Download Materials science and engineering a 286(2000)16 23 sc nw synthesis, structure properties lee wang and more Lecture notes Manufacturing Processes in PDF only on Docsity!
Materials Science and Engineering A286 (2000) 16–
Semiconductor nanowires: synthesis, structure and properties
S.T. Lee * , N. Wang, C.S. Lee
Center Of Super - Diamond and Ad 6 anced Films ( COSDAF ) and Department of Physics and Materials Science , City Uni 6 ersity of Hong Kong , 83 Tat Chee A 6 enue , Kowloon Tong , Hong Kong , China
Abstract
Highly pure, ultra long and uniform-sized semiconductor nanowires in bulk-quantity have been synthesized by novel methods of laser ablation and thermal evaporation of semiconductor powders mixed with metal or oxide catalysts. Transmission electron microscopic study shows that decomposition of semiconductor sub-oxides and the defect structure play an important role in enhancing the formation and growth of high-quality semiconductor nanowires. The morphology, microstructure, optical and electrical properties of the nanowires have been characterized systematically by Raman scattering, photoluminescence and field emission. A new growth mechanism, namely oxide-assisted growth, is proposed based on the microstructure and different morphologies of the nanowires observed. © 2000 Elsevier Science S.A. All rights reserved.
Keywords : Silicon; Nanowire; Electron microscopy; Nanostructured materials
www.elsevier.com/locate/msea
1. Introduction
Using geometry scaling to improve the performance/ cost (including power consumption) ratio of electronic devices has been the objective of semiconductor indus- try. The quantum effect associated with nano-scale devices is an important aspect of microelectronic device design. In recent years, semiconductor wires with nanometer width have attracted much attention be- cause of their potential applications in mesoscopic re- search and nanodevices. Since the 1960s, Si whiskers grown from the well-known vapor–liquid–solid (VLS) reaction [1,2] have been extensively investigated. In the VLS reaction, metal particles, for example Au particles on Si substrate, are generally used as the mediating solvent. This is because Au and Si form molten eutectic alloy droplets at a relatively low temperature. Si in the vapor phase diffuses into the liquid alloy droplets and bonds to the solid Si at the liquid–solid interface. This reaction results in the growth of Si whiskers. The diameter of the whisker grown by this technique is determined by the diameter of the alloy droplet at its tip. It has been reported that single crystal Si whiskers
grew along é 111 è direction epitaxially on Si(111) sub- strates by VLS [1–3]. In different material systems, however, a variety of semiconductor whiskers was ob- tained, for example, GaP whiskers [4,5] displayed rota- tional twins around their é 111 è growth axes, while GaAs [6–8] whiskers grew in the form of wurtzite structure. The synthesis of one-dimensional nanostruc- tured materials in bulk-scale remains a challenge. In recent years, many efforts have been made to synthesize Si nanowires by employing different methods, such as the photolithography technique [9–11] and scanning tunneling microscopy [12,13]. Of particular interest is the recently developed method, namely, laser ablation of metal-containing semiconductor targets [14–16], by which bulk-quantity semiconductor nanowires can be readily obtained. Our recent studies [17–20] show that oxides play a dominant role for the nucleation and growth of high quality semiconductor nanowires in bulk-quantity by laser ablation, thermal evaporation or chemical vapor deposition. A new growth mechanism named oxide-assisted nanowire growth is therefore pro- posed. The ability to synthesize large quantities of highly pure (contamination-free), ultra long (in millime- ters) and uniform-sized semiconductor nanowires from this new technique offers exciting possibilities in funda- mental and applied research.
