Sub-5 nm Pd-Ru nanoparticle alloys as efficient catalysts for formic acid electrooxidation

Article abstract of DOI:10.1002/cctc.201400086

Both Pd and Ru are important elements in electrochemistry. However, Pd-Ru nanoparticle (NP) alloys are rarely reported so far, owing to the limitations described by Hume-Rothery. Herein, we successfully synthesized Pd-Ru NP alloys over the whole compositi

Full text of DOI:10.1002/cctc.201400086

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201400086
Sub-5 nm Pd–Ru Nanoparticle Alloys as Efficient Catalysts
for Formic Acid Electrooxidation
[
a]
Dongshuang Wu, Minna Cao, Min Shen, and Rong Cao*
Both Pd and Ru are important elements in electrochemistry.
However, Pd–Ru nanoparticle (NP) alloys are rarely reported so
far, owing to the limitations described by Hume-Rothery.
Herein, we successfully synthesized Pd–Ru NP alloys over the
whole composition range using a facile atomic-diffusion-based
strategy, and we carefully investigated the corresponding for-
mation mechanism. For the first time, we confined the size of
all alloy NPs at 3–5 nm, which is favorable in direct formic acid
fuel cells. Experimental data indicate that the core-level bind-
ing energy of Pd positively shifts as a result of electron transfer
from Pd to Ru. Such an electronic effect can effectively tune
the adsorption intensity of reactive molecules during electroca-
talysis. Driven by all these effects, these Pd–Ru alloys are
highly active and stable towards formic acid oxidation. Compo-
sition–activity study further indicates that Pd Ru exhibits the
60
40
best mass activity, which is 4.1 times higher than that of com-
mercial Pd/C.
Introduction
Formic acid oxidation reaction (FAOR), as the anodic reaction
in direct formic acid fuel cells, has attracted considerable atten-
loying process between Pd and Ru. Especially, for example, Pd
and Ru belong to different crystalline phase (face-centered
cubic, fcc, vs. hexagonal close-packed, hcp), and, thus, they are
[
1]
tion. It is believed that Pd is the most efficient metal in cata-
lyzing this reaction because of its low oxidation overpotential
[13b,14]
immiscible even at 773 K.
Moreover, the standard reduc-
[
2]
2À
and substantially high performance. However, moving to-
wards its commercialization requires more effort to improve
activity and stability. This promotion can be achieved by intro-
ducing another metal to form Pd-based bimetallic alloys.
Adding alloyed elements can provide a bifunctional or elec-
tion potential of the redox pair PdCl4 /Pd (0.95 V vs. standard
3
+
hydrogen electrode, SHE) is much higher than that of Ru /Ru
(0.35 V vs. SHE), and, thus, there is a huge gap in the nuclea-
tion and growth rate of the two metals. Until recently, Kitaga-
[15]
wa’s group has prepared Pd–Ru alloy NPs with the average
size ranging from 6.4 to 12.5 nm over the whole composition
range. These NPs display enhanced CO-oxidizing catalytic activ-
ity. However, the formation mechanism of Pd–Ru alloy NPs re-
mains fuzzy, and the previously reported nanoparticle size is
still relatively large for fuel cell reactions.
[
3]
tronic effect for catalysis. For example, the latter can dramati-
cally change the electronic structure of Pd, thereby tuning the
[
4]
adsorption intensity of reactive molecules. To date, various
[
5]
[6]
[7]
Pd-based alloy NPs, such as Pd–Fe, Pd–Co, Pd–Ni, Pd–
[
8]
[9]
[10]
[11]
Cu, Pd–Au, Pd–Ag, Pd–Pt etc., have been conveniently
obtained and showed enhanced catalytic properties.
Herein, using a facile one-pot polyol strategy, we present
a successful preparation of Pd–Ru alloy NPs over the whole
composition range. For the first time, all these Pd–Ru alloy NPs
are well dispersed in a narrow size distribution ranging from 3
to 5 nm, which is favorable in direct formic acid fuel cells.
Moreover, we carefully investigated the corresponding forma-
tion mechanism. The as-synthesized Pd–Ru alloy NPs not only
