J. Phys. Chem. C 2007, 111, 13367-13371
13367
Atomic Alignment in Particle Crystals of Au Nanoparticles Grown at an Air/Water Interface Seiichi Sato, Suhua Wang, and Keisaku Kimura* Department of Material Science, Graduate School of Science, UniVersity of Hyogo, 3-2-1 Koto, Kamigori-cho, Ako-gun, Hyogo 678-1297, Japan ReceiVed: May 28, 2007; In Final Form: July 17, 2007
A quality superlattice made of hydrophilic Au nanoparticles was grown at an air/water interface over several days or a week; during this period, it is possible for a self-correcting process to take place, which revealed both narrow angle and wide angle electron diffraction, indicating that there were both translational and orientational ordering in a superlattice. The model structure of this superlattice was constructed on the basis of a truncated octahedral shape at an atomic level. The ordering misfit angle was derived from the analysis of the crescent pattern in a wide-angle diffraction region.
Introduction The precise control of the stacking of building units and the atomic arrangement in a building block is required for the nanometer-scale architecture such as quantum dot array, monolayer stacking, and nanoparticle self-assembly.1 Specifically, the atomic arrangement throughout the architecture is recommended because it is strongly related to the coherence of dynamics of the system.2 Still, it is a current research target to establish such atomically regulated structure and examine the stacking structure in atomic resolution. Diffraction data are commonly used to construct the molecular structure and superlattice architecture in atomic and molecular dimensions. We emphasize that it is essential to show two kinds of diffraction in one image field to avoid the arbitrary selection of the diffraction area at different scales of angstrom (atomic periodicity) and nanometer (superlattice periodicity) orderliness. The translational ordering of particles reflects in small angle electron diffraction (SAED), and the orientational ordering of atoms in a particles appears in wide angle electron diffraction (WAED) region. Both, together, establish a complete directional ordering of atoms in a whole superlattice extending to the micrometer scale. Figure 1 illustrates how atomic lattice alignment in a superlattice reflects to ED. Case a stands for the situation of both translational and orientational orderliness; central spots in the small angle region are arising from the translational alignment of particles, which are surrounded by peripheral spots from atomic scale regularity. Case b is the translational ordering with no atomic orientational alignment, which comprises peripheral rings from the random orientation of atoms. Note that rings or spots in the WAED region can classify these two cases. There are numerous reports that show case b type diffraction. Both small angle and wide angle diffraction data are inevitable to prove that these structures are really constructed in a superlattice. However, only arbitrary sampling data have appeared in the literature such as small angle diffraction in different field from wide-angle diffraction data without the precise description of the positioning of the electron beam. Collier et al. mentioned the basic idea of orientational alignment of the atomic plane in the metallic nanoparticle * Corresponding author. E-mail:
[email protected].
Figure 1. Schematic diagram of two possible atomic alignments in real (above) and inverse (below) space. (a) Translational and orientational alignments in a particle crystal reflect to central (SAED) and outer (WAED) diffraction spots. Central part reflects the nanoscale ordering whereas the outer part reveals the atomic arrangement in Angstrom level. (b) No orientational alignment. Translational alignment reflects to central small angle diffraction spots. Random orientational alignments reflect to peripheral ring diffractions.
