Letter pubs.acs.org/NanoLett
Understanding and Controlling Nanoporosity Formation for Improving the Stability of Bimetallic Fuel Cell Catalysts Lin Gan,† Marc Heggen,‡ Rachel O’Malley,§ Brian Theobald,§ and Peter Strasser*,† †
The Electrochemical Catalysis, Energy and Materials Science Laboratory, Department of Chemistry, Technical University Berlin, 10623 Berlin, Germany ‡ Ernst Ruska Center for Microscopy and Spectroscopy with Electrons, Forschungszentrum Juelich GmbH, 52425 Juelich, Germany § Johnson Matthey Fuel Cells Ltd., Lydiard Fields, Great Western Way, Swindon SN5 8AT, United Kingdom S Supporting Information *
ABSTRACT: Nanoporosity is a frequently reported phenomenon in bimetallic particle ensembles used as electrocatalysts for the oxygen reduction reaction (ORR) in fuel cells. It is generally considered a favorable characteristic, because it increases the catalytically active surface area. However, the effect of nanoporosity on the intrinsic activity and stability of a nanoparticle electrocatalyst has remained unclear. Here, we present a facile atmosphere-controlled acid leaching technique to control the formation of nanoporosity in Pt−Ni bimetallic nanoparticles. By statistical analysis of particle size, composition, nanoporosity, and atomic-scale core−shell fine structures before and after electrochemical stability test, we uncover that nanoporosity formation in particles larger than ca. 10 nm is intrinsically tied to a drastic dissolution of Ni and, as a result of this, a rapid drop in intrinsic catalytic activity during ORR testing, translating into severe catalyst performance degradation. In contrast, O2-free acid leaching enabled the suppression of nanoporosity resulting in more solid core−shell particle architectures with thin Pt-enriched shells; surprisingly, such particles maintained high intrinsic activity and improved catalytic durability under otherwise identical ORR tests. On the basis of these findings, we suggest that catalytic stability could further improve by controlling the particle size below ca. 10 nm to avoid nanoporosity. Our findings provide an explanation for the degradation of bimetallic particle ensembles and show an easy to implement pathway toward more durable fuel cell cathode catalysts. KEYWORDS: Pt bimetallic alloys, dealloying, porous nanoparticles, durability, size effect, oxygen reduction reaction
T
and control metal leaching are therefore believed to be critical to achieve high catalytic stability. Recent advanced electron microscopic and spectroscopic studies revealed that two distinctly different structures exist in dealloyed Pt bimetallic nanoparticle catalysts: on the one hand, solid core−shell nanoparticles with a Pt-rich shell of given thickness surrounding a nonporous Pt-poor alloy core and on the other hand nanoscale particles with nanoporosity characterized by highly irregular pores, pits, and voids in their interior (also called “spongy” or “swiss cheese” type particles).17−23 Recent studies from our own and other laboratories show that nanoporosity forms spontaneously during the dealloying (leaching) of Pt alloy nanoparticles if and only if the particle size exceeds a critical value due to the competed surface atomic processes during dealloying (i.e., leaching of transition metals vs surface diffusion of residual Pt
he exploration of a highly active and stable catalyst for the cathodic oxygen reduction reaction (ORR) is one of the main challenges for the development of proton-exchangemembrane fuel cells (PEMFCs).1−4 By surface structural designs, Pt bimetallic catalysts containing transition metals (such as Fe, Co, Ni or Cu etc.) have been recently developed and show significantly enhanced mass activities compared to the state-of-the-art pure Pt catalyst.3,5−10 In particular, through dealloying transition-metal-rich Pt alloys, a Pt-rich shell surrounding an alloy core can be formed and the lattice-strain in the Pt-rich shell gave rise to very high catalytic ORR activities;9,11−13 recently ORR Pt mass activity gains up to 6x versus commercial Pt catalysts was reported for dealloyed Pt− Ni catalysts in industrial membrane-electrode-assembly (MEA) tests.14 Despite these recent achievements, the realization of Pt alloy electrocatalysts that maintain their initial Pt mass based activity has remained a major barrier. A key reason for their activity degradation is the continuous leaching of the lessnoble-metals during long-term stability test in acidic environment.15,16 A clear understanding and a path forward to suppress © XXXX American Chemical Society
Received: December 5, 2012 Revised: January 26, 2013
A
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 1. (A) TEM image and (B) HAADF-STEM image of Air−D−PtNi3 catalyst that was chemically dealloyed in air, showing the formation of nanoporous particles with larger sizes. (C) TEM image and (D) HAADF-STEM image of N2−D−PtNi3 catalyst that was chemically dealloyed in N2. All the nanoparticles exhibited nonporous structures. (E) Particle size distributions of as-received PtNi3 catalyst, Air−D−PtNi3 and N2−D−PtNi3 catalyst. (F) Correlations between particle size, composition, and porosity in the three catalysts, where the horizontal dot lines represent the average compositions. Only the particles well focused in the TEM images were used for the size-porosity-composition analysis in (E) and (F).
