Recent Advances in the Stabilization of Platinum Electrocatalysts for Fuel-Cell Reactions


  • Dr. Minna Cao,

    1. State Key Laboratory of Structural Chemistry, Key Laboratory of Coal to Ethylene Glycol and Its Related Technology, CAS, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Science, 155 Yangqiao West Road, Fuzhou Fujian 350002 (P. R. China), Fax: (+86) 591-83796710
    Search for more papers by this author
  • Dongshuang Wu,

    1. State Key Laboratory of Structural Chemistry, Key Laboratory of Coal to Ethylene Glycol and Its Related Technology, CAS, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Science, 155 Yangqiao West Road, Fuzhou Fujian 350002 (P. R. China), Fax: (+86) 591-83796710
    Search for more papers by this author
  • Prof. Rong Cao

    Corresponding author
    1. State Key Laboratory of Structural Chemistry, Key Laboratory of Coal to Ethylene Glycol and Its Related Technology, CAS, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Science, 155 Yangqiao West Road, Fuzhou Fujian 350002 (P. R. China), Fax: (+86) 591-83796710
    • State Key Laboratory of Structural Chemistry, Key Laboratory of Coal to Ethylene Glycol and Its Related Technology, CAS, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Science, 155 Yangqiao West Road, Fuzhou Fujian 350002 (P. R. China), Fax: (+86) 591-83796710

    Search for more papers by this author


Polymer electrolyte membrane fuel cells (PEMFCs) feature high energy densities, low operating temperatures, and low environmental impact, which make them a promising technology for power applications. As a key component of PEMFCs, Pt-based catalysts are still under widespread investigation and have shown exciting performance; however, to move towards their successful commercialization, focusing solely on their catalytic activity is not sufficient. Instead, more effort is required to improve their stability and to decrease costs. Herein, we provide a comprehensive review of current research activities that have concentrated on how to stabilize the Pt-based catalysts. We devote the most attention to the structure-optimization of the Pt-based catalysts and the development of advanced supports. The feasible strategies for structure optimization are subdivided into three groups: 1) dimension effects; 2) electronic and bifunctional effects; and 3) steric effects. Then, we discuss the techniques that have been developed for improving carbon black and for generating various types of carbon-free supports and composites supports (e.g., graphite, carbon nanotubes, new-type oxides and nitrides, and macromolecules). An outlook on the future trends and developments in this area is also provided at the end of the review.

1. Introduction

Among the various types of fuel cells, polymer electrolyte membrane fuel cells (PEMFCs) are generally considered to be viable candidates for various power applications.1 The commercialization of PEMFCs is the most promising solution for the rapid growth in energy demands and the worsening environmental pollution.2 Figure 1 shows a typical PEMFC with H2 as a fuel. PEMFCs generate power by coupling the catalyzed oxidation of fuels at the anode to the reduction of oxygen at the cathode, thereby producing an electrical current through an external circuit. In addition to H2, the fuel can be MeOH, EtOH, formic acid, and other small organic molecules. State-of-the-art electrocatalysts for PEMFC reactions primary consist of nanostructured Pt-group metals that are loaded onto a carbon support and then embedded into a polymer matrix.3

Figure 1.

A simplified representation of a H2-fed PEMFC.

Until now, Pt and Pt-based alloys have been the only available electrocatalysts for practical PEMFCs. The short-term performance of Pt-based catalysts has met the demands of such applications. However, to move towards their successful commercialization, solely focusing on catalytic activity will not be sufficient. Instead, more efforts are needed to improve their long-term stability and to decrease costs.4 Hoping to achieve a breakthrough, researchers have devoted great efforts to studying the degradation mechanism of these catalysts. These reports have indicated that corrosion of the carbon support and metal dissolution/aggregation are the major contributors to catalyst degradation.1d, 3, 5 Therefore, a rational construction of catalysts from metals with optimized activities and durabilities and from supports with enhanced stabilities is a key point for improving PEMFCs.1d, 6

In recent years, increasing efforts have focus on the preparation of stable and long-lifetime Pt-based electrocatalysts. In this review, we summarize the current strategies for stabilizing Pt-based catalysts in reactions related to PEMFCs and classify these approaches into two groups: 1) Structure optimization of the metal itself; dimensional effects, steric effects, and introducing multiple-metal elements are shown to afford enhanced stability. (2) The development of advanced support materials, such as graphite, carbon nanotubes, new-type oxides and nitrides, and composite supports. We have attempted to elaborate on these approaches by using up-to-date examples in such a rapidly growing field. Finally, an outlook on the future trends and developments in this area is presented.

2. Structure Optimization

Composition, size, and shape are three typical parameters for tailoring catalytic performance.7 Composition, including the constituent elements and their ratio, is the basic essence of a catalyst and is directly related to a catalyst′s shape and size. The size determines the specific surface area and the ratio of surface to bulk atoms. The shape controls the facets and, thus, the surface structure of a nanocrystal, as well as the fractions of atoms at its corners and edges. It is generally accepted that the activity for the oxygen-reduction reaction (ORR) decreases in the order: Pt (110)>Pt (111)>Pt (100) for pristine single-crystal metal particles in HClO4 solution.8 Moreover, Pt (111) has the best stability and significantly lower poisoning rate. However, extension of the results from single-crystal substrates to nanocrystals can be quite problematic in some case, owing to the involvement of atoms at the corner and edges, the spatial shapes, structural defects, etc. For example, structure defects may be beneficial for activity in certain reactions but disadvantageous for the stability.2b, 9

To date, high-quality Pt and Pt-based nanomaterials with different shapes have been conveniently obtained through a kinetics-controlled or thermally controlled process. These rich architectures include nanocubes,8b, 10 nanorods,11 nanowires,12 nanotubes,13 nanothorns,14 nanocages,15 polyhedra,9a, 16 dendrites,17 hollow structures,18 porous structures,19 and concave structures.20 However, not all of the shapes are stable in an electrochemical environment. In terms of stability, we should take full account of all of the factors that are involved in structural optimization, which can be classed into three aspects: dimension effects, electronic and bifunctional effects, and steric effects.

2.1. Dimension effects

Dimensionality plays a critical role in determining the properties of materials, owing to, for example, the different ways that electrons interact in the structures.2b, 21 In general, 0D ultrasmall structures make the processes of dissolution and ripening easier during the fuel-cell operation, which results in the rapid loss of activity. In contrast, 1D structure motifs, 2D nanosheets, and 3D array networks could overcome these shortcomings and preserve their activity for a long time, thereby improving the stability.

