The Solution Growth of Copper Nanowires and Nanotubes is Driven

Dec 5, 2011 - Nano Lett. , 2012, 12 (1), pp 234–239 ... Here we show that the one-dimensional anisotropic growth of Cu NWs and nanotubes (NTs) in so...
3 downloads 9 Views 3MB Size
Letter pubs.acs.org/NanoLett

The Solution Growth of Copper Nanowires and Nanotubes is Driven by Screw Dislocations Fei Meng and Song Jin* Department of Chemistry, University of Wisconsin-Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States S Supporting Information *

ABSTRACT: Copper (Cu) nanowires (NWs) are inexpensive conducting nanomaterials intensively explored for transparent conducting electrodes and other applications. However, the mechanism for solution growth of Cu NWs remains elusive so far. Here we show that the one-dimensional anisotropic growth of Cu NWs and nanotubes (NTs) in solution is driven by axial screw dislocations. All three types of evidence for dislocationdriven growth have been conclusively observed using transmission electron microscopy (TEM) techniques: rigorous twobeam TEM analysis that conclusively characterizes the dislocations in the NWs to be pure screw dislocations along ⟨110⟩ direction, twist contour analysis that confirms the presence of Eshelby twist associated with the dislocation, and the observation of spontaneously formed hollow NTs. The reduction− oxidation (redox) electrochemical reaction forming the Cu NWs presents new chemistry for controlling supersaturation to promote dislocation-driven NW growth. Using this understanding to intentionally manipulate the supersaturation, we have further improved the NW growth by using a continuous flow reactor to yield longer Cu NWs under much milder chemical conditions. The rational synthesis of Cu NWs with control over size and geometry will facilitate their applications. KEYWORDS: Copper, nanowire, nanotube, screw dislocation, supersaturation, reduction−oxidation

T

growth: the presence of dislocation sources and low supersaturation for crystal growth.17−19 In addition to the chemical vapor deposition growth of PbS14,15 and PbSe21 pine tree NWs, GaN NWs,22 Fe-doped ZnO NWs,23 ZnkIn2Ok+3 pyramids with pin holes,24 and In2O3 NTs,25,26 a variety of oxides/hydroxides, such as ZnO,16,17 FeOOH,19 and Co(OH)227 can be readily grown into NWs and NTs morphologies in low-temperature aqueous solutions and their growth has been confirmed to be driven by screw dislocations. Two-dimensional nanoplate morphology has also recently been explained by the screw dislocation-driven growth.28 As a metal, Cu would represent an entirely different class of materials to be grown via the dislocation-driven mechanism. Unlike previously reported dislocation-driven growth of oxides/hydroxides NWs/NTs in aqueous solutions, which only require the hydrolysis of a single precursor metal ion, the formation of Cu NWs proceeds via a reduction−oxidation (redox) reaction that requires the involvement of both a reducing reagent and an oxidizing reagent. The necessity for multiple reactant species increases the complexity of controlling the supersaturation; nevertheless, through the understanding of this complexity we can achieve better control of the supersaturation of such redox reactions to promote dislocation-driven NW growth.

here is growing interest in pursuit of one-dimensional (1D) metal nanostructures1−4 such as nanorods (NRs), nanowires (NWs), and nanotubes (NTs) for a variety of applications including plasmonics, nanoelectronics, chemical sensors, and biotechnology.3,5 In particular, copper (Cu) NWs are being explored as inexpensive, electrically conducting additives to composites and transparent conducting electrodes for solar cell and flexible electronic devices.6−8 Comparing with methods such as electrochemical deposition into porous templates9 and vapor deposition,10,11 the more attractive approach to synthesize Cu NWs for practical applications is the recently reported “spontaneous” formation of high aspect ratio Cu NWs in aqueous solution6,12,13 because of the intrinsic low-cost and large-scale advantages of the solution synthesis. Unlike the solution growth of noble metal (Au, Ag) NWs/NRs that is widely explained by the penta-twinning,1−3 the driving force of the anisotropic growth of these solution-grown Cu NWs remains a mystery.6 In this report, we show that axial screw dislocations are responsible for the 1D solution growth of Cu NWs and we also report the formation of Cu NTs driven by screw dislocations. Furthermore, we utilize the understanding of the growth mechanism to improve Cu NW synthesis by manipulating the supersaturation using a continuous flow reactor under milder chemical conditions. NW growth can be driven by axial screw dislocations, which provide self-perpetuating crystal growth spiral steps at the tip of the NW and break the symmetry of crystal growth.14−20 Only two conditions are required to enable dislocation-driven 1D © 2011 American Chemical Society

