Facile Synthesis of Highly Efficient Pt/N-rGO/N-NaNbO3 Nanorods toward Photocatalytic Hydrogen Production

Fengli Yang, Quan Zhang, Lu Zhang, Mengting Cao, Qianqian Liu, Wei-Lin Dai* Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433, P. R. China

*Corresponding author: Prof. & Dr. Wei-Lin Dai Fax: +86 312402978 Tel: +86 3124 9122 E-mail: wldai@fudan.edu.cn

Graphical abstract

Highlights Ø

Pt/N-rGO/N-NaNbO3 photocatalyst prepared by two-step hydrothermal method; Ø Excellent hydrogen evolution rate of 2342 μmol g -1h -1 is attained; Ø 0.5%Pt/0.5%N-rGO/N-NaNbO3 shows 9-fold higher activity than pure NaNbO3; Ø 0.5%Pt/0.5%N-rGO/N-NaNbO3 exhibited the super stability (more than 4 cycles); Ø The light absorption, band gap and electron transfer properties were improved.

Abstract

Highly efficient and stable Pt/N-rGO/N-NaNbO3 nanorod photocatalysts were prepared by a facile two-step hydrothermal method for the first time. An excellent hydrogen evolution rate of 2342 μmol g -1 h -1 is attained by 0.5%Pt/0.5%N-rGO/NNaNbO3 under mimic sunlight irradiation, which is evidently higher than those of both the pristine NaNbO3 nanorod (241 μmol g -1 h -1 ) presented in this work and several modified NaNbO3 catalysts reported in the preceding literatures. Such a huge improvement in the photocatalytic performance is mainly attributed to the narrowed band gap of NaNbO3 nanorod by the doping of nitrogen element, the increased light absorption over the entire range of wavelength by the introduction of N-rGO, and the reduced electron-hole recombination rate by the excellent electron transfer properties of N-rGO and Pt nanoparticles. This innovative work may suggest a meaningful approach for developing a more appealing and practical method to modify sodium niobate for photocatalytic hydrogen generation.

Keywords: N-doped NaNbO3,N-doped graphene,Pt nanoparticles,photocatalytic hydrogen evolution

1. Introduction

Nowadays, the environmental pollution and the scarcity of fossil fuels have become two conflicting concerns that threaten the development and even survival of human beings. With the substantial growth of economy and the improvement of environmental awareness, people worldwide are striving to search for alternative energy sources. The research on hydrogen production, as one of them, has therefore been popularizing, while photocatalytic water-splitting hydrogen evolution is a promising technology to generate it. Accordingly, looking for new and highly efficient photocatalysts has been a research hotspot. So far, there have been a large number of semiconductor materials for photocatalytic hydrogen production, such as TiO2 [1], gC3N4 [2,3], CdS [4], SrTiO3 [5]. Among these, environmentally perovskites have garnered quite noticeable attention, for its outstanding dielectric properties and perovskite structure [6-8].

Among perovskites, sodium niobate (NaNbO3) has attracted considerable attention in both scientific and engineering fields for good chemical stability, low cost, high abundance, and environmental friendliness [9,10]. Recent researches show that NaNbO3 is able to play a noticeable role in H2 generation, CO2 reduction and degradation of pollutants due to its unique structures and properties [11-14]. Nevertheless, its inherent rapid recombination rate of photoexcited electron-hole pairs and wide band gap both restrict its photocatalytic activity. To address this issue, various strategies have been devoted to narrow its band gap and speed up the separation of electron-hole pairs, improving its photocatalytic activity including coupling with low bandgap semiconductors, self-doping or doping with other elements [15,16]. Coupling NaNbO3 with other materials (e.g., CdS [17], BiOI [18], CaZrO3 [19]) to develop composite photocatalysts is presented to be an efficient approach for improving its photocatalytic activity.

Metal particles (such as Au, Ag and Pt) loaded semiconductors have exhibited a shining prospective for improving the photocatalytic activity of niobates. On the one hand, due to the energy band alignment, the loading of metal particles on the semiconductor improves the charge separation efficiency. On the other hand, owing to the lower Fermi energy level of the metal conduction potential, the conduction electrons of the semiconductor can be injected into noble metals through the metal/semiconductor interface, and these electrons will effectively participate in the catalysis [20,21]. Among metal loaded semiconductor system, Pt-based composite is currently recognized as one of the most effective catalysts for H2 evolution, primarily because Pt particles can improve electron-hole separation through acting as electron traps, and decrease overpotential for water splitting as well.

