Encyclopedia of Nanoscience and Nanotechnology
www.aspbs.com/enn
Nanocrystals Assembled from the Bottom Up Edson Rober...
5 downloads
417 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Encyclopedia of Nanoscience and Nanotechnology
www.aspbs.com/enn
Nanocrystals Assembled from the Bottom Up Edson Roberto Leite Federal University of São Carlos, Carlos, SP, Brazil
CONTENTS 1. Introduction 2. Transition Metal Nanocrystals 3. Metal Oxide Nanocrystals 4. Nanocrystalline Composites 5. Characterization of Nanocrystals 6. Summary Glossary References
1. INTRODUCTION The physical and chemical properties of materials on the nanoscale (usually defined in the 1–100 nm range) are of immense interest and increasing importance for future technological applications. Nanoparticles or nanocrystals (in this work the words nanoparticles and nanocrystals are synonymous) generally display properties that differ from those of bulk material. The literature provides several examples of properties, such as magnetic and optical properties, melting point, specific heat, and surface reactivity, which can be affected by particle size [1–5]. A material’s properties are usually very substantially modified in the 1–10 nm sizes. These changes are known as quantum size effects and their origin is directly related to the type of chemical bond in the crystal [6]. The correlation between properties and particle size has been known since the 19th century, when Faraday demonstrated that the color of colloidal Au particles can be modified, changing the Au particle size [7]. However, despite the subject’s long history, interest in nanoparticles has grown considerably over the last decade. The driving force for this increase in research activity is the ability to control a material’s properties by controlling the size and shape of crystals and the arrangement of such particles. These developments can lead to new technologies, including energy conversion [8–12], catalysis and sensors [13, 14],
ISBN: 1-58883-062-4/$35.00 Copyright © 2004 by American Scientific Publishers All rights of reproduction in any form reserved.
ultrahigh density data storage media [15–17], nanoparticle light-emitting diodes [18, 19], and special pigments [20]. The future of these new technologies is strictly dependent on the development of synthetic routes to process metal, metal oxides, and semiconductor nanoparticles, as well as processes that allow such nanoparticles to be manipulated and controlled. This chapter focuses precisely on these two topics (i.e., the synthesis and control of nanoparticles). Special attention is given to particle growth control during high temperature heat treatment. This chapter is dedicated to presenting a concise description of different wet chemical processes to synthesize metal and metal oxide nanoparticles by “bottom-up” methods and of how to control the shape, size, and particle growth of such nanocrystals. This chapter will also analyze the synthesis of nanocomposites. Nanocomposite processing will be considered as a route to control the particle growth, surface oxidation, and agglomeration of nanoparticles. The fact that the combination of different phases may result in a material with superior performance will be considered a secondary, albeit no less important, effect. Indeed, these two factors, control and synergetic effect, must be considered during the selection of material for nanocomposite processing. Neither the synthesis of metal and metal oxide nanoparticles by high energy mechanical milling (“top-down” methods) or vapor phase nor the synthesis of semiconductor nanoparticles, such as II–IV semiconductors crystals, is discussed in this chapter, since the literature contains excellent papers and reviews on these subjects [21–23].
2. TRANSITION METAL NANOCRYSTALS The “bottom-up” methods of wet chemical nanocrystal synthesis are based on the chemical reduction of the salts, or the controlled decomposition of metastable organometallic compounds in an organic or water solution. These reactions are always carried out in the presence of a large variety of stabilizers, which are used basically to control the growth of the initial nanocluster and to avoid particle coagulation
Encyclopedia of Nanoscience and Nanotechnology Edited by H. S. Nalwa Volume 6: Pages (537–554)
538 or agglomeration. The mechanism of nanoparticle formation is generally based on a process of nucleation, growth, and coagulation. This process was proposed by Turkevich and is based on the synthesis of metal nanoparticles by salt reduction [24–26]. This model still is valid and has recently been refined. A recent review authored by Bönnemann and Richards [27] contains a good discussion about the refined model and supplementary references on this subject are also available. Since nanocrystals are unstable from the standpoint of agglomeration and bulk, coagulation and agglomeration are the paths that nanoparticles follow to decrease their high surface area, thus becoming more stable. In the absence of any extrinsic impediment, the unprotected particle coagulates, basically under the action of van der Waals forces. To prevent the coagulation process from occurring, the particle surface can be protected by electrostatic stabilization and/or steric stabilization [28]. Electrostatic stabilization is based on the Coulombic repulsion between particles, promoted by a double layer composed of ions adsorbed on the particle surface. The electrostatic stabilization process can be modified by several parameters, such as ionic strength of the dispersing media, ion concentration, and the presence of neutral adsorbate, which may replace the adsorbed ion on the particle surface. Steric stabilization is based on the steric hindrance caused by organic molecules that are attached to the particle surface, forming a protective layer that prevents particle coagulation or agglomeration. This type of stabilizing system can be viewed as a nanocomposite material, since the organic layer forms a nanometric scale second phase [29, 30]. Several kinds of protective groups can be used as steric stabilization agents, among them polymers and block polymers, P, N, and S donors (phosphanes, amines, thioethers), surfactants, organometallic compounds, and solvents. A detailed description of the several types of steric stabilizers used during the synthesis of metal nanocrystals is given in Bradley’s review [31]. The synthesis of transition metal nanocrystals can be divided basically into two major groups: salt reduction and decomposition method. Examples of these methods are described in the following section.
2.1. Synthetic Methods The salt reduction method is a process by which a reduction agent reduces the metal salt, in solution, to metal. These reactions can be done in water or in an organic solution. In an organic solution, the solvent can also act as a reduction agent. Alcohols are generally useful reduction agents, particularly alcohols containing -hydrogen. In this process, the alcohol is oxidized to the corresponding carbonyl group. An example of this kind of synthesis is the processing of palladium nanoparticles through the reduction of palladium acetate by methanol [32]. Teranishi and Miyake [33] reported on the reduction of H2 PdCl4 by alcohols to synthesize Pd nanoparticles, demonstrating that the mean diameter of Pd nanocrystals can be controlled from 1.7 to 3.0 nm in a one-step process by changing the amount of protective polymer, poly(N -vynil-2-pyrrolidone) (PVP) and the kind and/or concentration of alcohol in the solvent. The
Nanocrystals Assembled from the Bottom Up
solvent they used was water. They also showed that the reduction rate of [PdCl4 ]− ions is an important factor in the production of smaller Pd particles. The reduction rate was controlled using different kinds of alcohol. The reduction of metal salts by the addition of a reducing agent in a nonreducing solvent is a well-established synthetic route for the preparation of aqueous suspensions of metal nanocrystals. Faraday, for instance, used phosphorous vapor to promote the reduction of [AuCl4 ]− in aqueous solution to synthesize gold nanoparticles [7]. Different kinds of reducing agents have been used to process gold nanocrystals, allowing for the processing of particles ranging from 1 to 100 nanometers in diameter. Turkevitch and co-workers [24, 26] established the first reproducible standard protocol for the synthesis of gold nanoparticles. Their processing of gold nanoparticles by the reduction of [AuCl4 ]− with sodium citrate, for example, became a standard for histological staining applications [34] and for undergraduate experiments in surface and nanomaterials chemistry [35]. Platinum nanoparticles can also be synthesized by the reduction of metal salts, using a reducing agent [36, 37]. Van Rheenen et al. [37] demonstrated that the morphology of platinum particles could be controlled by controlling the synthetic parameters, such as temperature, protective polymer, time, pH, reagent concentration, and the sequence of reagent additions. These authors used various reducing agents and chloroplatinic acid as platinum salt. An interesting synthetic route was recently developed based on the reduction of organometallic compounds by dihydrogen at low pressure and temperature [38–42]. The organometallic compounds used were low-valent alkene or a poly-ene complex of the desired metal. Using this process, well-dispersed nanoparticles of Ru, Pt, Ni, and Co with a narrow size distribution were synthesized. The particles were stabilized by the presence of PVP. Based on a similar process, Ould Ely et al. [43] synthesized nanoscale bimetallic Cox Pt1−x particles, using Co(3C8 H13 )(4-C8 H12 ) and Pt2 (dba)3 (dba = bis-dibenzylidene acetone) as organometallic compounds. They found that the alloy’s composition was determined by the initial ratio of the two organometallic precursors. Recently, the so-called “polyol process” [44] has been used successfully to process magnetic nanoparticles with a very narrow particle size distribution [44–46]. This process is based on the reduction of metallic salt in solution, at a high temperature (100 < T < 300 C), by the addition of a polyol (such as ethylene glycol), resulting in nanometric particles. In this process, surfactants such as oleic acid are used to control particle growth and stabilize the nanoparticles. Park and Cheon [47] discussed an interesting synthetic route to process solid solution and core–shell type cobalt– platinum nanoparticles via a redox transmetallation reaction, reporting they had obtained nanoparticles of solid solution and core–shell structures smaller than 10 nm. These alloys were formed by redox transmetallation reactions between the reagents without the addition of reducing agents. The reaction between Co2 (CO)8 and Pt(hfac)2 (hfac = hexafluoroacetylacetonato) resulted in the formation of solid solution, while the reaction between Co nanoparticles and Pt(hfac)2 in solution resulted in “Cocore –Ptshell ”
Nanocrystals Assembled from the Bottom Up
type nanoparticles. Narrow particle size distributions were achieved in both processes. The organometallic compounds of transition metals usually display low thermal stability, decomposing into their respective metals even under mild conditions. Owing to these properties, organometallic compounds can be considered good sources to process metal nanoparticles. Metal carbonyl pyrolysis has been used for the synthesis of several metal nanoparticles, although a broad particle size distribution is usually obtained [48, 49]. Park et al. [50] reported on the synthesis of iron nanorods and spherical nanoparticles using the thermal decomposition of Fe(CO)5 , in the presence of surfactant. They found that rodlike particles, with a higher aspect ratio, could be obtained by changing the concentration of didodecyldimethylammonium bromide during the reaction process. Alivisatos et al. [51] recently reported on the control of the size and shape of Co nanocrystals. A synthetic route, based on the principles applied to the synthesis and control of CdSe nanocrystals, was used [52]. These authors discussed the synthesis of Co nanoparticles with high crystallinity, narrow particle size distribution, and a high degree of shape control. The nanocrystals are produced by injecting an organometallic precursor [Co2 (CO)8 ] into a hot (T ∼ 180 C) surfactant mixture [oleic acid and trioctylphosphine oxide (TOPO)] under an inert atmosphere. An interesting approach to synthesize metal alloy nanocrystals is the use of simultaneous salt reduction and thermal decomposition processes. Sun et al. [53] reported on the synthesis of iron–platinum (FePt) nanoparticles through the reduction of platinum acetylacetonate by a diol, and decomposition of iron pentacarbonyl [Fe(CO)5 ] in the presence of a surfactant mixture (oleic acid and oleyl amine). Based on a similar approach, Chen and Nikles [54] synthesized ternary alloy nanoparticles (Fex Coy Pt100 − x − y ), using a simultaneous reduction of acetylacetonate and platinum acetylacetonate and thermal decomposition of Fe(CO)5 and obtaining an average particle diameter of 3.5 nm and narrow particle size distribution. The decomposition of the organometallic compounds can be promoted by different energy sources, such as acoustic waves (sonochemistry). Sonochemistry stems from acoustic cavitation—the formation, growth, and implosive collapse of bubbles in the liquid. These phenomena can generate high temperatures and high cooling rates [55, 56], creating a favorable environment to promote the decomposition of organometallic compounds and the synthesis of nanoparticles. Suslick et al. [57] used this approach to synthesize Fe nanocrystals. They used Fe(CO)5 as the precursor, promoting sonochemical decomposition (by ultrasonic irradiation) in the presence of PVP as stabilizer and obtaining Fe nanoparticles with a mean particle size of 8 nm. As previously described, the synthesis of metal nanocrystals is based on a colloidal process whereby a metallic salt or an organometallic compound, in solution, is reduced to a metal by a reducing agent (salt reduction method) or by thermal decomposition (thermal decomposition method). The reduction is developed in the presence of a stabilizer (electrostatic stabilization and or steric stabilization). These additives have two basic functions: to control particle growth and to prevent particle agglomeration.
