Flame synthesis of aligned tungsten oxide nanowires - Semantic Scholar
APPLIED PHYSICS LETTERS 88, 243115 共2006兲
Flame synthesis of aligned tungsten oxide nanowires Fusheng Xu and Stephen D. Tsea兲 Department of Mechanical and Aerospace Engineering, Rutgers University, Piscataway, New Jersey 08854
Jafar F. Al-Sharab and Bernard H. Kear Department of Materials Science and Engineering, Rutgers University, Piscataway, New Jersey 08854
共Received 8 March 2006; accepted 28 April 2006; published online 15 June 2006兲 Aligned single-crystal WO2.9 nanowires are grown directly from tungsten substrates at high rates using a flame synthesis method. The nanowires have diameters of 20– 50 nm, lengths ⬎10 m, coverage density of 109 – 1010 cm−2, and growth rates ⬎1 m / min. Growth occurs by the vapor-solid mechanism, with local gas-phase temperature 共⬃1720 K兲 and chemical species 共O2, H2O, and H2兲 strategically specified at the substrate for self-synthesis. Advantages of this synthesis method are reduced processing times, absence of necessity for substrate pretreatment or catalysts, scalability for large-area surface coverage, high purity and yield of oriented nanowires, and continuous processing conditions. © 2006 American Institute of Physics. 关DOI: 10.1063/1.2213181兴 Recently, nanostructured tungsten oxide materials have attracted considerable attention due to their unique properties 共electrochromic, gaschromic, optochromic, and magnetic兲 along with their potential for incorporation into nanodevices and sensors. As such, several techniques for the fabrication of one-dimensional tungsten oxide nanostructures, i.e., nanowires, have been explored.1–18 Gu et al.2 obtained tungsten oxide nanowires on tungsten tips, pretreated with electrochemical etching, by heating them to 700 ° C in argon flow. The nanowires had diameters of 10– 30 nm and an average length of 0.3 m after 10 min. Later, Li et al.4 grew quasialigned single-crystalline W18O49 nanotubes and nanowires by heating tungsten foils as targets 共1000– 1050 ° C兲 and using a Ta wafer 共650 ° C兲 as substrate at low pressures 共0.2– 10 torr兲. After 2 h, the nanowires were 20– 100 nm in diameter and ⬍3 m long. Qi et al.10 obtained potassiumdoped tungsten oxide nanowires by heating a tungsten plate 共625– 650 ° C兲 covered with a layer of potassium halide salts for 2 h. The heavily doped wires had diameters of ⬃400 nm and average lengths of 10 m. Liu et al.16 heated a thin tungsten filament 共1400 ° C兲 in a vacuum chamber with some air leakage 共oxygen partial pressure of ⬃2.7 ⫻ 10−5 torr兲 for 48 h to grow undoped/doped tungsten oxide nanowires. The wires were found standing straight and clean on the filament, being ⬃30 nm in diameter and up to a few tens of microns long. Wang et al.17 synthesized dense and well-crystallized monoclinic W18O49 共010兲 nanowires with diameters of 10– 20 nm and lengths of 0.15– 0.2 m in 1 h from annealing 共680 ° C兲 and then oxidizing in O2 共450 ° C兲 WCx films. As can be seen above, processing can be complex, incatalysts,10,15 and vacuum volving pretreatment,2 1,4,8,16,18 systems, while still characterized by low singlenanowire growth rates and low total yield densities. Consequently, studies on nanoscale WOx materials and their applications are presently limited due to lack of easy processes for high- rate, yield, purity, and orientation synthesis of such materials. The growth of tungsten oxide nanowires over large a兲
Author to whom correspondence should be addressed; electronic mail: [email protected]
areas remains especially challenging. In this work, a flame synthesis method is employed to grow well-aligned, singlecrystal nanowires with diameters ranging from 20 to 50 nm, coverage density of 109 – 1010 cm−2, and growth rates of microns per minute, without any pretreatment or catalysts and in open environments. The method is also robust in that the combustion process inherently provides for 共i兲 an elevated enthalpy source to evaporate the metal substrate atoms, 共ii兲 the gas-phase chemical species 共e.g., oxidizer, water vapor, and hydrogen兲 to produce the requisite oxide, and 共iii兲 a favorable temperature gradient for growth of the nanowires. The experiment utilizes the quasi-one-dimensional counterflow diffusion flame, with an air jet impinging onto an opposed jet of nitrogen-diluted methane 共1.32 l / min fuel from a 19-mm-diam nozzle兲 at atmospheric pressure. The flame is aerodynamically well defined, with gradients existing only in the axial direction, and can be easily probed and compared with simulations involving detailed chemical kinetics and transport. Figure 1 displays the flame profiles of temperature and relevant species mole fractions as predicted computationally using GRI-Mech 1.2,19 and verified using spontaneous Raman spectroscopy.20 A tungsten wire sub-
FIG. 1. 共Color online兲 Gas-phase flame structure of counterflow diffusion flame. TC represents thermocouple measurement assessing substrate temperature. Tungsten substrate position is marked.
