Synthesis of Nanosized Silver Platelets in ... - ACS Publications

8692

Langmuir 2002, 18, 8692-8699

Synthesis of Nanosized Silver Platelets in Octylamine-Water Bilayer Systems Doruk O. Yener,† Ju¨rgen Sindel,‡ Clive A. Randall, and James H. Adair* Materials Research Institute, The Pennsylvania State University, University Park, Pennsylvania 16802 Received August 2, 2001. In Final Form: April 29, 2002 The synthesis of nanosized powders with targeted properties and their assembly is of considerable importance to the microelectronics industry because of the pervasive drive to miniaturize components. In the current work, synthesis of nanosized silver (Ag) metal platelets was studied. The platelets were synthesized in the lamellar bilayer phase region (neat phase D) of the octylamine-water binary system. Amylamine was used as cosurfactant to increase the phase stability of the lamellar region. The influences of the synthesis conditions and the concentrations of the system components on morphology and size of the platelets were examined. Low-angle X-ray diffraction studies revealed that the water/surfactant molar ratio (R) has a strong influence on the thickness of bilayers (d1 + d2) and consequently particles grown in them. Atomic force microscopy (AFM), scanning electron microscopy (SEM), and high-resolution transmission electron microscopy (HRTEM) studies have shown the effect of R on platelet thickness (t) and face size (a). Thickness and face area of platelets is controlled by regulating the bilayer dimensions by changing R. Platelets with face sizes ranging from 12 to 83 nm and thicknesses from 0.3 to 5 nm have been synthesized at room temperature.

Introduction The miniaturization of electronic components is one of the most important goals to increase the volume efficiency of electronic circuits. For components such as multilayer capacitors, the thicknesses of the alternating dielectric and electrode layers can be minimized to reduce the volume without compromising performance.1-4 The thickness of the dielectric layers has been successfully reduced to below 1.0 µm.5 Relative to the shrinkage in the ceramic layers, only modest success has been achieved in decreasing the thickness of the electrode layers. This subject is of particular importance, because of the high cost for the noble metals such as silver and their alloys commonly used in this application. The replacement of equiaxedshaped metal particles with metal platelets offers the opportunity to form thinner layers due to the efficiency of the surface coverage/unit volume of the anisotropically shaped platelets relative to spherical particles. A wide variety of methods have been used to prepare silver colloid monolayers, anisotropically shaped silver and silver composite platelets, including self-seeding, the vertical freezing method,6 ball milling of oxides,7 epitaxial growth,8,9 chemical precipitation,10 gas evaporation tech* Corresponding author. Fax: (814)863-9704. Tel: (814)8636047. E-mail: [email protected]. † Present address: Saint-Gobain Ceramics & Plastics, Northboro R&D Center, Goddard Rd., Northboro, MA 01532. ‡ Present address: Robert Bosch GmbH, K3/ESV, Postfach 30 02 20, 70442 Stuttgart, Germany. (1) Pepin, J. G. Adv. Ceram. Mater. 1988, 3, 517. (2) LaBranche, M. H.; Pepin, J. G.; Borland, W. Proceedings of the ASM Thick Film Conference, Atlanta, GA, June 8-9, 1988; ASM International: Metals Park, OH, 1988. (3) Pepin, J. G. J. Mater. Sci. 1991, 2, 34. (4) Pepin, J. G.; Borland, W.; Callaghan, P. O.; Young, R. J. S. J. Am. Ceram. Soc. 1989, 72, 2287. (5) Kuo, C. C. Y. ASM Engineered Materials Handbook; ASM International: Metals Park, OH, 1991; Vol. 4, p 1140. (6) Niwa, E.; Masumoto, K. J. Cryst. Growth 1998, 192, 354. (7) Szwarc, H.; Gasgnier, M. J. Solid State Chem. 1998, 136, 51. (8) Schieffer, P.; Krembel, C.; Hanf, M. C.; Gewinner, G. Surf. Sci. 1998, 398, 332.

