Fabrication of platinum electrocatalysts on carbon nanotubes using atomic layer deposition for proton exchange membrane fuel cells

Article abstract of DOI:10.1016/j.electacta.2012.04.084

Platinum (Pt) has been synthesized by atomic layer deposition (ALD) onto the surface of multi-layered carbon nanotubes (CNTs) in order to investigate catalytic activity and durability of the ALD-Pt catalyst electrode in proton exchange membrane fuel cells (PEMFCs). Methylcyclopentadienyl-(trimethyl) platinum (MeCpPtMe₃) and oxygen serve as precursors at a deposition temperature of 250 °C. One growth mechanism involving two self-limiting reactions is proposed to clarify the decoration of ALD-Pt nanoparticles over the CNT support. The resultant ALD-Pt catalyst displays not only the fairly good electrochemical activity (electrochemically active surface area) but also the stability after potential cycling >1000 cycles. Experimental results also show that the CNTs decorated with the ALD-Pt catalysts exhibit low-equivalent series resistance (～3.55 Ω), high-power density (～2.95 kW g Pt^-1 at 80 °C) and long-term durability. On the basis of the results, the ALD approach is capable of synthesizing well-dispersed Pt nanoparticles onto CNTs, forming the advanced design of catalyst electrode for PEMFCs.

Full text of DOI:10.1016/j.electacta.2012.04.084

Electrochimica Acta 75 (2012) 101–107
Contents lists available at SciVerse ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Fabrication of platinum electrocatalysts on carbon nanotubes using atomic layer
deposition for proton exchange membrane fuel cells
Ting Shua, Shi-Jun Liaoa, Chien-Te Hsiehb,∗, Anup Kumar Royb, Yung-Ying Liub, Dong-Ying Tzoub,
Wei-Yu Chenb
a School of Chemistry and Chemical Engineering, South China University of Technology, Guangdong 510640, China
b Department of Chemical Engineering and Materials Science, Yuan Ze Fuel Cell Center, Yuan Ze University, Taoyuan 32003, Taiwan
a r t i c l e i n f o
a b s t r a c t
Article history:
Platinum (Pt) has been synthesized by atomic layer deposition (ALD) onto the surface of multi-layered
carbon nanotubes (CNTs) in order to investigate catalytic activity and durability of the ALD-Pt catalyst
electrode in proton exchange membrane fuel cells (PEMFCs). Methylcyclopentadienyl-(trimethyl) plat-
inum (MeCpPtMe3) and oxygen serve as precursors at a deposition temperature of 250 ◦C. One growth
mechanism involving two self-limiting reactions is proposed to clarify the decoration of ALD-Pt nanopar-
ticles over the CNT support. The resultant ALD-Pt catalyst displays not only the fairly good electrochemical
activity (electrochemically active surface area) but also the stability after potential cycling >1000 cycles.
Experimental results also show that the CNTs decorated with the ALD-Pt catalysts exhibit low-equivalent
series resistance (∼3.55 ꢀ), high-power density (∼2.95 kW g Pt−1 at 80 ◦C) and long-term durability. On
the basis of the results, the ALD approach is capable of synthesizing well-dispersed Pt nanoparticles onto
CNTs, forming the advanced design of catalyst electrode for PEMFCs.
Received 4 April 2012
Accepted 21 April 2012
Available online 27 April 2012
Keywords:
Atomic layer deposition
Pt catalysts
Carbon nanotubes
Electrochemical activity
Fuel cells
© 2012 Elsevier Ltd. All rights reserved.
1. Introduction
carbon electrodes (e.g., CNTs) can be considered as one of the key
factors in determining the performance of PEMFCs.
