Two-dimensional (2D) graphene-analogue semiconductor materials, based on the typical layer structured semiconductors, exhibit peculiar and fascinating properties in contrast with those of their bulk parent compounds ^[1]. The exceptional properties of 2D semiconductors, such as piezoelectric coupling and band energy transition, will enable new breakthroughs in nanomaterials science ^{[2, 3]}. Inevitably, the photochemistry of 2D semiconductor photocatalysts has received increasing attention to address the problem of the usage of solar energy. Previous studies have reported the fabrication of semiconductors into 2D nanosheets for photocatalytic application, such as layered MoS₂, WS₂, SnS₂, TiO₂ and so forth ^[4-7]. Theoretical analysis has shown that 2D graphene-analogue materials possess many advantages as catalysts in a visible-light photocatalytic reaction, such as increasing the absorption of light, surface area, speed of carrier transport and so forth.

Recently, 2Dmetal-free materials, as key determinants of cost-effective hydrogen generation on a large-scale, have been identified as newcomers to the family of photocatalysts ^[8]. Graphitic carbon nitride (g-C₃N₄), a sustainable conjugated 2D polymer semiconductor with a medium band gap of 2.7 eV, has been shown to exhibit great potential to scientists searching for new materials for energy conversion in the future ^[9-11]. As is desired for analogues of graphene, g-C₃N₄ possesses a graphitic-like layered structure and involves van der Waals interactions between adjacent C-N layers with strong covalent bonding between each layer. Therefore, its 2D layers present promising opto-electronic properties and facilitate coupling with various functional materials to enhance the performance ^[12-14]. However, the preparation of their nanosheet structures with atomic scale thickness remains difficult owing to the experimental challenges.

Motivated by the intriguing graphene chemistry, the exfoliation of bulk g-C₃N₄ into single- or few layer-2D layers has been actively pursued. Mechanical liquid ultrasonication- assisted exfoliation in water and organic solvents, such as ethanol and isopropanol, has been used to achieve g-C₃N₄ nanosheets ^{[15, 16]}. The chemical oxidation etching method has generally been regarded as an efficient preparation method for 2D g-C₃N₄ nanosheets with a unilamellar structure. With the progress in this field, oxidant HNO₃ or H₂SO₄ have been used as intercalation compounds ^{[17, 18]}, and other intercalation compounds have also been reported recently ^{[19, 20]}. However, the obtained catalysts were often doped materials. For example, Song et al. ^[21] obtained few-layer-thick g-C₃N₄ nanosheets from liquid ammonia (LA)-assisted lithiation, and the samples showed an enhanced photocatalytic redox activity with respect to both photocatalytic H₂ evolution and hydroxyl radical generation. Dong et al. ^[22] prepared a stable colloidal suspension of g-C₃N₄ nanosheets through an H₂SO₄ exfoliation route. Recently, Zhang et al. ^[23] reported a one-step electrochemical method to prepare ultrathin g-C₃N₄ nanosheets from melamine and the as-synthesized materials showed intrinsic peroxidase-like activity. However, structure disorder was observed owing to the partial destruction of the triazine units under the electrolytic process. Thus, developing a facile method for the large-scale production of well-dispersed g-C₃N₄ nanosheets is urgently required. The approach of thermal oxidation exfoliation i s regarded as highly efficient and environmentally friendly ^[24]. Currently, the most direct and simple method to prepare g-C₃N₄ composed with nanosheet- like morphology is direct pyrocondensation of urea or thiourea ^[25]. Nevertheless, its microstructure and ability for photocatalytic hydrogen evolution still needs optimization ^[26].

The introduction of nanopores into the bulk structure of g-C₃N₄ could effectively increase the specific surface area, thereby increasing the number of surface active sites and photocatalytic activities of the catalysts. A number of methods have been discussed for the preparation of porous g-C₃N₄, including replicating the synthesis from SiO₂ templates or soft templates ^[27-29], using pore-formingagents ^[30], and protonation processes ^[31]. However, during the synthetic progress, the compellent pore-creation leads to incomplete condensation of carbon nitride and increases the disorder of the layer stacking structure. Hence, this becomes unfavorable to the improvement of electron transmission and photocatalytic performance. Therefore, a direct route for 2D g-C₃N₄ nanosheets with high polymerization degree and porosity needs to be developed to further improve the photocatalytic activity. Recently, ammonia thermopolymeriza- tion was developed by our group to prepare g-C₃N₄ with a high degree of condensation polymerization ^[32]. Based on this method, porous g-C₃N₄ nanosheets are expected to be achieved by a one-step process.

