Abstract
A new visible light response photocatalyst has been developed for H2 evolution from methanol solution by elemental doping. With lanthanum and cobalt dopants, the photoabsorption edge of NaNbO3 was effectively shifted to the visible light region. It is also found that the photoabsorption edge is effectively controlled by the dopant concentration. Under visible light irradiation, H2 was successfully generated over the doped NaNbO3 samples and a rate of 12 μmol·h−1 was achieved over (LaCo)0.03(NaNb)0.97O3. Densityfunctional theory calculations show that Co-induced impurity states are formed in the band gap of NaNbO3 and this is considered to be the origin of visible-light absorption upon doping with La and Co.
1. Introduction
Because of the current energy crisis and environmental pollution from the consumption of fossil fuels, new source which can provide a big amount of maintainable energy must be developed in hurry. H2 is considered as a candidate of the next generation energy source because of its renewable, unlimited, and environmental friendly performances [1, 2]. However, there are still several barriers to realize the practical utilization of H2 energy, and the produce of H2 is the most serious one. As the present H2 is mostly generated from the reformation of fossil fuel, a new method which can produce H2 with clean energy should be developed [3]. Photocatalysis has been developed as a candidate that can satisfy the demand of supplying H2 by splitting water with solar energy. In the past decades, a lot of photocatalysts were developed for producing H2 with high efficiency. But most of the photocatalysts, such as TiO2, SrTiO3, and NaTaO3, have only UV light responsibility, and the low visible light utilization limited the practical use of photocatalysis with solar light [4–6]. To improve the visible light absorption, the common method is doping with cations to adjust the electronic structures of photocatalysts [7]. When the cation dopants replace the positions of lattice cations or occupy the interstices in the crystal lattice, impurity energy levels might be generated within the band gap of the photocatalyst, which can extend the responsive region of photocatalytic reactions into visible light [8, 9].
NaNbO3 is a typical nontoxic and highly stable semiconductor which has abundant applications in photocatalysis. In many reports, NaNbO3 has been demonstrated to be a high efficiency photocatalyst for H2 generation [10–17]. Under the irradiation of UV light, NaNbO3 nanoparticles could reduce H2O to H2 with quite high efficiency with sacrificial agents [12]. Fiber-structured NaNbO3 was also verified to be useful in splitting pure H2O into H2 and O2 [10]. However, almost all the reported NaNbO3 photocatalysts are only sensitive to the UV light. Although iridium doped NaNbO3 was proved to be active in water splitting under visible light irradiation, the efficiency is still low and this method needs precious metal [18]. To achieve visible light photoactivity of NaNbO3 without previous metal dopant is still a big challenge. Cobalt, which is a typical transition element with partially occupied state, is commonly used as dopant to improve the visible light responsibility of wide band-gap photocatalysts [19–22]. However, simply doping binary oxide with cobalt may increase the defect concentration and negatively affect the photocatalytic performance. Thus, codoping is more popular to balance the charge state and decrease the defects [23, 24]. In this work, we developed a series of NaNbO3 doped with lanthanum and cobalt with H2 evolution activity under visible light irradiation. The further theoretical study indicates that the cobalt dopant creates new states in the band gap of NaNbO3 and provides the visible light absorption.
2. Experimental Section
2.1. Material Preparation
The samples were synthesized via a hydrothermal method [12]. In a typical synthesis of NaNbO3, 1.0 g of (C2H5O)5Nb and 0.24 g of C2H5ONa were added into 10 mL of 2-methoxyethanol and stirred at room temperature to form a clear colloid. Next, the mixture was stirred for 30 minutes and then heated to 120°C with a rate of 1°C·min−1 and maintained at this temperature until a dry gel was obtained. After that, 40 mL of 6 M NaOH solution was added to the powdered dry gel and stirred at room temperature to form a uniform precursor. Then, the mixture was transferred into a 50 mL Teflon sealed autoclave and heated at 180°C for 24 h. Finally, the product was washed with distilled water until pH was lower than 8.0 and the obtained powder was dried at 70°C overnight. To synthesize La, Co codoped NaNbO3, the dopant reagent La(CH3COO)3, and Co(CH3COO)2 were added in the first step and all the other procedures were the same.
2.2. Sample Characterization
The crystal structure of NaNbO3 powder was determined by an X-ray diffractometer (Rint-2000, Rigaku Co., Japan) with Cu-Ka radiation. The optical absorption spectra were measured with a UV-visible spectrophotometer (UV-2500PC, Shimadzu Co., Japan) using a BaSO4 reference. Scanning electron microscopy images were recorded with a field emission scanning electron microscopy (JSM-6701F, JEOL Co., Japan) operated at 15 kV.
