One-step solution synthesis of bismuth sulfide (Bi₂S₃) with various hierarchical architectures and their photoresponse properties

Abstract

In this paper, we introduce a facile and phosphine-free one-step solution method to synthesize size- and shape-controlled bismuth sulfide (Bi₂S₃) with hierarchical architectures. Changing variables, such as the reaction temperature, the ratio of precursors, and the concentration of oleic acid were observed to influence the resultant shape of Bi₂S₃ microstructures. For the formation of Bi₂S₃ hierarchical architectures, the crystal splitting growth mechanism played the dominant role. The absorption spectra were recorded at room temperature, which revealed that the obtained Bi₂S₃ product was a direct band gap semiconductor and the band gap Eg was estimated to be about 1.9 eV. Furthermore, the I–V characteristics of the Bi₂S₃-based device show a significant increase by ca. 1 order of magnitude compared with the dark state, indicating an enhanced conductivity and high sensitivity. The response and decay times are estimated to be about 0.5 and 0.8 s, respectively, which are short enough for it to be an excellent candidate for high-speed and high-sensitivity photodetectors or optical switches. Thus the Bi₂S₃ hierarchies as building blocks may offer the potential for monolithic, low-cost and large-scale integration with CMOS electronics.

---

1. Introduction

During the past decades, chalcogenide semiconductor crystals have attracted broad attention due to their unique shape- and size-dependent physical and chemical properties, and much effort has been dedicated to achieving rational control over the morphology of the materials.^1–3 Therefore, the ability to manipulate the morphology, size and size distribution of inorganic materials is still an important goal in modern materials physics and chemistry. As an important V–VI semiconductor material, bismuth sulfide (Bi₂S₃) has aroused extensive research interest for various potential applications in expanding the use of X-ray computed tomography imaging (CT),⁴ Schottky diode,⁵lithium ion battery⁶ and gas sensors.⁷ Moreover, Bi₂S₃ is a significant layered semiconductor with a direct band gap of 1.3 eV and large absorption coefficient,⁸ making it an ideal candidate for solution-processed visible-wavelength photodetectors.^9–11 Hierarchical architectures possess a high surface to volume ratio with a large proportion of free open edges, and it is anticipated that such complex structures would offer new opportunities to explore their unique properties to fabricate electronic and photonic devices.¹² Since the properties and performance of Bi₂S₃ are largely determined by their morphology, thus exploring more facile routes to synthesize Bi₂S₃ micro- and nano-scale structures are always desired. Many groups have focused on preparing Bi₂S₃ crystals by using various strategies, including template-assisted route,^13–15solvothermal process,^16–19 sonochemical approach,²⁰ and vapor deposition method.²¹

Herein, based on one-step solution-phase method, we systematically synthesize a series of Bi₂S₃ hierarchical architectures through changing the reaction parameters in the synthetic system. According to these experimental results, the crystal splitting growth mechanism was contributed to forming these Bi₂S₃ hierarchical architectures. In addition, the I–V characteristics of a Bi₂S₃–based device show a significant increase by ca. 1 order of magnitude compared with the dark state, indicating an enhanced conductivity and high sensitivity. The response and decay times were estimated to be about 0.5 and 0.8 s, respectively, which are short enough to enable it as an excellent candidate for high-speed and high-sensitivity photodetectors or optical switches.

2. Experimental section

2.1. Chemicals

Bismuth(III) acetate (Bi(Ac)₃, ≥ 99.99%), sulfur (S, 99.998%), oleic acid (OA, 90%) and 1-octadecene (ODE, 90%) were purchased from Aldrich. Toluene, methanol, chloroform and acetone were obtained from commercial sources. All chemicals were used without further purification.

2.2. Synthesis

All experiments were carried out using standard airless techniques: a vacuum/dry nitrogen gas Schlenk line was used for the synthesis and a nitrogen glove box for storing and handling air- and moisture-sensitive chemicals. Typically, 0.2 mmol of Bi(Ac)₃, 1 mL of OA and 9 mL of ODE were loaded into a 50 mL three-neck flask in a glove box. Then the flask was sealed and taken out to connect to the Schlenk line equipped with thermocouple and mantle. The mixture was heated to 130 °C and degassed for 20 min to form a colorless solution and remove the moisture under a nitrogen flow and vigorous magnetic stirring conditions. Then the colorless solution was cooled to room temperature naturally. The sulfur precursor, a nearly colorless solution, containing 0.3 mmol of elemental sulfur and 1.5 mL of 1-octadecene, was added to the bismuth acetate solution. Finally, the mixed solution was heated to 180 °C rapidly and the reaction time was last for 2 h. The mixture solution changed color from colorless to brown gradually, indicating the formation of Bi₂S₃. The black precipitate was purified with toluene and methanol for the first time, and then a little of chloroform and an excess amount of acetone were added to be washed for another twice.

