1 Introduction

The photocatalytic conversion of carbon oxide (CO2, green house gas) into valuable chemicals (such as methanol or methane) is of importance in view of preventing the continuous rise in global temperature and providing the alternative fuels [14]. In particular, since the first report that the photocatalytic reduction of CO2 into organic compounds over suspending semiconductor particles in water [1], many efforts have been expanded to develop efficient photocatalysts [58]. To date, a large number of reports have focused on the photoreduction of CO2 over TiO2-based photocatalysts [915]. Recently, some mixed metal oxides such as FeCaO4 [16], InTaO4 [17], and BiVO4 [18] have been developed as photocatalysts for conversion of CO2 into carbon products. It should be pointed out that the development of new photocatalyst could enhance our understanding of searching for a highly efficient photocatalyst. Thus, developing novel active photocatalysts for CO2 reduction is still of great interest and urgency.

Environmentally friendly sodium niobate (NaNbO3) is known as a promising lead-free piezoelectric material. More recently, NaNbO3 was developed as an efficient photocatalyst for hydrogen generation [19]. However, to the best of our knowledge, there is no report about the photocatalytic reduction of CO2 properties of NaNbO3. In addition, the morphology of a photocatalyst strongly affects its photocatalytic activity since the photoreduction reaction occurs on the catalyst surface [20, 21]. Hence, in this study, NaNbO3 with different morphologies was tested as a photocatalyst for photoreduction of carbon dioxide with water under light irradiation. The physical characteristics of samples were examined by the techniques, such as XRD, BET measurement, and UV–vis diffuse reflectance spectroscopy. The electronic structures of NaNbO3 were investigated using plane-wave based density functional theory calculations.

2 Experimental Section

2.1 Catalyst Synthesis

NaNbO3 bulk powders (SSR) were prepared by a conventional solid-state reaction method according to a previous report [22]. As for the preparation of NaNbO3 nanowires, we first successfully synthesized Na2Nb2O6·H2O nanowires via a facile hydrothermal route and then converted the precursors into NaNbO3 with similar nanostructures by a heating treatment. In a typical case, 1 g of P123 (EO20PO70EO20, BASF, USA) was firstly added to 25 ml of distilled water under continuous stirring at 40 °C for 2 h, followed by addition of 5 g of Nb(OC2H5)5 (Aldrich, USA). Then NaOH (Wako, Japan) solution (20 mol L−1, 10 ml) was added drop by drop into the above solution. After being stirred for 1 h at 40 °C, the so-obtained suspension was transferred into a Teflon-lined autoclave and thermally treated at 200 °C for 24 h. The so-obtained white precipitate was washed with distilled water and ethanol, and subsequently dried in an oven at 70 °C over night. The so-obtained precursors were heated at 550 °C for 4 h to synthesis the NaNbO3 orthorhombic phase. Platinum-loaded NaNbO3 samples (Pt-NaNbO3) were prepared by a photodeposition method. In a typical case, an aqueous methanol (20 vol.%) solution containing the NaNbO3 powders and hexachloroplatinic acid (0.5% H2PtCl6·6H2O) was irradiated by a 400 W high-pressure Hg lamp. After 1 h of irradiation, the suspension was filter and dried for 12 h.

2.2 Sample Characterization

The crystal structure of NaNbO3 was determined by an X-ray diffractometer (Utilma III., Japan) with Cu–Kα radiation (λ = 1.54056 Å) in the 2θ range of 20–80°. The UV–visible diffuse reflectance spectra of the sample were measured in the range of 250–800 nm using an UV–vis spectrophotometer (UV-2550, Shimadzu., Japan) equipped with an integrating sphere attachment. Morphologies of NaNbO3 samples were characterized by a field-emission scanning electron microscope (FE-SEM; JSM-6700, JEOL Co., Japan). The specific surface area was deduced according to Brunauer–Emmett–Teller method by a nitrogen adsorption apparatus (TriStar-3000, Micrometrics., USA) at 77 K.

