Characterization and application of electrodeposited Pt, Pt/Pd, and Pd catalyst structures for direct formic acid micro fuel cells

Article abstract of DOI:10.1016/j.electacta.2005.02.018

In this work, we study the preparation, structural characterization, and electrocatalytic analysis of robust Pt and Pd-containing catalyst structures for silicon-based formic acid micro fuel cells. The catalyst structures studied were prepared and incorporated into the silicon fuel cells by a post CMOS-compatible process of electrodeposition, as opposed to the more common introduction of nanoparticle-based catalyst by ink painting. Robust, high surface area, catalyst structures consisting of pure Pt, pure Pd, and Pt/Pd = 1:1 were obtained. In addition, Pt/Pd catalyst structures were obtained via spontaneous deposition on the electrodeposited pure Pt structure. The catalyst structures were characterized electrochemically using cyclic voltammetry and chronoamperometry. All Pd-containing catalyst structures facilitate formic acid oxidation at the lower potentials and deliver higher oxidation currents compared to pure Pt catalyst structures. Fuel cells of these catalyst structures show that pure Pd catalyst structures on the anode exhibit the highest peak power density, i.e. as high as 28.0 mW/cm². The MEMS compatible way of catalyst electrodeposition and integration presented here has yielded catalyst structures that are highly active towards formic acid oxidation and are sufficiently robust to be compatible with post-CMOS processing.

Full text of DOI:10.1016/j.electacta.2005.02.018

Electrochimica Acta 50 (2005) 4674–4682
Characterization and application of electrodeposited Pt, Pt/Pd, and Pd
catalyst structures for direct formic acid micro fuel cells
R.S. Jayashreea, J.S. Spendelowa, J. Yeomb, C. Rastogia, M.A. Shannonb, P.J.A. Kenisa,b,∗
a
Department of Chemical and Biomolecular Engineering, University of Illinois at Urbana Champaign, 600 S. Mathews Avenue, Urbana, IL 61801, USA
Department of Mechanical and Industrial Engineering, University of Illinois at Urbana Champaign, 600 S. Mathews Avenue, Urbana, IL 61801, USA
b
Received 29 December 2004; received in revised form 21 February 2005; accepted 27 February 2005
Available online 23 March 2005
Abstract
In this work, we study the preparation, structural characterization, and electrocatalytic analysis of robust Pt and Pd-containing catalyst
structures for silicon-based formic acid micro fuel cells. The catalyst structures studied were prepared and incorporated into the silicon fuel
cells by a post CMOS-compatible process of electrodeposition, as opposed to the more common introduction of nanoparticle-based catalyst
by ink painting. Robust, high surface area, catalyst structures consisting of pure Pt, pure Pd, and Pt/Pd = 1:1 were obtained. In addition,
Pt/Pd catalyst structures were obtained via spontaneous deposition on the electrodeposited pure Pt structure. The catalyst structures were
characterized electrochemically using cyclic voltammetry and chronoamperometry. All Pd-containing catalyst structures facilitate formic acid
oxidation at the lower potentials and deliver higher oxidation currents compared to pure Pt catalyst structures. Fuel cells of these catalyst
structures show that pure Pd catalyst structures on the anode exhibit the highest peak power density, i.e. as high as 28.0 mW/cm2. The MEMS
compatible way of catalyst electrodeposition and integration presented here has yielded catalyst structures that are highly active towards
formic acid oxidation and are sufficiently robust to be compatible with post-CMOS processing.
© 2005 Elsevier Ltd. All rights reserved.
Keywords: Direct formic acid fuel cell; Pt/Pd catalysts; Spontaneous deposition; Micro fuel cell; Electrodeposition
1. Introduction
structures that (i) are compatible with post-CMOS process-
ing; (ii) have high surface area; (iii) exhibit sufficient catalytic
activity towards the fuel used; and (iv) can withstand oper-
ation conditions including temperature changes, shock, and
vibration. In this work, we study the preparation and per-
formance of robust high-surface-area Pt and Pd-containing
catalyst structures for silicon-based formic acid micro fuel
cells that fulfill the first three and possibly also the fourth of
the aforementioned requirements.
