ABSTRACT

Highly-dispersed platinum and platinum-based catalysts on a conductive support are commonly used as electrode materials in low-temperature fuel cells, particularly the hydrogen PEMFC and the direct methanol PEMFC. The performance and durability/stability of these catalysts strongly depend on the characteristics of the support. Catalysts supported on high surface area carbon black are widely used in low-temperature fuel cells. However, the corrosion of carbon black has been recognized as one of major causes of performance degradation and durability issues of low-temperature fuel cells under high-potential conditions. So the need for alternative supports with outstanding physical and mechanical properties to carry out the successful reaction in catalyst layer and give a longer lifetime for the electrocatalysts is inevitable. The emergence of nanotechnology and development of nanostructure materials in recent years has opened up new avenues of materials development for low-temperature fuel cells. This paper presents the performance with a variety of carbon-based nanostructured materials such as carbon nanotubes (CNT), carbon nanofibers (CNF), carbon aerogels, nanoplates of graphene, etc. So the present paper provides an overview of these nanostructured materials as low-temperature fuel cell catalyst supports. The improved characteristics of the nanostructured supports with respect to commercially used carbon black (Vulcan XC-72) and their effect on the electrochemical activity are highlighted. Additionally, it reviews the literature on the synthesis of nanostructured-supported Pt electrocatalysts for proton exchange membrane (PEM) fuel cell catalyst loading reducing through the improvement of catalyst utilization and activity. The features of each synthetic method were also discussed based on the morphology of the synthesized catalysts.

1. Introduction

In recent years, the concept of green energy has attracted considerable attention. The aim of green energy is to explore techniques and methodologies that reduce or eliminate the generation of products or byproducts that are hazardous to human health or the environment. Nowadays, green energy is becoming one of the main goals of designing new processes and reactions in the emerging areas of nanoscience and nanotechnology. Fuel cells are electrochemical conversion devices used for power generation in portable, stationary/residential and transportation applications. As a “green/clean” electric power sourcefuel cells can be used to power vehicles, back-up the power supply for electric devices, and store electricity in power stations by converting water into hydrogen and oxygen during off-peak hours. The only by-products are water and heat. Among various types of fuel cells, proton exchange membrane fuel cells (also called polymer electrolyte membrane fuel cell, PEMFCs) have been considered to be a promising energy source because they offer a highly efficient, low operation temperature (below 100˚C), higher power density, rapid start-up/shut-down, low weight, compactness, potential for low cost and volume, long stack life, suitability for discontinuous operation and environment-friendly technology for energy conversion without the Carnot limitations of combustion engines. Although great progress has been made in the last several decades in the research and development of PEM fuel cells, several technical challenges should be overcome for wide-scale commercialization of PEMFCs. These challenges include reducing the cost of component, optimizing the utilization of electrocatalyst at all current densities, improving the performance and durability/stability of the membrane electrode assembly (MEA), the crucial part of the fuel cell system. As the key component in PEMFC, the electrocatalyst layer is the place where the electrochemical reactions take place and the electrical energy is produced, exhibiting a great influence on the total performance and durability. The main processes that occur in the catalyst layer include mass transport, interfacial reactions at electrochemically active sites, proton transport in the electrolyte phase, and electron conduction in the electronic phase. The PEM fuel cell MEA utilizes precious metal catalyst (such as Pt or Pt alloy) dispersed on support. These catalysts are estimated to contribute to as much as 50% of total fuel cell cost. It was emphasized that in order to reach the DOE (department of energy) cost reduction objects, among other component cost reductions, the current state of the art platinum loading (~0.4 mg/cm²) of PEM fuel cell electrodes, the total Pt catalyst loading in a MEA must be reduced to levels lower than 0.1 mg/cm² to come across the target for commercialization [1]. Similarly, durability targets for these applications have not been mentioned for catalysts, but MEAs are expected to endure for >40,000 and 3000 hours for stationary and portable applications respectively, with low degradation rates.

Several strategies have been explored with the purpose of reducing the cost and increase the performance of a fuel cell. These active approaches toward the reduction of Pt usage in PEMFC catalyst can generally be listed as:

1) Reducing Pt usage by alloying with other transition metals (either as bimetallic or ternary catalyst systems) and non-noble electrocatalysts.

2) Reducing the Pt loading in fuel cell electrodes by using ultra-low loading catalyst methods.

3) Reducing the electrocatalyst nanoparticle size and improving PPEMFC performance by optimization of operating parameters.

4) Improving electrocatalyst dispersion by using novel fabrication techniques (e.g. The National Research Council of Canada’s Institute for Fuel Cell Innovation (NRCIFCI), Reactive spray deposition technology (RSDT)).

5) Developing MEA fabrication methods to enable better catalyst dispersion and utilization.

6) Improving the performance of carbon-based electrocatalyst support and exploring novel non-carbonaceous electrocatalyst support materials by employing the concept of nanotechnology for increasing the surface area and other properties.

It is very difficult to keep or improve catalyst layer performance when the catalyst loading is reduced. In order to attain that, great effort must be put on Pt utilization and the enhancement of catalyst activity. Two different routes have been explored for enhancing the utilization of platinum catalyst in PEMFCs, either by reducing the particle size of catalyst or by uniforming distribution on the surface of support materials. The prominence of the support materials in fuel cell catalysis is well documented. Typically, the support provides a physical surface for dispersion of small metal particles, which is necessary for accomplishing high surface area. Carbon is commonly used for supporting nanosized electrocatalyst particles in low temperature fuel cells because of its large specific surface area, high electrical conductivity, and pore structures. In spite of the high surface area of the carbon black particle, it has two main problems. Carbon black persuades significant mass transfer limitations due to its dense structure (the pore size and pore distribution of carbon black affects the interaction between Nafion ionomer and the catalyst nanoparticles), resulting failure of launching three-phase boundaries among gas, electrolyte and electrocatalyst which lead to a very low Pt utilization. Additionally, carbon black is known to endure sever electrochemical oxidation, forming surface oxides which react finally to CO₂ (C + 2H₂O → CO₂ + 4H⁺ + 4e⁻) at the cathode of the fuel cell. The formation of CO₂ or corrosion can be accelerated at lower pH and higher potential, humidity and temperature so the life time of potential electrocatalyst can be reduced by poisoning and sintering. The standard potential for the electrochemical oxidation of carbon to carbon dioxide is 0.207 V vs. RHE at 25˚C [2]. During the start-up/shut-down cycle, the cathode potential of vehicle’s fuel cell because of slow oxygen reduction reaction (ORR) may increase to 1.2 - 1.5 V or even higher, therefore, under such PEMFC cathode operation conditions, carbon corrosion is not only thermodynamically viable due to the high potentials (>1.2 V) and high O₂ concentrations, but also kinetically enhanced by the elevated temperatures (50˚C - 90˚C). As carbon black (Vulcan XC-72) corrodes, noble metal nanoparticles (e.g., Pt) on carbon black will isolate from the electrode and possibly combine to larger particles, resulting in Pt surface area loss, which subsequently lowers the durability of PEM fuel cells [3-5]. Accordingly, corrosion of the support may affect the performance of the PEMFC. Therefore, prompted by the thermodynamic instability of carbon, tremendous efforts have been made to identify and develop alternative support materials that show good stability up to voltages of 1.4 - 1.6 V and improve durability and performance.

