Enhancing Visible-Light Photodegradation of TC-HCl by Doping Phosphorus into Self-Sensitized Carbon Nitride Microspheres

Abstract

SSCN is a new type of self-sensitive photocatalyst. It consists of oxygenated carbon nitride-containing microspheres inside and polymerized triazine dye (TBO) formed on its surface by in situ polymerization. The presence of TBO endows SSCN with a wide range of optical responses. However, the TBO would self-degrade under light, making SSCN extremely unstable in photocatalytic reactions and limiting the practical application of SSCN. The introduction of phosphorus into the structure of SSCN significantly improved the electron–hole separation efficiency and reduced the self-degradation of surface TBO. Phosphorus-doped self-sensitive carbon nitride microspheres (P-SSCN) are easily synthesized by a one-pot solvothermal method—the phosphorus source was added to the precursor solution of SSCN. This resulting material was used for the photodegradation of tetracycline hydrochloride (TC-HCl) for the first time, giving improved visible light sensitivity and high stability in the photocatalytic process. This provides a new method for modifying self-sensitive carbon nitride carbon.

1. Introduction

Recently, water pollution from the overuse/abuse of antibiotics has aroused widespread concerns [1,2]. Photocatalysis directly utilizes solar energy to degrade antibiotic residues into harmless components in an aqueous solution, which has become a robust method to address antibiotic pollutants in water [3,4]. Studies have been conducted extensively on graphitic carbon nitride (g-C₃N₄), which is a non-toxic, metal-free, and economical photocatalyst with good performance. Generally, g-C₃N₄ is prepared by annealing the molecular precursors such as cyanamide, urea, and melamine at a high temperature of 400–600 °C [5,6]. Its photocatalytic activity is significantly constrained by the flaws in bulk g-C₃N₄, such as high carrier complexation efficiency and few active sites. Solvothermal synthesis has been put forward as an alternative approach for the preparation of g-C₃N₄, and it has been shown that it is an easy, practical, and efficient way to control the morphology and structure of g-C₃N₄ [7,8]. For example, Gu et al. [9] successfully synthesized an oxygen-containing carbon nitride microsphere as the core and the linked s-triazine oligomer (TBO) covalently as the shell by the solvothermal method. The obtained core/shell structure exhibits enhanced photocatalytic hydrogen production performance, in which the TBO shell acts as the sensitizer to harvest the visible light. However, the TBO shell may be self-degraded by the photogenerated carriers, which reduces the photocatalytic efficiency and stability. Several methods, such as hetero-element doping [10,11] and morphology modification [12,13], to enhance the photocatalytic performance of g-C₃N₄ and compound coupling have been used. Among these techniques, phosphorus doping has emerged as one of the most promising means of modifying the characteristics of g-C₃N₄. According to a study by Liu et al. [14,15], the doped P atoms into the C-N structural framework of g-C₃N₄ produced delocalized lone electrons. Such delocalized lone electrons can efficaciously promote the switch of photogenerated electron–hole pairs and decrease their recombination. After being doped with P atoms, Zhou et al. [16] discovered that the electronic characteristics of g-C₃N₄ had entirely changed. Due to the fact that P atoms were in the pentavalent state while C atoms were in the tetravalent state, zeta potential revealed that pure g-C₃N₄ was positively charged while P-doped g-C₃N₄ was negatively charged. The four pentavalent P atoms form covalent bonds with the surrounding N and the P atom is inserted into the C-N structural framework. The remaining free electrons of the P atom will thus dissociate from them, forming an electron-rich state. Hydroxyethylidene phosphonic acid (HEDP) and melamine mixtures were calcined at 500 °C in N₂ flow by Yuan et al. [17], and the resulting catalysts demonstrated improved photocatalytic H₂ evolution performance. The findings showed that adding P atoms to the g-C₃N₄ framework resulted in a delocalized-conjugated system, improving the system’s conductivity and ability to transmit electrons. The aforementioned findings suggest that adding P atoms to the g-C₃N₄ structure may modify the intrinsic electrical and band gap structure, hence enhancing the photocatalytic activity. From the energy point, a band gap impurity energy level is produced by the doping of P elements in the carbon nitride structure, and this impurity energy level narrows the band gap and increases the absorption of visible light [18]. One-step solvothermal procedures for the doping of P atoms into g-C₃N₄ are predicted to produce a delocalized lone electron that would result in an internal and external potential difference and enable a mutually attractive Coulomb force between the external TBO and the internal C-N structural framework.

