Environ Sci Pollut Res DOI 10.1007/s11356-017-8497-4

RESEARCH ARTICLE

A quasi-hexagonal prism-shaped carbon nitride for photoreduction of carbon dioxide under visible light Zhiqiao He 1 & Danfen Wang 1 & Juntao Tang 1 & Shuang Song 1 & Jianmeng Chen 1 & Xinyong Tao 2

Received: 15 September 2016 / Accepted: 20 January 2017 # Springer-Verlag Berlin Heidelberg 2017

Abstract A quasi-hexagonal prism-shaped carbon nitride (HC3N4) was synthesized from urea-derived C3N4 (U-C3N4) using an alkaline hydrothermal process. U-C3N4 decomposition followed by hydrogen bond rearrangement of hydrolyzed products leads to the formation of a quasi-hexagonal prism-shaped structure. The H-C3N4 catalysts displayed superior activity in the photoreduction of CO2 with H2O compared to U-C3N4. The enhanced photocatalytic activities can be attributed to the promotion of incompletely coordinated nitrogen atom formation in the C3N4 molecules. Keywords Graphitic carbon nitride . Hexagonal prism shape . Hydrogen bond . Carbon dioxide . Photocatalytic reduction

Introduction The continuous use of fossil fuels contributes to the generation of CO2 as a greenhouse gas (Wang et al. 2016). Conversion of CO2 into valuable organics and fuels could minimize the effects of CO2 emission and reduce mankind’s strong dependence on nonrenewable energy sources (Whipple and Kenis 2010; Ming et al.

Responsible editor: Suresh Pillai * Shuang Song [email protected] * Xinyong Tao [email protected] 1

College of Environment, Zhejiang University of Technology, Hangzhou 310032, People’s Republic of China

2

College of Materials Science and Engineering, Zhejiang University of Technology, Hangzhou 310014, People’s Republic of China

2016). The photocatalytic reduction of CO2, which can directly utilize clean and renewable solar energy, seems to be an attractive method for the processing and recovery of its carbon-based sources (Goncalves et al. 2007). Since the pioneering report on photocatalytic reduction of CO2 using semiconductor powders in aqueous suspension systems (Inoue et al. 1979), numerous efforts have been conducted on the development of photocatalysts, such as TiO2, ZnO, SnO2, WO3, CdS, and ZnS, for CO2 photoreduction in the liquid or gas phase (Habisreutinger et al. 2013; Mao et al. 2013). Nonetheless, a variety of research strategies has primarily focused upon the investigation of wide band gap metal oxides, which can only respond to ultraviolet light. However, the efficiency of solar energy utilization of the reported photocatalysts is still low (Ong et al. 2014; Shi et al. 2014). As a result, it is necessary to find more suitable photocatalysts with highly efficient utilization of visible light for CO2 conversion. Recently, graphitic carbon nitride (g-C3N4) has attracted considerable attention owing to its excellent characteristics such as high thermal and chemical stability, appropriate band positions, and special optical, electronic, and catalytic properties (Wang et al. 2009b; Liu et al. 2011). As such, it possesses a wide array of potential applications in energy conversion, photocatalytic splitting of water, organic pollutant degradation, hydrogen evolution, and CO2 reduction (Vinu 2008; Maeda et al. 2009; Wang et al. 2009a; Dong and Zhang 2012; Shi et al. 2014). g-C3N4 has a laminar structure with weak van der Waals interactions between layers, which is analogous to the structure of graphite (Zhang and Yu 2014). However, different from graphite, g-C3N4 contains twofold and threefold coordinated nitrogen atoms. Each carbon atom is sp2 hybridized and is bound to three nitrogen atoms in threefold coordination (Zhang et al. 2001). Two basic building blocks of g-C3N4 have been mentioned often, these include a melamine and a tri-s-triazine unit (Lyth et al. 2009). The tri-striazine structure is energetically more stable than the melamine

Environ Sci Pollut Res

structure according to density functional theory calculations (Merschjann et al. 2013). Upon light irradiation, the g-C3N4 could be regarded as an optical quasi-monomer, and the primary optical excitations in g-C3N4 are mostly molecular singlet excitons formed in tri-s-triazine units (Merschjann et al. 2013). Two types of electronic transitions, including π–π* transitions and n–π* transitions occur, which are responsible for ultraviolet and visible light absorption by g-C3N4, respectively (Zhang and Yu 2014). In addition to light absorption, the edge defects of g-C3N4 function as active sites in photocatalysis. The defects consist mainly of uncondensed amino groups, including primary (-NH2) and secondary (-NH-) amino groups (Su et al. 2014). Taking into account the aforementioned considerations, it is a challenge to increase the number of edge defects in g-C3N4 without damaging the tri-s-triazine units to improve the photocatalytic performance of g-C3N4. Surface morphology is another key factor in determining the catalytic performance of g-C3N4 due to the fact that different morphologies present different surface atomic arrangements, which results in distinct geometric structures that influence the catalytic activity (Liu et al. 2015). To date, g-C3N4 structures of various shapes, including nanosheets (Niu et al. 2012; Zhang et al. 2013), nanoporous (Zheng et al. 2011), nanospheres (Zimmerman et al. 2001; Zheng et al. 2015), hollow vessels (Li et al. 2010), and microcones (Wu et al. 2015), have been successfully synthesized and investigated in order to achieve higher photocatalytic performance and stability. More recently, C3N4 nanorods assembled with irregular nanoplates were prepared by a simple reflux method (Bai et al. 2013). However, to our knowledge, no study has reported the synthesis of C3N4 with hexagonal prisms, although hexagonal nanorods have been obtained for Zn2GeO4 (Yan et al. 2011), α-Fe2O3 (Liu et al. 2013b), ZnO (Qurashi et al. 2011), and WO3 (Salmaoui et al. 2011). Herein, we report for the first time the synthesis of quasihexagonal prism-shaped C3N4 with highly exposed edge defects (H-C3N4) through thermal condensation of urea, followed by alkaline hydrothermal treatment. A detailed structural and chemical characterization of H-C3N4 was completed and a reasonable growth mechanism was proposed. In addition, its photocatalytic performance in the reduction of CO2 with water vapor under visible light was investigated and compared with that of urea-derived g-C3N4 (U-C3N4).

Experimental methods Photocatalyst preparation U-C 3N 4 was synthesized by heating urea directly in a semisealed alumina crucible with a cover at 550 °C for 2 h. After cooling to room temperature, the obtained yellow powder was collected and ground in an agate mortar.

