Solution-Processed Zinc Oxide Field-Effect Transistors Based on Self

OTFTs with mobilities of 0.01-0.1 cm2/Vs, good reliability, stability, and ..... the TFTs made from nanosphere films subjected to hydro- thermal growt...
0 downloads 0 Views 486KB Size
NANO LETTERS

Solution-Processed Zinc Oxide Field-Effect Transistors Based on Self-Assembly of Colloidal Nanorods

2005 Vol. 5, No. 12 2408-2413

Baoquan Sun* and Henning Sirringhaus CaVendish Laboratory, UniVersity of Cambridge, Madingley Road, Cambridge, CB3 0HE, U.K. Received August 11, 2005; Revised Manuscript Received October 23, 2005

ABSTRACT Colloidal zinc oxide (ZnO) nanocrystals are attractive candidates for a low-temperature and solution-processible semiconductor for highperformance thin-film field-effect transistors (TFTs). Here we show that by controlling the shape of the nanocrystals from spheres to rods the semiconducting properties of spin-coated ZnO films can be much improved as a result of increasing particle size and self-alignment of the nanorods along the substrate. Postdeposition hydrothermal growth in an aqueous zinc ion solution has been found to further enhance grain size and connectivity and improve device performance. TFT devices made from 65-nm-long and 10-nm-wide nanorods deposited by spin coating have been fabricated at moderate temperatures of 230 °C with mobilities of 0.61 cm2V-1s-1 and on/off ratios of 3 × 105 after postdeposition growth, which is comparable to the characteristics of TFTs fabricated by traditional sputtering methods.

There is currently significant interest in realizing highperformance thin-film transistors (TFTs) based on solutionprocessible semiconducting materials for applications requiring low-cost, low-temperature manufacturing on largearea flexible substrates.1,2 Much effort has been devoted to low-temperature solution processible organic semiconductors as a potential alternative to traditional inorganic semiconductors. OTFTs with mobilities of 0.01-0.1 cm2/Vs, good reliability, stability, and device-to-device uniformity have been demonstrated.3-8 There are also various approaches to realizing solution-processible inorganic semiconductors, which provide a potential route to significantly higher mobilities, but for which control of electronic defect states when processed from solution at low temperatures can be challenging.9 Inorganic semiconductors might also provide a route to high performance n-type TFTs required for complementary circuits, which are traditionally difficult to realize with organic TFTs, although much progress has been made recently.10,11 A variety of solution-processible inorganic semiconductors for TFTs have been reported.9 These include tin(II) iodide based organic-inorganic hybrids,12 chalcogenide semiconductors,13 semiconductor nanowires,14 and nanocrystals.15 The semiconductor nanowire/nanocrystal approach is very promising because it allows one to decouple the high-temperature growth/synthesis of the nanowire from the low-temperature device fabrication process and achieve high performance. However, it is a tough challenge to disperse micrometer* Corresponding author. E-mail: [email protected]. 10.1021/nl051586w CCC: $30.25 Published on Web 11/25/2005

© 2005 American Chemical Society

long nanowires in a solution for simple solution-coating or printing-based deposition. This problem is easier to solve with smaller colloidal nanocrystals, such as CdSe nanocrystals, which can be drop-cast onto a substrate and can remelt to form a uniform film after annealing at 350 °C because of lowering of the melting point for these ultra-small nanocrystals.16 However, the mobility of a thin film of nanocrystals is significantly lower than the maximum achievable bulk mobility of the semiconducting material because of grain boundaries in the sintered nanocrystal network. As-prepared CdSe nanocrystal devices show n-type behavior with a mobility of 1 cm2V-1s-1 and an on/off ratio of 3.1 × 104. Some of the important requirements for using semiconductor colloidal nanocrystals in this application include a good dispersing capacity (>50 mg/mL) and adequate stability of the dispersion (at least one week). Although there are many reports to synthesize different kinds of nanocrystals, there are few colloid nanocrystal systems that can meet these requirements.17 Zinc oxide (ZnO) is an environmentally friendly transparent semiconductor with a large band gap of 3.37 eV. TFT devices based on polycrystalline ZnO as active layer have been reported with mobilities of around 0.2-3 cm2V-1s-1.18-21 Most fabrication methods use a sputtering process to grow ZnO films. Solution-processing techniques have also been used to fabricate ZnO devices but have suffered from poor device performance20 or the need to use a high annealing temperature (700 °C).21 Here we use spin-coated ZnO nanocrystals films to fabricate high-performance TFT devices