crystals in 1998.3 The formation of quality crystals comprised of alkylthiolate modified Ag nanoparticles of diameter of 5 nm was reported, and they insisted that the orientational order formed in nanocrystal lattices.4 However, there was no presentation of both wide and small angle electron diffraction patterns from the same field at one time; it only showed an arbitrarily selected diffraction profile. Moreover, the diffraction from fcc Ag atomic (111) lattice showed the ring patterns but not the spot pattern. Wang also presented a detailed packing model of truncated octahedra mainly using high resolution transmission electron microscopy (HRTEM) images and discussed the orientational orderliness.5 The WAEDs cited in his paper all showed ring patterns. It may be said that the evidence of atomic alignment over a superlattice is not enough in the observation. The orientational ordering in a two-dimensional Au superlattice with a core diameter of 8.3 nm was reported without showing
10.1021/jp074126v CCC: $37.00 © 2007 American Chemical Society Published on Web 08/22/2007
13368 J. Phys. Chem. C, Vol. 111, No. 36, 2007 SAED and WAED in the same beam irradiation area.6 Stoeva et al. showed the SAED and WAED patterns not in the same image field and insisted orientational ordering in 4.7 nm Au particles prepared by the inverse micelle method.7 They were presenting a spot image in SAED with rings in WAED from Au 4.5 nm particles by the SAMD method. Hence, it is not plausible to conclude the formation of orientational alignment over whole superlattices in all cases. Here, we show a clear evidence of translational and orientational ordering, which was attained in the superlattices made of water soluble Au nanoparticles, and we present a model structure for the Au superlattice. Applying ED method for this material, we have shown clear diffraction from both small and wide angle regions in a single diffraction image. We used hydrophilic mercaptosuccinic acid (MSA)-capped Au nanoparticles as an example8,9 and reported a facile quality superlattice formation with hydrogen chloride (HCl). Chen et al.8 reported that a single water molecule is bound to a single surface MSA molecule even in its dry powdered state elucidated by elemental analysis, and they pointed out the possibility of forming nanoparticle arrays via hydrogen bonding. Because of a selfcorrecting process taking place in the aqueous solution to give a high quality superlattice, Au particle crystals were formed by direct injection of HCl into the sample solution as a precipitate agent or a pH regulator within several days.10 This method is called a self-correcting process.11 In 2003, we presented the formation of a quality particle crystal grown at an air/water interface, which shows a well-resolved OH stretching bond in the IR region suggesting the existence of a water cluster in interstitial space at room temperature.12 In this case, the regular lattice is responsible for allowing the ice cluster formation. Since an MSA-capped nanoparticle could be regarded as an artificial macromolecule, the method for the formation of colloidal crystals from bulk solution has common features with the practical procedures for protein crystallization. All of these methods are utilizing water surface as a substrate for crystallization; that is, crystal growth via stacking false free process contrast to a normal crystallization process where solid substrate is engaged. This is beneficial to obtain a stress-free superlattice because liquid has no special periodicity enabling that lattice to grow according to its own regularity.
Sato et al.
Figure 2. Optical microscopic image of quality Au-MSA particle crystals grown at an air/water interface. Clear crystal habit is seen with pyramidal, triangular, and hexagonal shape. Scale bar is 10 µm.
and the pH value of the dispersion was adjusted with 6.0 M HCl aqueous solution. Then the solution sample was filtered through a syringe-driven microfilter with a 0.22 µm pore size immediately before they were stored in a sealed glass vial of 10 mL. After 3-5 days under room temperature, the crystallization took place in a wide range of HCl concentrations (0.3 ( 0.2M) giving numerous faceted crystals with micrometer sizes (Figure 2). These gold nanoparticle crystals were transferred to silicon, glass, amorphous carbon, and NaCl substrates before analysis. Instruments. TEM images and corresponding selected area electron diffraction (SAED) were obtained at 200 kV acceleration voltages. The Au nanoparticle superstructures formed at the air-water interface were directly transferred to a carboncoated copper support grid for TEM and SAED experiments. The mean particle size of Au nanoparticles was estimated by measuring the diameters of at least 200 individual particles. Scanning electron microscopic (SEM) images were taken at 15 kV acceleration voltages. X-ray diffraction (XRD) was performed using Cu KR1 radiation (λ ) 1.540 56 Å) operated at 40 kV and 20 mA. An optical microscopic image of the particle crystals was recorded with a Keyence VK-9500 violet laser microscope. Results and Discussion
Experimental Section Large Scale Synthesis of Gold Nanoparticles Dispersible in Water. MSA-coated gold nanoparticles were prepared using a procedure basically similar to that described in previous work but largely modified for mass production.8 Under vigorous stirring and ultrasonic irradiation, 80 mL of freshly prepared 0.3 M NaBH4 aqueous solution was added to a water-methanol mixture containing 1.0 g of HAuCl4‚4H2O and 0.73 g of MSA (97%) at ice/water temperature (The molar ratio of MSA to gold was fixed to 2.0). After the reduction reaction, a flocculent precipitate was washed with a water-methanol mixture by repeating re-suspension and re-centrifugation processes, followed by dialyzing against the flow of distilled water. A powder product of 0.6 g was obtained through lyophilization and evacuating on a vacuum line. All operations were performed under the ice-freezing temperature. The mean diameter of the MSA-coated gold nanoparticles was determined to be 3.7 nm with fwhm of 0.5 nm using TEM, as shown in Figure S1 (Supporting Information). Preparation of Superlattice Assembly from Aqueous Solution. 7.2 mg of MSA-coated Au nanoparticle powder was dispersed in 4.0 mL of distilled water to form brown solutions,
