Systematic Structure-Property Relationship Studies in Palladium-Catalyzed Methane Complete Combustion

Article abstract of DOI:10.1021/acscatal.7b02414

To limit further rising levels in methane emissions from stationary and mobile sources and to enable promising technologies based on methane, the development of efficient combustion catalysts that completely oxidize CH₄ to CO₂ and H₂O at low temperatures in the presence of high steam concentrations is required. Palladium is widely considered as one of the most promising materials for this reaction, and a better understanding of the factors affecting its activity and stability is crucial to design even more improved catalysts that efficiently utilize this precious metal. Here we report a study of the effect of three important variables (particle size, support, and reaction conditions including water) on the activity of supported Pd catalysts. We use uniform palladium nanocrystals as catalyst precursors to prepare a library of well-defined catalysts to systematically describe structure-property relationships with help from theory and in situ X-ray absorption spectroscopy. With this approach, we confirm that PdO is the most active phase and that small differences in reaction rates as a function of size are likely due to variations in the surface crystal structure. We further demonstrate that the support exerts a limited influence on the PdO activity, with inert (SiO₂), acidic (Al₂O₃), and redox-active (Ce_0.8Zr_0.2O₂) supports providing similar rates, while basic (MgO) supports show remarkably lower activity. Finally, we show that the introduction of steam leads to a considerable decrease in rates that is due to coverage effects, rather than structural and/or phase changes. Altogether, the data suggest that to further increase the activity and stability of Pd-based catalysts for methane combustion, increasing the surface area of supported PdO phases while avoiding strong adsorption of water on the catalytic surfaces is required. This study clarifies contrasting reports in the literature about the active phase and stability of Pd-based materials for methane combustion.

Full text of DOI:10.1021/acscatal.7b02414

Subscriber access provided by LAURENTIAN UNIV
Article
Systematic Structure-Property Relationship Studies in
Palladium-Catalyzed Methane Complete Combustion
Joshua J. Willis, Alessandro Gallo, Dimosthenis Sokaras, Hassan Aljama,
Stanis#aw Henrik Nowak, Emmett D. Goodman, Liheng Wu, Christopher J.
Tassone, Thomas F. Jaramillo, Frank Abild-Pedersen, and Matteo Cargnello
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b02414 • Publication Date (Web): 09 Oct 2017
Downloaded from http://pubs.acs.org on October 9, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 14
ACS Catalysis
1
2
3
4
5
6
7
8
9
Systematic Structure-Property Relationship Studies in Palla-
dium-Catalyzed Methane Complete Combustion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Joshua J. Willis†, Alessandro Gallo†,‡, Dimosthenis Sokaras‡, Hassan Aljama†, Stanislaw H.
Nowak‡, Emmett D. Goodman†, Liheng Wu†,‡, Christopher J. Tassone‡, Thomas F. Jaramillo†,
Frank Abild-Pedersen†, Matteo Cargnello*,†
†Department of Chemical Engineering and SUNCAT Center for Interface Science and Catalysis, Stanford University,
Stanford, CA 94305, USA.
‡Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, Menlo Park, CA 94025, USA.
ABSTRACT: To limit further rising levels in methane emissions from stationary and mobile sources and to enable promis-
ing technologies based on methane, development of efficient combustion catalysts that completely oxidize CH4 to CO2
and H2O at low temperatures in the presence of high steam concentrations is required. Palladium is widely considered as
one of the most promising materials for this reaction, and a better understanding of the factors affecting its activity and
stability is crucial to design even more improved catalysts that efficiently utilize this precious metal. Here we report a
study of the effect of three important variables (particle size, support, and reaction conditions including water) on the
activity of supported Pd catalysts. We use uniform palladium nanocrystals as catalyst precursors to prepare a library of
well-defined catalysts to systematically describe structure-property relationships with the help from theory and in-situ X-
ray absorption spectroscopy. With this approach, we confirm that PdO is the most active phase, and that small differ-
ences in reaction rates as a function of size are likely due to variations in the surface crystal structure. We further demon-
strate that the support exerts a limited influence on the PdO activity, with inert (SiO2), acidic (Al2O3), and redox-active
(Ce0.8Zr0.2O2) supports providing similar rates, while basic (MgO) supports shows remarkably lower activity. Finally, we
show that introduction of steam leads to a considerable decrease in rates that is due to coverage effects, rather than struc-
tural and/or phase changes. Altogether, the data suggest that to further increase activity and stability of Pd-based cata-
lysts for methane combustion, increasing the surface area of supported PdO phases while avoiding strong adsorption of
water on the catalytic surfaces is required. This study clarifies contrasting reports in the literature about active phase and
stability of Pd-based materials for methane combustion.
KEYWORDS: nanocrystals, methane complete combustion, palladium catalysts, in-situ XAS, structure-property relation-
ships
1. INTRODUCTION
rates to be improved especially in the presence of steam.8–
Also, homogeneous combustion systems are far from
16
Important advances in the extraction of natural gas
from shale have allowed the U.S. to become the world’s
leading natural gas producer and to achieve a new record
of >300 trillion cubic feet in gas reserves. The increased
availability of cheap natural gas has led to renewed inter-
est in the use of methane, its main constituent, to develop
more efficient engines for transportation,1 lower-
temperature fuel cells for the effective generation of elec-
trical energy,2,3 and catalysts for the production of olefins
and aromatics directly from methane.4,5 In all these appli-
cations, methane combustion at low temperatures is cru-
cial for the technologies to perform while limiting harm-
ful methane emissions.6 The main problems with me-
thane utilization are its high greenhouse gas potential,
practical due to the production of toxic gas, such as CO
and NOx.17,18 The challenges posed by methane combus-
tion lie in activating the strong C-H bonds in methane,
while also coupling this step with activation of oxygen in
sites being preferably different from those that activate
the C-H bond, yet in close proximity.11 The discovery of
materials that could deliver high rates at temperatures of
~300 °C or less under demanding conditions and in the
presence of poisoning species would represent a break-
through. Palladium (Pd) is widely accepted as one of the
most active methane combustion catalyst and has re-
ceived continuous attention for the past few
decades.13,14,19–21 While studies in the high-temperature
regime (>600 °C) have studied the kinetics and intercon-
version of PdO and Pd phases, there are still debated
questions in the low-temperature regime.16,22,23 Specifical-
ly, there is no consensus on the reaction active site, but
strong indications have recently emerged in the literature.
One challenge in identifying the active site is that the
7
much higher than CO2 and the stability of the molecule
that makes it rather unreactive. For these problems to be
solved, methods must be developed for the activation of
methane at low temperatures (<300 °C) with stable cata-
lysts. In recent years, there have been important advanc-
es, but supported heterogeneous catalysts still require
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 14
reaction is highly dependent on the Pd oxidation state, the reaction mechanism and significantly increase reac-
1
2
3
4
5
6
7
8
which is known to vary from metallic Pd to an intermedi-
ate chemisorbed oxygen state to bulk PdO depending
upon reaction conditions (and in particular, on the oxy-
gen chemical potential).21 While it is generally accepted
that PdO phase is the most active phase for methane
combustion,14,21,24,25 there are still questions regarding the
structure of the active site, the role of the support, and
the effect of steam on the structural properties of sup-
ported Pd catalysts. A systematic study of these variables
is needed to precisely understand the location and nature
of active sites, and in order to guide future studies to pre-
pare more active and stable catalysts for low-temperature
(<300 °C) methane combustion.
tion rates, understanding the supports’ role in the mech-
anism for Pd–catalyzed methane combustion could help
to engineer more active and stable catalysts.
