Letter pubs.acs.org/NanoLett
In Situ Study of Noncatalytic Metal Oxide Nanowire Growth Simas Rackauskas,*,† Hua Jiang,† Jakob B. Wagner,‡ Sergey D. Shandakov,§ Thomas W. Hansen,‡ Esko I. Kauppinen,† and Albert G. Nasibulin*,†,∥ †
Department of Applied Physics, Aalto University School of Science, Puumiehenkuja 2, 00076 Espoo, Finland Center for Electron Nanoscopy, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark § Kemerovo State University, 6, Krasnaya Street, Kemerovo 650043, Russia ∥ Skolkovo Institute of Science and Technology, 100 Novaya Street, Skolkovo, Odintsovsky District, Moscow Region 143025, Russia ‡
S Supporting Information *
ABSTRACT: The majority of the nanowire synthesis methods utilize catalyst particles to guide the nanowire geometry. In contrast, catalyst-free methods are attractive for facile fabrication of pure nanowires without the need for catalyst preparation. Nonetheless, how nanowire growth is guided without a catalyst is still widely disputed and unclear. Here, we show that the nanowire growth during metal oxidation is limited by a nucleation of a new layer. On the basis of in situ transmission electron microscope investigations we found that the growth occurs layer by layer at the lowest specific surface energy planes. Atomic layers nucleate at the edge of twin boundary ridges and form a long-range ordering along the twin boundary. We anticipate our study to be a starting point to employ defects for nanowire growth control and consequently shaping the geometry of nanowires in a similar manner as in the catalyst-assisted growth method. KEYWORDS: nanowires, in situ TEM, noncatalytic growth, metal oxide, nucleation, atomic layer
S
ynthesis mechanisms, such as vapor−liquid−solid1,2 or selfcatalytic3 growth, are based on the use of catalyst particle, which determine the nanowire (previously called whisker) geometry. In contrast, catalyst particles are not used in vapor− solid4 or metal oxide and sulfide5−9 nanowire growth methods. This gives an advantage of pure nanowires and minimizes the number of technological steps. Spontaneous nanowire formation by noncatalytic methods is explained by the dislocations present in specific directions10,11 or the growth anisotropy of various facets.12 Moreover it has been recently shown that even if dislocations are present during the growth, they might disappear from the final product.13 Frank et al.10,11 pointed out that dislocations or line defects, such as the self-perpetuating steps around a screw dislocation, provided facile spiral growth fronts for crystal growth, and recently, screw dislocations have been demonstrated to be responsible for tubular and cylindrical growth of nanowires.14−16 However, self-perpetuating steps fail to explain the growth in case of planar defects, such as twin boundaries. Nanowires were explained to form in preferential direction, determined by surface energy and chemical activity; however, the details of the process, that is, how atoms can be rationally assembled into wire-like structures, are not fully understood.17 The influence of twin boundaries was explained by the twin-plane re-entrant mechanism only for the formation of platelets18 and for catalytic nanowire growth.19 Dynamical studies of the nanowire growth at the atomic level will give invaluable input in the understanding of the growth mechanics in catalyst-free growth. Control of the defects acting as © XXXX American Chemical Society
nucleation centers provides a possibility to regulate the growth of 1D structures. This work is to the best of our knowledge the first lattice resolved in situ investigation of noncatalytic growth of nanowires. Here, on the basis of in situ transmission electron microscope (TEM) observations of CuO nanowire growth, we present a new insight into the nanowire growth. CuO nanowires were synthesized in an environmental TEM by oxidation of pure metallic Cu in the presence of 3.2−7 mbar of O2 in the temperature range of 350−475 °C. The nanowires were grown directly on the surface of copper without surface treatment. In situ TEM observations revealed layer by layer growth at the tip of the nanowire. Figure 1a and g show a schematic view of two nanowire orientations as observed during the in situ CuO nanowire growth in Figure 1b−f and h−m, respectively (Supporting Information Movies 1 and 2). The resulting nanowires have a twin plane in the middle, as confirmed after the growth. As the twin plane is not parallel to the viewing direction during growth, Moire pattern is observed in the center region. The kinetics of the growth is followed at a spatial resolution revealing the atomic steps and lattice fringe spacing of 0.23 nm attributed to the CuO (111) planes. Furthermore, Received: July 15, 2014 Revised: August 28, 2014
A
dx.doi.org/10.1021/nl502687s | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 2. TEM characterization of CuO nanowires. (a) Bright-field TEM image of a CuO nanowire grown at 400 °C at atmospheric pressure, showing a twin boundary, clearly visible in the middle of the nanowire on its growth axis. The left crystal is labeled “(L)” and the right crystal “(R).” (b) Selected-area electron diffraction pattern of the nanowire taken from inside the region indicated by the red dotted circle in (a). (c) High-resolution TEM image of the twin boundary structure corresponding to the region in the square red frame in (a). The arrow marks the axial direction of the nanowire, which is not associated with any well-defined crystallographic direction in the structure. (d, e) Dark-field TEM images of the nanowire taken from the left crystal using the (1̅1̅1)L reflection d, and from the right crystal using the (11̅1)R reflection (e). (f) The same structure is found in a CuO nanowire grown at 450 °C, showing (11̅1̅) and (1̅1̅1) lattice planes on each side of the tip of the nanowire. The insets show fast Fourier transform images from the areas of the crystals indicated by red squares.
