pubs.acs.org/NanoLett
Synthesis of Diamond Nanowires Using Atmospheric-Pressure Chemical Vapor Deposition Chih-Hsun Hsu,† Sylvain G. Cloutier,‡,§ Steven Palefsky,| and Jimmy Xu*,†,| †
School of Engineering, Brown University, Providence, Rhode Island 02912, ‡ Department of Electrical & Computer Engineering, University of Delaware, Newark, Delaware 19716, § Delaware Biotechnology Institute, University of Delaware, Newark, Delaware 19716, and | Department of Physics, Brown University, Providence, Rhode Island 02912 ABSTRACT We report diamond nanowires grown in an atmospheric pressure chemical vapor deposition process. These diamond nanowires are straight, thin and long, and uniform in diameter (60-90 nm) over tens of micrometers. Spectroscopic analysis, electron diffraction, and transmission electron microscopy provided confirmation that these nanowires are diamond with high crystallinity and high structural uniformity. They further revealed that these diamond nanowires are encased within multiwalled carbon nanotubes. KEYWORDS Diamond nanowire, carbon allotropes, carbon nanotube, Young-Laplace pressure, APCVD
I
n light of the recent discovery of many forms of carbon, from C601 and carbon nanotubes2 to graphene,3 it is natural and tempting to ask “what else may carbon have to offer? Can other crystallographic forms, or allotropes, of carbon exist or be discovered?” Among the discovered carbon allotropes, diamond has long been a material of interest for research due to its unique properties. Of all known materials, it is the hardest, and has the highest thermal conductivity.4 Its wide band gap, high electron and hole mobility, and negative electron affinity (NEA) make it an attractive candidate for use in ultraviolet (UV) light detectors and emitters,5 radiation particle detectors,6 field effect transistors,7 electron field emission sources,8 position-sensitive biochemical substrates,9 and many other possible applications. Due to diamond’s robustness, these technologies could be applied in harsh environments such as high temperatures or high-power devices for space applications. Techniques for growing crystalline diamond have evolved from the high-temperature high-pressure (HTHP) method10 to plasma enhanced chemical vapor deposition (PECVD) techniques that typically operate at 120-220 Torr and 900-1500 °C followed by microwave annealing at high temperature.11 Diamond microwires with 25 µm diameter and 400 µm length were synthesized in 1968 using a radiation heating unit developed from a superhigh-pressure xenon tube.12 Some top-down approaches employed include reactive ion etching to fabricate diamond nanowires 3-10 nm in length,13 and using porous anodic aluminum oxide as a template to form a diamond cylinder array.14 Post-treatment of carbon species and PECVD tech-
niques have been used to make diamond nanorods of low crystallinity of up to 200 nm in length.15-17 Diamond nanowires have also been a subject of great interest in various theoretical studies and structural simulations with findings providing insights into the structural stabilities and inspiration for potential applications.18-20 Synthesis of crystalline diamond nanowires is of major interest since they offer the potential for enabling applications across many disciplines, for advancing the science of material synthesis at the nanoscale and atomic scale, and for validating the search for new forms of carbon. However, the fabrication of long, single crystalline diamond nanowires using conventional thermal CVD methods has so far proven elusive, despite the potential benefits. In this work, we report that by altering an otherwise standard carbon nanotube (CNT) growth process with the introduction of sample cooling under a pure hydrogen flow, diamond nanowires were grown on the silicon substrate. They were straight, thin and long, and longitudinally uniform in exterior diameter (60-90 nm) along the entire lengths of tens of micrometers (Figure 1). The growth process began with methane and hydrogen flow over an Fe catalyst solution dispersed on a silicon substrate21 under conventional chemical vapor deposition (CVD) conditions at 900 °C.22 Unlike the diamond film growth processes that have been documented, no plasma or energy radiation is used during the CVD growth. After this process was completed, and without pumping the residual methane from the chamber, pure hydrogen was flowed through the quartz tube chamber at the rate of 200 sccm while the temperature was slowly lowered to ambient at the rate of ∼1.2 °C/min over a period of 12 h. Scanning electron microscope (SEM) imaging (Figure 1A) shows sparsely distributed straight nanowires that look
* Corresponding author:
[email protected], 184 Hope St, Providence, RI 02912, telephone (401) 863-1439, fax (401) 863-9107. Received for review: 02/20/2010 Published on Web: 08/02/2010 © 2010 American Chemical Society
3272
DOI: 10.1021/nl100616x | Nano Lett. 2010, 10, 3272–3276
FIGURE 2. HRTEM images of a single diamond nanowire. (A) HRTEM image showing the single-crystallinity of a diamond nanowire with domain size at least 30 × 30 nm in this micrograph. (B) The selective area electron diffraction (SAED) pattern taken from the same diamond nanowire also shows a crystalline cubic diamond (cdiamond) structure. (C) A zoomed-in view of the crystalline structure displaying clear lattice planes of cubic diamond (111) surfaces with a lattice constant of 0.21 nm.
