Measuring electrocatalytic activity on a local scale with scanning differential electrochemical mass spectrometry

Article abstract of DOI:10.1149/1.1570823

This paper describes a new technique entitled scanning differential electrochemical mass spectrometry (SDEMS) that combines a quadrupole mass spectrometer with a membrane-covered capillary inlet and a high resolution positioning system that is designed to perform spatial mapping in solution near an electrode interface. Potential applications of this technique include the local characterization of anode catalysts for fuel cells as well as a range of analytical measurements and combinatorial screening studies. The capabilities of this technique are demonstrated by monitoring product evolution in several model electrocatalytic reactions, including the hydrogen evolution reaction, carbon monoxide oxidation, and the direct oxidation of methanol on platinum and platinum-ruthenium electrodes. The inlet of the SDEMS is based upon a small diameter capillary tube to which a nanoporous, hydrophobic membrane is attached. The capillary inlet is positioned near a substrate electrode using a three-dimensional positioning system. The effect of capillary substrate separation and substrate current on the sensitivity and time response of mass spectrometer's ion current are illustrated during hydrogen evolution at a platinum substrate. The sensitivity is demonstrated further by detection of carbon dioxide evolution during the oxidation of a monolayer of carbon monoxide adsorbed on platinum. The ability to address more complex reactions involving complete and partial oxidation products is illustrated with methanol oxidation. In order to demonstrate the ability of this technique to perform spatial mapping, an eight-element band electrode was interrogated for hydrogen evolution and methanol oxidation. Detection of ion currents associated with complete and partial oxidation products of methanol on a set of platinum-ruthenium band electrodes illustrates the ability of this method to spatially discriminate between various reactive sites on a surface, which has potential utility in analytical characterization as well as application as a screening tool in combinatorial catalysis studies.

Full text of DOI:10.1149/1.1570823

E312
Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
0013-4651/2003/150͑6͒/E312/9/$7.00 © The Electrochemical Society, Inc.
Measuring Electrocatalytic Activity on a Local Scale with
Scanning Differential Electrochemical Mass Spectrometry
*
**,z
K. Jambunathan and A. C. Hillier
Department of Chemical Engineering, University of Virginia, Charlottesville, Virginia 22903-2442, USA
This paper describes a new technique entitled scanning differential electrochemical mass spectrometry ͑SDEMS͒ that combines a
quadrupole mass spectrometer with a membrane-covered capillary inlet and a high resolution positioning system that is designed
to perform spatial mapping in solution near an electrode interface. Potential applications of this technique include the local
characterization of anode catalysts for fuel cells as well as a range of analytical measurements and combinatorial screening studies.
The capabilities of this technique are demonstrated by monitoring product evolution in several model electrocatalytic reactions,
including the hydrogen evolution reaction, carbon monoxide oxidation, and the direct oxidation of methanol on platinum and
platinum-ruthenium electrodes. The inlet of the SDEMS is based upon a small diameter capillary tube to which a nanoporous,
hydrophobic membrane is attached. The capillary inlet is positioned near a substrate electrode using a three-dimensional position-
ing system. The effect of capillary substrate separation and substrate current on the sensitivity and time response of mass
spectrometer’s ion current are illustrated during hydrogen evolution at a platinum substrate. The sensitivity is demonstrated further
by detection of carbon dioxide evolution during the oxidation of a monolayer of carbon monoxide adsorbed on platinum. The
ability to address more complex reactions involving complete and partial oxidation products is illustrated with methanol oxidation.
In order to demonstrate the ability of this technique to perform spatial mapping, an eight-element band electrode was interrogated
for hydrogen evolution and methanol oxidation. Detection of ion currents associated with complete and partial oxidation products
of methanol on a set of platinum-ruthenium band electrodes illustrates the ability of this method to spatially discriminate between
various reactive sites on a surface, which has potential utility in analytical characterization as well as application as a screening
tool in combinatorial catalysis studies.
© 2003 The Electrochemical Society. ͓DOI: 10.1149/1.1570823͔ All rights reserved.
Manuscript submitted September 6, 2002; revised manuscript received January 2, 2003. Available electronically April 21, 2003.
The inability of existing anode catalysts in polymer electrolyte
Scanning mass spectrometry has also been exploited in both gas and
liquid systems.25-27 In electrochemical systems, the application of
electrode arrays with individually addressable electrodes allows for
quantitative measurement of individual reaction currents.28 The ease
of constructing microfabricated electrode arrays via traditional mi-
croelectronics processing steps has facilitated these studies. Product
detection schemes have also been developed. A pH-sensitive optical
screening method was employed to identify the onset of methanol
oxidation through detection of proton evolution.20 Although this is a
rapid and parallel method, it suffers from limited spatial resolution
and an inability to discriminate between partial and complete oxida-
tion processes. In addition, quantitative information concerning re-
action rates, reaction pathways, and interfacial chemistry is absent in
this method. A tactic that is related to the optical screening is to
probe the solution environment near an electrode sample using scan-
ning probe techniques. We recently presented a screening method
that quantitatively detects the rate of hydrogen oxidation with high
spatial resolution using the scanning electrochemical microscope.29
Although this method provides quantitative rate measurements for
hydrogen oxidation and can be used to directly probe the generation
of protons near a substrate during methanol oxidation,30 the useful
pH range of this technique is limited.
membrane fuel cells to oxidize fuels other than hydrogen at suffi-
cient levels has severely limited the development of this technology
for practical applications.1-3 Electrode-deactivating species such as
carbon monoxide, that are present in fuels either as by-products
from upstream reforming or that appear as partial oxidation products
during direct oxidation of liquid fuels such as methanol, pose a
significant problem for platinum-based catalysts. The strong binding
of carbon monoxide to platinum sites leads to catalyst deactivation
and decreases reactivity to unsuitable levels.4,5 Although improve-
ments in catalytic activity may be achieved over pure platinum by
the addition of oxophilic species such as ruthenium,6-9 tin,10,11 and
molybdenum12-14 that promote CO oxidation at lower potentials,
necessary levels of performance are yet to be achieved.
One of the barriers impeding the discovery of more active and
poison-tolerant anode catalysts is the vast parameter space that must
be sampled in order to thoroughly evaluate the ranges of composi-
tion and structure that form the basis of the next generation cata-
lysts. The multifunctional requirements of catalysts for the direct
oxidation of hydrocarbons, which include the ability to activate
C-H, C-O, and even C-C bonds, suggest that optimum performance
will require multicomponent binary, ternary, or even quaternary
catalysts.15-17 Combinatorial methods pose great promise to effi-
ciently identify candidate materials or sample vast regions of com-
position space for further exploration.18,19 This was recently demon-
strated in a combinatorial study that discovered a new, highly active,
four-component anode catalyst for methanol oxidation via an inkjet
printer-based array fabrication method combined with an optical
screening technique.20
In this manuscript, we describe a recently developed technique
entitled scanning differential electrochemical mass spectrometry
͑SDEMS͒ that functions as a reactivity screening tool and general
characterization method for local measurement of dissolved gases
and volatile liquid species near electrode surfaces. It consists of a
scanning probe microscope that utilizes a mass spectrometer as a
detector with a membrane-coated capillary tip as the inlet. This
membrane-covered capillary inlet is reminiscent of an earlier paper
describing a stationary, membrane-coated pin-hole inlet that was
used to indirectly sample single crystal electrodes in a hanging me-
niscus configuration.31 In addition, a recent study describes a scan-
ning mass spectrometry tool that utilizes a membrane-free, scanning
capillary tube with a thermospray inlet to completely vaporize a
continuously pumped electrolyte sample and detect dissolved solu-
A critical component in the application of combinatorial ap-
proaches for catalyst discovery is the development of in situ, chemi-
cally sensitive screening tools to ascertain reactivity parameters and
spatially evaluate catalytic activity.21,22 A limited number of screen-
ing tools have been developed for catalyst characterization in gas or
liquid phases. In gas-phase systems, infrared spectral and thermo-
graphic imaging have proven useful for rapid catalyst screening.23,24
tion species.32 Our device has a different sampling scheme that is
based upon modification of the differential electrochemical mass
spectrometer ͑DEMS͒. This technique is based upon the original
membrane-inlet, electrochemical mass spectrometry ͑EMS͒ design
* Electrochemical Society Student Member.
