Disordered layers on WO₃ nanoparticles enable photochemical generation of hydrogen from water

Abstract

Tailored defects on a semiconductor surface can provide active catalytic sites and effectively tune the electronic structure for suitable optical properties. Herein, we report that surface modification of WO₃ with a disordered layer enables the photochemical hydrogen production from water. A simple room temperature solution process with lithium-ethylenediamine (Li-EDA) alters the surface of WO₃ with localized defects that form a thin disordered layer. Both structural and optical characterization reveal that such a disordered layer induces an upshift in the Fermi level and the elevation of the conduction band of WO₃ above the hydrogen reduction potential. Using an alkaline sacrificial agent, Li-EDA treated WO₃ shows a co-catalyst-free photochemical hydrogen evolution rate of 94.2 μmol g⁻¹ h⁻¹ under simulated sunlight. To the best of our knowledge, this is the first example of using WO₃ as a direct photocatalyst for hydrogen generation from water via simple surface defect engineering.

---

Introduction

Solar energy-driven hydrogen generation from water can offer a clean and sustainable solution to address the growing energy demand with reduced fossil fuel dependence and associated environmental issues.^1–3 In recent years, much effort has been made to develop effective photocatalysts based on abundant and economical elements.^4–7 However, their overall photocatalytic efficiencies are insufficient to replace those made of noble metals. Tuning the light absorption, charge kinetics, and surface activity of catalysts can elevate the photocatalytic performance, but it is still challenging to achieve the activity of noble metal-based ones.

Semiconductor metal oxides, with advantages of natural abundance, chemical stability, and low cost, have been widely used as photocatalysts and electrode materials for energy conversion and storage devices.⁸ Both electronic and optical properties of metal oxides can be regulated by the type and population of defect introduced. The creation of intrinsic defects, such as oxygen vacancies, and doping of low valence metal centers or non-metal atoms are the most commonly used defect engineering methods for tailoring the photocatalytic activity.^9–12 In particular, oxygen vacancies are known to serve as shallow donors for metal oxides, such as TiO₂,¹³ WO₃,¹⁴ and SnO₂.¹⁵ It is generally accepted that the increased carrier density improves the electrical conductivity and facilitates the charge transfer at the interface between the metal oxide and substrate.

The introduced defects can be extended to form an internal boundary⁷ or even a disordered layer,¹⁰ and are relatively less studied than atomic point defects. A surface disordered layer arises from the large amount of lattice disorder on the crystalline surface after being treated under reducing conditions. Such layers usually contain a significant amount of oxygen vacancies and low valence metal centers (self-doped), which may create mid-gap states below the Fermi level of the metal oxide.¹⁶ The mid-gap states form a continuum that extends to and overlaps with the conduction band (CB) or merges with the valence band (VB). This can effectively reduce the band gap and enhance the light absorption in the visible and near-IR regions.^17,18 These extended energy states and the newly produced energy levels may become dominant for optical excitation and relaxation, and thus new applications may be obtained. Furthermore, the localized holes generated in the mid-gap state of the disordered layer efficiently inhibit the charge recombination,^16,19–22 where the low valence metal centers act as trapping sites for photogenerated carriers.

The pioneering work on black TiO₂ with disordered surface layers¹⁶ demonstrated a significant enhancement in the photocatalytic hydrogen evolution reaction (HER) rate and has sparked the investigations on creating disordered layers in other metal oxides, including Fe₂O₃,²³ ZnO,²⁴ BiVO₄,²⁵ and WO₃.^26,27 Among them, tungsten oxide (WO₃) is an important n-type semiconductor with various applications ranging from dye-sensitized solar cells²⁸ to electrochromic devices,²⁹ gas sensors,³⁰ and photoelectrochemical water splitting,³¹ owing to its low cost and good conductivity. With a desirable bandgap energy (2.5–2.8 eV) for capturing visible light and a number of stable oxygen-deficient sub-stoichiometric structures (WO_3−x, where 0 < x ≤ 0.28) available, WO₃ is one of the most attractive candidates for photocatalytic water splitting.³² Nonetheless, restricted by its intrinsic electronic structure, WO₃ is only suitable for water oxidation.³³ The position of the valence band maximum (VBM) of WO₃ is located at a relatively high energy level (3.0 eV vs. reversible hydrogen electrode, RHE), whereas the conduction band minimum (CBM) lies at a position lower than the hydrogen reduction potential (0 V vs. RHE).

