The global energy shortage and environmental pollution are two major challenges facing humanity^[1-6]. To address these issues, various clean energy sources with environmentally friendly characteristics are being explored^[7]. Among them, hydrogen is regarded as one of the most promising clean energy sources^[8-15]. However, industrial hydrogen production currently relies heavily on fossil resources such as oil, natural gas, and coal. Therefore, the development of green hydrogen production technologies is both necessary and challenging^[16-17]. Among various methods, photocatalytic hydrogen production has emerged as a promising solution^[18-24], and the key to its industrial application lies in identifying suitable semiconductors that can achieve highly efficient hydrogen evolution rates.

Metal-free graphite carbon nitride (g-C₃N₄) has garnered significant attention in the field of photocatalysis since its photocatalytic hydrogen evolution activity was first discovered in 2009^[25]. g-C₃N₄ possesses a suitable band structure for water splitting under visible light irradiation, along with excellent physical and chemical stability^[26-27]. However, conventional g-C₃N₄ faces limitations in industrial hydrogen production due to its low efficiency in utilizing photogenerated electron-hole pairs^[28-30]. To overcome this problem, alkali metal doping has been identified as an effective approach. For instance, Wu et al.^[31] demonstrated enhanced photocatalytic hydrogen evolution in g-C₃N₄ through K doping. Zhu et al.^[32] improved the charge carrier separation efficiency of g-C₃N₄ by incorporating Na⁺ into its framework. Zhao et al.^[33] enhanced the photocatalytic activity of porous graphitic carbon nitride nanorods by adding functional Na⁺. Han et al.^[34] improved photocatalytic hydrogen evolution in Na-doped g-C₃N₄ using a NH₄Cl-assisted method. Similarly, Dou et al.^[35] developed willow-shaped g-C₃N₄ with sodium doping, achieving enhanced photocatalytic performance under visible light. There is a consensus that Na doping is an effective strategy for improving the photocatalytic activity of g-C₃N₄. Meanwhile, O doping has also been reported as another promising approach^[36-37]. Hence, designing Na and O co-doped g-C₃N₄ is expected to further enhance photocatalytic activity. Ren et al.^[38] synthesized Na and O co-doped g-C₃N₄ for H₂O₂ production, indirectly supporting this hypothesis. However, the potential of Na and O co-doped g-C₃N₄ in photocatalytic hydrogen production remains unclear, particularly regarding the synergistic effects between Na and O atoms, which requires further experimental exploration.

In this work, carbon nitride co-doped with Na and O (Na/O-CN_x) was synthesized using NaCA as sodium source. The enhanced photocatalytic activities of Na/O-CN_x samples were thoroughly analyzed through structural, compositional, and optical performance characterizations. Additionally, the synergistic effects in Na/O-CN_x were elucidated through comparative studies of pure CN, Na/O-CN_x, and Na/O-CN_3.0-yO₂ samples by using carefully planned experiments.

Melamine (C₃H₆N₆, Aladdin Chemistry, ≥99%), sodium citrate tribasic dihydrate (C₆H₅Na₃O₇·2H₂O, Aladdin Chemistry, ≥99%), triethanolamine (C₆H₁₅NO₃, Sinopharm Reagent, ≥99%) and chloroplatinic acid hexahydrate (H₂PtCl₆·6H₂O, J&K Scientific Co., Ltd., ≥99.9%) were used without any purification. Deionized water was produced by Unique-R10 with a specific resistance of 18.2 MΩ∙cm.

Pure CN powder was synthesized following a reported method^[34,39]. Melamine (10 g) was heated in a muffle furnace at 550 ℃ for 4 h in static air with a ramp rate of 10 ℃/min. The reaction took place in a covered crucible, and the resulting CN powder was ground using an agate mortar.

The synthesis process of Na/O-CN_x samples is shown in Fig. 1. Pure CN (1.0 g) was mixed with x g (x=1.0, 2.0, 3.0, 4.0) of sodium citrate, and the mixture was ground for 10 min using a mortar. The resulting powders were placed in a Teflon-sealed autoclave (25 mL) and heated in the oven at 180 ℃ for 6 h. This synthesis temperature is lower than that used in many other alkali metal-doped carbon nitride processes (Table S1). After cooling to room temperature, the samples were washed with water three times and dried at 50 ℃ for 12 h, yielding Na/O-CN_x samples (x=1.0, 2.0, 3.0, 4.0).

