1 Introduction
Semiconductor-based photocatalysis can directly convert solar energy into chemical energy in hydrogen molecules via the water-splitting process, which has been widely accepted as a promising approach to meet the increasing energy demand [
1–
11]. Despite many processes made in recent years, the photocatalytic activity for H
2 evolution is still far from practical application, mainly due to the severe charge combination in the bulk and/or at the surface of semiconducting photocatalysts [
12–
15]. Thus, it is of great importance to make charge efficiently separate during the photocatalytic process [
16–
18].
To date, various strategies have been developed to promote charge separation in the bulk and/or at the surface of photocatalysts, such as constructing heterojunction [
19–
40], phase junction [
41–
45], and surface disordering engineering [
46–
49]. Among these approaches, the crystal-facet engineering has been widely reported to improve the separation of photo-generated charge carriers [
41,
50–
52]. For the faceted semiconductor nanocrystal enclosed by two different families of facets, the photo-induced electrons would be accumulated on one family of facets with a lower conduction band bottom, while the photo-induced holes would be accumulated on the other family of facets, due to the band level difference of the adjacent facets, thus leading to the spatial electron and hole separation for photoactive reduction and oxidation reactions, respectively. For example, Li et al. synthesized a water-oxidation photocatalyst of BiVO
4 decahedron enclosed by {010} and {110} facets, and found that the photo-excited electrons and holes could be transferred to the {010} and {110} facets, respectively [
53,
54]. They further investigated the anisotropic charge distribution on different facets of BiVO
4 decahedron via the spatially resolved SPV spectroscopy with nanoscale spatial resolution and mV sensitivity, demonstrating that the built-in electric fields in this faceted BiVO
4 resulted in an anisotropic charge transfer [
55]. The similar phenomenon was also reported on a facet-engineered chalcogenide photocatalyst [
56,
57]. Taking Cu
2WS
4 decahedra as model photocatalysts, Li et al. revealed that Pt nanoparticles were only deposited on the {001} facets of Cu
2WS
4 via the
in-situ photo-deposition method, and the {101} facets were etched during the photocatalytic reaction, indicating that the photogenerated electrons and holes could be transferred to the {001} and {101} facets of Cu
2WS
4 decahedra, respectively [
56].
Previously, cube and tetrahexahedron SrTiO
3 (STO) nanocrystals were synthesized enclosed with {023} and {001} facets [
58]. By tracking the deposition positions of Pt and PbO
2 in the
in-situ photo-deposition process, it was demonstrated that the photogenerated electrons and holes could be transferred to the low-indexed {001} facets and high-indexed {023} facets, respectively [
58]. Moreover, this anisotropic charge distribution on {001} and {023} facets still maintained even after the surface reconstruction of STO by hydrogenation [
49]. However, in these cases, only one family of facets was used for photocatalytic hydrogen evolution, while the other family of facets could not be involved in the hydrogen evolution reaction. As is known, photocatalytic water splitting is a heterogeneous catalysis process, which occurs only on the surface of photocatalyst. Thus, as the efficient facet utilization of the faceted photocatalyst for hydrogen evolution increases, the photocatalytic activity increases.
In this work, taking SrTiO3 nanocrystals enclosed by {023} and {001} facets as a model photocatalyst, a strategy was proposed to achieve the full-facets-utilization of the nanocrystal for photocatalytic hydrogen reaction by depositing Pt nanoparticles on all the exposed facets. The results of photo-deposition experiment of CdS demonstrated that the function of the {023} facets, which were used for oxidation reaction, could be reversed for photocatalytic hydrogen evolution after depositing Pt nanoparticles. Thus, the function-reversible {023} facets, together with the {001} facets, led to a higher activity of SrTiO3 nanocrystal for photocatalytic hydrogen, in contrast to that of SrTiO3 with Pt deposited by a photo-deposition method.
2 Materials and methods
2.1 Chemicals
Ethanolamine (EA) (NH2CH2CH2OH), sodium hydroxide (NaOH), strontium nitrate (Sr(NO3)2), sodium borohydride (NaBH4), methanol (CH3OH), ethylene glycol (EG), surfer powder, toluene (C7H8), cadmium nitrate tetrahydrate (Cd(NO3)2·4H2O) and tetrabutyl titanate (TBOT) (Ti(OC4H9)4) were purchased from Sinopharm Chemical Reagent Co., Ltd. Ammonium hexachloroplatinate ((NH4)2PtCl6) was offered by Alfa Aesar, A Johnson Matthey Company. All the materials were analytical grade and used without further purification, and the water used in all syntheses was deionized water with a resistivity of 18.25 MΩ·cm.