Nanostructure materials have attracted great interest in recent years because of their exceptional characterization such as high catalytic activity, mechanical and electrical properties, all of which are obtained by obstructing their dimensions. It is well known that the physical, electrochemical, and electronic properties of nanostructured materials are not the same as those of bulk materials. In terms of nanotechnology methodologies towards improving clean energy solution, numerous innovative nanomaterials as catalyst supports for PEMFCs have been actively investigated. In this context, the role of the nanostructure in carbon supports like carbon nanotubes (CNTs), carbon nanofibers (CNFs), carbon aerogels, and nano-plates of grapheme was much highlighted. The main purposes for using nanostructure-supported Pt catalysts are to reduce Pt loading through increasing the catalyst utilization, and improving the catalyst activity and total performance. The higher catalytic activity of Pt and Pt-based catalysts supported on CNTs and CNFs than that of the same catalysts supported on carbon blacks was ascribed to their unique structure and properties such as high surface area, good electronic conductivity and chemical stability [6-10]. Tests carried out in PEM fuel cell conditions indicated that these materials can be more durable and can outlast the lifetime of most widely used Vulcan XC-72.

This review will focus on some recent developments and investigations of nanostructure materials for application as support for preparation of electrocatalyst for MEAs purpose, excluding electrolyte membrane and other parts of fuel cell. The structure, dispersity and morphology of the supported catalysts, which are closely related to utilization and performance, are strongly influenced both by the support property and the method by which Pt nanoparticles can be deposited into the support. In this study, the development of low-temperature fuel cell catalysts in recent years is reviewed, primarily focusing on the two most active areas, i.e. support materials especially supports with tailored nanostructural, surface chemical and electronic properties by employing the concept of nanotechnology and nanoscience in designs and production techniques and Pt-based material deposition approaches.

Considering the recent move of a large variety of the public converging to CNTs, CNFs, and GNFs (graphene nanoplates) related synthesis techniques (also surface modification), we did not discuss the production methods of these materials in the present study, which would have been another useful option if one considers numerous research articles about synthesis and processing techniques of the nanostructure of these materials that have been published in this 10-years span. So where necessary, the reader will be addressed to more specific recent review papers and books. After some simple background information, the structure of each of these materials, the basic deposition techniques of Pt-based metals onto the support materials along with the surface functionalization of these nanostructures, characterization and other features of resulted supported electrocatalyst such as morphology, electrochemical activity as well as electrochemical surface area will be described. Next, it will be continued by describing fabrication processes of MEA fuel cells containing nanostructures. This will be followed by investigating the improvement of utilization of electrocatalyst and performance in single cell measurements. Finally, we give a summary of the stability/durability of the nanostructure-based fuel cell electrodes. Our focus will be on low-temperature fuel cells including only hydrogen and methanol fuelled polymer electrolyte membrane fuel cells (H-PEMFCs [hydrogen fuel cell] and DMFCs) direct methanol fuel cells. The term “PEMFC” has been used often to cover the circumstances where DMFC is also applicable unless specifically mentioned separately.

2. CNT/CNF-Supported Pt Nanoparticles: Characterization and Electrochemical Activity Performance as Anode and Cathode Catalysts for PEM Fuel Cell

The potentiality of using CNT or CNF for fuel cell use, usually as supports for Pt-based particles for PEMFC testing has already been investigated [11-13]. In comparison with the conventional Vulcan XC-72R carbon support which has an electronic conductivity of 4.0 S/cm and specific surface area of 237 m²/g [14], CNT and CNF have significantly higher electronic conductivities of 10⁴ and 10³ - 10⁴ S/cm, respectively [15,16] and extremely high specific surface areas of 200 - 900 m²/g [16]. Apart from that, Vulcan XC-72R has a large ratio of micropores which are smaller than 2 nm, while CNT and CNF have no micropores smaller than 2 nm [14]. For Vulcan XC-72R support, the Pt nanoparticles may sink into the micropores, which will reduce the number of three-phase boundary reactive sites, thus reducing the Pt utilization [17,18]. These potential benefits for fuel cell use that have been suggested include higher utilization of active metal due to the lack of smaller porosity and higher corrosion resistance due to the (theoretically) inert surfaces. The main reason for using CNTsand CNFssupported Pt catalysts is to reduce Pt loading through increasing the catalyst utilization, and improving the catalyst activity. Electrocatalysts supported on CNTs or CNFs as electrodes in either PEMFCs or direct methanol fuel cells have been examined extensively and the resulting supported catalysts gave better performance than those supported on commercial carbon black [19]. The attractive features of CNT-based CLs include improved thermal and charge transfer, and maximum exposure of the catalyst sites to the gas reactant through uniform support geometry and parallel alignment. Figure 1 shows

the structure of an aligned CNT-based MEA. A mass activity of 250 A/mg Pt at 0.6 μg Pt/cm² was obtained by Tang et al. [20]. It is well known that the activity of a catalyst depends considerably on the size of the Pt particles and their dispersion configuration over the support structures as well as the particle interactions with the support substrates. It has been found that the optimal dispersion pattern and Pt particle size can be obtained by using an appropriate preparation procedure on an ideal supporting material [21]. So far a large variety of methods like Impregnation, Electrochemical, Colloidal, Ion-exchange, Microwave Heated Polyol, Sputtering, Supercritical Fluid Method have been reported for the deposition of Pt on the surface of CNTs as a support in the literature. All these methods have been successful in yielding Pt nanoparticles but with wide particle size ranges, basically due to agglomeration or inefficient control on the growth of nuclei. However, colloidal process has been well recognized to produce uniform particles with excellent control on nuclei growth and morphology [22]. Different growth control mechanisms and strategies are used in each of the different deposition methods. For example, in the impregnation method, the size of the Pt nanoparticles is controlled by the structure of the support material which acts as the confining intermediate to restrict reaction, diffusion, and aggregation processes. In the colloidal method, the Pt size is controlled either by electrostatic interruption or the addition of a protecting agent, which will adhere onto the surface of Pt nanoparticles. For the ion-exchange method, the surface groups of the support material provide the anchorage sites for the Pt particles and control the dispersion and distribution of the Pt nanoparticles [23]. In this section, some examples of Pt deposition methods will be discussed. Since the pristine surface of CNTs is relatively inert it is impossible to have a high dispersion of Pt nanoparticles with controlled catalyst loading on the surface of support. So surface modification of carbon nanotubes commonly in order to make the surface more hydrophilic and improvement the catalyst support interaction towards an active fuel cell cathode catalyst is compulsory. Essentially, different covalent and non-covalent modification methods have been investigated to functionalize CNTs. Thus, functionalization to introduce surface oxygen groups such as hydroxyl, carboxyl carbonyl and sulfate (using either strong acids like HNO₃, H₂SO₄ or mild acids like citric acid) and other solution like, KMnO₄, K₂Cr₂O₇, KOH and H₂O₂ have been performed on the surface of CNTs. It is believed that these surface functional groups function as metal-anchoring sites to facilitate metal nuclei formation and electrocatalysts deposition [24]. The surface modification of CNTs could also be performed via an electrochemical method [25,26].