Na/O-CN_3.0-yO₂ samples were prepared using 3.0 g of sodium citrate under varying oxygen contents (y=0, 20%, 40%, 60%, 80%, 100%, in volume) in a mixture of oxygen and nitrogen gases instead of air. The resulting samples were labeled as Na/O-CN_3.0-0O₂, Na/O-CN_3.0-20%O₂, Na/O-CN_3.0-40%O₂, Na/O-CN_3.0-60%O₂, Na/O-CN_3.0- 80%O₂, and Na/O-CN_3.0-100%O₂.

Details on the photocatalytic hydrogen evolution test and materials characterization can be found in the supporting materials.

The structures of the synthesized Na/O-CN_x samples were analyzed via X-ray diffractometer (XRD), as shown in Fig. 2(a, b). All peaks of Na/O-CN_x samples align with those of pure CN at 2θ=13.0° and 27.5°^{[33,40 -41]}, indicating that the in-plane and interlayer structures of carbon nitride remain intact. However, the peak intensity at 2θ=27.5° of Na/O-CN_x samples increases compared to pure CN, suggesting that Na and O co-doping enhances the crystallinity of Na/O-CN_x^[42-44]. Similar phenomena are observed in Na/O-CN_3.0-yO₂ samples depicted in Fig. 2(b), indicating that main structures of Na/O-CN_3.0-yO₂ remain intact, with only the crystallinity of samples being affected by the doping Na and O atoms. Fourier transform infrared (FT-IR) spectra of samples have also been operated for analyzing their structures, as shown in Fig. 2(c). Na/O-CN_3.0 and Na/O-CN_3.0-40%O₂ samples exhibit the same peaks in the range of 807-880 cm^-1, corresponding to the out-of-plane bending mode of the triazine ring in carbon nitride^[45]. Additionally, Na/O-CN_3.0 and Na/O-CN_3.0-40%O₂ samples also show the same absorption peaks in the range of 1200-1700 cm^-1 compared to pure CN, corresponding to the aromatic C-N heterocycle structures in carbon nitride^[46]. In contrast, the peaks in the range of 3100-3400 cm^-1 for Na/O-CN_x slightly shift compared to pure CN, indicating that the stretching vibrations of -N-H groups in Na/O-CN_x are altered due to Na and O doping^[47-49]. Consequently, the electronic structure of Na/O-CN_x is tuned by the synergistic effects of Na and O co-doping^[50].

Scanning electron microscope (SEM) images (Fig. 3(a, b)) reveal that Na/O-CN_3.0 has a block-like structure similar to pure CN with no structural damage after the reaction in the Teflon-sealed autoclave. Transmission electron microscope (TEM) images (Fig. 3(c, d)) confirm that the nanosheet structure of Na/O-CN_3.0 is also similar to that of pure CN^[51]. High-resolution TEM (HRTEM) image (Fig. 3(d)) indicates a lattice spacing of 0.33 nm for Na/O-CN_3.0, consistent with (002) plane (JCPDS #87-1526)^[26,34,52]. However, slight structural differences between Na/O-CN_3.0 and pure CN are evident through Brunauer-Emmett-Teller (BET) analysis (Fig. S1). The surface area of Na/O-CN_3.0 is 18.8 m²/g, 60.7% larger than that of pure CN (11.7 m²/g). This value is even larger than some alkali metal-doped carbon nitrides synthesized at higher temperatures, as indicated in Table S1. Additionally, Na/O-CN_3.0 exhibits a larger pore volume (0.039 cm³/g) compared to pure CN (0.023 cm³/g). These changes in the surface area of Na/O-CN_3.0 may result from oxidation reactions with O₂ from the air or the solid-phase reaction^[53] between pure CN and sodium citrate in the Teflon-sealed autoclave under high temperature and pressure. As a result, Na/O-CN_3.0 with increased reactive sites is expected to enhance photocatalytic hydrogen evolution activity.