Fuel cells such as proton exchange membrane fuel cells (PEM-
FCs) are potential candidates for clean energy due to their
eco-friendly system, high energy conversion efficiency, and high
power density [1–3]. Despite considerable advances, some techni-
cal and economical barriers still exist, hindering the widespread
use of low-temperature PEMFCs. The challenges to commercialize
PEMFCs include (i) reducing the amount of noble metals (e.g., Pt)
used as electrocatalysts, (ii) enhancing the Pt utilization efficiency,
and (iii) preventing the performance decay during long-term oper-
ation [4]. Recently, there has been increasing interest in carbon
nanotubes (CNTs) as heterogeneous catalyst support in fuel cell
applications. Pioneer studies have pointed out that Pt [5–7] or
bimetallic Pt-based [8,9] nanoparticles supported on CNTs show
an excellent catalytic activity. One-dimensional CNTs serve as a
good electronic conductor and provide a high surface area for the
attachment of catalysts with well dispersion. However, the activity
of Pt-based catalysts is generally reduced after cycling due to three
mechanisms: (i) Pt agglomeration on carbon supports upon the Pt
migration, (ii) growth of Pt particle size due to Ostwald ripening;
and (iii) appearance of unsupported Pt particles from corrosion of
carbon supports [10]. Accordingly, the durability of Pt catalysts in
Atomic layer deposition (ALD) is a thin film growth technique
that enables an excellent conformal and uniform growth with pre-
cision control over both thickness and composition [11]. The ALD
technique basically depends on the principle of splitting a binary
reaction into two half-reactions and alternating the exposure of
gas-phase precursors for each half-reaction to a surface [12]. Such
conformal growth is attributable to the atomic level control of thin-
film growth using sequential and self-limiting surface reactions.
Thus, two-dimensional nano-architectures, such as Ir thin film [13],
Pt layers onto metallic cores (e.g., Ag, Pd, and Ru) and stabilized
zirconia substrate [14], have been successfully fabricated by ALD.
More recently, the ALD approach was applied to deposit Pt nanopar-
ticles on TiO2 catalysts [11], mesoporous silica [12], carbon aerogels
[15], and CNTs [16]. The latter two references have pointed out
the feasibility of ALD-Pt catalysts on different support materials in
fuel cell applications. However, a detailed investigation on ALD-Pt
catalyst electrodes with higher power efficiency and better cycle
stability is still required.
Accordingly, one strategy associated with the design of elec-
trode materials is proposed here to prepare ALD-Pt catalysts on
CNT-carbon paper (CP) composite support that can raise the power
efficiency and the stability of Pt catalyst electrodes. Multi-layered
CNTs were well dispersed over the CP substrate, forming com-
posite supports. The ALD technique could be applied to deposit
the Pt nanoparticles inside the high-aspect-ratio carbon structures,
∗
Corresponding author. Tel.: +886 3 4638800x2577; fax: +886 3 4559373.
E-mail address: cthsieh@saturn.yzu.edu.tw (C.-T. Hsieh).
0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.electacta.2012.04.084

102
T. Shu et al. / Electrochimica Acta 75 (2012) 101–107
e.g., CNT forest. To examine the performance of ALD-Pt catalyst
electrodes, the electrochemical surface area (ESA) after potential
cycling and the power density of single cells at different operat-
ing temperatures are systematically investigated. The effort of the
present work would shed some lights on the feasibility of the ALD
technique on the synthesis of Pt catalysts for high-performance
PEMFCs.
Cyclic voltammetry (CV) measurement was performed at five
sweep rates (i.e., 5, 15, 25, 35, and 45 mV s−1) under N2 atmosphere
at ambient temperature using 1 M H2SO4 as the electrolyte solution
using an electrochemical analyzer (CH Instrument, Inc., CHI 608).
The area of ALD-Pt catalyst electrodes was set at 1 cm2. Herein
the Pt wire and the Ag/AgCl electrode were used as the counter
and the reference electrodes, respectively. The working electrodes
were constructed by pressing the CNT-CP composites onto a stain-
less steel foil, which served as the current collector. The potential
scan rate and scan number were set at 30 mV s−1 and 1000 cycles,
respectively, confirming the stability of ALD-Pt catalysts in acid
electrolyte. The electrochemical analyzer was also used to mea-
sure ac electrochemical impedance spectra of Pt catalyst electrodes
before and after 1000 CV cycles. In this study, the potential ampli-
tude of ac was equal to 5 mV, and the frequency was from 100 kHz
to 1 MHz.
