Research Article pubs.acs.org/acscatalysis
Role of Chemical Composition in the Enhanced Catalytic Activity of Pt-Based Alloyed Ultrathin Nanowires for the Hydrogen Oxidation Reaction under Alkaline Conditions Megan E. Scofield,† Yuchen Zhou,† Shiyu Yue,† Lei Wang,† Dong Su,‡ Xiao Tong,‡ Miomir B. Vukmirovic,§ Radoslav R. Adzic,§ and Stanislaus S. Wong*,†,∥ †
Department of Chemistry, State University of New York at Stony Brook, Stony Brook, New York 11794-3400, United States Center for Functional Nanomaterials, Brookhaven National Laboratory, Building 735, Upton, New York 11973, United States § Chemistry Department, Brookhaven National Laboratory, Building 555, Upton, New York 11973, United States ∥ Condensed Matter Physics and Materials Sciences Division, Brookhaven National Laboratory, Building 480, Upton, New York 11973, United States ‡
S Supporting Information *
ABSTRACT: With the increased interest in the development of hydrogen fuel cells as a plausible alternative to internal combustion engines, recent work has focused on creating alkaline fuel cells (AFC), which employ an alkaline environment. Working in alkaline as opposed to acidic media yields a number of tangible benefits, including (i) the ability to use cheaper and plentiful precious-metal-free catalysts, due to their increased stability, (ii) a reduction in the amount of degradation and corrosion of Ptbased catalysts, and (iii) a longer operational lifetime for the overall fuel cell configuration. However, in the absence of Pt, no catalyst has achieved activities similar to those of Pt. Herein, we have synthesized a number of crystalline ultrathin PtM alloy nanowires (NWs) (M = Fe, Co, Ru, Cu, Au) in order to replace a portion of the costly Pt metal without compromising on activity while simultaneously adding in metals known to exhibit favorable synergistic ligand and strain effects with respect to the host lattice. In fact, our experiments confirm theoretical insights about a clear and correlative dependence between measured activity and chemical composition. We have conclusively demonstrated that our as-synthesized alloy NW catalysts yield improved hydrogen oxidation reaction (HOR) activities as compared with a commercial Pt standard as well as with our as-synthesized Pt NWs. The Pt7Ru3 NW system, in particular, quantitatively achieved an exchange current density of 0.493 mA/cm2, which is higher than the corresponding data for Pt NWs alone. Additionally, the HOR activities follow the same expected trend as their calculated hydrogen binding energy (HBE) values, thereby confirming the critical importance and correlation of HBE with the observed activities. KEYWORDS: hydrogen oxidation reaction, Pt-based alloy, nanowires, ligand effect, hydrogen binding energy
1. INTRODUCTION
fuel cell where the oxidation of hydrogen occurs, because unfortunately, the kinetics for this reaction process are inherently slower in alkaline media versus those in acid electrolytes. In particular, despite the fact that platinum is known to be the most active metal for initiating the hydrogen oxidation reaction (HOR) in alkaline media, it unfortunately still exhibits 2 orders of magnitude slower kinetics than those measured in corresponding acid electrolytes.2,3 Moreover, existing catalysts are particularly susceptible to CO poisoning. Therefore, to deal with all of these issues, there is a tangible need to create unconventional geometries possessing superior HOR kinetics in alkaline media whose performance inherently surpasses that of elemental, monometallic Pt.
