Langmuir 2006, 22, 2851-2855
2851
Sintering of Passivated Gold Nanoparticles under the Electron Beam Yu Chen,*,† Richard E. Palmer,† and Jess P. Wilcoxon‡ Nanoscale Physics Research Laboratory, UniVersity of Birmingham, Birmingham B15 2TT, United Kingdom and Nanostructures and AdVanced Materials Department, Sandia National Laboratory, Albuquerque, New Mexico 87185 ReceiVed December 7, 2005. In Final Form: January 19, 2006 Time-lapse studies of a film of passivated gold nanoparticles under electron beam irradiation have been performed using a transmission electron microscope, revealing the microscopic dynamics of the sintering process at the single nanoparticle level. It is found that the sintering of individual passivated gold nanoparticles under electron irradiation is local and mainly depends on the sensitivity of the passivating ligands to the electron beam. A multilayer film is less stable than monolayer film, consistent with the enhanced generation of secondary electrons. The observations also reveal a significant difference between the sintering of passivated nanoparticles and bare metal particles, especially regarding the size effect on the sintering rate. The formation of a neck between adjacent nanoparticles further indicates a mechanism driven by surface diffusion rather than Ostwald ripening at the initial sintering stage.
Passivated nanoparticles have attracted intensive interest in recent years. Advances in synthesis techniques enable the production of particles with controlled size and structure from various materials including metals, alloys, oxides, and semiconductors.1 The unique properties of the nanoparticles intrinsic to their nanoscale size suggest potential applications in, e.g., catalysis, optics, and electronic devices. Self-assembly of these nanoparticles in a controlled way2,3 demonstrates the possibility of using nanoparticles as building blocks to create nanoscale architectures and components of novel devices. The stability of the nanoparticles is critical for all these applications. Recent work has revealed that electron beam irradiation can modify a nanoparticle assembly. Indeed, electron beam writing has been employed to create lines and patterns in films of passivated nanoparticles.4-9 It was found that exposing a thin film of alkanethiol-passivated gold nanoparticles to an electron beam with an energy of 9 keV, with a typical dose of 5 × 105 µC/cm2, resulted in the formation, possibly through cross-linking between the ligands of adjacent particles, of a carbonaceous network in which gold cores were distributed.9 It was also reported that removal of the organic ligands by heating or electron beam irradiation leads to coalescence and sintering of metal cores.10 In this case, wires were fabricated through the * Corresponding author. E-mail:
[email protected]. † University of Birmingham. ‡ Sandia National Laboratory. (1) (a) Daniel, M. C.; Astruc, D. Chem. ReV. 2004, 104, 293. (b) Li, Z. L.; Yuan, J.; Chen, Y.; Palmer, R. E.; Wilcoxon, J. P. AdV. Mater. 2005, 17, 2885. (c) Li, Z. Y.; Yuan, J.; Chen, Y.; Palmer, R. E.; Wilcoxon, J. P. Appl. Phys. Lett. 2005, 87, 243103. (d) Erwin, C. S.; Zu, L.; Hafter, M. I.; Efros, A. L.; Kennedy, T. A.; Norris, D. J. Nature 2005, 436, 91. (2) Kiely, C. J.; Fink, J.; Brust, M.; Bethell, D.; Schiffrin, D. J. Nature 1998, 396, 444. (3) Mendes, P. M.; Chen, Y.; Palmer, R. E.; Nikitin, K.; Fitzmaurice, D.; Preece, J. A. J. Phys.: Condens. Matter 2003, 15, S3047. (4) Bedson, T. R.; Palmer, R. E.; Wilcoxon, J. P. Appl. Phys. Lett. 2001, 78, 2061. (5) Bedson, T. R.; Jenkins, T. E.; Hayton, D. J.; Wilcoxon, J. P.; Palmer, R. E. Appl. Phys. Lett. 2001, 78, 1921. (6) Chen, Y.; Palmer, R. E. In Encyclopaedia of Nanoscience and Nanotechnology; Nalwa, H. S., Ed.; American Scientific Publishers: Valencia, CA, 2004; Vol. 3, pp 17-28. (7) Lin, X. M.; Parthasarathy, R.; Jaeger, H. M. Appl. Plys. Lett. 2001, 78, 1915. (8) Werts, M. H. V.; Lambert, M.; Bourgoin, J.; Brust, M. Nano Lett. 2002, 2, 43. (9) Plaza, J. L.; Chen, Y.; Jacke, S.; Palmer, R. E. Langmuir 2005, 21, 1556.