2. Experimental
2.1. ALD of Pt on CNT-CP supports
The ALD substrates were prepared by coating the CNT ink
directly on the CP substrate (SGL 10 BC, Sigracet, Germany) through
a screen painting process. The CP sample is made of fibers having a
diameter of approximately 8–10 ␮m and the surface of the carbon
fiber is clean before the coating of CNTs. Prior to coating onto CP
substrate, the CNT samples were chemically oxidized by citric acid
treatment. 2.0 g of CNTs were mixed with 2.0 g of citric acid (99.5%)
and 50 ml distilled water with the assistance of ultrasonic vibration
for 1 h. After that, the mixture was heated in a muffle furnace at
300 ◦C for 0.5 h. No post-washing and filtering were required since
the thermal decomposition temperature of citric acid is 175 ◦C. The
purpose of the citric acid treatment was to implant surface oxides,
such as carboxyl, and hydroxyl groups, on the ends or the side-
wall of the CNTs [17]. The oxidized CNTs (50 mg) were mixed with
25 ml isopropyl alcohol (85%) and 0.15 ml polytetrafluoroethylene
(17 mg ml−1) solution. The CNT slurry was then ultra-sonicated for
30 min to obtain a homogenous suspension. The CP substrates were
cut into an area of 3 cm × 4 cm. The CNT ink was painted by a coater,
and the CNT-CP composites were placed in an oven and evaporated
at 100 ◦C overnight. The CP composites were weighted before and
after the CNT coating, and the loading of CNTs was approximately
2.3. Performance and durability test of a single cell
To examine the performance of a single cell, the membrane
electrode assembly (MEA) was fabricated using the ALD-Pt elec-
trode as the anode, the Nafion® 212 membrane (Dupont Inc.) as the
proton exchange membrane, and a commercial gas diffusion elec-
trode (HiSPEC® 4100, 40 wt.% Pt/C, 0.40 mg cm−2, Johnson Matthey)
as the cathode. The ALD-Pt catalyst electrode was painted with
5 wt.% Nafion 212 solution. The Nafion-impregnated electrodes and
membrane were then hot pressed together at 135 ◦C for 2 min
at a pressure of 30 atm to form the MEA. Subsequently, the MEA
was inserted between two graphite plates with a serpentine flow
pattern. The performance test was conducted using a 5-cm2 sin-
gle fuel cell operated at 40, 60, and 80 ◦C. High-purity hydrogen
(99.999%) and oxygen (99.999%) were fed at the anode and cathode,
respectively, both at flow rates of 200 cm3 min−1 at 100% relative
humidity. The polarization curve and the power density of the sin-
gle cell were then monitored and characterized. The durability of
ALD-Pt catalysts was carried out by operating the single cell with a
constant voltage of 0.7 V at 60 ◦C. The current density of single cell
was continuously recorded within 50 h. One MEA fabricated with
both commercial gas diffusion electrodes (HiSPEC® 4100, 40 wt.%
Pt/C, 0.40 mg cm−2, Johnson Matthey) was also conducted to com-
pare its durability with ALD-Pt electrode.
1.0 mg cm−2
.
