-900 is higher than that of C-900, we further investigated the role of ZnS in the PDS activation process. At first, we compared the DMPO trapped and 2,2,6,6-tetramethylpiperidine (TEMP) trapped EPR signals in C-900/PDS and ZnS-C-900/PDS systems. The EPR signals for DMPO-OH, DMPO-SO₄⁻, and TEMPO in both C-900 and ZnS-C-900 systems exhibited striking similarities (Fig. 5). This observation implies that ZnS has no impact on the generation of ^•OH, SO₄^•−, and ¹O₂. Consequently, it suggests that ZnS may contribute to other reaction pathways in this system for enhanced activity.

Fig. 5.

a, DMPO trapped EPR spectra of C-900/PDS and ZnS-C-900/PDS systems. b, TEMP trapped EPR spectra of C-900/PDS and ZnS-C-900/PDS systems. Reaction conditions: [Catalysts]₀ = 100 mg L⁻¹, initial pH₀ = 7, [PDS]₀ = 2.0 mM.

More recently, electron transfer, as an important non-radical reaction pathway, has drawn increasing attention in the PDS activation process [13,16]. Unlike the generation of radicals and ¹O₂, the electron transfer pathway depends on the interaction between the organic pollutant and catalyst-PDS intermediates. In this system, the electron transfer happens when ZnS-C-90, PDS, and BPA come into contact. Therefore, the sequence of adding catalyst, PDS, and BPA could potentially influence the reaction rate if the electron transfer pathway dominates. To investigate this, we compared the observed rate constant (k_obs) by employing two distinct sequences of reagent addition: (1) BPA, ZnS-C-90, and PDS, and (2) ZnS-C-90, PDS, and BPA (Fig. 4c). The second sequence (premixing of ZnS-C-90 and PDS for 5 min and then adding BPA) resulted in higher k_obs compared with the first sequence. This is because the formation of catalyst-PDS intermediates is crucial for subsequent electron transfer, and the premixing step enhances the generation of intermediates. However, the k_obs decrease with a longer premixing time (60 min). This reduction is primarily due to the consumption of the generated radicals and ¹O₂ during the premixing period before the BPA addition. The adding order experiments confirm the presence of an electron transfer pathway and indicate a limited contribution from the radical pathway to BPA removal.

The electron transfer pathway was further investigated by electrochemical tests, i.e., the chronoamperometric response and open circuit potential [[40], [41], [42]]. As shown in Fig. 4d, C-900 and ZnS-C-900 electrodes were used as the working electrodes, and chronoamperometric responses were recorded after the addition of PDS and BPA. The current first decreases when PDS is injected into the electrolyte solution due to the formation of the catalyst-PDS complex with instant electron transfer from the ZnS-C-900 to PDS. After adding BPA, the current suddenly rises, indicating that electrons are transferred from BPA to the catalyst-PDS complex [43]. In contrast, the C-900 electrode shows no responses in the current after the addition of PDS and BPA. This result confirms the electron transfer reaction pathway between the BPA and ZnS-C-900-PDS complex during the PDS activation process, which is negligible in the C-900/PDS system.

To exclude the influence of the external electric field, we conducted additional testing to assess the open circuit potential of the ZnS-C-900 electrode during the PDS activation reaction. As shown in Fig. 4e, the open circuit potential of the ZnS-C-900 electrode is stable in the blank electrolyte and the mixture of electrolyte and BPA. However, when introduced into an electrolyte solution containing PDS, the open circuit potential of the electrode experiences a significant increase and stabilizes at around 0.69 V due to the forming of the ZnS-C-900-PDS complex. While with both PDS and BPA in the electrolyte, the oxidation potential also increases and eventually maintains at 0.57 V, caused by the oxidation of BPA [44,45]. The stabilized potential reveals that the lowest potential for BPA oxidation in this system is 0.57 V. Then, we used the electrochemistry workstation to supply a high potential to mimic the role of PDS (Fig. S14), and the current responses without PDS were obtained. The potential was set at 0.57 V (the open circuit potential of the ZnS-C-900-PDS-BPA complex) and 0.16 V (the open circuit potential of the ZnS-C-900-BPA complex). At 0.16 V, no current change is detected with or without BPA. However, when the potential is set at 0.57 V, a noticeable current enhancement is observed in the presence of BPA. The results suggest that when the potential reaches 0.57 V, the electron transfer is initiated through an established electron migration highway with the organic substrate as the electron provider [43].