Recently, a thin disordered layer on the WO₃ surface has been reported to significantly increase the charge carrier density and facilitate the interface charge transfer, thus improving the charge separation efficiency and conductivity in photoelectrocatalytic water oxidation.^26,27 Localized surface atomic disorder was also engaged in turning the inert WO₃ into an active electrocatalyst for hydrogen generation from acidic water.³⁴ Nevertheless, it is still challenging to trigger solar water reduction using WO₃ and the effect of the surface disordered layer on the band structure of WO₃ has not been clearly understood yet. Herein, we demonstrate that a simple room-temperature solution processing method can modify the surface morphology and electronic structure of WO₃ to enable the unprecedented capability of facilitating the photocatalytic HER without a co-catalyst. Using Li-ethylenediamine (Li-EDA) as a disordering agent, a surface disordered layer was successfully created on the crystalline WO₃ nanoplates (see Scheme 1 for synthetic schematics). Experimental results confirm the formation of a mid-gap state that causes a blue-shift in both the CBM and Fermi level (E_f) with respect to those of pristine WO₃. The reformed band structure of Li-EDA treated WO₃ (LT-WO₃) allows the photocatalytic hydrogen evolution without the assistance of a co-catalyst at a steady rate of 94.2 μmol g⁻¹ h⁻¹ under simulated sunlight. With the first example of using WO₃ as a co-catalyst-free photocatalyst for the HER, we demonstrate that Li-EDA treatment can be applied as a simple yet effective strategy to tailor the optical band structure of metal oxides via the formation of a thin disordered surface layer.

Scheme 1 Schematic illustration of the synthetic procedure for WO₃ with a surface disordered layer (LT-WO₃).

Experimental section

Synthesis of WO₃ nanoplates

Sodium tungstate solution (0.4 mol L⁻¹) was added to excess hydrochloric acid solution (1 mol L⁻¹) under vigorous stirring to obtain the tungstic acid nanoparticle precipitate. The yellow precipitate was collected after centrifugation, and washed with water 3 times. Afterwards, the as-prepared solids were calcined at 550 °C under an Ar atmosphere for 3 h. The product was collected after the furnace cooled down to room temperature.

Synthesis of LT-WO₃ nanoplates

The as-synthesized WO₃ nanoplates were immersed in solvated electron solution, Li-EDA (1 mol L⁻¹), at room temperature with continuous stirring for 4 hours. After quenching the reaction with deionized water, the resulting dark blue precipitate was collected using centrifugation. The solids were washed with HCl solution (0.5 mol L⁻¹) and ethanol alternately in order to remove any physically adsorbed Li and EDA. It was then dried under vacuum at room temperature overnight.

Photocatalytic H₂ production

In a typical run, a 5 mg sample was suspended in a 25 mL mixed solution of Na₂S (0.35 M) and Na₂SO₃ (0.25 M). The solution was degassed with N₂ for 30 minutes. Afterwards, the as-prepared solution was irradiated with a Newport solar simulator (150 W Xe lamp, ozone free, Air Mass Filter, AM 1.5 Global) with 100 mW cm⁻² light intensity under stirring. The amount of generated H₂ was analyzed using an Agilent 7890B gas chromatograph system equipped with a TCD detector using N₂ as the carrier gas.

Sample characterization

The crystal structure of samples was identified using an X-ray diffractometer (XRD, Rigaku SmartLab, 9 kW) with a diffraction angle (2θ) ranging from 10 to 70°. TEM images were obtained using a STEM (JEOL JEM-2100F) operated with a field emission gun at 200 kV. The samples were first dispersed in ethanol and drop-cast onto holey carbon-coated 400 mesh copper grids. The UV-vis diffuse reflectance spectroscopy was performed using a UV-vis spectrophotometer (Cary 4000) with ultrapure BaSO₄ as the reference. Raman spectra were obtained using a Raman spectrometer (Renishaw inVia) with an excitation wavelength of 785 nm at room temperature. XPS measurements were performed using a K-alpha X-ray photoelectron spectrometer (Thermo Fisher Scientific, UK) with a monochromatic Al Kα X-ray source (excitation energy = 1468.6 eV) and a base pressure of 3 × 10⁻⁹ torr in the analytical chamber, and with an internal reference of C 1s at 284.0 eV. EPR measurements were performed on an EPR spectrometer (SPINSCAN X, ADANI) operating at 9.3 GHz (X band) using a 100 kHz field modulation. All the spectra were recorded at 298 K using a microwave power of 10 mW and accumulated three times.