H-C3N4 was prepared by the alkaline hydrothermal treatment of U-C3N4. U-C3N4 (1 g) was dispersed into 100 mL 0.1 mol L−1 NaOH solution at ambient temperature and then treated ultrasonically for 30 min. The mixture was transferred to a dried 200-mL Teflon-lined stainless steel autoclave and was heated at 150 °C for 24, 36, or 48 h. After cooling naturally, the obtained precipitates were collected by centrifugation and were rinsed with deionized water several times. The resultant samples were dried at 80 °C in an oven and were labeled H-C3N4-t, where t represents the hydrothermal time. Photocatalyst characterization Structural and chemical information on the synthesized samples was obtained by X-ray diffractometry (XRD, X’Pert Pro, Netherlands) with Cu-Kα radiation (λ = 0.154 nm), field emission scanning electron microscopy (FESEM, S-4800, Japan), transmission electron microscopy (TEM, Tecnai G2 F30 S-Twin, Netherlands), Brunauer–Emmett–Teller N2 adsorption (BET, ASAP 2010 analyzer, USA), X-ray photoelectron spectroscopy (XPS, PHI-5000C ESCA system, USA) with Mg-Kα radiation, Fourier transform infrared spectroscopy (FTIR, Nexus 670, USA), and ultraviolet/visible diffuse reflectance spectroscopy (UV/Vis DRS, TU-1901, China). The electrochemical properties of the samples were investigated by an electrochemical analyzer (CHI 660D, China) in a standard three-electrode system with a working electrode, a platinum wire counter electrode, and an Ag/AgCl (saturating KCl) reference electrode. The working electrodes were prepared as follows: 0.1 g of as-obtained photocatalyst was mixed with 0.05 mL acetyl acetone and 1 mL deionized water to form a slurry. The slurry was then evenly coated on a 2 × 2 cm indium tin oxide (ITO) glass electrode by a doctor blade technique. Finally, the electrode was dried in air and calcined at 300 °C for 1 h. The electrolyte was a 120 mL 0.1 mol L−1 Na2SO4 aqueous solution. A 300 W Xe lamp (CEL-HXUV300, China) with an Uvircut400 filters (λ = 400–780 nm) was utilized as a visible light source. Photocurrent–time (i–t) curves were measured in several on–off cycles of visible light irradiation to demonstrate the photocurrent responses of the photocatalysts. Photocatalytic reaction Photocatalytic performance tests were carried out in a CO2 photoreduction reaction system, which consisted mainly of a sealed stainless steel photoreactor (140 mL of inner volume) with a quartz window (80 mm in diameter), a 500 W Xe arc lamp (Beijing Electric Light Sources Research Institute, China) equipped with a UV cutoff filter (wavelength >400 nm) and a constant temperature water jacket. The integrated visible-light intensity measured with a visible-light radiometer (FZ-A, Beijing Normal University, China) was ∼10.75 mW cm−2. As the energy per photon is 4.2 × 10−19 J (Karbowski et al. 2016)

Environ Sci Pollut Res

and the Avogadro constant is 6.02 × 1023, the incident photon flux is calculated to be 2.551 × 10−6 Einstein min−1 cm−2. The water jacket temperature was controlled using a circulating water bath thermostatted at 25 ± 1 °C. In a typical run, 0.1 g of powdered catalyst was dispersed uniformly and evenly on a glass fiber pad (4 × 4 cm) located in the photoreactor. Prior to irradiation, the mixture of CO2 and water vapor was flushed through the reactor for 30 min at 1 L min−1 to remove other gases. After the reactor was filled with the gas mixture, the reactor was sealed tightly and immersed into the constant temperature water jacket. The total pressure inside the reactor was 0.1 MPa, and the partial pressure of the water vapor was 3.17 kPa. Then, the Xe lamp was switched on and magnetic stirring was continued during the light irradiation. At preset time intervals, gaseous samples were taken from a sampling port using a 1-mL gas-tight syringe and were injected manually into the gas chromatograph (GC). All the photocatalytic experiments were carried in triplicate and the data presented were mean ± standard errors.

Sample analysis Yield of the main product CH4 was measured using a GC (GC2014, Shimadzu, Japan) equipped with a flame ionization detector and an HP-PLOT/Q column (30 m × 0.32 mm × 20 μm). The amount of CO in gaseous products was analyzed using a GC (7890B, Agilent, USA) with a thermal conductivity detector and an HP-MOLESIEVE capillary column (30 m × 0.32 mm × 12 μm). The CH3OH concentration in Fig. 1 Typical FESEM images of U-C3N4 (a), H-C3N4-24 (b), HC3N4-36 (c), and H-C3N4-48 (d) as well as the TEM image of HC3N4-36 (c’)

the gas phase was determined using a GC (7890B, Agilent), equipped with a flame ionization detector and an HPINNOWAX capillary column (30 m × 0.25 mm × 25 μm). The quantum yield (QY) for CO2 reduction is defined as the number of electrons converted relative to the total number of photons incident in the reactor and was calculated by measuring the total amounts of carbon-containing products after 4 h of irradiation. The number of electrons required for the formation of one molecule of CH4, CH3OH, and CO from CO2 is 8, 6, and 2, respectively.

Results and discussion Catalyst characterization Figure 1 shows the FESEM images of the prepared C3N4 morphologies. The U-C3N4 presented an aggregated layered structure with several stacking layers (Fig. 1a). The alkaline hydrothermal process changes the surface structure of the C3N4 significantly. After hydrothermal treatment with alkaline solution for 24 h, quasi-hexagonal C3N4 plates of ~140-nm width were visible (Fig. 1b). The hexagonal plates extended in threedimensional space to generate a quasi-hexagonal prism geometric shape with increase in hydrothermal duration between 36 and 48 h (Fig. 1c, d). The H-C3N4-36 and H-C3N4-48 lengths were ~2 and ∼5 μm, and their widths were ∼0.7 and ∼1.2 μm, respectively. The SEM and TEM images of H-C3N4-36 in Fig. 1c, c’

Environ Sci Pollut Res

show that the C3N4 plates had a strong self-assembling tendency by face-to-face and edge-to-edge stacking. In order to gain more detailed structural information for the quasi-hexagonal prismshaped C3N4 catalysts, the samples were further characterized by high-resolution transmission electron microscopy (HRTEM). However, in this work, it is unfortunate that no effective data could be collected because the quasi-hexagonal prismshaped C3N4 samples were not stable and melted quickly under the focus of the strong electron beam, though the bubble-like structures of g-C3N4 were observed by Sano et al. in their HRTEM analysis (Sano et al. 2013). XRD was performed to determine the bulk crystalline phases of the catalysts, and the results are shown in Fig. 2. Two distinct diffraction peaks at ∼13.1 and ∼27.5° of U-C3N4, correspondingly indexed as the (100) and (002) planes, can be clearly distinguished. These peaks match well with the hexagonal phase of C3N4 (JCPDS 87-1526). The relatively weak diffraction peak at ∼13.1° reveals the in-plane structural packing motif with a repeated distance of 0.675 nm, whereas the stronger peak at ∼27.5° represents the interplanar structural stacking of π-conjugated carbon nitride layers with a value of 0.324 nm (Wang et al. 2009b; Dong et al. 2013). Differences in the diffraction peaks were observed when comparing the H-C3N4 with U-C3N4 patterns. Close observation shows that the (002) diffraction peaks of the alkaline-treated samples shifted to higher diffraction angles at a longer hydrothermal time, which suggests a decrease in interlayer distance. Different from the (002) peak, the (100) peak moves to lower angles, which implies that the periodicity in the direction parallel to the carbon–nitride layer was improved because of the hydrogen-bonded melem framework (Yuan et al. 2015). The calculated interlayer distances and in-plane repeated distances are 0.320 and 0.834 nm for all the H-C3N4 catalysts, respectively, based on Bragg’s equation (Eq. (1)). 2dsinθ ¼ nλ

