Synthesis and characterization of hierarchical TiO₂ microspheres composed of nanorods: effect of reaction conditions on nanorod density

Abstract

Urchin-like hierarchical TiO₂ microspheres consisting of numerous nanorods were successfully synthesized by a facile hydrothermal process. The effects of reaction conditions (such as additives and reactant concentrations) on TiO₂ morphology, particularly on rod density, were systematically investigated. The relationship between rod density and rod diameter was discussed. Based on the above results including a morphological evolution study, a possible mechanism was also proposed to elucidate the growth process of hierarchical TiO₂ microspheres using nanorods as building units.

---

Introduction

Titanium dioxide (TiO₂), regarded as one of the most important functional materials, has a wide range of applications, such as gas semiconductors, lithium ion batteries, photocatalysis, printing photoconductors, catalyst supports and solar-battery components.^1–9 Previous studies suggest that these applications are strongly related to the morphology and the structure of TiO₂.^1–4,7–12 Therefore, controllable synthesis of TiO₂ with novel well-crystallized nanostructures or morphology is highly desired.¹³ Up to now, various nanostructured TiO₂ materials, such as nanotubes, nanorods and nanosheets, have been designed for achieving improved properties.^6,7,12,14 Among these nanostructures, recently three-dimensional (3D) hierarchical TiO₂ nanomaterials, such as sea urchin-shaped nanostructures, have attracted much more attention than one- (1D) or two- dimensional ones due to their superior properties.^3,4,7,11,15 The 3D hierarchical structures are expected to play important roles in fabricating the next generation microelectronic and optoelectronic devices, due to their capability of being used as both building units and interconnections.¹⁶

However, controllable fabrication of complex 3D hierarchical architectures is still a significant challenge.^7,12 Conventional routes for preparing 3D TiO₂ materials are generally difficult and require hard/soft templates to facilitate the growth of TiO₂.¹³ However, the templates involved in synthesis are sacrificial and need subsequent removal by calcination or dissolution, which can increase production cost and introduce heterogeneous impurities.^4,13 Thus, various 3D hierarchical TiO₂ nanostructures have been developed using template-free self-assembly method.⁴ The typical route reported involves assembling 1D titanate precursors and annealing the formed titanate aggregates at 350–500 °C.³ Such 3D nanomaterials are micrometer-scale aggregates of 1D TiO₂ nanoparticles with polycrystalline structures.¹⁷

Besides above-mentioned synthesis route, many other template-free methods involving different Ti precursors are also reported. For example, Mendoza et al.¹⁸ prepared urchin-like rutile microstructures by thermal decomposition and oxidation of K₂TiF₆ at 640 °C. However, K₃TiOF₅ impurity and by-product F₂ gas were present in the thermal reaction. Previous work also reported the use of titanium chlorides to synthesis 3D TiO₂ microspheres. Park et al.¹ prepared carbon-painted urchin-like rutile TiO₂ microspheres by the hydrolysis and oxidation of TiCl₃. Hayashi et al.³ used TiCl₄ to synthesize urchin-shaped rutile TiO₂ having large solid cores (∼420 nm in diameter) with many short needle-like structure (∼15–30 nm in length). Rui et al.¹⁹ prepared rutile TiO₂ nanorod microspheres by salt-assisted hydrothermal treatment of TiCl₄ solution. In addition, Xu et al.²⁰ synthesized urchin-like TiO₂ nanostructure using hydrothermal treatment of emulsion (TiOSO₄, absolute alcohol, glycerol, etc.). In spite of the above efforts, some disadvantages in previous studies still exist, e.g. corrosive and toxic by-product F₂ gas,¹⁸ instability and flammability of TiCl₃,¹ strong irritation of TiCl₄ (ref. 19) and use of organic solvents,²⁰ etc. Therefore, a mild and environmental-friendly process for preparing 3D TiO₂ nanostructures is much desired for practical applications.

Recently, a mild synthesis route involving Ti-based precursors and HCl under hydrothermal treatment shows great potentials in rutile TiO₂ nanorods synthesis. Sarkar et al.¹⁶ reported TiO₂ nanoarchitectures by hydrothermally treating the reaction mixture containing titanium butoxide, oleic acid (as a solvent) and hydrochloric acid. Zhao et al.²¹ synthesized rutile nanorods by this method without oleic acid and the rutile nanorods keep excellent independence from each other. Also, Yu et al.²² fabricated vertically aligned rutile TiO₂ nanorods on fluorine-doped tin oxide (FTO) substrates using similar reaction system but in the presence of ethanol and FTO substrate. Thus, how to use Ti-based and HCl hydrothermal system to assembly TiO₂ nanorods into 3D hierarchical TiO₂ microspheres with tunable rod density is still not investigated yet.

