Vapor-deposited Pt and Pd-Pt catalysts for solid acid fuel cells: Short range structure and interactions with the CsH2PO4 electrolyte

Article abstract of DOI:10.1149/2.0371606jes

State-of-the-art cathodes for solid acid fuel cells (SAFCs) based on the crystalline electrolyte CsH₂PO₄ (CDP) are comprised of a proton-conducting CDP network coated by a vapor-deposited nanostructured catalyst. Pd-rich (85 at%Pd) Pt-Pd oxygen reduction catalysts vapor-deposited on CDP display both extraordinary activity for oxygen reduction and poor stability in cathodes for SAFCs operating at 250 °C. Similar catalysts with lower Pd content (57 at%Pd) are less active and more stable. Using X-ray absorption spectroscopy (XAS), we find that these catalysts are structurally similar and that structural variations are insufficient to explain the observed differences in activity. XAS and solid-state and solution nuclear magnetic resonance (NMR) also show that additional water-soluble chemical species are present in the Pd-rich electrode after fuel cell operation. We attribute the presence of these species to the reactivity of the Pd-rich catalyst with CsH₂PO₄ and suggest that these products are the cause of the observed deactivation.

Full text of DOI:10.1149/2.0371606jes

F464
Journal of The Electrochemical Society, 163 (6) F464-F469 (2016)
Vapor-Deposited Pt and Pd-Pt Catalysts for Solid Acid Fuel Cells:
Short Range Structure and Interactions with the CsH2PO4
Electrolyte
Alexander B. Papandrew,a,z Samuel St. John,a Ramez A. Elgammal,a David L. Wilson III,a
Robert W. Atkinson III,a Jamie S. Lawton,a Thomas M. Arruda,b
and Thomas A. Zawodzinski, Jr.a,c,∗
aDepartment of Chemical and Biomolecular Engineering, University of Tennessee, Knoxville, Tennessee 37996, USA
bDepartment of Chemistry, Salve Regina University, Newport, Rhode Island 02840, USA
cMaterials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA
State-of-the-art cathodes for solid acid fuel cells (SAFCs) based on the crystalline electrolyte CsH2PO4 (CDP) are comprised of a
proton-conducting CDP network coated by a vapor-deposited nanostructured catalyst. Pd-rich (85 at%Pd) Pt-Pd oxygen reduction
catalysts vapor-deposited on CDP display both extraordinary activity for oxygen reduction and poor stability in cathodes for SAFCs
operating at 250 ◦C. Similar catalysts with lower Pd content (57 at%Pd) are less active and more stable. Using X-ray absorption
spectroscopy (XAS), we find that these catalysts are structurally similar and that structural variations are insufficient to explain the
observed differences in activity. XAS and solid-state and solution nuclear magnetic resonance (NMR) also show that additional
water-soluble chemical species are present in the Pd-rich electrode after fuel cell operation. We attribute the presence of these species
to the reactivity of the Pd-rich catalyst with CsH2PO4 and suggest that these products are the cause of the observed deactivation.
© The Author(s) 2016. Published by ECS. This is an open access article distributed under the terms of the Creative Commons
Attribution Non-Commercial No Derivatives 4.0 License (CC BY-NC-ND, http://creativecommons.org/licenses/by-nc-nd/4.0/),
which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is not changed in any
way and is properly cited. For permission for commercial reuse, please email: oa@electrochem.org. [DOI: 10.1149/2.0371606jes]
All rights reserved.
Manuscript submitted September 30, 2015; revised manuscript received February 18, 2016. Published March 3, 2016.
Efficient, economical, reliable electricity generation from chemi-
their lifetime, probing catalyst structure as well as identifying inter-
actions with the CDP electrolyte.
cal fuels via electrochemical conversion has yet to be fully realized,
despite considerable efforts to inject fuel cells into the technological
mainstream. The slow adoption of these technologies can be traced
in part to the high cost and low durability of electrolyte membranes,
electrode materials, and other system components.1–4 Intermediate-
temperature fuel cells5 based on the crystalline proton conductor
CsH2PO4 (CDP)6 are no exception to these constraints, despite de-
sirable attributes that include fuel flexibility7 and the prospect of
platinum-free anodes.8,9
In the case of air-breathing cells based on CDP, which we hereafter
refer to as solid acid fuel cells (SAFCs), the major scientific challenges
concern the activity and durability of the cathode. As an entirely solid-
state technology, SAFC electrode performance is circumscribed by the
extent of the tri-phase contact of electrolyte, catalyst, and oxidant in
the electrode, a characteristic shared most closely with solid oxide
fuel cells.10 State-of-the-art SAFC cathodes rely on the activation of
all available cathode surface area by platinum coating, obtained via
vapor deposition.11 In this electrode architecture, the Pt catalyst serves
as both the oxygen reduction catalyst and the electronic conductor.