Received: September 28, 2011 Revised: November 18, 2011 Published: December 5, 2011 234

dx.doi.org/10.1021/nl203385u | Nano Lett. 2012, 12, 234−239

Nano Letters

Letter

introduced to bind with Cu2+ thus further lowering the supersaturation of the system. Transmission electron microscopy (TEM) characterization was carried out to further confirm the phase identity of the Cu NWs and to confirm that the NW growth is indeed driven by screw dislocations. Initial analysis was attempted on NWs from 90 min reactions, but the thickness of these NWs (∼100 nm) not only prevented us from collecting useful high-resolution TEM (HRTEM) images but also introduced features that proved to be too complicated to analyze (Supporting Information Figure S1). We reason that the thickness of these NWs could have been generated as a result of the convoluted growth modes that occur in prolonged static reactions due to the constantly changing supersaturation.17 By this reasoning, limiting the growth time would limit variation in the supersaturation levels and thus reduce the structural complexity of the NWs. Given the color change observed at 38 min, we would expect to find the thinnest diameter NWs to be present at that time. TEM samples were prepared from a shorter reaction (40 min) where the diameter of the NWs was less than 50 nm. A zero-beam bright-field TEM image reveals a faint contrast feature in the middle of the NW that is suggestive of dislocation contrast (Figure 2a). Selected area electron

To seek insight into the growth mechanism of Cu NWs, we first synthesized Cu NWs using a modified version of the approach developed by Zeng and co-workers.12 Under a typical condition, 10 mL of an aqueous solution composed of 5 mM Cu(NO3)2, 7 M NaOH, 0.2 M ethylenediamine (EDA), and 14 mM hydrazine (N2H4) (molar ratio 1:1400:40:2.8) was sealed in a 20 mL glass vial and heated at 60 °C in an oven for 40 to 90 min. The color of the solution changed from royal blue to bronze after 38 min, indicating the formation of the Cu NWs. Figure 1a shows a scanning electron microscope (SEM) image

Figure 1. (a,b) SEM images of as-synthesized Cu NWs, showing that NWs are grown from seed particles. (c) PXRD of as-synthesized NW samples in comparison with the reference Cu diffractogram.

of the NW products collected from a 90 min reaction. These NWs are approximately 100 nm in diameter and 5−10 μm in length, which is consistent with observations made in previous report.6 Most NWs clearly originate from a spherical “seed” particle (Figure 1a,b), which likely contains the dislocation sources responsible for initiating the NW growth. Powder X-ray diffraction (PXRD, Figure 1c) confirms that the as-synthesized NWs are face-centered cubic (fcc) Cu (PDF 04-0836, space group Fm3̅m, a = 0.3615 nm). The main reaction taking place to form Cu NWs is the reduction of Cu2+ ions (the oxidizing reagent) by the reducing reagent N2H4

2Cu 2 + + 4OH− + N2H4 = 2Cu + N2 + 4H2O

Figure 2. TEM characterization of as-grown Cu NWs from a 40 min reaction without etching. (a) Zero-beam bright-field TEM image of a NW growing from a seed. (b) SAED taken from the center of the NW in (a) showing single-crystal diffraction pattern with some impurity spots. (c) SAED taken from the seed in (a) showing its polycrystalline nature. (d) HRTEM image showing Moiré pattern of an edge dislocation. (e,f) HRTEM image (e) and its corresponding FFT (f) of a single NW, showing [11̅0] growth direction as well as the presence of Cu2O and its resulting Moiré patterns, indicated by the yellow arrows.

