Pd Bimetallic Nanoparticles in Reverse Micelles

The characterization and formation kinetics of particles suggested that (1) most of AuCl4- and .... The Journal of Physical Chemistry C 0 (proofing), ...
17 downloads 0 Views 292KB Size
Download PDF
Langmuir 2001, 17, 3877-3883

3877

Synthesis of Au/Pd Bimetallic Nanoparticles in Reverse Micelles Ming-Li Wu, Dong-Hwang Chen,* and Ting-Chia Huang Department of Chemical Engineering, National Cheng Kung University, Tainan, Taiwan 701, Republic of China Received January 10, 2001. In Final Form: March 14, 2001 The synthesis of Au/Pd bimetallic nanoparticles at various molar ratios by the coreduction of HAuCl4 and H2PdCl4 with hydrazine in the reverse micelles of water/sodium bis(2-ethylhexyl) sulfosuccinate/ isooctane at 25 °C was investigated. Transmission electron microscopy photographs showed that the Au/Pd bimetallic nanoparticles obtained at the feeding compositions of [HAuCl4]/[H2PdCl4] ) 9/1 to 1/9 essentially had uniform sizes and mean diameters of about 3 nm. Both the UV-vis absorption spectra and X-ray diffraction patterns for the bimetallic systems and the physical mixtures of individual metallic nanoparticles suggested the formation of bimetallic nanoparticles. The energy dispersive X-ray analysis on the single particle showed that the composition for each particle was roughly consistent with that of feeding solution. The X-ray photoelectron spectroscopy measurement revealed that Pd atoms were significantly enriched in the outer part of Au/Pd bimetallic nanoparticles. The characterization and formation kinetics of particles suggested that (1) most of AuCl4- and PdCl42- ions were reduced before the formation of nuclei, (2) the nucleation rate of Au was much faster than that of Pd, (3) the size of bimetallic nanoparticles was determined by the number of the nuclei formed at the very beginning of reaction, and (4) the nuclei for the bimetallic system might be composed of Au and Pd. According to these suggestions, a formation process of Au/Pd bimetallic nanoparticles was proposed.

Introduction Nanoparticles may possess unusual physical and chemical properties which are quite different from those of relatively larger particles of the same materials due to their extremely small size and large specific surface area.1-3 They are expected to have many potential applications in optoelectronics, semiconductors, catalysts, photocatalysts, magnetic materials, drug delivery, and so on.4-6 Bimetallic nanoparticles are attractive because of their improvement in the catalytic properties7 and the change in the surface plasma band energy8 relative to the separate metals. Many methods have been reported on their preparation, including alcohol reduction,7,9 citrate reduction,8b,10 polyol process,11 solvent extraction reduction,8a,12 * To whom correspondence should be addressed. Tel: 886-62757575 Ext. 62680. Fax: 886-6-2344496. E-mail: chendh@ mail.ncku.edu.tw. (1) (a) Ozin, G A. Adv. Mater. 1992, 4, 612. (b) Cahn, R. W. Nature 1992, 359, 591. (c) Ozin, G. A. Science 1996, 271, 920. (2) Fendler, J. H. Nanoparticles and nanostructured films: preparation, characterization and applications; Wiley-VCH: Weinhein, 1998. (3) Schimid, G. Clusters and Colloids: From Theory to Application; VCH: Weinhein, 1994. (4) Toshima, N.; Yonezawa, T. New J. Chem. 1998, 1179. (5) (a) Kamat, P. V. Chem. Rev. 1993, 93, 267. (b) Lewis, L. N. Chem. Rev. 1993, 93, 2693. (c) Gates, B. C. Chem. Rev. 1995, 95, 511. (6) (a) Hoffman, A. J.; Mils, G.; Yee, H.; Hoffman, M. R. J. Phys. Chem. 1995, 99, 4414. (b) Lee, A. F.; Baddeley, C. J.; Hardacre, C.; Ormerod, R. M.; Lambert, R. M.; Schmid, G.; West, H. J Phys. Chem. 1995, 99, 6096. (7) (a) Toshima, N.; Yonezawa, T.; Kushihashi, K. J. Chem. Soc., Faraday Trans. 1993, 89, 2537. (b) Toshima, N.; Harada, M.; Yamazaki, Y.; Asakura, K. J. Phys. Chem. 1992 96, 9927. (c) Wang, Y.; Toshima, N. J. Phys. Chem. B 1997, 101, 5301. (8) (a) Han, S. W.; Kim, Y.; Kim, K. J. Colloid Interface Sci. 1998, 208, 272. (b) Link, S.; Wang, Z. L.; El-Sayed, M. A. J. Phys. Chem. B 1999, 103, 3529. (9) (a) Toshima, N.; Wang, Y. Langmuir 1994, 10, 4574. (b) Yonezawa, T.; Toshima, N. J. Mol. Catal. 1993, 83, 167. (c) Harada, M.; Asakura, K.; Ueki, Y.; Toshima, N. J. Phys. Chem. 1993, 97, 10742. (d) Yonezawa, T.; Toshima, N. J. Chem. Soc., Faraday Trans. 1995, 91, 4111.

