Journal of Colloid and Interface Science 395 (2013) 85–90
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Tunable synthesis of ordered Zinc Oxide nanoflower-like arrays Fang Xie ⇑, Anthony Centeno 1, Bin Zou, Mary P. Ryan, D. Jason Riley, Neil M. Alford Department of Materials and London Centre for Nanotechnology, Imperial College London, London SW7 2AZ, UK
a r t i c l e
i n f o
Article history: Received 25 October 2012 Accepted 17 December 2012 Available online 11 January 2013 Keywords: ZnO nanoflower-like array Pulsed laser deposition Hydrothermal growth
a b s t r a c t A novel approach to fabricate an ordered array of ZnO nanoflowers, consisting of uniform polymer cores of 100s of nanometer diameter decorated with ZnO nanorods of 10s of nanometer diameter, is presented. The 2-stage method combines the formation of ZnO seed layer by pulsed laser deposition (PLD) onto a colloidally assembled polystyrene sphere monolayer and the subsequent hydrothermal growth of ZnO nanowires (NWs). The main advantages of this methodology are low cost and the large area scalability of perfectly ordered hierarchical structures. More importantly, the process enables a versatile control of dimensions and morphologies of ZnO NWs as well as control of the core diameter by changing the polystyrene sphere diameter. A strong improvement of light scattering by such arrays is observed, offering promise as building blocks in different types of solar cells and potentially useful for a wide variety of applications in optoelectronic devices. Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction Semiconducting ZnO-based nanomaterials have attracted significant research interest due to their wide bandgap, excellent chemical and thermal stability, and their specific electrical and optoelectronic properties such as optical transparency, high electron affinity, and high electronic conductivity. ZnO nanowires have found a broad range of potential applications, ranging from UV nanolasers [1], dye-sensitized and photovoltaic solar cells [2–4], photodetectors [5] electroluminescent, field emission devices [6,7], and gas sensors [8]. In recent years, tremendous efforts have been devoted to the research of oriented and ordered arrays of ZnO nanorods/nanowires by various chemical, electrochemical, and physical deposition techniques, with average diameters typically ranging over an order of magnitude from 20 to 200 nm, and length from a few hundred nanometers to 10 lm, yielding an aspect ratio up to 500. The methods which have been particularly successful in creating highly oriented arrays of nanowires and/or nanorods include catalytic growth via the vapor–liquid–solid epitaxial (VSLE) mechanism [9], metal–organic chemical vapor deposition (MOCVD) [10], pulsed laser deposition [11], templating within anodic alumina membranes [12], electrodeposition [13], and hydrothermal growth [14]. However, the generation of hierarchical structures via integration of 1D nanoscale building blocks into 3D ⇑ Corresponding author. E-mail address:
[email protected] (F. Xie). Present address: Malaysia Japan International Institute of Technology, University Technology Malaysia, 54100 Kuala Lumpur, Malaysia. 1
0021-9797/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2012.12.028
architectures is a more interesting alternative than simple arrays of nanowires due to the higher specific surface and porosity, especially for applications in dye and semiconductor-sensitized solar cells [2]. So far, there are three methods that have been reported in the literature to synthesize ZnO ‘‘nanoflower’’ structures: Liu and Zeng [15] reported a wet-chemical route using a modified Kirkendall process; a second strategy [16] is based on the calcination of metallic Zn microsphere powders at relatively high temperature (500–750 °C). With these two approaches, ZnO nanoflowers are often randomly distributed in size, which limits their practical applications as building blocks in nanodevices. The third strategy [17] is based on electrochemical deposition using polystyrene as a template. However, because of the requirement of a conducting layer beneath the nanoflower arrays for electrochemical processing, such materials have limited applications. In this paper, we report a novel and simple method to fabricate a well-ordered nanoflower ZnO arrays with independently tunable nanowire and core dimensions. Moreover, such 3D nanostructures could be fabricated on a wide variety of substrates of interest including hydrophobic, hydrophilic, organic, and inorganic surfaces. The ability to create such flexible hierarchical structures onto virtually any surface paves the way for their wide applications in nanodevices. The method combines the formation of a polystyrene (PS) sphere monolayer and the hydrothermal growth of ZnO nanowires after ZnO seed layer by pulsed laser deposition (PLD). In addition to the enhanced surface area of this material, we show that the light scattering properties of these 3D arrays exceed those of 1D ZnO nanowire arrays, which is a significant property in the application for solar cells.