2.3. Characterization

The field emission scanning electron microscopy (FESEM) measurements were carried out with a scanning electron microscope (JEOL, JSM-6700F) operated at an acceleration voltage of 8 kV. Composition of the specimens was analyzed using energy-dispersive X-ray (EDX) spectroscopy (INCA energy) attached to the JSM-6700F. An X-ray powder diffractometer (Shimadzu, XRD-6000) manipulated at 40 kV and 40 mA with Cu Kα radiation (λ = 1.5406 Å) was used to record the X-ray diffraction (XRD) patterns of the samples. Transmission electron microscopy (TEM) images were obtained on a Hitachi H-8100IV microscope operated at 200 kV. High-resolution transmission electron microscopy (HRTEM) images were measured via a JEM-2100 transmission electron microscopy at 300 kV. The absorption spectra were performed by a Shimadzu UV-3150 spectrometer.

The photoresponse devices were fabricated on cleaned indium tin oxide (ITO) patterned glass substrates. A certain thickness layer of Bi₂S₃ was deposited on top of the ITO by spin coating using chloroform as the solvent. The Bi₂S₃ films were sandwiched between two transparent ITO coated glass substrates with a device configuration of ITO/Bi₂S₃/ITO. The anode consisting of a glass substrate coated with ITO was cleaned by sequential ultrasonic treatment in detergent, deionized water, acetone, and isopropyl alcohol. All the current–voltage (I–V) characteristics of the photovoltaic devices were measured in ambient conditions recorded using Keithley 2400 Source Meter in the dark and under simulated AM 1.5 illumination (100 mW cm⁻²) by Solar Simulators (SCIENCETECH SS-0.5 K). The active device area was 5 mm². Film thickness was measured by Veeco DEKTAK 150 surface profilometer. Sensitivity of the photoresponse was recorded using electrochemical workstation in the dark and under illumination by halogen lamp.

3. Results and discussion

3.1. Structural characterization

The crystal structure and chemical composition of the as-prepared samples were confirmed by powder X-ray diffraction (XRD). Fig. 1a shows the typical XRD pattern of as-prepared Bi₂S₃ hierarchical architectures. All the diffraction peaks in the pattern are labeled and can be readily indexed to a pure orthorhombic phase bismuth sulfide with cell constants of a = 11.15 Å, b = 11.30 Å, c = 3.98 Å, which are in good agreement with the standard JCPDS card No.17-0320. The sharp diffraction peaks indicate that the products are highly crystalline. In addition to XRD result, the chemical composition of the as obtained product was further checked by energy dispersive X-ray analysis (EDX). The EDX spectrum given in Fig. 1b shows the stoichiometry of the products. The element proportion of bismuth to sulfide is found to be approximately 2 : 3 as expected (Bi: 38.60%; S: 61.40%), which is in good agreement with the stoichiometric ratio of the compound of Bi₂S₃. Therefore, both XRD and EDX analyses show that pure orthorhombic-phase Bi₂S₃ microstructure is successfully synthesized by the present one-step synthetic strategy.

Fig. 1 The typical XRD pattern (a) and EDX spectrum (b) of the as-prepared Bi₂S₃ architectures. The inset in (b) shows the table of composition.

