Tri-s-triazine-Based Crystalline Graphitic Carbon Nitrides for Highly

May 11, 2016 - Polymeric carbon nitride for solar hydrogen production. Xiaobo Li , Anthony F. Masters , Thomas Maschmeyer. Chemical Communications 201...
0 downloads 4 Views 7MB Size
Download PDF
Research Article pubs.acs.org/acscatalysis

Tri‑s‑triazine-Based Crystalline Graphitic Carbon Nitrides for Highly Efficient Hydrogen Evolution Photocatalysis Lihua Lin, Honghui Ou, Yongfan Zhang, and Xinchen Wang* State Key Laboratory of Photocatalysis on Energy and Environment, College of Chemistry, Fuzhou University, Fuzhou 350002, People’s Republic of China S Supporting Information *

ABSTRACT: Graphitic carbon nitride (g-CN) is an emerging metal-free photocatalyst for solar energy conversion via water splitting and CO2 fixation. Herein, we used preheated melamine as a starting material in combination with the salt melt method to synthesize a crystalline tri-s-triazine-based g-CN. The as-obtained sample exhibited high stability and photocatalytic activity toward hydrogen and oxygen production from water splitting. In addition, by adding phosphate to mimic natural photosynthetic environment, the apparent quantum yield (AQY) for the hydrogen production reached 50.7% at 405 nm, which is the highest value ever reported for conjugated carbon nitride polymers in hydrogen evolution photocatalysis. The results of this study demonstrate that crystalline covalent tri-s-triazine frameworks hold great promise for solar energy applications.

KEYWORDS: carbon nitride, salt melt, crystalline, photocatalysis, water splitting

■

INTRODUCTION Since the discovery of water photolysis on a TiO2 electrode by Fujishima and Honda in 1972, photocatalytic water splitting using sunlight, water, and a semiconductor to produce hydrogen gas has attracted much attention because it promises a sustainable and clean alternative that addresses global energy and environment issues.1−6 The key to achieving solar hydrogen production is to develop stable, efficient, and inexpensive photocatalysts that are capable of working in the visible spectrum, which occupies ca. 50% of the incoming solar irradiation on earth. In the past few decades, much effort has been devoted to the development of visible light responsive photocatalysts, including modified TiO2 and other metal oxides, metal (oxy)sulfides. and metal (oxy)nitrides.7−16 Meanwhile, other solar energy transducers have also been studied, including conjugated polymers with facile processability and finely tunable electronic structures by using organic synthesis methods.17,18 However, most traditional conjugated semiconductors are subject to chemical corrosion by light irradiation in the presence of water and air.19−22 Graphitic carbon nitride (g-CN) polymers, which are the most stable allotrope of binary carbon nitride materials under ambient conditions (Figure 1a,c), were successfully employed as metal-free visible light photocatalysts in 2009, due to their unique physicochemical properties.23 Therefore, this photocatalyst has been actively investigated in solar to chemical energy conversion processes, such as hydrogen production from water, CO2 conversion, and organic-selective synthesis.24−29 However, the pristine g-CN synthesized by the traditional thermal-induced polycondensation of nitrogen© 2016 American Chemical Society

containing precursors typically exhibits a lower crystallinity and moderate photocatalytic activity, which is most likely due to the kinetic problems encountered during the bulk condensation process. The incomplete deamination of the precursors primarily results in the formation of a tri-s-triazinebased melon structure which consists of in-plane infinite onedimensional chains of NH-bridged melem oligomers (Figure 1e). The chains adopt a zigzag motif, and the different chains were connected via hydrogen bonds.30 The existence of hydrogen bonds in the covalent carbon nitride framework may block electron conduction across the plane and lead to low conductivity. Therefore, the synthesis of a fully condensed, crystalline g-CN is desirable. In principle, the improved crystallinity of the photocatalysts would enhance the photocatalytic activity due to the defects (acting as recombination centers for electron−hole pairs) in the structure being minimized.31,32 To address this problem, a new approach was developed to synthesize crystalline g-CN from the self-condensation of dicyandiamide using a salt melt as the high-temperature solvent.33 This work opened up new opportunities for synthesizing crystalline g-CN. After that, many triazine-based crystalline g-CN were synthesized.34−38 Despite the high crystallinity, the use of a salt melt typically altered the structure of the carbon nitride polymer to a triazine-based motif (poly(triazine imide), PTI; Figure 1g). Therefore, the synthesis Received: March 31, 2016 Revised: May 6, 2016 Published: May 11, 2016 3921

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

Figure 1. Structure models and corresponding electronic properties of (a, b) triazine-based g-C3N4, (c, d) tri-s-triazine-based g-C3N4, (e, f) melon, and (g, h) PTI. The gray, blue, and white spheres denote the C, N, and H atoms, respectively.

of tri-s-triazine-based crystalline g-CN using an ionothermal method is still unprecedented. In comparison with triazinebased PTI, this tri-s-triazine-based g-CN (Figure 1c) is expected to be photocatalytically more promising due to its extended and fully condensed conjugation structure, which stabilizes the πelectron system for fast charge mobility and improves the sunlight-harvesting capability due to the decreased band gap. Herein, we use preheated melamine as the tri-s-triazine-based precursor to synthesize crystalline g-CN semiconductors with tri-s-triazine subunits using KCl/LiCl salts. The as-obtained sample exhibited high crystallinity in comparison to bulk g-CN, and the crystal structure is quite different from that of the triazine-based PTI. The results from the photocatalytic experiments demonstrate that the as-synthesized sample exhibits high activity toward hydrogen production with an AQY of 50.7% at 405 nm, which is thus far the highest reported value for hydrogen production by g-CN-based photocatalysts,

whereas the PTI exhibits rather low activity toward photocatalytic hydrogen production. In addition, a high photostability was maintained during prolonged operation. This study highlights the important influence of crystallinity and conjugated subunits on carbon nitride photocatalysis and provides a new approach for the design and synthesis of new types of g-CN crystals with stable tri-s-triazine tectons for polymeric photoredox catalysis.