2.2 Synthesis of SrTiO3 (STO) nanocrystals
The cube and tetrahexahedron STO nanocrystals enclosed with {023} and {001} facets were synthesized by solvothermal methods, as described in Refs. [
49,
58]. Briefly, 10 mmol of TBOT was dissolved into 20 mL of EA to form a transparent solution, and then a 50 mL of NaOH (3 mol/L) solution was dropped to obtain a suspension. An aqueous solution (10 mL) containing 10 mmol of Sr(NO
3)
2 was added to the suspension under stirring for 30 min. The reaction mixture was sealed in a 100 mL teflon lined stainless-steel autoclave, followed by heat treatment at 180 °C for 24 h. After cooling, the resulted white precipitate was centrifuged and washed with ethanol and deionized water five times. Finally, SrTiO
3 nanocrystals enclosed with {023} and {001} facets (STO) were obtained.
2.3 Photo-deposition (PD) of Pt and PbO2 on STO
Photo-deposition of 2 wt.% (wt., mass fraction) Pt on the surface of STO was conducted by using (NH4)2PtCl6 as the precursor. Briefly, the as-prepared STO sample (100 mg) was dispersed in an aqueous solution (220 mL) containing 20 vol.% (vol., volume fraction) methanol by a magnetic stirrer in a side irradiation Pyrex cell. A calculated (NH4)2PtCl6 aqueous solution was added to the solution. After evacuation with N2 gas, the suspension was irradiated by a 300 W Xe lamp for 5 h. Finally, the precipitate was centrifuged and washed with deionized water five times, and dried under vacuum at 80 °C overnight. The obtained sample was assigned as STO-PD. Similarly, photo-deposition of 5 wt.% PbO2 on the surface of STO was performed by using Pb(NO3)2 as the precursor and KIO3 as electron acceptor. The obtained sample was assigned as STO-PbO2.
2.4 Chemical deposition (CD) of Pt on STO
100 mg of the as-prepared STO was dispersed in 17 mL of EG in a three-necked flask by using sonication, and a calculated EG solution containing (NH4)2PtCl6 was then added into the three-necked flask under stirring for 30 min. Then, the three-necked flask was heated by an oil bath and maintained at 60 °C for 30 min. Finally, 167 mg of NaBH4 was then added to the suspension and the suspension was stirred for 10 h. After cooling, the precipitate was centrifuged and washed with deionized water several times and dried under vacuum at 80 °C overnight. The obtained sample was assigned as STO-CD.
2.5 Photo-deposition of CdS on STO-CD
To investigate the role of Pt on {023} facets in the photocatalytic hydrogen reaction, CdS was selected to be deposited on STO-CD by a photo-deposition method. Briefly, 50 mg of STO-CD was dispersed in 200 mL of methanol. Then, 0.44 mL of toluene containing 0.44 mg of S powder and 3 mL of Cd(NO3)2 solution (2 mg/mL) was injected into the suspension. After evacuation with N2 gas, the suspension was irradiated by a 300 W Xe lamp for 5 h. Finally, the precipitate was centrifuged and washed with toluene and methanol five times, and dried under vacuum at 80 °C overnight. The obtained sample was assigned as STO-CD-CdS.
2.6 Photocatalytic reaction
Photocatalytic hydrogen evolution was conducted in a Pyrex cell under full-wavelength irradiation. Briefly, 50 mg of photocatalyst powder was dispersed into 220 mL of 20 vol.% methanol solution by sonication for 30 min. After the reactor was purged with N2 gas to eliminate air for 30 min, the suspension was irradiated by a 300 W Xe lamp for 5 h. The amount of H2 gas was determined using an online thermal conductivity detector (TCD) gas chromatography (NaX zeolite column, TCD detector, N2 carrier). The apparent quantum efficiency (AQE) was measured under visible light irradiation (λ = 350 ± 10 nm), and was calculated by the equation: AQE = Ne/Np× 100%, where Ne is the number of reacted electrons and Np is the number of incident photons.