3.2. Morphological characterization and proposed formation mechanism

Fig. 2a,b show a series of FESEM micrographs of Bi₂S₃ MCs with different magnifications. For these FESEM investigations, uniform Bi₂S₃ hierarchical architectures with lengths of 2 μm are observed. The bundles of nanofilaments were observed by high magnification and possess the same symmetry, size and structure. It looks like a sheaf of the butted ear of ripe wheat in autumn. To provide further insight into the structures of the Bi₂S₃ hierarchical architectures, the morphologies of the as-prepared products have also been investigated by TEM and HRTEM, respectively. Fig. 2c represents an overall TEM image at low magnification; we can see that nearly monodisperse microstructures which seem to the butted spike were obtained. Fig. 2d depicts a typical TEM image of individual Bi₂S₃ hierarchies with an average size of 2 μm which is in correspondence with FESEM results. The HRTEM image near the node of a bundle was recorded as shown in Fig. S1 (ESI†). From the figure we can see that the HRTEM image contains different orientations corresponding to the different facets. While every orientation is a characteristic of individual filament, as a consequence, we can deduce that the bundles of Bi₂S₃ MCs were formed by crossing the nanofilaments. HRTEM image, Fig. 2e, reveals that the filament which has an average diameter of ∼15 nm is highly single crystalline in nature. Notably, the distance between fringes shown in Fig. 2f was measured to be 0.357 nm, corresponding to the (130) interplanar spacing of the orthorhombic bismuthinite. One can also deduce that the filament grows along the [001] direction²² which is perpendicular to the (130) crystal facet. The formation of Bi₂S₃ hierarchical architectures with [001] direction is related to the inherent crystal structure. Fig. S2 (ESI†) illustrates the unit cell prolonged 4 times along its c-axis branded with the lattice constants of orthorhombic bismuth sulfide. It is known that Bi₂S₃ MCs have a strong tendency to grow along its c-axis into 1D anisotropic morphology and it is in fact a lamellar structure composed of Bi₂S₃ bands elongated along the c-axis and kept together through van der Waals interactions.²³ The chain-like arrangement of Bi and S atoms parallel to the c-axis in the orthorhombic lattice contributes to the growth of Bi₂S₃ nanofilaments along the [001] direction.

Fig. 2 The typical low- and high-magnification FESEM and TEM images (a)–(d) of the as-prepared Bi₂S₃ architectures; (e) and (f) shows HRTEM images of an individual filament formed Bi₂S₃ architectures. A corresponding fast Fourier transform (FFT) of the lattice is shown in the inset of (e). The area marked with a white frame in (e) is displayed in (f).

The hyperbranched sheaf-like structures were attributed to the well-known crystal splitting phenomenon,¹⁸ which often occur in the formation of some minerals in nature.²⁴ As shown in Fig. 3a–d, a series of time-dependent morphologies were obtained to get insight into the growth mechanism of Bi₂S₃ MCs. As the reaction proceeds, the nanorods start to split into small nanofilaments along the growth direction. The individual nanofilaments continue to grow, mainly in the elongated direction and finally form the hyperbranched sheaf-like structures. Thus the observed temporal shape evolution provided strong evidence that the crystal splitting behavior is conducive to the formation of sheaf-like Bi₂S₃ hierarchies in the solution phase process. Additionally, the sheaf-like feature appeared as soon as the solution color changed, indicating the splitting course occurred. Along with time extension, the sheaf-like Bi₂S₃ MCs rapidly grew larger from 350 nm to 1 μm only over the course of ∼10 min. As a consequence, the crystal splitting behavior was believed to play the dominant role on the earlier formation stage of sheaf-like Bi₂S₃ MCs and the splitting process was generally associated with fast crystal growth. Fig. 3e illustrates the schematic of crystal splitting process. When the temperature was brought up to ∼170 °C, the nuclei and growth of Bi₂S₃ occurred rapidly, and then the quasi-rod-like Bi₂S₃ MCs were formed. Once small particles are formed in the reaction solution, they are active and proceed to the self-assembly process to form larger nanocrystals to minimize the surface energies, resulting in density stacking to form Bi₂S₃ hierarchical architectures.²⁵ As we know, the shape of a crystal is determined by the relative specific energy of each face.²⁶ The fast growing planes with high energy should disappear to leave behind the slowest growing planes with low energy as facets of the product. The surfactant can stabilize a certain surface by “selective adhesion” and change the corresponding free energy, thus the growth rates of different crystallographic directions are varied and lead to self-assembly of low-dimensional products.²⁷ Consequently, during the thermally induced aggregation process, the introduction of capping agents (OA) promotes the anisotropic growth of Bi₂S₃ microstructures, resulting in the hierarchical architectures driven by the cooperative effect of the surfactants along with the crystal splitting growth mechanism.

