BNL-112220-2016-JA

Fuel Cell Performance of Palladium-Platinum

Core-Shell Electrocatalysts Synthesized in

Gram-Scale Batches

Siddique Khateeb, Sandra Guerreo, Dong Su,

Robert M. Darling, Lesia V. Protsailo, Minhua Shao

Submitted to the ACS catalysis

April 2016

Center for Functional Nanomaterials

Brookhaven National Laboratory

U.S. Department of Energy USDOE Office of Science (SC),

Basic Energy Sciences (BES) (SC-22)

Notice: This manuscript has been authored by employees of Brookhaven Science Associates, LLC under

Contract No. DE- SC0012704 with the U.S. Department of Energy. The publisher by accepting the

manuscript for publication acknowledges that the United States Government retains a non-exclusive, paid-up,

irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others

to do so, for United States Government purposes.

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the

United States Government. Neither the United States Government nor any

agency thereof, nor any of their employees, nor any of their contractors,

subcontractors, or their employees, makes any warranty, express or implied, or

assumes any legal liability or responsibility for the accuracy, completeness, or any

third party’s use or the results of such use of any information, apparatus, product,

or process disclosed, or represents that its use would not infringe privately owned

rights. Reference herein to any specific commercial product, process, or service

by trade name, trademark, manufacturer, or otherwise, does not necessarily

constitute or imply its endorsement, recommendation, or favoring by the United

States Government or any agency thereof or its contractors or subcontractors.

The views and opinions of authors expressed herein do not necessarily state or

reflect those of the United States Government or any agency thereof.

Fuel Cell Performance of Palladium-Platinum Core-Shell

Electrocatalysts Synthesized in Gram-Scale Batches

Siddique Khateeb,1 Sandra Guerreo,

1 Dong Su,

2 Robert M. Darling,

1,† Lesia V. Protsailo,

1,†

Minhua Shao3, *

1UTC Power, 195 Governor’s Highway, South Windsor, CT 06074, USA

2Center for Functional Nanomaterials, Brookhaven National Laboratory, Upton, NY, 11794,

USA

3Department of Chemical and Biomolecular Engineering, The Hong Kong University of Science

and Technology, Clear Water Bay, Kowloon, Hong Kong

†Current address: United Technologies Research Center, East Hartford, CT 06108, USA

Email: kemshao@ust.hk; +852-34692269

Abstract This paper presents the performance of palladium-platinum core-shell catalysts

(Pt/Pd/C) for oxygen reduction synthesized in gram-scale batches in both liquid cells and

polymer-electrolyte membrane fuel cells. Core-shell catalyst synthesis and characterization, ink

fabrication, and cell assembly details are discussed. The Pt mass activity of the Pt/Pd core-shell

catalyst was 0.95 A mg-1

at 0.9 V measured in liquid cells (0.1 M HClO4), which was 4.8 times

higher than a commercial Pt/C catalyst. The performances of Pt/Pd/C and Pt/C in large single

cells (315 cm2) were assessed under various operating conditions. The core-shell catalyst showed

consistently higher performance than commercial Pt/C in fuel cell testing. A 20-60 mV

improvement across the whole current density range was observed on air. Sensitivities to

temperature, humidity, and gas composition were also investigated and the core-shell catalyst

showed a consistent benefit over Pt under all conditions. However, the 4.8 times activity

enhancement predicated by liquid cell measurements was not fully realized in fuel cells.

Keywords: Core-shell, polymer electrolyte membrane fuel cell, platinum monolayer, oxygen

reduction reaction

1. Introduction

Polymer electrolyte membrane (PEM) fuel cells face hurdles to commercialization in automotive

applications in part because of the high cost of Pt-based catalysts used in the anode and

cathode.1-3

Reducing the amount of Pt in PEM fuel cells has long been a focus among

automotive companies and researchers.4, 5

One of strategies to reduce the amount of Pt is to

increase the rate of oxygen reduction reaction per unit mass of Pt. Core-shell catalysts that

consist of an ultrathin Pt shell and a Pd core have been proposed as candidates to replace pure

Pt.6-10

In addition to the high Pt utilization and lower cost of the Pd core, strain and electronic

effects induced by the Pd core can enhance the activity of the Pt shell for the ORR,11

leading to a

greater activity per mass of Pt than expected based on the utilization improvement alone.

