The Multiple Effects of Precursors on the Properties of Polymeric Carbon Nitride

Polymeric graphitic carbon nitride (g-C 3 N 4 ) materials were prepared by direct pyrolysis of thiourea, dicyandiamide, melamine, and urea under the same conditions, respectively. In order to investigate the effects of precursors on the intrinsic physicochemical properties of g-C 3 N 4 , a variety of characterization tools were employed to analyze the samples. The photocatalytic activity of the sampleswas evaluated by the removal ofNO in gas phase under visible light irradiation.The results showed that the as-preparedCNT (from thiourea), CN-D (from dicyandiamide), CN-M (from melamine), and CN-U (from urea) exhibited significantly different morphologies andmicrostructures.The band gaps of CN-T, CN-D, CN-M, andCN-Uwere 2.51, 2.58, 2.56, and 2.88 eV, respectively. Both thermal stability and yield are in the following order: CN-M>CN-D>CN-T>CN-U. The photoactivity of CN-U (31.9%) is higher than that of CN-T (29.6%), CN-D (22.2%), and CN-M (26.8%). Considering the cost, toxicity, and yield of the precursors and the properties of g-C 3 N 4 , the best precursor for preparation of g-C 3 N 4 was melamine. The present work could provide new insights into the selection of suitable precursor for g-C 3 N 4 synthesis and in-depth understanding of the microstructure-dependent photocatalytic activity of g-C 3 N 4 .

In particular, g-C 3 N 4 material as a novel mental-free organic photocatalyst has triggered great attention in the fundamental and applied scientific researches due to its suitable electronic band structure, nontoxic nature, biocompatibility, high thermal and chemical stability, easily available at low cost, and amenability to chemical modification [11][12][13][14][15][16].Properties mentioned previously make it a promising organic photocatalyst for solar energy converting, organic photosynthesis, drug delivery, and environment remediation under visible light irradiation [17][18][19].
g-C 3 N 4 can be prepared by the direct pyrolysis of various organic precursors.Dong et al. prepared g-C 3 N 4 from thiourea and urea at 550 ∘ C for 2 h [20,21], respectively.Yan et al. obtained g-C 3 N 4 from melamine at 550 ∘ C for 4 h [22].Xu et al. prepared g-C 3 N 4 by directly heating dicyandiamide first at 350 ∘ C for 2 h and then 550 ∘ C for another 2 h [23], respectively.The different pyrolysis conditions for g-C 3 N 4 synthesis made the comparison of the precursors difficult.Until now, little information is known about the effects of precursors on intrinsic physicochemical properties of g-C 3 N 4 under the same pyrolysis conditions, including morphology, band gap, In this paper, g-C 3 N 4 materials were prepared by directly treating thiourea, dicyandiamide, melamine, and urea under the same condition (550 ∘ C for 3 h in air), respectively.The precursors are easily available in the chemical industry at low cost (see Table 1).Various characterization tools were utilized to analyze the effects of precursors on the intrinsic physicochemical properties and photocatalytic activity of g-C 3 N 4 .The results showed that the precursors had significant effect on the properties of g-C 3 N 4 .Considering the cost, toxicity, and yield of the precursors and the properties of g-C 3 N 4 , the best precursor was confirmed.

Experimental
2.1.Synthesis of g-C 3 N 4 .In a typical synthesis, 12 g of thiourea was put into a semiclosed alumina crucible with a cover and then heated to 550 ∘ C in a muffle furnace for 3 h at a heating rate of 10 ∘ C min −1 .Following the same procedure, g-C 3 N 4 samples were prepared by directly treating dicyandiamide, melamine, and urea, respectively.After the reaction, the alumina crucible was cooled to room temperature.The resultant g-C 3 N 4 were collected and ground into powder, respectively.The different g-C 3 N 4 prepared from thiourea, dicyandiamide, melamine, and urea were labeled as CN-T, CN-D, CN-M, and CN-U, respectively.

