Highly Ordered Superlattices from Polydisperse Ag Nanoparticles: A

Takuya Harada and T. Alan Hatton. Langmuir ... Seiichi Sato, Suhua Wang, and Keisaku Kimura. The Journal of ... Keisaku Kimura , Thalappil Pradeep. Ph...
1 downloads 0 Views 791KB Size
Download PDF
24442

J. Phys. Chem. B 2006, 110, 24442-24449

Highly Ordered Superlattices from Polydisperse Ag Nanoparticles: A Comparative Study of Fractionation and Self-Correction Yang Yang 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: July 30, 2006; In Final Form: October 3, 2006

We have examined two different routes to construct highly ordered two- or three-dimensional (2D or 3D) superlattice structures from hydrophilic polydisperse mercaptosuccinic acid (MSA)-modified Ag nanoparticles of the average size of 2.5 nm. First, polydisperse particles were fractionized by the polyacrylamide gel electrophoresis (PAGE) method. Due to the size-dependent migration under the electric field, the particles were isolated into a series of gel bands and each band contained particles with significantly narrow size distribution. Subsequent to phase transfer into chloroform by cationic surfactant, long-range 2D superlattices were simultaneously formed on the substrate upon evaporation of chloroform. Second, 3D superlattices were directly grown at an air-water interface from the polydisperse bulk dispersion by diffusion of HCl vapor without any pretreatment for the size narrowing. The influence of diffusion rate of HCl was also studied. The achievement of 3D superlattices via this route was ascribed as a long-time self-correction process. Furthermore, it was revealed that the superlattice structures obtained by the above two procedures exhibited distinct features though the starting material was the same MSA-Ag nanoparticles. The surface distance of core between component particles, the orientation of particles inside the superlattice, and the process of superlattice formation were comprehensively studied. We confirmed that each growth process depended on a corresponding selfassembly mechanism.

Introduction Organization of nanoparticles into ordered macroscopic architectures via a route called “bottom up” assembly has attracted tremendous interest from the science and technology community.1-16 Since metal nanoparticles exhibit unique size-, shape-, arrangement-induced surface plasmon resonance (SPR), the achievement of ordered superstructures would provide an opportunity to investigate and further tailor their collective physical properties, which is expected to have potential applications in future optical and electrical devices. Therefore, over the last several years, self-assembly of metal nanoparticles such as gold, silver, and copper into two- or three-dimensional (2D or 3D) ordered arrays has been a topic of considerably increasing interest in the field of nanotechnology.10-16 Among the plentiful strategies for the preparation of nanoparticle superlattices, the method of solvent evaporation from nanoorganosols has been widely applied to obtain 2D or 3D superlattice structure on a substrate.17-21 Recently, crystallization of nanoparticles from their bulk solution by sedimentation or precipitation has also demonstrated its preponderant advantages in superlattice formation.22,23 In most of the cases, the achievement of quality superlattice structures strongly depends on how to attain highly monodisperse nanoparticles as building blocks. For metal nanoorganosols, narrowing size distribution has been accomplished by size-selective precipitation, where two miscible solvents with different solvability for surfactant alkyl chains are employed.24,25 Through a successive addition of poor solvent, gradual size-dependent precipitation of metal nanoparticles can be obtained. The high monodispersity of each fraction will lead * Address correspondence to this author. E-mail: [email protected].

to the spontaneous 2D or 3D superlattice formation via the subsequent assembly process. Relatively, metal nanohydrosols stabilized by thioacid derivatives are a family of nanoparticles with good water-solubility owing to the abundant carboxylate groups anchored on the particle surface. Compared to their various preparative methods and plentiful species, few effective size-selective routes are available for further narrowing their size distribution until now, which obviously restricts the constitutionofsuperlatticefromthistypeofmetalnanoparticles.26-28 Girault and co-workers once suggested an interesting method of isoelectric focusing electrophoresis (IEF) for the size-selective separation of water-soluble gold nanoparticles capped by thioacid derivative, but the fraction composed of smaller nanoparticles still showed a broad size distribution such as 1.7 ( 0.4 nm.26 Recently, Hutchison and co-workers reported the rapid size separation of thioacid-derivative-capped gold nanoparticles by diafiltration.27 It was found that different-sized nanoparticles could be fractionized to a certain extent. However, the particle within each fraction still presented polydisperse size distribution. Thus, how to implement superlattice formation from polydisperse hydrophilic metal nanoparticles stabilized by molecules bearing carboxylate groups is still a very important subject requiring further exploration. In this report, we introduce two different approaches to prepare highly ordered superlattice structure from a hydrophilic and polydisperse system of mercaptosuccinic acid-modified silver (MSA-Ag) nanoparticles with an average size of less than 3 nm. First, we adopt a polyacrylamide gel electrophoresis (PAGE) method to size-selectively fractionize polydisperse Ag nanoparticles. The MSA-Ag nanoparticles in one dense fraction are extracted from the gel into the organic phase by virtue of cationic surfactant and further assembled into the 2D superlattice