niques,11 vacuum deposition,12 Langmuir-Blodgett films,13 and self-assembly systems.14a,b,c In the literature, selfassembly systems were also used for synthesis of anisotropic particles with different shapes including rodlike14d,e,f or filament14g morphologies. Synthesized particles via self-assembly systems are often almost monosized and nearly agglomerate-free.14a Adequate quantities of the powders can also be produced to permit comprehensive studies on processing characteristics and properties. Particles have been synthesized in sizes as small as 3 nm in diameter without compromising the ability to control shape.14a The current study focuses on the use of self-assembly (SA) systems for the synthesis conditions including the ratio of water to surfactant and Ag+ concentration that control thickness and aspect ratio of nanometer size, tabular silver. In this system, the amphiphilic octylamine and amylamine molecules form micellar bilayer structures which act as domains for the formation of platelets. Between the bilayer an aqueous solution is present. When cations are introduced into this system, it has been proposed that the cations associate with the polar headgroups of the amphiphilic molecules between the bilayers (9) Newstead, D. A.; Norris, C.; Binns, C.; Stephenson, P. C. J. Phys. C.: Solid State Phys. 1987, 20, 6245. (10) Medendorp, N. W., Jr.; Bowman, K. J.; Trumble, K. P. Mater. Sci. Eng., A 1996, A212, 222. (11) Tanaka, N.; Cowley, J. M. Advanced Photon & Particle Techniques for the Characterization of Defects in Solids Symposium; MRS Symposium Proceedings 155; MRS: Pittsburgh, PA, 1984. (12) Takayanagi, K. Jpn. J. Appl. Phys., Part 1 1983, 22, L4. (13) Bright, R. M.; Musick, M. D.; Natan, M. J. Langmuir 1998, 14, 5695. (14) (a) Adair, J. H.; Li, T.; Kido, T.; Havey, K.; Moon, J.; Mecholsky, J.; Morrone, A.; Talham, D. R.; Ludwig, M. H.; Wang, L. Materials Science and Engineering Report; Elsevier Science S.A.: Amsterdam, 1998; Vol. R23, Chapter 5. (b) Taleb, A.; Petit, C.; Pileni, M. P. Chem. Mater. 1997, 9, 950. (c) Fendler, J. H. Chem. Mater. 2001, 13 (10), 3196. (d) Kon-no, K.; Koide, M.; Kitahara, A. Nippon Kagaku Kaishi 1984, 6, 815. (e) Pileni, M. P.; Gulik-Krzywicki, T.; Tanori, J.; Filankembo, A.; Dedieu, J. C. Langmuir 1998, 14, 7359. (f) Peng, X. G.; Manna, L.; Yang, W. D.; Wickham, J.; Scher, E.; Kadavanich, A.; Alivisatos, A. P. Nature 2000, 404, 59. (g) Hopwood, J. D.; Mann, S. Chem. Mater. 1997, 9, 1819.

10.1021/la011229a CCC: $22.00 © 2002 American Chemical Society Published on Web 10/02/2002

Nanosized Silver Platelets

Langmuir, Vol. 18, No. 22, 2002 8693

due to coordination bonds with the nitrogen ligand in the amine groups with the metal ions.14a This step is followed by two-dimensional growth. Since the growth in the z axis is inhibited relative to the growth in the x and y axes, this process leads to predominantly two-dimensional growth. While there have been a number of studies on the processing features that control size for spherical14b and rod shaped particles,14e there have not been any studies providing a detailed systematic analysis of synthesis variables on particle dimensions for tabular-shaped, nanometer-size particles via self-assembly systems. Pileni et al.14e and Peng et al.14f have addressed size and shape control for rod-shaped nanoparticles. In particular, Pileni et al.14e determined the two-dimensional length characteristics of elongated Cu metal particles as a function of sodium chloride concentration on the basis of transmission electron microscopy observations. The major goal of this study is to determine the effect of process variables (i.e. [water]/[surfactant] (R) and n(Ag+)) on the thickness, face sizes, and aspect ratios of nanometer-size tabular particles. Care must be taken to avoid formation of the metal oxide or hydroxide even with a noble metal such as silver. To prevent oxidative precipitation of Ag2O, Ag+ has been reduced to Ag0 with reducing agents such as Nafion 117 dissolved in alcohol,15 ethylene glycol,16 poly(ethyleneimine) (PEI) in acetonitrile,17 alkyl acid phosphate,18 and hydrazine hydrate.14a,19 Hydrazine hydrate (N2H4‚xH2O) has the advantage of reducing silver without additional heat or chemical treatment19 and therefore was used as a reducing agent in this study. Hydrazine is only a modestly stable compound and reduces Ag+ via several different reaction routes. One possible reaction for the reduction of Ag+(aq) by N2H4(aq) is given as follows:20a,20b,21

