1 Introduction
Hydrogen is one of the cleanest renewable energy carriers, possibly replacing fossil fuels [1–6]. Hydrogen generation via photocatalytic water splitting can produce cheaper, clean, and renewable hydrogen driven by solar energy. Many semiconductor photocatalysts are found to be active for hydrogen evolution, such as CdS, ZnS [7,8], TiO2, WO3, ZnO [9–11], Ta3N5, GaN [12,13], and Ni(OH)2, In(OH)3 [14,15]. However, these catalysts can only absorb ultraviolet light or visible light shorter than 600 nm that accounts for less than 40% of the solar radiation on the earth. It is urgent to develop catalysts that can absorb longer wavelength visible light or infrared light for approaching effective use of solar energy, and catalyze water splitting to hydrogen with enough stability. At the same time, these catalysts have to hold suitable potential positions and good carrier mobilities for the reaction [16,17].
Since III-V semiconductors exhibit an excellently wide spectral response, absorption, and high charge mobility, the solar cell made of III-V semiconductors usually have a very high conversion efficiency from light to photocurrent [18–21]. Scientists have proposed that the III-V semiconductor materials are potential photocatalysts for water splitting. Lou and Lee suggest that a two-dimensional GeC/GaN heterostructural material might present excellent photocatalytic properties in which GaN and GeC monolayer are stacked together [19]. InP (1.34 eV) is a typical III-V semiconductor material whose band edge position is higher than the reduction potential of water. More importantly, its surface charge recombination is extremely low (170 cm/s) [22], leading to a higher carrier transfer efficiency. However, InP is unstable in many electrolyte solutions and undergoes chemical corrosion or photocorrosion. Therefore, it is necessary to construct a protection layer to stabilize InP in water dispersion [23]. Such a protection layer should exhibit a broadened spectral transmission window and permit the transfer of photogenerated charges [24]. At present, InP/Cu: ZnS QDs and Zn-InP QDs and other InP-based composite materials have been used for photocatalytic reduction of H2. However, these catalysts require the addition of sacrificial reagents to achieve a high photocatalytic activity [25,26]. SnO has a bandgap of 2.5–3.4 eV. It is a p-type wide bandgap semiconductor that allows visible light transmission and can be easily prepared under low-temperature conditions. Besides, it is very stable in water in the range of pH= 4–14 [27]. Nevertheless, due to the lattice matching problem between the semiconductor heterojunctions, the electron transport between these two semiconductors is significantly affected [28]. Therefore, it is considered to introduce a semiconductor between these two to reduce the mismatch between them as much as possible. As an n-type wide bandgap semiconductor, In(OH)3 has suitable conduction band (CB) (–0.93 V) and valence band (VB) (4.24 V) positions versus NHE. It has a strong redox ability, so it is a promising photocatalytic material [29,30]. For this reason, it is possible to consider building a SnO (110) - In(OH)3 (220) - InP (200) or SnO (211) - In(OH)3 (422) - InP (222) assembly to reduce inter-semiconductor lattice mismatch.
In this paper, a wide bandgap semiconductor SnO layer was deposited on the In(OH)3/InP surface to construct a SnO/In(OH)3/InP catalyst, which greatly improves the catalyst stability and activity by applying an anti-corrosion SnO layer and a In(OH)3 transition layer. Depositing the precious metal Pt on the surface of the catalyst further accelerates its charge separation and transfer. The obtained Pt/SnO/In(OH)3/InP exhibits a hydrogen production rate of 144.42 µmol/g in 3 h under visible light illumination without remarkable decay in the successive four cycle tests, which is 123 times higher than that of naked In(OH)3/InP without any sacrifice reagent under visible irradiation. The In(OH)3 transition layer remarkably reduces the crystal mismatch between SnO and InP and increases charge transfer. This work provides a new method to develop high efficiency and high stability photocatalysts made of III-V semiconductor materials.