10.1021/jp064876p CCC: $33.50 © 2006 American Chemical Society Published on Web 11/14/2006

Construction of 2D or 3D Superlattice Structures structure by the solvent evaporation method. Second, we directly grow 3D superlattice at an air-water interface from the bulk solution containing polydisperse MSA-Ag nanoparticles via a vapor diffusion route that we have recently developed.29 The above two methods are summarized as fractionation and selfcorrection routes on the basis of their distinct features. We will show that both routes are efficient in creating highly ordered 2D or 3D superlattice from this polydisperse small-sized nanoparticle system. In case where water-soluble nanoparticles are used as the starting materials, highly ordered superlattices of nanoparticles in a core size less than 3 nm are still rare to the best of our knowledge.30 As a final point, we precisely compare the superlattice structure formed through these two different self-assembly routes and briefly discuss the corresponding mechanisms. Experimental Section Materials. All chemicals used in this experiment were purchased from Wako Pure Chemicals without further purification. Distilled water was produced by an Advantec GS-200 automatic water-distillation supplier. Water-soluble MSA-Ag nanoparticles with an average diameter of ca. 2.5 nm were prepared by a standard procedure previously described by our group.31,32 Fractionation by the PAGE Method. A gel electrophoresis unit with a gel 2 mm thick (JAPAN ATTO COPORATION, AE-6200) was used to process the PAGE method. The total contents of acrylamide monomers (acrylamide:bis(acrylamide) ) 93:7) were 3% for condensation gel and 23% for separation gel. The eluting buffer was prepared by dissolving 192 mM glycine and 25 mM tris(hydroxymethyl)aminomethane into 1000 mL of distilled water. For preparation of the sample solution 20 mg of crude polydisperse MSA-Ag nanoparticles was dissolved in the mixture of 950 µL of 30% sugar (sucrose) solution and 50 µL of 1 M tris(hydroxymethyl)aminomethaneH3PO4 solution (pH 6.8). Then the sample solution was loaded onto the 2-mm gel and eluted for 6 h at a constant voltage mode (150 V) to achieve separation. The gel containing fraction 4 was cut out, ground, and dipped in iced distilled water (2 mL) for 30 min. Subsequently, the mixture was centrifuged at 12 000 rpm for 10 min, followed by filtering with 0.22-µm pores to remove the gel lumps suspended in the solution. The resultant solution including watersoluble MSA-Ag nanoparticles in fraction 4 was mixed with 1.0 mL of cetyltrimethylammonium bromide (CTAB) chloroform solution (1.0 × 10-5 M). By moderately shaking this hydrosol-chloroform biphasic solution, the deep brown color of the aqueous solution was quickly transferred to the chloroform phase. The solution was left quiet until two clearly separated layers appeared. The water phase was then removed. For achieving the self-assembly of Ag nanoparticles from the obtained organosols, 2 droplets (∼0.05 mL) of this chloroform solution were dropped on to a carbon-supported copper grid and dried under ambient conditions for the following transmission electron microscopy (TEM) observation. Self-Correction by the Vapor Diffusion Method. The crystallization of MSA-Ag nanoparticles from the bulk solution was carried out via a vapor diffusion approach.29 In glass vial A 5 cm in height and 1.5 cm in diameter, 6.0 mg of the asprepared MSA-Ag nanoparticles powder was dissolved in 3.0 mL of distilled water. In the same-dimension glass vial B, 3.0 mL of 3.0 M HCl solution was placed. Both vials A and B were stored in one closed vessel with suitable volume (50 mL) and left in the dark for 4 months. Due to the chemical potential

