Hydrogen-generating behavior of Pd-decorated gold nanoparticles via formic acid decomposition

Article abstract of DOI:10.1016/j.cattod.2018.06.044

Formic acid is a promising hydrogen storage material where hydrogen is generated via metal-catalyzed decomposition. Bimetallic catalysts are active for this reaction, but the mechanism has not been fully proven. Palladium metal supported on gold nanoparticles (Pd-on-Au NPs) has structural properties that are advantageous for studying aqueous-phase catalytic reactions. In this work, a series of Pd-on-Au NPs of varying Pd loadings (calculated in terms of Pd surface coverage, sc%) were synthesized, immobilized onto carbon, and studied for formic acid decomposition at room temperature. Pd-on-Au NPs were catalytically active, with a reaction rate constant as high as 137 m L-H₂/g_Pd/min (corresponding to an initial turnover frequency TOF of 123 h^?1) at a Pd loading of 300 sc%. In contrast, Au NPs were inactive, and Pd NPs were slightly active (5 mL-H₂/g_Pd/min and TOF of 38 h^?1). The Pd metal of Pd-on-Au catalysts are partially oxidized, and is readily reduced without changing the metal-on-metal structure during reaction, according to in situ x-ray adsorption spectroscopy measurements. CO formation was inhibited at a Pd loading of 300 sc%, suggesting that three-dimensional Pd ensembles favored the desired dehydrogenation pathway while single-atom and small two-dimensional Pd ensembles are active for the undesired dehydration pathway.

Full text of DOI:10.1016/j.cattod.2018.06.044

CatalysisTodayxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Hydrogen-generating behavior of Pd-decorated gold nanoparticles via
formic acid decomposition
Zhun Zhaoa,1, Kimberly N. Hecka,1, Pongsak Limpornpipata, Huifeng Qiana, Jeffrey T. Millerb,
⁎
Michael S. Wonga,c,d,e,
a
Department of Chemical and Biomolecular Engineering, Rice University, 6100 S. Main Street, Houston, TX 77005, USA
Davidson School of Chemical Engineering, Purdue University, West Lafayette, IN 47907, USA
Department of Chemistry, Rice University, USA
Department of Civil and Environmental Engineering, Rice University, USA
b
c
d
e
Department of Materials Science and NanoEngineering, Rice University, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
Nanoparticles
Palladium
Gold
Bimetallic
Formic acid
Hydrogen
Catalyst
Formic acid is a promising hydrogen storage material where hydrogen is generated via metal-catalyzed de-
composition. Bimetallic catalysts are active for this reaction, but the mechanism has not been fully proven.
Palladium metal supported on gold nanoparticles (Pd-on-Au NPs) has structural properties that are advantageous
for studying aqueous-phase catalytic reactions. In this work, a series of Pd-on-Au NPs of varying Pd loadings
(calculated in terms of Pd surface coverage, sc%) were synthesized, immobilized onto carbon, and studied for
formic acid decomposition at room temperature. Pd-on-Au NPs were catalytically active, with a reaction rate
constant as high as 137 m L-H2/gPd/min (corresponding to an initial turnover frequency TOF of 123 h−1) at a Pd
loading of 300 sc%. In contrast, Au NPs were inactive, and Pd NPs were slightly active (5 mL-H2/gPd/min and
TOF of 38 h−1). The Pd metal of Pd-on-Au catalysts are partially oxidized, and is readily reduced without
changing the metal-on-metal structure during reaction, according to in situ x-ray adsorption spectroscopy
measurements. CO formation was inhibited at a Pd loading of 300 sc%, suggesting that three-dimensional Pd
ensembles favored the desired dehydrogenation pathway while single-atom and small two-dimensional Pd en-
sembles are active for the undesired dehydration pathway.
1. Introduction
known fuel cell catalyst poison. Noble metals have been long been
known to selectively catalyze FA dehydrogenation via reaction (1).
The chemical energy per mass and zero carbon emissions of hy-
drogen makes it a very attractive potential energy carrier. However, its
safe storage and transportation is problematic. Formic acid (HCOOH,
"FA"), a liquid at room temperature, is a potential source for in situ-
formed H2 for fuel cells and hydrogenation reactions [1–3], due to its
ability to decompose to form up to 4.4 wt% H2. The decomposition of
FA can occur through one of two exergonic reactions:
These catalysts can be divided into two main categories: homogeneous
metal-ligand complexes and supported metal catalysts. Homogeneous
catalysts, based mainly on Ru, Ir, and Rh metals, have been shown to
have excellent catalytic activities in FA decomposition at ambient
temperature [4]. Catalyst recovery issues, use of ligands and solvents,
and excessive additives limit their usefulness.
Monometallic Pd and Au catalysts have been both shown to exhibit
some activity for the FA decomposition reactions, with some reports
finding that PdAu bimetallic catalysts are more active than mono-
metallics [5–10]. Catalysts were typically tested at high FA con-
centrations, under gas or liquid phase, and at elevated temperatures
(> 50 °C). Many catalysts were shown to form high amounts of CO
(10–1000 ppm), which likely contributes to catalyst deactivation.
Wang et al. observed that a graphene-supported Au@Pd (core@
Dehydrogenation: HCOOH ↔ H2 + CO2ΔG = −48.4 kJ/mol
Dehydration: HCOOH ↔ H2O + COΔG = −28.5 kJ/mol
(R1)
(R2)
To produce continuous and stable H2 flow at near-ambient condi-
tions, reaction (1) is desired, whereas reaction (2) should be avoided
since it consumes FA without H2 generation and generates CO, a well-
⁎
Corresponding author at: Department of Chemical and Biomolecular Engineering, Rice University, 6100 S. Main Street, Houston, TX 77005, USA.