539 These processes are based on nucleation, growth, and agglomeration, or coagulation. As illustrated in Figure 1, the general process of metal nanocrystal synthesis can be divided, for didactic purposes, into five steps. The first step (step I) consists of the reduction of the metallic precursor (M+ X− ), which results in metal atoms (M ). These metallic atoms, ions, and metallic clusters will interact (step II), resulting in a metallic cluster growth process. Steps I and II are reversible. When the cluster grows to a critical size (step III), the process becomes irreversible (thermodynamic condition). Particle size can be controlled with the aid of stabilizers (step IV). The presence of only one stabilizer can result in a spherical particle. The origin of this morphology is thermodynamic. In fact, the cluster will grow in a geometrical arrangement in order to minimize the surface energy. The presence of two simultaneous stabilizers, on the other hand, may give rise to a preferential growth process caused by the preferential adsorption of one of the stabilizers. This process, which leads to the formation of anisotropic particles such as nanorods, occurs under a kinetic condition. Particle agglomeration, basically, is prevented by steric stabilization under the influence of the molecules attached to the particles’ surface (step V). Step V is essential to control nanoparticle deposition. Thus, colloidal metal dispersion can be used as a building block to produce functional materials [58]. This nanocrystal-based self-assembly process is governed by particle–particle and particle–substrate interaction. The nanocrystal self-assembly process requires a monodispersion system (particle size deviating by less than 10% from the average size) [59] and can be achieved by solvent evaporation [53, 60] or polymer-mediated nanocrystal assembly [61]. Needless to say, the process described in Figure 1 is ideal. One of the major obstacles to achieving good control over metal colloidal synthesis is in separating the nucleation from the growth process. A positive advance in this direction is the synthesis of CdSe nanocrystals, in which nucleation is separated from the growth process by rapidly injecting the precursor in a solvent at high temperature [62, 63]. This approach has become the most suitable thermal decomposition method and has been applied successfully in the synthesis of Co nanocrystals [51], as discussed in Section 2.1.
Figure 1. Schematic representation of the general process of metal nanocrystal synthesis.
540 2.2. Properties When we talk about metal nanocrystal, two major application fields come to mind: catalysis and magnetic materials. In this section, we will present some examples of metal nanocrystal properties in these fields. The magnetic properties of nanocrystals differ in several aspects from their bulk properties. These differences stem from the large fraction of atoms located on the surface or at the interface, leading to a local environment that differs greatly from that of the internal atoms. Hence, the intrinsic and extrinsic properties of nanostructured materials must be different. Besides these modifications, a second phenomenon must be considered as well. Unlike bulk ferromagnetic materials, which are composed of multiple ferromagnetic domains, a sufficiently small ferromagnetic particle consists of a single magnetic domain. These two effects can lead to new magnetic properties that can help the development of ultra-high-density magnetic recording, particularly on metal thin film media. Metal thin film media are generally processed by physical deposition techniques, which usually require postdeposition annealing. This annealing treatment can result in loss of control over particle size and particle size distribution due to particle growth [64]. The use of nanoparticles as building blocks to fabricate magnetic thin films, via self-assembly, appears to be a good alternative to process metal thin film media [53, 61]. Sun and co-workers [53] showed that ferromagnetic FePt nanocrystal superlattices are chemically and mechanically robust and can support high-density magnetization reversal transitions. The development of metal nanoparticles whose particle size, size distribution, and composition is strictly controllable can result in high performance catalysts. The preparation of metal nanoparticles prior to deposition on a support can provide fine control over the composition and microstructure [65]. Such control is difficult with the traditional impregnation/reduction catalyst preparation. Polymerprotected precious metal nanoparticles have been used with good results in the hydrogenation of unsaturated organic molecule catalysts [66–68]. Transition metal nanoparticles have recently become a fundamental component for the synthesis of carbon nanotubes. This synthesis can be done by noncatalytic and catalytic methods [69]. When the growth method of nanotubes by chemical vapor deposition (even of single-walled carbon nanotubes) is used, there is strong evidence that carbon nanotubes grow via the base-growth mode, with the catalytic nanoparticle acting as seed for the growth [70, 71]. In fact, the size of the nanoparticle determines the diameter of the nanotube [72]. In the catalyst method, Ni, Fe, and Co are the transition metals that can serve as catalysts for carbon nanotube growth [73]. The catalyst method produces large quantities of very long carbon nanotubes [74].
3. METAL OXIDE NANOCRYSTALS Metal oxides represent an important class of materials with a variety of technological applications. Several reports in the literature describe the effects of size on the various properties of this class of materials. Nanocrystalline metal oxide
Nanocrystals Assembled from the Bottom Up
semiconductors such as TiO2 , SnO2 , and ZnO, for example, display a quantum confinement effect, with enlargement of the bandgap as the particle size decreases [75–77]. Colloidal nanocrystals with quantum size effects are promising building blocks for novel electrical and optoelectronic devices [2, 78]. The ferroelectric phenomena in complex ferroelectric oxides are suppressed when the particles are reduced to a critical size [79–81]. Ceramic materials with nanometric scale microstructures display superior mechanical properties and enhanced electronic conductivity [82, 83]. Based on the above analysis, the development of metal oxides of nanometric dimensions can result in devices and materials with superior performance. However, these developments are directly related to the development of synthesization methods that allow for controlled particle size, particle morphology, and deposition. Once again, the “bottom-up” methods of wet chemical nanocrystal synthesis are apparently the most viable approach to achieve such control. Compared to the control attained in the synthesis of metal and II–IV semiconductor nanocrystals, the control of metal oxide nanocrystals is still in its earliest stage, particularly insofar as the synthesis of complex metal oxide nanocrystals (oxides formed of more than one cation) is concerned. The synthesis of metal oxide nanocrystals by wet chemical processes can be divided basically into two major groups: (a) chemical synthesis method based on the hydrolysis of metal alkoxides or metal halides; (b) chemical synthesis based on the nonhydrolytic method. Examples of these methods are described in the following section.
3.1. Synthetic Methods The chemical synthesis of metal oxide nanocrystals based on hydrolysis falls into two major groups: hydrolysis of metal alkoxides and hydrolysis of metal halides and other inorganic salts. Following I give a description of these two groups. Metal alkoxide compounds are defined as compounds that have metal–oxygen–carbon bonds. Si(OC2 H5 )4 , for instance, which is known as tetraethyl orthosilicate (TEOS), is an alkoxide compound. This class of compound is highly reactive with water. Because the hydroxyl ion (OH− ) becomes bonded to the metal of the organic precursor, this reaction is called hydrolysis. Reaction 1 shows a typical hydrolytic reaction of an alkoxide compound, M-OR + H2 O → M-OH↓ + ROH
(1)
where M represents Si, Ti, Zr, Al, and other metals, R is a ligand such as an alkyl group, and ROH is an alcohol. Hydrolytic reactions are strongly dependent on water content and catalysts. Due to the high reactivity of alkoxide compounds with water, hydrolytic reactions must be carried out in an atmosphere devoid of water vapor and the solvents used must have a very low water content. A partially hydrolyzed metal alkoxide molecule can react with other partially hydrolyzed molecules by a polycondensation reaction, as described in the following equations. M-OH + M-OR → M-O-M + ROH
(2)
M-OH + M-OH → M-O-M + H2 O
(3)
541
Nanocrystals Assembled from the Bottom Up
This type of reaction leads to the formation of an inorganic polymer or a three-dimensional network formed of metal oxianions. The described process is called metal alkoxide-based sol–gel. The literature contains excellent reports providing in-depth analyses of this method [84, 85]. The sol–gel process allows for very good chemical homogeneity and offers the possibility of obtaining metastable phases, including the amorphous phase. This process normally promotes the formation of amorphous metal oxides, which require thermal or hydrothermal treatment to promote crystallization. Several factors affect the sol–gel process, including the kind of metal alkoxide, pH of the reaction solution, water:alkoxide ratio, temperature, nature of the solvent, and stabilizers [84]. By varying these parameters, particles can be synthesized with controlled size, morphology, and agglomeration. When the metal alkoxide’s hydrolytic reaction rate is too fast, particle size and morphology are more difficult to control. A good alternative to overcome this problem is to use organic additives, which act as chelating ligands (carboxylic acids, -diketones, and others) and decrease the precursor’s reactivity [86]. The sol–gel process can generally be divided into three steps, as illustrated schematically in Figure 2: (1) precipitation of hydrous oxide particles; (2) control of hydrous oxide particle coagulation; and (3) crystallization of the hydrous oxide particle. Thus, the sol–gel process requires control of the particle size and morphology during the precipitation and coagulation steps and during the heat or hydrothermal treatment to promote crystallization. These three steps are now discussed in detail. The precipitation of amorphous metal oxide (step 1) is controlled by a nucleation growth mechanism. The nucleation mechanism used in this step is well described by the LaMer and Dinegar theory [87]. Following this model, the supersaturation of hydrous oxides increases continuously (by a change in temperature or pH) until a critical concentration is reached. In this condition, nucleation occurs very rapidly and leads to precipitation. Precipitation decreases the supersaturation to levels below the critical concentration,
preventing further nucleation and precipitation. After nucleation occurs, the nuclei thus formed grow, reducing the concentration until an equilibrium concentration is achieved. The solubility (S) of the particle formed during nucleation is related to the particle’s size (d) by the Ostwald– Freundlich equation, S = S0 exp4SL Vm /RTd
where S0 is the solubility of the flat surface, SL is the solid– liquid interfacial energy, Vm is the molar volume of the solid phase, R is the gas constant, and T is the temperature. This dependence controls the nucleation and the growth process. After the nucleation step, the smaller particles show a tendency to dissolve, increasing the supersaturation, which may cause reprecipitation of the larger particles. This mechanism, which controls particle growth, is known as Ostwald ripening. Controlled hydrolysis is one of the most popular methods for processing silica spheres in the range of 10–1000 nm. The method was developed by Stöber et al. [88] and is based on the hydrolysis of TEOS in a basic solution of water and alcohol. Particle size depends on the reactant concentration, that is, the TEOS/alcohol ratio, water concentration, and pH (>7). This method has been extended to other metal oxide systems with similar success, particularly for TiO2 synthesis [89, 90]. The hydrous oxide particles precipitated by the hydrolysis of an alkoxide compound have the same tendency to agglomerate as that described under item 2 for metal colloid systems. Different stabilizers can be used to stabilize these particles and prevent coagulation (step 2). These stabilizers control coagulation by electrostatic repulsion or by steric effects [28], similarly to the metal colloid systems. There is not only a similarity but also a fundamental difference between the approach used to control coagulation in the sol–gel process and that used for metal nanocrystal systems. In the sol–gel process, the surface charge is controlled by the protonation or deprotonation of the hydrous oxide particle surfaces (M-OH). Thus, the charge-determining ions are H+ and OH− . The ease with which protonation or deprotonation occurs depends on the metal atoms and can be controlled by the pH. The pH at which the particles are neutrally charged is called the isoelectric point. At the pH > isoelectric point, the particles are negatively charged [see Eq. (5)] while, at the pH < isoelectric point, the particles are positively charged [see Eq. (6)]: M-OH + OH− → M-O− + H2 O +
M-OH + H →
Figure 2. Schematic representation of the sol–gel process. For more details about each step, see Section 3.
(4)
M-OH+ 2
(5) (6)
Electrostatic stabilization is most commonly employed in the water solution system, while steric stabilization can be more effective in organic media [91, 92]. A steric stabilizer can be used to control the condensation reaction during the precipitation of hydrous oxides. In this case, the stabilizer is added during the hydrolysis step [93–96]. Peiró et al. [97] recently reported on the synthesis of TiO2 anatase phase with a nanorod morphology (9 × 5 nm size) using controlled hydrolysis of tetraisopropyl orthotitanate and tetrabutylammonium hydroxide as a steric stabilizer agent.