J. Phys. Chem. B, Vol. 110, No. 48, 2006 24443 difference of HCl between vials A and B, a portion of HCl slowly divorced from vial B as the vapor state and dissolved into vial A, which resulted in the gradual decrease of pH in the solution of MSA-Ag nanoparticles. During this period slow diffusion of water through the vapor phase also happened from vial A to B. The visible film formed at an air-water interface in vial A was scooped on a carbon-supported copper grid and dried in vacuum prior to TEM observation. For investigating the influence of HCl concentration on superlattice formation, we also employed 3.0 mL of 6.0 M HCl solution in vial B for comparison. Instruments. TEM images and corresponding selected area electron diffraction (SAED) were obtained on a Hitachi H-8100 microscope operated at 200 kV accelerating voltage. X-ray diffraction (XRD) was performed on a Rigaku RINT/DMAX2000 diffractometer, using Cu KR1 radiation (λ ) 1.54056 Å) operated at 40 kV and 20 mA. The scan speed and scan step were set at 2 deg/min and 0.01°, respectively. MSA-Ag nanoparticle powder (10 mg) was spread in a standard cavity mount. Fourier transform infrared spectroscopy (FT-IR) was recorded on a HORIBA FT-210 spectrophotometer by using a KBr disk (150 mg) dispersed with the MSA-Ag nanoparticles at a weight ratio of 0.5%. A reference “blank” KBr pellet was used for the background correction. Results and Discussion As-Prepared Polydisperse MSA-Ag Nanoparticles. Figure 1a presents a representative TEM image of the as-prepared MSA-Ag nanoparticles, in which a large number of nanoparticles in roughly spherical shape are found. On the basis of analysis of the particle-size-distribution histogram shown in Figure 1b, it is determined that the particles have an average diameter of 2.5 ( 0.98 nm with a standard size deviation of nearly 40%. Compared to the larger MSA-Ag nanoparticles in the average size of 4.8 ( 1.4 nm (size deviation is 29%) we studied previously,29 the current particle system is more polydisperse since similar derivations in core size will broaden the size distribution for small particles more significantly. Therefore, the orderly arrangement of small particles is easier to disrupte by size polydispersity and in general small particles do not selfassemble well relative to larger ones. The XRD pattern in Figure 1c indicates that the MSA-Ag nanoparticles hold a facecentered-cubic (fcc) phase. The average size estimated from XRD peaks by the Scherer equation is about 2.6 nm, almost consistent with the TEM result. The surface feature of the prepared Ag nanoparticles was characterized by FT-IR spectroscopy (Figure 1d). The strong peaks at ∼1575 and ∼1400 cm-1 are attributed to the asymmetric and symmetric stretching vibration of carboxylate ions (COO-), respectively. Besides, the ν(COOH) at ∼1679 cm-1 once existing in pure MSA molecules could not be detected.31 The above results demonstrate that a layer of MSA molecules in the form of carboxylate groups are anchored on the particle surface. The presence of a negatively charged carboxylate layer not only leads to the good water-solubility of MSA-Ag nanoparticles, but also stabilizes the dispersion state owing to the Coulomb repulsion force among particles. This is also a common feature of all the thioacidderivative-capped nanoparticles. For achieving the long-range ordered 2D or 3D self-assembly from MSA-Ag nanoparticles of the size 2.5 ( 0.98 nm, two fundamental factors should be taken into consideration. First, fabrication of well-defined superlattice structure strongly depends on the monodispersity of particles as the building blocks, so the size distribution of this nanoparticle system must be

24444 J. Phys. Chem. B, Vol. 110, No. 48, 2006

Yang and Kimura

Figure 1. Characterizations of as-prepared MSA-Ag nanoparticles: (a) TEM image; (b) histogram of particle size distribution; (c) XRD pattern; and (d) FT-IR spectroscopy.

drastically narrowed. Next, self-assembly of small nanoparticles would be mostly driven by interaction at molecular length scales such as surfactant chain interdigitation,30 hydrogen bonding,33 π-π interaction,34 and precisely controlled electrostatic force.35 If all the particles are stabilized by homogeneously negative charge (or positive) like as-prepared MSA-Ag nanoparticles, the inaccessible balance of the force field will force them to stick to each other instead of forming an ordered assembly after they are separated from the solvent.36 Therefore, effective neutralization or reduction of the surface charge of the MSAAg nanoparticle is also very important for the long-range organization. This goal can be achieved through surfactantinduced anion-cation pair formation on the particle surface or conversion of MSA molecules from the carboxylate to the acid state by acidification. Superlattice Formation via Fractionation by the PAGE Method. PAGE is a widely used biochemical method for protein separation. Recently, utilization of this method for the isolation of subnanometer-sized gold nanoclusters capped by glutathione monolayers has also been reported.37,38 Since the MSA-modified nanoparticle often behaves like a negatively charged biomacromolecule,39 we further extended the PAGE method to the sizeselective fractionation of the MSA-metal nanoparticle system. Figure 2a exhibits an instant photo of the gel when the electrophoresis process lasted for 6 h for the current polydisperse MSA-Ag nanoparticles. Interestingly, well-defined bands (partially indexed as 1-7) can be clearly visualized, indicating that PAGE is also able to isolate polydisperse MSA-Ag nanoparticles into a series of distinct fractions. TEM investigation reveals that the principal part of MSA-Ag nanoparticles located in each fraction hold a quite uniform size compared to the starting dispersion and generally the fractionized size increases slowly

Figure 2. (a) PAGE result of polydisperse MSA-Ag nanoparticles eluted for 6 h at a constant 150 V voltage. (b) TEM image of MSAAg nanoparticles in band 4 after phase-transfer into chloroform by cationic surfactant CTAB. The inset shows a magnified area.

from the bottom to the top of the gel, i.e., from band 1 to 7. However, a small number of large nanoparticles whose size is unexpected under the circumstances could also be detected in most of the bands, indicating the existence of bimodal size distribution in each fraction. In the current research, we only concentrate on the MSA-Ag nanoparticles distributed in band 4. Systematic analysis of population, thermodynamic stability, and optical properties of MSA-Ag nanoparticles in each band fractionized by the PAGE method will be discussed elsewhere. Two reasons attract us to select particles in band 4 as the objects for further superlattice fabrication. First, from the intensities of the bands shown in Figure 2a, it can be found that the particle concentration in band 4 is relatively high. Second, the particle size in this band is ∼2.5 nm by TEM observation, agreeing with the average size of the as-prepared MSA-Ag nanoparticles.