E-mail address: mswong@rice.edu (M.S. Wong).
These authors contributed equally to the manuscript.
1
https://doi.org/10.1016/j.cattod.2018.06.044
Received 20 December 2017; Received in revised form 5 June 2018; Accepted 28 June 2018
0920-5861/©2018ElsevierB.V.Allrightsreserved.
Pleasecitethisarticleas:Zhao,Z.,CatalysisToday(2018),https://doi.org/10.1016/j.cattod.2018.06.044

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
shell) catalyst had greater activity compared to an alloyed AuPd cata-
lyst, but the identity of the active site was not investigated [10]. Mullins
and coworkers recently studied single-crystal model Au surfaces cov-
ered with various amounts of Pd, and found that the FA decomposition
rate increased with increasing Pd loading up to 4 monolayers, and that
H
2 production was maximized between 2–3 Pd monolayers [11]. It was
proposed that Pd-metal centers were active for the initial decomposi-
tion reaction, and that interfacial Au was responsible for dehydration
selectivity. This study was conducted in ultrahigh vacuum, far removed
from the practical aqueous-phase conditions of H2 generation from FA.
We suggest Pd-on-Au NP model catalysts (composed of Au NPs
decorated with controlled amounts of Pd) can provide insights into FA
decomposition chemistry. This bimetallic material is highly active for
the aqueous phase hydrodechlorination (HDC) of chlorinated com-
pounds [12–19], aqueous reduction of nitrite [20] and nitrophenol
[21], as well as glycerol oxidation [22,23]. The activity of Pd-on-Au
NPs is dependent on Pd content, which is quantified and metered by
calculated Pd surface coverage (sc%), such that a “volcano-shaped”
dependence of activity on Pd surface coverage was observed for the
various reactions. Using X-ray absorption fine structure (XAFS) spec-
troscopy, we identified Pd surface species, ranging from single atoms to
two-dimensional Pd ensembles (in which all Pd atoms are in contact
with the Au) and three-dimensional Pd ensembles (in which a fraction
of the Pd atoms do not contact the Au). Pd-on-Au NPs have a core-
subshell structure where all Pd atoms are located on the surface of the
Au-rich core and up to ∼20% of them were oxidized, depending on
surface coverage; monometallic Pd NPs had 25–35% of its Pd atoms as
surface atoms and nearly all were oxidized [13,17,18]. Au NPs are es-
sentially able to keep the surface Pd atoms in zerovalent form, leading
to active sites that would not be present in Pd NPs under reaction
conditions [17]. In this study, we synthesized carbon-supported Pd-on-
Au NPs of a range of different Pd sc% and evaluated their catalysis for
FA decomposition in a semi-batch reactor. The reaction products CO2,
CO, and H2 were quantified, and selectivities were calculated. From ex
situ and in situ XAS measurements of selected samples, bulk catalytic
activity was correlated with Pd surface species, with the NPs with the
highest Pd surface coverage as the most active and selective for FA
dehydrogenation.
Fig. 1. Volume of gas released profile for Au/C, Pd/C, and 30, 60, 150 and 300
sc% Pd-on-Au/C catalysts. Reaction conditions: 0.5 g catalyst (1 wt% Pd with
variable Au content), 23 °C, 1200 rpm stirring rate, 10 mL of 1 M formic acid.
slurry was collected and dried in a vacuum oven at 70 °C overnight until
no further mass loss from evaporated water was observed. The material
was then ground into powder form and stored in the dark at ambient
conditions. Activated carbon, in the untreated ("as-is" carbon) and
treated forms ("as-processed" carbon), were used in control experi-
ments. 60, 150, 300 sc% Pd-on-Au/C samples were prepared similarly
using 1407, 542, and 290 mL of sol. The resulting solids of 30, 60, 150,
300 sc% Pd-on-Au/C were calculated to have 13.5, 6.8, 2.5 and 1.4 wt%
Au, respectively. Carbon-supported with 1 wt% Au (Au/C) was pre-
pared in the same manner by mixing 204 mL of Au sol (49.7 mg Au/L)
with 1.0 g of activated carbon; carbon-supported with 1 wt% Pd (Pd/C)
was prepared in the same manner by mixing 314 mL of Pd sol (31.8 mg
Pd/L) with 1.0 g of activated carbon.
2.3. Catalytic testing
2. Experimental
Catalytic activity was evaluated in a homemade semi-batch reactor
by measuring the gas release rate from the reactor (Scheme S1) at room
2.1. Materials
temperature (23
1 °C). 500 mg of catalyst powder was loaded into a
Tetrachloroauric acid (HAuCl4·3H2O, > 99%), tannic acid
(C76H52O46, > 99.5%), potassium carbonate (K2CO3, > 99.5%), palla-
dium chloride (PdCl2, 99.99%), formic acid (HCOOH, 98%), and acti-
vated carbon (Darco-G60) were purchased from Sigma-Aldrich. Sodium
citrate dihydrate (Na3C6H5O7·2H2O, > 99.5%) was purchased from
Fisher Scientific. Commercially available 1 wt% Au/Al2O3 was obtained
from Mintek. Helium (99.99%), hydrogen (99.99%), carbon monoxide
(99.99%), and carbon dioxide (99.99%) were purchased from
Matheson. All experiments were conducted in Nanopure water (> 18 M
Ω-cm, Barnstead NANOpure Diamond). All chemicals were used as
received unless otherwise noted.
glass vial (40 mL, VWR), which was then sealed with Teflon tape and a
Teflon-coated rubber septum. The total Pd charge to the reactor was
125 mg-Pd/L-fluid for all Pd-containing catalysts. The reactor was first
connected to a CO electrochemical detector (0–300 ppm with 0.1 ppm
precision, Environmental Sensors Co.), and then connected with a
horizontally laid plastic bottle (2 L, VWR) loaded with 1.8 L water
(acidified to pH 4 to prevent CO2 dissolution) in series. The amount of
gas formed over time was quantified by weighing the amount of water
displaced from the plastic bottle by the gaseous reaction products. A
series of experiments varying the amount of the most active catalyst
were conducted to ensure the reaction was not mass transfer limited
[16,22] (see Supporting Information).