An additional level of complication in the fundamental
understanding of Pd-catalyzed methane combustion is
represented by the presence of water (steam) during the
reaction at low temperatures (<400 °C). Water is well-
known to strongly deactivate Pd catalysts, even with op-
timized systems that show very good activity in the ab-
sence of water.46 The effect of water depends on the tem-
perature, and weaker effects are observed at increasing
temperatures (>450 °C).16 Kinetic experiments mostly in-
dicate a rate order of about -1 for water indicating strong
coverage of the catalytic surface during low-temperature
Pd-catalyzed methane combustion.13–16,24 A kinetic model
has suggested that the strong inhibition by water is
caused by its tendency to adsorb onto oxygen vacancies
on the PdO surface responsible for the rate-limiting me-
thane activation step.13 Experimental studies have ob-
served irreversible deactivation at low temperatures in the
presence of water, suggesting that it is caused by a slow
transition of the active PdO phase to an inactive Pd(OH)2
surface phase.47–49 Other studies invoke the participation
of the support, with water affecting the hydroxylation of
the support surface and the consequent inhibition of oxy-
gen exchange at the metal-support interface, especially
for redox-active supports.50 For a more complete under-
standing of the water poisoning effect, systematic studies
of the structure/size dependence, as well as support ef-
fects, on water poisoning at low temperatures would be
beneficial.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The effect of Pd particle size is of particular interest,
because a change in this parameter produces materials
with different exposed facets and fractions of under coor-
dinated sites that could result in structure-sensitivity for
the reaction of interest.26 The results of studies of the ef-
fect of Pd size on activity are contrasting: some conclude
that methane combustion on Pd catalysts is insensitive to
particle size,14,15,20,27 while others state that the reaction is
structure-sensitive due to differences in the bond energy
between palladium and oxygen at varying particle
size.13,21,28 The situation is certainly complicated by the
limited control in Pd particle size in conventional materi-
als, or by the fact that these catalysts are made using dif-
ferent Pd precursors, and they can be potentially contam-
inated from the precursor salts (e.g. chloride ions). Con-
clusions from studies on model single crystal surfaces also
disagree in which a study of multiple palladium oxide
facets showed similar activities,15 but others suggest that
the coordinatively unsaturated Pd cations on the PdO(101)
facets are crucial for methane dissociation and thus com-
bustion.29,30 Improved synthetic techniques for supported
size-controlled palladium particles would help to further
our understanding of the structure-sensitivity in Pd-
catalyzed methane combustion.
We herein report a systematic study of the effects of
size, support, and water inhibition on the low-
temperature activity of Pd-catalyzed methane combustion
with the goal of understanding these effects, explaining
contrasting reports, and learning how to improve the ac-
tivity of supported Pd catalysts for this reaction in the
low-temperature regime. This study relies on the synthe-
sis of size-controlled Pd nanocrystals from 2 to 9 nm de-
posited onto high-surface area supports with similar tex-
tural properties but different surface chemistries: Al2O3,
MgO, SiO2, and Ce0.8Zr0.2O2 (CZ80). Kinetic measure-
ments reveal that the activity of palladium is only mildly
sensitive to the nanocrystal size on all four supports, and
that there is a strong trend with respect to support chem-
istry: acidic (Al2O3), inert (SiO2) and reducible (CZ80)
supports have activities over an order of magnitude larger
compared to that of basic support MgO, but surprisingly
alumina and ceria-zirconia deliver similar rates despite
the known oxygen-donating capabilities of this latter
support. Further kinetic and rate order measurements in
the presence of steam demonstrate that these trends do
not change and that the rate limiting step of methane
combustion occurs on PdO.
The support is also known to be crucial for the activity
and stability of Pd catalysts. Numerous supports have
been utilized for Pd-catalyzed methane combustion that
can be classified as inert (e.g. silica,31 zirconia, 13,31–33), acid-
or base- (e.g. alumina and promoted alumina,21,27,28,31,33,34
and zeolites35,36), activeinthe sense that they participate in
8,32,33,37
38
the catalytic cycle (e.g. ceria,
tin oxide, and tita-
nium oxide32,33). These studies have shown that the sup-
port may serve many functions, such as stabilizing the
active PdO phase from thermal decomposition,31,33,39 in-
hibit sintering via strong interactions with the supported
PdO phase,36 and provide reactive oxygen species during
methane combustion.40–42 It is however hard to disentan-
gle the role of the support from other effects of Pd size
and structure because supports with very different mor-
phological properties lead to different Pd size distribu-
tions obtained by impregnations, which in turn can dras-
tically affect its chemistry, and systematic studies of Pd-
support interactions are still needed. As demonstrated in
other oxidation reactions such as CO oxidation43 and wa-
ter-gas-shift reaction,44,45 where the support can change
2. EXPERIMENTAL SECTION
2.1. Synthesis of Pd nanocrystals (NCs). Palladium
NCs syntheses were performed following previously re-
ACS Paragon Plus Environment

Page 3 of 14
ACS Catalysis
and centrifugation. The supernatant was discarded and
ported procedures43,51,52 with small modifications; in par-
1
2
3
4
5
6
7
8
ticular, solvents mixtures were used in place of pure 1-
octadecene (ODE) to achieve reflux conditions that are
known to improve size uniformity,53 and different surfac-
tant-to-palladium ratios were used to tune particle size
control and uniformity. Briefly, palladium (II) acety-
lacetonate (Pd(acac)2), solvents (1-dodecene (DDE), 1-
tetradecene (TDE), or ODE), 1-oleylamine (OLAM) and in
certain cases oleic acid (OLAC) are mixed in a three-neck
flask (see table S1 for further details) and evacuated at RT
for 15 min under stirring. Trioctylphosphine (TOP) was
then added under evacuation and the mixture was heated
to 50 °C for 30 minutes to remove all water and other im-
purities. At this point, the reaction mixture was a trans-
parent yellow-orange colored solution. The reaction flask
was then flushed with nitrogen and rapidly heated (~
40°C min-1) to the desired temperature. After 15 min of
reaction at the appropriate temperature (see Table S1 for
details) and under magnetic stirring, the solution was
quickly cooled to RT by blowing compressed air on the
outside of the flask and adding a water bath when the
temperature was below 170°C. The particles were precipi-
tated with isopropanol and ethanol, and separated by
centrifugation (8000 RPM, 3 min) three times, with redis-
solution in a hexanes/OLAM solution (20 mL Hexanes:
100 µL OLAM) after each centrifugation step. Finally, the
particles were dissolved in hexanes producing a deep
black solution
the final powders were dried at 80 °C overnight and
sieved below 180 µm grain size. Ligands were removed
from the deposited Pd nanoparticles using a previously
published rapid thermal annealing technique,56 in which
catalyst powders were carefully placed in a pre-heated
700°C furnace for 30 s (ATTENTION: the high tempera-
ture in the furnace can cause severe burns and the
insertion and removal of samples should be per-
formed carefully and with heat-resistant gloves and
with tongs). All powders were sieved below 180 µm grain
size after thermal annealing.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.4. Characterization Techniques. N2 physisorption
and CO Chemisorption experiments were carried out on a
Micromeritics 3Flex. For physisorption measurements,
support powders were degassed in vacuum at 350 °C for 12
h prior to N2 adsorption at liquid nitrogen temperature.
For CO chemisorption, catalyst powders were placed in a
U-shaped quartz reactor and then pretreated and de-
gassed in the following manner: evacuated at 110 °C for 30
min, heated in flowing 5% O2 in Ar at 300 °C for 30 min,
evacuated at 300 °C for 30 min, reduced in flowing 5% H2
in Ar at 300 °C for 1 h, and then evacuated at 300 °C for 4
h. CO adsorption experiments were conducted at 35 °C in
the pressure range from 100 to 450 torr for Al2O3, SiO2,
and MgO-based systems and at -70 °C in an ethanol-dry
ice bath in the pressure range from 2 to 20 torr for CZ80-
based systems using a double isotherm to remove the
contribution from physisorption.57 Adsorption values
were obtained by linear extrapolation to zero pressure.
2.2. Support Preparation. Alumina (Al2O3), silica (SiO2),
ceria-zirconia (Ce0.8Zr0.2O2, or CZ80), and magnesia
(MgO) were synthesized and calcined to obtain similar
surface areas (ca. 100 m2 g-1) pore size (ca. 10 nm), and
pore size distributions (centered at ~10 nm). Alumina was
prepared by calcining Pluralox TH100/150 at 900 °C for 24
h using heating and cooling ramps of 3 °C min-1 in static
air. Silica was prepared by calcining silica gel at 900 °C for
8 h using heating and cooling ramps of 3 °C min-1 in static
air. Ceria-Zirconia was synthesized by inverse co-
precipitation of cerium nitrate hexahydrate (Ce(NO3)3 •
6H2O) and zirconium (IV) oxynitrate hydrate
(ZrO(NO3)2•xH2O) using a previously published proce-
dure54 and then calcined at 700 °C for 5 h using heating
and cooling ramps of 3 °C min-1 in static air. Magnesia was
prepared following a previously published procedure55 by
calcining basic magnesium carbonate at 600 °C for 2 h
using heating and cooling ramps of 10 °C min-1. All sup-
ports were sieved below 180 μm grain size after calcina-
tion.
Transmission electron microscopy (TEM) images were
recorded on a FEI Tecnai transmission electron micro-
scope equipped with an Orius CCD and a FEI Titan envi-
ronmental transmission electron microscope equipped
with a spherical aberration corrector in the image forming
lens and a Gatan OneView camera operating at 200kV.
