Regulation of Thin-Layered g-C₃N₄ for Efficient Persulfate Photocatalysis of Ibuprofen Contaminated Groundwater

Abstract

The rapid and effective removal of pharmaceuticals and personal care products (PPCPs) from groundwater is challenging. In this paper, porous cyano group-rich g-C₃N₄ catalysts were prepared by urea (U) and a KOH-assisted thermal polymerization strategy. The thickness, active sites, and pores of g-C₃N₄ were successfully modulated by urea and KOH-assisted thermal polymerization. In addition, the charge separation efficiency of g-C₃N₄ was effectively improved by the above methods. We combine the g-C₃N₄ photocatalyst with peroxymonosulfate (PMS) to achieve the efficient degradation performance of ibuprofen. Meanwhile, we also explored the reaction mechanism of g-C₃N₄ in the photocatalytically coupled persulfate system, which illustrated the active roles of singlet oxygen and holes in the system in degrading pollutants. Our work demonstrates that the photocatalytically coupled persulfate system is an advanced technology necessary for the deep treatment of PPCPs in groundwater and suggests a feasible strategy for catalyst modulation.

1. Introduction

In recent years, commonly used pharmaceuticals and personal care products (PPCPs) have been widely detected in groundwater around the world. PPCPs are also recognized as emerging environmental contaminants, which typically spread with groundwater flow and pose a serious threat to the sustainable use of groundwater [1,2,3,4]. Ibuprofen, a commonly used nonsteroidal anti-inflammatory drug (NSAID), is now one of the most widely used over the counter PPCPs [5,6,7]. Ibuprofen is detected in groundwater, surface water, and plant effluents at concentrations of nanograms per liter (ng L⁻¹) or milligrams per liter (mg L⁻¹) frequently [8]. Ibuprofen has a complex chemical structure and chemical stability, so it can accumulate in ecosystems and its residual biotoxicity back severe effects. Therefore, residues of ibuprofen in the groundwater are of serious concern [9,10]. The development of rapid, cost-effective, and efficient methods to remove ibuprofen from groundwater will be effective in enhancing the sustainable use of groundwater. Recently, heterophase photocatalysis has been recognized as a possible potential technology for removing the presence of contaminants in groundwater, which can effectively degrade, transform, and convert the contaminants in water into low or non-toxic substances [11,12,13]. Among the many photocatalytic materials, graphitic carbon nitride (g-C₃N₄) has good chemical stability, wide visible light absorption range, molecular structure that can be adjusted, etc. [14,15,16]. However, even so, the pristine g-C₃N₄ still faces problems such as low separation efficiency of photogenerated carriers and low light utilization, which seriously limit the degradation effect and practical application of g-C₃N₄ [17,18,19].

To solve these problems, g-C₃N₄ has been improved through several strategies, which include elemental doping, construction of vacancies, modulation of catalyst morphology, etc. Zhiyi Lu prepared g-C₃N₄ (Cu-C₃N₄) with atomically dispersed Cu-N₄ sites by introducing Cu atoms into g-C₃N₄ and utilized the material to achieve the degradation of 10 ppm ciprofloxacin (CIP) in water within 30 min. The degradation of 10 ppm ciprofloxacin (CIP) in water was achieved within 30 min, and the degradation rate was higher than 99% [20]. Jiafeng Ding synthesized (Cu/Mo@CN) using Cu and Mo co-participation with g-C₃N₄ materials, and the tests and results showed that the doping of Cu and Mo could effectively improve the charge separation efficiency of the catalyst [21]. Jinjiang Zhu synthesized a g-C₃N₄ material (CN-K-V_N) with N vacancies using potassium citrate and found that introducing nitrogen vacancies could effectively enhance the separation of electron-hole pairs [22]. Zhang, H. et al. synthesized tubular P-C₃N₄ by doping P elements into g-C₃N₄ with morphology modification and found that tubular P-C₃N₄ had a larger specific gravity than g-C₃N₄. g-C₃N₄ has a larger specific surface area, which can provide more reactive sites during the reaction, and enhance the reaction by 30-fold in the degradation of tetracycline compared to bulk g-C₃N₄ [23]. Although these methods can effectively enhance the charge separation efficiency and photocatalytic degradation of organic pollutants by g-C₃N₄, the usual g-C₃N₄ materials can only rely on self-generated holes (h⁺) or generated superoxide radicals (•O₂⁻) for pollutant degradation due to the relative fixation of the energy band positions. However, the oxidizing ability of these oxidizing active substances is generally not strong, so the application of g-C₃N₄ photocatalytic materials is greatly limited.