Fig. 3 The shape evolution process of sheaf-like Bi₂S₃ hierarchical architectures (a)–(d) and corresponding schematic illustration of the proposed splitting mechanism (e). (a)–(d) Depict the time-dependent morphologies at 165, 170, 175 and 180 °C, respectively over the course of ~10 minutes.

3.3. The investigation of several reaction parameters

To further understand the formation mechanism of Bi₂S₃ by the facile one-step process, we systematically investigated a series of reaction parameters (initial added temperature, the ratio of precursors and the concentration of OA). We found that all these factors have significant influence on the morphologies, and the resulting products can be tuned from wheat-like to clover-like, dandelion-like and other novel architectures. We examined the effect of initial added temperature on the morphologies of as-prepared products. It was found that when the reaction was carried out at room temperature (∼27 °C), wheat-like Bi₂S₃ was produced as shown in Fig. 4a. With increasing temperature, chrysanthemum-like and dandelion-like shapes were obtained (Fig. 4b, c). From the FESEM images as presented in Fig. 4, one can see that the Bi₂S₃ microstructures firstly turn closely packed, and then loose with the rise of temperature. Moreover, we found that the ratio of precursors had a key impact on the morphologies of the products as well as the initial added temperature. We conducted the synthetic process by varying the ratio of Bi(Ac)₃ relative to element S from 2 : 1, 2 : 3 to 2 : 5 at 130 °C, while keeping all other parameters the same. It is interesting to note that when the ratio is 2 : 1, the clover-like Bi₂S₃ was obtained given in Fig. 4d. On the contrary, when the ratio is increased to 2 : 3 and 2 : 5, the flowers of Bi₂S₃ turn to cleave into loose structures which are composed of more filaments (Fig. 4e, f). From the images in Fig. 4d–f, one can see that the arm's orientation decreased with the increase of molar ratio of precursors. When the precursor ratios are increased, the concentration of monomers in the solution becomes higher in local regions, which provides a large supersaturation favoring the formation of Bi₂S₃ nuclei. Furthermore, the high concentration of S-complex around a Bi₂S₃ nucleus also favors kinetic crystal growth to an anisotropic structure and ultimately promotes branching.²⁸ Besides, the low concentration of precursors presumably leads to a slow diffusion rate of the reactant which facilities the formation of close-packed structures of Bi₂S₃. Compared to the initial added temperature and the molar ratio of precursors, the amount of oleic acid ligands also played an important role in the reaction on the morphologies of the as-obtained products. Upon increasing the concentration of oleic acid, the branched structures of Bi₂S₃ were decreased and the filaments tended to be closely packed shuttle-like architectures, as shown in Fig. 4g–i. These observations indicated that the well-known “limited ligand protection” (LLP)²⁹ strategy would be a reasonable formation mechanism of Bi₂S₃ in our synthesis. When 1 mL of OA was used as ligand for the Bi₂S₃ MCs, the number of coordinating ligands was lower, so that insufficient protection would result in close-packed chrysanthemum-like shape. However, the addition of a sufficient amount of OA, such as 3 mL and 5 mL, significantly enhanced the degree of ligand protection and stabilized the Bi₂S₃ hierarchical architectures in a dispersed form, which leads to the formation of loose-packed structures. Thus, with careful choice of these reaction parameters, it is easy to achieve a large-scale synthesis of shape-controlled monodisperse Bi₂S₃ hierarchical architectures.

Fig. 4 Size- and shape-controlled Bi₂S₃ hierarchical architectures upon changing the parameters in the synthetic system; (a)–(c) 27, 80 and 130 °C with 1 mL of OA and the ratio of 2 : 3 (Bi : S), respectively; (d)–(f) the ratio of 2 : 1, 2 : 3 and 2 : 5 (Bi : S) with 1 mL of OA at 130 °C, respectively; (g)–(i) 1, 3 and 5 mL of OA with the ratio of 2 : 3 at 80 °C, respectively.