atoms).17,19 For instance, Snyder et al. reported that in order to evolve fully formed porosity, PtNi3 nanoparticles must have a minimum particle size of ca. 15 nm;19 Oezaslan et al. reported a critical particle diameter of ca. 30 nm in PtCu3 and PtCo3 nanoparticle for the pore formation. 17 The deliberate introduction of nanoporosity in bimetallic alloys has always been reported and perceived to be an effective way to enhance the catalytic activity due to enhanced active surface area.18,19,24−30 However, molecular understanding of the impact of nanoporosity beyond increased surface areas, in particular its influence on the particle composition and surface lattice strain, and, related to these, the intrinsic ORR activity and stability of a nanoparticle electrocatalyst has remained unclear. Here, we address this very question using atomic scale electron microscopy and spectroscopy and provide an answer as to which of the two dealloyed nanoparticle architectures, solid core−shell or nanoporous, is desirable for the design of an active and stable Pt alloy ORR nanoparticle electrocatalyst. In particular, we present a facile method to control the formation of nanoporosity in the Pt alloy nanoparticle catalyst via atmosphere-controlled chemical dealloying (ACD). This ACD protocol allows us to control the relative rates of surface dissolution and surface diffusion, a critical parameter in the formation of nanoporosity. By studying the evolution of the intraparticle structure of the alloy nanoparticles across different stages, we uncover how nanoporosity can influence the activity as well as the stability of Pt alloy catalysts. This methodology is applicable to a broad range of bimetallic catalyst nanoparticles, leading to unique insights concerning the correlation between nanoporosity and catalytic properties.
We studied the formation of nanoporosity in a single-phase face-center-cubic PtNi3 bimetallic alloy catalyst, as evidenced from X-ray diffraction analysis (Supporting Information Figure S1A). Transmission electron microscopy (TEM) images suggest a size distribution between 3−25 nm of the alloy nanoparticles (Supporting Information Figure S1B). Such a broad particle size distribution was ideal to monitor the size dependence of nanoporosity formation. Compared to electrochemical dealloying adopted in previous reports,9,19 chemical dealloying through acid leaching becomes more practical and could be used at large-scale in catalyst industry.31 To control the formation of nanoporosity in the PtNi3 catalyst, we found that the atmosphere during the chemical dealloying plays a key role. The ACD of the PtNi3 catalyst was performed in 0.5 M H2SO4, first under an air atmosphere, and then for comparison under an N2 atmosphere for 24 h. When dealloyed in air for 24 h (the obtained catalyst is hereafter referred to Air−D−PtNi3), nanoporous particles are readily formed. Figure 1A,B present the typical TEM and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of the Air−D−PtNi3 catalyst, respectively. Obviously, there are many larger particles showing the nanoporous structure with pore sizes of ca. 1−2 nm, while smaller particles remained nonporous due to their small size.17,19 Particle size distribution analysis (Figure 1E) further confirms that, in agreement with previous studies,19 the Air−D−PtNi3 nanoparticles show a nanoporous structure without exception when the size is above ∼13 nm. To correlate the nanoporosity to the particle size and composition, we performed statistical size-composition-porosity analysis using energy dispersive spectroscopy (EDX) over 30 B
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
size,17,19 (ii) a minimum composition in the less noble transition metal (parting limit),33 and (iii) a minimum anodic (critical) electrode potential.33,34 On a molecular level, these three conditions ensure that the surface diffusion rate of the more noble metal is slower than the surface dissolution rate of the less noble metal; this allows continuous dealloying deep in to the bulk and eventually leads to nanoporosity. To explain our results in Figure 1, we note that the gaseous environment has great impact on the free corrosion potential (or open circle potential, OCP) of a corroding alloy. In particular, the presence of molecular O2 created a complex mixed electrode potential between the Ni corrosion potential
