Hydrothermal-Induced Assembly of Colloidal Silver Spheres into

Mar 8, 2005 - the Basis of HTAB-Modified Silver Mirror Reaction. Dabin Yu and Vivian Wing-Wah Yam*. Center for Carbon-Rich Molecular and Nano-Scale ...
0 downloads 0 Views 879KB Size
J. Phys. Chem. B 2005, 109, 5497-5503

5497

Hydrothermal-Induced Assembly of Colloidal Silver Spheres into Various Nanoparticles on the Basis of HTAB-Modified Silver Mirror Reaction Dabin Yu and Vivian Wing-Wah Yam* Center for Carbon-Rich Molecular and Nano-Scale Metal-Based Materials Research and Department of Chemistry, The UniVersity of Hong Kong, Pokfulam Road, Hong Kong, P.R. China ReceiVed: NoVember 10, 2004; In Final Form: January 25, 2005

Small colloidal silver spheres (diameter < 10 nm) were found to assemble into various silver nanoparticles including cubes, triangles, wires, and rods in water in the presence of HTAB (n-hexadecyltrimethylammonium bromide) at 120 °C, while the colloids were generated in situ on the basis of a HTAB-modified silver mirror reaction during the synthesis process. Adjustment of the synthesis parameters, in particular the concentrations of HTAB and [Ag(NH3)2]+, led to an obvious shape evolution of silver nanoparticles, thus resulting in the shape-selective formation of the silver nanoparticles. The monodisperse nanocubes with a well-defined crystallographical structure (a single crystal bounded by six {200} facets) have a strong tendency to assemble into two-dimensional arrays on substrates. The nanowires with uniform diameter usually existed in the form of two-dimensional alignments. The findings suggested that hydrothermal-induced assembly of small silver colloidal particles should be a convenient and effective approach to the preparation of various silver nanoparticles.

1. Introduction The preparation of silver nanoparticles has attracted much attention and has been intensely investigated in recent years because of their unique shape-dependent optical and electrical properties that confer them with the ability to function as potential structural blocks for a new generation of electronics,1 photonics,2 and sensor materials.3 Apart from the onedimensional (1D) silver nanoparticles such as nanowires that have been studied for many years, most recently, a few of the other novel nanostructures, such as cubes,4 prisms,5 disks, and plates,6 have attracted particular attention. To provide further insights into the understanding of their intrinsic shape-dependent properties,7 the elucidation of their particle growth mechanism, and development of new template systems,8 there is still a need to develop new shape-selective chemical synthesis strategies to yield these nanoparticles. Although a variety of methods have been developed for the preparation of silver nanoparticles, the transformation of silver colloids into nanoparticles has become one of the most interesting subjects in recent years. Similar to the spontaneous organization of CdTe nanoparticles into nanowires,9 during which the strong dipole-dipole interaction acted as the driving force for nanoparticles self-organization with the removal of the protective shell of organic stabilizer, the strong metallic bonding would help to drive nanocrystal adhesion and subsequent coalescence if the metal particles were allowed to approach each other within a suitable distance.10 For example, Fitzmaurice’s group reported the self-assembly of silver nanoparticles into two-dimensional (2D) nanowire arrays by carefully reducing the capping ligand coverage of size-monodisperse prolate nanocrystals.11 More recently, the assembly of silver colloids (diameter < 10 nm) into nanoparticles induced by light or thermal energy appears to be a promising and attractive alternative. For example, Mirkin and co-workers reported the * Corresponding author. E-mail: [email protected].