N2H4(aq) + 4OH- h N2(g) + 4H2O + 4eE°oxidation ) +1.16 V (1a) 4Ag+(aq) + 4e- h 4Ag0(s) E°reduction ) +0.80 V (1b) 4Ag+(aq) + N2H4(aq) + 4OH-(aq) h 4Ag0(s) + N2(g) + 4H2O E°cell ) +1.96 V

(1c)

Oxidation of the silver platelets after synthesis can be minimized with a corrosion inhibitor. In previous studies, D-sorbitol22 and poly(vinylpyrrolidone) (PVP)18,22-23 were used for this purpose during the metal particle synthesis. Since PVP was successfully used specifically for silver particles, it was selected as the corrosion inhibitor in the current study.18,22,23 Several approaches were used in the current study to determine the influence of the water/surfactant molar (15) Huang, Z. Y.; Mills, G.; Hajek, B. J. Phys. Chem. 1993, 97, 11542. (16) Torigoe, K.; Nakajima, Y.; Esumi, K. J. Phys. Chem. 1993, 97, 8304. (17) Duff, D G.; Curtis, A. C.; Edwards, P. P.; Jefferson, D. A.; Johnson, B. F. G.; Logan, D. E. J. Chem. Soc., Chem. Commun. 1987, 1264. (18) Zhang, Z.; Zhao, B.; Hu, L. J. Solid State Chem. 1996, 121, 105. (19) Burshtain D.; Zeiri L.; Efrima S. Langmuir 1999, 15, 3050. (20) (a) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; John Wiley and Sons: New York, 1988; p 316. (b) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; John Wiley and Sons: New York, 1988; p 938. (21) Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions; National Association of Corrosion Engineers: Houston, TX, 1974; p 393. (22) Fievet, F.; Fievet-Vincent, F.; Lagier, J. P.; Dumond, B.; Figlarz, M. J. Mater. Chem. 1993, 3, 627. (23) Ducamp-Sanguesa, C.; Herrera-Urbina, R.; Figlarz, M. J. Solid State Chem. 1992, 100, 272.

Table 1. Selected Composition Points for the Synthesis Experimentsa sample no.

n(Ag+) (mol)

R

W

a (nm)

t (nm)

d1 (nm)

d2 (nm)

Ag-Oct-1 Ag-Oct-2 Ag-Oct-3 Ag-Oct-4 Ag-Oct-5

9.4 × 10-4 1.6 × 10-3 2.5 × 10-3 3.8 × 10-3 5.7 × 10-3

1.7 2.9 4.6 6.8 10.3

0.45 0.69 1.04 1.56 2.45

12 16 26 58 83

0.3 0.3 0.7 2 5

0.3 0.8 1.0 2.7 6.1

0.98 0.99 1.01 1.02 1.03

a R ) molar ratio of [water]/[octylamine], and W ) water (wt %)/octylamine (wt %). The octylamine to amylamine weight ratio is 25:1. n(Ag+) ) number of moles of Ag+ ion, d1 ) aqueous layer thickness, d2 ) amphiphilic molecular layer thickness, a ) platelet face size, and t ) platelet thickness. The concentration of [Ag+] is 0.04 M.