Despite the utility of this configuration, the mass-normalized activity
of Pt in SAFC cathodes remains unsatisfactory.12 Furthermore, the
nature of the Pt-CDP interface is not well understood, and early stud-
ies have revealed the presence of unusual phenomena in this case.13
Understanding catalyst-CDP interactions is key to optimizing the per-
formance of the SAFC cathode.
A fleeting breakthrough for the SAFC cathode resulted from the
deposition of Pt-Pd alloy catalysts on CDP.7,12 Pd-rich electrodes dis-
played activity enhancements of 450% compared with baseline Pt
electrodes, for reasons that remain poorly understood. Furthermore,
Pd-rich (>70 at% Pd) compositions are particularly unstable, deac-
tivating to levels below the Pt baseline in approximately 12 hours of
operation.12 These results motivated a deeper inquiry into the origins
of the enhanced activity and reduced durability of vapor-deposited
Pd-Pt@CDP electrode materials in SAFCs. In this contribution, we
investigate Pt@CDP and Pd-Pt@CDP electrodes at various stages of
Experimental
Electrode materials consisting of Pt@CDP and Pd-Pt@CDP were
prepared as described previously.11,12 Briefly, fine CDP powder (BET
surface area approximately 2.4 m2/g) was combined with solid
Pt(acac)2 or a mixture of Pt(acac)2 and Pd(acac)2 (ACROS Organ-
ics) in the appropriate ratio in a fixed bed reactor and heated to 210 ◦C
in a N2-water vapor atmosphere. Experimental samples of 85 at% Pd
and 57 at% Pd, as determined by EDS (hereafter Pd85Pt15 and Pd57Pt43,
respectively) and a control sample of 100 at% Pt were synthesized.
All samples consisted of 20wt% metal on CDP.
SAFC cathodes were fabricated by spreading the powders over
an anode-supported (1.0 mgPt /cm2) assembly comprising current col-
lector, anode, and a dense CDP membrane, followed by lamination
at 24 MPa, forming a membrane-electrode assembly (MEA) similar
to those reported previously.14 Fuel cell testing proceeded at 250 ◦C
using anode and cathode gas feeds of 30 sccm hydrogen and 75 sccm
◦
air, respectively, each humidified to a 75 C dew point. Polarization
curves were recorded at 30 minute intervals over 8 hours by sweeping
the cell potential from the open circuit voltage (OCV) to 0.4 V at
2 mV/s. Curves free of the membrane ohmic resistance were derived
by correction using the high-frequency intercept from electrochemi-
cal impedance spectroscopy, typically found at 10-20 kHz in spectra
recorded in the range of 200 kHz to 200 mHz. Between polarization
curves, and after the first 8 hours of testing, cells were potentiostati-
cally held at 0.6 V. All electrochemical tests were performed with a
Bio-Logic VSP potentiostat.
We further characterized the as-synthesized materials as well as
samples recovered from SAFC cathodes after testing. A fraction of
each of these sets was also washed with copious deionized water
to dissolve the water-soluble CDP electrolyte and any other soluble
components. In the case of XAS and XRD measurements, this water
was exchanged several times to remove any traces of the dissolved
compounds in the dried material. Samples recovered without washing
were also evaluated for comparison.
X-ray absorption spectra data were acquired at beamline X11A
of the National Synchrotron Light Source (NSLS) at Brookhaven
∗
Electrochemical Society Fellow.
zE-mail: apapandrew@utk.edu
Downloaded on 2016-06-17 to IP 109.163.234.5 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, 163 (6) F464-F469 (2016)
F465
1.0
0.9
0.8
0.7
0.6
0.5
(2)
a
b
400
300
200
100
(1)
Pd85Pt15
Pd85Pt15 peak
Figure 1. (a) iR-free polarization curves and (b)
chronoamperometric data acquired at 0.6 V cell volt-
age for 20%metal@CDP SAFC cathodes with varying
compositions. Polarization curves denoted (1), (2), and
(3) were recorded at times corresponding to the noted
locations in panel (b). All tests were performed at 250 ◦C
with 30 sccm hydrogen and 75 sccm air humidified to a
75 ◦C dew point.
Pd57Pt43
(2)
Pd85 peak
Pt
(1)
Pd85 initial
Pt
Pd57
(3)
Pd85
(3)
20
5
10
15
0
200
400
600
800
Current Density [mA cm-2]
Elapsed Time [h]
National Laboratory. Spectra were obtained in fluorescence or trans-
mission mode at the Pt L3 and Pd K edges, using a Si(111) double
crystal monochromator with a 10% detuning and gas ionization cham-
bers for beam intensity monitoring. Simultaneous edge fitting was
performed using the iXAFS software. A complete table of EXAFS
fits can be found in Table II.