photoinduced synthesis of silver nanoprisms by preparing small spheres, followed by conversion of the spheres to large prisms with visible light that is surface-plasmon directed.12 Xia’s group also prepared silver nanobelts and triangular nanoplates by refluxing an aqueous dispersion of spherical colloids in the presence of natural light.13 Recently, we have developed a new approach to the preparation of silver nanocubes on the basis of a HTAB-modified silver mirror reaction at 120 °C.14 Detailed investigation on their formation process indicated that the growth of the nanocubes was, in fact, a result of the assembly of silver colloids (diameter < 10 nm) in the hydrothermal process. Unlike the methods as described above where the silver colloids were prepared by an additional experimental process, in our experiments the colloids were actually generated in situ during the synthesis process. Also, unlike those methods where the light energy seems to be crucial for the assembly of silver colloids,12,13 our experiments were carried out in the absence of light. Most importantly, besides the silver nanocubes, other high-quality silver nanoparticles including triangular particles, rods, and wires were also obtained through the assembly of the silver colloids simply by adjusting synthesis parameters, in particular the concentrations of HTAB and the reagents. This paper reports the synthesis of various silver nanoparticles, the investigations on the reaction process, as well as the formation mechanism of the nanoparticles, with the particular emphasis on the silver nanocubes. 2. Experimental Section Materials. AgNO3 (99.95%) and D-glucose (99%) were obtained from Lancaster Synthesis Ltd. NH3‚H2O (25%) was purchased from Merck. n-Hexadecyltrimethylammonium bromide (HTAB, 98%) was obtained from ABCR GmbH & Co. All reagents were used as received. Preparation of [Ag(NH3)2]OH Aqueous Solution. Analytical grade AgNO3 (0.51 g, 0.003 mol) was dissolved in ca. 50

10.1021/jp0448346 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/08/2005

5498 J. Phys. Chem. B, Vol. 109, No. 12, 2005 mL of deionized water in a beaker. To the AgNO3 solution was added aqueous ammonia (1.0 M) in a dropwise manner under vigorous magnetic stirring until a clear colorless solution was obtained. The solution was then quantitatively transferred into a 100-mL volumetric flask, and was made up to 100 mL with deionized water. Synthesis of Silver Nanoparticles. Silver nanoparticles were synthesized using a procedure as described elsewhere with some minor modifications by adjusting the initial concentrations of HTAB and [Ag(NH3)2]OH to control the morphology and dimension of the silver nanoparticles.14 In a modified synthesis procedure of silver nanocubes, 2.5 mL of freshly prepared [Ag(NH3)2]OH (30 mM) solution was added into a Teflon-lined stainless steel autoclave of 22 mL capacity, followed by the addition of 10 mL of deionized water, 2.5 mL of glucose (20 mM), and 3 mL of HTAB (50 mM) solutions under vigorous stirring. The autoclave was sealed and heated at 120 °C in a furnace for 8 h. (The synthesis time was taken to be the difference between the time the autoclave was put into and taken out of the furnace). It was then allowed to cool naturally to room temperature. After that, the solution was filtered, and then centrifuged at 6000 rpm for 20 min to obtain a precipitate, which was collected and re-dispersed in water (2 mL) for characterization. To investigate the intermediates of the silver nanoparticles, the synthesis was stopped at different stages during the synthesis process. When the autoclave was cooled naturally to ca. 100 °C, it was cooled to room temperature as soon as possible by water, and the sample was then poured out from the autoclave and directly used for characterization without filtering and centrifuging treatment. Characterization. Small drops of the dispersions were placed onto glass substrates and copper grids, and dried at room temperature for X-ray diffraction (XRD) and transmission electron microscopy (TEM) measurements, respectively. The XRD patterns were recorded on a Siemens D-500 (λ ) 1.54056 Å) diffractometer in the 2θ range of 20-70° with a scanning rate of 0.05°/s. TEM images and the corresponding selectedarea electron diffraction (SAED) patterns were taken using a Philips microscope (Tecnai 20) operated at an acceleration voltage of 200 kV. High-resolution transmission electron microscopy (HRTEM) images and the corresponding SAED patterns were obtained on a microscope (JEM2010F) operated at an acceleration voltage of 200 kV. UV/vis absorption spectra were recorded on a Hewlett-Packard 8452A diode array spectrophotometer using quartz cuvettes with an optical path length of 1 cm at room temperature.