sonochemical method,13 photolytic reduction,14 decomposition of organometallic precursors,15 and electrolysis of bulk metal.16 Their sizes are usually controlled by the addition of protective agents such as soluble polymers, surfactants, and organic ligands. The size, structure, and composition distribution of the resultant particles were all affected by the preparation conditions. Reverse micellar solutions are transparent, isotropic, thermodynamically stable water-in-oil microemulsions with nanosized water droplets which are dispersed in a continuous oil phase and stabilized by surfactant molecules at the water/oil interface. The surfactant-stabilized water pools not only act as microreactors for processing reactions but also inhibit the excess aggregation of particles because the surfactants could adsorb on the particle surface when the particle size approaches that of water pool. Consequently, the particles obtained in such a medium are generally very fine and uniform. A number of nanoparticles have been prepared in reverse micelles, including metals,17-21 metal oxides and hydroxides,22-26 metal sulfides and selenides,27-31 metal borides,32 metal (10) Miner, R. S.; Namba, S.; Turkevich, J. In Proc. Int. Cong. Catal., 7th Tokyo, 1981. (11) Silvert, P.-Y.; Vijayakrishnan, V.; Vibert, P.; Herrera-Urbina, R.; Elhsissen, K. T. Nanostruct. Mater. 1996, 7, 611. (12) Esumi, K.; Shiratori, M.; Ishizuka, H.; Tano, T.; Torigoe, K.; Meguro, K. Langmuir 1991, 7, 457. (13) Mizukoshi, Y.; Okitsu, K.; Maeda, Y.; Yamamoto, T. A.; Oshima, R.; Nagata, Y. J. Phys. Chem. B 1997, 101, 7033. (14) Remita, S.; Mostafavi, M.; Delcourt, M. O. Radiat. Phys. Chem. 1996, 47, 275. (15) (a) Bradley, J. S.; Hill, E. W.; Klein, C.; Chaudret, B.; Duteil, A. J. Chem. Mater. 1993, 5, 254. (b) Pan, C.; Dassenoy, F.; Casanove, M. J.; Philippot, K.; Amiens, C.; Lecante, P.; Mosset, A.; Chaudret, B. J. Phys. Chem. B 1999, 103, 10098. (16) (a) Reetz, M. T.; Helbig, W.; Quaiser, S. A. J. Chem. Mater. 1995, 7, 2227. (b) Reetz, M. T.; Quaiser, S. A. Angew. Chem., Int. Ed. Engl. 1995, 34, 2240. (17) Boutonnet, M.; Kizling, J.; Stenius, P.; Maire, G. Colloids Surf. 1982, 5, 209. (18) Petit, C.; Lixon, P.; Pilei, M.-P. J. Phys. Chem. 1993, 97, 12974. (19) Qi, L.; Ma, J.; Shen, J. J. Colloid Interface Sci. 1997, 186, 498.

