Integrated platinum carbon nanotube based electrocatalyst for high efficiency proton exchange membrane fuel cells 1

1

Chapter 1 Introduction

1.1 Background
Fuel cells are deemed as one of the most important alternative power sources
utilizing sustainable energy to replace conventional combustion generators. A fuel
cell is an energy conversion device that generates electricity directly from an
electrochemical reaction, in which oxygen (or air) and fuels (e.g. hydrogen) combine
to form water [1, 2]. Fuel cells differ from ordinary batteries in that they are able to
continuously produce electricity as long as fuels are supplied. Moreover, fuel cells
convert fuels directly into electricity via an electrochemical process that does not
require fuel combustion. Therefore fuel cells are intrinsically more efficient and
environment friendly than combustion engines [1, 2].

1.2 Main Types of Fuel Cells
In the last two decades of the 20
th
Century, fuel cell technologies have achieved
significant breakthrough in their applications when the world is facing the shortage of
fossil fuels, coal and oil [2]. Owing to the tremendous research effort on fuel cell
technologies, several types of fuel cells have been developed for a variety of
applications to meet future energy demand, including transportation, stationary, and
portable electronic devices.

In general, fuel cells can be categorized into five main types according to the
specific electrolyte used. The five main types of fuel cells are: proton exchange
membrane fuel cells (PEMFC), alkaline fuel cells (AFC), phosphoric acid fuel cells
(PAFC), molten carbonate fuel cells (MCFC) and solid oxide fuel cells (SOFC) [1, 2].
2


The different characteristics of their respective electrolytes used in the cell
configuration determine the particular materials and fuels required for them, as well
as their unique properties and applications. Table 1.1 demonstrates the main types of
fuel cells and their characteristics and applications [1]. Among the five types of fuel
cells, PEMFCs have the most promising potential for domestic uses and other
applications in small-scale power generations due to their relatively low operating
temperature and compact structure [1, 2].

Fuel Cell
Type
Electrolyte
Operating
Temperatu
re
Fuel
Electric
Efficiency
Applications
Proton
Exchange
Membrane
FC
(PEMFC)
Solid
polymer
50−100 °C Pure H
2
35−45%
Vehicles and
portable

applications,
lower power
station.
Alkaline
FC (AFC)
KOH 60−120 °C Pure H
2
35−55%
Used in
military &
space
vehicles.
Phosphoric
Acid FC
(PAFC)
Phosphoric
acid
~220 °C Pure H
2
~40%
Suitable for
200 kW
power
stations.
Molten
Carbonate
FC
(MCFC)
Lithium &
potassium

carbonate
~ 650 °C
H
2
, CO,
CH
4
, other
hydrocarb
on
>50%
Suitable for
medium to
large-scale
power
stations, up to
MW capacity.
Solid
Oxide FC
(SOFC)
Solid oxide ~1000 °C
H
2
, CO,
CH
4
, other
hydrocarb
on
>50%

Suitable for
all sizes of
power
stations, 2 kW
to multi-MW.


Table 1.1 Main types of fuel cells and their characteristics and applications
3

1.3 Proton Exchange Membrane Fuel Cells (PEMFCs)
Proton Exchange Membrane Fuel Cells are also known as polymer electrolyte
membrane fuel cells or polymer electrolyte fuel cells (PEFCs). Unlike other types of
fuel cells, a typical PEMFC uses a thin solid polymer membrane as its electrolyte.
This membrane is not only an electronic insulator but also an excellent conductor for
hydrogen ions, i.e. protons. Owing to its thin and solid properties, the polymer
membrane is able to greatly diminish the electrochemical corrosion of electrodes and
improve the power densities compared to liquid electrolytes. Furthermore, as it needs
water for proton conduction, a PEMFC system usually operates at a relatively low
temperature of about 80
o
C, which allows rapid start-ups of the system from ambient
temperature. These distinct advantages of PEMFCs make them particularly suitable
for automotive and portable applications.

1.3.1 History of PEMFCs
The history of fuel cells can date back to one and a half centuries ago when
William R. Grove invented the first fuel cell in 1839 [1]. His setup, so called
“gaseous voltaic battery”, included two platinum electrodes covered with inverted
tubes that were halfway submerged in a beaker of aqueous sulfuric acid. One of the

tubes was filled with hydrogen gas and the other was filled with oxygen. When these
electrodes were immersed in a dilute sulfuric acid, a current began to flow between
the two electrodes and water was formed in the inverted tubes. In order to increase the
voltage produced, Grove linked several of these devices in series and produced a “gas
battery”. The name of “fuel cell” had not been introduced until 1889 the chemists
Ludwig Mond and Charles Langer used platinum black supported on platinum or gold
electrodes as catalyst and introduced a diaphragm to contain the electrolyte between
4