86
F. Xie et al. / Journal of Colloid and Interface Science 395 (2013) 85–90
2. Experimental section 2.1. Materials The following materials were purchased from Alfa Aesar and used as-received: hexamethylenetetramine (HMT, also termed methenamine, 99+%), zinc nitrate (Zn(NO3)2, 99%, ZnO powder (99%), ethanol (99+%), and acetone (99%). Polystyrene microspheres (10 wt.%) were purchased from Microparticles GmbH. Ptype silicon wafers, boron-doped (with resistivity of 1–5 X cm), were purchased from MMRC Inc. Nanopure water (>18.2 MX), purified using the Millipore Mili-Q gradient system, was used in all experiments. 2.2. Preparation of polystyrene monolayer using colloidal lithography (Fig. 1a) The sequential deposition process is shown schematically in Fig. 1. The polystyrene monolayers were prepared by using a modified colloidal lithography method, which is a self-assembly technique, described in detail in Ref. [17]. Briefly, monodisperse polystyrene (PS) particles with a diameter from 100 nm to 1000 nm were diluted by mixing with an equal amount of ethanol, respectively. Silicon substrates were cut into small pieces (10 by 10 or 5 by 5 mm2). Each square was thoroughly cleaned with acetone in an ultrasonic bath for 20 s and then rinsed in deionized water. Approximately 3–5 ll of the prepared PS solutions was applied onto the surface of a clean and large (30 by 20 mm) silicon wafer, which had been kept in 10% dodecylsodiumsulfate solution for 24 h previously. The wafer was then slowly immersed in a 15 cm glass vessel filled with 150 ml of Milli-Q water; the PS particles form a disordered monolayer on the water surface. To consolidate the particles, the water surface tension was changed by the addition of 4 ll of 2% dodecylsodiumsulfate solution, and a large monolayer with highly ordered areas was obtained. Such monolayers were then lifted off from the water surface using the silicon squares or glass substrates. A control sample without polystyrene monolayer was also fabricated by the same procedure, in order to generate free-standing ZnO nanowire array.
from ZnO powder, pressed, and sintered at 1000 °C under flowing oxygen for 10 h. Laser frequency, target-substrate distance and laser beam power density were fixed for all experiments at 8 Hz, 50 mm, and 0.85 J cm2, respectively. The chamber was evacuated to a background pressure of 3 105 torr, and the oxygen pressure was set at 50 103 torr. The substrates (in this case Si or glass with PS monolayer) were mounted onto a resistive heater using silver paint and were held at a fixed temperature of 50 °C (which is below the glass transition temperature of PS) for 30 min prior to deposition. For all the experiments, the number of laser pulses incident on the ZnO target was optimized at 200. After the deposition, the chamber was backfilled with 600 103 torr oxygen, and the samples were cooled to room temperature. 2.4. ZnO nanoflower-like array by hydrothermal growth (Fig. 1c) The ZnO film-coated Si/PS substrates were used as substrates for subsequent growth of nanoflower-like arrays, using modified hydrothermal method [14]. Briefly, 25 mM solutions of zinc nitrate and HMT in water were prepared and then maintained at 90 °C in a thermostatically controlled oil bath for 30 min. 50 ml aliquots of each solution were then mixed in another glass bottle, and the preheated substrate of interest was then immersed in the resulting reactive mixture solution. The bottle was sealed and held at constant temperature of 90 °C for 2 h, without stirring, throughout the growth process. The as-grown samples were then rinsed in deionized water and dried in air. 2.5. Characterization Scanning electron microscopy (SEM) images of the Au coated polystyrene monolayers and ZnO nanoflower-like arrays were collected using a LEO Gemini 1525 field emission gun scanning electron microscope (FEG-SEM), equipped with a Genesis 4000 EDAX for collection of energy dispersive X-ray spectra of the samples. The optical reflectivity (R) was measured at room temperature with a Bentham PVE300 PV characterization system fitted with an integrating sphere, from 300 nm to 1100 nm. Transmission electron microscopic (TEM) images were recorded by an FEI 80–300 kV Titan operated at 300 kV. It is equipped with a monochromator. The spatial resolution of the microscope is 0.3.
2.3. ZnO seed layer by pulsed laser deposition (Fig. 1b) 3. Results and discussion The ZnO seed layer was grown by PLD, using apparatus and methods that have been described elsewhere [18]. In summary, a KrF laser with wavelength 248 nm was used with a pulse duration of 25 ns. A ZnO target with a diameter of 22 mm was prepared
Self-assembled polystyrene colloidal crystals have been widely used as effective intermediate for the fabrication of nanostructure arrays due to several outstanding advantages, such as large area,
Fig. 1. Schematic view of the fabrication process of ZnO NW nanoflower-like arrays. (a) PS sphere monolayer was deposited by colloidal lithography onto a Si substrate; (b) ZnO thin film deposition on PS spheres by PLD; (c) hydrothermal growth of ZnO NWs onto the spheres to produce nanoflower morphology.
