Metal-Catalyzed CVD Method to Synthesize Silicon Nanobelts - The

Sep 6, 2008 - Norihisa Uesawa , Susumu Inasawa , Yoshiko Tsuji and Yukio Yamaguchi. The Journal of Physical Chemistry C 2010 114 (10), 4291-4296...
0 downloads 0 Views 1MB Size
Download PDF
J. Phys. Chem. C 2008, 112, 15129–15133

15129

Metal-Catalyzed CVD Method to Synthesize Silicon Nanobelts D. P. Wei and Q. Chen* Key Laboratory for the Physics and Chemistry of NanodeVices and Department of Electronics, Peking UniVersity, Beijing 100871, P. R. China ReceiVed: May 2, 2008; ReVised Manuscript ReceiVed: July 29, 2008

On the basis of the metal-catalyzed chemical vapor deposition method, a general route was developed to synthesize single crystal silicon nanobelts using Au, Cu, and Zn as catalysts, avoiding any template and additive. The nanobelts were characterized using electron microscopy. The silicon nanobelts, which have great length, smooth edges, flat surfaces, and tiny thicknesses, grow along the 〈112〉 direction with the top surface of the {111} facet. Evidence shows that the growth of the silicon nanobelt may follow the vapor-liquid-solid process. Furthermore, the growth conditions of the nanobelts were preliminarily investigated. These studies not only provide novel materials for nanodevices but also offer helpful instructions for the metal-catalyzed synthesis of other semiconductor nanobelts. 1. Introduction

2. Experimental Section

In recent years, silicon nanostructures, being possible building blocks for nanoelectronics, have attracted considerable interest in the controllable synthesis1-6 and application of nanodevices.7-12 Especially, silicon nanowires also have great potential in many other applications, such as chemical or biological nanosensors,13,14 solar cells,15 and photoelectrochemical cells.16 In such applications, the nanobelts may have some advantages over the nanowires because the ribbon geometry provides large planar surfaces, which can increase the sensing and detecting area, reduce diffusion time, and can efficiently fill the channel regions of transistors.17,18 Field effect transistors based on silicon nanobelts can have mobilities of ∼190 cm2 V-1 S-1 and on/off ratios of ∼3 × 104.17 Biological sensors based on silicon nanobelts have exhibited sensitivity extending to below the picomolar concentration regime.18 Despite such a fascinating foreground, study on silicon nanobelts is still rare, partially due to the lack of a general synthesis method. To date, three main methods have been developed to synthesize silicon nanobelts. Specialized lithography and etching procedures have been used to produce Si nanobelts from bulk Si wafers via a top-down approach.17 Nanochannels made by lithography have been used as templates to fabricate nanowires/nanobelts by a grow-in-place approach.19 However, these methods are expensive and complicated. The third method is an oxide-assisted growth (OAG) method,20,21 which is simple and easy but brings bulk defects and a thick surface oxide layer.

Synthesis of Si Nanobelts Using Au and Cu Catalysts. The synthesis of silicon nanobelts is similar to the synthesis of silicon nanowires via the metal-catalyzed VLS process.22,23 Si (100) wafer, on which 5 nm Au or 10 nm Cu film has been deposited by electron beam evaporation (using a K. J. Lesker thin film evaporator), was placed into a quartz tube in a furnace. First, the quartz tube was flushed by high-purity Ar for 15 min with a flowing rate of 2000 sccm (standard cubic centimeter per minute). Second, under a constant flow of 300 sccm H2, the furnace was heated to 860 °C at a rate of 30 °C/min, and then the substrate was moved to the highest temperature region of the furnace. Meanwhile, additional 60 sccm H2, which has passed through a bubbler filled with SiCl4 at 19 °C, also flowed through the quartz tube making the total flux of H2 gas in the tube 360 sccm. The reaction lasted 15 min under atmospheric pressure. After the reaction, the quartz tube was moved back to its original position, and was flushed with 300 sccm Ar. Finally, the system cooled down to room temperature, and yellow products were collected. Synthesis of Si Nanobelts Using the Zn Catalyst. The experimental setup is similar to that using the Au or Cu catalyst. The quartz boat, containing Zn powders (purity 99.99%), was placed into a quartz tube in a furnace. When the temperature at the center of the furnace reached 860 °C, the quartz tube was gradually moved in until the quartz boat was in the highest temperature region of the furnace. Meanwhile, SiCl4 gas was brought into the quartz tube by additional 60 sccm H2. The reaction lasted 15 min under atmospheric pressure. Finally, when the system cooled down to room temperature, yellow products in the region of 650-700 °C were collected. Other experimental conditions and detailed procedures were the same as that using Au or Cu as catalyst. Characterization of the Products. The products were characterized using scanning electron microscope (SEM, FEI XL 30F), transmission electron microscope (TEM, FEI Tecnai G20 and Tecnai F30), and energy-dispersive X-ray spectrometry (EDS, EDX spectroscope attached to the TEM).