The Cu-mediated-Pt-displacement method developed by Adzic et al. can theoretically

produce a Pt monolayer on a Pd core.7 In this method, a Cu monolayer deposited on a Pd core is

displaced by Pt via a surface limited redox replacement (SLRR) reaction: Cu/Pd + PtCl42-

→

Pt/Pd + Cu2+

+ 4Cl-. Most core-shell catalysts made by the Cu-Pt-displacement method were

synthesized on a microgram scale, i.e., with the Cu UPD and Pt displacement reactions on a

rotating disk electrode (RDE) tip. When this process was transferred to the gram scale, the high

activity of core-shell catalysts was not fully realized. So far, only a limited amount of core-shell

materials have been synthesized at a gram scale and tested in real fuel cells.12-19

For instance,

Naohara et al.13

firstly reported the synthesis of Pt/Pd/C on a gram scale with a Pt mass activity

around 0.6 A mg-1

at 0.9 V (measured in RDE), which was lower than that of the same type of

material synthesized at a RDE tip and measured by the same group (0.96 A mg-1

).20

Furthermore,

the Pt mass activity (0.3 A mg-1

) of the Pt/Pd/C catalyst measured in a fuel cell was only half of

that measured in RDE.15

Various issues related to scaling the batch size have resulted in less

active and stable catalysts. Other possible reasons include differences in testing protocols and

test conditions between RDE and fuel cell.17

Thus, an increasing focus on synthesis at large

scales and understanding the fuel-cell performance is necessary.

Ideally, a uniform Pt monolayer is deposited on the core after all Cu atoms are replaced in the

SLRR reaction. However, our in situ XRD results in an H2 environment demonstrated that Pt

clusters rather than a uniform Pt shell was formed on Pd nanoparticles. 12

The mechanism

underlying this observation is that the SLRR process involves electron transfer from the substrate

(Pd in this case) to PtCl42-

ions, rather than direct electron exchange from Cu.21

This means that

electrons generated anywhere on the surface can move freely through the Pd substrate, reducing

PtCl42-

ions wherever their activity and surface energy are greatest. In other words, the Pt atom

may not deposit on the same site left by Cu dissolution, but rather on Pt that was already

deposited on the core leading to the formation of Pt clusters. Due to incomplete Pt coverage and

low Pt usage, clustered catalysts are expected to have lower activity and stability than perfect

core-shell structures. Thus, the key to synthesizing core-shell materials with good quality on a

large scale is to force Pt atoms to deposit on the surface of the core rather than on Pt atoms

already deposited by manipulating the Cu-Pt displacement reaction mechanisms.

This paper reports the synthesis of highly active Pt/Pd/C core-shell catalysts on a gram scale

by employing additives in the Cu-Pt displacement solution. Their characterization and evaluation

in both liquid and fuel cells are also reported.

2. Experimental Section

2.1 Core-Shell Catalyst Synthesis

A custom reactor was built to synthesize gram-scale batches of core-shell catalysts using

commercial Pd nanoparticles supported on Ketjen Black (Pd/C) as the core materials (TKK, 35

wt.%). The reactor comprised a graphite sheet as the working electrode, carbon cloth as the

counter electrode, and an Ag/AgCl leak-free reference electrode (BASi). The procedure of core-

shell catalyst synthesis is illustrated in Figure 1. The Pd/C nanoparticles were dispersed in a 50

mM H2SO4 solution, and then added to the reactor. Pretreatment of the Pd/C was conducted as

needed to wet the powder. After pretreatment, a deaerated aqueous CuSO4 solution in 50 mM

H2SO4 was added to the reactor to obtain a Cu2+

concentration of 50 mM. After addition of the

CuSO4, the potential was held at approximately 0.37 V vs. Ag/AgCl to deposit an atomically thin

layer of Cu on the Pd nanoparticles. Immediately upon completion of the Cu UPD (under

potential deposition), a deaerated aqueous solution of K2PtCl4 was added dropwise to the reactor

to perform galvanic replacement of the Cu by the Pt. In addition to the Pt precursor, citric acid

was also included in the Pt precursor solution to minimize the formation of Pt clusters.12, 22

After

the reaction was complete, typically in 30 minutes, catalysts were filtered and rinsed with

ultrapure water (Milli-Q® UV-plus water). The resulting catalyst powder, denoted as Pt/Pd/C,

was dried in a vacuum at 60°C. Only a trace amount of Cu (less than 0.4%) remained in the final

product based on inductively coupled plasma – mass spectrometry (ICP-MS) measurements.