F. Xie et al. / Journal of Colloid and Interface Science 395 (2013) 85–90
low cost, multi-scale, and flexible tuning parameters. More importantly, the final product of the nanostructure arrays can be further adjusted through the combination of colloidal lithography and subsequent growth methods. Such flexibility makes it a very versatile method for preparation of desired nanostructures. There are several colloidal crystallization strategies that have been welldocumented, including drop coating [19], dip-coating [20], spincoating [21], and nano-robotic manipulation [22]. In this paper, a monolayer of commercially available polystyrene spheres with diameter of 1 lm was deposited directly on a silicon surface by using a self-assembly technique on a water surface, a modified method [17] for forming large area and highly hexagonal ordered polystyrene monolayer. The transfer of ordered polystyrene array to the water surface enables various substrates including hydrophobic surfaces to be used for subsequently ‘‘fishing’’ the monolayer. Fig. 2a shows a top view SEM image of a typical polystyrene monolayer in a hexagonal (hcp) array. A wellorganized monolayer of PS microspheres can be observed in addition to occasional ‘‘line’’ defects in some regions due to grain boundary effects. Such a well-ordered layer was observed across the whole Si surface of 1 cm 1 cm, as demonstrated in Fig. 2c, the sample color is perfectly homogeneous, reflecting the presence of only one PS domain on the substrate. The detailed organization of the spheres was investigated using higher magnification SEM (Fig. 2b), which shows a relatively large area and perfectly ordered polystyrene spheres. The ZnO seed layer was grown by PLD. It has been reported that a 10–15 nm seed layer of ZnO, either from ZnO quantum dot by dip-coating [23] or by PLD deposition [18], will provide nucleation sites during the hydrothermal growth of ZnO nanowires. The precursor PLD film consists of many nanocrystallites, with a preferred c-axis alignment [24], which hence serves to nucleate the hydrothermal growth of c-axis aligned ZnO nanowires. It has been reported [25] that ZnO targets become Zn rich during PLD, owing to the loss of oxygen under vacuum, in addition to backscattering of Zn rich particles onto the target surface. The inclusion of oxygen in the chamber is necessary, so that additional oxygen could be incorporated into the depositing film ensuring correct stoichiometry of the thin seed layer. Hydrothermal growth of ZnO NWs was pioneered by Vayssieres [24]. Key reactions in the formation of ZnO NWs arrays are the thermal decomposition of HMT to formaldehyde and ammonia, with the latter acting as a base in aqueous solution. The reaction proceeds as:
Fig. 2. SEM images of a self-assembled monolayer of polystyrene spheres (1 lm in diameter) on a silicon substrate. (a) a top view of PS monolayer on Si substrate; (b) a higher magnification of PS monolayer; (c) a digital picture of the prepared sample.
C6 H12 N4 þ 10H2 O $ 6HCHO þ 4NHþ4 þ 4OH
87
ð1Þ
followed by the pH-driven precipitation of Zn(OH)2 or ZnO:
Zn2þ þ 2OH $ ZnO þ H2 O
ð2Þ
or
Zn2þ þ 2OH $ ZnðOHÞ2 þ ZnO þ H2 O
ð3Þ
Fig. 3a shows a top view SEM image of the ZnO nanowires uniformly assembled into an array, which has the original hcp arrangement of PS sphere template. All PS spheres which are covered by ZnO nanowires exhibit the characteristic nanoflower-like shape. They are close-packed, comprise a high density of hexagonal NWs grown perpendicular to the original PS surface, and exhibit a much higher developed surface area. Fig. 3b shows the structure of the individual ‘‘nanoflower’’ at higher magnification. The diameters of PS sphere cores are uniform with average values of 950 ± 11 nm. The diameters of the NWs range from 50 nm to 120 nm with
Fig. 3. SEM images of ZnO nanoflower-like arrays on PS monolayer. (a) a top view of the array; (b) a higher magnification of individual ZnO nanoflower titled from its original position; (c) a single ZnO with hexagonal cross section; (d and e) higher magnification SEM image of ZnO nanowires interlaced between the two closepacked PS; (PS diameter: 950 nm; ZnO NWs have diameter of 100 nm and aspect ratio of 6).