2.3 Evaluation of Photocatalytic Activity

In this study, we applied the photoreduction of CO2 into CH4 to evaluate the property of Pt-NaNbO3 catalyst. Photocatalytic reaction was performed in a Pyrex glass vessel. The NaNbO3 powder sample (0.03 g nanowires; 0.1 g particles) was uniformly and evenly dispersed on the bottom of a small glass cell that was located in the bottom of a Pyrex glass cell, which was connected with a closed system. After replacing the system air into CO2, 3 ml of H2O was added into the reactor by a liquid syringe. Then the reactor was stored in the dark condition for 2 h until an adsorption–desorption equilibrium was reached. Finally, the light was irradiated from a 300-W Xe lamp (ILC Technology, CERMAX LX-300). Sample was periodically extracted from the reaction cell to analyze the concentration of CH4 using a gas chromatograph (GC-14B, Shimadzu., Japan).

2.4 Calculation of Electronic Structure

The ab initio calculations described here were performed with a CASTEP program package based on the density functional theory [23]. A plane wave basis set was used to describe the electronic wave functions with a kinetic energy cutoff of 800 eV. The interactions between ionic cores and valence electrons were represented by ultrasoft pseudopotentials. Exchange–correlation potentials were described by generalized gradient approximations (GGA-PBE). The Brillouin-zone (BZ) integrations of total energy were calculated using the special 5 × 3 × 2 k point generated by the Monkhorst–Pack scheme.

3 Results and Discussion

Figure 1a and b displays the ball-and-stick model and the polyhedron model of NaNbO3, respectively. It was found that NaNbO3 was a typical ABO3-type pervoskite structure, with an orthorhombic crystal structure (space group, Pbcm). The crystal structure of NaNbO3 (Fig. 1b) consisted of NbO6 octahedron sharing their vertex. As shown in Fig. 1c, It was evident that the NaNbO3 products prepared by a hydrothermal route consisted of straight, smooth, and homogeneous nanowires with a uniform diameter of about 100 nm and length of up to several tens micrometers. The SEM picture of NaNbO3 prepared by a solid-state reaction method (Fig. 1d) revealed that it mainly contained a number of micron-sized particles in addition to some big particles. Figure 2 shows a comparison of the XRD patterns of NaNbO3 samples. All diffraction peaks could be indexed to an orthorhombic structured NaNbO3, in good agreement with those in the JCPDS Card (No. 01-072-7753).

Fig. 1
figure 1

a The ball-and-stick model, b the polyhedron model of NaNbO3, c SEM image of NaNbO3 prepared by a hydrothermal method, and d SEM image of NaNbO3 SSR sample

Fig. 2
figure 2

X-ray diffraction patterns of NaNbO3 samples. As a reference, the standard pattern for NaNbO3 was shown at the bottom

Figure 3 shows the UV–vis diffuse reflectance spectrum of the sample. The optical band gap Eg of a semiconductor could be deduced according to the following equation

$$ (\alpha h\nu )^{n} = A\left( {h\nu - E_{g} } \right) $$
(1)

where α is the absorption coefficient, is the incident photo energy, the value of the index n depends on the electronic transition of the semiconductor (n direct = 2; n indirect = 1/2), A is a proportionality constant related to the material, and E g is the band gap energy of the semiconductor, respectively. The band gap energy was obtained from the intercept of the tangent line in the plot of (αhν)2 versus energy, and the value was determined to be ~3.4 eV for NaNbO3.

Fig. 3
figure 3

UV–vis diffuse reflectance spectra of NaNbO3 samples, the inset is a plot of (αhν)1/2 versus energy (eV) for NaNbO3

Figure 4 represents the evolved CH4 concentration over Pt-NaNbO3 catalysts under light irradiation. It was observed that the CH4 generated and increased with the prolongation of irradiation time over Pt-NaNbO3 sample, while there was almost no CH4 to be detected over NaNbO3 under the same reaction conditions. In addition, there was hardly any CH4 to be observed either under the dark condition or without any catalyst. This implied that the reaction strongly depended on catalyst and light. Namely, the reaction was a photocatalytic reduction process. It was worthwhile to note that the activity of NaNbO3 nanowires is higher than that of NaNbO3 particles, indicating that the NaNbO3 nanowires catalyst has fairly good activity. In order to investigate the photocatalytic repeatability of the catalyst, the photocatalytic reduction of CO2 over used NaNbO3 nanowires sample was performed again. It was revealed that the photocatalytic activity of used sample was similar to that of the fresh sample.