0003-6951/2006/88共24兲/243115/3/$23.00 88, 243115-1 © 2006 American Institute of Physics Downloaded 16 Jun 2006 to 128.6.73.78. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
243115-2
Appl. Phys. Lett. 88, 243115 共2006兲
Xu et al.
FIG. 2. 共Color online兲 共a兲 Low magnification FESEM image of as-grown tungsten oxide nanowires showing high density of yield. 共b兲 Typical FESEM image of nanowires from a magnified top view. 共c兲 Typical FESEM image of nanowires from a side view. 共d兲 EDXS spectra of as-grown nanowires.
strate 共99.95% purity, 0.5 mm diameter兲, without any pretreatment, is inserted radially into the flame structure at z = 0.88 cm 共Fig. 1兲, where the temperature and the oxygen supply are sufficient to promote reactions leading to tungsten oxide. Temperatures are measured using a 125 m Pt/ Pt– 10% Rh thermocouple 共S-type兲 coated with BeO – Y2O3 共see Fig. 1兲 to establish the actual tungsten substrate temperature, which is expected to differ from the gasphase temperature due to radiative effects and conductive losses along the length of the substrate probe. The morphologies of as-grown tungsten oxide nanowires are examined using field emission scanning electron microscopy 共FESEM兲. Elemental analysis is conducted using energy dispersive x-ray spectroscopy 共EDXS兲 attached to FESEM. Structural features of the nanomaterial are investigated using high resolution transmission electron microscopy 共HRTEM兲, along with selected area electron diffraction 共SAED兲. Figure 2共a兲 shows a low magnification FESEM image of a dense yield of nanomaterials grown directly on a tungsten substrate. A magnified image of a top view 关Fig. 2共b兲兴 shows vertically oriented nanowires. A side view 关Fig. 2共c兲兴 reveals individual nanowires, as well as those grouped in small bundles. The as-grown nanowires have diameters of 20– 50 nm, with lengths of more than 10 m for a sampling duration of 10 min. EDXS mappings for different growth regions of various sizes, from 10⫻ 10 m2 to single nanowires, are essentially the same with no discernable differences. The EDXS spectra 关Fig. 2共d兲兴 show that the nanowires are composed of W and O, along with negligible amounts of C. The presence of C results most likely from deposition of C on the nanowires from carbon-containing species that diffuse from the reaction zone during processing. Nevertheless, the findings indicate that the as-synthesized materials are nanosized tungsten oxide wires. Tungsten oxide can exhibit different crystal structures, such as cubic and monoclinic WO3, tetragonal WO2.9, and monoclinic W18O49. Figure 3共a兲 presents a low magnification bright-field TEM image of a 20 nm nanowire. Nanocrystallized tungsten oxide platelets 共as indicated by the arrow兲 which come off the surfaces of bundled nanowires during TEM sample preparation can also be observed. The inset in Fig. 3共a兲 is a SAED pattern from the tungsten oxide nano-
FIG. 3. 共Color online兲 共a兲 TEM image showing a 20 nm nanowire and platelets, along with the selected area electron diffraction pattern 共inset兲. 共b兲 HRTEM image of a single tungsten oxide nanowire, along with lattice-plane spacing 共inset兲. 共c兲 HRTEM of a nanowire tip, with ledge growth suggesting VS mechanism.