Characterization methods that were employed to analyze the CNT and CNF supported Pt catalysts included X-ray diffraction (XRD) to measure average Pt particle size, transmission electron microscopy (TEM) to determine Pt particle size and size distribution, thermal gravimetric analysis (TGA) to determine Pt metal loading and cyclic voltammetry (CV) to obtain information about their electrochemical active surface area (ECSA).

2.1. Deposition Methods of Pt Nanoparticles onto CNTs and Their Morphology

2.1.1. Impregnation Method

Due to its simplicity, impregnation is one of the most commonly used techniques for depositing Pt-based metal on the carbonacouse support for the preparation of the PEM fuel cell electrode. High-surface-area carbon nanotube can be impregnated with catalyst precursors (e.g.) by mixing the two in an aqueous solution. This is followed by a reduction step which is required to reduce the catalyst precursor to its metallic state. As reduction occurs after the impregnation step, the nature of the support plays a crucial role in controlling particle size. The reduction step can be chemical or electrochemical. Common liquid phase reducing agents are Na₂S₂O₃, sodium borohydride (NaBH₄), Na₄S₂O₅, N₂H₄, and formic acid or hydrazine. H₂ is the predominant gas phase reducing agent under higher temperatures. The impregnation technique is a comparatively green method to synthesize catalysts; reduction reactions occur either at low or room temperatures, minimizing energy consumption, and organic solvents are avoided by using aqueous media. Drawbacks of the impregnation technique are related to using liquid solutions as the processing medium. The easily agglomeration of particles in solution has been observed.

Matsumoto and co-workers [18] synthesized Pt nanoparticles supported on CNTs using two Pt precursors; H₂PtCl₆·6 (H₂O) or K₂PtCl₄. In their research, heated CNTs stirred with Pt precursor in ethylene glycol at 40˚C for 2 h. Then this solution was added into 50 ml of 14 N HNO₃ and filtered. A commercial carbon black supported Pt (Pt-CB) electrocatalyst was used as a control to compare the electrode’s performance. TEM images show 2 - 4 nm Pt nanoparticles dispersed on the CNT surfaces. Figure 2(a) shows the TEM image of the HPt-CNT. As can be seen the small dark particles assigned to Pt particles were well dispersed on the CNT surfaces in the TEM image, while Pt particles were deposited locally and agglomerated each other on the KPt-CNT in Figure 2(b).

Cui and coworkers [27] presented a simple and rapid synthesis method (denoted as modified impregnation method, MI) for PtRu/CNTs (MI) and PtRu/C (MI). In their experiments the surface-oxidized CNTs were suspended in deionized water and treated in an ultrasonic bath. Then H₂PtCl₆ and RuCl₃ solution were added dropwise under mechanically agitated conditions for 2 h. Then 5% (v/v) NH₃·H₂O solution was added to adjust PH of the mixture solution to above 12, therefore and Ru³⁺ were precipitated due to the formation of (NH₄)₂PtCl₆ and Ru(OH)₃ and the precipitate was adsorbed on CNTs surface. With dropwise addition of NaBH₄ solution, (NH₄)₂PtCl₆ and Ru(OH)₃ were reduced to Pt and Ru nanoparticles. Consequently, the solution was filtered and washed with hot deionized water until no Cl⁻ was perceived. Finally, the resulting catalysts were dried at 100˚C in a drying oven for 24 h and then transferred to desiccator. The PtRu catalysts supported on CNTs and carbon black were noted Pt-Ru/CNTs (MI) and Pt-Ru/C (MI), respectively. From the transmission electron microscopy (TEM) images it was shown that PtRu particles with small average size (2.7 nm) were uniformly dispersed on carbon supports. Cyclic voltammetry showed that the Pt-Ru/CNTs (MI) catalyst exhibited better methanol oxidation activities than Pt-Ru/C (MI) catalyst and commercial Pt-Ru/C (E-TEK) catalyst (Figure 3).

2.1.2. Electrochemical Method

Electrochemical deposition has been used to deposit Pt and Pt-based nanoparticles on a wide verity of substrates including carbon nanotubes [24] selectively, at desired locations in the substrate through easily controlling the nucleation and growth of the metal nanoparticles. Pulse current, direct current and cyclic voltammetry are utilized to deposit Pt nanoparticles using the electrochemical methods. There are five stages to electrochemical deposition of metals: 1) transport of metal ions in solution to the support surface, 2) electron transfer, 3) formation of metal ad-atoms via adsorption, 4) nucleation and growth, two-or three-dimensional, of metal particles, and 5) growth of the

three dimensional bulk metal phase. With the electrochemical method, the Pt loading can be reduced ten-fold without sacrificing the cell performance in comparison to the conventional deposition technique [28]. Electrodeposition of Pt nanoparticles on CNTs has recently been reported by several groups. With the vigilant choice of substrate, deposition potential, and growth time, it is possible to deposit nanoparticle with a high degree of monodispersity and narrow particle size distribution [29-31]. Direct potential electrodeposition has been used for the formation of metal nanoparticles that are supported on CNTs. For example, Day and co-workers [32] studied the nucleation and growth mechanisms of the electrodeposition of metals such as Ag and Pt onto CNT s. In their studyPt electrodeposition was performed by stepping the potential from 0 V to −0.4 V (vs Ag/AgCl) for a period of 30 s backing to an open circuit. According the authors this procedure allows the density and size of metal nanoparticles to be controlled by cautious choice of the applied potential and deposition time.