Electrochemical measurements

All the measurements were performed on a CHI1030A electrochemical workstation. 4 mg of catalyst and 200 μL of 5 wt% Nafion solution were dispersed in 4 mL of water/ethanol (3 : 1 v/v) by 10 min sonication to form a homogeneous suspension. Then, 10 μL of catalyst ink was drop-cast onto a glassy carbon electrode of 3 mm diameter (surface area = 7.07 mm²) as the working electrode with a constant mass loading of 0.14 mg cm⁻². A linear sweep voltammogram from 0 to −0.6 V was measured at a scan rate of 1 mV s⁻¹ in 0.5 M H₂SO₄ using a saturated calomel electrode and a Pt wire as the reference electrode and counter electrode, respectively. Pt/C (20 wt% platinum on graphitized carbon) is commercially available from Sigma-Aldrich. The Mott–Schottky plot was constructed based on the electrochemical data obtained using a typical three electrode system in phosphate buffer solution of pH 7.4. The Ag/AgCl (in 3 M NaCl solution) electrode and graphite were used as a reference and counter electrode, respectively. The data were collected by measuring the space-charge layer capacitance over a range of applied potentials (from −1.0 to 1.0 V) at 1000 Hz using a CHI760E electrochemical station. The recorded potential was referenced to the reversible hydrogen electrode (RHE) according to the following equation: E (vs. RHE) = E (vs. Ag/AgCl) + 0.197 V + 0.0592 × pH.

Results and discussion

Monoclinic WO₃ nanoplates were first synthesized by directly annealing tungstic acid at 550 °C, which was afforded by a simple co-precipitation of sodium tungstate solution with HCl. To introduce a disordered layer, the as-prepared WO₃ nanoplates were dispersed and stirred in Li-EDA solution³⁵ at room temperature for 4 h. During the Li-EDA treatment, a distinct color change from pale yellow to dark blue was observed (Fig. 1a). The UV-vis diffuse reflectance spectra (UV-vis DRS, Fig. 1b) reveal that the light response of Li-EDA treated WO₃ (LT-WO₃) has been extended, as evident from a significant absorption tail in the visible region (450 to 800 nm).

Fig. 1 (a) Photographs of pristine WO₃ and LT-WO₃ obtained via Li-EDA treatment. (b) UV-vis diffuse reflectance spectra (UV-vis DRS) of WO₃ and LT-WO₃. (c) XRD and (d) Raman spectra of WO₃ and LT-WO₃.

Such a change in optical properties can arise from the oxygen vacancies created on the surface or modified crystal phase.^36,37 X-ray diffraction (XRD) analysis was conducted to investigate the structural changes induced by Li-EDA treatment. As shown in Fig. 1c, the XRD pattern of pristine WO₃ matches well with that of the monoclinic phase of WO₃ (JCPDS no. 43-1035) where the sharp diffraction peaks indicate its crystalline nature. LT-WO₃ also shows a similar pattern of diffraction peaks, indicating that the crystal structure is retained. However, the peaks are broader with much weaker intensities compared to those observed for pristine WO₃, suggesting reduced crystallinity due to the localized modification of the crystal structure.³⁶

A similar change was observed in the Raman spectra (Fig. 1d), where the characteristic Raman peaks of WO₃ are broadened after the Li-EDA treatment. The two peaks at 273 and 326 cm⁻¹, which are assigned to the W⁶⁺–O–W⁶⁺ bending vibration mode of bridging oxygen in the WO₃ crystal lattice,³⁸ almost disappeared in the LT-WO₃ spectrum, suggesting the partial removal of oxygen atoms. The peak at 716 cm⁻¹ that is attributed to the long W⁶⁺–O–W⁶⁺ bond stretching mode was also markedly reduced with a slight shift. In addition, the peak for the short W⁶⁺–O–W⁶⁺ bond stretching at 808 cm⁻¹ shows an appreciable broadening that is correlated to the weakening of the W–O bond.³⁸ Furthermore, a small peak at 430 cm⁻¹ appeared after the Li-EDA treatment, indicating the presence of W species at lower oxidation states.^38–40 The reduction of W⁶⁺ to a lower oxidation state by Li-EDA is believed to weaken the W–O bond, which in turn leads to the partial loss of oxygen from the lattice. This is supported by the electron paramagnetic resonance (EPR) studies where only LT-WO₃ shows a signal at a g-value of 2.002, an indication of oxygen vacancies (Fig. 2).¹⁴ Both Raman and EPR results are in good agreement with the observations from XRD measurements.