542 The crystallization process (step 3) can be considered the critical step in the sol–gel process when a crystalline phase is desirable. If an amorphous phase is the final target, as in the case of SiO2 nanoparticle processing, the synthesis ends with step 2. However, for crystalline materials, a heat or hydrothermal treatment is necessary to promote the crystallization of the amorphous hydrous oxide which is formed during hydrolysis. Such subsequent treatments can lead to particle growth and modify the particles’ morphology. In the case of crystallization by heat treatment, the hydrous oxide colloidal suspension must be dry before the treatment. During the heat treatment, normally done in an electric furnace, crystallization occurs by a nucleation growth process and can be described by the standard nucleation growth theory [84, 98]. Since the amorphous phase crystallizes via the nucleationgrowth process, particle size and growth can be controlled based on the separation of the nucleation phenomena from the growth process. However, during crystallization, each hydrous metal oxide particle can generate several nuclei, rendering it very difficult to control particle morphology and shape. Figure 3 shows a high-resolution transmission microscopy (HRTEM) image of the nucleation of several PbTiO3 nuclei in an amorphous inorganic nanoparticle. This figure shows the particle before and after crystallization induced by an electronic beam. The polynuclei process generates polycrystalline particles rather than freestanding ones. Controlling the generation of polynuclei in a single amorphous particle is the main challenge involved in obtaining crystalline metal oxide freestanding nanoparticles through the sol–gel process or by any other process that requires crystallization by heat treatment at high temperatures. The origin of the polynuclei process occurring during sol–gel amorphous precursor crystallization is assumed to be related to the preferential heterogeneous nucleation process (surface and interface nucleation) in detriment to homogeneous crystallization (bulk nucleation). The presence of hydroxyl groups and other defects on the particle surface can contribute to reduce the Gibbs free energy for crystallization, rendering the surface crystallization more favorable than the bulk crystallization. Since crystallization occurs in a scenario of high driving force (high temperature of heat treatment),
Figure 3. HRTEM image of the nucleation of several PbTiO3 nuclei in an amorphous inorganic nanoparticle. (a) Amorphous particle before the crystallization (inset shows the electron diffraction pattern typical of amorphous materials) induced by the electron beam (300 kV); (b) the same particle after the crystallization process (inset shows the electron diffraction pattern of the crystalline materials).
Nanocrystals Assembled from the Bottom Up
surface crystallization must occur first, followed by bulk crystallization, giving rise to a particle with several nuclei. A possible way to avoid this problem is to suppress surface crystallization by using an inhibitory surface layer. If crystallization occurs at a temperature that favors bulk crystallization, a single nucleus can be generated, resulting in a freestanding particle. This approach was used recently to process freestanding lead zirconate titanate (PZT) nanoparticles [99]. Liu and co-workers [99] used a sol–gel process based on controlled hydrolysis and a two-step heat treatment. They first applied a 12-h treatment in Ar atmosphere at 700 C, which formed a surface layer rich in carbonaceous materials on the nanoparticles, inhibiting surface nucleation. A second treatment was carried out at 500–600 C, in air, to burn out the carbon residue. Freestanding PZT nanoparticles with a mean particle size of 17 nm were reported. Freestanding particles are desirable in a variety of fundamental studies and in some technologies, particularly for ferroelectric metal oxides such as PbTiO3 (PT), Pb(Zr,Ti)O3 (PZT), BaTiO3 (BT), among others. Freestanding and single crystalline nanorods of BT and SrTiO3 (ST) were recently obtained [100, 101]. The approach used in both these studies to obtain this type of material was the injection of a bimetallic alkoxide compound into a solvent at high temperature (100–280 C), in which the hydrolysis took place (injection–hydrolysis method). Murray et al. [100] synthesized BT nanoparticles with diameters ranging from 6 to 12 nm based on this approach, controlling the particle size by the bimetallic alkoxide (BaTi(OR)6 )/oleic acid ratio. Park and co-workers [101] reported on the synthesis of BT and ST single crystalline nanorods using a similar process. The origin of nanorod morphology is not yet well understood. Nonetheless, the above-described approach appears to promote and control crystallization with no extra heat treatment, allowing for good control of particle size and morphology. Another alternative approach to avoid the heat treatment process is to promote crystallization under hydrothermal conditions, a process that is widely used in the synthesis of zeolites [102, 103]. Ying and Wang [104] used hydrothermal treatment to promote the crystallization of anatase and rutile phases, using an alkoxide sol–gel route and achieving the crystallization of anatase TiO2 phase with a mean particle size of 10 nm at 180 C, as well as the synthesis of ultrafine rutile TiO2 phase obtained by hydrothermal treatment in an acidic medium. The hydrolysis of metal halides and other inorganic salts is a method widely employed to process metal oxide nanoparticles, such as TiO2 [105, 106], doped and undoped SnO2 [107–111], ZnO [112, 113], ZrO2 [114, 115], Y2 O3 [116], and others. This process is less sensitive to water content, requiring less control than the hydrolysis of metal alkoxide. In fact, the hydrolytic process normally occurs in a water solution. In solution, the metallic salt generates the anion (Cl− , n+ F− , NO− 3 , and others) and the cation (M ). The cation is normally hydrolyzed by pH chancing. Hydrolysis promotes the precipitation of an insoluble amorphous hydrous metal oxide. Thus, the steps illustrated in Figure 2 to describe the metal alkoxide hydrolysis-based sol–gel process can also be used to describe the sol–gel process based on inorganic salts. Equations (7) and (8) describe the reaction involved during
543
Nanocrystals Assembled from the Bottom Up
the hydrolysis of inorganic salt in water solution. Precipitation will occur by increasing the pH or the [OH− ] concentration MX − H2 O→ M+ + X− M+ + X− + −pH
control→
MOH↓ + X− + H+
(7) (8)
The synthesis based on this approach requires the same control described earlier for the sol–gel method related to the hydrolysis of metal alkoxide. However, control of the atmosphere and water content in the solvents is much less demanding. Again, the major problem of the metal salt hydrolysis approach is the crystallization step, since a heat treatment or hydrothermal treatment is necessary to promote crystallization. The synthesis of metal oxides based on the hydrolysis of inorganic salts or metal alkoxides, with a high degree of crystallinity at room temperature, still represents a challenge. The best route to obtain good crystallinity at low temperatures with minimum particle growth is the hydrothermal treatment. Nütz and Haase [107] synthesized wellcrystallized Sb-doped SnO2 nanocrystals, with particles in the range of 4–9 nm, using a hydrothermal treatment of colloidal gel. The gel was treated in an autoclave at temperatures in excess of 250 C. These researchers used a solution of SnCl4 and SbCl3 or SbCl5 in fuming HCl as precursors and promoted hydrolysis by increasing the pH (using aqueous ammonium). Goebbert et al. [110] also reported on the synthesis of well-crystallized Sb-doped SnO2 using the hydrothermal process. However, they used a solution of SnCl4 and SbCl3 or SbCl5 in ethanol, promoting hydrolysis by raising the pH (using aqueous ammonium). The hydrothermal treatment was carried out at 150 C using 10 bar of pressure. This synthesization route produced nanocrystals in the range of 5 nm. Leite [117] recently demonstrated that well-crystallized SnO2 nanocrystals could be produced at room temperature with no hydrothermal treatment. This process is based on the hydrolysis of SnCl2 in an ethanolic solution, followed by dialysis to remove the Cl− ions. The result of this dialysis is a transparent colloidal suspension, as depicted in Figure 4a. Figure 4b shows an HRTEM image of typical SnO2 particles obtained by this process (particles in the range of 2–5 nm).
Figure 4. (a) Picture of the SnO2 colloidal suspension before and after the dialysis (transparent colloidal suspension); (b) HRTEM image of typical SnO2 particles obtained by the hydrolysis of SnCl2 in an ethanolic solution (particles in the range of 2–5 nm).
Zinc oxide (ZnO) nanocrystals have also been synthesized at room temperature. The process, developed by Bahnemann et al. [118], consists of hydrolyzing zinc acetate dihydrate dissolved in 2-propanol by the addition of NaOH in a 2-propanol solution. A colloidal suspension of crystalline ZnO nanoparticles is obtained without hydrothermal treatment. Similar results were obtained by Spanhel and Anderson [119] and by Meulenkamp [112]; however, they dissolved the zinc acetate dihydrate (Zn(Ac)2 · xH2 O) in ethanol and used LiOH to promote the Zn+2 hydrolysis. Particles in the range of 3–6 nm were reportedly obtained by both of these processes. Rusakova et al. [120] used an interesting approach to control the particle growth of hydrous metal oxide gels. They showed that the growth could be inhibited by replacing the surface hydroxyl group, before the crystallization step, with a functional group that does not condense and that can produce small secondary-phase particles which restrict boundary mobility at high temperatures. These authors reported that fully crystalline SnO2 , TiO2 , and ZrO2 nanocrystals (ranging in size from 1.5 to 5 nm) can be obtained after heat treating the precipitate gel at 500 C, by replacing the hydroxyl group with the methyl siloxyl group before firing. The development of metal oxide nanocrystals by nonhydrolytic synthesization routes results in materials whose surfaces are free of OH− groups and nanocrystals with different properties particularly suitable for catalytic and sensor applications. Several nonhydrolytic processes have been developed to process metal oxides, and the molecular chemistry of these various methods is discussed by Vioux [121]. Based on the strategy used to process II–IV semiconductor nanocrystals, using the rapid decomposition of molecular precursor in the presence of strong coordinating agents, Colvin and co-workers [122] proposed an interesting route to process TiO2 nanocrystals, based on the reactions TiX4 + TiOR4 → 2TiO2 + RX
(9)
TiX4 + 2ROR → TiO2 + 4RX
(10)
where X is a halide ion (Cl− , F− , Br− , I− ) and R is an alkyl group. The synthetic route involved injecting the metal alkoxide [Ti(OR)4 ] into a titanium halide mixed with TOPO and a solvent at high temperature (300 C). Nanoparticles with a mean particle size of 7.3 nm and anatase phase were obtained. Alivisatos et al. [123] demonstrated that transition metal oxide nanocrystals (-Fe2 O3 , Mn3 O4 , Cu2 O) could be prepared using a nonhydrolytic process based on the thermal decomposition of metal Cupferron complexes Mx Cupx [M = metal ion, Cup = C6 H5 N(NO)O− ] in a hot solvent with surfactant. Their results suggest that a good level of control can be achieved when this approach is used to process metal oxide nanoparticles. Camargo et al. [124–127] recently developed a new route to synthesize lead-based perovskite nanoparticles, such as PT [124], PZT [125], PbZrO3 (PZ) [126], and PbHfO3 (PH) [127]. This method, which apparently involves no hydrolytic reaction and is carbon- and halide-free, is called the oxidant peroxo method (OPM) because it is based on the oxidation– reduction reaction between Pb(II) ion and water-soluble
544 metal–peroxide complexes with high pH. This process results in an inorganic amorphous precursor that requires subsequent thermal treatment to promote crystallization of the desired phase. Figure 5 shows an HRTEM image of a PT nanoparticle obtained by this process. The low crystallization temperatures (400–450 C for the PT phase) of the amorphous precursor suggest that the OPM method favors the formation of a homogeneous inorganic compound. An important nonhydrolytic chemical process is the so-called Pechini process [128], or in-situ polymerizable complex (IPC) [129]. This process is based on the ability of polycarboxylic acids, particularly citric acid (CA), to form very stable water-soluble chelate complexes. Even cations with a high tendency to become hydrolyzed, such as Ti+4 and Nb+5 , can be chelated by CA in a water solution, preventing the hydrolysis and precipitation of hydrous metal oxide. The CA complex thus formed can be immobilized in a solid organic resin through a polyesterification reaction with ethylene glycol (EG). This process leads to the formation of a polymeric precursor with the cations of interest randomly distributed in a three-dimensional solid network, avoiding precipitation or phase segregation during the synthesis of the metal oxide compound [130]. Figure 6 schematically represents the IPC process used in the synthesis of very fine particles. This method is widely used to process titanatates [131–137], niobates [138–140], and other kinds of polycationic or single cationic metal oxides [141, 142]. In the last five years, this method has also proved suitable to process oxide thin films with superior performance [143–148]. Using this process, PbTiO3 thin film [148] and nanometric powder [149], for example, can be synthesized at temperatures as low as 450 C, resulting in a metastable cubic PbTiO3 phase. Crystallization was observed at a temperature at which longrange diffusion had to be constrained, and the thermodynamic equilibrium configuration was kinetically suppressed [149]. The ability to form complex metal oxides at low crystallization temperatures and metastables phases is not yet well understood, but it is generally assumed to be associated with the tendency of a polycationic CA complex to develop during the chelation step in water solution [141], and/or the tendency to form an inorganic amorphous phase, with a local
Figure 5. HRTEM image of a PT nanoparticle obtained by the OPM.