Another interesting approach to decrease the photoinduced electron-hole pair recombination rate is to refine the electron transfer by utilizing conductive carbon materials. In recent years, reduced graphene oxide (rGO) has attracted considerable attention for its various potential applications thanks to its superior charge carrier mobility properties, ultrathin thickness, large specific surface, structural flexibility and physicochemical stability [22,23]. Due to its excellent electron mobility and lowresistance pathways, efficient separation of photoexcited electron-hole pairs at interface can be easily achieved, causing the boost of the photocatalytic activity [24,25]. However, rGO does not display high electronic conductivity because of its incomplete sp2-hybridized network. So far, theoretical and experimental studies have shown that the electronic properties of rGO can be rationally tuned through various chemical methods. Heteroatom doping, one of the most significant methods, is wellknown to modify the possibly optimized electronic structure and properties of graphene [26-28]. In particular, N-doping can effectively modulate the chemical reactivity and electron-donor properties of graphene, for the reason that nitrogen has stronger electron negativity than carbon and the nitrogen lone electron pairs are conjugated to the graphene π-system. With such particular properties, nitrogen-doped rGO is currently applied in many fields including lithium ion battery [29] and catalytic reactions [30,31].

Herein, in this study, NaNbO3 and rGO are employed as two candidates for constructing N-rGO/N-NaNbO3 composite photocatalysts, and then Pt NPs are evenly distributed on the surface of N-rGO/N-NaNbO3. This work introduces the preparation and characterization of novel Pt/N-rGO/N-NaNbO3 composite through a facile twostep method for the first time, and the nanocomposites are characterized utilizing Xray diffraction (XRD), Raman spectroscopy, X-ray photoelectron spectroscopy (XPS) and transmission electron microscopy (TEM). The photocatalytic activity of the Pt/NrGO/N-NaNbO3 nanorod composites towards the hydrogen evolution has been investigated as well. It is interesting to find that an excellent hydrogen evolution rate of 2342 μmol g -1 h -1 is attained by 0.5%Pt/0.5%N-rGO/N-NaNbO3 under sunlight irradiation, much higher than those of both the pristine NaNbO3 nanorod (241 μmol g - 1 h -1 ) presented in this work and several modified NaNbO3 catalysts reported in the literatures elsewhere.

2. Experimental

GO is purchased from XFNASNO, and other reagents of analytical grade are purchased from Sinopharm Chemical Reagent Co., Ltd. All reagents are used without further purification.

2.1 Synthesis of NaNbO3 composite

Precipitation method has been used for the synthesis of NaNbO3 nanorods for which 1 g of niobium pentaoxide (Nb2O5) has been added to 11 M NaOH aqueous solution (60 mL) and kept it magnetic stirring with a moderate speed for 30 min. Then the intermixture has been transferred into a 100 mL Teflon-lined autoclave with a hydrothermal reaction (180°C for 2 h). Subsequently, after filtering the solution, the filtrate has been washed by using deionized (DI) water and pure ethanol, and then dried at 100°C overnight. Finally, the precipitates are calcined at 500°C for 12 h.

2.2 Synthesis of N-rGO/N-NaNbO3 composite

N-rGO/N-NaNbO3 composites were firstly synthesized by a one-step hydrothermal calcinations method. Typically, GO (1 mg) has been dispersed in 50 mL DI water, then the above as-prepared NaNbO3 (0.5 g) and urea (0.36 g) have been mixed with the suspension, followed by 1 h stirring. The mixture has been transferred into a 100 mL Teflon beaker with a hydrothermal reaction (180°C for 12 h). Finally, the ground sample was washed thoroughly with DI water, dried at 100°C and then calcined at 400°C for 2 h under nitrogen atmosphere. A series of N-rGO/N-NaNbO3 composites have been synthesized through varying the amounts of GO and urea while using the same method.

2.3 Synthesis of Pt/N-rGO/N-NaNbO3 composite

The platinum nanoparticles-modified N-rGO/N-NaNbO3 composites have been prepared by using the photo-deposition method. Typically, as-prepared N-rGO/NNaNbO3 composites (0.3 g) and 10 mg/mL aqueous H2PtCl6·6H2O solution (0.42 mL) have been put into a beaker with 40 mL DI water and 9.5 mL methanol. And then the suspension has been irradiated by 300 W Xe light at 25°C for 4 h under magnetic stirring. The sample was filtered, washed with DI water and then dried in a vacuum oven at 80°C. Using the same method, the Pt NPs loaded NaNbO3 catalyst has also been synthesized.

