Yajun Zhou, Lingxia Zhang* , Weimin Huang, Qinglu Kong, Xiangqian Fan, Min Wang, Jianlin Shi** State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Science, 1295 Ding-xi Road, Shanghai 200050, China

Abstract

Described herein is a facile one-pot strategy to synthesize N-doped graphitic carbon-incorporated g-C3N4 by adding slight amount of citric acid into urea as the precursor during thermal polymerization. The obtained materials retained the original framework of g-C3N4 and show remarkably enhanced visible light harvesting and promoted photo-excited charge carrier separation and transfer. The high-resolution N 1s spectrum of XPS showed a graphitic N peak, which could be attributed to N-doped graphitic carbon. In addition to the common-recognized light harvesting enhancement and charge carrier recombination inhibition, the incorporation of N-doped graphitic carbon into the planar framework of g-C3N4 is suggested to result in extended and delocalized p-conjugated system of this copolymer, thus greatly elevating the photocatalytic performance for H2 evolution by water splitting under visible light. The H2 evolution rate on N-doped graphitic carbon-incorporated g-C3N4 reached 64 mmol h
1 , which is almost 4.3 times the rate on pure g-C3N4. This approach may provide a promising route for rational design of high performance, cost-effective and metal-free photocatalysts. © 2

1. Introduction

   The ever-increasing demand for clean and renewable energy has stimulated extensive researches in the photosynthesis of clean fuels [1]. Semiconductor-based photocatalysis and photoelectrocatalysis hold promise as alternative approaches for solar energy harvesting and storage [2]. As a novel and promising photocatalyst, graphitic carbon nitride (g-C3N4) has sparked significant excitement in energy applications ranging from fuel cells [3,4], supercapacitors [5], electrocatalytic water splitting (HER [6], OER [7]), to photocatalysis for solar water splitting and pollutants photodegradation [8], etc. Besides, g-C3N4 can be applied as a chemosensor for selective optical sensing of metal ions, because its surface exposed terminal amino groups (-NH2 or ¼ NH groups) can act as Lewis basic sites [9]. Among its various applications, photocatalysis using g-C3N4 has attracted researchers' great attention due to its unique physicochemical properties. As an organic semiconductor consisting of carbon and nitrogen, which are among the most abundant elements on the Earth, g-C3N4 is metal-free and sustainable, and can be cheaply obtained from simple precursors like urea, cyanamide, dicyandiamide, melamine, etc. It is featured with 2D conjugated planes, in which tri-s-triazine units periodically repeat themselves, and stacked together through van der Waals interactions. The advantages of g-C3N4, such as high thermal and chemical stability, suitable band gap (~2.7 eV) and simple preparations, are extremely favorable for practical applications in photocatalysis [10,11]. However, its known poor visible-light utilization and fast charge carrier recombination are the serious limitations. Significant progresses have recently been made in the discovery and design of optimized geC3N4ebased photocatalysts by band engineering [12,13], micro-/nano-structure construction [14,15], bionic synthesis [16], co-catalyst combination [17], surface/ interface modification [18], etc