Nanocrystals Assembled from the Bottom Up
Figure 6. Schematic representation of the IPC process: (a) CA complex formation in water solution; (b) polymerization promoted through a polyesterification reaction between the CA complex and EG; (c) prepyrolysis of the polymeric precursor. In this step an inorganic amorphous precursor is obtained (in detail electron diffraction pattern of an amorphous inorganic precursor); (d) crystallization process.
symmetry close to that of the crystalline phase, during the crystallization step [150]. The major problem with this process is maintaining control over the particle size and morphology. During the crystallization process, it is very difficult to keep the nucleation and growth processes separate, resulting in agglomerates made up of nanocrystals. The particle growth process, which was studied in the final stage of the crystallization of nanometric powder processed by the IPC method, showed that growth occurs in two different stages [151]. At heat treatment temperatures of <800 C, the growth process was associated with the surface diffusion mechanism, with an activation energy in the range of 40–80 kJ/mol. At a temperature of >800 C, particle growth is controlled by densification of the agglomerate formed by nanometric particles and by the neck-size-controlled growth mechanism [152]. Basically, two methodologies have been used to control the particle size of metal oxides processed by the IPC method. Quinelato et al. [142] demonstrated that the particle size and morphology of CeO2 -doped ZrO2 could be controlled by controlling the metal/CA ratio. A high concentration of CA leads to smaller particles with a soft agglomeration. Leite et al. [153, 154] showed that the particle size and morphology of SnO2 could be controlled by the addition of dopants such as Nb2 O5 and rare earths. The same authors [154] also showed that doped SnO2 nanocrystals are highly stable against particle growth, even at high temperatures. The technique used to achieve this high stability was to process supersaturated solid solution between the SnO2 and the dopant. Segregation of the dopant on the nanocrystal surface occurs during the heat treatment, decreasing the particle boundary mobility or the surface energy. This approach was originally developed to control the particle growth of metal nanocrystals [5, 155] and was used successfully to control the growth of metal oxide nanocrystals.
545
Nanocrystals Assembled from the Bottom Up
3.2. Properties The development of metal oxide nanocrystals can lead to significant improvements of their properties, resulting in materials with superior chemical, electrical, and magnetic performance. An example of how metal oxide-based nanostructured materials can very substantially improve performance is in semiconductor gas sensors. The sensing characteristic of a semiconductor gas sensor using SnO2 can be improved by controlling basic parameters relating to the receptor and transducer functions. Stated simply, chemical sensors consist of two functions, that is, a receptor function, which recognizes a chemical substance, and a transducer function, which converts the chemical signal into an output signal. The transducer function is associated directly with the grain size of the metal oxide and the depth of the surfacecharge layer (Debye length) [14]. The semiconductor’s sensitivity should improve dramatically when the grain size has the same order of magnitude as the Debye length (LD ) (e.g., when the grain size is comparable to 2LD ). The LD for SnO2 is ∼3 nm [156]. The sensitivity can therefore be improved by decreasing the grain size or modifying the LD using impuredoped elements [14]. The sensor’s response time can also be improved by controlling the particle size. Leite et al. [153, 157] demonstrated that the addition of Nb2 O5 resulted in a smaller SnO2 particle size and a shorter response time for the detection of methanol. The receptor function can be modified by the introduction of dopants on the SnO2 surface. Yamazoe [14] showed that the addition of precious metal (Pd and Ag) on SnO2 can improve gas sensor performance. Carreño et al. [158] showed that the surface reactivity of doped SnO2 nanoparticles, in terms of methanol oxidation, can be modified by controlling the dopant’s surface segregation. Electrochemical devices based on the Li+ intercalation process, such as lithium batteries and electrochromic windows, constitute another area in which the development of nanostructured materials based on metal oxide semiconductors can result in high performance devices. Aurbach and co-workers [159] tested SnO nanoparticles processed by the sonochemical method for use as anodes in rechargeable Li batteries. The SnO nanocrystal was found to be an effective material for electrodes. These electrodes reached almost their theoretical capacity (∼790 mAh/g) in electrochemical lithiation–delithiation processes as opposed to a Li counterelectrode in an aqueous Li salt solution. Felde et al. [160] reported the remarkable electrochromic effect of layers of nanocrystalline SnO2 strongly doped with antimony (Sb). The layers were prepared by spin coating of SnO2 :Sb colloids onto conductive substrates. These researchers found that the rate of the color change promoted by the Li+ intercalation was determined mainly by the substrate’s conductivity. Very fast coloration and decoloration were observed (<10 ms). Cummins et al. [161] built a dynamic color display based on a electrochromic cell, using two porous metal oxide thin films sandwiched between glass electrodes. The negative electrode was coated with a layer of TiO2 nanocrystals with viologen molecules anchored to the nanocrystal surface. Viologen molecules turn blue when injected with charge (e.g., Li+ ). The positive electrode consists of a nanocrystalline SnO2 :Sb film, which is linked to
phenothiazine molecules that turn red when oxidized. The prototype built by Cummins and co-workers showed that the color in the display could be switched on and off in less than 250 milliseconds. Even in ion-conducting polymer membranes used as electrolytes for lithium batteries, the nanocrystals can play an important role. Scrosati and co-workers [12] showed that ceramic nanoparticles can act as solid plasticizers for polyethylene oxide (PEO), an ion-conducting polymer membrane. They demonstrated Li+ ionic conductivities of around 10−4 S cm−1 at 50 C and 10−5 S cm−1 at 30 C in a PEO– LiClO4 mixture containing TiO2 and Al2 O3 powders with 5.8–13 nm particle sizes. This may lead to the development of polymer electrolytes having a true solid-state configuration combined with high conductivity, a high lithium ion transference number, and high interfacial stability with the lithium metal electrode. In fact, the characteristics of the PEO–nanoparticle electrolyte render this material an ideal candidate for use in rechargeable lithium batteries. In photovoltaic cells (i.e., devices that convert sunlight into electrical power), nanocrystals are responsible for a new and promising generation of devices [8, 162, 163]. The creation of these new devices is based on metal oxide nanocrystals, such as TiO2 , SnO2 , ZnO, and others, and on conducting polymers. Generally, the device is composed of a mesoporous oxide film and the pores are filled with a conducting medium (conducting polymers).
4. NANOCRYSTALLINE COMPOSITES Several metal or metal oxide nanocrystal applications or processing steps require high temperature environments. A typical example is materials for heterogeneous catalysis processes, which usually require temperatures of over 200 C coupled with high pressure [164]. Under such conditions, most nanocrystals are unstable, resulting in particle growth and thus losing their nanoscale dimensions. As discussed earlier, a good way to improve the thermal stability of nanocrystals and prevent particle growth is to use the supersaturated solid solution process. This process, however, has only been applied for tin oxide [154, 157, 165] and in some restricted metal systems [5, 155]. A satisfactory way of improving thermal stability against particle growth may be the development of nanocomposites (i.e., materials composed of two or more phases with nanometric scale sizes). These nanocomposites must display a special phase distribution, in which the dispersed active phase is isolated from the others, preventing particle growth by coagulation or coalescence at high temperature. This concept is based on the fact that the dispersed solid phase in a solid matrix grows preferentially by coalescence, since a growth process such as Ostwald ripening does not occur due to the low solubility of the dispersed phase in the solid matrix. Based on this concept, two types of nanocomposite microstructures can be proposed to prevent particle growth: core–shell (CS) nanocomposites and embedded dispersed phase (EDP) nanocomposites. Figure 7 illustrates the two types of nanocomposites proposed. These two types of nanocomposite are now discussed in detail.
546
Figure 7. Nanocomposite microstructures proposed to prevent particle growth: CS nanocomposites and EDP nanocomposites.
4.1. Synthetic Methods CS nanocomposites composed of an organic shell are basically nanocrystals protected by a surfactant layer that controls particle growth during synthesis and particle coagulation by the steric effect [29, 30]. This process is useful to improve the stability of the nanocrystals in suspension; however, the organic layer is ineffective at high temperature due to its low thermal stability. CS nanocomposites with inorganic shells can be divided in two subgroups (i.e., crystalline inorganic shells and amorphous inorganic shells). Good reviews about the CS nanocomposite with an inorganic layer are available in the literature [6, 166]. CS nanocomposites with amorphous inorganic shells are composed of metals or metal oxides nanocrystals (core) with an amorphous silica layer (shell) usually synthesized by the sol–gel process [6, 167, 168]. Silica shells have been used on gold nanoparticles to stabilize them and to modulate their optical properties. Liz-Marzán et al. [168] synthesized CS nanocomposites composed of gold nanoparticles and silica shell (these authors introduced the Au@SiO2 nomenclature to signify gold core and silica shell), using gold nanocrystals prepared by citrate reduction of HAuCl4 . The silica shell is prepared in two steps. First, the gold particles are coated with (3-aminopropyl)trimethoxysilane (primer layer). The organometallic compound becomes complexed to the gold surface (using the amine groups), replacing the adsorbed citrate ions. This primer layer prevents particle coagulation. The second step consists of adding sodium silicate to form a thin silica layer over the gold particle. The primer layer (formed in the first step) acts as an anchor upon which the silica layer is formed. Further growth of the silica layer can be achieved by transferring the gold-coated particles to an ethanol/water solution with a controlled amount of TEOS. Slow TEOS hydrolysis and a polycondensation reaction cause the silica layer to increase. This technique has been adapted successfully to process several CS nanocomposites such as CdS@SiO2 [169] and Ag@SiO2 [170]. Because the silica shell thickness can be controlled, these CS
Nanocrystals Assembled from the Bottom Up
nanocomposites can be used to process dense thin films with controlled intercore distances [171]. Lu et al. [167] used the sol–gel approach to coat iron oxide nanoparticles with a uniform shell of amorphous silica. The thickness of the silica layer was controlled within the range of 2–100 nm by changing the concentration of the sol–gel solution. In their study, TEOS was used as alkoxide precursor for the silica layer. A primer layer was not necessary to coat the iron oxide. Oldfield et al. [172] also reported the formation of a tin oxide crystalline layer over gold nanoparticles (Au@SnO2 ) without a primer layer. The crystalline shell was produced through heat treatment and the gold nanocrystals were coated by adding Na2 SnO3 · 3H2 O (sodium stanate) to a gold sol with high pH (10.5). Bedja and Kamat [173] successfully prepared a TiO2 layer on SnO2 nanocrystals (TiO2 @SnO2 using a sol–gel-like process. They reported on the improved photocatalytic efficiency of the TiO2 @SnO2 nanocomposite compared to pure SnO2 nanocrystals. As can be seen in Figure 7, the structure of EDP nanocomposites differs from that of CS nanocomposites. In this kind of nanocomposite, the dispersed phase (nanocrystals) is homogenously distributed in a matrix. In fact, the nanocrystals are embedded in a matrix that can be either porous or dense. Basically, EDP nanocomposites can be divided into two classes: EDP nanocomposites formed by an in-situ process and EDP nanocomposites formed by an ex-situ process. With the advent of mesoporous molecular sieves [174, 175], new possibilities for the development of EDP nanocomposites formed by ex-situ processing have arisen. These mesoporous sieves present a well-defined pore structure and large surface area, which makes them excellent matrixes for EDP nanocomposites. When these mesoporous materials are used as matrixes for EDP nanocomposites, a second processing step is needed to incorporate the dispersed phase (nanocrystals). This two-step process is called ex-situ. The dispersed phase is normally incorporated by liquid infiltration of metal salt solution with subsequent heat treatment in a hydrogen atmosphere, or by vapor phase [176]. The vacuum or electrophoretic method also can be used to fill the channels of mesoporous matrixes with colloidal suspension [177–179]. An alternative way to process EDP nanocomposites is through an in-situ process whereby the dispersed phase and the matrix are formed in a single step process. Rolison et al. [180–182] proposed an interesting synthetic route to process EDP nanocomposites by an in-situ process. They used silica sol (silica nanoparticles dispersed in a liquid) as a nanoscale glue to prepare composite aerogel by dispersing a nanometric solid together with the silica sol. With the condensation of the silica sol, the second phase was incorporated in a three-dimensional silica network. After gelation, the materials were dried by a supercritical process, keeping their three-dimensional structure and the dispersed phase (metal nanoparticles, carbon black, zeolites, or TiO2 embedded in a mesoporous matrix. Leite et al. [183, 184] recently described a new in-situ route for the synthesis of EDP nanocomposites. A mesoporous silica matrix, with a surface area of more than 200 m2 /g, a narrow pore size distribution, and transition
547
Nanocrystals Assembled from the Bottom Up
metal nanocrystals (Ni, Fe, Co, Ag) embedded in the matrix, was obtained in a direct process. Figure 8a shows a bright field (BF) transmission electron microscopy (TEM) image of Ni nanoparticles embedded in an amorphous mesoporous silica matrix processed by this method. A different approach was adopted to process such nanocomposites. This new approach is based on the formation of a hybrid polymer composed of Si, O, C, H, and metal cations arrested in the macromolecule chain and on the control of the pyrolysis step. The CO/CO2 atmosphere resulting from the pyrolysis of the organic material leads to the reduction of the metallic salt. In this process, CA is used to chelate the Si, resulting in a Si–citric acid complex. EG and the metallic salt are then added to the citrate solution to promote a polyesterification reaction, resulting in a polymeric precursor. The polymeric precursor is then pyrolyzed in an N2 atmosphere to promote the formation of the EDP nanocomposite. This process leads to the formation of nanocomposites with good catalytic [183, 185] and magnetic properties [186]. The same approach can also be used to process nanocomposites with matrixes other than silica, as illustrated in Figure 8b, which
Figure 8. (a) BF TEM image of Ni nanoparticles (dark spots) embedded in an amorphous mesoporous silica matrix; (b) BF-TEM image of a nanocomposite composed of Ni nanocrystals (dark spots) with a mean particle size of 8.8 nm embedded in a matrix composed of amorphous and crystalline (anatase) TiO2 phase (inset shows a HRTEM image of the nanocomposite, in which can be visualized the anatase, amorphous TiO2 , and the Ni nanoparticles).