the electrodes [1]. Historically, the first major application of PEMFCs was to serve as
an auxiliary power source in NASA’s Gemini space flights in the 1960s [1]. However,
the early version of PEMFCs with a polystyrene sulfonate ion-exchange membrane as
electrolyte was unable to give adequate power densities and required operational
lifetime. Thereafter the development of this technology remain stagnant for more than
ten years until the polystyrene sulfonic acid membrane was replaced by Du Pont’s
perflourosulfonic acid membrane (Nafion
®
) in the 1970s [1]. The utilization of Nafion
membrane in PEMFCs boosted the power densities by ten times and the lifetime from
two thousand hours to one hundred thousand hours. In the late 1980s and early 1990s,
another technical breakthrough was achieved via the 10-fold reduction of platinum
loading in PEMFC electrodes. This achievement was first realized by using high
surface area carbon particles as platinum support instead of using pure Pt black as
electrocatalyst, and also by impregnating a small amount of Nafion ionomers into the
catalyst layer of the porous gas diffusion electrode in the Gemini fuel cells [1].

1.3.2 Structure of PEMFCs
As shown in Fig. 1.1 [3], a typical single PEMFC usually consists of four main
components: tow current collectors, two gas diffusion layers, two catalyst layers and a
solid polymer electrolyte membrane.

5

















Fig. 1.1 Schematic diagrams of (a) a single PEMFC and (b) a 3-cell PEMFC stack [3].

Current Collectors
Current collectors, which are placed at two cell ends to collect current, are used to
separate reactant gases, and provide mechanical support and gas flow channels (see
Fig. 1.1) [4]. They are also called bipolar plates in multi-cell stacks when each plate is

electrically connecting the anode of one cell to the cathode of the adjacent cell. A
desirable material for current collector must be electrically and thermally conductive,
as well as impermeable to gases. It should also have high resistance to the reactant

(a)
(b)
6

gases to avoid corrosion. Presently graphite and stainless steel are the most commonly
used materials for current collectors, which provide both high electronic conductivity
and corrosion resistance. Current research on current collectors has been focusing on
developing novel materials with superior corrosion resistance and lower cost.

Gas Diffusion Layers (GDLs)
Gas diffusion layers are a part of a PEMFC electrode and they are in contact with
current collectors. They usually consist of a macroporous backing layer and a
microporous gas diffusion layer [4]. The macroporous backing layer is either carbon
cloth or carbon paper, with thickness ranging from 100 to 300 μm. And the
microporous gas diffusion layer is usually composed of nanosized carbon blacks
spread on the backing layer. These high porosity media can provide mechanical
support as well as pathways for electrons, gas and water in PEMFC electrodes. When
reactant gases flow out from the channels in current collectors and reach the gas
diffusion layer, they are evenly distributed by the porous structure before reacting in
the catalyst layer. In addition, the carbon backing layer is the electronic connection
between current collector and electrode. Furthermore, GDLs also provide
hydrophobicity for electrodes to facilitate liquid water removal by impregnating
Polytetrafluoroethylene (PTFE) into them. Hence the gas diffusion layer is a critical
factor in electrode structure design to ensure efficient mass transport in PEMFC
electrodes.

Catalyst Layers (CLs)
Catalyst layers are attached to the microporous gas diffusion layers. They are the
most essential part of a PEMFC where the electrochemical reactions take place [4].
7

The electrochemical reactions in a PEMFC consist of two separate reactions:
hydrogen oxidation reaction (HOR) occurring at anode and oxygen reduction reaction
(ORR) at cathode. At anode catalyst layer, gaseous hydrogen splits into two protons
and the protons then pass through the electrolyte membrane to reach the cathode.
While at cathode catalyst layer, oxygen combines with these protons from anode and
electrons from external circuit to form water and excess heat. Typically, these two
half-reactions would take place very slowly at PEMFC operating temperature.
Therefore electrocatalysts are necessary on both anode and cathode to increase the
reaction rate of each half-reaction. The best electrocatalyst for each reaction to date is
noble metal platinum, a very expensive material. In the 1970s and 1980s the catalyst
layer of PEMFCs was made up of pure platinum black and PTFE suspension with a Pt
loading up to 4 mg cm
-2
[1]. However, it was revealed in later research that it is the
catalyst surface area that determines the reaction rate rather than catalyst weight.
Thereafter a significant Pt loading reduction to less than 0.4 mg cm
-2
was achieved in
the late 1990s by synthesizing carbon supported Pt catalyst for PEMFC applications
[5]. Porous carbon blacks are widely used as Pt support for their high surface ratio and
excellent electrical conductivity. In the most prevalent catalyst layer, composite
catalysts consisting of Pt nanoparticles (4nm or smaller) supported on carbon black
Vulcan XC72R (ca. 40nm) are usually used with Nafion
®
ionomers impregnated. The

catalyst layers are usually very thin with a thickness of around 10−50 μm, containing
electrochemically active regions where three phases − catalysts, ionomers, and
reactant gases coexist. In order to improve the utilization of Pt catalyst, optimum
catalyst layer structure should be well-maintained to obtain maximum three-phase
zones [4].