10.1021/la010060y CCC: $20.00 © 2001 American Chemical Society Published on Web 06/02/2001

3878

Langmuir, Vol. 17, No. 13, 2001

carbonates,33 and organic polymer.34 However, the preparation of bimetallic nanoparticles in reverse micelles has not been tried as much except for the Pt/Pd35 and Cu/Au36 systems. Au/Pd bimetallic nanoparticles have received significant attention due to their special catalytic properties. Their preparation has been achieved by some methods such as alcohol reduction,7b,37 sonochemical method,38 citrate reduction,39 and vacuum vapor deposition technique.40 However, they have not been prepared in reverse micelles until now. In this paper, the synthesis of Au/Pd bimetallic nanoparticles in the reverse micelles of water/sodium bis(2-ethylhexyl) sulfosuccinate (AOT)/isooctane by the coreduction of HAuCl4 and H2PdCl4 with hydrazine at 25 °C was studied. The size, structure, optical properties, and composition distribution of the resultant nanoparticles were characterized by transmission electron microscopy (TEM), UV-vis spectroscopy, X-ray diffraction (XRD), energy-dispersive X-ray analysis (EDX), and X-ray photoelectron spectroscopy (XPS). The effects of preparation conditions including the molar ratio of water to surfactant and the composition of feeding solution were investigated. The formation process of bimetallic nanoparticles was also discussed. Experimental Section Materials. Palladium(II) chloride, hydrogen tetrachloroaurate(III) hydrate, hydrochloric acid solution, and hydrazinium hydroxide were the guaranteed reagents of E. Merck (Darmstadt). Sodium bis(2-ethylhexyl) sulfosuccinate (AOT) purchased from Sigma Chemical Co. (St. Louis, MO) was vacuum-dried at 60 °C for 24 h and stored in a vacuum desiccator prior to use. HPLCgrade isooctane supplied by TEDIA (Fairfield) was dehydrated with 4 Å molecular sieves (8-12 mesh, Janssen) for at least 24 h and kept in a vacuum desiccator prior to use. The residual water of the AOT/isooctane solution was recognized to be negligible by a Karl-Fisher moisture titrator (Kyoto Electronics (20) (a) Chen, D. H.; Wang, C. C.; Hung, T. C. J. Colloid Interface Sci. 1999, 210, 123. (b) Chen, D. H.; Yeh, J. J.; Hung, T. C. J. Colloid Interface Sci. 1999, 215, 159. (21) Chen, D. H.; Wu, S. H. Chem. Mater. 2000, 12, 1354. (22) Chang, C. L.; Fogler, H. S. Langmuir 1997, 13, 3295. (23) Osseo-Asare, K.; Arriagada, F. J. Colloid Interface Sci. 1990, 50, 321. (24) Pillai, V.; Kumar, P.; Multani, M. S.; Shah, D. O. Colloid Surf., A 1993, 80, 69. (25) Joselevich, E.; Willner, I. J. Phys. Chem. 1994, 98, 7628. (26) Chhabra, V.; Lal, M.; Maitra, A. N.; Ayyub, P. Colloid Polym. Sci. 1995, 273, 939. (27) Lianos, P.; Thomas, J. K. J. Colloid Interface Sci. 1987, 117, 505. (28) Kortan, A. R.; Hull, R.; Opila, R. L.; Bawendi, M. G.; Steigerwald, M. L.; Carroll, P. J.; Brus, L. E. J. Am. Chem. Soc. 1990, 112, 1327. (29) Ward, A. J.; O’Sullivan, E. C.; Rang, J.-C.; Nedeljkovic, J.; Patel, R. C. J. Colloid Interface Sci. 1993, 161, 316. (30) Hirai, T.; Shiojiri, S.; Komasawa, I. J. Am. Chem. Eng. Jpn. 1994, 27, 590. (31) Haram, S. K.; Mahadeshwar, A. R.; Dixit, S. G. J. Phys. Chem. 1996, 100, 5868. (32) Nagy, J. J. Colloids Surf. 1989, 35, 201. (33) Kandori, K.; Kon-No, K.; Kitahara, A. J. Colloid Interface Sci. 1988, 122, 78. (34) Antonietti, M.; Basten, R.; Lonmann, S. Macromol. Chem. Phys. 1995, 196, 441. (35) Touroude, R.; Girard, P.; Maire, G.; Kizling, J.; BoutonnetKizling, M.; Stenius, P. Colloids Surf. 1992, 67, 9. (36) Sangregorio, C.; Galeotti, M.; Bardi, U.; Baglioni, P. Langmuir 1996, 12, 5800. (37) (a) Harada, M.; Asakura, K.; Ueki, Y.; Toshima, N. J. Phys. Chem. 1993, 97, 5103. (b) Liu, H.; Mao, G.; Meng, S. J. Mol. Catal. 1992, 74, 275. (38) Mizukoshi, Y.; Okitsu, K.; Maeda, Y.; Yamamoto, T. A.; Oshima, R.; Nagata, Y. J. Phys. Chem. B 1997, 101, 7033. (39) Schmid, G.; West, H.; Malm, J.-O.; Bovin, J.-O.; Grenthe, C. Chem. Eur. J. 1996, 2, 1099. (40) Deki, S.; Akamatsu, K.; Hatakenaka, Y.; Mizuhata, M.; Kajinami, A. Nanostruct. Mater. 1999, 11, 59.