at moderately low temperatures. ZnO nanospheres can be dispersed at high concentration beyond 75 mg/mL for solar cell applications as shown in recent reports.22,23 Furthermore, the shape of the nanocrystals can be controlled from nanosphere to nanorods by adjusting the growth time.24 It is possible to achieve a nanorod suspension with a concentration of up to 85 mg/mL with good stability by adding a small amount of alkylamine. We investigate the role of the shape of the nanocrystal on the colloidal self-assembly of the nanoparticles on the substrate and on the resulting device performance. We also show that the TFT device performance and mobility can be further improved without sacrificing the on-off ratio by simple hydrothermal growth in an open beaker with aqueous zinc ion solution to increase the domain size and minimize the effect of grain boundaries. Experimental Section. ZnO nanorods are prepared according to a literature method developed by Pacholski23,24 with some modification. Zinc acetate (Zn(Ac)2, 0.8182 g, 4.46 mmol) and 250 µL of water was added into a flask containing 42 mL of methanol. The solution was heated to 60 °C with magnetic stirring. Potassium hydroxide (KOH, 0.4859 g, 7.22 mmol, purity 85%) was dissolved into 23 mL of methanol as the stock solution that is dropped into the flask within 10-15 min. At a constant temperature of 60 °C, it takes 2 h and 15 min to obtain 6-nm-diameter nanospheres. A small amount of water was found helpful to increase the ZnO nanocrystal growth rate. To grow the nanorods, the solution is condensed to about 10 mL. This was found helpful before further heating to decrease the growth time of the nanorods. Then it is reheated for another 5 h before stopping the heating and stirring. The upper fraction of the solution is removed after 30 min. Methanol (50 mL) is added to the solution and stirred for 5 min. The upper fraction of the solution is discarded again after 30 min. This process is repeated twice. For the second washing, the upper fraction of the solution is taken away after overnight staying. Finally, 3.3 mL of chloroform and 100 µL of n-butylamine are used to disperse the nanorods. The nanorod concentration is about 85 mg/mL, and the suspension is stable for more than 2 weeks. Using the modified method reported here, it takes only 5 h to obtain 65-nm-long nanorods instead of several days as reported in the literature.24 Nanocrystal films and devices are fabricated on SiO2(300 nm)/Si substrates with photolithographically patterned interdigitated Cr(3 nm)/Au(12 nm) electrodes. The device structure is shown in Scheme 1. Before spin-casting the ZnO solution, the substrate is cleaned in an oxygen plasma at a power of 150 W for 2 min. The film is spin-coated from the filtered (0.45-µm PTFE filter) ZnO solution with a speed of 2000 rpm. Then the devices are annealed at 230 °C in N2/H2(V/V, 95:5) for 30 min. For additional hydrothermal growth, the substrates are immersed upside down into a glass beaker filled with an aqueous solution containing zinc nitrate (0.025 M) and ethyldiamine (0.04 M) with slow stirring at 90 °C. The devices are taken out after 50 min and rinsed with deionized water. Finally, the device is annealed at 200 °C for 15 min in an N2/H2 atmosphere after drying with nitrogen. Nano Lett., Vol. 5, No. 12, 2005

Scheme 1.