where d corresponds to the interplanar spacing, θ represents the measured diffraction angle, λ refers to the wavelength of the X-ray diffraction, and n is the diffraction order. An intramolecular hydrogen bond link is expected to occur between the polar groups produced in in-plane and interplanar structural stacking (Sano et al. 2013). Since the hydrogen bond is stronger than a van der Waals force but weaker than a covalent bond, a compression of the average interlayer distance accompanied by in-plane spacing expansion appeared compared with UC3N4. Extra H-C3N4 peaks in the XRD pattern may be attributed to low polycondensation intermediates that originate from the decomposition of U-C3N4 during the alkaline hydrothermal process (Li et al. 2015a). FTIR spectra were used to verify conformational changes of the as-prepared C3N4 samples (Fig. 3). The broad set of peaks between 3000 and 3500 cm−1 originated from N-H stretching and hydrogen-bonding interactions (Zhang et al. 2013; Sun et al. 2014). This result demonstrates that the prepared C3N4 is not “perfect g-C3N4” and contains one-dimensional chains of NH-bridged tri-s-triazine in the carbon–nitride sheet (Sano et al. 2013). Compared with U-C3N4, intense bands assignable to the OH stretching of hydroxyl groups at 1205 and 3400 cm−1, respectively, as well as a characteristic band of the C-O vibrations in C-OH functional groups at 1060 cm−1 appeared on the HC3N4 samples, which illustrates the formation of -OH groups during the hydrothermal process (Li et al. 2012; Ong et al. 2015). Moreover, the sharp peak at ∼809 cm−1 is associated with the in-plane vibration of the tri-s-triazine ring system, whereas several absorption bands in the framework vibration region from 900 to 1800 cm−1 are ascribed to typical skeletal stretching vibrations for CN heterocycles (Li et al. 2012; Li et al. 2015b). When the U-C3N4 underwent hydrothermal treatment, the vibration mode at 809 cm−1 shifted slightly to lower wavenumber caused by the hydrogen-bonding interactions between hydroxyl

ð1Þ

tri-s-triazine ring

(iv) (iii) (ii) (i) 10

11

12

13

Intensity (a.u.)

(002)

2 Theta (degree)

(iii)

(ii) (i)

H-C3N4-48

Transmittance (%)

Intensity (a.u.)

(ii)C3H-C3N4-24 (i)C3 U-C3N4 (iv)

(100)

Intensity (a.u.)

(iv)CH-C N -48 3 3 4 (iii)C3H-C3N4-36

(100)

-OH

H-C3N4-36 H-C3N4-24 U-C3N4

(iv) (iii)

C-OH

-OH

(ii) (i)

10

20

30

40

50

2 Theta (degree)

60

70

80

26

27

28

29

2 Theta (degree)

Fig. 2 XRD patterns of U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N448

4000

3000

2000

1000 −1

Wavenumber (cm ) Fig. 3 FTIR spectra of U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N448

Environ Sci Pollut Res

and amino groups, and peaks ranging from 900 to 1800 cm−1 shifted to higher wavenumbers because of the electron donor effect of the amino and hydroxyl groups adjacent to CN heterocycles (Shalom et al. 2013). XPS was conducted to determine the surface electronic state of each sample prepared under various conditions. The core-level binding energies of different elements were corrected by referencing the C 1s peak from the adventitious carbon at 284.6 eV. The wide-scan XPS survey spectra showed that all as-synthesized sample surfaces were composed of elemental C, N, and O (Fig. 4a). Figure 4b displays high-resolution XPS spectra of C 1s after Gaussian curve fitting. Generally, the three separate peaks at 284.6, 286.2, and 288.4 eV correspond to common carbon species of adventitious carbon, C-NH2 species and sp2bonded carbon in N-containing aromatic rings (N-C=N), respectively (Li et al. 2012; Chen et al. 2014b). A number of studies have claimed that the adventitious carbon is identified as originating from the XPS instrument itself (López et al. 2009; Li et al. 2010; Bai et al. 2011; Li et al. 2013; Wang et al. 2014). However, in this work, the intensity of the C 1s XPS peak at ∼284.6 eV increased with increasing hydrothermal time. According to the literature (Nishimiya et al. 1998), the peaks at 284.6 and 285.0 eV can be assigned to graphite and aliphatic carbon, respectively. Thus, the alkaline hydrothermal treatment would result in the formation of graphite and aliphatic carbon.

N 1s H-C3N4-48 H-C3N4-36

Intensity (a.u.)

(b) C 1s C 1s

Intensity (a.u.)

O 1s

(a)

H-C3N4-48 H-C3N4-36 H-C3N4-24

H-C3N4-24

U-C3N4

U-C3N4

1400

1200

1000

800

600

400

200

0

292

290

Binding energy (eV)

288

286

284

282

Binding energy (eV)

(c) N 1s

(d) O 1s H-C3N4-48

H-C3N4-36 H-C3N4-24

H-C3N4-48

Intensity (a.u.)

Intensity (a.u.)

Fig. 4 XPS spectra of obtained U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48. a Survey spectra. b C 1s. c N 1s. d O 1s

Besides the common responses, a newly generated signal at 289.0 eV appeared on treated samples, which is identified as the C-O bond (Li et al. 2012). The presence of this bond indicates that some oxygen species could be bound to sp2hybridized carbon after alkaline hydrothermal treatment. The component distribution of C 1s spectra is summarized in Table 1. The percentages of C-NH2 and C-O species increased simultaneously by extending the hydrothermal duration. After deducting the effect of adventitious carbon, the N/C ratios decreased as a dependence of hydrothermal times, with values of 1.340, 1.295, 1.165, and 0.959 for U-C3N4, H-C3N4-24, HC3N4-36, and H-C3N4-48, respectively. The results imply that the N in C3N4 molecules was removed gradually during hydrothermal alkaline treatment High-resolution XPS spectra of N 1s in Fig. 4c were deconvoluted into three Gaussian peaks. The major peak at 398.3 eV is attributable to the sp2-hybridized nitrogen in triazine rings (C-N=C); the other relatively smaller peaks at 399.4 and 400.7 eV are normally assigned to the sp3-hybridized nitrogen (N-(C)3) and amino functional groups carrying hydrogen (CNHx), respectively (Sun et al. 2010). As the hydrothermal duration increased, the calculated N-(C)3 percentage decreased accompanied by an increase in the amount of C-NHx (Table 1). This implies that the U-C3N4 was partially hydrolyzed at the N-(C)3 bond. Figure 4d shows a high-resolution XPS scan over the O 1s core level region. No obvious signals in addition to chemisorbed