Wu et al. MKC-50). The water used throughout this work was the reagentgrade water produced by Milli-Q SP ultrapure-water purification system of Nihon Millipore Ltd., Tokyo. The aqueous solutions of H2PdCl4 and HAuCl4 were prepared by dissolving palladium(II) chloride in 0.2 N HCl solution and hydrogen tetrachloroaurate(III) hydrate in water, respectively. The reverse micellar solutions containing hydrazine or HAuCl4/ H2PdCl4 mixture at a specified molar ratio were prepared by injecting the required amounts of the corresponding aqueous solution into the AOT/isooctane solution and then were used for the preparation of nanoparticles within a few minutes. Preparation of Nanoparticles. The preparation of monometallic and bimetallic nanoparticles was achieved by mixing equal volumes of two reverse micellar solutions at the same molar ratio of water to AOT (ω0), one containing an aqueous solution of metal salts and the other containing an aqueous solution of hydrazine. The reductions of H2PdCl4 and HAuCl4 were

2H2PdCl4 + N2H5OH f 2Pd + 8HCl + N2 + H2O (1) 4HAuCl4 + 3N2H5OH f 4Au + 16HCl + 3N2 + 3H2O (2) The preliminary study indicated that the particles reached their final sizes within 1 h. Accordingly, the samples for various analyses were taken after 3 h. In this work, the AOT concentration based on the overall volume of reverse micellar solution was fixed at 0.1 M. The concentrations of total metal ions and hydrazine were kept at 0.1 and 1.0M, respectively, on the basis of the volume of aqueous solution added in the reverse micellar solution. The temperature was fixed at 25 °C and, unless otherwise specified, the ω0 value was 6.0. Characterization. The particle sizes were determined by TEM using a JEOL model JEM-1200EX at 80 kV. The sample for TEM analysis was prepared by placing a drop of colloidal solution onto the Formvar-covered copper grid and evaporating it in air at room temperature. For each sample, usually over 100 particles from different parts of the grid were used to estimate the mean diameter and size distribution of particles. The samples were also used to determine the elemental ratios of particles by EDX analysis with a Noran model Voyager 1000 system attached to a Hitachi model HF-2000 field emission transmission electron microscope. XRD measurements were performed on a Rigaku D/max III.V X-ray diffractometer using Cu KR radiation (λ ) 0.1542 nm). The samples for XRD analysis were prepared by first adding thiophenol to the reverse micellar solutions (i.e., [thiophenol]/[metal] ) 2/1) to cause phase separation, then centrifuging the mixtures, washing the precipitates with isooctane and ethanol, and finally drying the obtained precipitates at room temperature. After the precipitates were redispersed in 1,2-dichloroethane and the thiophenol-capped particles were deposited on a graphite support by solvent evaporation at room temperature, the obtained samples were further used for the XPS measurements by the Fison (VG) ESCA 210 spectrometer equipped with a Mg KR X-ray source. The UV-vis spectra of the reverse micellar solutions containing various nanoparticles were measured after 3 h by a Hitachi U-3000 spectrophotometer with a 10 mm quartz cell.