In this work, porous g-C₃N₄ nanosheets were synthesized by the direct ammonia thermopolymerization method at different temperatures. Experimental results confirmed that a higher condensation temperature increases the surface area and pore structure, optimizes the photoelectric properties of catalysts to inhibit the recombination of carriers, and therefore, enhances the photocatalytic activity for the H₂ evolution under visible-light irradiation.

All materials were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) and used as received without further purification.

Porous g-C₃N₄ nanosheet photocatalysts (pg-C₃N₄) were prepared using the ammonia-polymerization approach ^[32]. In a typical procedure, 10 g of ammonium thiocyanate (NH₄SCN) was placed in a tube furnace and then heated to a certain temperature in the range of 450-600 ℃ for 2 h with a ramp rate of 5 ℃/min in an ammonia atmosphere, followed by naturally cooling to room temperature. Air was removed for 30 min through ammonia before heating. Final products were obtained after grinding the products into powders and were denoted as CN-T, where T refers to the calcination temperature (℃).

X-ray diffraction (XRD) patterns were collected on a Bruker D8 Advance diffractometer with Cu K_α radiation (λ = 1.5406 Å). Fourier transformed infrared (FTIR) spectra were obtained using a Nicolet Magna 670 FTIR spectrometer in KBr at a concentration of approximately 1 wt%. Nitrogen adsorption- desorption isotherms were collected at -196 ℃ using a Micromeritics ASAP 2020 surface area and porosity analyzer. Ultraviolet-visible (UV-vis) diffuse reflectance spectroscopy (DRS) was performed on a Varian Cary 500 Scan UV-visible system. Scanning electron microscopy (SEM) was performed using a JEOL model JSM-6700F instrument. The photoluminescence (PL) emission spectra were recorded on an Edinburgh FI/FSTCSPC 920 fluorescence spectrometer. Electron paramagnetic resonance (EPR) measurements were carried out on a Bruker model A300 spectrometer.

Photocatalytic activity was evaluated by the photocatalytic hydrogen evolution from water under visible light irradiation (λ > 420 nm). Reactions were carried out in a Pyrex top- irradiation reaction vessel connected to a glass closed gas system. In each experiment, 50 mg of photocatalyst powder was dispersed in an aqueous solution (100 mL) containing triethanolamine (10 vol%) as a sacrificial electron donor. Pt (3 wt%) was loaded on the surface of the catalyst by in situ photodeposition using H₂PtCl₆. The reaction solution was evacuated several times to completely remove air prior to irradiation under a 300-W xenon lamp. The wavelength of the incident light was controlled by applying the appropriate cut-off filters. The temperature of the reaction solution was maintained at room temperature by the flow of cooling water during the reaction. The evolved gases were analyzed by gas chromatography equipped with a thermal conductivity detector (TCD) and a 5-Å molecular sieve column, using argon as the carrier gas.

Fig. 1 exhibits the XRD patterns of the g-C₃N₄ samples condensed at different heating temperatures in ammonia. It can be observed at 450 ℃, graphitic-like networks are incompletely formed. Upon increasing the calcination temperature to 500 ℃, the distinctive graphitic-like structure (JPCDS 87-1526) is evolved. The strong diffraction peaks at approximately 27.5° and 13.0° are indexed as (002) of the inter-layer stacking reflection and (100) of the in-plane structure repeating motif of heptazine-based g-CN, respectively. Obviously, the polymerization temperature was increased compared with our previous work in which a graphitic C₃N₄ structure could be formed at 450 ℃ with NH₄SCN heating in nitrogen. This may be attributed to the inhibitory effect of deamination of NH₄SCN in extra ammonia, which is beneficial for the condensation degree of NH₄SCN to g-C₃N₄.