Construction of 2D or 3D Superlattice Structures

J. Phys. Chem. B, Vol. 110, No. 48, 2006 24445

Figure 3. Highly ordered 2D superlattice structure obtained from the area consisting of 2.5-nm CTA-MSA bilayer-capped Ag nanoparticles after chloroform evaporation (band 4): (a) TEM image of the sample in a large area; (b) the magnified image and (c) size distribution histogram of (a); (d) SAED and (e) FFT pattern obtained from (b).

Aqueous MSA-Ag nanoparticles extracted from band 4 were transferred into the chloroform phase under the assistance of cation surfactant CTAB through electrostatic interaction. The formation of MSA--CTA+ anion-cation pairs not only alters the hydrophilic particle surface to the hydrophobic, but also effectively neutralizes the charge coming from carboxylate groups.40 The large scale and enlarged images of CTA-MSAbilayer-capped Ag nanoparticles deposited on the TEM grid are shown in Figure 2b and its inset, respectively, in which two distinct types of particles could be found. Dense and uniform particles with a size of 2.5 nm make up the background while some large particles (6-8 nm) are dispersed in the interior. It must be pointed out that even though the distribution of large nanoparticles shown in this image is uppermost in all the areas observed, their proportion obtained from this photo is less than 5%. Therefore, we can conclude that the majority in band 4 are monodisperse 2.5 nm MSA-Ag nanoparticles. It is known that the PAGE separation process depends on the molecular size as well as the charge number. The MSA molecule is a dicarboxylic acid with two equilibrium constants K1 ) 6.45 × 10-5 and K2 ) 2.29 × 10-6. When the nanoparticle solution is introduced on the gel at pH 8.8, the dissociation degree of MSA anchored on particle surfaces is near 100% calculated by the equation of dissociation equilibrium. Thus, the particle surfaces can be assumed to be fully negatively charged. Moreover, the charge number is size dependent allowing for a 0.156 nm2 constant occupation area per MSA molecule on the particle surface.39 Under a constant electric field all the particles can gradually migrate from the cathode to the anode and those with a smaller size are supposed to have a

higher migration rate owing to the larger specific charge, which allows an effective separation of the particles depending on their size. Such a size-related migration trend has also been confirmed by the previous report on isolation of magic-numbered Aun clusters via the PAGE approach.38 However, there is no obvious reason why a small portion of larger particles coexist in band 4 and a similar situation takes place in most bands only allowing for the PAGE method. Recently, we found that the larger particles observed in each band probably formed during the phase transfer process. Further research is required to clarify this phenomenon. Focusing on the area made of only 2.5-nm MSA-Ag particles (Figure 3a), we can find that the particles spontaneously organize into a 2D compact network over a long distance after chloroform is vaporized. Panels b and c of Figure 3 show its magnified image and corresponding histogram of particle size distribution, respectively. Obviously, formation of such a highly ordered 2D superlattice is accompanied by the greatly narrowed size polydispersity (reduced to 12%) of MSA-Ag nanoparticles derived from the PAGE fractionation. The SAED pattern (Figure 3d) exhibits the typical diffraction rings indexed to (111), (200), (220), and (311) planes, indicative of the fcc atom orientation of Ag nanoparticles existing in the superlattice structure. The fast Fourier transform (FFT) of Figure 3b shown in Figure 3e further proves the long-range hexagonal order of the formed 2D structure. It is well accepted that the colloidal pair interactions like chain interdigitation for the formation of superlattice are mainly isotropic and centrosymmetric, just as the atomistic Lennard-Jones potential, which imposes the nanoparticles to arrange in a hexagonal monolayer with a close-packed structure

