TiO₂-pillared clays with well-ordered porous structure and excellent photocatalytic activity

Abstract

TiO₂-pillared clays with well-ordered porous structures are successfully prepared via incorporating TiO₂ nanosol particles into the clays, where empty octahedral sites are partially modified with divalent metal ions such as Mg²⁺ and Fe²⁺. The prepared TiO₂-pillared mica exhibits excellent photocatalytic activity, which can be controlled by tuning the optical transparency of clay support rather than the specific surface area of the hybrid catalysts.

---

Layered silicates including cationic clays have been widely employed as adsorbents,^1,2 fillers,³ catalysts,^4,5 catalyst supports,^6,7 drug delivery carriers,⁸ etc. To introduce better physicochemical functions on the layered silicates, organic or inorganic functional molecules were incorporated into the interlayer space via an intercalation and/or grafting reaction. In the last decades, however, several studies have focused on nanoporous clays, namely pillared clays, whose pore sizes are larger than those of the typical zeolites. However, most of the pillared clays appearing in the literature so far have been prepared by intercalating nanosized metal oxides, such as SiO₂, TiO₂, Al₂O₃, ZrO₂, SiO₂–TiO₂, SiO₂–CoO, SiO₂–Fe₂O_3, SiO₂–Cr₂O₃, etc., into the clay layers followed by the calcination process.^8–16 In such a manner, pillared clays have been eventually realized with a large specific surface area, a high thermal stability and enhanced catalytic activity.

Among various pillared clays, TiO₂- and Fe₂O₃-pillared clays have been applied to the photocatalytic reactions involving photodecomposition of environmental contaminants such as organic dyes, halogenated hydrocarbons, pesticides, and the photolysis of water.^7,13,14 The TiO₂- and Fe₂O₃-pillared clays exhibit interesting two-dimensional (2D) heterostructure, which makes them unique for photocatalysis because the photocatalytic property of semiconductors,¹⁵ such as bulk titania or ferric oxide, is mainly dictated by their electronic structure (band gap energy), particles size, and specific surface area. As the 2D clay lattice structure can help to avoid the agglomeration of the nanoparticles, the pillared clays with high specific surface area and porosity, and enhanced catalytic property can be easily realized.

In the heterogeneous catalysis, a well-ordered pillared structure with regular pore size such as the Al₂O₃-pillared clay is considered to be an attractive catalyst due to size, shape and a selective catalytic reaction similar to that of the zeolite catalyst. However, the reports on the preparation of well-ordered porous TiO₂-pillared clays are quite limited,¹⁰ even though several studies have reported the TiO₂-pillared clays.¹³ In pillaring TiO₂ nanoparticles into the clay layers, the layer charge density of clay support as well as the regular size of positively charged TiO₂ nanoparticles is an important factor as it controls the intercalation of guest into the clay layers. In this study, we intend to demonstrate a synthetic route to well-ordered TiO₂-pillared clays with different transparency and porosity by controlling the layer charge density of the host clay, and finally to correlate them with photocatalytic activities.

The layer charge density of optically transparent Na-mica could be modified by partially incorporating Mg²⁺ or Fe²⁺ ions into the vacant octahedral Mg²⁺ defect sites through the ion-exchange reaction (Scheme 1) and the subsequent calcination (hereafter denoted as MgM and FeM, respectively). The cation exchange capacity (CEC) of Na-mica (100 meq g⁻¹) was reduced to 75 and 67 meq g⁻¹ for MgM and FeM, respectively, due to the decrease in layer charge density caused by the fixation of Mg²⁺ or Fe²⁺ ions in vacant octahedral defect sites (Hofmann–Klemen effect) (Scheme 1 and Fig. S1, ESI†).¹⁶

Scheme 1 Synthesis of well-ordered TiO₂-pillared clay.