Results and Discussion Particle Size Analysis. It has been known that the size of reverse micelles increased upon increasing the ω0 value. As indicated in Figure 1, the sizes of Au, Pd, and Au/Pd(1/1) bimetallic nanoparticles all also increased with the increase of ω0 value. This could be reasonably interpreted by the fact that the surfactant molecules might adsorb on the surface of the particle formed therein and restricted the growth of nanoparticles. The conditions for ω0 > 12 were not investigated because the aqueous solution of metal salts could not be solubilized completely into the AOT/isooctane solutions. In addition, it was found that Au/Pd(1/1) bimetallic nanoparticles were significantly smaller than both Au and Pd nanoparticles over the whole

Synthesis of Au/Pd Bimetallic Nanoparticles

Figure 1. Effect of ω0 value on the size of Au, Pd, and Au/Pd bimetallic nanoparticles synthesized in water/AOT/isooctane reverse micelles. [total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; [AOT] ) 0.1 M; reaction time ) 3 h.

ω0 range investigated. This might be referred to the nucleation process of particles and will be further discussed later. Figure 2 shows the typical TEM photographs and the size distributions of nanoparticles obtained at various molar ratios of Au/Pd. It revealed that Au, Pd, and the Au/Pd bimetallic nanoparticles essentially all were very fine and uniform. From their mean diameters as illustrated in Figure 3, it was obvious that the sizes of bimetallic nanoparticles were significantly smaller than those of individual metallic nanoparticles. The uniform and smaller sizes of bimetallic systems implied that they were not the physical mixtures of individual metallic nanoparticles and bimetallic nanoparticles were really formed. Moreover, Yonezawa and Toshima found that the mean diameters of bimetallic nanoparticles depended on the compositions of feeding solution and exhibited a negative deviation for the Au/Pd, Au/Pt, and Pd/Pt systems synthesized by alcohol reduction,9d while Esumi et al. observed a positive deviation for the Pd/Pt system prepared by solvent extraction reduction.12 Although these results were not consistent completely, they all could be attributed to the difference in the nucleation process which was dependent on the kinds of materials, the reaction media, and the preparation conditions. As has been known, except for the collision energy and the sticking coefficient, the rates of nucleation and growth were determined mainly by the probabilities of the collisions between several atoms, between one atom and a nucleus, and between two or more nuclei. The former kind of collision related to the nucleation and the latter two kinds of collision to the growth process. When the reduction rate was so large that most ions were reduced before the formation of nuclei and the probability of the effective collision between one atom and a nucleus were much higher than those of the other two collisions, the size of the resultant particles would be uniform and determined by the number of the nuclei formed at the very beginning of reaction. Thus, in this work, most of AuCl4- and PdCl42- ions might be reduced before the formation of nuclei. Also, the nucleation process determined the sizes of Au, Pd, and Au/Pd bimetallic nanoparticles. For a fixed concentration of total metal ions,