Guo and coworkers [33] described a novel method (potential-step method) for preparation of Pt nanoparticles through molecule-level design, similar to the procedure employed to disperse Ptparticles on other carbon substrates such as highly oriented pyrolytic graphite (HOPG) in the literature [34]. Their TEM images (Figure 4) reveled that the surface of MWNTs was uniformly covered with a certain amount of platinum nanoparticles with an average size of 1 - 3 nm [33].

2.1.3. Colloidal Method

In recent years, there has been considerable interest in the development of colloidal methods to prepare Pt catalysts supported on CNTs with a narrow particle size distribution. Usually, this method comprises the preparation of a platinum metal colloid, followed by adsorption on the support, or in the formation of a Pt oxide colloid, followed by coincident reduction and adsorption, or adsorption followed by chemical reduction. In the colloidal method, the size of the Pt nanoparticles is largely controlled or stabilized by the protecting agents, such as ligands, surfactants or polymers [35]. For example, Li and co-worker [36] used surfactant 3-(N,N-dimethyldodecylammonio) propanesulfonate (SB12) as a stabilizer to prepare Pt nanoparticles supported on the functionalized CNTs by methanol reduction of H₂PtCl₆ [36,37]. Uniform Pt nanoparticles of an average size of 2.2 nm were dispersed on the external walls of the CNTs [36].

Though the colloidal method can provide a narrow size distribution of metal nanoparticles, the foremostproblem is the presence of a protecting agent, which can be difficult to remove once the particles are adsorbed onto the

support and may impede the catalytic performance of the nanoparticles. One further disadvantage is that the colloidal particles are prepared at high dilution (typically milimolar concentrations-for example, 0.2 g Pt/L), which is not suitable for its scale-up. It was recently reported that the colloidal Pt nanoparticles may be protected by glycol, which works as both a solvent and the protecting agent [14,38]. For example, Kongkanand and co-workers [38] used ethylene glycol as a reducing agent to prepare welldispersed Pt catalysts supported by polymer-wrapped CNTs. They used polymers such as polystyrene sulfonate wrapping around the individual CNTs by means of a thermodynamic driving force in order to reduce the hydrophobic interface between the tubes and polar solvent (i.e. N,N-dimethylformamide). The Pt nanoparticles were dispersed on these polymer-wrapped CNTs by the colloidal method. The Pt colloidal solution was synthesized using the reduction of by ethylene glycol at a pH of 12.5. Cyclic voltammograms measurements reveled that these polymer-wrapped CNTs with Pt nanoparticles deposited on them have very high electrochemically active surface areas as shown in Figure 5. The increase in utilization efficiencies for Pt catalysts on these CNT supports can be ascribed to the increased surface areas and also the well-dispersed nature of the carbon support and catalyst [38].

2.1.4. Ion-Exchange Method

An ion-exchange method is an operative technique for depositing Pt metal particles on the CNTs without using protective colloidal agents, reduction agents or precursor complexes which are not easy to decompose [26,39,40]. Wang and co-workers used an in situ ion-exchange method for preparing the electrode composed of Pt particles on the CNTs [40]. In their study, the CNT electrode was electrochemically functionalized to introduce the carboxyl

functional groups on the surface of CNTs. Then this electrode was immersed in a solution of the platinum cationcomplex salt for 48 h. As a result of this immersion, the hydrogen ions of the functional group on the surface of the CNT electrode undergo an ion-exchange process with the platinum cation complex. The immersed CNT electrode was then filtrated and washed sufficiently with distilled water. Reduction of the adsorbed platinum complex to its metallic state was carried out by the treatment with hydrogen gas at 190˚C. They observed the Pt nanoparticles to be highly dispersed on the CNTs and their dispersion was much superior to the catalyst prepared by the borohydride method [40].

Shao et al. [26] prepared Multi-walled carbon nanotubes based Pt electrodes (Pt/MWNTs) for oxygen reduction reaction in PEMFCsby depositing Pt nanoparticles on electrochemically functionalized MWNTs with in situ ion exchange method. Figure 6 shows representative SEM images of the MWNT electrodes before (A) and after (B) Pt loading by in situ ion exchange. It can be seen from the images Multi-walled carbon nanotubes

have around 20 nm in diameters. Pt nanoparticles are highly dispersed on carbon nanotubes. They reported that the electrochemical surface area of the ion-exchanged electrode and the conventional electrode was 688.3 and 542.6 cm²/mg Pt, respectively. Also the Pt utilization, which is the ratio of the electrochemical surface area and the chemical surface area, was greatly enhanced on the ion-exchanged electrode (98.2%), as compared with 48.4% for the conventional one. Moreover the activity of the ion-exchanged electrode to ORR was higher than the conventional one.

In other study that was carried out by the Wang and et al. [41] highly dispersed platinum supported on single-walled carbon nanotubes (SWNTs) as electocatalyst was prepared by ion exchange method. The homemade Pt/SWNTs underwent a repetition of ion exchange and reduction process in order to achieve an increase of the metal loading. For comparison, the similar loading of Pt catalyst supported on carbon nanotubes was prepared by borohydride reduction method. The particle morphology, size, and dispersion of Pt nanoparticles deposited on the carbon nanotubes were characterized by transmission electron microscopy it is reported that the particle sizes of Pt catalysts prepared by ion exchange method was remarkably uniform and the dispersions of Pt nanoparticles on the SWNT was good. The average sizes of the particles of all resulted catalysts calculated from the XRD data are consistent with the results by TEM micrographs. As can be seen from XRD (Figure 7) observations, the small platinum particles with a well-defined particle size distribution were deposited on the carbon support by ion exchange method. The formation of small size of Pt nanoparticles attributed to the “anchor effect” of the ion exchange reaction. Compared with the Pt/ SWNTs catalyst prepared by borohydride method, higher Pt utilization was attained on the SWNTs by ion exchange method. Furthermore, in comparison to the ETEK 20 wt% Pt/C catalyst with the support of carbon black, the results from electrochemical measurement (Figure 8) indicated that the Pt/SWNTs prepared by ion exchange method displayed a higher catalytic activity for methanol oxidation and higher Pt utilization, while no substantial increasing in the catalytic activity of the Pt/SWNTs catalyst obtained by borohydride method.