Fig. 2 EPR spectra of WO₃ and LT-WO₃.

Fig. 3 compares the transmission electron microscopy (TEM) images of the as-prepared WO₃ and LT-WO₃ nanoplates. The typical morphology of WO₃ is rectangular nanoplates (with a mean side length of ca. 130 nm) with some smaller nanoplates of irregular shapes (Fig. 3a). A high resolution TEM (HR-TEM) image shown in Fig. 3b clearly reveals well-ordered lattice fringes of 0.366 and 0.378 nm that can be indexed to the (200) and (020) planes of monoclinic phase WO₃, respectively. The corresponding selected area electron diffraction (SAED) pattern (inset in Fig. 3b) shows the single-crystal spots that are in good agreement with the (200) and (020) planes. On the other hand, the TEM image of LT-WO₃ (Fig. 3c) shows a similar but more round-shaped morphology with a slightly reduced average side length (ca. 77 nm). Moreover, in the HR-TEM image (Fig. 3d), a thin disordered layer (average thickness = ca. 2 nm) is evident along the edge, which is supported by the disappearance of lattice fringes and the corresponding SAED pattern (inset in Fig. 3d). The formation of the disordered layer is attributed to the removal of surface lattice oxygen by solvated electrons during the treatment with Li-EDA, which is a strong reducing agent.⁴¹ By reacting with Li-EDA, W⁶⁺ is reduced to a lower oxidation state, and oxygen atoms are removed from the surface lattice.^18,35,42 The removal of oxygen atoms would lead to the crystal lattice deformation on the surface, leading to the decrease in both crystallinity and size of the nanoplates as observed from the XRD, TEM, and Raman spectroscopic studies.

Fig. 3 Electron microscopy analysis of WO₃ and LT-WO₃. (a) Low-resolution and (b) high-resolution TEM images of WO₃. (c) Low-resolution and (d) high-resolution TEM images of LT-WO₃. The insets in (b) and (d) show the FFT patterns of selected areas (orange squares: monoclinic phase core; green squares: disordered layer).

It is interesting that the core part of LT-WO₃ still retained the crystallinity as manifested by the clear lattice fringes (Fig. 3d). We have conducted a series of time-resolved Li-EDA treatments to investigate the formation of the disordered layer and control its thickness (Fig. 4). The disordered layer starts to form on the edges after 2 h of reaction and gradually grows thicker with treatment time. The thickness of the disordered layer reached up to 5 nm after 7 h; however, further prolonged reaction did not change the thickness, but only reduced the nanoplate size. The reaction between Li-EDA and WO₃ initially occurs at the Li-EDA/WO₃ interface. In the early stage, solvated electrons can diffuse through the thin disordered layer formed on the surface to reduce the W⁶⁺ ions in the inner crystal lattice. As the disordered layer becomes thicker, it acts as a passivation layer that hinders the diffusion of the electrons into the inner part. With further extended reaction time, the size of the core shrinks, as the outer layer is lost. These results suggest that the surface disordered layer is closely associated with the formation of oxygen vacancies.

Fig. 4 Time-resolved TEM analysis of the thickness of the disordered layer in LT-WO₃ nanoplates treated for 2, 4, 5, and 7 h.

