Synthesis of Pt-based hollow nanoparticles using carbon-supported Co@Pt and Ni@Pt core–shell structures as templates: Electrocatalytic activity for the oxygen reduction reaction
Graphical abstract
Highlights
► Electrocatalysts exhibited hollow-induced lattice contraction. ► Ligand effect played important role on the activity for the ORR. ► Hollow@NiPt/C presented high specific activity. ► High activity of Hollow@NiPt/C associated to the faster reduction of oxygenated species.
Introduction
Platinum is the best single metal catalyst for the oxygen reduction reaction (ORR) for low temperature polymer electrolyte membrane fuel cells (PEM), in both acid and alkaline electrolytes [1], [2], [3], [4], [5]. Aiming at increasing the reaction rate, several bimetallic electrocatalysts composed of Pt and 3d-transition metals (Co, Ni, Fe) have been investigated [6], [7], [8], [9]. These electrocatalysts exhibit improved catalytic performance for the ORR when compared to Pt alone, both in terms of mass activity (per g of Pt) or specific activity (per real cm2 of Pt) [10], [11], [12]. Previous works have shown that the ORR activity of these bimetallic electrocatalysts can be further improved by enriching the surface of the catalyst with Pt (surface segregation), induced by annealing a Pt3M surface at elevated temperatures [13], [14], [15] and/or induced by dissolution of the 3d-transition metal in an acidic environment [16], [7]. The enhancement in ORR specific activity was achieved by using the mismatch-induced lattice contraction and/or the ligand effect [17], [18], [19], [20], [21]. In this type of electrocatalyst, Pt atoms suffer a significant d-band center down-shift and, therefore, chemisorb oxygenated and spectator species more weakly than pure Pt.
Another strategy to enhance the Pt electrocatalytic activity that has been explored is the so-called de-alloying method [22]. In this procedure, the formation of a Pt “skeleton” surface in a bimetallic particle is achieved via dissolution of the alloying elements, such as Cu, Ni and Co in an acid electrolyte. The increased activity for the ORR has been ascribed to variations in the number and kind of neighbors surrounding the Pt atoms, as well compression or expansion in the Pt–Pt bond distance, which induces a structural change that directly affects the Pt electronic density of occupied states in the Pt 5d-band, weakening the adsorption strength of reaction intermediates and spectators.
Furthermore, in order to enhance the Pt catalytic activity and dramatically decrease the total mass of Pt in the electrocatalyst, some works have focused on the study of core–shell nanostructures, in which an ultra-thin platinum shell is either deposited on a noble [23], [24], [25] or non-noble/noble metal nanoparticle core [19], [26], [27]. Therefore, the total mass of Pt is reduced, while the metal core plays an important role in modifying the Pt electronic structure, decreasing the Pt d-band center.
The use of core–shell and bi-metallic alloys with non-noble metal in their composition or in their cores (such as Co, Fe, Ni or Cu atoms) as electrocatalysts would be an important way for enhancing both the Pt specific catalytic activity and the Pt mass activity for oxygen reduction in fuel cells operating with acid electrolytes [13], [28]. However, these materials show a loss of the non-noble metal due to dissolution in acidic media, which would decrease the ligand effect (electronic interaction between Pt and the non-noble metal) and, consequently, the electrocatalytic performance of the electrocatalyst. However, other nanoparticle compositions can be stable upon potential cycling, as in the case of IrNi [29] and AuNi0.5Fe [30] core–shell structures, with Pt deposited in the outermost layer.
In the case of core–shell or bi-metallic structures with unstoppable non-noble metal dissolution, these materials can be used as precursors for the synthesis of noble metal hollow nanoparticles. Recently, Wang et al. [31] and Dubau et al. [32] studied the ORR catalyzed by Ni@Pt/C core–shell nanostructures and Pt3Co/C alloys, respectively. They revealed the formation of Pt hollow nanoparticles after electrochemical potential cycling in acid media. The authors ascribed the cause of Pt hollow formation to the diffusion of the non-noble metal atoms from the bulk to the surface of the material, induced by the Kirkendall effect on the nanoscale.