With the increased interest in the development of hydrogen fuel cells as a plausible alternative to internal combustion engines, recent work has focused on creating viable alkaline fuel cells (AFC), which employ an alkaline medium as opposed to acid as the primary electrolyte. In effect, AFCs possess a number of important benefits associated with the presence of a more favorable and desirable alkaline electrolyte medium. Specifically, these include (i) the ability to use non-preciousmetal catalysts due to their increased stability, (ii) a diminished degree of degradation and corrosion of Pt-based catalysts, and (iii) a general reduction in the amount of deterioration inherent to the overall fuel cell configuration.1−3 Additionally, the reaction that occurs at the cathode, namely the oxygen reduction reaction (ORR), tends to possess faster kinetics in alkaline media.1,4 However, there remains a significant need for improvement at the anode side of the © XXXX American Chemical Society
Received: February 2, 2016 Revised: May 3, 2016
3895
DOI: 10.1021/acscatal.6b00350 ACS Catal. 2016, 6, 3895−3908
Research Article
ACS Catalysis
achieving hydrogen binding with an “optimal” HBE: i.e., a value closer to 0 by comparison with Pt bulk itself. Additional work performed by Wang and co-workers17 implied that by alloying Pt with a metal possessing a comparatively stronger hydrogen binding energy, one could conceivably shift the overall HBE to more favorable values: i.e., to weaker HBE values by comparison with that of pure Pt itself. For optimal HOR performance, we ultimately seek a weakening in the hydrogen binding energy by comparison with that of Pt, an assertion supported by others.17 In effect, there is a strengthening of the HBE as the d band center moves closer to the Fermi level of a metal such as Pt.10,18 Our goal is the opposite, and therefore, our objective has been to create alloyed nanomaterials that demonstrate a weakening in HBE by moving the d band center away from the Pt Fermi level. Hence, on the basis of previous HBE calculations associated with various transition metals,16,19 we can potentially tailor novel electrocatalysts with improved HOR kinetics by deliberatively and systematically altering the alloy composition and, therefore, control the corresponding variations in HBE. Therefore, our goal has been to correlate composition with activity. Additionally, our objective of replacing expensive noble metals with cheaper, more abundant metals is essential for designing electrocatalysts for mass production. Apart from rationally varying the chemical compositions of our Pt-based alloy structures, we have also tailored the morphology of our catalysts. In effect, crystalline onedimensional (1D) catalysts have previously been shown to possess high aspect ratios, fewer potentially deleterious defect sites, and short segments of crystalline planes, all of which contribute to the enhanced activity of 1D systems as compared with their zero-dimensional (0D) counterparts.20−22 Furthermore, as per our recent Perspective article on this issue,23a anisotropic nanostructures such as Pt NWs maintain a favorable downshift in the Pt d band, which contributes to a weaker d−π* interaction with the adsorbed CO, thereby improving Pt’s ability to oxidize adsorbed CO at potentials closer to the thermodynamic potential for the methanol oxidation reaction (MOR). Moreover, the surfaces of 1D morphologies can be tuned so as to preferentially display different crystal facets.8,21 In addition, the rates of dissolution and ripening processes have been demonstrated to be significantly slower in the case of 1D nanostructures, by comparison with those of commercial Pt NP/C. All of these findings suggest that 1D architectures represent promising motifs for HOR catalysts. Our last novel variation for HOR has been to reduce the average diameters of our 1D nanowires tested to the ultrathin size regime (0.55 V vs RHE) inhibits HOR activity54 by blocking the Pt active sites upon subsequent deposition onto the surface.53,55 3906
DOI: 10.1021/acscatal.6b00350 ACS Catal. 2016, 6, 3895−3908
ACS Catalysis
■
catalyst exhibits the highest HOR activity and, specifically, an exchange current density of 0.493 mA/cm2, a value even better than that of pure Pt NWs alone. Additionally, all of our binary alloy NW systems, with the exception of the PtCu and PtAu NW systems, outperformed our as-synthesized, monometallic Pt NWs, likely due to the addition of M to the Pt lattice, thereby imparting electrochemically favorable ligand and lattice strain effects on the bare Pt structure. In addition, we have been able to correlate the trend found in the calculated HBE values of our various alloyed structures with their actual HOR activities, as previously discussed in section 3.3. We find that all catalysts generally follow the expected theoretical trend (within experimental error).15 To summarize, the significance of this work is several-fold. First, we have systematically synthesized a variety of chemically well-defined Pt-based binary alloy systems and correlated their HOR activities in alkaline media as a function of composition. Our work highlights our capability of reproducibly and reliably achieving electrocatalytic enhancements over that of commercial Pt NPs alone with specific, well-chosen alloy compositions. Second, in our alloy NW samples, we have experimentally elucidated and confirmed the presence of desirable electronic interactions associated with the coupling of the second metal with Pt. These critically relevant ligand and lattice strain effects have manifested themselves in not only perceptible shifts in the Pt 4f region associated with the XPS spectra for each binary catalyst tested but also corresponding shifts in both the hydrogen adsorption and oxide regions with respect to Pt. In the case of PtRu, PtFe, PtCo, and PtCu NW alloyed systems, these phenomena may collectively explain the expected decrease in HBE values (versus Pt) because of a transitioning of the d band center away from the Pt Fermi level. On the other hand, for PtAu NW alloys, one would expect a corresponding increase in HBE (versus Pt), which can be ascribed to the d band center moving closer to the Pt Fermi level. In doing so, these results have supported the notion that electronic effects substantively control HBE values and, therefore, HOR activity. Third, we have correlated our observed current densities and, hence, the resulting measured experimental activities with trends predicted on the basis of theoretical calculations of nearsurface alloys possessing identical chemical compositions and in doing so, highlighted possible limitations and caveats associated with existing models. Finally, all of our data have confirmed our assertion that morphology, size, and chemical composition need to be rationally and collectively tuned in order to achieve optimal electrochemical performance. Specifically, we have demonstrated that 1D anisotropic motifs, characterized by