ratio, R, on the particle size and morphology. Highresolution transmission electron microscope (HRTEM) and atomic force microscope (AFM) were used to determine the face area and thickness, respectively. X-ray diffraction (XRD) with both low-angle and wide-range scans was used to determine the aqueous layer and amphiphilic molecular layer thicknesses as well as to confirm phase analysis of the precipitated silver particles. Materials and Methods All chemicals used were reagent grade. Silver nitrate (greater than 99% purity, Aldrich Chemical Co.) was used as the silver source. Octylamine (OA) (99%, Aldrich Chemical Co.), amylamine (AA) (99%, Aldrich Chemical Co.), hydrazine hydrate (Aldrich Chemical Co.), ammonium hydroxide (28-30% NH3, Aldrich Chemical Co.), and ethanol (94.4% 200 proof, J. T. Baker Chemicals) were used as received without further purification. N-Vinylpyrrolidone (NVP) (99+%, Aldrich Chemical Co.), and hydrogen peroxide (30 wt % in water, Aldrich Chemical Co.) were kept refrigerated before their use. Deionized water (specific conductivity ) 0.4 × 10-7 S/m) was used for all experiments. In this study OA + AA-water microemulsion systems were used to form lamellar bilayer phases. The amphiphilic molecules octylamine and amylamine were used as surfactant and cosurfactant in a weight ratio of 25:1, respectively. Preliminary studies showed that mixing an aqueous solution of Ag(NO3) with OA formed poorly defined edges for the platelets as well as anisotropically shaped clusters instead of the desired well-defined platelets. The addition of cations probably compromises the lamellar layers. For this reason AA, a shorter chain amine molecule, was used as a cosurfactant to increase the amphiphilicity by increasing the amount of hydrophilic functional groups and hydrophobic tail groups without causing the formation of any amphiphilic molecule clusters.24 This leads to an increase in the thermodynamic phase stability in the micellar and lamellar phase regions. The synthesis compositions were selected within the lamellar bilayer phase (D) regions in the binary phase diagram for the water OA system (Table 1).25 The effect of AA addition on the change in the phase stability regions at 20 °C was examined. In all cases reported (Table 1), polarized light microscopy verified that the birefringement bilayer phase was present. The flowchart in Figure 1 was used to prepare anisotropic particles in the bilayer system. A defined amount of an aqueous Ag(NO3) stock solution (0.04 M) was slowly added to a specific amount of a surfactant or surfactant/cosurfactant mixture. The R ratios were varied in the range between 1.7 and 10.3 to control its effect on the bilayer thickness (d1), on the platelet face size (a), and on the thickness (t). During the synthesis steps, the temperature was maintained at 20 ( 0.2 °C with a thermal bath. After the aqueous solution addition, the mixture was shaken vigorously for 15 min to equilibrate the system. This leads to the formation of a transient white precipitate of Ag2O in solution which was verified by XRD. This step was followed by the addition (24) Wormuth K. R.; Kaler E. W. J. Phys. Chem. 1987, 91, 611. (25) Sjoblom, J.; Stenius, P. Surfactant Science Series, Vol. 23; Marcel Dekker: New York, 1987; p 398.

8694

Langmuir, Vol. 18, No. 22, 2002

Yener et al. Japan), and with a high-resolution transmission electron microscope (HRTEM) (HF2000, Hitachi, Japan). The SEM samples were prepared by adding drops of suspension onto a specially polished sample holder. For the AFM anlyses, the samples were prepared by placing drops of the Ag particles in ethanol on an atomically flat, freshly cleaved mica substrate surface. Typical root-mean-square (Rg) surface roughness for molecularly cleaved mica is 0.04-0.05 nm. Thus, AFM determination of the thickness of even the thinnest particles, as seen in Table 1, 0.4 nm for sample Ag-Oct-1, on the molecularly cleaved mica was not difficult to evaluate because of the low roughness associated with the substrate. For the TEM studies, carbon-Formvar film supported Cu grids were used as sample holders. A single drop of suspension was placed on each TEM grid.

Results and Discussion

Figure 1. Synthesis and process diagram used in the synthesis of tabular, nanometer-size particles. of hydrazine hydrate and another vigorous shaking step, before allowing the system to statically equilibrate for 24 h. The second equilibration ensured complete reduction of the silver particles. The color change of the precipitate from white to dark gray indicated the transformation from Ag+ precursors to Ag0. The silver particles were recovered by a solvent extraction process carried out in separation funnels. All solid-solution separation steps were carefully conducted to avoid direct exposure of the reacted solution to light. The equilibrated bilayers were first broken down by the addition of ethanol into the solution with a 3:1 volume ratio for the ethanol and bilayer, respectively. Breakdown of the bilayers after ethanol addition took about 24 h to complete. Gravitational sedimentation of the particles occurred during this step. The sedimented particles were transferred directly into centrifuge tubes and washed three times with ethanol using 3000 rpm at 10 min for each. To inhibit further oxidation and to achieve dispersion of the platelets in the ethanol solution, the particle surfaces were coated by PVP. Preliminary studies in our laboratory showed that PVP can be polymerized by mixing N-vinylpyrrolidone (NVP) with H2O2 in an ethanol solution. Appropriate amounts of NVP were added to the system to keep the Ag0 to NVP weight ratio at 2:3 based on the Ag values in Table 1. After NVP addition, the solution was vigorously shaken. Then a premixed 2.53 M NH3/EtOH solution with H2O2 (NVP:H2O2 weight ratio is 10:1) as initiator was added to the system followed by mechanical agitation for 5 min. To determine bilayer thickness, small samples were taken from the batches immediately prior to the solvent extraction. Without the samples being dried, X-ray diffraction (XRD) measurements (Pad V, Scintag) were performed between 3 and 25° in 2θ angles for a low-angle θ scan to determine bilayer thickness. The bilayer solutions were placed in quartz capillaries mounted on a single-crystal Si substrate to minimize water loss through evaporation during XRD analyses. The solid phase was verfied as fcc metallic silver for all reported samples with wide angle XRD scans. The morphologies and dimensions of the particles were examined with an atomic force microscope (AFM) (MultiMode, Digital Instruments) using the tapping mode, with a scanning electron microscope (SEM) (JSM 6300F, JEOL,