TFT Device Structure

Results and Discussions. Transmission electron microscopy images of ZnO nanospheres (a) and nanorods (b) synthesized as above are shown in Figure 1. The diameter of the nanospheres is about 6 nm. The nanorods have an average width of 10 nm and length of 65 nm. The nanorod length can be tuned by the reaction time. However, long nanorods (longer than 100 nm) are quite difficult to disperse into any solution. The UV-Vis absorption spectrum of nanospheres and nanorods are shown in Figure 1S. The absorption peak of the nanorods shifts to a longer wavelength, reflecting the larger diameter of nanorods compared to that of nanospheres. In the synthesis process, it is critical to have the correct mole ratio between KOH and Zn(Ac)2. The chemical composition of as-prepared nanorods is determined by the initial mole ratio. Variations in stoichiometry affect the conductivity of the films and the mobility and on-off current ratio of the TFTs. The characteristics for three TFT devices made from ZnO nanorods synthesized with different mole ratios are summarized in Table 1. All of the devices exhibit n-type field-effect conduction. The optimized mole ratio is 1.62. It is found that the conductivity increases and the mobility decreases as the stoichiometry is varied from the experimentally determined optimum mole ratio in both directions. The stoichiometry can be characterized by X-ray diffraction. The (002) diffraction signal of the ZnO nanocrystals comes only from zinc atoms in the wurtzite crystal structure. It has been found that the (002) signal of the ZnO nanocrystals in their X-ray diffraction patterns is maximum if the mole ratio is near to its stoichiometric value,25 which means that there will be the lowest concentration of oxygen vacancies in this crystal structure at this ratio. ZnO films containing a low concentration of oxygen vacancies should exhibit low conductivity because oxygen vacancies behave as deep donors.26 Consistent with this expectation, TFT devices based on this ratio show the lowest conductivity. It is worth mentioning that small variations in the mole ratio do not appear to have a significant effect on the shape and size of the nanorods but do affect the TFT device performance greatly. We believe that the large difference of the TFTs’ characteristics originates in small changes of the stoichiometry of the ZnO films. 2409

Figure 1. Transmission electron microscopy images of ZnO nanocrystals. (a) nanosphere with average diameter of 6 nm. (b) Nanorods with average lengths of 65 nm and diameters of 10 nm. Table 1. TFT Device Characteristics of As-Deposited ZnO Nanorod Films, Which Were Synthesized from Different Mole Ratios between Potassium Hydroxide and Zinc Acetatea film a b c

mole ratio KOH/Zn(Ac)2

µsat (cm2V-1s-1)

on/off ratio

|V0| (V)

1.5 1.62 1.7

1.9e-4

50 mg/mL). The as-prepared ZnO nanocrystals comprise acetate (CH3COO-) ligand groups chelating with zinc atoms on the surface of nanocrystals. The ligands are very important to facilitate the dispersion of the nanocrystals in the solvent. For small nanospheres (approximately 6 nm), it is quite easy to achieve a high-concentration suspension using just the short acetate ligands. However, the acetate ligands are not sufficient to achieve high-concentration dispersions of the longer nanorods. Long-chain alkylamines (dodecylamine) have been used as ligands to help ZnO nanocrystal suspension.27 Here we use butylamine as a ligand with a shorter chain and a low boiling point (78 °C) instead of dodecylamine. When butylamine is added to the suspension, the 10 × 65 nm2 nanorods can be dispersed into chloroform with concentration as high as 90 mg/mL. At the same time, the ligand can increase the interaction between nanocrystals and favor formation of a uniform film.28 For the nanospheres, there is a large number of microcracks in the spin-cast films if butylamine is not added to the suspension. Butylamine was used in all of the spin-cast films reported here unless stated otherwise. Furthermore, residual ligands can be removed easily by annealing because of their low boiling points. It is found that most ligands composed of mixtures of acetate groups and butylamine groups can be removed easily by annealing. Fourier transform infrared spectroscopy (FTIR) was used to check the presence of the ligand as shown 2410