In recent years, more studies have found that the activation of Peroxymonosulfate (PMS) can generate sulfate radicals (SO₄^•−) with strong oxidizing ability (E₀ = 2.5–3.1 eV, vs. NHE (Normal Hydrogen Electrode)). It has been extensively researched in the elimination of organic pollutants [24,25,26], while the sulfate radicals have a wider pH adaption range and longer half-life (t_1/2 = 30–40 μs for SO₄^•− vs. t_1/2 ≤ 1 μs for ^•OH) [27], resulting in more effective and sustained removal of pollutants. Although PMS is capable of generating SO₄^•− with strong oxidizing properties, the activation of PMS usually requires an external electronic or energy input, such as Fe²⁺ ions or UV light activation [28,29,30]. The photogenerated electrons generated by g-C₃N₄ during photoexcitation have a great potential to stimulate PMS, and PMS consumes the photogenerated electrons and promotes the release of photogenerated holes, which implies that coupling photocatalytic technology with PMS is a feasible strategy [31,32]. Nevertheless, few studies have been devoted to enhancing the activation ability of PMS by modulating the morphology and defects of g-C₃N₄.

In this work, we attempted to prepare defect-modified thin, porous g-C₃N₄ by facile urea (U) and KOH-assisted thermal polymerization method. These thin and porous structures provide many reaction sites for the reaction. We constructed a photocatalytic PMS co-degradation system using g-C₃N₄ with thin and porous layers, which effectively enhanced the activation effect of g-C₃N₄ for PMS. Finally, in the presence of co-catalyst Co, 10KCN further injected photogenerated carriers into Co to realize the efficient activation of PMS, indicating that the combination of the g-C₃N₄ photocatalyst and PMS is an effective method for removing organic pollutants from water. In conclusion, our work provides an effective method and strategy to protect the sustainability of groundwater.

2. Materials and Methods

2.1. Chemicals and Equipments

The drugs and reagents used in the experiment were commercial reagents, of which urea (AR, ≥99%, CH₄N₂S), melamine (AR, ≥99%, C₃H₆N₆), and methanol (GR, ≥99.9%, CH₃OH) were procured from Aladdin Company (Shanghai, China). Peroxymonosulfate (PMS) (AR, ≥42%, KHSO₅), ibuprofen (GC, ≥99%, C₁₃H₁₈O₂), EDTA (AR, ≥99%, C₁₀H₁₆N₂O₈), furfuryl alcohol (AR, ≥99%, C₄H₃OCH₂OH), and p-benzoquinone (AR, ≥99%, C₆H₄O₂) were procured from McLean Company (Shanghai, China). Potassium hydroxide was purchased from Best Company (Ningxia, China), and tert-butanol was purchased from Beijing Bailing Wei Technology Company (Beijing, China). The above drugs were used directly without secondary purification.

A muffle furnace (FURNACE 1200 °C) was used for catalyst preparation, and a diaphragm vacuum pump (GM-0.33A, Jinten company, Tianjin, China) and an electric blast drying oven (DZF-6020, Shanzhi company, Shanghai, China) were used to dry the catalysts. The Discover multi-channel photocatalytic reaction system (PCX50C, PerfectLight, Beijing, China) was used to provide the light source for photocatalytic reaction. electrochemical workstation (CHI 660E, Inc., Shanghai, China) for testing photocurrent and AC impedance.