Metal-catalytic chemical vapor deposition (CVD) is a general method to produce one-dimensional nanomaterials; however, it has not been reported to produce silicon nanobelts. Here, we successfully synthesized free-standing high quality silicon nanobelts by the metal-catalytic CVD method for the first time using various metal catalysts. The nanobelts synthesized here have great length, smooth edges, flat surfaces, and tiny thicknesses, providing novel materials for nanodevices. Our results suggested that these silicon nanobelts grow via the vaporliquid-solid (VLS) process. * Corresponding author. E-mail: [email protected].

10.1021/jp8038785 CCC: $40.75  2008 American Chemical Society Published on Web 09/06/2008

15130 J. Phys. Chem. C, Vol. 112, No. 39, 2008

Figure 1. SEM images of silicon nanobelts synthesized using (a) Au and (b) Cu as catalysts. Insets: higher magnification images of silicon nanobelts. The scale bars in the insets are 2 µm. (c and d) TEM images of two silicon nanobelts showing the thickness of the nanobelts.

3. Results and Discussion Figure 1a and b shows typical SEM images of the products synthesized using Au and Cu as catalysts, respectively. It shows that the products consist of wire-like nanostructure and ribbonlike nanostructure with apparent waving and twisting shapes. Most of the nanobelts are longer than 100 µm. All the nanobelts have smooth edges along their entire length, and the typical width of the nanobelts is in the range of 0.2 to 2 µm. From the cross sectional TEM images (Figure 1c and d), the typical thickness of the nanobelts was measured to be about 20 to 60 nm. Figure 2a shows a low magnification TEM image of a silicon nanobelt with 1.5 µm width. The ripple-like contrast is due to slight bending of the nanobelt. The corresponding selected area electron diffraction (SAED) pattern in Figure 2b can be indexed to the [111] zone axis of Si single crystal plus a set of anomalous diffraction spots at 1/3{224j} positions. The EDS spectrum taken from the individual nanobelt (shown in Figure 3d; the Cu peak comes from the Cu grid) confirms that the nanobelt is Si. The single crystal Si nanobelts grow along [112j] and lies on the (111) plane, which is different from the nanobelt grown by OAG method.20 The 1/3{224j} reflections should be forbidden for bulk Si, but have been observed in silicon nanowires24,25 and other nanomaterials such as Au or Ag nanostructures.26-28 There are several possible reasons for the appearance of these forbiddon reflections. For a perfect crystal with clean surfaces, an incomplete ABC stacking sequence causes a fractional unit cell along the [111] direction and gives rise to the extra reflections. This is also called the surface termination effect or surface step effect.29 Besides, stacking fault, twin or surface reconstruction on the (111) plane may also cause the appearance of the forbidden 1/3{224j} reflections.26 Here, experiments were performed to find the reason of the forbidden reflections in the present case. Convergent beam electron diffractions were taken along the [111] zone axis, and the same 3-fold symmetry as bulk Si was observed. A sequence of tilting experiments along the [022j]* axis was also performed. The [211] zone axis was obtained after the sample was tilted 19.5° from the [111] direction. The 1/3(224j) forbidden spot never disappeared during the tilting, but now it is indexed to 1/2(113j), as shown in Figure 2c. When the nanobelt was tilted -35.3° from [111], the [011] zone axis appeared, as shown in Figure 2d. The diffraction spots are all very sharp and round in these SAED patterns. Besides