** Electrochemical Society Active Member.
z E-mail: ach3p@virginia.edu
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
E313
of Grambow and Bruckenstein,33 which was later modified by
Wolter and Heitbaum to achieve a faster response time via differen-
tial pumping of the inlet.34
A variety of modifications and upgrades to the original DEMS
design have been made to expand the capabilities and application of
this technique.35-38 DEMS has proven a highly valuable method for
direct, quantitative measurement of dissolved gases and volatile liq-
uid species present at the electrode-electrolyte interface.39 The abil-
ity of DEMS to measure complete and partial oxidation products
associated with the oxidation of fuels such as methanol,40 formic
acid,41 and ethanol35,42 is a particularly appealing feature in the
characterization of anode catalysts. DEMS can provide quantitative
evidence of product distribution and reaction pathways.40,43,44 Iden-
tification of the nature of decomposition products and dominant sur-
face bound species45-47 has also been demonstrated through careful
product analysis as well as by the application of isotopic reagents.48
In this work, the capabilities of SDEMS are demonstrated for
several model electrocatalytic reactions, including hydrogen evolu-
tion, carbon monoxide oxidation, and the direct oxidation of metha-
nol. General system characteristics as well as time response and
sensitivity are illustrated using hydrogen evolution at a platinum
substrate. The ability to detect monolayer quantities of product fol-
lowing carbon monoxide oxidation in addition to visualizing com-
plete and partial oxidation products of methanol oxidation are illus-
trated. The screening capabilities of SDEMS are demonstrated by
mapping product evolution at an eight-element band electrode con-
taining electrodeposited platinum-ruthenium catalysts of variable
composition.
applying a potential pulse program between Ϫ0.15 and Ϫ1.57 V vs.
Hg/Hg2SO4 at a frequency of 100 Hz via a potentiostat ͑model 283,
EG&G Instruments, Oak Ridge, TN͒ using a function generator
͑Global Specialties, New Haven, CT͒. The various band electrodes
were created serially starting with a pure Ru band followed by bands
with increasing Pt content. An individual band was coated with a
PtxRuy electrodeposit and then the solution was removed. The cell
was then rinsed with DI water and a new Pt/Ru solution was injected
for the next band. Creating the pure Pt electrode last was necessary
to avoid exposing the Pt electrode to RuCl3 solution and, thus, pre-
venting the spontaneous deposition of Ru on the Pt electrode.49,50
However, the spontaneous deposition of Pt on Ru can also occur, but
the data for our samples suggested that this process did not occur to
a significant degree.51,52 The electrode compositions were verified
using both Auger electron spectroscopy ͑AES͒ and energy disper-
sive spectroscopy ͑EDS͒, which gave identical composition mea-
surements within measurement error.
Before use, the working electrodes were sonicated in an ethanol/
water solution and then rinsed with copious amounts of DI water.
Further electrode pretreatment included electrochemical potential
cycling between water oxidation and hydrogen evolution in an aque-
ous solution of 0.5 M H2SO4 until a characteristic voltammetric
signature of a polycrystalline platinum surface was obtained. This
electrochemical treatment was not performed for the PtxRuy elec-
trodes to avoid the oxidative dissolution of Ru that occurs at positive
potentials. A length of Pt0.8Ir0.2 wire ͑Goodfellow Corporation, Ber-
wyn, PA͒ was used as a counter electrode. The reference electrode
consisted of a AgCl-coated Ag wire. The potential scale of this elec-
trode was verified against a standard calomel electrode and con-
verted to the RHE scale. All data are presented vs. the RHE.
Two types of electrochemical cells were used. For experiments
performed with the Pt disc electrode, the electrochemical cell was
made from Teflon with a cell volume of 100 mL. The cell contained
slots for reference, counter, and working electrodes. The working
electrode was positioned in a slot at the bottom of the cell with a
vertical orientation and the electrode surface facing up. The cell was
made leak-proof by pressure fitting all electrodes with Teflon tape.
The cell also contained a slot for a right angle glass prism ͑Edmond
Scientific, Barrington, NJ͒, which was used to visualize the gap
between capillary tip and working electrode. For experiments per-
formed with the band electrodes, a 1 ϫ 0.5 in. rectangular glass
tube ͑Whale Apparatus Company, Inc., Hellertown, PA͒ was glued
directly to the band electrode assembly using an insulating lacquer
͑WB-018 Stop-Off Coating, Michigan Chrome & Chemical Com-
pany, Hope, AK͒. The eight working electrodes were connected to a
twenty pin dip-clip ͑Pomona Electronics, Pomona, CA͒ to allow
potential control of individual electrodes or the entire assembly.
Experimental
Reagents.—All experiments were performed using electrolytes
dissolved in 18 M⍀ deionized ͑DI͒ water ͑E-Pure, Barnstead, Dubu-
que, IA͒. The solutions were deaerated with nitrogen ͑BOC Gases,
Murray Hill, NJ͒ prior to measurement. Carbon monoxide studies
were achieved by delivering pure CO gas ͑BOC Gases, Murray Hill,
NJ͒ to the electrochemical apparatus by bubbling through a porous
ceramic frit ͑Ace Glass, Inc., Vineland, NJ͒ into the electrolyte so-
lution for a period of 5 min with the working electrode held under
potential control. The CO was then removed from solution by purg-
ing with nitrogen. Electrochemical measurements were performed in
solutions containing as-received sulfuric acid (H2SO4), sodium sul-
fate (Na2SO4), and methanol (CH3OH, 99.9%͒ ͑Aldrich, Milwau-
kee, WI͒. Electrodeposits of Pt and Ru were obtained from solutions
containing chloroplatinic acid (H2PtCl6) and ruthenium chloride
(RuCl3) ͑Strem Chemicals, Newburyport, MA͒.
Electrodes.—Two types of working electrodes were used that
included a 2 mm diam polycrystalline Pt disc electrode ͑Bioanalyti-
cal Systems, West Lafayette, IN͒ encased in poly ether ether ketone
͑PEEK͒ and an eight-element band electrode coated on a frosted
glass slide ͑Erie Scientific Company, Portsmouth, NH͒. Preparation
of the Pt disc electrode consisted of polishing with successively finer
grades of energy paper to achieve a planar surface, followed by
diamond polishing paste and alumina slurry down to 0.05 ␮m par-
ticle size to produce electrodes that exhibited a mirror finish. The
band electrodes were created by metal evaporation through a pat-
terned, aluminum mask containing 8 slits ͑1 mm width and 2.5 mm
period͒. A coating of 10 nm Ti followed by 100 nm Au ͑Ernest F.