where a portion of the nanotube shell was stripped away by a focused 532 nm laser beam in the micro-Raman spectrometer as shown in Figure 1B. Note that the diamond nanowire remained intact even after being subjected to prolonged laser exposure, indicative of the difference in light absorption between the transparent wide band gap diamond and the narrow band gap graphitic shell. This also suggests that the diamond nanowire has good crystallinity. High-resolution transmission electron microscopy (HRTEM) imaging of the nanowire was performed, as shown in Figure 1C. It confirms that the structure of the nanowire indeed consists of a diamond core encased within a graphitic shell. In some cases, we observed amorphous carbon layers instead of graphitic shells; however, the diamond nanowires contained within were of the same high-crystalline quality. Information about the diamond structure of the nanowires was extracted from transmission electron microscopy. Figure 2A shows a HRTEM image of one of these nanowires, which reveals a crystalline diamond wire structure with a lattice constant of 0.21 nm, corresponding to the (111) orientation of diamond. This image indicates that the size of the crystal domain is at least 30 × 30 nm. In fact, in some cases, single crystal with low defects was seen along the length of the entire nanowire. Figure 2B presents a selective area electron diffraction pattern (SAED) taken from the diamond nanowire along the
FIGURE 1. Electron microscopy of diamond nanowires encased within a carbon nanotube shell. (A) Diamond core enclosed in a CNT sheath, typically tens of micrometers in length and 60-90 nm in exterior diameter. A magnified SEM image of the nanowire tip in the upper inset shows a catalyst embedded inside the tip of the nanowire. The SEM image in the lower inset shows the straight and uniform nanowires with flat terminations. (B) SEM image of a core-shell diamond nanowire with a portion of the CNT shell stripped away selectively by laser burning. The blurring of the diamond nanowire tip is due to oscillations caused by charge accumulation from the impinging electron beam. (C) The transmission electron microscopy (TEM) image shows the diamond core and CNT walls of the nanowire. The diamond (111) lattice (0.21 nm) is enclosed by the outer CNT shell with a graphite (002) surface (0.34 nm) whose surface stress provides a very high effective pressure (on the order of gigapascals) in which the diamond phase is stable. The low-magnification TEM image in the inset shows the core-in-shell wire structure more clearly. The scale bar in the inset is 5 nm.
qualitatively different from the curly carbon nanotubes normally grown in CVD and seen aggregated in a crowded matrix elsewhere in the sample, and from dense agglomerations of mixed diamond and graphite seen in previous studies.15-17 The nanowires measure tens of micrometers in length and 60-90 nm in outer diameter. Extensive spectroscopic and imaging analysis was performed and is presented below. The core-shell structure of a diamond nanowire encased in a graphitic shell also manifested itself in an SEM image © 2010 American Chemical Society
3273
DOI: 10.1021/nl100616x | Nano Lett. 2010, 10, 3272-–3276
monly used in diamond synthesis, it is likely that the growth mechanism differs from the traditional CVD diamond growth mechanism. There is an existing theoretical framework within which this discovery may be placed; stable diamond nanowires have been shown to be possible in computational models and simulations for some time (albeit under idealized theoretical settings and without considering the actual growth conditions and environmental limitations). More specifically, density functional models have shown that diamond nanowires might exist in cubic, cylindrical, or dodecahedral crystalline forms.18 It seems reasonable to expect that the nanoscale curvature has played a significant role, as it is known that a small diameter is accompanied by a large surface stress, which could give rise to a high equivalent pressure on the core of the nanowire. Indeed, an empirical thermodynamic model argued that it is possible for a diamond nanowire to grow inside a tubular