H-C3N4-36 H-C3N4-24 U-C3N4

U-C3N4 404

402

400

398

Binding energy(eV)

396

394

536

534

532

530

Binding energy (eV)

528

526

Environ Sci Pollut Res Table 1

Surface functionalities based on peaks fit to XPS spectra

Sample

U-C3N4 H-C3N4-24 H-C3N4-36 H-C3N4-48

C 1s

N 1s

O 1s

C-NH2 (%)

N-C=N (%)

C-O (%)

C=N-C (%)

N-(C)3 (%)

C-NHx (%)

C-O-H (%)

H-O-H (%)

1.85 2.32 3.42 5.71

98.15 60.46 57.54 54.28

– 37.22 39.04 40.01

59.11 58.19 59.02 58.45

33.22 30.10 24.31 23.94

7.67 11.71 16.67 17.61

– 71.43 74.81 78.12

100 28.57 25.19 21.88

H2O molecules were visible in the O 1s spectra of U-C3N4. In contrast, for all alkaline-treated samples, the O 1s core level spectra is split into two peaks at 531.4 and 532.5 eV, which are assigned to surface OH species and H-O-H. Also, an increase in hydrothermal time leads to an increase in the amount of -OH species. This trend, together with the analysis of the C 1s peak, suggests the formation of a C-OH configuration in C3N4 under our hydrothermal conditions. In general, structural changes affect the optical properties of photoresponse materials. Figure 5 shows that all the samples have an obvious optical absorption in the blue region, which is identified as the exclusive excitation of π–π* transitions of the conjugated ring systems (Sun et al. 2010). Prolonging the hydrothermal time shifts the absorption edge to a shorter wavelength and decreases the π–π* electronic transition absorbance. This blue shift results from a decomposition of single-layer networks, which is not beneficial for extended electron delocalization in aromatic layers (Fan et al. 2015). The optical band gaps (Eg, eV) of the synthesized U-C3N4 and H-C3N4 samples were determined by Tauc’s equation (Eq. (2)). αhv ¼ A hv−E g


n

ð2Þ

where α, h, v, and A are the absorption coefficient, Planck’s constant, the frequency of the incident photon, and a proportional constant, respectively (Wood and Tauc 1972; Wang et al. 2010). Among them, n is determined from different electronic transition modes (n = 1/2 for direct transition and n = 2 for indirect transition). As reported previously, g-C3N4 has been classified as a direct band-to-band transition mode (n = 1/2) (Wang et al. 2010). The corresponding Tauc plots of (αhv)1/2 vs. (hv) were depicted in the inset of Fig. 5. The Eg values of U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48 were estimated to be approximately 2.52, 2.54, 2.70, and 2.78 eV, respectively. The band structure of photocatalysts is an important physical parameter that affects the photocatalytic properties, which determine the reducing/oxidizing ability of photogenerated electrons/ holes for CO2 reduction and H2O oxidation. The positions of the conduction band (CB) and valence band (VB) of the U-C3N4 and

H-C3N4 samples can be calculated by the following equations (Eqs. (3) and (4)) (Nethercot 1974): E CB ¼ X −Ee −0:5E g

ð3Þ

E VB ¼ X þ E g

ð4Þ

where ECB represents the CB potential, EVB denotes the VB potential, Ee refers to the energy of free electrons on the hydrogen scale (Ee = 4.5 eV), Eg corresponds to the band gap energy of the semiconductor, and X is the absolute electronegativity of the semiconductor (the geometric mean of the absolute electronegativity of the constituent atoms). Herein, the Eg values of U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48, derived from Tauc’s extrapolation, were found to be approximately 2.52, 2.54, 2.70, and 2.78 eV, respectively, and the X value for g-C3N4 was found to be 4.73 eV (Chen et al. 2014a). Thus, the ECB values of U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48 could be separately estimated to be −1.03, −1.04, −1.12, and −1.16 eV, respectively, and the corresponding EVB were calculated to be 1.49, 1.50, 1.58, and 1.62 eV, respectively. Consequently, the photoinduced carriers could undergo interfacial transfer and react with the adsorbed CO2 and H2O owing to the suitable redox potentials of the photogenerated electrons and holes. Transient photocurrent measurements were conducted to evaluate the photoelectric conversion ability of the asprepared C3N4 samples. Figure 6 displays the i–t curves of the U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48 samples for three on–off cycles of intermittent visible light irradiation. It can be seen that the photocurrent quickly reaches a constant value as soon as the light turns on, while the photocurrent decreases to background level once the light turns off, therefore exhibiting good reproducibility. All the C3N4 samples present a quick response to the light either on or off, but with different photocurrent gain values. U-C3N4 exhibits the highest photocurrent value of ∼0.2 μA, while the H-C3N4 samples show relatively low photocurrent values, indicating that U-C3N4 is superior to H-C3N4 in separating electron–hole pairs in Na2SO4 aqueous solution. It should be noted that the

Environ Sci Pollut Res

0.5 μm

1.0 0.8

1/2

cm

− 1/2

)

1.2

( α hυ )

Intensity (a.u.)

1/2

(eV

0.5 μm

0h

Hydrothermal me N

N

N

N

2.4

2.6

2.8

3.0

N

hυ (eV)

U-C3N4

N

N

H-C3N4-24

N

N

N

N

N

N

H-C3N4-48 500

600

700

Fig. 5 UV/Vis diffuse reflectance and Tauc plot (inset) of U-C3N4, HC3N4-24, H-C3N4-36, and H-C3N4-48

transient photocurrent measurement only reflects the photocatalytic activity of a catalyst in Na2SO4 aqueous solution. In fact, in this work, the reduction of CO2 would compete with the reduction of H2O by trapping the photogenerated electrons (Zhai et al. 2013). As shown in Fig. 6, the generated photocurrent values on the H-C3N4 samples generally decrease with increasing hydrothermal time. This result implies that prolonged hydrothermal time is beneficial to suppressing the reduction of H2O, thereby accelerating the reduction of CO2. Similar results were also obtained by other researchers. during photocatalytic reduction of CO2 in the presence of H2O (Zhai et al. 2013; Xie et al. 2014).

-2

Photocurrent ( μ A cm )

N N

N

N

N

N

N

N N

N

N

N

N

N N

N

off

(i) U-C3N4

N

N

Assembly

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N N

N N

N N

N N

N

N

N

N

N N

N

N

N

N

N

+ H2O

(ii) (iii) (iv)

0.05

0.00 300

400

+

N

N

N

N

H

N

N

N

N N

N

N

N

N N

N N

N

O

N

N

N

H

N

N

N

N N

N N

N

N

(5)

N N

N N

N

N

N N N

H

N N

N

N

N

N N

N

N

N

N

N

+ H2O

N N

N

N

O

N

H

H

N

(6)

N

N

N

H N

N

N

+

N

N

N

N

N

N

+ H2O

N

N

N N

N

N

N

N

N

N

N N

N

N

H

N N

H N

N

+ NH3

N N

N N

(7)

O H

(ii) H-C3N4-24

(iii) H-C3N4-36 (iv) H-C3N4-48

(i)

200

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

From physical–chemical characterization and based on the morphology evolution process with reaction time, a possible

100

N

N N

N

0.10

Hydrolysis

formation mechanism of alkaline hydrothermal conversion of layered U-C3N4 into the three-dimensional quasi-hexagonal prism morphology is provided in Scheme 1 and Eqs. (5)–(7).

