Thin Solid Films 514 (2006) 20 – 24 www.elsevier.com/locate/tsf
The growth of silicon nanowires by electroless plating technique of Ni catalysts on silicon substrate Jung-Fu Hsu a , Bohr-Ran Huang b,⁎ a b
Graduate School of Engineering Science and Technology, National Yunlin University of Science and Technology, Yunlin 640, Taiwan, ROC Department of Electronic Engineering, College of Engineering, National Yunlin University of Science and Technology, 123 University Road, Section 3, Touliu, Yunlin, Taiwan 640, ROC Received 25 July 2005; received in revised form 13 January 2006; accepted 7 February 2006 Available online 14 March 2006
Abstract The silicon nanowires (SiNWs) in this research were synthesized on silicon substrates via a catalytic reaction under N2 atmosphere by the thermal chemical vapor deposition system. Nickel catalyst was deposited on the silicon substrates by electroless nickel plating technique. It was found that the Ni content was increased from 0.31 wt.% (30 s, 75 nm) to 15.52 wt.% (300 s, 370 nm) from the energy dispersive X-ray spectroscopy analysis. It was also shown that the sizes of the Si–Ni alloy droplets and the growth density of SiNWs were both increased as the thickness of the electroless plating layer increased. It was concluded that the diameters, lengths and growth densities of SiNWs could be controlled by the Ni content of the electroless plating layer on the silicon substrate. © 2006 Elsevier B.V. All rights reserved. Keywords: Silicon nanowires (SiNWs); Electroless nickel plating; Si-Ni alloy droplets
1. Introduction Since the discovery of silicon nanowires (SiNWs) [1], it has been anticipated that they should exhibit potentially useful electrical, optical, mechanical and chemical properties due to their small dimensions, unique shapes and high surface-to-volume ratio. Many efforts have been made to improve the synthesis of SiNWs by employing different techniques, such as excimer laser ablation [2], chemical vapor deposition (CVD) [3] and other methods [4–6]. In most of these previous studies, the SiNWs were synthesized using metallic catalysts. However, these catalyst materials are usually prepared using expensive facilities, such as sputter coaters or evaporators, causing the cost of silicon nanowires fabrication to increase. The electroless deposition process experienced numerous modifications to meet the challenges needs of a variety of industrial applications since Brenner and Riddell invented the
⁎ Corresponding author. Tel.: +886 5 5342601x4315; fax: +886 5 5312063. E-mail address:
[email protected] (B.-R. Huang). 0040-6090/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2006.02.013
process in 1946 [7]. The electroless plating technique has many advantages [8,9], such as low temperature processing, simple process with non-expensive deposition facilities and simpler control of the composition of the deposited thin films. Recently, electroless plating technique has become one of the most attractive manufacturing methods in mass-production of nanostructures [10]. Tsai et al. [11] have investigated the catalytic effect of electroless Ni–P alloy on Si wafer for the growth of carbon nanofibers. In this work, the electroless plating technique was adapted to prepare the metal catalysts on silicon substrates for the synthesis of SiNWs by the solid–liquid–solid (SLS) mechanism method [5,12–14]. The correlation between diameters, lengths, nucleation densities of SiNWs and process of the electroless plating treatment were investigated. 2. Experimental details The substrates used were n-type (resistivity about 3–5 Ω cm) Si(100) wafers. The silicon substrates were cleaned ultrasonically in acetone and in ethanol in turn for 10 min each, and then leached with in deionized water. In order to help the Ni
J.-F. Hsu, B.-R. Huang / Thin Solid Films 514 (2006) 20–24
Chemicals
Concentration (g/l)
NiSO4·6H2O 87 24 NaH2PO2·H2O C2H2(COONa)2·6H2O 4.1 C3H4(OH)(COOH)3·H2O 2 Pb(CH3COO)2·3H2O 1.5 × 10− 3 CH3COONa·3H2O 30 Operating conditions: pH: 4.6 (adjusted with H2SO4). Bath temperature: 85 °C.
