Correlation between ZnO Nanowire Growth and the Surface of AlN Substrate Sang Hyun Lee,*,† In-Ho Im,‡ Hyun Jung Lee,§ Zahra Vashaei,‡ Takashi Hanada,‡ Meoung-Whan Cho,‡ and Takafumi Yao†,‡ Center for Interdisciplinary Research, Tohoku UniVersity, Aramaki, Aoba-ku, Sendai 980-8578, Japan, Institute for Materials Research, Tohoku UniVersity, Katahira 2-1-1, Aoba-ku, Sendai 980-8577, Japan, and Graduate School of EnVironmental Studies, Tohoku UniVersity, 07 Aramaki-Aoba, Aoba-Ku, Sendai 980-8579, Japan
CRYSTAL GROWTH & DESIGN 2006 VOL. 6, NO. 12 2640-2642
ReceiVed July 8, 2006; ReVised Manuscript ReceiVed October 5, 2006
ABSTRACT: Single-crystalline ZnO nanowires are fabricated by thermal chemical vapor transport and condensation on AlN epilayers without employing any metal catalyst. Before the growth of ZnO nanowires, the surface of AlN epilayers was treated by HF solution and then changed to a mixture of flat plane and hillocks according to dipping time. ZnO nanowires along the c-axis direction of hexagonal structures were synthesized only on the HF treated AlN surface. ZnO nanowire arrays have been obtained by selective etching by photolithography process. A mechanism for nanowire growth on modified AlN epilayer is proposed. Recently, many efforts have been directed toward a new synthesis method of nanostructures that makes it possible to control shape, growth site, and direction; the study of the growth mechanism; and an array technique using semiconductor materials. Zinc oxide (ZnO) is investigated as one of the most important oxide semiconductor materials because of its multifunctional properties such as a direct wide band gap (3.37 eV), large excitation binding energy (60meV),1 and piezoelectricity attributable to noncentral symmetry.2 Especially, ZnO nanostructures of various shape such as nanowires,3,4 nanobelts,5 and nanorings6 have been proven as useful materials in applicable devices such as UV detectors,7 lasers,4 light-emitting diodes,8 and the electron source in field-emission devices.9 Aligned one-dimensional ZnO nanostructures have been studied using a plane-oriented Al2O3 substrate4 or AlxGa1-xN epilayer with the assistance of Au particles as catalysts.10 However, this technique has limitations for device applications. First, Al2O3 is a nonconductive material, which makes it difficult for ZnO nanostructures to be fabricated in electric and optoelectric devices. Another limitation is that the metal catalyst could diffuse into the nanowires; Au is usually trapped in the tip of the nanowires in vapor-liquidsolid (VLS) growth and generates unintentional defects. Using a heterostructure substrate, other applications are also possible. In the case of ZnO films, polarity control of ZnO and application as a surface acoustic wave (SAW) device using aluminum nitride (AlN) film have been reported also.11,12 In this communication, we present ZnO nanowires grown on surface-modified AlN thin-film substrates without any catalysts. The grown nanowires are aligned vertically with low density on a HF-treated AlN epilayer surface. Our results show that ZnO nanowires can be epitaxially grown on the heterostructure substrate and suggest them as a candidate structure for single nanowire devices such as vertical type light-emitting diodes (LED), singleelectron transistors (SET), and field-effect transistor (FET). In experiment, undoped c-plane AlN was epitaxially grown by molecular beam epitaxy (MBE) on one side-polished c-planeoriented sapphire (0001) substrate to a thickness of about 200 nm. For surface cleaning and treatment, we used acetone and methanol as a solvent and HF as an acid. The morphology variation of the AlN epilayer by the HF treatment was examined using atomic force microscopy (AFM, Dimension 3000 (Veeco)). As shown in Figure 1, the root-mean-square (rms) roughness and height of hillock was obtained by AFM. The mean height of the hillocks increased from 2.3 to 9.5 nm with etching time, whereas rms values are similar to * Corresponding author. E-mail:
[email protected]. Tel: 81-22795-4404. Fax: 81-22-795-7810. † Center for Interdisciplinary Research, Tohoku University. ‡ Institute for Materials Research, Tohoku University. § Graduate School of Environmental Studies, Tohoku University.