Fullam, Inc., Latham, NY͒ was evaporated onto the glass slide
through the mask using a thermal evaporator ͑Bench Top Turbo III
high vacuum evaporator, Denton Vacuum, LLC, Moorestown, NJ͒.
Pt was then electrodeposited onto these Au bands from a solution
containing 5 mM H2PtCl6 and 0.1 M Na2SO4 at a constant potential
of 50 mV ͑vs. reversible hydrogen electrode, RHE͒ for 5 min using
a eight channel multielectrode potentiostat ͑model CH1030, CH In-
struments, Inc. Austin, TX͒. Platinum-ruthenium electrodes were
electrodeposited from 0.1 M Na2SO4 solutions neutralized to pH 7
with NaOH containing different amounts of H2PtCl6 and RuCl3 at a
total metal content of 10 mM. Electrodeposition was achieved by
SDEMS.—The SDEMS used here consisted of an electrochemi-
cal cell and potentiostat combined with a mass spectrometer and
three-dimensional ͑3-D͒ positioning system, which were all under
computer control ͑Fig. 1͒. The mass spectrometer ͑MS͒ was a QMS
200 ͑Pfeiffer Vacuum, Nashua, NH͒ consisting of a quadrupole mass
analyzer using a Channeltron detector with mass resolution of
m/z ϭ 0.5 atomic mass units ͑amu͒ and a mass range up to 200
amu. The spectrometer utilized an electron impact ͑EI͒ ionization
source. Pumping of the detector’s ionization chamber was achieved
with a combination of mechanical diaphragm roughing pump and a
turbomolecular pump, which were capable of an ultimate pressure
approaching 10Ϫ7 Torr. The inlet to the MS consisted of a sampling
probe that was a 10 cm long, 150 ␮m inner diam ͑i.d.͒ poly͑tet-
rafluoroethylene͒ ͑PTFE͒ capillary tube ͑Valco Instruments, Hous-
ton, TX͒ to which a 20 nm pore size Gore-Tex membrane was se-
cured with shrink tubing. This capillary was connected to a
T-connector fitting ͑Valco Instruments, Houston, TX͒ that was fur-
ther connected to the mass spectrometer ionization chamber with a
10 cm, 1 mm diam stainless steel ͑SS͒ rod and a rotary vane pump
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E314
Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
Figure 2. Amplitude of hydrogen ion current (IH2 ,m/z ϭ 2) as a function
of substrate current at a capillary/substrate separation of approximately 100
␮m. Solution: deaerated 0.5 M H2SO4 .
conditions existing within the ionization chamber of the mass spec-
trometer. The hydrophobic nature of the membrane coating is criti-
cal to the operation of this detector and in particular served to reduce
the quantity of water within the spectrometer. The position of the
capillary inlet was controlled with high precision using a computer-
controlled, 3-D positioning stage. For measurements that involve
detecting species evolved from a specific location on a substrate
electrode, the tip-substrate gap was reduced to the smallest possible
value of typically less than 100 ␮m in order to maximize the col-
lection of surface generated species and minimize the time delay
between the evolution of a species at the electrode surface and its
detection in the spectrometer.
Figure 1. Schematic of SDEMS showing major system components. Inset
depicts magnified view of tip/sample interface and capillary/membrane inlet.
Hydrogen evolution and SDEMS characteristics.—The ability of
the MS to detect dissolved gases or other volatile species when
immersed in an aqueous solution is key to using this method for the
characterization of electrochemical processes. A model reaction that
illustrates the various features and capabilities of this technique is
the hydrogen evolution reaction ͑HER͒
͑RV5, BOC, Edwards, Wilmington, MA͒ to provide differential
pumping of the inlet. The rotary vane pump provided an ultimate
vacuum of 0.1 Torr within the T connector.
The capillary sampling probe was mounted to a holder attached
to three piezoelectric inchworm motors ͑model IW-710, Burleigh
Instruments, Inc., Fishers, NY͒. The motors were mounted directly
onto one TS-300 ͑z͒ and two TS-100 ͑xy͒ translation stages config-
ured for 3-D orthogonal motion. The complete assembly was
mounted directly to an optical breadboard ͑Newport Corporation,
Irvine, CA͒. The inchworm motors had a 25 mm range of motion, a
mechanical resolution of 4 nm, and optical encoders giving an ab-
solute position reproducibility of 0.5 ␮m. Position control of the
capillary inlet was achieved with a Series 6200 ULN Controller
͑Burleigh Instruments, Inc., Fishers, NY͒. Electrochemical measure-
ments were performed using a bipotentiostat ͑AFDRE, Pine Instru-
ment Company, Grove City, PA͒. Data collection was achieved us-
ing a commercial mass spectrometer controller program ͑Quadstar,
Balzers Limited, FL͒ with additional analog input channels to allow
synchronization of the electrochemical signals ͑substrate current and
potential͒ with up to 64 additional mass channels.
2Hϩ ϩ 2eϪ → H2
͓1͔
where the hydrogen that is evolved at a substrate electrode can be
monitored with the MS. In order to verify the differential mode of
detection and to provide an illustration of the sensitivity, steady-state
ion current values were monitored as a function of the substrate
current during hydrogen evolution. Figure 2 shows the steady-state
ion current values for m/z ϭ 2 at a tip-substrate separation of ϳ100
␮m for several substrate current values between Ϫ5 and Ϫ35 ␮A.
Although there is some scatter in the data, a near-linear relationship
between the ion current and substrate current is observed. This is the
expected response for a DEMS configuration and indicates that the
tip response is proportional to the flux of hydrogen evolving from
the substrate surface. Thus, the spectrometer is detecting the differ-
ential response as is typical for DEMS.34 In addition, the current
sensitivity at the smallest tip-substrate gap is approximately 1 nA
per 10 ␮A of substrate current at a 0.03 cm2 electrode. This gives an
area-normalized sensitivity for hydrogen of 3 ϫ 10Ϫ6 nA ͑hydro-
gen ion current͒/nA cmϪ2 ͑substrate current͒.
In order for a species like hydrogen that is produced at the elec-
trode surface to be detected by the mass spectrometer, it must dif-
fuse from the substrate to the membrane inlet, pass through the
membrane and then be carried down the length of the capillary tube
into the ionization chamber. The ion current (Ii) for any species i
will be related to its flux (Ni) into the ionization chamber times the
capillary area (Ac) and a calibration constant (⌳i) according to
Results and Discussion
The configuration of the SDEMS is depicted in Fig. 1. The key
components included an electrochemical cell with a substrate elec-
trode connected to a potentiostat, a high resolution micropositioning
system, a quadrupole mass spectrometer with a differentially
pumped, capillary inlet and a computer for data acquisition and
control ͑Fig. 1͒. The inlet to the spectrometer comprised a capillary
tube with an i.d. of 150 ␮m. The entrance of this capillary tube was
covered with a nanoporous, hydrophobic membrane. The small inlet
diameter increased the spatial resolution and the hydrophobic mem-
brane coating separated the liquid electrolyte from the high vacuum
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
E315
where ␧ is the membrane void fraction representing the open pore
area, DKi n is the Knudsen diffusion coefficient for species i, Lm is the
membrane thickness, T is the temperature, and R is the gas constant.