- Corresponding author. Fax: +852-27887830. E - mail address : apannale@cityu.edu.hk (S.T. Lee)
0921-5093/ 00 /$ - see front matter © 2000 Elsevier Science S.A. All rights reserved. PII: S 0 9 2 1 - 5 0 9 3 ( 0 0 ) 0 0 6 5 8 - 4
2. Nanowires from laser ablation
Compared with the classical VLS method, nanoparti- cles of metal or metal silicide in large quantity are rather easy to obtain from the high-temperature laser ablation method using metal-containing Si target. As- sisted by laser ablation, these nanoparticles can act as the critical catalyst needed for the nucleation and growth of nanowires. Si nanowires, for example, have been synthesized by this technique [14–16]. The typical experiment was carried out by using an excimer laser to ablate the target in an evacuated quartz tube filled with Ar gas (see Fig. 1) [14]. The temperature around the target was in the range between 1200 and 1400°C. The solid target could be highly pure Si powder mixed with metals (Fe, Ni, or Co). The temperature of the area around the substrate on which the nanowire grew was 900–1100°C. Our transmission electron microscopy (TEM) investi- gations showed that Si nanowires obtained by this method were extremely long and highly curved with a typical diameter of 20 nm (see Fig. 2a). Each wire consisted of an outer layer of Si oxide and a crystalline Si core. A high density of defects, such as stacking faults and micro-twins, has been observed in the crys- talline Si core. As identified by electron diffraction and high-resolution transmission electron microscopy (HRTEM), the axis of the nanowires was generally along é 112 è direction and the {111} surfaces of Si crystalline cores were parallel to the axis of the nanowire [17]. It is widely acknowledged that for VLS growth mech- anism, metal or metal silicide particles should exist at the tips of Si nanowires and dictate the one-dimen- sional growth phenomenon of nanowires. However, our extensive HRTEM investigation found no evidence of metals in any of the samples synthesized from all targets used (Fe-, Ni- and Co-containing targets). The chemical composition of Si nanowires was studied by electron energy dispersive spectroscopy (EDS) in TEM and the results showed only the presence of silicon and oxygen in the nanowires. Our extensive investigations Fig. 2. (a) TEM micrograph of Si nanowires; (b) the morphology; and (c) a typical HRTEM image of the Si wires formed at the high temperature zone.
Fig. 1. Experimental set-up for the synthesis of Si nanowires by laser ablation.
revealed that Si nanowire tips were generally round and covered by a relatively thick Si oxide layer (2–3 nm). The Si crystal core near the tip contained a high density of stacking faults and micro-twins [19]. Again, the stacking-faults were generally along the axis of the nanowire in é 112 è direction and were consistent with those defects observed previously in Si nanowires. No metal was detected in the tips regardless of the type of
Table 1 Summary of Si nanowire forms and their formation mechanisms
Two-step growth model Spring Fishbone Frog egg Necklace
Nucleation Monocentric Polycentric
Growth Periodic stable Periodic unstable
SiO 2 -containing Si powder targets were used. Fig. 3 shows the product yield of Si nanowire obtained from SiO2-containing targets is always higher and up to 30 times (at 50 wt.