Solid state NMR were recorded on a Varian Innova (9.4 T) with a
31P spectral frequency of 162.3 MHz at 25 ◦C. The samples were spun
at a MAS spin rate of 6.1 kHz in a 5.0 mm zirconia rotor (Agilent
Technologies). Liquid state NMR experiments were performed on a
Varian VNMRS (11.7 T) instrument with a 31P spectral frequency of
202.9 MHz at 25 ◦C. Chemical shifts are reported with respect to an
external standard containing 85% H3PO4.
this topic, we note the shift of fcc(111) peak of the as-synthesized
Pd-containing samples to lower 2θ values. This shift is due to the
presence of interstitial carbon in Pd-rich vapor-deposited catalysts,12
which we have observed in the SAFC system and in other applica-
tions of this deposition technique.15,16 At peak activity, the Pd85Pt15
sample displays a lattice parameter (0.3898 nm) consistent with a
Pt-Pd solid solution of the nominal composition, indicating that inter-
stitial carbon has been removed, in agreement with our earlier report.
Analysis of the fcc(111) peak width for the Pd85Pt15 peak-activity sam-
ple yields a Scherrer coherence length of 11.8 nm, intermediate be-
tween the as-synthesized state (4.7 nm) and the final, deactivated state
(16.5 nm).
Since X-ray diffraction is sensitive primarily to long-range order,
we used extended X-ray absorption fine structure (EXAFS) spec-
troscopy to study the experimental materials. EXAFS is able to reveal
short-range structural motifs responsible for catalyst activity enhance-
ments, such as core-shell structures with surface reactivity modified by
strain and ligand effects.17–19 This has been shown in multiple exam-
ples of ORR catalysts for low-temperature fuel cell systems, where Pd
has been used extensively as a promoter of Pt activity for the oxygen
reduction reaction,20–25 often in a Pd-core/Pt-shell morphology.
Based on particle sizes derived from the XRD data, we stipulate
that the catalyst particles are large enough to be nearly fully coordi-
nated (N = 12 for fcc materials) .26 Consequently, we have restrained
the sum of the coordination numbers for the three types of scatter-
ing paths (Pt-Pt, Pd-Pd, and Pt-Pd) in each catalyst to approximately
Results and Discussion
Polarization curves corrected for membrane ohmic resistance (“iR-
free”) and cell current density at 0.6 V for the fuel cells tested are
shown in Figure 1, and summarized in Table I. In the case of the
Pd85Pt15 composition, we prepared electrode materials in different
states of activity by operating cells for varying amounts of time.
The cell operated until it reached peak activity is denoted as such
hereafter.
We note in the case of the Pd85Pt15 electrodes a large enhance-
ment in current density with respect to the Pt control, and the rapid
decline in this enhancement over the course of the test. Also evident
is a relatively brief “break-in” period experienced by the Pd85Pt15
electrodes at the start of the test, over which the current density in-
creases by approximately 15% over 30 minutes at 0.6 V. Samples
with the Pd57Pt43 composition show very little initial current increase,
much lower peak current density, and a much lower degradation rate
than the Pd85Pt15 sample (Figure 1b). The origins of activity enhance-
ment, the break-in phenomenon, and of catalyst instability are thus
composition-dependent and restricted to the case of a Pd-rich catalyst
composition.
(a) Pre-FC
(b) Post-FC
Pd85Pt15
Pd57Pt43
Pt
X-ray probes were applied to the cathode materials to detect struc-
tures capable of explaining the observed phenomena. X-ray diffrac-
tion patterns of washed samples free of the intense CDP diffraction
peaks are shown in Figure 2. In agreement with our earlier work on
Table I. Initial (ji ), peak (jp), and final (j f ) current density (in
mA cm−2) at 0.6 V for the electrodes investigated. Acquired at
0.6 V cell voltage at 250 ◦C with 30 sccm hydrogen and 75 sccm air
humidified to a 75 ◦C dew point.
20 30 40 50 60 70
20 30 40 50 60 70
Sample
ji
jp
j f
2
[deg.]
2
[deg.]
Pd57Pt43
Pd85Pt15
Pd85Pt15 peak
Pt
192
358
402
179
201
447
457
208
160
129
457
208
Figure 2. Cu Kα X-ray diffraction patterns for vapor-deposited catalysts af-
ter washing with H2O (a) before and (b) after testing in SAFC cathodes at
250 ◦C. Vertical lines indicate the position of the Pt(111) peak.
Downloaded on 2016-06-17 to IP 109.163.234.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

F466
Journal of The Electrochemical Society, 163 (6) F464-F469 (2016)
Table II. EXAFS parameters derived from fitting for experimental samples.
