NANO LETTERS
Self-Assembled Boron Nanowire Y-Junctions
2006 Vol. 6, No. 3 385-389
Sang H. Yun* Department of Physics and Astronomy, UniVersity of Kansas, Lawrence, Kansas 66045, and Department of Microelectronics and Information Technology, Royal Institute of Technology, Kista, SE-164 40, Sweden
Judy Z. Wu* and Alan Dibos Department of Physics and Astronomy, UniVersity of Kansas, Lawrence, Kansas 66045
Xiaodong Zou Department of Structural Chemistry, Stockholm UniVersity, Stockholm, SE-106 91, Sweden
Ulf O. Karlsson Department of Microelectronics and Information Technology, Royal Institute of Technology, Kista, SE-164 40, Sweden Received October 29, 2005; Revised Manuscript Received January 1, 2006
ABSTRACT In this work, we demonstrate that boron nanowire Y-junctions can be synthesized in a self-assembled manner by fusing two individual boron nanowires grown inclined toward each other. We show that the presence of a second liquid, in addition to the liquid Au catalyst, is critical to the inclination of the boron nanowire. The structure of the BNYJ arrays that we report here may allow construction of three- or multipleterminal nanowire devices directly on Si-based readout circuits through controlled nanowire growth.
Nanowires are attractive building blocks for functional nanoscale electronics, such as diodes, transistors, and Josephson junctions, if these nanowires can be assembled in a highly controllable fashion. To date, many individual semiconductor nanowires have already been configured into device structures with novel and exciting device performance.1-4 This, however, has presented a major challenge in the development of nanowire electronics. Among the other schemes developed so far, fabrication of nanowire junctions via direct growth can provide unique advantages in generating large-area monolithic integrated nanowire electronics on Si-based readout circuits. In this paper, we investigate the fusion of boron nanowires (BNWs) grown on Si substrates. Boron and boron-rich icosahedral compounds have potentials for high-temperature electronics because of their unique chemical and physical properties, such as chemical inertness, hardness, and adjustable energy band-gap.5-7 Although boron has a low electrical * Corresponding authors. Sang H. Yun, e-mail:
[email protected]; tel: +46-8-7904385; fax: +46-8-7527782. Judy Z. Wu, e-mail:
[email protected]; tel: (785) 864-3240; fax (785) 864-5262. 10.1021/nl052138r CCC: $33.50 Published on Web 02/15/2006
© 2006 American Chemical Society
conductivity in the semiconductor regime,8 high conductivity in the metal regime may be achievable in nanostructures of layered, tubular, and fullerene-like solids.9,10 The recent discovery11 of superconductivity above 39 K in MgB2 has further intensified the interest in boron-related materials. A key step in fusing two nanowires is to make the two nanowires grow inclined toward each other at large inclination angles. Motivated by this, we have recently developed an oxide-assisted vapor-liquid-solid (VLS) process for growth of inclined BNWs.12 The VLS process13 has been used widely for fabrication of single nanowires of various semiconductors.14-20 It generally employs a low-melting metal catalyst that provides a solid/liquid alloy interface for nucleation and anisotropic crystal growth of semiconductor nanowires. The nucleation and growth mechanism of nanowires may be altered greatly when a second liquid is introduced into the solid/liquid alloy interface.12 In this oxide-assisted VLS process, the low-melting B2O3 (with a melting temperature of ∼450 °C) was present during the BNW growth. The interplay between the low-melting B2O3 with the Au liquid metallic catalyst resulted in inclination of BNWs at
Figure 1. (A) SEM image of BNYJs grown on a silicon substrate coated with 10-nm-thick gold film. Scale bar: 1 µm. A mixture of pure boron and boron oxide was used as the vapor source. Inset: fusing point of two bottom BNW stems. (B) Elemental EDS line scans across a BNYJ removed from the substrate. O peaks were detected on the wall and at the merging point of two bottom stems. Scale bar: 50 nm.