2.4Materials characterization

The phase composition and crystalline nature were obtained on a D2 phaser Xray diffractometer. TEM and HRTEM images were attained using a transmission electron microscopy (JOEL JEM 2010) operated at an accelerating voltage of 200 kV. XPS measurements were performed with a RBD 147 upgraded Perkin Elmer PHI 5000C ESCA system using Mg Kα radiation. And contaminant carbon (C 1s=284.6 eV) was used to calibrate all binding energies. Elemental analysis (N, C and H) was determined by a VARIO EL III microanalyzer. FT-IR experiment was carried out with an AVATAR-360 Fourier Transform Infra-Red spectrometer. The band gaps of the samples were measured by Ultraviolet visible diffuse reflectance spectroscopy (UVVis. DRS) carried out by a SHIMADZU UV-2450 spectrophotometer with BaSO4 used as a reference. Photoluminescence (PL) spectra of powders were obtained with a JASCO FP-6500 fluorescence spectrophotometer at the emission wavelength of 300 nm.

2.5Photocatalytic tests

The photocatalytic hydrogen evolution reaction was carried out in a gas recirculating closed system using a CEL-HXF300 300 W Xe arc lamp without a UV cutoff filter at room temperature. 20 mg of the photocatalyst was suspended in 100 mL aqueous solution of methanol (CH3OH) (20 vol.%). Then, the reactant solutions were degassed and irradiated. And the gases produced were analyzed through gas chromatograph fitted out with a 5 Å molecular sieve column and a thermal conductivity detector.

2.6 Photoelectrochemical tests

The electrochemical measurements were carried out by using an electrochemical station and an electrochemical analyzer (CHI660E) in a standard three-electrode quartz cell. The working electrode of photocurrent was prepared as follows: the sample (3 mg) was ultrasonically dispersed in 1.0 mL of ethanol solution and then the mixture was applied to a clean FTO glass with a working area of 1 cm2 . After being dried at room temperature, the FTO glass was used as the working electrode, with a platinum plate as a counter electrode and a saturated calomel reference electrode. The transient photocurrent response of the different samples was determined in a 0.5 M Na2SO4 aqueous solution under irradiation of a 300 W Xe lamp without filter. The EIS plots were detected in mixed solution of potassium ferricyanide (0.025 M) and muriate of potash (0.1 M).

3. Results and discussion

3.1 Characterization of Pt/N-rGO/N-NaNbO3

The morphology of the Pt/N-rGO/N-NaNbO3 composites with different mass ratios is observed by TEM analyses, which discerns the presence of N-rGO and Pt NPs. The TEM image (Fig. 1a) shows that the synthesized pure NaNbO3 are nanorods with a size of about 150 nm, and possess a smooth surface. In addition, for the 0.5%Pt/0.5%N-rGO/N-NaNbO3 (Fig. 1b and 1c), the layered structure of the stacked N-rGO sheets can be clearly seen. And the N-NaNbO3 nanorods and Pt NPs (2.5 nm) are partly well distributed on the N-rGO nanosheets, as further evidenced by the clear lattice fringes of N-NaNbO3 and Pt NPs with the interplanar spacing of 0.39, 0.27 and 0.22 nm, which corresponds to the (101), (200) facets of N-NaNbO3 and (111) facet of Pt NPs, respectively. Furthermore, by utilizing energy dispersive X-ray (EDX) elemental mapping (Fig. 1e), Na, Nb, O, C, N, and Pt elements clearly emerge in the 0.5%Pt/0.5%N-rGO/N-NaNbO3 composite, offering the convincing evidence of the successful combination Pt NPs, N-rGO and N-NaNbO3. Especially, the N signal minor overlaps with the main Na, Nb and O signals, and most of them overlap with the main carbon signal, indicating that a small part of nitrogen element is successfully doped into NaNbO3, while most of it is doped into the graphene. Based on the above results, it is suggested that Pt NPs, N-rGO and N-NaNbO3 are well crystallized, and a close contact between Pt NPs, N-rGO and N-NaNbO3 is noticed, which contribute to the fast separation of photogenerated electrons and holes.

Fig. 1 (a) TEM, (b) and (c) HRTEM images of 0.5%Pt/0.5%N-rGO/N-NaNbO3; (e) The corresponding EDX mapping images of C, N, Pt, Na, Nb and O in the area of (d).