shows a BF-TEM image of a nanocomposite composed of Ni nanocrystals with a mean particle size of 8.8 nm (see inset of Fig. 8b) embedded in a matrix composed of amorphous and crystalline (anatase phase) TiO2 . In this case, a polymeric precursor composed of Ti, O, C, and H was used. Chiodini [77] and co-workers reported that they had synthesized SnO2 nanocrystals embedded in a dense SiO2 matrix. A modified sol–gel process was used to prepare the nanocomposite. The main step of this process consists of obtaining a starting low-density skeleton from the hydrolysis of organometallic precursors (TEOS as SiO2 precursor and dibutyl tin diacetate as SnO2 precursor). With controlled hydrolysis of the precursors and thermal treatment in a controlled atmosphere up to 1050 C, an EDP nanocomposite is obtained, with SnO2 nanocrystals (2–8 nm) dispersed in a dense SiO2 matrix.
4.2. Properties The properties of nanocomposites materials are directly related to the combination of different phases, which potentially result in a material with superior performance. This synergetic effect should, in fact, be considered in the selection of material for nanocomposite processing. Mulvaney et al. [171] showed that the optical properties of Au nanoparticles can be modulated in a CS nanocomposite, whose shell consists of a silica layer. These authors found that the shell thickness controls the dipole coupling between particles and, hence, the color of the film composed of this CS nanocomposite [171]. The same group found [172] that CS nanocomposites composed of SnO2 shells and Au cores have more advantageous capacitor properties than the uncoated gold particle. This superior performance derives from the considerable difference between the Fermi level of the core and the shell. This CS nanocomposite is potentially applicable for electron storage and electronic circuit components. An interesting application for the SiO2 @Fe2 O3 nanocomposite was proposed by Levy and co-authors [187], who used this nanocomposite for targeted diagnostics and therapy. The main purpose of their development was to use the CS material for enhanced contrast magnetic resonance imaging and magnetic-induced cancer therapy. Nanocomposites also display excellent performance as catalytic material. Leite et al. [183] showed that EDP nanocomposites composed of Ni nanoparticles embedded in a mesoporous silica matrix possess good catalytic properties. This nanocomposite performed well in methanol decomposition and conversion to H2 and CO. A high methanol conversion was observed (100%), as well as high selectivity for H2 formation at a temperature of 263 C, with 53 mg of catalyst mass, a total flow (Ar + CH3 OH) of 30 cm2 /min, and a weight hourly space velocity of 34,000 ml/h g. The same nanocomposite showed excellent catalytic performance for -pinene hydrogenation. This catalytic material displayed high selectivity in relation to the -pinene hydrogenation, with preferential formation of cis-pinane. The same Ni–SiO2 EDP nanocomposite showed interesting results as a catalyst for the conversion of methane into synthesis gas [185]. Carreño and co-workers reported that this catalyst presents singular properties to control
548 the carbon deposition and deactivation of active sites. A comparative study of EDP nanocomposites and conventional catalysts (prepared by impregnation) revealed that impregnated material showed preferential encapsulation and growth of carbon nanotubes on the metal surface. The impregnated catalyst displayed a higher tendency to form carbon nanotubes and whiskers. This tendency was not observed in the EDP nanocomposite. Rolison et al. [182] have developed a highly active electrocatalytic EPD nanocomposite composed of Pt nanoparticles dispersed in a carbon–silica composite aerogel. They reported electrocatalytic activity for methanol oxidation four orders of magnitude per gram of Pt higher than native Pt-modified carbon powder. EDP nanocomposites can also be used to process photonic materials for telecommunications. Chiodini et al. [77] developed an EDP nanocomposite with SnO2 nanoparticles embedded in a dense SiO2 matrix. The SnO2 nanoparticles acted as wide-bandgap (>4 ev) quantum dots. In this system, unlike other semiconductor-doped glasses, both the glassy host and the nanophase are oxides of IV-group elements. Due to their thermochemical compatibility, the two phases give stable optical-grade glass ceramics with potential photonic applications.
5. CHARACTERIZATION OF NANOCRYSTALS Because of the finite size of nanocrystals, their surface structure and composition are easily affected, resulting in materials with different chemical and physical properties. Thus the characterization of nanostructured materials is a key step in the development of nanocrystals. This characterization should, basically, answer the following questions: (a) What is the material’s particle size and particle size distribution? (b) What is its composition? (c) What is its structure and how does its particle size affect this structure? The answers to these questions therefore require the use of an arsenal of techniques that allow for characterizations of a morphological, chemical, and structural nature. However, the greatest challenge involves determining the superficial structure of nanocrystals and how the atoms located on their surface change their position (surface reconstruction) to accommodate high surface areas in finite sizes and the interaction of nanocrystals with additives such as surfactants. This section describes characterization techniques that provide answers to the above questions, particularly techniques that can provide information on the surface structure. The first technique described here is TEM and scanning electron microscopy (SEM). These are quite straightforward techniques to determine the size and shape of nanostructured materials, as well as to obtain chemical and structural information. TEM uses transmitted and diffracted electrons, which generate a two-dimensional projection of the sample. The principal contrast in this projection (or image) is provided by diffracted electrons. In the so-called bright-field images, the transmitted electrons generate bright regions while the diffracted electrons produce dark regions. In the so-called dark field image, the diffracted electrons preferentially form
Nanocrystals Assembled from the Bottom Up
the image. In this case, the bright areas in the dark field image represent the regions that produced the diffraction. This technique is useful to build the particle size distribution based on a sample composed of an amorphous phase and crystalline phases, as in EDP nanocomposites. Figure 9a shows a dark field TEM image of an EDP nanocomposite consisting of Ni nanoparticles embedded in an amorphous silica matrix. The bright points represent the Ni nanoparticles. The particle size distribution can be built from the TEM image, using an image analyzing software program that allows one to define the particle area of several particles [several particles (total number of particles >300) are normally necessary to obtain a good statistic] and to determine an equivalent particle size based on these areas. Figure 9b shows an example of particle size distribution of Ni nanoparticles in a Ni–SiO2 EDP nanocomposite built from dark field TEM images. In addition to morphological information, TEM provides crystallographic and chemical information. The crystallographic information can be obtained from the diffraction pattern of a selected area, using the so-called select area electron diffraction (SAED) technique. This technique allows for the indexation of crystallographic structures of small or even nanometric regions in the samples. Besides SAED, HRTEM can generate lattice images of the crystalline material, allowing for the direct characterization of a sample’s atomic structure. Figures 4b and 5 contain examples of HRTEM images. HRTEM images are particularly sensitive to changes in a phase of the incoming parallel electron wave as it passes through the sample. The atomic structure can also be characterized by scanning transmission electron microscopy (STEM), using a high angle annular dark field detector (HAADF). In fact, the image generated by a HAADF detector, which is sensitive to the atomic number (Z), is also known as a Z-contrast image [188]. The high-resolution image generated by HAADF detector is sensitive to the chemical composition. In TEM, the chemical information is provided by the X-rays generated through the interaction of the electronic
Figure 9. (a) Dark field TEM image of an EDP nanocomposite consisting of Ni nanoparticles embedded in an amorphous silica matrix. The bright points represent the Ni nanoparticles. (b) Particle size distribution of Ni nanoparticles in a Ni–SiO2 EDP nanocomposite built from dark field TEM images.
549
Nanocrystals Assembled from the Bottom Up
beam with the sample. This technique, known as energy dispersive X-ray spectroscopy (EDS), allows regions as small as 2 nm in a TEM configuration or lower than 0.8 nm to be chemically characterized, using STEM equipment. In addition to EDS analyses, nanostructured materials can also be chemically characterized by electron energy loss spectroscopy (EELS), which allows for a better characterization of light elements, which usually pose a major problem in X-ray spectroscopy-based characterization. Another advantage of EELS spectroscopy is the fact that it is sensitive to the element’s chemical environment, allowing information to be obtained on the local structure of the analyzed element. EDS analysis is insensitive to the chemical environment. Detailed information about TEM characterization is given in the book by William and Carter [189]. TEM characterization as been used more recently for in-situ studies of materials in the nanoscale range [190, 191]. Using in-situ TEM characterization to obtain atom-resolved images of copper nanocrystals on different supports, Hansen and co-workers [190] found that Cu nanocrystals undergo dynamic reversible shape changes in response to changes in the gaseous environment. SEM is carried out by scanning an electron beam over the sample’s surface and detecting the yield of low energy electrons (secondary electrons) and high-energy electrons (backscattered electrons) according to the position of the primary beam. The secondary electrons, which are responsible for the topologic contrast, provide mainly information about the surface morphology. The backscattered electrons, which are responsible for the atomic number contrast (Z-contrast), carry information on the sample’s composition. Regarding this type of contrast, because the heavy elements (high Z value) have the ability to scatter the electrons more efficiently, an image composed of bright areas (consisting of high Z elements) and dark ones (consisting of low Z elements) is generated. A new generation of SEM has recently emerged as an important tool to characterize nanostructured materials. This state-of-the-art equipment, called a field emission gun (FEG)-SEM, has a FEG that provides the electron beam, and it uses in-lens or semi-in-lens secondary electron detectors to obtain images with a low acceleration voltage and a resolution as high as 1 nm. A STEM detector can be used in the FEG-SEM device, enabling one to obtain the transmitted dark and bright field images of thin specimens or nanoparticles, similar to images obtained with a dedicated STEM. Figure 10 shows images of SnO2 nanoparticles, with rather good bright and dark field images, which were taken with a FEG-SEM instrument, using a STEM detector. SEM can provide not only morphological but also chemical information about the sample; however, the spatial resolution is only about 1 m3 . This limitation has rendered the chemical characterization of nanostructured materials using SEM practically useless. The second characterization technique to be analyzed is X-ray diffraction (XRD). XRD is the most popular technique for structural characterization, providing information about the phases present in the material, as well as refined crystallographic parameters and morphological information. XRD consists of the elastic scattering of X-ray photons by
Figure 10. STEM images of SnO2 nanoparticles, which were taken with a FEG-SEM instrument, using a STEM detector. (a) Bright field image; (b) dark field image.