Water-soluble MSA modified Au nanoparticles in the average size of 3.7 nm, see Figure S1 (Supporting Information), was prepared by a standard procedure previously reported by our group and described in Experimental Section. Superlattice formation of Au-MSA was followed by an established method by dropping hydrochloric acid directly into the sample solution. It took around 4 days to form the superlattice at an air/water interface whose size was several micrometers. Figure 2 represents the optical micrographic image of the Au quality superlattice of 10 µm in size. Several morphologies are noticeable such as the pyramidal, triangular, and hexagonal plates. Figure 3a is the magnified TEM image of one thin plate, in which sixfold symmetry is obvious as indicated by 60° arrows. Figure 3b is WAED from the sample of Figure 3a with camera length of 1 m. There are central spots in SAED as well as the sixfold crescent in the WAED region, which is successfully indexed as fcc metallic gold. Another diffraction image without the use of a beam stopper on another thin plate is shown in Figure S2 (Supporting Information). The central SAED patterns can be clearly noticed in this figure than in Figure 3b. To clarify the SAED more precisely, magnified images are presented in Figure 3c,d, in which even the 5th Bragg diffraction spots can be seen,
Atomic Alignment of Au Nanoparticles
J. Phys. Chem. C, Vol. 111, No. 36, 2007 13369
Figure 4. Model structure of gold nanoparticles; (a) regular octahedron (thin line), l × m truncated cuboctahedron (TO; bold line), and (b) full shell polyhedron (m ) (l - 1)/2). Unit polyhedron is a regular octahedron (a) with l atoms (here 8) on the edge. Length m atoms (here 2) on the six corners are truncated to give l × m TOs. In case of m < (l - 1)/2, we have six squares and eight hexagons and m g(l - 1)/2 (full shell 14 hedron), there are six squares and eight triangles. (c) Structure of 15 × 6 TO made of 1709 Au atoms.
TABLE 1: Magic Numbered Truncated Cuboctahedron Related to Au-MSA Nanoparticlesa Au core (Å)
Figure 3. Electron micrograph and diffraction from a Au-MSA quality particle crystal. (a) Magnified bright field image of a single particle crystal with lattice fringes. Angle made by arrows in top-left corner shows 60° rotation. (b) TED from whole single particle crystal with CL ) 1.0 m. Both small and wide angle diffraction spots are clearly seen. Sixfold symmetry is clear in both ED. (c) Magnified image of SAED. Figures are indexing spots. (d) Magnified and deep development SAED, indicating fifth order Bragg diffraction.
showing the completeness of the particle alignment in the superlattice. In both SAED and WAED, hexagonal symmetry is obvious and the relative diffraction position is apparent in panel b of Figure 3. On the basis of the ED pattern, we constructed a structural model of a Au nanoparticle superlattice and present it in the following section. Figure S3 (Supporting Information) is the fast Fourier transform (FFT) of a superlattice shown in Figure 3a. The location of FFT spots is the same as for WAED which also has sixfold symmetry. Hence, superlattice orientation must satisfy a simple relation to atomic alignment. Let us now construct a three-dimensional (3D) stacking model of the Au nanoparticle superlattices based on the TEM and ED data. If we use an empirical relation n ) kD3 as the number of gold atoms, n is 1690 for the diameter D ) 3.75 nm, in which k is fitted to the mass spectrum data of gold clusters by Whetten et al.13 and our’s14 and is determined to be 32, which is close to the bulk value 31. If we accept the truncated octahedron (TO) morphology4,13 as the shape of the Au cores (Figure 4a) rather than a full shell structure (Figure 4b), the component Au nanocrystals of average core diameter 3.65 nm consist of 1709 atoms. Table 1 shows some candidates of TOs in which the regular octahedron with l atoms on the edge is truncated at six corners by m atoms (Figure 4a,c). All of the TOs listed are a magic number family. Note that full shell structures, Figure 4b, such as Au561, Au923, and Au1415, are not close to the spacefilling shapes since these TOs do not have regular hexagonal faces because of large {100}A faces. It should be noted that the ratio of x to y is close to that of the regular hexagon, 0.866 for the species marked by a circle in Table 1. That is, closest packing is possible for this case, which guarantees the equispacing filling. The 60° rotation is possible for the case of x ∼ w, which is marked by a triangle in the table. Once we set the size of the cuboctahedron such as 15 × 6 and set the orientation of the core, the possible superlattice structure is uniquely
# of atom
l
m
y
x
w
n
14 14 14 15 15 15 15 16 16 16
3 4 5 4 5 6 7 6 7 8
37.44 37.44 37.44 40.32 40.32 40.32 40.32 43.20 43.20 43.20
40.80 36.72 32.64 40.80 36.72 32.64 28.56 36.72 32.64 28.56
30.68 30.68 30.68 33.04 33.04 33.04 33.04 35.4 35.4 35.4
1750 1654 1504 2075 1925 1709 1415 2190 1896 1512
O O4 O4
a For the definition of l and m, see Figure 4a, and n is the number of Au atoms in a cuboctahedron. Bolded figures fit well to the experimental data. As for the definition of x, y, and w, see Figure 5b. Note that the ratio of x to y is close to that of regular hexagon, 0.866 for the cuboctahedron with a circled mark, and the difference of x from w is less than 5% for triangle, enabling rotational misfit for TOs.