Nanoporous TiO₂-pillared clays could be prepared by incorporating TiO₂ nanoparticles into the metal-ion modified mica (MgM and FeM) via the ion-exchange reaction and subsequent calcination. The resulting TiO₂-pillared mica, MgM and FeM calcined at 500 °C are denoted as TM-500, TMgM-500 and TFeM-500, respectively. According to the XRD patterns of TiO₂-pillared clays, as shown in Fig. 1(A), the basal spacings increased to 24.5 Å, 24.3 Å and 24.7 Å for TM-500, TMgM-500 and TFeM-500, respectively, compared to that of the pristine Na-mica (9.6 Å), MgM and FeM. Such a large basal expansion along the c-axis provides a good evidence of the intercalation of TiO₂ nanoparticles into clay layers to form porous TiO₂-pillared clays. The size of the pores was estimated to be ∼15 Å, which was obtained by subtracting the basal spacing of the pristine (∼9.6 Å) from those of the TiO₂-pillared clays. Moreover, no impurity peaks corresponding to the anatase-, rutile- or brookite-structure of TiO₂ were observed (Fig. S2, ESI†). For TMgM-500 and TFeM-500, well-developed (00l) diffraction peaks (at least second order) appeared, which are attributed to the regular stacking of TiO₂ nanoparticles between the silicate layers. These results indicate that the lower layer charge density of the pristine is critical to obtain the TiO₂-pillared clay with a better crystallinity. As can be seen in Fig. 1(B), the periodically ordered layered structure along the crystallographic c-axis was reconfirmed by the cross-sectional TEM images and Fourier filtered images of pristine clay (MgM) and TiO₂-pillared clay (TMgM-500). The basal spacings measured between the two lattice fringes for MgM(a) and TMgM-500(b) were 10 Å and 26 Å, respectively, which are in excellent agreement with those determined from XRD measurements.

Fig. 1 (A) XRD patterns of (a) Na-mica, (b) TM-500, (c) TMgM-500 and (d) TFeM-500 (♦ corresponds to non-swellable mica), and (B) TEM images and their corresponding FFT diffractograms (insets) of the cross-sectioned (a) MgM and (b) TMgM-500.

The diffuse reflectance UV-vis spectra of Na-mica modified with Mg²⁺ and Fe²⁺ ions, MgM and FeM, are represented in Fig. 2(A) along with those of the pristines, Na-mica and Kunipia-G (natural montmorillonite containing Fe ions). As can be clearly seen, Na-mica and MgM can hardly absorb the light in the wavelength range of 300 nm–350 nm due to their optical transparency; however, Kunipia-G and FeM absorb the light in the same range because the transition metal ion like Fe²⁺, is stabilized in the octahedral site of the clay lattice, and eventually absorb the UV and visible lights. Therefore, Na-mica and MgM could be considered as the good photocatalyst supports for immobilizing photo-active species like TiO₂ due to their good transparency in the range of UV and visible light. As shown in Fig. 2(B), the absorption edges of TM-500, TMgM-500 and TFeM-500 are blue-shifted from that of the anatase-type TiO₂ owing to the quantum size effect of TiO₂ nano-pillars stabilized in the interlayer space of 2D-clay lattice.¹⁷ The bandgap energy calculated from the absorption edge was determined to be 3.46 eV for TM-500 and TMgM-500, and 3.38 eV for TFeM-500, respectively, which is higher than that for anatase-type TiO₂ (3.28 eV). All these results are in good agreement with the XRD analyses because the size of TiO₂ nanopillars in the 2D-clay lattice was determined to be even <2 nm. The small difference in bandgap energy among TiO₂-pillared clays is mainly due to the slight size difference (∼1 Å) of TiO₂ pillars in the 2D-clay lattice.

Fig. 2 Diffuse reflectance UV-vis spectra for (A) various layered silicates; Kunipia-G, Na-mica, MgM and FeM, and (B) anatase TiO₂, TM-500, TMgM-500 and TFeM-500.