88
F. Xie et al. / Journal of Colloid and Interface Science 395 (2013) 85–90
Fig. 4. (a) TEM image of ZnO nanowires scratched from a nanoflower-like structure; (b) HRTEM image shows lattice fringes with inter-planar spacing of 0.52 nm.
Fig. 5. EDX spectra of ZnO (a) upper part of PS after PLD deposition and hydrothermal growth; (b) upper part of PS after PLD deposition; (c) lower part of PS after hydrothermal growth.
average values of 100 nm. The average length of the NWs is 600 ± 34 nm under our experimental condition of 2 h hydrothermal growth. The circled region in Fig. 3d and e shows that the ZnO nanowires between the two closely packed spheres become interlaced, and we expect that growth in this direction would become limited for longer deposition times. ZnO NWs assembled on PS surface retain columnar growth and exhibit near-perfect hexagonal cross sections, as shown in Fig. 3c. The basal seed layer particles are presumed to evolve from nucleation on the PS surface and to guide the subsequent formation of the orthogonal NWs, including their polarities and morphologies. The detailed morphology and crystalline structure of the ZnO has been studied by transmission electron microscopy (TEM). Fig. 4 shows a high-resolution TEM image of ZnO NWs from the nanoflower-like structure. Crystal planes aligned perpendicularly to the growth direction are clearly visible. The measured inter-plane spacing is 0.52 nm, which matches well with the literature value for (0 0 0 1) planar separation in the ZnO wurtzite crystal, showing that the NWs grow along the [0 0 0 1] direction. The morphology of ZnO nanostructures formed by hydrothermal methods can depend sensitively upon the growth time, because of the evolving concentrations of OH, Zn2+, and HMT in
the reactive solution [14]. The final morphology of ZnO nanostructures is extremely sensitive to the local solution-phase composition when working with low reactant concentrations. Under such conditions, even the termination of the polar surface of ZnO NWS can affect the product morphology by influencing the ratio of Zn2+/OH in the double layer at the growing polar surface. Previous studies [27,28] demonstrated ZnO NWs hydrothermally grown on PLD films to be Zn-polar and Zn-ion terminated instead of O-ion terminated. Zn-ion termination causes a reduction in the local Zn2+/OH ratios relative to those in bulk solution, thereby encouraging tapered NWs growth and, as the Zn+ concentration falls further, the emergence of volcano-like structures on the polar surface which seed the subsequent growth of ZnO nanotubes. It has also been demonstrated that ZnO NWs grown on thin PLD ZnO films at early times (t < 3 h) have perfect hexagonal morphology, while nanotubes were seen to emerge at extended growth time (t > 10 h) [14]. Under experiment conditions set herein, that is the reaction time of 2 h, we only observed ZnO NWs with hexagonal morphology. By changing the hydrothermal growth conditions such as the initial concentration of HMT and Zn2+ZnO nanotubes will readily form, which makes this approach even more versatile. The requirement for the PLD seed layer was demonstrated by investigating Si substrate with PS monolayer without the PLD ZnO. We observed no ZnO NWs growth under the same hydrothermal experimental conditions. This clearly indicates that the nucleation sites supplied by the PLD film are crucial for formation of ZnO NWs. The precursor PLD film consists of many nanocrystallites, with a preferred c-axis alignment [26], which served to nucleate the hydrothermal growth of c-axis aligned NWs. This is in agreement with previous literature [14]. Fig. 3b shows an individual nanoflower that has been lifted and tilted relative to its original position; it clearly shows that ZnO nanowires only grow on the upper half of the PS sphere, where the seed layer has been deposited. As described above and illustrated in Fig. 1, the key point for the formation of the ZnO nanoflower structure is the ZnO seed layer by PLD. The absence of ZnO from the lower parts was confirmed by energy dispersive Xray (EDX) analysis carried out on PS upper surface (Fig. 5a and b) and lower part (Fig. 5c). The weak Zn peak (Fig. 5b) on the upper surface of PS indeed proved the existence of ZnO thin film on PS after PLD. With hydrothermal growth of ZnO nanoflower-like array, it shows that Zn peak intensity enhanced dramatically in EDX spectra (Fig. 5a). In order to validate that such a hierarchical nanostructure system is indeed tunable, we demonstrate that both variations in sphere and rod dimensions are readily achieved. For example, Fig. 6 shows ZnO nanoflower-like arrays with core size of 500 nm
F. Xie et al. / Journal of Colloid and Interface Science 395 (2013) 85–90
89
Fig. 6. SEM images of ZnO nanoflower-like arrays on PS monolayers. (a) A top view of the array; and (b) a higher magnification of individual ZnO nanoflower (PS diameter: 500 nm; ZnO NWs have diameter of 50 nm and aspect ratio of 8).