The hypothesis that the polar headgroup of the amphiphilic molecule acts as a binding site to promote templating of platelet particles was tested by systematic variations in bilayer thickness. Careful characterization was conducted to establish correlations between bilayer thickness and platelet thickness. Characterization of Bilayers. Mixing an aqueous solution of Ag(NO3) with OA + AA solution changes the physical nature of the solution due to variations in the water/surfactant molar ratio R. When the water content is too high or too low (R less than 1.7 or greater than 10.3), clear solutions of low viscosity are obtained. The interfacial surface tension among the different phases has a strong effect on viscosity.26,27 Increasing R to 4.5 leads to an increase in viscosity due to gelation. The gel is formed by the lamellar liquid crystals in the system. For R ) 1.7 and 10.3, initially the solutions were transparent due to the micellar structures that form in very small areas. In all cases of R summarized in Table 1 other than R ) 1.7 and 10.3, the solutions were white and opaque. After introduction of the reducing agent and equilibrating, the reduction to silver metal causes a change in the light absorption of the samples and all solutions become dark gray. Immediately after the 24 h equilibration, the samples were analyzed by X-ray diffraction to determine the interlamellar spacings. During the XRD evaluations, a low-angle scan was used as shown in Figure 3 (with minimum angle at 1.5°). The interlamellar spacings obtained were calculated using Bragg’s law:28

nλ ) 2d sin θ

(2)

Here n is the order of refraction, λ is the X-ray wavelength (i.e. 1.54 Å for Cu KR), d is the interplanar spacing, and θ is the diffraction angle. The d spacings are summarized in Figure 3 and Table 1. Several previous studies assumed the d spacing values to be the sum of the aqueous layer thickness (d1) and organic layer thickness (d2).29-32 Such an assumption does not permit the separate calculation of d1 and d2 and can possibly lead to incorrect assignments for thicknesses of (26) Nyvlt, J.; Sohnel, O.; Matuchova, M.; Broul, M. Chemical Engineering Monographs 19; Elsevier: New York, 1985; p 56. (27) Biasio, D.; Cametti, C.; Codastefano, P.; Tartaglia, P.; Rouch, J.; Chen, S. H. Prog. Colloid Polym. Sci. 1993, 93, 191. (28) Cullity, B. D. Elements of X-ray Crystallography, 2nd ed.; Addision-Wesley Publishing Co.: New York, 1978; p 86. (29) Auvray, L.; Cotton, J. P.; Ober, R.; Taupin, C. Surfactant Science Series, Vol. 24; Marcel Dekker: New York, 1987; p 225. (30) Rico, A.; Lattes, A. Surfactant Science Series, Vol. 24; Marcel Dekker: New York, 1987; p 357. (31) Marigo, A.; Marega, C.; Zannetti, R.; Ajroldi, G. Macromolecules 1996, 29, 2197. (32) Wang, W.; Efrima, S.; Regev, O. J. Phys. Chem. B 1999, 103, 5613.


Langmuir, Vol. 18, No. 22, 2002 8695

Figure 3. Variation of d spacing due to the R value change. The data were obtained by low-angle X-ray scans. d1 ) d spacing of the water layer thickness, and d2 ) d spacing of the organic layer thickness.

Figure 2. (a) Schematic of the proposed layered two-phase system. The schematic is based on X-ray diffraction analyses of inorganic superlattices in refs 34a,b. (b) Low-angle X-ray diffraction pattern obtained from the Ag-Oct-4 system. (See Table 1 for details.) (c) d spacing values that were calculated from the low-angle X-ray diffraction pattern for the Ag-Oct-4 system.