Sample
Path*
Edge
R (Å)
σ2 (Å2)
ꢀE0 (eV)
N
Pt foil
Pd foil
Pt-Pt
Pd-Pd
L3
K
2.760 0.004
2.735 0.003
0.0054 0.0002
0.0065 0.0002
5.01 0.73
1.74 0.55
12
12
Pt pre-FC, washed
Pt post-FC, washed
Pt-Pt
Pt-Pt
L3
L3
2.754 0.012
2.762 0.004
0.0096 0.0007
0.0063 0.0005
7.51 1.17
6.77 0.48
12
12
Pd57 pre-FC
Pt-Pt
L3
2.739 0.011
2.740 0.022
2.740 0.022
2.780 0.004
0.0078 0.0005
0.0078 0.0005
0.0078 0.0005
0.0078 0.0005
1.34 0.41
1.34 0.41
1.34 0.41
1.34 0.41
4.21 0.65
1.02 0.25
1.02 0.25
6.58 0.45
Pt-Pd
Pd-Pt
Pd-Pd
L3/K
L3/K
K
Pd57 post-FC
Pd-O
Pd-Pt
Pd-Pd
K
K
K
-
0.0064 0.0008
0.0064 0.0008
0.0064 0.0008
1.47 0.71
1.47 0.71
1.47 0.71
-
2.733 0.013
2.736 0.007
3.72 0.39
6.32 0.66
Pd57 post-FC, washed
Pt-Pt
L3
2.742 0.009
2.730 0.013
2.730 0.013
2.750 0.008
0.0066 0.0006
0.0066 0.0006
0.0066 0.0006
0.0066 0.0006
3.25 1.03
3.25 1.03
3.25 1.03
3.25 1.03
3.80 0.35
1.75 0.20
1.75 0.20
6.24 0.65
Pt-Pd
Pd-Pt
Pd-Pd
L3/K
L3/K
K
Pd85 pre-FC
Pt-Pt
L3
2.736 0.006
2.733 0.007
2.733 0.007
2.791 0.006
0.0074 0.0006
0.0074 0.0006
0.0074 0.0006
0.0074 0.0006
1.83 0.70
1.83 0.70
1.83 0.70
1.83 0.70
4.03 0.78
0.74 0.39
0.74 0.39
7.03 0.76
Pt-Pd
Pd-Pt
Pd-Pd
L3/K
L3/K
K
Pd85 post-FC
Pd-O
Pd-Pt
Pd-Pd
K
K
K
1.956 0.067
2.736 0.006
2.734 0.019
0.0077 0.0005
0.0077 0.0005
0.0077 0.0005
1.85 1.73
1.85 1.73
1.85 1.73
1.41 0.96
2.50 1.10
7.29 2.90
Pd85 post-FC, washed
Pt-Pt
L3
2.735 0.007
2.730 0.007
2.730 0.007
2.740 0.005
0.0050 0.0005
0.0050 0.0005
0.0050 0.0005
0.0050 0.0005
3.57 0.90
3.57 0.90
3.57 0.90
3.57 0.90
2.79 0.20
2.01 0.14
2.01 0.14
7.00 0.51
Pt-Pd
Pd-Pt
Pd-Pd
L3/K
L3/K
K
Pd85 peak-FC
Pd-O
Pd-Pt
Pd-Pd
K
K
K
1.956 0.049
-
2.738 0.011
0.0063 0.0007
0.0063 0.0007
0.0063 0.0007
3.46 1.94
3.46 1.94
3.46 1.94
0.85 0.42
-
6.00 0.14
Pd85 peak-FC, washed
Pt-Pt
L3
2.742 0.014
2.730 0.017
2.730 0.017
2.740 0.007
0.0054 0.0008
0.0054 0.0008
0.0054 0.0008
0.0054 0.0008
5.51 1.15
5.51 1.15
5.51 1.15
5.51 1.15
3.69 0.57
1.93 0.35
1.93 0.35
6.17 0.80
Pt-Pd
Pd-Pt
Pd-Pd
L3/K
L3/K
K
12 while the Debye-Waller factor was used as a fitting parameter. A
complete summary of all EXAFS fitting parameters obtained using
this approach is presented in Table II.
Radial distributions for the H2O-washed Pt sample (Figure 3) show
that the vapor-deposited catalysts agree well with the Pt foil standard.
As-deposited Pt shows slight lattice contraction and a marginally en-
hanced Debye-Waller factor that is likely related to the presence of Pt
oxides. The radial distribution of the post-FC material recovered from
the SAFC cathode shows increased intensity resulting from particle
coalescence and a lattice parameter nearly identical to that of the foil
standard.
For both the as-synthesized Pd85Pt15 and Pd57Pt43 samples, radial
distributions obtained from Fourier transformation of the Pd K edge
EXAFS show excellent sensitivity to the bulk lattice dilatation induced
by lattice carbon (Figure 4). First-shell fits to these data return an initial
Pd-Pd spacing of 2.791 0.006 Å for Pd85Pt15 and 2.780 0.004
Å for Pd57Pt43, significantly larger than the Pd-Pd nearest neighbor
path of the Pd foil standard (2.735 0.003 Å). After operation these
Pd-Pd paths contract to the Pd-Pd spacing of elemental Pd.
2.5
Pt foil
2.0
Pt post-FC
Pd K
Pt L3
Pd85
Pd57
a
b
20
15
10
5
pre-FC
post-FC
pre-FC
post-FC
0.8
0.6
0.4
0.2
0.0
1.5
1.0
Pd85
Pd57
Pt pre-FC
0.5
0
0.0
1.0
2.0
R [Å]
3.0
1.0
2.0
R [Å]
3.0
1.0
2.0
3.0
R [Å]
Figure 4. Forward Fourier transforms of the (a) Pd K EXAFS and (b) Pt
L3 EXAFS for Pd57Pt43 and Pd85Pt15 before (thin line) and after (bold line)
testing in SAFC cathodes at 250 ◦C. Pt L3 data were obtained from samples
washed with H2O to remove the CDP electrolyte and other soluble species.