Fig. 7. Total reflectance spectra and diffuse reflectance spectra for free-standing (black and gray) and nanoflower (green and purple) ZnO nanowires on glass substrate. The dimensions of nanowires are similar for both samples. Each inset SEM image corresponds to its spectra below. (The line on the spectra indicates 420 nm). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
and ZnO nanowire of approximately 50 nm in diameter with aspect ratio of 8, made by changing the size of PS spheres and reducing the reactant concentrations for hydrothermal growth. This clearly demonstrates that the process enables a versatile control of dimensions and morphologies proving its flexibility and tunability in a controlled way. In addition, we have used this approach for depositing onto another substrate, for example, glass (data are not shown). The method established has enabled the production of a wide range of ordered nanoflower-like arrays of ZnO on relatively large surfaces of any substrates including conducting, dielectric, and semi conducting surfaces. We expect that these structures will find many applications in devices where increased scattering and reduced optical reflection would be beneficial [27–30,3]. To demonstrate the scattering properties of nanoflower-like ZnO arrays, we investigated the geometric influence on solar light scattering by measuring total and diffuse reflectance spectra of the free-standing nanowires and the nanoflower-like ZnO array on glass substrates, respectively. The results are shown in Fig. 7. All spectra show UV absorbance edges close to 370 nm (Eg = 3.37 eV), corresponding to the band-gap energy of bulk ZnO. An increase (more than 180% and 300% at 420 nm for total and diffuse reflectance, respectively) in the reflectance was observed throughout the entire visible wavelength range for the
nanoflower-like arrays, compared to the free-standing arrays. It can be seen from Fig. 7 that there is a relatively small difference between the total and the diffuse scattering for the nanoflower-like ZnO arrays. The ZnO nanowire density on nanoflower array and free-standing array can be estimated based on SEM images. The densities of ZnO nanowires are as follows: 53.82 ± 2.76 rods lm2 for nanoforest-like array; and 48.73 ± 3.47 rods lm2 for nanoflower-like array. Hence, the observed increase in scattering cannot be accounted for by an increased surface area. To ensure there is no impact on the optical properties from the PS template, we performed characterization of the self-assembled template prior to ZnO deposition (data are not shown). The results showed a negligible contribution to both total and diffuse reflectance. This implies that the scattering is predominantly diffuse in nature, extremely important for improving light trapping in photovoltaic devices and increased light conversion efficiency in dye-sensitized solar cells. To provide a better insight into the increased diffuse light scattering by ZnO nanoflower-like arrays, the ratio of diffuse to total reflectance at 420 nm for the free-standing NW array and the nanoflower-like array were calculated to be 91% and 57%, respectively, indicating a significant improvement in light trapping and scattering abilities of nanoflower-like arrays. This could be explained by the difficulty of increasing light scattering using vertically-aligned ZnO NWs due to the limited surface area, periodic spacing of less than a wavelength and high verticality. It is interesting to note that despite not being optimized, the ZnO nanoflowerlike array measured here shows a significant increase in scattering which, compared to the free-standing array, is mostly diffuse. The ZnO nanoflowers are arranged in an array which lacks periodicity and so can be considered quasi-random. In fact, diffuse scattering is a measure of the randomness of the scattering surface. With an aligned array, the periodicity is between adjacent nanowires, causing scattering from the surface to constructively interfere. When measured some distance away the scattering will appear specula. Future work will use numerical analysis to suggest optimized dimensions of nanoflowers for both reduced reflection and increased scattering.