Yu and Yam TABLE 1: Various Silver Nanoparticles Synthesized under Different Conditionsa [HTAB] [[Ag(NH3)2]+] temperature (mM) (°C) (mM) 15 8.3 8.3 8.3 8.3 8.3 8.3 8.3 8.3

2.8 2.8 2.8 2.8 4.3 5.3 5.8 6.5 7.0 7.5 4.3 4.3

120 120 120 120 120 120 120 120 120 120 140

shape/profileb sphere cube wire sphere cube cube triangle wire rod sphere

average dimensionc (nm) 25 55 23 >20 58 72 82 28 35 20

irregular particle

a

The initial concentrations were calculated on the basis of the total volume of the reaction mixture, i.e., the mixture of the solutions of HTAB and the reagents together with the water in the autoclave. The concentration of glucose was fixed at 0.75 times that of [Ag(NH3)2]+ in the corresponding synthesis process. The synthesis time was 8 h. b The product was dominated by particles of such shape. c For cubes, this corresponds to the mean edge length; for a triangular profile, this corresponds to the longest edge length. The dimensions were reproducible within (8% of the given value.

Figure 1. TEM images of silver nanoparticles: (A) cubes; (B) triangles; (C) wires; (D) an alignment of wires.

3. Results and Discussion 3.1. Characterization of the Silver Nanoparticles. The shape and size of the Ag nanoparticles depend on the concentrations of surfactant, reagents, synthesis temperature, and time. All the parameters were found to be interdependent, thus resulting in interesting combinations for the shape-selective synthesis of various silver nanoparticles (Table 1). Among the various parameters, the concentrations of HTAB and [Ag(NH3)2]+ are particularly crucial for the control of both the morphology and the size of silver nanoparticles. As shown in Table 1, with the concentration of [Ag(NH3)2]+ fixed at 2.8 mM, an increase in the concentration of HTAB led to a shape evolution of silver nanoparticles from spheres to cubes, wires, and spheres. With the concentration of HTAB fixed at 8.3 mM, increasing the concentration of [Ag(NH3)2]+ from 2.8 to 5.3 mM led to an increase in the average dimensions of the cubes from 55 to 72 nm, accompanied by the formation of a small

amount of wires. Upon further increasing the concentration of [Ag(NH3)2]+, the product was dominated by triangular particles, wires, and rods, and finally dominated by spheres. The nanocubes were characterized by having a regular cubic shape, monodisperse size, and smooth surfaces (Figure 1A). They had a strong tendency to assemble into 2D arrays with a regular checked pattern as communicated previously;14 further details of which will be reported here. The triangular particles exhibited the shape of right-angled triangles or isosceles triangles (Figure 1B), unlike the nanoprisms with a shape of equilateral triangles that were obtained by plasmon excitation.12 The nanowires with uniform diameter of ca. 28 nm and length of tens of microns could be obtained in high yield (Figure 1C), and they tended to exist in the form of 2D alignments with the spacings between adjacent wires of ca. 3 nm (Figure 1D). In addition, some nanowires with some interesting structures and large-scale

Assembly of Colloidal Ag Spheres into Nanoparticles

Figure 2. (A) A typical TEM image of the 2D arrays of silver nanocubes and a SAED pattern (inset) recorded on the area marked with the circle; (B) a typical XRD pattern of the arrays of silver nanocubes; (C) and (D) TEM images with higher magnifications of the array marked with the circle as indicated in (A), image D recorded after rotating the array by an angle of 30°.

nanorods with diameters in the range of 30-50 nm were also obtained. By carefully evaporating the solvent (water), the nanocubes assembled into 2D closed-packed arrays on solid substrate (Figure 2A) since in a controlled drying process particles were transported to and concentrated at the drying edge by a fluid flow induced by the evaporation.15 The nanocubes assembled so well that, by directing the incident electron beam perpendicular to one of square facets of the nanocubes, the ED pattern of a large selected area of an array as marked with the circle in Figure 2A, which contains tens of cubes, seems to be that resulting from an individual cube (inset in Figure 2A). This is probably due to the facts that the cubes not only formed arrays on the substrate with the c-axis strictly perpendicular to the substrate surface, as reflected by the intensity of the (200) diffraction peak being strongest in the XRD pattern as described elsewhere (Figure 2B),14 but also arranged into a checked pattern with uniform spacings between adjacent cubes, which can be observed from the TEM images with higher magnifications (Figure 2C). The major factor that allowed the nanocubes to assemble into such close-packed arrays lies in their monodisperse size and crystallographically well-defined shape. Interestingly, if a TEM image was recorded after rotating an array of cubes by an angle, there appeared an array of particles with cubic morphology emerging from the remarkable contrast (Figure 2D). Figure 3A shows a HRTEM image of a nanocube (bottom right inset) and the corresponding SAED pattern (top right inset). The SAED pattern was obtained by directing the incident electron beam perpendicular to one of the square facets of the cube, and its square spot array was indexed to [200] and [020] of the fcc silver. The perpendicular lattice fringes of the HRTEM image were examined to be 0.205 nm, in agreement with the {200} lattice spacing of the fcc silver. It can be seen that the fringes are either parallel to or orthogonal to the edges of the cube, directly suggesting that the nanocube is of a well-defined shape, that is, a single-crystal bounded by six {200} facets.