Duan et al. reported that furfuryl alcohol (FFA) could inhibit the electron transfer route completely [1]. The FFA quenching experiment confirms that the FFA almost completely inhibits the BPA removal (Fig. 4a). Therefore, we propose that the BPA degradation is dominated by the electron transfer route in this ZnS-C-900/PDS system. Considering that FFA is also a ¹O₂ scavenger, we further tested the quenching effect of l-Histidine (typical ¹O₂ scavenger) on BPA degradation. Although l-Histidine can react with PDS and consequently affect BPA degradation, the inhibition of l-Histidine is still limited, indicating that the role of ¹O₂ is limited (Fig. S15). To study the potential involvement of dissolved oxygen in the generation of ¹O₂, BPA degradation experiments were conducted under conditions of agitation and N₂ bubbling. The results reveal that N₂ bubbling has a very small impact on BPA degradation (Fig. S16). This observation suggests that the dissolved oxygen does not hold a pivotal role in ¹O₂ generation.

In the ZnS-C-900/PDS system, the catalyst works as a conductive bridge to facilitate the electron transfer between organic pollutants and PDS that absorbs on the catalyst surface and regulates the reaction pathway to electron transfer. Thus, the electron transfer ability of the catalyst is crucial for the degradation efficiency. Based on the Tafel results in Fig. 4f, the corrosion current of the ZnS-C-900 electrode is 1.225 × 10⁻³ A, while the corrosion current of the C-900 electrode is 4.342 × 10⁻⁴ A. The higher corrosion current of ZnS-C-900 indicates a higher electron transfer ability than that of C-900 [44,45]. The electrochemical impedance (EIS) tests were also carried out. As shown in Fig. 4g, the Nyquist diagrams of the ZnS-C-900 electrode and C-900 electrode both show semicircular and linear parts, and the semicircle diameter represents the charge transfer resistance at the electrode interface. It is found that the semicircle radius of the ZnS-C-900 electrode is much smaller than that of the C-900 electrode, indicating lower charge transfer resistance [46]. These results further confirm that ZnS can enhance the electron transfer ability of the catalyst.

A galvanic battery with a two-cell configuration was constructed to study the direction of electron transfer, and the device setup is shown in Fig. 4i. The GOP experiment was conducted in a two-cell reactor with BPA and PDS solution. The proton exchange membrane connected the two cells during the GOP. The multimeter connects and records the current through the two cells. As shown in Fig. 4h, the device current is very small with blank carbon paper electrodes after adding PDS. In contrast, the device's current with ZnS-C-900 coated carbon paper electrodes increases instantaneously after PDS is added to one cell. The current firstly increases and then gradually reduces along the reaction. The current direction indicates that the electrons are transferred from BPA to PDS through ZnS-C-900. In summary, the electron transfer pathway is the main reaction pathway in this ZnS-C-900/PDS system with other minor pathways including ^•OH, SO₄^•−, and ¹O₂.

3.4. The role of ZnS in reaction pathway regulation

The study of active sites on the PDS activation catalyst is necessary to fully understand the reaction mechanism. The XPS spectrum is used to study the valence state of the elements in ZnS-C-900 before and after the reaction. As shown in Fig. 6a, the XPS C 1s spectrum could be fit into four peaks, in which the peak at 162.7, 163.7, 164.7, and 169.2 eV belongs to C(sp²), C(sp³), C(sp²)-S, and π-π∗, respectively [47]. Notably, the proportion of C(sp²)-S is significantly reduced, and C(sp³) increases correspondingly (Table S1). The S 2p spectra of ZnS-C-900 show peaks at 161.1, 162.1, 163.1, and 164.3 eV, corresponding to S(2p_1/2)-Zn, S(2p_3/2)-Zn, S(2p_1/2)-C and S(2p_3/2)-C, respectively, indicating the presence of ZnS and C-S (Fig. 6b) [48]. From the structure of Zn-MOF-74 and XRD patterns at different annealing temperatures, S atoms will replace O atoms, and ZnS forms a heterojunction structure through the Zn-S-C bond. After the reaction, the total percentage of C-S decreases from 43.3% to 41.3% (Table S1), indicating that the C-S bond is broken during the reaction process. The cleavage of C-S bonds might occur during the generation of oxidative species (minor pathway). The Zn 2p high-resolution XPS spectra before and after the reaction both reflect no change of valence states (Fig. 6c). A careful comparison reveals that the locations of Zn 2p peaks shift to lower binding energy after the reaction. As we know, the shifts result from valence and electron density changes, and the decrease in binding energy is due to the increased number of electrons [49]. Therefore, the binding energy shifts reflect that the ZnS active sites obtain electrons after the reaction, indicating the occurrence of the electron transfer.

