Abstract
While nanocatalysis is a very active field, there have been very few studies in the size/shapedependent catalytic properties of transition metals from a thermodynamical approach. Transition metal nanoparticles are very attractive due their high surface to volume ratio and their high surface energy. In particular, in this paper we focus on the PtPd catalyst which is an important system in catalysis. The melting temperature, melting enthalpy, and catalytic activation energy were found to decrease with size. The face centered cubic crystal structure of platinum and palladium has been considered in the model. The shape stability has been discussed. The phase diagram of different polyhedral shapes has been plotted and the surface segregation has been considered. The model predicts a nanoparticle core rich in Pt surrounded by a layer enriched in Pd. The Pd segregation at the surface strongly modifies the catalytic activation energy compared to the nonsegregated nanoparticle. The predictions were compared with the available experimental data in the literature.
PACS
65.80g; 82.60.Qr; 64.75.Jk
Introduction
Bimetallic nanoparticles exhibit unusual physicochemical properties different from those of the bulk material or their individual constituents [1,2]. They are very used in catalysis, fuel cells, and hydrogen storage. These unusual properties are determined by their size, shape, and composition. When considering metallic catalysts, platinum is a standard material but this material is most expensive than gold [3]. Therefore, to reduce the amount of platinum and then the cost of the application, one possible way is to use an alloy of platinum with another metal. In the present study, the chosen alloy is the binary PtPd system [4] that we propose to theoretically study from a thermodynamic approach [5,6], as well as its pure components. It has been shown previously [5,6] that thermodynamics may provide useful insights in nanotechnology where the size of the considered nanoparticles is higher than approximately 4 nm. Within this approach, the size and shape effects on the melting temperature, melting enthalpy, phase diagram, and catalytic activation energy of this system are investigated.
As facecentered cubic (fcc) metals, Pt and Pd can exhibit a variety of geometrical shapes. Therefore, to address the shape effect on the materials properties of these metals at the nanoscale [7,8], the following shapes have been considered: sphere, tetrahedron, cube, octahedron, decahedron, dodecahedron, truncated octahedron, cuboctahedron, and icosahedron.
Sizedependent melting properties of Pt and Pd
At the nanoscale, the melting temperature T_{m }and melting enthalpy ΔH_{m}, for freestanding nanostructures can be expressed as function of their bulk corresponding property, the size of the structure and one shape parameter [9].
where the shape parameter, α_{shape}, is defined as α_{shape }= AD(γ_{s}γ_{l})/(VΔH_{m, ∞}); D being the size of the structure (i.e. for a sphere, D is the diameter), A (meter squared) and V (cubic meter) are the surface area and volume of the nanostructure, respectively. ΔH_{m,∞ }is the bulk melting enthalpy (Joule per cubic meter), whereas γ_{l }and γ_{s }are the surface energy in the liquid and solid phases (Joule per square meter), respectively. γ_{l }and γ_{s }are considered size independent. This is justified by the fact that the size effect on the surface energies is less than 4% for sizes higher than 4 nm [10,11]. Indeed, below this size, edges, and corners of the structures begin to play a significant role in the surface energy [12].
The sizedependent melting temperatures of platinum and palladium are plotted in Figures 1 and 2 respectively. The materials properties of the considered materials are indicated in Table 1. The melting properties for the sphere have been calculated using for the solid surface energy the mean value of experimental data [13]. For the other polyhedra shapes, we have considered the fcc crystal structure of the metals and the respective solid surface energy for each face [14]. Tables 2 and 3 indicate the parameters used for the calculation of the melting properties. Experimentally, the melting of agglomerated Pt nanocrystals (tetrahedrons and cubes) with an average size around approximately 8 nm starts at approximately 900 K [15] in relative good agreement with our theoretical predictions. Molecular dynamics simulations [16] have calculated the size effect on the melting temperature of Pd and found α_{sphere }= 0.95 nm while our theory predicts 1.68 nm.
Discussion
At the nanoscale, the shape which exhibits the highest melting temperature is the one which minimizes the most the Gibbs' free energy (G = H  TS); and is then the favored one. From Figures 1 and 2, the four moststable shapes among the ones considered are the dodecahedron, truncated octahedron, icosahedron, and the cuboctahedron. Experimentally, truncated octahedron and cuboctahedron are observed for platinum nanoparticles [8] whereas icosahedron, decahedron, truncated octahedron and cuboctahedron are observed for palladium nanoparticles [8]. Therefore, our predictions are in relative good agreement with the observations for palladium and platinum except that dodecahedron and icosahedron are not observed for platinum. Other theoretical calculations confirmed that the dodecahedron is a stable shape for palladium [17]. More generally, according to Yacaman et al. [8], the most often observed shapes at the nanoscale are the cuboctahedron, icosahedron, and the decahedron.
Furthermore, care has to be taken when we compare theoretical results with experimental ones due those materials properties depend on the synthesis process [18,19]. And then predicted properties from thermodynamics may differ from the experimentally observed if the synthesis process is not running under thermodynamical equilibrium. Moreover, thermal fluctuations are often observed in nanoparticles [20] meaning that the shape stability is much more complicated than just a minimisation of the A/V ratio with faces exhibiting the lowest surface energy.
Nanophase diagram of PtPd
According to the HumeRothery's rules, platinum and palladium forms an ideal solution [21]. In this case, considering no surface segregation, the liquidus and solidus curves of bulk and nanostructures are calculated from the following equations [2224]:
where x_{solidus }(x_{liquidus}) is the composition in the solid (liquid) phase at a given T, respectively. is the sizedependent melting temperature of the element i. is the sizedependent melting enthalpy of the element i.
The phase diagram of the PtPd alloy is plotted in Figure 3. We note that the lens shape of the phase diagram is conserved at the nanoscale; however, the lens width increases for the shapes characterized by a small melting enthalpy and melting temperature, i.e., exhibiting a strong shape effect. Moreover, the melting temperature increases with the concentration of Pt in agreement with Ref. [25].
Figure 3. Phase diagram of the PtPd system for different shapes. Different shapes at a size equal to 4 nm and at the bulk scale. The solid lines indicate the liquidus curves whereas the dashed lines indicate the solidus ones.
In order to predict nanomaterials properties more accurately, we are considering a possible surface segregation which is known as the surface enrichment of one component of a binary alloy. At the nanoscale, surface segregation leads to a new atomic species repartition between the core and the surface. According to Williams and Nason [26], the surface composition of the liquid and solid phase are given by:
where z_{1 }is the first nearest neighbor atoms; z_{1ν }is the number of first nearest atoms above the same plane (vertical direction). In the case of facecentered cubic (fcc) crystal structure of Pt and Pd materials, we have z_{1 }= 12, z_{1ν }= 4 for (100) faces and three for (111) faces. ΔH_{vap }is the difference between the bulk vaporization enthalpies of the two pure elements, . ΔH_{sub }is the difference between the bulk sublimation enthalpies of the two pure elements, . Element A is chosen to be the one with the highest sublimation and vaporization enthalpies. If the two components are identical, ΔH_{sub }= 0 and ΔH_{vap }= 0, there is no segregation and we retrieve Equation 3. x_{solidus }and x_{liquidus }are obtained from solving Equation 3. Assuming an ideal solution, only the first surface layer will be different from the core composition.
Considering the surface segregation in the PtPd system, we can see in Figure 4 that the lens shape of the surface liquidus/solidus curves is deformed compared to the core. At a given temperature, the liquidus and solidus curves of the surface are enriched in Pd compared to the core; meaning that the surface is depleted of Pt (the higher bond energy element) which is in agreement with experimental observations[2729] and other theoretical calculations[2931]. This is due to the fact that Pd has a lower solid surface energy, a lower cohesive energy compared to Pt and also because diffusion is enhanced at the nanoscale [32].
Figure 4. Phase diagram of the PtPd system considering the surface segregation effect. Surface segregation effect at a size equal to 4 nm for a spherical nanoparticle.
Sizedependent catalytic activation energy of PtPd
The catalytic activation energy is the energy quantity that must be overcome in order for a chemical reaction to occur in presence of a catalyst. The low the catalytic activation energy is, the most active the catalyst is. It is thus an important kinetic parameter linked to the chemical activity. Indeed, the catalytic activation energy is a linear function of the work function [3335]. For pure materials, the catalytic activity depends on the fraction of surface atoms on corners and edges while for binary compounds it depends also on the surface segregation. Recently, it has been showed by Lu and Meng in Ref. [36] that the sizedependent catalytic activation energy, E_{ca }could be obtained from the following relation:
Therefore, it means that the sizedependent catalytic activation energy decreases with size.
To compare with experimental results, the ratio of the catalytic activation energies between tetrahedral (D = 4.8 nm) and spherical (D = 4.9 nm) pure platinum nanoparticles has been determined around 0.66 in excellent agreement with the experimental value of 0.62 ± 0.06 announced by Narayanan and ElSayed [3739]. Moreover, the ratio of the catalytic activation energies between cubic (D = 7.1 nm) and spherical (D = 4.9 nm) pure platinum nanoparticles is around 1.01 in relative good agreement with the experimental value of 1.17 ± 0.12 [3739].
From the sizedependent PtPd phase diagram, the melting temperature of the alloy can be deduced. Equation 6 describes the melting temperature of the bulk PtPd while Equations 7 and 8 describe the nanoscaled melting temperature of a nonsegregated and segregated spherical nanoparticle (with a diameter equal to 4 nm), respectively.
where x represents the alloy composition. For a spherical PtPd nanoparticle with a diameter equal to 4 nm, by combining Equations 58, E_{ca }seems to evolve quadratically with the composition when the segregation is not considered; which is not the case when the segregation is considered (Figure 5). For the segregated PtPd nanoparticle, a maximum in the catalytic activation energy is reached around 16% of Pt composition.
Figure 5. Composition dependency of the catalytic activation energy for a spherical nanoparticle of PtPd. Nanoparticle of PtPd with a size equal to 4 nm.
Conclusions
In conclusion, it has been shown that thermodynamics can still provide useful insights in nanoscience and more specifically in catalysis. The future development of catalysts and fuel cells is dependent upon our ability to control the size, shape, and surface chemistry of individual nanoparticles. Future theoretical work will have to consider the environment in which the particles are synthesized as well as the preparation method because these parameters can have a great influence on the shape stability and on the catalytic properties.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
GG carried out the calculations on the size and shape effects on the melting temperature, phase diagrams and catalytic activation energy; drafted the manuscript. GA carried out the calculations on the phase diagrams (shape effect) in collaboration with GG. DH carried out the calculations on the phase diagrams (segregation effect) in collaboration with GG. All authors read and approved the final manuscript.
Acknowledgements
G. Guisbiers would like to thank the Belgian Federal Science Policy Office (BELSPO) through the "Mandats de retour" action for their financial support.
References