■

EXPERIMENTAL SECTION Materials. Melamine (C3H6N6, 99%), KCl (99%) and LiCl (99%) were purchased from Alfa Aesar Chemicals Co. Ltd. (China) and used without further purification. Synthesis of Samples. Melamine (8 g) was heated to 500 °C for 4 h at a rate of 12 °C min−1 in a muffle furnace in an air atmosphere. After that, 600 mg of the preheated sample was ground with KCl (3.3 g) and LiCl (2.7 g) in a glovebox. Then, 3922

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis the mixture was heated to 550 °C for 4 h under a N2 atmosphere (2 L min−1) in a muffle furnace. After it was cooled to room temperature, the product was washed with boiling deionized water several times and collected by filtration, followed by drying at 60 °C under vacuum. This sample is denoted as g-CN-1. For comparison, 1.2 g of melamine was heated to 550 °C for 4 h in an air atmosphere, which was referred to as bulk g-CN. A total of 600 mg of the 500 °C preheated sample was directly heated to 550 °C for 4 h without the KCl/LiCl molten salt, and this sample is referred to as g-CN-2. Another reference sample was prepared by mixing 1.2 g of melamine with KCl (6.6 g) and LiCl (5.4 g) in a glovebox; the mixture was then heated to 550 °C under a N2 atmosphere (2 L min−1) in a muffle furnace. After it was cooled to room temperature, the product was washed with boiling deionized water several times and collected by filtration followed by drying at 60 °C under vacuum. The sample is referred to as PTI/Li + Cl − . For PTI/Li + Cl − synthesized under vacuum conditions, 1.2 g of melamine was ground with KCl (6.6 g) and LiCl (5.4 g) in glovebox. Then, the mixture was transferred to an open glass tube and heated to 450 °C. After the mixture was cooled to room temperature, the tube was sealed and heated to 550 °C for 24 h at a rate of 1 °C min−1 in a tube furnace. After it was cooled to room temperature, the sample was washed with boiling deionized water several times and dried at 60 °C under vacuum. This sample is referred to as PTI/Li+Cl−Vac. Mesoporous graphitic carbon nitride (mpg-CN) was synthesized according to a previously reported procedure. Typically, cyanamide (3 g, 72 mmol, Aldrich) was dissolved in different amounts of a 40% dispersion of 12 nm SiO2 particles (Ludox HS40, Aldrich) in water (1.5, 3.75, 7.5, 12.25 g of Ludox for r = 0.2, 0.5, 1, 1.5) with stirring at 60 °C overnight. The resulting transparent mixtures were then heated at a rate of 2.3 °C min−1 over 4 h to reach a temperature of 550 °C and then tempered at this temperature for another 4 h. The resulting brown-yellow powder was treated with 4 M NH4HF2 for 24 h to remove the silica template. The powders were then centrifuged and washed three times with distilled water and twice with ethanol. Finally, the powders were dried overnight at 70 °C under vacuum. Characterization. X-ray diffraction (XRD) measurements were performed on a Bruker D8 Advance diffractometer with Cu Kα1 radiation (λ = 1.5406 Å). The solid-state 13C CP-MAS nuclear magnetic resonance (NMR) spectra were recorded on a Bruker Advance III 500 spectrometer. The Fourier transform infrared (FTIR) spectra were recorded on a BioRad FTS 6000 spectrometer. The Raman spectra were recorded out on an inVia-Reflex Raman spectroscopy system under 325 nm excitation. The UV−vis diffuse reflectance spectra (DRS) were performed on a Varian Cary 500 Scan UV−visible system. The photoluminescence (PL) spectra were recorded on an Edinburgh FI/FSTCSPC 920 spectrophotometer. The electron paramagnetic resonance (EPR) measurements were carried out on a Bruker Model A300 spectrometer. The X-ray photoelectron spectroscopy (XPS) data were obtained using a Thermo ESCALAB250 instrument with a monochromatized Al Kα line source (200 W). The morphology of the samples was investigated using Hitachi SU8010 field emission scanning electron microscopy (SEM). Transmission electron microscopy (TEM) and energy-dispersive X-ray (EDX) were performed on a FEI Tencai 20 microscope. The nitrogen adsorption− desorption isotherms were collected at 77 K using a Micromeritics ASAP 2020 surface area and porosity analyzer.

Electrochemical impedance spectroscopy and photocurrent performance analysis were performed using a BioLogic VSP300 electrochemical system. Photocatalytic Activity Test. The reactions were carried out in a Pyrex top-irradiation reaction vessel connected to a glass closed gas system. Hydrogen production was performed by dispersing 50 mg of the catalyst powder in an aqueous solution (100 mL) containing 10 mL of the 10 vol % sacrificial agent as an electron donor. A 3 wt % portion of Pt was loaded on the surface of the photocatalyst as a cocatalyst using an in situ photodeposition method with H2PtCl6. Oxygen production was carried out by dispersing 50 mg of the catalyst powder in an aqueous silver nitrate solution (0.01 M, 100 mL). La2O3, which is a basic metal oxide, buffers the pH of the reactant solution to a pH of 8−9 during the reaction. The reaction solution was evacuated several times to completely remove the air prior to irradiation under a 300 W Xe lamp. The wavelength of the incident light was controlled by applying appropriate long-pass cutoff filters. The temperature of the reaction solution was maintained at room temperature using a flow of cooling water during the reaction. The evolved gases were analyzed by a gas chromatograph equipped with a thermal conductive detector (TCD) and a 5A molecular sieve column, using argon as the carrier gas. The wavelength-dependent experiments were carried out similarly to previous experiments, only changing the cutoff filters to a different cutoff wavelength. The biomimetic experiment for hydrogen production was carried out by adding additional K2HPO4·3H2O (2.28 g, 0.01 mol) to the reaction solution. The apparent quantum yield (AQY) for the H2 evolution was determined by replacing the Xe lamp with a 405 nm semiconductor laser. The irradiation area was 7.92 cm2. The total intensity irradiation was measured by averaging 40 points in the irradiation area. For the 405 nm monochromatic light, the average intensity was 4.24 mW cm −2 (ILT 950 spectroradiometer). The AQY was calculated as AQY =