2.2. Preparation of Catalysts

Firstly, 10 mg of KOH was dissolved in 20 mL of deionized water. Then, 0.75 g of urea (CO(NH₂)₂, U) and 2.25 g melamine (C₃H₆N₆, M) were dispersed into the above KOH solution and stirred for 30 min. The above solution was stored at 80 °C for 12 h to remove the water. The resulting powder was then transferred to a crucible and calcined at 520 °C and 500 °C for 2 h, respectively. At the end of the reaction, the solid was ground to powder to obtain bulk g-C₃N₄ (KCN-B/KOH). Then, we washed the g-C₃N₄ (KCN-B/KOH) three times with ionized water to remove the residual KOH, then we dried it at 60 °C, and named the resulting solid 10KCN-B. The obtained 10KCN-B was first laid flat in an alumina ceramic boat and then calcined at 500 °C for 2 h. Finally, the thin layer of mesoporous 10KCN was obtained. The KOH-free precursor samples were prepared in the same way and named M₁U₃-B, M₁U₃, (0.75 g U and 2.25 g M), and U-free samples were named 10KMCN-B (10 mg KOH and 2.25 g M). KOH and U free samples were named MCN (2.25 g M), All comparison samples were synthesized in the same steps as for the 10KCN material. Scheme 1 shows the synthesis of g-C₃N₄. The bulk sample (mKCN-B) was synthesized by calcining a mixture of M, U, and KOH. The resulting mKCN-B/KOH bulk sample was ground, rinsed with H₂O to remove the residual KOH, and finally, calcined in air at 500 °C for 2 h. The 10KCN with a thin mesoporous layer structure was obtained.

2.3. Test Methods and Equipment

The concentration of ibuprofen was measured by ultra-high performance liquid chromatography (DEAEQ53155, Agilent Technologies Inc., Santa Clara, CA, USA). The ratio of the flow phase was V_acetonitrile:V_water = 8:2, the volume of each sample was 20 μL, and the detection time was 10 min. The activation of PMS by the prepared catalysts was evaluated by degrading ibuprofen under ambient conditions and neutral pH (pH = 7). During the degradation process, we took samples at specific times and filtered the samples through PTFE filter tips. The concentration of ibuprofen in solution was subsequently determined by high-performance liquid chromatography. Finally, the catalytic activity of the catalyst was evaluated by comparing the concentration-time curves.

The transient photocurrent response properties and electrochemical impedance spectra (EIS) of the samples were characterized by an electrochemical workstation (CHI 660E, Inc., Shanghai, China). A standard three-electrode system was chosen for the experiments, including an Ag/AgCl reference electrode, a sample film working electrode, and a Pt sheet counter electrode, and all the tests were performed in a room-temperature condition in 0.1 M Na₂SO₄ solution was accomplished. A 300 W xenon lamp (PLS-SXE300/300UV, PerfectLight, Beijing, China) equipped with a UV filter (λ > 420 nm) was used as the light source.

The micro configurations of the synthesized samples were characterized by field emission scanning electron microscopy (FESEM, HITACHI, Tokyo, Japan) and high-resolution transmission electron microscopy (HRTEM, HITACHI, Japan), and the elemental compositions of the surfaces of the samples were analyzed using energy dispersive spectroscopy (EDX, HITACHI, Japan). The crystal structures of the synthesized samples were characterized and analyzed using an X-ray powder diffractometer with Cu Kα as the radiation source. Fourier transform infrared spectroscopy (FTIR) was tested by an FTIR spectrometer. Adsorption and desorption experiments of nitrogen from the samples were carried out by a Specific Surface and Porosity Analyzer, using the BET (Brunauer–Emmett–Teller) method and the Barrett–Joyner–Halenda (BJH) model used to analyze the specific surface area and porosity distribution of the samples, respectively. The elemental content of C and N in the samples was analyzed by an elemental analyzer (Elementar, Vario EL III, Elementar, Langenselbold, Germany). The valence states of the atoms on the surface of the samples were characterized by an X-ray photoelectron spectrometer (Thermo VG, ESCALAB 250, Thermo Fisher Scientific, Waltham, MA, USA, America with Al Kα as the irradiation source and the positions of the valence bands of the samples were characterized by XPS valence band spectra. The light absorption properties of the samples were tested and analyzed using a UV–visible spectrophotometer (TU1950, HITACHI, Japan), which utilized the integrating sphere method with BaSO₄ as a reference. The steady-state PL and TRPL spectra of the samples were tested using a time-resolved visible-UV spectrophotometer (Hitachi High-Tech Ltd., Tokyo, Japan).