In the ALD of Pt catalysts on CNT-CP supports, the Pt nanopar-
ticles were grown on carbon supports, using (methylcyclopen-
tadienyl) trimethylplatinum (MeCpPtMe3, 99%) and high-purity
oxygen (99.9995%) as ALD precursors. Prior to the ALD process, the
O2 gas was dried through a desiccant dryer to a dew point tem-
perature of −45.5 ◦C before use, and the MeCpPtMe3 was kept at
60 ◦C. An ultra-high-purity N2 (99.9995%) was used as a carrier and
a purging gas. The regular ALD cycle consisted of a 1-s exposure
to MeCpPtMe3, a 20-s N2 purge, a 5-s pumping, a 2.5-s exposure
to O2, a 20-s N2 purge and a 5-s pumping. The purge and pump-
ing steps were used to remove any unreacted precursors from the
CNT supports. The Pt nanoparticles were deposited with the sup-
port temperature at 250 ◦C at a reactor pressure of 700 mTorr after
100 ALD cycles.
3. Results and discussion
3.1. Characterization of ALD-Pt catalysts on CNTs
A top-view FE-SEM image of the resultant ALD-Pt catalysts on
CNT-CP composite is depicted in Fig. 1(a). Generally, the CP sub-
strate consists of an irregular interlacing of microscale carbon fibers
(diameter: approximately 8 ␮m). A large number of curved nano-
tubes are deposited over the surface of carbon fibers after surface
coating of CNTs. Each nanotube appears as an average diameter
of 30–60 nm and a length of several micrometers. Due to their
small particle size, it is difficult to observe the dispersion of ALD-Pt
nanoparticles over the CNT forest. Figs. 1(b) and 2(c) show the HR-
TEM images of Pt nanoparticles on the sidewall of CNTs at low and
high magnifications, respectively. The Pt nanocatalysts, having an
average sizeof3–5 nm, are nucleatedheterogeneously onto thesur-
face of CNTs through the ALD process. The ALD-Pt particles exhibit a
well dispersion with a narrow size distribution on the CNT support,
confirming the suitability of ALD in depositing Pt over high-aspect-
ratio CNTs. The selected-area electron diffraction pattern from an
area focusing on the ALD-Pt nanoparticle is illustrated in the inset
of Fig. 1(b), reflecting the presence of bright diffraction spots along
2.2. Characterization of ALD-Pt catalysts
The topography and distribution of ALD-Pt nanoparticles on the
CNT support were observed using a field-emission scanning elec-
tron spectroscope (FE-SEM, JEOL JSM-5600) and a high-resolution
transmission electron microscope (HR-TEM, JEOL, JEM-2100). A
thermogravimetric analyzer (TGA, Perkin Elmer TA7) was adopted
to analyze the amount of ALD-Pt catalysts deposited on the car-
bon composites. The Pt-loaded carbon composite was carefully
cut into an area of 1 cm2 and then placed in the TGA. The sam-
ple was heated to 1000 ◦C at a heating rate of 10 ◦C min−1 in air
atmosphere. To ensure the accuracy of the Pt loading, the TGA anal-
ysis was conducted for each sample twice. The crystalline structure
of the ALD-Pt catalysts was examined by grazing incident X-ray
diffraction (XRD) with Cu-K␣ radiation, using an automated X-ray
diffractometer (Rigaku, D/MAX 2500).

T. Shu et al. / Electrochimica Acta 75 (2012) 101–107
103
Fig. 1. (a) FE-SEM photographs of ALD-Pt nanoparticles over CNT forest. HR-TEM images of ALD-Pt nanoparticles onto the surface of CNTs with (b) low- and (c) high-
magnification. The inset of (b) shows selected-area diffraction pattern on the ALD-Pt nanoparticle.
with diffraction rings. The existence of such rings suggests that the
Pt grain consists of Pt nanocrystallites.
information regarding the chemical composition of oxidized CNTs.
The EDS analysis showed that only C and O elements are present
on the oxidized CNTs. The O/C ratio on the CNTs exceeded 15.5%,
reflecting a number of surface oxide groups attached to both ends
and sidewalls of the nanotubes. The presence of oxygen function-
alities has been demonstrated to act as the key role in affecting the
distribution of Pt monolayer through self-limiting reactions [16].
Because surface oxides (e.g., carbonyl and carboxylic groups) are
attached to the surface of CNTs, these groups serve as active sites
for chemically adsorbing with Pt precursors (i.e., MeCpPtMe3) in
gas phase, thus forming the first atomic Pt layer.
The schematic diagram for growing the second Pt layer deco-
rated to sidewalls of CNTs is illustrated in Fig. 3. Basically, one ALD
cycle is composed of four steps including (i) MeCpPtMe3 pulse, (ii)
system purge, (iii) oxygen pulse, and (iv) system purge. The ALD
cycle involves two self-limiting chemical reactions [11,12]:
3.2. Crystalline ALD-Pt and proposed growth mechanism
Fig. 2 depicts the typical XRD pattern of the ALD-Pt nanoparti-
cles over CNTs, confirming the face-centered cubic (fcc) Pt crystal.