2.1.5. Microwave Heated Polyol Method

Microwave irradiation through dielectric heating loss is fast, simple, uniform, energy efficient and has been widely used in materials synthesis [42]. The polyol process, in which an ethylene glycol solution of the metal precursor salt is slowly heated to produce colloidal metal, has recently been extended to produce metal nanoparticles supported on various substrates. Recently there have been more reported achievements in preparing high purity

Pt or Pt-based nanoparticles supported on CNTs by a microwave-assisted polyol process in literature [43-47]. The polyol process is normally used in preparing colloidal metal particles [48]. In this process, a polyol (most commonly ethylene glycol) solution containing the metal precursor salts is refluxed at 393 - 443 K, where the polyol decomposes homogeneously to release the reducing agent for metal ion reduction [49]. A support material may be optionally present to capture the depositing metal particles [50]. For example, Lordi et al. [51] used the polyol process to prepare Pt supported on single-walled carbon nanotubes for selective hydrogenation. Li et al. [52] used an analogous technique to prepare Pt/CNTs catalyst for fuel cell applications. Conventional heating, with its intrinsic heterogeneous temperature distribution, was used to thermally activate the polyol in these instances. On the other hand, as a rapid, uniform and effective heating method, microwave heating has been used as an attractive alternative for nanosize materials synthesis. Microwave radiations can increase the reaction kinetics by one to two orders of magnitude and promote the formation of uniform metal nanoparticles.

The pH value of the Pt precursor solution is an important factor in controlling the Pt particle size and distribution by the polyol process. The effect of the pH on the Pt particles’ size and distribution were investigated by Li et al. [47]. In their experiments, the pH of the solution was adjusted to the desired value (3.4 - 9.2) by the addition of a KOH solution. Figure 9 shows TEM images of Pt/ CNTs prepared by the microwave polyol process using different values of pH. At a lower pH range (pH 3.6 - 5.8), the Pt nanoparticles are agglomerated and not as well dispersed on the CNT surfaces. In contrast in the pH range of 7.4 - 9.2 less agglomeration has been formed and the Pt nanoparticles dispersed more uniformly on the CNT surface. The direct evidence of the pH effect on the size of Pt was given by TEM images of Pt/CNTs prepared by microwavepolyol heating. As it is shown pH value in the solution plays an important role in controlling the Pt particle size and distribution.

2.1.6. Sputtering Techniques

Sputtering techniques as efficient deposition method can produce anextremely uniform particle distribution and the fabricating process is quite simple. The particle size is determined by controlling the sputtering time and current. Recently, sputtering deposition techniques were used for producing thin platinum films on the surface of CNTs. According to Chen et al. [53], the sputtering deposition method could generate highly uniformed Pt nanoparticles on CNTs compared to those produced by the impregnation deposition method. In their research, the CNTs were synthesized directly on the carbon cloth by a bias-assisted microwave plasma enhanced CVD method and the Pt nanoparticles were subsequently deposited on the CNTs. Wang et al. [54] deposited Pt-Ru electrocatalysts on the directly grown on a carbon cloth(DGCNTs) using sputtering process (Pt-Ru/DGCNT), which were acted as an anode in the DMFC. The diameters of Pt-Ru nanoparticles deposited on the DGCNTs was ca. 3.54 nm, which was the high-degree alloy by X-ray diffraction analysis. The HRSEM and TEM images of Pt-Ru/DGCNT are shown in Figures 10(a) and (b), respectively. From Figure 10(a), the Pt-Ru electrocatalysts are successively deposited on the DGCNTs. It is clear that the electrocatalysts are highly dispersed on the DGCNT, as shown in Figure 10(b). The diameter of electrocatalyst is around 3.5 nm. The electrochemical property was measured by cyclic voltammogram. Comparing of the Pt-Ru/ DGCNT and the conventional Pt/C, Pt-Ru/DGCNT shows a high-

er methanol-oxidation activity.

The advance variation of the sputtering deposition methods found in the literature is to deposit Pt nanoparticles on nitrogen-containing CNTs (N-CNTs) for microDMFC applications [55]. A lot of efforts have been dedicated to the NCNTs [56,57]. It has been reported that the N-CNTs have many advantages over general CNTs including the enhancement of interaction between the nanotube surface and the deposited Pt nanoparticles [58].

While the sputtering technique works well for the direct deposition method, the main drawback is the durability. In most cases the deposition has not strongly adherence to the substrate and under the variable conditions of load and temperature there is a greater probability that the deposits will sinter or dissolute. As discussed, the sputtering technique is a very useful method to achieve an ultra-low Pt loading in the catalyst layer. A unique advantage of the sputtering technique is that the catalyst utilization is extremely high. In addition, the sputtering method accelerates the preparation of a nano-scale Pt catalyst layer with a defined thickness, which could streamline fuel cell water/thermal management and eliminate mass transfer loss. Furthermore, the sputtering process allows the deposition of Pt on various substrates, such as the GDL, membrane, and other supports. However, although the sputtering CLs showed remarkably higher mass activity than the conventional layer, the validated output power density is still low. For example, Tang et al. [59] sputter deposited nanosized Pt nanodots (2 - 3 nm) as catalyst on CNT prepared in-situ on carbon paper. The system acting as a combined gas diffusion layer (GDL) and catalyst layer displayed a high maximum power density of 595 mW/ cm² using a catalyst loading of 0.04 mg/cm² Pt on the cathode side. This was significantly higher than the Pt/Vulcan XC 72R-based electrode (435 mW/cm²) with equal Pt loadings and that of the reference electrode with sputtered Pt on CNT/CB blend layer (530 mW/cm).

2.1.7. Supercritical Fluids

The use of supercritical fluids (SCF) as “green” solvents for the synthesis and processing of nanomaterials has gained considerable interest in recent years. Supercritical carbon dioxide (scCO₂) is particularly attractive for electrocatalysts processing as it is accessible under moderate conditions (T_c =31˚C and P_c = 7.38 MPa), abundant, low-priced, non-flammable, non-toxic, and environmentally benign [60-62]. This process involves the dissolution of a metallic precursor in a supercritical fluid and the introduction of a porous support to the solution. After adsorption of the precursor on the support, the metallic precursor is converted to its metal form by chemical or thermal reduction. Using a SCF as the processing medium for synthesis of electrocatalysts has many advantages which are directly related to the special properties of the SCFs. Supercritical carbon dioxide (scCO₂) allows reactive components to penetrate inside the porous materials themselves, partitioning into the inner regions of the porous supports. This promising catalyst preparation technique results in small particle sizes and homogeneous dispersions. An additional advantage of this technique is the ability to thermodynamically control the metal loading [63].