determined. It should be noted that there is no fitting parameter to construct a superlattice structure once one set the shape and size of the l × m cuboctahedron and requested translational and orientational symmetry. It should also be taken into account that the chain length of MSA is so short as much as 0.7 nm so that packing of particles is primarily determined by the Au core shape. Figure 5a,b (monolayer stacking projection) and c (3D stacking model) are the packing model we propose here. Each {111}A surface is faced with a {111}A surface of a neighboring nanoparticle, and {100}A is faced with {100}A in the same way. As shown in the model, the sixfold symmetric structure appears in superlattice packing with the equal intersurface distance. We use 1.4 nm in this case. The distance, 1.4 nm is corresponding to twice the length of an MSA molecule (∼0.7 nm). This is consistent with the fact that there is almost no room for interdigitation for both {111}A and {100}A surfaces because of rigidity and short length ligand of MSA molecules with hydrogen-bonding networks. This is a severe constraint on the selection of magic numbered TOs. Because of geometric limitation, the possible stacking sequence of this monolayer is restricted to ABABAB... (Figure 5c) which corresponds to hcp packing. This agrees with our experimental results reported previously: The MSA modified Au nanocrystals are assembled into hcp arrangements, and the surface particle distance is 1.4 nm.8 The ED pattern obtained from this model is shown in Figure 5d, which is not consistent with the observations in Figure
13370 J. Phys. Chem. C, Vol. 111, No. 36, 2007
Figure 5. (a) Schematic representation of an Au particle crystal composed of 15 × 6 truncated octahedron (TO). The yellow and orange areas are {111}A and {100}A faces, respectively. (b) A hexagonally close-packed monolayer of the A-oriented nanocrystals. The center distance, a ) b ) 3.3 + 1.4 + 0.3 nm ) 5.0 nm. (c) Stacking of the close-packed monolayers. (d) Expected WAED and SAED pattern. Letter A stands for atomic alignment, and SL stands for superlattice. (e) Expected ED pattern with orientational misfit. (f) A side view of three-dimensional stacking. In all model structures, small dots represent gold atoms.
