Improving the Visible-Light Photocatalytic Activity of Graphitic Carbon Nitride by Carbon Black Doping.

Hydrogen production by water splitting and the removal of aqueous dyes by using a catalyst and solar energy are an ideal future energy source and useful for environmental protection. Graphitic carbon nitride can be used as the photocatalyst with visible light irradiation. However, it typically suffers from the high recombination of carriers and low electrical conductivity. Here, we have developed a facile mix-thermal strategy to prepare carbon black-modified graphitic carbon nitrides, which possess high electrical conductivity, a wide adsorption range of visible light, and a low recombination rate of carriers. With the help of carbon black, highly crystallized graphitic carbon nitrides with built-in triazine and heptazine heterojunctions are obtained. Improved photocatalytic activities have been achieved in carbon black-modified graphitic carbon nitride. The dye removal rate can be three times faster than that of pristine graphitic carbon nitride and the photocatalytic H2 generation is 234 μmol h-1 g-1 under visible light irradiation.

Introduction

Clean hydrogen generation from water splitting with sunlight plays an important role in developing sustainable green energy. Three factors are crucial to achieving highly efficient photocatalysts: namely, an appropriate band gap to absorb maximum solar radiations, high electrical conductivity, and a low recombination rate of carriers.[1] In addition, the materials should be nontoxic, abundant, and easily processable into desired shapes. To date, various oxide,[2,3] sulfide,[4,5] and oxynitride[6,7] semiconductors have been applied as photocatalysts. However, they typically suffer from the limited concentration of active sites and the toxicity of heavy metals. Graphitic carbon nitride (g-C3N4) has been widely studied for photocatalytic water splitting and organic pollutant degradation due to its chemical and thermal stability and nontoxicity. However, the efficiency of pristine g-C3N4 under visible light is relatively low due to the marginal absorption of visible light and grain boundary effects.[8] Creation of heterojunctions with other semiconductors is a feasible strategy to promote the photocatalytic activity of g-C3N4.[9] Moreover, attempts to improve the photocatalytic efficiency of g-C3N4 have also focused on heteroatom doping with S,[10,11] Ag,[12−15] K,[16] C,[17,18] and P,[18,19] which result in the modification of the electronic structure and therefore improvement of the photocatalytic performance. Nevertheless, heteroatom doping entails a relatively high cost with the complicated process. Moreover, some heavy metal dopants probably leak out from the catalyst, causing additional pollutants.

Due to the high electronic transportability, carbon allotropes (e.g., carbon quantum dots,[17] graphene[20,21]) are optimal materials to accelerate electron transfer from the g-C3N4 photocatalyst, to improve its separation of electron–hole pairs and to enhance the photocatalytic activity. In addition, amorphous g-C3N4 with a disordered structure and an increased surface area has been reported numerously in improving the photocatalytic activity.[22−26] Nevertheless, there are limited reports on g-C3N4 with high crystallinity and condensation and at the same time with improved photocatalytic activity. Herein, with the introduction of carbon black (CB), we succeeded in a one-step facile mix-thermal strategy to obtain highly crystalline g-C3N4. The triazine and (tri-s-triazine) heptazine phases of g-C3N4 are generated simultaneously and the efficient heterostructure is built between them in the g-C3N4 matrix. According to the literature, a conjugated heterostructure is also built in carbon nitride when it is prepared from a high-nitrogen-content precursor.[27] Benefiting from the high surface area-to-volume ratio and being an excellent electricity conductor, the heterostructure has high electrical conductivity, a wide adsorption range of visible light, and a low recombination rate of carriers, and therefore displays high photocatalytic activity. In this study, we describe several such materials made from the pyrolysis of urea mixed with CB, during which both crystallinity and condensation degrees of g-C3N4 are improved. The prepared CB-modified g-C3N4 shows excellent photocatalytic activities, including a higher methylene blue (MB) removal ratio of 70.2% as compared with pristine g-C3N4 (26.5%) and a higher photocatalytic H2 evolution rate of 234 μmol h–1 g–1 compared with pristine g-C3N4 (163 μmol h–1 g–1) under visible light irradiation.