wire and platelets. The indexed SAED pattern with the first three highest intensities of 3.779, 3.126, and 2.67 Å matches very well with the tetragonal phase of WO2.9 with lattice constants of a = 5.3 Å, b = 5.3 Å, and c = 3.83 Å 共PDF card No. 18-1417兲. These d spacings correspond to 兵110其, 兵101其, and 兵200其, respectively. HRTEM imaging of a typical flame-grown nanowire is shown in Fig. 3共b兲, revealing its dislocation-free, singlecrystalline nature. Analysis of the two-dimensional Fourier transform pattern of the TEM image gives an average spacing for the lattice planes of 3.78 Å, which corresponds to the reflections from d spacings of 共110兲 planes of the tetragonal WO2.9 phase. Thus, these as-grown single-crystalline nanowires have preferable growth along the 关110兴 direction, as indicated in Fig. 3共b兲. The mechanism of formation of these nanowires appears to be by vapor-solid 共VS兲 growth. In the vapor-liquid-solid 共VLS兲 process, the growth is promoted by a liquid-solid interface, generally marked by the presence of droplets at the tips of the nanowires. The morphology of a nanowire is shown in Fig. 3共c兲, evincing no metal nanoparticle at its tip. The TEM image further reveals thickening by a ledgegrowth mechanism, which is additional evidence for vaporphase transport and deposition. Figure 1 shows that the gas-phase temperature at the substrate is ⬃1720 K, with the tungsten substrate temperature estimated to be ⬃1600 K, which is sufficiently high to evaporate tungsten/oxygen species. With available oxygen at the probe location, as displayed in Fig. 1, conditions are then favorable for evaporated tungsten/oxygen species to react with oxygen to form tungsten oxide. At the same time, oxygen can react with the tungsten substrate to form tungsten oxide, which may then evaporate due to its low melting temperature 共⬃1870 K兲. In fact, formation of vapor-phase tungsten oxide species can be quite spontaneous, e.g.,11 共g兲 W共s兲 + O共g兲 2 → WO2 ,
⌬G = − 40.08 kJ/mol K
共1兲 共1723 K兲.
Initial formation of tungsten oxide nanoparticles on the surface of the substrate serves to nucleate the tungsten oxide
Downloaded 16 Jun 2006 to 128.6.73.78. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
243115-3
Appl. Phys. Lett. 88, 243115 共2006兲
Xu et al.
nanowires. Subsequent elongated growth along the 关110兴 direction results from preferential diffusion and condensation of tungsten oxide vapor as adatoms at the tip of a nanowire. With solid-phase WO2 readily transformed into WO3, i.e.,11 1 共g兲 共s兲 WO共s兲 2 + 2 O2 → WO3 ,
⌬G = − 108.24 kJ/mol K
共2兲 共1273 K兲,
we hypothesize that, given the elevated temperature in the probe region 共near the tungsten oxide decomposition temperature of ⬃1720 K兲, decomposition of solid-phase WO3 results in the formation of the final WO2.9 tetragonal phase. The positive temperature gradient extending outward from the 共cooler兲 substrate surface into the 共hotter兲 surrounding gas likely promotes the vertical orientation of the nanowires. The high growth rates that are observed in the flame process are due to favorable local conditions allowing for various mechanisms to occur simultaneously. The presence of water vapor 共a combustion by-product兲 at the probe location 共Fig. 1兲 may further enhance the nanowire growth process. Water vapor cannot only form tungsten oxide, e.g.,15 共g兲 W共s兲 + 3H2O共g兲 → WO共s兲 3 + 3H2 ,
共3兲
but it can also increase the rates of tungsten oxide evaporation and hydrate species formation, e.g.,14 共g兲 WO共s兲 → WO2共OH兲共g兲 3 + H 2O 2 .