atoms in a periodic lattice. The scattered X-rays (monochromatic) that are in phases provide constructive interference. The lattice space (dhkl , or the distance between two lattice planes in a crystal, is described by the Bragg relation 2 dhkl sin = n
(11)
where is the X-ray wavelength, is the diffracted angle, and n is the integer called order of the reflection. A dhkl set is characteristic of a given compound. Powder XRD has been used extensively to characterize nanoparticles; however, due to the small particle size, broadening of the diffraction lines occurs at the usual X-ray wavelengths, rendering analyses more complex. These analyses are more difficult if structural parameters are required. In such cases, special X-ray surges like synchrotron radiation and special fit analyses are required to achieve a good structural characterization [192]. X-ray line broadening analysis provides a fast and usually reliable estimate of the particle size [153, 154, 183, 184, 193]. The particle size or, more precisely, the crystallite size (Dhkl ), can be calculated based on Scherer’s equation, Dhkl = K / cos
(12)
where K is a constant (0 9 ≤ K ≤ 1 0), and is the corrected full width at half maximum of the diffraction peak. This method of determining the crystallite size takes into account only the X-ray line broadening promoted by the
550
Raman Intensity (counts/s)
0,005
a)
SnO 2 5% La
A1g
0,004
Bands associated to superficial disorder
Eg
0,003
0,002
0,001
0,000 400
500
600
700
800
900
-1
Raman shift (cm )
3,0
Raman Intensity (counts/s)
size effect. Other effects, such as the broadening induced by lattice strain, are not considered. However, considerable discrepancies can occur in metallic nanocrystalline systems [194], even in metal oxide nanoparticles [154]. In such situations, correction of the strain broadening is indispensable for a reliable characterization of the crystallite size. XRD refinement methods, such as the Rietveld method, can be used to correct the strain broadening [195]. Electron microscopy and XRD can provide good morphological, chemical, and structural characterization of nanocrystals; however, these techniques do not provide information about the disordered part of nanocrystals, such as their surfaces. An alternative technique to obtain information about structural disorder in nanocrystals, including surface information of nanocrystals, is X-ray absorption spectroscopy (XAS). This technique is based on the absorption of X-rays and the creation of photoelectrons that are scattered by nearby atoms in a structure. This technique provides information about local structures, such as distance and the number and type of neighbors of the absorbing atom, and can be used to characterize the bulk and surface of solid-state materials, including amorphous materials and catalysts [196, 197]. XAS techniques [X-ray absorption near-edge spectroscopy and extended X-ray absorption fine structure (EXAFS)] have been used to study the size dependence of structural disorder in semiconductors and metal oxide nanocrystals. The results of these techniques have demonstrated that this spectroscopic analysis is sensitive to the structure of the entire nanocrystal, including the surface [198, 199]. By EXAFS analyses of CdS nanoparticles (1.3–12 nm diameter), Rockenberger et al. [200] showed that the stabilization of nanoparticles influences the mean Cd–S distance. Thiol-capped nanoparticles display an expansion of their mean interatomic Cd–S distance, whereas polyphosphate-stabilized particles contract in terms of their CdS bulk. EXAFS has also proved to be highly sensitive in the identification of the surface segregation of foreign cations [201, 202]. Raman spectroscopy can also provide information about the overall nanocrystal structure. Raman spectroscopy is based on the inelastic scattering of photons, which lose energy by exciting vibration in the sample. This technique is very sensitive to the crystal’s symmetry and any change can cause modifications in the spectrum, such as shift of the line position, broadening of the line, and new lines. The crystal symmetry of the nanocrystal surface must present differences in relation to the symmetry of the nanocrystal core. Hence, the Raman spectrum of a nanostructured material must be different from the Raman spectrum of the same material with bulk characteristics. In fact, several reports have described the influence of particle size on the Raman spectrum [203–206]. Figure 11a shows Raman spectra of a SnO2 nanocrystal processed by the IPC process (particle size ∼9 nm) compared with a spectrum of a SnO2 single crystal (Fig. 11b). In addition to the classic modes of the rutilelike structure, the figure clearly shows bands associated with superficial disorder. A detailed discussion of the influence of particle size on the Raman spectrum of SnO2 is given in the literature [205]. In that work, Dieguez and co-workers proposed a model in which the additional band is due to a
Nanocrystals Assembled from the Bottom Up
A1g
b)
2,5
2,0
B2g
1,5
Eg
1,0
0,5 200
300
400
500
600
700
800
900
-1
Raman Shift (cm )
Figure 11. Raman spectra of a SnO2 : (a) nanocrystal processed by the IPC process (particle size ∼9 nm); (b) SnO2 single crystal spectrum.
surface layer of SnO2 with a dissimilar symmetry. The thickness of this layer was calculated at ∼1.1 nm. Analyses of the Raman line can provide important information about lattice disorder. Such analyses can be performed using the spatial correlation model described by Parayanthal and Pollak [207]. Basically, this model describes the crystal’s quality by introducing a parameter called correlation length (L), which can be interpreted as the average size of the region of homogeneous materials. According to this model, the Raman line intensity [I] can be written as I =
1 0
exp −q2L2/4 d 3 q/ − q2 + 0 /22 (13)
where is the frequency, q is the wave vector, 0 is the halfwidth of the Raman line, and q is the Raman phonon dispersion. Based on the spatial correlation model, Kosacki et al. [208] demonstrated that the Raman spectra of CeO2 and Y-doped ZrO2 thin film are influenced by defects attributed to grain size-controlled nonstoichiometry. Indeed, they proved that Raman spectroscopy could be an effective method to determine the concentration of oxygen vacancies in fluorite-structured oxides.
6. SUMMARY An analysis of the various synthesization routes described above leads to the conclusion that the development of new synthesization routes or the improvement of routes currently under study are necessary to achieve greater control over
551
Nanocrystals Assembled from the Bottom Up
the synthesis of nanocrystals, particularly of metal oxide nanocrystals. The advances achieved over the last year in expanding our understanding of the synthesis and control of metal nanocrystals have led to the development of nanoparticles with strictly controlled particle size and particle size distribution, and good control of particle shape. Part of this progress can be attributed to the application of concepts used in the synthesis of semiconductor nanocrystals, such as the thermal decomposition of organometallic compounds and the use of surfactants for the control of particle growth and shape. The same progress, however, has not been achieved in the synthesis of metal oxide nanocrystals, particularly of metal oxides composed of two or more cations. Part of the reason for this lack of progress is the fact that, during the process of hydrolysis and precipitation, a hydrated amorphous metal oxide phase is formed, requiring heat treatment or a hydrothermal process to promote crystallization of the desired phase. This crystallization step may lead to particle growth, particle shape modification, and particle agglomeration. The use of dopants and surface modifying agents during the synthesis of metal oxide nanoparticles may offer a good alternative to increase the thermal stability of nanocrystals, thus preventing particle growth. The use of nanocomposites to prevent the growth of nanocrystals, particularly in the case of materials for use at high temperature, appears to be a good strategy. This approach can be combined with the synergism between phases to design nanocomposite materials with superior performance. There is no doubt that considerable progress has been made in nanocrystal synthesis over the last few years. However, further studies are necessary to achieve improved control over particle size, particle shape, and particle size distribution, as well as to obtain nanoparticles with greater stability against particle growth at high temperatures.
GLOSSARY Nanocomposite Nanostructured material formed by two or more phases with nanometric scale (1–100 nm). Nanocrystal Crystalline particle with size in the 1–100 nm range. Particle growth Physical–chemical process promoted by the high superficial energy of the particles. Self-assembled Controlled (ordered) particle deposition governed by particle–particle and particle–substrate interaction. Sol–gel Chemical synthesis process in which a sol (often defined as a suspension of small particles or molecules in a liquid phase) undergoes a transition to a gel (a threedimensional network structure spreading throughout the liquid medium).
ACKNOWLEDGMENTS The author thanks many colleagues for stimulating discussions and collaboration, especially Elson Longo, J. A. Varela, Carlos A. Paskocimas, Emersom R. Camargo, Fenelon M. Pontes, N. L. V. Carreño, and Adeilton
P. Maciel. The following Brazilian agencies have provided support for my work: FAPESP, CNPq, and CAPES.
REFERENCES 1. L. V. Interrante and M. J. Hampden-Smith, “Chemistry of Advanced Materials.” Wiley–VCH, New York, 1998. 2. A. P. Alivisatos, J. Phys. Chem. 100, 13226 (1996). 3. A. P. Alivisatos, Science 271, 933 (1996). 4. S. Morup, “Nanomagnetism” (A. Hernando, Ed.), p. 93. Kluwer Academic, Boston, 1993; K. O’Grady and R. W. Chantrell, “Magnetic Properties of Fine Particles,” p. 93. Elsevier, Amsterdam, 1992. 5. H. Gleiter, Acta Mater. 48, 1 (2000). 6. P. Mulvaney, Mater. Res. Soc. Bull. 12, 1009 (2001). 7. M. Faraday, Philos. Trans. Roy. Soc. London 147, 145 (1857). 8. B. O’Regan and M. Grätzel, Nature 353, 737 (1991). 9. W. Li, H. Osora, L. Otero, D. C. Duncan, and M. A. Fox, J. Phys. Chem. A 102, 5333 (1998). 10. I. Bedja, P. V. Kamat, A. G. Lapin, and S. Hotchandani, Langmuir 13, 2398 (1997). 11. L. Kavan, K. Kratochvilova, and M. Grätzel, J. Electroanal. Chem. 394, 93 (1995). 12. F. Crou, G. B. Appetechi, L. Persi, and B. Scrosati, Nature 394, 456 (1998). 13. A. Hagfeldt and M. Grätzel, Chem. Rev. 95, 49 (1995). 14. N. Yamazoe, Sensors Actuators B 5, 7 (1991). 15. D. Weller and A. Moser, IEEE Trans. Magn. 35, 4423 (1999). 16. S. Sun and D. Weller, J. Magn. Soc. Jpn. 25, 1434 (2001). 17. J. Daí, J. Tang, S. T. Hsu, and W. Pan, J. Nanosci. Nanotech. 2, 281 (2002). 18. C. Wei, D. Grouquist, and J. Roark, J. Nanosci. Nanotech. 2, 47 (2002). 19. V. L. Colvin, M. C. Schlamp, and A. P. Alivisatos, Nature 370, 354 (1994). 20. C. Feldamann, Adv. Mater. 17, 1301 (2001). 21. H. Bakker, G. F. Zhou, and H. Yang, Progr. Mater. Sci. 39, 159 (1995). 22. E. R. Leite, L. P. S. Santos, N. L. V. Carreño, E. Longo, C. A. Paskocimas, J. A. Varela, F. Lanciotti, C. E. M. Campos, and P. S. Pizani, Appl. Phys. Lett. 78, 2148 (2001); T. Trindade, P. O’Brien, and N. L. Pickett, Chem. Mater. 13, 3843 (2001). 23. J. R. Blackborrow and D. Young, “Metal Vapor Synthesis.” SpringVerlag, New York, 1979. 24. J. Turkevich, P. C. Stevenson, and J. Hillier, Disc. Faraday Soc. 11, 55 (1951). 25. J. Turkevich and G. Kim, Science 169, 873 (1970). 26. J. Turkevich, Gold Bull. 18, 86 (1985). 27. H. Bönnemann and R. M. Richards, Eur. J. Inorg. Chem. 2455 (2001). 28. R. J. Pugh, in “Surface and Colloid Chemistry in Advanced Ceramics Processing” (R. J. Pugh and L. Bergström, Eds.). Dekker, New York, 1994. 29. E. Bourgeat-Lami, J. Nanosci. Nanotech. 1, 1 (2002). 30. C. Sanchez, G. J. de, A. A. Soller-Illia, F. Ribot, T. Lalot, C. R. Mayer, and V. Cabuil, Chem. Mater. 13, 3061 (2001). 31. J. S. Bradley, in “Clusters and Colloids” (G. Schmid, Ed.). VCH, Weinheim, 1994. 32. J. S. Brandley, J. M. Millar, and E. W. Hill, J. Am. Chem. Soc. 113, 4016 (1991). 33. T. Teranishi and M. Miyake, Chem. Mater. 10, 594 (1998). 34. M. A. Hayat, “Colloidal Gold: Methods and Applications.” Academic, San Diego, 1989. 35. C. D. Keating, M. D. Musik, M. H. Keefe, and M. J. Natan, J. Chem. Education 76, 949 (1999).