To confirm the formation of the disordered layer, the surface species was characterized using X-ray photoelectron spectroscopy (XPS). Similar survey XPS spectra were obtained from pristine WO₃ and LT-WO₃, where only W and O were identified, with a small C 1s peak due to the adsorbed carbon-containing species, eliminating the possibility of Li contamination (binding energy of Li 1s is ca. 55.0 eV), especially in LT-WO₃ (Fig. 5). In the normalized W 4f core level spectra (Fig. 6a), only W⁶⁺ ions are detected for pristine WO₃, with the peaks at binding energies of 35.7 and 37.8 eV assigned to W 4f_5/2 and W 4f_7/2, respectively.¹⁴ In contrast, the W 4f spectrum of LT-WO₃ shifts to a lower binding energy and becomes broader, indicating the presence of lower oxidation states. In addition to the W⁶⁺ state, W⁵⁺ ions appear as a shoulder at binding energies of 34.5 and 36.4 eV. This is in good accordance with the Raman studies and further validates the reduction of W⁶⁺ to a lower oxidation state by Li-EDA. Additionally, a broad peak was seen in the O 1s spectrum of LT-WO₃ (Fig. 6b), which can be deconvoluted into two peaks. The peak centered at 530.4 eV is ascribed to the lattice O²⁻ species (W–O–W),¹⁴ while the other peak centered at 531.6 eV is assigned to O²⁻ in the vicinity of oxygen vacancies.⁴³ Meanwhile, the O 1s spectrum of pristine WO₃ shows a major peak for lattice oxygen atoms at 530.3 eV with a small peak at 531.6 eV, presumably due to the surface contamination during sample preparation. It is noted that both W 4f and O 1s peaks of LT-WO₃ exhibit binding energies slightly shifted to the lower values, ca. 0.2 and 0.1 eV, respectively. Similar shifts were observed for WO₃ after the chemical treatment as a result of the change in the electronic structure.^26,44

Fig. 5 XPS survey spectra of (a) WO₃ and (b) LT-WO₃.

Fig. 6 (a) W 4f and (b) O 1s XPS spectra of LT-WO₃ (top) and WO₃ (bottom).

A disordered atomic arrangement is known to affect the electronic structure, and thus the physicochemical properties of WO₃.^16,25,26 The observed change in the optical absorption of LT-WO₃ is related to the surface disordered layer. The band gap energies calculated by fitting the UV-vis DRS to the Tauc equation are 2.68 and 2.74 eV for WO₃ and LT-WO₃, respectively (Fig. S1†). Despite the obvious color change, the optical band gap energies (E_g) did not show much difference. However, the valence band XPS (VB-XPS) spectra reveal the changes in the electronic structure and band offset (Fig. 7a). The position of the VB maximum (VBM) was first evaluated by linearly extrapolating the onset of XPS valence band spectra to the baseline, which reflects a band edge position with respect to the Fermi level (E_f). The VBM edge levels of WO₃ and LT-WO₃ were determined to be 2.63 and 2.43 eV, respectively, indicating the upshift of the VBM by the induced oxygen vacancies. Interestingly, the VB-XPS spectrum of LT-WO₃ also exhibits the VB tail state (1.31 eV with respect to E_f), suggesting the formation of an additional electronic state. This mid-gap state would essentially broaden the VB width to create a reduced band gap and consequently allows the wider visible light absorption and change in sample color to dark blue.⁴⁵ We also used the Mott–Schottky plots to determine the E_f levels and thus obtain the band structures of samples.^46,47Fig. 7b shows the flat band potentials, which shifted from 0.5 to 0.25 V after Li-EDA treatment. The flat band potential reflects the difference between the E_f level and the water reduction potential of the sample. This shift of E_f confirms again that the surface disordered layer reconstructs the electronic structure of WO₃ by introducing oxygen vacancies. The formation of the mid-gap is closely related to the disordered surface layer on LT-WO₃, which contains W⁵⁺ ions and oxygen vacancies (Fig. 6). The self-doping of W⁵⁺ ions also modifies the electronic structure by creating a mid-gap and enhances the surface electron density, contributing to the upshift of the E_f.^16,48,49

Fig. 7 (a) VB-XPS and (b) Mott–Schottky plots of WO₃ and LT-WO₃. Inset in (a) is the enlarged area near the band edges indicated with a dotted box. (c) Schematic illustration of the energy levels of WO₃ and LT-WO₃ nanoplates.