From the nitrogen adsorption–desorption isotherm analysis, as shown in Fig. 3 and Table S2 (ESI†), it is clearly seen that the amount of nitrogen adsorbed at the lower relative pressure is quite negligible at 77 K. However, after TiO₂ pillaring, the absorption amount of nitrogen for Na-mica at the lower relative pressure, which is directly related to the specific surface area of the samples, is significantly improved. This is because the incorporation of the TiO₂ nanoparticles into the silicate layers creates the nanoporosity that is responsible for the enhancement in the amount of adsorption. The isotherms of TiO₂-pillared clays, TM-500, TMgM-500 and TFeM-500, are of type I according to the Brunauer, Deming, Deming, and Teller (BDDT) classification,¹⁸ and free from a hysteresis loop. It is interesting to note that the adsorption and desorption behaviors are quite similar to that of zeolite (a typical microporous material), indicating that all the samples are microporous in nature.^18,19 This result is different from other reports, in which the isotherm curves of TiO₂ pillared clay was classified as Type IV (typical isotherm of mesoporous materials),^10,13 indicating that TiO₂ nanoparticles with homogeneous size (<2 nm) are regularly stacked in-between the mica layers. Based on the t-plot analyses with the statistical t thickness from the Harkins–Jura equation,^18,19 the micropore volume for TM-500, TMgM-500 and TFeM-500 was determined as 0.10 mL g⁻¹, 0.15 mL g⁻¹ and 0.15 mL g⁻¹, respectively. Each value corresponds to more than ∼50% of each total pore volume. The Brunauer–Emmett–Teller (BET) specific surface area and total pore volume of all the TiO₂-pillared clays are in the range of 242–347 m² g⁻¹ and 0.21–0.29 mL g⁻¹, respectively. These values are significantly higher as compared to those of Na-mica (9 m² g⁻¹ and 0.03 mL g⁻¹), confirming the formation of nanoporosity in the samples. In particular, it is interesting to note that the specific surface area and the pore volume of TMgM-500 (335 m² g⁻¹ & 0.29 mL g⁻¹) and TFeM-500 (347 m² g⁻¹ & 0.27 mL g⁻¹) are much higher than those of TM-500 (242 m² g⁻¹ & 0.21 mL g⁻¹), which is mainly due to the formation of highly ordered TiO₂-pillars with nanoporous structure. Furthermore, the Horbáth-Kawazoe pore size distribution²⁰ of all the samples shows a broad peak at ∼15 Å, which is also well consistent with the gallery height of the TiO₂-pillared clays obtained from the XRD results. Moreover, a sharp peak at ∼8 Å for well-ordered TiO₂ pillared clays, such as TMgM-500 and TFeM-500, is observed, which is attributed to the inter-particle distance among TiO₂ nanoparticles regularly incorporated between 2D-clay lattices.

Fig. 3 N₂ adsorption–desorption isotherms of various TiO₂-pillared clays together with pristine Na-mica (square ■); TM-500 (triangle ▲), TMgM-500 (inverted triangle ▼), TFeM-500 (circle ●). Solid and open symbols indicate adsorption and desorption curves, respectively. TMgM-500, was plotted by y-axis off-set of 25. (Inset: Horváth-Kawazoe pore size distribution of TM-500, TMgM-500 and TFeM-500).

To elucidate the local symmetry of titanium atoms in the samples, the X-ray absorption near-edge structure (XANES) analysis was also performed for all the TiO₂-pillared clays.²¹ As shown in Fig. 4, the overall feature of Ti K-edge XANES spectra for TM-500, TMgM-500 and TFeM-500 was similar to that for anatase-type TiO₂ rather than that for rutile-type one. The weak peaks (P₁, P₂ and P₃) corresponding to the dipole-forbidden transitions from 1s to 3d-4p hybrid orbital, t_2g and e_g, respectively, were observed for all the TiO₂-pillared samples and the reference anatase-TiO₂ and rutile-type one, indicating that titanium ions in the TiO₂ pillars were in pseudo-octahedral symmetry.^12,22–24 The data of TiO₂-pillared clays were also analyzed based on the previous study by Farges et al.²² on the position and intensity of the P₂ peak. It was found that ∼90% of Ti⁴⁺ ions in the TiO₂ pillar were high in symmetry, six coordinated, but ∼10% of them were stabilized in a lowered symmetric site, five coordinated site. The coordinatively unsaturated symmetry for Ti⁴⁺ ions definitely originated from the surface contribution of nano-sized TiO₂ pillars, where the oxygen defects are abundant. In the main edge region, three spectral features denoted as A, B and C were observed, where the lower energy peak A was related to the transition with the shakedown process, whereas the higher energy peaks, B and C, were assigned as the transition from core 1 s to out-of-plane 4p_z orbital and to in-plane 4p_x,y orbitals.²⁴ The peak positions and intensities for all the TiO₂-pillared clays were very similar to those for the anatase-type bulk TiO₂, whereas the overall spectral feature and the peak C intensity, in particular, for rutile-type bulk TiO₂, were completely different from others owing to the fact that the local symmetry around Ti atoms in rutile TiO₂ is tetragonally less distorted than that in anatase and TiO₂-pillared clays.