The Kirkendall effect is a diffusion mechanism in which there is net mass flow of a faster diffusing species balanced by the opposing flow of vacancies that condense into voids in solids and, therefore, it is responsible for solid-to-hollow conversions [33], [34], [35], [36]. The phenomena of solid state diffusion are more frequently described by using a unique diffusion coefficient which describes the migration of one component through the structure of another (which composes the matrix). This is a typical approach to describe the diffusion of a low content solute. When the amount of both components is considerably high, solid state diffusion is usually described using the diffusion coefficients of one species into the other (i.e., diffusivity of component A into component B, DAB, and diffusivity of component B into component A, DBA). These diffusion coefficients are called “intrinsic diffusion coefficients”. In these metallic systems, diffusion might be described through a combination of both intrinsic diffusion coefficients, describing the so-called interdiffusion. In the 1940s, Kirkendall studied interdiffusion using copper–brass diffusion couples and observed that these intrinsic diffusion coefficients usually have different values. Thus, in order to obey the laws of continuity, vacancies are created and migrate in the opposite direction to the net atomic flux caused by the asymmetry of the intrinsic diffusion coefficients during interdiffusion. These vacancies might condense into the solid, creating voids [33], [37].
In nanoparticles, if the core species (a non-noble metal) has the faster intrinsic diffusion coefficient, an outward atomic flux is created in response to a reactive environment to which the material is exposed, and this is balanced by an inward flux of vacancies. This effect is also responsible for the dissolution of the non-noble metal in nanoparticles in acid media (or oxide formation on the material surface in the case of alkaline media). As a result, the formation of a Pt structure with a void in the center occurs, forming the so-called hollow nanoparticle. Some amount of the non-noble metal may remain in the hollow wall layers due to the formation of solid solution, in which diffusion, followed by dissolution, is kinetically very low. This would decrease the rate of non-noble metal diffusion/dissolution on the particle surface.
Among the numerous investigations concerning the synthesis of new electrocatalysts for enhancing both the electrocatalytic activity and durability regarding the ORR in PEM fuel cells, hollow nanostructures have shown novel properties for controlling the electrocatalytic activity [36], [46]. However, little is known about how the electrocatalytic activity of these nanoparticles is influenced by varying the size, shape and the remaining activity-enhancing non-noble metal components, which is controlled by the nature of the template through which the hollow structure is produced.
In this work, the synthesis and ORR activity of Pt hollow nanoparticles were studied. The effect of two different non-noble metals as cores (Ni and Co) was investigated on the shape and size of the resulting Pt hollow nanoparticle, in addition to their effect on the electrocatalytic activity for the ORR.
Section snippets
Synthesis of carbon-supported Ni@Pt and Co@Pt core–shell and Pt hollow nanoparticles
The investigated materials consisted of carbon-supported Pt hollow nanoparticles, synthesized using Ni@Pt/C and Co@Pt/C core–shell nanostructures as templates. The hollow electrocatalysts are represented by Hollow@NiPt/C and Hollow@CoPt/C, indicating that these two materials were synthesized using Ni or Co as metal cores. Commercial carbon-supported Pt nanoparticles (Pt/C E-TEK) were also used for comparison.
The core–shell nanoparticle templates (Ni or Co core and Pt shell) were synthesized
Results and discussion
The effect of the nature of the non-noble metal on the particle distribution on the carbon support, particle size and average atomic composition of the Ni@Pt/C and Co@Pt/C templates were investigated by STEM, X-EDS and XRD. Fig. 1 show the representative dark-field (DF) and bright-field (BF)-STEM images for the as-prepared Ni@Pt/C (a) and Co@Pt/C (b) electrocatalysts, respectively. The images reveal different particles sizes when Ni@Pt and Co@Pt were used as the material templates. The particle
Conclusions
The results presented in this work show considerable differences in nanoparticle structure/composition and activity for the ORR depending on the nature of the non-noble metal in the nanoparticle core. The Pt hollow nanostructures showed higher specific catalytic activity than that of the state-of-the-art Pt/C electrocatalyst. This was attributed to three main effects: (i) hollow-induced lattice contraction in the multilayer Pt shells, (ii) mismatch-induced lattice contraction of the thick Pt
Acknowledgements
The authors thank FAPESP (Fundação de Amparo à Pesquisa do Estado e São Paulo) (Grants: 2009/07629-6, 2011/50727-9 and 2009/11073-3) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) for financial support and LNLS (Brazilian Synchrotron Light Laboratory) for assisting with the XAS experiments.
References (49)
Electrochimica Acta
(1984)- et al.
Applied Catalysis B: Environmental
(2005) - et al.
Journal of Power Sources
(2011) - et al.
Electrochimica Acta
(2002) - et al.
Vacuum
(1990) - et al.
Journal of Power Sources
(2009) - et al.
Electrochimica Acta
(2012) - et al.
Applied Catalysis B: Environmental
(2008) - et al.
Electrochemistry Communications
(2010) - et al.
Electrochimica Acta
(2006)