Herein, we synthesized P-doped self-sensitive graphitic nitride microspheres (P-SSCN) by a one-step solvothermal method. Along with the C-N structural framework, the P atoms were doped into the TBO by replacing the C atoms under high pressure at 200 °C. Compared with unmodified SSCN, P-SSCN exhibited stronger photocatalytic stability in the degradation of TC-HCl. This work gives a feasible scheme to develop carbon nitride-containing photocatalysts with simple phosphorus doping for efficient solar-driven degradation of antibiotics with good stability.

2. Materials and Methods

2.1. Preparation of P-SSCN

When cyanuric chloride is exposed to air, some of it will be converted to cyanuric acid due to its strong tendency to hydrolyze. In light of the aforementioned hydrolysis, the molar ratio of cyanuric acid and chloride in the solvothermal process was set at 1.8 [9]. In a typical solvothermal process for x%P-SSCN (x% represents the mass ratio of solid phosphorous acid and cyanuric chloride), in 75 mL of acetonitrile, 0.75 g of cyanuric chloride and 15, 39, 83, and 132 mg of solid phosphorous acid were dissolved. The mixture was placed in a 100 mL sealed Teflon-lined autoclave (Shenghua, China) and heated to 200 °C and held for 20 h. The solution was removed and centrifuged to obtain a dark brown powder, after which washing was performed with anhydrous ethanol followed by deionized water. After that, the obtained powder was kept at 70 °C for 8 h drying using a vacuum oven. In the same conditions, pure SSCN without P doping was also produced without the addition of solid phosphoric acid. The TBO on the surface of the acquired SSCN and P-SSCN samples were removed for comparison by treating them with the Piranha solution [9]. As a result, the samples (designated SSCN-R and P-SSCN-R, respectively) were produced after being washed with anhydrous ethanol followed by deionized water. After that, the obtained powder was kept at 70 °C for 8 h drying using a vacuum oven.

2.2. Characterization

On a DX-2700 X-ray diffractometer (XRD), the phase structure of the prepared SSCN and x%P-SSCN (x = 2, 5, 10, 15, 20) samples was examined with Cu Kα (λ = 1.5406 Å) radiation at 2 theta range from 10 to 90°. On a Nexus 410 spectrometer, Fourier transform infrared (FTIR) measurements were taken (Nicolet, Green Bay, WI, USA). The AXIS Supra XPS equipment (Thermo Kalpha, Chelmsford, MA, USA) was used to run the X-ray photoelectron spectra with Al K radiation as the source. BaSO₄ was used as the reference material when recording UV/Vis diffuse reflectance spectra on the Lambda 750S UV/Vis/NIR spectrophotometer (PerkinElmer, Waltham, MA, USA). A JEM-2100F TEM, transmission electron microscope (TEM) (JEOL, Akishima, Japan) was used to analyze the morphological structure of the samples. The prepared samples were scanned for morphology using a scanning electron microscope (SEM) equipped with energy-dispersive X-ray spectroscopy (EDS) (Hitachi, Chiyoda City, Japan) operating at an accelerating voltage of 10 kV.