2 Experimental section
2.1 Preparation of In(OH)3/InP
An improved hydrothermal method is used to prepare In(OH)3/InP catalyst [31]. Typically, 1.00 g and 1.09 g of InCl3 and NaOH, respectively, are dissolved in 27.2 mL of water, into which, 1.24 g of hexadecyl trimethyl ammonium bromide (CTAB), 9 mL of n-hexanol and 2.72 mL of n-octane are added. After stirring for 60 min, it is put in a 50 mL reactor. Next, 0.54 g and 1.73 g of red phosphorus (P) and iodine element (I2), respectively, are added. Then it is transferred to an autoclave which is maintained at 160°C for 24 h. In the end, it is cooled at room temperature, centrifuged and washed (xylene, ethanol, dilute hydrochloric acid (0.1 mol/L), and deionized water). The obtained product is dried at 60°C and is ready for the test.
2.2 Preparation of SnO/In(OH)3/InP catalyst
The SnO/In(OH)3/InP catalyst is prepared by the hydrothermal method. First, 200 mg of In(OH)3/InP powder and a calculated amount of SnCl2·2H2O are added to the beaker, then 30 mL of water is added, after stirring for 30 min, 0.7 mol/L KOH is added, and the dispersion is stirred ultrasonically for 30 min. It is then transferred to an autoclave which is reacted at 80°C for 15 h. The resulting products were collected by filtering, and washed with water and ethanol, respectively. The powder is dried at 60°C to obtain the final product. The obtained samples are donated respectively according to the different ratios of SnO (3%, 5%, 7%, and 10%, respectively).
2.3 Photocatalytic activity and AQE measurement
The photocatalytic activities are tested in a quartz reactor (approximately 190 mL). The effective irradiation area of the reactor is 11.9 cm2. For each run, 50 mg of SnO/In(OH)3/InP powder is dispersed in water, and a calculated amount of H2PtCl6 solution (1% (mass fraction)) is added under stirring. Ar is bubbled for 20 min before the reaction. The photocatalytic activity is measured by gas chromatography (Agilent 6820, Ar support). The light source is 300 W Xenon lamp equipped with different wavelength optical cut-off filters. For comparison, In(OH)3/InP and SnO are also tested under similar conditions.
To measure the AQE (apparent quantum efficiency), the different wavelength bandpass filters (440, 460, 490, 520, and 550 nm) are used. np is measured by a calibrated FU100 radiometer. The AQE of photocatalytic hydrogen production is calculated according to Eq. (1) without considering refraction and scattering loss of photons.
where is the production of hydrogen (μmol), and nP is the number of incident photons (μmol–1m–2s–1), nP = t× S× Q (t is the reaction time, S is the effective illumination area of the reactor, and Q is the number of incident photons measured by the radiometer).
2.4 Stability test
The stability of the catalyst is studied by successive reaction cycles under similar reaction conditions. After each 3 h, the reaction solution is centrifuged, washed with water, and then re-dispersed in water for cyclic reaction. More detailed information for working electrode preparation, electrochemical measurements and sample characterization are presented in the Electronic Supplementary Material.
3 Results and discussion
3.1 Structure and morphology characterization
The XRD patterns of the prepared samples show typical distinct diffraction peaks, which can be ascribed to In(OH)3, InP, and SnO (see Fig. 1 and Figs. S1–S2). A series of diffraction peaks centered at 22.2°, 31.6°, 35.5°, 39.0°, 45.4°, 51.1°, 56.4°, 66.8°, 70.8°, 75.3°, and 79.7° correspond to the (200), (220), (310), (222), (400), (420), (422), (440), (600), (620), and (622) planes of In(OH)3 (PDF#16-0161) [32], while the peaks centered at 26.2°, 30.4°, 51.5° and 79.9° correspond to the (111), (200), (311), and (422) planes of InP sphalerite structure (PDF#65-0233) [33]. In addition, the typical peaks of SnO are well consistent with the standard card (PDF#06-0395) [34]. The XRD patterns of the SnO/In(OH)3/InP and Pt/SnO/In(OH)3/InP composite photocatalysts show similar diffraction peak characteristics as In(OH)3/InP, indicating that the protective layer SnO and co-catalyst Pt do not change the crystal structure of the catalyst [35]. No obvious diffraction peak belonging to SnO is observed in XRD patterns due to the low loading of SnO in the prepared catalysts.