The representative diffraction peaks at 39.8◦, 46.2◦, and 67.5◦ can be
observed, assigning to Pt (1 1 1), (2 0 0), and (2 2 0), respectively. The
Pt grains are found to show the (1 1 1) plane with the highest inten-
sity. Herein the intensity ratios of (1 1 1)/(2 2 0) and (2 0 0)/(2 2 0)
are estimated to be 4.91 and 2.64, respectively. This can be ascribed
to the fact that the growth of the ALD-Pt nanoparticles on CNTs
mostly favors a specific crystallographic direction (i.e., Pt (1 1 1)
plane), indicating an anisotropic growth of Pt crystals. Electron dis-
persive X-ray spectrometer (EDS) was applied to clarify the crucial

104
T. Shu et al. / Electrochimica Acta 75 (2012) 101–107
increase with ALD cycle number. Data determined by TGA anal-
Pt
C
ysis, the amounts of ALD-Pt were found to have 0.26, 0.40 and
0.58 mg cm−2 after 100, 150 and 200 cycles, respectively. The
weight loading of Pt deposits as a linear function of the ALD cycle
number can be obtained, and the growth rate is approximately
0.0031 mg cm−2 cycle−1 under the ALD operating conditions. This
information is beneficial for an estimation of Pt loading on other
carbon-based supports, based on the similar surface chemistry.
Accordingly, the linearity relation proves the presence of self-
limiting reaction during the ALD process.
(111)
(200)
(100)
(220)
3.3. Electrochemical behavior of the ALD-Pt electrode
Fig. 4(a) shows cycle voltammograms for ALD-Pt electrodes in
1 M H2SO4 at different sweep rates (5, 15, 25, 35, and 45 mV s−1).
The CV scans start from an open-circuit potential and sweep
within the entire potential region between −0.2 V and 0.8 V vs.
Ag/AgCl. Generally, the ALD-Pt catalyst shows well-defined hydro-
gen adsorption/desorption peaks in the potential region[18,19], i.e.,
hydrogen was adsorbed in the cathodic direction and then oxidized
in the anodic direction. Additionally, an increase in the peak cur-
rents is nearly proportional to the scanning rates, indicating that (i)
a good dispersion of ALD-Pt catalysts over CNTs leads to fast redox
kinetic, and (ii) the presence of CNTs not only acts as the catalyst
support but also allows fast electron transportation and provides
available adsorptive sites for hydrogen storage and double-layer
formation [20].
The ESA (=Qcat/QPt) value is one crucial index in evaluat-
ing catalytic activity of Pt nanoparticles, determined through
the charge transfer due to hydrogen adsorption/desorption (i.e.,
Pt + H+ + e− ↔ Pt − Hads) on Pt catalysts [21]. Here Qcat represents
specific charge transfer contributed from Pt catalysts on carbon
support, and QPt represents an electrical charge associated with
30
40
50
60
70
80
Scattering angle (º)
Fig. 2. Typical XRD pattern of ALD-Pt nanoparticles on the surface of CNTs.
The 1st half cycle:
(R1)Pt–O ∗ (ads) + MeCpPtMe3 (g) → Pt–MeCpPtMe2∗ (ads)
+ CO2 (g) + H2O (g) + other products
The 2nd half cycle:
Pt–MeCpPtMe2∗ (ads) + O2 (g) → Pt–Pt–O ∗ (ads) + CO2 (g)
+ H2O (g) + other products
(R2)
where “*” represents active surface species and Pt–O* represents
oxygen chemically adsorbed on the platinum surface. Due to the
limiting reaction steps, the amount of Pt nanoparticles tends to
Fig. 3. Schematic diagram of the growth of ALD-Pt nanoparticles on the surface of CNTs, involving (a) 1st half cycle and (b) 2nd half cycle.

T. Shu et al. / Electrochimica Acta 75 (2012) 101–107
105
40
20
0
45 mV s-1
35 mV s-1
(a)
25 mV s-1
15 mV s-1
5 mV s-1
-20
-40
-0.2
0
0.2
0.4
0.6
0.8
Potential (V vs. Ag/AgCl)
30
20
(b)
1st cycle
1000th cycle
10
0
Fig. 5. (a) Nyquist impedance plots of ALD-Pt electrode in 1 M H2SO4 before and
after potential cycling within the entire frequency range. The inset is high-frequency
Nyquist plots for ALD-Pt electrode. (b) Proposed equivalent circuit for the impedance
behavior of ALD-Pt electrode.
-10
-20
-30
the rigid architecture by ALD is capable of reducing catalyst loss or
crystal mergence in acidic electrolyte.