Wei and Chen the forbidden spots, no other extra spots or streaks can be observed. Furthermore, we have tilted the nanobelt for more than 60° and examined the image; no plane defect contrast (such as that due to stacking fault or twin) can be observed. The above evidence rules out the possibility that the appearance of the forbidden spots is due to stacking fault or twin. Surface reconstruction is not possible in the present case because the nanobelt has never been treated in ultrahigh vacuum. Therefore, the appearance of the extra diffraction spots is due to the surface termination effect. The highresolution TEM (HRTEM) image shown in Figure 2e reveals the lattice spacing of 0.19 nm corresponding to {022j } planes and the lattice spacing of 0.33 nm corresponding to 1/3{224j } planes. All of the above evidence indicate that the present nanobelt is a very flat and thin Si single crystal.30 To investigate the role of metal catalyst, an additional experiment was done without using a metal catalyst in the reaction and keeping all the other experimental conditions unchanged. Neither nanowires nor nanobelts can be obtained, and the substrate remained clean after the reaction. Obviously, the metal catalyst is vitally important to the growth of the silicon nanobelts, which cannot be interpreted using OAG or the vapor-solid (VS) mechanism. Metal particles were often observed at the front end of the present silicon nanobelts, as shown in Figure 3a and the inset of Figure 3d. The EDS spectra shown in Figure 3d verify that the metal particle is the Au-Si alloy and the rest of the nanobelt is Si. Cu and C signals come from the supporting grid and the supporting film. The presence of a catalyst nanoparticle at one of the ends of the nanobelt is a clear evidence supporting the VLS mechanism.31 Similar to the growth of ZnS and PbO2 nanobelts,32,33 the alloy particle provides a catalytic active site at the growth front and is crucial in determining the morphology of the nanostructures.32,34 Interestingly, we sometimes observed that a small portion of the nanobelts shrunk abruptly at the growth front to form nanowires, as shown in Figure 3b and c. This phenomenon might result from mass loss of the metal catalyst35,36 or abrupt environment fluctuations during the synthesis process. In general, the growth direction via VLS process depends on the minimum of the total system energy including the surface and interface free energy.37,38 Figure 4 shows a silicon nanowire obtained simultaneously with the nanobelts. The silicon nanowires grew primarily along the 〈111〉 direction versus 〈112〉 direction for nanobelts. The contribution of the solid-liquid interface energy to the total energy has been demonstrated to be dominant for the growth of ZnSe nanowires with large diameter.37 The same rule may also apply to Si nanowires. The silicon nanowires, with diameters larger than 30 nm, grow along the 〈111〉 direction because the {111} plane is the minimumenergy facet.38 However, the contribution of surface free energy to the total energy is dominant for silicon nanobelts because of their large surface. In order to minimize the system energy, silicon nanobelts thus tend to grow predominantly along the 〈112〉 direction with the top surface terminated on the {111} facet. In order to manifest that the above method can be extended easily, another metal catalyst (Zn) was also used to synthesize silicon nanobelts. At first, Zn film, which was deposited on a substrate by electron beam evaporation, was used as catalyst. But such a Zn film could not effectively catalyze the growth of silicon nanostructure. It could be explained as follows: as the melting point of Zn (419 °C) is low, the thin Zn film may evaporate at high temperature and may not be able to catalyze the reaction. Hence, maintaining Zn droplets is vitally important

Metal-Catalyzed CVD Method

J. Phys. Chem. C, Vol. 112, No. 39, 2008 15131

Figure 2. TEM image of a silicon nanobelt (a) and the corresponding SAED pattern (b). (c and d) SAED patterns at the indicated tilting angle. (e) High-resolution TEM image of a nanobelt taken along the [111] zone axis. The arrow points out the growth direction of the nanobelt.

