Recent progress in the development of anode and cathode catalysts for direct methanol fuel cells

(1)

journal homepage: www.elsevier.com/locate/nanoenergy Available online at www.sciencedirect.com

REVIEW

Recent progress in the development of anode

and cathode catalysts for direct methanol

fuel cells

Jitendra N. Tiwari

a,n

, Rajanish N. Tiwari

b

, Gyan Singh

c

,

Kwang S. Kim

a

a_{Center for Superfunctional Materials, Dept. of Chemistry, Pohang University of Science and Technology,}

San 31, Hyojadong, Namgu, Pohang 790-784, Korea

b

Surface Science Laboratory, Toyota Technological Institute, 2-12-1 Hisakata Tempaku, Nagoya 468-8511, Japan

c

Department of Biological Science and Technology, National Chiao Tung University, 1001 Ta Hsueh Road, Hsinchu, Taiwan, 300, R.O.C.

Received 14 March 2013; received in revised form 18 June 2013; accepted 18 June 2013 Available online 3 July 2013

KEYWORDS Methanol oxidation; Oxygen reduction; Anode catalysts; Cathode catalysts; Direct methanol fuel cell

Abstract

Continuous growth in global energy demand has sparked concerns about energy security and environmental sustainability. In the past two decades, attempts have been made in the development of innovative energy technologies. The direct methanol fuel cell (DMFC) is among the most promising alternative energy sources for the near future. Simple construction, compact design, high energy density and relatively high energy-conversion efﬁciency give the DMFC an advantage over other promising power sources in terms of portability. However, the translation of DMFCs into commercially successful products is precluded due to poor performance. In addition, low activity, poor durability and reliability and an expensive anode and cathode further discourage the application of DMFCs. In this regard, the present review article focuses on recent progress in the development of anode and cathode catalysts for DMFCs. The ﬁrst part of the review discusses the recent developments in the synthesis of single-, double-, and multiple-component catalysts and new catalyst supports for anode electrodes. The section is followed by the chemical approaches employed to make alloys and composite catalysts, aiming to enhance their activity, reliability and durability for the methanol oxidation reaction. Finally, exciting new research that pushes the development of single-, double-, and multiple-component catalysts and new catalyst supports for cathode electrodes is introduced. In addition, size-, shape- and composition-dependent electrocatalysts that are advocated for methanol

n_{Corresponding author. Tel.:}_{+82 54 279 2110;}

fax:+82 54 279 8137.

(2)

oxidation at the anode and oxygen reduction at the cathode are highlighted to illustrate the potential of the newly developed electrocatalysts for DMFC applications. Moreover, this article provides a comprehensive review of the experimental work that is devoted to understanding the fundamental problems and recent progress in the development of anode and cathode catalysts for DMFCs.

Introduction . . . 554

Overview . . . 554

The scope of this review . . . 555

Direct methanol fuel cell (DMFC). . . 555

Principle of DMFC operation . . . 555

Challenges . . . 555

Reaction mechanism for methanol oxidation. . . 556

Reaction mechanism for oxygen reduction . . . 556

Anode catalysts of DMFCs . . . 556

Single-component catalysts. . . 556

Double-component catalysts . . . 558

Multiple-component catalysts . . . 561

New catalyst supports . . . 563

Cathode catalysts of DMFCs . . . 565

Single-component catalysts. . . 565

Double-component catalysts . . . 568

Multiple-Component Catalysts . . . 571

New Catalyst Supports. . . 572

Conclusions and future outlook . . . 574

Acknowledgments . . . 575

References . . . 575

Introduction

Overview

Direct methanol fuel cells (DMFCs) have attracted consider-able recent interest because of various advantages, includ-ing high power density, zero or low exhaust, ease of recharging, simple structure, and quick startup at low temperature[1–3]. However, one of the major drawbacks of the DMFC is its high manufacturing cost, which prevents their successful commercialization [3]. By initiating cost-effective steps from prototypes to mass production, manu-facturing expenses can be reduced; however, the expensive materials required for the production of DMFCs remain a challenge because DMFCs operating at low temperature use platinum (Pt) and its alloys for the conversion of fuel at the anode and reduction of oxygen at the cathode. The high cost of Pt significantly increases the total price of the fuel cell devices. The problem can be solved in the short term by using single-, double-, and multiple-component catalysts and new catalyst supports. Therefore, the study of single-, double-, and multiple-component catalysts and new cata-lyst supports not only has emerged as a significant area of research and development but also has greatly influenced everyday life, as more products based on single-, double-, and multiple-component catalysts and new catalyst sup-ports are increasingly introduced to the market. Extensive research in the past two decades has established that

the physical and chemical properties of materials show signiﬁcant changes if their dimensions are on the nanometer scale, which opens new avenues for a wide range of future applications[3,4]. Especially, physical properties of nanos-tructures, such as large surface area and novel size effects, markedly improve the efﬁciency of DMFCs[3,4]. Therefore, single-, double-, and multiple-component catalysts and new catalyst supports have become increasingly important in the development of DMFCs in recent years. In addition, there are two main types of effects that result from single-, double-, and multiple-component catalysts and new cata-lyst supports: (1)‘trivial size effects’, which rely solely on the increased surface-to-volume ratio and decreased layer thickness and volume of the nanoparticles and (2)‘true size effects’, which also involve changes in local material properties [5,6]. Therefore, nanoscale engineering of the material appears to be critical in the next stage of advancement of DMFCs. Development in recent years has shown that single-, double-, and multiple-component catalysts and new catalyst supports have great potential for innovative new technology for the recurring energy demand.

In this regard, the review article is mainly focused on recent developments in the ﬁeld of DMFC anode and cathode catalysts; the role of single-, double-, and multiple-component catalysts and new catalyst supports for non-alloy and alloy nanoparticles is particularly dis-cussed in more detail. In addition, challenges involved in the development of DMFCs and the reaction mechanism for

(3)

methanol oxidation and oxygen reduction are subsequently discussed in more detail.

The scope of this review

The review article briefly describes the principle of DMFC operation. Subsequently, challenges involved in DMFC anode and cathode electrocatalysts are described. In Section 3, a brief summary of the reaction mechanism for methanol oxidation is reported. Section 4 describes the reaction mechanism for oxygen reduction.Sections 5 and6address recent progress in the development of anode and cathode electrocatalysts for DMFCs. Sections 5 and 6 particularly focus on the role of single-, double-, and multiple-component catalysts and new catalyst supports with non-alloy and non-alloy nanoparticles in the development of highly efficient DMFCs. The final section will describe the conclu-sions. Some open problems and continuing challenges are also highlighted in thefinal section.

Direct methanol fuel cell (DMFC)

Principle of DMFC operation

The contemporary science and technology of DMFCs have already been extensively reported in many review papers and articles; interested readers can refer to references

[7–9] for more details. In short, the main active components of a DMFC are the fuel electrode (anode), oxidant electrode (cathode), and an electrolyte sandwiched between them.

Figure 1shows the basic operational principle of a fuel cell with its reactant or product gases and ion conduction ﬂow directions. A schematic drawing of the DMFC system is given inFigure 1, displaying the principle of DMFC operation. The DMFC converts chemical energy into electrical energy by oxidizing methanol to CO2 and H2O. A proton-conducting

solid membrane, used both as an electrolyte and separator between anode and cathode, is sandwiched between porous structures (such as carbon). The latter serve both as current collectors and as a support for catalyst particles. Before catalyst deposition, the current collectors are impregnated with polymer electrolyte to provide intimate contact of the metal particles with both electron and proton conductors.

At the anode, a methanol molecule reacts with an H2O

molecule and liberates CO2, six protons that are free to

migrate through the electrolyte towards the cathode and six electrons that can travel through the external load. The CO2

produced in the reaction is rejected by the acid electrolyte solution. The protons migrating through the electrolyte and electrons moving via the external loaded circuit must reach a particle of catalyst on the cathode, where O2is

electro-catalytically reduced and H2O is produced. An electric

potential occurs between the electrodes because of the excess of electrons at the anode (where they are generated) compared with the cathode (where they are consumed). This potential difference drives current through the exter-nal load, making the fuel cell a real source of power. The maximum theoretical voltage attainable from the overall reaction in the methanol–air fuel cell is ∼1.21 V with a theoretical efﬁciency of 96.5%, but in practice, this voltage is not obtained due to poor electrode kinetics and ohmic losses in the electrolyte[10,11]. The relevant electroche-mical reactions at the electrode surface in acidic media are

[12]as follows:

Anode reaction: CH3OH+H2O-CO2+6H++6e ,

E0=0.016 V/SHE (1)

Cathode reaction: 3/2 O2+6H+_{+6e -3H2O,}

E0=1.229 V/SHE (2)

Overall reaction: CH3OH+3/2O2-CO2+2H2O, E0=1.21 V (3)

for which the standard hydrogen electrode (SHE) is used as a reference electrode.