8

Polymer Electrolyte Membrane (PEM)
A polymer electrolyte membrane is a solid polymer film that separates the anode
and cathode catalyst layers. It is a pivotal component of a PEMFC as it not only
permits protons to transport from anode to cathode but also insulates electrons to
travel through that the free electrons can only reach cathode through external circuit,
generating useful electricity. It also separates fuel and oxidant gases from each other
thus direct combustion of fuels can be avoided. In addition, a good PEM material
should remain chemical and mechanical stability in the hostile fuel cell environment
to ensure long-term operation durability. The most commonly used PEM at present is
Nafion
®
series invented by Dupont in 1960’s, due to its high proton conductivity and
chemical inertness [2]. The chemical structure of a typical Nafion PEM is composed
of a polytetrafluoroethylene (PTFE) chain and a side chain ending with sulphonic acid
HSO
3
. A micro-view of the PEM structure is shown in Fig. 1.2. The fluorocarbon
chain usually has a repeating structural unit, i.e. —[CF
2
–CF
2
]

n
—, where n is very
large. This chain can provide the PEM with good mechanical strength and chemical
stability. On the other hand, the sulphonic acid group HSO
3
is highly hydrophilic and
is ionically bonded with a SO
3
-
ion and a H
+
ion. This is why such structure is called
ionomer. When the PEM absorbs water and becomes hydrated, the ionically bonded
H
+
ions are relatively weakly attracted to the SO
3
-
group. As a result, the H
+
ions are
able to move through concentration gradient within the well-hydrated regions of PEM,
making this material a very good proton conductor. The proton conductivity of this
PEM is strongly correlated with its water content, thus making humidification of the
fuel gases a requirement during cell operation. Another requirement of using this
PEM is that the operating temperature is limited to the boiling temperature of water to
maintain its liquid water content. However, this requirement can lead to a severe mass
9

transport limitation when excess water accumulates in electrode and blocks gas

diffusion pores. Therefore water management is a very important topic in current
PEMFC research, especially in developing high-temperature PEM materials as well as
optimizing electrode structure [2].







Fig. 1.2 Chemical structure of Nafion membrane [2].

1.3.3 Basic Thermodynamics and Electrochemistry of PEMFCs
As shown previously, a typical PEMFC consists of three core parts: two
electrodes (anode and cathode) sandwiched with a solid polymer electrolyte
membrane between them. The combination of anode, cathode and membrane
corresponds to the heart of a PEMFC, known as membrane electrode assembly
(MEA). In a working PEMFC, hydrogen fuel flows into the anode catalyst layer
through gas diffusion layer, and it is then split by platinum catalysts into two electrons
and two protons (hydrogen ions). The protons can pass through the polymer
electrolyte membrane to cathode catalyst layer, whereas the electrons have to travel
from the anode to cathode through an external circuit consuming the power generated
by the cell. Again with the help of platinum catalysts, the protons and electrons
combine with oxygen within the cathode catalyst layer, producing pure water. The

10

overall electrochemical process of a PEMFC is called "reverse hydrolysis",
corresponding to the opposite reaction of hydrolyzing water to form hydrogen and
oxygen. The illustration of this process is shown below in Fig. 1.3.

Fig. 1.3 Illustration of electrochemical processes in PEMFCs [2].

As shown in Fig. 1.3, the electrochemical reactions occurring at anode and
cathode in a PEMFC can be illustrated as Eq. 1-1 to 1-3 [2]:
Anode: H
2
2H
+
+ 2e
-
(1-1)
Cathode: 2H
+
+ 2e
-
+ 1/2O
2
H
2
O (1-2)
Overall: H
2

+ 1/2O
2
H
2
O + electricity + heat (1-3)
According to Eq. 1-3, the electrochemical energy in hydrogen fuel is directly
converted into electricity through the PEMFC reaction, producing pure liquid water
and heat as the only by-products. As an electrochemical energy converter, a PEMFC
must obey the laws and principles of thermodynamics.





11

Open Circuit Voltage
In the H
2
-O
2
reaction of a PEMFC (Eq. 1-3), only the free energy part (
△
G) of
the total enthalpy (
△
H) can be converted into electricity due to irreversible entropy
changes (T
△
S), i.e.