Figure 3. Forward Fourier transform of the Pt L3 EXAFS for the Pt control
sample before and after testing in SAFC cathodes at 250 ◦C. Pt L3 data were
obtained from samples washed with H2O to remove the CDP electrolyte.
Downloaded on 2016-06-17 to IP 109.163.234.5 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, 163 (6) F464-F469 (2016)
F467
18
16
14
12
10
8
Pd K
Pd K
Pt L3
b
Pd K
Pd85 peak-FC
a
a
b
20
15
10
5
12
10
8
0.6
0.5
0.4
0.3
0.2
0.1
0.0
washed
Pd85 post-FC
6
Pd85 post-FC
Pd85 peak-FC
Pd85 post-FC
Pd57 post-FC
washed
6
4
4
2
2
0
0
0
1.0
2.0
R [Å]
3.0
1.0
2.0
R [Å]
3.0
1.0
2.0
R [Å]
3.0
1.0
2.0
R [Å]
3.0
Figure 5. Forward Fourier transforms of the (a) Pd K EXAFS and (b) Pt L3
EXAFS for Pd85Pt15 sampled at peak activity (1.5 hrs) and after deactivation
(20 hrs) after variable testing duration in SAFC cathodes at 250 ◦C. Data were
obtained from samples washed with H2O to remove the CDP electrolyte and
other soluble species.
Figure 6. Forward Fourier transforms of the (a) Pd K EXAFS and (b) Pt L3
EXAFS for Pd85Pt15 sampled at peak activity (1.5 hrs) and after deactivation
(20 hrs) after variable testing duration in SAFC cathodes at 250 ◦C. Each panel
compares the as-recovered (post-FC) sample to its H2O-washed counterpart.
with electrolytes. For example, the performance of liquid phosphoric
acid fuel cell cathode is dominated by phosphate anion adsorption on
platinum.27 To probe the possibility of an interaction with the CDP
electrolyte, we compared the Pd K edge EXAFS radial distributions of
H2O-washed samples to unwashed samples. These results are shown
in Figure 6. In both the case of Pd85Pt15 and Pd57Pt43, we observe
increased intensity in the unwashed samples at low R values in com-
parison with the washed samples. We also note in the case of the
washed Pd85Pt15 an increase in the Pd-Pd scattering intensity rela-
tive to the unwashed sample. By including Pd-O scattering paths we
were able to fit these low-R features with Pd-O bond lengths from
1.90–1.94 Å.
The◦presence of Pd-O species in cells operated in humidified air
at 250 C was not surprising, but the lowered intensity of the Pd-O
species after washing implies solubility of these species in water, rul-
ing out simple Pd oxides. Examination of the supernatant solution
resulting from washing of Pd85Pt15 revealed a translucent yellow so-
lution, shown in Figure 7. Chemical analysis of this solution using
ICP-AES confirmed the presence of Pd.
We used NMR to further analyze the chemical environments
present in the experimental electrodes. Figure 8 shows the 31P MAS
NMR spectra of the Pd85Pt15 electrode and the 100% Pt-CDP control
prior to and after 24 hours of fuel cell testing, in comparison with a
bulk CDP standard. Spectra of the standard and as-synthesized mate-
rials are in good agreement with other published measurements.28–30
In the as-synthesized samples, we find that neither the conditions of
the vapor-deposition process nor the presence of the catalyst nanopar-
ticles significantly affect the linewidth of the as-synthesized electrode
No such evolution in atomic spacings is present in the case of Pt-Pt
paths derived from the Pt L3 edge EXAFS, shown in Figure 4. We find
Pt-Pt spacings of 2.736 0.006 Å and 2.739 0.011 Å for Pd85Pt15
and Pd57Pt43, respectively, and negligible shifts of these spacings in
the post-test state. These values are slightly contracted with respect
to the Pt-Pt spacing of 2.762 0.004 Å for the post-test Pt control
sample and 2.760 0.004 Å for the Pt foil standard.
First-shell coordination numbers derived from the fits indicate
atomic clustering in the as-synthesized samples rather than random
alloy formation. For a random alloy with the nominal composition
Pd85Pt15, a Pt-Pt coordination number NPt−Pt of ∼0.3 is expected, but
here we find NPt−Pt = 4.03 0.78. The heterometallic coordination
number NPt−Pd is also lower than expected for random ordering. We
find NPt−Pd = 0.76 0.39 for Pd85Pt15, versus the expected value for
random ordering of ∼3.0. In the case of Pd57Pt43 we similarly note
enhanced Pt-Pt coordination and fewer Pt-Pd paths than expected from
a random subsitutional alloy.