Challenges

There are some crucial obstacles to overcome before large-scale commercialization of DMFC is achievable[13–15]: (a) the greatest present concern is the low activity and high cost of electrocatalysts in the anode and cathode; (b) the anode reaction has poor kinetics at lower temperatures, making identiﬁcation of improved electrocatalysts and higher working temperatures highly desirable; (c) at the cathode, the oxygen reduction reaction is also slow: this problem is particularly serious with aqueous mineral acids, although studies show the issue is not quite as serious with acidic polymer membranes (Methanol vapor also appears in the cathode exhaust, from which it must be removed.); (d) the high cost of the currently used Naﬁon membrane (price=US$ 800–2000/m2

); (e) and the permeability of the current perﬂuorosulfonic acid membranes (Naﬁon, Figure 2) to methanol, which allows considerable crossover of methanol from the anode region to the cathode region. This crossover causes degradation of the performance of both electrodes, as mixed potential develops at the cathode, and deterioration of fuel utilization.

Figure 1 Schematic diagram of a DMFC. PEM=polymer elec-trolyte membrane.

Figure 2 Chemical structure of Naﬁon, where n= 6.5 13.5; m=1, 2, 3…; x=200 1000; and M+is the exchangeable cation.

(4)

Reaction mechanism for methanol oxidation

Over the past several decades, materials scientists have sought to improve their knowledge of methanol oxidation mechanisms at different electrocatalysts under perfectly well-controlled conditions, such as different single crystal orientations and foreign metal clusters on polycrystalline or single crystal surfaces[10,16–18]. The basic mechanism for the methanol oxidation reaction (MOR) wasﬁrst reviewed in 1988[19]. Based on previous reports, the reaction can be summarized as follows[16–20]:

(1) Electrosorption of methanol.

(2) C–H bond activation (methanol dissociation). (3) H2O adsorption and activation.

(4) Addition of oxygen to adsorbed carbon-containing inter-mediates to generate CO2, which can be facilitated by

addition of a second metal in alloy systems.

The catalytic oxidation mechanism of methanol by Pt is schematically described inFigure 3. As shown in Figure 3, during the MOR, COadsis formed and strongly adsorbs onto

the Pt electrocatalyst, reducing the surface area and thus the performance of the DMFC. Three reaction paths for methanol oxidation to CO2 have been investigated to

date. One is an indirect mechanism that involves a CO intermediate (path 1, Figure 3), and the other two are direct mechanisms in which methanol is oxidized to CO2

without the formation of a CO intermediate (paths 2 and 3,

Figure 3).

Reaction mechanism for oxygen reduction

The oxygen reduction reaction (ORR) is another important reaction in energy-converting systems such as DMFCs. How-ever, the mechanism of the electrochemical ORR is very complicated and involves many intermediates. In addition, the ORR is highly dependent on the nature of the electrode material, catalyst, and electrolyte solution. The electro-chemical reduction of O2 in solution occurs by two main

pathways: one involving the gain of two electrons to produce H2O2and another producing H2O by a direct

four-electron pathway. The 2e and 4e reduction pathways in both acidic and alkaline media are given inFigure 4 [21–23]. To achieve maximum energy capacity, reduction of O2must

occur via the 4e pathway. The four-electron reduction pathway is thus used in fuel cell systems; however, the two-electron reduction pathway is used in H2O2production.

Anode catalysts of DMFCs

The success of DMFC technology depends on several factors, such as membrane, anode and cathode electrocatalysts. Among these, the anode electrocatalyst suffers from slow reaction kinetics that can only be overcome through devel-oping new electrocatalyst types. With regard to new fuel cell anode electrocatalysts, there are two major concerns: performance, including activity, reliability and durability, and cost reduction. In the following section, recent progress in the development of DMFC anodes based on single-, double- and multiple-component catalysts is discussed in detail. In addi-tion, a comprehensive overview of recently developed support materials for DMFC applications is covered.

Single-component catalysts

As a common catalyst for methanol oxidation, noble metals (such as Pt, Pd, and Au) can contribute signiﬁcantly to DMFCs [24–27]. Therefore, a large variety of noble metal catalysts have been tested in DMFCs[24–27]. Reduced cost and improved methanol oxidation activity are especially desirable. The methanol oxidation activity of noble metal catalysts is strongly dependent on particle shape, size and surface structure [25]. Therefore, to increase methanol oxidation activity and reduce metal loading, many research-ers have recently synthesized noble metal catalysts with different sizes, shapes and surface structures[25–27]. More recently, Lu et al. [26]synthesized ultrathin Au nanowires with high density stacking faults (HDSF) and studied their methanol oxidation properties in acidic and alkaline media.

Figure 4 Schematic representation of the oxygen reduction reaction[21]. Figure 3 Schematic representation of the methanol oxidation

(5)

The ultrathin Au nanowires with an HDSF structure were synthesized from an Au precursor using oleylamine at 801C; in this reaction, oleylamine served as the solvent, reducing agent and surface capping agent. The Au nanowires exhibit higher electrocatalytic activity for methanol than do poly-Au nanoparticles and bulk Au in both acidic and alkaline media.

Generally, the MOR on an Au electrode proceeds in two distinct potential regions, with different mechanisms invol-ving the two equations shown below:[28–30]

CH3OH+5OH -HCOO +4H2O+4e (4)

CH3OH+8OH -CO32 +6H2O+6e (5)

At lower potentials, methanol was mainly oxidized to formates via an overall four-electron transfer reaction (Eq. 4). At higher potentials, the methanol molecules were oxidized to carbonates with the exchange of a six-electron transfer reaction (Eq. 5).

In 2009, Li et al.[31]fabricated Au nanoprism thinfilms on indium tin oxide (ITO)-coated glass substrates via wet chemical methods. The synthesized Au nanoprism thinfilms are single crystalline, whose basal and lateral surfaces are atomically flat {111} and {110} planes, respectively. The electrochemical catalytic activity of Au nanoprism thinfilms (Au surface modification with Pt) towards methanol oxida-tion was found to be much higher than that of commercial Pt-based catalysts due to the synergistic effect. Thus, even Au can exhibit high catalytic activity, but that performance is not sufficiently high for use in DMFCs. Therefore, researchers are looking for another metal. To overcome this problem, Wang et al. [32] synthesized nanoporous palladium rods and studied their electrochemical catalytic activity behavior in a methanol–KOH solution. The nanopor-ous palladium rods were fabricated through the chemical dealloying of rapidly solidified Al80Pd20alloy in a 5 wt% HCl

aqueous solution under free corrosion conditions. Individual nanoporous palladium rods were several microns in length and several hundred nanometers in diameter. In addition, all nanoporous palladium rods had a three-dimensional (3D) bicontinuous interpenetrating ligament-channel structure with a length scale of 15–20 nm. Moreover, the nanoporous palladium rods with high specific surface areas revealed superior electrocatalytic performance (223.52 mA mg 1) toward methanol oxidation in alkaline media [32]. Recent investigations showed that Pt nanoparticles are more pro-mising electrocatalysts than Pd and Au nanoparticles for the DMFC anode[33-42]. Nevertheless, one of the most impor-tant goals in nanoparticle-based electrochemistry is to understand how electrocatalytic activity is influenced by surface structure and nanoparticle shape [35]. In this regard, wefirst discuss the reported electrocatalytic prop-erties of shape-controlled nanoparticles. Han et al. [36]

synthesized Pt nanocube catalysts approximately 3.6 nm in size by a polyol process in the presence of polyvinylpyrro-lidone (PVP) as a stabilizer and Fe3+ ions as a kinetic controller. The Pt nanocubes were single-crystalline with exposed (100) planes. These Pt nanocube catalysts showed a lower onset potential and higher current density for metha-nol oxidation than polycrystalline Pt. In addition, Han et al. found that the edge of the stepped (100) planes in the Pt nanocube catalyst was more preferable for the easy

breaking of the methanol C–C bond than polycrystalline Pt. Recently, Lee et al.[37]reported Pt nanocubes∼4.5 nm in size that were synthesized by a modiﬁed polyol method with the assistance of a thermal reduction process. These Pt nanocubes with well-deﬁned (100) planes provided much higher electrocatalytic activity toward the MOR than the spherical Pt/C with a polycrystalline structure. The excel-lent electrocatalytic activity of Pt nanocubes for the MOR may be due to predominantly exposed (100) planes. More recently, our group also reported the synthesis of shape-controlled perfect Pt nanoparticles via a pair of low-resistivity fastened silicon (FS) wafers at room temperature

[38]. We also found that perfect Pt nanocubes deposited on FS wafers exhibited much higher electrocatalytic activity and stability for methanol oxidation than truncated Pt nanocubes, truncated Pt (cubes-tetrahedra) or spherical Pt nanoparticles. We reported that the high electrocatalytic activity of perfect Pt nanocubes is mainly due to the exposed Pt (100) planes of the single crystals. Liang et al.