Methanol and formic acid are promising liquid fuels for
micro fuel cells. Several miniaturized proton exchange mem-
brane fuel cells (PEM-FC) running on methanol or on formic
acid have been reported [3–7,8]. Kelley et al. were the first to
report methanol/air miniaturized fuel cells fabricated using
MEMS and microfabrication techniques [7]. While methanol
has been a fuel of choice for PEM-FCs for several years, Rice
et al. have recently shown the viability of formic acid in a
cm-scale fuel cell [8]. Compared to methanol, formic acid
Micro fuel cells are emerging as promising high energy
density power sources for portable applications [1]. Fuel
cells are capable of delivering higher energy densities than
the most competitive rechargeable batteries, like Li-ion and
Ni–Cd. One of the key challenges for integration of high en-
ergy density fuel cells in miniature electronic applications
is the development of fabrication procedures that are com-
patible with MEMS and/or post-CMOS processes. Along
these lines, we have reported earlier on the successful fabrica-
tion of silicon-based fuel cell membrane electrode assemblies
(MEAs) that efficiently integrate flow field channels, current
collectors (Au), and catalysts, on a single chip [2]. However,
several challenges still remain in developing robust catalyst
∗
Corresponding author. Tel.: +1 217 265 0523; fax: +1 217 333 5052.
E-mail address: kenis@uiuc.edu (P.J.A. Kenis).
0013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.02.018

R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
4675
has faster room temperature kinetics and has less of a ten-
dency to crossover, but the energy density of formic acid is
only one-third that of methanol [9].
There are two proposed pathways for formic acid or
methanol oxidation, namely the ‘direct pathway’ and the ‘CO
pathway’. In the ‘CO pathway’, formic acid or methanol is
oxidized to an intermediate like CO, which bonds strongly
to the Pt surface and impedes the adsorption of fuel species
(catalyst poisoning). For methanol electro-oxidation, COpoi-
soning is reduced by the use of Pt/Ru catalysts [10]. Formic
acid oxidation on Pt surfaces occurs through both the ‘direct
pathway’ and the ‘CO pathway’ [8,11–13]. Lu et al. have
shown that formic acid oxidation typically proceeds through
the preferred ‘direct pathway’ on Pd-modified Pt surfaces
[11], and the suppression of CO formation on Pd-modified
Pt has been confirmed by FT-IR studies [14].
Generally, Pt and Pt-based metal nanoparticles with high
surface areas are used as catalysts in PEM-FCs. These
nanoparticles are typically painted onto the current collec-
tors with an ink containing the nanoparticles as well as a
certain percentage of Nafion solution to act as the binder
upon evaporation of the solvent [15,16]. The presence of
Nafion as a binder in catalyst inks provides good adhesion
as well as good protonic conductivity, but hampers the elec-
tronic conductivity of the matrix. McGovern et al. reported a
13% decrease in surface area and a 40% decrease in oxida-
tion current density resulting from inclusion of Nafion in Pt
black nanoparticle-based inks [17]. This method of catalyst
application thus inhibits the complete utilization of the cat-
alyst. Alternatively, Pt-based metal catalysts can be applied
to carbon-based high surface area supports via electrodepo-
sition [7,18,19]. Here we obtain high surface area catalyst
structures on smooth supports by electrodeposition at high
deposition rates leading to rough deposits [20,21]. A key ad-
vantage of using electrodeposition as the catalyst deposition
technique is its compatibility with MEMS and post-CMOS
processing, enabling more straightforward integration of mi-
cro fuel cells in silicon-based microelectronic applications.
In this paper, we report on the preparation, structural
characterization, and the electrochemical activity of elec-
trodeposited dendritic Pt (pure Pt), spontaneously deposited
Pd on dendritic Pt (Pt/Pd-SD), co-electrodeposited Pt/Pd
(Pt/Pd = 1:1), and electrodeposited Pd (pure Pd) catalyst
structures towards formic acid oxidation. The performance
of silicon-based micro fuel cells equipped with these catalyst
structures will also be studied and compared.
Fig. 1. Schematic diagram of micro fabricated Si-based MEA. In this paper,
the cathode catalyst is always pure Pt, and the anode catalyst is pure Pt,
Pt/Pd = 1:1, pure Pd, or Pt/Pd-SD.
˚
front side with a 1000 A Au layer by dc magnetron sputter-
ing (∼10−2 Torr of Ar background pressure). Subsequently,
this Au layer, the eventual current collector, is further pat-
terned into circular grids using a multistep procedure involv-
ing photolithography and liftoff, as reported previously [22].
In brief, the same side was then covered via spincoating with
a polyimide layer (PMDA-ODA PI2808, HD Microsystem)
that was further patterned and etched using deep reactive ion
etching (Plasma-Therm SLR 770) from the top and the bot-
tom to yield the 50-␮m thick Si-grids with 100-␮m square
holes. Si-MEAs (Fig. 1) were prepared by bonding two Si-
grids to a Nafion-112 membrane (Fuel Cell Scientific, Stone-
ham, MA) in a sandwich configuration. Bonding of two Si-
grids to the Nafion membrane was achieved using a hot press-
ing method in an anodic bonder (EV, 501 series) at 120 ◦C
and a pressure of ∼200 N/cm2. A polyimide adhesion pro-
moter (VM652, HD Microsystem) was employed to enhance
the adhesion between the Nafion membrane and the PMDA-
ODA-covered surface of the Si-grids. A detailed description
of the fabrication of these Si-MEAs can be found elsewhere
[2,22].