the aqueous and organic phases. In this study, the procedures relevant to the interpretation and analysis of X-ray diffraction from inorganic superlattices33 including amorphous layers33d were used to guide evaluation of the thicknesses of the aqueous and organic layers. Also in this study, the d spacing calculations were carried out by assuming the formation of a two-phase, layered system33a,b as shown in Figure 2a. Since each lamellar phase composed of either the aqueous layer or amphiphilic molecular layer has its own order in Bragg’s equation, from low angle to high angle every peak in the diffraction pattern was taken (33) (a) Lowe, W. P., Barbee, Jr., T. W., Geballe, T. H., McWhan, D. B. Phys. Rev. B 1981, 24 (10), 6193. (b) Underwood, J. H.; Barbee, T. W., Jr. Appl. Opt. 1981, 20 (17), 3027. (c) Rodmacq, B. J. Appl. Phys. 1991, 70 (8), 4194. (d) Fujimoto, T.; Li, B.; Xu, W.; Kojima, I. Characterization and Metrology for ULSI Technology: 2000 International Conference; Seilor, D. G. A., Diebold, C., Shaffner, T. J., McDonald, R., Bullis, W. M., Smith, P. J., Secula, E. M., Eds.; American Institute of Physics: Secaucus, NJ, 2001; pp 586-590.

as n ) 1, with the peaks obtained at higher angles calculated with the appropriate reflection number (n ) 2, 3, 4, ..., and so on). The remaining calculations for d spacings were carried out by following this order. The d spacing results obtained for each peak series were compared and designated as d1 and d2 for the aqueous and the amphiphilic molecular phase, respectively.33a,b Change in the d spacing with R permitted assignment of the reflections due to the aqueous phase and amphiphilic layers, respectively, as shown in Figure 3. The results of the XRD analyses showed only a slight increase in the amphiphilic molecular layer (d2) values with increasing R. In all experiments amphiphilic molecular layer thickness remained around 1 nm which correlates with the extended length of a single octylamine molecule (0.97 nm due to ACD ChemSketch ver. 4.01 software output34). This analysis indicates that the amphiphilic molecules are arranged in an overlapping tail conformation as shown in Figure 2a instead of forming a tail to tail double layer. In contrast to the insensitivity of the organic layer thickness to changes in R, a reasonably linear increase is present in the aqueous phase thickness (d1) due to increases in R (Figure 3). This increase occurred due to the swelling of the bilayers with the introduction of increasing amounts of water into the system. When the surfactant content is increased, the excess surfactant content increases the surface area of the bilayers but not the thickness. Since the volume of water between the bilayers is constant, an increase in the surface area results in a slight decrease of the bilayer thickness. This explanation is also consistent with the trends observed for silver particle face areas discussed below. After polymerization of NVP to PVP on the particle surfaces, a small sample batch was taken from the main batch and dried at ambient atmospheric conditions in a fume hood. The diffraction pattern in Figure 4c matched the standard pattern for phase pure silver (JCPDS ICDD 4-783). Characterization of Platelets. To verify the tabular shape of the particles, morphological analysis was initially conducted using high-resolution field emission SEM. The SEM analyses with albeit poor resolution (not shown) indicated that the particles produced from all of the compositions in Table 1 were platelets although the composition point R ) 1.7 lies outside the lamellar bilayer phase region. It is possible that the lamellar bilayer stability phase region is enlarged with the addition of AA as a cosurfactant. In particular, the particles that were obtained for R ) 1.7 were only a few unit cells in size and resolution with the field emission SEM was extremely (34) ACD ChemSketch ver. 4.01, http://www.acdlabs.com, 2001.

8696


Figure 4. (a) HRTEM micrographs of silver platelets. Lattice spacing shown in the insert of the right-hand photomicrograph for one of the particles correlates with the (111) plane of silver. The 95% confidence interval is 2.5 ( 0.3 Å. (b) Electron diffraction pattern of a platelet aggregate from the sample batch Ag-Oct-4 (see Table 1 for details), which is consistent with the phase pure silver X-ray diffraction pattern (JCPDS-ICDD 4-783). (c) X-ray diffraction patterns indicating that the tabular particles do not have a preferred texture.

poor. Therefore, both HRTEM and AFM were used to provide characterization to determine face size (a) and thickness (t), respectively. HRTEM was used to obtain detailed data about the morphology and the crystal structure of the particles. It was observed that the platelets have circular or ellipsoidal face morphologies. The lattice fringes observed in the highresolution bright-field images demonstrated that the many of the platelets were single crystal (Figure 4a). The distance of 2.5 ( 0.3 Å between these fringes is in reasonable correlation to the d111 of 2.36 Å for the fcc crystal structure (a ) 4.09 Å) for one of the particles. However, even though some of the particles are single crystal, there is no preferred orientation or texture indicated in the XRD pattern (with an example shown in Figure 4c) or the selected area diffraction pattern (SADP) for the electron image shown in Figure 4b for a cluster of particles. To obtain a more detailed phase analysis, the SAPD was