% SiO 2 ) greater than that generated from a metal-containing target (0.03 mg h−^1 ). No nanowires were obtained from a pure SiO 2 target. The possible contributing factors to Si nanowire growth are: (1) oxygen atoms desorbed from SiO 2 during laser ablation at high temperature may contribute to the growth; (2) different species of Si oxides may have been formed in the presence of oxygen; (3) SiO 2 may change the ability of the target to absorb energy from the laser beam; and (4) reaction of Si and SiO 2 to form oxides in vapor phase. We therefore investigated systematically the influence of oxygen on the formation and growth of Si nanowires. When a pure Si target was ablated in an oxygen–argon atmosphere (minimum ratio of O:Ar= 1:50), no formation of Si nanowires was observed. To ascertain the necessity of SiO 2 for the nucleation and growth of nanowires, the following experiment was performed. The laser was first used to ablate a SiO 2 - containing Si target to form Si nanowire nuclei and to initiate the growth process. Afterwards, the laser was directed onto a pure Si target to induce further growth. However, no additional growth of Si nanowires was observed, indicating that the growth was not VLS. Because in the VLS mechanism once the molten eutec- tic metal–Si alloy is present at the tip of Si nanowire and acts as the catalyst, it would continue to catalyze Si nanowire growth as long as the supply of Si vapor source continues. Clearly, in the laser ablation method, SiO 2 was continually needed throughout the process of Si nanowire nucleation and growth. In the present experiment, the vapor phase generated from the mixture of Si and SiO 2 at 1200°C mainly consisted of Si monoxide (Si (s)+SiO 2 (s)ì2SiO (g), where (s) and (g) represent solid and gas, respectively). This was proven by the EDS observation that the material collected on the water-cooled Cu finger was Si m O n ( m =0.51, n =0.49). Si monoxide (SiO) is an amorphous semiconductor of high electrical resistivity, which can be readily generated from the powder mix- ture (especially in 1:1 ratio) of Si and SiO 2 by heating [21–23]. By heating the Si monoxide sample, Si precip- itation was observed. Such precipitation of Si nanopar- ticles from annealed Si monoxide is quite well-known
[23]. According to the above observations, we propose that the growth mechanism of Si nanowires is oxide-as- sisted. The vapor phase of Si x O ( x \1) generated by thermal evaporation or laser ablation is the key factor. The precipitation, nucleation and growth of Si nanowires always occurred at the area near the cold finger, which suggested that the temperature gradient provided the external driving force for nanowire forma- tion and growth. The nucleation of nanoparticles is assumed to occur on the substrate by different decom- positions of Si oxide as shown in Eqs. (1) and (2).
Si x O (s)ìSi x − 1 (s)+SiO (s) ( x \1) (1)
and
2SiO (s)ìSi (s)+SiO 2 (s) (2)
Our TEM results suggested that these decompositions result in the precipitation of silicon nanoparticles, which are the nuclei of Si nanowires, clothed by shells of silicon oxide. The oxide-assisted Si nanowire growth found direct support in the experiment where highly pure SiO pow- der was used as the solid source. The morphology and structure of Si nanowires obtained from SiO were simi- lar to those grown from Si+SiO 2 solid source. The yield of Si nanowires increased with increasing thermal evaporation temperature and pressure (see Fig. 4). Us- ing highly pure Si monoxide powders, we obtained a high yield of Si nanowires at temperatures ranging from
Fig. 7. (a) TEM micrograph showing that the Si crystals in the knots of the Si nanowire chain have different orientations. The arrows show the é 112 è direction.
Fig. 8. (a) Raman spectra taken from the as-grown Si nanowires, Si monoxide and fully oxidized Si nanowires. PL spectra taken from the as-grown Si nanowires, Si monoxide and fully oxidized Si nanowires.
at 830°C (Morales and Lieber [15] used metal-contain- ing targets). TEM studies showed that the morphology of Ge nanowires is similar to that of Si nanowires. Each Ge nanowire also consisted of a crystalline Ge core and a thick amorphous oxide shell. This may be expected since Ge and Si have similar structures and properties. Unlike Si nanowires, the diameters of Ge nanowires are not uniform (ranging from 16 to 367 nm).
Fig. 9. I – V results from (a) Si and (b) SiC nanowires. The insets are Fowler–Nordheim plots of ln( I / E^2 ) against (1/ E ).