graphitic enclosure (or carbon nanotube).20 A key aspect of this argument, independent of the specifics of the model, is that for a smalldiameter wire a capillary pressure (or so-called YoungLaplace pressure) ∆P ≈ 2γ/r (where r is the radius and γ is the surface energy of diamond) can be sufficiently high to allow the diamond phase to exist in the inner core while the outer shell with less equivalent pressure takes on the graphitic phase. Depending on the specific surface and interface energy parameters, within a thick nanotube-like shell of an interior radius around 10-25 nm, or for even smaller dimensions occurring at a defective spot of the nanotube wall, the equivalent pressure could reach the order of 1-5 GPa.27 Similar phenomena were observed experimentally in diamond nucleation inside carbon onion high curvature enclosures.28 Molecular dynamic (MD) simulations18,19 estimate that a stable diamond nanowire could have a diameter on the order of ∼10 nm. Though they indicate that growth along the [110] principal axis (the growth direction of our diamond nanowire) may not be stable, these simulations were carried out under boundary conditions that differ from our experimental conditions. Therefore, our diamond nanowire may be in a different regime. However, the order of magnitude estimate corresponds well to the experimental findings of the wire diameter and the crystal domain size. A large number of trials were performed to narrow down the conditions under which diamond nanowires can grow. For example, when the final step of cooling under hydrogen flow was omitted, carbon nanotubes and amorphous carbon were synthesized but no diamond. Hydrogen has been known to produce a dilution effect that facilitates the transformation of sp and sp2 bonds into sp3 bonds.29 It has also been shown that hydrogen locally decomposes amorphous and graphitic deposits more quickly than diamond at process temperatures similar to ours,4,30 allowing methane feedstock to grow diamond preferentially31 Meanwhile, transition metals (iron in our case) are known to facilitate the dissociation of hydrogen molecules into atomic hydrogen at signifi-
FIGURE 3. Micro-Raman spectrum taken from the same nanowire shown in the SEM image (Figure 1A). The diamond signature peak at 1332 cm-1 is narrow and pronounced, and the two Raman peaks at 1564 and 1603 cm-1 demonstrate the presence of a CNT shell. The inset is the optical microscope image of the diamond nanowire placed in the Raman system corresponding to the SEM provided in Figure 1A. The cross hair in the inset marks the location where the Raman laser beam (spot size ∼1 µm) impinged on the diamond nanowire sample. This figure also includes (top left) a graphic impression showing formation of the nanowire within a shell.
[-1,1,2] zone axis, showing that the nanowire has a crystalline cubic diamond structure. Electron diffraction patterns taken along multiple zone axes were further examined from the same sample, and again the cubic diamond structure was observed. The SAED experiment also indicated that diamond nanowire has the same lattice parameter as bulk cubic crystalline diamond. The diffraction pattern suggests that the growth direction of this nanowire is along the [2,2,0] principal axis. Micro-Raman spectra, taken with a 532 nm laser beam focused on randomly selected microspots of individual isolated nanowires (e.g., at the point indicated by a red cross in the inset of Figure 3), reveal the signature diamond peak23 at 1332 cm-1, dominant in intensity and extremely narrow in line width. Its fwhm of 14.5 cm-1 is comparable to that of high-quality CVD-grown bulk diamond (5-10 cm-1) and is much narrower than that of other diamond nanocrystallite films found in literature.24 The two additional peaks at 1564 and 1603 cm-1 are those of the well-known graphitic Gbands.25 Since the micro-Raman spectroscopy was performed on individual isolated nanowires as shown in Figure 3, these G-band peaks support the HRTEM evidence of a graphitic shell and a diamond core-in-nanotube shell structure. No such spectral lines were observed from the substrate when moving the excitation beam away from the wire. A Raman line at 1150 cm-1 often seen in CVD-grown nanodiamond films is absent here, as is its associated 1450 cm-1 peak. These lines, if present, are usually indicative of poor quality diamond crystals.26 Since fabrication was carried out under atmospheric pressure and low temperature compared with the thermodynamic nonequilibrium plasma systems that are com© 2010 American Chemical Society