2.7 Characterization
The optical properties of the samples were measured by a UV-vis-near-IR spectrophotometer (HITACHI U4100) equipped with a labsphere diffuse reflectance accessory within a 240−800 nm range, using BaSO4 as the reference. The powder X-ray diffraction (XRD) characterization were conducted on a PANalytical X’pert MPD Pro diffractometer with Ni-filtered Cu Kα irradiation (λ = 1.5406 Å). The transmission electron microscopy (TEM) images, high-angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) images, and high-resolution TEM (HRTEM) images were observed by a JEOL JEM-F200 transmission electron microscope. X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo Fisher ESCALAB Xi+ with monochromatic Al Kα radiation (hν = 1486.69 eV). All binding energies were referenced to the C 1s peak at 284.8 eV. The elemental analysis of the as-prepared samples was performed by an OXFORDMAX-80 energy-dispersive X-ray detector (EDX) which was mounted in the JEOL JEM-F200 transmission electron microscope.
3 Results and discussion
The cube and tetrahexahedron STO nanocrystals enclosed with {023} and {001} facets were synthesized by the solvothermal method, as shown in Fig. S1(a). Figures S1(b) and S1(c) show the scanning electron microscopy (SEM) image of STO loaded with Pt and PbO
2 nanoparticles by the photo-deposition method, respectively. It was clearly observed that Pt nanoparticles and PbO
2 nanoparticles were deposited on the surface of {001} and {023} facets of STO, respectively. This result revealed that the {023} and {001} facets provided active sites for photo-oxidation reaction and photo-reduction reaction (Fig. S1(d)), respectively. More details about the charge separation on the {001} and {023} facets of STO could be obtained in Refs. [
49,
58]. During the photocatalytic process, the photoexcited electrons would be trapped by the Pt nanoparticles deposited on the {001} facets of STO nanocrystals and take part in the photocatalytic hydrogen reaction (Fig.1(a)) [
49,
58]. As is known, the photocatalytic reaction occurs on the surface of photocatalyst. In the system of photocatalytic hydrogen evolution with hole scavengers (i.e., CH
3OH), the sacrificial agent would rapidly react with photo-generated holes, which was not the rate-determining step during the photocatalytic process, thus only leaving the photo-generated electrons in photocatalyst for hydrogen evolution. Consequently, the photocatalytic H
2 evolution reaction should be the rate-determining step in such a photocatalytic system. Hence, as the facets designed for the photocatalytic hydrogen evolution increase, the performance the photocatalyst possesses increases. Therefore, if the photo-reduction active sites (i.e., Pt cocatalyst) were constructed on all facets of STO, the {023} facets utilized for oxidation reaction would be reversibly activated for hydrogen evolution, together with the {001} facets, leading to the higher hydrogen evolution performance (Fig.1(b)).
Inspired by this assumption, Pt cocatalysts were deposited on both {001} and {023} facets of STO via a chemical deposition method (STO-CD). As a reference, the STO with Pt cocatalysts selectively deposited only on {001} facet (STO-PD) was synthesized by the photo-deposition method. XRD diffraction patterns in Fig. S2(a) revealed that both STO-PD and STO-CD possessed cubic-phase perovskite structure (JCPDS No. 00-035-0734, space group: Pm-3m), and no peaks for Pt were observed due to the relatively low content of Pt in these two samples. The SEM in Fig. S3 indicated that islanded Pt nanoparticles were distributed on both {001} and {023} facets of STO (Fig. S3), while for STO-PD, Pt nanoparticles were mainly deposited on the {001} facets of STO via the photo-deposition method (Fig. S1(b)), which was consistent with Refs. [
49,
58]. Furthermore, the XPS characterization in Fig. S4 revealed that the Pt nanoparticles deposited by both chemical deposition and photo-deposition methods were in the form of Pt(0), suggesting that the chemical states of Pt cocatalyst in these two samples should have no influence on their photocatalytic activities.