3b showing hexagonal symmetry. If there is a mismatch, such as mixing with 60° rotation configuration and allowing a misfit with small rotational angle among nearest neighbor particles as stated in a later section, the ED spot may be diffused. By changing the WAED spots to crescents and by mixing 60° rotation in the stacking since the 15 × 6 TO is almost a regular hexagon in this projection (see caption of Table 1), we find the pattern appears as in Figure 5e. Note the mutual direction of spots in WAED and SAED. This resembles the experimental patterns shown in Figure 3b-d. One may argue that the component nanocrystals are also likely A oriented since they are packed into sixfold symmetric arrangements. However, this model contradicts our experimental WAED results: the WAED patterns show clear diffraction from {111}A and {200}A planes, which are not observed from A-oriented crystals. Furthermore, the stacking of the A-oriented TO results in an fcc packing (the stacking of ABCABC...), which is also inconsistent with our experimental results. Since TOs having regular hexagonal faces and regular square faces tessellate 3D space, many are good components to build up stable 3D superlattices. However, there are a limited number of magic-numbered Au nanocrystals whose shapes are close to the space-filling shapes; for example, Au201, Au586, Au1289, and so forth are very close to the space-
Sato et al. filling TOs. In addition, not everything is able to construct a superlattice because of the constraint described above. All of the observation data, such as the number of gold atoms estimated at 1690 (1709 from the model), the center distance of particles of 5.1 nm (5.0 nm from the model), and the hcp stacking structure, are consistent with the model structure proposed (the value in the parentheses) in Figure 5b constructed solely by ED data. Figure 5f stands for a side view of the present model. The WAED structure shows that nanoparticles are in preferential orientation with clear superlattice arrangement. Therefore, it is possible to analyze quantitatively the degree of misfit in the atomic level in the superlattice formation based on crescent shape. The misfit of a particle level was already discussed in literature.7,15 If the shape of the spot is a circle, there is no orientational misfit among neighboring particles. The presence of the crescent structure suggests a slight misfit other than exact 60° rotation among neighboring particles. This is the first trial to analyze the crescent structure in WAED of nanoparticle superlattices. We noticed, at a glance in Figure 3b, there are six crescents but not spots or streaks in the atomic diffraction position. This indicates that the lattice constant does not change (if it does, this could cause diffusive streak structure in the radial direction of ED), but the orientation of lattice plane varies from particle to particle. Orientational misfit at the contact part of nanoparticles induces arc diffraction spots if the shape of the beam is a point. The following analysis is based on a one-dimensional random walk process. Set δθ as the variation of orientation of atomic lattice between neighboring particles and N as the number of particles in one superlattice at the given direction. Here, N can be estimated from the sizes of superlattice and particles. The probability of finding the orientation of lattice plane at angle θ is assumed to obey the Gaussian distribution as follows,
p(θ) )
1
x2πσθ
exp[-θ2/2σθ2]
σθ2 ) N〈δθ2〉
(1)
in which 〈δθ2〉 stands for the mean square of δθ. For the gold superlattice, N is around 1000. This relation gives rise to the arc shape. The true shape of the diffraction from a crystal is not a point but a full circle because of several broadening processes (an electron bean is not a point) such as homogeneous defects distributed in the crystals, finite size effect of the crystal, and development process of the film. When the homogeneous broadening effect is incorporated, eq 1 should be modified at the diffraction point (r0, θ0) as
p(r, θ) ) 1 exp[-(r - r0)2/2σ2r - (θ - θ0)2/2σ2θ) (2) 2πσrσθ Equation 2 gives rise to the crescent shape in the ED structure. We can separate homogeneous contributions from the inhomogeneous broadening induced by the misfit alignment of the superlattice. Figure 6a illustrates the case of homogeneous broadening, and Figure 6b illustrates both homogeneous and inhomogeneous cases. Since the homogeneous effect is the same for the r direction as it is for the θ direction, we can determine σr as illustrated in Figure 6c. Typical example of the analysis is depicted in Figure 7 taken for the case of the 12 o’clock direction. The results are σθ ) 14.2 and σr ) 2.77. For other directions (we exclude 2 and 6 o’clock directions on account
Atomic Alignment of Au Nanoparticles
J. Phys. Chem. C, Vol. 111, No. 36, 2007 13371 step misfit 〈δθ2〉0.5 equals 0.45°. This value may largely decrease if we select much smaller crystal or apply the vapor diffusion method to prepare quality crystals.16 It will be a future work to discuss the quality of superlattices more precisely based on the orientational disorder. Conclusion
Figure 6. Crescent pattern analysis. (a) Homogeneous broadening case (no orientational disorder). The shape of the diffraction spot is a circle. (b) Inhomogeneous broadening (orientational disorder) results in crescent structure. (c) Separation of homogeneous and inhomogeneous broadening. Orientational disorder appears on θ direction and homogeneous broadening effect is apparent on azimuthal direction.