J. Phys. Chem. B, Vol. 109, No. 12, 2005 5499 Figure 3B shows a typical HRTEM image of silver nanowires and the corresponding SAED pattern (inset), and the lattice fringes of the HRTEM image were examined to be 0.236 nm, close to the {111} lattice spacing of the fcc silver, indicating that the individual wire was also a single crystal. 3.2. Reaction and Formation Process of the Silver Nanoparticles. 3.2.1. Synthesis Reactions. The synthesis reactions were monitored using XRD and UV/vis absorption spectra. According to one synthesis procedure of silver nanocubes, the concentrations of [Ag(NH3)2]+ and HTAB were fixed at 4.3 and 8.3 mM, respectively, unless indicated otherwise in the following sections. Figure 4 shows the XRD patterns of the samples obtained at different stages during the synthesis process. In Figure 4A, the peaks were assigned to the diffraction of cubic AgBr accordingly (PCPDF no. 790148) except for two weak peaks marked with asterisks, suggesting that after the synthesis had proceeded for 1.25 h Ag(I) existed in AgBr, which resulted from a reversible reaction (eq 1) with the mixture of the solutions of HTAB and [Ag(NH3)2]+.14 As the formation of AgBr resulted in a cathodic shift of the reduction potential of [Ag(NH3)2]+/ Ag,16 the silver mirror reaction (eq 2) could be performed at a suitable rate in a controlled manner despite the fact that a much higher reaction temperature (120 °C) was used for the silver mirror reaction.17

[Ag(NH3)2]+(aq) + Br-(aq) h AgBr(s) + 2NH3(aq)

(1)

[Ag(NH3)2]+(aq) + 2RCHO(glucose)(aq) f Ag (NPs) + 2RCOO-(aq) + 2NH4+(aq) (2) In Figure 4B, a weak peak marked with rhomboid at 38.12° was assigned to the diffraction of the (111) plane of fcc silver metal, indicating that silver metal began to form as described in eq 2 after the synthesis had proceed for 1.5 h, while further support for the formation of silver metal was provided by the evolution of this diffraction peak in Figure 4C,D and the UV/ vis absorption spectra of the corresponding samples. It is likely that the diffraction peak of the (200) plane of silver metal, as marked with rhomboids in the spectra, also appeared at 44.30°. However, its position was so close to that of the (220) peak of cubic AgBr (2θ, 44.61°) that the two peaks were not differentiated under the experimental conditions. As the synthesis had proceeded for 2 h, the (111) and (222) diffraction peaks of AgBr together with the two peaks marked with asterisks in the XRD pattern all disappeared (Figure 4C), and the (200) peak of AgBr became much weaker, suggesting that almost all the AgBr was converted into silver metal. At first glance, it appears as if the intensity of the (220) peak of AgBr became stronger (Figure 4C), but in fact it is due to the formation of silver metal, thus resulting in an increase of the intensity of the (200) peak of silver. A further increase in the reaction time led eventually to a pure phase of silver metal (Figure 4D). Figure 5 shows the photographs and UV/vis absorption spectra of the dispersions obtained after the synthesis had proceeded for various times. As the synthesis had proceeded for less than 1.5 h, the dispersions were almost colorless (Figure 5A, vials a-c), due to the fact that the solutions had little absorption in the visible range (Figure 5B, curves a-c), but when the synthesis proceeded for 1.25 h the dispersion exhibited a broad absorption with band maximum at around 350 nm, indicating that some silver clusters were formed.18 As the synthesis had proceeded for 1.5 h, the dispersion developed a sharp band at 378 nm (Figure 5B, curve c), indicating that truly metallic particles were formed,16 which further confirms the

5500 J. Phys. Chem. B, Vol. 109, No. 12, 2005

Yu and Yam

Figure 3. (A) A typical HRTEM image of Ag nanocubes, insets: the corresponding SAED pattern (top right) and the nanocube used for the SAED and HRTEM analyses (bottom right); (B) a typical HRTEM image of individual nanowires and the corresponding SAED pattern.