The microstructure and morphology of prepared SnO/In(OH)3/InP catalyst are investigated using TEM (Fig. 2(a)). The nanoparticles exhibit a typical aggregation due to the limited passivation during the hydrothermal reactions. In the HRTEM image of Fig. 2(b), three sets of ordered lattice structures show that the composite catalyst is constructed by the order of InP, In(OH)3, and SnO. Moreover, the multi-group lattice spacing of 0.29, 0.27, and 0.26 nm can be respectively indexed to the InP (200), In(OH)3 (220), and SnO (110) plane, indicating that the SnO/In(OH)3/InP photocatalyst is successfully synthesized. The lattice fringes of InP, In(OH)3, and SnO are 0.29, 0.27, and 0.26 nm respectively, indicating that the In(OH)3 transition layer can reduce the lattice mismatch between SnO and InP. It accelerates the charge separation and transfer during the reaction process. The Pt/SnO/In(OH)3/InP catalyst with tightly contacted interfaces shows similar morphological characteristics as SnO/In(OH)3/InP (Fig. S3(a)). Many Pt particles with 0.18 nm lattice fringes are observed on the catalyst surface (Fig. S3(b)), indicating the co-catalyst Pt is well deposited on the catalyst. This consequence can be further proved by the elemental mapping images of Pt/SnO/In(OH)3/InP due to relatively homogeneous element spots of In, P, Sn, O, and Pt (Fig. S4).
XPS is used to analyze the element chemical state of as-prepared catalysts. The survey XPS spectra of samples show typical Sn 3d, O 1S, In 3d, P 2P and Pt 4f peaks. No other peaks of impurities appeared (see the result in Figs. S5 and S6). Figure 3(a) shows clear peaks centered at 486.9 eV to 487.2 eV, which belong to typical characte-ristic Sn 3d5/2 peaks of SnO [36]. The O 1s peak shifts to the high energy side about 0.5 eV when SnO loaded onto In(OH)3/InP, indicating the strong interaction between SnO and InP via hydroxide species. The In 3d5/2 peaks located at 444.7 eV and 445.2 eV, indicating In atom neighbored with P and O atoms, possibly as InP and In(OH)3 like species [37–39]. Very clear P 2p peaks are observed, centered at 129.3 eV, 130.0 eV, and 133.2 eV, respectively. The former two peaks can be ascribed to the P in InP, the latter can be ascribed to the P neighbored with O (Fig. 3(d)) [40]. The Pt 4f centered at 72.6 eV indicates that the metallic Pt is located near the O atoms of SnO [41].
3.2 UV-vis diffuse reflection spectra
The light absorption properties of Pt/SnO/In(OH)3/InP series photocatalysts are studied by utilizing the ultraviolet-visible absorption technique (Fig. 4). According to Fig. 4(a), the Pt/SnO/In(OH)3/InP catalyst exhibits significantly enhanced light absorption and exhibits redshift of the absorption edge compared with In(OH)3/InP and SnO/In(OH)3/InP. It is known that SnO is a wide bandgap semiconductor (2.59 eV) and is transparent to visible light [42]. Coating SnO on catalyst surface will permit input light to reach InP and generate excited charges for the following reaction. The optical band gap of the prepared catalysts can be deduced from αhυ = А(hυ‒Eg)n/2 according to the absorption spectra in Fig. 4(a) [43], where α is the absorption coefficient, hυ is the light energy, А is the constant, Eg is the optical band gap, and n = 1 [35]. The Eg of In(OH)3/InP, SnO/In(OH)3/InP, and Pt/SnO/In(OH)3/InP is approximately 1.90, 1.84, and 1.76 eV, respectively, indicating that the Pt/SnO/In(OH)3/InP catalyst can absorb most of input light with a wavelength of up to 700 nm.