1130 to 1400°C. This provided the direct evidence for the oxide-assisted growth mechanism. Laser ablation of the stable phase SiO 2 did not produce Si nanowires, but SiO 2 powder mixed with Ge produced Si nanowires at 1200°C. This was due to the following reaction
SiO 2 (s)+Ge (s)ìSiO (g)+GeO (g) (3)
This reaction also offers the possibility to grow Si or Ge nanowires using the same target. Similar to the production of Si nanowires, we used a GeO 2 -containing Ge target (to generate GeO) to synthesize Ge nanowires
consists of a row of knots and necks with approxi- mately equal distances between them. Finally, the for- mation of the frog egg-shaped and the pearl-shaped Si nanowires can also be explained by the two-step model of the VLS growth mechanism of polycentric nucle- ations. The black dots of the frog egg-shaped and necklace-shaped Si nanowires are the sites of nuclei. Periodic unstable growth makes the diameter of the Si core in the necklace-shaped nanowire change regularly while periodic stable growth gives the frog egg-shaped SINW possess a continuously uniform Si core. The nucleation sites on the frog egg-shaped nanowire will keep growing to become strings of Si nanowires. Fig. 6d exhibits that some nanowires grow from the nucle- ation sites (pointed by arrows) of the nanowires. The high-resolution TEM image shown in Fig. 7 supports the mechanism proposed. Each Si knot connected by Si oxide necks has its own crystalline orientation and its é 112 è axis is generally not parallel to the growth direction. The Raman spectrum of Si nanowires (Fig. 8a) shows a broad and symmetric peak at 521 cm−^1 compared to that of a bulk single crystal Si. The peak profile may be associated with the effect of the small-sized Si nanocrystals or defects. The presence of nonstoichio- metric Si sub-oxide may also contribute to the peak asymmetry. For comparision, the spectrum from a Si monoxide film contains a broad peak at about 480 cm−^1 , while that of the fully oxidized Si nanowires (prepared by annealing in air) shows no Raman scatter- ing (Fig. 8a). Si monoxide has a strong photoluminescence (PL) at 740 nm, while the oxidized nanowire gives a weak PL peak at 600 nm (Fig. 8). However, the PL from the Si nanowire product is weak and complicated. A typical PL spectrum from Si nanowires covers the range of 600–800 nm. Clearly, the Si monoxide and Si sub-oxide components in the nanowires are the main contributors to this spectrum. The SiO generated by thermal evapo- ration is indeed a mixture of various oxides of Si. Si nanoparticles also co-exist with the SiO generated. It is well-known that nanotubes and nanowires with sharp tips are promising materials for applications as cold cathode field emission devices. Fig. 9 shows the results of current versus voltage ( I – V ) with the inset being the Fowler–Nordheim representation of the data. Both Si and SiC nanowires exhibit well-behaved and robust field emission. The turn-on fields for Si and SiC nanowires were 15 and 20 V mm−^1 , respectively [27,28] and current density of 0.01 mA cm−^2 which are com- parable with those for other field emitters including carbon nanotubes and diamond. The field emission characteristics may be improved by further optimiza- tion, such as oriented growth or reducing the oxide shells and will be most promising for applications.
6. Summary
A novel method based on oxide-assisted growth has been developed that is capable of producing high-qual- ity and bulk-quantity of various semiconductor nanowires including Si, Ge, amorphous carbon and SiC. The presence of oxides in the target is a common and essential ingredient for the synthesis using laser ablation or thermal evaporation, so that the targets are capable of generating semiconductor oxides in the va- por phase. Subsequent decomposition of the vapor phase oxides at high temperature plays a crucial role in the nucleation and growth of high-quality nanowires. Si (or Ge) nanowires can be synthesized from metal- or oxide-containing Si (or Ge) targets or directly from Si monoxide powders by laser ablation or thermal evapo- ration. The nanowires grown by metal catalyst and oxide show significant difference in their morphology and microstructure. Especially, metal catalyst-assisted growth results in the growth of thick Si wires at a higher temperature condition. Semiconductor nanowires show unusual optical and field emission characteristics, which may be exploited for potential applications.
Acknowledgements
We thank Y.F. Zhang, Y.H. Tang, I. Bello, H.Y. Peng, F.C.K. Au and K.W. Wong for their contribu- tion to the work presented in this review. This work was supported by Strategic Research Grants of the City University of Hong Kong and the Research Grant Council of Hong Kong.