For 1D/2D systems, such as nanowires, nanotubes, and nanosheets, the inherent anisotropy is central to the unique properties that these materials exhibit, thus making them particularly appealing in fuel cells. In contrast to their corresponding 0D nanomaterials, highly anisotropic growth endow these materials with many unique properties, such as long segments of smooth crystal planes, the preferential exposure of highly active/low-energy facets, higher aspect ratios, fewer defect sites, and enhanced electron transport.2b, 11a In particular, such structural anisotropy largely contributes to the enhanced stability by removing defect sites, which are susceptible to oxidation and decomposition. That is, the asymmetry of this structure suppresses the physical ripening process and, thus, these nanomaterials are inherently stabilized from dissolution and Ostwald ripening.2b, 22 Therefore, in theory, these low-dimensional nanostructures would be less prone to require a carbon support, because they do not suffer from the same propensity to ripen and aggregate.22

Because 1D and 2D nanomaterials have many advantages, they are currently the main research subject of many groups. Recently, Yamauchi and co-workers11c synthesized self-supported 1D Pt nanorods with various lengths from 0.6–3.2 μm (Figure 2 a, b). More importantly, the 1D nanorods showed a high density of mesopores; the size and wall-thickness of the mesopores were approximately 6–8 nm and 2–3 nm, respectively. The as-prepared mesoporous nanorods exhibited significant stability; they only lost 25 % of the initial Pt electrochemical active surface area (ECSA) after 1000 cycles and 31 % after 2000 cycles. Besides single-component 1D Pt nanowires, multicomponent Pt-based 1D materials have also been widely studied. Sun and co-workers12c prepared FePt and CoPt nanowires (NWs) with tunable compositions by using a simple organic-phase decomposition of the corresponding precursors (Figure 2 c, d). On treatment with acetic acid, the FePt NWs became highly active and stable towards the ORR. After 4000 cycles in O2-saturated HClO4, the Fe20Pt80 NWs lost a very small amount of their initial ECSA and their ORR polarization curve overlapped with that before the stability test. Ding et al.13b synthesized porous PtNiP composite nanotube arrays (NTAs), which incorporated non-metal elemental P onto a conductive substrate (Figure 2 e, f). In addition to enhanced activity, the as-prepared catalysts showed good long-term stability for the methanol-oxidation reaction (MOR). Recently, 2D nanostructures have emerged as promising long-lasting fuel-cell catalysts. Yang and co-workers23 presented an in situ chemical-vapor reaction in a molten salt for the synthesis of 2D Pt nanosheets (Figure 2 i). Such nanosheets were surfactant-free and carbon-free and, hence, were highly active for the ORR. Importantly, as shown in Figure 2 j, k, the as-prepared Pt nanosheets showed little loss in their ECSA during the accelerated durability test (after 30 000 cycles), which is a unique advantage of 2D materials and affords them with great potential for their practical applications.

Figure 2.

a, b) SEM images of the surface (a) and cross-section (b) of 1D mesoporous Pt nanorods (length: 3.2 μm). Inset in (a) shows a geometric model (Adapted from Ref. 11c). c, d) TEM images of 1D FePt (c) and CoPt nanowires (d; Reproduced from Ref. 12c). e, f) SEM image of PtNiP 1D porous hollow nanotubes (e) and a broken nanotube (f; Reproduced with permission from Ref. 13b. Copyright 2012, American Chemical Society). g) FESEM and h) TEM images of 3D ultrathin Pt nanowire assemblies. Inset in (g) shows a photograph of a Pt nanowire membrane (length: 2 cm; reproduced with permission from Ref. 12e. Copyright 2013, American Chemical Society). i) SEM image of a 2D Pt nanosheet. Inset shows the thickness of the sheets. j) CVs and k) polarization curves of the nanosheets before and after 30 000 potential cycles (Reproduced with permission from Ref. 23. Copyright 2012, the Royal Society of Chemistry).

The dimensions of some 3D nanostructures, such as highly branched structures,17b, 19e porous networks,19f and hollow nanocages,18a are much larger (usually 50–200 nm) than those of conventional nanocatalysts particles. However, these structures have low Pt consumption and broad spatial distributions of the Pt sites, thereby exhibiting good resistance to aggregation. For example, PtNi bimetallic nanobundles (NBs) with a highly branched morphology and stepped surfaces have recently been prepared by Li′s group.17c The size of PtNi NBs can reach 100 nm and the arm length can be tuned to a maximum of (54.3±10.3) nm. With an aggregation-resistant star-like morphology, the PtNi NBs retained 55 % of their initial peak current density after 4000 cycles for the MOR, whilst the conventional Pt nanoparticles (NPs) only retained 10 % of their current density. Xia et al.12e demonstrated a facile, template-free solvothermal synthesis of nanowire assemblies that were composed of ultrathin (about 3 nm) and ultralong Pt nanowires (up to 10 μm; Figure 2 g, h). The amine molecules that were generated during the reaction induced the nanowires to self-assemble into 3D interconnected networks that featured high surface areas and large porosities. The Pt nanowires exhibited remarkably high stability for the formic-acid-oxidation reaction (FAOR), the MOR, and the ORR. After 1000 and 3000 accelerated cycles, the Pt nanowires only lost 5 % and 10 % of its initial ECSA, respectively, whereas commercial Pt/C lost 75 % and 93 %, respectively. Impressively, the Pt nanowire only showed 37 % and 31 % loss of activity for the FAOR and MOR, respectively. In vast contrast, the data for commercial Pt/C were 79 % and 95 %, respectively. Such enhanced stabilities benefited from the anisotropic interconnected nanowires, which preferentially improved electron transport and catalyst utilization.

2.2. Electronic and bifunctional effects: Pt-based bimetallic/trimetallic catalysts

To decrease Pt utilization and further improve catalytic performance, including both activity and stability, the introduction of a foreign metal is the most promising strategy. Bifunctional and electronic effects (ligand effect) are proposed to account for enhanced bimetallic/trimetallic electrocatalytic performance. In terms of bifunctional effects, such as in the MOR on PtM, the foreign metal is employed for water dissociation at lower potentials, thereby promoting the oxidation of CO on Pt sites and enhancing the durability. Regarding electronic effects, foreign metals can alter the electronic properties of Pt and lower the adsorption energy of CO, thereby facilitating the oxidation of CO at lower potentials. It is difficult to define the exact mechanism for a bimetallic system. In most cases, bifunctional and electronic effects would simultaneously contribute to an electrochemical process. For example, Ataee-Esfahani et al.19h prepared hollow mesoporous PtRu alloy particles as highly CO-tolerable catalysts for the MOR. They indicated that such superiority originated from two aspects, a downshifted d-band center (electronic effect) and adsorbed oxygen-containing species on Ru atoms (bifunctional effect).

In general, there are three possible types of mixing patterns between two or three metals:24, 25 1) Homogeneously mixed alloys, in which the two metals are mixed in either an atomically ordered (Figure 3 a, top) or a statistically random manner (Figure 3 a, bottom). Ordered intermetallic alloys have long-range atomic order and are less-commonly reported than random alloys. 2) Heterostructures (also called sub-cluster-segregated nanoalloys), in which the two metal components share the same interface (Figure 3 b). 3) Core–shell structures, which consist of a shell (single shell) of one type of atom that surrounds a core of another type of atom (Figure 3 c<xfigr3, top). Multiple-shell structures with layered or onion-like shells have also been experimentally produced (Figure 3 c, bottom). With the development of various theoretical and synthetic techniques, an abundance of composites structures have been experimentally prepared, such as a core–shell structures that consists of an alloyed core or shell (e.g., Pt3Fe2@Pt core–shell structures with an alloyed core26 and Au@FePt3 core–shell structures with an alloyed shell27). Such composite structures derive good qualities from its constituents and, hence, are of great interest and importance for various applications.

Figure 3.

Bimetallic/trimetallic nanocrystals with different mixing patterns: a) Alloy structures: ordered intermetallic alloy (top) and random alloy (bottom); (b) heterostructures; c) core–shell structures: single-shell structure (top) and multi-shell structure (bottom; Adapted with permission from Ref. 24. Copyright 2008, American Chemical Society).