To directly observe the Pt distribution on the surface of STO by chemical deposition method in detail, the TEM image of a typical STO-CD nanocrystal was recorded (Fig.2(a)). Viewed from [100] direction, the STO-CD nanocrystal showed a dodecagon shape, consistent with Refs. [
49,
58]. It can be seen that there were some small particles with a size of less than 5 nm on both {023} and {001} facets of STO-CD nanocrystal, with the corresponding model of STO-CD illustrated by the inset
i in Fig.2(a). To further investigate the small particles in detail, high-resolution TEM (HRTEM) images were recorded from the rectangle I, II, IV, and III regions, as shown in Fig.2(b)–Fig.2(e), respectively. The interplanar space of the nanoparticles shown in Fig.2(b)–Fig.2(e) was measured to be 2.27 Å, corresponding to the (111) plane of Pt metal. Fig.2(f) shows the energy-dispersive X-ray (EDX) spectrum of the point labeled in inset
iii (Fig.2(a)), which was recorded from the rectangle V region in inset
ii in Fig.2(a). The Cu signal came from the Cu grid, while the signals of Ti, Sr, and O came from SrTiO
3. The obvious Pt signal provided direct evidence confirming the existence of the Pt nanoparticles deposited on {023} facets of STO-CD nanocrystal. Therefore, it could be concluded that the Pt cocatalysts were successfully deposited on both {001} and {023} facets of STO by using the chemical deposition method. In addition, TEM and HRTEM images of STO-PD were also recorded to observe the Pt nanoparticles on the surface of STO-PD, as shown in Fig. S5. It can be seen that Pt nanoparticles are only deposited on the {001} facets of STO, and the size of Pt in STO-PD is less than 5 nm, which is similar to that of Pt in STO-CD, suggesting that the size of Pt nanoparticles should not contribute to the different photocatalytic activities of STO-PD and STO-CD. Moreover, without other variables introduced in the experiment, the only difference between these two samples is the position of deposited Pt, one only on {001} facets (STO-PD) and the other on both {001} and {023} facets (STO-CD).
Fig.3(a) shows the time-course photocatalytic hydrogen evolution of STO-PD and STO-CD, respectively. Obviously, the samples with Pt deposited on both {001} and {023} facets, i.e., STO-CD, exhibit a higher photocatalytic activity, which is ca. 2 times that of STO with Pt deposited on only {001} facets by the photo-deposition method (STO-PD) (Fig.3(b)). In addition, the AQE at 350 nm of the as-prepared STO-CD was calculated to be 0.83%. For the STO-PD sample, only {001} facets were used for the photocatalytic hydrogen evolution. In comparison, for STO-CD, Pt nanoparticles were deposited on both {001} and {023} facets, which may make all of the facets be activated for hydrogen evolution. Moreover, no obvious decay in photocatalytic H2 evolution activity was observed for STO-CD over six consecutive cycles (Fig. S6), indicating the high stability of STO-CD.
To further demonstrate that the {001} and {023} facets deposited with Pt cocatalyst can be used for photocatalytic reduction reaction, i.e., hydrogen evolution, a series of ex-situ experiments were designed to identify the successful activation of {023} facets by Pt deposition. Briefly, the as-prepared STO-CD were dispersed in a methanol solution containing S powder and Cd(NO3)2 under the irradiation from a 300 W Xe lamp, and the obtained sample was assigned as STO-CD-CdS (Fig.4). As is known, photo-induced electrons and holes would be generated in SrTiO3 when it is excited by the photons with an energy higher than its bandgap. During the photocatalytic process, the photo-excited electrons would be trapped by the Pt nanoparticles on the surface of photocatalyst to participate in the reduction reaction. Thus, in this designed ex-situ experiment, the electrons trapped by Pt nanoparticles are supposed to react with S to produce S2− (S + 2e−→ S2−). The produced S2− would then react with Cd2+ near Pt nanoparticles to generate CdS (S2− + Cd2−→ CdS). Consequently, these CdS monomers would be nucleated on the surface of Pt nanoparticles to form a core-shell structured Pt@CdS on the surface of STO. Therefore, if both {023} and {001} facets deposited with Pt were used for photolytic hydrogen evolution, the CdS nanoshell would be deposited on the surface of Pt on both two facets of STO-CD (right top in Fig.4). In other words, if only the {001} facets deposited with Pt nanoparticles was used for photocatalytic hydrogen reaction, the CdS nanoshell would be only deposited on the surface of Pt on {001} facets (right bottom in Fig.4).