Lin et al. [64] used a supercritical fluids (SCFs) method as a rapid, direct, and clean approach to prepare Pt/CNT catalyst for DMFCs. Bimetallic platinum-based electrocatalyst nanoparticles supported on carbon nanotubes for direct methanol fuel cells have also been produced using scCO₂ [65,66]. Bayrakceken et al. [67] used the scCO₂ deposition method to disperse Pt-based electrocatalysts on various carbon supports such as MWCNTs, Vulcan XC 72R (VXR) and black pearl 2000 (BP2000). Commercially available Pt/C-ETEK was used for comparing the synthesized catalysts. XRD and HRTEM results demonstrated that the scCO₂ deposition technique enables a high surface area metal phase to be deposited, with the size of the Pt particles ranging from 1 to 2 nm. HRTEM images obtained from the synthesized and commercial catalysts are presented in Figure 11 for Pt/ MWCNT (a), (b); Pt/VXR (c), (d); Pt/BP2000 (e), (f) and ETEK Pt/C (g)-(j).

The electrochemical surface areas (ESAs) of the prepared electrocatalysts were compared to the surface areas of commercial ETEK Pt/C (10 wt% Pt) and Tanaka Pt/C (46.5 wt% Pt) catalysts (Figure 12). The CV data indicated that the ESAs of the prepared Pt/VXR and Pt/ MWCNT catalysts was about three times larger than that of the commercial ETEK catalyst for same (10 wt% Pt) loadings. Oxygen reduction activity was investigated by hydrodynamic voltammetry. From the slope of KouteckyLevich plots, the average number of electrons transferred in the oxygen reduction reaction (ORR) was 3.5, 3.6 and 3.7 for Pt/BP2000, Pt/VXR and Pt/MWCNT, correspondingly, which indicated almost complete reduction of oxygen to water.

2.2. Fabrication and Performance of Carbon Nanotubes Based Membrane Electrode Assembly

The conventional process for constructing CNT-based catalyst layers in PEMFC is to disperse CNTs in a binder (such as Nafion) to form a slurry, and then coat on a carbon cloth (gas diffusion layer) [68]. A common problem in this process is that the addition of the binder has a tendency to isolate carbon nanotubes in the electrocatalyst layer, resulting in poor electron transport and reduces the Pt active surface. To overcome this problem, many efforts have recently been made. For instance, researchers tried to grow or filter carbon nanotubes directly onto a carbon cloth and then subsequently depositing Pt selectively on the carbon nanotubes [69-72]. In these efforts, the contact resistance between the two layers is minimized and the utilization rate of the noble metal catalyst can be increased by abolishing the separation of the carbon particles from the electrode support. A thin film of CNTs loaded with nanosized Pt prepared by filtration method was applied as a cathode catalyst layer by Haddon et al. [73], which showed high performance at ultra low Pt loading due to the efficient proton transport in a small thickness.

Li and co-workers [72] reported a simple preparation procedure (filtration method) for the fabricating MEA using CNTs films as a support for PEM fuel cells. They deposited the Pt nanoparticles on the CNTs before the film formation and then used filtration as a means to form an oriented film of Pt/CNTs on the Nafion membrane. Pt particles were dispersed on the CNT’s surface through a chemical reduction of H₂PtCl₆ by ethylene glycol. The Pt/CNT architecture was then filtered through 0.2-µmpore hydrophilic nylon filter paper. The dispersed film was then transferred onto a Nafion membrane by pressing the CNT-coated side of the filter onto the Nafion membrane. Finally, the oriented Pt/CNT-film-coated Nafion membrane, and a gas diffusion layer were hot pressed to obtain an MEA. A higher performance was gained with oriented Pt/CNT-based MEA than Pt/C and non-oriented Pt/CNT-based MEAs.

Recently, a new style for composite electrode fabrication involving the growth of CNTs directly on the carbon

fibers of a fuel cell baking with subsequent deposition of Pt nanoparticles was established [69,74,75] Randomly oriented single or multiwalled carbon nanotubes can be easily converted into highly conducting thin porous networks appropriate for gas diffusion electrodes in fuel cells. High dispersion and electrocatalytic properties of Pt nanoparticles on SWNT bundles [76,77] Partially aligned MWNTs grown on commercial carbon fibres [78] combine the well known properties of conventional carbon electrodes with the high specific surface area of MWNTs (up to 1000 m²/g [79]). Totally aligned nanocarbon materials can be prepared by the vertical growth of MWNTs on thermally insulating substrates allowing well-defined three-dimensional distribution of properties. Aligned MWNTs are discussed as gas diffusion layers [80] and catalyst support by several authors [81,82].

2.3. Test in a Single Fuel Cell

From the practical point of view, the single cell test is the ultimate evaluation criterion for new electrocatalysts materials. A lot of  experiments have been performed on the single cell in real fuel cell condition in order to evaluation the fuel cell performance of Pt nanoparticles supported on CNTs at the cathode of H₂/O₂ fuel cells or direct methanol fuel cells by many authors [14,18,52,69,72,75,83-88]. Without exceptional it has been shown that the performance of a CNT-based MEA is better than that of conventional Pt/C MEA. For example, Matsumoto and coworkers [18] conducted the fuel cell tests of the MEA with CNTs used as a support for Pt catalysts with different two Pt precursors; H₂PtCl₆·6(H₂O) or K₂PtCl₄ at the cathode. The performance of the fuel cell for the KPt-CNT, HPt-CNT, and Pt-CB electrodes are shown in Figure 13. The performance of the Pt supported CNT electrodes is comparable to the Pt-CB electrodes in Figure 13(a), and

the KPt-CNT electrodes showed the highest voltages below 500 mA/cm² in Figure 13(b). They suggest that the higher voltages of the KPt-CNT related to the formation of more effective triple-phase boundaries on the KPt-CNT electrodes than the other two electrodes.

In another study Matsumoto et al. [88] found that a 12 wt% Pt deposited CNT electrode provides 10% higher voltage than 29 wt% Pt deposited on Vulcan and reduces Pt usage by 60% in PEMFCs when using hydrogen and oxygen. With respect to comparing the performance of Pt-Ru/CNTs (MI) with Pt-Ru/C (E-TEK) catalyst that were conducted by Cui and coworkers [27] the single cells with Pt-Ru/CNTs (MI) catalyst exhibited a power density of 61 mW/cm², about 27% higher than those single cells with commercial Pt-Ru/C (E-TEK) catalyst (Figure 14).

Saha and co-workers [89] developed a synthetic method to deposit Pt nanoparticles on CNTs that were grown on commercially-used carbon paper by the reduction of the Pt precursor with glacial acetic acid. Figure 15 shows the single cell performance for the MEA made with a CNTbased electrode as the cathode and standard E-TEK electrode as the anode for H₂/O₂ fuel cells. Compared with the standard Pt/C electrode, the CNT-based electrode exhibited higher single-cell performance in a H₂/O₂fuel cell. According to Figure 16 at a constant cell voltage of 0.6 V, the current density of the Pt/CNT/carbon paper composite is 1.58 A/cm² which is 27% higher than the standard E-TEK electrode (1.16 A/cm²). The corresponding power densities normalized on the basis of Pt loading were 2.19 W/mgPt (Pt/CNT/carbon paper composite) and 1.42 W/mgPt (standard E-TEK electrode) at 0.6 V, showing a significant increase in the power density of about 0.77 W/mgPt for the Pt/CNT/carbon paper composite. It is clearly obvious that the current density at 900 mV for the Pt/CNT/carbon paper composite is meaningfully higher than the E-TEK electrode (approximately 66%), which attributed to the better dispersion of Pt nanoparticles on the CNT’s surface and to the 3D structure of CNT-based electrodes.