in Figure 2S. We do not detect a signal from butylamine, probably because most of the butylamine volatilizes in a short time during sample preparation and the small amount of butylamine on the surface is difficult to detect by FTIR because of the spectral overlap between the characteristic amine absorption and that of residual water. The presence of the acetate groups can be checked by observing the -CdO stretching vibrations in the range of 1600-1300 cm-1 and the -CH3 stretching vibrations in the range of 30002800 cm-1. Both signals disappear after annealing if butylamine has been added as a ligand. In the presence of butylamine, the chelating acetate groups are dissociated from the surface after being replaced with butylamine. They remain present in the film after spin-coating (Figure 2S.II), but can be removed easily during annealing (Figure 2S.IV). Without butylamine, the acetate groups, which remain chelated on the surface of ZnO nanocrystals after spincoating, are difficult to remove by annealing (Figure 2S.III). The as-prepared films are annealed under nitrogen/hydrogen atmosphere to increase the mobile carrier concentration and field-effect mobility. It has been reported that hydrogen can be incorporated into ZnO films in high concentration at annealing temperatures of 200 °C and behave as a shallow donor acting as a source of conductivity.26 The characteristics of TFTs made from ZnO nanospheres (a) and nanorods (b) are shown in Figure 2. Both TFT devices exhibit clean n-type transistor behaviors with low turn-on voltages |V0| ) 0-8 V and good operating stabilities. For the device made from 6-nm nanospheres, the on-off ratio is 5 × 103 and the linear and saturated fieldeffect mobilities are 2.37 × 10-4 cm2V-1s-1 and 4.62 × 10-4 cm2V-1s-1, respectively. The TFT device performance is improved significantly when nanorods are used as the active layer instead of nanospheres. These devices exhibit an on-off ratio of 1.1 × 105 and higher mobility of 0.023 cm2V-1s-1 derived from the saturated operating region and 0.013 cm2V-1s-1 derived from the linear region. The mobility is improved by almost two orders from devices made from 6-nm nanospheres to those from 65-nm-long nanorods. The approximately 10x larger size of the nanorod particles Nano Lett., Vol. 5, No. 12, 2005

Figure 2. Log-linear scale plots of linear (Vd ) 5 V) and saturated(Vd ) 60 V) transfer characteristics for as made TFT device with (a) 6-nm nanospheres and (b) 10 × 65 nm2 nanorods without postdeposition hydrothermal growth measured in nitrogen atmosphere. The devices have been annealed at a temperature of 230 °C before measuring. The channel length (L) and width (W) are 20 µm and 1 cm, respectively. The capacitance value of the gate dielectric is 11.4 nF/cm2.

compared to the nanospheres will significantly reduce the number of interparticle hopping events that an electron has to undergo when moving from source to drain electrode. This will result in an increase of the mobility even if the nanorods are not uniaxially aligned along the direction of current flow. If we assumed somewhat simplistically that the conduction process involves many electron hopping steps from one nanocrystal to its neighbors, then for a channel length of 20 µm and 6-nm-wide nanospheres an electron requires at least 3000 internanocrystal hops to transfer from one electrode to the other. In contrast, for 10 × 65 nm2 nanorods, the number of hops would be reduced to about 500 (assuming an isotropic orientation of the nanorods in the plane). Therefore, the electron hopping steps to cross the channel should be much less in the nanorod TFT than in the in the nanosphere device. However, another important reason for the improved performance of the nanorod device is believed to be related to the favorable in-plane self-alignment of neighboring colloidal nanorods with respect to each other when spincoated onto the substrate as discussed below (see Figure 4a). The TFT device performances can be further enhanced by the postdeposition hydrothermal growth step in aqueous solution. The corresponding TFT device characteristics are shown in Figure 3b. Devices composing of nanorods have achieved a mobility of 0.61 cm2V-1s-1 and 0.24 cm2V-1s-1 as extracted from the saturated and linear transfer characteristics, respetively. The on-off ratio measured between -50V and +60 V is 3 × 105. It has been observed that the threshold voltage shifts to high negative values, which we explain with a possible overdoping with hydrogen after postdeposition annealing. After hydrothermal growth, but before annealing in hydrogen atmosphere, the threshold voltage is still around zero volts as in spin-coated films and the on-off ratio is 1.6 × 105 (between 0 and 60 V, see Figure 3S). The saturation mobility value before annealing is 0.10 cm2V-1s-1. The annealing in hydrogen leads to a further improvement of mobility but also to a shift of the threshold voltage to large negative values. The device performance has also been evaluated when measured in air without encapsulation because ZnO is known Nano Lett., Vol. 5, No. 12, 2005

Figure 3. (a) Transfer characteristics of a device composed of nanospheres in the linear region (Vd ) 5 V) and saturated region (Vd ) 60 V) after postdeposition hydrothermal growth measured in nitrogen. (b and c) Transfer and output characteristics of a device made from nanorods after postdeposition hydrothermal growth measured in nitrogen. (d) Transfer characteristics of the device shown in b and c measured in air after it had been stored in air for 3 days.