(1)

The OH− ion plays dual roles in the reaction. First, it provides a suitable pH environment for the redox reaction to occur (the electrochemical potential is dependent on the solution pH); and second, it is intrinsically a ligand capable of complexing with Cu2+ to form Cu(OH)42−. This complexation lowers the activity (supersaturation) of the Cu2+ ions, which promotes the dislocation-driven growth mechanism. Moreover, EDA, as a strong complexing ligand (rather than as a surfactant), is also

diffraction (SAED) of a NW (Figure 2b) shows the NW to be single-crystalline yet with extra diffraction spots that can be indexed to Cu2O (PDF 05-0667, space group Pn3̅m, a = 0.4270 nm). The SAED of the “seed” particle (Figure 2c) shows it to be polycrystalline. Additional HRTEM images of these seed particles (Supporting Information Figure S2) show that they 235

dx.doi.org/10.1021/nl203385u | Nano Lett. 2012, 12, 234−239

Nano Letters

Letter

are highly defective but have coherent crystal lattice with the NW stems, similar to what was previously observed for the seed particles in the ZnO NW growth.17 These highly defective seed particles with multiple slightly misoriented grains likely provide the source of screw dislocations for driving the NW growth. However, even with much smaller diameters we were only able to obtain HRTEM images with poorly resolved lattice fringes (Figure 2e). The corresponding fast Fourier transform pattern (FFT, Figure 2f) indicates the growth direction of the NW to be [11̅0], which is consistent with previous reports.12 The FFT also exhibits two sets of patterns that can be indexed to Cu and Cu2O. It is well-known that copper oxides spontaneously form at the surface of Cu upon exposure to air. Furthermore, previous studies have shown that Cu nanocrystals suspended in solvent were readily oxidized to Cu2O nanocrystals by dissolved oxygen.29,30 Given this knowledge, the crystalline Cu2O layer observed on the NWs is likely formed while the NWs are in the growth solution and the subsequent cleaning solution. Another interesting feature observed in the HRTEM of these NWs is the numerous Moiré patterns, marked by the yellow arrows/dashed line in Figure 2d,e. A Moiré pattern is an interference pattern created when two gratings are overlaid with a certain angle and/or when they have different periodicities.31 During the development of electron microscopy, before HRTEM technique became widely available, Moiré patterns were often utilized as indirect measurements for identifying microstructures, for example, lattice spacing, dislocations, precipitations, and so forth.32,33 Here the Moiré patterns were generated through the superimposition of the crystalline Cu2O overcoating layer lattice on top of the Cu lattice. Interestingly, we observed Moiré patterns suggestive of edge dislocations in some NWs (Figure 2d, the zoom-in box marked by a yellow dashed line), confirming the defective nature of the as-synthesized NWs. However, despite the fact that Moiré patterns can provide limited information about the microstructures, they conceal details that are essential for the dislocation analysis. Thus, in order to analyze the crystal defects present in the Cu NWs we must remove the surface Cu2O layer. We used glacial acetic acid (HAc) to remove the oxide layer without damaging the Cu NWs.34 In a typical etching process, a TEM grid with as-synthesized Cu NWs was dipped into glacial HAc for three minutes, dried using a piece of Kimwipe paper, and then quickly pumped into the TEM under high vacuum to minimize the regeneration of oxide layer in the air. After the removal of the oxides, both a clean SAED pattern and a clearly lattice-resolved HRTEM image were readily acquired as shown in Figure 3. They confirm the NWs to be purely single-crystal Cu. The HRTEM image (Figure 3b), now free of Moiré patterns, displays clear lattice fringes and exhibits the same [11̅0] growth direction observed prior to etching. More importantly, after the etching treatment, dislocation contrast becomes readily observable. A zero-beam bright-field TEM image (Figure 3a) taken from the [110] zone axis clearly shows a contrast line going through the center of the NW, which is strong evidence for an axial screw dislocation. Alternatively, using diluted N2H4 aqueous and ethanol solutions during the postgrowth treatments and storage can minimize the formation of crystalline Cu2O layer and therefore enable direct HRTEM observation (without the need for an etching step) of the lattice fringes and the dislocation contrast as shown in Figure S2 and S3 in the Supporting Information. Clear dislocation contrast was observed in many NWs that are free of crystalline surface

Figure 3. TEM characterization of surface-etched Cu NWs from a 40 min reaction. (a) Low-resolution bright-field TEM image of a single NW, revealing the dislocation contrast at the center of the NW. The inset shows the corresponding SAED pattern without impurity spots. (b) HRTEM image of a single NW, confirming growth along the [11̅0] direction (inset is the corresponding indexed FFT). (c) Lowresolution bright-field TEM image of a single NW with complicated textures (inset is the corresponding SAED pattern showing superlattice feature). (d) HRTEM image of another NW with superlattice structures (inset is the indexed selected-area FFT).