△
G =
△
H - T
△
S (1-4)
The maximum electrical work that can be done by a PEMFC is given by the Gibbs
free energy change of the electrochemical reaction, and it relates to the reversible
open circuit potential E
0
by following equation:
W
e
=
△
G = -zE
0
F (1-5)
where z is the number of moles of electrons transferred in the reaction (here in
PEMFC z = 2), and F is the Faraday’s constant (96,485 C mol
-1
).
The reversible open circuit potential generated by a H
2
-O
2
PEMFC under standard-
state conditions (298.15 K, 1 atm) is thus
E

θ
= -
△
G
θ
/ zF =1.23 V (1-6)
where E
θ
is the standard-state reversible thermodynamic potential and
△
G
θ
(-237
kJ/mol) is the standard-state free energy change for H
2
-O
2
reaction.
At a given condition, the Gibbs energy change of the reaction becomes:

△
G =
△
G
θ
– RT In[p
H
(p
O
)

1/2
] (1-7)
Here p
H
and p
O
are the hydrogen and oxygen partial pressures (atm), respectively. T
denotes the cell temperature (K), R is the universal gas constant (8.314 J mol
-1
K
-1
).
Therefore the theoretical open circuit potential of a PEMFC working at a given
condition is determined by the following Nernst equation [4]:
E
0
= E
θ
– RT In[p
H
(p
O
)
1/2
]/zF (1-8)

12

In practice, the open circuit potential is usually about 0.2 V lower than the
theoretical value. One main reason causing this deviation is the formation of hydrogen

peroxide as an intermediate of the oxygen reduction reaction. Another factor that
attributes to it lies in the diffusion of hydrogen from anode to cathode through
polymer electrolyte membrane. As a result, the open circuit potential is difficult to
determine purely by estimation from the theoretical equations shown above [4].

Theoretical Efficiency
The efficiency of any energy conversion device is defined as the ratio between
useful energy output and energy input [6]. A simplified way to compare the
efficiencies of energy conversion devices is to examine their maximum theoretical
efficiencies. For the H
2
-O
2
reaction in PEMFCs, it is the changes in Gibbs free energy
of water formation that are converted into electrical energy [4]. The maximum
efficiency for a PEMFC can be calculated based on the changes in Gibbs free energy
△
G and the changes in enthalpy
△
H of the reaction:
η
max
=
△
G/
△
H (1-9)
The total electrochemical energy can be obtained from any electrochemical reaction
is determined by the total Gibbs free energy change
△

G, which is approximately -237
kJ mol
-1
for the H
2
-O
2
reaction at the standard-state condition. If the process is
reversible, all the Gibbs free energy will be converted into electrical energy. In
practice, however, the process is usually not reversible and some Gibbs free energy
will be released as heat. The enthalpy change
△
H represents the total thermal energy
available from the reaction, and its value is varied depending on whether the product
water is in vapor or in liquid phase. If the produced water is in liquid phase, then
△
H
is higher due to the release of heat from water condensation. The higher
△
H value is
13

called higher heating value (HHV) with the amount of -286 kJ mol
-1
, and the lower
△
H value is about -241 kJ mol
-1
, denoted as lower heating value (LHV). Assuming
that all of the Gibbs free energy can be converted into electrical energy, the maximum

theoretical efficiency of a PEMFC is given by:
η
max
=
△
G/
△
H = -237 / -286 = 83% (1-10)
However, the actual efficiency of a real fuel cell is lower than this theoretical value
shown above due to various irreversible losses under certain operating conditions. It is
also proportional to the cell potential that corresponds to the cell output power [4].
Denote E
HHV
as the potential corresponding to hydrogen’s higher heating value, or the
thermoneutral potential:
E
HHV
= -
△
H

/ zF =1.48 V (1-11)
The actual efficiency of a working PEMFC can be estimated as the ratio of the cell
potential E and the potential corresponding to hydrogen’s higher heating value E
HHV
:
η = E / E
HHV
= E / 1.48 (1-12)

Polarization / Overpotential
When a current is drawn in a working PEMFC, a deviation of cell potential from
the thermodynamic equilibrium potential occurs according to the electrical work done
by the cell [6]. The deviation from the equilibrium value is called overpotential or
polarization, denoted as V. There are three major overpotentials causing different
irreversible voltage losses during cell operation: activation overpotential, ohmic
overpotential and mass transport overpotential. Therefore, the output voltage of a
single fuel cell can be determined by the sum of the open circuit potential and the
three voltage losses:
E = E
0
+ V
act
+ V
ohm
+ V
mt
(1-13)
14

Where V
act
is the activation polarization, V
ohm
is the ohmic polarization and V
mt
is the
mass transport polarization.