The chemical compositions of Na/O-CN_3.0, pure CN, and Na/O-CN_3.0-40%O₂ were analyzed using X-ray photoelectron spectroscopy (XPS). As shown in Fig. S2, XPS survey spectra confirm the presence of C, N, and O elements in all samples, while Na is detected only in Na/O-CN_3.0 and Na/O-CN_3.0-40%O₂ samples. To further investigate the chemical bonds, high-resolution XPS spectra for C1s, N1s, O1s, and Na1s were presented in Fig. 4. In C1s XPS spectra (Fig. 4(a)), three distinct peaks are observed at 284.6, 286.0, and 288.0 eV, corresponding to C-C/C=C bonds, C-NH_x species in CN, and sp²-bonded carbon (N-C=N), respectively^{[32,39,42,54]}. N1s XPS spectra (Fig. 4(b)) show two common peaks across all samples at 398.4 and 401.0 eV, associated with sp² hybridized C=N-C in the triazine ring and amino groups (C-N-H), respectively^[55]. Notably, the nitrogen- related peak at 399.9 eV shifts to a lower binding energy in Na/O-CN_3.0 sample, indicating slight disruption of N-(C)₃ moieties structure^[39,43]. The C-O bond and Na element are detected only in Na/O-CN_3.0 and Na/O-CN_3.0-40%O₂ samples (Fig. 4(c, d)), confirming the successful incorporation of Na and O atoms into these samples. In summary, the chemical bonds in Na/O-CN_3.0 are adjusted by doping Na and O atoms.

Additionally, electron spin resonance (ESR) measurement (Fig. S3) was conducted to analyze the defect structures. All samples exhibit the same Lorentzian line centered around a g-factor of 2.004, corresponding to unpaired electrons from sp² carbon atoms in the aromatic rings^[11,13,15]. However, the peak intensities for Na/O-CN_3.0 and Na/O-CN_3.0-40%O₂ are higher than those for pure CN, which is attributed to defects introduced by Na and O doping in Na/O-CN_x samples^[56].

To elucidate the impact of doping defects, the band structures of prepared samples were characterized using ultraviolet-visible diffuse reflectance spectroscopy (UV-Vis DRS), as shown in Fig. 5(a). Na/O-CN_3.0 exhibits a stronger absorption intensity compared to pure CN, which enhances the generation of photogenerated electron-hole pairs. The bandgap of Na/O-CN_3.0 is calculated to be 2.68 eV using the Tauc equation^[57] (Fig. 5(b)), which is slightly lower than that of pure CN (2.70 eV). This reduction in bandgap improves its visible light absorption. To further investigate the band structures, valence band (VB) XPS spectra of the samples were measured (Fig. 5(c)). The VB values for pure CN and Na/O-CN_3.0 are determined to be 1.97 and 2.01 eV, respectively. The potential band structures are depicted in Fig. 5(d). The conduction band (CB) position of Na/O-CN_3.0 is slightly lower than that of pure CN but remains higher than the hydrogen reduction potential (H⁺/H₂, 0 eV), making it suitable for water splitting applications^[58].

Photocurrent and electrochemical impedance spectroscopy (EIS) measurements (Fig. 6(a, b)) were conducted to analyze the photocarrier separation and transfer capabilities. Na/O-CN_3.0 displays a higher photocurrent density compared to pure CN, indicating enhanced formation of photogenerated electron-hole pairs. Similarly, Na/O-CN_3.0 exhibits a smaller arc radius in the Nyquist plot than pure CN, reflecting improved conductivity. Additionally, photoluminescence (PL) tests of Na/O-CN_3.0 and pure CN, presented in Fig. 6(c, d), reveal that the PL intensity of Na/O-CN_3.0 is lower than that of pure CN, suggesting a reduced recombination rate of photogenerated electron-hole pairs in Na/O-CN_3.0 compared to pure CN^[52]. The fluorescence lifetime of Na/O-CN_3.0 is 1.26 ns, slightly shorter than pure CN (1.38 ns), as shown in Fig. 6(d), corresponding to more efficient charge separation and transfer capabilities^[59]. Therefore, doping Na and O atoms in Na/O-CN_3.0 improves photocarrier separation and electron transfer, and enhances its photocatalytic hydrogen evolution.

The photocatalytic hydrogen evolution performances of prepared samples are shown in Fig. 7. As depicted in Fig. 7(a), all Na/O-CN_x samples exhibit an increased photocatalytic hydrogen evolution rate (PHER). The PHER of Na/O-CN_x samples initially rises with the increasing content of sodium citrate, but decreases when x exceeds 3.0. These results indicate that suitable co-doping with Na and O atoms is beneficial for the photocatalytic activity of Na/O-CN_x samples.