24446 J. Phys. Chem. B, Vol. 110, No. 48, 2006 and further grow into a fcc or a hcp 3D network. However, we failed to find any ordered 3D superlattice structure via this route even when the initial concentration of Ag nanoorganosols was greatly amplified. Superlattice Formation via Self-Correction by the Vapor Diffusion Method. The growth of colloidal crystals of MSAcapped nanoparticles such as Au has been achieved by direct addition of a concentrated or 6 M HCl solution into their bulk aqueous solution.41,42 Experientially, the superlattice is likely to appear in 3-5 days when the concentration of HCl is 0.20.5 M in the bulk solution of MSA-Au nanoparticles (2.0 mg/ mL). Full conversion of MSA molecules on the particle surface from the charged anion to the acid state can transform MSAcapped particles from the charged state to the neutral state, which makes the particles divorced from the bulk solution and assembly at the air-water interface. The decrease of the repulsive forces as well as the present hydrogen bonding among the particles is responsible for the superlattice formation. However, formation of colloidal crystals by this method strongly depends on the monodispersity of starting nanoparticles and its formation conditions are quite sensitive to the surrounding atmosphere with many parameters. Recently, we further improved this approach by gradual diffusion of HCl vapor through the gas phase into the bulk solution of MSA-Ag nanoparticles with a size of 4.8 ( 1.4 nm.29 It was interesting to find that highly ordered 2D or 3D superlattices composed of 4.8-nm particles could also be formed at the air-water interface for this system with an initial size distribution up to 30%. We have ascribed its possible mechanism as the self-correction process favored by the air-water interface during the time-related equilibrium growth. For the current MSA-Ag system with smaller size and broader size distribution (2.5 ( 0.98 nm, 40% polydispersity), a similar vapor diffusion strategy was applied for superlattice creation. After sealing the crystallization vessel with two vials, A, the bulk solution, and B, the 3.0 M HCl reservoir solvent, for 4 months, the dark-brown color of the starting MSA-Ag nanoparticle solution obviously faded. Besides, assembled structures visible to the naked eye were found at the air-water interface of vial A, whereas some black precipitations were also formed at its bottom. Figure 4a is the TEM image of the assembly formed at the air-water interface, in which large domains of highly ordered arrays are observed. The obvious difference in the contrast indicates the particle arrangement in a 3D structure. The inset of Figure 4a presents a magnified image of one array boundary. It can be seen that the close-packed structure is stacked by component nanoparticles in a uniform size of 2.5 nm. Compared to the starting materials with a high polydispersity, the size distribution of MSA-Ag nanoparticles is surprisingly narrowed inside the superlattice via this route though no fractionation process is performed. Because formation of a highly ordered particle array is often beneficial from the interface such as air-water or oil-water during a highly dynamic process,43,44 we proposed the air-water interface in this case not only offers an alternative scaffold for the organization of MSA-Ag nanoparticles, but retains sufficient reversibility to correct the errors during the assembly, which we considered a self-correction process. With the pH value in vial A decreased, MSA-Ag nanoparticles gradually assemble on the suspension surface. However, the aggregates formed at this initial stage are random owing to the size polydispersity. Gradual vapor diffusion of HCl into the bulk solution can produce aggregates at a very mild rate, which provides sufficient opportunities for

Yang and Kimura

Figure 4. (a) TEM image of a superlattice structure grown at the airwater interface through the gradual diffusion of HCl vapor into the polydisperse bulk solution of MSA-Ag nanoparticles after 4 months. The inset shows a magnified image of one array boundary. (b, c) Representative images of assembled structures exhibiting the formation process of 3D superlattices.

particles to reassemble and order themselves into a crystalline lattice relying on their diffusive nature in a suspension (inadequate solvation). Along with the growth process of colloidal crystals, a small portion of the particle ensemble in unmatched size will be gradually repelled toward the boundary of the superlattice and separated. Moreover, even the particle in the suitable size will also be released back into bulk solution if it is trapped inside a metastable site. This particle can re-enter and remain in the lattice when the most stable position is encountered. Direct inspections to testify to this phenomenon are provided in Figure 4, panels b and c. The areas denoted by arrow 1 in both images are the highly ordered multilayers of tightly packed 2.5-nm particles, while the particles around the border of the superlattice hold a large size of 4-6 nm (denoted by arrow 2 in Figure 4c). Alternatively, some bigger nanoparticles constitute the superlattice structure in very small domains by themselves (denoted by arrow 2 in Figure 4b). As a result, most 3D superlattices found at the air-water interface via this self-correction route are built by 2.5-nm MSA-Ag particles, the preponderant component of the starting material. Details on the morphology and structure of the 3D superlattices formed are shown in Figure 5, which will be discussed in the next section. To investigate the influence of the HCl diffusion rate on superlattice formation, we increased the initial concentration of HCl in vial B from 3.0 to 6.0 M. The time was obviously shortened for the appearance of assemblies at the air-water interface, indicating a faster rate of HCl diffusion from vial B to vial A due to the increase of chemical potential difference (HCl) between the two vials. However, no highly ordered 2D or 3D superlattices were once or finally obtained during the growth process of more than 4 months. Panels a-d of Figure 6 show the instant morphology of the assembled structures formed at the air-water interface over a growth period of 4, 10, 15,

Construction of 2D or 3D Superlattice Structures

J. Phys. Chem. B, Vol. 110, No. 48, 2006 24447

Figure 5. Highly ordered 3D superlattices grown at the air-water interface by vapor diffusion after 4 months: (a) TEM image of several 3D superlattices; (b) the enlargement of the boxed area selected from (a); (c) SAED and (d) FFT pattern obtained from (b).