552 36. D. N. Furlong, A. Launikonis, W. H. F. Sasse, and J. V. Saunders, J. Chem. Soc., Faraday Trans. 80, 571 (1988). 37. P. R. van Rheenen, M. J. Mckelvey, and W. S. Glaunsinger, J. Solid State Chem. 67, 151 (1987). 38. T. Ould Ely, C. Amiens, B. Chaudret, E. Snoeck, M. Verelst, M. Respaud, and J. M. Broto, Chem. Mater. 11, 526 (1999). 39. F. Dessenoy, K. Philippot, T. Ould Ely, C. Amiens, and B. Chaudret, New J. Chem. 22, 703 (1998). 40. J. Osuma, D. de Caro, C. Amiens, B. Chaudret, E. Snoeck, M. Respaud, J. M. Broto, and A. Fert, J. Phys. Chem. 100, 14571 (1996). 41. A. Rodriguez, C. Amiens, B. Chaudret, M. J. Casanove, P. Lecante, and J. S. Bradley, Chem. Mater. 8, 1978 (1996). 42. A. Duteil, R. Queau, B. Chaudret, R. Mazel, C. Roucau, and J. S. Bradley, Chem. Mater. 5, 341 (1993). 43. T. Ould Ely, C. Pan, C. Amiens, B. Chaudret, F. Dassenoy, P. Lecante, M. J. Casanove, A. Mosset, M. Respaud, and J. M. Broto, J. Phys. Chem. B 104, 695 (2000). 44. F. Fiévet, J. P. Lagier, and M. Figlarz, Mater. Res. Soc. Bull. 12, 29 (1989). 45. G. Viau, F. Ravel, O. Archer, F. Fiévet-Vicente, and F. Fiévet, J. Magn. Magn. Mater. 377, 140 (1995). 46. C. B. Murray, S. Sun, H. Doyle, and T. Betley, Mater. Res. Soc. Bull. 12, 985 (2001). 47. J.-I. Park and J. Cheon, J. Am. Chem. Soc. 123, 5743 (2001). 48. D. P. Dinega and M. G. Bawendi, Angew. Chem. Int. Ed. 38, 1788 (1999). 49. J. R. Thomas, J. Appl. Phys. 37, 2914 (1966). 50. S.-J. Park, S. Kim, S. Lee, Z. G. Khim, K. Char, and T. Hyeon, J. Am. Chem. Soc. 122, 8581 (2000). 51. V. F. Puntes, K. M. Krishnan, and A. P. Alivisatos, Science 291, 2115 (2001). 52. X. Peng, L. Manna, W. Yang, J. Wickham, E. Scher, A. Kadavanich, and A. P. Alivisatos, Nature 404, 665 (2000). 53. S. Sun, C. B. Murray, D. Weller, L. Folks, and A. Moser, Science 287, 1989 (2000). 54. M. Chen and D. E. Nikles, Nano Lett. 3, 211 (2002). 55. K. S. Suslick, Science 247, 1439 (1990). 56. E. B. Flint and K. S. Suslick, Science 253, 1397 (1991). 57. K. S. Suslik, M. Fang, and T. Hyeon, J. Am. Chem. Soc. 118, 11960 (1996). 58. S. R. Whaley, D. S. English, E. L. Hu, P. F. Barbara, and A. M. Belcher, Nature 405, 665 (2000). 59. L. O. Brown and J. E. Hutchison, J. Phys. Chem. B 105, 8911 (2001). 60. M. Chen and D. E. Nikles, Nano Lett. 3, 211 (2002). 61. S. Sun, S. Anders, H. F. Hamann, J.-U. Thiele, J. E. E. Baglin, T. Thomson, E. E. Fullerton, C. B. Murray, and B. D. Terris, J. Am. Chem. Soc. 124, 2884 (2002). 62. X. Peng, J. Wickham, and A. P. Alivisatos, J. Am. Chem. Soc. 120, 5343 (1998). 63. V. F. Puntes, K. M. Krishnan, and A. P. Alivisatos, Science 291, 2115 (2001). 64. R. A. Ristau, K. Barmak, L. H. Lewis, K. R. Coffey, and J. K. Howard, J. Appl. Phys. 86, 4527 (1999). 65. J. B. Michel and J. T. Schwartz, in “Catalyst Preparation Science, IV” (B. Delmon, P. Grange, P. A. Jacobs, and G. Poncelet, Eds.), p. 669. Elsevier Science, New York, 1987. 66. H. Hirai, Y. Nakao, N. Thosima, and K. Adachi, Chem. Lett. 905 (1976). 67. M. Boutonnet, J. Kizling, P. Stenius, and G. Maire, Colloids Surf. 5, 209 (1982). 68. M. Boutonnet, J. Kizling, R. Touroude, G. Maire, and P. Stenius, Appl. Catal. 20, 163 (1986). 69. P. J. F. Harris, “Carbon Nanotubes and Related Structures,” p. 16. Cambridge Univ. Press, Cambridge, UK, 1999.
Nanocrystals Assembled from the Bottom Up 70. A. Cassel, J. Raymakers, J. Kong, and H. Dai, J. Phys. Chem. 103, 6484 (1999). 71. J. Kong, H. Soh, A. Cassel, C. Quate, and H. Dai, Nature 395, 878 (1998). 72. Y. Zhang, Y. Li, W. Kim, D. Wang, and H. Dai, Appl. Phys. A 74, 325 (2002). 73. Z. P. Huang, D. Z. Wand, J. G. Wen, M. Sennett, G. Gibson, and Z. F. Ren, Appl. Phys. A 74, 387 (2002). 74. Z. W. Pan, S. S. Xie, B. H. Chang, C. Y. Wang, L. Lu, W. Liu, W. Y. Zhou, W. Z. Li, and L. X. Qian, Nature 394, 631 (1998). 75. E. M. Wong, J. E. Bonevich, and P. C. Searson, J. Phys. Chem. B 102, 7770 (1998). 76. A. L. Roest, J. J. Kelly, and D. Vanmaekelbergh, Phys. Rev. Lett. 89, 36801 (2002). 77. N. Chiodini, A. Paleari, D. DiMartino, and G. Spinolo, Appl. Phys. Lett. 81, 1702 (2002). 78. H. Weller, Adv. Mater. 5, 88 (1993). 79. K. Uchino, E. Sadanaga, K. Oonishi, T. Morohashi, and H. Yamamura, in “Ceramic Dielectrics: Composition, Processing, and Properties,” Ceramic Transactions, Vol. 8. American Ceramic Society, Westerville, OH, 1990. 80. M. H. Frey and D. A. Payne, Phys. Rev. B 54, 3158 (1996). 81. T. Maruyuma, M. Saitoh, I. Skai, and T. Hidaka, Appl. Phys. Lett. 73, 3524 (1998). 82. M. Sternitzke, J. Eur. Ceram. Soc. 17, 1061 (1997). 83. Y.-M. Chiang, E. B. Lavik, I. Kosacki, H. L. Tuller, and J. Y. Ying, J. Electroceram. 1, 7 (1997). 84. C. J. Brinker and G. W. Scherrer, “Sol–Gel Science.” Academic Press, Boston, 1990. 85. J. Y. Ying, Special Issue: Sol–Gel Derived Materials, in Chem. Mater. 9, 2247 (1997). 86. C. Sanchez, J. Livage, M. Henry, and F. Babonneau, J. Non-Cryst. Solids 100, 65 (1988). 87. V. K. LaMer and R. H. Dinegar, J. Am. Chem. Soc. 72, 4847 (1950). 88. W. Stöber, A. Fink, and E. Bohn, J. Colloid Interface Sci. 26, 62 (1968). 89. E. A. Barringer and H. K. Bowen, Langmuir 1, 414 (1985). 90. M. T. Harris and C. H. Byers, J. Non. Cryst. Solids 103, 49 (1988). 91. T. Sato and R. Ruch, in “Stabilization of Colloidal Dispersion by Polymer Adsorption.” Dekker, New York, 1980. 92. D. H. Napper, “Polymer Stabilization of Colloidal Dispersion.” Academic Press, New York, 1983. 93. T. E. Mates and T. A. Ring, Colloids Surf. 24, 299 (1987). 94. J. H. Jean and T. A. Ring, Langmuir 2, 251 (1986). 95. J. L. Look and C. F. Zukoski, J. Am. Ceram. Soc. 75, 1587 (1992). 96. J. L. Deiss, P. Anizan, S. El Hadigui, and C. Wecker, Colloids Surf. A 106, 59 (1996). 97. A. M. Peiró, J. Peral, C. Domingo, X. Domènech, and J. A. Ayllón, Chem. Mater. 13, 2567 (2001). 98. R. W. Schawartz, Chem. Mater. 9, 2325 (1997). 99. C. Liu, B. Zou, A. J. Rondinone, and Z. J. Zhang, J. Am. Chem. Soc. 123, 4344 (2001). 100. S. O’Brien, L. Brus, and C. B. Murray, J. Am. Chem. Soc. 123, 12085 (2001). 101. J. J. Urban, W. S. Yun, Q. Gu, and H. Park, J. Am. Chem. Soc. 124, 1186 (2002). 102. R. M. Barrer, “Hydrothermal Chemistry of Zeolites.” Academic Press, London, 1982. 103. S. Mintova, N. H. Olson, V. Valtchev, and T. Bein, Science 283, 958 (1999). 104. C.-C. Wang and J. Y. Ying, Chem. Mater. 11, 3113 (1999). 105. J. Ragai and W. Lotfi, Colloid Surf. 61, 97 (1991). 106. L. Cao, H. Wan, L. Huo, and S. Xi, J. Colloid Interface Sci. 244, 97 (2001). 107. T. Nütz and M. Haase, J. Phys. Chem. B 104, 8430 (2000).
Nanocrystals Assembled from the Bottom Up 108. G. Pang, S. Chen, Y. Koltypin, A. Zaban, S. Feng, and A. Gedanken, Nano Lett. 1, 723 (2001). 109. R. S. Hiratsuka, S. H. Pulcinelli, and C. V. Santilli, J. Non-Cryst. Solids 121, 76 (1990). 110. C. Goebbert, R. Nonninger, M. A. Aegerter, and H. Schmidt, Thin Solid Films 351, 79 (1999). 111. A. P. Rizzato, L. Broussous, C. V. Santilli, S. H. Pulcinelli, and A. F. Craievich, J. Non-Cryst. Solids 284, 61 (2001). 112. E. A. Meulenkamp, J. Phys. Chem. B 102, 5566 (1998). 113. E. M. Wong, J. E. Bonevich, and P. C. Searson, J. Phys. Chem. B 102, 7770 (1998). 114. M. Z. C. Hu, R. D. Hunt, E. A. Payzant, and C. R. Hubbard, J. Am. Ceram. Soc. 82, 2313 (1999). 115. G. Pang, S. Chen, Y. Zhu, O. Palchik, Y. Koltypin, A. Zaban, and A. Gedanken, J. Phys. Chem. B 105, 4647 (2001). 116. M. D. Fokema, E. Chiu, and J. Y. Ying, Langmuir 16, 3154 (2000). 117. E. R. Leite, unpublished work, 2002. 118. D. W. Bahnemann, C. Kormann, and M. R. Hoffmann, J. Phys. Chem. 91, 3789 (1987). 119. L. Spanhel and M. A. Anderson, J. Am. Chem. Soc. 113, 2826 (1991). 120. N.-L. Wu, S.-Y. Wang, and I. A. Rusakova, Science 285, 1375 (1999). 121. A. Vioux, Chem. Mater. 9, 2292 (1997). 122. T. J. Trentler, T. E. Denler, J. F. Bertone, A. Agrawal, and V. L. Colvin, J. Am. Chem. Soc. 121, 1613 (1999). 123. J. Rockenberger, E. C. Scher, and A. P. Alivisatos, J. Am. Chem. Soc. 121, 11595 (1999). 124. E. R. Camargo and M. Kakihana, Chem. Mater. 13, 1181 (2001). 125. E. R. Camargo, J. Frantti, and M. Kakihana, J. Mater. Chem. 11, 1875 (2001). 126. E. R. Camargo, J. Frantti, and M. Kakihana, Chem. Mater. 13, 3943 (2001). 127. E. R. Camargo and M. Kakihana, J. Am. Ceram. Soc. 85, 2107 (2002). 128. M. P. Pechini, U. S. Patent 3, 330, 697, 1967. 129. M. Kakihana, J. Sol–Gel Sci. Technol. 6, 7 (1996). 130. P. A. Lessing, Am. Ceram. Soc. Bull. 168, 1002 (1989). 131. S. G. Cho, P. F. Johnson, and R. A. Condrate, Jr., J. Mater. Sci. 25, 4738 (1990). 132. E. R. Leite, C. M. G. Sousa, E. Longo, and J. A. Varela, Ceram. Int. 21, 143 (1995). 133. E. R. Leite, C. A. Paskocimas, E. Longo, and J. A. Varela, Ceram. Int. 21, 153 (1995). 134. M. Kakihana, T. Okubo, Y. Nakamura, M. Yashima, and M. Yoshimura, J. Sol–Gel Sci. Technol. 12, 95 (1996). 135. S. Kumar, G. L. Messing, and W. White, J. Am. Ceram. Soc. 76, 617 (1993). 136. M. Kakihana, T. Okubo, M. Arima, O. Uchiyama, M. Yashima, M. Yoshimura, and Y. Nakamura, Chem. Mater. 9, 451 (1997). 137. M. Cerqueira, R. S. Nasar, E. Longo, E. R. Leite, and J. A. Varela, Mater. Lett. 22, 181 (1995). 138. H. Takahashi, M. Kakihana, Y. Yamashita, K. Yoshida, S. Ikeda, M. Hara, and K. Domen, J. Alloys Compounds 285, 77 (1999). 139. M. A. L. Nobre, E. Longo, E. R. Leite, and J. A. Varela, Mater. Lett. 28, 215 (1996). 140. E. R. Camargo, E. Longo, and E. R. Leite, J. Sol–Gel Sci. Technol. 17, 111 (2000). 141. M. Kakihana and M. Yoshimura, Bull. Chem. Soc. Jpn. 72, 1427 (1999). 142. A. L. Quinelato, E. Longo, E. R. Leite, and J. A. Varela, Appl. Organomet. Chem. 13, 501 (1999). 143. S. M. Zaneti, E. R. Leite, E. Longo, and J. A. Varela, Appl. Organomet. Chem. 13, 373 (1999). 144. S. M. Zaneti, E. R. Leite, E. Longo, and J. A. Varela, J. Mater. Res. 13, 2932 (1998).