Fig. 4 Normalized Ti K-edge XANES spectra for (a) rutile-type TiO₂, (b) anatase-type TiO₂, (c) TM-500, (d) TMgM-500 and (d) TFeM-500.

In addition to the XANES study, we conducted an extended X-ray absorption fine structure (EXAFS) analysis for the pillared samples because the local structural symmetry of TiO₂ pillars is closely related to the photocatalytic activity. As shown in Fig. 5, the Fourier transforms (FTs) of the Ti K-edge k³-weighted EXAFS spectra of the pillared clays were investigated along with that of the bulk anatase TiO₂ as a reference material. In the bulk anatase, the FT peaks at 1.6, 2.7, and 3.5 Å are attributed to the contribution of (Ti–O), (Ti–Ti_edge), and (Ti–Ti_corner) bonds, respectively, and the structural parameters are listed in Table S3 (ESI†). Even though the bond distances of TiO₂ pillared clays are determined to be virtually identical to those of the bulk anatase TiO₂, the coordination numbers of the edge- and corner-shared (Ti–Ti) shells for the formers are smaller than those for the latter, which are in good agreement with the XANES results. These results indicate the structural disorder of oxygen atoms around Ti atoms due to the formation of nanosized TiO₂ pillars in the interlayer space of clays, as evidenced by the XRD analysis, nitrogen adsorption measurements, and XANES results. In contrast, the FT peaks beyond 4 Å, attributed to the multiple scattering effect from the surrounding atoms, are also reduced in the pillared clays, highlighting the damping of the EXAFS signal because of the nanosized flexible structure of TiO₂ pillars.

Fig. 5 k³-weighted Ti K-edge EXAFS spectra for (a) anatase-type TiO₂, (b) TM-500, (c) TMgM-500 and (d) TFeM-500. Their best-fitted spectra (open circles) are compared with experimental spectra (solid lines). The Fourier fitting range is R = 1.0–3.85 Å.

To evaluate the photocatalytic activity of the well-ordered TiO₂-pillared clays with nanoporous structure, the time-dependent concentration variation of methyl orange (MO) was measured in an aqueous solution suspended with TiO₂-pillared clays under UV irradiation (λ > 290 nm) because MO is one of the model compounds largely used for the evaluation of photodecomposition activity.⁷ Fig. 6 shows the fractional concentration of MO in the TiO₂-pillared clay suspensions depending on the UV irradiation time compared to that in TiO₂-pillared Kunipia-G (denoted as TK-500), in which the photodecomposition kinetic constants were determined by fitting the curves based on the pseudo-first order kinetic equation.^7,25 Under this photocatalytic reaction condition, the concentration of MO with the TiO_2-pillared clay catalysts decreased drastically upon increasing the UV light irradiation time, whereas no photocatalytic degradation of MO was observed without the photocatalyst within 120 min. The order of the photodecomposition activity is as follows; TMgM-500 > TM-500 > TFeM-500 > TK-500. In general, the catalysts with larger specific surface area show higher activity, confirming the importance of nanoporosity in the catalyst. Comparing TMgM-500 and TFeM-500 with the similar specific surface area (∼340 m² g⁻¹) and TiO₂ contents (∼32 wt%), the photocatalytic activity of TMgM is higher than that of TFeM-500.

Fig. 6 Time-dependent curves of the concentration of methyl orange (MO) upon photocatalytic degradation reaction, under UV irradiation (λ > 290 nm, 0.1 W cm⁻²), with and without TiO₂-pillared clays; TK-500, TM-500, TMgM-500, TFeM-500 and blank (without catalyst). Initial concentration of MO was 5 × 10⁻⁵ mol L⁻¹ and 30 mg of catalyst was dispersed in 30 mL MO solution.