Yener et al.

obtained from an aggregate of platelets giving a pattern in which the diffraction rings could be indexed. The electron diffraction pattern (see Figure 4b) gave similar results to that for the XRD verifying that each of the platelets is phase pure Ag metal but without particular texture (Figure 4b and c). It is likely that the amphiphlic layer is too mobile at 25 °C to produce a truly templated collection of particles all with the same habit. One way to check this is to reduce the Ag in the bilayer as a function of temperature. One would expect the bilayer to become less mobile as temperature is reduced. While the influence of temperature is beyond the scope of the current study, the effect of particle morphology and crystallographic orientation should be evaluated in future work. Since the thickness fringe studies by HRTEM did not provide results precise enough to determine platelet thickness (t), thickness measurements were conducted using AFM. During the AFM studies, the tapping mode was used to determine the face diameter and thickness of the platelets on atomically flat freshly cleaved mica substrate surface. TEM holography of the platelets was used to verify the AFM thickness determinations. Details of this analysis will be provided in a separate report.35 Platelets were washed three times with ethanol to dissolve and remove the organic layer. The particles presented in Figure 5, shown lying on the molecularly cleaved mica substrate, have a flat cylindrical shaped morphology. The measured face dimensions are in good agreement with the XRD, SEM, and HRTEM results. In general, an increase in R leads to an increase in both “a” and “t”. This causes a decrease in the aspect ratio of platelets, which is the ratio of platelet face size (a) to its thickness (t) as summarized in Figure 6 discussed below. However, the particles produced in the thinnest bilayers (samples AgOct-1 and -2, Table 1) both have thicknesses of ∼0.3 nm or only about one atomic layer thick for either water or silver atoms, with corresponding thicknesses for the aqueous phase of 0.3 and 0.8 nm. It is likely that up to the amount of Ag+ indicated for these thinner layers of 9.4 × 10-4 and 1.6 × 10-3 mol the bilayers polar headgroups are undersaturated with respect to the silver. Thus, increases in the amount of silver above these values leads to multilayers of Ag+ in the bilayer and thicker silver particles upon reduction. Relationship of Aqueous Bilayer Thickness to Platelet Thickness. Definitive kinetic evaluations are currently being conducted to determine the rate-determining mechanisms for nucleation and growth of the tabular particles. However, the aqueous layers are likely to act as domains for nucleation and growth. Ag+ clusters introduced into the aqueous solution are considered to be bonded to the polar headgroups of the amphibilic molecules. The type of bonds between Ag+ and the nitrogen in the amine groups are likely to form a covalent coordination bond due to the strong complex generally formed by amines with Ag+.14a With such bonding, clustered Ag+ ions at the amine/water phase boundary can act as intrinsic nuclei. During the growth process, it will be assumed that particles grow both parallel and perpendicular to the phase boundaries. It is proposed that covalently coordinated neighbor silver ions on the amine headgroup form the silver layer on the polar headgroups of the organic layer. The free ions in aqueous solution are likely to be nucleated on the Ag-clustered regions of the phase boundary when reducing agent is introduced. This leads to formation of (35) Yener, D. O.; Carim, A. H.; Adair, J. H. J. Phys. Chem., submitted for publication.


Langmuir, Vol. 18, No. 22, 2002 8697

Figure 5. AFM pictures of a sample from the batch Ag-Oct-4. (See Table 1 for details.) The tapping mode was used for the AFM analysis. The samples were prepared by placing drops of the Ag particles in ethanol on an atomically flat, freshly cleaved mica substrate surface. The 2D profile image shows the even thickness of the platelets.

tabular crystals on this plane as shown in the HRTEM micrograph (Figure 4a). This proposed mechanism is also consistent with the relationships between platelet face size (a) and platelet thickness (t) to number of moles of silver ion (nAg+) and platelet thickness (t) to bilayer thickness (d1) as discussed below. A linear relationship between the platelet face size (a) and number of moles of silver ions (nAg+) is observed (Figure 7). This suggests that lateral growth is linearly related to nAg+. This conclusion will be tested in current studies by varying the concentration of the Ag+ in the bilayer. Lateral growth should be increased while platelet thickness should remain constant for a given R value. In contrast, thickness of the platelets is parabolically related to (nAg+) (Figure 8), which may indicate a diffusion mechanism is the ratelimiting step for growth perpendicular to the amphiphilic surfaces. The bilayer thickness (d1) is the controlling variable for the platelet thickness (t) as shown in Figure 9, which gives a linear correlation between t and d. On the basis of the linear correlation, the platelet thickness is about 84% of the aqueous layer thickness. To put this into perspective, for the 2 nm thick platelets, there is a monolayer of water molecules on each face for the overall aqueous layer thickness of 2.7 nm. The face area of the newly formed structures change due to the variation with respect to R. There is a linear relationship between the face size (a) and R (Figure 10)