Maye MM, Kariuki NN, Luo J, Han L, Njoki P, Wang L, Lin Y, Naslund HR, Zhong CJ: Electrocatalytic reduction of oxygen: gold and goldplatinum nanoparticle catalysts prepared by two phase protocol.
Gold Bulletin 2004, 37:217. Publisher Full Text

Deng L, Hu W, Deng H, Xiao S: Surface segregation and structural features of bimetallic AuPt nanoparticles.
Journal of Physical Chemistry C 2010, 114:11026. Publisher Full Text

Sealy C: The problem with platinum.
Materials Today 2008, 11:6568. Publisher Full Text

Tao F, Grass ME, Zhang Y, Butcher DR, Renzas JR, Liu Z, Chung JY, Mun BS, Salmeron M, Samorjai GA: Reactiondriven restructuring of RhPd and PtPd coreshell nanoparticles.
Science 2008, 322:932934. PubMed Abstract  Publisher Full Text

Guisbiers G, Buchaillot L: Universal size/shapedependent law for characteristic temperatures.
Physics Letters A 2009, 374:305308. Publisher Full Text

Guisbiers G: Sizedependent materials properties towards a universal equation.
Nanoscale Research Letters 2010, 5:11321136. PubMed Abstract  Publisher Full Text  PubMed Central Full Text

Barnard AS: Using theory and modelling to investigate shape at the nanoscale.
Journal of Materials Chemistry 2006, 16:813815. Publisher Full Text

Yacaman MJ, Ascencio JA, Liu HB, GardeaTorresdey J: Structure shape and stability of nanometric sized particles.
Journal of Vacuum Science & Technology B 2001, 19:10911103. PubMed Abstract  Publisher Full Text

Guisbiers G, Buchaillot L: Modeling the melting enthalpy of nanomaterials.
Journal of Physical Chemistry C 2009, 113:3566. Publisher Full Text

Liang LH, Zhao M, Jiang Q: Melting enthalpy depression of nanocrystals based on surface effect.
Journal of Materials Science Letters 2002, 21:18431845. Publisher Full Text

Lu HM, Jiang Q: Sizedependent surface energies of nanocrystals.
Journal of Physical Chemistry B 2004, 108:56175619. Publisher Full Text

Barnard AS, Zapol P: A model for the phase stability of arbitrary nanoparticles as a function of size and shape.
Journal of Chemical Physics 2004, 121:4276. PubMed Abstract  Publisher Full Text

Vitos L, Ruban AV, Skriver HL, Kollar J: The surface energy of metals.
Surface Science 1998, 411:186202. Publisher Full Text

Abudukelimu G: Thermodynamical approach of phase diagrams of nanosystems: size and shape effects. In Ph D thesis. University of MonsHainaut; 2009.

Wang ZL, Petroski JM, Green TC, ElSayed MA: Shape Transformation and Surface Melting of Cubic and Tetrahedral Platinum Nanocrystals.
Journal of Physical Chemistry B 1998, 102:61456151. Publisher Full Text

Miao L, Bhethanabotla VR, Joseph B: Melting of Pd clusters and nanowires: a comparison study using molecular dynamics simulation.