Results and Discussion

Structure and Composition of Photocatalysts

Figure a shows the X-ray diffraction (XRD) patterns of pristine g-C3N4 and CB-modified g-C3N4. The peak at around 13° corresponds to the (100) diffraction plane, which indicates the in-plane repeated units of tri-s-triazine. For 1500CNC and 3000CNC, peaks at around 18° can be indexed to the (100) peak for the triazine phase in the modified g-C3N4.[28,29] Therefore, the XRD patterns for nCNCs reflect the coexistence of triazine and (tri-s-triazine) heptazine phases in g-C3N4 after the CB modification. The peaks at around 27° can be indexed to the (002) diffraction plane in the c-axis, which represents the periodic graphitic stacking of the conjugated system with an interlayer distance around 0.3 nm.[10,30] The (002) and (100) peaks for the tri-s-triazine phase are more intense and sharper when CB is introduced in the condensation of urea, which suggests that the crystallinity of g-C3N4 is improved after the CB modification. Therefore, in this work, g-C3N4 with the triazine and (tri-s-triazine) heptazine heterojunction is built in a one-pot synthesis with the help of CB (Figure S1 shows XRD patterns for CNC with different urea and CB ratios). Moreover, both (100) and (002) peaks shift to the higher degree for 1500CNC and 3000CNC as compared with that of CN, indicating narrower intralayer and interlayer distances. The interlayer distances of aromatic units from the (002) peak are listed in Table S1. Note that most reported dopants in the C–N cycles result in an increase of the interlayer distance due to the larger atomic radius.[16] Nevertheless, the decrease of the intralayer and interlayer distance in our nCNCs is presumably caused by the covalent bonding between interlayer dopant carbon atoms and the C, N on the aromatic rings.

Figure 1

(a) XRD patterns of CN, 1500CNC and 3000CNC; (b) Fourier transform infrared (FTIR) spectra of CN and 1500CNC and 3000CNC.

The Fourier transform infrared (FTIR) spectra for the CN and CNC samples are shown in Figure b. CNCs and CN display a similar FTIR pattern, indicating that the introduction of CB does not alter the basic chemical structure of CN. The slightly broader vibration bands at around 3200 cm–1 indicate a complex environment for the O–H and N–H stretching vibrations on the edges of g-C3N4 layers, suggesting more compact interlayer interaction.[29] The vibration bands at 1000–1750 cm–1 are attributed to the CN stretching vibrations in the CN heterocycles.[31] The strong band at 810 cm–1 is ascribed to the vibration mode of the triazine-based unit. The comparison of all nCNCs is provided in Figure S2.

The X-ray photoelectron spectroscopy (XPS) spectra were used to determine the compositions and chemical states of the constituents in the samples. The CN and nCNCs mainly consist of C and N elements with a small amount of oxygen as shown in the survey spectra in Figure S3. The higher atomic concentration of C in 3000CNC and 1500CNC suggests the successful doping of CB in 3000CNC and 1500CNC.