Growth mechanism

0.15

N

N N

N N

N

800

Wavelength (nm)

on

N

N

N N

N

N

N

N

N

N

N

N N

N

N

N

N

N

N

Scheme 1 Morphology evolution of C3N4 as a function of hydrothermal time and schematic illustration of H-C3N4 formation process. The inset is the TEM image of U-C3N4 hydrothermally treated in water for 36 h

H-C3N4-36

0.20

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N N

N

N

N

N

N

N

N

N N

N

N

N

N

N

N

N

N

N

N

N

N

3.2

N

N

N

N

2.2

N

0.25

48 h

0.2 0.0 2.0

400

36 h

24 h

0.4

N

300

0.5 μm

0.5 μm

0.5 μm

0.6

500

600

Time (s)

Fig. 6 Photocurrent responses of U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48 in 0.1 M Na2SO4 aqueous solution under visible light irradiation

With alkaline hydrothermal treatment, C3N4 undergoes hydrolysis and thus a ternary amine connected with three tri-s-triazine units (N-(C)3) is converted into a corresponding -NH- and -OH group (Eq. (5)). Further hydrolysis of a NH- group provides a -NH2 group (Eq. (6)). Excessive hydrothermal treatment may cause -NH2 substitution by an -OH group with NH 3 evolution (Eq. (7)). In other words, the alkaline hydrothermal treatment could fragment the laminar structure of g-C3N4 into relatively stable domains (Scheme 1) to form edge defects of the primary amino groups, secondary amino groups, and hydroxyl groups (Sano et al. 2013). To explore the role of NaOH in the hydrothermal process, hydrothermal treatment of U-C3N4 in pure water (without NaOH) was performed at 150 °C for 36 h. As shown in the inset in Scheme 1, the

Environ Sci Pollut Res

Photocatalytic activity The catalyst photocatalytic performance was evaluated by CO2 reduction with water vapor under visible light irradiation (λ ≥ 400 nm). No appreciable amount of hydrocarbon compounds could be detected in the blank experiment (without catalyst) and the dark experiment (without irradiation). Furthermore, no carbon-containing substances were produced when Ar (instead of CO2) and water vapor were used as the reaction gases, ruling out the possibility of carbon impurities as a source of reduction products. The yields of all carboncontaining products on the synthesized photocatalysts as a function of irradiation times are plotted in Fig. 7. The major reaction products are methane, methanol, and carbon monoxide. Methanol could not be observed in the early reaction stages because its concentration was below the detection limit. U-C3N4 with the highest BET surface of 72.86 m2 g−1 (Table 2) exhibited the weakest photocatalytic activity in CO2 reduction. Alkaline hydrothermal treatment improved the catalytic activity of the C3N4 catalysts significantly. After 4-h reaction, the quasi-hexagonal prism-shaped HC3N4-36 was most active with a product yield of 13.67, 1.22, and 7.30 μmol g−1 for CH4, CH3OH, and CO, respectively. It should be noted that a water film was inevitably covered on the surface of the catalysts. However, it is difficult to determine the products dissolved in the water film. Thus, the yields of the products mentioned in this study were obtained from the gas phase of the reactor. The mechanism of CO2 photoreduction is quite complex with a multiple-electron transfer process, which involves photoexcited electron–hole pair generation, electron and proton

25

CO CH4

20

ƹ

Ƶ U-C3N4 Ʒ H-C3N4-24 ƹ H-C3N4-36 ƽ H-C3N4-48

CH3OH

−1

Yield (μmol g )

resulting samples exhibited a similar morphology to the UC3N4 treated in 0.1 mol L−1 NaOH aqueous solution for 24 h; as a consequence, it is inferred that the role of NaOH was to promote the hydrolysis of U-C3N4. As a result of the high electronegativity of the N and O atoms, a hydrogen bond is formed because the hydrogen (H) atom binds to N or O could be attracted to some other nearby highly electronegative atom (mainly N and O). As a consequence, different from U-C3N4, where the fundamental structural units, tri-s-triazine units, are mainly cross-linked via tertiary N atoms in each single layer and the interlayers are packed together via van der Waals attractions, the hydrolyzed products with -NH-, -NH2 and -OH groups were assembled via hydrogen bond rearrangement. The quasihexagonal morphology of C3N4 without any controlling agents was formed through preferential directional growth of C3N4 crystallites owing to the limitation of the hexagonal spatial structure. With longer reaction times, the lateral and end surfaces were enlarged gradually via face-to-face stacking and edge-to-edge arrays, which form the hexagonal prism structure.

ƹ Ʒ

15

10

Ƶ 5

Ʒ

ƹ

Ƶ

0

ƽ

ƽ

ƹ Ʒ

Ƶ

Ʒ

ƽ

Ƶ

ƽ

1

2

3

4

Time (h) Fig. 7 Time dependence of the yields of CO, CH4, and CH3OH over UC3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48 under visible light

transfer, C-O bond breaking as well as C-H and O-H bond formation (Habisreutinger et al. 2013). The evolved products are closely related to the reduction potentials of CO2 to various products, as well as the number of electrons involved in the reduction process, which evidently depend on the photocatalyst properties and reaction conditions (Indrakanti et al. 2009; Mao et al. 2012). As in previous reports, the vast majority of aqueous photoreduction systems predominantly produced methanol and methane, whereas methane was found to be the major product for most gas phase photoreduction systems (Mao et al. 2013; Yu et al. 2014). Additionally, He and his coworkers demonstrated that pure g-C3N4 was a potential photocatalyst for visible light reduction of CO2 to CH4, CH3OH, and CO in their gas–solid reaction system (He et al. 2014). Thus, the products in this study are reasonable. The corresponding overall reaction process along with the thermodynamic potentials of different products (vs. the normal hydrogen electrode (NHE) at pH 7) can be described as follows: C3 N4 þ hv→ e− þ hþ

ð8Þ

CO2 þ 8Hþ þ 8e− → CH4 þ 2H2 O E 0 ¼ −0:24 V

ð9Þ

þ

−

CO2 þ 6H þ 6e → CH3 OH þ H2 O

E0 ¼ −0:38 V ð10Þ

CO2 þ 2Hþ þ 2e− →CO þ H2 O

E 0 ¼ −0:53 V

ð11Þ

According to the theoretical band structure analysis, the conduction band positions of each prepared C3N4 samples were more negative than all the standard reduction potentials of the aforementioned products, which further demonstrated that the formation of CH4, CH3OH, and CO is thermodynamically feasible. It is clear that a thermodynamically controlled reaction is more favorable to produce thermodynamically more stable molecules (Daley and Daley 2005). Thus,