2.2 Catalyst Characterization

XRD experiments were performed at Beamline X18A at the National Synchrotron Light

Source (NSLS) at Brookhaven National Laboratory (BNL) in Upton, NY. A detailed

experimental procedure can be found elsewhere.12

Briefly, helium (99.999%, Praxair) and

hydrogen (99.999%, Praxair) were metered through a home-made sample cell by two mass flow

controllers (Brooks 5850) at a flow rate of ~20 mL min-1

. The atmosphere was first purged with

He (helium) to prevent burning the catalyst powders via the combustion of H2. The XRD spectra

were recorded with a 2D detector (Perkin Elmer XRD 1621 N ES) at 20 keV (calibrated by the

Mo K absorption (λ = 0.7107 Å)) and subsequently processed by standard software. The dark

current was collected over 40 seconds (10 exposures at 4 s per exposure) and the diffraction

pattern was recorded over 400 seconds (100 exposures at 4 s per exposure). After recording the

XRD pattern under an inert atmosphere, the cell was purged with H2 for 10 minutes and the XRD

spectra were then recorded under this atmosphere. Further purging with H2 did not affect the

structure of the material. Finally, the cell was purged with He for 10 minutes and a post-H2

exposure XRD spectrum was recorded. The sample to detector distance was approximately 400

mm and was calibrated by the use of titanium dioxide (rutile) powder loaded into the sample

holder of the cell in an identical fashion as done for catalyst characterization. Performing a

calculation of error propagation suggests that an error in sample position of +/- 1 mm will result

in an error in the lattice constant of +/- 0.001 nm. Further, there is an internal calibration

standard for each sample corresponding to amorphous carbon peak, which is present in all of the

XRD spectra collected in this work, and which always lies at the same value of 2θ.

High angle annular dark-field scanning transmission electron microscopy (HAADF-STEM)

images and electron energy loss spectroscopy (EELS) data were collected using a Hitachi

aberration-corrected scanning transmission electron microscope (HD-2700C). An electron probe

diameter of 1.3 Å with a convergence angle of 28 mrad was used. At these experimental

conditions, the contrast present in the HAADF images is directly related to the atomic number

and thickness of the materials. The EELS line scans were collected with a pixel time of 0.02 s

and 1 s, respectively, using a Gatan Enfina-ER detector.

2.3 Electrochemical Evaluation

Approximately 5 mg of catalyst was ultrasonically dispersed in a solvent consisting of 5 mL

of water, 2.5 mL of isopropanol, and 30 µL of 5% Nafion (Aldrich) for 10 min. 5 or10 µL of the

suspension was deposited on a pre-cleaned glassy carbon rotating disk electrode (Pine

Instruments) and allowed to dry in air. The electrode was cycled between 0.02 and 1.2 V for 20

cycles in a nitrogen-saturated aqueous solution of 0.1 M HClO4 at 100 mV s-1

. Then a stable

cyclic voltammetry (CV) curve was recorded at 50 mV s-1

. The charges associated with

hydrogen adsorption/desorption were used to calculate the electrochemical areas (ECAs) of all

the samples evaluated in this paper, assuming 210 μC cm-2

for full HUPD coverage. The oxygen

reduction polarization curves were measured in an oxygen-saturated 0.1 M HClO4 (GFS

Chemicals) solution at a scan rate of 10 mV s-1

at 1600 rpm. The kinetic current jk at 0.9 V was

derived from the Koutecky-Levich equation:

where j is the measured current density, B is a constant, and ω is the rotation rate. A reversible

hydrogen electrode (RHE) and Pt gauze were used as the reference and counter electrodes,

respectively. All potentials were corrected to RHE in this paper. Long term durability of Pt/Pd/C

was tested by potential cycling. A square-wave signal with 5 s each at 0.65 and 1.0 V was

applied on the RDE with Pt/Pd/C in an oxygen-saturated 0.1 M HClO4 solution at the room

temperature. CV and ORR polarization curves were recorded after a certain number of cycles.

2.4 Fuel Cell Tests

Pt/Pd/C core-shell catalyst was mixed with ionomer (Ion Power) in a glass vial with a specific

ionomer/carbon (I/C) ratio of 0.8. Ethanol/water solvent was added into the vial and the ink was

dispersed and mixed by placing the vial in an ultrasonic water bath. The temperature of water in

the bath was maintained at the room temperature and subjected to sonication for a few hours

until the desired catalyst agglomerate size was achieved.

Once the catalyst ink achieved the desired agglomerate particle size, the ink was ready to be

sprayed on a membrane to make the membrane electrode assembly (MEA). The membrane was

obtained from a commercial vendor containing anode catalyst on one side with a Pt loading of

0.1 mg cm-2

and a geometric area of 310 cm2. The cathode catalyst ink was sprayed on the bare

2/1

111

Bjj k

membrane side using an automated sprayer manufactured by SonotekTM

. A few layers of catalyst

ink were uniformly sprayed to obtain a Pt loading of 0.1 mg cm-2

. During spraying, the solvent

evaporated inside the spray chamber. Next, the MEA was hot pressed in a pneumatic mechanical

press to ensure sufficient mechanical bonding between the membrane and the catalyst layer.