Figure 6. TEM images of structures assembled at the air-water interface from the polydisperse bulk solution of MSA-Ag nanoparticles with 6.0 M HCl solution in vial B: (a) 4, (b) 10, (c) 15, and (d) 18 weeks after introduction of HCl.

and 18 weeks, respectively. Besides some random aggregates, we found gradually developed shuttle-like particles existing in many areas, which were possibly evolved from the small MSAAg nanoparticles by dimeric fusing. After about 4 weeks, we could observe that necks had been formed between nanoparticles at some places, leading to particles connected to each other (Figure 6a). When the growth time is up to 10 weeks, many nanoshuttles appear and also self-assemble into locally ordered structures due to their uniform size (Figure 6b). This result is quite different from the one with MSA-Ag particles in 4.8 ( 1.4 nm as the starting blocks with the same condition where rhombic 3D superlattices could be successfully fabricated. On the one hand, rapid diffusion of HCl vapor will result in a quicker increase of the acidity in vial A. Hence, the production of MSA-Ag nanoparticles at the air-water interface will be fastened and the particles will more probably aggregate. Usually, small metal colloid particles are much more sensitive to the acidity than the large because of their very high surface energy. Thus, size-evolution might happen for most of the small MSAAg nanoparticles before the ordered superlattice structures are formed through a long-time self-correction route. It is noted that when we directly added a concentrated HCl solution in vial A, the assembled structures shown in Figure 6 did not appear and only many arbitrary particle aggregations could be found in the bulk solution by TEM observation. Therefore, the

formation of shuttle-like Ag nanoparticles in the regular dimension is still favored by the air-water interface that provides an important platform for the small nanoparticles to aggregate directionally. On the basis of the above discussions, we can suggest that for the superlattice creation from the more polydisperse MSA-Ag nanoparticles in a small size, a milder increase of acidity in bulk solution is necessary for the selfcorrection route. Comparison of Superlattice Structure Prepared via Fractionation and Self-Correction Routes. As illustrated in Figure 3b, the assembled superlattice obtained from organosols after the PAGE fractionation only displays perfect hexagonal closepacked 2D structures. The uniform distance between Ag nanoparticles is due to the presence of the organic modification layer adsorbed on the particle surface. The mean particle Ag core diameter is ∼2.5 nm and the center-to-center distance between the nearest neighboring particles is determined to be ∼6.6 nm by the FFT pattern. Thus, the interparticle gap could be estimated to be ∼4.1 nm. This separation distance is much longer than ∼1.2 nm possibly produced by two MSA molecules (∼0.6 nm), confirming the presence of another bonded layer of CTA+ on particle surface under cation-anion electrostatic interaction. Allowing for the contribution of the length of each CH2 unit as 0.127 nm, the length of a CTA+ chain was estimated to be ∼2.6 nm. Thus, the total thickness of MSA-CTA bilayers around the particle core could be speculated as ∼3.2 nm by summation of them both. It is obvious that the experimental value of the interparticle gap (∼4.1 nm) is considerably shorter than twice the layer thickness composed of the MSA-CTA chains (∼6.4 nm). The measured lesser value for the interparticle distance than that predicted by alkyl chain lengths implies the conformation change in the chain terminal and the following interdigitation between ligand chains on adjacent particles, which makes for the formation of this 2D superlattice structure. It can be calculated that the ultimate interparticle gap required for the hexagonal packing of the second layer is 3.7 nm for the particles in the size of 2.5 nm if no extraordinary acting force is present. In our case the gap between the nearest Ag core is up to 4.1 nm larger than the restricted maximum, which should be the reason why we failed in obtaining any 3D superlattices though the size of Ag nanoparticle is fairly uniform. On the contrary, a large number of 3D superlattices could be produced at the air-water interface by diffusion of HCl vapor (Figures 4 and 5). On this condition, the surface structure of Ag nanoparticles and the connecting mode of adjacent particles inside the superlattice are different from those capped by the MSA-CTA bilayer derived from water-chloroform phase transfer. Figure 5b presents a higher magnification TEM image of Figure 5a