2.2.1. Alloys

The alloys of Pt with various base metals, such as V,28 Cr,29 Co,30 Ti,31 Pb,17d, 32 Hg,33 Cu,34 and Ni,16b and of some noble metals, such as Pd,18a, 35 Ru,19i, 36 Au,37 and Ag,38 have been found to exhibit significantly higher electrocatalytic stabilities and better activities towards the ORR in PEMFCs than Pt alone. Such enhancement can be ascribed to different structural changes that are caused by alloying, such as geometrical factors (decreased Pt[BOND]Pt bond length),39 dissolution of the more oxidizable alloying component,40 changes in surface structure,41 or electronic factors (increased Pt d-band electrochemical vacancy of the Pt skin layer, which originates from the bulk alloy).28, 42

The catalytic stability and activity of PtM alloy catalysts not only depends the nature of Pt, but also on the second metal. In terms of catalytic activity for the ORR and the tolerance towards chemical corrosion, He et al.43 classified the PtM (M=base metal) alloy cathodic catalysts into four categories: 1) Highly corrosive and highly active (M=Fe, Co, V, Mn); 2) corrosive and highly active (M=Zn, Cu, Mo, Ni); 3) stable but less active (M=Zr, Cr, Ta); and 4) stable and active (M=W, Ti). However, the leaching of the base metals from the alloys should be carefully handled. On one hand, it is thought that the loss of the base metal from the PtM catalysts can increase their stability/activity through surface roughening and, hence, increase the Pt surface area, or by modification of the electronic structure of the Pt skin layer. For example, Yu′s group19b utilized the dissolution of the base metal to reconstruct more stable catalysts. They found that the ECSA of PtNi disordered alloyed nanotubes (PNPT) increased about 100-fold after 250 potential cycles, owing to the dissolution of Ni from the PtNi random alloy (Figure 4 c). The experimental data further indicated that such dissolution of Ni during potential cycling should be responsible for the persistent surface restructuring, increased surface roughness and lattice strain, and, hence, high catalytic activity, stability, and resistance to poisoning. As shown in Figure 4 a, b, the activity and CO tolerance of PNPT was much higher than that of Pt/C. More importantly, after the degradation test, the ECSA of PNPT only decreased by 10 % (Figure 4 d), whereas Pt black and Pt/C lost 35 % and 70 % of their initial ECSAs, respectively.

Figure 4.

a) CV and b) CA curves of Pt/C and PNPT catalysts for the MOR. c) Change in the CV of the PNPT catalyst with repeated cycling; the 1st and 250th cycles are indicated. d) ECSA of the PNPT catalyst after each 2000 s CA test in MeOH solution and 250 cycles in Ar-saturated HClO4. (Reproduced with permission from Ref. 19b. Copyright 2011, The Royal Society of Chemistry).

However, in most case, the loss of the base metal will result in decreased stability/activity because of the change in surface structure and contamination of the electrochemical environment. To obtain highly active and stable catalysts, a promising synergistic effect has been proposed, based on the four categories mentioned above. Active metals in categories 1 or 2 are introduced into stable and less-active binary alloys to enhance the activities. Similarly, stable metals in categories 3 or 4 are added to stabilize or passivate the active and corrosive binary alloys. As such, there has been a boom in the creation of multicomponent alloys, thus confirming that such synergistic effects are effective. For example, Jeon et al.44 introduced Cr into PtCo/C to form PtCoCr/C alloy catalysts. Over a 20 h chronoamperometric (CA) test, the MOR activity of a Pt20Co30Cr40 powdered alloy catalyst showed 160 % higher mass activity than that of a PtRu/C catalyst, thus suggesting greatly increased durability and activity. Sun and co-workers45 demonstrated that the ternary FePtCu alloy nanorods exhibited much better ORR activity and stability than commercial Pt/C catalysts and binary FePt nanorod catalysts. The CV polarization curves of the Fe10Pt75Cu15 nanorods before and after 5000 cycles almost overlapped, thus indicating significantly enhanced stability.

High-temperature thermal annealing28, 46 and the formation of ordered intermetallic alloys31, 47 have also been shown to minimize the base-metal contamination and to offer enhanced stability in PEMFCs. In general, high-temperature thermochemical treatment can give rise to the formation of atomic-scale, chemically ordered nanoalloys. For example, Wang and co-workers34 prepared ordered intermetallic Cu3Pt/C catalysts with a Pt-rich skin after annealing at 1273 K under a H2 atmosphere, whereas a random CuxPt/C alloy was obtained at 573 K. Wanjiala et al.46 observed that PtVCo catalysts went through a phase-transition on treatment at elevated temperatures, from a chemically disordered fcc-type structure to a long-range-ordered fct-type structure with an enhanced degree of alloying. An ordered alloy structure generally displays enhanced stability and activity. Specifically with regard to Pt alloys with first-row transition metals, Gonzalez and co-workers1a proposed that Pt[BOND]Cr and Pt[BOND]Co were generally more stable than Pt[BOND]V, Pt[BOND]Ni, and Pt[BOND]Fe in acidic media, because Cr and Co presented a higher degree of alloying with Pt than with V, Ni, and Fe. Quite recently, Botton′s group26 demonstrated a new intermetallic core–shell (IMCSs) regime in PtFe systems, which comprised an ordered Pt3Fe2 core that was encapsulated within a bilayer Pt-rich shell (about two atomic layers thick, induced by heat treatment; Figure 5 a–c). A minimal decrease (9 %) in catalytic activity, even after 6000 cycles, showed extended stability. Both the STEM-HAADFA image (particle center, Figure 5 e) and the XRD pattern (Figure 5 f) of the non-cycled Pt3Fe2 and after cycling about 10 000 times indicated that, despite a constant enrichment of the Pt shell over 10 000 electrochemical cycles, the particle still retained core-ordering. The changing tendency in the core and shell is directly reflected by the plots shown in Figure 5 g, h. Such endurance in the chemical order of the core is the major reason for the enhanced durability.

Figure 5.

a, b) STEM-HAADF images of Pt3Fe2 nanocatalysts, which show alternating bright and dark areas that correspond to Pt and Fe atomic columns, respectively. c) 3D model of a typical IMCS nanocatalyst (color code: Pt gray, Fe yellow). (d, e) Atomic-resolution STEM-HAADF images of the non-cycled (d) and cycled catalysts (e). f) XRD data of the bulk sample. g, h) Plot of ordered core volume (g) and shell volume (h) versus particle size. Particles within the size range 3.2–4.2 nm retained comparable core volumes, even after 10 000 cycles, albeit with an increase in their shell volume (Reproduced with permission from Ref. 26. Copyright 2013, American Chemical Society).

2.2.2. Heterogeneous structures

Pt-based heterogeneous structures, such as 1D heterojunction nanowires/dimers and 3D dendrites, are usually prepared by growing a small amount of Pt NPs on a less-expensive metal-NP substrate. Such structures can greatly lower the Pt consumption and broadly expand the spatially distribution of Pt sites, thereby exhibiting good resistance towards aggregation. In addition, owing to the third-body effect,48 electronic effects,49 increasing interface area,50 etc., heterogeneous bimetallic nanomaterials are highly stable and active towards fuel-cell reactions.