Figure 1. In situ TEM observation of CuO nanowire growth. Schematics of the nanowire viewing parallel (a) and perpendicular (g) to the twin boundary is followed from the in situ TEM of lateral growth: (b−f) atomic steps propagate from the tip and sideways, (h−j) from left to right and (k−m) from right to left. White arrows show atomic layer edge, black arrows show layer growth direction. Lattice fringe spacing of 0.23 nm attributed to the CuO (111) planes. Scale bar is 2 nm (see Supporting Information Movie 1 and 2).
diffraction pattern (Figure 2b) are caused by twinned crystals, indicated by labels “(L)” (left) and “(R)” (right) on either side of the twin boundary. The diffraction pattern is indexed using the monoclinic CuO structure (4.69 Å, 3.43 Å, 5.13 Å, 99.55°), giving the (002) twin plane to be parallel to the growth axis. The twinned crystals are well highlighted in the corresponding dark-field TEM images (Figure 2d and e), which were acquired using the (1̅1̅1)L and (111)R reflections, respectively. As found during the ex situ observations, the tip planes had crystallographic orientations of (111̅ )̅ and (11̅ 1̅ ), both inclined to the nanowire axis by 65.9° (Figure 2f). The tip of the nanowire retained the faceted shape during growth (see Supporting Information Movies 1−3). Our in situ observations show that the growth initiates at the edge of the nanowire ridge, where (11̅1̅), (1̅1̅1) facets and the twinning plane (002) intersect. Nucleation is energetically favorable on the surface with the lowest specific surface energy, which for CuO is (111) plane.20 On the basis of our kinetic studies of the atomic step dynamics (Figure 3a), we notice a few consistent patterns: (i) there is a delay in the formation of a new layer (an induction period), which varies randomly from layer to layer (Figure 3b); (ii) nanowire growth rate is constant over the entire growth period (Figure 3c). We found a strong correlation of the nanowire growth rate with the duration of the experiment (heating time of the sample
the thickness of a single layer (marked by arrow in Figure 1b and h) is 0.23 nm indicating an atomic layer-by-layer growth of the (111) plane. The steps are observed to propagate both from left to right (Figure 1b−j) and from right to left (Figure 1k−m). No catalyst particles are involved in the growth of the nanowires at any stage. We can unambiguously rule out catalytic, droplet-mediated, and the root-growth mechanisms to explain the nanowire growth. CuO nanowires were grown ex situ in a vacuum chamber in an O2 partial pressure ranging from 0.2 mbar to the ambient atmospheric partial pressure at similar temperatures as that of the in situ experiments. CuO nanowire growth was observed at O2 partial pressures ≥4 mbar (Supporting Information Figure S1). Higher O2 pressures resulted in the higher density of nanowires, but no further structure−pressure correlation was observed. Twin boundaries (planar defects) parallel to the nanowire growth axis were revealed in all examined nanowires as the only structural defects. As an example, Figure 2a shows a typical bright field TEM image of a CuO nanowire grown at 400 °C, revealing a clear twin boundary in the middle of the nanowire along its growth axis. The diffraction doublets in the electron B
dx.doi.org/10.1021/nl502687s | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 3. Kinetics of nanowire growth. (a) In situ TEM observation of CuO nanowires’ length dependence on time shows the linear behavior at any growth rate. (b) Induction period of 26 atomic layers, showing random variation from layer to layer (measured from Supporting Information Movie S1). (c) Lateral growth of atomic layers at the tip of the nanowire. (d) Model of a CuO nanowire with a critical nucleus at the edge of twin boundary ridge. Cu atoms are shown as blue spheres, oxygen as red.