Herein, we employ the above synthesis system (i.e., Ti(OBu)₄/HCl) but without the use of ethanol, oleic acid and FTO. The results suggest that, under suitable reaction conditions, the above 1-D rutile TiO₂ nanorods can be assembled into hierarchical TiO₂ microspheres with controllable rod density. Since nanorods are the building units of 3D hierarchical TiO₂ nanostructures, their density in the hierarchical spheres may have significant influences on the performances. This work mainly focused on the effects of reaction conditions (e.g., additives, reactant concentration and reaction time) on the TiO₂ morphology (particularly on the density of nanorods as building units), and the possible mechanism for the growth of hierarchical TiO₂ microspheres. To the best of our knowledge, it is the first time to report the study on the effects of synthesis conditions on TiO₂ morphological structures from the view of nanorod density. It is also believed that varying the density of nanorods will have significant impacts on the practical applications of urchin-like hierarchical TiO₂.

Results and discussion

Urchin-like hierarchical TiO₂ microspheres consisting of nanorods were synthesized via hydrothermal treatment of Ti(OBu)₄ in HCl aqueous solution. In a typical procedure, 3 mL Ti(OBu)₄ was first homogeneously mixed with 15 mL HCl (37%) and 15 mL H₂O followed by hydrothermal treatment in a 50 mL Teflon-lined stainless steel autoclave at 180 °C for 4 hours.

Fig. 1 confirms that the as-prepared urchin-like hierarchical TiO₂ microspheres are successfully fabricated under the above reaction conditions. Fig. 1a clearly shows that the hierarchical TiO₂ spheres are ∼5 μm in diameter. To verify the hierarchical structure consisting of nanorods, the samples were ground in a mortar to break away the nanorods from the microspheres. As shown in the inset of Fig. 1a, the nanorods radiate from the centre of the crystals to form a spherical structure. Thus, the as-prepared TiO₂ microspheres are similar to the sea-urchin structure. Selected area electron diffraction (SAED) pattern analysis shown in Fig. S1† indicates that the growth of TiO₂ nanorod is along (001) direction. From the FESEM image with a higher magnification of TiO₂ hierarchical microsphere surface (Fig. 1b), it can be clearly seen that numerous nanorods of about 20 nm in diameter can be considered as building units of the microsphere and the rod density is estimated to be 8.9 × 10⁸ rods per mm².

Fig. 1 (a) FESEM image of urchin-like hierarchical TiO₂ microsphere (synthesized with 3 mL Ti(OBu)₄ at 180 °C for 4 h; inset: fragment of the destroyed TiO₂ microspheres) and (b) enlarged FESEM image of TiO₂ microsphere surface.

XPS analysis was carried out to investigate the oxidation states of as-prepared TiO₂. Fig. 2 displays the Ti2p and O1s XPS spectra of TiO₂ hierarchical microspheres. The peaks at 459.7 eV and 465.4 eV of Ti2p are attributed to Ti⁴⁺2p_3/2 and Ti⁴⁺2p_1/2, respectively (i.e., the oxidation state of titanium is +4).^23,24 The O1s XPS spectra are asymmetric, indicating that both lattice oxygen and adsorbed oxygen species are present on the TiO₂ surface.^23,24 Therefore, XPS analysis confirms that the as-prepared sample is in the form of TiO₂.

Fig. 2 XPS spectra of urchin-like hierarchical TiO₂ microspheres: (a) Ti2p and (b) O1s.

The crystal phase of hierarchical TiO₂ microspheres was further analyzed by Raman and XRD techniques. As shown in Fig. 3a, the peaks located at 448 cm⁻¹ (E_g) and 612 cm⁻¹ (A_1g) and a two-phonon scattering band at 237 cm⁻¹ for hierarchical TiO₂ microspheres can be ascribed to the characteristic vibration modes of the rutile phase.† XRD result (Fig. 3b) also confirms that the TiO₂ microspheres are rutile phase (PDF no: 75-1749), consistent with XPS and Raman analysis. All the diffraction peaks, such as the peaks at 2θ = 27.4°, 36.1°, 41.2° and 54.3°, are indexed to the characteristic peaks of pure rutile phase,^{1,3,8–10,15} implying high purity of the TiO₂ microspheres.