Carmo and co-workers [90] conducted polarization curve measurement with noble metal catalysts supported on carbon nanotubes (MWNT and SWNT) and also on a high surface area carbon powder Vulcan XC-72, for proton exchange membrane fuel cells fed with hydrogen contaminated with CO and also for the direct methanol fuel cell. A high performance was obtained with PtRu supported on nanotubes for H₂ + 100 ppm CO, although it was analogous to that presented by PtRu on Vulcan XC-72 with an overpotential of 100 mV at 1 A/cm. Results for the DMFC showed power densities more than 100 mW/cm² at 90˚C and 0.3 MPa and the activity of the anodes followed the sequence: PtRu/MWNT > PtRu/ Vulcan XC-72 > PtRu/SWNT.

3. Methods for Depositing Pt Catalysts on Carbon Nanofibers (Pt/CNFs) and Performance of the CNFs Supported Electrode in PEM Fuel Cell

In recent years, Carbon nanofibers (CNFs) have attracted interest as electrocatalyst support for PEM fuel cells due to their excellent mechanical, electrical and thermal properties. Within several supports, carbon nanofibers blend

two properties that rarely coexist in a material: a high mesoporosity and a high electrical conductivity, due to their particular crystalline structure. Moreover, the ratio of edge atoms to basal atoms of CNFs is controllable, which provides a means to adjust the deposition of and the interaction with the metal [91-93]. Bessel et al. [94] utilized Pt/CNFs and Pt/(Vulcan XC-72) as the catalysts for methanol oxidation, and found that the platelet CNF (p-CNF) and tubular CNF (t-CNF) supports were superior to fish-bone CNF (f-CNF) or Vulcan XC-72 in terms of the activities. They also showed that 5 wt% Pt supported on p-CNT or t-CNT would be as active as 25 wt% Pt supported on f-CNF or XC-72. However, Ismagilov et al. [94] reported that the performance of the catalysts supported on p-CNF f-CNF was lower than those supported on Vulcan XC-72 and amorphous supermicroporous carbons (ASC). Tsuji and coworkers [95] deposited PtRu alloy nanoparticles supported on three types of carbon nanofibers (CNFs); platelet, herringbone, and tubular ones by using a microwave assisted-polyol method. The dependence of particles sizes and electrochemical properties on the structures of CNFs was examined. The methanol fuel cell activities of PtRu/CNFs measured at 60˚C were 1.7 - 3.0 times higher than that of a standard PtRu (29 wt%, Ru/Pt atomic ratio, 0.92) catalyst loaded on carbon black (Vulcan XC72R) support. Based on results the greatest electrocatalytic activity was obtained for the platelet CNF, which is characterized by its edge surface and high graphitization degree.

Since the publication of the works of Bessel et al. [93] and Steigerwalt et al. [96,97] several studies have shown the advantageous use of CNFs for the electrooxidation of methanol, the anodic reaction of direct methanol fuel cells (DMFCs) [98-100]. However, few works have dealt with the performance of CNF supported platinum for the oxygen reduction reaction (ORR) in a direct methanol fuel cell system. Ismagilov et al. [94] found an inferior ORR performance of platelet and parallel CNFs as support for Pt by comparison with the most used commercial carbon black, Vulcan XC-72R. Instead, Yang et al. [101] found a superior performance for the oxygen reduction when supporting Pd on CNFs. The irregularity within these results could be ascribed to the differences among the properties of different CNFs as well as the different deposition methods employed. Most of the deposition methods that were described for CNTs supported electrocatalyst previously are used for deposition of Pt and Pt-based metals on CNFs.

Sebastian and coworkers [102] prepared Pt catalyst supported on CNFs by different deposition methods and the obtained catalysts have been studied by different chemical techniques. In the first method, named sodium borohydride method (SBM), an aqueous solution of sodium borohydride (NaBH₄) was used as reduction agent. Firstly the precursor solution (an aqueous solution of H₂PtCl₆ 3.2 mmol/ L) slowly added to a dispersion of carbon in ultrapure water under sonication and then the pH of the dispersion is adjusted to 5.0. After that, the sodium borohydride aqueous solution (26.5 mmol /L) is dropwise added, keeping temperature under 18˚C and in the presence of sonication. Subsequently, the catalyst is filtered and thoroughly washed with ultrapure water, and then dried overnight at 60˚C. The second method, named formic acid method (FAM), consisted on the use of an aqueous solution of formic acid as reduction agent. Firstly, the carbon is dispersed in a 2 M formic acid aqueous solution, prepared from high purity reagents, and then heated up to 80˚C. Then an aqueous solution of H₂PtCl₆ 4.2 mmol/L was slowly added to the dispersion under continuous stirring at 80˚C ± 1˚C. Consequently like the first method, the catalyst was filtered and carefully washed with ultrapure water, and then dried overnight at 60˚C. Platinum has then been supported on functionalized carbon nanofibers by two different procedures, obtaining Pt crystal sizes between 5.4 nm and 8.1 nm and a complete reduction of the metallic phase. Two of the catalysts obtained by SB method presented higher current densities than those obtained for commercial catalysts, ranging from 32 (CNF LT-F) to 53 mA/cm² (CNF HT-F), whereas the Pt/Vulcan-ETEK presented 28 mA/cm². Samples obtained by FA method show, on the other hand, stationary current densities, achieving similar values that commercial catalyst when supporting Pt on the most conductive carbon nanofibers: CNF HT-F.SB method gave the impression to be better than FA method with respect to electrochemical activity, CO tolerance and performance whereas FA method offerings better stability.

The geometric effect of graphite nanofibers (GNFs) as a support for PtRu electrocatalysts on the oxidation of methanol for direct methanol fuel cells was studied using X-ray diffraction, field emission transmission electron microscopy (FETEM) and electrochemical measurements [103]. A high loading of 60 wt% PtRu catalystwas dispersed on GNFs. Further, the shape of the supported metal particles was affected by interactions with the GNFs. Electrochemical analysis indicated that GNFsupported PtRu catalysts resulted in an increased catalytic activity of about 100% over that of Vulcan XC-72 supported catalysts. FETEM data indicate that the enhanced activities result from a geometric modification of the catalyst particles by specific interactions between the GNFs and the supported PtRu nanoparticles.

