Colored TiO₂ hollow spheres for efficient water-splitting photocatalysts

Abstract

TiO₂ nanomaterials have attracted tremendous interest in photocatalytic hydrogen generation. Herein we report unique colored anatase TiO₂ hollow spheres consisting of crystalline-inner shell/amorphous-outer shell structured nanocrystallites, which exhibit remarkably enhanced photocatalytic performances for water splitting. The optimized TiO₂ spheres delivered a H₂ production rate of 0.182 mmol g⁻¹ h⁻¹, which was 2 times higher than that of pristine TiO₂. The enhanced activity can be ascribed to the generation of oxygen vacancies and Ti³⁺ sites.

---

Titanium dioxide (TiO₂) is a very promising photocatalyst for solar energy conversion including photovoltaics, photocatalysis and photoelectrochemical (PEC) water splitting for H₂ generation, due to its abundance, non-toxicity, and high activity and stability.¹ However, the application of TiO₂ is greatly limited by its wide bandgap (3.2 eV for anatase), which limits response to the UV range.² Enormous efforts, such as elemental doping,³ compositing with other semiconductors⁴ and sensitizing with dyes, have been devoted to extend TiO₂ absorption to the visible light range. Colored TiO₂ has demonstrated excellent visible light-response properties and efficient activity in photocatalytic and photoelectrochemical water splitting (PWS).^5–12 Several approaches including high pressure treatment, plasma assisted hydrogenation and high temperature Al reduction have been developed to prepare colored TiO₂.⁵ However, these methods involve high temperature³ (500 °C) or pressure and it may be difficult to control the crystallite size. Recently, Tan et al. developed a facile and mild solid-state chemical reduction method with NaBH₄ for the large-scale production of colored TiO₂ with remarkable enhancement in PWS.¹³

Besides the band structure, the PWS performance of TiO₂ also can be optimized by tailoring the morphology. To improve the PWS activity, colored TiO₂ including nanoparticles,^5–7 nanowires,¹⁰ nanotubes,¹¹ and nanosheets¹² have been reported. However, colored TiO₂ with a hollow sphere structure has not been studied yet. This hollow sphere composed of nanocrystallites with three-dimensional structure can provide much more reactive sites for the photocatalytic reactions.¹⁴ The porous hollow sphere would be beneficial for the photocatalytic reactions that occur on both the internal or outer surface of TiO₂ sphere shell. Furthermore, the hollow structure allows the multiple reflection in the internal shell resulting in the improvement on the utilization of light irradiation.^14,15 Thus, it can be expected that the PWS performance of colored TiO₂ can be boosted by the hollow sphere structure. In this study, colored TiO₂ spheres with hollow feature were fabricated via a NaBH₄ reduction approach at a mild temperature of 300 °C, which exhibited greatly improved PWS performance. Compared with conventional routes, this method is characterized with relatively low reaction temperature that can be readily used to control the color and activity of TiO₂.

TiO₂ hollow spheres were synthesized via a solvothermal method based on the Ostwald ripening (for experimental procedure, see ESI†).¹⁶ As prepared anatase hollow spheres (labeled as AHS) were further reduced by mixing NaBH₄ at 300 °C for 20 min, 40 min and 60 min in N₂ atmosphere, and the corresponding samples were named as AHS20, AHS40 and AHS60. Typical scanning electron microscopy (SEM) images (Fig. S1a, ESI†) reveal their uniform sphere structure with diameters of 400–500 nm. These spheres are comprised of a large quantity of TiO₂ nanocrystallites, resulting in a rough surface. A few broken TiO₂ spheres with open edge proved the hollow structure (Fig. S1c†). The shell thickness is ca. 50 nm. The morphology and size of the spheres were not impacted by the heat treatment with NaBH₄ reduction (Fig. S1b–d, ESI†). TEM images further prove the hollow sphere microstructure (Fig. 1a and c) with tiny observable pores on the surface, which is resulted by the assembling of TiO₂ nanoparticles. High resolution transmission electron microscopy (HRTEM) image of pristine TiO₂ sphere in Fig. 1b confirms that it exhibits highly crystalline and well-resolved lattice features. On the other hand, after reduction treatment, typical HRTEM image of AHS40 in Fig. 1d reveals a core–shell structure with a crystallized core and a highly disordered surface with thickness of ca. 2.5 nm. Interestingly, the core of nanocrystal is still well crystallized. The lattice distances of both AHS and AHS40 are 0.35 nm corresponding to the (101) lattice plane of anatase TiO₂. In fact, the activity of TiO₂ is greatly dependent on the crystal phase and anatase is superior to the rutile in water splitting reactions due to its conduction band bottom is more negative than the latter.¹⁷ The color change can be found in optical images in Fig. 1e. As prepared AHS is white, which turns to be light brown and dark brown with the increasing of reaction time. The BET surface areas of pristine and colored TiO₂ hollow spheres are 56–58 m² g⁻¹ with pore size in the range of 3–10 nm (Fig. S2, ESI†).