Ne 2MNAhc × 100% = × 100% Np SPtλ

where Ne is the amount of reaction electrons, Np is the amount of incident photons, M is the amount of H2 molecules, NA is Avogadro’s constant, h is the Planck constant, c is the speed of light, S is the irradiation area, P is the intensity of the irradiation, t is the photoreaction time, and λ is the wavelength of the monochromatic light. Theoretical Calculations. The periodic DFT calculations were performed using the pure PBE39 XC functional and the plane wave basis sets40 as implemented in the Vienna ab initio simulation package (VASP).41−43 C s2p2 and N s2p3 were treated as valence electrons. The cutoff energy for the plane wave basis set was 550 eV. The geometry optimizations were performed using the conjugate gradient technique, until the total energies converged to 10−4 eV and the Hellmann− Feynman forces on the atoms were less than 0.01 eV Å−1. For band calculations, the high-symmetry k paths were obtained from the literature.44 In addition to the four structures of carbon nitride, the structure of PTI/Li+Cl− was calculated. Although some programs have been developed, the theoretical calculation of random structures remains challenging. Therefore, for random occupied Li+ in the 6c site of PTI/Li+Cl−,45 we 3923

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

Figure 2. (a) XRD patterns of the samples. (b) Solid-state 13C CP-MAS NMR of the samples and melamine. Inset: the motif of triazine (left) and tri-s-triazine (right). (c) FTIR spectra of the samples. (d) Raman spectra (325 nm excitation) of the samples. (e) Ultraviolet−visible DRS of the samples. (f) PL spectra of the samples under 400 nm excitation.

tively. Although the current results underestimate the band gap, tri-s-triazine-based g-CN shows the smallest band gap when it is qualitatively compared to those of the other three motifs. Obviously, the tri-s-triazine substructure is more favorable in terms of light harvesting. For the three representative structures of PTI in Figure S1 in the Supporting Information, the results indicate that the incorporation of Cl and Li atoms in the PTI structure does not significantly influence the band composition. In addition, the total energy of tri-s-triazine-based g-CN was 0.415 eV/fu lower than that of triazine-based g-CN, which suggests that the former structure is more energetically favorable than the latter, consistent with previous results.46,47

constructed three representative structures as preliminary models.

■

RESULTS AND DISCUSSION On the basis of the results from the density functional theory (DFT) calculations, for the four types of g-CN structures, the valence band maximum (VBM) primarily consists of N p orbitals, whereas the conduction band minimum (CBM) originates from hybridization of the C p and N p orbitals (Figure 1b,d,f,h). In addition, the band gap energies of the four structures are quite different. The calculated band gap energies are 1.45, 1.17, 2.40, and 3.23 eV for the triazine-based g-CN, tri-s-triazine-based g-CN, melon, and PTI structures, respec3924

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

In addition, two peaks were observed in the NMR spectrum of PTI/Li+Cl−, which is most likely due to the coexistence of triazine and tri-s-triazine tectons under current synthetic conditions due to the high thermal stability of the tri-s-triazine structure.50 The chemical structure of g-CN-1 was further analyzed using FTIR spectra, which are shown in Figure 2c. Broad peaks between 3500 and 3000 cm−1 originated from the terminal amino groups. The peak located at 2150 cm−1 is attributed to terminal cyano groups CN, due to the loss of ammonia on the surface.51,52 The set of peaks between 1700 and 900 cm−1 is characteristic of tri-s-triazine derivatives.53 The FTIR spectrum of g-CN-1 is similar to that of bulk g-CN, indicating that the ionothermal method maintains a chemical core structure which is similar to that in bulk g-CN. Figure 2d shows the UV−Raman spectra of the samples that were obtained using laser (325 nm) excitation. The spectra of bulk g-CN contained a broad band in the 1100−1800 cm−1 region, which was assigned to C−N stretching vibrations. Three strong peaks were observed at 705, 765, and 972 cm−1. In addition, the peaks located at 705 and 765 cm−1 were due to the in-plane bending vibrations of the C−NC linked heptazine linkages. The peak at 972 cm−1 was due to the symmetric N-breathing mode of the tri-s-triazine units.35 Except for PTI/Li+Cl−, similar patterns were observed for the other samples even though the peak intensity was lower than that of bulk g-CN. For PTI/Li+Cl−, the intensities of the peaks located at 705 and 765 cm−1 were significantly lower than those observed for other samples. Moreover, the peaks located at approximately 1109 cm−1 completely disappeared, which is most likely due to a different connection pattern between the triazine-based PTI/Li+Cl− and the tri-s-triazine based g-CN. The optical absorption spectra are shown in Figure 2e. The g-CN-1 sample exhibits the typical absorption patterns of semiconductors, which is similar to that of bulk g-CN. The band edges of g-CN-1 and mpg-CN exhibited a slight red shift. In contrast, PTI/Li+Cl− exhibited a blue shift in comparison to bulk g-CN. In addition, the absorption curve of g-CN-2 is in good agreement with that of bulk g-CN, indicating that postannealing does not change the optical properties. The absorption of g-CN-1 is significantly higher than those of other samples, which is beneficial for harvesting light and enhancing the photocatalytic activity. The band gap energies estimated from the Tauc plot (Figure S5 in the Supporting Information) were 2.68, 2.72, 2.76, and 2.88 eV for mpg-CN, g-CN-1, g-CN2, and PTI/Li+Cl−, respectively. The photogenerated electron energy of these samples is sufficient to overcome the endothermic water splitting reaction with 1.23 eV redox potentials. The charge carrier separations and recombination rates of the photoexcited carriers were investigated using room-temperature photoluminescence spectra with an excitation wavelength of 400 nm. As shown in Figure 2f, bulk g-CN exhibited a broad emission peak that was centered at ∼460 nm due to the band− band PL phenomenon with the energy of light, which is approximately equal to the band gap energy of bulk g-CN. The emission peak of PTI/Li+Cl− shifted to ca. 440 nm, which was consistent with DRS results. The PL intensities for mpg-CN and g-CN-1 decreased significantly in comparison to that of bulk g-CN. In general, a decrease in the PL intensity indicates a suppressed electron−hole pair recombination due to the enhanced crystallinity of the extended conjugated system, which reduces the density of surface defects.54 The lower defect