The Knudsen diffusion coefficient can be estimated from the kinetic
theory of gases according to
1/2
d 2RT
DKi n
ϭ
͓6͔
ͫ
ͬ
3
Mi
where d is the pore diameter and Mi is the molecular weight of
species i.53 This equation indicates that species will diffuse through
the membrane at different rates that are determined by differences in
their molecular weights. However, control experiments ͑vide infra͒
suggest that the membrane does not present a substantial mass trans-
fer barrier for the diffusion of gases and, thus, fractionation at this
interface is limited.
The final mass transfer step involves transporting species from
the inner membrane surface down the capillary tube into the ioniza-
tion chamber of the mass spectrometer. This process can be best
described as viscous flow where the gas velocity is pressure-driven
and obeys the Hagen-Pouiseville law.53 The corresponding mass
transfer coefficient for this process has the form
Scheme 1. Schematic of mass spectrometer inlet in electrolyte solution dur-
ing hydrogen generation at a platinum substrate. Mass transfer coefficients
are shown corresponding to the electrolyte solution (ks), membrane film
(km), and capillary tube (kc) regions.
Ii ϭ ⌳iNi Ac ϭ ⌳iKi⌬pi Ac
͓2͔
d2⌬p
kc,i
ϭ
͓7͔
where ⌬pi is the pressure difference for species i between the sub-
strate electrode and in the ionization chamber and Ki is an overall
mass transfer coefficient. The flux can be described with a simple
model that considers three mass transfer steps occurring in series
͑Scheme 1͒. The individual steps include diffusion from the sub-
strate through solution to the membrane (ks,i), transport across the
membrane (km,i), and convection down the capillary tube to the
ionization chamber (kc,i). The overall mass transfer coefficient can
be written as a combination of the three individual mass transfer
coefficients according to
32 ␮Lc
where ⌬p is the total pressure difference between the membrane
surface and the ionization chamber, ␮ is the gas viscosity, and Lc is
the length of the capillary tube. Clearly, this mass transfer process is
independent of the species involved and is simply a function of the
pressure drop and the capillary dimensions. This process can be
accelerated by increasing the capillary diameter, increasing the pres-
sure drop, or by decreasing the capillary length.
The controlling mass transfer process in this system was evalu-
ated by comparing the response in liquid to that in laboratory air
with and without the membrane. The rate of mass transfer down the
length of the capillary was estimated by measuring the time re-
sponse of the detector to gas dosing in laboratory air in the absence
of the membrane. Following exposure of the capillary inlet to car-
bon dioxide, an almost instantaneous increase in ion current was
observed for m/z ϭ 44. The time constant for this response was
estimated at less than 1 s. This indicated that the value of the mass
transfer coefficient (kc,i) down the capillary tube was quite large. A
1/Ki ϭ 1/ks,i ϩ 1/km,i ϩ 1/kc,i
͓3͔
where the overall process can be influenced by all three steps or
dominated by one or two. Ultimately, the sensitivity and time re-
sponse of the SDEMS can be optimized by increasing the rates of all
mass transfer steps.
Steady-state mass transfer from the substrate to the membrane
can be approximated by diffusion through a thin gap
Di
similar experiment was performed in laboratory air with
a
ks,i
ϭ
͓4͔
HiZ
membrane-coated capillary inlet. The time response was again al-
most instantaneous, indicating a large value for the mass transfer
coefficient through the membrane (km,i). Thus, the membrane does
not present a substantial mass transfer limitation for the diffusion of
gases.
The time response of the system in solution was then evaluated
for the series of transient curves measured during hydrogen evolu-
tion at a substrate electrode. Figure 3 depicts measurement of the
hydrogen ion current (m/z ϭ 2) with the capillary inlet positioned
near a polycrystalline Pt disk electrode. The electrochemical sub-
strate current is stepped from zero to a constant value of Ϫ20 ␮A
while the ion current in the MS is monitored. Under these condi-
tions, protons are reduced at the substrate and hydrogen is evolved
at a constant rate. The MS ion current is plotted as a function of time
at separations of 60, 210, and 500 ␮m between the capillary inlet
and substrate electrode.a The shapes of the transient curves exhibited
where Di is the diffusion coefficient of species i in solution, Z is the
membrane/substrate gap distance, and Hi is the Henry’s law constant
for species i. Thus, the diffusion coefficient of each species in solu-
tion impacts upon the intensity as well as the time response of the
ion signal. A smaller diffusion coefficient will reduce the measured
mass intensity and create a slower time response for that species. A
decrease in the Henry’s law constant, which is equivalent to an
increase in gas solubility in the liquid phase, will serve to increase
k
s,i . Therefore, highly soluble gas species will exhibit a higher ks,i
value. Ultimately, reducing the influence of mass transfer in the
solution phase for all species can most effectively be achieved by
decreasing the membrane/substrate gap distance Z to the smallest
possible values.
Mass transfer across the membrane is best estimated by Knudsen
diffusion through a small, gas-filled pore
a The absolute separation between tip and substrate is determined in these measure-
ments by visual inspection using a charge-coupled device camera fit with a variable,
zoom lens. The intrinsic error in this measurement derives from the difficulty in deter-
mining absolute tip-substrate contact.
␧DKi n
km,i
ϭ
͓5͔
RTLm
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E316
Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
Figure 3. Plot of hydrogen ion current (IH2 ,m/z ϭ 2) evolving from a Pt
substrate electrode following a current step ͑0 ␮A to Ϫ20 ␮A͒ at capillary/
substrate separations of 60, 210, and 500 ␮m. Solution: deaerated 0.5 M
H2SO4 .
a characteristic response, where the hydrogen ion current ͑I͒ can be
represented by a first-order with dead-time model.54
Ͻ
␪
d
I ϭ 0
t
t Ϫ ␪ ͒
͑
d
Ͼ
I ϭ A 1 Ϫ exp Ϫ
t
␪
͓8͔
ͫ
ͩ
ͪ
ͬ
d
␶
As the tip/substrate separation is decreased from a gap distance of
500 to 60 ␮m, the amplitude ͑A͒ of the measured ion current in-
creases while the magnitude of both the dead time (␪d) and time-
constant ͑␶͒ decrease. The increase in ion current reflects an increase
in amount of sampled hydrogen, which is equivalent to an increase
in the collection efficiency, as the tip is positioned closer to the
substrate surface. The decrease in time constant and dead time are a
reflection of the increased speed at which products generated at the
substrate electrode reach the capillary tip with decreasing gap dis-
tance.
Figure 4. ͑A͒ Plot of hydrogen ion current (IH2 ,m/z ϭ 2) as a function of
capillary/substrate separation at a substrate current of Ϫ20 ␮A. ͑B͒ Open
circles show plot of time constant and filled squares show dead time corre-
sponding to fits of current-step experiments for hydrogen evolution as a
function of capillary/substrate separation. Solution: deaerated 0.5 M H2SO4 .