Langmuir, Vol. 17, No. 13, 2001 3879

the smaller sizes of Au/Pd bimetallic nanoparticles than the individual Au and Pd nanoparticles revealed more nuclei were formed at the very beginning of the reaction. This seemed to imply the number of atoms required to form a nucleus was dependent on the composition. It is known that the bond dissociation enthalpies for every dissociation step may be different during the atomization of a molecule. Similarly, considering the formation of a nucleus, the bond enthalpy also might change during the assembly of atoms and be different from the mean value for the bulk material. Thus, the total Gibbs energy for the formation of a nucleus might vary with the compositions due to the different bond enthalpies or interactions of AuAu, Pd-Pd, and Au-Pd and hence led to the number of atoms required for the formation of a nucleus, which are different at different compositions. Accordingly, it was suggested that the nuclei for the formation of Au/Pd bimetallic nanoparticles in this work might be composed of Au and Pd atoms. Another possible condition is that the nuclei were composed of only Au atoms but the number of Au atoms required to form a stable nucleus was reduced due to the presence of the Pd atoms in the solution. However, Pd atoms should have a stronger influence on the total Gibbs energy of an atomic assembly when they were present in a nucleus than when they were just present in the solution. On the basis of the reason mentioned above and the probability of the atomic assembly, the former suggestion that the nuclei might be composed of both Au and Pd atoms is favored. UV-vis Spectra. Figure 4a shows a series of the UVvis spectra for the physical mixtures of Au and Pd monometallic nanoparticles. The absorption peak around 520 nm was the surface plasma absorption due to Au, and its area increased with the increase of the Au/Pd ratio. The UV-vis spectra for the bimetallic systems, as illustrated in Figure 4b, showed no significant surface plasma absorption except that the Au/Pd(9/1) system exhibited a very small absorption peak. This phenomenon was consistent with the previous observation that the presence of a group 10 metal (d8s2) in the bimetallic nanoparticles suppressed the surface plasma energies of group 11 metal (d10s1)4,37 and implied the formation of Au/Pd bimetallic nanoparticles. Furthermore, the sudden disappearance of the surface plasma absorption of Au implied that the surface of the Au/Pd bimetallic nanoparticles obtained in this work had more Pd atoms than the inner core. The composition distribution will be further investigated by the analyses of EDX and XPS later. Particle Structure. As described in the Experimental Section, the thiophenol-capped nanoparticles were prepared for XRD analysis. It was found that the particles could be well redispersed in a mixture of pyridine and tetrahydrofuran and essentially were similar to original nanoparticles in size. This ensured that the particle structure was not affected due to the sample preparation. Figure 5a shows the XRD patterns of Au, Pd, and the physical mixture of Au and Pd (1:1), Au/Pd (1/1), bimetallic nanoparticles. The characteristic peaks for Au (2θ ) 38.2, 44.4) and those for Pd (2θ ) 40.1, 46.7), marked by their indices ((111), (200)), revealed that both the resultant Au and Pd particles essentially were face-centered cubic (fcc) structure. Also, the broad peak indicated the particles obtained were poor crystalline due to less ordered structures as usually observed for nanoparticles. For the physical mixture of Au and Pd nanoparticles, both the characteristic peaks of Au (2θ ) 38.2) and Pd (2θ ) 40.1) were observed without shift. However, only one broad peak between those of Au and Pd was shown for the Au/Pd(1/1) bimetallic system. This suggested the formation of bi-

3880

Langmuir, Vol. 17, No. 13, 2001

Wu et al.

Figure 2. Transmission electron micrographs and particle size distributions of the individual metallic and bimetallic nanoparticles: (a) Au; (b) Au/Pd (3/1); (c) Au/Pd (1/1); (d) Au/Pd (3/1); (e) Pd. [total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; ω0 ) 6; AOT ) 0.1 M; reaction time ) 3 h.

metallic nanoparticles. In addition, Figure 5b indicates a continuous drift of the peak from that of Au to that of Pd. This phenomenon revealed that the composition of each bimetallic particle was proportional to that of feeding solution. Composition Analysis. For each molar ratio of Au/ Pd, four particles on the copper grid were chosen randomly to analyze the composition in each particle by EDX attached to a high-resolution TEM. The result is indicated in Table 1. The deviation of composition among different particles at each [HAuCl4]/[H2PdCl4] ratio might result from the detection errors because the particles were too

fine. In addition, it was observed that the average compositions were roughly in agreement with those of feeding solutions. The deviation could be due to the detection errors and the fact that fewer particles was chosen for analysis. In any case, the EDX analysis provided direct and powerful evidence for the formation of Au/Pd bimetallic nanoparticles. To further investigate the composition distribution in a particle, the compositions in the outer part of bimetallic nanoparticles were measured. The XPS spectra of the Au4f and Pd3d regions of the thiophenol-capped nanoparticles are shown in Figure 6. The binding energy of Pd3d3 showed

Synthesis of Au/Pd Bimetallic Nanoparticles

Langmuir, Vol. 17, No. 13, 2001 3881

Table 1. Elemental Ratio in an Au/Pd Bimetallic Nanoparticle Chosen Randomly on the Copper Grid from the EDX Analysisa

a

feeding solution [HAuCl4]/[H2PdCl4]

1

3/1 1/1 1/3

71.6/28.4 48.8/51.2 29.8/70.2

elemental ratio (Au/Pd) of each particle 2 3 62.2/37.8 63.3/36.7 29.9/70.1

73.6/26.7 48.2/51.8 21.4/78.6

4

av

75.1/24.9 44.2/55.8 24.2/75.8

70.6/29.4 51.1/48.9 26.3/73.7

[total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; ω0 ) 6; [AOT] ) 0.1 M; reaction time ) 3 h.