Figure a–c are the high-resolution spectra of C, N, and O, respectively. In Figure a, all samples exhibit two bonding states of carbon species, which are evident with the C 1s binding energies of around 284.6 and 289 eV. In Figure b, the main peaks of N 1s binding energies are located around 399 and 401 eV for all samples. The broader N 1s, C 1s binding energy areas for samples 3000CNC and 1500CNC as compared with CN imply the complex bonding types, indicating the formation of a mixture of crystalline phases of g-C3N4. The higher O 1s bonding energy for samples 3000CNC and 1500CNC implies that the absorbed O2 and surface −OH and C=O generated on the edge of the CN cycle are located in a more intense chemical environment (Figure c,f). The high-resolution C 1s XPS spectrum of sample 1500CNC in Figure d shows three deconvolution peaks at 289.0, 286.2, and 284.6 eV, reflecting sp2-bonded carbon (N–C=N), C–O, and graphitic carbon (C–C), respectively. Figure e shows the peak deconvolution of the N 1s XPS spectrum for 1500CNC. Three peaks at 399.2, 400.7, and 401.6 eV are ascribed to sp2-hybridized aromatic N bonded to carbon atoms (C=N–C), the tertiary N bonded to carbon atoms in the form of N–(C)3, and N–H side groups, respectively.[31] Both the binding energies of sp2-bonded C and N shift to higher energies for 3000CNC and 1500CNC as compared with CN and reported g-C3N4.[32] The higher binding energy presumably comes from the incorporation of intercalated CB and the CN heterocycles. CB in the g-C3N4 matrix bonds with adjacent C and N atoms, which results in the increase of the bonding energy in the in-plane triazine and/or heptazine units. On the other hand, the bonding force from the graphitic interlayer is supposed to drag the g-C3N4 layers and to make them into more compact layers. This hypothesis from XPS results is consistent with the shrinking of the crystal lattice space for g-C3N4 as confirmed by the XRD (Tables and S1). The shifted XRD and XPS patterns have also been reported in other elemental modifications of g-C3N4.[16,19]

Figure 2

(a) High-resolution C 1s, (b) N 1s, and (c) O 1s XPS spectra (A, B, and C refer to CN, 1500CNC, and 3000CNC, respectively) and the deconvolution of high-resolution XPS spectra for C 1s (d), N 1s (e), and O 1s (f) for sample 1500CNC.

Table 1

Brunauer–Emmett–Teller (BET) Surface Area, Pore Volume, Band Gap and the Interplanar Distance for Samples and the Respective Pseudo-First-Order Rate Constants for MB and OrgII Photodegradation over Photocatalysts

samples	SBET (m2 g–1)	pore volume (cm3 g–1)	band gap (eV)	K (MB) (10–3 min–1)	K (OrgII) (10–3 min–1)	d(002) (nm)
CN	128.6	0.598	2.89	1.07	5.34	0.322
3000CNC	77.2	0.341	2.95	3.05	6.07	0.314
1500CNC	54.9	0.152	2.89	2.94	2.25	0.314
1000CNC	61.7	0.165	2.95	2.36	0.861	0.315
600CNC	69.4	0.170	2.90	2.24	0.693	0.316
200CNC	87.2	0.350	2.70	1.97	0.125	0.317

Microstructures and Texture Properties of Photocatalysts

The transmission electron microscopy (TEM) morphologies of 1500CNC and CN are presented in Figure a–d, respectively. Both samples exhibit the typical layered platelet-like surface morphology with wrinkling edges, where 1500CNC has more compact and intact flakes. On the contrary, CN possesses interstitial pores at the edge of the flakes. The porous structure of CN is generated due to the release of various gases during the pyrolysis of urea, including NH3, CO2, H2O, etc. For 1500CNC, CB is assumed to exert its absorption property, accommodating the gases generated during the heating process. Moreover, it serves as the condensation site for the generated intermediate gases. Therefore, 1500CNC is more intact and compact than CN, corresponding to the N2 adsorption–desorption results. As shown in Table , CN and 1500CNC have specific surface areas (SSAs) of 128.6 and 54.9 m2 g–1, respectively. However, 1500CNC has a higher surface area-to-volume ratio (3.6 × 106 cm–1) than CN (2.2 × 106 cm–1).

Figure 3

TEM images for 1500CNC (a) and CN (b); high-resolution TEM (HRTEM) images with inset selected area electron diffraction (SAED) patterns for 1500CNC (c) and CN (d).