3.4. Optical characterization

The optoelectronic properties of semiconducting materials are crucial for a large number of applications, while many unique electronic properties result from the band energy structures of semiconducting materials. To examine the optical properties of the obtained products, we performed absorption spectra in the range 300–1500 nm at room temperature. Fig. 5 shows the representative absorption spectrum of sheaf-like Bi₂S₃ hierarchical architectures dispersed in absolute tetrachloroethylene. The curve provides a good reading at the peak center of about 650 nm. It is obvious that the absorption edge of the Bi₂S₃ hierarchical architectures shows a blue shift compared with its bulk materials (1.3 eV).⁸ That could be attributed to the excitonic absorbance due to the quantum confinement effect at low dimension relative to its bulk counterparts.³⁰

Fig. 5 The typical absorption spectrum of Bi₂S₃ hierarchical architectures; the inset is the plot of (αhν)²versus hν according to Kubelka–Munk transformations.

Besides, it is well known that for direct band gap materials like Bi₂S₃, the relationship between the absorption coefficient α and optical band gap Eg for semiconductors obeys the following formula by performing Kubelka–Munk transformations:³¹

(αhν)² = A(hν − Eg) (1)

where A and hν are the edge-width parameter and the incident photon energy, respectively. Hence, the optical band gap Eg for the absorption peak can be determined with the x-axis intercept by extrapolating the linear portion of the (αhν)²versus hν plot to α = 0. Therefore, the fit of the linear portion corresponding to the direct band-gap transition and the optical band gap Eg of Bi₂S₃ is found to be ca. 1.9 eV according to the curve inset in Fig. 5, which is in general agreement with reported 1.85 eV band gap values of Bi₂S₃ nanowires.³²

3.5. Photoresponse properties

Photodetectorsor optical switches are indispensable elements in imaging techniques, lightwave communications, and possibly in future memory storage and optoelectronic circuits.³³ While conventional photodetectors are usually in film or bulk configurations, the unique and significant properties of micro- and nano-scale materials are expected to show better performances, such as enhanced optical gain and polarized lasing, and so forth.³⁴ In order to further evaluate the potential device applications with Bi₂S₃ MCs, the photoresponse properties of the wheat-like Bi₂S₃ architecture as a representative system were studied. To further investigate the device configuration in our system, we carried out the top-view and cross-sectional SEM characterization of a Bi₂S₃-based device, as shown in Fig. 6. The overall SEM images not only indicate that the surface of film was relatively uniform and smooth, but also reveal the close packing of the hierarchical Bi₂S₃ film on the ITO substrate. We observe that there are some holes and surface undulation, which may be attributed to the larger size of Bi₂S₃ MCs. Fig. S3 (ESI†) demonstrates the thickness and surface roughness characterization of hierarchical Bi₂S₃-based film performed by Veeco DEKTAK 150 surface profilometer. The profile manifests the average thickness of ∼1 μm and the surface roughness Ra of ∼200 nm. Fig. 7a shows the drastic increase of their photoconductivity with the elevation of voltage. A high photoexcited current of ∼46 μA was recorded at a low bias of 1.0 V and the slope of the I–V curve became much higher when the light was on, indicating an enhanced conductivity. A corresponding logarithmic plot (Fig. 7b) clearly shows that the photocurrent significantly increases by ca. 1 order of magnitude compared with the dark state, indicating high sensitivity. This was thought to result from the illumination promoting electrons from the valence band of Bi₂S₃ into the conduction band, increasing the charge carrier concentration via direct electron–hole pair creation and thus enhancing the conductivity.³⁵ It is worth noting that the photoresponsivity based on the Bi₂S₃ MCs has so far been found to be the best result in comparison with previous reports.^9,32 Interestingly, the I–V curves for both dark current and photocurrent exhibit an approximately linear shape with a bias from 0 to 1.0 V that indicates a good ohmic behavior. The schematic illustration of Bi₂S₃-based device fabrication used for photoelectrical measurements with a structure of ITO/Bi₂S₃/ITO is shown in the inset of Fig. 7a.

Fig. 6 Low- and high-magnification SEM images of Bi₂S₃ films obtained by the present solution process: top view (a and b) and cross-section (c and d).

Fig. 7 Photoresponsive sensitivity of the wheat-like Bi₂S₃ architectures as a representative system was studied. (a) The I–V characteristic of device based on Bi₂S₃ MCs in the dark and under simulated AM 1.5 illumination. (b) Logarithmic plot of (a) shows the large ratio of photo-excited to dark current. (c) Time dependence of the generated photocurrent of Bi₂S₃ MCs at a bias of 5 V in the dark and under illumination by halogen lamp. (d) The enlarged portions of the 799.5–802.5 s and 850.0–853.0 s ranges showing ultimately fast response and decay times. The inset in (a) illustrates the schematic of Bi₂S₃-based device fabrication used for photoelectrical measurements with a structure of ITO/Bi₂S₃/ITO.

