Construction of Sulfur-Doped and Cyanide-Modified Carbon Nitride Photocatalysts with High Photocatalytic Hydrogen Production and Organic Pollutant Degradation

Abstract

Element doping and functional group modification engineering serve as efficient approaches that contribute to the improvement of the functional efficiency in graphitic carbon nitride (CN) materials. A CN photocatalyst co-modified with sulfur (S) and cyano moieties was prepared through thermal condensation polymerization. The introduced S species modulated the band structure, increased charge carrier mobility, and significantly promoted charge separation and transport. Additionally, the introduction of cyano groups extended light absorption range and improved the material’s selective adsorption of reactant molecules. The as-prepared sulfur-modified CN photocatalyst obtained after a 6 h thermal treatment, which was capable of degrading organic pollutants and producing hydrogen (H₂) efficiently and stably, exhibited excellent catalytic performance. The photocatalyst’s photocatalyst exhibited a significantly enhanced photocatalytic activity, with a Rhodamine B (RhB) removal efficiency reaching 97.3%. Meanwhile, the H₂ production level reached 1221.47 μmol h⁻¹g⁻¹. Based on four-cycle experiments, the photocatalyst exhibited excellent recyclability and stability in both H₂ production processes and photocatalytic organic pollutant degradation. In addition, mechanistic studies confirmed the dominant role of ·OH and ·O₂⁻ as active species responsible for the reaction system’s performance. This study highlights that the co-decoration of heteroatoms and functional groups can markedly enhance the photocatalytic performance of CN-based materials, offering considerable potential for future applications in energy conversion and environmental remediation.

1. Introduction

The rapid expansion of the global economy and the steady rise in population have intensified the conflict between rising energy demands and the need for environmental protection [1,2]. Finding a sustainable energy supply method and an effective means of environmental pollution control has become the focus of global attention [3]. Photocatalytic processes harness solar energy to drive reactions like hydrogen (H₂) production via water splitting and the breakdown of organic contaminants [4,5,6]. Effectively coupling pollutant degradation with H₂ production requires the deliberate construction of high-performance photocatalysts [7,8]. Among various photocatalytic materials, graphitic carbon nitride (CN) is particularly appealing due to its desirable combination of economic viability, appropriate energy band gap, ease of synthesis, and superior chemical stability [9,10].

Although promising, CN demonstrates limited light-harvesting efficiency and rapid charge recombination, hindering practical applications [11,12]. To address these limitations, the photocatalytic capability of CN can be significantly improved via multiple modification routes, including morphology engineering, heterojunction, homojunction, metal/non-metal element doping, etc. [13,14,15]. Heteroatom incorporation into CN frameworks represents a particularly efficient modification strategy, simultaneously optimizing electronic band structure, suppressing charge recombination, and extending visible-light response [16,17,18]. Notably, sulfur (S) stands out among dopant elements due to its relatively large atomic radius, low electronegativity, high electron density, and its ability to form diverse covalent bonding configurations [19,20]. In a notable contribution, Luo’s team synthesized thin, S-modified CN sheets with controlled porosity using precursor-directed thermal condensation, where thiocyanuric acid served as both S donor and structure-directing agent. A hydrogen production rate of 6225.4 μmol g⁻¹h⁻¹ was obtained [21]. S doping enhances light absorption and suppresses charge recombination, thereby significantly improving the H₂ production photoactivity. Like heteroelement doping, rational modification by such as cyano group was regarded as one of strategies to modulate the activity of CN. The abundant presence of cyano groups facilitates enhanced light harvesting, promotes more efficient separation of charge carriers, and increases the availability of active sites on the catalyst surface [22,23]. In this study, Lee et al. used SiO₂ clusters as templates to prepare cyano-bonded mesoporous graphitic carbon nitride nanoclusters as highly active photocatalysts, which completely degraded tetracycline solution within 30 min [24]. While the modification of CN through cyano group functionalization and elemental doping has been widely recognized as an effective approach in photocatalysis, limited research has explored the potential synergistic effects of combining these two strategies to further improve photocatalytic activity.

Herein, we developed a new strategy based on synergistic effect to co-modify CN with S and cyano groups (denoted as CNS) to enhance the photodegradation ability and photocatalytic H₂ production. The obtained CNS-X has significantly improved RhB degradation and H₂ production performance. CNS-6 showed superior photocatalytic behavior. Multiple characterization techniques confirmed the successful synergistic modification of S atoms and cyano groups. This dual modification induces an active structure with an electron compensation effect and spatial coordination advantage. It shows a significant synergistic catalytic mechanism in regulating the electron density of metal centers, constructing efficient reaction sites, stabilizing key intermediates, and optimizing reaction pathways. Our work provides a strategy to achieve the co-decoration of CN with heteroelements and functional groups.