2.3. Photocatalytic Experiment

To investigate whether the prepared samples have good photocatalytic activity, an aqueous solution containing tetracycline hydrochloride (TC-HCl, 20 mg/L) was used as the target pollutant for photocatalytic degradation experiments. A 300 W xenon lamp with a 420 nm cut-off filter was used as the light source in the photocatalytic experiments. To ensure the establishment of an absorption–desorption equilibrium system, we added 10.0 mg of photocatalyst to a solution containing 100 mL of 20 mg/L TC-HCl before turning on the light. The suspension was then stirred for 60 min in a light-free environment. During the photodegradation test, 5 mL of the solution was collected every 10 min and centrifuged to remove any residual particles. The supernatant was filtered using a syringe to ensure that no impurities interfered with the absorbance measurements, and the intensity of the absorption peak at 356 nm for the TC-HCl solution was then measured using a UV-2600 spectrometer.

2.4. Photoelectrochemical Performance

The photocurrent measurements were characterized in a standard three-electrode electrochemical system using the prepared sample (1.4 cm × 2.8 cm) as the working electrode. The procedure for fabricating these working electrodes is shown below. First, the FTO conductive glass was cut into suitable sizes, washed and dried with water, ethanol, and isopropyl alcohol, in that order, and set aside. A 4 wt% catalyst dispersion was prepared by adding the sample to polyvinylidene fluoride and N-methyl pyrrolidone solvent with the help of ultrasonic waves. Then, the catalyst dispersion was loaded onto the FTO glass by the spin coating method, which was divided into two stages. First, a slow suspension coating at a speed of 500 rpm for 15 s, followed by a high-speed spin at 2000 rpm for 30 s. Finally, the spin-coated FTO conductive glass was dried at room temperature. A Pt sheet (1.0 cm × 1.0 cm) and an Ag/AgCl electrode were used as the counter and reference electrode, respectively. A 0.5 M Na₂SO₄ solution (50 mL) was used as the electrolyte. The photocurrent reaction was carried out by a 300 W Xe lamp and a filter (λ > 420 nm). The light/dark short-circuit photocurrent reactions were recorded on a CHI 660E electrochemical workstation with an open-circuit voltage as bias. The electrodes were impeded with a sinusoidal AC perturbation of 10 mV in the frequency range of 0.01–105 Hz. Mott–Schottky plots were generated from capacitance values derived from electrochemical impedance at 1000 Hz frequencies. All experiments were performed under the same environmental conditions.

3. Results and Discussion

3.1. Structure and Morphology

The crystal structure of the produced materials was investigated using X-ray diffraction (XRD), and Figure 1a shows the XRD patterns of SSCN and x%P-SSCN (x = 2, 5, 10, 15, 20). The broad and weak peak centered at 27.3° (d = 0.326 nm) corresponds to the long-range interplanar stacking of aromatic systems (002) planes of g-C₃N₄ [17,18,19,20]. However, the weaker diffraction peak around 13° from the in-plane tri-s-triazine motif is disappeared [21,22]. This is different from most of the reported g-C₃N₄, which indicates that the planar size of P-SSCN is smaller [22]. It is noteworthy that the diffraction peak broadens with increasing phosphorus content. The increase in phosphorus content implies that there are more defects in the structure. This may be due to the doping of phosphorus heteroatoms in the lattice of g-C₃N₄, which leads to greater lattice distortion and causes a decrease in crystallinity, which indicates that the crystal structure of SSCN microspheres is imperfect and disordered [23]. Since the atomic radius of phosphorus is larger than that of carbon atoms, the planar spacing of P-SSCN increases, making a shift in the diffraction peak.

Fourier transform infrared spectroscopy (FTIR) is used to validate a product’s polymeric structure. The spectra of SSCN and 15% P-SSCN do not significantly vary, as seen in Figure 1b. This suggests that they share comparable functional groups. The distinctive aromatic CN heterocycles make up the bands seen at 1637, 1402, and 1385 cm⁻¹ [24]. The N-H stretches and O-H are represented by two bands at 3423 and 3179 cm⁻¹, respectively, while the s-triazine ring modes are represented by the band at 802 cm⁻¹ [16,25]. Due to the low amount of P or the overlap with C-N vibration, the P-related group’s vibration is scarcely ever detected [26,27].