In room temperature acidic systems, Pt clusters under compressive
strain17,19,20,24 exhibit increased activity for the ORR by modification
of the chemisorption of oxygenated species, and it is possible that
such an effect is a possible explanation for activity enhancement on
Pd85Pt15. However, extant data opposes this hypothesis. First, in room
temperature aqueous systems, ORR activity enhancement of Pt due to
alloying is thought to be due to reduced surface coverage of oxygenate
spectators, freeing active sites for dioxygen binding. However we pre-
viously noted that Pd-rich catalysts for SAFCs are more oxophilic than
pure Pt catalysts.12 Furthermore, there is very close correspondence of
Pt-Pt spacings and coordination in the highly active Pd-rich Pd85Pt15
sample versus the unremarkable Pd57Pt43 sample, and additionally
the deactivated Pd85Pt15 sample displays a similarly compressed Pt-Pt
distance.
We find very little difference in the EXAFS first-shell radial dis-
tributions from the Pd K edge and Pt L3 edge for Pd85Pt15 in the
peak-activity (peak-FC) state versus the deactivated (post-FC) state.
These data are shown in Figure 5, and detailed in Table II. Qualita-
tively, the modulations of the radial distribution are superimposable,
with small differences in intensity present. We find nearly identi-
cal interatomic spacings for Pt-Pt, Pt-Pd, and Pd-Pd paths in both
samples, each approximately 2.73 Å, in close agreement with the
Pd-Pd path for the Pd foil standard. Only slight differences in coor-
dination are present, with fractionally more Pt-Pt coordination in the
peak-FC sample. Based on these results, we find it unlikely that a
fortuitous structural arrangement analogous to a core-shell or other
highly segregated architecture is alone responsible for SAFC activ-
ity enhancements. Similarly, we find no evidence for the destruc-
tion of a “magic” metallic nanostructure as a root cause of catalyst
deactivation.
Figure 7. Solution resulting from washing and sonication of the Pd85Pt15
electrode followed by centrifugation. The solid components of the electrode
are visible at the bottom of the tube.
Of course, an understanding of real fuel cell systems is inher-
ently incomplete without accounting for the interactions of catalysts
Downloaded on 2016-06-17 to IP 109.163.234.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

F468
Journal of The Electrochemical Society, 163 (6) F464-F469 (2016)
c
Pt post-FC
Pt pre-FC
*
b
*
*
*
b
Pd85 post-FC
Pd85Pt15-CDP post FC
Pd85 pre-FC
bulk CDP
*
a
a
20
10
0
-10
-20
31P chemical shift [ppm]
CsPO3
Figure 8. Proton-decoupled 31P solid state MAS NMR spectra of (a) bulk
CDP, (b) Pd85Pt15-CDP pre- and post-fuel cell testing, and (c) Pt-CDP pre-
and post-fuel cell testing.
40
20
0
-20
-40
31P chemical shift / ppm
Figure 9. 31P solid state MAS NMR spectra of (a) bulk CsPO3 and (b)
Pd85Pt15-CDP post fuel cell testing. Spinning sidebands are indicated by as-
terisks.
materials compared to pristine CDP. The lack of a pronounced Knight
shift and a small downfield shift of ∼0.2 ppm observed in the pres-
ence of the metal nanoparticles indicates that the metallic catalysts are
diamagnetic and only minimally change the magnetic susceptibility
of the composite.31
other phases to be poor proton conductors, passivating the catalyst-
electrolyte interface.
An additional conjecture that is self-consistent with all of the ex-
tant data is based on the well-known high solubility and diffusivity of
hydrogen in Pd. We hypothesize that Pd-rich catalyst particles may
act as a hydrogen shuttle, allowing transport of protons out of the
CDP bulk into the Pd lattice. The adsorbed hydrogen can then desorb
at the catalyst surface to react with oxygen in the familiar acid fuel
cell cathode half-reaction, producing water. As we demonstrated pre-
viously, Pd is a facile hydrogen oxidation/proton reduction catalyst in
the SAFC environment.8
This hypothesis also explains the lack of activity observed in more
Pt-rich compositions, since the solubility of H in Pt-Pd alloys is negli-
gible above 30 at% Pt,33 and we similarly find limited activity enhance-
ment above 30 at% Pt.12 The break-in phenomenon can be explained
by the desorption of carbon from the octahedral interstitial sites in
Pd, which are shared by H and C,34,35 and necessary for H transport.
Deactivation of this scheme can be explained by chemical reactivity
of CDP with Pd or Pd oxides to form a passivating layer comprised
of CsPO3 and Pd phosphates and pyrophosphates with low proton
conductivity that prevents proton traversal into the Pd bulk. It may be
possible, then, to retain the high activity of the Pd-rich catalysts by
protecting the CDP surface with a contiguous thin film of Pt or other
non-reactive species, or by using explicitly designed core-shell Pt@Pd
nanoparticles that prevent direct contact of CDP with Pd. We suggest
that a future in situ XAS experiment may be capable of verifying these
conjectures.