   Recently, researchers aimed to extend the light absorption and elevate the charge separation efficiency by copolymerization with organic monomers or constructing polymerepolymer composites [19e22]. Manas et al. incorporated a subtle amount of the pyrimidine moiety into the g-C3N4 layer structure using a strategy of combining super molecular aggregation with melted ionicpolycondensation, and the apparent quantum efficiency (AQE) of this photocatalyst reached approximately 7% at 420 nm [19]. Chen et al. adopted a molecular doping strategy to incorporate p-deficient pyridine ring entities into the conjugated matrix of g-C3N4 to relocate its p-electrons [20]. Zhang et al. grafted a variety of aromatic groups on g-C3N4 to extend the delocalization of p-electrons and optimize its photocatalytic performance [21]. Yan constructed composite catalysts of g-C3N4 and poly(3-hexylthiophene) and elevated the H2 evolution rate by up to 300 times [22]. Most recently, the construction of g-C3N4-based intramolecular donoracceptor conjugated copolymers has been reported by our group, which showed elevated activity of g-C3N4 for hydrogen evolution [23]. It is obvious that copolymerization modification and construction of composite polymer is a smart and efficient way to modulate the intrinsic electronic property and optimize the catalytic performance of g-C3N4. Whereas, we note that all the employed chemicals containing aromatic heterocycles in above reports are not environmentally friendly. Carbon or graphene composited photocatalysts have become a hot interest recently, especially carbon quantum dots (CQDs) or graphene quantum dots (GQDs) sensitized semiconductors [24,25]. Recently, Kang and his coworkers reported a very interesting result. They synthesized carbon nanodots/g-C3N4 photocatalyst with impressive performance for solar water splitting [26]. Citric acid, a weak organic acid with three carboxyl groups, is one of the commonly used precursors to synthesize carbon materials through dehydration and carbonization [27,28]. Qu and his coworkers prepared N-doped GQDs using citric acid and urea as carbon precursor and N-containing base, respectively [29]. They believed that citric acid had self-assembled into sheet structure through an intermolecular dehydrolysis reaction, and amide formed between eNH2 and eCOOH under the presence of amine, thus N-doped GQDs were obtained. In an earlier report by Schaber et al., urea was found to melt at about 133 C during the pyrolysis, accompanying the decomposition and vaporization [30]. As discussed above, only N-doped GQDs were obtained by hydrothermal treatments using urea-added citric acid as the carbon precursors in the above reports, however, we deduce that urea, the most commonly used precursor for g-C3N4 fabrication, would offer a liquid and amine-rich environment to react or copolymerize with citric acid during the process of high temperature treatment, thus offering an opportunity to yield g-C3N4 -based composite. Here we report, for the first time as far as we know, a facile onepot approach of calcining the mixture of urea and small amount of citric acid for the synthesis of N-doped graphitic carbonincorporated g-C3N4 composite, in which citric acid played the role of carbon source and reacted with urea to form N-doped graphitic carbon units in g-C3N4 matrix. The as-obtained composites showed remarkably enhanced photocatalytic performance for H2 evolution by water-splitting under visible light.

   2. Experimental section

   2.1. Chemicals Urea (AR) was purchased from Sinopharm Chemical Reagent Co.Ltd. (Shanghai, China), and citric acid monohydrate (ACS) was obtained from Alfa Aesar. All chemicals were used as received without further treatment.

   2.2. Synthesis of photocatalysts The pure g-C3N4 was prepared by directly heating urea (20 g) at 550 C for 4 h, with a ramp rate of 2 C min
   1 in air. The product denoted as CN. The modified samples were prepared by mixing urea (20 g) with different amounts (15 mg, 20 mg, or 25 mg) of citric acid monohydrate, subsequently calcining the mixture in a crucible with a cover at 550 C for 4 h with a ramp rate of 2 C min
   1 in air. The obtained samples were denoted as CN-x (x ¼ 15, 20, 25, respectively). 2.3. Characterization X-ray diffraction measurements were collected on a Rigaku Ultima IV diffractometer (Cu Ka radiation). Fourier transformed infrared (FTIR) spectra were recorded with a Nicolet iS10 FTIR spectrometer. Transmission electron microscopic (TEM) imaging and selected area electron diffraction (SAED) were performed on a JEOL 200CX electron microscope operated at 200 kV. Nitrogen adsorptionedesorption isotherms at 77 K were measured on a Micromeritics TriStar 3000 instrument. All the samples were degassed at 150 C for 8 h under flowing N2 before the measurement. Elemental analysis (C, H, N) was performed on a VARIO EL III microanalyzer. X-ray photoelectron spectroscopy (XPS) measurement was carried out on a Thermo Scientific ESCALAB 250 spectrometer with Al Ka radiation as the excitation source. Binding energies for the high resolution spectra were calibrated by setting C 1s to 284.6 eV. The UVevis absorption spectra were recorded on a UV-3600 Shimadzu spectroscope. Photoluminescence spectra (PL) of the samples were obtained at room temperature excited by incident light of 370 nm on FluoroMax®4 fluorescence spectrometer. PL spectra of CN and CN-20 dispersed in deionized water were measured on RF-5301 PC (Shimadzu).