Nanoparticle samples were dropcast onto TEM grids from
their native hexanes solutions onto ultrathin carbon films
supported on Cu TEM mesh grids. Supported catalyst
TEM grids were prepared by dry deposition by lightly
shaking a lacey carbon Cu-mesh TEM grid with catalyst
powder in a plastic tube.
Small Angle X-ray Scattering (SAXS) measurements
were performed at Beamline 1-5 at Stanford Synchrotron
Radiation Lightsource (SSRL) of SLAC National Accelera-
tor Laboratory using a Rayonix 165 SX CCD area detector.
Scattering patterns were analyzed by fitting to a quantita-
tive model using the IRENA package (available at
usaxs.xray.aps.anl.gov/staff/ilavsky/irena.html from the
APS) (Ilavsky, J. & Jemian, P. R. Irena: Tool suite for mod-
eling and analysis of small-angle scattering)58 to deter-
mine the size and size distribution of disperse nanocrys-
tals of the as synthesized particles.
2.3. Catalyst Preparation. An appropriate amount of
metal nanoparticles to give a nominal final loading of 0.5
wt. % Pd, ICP-OES was used to ensure actual weight load-
ings, (see Table S2 for details) was added to a dispersion
of the support (Al2O3, MgO, SiO2, or CZ80) in hexanes
under vigorous stirring. In a typical procedure, 3 g of sup-
port were dispersed by sonication in 30 mL of hexanes
and added with 1.6 mL of a 9.2 mg mL-1 solution of Pd NCs
in hexanes. The mixture was left stirring for 15 min. The
solid was recovered by centrifugation (8000 RPM, 3 min)
and washed once with hexanes (30 mL), with sonication
Quantitative elemental analysis of supported Pd cata-
lysts were analyzed by an ICP-OES (Thermo Scientific
ICAP 6300 Duo View Spectrometer). Catalyst powders
were digested with a mixture of nitric acid (710 µL) and
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 14
hydrochloric acid (660 µL) for seven hours, filtered, and Radiation Lightsource (SSRL) using a Si (220) mono-
then diluted before measurements.
1
chromator detuned by 50% and Lytle Detector. Athena
software was used for data normalization.59 Typically, 3 to
4 scans were averaged and a linear pre-edge was subtract-
ed. Normalization was performed by unit edge jump. k-
weighted EXAFS data in R-space were obtained by sub-
tracting a polynomial background function to the normal-
ized data and processing the resulting χ(k) signal through
Fourier transform. Data were fitted in R-space using the
Artemis software. Single scattering paths amplitude and
phase shift were calculated using structural parameters
from PdO and Pd crystal structure. A Be tube (OD 5mm,
ID 3.8mm, L 65 mm; PF-60 grade Materion) was used as
in-situ cell and an aluminum block was employed as heat-
ing element. Typically, 50 mg of catalyst (Pd 0.5 wt%, <80
mesh) were loaded into the reactor between quartz wool.
After inserting reactor into the flow system, XAS spectra
of the as-prepared catalyst was collected. The catalyst was
then pretreated under a 45 mL min-1 flow of 5% O2 in He
at 300 °C for 30 min and then cooled to 220 °C under He.
XAS spectra were again collected for the after-
pretreatment state of the catalyst. The catalyst was then
put under reaction flow conditions in the absence of wa-
ter (1% CH4 and 4% O2 in He at 80 mL min-1), while XAS
spectra were continually recorded. Once the XAS scans
were stable, water was introduced into the feed stream by
flow the reaction gas through a saturator maintained at
20 °C (0.75% CH4, 3% O2, and 2.3% H2O in He at 80 mL
min-1). Again, XAS spectra were recorded during this time
until stable scans were observed.
2
3
4
5
6
7
8
9
2.5. Catalytic Measurements. All catalytic measure-
ments were conducted at atmospheric pressure in a U-
shaped quartz microreactor with an internal diameter of 1
cm. Catalyst (~20 mg) and bare alumina powders in a 1:10
dilution ratio were physically mixed and loaded into the
reactor between two layers of granular acid-washed
quartz resulting in a bed length of ca. 1 cm. The reactor
was heated by a Micromeritics Eurotherm 2416 furnace
while the temperature of the catalyst was measured with
a K-type thermocouple inserted inside the reactor and
touching the catalyst bed. No appreciable conversions
were found when only quartz or bare supports were
placed in the reactor in the range of temperatures used
for this study.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reaction mixtures (0.75-6% CH4, 0.25-4% O2, 0-12%
H2O, balance Ar) were prepared by passing a mixture of
5% or 20% CH4 (certified standards, Airgas), 5% O2 in Ar
(certified standard, Airgas), and Ar (99.999%, Airgas)
controlled by electronic thermal mass flow controllers
(Brooks SLA5850) through a saturator filled with MilliQ
water and maintained at the appropriate temperature
with a J-KEM 210-K PID controller and K-type thermo-
couple in contact with the outside of the saturator. The
saturator was bypassed for measurements without steam.
Gas hourly space velocities (GSHV) were between 75,000-
750,000 mL g-1 h-1. An on-line gas chromatograph (GC)
(Buck Scientific model 910), equipped with a thermal
conductivity detector (TCD) and flame ionization detec-
tor (FIG) with a methanizer, and using Ar as the carrier
gas was used to monitor the reactant and product
streams.
2.7. DFT Calculations. All calculations were performed
with The QUANTUM ESPRESSO code60 using plane-wave
DFT employing Vanderbilt ultrasoft pseudopotentials.61
The BEEF-vdW exchange correlation functional62 was
used in all calculations. A 4x4x1 Monkhorst-Pack k-point
sampling to model the Brillouin zone and a periodic unit
cell containing 4 atomic layers were used, where the top
two layers, together with the adsorbates, were allowed to
relax whereas the bottom two layers were fixed in their
bulk positions. The unit cell sizes were (2x2) for Pd(100)
and PdO(101), and (3x3) for Pd(111) with 14 Å of vacuum
between successive slabs. The plane-wave and density
cutoff were 500 eV and 5,000 eV, respectively. Geometry
optimizations were performed with a quasi-Newton algo-
rithm as implemented in the Atomic Simulation Envi-
ronment (ASE).63 The convergence criterion for all struc-
tural optimizations was a maximum force of 0.05 eV/Å
per atom. Transition states were determined using the
Prior to measuring catalyst performance, each catalyst
was cleaned under a flow of O2 (5%) in Ar at 45 mL min-1
for 30 min at 300 °C. Then, the sample was cooled to ini-
tial measurement temperature under Ar flow and the re-
action mixture was then introduced. Kinetic rates were
measured at steady-state as determined by a stable CO2
production signal. For kinetic rates, conversions of the
limiting reactant were always kept below 2% conversion
to guarantee differential working conditions. Turnover
frequencies (TOF) were calculated on the basis of accessi-
ble metal surface area calculated from CO chemisorption
measurements using the following equation:
ꢀꢁ
ꢄ
ꢆꢇ
ꢂꢃ
ꢉꢊꢋ
ꢈ
TOF =
where ꢔ is the pressure of the reaction, ꢕꢖꢗ is the molar
mass of Pd, ꢘ is the volumetric flow rate of CH4, ꢙ is
ꢑꢒꢓ
(1)
ꢈ
ꢏ
ꢐ
ꢌꢍꢎ ꢌꢍꢎ
climbing
image-nudged
elastic
band
(CI-NEB)
approach.64
ꢒꢓ
ꢈ
the universal gas constant, ꢚ is temperature, ꢛꢜꢝꢞ is the
A self-consistent mean-field microkinetic model was
solved using CatMap,65 where rates were determined by
solving the coupled differential equations numerically
using the steady state approximation. Free energies were
calculated combining the contributions from DFT calcu-
lations, ZPE, and entropy. For gas species, the entropy
contribution was calculated using Shomate equations.66
Contributions for adsorbed and transition state species
due to vibrational entropy were calculated using the har-
mass of the catalyst in the bed, ꢟ is the weight fraction
ꢖꢗ
of Pd in the catalyst, ꢠ is the dispersion of the catalysts as
measured by CO chemisorption, and ꢑꢒꢓ is the conver-
ꢈ
sion of methane. Serial dilution experiments were per-
formed to ensure no thermal nor mass transport limita-
tions were present.
2.6. in-situ XAS Experiments. Pd K-edge (24350 eV)
XAS data were collected at BL 4-1 at Stanford Synchrotron
ACS Paragon Plus Environment

Page 5 of 14
ACS Catalysis
area supports by adsorption from solution and activated
monic approximation. Further details of the microkinetic
1
2
3
4
5
6
7
8
model can be found in the supplementary information.
via a fast thermal treatment to remove the ligands that
would hinder reactivity.56 As shown in Figure 2 and S1,
and Tables S2 and S4, TEM analysis and CO chemisorp-
tion measurements confirm that the high degree of uni-
formity of all sizes of Pd NCs on all supports is main-
tained through deposition, ligand removal, and after rate
measurements. Inspection of the x-large particles on SiO2
reveals that the particles sintered during chemisorption
measurements due to their poor dispersion (see Figure
S2).