particles across different sizes (Figure 1F). Interestingly, we found that the nanoporosity formation is strongly tied to dramatic particle composition changes: all the nanoporous particles exhibit a significant drop of the Ni composition from the original ∼75 atom % to 10−30 atom %, while nonporous nanoparticles below 13 nm mainly show only a mild decrease of the Ni composition, likely due to Ni dissolution from the near surface region only. This is in good agreement with the formation mechanism of nanoporosity where the surface diffusion of residual Pt atoms is slower than the leaching rate of Ni atoms on larger particles; therefore, Pt surface atoms are unable to efficiently fill and cover the surface corrosion pits formed by surface Ni leaching. As a result of this, free Ni corrosion can grow in depth (referred to as “Rayleigh surface instability”32), forms porous networks, and enhances additional Ni leaching. A very different picture is obtained when the free corrosion is performed under an inert atmosphere. When the PtNi3 catalyst was leached in N2-purged solution (hereafter referred as N2− D−PtNi3, all nanoparticles without exception exhibited a solid nonporous structure (Figure 1C and D). Statistical sizecomposition analysis shows a uniform and slight dropping of the Ni composition across different particle sizes. Most importantly, all particles larger than 13 nm maintained a high residual Ni composition around 60−70 at%, showing a sharp contrast to those in the Air−D−PtNi3 catalyst. The overall average compositions measured from large collections of nanoparticles (shown as the dash lines in Figure 1F) also suggests a higher average Ni-composition of N2−D−PtNi3 catalyst compared to Air−D−PtNi3. The sharply varying intraparticle structures and compositions, especially of the larger particles, are reflected in their distinctly different XRD patterns (Figure 2), where the Air−
Ni → Ni 2 + + 2e−
( −0.257 V/SHE)
(1)
the hydrogen evolution potential H2 + + 2e− → H 2
(0 V/SHE)
(2)
the water activation potential on Pt surfaces Pt − H 2Oad → Pt − OHad + H+ + e− (ca. +0.7 V/SHE)
(3)
and the oxygen redox potential O2 + 4e− + 4H+ → 2H 2O
( +1.23 V/SHE)
(4)
We probed the free corrosion potential, or equivalently the open circle potential (OCP), of the as-received PtNi3 catalyst under repeated N2 and O2 atmosphere (Figure S2 in the Supporting Information). The OCP of the catalyst under N2 was +0.82 V. Introducing O2 into the solution lead to a sudden increase of the OCP to a mixed potential of +0.97 V. Changing back to a N2 atmosphere decrease the OCP back to low values. From this we conclude that the lower OCP in absence of O2 slowed down the leaching rate of Ni, so that Pt surface diffusion prevented the formation of nanoporosity. In other words, the nitrogen atmosphere affected the corrosion potential such that condition (iii) was no longer fulfilled. Besides the effect on the OCP, enhanced O2 partial pressures favor the formation of Pt surface oxide species, which can slow down the surface diffusion rate of Pt, favoring nanoporosity. We then strived to gain insight into how nanoporosity affects the structure and catalytic properties of the surface layer. To this end, we studied the intraparticle compositional fine structures of the catalysts after dealloying by using aberrationcorrected STEM and electron energy loss spectroscopy (EELS) line scans. As reported previously, the subsurface compositional fine structure, involving the Pt shell thickness as well as the underlying Ni-compositional depth profile, plays important roles in resulting surface lattice-strain and the associated ORR activities.13,35−37 Figure 3A−C represents the EELS line profiles measured from selected Air−D−PtNi3 nanoparticles with different sizes. As expected, the smaller solid nanoparticles shown in Figure 3A exhibited a typical solid non porous core− shell structure with an average Pt shell thickness of ca. 0.6 nm (∼3 atomic layers). Here, Pt surface diffusion is fast enough to suppress nanoporosity formation inside the particle bulk. With increased size, however, nanoporosity emerged and the particles (Figure 3B,C) showed nonuniform and significantly increased Pt shell thickness associated with reduced total Ni content. This larger Pt shell thickness formed in porous nanoparticles can be ascribed to a clustering of the surface Pt atoms during dealloying, which could not redistributed on the
Figure 2. XRD patterns of as-received PtNi3 catalyst, N2−D−PtNi3, and Air−D−PtNi3 catalyst. All three show face-centered cubic structure. A shoulder peak arises near the pure Pt(111) peak, indicating the appearance of more Pt-rich phase in Air−D−PtNi3.
D−PtNi3 shows an obvious shift of the (111) peak compared to as-received PtNi3 and N2−D−PtNi3 due to the enhanced loss of Ni. In addition, a shoulder near the position of the pure Pt(111) reflection confirms the formation of Pt-richer phases in Air−D−PtNi3 sample. To understand the mechanism of our atmosphere−based porosity control technique we recall that there are three necessary conditions for porosity formation during dealloying of alloy nanoparticles: (i) a minimum (“critical”) particle C
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 3. Aberration-corrected STEM-EELS line profiles of (A−C) Air−D−PtNi3 catalyst and (D−F) N2−D−PtNi3 catalyst with different sizes, where the intensities were normalized to the elemental scattering factors and hence the intensity ratios represent thickness-projected compositions. The insets show the corresponding HAADF images of the nanoparticles with the white arrows indicating the line scan directions.