A classic example is PdPt bimetallic nanodendrites that were prepared through seeded-mediated growth by Xia′s group,17a which consisted of a dense array of Pt branches on a truncated octahedral Pd core. The Pt branches that were supported on faceted Pd NPs exhibited relatively large surface areas and particularly active facets toward the ORR and, hence, were much more active than commercial Pt catalysts. In addition, the stability of the nanodendrites was higher than that of commercial Pd catalysts. Yang′s50a group described a different approach for simultaneously addressing the activity and stability issues of Pt-on-Pd heterostructures. The Pd seeds were synthesized through an oleylamine-based procedure and, thus, were fairly monodispersed. The as-made catalysts exhibited superior performance for the ORR, especially long-term stability. After 30 000 cycles, the catalysts only lost about 12 % of the initial ECSA and showed a small degradation (9 mV) in the half-wave potential. The authors attributed the improved stability to the favorable interfacial structures between Pt and the Pd support. In addition, the heteronanostructures were larger than the typical overall particle size of Pt-on-Pd nanostructures, thus preventing the small Pt particles from dissolving during the ORR. Later, Gou et al.17b fabricated hybrid catalysts that consisted of 3D Pt-on-Pd nanodendrites that were loaded onto 2D graphene nanosheets that were decorated by PVP. Owing to the porous and high-surface-area properties of dendrite structures, the strong coordinative interactions between PVP and the nanodendrites, as well as the strong π[BOND]π interactions between graphene and PVP, the as-prepared catalysts exhibited much higher activity and stability toward the MOR than platinum black and commercial E-TEK Pt/C catalysts. Recently, Yamauchi′s51 group reported a facile one-pot formation of Pt-on-Pd nanodendrites by using Pluronic P123 as an additive, which exhibited high activity and durability for the MOR. Such an open dendritic structure of the Pt-on-Pd nanodendrites was highly beneficial for their superior resistance towards agglomeration of the active sites.

Recently, Zhang et al.49b demonstrated that Au clusters could be used to stabilize Pt ORR catalysts against dissolution (Figure 6 a). More importantly, there were no leakage issues on Au. The decoration of Au clusters on Pt greatly enhanced the stability of Pt, as shown in Figure 6 b, c; the Pt surface area changed little (only 4 % in ECSA) after a durability test (30 000 scans) in an O2-satrurated HClO4 solution. In situ X-ray absorption near-edge spectroscopy and voltammetry data suggested that the Au clusters conferred stability by raising the Pt-oxidation potential. Liu et al.48 modified tetrahexahedral Pt nanoparticles with Au adatoms. The Au decoration promoted the direct mechanism, thereby leading to highly active, stable, and poison-resistant anodic electrocatalysts for the FAOR.

Figure 6.

a) Electron micrographs of a Au/Pt/C catalyst. The different structure in the area indicated by the arrow is ascribed to Au clusters. Inset shows an STM (125 nm×125 nm) image. b, c) Polarization curves (b) and CVs (c) of the Au/Pt/C catalyst before and after 30 000 cycles. d) Comparison of the change in the absorption-edge peaks of the XANES spectra for Au/Pt/C and Pt/C catalysts as a function of potential, as obtained at four different potentials in 1 M HClO4 (Reproduced with permission from Ref. 49b. Copyright 2007, AAAS).

Trimetallic heterostructures, such as Pt-on-PdBi nanowires,52 with enhanced stability during electrocatalytic processes have also recently been reported.

2.2.3. Core-shell structures

Various methods have been developed for the preparation of core–shell NPs, such as galvanic replacement,53 seed-mediated growth,54 surface dealloying,34, 55 heat treatment,26, 34 and combinations of these.56 Core–shell structures represent an effective structural model, in which a monolayer or a few layers of Pt are arranged on less-expensive core materials.13d Such structure can greatly decrease the Pt utilization, because the majority of Pt atoms are distributed at the electrochemical reaction interface.3 More importantly, the structure and properties of a Pt shell can be effectively tailored by the core material, owing to structure-induced strain (geometry) and electronic effects (alloying). For example, the existence of subsurface Au atoms would make an additional contribution to the durability enhancement of the Pt skin layer by modifying the well-known core-hindered place-exchange mechanism.3, 57 Such a mechanism has been well-demonstrated in Au@FePt3 NPs,27 which were synthesized by mixing Au particles (7 nm) with [Fe(CO)5] and [Pt(acac)2] in a non-organic system (Figure 7 a, b). In addition to greatly enhanced activity, the Au@FePt3 NPs possessed superior durability and showed negligible morphology changes after a stability test of 60 000 cycles (Figure 7 c–e), much better than Pt/C and FePt3 NPs catalysts. The existence of thermodynamically stable Au in the subsurface layer makes the formation of subsurface oxide energetically harder and, thus, prevents dissolution.

Figure 7.

a) Schematic representation of the synthesis of core–shell Au@FePt3 NPs. b) HAADF-STEM elemental mapping of Au, Pt, and a single Au@FePt3 core–shell NP. c) Summary of the mass activities of different NPs before and after the cycling tests. d, e) TEM images of Au@FePt3 NP catalysts before (d) and after (e) stability tests over 60 000 cycles (Reproduced with permission from Ref. 27. Copyright 2011, American Chemical Society).

In recent years, the concept of Pt-monolayer catalysts for addressing the future challenges of limited Pt resources has been established. However, the lifetime that is required by the US Department of Energy (DOE) for these Pt catalysts requires the use of thick Pt layers or larger particles. To seek a balance between cost and long-term activity, a promising solution is controlling the core′s composition, size, and shape. Recently, Adzic′s group54b synthesized a NiN@Pt core–shell catalysts by encapsulating NiN core NPs inside 2–4-monolayer-thick Pt shells. The NiN@Pt catalysts had very good stability for the ORR, which only lost 11 mV in its half-wave potential after 35 000 cycles. Nitride caused a bifunctional effect, which facilitated the formation of the core–shell structures and improved the performance of the Pt shell through both geometric and electronic effects. Sasaki et al.58 demonstrated that a PtML/Pd/C catalyst (ML=monolayer) showed a small loss of ECSA after 100 000 potential cycles between 0.7–0.9 V during an accelerated durability test (ADT). On alloying the Pd core with Au to obtain a Pd9Au1 alloy core, the PtML/Pd9Au1/C catalyst showed a small loss of ECSA, even after 200 000 cycles, further showing the superiority of the composited structure. These experimental data and density functional theory (DFT) calculations indicated that the core could increase the stability of the Pt shell by positively shifting its oxidation potential and by preventing the cathode potential from reaching values at which Pt dissolution takes place. Further work on the PdAu alloy@PtML catalyst was performed.58b Although more severe ADT conditions (0.6–1.4 V) were employed in this case, there was no marked loss of Pt and Au, despite the dissolution of Pd. Such work clearly demonstrates the significant role of Au in the stability enhancement of the Pt monolayer.

Multi-shell-structured catalysts are also stable catalysts in an electrocatalytic environment. For example, Yamauchi′s group prepared a series of multilayered Au@Pd@Pt NPs with the assistance of polyvinylpyrrolidone (PVP)54a or Pluronic F12759 in a one-step process at room temperature. These trimetallic multilayered NPs showed higher activities and better durabilities than Au@Pt bimetallic core–shell structures and commercial Pt/C and Pt-black catalysts. Fang et al.60 studied the electrocatalytic performance for the FAOR on Au@Pd@Pt NPs that consisted of a gold core covered by a Pd shell, on which Pt clusters were deposited. By carefully investigating the synergistic effects between the three different nanostructure components (sphere, shell, and islands), they indicated that the optimized structure should only have two atomic layers of Pd and a half-monolayer equivalent of Pt. A further increase in the loading of Pd or Pt would actually decrease the catalytic performance.

2.3. Steric effects: Surface coating and nanoencapsulation

Recent studies have shown that a shell coating on a catalyst can act as a physical barrier to inhibit the migration and coalescence of metal particles and simultaneously provide mass-transfer channels during a reaction.61 Such steric effects can greatly enhance the stability and durability of a Pt catalyst, as demonstrated in several systems. Various materials have been explored as effective coating/nanoencapsulating agents, including inorganic mesoporous materials (e.g. silica,36a, 61a, 62 carbon shells,63 mesoporous niobium-oxide/carbon composites64) and organic materials (e.g. polymers,65 macrocyclic molecules66).