2. Results and Discussion

2.1. Material Characterizations and Structural Model Optimization

The nanoscale architecture of the as-prepared samples was investigated through scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The SEM image in Figure 1a shows that CN has a thick bulk structure. However, after thermal reaction, the thickness of the obtained CNS-6 is greatly reduced, showing a porous flake structure (Figure 1b). All the prepared CNS samples showed a layered structure from TEM observation. In Figure 2a, CNS-6 presents nanoscale-layered morphology. The sample surface is a loose and diffuse thin layer with abundant pores. The TEM image shown in Figure 2b also clearly indicates that CNS-6 possess macropores and presents an ultra-thin layer structure that is almost transparent. Further observations show that there are abundant planar nanopores of a few nanometers, proving the formation of porous CNS-6 sheets (Figure 2c). Furthermore, both CNS-4 and CNS-8 exhibited porous structures (Figure S1).

As shown in Figure 3a, the specific surface area (BET) and pore volume of CNS-6 are 109 cm²/g and 0.683 cm³/g, respectively (Table S1). The BET and pore volume of CNS-6 are significantly larger than those of CN. The increased BET provides more active sites and reaction interfaces, which is beneficial for improving catalytic activity. Figure 3b presents the X-ray diffraction (XRD) profiles of CN and CNS-X. Layered CN presents two distinct diffraction peaks around 13° and 27°, corresponding, respectively, to the (100) in-plane packing and (002) interlayer $\pi$-$\pi$ stacking pattern of the heptazine structure [25,26]. For CNS-X, the main peaks are still near 13° and 27°, indicating that their main structure is CN. Moreover, the magnified image shows that the diffraction peak assigned to the (002) plane of CNS-6 changes to a higher diffraction angle, indicating that there is a strong van der Waals attraction between adjacent heptaquinone unit layers, resulting in a reduction in the stacking distance between $\pi$-$\pi$ layers. The (002) plane of the sample becomes weaker and wider, indicating that the morphology of CNS-6 becomes thinner. This observation aligns with previous TEM images. As shown in Figure 3c, the functional groups present in both CN and CNS were characterized via Fourier transform infrared (FT-IR). Furthermore, the broad absorption region at 1200–1700 cm⁻¹ can be attributed to the characteristic skeletal vibrations of aromatic structures, including C=C stretching in benzene rings and C-N heterocycles [27,28]. Specifically, the peaks at 1241–1242 cm⁻¹ correspond to C-N-C and C-N stretching vibrations [29]. Broad absorption features observed between 3100 and 3500 cm⁻¹ originate from stretching modes of -NH and -OH functional groups, resulting, respectively, from terminal amine functionalities and surface-bound water species. Notably, characteristic -C≡N stretching vibrations appear at 2160–2180 cm⁻¹ in CNS-X materials, confirming cyano group incorporation [30], which likely emerge due to the deprotonation of amino groups during thermal treatment. To further illustrate the structural stability of the catalyst, thermogravimetric analysis was performed. As shown in Figure S2, CNS exhibits almost no mass loss at 520 °C, indicating that it is stable at this temperature.