Fig. 1 TEM and HRTEM images of (a and b) AHS, (c and d) AHS40 sample and (e) optical photographs of pristine and colored anatase powders.

Raman spectroscopy was used to investigate the structural evolution of pristine and reduced TiO₂ hollow spheres. As shown in Fig. 2a and b, these samples exhibit similar peaks and six typical Raman active modes (3E_g + 2B_1g + A_1g) of anatase phase can be detected.⁸ With the increasing of reduction time, the strongest E_g mode shows a clear blue shift from 144 cm⁻¹ to 149 cm⁻¹ (Fig. 2b). Meanwhile, the full width at half maximum intensity (FWHM) of the E_g peaks for reduced TiO₂ samples were broadened from 12.5 cm⁻¹ (AHS) to 14.5 cm⁻¹ (AHS60). According to the literature,^8,13 the blue shift and the broadening of the E_g peak may due to the decreasing of TiO₂ size or the shortening of the phonon correlation length. As no noticeable change of TiO₂ size has been observed in SEM and TEM images, the blue shift and broadening of the E_g mode at 144 cm⁻¹ can mainly be attributed to the formation of oxygen vacancies in the reduced samples. The X-ray diffraction (XRD) patterns in Fig. S3† reveal that the peaks of four samples correspond to the standard anatase TiO₂ (JCPDS no. 21-1272). No additional peak was introduced after the reduction treatment. However, the enlarged pattern (Fig. S2b†) indicates that, with the increasing of the treatment time, the (101) peaks shift to high angle, which might be caused by oxygen vacancies-induced lattice strains.¹³ The presence of oxygen vacancies has been confirmed by electron paramagnetic resonance (EPR) results. As shown in Fig. 2c, a very strong EPR signal at g = 1.996 were introduced after the reduction treatment and the intensity is greatly improved with treatment time. This signal can be assigned to Ti³⁺ (3d₁),^18,19 which are produced by the trapping of electrons at defective sites in TiO₂ due to the generation of oxygen vacancies. The increase of the EPR signal suggests the oxygen vacancy concentration can be controlled by the reducing time. The surface composition and the chemical states of the elements of the samples were studied by X-ray photoelectron spectroscopy (XPS). The peaks of Ti 2p_3/2 and Ti 2p_1/2 locate at 458.4 and 464.2 eV, respectively (Fig. 2d).²⁰ Compared with AHS, the peaks of the reduced samples show a distinct shift to the lower binding energy, indicating the formation of Ti³⁺ in the reduced samples.¹³ Except for AHS, the peaks of Ti 2p_3/2 for all of the reduced samples can be fitted into two Gaussian peaks at 458.3 and 457.9 eV, which could be assigned to Ti⁴⁺ and Ti³⁺, respectively (Fig. S3, ESI†).²¹ Moreover, with the increasing of the reduction time, the peak intensity at 457.9 eV becomes stronger, indicating the increasing of Ti³⁺ concentration.¹³ This is consistent with the Raman and EPR results. As for the valence-band XPS spectra shown in Fig. S4e,† there is an obviously band shifts were observed, indicating the impurity level from the oxygen vacancies formed on the top of the valence band. The band gap narrowing induced by the impurity level would enlarge the light absorption range and improve the activity consequently, as evidenced in UV-Vis absorption spectra in Fig. S5.† Compared to AHS sample, the light absorption of AHS40 sample was enhanced which will be beneficial for the improvement of photocatalytic activity.

Fig. 2 (a) Raman spectra of AHS and reduced samples, (b) enlargement of most intense E_g peak of all samples, (c) EPR spectra of AHS and reduced samples measured at 100 K, (d) Ti 2p XPS spectra.

Time courses of H₂ production obtained over TiO₂ spheres and the reduced samples were evaluated and a steadily increasing of H₂ amount with prolonged irradiation time was observed for all tests (Fig. S6, ESI†). The H₂ generation of AHS after the photocatalytic reaction for 4 h was about 0.38 mmol g⁻¹. As for the reduced TiO₂, an increased activity is observed for AHS20 and AHS40, while an obvious decrease is detected for AHS60, which can be attributed to the excessive oxygen vacancies served as recombination centers. A much higher photocatalytic activity with a H₂ production rate of 0.182 mmol g⁻¹ h⁻¹ is observed on AHS40, which is 2 times higher than that of AHS (0.094 mmol g⁻¹ h⁻¹) (Fig. 3a). The stability of AHS40 was tested by a cycling experiment. Interestingly, AHS40 did not exhibit any reduction of H₂ production activity after irradiation for 20 h, revealing the excellent stability of the sample for photocatalytic H₂ production (Fig. 3b). These results verified that the crystallized core with the highly disordered surface of TiO₂ can greatly enhance the photocatalytic activity while maintain outstanding cycling stability. In addition, the performance of colored TiO₂ nanowires was also tested for comparison. However, a H₂ production rate of 0.06 mmol g⁻¹ h⁻¹ was obtained, which is quite lower than that of colored AHS40 TiO₂ hollow spheres.