MontejanoCarrizales JM, RodriguezLopez JL, Pal U, MikiYoshida M, JoséYacaman M: The completion of the platonic atomic polyhedra: The dodecahedron.
Small 2006, 2:351355. PubMed Abstract  Publisher Full Text

Xiong Y, Xia Y: Shapecontrolled synthesis of metal nanostructures: the case of palladium.
Advanced Materials 2007, 19:33853391. Publisher Full Text

Cheong S, Watt JD, Tilley RD: Shape control of platinum and palladium nanoparticles for catalysis.
Nanoscale 2010, 2:20452053. PubMed Abstract  Publisher Full Text

Vollath D, Fischer FD: Structural fluctuations in nanoparticles.
Journal of Nanoparticle Research 2009, 11:433439. Publisher Full Text

Kittel C: Introduction to solid state physics. 7th edition. New York: Wiley; 1995.

Guisbiers G, Wautelet M, Buchaillot L: Phase diagrams and optical properties of phosphide, arsenide, and antimonide binary and ternary IIIV nanoalloys.

Guisbiers G, Liu D, Jiang Q, Buchaillot L: Theoretical predictions of wurtzite IIInitrides nanomaterials properties.
Physical Chemistry Chemical Physics 2010, 12:7203. PubMed Abstract

Guisbiers G, Abudukelimu G, Wautelet M, Buchaillot L: Size, shape, composition, and segregation tuning of InGaAs thermooptical properties.
Journal of Physical Chemistry C 2008, 112:1788917892. Publisher Full Text

Sankaranarayanan SKRS, Bhethanabotla VR, Joseph B: Molecular dynamics simulation study of the melting of PdPt nanoclusters.

Williams FL, Nason D: Binary alloy surface compositions from bulk alloy thermodynamic data.
Surface Science 1974, 45:377408. Publisher Full Text

Renouprez AJ, Rousset JL, Cadrot AM, Soldo Y, Stievano L: Structure and catalytic activity of palladiumplatinum aggregates obtained by laser vaporisation of bulk alloys.
Journal of Alloys and Compounds 2001, 328:50. Publisher Full Text

Rousset JL, Cadrot AM, Cadete Santos Aires FJ, Renouprez A, Mélinon P, Perez A, Pellarin M, Vialle JL, Broyer M: Study of bimetallic PdPt clusters in both free and supported phases.
Journal of Chemical Physics 1995, 102:85748585. Publisher Full Text

Ferrando R, Jellinek J, Johnston RL: Nanoalloys: From Theory to Applications of Alloy Clusters and Nanoparticles.
Chemical Reviews 2008, 108:845910. PubMed Abstract  Publisher Full Text

Massen C, MortimerJones TV, Johnston RL: Geometries and segregation properties of platinumpalladium nanoalloy clusters.
Journal of the Chemical Society, Dalton Transactions 2002, 43754388.

Luyten J, Creemers C: Surface segregation in ternary PtPdRh alloys studied with Monte Carlo simulations and the modified embedded atom method.
Surface Science 2008, 602:24912495. Publisher Full Text

Guisbiers G, Buchaillot L: Size and shape effects on creep and diffusion at the nanoscale.
Nanotechnology 2008, 19:435701. Publisher Full Text

Vayenas CG, Bebelis S, Pliangos C, Brosda S, Tsiplakides D: Electrochemical activation of catalysis. New York: Kluwer Academic; 2001.

Kemball C: Catalysis on evaporated metal films. I. The efficiency of different metals for the reaction between ammonia and deuterium.
Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences 1952, 214:413426. Publisher Full Text

Vayenas CG, Ladas S, Bebelis S, Yentekakis IV, Neophytides S, Yi J, Karavasilis C, Pliangos C: Electrochemical promotion in catalysis: nonfaradaic electrochemical modification of catalytic activity.
Electrochimica Acta 1994, 39:18491855. Publisher Full Text

Lu HM, Meng XK: Theoretical model to calculate catalytic activation energies of platinum nanoparticles of different sizes and shapes.
Journal of Physical Chemistry C 2010, 114:15341538. Publisher Full Text

Narayanan R, ElSayed MA: Shapedependent catalytic activity of platinum nanoparticles in colloidal solution.
Nano Letters 2004, 4:1343. Publisher Full Text

Narayanan R, ElSayed MA: Catalysis with transition metal nanoparticles in colloidal solution: nanoparticle shape dependence and stability.
Journal of Physical Chemistry B 2005, 109:1266312676. Publisher Full Text

Narayanan R, ElSayed MA: Some aspects of colloidal nanoparticle stability, catalytic activity, and recycling potential.
Topics in Catalysis 2008, 47:1521. Publisher Full Text

Martienssen W, Warlimont H: Springer Handbook of Condensed Matter and Materials Data. Berlin: Springer; 2005.

Arblaster JW: Vapour pressure equations for the platinum group elements.
Platinum Metals Review 2007, 51:130135. Publisher Full Text