The photoresponse as a function of time was measured with the light regularly chopped at a bias 5 V, which is shown in Fig. 7c. Upon exposure to the simulated light from the halogen lamp, the photocurrent rapidly increased to a maximum value of 28 μA, and then sharply returned to its initial one as soon as the light was turned off. The on–off sensing cycles can be repeated many times without any detectable degradation, thereby revealing its excellent stability and reproducible behavior. The response speed is a key parameter which determines the capability of a photodetector to follow a fast-varying optical signal. From the enlarged view (Fig. 7d) of one on–off cycle, the response and decay times (defined as the time required to reach 90% of the final equilibrium value) were estimated about 0.5 and 0.8 s, respectively. The characteristics of the photoconductive Bi₂S₃ MCs suggest that they are good candidates for light-sensitive materials or optoelectronic switches, with the light exposed conducting state as “ON” and the dark insulating state as “OFF”. Also it is worth noting that these Bi₂S₃ MCs, as well as the corresponding photoresponsive devices, were totally solution-processed, which may offer the potential for monolithic, low-cost and large-scale integration with CMOS electronics.

Conclusions

In summary, we demonstrated that the various size- and shape-controlled Bi₂S₃ hierarchical architectures were successfully prepared via a facile, low-cost, and environmentally-benign while effective one-step solution-based strategy. The morphological features of the resultant branched products can be tuned upon changing the parameters in the reaction system. By variation of the synthetic conditions, the branched Bi₂S₃ MCs from sheaf-like to clover, dandelion-like and other novel architectures were obtained in long-chain solvents 1-octadecene (ODE) in the presence of oleic acid (OA) as capping ligand. Thus with the careful choice of reaction parameters, it is possible to achieve a large-scale synthesis of shape-controlled monodisperse Bi₂S₃ microstructures. Furthermore, the crystal splitting behavior was believed to be responsible for the formation of Bi₂S₃ hierarchical architectures. It was demonstrated further that the as-prepared typical Bi₂S₃ microstructures possess a direct band gap of ∼1.9 eV ascribed to the quantum confinement effect. The photoresponse properties based on the sheaf-like Bi₂S₃ as a representative system were also investigated in an ambient environment. It is worth nothing that the photocurrent significantly increases by ∼1 order of magnitude compared with the dark state, indicating excellent sensitivity with respect to potential applications in photoelectric switch and light-sensitive devices. We believe that this one-step approach could be extended to the preparation of other metal chalcogenides and it is highly probable to provide new functional blocks for the design of novel devices.

Acknowledgements

This work was supported by NSFC (Nos. 21073071 and 51025206) and the National Basic Research Program of China (Nos. 2011CB808200 and 2007CB808000).