Fig. 3 (a) Photocatalytic water splitting of AHS and reduced TiO₂ hollow spheres under simulated sunlight (a 300 W xenon lamp without filter). (b) Cycling measurements of H₂ generation on AHS40 under simulated solar light.

The photoelectrochemical properties were characterized and the results show that, as the voltage sweeping towards a more positive bias, all sample electrodes had a steady increase in photocurrent density (Fig. 4a). Notably, the AHS40 electrode shows a significant increase in the photocurrent density that was about 1.7 times higher than that of AHS. Moreover, compared with AHS (an onset of photocurrent at −0.51 V_SCE), AHS40 exhibits a shift in the onset of photocurrent to −0.63 V_SCE. The higher photocurrent density and the lower onset potential of AHS40 indicates a more efficient charge separation and transport compared to that of AHS.⁸ Due to the excessive oxygen vacancies, the photocurrent density of AHS60 in PEC is lower than that of AHS leading to the photo-generated charges recombination and the lowered activity as observed in Fig. 3a. Further characterization of the behavior for the photo-generated charges was carried out on the transient photocurrent of AHS and AHS40 samples that were measured during repeated ON/OFF illumination cycles at 0.3 V_SCE (Fig. 4b). Both samples exhibit prompt and reproducible photocurrent responses on each illumination. Obviously, the transient photocurrent density of AHS40 is about 1.7 times higher than that of AHS. This result confirms that the higher amount of photoinduced carriers due to the more absorption from the colored TiO₂ and more efficient electron–hole separation takes place in AHS40. The increasing of carriers was also observed in the Mott–Schottky plots (Fig. S7, ESI†), which displays positive slopes characteristic of n-type semiconductors.⁸ Compared with AHS, AHS40 shows a smaller slope of the Mott–Schottky plot, which implies higher charge carrier density in AHS40.

Fig. 4 (a) Linear-sweep voltammograms collected under simulated solar light irradiation using a three-electrode setup in 0.1 M Na₂SO₄ electrolyte (pH ≈ 7) with a scanning rate of 10 mV s⁻¹. The prepared samples were used as a working electrode, a Pt plate as the counter, and a saturated calomel as a reference electrode. (b) Transient photocurrent responses of AHS and AHS40 samples at 0.3 V_SCE.

AHS40 exhibits excellent photocatalytic activity due to the formation of oxygen vacancy and reduced Ti sites such as Ti³⁺ after the reductive reaction. The oxygen vacancies are reported as the electron donor that can contribute to the enhanced donor density in TiO₂, accompanied by the shift of the Fermi level towards the conduction band of TiO₂.^13,22,23 This shift can facilitate the photogenerated electron–hole separation, enhance charge transportation, and effectively retard the photogenerated electron–hole recombination. However, high concentration of oxygen vacancies can also act as charge recombination centers so that free charges were reduced and thus low photocatalytic activity was observed for AHS60,¹³ as evidenced by the decrease of H₂ generation rate (Fig. 3a) and photocurrent density (Fig. 4b).

In summary, we reported for the first time that oxygen-deficient TiO₂ hollow spheres exhibited a stable and excellent photocatalytic performance in solar driven water splitting. The treated anatase hollow spheres are composed of many TiO₂ nanocrystallites with a crystalline core and a disordered shell. Both H₂ production rate and the photocurrent density of the optimized AHS40 sample are significantly higher that of pristine TiO₂. The generation of oxygen vacancy and reduced Ti sites are responsible to the enhanced photocatalytic activity. Our study demonstrates that defective anatase hollow spheres can be highly-efficient photocatalysts for solar energy conversion.

Acknowledgements

This work was supported by the Natural Science Foundation of Heilongjiang Province of China (A2015005), and National Natural Science Foundation of China (21473066, 21373071).