The series of transient, galvanostatic measurements depicted in
Fig. 3 were fit to the first-order model ͑Eq. 8͒ and the results are
summarized in Fig. 4. Figure 4A depicts the steady-state ion current
measured at a tip/substrate separation ranging from 0 up to 1200 ␮m
for a fixed substrate current of Ϫ20 ␮A. Each data point represents
the mass spectrometer ion current for hydrogen (m/z ϭ 2) mea-
sured 2 min following a step in the substrate current and, therefore,
reflects the steady-state hydrogen generation rate. The response
shows an increasing ion current as the tip/substrate separation is
decreased that follows a linear trend. This behavior indicates that the
efficiency of tip collection improves with decreasing tip/substrate
separation. The tip is able to capture a greater fraction of the
substrate-generated hydrogen as it is positioned closer to the source
of the hydrogen. In Fig. 4A, the steady-state current levels off at a
separation of approximately 200 ␮m and does not increase at
smaller separations. This is likely a result of the tip and substrate
making contact at this separation. In the present configuration, the
capillary tube used for the tip has an i.d. of 150 ␮m but an outer
diam ͑o.d.͒ of approximately 1.5 mm, which makes reducing the
tip/substrate separation to very small values difficult. It is expected
that improvements to this response may be achieved by constructing
smaller or conical tips, whose physical dimensions have been re-
duced to limit unwanted tip/substrate contact.
high of 40 s to a low of almost 0 s over this separation range. This
behavior can only be explained by mass transfer that is controlled by
the liquid phase where the value for ks,i is much smaller than km,i
and kc,i . Clearly, decreasing the tip/substrate separation increased
the value of ks,i . With the current experimental assembly, the fastest
time response that could be achieved indicated a time constant be-
tween 2 and 10 s with the tip immersed in liquid. Under these
conditions, it was estimated that over 90% of the resistance to mass
transfer is in the liquid phase. Improvements to this response may be
achieved by developing smaller tips that can be positioned closer to
the substrate surface without contact.
Carbon monoxide oxidation.—A variety of species in addition to
hydrogen can be detected with this SDEMS system just as has been
previously demonstrated with traditional DEMS.39,41,55 Of particular
interest are reagents relevant to anode reactants in fuel cells. DEMS
has been used to monitor the oxidation of various hydrocarbons,
including carbon monoxide, formic acid, methanol, ethanol, and a
variety of other potential fuels. Complete oxidation can be directly
measured by monitoring the evolution of carbon dioxide and a quan-
titative analysis provides a measure of the system’s current
efficiency.40 An example of CO2 detection is illustrated in Fig. 5. In
this experiment, a monolayer of CO is adsorbed onto the sur-
The values for the time constant and dead time are plotted in Fig.
4B at various separations. Both parameters decrease in magnitude
with decreasing tip/substrate separation. The time constant exhibits a
value of nearly 70 s at a separation of 1200 ␮m and approaches 10
s at a separation of less than 200 ␮m. The dead time varies from a
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
E317
Figure 5. Electrochemical and mass spectrometric response of Pt substrate
during oxidation of a CO monolayer in a nitrogen-purged solution containing
0.5 M H2SO4 at a scan rate of 10 mV sϪ1. ͑A͒ Electrochemical current and
͑B͒ CO2 ion current (ICO2 ,m/z ϭ 44).
face of a polycrystalline Pt electrode. This was achieved by expos-
ing that surface to aqueous electrolyte saturated with dissolved CO
for a period of 5 min while holding the electrode potential at 0.1 V
͑vs. RHE͒. The solution is then purged with nitrogen. Subsequently,
the adsorbed CO layer is oxidized to CO2 by scanning the substrate
electrode in the positive direction. The capillary inlet tip is posi-
tioned near the electrode surface, and the substrate current and CO2
ion current (m/z ϭ 44) are monitored.
Figure 6. Electrochemical and mass spectrometric response of Pt substrate
during methanol oxidation in a solution containing 1.0 M CH3OH, 0.5 M
H2SO4 at a scan rate of 10 mV sϪ1. ͑A͒ Electrochemical current, ͑B͒ CO2 ion
current (ICO2 ,m/z ϭ 44), C͒ methyl formate ion current (I60 ,m/z ϭ 60)
and D͒ hydrogen ion current (IH2 ,m/z ϭ 2). Arrows indicate direction of
potential scan. The ion currents are adjusted by 9, 15, and 3 s for the signals
in B, C, and D to compensate for the associated measurement delay times.
The electrochemical current observed in Fig. 5A is typical of CO
monolayer oxidation from polycrystalline Pt.4,5 A high coverage CO
layer forms following adsorption at substrate potentials in the hy-
drogen under potential deposition ͑UPD͒ region. Scanning the elec-
trode potential in the positive direction results in a prewave feature
starting just positive of 0.2 and extending up to 0.6 V. This prewave
is associated with oxidation of ‘‘weakly bound’’ CO via adsorbed
water and amounts to a loss of ϳ0.1 monolayer.4,5,56,57 The simul-
taneously measured CO2 ion current exhibits an increase in magni-
tude at potentials in the prewave region. Thus, the MS is detecting
CO2 that evolves from the surface during oxidation of the weakly
bound CO. Clearly, the detector is sensitive to submonolayer quan-
tities of CO2 . Further positive excursions in the substrate potential
produce the main oxidation wave with a peak centered at 0.8 V. The
remainder of the CO layer is oxidized to CO2 during this process.
The CO2 ion current signal responds to the main oxidation peak by
increasing in accordance with the formation of additional CO oxi-
dation product. Notably, a tail is observed in the ion current signal at
positive potentials that is absent in the electrochemical current. This
tail is related to the time required for CO2 generated at the substrate
to diffuse to the capillary inlet. The magnitude of this diffusional
tailing decreases at slower potential scan rates.
scanned in the positive direction. As the potential reaches ϳ0.7 V,
anodic current increases due to methanol oxidation. Methanol oxi-
dation presumably commences in concert with the formation of Pt
hydroxides, which serve to remove the surface-bound decomposi-
tion products. The methanol oxidation current peaks at 0.8 V and
then drops due to formation of a Pt oxide layer. Methanol oxidation
on the Pt oxide surface occurs at potentials in excess of 1.0 V.
During the reverse scan, a current peak is observed at 0.7 V corre-
sponding to methanol oxidation on clean Pt. This current subse-
quently decreases as methanol decomposition products adsorb and
deactivate the surface.
Simultaneous measurement of mass spectrometer ion currents
during methanol oxidation can be used to further characterize this
reaction. Figure 6B depicts the ion current for CO2 (m/z ϭ 44). A
comparison of the ion-current signal for CO2 with the electrochemi-
cal current indicates that they change in a similar fashion. An in-
crease in electrochemical current results in an increase in CO2 ion
current. This reflects a CO2 evolution rate that is proportional to the
methanol oxidation current. A few subtle differences are noted, how-
ever. The maximum electrochemical current is observed during oxi-
dation on Pt metal at 0.8 V during the positive potential scan while
the largest CO2 signal is observed during the negative potential scan
at 0.7 V. This would suggest that a greater current efficiency, which
corresponds to a greater conversion of methanol to carbon monox-
ide, occurs on the bare Pt metal during the negative potential scan.