3. RESULTS AND DISCUSSION
3.1. Synthesis of Well-Defined Pd-based Catalysts.
Uniform palladium nanocrystals (NCs) were synthesized
via a modified synthesis43,51,52 and deposited onto several
supports (Al2O3, MgO, SiO2, and Ce0.8Zr0.2O2 (CZ80)) with
similar textural properties (surface area, pore size distri-
butions) to systematically study the structure-property
relationships for Pd-catalyzed methane combustion. For
this study, the particular range of sizes (2 to 9 nm) was
selected because of the large change in the fraction of
different sites (corners edges facets, etc.) that occurs in
this size regime.43–45,67 For simplicity, particles of 2.5, 3.7,
4.3, 6, and 8.2 nm average diameter will be labeled as x-
small, small, medium, large, and x-large through the text.
These uniform Pd NCs were prepared via thermal decom-
position of Pd (II) acetylacetonate in mixtures of high
boiling point solvents (octadecene (ODE), tetradecene
(TDE), and/or dodecene (DDE)) in the presence of 1-
oleylamine (OLAM), trioctylphosphine (TOP), and, in
some samples, oleic acid (OLAC) (see Table S1 for synthe-
sis details) to control size and uniformity and provide
colloidal stability. Representative transmission electron
microscopy (TEM) images, particle size analysis, small
angle X-ray scattering (SAXS) measurements demonstrate
qualitatively and quantitatively the high degree of uni-
formity of the as-synthesized Pd NCs and the level of con-
trol over size achieved by varying the surfactant-to-Pd
precursor ratio (Figure 1, and Tables S1 and S4). The as-
synthesized NCs are then deposited onto high surface
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Control over the support textural properties was also
achieved in order to reduce the number of variables af-
fecting the final catalytic activity. In this study, the cho-
sen supports were calcined such that their surface area,
pore size, and pore volume were similar (see figure S3).
The supports investigated in this study span relatively
inert (SiO2) to acidic (Al2O3) to basic (MgO) to active,
oxygen-donating supports (CZ80). This element allowed
us to better understand the effect of surface chemistries
on the activity of Pd for methane combustion. Compari-
sons of the catalytic activity on the different supports al-
lowed us to understand the effects of acidity and oxygen
storage potential on reaction rates and mechanisms.
Overall, the structural characterization of the samples
confirmed that the library of materials prepared in this
work allowed us to unequivocally correlate structural fea-
tures of the particles (fraction of specific sites at varying
particle size), support chemistry, and reaction conditions
to changes in catalytic activity because we are only vary-
ing
one
of
these
parameters
at
a
time.
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 14
Figure 1. Representative TEM images (A-E), small angle X-ray scattering patterns (F-J), and particle size distributions (K-O) for
uniform Pd NCS (from left to right x-small, small, medium, large, and x-large, respectively). For TEM measurements, N ≥ 300
counts.
1
2
3
4
5
6
7
8
9
3.2. Catalytic Methane Combustion.
cle sintering or drastic shape effects given the characteri-
zation data of the particles after catalysis (see Figure 2
and Table S4). Turnover frequencies (TOFs) were calcu-
lated from rate measurements by measuring exposed Pd
surface area via CO chemisorption on all the samples.
Due to the various conditions used for measuring me-
thane combustion turnover rates in the literature, com-
parisons are often difficult. Using previously reported rate
orders13,15 and assuming no water inhibition, the TOFs in
our work, 4x10-4 to 3x10-2 s-1 at 220 °C, are comparable to
those seen previously in literature for Pd-based
catalysts13,14,24,28,68,69 (see table S5). Normalized rate meas-
urements revealed a weak size dependence for Pd-
catalyzed combustion for all four support series (Al2O3,
SiO2, MgO, and CZ80), in which the difference in activity
for different sizes is between 2 to 5 times (Figure 3 and S6,
and table S5). The difference was larger than the error in
rate measurement or particle size estimation (see figure
S7).
3.2.1. Size Effects. Kinetic rate measurements were per-
formed under lean conditions (O2:CH4 ratio of 4:1) initial-
ly in the absence of excess steam to replicate numerous
studies reported in the literature. To ensure neither mass
transport nor thermal diffusion limitations affected the
results, very low conversions (<2%), high space velocities,
and catalyst dilutions were used in each experiment. Tests
were performed to confirm that dilution was sufficient to
avoid temperature and concentration gradients for the
reactor used in our studies (see figure S4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Additionally, kinetic measurements were performed un-
der true steady-state conditions (see figure S5), when
rates were stable at each temperature step. Because of the
peculiar PdO-Pd interconversion and restructuring occur-
ring in Pd catalysts depending on reaction conditions,
steady state measurements would sometimes need to wait
long stabilization times (hours) before kinetic data could
be taken. This phenomenon is likely not related to parti-
ACS Paragon Plus Environment

Page 7 of 14
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Representative TEM images of Pd NCs before impregnation, after ligand removal, after dry methane combustion, and
after methane combustion in the presence of steam, and particle size distribution histograms (N ≥ 100 counts) for X-Small (A-E),
Small (F-J), Medium (K-O), Large (P-T), X-Large (U-Y) Pd NCs on Al2O3, respectively.
Interestingly, it was not the smallest particles that deliver
the highest rates; instead, particles of intermediate parti-
cle size around 4 to 5 nm showed the largest rates above
all the supports tested. Despite the slight differences in
turnover rates between Pd catalysts on different supports,
the apparent activation energy had similar values of about
80 to 100 kJ mol-1 across all sizes and supports, suggesting
that the rate limiting step for the reaction in all these cat-
alysts was similar, likely the C-H activation on the
Pd/PdO surface in accordance to other reports.13,21
We propose that the following are possible reasons for the
observed mild structure sensitivity: oxidation state of the
Pd phase,21,70 strain effects on the Pd active phase,71–73 pro-
portion of different sites,43–45,67 the relative proportion of
PdO (101) to PdO (100) facets,29,74 and chemisorption ef-
fects.15,24
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 14
that PdO is the thermodynamically stable phase under
1
2
3
4
the working conditions used in these experiments (Figure
S9). This data suggests that the oxidation state did not
change with particle size and thus is not the cause for the
observed structure sensitivity.
5
6
7
8
Strain effects have been shown to enhance catalytic
rates, such as in propene metathesis71 and oxygen reduc-
tion72,73 reactions, by adjusting the bonding strength of
reaction intermediates. As shown in Table S6, fits of
EXAFS spectra demonstrated the bond lengths for the Pd-
O, second shell Pd-Pd, and third shell Pd-Pd did not
change with particle size for alumina supported catalysts,
suggesting that strain effects are not the cause for the
structure sensitivity.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Rate order experiments for O2 and H2O, which are in
13,14,24
agreement with previous reports,
revealed that the
apparent rate orders do not change with particle size (see
figure S10 and table S7), suggesting that the reaction
mechanism is not affected by particle size to a great ex-
tent. The combination of this information and the
knowledge that the oxidation state of Pd (II) remains un-
changed under reaction conditions suggests that the
structure sensitivity may be caused by a small difference
in the proportion of sites43–45,67 or facets,29,74 or by differ-
ent relative proportion of exposed PdO (101) and (100)
facets. To assess the effect of different proportions of
sites, a previously developed physical model for cubo-
octahedral NCs67 was implemented. Fitting the model to
scaling relationships (see Figure S11) demonstrates that in
the size range of interest (2 to 9 nm) the fraction of sur-
face sites from facets (coordination number (CN) > 8) is
proportional to the diameter as ~ꢡꢢ.ꢣ, that of edge sites
(CN = 7) to ~ꢡꢤꢢ.ꢥ, and that of corner sites (CN = 6) to
~ꢡꢤꢦ.ꢧ. While this model assumes the use of a metallic
particle with an FCC crystal structure, whereas our mate-
rials are oxidized palladium during catalysis, we believe
that this model is robust as the scaling relations do not
drastically change with geometries and thus can be used
to assess the proportion of different sites as the cause for
the observed mild structure sensitivity.43–45,67 Plotting the
TOF for the different series of supported Pd catalysts re-
vealed that the physical model does not fit the TOF’s for
all supports (ꢙꢦ < 0.6; Figure S11 and Table S8). This ob-
servation suggests that the structure sensitivity is not the
only cause for the observed activity trend. It has to be
highlighted that to calculate the TOF for these catalysts,
CO chemisorption experiments were performed on the
reduced catalysts. As discussed previously, differences on
the order of 2-4 times have been attributed to the in-
crease in
Figure 3. Turn-over Frequency (TOF) at 220°C (A) and Ap-
parent Activation Energy (B) for all support Pd NC samples
as a function of size as calculated from Arrhenius fits. Rates
were measured under the following conditions: 1% CH4, 4%
O2 in Ar at 175,000 mL gcat-1 h-1.