   2.4. Photoelectrochemical measurements Electrochemical measurements were carried out on a CHI 660D electrochemical workstation (CH Instruments, Shanghai, China) with a standard three-electrode cell, which employed an FTO electrode deposited with samples as the working electrode, a platinum sheet as the counter electrode and saturated Ag/AgCl as the reference electrode.. The photocurrent was measured at a bias voltage of 0.1 V. The working electrodes were prepared as follows: First, FTO glasses were cleaned by sonication successively with distilled water, acetone and ethanol for 30 min. Then, 4 mg of sample powder CN, CN-15, CN-20, CN-25 was ultrasonically dispersed in 500uL ethanol with 20 uL Nafion solution (5%). One solution without sample powder is used for blank experiment. The resulting dispersion was drop-coated onto the FTO side. Then the electrodes were sealed with epoxy resin except for the 0.25 cm2 sample area left for photoexcitation experiments, and then dried in air. EIS Nyquist plots were obtained with an amplitude of 5 mV over the frequency range from 105 to 0.01 Hz at a bias voltage of 0 V.

   2.5. Photocatalytic water splitting test The visible light-induced H2 evolution was carried out in a Pyrex topeirradiation reaction vessel connected to a closed glass gascirculation system (Lab-Solar-III AG, Perfectlight Limited, Beijing). A 300 W xenon lamp (CEL-HXF300, Ceaulight, Beijing) with a 420 nm cut-off filter was chosen as a visible light source, and the light intensity was 230 mW cm
   2 (tested by FieldMaxII-TO, Coherent). Photocatalyst (50 mg) was suspended in an aqueous solution (100 mL) containing triethanolamine (10 vol%), and 3 wt% Pt was loaded on the surface of the catalyst by the in situ photodeposition method using H2PtCl6 as the starting material. The reactant solution was evacuated several times to remove air prior to the irradiation experiment. The temperature of the reaction solution was maintained at 10 C by a flow of cooling water during the photocatalytic reaction. The evolved gases were analyzed by gas chromatography (GC7900, Techcomp) equipped with a thermal conductive detector (TCD) and a 5 Å molecular sieve column, using nitrogen as the carrier gas.

   