Characterization.
The crystal phases of the samples were analyzed by X-ray diffraction with Cu K radiation (XRD: model D/max RA, Rigaku Co., Japan).FT-IR spectra were recorded on a Nicolet Nexus spectrometer on samples embedded in KBr pellets.The morphology and structure of the samples were further examined by transmission electron microscopy (TEM: JEM-2010, Japan).The UV-vis diffuse reflection spectra were obtained for the dry-pressed disk samples using a Scan UV-vis spectrophotometer (UV-vis DRS: UV-2450, Shimadzu, Japan) equipped with an integrating sphere assembly, using BaSO 4 as the reflectance sample.Nitrogen adsorption-desorption was conducted on a nitrogen adsorption apparatus (ASAP 2020, USA).The thermal stability was detected by using thermogravimetry analysis (TG-DSC, Netsch STA 449F3) under N 2 gas atmosphere.All the samples were degassed at 150 ∘ C prior to measurements.

Visible Light Photocatalytic
Performance.The photocatalytic activity was investigated by removal of NO at ppb levels in a continuous flow reactor at ambient temperature.The volume of the rectangular reactor, which was made of polymethyl methacrylate plastics and covered with quartz-glass, was 4.5 L (30 cm × 15 cm × 10 cm).A 100 W commercial tungsten halogen lamp (General Electric) was vertically placed outside and above the reactor.Four minifans were used to cool the flow system.Adequate distance was also kept from the lamp to the reactor for the same purpose to keep the temperature at a constant level.For the visible light photocatalytic activity test experiment, a UV cut-off filter (420 nm) was adopted to remove UV light in the light beam.
For each photocatalytic activity test, two sample dishes (with a diameter of 12.0 cm) containing photocatalyst powder were placed in the center of the reactor.The weight of the photocatalyst used for each dish was kept at 0.1 g. g-C 3 N 4 sample was added into 30 mL of H 2 O and sonicated for 10 min, and then photocatalyst samples were prepared by coating aqueous suspension of the samples onto the glass dishes.The coated dish was pretreated at 60 ∘ C to remove water in the suspension and then cooled to room temperature before photocatalytic testing.
The NO gas was acquired from a compressed gas cylinder at a concentration of 100 ppm of NO (N 2 balance, BOC gas) with the National Institute of Standards and Technology (NIST) standard.The initial concentration of NO was diluted to about 600 ppb by the air stream supplied by a zero air generator (Thermo Environmental Inc., model 111).The relative humidity at indoor environmental condition is 40%-80%.The desired relative humidity in the present system is controlled at 50% in the gas flow which could simulate the indoor environmental conditions.The desired relative humidity (RH) level of the NO flow was controlled at 50% by passing the zero air streams through a humidification chamber.The gas streams were premixed completely by a gas blender, and the flow rate was controlled at 2.4 L min −1 by a mass flow controller.After the adsorption-desorption equilibrium was achieved, the lamp was turned on.The concentration of NO was continuously measured by a chemiluminescence NO analyzer (Thermo Environmental Instruments Inc., model 42c), which monitors NO, NO 2 , and NOx (NOx represents NO + NO 2 ) with a sampling rate of 0.7 L min −1 .The removal ratio () of NO was calculated as  (%) = (1 − / 0 ) × 100%, where  and  0 are concentrations of NO in the outlet steam and the feeding stream, respectively.The kinetics of photocatalytic NO removal reaction is a pseudofirst order reaction at low NO concentration as ln( 0 /) = , where  is the apparent rate constant.