The crystalline characteristics of both samples are also characterized by the selected area electron diffraction (SAED), which are shown in insets in Figure c,d. Both SAED patterns show a full outer diffraction ring, which can be indexed as the (002) reflection of g-C3N4, and an inner weak diffraction ring, which can be indexed as the reflection of the (100) plane from g-C3N4, indicating that both samples consist of extremely small crystallites with random orientation. The clearer SAED pattern for 1500CNC than for CN suggests the improved crystallinity, which is consistent with the XRD data.

The adsorption–desorption isotherms (Figure S4) for all samples are type IV (BDDT Classification), suggesting the presence of mesopores (2–50 nm).[33] The hysteresis loops are type H3 (IUPAC classification), indicating the formation of slit-shaped pores from the aggregates of plate-like particles.[34,35] Overall, the SSAs and pore volumes are less for nCNCs compared with CN. Since 1500CNC and CN have a similar particle size based on the TEM, the Brunauer–Emmett–Teller (BET) results indicate that nCNCs should be more compact than CN, which is due to the improved condensation degree of urea resulting from the CB modification. The SSAs and pore volumes (listed in Table ) decrease with an increase in the amount of CB. However, with further increase in the CB amount, the SSAs and pore volume gradually increase. A small amount of dopant is supposed to improve the polymerization. However, further increase in the amount of CB will result in a higher surface area of nCNCs due to the excess porous CB. nCNCs also exhibit superior thermostability compared with pristine g-C3N4. The thermogravimetric differential scanning calorimetry analysis (TG-DSC) for photocatalysts is shown in Figure S5.

Optical and Electrical Properties

UV–vis diffuse reflectance absorption spectra (DRS) of the obtained g-C3N4 samples are presented in Figure a. CB-modified g-C3N4 and pristine CN show typical absorption patterns of semiconductors with absorption edges at ∼445 nm, indicating that samples can absorb solar energy in the blue-light region.[36] The incorporation of CB into the g-C3N4 matrix increases the UV–vis absorption over the entire wavelength range. The absorption tail in the longer wavelength has been reported in most doped CN.[37,38] Moreover, its intensity increases along with the amount of CB (Figures a and S6). The bandgap energies for all of the photocatalysts are in the range of 2.70–3.00 eV as presented in Figures b and S6, and listed in Table . 3000CNC has the largest band gap, which can be ascribed to the change in the valence band of g-C3N4 due to the increased bonding energy in in-plane units. It has been calculated and predicted that seven phases of C3N4 exist, which are α-C3N4, ß-C3N4, cubic C3N4, pseudocubic C3N4, g-h-triazine, g-o-triazine, and g-h-heptazine, with bandgap values of 5.49, 4.85, 4.30, 4.13, 2.97, 0.93, and 2.88 eV, respectively.[39,40] As we can see, the band gap for 3000CNC is 2.95 eV (between 2.97 and 2.88 eV), indicating the mixture phases of g-C3N4 with triazine and heptazine (tri-s-triazine). The CB in the matrix leads to the formation of more sub-band states in the band gap,[17] which results in the decrease in the band gap.

Figure 4

(a) UV–vis diffuse reflectance spectra and (b) estimated bandgap energies of CN, 3000CNC, and 1500CNC; (c) photoluminescence (PL) spectra of CN, 3000CNC, 1500CNC, 1000CNC, 200CNC, and 600CNC; (d) electrochemical impedance spectroscopy (EIS) of CN, CB, and 3000CNC; (e) XPS valence band spectra (a and b refer to CN and 3000CNC, respectively); (f) band structures of CN and 3000CNC.