It is important to note that the photocatalytic activity of TK-500 (with less transparent clay support; Kunipia-G) is found to be considerably lower than that of TM-500 (with transparent clay support; Na-mica), even though TK-500 has significantly higher TiO₂ content (56.9 wt%) and specific surface area (314 m² g⁻¹) than those of TM-500 (34.0 wt% and 242 m² g⁻¹). We believe that such unusual photocatalytic property of the TiO₂-pillared clays with different textural parameters is due to the difference in the optical transparency of the host clay layers. According to the UV-vis spectroscopic analysis (Fig. 2(A)), the order of the absorption value (Kubelka Munk (K.M.) value) at the absorption edge (∼350 nm) is Kuinipia-G (0.55) > FeM (0.23) > MgM (0.01) and Na-mica (0.00), indicating that some part of the irradiated photoenergy in the TiO₂-pillared clay is shielded in the case of less transparent clay such as Kunipia-G and FeM. Therefore, it is quite clear why the photodegradation activity of MO was higher in the presence of TiO₂-pillared in transparent clays; TMgM-500 (k = 0.025 min⁻¹) > TM-500 (k = 0.023 min⁻¹) > TFeM-500 (k = 0.018 min⁻¹) > TK-500 (k = 0.011 min⁻¹). This result is quite similar to the order of optical transparency of the pristine clay support. If the TiO₂ photocatalysts are supported on clays with similar optical transparency such as Na-mica and MgM, the photocatalytic activity of TM-500 should be equivalent to or slightly higher than that of TMgM-500. However, the specific surface area of TMgM-500 is quite higher than that of TM-500, resulting in slightly higher photocatalytic activity, because of the cancelling effect (optical transparency ⇆ specific surface area). All these results reflect that the optical transparency as well as the specific surface area of the catalyst support (clay) is an important factor that controls the photocatalytic performance of the photocatalytic immobilized clays.

Conclusions

The layer charge density and optical transparency of sodium fluoromica were modified by partially incorporating Mg²⁺ and Fe²⁺ cations into the empty octahedral sites of Na-mica using the HK effect. Heterogeneous microporous TiO₂ photocatalysts with well-ordered nanoporous structure were also successfully synthesized by intercalating TiO₂ nanoparticles into the clay layers via ion-exchange reaction and subsequent calcination (Scheme S1, ESI†). The prepared TiO₂-pillared clays were highly porous and exhibited the specific surface areas of 242–347 m² g⁻¹ and large pore volume of 0.21–0.29 mL g⁻¹, which were significantly higher as compared to those of the pristine clay (∼10 m² g⁻¹), in which more than ∼50% of total pore volume is composed of micropores. From the XANES and EXAFS analyses at the Ti K-edge, the local symmetry of Ti in TiO₂-pillared clays was fairly similar with that in anatase and was found to be slightly lower because of the formation of oxygen deficient TiO₂ nano-pillars between the clay sheets. The order of photocatalytic activity of TiO₂-pillared clays on photodegradation of methyl orange under UV irradiation was determined to be TMgM-500 > TM-500 > TFeM-500 > TK-500. It was found that the photocatalytic activity of TiO₂-pillared clays is significantly affected by the optical transparency of host clay lattice, as well as the specific surface area.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant, funded by the Korea government (MSIP) (2005-0049412 and 2013R1A1A2062239) and by the Ewha Womans University Research Grant of 2013. One of the authors A. V thanks Australian Research Council for the Future Fellowship.