For CN-500, CN-550 and CN-600, with a heating temperature increase, the intensities of the XRD peaks are not weakened and become narrower, indicating an optimized polycondensation of g-C₃N₄. Compared with previous reports that a higher temperature usually induces decomposition of the g-C₃N₄, a highly condensed and compact packing of the conjugated skeleton of g-C₃N₄ could be achieved during ammonia polymerization. In addition, the (002) peak displays an up-shift from 27.5° to 27.8°, corresponding to a reduction in the stacking distance from 0.324 to 0.320 nm of the graphitic layer. Hence, a denser layer packing structure was achieved at a higher condensation temperature. Considering the pronounced peak at approximately 13.0°, no obvious difference could be detected among the CN-500 to CN-600 samples. This indicates that the three samples possess virtually the same void-to-void distance (d = 0.681 nm) of the in-plane structural repeating motifs, and the in-plane connection structure of the samples is well maintained even if synthesized at 600 ℃.

FTIR spectra of CN synthesized at different temperatures in ammonia are shown in Fig. 2. Clearly, for the sample synthesized at 450 ℃, the weak bands in the 1200-1700 cm^-1 range and at 810 cm^-1 are attributable to the organic molecules containing s-triazine ring moieties. However, the strong peak near 480 cm^-1, corresponding to C-S vibration, also exist, which illustrates the incomplete condensation of the g-C₃N₄ samples.

When the temperature exceeds 500 ℃, a practically complete desulphurization takes place and the as-prepared samples show typical IR peaks of C₃N_4. The bands at 1200-1700 cm^-1 (1640, 1468, 1320, 1238 cm^-1) are regarded as stretching vibration peaks of the thiazine or heptazine heterocyclic ring (C₆N₇) units and broad absorption in the region of 3100-3300 cm^-1 is related to the stretching modes of residual secondary and primary amines or absorbed water. For CN-500 to CN-600, the slight shift (4 cm^-1) to a higher wavenumber of the sharp peak at ~810 cm^-1, which is considered as a characteristic breathing mode of C₆N₇ units, illustrates the improved conjugation of the tri-s-triazine structure ^[33]. The small change of all these bands indicates that the porous carbon nitride nanosheets obtained at different heat-treatment temperatures does not lead to notable changes in the basic unit of CN samples. These IR spectra clearly prove that NH₄SCN can undergo condensation processes to construct a heterocyclic tri-s- triazine ring system in ammonia, and higher temperatures enable the obtained samples to possess a higher content of the tri-s-triazine phase.

The textural properties of CN-T samples were analyzed using a nitrogen adsorption-desorption spectrometer. The isotherms and the Barrett-Joyner-Halenda (BJH) pore size distributions are shown in Fig. 3. The isotherms of the samples show typical hysteresis, proving the existence of mesopores connected by micropores. The Brunauer-Emmett-Teller (BET) surface area (A_BET), BJH pore sizes, and pore volumes are summarized in Table 1. The A_BET and nanopores of CN-T were shown to be closely dependent on the calcination temperature. For CN-450, the A_BET was only 9 m²/g, with the equivalent bulk CN catalyst prepared from NH₄SCN in nitrogen. When increasing the heating temperature from 500 to 600 ℃, the A_BET of CN-T increased from 20 to 52 m²/g. The average pore diameter and pore volume of the samples also increased with an increased calcination temperature. Interestingly, for the CN-500 to CN-600 samples, in addition to the enlarged mesopore volume, micropore sizes centered at approximately 3.8 nm appeared and these pore volumes also increased. The enlarged surface area and increased pore volumes could increase the surface mass transfer and improve the light absorption by reducing the light scattering, therefore enhancing the photocatalytic performance.

In this work, the grain morphology of CN-T catalysts obtained under different temperatures was investigated by SEM analysis, and clear pictures are shown in Fig. 4. It is evident that the CN-500 sample presents an irregular bulk stacking like the intrinsic samples prepared in nitrogen. With an increasing temperature, smaller particles and a sheet structure distribution could be observed for CN-550. Furthermore, a remarkable nanosheet morphology accompanied with nanopores was apparent for CN-600. The results are consistent with the increased surface area and nanopores of these samples. The formation of the 2D nanosheet structure is an important factor to improve the photocatalytic performance. Meanwhile, the formation of nanopores in a graphitic layer of g-C₃N₄ is predictable for the improvement of the photocatalytic activity.