2.2. Catalyst preparation
The composition of the gas phase products was analyzed by in-
jecting 250 μL of headspace sample into an Agilent Technologies
6890 N gas chromatography (GC) equipped with a thermal conductivity
detector (TCD) and a ResTek PC 3533 Hayesep Q. 60180 packed
column. A GC method with a run time of 4.5 min, helium carrier gas at
a flow rate of 10 mL/min, and an oven temperature of 35 °C was used.
Standard curves for H2, CO, and CO2 were prepared by injecting various
volumes of a gas standard (100 ppm CO, 49.995% H2, 49.995% CO2,
Matheson, see Fig. S2 for representative chromatogram). CO content in
ppm was directly read from the digital display of the CO electro-
chemical detector calibrated with known CO concentrations.
All NPs were synthesized using well-developed protocols reported in
our earlier studies (see Supporting Information) [24–26]. Four com-
positions representative of < 100 sc% and > 100 sc% NPs were chosen:
30 sc%, 60 sc%, 150 sc% and 300 sc% Pd-on-Au. The Pd loading for all
Pd-on-Au/C catalysts was kept constant at 1 wt%, while the Au loading
varied according to the calculated Pd surface coverage.
For the 30 sc% Pd-on-Au/C sample, 2.814 L of sol were mixed with
1 g of activated carbon. The mixture was stirred for ∼24 h at 700 rpm,
cooled to 4 °C, and centrifuged for 40 min at 14,000 rpm. The carbon
2

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Table 1
Catalytic activity results for carbon-immobilized Au, Pd, and Pd-on-Au NPs. Reaction conditions: 0.5 g catalyst, 23 °C, 1200 rpm stirring rate, 10 mL, 1 M formic acid.
Each reaction rate constant was the average of three runs.
Catalyst type Volume of gas released after 3 h Initial H2 generation rate (L h−1
kcat
Initial TOF
H2:CO2 molar
ratioc
CO:CO2 molar ratioc
b
−1
a
(mL)
gPd
)
(L gPd−1 h−1
)
mol-FA·mol-surface-
Pd−1·h−1
Creceived
Cprocessed
Au/C
30 sc%
60 sc%
150 sc%
300 sc%
Pd/C
0
0
0
0
0
0
–
–
–
–
–
–
–
–
–
0.99
1.0
1.0
1.0
1.0
–
–
–
< 0.2
1.0
22.5
58.2
9.8
1.6 × 10−7
0.024
0.162
0.157
0.043
1.6 × 10−9
2.6
1.03 × 10−2
3.01 × 10−4
6.86 × 10−6
0
0.06
7.98
8.22
0.30
35.2
123
38
3.70 × 10−6
a
Initial H2 generation rate based on H2 generation rate after 5 min reaction.
First-order reaction rate constants were calculated using data points in the first 2 h. Reaction rate constants were calculated from the first 0.5 h of the con-
b
centration-time profile for Pd/C, and for the first 1 h of the concentration-time profile for 100, 150, and 300 sc% Pd-on-Au/C catalysts.
c
After 20 min reaction time.
frequency (TOF) was defined as
TOF = kmeas
CFA,0CsPd
(5)
(with units of mol-FA mol-surface-Pd−1 h−1), where CsPd is the
surface metal content (mol-surface-Pd L−1) of the reactor, estimated
using the magic cluster model [22] (see Supporting Information for
details).
2.4. Ex situ and in situ X-ray absorption spectroscopy (XAS)
Catalyst samples were analyzed through both ex situ and in situ XAS
modes. XAS measurements were carried out on Au L3 (11.919 keV) or
Pd K (24.350 keV) edges on the insertion device (beamline 10-ID-B) of
the Materials Research Collaborative Access Team (MRCAT) at the
Advanced Photon Source at Argonne National Laboratory. Detailed
instrumentation can be found in our earlier studies [17,26,27].
Ex situ XAS spectra of each sample were collected first under air at
room temperature. The samples were then treated under flowing 4%
H2/He gas for ∼30 min at 200 °C and cooled to room temperature
under a He purge. Based on our previous studies, this treatment method
fully reduces oxidized Pd in the Pd-on-Au NPs without affecting the
bimetallic nanostructure [17]. The reduced samples were then analyzed
under He at room temperature to determine the extent of Pd oxidation
in the Pd/C and Pd-on-Au/C samples.
In situ XAS measurements were taken using a liquid cell fabricated
out of polyether ether ketone (PEEK) with Kapton windows. The cell
was equipped with gas inlet-outlet and liquid injection ports, and was
capped with a Swagelok VCR fitting with O-ring seal. Prior to mea-
surement, the cell was loaded with 200 mg of catalyst and purged with
He gas at 50 mL/min for 15 min. The gas inlet port was turned off and
the outlet port was connected to a plastic tube immersed in water to
avoid pressure build-up in the cell. The reaction was initiated by in-
jecting 0.5 mL of 0.5 M formic acid solution through the injection port,
and in situ XAS measurements were taken in transmission mode every
5 min for 30 min.