To further explore the synergistic effects between Na and O atoms, PHER of Na/O-CN_3.0-yO₂ samples synthesized under varying oxygen contents were analyzed (Fig. 7(b)). Na/O-CN_3.0-0O₂ sample shows higher photocatalytic hydrogen evolution than pure CN, indicating that Na doping is a key factor in improving photocatalytic activity. As oxygen content increases, PHER of Na/O-CN_3.0-yO₂ first rises, then decreases when oxygen content exceeds 40%, demonstrating that suitable O doping also positively impacts photocatalytic activity. Therefore, the enhanced photocatalytic activity of Na/O-CN_x samples is attributed to the synergistic effect of Na and O atom doping. Among the samples, Na/O-CN_3.0 shows the highest photocatalytic hydrogen evolution, with a PHER of 103.2 μmol∙g^-1∙h^-1 (Fig. 7(c)), which is 8.2 times higher than that of pure CN (11.2 μmol∙g^-1∙h^-1). Additionally, Na/O-CN_3.0 stands out as a strong photocatalyst for hydrogen evolution, compared to other reported alkali metal-doped carbon nitride samples, as shown in Table S1. Interestingly, PHER of Na/O-CN_3.0 synthesized under air (with approximately 21% O₂) surpasses those of all Na/O- CN_3.0-yO₂ samples, likely due to the presence of some unknown trace gases in the air.

Then Na/O-CN_3.0 was chosen for the stability test. As shown in Fig. 7(d), PHER of Na/O-CN_3.0 remains stable over a 12.5 h cycling experiment. After the first cycle, PHER of Na/O-CN_3.0 slightly increases, which may be due to in-situ formation of co-catalyst Pt during the photocatalytic reaction.

Based on the discussions above, the proposed photocatalytic hydrogen evolution mechanism for Na/O-CN_3.0 is illustrated in Fig. 8. Under visible light irradiation, Na/O-CN_3.0 is excited to generate photogenerated electrons, which combine with H⁺ ions to produce hydrogen gas. Simultaneously, the corresponding holes are involved in the oxidation of triethanolamine (TEOA). Due to the synergistic effect of Na and O atom doping, Na/O-CN_3.0 exhibits enhanced production of photogenerated electron-hole pairs while reducing their recombination rate. Consequently, Na/O-CN_3.0 demonstrates significantly higher photocatalytic hydrogen evolution activity compared to pure CN.

In summary, Na/O-CN_x were synthesized using NaCA as the sodium source at relatively low temperature. Na/O-CN_x samples exhibit an increased surface area, providing more active sites for reactions. The improved structure facilitates the transfer of photogenerated charges and enhances their production. Consequently, Na/O-CN_x samples achieve significantly higher photocatalytic hydrogen evolution rates compared to pure CN while maintaining excellent stability. This study offers a novel strategy for synthesizing alkali metal- doped carbon nitride at lower temperature, enabling highly efficient photocatalytic hydrogen evolution.

Supporting materials related to this article can be found at https://doi.org/10.15541/jim20240345.

Na and O Co-doped Carbon Nitride for Efficient Photocatalytic Hydrogen Evolution

CHEN Libo¹, SHENG Ying¹, WU Ming¹, SONG Jiling², JIAN Jian¹, SONG Erhong³

(1. School of Chemistry and Chemical Engineering, Hunan University of Science and Technology, Xiangtan 411201, China; 2. National Engineering Research Center for Compounding and Modification of Polymer Materials, Guiyang 550014, China; 3. Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China)

Photocatalytic hydrogen evolution test

Photocatalytic reactions of samples were conducted on an on-line photocatalytic hydrogen evolution system (CEL-PAEM-D8). 50 mg samples were dispersed in triethanolamine solution (50 mL, 10%, in volume) with 0.5 mL H₂PtCl₆ aqueous solution (1 mg/mL, Pt basis) as Pt source. Pt-CN and Pt-Na/O-CN_x samples were synthesized with in situ photo-deposition method by 30 min simulated sunlight irradiation. Photocatalytic hydrogen reaction was operated under vacuum condition with visible light irradiation of a 300 W xenon-lamp (CEL-HXF 300) equipped with a 420 nm cutoff filter (AULIGHT). Reaction solution temperature was kept below 6 ℃ with a flow of cooling water, and obtained hydrogen was analyzed by gas chromatography (GC7920- TFZA).