Figure 4. (A,B) ORR voltammograms of Air−D−PtNi3 and N2−D−PtNi3 catalyst in O2-saturated 0.1 M HClO4 at 5 mV/s before (solid) and after (dash) stability test (10 000 potential cycles between 0.6 and 1.0 V at 100 mV/s in N2-saturated 0.1 M HClO4). The inset shows the CVs in N2saturated 0.1 M HClO4 before and after 10K cycles. (C−E) Comparisons of ECSA, Pt-area normalized specific activity, and Pt-mass activity before and after stability test, respectively.
much more uniform Pt shell thickness of ca. 0.6 ± 0.2 nm (2−4 atomic layers) across the selected particle sizes (Figure 3D−F). Meanwhile, the Ni compositions at the core of the N2−D− PtNi3 catalyst are also higher compared to that for Air−D−
whole surface due to their relative slow diffusion rate. This is why a strong dependence of Pt shell thickness on the particle size is revealed in the Air−D−PtNi3 catalyst. In contrary, the solid nonporous N2−D−PtNi3 catalyst exhibited a smaller and D
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 5. Characterization of Air−D−PtNi3 (A−F) and N2−D−PtNi3 (G−H) after 10 000 cycles of stability test. (A) TEM image of Air−D−PtNi3 after stability test, showing porous nanoparticles with a pore size of of ca. 3−5 nm. More TEM images are provided in Supporting Information Figure S4. White arrows indicate larger particles with multipores (a and b, >13 nm) and medium size particles with single-pores (c and d, 9−13 nm). (B) Size-composition-porosity relationship in Air−D−PtNi3 before and after stability test, where the horizontal dash lines plotted the large area average compositions. The minimum particle size for porosity formation decreased from 13 to 9 nm. (C−F) STEM-EELS line scan analysis of the intraparticle compositional distributions in Air−D−PtNi3 nanoparticles with different sizes after stability test (G) TEM image of N2−D−PtNi3 after aging, indicating that most of particles remained nonporous. More TEM images are shown in Supporting Information Figure S5. (H) Sizecomposition-porosity relationship in N2−D−PtNi3 before and after stability test. Pores occur in a few of particles larger than ∼13 nm. (I) Size distributions of porous and nonporous particles in Air−D−PtNi3 and N2−D−PtNi3 catalyst after stability test analyzed from over 200 particles, which further confirms delayed formation of porosity in N2−D−PtNi3 catalyst.
provide a larger ESA for the catalysis; however, at the same time it also inevitably results in a reduced intrinsic specific activity. Therefore, nanoporosity can enhance the overall mass activity by increasing the surface area and is most effectively doing that in bulklike catalytic materials where surface area effects dominate. This explains the improved ORR activities observed in porous bulk alloys, such as porous leafs or foils26,29,30 or porous nanowires.25,27 However, in small dealloyed bimetallic nanoparticles where surface lattice strain rather than intraparticle surface area is the key factor for activity enhancement, the concomitant loss of specific (intrinsic) surface catalytic activity due to nanoporosity formation typically more than counterbalances the effect of increased surface area, and therefore in this context nanoporosity does not bring any benefit for the overall mass activity. Our insights in the role and effect of nanoporosity on the catalytic activity of a bimetallic particle ensemble are important as they provide understanding of and guidance for the design of more active nanostructure electrocatalysts. Now we will turn to the effect of nanoporosity on catalyst stability. Despite their similar initial mass activities and identical ORR testing conditions, the nonporous N2−D−PtNi3 catalyst proved more durable after 10 000 cycles compared to the porous Air−D−PtNi3. The ORR polarization curve of the Air− D−PtNi3 catalyst shows a 24 mV degradation in terms of halfwave potential (Figure 4A), resulting in 59% loss of mass
PtNi3 catalyst, demonstrating that the solid core−shell architecture does protect against significant loss of Ni atoms from subsurface regions. To investigate the activity and stability of the catalysts, we measured the ORR voltammograms of the N2−D−PtNi3 and Air−D−PtNi3 catalysts on a rotating disk electrode (RDE) before and after an electrochemical stability test (Figure 4); the stability test consisted of 10 000 potential cycles between 0.6 and 1.0 V, typical conditions to test the long-term stability of ORR fuel cell catalysts. Surprisingly, despite an identical initial ORR measurement, the N2−D−PtNi3 fully retained its nonporous nanoparticle structure (Supporting Information Figure S3). Despite their vastly different particle morphologies, the Air−D−PtNi3 and N2−D−PtNi3 catalysts exhibited similar initial mass activities around 0.75 A/mg(Pt). The electrochemical Pt surface area (ECSA) of the Air−D−PtNi3 catalyst is larger than that of N2−D−PtNi3 catalyst, as expected from the appearance of nanoporosity in the former. The specific activity of the Air−D−PtNi3 catalyst, however, was much lower than that of the N2−D−PtNi3 catalyst. This is consistent with its lower Ni content, larger Pt shell thickness and hence a lower compressive Pt surface strain and ORR activity in Air−D− PtNi3.9,13,38 The similar mass activity of Air−D−PtNi3 and N2− D−PtNi3 is therefore a coincidental offset of their differences in the ECSA and specific activity. We therefore consider that the role of the nanoporosity during the electrocatalysis is largely to E