Figure 1. Bright field TEM images of dodecanethiol-passivated gold nanoparticles with core diameter 4.8 nm (a) before and after focused 200 keV electron beam irradiation with a dose of (b) 7.1, (c) 16.4, (d) 33.7, (e) 73.7, and (f) 149.8 µC/µm2.
sintering of nanoparticles assembled on a linear template by annealing at 300 °C for 1 h. Thus, the sintering of nanoparticles has been both reported and exploited in nanowire fabrication. However, the mechanism of the sintering process is not clear. In this work, we report time-lapse studies of a passivated gold nanoparticle film under electron beam irradiation, performed using a transmission electron microscope (TEM), which disclose the microscopic dynamics of the sintering process at the single nanoparticle level. The experiment employed a 200 keV TEM with field emission gun (FEI TECNAI F20) which incorporates a novel liquid helium cooled cryo shield around the specimen region to achieve a local near ultrahigh vacuum (UHV) environment, significantly reducing contamination growth during electron beam irradiation. The passivated gold nanoparticles, with gold cores surrounded by dodecanethiol (C12H33S) ligands, were prepared using the inverse micelle method11 and deposited from solution onto a silicon (10) Hutchinson, T. O.; Liu, Y. P.; Kiely, C.; Kiely, C. J.; Brust, M. AdV. Mater. 2001, 13, 1800. (11) (a) Wilcoxon, J. P.; Provencio, P. P. J. Am. Chem. Soc. 2004 126, 6402. (b) Wilcoxon, J. P.; Williamson, R. L.; Baughman, R. J. J. Chem. Phys. 1993, 98, 9933.
10.1021/la0533157 CCC: $33.50 © 2006 American Chemical Society Published on Web 02/16/2006
2852 Langmuir, Vol. 22, No. 6, 2006
Chen et al.
Figure 2. Size (diameter) distributions and electron micrographs of passivated gold nanoparticles (a) before and after electron exposure with a dose of (b) 4 and (c) 12 mC/µm2. The histograms display the number of particles as a function of their diameters.
dioxide film on a TEM grid. The nanoparticle film consists of monolayer and multilayer regions; ordered close-packed areas were observed in the monolayer region. The gold cores have single-crystal fcc structure and an average diameter of 4.8 nm. After initial imaging, the electron beam was focused to expose a local area of about 50 nm2. Figure 1 shows bright field electron micrographs of a small region of nanoparticles before and after the 200 kV electron beam irradiation, with doses as labeled.12 Reorientation of the two particles at the center of the image was observed after an electron dose of 7.1µC/µm2, Figure 1b. At a dose of 16.4µC/µm2, these two nanoparticles still retain their original position without obvious translational movement across the support, Figure 1c. Next, in Figure 1d, a neck is formed (12) In this work, no obvious difference is observed in varying the electron beam current to achieve a given dose.
between these two particles after an exposure of 33.7 µC/µm2. Compared with the particle positions in Figure 1c, no obvious relative movement of the two mass centers is observed. Further irradiation results in the formation of a single, fully crystalline, new particle with a clear shrinkage of the agglomerate size resulting from condensation, as shown in Figure 1e. The lattice fringes from the Au (111) planes are clearly visible. Subsequently, a third particle coalesces with this new particle, and a crystalline gold island of approximately triangular shape is formed after an electron dose of about 150 µC/µm2, Figure 1f. The changes in the nanoparticle size distributions as a result of the sintering are shown in Figure 2, where the corresponding electron micrographs are also included. Evidently, the number of large particles increases as the electron dose increases. Significant broadening of the diameter distribution is observed
Sintering of PassiVated Gold Nanoparticles
Figure 3. Electron micrographs of passivated gold nanoparticles (a) before and after electron exposure with a dose of (b) 7.1, (c) 16.4, (d) 33.7, (e) 73.7, and (f) 149.8 µC/µm2, showing coalescence of two pairs of nanoparticles.