2.2. Pt and Pt/Pd deposition
2.2.1. Pt deposition
High surface area, dendritic Pt catalyst structures (pure Pt)
were prepared by electrodeposition on Au-coated Si-grids,
either potentiostatically (−1, −2, or −5 V) versus a Ag/AgCl
reference electrode (in 3 M NaCl, BAS, West-Lafayette, IN),
or galvanostatically (−230, −1135, or −1365 mA/cm2) in a
0.08 M H2PtCl6·6H2O solution (Alfa Aesar) for 120 s, using
an Autolab PGSTAT-30 potentiostat.
2.2.2. Spontaneous Pd deposition
The spontaneously deposited Pd on dendritic Pt (Pt/Pd-
SD) sample was obtained by depositing Pd onto the high sur-
face area dendritic Pt structures (pure Pt, Section 2.2.1) by
spontaneous deposition as described elsewhere [11]. In short,
the Pt-covered Si-MEAs were first cleaned electrochemically
in 0.1 M H2SO4 (GFS Chemicals, Powell, OH) by cycling
between 0 and 1.5 V versus RHE. Then the electrodes were
placed in a 5 mM Pd(NO3)2 (Aldrich, 10 wt.% solution in
10 wt.% HNO3) + 0.1 M H2SO4 solution for 5 min, followed
by thorough rinsing with Milli-Q water (Barnsted E-pure wa-
ter, 18 Mꢀ cm). These electrodes were electrochemically an-
nealed by cycling five times between 0.05 and 0.95 V in 0.1 M
2. Experimental
2.1. Microfabrication of Si-MEA assemblies
Silicon-based membrane electrode assemblies (Si-MEAs)
were prepared using standard microfabrication processes. A
100 mm N-doped double-sided polished Si-wafer (Silicon
Quest, 500 ␮m thick, ꢀ1 0 0ꢁ oriented) was covered on the

4676
R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
H2SO4. This spontaneous deposition procedure was repeated
five times to yield Pd-decorated dendritic Pt catalyst struc-
tures (Pt/Pd-SD).
6060LV microscope). High-resolution SEMs were obtained
using a Hitachi S-4700 microscope.
2.4.2. Electrocatalytic analysis
2.2.3. Co-deposition of Pt and Pd
Electro-oxidation of 0.1 M formic acid (Acros, 96%) in
0.1 M H2SO4 was studied using cyclic voltammetry and
chronoamperometry.
Pt and Pd co-deposited catalyst structures (Pt/Pd = 1:1)
were obtained by potentiostatic co-electrodeposition
at −2 V versus Ag/AgCl for 120 s using a 0.04 M
H2PtCl6·6H2O + 0.04 M PdCl2 (5 wt.% solution in 10 wt.%
HCl solution, Aldrich).
2.4.2.1. Cyclicvoltammetry. Electrochemicallycleanedcat-
alyst structures were cycled between 0.05 and 1.25 V versus
RHE in 0.1 M H2SO4 + 0.1 M HCOOH solution at a scan rate
of 10 mV/s.
2.2.4. Pure Pd deposition
High surface area Pd catalyst structures (pure Pd) were ob-
tained by potentiostatic deposition at −2 V versus Ag/AgCl
for 120 s using a 0.08 M PdCl2 solution.
2.4.2.2. Chronoamperometry. Potential steps between 0.95
and 0.01 V were applied for 1 s and these steps were repeated
four times to ensure complete removal of surface-bound CO.
After these cleaning steps the electrodes were held at 0.2, 0.3,
or 0.4 V versus RHE for 600 s and the current was recorded
as a function of time.
2.3. Surface area measurement of catalyst structures
The true surface areas of the Pt catalyst structures were
determined from the hydrogen adsorption/desorption or CO
stripping charge measured in 0.1 M H2SO4. The true surface
areas of Pt/Pd and pure Pd catalysts were determined using
CO stripping voltammetry. The geometric surface area of the
Si-grids used to calculate current densities was 0.44 cm2.
2.5. Pd quantification
2.5.1.1. Spontaneous deposition samples
After performing all electrochemical studies of formic
acid electro-oxidation on Pt/Pd-SD, the catalyst structure was
immersed in 1 mM NaI (Aldrich) for 5 min and then cycled in
1 M H2SO4 from 0.05 to 1.2 V versus RHE. The charge cor-
responding to the peak from 0.9 to 1.25 V due to anodic strip-
ping of Pd and a charge to surface area value of 452 ␮C/cm2
corresponded to a monolayer of Pd on Pt surface was used
to calculate the number of Pd monolayers on the Pt surface
[24].