3274
DOI: 10.1021/nl100616x | Nano Lett. 2010, 10, 3272-–3276
cant low hydrogen dissociation barrier.32,33 Such atomic hydrogen was demonstrated to enhance diamond nucleation as well as graphite etching and thus has been widely applied in the synthetic diamond growth processes. As written above, no diamond content was found without the hydrogen step, which suggests that hydrogen is instrumental in obtaining a yield of diamond nanowires. Other combinations of temperature, pressure, gas content, flow rates, and process time were attempted but did not yield any diamond content. As with the discovery of all new materials and systems, the technologies that will emerge from diamond nanowires cannot be anticipated at such an early stage. For example, carbon nanotubes are now pursued for many applications that were not anticipated at the time of their discovery, or predicted for years after. However, one may already see a number of potential applications of this new material. The use of diamond nanowires in field-effect transistors could be advantageous for high power, high temperature, and high speeds34 due to their one-dimensional structure, high mobility, large band gap and breakdown field, and high thermal conductivity.35 Solar-blind, radiation-hard diamond nanowire UV detectors6 and DNA sensors would also be possibilities of broad interest. The experimental findings reported here are both unprecedented and rather intriguing, especially in the light of the discoveries of other relatively new carbon phasessC60 and carbon nanotubes. The fact that diamond can develop with this diameter, length, morphology and crystallinity in these moderate growth conditions (low temperature, low pressure, thermal CVD) not only offers further confirmation for the modeling expectations of stable diamond nanowires18,19 and reduced-pressure transformations27,28 but also illuminates a pathway where nanoenclosures could provide growth conditions that might be otherwise unattainable, which in principle could be accessible to other material systems as well. These results also show that the existing understanding of diamond growth, or more generally carbon structures, is far from complete. As much as the results reported here appear to provide a basis for optimism, they themselves are far from optimal. They were achieved with a relatively low yield (∼10%) and so far offer no direct control for the diameter and length. More study is needed to find the optimal growth conditions. We hope the report of the existence of diamond nanowires will generate broad interest in developing a controlled and reproducible high-yield fabrication process and a good understanding of the growth mechanism.
Research, the WCU Program at Seoul National University, and NSERC-Canada. REFERENCES AND NOTES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22)
Acknowledgment. We thank Dr. C. Ni, Dr. C. Y. Wen, and A.W. McCormick for discussions on TEM and Dr. P. Jay for consultations. This research was enabled by the Office of Naval Research. Its technology developments have been supported in part by the Air Force Office of Scientific © 2010 American Chemical Society
(23) 3275
Kroto, H. W.; Heath, J. R.; Obrien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985, 318, 162–163. Iijima, S. Nature 1991, 354, 56–58. Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Science 2002, 306, 666–669. Spear, K. E. J. Am. Ceram. Soc. 1989, 72, 171–191. Koizumi, S.; Watanabe, K.; Hasegawa, F.; Kanda, H. Science 2001, 292, 1899–1901. Monroy, E.; Omnes, F.; Calle, F. Semicond. Sci. Technol. 2003, 18, R33–R51. Isberg, J.; Hammersberg, J.; Johansson, E.; Wikstro¨m, T.; Twitchen, D. J.; Whitehead, A. J.; Coe, S. E.; Scarsbrook, G. A. Science 2002, 297, 1670–1672. Okano, K.; Koizumi, S.; Silva, S. R. P.; Amaratunga, G. A. J. Nature 1996, 381, 140–141. Yang, W.; Auciello, O.; Butler, J. E.; Cai, W.; Carlisle, J. A.; Gerbi, J. A.; Gruen, D. M.; Knickerbocker, T.; Lasseter, T. L.; Russell, J. N.; Smith, L. M.; Hamers, R. J. Nat. Mater. 