oxide layer and representative TEM images are collected in Figure S4 in the Supporting Information. However, approximately 10% of the NWs examined do not display dislocation contrast; instead, the TEMs of these NWs show more complex texture and unusual structural characteristics suggestive of superlattices. Figure 3c shows one typical example for which the corresponding SAED appears to display superlattice diffraction with additional periodicity along [110] direction. HRTEM of another NW (Figure 3d) shows similar features and its corresponding FFT suggests a superlattice with 2a unit cell. The origin of these features is not fully understood at present. One hypothesis is that these superlattice structures are related to the micro/nanotwinning that is often seen in fcc metals,1−3 and which might evolve from the original screw dislocations. We have determined the dislocation Burgers vector (b) to be along the [11̅0] direction using diffraction contrast TEM under strong two-beam conditions.31 The bending of crystal planes near the dislocation core causes additional electron diffraction that creates a visible contrast around the dislocation in the TEM image. However, diffraction spots (g) that are normal to the Burgers vector produce no dislocation contrast, which is known as the “invisibility criterion”. Therefore the direction of 236

dx.doi.org/10.1021/nl203385u | Nano Lett. 2012, 12, 234−239

Nano Letters

Letter

is along the same direction as the NW growth axis, so this is a pure screw dislocation. Additional proofs that the Cu NWs form by dislocationdriven mechanism come from the presence of Eshelby twist and the spontaneous formation of hollow tubes. Axial screw dislocations create stress and strain in NWs, which can be alleviated by twisting the crystal lattice around the dislocation core, known as the Eshelby twist (α)36

a Burgers vector can be identified by taking the cross product of two noncollinear g vectors that make the dislocation contrast disappear. The NW shown in Figure 4 was tilted close to the

α=

b

(2) πr 2 where b is the Burgers vector magnitude and r is the NW radius. Even larger dislocation strain can be alleviated by hollowing out the dislocation core to create NTs.17 Here we quantify the lattice twist using twist contours analysis, whose principles have been described in previous reports.16,19 To perform this analysis, any pair of reciprocal lattice vectors (±g vectors) that are noncollinear with the growth axis can be chosen to excite, although it is more convenient to excite ±g vectors that are orthogonal to the growth axis so that no angle correction is needed. In this case, given that the Cu NWs grow along [11̅0] and the most common zone axis is [110], the most probable g vectors to excite are ±(002) with additional contribution to the SAED from ±(1̅13). A zero-beam brightfield image (Figure 5a) is then taken to record the locations of the contour bands so that the separations between them can be measured. Next, each twist contour is uniquely indexed to a specific g vector using displaced-aperture dark-field imaging (Figure 5b e). Two pairs of twist contour bands are clearly present in the bright-field image (Figure 5a), and they are indexed to ±(002) and ±(11̅3) g vectors, respectively. The real-space twisting angle is then calculated using the following equation λ g + − g− α= (3) 2L sin θ

Figure 4. Diffraction contrast TEM imaging of the dislocation in the Cu NW. (a,b) Schematic superposition of real and reciprocal space of a dislocated NW along [112] (a) and [001] (b) zone axis illustrating the process of finding g vectors that satisfy the “invisibility criterion”. (c−e) TEM images with corresponding SAED patterns under strong two-beam conditions. Panels c and e represent invisibility conditions when the dislocation contrast disappears; panel d represents g||b conditions when the dislocation contrast is the strongest, as highlighted in the zone axis diagram (a,b).

[112] zone axis and the [001] zone axis (Figure 4a,b, respectively). The image with the (22̅0) diffraction spot (Figure 4d) shows strong dislocation contrast (corresponding to the g||b contrast maximum), while the dislocation meets the invisibility criterion under the perpendicular (222̅) and (220) spots (Figure 4c,e). Therefore, taking the cross product of the (222̅) and (220) vectors yields a Burgers vector along the [11̅0] direction, which is known to be the most common Burgers vector direction in fcc metals.35 In addition, the Burgers vector

where λ is the wavelength of the electron (2.51 pm here), L is the measured real-space distance between each pair of twist contours [39.5 nm for ±(002), and 65.8 nm for ±(113̅ )], g+ and g− are two opposing reciprocal lattice vectors, and θ is the angle between ±g and the growth direction [90° for ±(002),