Figure 3. (a) SEM image of a silicon nanobelt with a catalyst metal at the end. Inset: Cross-sectional SEM image of this nanobelt. Scale bar is 1 µm. (b and c) SEM images showing that some Si nanobelts shrunk abruptly at the front to form nanowires. (d) EDS spectra taken from the area pointed by the arrows in the inset. The inset is a TEM image of silicon nanobelts.

to the VLS growth of silicon nanostructures. Here, we solved the problem by evaporating Zn powder at high temperature in the furnace during the reaction. Silicon and Zn vapors mixed together and alloyed in the appropriate temperature region. The steady concentration of Zn vapor helps to maintain alloy droplets. A mass of silicon nanowires and nanobelts were synthesized in the temperature region of 650-700 °C. Figure

5a shows a SEM image of the products. Similar to that obtained using Au or Cu catalysts, nanobelts and nanowires coexisted. TEM image and the SAED pattern of a silicon nanobelt shown in Figure 5b indicate that the nanobelt also has a smooth surface and single crystal structure, and that it grew along 〈112〉. With other metal catalysts, such as Fe,39 Ni,40 Al,41 Pt,42 and Ti,43 which have been successfully used to grow silicon

15132 J. Phys. Chem. C, Vol. 112, No. 39, 2008

Wei and Chen

Figure 4. (a) TEM image of an individual silicon nanowire obtained simultaneously with the silicon nanobelts. Inset: The corresponding SAED pattern reveals that the nanowire grew along the [111] direction. (b) High-resolution TEM image of the silicon nanowire shown in (a).

Figure 6. TEM images of silicon nanostructures obtained using Au as catalyst at 1000 °C. (a and b) TEM images of silicon nanostructures obtained when the flow rates of the additional H2 were (c) 40 and (d) 140 sccm, respectively. Figure 5. (a) SEM image of silicon nanobelts produced using Zn as catalyst. (b) TEM image of a silicon nanobelt. Inset: The corresponding SAED pattern.

nanowires, it should be possible to fabricate silicon nanobelts. The metal atoms diffused from catalyst droplets to the surface or the inside of Si nanostructures can result in weak doping. Proper metal catalysts may improve the electronic property of the silicon nanobelts. The growth temperature,44,45 the concentration of precursor vapor,46 the flow rate of the gases,34 and other factors may affect the morphology of the products. The experimental parameters used here to grow Si nanobelts are very similar to those used previously to synthesize Si nanowires.22,23 It is very important to precisely control the experimental conditions to grow special nanostructures. The growth conditions of silicon nanobelts and nanowires were investigated on our system using Au as catalyst. When the temperature of the furnace was 1000 °C and all the other conditions were unchanged, no nanobelts can be obtained. Instead, rough nanowires covered by many large particles were observed, as shown in Figures 6a and b. When the temperature of the furnace was lower than 500 °C, neither nanowires nor nanobelts could be obtained. The possible reason is that SiCl4 and H2 cannot react properly at low temperature. The optimal temperature for the nanobelt to grow was found to be about 860 °C for our experimental system. The flow rate of the additional H2 also affects the products. If the flow rate of the additional H2 (which passed through a SiCl4 bubbler) was 40 sccm, only nanowires and no nanobelts were produced, as shown in Figure 6c. If the flow rate of the additional H2 was 140 sccm and the total flux of H2 gas remains unchanged, silicon microrods were produced, as shown in Figure 6d. Some microrods interconnected, implying that the vapor-solid process might be involved. The above results indicate that the morphologies of the products are very sensitive to growth conditions. Different experimental systems may need different parameters to grow nanobelts or nanowires. Although the reaction temperature and the total flow rate used in the present experiments to grow nanobelts are roughly the same as those used to grow