X-ray photoelectron spectroscopy (XPS) analysis provides definitive evidence of both elemental constituents and their bonding configurations. The scanning survey results (Figure 4a) fully explains that the original CN contains carbon (C), nitrogen (N), and oxygen (O). In parallel with C, N, and O, CNS-6 contains S elements, and CNS-4 (Figure S3) and CNS-8 (Figure S4) also contain them. This shows that S atoms are successfully doped into CN. Figure 4b shows two significant peaks at 284.6 eV and 288.8 eV, which correspond to adventitious carbon (C=C) and sp²-hybridized C atom bonded to N (N-C=N) [31]. Compared with the original CN, the observed peak shifts in the CNS-6 further confirm the breaking of heterocyclic bonds and the generation of cyano-related defects within the aromatic framework during the thermal treatment of N-C=N linkages [32]. The samples reveal four distinct binding energy signals centered at 398.5, 400.8, and 401.2 eV, along with a higher-energy feature attributed to $\pi$* excitation (Figure 4c). The presence of the $\pi$* signal in the XPS spectrum is likely associated with the delocalized π-electron system inherent to the conjugated structure of the photocatalyst. Characteristic peaks of the $\pi$* excitation state may emerge in the N 1s spectrum as a result of the electronic transition of this excited state. The various N chemical conditions are represented by these peaks. The 398.5 eV signal in pristine CN is characteristic of C=N-C coordination, with additional features at 400.8 eV and 401.2 eV corresponding to N-(C)₃ and C-N-H moieties. With regard to CNS-6, a comparable peak is detected, with a slight shift in the N-(C)₃ component. The creation of cyano groups, whose binding energy falls between that of N-(C)₂ and N-(C)₃ configurations, is probably responsible for this change [33]. According to the findings displayed in Figure 4d, the peak at 164.4 eV corresponds to C-S bonding, suggesting the substitution of N atoms by S within the CN framework. To reconfirm the successful introduction of sulfur atoms, we tested the sulfur content in CNS-6 using organic elemental analysis (Table S2). These findings provide solid evidence for the formation of defective CN simultaneously modified with S-doped and cyano groups.

2.2. Photocatalytic Properties

2.2.1. Optical Properties and Band Structures

Figure 5a clearly demonstrates that the optical properties of CN and CNS-6 and their light capture ability were systematically studied through ultraviolet–visible diffuse reflectance spectroscopy (UV–vis DRS). Because of its $\pi$-$\pi$ electron transition, the CN sample shows a distinct absorption edge at roughly 474 nm, suggesting a typical band gap absorption feature. This indicates the restricted absorption capacity of CN in the visible-light region. In contrast, CNS-6 shows a broader and enhanced absorption range, especially showing a significant red shift. This improvement is attributed to the incorporation of S atoms and cyano groups, which help to regulate its electronic structure, expand the $\pi$ conjugated system, and reduce the energy required for $\pi$-$\pi$ transitions, thereby effectively broadening the photoresponse range. The increased absorption intensity suggests that CNS-6 is more efficient in harvesting sunlight and utilizing photogenerated charge carriers. As shown in Figure 5b, from the Kubelka–Munk equation $\alpha h\nu = {A (h\nu -Eg) }^2,$ the bandgap energies of the CN and CNS-6 samples are obtained to be 2.89 eV and 2.87 eV. Obviously, functional group modification and element doping have an important influence on the band gap of semiconductors. Figure 5c presents the valence band (VB) spectra obtained by XPS. The VBs are determined to be 2.08 eV for CN and 2.21 eV for CNS-6. The formula Eg = E_VB − E_CB was used. Combined with the Eg value, the conduction bands (CBs) of CN and CNS-6 are inferred to be −0.79 eV and −0.68 eV. The position of E_VB and E_CB is the decisive factor affecting the catalytic redox ability. Figure 5d shows the band diagrams of CN and CNS-6. Compared with the original CN, the positions of CB and VB in CNS-6 shift downward, indicating that CNS-6 has a stronger oxidation ability.

2.2.2. Characterization of Charge Separation

To investigate the impact of S doping and cyano groups on charge carrier dynamics, we conducted transient photocurrent measurements and Nyquist plot analysis. The experimental consequents show that CNS-6 exhibits a sensitive photocurrent response in the on/off lighting cycle (Figure 6a). Compared with other catalysts, CNS-6 exhibits the strongest signal, indicating that CNS-6 has a smaller carrier migration impedance and a higher photogenerated charge transfer efficiency [34]. The experimental data plotted in Figure 6b indicate that the smallest arc radius in EIS for CNS-6 compared to other catalysts suggests indicating that its charge transfer resistance is significantly reduced [35]. It is obvious that CN has the weakest carrier separation efficiency. It can be seen from Figure 6c that the sample has a PL emission peak at 475 nm, and CNS-6 has a lower fluorescence emission intensity. These findings further demonstrate that S doping and cyano group introduction regulate the electronic configuration of CN, thus facilitating more effective spatial separation of photoinduced charge carriers within CNS.