Konno′s group36a utilized silica-coupling agent (3-aminopropyl)trimethoxysiliane (APS) to stabilize PtRu/C catalysts. The catalytic activity of APS-coated PtRu/C for the MOR was unchanged, even after immersion in sulfuric acid solution for 1000 h, thus suggesting high stability of the PtRu catalyst. However, these silica layers probably decrease the electrochemical activity, because the reactant molecules, such as water and oxygen, need to be supplied through the silica layers and silica is an insulator. Therefore, the effects of thickness and density of the silica layers on the durability of such hybrid catalysts should be carefully investigated. Recently, Kishda and co-workers62c prepared multi-walled-carbon-nanotube-supported Pt (Pt/CNT) catalysts that were coated by silica with different thickness and densities. They indicated that Pt/CNT that was covered with silica layers (thickness: about 6 nm) exhibited a high ECSA, as well as excellent durability.

Recently, Galeano et al.63b developed a Pt NP catalyst (3–4 nm) that was encapsulated inside the pores of hollow graphitic spheres (HGS). The HGS had a specific surface area exceeding 1000 m2 g−1 and a precisely controlled pore structure (Figure 8 a–c). Such Pt catalysts were highly stable at 1123 K, as well as during simulated accelerated electrochemical aging in both half-cell measurements (Figure 8 d) and in a fully assembled PEMFC (Figure 8 e, f). HR-SEM and DF-STEM images (Figure 8 g–j) indicated that no significant particle growth occurred and only a small loss in the total particle number could be observed in comparison to commercial Pt/C (Figure 8 k, l). Zhao et al.63d synthesized highly ordered mesoporous platinum@grahitic carbon (Pt@GC) composites with well-graphitized carbon frameworks and uniformly dispersed Pt NPs embedded within the carbon pore walls. The integration of high surface area, regular mesopores, and graphitic nature of the carbon walls, as well as highly dispersed and spatially embedded Pt NPs in the mesoporous Pt@GC composites, made them highly active, extremely stable, and MeOH-tolerant electrocatalysts for the ORR.

Figure 8.

a) HRTEM image of the HGS support. b, c) Dark-field scanning transition electron microscopy (DF-STEM) images of cross-sectional cutting of Pt@HGS (b) and Pt@HGS (c) after thermal treatment at 1123 K. d) Ex situ and e, f) in situ electrochemical stability tests. g) Overlapping of the HR-TEM and DF-STEM images of Pt@HGS after 0 (g) and 3600 electrochemical degradation cycles (i). h, j) Simultaneously taken DF-STEM micrographs of the region highlighted by a white square in (g) and (i), respectively. k, l) Particle-size distributions at a typical location for Pt@HGS (k) and Pt/C (l) before and after 3600 degradation cycles (Reproduced with permission from Ref. 63b. Copyright 2012, American Chemical Society).

Genorio et al.66 demonstrated that chemically modified Pt with a self-assembled monolayer of calix[4]arene molecules could selectively block the ORR without affecting the activity and kinetics for the hydrogen-oxidation reaction (HOR). The surface coverage of calix[4]arene can lead to the formation of a critical ensemble of O2-tolerant Pt sites, which are active for the adsorption of H2 and consequent H[BOND]H bond breaking. Such selectivity could greatly improve the stability of a cathode Pt/C catalyst, which is strongly affected by an undesired ORR on the anode.

3. Development of Advanced Supporting Materials

The durability of a catalyst support is current a technical barrier for the commercialization of PEMFCs. At present, almost all PEMFCs employ carbon-supported catalysts, especially Vulcan XC-72 carbon black, because of its large surface area, high electrical conductivity, and well-developed pore structure. However, during fuel-cell start-up and shut-down, the carbon support (especially in the cathode) encounters severe corrosion, as indicated by the reaction shown in Equation (1).67

equation image(1)

Such poor resistance to corrosion can readily cause agglomeration and sintering of the Pt NPs, thereby resulting in low catalyst utilization and decreased fuel-cell performance. To overcome the shortcomings of the currently available Pt/C catalysts, the major aim is to develop potential supporting materials that can improve the Pt–support interactions, thereby simultaneously sustaining bulk-like catalytic activities with very highly dispersed particles. This latter aspect is key to meeting the 2015 DOE targets for platinum-group-metal loadings of 0.20 mg cm−2.68 To this end, the support materials should be low cost and have high surface areas, good electrical conductivity, and anticorrosive to not only maximize the availability of the electrocatalytic surface area for electron transfer, but also provide better mass transport of the reactants to the electrocatalyst.

Over the past few decades, increased efforts have been devoted to developing advanced supporting materials, mainly in two aspects. Firstly, various methods have been developed for boosting the currently used carbon materials, such as graphitization, nitrogen-atom doping, and surface treatment. Secondly, more anticorrosive carbon-free materials have been developed as alternative supports. In addition, hierarchical hybrid structures, combined with carbon and other non-carbon materials, which are complementary in both structure and properties, provide opportunities for the exploration of new types of supporting materials.

3.1. Carbon materials

3.1.1. Graphitization of the carbon supports

Carbon materials with a high degree of graphitization, such as graphitized carbon black, carbon nanotubes (CNTs), carbon fibers, carbon nanothorns, and graphene, are believed to be more corrosion-resistant than conventional carbon black.5a Enhanced graphitization leads to more delocalized π-bonds (sp2-hybridised carbon atoms), which strengthen the interactions with Pt and lead to anti-agglomeration of the Pt NPs on such supports.

The recent emergence of graphene nanosheets that are only one atom thick has opened up new avenues for utilizing 2D carbon materials as supports, because of their high conductivity (×103–×104 S m−1), huge surface area (calculated value: 2630 m2 g−1), unique graphitized basal-plane structure, high stability, and strong interactions with NPs.3, 5a, 69 Studies on fuel-cell catalysis indicate that a graphene support enhances both Pt activity and stability. Huang′s group70 inserted carbon black into reduced graphene oxide (RGO) sheets and used them as supports for Pt NPs in the ORR. An ADT revealed that the as-made catalysts only lost about 5 % of their ECSAs after 20 000 cycles. It is thought that the 2D profile of the RGO functions as a barrier to prevent Pt leaching. Carbon black in the vacancies acts as active sites to recapture/re-nucleate the dissolved Pt species. Similar work was also reported by He et al.,71 who synthesized nano-sandwiched graphene/carbon/graphene (GCG) composites as supports for Pt NPs, which exhibited very high activities and stabilities for the ORR. The authors ascribed the long-term performance of the catalysts to the presence of carbon nanospheres between the graphene nanosheet layers, which effectively prevented them from rapidly aggregating and improved the Pt dispersion during the electrochemical process. Maiyalagan et al.72 utilized 3D interconnected graphene monoliths as electrode supports for the pulsed electrochemical deposition of Pt NPs. The as-made catalysts exhibited significantly improved catalytic activities and stabilities for the MOR, owing to their 3D interconnected seamless porous structure, high conductivity, and the inexistence of junction resistance.

CNTs, including single-walled carbon nanotubes (SWCNTs) and multi-walled carbon nanotubes (MWCNTs), are made of rolled-up graphite sheets with a large 1D π-conjugated structure. Since the 1990s, CNTs have been recognized as a promising alternative support material for fuel-cell electrocatalysts, owing to its higher electrical conductivity, lower degree of impurities, higher catalyst-loading efficiency, higher durability, and higher surface-to-volume ratio compared to carbon black.73 Cao et al.74 prepared a Pt/RuO2x H2O/CNTs nanocomposite catalyst for the MOR. The experimental data indicated that the CNTs could compensate for the loss of electron conductivity that was caused by the RuO2x H2O coating and improved the microstructure and resistance of the electrode. Strasser′s group75 investigated the activity, stability, and degradation of Pt/MWCNTs for the ORR. Only trace amounts of migrating Pt NPs were present inside a MWCNT, compared to the Pt/Vulcan XC-72R catalysts, thus suggesting enhanced adhesion between the Pt atoms and the walls of the graphene tube.