nanowires by other groups, the temperature of the SiCl4 bubbler (19 °C) and the flow rate of the additional H2 (60 sccm) in the present case are lower than those in the previous report (which are 30 °C and 100 sccm, respectively).23 The present results also show that the morphology of the products changes with the flow rate of the additional H2, which passed through a SiCl4 bubbler and affected the amount of SiCl4 vapor being introduced into the furnace. Nanowires will grow at a relatively low concentration of SiCl4. A suitable amount of SiCl4 vapor benefits the growth of nanobelts. If there is too much SiCl4, microrods will grow. The growth conditions on a new experimental system need to be optimized to produce high-quality silicon nanobelts. 4. Conclusions In conclusion, we successfully synthesized high-quality single crystal silicon nanobelts by using Au, Cu, or Zn as metal catalyst. Generally, the silicon nanobelts are more than 100 µm long and from 0.2 to 2 µm wide. Importantly, the nanobelts have flat surfaces and tiny thicknesses. Evidence shows that the growth of silicon nanobelts may be a VLS process. Silicon nanobelts grow along the 〈112〉 direction with the top surface on the {111} facet, which minimizes the system free energy. Furthermore, the growth conditions of the nanobelts were investigated. Acknowledgment. This work was supported by NSF of China (60771005, 60728102). References and Notes (1) Morales, A. M.; Lieber, C. M. Science 1998, 279, 208. (2) Holmes, J. D.; Johnston, K. P.; Doty, R. C.; Korgel, B. A. Science 2000, 287, 1471. (3) Hu, J. Q.; Bando, Y.; Liu, Z. W.; Zhan, J. H.; Golberg, D.; Sekiguchi, T. Angew. Chem., Int. Ed. 2004, 43, 63. (4) Li, C.; Liu, Z. T.; Gu, C.; Xu, X.; Yang, Y. AdV. Mater. 2006, 18, 228. (5) Carroll, M. S.; Brewer, L.; Verley, J. C.; Banks, J.; Sheng, J. J.; Pan, W.; Dunn, R. Nanotechnol. 2007, 18, 315707. (6) Mangolini, L.; Kortshagen, U. AdV. Mater. 2007, 19, 2513.

Metal-Catalyzed CVD Method (7) Yang, C.; Zhong, Z. H.; Lieber, C. M. Science 2005, 310, 1304. (8) Huang, Y.; Duan, X. F.; Cui, Y.; Lauhon, L. J.; Kim, K. H.; Lieber, C. M. Science 2001, 294, 1313. (9) Cui, Y.; Lieber, C. M. Science 2001, 291, 851. (10) Goldberger, J.; Hochbaum, A. I.; Fan, R.; Yang, P. D. Nano. Lett. 2006, 6, 973. (11) Walters, R. J.; Bourianoff, G. I.; Atwater, H. A. Nat. Mater. 2005, 4, 143. (12) Walters, R. J.; Carreras, J.; Feng, T.; Bell, L. D.; Atwater, H. A. IEEE J. Sel. Top. Quantum. Electron. 2006, 12, 1647. (13) Cui, Y.; Wei, Q. Q.; Park, H. K.; Lieber, C. M. Science 2001, 293, 1289. (14) Zheng, G. F.; Patolsky, F.; Cui, Y.; Wang, W. U.; Lieber, C. M. Nat. Biotechnol. 2005, 23, 1294. (15) Tian, B. Z.; Zheng, X. L.; Kempa, T. J.; Fang, Y.; Yu, N. F.; Yu, G. H.; Huang, J. L.; Lieber, C. M. Nature 2007, 449, 885. (16) Goodey, A. P.; Eichfeld, S. M.; Lew, K. K.; Redwing, J. M.; Mallouk, T. E. J. Am. Chem. Soc. 2007, 129, 12344. (17) Ko, H. C.; Baca, A. J.; Rogers, J. A. Nano Lett. 2006, 6, 2318. (18) Elfstroˆm, N.; Karlstro¨m, A. E.; Linnros, J. Nano Lett. 2008, 8, 945. (19) Shan, Y.; Kalkan, A. K.; Peng, C. Y.; Fonash, S. J. Nano Lett. 2004, 4, 2085. (20) Shi, W. S.; Peng, H. Y.; Wang, N.; Li, C. P.; Xu, L.; Lee, C. S.; Kalish, R.; Lee, S. T. J. Am. Chem. Soc. 2001, 123, 11095. (21) Zhang, R. Q.; Lifshitz, Y.; Lee, S. T. AdV. Mater. 2003, 15, 635. (22) Hochbaum, A. I.; Fan, R.; He, R. R.; Yang, P. D. Nano Lett. 2005, 5, 457. (23) Ge, S. P.; Jiang, K. L.; Lu, X. X.; Chen, Y. F.; Wang, R. M.; Fan, S. S. AdV. Mater. 2005, 17, 56. (24) Gudiksen, M. S. Ph.D. Dissertation, Harvard University, Cambridge, MA, 2002. (25) Bell, D. C.; Wu, Y.; Barrelet, C. J.; Gradecak, S.; Xiang, J.; Timko, B. P.; Lieber, C. M. Microsc. Res. Tech. 2004, 64, 373. (26) Germain, V.; Li, J.; Ingert, D.; Wang, Z. L.; Pileni, M. P. J. Phys. Chem. B 2003, 107, 8717. (27) Jin, R.; Cao, Y. W.; Mirkin, C. A.; Kelly, K. L.; Schatz, G. C.; Zheng, J. G. Science 2001, 294, 1901.