Environ Sci Pollut Res

Sample

Product yield after 4 h reaction (μmol g−1) CH4

U-C3N4 H-C3N4-24 H-C3N4-36 H-C3N4-48

CO

SBET (m2 g−1)

QY (‰)

QY/SBET (10−3 m−2)

CH3OH

6.25

4.03

0.35

72.86

0.35

0.48

8.75 13.67 6.64

5.21 7.30 4.38

0.79 1.22 0.44

10.10 8.57 8.01

0.50 0.77 0.38

4.95 8.98 4.74

according to the order of reduction potentials of CO2/CH4, CO2/CH3OH, and CO2/CO, the yields of expected products in this study should be in the order of CH4 > CH3OH > CO. Conversely, from kinetic considerations, it is considerably easier to convert one CO2 molecule to CO via two proton/ two electron reaction rather than to reduce one CO2 molecule to CH3OH or CH4 by consuming more protons and electrons. In the present work, the observed yield order of the products was of the order CH4 > CO > CH3OH. Therefore, we speculate that both thermodynamic and kinetic controls play equivalent roles in the photocatalytic reduction of CO2 with water vapor. QY of the photocatalytic process is a more meaningful parameter than the yield of CO2 reduction products in evaluating the performance of a photocatalyst. The calculated QY for the catalysts U-C3N4, H-C3N4-24, H-C3N4-36, and H-C3N4-48 was approximately 0.35, 0.50, 0.77, and 0.38‰, respectively. These values indicate that the relative photocatalytic activity of the C3N4 catalysts decreased as H-C3N4-36 > H-C3N4-24 > H-C3N4-48 > U-C3N4. Generally, high-surface area photocatalytic materials are expected to give better photocatalytic performance. To exclude the influence of surface area, values of QY per unit BET surface area of the catalyst (defined as QY/ SBET) were calculated and the results are listed in Table 2. It could be found that the QY/SBET values followed a trend similar to that of QY, which implies that activity improvement of U-C3N4 was caused mainly by other factors in addition to the variable specific sample surface areas. As the overall photochemical reaction of CO2 reduction consists of reduction and oxidation half reactions, the photogenerated holes can also play a pivotal role in the CO2 photoreduction process. Accompanying CO2 photoreduction with photogenerated electrons, the photogenerated holes can react with adsorbed H2O molecules to produce oxygen and protons simultaneously (Eq. (12)). Therefore, to verify whether oxidation of H2O via the hole process occurred, it is necessary to analyze the formed O2 concentration in the gas phase. However, there was always background O2 and N2 at levels of a few hundred parts per million detected in the gas phase of the reactor after purging the mixture of CO2 and water vapor for 30 min. A similar background level was also observed by Liu et al. (2013a). In addition, the air in the syringe needle would inevitably enter into the reactor during the product

analysis process, the volume of the needle hole is approximately 2 μL. Compared with an inner volume of 140 mL of the reactor, the error caused by the air in the syringe needle can be neglected. Since the background of O2 and N2 could not be avoided, the O2/N2 volumetric ratio was used as an indicator to evaluate the variation of O2 production instead of a quantitative determination of O2 (Liu et al. 2012; Liu et al. 2013a). As depicted in Fig. 8, the O2/N2 ratio gradually increased with irradiation time, implying that the photocatalytic reduction of CO2 occurred simultaneously, accompanied by the generation of O2. 2H2 O þ 4hþ → O2 þ 4Hþ

ð12Þ

E 0 ¼ þ0:82 V

From physicochemical characterization, the main differences in the properties of the various catalysts are differences in their surface chemistry, morphology, and optical properties. In general, surface terminations and edge defects are deemed to act as active sites in defect-containing g-C3N4 (Wang et al. 2012). Su et al. found that -NH2 and -NH- served as the main active sites for CO2 activation and photoconversion (Su et al. 2014). The XPS and FTIR analysis and the proposed synthesis mechanism (Eqs. (5)–(7)) indicate that the number of edge defects in quasi-hexagonal prism-shaped C3N4 catalysts increased gradually during morphology evolution. The 0.32

Volume ratio of O2 /N2

Table 2 CO, CH4, and CH3OH production rates and SBET and QY of U-C3N4 and H-C3N4-t under visible light

0.30

0.28

0.26 0

1

2

3

4

Time (h)

Fig. 8 Time dependence of the O2/N2 volumetric ratio over H-C3N4-36 under visible light

Environ Sci Pollut Res

hexagonal prism-shaped C3N4 catalysts, which possess a larger number of edge defects, can provide a sufficient number of active sites for CO2 photoreduction reaction and are expected to have higher photocatalytic activities. Nonetheless, an unfavorable decrease in the light-harvesting efficiency and BET surface area also occurred after alkaline hydrothermal treatment. Considering all of the aforementioned considerations, there should be an optimal alkaline hydrothermal time for C3N4 modification. In this study, the optimal alkaline hydrothermal time was determined to be 36 h.

Conclusions In summary, we demonstrate the facile hydrothermal synthesis of quasi-hexagonal prism-shaped C3N4 with highly exposed edge defects using urea-derived C3N4 and NaOH solution. The hydrothermally treated U-C3N4 has a much higher photoactivity in catalyzing CO2 reduction with water vapor. The enhanced partially coordinated -NH2 and -NH- groups contribute mainly to the improved photocatalytic activities. The quasi-hexagonal prism-shaped C3N4 can be extended to other photocatalytic applications.

Acknowledgements This work was supported by the Program for Changjiang Scholars and Innovative Research Team in University (Grant IRT13096), the National Natural Science Foundation of China (Grants 21177115 and 21477117), and the Zhejiang Provincial Natural Science Foundation of China (Grants LR13B070002 and LR14E080001).

References Bai HW, Liu ZY, Sun DD (2011) Hierarchical ZnO/Cu “corn-like” materials with high photodegradation and antibacterial capability under visible light. Phys Chem Chem Phys 13:6205–6210 Bai XJ, Wang L, Zong RL, Zhu YF (2013) Photocatalytic activity enhanced via g-C3N4 nanoplates to nanorods. J Phys Chem C 117:9952–9961 Chen SF, Hu YF, Meng SG, Fu XL (2014a) Study on the separation mechanisms of photogenerated electrons and holes for composite photocatalysts g-C3N4-WO3. Appl Catal B-Environ 150:564–573 Chen Y, Wang B, Lin S, Zhang YF, Wang XC (2014b) Activation of n→ π* transitions in two-dimensional conjugated polymers for visible light photocatalysis. J Phys Chem C 118:29981–29989 Daley RF, Daley SJ (2005) Organic chemistry; Benjamin Cummings Publishing Company: Redwood City, CA Chapter 2:811–856 Dong GH, Zhang LZ (2012) Porous structure dependent photoreactivity of graphitic carbon nitride under visible light. J Mater Chem 22: 1160–1166 Dong F, Zhao ZW, Xiong T, Ni ZL, Zhang WD, Sun YJ, Ho WK (2013) In situ construction of g-C3N4/g-C3N4 metal-free heterojunction for enhanced visible-light photocatalysis. ACS Appl Mater Inter 5: 11392–11401 Fan XQ, Xing Z, Shu Z, Zhang LX, Wang LZ, Shi JL (2015) Improved photocatalytic activity of g-C3N4 derived from cyanamide-urea solution. RSC Adv 5:8323–8328