24448 J. Phys. Chem. B, Vol. 110, No. 48, 2006 (area inside the white frame). Its corresponding FFT pattern (Figure 5d) reveals the arrangement of particles in the observed area is consistent with that expected for [111] projection of an fcc superlattice. The interparticle distance obtained from Figure 5b is ∼1.2 nm, almost equal to double the length of the MSA molecule. This result demonstrates that the chain interdigitation is negligible here and it is the hydrogen bond among MSA molecules that connects the adjacent building nanoparticles and answers for the creation of 3D superlattices in this aciditymodulation process.33 Regarding the high-angle SAED patterns of the superlattice structure obtained by fractionation (Figure 3d) and selfcorrection (Figure 5c) approaches, we can find that the orientation order for the collective atomic lattice is quite different for these two samples. The SAED pattern of the 2D superlattice directly deposited from the chloroform solution exhibits a series of rings, indicating the orderliness of this superlattice only exists in the level of particle stacking (translational order) because the alignment of individual atomic lattices inside the superlattice is still random. Relatively, small arcs clearly appear in the (220) ring of the SAED pattern of the 3D superlattice grown at the air-water interface (Figure 5c), like those existing in a single crystal. This means the Ag nanoparticles inside this 3D superlattice are not only arranged in an orderly fashion, but also are highly oriented (orientation order), which is usually favored by the slow growth of colloidal crystals. In the crystallization process of MSA-Ag nanoparticles by diffusion of HCl vapor, a nucleus can be considered to first form at the air-water interface, followed by the gradual addition of all remaining moieties according to the self-correction mechanism. The faceted morphology of nanoparticles as well as the mediation of the interparticle actions of surfactant groups anchored on the facets will promote the emergence of orientation order.45,46 During the time-related equilibrium growth, if the latter inadequate solvation attached to the surface of the nucleus is in a misaligned fashion, it will soon return to the surrounding solution, while the one oriented with the foregoing core is more likely to remain attached to it and contribute to the further growth of the superlattice structure.30 The orientation order is very important for the collective properties of the superlattice because electronic transport with less scattering is expected to take place within an ordered crystalline structure concerning the possible applications in optical and electrical devices. As for the 2D superlattice prepared through the fractionation route, the rapid growth process by the evaporation of chloroform restrains the particles from finding the most stable position before an unchangeable structure is formed, so the assembly exhibits translational order only and orientation order of the metal cores cannot be achieved in this experiment. Finally, for the energy minimization of polydisperse particles of attracting hard spheres, the resulting structures organized at the liquid-gas or solid-gas interface usually involve the biggest particles at the center surrounded by successively smaller ones.47 However, during the experiment by the vapor diffusion method, the observable fact was different from the theoretical prediction that larger nanoparticles should be adsorbed much better to the air-water interface than smaller ones. Actually, both the smaller and the bigger in the minority whose size are different from other component particles can be gradually repelled toward the boundary of the superlattice during this long-term self-correction route, partially displayed in Figure 4b,c. It is reasonable to consider that in this crystallization process the stabilization energy obtained at the interface (proportional to the square of particle’s radius) is not the only factor to determine the stability

Yang and Kimura of a nanoparticle in a highly ordered 3D structure when additional interactions among nanoparticles are allowed. It is noted that formation of 3D colloidal crystals is also possibly accompanied by the irreversible changes of the stabilizing shells of the building blocks,23 which sometimes even affects properties such as solubility and stability of the component nanoparticles. In our case, when the 3D superlattices are formed at the airwater interface through the gradual diffusion of HCl vapor, the 2.5-nm MSA-Ag particles located inside the superlattices can resist higher acidity with the pH value further decreased in the bulk solution. But for the 2.5-nm MSA-Ag particles optionally assembled at the interface, the increase of acidity will inevitably lead to the aggregation and evolution of the small nanoparticles as reflected in Figure 6. Conclusions We have succeeded in preparing superlattice structures from hydrophilic polydisperse MSA-Ag nanoparticles with the average size of 2.5 nm by PAGE fractionation and self-correction methods. The key point of the former is the observation of a long-range 2D superlattice after solvent evaporation. A noticeable attribute of the latter is the successful preparation of 3D superlattices composed of 2.5-nm particles through gradual diffusion of HCl vapor. This result is especially important since it is still difficult to obtain large-scale highly ordered 3D superlattices by using such a small-sized particle as the building block until now. Comparison of the morphology and structure of the superlattice obtained from these two different routes was made and the reasons resulting in the diversities were discussed. We hope the current research can provide a general and practical strategy for 2D or 3D superlattice creation from other polydisperse thioacid-modified metal or semiconductor nanoparticles. Financial support and granting of the postdoctoral fellowship (P04402) of this work from the Japan Society for Promotion of Science is gratefully acknowledged. This work is also supported in part by a Grant-in-Aid for Scientific Research (S: 16101003) and Scientific Research in Priority Areas, Molecular Spins (15087210) from MEXT. References and Notes (1) Weller, H. Curr. Opin. Colloid Interface Sci. 1998, 3, 194. (2) Collier, C. P.; Vossmeyer, T.; Heath, J. R. Annu. ReV. Phys. Chem. 1998, 49, 371. (3) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu. ReV. Mater. Sci. 2000, 30, 545. (4) Dabbousi, B. O.; Murray, C. B.; Rubner, M. F.; Bawendi, M. G. Chem. Mater. 1994, 6, 216. (5) Ohara, P. C.; Heath, J. R.; Gelbart, W. M. Angew. Chem., Int. Ed. Engl. 1997, 36, 1078. (6) Redl, F. X.; Cho, K. S.; Murray, C. B.; O’Brien, S. Nature 2003, 423, 968. (7) Lin, Y.; Skaff, H.; Emrick, T.; Dinsmore, A. D.; Russell, T. P. Science 2003, 299, 226. (8) Binder, W. H. Angew. Chem., Int. Ed. Engl. 2005, 44, 5172. (9) Urban, J. J.; Talapin, D. V.; Shevchenko, E. V.; Murray, C. B. J. Am. Chem. Soc. 2006, 128, 3248. (10) Harfenist, S. A.; Wang, Z. L.; Whetten, R. L.; Vezmar, I.; Alvarez, M. M. AdV. Mater. 1997, 9, 817. (11) Heath, J. R.; Knobler, C. M.; Leff, D. V. J. Phys. Chem. B 1997, 101, 189. (12) Pileni, M. P. J. Phys. Chem. B 2001, 105, 3358. (13) Stoeva, S.; Klabunde, K. J.; Sorensen, C. M.; Dragieva, I. J. Am. Chem. Soc. 2002, 124, 2305. (14) Daniel, M.-C.; Astruc, D. Chem. ReV. 2004, 104, 293. (15) Lisiecki, I. J. Phys. Chem. B 2005, 109, 12231. (16) Shevchenko, E. V.; Talapin, D. V.; Kotov, N. A.; O’Brien, S.; Murray, C. B. Nature 2006, 439, 55. (17) Taleb, A.; Petit, C.; Pileni, M. P. Chem. Mater. 1997, 9, 950. (18) Taleb, A.; Petit, C.; Pileni, M. P. J. Phys. Chem. B 1998, 102, 2214.