Fig. 3 (a) Raman spectra and (b) XRD pattern of urchin-like hierarchical TiO₂ microspheres.

The effect of Ti(OBu)₄ amount on TiO₂ morphology and rod density was investigated. As shown in Fig. 4 and 5, when 1.5 mL Ti(OBu)₄ is added in the hydrothermal synthesis, sparse TiO₂ nanorods with low rod density can be observed on the microspheres surface. This may be due to the insufficient Ti(OBu)₄ precursors for the formation of TiO₂ nanorods, which are considered as the building units for constructing TiO₂ microspheres by self-assembly strategy. Moreover, the relatively large nanorod diameter can be the second reason for the low rod density. Interestingly, when Ti(OBu)₄ amount is increased to 3.0 mL, the rod density of TiO₂ microspheres is significantly increased to 8.9 × 10⁸ rods per mm² (Fig. 1 and 4). However, when the Ti(OBu)₄ amounts are further increased to 4.5 mL and 6.0 mL, the rod densities on TiO₂ microspheres are sharply decreased to 3.3 × 10⁸ and 1.7 × 10⁸ rods per mm², respectively (Fig. 4). In addition, the rough surface of TiO₂ microspheres is found in these two samples (Fig. 5c and d). The relatively low rod density at high Ti(OBu)₄ amounts is likely ascribed to the formation of TiO₂ debris on microspheres surface (Fig. 5c and d) and the increase in the diameter of TiO₂ nanorods (Fig. 4) caused by the fast growth rate at high reactant concentrations. As shown in Fig. S2,† it is found that rod diameter of TiO₂ microspheres showed exponential decrease with the increase of rod density. The relationship between rod diameter and rod density is further verified by TiO₂ samples prepared with 0.75 mL Ti(OBu)₄. It can be clearly seen from Fig. 5a that the hierarchical TiO₂ samples exhibit the largest rod diameter (ca. 400 nm) and the lowest rod density. It should be noted that the TiO₂ nanorod diameter and density are kinetically controlled by the nucleation and growth rate in the solution, which should be related to the precursor concentrations. Thus, an optimal amount of Ti(OBu)₄ (ca. 3 mL) is demonstrated for the formation of TiO₂ microspheres endowed with smooth surface (Fig. 1) and highest rod density (Fig. 4).

Fig. 4 Relationship between rod density, rod diameter and used amount of Ti(OBu)₄ precursor.

Fig. 5 FESEM images of urchin-like hierarchical TiO₂ microspheres synthesized with different amounts of Ti(OBu)₄ (a) 0.75 mL, (b) 1.5 mL, (c) 4.5 mL and (d) 6.0 mL at 180 °C for 4 h.

Fig. 6 illustrates the effects of the presence of various additives (ethylene glycol, ethanol, NaCl and Na₂SO₄) during hydrothermal synthesis on the morphology of TiO₂ microspheres. The rod densities of TiO₂ microspheres synthesized in the presence of ethylene glycol, ethanol, NaCl are estimated to be 8.9 × 10⁸, 6.9 × 10⁸ and 5.8 × 10⁸ rods per mm², respectively. In comparison with the rod density of 8.9 × 10⁸ rods per mm² obtained in the absence of additives, ethylene glycol does not show any influences on TiO₂ microsphere growth in terms of rod density. In addition, both ethanol and NaCl decrease the rod density of urchin-like TiO₂ microspheres, possibly due to the changes in Cl⁻ ion concentration or ionic strength. Interestingly, when Na₂SO₄ is introduced to the hydrothermal synthesis, only solid TiO₂ spheres with smooth surface are obtained (Fig. 6g and h). This suggests that the presence of foreign SO₄²⁻ has significant impacts on the TiO₂ microstructures. Moreover, XRD and electron diffraction patterns in Fig. S3 and S4† indicate that the TiO₂ microspheres obtained in the presence of SO₄²⁻ is in the form of polycrystalline anatase phase with no preferred crystallographic orientation. Thus, the synthesis of TiO₂ microspheres using nanorods as building blocks can be well tuned by adding different additives and varying the reactant concentrations.

Fig. 6 FESEM images of TiO₂ microspheres synthesized in the presence of different additives. (a) and (b): ethanol; (c) and (d): ethylene glycol; (e) and (f): NaCl; (g) and (h): Na₂SO₄.