Figure 1. Plot of average AFM rms roughness (dash line) and height (line) of hillock of AlN layer according to HF treatment time.
that of the as-grown sample. The surface of AlN became smooth with the etching except for the partial hillocks. The rms values of all parts except the hillock (shown in panels d and e of Figure S1 of the Supporting Information) decreased to 0.712 and 0.569 nm, compared to 0.846 nm for the as-grown sample. ZnO nanowires were grown on as-grown and HF-treated AlN layers by thermal chemical vapor transport and condensation using a mixture of equal amounts of ZnO and graphite powder (0.5 g each) without any metal catalysts.7 The source being loaded in an alumina boat was placed into the center of a quartz tube of an electrical furnace and then the substrates were typically placed at downstream ∼10 cm away from the source. Ar was injected as the carrier gas at a 300 sccm flow rate. The growth of the ZnO nanowires was conducted by evaporation of the ZnO source at 930980 °C for 30 min and condensation on the substrate at about 700750 °C. After the synthesis, the furnace was turned off, and the reactor tube was cooled to room temperature under an argon atmosphere. Both as-grown AlN films with flat surface morphology and hillocks by HF treatment were loaded into the furnace simultaneously. The inset images a and b of of Figure 2 indicate the morphologies of the AlN layers before loading into the furnace for the growth of the ZnO nanowires. The former AFM image represents a flat surface, whereas the latter reveals the hillock formation. Scanning electron microscopy (SEM, Hitachi S-4300E) was used to confirm whether the nanowires were synthesized on the substrates or not. Interestingly, the growth of the ZnO nanowires occurred on the HF-treated AlN layer only, whereas the as-grown AlN surface retained the flat morphology like before the growth,
10.1021/cg060435j CCC: $33.50 © 2006 American Chemical Society Published on Web 10/31/2006
Communications
Crystal Growth & Design, Vol. 6, No. 12, 2006 2641
Figure 4. (a) Cross-sectional SEM image of the ZnO nanowires with pyramid shaped bottom on the AlN layer. (b) Schematic diagram of the ZnO nanowire formation at a hillock site by nucleation and epitaxal growth.
Figure 2. Tilted SEM images after the ZnO growth on the AlN surfaces (a) without any treatments and (b) with the HF treatment for 4 min where the ZnO nanowires were grown. Inset images indicate the surface morphology of the AlN before the growth of ZnO. (c) Transmission electron microscopy (TEM) image of thin ZnO nanowires grown on the AlN layer. Upper left inset image shows selected area diffraction pattern (SADP) pattern corresponding to the lower right high-resolution TEM image of a single crystalline ZnO nanowire with [0001] growth direction.
Figure 3. (a) Schematic diagram for the growth of the ZnO nanowire arrays. (b) Plane and (b) cross-sectional SEM images of the arrays of the vertical ZnO nanowires on the AlN epilayer.