2.3.1. Hydrogen adsorption/desorption
Pure Pt catalyst structures were electrochemically cleaned
in 0.1 M H2SO4 by cycling from 0.05 to 1.5 V versus RHE at
a scan rate of 500 mV/s, followed by cycling in 0.1 M H2SO4
between 0.05 and 0.95 V versus RHE at a scan rate of 5 or
10 mV/s. The true surface area of the Pt catalyst was deter-
mined by calculating the charge associated with the hydrogen
adsorption/desorption peaks using a charge to surface area
value for hydrogen adsorption/desorption on polycrystalline
Pt of 210 ␮C/cm2 [23].
2.5.1.2. Co-deposition samples
The atomic ratios of Pt and Pd in the co-deposited cata-
lyst structures (Pt/Pd = 1:1) were determined during scanning
electron microscopy (JEOL 6060LV SEM) with an energy
dispersive X-ray analyzer (EDX, Oxford). The average com-
position of 15 different spots is reported. All measurements
were within 10 at.% deviation from the average.
2.3.2. CO stripping
The Si-grids with Pd-containing catalyst structures were
placed in 0.1 M H2SO4, and held at 0.1 V versus Ag/AgCl
while the solution was bubbled with carbon monoxide gas
(Matheson, 99.9% purity, Irving, TX) for 5 min. Then the
solution was purged with N2 (SJ Smith Welding supply,
99.9% purity, Davenport, IA) for 5 min to remove dissolved
CO. Next, the electrodes were cycled between 0.05 and
1.2 V versus RHE, and the true surface areas of the cata-
lyst structures were determined by calculating the charge
corresponding to the CO stripping peak using a charge
to surface area value for CO stripping on Pt or Pt/Pd of
420 ␮C/cm2 [23].
2.6. Fuel cell testing
The Si-grids with the desired catalyst (pure Pt, Pt/Pd = 1:1,
Pt/Pd-SD, or pure Pd) as the anode, and a Si-grid with pure
Pt deposited at −2 V versus Ag/AgCl for 120 s from 0.08 M
H2PtCl6·6H2O solution as the cathode were bonded to a
Nafion-112 membrane in a sandwich configuration to obtain
silicon-based membrane electrode assemblies (Si-MEAs)
with different anode catalyst structures (see also Section 2.1).
For testing, each Si-MEA was mounted on a test fixture as
described earlier [22]. Fuel cell testing was performed by
placing 2 ml of 10 M formic acid directly on the anode, and
the cathode was exposed to a stream of oxygen at a flow rate
of 10 sccm. The polarization curves were recorded by main-
2.4. Characterization of the catalyst structures
2.4.1. Morphology
The catalyst-covered Si-grids were mounted on conduc-
tive carbon tape for scanning electron microscopy (JEOL

R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
4677
taining a constant cell potential, and measuring the current in
the circuit after waiting for the cell to reach steady state.
were performed using the Pt catalyst structure deposited at
−2 V, as the sample deposited at −5 V, while exhibiting a
higher surface area, was mechanically unstable due to flak-
ing. The co-deposited Pt/Pd catalyst structure (Pt/Pd = 1:1)
and the Pd catalyst structure (pure Pd) were obtained by elec-
trodeposition at −2 V from appropriate solutions (see Section
2.2). The Pt/Pd-SD catalyst structure was obtained by dec-
orating pure Pt catalyst structures with Pd nanoislands by
spontaneous deposition (see Section 2.2) [11].
3. Results and discussion
3.1. Catalyst structure preparation on Si-grids
In previous work we reported on silicon-based membrane
electrode assemblies (Si-MEAs, Fig. 1), comprised of two
microfabricated Si-grids bonded to a Nafion-112 membrane
[22]. In these Si-MEAs the flow-field, the current collector,
and the catalyst structures are integrated on a single chip.
Here, we integrate different catalyst structures on these Si-
grids via electrodeposition. High surface area dendritic Pt cat-
alyst structures (pure Pt) were obtained by electrodeposition
on Au-covered Si-grids at different potentials and/or current
densities (Table 1) to identify conditions yielding mechani-
cally stable structures with high surface areas. Further studies
3.2. Structural characterization of Pt, Pt/Pd, and Pd
catalyst structures
The morphologies of the different catalyst structures were
studied using scanning electron microscopy. Catalyst struc-
tures deposited at −1 V show a smooth morphology in SEM
(Fig. 2a), while the catalyst structures deposited at −2 V ex-
hibited a dendritic morphology (Fig. 2b and c). The cata-
lyst structure deposited at −5 V exhibited dendritic morphol-
Table 1
Deposition conditions, surface areas, and current densities of Pt and Pd-containing catalyst structures
Catalyst structures
Deposition potential (V vs. Ag/AgCl)
True surface areaa (m2/g)
Current density (␮A/cm2)b
0.2 V vs. RHE
0.3 V vs. RHE
0.4 V vs. RHE
Pure Pt
Pure Pt
Pure Pt
Pt/Pd = 1:1
Pure Pd
Pt/Pd-SD
a
−1
−2
−5
−2
−2
−2
10.2
16.9
19.3
10.8
8.9
0.02
0.05
0.09
2.52
5.45
7.16
0.20
0.42
0.37
3.99
12.20
9.17
0.98
2.30
1.40
5.22
11.70
9.00
16.9
Determined using cyclic voltammetry.