2.3. Photocatalytic H2 Evolution
The H2 evolution experiments were carried out in a gas closed circulation system. In a typical experiment, 0.3 g catalyst was dispersed by a magnetic stirrer in a CH3OH solution (220 mL distilled water and 50 mL CH3OH) in a Pyrex cell with a side window. Calculated amount of H2PtCl6 solution (0.5 wt%) was added into the solution. The light source used for cocatalyst deposition was a 300 W Xe arc lamp without filter ( nm). After the H2 evolution rate became constant, the system was evacuated and an L-42 cutoff filter was added to the 300 W Xe arc lamp ( nm). The H2 evolution was measured by an in situ gas chromatograph (GC-8A, Shimadzu Co., Japan) with a thermal conductivity detector (TCD).
2.4. Theoretical Calculation
The band structures, densities of state (DOS), and partial densities of state (PDOS) of NaNbO3 and codoped NaNbO3 were calculated using the plane-wave density functional theory (DFT) with the CASTEP program package [25]. The doping concentration was set to 3.125% by, respectively, replacing a Na atom and a Nb atom by a La and a Co atom in a supercell. The electronic state of Co was [Ar] and high spin. The core electrons were replaced by ultrasoft pseudopotentials with a plane-wave basis cutoff energy of 410 eV, and the interactions of exchange and correlation were treated with Perdew-Burke-Ernzerhof parameterization (PBE) of the generalized gradient approximation (GGA). The FFT grids of basis in all the models were and the -point sets of were used.
3. Results and Discussions
The crystallographic structures of all the synthesized NaNbO3 samples were determined by X-ray diffraction (XRD) measurement (as shown in Figure 1(a)). All the observed diffraction peaks in the XRD patterns of NaNbO3 and doped NaNbO3 present good agreement with the reference data from the standard diffraction database (JCPDS-073-0803), showing that every sample was well crystalized in a single phase with the space group of Pbcm, which is the common phase of NaNbO3. However, slight shifts could be found when focusing on the particular diffraction peaks. Figure 1(b) gives the enlarged diffraction peaks with the highest intensity of NaNbO3 and doped NaNbO3. When doping NaNbO3 with La and Co, the diffraction peak shifts to the smaller diffraction angle, suggesting that the unit cell of NanbO3 has a slight expansion. As the radius changes from Na+ (102 pm) and Nb5+ (64 pm) to La3+ (103.2 pm) and Co3+ (61 pm), such expansion of cell volume is understandable [26]. The XPS measurement (as shown in Figure S1 in Supplementary Material available online at http://dx.doi.org/10.1155/2014/380421) gives obvious evidence that the valance state of Co is +3 as no evident peak of Co2+ is observed [27]. The detailed lattice parameters of the as-prepared doped and undoped NaNbO3 samples are shown in Table 1.
(a)
(b)
Since the morphology is an important factor which can greatly affect the photocatalytic performance, the scanning electron microscope (SEM) was further used to observe the morphology of the as-prepared samples and the SEM images of NaNbO3 and (LaCo)0.05(NaNb)0.95O3 are shown in Figure 2. The NaNbO3 sample is constituted by particles with the cubic morphology, and the cubic particles are generally 300~1000 nm in length. The obtained NanbO3 has the similar morphology as the sample synthesized by hydrothermal reaction in the previous report [12]. Although the crystal structure changes a little after doping with La and Co, the crystal growth process has almost no change. The doped sample has the same morphology as the pure NaNbO3.
(a)
(b)
UV-visible absorption spectra of NaNbO3 and La, Co codoped NaNbO3 powder samples are shown in Figure 3(a). The pure NabO3 sample only has an intense absorption with steep edges in the UV region. Different from the pure NaNbO3, the samples have evident absorptions in the visible light region. The optical band gaps of the as-prepared NaNbO3 samples were determined according to the following equation: in which , , , and are absorption coefficient, light frequency, proportionality constant, and optical band gap, respectively [28]. The value of index depends on the property of materials, whereas for the direct transition and for the indirect transition. For NaNbO3, the index was determined to be 1/2 according to the relationship between and . For La, Co codoped NaNbO3, the indexes were determined to be 2. The different indexes of NaNbO3 and doped NaNbO3 indicate that NaNbO3 is an indirect band-gap semiconductor, while the doped NaNbO3 samples have direct transitions with visible light absorptions. From Figure 3(b), the values of the optical band gaps for NaNbO3, (LaCo)0.01(NaNb)0.99O3, (LaCo)0.03(NaNb)0.97O3, and (LaCo)0.05(NaNb)0.95O3 are determined to be 3.42, 2.74, 2.70, and 2.65 eV, respectively. With the increasing of doping concentration, the optical band gap of NaNbO3 is continuously decreasing.