2.3. Photocatalytic Activity

2.3.1. Photocatalytic H₂ Generation Performance

The visible-light-induced H₂ evolution reaction was performed to evaluate the catalytic performance of CN and CNS-X. As shown in Figure 7a, CNS-6 exhibits a remarkably enhanced activity compared to the pristine CN, with the highest H₂ production rate reaching 1221.47 μmol h⁻¹g⁻¹. Compared with other catalysts, it has better hydrogen production performance (Table S3). The experiments indicated the hierarchical porous structure of CNS-6 effectively reduces electron–hole recombination and consequently shortens migration path of charge carriers. This is consistent with the large specific surface area of CNS-6, which facilitates the provision of more active sites. The durability of CNS-6 in H₂ evolution was further confirmed, showing minimal performance loss after four consecutive cycles (Figure 7b). These findings indicate that CNS-6 possesses excellent photostability.

2.3.2. Photocatalytic Degradation of RhB

The photocatalytic degradation capability of CN and CNS-X was systematically investigated using RhB dye under visible-light illumination. Different samples were subjected to dark adsorption treatment for 60 min, and the adsorption equilibrium was reached at 30 min (Figure S5). The RhB concentration at 30 min was selected as the initial concentration for the experiment. Compared with other photocatalysts (Figure 8a), CNS-6 achieved a removal efficiency of up to 97.3% for RhB by heating for 6 h, which is also higher than other catalysts reported before (Table S4). This is consistent with CNS-6 having a more positive valence band potential, which confers a stronger oxidative capacity. S doping in CNS enhances electron utilization by accelerating electron transfer to the target RhB, thereby achieving a higher degradation efficiency. Figure 8b,c show the fitting of the degradation data using the Equation (2). It can be seen from the figure that the CNS-6 photocatalyst shows the best rate constant of 0.03250 min⁻¹. Table S5 showed the experimental error values of Rhb degradation by different catalysts. The modified CNS-6 has better degradation performance than other catalysts. The data obtained here strongly suggest that doping S in CNS consistently increases the oxidative capacity of CNS and thus markedly increases the degradation rate of RhB. The stability of the CNS-6 sample was also tested (Figure 8d). After four cycles for a total of 8 h, the stability of the CNS-6 photocatalyst still maintained an extremely high degradation rate, indicating that CNS-6 has recyclability and excellent stability. At the same time, photolysis experiments were carried out under the same irradiation conditions but without the addition of a photocatalyst. Figure S6 shows that under visible light conditions, the degradation of RhB is negligible.

2.3.3. Photocatalytic Mechanism

In order to determine the dominant oxidative species responsible for the photocatalytic decomposition mechanism, radical trapping investigations were carefully carried out. Disodium ethylenediamine tetraacetic acid (EDTA-2Na), isopropyl alcohol (IPA), and 1,4-benzoquinone (BQ) were used as used as quenchers for holes (h⁺), hydroxyl radicals (OH), and superoxide radicals (O₂⁻), respectively. Figure 9a,b clearly illustrates that the photocatalytic process of CNS-6 degrading RhB with different quenchers. When EDTA-2Na was added to the system, no significant impact on degradation efficiency was observed, suggesting that h⁺ is not an active species for RhB degradation. The introduction of BQ led to significant decreases in degradation efficiency, with reductions of 68.72%, respectively. Therefore, it is inferred that O₂⁻ plays the major role. OH plays a role, not by directly oxidizing bulk water but through surface -OH oxidation or intermediate O₂⁻/H₂O₂ conversion. Due to the simultaneous action of other ROS (such as H⁺ and O₂⁻), IPA capture only partially inhibits degradation [36].

3. Experimental Section

3.1. Materials

All materials were obtained at analytical reagent (AR) grade and applied without further processing. Melamine (C₃H₆N₆), trithiocyanuric acid (C₃H₃N₃S₃), acetonitrile (C₂H₃N), and triethanolamine (TEOA) were purchased Shanghai McLean Biochemical Technology Co. (Shanghai, China). Isopropyl alcohol (C₃H₈O), ethanol (C₂H₅OH), and sodium sulfate (Na₂SO₄) were purchased from Sinopharm Chemical Reagent Co. (Shanghai, China).

3.2. Preparation of Photocatalysts

3.2.1. Synthesis of CN

For CN synthesis, 2.0 g of C₃H₆N₆ was calcined in a muffle furnace using the following program: heating at 2 °C/min to 520 °C with a 2 h plateau. The sample obtained was labeled as CN.

3.2.2. Synthesis of S-Doped and Cyanide-Modified Carbon Nitride Photocatalysts

In total, 2.0 g of C₃H₃N₃S₃ was heated in a ceramic crucible at 2 °C/min until reaching 520 °C in a muffle furnace, with subsequent thermal treatments maintained for 4, 6, or 8 h, respectively. The product was cooled and set aside for use. By changing the heating time during the preparation process, a series of samples with different defects were obtained, named CNS-X. The products are named CNS-4, CNS-6, and CNS-8 according to their time gradient.