[41]synthesized Pt hollow nanospheres on a large scale via a replacement reaction between Co nanoparticles and H2PtCl6. From electrocatalytic calculations, the

electroca-talytic activity of the Pt hollow nanospheres was twice that of the solid Pt nanoclusters. The high methanol oxidation current of the hollow-sphere catalysts was directly related to their high surface area. In addition, the incomplete shell of a hollow nanosphere may provide an interior surface for the electrocatalytic reaction. Liang et al. reported that the inner and outer surfaces of a hollow nanosphere can participate in the electrocatalytic reaction. Although Pt nanocubes and Pt hollow nanospheres have high electro-catalytic activity and stability, they suffer from a smaller electroactive surface area (ESA). Therefore, our group fabricated a two-dimensional (2D) continuous Pt island network by electrochemical deposition on a ﬂat silicon substrate [40]. The ESA of this 2D continuous Pt island network was 67 m2_g 1_{, which is much higher than that of a}

Ru-decorated Ptﬁlm electrode (21 m2g 1) and a blanket Pt electrode (16 m2_g 1_{). The 2D continuous Pt island network}

exhibited higher CO and methanol oxidation activities than did the Ru-decorated Ptﬁlm and blanket Pt electrodes. The improvement of the CO and methanol oxidation activities can be attributed to the synergistic effect of the Pt island catalyst and the SiO2 surface layer. In an endeavor to

produce 3D nanostructured materials for DMFCs [33], our group synthesized 3D Pt nanoflowers that had a higher surface area than that of the Pt thin film/Si electrode. Moreover, compared to blanket Pt thinfilms, 3D Pt nano-flowers/Si had much higher electrocatalytic activity, with a steady-state current density approximately 310 times higher than that of the thin film Pt/Si electrode. The electro-chemical characteristic curves of 3D Pt nanoflowers and blanket Pt thin films were obtained in 1 M H2SO4–1 M

methanol (Figure 5) [33]. The anodic peak in the reverse scan can be attributed to the removal of incompletely oxidized carbonaceous species (CO-like species) formed in the forward scan. Thus, the ratio of the forward anodic peak current density (If) to the backward anodic peak current

density (Ib), If/Ib, can be used to describe the CO tolerance

of a catalyst to accumulation of carbonaceous species

[33,40]. A low If/Ib indicates poor oxidation of methanol

(6)

accumulation of carbonaceous residues on the catalyst surface. However, high If/Ib indicates excellent oxidation

of methanol during the forward anodic scan and less accumulation of residue on the catalyst. As shown in

Figure 5, the If/Ibratio of 3D Pt nanoﬂowers (∼2.5) is higher

than that of Pt thinﬁlm (∼0.93), indicating less accumula-tion of CO-like species on the catalyst during the MOR and thus excellent catalytic activity.

Other groups also reported the effectiveness of 3D nanostructured materials on electrocatalytic properties

[34,42]. More recently, Rauber et al.[42]synthesized highly ordered 3D Pt nanowire networks by a method based on hard templates using electrodeposition within nanochannels of ion track-etched polymer membranes. The network structures consisted of well-deﬁned interconnected nano-wires with controlled morphologies and compositions. The electrocatalytic activity of the 3D Pt nanowire networks in the MOR was measured in H2SO4 (0.5 M) and methanol

(0.5 M) at room temperature. The peak current densities of the anodic peaks of 3D Pt nanowire networks, PtB, and Pt/C catalysts were 0.76, 0.39, and 0.24 mA/cm2, respec-tively, showing the high electrocatalytic activity of 3D Pt nanowire networks. The excellent electrocatalytic activity is a direct consequence of the inherent properties of the continuously organized, porous 3D architecture, resulting in excellent transport properties and efﬁcient access of the reactants to catalytic sites. Anode poisoning by the inter-mediates or CO-like species formed during the MOR on Pt catalysts is a familiar phenomenon. In addition, due to the high cost, commercial applications of noble Pt metal catalysts have been restricted. However, the modiﬁcation of Pt catalysts with other metals or metal oxides that have higher tendencies to form surface oxygenated species at lower potentials is one of the best methods to solve this problem[43,44].

Double-component catalysts

Another major problem for the efﬁcient conversion of methanol fuel to electric current in a DMFC is the sluggish MOR kinetics on the anode catalyst. This slow nature is mostly due to self-poisoning of the surface by reaction intermediates such as COads-like species that are generated

during the stepwise dehydrogenation of methanol [45-48]. Therefore, the MOR on Pt is possible only at potentials where adsorbed COads-like species and other poisoning

intermediates are effectively oxidized, leading to signiﬁ-cant overpotential and loss in DMFC efﬁciency [49]. This problem necessitates the search for binary Pt-based alloy (double-component) catalysts, such as PtRu [50,51], PtSn

[52,53], PdNi[54], PtMo[55,56], PtTiO2 [57,58], PtW[55],

PtOs[60], and PtMn[61]. These alloyed metals can provide OH species at more negative electrode potentials to react with COads-like poisoning species, thus improving the

electro-oxidation activities of methanol. Zhang et al. [62]

deposited Pt nanoparticles on a nanoporous Au substrate by a simple immersion-electrodeposition technique, forming porous nanostructured Pt–Au catalysts. These porous Pt–Au catalysts have better electrochemical activity, antipoisoning ability, and long-term structural stability than commercial Pt–Ru catalysts, which can be justiﬁed by the bifunctional mechanism of bimetallic catalysts. According to the bifunctional mechan-ism, the oxidation of surface COadsspecies at the bimetallic

porous Pt–Au catalysts would proceed through one of the following three routes:

Pt–COad+Au–OHad-CO2+H++Pt+Au+e (6)

Au–COad+Au–OHad-CO2+H++Au+e (7)

Pt–COad+Pt–OHad-CO2+H++Pt+e (8)

Wang et al. [63] investigated the MOR activity on an Au-modiﬁed Pt (Au/Pt) electrode together with phosphomolyb-dic acid. In the typical synthesis of the Au/Pt electrode, the ﬁrst step required forming an underpotential deposition (upd) Cu adlayer on the Pt substrate; the upd-Cu adlayer was then replaced by Au from HAuCl4aqueous solution. This

replacement process takes place by the following reaction: 3upd-Cu/Pt+2AuCl4=2Au/Pt+3Cu2++8Cl+ (9)

Further, the performance of Au-modiﬁed Pt as a catalyst for the MOR was investigated by Wang et al. [62]. Analysis of data indicated that the MOR on Au-modiﬁed Pt was mark-edly enhanced not only by phosphomolybdic acid solution but also by adatom Au. The authors reported that adsorbed hydrogen and intermediate CO from methanol dehydrogena-tion and oxidadehydrogena-tion were electrocatalytically oxidized by the oxidant state of phosphomolybdic acid in the presence of an Au catalyst. To improve activity and stability, Zhang et al.

[64] fabricated a novel nanoporous bimetallic Pt–Au alloy

nanocomposite by dealloying the rapidly solidiﬁed Al75

P-t15Au10 alloy in NaOH or HCl aqueous solution. The

deal-loying leads to the formation of nanoporous Pt60Au40

nanocomposites. This electrode showed superior catalytic activity towards the MOR in acidic media compared to the commercial JM-Pt/C catalyst. Lee et al.[65]more recently synthesized octahedral Pt–Pd nanoparticles by means of a polyol process with glycerol as a reducing agent. The dominant exposed surfaces of the octahedral Pt–Pd nano-particles were (111) facets, with well-defined alloy forma-tion between the Pt and Pd metallic phases. According to their analysis, the octahedral Pt–Pd alloy catalysts showed improved specific activity and long term stability compared to polycrystalline Pt/C in the MOR. Other researchers have also reported that Pt and Pd electrodes with (111) facets Figure 5 CV curves of 3D Pt nanoflowers/Si and thin film Pt/Si

in a solution of 1 M CH3OH–1M H2SO4 [33]. Reprinted by permission of the Royal Society of Chemistry.

(7)

exhibited lower onset potentials and higher current densi-ties for methanol than electrodes with other low-index facets [66]. Jiang et al. [67] made bimetallic PtmNin

electrocatalysts with different Pt/Ni atomic ratios through a polyol process. Brieﬂy, H2PtCl6 6H2O and Ni(NO3)2 6H2O

wereﬁrst mixed in 1 M NaOH/ethylene glycol solution and further stirred for 0.5 h to obtain the homogeneous solu-tion. The metal salts were reduced by EG through heating of the homogeneous solution in an oil bath at 1801C for 4 h in the presence of N2 gas. The resulting product was

subse-quentlyﬁltered and washed with a large amount of distilled water. The amounts of the metal (Pt and Ni) precursors were adjusted to maintain the atomic ratios of Pt and Ni. Moreover, due to charge transfer from Ni to Pt atoms in PtmNinclusters, bimetallic PtmNinelectrocatalysts exhibited

enhanced activities for the MOR and decreased CO adsorp-tion in alkaline media.

More recently, Qi et al.[54] fabricated a Pd40Ni60 alloy

catalyst by applying an approach similar to that of Zhang et al.[64]. The Pd40Ni60alloy consists of nanocrystals with

sizes of 5–10 nm, and nanocrystalline and amorphous zones and lattice distortion were observed in the Pd40Ni60 alloy.