Fig. 2. CO concentration profiles for Au/C, Pd/C, and 30, 60, 150 and 300 sc%
Pd-on-Au/C catalysts. Reaction conditions: 0.5 g catalyst, 23 °C, 1200 rpm
stirring rate, 10 mL, 1 M formic acid.
At 23 °C, the molar volume for an ideal gas is 24.3 L/mol. Since the
reaction generates equal volume amount of CO2 and H2, the converted
amount of FA (CFA,0-CFA, where CFA and CFA,0 were the time-dependent
and initial formic acid concentrations (mol-FA L-1) respectively) can be
calculated as V/(2 × 24.3), where V is the volume of the gas evolved in
liters. Thus, CFA can be further expressed as:
V
CFA = CFA,0
−
(2 × 24.3)
(1)
Formic acid decomposition reaction kinetics was modeled as a
pseudo-first-order reaction (2) with respect to formic acid.
dCFA
dt
= kmeas CFA
(2)
The apparent initial first-order reaction rate constant, kmeas (with
units of h-1), was calculated from fitting Eq. (3) to the first 2 h of the
concentration-time profiles, where t is the reaction time. For catalysts
observed with deactivation, kmeas was also calculated by fitting Eq. (2)
to the first 1 h of the concentration-time profiles as a comparison.
3. Results and discussion
−k
t
3.1. Activity of Pd-on-Au NPs for formic acid decomposition
meas
CFA = C
e
(3)
FA,0
The metal-normalized rate constant kcat (with units of L gPd−1 h−1
)
The activity of the catalysts for formic acid decomposition at room
temperature was evaluated for monometallic Au, Pd, and bimetallic Pd-
on-Au/C. Fig. 1 shows the volume of gas released versus time and
Table 1 lists the calculated reaction rate constants.
The Au/C catalyst had no activity as no gas formation was observed
after 3 h of reaction time (Fig. 1, Table 1), consistent with Sabatier and
co-worker's observations that Au was the least active for this reaction
was defined as kmeas divided by the Pd metal content charged to the
reactor (CPd, gPd L
−1):
kmeas
kcat
=
CPd
(4)
To represent catalytic activity at the particle surface, initial turnover
3

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Table 2
Ex situ EXAFS fit parameters for catalysts.
Catalyst
Treatment
Scattering Edge
Scattering path
CN ( 10%)
R, Å (
0.02 Å)
DWF
Eo (eV)
(× 103 Å2)
Au/C
Pd/C
None
H2 200 °C
None
Au
Pd
Au-Au
Au-Au
Pd-O
Pd-Pd
Pd-Pd
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
9.5
9.3
2.3
3.5
8.6
9.1
2.3
0.9
0.6
9.8
9.4
2.2
0.6
10.0
8.8
2.4
1.3
1.5
7.3
8.7
2.5
2.0
7.8
8.6
2.4
2.6
1.1
2.3
8.5
3.1
6.5
2.7
2.86
2.86
2.04
2.74
2.75
2.85
2.78
–
1.0
1.0
1.0
1.0
1.0
0.0
0.0
–
−0.6
−1.0
−5.6
−1.0
0.2
−0.4
2.6
–
0.2
−6.7
0.2
2.3
0.2
H2 200 °C
None
60 sc% Pd-on-Au/C
150 sc% Pd-on-Au/C
300 sc% Pd-on-Au/C
Au
Pd
2.75
2.78
2.86
2.80
2.75
2.78
2.85
2.78
2.04
2.75
2.78
2.85
2.78
2.75
2.78
2.85
2.78
2.04
2.75
2.78
2.85
2.78
2.75
2.78
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
1.0
1.0
0.0
0.0
1.0
1.0
0.0
0.0
1.0
1.0
1.0
0.0
0.0
1.0
1.0
H2 200 °C
None
Au
Pd
Au
Pd
−7.0
0.3
2.9
0.6
−3.9
−6.0
0.9
H2 200 °C
None
Au
Pd
Au
Pd
2.9
−2.0
−5.8
−0.6
4.8
−1.3
2.0
−2.5
0.4
4.1
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-Pd
Pd-Au
H2 200 °C
Au
Pd
2.1
−4.1
compared to other metals [28,29]. Bi et al. recently reported the very
high activity of Au supported on ZrO2 catalyst, which had a TOF of
1590 h−1 at 50 °C for formic acid decomposition in triethylamine [30].
They concluded that reaction proceeds unimolecularly via an amine-
assisted formate decomposition mechanism at the Au–ZrO2 interface.
Otherwise, Au generally has little activity [5–7].
Pd content increased catalytic activity (Fig. 1, Table 1). The 30 sc%
Pd-on-Au NPs had a very small rate constant, with trace amounts of gas
(< 0.2 mL after 3 h) were observed. The 60 sc% Pd-on-Au/C was
slightly more active with ∼1.0 mL gas released within 3 h, and had a
calculated TOF value of 2.6 h−1. The 150 sc% Pd-on-Au/C initially had
a faster gas release during the first 10 min after which the release rate
slowed down, with a total of 22.5 mL gas generated at 3 h. The TOF
value of 150 sc% Pd-on-Au/C was calculated to be 35.2 h−1, higher
than the 60 sc% one by 13.5 times.