3.3. Experimental

3.3.1. Characterizations

Scanning electron microscopy (SEM, JSM-7600F, JEOL, Tokyo, Japan) and transmission electron microscopy (TEM, JEM-1400, 120 kV, JEOL, Tokyo, Japan) were used to study the morphology. Fourier transform infrared spectroscopy (FT-IR) was employed to identify functional groups in the samples. X-ray diffraction (XRD) patterns were measured on the Ultima IV instrument (Rigaku, Tokyo, Japan) using Cu-Kα radiation. Fourier transform infrared (FT-IR) spectrum was acquired on Bio-Rad 575c FTIR spectrometer (Bio-Rad Laboratories, Hercules, CA, USA). X-ray photoelectron spectroscopy (XPS) was conducted using a ESCALAB 250Xi (Thermo Fisher Scientific, Waltham, MA, USA, Al-Kα = 1486.6 eV). UV–vis diffuse reflectance spectra (DRS) were measured with a Lambda 950 UV–vis spectrophotometer (PerkinElmer, Waltham, MA, USA). Photoluminescence (PL) spectra were captured by a HITACHI F-7100 instrument (Tokyo, Japan) with a Xe lamp.

3.3.2. Electrochemical Testing

The electrochemical impedance spectroscopy (EIS) and transient photocurrent response (I-T) curves were obtained using a conventional three-electrode system on the CHI760 electrochemical workstation. Anhydrous sodium sulfate was used as the electrolyte. In total, 5 mg of photocatalyst was added to ethanol, sonicated and mixed, and then dropped uniformly on the FTO glass. It was dried at 60 °C and subsequently used as working electrode. In addition, Pt foil was used as counter electrode, and Ag/AgCl was used as reference electrode. I-T measurements were conducted under a 300 W Xe lamp. The EIS was tested under identical conditions over a frequency range of 10 Hz to 1.0 kHz with a 1.6 V sweep frequency.

3.3.3. Photocatalytic Performance Experiments

Photocatalytic H₂ production: 20 mg of photocatalyst was added to 50 mL of deionized water, along with TEOA and Pt. Air was removed from the reactor using a vacuum pump. A 300 W xenon lamp (420 nm ≤ λ ≤ 780 nm) was used as the light source. H₂ production was measured using a gas chromatograph (GC 7920, argon carrier gas, TCD detector, Beijing China Education Au-light Co., Ltd., Beijing, China) with online detection. Stability testing requires re-preparing the sacrificial agent solution with the same concentration, resetting the apparatus, and re-evacuating the reactor for 4–10 cycles.

Rhodamine B (RhB) degradation: 10 mg of photocatalyst was added to a 10 mg/L RhB solution and sonicated. To eliminate adsorption effects, the solution was stirred in the dark for 30 min. Subsequently, samples were collected and filtered every 20 min under visible light. UV–visible absorbance was measured at the RhB maximum absorption peak (λmax = 554 nm). The degradation rate is defined as below in Equation (1):

$$Degradation rate=(1-{\displaystyle \frac{C}{C_0}})\ast 100\%$$

(1)

The catalytic kinetics of the RhB degradation process was analyzed using a pseudo-first-order kinetic model. The specific method is Equation (2):

$$ln{\displaystyle \frac{C}{C_0}} =kt$$

(2)

In this equation, C denotes the real-time concentration of RhB, C₀ denotes the initial concentration of RhB, and k represents the reaction rate constant. Stability experiments were then conducted. After each cycle, the RhB solution was completely replaced with an equal volume and concentration. The catalyst was recovered and repeated multiple times, with data from each cycle analyzed.

4. Conclusions

In summary, CNS with S-modified and endowed with cyano-functionalization was successfully prepared to achieve efficient photocatalytic degradation and decomposition of water for H₂ evolution. The results show that S atoms and cyano groups synergistically regulate, which not only modulates the band structure of samples but also enhances their oxidation capacity, improves visible-light absorption, and promotes the efficient separation of photogenerated charge carriers. The numerical results indicate that CNS-6 exhibits a degradation rate of 97.30% for RhB and H₂ yield of 1221.47 μmol h⁻¹g⁻¹. The findings of this work highlight the significant promise of CN in photocatalysis and offer a foundation for advancing novel CN-derived photocatalytic materials.