-0.2
0
0.2
0.4
0.6
0.8
The ac impedance spectroscopy was further applied to analyze
the electrochemical behavior of the ALD-Pt electrode. Fig. 5(a) illus-
trates the Nyquist plots of ALD-Pt electrode before and after 1000
CV cycles, showing a resemble configuration of impedance spec-
tra; that is, a depressed semicircle within the whole frequency
range. Basically, the configuration consists of (i) an intersection
in the real axis in high-frequency region, (ii) a single depressed
semicircle in the frequency region between 100 kHz and 1 Hz, and
(iii) a slant line with decreasing frequency. As shown in Fig. 5(b),
one equivalent circuit in the frequency region was proposed to
describe the presence of an RC loop involving a constant phase
element (CPE) in parallel with a charge-transfer resistance (RCT).
In the circuit, RS represents the bulk solution resistance, and the
plot transforms to a slant with decreasing frequency, showing the
appearance of Warburg impedance (W) [23]. The CPE can be formu-
lated to Z = 1/[Q(jw)n], where Q and n are the pseudo-capacitance
(with a unit of ꢀ−1 sn cm−2) and the non-homogeneity constant
(dimensionless), respectively [24]. A Z-view software was adopted
to simulate the impedance behavior. Herein the derivation <10% for
each component indicates a fairly good fit for the impedance behav-
ior. The values of n approximate 0.76–0.85, showing the typical
characteristic of a porous electrode [23].
The analysis on the equivalent circuit generally reveals to a
locus showing an equivalent series resistance (RES), which is com-
posed of bulk solution resistance and charge transfer resistance
(i.e., RES = RS + RCT) [25]. The RES values of ALD-Pt catalyst electrode
are estimated to be 3.55 ꢀ and 3.78 ꢀ before and after 1000 CV
cycles, respectively. Such low magnitude of RES value confirms that
the ALD-Pt nanocatalysts display electrochemically activity, lead-
ing to reversibly rapid redox kinetics. After potential cycling, there
is no obvious increase of internal resistance in the ALD-Pt catalyst
electrode, indicating an excellent durability in the electrochemical
Potential (V vs. Ag/AgCl)
Fig. 4. (a) Cyclic voltammetry profiles of ALD-Pt catalyst electrode in 1 M H2SO4 at
different scanning rates of 5, 15, 25, 35, and 45 mV s−1, and (b) Cyclic voltammetry
profiles of ALD-Pt catalyst electrode in 1 M H2SO4 at 30 mV s−1 within 1000 CV cycles.
monolayer adsorption of hydrogen on Pt, which has been estimated
to be 0.21 mC cm−2 [8]. Based on the CV curves at 30 mV s−1, the ESA
value of ALD-Pt catalysts can be estimated to 78.2 m2 g Pt−1, which
is about two times higher than commercial E-TEK Pt catalyst (i.e.,
43.3 m2 g Pt−1) [22]. To inspect the stability, the catalytic activity
of ALD-Pt electrode was investigated with various CV scans at a
scan rate of 30 mV s−1, as shown in Fig. 4(b). The redox current of
ALD-Pt catalysts displays a stable current within 1000 cycles, e.g.,
the ESA value of 70.4 m2 g−1 at the 1000th cycle. An ESA retention of
∼90% was observed after CV cycling, indicating that the ALD-Pt cat-
alyst still maintains a stable activity toward the acidic electrolyte.
The ESA stability of the ALD-Pt catalyst is significantly enhanced in
comparison with the commercial Pt catalyst, e.g., E-TEK: 57% ESA
retention after 1000 cycles [22]. The improved ESA value during the
potential cycling can be attributed to the design of catalyst elec-
trode, i.e., the ALD-Pt catalysts decorated with CNT support. First,
the CNT support offers an anti-corrosion ability in electrolysis reac-
tion, which has been confirmed by one previous study [20]. Second,
a strong connection between the ALD-Pt particles and the sidewall
of CNTs would effectively prevent the Pt migration, which origi-
nated from the chemical adsorption of Pt atoms onto the surface
oxides of CNTs (e.g., C OH and C OOH groups) during the ALD
process. The type of layer-by-layer occupancy tends to strongly
immobilize the Pt sites on the surface of CNTs. The formation of

106
T. Shu et al. / Electrochimica Acta 75 (2012) 101–107
1
0.8
0.6
0.4
0.2
4
3
2
1
0
2400
2200
2000
1800
1600
1400
1200
1000
(a)
ALD-Pt
Commercial Pt
40oC
60oC
80oC
0
1
0
400
800
1200
1600
2000
0
10
20
30
40
50
Current density (mA cm-2)
Time (hr)
3
2
1
0
(b)
Fig. 7. Constant-voltage durability test (at 0.7 V) obtained with MEAs made of ALD-
Pt and commercial Pt (Johnson Matthey) at 60 ◦C for 50 h.