The authors reported that Pd40Ni60 alloy had enhanced

electrocatalytic performance towards the MOR in alkaline media than nanoporous Pd. This enhanced electrocatalytic activity is due to the following reasons: (1) the electro-negativity of Pd (2.20) is larger than that of Ni (1.91), thus the transfer of electrons from Ni to Pd may occur, which can decrease the Pd–CO binding energy, improve CO oxidation from methanol dehydrogenation, and enhance the adsorp-tion and oxidaadsorp-tion of methanol molecules. In addiadsorp-tion, like Ru, Ni is an oxophilic element and has the capacity to generate OHads at a lower potential and facilitates the

oxidative desorption of the intermediate products, thus enhancing both the catalytic activity and stability of Pd catalysts[68].

Recently, core–shell nanoparticles have attracted exten-sive attention in materials science because of their many unique physical/chemical properties relative to their single-component counterparts, such as monodispersity, stability, maneuverability, and self-assembly. Long et al. [69] have recently synthesized Pt–Pd core–shell structures by a mod-iﬁed polyol method with the assistance of AgNO3. Their

cyclic voltammetry results showed that Pt–Pd core–shell

structures had excellent electrocatalytic activities com-pared with alloy nanoparticles as well as mixed nanoparti-cles of various single and bimetallic components. Khalid et al. [70] also synthesized bimetallic core–shell Au–Pt

nanoparticle assemblies on silicon and ITO-coated glass substrates. Core–shell Au–Pt nanoparticles were synthesized by the simultaneous reduction of surface-bound [AuCl4]

and [PtCl6]2 ions. In this process, Pt was ﬁrst reduced to

Pt2+ _{from Pt}4+ _{and then to Pt}0_{, with standard redox}

potentials of 0.775 V for [PtCl6]2 /[PtCl4]2 and 0.68 V for

[PtCl4] 2

/Pt0, compared to a single-step reduction for gold from Au3+ to Au0, with a standard reduction potential of 1.002 V for [AuCl4] /Au

0

[71]. On the basis of reduction potentials, Au will preferably nucleate ﬁrst to form the core, followed by Pt to form the shell. Such core–shell structures exhibited enhanced MOR activity, which was attributed to the electronic effect of the Au core Au on the Pt shell. Long et al. [69] prepared shape-controlled Pt–Pd core–shell bimetallic nanoparticles by a modiﬁed polyol process with the assistance of AgNO3. In the case of

epitaxial growth, the overgrowth of Pd shells on Pt cores was observed, but the overgrowth of the Pt shell on Pd cores was non-epitaxial growth. The size range of Pd–Pt core–shell nanoparticles was ∼18–25 nm. Based on electro-chemical measurements, the Pt–Pd core–shell (15 min) catalyst showed more enhanced activity than a Pt–Pd alloy and cluster for the oxidation of methanol. Long et al. reported that Pt–Pd core–shell (15 min) nanoparticles showed higher MOR activity due to their unique structure (shells as atomic monolayers on the cores) and homoge-neous size. Ghosh et al. [73] prepared a Pt Cd alloy by coreduction of K2PtCl6and CdCl2with NaBH4in 1:1 ratio and

annealing at 3001C. The Pt Cd nanoparticles showed much higher MOR activity than Pt due to the presence of the PtCd tetragonal ordered intermetallic phase. Xing et al. depos-ited Pt nanoparticles onto TiO2 nanotubes (Pt/TiO2NTs) by

cyclic voltammetry[58]. The TiO2NT arrays were fabricated

by potentiostatic anodization of the Ti foil. The Pt/TiO2NTs

catalyst exhibited better electrocatalytic activity and sta-bility toward the MOR due to the increase in the oxophilicity of TiO2NTs. More recently, Shi et al. [74] also prepared

Pt/CaδTiOx–Ti2O (Pt/TCT) catalysts by chemical methods.

The nano- and submicro-sized Ti2O and CaδTiOx were

produced by partial metallization of TiO2 nanoparticles

Figure 6 (A) CV curves of Pt/XC72 (dashed line) and Pt/TCT (solid line) in a solution of 1 M CH3OH – 0.5 M H2SO4 at 601C; (B) Chronoamperometry curves of Pt/TCT (ﬁrst and second tests on the same catalyst), Pt/XC72 and PtRu/XC72 catalysts in a solution of 1 M CH3OH– 0.5 M H2SO4at 0.6 V and 601C[74]. Reprinted by permission of the Wiley-VCH Verlag GmbH & Co. KGaA.

(8)

and an eutectic mixture of CaCl2and NaCl at 6001C. Based

on the CV results, Pt/TCT catalysts showed an oxidation peak current that is∼7.5–fold greater than that of Pt/XC72 (Figure 6(A)). In addition, the chronoamperometric results also exhibited that Pt/TCT had higher catalytic activity for the MOR at 0.6 V (Figure 6(B)). Electronic and bifunctional effects affected the comparable catalytic activities of CaδTiOxand Ti2O in partially electrometallized TiO2.

Kang et al. fabricated Mn Pt nanocubes from Pt acet-ylacetonate and manganese carbonyl in the presence of oleic acid and oleylamine [61]. The structure of the as-synthesized Mn Pt nanocubes was converted to an ordered MnPt3intermetallic phase after annealing at 6001C.

Accord-ing to their data, the Mn Pt nanocubes exhibited higher MOR activity than ETEK Pt and spherical Mn Pt nanoparti-cles. Wang et al.[76] synthesized Pt-on-Pd nanodendrites by spontaneous step-by-step depositions of Pd and Pt precursors in the absence of Pd seeds, organic solvent, or high temperature. The CV curves indicated that the current densities of Pt-on-Pd nanodendrites were higher than those of Pt nanodendrites and Pt black. Such a nanodendrite structure was very favorable for reduction of the electronic binding energy in Pt and facilitation of the C–H cleavage reaction in methanol decomposition. Thus, Pt-on-Pd nano-dendrites showed superior MOR activity. Cui et al. [77]

reported the synthesis of porous Pt–Ni nanoparticle tubes for the electrocatalytic oxidation of methanol. The porous Pt–Ni nanoparticle tubes were synthesized by thermally annealing the AAO template-supported Pt–Ni nanoparticle tubes. This porous catalyst exhibited high catalytic activity, improved stability and high resistance to CO poisoning due to compressive strain.

Magnetic materials such as Pt–Co have been used as an ORR

[78], but their suitability as a DMFC anode catalyst has not been explored in great detail. Zeng et al.[48]synthesized carbon-supported Pt–Co and Pt catalysts by NaBH4reduction of metal

precursors. The authors found that the size of nanoparticles changed with the pH; for example, the size of nanoparticles in an alkaline medium was∼12 nm, while the size of nanoparticles in un-buffered solution was∼3.7 nm. They also found that Pt– Co nanoparticle catalysts were more active than Pt-only catalysts in acidic media and that the increase in speciﬁc activity was more than a surface-area effect. They believed that oxophilic Co acts as a catalyst promoter. Recently, Yang et al.[80] prepared high-quality and (100)-facet-terminated Pt3Co and Pt nanocubes in the presence of oleylamine, oleic

acid and argon gas.Figures. 7(A, C and B, D) exhibited TEM

and HRTEM images of the Pt3Co and Pt nanocubes. The

catalytic activity of (100)-facet-terminated Pt3Co nanocubes

towards the MOR was found to be much higher than that of Pt nanocubes, which was attributed to weaker and slower CO adsorption.

Among the various types of double-component catalysts, the Pt–Ru alloy has been found to be the most active and is the state-of-the-art anode catalyst for DMFCs. The enhanced catalytic activity and improved CO tolerance of the Pt–Ru catalyst relative to Pt for CO and methanol oxidation has been ascribed to both a bifunctional mechan-ism [81] and a ligand (electronic) effect [82]. Peng et al.

[83] synthesized 3D nanoporous Pt–Ru bimetallic networks

by decorating nanoporous Pt networks with Ru using a hydrothermal process. Further, a CV curve was used to characterize the electrochemical properties of synthesized nanoporous Pt–Ru networks. The CV curves of a Pt elec-trode, nanoporous Pt and nanoporous Pt–Ru are exhibited in

Figures 8(A) and (B). As shown in Figures 8(A) and (B), nanoporous Pt–Ru signiﬁcantly enhanced the ECSA and MOR activity and reduced the onset potential.