b
Steady-state current densities (from chronoamperometry) in formic acid solution normalized to the true surface area of Pt catalyst structures.
Fig. 2. Scanning electron micrographs of the pure Pt catalyst structures electrodeposited at (a) −1 V and (b–d) −2 V vs. Ag/AgCl, at different magnifications.

4678
R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
Fig. 3. Scanning electron micrographs of Pt/Pd = 1:1 catalyst structures codeposited at −2 V vs. Ag/AgCl at different magnifications. The existence of spherical,
and wedge-shaped features can be seen in (b) and at higher magnification in (c and d) and (e and f).
ogy similar to catalyst structures deposited at −2 V. Such
metallic dendritic morphologies obtained by high current or
high potential electrodeposition have been reported by oth-
ers [20,21]. We have used similar dendritic Pt structures as
the catalysts in our previous work on Si-MEAs [22]. High-
resolution images of dendritic Pt deposited at −2 V show
the existence of nanopores (Fig. 2d). The presence of these
nanoporous features explains the high surface area deter-
mined for these catalyst structures [18]. Using cyclic voltam-
metry, we measured surface areas increasing from 10.2 to
19.3 m2/g for the pure Pt catalyst structures deposited at in-
creasingly negative potentials (Table 1). This trend of increas-
ing surface area corresponds to the increase in dendritic mor-
phological character of the structures at increasingly nega-
tive potentials. Increasingly negative currents had a less pro-
nounced effect; pure Pt catalyst structures deposited galvano-
statically at −230, −1135, or −1365 mA/cm2 resulted in sur-
face areas of 16.6, 16.8, and 14.5 m2/g, respectively.
(Fig. 3b). High-resolution images indicate a granular nature
of the spherical and dendritic particles (Fig. 3d and f). Energy
Dispersive X-ray analysis (EDX) shows tht the ratio of Pt to
Pd in this sample is 1:1. The surface area of co-deposited
Pt/Pd = 1:1 was determined to be 10.8 m2/g.
The pure Pd catalyst structures exhibit a variety of spher-
ical features (Fig. 4). The surfaces of the larger spherical
features that protrude further into the solution are made up
of small (10–20 nm) round crystallites (Fig. 4d), while the
smaller spherical entities in the recessed areas between them
show fibrous surface features (Fig. 4c). These pure Pd have
a surface area of 8.9 m2/g.
The Pt/Pd-SD sample (not shown) did not exhibit any mor-
phological differences compared to pure Pt samples (Fig. 2b),
as expected. Using selective and quantitative anodic stripping
of Pd, we determined that 0.23 monolayers of Pd islands were
distributed on the Pt dendrites [24].
The difference in morphology of the pure Pt and the pure
Pd catalyst structures is striking: The electrodeposition of
pure Pt shows a nominal grain size of approximately 2.5 ␮m
SEM images of the Pt/Pd = 1:1 catalyst structure (Fig. 3)
reveal the presence of spherical and dendritic features

R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
4679
Fig. 4. Scanning electron micrographs of pure Pd catalyst electrodeposited at −2 V vs. Ag/AgCl. The existence of fibrous and smooth particles can be seen in
(b) and at higher magnification in (c and d).
with a roughened surface with small granules of less than
100 nm, while the electrodeposition of pure Pd shows a larger
variation of grain size ranging from less than 2.5 ␮m to more
than 5 ␮m but all with a very smooth surface. At this point, we
canonlyspeculateabouttheoriginofthisdifference, sincethe
morphology of electrodeposited catalyst structures depends
on a number of deposition parameters including the deposi-
tion potential and current densities (as studied for pure Pt),
the duration, additives, and the chemical properties includ-
ing differences in diffusivities, of the electrodeposited metal.
Varying the potential and current we identified optimum de-
position conditions for pure Pt of −2 V versus Ag/AgCl (Sec-
tion 3.2). We used these conditions for the preparation of the
other catalyst structures, including pure Pd. Explaining the
origin of the different morphologies between the pure Pt and
pure Pd catalyst structures and identifying critical parame-
ters to control grain size for different catalysts would require
further study of the deposition parameter space.