The proportion of CO formed relative to CO2 decreased with in-
creasing Pd surface coverage, with 300 sc% Pd-on-Au/C (the most ac-
tive catalyst) generating no detectable amount of CO. The trend in CO
concentration is consistent with the observed activity order of the
catalysts, correlating catalyst deactivation to the dehydration pathway
(due to CO formation). This correlation of higher FA decomposition
activity and lower CO formation suggests dehydrogenation dominates
over dehydration with increasing Pd surface coverage (at which there
are less PdAu interfaces and larger Pd ensembles). This observation
agrees with what Tedsree et al. observed for Ag-Pd bimetallic NPs
catalyzed room temperature formic acid decomposition [31], and those
of Mullins and coworkers who suggested that PdAu interface sites are
active for the dehydration pathway [11].
3.3. XAS structures of catalysts
The 300 sc% Pd-on-Au/C was the most active catalyst, generating
58.2 mL of gas in 3 h and exhibiting no apparent deactivation.
Compared with published results, our Pd-on-Au catalyst was much
more active than other bimetallic PdAu catalysts (Table S1), and was in
The EXAFS fit parameters of the ex situ measurements of the cata-
lysts are shown in Table 2. For the Au/C catalyst, a Au-Au coordination
number (CN) of ∼9.5 (Table 2) corresponds to an average NP size of
∼4.3 nm based on TEM-EXAFS size correlation developed for sup-
ported Au materials [32]. Au atoms were all in metallic state with no
evidence of oxidation, and no change was observed after treating with
H2. This result is consistent with our previous studies [17,22,26].
All catalysts containing Pd, however, did contain some oxidized Pd,
which is likely due to handling and storage of the catalysts in ambient
air. The 60 sc% Pd-on-Au/C had coordination numbers similar to our
previous results [17,26], indicating a core-shell structure. A Pd-Pd CN
value of 0.6 and Pd-O CN value of 0.9 indicated the Pd atoms were
present as small 2-dimensional (2-D) ensembles and that ∼20% of
those atoms are oxidized. This untreated sample’s ex situ result is si-
milar to our previous results for 60 sc% Pd-on-Au NPs. [17,22]
Based on the high Au-Au CNs, the 150 sc% and 300 sc% Pd-on-Au/C
catalysts also exhibited Au-core and Pd-shell structure, and had
the range of homogeneous Ru catalysts with a TOF value of 123 h−1
.
3.2. Selectivity
The CO2 to H2 molar ratio was 1.00
0.05 at all reaction times for
all catalysts, consistent with the FA dehydrogenation pathway
(Table 1). Dehydration was co-occuring to a limited extent, with ppm-
level formation of CO for most Pd-containing catalysts (Fig. 2). Au/C
was not active and did not generate any CO. Pd/C had an increasing
amount of CO during reaction, reaching ∼1 ppm after 3 h of reaction.
The 30 sc% Pd-on-Au/C catalyst generated the most CO (∼8 ppm)
during the reaction. Increasing the Pd content to 60 sc% and 150 sc%
Pd decreased the amount of CO generated to ∼5 ppm and ∼1 ppm,
respectively.
4

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Table 3
In situ EXAFS fit parameters for catalysts under reaction conditions: 0.2 g catalyst, 23 °C, 0.5 mL, 0.5 M formic acid.
Catalyst
Time point
Scattering Edge
Scattering path
CN ( 10%)
R, Å (
0.02 Å)
DWF
Eo (eV)
(× 103 Å2)
Au/C
Pd/C
Baseline
5 min
Baseline
Au
Pd
Au-Au
Au-Au
Pd-O
Pd-Pd
Pd-O
9.5
9.3
2.3
3.5
–
2.86
2.86
2.04
2.74
–
1.0
1.0
1.0
1.0
–
−0.6
−0.3
−5.6
−1.0
–
5 min
Pd-Pd
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
Pd-Pd
Pd-Au
Au-Au
Au-Pd
Pd-O
7.3
9.1
2.3
0.9
0.6
9.8
8.8
2.4
0.9
0.4
9.5
8.8
2.4
1.3
1.5
7.3
8.8
2.4
1.2
1.3
6.2
8.6
2.4
2.6
1.1
2.3
9.1
2.2
1.2
4.5
4.9
2.76
2.85
2.78
–
2.75
2.78
2.85
2.78
–
2.75
2.78
2.85
2.78
2.04
2.75
2.78
2.85
2.78
2.05
2.75
2.78
2.85
2.78
2.04
2.75
2.78
2.85
2.78
2.05
2.75
2.78
1.0
0.0
0.0
–
0.0
0.0
0.0
0.0
–
1.0
1.0
0.0
0.0
1.0
1.0
1.0
0.0
0.0
1.0
1.0
1.0
0.0
0.0
1.0
1.0
1.0
0.0
0.0
1.0
1.0
1.0
0.1
−0.4
2.6
–
0.2
−6.7
−0.3
3.1
–
0.2
60 sc% Pd-on-Au/C
150 sc% Pd-on-Au/C
300 sc% Pd-on-Au/C
Baseline
Au
Pd
5 min
Au
Pd
−7.3
0.3
2.9
Baseline
5 min
Au
Pd
0.6
−3.9
−6.0
−0.3
3.1
Au
Pd
4.0
−3.6
−5.8
−0.6
4.8
−1.3
2.0
−2.5
0.1
3.5
Baseline
5 min
Au
Pd
Au
Pd
−0.6
−1.7
−2.8
Pd-Pd
Pd-Au
increasingly larger 2-D and 3-D Pd ensembles (Pd-Pd of 1.5 and 1.1, and
Pd-O of 1.3 and 2.6 for the 150 sc% and 300 sc% respectively (Table 2)
compared to the 60 sc% Pd-on-Au/C. The overall degree of oxidation
increased with the increase in Pd; the 150 sc% and 300 sc% had 38%
and 65% of the Pd oxidized, which is expected given the stabilizing
effect of Au on Pd. In comparison, the commercial monometallic Pd/C
was ∼60% oxidized. Treating the catalysts with H2 at 200 °C was suf-
ficient to reduce all the Pd, as evidenced by the disappearance of any
Pd-O in the samples. The structure of the catalysts was mostly un-
changed (that is, no alloying occurred) based on the negligible changes
in Au-Au, Au-Pd, and Pd-Au CN numbers before and after treatment.