0.8
0.6
0.4
0.2
0
Pt activity, future work is required to optimal ALD parameters and
to synthesize Pt nanoparticles with a size <3 nm.
To evaluate the Pt durability, both MEAs were operated at a con-
stant voltage of 0.7 V for 50 h. As shown in Fig. 7, both MEAs deliver
a current decay at the initial stage (∼1 h) and then maintain a stable
current density within the operating period. After 50 h, the MEAs
equipped with ALD-Pt and commercial Pt catalysts keep at 2155
and 1860 A g Pt−1, respectively. This satisfactory result confirms
that the ALD-Pt catalyst provides an excellent durability in compar-
ison with the commercial Pt catalyst. Again, this can be ascribed to
the chemical bonding between the Pt nanoparticles and the surface
of CNTs through the ALD route, inducing the long-term durability.
On the basis of the results, the ALD technique shows the commercial
feasibility to prepare Pt catalysts on CNT supports, exhibiting high
catalytic activity, low internal resistance, and excellent durability
for PEMFC application.
40oC
60oC
80oC
0
400 800 1200 1600 2000 2400 2800
Current density (mA cm-2)
Fig. 6. Performance curves obtained with MEAs made of (a) ALD-Pt and (b) com-
mercial Pt (Johnson Matthey) anodes at 40, 60, and 80 ◦C.
system. Herein the CNTs are not only a catalyst support but also
a charge transfer media, providing the rigidity for Pt catalysts and
gas diffusion layer (i.e., CP substrate).
4. Conclusions
The catalytic activity and durability of well-dispersed Pt
nanoparticles, prepared using the ALD technique onto the surface
of CNTs, have been investigated in the application of PEMFCs. The
ALD route, using MeCpPtMe3 and oxygen as precursors, was capa-
ble of synthesizing metallic Pt nanoparticles with an average size
of 3–5 nm over the CNT support. One growth mechanism, which
consisted of two self-limiting reaction steps, was proposed to clar-
ify the formation of ALD-Pt catalysts. The CV test and ac impedance
spectroscopy indicated that the ALD-Pt catalyst offer a stable ESA
value (90% retention) and a low RES value (∼3.78 ꢀ) after potential
cycling (>1000 cycles). The analysis of the single cell reflected that
the CNTs decorated with the ALD-Pt catalysts exhibit high power
density (∼2.95 kW g−1 at 80 ◦C) and long-term durability (>50 h).
Accordingly, the ALD approach delivered a promising potential to
synthesize high-performance Pt nanocatalysts onto CNTs, forming
the novel design of catalyst electrode for PEMFCs.
3.4. Performance of single cell
The polarization curves of ALD-Pt and commercial (i.e., Johnson
Matthey) catalyst electrodes in PEMFCs operated at different tem-
peratures of 40, 60, and 80 ◦C were investigated, as shown in Fig. 6. It
is worth noting that the Pt loadings of ALD-Pt and the commercial
electrodes are 0.26 and 0.40 mg cm−2, respectively. These curves
typically consist of three polarization regions: (i) activation polar-
ization, (ii) ohmic polarization, and (iii) concentration polarization.
Accordingly, the profile of ALD-Pt polarization curves is similar to
that of commercial Pt catalyst electrodes. The maximal power den-
sities obtained from the MEAs using ALD-Pt catalysts are calculated
to be 2.25, 2.74, and 2.95 kW g Pt−1 at 40, 60, 80 ◦C, respectively. For
comparison, the ALD-Pt catalyst offers the power efficiencies higher
than the commercial Pt one, i.e., their maximal power densities of
1.62, 1.83, and 1.90 kW g Pt−1 at 40, 60, 80 ◦C, respectively. Because
both single cells were operated under the same conditions, the ALD-
Pt offers an improved catalytic activity compared to the commercial
Pt. This improved power efficiency can be attributed to the config-
uration of the catalyst electrode. That is, well-dispersed Pt particles
over CNTs by the ALD technique serve as active sites for hydrogen
oxidation reactions, raising the Pt utilization. However, to enhance
Acknowledgment
The authors are very grateful for the financial support from the
National Science Council of the Republic of China under the con-
tracts NSC 100-2120-M-155-001 and NSC 100-2221-E-155-031.