3.3. Electrocatalytic analysis
3.3.1. Cyclic voltammetry (CV)
Fig. 5 shows the forward (a) and reverse (b) scans of
the four different catalyst structures of pure Pt, Pt/Pd-SD,
Pt/Pd = 1:1, and pure Pd in a 0.1 M HCOOH + 0.1 M H2SO4
solution. For the pure Pt catalyst structures, the first anodic
peak at 0.6 V corresponds to formic acid oxidation, while the
second peak at 0.92 V can be attributed to CO oxidation and
Fig. 5. (a) Forward and (b) reverse scans of cyclic voltammograms of
pure Pt, Pt/Pd = 1:1, pure Pd, and Pt/Pd-SD catalyst structures in 0.1 M
formic acid oxidation on sites that were previously blocked
by CO. At higher potentials formic acid oxidation is deacti-
HCOOH + 0.1 M H2SO4. The current densities are normalized to the true
surface area of the catalyst structures.

4680
R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
vated as a result of surface oxidation. The peak at 0.58 V in
the reverse scan is due to the oxidation of formic acid after
reduction of Pt oxides.
The formic acid oxidation peak shifts to lower values for
the Pt/Pd = 1:1, Pure Pd, and Pt/Pd-SD samples compared to
the pure Pt sample. This shift suggests that the Pt/Pd = 1:1,
pure Pd, and Pt/Pd-SD catalyst structures exhibit better cat-
alytic activity towards formic acid oxidation than pure Pt.
The trends observed for the electrodeposited catalyst struc-
tures studied here are in agreement with previous formic
acidelectro-oxidationstudiesonnanoparticle-basedcatalysts
[11].
3.3.2. Chronoamperometry (CA)
CA data of the pure Pt, Pt/Pd = 1:1, pure Pd and Pt/Pd-SD
catalyst structures (Fig. 6a–c) were recorded by maintaining
the electrode at a constant potential of 0.2, 0.3 or 0.4 V ver-
sus RHE in a solution containing 0.1 M HCOOH and 0.1 M
H2SO4. At an applied potential of 0.2 V versus RHE (Fig. 6a),
thePt/Pd-SDsampleexhibitsthehighestcurrentdensitycom-
pared to the pure Pt, Pt/Pd = 1:1, and pure Pd samples. In
this experiment, the current densities from the Pd-containing
samples are about two orders of magnitude higher than the
current density measured for the pure Pt catalyst structure
at 0.2 V (Table 1). Moreover, at potentials of 0.3 and 0.4 V
versus RHE, the pure Pd catalyst structure delivers a current
density that is, respectively, 30 and five times larger than that
for the pure Pt catalyst structure.
Formic acid oxidation can progress through two parallel
pathways: a ‘direct pathway’ in which formic acid is oxi-
dized directly into CO2 and a ‘CO pathway’, which goes
through the CO intermediate [11]. Higher potentials are nec-
essary to oxidize the surface-bound CO, which leads to a
decrease in cell potential in fuel cells [25]. In the case of
Pd-containing catalyst structures, formic acid oxidation oc-
curs on Pd sites through the ‘direct pathway’, without the
formation of surface-poisoning CO intermediates [11]. The
chronoamperometric data of the Pd-containing catalyst struc-
tures studied here (Fig. 6) confirm the higher electroactiv-
ity towards formic acid oxidation: all Pd-containing catalyst
structures exhibit higher steady-state currents than the pure Pt
catalyst structures. The higher current densities from Pt/Pd-
SD catalyst structures in comparison to Pt/Pd = 1:1 are most
likely due to the random distribution of Pt and Pd atoms
in the co-deposited Pt/Pd = 1:1 sample, whereas Pd nanois-
lands form on the Pt features in the Pt/Pd-SD catalyst struc-
tures [11]. On these Pd nanoislands formic acid oxidation
occurs via the ‘direct pathway’, leading to higher current
densities.
Fig. 6. Chronoamperometric data from pure Pt, Pt/Pd = 1:1, pure Pd, and
Pt/Pd-SD catalyst structures studied at (a) 0.2 V, (b) 0.3 V, and (c) 0.4 V vs.
RHE in 0.1 M HCOOH + 0.1 M H2SO4. The current densities are normalized
to the true surface area of the catalyst structures.
increases as a result of the increase in surface area of catalyst
structures deposited at increasingly negative potentials, −1
to −5 V.