Fig. 6.

a–c, The XPS C 1s (a), S 2p (b), and Zn 2p (c) spectra of ZnS-C-900 before and after the reaction. d, g, The differential charge density distribution for PDS adsorbed on C (d) and ZnS-C (g) models. e–f, The differential distribution of charge density of C/PDS/BPA from the top (e) and side (f) view. h–i, The differential distribution of charge density of ZnS-C/PDS/BPA from the top (h) and side (i) view. Yellow and blue regions represent electron accumulation and depletion moieties, respectively. The gray, yellow, red, blue, and white ball represents C, S, O, Zn, and H atoms, respectively.

The reusability of the catalyst is always a concern in heterogeneous catalysis systems. The SEM image of the catalyst after the reaction (Fig. S17) shows that the catalyst maintains its rod-like morphology during the reaction. The XRD patterns of the catalyst (Fig. S18) also indicate that the crystal structure of ZnS-C-900 remains unchanged after the reaction. Although there are no significant changes in morphology and crystalline structure, considering the slight decrease of C-S bonds (Table S1) and the adsorption of intermediate species on the catalyst, a re-sulfidation annealing was conducted to restore its activity. As shown in Fig. S19, the ZnS-C-900 catalyst demonstrates good cycling stability in three cycles.

The differential charge density distribution can directly reflect the interface charge conditions to confirm the electron transfer direction. Thus, DFT calculations on the differential distribution of charge density on the ZnS-C/PDS/BPA were conducted, and detailed calculation parameters and models (C and ZnS-C) are shown in Text S5 and Figs. S20–21. At first, we monitor the evolution of the peroxide O-O bond length (l_O-O) in different systems. The l_O-O of sole PDS is 1.222 Å, while the l_O-O of PDS adsorbed on C and ZnS-C is 1.310 and 4.089 Å, respectively. The longer the bond, the easier it is broken (higher activity). The peroxide O-O PDS on the ZnS-C-900 is largely stretched, indicating that the catalyst-PDS transition structure (ZnS-C-PDS∗) is generated. The DFT results reveal that the adsorption energy (E_ads) in C/PDS and ZnS-C/PDS is −6.3 and −11.1 eV, respectively. The more negative value reflects a stronger interaction between PDS and catalyst. Therefore, PDS is more active toward ZnS-C. Subsequently, the differential charge density distribution is monitored when BPA is added. The differential distribution of charge density of C/PDS/BPA and ZnS-C/PDS/BPA from top and side views are shown in Fig. 6d–i. The yellow and blue regions represent electron accumulation and depletion moieties, respectively. By comparing the two systems, we can observe that the electron cloud of C/PDS/BPA is smaller than that of ZnS-C/PDS/BPA, and more electrons accumulate around the ZnS-C/PDS complex. The ZnS/C heterojunction can optimize the interface electronic structure of the carbon network in the presence of BPA.

The partial density of states (PDOS) of ZnS-C/PDS and C/PDS clusters can provide valuable insights into the catalytic activity. The overlapping peaks in PDOS diagrams were employed to study the bonding between two closely located substances in space and the energy of bonding orbitals. As depicted in Fig. S22, strong overlapping peaks suggest a bonding between ZnS and PDS after ZnS doping. Compared to undoped ZnS (−5.1 eV), the lower energy (−6.8 eV) of the bonding orbitals reveals the enhanced adsorption of PDS. A comparison of the PDOS diagrams for PDS and C in ZnS-C/PDS and C/PDS clusters reveals that the peak shapes are identical, with the key distinction lying in the relative position to the Fermi energy level. After ZnS doping, the peaks of C shift leftward, indicating an increase in electron density within the ZnS-C/PDS system, thereby facilitating electron transfer. Additionally, the key difference between the transition metals and Zn on their reversibility is the occupancy of the 3d orbital. Compared with other transition metals, the d orbitals of Zn (ZnS-C) are fully occupied, making Zn the most stable in this state. On the other hand, the d-band of other transition metals that are not fully occupied can further lose electrons. Therefore, ZnS serves as an electron transfer agent, which aligns with the electron transfer mechanism of the ZnS-C-900/PDS system. The new results and discussions were added to the revised manuscript. The simulation results further elucidate the important role of ZnS as an electron pump in facilitating electron transfer and reaction pathway regulation.