3. Results and Discussion

3.1. Material Characterization

The morphology micromorphology and structure of the samples were tested using SEM and TEM. As shown in Figure 1a SEM image and Figure 1b TEM image, MCN and MCN-B are made of stacked g-C₃N₄ lamellae and show thick lamellar structures, whereas M₁U₃ and 10KCN after thiourea-assisted syntheses show ultrathin lamellar structures, and the morphological and structural differences between them are in sharp contrast. In addition, the morphology of the samples shows that the 10KCN and M₁U₃ materials are more loosely structured compared to the MCN and MCN-B. This also indicates that the addition of urea can effectively reduce the thickness of the samples so that the original dense layered bulk g-C₃N₄ can be transformed into loose ultra-thin g-C₃N₄ nanosheets. Secondly, a large number of mesoporous structures can be seen in the TEM image of 10KCN, which indicates that the addition of KOH during the synthesis process can effectively introduce mesoporous structures into the g-C₃N₄ structure. The addition of KOH can play an effective etching role to promote the formation of mesoporous structures. These mesoporous structures can on the one hand increase the scattering probability of absorbed light and improve the utilization of the material for light. At the same time, such structures are also conducive to the enhancement of the mass transfer process in the catalytic reaction process, so that the target reactive substances can migrate to the active site more quickly for the reaction. For the photocatalytic process, the thin-layer structure also enables the photogenerated carriers to migrate to the surface of the material faster, reduces the accumulation of photogenerated charges, decreases the carrier recombination rate, and increases the utilization rate of photogenerated carriers.

From the high-resolution XPS spectra, it can be seen that the C 1s XPS spectra of the samples can be categorized into three peaks: 284.8, 286.3, and 288.3 eV. The value 284.8 eV corresponds to the (C-C) peak, 286.3 eV corresponds to the C-NHx bond in the aromatic ring, and 288.3 eV corresponds to the Sp²-hybridized carbon in the N-containing aromatic ring (N-C=N) in g-C₃N₄ (Figure 2a). After the addition of KOH, the peak at the C-NHx bond (286.3 eV) was gradually enhanced, and the peak area ratio increased from 0.033 to 0.116, which indicated that the addition of KOH promoted the formation of cyano group (-C≡N). As shown in Figure 2b, the N 1s XPS spectra can be fitted to three peaks 398.7, 400.2, and 401.2 eV. The value 398.7 eV corresponds to the C-N=C of melon, 400.2 eV corresponds to the NC₃, and 401.2 eV corresponds to the NHx moiety. However, after the addition of KOH, we found that the peak position of NC₃ in 10KCN shifted to a lower binding energy. This is because the cyano group further reduces the electron cloud density of elemental N in NC₃, which makes the electron nucleus more attractive to the extranuclear electrons. Similarly, from the Fourier transform infrared spectroscopy (FTIR) it can be seen that all the bands corresponding to a typical g-C₃N₄ can be observed (Figure 2c), which indicates that the chemical structure of most of the triazine ring units remains unchanged even after different synthesis methods. The first peak at 810 cm⁻¹ corresponds to the out-of-plane bending of the heptazine ring, the strong band in the 900–1800 cm⁻¹ region corresponds to the absorption of the N-C=N heterocyclic ring, and the peaks in the 3000–3500 cm⁻¹ region are produced by N-H and O-H stretching vibrations. However, in MCN, M₁U_3, and 10KCN, it is evident that in 10KCN, to which KOH was added during synthesis, a new absorption peak at 2178 cm⁻¹ appears, which is caused by the asymmetric telescoping vibration of the cyano group’s C≡N. It can also be observed in the XRD pattern (Figure 2d) that the microcrystal structure of g-C₃N₄ has not changed significantly despite the introduction of new functional groups in g-C₃N₄. The weak peak at 13.2° (100) is associated with the in-face stacking of the tri-s-triazine unit. The strong peak at 27.4° is associated with the (200) facet of g-C₃N₄ which is due to the interplanar stacking of conjugated aromatic units. This indicates that the introduction of KOH successfully introduced the cyano group into g-C₃N_4, and did not destroy the structure of g-C₃N₄ due to the introduction of the cyano group.