Goncalves MC, Condeco JA, Pardal TC, Roncero DM, Santos DM, Sequeira CA (2007) Electrochemical reduction of carbon dioxide for fuel cell utilisation. ECS Trans 3:49–56 Habisreutinger SN, Schmidt-Mende L, Stolarczyk JK (2013) Photocatalytic reduction of CO2 on TiO2 and other semiconductors. Angew Chem Int Edit 52:7372–7408 He YM, Zhang LH, Teng BT, Fan MH (2014) New application of Zscheme Ag3PO4/g-C3N4 composite in converting CO2 to fuel. Environ Sci Technol 49:649–656 Indrakanti VP, Kubicki JD, Schobert HH (2009) Photoinduced activation of CO2 on Ti-based heterogeneous catalysts: current state, chemical physics-based insights and outlook. Energ Environ Sci 2:745–758 Inoue T, Fujishima A, Konishi S, Honda K (1979) Photoelectrocatalytic reduction of carbon dioxide in aqueous suspensions of semiconductor powders. Nature 277:637–638 Karbowski LM, Murugan NJ, Persinger MA (2016) Experimental evidence that specific photon energies are “stored” in malignant cells for an hour: the synergism of weak magnetic field-LED wavelength pulses. Biol Med 8:1–8 Li YG, Zhang J, Wang QS, Jin YX, Huang DH, Cui QL, Zou GT (2010) Nitrogen-rich carbon nitride hollow vessels: synthesis, characterization, and their properties. J Phys Chem B 114:9429–9434 Li JH, Shen B, Hong ZH, Lin BZ, Gao BF, Chen YL (2012) A facile approach to synthesize novel oxygen-doped g-C3N4 with superior visible-light photoreactivity. Chem Commun 48:12017–12019 Li TT, Zhao LH, He YM, Cai J, Luo MF, Lin JJ (2013) Synthesis of gC3N4/SmVO4 composite photocatalyst with improved visible light photocatalytic activities in RhB degradation. Appl Catal B-Environ 129:255–263 Li HJ, Qian DJ, Chen M (2015a) Templateless infrared heating process for fabricating carbon nitride nanorods with efficient photocatalytic H2 evolution. ACS Appl Mater Inter 7:25162–25170 Li HY, Gan SY, Wang HY, Han DX, Niu L (2015b) Intercorrelated superhybrid of AgBr supported on graphitic-C 3N4-decorated nitrogen-doped graphene: high engineering photocatalytic activities for water purification and CO2 reduction. Adv Mater 27:6906–6913 Liu JH, Zhang TK, Wang ZC, Dawson G, Chen W (2011) Simple pyrolysis of urea into graphitic carbon nitride with recyclable adsorption and photocatalytic activity. J Mater Chem 21:14398–14401 Liu LJ, Zhao HL, Andino JM, Li Y (2012) Photocatalytic CO2 reduction with H2O on TiO2 nanocrystals: comparison of anatase, rutile, and brookite polymorphs and exploration of surface chemistry. ACS Catal 2:1817–1828 Liu LJ, Zhao CY, Zhao HL, Pitts D, Li Y (2013a) Porous microspheres of MgO-patched TiO2 for CO2 photoreduction with H2O vapor: temperature-dependent activity and stability. Chem Commun 49: 3664–3666 Liu Z, Lv BL, Xu Y, Wu D (2013b) Hexagonal α-Fe2O3 nanorods bound by high-index facets as high-performance electrochemical sensor. J Mater Chem A 1:3040–3046 Liu H, Nosheen F, Wang X (2015) Noble metal alloy complex nanostructures: controllable synthesis and their electrochemical property. Chem Soc Rev 44:3056–3078 López R, Gómez R, Llanos ME (2009) Photophysical and photocatalytic properties of nanosized copper-doped titania sol-gel catalysts. Catal Today 148:103–108 Lyth SM, Nabae Y, Moriya S, Kuroki S, Kakimoto M, Ozaki JI, Miyata S (2009) Carbon nitride as a nonprecious catalyst for electrochemical oxygen reduction. J Phys Chem C 113:20148–20151 Maeda K, Wang XC, Nishihara Y, Lu DL, Antonietti M, Domen K (2009) Photocatalytic activities of graphitic carbon nitride powder for water reduction and oxidation under visible light. J Phys Chem C 113: 4940–4947 Mao J, Peng TY, Zhang XH, Li K, Zan L (2012) Selective methanol production from photocatalytic reduction of CO2 on BiVO4 under visible light irradiation. Catal Commun 28:38–41