The photoluminescence (PL) spectra (Figure c) are usually used as an indication of the recombination or separation efficiency of the photogenerated charge carriers.[41] A faster recombination of electron–hole results in a more intense PL spectrum.[42] All of the samples show only one dominant luminescence peak. This strong peak at 450 nm for pristine CN is attributed to the near-band-edge emission, which is related to the direct recombination of excitons through an exciton–exciton collision process.[43] The intensity of the emission peak appreciably decreases when g-C3N4 is modified with a small amount of CB, such as 3000CNC and 1500CNC. A comparative study of the overall intensity of all samples indicates a substantial inhibition of charge carriers’ recombination in the samples 3000CNC and 1500CNC. Low intensities of PL for 1000CNC, 600CNC, and 200CNC can be ascribed to the absorption of incident light by the extra CB. The photocatalytic activity is related to the concentration of free charge carriers generated in the photocatalytic process. Therefore, 3000CNC and 1500CNC are relatively effective for the separation of charge carriers, indicating the potentially higher photocatalytic activity.

In addition, electrochemical impedance spectroscopy (EIS) measurement was used to investigate the charge transfer resistance and the separation efficiency of the photoinduced charge carriers. Figure d demonstrates the diameter of the Nyquist semicircle for the 3000CNC, CN, and CB electrodes. The Nyquist plot semicircle of 3000CNC is located between those of pristine CN and CB. A small arc radius implies a high efficiency of charge transfer and separation.[30,44] Therefore, the EIS measurement is consistent with PL spectra, indicating that CB-modified g-C3N4 has an enhanced separation and transfer efficiency of charge carriers.

Adsorption and Photocatalytic Activities of Catalysts

The photocatalytic activities of the prepared samples were evaluated by monitoring the degradation of methylene blue (MB) and Orange II sodium salt (OrgII) solution under visible light irradiation (wavelength ≥ 400 nm). The adsorption equilibrium was established in 1 h before the photocatalytic test. Compared with CN, the adsorption properties of nCNCs are improved for cationic dye MB (Figure A), but inferior for anionic dye OrgII (Figure B). Although with the highest SSA and pore volume, due to the positive charge,[34] pristine g-C3N4 shows an inferior adsorption toward cationic dye MB. Exceptionally, 200CNC exhibited substantial adsorption toward both MB and OrgII, which can be attributed to the amount of existing CB in the g-C3N4 matrix.

Figure 5

Comparison of MB (A) and OrgII (B) adsorption and photodegradation in water under visible light over CN and nCNCs; first-order kinetic plots for the photodegradation of MB (C) and OrgII (D) over CN and nCNCs.

The MB removal ratios are 63.3, 70.2, and 26.5% on 1500CNC, 200CNC, and CN, respectively. However, the OrgII removal ratios are 80.1, 50.1, and 87.8% on 3000CNC, 1500CNC, and CN, respectively. The photodegradation experimental data are fitted by a first-order kinetic model as expressed by the equation ln(C/C0) = −kt, with the value of the rate constant k giving an indication of the activity of the photocatalysts (listed in Table ).[45] The photoactivity has been improved obviously for the degradation of MB over nCNCs, with the highest k value of 3.05 × 10–3 min–1 for 3000CNC (Figure C). Although CN still exhibits superior removal toward OrgII, 3000CNC suppresses it in photocatalytic activity with the highest k value of 6.07 × 10–3 min–1 (Figure D). The UV–vis absorption spectra of the MB solution and OrgII solution after different periods of adsorption and photocatalytic decomposition are illustrated Figure S7.

We also conducted photocatalytic hydrogen evolution from the water splitting for 3000CNC and CN under visible light (λ ≥ 420 nm) irradiation (Figure S8). In our system, electron donor triethanolamine (TEOA) works as a sacrifice to accelerate the hole oxidation process.[46] In situ photodeposited Pt acts as a cocatalyst to accelerate the surface electron transfer and hydrogen reduction, promoting photocatalytic activity.[47] The total hydrogen production of 3000CNC after 6 h reaction is 14.03 μmol, which is 1.4 times higher than that of CN (9.82 μmol). The H2 generation rate of 3000CNC reaches 234 μmol h–1 g–1, which is higher than that of CN (163 μmol h–1 g–1). Therefore, 3000CNC not only exhibits superior photocatalytic oxidation but also higher photocatalytic reduction than CN.