3.3 Photocatalytic activity and stability test
Figure 5 presents the activity results of catalysts for H2 generation from water under visible light irradiation. The SnO sample exhibits a very low activity, while SnO/In(OH)3/InP (3% (mass fraction) SnO) show a quite high activity for over-all water splitting to hydrogen and oxygen. The ratio of hydrogen to oxygen is nearly 2:1 (see the results in Fig. S7). Its activity is about twice higher than that of In(OH)3/InP. Although the In(OH)3/InP catalyst shows an initial activity for water splitting, it suffers corrosion. As mentioned later, its activity decays when the reaction is prolonged. The concentration of (PO4)3– in all catalyst dispersion were measured after 7 h of reaction, and it was found the concentration of (PO4)3– increased very significantly in In(OH)3/InP catalyst dispersion (see the results in Fig. S8). After 7 h of reaction, the concentration of (PO4)3– reached up to 5.0 μg/mL in the In(OH)3/InP catalyst. Although this datum is lower than that of the InP catalyst (about 6.0 μg/mL), the results indicate that serious corrosion occurs both in In(OH)3/InP and InP cases. However, after coating with SnO, only a trace amount of (PO4)3– is detected, implying that the SnO coating could inhibit catalyst photocorrosion significantly. Consequently, the activity-dependence on the amount of SnO over SnO/In(OH)3/InP catalyst (Fig. 6) was studied. When the SnO content increases from 3% to 10% (mass fraction), the activity is enhanced, but it decreases when SnO is overloaded. According to the stability tests of the multi-cycle test (see the results in Fig. 7), Pt/SnO/In(OH)3/InP exhibits a very good stability and a high activity for hydrogen generation in the successive cycle tests. After the fourth cycle test, that catalyst still gives 144.42 µmol/g of hydrogen in 3 h, which is almost the same compared to the hydrogen amount generated over fresh catalyst. The AQE of the Pt/SnO/In(OH)3/InP photocatalyst were also measured under different wavelengths of visible light (Fig. 8).
The hydrogen and oxygen production test of the Pt/SnO/In(OH)3/InP catalyst in pure water is conducted under visible light irradiation. In Fig. S7, the hydrogen and oxygen production generate continuously from catalyst dispersion at a hydrogen to oxygen ratio of nearly 2:1. To further prove that the Pt/SnO/In(OH)3/InP catalyst can decompose pure water photocatalytically, the isotope tracer experiment is performed. In Figs. 9(a) and 9(b), D2 is detected (m/z = 4) when D2O is used, and the signal of (m/z = 36) is detected when H218O is used.
3.4 Photoelectrochemical test
3.4.1 Photocurrent response
The photocurrents of synthesized samples are tested, whose results are presented in Fig. 10. The photocurrent density is closely connected with the applied bias voltage [44]. According to the results, Pt/SnO/In(OH)3/InP present the highest photocurrent both at 0 V and 1.23 V (RHE), which is about three times higher than that of the SnO/In(OH)3/InP sample at 1.23 V (RHE). The results indicate that coating SnO significantly enhances the separation and transfer of charges in a composite photocatalyst.
3.4.2 PL and TRPL
The synthesized samples are further studied by photoluminescence (PL) and time-resolved photoluminescence (TRPL) (see Fig. 11). According to the results in Fig. 11(a), In(OH)3/InP presents a strong photoluminescence emission peak centered at 797 nm under 532 nm excitation. Once the In(OH)3/InP is coated by SnO, the PL emission density is significantly decreased, indicating that the recombination of excited charges is remarkably inhibited. If the sample is further coated by the Pt particle, the PL emission density decreases, which might be caused by the charge transfer enhancement over catalyst by Pt loading. The time-resolved photoluminescence technique is used to get the lifetimes of synthesized samples, and the results are given in Fig. 11(b) and Table S1. The average photoluminescence lifetime of charges in different catalysts are quite different. Accordingly, the data of In(OH)3/InP, Pt/In(OH)3/InP and Pt/SnO/In(OH)3/InP catalysts are 0.050 ns, 0.061 ns, and 0.068 ns, respectively. Among them, the lifetime of photogenerated electrons of the Pt/In(OH)3/InP and SnO/In(OH)3/InP catalysts is shorter than that of the Pt/SnO/In(OH)3/InP catalysts, mainly due to the introduction of SnO and precious metal. Pt accelerates the migration of photogenerated carriers and makes electron-hole pairs effectively separate, so that the catalyst exhibits an excellent photocatalytic activity.