The negative potential scan provides a ‘‘fresh’’ platinum surface by
Methanol oxidation.—The ability to simultaneously detect sev-
eral species associated with an electro-oxidation reaction is illus-
trated by the oxidation of methanol on a Pt electrode ͑Fig. 6͒. The
electrochemical current depicts the characteristic behavior of metha-
nol oxidation on Pt under acidic conditions. The hydrogen UPD
current is suppressed at low potentials due to the formation of
strongly bound, electrode-deactivating species that are produced
from methanol decomposition at the electrode surface.58,59 The sub-
strate current maintains a low value as the electrode potential is
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E318
Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
reduction of Pt oxides while the positive scan corresponds to a sur-
face that has been covered with methanol decomposition products.
One particular advantage of using DEMS to characterize electro-
oxidation reactions is the ability to detect partial oxidation products
in addition to the complete oxidation product carbon dioxide.40 The
two primary partial oxidation products of methanol oxidation are
formaldehyde and formic acid. Both species represent undesirable
products and the source of losses in current efficiency. Formalde-
hyde can be formed by a two-electron oxidation of methanol while
formic acid forms by a four-electron oxidation. However, formalde-
hyde and formic acid are difficult to detect directly by mass spec-
trometry in a methanol-containing solution due to m/z overlap and
low volatility.40 A solution to this dilemma for formic acid is the
ability to detect it indirectly via methyl formate. Although methyl
formate is not itself a direct partial oxidation product of methanol
oxidation, it is created by the reaction of formic acid and methanol
in the electrolyte solution
Scheme 2. Schematic of electrodeposited Pt-on-Au band electrodes and cap-
illary inlet tip to mass spectrometer during mapping experiments.
HCOOH ϩ CH3OH → HCOOCH3 ϩ H2O
͓9͔
This reaction is quite fast in the presence of high methanol concen-
tration and allows the ion current for methyl formate at m/z ϭ 60 to
be used to deduce formic acid concentration. In principle, a similar
indirect detection is possible for formaldehyde through its reaction
with methanol to produce dimethoxymethane (m/z ϭ 75).
at Ϫ0.05 V ͑vs. RHE͒ to reduce protons in the solution to H2 . For
methanol oxidation the substrate was held at 0.7 V to oxidize
CH3OH to CO2 . The capillary inlet was positioned at a separation
of about 100 ␮m from the substrate and rastered perpendicular to
the band axis at a scan rate of 25 ␮m/s. In the methanol oxidation
experiments, the electrode potential was held for 15 min before
imaging in order to approach a steady oxidation current at the vari-
ous electrodes.
Figure 7 shows the ion currents corresponding to H2 and CO2 as
a function of position on the substrate. The H2 signal depicts the
variation in ion current as a function of spatial position as the sub-
strate bands are reducing protons to hydrogen at a potential of Ϫ0.5
V. An increase in ion current at m/z ϭ 2 is observed as the inlet is
scanned over each band. A minimum in current is observed between
the bands. Notably, the CO2 ion current during this measurement
was unchanged as a function of position ͑not shown͒. The maximum
and minimum H2 signals are approximately constant at and between
each band, which is expected as the bands are all the same size and
Pt content. A similar behavior was found for the variation in CO2
ion current during methanol oxidation. The CO2 signal varied be-
CH O ϩ 2CH OH → CH OCH ͒ ϩ H O
͓10͔
͑
2
3
2
3
2
2
However, this product has only been detected in DEMS studies at
high temperatures.60 Nevertheless, the formation of significant form-
aldehyde near ambient conditions has been demonstrated using fluo-
rescence detection.61
The ion current for methyl formate (m/z ϭ 60) follows a trend
that is similar to CO2 . Peaks in the methyl formate signal are ob-
served associated with peaks in the electrochemical current. The
methyl formate peaks tend to trail the electrochemical peaks due to
the lower diffusion coefficient of this species. Nevertheless, the ob-
servation of methyl formate in parallel with both electrochemical
current and CO2 production clearly indicates the evolution of formic
acid during methanol oxidation. Notably, formic acid is thought to
be a species in the parallel reaction pathway for methanol and its
detection here and in prior DEMS studies confirms this path.40,44 As
with the CO2 signal, the methyl formate signal exhibits a maximum
for methanol oxidation on clean Pt during the negative potential
scan.
The signal corresponding to hydrogen (m/z ϭ 2) is also mea-
sured during these experiments ͑Fig. 6D͒. Hydrogen is not expected
to be evolved during methanol oxidation. However, during negative
excursions in the electrode potential, hydrogen is detected as a result
of the HER. The observation of hydrogen evolution indicates that
this reaction proceeds through a partially deactivated surface. The Pt
substrate is coated with sufficient methanol decomposition products
to suppress hydrogen adsorption as seen in the substrate current
͑Fig. 6A͒.
Surface imaging and catalyst screening.—One of the benefits of
the SDEMS configuration over traditional DEMS is that the detector
inlet consists of a small diameter capillary tube that can locally
sample the solution environment near an electrode surface. When
coupled to a high resolution positioning system, this provides spatial
mapping capabilities. This capability was demonstrated here using a
band electrode assembly. Scheme 2 depicts the substrate construc-
tion and capillary inlet position used for imaging studies. The sub-
strate comprised a series of band electrodes of 1 mm width and 2.5
mm period. The bands comprised vapor deposited gold on glass with
a layer of electrodeposited platinum at the outer surface. The sub-
strate was placed in a solution containing 1.0 M CH3OH and 0.5 M
H2SO4 in order to sample, in sequential measurements, the products
of methanol oxidation and hydrogen evolution. During these experi-
ments, the Pt bands were shorted to each other and held at the same
potential. For hydrogen evolution, the electrode potential was held
Figure 7. Ion currents for hydrogen (IH2 ,m/z ϭ 2) and CO2 (ICO2 ,m/z
ϭ 44) as a function of spatial position for a capillary separation of approxi-
mately 100 ␮m and a raster rate of 25 ␮m sϪ1. For hydrogen evolution, the
substrate was held at a constant potential of Ϫ0.05 V ͑vs. RHE͒. For metha-
nol oxidation, the substrate was held at a constant potential of 0.5 V ͑vs.
RHE͒ for 15 min before imaging with the SDEMS tip. Both measurements
were performed in a solution containing 1 M CH3OH and 0.5 M H2SO4 .
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
E319
Scheme 3. Schematic of electrodeposited PtRu-on-Au band electrodes. The
compositions for the electrodes are given in percentages as determined by
Auger electron spectroscopy.
tween nearly constant high and low values on top and between the
Pt bands while held at a potential of 0.7 V, corresponding to metha-
nol oxidation. The hydrogen ion current was unchanged as a func-
tion of position during this latter measurement.
Clearly, the spatial resolution of the SDEMS tip is poorer than
that found with other scanning probe techniques such as the atomic
force or scanning electrochemical microscopes. This is expected as
the capillary inlet in this system had a diam of 150 ␮m, and could
not resolve features below this size. Smaller capillary inlets can be
used to improve imaging resolution. However, this will be achieved
at the expense of detection of the mass signals as a necessarily
smaller flux of gas will enter the ionization chamber with a smaller
inlet diameter. Thus, there is a trade-off between image resolution
and ion signal in this technique. In addition, the large o.d. of the
capillary tube at 1.5 mm used in these studies prevented very small
tip/sample separations, which also limits the spatial resolution. The
outer capillary diameter may also present a mass transfer problem
when studying reactions at lower concentrations where diffusion
within the capillary/substrate gap is limited. These issues are cur-
rently being addressed by reducing the capillary o.d. and also via
forced convection to the capillary inlet.