To substantiate the above assumption, XRD characterization is carried out. Fig. 2

shows the XRD patterns of the Pt/N-rGO/N-NaNbO3 composites with different mass ratios. As for pure NaNbO3, the main peak at 2θ of 22.9°, 32.6°, 46.5°, 52.6°, 58.1° and 68.1° are assigned to the diffraction planes (001), (110), (002), (021), (112) and (022) of orthorhombic crystal phase of NaNbO3 (JCPDS no. 33-1270). Compared to pure NaNbO3, it seems that the doping of N element does not lead to notable structural changes for NaNbO3 (Fig. S2). In addition, there is also no obvious signal of N-rGO phases detected in the XRD patterns of Pt/N-rGO/N-NaNbO3 with different N-rGO content, as a result of the low content and comparatively low diffraction intensity of N-rGO. Moreover, because of the ultrafine nanoparticles and well dispersion of Pt on the surface of the N-NaNbO3 and N-rGO, no signal about platinum can be detected in the XRD patterns of Pt/N-rGO/N-NaNbO3. Furthermore, there are no traces of any impurity phase under this resolution.

Fig. 2 XRD patterns of different photocatalysts

FT-IR and Raman spectra can further demonstrate the formation of Pt/N-rGO/NNaNbO3 photocatalysts. Raman scattering is an effectual approach to study the local distortions of crystal lattice. As shown in Fig. 3a, all Raman spectra have shown a strong characteristic peak near 600 cm-1 assigned to different Nb-O bond lengths [32,33], correlated with the stretching model of NbO6 octahedron [34,35]. However, compared to the pristine NaNbO3, the characteristic band of the Pt/N-rGO/N-NaNbO3 composites are shifted towards lower wavenumber and remarkably broadened (Fig. 3b), while the characteristic band of NaNbO3 treated with hydrothermal calcinations show no significant peak shifts (Fig. S3), revealing that the shift is caused by the interaction of N-rGO, Pt NPs and N-NaNbO3 rather than the additional hydrothermal treatment and calcinations of NaNbO3. The same conclusion can be drawn from the FT-IR spectra. As shown in Fig. 3c, the FT-IR spectra of the composites are in well lined with that reported in the literature. The broad peak near 600 cm-1 is assigned to the stretching model of Nb-O octahedron [36-38]. However, with the combination of N-rGO, Pt NPs and N-NaNbO3, the peak shifts to a lower wavenumber, as depicted in Fig. 3d. These shifts in the Raman and FT-IR characteristic peak mirror the different NbO6 octahedra local structure of the composites, which might affect the band structure of the NaNbO3, leading to changes in the optical band gaps.

Moreover, the Raman spectra could also confirm the existence of nitrogen doped rGO in the Pt/N-rGO/N-NaNbO3 composites. For comparison, the spectra of pristine GO and N-rGO under the same conditions are also collected (Fig. 4). As is expected, GO displays two prominent G and D bands centered at around 1350 and 1600 cm-1 . The G band of the N-rGO is shifted to 1602 cm-1 , while the D band of the N-rGO is shifted to 1355 cm-1 . The D band is well-known linked to structural defects and disorder in graphene, and the G band is associated with the E2g vibration mode of sp2 carbon [39,40]. Moreover, with the increase of N-rGO content, the Raman spectra of 0.5%Pt/2%N-rGO/NaNbO3 also exhibits the two typical D and G peaks (Fig. 3e). Compared with N-rGO, the D band of 0.5%Pt/2%N-rGO/N-NaNbO3 shifts from 1330 to 1355 cm-1 , whereas the G band shifts from 1590 to 1602 cm-1 , indicating the presence of strong interaction between N-rGO sheet and Pt/N-NaNbO3. Moreover, it is reported that the intensity ratio of D to G (ID/IG) provides the structural defect level in GO or graphene [41]. As expected, the ID/IG increased obviously from GO (0.90) to N-rGO (1.02), demonstrating a decrease of sp2 domain induced and a lower graphitic crystalline structure by the N-rGO [42]. Both the increasing of ID/IG and the shift of D and G bands demonstrate the existence of numerous defect sites and the successful introduction of nitrogen into GO in the Pt/N-rGO/N-NaNbO3 composites, which may also contribute to the photocatalytic activity of the composites. This result is further confirmed by the following XPS analysis.

Fig. 3 Raman spectroscopy (a,b) and FT-IR spectra (c,d) of (a) NaNbO3, (b) 0.5%Pt/N-NaNbO3, (c) 0.5%N-rGO/N-NaNbO3, (d) 0.5%Pt/0.1%N-rGO/N-NaNbO3, (e) 0.5%Pt/0.5%N-rGO/N-NaNbO3, (f) 0.5%Pt/1%N-rGO/N-NaNbO3, (g) 0.5%Pt/2%N-rGO/N-NaNbO3. (e) Raman spectra of GO, N-rGO and 0.5%Pt/2%NrGO/N-NaNbO3.