Figure 3. Mean diameters of Au/Pd bimetallic nanoparticles as a function of composition in feeding solution. [total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; ω0 ) 6; [AOT] ) 0.1 M; reaction time ) 3 h.

two peaks at 340.6 and 341.9 eV with respect to the C1s peak. The former was equal to that for the uncapped Pd and could be referred to the Pd in zero valent state. The latter decreased with the increase of Au/Pd ratio and should result from the formation of Pd thiolate. Au thiolate should also be formed on the surface of Au and Au/Pd particles. However, the binding energy of Au4f7 for Au thiolate was very close to that for the uncapped Au.8a,41 So, only one peak was observed around 84.0 eV for Au and Au/Pd particles. According to the intensities of XPS peaks, the elemental ratios of Au/Pd in the outer part of bimetallic nanoparticles could be obtained. The results are summarized in Table 2. It was obvious that the Pd atoms were enriched in the outer part of Au/Pd bimetallic nanoparticles, suggesting a structure of incomplete Au-core/ Pd-shell. This was consistent with the observation of UVvis absorption spectra. It was also noticed that the composition in the outer part of bimetallic nanoparticles varied linearly with the composition in the feeding solution within the composition range of [HAuCl4]/[H2PdCl4] ) 3/1 to 1/9 (see Supporting Information). The control of composition distribution by feeding composition should be interesting and important for the preparation of catalysts. Formation Process of Particles. For the preparation of Au nanoparticles, it was observed that the color of solution turned from yellow to red instantaneously. The Au nanoparticles reached their final sizes (4.2 ( 1.0 nm, see Supporting Information) so quickly that the required time could not be detected by conventional methods. (41) Chastain, J.; King, R. C. Handbook of X-ray Photoelectron Spectroscopy; Physical Electronics: Eden Prairie, MN, 1995.

Figure 4. UV-vis absorption spectra of the physical mixtures of individual Au and Pd nanoparticles (a) and Au/Pd bimetallic nanoparticles (b) at various molar ratios. [total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; ω0 ) 6; [AOT] ) 0.1 M; reaction time ) 3 h.

Therefore, a stopped-flow spectrophotometer was used to measure the formation rate of Au nanoparticles. The absorbance at 520 nm (characteristic peak of Au nanoparticles) increased first and then approached to a constant value after about 4 s (see Supporting Information). This

3882

Langmuir, Vol. 17, No. 13, 2001

Figure 5. XRD patterns of the physical mixtures of individual Au and Pd nanoparticles as well as the Au/Pd bimetallic nanoparticles at various molar ratios. [total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; ω0 ) 6; [AOT] ) 0.1 M; reaction time ) 3 h.

implied that the formation of Au nanoparticles, including the reduction of AuCl4- and the nucleation and growth of particles, was finished within 4 s. For the case of Pd, when the reaction was initiated, the yellow-brown color of solution disappeared instantaneously and then the solution remained colorless and transparent for about 15 s. In the following 30 min, the color of solution turned to brown gradually and reached a stable state. The disappearance of yellow-brown color meant the complete reduction of PdCl42-. The first 15 s could be considered as the time for the nucleation of Pd particles, and the following 30 min was referred to the period for particle growth. This was consistent with the suggestion mentioned in the investigation of particle size that the

Wu et al.