In a typical synthesis process, melamine was directly heated to 500 °C to form the melon structure featuring structural hydrogen in the forms of NH bridges and terminal amino groups. Then, this melon-based g-CN is applied as the starting material to cooperate with the molten salts at 550 °C to induce the thermodynamically favorable reconstruction of the chemical structure and promote crystallization kinetics by facilitating both breaking of the hydrogen bond and the deamination process to yield crystalline g-CN with tri-s-triazine subunits. After the system was cooled to room temperature, a yellow mixture was obtained. Next, the mixture was dispersed in water via sonication and the salts removed by filtration to afford the final g-CN-1 sample. The XRD patterns of the samples are shown in Figure 2a. gCN-2 exhibits a pattern similar to that of bulk g-CN, indicating that g-CN-2 has the melon structure as bulk g-CN and postannealing of the precursor did not significantly change the crystal phase structure and crystallinity. For mpg-CN, the peak positions were similar to those of the bulk counterpart. The broadened peaks of mpg-CN are probably due to the reduced correlation length of the interlayer periodicity.48 For PTI/ Li+Cl−, the XRD pattern is similar to that previously reported even though the crystallinity is lower than that of the sample calcined under vacuum for a prolonged period of time (i.e., 24 h; Figure S2 in the Supporting Information).34 Interestingly, the peak positions of g-CN-1 are different from those of other reference samples. The two main peaks of g-CN-1 were shifted in the opposite direction with respect to those of bulk g-CN. The narrowed full width at half-maximum (fwhm) of the peak for the samples (Figure S3 in the Supporting Information) indicates the well-developed and more condensed crystal structure of g-CN-1. The strongest peak in bulk g-CN, which was located at 27.4° with an interlayer distance of 0.326 nm, was shifted slightly to 28.3° for g-CN-1, corresponding to a d spacing of 0.316 nm. Because this sharp peak was due to the stacking of the conjugated aromatic system being similar to that in graphite,23 the change in the peak indicated a decreasing layer distance, which was most likely due to enhanced interaction between layers. The peak at 13.0° that corresponded to an in-plane repeat unit of 0.618 nm for bulk g-CN was shifted to 8.0° with a repeat unit of 1.099 nm, which was most likely due to an unfolded in-plane network that was associated with sufficient condensation of the conjugated framework. No peak corresponded to the KCl and LiCl crystals, indicating complete removal of the salts after washing with deionized water. The crystal structure characterizations confirm that the crystallinity of the g-CN-1 is higher than that of bulk g-CN, which is primarily due to the use of an ionothermal method under present synthetic conditions. The XRD pattern of g-CN-1 is obviously different from that of PTI/ Li+Cl−, indicating that the crystal structure of g-CN-1 is distinct from that of TI/Li+Cl−. The proposed chemical structure of the tri-s-triazine units in the g-CN-1 framework was investigated by solid-state 13C CPMAS NMR analysis, and the results are shown in Figure 1b. Two distinct peaks were observed at 163.1 and 156.5 ppm for g-CN-1. The first resonance was attributed to the C(e) atoms (N2-CN or terminal CN2(NHx)), and the second resonance was related to the C(i) atoms (CN3).30,49,50 The 13C signal is similar to those of melem (Figure S4 in the Supporting Information) and melon but different from that of triazinebased melamine, which confirms the presence of tri-s-triazine units in the g-CN-1 framework rather than triazine-based units. 3925

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

Figure 3. XPS results of (a) the survey spectrum and high-resolution spectra of (b) C 1s, (c) N 1s, (d) K 2p, (e) Cl 2p, and (f) Li 1s for g-CN-1.