large angles up to 50-60°. This provides an essential condition for fusing BNWs. In this article, we demonstrate the synthesis of boron nanowire Y-junctions (BNYJs) on Si substrates using this oxide-assisted VLS process by fusing the two BNWs inclined toward each other. Single-crystal Si(100) substrates were sputter-Au-coated in a range of 5-20 nm thickness. The source materials were mixed at a composition of boron and 40 wt % of B2O3 and pressed into a pellet. The source pellet and substrate were placed into a quartz tube. The source was in the higher temperature zone while the substrate, in a slightly lower one to allow condensation of the decomposed boron vapor on the substrate. The quartz tube was evacuated to about 10 mTorr, torch-sealed, and then heated to the processing temperature (Ts for source) in the range of 1100-1200 °C. The temperature difference between the source and the substrate was maintained in the range of 100-150 °C during processing. A field-emission scanning electron microscopy (FESEM, LEO 1550) image of a representative BNYJ is shown in Figure 1A. A large number of BNYJs were formed on the 386
substrates. Most of them have three branches with two at the bottom connected to the substrate and one on top stemmed from the fusing point of the two bottom ones. Overall, they looked like an inverted “Y”. It should be mentioned that the inverted Y-junctions have been achieved earlier in carbon nanotubes,21,22 while this work represents the first in nanowires. After joining, a BNW continued to grow from the joint and turned into the top branch of the BNYJ. The diameter of the BNWs in the BNYJs is in the range of 30-130 nm depending on the thickness of the Au catalyst layer. Most top branches of BNYJs were terminated with Au catalytic end-tips of either semispherical or spherical shapes (see Figure 1A and its inset), suggesting that the BNYJs grew from the supersaturated liquid B-Au phase via the VLS process. Elemental line scans were collected on BNYJ samples using energy-dispersive X-ray spectroscopy (EDS). Figure 1B shows an elemental scanning profile of an isolated BNYJ. Several elements were identified including B, O, and Si. Each element was then profiled on four lines across the BNYJ (Figure 1B). The first line (from left to right) was scanned directly across the junction (points 1 through 10), and the next three lines (points 11-25) were scanned on the stem of the BNYJ. B was distributed uniformly across the stem, whereas O displayed a strong peak near the surface of the upper part of the BNW stem (points 15, 20 and 25) and also at the junction (point 6) where the two bottom branches joined. However, no strong O peaks were observed from the bottom surface of the BNW stem due to a slight e-beam shift away from the BNW surface, which is not surprising because the oxide sheath thickness was found to be in the range of few nanometers.23-27 Finally, small traces of Si of uniform distribution were also detected in almost all BNWs grown on Si substrates, which might be due to Si diffusion at the high processing temperatures and is believed to be irrelevant to the formation of BNYJs because the same was also observed in single BNWs grown at different conditions for single BNW growth. Figure 2 shows transmission electron microscopy (TEM, JEOL-3010) images of several BNYJs removed from the Si substrate with distinctively different fusing angles. The fusion angle θ ≈ 20° in Figure 2A is nearly the smallest angle observed in this experiment. After fusion, the top branch of this BNYJ grew more or less symmetrically along the direction bisecting θ. This implies that the inclination angle for the two bottom BNWs is ∼θ/2 ) 10°. In Figure 2B and C, asymmetric growth of the top branch was observed with the top branch more aligned with one of the two bottom branches. In both of these cases, the θ values are much larger ranging from 60° (Figure 2B) to 105° (Figure 2C). The dimension of the top branch of the BNYJ also varies, with some slightly bigger than the larger of the two bottom branches (Figure 2A and C) and some nearly twice the average of the bottom branches (Figure 2B). Although it is difficult to predict the morphology of the Au catalyst tip before fusion, the change of the BNW diameter after fusion may be explained using the following argument. In VLS growth, the diameter of the nanowire is determined by that of the metal catalyst tip. Because both semispherical and Nano Lett., Vol. 6, No. 3, 2006
Figure 2. Low-magnification bright-field TEM images of three BNYJs with distinctively different angles of fusion: (A) 20°, (B) 60°, and (C) 105°. After fusion, the top branch of the BNYJ grows either symmetrically along the direction bisecting θ (Figure 2A) or asymmetrically with the top branch more aligned with one of the two bottom branches (Figure 2B and C). Also shown is dimensional variation of the top branch with some slightly bigger than the larger of the two bottom branches (Figure 2A and C) and some nearly twice of the average of the bottom branches (Figure 2B). Scale bar: 50 nm.
Figure 3. Surface morphology and structure of an isolated single boron nanowire Y-junction. Left and right insets indicate highresolution TEM and SAED pattern, respectively. Scale bar: 50 nm.