Proposed Mechanism under Visible Light Irradiation

It is accepted that •OH, •O2– and photogenerated holes are the main reactive species in photocatalytic oxidation, which are determined by the surface redox potential of the photocatalyst. The XPS valence band spectra of CN and 3000CNC are shown in Figure e. The valence band maxima (VBM) of CN and 3000CNC are 1.45 and 1.78 eV, respectively. The higher VBM value suggests a stronger oxidation of the photogenerated holes in 3000CNC. Considering that the band gaps of CN and 3000CNC are 2.89 and 2.95 eV (Figure b), the conduction band minima (CBM) are located at −1.44 and −1.17 eV, respectively. Therefore, the band structures of CN and 3000CNC are as provided in Figure f. The potential energies of the VBM for both CN and 3000CNC are lower than those of OH–/•OH (1.99 eV), whereas the potential energy of CBM is more negative than that of O2/•O2– (−0.28 eV). Therefore, the photogenerated holes cannot directly oxidize OH– into hydroxyl radicals •OH, whereas the electrons can reduce the O2 into superoxide radicals •O2–. Hence, the main reactive species in the photocatalytic degradation could be •O2– and photogenerated holes. Different electronic structures for the triazine and tri-s-triazine phases of g-C3N4 result in the broadened band gap and increased VBM for 3000CNC. The heterojunction between them facilitates the transfer and separation of carriers. Moreover, the CB in the g-C3N4 matrix plays a key role in conducting the photoelectron and reducing the recombination of electrons and holes. Therefore, improved photocatalytic activity has been realized in 3000CNC.

Conclusions

We modified graphitic carbon nitride via pyrolysis with a mixture of urea and carbon black in different ratios, to obtain graphitic carbon nitrides with built-in triazine and heptazine heterojunctions. The crystallization and condensation degrees of graphitic carbon nitride are improved, which is attributed to the carbon black working as the condensation site for the generated intermediate gases. The doping of carbon atoms in the graphitic carbon nitride matrix not only enhances light absorption but also quenches the recombination of charge carriers. Sample 200CNC exhibits a 70.2% removal ratio for MB, which is higher than the 26.5% of CN. Sample 3000CNC shows the most superior photocatalytic activity with the highest kinetic rate constant k in degradations of both MB and OrgII. It also exhibits a higher hydrogen liberation rate than pristine g-C3N4. The carbon black-modified graphitic carbon nitride with unique physicochemical structural features produced through a facile process has a promising application in energy conversion and environmental remediation.

Experimental Section

Chemicals and Materials

Urea, the cationic dye methylene blue (C16H18ClN3S), and the anionic dye Orange II sodium salt (C16H11N2NaO4S), purchased from Sigma-Aldrich, were of analytical grade and used without further purification. TEOA (AR, 98%) was purchased from Aladdin, China. Hydrochloroplatinic acid H2PtCl6 (ACS, 99.95%, Pt 3 g L–1) and carbon black were purchased from Alfa Aesar. Milli-Q water (ultrapure water) was utilized in all experiments.

Synthesis of g-C3N4 and Carbon Black-Modified g-C3N4

The g-C3N4 used in this study was prepared by urea pyrolysis at 550 °C for 3 h in an alumina crucible according to the literature.[48] In brief, a given amount of urea was dried at 80 °C for 24 h, and then calcined at 550 °C for 3 h (the heating rate was 2 °C min–1). After naturally cooling down, buff-colored powders were collected and washed with 10% HNO3 solution and ultrapure water three times by centrifugation. g-C3N4 samples were obtained and named as CN for short in this work.