Sebastian et al. [104] also studied the influence of carbon nanofiber support properties for the oxygen reduction reaction (ORR) in proton conducting electrolytebased direct methanol fuel cells. The synthesis and deposition of platinum nanoparticles was carried out by the microemulsion method, using a commercial surfactant (Brij_30, Sigma Aldrich), n-heptane as the non-polar phase and 2-propanol as co-surfactant. Briefly, the surfactant and n-heptane were blended and agitated. The platinum precursor (H₂PtCl₆) dissolved in aqueous solution (0.008 M) was then dropwise added to the mixture. Subsequently, 2-propanol was added until a visually transparent mixture was observed, demonstrating the formation of the microemulsion. After 4 h of stirring, NaBH₄ was gently added in 0.1 M aqueous solution to the microemulsion under continuous stirring. The suspension was stirred overnight and then it was slowly added to a suspension of carbon in ethanol under sonication and continuously stirred for 16 h. The quantities were adjusted for a Pt loading of 40 wt%. The catalyst was then carefully washed with ethanol and water and finally dried overnight at 70˚C. Half-cell studies determined that the ORR activity is enhanced when using a CNF with improved graphitization.

Knupp and coworkers [10] investigated the effects of water content, and metal loading on the average Pt particle size and electrochemically active surface area (ECSA) using conventional refluxing and microwave irradiation techniques on the deposition of CNF supported Pt. Optimization of deposition conditions leads to higher ECSA than seen in a commercially available carbon black supported catalyst.

Tang and co-workers [98] employed an electrochemical method for the deposition of Pt nanoparticles on graphitic carbon nanofibers (GCNFs). In their efforts, Pt nanoparticles were electrodeposited on a GCNF/graphite working electrode from an acidic solution of H₂PtCl₆ by cyclic voltammetry under the condition of a deposition potential of +0.1 - −0.25 V vs. SCE and a sweep rate of 15 mV/s. Before the Pt particle deposition, the GCNFs/ graphite electrode was pretreated in 30 wt% HNO₃ aqueous solution for 40 min to remove the metallic nickel catalyst. To further increase the electrochemical activity of the surface of the GCNFs in the water solution, the GCNFs/graphite electrode was then cycled in the range of −0.15 - +1.3 V at a sweep rate of 50 mV/s in 0.5 M H₂SO₄ for 20 cycles. They observed that the Pt nanoparticles are uniformly dispersed on the whole surface of the GCNFs with a diameter of about 40 - 50 nm, which is about 2 times smaller than that of the Pt/graphite electrode (100 nm).

In another investigation Kim and coworkers [105] employed electrochemical deposition method on Carbon nanofibers (CNFs) web supported platinium nanoparticles at sweep times of 5, 10, 20 and 40. In their experiments the CNF webs were attached onto glassy carbon substrate with the aid of 0.1% Naflon solution. A 10 mM hexachloroplatinic acid (H₂PtCl₆) was dissolved in 0.5 M HCl aqueous solution. A potential was swept from −0.7 to −0.2 V (versus Ag/AgCl) with a scan rate of 20 mV/s at four mentioned sweep for the deposition of Pt nanocrystalline particles over CNF webs. Particle size and loading level (wt%) of Pt were found to increase from 10.7 to 18.6 nm and 3.3 to 7.0 6%, respectively with the increase of sweep times (Figure 17). The results of elec-

trocatalytic activity of the nano Pt/CNF composite which was demonstrated by linear voltammetrty, cyclic voltammetry and impedance spectra showed increment in catalytic activity with the increase of sweep time.

Oh et al. [106] examined the effect of chemical oxidation of carbon nanofibers (CNFs) on the electrochemical carbon corrosion in polymer electrolyte membrane (PEM) fuel cells. In their experiments with increasing time of chemical oxidation treatment using an acidic solution, more oxygen functional groups are formed on the surface of CNF resulting in an increasingly hydrophilic carbon surface. This effect contributed to improvements in Pt loading and the distribution of Pt particles on carbon supports. However, the chemical oxidation treatment is found to accelerate electrochemical carbon corrosion. The oxygen functional group and the hydrophilic nature of CNFs after chemical oxidation treatment are believed to inspire the formation of CO₂. From the observed results, the authors concluded that the chemical oxidation of CNFs is beneficial for catalyst loading and distribution. On the other hand, however, it reduces the durability of the PEM fuel cells caused by the electrochemical carbon corrosion.

Steigerwalt et al. [96] reported the preparation and characterization of a Pt-Ru/GCNF nanocomposite where the GCNF support has the “herringbone” atomic structure. Comparative testing of this nanocomposite and unsupported Pt-Ru colloid as anode catalysts in a DMFC revealed 50% higher performance for the Pt-Ru/GCNF nanocomposite. Park et al. [107] reported the novel results regarding the effects of electrospun carbon nanofibers (e-CNF) as a catalyst support by comparison with the commercial Vulcan XC-72R. The e-CNF was synthesized by stabilizing and carbonizing the electrospun PAN-based fibers. The e-CNF showed an average diameter of 250 nm with a rough surface and was partially aligned along the winding direction of the drum winder. The characteristic morphology was fundamentally dependent on the shape of the carbon materials. The average pore size of the e-CNF was 2.36 nm, while that of the XC-72R was 10.92 nm as shown in Figure 18. The morphology of e-CNF was developed by shallow pores with rough surfaces due to the effects of electrospinning and carbonization, while that of the XC-72R was largely developed by mesopores rather than micropores due to the granular shape.

Moreover for performance optimization purpose tubular carbon nanofibers with an average diameter of 150 nm was investigated as a possible material for the electrodes preparation for polymer electrolyte membrane fuel cells [108]. Well-dispersed Platinum particles with an average crystallite size of 4.6 nm are deposited on sur-

face-oxidised fibers to be used as a catalyst support with an impragnation method. The carbon nanofiber-based electrodes are prepared by a sedimentation method without the use of organic solvents. This method allowed an exact setting of the fiber and binder content and the catalyst loading. The electrodes were optimized by varying the thickness of the gas diffusion layer and its binder content as well as the thickness of the active layer. These optimized electrodes showed a considerably better performance comparing to carbon black based electrodes with the same catalyst loading prepared by a spraying process using the same type and amount of electrolyte in the membrane electrode assembly. By reducing the Platinum content from 0.7 to 0.2 mg/cm², catalyst utilization is considerably increased.

Figure 19 exemplifies the CVs of an MEA with CNFbased electrodes.