An increase in current density in formic acid oxidation can
be due to differences in catalyst composition as explained
in the previous paragraphs. In addition, an increase in the
surface area of the catalyst structures will lead to an increase
in the current densities measured in chronoamperometry. For
example, the steady-state current densities normalized to the
true surface area for the different pure Pt structures (Table 1)
3.4. Fuel cell testing
To test these catalyst structures in actual fuel cells, we
assembled Si-MEAs (Fig. 1) comprised of Si-grids with pure
Pt, Pt/Pd = 1:1, Pt/Pd-SD, and pure Pd catalyst structures on
the anode side, and Si-grids with pure Pt catalyst structures

R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
4681
these trends are all in full agreement with trends observed for
fuel cells using nanoparticle-based catalysts [26,27].
4. Conclusions
High surface area Pt, Pd, and Pt/Pd catalyst structures were
prepared by electrodeposition at large overpotentials, as well
asthroughspontaneousdepositionofPdonpurePtstructures.
The resulting structures provide high surface area electrocat-
alysts on microfabricated silicon substrates via MEMS com-
patible preparation methods. The activity of these catalyst
structures for formic acid oxidation was studied using elec-
trochemical techniques, and improved activity toward formic
acid oxidation was observed as a result of increased surface
areas and/or different chemical composition of the catalyst
structures. Catalyst structures with dendritic as well as other
high surface area morphologies exhibited increased activity
for formic acid electro-oxidation due to their larger number
of active sites. Similarly, Pd-containing catalysts were found
to be significantly more active than the pure Pt catalyst struc-
tures. In chronoamperometry at low anode potentials (which
corresponds to high cell potentials in a fuel cell), Pt structures
modified by spontaneously deposited Pd (Pt/Pd-SD) were
found to be the most active electrocatalysts, while pure Pd
structures showed higher activity at higher anode potentials,
i.e. at lower cell potentials. Pure Pd catalyst structures may
thus be a good choice for fuel cells continuously operating
under high load.
Parallel fuel cell testing studies using formic acid as the
fuel confirmed that cells constructed with Pd anodes gave the
highest power density, although at the highest fuel cell po-
tentials the performance of cells containing Pt/Pd-SD anodes
was superior. Note that the use of formic acid, rather than
methanol, in these fuel cells has several advantages over di-
rect methanol fuel cells, including faster anode kinetics and
smaller amounts of crossover through the Nafion membrane.
The lower tendency for crossover allows for fuel delivery at
high concentration directly on the Si-MEA, which greatly
simplifies the fuel cells design, reduces parasitic losses, and
thus increases the attainable overall energy density.
The electrodeposited catalyst structures studied here ex-
hibit power densities as high as 28.0 mW/cm2 when using
formic acid as the fuel on pure Pd catalyst structures. For
comparison, others have reported microfabricated fuel cells
using methanol as the fuel with power densities ranging from
the ␮W/cm2 level to 20 mW/cm2 at T < 30 ◦C [3–7]. In ad-
dition, the intimate bonding of these catalyst structures on
the silicon-based current collector grids as well as the den-
dritic, rigid morphology of the catalyst structures themselves,
fulfills key compatibility requirements with respect to ro-
bustness for post-CMOS processing, as well as use for its
intended application in miniature microelectronics. Further
testing is needed to determine whether these silicon-based
MEAs also will withstand temperature changes, shock, and
vibration.
Fig. 7. (a) Polarization and (b) power density curves from Si-MEAs with
purePt, Pt/Pd = 1:1, purePd, orPt/Pd-SDcatalyststructuresastheanode, and
pure Pt as the cathode. The current densities are normalized to the geometric
surface area of the electrodes.
on the cathode side. The polarization and power density
curves of these integrated Si-MEAs are shown in Fig. 7. The
open circuit potentials (OCPs) of the Si-MEAs with pure
Pt, Pt/Pd = 1:1, Pt/Pd-SD, and pure Pd, catalyst structures
on the anode were 0.46, 0.62, 0.68, and 0.63 V, respectively,
while the peak power densities were 3.0, 10.1, 20.6, and
28.0 mW/cm2, respectively. At cell potentials ≥0.5 V, the
Pt/Pd-SD sample exhibited the highest current density,
which is in agreement with the results from CA, on the same
sample at an applied electrode potential of 0.2 V versus
RHE. Note that a lower applied potential in CA corresponds
to a higher measured cell potential in fuel cell measurements.
At cell potentials ≤0.5 V, the highest current densities were
obtained from the Si-MEA with pure Pd catalyst structures
at the anode.
Formic acid oxidation on pure Pt catalyst structures re-
quires concomitant CO oxidation, which leads to an increase
in anodic potential and thus a decrease in cell potential [25].