Figure 4. XRD patterns of the samples obtained after the synthesis had proceeded for (A) 1.25 h, (B) 1.5 h, (C) 2 h, (D) 4 h. All the peaks were assigned to the diffraction of cubic AgBr except for those marked with rhomboids which were assigned to the diffraction of fcc Ag, and those marked with asterisks which could not be indexed according to JCPDF cards.

analysis of the XRD pattern (Figure 4B). When the synthesis proceeded for 2 h, a broad absorption at 380-418 nm developed, indicating that large silver nanoparticles began to form, and also suggesting that during this stage the silver nanoparticles had a wide size distribution. It is interesting to note that a broad and weak absorption peak centered at ca. 540 nm developed when the synthesis had proceeded for 1.5 h (Figure 5B, curve c), which was a result of the longitudinal plasmon resulting from the aggregation of the silver clusters.19 This peak eventually disappeared as the synthesis time was further increased (Figure 5B, curves d and e), suggesting that such an aggregation did not result in the 1D growth of silver nanoparticles under the experimental conditions. According to the XRD analysis, the chemical reactions were almost completed within 2 h, but according to the UV/vis spectra (Figure 5B, curves d-f), there was still an obvious evolution of the silver nanoparticles in both the shape and size because the absorption band of the samples gradually became sharper with a narrower bandwidth and exhibited an obvious red-shift with an increase in the synthesis time, which will be further confirmed by TEM investigations. 3.2.2. Growth of SilVer Nanoparticles. As described above, the reactions were almost completed within 2 h and were accompanied by the formation of silver metal. After that, TEM images were used to follow the growth process of silver nanocubes (Figure 6). As the synthesis had proceeded for 2 h, a number of small silver particles with an average size of approximately 6 nm were generated (Figure 6A). Meanwhile, a few large silver nanoparticles with spherical or nearly cubic shape were also observed (Figure 6B), indicating that the silver nanocubes began to form. When the synthesis had proceeded

Figure 5. (A) Photographs and (B) UV/vis absorption spectra of the dispersions of the samples obtained after the synthesis had proceeded for (a) 0 h, (b) 1.25 h, (c) 1.5 h, (d) 2 h, (e) 4 h, and (f) 8 h. The samples used for (b), (c), (d), and (e) correspond to those of (A), (B), (C), and (D) in Figure 4, respectively.

for 2.5 h, an increasing number of large silver nanoparticles with nearly cubic shape appeared, while some of them had bent facets and some of them still had a nearly spherical shape, as indicated by the arrows (Figure 6C). When the synthesis time was increased to more than 3 h, silver nanoparticles with regular cubic structure gradually formed (Figure 6D-F). It was found that once the silver particles achieved a shape with six smooth {200} facets and sharp edges (Figure 6F), their size did not show an obvious increase with an increase in synthesis time, which was in contrast to the silver nanocubes synthesized in ethylene glycol where a longer growth time resulted in larger particle size.4 The TEM images show an obvious evolution of silver particles in both size and shape during the growth process of nanocubes, in agreement with the analysis of the evolution of the UV/vis spectra. The formation process of the silver nanocubes is proposed as illustrated schematically in Figure 6G. The growth of silver

Assembly of Colloidal Ag Spheres into Nanoparticles

J. Phys. Chem. B, Vol. 109, No. 12, 2005 5501

Figure 6. (A-F) TEM images of the samples obtained after the synthesis had proceeded for (A) and (B), 2 h; (C) 2.5 h; (D) 3 h; (E) 4 h; and (F) 8 h. The inset in image A shows the corresponding SAED pattern, the diffraction rings of which were assigned to (111), (200), (220), and (311) of fcc silver metal, respectively. (G) Schematic illustration of the proposed formation process of silver nanocubes.