The surface compositions and chemical states of the as-prepared ternary nanocomposites are investigated through XPS measurements of GO, N-rGO, NaNbO3, N-NaNbO3 and Pt/N-rGO/N-NaNbO3 (Fig. 4a). It should be emphasized that the obvious N 1s signal in N-rGO and 0.5%Pt/0.5%N-rGO/N-NaNbO3 confirms the successful doping of N element into GO and NaNbO3. According to the XPS results, the N content of composites is listed in Table S1. The high resolution N 1s XPS spectra of the N-rGO shown in Fig. 4b is deconvoluted into four peaks centered at 399.3, 400.7, 401.6 and 402.5 eV, assigned to pyridinic N, pyrrolic N, graphitic N and oxidized pyridinic N, correspondingly [43], while the N 1s XPS spectra of NNaNbO3 in Fig. 4b is divided into two peaks at 399.0 and 395.4 eV, belonged to oxidized Nb-N (Nb-O-N) and Nb-N, correspondingly. The presence of N 1s peak of N-rGO and N-NaNbO3 further confirms that nitrogen is successfully introduced into rGO and NaNbO3. As shown in the N 1s XPS of the 0.5%Pt/0.5%N-rGO/N-NaNbO3 composite, in addition to the above five typical types of nitrogen, the sixth N can be observed at a much low binding energy of 396.5 eV [44,45], probably attributing to the Pt-N formed by the interaction of Pt and N-rGO [46]. Moreover, it is found that the N 1s of 0.5%Pt/0.5%N-rGO/N-NaNbO3 shifts to a lower binding energy, as a result of the interaction among the Pt, N-rGO and N-NaNbO3 (Table S2). Due to the interaction in the composites, the electrons generated by the photo-excitation can be easily transmitted to the electron-rich center N atoms, resulting in a significant decrease in the N 1s binding energy.

The C 1s XPS peaks of the N-rGO (Fig. 4c) is centered at 284.6, 286.6 and 288.3 eV, corresponded to the C-C, C-N/C=O and C=N/C-O bond respectively [47,48]. The Pt 4f XPS peaks of the 0.5%Pt/N-NaNbO3 and 0.5%Pt/0.5%N-rGO/N-NaNbO3 catalysts are fitted with two transitions corresponding to the different oxidation states of Pt (Pt0 and Pt2+). The principle peaks of 0.5%Pt/0.5%N-rGO/N-NaNbO3 are assigned to Pt0 at 73.1 eV (4f7/2) and 76.6 eV (4f5/2), while the peaks at 74.9 and 78.3 eV are corresponded to Pt in the 2+ states [49]. The ratio results of different Pt species are calculated based on above data and listed in Table S3. The Pt0 content of 0.5%Pt/0.5%N-rGO/N-NaNbO3 sample is significantly higher than that of 0.5%Pt/NNaNbO3 sample, indicating that N species in N-rGO and N-NaNbO3 act as electronrich active sites to promote the reduction of Pt precursor. Moreover, a shift to higher energy is observed in the binding energies of the 0.5%Pt/0.5%N-rGO/N-NaNbO3 catalyst, due to part of Pt donates electrons to N of N-rGO and N-NaNbO3. This is consistent with the report that the polar functional groups induced by nitrogen doping can enhance the electronic affinity of the substrate and further promote the electronic donation behavior of platinum nanoparticles [50,51].

The Nb 3d XPS peaks of NaNbO3 (Fig. 4e) centered at 206.7 eV (Nb 3d5/2) and 209.4 eV (Nb 3d3/2) are assigned to the Nb (+5) chemical state of NaNbO3. Compared to the pure NaNbO3, the Nb 3d peaks of N-NaNbO3 migrate to lower binding energy, revealing the effect of N element doping into NaNbO3. Furthermore, the Nb 3d XPS peaks of 0.5%N-rGO/N-NaNbO3 and 0.5%Pt/0.5%N-rGO/N-NaNbO3, centered at 206.7 and 209.4 eV, are fitted with the Nb (+5) chemical state of NaNbO3, and two other new peaks with higher binding energy appear. The migration to the direction of higher binding energy indicates that N species in N-rGO and N-NaNbO3 act as electron-rich active sites and the presence of Pt NPs and N-rGO both strongly interact with NaNbO3, which drastically promotes the transfer of electrons. Due to the interaction in the composites, the electrons generated by the photo-excitation can be easily transmitted to the electron-rich center N atoms, resulting in a significant decrease in the N 1s binding energy and the increase in the Pt 4f and Nb 3d binding energy, revealing the interaction among the Pt, N-rGO and N-NaNbO3. These XPS results reveal the successfully doped N element in rGO and NaNbO3, and the effective interaction among Pt NPs, N-rGO and N-NaNbO3.

...