(a)
(b)
The H2 evolutions from aqueous CH3OH solution (50 mL CH3OH + 220 mL H2O) over NaNbO3 and La, Co codoped NaNbO3 (0.3 g) with 0.5 wt% Pt loading under the irradiation of visible light ( nm) are presented in Figure 4(a). As shown by the UV-visible absorption in the previous part, NaNbO3 has no visible light absorption. Under the irradiation of visible light, there is no H2 detected during the experiment in 8 hours, while the doped NaNbO3 samples exhibit photoactivities for H2 evolution in the presence of methanol as sacrificial reagent. H2 was generated almost linearly over all the doped samples in 8 hours. As plotted in Figure 4(b), the H2 evolution rates are significantly different: (LaCo)0.03(NaNb)0.97O3 > (LaCo)0.05(NaNb)0.95O3 > (LaCo)0.01(NaNb)0.99O3. Over the best catalyst (LaCo)0.03(NaNb)0.97O3, 11.9 μmol H2 could be produced every hour.
(a)
(b)
To understand the mechanism of visible light photocatalytic activity of La, Co codoped NaNbO3, theoretical calculation based on density functional theory (DFT) was carried out. The density of states (DOS) in Figure 5 indicates that the undoped NaNbO3 has simple valence band maxima (VBM) and conduction band minima (CBM). Its VBM and CBM are mainly composited by O states and Nb states. Under light irradiation, the electrons are excited from O orbitals to Nb orbitals and the holes are left in O orbitals. Then, the photogenerated electrons and holes migrate to the surface and react with water and sacrificial reagent, respectively. With La and Co doping, significant changes could be found with VBM and CBM. Two dopant states are observed between the original VBM and CBM, which narrow the band gap of doped NaNbO3 and induce the visible light absorption and visible light response H2 evolution activity. However, these two states are hybrid by Co states and O states and Co states have larger combination ratio. Thus, the improved visible light absorption is mostly caused by the transition of Co. Since the electrons excited from states to states have a high backward transition rate, the photogenerated electrons could hardly migrate to the surface and perform photocatalytic reactions. This is the reason why the photoactivity of La, Co codoped NaNbO3 under visible light is not as high as pure NaNbO3 under UV light. The general mechanism of the visible light activity over La, Co codoped NaNbO3 could be concluded in Figure 6. The doping with Co element creates new occupied and unoccupied energy levels in the band gap of NaNbO3. The transition between the new CBM and VBM could absorb visible and make the visible light photocatalytic reaction possible.
4. Conclusions
In conclusion, La, Co codoped NaNbO3 were synthesized to realize the visible light response photocatalytic H2 evolution. The doped NaNbO3 samples showed narrower optical band gaps (2.65, 2.70, and 2.74 eV for (LaCo)0.05(NaNb)0.95O3, (LaCo)0.03(NaNb)0.97O3, and (LaCo)0.01(NaNb)0.99O3, resp.) than the pure NaNbO3 (3.42 eV). In photocatalytic H2 evolution experiments, the doped NaNbO3 samples showed activity under the visible light irradiation, while the undoped NaNbO3 was not active. According to the theoretical calculation, the visible light activity of La, Co codoped NaNbO3 could be attributed to the new impurity electronic states of Co dopant. Therefore, this work presented a new material for visible light photocatalytic H2 evolution.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
The authors thank Professor Naoto Umezawa for the result discussion and Dr. Akihiro Tanaka and Dr. Hideo Iwai of Materials Analysis Station of NIMS for the XPS measurement and analysis. This work was supported by Japan Science and Technology Agency (JST) and Precursory Research for Embryonic Science and Technology (PRESTO) program.
Supplementary Materials
In the XPS measurement of (LaCo)0.05(NaNb)0.95O3, the peaks with binding energy of about 780 eV and 797 eV were observed, while their shoulder peaks were missing, suggesting the cobalt dopant exists in (LaCo)0.05(NaNb)0.95O3 as Co3+ [27].