Yoo et al. [51] fabricated multilayered Pt/Ru nanorods with controllable bimetallic sites by the oblique angle deposition (OAD) technique. They synthesized single-segmented Pt nanorods and multi-single-segmented PtRu nanorods with 3, 7, and 13 layers via OAD. The electrochemical results suggest that the MOR over 13-layered PtRu nanorods shows enhanced catalytic activity compared to CeO2

7-layered PtRu nanorods, 3-7-layered PtRu nanorods and Pt nanorods. The enhanced catalytic activity can be attributed to the electronic effect, which promotes the weakening of the bond of a CO-like species. In addition, the catalytic performance of 13-layered PtRu nanorods was also affected by the number of bimetallic pair sites, degree of alloying, and d-band vacancy. More recently, Şen et al. [85] fabri-cated carbon-supported PtRu nanoparticles (Ru/Pt: 0.25) by three different reduction processes: simultaneous reduction of PtCl4 and RuCl3 (catalyst A) and changing the reduction

order of PtCl4 and RuCl3 (catalysts B and C). Their results

suggest that catalyst C had better tolerance to CO poisoning than catalysts A and B. Moreover, catalyst C displayed superior performance relative to catalysts A and B in terms of catalytic activity and stability. Although double-component catalysts such as Pt–Ru exhibited better CO-tolerance and higher MOR activity in acidic and alkaline media, the commercial viability of DMFCs remains hindered due to the high costs (Pt–Ru catalysts) and low abundance of

Figure. 7 (A) TEM and (B) HRTEM images of Pt3Co nanocubes; (C) TEM and (D) HRTEM images of Pt nanocubes[76]. Reprinted by permission of the Wiley-VCH Verlag GmbH & Co. KGaA.

(9)

noble metal catalysts. Furthermore, the supply of Ru and the CO poisoning effect of Ru remain questionable. The research for a low-cost, durable and more active catalyst for the MOR in acidic and alkaline medium is thus of great importance.

Multiple-component catalysts

As discussed previously, we found that the addition of Ru has proven effective in improving anti-poisoning performance. However, the high cost of the Pt–Ru catalyst is also one of the main obstacles to its scaled application [86]. The wide

commercialization of DMFCs mainly depends on decreasing the cost of the catalysts[87]. The above-mentioned problems can be solved by modifying Pt catalysts with other metals or metal oxides that have higher tendencies to form surface-oxygenated species at lower potentials. To this end, serious efforts have been made by several groups to synthesize new catalyst systems for methanol oxidation based on multiple-component catalysts such as PtRuNi[88–90], PtRuMo[89,91,92], and PtMOx (M=Ti, V, Mn, W)[93–96]. Wang et al.[92]prepared a PtRuMo catalyst by an impregnation reduction process. In the synthesis of the PtRuMo catalyst, sodium borohydride was used to chemically reduce the precursors of H2PtCl6, RuCl3 and

(NH4)6Mo7O24, in different atomic ratios. The authors found

Figure. 8 CV curves of polycrystalline Pt electrode, nanoporous Pt (A) and Pt–Ru (B): (a) in 0.5 M H2SO4at 20 mV s 1; and (b) in 0.1 M CH3OH-0.5 M H2SO4at 20 mV s 1[83]. Reprinted by permission of IOP Publishing Ltd.

Figure. 9 (A, B) SEM and corresponding HRSEM images of Pt nanowire–Sn@CNT. (C, D) TEM and corresponding HRTEM images of Pt nanowire–Sn@CNT[98]. Reprinted by permission of the Wiley-VCH Verlag GmbH & Co. KGaA.

(10)

that PtRuMo in a molar ratio of 6:3:1 had higher catalytic activity and durability for the MOR than homemade PtRu/C and Pt/C. The promotion effect of Mo in the PtRuMo catalyst could be ascribed to the bifunctional mechanism and ‘electronic effect’ of Mo metal. According to the bifunctional mechanism, MoOx mixed with Ru further supports H2O activation, in

comparison to Ru alone, which subsequently facilitates CO oxidation, resulting in enhanced MOR activity. Chai et al.[97]

prepared Pt–Ru–Co–W quaternary anode electrocatalysts on a conductive substrate by using a robotic dispensing system. According to data analysis, the Pt–Ru–Co–W quaternary anode electrocatalysts exhibited markedly increasing current densi-ties due to the increase in COadsoxidation by addition of more

oxophilic elements to Pt. Sun et al. [98] reported a newly fabricated 3D nanocomposite composed of Pt nanowires and a coaxial nanocable support consisting of a tin nanowire and a CNT (Figure 9). The ultrathin single-crystal Pt nanowire/ Sn@CNT was synthesized via a facile surfactant-free aqueous solution method. This 3D Pt nanowire/Sn@CNT electrode exhibited higher MOR activity and improved CO tolerance due to its higher gas permeability, improved metal-support inter-actions, and enhanced mass transport. Sun et al.[99] synthe-sized Pt–Ru/CeO2/MWNT catalysts by a sonication deposition

process. They reported a higher electrochemically active sur-face area of Pt Ru/CeO2/MWNT because of a synergistic effect

between Pt and the ionic conductor CeO2,which may promote

the hydrogen spillover rate of Pt H and thus increase the dissociation of hydrogen adsorption.

In addition, Pt Ru/CeO2/MWNT displayed the highest

oxidative current, further indicating its superior electro-chemical performance for the MOR compared to Pt Ru/ MWNT and Pt/MWNT catalysts. This enhancement can be attributed to the bifunctional mechanism, in which CeO2

may promote the dissociation of coordinated water in the manner of Ru, forming more OH species to oxidize COadand

release more Pt active sites. Yamazaki et al. developed a CO-tolerant anode catalyst composed of PtRu and a Rh porphyrin [100]. By using this catalyst, hydrogen in the anode gas was oxidized by PtRu, and CO in the anode gas was oxidized via the Rh porphyrin. This porphyrin reduces the CO concentration around Pt Ru through the CO oxida-tion process. Kim et al. fabricated shape- and composioxida-tion- composition-controlled Pt–Fe–Co nanoparticles (nanocubes, branched nanocubes, nanoparticles with low cobalt content and nanoparticles with high cobalt content) for enhanced MOR activity [101]. Pt–Fe–Co branched nanocubes showed the highest activity and durability toward MOR. This substantial increase in MOR activity and durability is attributed to the presence of a small amount of Co, which may have affected the electronic structure of the nanoparticles. Ahn et al.

[102]prepared a CoPtRu catalyst through electrochemical methods on a carbon paper substrate. In a typical synthesis of CoPtRu, Co particles wereﬁrst deposited on carbon paper via an electrodeposition method by changing the deposition potential and time. Subsequently, Pt and Ru galvanic displacements were carried out by controlling displacement time. Compared with other methods, electrochemical pro-cesses have many advantages, such as high deposit purity, short preparation time, and facile control of size and composition of the CoPtRu catalyst. Ahn et al. found that the CoPtRu catalysts showed superior catalytic activity for the MOR and better CO tolerance than a commercial PtRu/C

catalyst. Sarkar et al. [103] fabricated carbon-supported PtPdCo nanoalloy electrocatalysts by a microwave-assisted solvothermal method. These materials were synthesized at 3001C without any post-annealing in reducing gas atmo-spheres. The PtPdCo nanoalloy electrocatalysts showed high tolerance to methanol due to surface strain effects, which developed on the outermost Pt and Pd atomic layers. Saida et al. reported the use of TiO2 nanosheets (TiO2ns,

[Ti4O9]2 ) as an additive to a PtRu/C anode catalyst for

the MOR[104]. The thickness of two-dimensional nanocrys-tallite TiO2ns was approximately 1 nm. When TiO2ns content

was low, the mixed TiO2ns PtRu/C nanocomposite

exhib-ited high activity and durability due to an increase in the interphase between the electrolyte and PtRu nanoparticles and high ECSA. However, a high content of TiO2ns tended to

have poor current efﬁciency compared with pristine PtRu/ C. Jeon et al. reported a new and economic method to improve the MOR activity of the PtRu catalyst, which was accomplished by mixing the PtRu catalyst with Fe2O3

nanoclusters[105]. These authors found that the composite catalyst was more active if 10 wt% Fe2O3was mixed with the

PtRu catalyst. The MOR activity was increased by 80% over that of the pure PtRu catalyst. The enhanced MOR activity was attributed to the change in electronic state of Pt through the Fe2O3nanoclusters. Eguiluz et al.[106]

synthe-sized carbon-supported Ptx (RuO2 M)1 x composite

tern-ary catalysts (M=CeO2, MoO3, or PbOx) by the sol–gel

process. Carbon-supported Ptx (RuO2 M)1 x composite

ternary catalysts were then used as catalysts for the MOR studies. The current density of carbon-supported Pt0.50(RuO2 CeO2)0.50 catalysts was much higher than that

of commercial Pt/C at 450 mV. Due to the oxygen capture and release capacity of ceria, the carbon-supported Pt0.50(RuO2 CeO2)0.50 catalysts exhibited the highest

elec-trocatalytic activity toward the MOR (onset potential ∼207 mV). Li et al. synthesized a tri-component hybrid, such as Pt nanoparticle-polyoxometalate-CNT, via a green approach at room temperature[107]. The polyoxometa-lates act as both reducing and bridging molecules. The nanohybrid exhibited higher MOR activity than a tradi-tional Pt–C catalyst and other reported Pt/CNT systems due to the following three major factors: (i) the high-conductivity of CNTs facilitated rapid electron transfer between the target molecule and electrode, (ii) the small size of the Pt nanoparticles was beneficial to MOR activities, and (iii) the existence of the layer of poly-oxometalate around Pt nanoparticles may be helpful for the catalytic activities and anti-CO-poisoning properties of the catalyst. More recently, Wu et al. fabricated FePt– Au heterostructured nanocrystals (HNCs), such as tad-pole-, dumbbell-, bead-, and necklace-like nanostruc-tures by heteroepitaxial growth of Au nanocrystals (NCs) onto FePt nanorods (NRs)[108]. The tadpole-like FePt/C exhibited significantly higher MOR catalytic activity than that of commercial Pt catalysts; this activity was also significantly higher than that of FePt/C catalysts (Figure 10).