The Pd nanoparticle size is estimated to be 3–4 nm using the EXAFS-size
relationship established in previous studies [32].
To determine what effect the reaction conditions had on the cata-
lysts, in situ XAS was performed on the catalysts before reaction and
after injecting the formic acid solution. Table 3 shows the results of the
EXAFS fitting on the samples before reaction and at the first time point
taken after adding the formic acid solution (5 min), after which no
change in CNs was observed. The Au/C catalyst CN remained virtually
unchanged (Table 3). Similarly, the 60 sc% Pd-on-Au/C CNs also stayed
constant and the percentage of oxidized Pd stayed 20% throughout the
30 min (Fig. 3b).
time point taken (30 min, Fig. 3d). The 300 sc% Pd-on-Au/C also had
insignificant change in most of its coordination numbers (Table 3) but
had a dramatic decrease in percentage of oxidized Pd, from 62.5% for
as-synthesized sample to ∼18% after 5 min (Fig. 3h).
Pd/C was similarly affected by the reaction conditions. As can be
seen in Table 3 and Fig. 3g-h, the amount of Pd-O was decreased to
∼10% of total Pd within the first 5 min of reaction time, and the cat-
alyst remained reduced throughout the measurements. The Pd particle
size after 5 min reaction (2–3 nm, corresponding to the measured Pd-Pd
CN of 7.3; Table 2) is smaller than that seen after reduction in H2 at
200 °C (3–4 nm, Pd-Pd CN of 8.6), due to the higher temperature used in
the H2 treatment causing the particles to sinter. The similarity between
the surface Pd of the 300 sc% Pd-on-Au/C and monometallic Pd/C
samples suggests that the electronic interaction between Pd and Au, in
addition to the geometric effect, is important for enhanced activity and
dehydrogenation selectivity for the bimetallic catalyst.
3.4. Pd surface structure, activity and selectivity
Tedsree and co-workers studied the structure-activity relationship of
Ag-Pd core-shell NPs for formic acid decomposition, and attributed the
enhanced activity (relative to monometallic Pd NPs) to electronic and
geometric effects, with a TOF of 192 h−1 observed for the most active
catalyst which consisted of ∼1 layer of Pd on an Ag NP core, with
catalysts containing more or less Pd exhibiting lower activity
[31,33,34]. The size of Pd ensembles in their catalysts was important to
the chemisorption geometry of the formate species and the subsequent
surface reaction pathway. Formate that chemisorbs linearly to smaller
Pd ensembles and edge atoms undergo dehydration, and formate that
The higher loaded Pd-on-Au/C showed more subtle changes during
reaction. The 150 sc% Pd-on-Au/C did not have significant changes in
most of its CNs (Table 3), with the exception of Pd-O, which suggests
the surface structure of the catalyst was maintained. However, the
percentage of oxidized Pd dropped from 37.5% to ∼20% after 5 min,
presumably due to reduction by either formic acid or H2 generated by
the formic acid reaction in situ, which was maintained through the last
5

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 3. XAS spectra at the Pd K edge and the percentage of oxidized Pd for (a, b) 60 sc% Pd-on-Au/C, (c, d) 150 sc% Pd-on-Au/C, (e, f) 300 sc% Pd-on-Au/C and (g, h)
Pd/C at 0, 5 and 30 min reaction times. 30 min scan is offset in e and g for clarity.
chemisorbs in a bridging mode on larger Pd ensembles undergo the
dehydrogenation pathway. Our results suggest that FA goes through a
similar mechanism over the Pd-on-Au catalysts studied here (Scheme
1).
From our XAS analysis, we found the 30 sc% Pd-on-Au NPs had
mostly isolated metallic Pd atoms, and generated the most CO,
consistent with formate binding linearly on the small Pd ensembles,
leading to the dehydration pathway. We hypothesize that the low ac-
tivity observed for this catalyst was likely due to in situ CO poisoning of
the Pd atoms. As the sc% of the Pd-on-Au NPs increases to 60 sc% and
150 sc%, we Pd ensemble size incread and decrease CO decreased,
which is consistent with the hypothesis that the larger Pd ensembles
6

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Scheme 1. Proposed reaction pathways on (a) low sc% Pd-on-Au NP surface, (b) intermediate sc% Pd-on-Au NP surface, and (c) high sc% Pd-on-Au NP surface.
adsorb formate in a bridging conformation and is more selective to the
dehydrogenation pathway. With 300 sc%, the Pd-on-Au NPs likely have
large ensembles of Pd conducive to the bridging conformation, which is
reflected in the very low selectivity to CO and lack of deactivation. The
fact that this catalyst, which has a surface composed primarily of Pd, is
much more active than monometallic Pd catalysts suggests that the
electronic effect of the Au on the Pd still has a significant role in the
increased activity.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.cattod.2018.06.044.
References
[1] B. Loges, A. Boddien, F. Gärtner, H. Junge, M. Beller, Catalytic generation of hy-
drogen from formic acid and its derivatives: useful hydrogen storage materials, Top.
Catal. 53 (2010) 902–914.