T. Shu et al. / Electrochimica Acta 75 (2012) 101–107
107
References
[13] T. Aaltonen, M. Ritala, V. Sammelselg, M. Leskela, Journal of the Electrochemical
Society 151 (2004) G489.
[14] X. Jiang, S.F. Bent, Journal of the Electrochemical Society 154 (2007)
D648.
[1] C. Wang, M. Waje, X. Wang, J.M. Tang, R.C. Haddon, Y. Yan, Nano Letters 4 (2003)
345.
[2] J.H. Wee, K.Y. Lee, S.H. Kim, Journal of Power Sources 165 (2007) 667.
[3] A. Miura, H. Wang, B.M. Leonard, H.D. Abruna, F. DiSalvo, Chemistry of Materials
21 (2009) 2661.
[4] M.H. Shao, Journal of Power Sources 196 (2011) 2433.
[5] X.P. Gao, Y. Lan, G.L. Pan, F. Wu, J.Q. Qu, D.Y. Song, P.W. Shen, Journal of Elec-
trochemical and Solid-State Letters 4 (2001) A173.
[6] H.S. Kim, H. Lee, K.S. Han, J.H. Kim, M.S. Song, M.S. Park, J.Y. Lee, J.K. Kang, Journal
of Physical Chemistry B 109 (2005) 8983.
[7] A.K.M. Fazle Kibria, Y.M. Mo, K.S. Park, K.S. Nahm, M.H. Yun, International Jour-
nal of Hydrogen Energy 26 (2001) 823.
[8] C.T. Hsieh, J.Y. Lin, J.L. Wei, International Journal of Hydrogen Energy 34 (2009)
685.
[9] C.T. Hsieh, J.Y. Lin, Journal of Power Sources 188 (2009) 347.
[10] T. Sakae, M. Hiroshi, M. Hideki, T. Eishi, K. Masahiro, Journal of the Electrochem-
ical Society 155 (2008) B929.
[11] Y. Zhou, D.M. King, X.H. Liang, J.H. Li, A.W. Weimer, Applied Catalysis B: Envi-
ronmental 101 (2010) 54.
[15] J.S. King, A. Wittstock, J. Biener, S.O. Kucheyev, Y.M. Wang, T.F. Baumann, S.K.
Giri, A.V. Hamza, M. Baeumer, S.F. Bent, Nano Letters 8 (2008) 2405.
[16] C. Liu, C.-C. Wang, C.-C. Kei, Y.-C. Hsueh, T.-P. Perng, Small 5 (2009) 1535.
[17] C.K. Poh, S.H. Lim, H. Pan, J. Lin, J.Y. Lee, Journal of Power Sources 176 (2008)
70.
[18] Y. Xing, Journal of Physical Chemistry B 108 (2004) 19255.
[19] Y.L. Yao, Y. Ding, L.S. Ye, X.H. Xia, Carbon 44 (2006) 61.
[20] C.T. Hsieh, W.M. Hung, W.Y. Chen, International Journal of Hydrogen Energy 35
(2010) 8425.
[21] S. Kim, S.J. Park, Journal of Power Sources 159 (2006) 42.
[22] S. Sun, G. Zhang, D. Geng, Y. Chen, R. Li, M. Cai, X. Sun, Angewandte Chemie
International Edition 50 (2011) 422.
[23] B.E. Conway, Electrochemical Supercapacitors: Scientific Fundamentals and
Technological Applications, Kluwer Academic/Plenum Publishers, New York,
1999.
[24] Y.H. Chen, K.C. Huang, J.G. Chen, R. Vittal, K.C. Ho, Electrochimica Acta 56 (2011)
7999.
[25] S. Koh, J. Leisch, M.F. Toney, P. Strasser, Journal of Physical Chemistry B 111
(2007) 3744.
[12] J.H. Li, X.H. Liang, D.M. King, Y.B. Jiang, A.W. Weimer, Applied Catalysis B:
Environmental 97 (2010) 220.