3.5. BPA removal in natural waters

The coexisting Cl⁻, HCO₃⁻, and HA in water bodies may negatively impact the PS activation efficiency. Therefore, we further investigated the BPA removal performance of the ZnS-C-900/PDS system for real water samples. As shown in Fig. 7a, the anions and HA have almost no impact on the removal rate of BPA. The results align with the outstanding anti-interference characteristic of the electron transfer pathway to background substances [13,16]. The ZnS-C-900/PDS system also achieves similar BPA removal rates in ultrapure water, tap water, pond water, and lake water (Fig. 7b). However, the reaction rate constant reduces a little bit from 0.0516 min⁻¹ (ultrapure water) to 0.0351 and 0.0311 min⁻¹ in pond water and lake water, respectively (Fig. 7c). The water quality parameters of actual water samples are listed in Table S2. Notably, the initial pH value of river water, pond water, and tap water is 6.00, 6.17, and 6.40, respectively, below the neutral pH of 7. Comparatively, there is a slight delay in degradation observed in real water in contrast to that in ultrapure water. In real water, impurities, such as natural organic matter and NO₂–N, can react with radicals and reduce the reaction rate [50]. In line with the reported non-radical pathway, the ZnS-C/PDS system has been demonstrated to have high anti-interference properties for real water samples with rich background substances, indicating its potential for organic pollutants removal in various water bodies.

Fig. 7.

a, BPA removal rates in the presence of Cl⁻, HCO₃⁻, and HA. b, BPA removal rates in natural water samples with ZnS-C-900/PDS system. c, Comparison of the reaction rate constants at different systems. Reaction conditions: [BPA]₀ = 20 mg L⁻¹, [Catalyst]₀ = 100 mg L⁻¹, [PDS]₀ = 1.0 mM, [Cl⁻]₀ = 5.0 mM, [HCO₃⁻]₀ = 5.0 mM, [HA]₀ = 10 mg L⁻¹, initial pH₀ = 7.

4. Conclusion

Metal sulfide catalysts (e.g., FeS, CuS, and CoS) have high activity toward PDS activation. In this context, the valence cycle of metal ions assumes a pivotal role in the reaction process. Conversely, ZnS lacks reversible valence transitions, yet its notable conductivity enables efficient electron transfer between PDS and organic molecules during PDS activation. We are surprised to find that the enhanced degradation efficiency of BPA resulted from the presence of ZnS rather than the doping of sulfur on the porous carbon network. The optimized ZnS-C-900 shows an impressively high activity toward PDS in a wide pH range. This results in substantial mineralization of BPA while effectively mitigating interference from background substances. Through the innovative GOP measurement and the theoretical simulation of differential charge density distribution on ZnS-C/PDS/organic pollutants, we have confirmed the direction of electron transfer from organic molecules to PDS through the ZnS-C catalyst. This study represents the first experimental and theoretical exploration highlighting ZnS as an electron pump, thereby regulating the non-radical electron transfer pathway in the PDS activation process. The results reported herein offer a novel strategy in reaction pathway tuning in PS-AOP for promoted organic pollutant removal.

CRediT authorship contribution statement

Ying Liu: Conceptualization, Methodology, Investigation, Writing - Original Draft, Writing - Review & Editing. Ningjie Du: Validation, Investigation. Xinru Liu: Resources, Data Curation. Ducheng Yao: Software. Deli Wu: Methodology. Zhuo Li: Supervision. Shun Mao: Writing - Review & Editing, Supervision, Project Administration, Funding Acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is/are the supplementary data to this article.