Result and Discussion
The XRD patterns (Figure 1(a)) of the samples prepared from thiourea, dicyandiamide, melamine, and urea can be indexed to g-C 3 N 4 [24][25][26].The strongest interplanar stacking peaks of the conjugated aromatic systems at around 27.6 ∘ are indexed to graphitic materials as the (002) peak.The small angle peaks at around 13.0 ∘ can be indexed to (100) peak of graphitic materials, corresponding to the in-plane tri-striazine units which formed one-dimensional (1D) melon strands.Further observation implies (Figure 1(b)) that the typical (002) peak is 27.43 ∘ for CN-T, 27.47 ∘ for CN-D, 27.60 ∘ for CN-M, and 27.72 ∘ for CN-U, respectively, indicating that the interplanar distance tends to decrease and g-C 3 N 4 structure becomes more compact.The dense structure can be ascribed to the localization of the electrons and stronger binding between the layers [27,28].The dominant (002) diffraction peaks shifted toward higher diffraction angles for CN-U compared with CN-T, indicating that O-containing precursors (urea) could improve the polycondensation of g-C 3 N 4 .In addition, the intensity and breadth of the XRD patterns (Figure 1) are different, due to the presence of different nanostructures and morphologies of the as-prepared g-C 3 N 4 [29].
The chemical structures of dicyandiamide and melamine containa C=N bond which plays a key role in the formation of g-C 3 N 4 .However, both the S-containing thiourea and O-containing urea do not containa C=N bond.The fact indicates that the formation mechanisms and condensation degrees of thiourea and urea are different from that of dicyandiamide and melamine [20][21][22]30].
Figure 2(a) shows that the weak absorption at the 700-800 cm −1 region is assigned to the bending vibration mode of CN heterocycles and the characteristic out of plane bending vibration mode of the triazine units at 810 cm −1 for all the samples [31].All the samples reveal several bands in the 1200-1650 cm −1 region, which corresponds to the typical stretching vibration modes of the heptazine heterocyclic ring (C 6 N 7 ) units [32].
The stretching vibration modes of N-H and O-H at 3000-3500 cm −1 are also observed (Figure 2 precursors could improve the polycondensation of g-C 3 N 4 [34]. The morphologies and microstructures of CN-T, CN-D, CN-M, and CN-U were observed by TEM.As shown in Figure 3, all the samples consist of large amounts of packing layers with different sizes of nanosheets and nanoparticles, which exhibit obviously wrinkles and irregular shapes.Figure 3(a) shows that CN-T is composed of smooth and flat layers.Numerous large buckle nanosheets with aggregated structures which formed smooth layers in the CN-D sample can be seen in Figure 3(b).By observing carefully for CN-M (Figure 3(c)), the layers consist of dense and thick nanosheets with irregular shape.Figure 3(d) shows that CN-U is composed of smooth and thin layers with typical porous morphology and loose structure.The result clearly shows that CN-U prepared from urea favors the formation of typical porous structure.The fact indicates that the morphologies and microstructures of the resultant g-C 3 N 4 strongly depend on the different heteroatom-containing precursors.
As shown in Figure 4(a), all the samples exhibit excellent visible light absorption, and the absorption edges of the samples shift apparently to longer wavelengths from CN-U to CN-T.The band energies (  ) which can be estimated from the intercept of the tangents to the plots of (ℎ]) 1/2 versus photo energy (Figure 4(b)) are 2.51, 2.58, 2.56, and 2.88 eV for CN-T, CN-D, CN-M, and CN-U, respectively.The fact indicates that the precursors could affect the band gap and absorption edge of g-C 3 N 4 , which may be caused by the different local structures, defects, and degrees of condensation during the pyrolysis [8,11].
Figure 5 shows the nitrogen adsorption-desorption isotherms and corresponding curves of the pore size distribution for CN-T, CN-D, CN-M, and CN-U.The isotherms of CN-T (Figure 5(a)), CN-D (Figure 5(c)), and CN-M (Figure 5(e)) can be classified to type IV, which indicates the presence of mesopores.The CN-U exhibits type III behavior (Figure 5(g)), which can be ascribed to the weak adsorbent-adsorbent interaction [35].The specific surface areas and pore volumes of CN-T (23 m 2 /g and 0.14 cm 3 /g) and CN-U (153 m 2 /g and 0.40 cm 3 /g) are significantly higher than that of CN-D (18 m 2 /g and 0.09 cm 3 /g) and CN-M (14 m 2 /g and 0.06 cm 3 /g).The data illustrates that the heteroatoms of sulfur and oxygen play a key role in increasing the specific surface area and enlarging the pore volume during the