In the case of post-FC Pd85Pt15@CDP (Figure 8b) we observe the
appearance of three distinct new peaks. We assign the peak centered
at 10.78 ppm as a spinning side band of CsPO3 (see Figure 9). We
observe neither the signature of the partially dehydrated Cs2H2P2O7,
nor of species with differing degrees of condensation of the phos-
phate groups. Instead, there are two downfield resonances at 3.36 ppm
and 6.92 ppm. We tentatively assigned these peaks to Pd-containing
decomposition products, based on the downfield shift observed for
other palladium phosphates.32 Table III summarizes the locations and
widths of all observed peaks derived from fitting the peaks with a
single Lorentzian.
Given the detection of soluble Pd moieties in the supernantant
solution resulting from washing of the Pd85Pt15 electrode, we also
conducted a solution state 31P NMR experiment. Those results are
illustrated in Figure 10. Again we find two distinct species in the
case of the Pd85Pt15@CDP sample, with chemical shifts at 17.00 and
−10.65 ppm. In the case of post-FC Pt@CDP, we observe a spectrum
identical to aqueous bulk CDP.
As we declared earlier in this contribution, we have sought an
explanation for three key phenomena in the Pd85Pt15@CDP system.
To reiterate, these are the anomalously high activity of the electrode,
the substantial initial increase in its performance, and the subsequent
rapid loss of the excess activity. Based on the XRD and XAS data, we
reject the hypothesis of exquisitely nanostructured catalyst particles
that provide activity enhancement due to a promotional effect of Pd
on Pt. We do find from XAS, however, the presence of a soluble Pd-
containing product in deactivated catalysts. NMR confirms that this
is accompanied by the formation of multiple other chemical species,
including CsPO3. The apparent chemical reactivity of the Pd-rich
material explains the deactivation of the catalyst, since we expect these
Conclusions
Using X-ray and nuclear magnetic resonance methods, we char-
acterized the structural and chemical evolution of vapor-deposited
Table III. 31P NMR parameters derived from Lorentzian fits to solid-state MAS spectra.
Peak 1 Peak 2
Peak 3
Sample
δ31 [ppm]
ꢀν1/2 [Hz]
δ31 [ppm]
ꢀν1/2 [Hz]
δ31 [ppm]
ꢀν1/2 [Hz]
P
P
P
CDP
−5.65
−5.32
−5.21
−5.46
−5.67
376
490
503
371
394
Pt-CDP pre-FC
Pt-CDP post-FC
Pd85Pt15-CDP pre-FC
Pd85Pt15-CDP post-FC
3.36
84
6.92
604
Downloaded on 2016-06-17 to IP 109.163.234.5 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, 163 (6) F464-F469 (2016)
F469
3. D. J. L. Brett, A. Atkinson, N. P. Brandon, and S. J. Skinner, Chem. Soc. Rev., 37,
1568 (2008).
4. R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan, N. Garland, D. Myers,
M. Wilson, F. Garzon, D. Wood, P. Zelenay, K. More, K. Stroh, T. Zawodzinski,
J. Boncella, J. E. McGrath, M. Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi,
S. Miyata, A. Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K.-I. Kimijima, and
N. Iwashita, Chem. Rev., 107, 3904 (2007).
5. D. Boysen, T. Uda, C. Chisholm, and S. Haile, Science (80-. ), 303, 68 (2004).
6. S. Haile, C. Chisholm, K. Sasaki, D. Boysen, and T. Uda, Faraday Discuss., 134, 17
(2007).
c
b
7. C. R. I. Chisholm, D. A. Boysen, A. B. Papandrew, S. K. Zecevic, S. Cha,
K. A. Sasaki, A. Varga, K. P. Giapis, and S. Haile, Electrochem. Soc. Interface,
18, 53 (2009).
8. A. B. Papandrew, D. L. Wilson, N. M. Cantillo, S. Hawks, R. W. Atkinson,
G. A. Goenaga, and T. A. Zawodzinski, J. Electrochem. Soc., 161, F679 (2014).
9. A. B. Papandrew, R. W. Atkinson III, R. R. Unocic, and T. A. Zawodzinski, J. Mater.
Chem. A, 3, 3984 (2015).
-10.0 -10.4 -10.8
17.2 16.8
Pd85 post-FC
supernatant
Pt post-FC supernatant
aqueous CDP
a
20
10
0
-10
-20
10. S. B. Adler, Chem. Rev., 104, 4791 (2004).
11. A. B. Papandrew, C. R. Chisholm, R. A. Elgammal, M. M. zer, and S. K. Zecevic,
31P chemical shift [ppm]
Chem. Mater., 23, 1659 (2011).
12. A. B. Papandrew, C. R. I. Chisholm, S. K. Zecevic, G. M. Veith, and
T. A. Zawodzinski, J. Electrochem. Soc., 160, F175 (2013).
13. M. W. Louie and S. M. Haile, Energy Environ. Sci., 4, 4230 (2011).
14. T. Uda and S. Haile, Electrochem. Solid-State Lett., 8, A245 (2005).
15. R. W. Atkinson, S. St. John, O. Dyck, K. A. Unocic, R. R. Unocic, C. S. Burke,
J. W. Cisco, C. A. Rice, T. A. Zawodzinski, and A. B. Papandrew, ACS Catal., 5,
5154 (2015).