Figure 6. Variation of platelet thickness (t) as a function of face diameter (a) and R ([water]/[octylamine]) value. The polynomial fit is expressed with an equation that is given in the box. The two lines around the polynomial fit represent the confidence bands with 95% confidence interval. Note that the platelet face diameter (a) to platelet thickness (t) aspect ratio decreases by increasing R.

8698


Figure 7. Platelet face size (a) as a function of amount of cation used. (Number of moles of Ag+ ) n(Ag+).). The linear fit is expressed with an equation that is given in the box. The two lines around the linear fit represent the confidence bands with 95% confidence interval.

Yener et al.

Figure 9. Aqueous layer thickness (d1) vs platelet thickness (t) plot. These experimental results were obtained from AFM and XRD. The linear fit is expressed with an equation that is given in the box. The two lines around the linear fit represent the confidence bands with 95% confidence interval.

Figure 8. Platelet thickness (t) as a function of moles of silver used (n(Ag)+). The polynomial fit is expressed with an equation that is given in the box. The two lines around the polynomial fit represent the confidence bands with 95% confidence interval.

Figure 10. Platelet face size (a) as a function of R. The linear fit is expressed with an equation that is given in the box. The two lines around the linear fit represent the confidence bands with 95% confidence interval.

which is similar to the relationship between the face size (a) and [Ag+] (Figure 7). This correlation between Figure 7 and Figure 10 is due to the relation between [Ag+] and R. While R is increased, [Ag+] increases because the water amount in R is equal to the amount of aqueous Ag stock solution used. The wide size distribution as seen in TEM and AFM images results from the homogeneous nucleation and growth that forms over the surfaces of the amphiphilic molecules. Since the water in the microemulsion system came from the aqueous silver solution, the increase in the water content also increased the cation content in the system. This change led to the formation of larger platelets in a limited space.

amphiphilic octylamine and amylamine bilayers at a 25:1 ratio, respectively. The thickness of the aqueous layers in the bilayers increased with increasing water/surfactant ratios, R, and was shown to directly influence the thickness and face diameter of platelets formed in the bilayers. Variations in R from 1.7 to 10.3 lead to the formation of particles with face diameters ranging from 12 to 110 nm, with a corresponding thickness range from 0.3 to 5 nm. Thus, the thickness of the platelets can be controlled by adjusting the thickness of the aqueous layers. The thickness of the platelets is about 84% that of the aqueous layers. HRTEM and XRD diffraction patterns also verified that the synthesized platelets were phase pure silver demonstrating that the synthesis of nanosized phase pure silver with controlled thickness and face diameter is possible in the octylamine/amylamine self-assembly system with hydrazine hydrate as the reducing agent. However, the particles were not templated by the polar headgroup to produce oriented single crystals as originally

Conclusions Nanosized Ag platelets have been synthesized in selfassembled lamellar bilayers at room temperature. Thin lamellar aqueous layers were produced with the use of


hypothesized by Adair et al.14a It is suggested that the polycrystalline particles that show a variety of closepacked habit planes are created by aggregation of nuclei (36) (a) Ostwald, W. Lehrbruck der Allgemeinen Chemie; Leipzig, Germany, 1896; Vol. 2, Part 1, p 17. (b) Ostwald, W. Z. Phys. Chem. 1897, 22, 289. (c) Boistelle, R.; Astier, J. P. J Cryst. Growth 1988, 90, 14. (37) (a) Yim, H.; Foster, M. D.; Engelking, J.; Menzel, H.; Ritcey, A. M. Langmuir 2000, 16, 9792.

Langmuir, Vol. 18, No. 22, 2002 8699

due to Ostwald coarsening36 while the single-crystal particles are templated but with a range of habit planes. A possible reason for particle formation without any particular texture is the mobility of the amphiphilic layers. It is likely that the amphiphilic layers are too mobile at 20 °C to produce a templated collection of particles with the same habit.37 LA011229A