Combining the results from UV-vis DRS, VB-XPS, and Mott–Schottky plots, we have constructed an energy level diagram qualitatively representing the changes in the band structure of LT-WO₃, and presented in Fig. 7c. The conduction band minimum (CBM) was estimated according to the equation E_CBM = E_VBM − E_g. The upshift of the E_f induced by the disordered layer on LT-WO₃ caused the VBM shift from 3.13 to 2.68 V. With an E_g of 2.74 eV, the CBM in LT-WO₃ is determined to be −0.06 V, which satisfies the thermodynamic requirement for H₂ generation.

To verify this reformed band structure of LT-WO₃, we have carried out the photocatalytic HER using a mixed solution of Na₂S and Na₂SO₃ as a sacrificial agent under one sun illumination. LT-WO₃ indeed shows a remarkable performance in the photocatalytic HER with an average rate of 94.2 μmol g⁻¹ h⁻¹ (Fig. 8). The H₂ generation rate was stable during the continuous 25 h reaction, showing no discernable degradation in the catalytic activity. The control experiment conducted in the dark generated no detectable H₂. To the best of our knowledge, this is the first report of photocatalytic H₂ evolution using WO₃ without any co-catalysts. The creation of a disordered surface layer on WO₃ is demonstrated to modify its electronic structure, thus endowing it with a new capability of photocatalytic H₂ generation. It also seems to affect the generation and lifetime of charge carriers. The steady state photoluminescence (PL) spectra (Fig. S2a†) recorded at room temperature show a much enhanced PL emission intensity, attributed to the higher concentration of photogenerated excitons. The time-resolved PL spectra (Fig. S2b†) indicate that the charge carriers in LT-WO₃ have more than 4 times longer lifetime (1.05 ± 0.05 ns) than those in WO₃ (0.25 ± 0.05 ns).

Fig. 8 Photocatalytic H₂ production using LT-WO₃ as a photocatalyst for 25 h reaction of five cycles.

In order to gain further insight on the disordered surface layer of LT-WO₃, the surface reactivity was studied by examining the electrocatalytic activity in H₂ production. Fig. 9a shows the polarization curves where the normalized current densities are plotted against applied potential without iR correction. LT-WO₃ shows a much lower overpotential (η) of −176 mV (vs. RHE, the same below) at a current density of 10 mA cm⁻² compared with that of pristine WO₃ (−667 mV). The corresponding Tafel slope of LT-WO₃ is 59 mV dec⁻¹ (Fig. 9b), lower than that of WO₃ (88 mV dec⁻¹), implying its higher activities towards H₂ production. These results indicate that the disordered layer on the WO₃ surface can significantly improve the surface reactivity towards H₂ production. The increased population of trapping sites for electron carriers (i.e., oxygen vacancies) is believed to promote the electron transfer, which effectively reduces the energy barrier for proton reduction, and activates the surface reactivity towards H₂ production.^26,27

Fig. 9 (a) Polarization curves and (b) Tafel plots of WO₃ and LT-WO₃, with Pt/C electrodes included for comparison.

Conclusions

In conclusion, the band structure of WO₃ has been successfully tuned by a simple surface defect engineering method, Li-EDA treatment. The formation of oxygen vacancies and self-doping of W⁵⁺ onto the surface of WO₃ through the generation of a disordered surface layer create a mid-gap level in the electronic structure and shift the CBM above the H₂ reduction potential. This is the first example of tuning the band structure of WO₃ to realize co-catalyst-free photocatalytic H₂ evolution. Electrochemical investigations also reveal that the oxygen vacancy-rich disordered layer can provide efficient electron trapping sites that promote the surface activity for the HER. Our work demonstrates that a disordered layer on a semiconductor surface can be utilized to effectively tune the band structure and extend its catalytic applications.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Innovation and Technology Commission, The Hong Kong Polytechnic University (Grant No. G-YBSZ and 1-BE0Y), and the National Natural Science Foundation of China (Grants No. 21701135). L. W. acknowledges the award of a postdoctoral fellowship by The Hong Kong Polytechnic University. K. Y. W. acknowledges the support by the Patrick S. C. Poon endowed professorship.