emerges spontaneously (Figure 3b) and at the same time the induction period is longer than the lateral and ridge growth time. In other words, time required for the filling of a layer of CuO in the lateral and ridge direction is shorter than the time needed to nucleate a new layer (Figure 3c). In our case the nanowire growth is not determined by diffusion, since the time dependence of layer filling and individual nanowires’ length change does not follow the parabolic law as can be seen from Figure 3a (Supporting Information Section III). Therefore, the nanowire growth should be described as the layer by layer formation process, in which nucleation of a new layer is the growth limiting stage. We believe that twin boundaries in the nanowires grown by noncatalytic methods serve as a point for preferential nucleation. It was recently reported19 for catalytic nanowire synthesis that a twin boundary effectively reduces the nucleation barrier energy. Nucleation is also preferred on the edge of the crystal, when compared to the center of the plane because of the Berg effect,24−27 which results in the higher CuO concentration at the edges of twin boundary. To summarize, on the basis of our in situ environmental TEM observations of CuO NW growth, we show a lattice resolved atomic step kinetics of the noncatalytic, twin boundary
at a set temperature) and temperature. The experimental duration determines the thickness of Cu2O and CuO layers underneath the nanowires (Supporting Information Figure S2), therefore the diffusion length of Cu, which in turn defines the concentration of Cu atoms on the surface of CuO.21 It is worth noting that the typical nanowire growth time was in the range of a few minutes, whereas the whole experiment might take up to 2 h. The experimental results show a temperature dependence of the nanowire growth rate (at the same duration of experiment) with the activation energy of 37 kJ/mol (see Supporting Information Figure S3). Apparently, this activation energy is attributed to the process, which limits the nanowire formation and therefore determines its growth rate. During the thermal metal oxidation, the nanowire growth can be limited by copper diffusion (along the grain boundaries of the oxide layers19,22 and to the nanowire tip) or by nucleation at the growing tip of the nanowire.23 We believe that the nanowire growth is limited by nucleation of a new layer. The induction period is caused by the time required to form a nucleus on a growing plane. Indeed, we estimate the energy of this atomic layer formation to be equal to a nucleus comprising of six CuO molecules as it is shown in Figure 3d. (see Supporting Information Section IV). A new atomic layer C
dx.doi.org/10.1021/nl502687s | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
(12) Cao, G.; Wang, Y. Nanostructures and Nanomaterials: Synthesis, Properties, and Applications; World Scientific: Singapore, 2011; p 581. (13) Meng, F.; Estruga, M.; Forticaux, A.; Morin, S. A.; Wu, Q.; Hu, Z.; Jin, S. ACS Nano 2013, 7, 11369−11378. (14) Morin, S. A.; Bierman, M. J.; Tong, J.; Jin, S. Science 2010, 328, 476−480. (15) Bierman, M. J.; Lau, Y. K. A.; Kvit, A. V.; Schmitt, A. L.; Jin, S. Science 2008, 320, 1060−1063. (16) Meng, F.; Morin, S. A.; Forticaux, A.; Jin, S. Acc. Chem. Res. 2013, 46, 1616−1626. (17) Wang, Z. L. Adv. Mater. 2003, 15, 432−436. (18) Faust, J. W.; John, H. F. J. Phys. Chem. Solids 1964, 25, 1407− 1415. (19) Gamalski, A. D.; Voorhees, P. W.; Ducati, C.; Sharma, R.; Hofmann, S. Nano Lett. 2014, 14, 140214104227007. (20) Hu, J.; Li, D.; Lu, J. G.; Wu, R. J. Phys. Chem. C 2010, 114, 17120−17126. (21) Young, J. High Temperature Oxidation and Corrosion of Metals; Corrosion Series; Elsevier: Amsterdam, 2008; Vol. 1, pp 29−79. (22) Yuan, L.; Wang, Y.; Mema, R.; Zhou, G. Acta Mater. 2011, 59, 2491−2500. (23) Tang, W.; Picraux, S. T.; Huang, J. Y.; Gusak, A. M.; Tu, K.-N.; Dayeh, S. a. Nano Lett. 2013, 13, 2748−2753. (24) Berg, W. F. Proc. R. Soc. A Math. Phys. Eng. Sci. 1938, 164, 79− 95. (25) Golovin, A. A.; Davis, S. H.; Nepomnyashchy, A. A. Phys. D, Nonlinear Phenom. 1998, 122, 202−230. (26) Nanev, C. N. Prog. Cryst. Growth Charact. Mater. 1997, 35, 1− 26. (27) Kuroda, T.; Irisawa, T.; Ookawa, A. J. Cryst. Growth 1977, 42, 41−46.