In-situ X-ray absorption spectroscopy (XAS), DFT cal-
culations, rate order experiments, a surface site model,68
and post-catalysis characterization were implemented to
understand which of these effects are likely contributing
to the mild structure sensitivity. The oxidation state is
extremely important for Pd-catalyzed methane combus-
tion, controlling both the mechanism and reactivity of the
Pd-based catalyst, in which the PdO phase is significantly
more active than the Pd phase.21,74 In-situ XAS measure-
ments revealed that the palladium phase was PdO for
three sizes tested (x-small, medium and large) of alumina
supported Pd NCs, as evidenced qualitatively by XAS
spectra in the X-ray absorption near-edge structure
(XANES, Figure S8) and extended X-ray absorption fine
structure (EXAFS, Figure 4) regions, and quantitatively by
fits to the EXAFS spectra (see Figure 4 and Table S6). DFT
calculations of the equilibrium state further corroborate
ACS Paragon Plus Environment

Page 9 of 14
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Comparison of EXAFS Fourier Transform (FT) magnitudes (A) in k3-weighted, nonphase-corrected R-space for Al2O3
supported x-small, medium, and large nanocrystal samples under working conditions (1% CH4 and 4% O2 in Ar at 220 °C) and Pd
and PdO standards. Curve-fits of EXAFS data for x-small Pd/Al2O3 (B), medium Pd/Al2O3 (C), and large Pd/Al2O3 (D) under
working conditions (1% CH4 and 4% O2 in Ar at 220 °C) is shown for both the FT magnitude and imaginary components, with
curve-fit parameters in Table S5.
surface area during oxidation of metallic Pd.14,15,24 We be-
lieve that the fact that the size of the supported PdO
nanocrystals is maintained after catalysis (both in the
absence and presence of steam) (see Figure 2 and Table
S4) suggest that the chemisorption results in this work
are an accurate representation of the exposed surface area
of the active phases. Thus, the observed mild structure
sensitivity is not only caused by inconsistencies caused by
performing CO chemisorption on the reduced catalysts.
phology may show preference for one exposed facet com-
pared to another.
In conclusion, for these supported catalysts used in this
study, the data showed that PdO is the most active phase
under lean-burn conditions and that there is a mild sensi-
tivity of the reaction to the particle size possibly due to
the proportion of relatively extended PdO facets with a
preferential crystallographic structure. This observation is
in agreement with previous literature reports for support-
ed Pd catalyst systems.29,30,74 We believe that these results
also suggest that previous contrasting reports of struc-
ture-sensitivity in Pd methane combustion may have also
been a result of a combination of factors including vary-
ing proportions of sites and facets.
Finally, a difference in proportion of exposed PdO(101)
and (100) facets can also contribute to the differences in
activity between samples. These two facets, which are the
most thermodynamically stable,75,76 are indeed known to
have different performance.29,30,74 In particular, it has been
theoretically shown that the PdO(100) facet is more stable
but less active than the PdO(101) facet.77 In our case, it is
rather difficult to obtain information about exposed facets
during reaction conditions. Attempts to use HR-TEM to
characterize the post-catalysis samples resulted in beam-
induced artifacts that did not allow us to gain any trusta-
ble conclusion from this characterization. However, we
cannot exclude that particles of different sizes and mor-
3.2.2. Support Effects. Comparison of the kinetic rate
measurements for Pd deposited onto different supports
showed that Al2O3, SiO2, and CZ80 gave very similar per-
formance at a given particle size, whereas MgO support
drastically reduced the catalytic rates of supported Pd,
but there were no substantial differences in the activation
energies in the absence of water in the feed (see Figures 3
and S6). As shown from catalytic rates in Figures 3 and
S6, a reducible support such as CZ80 appeared to not sig-
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 14
nificantly affect intrinsic rates of the Pd phase. However, Pd/MgO showed large initial rates, likely due to a clean
1
2
3
4
5
6
7
8
the significantly lower activity of MgO-supported catalyst
suggests that acidity may have a strong effect on intrinsic
activity.
surface where a large fraction of oxide ions was available
for the reaction, followed by a slow deactivation trend
that lasted for hours, likely due to poisoning of the sur-
face by either oxygen, formed water, or reaction interme-
diates (Figure S15). This behavior has been previously ob-
served for pre-oxidized supported Pd particles and pro-
posed to be caused by slight reduction of PdO.13,31 Howev-
er, while Al2O3-, SiO2-, CZ80-supported Pd catalysts stabi-
lize in 3 hours, the Pd/MgO samples typically took at least
12 hours before reaching steady-state conversions. This
long deactivation period may have been caused by the
slow adsorption of CO2 and formation of magnesium car-
bonate, then responsible for decrease in activity of the
supported Pd on basic supports such as MgO.
Catalytic rates reported in Figures 3 and S6 demon-
strated that the reaction mechanism is similar for each
system and that CZ80 did not enhance Pd-catalyzed me-
thane combustion activity. In order to test whether the
ceria-zirconia support used in this study was reducible
and could participate in oxidation reactions, we per-
formed CO oxidation tests on Pd/CZ80 and Pd/Al2O3 cat-
alysts prepared from the same Pd NC precursor. As pre-
viously shown,43 CZ80-supported palladium catalysts
showed significantly higher rates (>10 times, Figure S12)
compared to alumina-supported Pd catalysts under CO
oxidation conditions, thus demonstrating that the CZ80
used in this work could participate in the catalytic cycle
with its reactive surface oxygens. The surprising fact that
CZ80 did not increase Pd methane combustion rates sug-
gests that activation of oxygen is not a limiting step under
the conditions used in this study (excess oxygen). Further
corroboration to the hypothesis that CZ80 did not direct-
ly affect rates came from rate order experiments which
were also similar between Al2O3-supported catalysts and
CZ80-supported catalysts (see Table S7). In comparison,
reaction rate orders are rather different in CO oxidation
for ceria- and alumina-supported Pd.78
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Overall, the catalytic characterization and in-situ XAS
experiments demonstrate that the support chemistry and
its interaction with the Pd does not strongly affect the
intrinsic activity, nor the mechanism of supported Pd
catalysts.
3.2.3. Water Effects. To understand the effect of water
on Pd-catalyzed methane combustion, kinetic measure-
ments were performed in the presence of excess steam for
the Al2O3- and CZ80-supported samples. We focus on
these two supports because they are the most promising
in terms of catalytic rates, and because of the interesting
comparison between alumina, a redox-inert support, and
ceria-zirconia, which is instead considered as a support
that participates in oxidation reactions. Similar to the
experiments performed in the absence of steam, these
kinetic measurements were performed under steady-state
conditions when the catalysts provided stable rates (see
figure S5B). As expected,13,14,24 the rates were significantly
lower in the presence of steam by over one order of mag-
nitude, further demonstrating the severe deactivating
effect of water on Pd catalysts for methane combustion.