Activation overpotential is caused by the slow charge transfer rate at the

interface between electrocatalyst and electrolyte. This voltage loss is utilized to
activate the electron transfer in the reaction. Activation overpotential dominates
particularly at low current density region where the ohmic and mass transport losses
are negligibly small. It can be mostly attributed to the sluggish cathode reaction
whereas the contribution of anode activation polarization is negligible. The main
factors that affect activation overpotential are the electrochemical kinetics of cathode
reaction, including reaction temperature, catalyst surface area, reaction Tafel slope,
exchange current density and so forth [4].

As current density increases, ohmic overpotential becomes predominant for the
voltage losses occurring in the PEMFC. It can be ascribed to the overall ohmic
resistance of the PEMFC, including the electrical resistance of electrodes and other
cell components as well as the ionic resistance of PEM. Usually the PEMFC
components are made up of materials with very high electrical conductivity so that the
ohmic resistance of a PEMFC is mainly derived from the relatively much slower ionic
conductivity of the electrolyte. Ohmic resistance can be reduced by decreasing
membrane thickness or increasing ionic conductivity of electrolyte. However,
thinning membrane thickness is limited by losing the mechanical strength required for
the PEM. To alleviate the ohmic loss from ionic resistance of PEM, fuel gases are
usually humidified to sustain high water content in the electrolyte [4].

15

Mass transport overpotential arises from mass transport limitations of reactant
gases in PEMFC electrodes, mainly occurring at large current densities. When the
supply of reactant gases is not rapid enough to meet the fast reaction rate at large
current densities, a part of the generated energy is lost to drive mass transport process,
yielding a corresponding loss in output voltage. This loss is mostly determined by the
structure characteristics of GDLs and CLs, as well as the concentration and pressure
of reactant gases. In addition, mass transport overpotential can also occur when water

accumulates in the electrode pores and obstructs the diffusion paths or dilutes the
reactants, thus exacerbating the voltage loss [4].

Polarization curve is a characteristic diagnostic technique to evaluate fuel cell
performance, illustrating a plot of cell potential vs. current density. Polarization curve
is usually obtained by increasing cell current starting from open circuit potential and
recording I-V measurements at prescribed potential or current intervals. A typical
polarization curve of a PEMFC is shown in Fig. 1.4. As can be seen in Fig. 1.4, the
cell potential decreases with increasing current densities and three distinct regions of a
PEMFC polarization curve are noticeable according to the characteristics of voltage
losses. At low current densities, a dramatic voltage drop is drawn from open circuit
potential due to the activation overpotential. As current density increases, the
polarization curve shows a nearly linear region at intermediate current densities,
where voltage loss is dominated by ohmic overpotential. At large current densities,
the decrease of cell potential deviates from linear relationship with current and
aggravates dramatically till the maximum current limit is reached. This additional
sharp voltage drop stems from the mass transport overpotential, as described in
previous sections. In general, polarization curve measurement is commonly
16

Fig. 1.4 A typical polarization curve of PEM fuel cell.
performed as the first characterization step to evaluate fuel cell performance for its
convenience and ease of data acquisition and interpretation [1].

1.3.4 Applications of PEMFCs
PEMFCs are particularly attractive for applications ranging from low (less than 1
kW) to intermediate (up to 50 kW) power levels, due to their inherent advantages
such as low emissions, high efficiency, low operation temperature and rapid start-up
as described in the above introduction. The main driving force for the
commercialization of PEMFCs is from the automotive industry. Presently the world’s
major automobile manufacturers, such as General Motors, DaimlerChrysler, Toyota
Motor Corporation, Ford and etc., have been developing PEMFC powered vehicles
for their commercial viability to replace combustion engines, since the Canadian fuel
cell company Ballard invented the first automobile powered by a PEMFC in 1993.
0 400 800 1200 1600
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Cell potential / V
Current density / mA cm
-2
theoretical equilibrium potential 1.23 V
actual open circuit potential

polarization curve
activation polarization
ohmic overpotential
mass transport overpotential

17

The latest PEMFC vehicle developed by DaimlerChrysler and Ballard has grown to
the six generation. It has the capability of traveling 150 km with one fill of fuel and its
highest speed can reach up to 140 km/h [1]. However, barriers to the
commercialization of PEMFCs to date remain formidable due to their exiguous fuel
sources and high economic costs. One main problem lies in that expensive platinum
must be used as catalyst for hydrogen oxidation reaction at low temperature.
Moreover, the platinum catalyst is extremely sensitive to CO contained in the
hydrogen fuel. Therefore intensive research on advanced PEMFC electrocatalysts
with higher electrocatalytic performance and lower cost are pressingly needed to
provide competitive price and reliable performance for PEMFCs [4].