The optical properties of the resulting samples were investigated by UV–Vis diffuse reflectance spectroscopy (DRS). As shown in Figure 3a,b. The light absorption band edges of the samples appeared to be gradually blueshifted with the increase in the proportion of U added in the secondary calcination and precursor mixtures. Usually, due to the quantum size effect, the smaller the size of the nanostructures the wider their bandgap will be, so the degree of exfoliation of g-C₃N₄ will directly determine the size of its bandgap. We calculated the forbidden bandwidths of MCN-B (2.43 eV), MCN (2.5 eV), M₁U₃ (2.72 eV), and 10KCN (2.49 eV) for the different materials from the Kubelka–Munk plots. Based on this, we further combined with the XPS valence band test to find the valence band position of the catalysts with CB = 1.81 eV (as shown in Figure 3c). In semiconductor theory, the conduction band (CB) position can be obtained from the forbidden bandwidth and the valence band (VB) position after calculation by E_CB = E_VB − Eg [33,34]. We finally obtained the energy band structures of all materials as shown in Figure 3d. This indicates that the secondary calcination and the auxiliary thermal polymerization process of adding U did not change the valence band position of the samples, and the change in the band gap of the samples was mainly caused by the change in the conduction band position. It can be seen that with the addition of urea and KOH in the precursor, the degree of stripping of the sample is gradually enhanced, and its conduction band is gradually moved upward, which leads to the gradual increase in the band gap of the sample.

Through the isothermal adsorption curve test, we found that the urea-assisted synthesized M₁U₃ and 10KCN have a larger surface area, while M₁U₃ and 10KCN have pore structures ranging from 10 to 100 nm. This indicates that the urea-assisted synthesis effectively stripped the original bulk g-C₃N₄ and formed g-C₃N₄ with a thin-layer structure and mesoporous structure (Figure 4a,b). By testing the AC impedance spectra of the materials, we found that M₁U₃ and 10KCN synthesized using urea-assisted synthesis have a smaller impedance. This indicates that reducing the thickness of g-C₃N₄ can effectively enhance the migration of photogenerated carriers in g-C₃N₄ and reduce the resistance of photogenerated carrier transport in the material. Secondly, from the steady-state fluorescence spectra, it can be seen that 10KCN and M₁U₃ possess smaller steady-state fluorescence intensities (Figure 4c), indicating that the thin-layer structure can effectively prolong the lifetimes of the photogenerated carriers and inhibit the recombination of photogenerated electrons and photogenerated holes. This may be a result of the thin-layer structure promoting the directional transport of electrons between layers. Meanwhile, we observed that 10KCN has a smaller AC impedance value than M₁U₃ and a weaker steady-state fluorescence intensity (Figure 4d), which may also be related to the cyano group. It is well known that the cyano group is a strong electron-absorbing group. When photogenerated electrons are generated, the built-in electric field will be formed between the cyano group and the triazine ring to promote the migration of carriers. Not only does it promote the carrier migration rate but avoids the compounding of electrons and holes to a certain extent. At the same time, the cyano group will gather a large number of electrons after the photogenerated charge is generated, which becomes the active center of the reduction reaction.