The apparent activation energies also significantly in-
creased from 80-100 to 120-185 kJ mol-1 (see Figure 5 and
S16 and Table S9). This increase in activation energy is
expected because the measurements performed in dry
conditions neglected water poisoning and its negative
rate order. A similar weak size dependence to what was
observed in the absence of steam was found for Pd/Al2O3
catalysts, with rates on intermediate Pd particles about 4-
5 times those on smaller or larger sizes, whereas similar
rates were observed for Pd/CZ80 materials (see Figure 5
and S16 and Table S9). Interestingly, the activation ener-
gies increased at increasing particle size for the Al2O3-
supported Pd catalysts, while they did not change appre-
ciably for the CZ80-support series. In literature, the re-
ported activation energy varies from 135 to 185 kJ mol-1
when the water inhibition effect is accounted for.13,14,24
Repeated kinetic rate measurements for the Pd/Al2O3
study demonstrated that the error in the values of the
apparent activation energies in this study range between 1
and 15 kJ mol-1 (see figure S17), suggesting that the differ-
ence between activation energies is statistically signifi-
cant. Despite these differences, the rate order for water
across different Pd NC sizes and on both CZ80 and Al2O3
With respect to support acid/base properties, MgO-
supported catalysts showed significantly lower activities
compared to the other three supports (Al2O3, SiO2, and
CZ80) (see Figures 3 and S6). Despite this lower activity,
the activation energies were similar, suggesting that the
active site or phase is likely to be similar for all four sup-
ports. There are two potential explanations for the lower
rates observed on Pd/MgO catalysts that are related to
either the electronic state of the PdO phase, or to strong
adsorption of CO2 on the catalytic surface. It has been
suggested previously that MgO may stabilize an electron-
rich PdO phase where the oxide anion is strongly
basic.31,79 This situation would lead to an increased stabil-
ity of the PdO phase, and an increased oxygen vacancy
formation energy. Since efficient C-H activation on PdO
follows a Mars-van Krevelen mechanism with the oxide
ion abstracting a proton and being further converted to
water and desorbed,13 the formation of oxygen vacancies
plays a crucial part in the mechanism. Stabilization of the
oxide ion by MgO support therefore leads to a decrease in
rate. A second possibility is the formation of a magnesium
carbonate phase in close contact with the palladium par-
ticles. MgO is known to adsorb CO2 to form transient
magnesium carbonate species.80,81 These species may af-
fect the surface of supported Pd particles, or the metal-
support interface, thus blocking the active sites. Both hy-
potheses are consistent with the observed transient be-
havior of the catalysts on different supports. Despite there
were no changes in the structure and oxidation state of
the particles from in-situ XAS experiments when contact-
ing the reaction mixture with the catalyst (figures S13 an
S14), the transient methane conversion behavior for
ACS Paragon Plus Environment

Page 11 of 14
ACS Catalysis
effect was observed across different sizes of palladium
is -1 in agreement with literature studies (see Table S7).13–
16,24
1
This discrepancy may be explained by different PdO
NCs (x-small, medium, and large) supported on Al2O3
(figures S13 and S14). Previous work has suggested that
the formation of surface hydroxyls and subsequent diffu-
sion to sub-PdO layers leads to deactivation in the pres-
ence of steam.49 Our in-situ XAS experiments do not sup-
port this argument for the deactivation because no major
differences were observed in the bulk Pd structure once
water was introduced in the reaction atmosphere. Ex-situ
TEM characterization after catalytic reactions both in the
absence and presence of steam demonstrated that the
particles maintain their size after the kinetic studies (Fig-
ure 2 and table S4). This element further demonstrated
that the reaction atmosphere did not significantly affect
the structure, nor cause severe sintering under the condi-
tions of our study. Thus, the observed poisoning effect
and activation energy trends are likely not caused by the
formation of Pd(OH)2 in the bulk or other changes in the
Pd oxidation state and/or structure.
2
3
4
5
6
7
8
9
particle morphologies on the two supports. The particles
may have different proportions of sites on each support.
In the case of alumina, the proportion of sites could
change more drastically than the particles on CZ80 which
would result in the increasing activation energy pattern
on alumina. This element would also suggest that on ce-
ria-zirconia, all sites are uniformly affected by water,
whereas on alumina, the sites may be affected by water to
different extents (hence the different activation energy),
but that overall produce the same rates probably due to
different intrinsic activities.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Another possible explanation for water poisoning is the
competitive adsorption of water on sites for both oxygen
and methane activation, as presented in previous models
and reports.13,14,47 According to these models, water com-
petes with methane for sites and due to the strong bind-
ing of water on PdO, the surface would likely be saturated
with hydroxyl species leaving few sites for methane and
oxygen activation, thus resulting in a significant decrease
in combustion rates. Rate order studies were consistent
with this mode of water poisoning. As discussed previous-
ly in literature, rate orders provide an idea of the coverage
of species participating in catalytic reactions.44,82 The
strong negative rate order, -1, for water on multiple sizes
of supported Pd NCs and on both Al2O3 and CZ80 (Table
S7) implied that water is strongly bound to the PdO sur-
face, and that water surface coverage is near unity. Differ-
ences in transient behavior between experiments in the
absence and presence of steam are also consistent with
this conclusion.
Figure 5. TOF at 275°C (A) and Apparent Activation Energy
(B) for all support Pd NC samples as a function of size as
calculated from Arrhenius fits. Rates were measured under
the following conditions: 0.75% CH4, 3% O2, 4.2% H2O in Ar
at 75,000 mL gcat-1 h-1.
There are two hypotheses to explain the observed deacti-
vation behavior: i) changes in Pd oxidation state or struc-
ture in the presence of water; ii) poisoning and competi-
tion for sites between water and oxygen. We first investi-
gated how structural changes and variation in oxidation
state may have affected the activity. In-situ XAS radial
distributions demonstrated that the state and fine struc-
ture of the Pd particles was not significantly modified
upon introduction of steam to the feed stream. This same
Figure 6. The coverage results from microkinetic model on
PdO(101) under dry conditions (A) (1% CH4, 4% O2, and bal-
ance inert gas) and under wet conditions (B) (0.75% CH4, 3%
O2, 4.2% H2O, and balance inert gas).
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 14
ported catalysts; N2 Physisorption Isotherm and Pore Size
As shown in Figure S18, the catalytic activity in the
presence of steam stabilized typically within approximate-
ly 45 min, in comparison to at least 180 min needed in the
absence of steam. In the presence of large concentrations
of steam, the strong affinity of PdO for water causes rapid
coverage of the surface with hydroxyls such that equilib-
rium is rapidly achieved, resulting in a significantly short-
er time to reach steady state. However, in the absence of
steam, the water produced by the reaction would slowly
adsorb onto the PdO surface and longer times are needed
to reach equilibrium, in accordance to other studies per-
formed in the absence of water,21,70 where long equilibri-
um times were required before collecting kinetic data.
DFT calculations were used to produce a microkinetic
model that further supported the water adsorption model
proposed with the experimental data. Figure 6 shows the
coverage results from the microkinetic model calculations
(further details on the all the elementary steps in micro-
kinetic model are shown in Table S3). In the presence of
high levels of steam, there was a significant increase in
occupation of adsorption sites by intermediate species
(hydrogen and hydroxyls). This phenomenon led to a
drop in the turnover frequency of methane combustion
by an order of magnitude at 500 K (as shown in Figure
S19).
1
2
3
4
5
6
7
8
Distributions; Methane combustion Arrhenius Plots; Transi-
ent Methane Combustion Conversion Data; Additional
XANES and EXAFS Spectra; Pd Phase Diagram; Rate Order
Kinetic Measurement; Physical Model Analysis Plots; CO
Oxidation Lightoff and Kinetic Measurements; Plot of Ratio
of Methane Combustion Rates in the presence and absence
based on Microkinetic Modeling; Tabulated details of NC
synthetic parameters; Tabulated TEM and SAXS NC size da-
ta; Tabulated methane combustion kinetic data in presence
and absence of water; and Tabulated CO chemisorption data;
Tabulated Microkinetic Model Reactions; Tabulated EXAFS
Fit Parameters; Tabulated Apparent Rate Orders; Tabulated
Physical Model Fit Parameters. The Supporting information
is available free of charge on ACS Publications website at
DOI :
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
AUTHOR INFORMATION
Corresponding Author
*mcargnello@stanford.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENT
We gratefully acknowledge support from the U.S. Depart-
ment of Energy, Office of Sciences, Office of Basic Energy
Sciences, to the SUNCAT Center for Interface Science and
Catalysis. M.C. acknowledges support from the School of
Engineering at Stanford University and from a Terman Facul-
ty Fellowship. E. D. G. acknowledges support from the Na-
tional Science Foundation Graduate Research Fellowship
under Grant DGE-1656518. L. W. was supported by the US
Department of Energy (DoE) under the Laboratory Directed
Research and Development. Part of this work was performed
at the Stanford Nano Shared Facilities (SNSF), supported by
the National Science Foundation under award ECCS-1542152.
Part of this work was performed at the Beamline 1-5 at Stan-
ford Synchrotron Radiation Lightsource (SSRL) of SLAC Na-
tional Accelerator Laboratory, and use of the SSRL is sup-
ported by the U.S. Department of Energy, Office of Science,
Office of Basic Energy Sciences under Contract No. DE-
AC02-76SF00515. We would like to thank Prof. Curtis W.
Frank and Prof. Stacey Bent (Stanford University) for shared
laboratory space. We also thank Dr. Ai Leen Koh for her sup-
port while performing TEM experiments.