   3. Results and discussion

   3.1. Structure and morphology characterization As shown in Fig. 1a of the XRD patterns of as-synthesized materials, two distinct diffraction peaks are found for all the samples, which can be ascribed to the typical diffraction peaks of g-C3N4 (JCPDS 87e1526). The stronger peak at around 27.6 represents the (002) inter-planar graphitic stacking with an interlayer distance of d ¼ 0.322 nm [31]. The other peak at around 12.8 corresponds to the (100) reflection presenting in-plane structural packing motif of tri-s-triazine [32]. The small amount of citric acid addition led to slight change in the location, intensity and shape of these two diffraction peaks, indicating that the samples obtained by citric acid-assisted calcination well retain the molecular framework of pristine g-C3N4. Furthermore, the FT-IR spectra in Fig. S1 confirm this conclusion, all of these samples exhibit several typical absorption bands reveal the characteristic structure of gC3N4. The absorption bands located at 1200-1600 cm
   1 relate to the stretching modes of aromatic CeN heterocycles; The peak at ca. 807 cm
   1 represents the breathing mode of the triazine units; While the broad band at 3000-3500 cm
   1 belongs to the NeH vibration due to the surface uncondensed amine groups [33]. However, for these samples CN-x in Fig. 1b, there are two extra peaks sited at 1557 cm
   1 and 1635 cm
   1 , respectively, enhanced with increasing addition of citric acid, suggesting the presence of a C]C skeletal vibration band of aromatic domains [34]. Thus, it can be concluded that this additional carbon species has been copolymerized into the molecular structure of g-C3N4, without changing its original framework because of the ultra-low dosage of citric acid. In Fig. 1c and d of TEM images, both samples exhibit typical nanosheet structure and CN-20 is more curly at the edge than CN. Their corresponding selected area electron diffraction pattern inserted in Fig. 1c and d also confirms their polycrystalline nature. As mentioned above, citric acid can be used to synthesize CQDs or GQDs through a “bottom-up” approach [35], but in the present study no CQDs or GQDs can be found even under high resolution TEM imaging. Table S1 summarizes the BET specific surface areas and elemental compositions of the obtained samples. It can be inferred that the small amount of citric acid addition can slightly increase the specific surface area of g-C3N4, which may be caused by the increased gaseous by-product like CO2, H2O. Elemental analysis (C, H, N) was performed to investigate the elemental composition of the samples. As listed in Table S1, the four samples exhibit similar C, H, N contents. Based on the above results, including XRD, FT-IR, TEM and elemental analysis, we can conclude that in the resultant CN-x samples no isolated carbon particles have been formed and the additional carbon species must have been incorporated into the network of g-C3N4, without changing its original planar triazine molecular structure. In order to further reveal the surface chemical compositions of the obtained samples, X-ray photoelectron spectra (XPS) were recorded. The survey spectra in Fig. S2 demonstrate that the two samples CN and CN-20 are mainly composed of C and N elements, in line with the results of elemental analysis. The high-resolution C 1s spectra are shown in Fig. 2a. CN and CN-20 have similar C 1s spectra with two peaks located at about 284.6 eV and 288.1 eV, respectively. In a previous report, the peak at 284.6 eV has been attributed to the signal of CeC bonds of graphitic carbon impurities in g-C3N4 [36], while where and how these carbon species were incorporated in g-C3N4 were not clarified. The peak with a binding energy of 288.1 eV can be further deconvoluted into two GaussianLorenzian peaks. The main contribution peak at 288.1 eV is attributed to the sp2 -hybridized carbon bonded to N atom inside the triazine rings, while the minor peak at 288.8 eV is assigned to the sp2 -hybridized carbon in the triazine ring bonded to the amino group [37]. It is worth noting that the two peaks at 284.6 eV (5.89%) and 288.8 eV (13.09%) of sample CN-20 are much stronger compared to those (284.6 eV: 4.38%, 288.8 eV: 5.71%) of CN. The high-resolution N 1s spectrum of CN-20 is a little different from that of CN as shown in Fig. 2b. The N 1s spectrum of CN can be mainly deconvoluted into four peaks with binding energies at about 398.6 eV, 399.6 eV, 400.7 eV and 404.4 eV. The dominant peak at 398.6 eV corresponds to the sp2 -hybridized nitrogen in Ccontaining triazine rings (CeN]C), whereas the peak located at 399.6 eV is usually attributed to the bridging N atoms in N-(C)3 groups. The peak at 400.7 eV indicates the amino groups (NeH), and the peak at 404.4 eV is attributed to charging effects [38]. In sample CN-20, an extra peak located at 401 eV, which corresponds to the graphitic N (Fig. S3) [39], can be observed, indicating that Ndoped graphitic carbon has been formed and introduced into the gC3N4 matrix. This can be further confirmed by the enhanced graphitic CeC signal at 284.6 eV (5.89%) in C 1s spectrum of CN-20. TG curves obtained during the thermal polymerization of urea and urea/citric acid mixture are shown in Fig. S4. It can be observed that the addition of citric acid leads to the less weight loss above the melting point (~132.7 C) of urea, implying that reaction/condensation has taken place between citric acid and urea. Considering that carboxyl groups of citric acid can react with amine groups of urea, the copolymerization between citric acid and urea will take place during the calcination process to form N-doped graphitic carbon, which could be dangled onto the surface or inserted in tris-triazine network of g-C3N4 matrix (Fig. 3). However, because of the ultra-low dosage of citric acid and the similar properties of gC3N4 and N-doped graphitic carbon, it's hard to distinguish Ndoped graphitic carbon from the g-C3N4 matrix based on XRD, Raman spectra (Fig. S5) and even NMR [26,40]