The MEA was finally assembled with other components as shown in Figure 2. The entire cell

package after assembly was loaded to a pressure of 400 kPa to ensure good electrical and

mechanical contact between various components to minimize Ohmic losses and ensure uniform

pressure distribution.

To understand effects of temperature and air humidity on performance, cells were tested

under various operating conditions as shown in Table 1. Operating a PEM fuel cell using dry air

without an external humidifier is important for automotive applications to reduce system costs

and complexity. Thus, numerous tests were performed with dry air at different temperatures and

humidity to map the responses of cells. Limited test results are reported here and more results

will be published in a separate paper.

3 Results and Discussion

3.1 Pt/Pd/C Core-Shell Catalyst Characterization

Typical Pt and Pd loadings of the Pt/Pd/C catalyst are 18% and 27%, respectively, measured

by ICP-MS. If one monolayer of Pt were deposited on the Pd/C core, the corresponding Pt

loading should be ~12% based on the Cu UPD charge on the Pd core. The higher Pt loading in

the final core-shell product indicates that the thickness of the Pt shell is between 1 and 2 atomic

layers. Figure 3a shows typical core-shell catalyst dispersion in a HAADF-STEM image. The

inset of Figure 3b shows the HAADS-STEM image of a single core-shell particle and its EELS

line profile of Pt across the particle (Figure 3b). The sharp peaks at the two sides of the profile

indicate the Pt shell on the Pd core. The thickness of the Pt shell determined in the line scan is

0.5±0.15 nm, which is equivalent to ~2 atomic layers and is consistent with the ICP data.

The quality of the Pt shell was also evaluated using a qualitative method developed recently,12

which utilizes the ability of Pd to absorb hydrogen to form a Pd hydride structure. Figure 4

compares the XRD pattern of Pt/Pd/C in He and H2 atmospheres measured in a customized cell.

The absence of large Pt clusters in Pt/Pd/C was confirmed by the uniform shift of the XRD peak.

One would not expect the Pt peak (same position in the He atmosphere) to shift since Pt does not

absorb hydrogen resulting in a shoulder in the higher angle side of the main peak. A shoulder at

the position of Pt peak in H2 was clearly observed in a previous study of Pt/Pd/C catalysts

synthesized without citric acid in the Cu-Pt displacement step. 12

3.2 Pt/Pd/C Electrochemical Evaluation

Figure 5a compares CVs (cyclic voltammograms) of Pt/Pd/C and its core Pd/C in a nitrogen-

saturated 0.1 M HClO4 solution. The oxide layer formation and its reduction shifted to a more

positive potential when a Pt shell was deposited on the Pd/C. The delayed oxide layer formation

on the catalyst surface resulted in a significant improvement in ORR activity.

Figure 5b compares the oxygen reduction polarization curves (positive scan) of Pt/Pd/C and

state-of-the-art Pt/C (TEC10E50E, TKK) in an oxygen-saturated 0.1 M HClO4 solution. The

ORR curves were very similar for these two catalysts even though the Pt loading of the Pt/C (24

µg cm-2

) was 6 times higher than that of the Pt/Pd/C (4.0 µg cm-2

). The Pt mass and area specific

activities (kinetic current normalized to electrochemical active area (ECA)) were compared in

Figure 5c . The Pt mass activity of Pt/Pd/C was 0.95 ± 0.10 A mg-1

Pt, which was 4.8 times higher

than that of Pt/C (0.2 ± 0.02 A mg-1

Pt) at 0.9 V. The ECA of Pt/Pd/C was 160 ± 10 m2 g

-1Pt,

while the ECA of Pt/C was 85 ± 2 m2 g

-1Pt. Since the theoretical ECA of a Pt monolayer is 240

m2/g, the measured Pt/Pd ECA is consistent with a 1.5 monolayer thick shell. In addition to a

higher ECA, the core-shell also has a higher specific activity (0.6 mA cm2

Pt) than Pt (0.24 mA

cm2

Pt). The price of platinum is currently ~1.6 times the price of palladium making the core shell

2.6 times more active than pure Pt on a cost basis. It is worth noting that the ratio of platinum to

palladium prices has varied from 0.5 to 5.5 since 1986.

The stability of the core-shell catalyst was tested with a square-wave signal between 0.65 and

1.0 V. After 10000 cycles, the Pt mass activity dropped by 16% from 0.95 ± 0.10 to 0.8 ± 0.10 A

mg-1

. Pt/Pd/C was more stable than Pt/C, which degraded 42% under the same conditions (Figure

5d).