共4兲
In addition, the elevated temperatures and radical species, e.g., OH and H, innate in the flame permit nanowire growth without substrate pretreatment or external catalysts. Gu et al.2 observed that a clean and unoxidized tungsten surface was necessary for tungsten oxide nanowire growth. However, similar to Liu et al.,16 we grow tungsten oxide nanowires with high quality, orientation, and purity from oxidized surfaces. Other works have found that H2 reduction2 prior to heat treatment is necessary to grow nanowires on tungsten substrates. Again, this is not necessary in our experiments. As can be seen in Fig. 1, small concentrations of H2 are inherently present at the probe location, which, in combination with the elevated temperatures, can aid in initial nanoparticle formation. Finally, the high characteristic temperature in the flame process induces nanowire growth without the help of catalysts. External catalysts are needed to produce tungsten oxide nanowires in the temperature range of 873– 973 K,2,10,15 while neither catalyst nor pretreatment of substrate is necessary in the temperature range of 1073– 1873 K.1,4,5,7,8,12,16
In summary, we demonstrate the synthesis of vertically aligned, single-crystalline, tetragonal WO2.9 nanowires grown directly from tungsten substrates at high rates in a flame process. The growth mechanism appears to be vaporsolid based, with key parameters being radical species present, oxidizer and water-vapor concentrations, substrate temperature, and gas-phase temperature. Systematic variation of these parameters is readily achieved in our flame technique, which would be very time consuming and tedious in other heat-treatment and chemical vapor deposition 共CVD兲 methods. Finally, our technique is promising for large-scale applications due to its simplicity, scalability, and economy. This work was supported by the National Science Foundation 共NSF-CTS-0213929 and NSF-CTS-0325057兲. 1
Y. Q. Zhu, W. Hu, W. K. Hsu, M. Terrones, N. Grobert, J. P. Hare, H. W. Kroto, D. R. M. Walton, and H. Terrones, Chem. Phys. Lett. 309, 327 共1999兲. 2 G. Gu, B. Zheng, W. Q. Han, S. Roth, and J. Liu, Nano Lett. 2, 849 共2002兲. 3 K. Lee, W. S. Seo, and J. T. Park, J. Am. Chem. Soc. 125, 3408 共2003兲. 4 Y. Li, Y. Bando, and D. Golberg, Adv. Mater. 共Weinheim, Ger.兲 15, 1294 共2003兲. 5 Y. B. Li, Y. Bando, D. Golberg, and K. Kurashima, Chem. Phys. Lett. 367, 214 共2003兲. 6 X.-L. Li, J.-F. Liu, and Y.-D. Li, Inorg. Chem. 42, 921 共2003兲. 7 Z. Liu, Y. Bando, and C. Tang, Chem. Phys. Lett. 372, 179 共2003兲. 8 J. Liu, Y. Zhao, and Z. Zhang, J. Phys.: Condens. Matter 15, L453 共2003兲. 9 X. W. Lou and H. C. Zeng, Inorg. Chem. 42, 6169 共2003兲. 10 H. Qi, C. Wang, and J. Liu, Adv. Mater. 共Weinheim, Ger.兲 15, 411 共2003兲. 11 S. Vaddiraju, H. Chandrasekaran, and M. K. Sunkara, J. Am. Chem. Soc. 125, 10792 共2003兲. 12 Y. Z. Jin, Y. Q. Zhu, R. L. D. Whitby, N. Yao, R. Ma, P. C. P. Watts, H. W. Kroto, and R. M. Walton, J. Phys. Chem. B 108, 15572 共2004兲. 13 H. G. Choi, Y. H. Jung, and D. K. Kim, J. Am. Ceram. Soc. 88, 1684 共2005兲. 14 M. Gillet, R. Delamare, and E. Gillet, Eur. Phys. J. D 34, 291 共2005兲. 15 K. Hong, W. Yiu, H. Wu, J. Gao, and M. Xie, Nanotechnology 16, 1608 共2005兲. 16 K. Liu, D. T. Foord, and L. Scipioni, Nanotechnology 16, 10 共2005兲. 17 S.-J. Wang, C.-H. Chen, R.-M. Ko, Y.-C. Kuo, C.-H. Wong, K.-M. Uang, T.-M. Chen, and B.-W. Liou, Appl. Phys. Lett. 86, 263103 共2005兲. 18 J. Zhou, Y. Ding, S. Z. Deng, L. Gong, N. S. Xu, and Z. L. Wang, Adv. Mater. 共Weinheim, Ger.兲 17, 2107 共2005兲. 19 M. Frenklach, H. Wang, C.-L Yu, M. Goldenberg, C. T. Bowman, R. K. Hanson, D. F. Davidson, E. J. Chang, G. P. Smith, D. M. Golden, W. C. Gardiner, and V. Lissianski, http://www.me.berkeley.edu/gri mech/; and Gas Research Institute Topical Report: M. Frenklach, H. Wang, M. Goldenberg, G. P. Smith, D. M. Golden, C. T. Bowman, C. T. Bowman, R. K. Hanson, W. C. Gardiner, and V. Lissianski, GRI-Mech-An Optimized Detailed Chemical Reaction Mechanism for Methane Combustion, Report No. GRI–95/0058, November 1, 1995. 20 F. Xu, X. Liu, and S. D. Tse, Carbon 44, 570 共2006兲.