Fig. 4
figure 4

Photocatalytic methane evolution over Pt-NaNbO3 samples. Reaction conditions: catalyst, 0.1 g SSR NaNbO3 (SSR NNO); 0.03 g nanowires NaNbO3 (NW NNO)

Figure 5 shows a comparison of the XRD patterns of Pt-NaNbO3 nanowires before and after the photocatalytic reaction. There was no obvious peak change in the XRD patterns of NaNbO3 before and after the photocatalytic reaction, indicating that the phase of Pt-NaNbO3 catalyst was a stable photocatalyst without occurrence of structural degradation during the present photocatalytic reaction process.

Fig. 5
figure 5

X-ray diffraction patterns of NaNbO3 nanowires samples before (a) and after (b) the photocatalytic reaction

The optical properties and the photocatalytic performances of semiconductor photocatalysts are closely related to their electronic structures. We next applied plane-wave based density functional theory calculations (CASTEP program package) to study the band structures of NaNbO3. The DOS (Fig. 6b) displayed that the conduction bands of NaNbO3 are composed of the Nb 3d orbital, while the valence bands are constructed by the hybridized O 2p orbital. The correlation of E and v is known as \( v = \Updelta_{k} E(k)/\hbar \), where v is the velocity of electron and E is the energy of electron. If \( \Updelta_{k} E(k) \) is large, namely the energy band is dispersive, v will be high. That is, the more dispersive the energy is, the less localized the charge carriers are. As for the present NaNbO3, both calculated CB was abrupt, namely, v of the holes and the electron was high. Hence, the mobility of both charge carriers are high in NaNbO3, leading to promote the photocatalytic activity. On the basis of DFT results, one can see that NaNbO3 has highly mobile charge carriers, which is helpful to carrier transposition in the photocatalysts.

Fig. 6
figure 6

DFT calculations for NaNbO3, a energy band dispersion, and b density of states (DOS)

The photocatalytic CO2 reduction and physical properties of the NaNbO3 photocatalysts were summarized in Table 1. During the first 1.5 h irradiation, the rate of CH4 generation of NaNbO3 nanowires was as high as 653 ppm h−1 g−1 in contrast to that of NaNbO3 particle (22 ppm h−1 g−1). Namely, the nanowires structured photocatalyst notably enhanced the photocatalytic performance compared with particle photocatalyst. In a previous report, the phenomena that NaNbO3 nanowires showed the best photocatalytic activity for hydrogen generation among several morphologies photocatalysts was possibly ascribe to its good crystallinity, large surface-to-volume ratio (S/V) and anisotropic aspect [19]. In the present case for CO2 reduction, nanowires photocatalyst also showed higher activity. It is implied the good crystallinity, large surface-to-volume ratio (S/V) and anisotropic aspect are helpful to improve the photocatalytic activity for CO2 reduction. All these suggest some useful information to develop the photocatalyst for CO2 reduction and enhance efficiency of a photocatalytic material.

Table 1 Photocatalytic activities and physical properties of NaNbO3 photocatalysts

4 Conclusion

In this study, NaNbO3 had been successfully developed as a photocatalyst for CO2 reduction. The DFT calculations revealed that the top of VB consisted of O 2p orbital, while the bottom of CB was constructed by Nb 3d orbital, respectively. A band structure calculation showed the charge carrier in the CB was highly mobile. Compared with bulk NaNbO3 samples, NaNbO3 nanowires exhibited a much higher photocatalytic activity for CH4 production. It was possibly due to crystallinity, surface-to-volume ratio and anisotropic aspect. In summary, the present research is positive, and is expected to supply some useful information for enhancing the performance of photocatalysts and developing a new serial of photocatalysts for CO2 conversion.