3.1.2. Nitrogen-doped carbon supports

The doping of carbon nanostructures with nitrogen atoms has drawn much attention because conjugation between the nitrogen lone-pair and the graphene π system may create nanomaterials with tailored electronic and mechanical properties.3, 76 Recently, various N-doped carbon nanomaterials (NCNMs) were studied as efficient metal-free ORR catalysts in alkaline fuel cells. These NCNMs showed excellent electrocatalytic activity and better durability than commercial Pt-based catalysts.77 However, only a few papers have examined Pt-based catalysts that were supported on NCNM substrates. Sun et al.78 indicated that the appropriate N-doping of CNTs could prevent Pt agglomeration and result in monodispersed PtRuNi particles with sizes of about 3–4.5 nm. Chen et al.79 investigated the electrochemical stability of Pt on CNTs and N-doped CNTs (CNx) with different N content by performing accelerated durability tests. The experimental data indicated that Pt/CNx exhibited much higher stability than Pt/CNTs and that the stability increased with increasing N content (from 1.5–8.4 at. %). The more delocalized π-bonding, lone electron pair, and strong binding between Pt and CNx were responsible for such enhanced stability.

Disappointingly, most of the N-doped carbon materials showed poor stability/activity in acid electrolytes, which greatly hampered their application in PEMFCs. In addition, further studies are necessary to fully understand the implications of N-doping as it pertains to the improved kinetics/thermodynamics for fuel-cell reactions.

3.1.3. Surface-treatment of carbon supports

Pt/carbon-support interactions are beneficial for the enhancement of catalytic activity and stability of an electrocatalyst.80 Strategies for enhancing such interactions in Pt/C catalysts are mainly based on developing strong anchoring sites at the Pt/C interface through modification of the carbon surface. Two approaches have been adopted in the literature for achieving such a goal.

The first approach is based on the oxidative treatment of carbon. The oxidative treatment of a carbon surface, which can be performed through chemical treatment with various oxidants, including HNO3, H2O2, O2 or O3, and KMnO4, gives rise to the formation of surface acidic sites.80a Stable surface oxygen groups on a carbon support can offer effective anchoring sites for Pt NPs, thereby enhancing the stability of the catalyst. For example, Prado-Burguete et al.81 treated high-surface-area carbon black with solutions of H2O2 and Pt dispersion increased with increasing amount of surface oxygen groups. The chemical-oxidation method has also been widely used for the surface treatment of carbon nanotubes. Pristine CNTs are chemically inert and typically do not provide enough surface charge for the deposition of metal nanoparticles.69a By heating at reflux in strongly acidic oxidants to produce functionalities that are more suitable anchoring sites for Pt NPs, such as COOH, OH, and C[DOUBLE BOND]O groups, more uniform distribution and higher loadings of Pt NPs are obtained. Li and Hsing82 indicated that the electrocatalytic performance of Pt/CNTs catalysts for the ORR was greatly enhanced by the presence of oxygen-containing groups on the CNT surface.

The second approach is based on heat treatment. A previous discussion indicated that heat treatment had an impact on the properties of Pt and PtM species, such as size, surface composition, crystallinity, and degree of alloying. Heat treatment of carbon materials allows the removal of undesired impurities that are formed during the early preparation stages, thereby allowing uniform dispersion and stable distribution of the metal species over the support.43b Moreover, heat treatment also stabilizes the carbon supports against electrochemical corrosion, which most probably results from the induced increases in porosity and the development of graphitic structures at elevated temperatures.83 Therefore, Pt NPs on heat-treated carbon supports are more stable than those on untreated supports, owing to the strong interactions between the Pt NPs and the carbon surface. For example, Kim et al.84studied the catalytic behavior of Pt/NiO[BOND]C electrocatalysts after heat-treatment at different temperatures from 200–800 °C under a flow of N2. According to the experimental data, heat treatment can reduce NiO into Ni and then form an alloy with Pt, thereby resulting in a down-shift of the Pt 4f electrons. Pt/NiO[BOND]C electrocatalysts that were heat-treated at 400 °C showed the highest activity towards the MOR and minimal aggregation of the Pt NPs, thus indicating that heat treatment played a significant role in the improvement of both activity and stability.

3.2. Macromolecules

Conjugated heterocyclic conducting polymers, such as polypyrrole (PPy), polyaniline (PANI), polyacetylene (PA), and polythiophene (PTh), have received more and more attention since the late 1970s, because of their unique metallic/semiconductor characteristics and potential use in various areas, such as electronics, electrochemistry, and elctrocatalysis.85 In particular, extensive research has focused on PPy-containing anode and cathode catalysts in the last twenty years, because PPy has excellent environmental stability, facile synthesis, and high conductivity. For example, Huang et al.86 synthesized PPy through an in situ chemical oxidative polymerization method that was directed by modified self-degradable templates with sodium dodecyl sulfate as a surfactant. Electrochemical experiments revealed that the obtained PPy was much more resistive than conventional carbon Vulcan XC-72. Because the oxidation of PPy occurs at high positive potentials, this leads to better electrochemical stability of Pt/PPy than Pt/C.

In addition to the conductive polymers, non-conductive macrocyclic host molecules, which are central to supramolecular chemistry, such as calixarenes,87 cyclodextrins,88 and cucurbiturils,89 are emerging as promising materials for the synthesis and assembly of metal NPs. However, in the case of calixarenes and cyclodextrins, their use as supports for Pt NPs for fuel-cell reactions are seldom reported. Our group recently utilized cucurbit[6]uril (CB[6]) as both a stabilizing agent and a support for the synthesis of sub-10 nm CB[6]-Pt NPs.90

CB[6] has a rigid pumpkin-shaped structure and features high thermal and chemical stabilities. As shown in Figure 6 a, the surface of the hydrophobic cavity of CB[6] is negatively charged, whereas the outer surface is somewhat positively charged. In addition, the two identical portals of CB[6] that are lined by polar carbonyl groups provide two negative fringes that are capable of binding to other molecules or atoms through electrostatic interactions. The electrostatic interactions between CB[6] and the surface atoms of metal NPs, as well as volume effects, could stabilize metal NPs from aggregation. Moreover, such weak interactions do not block the active sites on the surface of metal NPs and, hence, endow them with high catalytic activity. Owing to the presence of CB[6], all three CB[6]-Pt NPs catalysts exhibited unprecedented catalytic activity and excellent resistance to poisoning, as well as much improved stability compared to commercial Pt/C (Figure 9 b). Such enhanced performance was attributed to the ability of CB[6] to bind small molecules, such as CO, which decreased the bond energy between small molecules and Pt sites and, finally, promoted the oxidation of MeOH and CO. Furthermore, the unique ability of CB[6] can even promote the poisoning tolerance of commercial Pt/C through simple physical mixing, as shown in Figure 9 c.

Figure 9.

a) Representation of a CB[6] molecule with two identical negatively charged portals and positively charges outside walls. b) CVs of the three kinds of CB[B]-Pt NPs and commercial Pt/C for the MOR. c) CVs of commercial Pt/C mixed with CB[6] (straight line) and without CB[6] (dashed line; Reproduced from Ref. 90). If and Ib are the corresponding peak current values of the forward and backward scans, respectively.