Pan et al. [109] fabricated Pt–Sb-doped tin oxide

nano-particles on carbon black (Pt–ATO/C) by using an in situ co-precipitation method and polyol process. The Pt–ATO/C catalyst showed higher MOR activity than the Pt–SnO2/C or

(11)

was attributed to the superior electrical conductivity of Sb-doped SnO2, which induced the electronic effects with Pt

catalysts. Liu et al. [110] prepared nanoporous Pt-based multimetallic alloy nanowires by using a template-assisted electrodeposition process. The TEM images and electron diffraction patterns of nanoporous ternary PtCoNi and PtCoAu, as well as quaternary PtRuCoNi nanowires, are shown inFigure 11. As shown in Figure 11, the nanowires were continuous and dense, with uniform diameters along the entire length of the wires. The average diameters for PtCoNi, PtCoAu and PtRuCoNi nanowires were 50, 48 and 45 nm, respectively. Electrochemical measurements indi-cated that nanoporous Pt57Co31Ni12 and Pt18Ru15Co48Ni19

nanowires markedly enhanced their durability upon contin-uous potential cycling.

We have shown experimentally that MOR activity and stability could be enhanced through multiple-component cata-lysts. However, the present design of multiple-component catalysts may not be optimal because researchers have explored only a limited number of such catalysts. In addition, although these multiple-component catalysts exhibited excel-lent performance toward the MOR, they still contained high Pt content of more than 50%. Therefore, further research and studies are required to investigate approaches for designing multiple-component catalysts.

New catalyst supports

Multiple-component catalysts are restricted by their low MOR activity, which is due to their low surface area. Because low surface area may restrict widespread use, researchers seek to increase the surface area and improve electrocatalytic activity and utilization. One approach to this problem is use of high surface area materials instead of multiple-component catalysts. The cost of Pt is also an issue of concern. For instance, in 2008, it was estimated that, on the basis of a recent peak price of Pt of over $2200/oz, the cost of Pt alone used in a 100-kW PEMFC engine (∼0.8 g/kW) is substantially greater than the current price of an entire internal combustion gasoline engine of equal power [111]. In the last several years, many high surface area based materials have been proposed as Pt supports for DMFCs

[112–116]. Regarding this issue, Li et al. [107] synthesized Pt nanoparticle-decorated carbon nanotubes (CNTs) via a green chemistry approach in the presence of polyoxometa-lates (POMs). The POMs functioned as reducing and bridging molecules in nanohybrids. In addition, due to the superior catalytic properties of POMs, the metal nanoparticles-POM Figure 10 CV curves of the MOR on different catalysts in a

0.5 M HClO4–0.5 M CH3OH solution[108]. Reprinted by permis-sion of the Tsinghua University Press and Springer-Verlag Berlin Heidelberg.

Figure 11 TEM images of the as-prepared nanowires (A) Pt1Co74Ni25, (B) Pt4Co94Au2 and (C) Pt0.6Ru0.9Co76.4Ni22.1, and the dealloyed nanoporous (D) Pt57Co31Ni12, (E) Pt30Co56Au14, and (F) Pt18Ru15Co48Ni19. Electron diffraction patterns of nanowires showed a single set of fcc-like rings (insets of (D)–(F)), suggesting that all nanowires consisted of a single-phase alloy[110]. Reprinted by permission of IOP Publishing Ltd.

(12)

composite supported on CNTs may provide enhanced catalytic activities [50]. According to their data analysis, the Pt nanoparticles-POM/CNT exhibited higher electrocatalytic activity towards the MOR than the traditional Pt–C catalyst and other reported Pt/CNT systems. The excellent electro-catalytic performance of the Pt nanoparticles–POM/CNT electrode can be explained as follows. First, the high conductivity of CNTs accelerated electron transfer between the target molecule and electrode. Second, the well-deﬁned small Pt nanoparticles have enhanced electrocatalytic activ-ities. Third, the existence of the layer of POM around Pt nanoparticles may be helpful for the electrocatalytic activity and anti-poisoning properties of the catalyst. Hsieh et al.[119] also assembled bimetallic Pt–M (M=Fe, Co, and Ni) on CNT via a two-step chemical reﬂux method. The CNTs were prepared by catalytic chemical vapor deposition using ethylene and Ni particles as the carbon precursor and catalyst, respectively. On the basis of the electrochemical measurements, the authors found that Pt– Co/CNT catalyst had superior electrochemical activity, anti-poisoning ability, and long-term cycleability relative to Pt–Fe/CNT and Pt–Ni/CNT catalysts due to the bifunc-tional mechanism of bimetallic catalysts. They reported that two types of mechanisms were involved in enhancing CO tolerance. First, CO adsorption should occur mainly on Pt, while OH should interact preferentially with Co. Thus, the proximity of CO and OH-adsorbed species could explain the promoter effect of Co toward CO oxidation on Pt-Co anode catalysts, leading to a high level of CO tolerance in methanol oxidation. Second, the presence of Co could promote the combined effects of H2O dissociation and CO

oxidation, creating a larger number of active sites for methanol oxidation. Ding et al.[120] synthesized a highly efﬁcient porous hollow carbon nanostructure supporting PtRu catalysts by a one-step pyrolysis process. The PtRu (with 18.5 wt% Pt and atomic ratio of Pt/Ru=1:1) catalysts supported on hollow carbon nanospheres (HCNS) exhibited high electrochemical activity and stability toward the MOR. This enhancement of the MOR catalytic abilities can be attributed to the unique structure of the carbon nanostructure and the pyrolysis-induced high stability and alloying degree of the loaded metallic catalysts. In addi-tion, the PtRu/HCNS catalysts also displayed higher CO tolerance due to the high alloying degree-enhanced bifunctional mechanism, in which Ru supplies an oxyge-nated surface species by dissociating H2O at lower

poten-tials with respect to bulk Pt or separated Pt sites. Our groups synthesized a low cost 3D nanoporous graphitic carbon (g–C) material by using an adamantane (C10H16)

ﬂame [121]. The synthesized g–C was used as a Pt–Ru

catalyst support because of its very high surface area. The electrochemical measurements showed that the supported Pt–Ru has higher activity towards the MOR, which is attributed to the presence of 3D nanopores in the g–C support by virtue of which easy transport of methanol and the oxidation products is possible. Joo et al.[122] synthe-sized highly crystalline graphitic nanocarbons (GNC) by the wet air treatment of hydrothermally derived graphitic porous carbon. They reported that the morphology and degree of graphitic crystallinity changed with tempera-ture. For instance, GNCs consisted of aggregates of silkworm-shaped carbon nanoparticles with enhanced

graphitic characteristics at 4501C. GNC was tested as a Pt catalyst support in the MOR. Based on their results, the Pt/GNC catalyst had a higher electrochemically active surface area than the Pt/C catalyst. These composite catalysts have a positive effect on Pt dispersion, crystal-lographic orientation, electrical conductivity and electro-chemical stability, resulting in the high MOR activity. Sharma et al. [123] prepared reduced graphene oxide/Pt supported electrocatalysts (Pt/RGO) by a microwave-assisted polyol process. The resulting products Pt/RGO exhibited excellent catalytic activity for MOR. They have also found that the CO oxidation activity of the Pt/RGO hybrid electrode was higher than that of the commercial Pt/C catalysts due to the presence of residual oxygen groups on RGO.

Three possible mechanisms were involved in enhancing the CO tolerance and methanol oxidation of these materials.

First, RGO promotes water activation due to its hydro-philic nature. Thus, the adsorbed OH species at the Pt edge promote CO oxidation.

Second, strong interaction between Pt and RGO was found, which can induce some modulation in the electronic structure of Pt clusters, modifying the Pt CO binding energy and, as a result, minimizing CO adsorption on Pt.

Third, Pt/RGO catalysts can promote hydrogen spillover (the process involving dissociative chemisorption of molecular hydrogen on a supported Pt catalyst surface, followed by the diffusion of atomic hydrogen onto the surface of RGO). The RGO species on the Pt surface promote the formation of GO) species adjacent to CO-poisoned Pt sites, which combine with adsorbed CO to strip it from the surface as CO2.

Pt/TiC was prepared by using a simple electrodeposition process to load Pt nanoparticles on TiC nanocomposite by Ou et al. [124]. They reported that the Pt/TiC catalyst is able to reduce the risk of the CO poisoning effect for the MOR. The enhancement in the MOR catalytic activity was ascribed to OH groups (formed by water discharge on the TiC surface) promoting CO removal near the Pt-oxide inter-face and strong metal support interaction. Tiwari et al.