Figure 5. (a−e) Twist contour analysis of Cu NWs and (f) TEM image of a hollow Cu NT. (a) Zero-beam bright-field TEM image of a representative NW, showing four indexed twist contours (inset is the SAED pattern). (b−e) Displaced-aperture dark-field TEM images used to index the four labeled contours. 237

dx.doi.org/10.1021/nl203385u | Nano Lett. 2012, 12, 234−239

Nano Letters

Letter

and 65° for ±(113̅ )]. The calculated twist for the NW shown in Figure 5a is approximately 20°/μm using data from either ±(002) or ±(11̅3) contours. This is in agreement with the Eshelby twist predicted for a NW with a radius of 22.5 nm and a Burgers vector magnitude of approximately 0.5 nm, which is about twice the magnitude of the elementary Burgers vector in fcc Cu (1/2[11̅0] = 0.255 nm for Cu metal).35 Although most of the reaction products observed are solid NWs, hollow single crystal NTs were occasionally present (Figure 5f and Supporting Information Figure S5). In addition to twisting, hollowing of NWs can result from the relaxation of dislocation strain energy for dislocations with large Burgers vectors. This was observed and clearly explained for the spontaneous formation of ZnO NTs17 and has been further observed for the formation of GaN22 and In2O3 NTs.25,26 Note that the NT shown in Figure 5f was imaged without the oxide etching treatment, and as such a fuzzy oxide overcoating layer is visible. At this point, all three types of evidence for dislocationdriven 1D growth have been observed for the Cu NWs: dislocation contrast, Eshelby twist, and the formation of hollow NTs. These experiments unequivocally confirm that the growth of these Cu NWs is indeed driven by screw dislocations. With the understanding that the growth of Cu NWs is driven by dislocation, we then re-examined and improved the synthesis conditions of Cu NWs. While the previously reported static reaction produces a relatively high yield of NWs with a decent aspect ratio, it still has two drawbacks. First, an extremely high concentration of NaOH (7 M) is required; and second, NW lengths are limited due to precursor depletion. Understanding and controlling the supersaturation of the crystal growth is the key for promoting dislocation-driven growth.17−19 In a redox reaction system such as the formation of Cu NWs herein, the best way to represent the equilibrium position of the chemical reaction (and crystal growth) is the electrochemical potential; chemical equilibrium is reached when the electrochemical potential of the reaction is zero. The supersaturation of the redox reaction system is simply the electrochemical overpotential. By tuning the electrochemical potential to be slightly over zero, low supersaturation condition is achieved, which favors the dislocation crystal growth regime. For the redox reaction between Cu2+ and N2H4, changing the concentration of Cu2+, either by changing the precursor concentrations or by complexation to suppress the concentration of free Cu2+, only represents one-half cell reaction for supersaturation control. Tuning the concentration of N2H4 also influences the supersaturation of the reaction system and thus the morphology of the final products as shown in Supporting Information Figure S6. With regard to the high pH, since one of the roles of NaOH is complexing Cu2+ and reducing the concentration of the free Cu2+ ions, thus maintaining the low supersaturation required to promote dislocation-driven growth, we reason that by compensating the reduction in NaOH concentration with an increase in the other complexing ligand, EDA, the low supersaturation can be maintained and successful dislocation-driven NW growth may still be achieved. Indeed, a modified static reaction with a much lower concentration of NaOH (pH ≈ 12) still led to the formation of Cu NWs: 10 mL aqueous solution of 0.5 mM Cu(NO3)2, 10 mM NaOH, 37.5 mM EDA, and 4 mM N2H4 (molar ratio 1:20:75:8) was sealed in a 20 mL glass vial and heated at 60 °C for 4 h in a oven. The color of this solution was purple [color of Cu(EDA)2], instead of royal blue [color of Cu(OH)42‑] observed at higher pH as stated above, which suggests that the primary forms of Cu(II)

are different in these two conditions since the ratios of the complexing ligands varied. We note that EDA is a much stronger ligand (β[Cu(EDA)2] = 1020) than OH− (β[Cu(OH)42−] = 1016); therefore, a modest increase in EDA concentration can sufficiently compensate for a two-order of magnitude drop in NaOH concentration. The often repeated argument of treating EDA as a surfactant is not essential for the growth of these NWs and does not explain the observation discussed above. Furthermore, according to reaction 1, OH− ions participate in the redox reaction that forms Cu, therefore reducing the NaOH concentration also lowers the redox overpotential and crystal growth supersaturation, thus favoring the dislocation-driven growth. NWs yielded from this protocol are typically ∼5 μm in length and 200−300 nm in diameters with a little tapering (Figure 6a,b), indicating precursor depletion is occurring as the reaction proceeds.