References
[1] R.S. Wagner, W.C. Ellis, Appl. Phys. Lett. 4 (1964) 89. [2] E.I. Givargizov, J. Cryst. Growth 32 (1975) 20. [3] G.A. Boostma, H.J. Gassen, J. Cryst. Growth 10 (1971) 223. [4] M. Gershenzo, R.M. Mikulyak, J. Electrochem. Soc. 108 (1961)
[5] N. Holonyak Jr, C.M. Wolfe, J.S. Moore, Appl. Phys. Lett. 6 (1965) 64. [6] E.N. Laverko, V.M. Marakhonov, S.M. Polyyakov, Sov. Phys. Crystallogr. 10 (1966) 611. [7] K. Hiruma, T. Katsuyama, K. Ogawa, G.P. Morgan, M. Koguchi, H. Kakibayashi, Appl. Phys. Lett. 59 (1991) 431. [8] K. Hiruma, M. Yazawa, T. Katsuyama, K. Ogawa, K. Haraguchi, M. Koguchi, H. Kakibayashi, J. Appl. Phys. 77 (1995) 447. [9] H.I. Liu, N.I. Maluf, R.F.W. Pease, J. Vac. Sci. Technol. B (1992) 2846. [10] H. Namatsu, S. Horiguchi, M. Nagase, K. Kurihara, J. Vac. Sci. Technol. B15 (1997) 1688. [11] Y. Wada, T. Kure, T. Yoshimura, Y. Sudou, T. Kobayashi, Y. Gotou, S. Kondo, J. Vac. Sci. Technol. B12 (1994) 48. [12] T. Ono, H. Saitoh, M. Esashi, Appl. Phys. Lett. 70 (1997) 1852.
[13] R. Hasunuma, T. Komeda, H. Mukaida, H. Tokumoto, J. Vac. Sci. Technol. B15 (1997) 1437. [14] Y.F. Zhang, Y.H. Zhang, N. Wang, D.P. Yu, C.S. Lee, I. Bello, S.T. Lee, Appl. Phys. Lett. 72 (1998) 1835. [15] A.M. Morales, C.M. Lieber, Science 279 (1998) 208. [16] D.P. Yu, Z.G. Bai, Y. Ding, Q.L. Hang, H.Z. Zhang, J.J. Wang, Y.H. Zou, W. Qian, G.C. Xiong, H.T. Zhou, S.Q. Feng, Appl. Phys. Lett. 72 (1998) 3458. [17] N. Wang, Y.H. Tang, Y.F. Zhang, D.P. Yu, C.S. Lee, I. Bello, S.T. Lee, Chem. Phys. Lett. 283 (1998) 368. [18] N. Wang, Y.F. Zhang, Y.H. Tang, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 73 (1998) 3902. [19] N. Wang, Y.H. Tang, Y.F. Zhang, C.S. Lee, S.T. Lee, Phys. Rev. B 58 (1998) 16024.
[20] N. Wang, Y.H. Tang, Y.F. Zhang, C.S. Lee, I. Bello, S.T. Lee, Chem. Phys. Lett 299 (1999) 237. [21] S.W. Roberts, G.J. Parker, M. Hempstead, Opt. Mater. 6 (1996)
[22] U. Setiowati, S. Kimura, J. Am. Ceram. Soc. 80 (1997) 757. [23] G. Hass, C.D. Salzberg, J. Opt. Soc. Am. 44 (1954) 181. [24] Ph. Buffat, J.P. Borel, Phys. Rev. A 13 (1976) 2287. [25] J.P. Borel, Surf. Sci. 106 (1981) 1. [26] Y.H. Tang, Y.F. Zhang, N. Wang, I. Bello, C.S. Lee, S.T. Lee, J. Appl. Phys. 85 (1999) 7981. [27] K.W. Wong, X.F. Zhou, F.C.K Au, H.L. Lai, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 75 (1999) 2918. [28] F.C.K. Au, K.W. Wong, Y.H. Tang, Y.F. Zhang, I. Bello, S.T. Lee, Appl. Phys. Lett. 75 (1999) 1700.
.