electroless plated on silicon substrates. Prior to the electroless plating process, the etched silicon substrates were first sensitized by immersion in SnCl2/HCl solution (10 g/l SnCl2 + 40 ml/l HCl) for 20 min and then activated by immersion in PdCl2/HCl solution (0.3 g/l PdCl2 + 2.5 ml/l HCl) for 5 min. For the electroless plating of the nickel, the simple bath consisted of a mixture of NiSO4, NaH2PO2, C2H2(COONa)2, C3H4(OH)(COOH)3, Pb(CH3COO)2 and CH3COONa as listed in Table 1. During plating, the bath was maintained at a temperature of 85 °C and the pH of the bath was kept constant at 4.6 adjusted with H2SO4 aqueous solution. The deposition times were varied from 30 s to 300 s. Then, for the synthesis of SiNWs, the samples were placed into a thermal CVD system under an N2 ambient pressure of approximately 2.67 × 104 Pa at 955 °C for 1 h. After the growth of SiNWs process, the samples were cooled down to 25 °C under the same N2 ambient. An energy dispersive X-ray spectroscopy (EDS) (Oxford Inca Energy 400, operated at 15 keV) was used for elemental analysis of the electroless Ni catalyst layer. A field emission scanning electron microscope (FESEM) (Jeol JSM 6700F, operated at 3–5 kV) was used for the SiNWs examination using secondary electron (SE) imaging and backscattered electron (BSE) imaging. 3. Results The energy dispersive X-ray (EDS) spectra of the electroless nickel plating on silicon substrates are shown in Fig. 1. Fig. 1(a)–(c) illustrate the EDS spectra of the deposit plating for 30 s, 180 s and 300 s, respectively. Fig. 1(a) is the EDS spectrum of 30-s Ni plating sample, in which the Si, P, S, Ni and O signals were present with a small Ni content of 0.31 wt.%. The P signal in the EDS spectrum reveals that P co-precipitated with Ni in the initial stage of deposition. In addition, the S signal is probably attributable to the NiSO4 or the pH adjusting solution H2SO4, which was also responsible for some sample oxidation. As deposition time increased, it was found that the Ni content of 180-s and 300-s Ni plating samples increased to 3.5 wt.% and 15.52 wt.%, as shown in Fig. 1(b) and (c), respectively. The Ni content was clearly increased as the deposition time increased. The thickness of the Ni electroless plating layer on the silicon substrate also were increased from approximately 75 nm to 370 nm. In other words, the Ni content increased as the thickness of the Ni electroless plating layer on the silicon substrate increased.
Fig. 2 shows secondary electron (SE) images of the SiNWs grown on the Si substrate. Fig. 2(a)–(c) illustrate the growth morphologies of the SiNWs with deposit plating for 30 s, 180 s and 300 s, respectively. Fig. 2(a) is the SE image of 30-s Ni plating sample, which shows short length and a low growth density of the SiNWs. The average length and growth density of SiNWs were both increased as the electroless plating time increased, as shown in Fig. 2(b) and (c), for 180-s and 300-s Ni plating samples, respectively. The detailed properties of the SiNWs were summarized in Table 2. 30-s Ni plating sample showed that the SiNWs have average diameter ∼ 8 nm, average length < 1 μm and low growth density; 180-s Ni plating sample showed that the SiNWs have average diameter ∼ 14 nm, average length <10 μm and medium growth density; and 300-s Ni plating sample showed that the SiNWs have average diameter ∼ 20 nm, average length > 10 μm and the highest growth density. Backscattered electron (BSE) images do not offer as many morphological details on contrasting domains as the secondary electron (SE) images, but they can provide more explicit information on chemical differences between different domains in a sample. The contrast of a BSE image depends on the backscattered electron generation rate, which increases with the mean atomic number of the specimen [15]. Fig. 3(a) and (b) illustrate SE images with BSE images of 30-s and
Elements
(a)
O Ni
Wt% 16.6 82.6 0.3 0.2 0.3
O Si P S Ni
Si
Intensity (arb. units)
Table 1 Chemical compositions and operating conditions of bath
21
PS
Ni
Ni
Elements Si
Ni
19.9 72.6 2.5 1.5 3.5
O Si P S Ni
(b)
O
Wt%
P S
Ni Ni
O
Elements
Si P
Wt% 35.7 37.0 8.4 3.4 15.5
O Si P S Ni
(c)
Ni S
Ni Ni
0
1
2
3
4
5
6
7
8
9
10
Energy (keV) Fig. 1. EDS spectra of the electroless nickel plating on silicon substrates for the deposition time of (a) 30 s, (b) 180 s and (c) 300 s.
22
J.-F. Hsu, B.-R. Huang / Thin Solid Films 514 (2006) 20–24
This indicated that the Si–Ni alloy droplets were present on both of the 30-s and 300-s Ni plating samples. It was observed that the average diameter of the Si-Ni alloy droplets were 30–100 nm and 100–300 nm for 30-s Ni plating sample and 300-s Ni plating sample, respectively.
(a)
(a) 1µm
(b)
100 nm SE
1µm
(c)
100 nm BSE
(b) 1µm
Fig. 2. SE images of the silicon nanowires for the (a) 30-s, (b) 180-s and (c) 300-s Ni plating samples, respectively.