after an electron beam exposure of 4 mC/µm2. Particles with a core size larger than 10 nm begin to appear. Alongside the formation of these large islands, a new set of small (∼2 nm) particles also appears, possibly due to electron-stimulated desorption of gold atoms.13 A further phenomenon observed is the sintering together of two pairs of particles. Figure 3 shows a series of bright field TEM images which illustrate the dynamic aggregation and sintering of four particles during electron beam exposure. The sintering of two particles in the center of the images is evident in Figure 3b-d to form a single larger particle after an electron dose of 33.7 µC/µm2. Sintering between this new particle and an adjacent pair of nanoparticles is observed at a dose of 73.7 µC/µm2, Figure 3e. An almost rectangular island emerges after a dose of 149.8 µC/µm2, Figure 3f. The sintering of bare metal nanoparticles on a support, e.g., catalyst particles, is often modeled in terms of “Ostwald ripening”, in which individual metal atoms leave a metal particle, diffuse over the support, and attach to another metal particle.14,15 Large particles grow at the expense of small particles. However, when the interaction between the nanoparticles and the support is weak, the diffusion of a whole particle becomes possible. An alternative sintering mechanism thus arises in which particles diffuse across the surface and collide with other particles, leading to coalescence.16 The neck formed between adjacent nanoparticles as observed in Figure 1 reveals a sintering mechanism dominated by surface atom diffusion rather than Ostwald ripening. This surface diffusion is driven by the large surface tension resulting from the small particle size.17 While the surface diffusion mechanism dominates the mass transport process at the initial sintering stage (Figure 1d), the bulk transport begins to contribute at later times, as demonstrated by condensation of the agglomerate (Figure 1e). It is found that sintering occurs only between particles spaced apart by a narrow gap of about 1-2 nm. Even after extended electron beam exposure, the sintering remains localized in a small area and often involves only three or four adjacent nanoparticles. Sustained sintering is hindered by the increased gaps between coalescence particles resulting from sintering, confirming a limited particle mobility. The removal of the passivating ligands by electron beam irradiation will be discussed later. (13) Tanaka, M.; Takeguchi, M.; Furuya, K. Micron 2002, 33, 441. (14) Campbell, C. T.; Parker, S. C.; Starr, D. E. Science 2002, 298, 811. (15) Lo, A.; Skodje, R. T. J. Chem. Phys. 2000, 112, 1966. (16) Bartholomew, C. H. Appl. Catal. A 2001, 212, 17.
Langmuir, Vol. 22, No. 6, 2006 2853
Figure 4. Bright field TEM images of dodecanethiol-passivated gold nanoparticles with core diameter of 2.8 nm (a) before and after electron exposure with a dose of (b) 30, (c) 60, (d) 120, and (e) 240 µC/µm2. The arrow in Figure 4c indicates the formation of a neck between two adjacent nanoparticles.
Figure 5. Bright field TEM images of dodecanethiol-passivated gold nanoparticles with core diameter of 2.8 nm (a) before and after electron exposure with a dose of (b) 60 and (c) 120 µC/µm2. Arrows in Figure 5b indicate breaks in the snakelike structure as a result of the sintering between nanoparticles in the first and second layers. The coalescence between nanoparticles in the first layer itself is marked with broken arrows in Figure 5c.
To explore the effect of the nanoparticle size at the sintering process, gold nanoparticles of smaller core size, 2.8 nm, but with the same passivating ligands were examined. Figure 4 records the electron-induced sintering process in this case. The ordered nanoparticle film, as shown in Figure 4a, exhibits a first layer with closed-packed structure and a partially complete second layer in which nanoparticles sit on the bridge sites of the first layer to form snakelike structures.18 No coalescence between monolayer particles is observed after electron irradiation with a dose of 30 µC/µm2, Figure 4b. After an exposure of 60 µC/µm2, a neck is formed between two nanoparticles, as marked by an arrow in Figure 4c. This observation again reveals a surface diffusion-dominated sintering mechanism as observed in the case of the large 4.8 nm particles, Figure 1d. However, the electron dose needed to cause sintering between nanoparticles in the first (17) Lewis, L. J.; Jensen, P.; Barrat, J.-L. Phys. ReV. B 1997, 56, 2248. (18) Wellner, A.; Nellist, P. D.; Palmer, R. E.; Aindow, M.; Wilcoxon, J. P. J. Phys. D. 2000, 33, L23.
2854 Langmuir, Vol. 22, No. 6, 2006
Chen et al.
Figure 6. Monto Carlo simulations of 200 keV elastic electron beam scattering trajectory in (a) a silicon dioxide film, (b) monolayer and (c) bilayer of 2.8 nm passivated gold nanoparticles, and (d) monolayer of 4.8 nm passivated gold nanoparticles.
layer is higher (nearly double) for the 2.8 nm core particles than the 4.8 nm particles (60 vs 33.7 µC/µm2). This observation indicates a significant difference between the sintering behavior of passivated particles and bare metal nanoparticles. In the case of bare particles, the surface energy increases substantially as the radius of a metal particle decrease below ∼3 nm.14 The small particles also have a lower melting temperature (Tm) than large particles.19 The reduction of the melting point enhances surface diffusion, as Dsurf ≈ a exp(-bTm/T).20 As a result, small particles have a higher sintering rate than large particles. However, in the case of passivated nanoparticles described here, electron irradiation must first remove the organic ligands that keep the metal cores apart to allow coalescence and sintering of the metal cores themselves. We suggest that the lower electron scattering cross section of small gold particles relative to large particles, as will be discussed later, together with an increased surface molecule/volume ratio with reducing size reduces the efficiency of ligand removal. Our results thus suggest that the rate of sintering of the passivated nanoparticles under the electron beam depends principally on the stability of the organic ligands to e-beam irradiation. To compare the sintering rates of nanoparticles in a monolayer and bilayer, Figure 5 shows the electron-induced sintering in both regions. A typical bilayer region is characterized as a snakelike structure, as shown in Figure 5a. Breaks in this snakelike structure were observed in Figure 5b (marked with arrows) as a result of the sintering between nanoparticles in the first and second layers, which occurs prior to the coalescence between nanoparticles in the first layer itself (marked with broken arrows in Figure 5c). This behavior suggests that electron scattering in the nanoparticle film plays an important role in initiating (19) Castro, T.; Reifenberger, R.; Choi, E.; Andres, R. P. Phys. ReV. B 1990, 42, 8548. (20) Morgenstern, K.; Lægsgaard, E.; Stensgaard, I.; Besenbacher, F. Phys. ReV. Lett. 1999, 83, 1613.