The time-dependent morphology evolution of hierarchical TiO₂ microspheres was also investigated. In this work, the hydrolysis reactions involved in hydrothermal synthesis are suggested as follows:^25,26

Ti(OBu)₄ + nH₂O → (OBu)_4−nTi(OH)_n + nBuOH (n = 1–4) (1)

≡Ti–OH + HO–Ti≡ → ≡TiOTi≡ + H₂O (2)

≡Ti–OBu + HO–Ti≡ → ≡TiOTi≡ + BuOH (3)

Previous work indicated the hydrolysis of alkoxide precursors was fast so that it was difficult to control the growth of TiO₂ particles.²⁷ Bleta et al.²⁸ used titanium isopropoxide (Ti(iOPr)₄) to prepare mesoporous TiO₂ and found nucleation and growth were completed within a few seconds due to the fast hydrolysis, leading to rapid agglomerate and the formation of large precipitates. The presence of acid in the TiO₂ synthesis can effectively slow down the TiO₂ growth rate. He et al.²⁹ prepared mesoporous TiO₂ with Ti(OBu)₄, triethanolamine, ethanol, H₂O, etc. They found that H₂SO₄ can reduce the hydrolysis rate of Ti(OBu)₄ to produce narrow pore size distributions. Thus, it is concluded that the morphology controlling agent HCl in this work can slow down the hydrolysis of Ti-based precursors. Indeed, the reaction solution with hydrothermal treatment of 15 min is still clear and no obvious solid product is found.

When hydrothermal treatment is prolonged to 30 min, the products with flower-like structures are produced and many short rods of ∼100 nm in diameter and ∼200 nm in length are formed (Fig. 7a and b). Then the morphology becomes more and more regular with the increase in the reaction time, as shown in Fig. 7. At 1 h and 2 h, cauliflower-like TiO₂ spheres with rough surface are observed (Fig. 7c–f). When the reaction time is further prolonged to 3 h, hierarchical TiO₂ microspheres are basically formed and the nanorods as building blocks are uniformly distributed in the microspheres (Fig. 7g and h). Until the reaction time is extended to 4 h, the TiO₂ hierarchical microspheres became fully round as shown in Fig. 1. It should be mentioned that the rod density of TiO₂ microspheres increases with the prolonged hydrothermal reaction time. The time-dependent morphology study confirms the evolution of rods to microspheres under hydrothermal conditions.

Fig. 7 Morphology evolution of urchin-like hierarchical TiO₂ microspheres obtained at different stages (a) and (b): 30 min (inset: enlarged SEM image of rod-like structure); (c) and (d): 1 h; (e) and (f): 2 h; (g) and (h): 3 h. Scale bar in the inset of Fig. 7a is 200 nm.

It is generally popular to explain the formation of 3D hierarchical structure with a two-stage growth mechanism involving the initial formation of primary 1D nanostructure and the subsequent self-assembly.^16,30 The previous proposed mechanism is not consistent with our results, as no well-defined 1D nanostructure of small sized nanorods is formed at initial stage (Fig. 7b, d, 7f, 7h and 1b). On the contrary, the hierarchical TiO₂ microspheres using nanorods as the building block became well-defined at the later stage of the reaction (Fig. 1 and 7). This process is called dissolution and recrystallization. The presence of Cl⁻ ions is beneficial for converting ∼200 nm nanorods formed at the beginning into smaller sized nanorods (∼20 nm in diameter). It is reported that the presence of Cl⁻ ions can suppress the growth of the (110) planes and improve the growth along the (001) direction due to the selective adsorption of Cl⁻ ions onto the (110) plane of rutile TiO₂,^1,15,16 which is also verified by SEAD result shown in Fig. S1.† Thus, the formation of hierarchical TiO₂ microspheres in this work is likely attributed to a self-templating dissolution/recrystallization process whose recrystallization directs the growth of nanorods along the (001) direction due to the selective adsorption of Cl⁻ ions. However, when foreign SO₄²⁻ ions are present in hydrothermal synthesis, only solid TiO₂ spheres can be obtained (Fig. 6g and h). This is possibly because the (001) orientation is disrupted by the presence of SO₄²⁻ ions resulting from the strong affinity of SO₄²⁻ ions to TiO₂ surface.³¹ The competitive adsorption of SO₄²⁻ ions can weaken the interaction between Cl⁻ and TiO₂ surface, leading to the morphology and crystal phase transformation from urchin like rutile structure to solid anatase spheres.