The carbon black-modified g-C3N4 samples were prepared by heating the mixture of urea and CB at 550 °C for 3 h in an alumina crucible. In brief, the mixtures of urea and CB at different ratios were dried at 80 °C for 24 h independently, followed by calcination at 550 °C for 3 h (the heating rate was 2 °C min–1). After cooling down naturally, dark gray powders were collected and washed with 10% HNO3 solution and ultrapure water three times by centrifugation. For abbreviation, the samples are denoted as nCNC, where n is the weight ratio of urea to CB.

Characterization

XRD patterns were analyzed on a Shimadzu diffractometer (XRD-6000, Tokyo, Japan) in the reflection mode under Cu Kα radiation (λ = 0.15418 nm) at a scanning rate of 0.02° s–1 with 2θ ranging from 5 to 80°. FTIR spectra were recorded on a VERTEX 80/80 v FTIR spectrometer over a wavenumber range of 4000–400 cm–1. Surface chemical states for samples were investigated by XPS, which were measured on a Kratos Axis ULTRA X-ray photoelectron spectrometer with a 165 mm hemispherical electron energy analyser and monochromatic Al Kα X-ray source (1486.6 eV) at 225 W (15 kV, 15 mA) with a charge neutralizer and adventitious C 1s peak (284.6 eV) as the reference. The TEM images were taken using a JEOL JEM-1010 TEM. The HRTEM images and SAED patterns were obtained with an analytical transmission electron microscope (The Philips Tecnai F20 FEG-S/TEM). The SSA was calculated using the BET method, and the pore volumes were calculated using the Barrett–Joyner–Halenda method from the adsorption isotherms of N2 measured at 77 K on a Micromeritics TriStar II 3020. The TG-DSC data were obtained on a Shimadzu DTG-60A analyzer. In detail, 5–10 mg dry sample was loaded in a Pt crucible without a lid and scanned at a rate of 2 or 10 °C min–1. DRS was measured using a Shimadzu UV-2600 spectrometer equipped with an integrating sphere ISR-3100 using BaSO4 as the reference. Fluorescence spectra were obtained using a photoluminescence spectrometer (FLS 920, Edinburgh Instruments). EIS was measured in 6 M KOH aqueous solution with a sinusoidal ac perturbation of 5 mV over a frequency range of 0.1–1 × 106 Hz.

Photocatalytic Performance Evaluation

Adsorption and Photodegradation Experiment

The dyes’ adsorptions were measured on a batch mode. For the adsorption of OrgII over the prepared samples (including CN and 3000CNC, 1500CNC, 1000CNC, 600CNC, and 200CNC), 10 mg solid sample was added to 50 mL of an OrgII solution (the initial concentration was 10 mg L–1) in the dark under stirring. At a certain time interval, 5 mL of solution was extracted by a syringe and filtered through a membrane (0.45 μm) and then the solution was analyzed using a Shimadzu UV-2600 UV–vis spectrophotometer. After the adsorption equilibrium, the photocatalytic degradation of OrgII was carried out in an open thermostatic photoreactor under visible light irradiation. The visible light source was obtained from a 200 W mercury lamp by adding a 400 nm cut filter.

For the adsorption and photocatalytic degradation of MB under visible light, the initial concentration of MB was the same as OrgII, and the test processes were the same as mentioned above.

H2 Generation from Water Splitting

For H2 generation, 10 mg of the obtained samples, 10 mL triethanolamine (TEOA), and 100 μL H2PtCl6 solutions (Pt, 3 g L–1) were mixed and diluted with ultrapure water to 100 mL. After sonication for 10 min, the solution was transferred to the reaction cell and vacuumed under stirring before being irradiated with visible light. A Xe lamp (CEAULIGHT, CEL-HXF30) was the light source; a 420 nm cut filter (CEAULIGHT, CEL-UVIRCUT420) was used (light intensity in the bottom center was adjusted to 100 mW cm–2). The generated hydrogen was monitored by gas chromatography (TCD, CEAULIGHT, GC-7920) every 30 min with ultrapure N2 worked as the carrier gas.