spherical Au catalytic tips were present, the smallest top branch diameter will occur when both bottom branches have semispherical tips while the top branch has a spherical tip. In this case, the top branch diameter is between those of the two bottom branches. If the two bottom branches have similar tip diameters, then the top branch will have the same. When both of the two bottom branches have spherical tips while the top branch has semispherical, the top branch will be bigger than both bottom ones. For example, the top branch diameter will be 1.6 times of that of the bottom branches if the two bottom branches have the same diameter. If one of the bottom branches is much bigger than the other, then the top branch diameter will be around twice that of the bigger bottom branch. No crystallinity has been confirmed in the BNYJs. Figure 3 depicts the TEM image of a BNYJ with the high-resolution TEM image (left inset) and selected area electron diffraction (SAED, right inset) pattern. Both the bottom and top branches are amorphous. This is not surprising considering that the single BNWs made in this temperature range using a similar method are also amorphous. In a prior work, we have found that the crystallinity, as well as the alignment, of the single BNWs can be improved significantly by thermal quenching.28 It remains to be seen whether a highly crystalline BYNJ array could be obtained using this method. In addition, the dirty Nano Lett., Vol. 6, No. 3, 2006
surface of the sample shown in Figure 3, which was observed on only few BNYJs, may not be a common feature of the BNYJ and could be caused in the TEM sample preparation. The microscopic mechanism of BNYJ formation is still under investigation. It has been found that the low-melting B2O3 involvement has played a critical role in causing inclined growth of the single BNWs.12 In fact, no BNYJs were formed if B2O3 was not included in the vapor source. To understand the role of B2O3, elemental mapping was performed on some top branches of the BNYJs and the representative result is shown in Figure 4A and B. Interestingly, a strong O peak was detected at the interface between the Au catalytic tip and the top branch stem. To confirm that this O peak is not from the oxide sheath, we removed the Au tip and took an elemental map on the exposed interface between the Au tip and the top branch stem. The exposed interface was tilted at ∼45° with respect to the electron beam that was scanned vertically back and forth over an area extended from the stem (on the left) to the interface (on the right). Fourteen points were taken on each line scan. On the first seven line scans (point 1-98) across the stem of the BNYJ top branch, O peaks appear only on the sheath of the top branch as expected. On the interface, however, additional broad O peaks centered on the interface were observed, suggesting that a thin disk of oxide was present directly underneath the Au catalytic end-tip. Because the B signal oscillation synchronizes with the O’s, we suspect that this thin disk is boron oxide. The amount of Au is negligible at the center of the interface, whereas small peaks are observable on the edges most probably from residues of the Au tip. B2O3 has a low melting point around 450 °C, much below the growth temperature for BNYJs employed in this experiment. The BOx disk sitting between the BNW top branch and the Au tip was most probably in its liquid state (if it is B2O3) during the BNYJ growth and therefore modified the growth mechanism of BNWs. In fact, BNWs grown with B2O3 in vapor source at temperatures between 800 and 1050 °C12 form bundles. The BNWs in the bundle incline with 387
Figure 4. EDS elemental line scans of an isolated boron nanowire Y-junction top branch. (A) Elemental line scan along the top branch of BNYJ. A strong O peak can be seen at the interface between the Au end-tip and the BNW top branch. (B) Elemental line scan across the interface between the Au tip and the top BNYJ branch along the direction shown in the drawing that covered the wall and interface area (see the rectangular mark in Figure 1B) after removing gold end-tip. The dashed line in Figure 2B indicates the boundary of the wall and the interface. A periodic distribution of strong oxygen peaks was observed in the middle of the top branch, indicating that the top branch is separated by an oxide disk from the Au tip. Scale bar: 50 nm. (C) Schematic description of the proposed BNYJ growth mechanism: step 1, B dissolves into Au catalyst while O attaches to Au surface; steps 2 and 3, BOx droplets form and accumulate near the Au-B interface; step 4, BNW branches form from the modified Au-B interface; and step 5, BNW fusion.
respect to the normal of the substrate. This differs drastically from the single BNW growth in the absence of boron oxide28 and results in multiple initiations of BNWs from the same nucleation site (see Figure 4C). Another experimental observation worth mentioning is that the aerial density of the BNYJ increases dramatically with the processing temperature, Ts. High Ts values contribute to BNYJ formation in at least two ways. First, the BNW grow rate is much higher at higher Ts values, allowing a BNW to grow to lengths needed to reach another nearby one for fusion within the sample growth time (∼30 min in this experiment at Ts ≈ 1100-1200 °C). At lower Ts ≈ 900 °C, a much longer growth time around 4-5 h was observed for BNYJs to form. The other effect of the high Ts values is enhancement of the solubility of B in Au, which in turn enhances the formation of liquid BOx inside the Au tip. Consequently, both the increased number of BNW branches in the bundle and the enlarged tip dimension could increase the probability for BNW fusion. In summary, the boron nanowire Y-junctions have been synthesized by introducing an oxide-assisted VLS growth 388
mode. In this oxide-assisted VLS process, the low-melting B2O3 was present during BNW growth. The interplay between the low-melting B2O3 with the Au liquid metallic catalyst resulted in inclination of BNWs. This provides an essential condition for fusing BNWs. Although further investigation toward synthesizing single-crystalline BNYJs is necessary, the approach of the self-assembled growth of the BNYJs reported in this work has provided a viable strategy for the formation of large-scale nanowire functional devices. Acknowledgment. This research was supported in part by DOE, NSF, and Swedish Research Council (VR). References (1) Gudiksen, M. S.; Lauhon, L. J.; Wang, J.; Smith, D. C.; Lieber, C. M. Nature 2002, 415, 617-620. (2) Tong, L.; Gattass, R. R.; Ashcom, J. B.; He, S.; Lou, J.; Shen, M.; Maxwell, I.; Mazur, E. Nature 2003, 426, 816-819. (3) Gu, Y.; Kwak, E.-S.; Lensch, J. L.; Allen, J. E.; Odom, T. W.; Lauhon, L. J. Appl. Phys. Lett. 2005, 87, 043111. (4) Bielejec, E.; Seamons, J. A.; Reno, J. L.; Lilly, M. P. Appl. Phys. Lett. 2005, 86, 083101.