Conclusions

In summary, a facile hydrothermal treatment of Ti(OBu)₄/HCl aqueous solution is successfully applied to fabricate hierarchical TiO₂ microspheres using nanorods as building blocks. The morphology of TiO₂ microspheres, such as the rod density, can be well tuned by changing the amounts of precursor, additives and reaction time. Under the applied reaction conditions, the rod diameter of TiO₂ microspheres shows an exponential decrease with the increase of rod density. The morphological analysis indicates that hierarchical TiO₂ microspheres form via a self-templating dissolution/recrystallization process. As the reaction proceeds, the microspheres and their building units (i.e., nanorods) become more and more well-defined, probably resulting from the recrystallization which orients towards (001) direction due to the presence of Cl⁻ ions. When SO₄²⁻ ions are added in TiO₂ synthesis, the recrystallization towards (001) direction is disrupted, leading to the production of solid microspheres with the change of crystal phase from rutile to anatase. Owing to the hierarchical structure of TiO₂ microspheres, the as-prepared materials have great potential applications in energy storage devices and photocatalysis applications. The further studies on the practical applications of the urchin-like TiO₂ hierarchical microspheres endowed with different rod densities and rod diameters will be performed in the future work.

Experimental

Urchin-like hierarchical TiO₂ microspheres were prepared by a facile hydrothermal process. In a typical procedure, 3 mL Ti(OBu)₄ (97%, Sigma Aldrich) and a mixture of 15 mL HCl (37%, Sigma Aldrich) and 15 mL H₂O (Millipore, Inc.,18 MΩ cm⁻¹) were added in a dry 50 mL Teflon-lined stainless steel autoclave. The hydrothermal reaction was performed at 180 °C for 4 hours. After the reaction was completed, the precipitation was collected by centrifuge and sequentially washed separately with H₂O and ethanol for three times. Finally, the white product was dried in a vacuum oven at 60 °C for 12 hours. In addition, a series of TiO₂ samples were prepared following the above procedures with 0.75 mL, 1.5 mL, 4.5 mL and 6.0 mL Ti(OBu)₄. To study the effect of additives on TiO₂ morphology, 1.5 mL ethylene glycol (99.8%, Sigma Aldrich), 1.5 mL ethanol (99.5%, Sigma Aldrich), 15 mmol NaCl (99.5%, Sigma Aldrich) and 15 mmol Na₂SO₄ (99.0%, Sigma Aldrich) were also separately added in the above-mentioned solution for TiO₂ synthesis. For morphology evolution study, TiO₂ samples were collected at different reaction stages (15 min, 30 min, 1 h, 2 h, and 3 h).

Scanning electron microscopy (SEM) images were obtained on a Hitachi S4800 field-emission scanning electron microscope. Transmission electron microscope (TEM) and selected area electron diffraction (SAED) were performed by a JEM-2010 transmission electron microscope. Raman spectrum was recorded on a Raman inVia-Reflex microscopy system (Renishaw, Wotton-under-Edge, UK) equipped with a He–Ne laser as the excitation source (633 nm, 17 mW). X-ray diffraction (XRD) patterns were measured with a D8 diffractometer (Bruker, Germany) with Cu Kα radiation at a step length of 0.05° s⁻¹ in the range of 20–80° using a LynxEye detector. The X-ray photoelectron spectroscopy (XPS, Model PHI5600) measurements were performed to analyze as-prepared TiO₂ samples. The binding energy was calibrated using C1s core line (284.6 eV) as reference. XPS data was treated on software XPSPEAK 4.1.

Acknowledgements

The research works presented in this paper are funded by Technological and Higher Education Institute of Hong Kong. Financial support from the Program for Liaoning Innovative Research Team in University (LNIRT; Project no. LT2013010) is also gratefully acknowledged.