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 6. Stability test of N2−D−PtNi3−B catalyst with more uniform particle size distribution of 3−12 nm. (A,B) TEM images before and after 30 000 cycles between 0.6 and 1.0 V, showing that all the particles remained nonporous. (C) ORR voltammograms in O2-saturated 0.1 M HClO4 at 5 mV/s and CVs in N2-saturated 0.1 M HClO4 (inset) before and after stability test. (D) Mass and specific activity changes after 30K cycles, demonstrating both high mass activity and stability that fulfill the DOE 2017 target.
thickness than the nonporous particles of 7 nm (Figure 5C) and 9 nm (Figure 5D), are clearly demonstrated. In contrast to the Air−D−PtNi3 catalyst and despite identical stability test, the structure of the N2−D−PtNi3 electrocatalyst, were much more durable against porosity formation. Although a few large particles changed to be porous and hence showed largely decreased Ni-composition, all the particles below 13 nm retained their nonporous structure and showed a uniform and mild decrease in Ni content (Figure 5G and H, and more TEM images in Supporting Information Figure S5). A detailed size and porosity analysis of over 200 particles (Figure 5I) clearly confirmed that the porosity formation in the N2-leached sample was delayed and restricted to a few large particles. As a result, the N2−D−PtNi3 catalyst possessed a higher percentage of solid core−shell nanoparticles associated with higher retained Ni content, which is fully consistent with its improved intrinsic and Pt mass based catalytic stability in Figure 4C. In addition to the continuous leaching of the less noble alloy component,15,16,42 catalyst degradation though Ostwald-ripening and particle coalescence were also recently observed in Pt alloy fuel cell catalysts.43,44 In a Pt3Co catalyst aged by voltage cycling (0.6−1.0 V) in MEA at 80 °C, Xin et al. found that the Ostwald-ripening caused preferential dissolution of Co and Pt atoms from smaller particles followed by Pt redeposition onto larger particles;43 this led to thick Pt shells, reduced the compressive surface lattice strain, and hence lowered specific activity. Using the same Pt3Co catalyst, Yu et al. showed that particle coalescence that leads to decreased ECSA became a major role when aged on RDE at room temperature, while the Ostwald-ripening mechanism was a minor effect.44 In this study, we could not fully rule out the Ostwald-ripening mechanism, but if it is significant it would have equally affected the Air−D−PtNi3 and N2−D−PtNi3 catalyst, which cannot explain their different stability observed here. The particle-
activity from 0.74 to 0.30 A/mg (Figure 4E). In comparison, the N2−D−PtNi3 catalyst exhibits a 16 mV decrease in the halfwave potential (Figure 4B) and a less extent of activity degradation (45% loss from 0.76 to 0.41 A/mg). To understand the origin of the difference in stability of the two leached catalysts, we further carried out size-compositionporosity analysis over tens of catalyst particles after 10 000 cycles. Compared to the starting state before stability test, two main structural features can be observed in the tested Air−D− PtNi3 catalyst (Figure 5A,B, tested catalyst labeled “AFTER 10K”, more TEM images are provided in Supporting Information Figure S4): (i) All the porous nanoparticles larger than 13 nm underwent a continued and significant decrease of Ni content after the stability test. Some of the particles had even become pure Pt. Meanwhile, the multipores of ca. 1−2 nm in the particles before stability test now changed to few large pores of 3−10 nm (e.g., particles a and b in Figure 5A) suggesting a pore-coarsening effect associated with the continuous leaching of Ni during the stability test. (ii) The minimum particle size for the appearance of porosity, as indicated by the arrows in the x-axis of Figure 5B, decreased from ca. 13 to 10 nm. These porous particles between 10 and 13 nm, which became porous after stability testing, generally showed a single-pore structure (e.g., particles c and d in Figure 5A) and again significant decrease in Ni content. This singlepore structure quite resembles the hollow nanoparticle structure resulted from Kirkendall effect.39−41 However, whether the single-pores are open pores or closed hollow structure and their detailed formation mechanism need further studies. Figure 5C−F present the EELS line scans of the intraparticle compositional fine structures of the aged Air−D− PtNi3 particles with different sizes. The formation of nanopores in the particles of 11 nm (Figure 5E) and 20 nm (Figure 5F), associated with their much lower Ni content and larger Pt shell F
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
potential cycling, which became the major reason for the ORR activity degradation. Under O2-free leaching, nanoporosity formation was effectively delayed, showing more solid core− shell nanoparticles with Pt rich shells and a higher Ni composition retained at the core, which finally resulted in improved catalytic durability. Our study therefore highlights the more beneficial role of solid nonporous core−shell nanoparticles compared to porous nanoparticles in the ORR catalytic stability. On the basis of these findings, we further demonstrate that the ORR stability could further improve by controlling the particle size below ca. 10 nm to avoid nanoporosity. We are confident that these results will have important implications for the future development of durable Pt alloy fuel cell cathode catalysts.