to interact with CO2 and other atmospheric species. Before measurements, the device was stored in air for 3 days. The transfer characteristics of the device shown in Figure 3b were measured in air and are shown in Figure 3d. A mobility of 0.23 cm2V-1s-1 and 0.20 cm2V-1s-1 was extracted from the saturated and linear transfer characteristics, respectively. The on-off ratio is 1.6 × 106. The threshold voltage of devices measured in air tends to be more positive than that of devices measured in nitrogen. The general device performance is comparable to that of TFT devices fabricated in the same device structure by sputtering methods (mobility: 1.2 cm2V-1s-1; on/off: 1.6 × 106).18 For comparison, single ZnO nanowire transistors with mobilities of 1-5 cm2V-1s-1 have been reported.29 For the solution-based method reported here, the raw materials and deposition methods are low-cost, and the aqueous hydrothermal growth in an open vessel should be applicable to large-area substrates. To investigate the relationship between film microstructure and device performance and to identify the mechanisms for the observed improvements of device performance, we have performed atomic force microscopy (AFM) (Figure 4) and scanning-electron microscopy (SEM) (Figure 5). From AFM images such as Figure 4a and the inset SEM image in Figure 5a, it is clear that in as-spun films the nanorods are preferentially oriented with their long axis in the substrate plane. The interactions between the colloidal nanorods during solution growth lead to the formation of small liquid crystalline-like domains with a size on the order of 100 nm in which the nanorods are oriented parallel to each other (see inset of Figure 5a). Similar colloidal self-organization into nematic and smectic-A ordered solids has been reported 2411

Figure 4. Atomic force microscopy topograph of as-spin-coated nanorod (a) and nanosphere (b) films. The scanning range is 500 × 500 nm2. The z topography scale is indicated. The root-mean-square of the film roughness derived from AFM topography images are 8.3 (nanorods) and 1.5 nm (nanospheres), respectively.

Figure 5. Scanning electron microscope images of a ZnO film after postdeposition hydrothermal growth: (a) nanorods and (b) nanosphere. The inset in Figure 5a shows a scanning electron microscope image of the as-spin-cast nanorod film. Figure c is a plan-view scanning electron microscope image at the edge between the ZnO nanorod film and the bare SiO2 substrate after postdeposition hydrothermal growth showing the in-plane orientation of the nanorods in the thin region near the edge. Figure d is a plan-view image of a film deposited from a dilute concentration of nanorods after postdeposition hydrothermal growth.

for CdSe and BaCrO4 nanorods.28,30-31 Because of this selfalignment of the rods, the probability of encountering highangle domain boundaries is reduced. We believe that this oriented in-plane self-assembly of the colloidal nanorods is an important factor contributing to the enhanced mobility of the as-deposited nanorod films compared to nanosphere films. The films exhibit good uniformity over large areas as evidenced by large-scale AFM (Figure 4S) and optical microscopy images (Figure 5S), which is important to achieving good device uniformity and reproducibility. 2412

During postdeposition hydrothermal growth, the nanorods grow further along their c axis forming longer rods, as shown in Figure 5c and d. The average final nanorod length is as long as 300 nm. SEM images of the nanorod film near a thin edge (Figure 5c) clearly show that near the interface with the substrate the nanorods retain their favorable in-plane orientation during the postdeposition hydrothermal growth, whereas on the surface of the film an increasing number of nanorods grow preferably normal to the film plane. We also observe an increase in the diameter of the nanorods, which appears to be occurring mainly because of the fusing of several nanorods (Figure 5d). A similar increase in nanorod size has also been observed in vertically oriented ZnO arrays where microwires are formed by the fusing of many nanowires.32,33 The diameter of individual nanorods also increases slightly to about 15 nm after hydrothermal growth, but by the fusing of several closely packed nanorods the diameter can become as large as ∼60 nm. Generally, the fusing process prefers to take place in the densely packed regions of the films in which nanorods are oriented parallel to each other. This increase in nanowire diameter and length is responsible for the observed improvement of TFT device performance after postdeposition hydrothermal growth. Here we use a two-step approach to obtain self-aligned 300-nmlong nanorods. It is quite difficult to achieve this in a single step because of the poor dispersion properties of long nanorods or nanowires. If the film is made from nanospheres and subjected to postdeposition hydrothermal growth, then the TFTs show only a small improvement of mobility to 0.0024 cm2V-1s-1 and the on-off ratio of 5 × 104. The mobility is more than 2 orders of magnitude lower than that of TFTs made from nanorods by the same fabrication process. During postdeposition hydrothermal growth, the as-spin-cast nanosphere-seed film can grow into an array of ZnO wires, which are, however, aligned randomly with respect to the substrate normal.32 This is less favorable than the in-plane orientation Nano Lett., Vol. 5, No. 12, 2005