Notes and references

1. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520–7535.
2. J. H. Montoya, L. C. Seitz, P. Chakthranont, A. Vojvodic, T. F. Jaramillo and J. K. Nørskov, Nat. Mater., 2016, 16, 70.
3. Y. Wang, H. Suzuki, J. Xie, O. Tomita, D. J. Martin, M. Higashi, D. Kong, R. Abe and J. Tang, Chem. Rev., 2018, 118, 5201–5241.
4. S. Fukuzumi, Y.-M. Lee and W. Nam, Coord. Chem. Rev., 2018, 355, 54–73.
5. E. Ha, L. Y. S. Lee, J. Wang, F. Li, K.-Y. Wong and S. C. E. Tsang, Adv. Mater., 2014, 26, 3496–3500.
6. J. Zhang, W. Yao, C. Huang, P. Shi and Q. Xu, J. Mater. Chem. A, 2017, 5, 12513–12519.
7. W. Liu, E. Ha, L. Wang, L. Hu, L. Y. S. Lee and K.-Y. Wong, Adv. Mater. Interfaces, 2018, 5, 1800611.
8. I. Concina, Z. H. Ibupoto and A. Vomiero, Adv. Energy Mater., 2017, 7, 1700706.
9. G. Wang, Y. Yang, D. Han and Y. Li, Nano Today, 2017, 13, 23–39.
10. P. Kofstad, Phase Transitions, 1996, 58, 75–93.
11. J. Nowotny, M. A. Alim, T. Bak, M. A. Idris, M. Ionescu, K. Prince, M. Z. Sahdan, K. Sopian, M. A. Mat Teridi and W. Sigmund, Chem. Soc. Rev., 2015, 44, 8424–8442.
12. J. Jia, C. Qian, Y. Dong, Y. F. Li, H. Wang, M. Ghoussoub, K. T. Butler, A. Walsh and G. A. Ozin, Chem. Soc. Rev., 2017, 46, 4631–4644.
13. X. Pan, M.-Q. Yang, X. Fu, N. Zhang and Y.-J. Xu, Nanoscale, 2013, 5, 3601–3614.
14. J. Meng, Q. Lin, T. Chen, X. Wei, J. Li and Z. Zhang, Nanoscale, 2018, 10, 2908–2915.
15. Y. Yang, Y. Wang and S. Yin, Appl. Surf. Sci., 2017, 420, 399–406.
16. X. Chen, L. Liu, P. Y. Yu and S. S. Mao, Science, 2011, 331, 746–750.
17. Y. Liu, Y. Li, S. Yang, Y. Lin, J. Zuo, H. Liang and F. Peng, ChemSusChem, 2018, 11, 2766.
18. G. Yin, X. Huang, T. Chen, W. Zhao, Q. Bi, J. Xu, Y. Han and F. Huang, ACS Catal., 2018, 8, 1009–1017.
19. P. Yan, G. Liu, C. Ding, H. Han, J. Shi, Y. Gan and C. Li, ACS Appl. Mater. Interfaces, 2015, 7, 3791–3796.
20. G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang, J. Z. Zhang and Y. Li, Nano Lett., 2011, 11, 3026–3033.
21. S. W. Kwon, M. Ma, M. J. Jeong, K. Zhang, S. J. Kim and J. H. Park, Chem. Commun., 2016, 52, 13807–13810.
22. Y. Cho, S. Kim, B. Park, C. L. Lee, J. K. Kim, K. S. Lee, I. Y. Choi, J. K. Kim, K. Zhang, S. H. Oh and J. H. Park, Nano Lett., 2018, 18, 4257–4262.
23. S. Sun, T. Zhai, C. Liang, S. V. Savilov and H. Xia, Nano Energy, 2018, 45, 390–397.
24. T. Xia, P. Wallenmeyer, A. Anderson, J. Murowchick, L. Liu and X. Chen, RSC Adv., 2014, 4, 41654–41658.
25. K. J. Kyu, C. Yoonjun, J. M. Jin, L.-W. Ben, S. Dongguen, Y. Yeonjin, W. D. Hwan, Z. Xiaolin and P. J. Hyeok, ChemSusChem, 2018, 11, 933–940.
26. M. Ma, K. Zhang, P. Li, M. S. Jung, M. J. Jeong and J. H. Park, Angew. Chem., Int. Ed., 2016, 55, 11819–11823.