Figure 10. (a) Solution state 31P NMR in water of aqueous CDP and the
aqueous abstractions of the Pd85Pt15 post-FC fuel material and post-FC Pt
-CDP Inset (b) and (c) show enlarged views of the additional peaks observed
in the case of the Pd85Pt15 material.
16. S. St. John, R. W. Atkinson, R. R. Unocic, T. A. Zawodzinski, and A. B. Papandrew,
Pt-Pd oxygen reduction catalysts supported on the proton conduc-
tor CsH2PO4. We find little structural difference between the highly
active Pd85Pt15 catalyst and the unremarkable Pd57Pt43 composition.
Both catalysts show atomic-scale clustering of Pt and Pd and com-
pression of Pt-Pt interatomic spacings consistent with an alloy hosted
on the Pd lattice. Both EXAFS and NMR measurements indicate the
formation of additional water-soluble chemical species resulting from
the reaction of the Pd-rich catalysts with the CsH2PO4 electrolyte. We
conclude that this reactivity is the source of the instability observed
in the highly active, Pd-rich composition.
J. Phys. Chem. C, 119, 13481 (2015).
17. P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C. Yu, Z. Liu, S. Kaya,
D. Nordlund, H. Ogasawara, M. F. Toney, and A. Nilsson, Nat. Chem., 2, 454 (2010).
18. V. R. Stamenkovic, B. S. Mun, M. Arenz, K. J. J. Mayrhofer, C. A. Lucas, G. Wang,
P. N. Ross, and N. M. Markovic, Nat. Mater., 6, 241 (2007).
19. P. S. Mehtap Oezaslan and Frdric Hasch, J. Phys. Chem. Lett., 4, 3273 (2013).
20. J. Zhang, Y. Mo, M. B. Vukmirovic, R. Klie, K. Sasaki, and R. R. Adzic, J. Phys.
Chem. B, 108, 10955 (2004).
21. Z. Peng and H. Yang, J. Am. Chem. Soc., 131, 7542 (2009).
22. B. Lim, M. Jiang, P. Camargo, E. Cho, J. Tao, X. Lu, Y. Zhu, and Y. Xia, Science
(80-. ), 324, 1302 (2009).
23. K. Sasaki, H. Naohara, Y. Cai, Y. M. Choi, P. Liu, M. B. Vukmirovic, J. X. Wang,
and R. R. Adzic, Angew. Chemie - Int. Ed., 49, 8602 (2010).
24. C. Koenigsmann, A. C. Santulli, K. Gong, M. B. Vukmirovic, W.-P. Zhou, E. Sutter,
S. S. Wong, and R. R. Adzic, J. Am. Chem. Soc., 133, 9783 (2011).
25. Y. Kim, Y. Noh, E. J. Lim, S. Lee, S. M. Choi, and W. B. Kim, J. Mater. Chem. A, 2,
6976 (2014).
Acknowledgments
Financial support for this work was provided by the National Sci-
ence Foundation through TN-SCORE (NSF EPS-1004083). Use of
the National Synchrotron Light Source, Brookhaven National Lab-
oratory, 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 also thank Carlos Steren for his
assistance with NMR measurements.
26. A. Jentys, Phys. Chem. Chem. Phys., 1, 4059 (1999).
27. D. Strmcnik, M. Escudero-Escribano, K. Kodama, V. R. Stamenkovic, A. Cuesta,
and N. M. Markovic´, Nat. Chem., 2, 880 (2010).
28. K. Yamada, T. Sagara, Y. Yamane, H. Ohki, and T. Okuda, Solid State Ionics, 175,
557 (2004).
29. G. Kim, F. Blanc, Y.-Y. Hu, and C. P. Grey, J. Phys. Chem. C, 117, 6504 (2013).
30. A. Gon˜i Urtiaga, K. Scott, S. Cavaliere, D. J. Jones, and J. Rozie´re, J. Mater. Chem.
A, 1, 10875 (2013).
31. L. E. Marbella and J. E. Millstone, Chem. Mater., 27(8), 2721 (2015).
32. K. Panagiotidis, W. Hoffbauer, and R. Glaum, Zeitschrift fu¨r Anorg. und Allg. Chemie,
635, 1969 (2009).
33. A. G. Knapton, Platin. Met. Rev., 21, 44 (1977).
34. J. A. McCaulley, J. Phys. Chem., 97, 10372 (1993).
References
35. M. W. Tew, M. Nachtegaal, M. Janousch, T. Huthwelker, and J. A. van Bokhoven,
1. M. K. Debe, Nature, 486, 43 (2012).
2. A. J. Jacobson, Chem. Mater., 22, 660 (2010).
Phys. Chem. Chem. Phys., 14, 5761 (2012).
Downloaded on 2016-06-17 to IP 109.163.234.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).