condensation.The formation of H 2 S and CO 2 , release of NH 3 , and generation of additional H 2 O vapor during the pyrolysis of thiourea and urea favor the expansion of the packing layers and porous structure [32,35].Further observation implies that all the samples have micropores and mesopores.The interconnected porous network could mainly contribute to the formation of micropores and mesopores of CN-T (Figure 5     In order to understand the thermal stability of CN-T, CN-D, CN-M, and CN-U, the thermal analysis was carried out by using TG-DSC and the heating rate of 20 ∘ C/min under N 2 gas atmosphere.It can be found that the CN-T (Figure 6(a)), CN-D (Figure 6(b)), and CN-M (Figure 6(c)) became unstable when the heating temperature was above 550, 563, and 575 ∘ C, respectively.The exothermic peaks can be seen at 672.4 for CN-T, 683.1 for CN-D, and 649.5 ∘ C for CN-U, which should be attributed to the decomposition for the samples.Figure 6(d) shows that no significant weight loss of CN-U is recorded when the temperature is below 530 ∘ C, and complete decomposition of CN-U occurred at 750 ∘ C. Further observation reveals that there are two strong exothermic peaks at 550.1 and 682.4 ∘ C of CN-U, which can be ascribed to the sublimation and thermal decomposition [36], respectively.It should be noted that the thermal stability of the samples is different, due to the different degrees of condensation and the packing between the layers during the polymerization [37,38].
The photocatalytic performance of the as-prepared samples was further evaluated by removal of NO in gas phase in order to demonstrate their potential ability for indoor air purification under visible light irradiation at room temperature (Figure 7).There are four reactions of the photocatalytic materials, which involved that NO reacted with the photo-generated reactive radicals and produced HNO 2 and HNO 3 displayed in the following [39]: Figure 7(a) shows the variation of NO concentration (/ 0 %) with irradiation time over the samples under visible light irradiation ( > 420 nm).Here,  0 is the initial concentration of NO, and  is the concentration of NO after photocatalytic reaction for time .
As shown in Figure 7(a), the concentration of NO for all samples decreased rapidly due to the photocatalytic degradation in 5 min.However, the reaction intermediates, and final products generated during irradiation may occupy the active sites of photocatalyst, which result in the decrease in activity.After 40 min irradiation, the removal rates and apparent rate constants of CN-T (29.2% and 0.310 min −1 ) and CN-U (32.2% and 0.384 min −1 ) are higher than that of CN-D (22.2% and 0.079 min −1 ) and CN-M (26.2% and International Journal of Photoenergy International Journal of Photoenergy 0.298 min −1 ), respectively.Generally, CN-T, CN-D, and CN-M with narrower band gaps than CN-U (see Table 2) are favorable for photocatalytic reaction, due to the enhanced visible light absorption with more available photogenerated electron hole.In fact, the high surface areas and large pore volumes can enhance the adsorption of reactants and diffusion of the reaction products and provide more efficient active sites for photochemical reaction [40,41].Thus, the surface area and pore volume of CN-U play a more significant role in NO degradation than the electronic properties.However, it is interesting to find that the CN-M exhibits higher photocatalytic activity than that of CN-D. Figure 4(b) shows that the band energy of CN-M is narrower than that of CN-D, which can enhance visible light absorption in CN-M sample.On the other hand, the enhanced condensation of CN-M can effectively reduce the number of structure defects (e.g., uncondensation groups of -NH 2 and -NH) which always capture the electrons or holes to prevent the following photoredox reaction [8].The fact indicates that the photocatalytic efficiency strongly depends on the surface area, pore volume, band energy, and the degree of condensation of g-C 3 N 4 .
To ensure the large-scale application, four evaluation indicators were proposed for screening the most suitable precursors, including the cost, toxicity, and yield of the precursors and the removal rate for the contaminant.As shown in Tables 1 and 2, melamine exhibits the middle cost, toxicity, and removal ratio for the degradation of NO.However, it possesses the highest yield among the four precursors.Based on the advantages analysis previously mentioned, the best precursor is melamine.

Table 2 :
The  BET , pore volume, band gap, removal ration, initial rate constant , yield, and thermal stability of CN-T, CN-D, CN-M, and CN-U.