J. Phys. Chem. B 2005, 109, 23543-23549
23543
Nanofaceted Platinum Surfaces: A New Model System for Nanoparticle Catalysts Vladimir Komanicky, Andreas Menzel, Kee-Chul Chang, and Hoydoo You* Materials Science DiVision, Argonne National Laboratory, 9700 South Cass AVenue, Argonne, Illinois 60439 ReceiVed: July 27, 2005; In Final Form: October 10, 2005
We present a novel model system for nanoparticle electrocatalysts. A surface consisting of alternating (100) and (111) facets, several nanometers across and nearly 1 µm long, were self-assembled by annealing Pt single crystal surfaces initially cut at the midpoint between [111] and [100] directions, i.e., Pt(1+x3 1 1). The formation of these self-assembled arrays of nanofacets was monitored by in-situ surface X-ray scattering. These surfaces were further characterized with scanning probe microscopy and cyclic voltammetry. We found that the Pt(1+x3 1 1) surface is flat with less than 1 nm rms roughness when it was annealed in argon/ hydrogen atmosphere. Then the surface forms nanofacets when it is annealed in pure air. This nanofaceting transition was completely reVersible and reproducible. We investigated effects of CO adsorption on the voltammetric characteristics of both hydrogen-annealed and air-annealed surfaces. We found that CO-adsorption/ desorption cycles in CO containing electrolyte solution result in considerable modification of blank cyclic voltammograms for the both surfaces. We attributed these differences to the electrochemical annealing of surface defects due to the increased mobility during the cycles.
1. Introduction Platinum metal and its alloys with other transition metals are important electrocatalysts for low-temperature fuel cells. The catalysts are typically dispersed in a form of nanoparticles on conductive supports for a minimum loading of platinum. Although it is known that Pt nanoparticle morphologies vary greatly depending on details of synthesis,1-3 a common feature is the prevalence of (111) and (100) surfaces, as in the case of cubooctahedra, one of the most frequently found geometries of carbon-supported platinum nanoparticles.4,5 While molecularlevel studies of electrocatalytic reactions on the nanoparticle surfaces are difficult, crystallographically well-defined surfaces of platinum, in particular, (111) and (100) surfaces, have been extensively studied as models for the nanoparticle surfaces, and a great deal of progress has been made in the past decades. However, the question of how well the individual single-crystal surfaces truly represent realistic nanoparticle surfaces still remains. In this work we describe a novel approach to the preparation of arrays of “one-dimensional platinum nanoparticles” consisting of narrow and long (111) and (100) facets. In this way, the surfaces resemble, at least in one dimension, those of real nanoparticles but are sufficiently well defined for systematic studies. Platinum single crystals were cut at the midpoint between (111) and (100) surfaces and polished to expose Pt(1+x3 1 1), one of the least thermodynamically favorable surfaces. After annealing in hydrogen gas or air, surface morphologies of the platinum crystals were studied by atomic force microscopy (AFM), scanning tunneling microscopy (STM), and surface X-ray scattering (SXS). Cyclic voltammogram (CV) measurements in sulfuric acid were also used to characterize the different surface morphologies. We will also show that adsorption and oxidative desorption cycles of CO can lead to significant modifications of the * Corresponding author. E-mail:
[email protected].
nanofaceted surfaces. We will discuss the results of the CO cycling, indicating that how the modifications make the adsorbed CO layers become more resistive to oxidation and thus bind more tightly to the surfaces. 2. Experimental Section Platinum crystals were prepared by melting the tip of a 99.999% platinum wire, forming a bead with a diameter of ∼4 mm. The beads were mounted in a goniometer, oriented to the desired direction by X-ray diffraction, and cast in epoxy resin. Platinum crystals were polished on sandpaper ranging from 400 to 1200 grit, then with 9 µm diamond suspension on Texmet followed by polishing with 3 µm and 1 µm diamond paste on Nanocloth. Final finish was obtained by polishing with 0.3 µm and 0.05 µm alumina suspension on Nanocloth. The singlecrystal surfaces prepared in this way have a miscut typically less than 0.2°. A Ag/AgCl (3 M KCl in gel) reference electrode placed in Luggin capillary has been used as a reference. All potentials in the text are given with respect to this reference unless mentioned otherwise. Bulk annealing of the platinum crystals was performed by an induction heater setup for 10 h at ∼1900 K in argon/hydrogen flow (3% hydrogen, high purity). This was sufficient to restore a bulk mosaic of Bragg peaks to