was the main contribution. The peaks at 286.4 and 288.3 eV were related to the sp2-hybridized carbon in the N-containing aromatic ring (N−CN), which is considered to be the major aromatic carbon species in the polymeric g-CN framework. The C 1s peak at 284.8 eV in the sample is typically assigned to impurity carbon, such as graphitic CC or grease. The N 1s spectrum in Figure 3c contains four peaks. The strongest peak (i.e., 398.8 eV) was due to the sp2-hybridized nitrogen involved in the tri-s-triazine rings (C−NC). The peak located at 400.5 eV corresponds to the tertiary nitrogen N-(C)3 groups. Both of these groups along with sp2-hybridized carbon (N−CN, 288.3 eV) comprise the tri-s-triazine heterocyclic ring units to form the basic substructure units of the g-CN polymers. The peak at 401.3 eV indicates the presence of amino functions (C−N−H), originating from the terminal amino groups on the

density of g-CN-1 was further confirmed by room-temperature EPR spectra (Figure S6 in the Supporting Information). A single Lorentzian line was centered at a g value of 2.0034 for both bulk g-CN and g-CN-1 samples, originating from the unpaired electrons in the aromatic rings of carbon atoms.55,56 The peak intensity of g-CN-1 is much lower than that of bulk gCN, indicating a decreased unpaired electron density due to the enhanced crystallinity and low defect density. Figure 3 shows the XPS results for g-CN-1. Signals corresponding to the elements C, N, K, and O were observed in the spectrum survey (Figure 3a and Table S1 in the Supporting Information). Higher-resolution spectra were recorded in the C 1s, N 1s, Cl 2p, Li 1s, and K 2p regions. The C 1s spectra in Figure 3b can be resolved into three peaks centered at 284.8, 286.4, and 288.3 eV, and the 288.3 eV peak 3926

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

Figure 4. (a) N2-sorption isotherms collected at 77 K and (b) pore size distribution for the samples. The pore size distribution was determined from the desorption branch of the isotherm. (c) Nyquist plots of electrochemical impedance spectroscopy from g-CN-1 and bulk g-CN at +0.6 V vs Ag/ AgCl reference electrode using Pt as counter electrode. (d) Photocurrent response of g-CN-1 and bulk g-CN at +0.6 V vs Ag/AgCl reference electrode using Pt as counter electrode.

surface. The peak located at 404.1 eV was due to the charging effects or positive charge localization in the heterocycles.36,53,57 The results in Figure 3d indicate that two small peaks can be assigned to the K 2p orbitals. No signals corresponding to Cl 2p and Li 1s were observed after washing with deionized water (Figure 3e,f), and this result is in contrast to the results obtained from those of the PTI/Li+Cl− sample.36 Because the potassium ions exist in the carbon nitride framework, the charge compensation may occur similarly to that in potassium melonate featuring CN as a surface terminal group (Figure S7 in the Supporting Information), which is in good agreement with the FTIR results.58 The porosity and surface area of bulk g-CN, mpg-CN, and gCN-1 were investigated using the nitrogen adsorption− desorption isotherms. As shown in Figure 4a, a typical type IV isotherm featuring a pronounced H1-type hysteresis loop in the 0.6−0.9 relative pressure (P/P0) region was observed for mpg-CN, and typical type-IV isotherms with H3-type hysteresis loops were observed for g-CN-1. In contrast, no obvious hysteresis loop was observed for bulk g-CN. The textural parameters are summarized in Table S2 in the Supporting Information, and the results indicate that the surface area of gCN-1 is higher than that of bulk g-CN, which may be due to the pores produced by the salt.59 As shown in Figure 4b, the pore size distributions determined by the BJH method indicate that the mean pore sizes are centered at approximately 12.2 nm for mpg-CN, whereas for the g-CN-1 sample the pore sizes

were 14.6 nm, which is related to the ionothermal method in present work. Electrochemical impedance spectroscopy was performed to investigate the charge-transfer rate, which resulted in the expected semicircular Nyquist plots for both the g-CN-1 and bulk g-CN samples (Figure 4c) but with a significantly decreased diameter for g-CN-1. The reduced diameter indicates improved charge mobility in the electrodes, which is further confirmed by the enhanced photocurrent generated on g-CN-1 (Figure 4d). Therefore, improved photocatalytic performance can be expected. Figure 5 and Figure S8 in the Supporting Information show typical SEM and TEM images of the samples. The morphology of g-CN-2 (Figure S8b) is similar to that of bulk g-CN (Figure S8a), which was synthesized by the traditional route. PTI/ Li+Cl− exhibits two types of morphologies as follows: hollow tubes with a diameter of approximately 500 nm and solid nanorods with diameters ranging from 100 to 200 nm (Figures S8d,e). These results are consistent with previously reported results.34 The g-CN-1 sample exhibits a nanosheet to nanorod structure with a porous structure (Figure 5a and Figure S8c), which is different from the structure of PTI/Li+Cl−. Figure 5b shows the TEM images of g-CN-1, which consists of several stacked layers of nanosheets. The high-resolution TEM images of g-CN-1 are shown in Figures 5c,d and indicated a clear hexagonal lattice structure of the crystal with two lattice fringes. The lattice fringe corresponding to 0.33 nm may be assigned as 3927

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

Figure 5. (a) SEM image of g-CN-1. (b−d) TEM images of g-CN-1.

as the electron donor. Because the band gap does not substantially change, the significantly enhanced hydrogen production for g-CN-1 is primarily due to significant improvement in the crystallinity, which can efficiently enhance the charge conductivity and reduce detrimental recombination sites. Moreover, the surface area of g-CN-1 remains lower than that of mpg-CN, indicating that the photocatalytic performance is not primarily correlated with the surface area, because in most solid−liquid phase photocatalyses the reaction rate is basically limited by charge separation rather than mass transfer.60 When TEOA was used as sacrificial agent, the HER dramatically increased to ca. 73.6 μmol h−1. This HER was further increased to 770 μmol h−1 (Figure 6e) by addition of phosphates (K2HPO4), which mimic the natural photosynthetic environment where the photosynthetic centers are embedded in thylakoid membranes featuring phospholipids that promote migration of electrons and chemical species.61 The AQY of the biomimetic photocatalytic hydrogen evolution system was determined to be 50.7% at 405 nm, which is the highest value ever established for conjugated carbon nitride polymers in hydrogen evolution photocatalysis, using TEOA as the electron donor. In addition, when the photocatalytic test was carried out with bulk g-CN by adding KCl to the reaction solution, no obvious difference was observed, which indicates that K ion in the solution does not influence the activity. However, the structure of the K ion on the surface may alternate surface properties (such as surface junctions) to speed up the migration of the photogenerated carriers to the surface. In this regard, further investigation is required to identify the role of the structure of K ion.