1.4 Electrocatalytic Material Studies of PEMFCs
The electrochemical reactions that take place in PEMFCs are the hydrogen
oxidation reaction (HOR) at anode and the oxygen reduction reaction (ORR) at
cathode, respectively. Although these reactions are thermodynamically favorable,
they cannot occur spontaneously at normal conditions due to their sluggish kinetics,
especially for the ORR. A suitable electrocatalyst is thus necessary to speed up the
reaction kinetics to meet the requirement for practical uses of PEMFCs. In the state-
of-the-art PEMFCs, platinum (Pt)-based catalysts are the most effective
electrocatalysts for both HOR and ORR, owing to the high electrochemical activity
and stability of platinum under PEMFC operating conditions. However, these Pt-
based catalysts consisting of noble metals are too expensive to make commercially
viable PEMFCs; hence extensive research over the past several decades has been

focusing on reducing Pt loading and improving Pt utilization in PEMFC electrodes.
To achieve these aims, novel electrocatalytic materials are developed mainly through
18

two strategies: Pt-based alloy catalysts and Pt catalysts on carbon support. In the
following sections, an overview of supported Pt catalysts will be given including Pt
catalysts supported on carbon black (CB) and carbon nanotube (CNT) as well as their
diverse synthesis methods.

1.4.1 Carbon Supported Pt Electrocatalysts for ORR
Platinum has been used as the electrocatalyst for PEMFC reactions since the
invention of the first prototype of PEMFC. It is believed that Pt has the best
electrochemical activity on the oxygen reduction reaction (ORR) over any other pure
metal. In earlier PEMFC models, pure Pt black was used as the electrocatalyst in
PEMFC electrodes with a high Pt loading of 4 mg cm
-2
[1]. It is only till late 1990’s
that a significant Pt loading reduction to less than 0.4 mg cm
-2
was achieved by
synthesizing supported Pt catalysts as PEMFC electrocatalyst [5]. Practical
electrocatalysts for current PEMFCs are typically nanosized Pt particles dispersed on
high-surface-area carbon supports. The most commonly used support material is
carbon black Vulcan XC-72R (VXC72R), which has an average particle size of 30
nm and a high surface area of 250 m
2
g
-1
[4]. Despite of the significant Pt loading
reduction by utilizing such composite catalysts, the costly amount of Pt catalyst in

state-of-the-art PEMFCs remains one of the major barriers for PEMFC
commercialization. Thus further reducing Pt loading has been a main driving force for
PEMFC research in recent years. There are a variety of synthesis methods for Pt/C
composite catalysts, including wet-chemical processes and physical vapor deposition
(PVD) techniques. Each method has its particular advantages over the others and also
disadvantages that limit its ubiquitous application. A general overview of the most
prevailing synthesis methods for carbon supported Pt catalysts is presented below.
19

Particular emphasis will be given to the sputter-deposition technique that was adopted
in this study.

Chemical Precipitation
Chemical precipitation method for supported Pt catalyst synthesis has been used
for several decades [7]. The main advantage of this method is that it can prepare
nanosized Pt catalysts at low temperatures. In principle, the preparation process is
basically reducing a Pt salt solution by adding a reducing agent into it. Typically, the
precursor Pt salt is firstly reduced to Pt metallic state. The reduced Pt particles
precipitate out of solution and deposit onto support particles when carbon supports are
mixed in the solution. Aqueous Pt salt solution such as H
2
PtCl
6
is commonly used as
precursor solution and sodium borohydride and hydrazine hydrate are popular
reducing agents. Continuous stirring must be conducted during Pt reduction to ensure
sufficient mixing of the precursor solution and the reducing agent. After reduction
process, the supported Pt catalysts will undergo a series of filtration, washing and
stoving processes. In general, the average Pt particle sizes synthesized through
chemical precipitation method are 2−5 nm [8, 9]. The main disadvantages of this

method are its time-consuming preparation process and incomplete Pt reduction that
additional reduction process under H
2
may be required for the as-synthesized Pt/C
catalysts.

Colloidal
The colloidal method is a common synthesis method for present Pt/C catalysts in
PEMFCs [10]. It is similar to the chemical precipitation method whereas it can
provide size control of the catalyst particles by adding a capping agent to prevent their
20

agglomeration. An effective capping agent usually contains molecules that can be
adsorbed onto the catalyst particles, such as H
2
O
2
. In the colloidal preparation process,
the Pt precursor solution (H
2
PtCl
6
) is mixed with a reducing agent (NaHSO
3
) and a
capping agent (H
2
O
2
) together. To synthesize supported Pt catalysts, carbon support is

added to the mixture before or after the formation of the catalyst particles. The Pt
precursor first forms PtO
2
colloidal from H
2
PtCl
6
with the reduction of NaHSO
3
at a
fixed pH of 5 [11]. The Pt colloidal then deposits onto carbon support and Pt/C can be
obtained after reduction under H
2
at 300 °C. The synthesized Pt catalysts usually have
relatively uniform particle sizes ranging from 2−4 nm. Bimetallic catalysts can also be
synthesized with the colloidal method by co-reduction of two metal precursor
solutions [12].