3.3 Cell Test Results

The performances of cells containing Pt/Pd/C core-shell catalyst (18% Pt, 27% Pd) and the

conventional Pt/C catalyst (20% Pt, TKK) on oxygen and air are compared in Figure 6. Both

cathode catalyst layers contain 0.1±0.03 mg cm-2

Pt. The coolant exit temperature was

maintained at 80oC, the coolant flow rate was 300 mL min

-1, gases entered the cell with dew

points of 80oC (yielding approximately 100% relative humidity at the local cell temperature), and

gases exited the cell at 140 kPaabs. Oxygen utilization was 50% in the pure oxygen experiment

and 67% in the air experiment, while hydrogen utilization was 67% in the pure oxygen

experiment and 83% in the air experiment. These are tests 1 and 2 in Table 1. The core-shell

cell is consistently better than the Pt cell on both oxygen and air. The oxygen curves are nearly

vertically offset, which is consistent with improved oxygen-reduction kinetics. The carbon

support, ionomer, and ionomer to carbon mass ratio are identical in the two catalyst layers. The

estimated electrode thicknesses, assuming 50% porosity, are 5.2 and 6.5 m for the core-shell

and Pt catalyst layers, respectively. The visible similarity of the polarization curves measured on

air and oxygen indicates that the two electrodes achieved similar Ohmic and oxygen-transport

losses, consistent with design intent. Switching from air to oxygen should boost performance by

45 mV in the kinetic region for a Tafel slope of 67 mV/decade. The platinum electrode showed

this expected behavior, but an unexpectedly small boost was observed on the core-shell

electrode. This discrepancy could not be reproduced at the RDE level by switching from oxygen

to air and remains unexplained.

Table 2 compares electrochemical areas and mass activities at 0.9 V in RDE and in cell. Mass

activities were estimated for air and oxygen at the temperature of the cell assuming Tafel slopes of

65 mV decade-1

. Mass activities in H2/air cells were corrected for oxygen concentration. For the Pt

cell, mass activities estimated from the cell data on air and oxygen were similar and consistent with

the corresponding RDE measurements. For the core-shell cell, a substantially higher mass activity

was estimated from the air data than from the oxygen data. The mass activity determined with RDE

exceeded both of these values. Hydrogen adsorption measurements on smaller (25 cm2) electrodes

gave electrochemical surface areas of 70 m2

gPt-1

for the core-shell and 50 m2 gPt

-1 for Pt. Both of

these ECAs are significantly lower than those measured in RDE, but the difference is more drastic

for the core shell. Certainly, the low core-shell ECA observed in the cell is partially responsible for

not achieving the expected mass activity. The relatively low ECA of the core shell in the electrode

could be due to an inability of the polymer electrolyte to contact some sites that can be contacted by

liquid electrolyte, or damage occurring during ink fabrication or spraying.

Figure 7 shows voltage differences between the core-shell cell and the Pt cell as a function of

current density on oxygen, air, and helox (21% O2 in He). Application of various gases enables a

rough separation of various losses. Transport in the catalyst layer is complex, involving Knudsen

diffusion of gases in small (< 100 nm) pores, permeation of oxygen through ionomer, proton

movement in ionomer, and multiphase water transport.23

The performance on oxygen is expected to

be dominated by kinetic and Ohmic losses. Switching from pure oxygen to helox activates oxygen

transport losses in ionomer and in nanopores. Normal gas-phase diffusion losses become important

when the balance gas is switched from helium to nitrogen. The differences are always positive as

the core-shell electrode is more active than the Pt electrode. The voltage difference on oxygen is

approximately 10 mV. This is consistent with the core-shell electrode having a mass activity that is

1.4X higher than the Pt electrode. A difference of 44 mV would be consistent with the 4.8X higher

activity observed in RDE. A higher Ohmic loss is expected in the Pt catalyst layer because it is ~1

m thicker. The sheet resistance of the catalyst layer should increase by ~2 m-cm2 for an ionomer

volume fraction of 20% and an ionomer conductivity of 0.17 S cm-1

. The slope of the oxygen

difference curve in Figure 7 is about 2.5 m-cm2.

Oxygen transport losses are minimal when pure oxygen is fed to the cell. Normal gas-phase

transport losses remain unimportant when the oxidant is switched to helox (21% O2 in He), however

oxygen transport losses in the ionomer and in small pores governed by Knudsen diffusion become

important. The Ohmic losses in the cathode can also increase as the current shifts away from the

membrane. The slope of the difference curve on helox is approximately 4X larger than the slope on

oxygen, indicating that additional transport losses become important. The offset also increases from

approximately 10 mV on oxygen to 20 mV on helox. The activity of the core-shell electrode appears

to be higher on helox than oxygen. A 20 mV offset corresponds to an activity improvement of 2X,

which is less than half of the gain observed on RDE. Normal gas diffusion losses become important

when the balance gas is switched from helium to nitrogen. The air and helox curves are similar

below 1.2 A/cm2, but they diverge at higher current densities. The activities on air and helox thus

appear to be similar; however additional transport losses appear on air at high current densities. The

data indicates that transport losses are lower in the core-shell electrode than the Pt electrode. Thus,

the inability of the core shell to achieve the activity observed in RDE does not carry over to other

types of polarization.