Downloaded 16 Jun 2006 to 128.6.73.78. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
Flame synthesis of aligned tungsten oxide nanowires Fusheng Xu and Stephen D. Tsea兲 Department of Mechanical and Aerospace Engineering, Rutgers University, Piscataway, New Jersey 08854
Jafar F. Al-Sharab and Bernard H. Kear Department of Materials Science and Engineering, Rutgers University, Piscataway, New Jersey 08854
共Received 8 March 2006; accepted 28 April 2006; published online 15 June 2006兲 Aligned single-crystal WO2.9 nanowires are grown directly from tungsten substrates at high rates using a flame synthesis method. The nanowires have diameters of 20– 50 nm, lengths ⬎10 m, coverage density of 109 – 1010 cm−2, and growth rates ⬎1 m / min. Growth occurs by the vapor-solid mechanism, with local gas-phase temperature 共⬃1720 K兲 and chemical species 共O2, H2O, and H2兲 strategically specified at the substrate for self-synthesis. Advantages of this synthesis method are reduced processing times, absence of necessity for substrate pretreatment or catalysts, scalability for large-area surface coverage, high purity and yield of oriented nanowires, and continuous processing conditions. © 2006 American Institute of Physics. 关DOI: 10.1063/1.2213181兴 Recently, nanostructured tungsten oxide materials have attracted considerable attention due to their unique properties 共electrochromic, gaschromic, optochromic, and magnetic兲 along with their potential for incorporation into nanodevices and sensors. As such, several techniques for the fabrication of one-dimensional tungsten oxide nanostructures, i.e., nanowires, have been explored.1–18 Gu et al.2 obtained tungsten oxide nanowires on tungsten tips, pretreated with electrochemical etching, by heating them to 700 ° C in argon flow. The nanowires had diameters of 10– 30 nm and an average length of 0.3 m after 10 min. Later, Li et al.4 grew quasialigned single-crystalline W18O49 nanotubes and nanowires by heating tungsten foils as targets 共1000– 1050 ° C兲 and using a Ta wafer 共650 ° C兲 as substrate at low pressures 共0.2– 10 torr兲. After 2 h, the nanowires were 20– 100 nm in diameter and ⬍3 m long. Qi et al.10 obtained potassiumdoped tungsten oxide nanowires by heating a tungsten plate 共625– 650 ° C兲 covered with a layer of potassium halide salts for 2 h. The heavily doped wires had diameters of ⬃400 nm and average lengths of 10 m. Liu et al.16 heated a thin tungsten filament 共1400 ° C兲 in a vacuum chamber with some air leakage 共oxygen partial pressure of ⬃2.7 ⫻ 10−5 torr兲 for 48 h to grow undoped/doped tungsten oxide nanowires. The wires were found standing straight and clean on the filament, being ⬃30 nm in diameter and up to a few tens of microns long. Wang et al.17 synthesized dense and well-crystallized monoclinic W18O49 共010兲 nanowires with diameters of 10– 20 nm and lengths of 0.15– 0.2 m in 1 h from annealing 共680 ° C兲 and then oxidizing in O2 共450 ° C兲 WCx films. As can be seen above, processing can be complex, incatalysts,10,15 and vacuum volving pretreatment,2 1,4,8,16,18 systems, while still characterized by low singlenanowire growth rates and low total yield densities. Consequently, studies on nanoscale WOx materials and their applications are presently limited due to lack of easy processes for high- rate, yield, purity, and orientation synthesis of such materials. The growth of tungsten oxide nanowires over large a兲
Author to whom correspondence should be addressed; electronic mail: [email protected]
areas remains especially challenging. In this work, a flame synthesis method is employed to grow well-aligned, singlecrystal nanowires with diameters ranging from 20 to 50 nm, coverage density of 109 – 1010 cm−2, and growth rates of microns per minute, without any pretreatment or catalysts and in open environments. The method is also robust in that the combustion process inherently provides for 共i兲 an elevated enthalpy source to evaporate the metal substrate atoms, 共ii兲 the gas-phase chemical species 共e.g., oxidizer, water vapor, and hydrogen兲 to produce the requisite oxide, and 共iii兲 a favorable temperature gradient for growth of the nanowires. The experiment utilizes the quasi-one-dimensional counterflow diffusion flame, with an air jet impinging onto an opposed jet of nitrogen-diluted methane 共1.32 l / min fuel from a 19-mm-diam nozzle兲 at atmospheric pressure. The flame is aerodynamically well defined, with gradients existing only in the axial direction, and can be easily probed and compared with simulations involving detailed chemical kinetics and transport. Figure 1 displays the flame profiles of temperature and relevant species mole fractions as predicted computationally using GRI-Mech 1.2,19 and verified using spontaneous Raman spectroscopy.20 A tungsten wire sub-
FIG. 1. 共Color online兲 Gas-phase flame structure of counterflow diffusion flame. TC represents thermocouple measurement assessing substrate temperature. Tungsten substrate position is marked.