[2] J.D. Scholten, M.I. Qadir, V.S. Souza, Nanoparticle-catalysts for hydrogen storage
4. Conclusions
based on small molecules, Recycl. Catal. 2 (2015).
[3] M. Grasemann, G. Laurenczy, Formic acid as a hydrogen source–recent develop-
ments and future trends, Energy Environ. Sci. 5 (2012) 8171–8181.
[4] S. Enthaler, J. von Langermann, T. Schmidt, Carbon dioxide and formic acid—the
couple for environmental-friendly hydrogen storage? Energy Environ. Sci. 3 (2010)
1207.
[5] M. Ojeda, E. Iglesia, Formic acid dehydrogenation on Au-based catalysts at near-
ambient temperatures, Angew. Chem. Int. Ed. 48 (2009) 4800–4803.
[6] D.A. Bulushev, S. Beloshapkin, J.R.H. Ross, Hydrogen from formic acid decom-
position over Pd and Au catalysts, Catal. Today 154 (2010) 7–12.
[7] Y. Huang, X. Zhou, M. Yin, C. Liu, W. Xing, Novel PdAu@Au/C core−shell catalyst:
superior activity and selectivity in formic acid decomposition for hydrogen gen-
eration, Chem. Mater. 22 (2010) 5122–5128.
[8] X. Zhou, Y. Huang, C. Liu, J. Liao, T. Lu, W. Xing, Available hydrogen from formic
acid decomposed by rare earth elements promoted Pd-Au/C catalysts at low tem-
perature, ChemSusChem 3 (2010) 1379–1382.
[9] Ö. Metin, X. Sun, S. Sun, Monodisperse gold–palladium alloy nanoparticles and
their composition-controlled catalysis in formic acid dehydrogenation under mild
conditions, Nanoscale 5 (2013) 910–912.
[10] Z.-L. Wang, J.-M. Yan, H.-L. Wang, Y. Ping, Q. Jiang, Au@ Pd core–shell na-
noclusters growing on nitrogen-doped mildly reduced graphene oxide with en-
hanced catalytic performance for hydrogen generation from formic acid, J. Mater.
Chem. A 1 (2013) 12721–12725.
[11] W.-Y. Yu, G.M. Mullen, D.W. Flaherty, C.B. Mullins, Selective hydrogen production
from formic acid decomposition on Pd–Au bimetallic surfaces, J. Am. Chem. Soc.
136 (2014) 11070–11078.
[12] M.O. Nutt, J.B. Hughes, M.S. Wong, Designing Pd-on-Au bimetallic nanoparticle
catalysts for trichloroethene hydrodechlorination, Environ. Sci. Technol. 39 (2005)
1346–1353.
[13] M.O. Nutt, K.N. Heck, P.J.J. Alvarez, M.S. Wong, Improved Pd-on-Au bimetallic
nanoparticle catalysts for aqueous-phase trichloroethene hydrodechlorination,
Appl. Catal. B 69 (2006) 115–125.
Pd-on-Au/C catalysts with various Pd surface coverages were stu-
died for the room-temperature decomposition of formic acid solutions.
The most active catalyst had a calculated theoretical Pd surface cov-
erage of 300 sc% with a TOF value of 123 h−1, much more active than
monometallic Au and Pd counterparts and literature values measured in
similar conditions. Pd surface coverage played an important role in the
reactivity, with a catalytic activity order of 30 sc% < 60 sc% < 150 sc
% < 300 sc%. The 300 sc% catalyst did not deactivate after 3 h reaction
while lower sc% catalysts did. GC-TCD analysis of headspace gas in-
dicated that the H2 to CO2 ratios for all catalysts were nearly 1:1, and
that the CO concentration increased with decrease in Pd surface cov-
erage. 300 sc% Pd-on-Au had no detectable amount of CO while 30 sc%
one had the highest. In situ XAFS analysis showed that oxidized Pd was
partially reduced for 150 and 300 sc% Pd-on-Au and Pd/C, while the
core-shell structures of all Pd-on-Au catalysts were preserved.
Electronic effects (i.e. Au stabilizing Pd from oxidation) and geometric
effects (i.e. the structure of surface Pd domains as well as the formation
of PdAu interfacial sites) were proposed to explain the observed activity
dependence on Pd surface coverage. Large 3-D Pd ensembles on 300 sc
% Pd-on-Au NPs favor the dehydrogenation pathway of formic acid
decomposition via bridging binding, while scattered Pd atoms as well as
small 2-D Pd ensembles on Pd-on-Au NPs with lower surface coverages
favor the dehydration pathway via linear binding, leading to CO for-
mation and active site poisoning. These results give further insights into
the role of Pd-ensemble size and Au electronic effects in the conversion
of FA to H2.
[14] K.N. Heck, M.O. Nutt, P.J.J. Alvarez, M.S. Wong, Deactivation resistance of Pd/Au
nanoparticle catalysts for water-phase hydrodechlorination, J. Catal. 267 (2009)
97–104.
[15] M.S. Wong, P.J.J. Alvarez, Y.L. Fang, N. Akcin, M.O. Nutt, J.T. Miller, K.N. Heck,
Cleaner water using bimetallic nanoparticle catalysts, J. Chem. Technol. Biot. 84
(2009) 158–166.
[16] Y.L. Fang, K.N. Heck, P.J.J. Alvarez, M.S. Wong, Kinetics analysis of Palladium/
Gold nanoparticles as colloidal hydrodechlorination catalysts, ACS Catal. 1 (2011)
128–138.