3.3. Transition-metal oxides

Improved catalytic properties and stabilities can be achieved by supporting metal NPs on conductive or semiconductive transition-metal oxides, such as CeO2, TiOx, WOx, and NbO2, which strongly interact with late transition metals.62a, 91 Compared to a common carbon support, lower surface area but greater stability is observed in those alternative support materials in an electrochemical environment, thereby conferring enhanced stability on the Pt NPs. For example, Adzic′s group92 demonstrated that the use of NbO2 as a support caused the Pt NPs to exhibit much greater stability against Pt dissolution after 30 000 cycles for the ORR.

Thus, the tactics of doping, hybridization, and nanometer-sized morphology have been widely adopted for improving the electric conductivity and available surface area of non-carbon supports. As such, TiO2 has received increasing attention because of its inherent stability, mechanical resistance, beneficial interactions with metal catalysts, and tunable morphologies. Popov and co-workers67 developed mesoporous TiO2 supports for Pt NPs, which exhibited excellent fuel-cell performance as well as ultrahigh stability, even at high positive potentials of 1.2 V. Figure 10 a, b shows polarization curves for the Pt/TiO2 and Pt/C electrocatalysts after the potential was held at 1.2 V for 0–200 and 0–80 h, respectively. The polarization curves of Pt/TiO2 were similar, even after a corrosion time of 200 h, whereas the Pt/C electrocatalyst showed a significant decrease in performance after 50 h, owing to carbon corrosion and subsequent detachment and agglomeration of the catalyst particles. Figure 10 c, d shows the total loss of ECSA and potential loss as a function of time, which further confirmed the excellent stability of Pt/TiO2. The ultrahigh stability of the Pt/TiO2 electrocatalyst was attributed to strong metal–support interactions between the Pt particles and the TiO2 support, which effectively inhibited Pt migration and agglomeration.

Figure 10.

Polarization curves for the ORR with a) Pt/TiO2 and b) Pt/C electrocatalysts after 0–200 and 0–80 h, respectively. c, d) Plots of normalized ECSA (c) and potential loss (d) at 0.5 A cm−2 as a function of corrosion time (Tc) for Pt/TiO2 and Pt/C electrocatalysts (Reproduced with permission from Ref. 67. Copyright 2009, American Chemical Society).

Composite metal oxides, such as BaTiO3,94 Na2Ti3O7,95 Nb0.07Ti0.93O2,91f and Sn-doped indium tin oxide (ITO),96 are also good supports for Pt NPs. Liu and Mustain96 prepared highly stable and conductive ITO supports for Pt ORR catalysts. The Pt/ITO catalysts showed very impressive stabilities under very harsh conditions: The ECSA was unchanged and the Pt half-wave potential was only shifted by 4 mV (to 1.4 V versus NHE) over 1000 cycles. Recently, Wang et al.97 reported the unique CO-tolerant properties of Pt NPs that were supported on conducting 50 nm Ti0.7W0.3O2. Owing to the W-doping, the conductivity of TiO2 was increased. The loss in the integrated columbic charge for Pt/Ti0.7W0.3O2 was only 5 % after 500 cycles, whereas it was over 30 % for a commercial E-TEK PtRu/C catalyst. Hwang′s group93 explored the use of robust non-carbon Ti0.7Mo0.3O2 as a functionalized co-catalytic support for Pt. Ti0.7Mo0.3O2 featured high surface area (232 m2 g−1) and electronic conductivity (2.8×10−4 S cm−1), as well as high stability in acidic solution and oxidative environments compared to TiO2, MnO2, and carbon black. In addition to high activity (Figure 11 c), as expected, Pt/Ti0.7Mo0.3O2 demonstrated extremely high stability, with no significant drop in activity after 5000 cycles and only about 8 % performance degradation at 0.9 V (Figure 11 d). Such enhanced stability and activity are shown in Figure 11 a, b: Electron transfer from Ti0.7Mo0.3O2 to Pt can modify the surface electronic structure of Pt, thereby resulting a shift in the d-band center of the surface Pt atoms and strong metal–support interactions.

Figure 11.

a) Pt LIII-edge XANES spectra and b) variation in unfilled d-states for Pt foil and various catalyst samples. Inset in (a) shows magnified regions of the peaks for the Pt LIII-edge XANES white line. c) Polarization curves, which show the ORR currents of a Pt/Ti0.7Mo0.3O2 catalyst and two commercial Pt/C (E-TEK) and PtCo/C (E-TEK) catalysts. d) Stability tests of the three catalysts before and after 5000 cycles at 0.9 V (Reproduced with permission from Ref. 93. Copyright 2011, American Chemical Society).

3.4. Tranistion-metal nitrides and carbides

Early-transition-metal nitrides and carbides have received considerable attention because of their noble-metal-like catalytic properties.98 Most transition-metal nitrides and carbides have high melting points and are good electrical conductors.99 For instance, the reported conductivity of molybdenum nitride (Mo2N) is over 7×103 s cm−1. In particular, some nitrides and carbides are active for some hydrogen-involving reactions, which is vital for the electro-oxidation of MeOH and formic acid, as well as the ORR.98b

For example, TiN-supported Pt catalysts have been reported to show enhanced catalytic activity for the ORR, owing to the superior dispersion of Pt NPs over a TiN NPs support, enhanced interactions between the metal NPs and the support, good conductivity, and resistance to oxidation and acid-corrosion.100 Cui et al. recently demonstrated a new and efficient method for the preparation of Mo2N nanoparticles (2–5 nm) that were loaded onto carbon black,98a which were used as support materials for Pt NPs for the oxidation of small organic molecules in fuel cells. The electrocatalytic curves demonstrated that the hybrid Pt[BOND]Mo2N/C exhibited much higher catalytic activities and durabilities than Pt/C catalysts for both the MOR and the FAOR. Quite recently, DiSalvo′s group developed mesoporous CrN with pore sizes ranging from 10–20 nm as a conductive support for Pt NPs.99 The Pt/CrN catalysts demonstrated high tolerance to corrosion and were a candidate for replacing carbon black.

Since their Pt-like behavior was first observed by Levy and Bounart in 1973, WxC have become the most widely studied transition-metal carbides in the field of energy conversion.101 Owing to electron transfer between the catalyst and the support,5a synergistic effects,102 and improved dispersion,103 WxC-loaded Pt catalysts exhibited enhanced electrocatalytic performance, such as high activity or immunity to MeOH. However, Shao and co-workers104 have argued that one monolayer of Pt on WC does not aid the ORR activity, because of the formation of WOx species in the presence of excess oxygen and high oxidizing potentials during the ORR. If the WC surface is oxidized, it can cause the Pt to either detach or agglomerate. Therefore, the choice of pH value (alkaline environment) and potential window (below 0.8 V) in which WC is electrochemically stable is of great importance for WC-containing catalysts. Recently, Lin′s group prepared Pt-doped WxC catalysts by a simple co-impregnation and thermal-reduction method.105 Unlike previous studies that used WxC as the support for Pt deposition, in this study, Pt and WxC played the role of structure promoter by preventing the agglomeration and oxidation of each other during the synthesis and electrocatalytic process. The Pt loading could be decreased to as low as 5 wt %, yet the catalysts still exhibited comparable activities and better durabilities than commercial available 20 wt % Pt/C, even if the upper potential was set at 1.2 V.

Other transition-metal carbides, including TiC,106 NbC,107 VC,108 and Mo2C,109 have also been used as Pt supports. For example, NbC nanowires have recently been synthesized by using a bamboo-based carbothermal method and used as Pt supports for direct methanol fuel cells (DMFCs).107 1D NbC NWs showed highly oriented growth behavior, high electrical conductivity, and outstanding oxidation resistance, thereby leading increased activity and excellent stability towards the MOR.