Environ Sci Pollut Res Mao J, Li K, Peng TY (2013) Recent advances in the photocatalytic CO2 reduction over semiconductors. Catal Sci Technol 3:2481–2498 Merschjann C, Tyborski T, Orthmann S, Yang F, Schwarzburg K, Lublow M, Lux-Steine MC, Schedel-Niedrig T (2013) Photophysics of polymeric carbon nitride: an optical quasimonomer. Phys Rev B 87:205204 Ming TZ, Richter RD, Shen S, Caillol S (2016) Fighting global warming by greenhouse gas removal: destroying atmospheric nitrous oxide thanks to synergies between two breakthrough technologies. Environ Sci Pollut Res 23:6119–6138 Nethercot AH (1974) Prediction of fermi energies and photoelectric thresholds based on electronegativity concepts. Phys Rev Lett 33: 1088–1091 Nishimiya K, Hata T, Imamura Y, Ishihara S (1998) Analysis of chemical structure of wood charcoal by X-ray photoelectron spectroscopy. J Wood Sci 44:56–61 Niu P, Zhang LL, Liu G, Cheng H-M (2012) Graphene-like carbon nitride nanosheets for improved photocatalytic activities. Adv Funct Mater 22:4763–4770 Ong WJ, Tan LL, Chai SP, Yong ST, Mohamed AR (2014) Highly reactive {001} facets of TiO2-based composites: synthesis, formation mechanism and characterization. Nanoscale 6:1946–2008 Ong WJ, Tan LL, Chai SP, Yong ST (2015) Graphene oxide as a structure-directing agent for the two-dimensional interface engineering of sandwich-like graphene-g-C3N4 hybrid nanostructures with enhanced visible-light photoreduction of CO2 to methane. Chem Commun 51:858–861 Qurashi A, Faiz M, Tabet N, Alam W (2011) Low temperature synthesis of hexagonal ZnO nanorods and their hydrogen sensing properties. Superlattice Microst 50:173–180 Salmaoui S, Sediri F, Gharbi N, Perruchot C, Aeiyach S, Rutkowsha A, Kulesza PJ, Jouini M (2011) Hexagonal nanorods of tungsten trioxide: synthesis, structure, electrochemical properties and activity as supporting material in electrocatalysis. Appl Surf Sci 257:8223–8229 Sano T, Tsutsui S, Koike K, Hirakawa T, Teramoto Y, Negishi N, Takeuchi K (2013) Activation of graphitic carbon nitride (g-C3N4) by alkaline hydrothermal treatment for photocatalytic NO oxidation in gas phase. J Mater Chem A 1:6489–6496 Shalom M, Inal S, Fettkenhauer C, Neher D, Antonietti M (2013) Improving carbon nitride photocatalysis by supramolecular preorganization of monomers. J Am Chem Soc 135:7118–7121 Shi HF, Chen GQ, Zhang CL, Zou ZG (2014) Polymeric g-C3N4 coupled with NaNbO3 nanowires toward enhanced photocatalytic reduction of CO2 into renewable fuel. ACS Catal 4:3637–3643 Su Q, Sun J, Wang JQ, Yang ZF, Cheng WG, Zhang SJ (2014) Urea-derived graphitic carbon nitride as an efficient heterogeneous catalyst for CO2 conversion into cyclic carbonates. Catal Sci Technol 4:1556–1562 Sun YQ, Li C, Xu YX, Bai H, Yao ZY, Shi GQ (2010) Chemically converted graphene as substrate for immobilizing and enhancing the activity of a polymeric catalyst. Chem Commun 46:4740–4742 Sun HQ, Zhou GL, Wang YX, Suvorova A, Wang SB (2014) A new metal-free carbon hybrid for enhanced photocatalysis. ACS Appl Mater Inter 6:16745–16754 Vinu A (2008) Two-dimensional hexagonally-ordered mesoporous carbon nitrides with tunable pore diameter, surface area and nitrogen content. Adv Funct Mater 18:816–827 Wang XC, Maeda K, Chen XF, Takanabe K, Domen K, Hou YD, Fu XZ, Antonietti M (2009a) Polymer semiconductors for artificial photosynthesis: hydrogen evolution by mesoporous graphitic carbon nitride with visible light. J Am Chem Soc 131:1680–1681 Wang XC, Maeda K, Thomas A, Takanabe K, Xin G, Carlsson JM, Domen K, Antonietti MA (2009b) A metal-free polymeric

photocatalyst for hydrogen production from water under visible light. Nat Mater 8:76–80 Wang Y, Di Y, Antonietti M, Li HR, Chen XF, Wang XC (2010) Excellent visible-light photocatalysis of fluorinated polymeric carbon nitride solids. Chem Mater 22:5119–5121 Wang Y, Wang XC, Antonietti M (2012) Polymeric graphitic carbon nitride as a heterogeneous organocatalyst: from photochemistry to multipurpose catalysis to sustainable chemistry. Angew Chem Int Ed 51:68–89 Wang YA, Wang F, Chen YT, Zhang DF, Li B, Kong SF, Li X, Cui LF (2014) Enhanced photocatalytic performance of ordered mesoporous Fe-doped CeO2 catalysts for the reduction of CO2 with H2O under simulated solar irradiation. Appl Catal B-Environ 147:602–609 Wang D, Yu Y, Zhang ZP, Fang HY, Chen JM, He ZQ, Song S (2016) Ag/ Ag2SO3 plasmonic catalysts with high activity and stability for CO2 reduction with water vapor under visible light. Environ Sci Pollut Res 23:18369–18378 Whipple DT, Kenis PJA (2010) Prospects of CO2 utilization via direct heterogeneous electrochemical reduction. J Phys Chem Lett 1: 3451–3458 Wood DL, Tauc JS (1972) Weak absorption tails in amorphous semiconductors. Phys Rev B 5:3144–3151 Wu XT, Liu CG, Li XF, Zhang XG, Wang C, Liu YK (2015) Effect of morphology on the photocatalytic activity of g-C3N4 photocatalysts under visible-light irradiation. Mat Sci Semicon Proc 32:76–81 Xie SJ, Wang Y, Zhang QH, Deng WP, Wang Y (2014) MgO- and Ptpromoted TiO2 as an efficient photocatalyst for the preferential reduction of carbon dioxide in the presence of water. ACS Catal 4: 3644–3653 Yan SH, Wan LJ, Li ZS, Zou ZG (2011) Facile temperature-controlled synthesis of hexagonal Zn2GeO4 nanorods with different aspect ratios toward improved photocatalytic activity for overall water splitting and photoreduction of CO2. Chem Commun 47:5632–5634 Yu JG, Jin J, Cheng BC, Jaroniec M (2014) A noble metal-free reduced graphene oxide–CdS nanorod composite for the enhanced visiblelight photocatalytic reduction of CO2 to solar fuel. J Mater Chem A 2:3407–3416 Yuan Y, Zhang L, Xing J, Utama MI, Lu X, Du K, Li Y, Hu X, Wang S, Genc A, Dunin-Borkowski R, Arbiol J, Xiong Q (2015) High-yield synthesis and optical properties of g-C3N4. Nanoscale 7:12343–12350 Zhai QG, Xie SJ, Fan WQ, Zhang QH, Wang Y, Deng WP, Wang Y (2013) Photocatalytic conversion of carbon dioxide with water into methane: platinum and copper(I) oxide co-catalysts with a core-shell structure. Angew Chem Int Edit 125:5888–5891 Zhang HY, Yu AC (2014) Photophysics and photocatalysis of carbon nitride synthesized at different temperatures. J Phys Chem C 118: 11628–11635 Zhang ZH, Leinenweber K, Bauer M, Garvie LAJ, McMillan PF, Wolf GH (2001) High-pressure bulk synthesis of crystalline C6N9H3·HCl: a novel C3N4 graphitic derivative. J Am Chem Soc 23:7788–7796 Zhang XD, Xie X, Wang H, Zhang JJ, Pan BC, Xie Y (2013) Enhanced photoresponsive ultrathin graphitic-phase C3N4 nanosheets for bioimaging. J Am Chem Soc 135:18–21 Zheng Y, Jiao Y, Chen J, Liu J, Liang J, Du AJ, Zhang WM, Zhu ZH, Smith SC, Jaroniec M, Li GQ, Qiao SZ (2011) Nanoporous graphitic-C3N4@ carbon metal-free electrocatalysts for highly efficient oxygen reduction. J Am Chem Soc 133:20116–20119 Zheng DD, Pang CY, Liu YX, Wang XC (2015) Shell-engineering of hollow g-C3N4 nanospheres via copolymerization for photocatalytic hydrogen evolution. Chem Commun 51:9706–9709 Zimmerman JL, Williams R, Khabashesku VN, Margrave JL (2001) Synthesis of spherical carbon nitride nanostructures. Nano Lett 1: 731–734