Impregnation
Impregnation method is also a commonly used technique due to its ease of
preparation process [13]. In this process, high-surface-area carbon black VXC72R is
firstly impregnated into an aqueous Pt solution such as Pt(NH
3
)
2
(NO
2
)
2
[14]. After air

drying, Pt/VXC72R catalysts are synthesized by reducing the mixture under H
2
at
300 °C for 3 h. As the reduction process is subsequent to the impregnation process,
the support material plays a crucial role in the size control of catalyst particles. The
particle size of Pt catalysts prepared by impregnation method usually has a relatively
large distribution range due to particle agglomeration. In addition, the high surface
tension of liquid solution may cause carbon support to collapse that limits the overall
effectiveness of this method [15].




21

Electrochemical Deposition
Electrochemical deposition has the advantage of being able to deposit Pt
nanoparticles on various catalyst supports [16-21]. Generally it takes place at the
interface between an electronically conductive substrate and an electrolyte solution
containing the ions of the catalyst metal. The particle size of Pt catalyst can be
controlled by deposition time and current. Typically, this method is difficult to
deposit Pt particles on powder substrate thus gas diffusion electrodes spread with
VXC72R GDL are usually used as deposition substrate [18]. In addition, this method
can be greatly affected by the infiltration of the electrolyte solution with the substrate
material. Deposition of Pt catalysts is usually very poor when the contact angle is high
at the electrolyte-substrate interface [21].

Sputter-deposition
Sputter-deposition technique is well-known to be one of the most effective thin-
film deposition techniques for having excellent control in film thickness and

uniformity. This technique is widely employed in semiconductor and hard-disk
industry and very recently PEMFC researchers found that ultra-low Pt loading with
high Pt utilization can be achieved for PEMFC applications by radio-frequency (R.F.)
magnetron sputtering technique [22-27]. R.F. magnetron sputtering is a physical
vapor deposition (PVD) process in which the atoms of a solid material (target) are
ejected by the knock of plasma of an inert gas (e.g. Ar). It is the most common
sputtering technique used in practice as it is able to confine the plasma close to the
target surface. Another important advantage of this magnetron sputtering is its high
deposition rate thus it is capable of forming denser layers than the alternative PVD
methods. In catalyst deposition for PEMFCs, the metal target acts as the catalyst
22

source, emitting catalyst atoms onto the substrates, usually electrolyte membranes or
gas diffusion layers. An ultra-low Pt loading of 0.014 mg cm
-2
can be obtained by
sputtering a 5 nm Pt layer onto the catalyst layer [22]. In addition, various catalyst
support materials can be used as substrates in sputtering process unlike in the wet-
chemical processes described previously. Furthermore, a large range of Pt-based
binary, ternary and quaternary alloy catalysts can be easily synthesized by sputtering
multiple metal targets consisting of different catalyst elements [27]. It can provide an
enormous versatility in fabricating supported Pt-based catalysts for PEMFC
applications.

Sputter-deposited Pt catalysts for fuel cell applications were firstly investigated
by Cahan and Bockris in the late 1960s [28]. This method was then improved by
Asher and Batzold several years layer whereas the performance of the fuel cell with
sputter-deposited catalyst layers was not adequate [29]. In 1987, Weber et al. stated in
their study that electrode fabrication could be greatly streamlined by sputter-
deposition of Pt catalyst onto wet-proofed, porous substrates [28]. They found that

their sputtered Pt catalysts on cathode performed considerably better than those on
anode, giving rise to current densities as high as 500 mA cm
-2
. They also found that
the performance of their sputtered electrodes depended more on the substrate structure
and properties than on the sputtering process. Substrates with impregnated PTFE
showed a significant influence on cell performance. In their study, the typical Pt
loading obtained by sputtering was 0.15 mg cm
-2
supported on porous carbon black-
based electrodes. In 1997, Hirano et al. prepared their sputter-deposited catalyst
layers for high performance PEMFCs [30]. The performance of their cells with
sputtered Pt catalysts of 0.1 mg cm
-2
at cathode was nearly equivalent to that of a
23

commercial electrode. However, a notable performance drop was observed at large
current densities for their sputtered electrodes. Moreover, they also observed that
when the sputtered Pt loading was reduced to 0.04 mg cm
-2
the cell performance was
visibly lower, presenting a power density of 160 mW cm
-2
at the current output of 200
mA cm
-2
. In 2002, O’Hayre et al. [22] reported their high performance sputter-
deposited electrodes with an ultra-low Pt loading about 0.04 mg cm
-2