[125]prepared amorphous carbon-coated silicon nanocones (SiNCs) by an anodic aluminum oxide (AAO) templation method and microwave plasma chemical vapor deposition (MPCVD). The Pt nanoparticles were electrodeposited on amorphous carbon-coated SiNCs (Pt/ACNC) and used as the catalyst for the MOR. The Pt/ACNC catalyst showed superior MOR performance in terms of mass activity and current density. The higher MOR activity of the Pt/ACNC catalyst could be attributed to the large ECSA. In addition, the small (nm scale) Pt catalyst nanoparticles could increase CO oxidation via the bi-functional and L–H mechanisms, produ-cing more active Pt sites for the MOR and thus enhanprodu-cing MOR efficiency. To further enhance MOR activity, Tiwari et al. synthesized 3D Pt nanopetals on SiNCs by pulse-electrodeposition [126]. A CV study on the 3D Pt nanope-tals/SiNCs electrode revealed higher current density than those of Pt nanoparticles/flat Si and Pt nanoflowers/flat Si electrodes. The higher MOR activity was attributed to abundance of a large ECSA for facile transport of methanol, SiO2sites in the vicinity of the SiNCs, as well as less contact

area between the Pt nanopetal catalyst and SiNCs. Fang et al. [127] synthesized Pt nanoparticles with an average size of 3.14 nm on N-doped CNTs by a chemical approach.

(13)

The Pt/N-doped CNTs exhibited the best catalytic activity because the so-called N incorporation could be used to produce preferential sites of CNTs with low interfacial energy for immobilizing Pt nanoparticles.

More recently, graphene has been studied as a catalyst support in DMFCs due to its basal plane structure with high surface area (calculated value, 2630 m2_{/g), high}

conductiv-ity (103 104S/m) and potentially low manufacturing cost

[128–131]. Zhou et al.[132]fabricated graphene-supported Pt nanoparticles via a one-step electrochemical approach. According to their data, Pt nanoparticles on graphene exhibited much higher catalytic activity and long-term stability towards the MOR than the Pt nanoparticles on Vulcan. Guo et al.[133]also fabricated high-quality 3D Pt-on-Pd bimetallic nanodendrites/graphene nanosheets by a wet-chemical approach. Due to their high ECSA, the gra-phene/bimetallic nanodendrite hybrids showed much higher electrocatalytic activity toward the MOR than Pt black and commercial E-TEK Pt/C catalysts. They also reported that the number of branches for Pt-on-Pd bimetallic nanoden-drites on graphene nanosheets was controlled by a few experimental parameters, thus resulting in increased cata-lytic properties. Joo et al.[122]synthesized highly crystal-line graphitic nanocarbons (GNC) by the wet-air treatment of hydrothermally (4501C) derived graphitic porous carbon. At 4501C, the products consisted of aggregates of silkworm-shaped carbon nanoparticles with enhanced graphitic prop-erties. The Pt nanoparticles were coated on GNC by a reduction method (sodium ethoxide as the reducing agent). Pt/GNC exhibited the highest MOR activity (Figure 12) due to enhanced graphitic characteristics with highly dispersed Pt nanoparticles on the graphitic layers.

Zhang et al. prepared large-scale single-crystalline hol-low nanobowls of pure C60 by applying a sonophysical

strategy in a binary organic solution of m–xylene and acetonitrile[135]. The Pt nanoparticles were deposited by electrolytic reduction of an aqueous solution of PtCl62 .

Figure 13(A, B) shows the SEM images of the resulting Pt/ C60 hollow nanobowl. As shown in Figure 13 (A, B) the Pt

nanoparticles were deposited well on both the outside and inside surfaces of the C60hollow nanobowls. Due to the high

surface area of the Pt/C60 hollow nanobowl, these hollow

nanobowls exhibited signiﬁcantly enhanced catalytic activ-ity toward the MOR (at 0.65 V,Figure 13C).

A quick literature survey can easily give readers many articles on support materials with highly active surface areas, such as 2D or 3D nanostructures, for good dispersion of catalysts[136,137]. Oh et al.[137]fabricated a 3D TiO2

nanostructure support by a seeding process with 1D TiO2

nanowires as a seed. The Pt/3D TiO2 catalysts exhibited

much-enhanced catalytic activity and stability toward the MOR due to its high speciﬁc surface area and improved electronic transfer efﬁciency via 3D TiO2 nanostructure

support.

Only a few supported materials (CNT, GNC, g-C, RGO, SiNC, TiC, TiO2) have been invented for anode catalysts,

which is more crucial in enhancing DMFC performance. Therefore, materials scientists mustﬁnd a newer and better supported material that converts a fuel source directly into electrical energy.

Cathode catalysts of DMFCs

Like the anode electrode, the cathode electrode of DMFCs also lacks an adequate electrocatalyst. Therefore, it is necessary to develop new cathode electrocatalysts (low cost and better durability) that have high electrocatalytic activity for the oxygen reduction reaction at low tempera-tures. In the following section, recent progress in the development of cathode electrocatalysts for DMFCs is discussed in detail. For a detailed discussion of cathode electrocatalysts, we collected more than 100 recently published works from different journals.

Single-component catalysts

As a common catalyst for the ORR, noble metals contribute significantly to DMFCs [138–145]. However, there are still some problems, such as sluggish kinetics and poor electro-catalyst durability of the ORR at the cathode, that limit the efficiency of DMFCs. Therefore, development of a highly active and durable catalyst is greatly sought to improve the ORR performance of DMFCs. The goal for materials scientists is tofind various kinds of catalysts to enhance the activity for the ORR. In this regard, researchers began their search with single-component catalysts. In the past decade, sev-eral single-component catalysts were reported by materials scientists. There have been several single-component cat-alysts (such as Ag [138], Au [138,139], Pt [140], Pd

[141,142], nitrogen (N)-doped graphitic carbon [135], N-based carbon nitride[144], MoN [145], N-doped carbon nanotubes (CNTs)[146]) that have shown improvements of ORR activity. In addition, carbon materials have been identiﬁed as some of the most promising materials for DMFCs in acidic media due to their high chemical stabilities, high electric conductivities and enhanced mass transport capabilities[147]. Modifying carbon materials by different N-functional groups is known to enhance their activities for the ORR[148]. Various types of N-functional groups can be initiated on carbon surfaces by varying the experimental procedure and parameters[149]. N-doped carbon materials Figure 12 CV curves of the three catalysts were collected in

a 0.5 M H2SO4–1 M CH3OH solution. Reprinted by permission of the Tsinghua University Press and Springer-Verlag Berlin Heidelberg.

(14)

have pyridinic and pyridinic-N-oxides as nitrogen species, which are responsible for the enhanced ORR activity[150]. N-functional groups transform to more thermally stable structures during heat treatment[149]. Furthermore, nitro-gen is also known to efﬁciently create defects on carbon materials, which may increase the edge plane exposure and thus enhance the ORR activity [151]. Nagaiah et al. [146]

synthesized N-doped CNTs (NCNTs) by treating HNO3

-oxi-dized CNTs (OCNTs) in ammonia ﬂow (25 sccm) at 200 1C (NCNT-200), 4001C (NCNT-400), 600 1C (NCNT-600) and 8001C (NCNT-800) for 6 h. They found that NCNT-800 showed greatly enhanced ORR activity in alkaline media. In addition, NCNT-800 also showed a favorable positive onset potential for the ORR, increased reduction currents, and high stability. Liu et al. [143] fabricated N-doped ordered mesoporous graphitic arrays (NOMGAs) on the basis of a metal-free casting technology. Mesoporous silica (SBA– 15) and N, N′-bis(2,6-diisopropyphenyl)- 3,4,9,10-perylene-tetracarboxylic diimide (PDI) were used as the template and carbon precursor, respectively. NOMGAs with different compositions were fabricated by the carbonization of PDI/SBA-15 composites at 600, 750, and 9001C; the result-ing materials are represented as 600, 750, and 900, respectively. The ORR activities of Pt–C, 600, PDI-750, and PDI-900 catalysts are shown inFigure 14. As shown

inFigure 14, PDI-900 catalysts showed greater ORR activity than did PDI-600, PDI-750, and commercially available Pt–C catalysts. According to this report, the electron transferred value of PDI-900 is∼3.89. The results have shown that the PDI-900 catalyst acts as a 4-electron transfer for the ORR. The unique features of the PDI-900 catalyst, including high surface area and a graphitic framework with a moderate N content, led to high ORR activity, excellent stability, and resistance to crossover effects for the ORR. More recently, Yang et al.[144]fabricated graphene-based carbon nitride (G–CN) nanosheets and studied their ORR properties. They found that the G–CN nanosheets not only possess high N contents, low thicknesses, high surface areas, and large aspect ratios but also show enhanced electrical conductiv-ity. The above properties are favorable for the access of O2

to the catalyst surface and can facilitate the fast diffusion of electrons in the electrode during the ORR process. As a result, the G–CN nanosheets exhibited excellent ORR activ-ity, high electrocatalytic activactiv-ity, long-term durabilactiv-ity, and high selectivity, compared to those of CN sheets and Pt–C catalysts (Figure 14).