driven nanowire growth. Our results reveal the essence of the existing defects, acting as preferential nucleation site, during the growth of nanowires by metal oxidation process. One can anticipate that it will be possible to use defects to drive and even control the process of growth and the geometry of nanowires in a similar manner to what can be done with the help of catalyst particles. Although we have investigated the formation of only CuO nanowires, similar growth mechanisms could be also extended to noncatalytic growth of other nanowires and ionic crystals grown by oxidation−reduction reactions.
■
ASSOCIATED CONTENT
* Supporting Information S
Three in situ TEM movies, SEM images, Arrhenius plot, schematic of the nanowire ridge, and additional information regarding the methods, nanowire growth kinetic regime, and nucleation of atomic layers. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected] *E-mail:
[email protected]. Author Contributions
A.G.N., S.R., and E.I.K. conceived and designed the experiments. J.B.W., T.W.H., S.R., H.J., and A.G.N. performed in situ growth experiments. S.D.S., S.R., A.G.N., and H.J. analyzed and interpreted the data. S.R., A.G.N., S.D.S., and H.J. cowrote the paper. All authors discussed the results and commented on the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The A.P. Møller and Chastine Mc-Kinney Møller Foundation is gratefully acknowledged for their contribution towards the establishment of the Center for Electron Nanoscopy in the Technical University of Denmark. S.D.S. acknowledges the Ministry of Education and Science of Russia (State assignment No 3.392.2014K).
■
REFERENCES
(1) Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89. (2) Hofmann, S.; Sharma, R.; Wirth, C. T.; Cervantes-Sodi, F.; Ducati, C.; Kasama, T.; Dunin-Borkowski, R. E.; Drucker, J.; Bennett, P.; Robertson, J. Nat. Mater. 2008, 7, 372−375. (3) Oh, S. H.; Chisholm, M. F.; Kauffmann, Y.; Kaplan, W. D.; Luo, W.; Rühle, M.; Scheu, C. Science 2010, 330, 489−493. (4) Pan, Z. W.; Dai, Z. R.; Wang, Z. L. Science 2001, 291, 1947− 1949. (5) Takagi, R. J. Phys. Soc. Jpn. 1957, 12, 1212−1218. (6) Nasibulin, A. G.; Rackauskas, S.; Jiang, H.; Tian, Y.; Mudimela, P. R.; Shandakov, S. D.; Nasibulina, L. I.; Jani, S.; Kauppinen, E. I. Nano Res. 2009, 2, 373−379. (7) Wang, R.; Chen, Y.; Fu, Y.; Zhang, H.; Kisielowski, C. J. Phys. Chem. B 2005, 109, 12245−12249. (8) Hiralal, P.; Unalan, H. E.; Wijayantha, K. G. U.; Kursumovic, A.; Jefferson, D.; Macmanus-Driscoll, J. L.; Amaratunga, G. A. J. Nanotechnology 2008, 19, 455608. (9) Zhang, W.; Yang, S. Acc. Chem. Res. 2009, 42, 1617−1627. (10) Frank, F. C. Discuss. Faraday Soc. 1949, 5, 48. (11) Burton, W. K.; Cabrera, N.; Frank, F. C. Philos. Trans. R. Soc. London. Ser. A, Math. Phys. Sci. 1951, 243, 299−358. D
dx.doi.org/10.1021/nl502687s | Nano Lett. XXXX, XXX, XXX−XXX