Additionally, the scanning electron microscope (SEM) images used to study the morphology of 15% P-SSCN are displayed in Figure 2. Almost every prepared particle is a well-defined microsphere with dispersed sheet-like layers on its surface. Statistical particle size analysis of 40 samples taken from the figure shows that the size of microspheres is uniformly distributed in the range of 0.8–1.2 μm (Figure S4). Additionally, the energy dispersive spectrometer (EDS) mappings of 15% P-SSCN in Figure 3 demonstrated that all the elements of C, N, Cl, and P exist and homogeneously spread on the surface of the sample, suggesting the successful preparation of P-SSCN.

The transmission electron microscope (TEM) images of 15% P-SSCN show the carbon nitride as spheres with diameters between 1 and 2 μm (Figure 4a,b), which is consistent with the SEM image. Many laminated materials were observed on the microsphere surface, which may be TBO in combination with the analysis of the SEM image. Additionally, we want to further observe TBO on the surface of the microsphere by high-resolution transmission electron microscope (HRTEM), but, unfortunately, only the boundary of the microsphere can be observed and the TBO may be destroyed under the electron beam with high energy. From this, it can be inferred that the lamellar material on the surface of the microsphere should be TBO.

Further detailed electronic structures and the chemical oxidation state of 15% P-SSCN were characterized by a high-resolution XPS probe technique, as shown in Figure 5. The investigated spectra showed that the prepared samples consisted mainly of C, N, O, Cl, and P elements, which corresponded well to the EDS analysis. The fine XPS spectrum of C 1s (Figure 5b) can be fitted into two peaks at 288.0 and 284.6 eV, which are attributed to sp²-bonded C in the unit of N-containing aromatic rings and the adventitious or contaminant carbon, respectively. In addition, the C 1s peak at 286.1 eV might belong to the C-Cl species [9]. The N 1s spectrum (Figure 5c) shows three peaks at 401.4, 400.3, and 398.8 eV, which correspond to N-H, tertiary nitrogen N-(C)₃ group, and sp²-hybridized nitrogen in triazine rings (C-N=C), respectively [9,24,28]. The O 1s signals (Figure 5d) show two peaks centered at 533.5 and 531.7 eV, which are assigned to C-OH and C=O [10], respectively [8,9,29]. The peaks of Cl 2p signals (Figure 5e) centered at 201.9 and 200.3 eV belong to Cl 2p1/2 and Cl 2p3/2 for the C-Cl species, respectively. Additionally, the two additional peaks, located at 198.6 and 197.0 eV, are caused by ionic Cl species [9]. P=N and P-N are responsible for the P 2p peaks at 133.9 and 133.0 eV, respectively. [15,16]. These results suggest that precursors ultimately polymerize and condense to form an oxygen-containing C-N structural framework made up of repeating tri-s-triazine units, and P atoms are successfully integrated into this framework [8,9].

To determine the distribution of P-doped sites and Cl, the surface TBO dye of 15% P-SSCN (denoted as 15% P-SSCN-R) was removed by Piranha treatment [8,9,30], and the corresponding survey scans are shown in Figure S2a. The Cl 2p and P 2p signals of 15% P-SSCN-R were weaker than those of 15% P-SSCN (Figure S2b,c). In addition, the EDS spectra by 15% P-SSCN and 15% P-SSCN-R (Figure S3) showed that the number of flakes covering the surface of the microspheres was greatly reduced, which echoed the previous SEM results and proved the presence and distribution of TBO. And when the surface TBO dye was treated with Piranha solution. The corresponding XRD patterns as well as FTIR patterns are shown in Figures S1 and S10. No significant changes in the characteristic peaks and functional groups of P-SSCN were found, which is consistent with the previous reports [8,9].