Figure 7. TEM images: (A) 1D array of Ag colloids with diameter of 4 nm. The inset shows the corresponding SAED pattern, the diffraction rings of which were indexed to the faces of fcc Ag metal; (B) two wires with uniform diameter assembled from Ag colloids. The samples were obtained at 120 °C after the synthesis had proceeded for 4 h, and the initial concentrations of [Ag(NH3)2]+ and HTAB were 6 and 8.3 mM, respectively.

nanoparticles should be a result of the interplay of the faceting tendency of the stabilizing agent of HTAB and the growth kinetics of silver metal. On one hand, [Ag(NH3)2]+ reacted with Br- from HTAB in reaction 1 to produce AgBr colloids, which were stabilized by HTAB. Subsequently, with the proceeding of reaction 2, silver colloids were formed and temporarily stabilized by HTAB, but they were unstable due to their high energy resulting from the small particle size. Driven by thermal movement, they assembled into larger particles with spherical or near-cubic shapes, on the basis of which nanocubes were eventually formed. Obviously, from the spherical or near-cubic silver particles to the final nanocubes, the {111} and {110}

facets of the particles had to grow faster than their {200} facets so as to form the sharp vertexes and edges of the cubes. That is, the silver colloids should have preferential {111} and {110} assembling orientation. Similar to the case of the formation of cubic gold nanoparticles in the presence of cetyltrimethylammonium bromide (CTAB) as reported by Murphy et al.,20 HTAB molecules probably bound more strongly to the {200} than to the {111} and {110} facets, which led to an increase in the growth rate along the 〈111〉 direction and a reduction in the growth rate along the 〈100〉 direction, thus resulting in the formation of the nanocubes.4 On the other hand, from the kinetics point of view, the lowest-energy {111} facets were

5502 J. Phys. Chem. B, Vol. 109, No. 12, 2005

Yu and Yam

Figure 8. TEM images of Ag colloids synthesized at 120 °C for 8 h using the same concentration of [Ag(NH3)2]+ (4.3 mM) and different concentrations of HTAB: (A) 3 mM; (B) 30 mM.

favorable for the colloids to assemble, leading to their disappearance and the formation of {200} facets till a cubic structure bounded by the six {200} facets was achieved. Although more direct intermediates of silver nanowires were not observed, a number of 1D arrays of aligned silver colloids with a size of approximately 4 nm were found in the prolate HTAB micelles after the synthesis had proceeded for 4 h (Figure 7A). Meanwhile, some nanowires that formed along the arrays were also observed (Figure 7B). Of particular interest is to note that the corresponding spaces left alongside the nanowires suggested that the nanowires were formed probably via the assembly of small spherical particles. Unlike the nanowires that were assembled from silver nanoparticles in a polyol/toluene medium which revealed a polycrystalline texture,21 these nanowires were characterized by the nature of a single crystal. If more than one nanowire were formed along an array of aligned small spheres, strong van der Waals forces would exist between adjacent nanowires because of their closed packing, and may lead to the formation of the 2D alignments of silver nanowires. Since the morphology of the surfactant micelles may range from spherical shape to prolate or rodlike, and lamellar phase, depending on the concentration of the surfactant and additives,22 the assembly of the silver colloids into wires other than the cubes could probably be attributed to the template action of the prolate micelles of HTAB under these experimental conditions. Because the as-synthesized silver nanoparticles were wellcrystallized single crystals, the silver colloids had probably assembled through an adhesion process followed by a subsequent coalescence process, similar to the 2D silver nanowire array formed by reducing the capping ligand coverage on small silver particles.11 However, this process was driven by hydrothermal energy, which could be supported by the fact that if the autoclave was cooled to room temperature the assembly of the colloids stopped during the growth process of the nanocubes. Thus, it is proposed that the observed formation of the silver nanoparticles occurs through a hydrothermal-induced assembly mechanism. Similar to other studies reported in the literature,11,23 the prerequisites for the assembly of small spheres into silver nanoparticles are the formation of the small colloidal particles ( 10 nm), which could not assemble into the nanoparticles because of their low energy (Figure 8A). Although a much higher HTAB concentration would be favorable for the forma-

tion of small silver colloids (