Although metal-free catalysts such as G–CN nanosheets, PDI-900 catalysts and NCNT-800 have low costs, the com-mercial viability of DMFCs is still hindered by poor kinetics and slow ORR activity.

Figure 13 SEM images of Pt-deposited C60hollow nanobowls (A) outside Pt deposition and (B) inside Pt deposition. (C) CV curves of (1) Pt/C60hollow nanobowl and (2) Pt/C60solid nanoball (2) electrodes in a 0.1 M H2SO4–2 M CH3OH solution[135]. Reprinted by permission of the Wiley-VCH Verlag GmbH & Co. KGaA.

(15)

The search for a stable and more active electrocatalyst for ORR activity in acidic and alkaline media is thus of great importance. As a consequence, the development of single-component (such as Pt, Ag, Pd or Au) catalysts with high ORR activity has recently become a major focus of DMFC research. Kuai et al. [138] reported uniform, high-yield (490%) icosahedral Ag and Au nanoparticles by using a hydro-thermal system in the presence of PVP and ammonia. The Ag and Au icosahedra showed excellent stability and much higher electrocatalytic activity than spherical nanoparticles; they exhibit positive shifts in the reduction peak potential for oxygen of 0.14 and 0.05 V, while the reduction peak currents of the Ag and Au icosahedra were 1.5- and 1.6-fold, respectively, better than those of spherical nanoparticles and Pt/C catalysts. Jeyabharathi et al.[139] electrodeposited Au atomic clusters in the presence of cetyltrimethylammonium bromide (CTAB) and studied their ORR catalytic activities. Some interesting properties were exhibited by such clusters: (a) molecule-like voltammetric features; (b) electrocatalysis of the oxygen reduction to H2O via a 4-electron pathway in acidic media;

and (C) representation of a transition of the ORR mechanism from four-electron to two-electron reduction[139]. At present, the most widely used cathode catalyst consists of Pt nano-particles due to its high ORR activity. Therefore, many researchers synthesized metal-based nanostructured materials for kinetically enhancing the sluggish electrode reaction

[138,139,142,152]. Wang et al.[153]reported a simple method

to monodisperse Pt nanoparticles with a nanometer size range (3–7 nm) and controlled polyhedral, truncated cubic, or cubic shapes; they also studied their ORR activities. In their synthesis, the sizes and shapes of the Pt nanoparticles were controlled by the reaction temperature.Figure 15displays TEM images of the 3, 5, and 7 nm Pt nanoparticles.

The inset of each TEM image shows a high-resolution TEM image of a single nanoparticle. The inset of Figure 15

(A) corresponds to Pt (111) lattice fringe. However, the insets ofFigure 15(B, C) correspond to Pt(100) and Pt(100) lattice fringes. According to their data, Wang et al. found that the current density from the ORR for Pt nanocubes is 4-fold that of polyhedral Pt or truncated cubic Pt nanoparti-cles, showing that ORR activity is indeed dependent on the shape, not on the size, of the Pt nanoparticles. More recently, Yu et al.[154] fabricated Pt concave nanocubes enclosed by high-index facets, including (510), (720), and (830), via addition of a NaBH4 solution and a mixture

containing K2PtCl4, KBr, and Na2H2P2O7 into DI water. In a

typical synthesis, the pyrophosphato complex and the slow addition of this precursor by a syringe pump contributed to the formation of Pt concave nanocubes. In this regard, the seeds selectively overgrow from corners and edges, and the Br ion acts as a capping agent to block the (100) facets. The Pt concave nanocubes showed substantially enhanced ORR activity relative to that of Pt nanocubes, cuboctahe-dra, and commercial Pt/C catalysts that are bounded by Figure 14 (A) CV curves of G–CN800 in O2-saturated 0.1 M KOH with different rotation speeds at a constant scan rate of 5 mV s 1. (B) Koutecky–Levich plots of G–CN800 obtained from CV curves in (A) at various electrode potentials. (C) CV curves of G–CN and CN nanosheets at a constant rotation rate of 1600 rpm. (D) Electrochemical activity, shown as the kinetic-limiting current density (JK) at

(16)

low-index facets such as (100) and (111). Further, ORR activity can be enhanced through free-standing Pt-nano-wires because these nanoPt-nano-wires possess many novel struc-tural characteristics, including flexibility, large area per unit volume, high stability, preferential exposure of highly active crystal facets, and easy electron transport[155–157]. To address this issue, Liang et al.[158]used a templating method to prepare a free-standing Pt nanowire, and its ORR activity was studied. The free-standing Pt nanowire catalyst exhibited high stability compared with Pt/C and Pt black. The free-standing Pt catalyst exhibited remarkably high stability because of its unique supportless nanowire network structure (rapid electron transport and gas diffusion) and the preferential exposure of certain crystal facets in 1D Pt nanowires. Wang et al.[159]synthesized a Pt nanoparticle netlike assembly through a hydrothermal method. The Pt nanoparticle netlike assembly exhibited higher durability and 2.9 times higher mass activity for ORR than the commercial Pt black catalyst due to its high specific surface area and large overall size. Porous nanotubes exhibited several novel structural characteristics, including high por-osity,flexibility, large area per unit volume, and an inter-connected open pore structure. Alia et al.[39]fabricated porous Pt nanotubes with a thickness of 5 nm, an outer diameter of 60 nm, and a length of 5–20 μm. The porous Pt nanotubes were synthesized by galvanic displacement with Ag nanowires, which were obtained by the ethylene glycol reduction of silver nitrate. They also evaluated the ORR activity and durability of porous Pt nanotubes, and they found that the stability and ORR activity of porous Pt nanotubes are much higher than those of Pt/C and bulk polycrystalline Pt catalysts. Although single component metals (such as Pt, Pd or Au) have high ORR activity and stability, dissolution/agglomeration of these single-component metals from cathodic catalysts contribute to the perfor-mance decay of DMFCs. Furthermore, researchers believe that the stability and activity of the Pt-based fibrous membrane catalyst can be further improved by alloying with other metals to form binary catalysts.

Double-component catalysts

As we discussed above, single-component nanostructured catalysts such as Pt, Pd, and Au can increase ORR activity

relative to Pt, Pd, and Au bulk metals[161]. Among these catalysts, nanostructured Pt is a traditionally great electro-catalyst for ORR but is very costly for commercialization in DMFCs [162]. To utilize Pt loading in a limited resource, double-component nanostructures are developed. In 2005, Wang et al. [163] showed that ORR activity would be enhanced by coupling a metal X (X=Co, Ni, Cr, or V) with a low occupancy of d-orbitals with another metal X′ (such as X′=Pd, Ag, and Au) with fully occupied d-orbitals [154]. The d-orbital coupling effect between metals can signiﬁ-cantly decrease the Gibbs free energy for the electron transfer steps in the ORR, resulting in enhanced ORR kinetics

[163]. Xia et al.[145]synthesized a MoN electrocatalyst via heat treatment of molybdenum tetraphenylporphyrin in the presence of a gaseous atmosphere of ammonia. The synthesis was followed by heat treatment at various temperatures in the presence of ammonia gas. In the presence of ammonia, the following type of reaction may occur and provide the source of N for MoN formation:

2NH3=3H2+2[N] (10)

The electrochemical measurements showed that the MoN catalyst has strong ORR activity, proceeding by an approx-imate four-electron pathway (the electron transfer number calculated based on the measured slope value was ∼3.8, which is very close to 4), through which molecular oxygen is directly reduced to water by accepting four electrons. Sarkar et al.[165]synthesized Pd–W nanoalloy

electrocata-lysts by simultaneous thermal decomposition of palladium acetylacetonate and tungsten carbonyl in o-xylene in the presence of Vulcan XC–72R carbon, followed by annealing up to 8001C in a hydrogen environment. Low cost Pd–W nanoalloy is found to increase both the catalytic activity for ORR and the catalyst durability. In addition, Pd–W nanoalloy catalysts offer the important advantage of high tolerance to methanol compared to Pt. The origin of the enhanced ORR activity has been attributed to the following characteristics: (i) modiﬁcation of the electronic structure of Pt (5d-orbital vacancies), (ii) changes in the Pt–Pt bond distance and coordination number, and (iii) inhibition of adsorbed oxygen-containing species from the electrolyte onto Pt. In 2008, Camargo et al.[166]synthesized RuSe2+d

nanotubes as an electrocatalyst for the ORR. These nano-tubes were synthesized by template-engaged replacement Figure 15 TEM and HRTEM images of (A) 3 nm, (B) 5 nm, and (C) 7 nm Pt nanoparticles[153]. Reprinted by permission of the Wiley-VCH Verlag GmbH & Co. KGaA.