Notes and references

1. (a) M. Graetzel, R. A. J. Janssen, D. B. Mitzi and E. H. Sargent, Nature, 2012, 488, 304–312; (b) A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663; (c) X. B. Chen, S. H. Shen, L. J. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570; (d) K. Maeda and K. Domen, J. Phys. Chem. Lett., 2010, 1, 2655–2661.
2. (a) A. B. Murphy, P. R. F. Barnes, L. K. Randeniya, I. C. Plumb, I. E. Grey, M. D. Horne and J. A. Glasscock, Int. J. Hydrogen Energy, 2006, 31, 1999–2017; (b) R. Su, R. Tiruvalam, Q. He, N. Dimitratos, L. Kesavan, C. Hammond, J. A. Lopez-Sanchez, R. Bechstein, C. J. Kiely, G. J. Hutchings and F. Besenbacher, ACS Nano, 2012, 6, 6284–6292; (c) D. Y. C. Leung, X. L. Fu, C. Wang, M. Ni, M. K. H. Leung, X. X. Wang and X. Z. Fu, ChemSusChem, 2010, 3, 681–694.
3. (a) W. Choi, A. Termin and M. R. Hoffmann, J. Phys. Chem., 1994, 98, 13669; (b) M. Anpo and M. Takeuchi, J. Catal., 2003, 216, 505; (c) R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science, 2001, 293, 269; (d) X. Chen and C. Burda, J. Am. Chem. Soc., 2008, 130, 5018.
4. (a) D. R. Baker and P. V. Kamat, Adv. Funct. Mater., 2009, 19, 805; (b) H. Wang, G. M. Wang, Y. C. Ling, M. Lepert, C. C. Wang, J. Z. Zhang and Y. Li, Nanoscale, 2012, 4, 1463–1466.
5. X. B. Chen, L. Liu and F. Q. Huang, Chem. Soc. Rev., 2015, 44, 1861–1885.
6. X. B. Chen, L. Liu, P. Y. Yu and S. S. Mao, Science, 2011, 331, 746–749.
7. C. H. Sun, Y. Jia, X. H. Yang, H. G. Yang, X. D. Yao, G. Q. Lu, A. Selloni and S. C. Smith, J. Phys. Chem. C, 2011, 115, 25590–25594.
8. Z. Wang, C. Y. Yang, T. Q. Lin, H. Yin, P. Chen, D. Y. Wan, F. F. Xu, F. Q. Huang, J. H. Lin, X. M. Xie and M. H. Jiang, Energy Environ. Sci., 2013, 6, 3007–3014.
9. C. Zhen, L. Z. Wang, L. Liu, G. Liu, G. Qing Lu and H. M. Cheng, Chem. Commun., 2013, 49, 6191–6193.
10. G. M. Wang, H. Y. Wang, Y. C. Ling, Y. C. Tang, X. Y. Yang, R. C. Fitzmorris, C. C. Wang, J. Z. Zhang and Y. Li, Nano Lett., 2011, 11, 3026–3033.
11. W. D. Zhu, C. W. Wang, J. B. Chen, D. S. Li, F. Zhou and H. L. Zhang, Nanotechnology, 2012, 23, 455204.
12. W. Wei, Y. N. Yaru, C. H. Lu and Z. G. Xu, RSC Adv., 2012, 2, 8286–8288.
13. H. Q. Tan, Z. Zhao, M. Niu, C. Y. Mao, D. P. Cao, D. J. Cheng, P. Y. Feng and Z. C. Sun, Nanoscale, 2014, 6, 10216–10223.
14. H. X. Li, Z. F. Bian, J. Zhu, D. Zhang, G. S. Li, Y. N. Huo, H. Li and Y. F. Lu, J. Am. Chem. Soc., 2007, 129, 8406–8407.
15. X. Lai, J. E. Halpert and D. Wang, Energy Environ. Sci., 2012, 5, 5604–5618.
16. S. Q. Shang, X. Jiao and D. R. Chen, ACS Appl. Mater. Interfaces, 2012, 4, 860–865.
17. N. G. Park, J. Van de Lagemaat and A. J. Frank, J. Phys. Chem. B, 2000, 104, 8989–8994.
18. Z. Lin, A. Orlov, R. M. Lambert and M. C. Payne, J. Phys. Chem. B, 2005, 109, 20948–20952.
19. T. M. Salama, H. Hattori and H. Kita, J. Chem. Soc., Faraday Trans., 1993, 89, 2067–2073.
20. A. Naldoni, M. Allieta, S. Santangelo, M. Marelli, F. Fabbri, S. Cappelli, C. L. Bianchi, R. Psaro and V. D. Santo, J. Am. Chem. Soc., 2012, 134, 7600–7603.
21. E. McCafferty and J. P. Wightman, Appl. Surf. Sci., 1999, 143, 92–100.
22. A. Janotti, J. B. Varley, P. Rinke, N. Umezawa, G. Kresse and C. G. Van de Walle, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 085212.
23. M. Ye, J. J. Gong, Y. K. Lai, C. J. Lin and Z. Q. Lin, J. Am. Chem. Soc., 2012, 134, 15720–15723.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedures, SEM images (Fig. S1), nitrogen isotherm profiles (Fig. S2), XRD patterns (Fig. S3), XPS results (Fig. S4), UV-Vis spectra (Fig. S5), time courses of H₂ production (Fig. S6) and Mott–Schottky plots (Fig. S7). See DOI: 10.1039/c6ra23151a

---