With the model of STO-CD-CdS proposed in Fig.5(a), after the photodeposition reaction, there were many islanded nanoparticles on both {023} and {001} facets of STO (Fig.5(b)) in the size of ca. 10 nm, which was much larger than that of Pt on STO-CD (Fig. S3), suggesting that CdS nanoshells were deposited on the surface of Pt nanoparticles on both {023} and {001} facets. The TEM and HADDF-STEM images of STO-CD-CdS were recorded to investigate its crystal structure in detail, as shown in Fig.5(c) and Fig.5(d). The typical STO-CD-CdS particle showed a 12-side polygon with lots of nanoparticles attached on its outline viewed from [001] direction (Fig.5(c)). The dark contrast of the inside core (indicated by red arrow) and the light contrast of the shell (indicated by blue arrow) of the nanoparticles indicated the successful deposition of CdS nanoshell on the surface of Pt nanoparticles. Furthermore, HRTEM images were recorded from the rectangle I, II, IV, and III regions in Fig.5(c) to investigate the structure of the deposited Pt@CdS nanoparticles in STO-CD-CdS. As shown in Fig.5(e)–Fig.5(h), the interplanar spacings of 3.36 and 2.27 Å corresponded to the (111) plane of CdS shells and (111) plane of Pt nanoparticles, respectively, indicating that the Pt@CdS core@shell structure was successfully deposited on the surface of STO-CD. The obvious S and Cd signals in Fig.5(i) also confirmed the existence of CdS nanoshells deposited on the surface of Pt nanoparticles.
In the XRD pattern (Fig. S2(b)), a weak peak can be observed at 2θ = 26.5°, which can be indexed to the (111) plane of CdS, also confirming the existence in STO-CD-CdS. Furthermore, as shown in Fig. S7, the obvious absorption band in the range of ca. 400–550 nm for STO-CD-CdS should be ascribed to the light absorption from CdS shells on Pt nanoparticles, providing another evidence for the existence of CdS in STO-CD-CdS. The Cd 3d and S 2p XPS spectra (Fig. S8) revealed that the chemical states of Cd and S were +2 and −2, respectively, further verifying the existence of CdS in STO-CD-CdS. Based on these analysis results, it is convincing that the CdS nanoshells have been photo-deposited on the surface of Pt nanoparticles, which are chemically deposited on both {001} and {023} facets of STO. That is to say, the function of {023} facets, which are responsible for photooxidation reaction, can be reversibly activated by chemical deposition of Pt for photocatalytic reduction reaction for hydrogen evolution. Thus, the full utilization of {001} and {023} facets are believed to contribute to the great improvement in photocatalytic hydrogen activity for STO-CD, as compared to STO-PD. A plausible mechanism of charge transfer in STO-CD is proposed in Fig.6. During the photocatalytic process, photo-induced electrons and holes would be generated on both {001} and {023} facets of STO when it is excited by the photons with energy higher than its bandgap. For the STO-CD, Pt nanoparticles were deposited on all the {001} and {023} facets. The photoexcited electrons on both {001} and {023} facets would be trapped by the Pt nanoparticles, and take part in the hydrogen evolution, while the photoexcited holes would be transferred from the valence band maximum (VBM) of {001} facets to VBM of {023} facets, and participate in the CH3OH oxidation reaction during the photocatalytic process.
4 Conclusions
In summary, taking cube and tetrahexahedron STO nanocrystal enclosed with {023} and {001} facets as a model photocatalyst, full-facets-utilization of the nanocrystal for enhanced photocatalytic hydrogen reaction were achieved by depositing Pt nanoparticles on all the exposed facets. The activity of STO with Pt deposited on both {001} and {023} facets was ca. 2 times that of SrTiO3 with Pt deposited on only {001} facets. The successful photodeposition of CdS shell on the surface of Pt deposited on all facets demonstrated that the function of the {023} facets, which were used for oxidation reaction, could be reversed for photocatalytic hydrogen evolution after depositing Pt cocatalyst. Thus, the function-reversible {023} facets, together with the {001} facets, led to the higher activity of SrTiO3 nanocrystal for photocatalytic hydrogen evolution.