Notes and references

1. M. Park, D. H. Lee, C. L. Choi, S. S. Kim, K. S. Kim and J. Choi, Chem. Mater., 2002, 14, 2582.
2. Y. Ide, N. Ochi and M. Ogawa, Angew. Chem., Int. Ed., 2011, 50, 654.
3. E. Ruiz-Hitzky, M. Darder, F. M. Fernandes, B. Wicklein, A. C. S. Alcantara and P. Aranda, Prog. Polym. Sci., 2013, 38, 1392.
4. J. M. Adams and R. W. McCabe, in Handbook of Clay Science, ed. F. Bergaya, B. K. G. Theng and G. Lagaly, Elsevier, Netherlands, 2006, vol. 1, pp. 541–581.
5. A. Olaya, S. Moreno and R. Molina, Appl. Catal., A, 2009, 370, 7.
6. P. Cool and E. F. Vansant, in Handbook of Layered Materials, ed. S. M. Auerbach, K. A. Carrado and P. K. Dutta, Marcel Dekker Inc, New York, 2004, pp. 261–311.
7. S. Liu, J. H. Yang and J. H. Choy, J. Photochem. Photobiol., A, 2006, 179, 75.
8. (a) H. Jung, H. M. Kim, Y. B. Choy and J. H. Choy, Int. J. Pharm., 2008, 349, 283; (b) J. H. Yang, H. Jung, S. Y. Kim and J. H. Choy, J. Nanosci. Nanotechnol., 2013, 13, 7331.
9. (a) T. J. Pinnavaia, Science, 1983, 220, 365; (b) K. Ohtsuka, Chem. Mater., 1997, 9, 2039.
10. (a) S. Yamanaka, T. Nishihara and M. Hattori, Mater. Chem. Phys., 1987, 17, 87; (b) S. Yamanaka and K. Makita, J. Porous Mater., 1995, 1, 29.
11. J. H. Park, J. H. Yang, J. B. Yoon, S. J. Hwang and J. H. Choy, J. Phys. Chem. B, 2006, 110, 1592.
12. (a) J. H. Choy, H. Jung, Y. S. Han, J. B. Yoon, Y. G. Shul and H. J. Kim, Chem. Mater., 2002, 14, 3823; (b) J. H. Yang, H. S. Lee, S. M. Paek and Y. S. Han, Chem. Lett., 2011, 40, 1242; (c) Y. S. Han, S. Yamanaka and J. H. Choy, J. Solid State Chem., 1999, 144, 45.
13. (a) C. Ooka, H. Yoshida, K. Suzuki and T. Hattori, Microporous Mesoporous Mater., 2004, 67, 143; (b) C. Ooka, S. Akita, Y. Ohashi, T. Horiuchi, K. Suzuki, S. Komai, H. Yoshidac and T. Hattori, J. Mater. Chem., 1999, 9, 2943; (c) X. Ding, T. An, G. Li, S. Zhang, J. Chen, J. Yuan, H. Zhao, H. Chen, G. Sheng and J. Fu, J. Colloid Interface Sci., 2008, 320, 501; (d) D. Chen, G. Du, Q. Ahu and F. Zhou, J. Colloid Interface Sci., 2013, 409, 151.
14. J. Pires and M. L. Pinto, etc. in Pillared Clays and Related Catalysts, ed. A. Gil, S. A. Korili, R. Trujillano and M. A. Vicente, Springer Science & Business Media, New York, 2010, pp. 23–301.
15. (a) W. J. Ong, L. L. Tan, S. P. Chai, S. T. Yong and A. R. Mohamed, ChemSusChem, 2014, 7, 690; (b) M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69.
16. (a) M. J. Wilson, A Handbook of Determinative Methods in Clay Mineralogy, Chapman and Hall, New York, 1987, p. 60; (b) H. Tateyama, H. Noma, S. Nishimura, Y. Adachi, M. Ooi and K. Urabe, Clays Clay Miner., 1998, 46, 245.
17. (a) M. Anpo, N. Aikawa, Y. Kubokawa, M. Che, C. Louis and E. Giamello, J. Phys. Chem., 1985, 89, 5017; (b) J. H. Choy, J. H. Park and J. B. Yoon, J. Phys. Chem. B, 1998, 31, 5991.
18. F. Rouquerol, J. Rouquerol and K. Sing, Adsorption by Powders and Porous Solids, Academic press, New York, 1999.
19. T. Allen, Particle size measurement, Chapman & Hall, London, New York, 4th edn, 1997, pp. 540–652.
20. (a) G. Horvath and K. Kawazoe, J. Chem. Eng. Jpn., 1983, 16, 470; (b) R. J. Dombrowski, C. M. Lastoskie and D. R. Hyduke, Colloids Surf., A, 2011, 187–188, 23.
21. X-ray absorption spectroscopic (XAS) experiments at the Ti K-edge were performed with the X-ray absorption fine structure (XAFS) facility installed at beam line 10C in Pohang Accelerator Laboratory (PAL), Republic of Korea.
22. F. Farges, G. E. Jr Brown and J. J. Rehr, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 56, 1809.
23. V. Luca, J. Phys. Chem. C, 2009, 113, 6367.
24. (a) M. J. Paek, H. W. Ha, T. W. Kim, S. J. Moon, J. O. Baek, J. H. Choy and S. J. Hwang, J. Phys. Chem. C, 2008, 112, 15966; (b) T. W. Kim, M. W. Ha, M. J. Paek, S. H. Hyun, I. H. Baek, J. H. Choy and S. J. Hwang, J. Phys. Chem. C, 2008, 112, 14853; (c) S. M. Paek, H. Jung, Y. J. Lee, M. Park, S. J. Hwang and J. H. Choy, Chem. Mater., 2006, 18, 1134.
25. Y. Lin, C. Ferronato, N. Deng, F. Wu and J. M. Chovelon, Appl. Catal., B, 2009, 88, 32.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra12880j

---