Figure 8 shows how the performances of the two cells change with temperature. While the core-

shell cell is consistently better than the Pt/C cell, the difference is relatively insensitive to

temperature and in general agreement with the polarization curves. Much of the difference can be

attributed to the kinetic improvement that attends changing from Pt to core-shell. The relative

humidity in the cell drops as temperature increases because the inlet dew point is constant.

Consequently, Ohmic resistances of the cells increase and the voltages decrease above ~75oC. The

optimum temperature increases with increasing current density because the water concentration

profile polarizes and rises in the MEA. The voltage declines at 60oC and 1.6 A cm

-2 because the

saturation of the various porous media with liquid water increasing.

Figure 9 shows how performances of the two cells change with the relative humidity of the inlet

air and fuel. The kinetic benefit of the core-shell catalyst is apparent as a vertical offset between the

Pt and core-shell cells. There appears to be a slight tendency for the gap between the two cells to

increase when the feed gases are dry at all current densities.

4. Conclusions

Core-shell catalysts consisting of an ultrathin Pt shell and a Pd core were successfully

synthesized using a gram-scale reactor involving the displacement of a Cu UPD layer. The XRD,

ICP, and TEM results showed that Pt/Pd core-shell catalyst had a nominal composition of 18

wt.% Pt and 27 wt.% Pd. Uniform Pt coverage 1-2 atomic monolayers thick (0.5 ± 0.15 nm) on

Pd was demonstrated by HAADF-EELS and in situ XRD. The mass activity of Pt/Pd/C core-

shell was measured to be 0.95 ± 0.10 A mg-1

, which was 4.8 times higher than that of Pt/C at 0.9

V using RDE in a 0.1 M HClO4 solution. The electrochemical area was 160 ± 10 m2 g

-1Pt, which

is double that of commercial Pt/C. Full scale fuel cell tests were performed to compare the

Pt/Pd/C core-shell catalyst with regular Pt/C catalyst under various temperature and humidity

conditions. The core-shell catalyst was approximately 20 and 60 mV higher than Pt/C in the low

and higher current density regions on H2-air, respectively. The cell with core-shell catalyst did

not show the improvement at the low current densities as observed in RDE. The lower than

predicted performance may be due to ink fabrication and spraying steps. Test results with

different ink preparation and MEA fabrication processes will be summarized in a separate

publication. Overall, Pt/Pd/C core-shell catalyst performs better than regular Pt/C catalyst under

both high and low humidity conditions at all temperatures tested (60 – 85ºC).

Acknowledgments

Use of the National Synchrotron Light Source, Brookhaven National Laboratory, was supported

by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under

Contract No. DE-AC02-98CH10886. The authors would like to thank Marianne Pemberton and

Michael Humbert for helping synthesis and evaluate the catalysts. The work at the Hong Kong

University of Science and Technology was supported by Research Grant Council of the Hong

Kong Special Administrative Region (IGN13EG05 and 26206115) and a startup fund from the

Hong Kong University of Science and Technology. Use of the National Synchrotron Light

Source (NSLS) and Center for Functional Nanomaterials (CFN) at Brookhaven National

Laboratory was supported by the U.S. Department of Energy, Office of Science, Office of Basic

Energy Sciences, under Contract No. DE-AC02-98CH10886. Beamline X18A at the NSLS is

supported in part by the Synchrotron Catalysis Consortium (DOE BES grant DE-FG02-

03ER15688).