Figure 6. (a,b) SEM images of Cu NWs from the low pH static reaction. (c) Schematic of the CFR for Cu NW growth. (d,e) SEM images of Cu NWs from the CFR reaction.

A continuous flow reactor (CFR, Figure 6c) was then employed to solve the precursor depletion problem. Precursor solutions of the same concentration of each component at lower pH condition identified above were flowed through the CFR at 60 °C for 4 h (see Supporting Information for experimental details.). Much longer Cu NWs (about 30 μm on average, up to 50 μm) were obtained (Figure 6d,e). The new lengths are 6 times longer than the products from the static reaction under comparable conditions. In addition to increased lengths, the diameters of NWs from the CFR reaction are more uniform (Figure 6e), because the precursor concentrations are maintained at constant low level.17,18 Even longer NWs can be achieved if the reactions are carried out for longer time. 238

dx.doi.org/10.1021/nl203385u | Nano Lett. 2012, 12, 234−239

Nano Letters

Letter

(15) Lau, Y. K. A.; Chernak, D. J.; Bierman, M. J.; Jin, S. J. Am. Chem. Soc. 2009, 131, 16461−16471. (16) Morin, S. A.; Jin, S. Nano Lett. 2010, 10, 3459−3463. (17) Morin, S. A.; Bierman, M. J.; Tong, J.; Jin, S. Science 2010, 328, 476−480. (18) Jin, S.; Bierman, M. J.; Morin, S. A. J. Phys. Chem. Lett. 2010, 1, 1472−1480. (19) Meng, F.; Morin, S. A.; Jin, S. J. Am. Chem. Soc. 2011, 133, 8408−8411. (20) Sears, G. W. Acta Metall. 1955, 3, 367−369. (21) Zhu, J.; Peng, H. L.; Marshall, A. F.; Barnett, D. M.; Nix, W. D.; Cui, Y. Nat. Nanotechnol. 2008, 3, 477−481. (22) Jacobs, B. W.; Crimp, M. A.; McElroy, K.; Ayres, V. M. Nano Lett. 2008, 8, 4353−4358. (23) Aleman, B.; Ortega, Y.; Garcia, J. A.; Fernandez, P.; Piqueras, J. J. Appl. Phys. 2011, 110, 014317. (24) Bartolome, J.; Maestre, D.; Amati, M.; Cremades, A.; Piqueras, J. J. Phys. Chem. C 2011, 115, 8354−8360. (25) Maestre, D.; Haussler, D.; Cremades, A.; Jager, W.; Piqueras, J. Cryst. Growth Des. 2011, 11, 1117−1121. (26) Maestre, D.; Haussler, D.; Cremades, A.; Jager, W.; Piqueras, J. J. Phys. Chem. C 2011, 115, 18083−18087. (27) Li, Y.; Wu, Y. Chem. Mater. 2010, 22, 5537−5542. (28) Morin, S. A.; Forticaux, A.; Bierman, M. J.; Jin, S. Nano Lett. 2011, 10, 4449−4455. (29) Hung, L. I.; Tsung, C. K.; Huang, W. Y.; Yang, P. D. Adv. Mater. 2010, 22, 1910−1914. (30) Han, M. Y.; Ye, E. Y.; Zhang, S. Y.; Liu, S. H. Chem.Eur. J. 2011, 17, 3074−3077. (31) Williams, D. B.; Carter, C. B. Transmission Electron Microscopy: a Textbook for Materials Science, 1st ed.; Plenum Press: New York, 1996. (32) Bassett, G. A.; Menter, J. W.; Pashley, D. W. Proc. Royal Soc. London, Ser. A 1958, 246, 345−368. (33) Pashley, D. W.; Menter, J. W.; Bassett, G. A. Nature 1957, 179, 752−755. (34) Chavez, K. L.; Hess, D. W. J. Electrochem. Soc. 2001, 148, G640−G643. (35) Nabarro, F. R. N. Dislocation in Solids, Vol. 2, Dislocations in Crystals, 1st ed.; North-Holland Publishing Company: New York, 1979. (36) Eshelby, J. D. J. Appl. Phys. 1953, 24, 176−179.