the interlayer distance, and the other corresponds to 0.98 nm, which likely originated from the in-plane periodicity. The d spacing of the lattice fringes is in agreement with the XRD results, which further confirms the formation of crystalline gCN during the ionothermal synthesis. In combination with the above results, we supposed that the main structure of g-CN-1 is most likely based on tri-s-triazine-based g-CN (Figure 1c). No other elements were found in the sample by EDX spectroscopy (Figure S9 in the Supporting Information). The element mapping shows a uniform distribution of C, N, and K throughout the whole selected area (Figure S10 in the Supporting Information). Hydrogen evolution studies were carried out to examine the photocatalytic activity of the samples. Three sacrificial agents, i.e., triethanolamine (TEOA), ethanol (EtOH), and methanol (MeOH), were employed to compare the photocatalytic behaviors in the presence of different sacrificial agents as electron donors. As shown in Figure 6a,c and Figure S11a in the Supporting Information, g-CN-1 shows a steady hydrogen production rate (HER) and the highest activity using visible light irradiation under the same reaction conditions using different sacrificial agents. A small amount of N2 (2 μmol) was observed in the first 1 h due to the slight decomposition of the terminal amino group. After that, no additional N2 or other gas was detected. The order of the HER of g-CN-1 was HERTEOA > HEREtOH > HERMeOH due to the different chemical reactivities of the sacrificial agents. When MeOH was used as a sacrificial agent, the average HER of g-CN-1 reached 27.5 μmol h−1, which was improved by a factor of 10 over the hydrogen production rate (2.3 μmol h−1) of the mpg-CN sample. In contrast, bulk g-CN exhibits a rather low activity using MeOH 3928

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis

Figure 6. H2 production of the samples and stability test of H2 production for g-CN-1 under visible light (>420 nm) irradiation using TEOA (a, b) and MeOH (c, d) as the sacrificial agents. (e) Time-dependent H2 production of g-CN-1 under visible light irradiation (>420 nm) by adding 0.01 M K2HPO4·3H2O to the reaction solution, using TEOA as sacrificial agent and 3 wt % of Pt as cocatalyst. (f) Oxygen production of bulk g-CN and gCN-1 under full arc irradiation using AgNO3 as sacrificial agent and La2O3 as pH buffer.

infer that this peak was related to the vibration of K ions in the surface of carbon nitride. After the long-time photocatalytic hydrogen production, the K ions were removed from the surface of carbon nitride due to the proton exchange. However, the XRD pattern shows negligible change before and after the photocatalytic experiments (Figure S12b in the Supporting Information), indicating the stable structure of the g-CN-1 backbone. In addition, wavelength-dependent H2 production was also examined. In Figure S13 in the Supporting Information, the H2 production trend was basically identical with the optical absorption of the photocatalyst, suggesting that

The stability of g-CN-1 for hydrogen evolution was studied by performing five consecutive operations in the presence of three sacrificial electron donors. As shown in Figure 6b,d and Figure S11b in the Supporting Information, the H2 evolution increased steadily as the irradiation time increased with slight fluctuations. After the photoreaction, the structure of g-CN-1 was checked by XRD and FTIR spectra. As shown in Figure S12a in the Supporting Information, overall similar spectra of gCN-1 before and after photoreaction were observed by FTIR and a peak around 1000 cm−1 disappeared. The peak located at 1000 cm−1 was observed in most salt treatment samples but was absent in the bulk g-CN. Considering the XPS results, we 3929

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

ACS Catalysis

■

ACKNOWLEDGMENTS This work was financially supported by the National Basic Research Program of China (2013CB632405) and the National Natural Science Foundation of China (21425309 and 21373048).

the H2 evolution reaction is induced by photoexcitation of the polymer sample. In the final set of experiments, photocatalytic oxygen production by g-CN-1 was performed. This process was more challenging than the hydrogen production process from water due to the four-electron process and high reaction barriers. As shown in Figure 6f, the oxygen evolution rate (∼7.0 μmol h−1) of crystalline g-CN-1 without the use of any cocatalyst was much higher than that of bulk g-CN (∼1.8 μmol h−1). This result was due to the enhanced crystallinity, which promotes charge migration and separation at the surface. This promotion of the charge separation and reaction kinetics may also favor the reduction of charge buildup at the surface, preventing photocorrosion/decomposition of g-CN-1 that releases N2 gas during the reaction (Figure S14 in the Supporting Information). The decrease in activity with reaction time was primarily due to the deposition of metallic silver on the catalyst surface, which blocks light absorption and blocks the active sites. Remarkably, metal-free crystalline g-CN-1 effectively photocatalyzed oxygen evolution by water splitting in the absence of any cofactors, and further optimization of the system for sustainable solar energy conversion via water splitting is ongoing in our laboratory.