coalescence between nanoparticles. In comparison with the SiO2 substrate, gold atoms in the nanoparticle film have a higher secondary electron yield. Thus, it appears that secondary electron generation, rather than thermal heating, is the principal driver of the removal of the organic molecules from the nanoparticles.21-24 To explore further the electron scattering process, Monte Carlo simulations25 of the electron scattering were undertaken for the various nanoparticle films employed in this work and the silicon dioxide substrate. The nanoparticle film was modeled as a continuum with the calculated average atomic number, atomic weight, and density values of the film.4,25 The number of Au atoms was calculated based on average core sizes and bulk density. The number of carbon atoms was calculated assuming a constant surface coverage of organic molecules on the surfaces of both large and small Au cores. Some of the many simulations are illustrated in Figure 6, which shows the results for a silicon dioxide film, Figure 6a, a monolayer and bilayer of passivated gold nanoparticles with core diameter of 2.8 nm, Figure 6b and c, and a monolayer of 4.8 nm passivated gold nanoparticles, Figure 6d. In the case of the monolayer of small nanoparticles (2.8 nm gold core), Figure 6b, the trajectories generally form a smooth cone that is closely confined around the incident beam axis and most electrons are only scattered at most once during their passage through the foil, which is similar to the result for the silicon dioxide film, Figure 6a. In the case of a bilayer film, (21) Heister, K.; Zhamikov, M.; Grunze, M. Langmuir 2001, 17, 8. (22) Zharnikov, M.; Frey, S.; Heister, K.; Grunze, M. Langmuir 2000, 16, 2697. (23) Chen, Y.; Schmidt, J.; Siller, L.; Barnnard, J. C.; Palmer, R. E.; Burke, T. M.; Smith, M. P.; Brown, S. J.; Ritchie, D. A.; Pepper, M. Appl. Phys. Lett. 2000, 76, 3034. (24) (a) Olsen, C.; Rowentree, P. A. J. Chem. Phys. 1998, 108, 3750. (b) Zharnikov, M.; Geyer, W.; Golzhauser, A.; Frey, S.; Grunze, M. Phys. Chem. Chem. Phys, 1999, 1, 3163. (25) Joy, D. C. Monte Carlo Modelling for Electron Microscopy and Microanalysis; Oxford University Press: Oxford, 1995.
Sintering of PassiVated Gold Nanoparticles
Figure 6c, large angle scattering is somewhat more prominent as a result of scattering at least once or twice. In the monolayer of large nanoparticles (4.8 nm gold core), Figure 6d, nearly all electrons are scattered once or twice and the beam profile becomes broader as the occasional electron is scattered through a large angle to travel almost horizontally through the film. The increased electron scattering cross section for the film of large gold nanoparticles and also the bilayer film in comparison with the small nanoparticles and the silicon dioxide film indicates an enhanced electron-material interaction and thus by implication increased secondary electron generation. This behavior accounts for the higher efficiency of ligand removal and thus the lower electron dose required to produce sintering between the nanoparticles which emerges from the experimental results. In summary, this work demonstrates that the sintering of individual passivated gold nanoparticles under electron irradia-
Langmuir, Vol. 22, No. 6, 2006 2855
tion is local and mainly depends on the sensitivity of the passivating ligands to the electron beam. It is found that a multilayer film is less stable than monolayer film, consistent with the enhanced generation of secondary electrons. A size effect on the sintering of passivated Au nanoparticles is revealed where nanoparticles of large cores sinter prior to those of small cores, significantly different from that of bare metal particles. In particular, the formation of a neck between adjacent nanoparticles indicates a mechanism driven by surface atom diffusion rather than Ostwald ripening or nanoparticle diffusion-collision at the initial sintering stage. An understanding of the sintering process for passivated nanoparticles is important for the long lifetime performance of novel devices and materials based on nanoparticles. LA0533157