An inherent advantage of the SDEMS imaging system for com-
binatorial studies is that it can quantitatively detect a number of
chemical species in liquid environments. In particular, as was dem-
onstrated in Fig. 6, SDEMS was able to detect some of the partial
and complete oxidation products of interest to anode reactions in
low temperature fuel cells. The ability of SDEMS to function as a
screening tool for methanol oxidation was demonstrated here by
measuring the evolution of CO2 and methyl formate on a series of
electrodeposited Pt-Ru electrodes. Scheme 3 depicts the test sample
used in this example. The substrate consists of the gold-on-glass
band electrodes as depicted in Scheme 2, but with a range of Pt-Ru
electrodeposits on the electrode surfaces. The compositions of these
electrodes range from pure Pt to pure Ru that were created by pulsed
electrochemical deposition from solutions containing H2PtCl6 and
RuCl3 at various concentration ratios ͑vide supra͒. The composi-
tions of these electrodes as listed in Scheme 3 were determined with
AES. Details of these electrodes and further electrochemical and
mass spectrometry studies will be presented in a subsequent publi-
cation.
Figure 8. Ion currents detected for ͑A͒ CO2 (ICO2 ,m/z ϭ 44), and ͑B͒
methyl formate (I60 ,m/z ϭ 60) during methanol oxidation at electrodepos-
ited PtxRuy samples as a function of applied substrate potential in a solution
containing 1 M CH3OH and 0.5 M H2SO4 . Schemes 2 and 3 depict the band
electrode structure and compositions. The electrode potentials were held for
15 min to obtain a constant current value before imaging.
the band containing 6% Ru gave the highest CO2 current, suggesting
that it had the highest activity under those conditions.
SDEMS scans over the electrodeposited PtxRuy band electrodes
are shown in Fig. 8. In these measurements, all 8 bands were elec-
trically shorted and placed in a 1.0 M methanol and 0.5 M H2SO4
solution. The MS capillary inlet was placed in close proximity to the
substrate at a separation of approximately 100 ␮m. The bands were
then held at different potentials for a period of 15 min before the
capillary was scanned over the bands at a rate of 25 ␮m sϪ1. The ion
currents corresponding to CO2 (m/z ϭ 44) and methyl formate,
HCOOCH3 (m/z ϭ 60) were monitored as a function of position
on the substrate. Figure 8A shows the CO2 ion currents as a function
of position and potential. In these results, the left most band contains
100% Ru and the right most band contains 100% Pt with the bands
in between containing varying compositions of Pt and Ru. At poten-
tials below 0.4 V, no CO2 was detected. However, at 0.40 V, a broad
CO2 signal was detected over the bands containing between 6 and
40% Ru. This suggests that methanol oxidation had begun at these
electrodes. With increasingly positive potentials, the broad CO2 sig-
nal increased and features associated with individual electrodes
emerged. At 0.55 V and above, distinct CO2 evolution from the
bands containing 34, 25, 22, and 6% Ru was indicated. At 0.65 V,
Simultaneous detection of methyl formate ͑Fig. 8B͒ evolution at
the PtxRuy band electrodes indicates a similar trend. Methyl formate
is not observed at potentials below 0.4 V. At 0.4 V, a broad signal is
detected at m/z ϭ 60 over the bands containing less than 40% Ru.
The magnitude of the methyl formate current also increased as the
electrode potential increased. The magnitude of the methyl formate
current at the various electrodes mimiced that of the CO2 signal. The
6% Ru band gave the highest methyl formate current at 0.65 V.
Notably, these results are consistent with measurements in which the
SDEMS tip was held stationary over each electrode while the elec-
trode potential was scanned or stepped to various potentials. A more
detailed analysis of these results is the subject of a subsequent pub-
lication.
Conclusions
This manuscript demonstrates a new imaging tool, SDEMS. This
technique is based upon the original design of the DEMS but with a
modified inlet in the form of a membrane-coated capillary and 3-D
positioning system that provides surface imaging capabilities. This
technique can be applied to a range of problems of importance to
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E320
Journal of The Electrochemical Society, 150 ͑6͒ E312-E320 ͑2003͒
20. E. Reddington, A. Sapienza, B. Gurau, R. Viswanathan, S. Sarangapani, E. S.
Smotkin, and T. E. Mallouk, Science, 280, 1735 ͑1998͒.
fuel cell catalysis and provides quantitative information about the
evolution of partial and complete oxidation products. In addition,
the ability to measure species in solution near an electrode surface at
specific spatial locations has additional analytical capabilities of par-
ticular relevance for combinatorial studies. In this work, we have
demonstrated the ability of SDEMS to be used as a screening tool to
detect carbon dioxide and methyl formate evolution during methanol
oxidation at a series of PtxRuy electrodes. We are currently extend-
ing these studies to characterize a range of catalyst systems in array
and gradient form with a diverse set of anode fuels.
21. T. Bein, Angew. Chem. Int. Ed. Engl., 38, 323 ͑1999͒.
22. P. Claus, D. Honicke, and T. Zech, Catal. Today, 67, 319 ͑2001͒.
23. C. M. Snively, G. Oskarsdottir, and J. Lauterbach, Catal. Today, 67, 357 ͑2001͒.
24. A. Holzwarth, H.-W. Schmidt, and W. F. Maier, Angew. Chem. Int. Ed. Engl., 37,
2644 ͑1998͒.
25. P. Cong, R. D. Doolen, Q. Fan, D. M. Giaquinta, S. Guan, E. W. McFarland, D. M.
Poojary, K. Self, H. W. Turner, and W. H. Weinberg, Angew. Chem. Int. Ed. Engl.,
38, 484 ͑1999͒.
26. S. Senkan, K. Krantz, S. Ozturk, V. Zengin, and I. Onal, Angew. Chem. Int. Ed.
Engl., 38, 2794 ͑1999͒.
27. A. Nayar, R. Liu, R. J. Allen, M. J. McCall, R. R. Willis, and E. S. Smotkin, Anal.
Chem., 74, 1933 ͑2002͒.
Acknowledgments
28. M. G. Sullivan, H. Utomo, P. J. Fagan, and M. D. Ward, Anal. Chem., 71, 4369
͑1999͒.
The authors would like to acknowledge the Office of Naval Re-
search for a Young Investigator Award, the National Science Foun-
dation for a CAREER Award, the Camille and Henry Dreyfus Foun-
dation for a New Faculty Award, and the Donors of The Petroleum
Research Fund as administered by the American Chemical Society,
for partial support of this research.
29. S. Jayaraman and A. C. Hillier, Langmuir, 17, 7857 ͑2001͒.
30. B. C. Shah and A. C. Hillier, J. Electrochem. Soc., 147, 3043 ͑2000͒.
31. Y. Gao, H. Tsuji, H. Hattori, and H. Kita, J. Electroanal. Chem., 372, 195 ͑1994͒.
32. A. D. Modestov, S. Srebnik, O. Lev, and J. Gun, Anal. Chem., 73, 4229 ͑2001͒.
33. L. Grambow and S. Bruckenstein, Electrochim. Acta, 22, 377 ͑1977͒.
34. O. Wolter and J. Heitbaum, Ber. Bunsenges. Phys. Chem., 88, 2 ͑1984͒.
35. V. M. Schmidt, R. Ianniello, E. Pastor, and S. Gonzalez, J. Phys. Chem., 100,
17901 ͑1996͒.