Notes and references

1. K. Park, K. Min, Y. Jin, S. Seo, G. Lee, H. Shim and D. Kim, J. Mater. Chem., 2012, 22, 15981.
2. T. Fukuoka and A. Takasu, RSC Adv., 2014, 4, 159834.
3. K. Hayashi, M. Nakamura, Y. Makita, R. Fujiwara, T. Kori and K. Ishimura, Mater. Lett., 2011, 65, 3037.
4. Y. Zhou, Y. Huang, D. Li and W. He, Mater. Res. Bull., 2013, 48, 2420.
5. J. Su, X. Zou and J. Chen, RSC Adv., 2014, 4, 13979.
6. R. S. Hyam and D. Choi, RSC Adv., 2013, 3, 7057.
7. F. Huang, Z. Fu, A. Yan, W. Wang, H. Wang, Y. Wang, J. Zhang, Y. Cheng and Q. Zhang, Cryst. Growth Des., 2009, 9, 4017.
8. Q. Cheng, V. Pavlinek, Y. He, Y. Yan, C. Li and P. Saha, Colloid Polym. Sci., 2011, 289, 799.
9. J. Archana, M. Navaneethan and Y. Hayakawa, Mater. Lett., 2013, 98, 38.
10. M. Ge, J. W. Li, L. Liu and Z. Zhou, Ind. Eng. Chem. Res., 2011, 50, 6681.
11. H. Qiao, Y. Wang, L. Xiao and L. Zhang, Electrochem. Commun., 2008, 10, 1280.
12. L. Liu, H. Liu, Y. Zhao, Y. Wang, Y. Duan, G. Gao, M. Ge and W. Chen, Environ. Sci. Technol., 2008, 42, 2342.
13. L. Zhu, C. Huang, J. Liu, N. Bing, H. Jin and L. Wang, Mater. Lett., 2013, 106, 348.
14. K. Fan, T. Peng, J. Chen, X. Zhang and R. Li, J. Power Sources, 2013, 222, 38.
15. H. Bai, Z. Liu, S. S. Lee and D. D. Sun, Appl. Catal., A, 2012, 447–448, 193.
16. D. Sarkar, C. K. Ghosh and K. K. Chattopadhyay, CrystEngComm, 2012, 14, 2683.
17. Y. Mao, M. Kanungo, T. Hemraj-Benny and S. S. Wong, J. Phys. Chem. B, 2006, 110, 702.
18. C. I. VillaVelázquez-Mendoza, J. L. Rodríguez-Mendoza, R. P. Hodgkins, V. Ibarra-Galván, A. L. Leal-Cruz, A. López-Valdivieso and M. I. Pech-Canul, Mater. Lett., 2014, 121, 191.
19. Y. Rui, Y. Li, Q. Zhang and H. Wang, CrystEngComm, 2013, 15, 1651.
20. S. Xu, Y. Zhang, S. Wang, J. Xu, H. Ding and G. Li, Eur. J. Inorg. Chem., 2013, 2601.
21. B. Zhao, L. Lin and D. He, J. Mater. Chem. A, 2013, 1, 1659.
22. H. Yu, J. Pan, Y. Bai, X. Zong, X. Li and L. Wang, Chem.–Eur. J., 2013, 19, 13569.
23. H. Guo, M. Kemell, M. Heikkilä and M. Leskelä, Appl. Catal., B, 2010, 95, 358.
24. J. F. Marco, A. Cuesta, M. Gracia, J. R. Gancedo and P. Panjan, Thin Solid Films, 2005, 492, 158.
25. X. Lü, F. Huang, J. Wu, S. Ding and F. Xu, ACS Appl. Mater. Interfaces, 2011, 3, 566.
26. J. Ye, W. Liu, J. Cai, S. Chen, X. Zhao, H. Zhou and L. Qi, J. Am. Chem. Soc., 2011, 133, 933.
27. X. Yao, X. Liu, T. Liu, K. Wang and L. Lu, CrystEngComm, 2013, 15, 10246.
28. R. Bleta, P. Alphonse and L. Lorenzato, J. Phys. Chem. C, 2010, 114, 2039.
29. J. He, Z. Zhang, J. Shen, Y. Zhang and Y. Chen, J. Inorg. Mater., 2009, 24, 43.
30. L. Xiang, X. Zhao, J. Yin and B. Fan, J. Mater. Sci., 2012, 47, 1436.
31. Y. Li, N. Lee, D. Hwang, J. S. Song, E. G. Lee and S. Kim, Langmuir, 2004, 20, 10838.

---

Footnotes

† Electronic supplementary information (ESI) available: TEM image and SAED pattern of hierarchical TiO₂ microspheres, relationship between TiO₂ rod density and rod diameter, TEM image, SAED and XRD patterns of solid TiO₂ microspheres. See DOI: 10.1039/c4ra04413d

‡ These authors are equally contributed to this article.

---