4. CONCLUSIONS
In conclusion, we have developed a fundamental un-
derstanding of structure-property relationships for sup-
ported Pd catalysts through the use of uniform Pd NCs
deposited on a variety of supports with similar textural
properties but different surface chemistry. Through kinet-
ic measurements, characterization experiments, and im-
plementation of a model, we demonstrated that PdO is
the most active phase for methane complete combustion
and its activity is mildly structure-sensitive likely due to
multiple factors such as proportion of sites and ratio of
PdO(101) facets to PdO(100) facets. Comparison of activity
of Pd NC deposited on different supports revealed a
strong deactivation effect by basic supports (MgO) and
that supports do not participate in the rate limiting step
of Pd-catalyzed methane combustion in other supported
nanoparticle systems. Further kinetic measurements,
characterization experiments, and DFT-calculations
demonstrated that water caused severe decrease in activi-
ty due to dissociation and high coverage levels on the
PdO surface at high water chemical potentials and low
temperatures. We believe this work settles the debated
issues of structure sensitivity, support effects, and water
inhibition for similar supported Pd methane complete
combustion catalysts and acknowledge the difficulty of
extending this understanding to more complicated Pd-
based catalysts.83–85 Overall, this study provides insights
into the design features required for more active methane
combustion catalysts.
REFERENCES
(1)
Mallapragada, D. S.; Duan, G.; Agrawal, R. Energy Policy
2014, 67, 499–507.
(2)
(3)
(4)
Lee, J. G.; Jeon, O. S.; Hwang, H. J.; Jang, J.; Lee, Y.; Hyun, S.
H.; Shul, Y. G. Electrochim. Acta 2016, 191, 677–686.
Qu, J.; Wang, W.; Chen, Y.; Deng, X.; Shao, Z. Appl. Energy
2016, 164, 563–571.
Guo, X.; Fang, G.; Li, G.; Ma, H.; Fan, H.; Yu, L.; Ma, C.; Wu,
X.; Deng, D.; Wei, M.; Tan, D.; Si, R.; Zhang, S.; Li, J.; Sun,
L.; Tang, Z.; Pan, X.; Bao, X. Science. 2014, 344, 616–619.
Horn, R.; Schlögl, R. Catal. Letters 2015, 145, 23–39.
Camuzeaux, J. R.; Alvarez, R. A.; Brooks, S. A.; Browne, J. B.;
Sterner, T. Environ. Sci. Technol. 2015, 49, 6402–6410.
(5)
(6)
ASSOCIATED CONTENT
(7)
(8)
Overview
of
Greenhouse
Gases
Supporting Information. Additional experimental details
on materials, microkinetic model, and phase diagrams; Addi-
tional TEM images and particles size distributions of sup-
https://www.epa.gov/ghgemissions/overview-greenhouse-
gases, accessed October 6, 2017.
Colussi, S.; Gayen, A.; Farnesi Camellone, M.; Boaro, M.;
ACS Paragon Plus Environment

Page 13 of 14
ACS Catalysis
Llorca, J.; Fabris, S.; Trovarelli, A. Angew. Chemie Int. Ed.
8571–8578.
2009, 48, 8481–8484.
(43)
(44)
(45)
Cargnello, M.; Doan-Nguyen, V. V. T.; Gordon, T. R.; Diaz,
R. E.; Stach, E. a.; Gorte, R. J.; Fornasiero, P.; Murray, C. B.
Science. 2013, 341, 771–773.
Shekhar, M.; Wang, J.; Lee, W.-S.; Williams, W. D.; Kim, S.
M.; Stach, E. A.; Miller, J. T.; Delgass, W. N.; Ribeiro, F. H. J.
Am. Chem. Soc. 2012, 134, 4700–4708.
Williams, W. D.; Shekhar, M.; Lee, W. S.; Kispersky, V.;
Delgass, W. N.; Ribeiro, F. H.; Kim, S. M.; Stach, E. A.;
Miller, J. T.; Allard, L. F. J. Am. Chem. Soc. 2010, 132, 14018–
14020.
Monai, M.; Montini, T.; Chen, C.; Fonda, E.; Gorte, R. J.;
Fornasiero, P. ChemCatChem 2015, 7, 2038–2046.
Gélin, P.; Urfels, L.; Primet, M.; Tena, E. Catal. Today 2003,
83, 45–57.
Burch, R.; Urbano, F. J.; Loader, P. K. Appl. Catal. A Gen.
1995, 123, 173–184.
Roth, D.; Gélin, P.; Primet, M.; Tena, E. Appl. Catal. A Gen.
2000, 203, 37–45.
Schwartz, W. R.; Ciuparu, D.; Pfefferle, L. D. J. Phys. Chem.
C 2012, 116, 8587–8593.
Cargnello, M.; Doan-Nguyen, V. V. T.; Murray, C. B. AIChE
J. 2016, 62, 392–398.
Kim, S. W.; Park, J.; Jang, Y.; Chung, Y.; Hwang, S.; Hyeon,
T.; Kim, Y. W. Nano Lett. 2003, 3, 1289–1291.
Lynch, J.; Zhuang, J.; Wang, T.; LaMontagne, D.; Wu, H.;
Cao, Y. C. J. Am. Chem. Soc. 2011, 133, 12664–12674.
Cargnello, M.; Montini, T.; Polizzi, S.; Wieder, N. L.; Gorte,
R. J.; Graziani, M.; Fornasiero, P. Dalt. Trans. 2010, 39, 2122–
2127.
Bartley, J. K.; Xu, C.; Lloyd, R.; Enache, D. I.; Knight, D. W.;
Hutchings, G. J. Appl. Catal. B Environ. 2012, 128, 31–38.
Cargnello, M.; Chen, C.; Diroll, B. T.; Doan-Nguyen, V. V.
T.; Gorte, R. J.; Murray, C. B. J. Am. Chem. Soc. 2015, 137,
6906–6911.
Tanabe, T.; Nagai, Y.; Hirabayashi, T.; Takagi, N.; Dohmae,
K.; Takahashi, N.; Matsumoto, S.; Shinjoh, H.; Kondo, J. N.;
Schouten, J. C.; Brongersma, H. H. Appl. Catal. A Gen. 2009,
370, 108–113.
Ilavsky, J.; Jemian, P. R. J. Appl. Crystallogr. 2009, 42, 347–
353.
Ravel, B.; Newville, M. J. Synchrotron Rad. 2005, 12, 537–541.
Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.;
Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni, M.;
Dabo, I.; Dal Corso, A.; de Gironcoli, S.; Fabris, S.; Fratesi,
G.; Gebauer, R.; Gerstmann, U.; Gougoussis, C.; Kokalj, A.;
Lazzeri, M.; Martin-Samos, L.; Marzari, N.; Mauri, F.;
Mazzarello, R.; Paolini, S.; Pasquarello, A.; Paulatto, L.;
Sbraccia, C.; Scandolo, S.; Sclauzero, G.; Seitsonen, A. P.;
Smogunov, A.; Umari, P.; Wentzcovitch, R. M. J. Phys.
Condens. Matter 2009, 21 (39), 395502.
Laasonen, K.; Car, R.; Lee, C.; Vanderbilt, D. Phys. Rev. B
1991, 43, 6796–6799.
Wellendorff, J.; Lundgaard, K. T.; Møgelhøj, A.; Petzold, V.;
Landis, D. D.; Nørskov, J. K.; Bligaard, T.; Jacobsen, K. W.
Phys. Rev. B 2012, 85, 235149.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(9)
Shan, J.; Huang, W.; Nguyen, L.; Yu, Y.; Zhang, S.; Li, Y.;
Frenkel, A. I.; Tao, F. F. Langmuir 2014, 30, 8558–8569.
Shi, L.; Yang, G.; Tao, K.; Yoneyama, Y.; Tan, Y.; Tsubaki, N.
Acc. Chem. Res. 2013, 46, 1838–1847.
Tao, F. F.; Shan, J.; Nguyen, L.; Wang, Z.; Zhang, S.; Zhang,
L.; Wu, Z.; Huang, W.; Zeng, S.; Hu, P. Nat. Commun. 2015,
6, 7798.
Xiao, T.; Ji, S.; Wang, H.; Coleman, K. S.; Green, M. L. . J.
Mol. Catal. A Chem. 2001, 175, 111–123.
Fujimoto, K.; Ribeiro, F. H.; Avalos-borja, M.; Iglesia, E. J.
Catal. 1998, 179, 431–442.
Ribeiro, F. H.; Chow, M.; Dalla Betta, R. A. Journal of
Catalysis. 1994, pp 537–544.
Zhu, G.; Han, J.; Zemlyanov, D. Y.; Ribeiro, F. H. J. Am.
Chem. Soc. 2004, 126, 9896–9897.
Zhu, G.; Han, J.; Zemlyanov, D. Y.; Ribeiro, F. H. J. Phys.
Chem. B 2005, 109, 2331–2337.