J. Phys. Chem. C, Vol. 112, No. 39, 2008 15133 (28) Kirkland, A. I.; Jefferson, D. A.; Duff, D. G.; Edwards, P. P.; Gameson, I.; Johnson, B. F. G.; Smith, D. J. Proc. R. Soc. London A 1993, 440, 589. (29) Cherns, D. Phil. Mag. 1974, 30, 549. (30) Gibson, J. M.; Lanzerotti, M. Y.; Elser, V. Appl. Phys. Lett. 1989, 55, 1394. (31) Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89. (32) Kar, S.; Chaudhuri, S. J. Phys. Chem. B 2005, 109, 3298. (33) Pan, Z. W.; Dai, Z. R.; Wang, Z. L. Appl. Phys. Lett. 2002, 80, 309. (34) Wei, D. C.; Cao, L. C.; Fu, L.; Li, X. L.; Wang, Y.; Liu, Y. Q. AdV. Mater. 2007, 17, 386. (35) Hannon, J. B.; Kodambaka, S.; Ross, F. M.; Tromp, R. M. Nature 2006, 440, 69. (36) Cao, L. Y.; Garipcan, B.; Atchison, J. S.; Ni, C. Y.; Nabet, B.; Spanier, J. E. Nano Lett. 2006, 6, 1852. (37) Cai, Y.; Chan, S. K.; Sou, I. K.; Chan, Y. F.; Su, D. S.; Wang, N. AdV. Mater. 2006, 18, 109. (38) Wu, Y.; Cui, Y.; Huynh, L.; Barrelet, C. J.; Bell, D. C.; Lieber, C. M. Nano. Lett. 2004, 4, 433. (39) Yu, D. P.; Bai, Z. G.; Ding, Y.; Hang, Q. L.; Zhang, H. Z.; Wang, J. J.; Zou, Y. H.; Qian, W.; Xiong, G. C.; Zhou, H. T.; Feng, S. Q. Appl. Phys. Lett. 1998, 72, 3458. (40) Zhang, Y. J.; Zhang, Q.; Wang, N. L.; Yan, Y. J.; Zhou, H. H.; Zhu, J. J. Crystal. Growth. 2001, 226, 185. (41) Wang, Y. W.; Schmidt, V.; Senz, S.; Go¨sele, U. Nat. Nanotechnol. 2006, 1, 186. (42) Garnett, E. C.; Liang, W. J.; Yang, P. D. AdV. Mater. 2007, 19, 2946. (43) Kamins, T. I.; Williams, R. S.; Basile, D. P.; Hesjedal, T.; Harris, J. S. J. Appl. Phys. 2001, 89, 1008. (44) Ma, C.; Wang, Z. L. AdV. Mater. 2005, 17, 2635. (45) Fang, X. S.; Ye, C. H.; Zhang, L. D.; Wang, Y. H.; Wu, Y. C. AdV. Funct. Mater. 2005, 15, 63. (46) Yin, Y. D.; Zhang, G. T.; Xia, Y. N. AdV. Funct. Mater. 2002, 12, 293.

JP8038785