300-s Ni plating samples, respectively. Fig. 3(a) shows the SE images along with the corresponding BSE image on the same area for the surface of 30-s Ni plating sample. It was found that the brightness of different regions in the SE image differs from that, in their BSE counterparts, the brighter regions in the BSE image possibly representing the Ni content (the mean atomic number of Ni is double that of Si). A similar situation was observed for 300-s Ni plating sample in Fig. 3(b) with a higher Ni content on top of the SiNWs.
1µm SE
Table 2 Summary of SiNW properties for 30-s, 180-s and 300-s Ni plating samples
1µm
SiNWs properties
Samples 30 s
180 s
300 s
Growth density Average diameter Average length
Low ∼ 8 nm <1 μm
Mid ∼ 14 nm <10 μm
Highest ∼ 20 nm >10 μm
BSE Fig. 3. Secondary electron (SE) images (top) and backscattered electron (BSE) images (bottom) taken in the same region for the (a) 30-s Ni plating sample and (b) 300-s Ni plating sample.
J.-F. Hsu, B.-R. Huang / Thin Solid Films 514 (2006) 20–24
23
Growing process (955 °C)
4. Discussion In the previous work for SiNWs synthesis, laser ablation [2,16] or chemical vapor deposition [3] followed VLS growth mechanism [1,17,18], where the catalysts (usually Ni, Au or Fe as impurity) act as a liquid-forming agent, which reacts with the vapor phase atomic Si, and forms the metal silicide eutectic liquid droplets. With the further absorption of Si atoms into the droplets from the vapor phase, the droplets become supersaturated, resulting in the precipitation of SiNWs from the droplets. Yan et al. [5] have reported the SLS growth mechanism, in which the Ni film can react with the Si substrate at high temperature, and formed Si2Ni eutectic liquid alloy droplets. Because of the relatively high solubility of Si in Si2Ni eutectic alloy, more Si atoms will diffuse through the solid (the substrate)–liquid interface into the liquid phase (the Si2Ni droplets). The next liquid–solid (nanowire) interface will be formed when the liquid phase becomes supersaturated due to thermal or compositional fluctuations, resulting in the growth of SiNWs. In this work, the electroless plating method was used to prepare the Ni catalysts layer on the silicon substrate with the synthesis SiNWs by the SLS growth mechanism [5,12–14]. The EDS analysis results indicated not only the Ni but also other elements, such as P and S exist in the plating layer. As shown in Fig. 4, the growth process is depicted as follows: (I) plating of a Ni catalysts layer on the Si substrate; (II) other elements were vaporized at high temperature when the Si–Ni alloy eutectic liquid droplets were formed; (III) more Si atoms will diffuse through the substrate–liquid interface into the liquid droplets and growth of SiNWs through the liquid–wire interface. That means the alloy droplet will remain attached to the silicon substrate. However, in Figs. 2(c) and 3(b), it was shown that the Si–Ni alloy droplets were present on top of nanowires for 300-s Ni plating sample. And, the growth temperature of nanowires at 955 °C was lower than the melting point of bulk silicon. It was also known that the nanometer-scale amorphous Si would be vaporized above 1000 °C [19,20]. Therefore, it was suggested that the Si–Ni alloy droplets might be detached from the silicon substrate by the thermal stresses for the long length and high
(I)
(II)
Electroless Ni plating layer
Ni-Si alloy
(III) Growth of SiNWs
Si substrate Diffusion of the Si atoms Fig. 4. Schematic representation for the growth of SiNWs by electroless plating technique of Ni catalysts on silicon substrate.
Si substrate
Cooling process (Down to 25 °C)
Si substrate Fig. 5. Schematic representation for the growth of SiNWs during the cooling process. The inset figure shows a SE image of the Si–Ni alloy droplets on top of nanowires for 300-s Ni plating sample.