coalescence mechanism and hence lower ECSA should be also a minor effect; instead, looking at its sharp drop, it is the specific intrinsic activity that largely causes the loss in overall mass activity. It should be noted here that the dominant degradation mechanisms of alloy nanoparticles are quite dependent on their initial composition. In most of the studies on Pt-rich alloys, the formation of nanoporous particles is scarce and only accounts for a very small fraction of the particles.16,42−44 Here, we address the degradation of initially less noble-metal-rich (here Ni-rich) alloy nanoparticles due to their higher initial intrinsic activity.8,9,13 In this case, nanoporosity and its implication on surface and bulk composition become significant, consistent to an earlier report.23 Our data evidence that nanoporosity formation is inseparably tied to a drastic continuous dissolution of Ni and hence a rapid drop in intrinsic specific activity. Moreover, nanoporosity appears to gradually extend from larger particles to smaller particles during extended electrode potential cycling, which became a major reason for the rapid and severe catalyst performance degradation. Our O2-free acid leaching protocol, however, can significantly and effectively delay nanoporosity formation resulting in Ni-rich solid core−shell N2−D−PtNi3 catalysts even at large particle sizes. Delaying the Ni loss, the intrinsic activity remained higher and the catalyst exhibited improved stability. In terms of practical catalyst design guidelines, our study highlights the important catalytic role of solid nonporous core− shell alloy nanoparticles and suggests a facile preparation strategy under oxygen-free leaching conditions. Also, considering that all particles smaller than 10 nm remained solid nonporous and hence higher Ni composition after the stability test regardless of their pretreatment (Figure 5B,H), we also suggest that the catalytic stability should further improve by controlling the particle size in the range of 5−10 nm. As a demonstration of the usefulness of our strategy, we studied the stability of another face-center cubic PtNi3 catalyst (PtNi3−B) with more uniform particle size distribution mainly between 3 and 12 nm (Supporting Information Figure S6A) and similar composition distribution (Supporting Information Figure S6B) as the 3−25 nm PtNi3 catalyst. Using oxygen-free chemical dealloying (N2 acid leaching), the obtained N2−D−PtNi3−B catalyst showed high mass activity combined with exceptionally high stability (Figure 6C,D); after the same testing protocol, its mass activity loss is only 11% after 10K cycles and 24% after 30K cycles, which fulfill the DOE 2017 target for catalyst stability. Indeed, we found that all the nanoparticles remained nonporous even after 30K cycles (Figure 6B) and showed a mild decrease in Ni content from Pt50Ni50 (N2-Dealloyed) to Pt61Ni39 (after 30K), which again supports our claim that the nonporous solid core−shell nanoparticles are more stable during the long-term operation. In conclusion, we have reported a facile approach to control the nanoporosity formation in a Pt bimetallic alloy catalyst under free acid leaching by controlling the effective OCP and hence the leaching rates using different atmospheres. While leaching in the presence of oxygen enabled porous nanoparticles at larger sizes (>13 nm), O2-free leaching condition effectively avoided nanoporosity even in large particles and kept them in a solid alloy core−Pt shell architecture. The oxygeninduced nanoporosity was intrinsically associated with a drastic loss of Ni and thick Pt shells and progressed from larger particles to smaller ones during the stability test by long-term
■
ASSOCIATED CONTENT
S Supporting Information *
The contents of Supporting Information include the following: Figure S1: Structural characterization of 3−25 nm PtNi3 catalsyt received from Johnson Matthey Fuel Cells. Figure S2: Evolution of open circle potential of PtNi3 catalyst in 0.1 M HClO4 solution under different atmosphere. Figure S3: TEM image of N2−D−PtNi3 catalyst after initial ORR test. Figure S4 and S5: More TEM and HRTEM images of Air−D−PtNi3 and N2−D−PtNi3 catalyst after 10K cycles, respectively. Figure S6: TEM image and size-composition analysis of as-received 3−12 nm PtNi3−B catalyst from Johnson Matthey Fuel Cells Ltd. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank the Zentraleinrichtung für Elektronenmikroskopie (Zelmi) of the Technical University Berlin for their support with TEM and EDX techniques. This work was supported by U.S. DOE EERE award DE-EE0000458 via subcontract through General Motors and by Ernst Ruska Center for Microscopy and Spectroscopy with Electrons, Forschungszentrum Juelich GmbH, Germany. P.S. acknowledges financial support through the Cluster of Excellence in Catalysis (UniCat) funded by DFG and managed by TU Berlin.