27. R. Zhang, F. Ning, S. Xu, L. Zhou, M. Shao and M. Wei, Electrochim. Acta, 2018, 274, 217–223.
28. H. Zheng, Y. Tachibana and K. Kalantar-zadeh, Langmuir, 2010, 26, 19148–19152.
29. S.-H. Lee, R. Deshpande, P. A. Parilla, K. M. Jones, B. To, A. H. Mahan and A. C. Dillon, Adv. Mater., 2006, 18, 763–766.
30. R. Xing, Y. Du, X. Zhao and X. Zhang, Sensors, 2017, 17, 710.
31. X. Shi, I. Y. Choi, K. Zhang, J. Kwon, D. Y. Kim, J. K. Lee, S. H. Oh, J. K. Kim and J. H. Park, Nat. Commun., 2014, 5, 4775.
32. H. Zhen-Feng, S. Jiajia, P. Lun, Z. Xiangwen, W. Li and Z. Ji-Jun, Adv. Mater., 2015, 27, 5309–5327.
33. C. A. Bignozzi, S. Caramori, V. Cristino, R. Argazzi, L. Meda and A. Tacca, Chem. Soc. Rev., 2013, 42, 2228–2246.
34. Y. H. Li, P. F. Liu, L. F. Pan, H. F. Wang, Z. Z. Yang, L. R. Zheng, P. Hu, H. J. Zhao, L. Gu and H. G. Yang, Nat. Commun., 2015, 6, 8064.
35. K. Zhang, L. Wang, J. K. Kim, M. Ma, G. Veerappan, C.-L. Lee, K.-j. Kong, H. Lee and J. H. Park, Energy Environ. Sci., 2016, 9, 499–503.
36. G. Wang, Y. Ling, H. Wang, X. Yang, C. Wang, J. Z. Zhang and Y. Li, Energy Environ. Sci., 2012, 5, 6180–6187.
37. J. Song, Z.-F. Huang, L. Pan, J.-J. Zou, X. Zhang and L. Wang, ACS Catal., 2015, 5, 6594–6599.
38. S. Bai, K. Zhang, L. Wang, J. Sun, R. Luo, D. Li and A. Chen, J. Mater. Chem. A, 2014, 2, 7927–7934.
39. S.-H. Lee, H. M. Cheong, J.-G. Zhang, A. Mascarenhas, D. K. Benson and S. K. Deb, Appl. Phys. Lett., 1999, 74, 242–244.
40. W. Chun-Kai, D. R. Sahu, W. Sheng-Chang, L. Chung-Kwei and H. Jow-Lay, J. Phys. D: Appl. Phys., 2012, 45, 225303.
41. R. S. Monson, in Advanced Organic Synthesis, ed. R. S. Monson, Academic Press, 1971, pp. 25–30,  DOI:10.1016/B978-0-12-504950-4.50006-X.
42. M. Ming, Z. Kan, L. Ping, J. M. Sun, J. M. Jin and P. J. Hyeok, Angew. Chem., Int. Ed., 2016, 55, 11819–11823.
43. Y. Sun, W. Wang, J. Qin, D. Zhao, B. Mao, Y. Xiao and M. Cao, Electrochim. Acta, 2016, 187, 329–339.
44. H. Cheng, M. Klapproth, A. Sagaltchik, S. Li and A. Thomas, J. Mater. Chem. A, 2018, 6, 2249–2256.
45. J. Liu, L. Han, N. An, L. Xing, H. Ma, L. Cheng, J. Yang and Q. Zhang, Appl. Catal., B, 2017, 202, 642–652.
46. J. Liu, J. Ke, D. Li, H. Sun, P. Liang, X. Duan, W. Tian, M. O. Tadé, S. Liu and S. Wang, ACS Appl. Mater. Interfaces, 2017, 9, 11678–11688.
47. L. Liao, Q. Zhang, Z. Su, Z. Zhao, Y. Wang, Y. Li, X. Lu, D. Wei, G. Feng, Q. Yu, X. Cai, J. Zhao, Z. Ren, H. Fang, F. Robles-Hernandez, S. Baldelli and J. Bao, Nat. Nanotechnol., 2013, 9, 69.
48. F. Zuo, L. Wang, T. Wu, Z. Zhang, D. Borchardt and P. Feng, J. Am. Chem. Soc., 2010, 132, 11856–11857.
49. W. Huang, J. Wang, L. Bian, C. Zhao, D. Liu, C. Guo, B. Yang and W. Cao, Phys. Chem. Chem. Phys., 2018, 20, 17268–17278.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8ta09446b

---