0003-6951/2006/88共24兲/243115/3/$23.00 88, 243115-1 © 2006 American Institute of Physics Downloaded 16 Jun 2006 to 128.6.73.78. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
243115-2
Appl. Phys. Lett. 88, 243115 共2006兲
Xu et al.
FIG. 2. 共Color online兲 共a兲 Low magnification FESEM image of as-grown tungsten oxide nanowires showing high density of yield. 共b兲 Typical FESEM image of nanowires from a magnified top view. 共c兲 Typical FESEM image of nanowires from a side view. 共d兲 EDXS spectra of as-grown nanowires.
strate 共99.95% purity, 0.5 mm diameter兲, without any pretreatment, is inserted radially into the flame structure at z = 0.88 cm 共Fig. 1兲, where the temperature and the oxygen supply are sufficient to promote reactions leading to tungsten oxide. Temperatures are measured using a 125 m Pt/ Pt– 10% Rh thermocouple 共S-type兲 coated with BeO – Y2O3 共see Fig. 1兲 to establish the actual tungsten substrate temperature, which is expected to differ from the gasphase temperature due to radiative effects and conductive losses along the length of the substrate probe. The morphologies of as-grown tungsten oxide nanowires are examined using field emission scanning electron microscopy 共FESEM兲. Elemental analysis is conducted using energy dispersive x-ray spectroscopy 共EDXS兲 attached to FESEM. Structural features of the nanomaterial are investigated using high resolution transmission electron microscopy 共HRTEM兲, along with selected area electron diffraction 共SAED兲. Figure 2共a兲 shows a low magnification FESEM image of a dense yield of nanomaterials grown directly on a tungsten substrate. A magnified image of a top view 关Fig. 2共b兲兴 shows vertically oriented nanowires. A side view 关Fig. 2共c兲兴 reveals individual nanowires, as well as those grouped in small bundles. The as-grown nanowires have diameters of 20– 50 nm, with lengths of more than 10 m for a sampling duration of 10 min. EDXS mappings for different growth regions of various sizes, from 10⫻ 10 m2 to single nanowires, are essentially the same with no discernable differences. The EDXS spectra 关Fig. 2共d兲兴 show that the nanowires are composed of W and O, along with negligible amounts of C. The presence of C results most likely from deposition of C on the nanowires from carbon-containing species that diffuse from the reaction zone during processing. Nevertheless, the findings indicate that the as-synthesized materials are nanosized tungsten oxide wires. Tungsten oxide can exhibit different crystal structures, such as cubic and monoclinic WO3, tetragonal WO2.9, and monoclinic W18O49. Figure 3共a兲 presents a low magnification bright-field TEM image of a 20 nm nanowire. Nanocrystallized tungsten oxide platelets 共as indicated by the arrow兲 which come off the surfaces of bundled nanowires during TEM sample preparation can also be observed. The inset in Fig. 3共a兲 is a SAED pattern from the tungsten oxide nano-
FIG. 3. 共Color online兲 共a兲 TEM image showing a 20 nm nanowire and platelets, along with the selected area electron diffraction pattern 共inset兲. 共b兲 HRTEM image of a single tungsten oxide nanowire, along with lattice-plane spacing 共inset兲. 共c兲 HRTEM of a nanowire tip, with ledge growth suggesting VS mechanism.