■
REFERENCES
(1) Gasteiger, H. A.; Markovic, N. M. Science 2009, 324, 48. (2) Wagner, F. T.; Lakshmanan, B.; Mathias, M. F. J. Phy. Chem. Lett. 2010, 1, 2204. (3) Debe, M. K. Nature 2012, 486, 43. (4) Liu, Z.; Yu, C.; Rusakova, I.; Huang, D.; Strasser, P. Top. Catal. 2008, 49, 241. (5) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493. (6) Wang, C.; Chi, M. F.; Li, D. G.; Strmcnik, D.; van der Vliett, D.; Wang, G. F.; Komanicky, V.; Chang, K. C.; Paulikas, A. P.; Tripkovic, D.; Pearson, J.; More, K. L.; Markovic, N. M.; Stamenkovic, V. R. J. Am. Chem. Soc. 2011, 133, 14396. (7) Adzic, R. R.; Zhang, J.; Sasaki, K.; Vukmirovic, M. B.; Shao, M.; Wang, J. X.; Nilekar, A. U.; Mavrikakis, M.; Valerio, J. A.; Uribe, F. Top. Catal. 2007, 46, 249. G
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
(8) Debe, M. K.; Steinbach, A. J.; Vernstrom, G. D.; Hendricks, S. M.; Kurkowski, M. J.; Atanasoski, R. T.; Kadera, P.; Stevens, D. A.; Sanderson, R. J.; Marvel, E.; Dahn, J. R. J. Electrochem. Soc. 2011, 158, B910. (9) Strasser, P.; Koh, S.; Anniyev, T.; Greeley, J.; More, K.; Yu, C. F.; Liu, Z. C.; Kaya, S.; Nordlund, D.; Ogasawara, H.; Toney, M. F.; Nilsson, A. Nature Chem. 2010, 2, 454. (10) Cui, C.; Gan, L.; Li, H.-H.; Yu, S.-H.; Heggen, M.; Strasser, P. Nano Lett. 2012, 12, 5885. (11) Strasser, P. Rev. Chem. Eng. 2009, 25, 255. (12) Oezaslan, M.; Strasser, P. J. Power Sources 2011, 196, 5240. (13) Gan, L.; Heggen, M.; Rudi, S.; Strasser, P. Nano Lett. 2012, 12, 5423. (14) Wagner, F. T. High-activity Dealloyed Catalysts. 2012 DOE Hydrogen Program Annual Merit Review FC-087, http://www. hydrogen.energy.gov/pdfs/review12/fc087_wagner_2012_o.pdf; accessed Nov 13, 2012. (15) Mayrhofer, K. J. J.; Hartl, K.; Juhart, V.; Arenz, M. J. Am. Chem. Soc. 2009, 131, 16348. (16) Carlton, C. E.; Chen, S.; Ferreira, P. J.; Allard, L. F.; Shao-Horn, Y. J. Phy. Chem. Lett. 2012, 3, 161. (17) Oezaslan, M.; Heggen, M.; Strasser, P. J. Am. Chem. Soc. 2012, 134, 514. (18) Wang, D.; Yu, Y.; Xin, H. L.; Hovden, R.; Ercius, P.; Mundy, J. A.; Chen, H.; Richard, J. H.; Muller, D. A.; DiSalvo, F. J.; Abruña, H. D. Nano Lett. 2012, 12, 5230. (19) Snyder, J.; McCue, Ian; Livi, Ken; Erlebacher, J. J. Am. Chem. Soc. 2012, 134, 8633. (20) Chen, S.; Sheng, W. C.; Yabuuchi, N.; Ferreira, P. J.; Allard, L. F.; Shao-Horn, Y. J. Phys. Chem. C 2009, 113, 1109. (21) Yu, Z.; Zhang, J.; Liu, Z.; Ziegelbauer, J. M.; Xin, H.; Dutta, I.; Muller, D. A.; Wagner, F. T. J. Phys. Chem. C 2012, 116, 19877. (22) Dutta, I.; Carpenter, M. K.; Balogh, M. P.; Ziegelbauer, J. M.; Moylan, T. E.; Atwan, M. H.; Irish, N. P. J. Phys. Chem. C 2010, 114, 16309. (23) Liu, Z.; Xin, H.; Yu, Z.; Zhu, Y.; Zhang, J.; Mundy, J. A.; Muller, D. A.; Wagner, F. T. J. Electrochem. Soc. 2012, 159, F554. (24) Snyder, J.; Fujita, T.; Chen, M. W.; Erlebacher, J. Nat. Mater. 