2002, 1, 253–257. Bundy, F. P.; Hall, H. T.; Strong, H. M.; Wentorf, R. H. Nature 1955, 176, 51–55. Meng, Y. F.; Yan, C. S.; Lai, J.; Krasnicki, S.; Shu, H.; Yu, T.; Liang, Q.; Mao, H. K.; Hemley, R. J. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 17620–17625. Derjaguin, B. V.; Fedoseev, D. V.; Lukyanovich, V. M.; Spitzin, B. V.; Ryabov, V. A.; Lavrentyev, A. V. J. Cryst. Growth 1968, 2, 380–384. Yang, N. J.; Uetsuka, H. S.; Osawa, E.; Nebel, C. E. Nano Lett. 2008, 8, 3572–3576. Masuda, H.; Yanagishita, T.; Yasui, K.; Nishio, K.; Yagi, I.; Rao, T. N.; Fujishima, A. Adv. Mater. 2001, 13, 247–249. Sun, L. T.; Gong, J. L.; Zhu, D. Z.; Zhu, Z. Y.; He, S. X. Adv. Mater. 2004, 16, 1849–1853. Vlasov, I. L.; Lebedev, O. I.; Ralchenko, V. G.; Goovaerts, E.; Bertoni, G.; Van Tendeloo, G.; Konov, V. I. Adv. Mater. 2007, 19, 4058–4062. Shang, N.; Papakonstantinou, P.; Wang, P.; Zakharov, A.; Palnitkar, U.; Lin, I. N.; Chu, M.; Stamboulis, A. ACS Nano 2009, 3, 1032–1038. Barnard, A. S.; Russo, S. P.; Snook, I. K. Nano Lett. 2003, 3, 1323– 1328. Barnard, A. S.; Snook, I. K. J. Chem. Phys. 2004, 120, 3817–3821. Wang, C. X.; Yang, G. W. Mater. Sci. Eng., R 2005, 49, 157–202. Li, Y. M.; Kim, W.; Zhang, Y. G.; Rolandi, M.; Wang, D. W.; Dai, H. J. J. Phys. Chem. B 2001, 105, 11424–11431. The catalyst for the nanowire synthesis consists of iron cations dispersed in a sol-gel, made of 11.7 g of tetraethoxysilane (TEOS, Aldrich, 98% purity), 25.9 g of ethanol, 25.3 g of deionized water, and 11 mL of hydrochloric acid (HCl). The sol-gel mixture was stirred for 3 h at room temperature and was adjusted to pH 5.8 using 0.05 M of sodium hydroxide (NaOH). A 0.1 M iron nitrate-9 hydrate (Fe(NO3)3·9H2O, Aldrich, 98%) 1-propanol solution was then mixed with the as-prepared sol-gel system (volume ratio 1:1). After 30 min of gelation, the sol-gel based catalyst was spincoated onto a p-type silicon substrate followed by baking at 70 °C for 30 min. The resultant distribution of catalyst particle sizes was in the range of 10-30 nm after 700 °C high temperature annealing, as observed by SEM. Diamond nanowire growth was done in an atmospheric pressure chemical vapor deposition (CVD) chamber under isothermal conditions (900 °C). The ratio of methane to hydrogen flow was set at 500:100 standard cubic centimeters per minute (sccm) for 5 min and was then switched to 100:200 sccm for 25 min. After the standard nanotube process was completed, pure hydrogen was flowed through the quartz tube (2.5 cm wide and 1.4 m long) at 200 sccm without having pumped the methane from the chamber. The chamber was then slowly cooled over 12 h to ambient temperature. Knight, D. S.; White, W. B. J. Mater. Res. 1989, 4, 385–393. DOI: 10.1021/nl100616x | Nano Lett. 2010, 10, 3272-–3276
(24) Ferrari, A. C.; Robertson, J. Philos. Trans. R. Soc. London, Ser. A 2004, 362, 2477–2512. (25) Jorio, A.; Souza Filho, A. G.; Dresselhaus, G.; Dresselhaus, M. S.; ¨ nlu¨, M. S.; Goldberg, B. B.; Pimenta, M. A.; Hafner, Swan, A. K.; U J. H.; Lieber, C. M.; Saito, R. Phys. Rev. B 2002, 65, 155412-1-9. (26) Ferrari, A. C.; Robertson, J. Phys. Rev. B 2001, 63, 121405-1-4. (27) Wang, C. X.; Chen, J.; Yang, G. W.; Xu, N. S. Angew. Chem., Int. Ed. 2005, 44, 7414–7418. (28) Banhart, F.; Ajayan, P. M. Nature 1996, 382, 433–435. (29) Lambrecht, W. R. L.; Lee, C. H.; Segall, B.; Angus, J. C.; Li, Z. D.; Sunkara, M. Nature 1993, 364, 607–610.
© 2010 American Chemical Society
(30) Angus, J. C.; Will, H. A.; Stanko, W. S. J. Appl. Phys. 1968, 39, 2915–2922. (31) Chauhan, S. P.; Angus, J. C.; Gardner, N. C. J. Appl. Phys. 1976, 47, 4746–4754. (32) Gross, A. Appl. Phys. A: Mater. Sci. Process. 1998, 67, 627–635. (33) Berger, H. F.; Grosslinger, E.; Rendulic, K. D. Surf. Sci. 1992, 261, 313–320. (34) Barnard, A. S.; Russo, S. P.; Snook, I. K. Phys. Rev. B 2003, 68, 073406-1-4. (35) Padgett, C. W.; Shenderova, O.; Brenner, D. W. Nano Lett. 2006, 6, 1827–1831.
3276
DOI: 10.1021/nl100616x | Nano Lett. 2010, 10, 3272-–3276