as shown in images a and b of Figure 2. The nanowires grow predominantly perpendicular to the AlN (0001) substrate, with diameters varying from 15 to 24 nm and average height of ∼5 µm. The density of the ZnO nanowires is about 0.8-0.16 µm-2, similar to that of the hillocks (average value 0.885 µm-2) existing on the AlN surface before the growth of the nanowires. To analyze the size and crystal structure of the ZnO nanowires, we operated a transmission electron microscope (TEM, JEM-3000F) at 300 kV. We detached the wires from the substrate by just scratching the surface of the substrate in ethanol. As shown in the lower-magnification TEM image of the ZnO nanowires grown on the AlN layer in Figure 2c, all of the nanowires exhibit a fairly uniform diameter of about 15-20 nm along their entire length except for the bottom of the nanowires. Insets in Figure 2c represent SADP and the lattice image at the middle part of the ZnO nanowire. It is shown that the straight nanowire is a single crystal and grown primarily along the [0001] direction. Arrays of the ZnO nanowires on the AlN layer were made by HF-selective treatment using a photolithography process shown in Figure 3. Positive photoresist (S1818, Shipley) was coated on the AlN surface by spin coating and polymerized by UV radiation to a strip-typed pattern using a photolithomask. After that, the UV irradiated part was removed using developer. The exposed AlN surface was immersed in the HF solution for 4 min and then any
remaining photoresists were removed by acetone. The selectively HF-treated AlN layer after all these processes was loaded in the furnace after cleaning by ethanol. The ZnO nanowires were synthesized on these stripe-patterned substrates by the previously mentioned method. As a result of the growths, arrays of vertical ZnO nanowires within a 20 µm linewidth can be confirmed, as indicated in images b and c of Figure 3. Here, we propose a growth mechanism for the ZnO nanowire on the HF treated AlN layer. Three possible factors, chemical binding, surface energy, and morphological effects, could be considered. The first is compositional variation at the hillock due to HF treatment. In the case of an untreated AlN layer, a natural oxide layer exists as a stable surface on the flat (0001) face of AlN. Therefore, thermal energy at substrate temperature, 700 °C, during the growth of ZnO is not enough to break the Al-O bonds on the untreated AlN surface. Previous studies reported on the wet chemical etching of AlN films.13,14 They suggested from the XPS results that the natural oxide on the AlN surface was effectively removed by the HF solution. Though the AlN surface became strongly passivated by hydrogen, the oxide layer partially remained on the AlN surface. Remaining oxide islands could work as an activation site for the formation of the ZnO seeds because of the chemical bonding energy difference. The substrate surface surrounding the hillocks became more flat with increasing HF etching time, as shown in Figure 1. The atoms of Zn and oxygen dropping on the flat AlN surface are considered to diffuse into the edge of the hillocks. In order to clearly support the proposed mechanism, we observed the boundary morphology between the ZnO nanowires and the AlN surface. Figure 4 shows a cross-section SEM image of the nanowires grown on the AlN surface and a schematic diagram of the ZnO nanowire formation at a hillock on the AlN surface. The pyramid-shaped structure with a (101h1) or (101h2) face is observed at the bottom of the nanowires in Figure 4a. The pyramid is bigger than the hillocks on the AlN surface observed by AFM before the growth of the ZnO nanowires. It is reported that hillocks with the six {101h1} faces of AlN are easy to form by etching.15 If HF is used as etchant, the AlN (0001) substrate surface would become energetically more stable by H termination of the surface dangling bonds. As a result, zinc and oxygen atoms are more mobile on the (0001) surface and attached on the (101h1) or (101h2) face to form the base pyramid at the initial hillock site. Recently, Kametani et al. reported the formation of ZnO nanodot array along the step edge on the R-face sapphire at 600 °C by metalorganic chemical vapor deposition (MOCVD).15 They considered that the ZnO nanodots aligned linearly along the steps because atoms are trapped only at the step edges because of large surface diffusion length on the R-plane terrace, which has less density of dangling bonds than the a-plane and the c-plane of sapphire. In our experiment, ZnO nanowire along the c-axis growth direction was presumably grown at the AlN hillock site. The epitaxial relationship between the ZnO nanowire and the AlN substrate is expected to be ZnO(0001) // AlN(0001) and ZnO[101h0] // AlN[101h0] because of the same wurtzite structure and small lattice mismatch of about 4.3%. It is considered that the surface energy of the (101h0) side wall of the ZnO nanowire is much lower than that of the top (0001) face at our growth condition because, on the nonpolar (101h0) surface, the dangling bond of a surface cation atom
2642 Crystal Growth & Design, Vol. 6, No. 12, 2006 can be emptied by charge transfer to the nearest-neighbor anion dangling bond, which will be filled with a lone pair, whereas the top ZnO (0001) surface is not terminated by H in contrast to the AlN substrate. As a result, diffusion length of the atoms on the (101h0) side wall is expected to be on the order of micrometers and the atoms are trapped at the top (0001) face to form the nanowire. Single-crystalline ZnO nanowires with c-axis growth direction have been grown on the HF-treated AlN film via the thermal chemical vapor transport and condensation method without metal catalyst. The growth of the ZnO nanowires occurred at the AlN surface with the HF treatment, and their density was similar to that of the hillocks formed in the HF-treated area. In addition, the array of the ZnO nanowires was obtained by the selective HF treatment using photolithography. Nucleation of ZnO is attributed to the existence of the hillocks on the HF-treated AlN surface. Supporting Information Available: AFM images of the surface roughness evolution (PDF). This material is available free of charge via the Internet at http://pubs.acs.org.