3.5. Hierarchical composite supports: Hybridization materials

Hybridization provides an effective strategy for enhancing the functionality of a material, owing to complementarity between their structures and physicochemical properties. Compared with pure matrix materials, the properties of hierarchical composites are enhanced over a wide range to address the increasing demands of stability and durability. In terms of supports for Pt NPs for fuel-cells reactions, there are several possible combinations that can be classified as follows: carbon-black–graphene/CNTs,62d, 70 carbon-materials–conductive-polymers,110 carbon-materials–transition-metal-oxides/nitrides/carbides,111 and more complex multicomponent composites.

An alternative approach to using the high conductivity/surface area of carbon and the high stability of metal oxides is to develop composites of the two materials. Kou et al.112 reported a new method to directly grow ITO on functionalized graphene sheets to form ITO/graphene hybrids and studied their application as supports for Pt electrocatalysts. The Pt NPs were stabilized at ITO–graphene junctions, thereby forming a special Pt-ITO-graphene triple-junction structure (Figure 12 a–c), which exhibited greatly enhanced performance, especially in terms of durability. DFT calculations suggested that the defects and functional groups on graphene played an important role in stabilizing the catalyst. Moreover, the ITO coating made the graphene substrates very resistant towards oxidation. Such triple-junction structures have also been reported with different compositions, such as Pt-TiO2-carbon (Figure 12 d–f),63a Pt-TiO2-RGO (Figure 12 g–i),111b and Pt-CeO2-carbon.91b The stability of carbon materials and the poor electronic conductivity of metal oxides are simultaneously robust in such an array of structures, thereby leading enhanced catalytic performance.

Figure 12.

Geometric model and TEM images of a–c) Pt-ITO-graphene (Adapted with permission from Ref. 112. Copyright 2011, American Chemical Society); d–f) Pt-TiO2-carbon (Reproduced with permission from Ref. 63a. Copyright 2011, The Royal Society of Chemistry); and g–i) Pt-TiO2-RGO (Reproduced with permission from Ref. 111b. Copyright 2012, PCCP Owner Societies); and j–l) Pt/C@PANI, (k) and (l) show PANI shells with thicknesses of 5 and 14 nm, respectively (Adapted with permission from Ref. 113. Copyright 2012, American Chemical Society).

Recently, Ma et al.111a synthesized nanocomposites that consisted of bimetallic carbide Co6Mo6C2 supported on graphitic carbon (gC) by using an in situ anion-exchange method. The Co6Mo6C2/gC nanocomposite was not only chemically stable but also electrochemical stable. Therefore, the catalysts that were prepared by loading Pt NPs onto Co6Mo6C2/gC showed superior stability and activity for the ORR in acid solution: Only a small (6 mV) degradation in the half-wave potential and no recordable loss of ECSA after 1000 cycles within the potential range 0.05–1.1 V were observed.

Chen et al.113 designed and synthesized a PANI-decorated Pt/C@PANI core–shell catalyst through the direct polymerization of a thin layer of PANI on the carbon surface of the Pt/C catalyst (Figure 12 j–l). In previous reports on Pt/PANI/C composites, most of the Pt NPs were wrapped in PANI and could not be utilized in the fuel-cell reactions. However, in this case, the PANI shell layer in the Pt/C@PANI core–shell catalyst preferentially and selectively covered the surface of the carbon support, rather than the Pt species. The as-made catalysts with a thickness of the PANI shell of 5 nm had the best ORR activity, which was much higher than that of non-decorated Pt/C. In particular, the Pt/C@PANI catalysts showed super stability and long-term durability, even in a single cell by using an accelerated stress test. The improved activity and stability were attributed to the PANI-coated core–shell structure, which resulted in electron delocalized between the Pt d orbitals and the PANI. In addition, the stable PANI shell protected the carbon support from direct exposure to the corrosive environment. Similar concepts have also been applied to other systems with different compositions, such as Pt/PPy-carbon black,114 Pt/PANI-CeO2 nanofiber membranes,115 and PANI/Pt/MWCNTs multilayer composites.116

4. Summary and Outlook

15 Years of intense development of PEMFCs have brought this technology close to pre-commercial viability.2e However, for its genuine commercialization, a few remaining barriers still need to be overcome: lower costs and established stability. As a key component of PEMFCs, Pt-based catalysts face an urgent need to enhance their long-term performance and stability. This review has provided a survey of the recent developments in the stabilization of Pt-based catalysts in reactions related to fuel cells and we have tried to present an up-to-date overview of such a rapidly growing field.

The design strategies for the construction of stable and lower-cost Pt-based catalysts can be classified into the following categories: 1) The construction of highly anisotropic 1D, 2D, and 3D Pt-based nanomaterials as carbon-free and self-supported catalysts; 2) alloying a second metal with Pt to modify the surface structure and d-band center; 3) enlarging the interface in the heterostructure; 4) making the shell rich in Pt and the core a transition metal/ordered intermetallic alloy; 5) coating the Pt NPs with stable/anti-corrosion silica or the carbon shell with controllable channels; and 6) using alternative materials other than carbon black as the catalyst supports.

In the foreseeable future, the area of Pt-based nanocrystals will continue to expand in terms of new methods; new materials; control of size, shape, structure, and composition; and the understanding of structure–property relationships. Moreover, fundamental theory and practical applications still have huge space for development. How to efficiently and cost-effectively use the catalytic properties that are associated with Pt and other metals, as well as supports, to meet the demands of future industrial applications is also not clear at present. For this purpose, efforts should be directed towards obtaining a better fundamental understanding of Pt-based nanocrystals and the reaction mechanisms through theoretical methods, such as DFT calculations. In addition, the development of characterization methods for the microstructures between Pt and metals/supports is also important for rationally designing and utilizing more active and durable catalysts for future PEMFCs.


We acknowledge financial support from the 973 Program (2011CB932504 and 2012CB821705), the NSFC (21221001, 21203199, and 21331006), the Fujian Key Laboratory of Nanomaterials (2006L2005), the Key Project from the CAS, the K. C. Wong Education Foundation (Hong Kong), and the China Postdoctoral Science Foundation (2012M521280 and 2013T60642).

Biographical Information

Rong Cao received his BS from the University of Science and Technology of China (USTC, Hefei, China) in 1986 and his MS and PhD from the Fujian Institute of Research on the Structure of Matter (FJIRSM) from 1986–1993. After a two-year stint as a Postdoctoral Fellow at Hong Kong Polytechnic University, he started as a Professor at the FJIRSM in 1998. From 2000–2002, he was a JSPS Follow at Nagoya University. He was promoted to the Deputy Director of the FJIRSM in 2008. His current research interests include the synthesis and assembly of nanomaterials, supramolecular chemistry, inorganic–organic hybrid materials, and their applications in catalysis and energy storage.

original image

Biographical Information

Minna Cao received her PhD in inorganic chemistry from the University of Science and Technology of China (USTC, Hefei, China) in 2011. She is currently a Postdoctoral Fellow in Prof. Rong Cao′s group at the FJIRSM. Her current research interests are the synthesis and assembly of nanostructure for applications in fuel cells and organic catalysis.

original image

Biographical Information

Dongshuang Wu received her BS in material physics and chemistry from Heilongjiang University in 2009. She is pursuing her PhD in physical chemistry with Prof. Cao at the FJIRSM. Her research focuses on the synthesis of nanomaterials for energy applications.

original image