In contrast, formic acid oxidation on Pd-containing catalysts
occurs on Pd sites through the ‘direct pathway’ without the
formation of any CO species, thus enabling formic acid oxi-
dation to occur at lower anodic potentials and thus higher cell
potentials, which is better for fuel cell performance. Again,

4682
R.S. Jayashree et al. / Electrochimica Acta 50 (2005) 4674–4682
[8] C. Rice, S. Ha, R.I. Masel, P. Waszczuk, A. Wieckowski, T. Barnard,
Acknowledgements
J. Power Sources 111 (2002) 83.
[9] Y.W. Rhee, S.Y. Ha, R.I. Masel, J. Power Sources 117 (2003) 35.
[10] J. Munk, P.A. Christensen, A. Hamnett, E. Skou, J. Electroanal.
Chem. 401 (1996) 215.
[11] G.-Q. Lu, A. Crown, A. Wieckowski, J. Phys. Chem. B 103 (1999)
9700.
TheauthorsthankDARPAforfinancialsupportunderU.S.
Air Force Grant F33615-01-C-2172. The scanning electron
microscopy work was carried out in the Center for Microanal-
ysis of Materials, University of Illinois, which is partially
supported by the U.S. Department of Energy under grant
DEFG02-91-ER45439. We acknowledge V. Petrova and J.
Mabon for helpful discussions. JSS acknowledges support
from a National Science Foundation Graduate Research Fel-
lowship.
[12] R. Parsons, T. VanderNoot, J. Electroanal. Chem. Interfacial Chem.
257 (1988) 9.
[13] X.H. Xia, T. Iwasita, J. Electrochem. Soc. 140 (1993) 2559.
[14] M. Arenz, V. Stamenkovic, T.J. Schmidt, K. Wandelt, P.N. Ross,
N.M. Markovic, Phys. Chem. Chem. Phys. 5 (2003) 4242.
[15] M.S. Wilson, S. Gottesfeld, J. Electrochem. Soc. 139 (1992) 28.
[16] M.S. Wilson, S. Gottesfeld, J. Appl. Electrochem. 22 (1992) 1.
[17] M.S. McGovern, E.C. Garnett, C. Rice, R.I. Masel, A. Wieckowski,
J. Power Sources 115 (2003) 35.
References
[18] T.L. Knutson, W.H. Smyrl, 206th Meeting of the Electrochemical
Society, Honolulu, 3–8 October, 2004.
[1] W.Y. Sim, G.Y. Kim, S.S. Yang, IEEE MEMS (2001) 341.
[2] J. Yeom, G.Z. Mozsgai, A. Asthana, B.R. Flachsbart, P. Waszczuk,
E.R. Choban, P.J.A. Kenis, M.A. Shannon, in: First International
Conference on Fuel Cell Science, Engineering and Technology,
Rochester, NY, USA, 21–23 April 2003, Fuel Cell Science, Eng.
Technol. (2003) 267.
[3] G.Q. Lu, C.Y. Wang, T.J. Yen, X. Zhang, Electrochim. Acta 49
(2004) 821.
[4] J. Li, C. Moore, P.A. Kohl, J. Power Sources 138 (2004)
211.
[19] S. Litster, G. McLean, J. Power Sources 130 (2004) 61.
[20] A.C. Hill, R.E. Patterson, J.P. Sefton, M.R. Columbia, Langmuir 15
(1999) 4005.
[21] B. Ilic, D. Czaplewski, P. Neuzil, T. Stanczyk, J. Blough, G.J.
Maclay, J. Mater. Sci. 35 (2000) 3447.
[22] J. Yeom, G.Z. Mozsgai, B.R. Flachsbart, E.R. Choban, A. Asthana,
M.A. Shannon, P.J.A. Kenis, Sens. Actuators B, in press.
[23] E. Yeager, J.O. Bockris, B.E. Conway, S. Sarangapani (Eds.), Com-
prehensive Treatise of Electrochemistry, Electrodics: Experimental
Techniques, vol. 9, 1984.
[5] S. Motokawa, M. Mohamedi, T. Momma, S. Shoji, T. Osaka, Elec-
trochem. Commun. 6 (2004) 562.
[24] E.A. Lafferty, Y.G. Kim, M.P. Soriaga, Electrochim. Acta 44 (1998)
1031.
[6] T.J. Yen, N. Fang, X. Zhang, G.Q. Lu, C.Y. Wang, Appl. Phys. Lett.
83 (2003) 4056.
[7] S.C. Kelley, G.A. Deluga, W.H. Smyrl, Electrochem. Solid-State
Lett. 3 (2000) 407.
[25] T.R. Ralph, M.P. Hogarth, Platinum Met. Rev. 46 (2002) 117.
[26] Y. Zhu, Z. Khan, R.I. Masel, J. Power Sources 139 (2005) 15.
[27] C. Rice, S. Ha, R.I. Masel, A. Wieckowski, J. Power Sources 115
(2003) 229.