3.2. Optical Performance

The band structure of the produced samples was examined using UV–Vis and valence band XPS spectra. As shown in Figure 6a, all the samples have absorption in the visible range. According to the Tauc formula [31,32,33,34], the band gaps of the prepared samples (Figure 6b) were 1.19, 1.27, 1.16, and 1.28 eV for SSCN, 10% P-SSCN, 15% P-SSCN, and 20% P-SSCN, respectively. To further analyze the energy band structure of the materials, the XPS valence band test (Figure 6c) was used to calculate the 10%, 15%, and 20% of the P-SSCN valence band (VB) were calculated to be 1.19, 1.07, and 1.0 eV. The slopes of the curves, as shown by the Mott–Schottky test results (Figure S9), are all positive, indicating that P-SSCN is an n-type semiconductor. The conduction band positions of 10% P-SSCN, 15% P-SSCN, and 20% P-SSCN are calculated to be −0.26, −0.34, and −0.35 eV. Therefore, combining the above two test results, the band gap values of 10%, 15%, and 20% P-SSCN are 1.45, 1.41, and 1.35 eV, respectively. These values have approached the results obtained from the Tauc formula, and the band structure is shown in Figure 6d [35].

The transient photocurrent response was also used to evaluate the performance of the prepared SSCN, 10% P-SSCN, 15% P-SSCN, and 20% P-SSCN in terms of space charge separation. As shown in Figure 7a, the P-doped samples show a much higher transient photocurrent response than the undoped SSCN, and 15% P-SSCN exhibits the highest light response. Additionally, 15% P-SSCN has a lower radius than pure SSCN, according to electrochemical impedance spectroscopy (Figure 7b). This indicates that 15% P-SSCN has a larger capacity for charge transfer, which is advantageous for efficiently separating photogenerated electron/hole pairs. These findings imply that the photocatalytic degradation of P-doped materials is accelerated by carrier characteristics.

3.3. Photocatalytic Performance and Mechanism

To evaluate the photocatalytic activity of the photocatalyst for the decomposition of TC-HCl, a xenon lamp was used as visible light (λ > 420 nm), Figure 8a demonstrates that there was no appreciable drop in TC-HCl concentration in the control experiment, demonstrating this substance’s high degree of stability and inability to self-degrade in the presence of visible light. The SSCN can degrade TC-HCl; however, the photocatalytic performance is low and the recycling stability is poor (Figure S5). It is impressive to see how the degrading efficiency of TC-HCl grows with the amount of doped P, reaching a peak efficiency of 40% on 15% P-SSCN. We used first-order kinetic fitting curves to analyze the kinetic mechanism of photocatalytic degradation, as shown in Figure S11, and the fitting curves were linearly correlated, indicating that the photocatalytic degradation in this study followed first-order kinetic principles. The degradation reaction constants k were obtained by calculation, and the values of k for SSCN, 2% P-SSCN, 10% P-SSCN, 15% P-SSCN, and 20% P-SSCN were 0.00261, 0.004843, 0.00754, 0.00880, and 0.0054, respectively. Among them, the kinetic performance of 15% P-SSCN was the best, which was 3.37 times higher than that of SSCN. However, the efficiency dropped when the P doping level was raised to 20%, demonstrating a volcano-like relationship between the concentration of P doping and the effectiveness of the photocatalytic reaction. The high efficiency of 15% P-SSCN can be ascribed to improved charge separation efficiency and narrow band gap based on the above electrochemical characterization and band structure (Figure S7). It is worth mentioning that 15% P-SSCN-R without TBO dye displayed low photocatalytic degradation efficiency, which is similar to that of the undoped SSCN (Figure S6), demonstrating the vital role of the TBO sheets.