of nanowires obtained near the substrate interface in films deposited from a nanorod dispersion. The nanosphere films exhibit good uniformity, as shown in Figure 4b. When immersed in the aqueous solution, the nanospheres grow along a random direction. Some rods are perpendicular to the substrate and appear as bright spots of high electron density in the SEM image; and some rods are growing at an angle to the substrate normal, achieving a limited length of about 50 nm, as shown in Figure 5b. Although this length is comparable to that of the as-prepared nanorod films (65 nm), the TFTs made from nanosphere films subjected to hydrothermal growth exhibit about 1 order of magnitude lower mobility than those made from as-prepared nanorod films. This is further evidence that the orientation of the nanorods is an important factor responsible for the improved performance of devices made from nanorods. Conclusions. High-performance n-type TFTs have been fabricated from colloidal ZnO nanorods deposited by spincoating. Careful control of stoichiometry during nanocrystal growth is crucial for achieving sufficiently low film conductivity and high field-effect mobility. As-deposited nanorods preferentially adopt an in-plane orientation with small domains on the order of 100 nm in which rods are aligned parallel to each other. Postdeposition hydrothermal growth leads to an increase of nanorod diameter and length and results in a significant improvement of device performance. Field-effect mobilities of 0.6 cm2/Vs were achieved in spincast ZnO nanorod films subjected to postdeposition hydrothermal growth. The use of nanorods instead of nanospheres as a seed layer for the hydrothermal growth results in long nanorods oriented preferentially in the plane of the substrate near the interface with the gate dielectric, which is favorable for the charge transport in a TFT. Making use of the selfassembly processes in colloidal nanocrystals is an attractive and simple route for controlling the microstructure and electronic properties of solution-processed semiconductor nanocrystal films. Acknowledgment. We thank Dr. Neil Greenham for use of his synthetic facility and Dr. Peng Wang for helpful discussions. This work was supported by Interdisciplinary Research Collaboration (IRC) in Nanotechnology. Supporting Information Available: UV-vis absorption and FTIR spectra, large-area topographs and optical microscopy images of ZnO nanocrystals, and the transfer characteristics of ZnO nanorod TFTs after postdeposition hydrothermal growth without further annealing. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Katz, H. E. Chem. Mater. 2004, 16, 4748-4756.