Acknowledgments
[17] Y.L. Fang, J.T. Miller, N. Guo, K.N. Heck, P.J.J. Alvarez, M.S. Wong, Structural
analysis of palladium-decorated gold nanoparticles as colloidal bimetallic catalysts,
Catal. Today 160 (2011) 96–102.
[18] H.J.S.L.A. Pretzer, Y. Fang, Z. Zhao, N. Guo, T. Wu, I. Arslan, J.T. Miller,
M.S. Wong, Regulating the hydrodechlorination activity of Pd-on-Au nanoparticles
with Au particle size, J. Am. Chem. Soc. (2012) submitted.
[19] J.C. Velázquez, S. Leekumjorn, Q.X. Nguyen, Y.-L. Fang, K.N. Heck, G.D. Hopkins,
M. Reinhard, M.S. Wong, Chloroform hydrodechlorination behavior of alumina-
supported Pd and PdAu catalysts, Aiche J. 59 (2013) 4474–4482.
[20] H. Qian, Z. Zhao, J.C. Velazquez, L.A. Pretzer, K.N. Heck, M.S. Wong, Supporting
palladium metal on gold nanoparticles improves its catalysis for nitrite reduction,
We acknowledge financial support from the National Science
Foundation (CBET-1134535), Sigma Xi Grants-in-Aid of Research
(GIAR) program (G20111015157503), and Rice University. JTM was
supported by the National Science Foundation under the Cooperative
Agreement No. EEC-1647722. We thank use of the Advanced Photon
Source was supported by the U. S. Department of Energy, Office of
Science, and Office of Basic Energy Sciences (DE-AC02-06CH11357).
MRCAT operations are supported by the Department of Energy and
MRCAT member institutions.
7

Z. Zhao et al.
CatalysisTodayxxx(xxxx)xxx–xxx
selectivity of palladium-decorated gold for glycerol oxidation, J. Am. Soc. Chem.
(2013) to be submitted.
Nanoscale 6 (2014) 358–364.
[21] L.A. Pretzer, K.N. Heck, S.S. Kim, Y.-L. Fang, Z. Zhao, N. Guo, T. Wu, J.T. Miller,
M.S. Wong, Improving gold catalysis of nitroarene reduction with surface Pd, Catal.
Today 264 (2016) 31–36.
[28] A. Mailhe, P. Sabatier, Catalytic decomposition of formic acid, Comp. Rend. Acad.
Sci. 152 (1911) 1212–1215.
[22] Z. Zhao, J. Arentz, L.A. Pretzer, P. Limpornpipat, J.M. Clomburg, R. Gonzalez,
N.M. Schweitzer, T. Wu, J.T. Miller, M.S. Wong, Volcano-shape glycerol oxidation
activity of palladium-decorated gold nanoparticles, Chem. Sci. 5 (2014)
3715–3728.
[23] Z. Zhao, J. Miller, T. Wu, N. Schweitzer, M. Wong, EXAFS characterization of
palladium-on-gold catalysts before and after glycerol oxidation, Top. Catal. 58
(2015) 302–313.
[24] M.O. Nutt, K.N. Heck, et al., Improved Pd-on-Au bimetallic nanoparticle catalysts
for aqueous-phase trichloroethene hydrodechlorination, Appl. Catal. B-Environ. 69
(2006) 115–125.
[29] O. Deutschmann, H. Knözinger, K. Kochloefl, T. Turek, Heterogeneous Catalysis and
Solid Catalysts, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag
GmbH & Co. KGaA, 2000.
[30] Q.-Y. Bi, X.-L. Du, Y.-M. Liu, Y. Cao, H.-Y. He, K.-N. Fan, Efficient subnanometric
gold-catalyzed hydrogen generation via formic acid decomposition under ambient
conditions, J. Am. Chem. Soc. 134 (2012) 8926–8933.
[31] K. Tedsree, T. Li, S. Jones, C.W.A. Chan, K.M.K. Yu, P.A.J. Bagot, E.A. Marquis,
G.D.W. Smith, S.C.E. Tsang, Hydrogen production from formic acid decomposition
at room temperature using a Ag-Pd core-shell nanocatalyst, Nat. Nano 6 (2011)
302–307.
[25] Y.-L.F.Z. Zhao, P.J.J. Alvarez, M.S. Wong, Degrading perchloroethene at ambient
conditions using Pd and Pd-on-Au reduction catalysts, Appl. Catal. B-Environ. 140-
141 (2013) 468–477.
[26] L.A. Pretzer, H.J. Song, Y.-L. Fang, Z. Zhao, N. Guo, T. Wu, I. Arslan, J.T. Miller,
M.S. Wong, Hydrodechlorination catalysis of Pd-on-Au nanoparticles varies with
particle size, J. Catal. 298 (2013) 206–217.
[32] J.T. Miller, A.J. Kropf, Y. Zha, J.R. Regalbuto, L. Delannoy, C. Louis, E. Bus, J.A. van
Bokhoven, The effect of gold particle size on AuAu bond length and reactivity to-
ward oxygen in supported catalysts, J. Catal. 240 (2006) 222–234.
[33] K.S. Kim, M.A. Barteau, Structural dependence of the selectivity of formic acid
decomposition on faceted titania (001) surfaces, Langmuir 6 (1990) 1485–1488.
[34] N. Hoshi, M. Nakamura, K. Kida, Structural effects on the oxidation of formic acid
on the high index planes of palladium, Electrochem. Commun. 9 (2007) 279–282.
[27] Z. Zhao, J. Arentz, L.A. Pretzer, P. Limpornpipat, J.M. Clomburg, R. Gonzalez,
N. Schweitzer, T. Wu, J.T. Miller, M.S. Wong, Tuning the catalytic activity and
8