References

1. Y. Huang, X. Duan, Q. Wei and C. M. Lieber, Science, 2001, 291, 630–633.
2. C. Murray, C. Kagan and M. Bawendi, Annu. Rev. Mater. Sci., 2000, 30, 545–610.
3. Y. N. Xia, P. D. Yang, Y. G. Sun, Y. Y. Wu, B. Mayers, B. Gates, Y. D. Yin, F. Kim and H. Q. Yan, Adv. Mater., 2003, 15, 353–389.
4. O. Rabin, J. M. Perez, J. Grimm, G. Wojtkiewicz and R. Weissleder, Nat. Mater., 2006, 5, 118–122.
5. H. F. Bao, C. M. Li, X. Q. Cui, Y. Gan, Q. L. Song and J. Guo, Small, 2008, 4, 1125–1129.
6. J. M. Ma, Z. F. Liu, J. B. Lian, X. C. Duan, T. Kim, P. Peng, X. D. Liu, Q. Chen, G. Yao and W. J. Zheng, CrystEngComm, 2011, 13, 3072–3079.
7. K. Yao, W. W. Gong, Y. F. Hu, X. L. Liang, Q. Chen and L. M. Peng, J. Phys. Chem. C, 2008, 112, 8721–8724.
8. M. Nair and P. Nair, Semicond. Sci. Technol., 1990, 5, 1225–1230.
9. G. Konstantatos, L. Levina, J. Tang and E. H. Sargent, Nano Lett., 2008, 8, 4002–4006.
10. X. L. Yu and C. B. Cao, Cryst. Growth Des., 2008, 8, 3951–3955.
11. H. Bao, X. Cui, C. M. Li, Y. Gan, J. Zhang and J. Guo, J. Phys. Chem. C, 2007, 111, 12279–12283.
12. A. Mieszawska, R. Jalilian, G. Sumanasekera and F. Zamborini, Small, 2007, 3, 722–756.
13. L. S. Li, N. J. Sun, Y. Y. Huang, Y. Qin, N. N. Zhao, J. N. Gao, M. X. Li, H. H. Zhou and L. M. Qi, Adv. Funct. Mater., 2008, 18, 1194–1201.
14. J. Jiang, S. H. Yu, W. T. Yao, H. Ge and G. Z. Zhang, Chem. Mater., 2005, 17, 6094–6100.
15. Q. Y. Lu, F. Gao and S. Komarneni, J. Am. Chem. Soc., 2004, 126, 54–55.
16. Z. P. Liu, S. Peng, Q. Xie, Z. K. Hu, Y. Yang, S. Y. Zhang and Y. T. Qian, Adv. Mater., 2003, 15, 936–940.
17. L. Cademartiri, R. Malakooti, P. G. O'Brien, A. Migliori, S. Petrov, N. P. Kherani and G. A. Ozin, Angew. Chem., Int. Ed., 2008, 47, 3814–3817.
18. J. Tang and A. P. Alivisatos, Nano Lett., 2006, 6, 2701–2706.
19. C. X. Song, D. B. Wang, T. Yang and Z. S. Hu, CrystEngComm, 2011, 13, 3087–3092.
20. H. Wang, J. J. Zhu, J. M. Zhu and H. Y. Chen, J. Phys. Chem. B, 2002, 106, 3848–3854.
21. J. Waters, D. Crouch, J. Raftery and P. O'Brien, Chem. Mater., 2004, 16, 3289–3298.
22. T. Wu, X. G. Zhou, H. Zhang and X. H. Zhong, Nano Res., 2010, 3, 379–386.
23. Y. W. Koh, C. S. Lai, A. Y. Du, E. R. T. Tiekink and K. P. Loh, Chem. Mater., 2003, 15, 4544–4554.
24. D. P. Grigorev, Ontogeny of minerals, Israel Program for Scientific Translations, 1965.
25. D. F. Zhang, L. D. Sun, J. L. Yin and C. H. Yan, Adv. Mater., 2003, 15, 1022–1025.
26. Z. L. Wang, J. Phys. Chem. B, 2000, 104, 1153–1175.
27. L. W. Qian, J. Zhu, Z. Chen, Y. C. Gui, Q. Gong, Y. P. Yuan, J. T. Zai and X. F. Qian, Chem.–Eur. J., 2009, 15, 1233–1240.
28. H. B. Chu, X. M. Li, G. D. Chen, Z. Jin, Y. Zhang and Y. Li, Nano Res., 2008, 1, 213–220.
29. A. Narayanaswamy, H. F. Xu, N. Pradhan and X. G. Peng, Angew. Chem., Int. Ed., 2006, 45, 5361–5364.
30. M. Salavati-Niasari, D. Ghanbari and F. Davar, J. Alloys Compd., 2009, 488, 442–447.
31. A. Hagfeldt and M. Graetzel, Chem. Rev., 1995, 95, 49–68.
32. A. A. Tahir, M. A. Ehsan, M. Mazhar, K. G. U. Wijayantha, M. Zeller and A. D. Hunter, Chem. Mater., 2010, 22, 5084–5092.
33. H. Kind, H. Q. Yan, B. Messer, M. Law and P. D. Yang, Adv. Mater., 2002, 14, 158–160.
34. J. S. Jie, W. J. Zhang, Y. Jiang, X. M. Meng, Y. Q. Li and S. T. Lee, Nano Lett., 2006, 6, 1887–1892.
35. L. S. Li, R. G. Cao, Z. J. Wang, J. J. Li and L. M. Qi, J. Phys. Chem. C, 2009, 113, 18075–18081.

---

Footnote

† Electronic Supplementary Information (ESI) available. See DOI: 10.1039/c1ra00289a/

---