Furthermore, the photocatalytic process was studied using trapping methods [36,37]. The active species of OH, h⁺, and O₂ were quenched using tert-butanol (TBA), ethylenediaminetetraacetic acid disodium salt (EDTA-2Na), and 4-Hydroxy-TEMPO(Tempol), respectively. Figure 8b shows that TBA did not significantly reduce the photocatalytic effect of 15% P-SSCN on TC-HCl. However, the effect of photocatalysis was significantly inhibited by the presence of both EDTA-2Na and Tempol. We also performed further kinetic calculations of the degradation process for the capture experiments, as shown in Figure S13, and the fitted curves were linearly correlated, indicating that the capture experiments followed first-order kinetic principles. The degradation reaction constants k were calculated to be 0.0088, 0.00387, 0.00134, and 0.00807 for 15% P-SSCN, EDTA-2Na, TBA, and Tempol, respectively, where the kinetic constants of the photocatalytic reaction were significantly reduced by the addition of a superoxide radical trapping agent as well as a hole trapping agent. These findings suggest that O₂ and h⁺ are essential to photodegradation. But OH, has no impact on the photocatalytic process. We suggest a potential mechanism to account for the increased activity of P-SSCN in light of these findings. As shown in Scheme 1, the surface photosensitive dye TBO allows the material to generate electron–hole pairs due to its strong photo-responsive properties. The photo-indued electrons move to the surface of P-SSCN for redox reaction with the O₂ molecules adsorbed on the surface, forming O₂⁻ with a strong oxidation ability to degrade the TC-HCl. The 15% P-SSCN has a reduced band gap value compared to SSCN, which is mainly due to the doping of P elements. In addition, the introduction of P atoms leads to the introduction of an impurity energy level between VB and CB of SSCN, which reduces the original band gap value of SSCN [38]. Meanwhile, due to the substitution of C atoms by the doped P atoms, off-domain electrons are generated, and the off-domain effect can effectively improve the ability to separate electron–hole pairs and increase carrier mobility [14,15]. In summary, the doping of P elements in the structure of SSCN can not only reduce the forbidden band width of SSCN but also have the dual effect of improving the electron transfer efficiency and reducing the possibility of electron-hole recombination [39].

3.4. Cyclic Stability

To evaluate the photocatalytic stability of 15% P-SSCN, we mezasured the photocatalytic stability using a cycling test to observe the change in the degradation performance of the photocatalyst against TC-HCl under xenon light irradiation for 5 cycles (λ > 420 nm). The photocatalytic efficiency does not significantly decline throughout the five cycles, as seen in Figure 9, while the catalytic efficiency somewhat declines with time but increases in the 5th cycle. It is speculated that the TBO on the surface of the sample was partially degraded with the cycle. It is reasonable to speculate that the optimal TBO content may have been reached with the continuation of the cycle test, thus achieving an increase in photocatalytic efficiency in the 5th cycle. The kinetic constants k values for the five cycles were 0.00738, 0.00802, 0.00671, 0.00568, and 0.00709, respectively. These results indicate that P-SSCN has good photocatalytic cycling stability and exhibits similar degradation rates in all five cycles (Figure S12). The samples were retrieved after the cycle test for a second XRD and FTIR analysis, and the findings of the experiments showed that neither the physical phase structure nor the elemental analysis of P-SSCN changed considerably after the reaction (Figure S8).

Compared with the SSCN (Figure S5), the 15% P-SSCN revealed an enhanced recycle stability, which might be ascribed to the doping of the P atom. The doping P makes the attraction between the surface TBO and the internal CN backbone, which reduces the self-degradation of the surface TBO dye during the photocatalytic process thus further improving the photocatalytic efficiency.

4. Conclusions

In conclusion, we have successfully synthesized P-doped self-sensitive carbon nitride microspheres (P-SSCN) by an easy one-pot solvothermal method—adding a phosphorus source to the SSCN precursor solution and heating it at a low temperature. The phosphorus atoms are uniformly introduced into the carbon nitride skeleton of SSCN as well as in the TBO formed in situ on its surface. Photoelectrochemical tests as well as cyclic stability tests confirm that phosphorus doping enhances the separation efficiency of photogenerated electron–hole as well as catalytic stability. This work provides a new method to modify self-sensitized carbon nitride. The obtained P-SSCN may be an efficient photocatalyst for the degradation of organic pollutants in the field of water environmental protection and engineering.