■

CONCLUSION In summary, we have used preheated melamine as the tri-striazine-based precursor combined with the salt melt method to synthesize crystalline g-CN with tri-s-triazine subunits. A detailed characterization indicated that this synthetic method not only improved the crystallinity of the sample but also enhanced the photogenerated charge carrier mobility, in addition to increasing the surface area. The as-obtained gCN-1 shows a quite different structure in comparison with PTI synthesized by traditional precursors with the molten salts. The results from the photocatalytic experiments indicated that the as-synthesized sample exhibited high activity for hydrogen production. When a natural photosynthetic environment was mimicked by addition of phosphate to the reaction solution, the AQY for hydrogen production was 50.7% at 405 nm, which is the highest value ever reported for conjugated carbon nitride polymers for hydrogen evolution photocatalysis. Crystalline gCN-1 also exhibits enhanced oxygen evolution in comparison to bulk g-CN. This study highlights the important role of crystallinity and conjugated subunits on carbon nitride photocatalysis and indicates the potential for use of crystalline conjugated polymer photocatalysts for solar energy hydrogen production and can be extended to other solar energy applications, such as CO2 photofixation, organic photosynthesis, and pollutant degradation. ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.6b00922. Additional photocatalyst characterization data (PDF)

■

REFERENCES

(1) Fujishima, A.; Honda, K. Nature 1972, 238, 37−38. (2) Maeda, K.; Teramura, K.; Lu, D.; Takata, T.; Saito, N.; Inoue, Y.; Domen, K. Nature 2006, 440, 295−295. (3) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15729−15735. (4) Zou, Z.; Ye, J.; Sayama, K.; Arakawa, H. Nature 2001, 414, 625− 627. (5) Yi, Z.; Ye, J.; Kikugawa, N.; Kako, T.; Ouyang, S.; StuartWilliams, H.; Yang, H.; Cao, J.; Luo, W.; Li, Z.; Liu, Y.; Withers, R. L. Nat. Mater. 2010, 9, 559−564. (6) Kato, H.; Asakura, K.; Kudo, A. J. Am. Chem. Soc. 2003, 125, 3082−3089. (7) Yang, H. G.; Sun, C. H.; Qiao, S. Z.; Zou, J.; Liu, G.; Smith, S. C.; Cheng, H. M.; Lu, G. Q. Nature 2008, 453, 638−641. (8) Chen, X.; Liu, L.; Yu, P. Y.; Mao, S. S. Science 2011, 331, 746− 750. (9) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269−271. (10) Hitoki, G.; Ishikawa, A.; Takata, T.; Kondo, J. N.; Hara, M.; Domen, K. Chem. Lett. 2002, 31, 736−737. (11) Maeda, K.; Takata, T.; Hara, M.; Saito, N.; Inoue, Y.; Kobayashi, H.; Domen, K. J. Am. Chem. Soc. 2005, 127, 8286−8287. (12) Hara, M.; Kondo, T.; Komoda, M.; Ikeda, S.; Kondo, J. N.; Domen, K.; Hara, M.; Shinohara, K.; Tanaka, A. Chem. Commun. 1998, 357−358. (13) Hisatomi, T.; Kubota, J.; Domen, K. Chem. Soc. Rev. 2014, 43, 7520−7535. (14) Lei, F.; Sun, Y.; Liu, K.; Gao, S.; Liang, L.; Pan, B.; Xie, Y. J. Am. Chem. Soc. 2014, 136, 6826−6829. (15) Liu, G.; Shi, J.; Zhang, F.; Chen, Z.; Han, J.; Ding, C.; Chen, S.; Wang, Z.; Han, H.; Li, C. Angew. Chem. 2014, 126, 7423−7427. (16) Pan, C.; Takata, T.; Nakabayashi, M.; Matsumoto, T.; Shibata, N.; Ikuhara, Y.; Domen, K. Angew. Chem. 2015, 127, 2998−3002. (17) Slater, A. G.; Cooper, A. I. Science 2015, 348, aaa8075. (18) Sprick, R. S.; Jiang, J.-X.; Bonillo, B.; Ren, S.; Ratvijitvech, T.; Guiglion, P.; Zwijnenburg, M. A.; Adams, D. J.; Cooper, A. I. J. Am. Chem. Soc. 2015, 137, 3265−3270. (19) Yanagida, S.; Kabumoto, A.; Mizumoto, K.; Pac, C.; Yoshino, K. J. Chem. Soc., Chem. Commun. 1985, 474−475. (20) Kawai, T.; Kuwabara, T.; Yoshino, K. J. Chem. Soc., Faraday Trans. 1992, 88, 2041−2046. (21) Spiller, W.; Wöhrle, D.; Schulz-Ekloff, G.; Ford, W. T.; Schneider, G.; Stark, J. J. Photochem. Photobiol., A 1996, 95, 161−173. (22) Suzuki, M.; Ohta, Y.; Nagae, H.; Ichinohe, T.; Kimura, M.; Hanabusa, K.; Shirai, H.; Wohrle, D. Chem. Commun. 2000, 213−214. (23) Wang, X.; Maeda, K.; Thomas, A.; Takanabe, K.; Xin, G.; Carlsson, J. M.; Domen, K.; Antonietti, M. Nat. Mater. 2009, 8, 76−80. (24) Zheng, Y.; Lin, L.; Wang, B.; Wang, X. Angew. Chem., Int. Ed. 2015, 54, 12868−12884. (25) Wang, X.; Blechert, S.; Antonietti, M. ACS Catal. 2012, 2, 1596−1606. (26) Cao, S.; Low, J.; Yu, J.; Jaroniec, M. Adv. Mater. 2015, 27, 2150−2176. (27) Zhang, J.; Chen, Y.; Wang, X. Energy Environ. Sci. 2015, 8, 3092−3108. (28) Chu, S.; Wang, Y.; Guo, Y.; Feng, J.; Wang, C.; Luo, W.; Fan, X.; Zou, Z. ACS Catal. 2013, 3, 912−919. (29) Shi, H.; Chen, G.; Zhang, C.; Zou, Z. ACS Catal. 2014, 4, 3637−3643. (30) Lotsch, B. V.; Döblinger, M.; Sehnert, J.; Seyfarth, L.; Senker, J.; Oeckler, O.; Schnick, W. Chem. - Eur. J. 2007, 13, 4969−4980.