The optical features of the as-prepared CN-T materials were examined by UV-vis DRS, as shown in Fig. 5. All samples demonstrated semiconductor-like absorptions in the blue visible light range. As expected, increasing the heating temperature from 400 to 600 ℃, the absorption edges of products showed a hypochromatic shift from 452 to 424 nm, corresponding to E_g = 2.74 to 2.92 eV for CN-500 to CN-600, owing to the strong quantum size confinement effect. Although a decreased optical absorption in the visible light region was generated, the formation of a layer structure, which was compressed to a smaller inter-plane distance, together with the enlarged surface area are believed to promote the photocatalytic redox functions.

Room-temperature PL spectra were obtained using excitation light of 400 nm to investigate the efficiency of charge- carrier separation/recombination in the materials. As shown in Fig. 6, all the samples exhibited a broad emission peak centered around 450 nm, which can be attributed to the band-band PL phenomenon with the light energy approximately equal to the band gap energy of g-C₃N₄. The lower PL emission peak intensity for CN-450 was attributed to the rich-defected heterocycle formed from incomplete condensation. The PL intensity decreased with increasing the polymeric temperature from 500 to 600 ℃, which indicates a suppressed recombination rate of the photo-induced charge carriers. The main benefit derives from the optimized construction of the layer stacking structure, which hastens the mobility of the free charge carriers and facilitates a high photocatalytic rate.

In addition, room-temperature EPR analysis was carried out to investigate the electronic band structure of the porous g-C₃N₄ samples. In Fig. 7, one single Lorentzian line, centered at a g value of 2.0034 originating from the unpaired electrons in the aromatic rings of carbon atoms, is observed for all of the g-C₃N₄ samples.

For samples obtained under a gradually increasing temperature, the enhanced spectral line intensity of EPR, demonstrates the progressive development of the electronic band structure. A larger surface area is a key factor for the enhancement of the EPR Lorentzian line, therefore, increasing the density of the surface sites of the catalysts ^[34].

The visible light photocatalytic activity of CN-T samples for H₂ evolution was evaluated in an aqueous proton solution under visible light irradiation (λ > 420 nm). Chloroplatinic acid (Pt nanoparticles) and triethanolamine were used as co-catalyst and electron donors, respectively. The H₂ evolution rates (HER) for samples obtained at different temperatures are shown in Fig. 8(a). We observe that treatment at a higher temperature is beneficial for the photocatalytic activity of the CN catalyst. The HER of CN-600 reaches 340 μmol/h, nearly 2.2 times that of CN-500. In addition, the HER of the obtained CN samples obeys a direct correlation with the corresponding surface area. A 16-h recycling experiment with intermittent evacuation every 4 h was performed on CN-600 under visible light (λ > 420 nm), and acceptable (photo)-chemical stability of CN could be obtained.

Compared with the previously prepared CN nanosheets through various methods, the enlarged surface area of our prepared porous g-C₃N₄ nanosheets is not conspicuous. However, the HER of CN samples obtained in ammonia condensation on each unit of their specific surface area was greatly improved. Hence, in addition to the introduction of a porous structure, the improved polymerization degree of materials plays a significant role in the photocatalytic H₂ evolution activity. The conclusion can be derived that the novel g-C₃N₄ material with improved polymerization degree and nanopores exhibits excellent photocatalytic activity and stability for H₂ evolution.

Porous g-C₃N₄ nanosheets were successfully prepared through a one-step ammonia thermopolymerization method at high temperatures. The inhibitory effect of ammonia from NH₄SCN decomposition under high temperature optimizes the polymeric layer structure of the g-C₃N₄ nanosheet. The higher condensation temperature results in more nanopores, larger surface area, wider bandgap and a lower recombination rate of photo-induced carriers in the samples, thereby improving the photocatalytic activity for H₂ evolution under visible light irradiation. This route for the preparation of 2D porous CN-based polymer semiconductors nanosheets is easily-conducted, and the physicochemical properties of the samples are conveniently designed and adjusted.