growth density of SiNWs (Fig. 5) during the cooling process (down to 25 °C). Inset of Fig. 5 shows a SE image of the Si–Ni alloy droplets on top of nanowires for 300-s Ni plating sample. In previous studies, sputtering and thermal evaporating methods are the most widely used metal catalyst preparation techniques in the synthesis of SiNWs. The role of the metal catalyst is not only to form a liquid alloy droplet of relatively low solidification temperature but also controlled the diameters and growth densities of nanowires. Paulose et al. [14] reported that the growth of SiNWs by heating gold-coated silicon substrate, which was prepared by thermal evaporation. With increasing gold layer thickness, the nucleation density of the Au–Si islands is expected to decrease with increasing grain size [21]. It was also found that the growth density of nanowires increased when the metal catalyst thickness was reduced. However, in this study, it was found that not only the Ni but also others, such as P and S, were also existed in the electroless Ni plating layer. In other words, if in the same thickness, the content of Ni by the electroless Ni plating technique would be less than that by the sputtering or thermal evaporating technique. In this work, 30-s Ni plating sample shows a low growth density of SiNWs, which has only one nanowire grown on each Si–Ni alloy droplet (Fig. 3(a)). Moreover, Fig. 3(b) shows 300-s Ni plating sample has a higher growth density of SiNWs, with several nanowires grown on each Si–Ni alloy droplet. A much smaller size (about 30–300 nm) of the Si–Ni alloy droplets was achieved when the electroless Ni plating technique was used. From the EDS analysis, the Ni content was
24
J.-F. Hsu, B.-R. Huang / Thin Solid Films 514 (2006) 20–24
increased as the thickness of the electroless plating layer increased from 0.31 wt.% (∼ 75 nm) to 15.52 wt.% (∼ 370 nm). It was indicated that the growth density of SiNWs was increased as the thickness of the electroless plating layer increased since the Ni content increased. Therefore, it was then concluded that the diameters, lengths and growth densities of SiNWs were controlled by the Ni content of the electroless plating layer on the silicon substrate. However, more detailed characterization and research on the growth mechanism need to be carefully studied in the future research.
ing of National Chung Hsing University for helping on the measurements of the EDS spectra and the SEM images. This research was supported by the National Science Council of ROC under grant nos. NSC 94-2216-E-224-002 and NSC 942120-M-224-001. References [1] [2] [3] [4]
5. Conclusions
[5]
In this research, SiNWs were successfully synthesized using the electroless plating technique to prepare Ni catalysts on silicon substrates with the solid–liquid–solid growth mechanism. From the EDS analysis, it was indicated that the growth density of SiNWs was increased as the thickness of the electroless plating layer increased. Backscattered electron images also showed that 300-s Ni plating sample had higher Ni content in the Si–Ni alloy droplets than that of 30-s Ni plating sample. Thus, it was then concluded that the diameters, lengths and growth densities of SiNWs were controlled by the Ni content of the electroless plating layer on the silicon substrate. The electroless Ni plating technique has the intrinsic advantages of simplicity and the potential of large-scale production at low cost.
[6]
Acknowledgments The authors would like to acknowledge Dr. Gordon Turner Walker for his precious opinions. The authors would also like to acknowledge Mr. Lu in the Department of Materials Engineer-
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
R.S. Wagner, W.C. Ellis, Appl. Phys. Lett. 4 (1964) 89. A.M. Morales, C.M. Lieber, Science 279 (1998) 208. S.W. Chung, J.Y. Yu, J.R. Heath, Appl. Phys. Lett. 76 (2000) 2068. N. Wang, Y.H. Tang, Y.F. Zhang, C.S. Lee, I. Bello, S.T. Lee, Chem. Phys. Lett. 299 (1999) 237. H.F. Yan, Y.J. Xing, Q.L. Hang, D.P. Yu, Y.P. Wang, J. Xu, Z.H. Xi, S.Q. Feng, Chem. Phys. Lett. 323 (2000) 224. C.P. Li, X.H. Sun, N.B. Wong, C.S. Lee, S.T. Lee, B.K. Teo, Chem. Phys. Lett. 365 (2002) 22. A. Brenner, G.E. Riddell, J. Res. Natl. Bur. Stand. 37 (1) (1946) 31. M.D. Feldstein, Plat. Surf. Finish. 85 (1998) 248. X. Haowen, Z. Bangwei, J. Mater., Prog. Technol. 124 (2002) 8. Y. Shacham-Diamad, Y. Sverdlov, Microelectron. Eng. 50 (2000) 525. T.K. Tsai, C.C. Chuang, C.G. Chao, W.L. Liu, Diamond Relat. Mater. 12 (2003) 1453. D.P. Yu, Y.J. Xing, Q.L. Hang, H.F. Yan, J. Xu, Z.H. Xi, S.Q. Feng, Physica, E 9 (2001) 305. Y.J. Xing, D.P. Yu, Z.H. Xi, Z.Q. Xue, Appl. Phys., A Mater. Sci. Process. 76 (2003) 551. M. Paulose, O.K. Varghese, C.A. Grimes, J. Nanosci. Nanotech. 3 (2003) 341. C.A.P. Leite, F. Galembeck, J. Colloid Interface Sci. 235 (2001) 4. E.I. Givargizov, J. Cryst. Growth 31 (1975) 20. G.A. Boostma, H.J. Gassen, J. Cryst. Growth 10 (1971) 223. 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. P. Buffat, J.P. Borel, Phys. Rev., A 13 (1976) 2287. J.P. Borel, Surf. Sci. 106 (1981) 1. J.S. Wu, S. Dhara, C.T. Wu, K.H. Chen, Y.F. Chen, L.C. Chen, Adv. Mater. 14 (2002) 1847.