. In their study,
thin Pt catalyst layers were directly sputter-deposited onto both sides of Nafion
membranes. Their MEA obtained by sputter-deposition could give rise to a maximum
power output up to three-fifths that of a commercial E-TEK MEA, of which the Pt
loading was 0.4 mg cm
-2
. Later Brault et al. [31] found that the maximum size of the
Pt nano-clusters formed on electrodes by sputtering was about 4 nm and the optimum
Pt loading was about 0.1 mg cm
-2
to give comparable performance as a commercial
MEA with a Pt loading of 0.4 mg cm
-2
. Their results also showed that the influence of
the sputter-deposited electrode was greater at cathode than at anode, confirming the
results of Weber et al. [28]. They suggested that the sputter-deposition technique can
greatly enhance Pt utilization due to the optimum location and distribution of the Pt
nanoparticles at the electrolyte-electrode interface.

Despite of the high Pt utilization obtained by sputter-deposition technique,
however, this method is limited by its nature of surface deposition. It is believed that
the electrochemical activity of Pt catalysts significantly depends on their surface area.
As such, porous CB-based GDLs are usually fabricated for Pt sputtering process to
increase the surface area of the sputter-deposited Pt catalysts. When Nafion
membrane was used as substrate, the optimum Pt loading correspond to a Pt film
5−10 nm think, which is about only 0.01−0.02 mg cm
-2
[22]. Pt loadings higher than
24


this amount would give rise to a dense Pt film on the membrane, which tremendously
led to a reduced Pt surface area and thus a considerably impaired cell performance.
Therefore, it was suggested that using suitable substrates with high surface area and
porosity for sputtered Pt catalysts is vitally important to optimize the effectiveness of
sputter-deposition technique for PEMFC applications.

1.4.2 Carbon Nanotubes as Support for Pt-based Electrocatalysts
It is well-known that the electrochemical activity of Pt catalysts greatly depends
on the Pt particle size as well as their dispersion on the support material [32]. The
intrinsic properties of support materials play an important role in the catalytic
performance of Pt catalysts by affecting their structure and morphology. An ideal
catalyst support should provide several desirable properties such as: (i) high electrical
and thermal conductivity, (ii) high surface area and porosity to ensure reactant gas
access to electrocatalysts, and (iii) high electrochemical stability under fuel cell
operating conditions [4]. Currently the most prevalent electrocatalyst for PEMFCs is
Pt nanoparticles dispersed on carbon black support. In spite of the high surface area of
carbon black particles, there remain two main problems for this support material [33]:
(i) carbon black-based catalyst may suffer significant mass transport limitations due to
its dense structure, which leads to low Pt utilization; (ii) carbon black is susceptible to
the hostile PEMFC environment and it would undergo electrochemical oxidation to
surface oxides and eventually, to CO
2
at cathode, where it is subjected to low pH,
high potential, high humidity and high temperatures (80 °C). As carbon black
corrodes, the Pt nanoparticles on carbon black will fall off from the support, and
consequently dissolve into water or aggregate to larger particles and attach onto
adjacent support. This is one of the major mechanisms that cause cell performance
25

degradation due to support oxidation and Pt catalyst loss [34]. Therefore, many

research efforts have been devoted to exploring for new catalyst support materials for
PEMFC applications.

Recently, a number of studies have shown that carbon nanotubes (CNTs) are
promising alternative support materials for PEMFC electrocatalysts due to their
distinct graphitic characteristics compared to those of carbon blacks [35-44]. CNTs
with higher graphite component exhibit superior electronic conductivity, high
electrochemical stability and excellent hydrophobility; therefore, they have been
proposed as alternative catalyst supports to replace carbon blacks in recent PEMFC
research. Many researchers have reported that the CNT supports were able to provide
higher corrosion resistance and mass transport capability for PEMFC electrodes [45-
48]. In addition, Tauster et al. [49] found an effective atomic interaction between the
Pt particles and the CNT supports, which could give rise to improved catalytic
performance by facilitating the electron transfer process for the PEMFC reactions.
Enhanced Pt utilization was thus obtained by reducing the Pt catalyst loading without
performance losses.

Previous studies on the electrochemical performance of Pt/CNT composite
catalysts were performed predominantly via cyclic voltammetry (CV) tests in a
simulated PEMFC environment [40-44]. The electrochemical performance of Pt/CNT
catalysts is usually compared with that of commercial Pt/VXC72R catalysts. When
this is conducted by cyclic voltammetry studies, the performance of Pt/CNT catalysts
always show notable activity enhancement for the oxygen reduction reaction (ORR)
compared to that of Pt/VXC72R catalysts [40-44]. For example, Kim and coworkers