Figure 6. XPS spectra of Au4f (a) and Pd3d (b) regions for the thiophenol-capped individual Au and Pd nanoparticles as well as Au/Pd bimetallic nanoparticles at various molar ratios. [total metal salts] ) 0.1 M; [N2H5OH] ) 1.0 M; ω0 ) 6; [AOT] ) 0.1 M; reaction time ) 3 h; [thiophenol]/[metals] ) 2/1. Table 2. Quantitative XPS Data of Thiophenol-Capped Au/Pd Bimetallic Nanoparticlesa feeding solution [HAuCl4]/[H2PdCl4]

elemental ratio (Au/Pd) in the outer part of particlesb

9/1 4/1 3/1 1/1 1/3 1/9

3.96/1 1.19/1 1/1.25 1/2.71 1/7.64 1/109.2

a [total metal salts] ) 0.1 M; [N H OH] ) 1.0 M; ω ) 6; [AOT] 2 5 0 ) 0.1 M; reaction time ) 3 h. b The elemental ratios of Au/Pd were estimated based on Au4f and Pd3d3 peaks.

reduction of PdCl42- ions was almost finished before the formation of nuclei. For the two cases of Au/Pd (1/1) and Au/Pd (1/9), the color of solutions turned from yellow to

Synthesis of Au/Pd Bimetallic Nanoparticles

Langmuir, Vol. 17, No. 13, 2001 3883

brown instantaneously. The kinetic data showed that Au/ Pd (1/1) nanoparticles reached their final size very quickly, but the Au/Pd (1/9) nanoparticles reach their final size after about 5 min (see Supporting Information). This indicated the nucleation and growth of Pd in the bimetallic system were accelerated, although they were still slower than those of Au. This could be attributed to the earlier formation of nuclei due to the presence of Au. Although it was not known whether the reduction of AuCl4- ions was faster than that of PdCl42- ions, their reduction potentials shown below revealed that AuCl4ions had the priority in reduction.4,42

AuCl4- + 3e- f Au + 4Cl-

1.00 V

(3)

PdCl42- + 2e- f Pd + 4Cl-

0.63 V

(4)

In addition, the above investigation on particle size suggested that most of the AuCl4- and PdCl42- ions were reduced before the formation of nuclei and the nuclei for the formation of Au/Pd bimetallic nanoparticles might be composed of Au and Pd atoms in this work. Thus, the formation process of Au/Pd bimetallic nanoparticles could be described as follows. First, most of the AuCl4- and PdCl42- were reduced in a very short time. In the meantime, Au and Pd atoms started to aggregate to form the nuclei and all nuclei might be formed almost at the same time. Since the nucleation rate of Au was much faster than that of Pd, the nuclei of the bimetallic system should be formed by the initiation of Au atoms and the composition of the nuclei might have higher Au content than that of the feeding solution. Then Au and Pd atoms codeposited (42) Fujishima, A.; Aizawa, M.; Inoue, T. Denki Kaqaku Sokuteiho; Gihodo Pupl. Co.: Tokyo, 1984; Vol. 2.

onto the nuclei and grew to their final sizes. The faster deposition rate of Au than Pd led to the enrichment of Pd in the outer part of Au/Pd bimetallic nanoparticles. Conclusions This study proposed a practical method for the preparation of Au/Pd bimetallic nanoparticles. The TEM, UVvis, and XRD analyses all suggested the formation of bimetallic nanoparticles. The EDX analysis on a single particle directly confirmed this fact and indicated the compositions of bimetallic nanoparticles were roughly consistent with those of feeding solutions. The XPS data showed the resultant bimetallic nanoparticles had a structure with enriched Pd atoms in the outer part of Au/Pd bimetallic nanoparticles, and their outer part compositions could be adjusted by feeding compositions. According to the size, composition distribution, and kinetic analyses of particles, a formation process of Au/Pd bimetallic nanoparticles could be suggested. This was helpful for the clarification of the formation process of Au/Pd bimetallic nanoparticles in reverse micelles. Acknowledgment. This work was performed under the auspices of the National Science Council of the Republic of China, under Contract Number NSC 89-2214-E006021, to which the authors wish to express their thanks. Supporting Information Available: Three figures showing (1) the relationship between Pd content in the outer part of particles and that in feeding solution, (2) the variation of mean diameter with reaction time for Pd, Au, Au/Pd (1/1), and Au/Pd (1/9), and (3) the absorbance change with reaction time at a wavelength of 520 nm measured by a stopped-flow spectrophotometer. This material is available free of charge via the Internet at http://pubs.acs.org. LA010060Y