553 145. V. Bouquet, M. B. I. Bernaardi, S. M. Zaneti, E. R. Leite, E. Longo, J. A. Varela, M. G. Viry, and A. Perrin, J. Mater. Res. 15, 2446 (2000). 146. F. M. Pontes, E. R. Leite, E. Longo, J. A. Varela, E. B. Araujo, and J. A. Eiras, Appl. Phys. Lett. 76, 2433 (2000). 147. F. M. Pontes, E. R. Leite, E. J. H. Lee, E. Longo, and J. A.Varela, Thin Solid Films 385, 260 (2001). 148. F. M. Pontes, J. H. G. Rangel, E. R. Leite, E. Longo, J. A. Varela, E. B. Araujo, and J. A. Eiras, Thin Solid Films 366, 232 (2000). 149. E. R. Leite, E. C. Paris, E. Longo, and J. A. Varela, J. Am. Ceram. Soc. 83, 1539 (2000). 150. E. R. Leite, E. C. Paris, E. Longo, F. Lanciotti, C. E. M. Campos, P. S. Pizani, V. Mastellaro, C. A. Paskocimas, and J. A. Varela, J. Am. Ceram. Soc. 85,2166 (2002). 151. E. R. Leite, M. A. L. Nobre, M. Cerqueira, E. Longo, and J. A. Varela, J. Am. Ceram. Soc. 80, 2649 (1997). 152. C. Greskovich and K. W. Lay, J. Am. Ceram. Soc. 55, 142 (1972). 153. E. R. Leite, I. T. Weber, E. Longo, and J. A. Varela, Adv. Mater. 12, 965 (2000). 154. E. R. Leite, A. P. Maciel, I. T. Weber, P. N. Lisboa-Filho, E. Longo, C. O. Paiva-Santos, A. V. C. Andrade, C. A. Paskocimas, Y. Maniette, and W. H. Schreiner, Adv. Mater. 14, 905 (2002). 155. J. Weissmuller, J. Mater. Res. 9, 4 (1994). 156. H. Onaga, M. Nishikawa, and A. Abe, J. Appl. Phys. 53, 4448 (1982). 157. I. T. Weber, R. Andrade, E. R. Leite, and E. Longo, Sensors Actuators B 72, 180 (2001). 158. N. L. Carreño, A. P. Maciel, E. R. Leite, P. N. Lisboa-Filho, E. Longo, A. Valentini, L. F. D. Probst, C. O. Paiva-Santos, and W. H. Schreiner, Sensors Actuators B 86, 185 (2002). 159. D. Aurbach, A. Nimberger, B. Markovsky, E. Levi, E. Sominski, and A. Gedanken, Chem. Mater. 14, 4155 (2002). 160. U. Zum Feld, M. Haase, and H. Weller, J. Phys. Chem. B 104, 9388 (2000). 161. D. Cummins, G. Boschloo, M. Ryan, D. Corr, S. N. Rao, and D. Fitzmaurice, J. Phys. Chem. B 104, 11449 (2000). 162. M. Grätzel, Nature 414, 338 (2001). 163. U. Bach, D. Lupo, P. Comte, J. E. Moser, F. Weissörtel, J. Salbeck, H. Spreitzer, and M. Glätzel, Nature 395, 583 (1998). 164. M. Boudart and G. Djega-Mariadasson, “Kinetics of Heterogeneous Reaction.” Princeton Univ. Press, Princeton, NJ, 1981. 165. I. T. Weber, A. P. Maciel, P. N. Lisboas-Filho, E. Longo, E. R. Leite, C. O. Paiva-Santos, Y. Maniette, and W. H. Schreiner, Nano Lett. 9, 969 (2002). 166. R. A. Caruso and M. Antonietti, Chem. Mater. 13, 3272 (2001). 167. Y. Lu, Y. Yin, B. T. Mayers, and Y. Xie, Nano Lett. 2, 183 (2002). 168. L. M. Liz-Marzán, M. Giersig, and P. Mulvaney, Langmuir 12, 4329 (1996). 169. M. A. Correa-Duarte, M. Giersig, and L. M. Liz-Marzán, Chem. Phys. Lett. 286, 497 (1998). 170. M. Giersig, T. Ung, L. M. Liz-Marzán, and P. Mulvaney, Adv. Mater. 9, 570 (1997). 171. P. Mulvaney, L. M. Liz-Marzán, M. Giersig, and T. Ung, J. Mater. Chem. 10, 1259 (2000). 172. G. Oldfield, T. Ung, and P. Mulvaney, Adv. Mater. 12, 1519 (2000). 173. I. Bedja and P. V. Kamat, J. Phys. Chem. 99, 9182 (1995). 174. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, and J. S. Beck, Nature 359, 710 (1992). 175. C. J. Brinker, Y. Lu, A. Sellinger, and H. Fan, Adv. Mater. 11, 579 (1999). 176. C. P. Mehnert, D. W. Weaver, and J. Y. Ying, J. Am. Chem. Soc. 120, 12289 (1998). 177. T. Hanaoka, H.-P. Kormann, M. Kröll, T. Sawitowski, and G. Schmid, Eur. J. Inorg. Chem. 807 (1998). 178. G. Hornyak, M. Kröll, R. Pugin, T. Sawitowski, G. Schmid, J.-O. Bovin, G. Karsson, H. Hofmeister, and S. Hopfe, Chem. Eur. J. 3, 1951 (1997).
554 179. G. Schmid, J. Chem. Soc. Dalton Trans. 1077 (1998). 180. M. L. Anderson, C. A. Morris, R. M. Stroud, C. I. Merzbacher, and D. R. Rolison, Langmuir 15, 674 (1999). 181. C. A. Morris, M. L. Anderson, R. M. Stroud, C. I. Merzbacher, and D. R. Rolison, Science 284, 622 (1999). 182. M. L. Anderson, R. M. Stroud, and D. R. Rolison, Nano Lett. 3, 235 (2002). 183. E. R. Leite, N. L. V. Carreño, E. Longo, A. Valentini, and L. F. D. Probst, J. Nanosci. Nanotech. 2, 89 (2002). 184. E. R. Leite, N. L. V. Carreño, E. Longo, F. M. Pontes, A. Barison, A. G. Ferreira, Y. Maniette, and J. A. Varela, Chem. Mater. 14, 3722 (2002). 185. N. L. V. Carreno, E. R. Leite, E. Longo, P. N. Lisboa-Filho, A. Valentini, L. F. D. Probst, and W. H. Schreiner, J. Nanosci. Nanotech. 5, 491 (2002). 186. F. C. Fonseca, G. F. Goya, R. F. Jardim, R. Muccillo, N. L. V. Carreño, E. Longo, and E. R. Leite, Phys. Rev. B 66, 104406 (2002). 187. L. Levy, Y. Sahoo, K.-S. Kim, E. J. Bergey, and P. N. Prasad, Chem. Mater. 14, 3715 (2002). 188. S. J. Pennycook and D. E. Jesson, Ultramicroscopy 37, 14 (1991); P. D. Nellist and S. J. Pennycook, Ultramicroscopy 78, 111 (1999). 189. D. B. Willian, “Transmission Electron Microscopy.” Plenum Press, New York, 1996. 190. P. L. Hansen, J. B. Wagner, S. Helveg, J. R. Rostrup-Nielsen, B. S. Clausen, and H. Topsoe, Science 295, 2053 (2002). 191. V. Rodrigues, T. Fuhrer, and D. Ugarte, Phys. Rev. Lett. 85, 4124 (2000). 192. C. L. Cleveland, U. Landman, T. G. Schaaaff, M. N. Shafigullin, P. W. Stephens, and R. L. Whetten, Phys. Rev. Lett. 79, 1873 (1997). 193. R. L. Penn and J. F. Banfield, Geochim. Cosmochim. Acta 63, 1549 (1999).
Nanocrystals Assembled from the Bottom Up 194. J. Weissmüller, “Characterization by Scattering Techniques and EXAFS in Nanomaterials: Synthesis, Properties and Applications” (A. S. Edelstein and R. C. Cammarata, Eds.), p. 219. IOP, Bristol, 1996. 195. C. O. Paiva-Santos, H. Gouveia, W. C. Las, and J. A. Varela, Mater. Struct. 2, 111 (1999). 196. R. F. Pettifer, “X-ray absorption studies of glass structure, in Glasses and Glass-Ceramics” (M. H. Lewis, Ed.), p. 41. Chapman and Hall, London, 1989. 197. J. W. Niemantsverdriet, “Spectroscopy in Catalysis, An Introduction,” p. 137. VCH, Weinhein, 1993. 198. K. S. Hamad, R. Roth, J. Rockenberger, T. van Buuren, and A. P. Alivisatos, Phys. Rev. Lett. 83, 3474 (1999). 199. T. Shido and R. Prins, J. Phys. Chem. B 102, 8426 (1998). 200. J. Rockenberger, L. Tröger, A. Kornowski, T. Vossmeyer, A. Eychmüller, J. Feldhaus, and H. Weller, J. Phys. Chem. B 101, 2691 (1997). 201. C. V. Santilli, S. H. Pulcinelli, G. E. S. Brito, and V. Briois, J. Phys. Chem. B 103, 2660 (1999). 202. C. M. Wang, G. S. Cargill, H. M. Chan, and M. P. Harmer, J. Am. Ceram. Soc. 85, 2492 (2002). 203. M. Yoshikawa, Y. Mori, H. Obata, M. Maegawa, G. Katagiri, H. Ishida, and A. Ishitani, Appl. Phys. Lett. 67, 694 (1995). 204. A. Dieguez, A. Romano-Rodriguez, J. R. Morante, N. Barsan, U. Weimar, and W. Göpel, Appl. Phys. Lett. 71, 1957 (1997). 205. A. Dieguez, A. Romano-Rodriguez, A. Vila, and J. R. Morante, J. Appl. Phys. 90, 1550 (2001). 206. J.-M. Jehng and I. E. Wachs, Chem. Mater. 3, 100 (1991). 207. P. Parayanthal and F. H. Pollak, Phys. Rev. Lett. 52, 1822 (1984). 208. I. Kosacki, V. Petrovsky, and H. U. Anderson, J. Am. Ceram. Soc. 85, 2646 (2002).