We have constructed a model structure of Au-MSA superlattice in the atomic resolution based on WAED and SAED analyses with the size of a particle, 3.7 nm. It was found that the hcp structure observed is solely based on the shape of a truncated cuboctahedron gold particle ligated with a short chain length molecule, MSA. Crescent diffraction spot analysis shows that there is 0.5° rotational misfit among neighboring particles. Acknowledgment. This work is supported in part by a Grant-in-Aid for Scientific Research (S: 16101003) and Scientific Research in Priority Areas, Molecular Spins (15087210) from MEXT. Supporting Information Available: The TEM image and size histogram of gold-MSA nanoparticle, ED of superlattice, and FFT of superlattice of Figure 3a. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes
Figure 7. Example of the crescent pattern analysis of Au-MSA superlattice. (a) Whole ED pattern (Figure 3b) is cut into five parts (12, 2, 4, 8, and 10 o’clock directions) and (b) 12 o’clock direction is cut both for θ direction and azimuthal direction (boxed region in the figure) as an example. (c) θ direction is expanded with angle markers in both ends. (d) Digital mapping of the concentration of a crescent. Least-squares fitting of the plot (d) by Gaussian distribution function (solid line) along with a tangential direction of the crescent, giving σθ. (e) Fitting along with radial direction of the arc (see boxed position in Figure 7b), giving homogeneous broadening contribution to a crescent. This corresponds to σr term in eq 2.
of scratch or shadow noise on the negafilm) of σθ and σr, we have 12.9, 2.81; 11.4, 3.3; and 18.8, 2.87 for 4, 8, and 10 o’clock directions, respectively. The average value is 14.3 for σθ and 2.94 for σr. Since σθ is correlated to 〈δθ2〉, the extent of one-
(1) (a) Shevchenko, E. V.; Talapin, D. V.; Kotov, N. A.; O’Brien, S.; Murray, C. B. Nature 2006, 439, 55-59. (b) Shon, B-H.; Choi, J-M.; Yoo, S. II; Yun, S-H.; Zin, W-C.; Jung, J. C.; Kanehara, M.; Hirata, T.; Teranishi, T. J. Amer. Chem. Soc. 2003, 125, 6368-6369. (c) Murray, C. B.; Kagan, C. R.; Bawendi, M.G. Science, 1995, 270, 1335-1338. (2) Oonishi, T.; Sato, S.; Yao, H.; Kimura, K. J. Appl. Phys. 2007, 101, 114314-1-114314-5. (3) Collier, C. P.; Vossmeyer, T.; Heath, J. R. Annu. ReV. Phys. Chem. 1998, 49, 371-404. (4) Harfenist, S. A.; Wang, Z. L.; Alvarez, M. M.; Vezmar, I.; Whetten, R. L. J. Phys. Chem. 1996, 100, 13904-13910. (5) Wang, Z. L. AdV. Mater. 1998, 10, 13-30. (6) Brown, L. O.; Hutchison, J. E. J. Phys. Chem. B 2001, 105, 89118916. (7) Stoeva, S. I.; Prasad, B. L. V.; Uma, S.; Stoimenov, P. K.; Zaikovski, V.; Sorensen, C. M.; Klabunde, K. J. J. Phys. Chem. B 2003, 107, 7441-7448. (8) Chen, S. H.; Kimura, K. Langmuir, 1999, 15, 1075-1082. (9) Chen, S. H.; Kimura, K. Chem. Lett. 1999, 28, 1169-1170. (10) (a) Kimura, K.; Sato, S.; Yao, H. Chem. Lett. 2001, 30, 372-373. (b) Wang, S. H.; Sato, S.; Kimura, K. Chem. Mater. 2003, 15, 2445-2448. (11) Sato, S.; Yao, H.; Kimura, K. Physica E 2003, 17, 521-522. (12) Wang, S. H.; Yao, H.; Sato, S.; Kimura, K. J. Am. Chem. Soc. 2004, 126, 7438-7439. (13) Whetten, R. L.; Khoury, J. T.; Alvarez, M. M.; Murthy, S.; Vezmar, I.; Wang, Z. L.; Stephens, P. W.; Cleveland, C. L.; Luedtke, W. D.; Landman, U. AdV. Mater. 1996, 8, 428-433. (14) (a) Negishi, Y.; Takasugi, Y.; Sato, S.; Yao, H.; Kimura, K.; Tsukuda, T. J. Amer. Chem. Soc. 2004, 126, 6518-6519. (b) Negishi, Y.; Nobusada, K.; Tsukuda, T. J. Amer. Chem. Soc. 2005, 127, 5261-5270. (15) Yao, H.; Minami, T.; Hori, A.; Koma, M.; Kimura, K. J. Phys. Chem. B 2006, 110, 14040-14045. (16) Yang, Y.; Liu, S.; Kimura, K. Angew. Chem., Int. Ed. 2006, 45, 5662-5665. (17) Yang, Y.; Kimura, K. J. Phys. Chem. B 2006, 110, 24442-24449.