Nano Lett., Vol. 6, No. 3, 2006
(5) Shaw, W. C.; Hudson, D. E.; Danielson, G. C. Phys. ReV. 1957, 107, 419-427. (6) Greiner, E. S.; Gutowski, J. A. J. Appl. Phys. 1957, 28, 1364-1365. (7) Nakayama, T.; Kimura, K. J. Jpn. Inst. Met. 1999, 63, 1400-1406. (8) Berezin, A. A. AIP Conference Proceedings on Boron-Rich Solids; Albuquerque, NM, 1985; pp 224-233. (9) Gindulyte, A.; Lipscomb, W. N.; Massa, L. Inorg. Chem. 1998, 37, 6544-6545. (10) Boustani, I.; Quandt, A.; Herna´ndez, E.; Rubio, A. J. Chem. Phys. 1999, 110, 3176-3185. (11) Nagamatsu, J.; Nakagawa, N.; Muranaka, T.; Zcnitani, Y.; Akimitsu, J. Nature 2001, 410, 63-64. (12) Yun, S. H.; Dibos, A.; Gao, X.; Wu, J. Z.; Karlsson, U. O. Appl. Phys. Lett. 2005, 87, 113109. (13) Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89-90. (14) Hiruma, K.; Yazawa, M.; Katsuyama, T.; Ogawa, K.; Haraguchi, K.; Koguchi, M.; Kakibayaashi, H. J. Appl. Phys. 1995, 77, 447462. (15) Westwater, J.; Gosain, D. P.; Tomiya, S.; Usui, S.; Ruda, H. J. Vac. Sci. Technol., B 1997, 15, 554-557. (16) Zhang, Y.; Zhang, Q.; Wang, N.; Yan, Y.; Zhou, H.; Zhu, J. J. Cryst. Growth 2001, 226, 185-191. (17) Zhang, J.; Peng, X. S.; Wang, X. F.; Wang, Y. W.; Zhang, L. D. Chem. Phys. Lett. 2001, 345, 372-376. (18) Wang, Y.; Zhang, L.; Liang, C.; Wang, G.; Peng, X. Chem. Phys. Lett. 2002, 357, 314-318.
Nano Lett., Vol. 6, No. 3, 2006
(19) Lew, K. K.; Pan, L.; Dickey, E. C.; Redwing, J. M. AdV. Mater. 2003, 15, 2073-2076. (20) Nguyen, P.; Ng, H. T.; Meyyappan, M. AdV. Mater. 2005, 17, 549553. (21) Li, J.; Papadopoulos, C.; Xu, J. Nature 1999, 402, 253-254. (22) Li, W. Z.; Wen, J. G.; Ren, Z. F. Appl. Phys. Lett. 2001, 79, 18791881. (23) Shi, W.; Peng, H.; Wang, N.; Li, C. P.; Xu, L.; Lee, C. S.; Kalish, R.; Lee, S. T. J. Am. Chem. Soc. 2001, 123, 11095-11096. (24) Cao, L. M.; Zhang, Z.; Sun, L. L.; Gao, C. X.; He, M.; Wang, Y. Q.; Li, Y. C.; Zhang, X. Y.; Li, G.; Zhang, J.; Wang, W. K. AdV. Mater. 2001, 13, 1701-1704. (25) Wang, N.; Fung, K. K.; Wang, S.; Yang, S. J. Cryst. Growth 2001, 233, 226-232. (26) Wang, S.; Yang, S.; Dai, Z. R.; Wang, Z. L. Phys. Chem. Chem. Phys. 2001, 3, 3750-3753. (27) Lee, S. T.; Wang, N.; Zhang, Y. F.; Tang, Y. H. MRS Bull. 1999, 24, 36-42. (28) Yun, S. H.; Dibos, A.; Wu, J. Z.; Kim, D. K. Appl. Phys. Lett. 2004, 84, 2892-2894. (29) Yang, P.; Wu, Y.; Fan, R. Int. J. Nanosci. 2002, 1, 1-39.
NL052138R
389