References

1. M. K. Debe, Nature, 486, 43 (2012). 2. H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, Appl. Catal. B, 56, 9 (2005). 3. F. T. Wagner, B. Lakshmanan and M. F. Mathias, J Phys. Chem. Lett., 1, 2204 (2010). 4. , in Electrocatalysis in Fuel Cells: A Non- and Low- Platinum Approach, M. Shao Editor, Springer, London (2013). 5. M. Shao, Q. Chang, J.-P. Dodelet and R. Chenitz, Chem. Rev., 116, 3594 (2016). 6. R. R. Adzic, Electrocatalysis, 3, 163 (2012). 7. R. R. Adzic, J. Zhang, K. Sasaki, M. B. Vukmirovic, M. Shao, J. X. Wang, A. U. Nilekar, M. Mavrikakis, J. A. Valerio and F. Uribe, Top. Catal., 46, 249 (2007). 8. L. Liu, G. Samjeske, S.-i. Nagamatsu, O. Sekizawa, K. Nagasawa, S. Takao, Y. Imaizumi, T. Yamamoto, T. Uruga and Y. Iwasawa, J. Phys. Chem. C, 116, 23453 (2012). 9. R. Choi, S.-I. Choi, C. H. Choi, K. M. Nam, S. I. Woo, J. T. Park and S. W. Han, Chemistry – A European Journal, 19, 8190 (2013). 10. J. Park, L. Zhang, S.-I. Choi, L. T. Roling, N. Lu, J. A. Herron, S. Xie, J. Wang, M. J. Kim, M. Mavrikakis and Y. Xia, ACS Nano, 9, 2635 (2015). 11. J. L. Zhang, M. B. Vukmirovic, Y. Xu, M. Mavrikakis and R. R. Adzic, Angew. Chem. Int. Ed., 44, 2132 (2005).

12. M. Humbert, B. Smith, Q. Wang, S. Ehrlich and M. Shao, Electrocatalysis, 3, 298 (2012). 13. H. Naohara, K. Sasaki and R. R. Adzic, in 216th ECS Meeting, Vienna, Austria (2009). 14. W.-P. Zhou, K. Sasaki, D. Su, Y. Zhu, J. X. Wang and R. R. Adzic, J Phys. Chem. C, 114, 8950 (2010). 15. K. Sasaki, H. Naohara, Y. Cai, Y. M. Choi, P. Liu, M. B. Vukmirovic, J. X. Wang and R. R. Adzic, Angew. Chem. Int. Ed., 49, 8602 (2010). 16. K. Sasaki, H. Naohara, Y. Choi, Y. Cai, W.-F. Chen, P. Liu and R. R. Adzic, Nat. Commun., 3 (2012). 17. S. Ball, in Electrocatalysis in Fuel Cells: A Non- and Low- Platinum Approach, M. Shao Editor, Springer, London (2013). 18. S. C. Ball, S. L. Burton, E. Christian, A. Davies, J. Fisher, R. O'Malley, S. Passot, B. C. Tessier, B. Theobald and D. Thompsett, ECS Transactions, 25, 1011 (2009). 19. K. Sasaki, J. Wang, H. Naohara, N. Marinkovic, K. More, H. Inada and R. Adzic, Electrochim. Acta, 55, 2645 (2010). 20. J. X. Wang, H. Inada, L. Wu, Y. Zhu, Y. Choi, P. Liu, W.-P. Zhou and R. R. Adzic, J. Am. Chem. Soc., 131, 17298 (2009). 21. C. Thambidurai, D. K. Gebregziabiher, X. Liang, Q. Zhang, V. Ivanova, P.-H. Haumesser and J. L. Stickney, J. Electrochem. Soc., 157, D466 (2010). 22. M. Shao, B. H. Smith, S. Guerrero, L. Protsailo, D. Su, K. Kaneko, J. H. Odell, M. P. Humbert, K. Sasaki, J. Marzullo and R. M. Darling, Phys. Chem. Chem. Phys., 15, 15078 (2013). 23. N. Nonoyama, S. Okazaki, A. Z. Weber, Y. Ikogi and T. Yoshida, J. Electrochem. Soc., 158, B416 (2011).

Figure Captions

Figure 1. Steps involved in core-shell catalyst fabrication: (1) Pd/C dispersion; (2) Pd/C

reduction; (3) Cu UPD; (4) Cu displacement by Pt; (5) washing, filtering, and drying; (6) catalyst

characterization.

Figure 2. A cross-sectional view of a fuel-cell assembly. GDL: gas diffusion layer; MPL:

microporous layer.

Figure 3. A typical HAADF-STEM image of Pt/Pd/C core-shell catalysts (a) and EELS line scan

profiles of Pt for a single Pt/Pd core-shell particle (inset) (b).

Figure 4. The XRD peak positions of (111) reflection of Pt/Pd/C in He and H2 atmospheres. The

Mo Kα (λ = 0.7107 Å) incident radiation was used in XRD measurements.