Cook, A. K.; Schimler, S. D.; Matzger, A. J.; Sanford, M. S.
Science. 2016, 351, 1421–1424.
Smith, K. T.; Berritt, S.; Gonzalez-Moreiras, M.; Ahn, S.;
Smith, M. R.; Baik, M.-H.; Mindiola, D. J. Science. 2016, 351,
1424–1427.
Anderson, R. B.; Stein, K. C.; Feenan, J. J.; Hofer, L. J. E. Ind.
Eng. Chem. 1961, 53, 809–812.
Cullis, C. F.; Willatt, B. M. J. Catal. 1983, 83, 267–285.
Chin, Y. H.; Buda, C.; Neurock, M.; Iglesia, E. J. Am. Chem.
Soc. 2013, 135, 15425–15442.
Lyubovsky, M.; Pfefferle, L. Catal. Today 1999, 47, 29–44.
Farrauto, R. J.; Hobson, M. C.; Kennelly, T.; Waterman, E.
M. Appl. Catal. A Gen. 1992, 81, 227–237.
Monteiro, R. .; Zemlyanov, D.; Storey, J. .; Ribeiro, F. . J.
Catal. 2001, 199, 291–301.
Graham, G. W.; Poindexter, B. D.; Remillard, J. T.; Weber,
W. H. Top. Catal. 1999, 8, 35–43.
Boudart, M.; Djega-Mariadassou, G. Kinetics of
Heterogeneous Catalytic Reactions; Princeton University
Press, 1984.
Baldwin, T. R.; Burch, R. Appl. Catal. 1990, 66 (1), 337–358.
Hicks, R. F.; Qi, H.; Young, M. L.; Lee, R. G. J. Catal. 1990,
122, 280–294.
(10)
(11)
(12)
(13)
(46)
(47)
(48)
(49)
(50)
(51)
(14)
(15)
(16)
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(17)
(18)
(52)
(53)
(54)
(19)
(20)
(21)
(22)
(23)
(55)
(56)
(24)
(25)
(26)
(57)
(27)
(28)
(58)
(59)
(60)
(29)
Martin, N. M.; Van den Bossche, M.; Hellman, A.;
Grönbeck, H.; Hakanoglu, C.; Gustafson, J.; Blomberg, S.;
Johansson, N.; Liu, Z.; Axnanda, S.; Weaver, J. F.; Lundgren,
E. ACS Catal. 2014, 4, 3330–3334.
Weaver, J. F.; Hakanoglu, C.; Hawkins, J. M.; Asthagiri, A. J.
Chem. Phys. 2010, 132, 24709.
(30)
(31)
Yoshida, H.; Nakajima, T.; Yazawa, Y.; Hattori, T. Appl.
Catal. B Environ. 2007, 71, 70–79.
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
Hoflund, G. B.; Li, Z.; Epling, W. S.; Göbel, T.; Schneider, P.;
Hahn, H. React. Kinet. Catal. Lett. 2000, 70, 97–103.
Farrauto, R. J.; Lampert, J. K.; Hobson, M. C.; Waterman, E.
M. Appl. Catal. B Environ. 1995, 6, 263–270.
Persson, K.; Ersson, A.; Jansson, K.; Iverlund, N.; Järås, S. J.
Catal. 2005, 231, 139–150.
Neyestanaki, A. K.; Kumar, N.; Lindfors, L.-E. Appl. Catal. B
Environ. 1995, 7, 95–111.
Okumura, K.; Matsumoto, S.; Nishiaki, N.; Niwa, M. Appl.
Catal. B Environ. 2003, 40, 151–159.
Colussi, S.; Trovarelli, A.; Groppi, G.; Llorca, J. Catal.
Commun. 2007, 8, 1263–1266.
Widjaja, H.; Sekizawa, K.; Eguchi, K. Bull. Chem. Soc. Jpn.
1999, 72, 313–320.
Groppi, G.; Cristiani, C.; Lietti, L.; Ramella, C.; Valentini,
M.; Forzatti, P. Catal. Today 1999, 50, 399–412.
Ciuparu, D.; Altman, E.; Pfefferle, L. J. Catal. 2001, 203, 64–
74.
Ciuparu, D.; Bozon-Verduraz, F.; Pfefferle, L. J. Phys. Chem.
B 2002, 106, 3434–3442.
(61)
(62)
(63)
(64)
(65)
Bahn, S. R.; Jacobsen, K. W. Comput. Sci. Eng. 2002, 4, 56–
66.
Henkelman, G.; Uberuaga, B. P.; Jónsson, H. J. Chem. Phys.
2000, 113, 9901–9904.
Medford, A. J.; Shi, C.; Hoffmann, M. J.; Lausche, A. C.;
Fitzgibbon, S. R.; Bligaard, T.; N?rskov, J. K. Catal. Letters
2015, 145, 794–807.
Shomate, C. H. J. Phys. Chem. 1954, 58, 368–372.
Van Hardeveld, R.; Hartog, F. Surf. Sci. 1969, 15, 189–230.
Shen, J.; Hayes, R. E.; Wu, X.; Semagina, N. ACS Catal. 2015,
5, 2916–2920.
(66)
(67)
(68)
(69)
(70)
Roth, D.; Gélin, P.; Kaddouri, A.; Garbowski, E.; Primet, M.;
Tena, E. Catal. Today 2006, 112, 134–138.
Chin, Y.-H.; García-Diéguez, M.; Iglesia, E. J. Phys. Chem. C
2016, 120, 1446–1460.
Schwartz, W. R.; Pfefferle, L. D. J. Phys. Chem. C 2012, 116,
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 14
(71)
Amakawa, K.; Sun, L.; Guo, C.; Hävecker, M.; Kube, P.;
2015, 5, 6187–6199.
Wachs, I. E.; Lwin, S.; Frenkel, A. I.; Patlolla, A.; Hermann,
K.; Schlögl, R.; Trunschke, A. Angew. Chemie Int. Ed. 2013,
52, 13553–13557.
Wang, H.; Xu, S.; Tsai, C.; Li, Y.; Liu, C.; Zhao, J.; Liu, Y.;
Yuan, H.; Abild-Pedersen, F.; Prinz, F. B.; Nørskov, J. K.;
Cui, Y. Science. 2016, 354, 1031–1036.
Friebel, D.; Viswanathan, V.; Miller, D. J.; Anniyev, T.;
Ogasawara, H.; Larsen, A. H.; O’Grady, C. P.; Nørskov, J. K.;
Nilsson, A. J. Am. Chem. Soc. 2012, 134, 9664–9671.
Hellman, A.; Resta, A.; Martin, N. M.; Gustafson, J.;
Trinchero, A.; Carlsson, P.-A.; Balmes, O.; Felici, R.; van
Rijn, R.; Frenken, J. W. M.; Andersen, J. N.; Lundgren, E.;
Grönbeck, H. J. Phys. Chem. Lett. 2012, 3, 678–682.
Weaver, J. F. Chem. Rev. 2013, 113, 4164–4215.
Rogal, J.; Reuter, K.; Scheffler, M. Phys. Rev. B 2004, 69,
75421.
Van den Bossche, M.; Grönbeck, H. J. Am. Chem. Soc. 2015,
137, 12035–12044.
Bunluesin, T.; Putna, E. S.; Gorte, R. J. Catal. Letters 1996,
41, 1–5.
Yazawa, Y.; Yoshida, H.; Hattori, T. Appl. Catal. A Gen.
2002, 237, 139–148.
Lee, S. C.; Choi, B. Y.; Lee, T. J.; Ryu, C. K.; Ahn, Y. S.; Kim,
J. C. Catal. Today 2006, 111, 385–390.
(84)
(85)
Senftle, T. P.; van Duin, A. C. T.; Janik, M. J. ACS Catal.
2017, 7, 327–332.
Cargnello, M.; Jaen, J. J. D.; Garrido, J. C. H.; Bakhmutsky,
K.; Montini, T.; Gamez, J. J. C.; Gorte, R. J.; Fornasiero, P.
Science. 2012, 337, 713–717.
1
2
3
4
5
6
7
8
(72)
(73)
(74)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(75)
(76)
(77)
(78)
(79)
(80)
(81)
Singha, R. K.; Yadav, A.; Agrawal, A.; Shukla, A.; Adak, S.;
Sasaki, T.; Bal, R. Appl. Catal. B Environ. 2016, 191, 165–178.
Borodzinski, A.; Bond, G. C. Catal. Rev. 2008, 50, 379–469.
Senftle, T. P.; van Duin, A. C. T.; Janik, M. J. ACS Catal.
(82)
(83)
Table of Contents (TOC) graphic.
ACS Paragon Plus Environment