References (1) Chen, Y.; Bagnall, D. M.; Koh, H.; Park, K.; Hiraga, K.; Zhu, Z.;Yao, T. J. Appl. Phys. 1998, 84, 3912. (2) Yamamoto, T.; Shiosaki, T.; Kawabata, A. J. Appl. Phys. 1980, 51, 3113.
Communications (3) Kind, H.; Yan, H.; Messer, B.; Law, M.; Yang, P. D. AdV. Mater. 2002, 14, 158. (4) Huang, M.; Mao, S.; Feick, H.; Yan, H. Q.; Wu, Y. Y.; Kind, H.; Weber, E.; Russo, R.; Yang, P. D. Science 2001, 292, 1897. (5) Pan, Z, W.; Dai, Z. R.; Wang, Z. L. Science 2001, 291, 1947. (6) Kong, X. Y.; Wang, Z. L. Nano Lett. 2003, 3, 1625. (7) Lee, S. H.; Lee, H. J.; Oh, D.; Lee, S. W.; Goto, H.; Buckmaster, R.; Yasukawa, T.; Matsue, T.; Hong, S.-K.; Ko, H.; Cho, M.-W.; Yao, T. J. Phys. Chem. B 2006, 110, 3856. (8) Liu, C. H.; Zapien, J. A.; Yao, Y.; Meng, A. M.; Lee, C. S.; Fan, S. S.; Lifshitz, Y.; Lee, S. T. AdV. Mater. 2003, 15, 839. (9) Jo, S. H.; Banerjee, D.; Ren, Z. F. Appl. Phys. Lett. 2004, 85, 1407. (10) Wang, W.; Song, J.; Li, P.; Ryou, J. H.; Dupuis, R. D.; Summers, C. J.; Wang, Z. L. J. Am. Chem. Soc. 2005, 127, 7920. (11) Yong, Y.-Y.; Kang, Y.-S.; Lee, P. S.; Lee, J.-Y. J. Vac. Sci. Technol., B 2002, 20, 42. (12) Wang, Y.; Du, X. L.; Mei, Z. X.; Zeng, Z. Q.; Ying, M. J.; Yang, H. T.; Jia, J. F.; Xue, Q. K.; Zhang, Z. Appl. Phys. Lett. 2005, 87, 051901. (13) Lee, K. N. Lee; Donovan, S. M.; Gila, B.; Overberg, M.; Mackenzie, J. D.; Abernathy, C. R.; Wilson, R. G. J. Electrochem. Soc. 2000, 147, 3087. (14) King, S. W.; Barmak, J. P.; Bermser, M. D.; Tracy, K. M.; Ronning, C.; Davis, R. F.; Nemanich, R. J. J. Appl. Phys. 1998, 84, 5248. (15) Kametani, K.; Imamoto, H.; Fujita, S. Physica E 2006, 32, 33.
CG060435J