2010, 9, 904. (25) Cui, C. H.; Li, H. H.; Yu, J. W.; Gao, M. R.; Yu, S. H. Angew. Chem., Int. Ed. 2010, 49, 9149. (26) Wang, R. Y.; Xu, C. X.; Bi, X. X.; Ding, Y. Energy Environ. Sci. 2012, 5, 5281. (27) Shui, J. I.; Chen, C.; Li, J. C. M. Adv. Funct. Mater. 2011, 21, 3357. (28) Wang, D.; Zhao, P.; Li, Y. Sci. Rep. 2011, 1, 37. (29) Ding, Y.; Chen, M. W.; Erlebacher, J. J. Am. Chem. Soc. 2004, 126, 6876. (30) Ding, Y.; Kim, Y. J.; Erlebacher, J. Adv. Mater. 2004, 16, 1897. (31) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal., B 2005, 56, 9. (32) Erlebacher, J. Phys. Rev. Lett. 2011, 106, 225504. (33) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Nature 2001, 410, 450. (34) Erlebacher, J. J. Electrochem. Soc. 2004, 151, C614. (35) Wang, C.; Chi, M. F.; Wang, G. F.; van der Vliet, D.; Li, D. G.; More, K.; Wang, H. H.; Schlueter, J. A.; Markovic, N. M.; Stamenkovic, V. R. Adv. Funct. Mat. 2011, 21, 147. (36) Heggen, M.; Oezaslan, M.; Houben, L.; Strasser, P. J. Phys. Chem. C 2012, 116 (36), 19073−19083. (37) Gan, L.; Yu, R.; Luo, J.; Cheng, Z. Y.; Zhu, J. J. Phy. Chem. Lett. 2012, 3, 934. (38) Zhang, J. L.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R. Angew. Chem., Int. Ed. 2005, 44, 2132. (39) Yin, Y.; Rioux, R. M.; Erdonmez, C. K.; Hughes, S.; Somorjai, G. A.; Alivisatos, A. P. Science 2004, 304, 711. (40) Dubau, L.; Durst, J.; Maillard, F.; Guetaz, L.; Chatenet, M.; Andre, J.; Rossinot, E. Electrochim. Acta 2011, 56, 10658.
(41) Wang, J. X.; Ma, C.; Choi, Y.; Su, D.; Zhu, Y.; Liu, P.; Si, R.; Vukmirovic, M. B.; Zhang, Y.; Adzic, R. R. J. Am. Chem. Soc. 2011, 133, 13551. (42) Chen, S.; Gasteiger, H. A.; Hayakawa, K.; Tada, T.; Shao-Horn, Y. J. Electrochem. Soc. 2010, 157, A82. (43) Xin, H. L.; Mundy, J. A.; Liu, Z. Y.; Cabezas, R.; Hovden, R.; Kourkoutis, L. F.; Zhang, J. L.; Subramanian, N. P.; Makharia, R.; Wagner, F. T.; Muller, D. A. Nano Lett. 2012, 12, 490. (44) Yu, Y.; Xin, H. L.; Hovden, R.; Wang, D.; Rus, E. D.; Mundy, J. A.; Muller, D. A.; Abruña, H. D. Nano Lett. 2012, 12, 4417.
H
dx.doi.org/10.1021/nl304488q | Nano Lett. XXXX, XXX, XXX−XXX