36. W. Z. Lu, X. M. Xu, and R. B. Cole, Anal. Chem., 69, 2478 ͑1997͒.
37. Z. Jusys, H. Massong, and H. Baltruschat, J. Electrochem. Soc., 146, 1093 ͑1999͒.
38. R. Ianniello and V. M. Schmidt, Ber. Bunsenges. Phys. Chem., 99, 83 ͑1995͒.
39. B. Bittins-Cattaneo, E. Cattaneo, P. Konigshoven, and W. Vielstich, in Electroana-
lytical Chemistry. A Series of Advances, A. J. Bard, Editor, p. 180, Marcel Dekker,
Inc., New York, ͑1991͒.
The University of Virginia assisted in meeting the publication costs of
this article.
References
1. P. N. Ross, in Electrocatalysts, P. N. Ross and J. Lipkowski, Editors, p. 1, Wiley,
New York ͑1998͒.
2. A. J. Appleby and F. R. Foulkes, Fuel Cell Handbook, Van Nostrand Reinhold,
40. H. Wang, T. Loffler, and H. Baltruschat, J. Appl. Electrochem., 31, 759 ͑2001͒.
41. S. Wasmus and W. Vielstich, Electrochim. Acta, 38, 185 ͑1993͒.
42. N. Fujiwara, K. A. Friedrich, and U. Stimming, J. Electroanal. Chem., 472, 120
͑1999͒.
43. Z. Jusys and R. J. Behm, J. Phys. Chem., 105, 10874 ͑2001͒.
44. Z. Jusys, J. Kaiser, and R. J. Behm, Langmuir, Submitted.
45. J. Willsau, O. Wolter, and J. Heitbaum, J. Electroanal. Chem. Interfacial Electro-
chem., 185, 163 ͑1985͒.
46. S. Wilhelm, W. Vielstich, H. W. Buschmann, and T. Iwasita, J. Electroanal. Chem.
Interfacial Electrochem., 229, 377 ͑1987͒.
47. T. Iwasita, W. Vielstich, and E. Santos, J. Electroanal. Chem. Interfacial Electro-
chem., 229, 367 ͑1987͒.
New York ͑1989͒.
3. J. Larminie and A. Dicks, Fuel Cell Systems Explained, John Wiley & Sons, Ltd.,
New York ͑2000͒.
4. H. Kita, H. Naohara, T. Nakato, S. Taguchi, and A. Aramata, J. Electroanal. Chem.,
386, 197 ͑1995͒.
´
5. N. M. Markovic, B. N. Grgur, C. A. Lucas, and P. N. Ross, J. Phys. Chem. B, 103,
487 ͑1999͒.
6. M. Watanabe, M. Uchida, and S. Motoo, J. Electroanal. Chem. Interfacial Electro-
chem., 229, 395 ͑1987͒.
7. H.-F. Oetjen, V. M. Schmidt, U. Stimming, and F. Trila, J. Electrochem. Soc., 143,
3838 ͑1996͒.
8. Y. Y. Tong, H. S. Kim, P. K. Babu, P. Waszczuk, A. Wieckowski, and E. Oldfield,
J. Am. Chem. Soc., 124, 468 ͑2002͒.
9. C. Roth, N. Martz, F. Hahn, J.-M. Leger, C. Lamy, and H. Fuess, J. Electrochem.
48. J. L. Rodriguez and E. Pastor, Electrochim. Acta, 44, 1173 ͑1998͒.
49. W. Chrzanowski and A. Wieckowski, Langmuir, 13, 5974 ͑1997͒.
50. E. Herrero, J. M. Feliu, and A. Wieckowski, Langmuir, 15, 4944 ͑1999͒.
´
Soc., 149, E433 ͑2002͒.
10. H. A. Gasteiger, N. M. Markovic, and P. N. Ross, J. Phys. Chem., 99, 8945 ͑1995͒.
11. Y. Morimoto and E. B. Yeager, J. Electroanal. Chem., 444, 95 ͑1998͒.
12. B. N. Grgur, N. M. Markovic, and P. N. Ross, J. Phys. Chem. B, 102, 2494 ͑1998͒.
13. B. N. Grgur, G. Zhuang, N. M. Markovic, and P. N. Ross, J. Phys. Chem. B, 101,
3910 ͑1997͒.
14. H. Nakajima and H. Kita, Electrochim. Acta, 35, 849 ͑1990͒.
15. B. Gurau, R. Viswanathan, R. Liu, T. S. Lafrenz, K. L. Ley, E. S. Smotkin, E.
Reddington, A. Sapienza, B. C. Chan, T. E. Mallouk, and S. Sarangapani, J. Phys.
Chem. B, 102, 9997 ͑1998͒.
ˇ ´
51. S. R. Brankovic, J. X. Wang, Y. Zhu, R. Sabatini, J. McBreen, and R. R. Adzic, J.
Electroanal. Chem., 524, 231 ͑2002͒.
52. S. R. Brankovic, J. X. Wang, and R. R. Adzic, Electrochem. Solid-State Lett., 4,
A217 ͑2001͒.
53. E. L. Cussler, Diffusion Mass Transfer in Fluid Systems, Cambridge University
Press, Cambridge, UK, ͑1997͒.
54. B. A. Ogunnaike and H. W. Ray, Process Dynamics, Modeling, and Control, Ox-
ford University Press, New York ͑1994͒.
55. H. Baltruschat and U. Schmiemann, Ber. Bunsenges. Phys. Chem., 97, 452 ͑1993͒.
56. I. Villegas and M. J. Weaver, J. Chem. Phys., 101, 1648 ͑1994͒.
57. C. A. Lucas, N. M. Markovic, and P. N. Ross, Surf. Sci., 425, L381 ͑1999͒.
58. V. S. Bagottzky and Y. B. Vassilyev, Electrochim. Acta, 12, 1323 ͑1967͒.
59. C. P. Wilde and M. Zhang, Electrochim. Acta, 39, 347 ͑1994͒.
60. S. Wasmus, J. T. Wang, and R. F. Savinell, J. Electrochem. Soc., 142, 3825 ͑1995͒.
61. C. Korzeniewski and C. L. Childers, J. Phys. Chem. B, 102, 489 ͑1998͒.
¨
16. Z. Jusys, T. J. Schmidt, L. Dubau, K. Lasch, L. Jorissen, J. Garche, and R. J. Behm,
J. Power Sources, 105, 297 ͑2002͒.
17. A. Lima, C. Coutanceau, J. M. Leger, and C. Lamy, J. Appl. Electrochem., 31, 379
͑2001͒.
18. W. F. Maier, Angew. Chem. Int. Ed. Engl., 38, 1216 ͑1999͒.
19. B. Jandeleit, D. J. Schaefer, T. S. Powers, H. W. Turner, and W. H. Weinberg,
Angew. Chem. Int. Ed. Engl., 38, 2495 ͑1999͒.
Downloaded on 2015-05-28 to IP 128.255.6.125 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).