wire and platelets. The indexed SAED pattern with the first three highest intensities of 3.779, 3.126, and 2.67 Å matches very well with the tetragonal phase of WO2.9 with lattice constants of a = 5.3 Å, b = 5.3 Å, and c = 3.83 Å 共PDF card No. 18-1417兲. These d spacings correspond to 兵110其, 兵101其, and 兵200其, respectively. HRTEM imaging of a typical flame-grown nanowire is shown in Fig. 3共b兲, revealing its dislocation-free, singlecrystalline nature. Analysis of the two-dimensional Fourier transform pattern of the TEM image gives an average spacing for the lattice planes of 3.78 Å, which corresponds to the reflections from d spacings of 共110兲 planes of the tetragonal WO2.9 phase. Thus, these as-grown single-crystalline nanowires have preferable growth along the 关110兴 direction, as indicated in Fig. 3共b兲. The mechanism of formation of these nanowires appears to be by vapor-solid 共VS兲 growth. In the vapor-liquid-solid 共VLS兲 process, the growth is promoted by a liquid-solid interface, generally marked by the presence of droplets at the tips of the nanowires. The morphology of a nanowire is shown in Fig. 3共c兲, evincing no metal nanoparticle at its tip. The TEM image further reveals thickening by a ledgegrowth mechanism, which is additional evidence for vaporphase transport and deposition. Figure 1 shows that the gas-phase temperature at the substrate is ⬃1720 K, with the tungsten substrate temperature estimated to be ⬃1600 K, which is sufficiently high to evaporate tungsten/oxygen species. With available oxygen at the probe location, as displayed in Fig. 1, conditions are then favorable for evaporated tungsten/oxygen species to react with oxygen to form tungsten oxide. At the same time, oxygen can react with the tungsten substrate to form tungsten oxide, which may then evaporate due to its low melting temperature 共⬃1870 K兲. In fact, formation of vapor-phase tungsten oxide species can be quite spontaneous, e.g.,11 共g兲 W共s兲 + O共g兲 2 → WO2 ,
⌬G = − 40.08 kJ/mol K
共1兲 共1723 K兲.
Initial formation of tungsten oxide nanoparticles on the surface of the substrate serves to nucleate the tungsten oxide
Downloaded 16 Jun 2006 to 128.6.73.78. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
243115-3
Appl. Phys. Lett. 88, 243115 共2006兲
Xu et al.
nanowires. Subsequent elongated growth along the 关110兴 direction results from preferential diffusion and condensation of tungsten oxide vapor as adatoms at the tip of a nanowire. With solid-phase WO2 readily transformed into WO3, i.e.,11 1 共g兲 共s兲 WO共s兲 2 + 2 O2 → WO3 ,
⌬G = − 108.24 kJ/mol K
共2兲 共1273 K兲,
we hypothesize that, given the elevated temperature in the probe region 共near the tungsten oxide decomposition temperature of ⬃1720 K兲, decomposition of solid-phase WO3 results in the formation of the final WO2.9 tetragonal phase. The positive temperature gradient extending outward from the 共cooler兲 substrate surface into the 共hotter兲 surrounding gas likely promotes the vertical orientation of the nanowires. The high growth rates that are observed in the flame process are due to favorable local conditions allowing for various mechanisms to occur simultaneously. The presence of water vapor 共a combustion by-product兲 at the probe location 共Fig. 1兲 may further enhance the nanowire growth process. Water vapor cannot only form tungsten oxide, e.g.,15 共g兲 W共s兲 + 3H2O共g兲 → WO共s兲 3 + 3H2 ,
共3兲
but it can also increase the rates of tungsten oxide evaporation and hydrate species formation, e.g.,14 共g兲 WO共s兲 → WO2共OH兲共g兲 3 + H 2O 2 .