In summary, we have shown that the solution growth of Cu NWs and NTs is driven by screw dislocations. The presence of axial screw dislocations is confirmed by rigorous diffraction contrast TEM under strong two-beam conditions, twist contour analysis, and the presence of hollow NTs. By tuning the concentration of complexing ligands (OH− and EDA), low supersaturation condition of this redox reaction system was created to favor dislocation-driven crystal growth, enabling more effective Cu NW growth under a milder pH environment (pH ≈ 12). Furthermore, longer Cu NWs were achieved using the CFR. These results further support the generality of dislocation-driven nanomaterial growth by expanding it to both a new class of materials, metals, and a new type of chemistry, the redox reactions. The understanding of metal NW growth and the improved rational design of their large scale, solution synthesis will promote their utility for practical applications.



ASSOCIATED CONTENT

S Supporting Information *

Details of experimental procedures, SEM images showing the influence of hydrazine concentration, and additional TEM images of the Cu NWs and NTs. This material is available free of charge via the Internet at http://pubs.acs.org.

■ ■

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected]. ACKNOWLEDGMENTS This research is supported by NSF Grant DMR-1106184. S.J. also thanks the Sloan Research Fellowship, Research Corporation SciaLog Award, and Honeywell University Affiliate Fund for support. We thank Ms. Rachel Selinsky for her comments and revisions on the manuscript.



REFERENCES

(1) Murphy, C. J.; Sau, T. K.; Gole, A. M.; Orendorff, C. J.; Gao, J.; Gou, L.; Hunyadi, S. E.; Li, T. J. Phys. Chem. B 2005, 109, 13857− 13870. (2) Xia, Y. N.; Chen, J. Y.; Wiley, B. J. Langmuir 2007, 23, 4120− 4129. (3) Xia, Y.; Xiong, Y.; Lim, B.; Skrabalak, S. E. Angew. Chem., Int. Ed. 2009, 48, 60−103. (4) Yang, P. D.; Tao, A. R.; Habas, S. Small 2008, 4, 310−325. (5) Murphy, C. J.; Gole, A. M.; Stone, J. W.; Sisco, P. N.; Alkilany, A. M.; Goldsmith, E. C.; Baxter, S. C. Acc. Chem. Res. 2008, 41, 1721− 1730. (6) Rathmell, A. R.; Bergin, S. M.; Hua, Y. L.; Li, Z. Y.; Wiley, B. J. Adv. Mater. 2010, 22, 3558−3562. (7) Wu, H.; Hu, L.; Rowell, M. W.; Kong, D.; Cha, J. J.; McDonough, J. R.; Zhu, J.; Yang, Y.; McGehee, M. D.; Cui, Y. Nano Lett. 2010, 10, 4242−4248. (8) Kang, M.-G.; Park, H. J.; Ahn, S. H.; Guo, L. J. Sol. Energ. Mat. Sol. Cells 2010, 94, 1179−1184. (9) Neumann, R.; Toimil-Molares, M. E.; Buschmann, V.; Dobrev, D.; Scholz, R.; Schuchert, I. U.; Vetter, J. Adv. Mater. 2001, 13, 62−65. (10) Choi, H.; Kim, C.; Gu, W. H.; Briceno, M.; Robertson, I. M.; Kim, K. Adv. Mater. 2008, 20, 1859−1863. (11) Choi, H.; Park, S. H. J. Am. Chem. Soc. 2004, 126, 6248−6249. (12) Chang, Y.; Lye, M. L.; Zeng, H. C. Langmuir 2005, 21, 3746− 3748. (13) Konya, Z.; Mohl, M.; Pusztai, P.; Kukovecz, A.; Kukkola, J.; Kordas, K.; Vajtai, R.; Ajayan, P. M. Langmuir 2010, 26, 16496−16502. (14) Bierman, M. J.; Lau, Y. K. A.; Kvit, A. V.; Schmitt, A. L.; Jin, S. Science 2008, 320, 1060−1063. 239

dx.doi.org/10.1021/nl203385u | Nano Lett. 2012, 12, 234−239