Construction of 2D or 3D Superlattice Structures (19) Rogach, A. L.; Talapin, D. V.; Shevchenko, E. V.; Kornowski, A.; Haase, M.; Weller, H. AdV. Funct. Mater. 2002, 12, 653. (20) 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. (21) Sau, T. K.; Murphy, C. J. Langmuir 2005, 21, 2923. (22) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Science 1995, 270, 1335. (23) Talapin, D. V.; Shevchenko, E. V.; Kornowski, A.; Gaponik, N.; Haase, M.; Rogach, A. L.; Weller, H. AdV. Mater. 2001, 13, 1868. (24) Murray, C. B.; Noris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115, 8706. (25) Wilson, W. L.; Szajowski, P. F.; Brus, L. E. Science 1993, 262, 1242. (26) Arnaud, I.; Abid, J.-P.; Roussel, C.; Girault, H. H. Chem. Commun. 2005, 787. (27) Sweeney, S. F.; Woehrle, G. H.; Hutchison, J. E. J. Am. Chem. Soc. 2006, 128, 3190. (28) Yang, Y; Kimura, K. Chem. Lett. 2006, 35, 462. (29) Yang, Y; Liu, S. M.; Kimura, K. Angew. Chem., Int. Ed. 2006, 45, 5662. (30) Brown, L. O.; Hutchison, J. E. J. Phys. Chem. B 2001, 105, 8911. (31) Chen, S. H.; Kimura, K. Langmuir 1999, 15, 1075. (32) Chen, S. H.; Kimura, K. Chem. Lett. 1999, 28, 1169. (33) Wang, S. H.; Yao, H.; Sato, S.; Kimura, K. J. Am. Chem. Soc. 2004, 126, 7438.

J. Phys. Chem. B, Vol. 110, No. 48, 2006 24449 (34) Yamada, M; Shen, Z.; Miyake, M. Chem. Commun. 2006, 2569. (35) Kalsin, A. M.; Fialkowski, M; Paszewski, M.; Smoukov, S. K.; Bishop, K. J. M.; Grzybowski, B. A. Science 2006, 312, 420. (36) Zhong, Z.; Chen, F.; Subramanian, A. S.; Lin, J.; Highfield, J.; Gedanken, A. J. Mater. Chem. 2006, 16, 489. (37) Schaaff, T. G.; Knight, G.; Shafigullin, M. N.; Borkman, R. F.; Whetten, R. L. J. Phys. Chem. B 1998, 102, 10643. (38) Negishi, Y.; Takasugi, Y.; Sato, S.; Yao, H.; Kimura, K.; Tsukuda, T. J. Am. Chem. Soc. 2004, 126, 6518. (39) Kimura, K.; Takashima, S.; Ohshima, H. J. Phys. Chem. B 2002, 106, 7260. (40) Yao, H.; Momozawa, O.; Hamatani, T.; Kimura, K. Bull. Chem. Soc. Jpn. 2000, 73, 2675. (41) Kimura, K.; Sato, S.; Yao, H. Chem. Lett. 2001, 30, 372. (42) Wang, S. H.; Sato, S.; Kimura, K. Chem. Mater. 2003, 15, 2445. (43) Whitesides, G. M.; Boncheva, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4769. (44) Sakata, J. K.; Dwoskin, A. D.; Vigorita, J. L.; Spain, E. M. J. Phys. Chem. B 2005, 109, 138. (45) Harfenist, S. A.; Wang, Z. L.; Alvarez, M. M.; Vezmar, I.; Whetten, R. L. J. Phys. Chem. B 1996, 100, 13904. (46) Wang, Z. L. AdV. Mater. 1998, 10, 13. (47) Ohara, P. C.; Leff, D. V.; Heath, J. R.; Gelbart, W. M. Phys. ReV. Lett. 1995, 75, 3466.