共4兲
In addition, the elevated temperatures and radical species, e.g., OH and H, innate in the flame permit nanowire growth without substrate pretreatment or external catalysts. Gu et al.2 observed that a clean and unoxidized tungsten surface was necessary for tungsten oxide nanowire growth. However, similar to Liu et al.,16 we grow tungsten oxide nanowires with high quality, orientation, and purity from oxidized surfaces. Other works have found that H2 reduction2 prior to heat treatment is necessary to grow nanowires on tungsten substrates. Again, this is not necessary in our experiments. As can be seen in Fig. 1, small concentrations of H2 are inherently present at the probe location, which, in combination with the elevated temperatures, can aid in initial nanoparticle formation. Finally, the high characteristic temperature in the flame process induces nanowire growth without the help of catalysts. External catalysts are needed to produce tungsten oxide nanowires in the temperature range of 873– 973 K,2,10,15 while neither catalyst nor pretreatment of substrate is necessary in the temperature range of 1073– 1873 K.1,4,5,7,8,12,16
In summary, we demonstrate the synthesis of vertically aligned, single-crystalline, tetragonal WO2.9 nanowires grown directly from tungsten substrates at high rates in a flame process. The growth mechanism appears to be vaporsolid based, with key parameters being radical species present, oxidizer and water-vapor concentrations, substrate temperature, and gas-phase temperature. Systematic variation of these parameters is readily achieved in our flame technique, which would be very time consuming and tedious in other heat-treatment and chemical vapor deposition 共CVD兲 methods. Finally, our technique is promising for large-scale applications due to its simplicity, scalability, and economy. This work was supported by the National Science Foundation 共NSF-CTS-0213929 and NSF-CTS-0325057兲. 1
Y. Q. Zhu, W. Hu, W. K. Hsu, M. Terrones, N. Grobert, J. P. Hare, H. W. Kroto, D. R. M. Walton, and H. Terrones, Chem. Phys. Lett. 309, 327 共1999兲. 2 G. Gu, B. Zheng, W. Q. Han, S. Roth, and J. Liu, Nano Lett. 2, 849 共2002兲. 3 K. Lee, W. S. Seo, and J. T. Park, J. Am. Chem. Soc. 125, 3408 共2003兲. 4 Y. Li, Y. Bando, and D. Golberg, Adv. Mater. 共Weinheim, Ger.兲 15, 1294 共2003兲. 5 Y. B. Li, Y. Bando, D. Golberg, and K. Kurashima, Chem. Phys. Lett. 367, 214 共2003兲. 6 X.-L. Li, J.-F. Liu, and Y.-D. Li, Inorg. Chem. 42, 921 共2003兲. 7 Z. Liu, Y. Bando, and C. Tang, Chem. Phys. Lett. 372, 179 共2003兲. 8 J. Liu, Y. Zhao, and Z. Zhang, J. Phys.: Condens. Matter 15, L453 共2003兲. 9 X. W. Lou and H. C. Zeng, Inorg. Chem. 42, 6169 共2003兲. 10 H. Qi, C. Wang, and J. Liu, Adv. Mater. 共Weinheim, Ger.兲 15, 411 共2003兲. 11 S. Vaddiraju, H. Chandrasekaran, and M. K. Sunkara, J. Am. Chem. Soc. 125, 10792 共2003兲. 12 Y. Z. Jin, Y. Q. Zhu, R. L. D. Whitby, N. Yao, R. Ma, P. C. P. Watts, H. W. Kroto, and R. M. Walton, J. Phys. Chem. B 108, 15572 共2004兲. 13 H. G. Choi, Y. H. Jung, and D. K. Kim, J. Am. Ceram. Soc. 88, 1684 共2005兲. 14 M. Gillet, R. Delamare, and E. Gillet, Eur. Phys. J. D 34, 291 共2005兲. 15 K. Hong, W. Yiu, H. Wu, J. Gao, and M. Xie, Nanotechnology 16, 1608 共2005兲. 16 K. Liu, D. T. Foord, and L. Scipioni, Nanotechnology 16, 10 共2005兲. 17 S.-J. Wang, C.-H. Chen, R.-M. Ko, Y.-C. Kuo, C.-H. Wong, K.-M. Uang, T.-M. Chen, and B.-W. Liou, Appl. Phys. Lett. 86, 263103 共2005兲. 18 J. Zhou, Y. Ding, S. Z. Deng, L. Gong, N. S. Xu, and Z. L. Wang, Adv. Mater. 共Weinheim, Ger.兲 17, 2107 共2005兲. 19 M. Frenklach, H. Wang, C.-L Yu, M. Goldenberg, C. T. Bowman, R. K. Hanson, D. F. Davidson, E. J. Chang, G. P. Smith, D. M. Golden, W. C. Gardiner, and V. Lissianski, http://www.me.berkeley.edu/gri mech/; and Gas Research Institute Topical Report: M. Frenklach, H. Wang, M. Goldenberg, G. P. Smith, D. M. Golden, C. T. Bowman, C. T. Bowman, R. K. Hanson, W. C. Gardiner, and V. Lissianski, GRI-Mech-An Optimized Detailed Chemical Reaction Mechanism for Methane Combustion, Report No. GRI–95/0058, November 1, 1995. 20 F. Xu, X. Liu, and S. D. Tse, Carbon 44, 570 共2006兲.
Downloaded 16 Jun 2006 to 128.6.73.78. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