Catalyst characterization

Powder X-ray diffraction (XRD) patterns were performed on a Bruker D8 Advance X-ray diffractometer with Cu Kα irradiation (λ=1.5406 Å) by a scan rate of 5 (°)/min. Transmission electron microscope (TEM) images were acquired using the JEOL JEM-F200. Scanning electron microscope (SEM) images were obtained utilizing the ZEISS Sigma 300. X-ray photoelectron spectroscopy (XPS) was measured using the Thermo Scientific K-Alpha spectrometer, with the C1s reference peak positioned at 284.6 eV. The Shimadzu UV-3600 spectrophotometer was employed to acquire the ultraviolet-visible diffuse reflectance spectrum (UV-Vis DRS) of prepared samples. The fluorescence lifetime and photoluminescence (PL) spectra of the sample were measured on the FLS1000 transient fluorescence spectrometer. Fourier transform infrared (FT-IR) spectra were acquired using the Thermo Scientific iN10 instrument. The Micromeritics ASAP 2460 analyzer was utilized to determine the Brunauer- Emmett-Teller (BET) surface areas through N₂ adsorption. Electron spin resonance (ESR) spectra of samples were recorded on JEOL JES-FA200 ESR spectrometer.

The electrochemical measurements were performed in a conventional three-electrode cell on a CHI-760E electrochemical workstation (Shanghai Chenhua Instrument Co., Ltd., China). During the photocurrent measurement, a saturated calomel electrode (SCE) was used as the reference electrode and a Pt foil electrode acted as the counter electrode. The working electrodes were designed using resulting samples covered on the surface of fluoride tin oxide (FTO) conductor glass. A quartz cell filled with 0.5 mol·L^-1 Na₂SO₄ electrolyte was used as the measuring system. For electrochemical impedance spectroscopy (EIS) measurements, the amplitude of the sinusoidal wave was 5 mV, and the frequency range was from 100 kHz to 0.1 Hz.

References:

[S1] CHEN Y, WANG X C. Template-free synthesis of hollow G-C₃N₄ polymer with vesicle structure for enhanced photocatalytic water splitting. J. Phys. Chem. C, 2018, 122(7): 3786.

[S2] WANG Y Y, ZHAO S, ZHANG Y W, et al. One-pot synthesis of K-doped g-C₃N₄ nanosheets with enhanced photocatalytic hydrogen production under visible-light irradiation. Appl. Surf. Sci., 2018, 440: 258.

[S3] CHANG X Y, FAN H Q, ZHU S W, et al. Engineering doping and defect in graphitic carbon nitride by one-pot method for enhanced photocatalytic hydrogen evolution. Ceram. Int., 2023, 49(4): 6729.

[S4] BAI P, WANG P, LI T, et al. Alkali functionalized carbon nitride with internal van der Waals heterostructures: directional charge flow to enhance photocatalytic hydrogen production. J. Colloid Interface Sci., 2023, 644: 211.

[S5] SHU Z, WANG Y, WANG W B, et al. A green one-pot approach for mesoporous g-C₃N₄ nanosheets with in situ sodium doping for enhanced photocatalytic hydrogen evolution. Int. J. Hydrog. Energy, 2019, 44(2): 748.

[S6] ZHANG L S, DING N, HASHIMOTO M, et al. Sodium-doped carbon nitride nanotubes for efficient visible light-driven hydrogen production. Nano Res., 2018, 11(4): 2295.

[S7] SHANG Y Y, MA Y J, CHEN X, et al. Effect of sodium doping on the structure and enhanced photocatalytic hydrogen evolution performance of graphitic carbon nitride. Mol. Catal., 2017, 433: 128.

[S8] ZHANG Y Z, ZONG S C, CHENG C, et al. One-pot annealing preparation of Na-doped graphitic carbon nitride from melamine and organometallic sodium salt for enhanced photocatalytic H₂ evolution. Int. J. Hydrog. Energy, 2018, 43(30): 13953.

[S9] FANG W J, LIU J Y, YU L, et al. Novel (Na, O) co-doped g-C₃N₄ with simultaneously enhanced absorption and narrowed bandgap for highly efficient hydrogen evolution. Appl. Catal. B Environ., 2017, 209: 631.

[S10] DOU Q, HOU J H, HUSSAIN A, et al. One-pot synthesis of sodium-doped willow-shaped graphitic carbon nitride for improved photocatalytic activity under visible-light irradiation. J. Colloid Interface Sci., 2022, 624: 79.

[S11] HU S Y, YU A C, LU R, A comparison study of sodium ion- and potassium ion-modified graphitic carbon nitride for photocatalytic hydrogen evolution. RSC Adv., 2021, 11(26): 15701.

[S12] ZHANG G Q, ZHU J Y, XU Y S, et al. In-plane charge transport dominates the overall charge separation and photocatalytic activity in crystalline carbon nitride. ACS Catal., 2022, 12(8): 4648.