3.4.3 CV and LSV
Figure 12(a) gives a linear scan of the In(OH)3/InP, SnO/In(OH)3/InP, Pt/In(OH)3/InP, and Pt/SnO/In(OH)3/InP photocatalysts. The current intensity of each catalyst increases with the applied bias, indicating that the prepared samples are typical n-type semiconductors [45]. The Pt/SnO/In(OH)3/InP photocatalyst has a stronger current density than In(OH)3/InP, SnO/In(OH)3/InP, Pt/In(OH)3/InP. In addition, CV is also used to study the photoelectrochemical performance of the photocatalyst. As shown in Fig. 12(b), the current density of the Pt/SnO/In(OH)3/InP electrode is larger than that of In(OH)3/InP, SnO, and SnO/In(OH)3/InP, indicating that the electron transfer rate is significantly improved on Pt/SnO/In(OH)3/InP.
3.4.4 Mott-Schottky analysis
To measure the flat band potential of catalysts, Mott-Schottky analysis is performed (Fig. 13). InP is a typical p-type semiconductor [31]. According to Fig. 13, the flat band potential of In(OH)3/InP is −1.10 V versus SCE. The corresponding standard hydrogen electrode potential can be calculated according to Nernst equation [46], expressed as
where EFB is flat band potential, the pH value of the electrolyte is about 7, and ESCE = 0.24 V. The calculated EFB of the catalyst is –0.45 V versus NHE. The calculated CB and VB of the catalyst are –0.55 V and 0.88 V versus NHE, respectively. Based on the Hall effect test results (Table S2), SnO is a p-type semiconductor [34]. The EFB of SnO/In(OH)3/InP and Pt/SnO/In(OH)3/InP shifts downside due to the formation of heterojunction.
3.5 I-V curve
To measure the conductivity of the catalyst, the I-V test is conducted. The I-V curve of the sample (Fig. 14) indicates that the catalyst Pt/SnO/In(OH)3/InP presents the best conductivity compared with In(OH)3/InP and SnO/In(OH)3/InP [47], which is also in good agreement with activity dependence. At the same time, when the voltage is scanned from –10V to 10V, the current curve of In(OH)3/InP will fluctuate and shift dramatically, possibly due to the high resistance of In(OH)3/InP and defects in the material itself [48].
3.6 Photocatalytic mechanism
Based on the above results and analysis, a possible photocatalytic mechanism of Pt/SnO/In(OH)3/InP photocatalyst for water splitting is proposed (Fig. 15). When the p-type SnO is coated on In(OH)3/InP, the Fermi level of n-type In(OH)3/InP decreases until it reaches an equilibrium state on p-n heterojunction crystal interface and establishes a built-in electric field between the SnO and In(OH)3 [49]. Under the irradiation of visible light, SnO and InP simultaneously generate electrons and holes [31,34]. At the same time, the In(OH)3 with a wide bandgap cannot be excited under visible light, and it can only be used as a transition layer to transfer electrons because the CB of catalysts decreases in the order of SnO>In(OH)3>InP. The photogenerated electrons are transferred from the CB of SnO to the CB of In(OH)3, and then to the CB of InP. Finally, Pt acts as an electron capture center, capturing photo-generated electrons and catalyzing protons to generate H2. Because the valence band position of In(OH)3 is much lower than that of SnO and InP, most of the holes still remain in SnO and InP. As a consequence, the separation and transfer of electrons and holes are realized. Then the holes in SnO oxidize H2O to generate O2. Moreover, the In(OH)3 transition layer can effectively reduce the lattice matching between the SnO and InP, accelerating the transfer of photo-generated charge. The SnO anti-corrosion layer can effectively suppress the photocorrosion of In(OH)3 and InP because it blocks the direct contact of the catalysts from water. Therefore, the In(OH)3 transition layer and the SnO anti-corrosion layer accelerate the separation and transfer of charges and suppress the photocorrosion of the catalyst, improving the photocatalytic activity.
4 Conclusions
In summary, this study successfully constructs the SnO/In(OH)3/InP system by utilizing the two-step hydrothermal method. The anti-corrosion SnO layer and the In(OH)3 transition layer can effectively improve the stability and activity of the photocatalyst, thereby reducing the crystal mismatch between SnO and InP and increasing the charge transfer. The further introduction of the precious metal Pt accelerates its charge separation and transfer. The obtained Pt/SnO/In(OH)3/InP photocatalyst, under visible light irradiation without adding any electron donor, has a hydrogen production rate of 144.42 µmol/g within 3 h, which is 123 times higher than that of bare InP without obvious decay.