Nano Lett., Vol. 5, No. 12, 2005

(2) Gamota, D. R.; Brazis, P.; Kalyanasundaram, K.; Zhang, J. Printed Organic and Molecular Electronics; Kluwer: New York, 2003. (3) Bao, Z. N. ACS Symp. Ser. 2004, 874, 1-14. (4) Dimitrakopoulos, C. D.; Malenfant, P. R. L. AdV. Mater. 2002, 14, 99-117. (5) Sirringhaus, H.; Kawase, T.; Friend, R. H.; Shimoda, T.; Inbasekaran, M.; Wu, W.; Woo, E. P. Science 2000, 290, 2123-2126 (6) Chang, J. F.; Sun, B. Q.; Breiby, D. W.; Nielsen, M. M.; Solling, T. I.; Giles, M.; McCulloch, I.; Sirringhaus, H. Chem. Mater. 2004, 16, 4772-4776. (7) Ong, B. S.; Wu, Y.; Liu, P.; Gardner, S. AdV. Mater. 2005, 17, 11411144. (8) Payne, M. M.; Parkin, S. R.; Anthony, J. E.; Kuo, C. C.; Jackson, T. N. J. Am. Chem. Soc. 2005, 127, 4986-4987. (9) Mitzi, D. B. J. Mater. Chem. 2004, 14, 2355-2365. (10) Chua, L. L.; Zaumseil, J.; Chang, J. F.; Ou, E. C.; Ho, W. P. K. H.; Sirringhaus, H.; Friend, R. H. Nature 2005, 434, 194-199. (11) Waldauf, C.; Schilinsky, P.; Perisutti, M.; Hauch, J.; Brabec, C. J. AdV. Mater. 2003, 15, 2084-2088. (12) Kagan, C. R.; Mitzi, D. B.; Dimitrakopoulos, C. D. Science 1999, 286, 945-947. (13) Mitzi, D. B.; Kosbar, L. L.; Murray, C. E.; Copel, M.; Afzali, A. Nature 2004, 428, 299-303. (14) Duan, X. F.; Niu, C. M.; Sahi, V.; Chen, J.; Parce, J. W.; Empedocles, S.; Goldman, J. L. Nature 2003, 425, 274-278. (15) Ridley, B. A.; Nivi, B.; Jacobson, J. M. Science 1999, 286, 746749. (16) Goldstein, A. N.; Echer, C. M.; Alivisatos, A. P. Science 1992, 256, 1425-1427. (17) Xia, Y. N.; Yang, P. D.; Sun, Y. G.; Wu, Y. Y.; Mayers, B.; Gates, B.; Yin, Y. D.; Kim, F.; Yan, Y. Q. AdV. Mater. 2003, 15, 353389. (18) Carcia, P. F.; McLean, R. S.; Reilly, M. H.; Nunes, G. Appl. Phys. Lett. 2003, 82, 1117-1119. (19) Hoffman, R. L.; Norris, B. J.; Wager, J. F. Appl. Phys. Lett. 2003, 82, 733-735. (20) Ohya, Y.; Niwa, T.; Ban, T.; Takahashi, Y. Jpn. J. Appl. Phys., Part 1 2001, 40, 297-298. (21) Norris, B. J.; Anderson, J.; Wager, J. F.; Keszler, D. A. J. Phys. D: Appl. Phys. 2003, 36, L105-L107. (22) Meulenkamp, E. N. J. Phys. Chem. B 1998, 102, 5566-5572. (23) Beek, W. J. E.; Wienk, M. M.; Janssen, R. A. J. AdV. Mater. 2004, 16, 1009-1013. (24) Pacholski, C.; Kornowski, A.; Weller, H. Angew. Chem., Int. Ed. 2002, 41, 1188-1191. (25) Pacholski, C. Selbstorganisation von ZnO nanoteilchen und Untersuchung der Ladungstra¨gerdynamik, Thesis, Hamburg University, 2002. (26) Van de Walle, C. G. Phys. ReV. Lett. 2000, 85, 1012-1015. (27) Norberg, N. S.; Kittilstved, K. R.; Amonette, J. E.; Kukkadapu, R. K.; Schwartz, D. A.; Gamelin, D. R. J. Am. Chem. Soc. 2004, 126, 9387-9398. (28) Li, L. S.; Alivisatos, A. P. AdV. Mater. 2003, 15, 408-411. (29) Law, M.; Greene, L. E.; Johnson, J. C.; Saykally, R.; Yang, P. D. Nat. Mater. 2005, 4, 455-459. (30) Talapin, D. V.; Shevchenko, E. V.; Murray, C. B.; Kornowski, A.; Forster, S.; Weller, H. J. Am. Chem. Soc. 2004, 126, 12984-12988. (31) Kim, F.; Kwan, S.; Akana, J.; Yang, P. D. J. Am. Chem. Soc. 2001, 123, 4360-4361. (32) Greene, L. E.; Law, M.; Goldberger, J.; Kim, F.; Johnson, J. C.; Zhang, Y. F.; Saykally, R. J.; Yang, P. D. Angew. Chem., Int. Ed. 2003, 42, 3031-3034. (33) Li, Q. C.; Kumar, V.; Li, Y.; Zhang, H. T.; Marks, T. J.; Chang, R. P. H. Chem. Mater. 2005, 17, 1001-1006.

NL051586W

2413