■

■

Research Article

AUTHOR INFORMATION

Corresponding Author

*E-mail for X.W.: [email protected]. Notes

The authors declare no competing financial interest. 3930

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931

Research Article

ACS Catalysis (31) Diebold, U. Nat. Chem. 2011, 3, 271−272. (32) Maeda, K. ACS Catal. 2013, 3, 1486−1503. (33) Bojdys, M. J.; Müller, J.-O.; Antonietti, M.; Thomas, A. Chem. Eur. J. 2008, 14, 8177−8182. (34) Wirnhier, E.; Döblinger, M.; Gunzelmann, D.; Senker, J.; Lotsch, B. V.; Schnick, W. Chem. - Eur. J. 2011, 17, 3213−3221. (35) Jorge, A. B.; Martin, D. J.; Dhanoa, M. T. S.; Rahman, A. S.; Makwana, N.; Tang, J.; Sella, A.; Corà, F.; Firth, S.; Darr, J. A.; McMillan, P. F. J. Phys. Chem. C 2013, 117, 7178−7185. (36) Schwinghammer, K.; Mesch, M. B.; Duppel, V.; Ziegler, C.; Senker, J.; Lotsch, B. V. J. Am. Chem. Soc. 2014, 136, 1730−1733. (37) Bhunia, M. K.; Yamauchi, K.; Takanabe, K. Angew. Chem., Int. Ed. 2014, 53, 11001−11005. (38) Algara-Siller, G.; Severin, N.; Chong, S. Y.; Björkman, T.; Palgrave, R. G.; Laybourn, A.; Antonietti, M.; Khimyak, Y. Z.; Krasheninnikov, A. V.; Rabe, J. P.; Kaiser, U.; Cooper, A. I.; Thomas, A.; Bojdys, M. J. Angew. Chem., Int. Ed. 2014, 53, 7450−7455. (39) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865−3868. (40) Blöchl, P. E. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953−17979. (41) Kresse, G.; Furthmüller, J. Comput. Mater. Sci. 1996, 6, 15−50. (42) Kresse, G.; Furthmüller, J. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (43) Kresse, G.; Joubert, D. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 1758−1775. (44) Setyawan, W.; Curtarolo, S. Comput. Mater. Sci. 2010, 49, 299− 312. (45) Chong, S. Y.; Jones, J. T. A.; Khimyak, Y. Z.; Cooper, A. I.; Thomas, A.; Antonietti, M.; Bojdys, M. J. J. Mater. Chem. A 2013, 1, 1102−1107. (46) Kroke, E.; Schwarz, M.; Horath-Bordon, E.; Kroll, P.; Noll, B.; Norman, A. D. New J. Chem. 2002, 26, 508−512. (47) Gracia, J.; Kroll, P. J. Mater. Chem. 2009, 19, 3013−3019. (48) Zhang, J.; Guo, F.; Wang, X. Adv. Funct. Mater. 2013, 23, 3008− 3014. (49) Jürgens, B.; Irran, E.; Senker, J.; Kroll, P.; Müller, H.; Schnick, W. J. Am. Chem. Soc. 2003, 125, 10288−10300. (50) Holst, J. R.; Gillan, E. G. J. Am. Chem. Soc. 2008, 130, 7373− 7379. (51) Khabashesku, V. N.; Zimmerman, J. L.; Margrave, J. L. Chem. Mater. 2000, 12, 3264−3270. (52) Gao, H.; Yan, S.; Wang, J.; Huang, Y. A.; Wang, P.; Li, Z.; Zou, Z. Phys. Chem. Chem. Phys. 2013, 15, 18077−18084. (53) Zhang, J.; Zhang, M.; Zhang, G.; Wang, X. ACS Catal. 2012, 2, 940−948. (54) Dontsova, D.; Pronkin, S.; Wehle, M.; Chen, Z.; Fettkenhauer, C.; Clavel, G.; Antonietti, M. Chem. Mater. 2015, 27, 5170−5179. (55) Tabbal, M.; Christidis, T.; Isber, S.; Mérel, P.; El Khakani, M. A.; Chaker, M.; Amassian, A.; Martinu, L. J. Appl. Phys. 2005, 98, 044310. (56) Zhang, G.; Zhang, M.; Ye, X.; Qiu, X.; Lin, S.; Wang, X. Adv. Mater. 2014, 26, 805−809. (57) Cui, Y.; Zhang, J.; Zhang, G.; Huang, J.; Liu, P.; Antonietti, M.; Wang, X. J. Mater. Chem. 2011, 21, 13032−13039. (58) Schwarzer, A.; Saplinova, T.; Kroke, E. Coord. Chem. Rev. 2013, 257, 2032−2062. (59) Dai, F.; Zai, J.; Yi, R.; Gordin, M. L.; Sohn, H.; Chen, S.; Wang, D. Nat. Commun. 2014, 5, 3605. (60) Huang, C.; Chen, C.; Zhang, M.; Lin, L.; Ye, X.; Lin, S.; Antonietti, M.; Wang, X. Nat. Commun. 2015, 6, 7698. (61) Liu, G.; Wang, T.; Zhang, H.; Meng, X.; Hao, D.; Chang, K.; Li, P.; Kako, T.; Ye, J. Angew. Chem., Int. Ed. 2015, 54, 13561−13565.

3931

DOI: 10.1021/acscatal.6b00922 ACS Catal. 2016, 6, 3921−3931