Figure 5. Comparison of (a) cyclic voltammetry curves and (b) oxygen reduction polarization

curves (positive-going) between Pt/Pd/C and Pt/C in N2- and O2-saturated 0.1 M HClO4

solutions, respectively. Scan rates = 50 and 10 mV s-1

in (a) and (b), respectively. The currents

were normalized to the geometric area of the rotating disk electrode (0.196 cm2). Pd loadings in

Pd/C and Pt/Pd/C were 10 and 7.7 µg cm-2

, respectively. The Pt loadings in Pt/Pd/C and Pt/C

were 4 and 24 µg cm-2

, respectively. (c) Comparison of Pt mass activity and specific activity

between Pt/Pd/C and Pt/C at 0.9 V. (d) Comparison of Pt mass activity at 0.9 V before and after

potential cycling between 0.65 and 1.0 V. The electrochemical areas were calculated from the

charges associated with HUPD assuming 210 µC cm-2

.

Figure 6. Comparisons of fuel-cell performance between Pt/Pd/C core-shell catalyst (18% Pt,

27% Pd) and regular Pt/C catalyst (20% Pt, TKK) under 100% inlet humidity conditions with

pure oxygen (solid lines) and air (dashed lines). Circle: Pt/Pd/C; triangle: Pt/C.

Figure 7. Voltage of core-shell cell minus voltage of Pt cell as a function of current density on

oxygen, air and helox.

Figure 8. Temperature dependence of Pt/Pd/C core-shell and regular Pt/C catalyst cells. Circle:

Pt/Pd/C; Triangle: Pt/C.

Figure 9. Humidity dependence of Pt/Pd/C core-shell and regular Pt/C catalyst cells. Circle:

Pt/Pd/C; triangle: Pt/C.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

0 1 2 3 4 5 6 7 8 9 10 11 12

No

rmal

ized

inte

rsit

y (a

.u.)

Position (nm)

Pta b

10 nm

5 nm

13.0 13.5 14.0 14.5 15.0 15.5 16.0

Inte

nsi

ty /

(a.

u.)

2q /degree

He

H2

Figure 5.

-1.2

-0.8

-0.4

0

0.4

0.8

0 0.2 0.4 0.6 0.8 1 1.2

j(m

A c

m-2

)

E (V vs RHE)

Pt/Pd/CPd/C

-7

-6

-5

-4

-3

-2

-1

0

0 0.2 0.4 0.6 0.8 1

j(m

A c

m-2

)

E (V vs RHE)

Pt/Pd/C

Pt/C

0

0.4

0.8

1.2

1.6

Pt/Pd/C Pt/C

Mas

s ac

tivi

ty (A

mg-1

)Sp

ecif

ic a

ctiv

ity

(mA

cm

-2)

Mass activity

Specific activity

0

0.4

0.8

1.2

1.6

Pt/Pd/C Pt/C

Mas

s ac

tivi

ty (A

mg-1

)

Initial

10000 cycles

a b

c d

Figure 6.

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Cell

Vo

lta

ge

(V

)

Current Density (A cm-2)

Regular Pt: Air

Pt/Pd Core-Shell: Air

Regular Pt: Oxygen

Pt/Pd Core-Shell: Oxygen

Figure 7.

0

10

20

30

40

50

60

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Vo

ltag

e d

iffe

ren

ce, V

cs-V

Pt (

mV

)

Current density (A cm-2)

Air

Oxygen

Helox

Figure 8.

Figure 9.

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

50 60 70 80 90

Ce

ll V

olta

ge

(V

); C

ell

"IR

" V

olta

ge

dro

p (

V)

Temperature (°C)

0.2 A cm-2

1.0 A cm-2

1.6 A cm-2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

-0.2 0 0.2 0.4 0.6 0.8 1 1.2

Cell

Vo

lta

ge

(V

)

Relative Humidity

0.2 A cm-2

1.0 A cm-2

1.6 A cm-2

Table 1 Cell operating conditions

Test

ID

Cell

temperature

(oC)

Gas exit

pressure

(kPa)

Cathode Anode

Reactant Utilization

(%)

Dew-Point

(oC)

Utilization

(%)

Dew-

Point

(oC)

1 80 140 Oxygen 50 80 67 80

2 80 140 Air 67 80 83 80

3 80 140 Helox* 67 80 83 80

4 80 140 Air 67 53 83 53

5 80 140 Air 67 Dry 60 53

6 85 140 Air 67 Dry 60 53

7 75 140 Air 67 Dry 60 53

8 70 140 Air 67 Dry 60 53

9 60 140 Air 67 Dry 60 53

*Helox: 21% oxygen, 79% Helium

Table 2: Comparisons of ECA and mass activities between Pt/C and Pt/Pd/C in liquid cells and MEAs

(Mass activities on MEA-Air were corrected for oxygen concentration).

Catalysts ECA (m2 g

-1) Mass Activity (A mg

-1)

Liquid MEA Liquid MEA-O2 MEA-Air

Pt/Pd/C 160 70 0.95 0.40 0.58

Pt/C 85 50 0.20 0.28 0.26