4. Conclusion In summary, we report on a novel approach to fabricate hierarchical ZnO nanoflower-like arrays. The method combines the formation of ZnO seed layer by pulsed laser deposition (PLD) on a polystyrene sphere monolayer and the hydrothermal growth of ZnO nanowires. This is a simple way to integrate 1D nanoscale building blocks into 3D architectures. The synthesis enables the generation of large-scale, low cost and importantly, homogeneous
90
F. Xie et al. / Journal of Colloid and Interface Science 395 (2013) 85–90
functional metal oxide nanostructures with controlled complexity. The ZnO nanoflower-like arrays combine the properties of 3D and 1D materials and are thus a more interesting alternative than simple arrays of ZnO NWs due to the higher specific surface area, especially for application in dye and semiconductor-sensitized solar cells. The process enables a versatile control of dimensions and morphologies of ZnO NWs as well as control of the core diameter by changing the size of polystyrene sphere diameter. The preliminary experimental results have confirmed a strong improvement of light scattering by such arrays, offering promising applications as building blocks in different type of solar cells and potentially useful for a wide variety of applications in optical devices, photonic crystals, and nanodevices. References [1] M.H. A Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo, P. Yang, Science 292 (2001) 1897. [2] I. Gonzalex-Valls, M. Lira-Cantu, Energy Environ. Sci. 2 (2009) 19. [3] Y.J. Lee, D.S. Ruby, D.W. Peters, B.B. McKenzie, J.W.P. Hsu, Nano Lett. 8 (2008) 1501. [4] Y. Liu, A. Das, S. Xu, Z. Lin, C. Xu, Z.L. Wang, A. Rohatgi, C.P. Wong, Adv. Energy Mater. 2 (2012) 47. [5] H. Kind, H. Yang, B. Messer, M. Law, P. Yang, Adv. Mater. 14 (2002) 158. [6] W.I. Park, G.C. Yi, Adv. Mater. 16 (2004) 87. [7] Y.K. Tseng, C.J. Huang, H.M. Cheng, I.N. Lin, K.S. Liu, L.C. Chen, Adv. Funct. Mater. 13 (2003) 811.
[8] J.B.K. Law, J.T.L. Thong, Appl. Phys. Lett. 88 (2006) 133114. [9] M. Huang, Y. Wu, H. Feick, N. Tran, E. Weber, P. Yang, Adv. Mater. 13 (2001) 113. [10] S. Liu, J.J. Wu, Mater. Res. Soc. Symp. Proc. 703 (2002) 241. [11] J.H. Choi, H. Tabata, T.J. Kawai, Cryst. Growth 226 (2001) 493. [12] M. Lai, D.J. Riley, Chem. Mater. 18 (2006) 2233. [13] R. Liu, A.A. Vertegel, E.W. Bohannan, T.A. Sorenson, J.A. Switzer, Chem. Mater. 13 (2001) 508. [14] Y. Sun, G.M. Fuge, N.A. Fox, D.J. Riley, M.N.R. Ashfold, Adv. Mater. 17 (2005) 2477. [15] B. Liu, H.C. Zeng, J. Am. Chem. Soc. 126 (2004) 16744. [16] Z. Gu, M.P. Paranthaman, J. Xu, Z.W. Pan, ACS Nano 3 (2009) 273. [17] J. Rybczynski, U. Ebels, M. Giersig, Eng. Aspects 219 (2003) 1. [18] C.F. Zhang, Z.W. Dong, K.J. Liu, Y.L. Yan, S.X. Qian, H. Deng, Appl. Phys. Lett. 91 (2007) 142109. [19] P. Jiang, J.F. Bertone, K.S. Hwang, V.I. Colvin, Chem. Mater. 11 (1999) 2123. [20] Z.Z. Gu, A. Fujishima, O. Sato, Chem. Mater. 13 (2002) 760. [21] D. Wang, H. Mohwald, Adv. Mater. 16 (2004) 244. [22] F. Garcia-Santamaria, H.T. Miyazaki, A. Urquia, M. Ibisate, M. Belmonte, N. Shinya, F. Meseguer, C. Lopez, Adv. Mater. 14 (2002) 1144. [23] M. Law, L.E. Greene, J.C. Johnson, R. Saykally, P. Yang, Nat. Mater. 4 (2005) 455. [24] L. Vayssieres, Adv. Mater. 15 (2003) 464. [25] J.B. Franklin, B. Zou, D.W. McComb, P. Petrov, M.P. Ryan, M.A. McLachlan, J. Mater. Chem. 21 (2011) 8178. [26] S.J. Henley, M.N.R. Ashfold, D.P. Nichols, P. Wheatley, D. Cherns, Appl. Phys. A 79 (2004) 1169. [27] H.A. Atwater, A. Polman, Nat. Mater. 9 (2010) 205. [28] A. Centeno, F. Xie, N. Alford, JOSA B 28 (2011) 325. [29] J.Y. Chen, K.W. Sun, Sol. Energy Mater. Sol. Cell 94 (2010) 930. [30] Z. Jehl, J. Rousset, F. Donsanti, G. Renou, N. Naghavi, D. Lincot, Nanotechnology 21 (2010) 395603.