Structural Transformation of Au-Pd Bimetallic Nanoclusters on Thermal Heating and Cooling: A Dynamic Analysis H. B. Liu,* ,† U. Pal,* ,‡ R. Perez, and J. A. Ascencio Programa de InVestigacio ´ n en Ductos, Materiales y Corrosio ´ n, Instituto Mexicano del Petro ´ leo, Eje Central La ´ zaro Ca ´ rdenas No. 152, Col. San Bartolo Atepehuacan, C.P.07730, Me ´ xico D.F., Mexico, and Instituto de ´sica, UniVersidad Auto ´ noma de Puebla, Apdo. Postal J-48, Puebla, Pue. 72570, Mexico ReceiVed: October 22, 2005; In Final Form: January 13, 2006 Classical molecular dynamics simulation is used for structural thermodynamic and dynamic analysis of Au- Pd bimetallic clusters. It is observed that the Pd-core/Au-shell structure is the most stable, and can be formed through annealing of other structures such as Au-core/Pd-shell, eutecticlike, or solid solution. Depending on the starting temperature and initial composition, three-layer icosahedral nanorod, face-centered cubic (fcc) nanorod, and fcc cluster can be obtained on slow cooling. The three-layer icosahedral nanorod structure is not as stable as the Pd-core/Au-shell decahedron; however it is more stable than the solid-solution decahedron structure up to 400 K. Our findings provide valuable insight into catalysis using Au-Pd and other similar bimetallic clusters. I. Introduction Bimetallic nanoclusters have received considerable attention recently 1 for their unique properties, which are very different from those of pure clusters of their constituents 2,3 and for a wide range of technological applications, ranging from catalysis to optics. 4,5 The properties of bimetallic nanoclusters vary not only with their size but also with their chemical compositions. Recent success in application of bimetallic nanoclusters has generated great interest on their design, synthesis, and characterization. 6-9 Among the bimetallic nanoclusters, Au-Pd is the most attractive catalyst for the hydrodesulfurization of thiophene, 10 direct synthesis of hydrogen peroxide from H 2 and O 2, etc. 11-13 Recently they have been developed for nanocontact applica- tions. 14 Catalytic performance of Au-Pd bimetallic nanoclusters benefits to a great extent from their structural diversities, such as alloy cluster, 15-22 wire, 23 lithography, 24 and core/shell. 25-27 From the point of view of the phase diagram, Au-Pd is likely to form unlimited solid solutions at low temperatures. 28 The dilute-limit heat of solution of Au in Pd host is about -0.33 eV and of Pd in Au host is about -0.29 eV, 29 which are consistent with the formation of a variety of bimetallic structures. While the size, shape, composition, and atomic distribution of the nanoclusters are the key parameters for catalyst design, studies of structural dynamics and thermodynamic behavior are essential for their applications. In our previous work, 30 structural incoherency and thermodynamic stability of Au-Pd nanoclus- ters were studied systematically. In the present work, we report on thermal effects and dynamic analysis of structural transfor- mation of Au-Pd nanoclusters through molecular dynamics (MD) simulation. II. Research Methods A simple analytical embedded-atom model (EAM) developed by Cai et al. 31 is used to describe the interatomic interaction potential in the classical molecular dynamics simulation, which includes a long-range force. In this model, we considered an exponentially decreasing electron-density function, a two-body potential function as defined by Rose et al., 32 and assumed the embedding energy to be a universal form as suggested by Banerjee and Smith. 33 The alloy model of Johnson 34 is applied and an extra parameter is introduced in order to fit dilute-limit heats of solution. The predicted heats of formation for the three possible compounds, namely, AuPd 3 , AuPd, and Au 3 Pd are -0.06, -0.08, and -0.06 eV. The corresponding values from first principle calculations are -0.06, -0.11, and -0.10 eV, respectively. The reasonable agreement between the heats of formation ensures that the Au-Pd alloying potential can be used for a wide range of compositions with good accuracy. Therefore, all the potential forms and parameters in Cai’s work are used directly without any modification (see details in ref 31). MD simulations are performed using XMD developed by J. Riffkin. 35 The program employs a predictor-corrector algorithm to integrate the equation of motion and velocity rescaling constant temperature to control system temperature. A time step of 5 × 10 -15 seconds (5 fs) is used. III. Results and Discussion A. Structural Transformation during Heating. To study the thermal effect on Au-Pd nanocluster, 262-atom decahedral clusters with different Au/Pd compositions and a few structural types are constructed. The gyration diameters of these clusters are about 26 angstrom. These structures include solid solution clusters such as Au 131 Pd 131 and Au 163 Pd 99 , eutecticlike clusters such as Au 33 Pd 229 , Au 229 Pd 33 , Au 84 Pd 178 , and Au 178 Pd 84 , and core/shell structures such as Pd 99 -core/Au 163 -shell and Au 99 - core/Pd 163 -shell (see Figure 2). We must mention that the decahedral structures are the most common type of clusters observed in bimetallic Au-Pd colloids. 36 Figure 1 presents structural transformation for the Au-core/Pd-shell, eutecticlike, and solid solution clusters during heating process at a heating rate of 5 × 10 11 K/s. Though our heating rate is relatively high and may lead to some overheating causing a shift in transition * To whom correspondence should be addressed. E-mail: h_b_liu@ yahoo.com (H.B.L.); upal@venus.ifuap.buap.mx (U.P.). Fax: +52-222- 2295611. Instituto Mexicano del Petro ´leo. Universidad Auto ´noma de Puebla. 5191 J. Phys. Chem. B 2006, 110, 5191-5195 10.1021/jp056060e CCC: $33.50 © 2006 American Chemical Society Published on Web 02/22/2006