3.2. Activity of Photo-Catalytically Coupled PMS System

The degradation of ibuprofen was used to evaluate the activity of the photo-catalytically coupled PMS system. As shown in Figure 5a, M₁U₃ and 10KCN can effectively activate PMS for ibuprofen degradation during light irradiation. It can be seen that the combination of M₁U₃ and 10KCN with PMS enhances the degradation of ibuprofen effectively, the degradation rate can reach 50% and 80% within 2 h, respectively. On the other hand, MCN, MCN-B, and M₁U₃-B were not as effective as M₁U₃ and 10KCN in the degradation of ibuprofen. Meanwhile, although both M₁U₃ and 10KCN have enhanced PMS activation ability due to their thin layer and mesoporous structure, 10KCN is synthesized with the introduction of cyano group. This gives 10KCN a more pronounced electron-rich center of activity, and therefore, has a better effect on PMS activation than M₁U₃. It also has a greater degradation rate for the degradation of ibuprofen. In Figure 5b, we further determined the activation mechanism of 10KCN on PMS, and we can see that 10KCN has no activation effect on PMS under dark conditions. Only under light irradiation, 10KCN can effectively stimulate PMS and thus produce a better degradation effect on ibuprofen. This indicates that the PMS can only be activated effectively when 10KCN is excited by light and generates photogenerated carriers, which also indicates that the photocatalytically coupled PMS system is feasible for the degradation of pollutants in water. The mesoporous structure and cyano groups can effectively enhance the activation of PMS [35,36]. We can see from the degradation effect that when changing the KOH in the synthesis process can affect the performance of the catalyst in (Figure 5c). When the amount of KOH was increased from 1 mg to 10 mg, the activation of PMS by KCN was in an increasing trend, but the performance of the catalyst decreased after exceeding 10 mg. However, too much mesoporous structure and the number of cyano groups can also change the structure of g-C₃N₄, thus affecting the catalytic performance [37].

In addition, we explored the reaction mechanism of this system during degradation through species capture experiments. The addition of furfuryl alcohol to burst the singlet oxygen (¹O₂) and the addition of EDTA-2Na to burst the photogenerated holes (h⁺) play a major contribution in the catalytic degradation of ibuprofen in the original system. Since methanol bursts both sulfate radicals (SO₄^•−) and hydroxyl radicals (•OH) and tert-butanol bursts •OH [38,39]. As shown in Figure 5d, we can see that the degradation efficiency of ibuprofen was significantly reduced by the addition of furfuryl alcohol and EDTA-2Na in the species capture experiments. The catalytic degradation efficiency of ibuprofen was lower with the addition of methanol than with the addition of tert-butanol, i.e., in addition to the co-bursting of •OH, more were burst by methanol than by tert-butanol to contribute to the catalytic degradation, but the contributions of s •OH and SO₄^•− were smaller than that of h⁺ and ¹O₂. The catalytic degradation of ibuprofen in the system after the addition of p-benzoquinone showed a paradoxical and significant enhancement. This was attributed to the fact that although benzoquinone bursts the superoxide radicals (•O₂⁻) in the system, benzoquinone itself generates ¹O₂ with the PMS, and the degradation will be enhanced instead by the addition of benzoquinone if there are no •O₂⁻ in the PMS system or the content of •O₂⁻ is very low. The above results indicate that singlet oxygen and photogenerated cavities are the main degradation active species in this system [40], and hydroxyl and superoxide radicals as well as SO₄^•− act as auxiliary roles in the degradation process.

Based on the above work, we combined Co(II) with 10KCN material to effectively enhance the activation of PMS. From the degradation effect, we find that it can be seen that the activation ability of 10KCN on PMS was further enhanced after the addition of Co, and ibuprofen, which could be removed only in two hours, could be completely degraded in one hour. This suggests that the photogenerated electrons generated by 10KCN by light excitation can be effectively injected into Co(III) to realize the reduction in Co(III) to Co(II), and the generation of Co(II) can synergistically activate the PMS with 10KCN to generate active species for the degradation of organic pollutants (Figure 6).

4. Conclusions

In this paper, the thin porous-like g-C₃N₄ enriched with cyano group was prepared by the strategy of U and KOH-assisted thermal polymerization. We finally achieve the modulation of thickness, defect pores, and functional groups in the g-C₃N₄ structure. The efficient catalytic degradation efficiency of ibuprofen was achieved by coupling the g-C₃N₄ photocatalytic system with PMS. In addition to this, the reaction mechanism of g-C₃N₄ in the photo catalytically coupled persulfate system was also explored, and the activation of PMS was further enhanced by combining the 10KCN material with Co. In conclusion, our technology effectively removes ibuprofen present in water, realizing water purification and protecting the safety of water resources. Our work demonstrates that the photo catalytically coupled PMS system is advanced and necessary for the deep treatment of PPCPs in groundwater, offers a potential strategy for groundwater security and sustainability.