exhibit higher activity, but also better stability towards FAOR, if
compared with both home-made Pd catalysts and commercial
Pd/C catalysts.
Ru, as a chemical element that can easily adopt various oxi-
dation states, is another important element in electrocataly-
[
12]
sis. Ru proves to greatly enhance the catalytic properties of
[4-
Pt through the bifunctional effect, electronic effect, or both.
d,e,12c]
Therefore, we deduce that great performance improve-
ment can be available with Pd–Ru alloys in FAOR. However,
a facile synthesis of Pd–Ru alloy nanoparticles (NPs) with tuna-
ble composition, homogenous shape, and narrow size distribu-
tion is currently limited. Owing to the limitations described by
[
13]
Hume-Rothery, there are several problems hindering the al-
Results and Discussion
[
a] D. Wu, Dr. M. Cao, M. Shen, Prof. R. Cao
State Key Laboratory of Structural Chemistry
Fujian Institute of Research on the Structure of Matter
Chinese Academy of Science
Characterization of Pd–Ru alloy NPs
In a typical experiment, Pd–Ru alloy NPs were prepared in
a boiling ethylene glycol (EG) solution by reduction of RuCl3
1
55 Yangqiao West Road Fuzhou Fujian 350002 (P. R. China)
Fax: (+86)591-83714946
E-mail: rcao@fjirsm.ac.cn
first and then addition of K PdCl in the presence of polyvinyl-
2
4
pyrrolidone (PVP). Such a feeding sequence can bridge the
gap of nucleation and growth rate to facilitate alloying. We
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201400086.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
1
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
0
have testified that K PdCl can be fast reduced into Pd NPs at
firms formation of the alloy (Figure 1 f and inset). Moreover,
the EDX line-scanning profile reveals the particles contain
a slightly higher proportion of Pd than Ru, which is in good
agreement with EDX elemental analysis revealing 57.2% of Pd
and 42.8% of Ru and inductively coupled plasma mass spec-
trometry (ICP–MS) revealing 60.5% of Pd and 39.5% of Ru. The
alloy formation in the other Pd–Ru NP alloys was also con-
firmed by HAADF–STEM and elemental mapping/line-scan
analyses (see the Supporting Information, Figure S1).
2
4
8
08C in EG solution during the experiment. Nevertheless, the
RuCl will not obviously undergo reduction until the tempera-
3
ture is elevated to 1708C under the same conditions. Thus the
reduction of Pd salts will be much faster and easier than that
of Ru salts in the same environment. The probable reason for
this is that the standard reduction potential of the redox pair
2
À
3+
PdCl4 /Pd (0.95 V vs. SHE) is higher than that of Ru /Ru
[
13b]
(0.35 V vs. SHE).
Therefore, herein the RuCl was added first
3
to earn itself more time to nucleate. The Pd/Ru molar ratio can
be adjusted over the whole composition range by solely mod-
ulating the feeding ratio between the two metal precursors.
In Figure 1a and b, typical bright-field TEM and high-angle
annular dark-field scanning transmission electron microscopy
The XRD pattern was used to characterize the purity of the
as-synthesized Pd–Ru alloys. As shown in Figure S3, each
sample exhibits a strong diffraction peak, which locates be-
tween the most intense peak of Pd(111) at 40.18 and Ru(101)
at 44.08, and shifts to high angles (closer to the Ru(101)) with
Pd/Ru composition decreasing. Such diffraction peaks match
[
17]
closely with the previously reported Pd–Ru alloy,
which
clearly states the formation of Pd–Ru random alloy. The XRD
results are exactly consistent with HRTEM and STEM analyses,
which do not provide evidence of core-shell or Janus-type par-
ticle, indicating that both metals are present within a single
particle.
X-ray photoelectron spectroscopy (XPS) was performed to
probe the alternations in surface electronic structure and to
clarify whether oxidation has impacted the observed crystal
structures. For 100% Ru NPs, the deconvolution of high-resolu-
tion XPS spectra of Ru3d, 3p3/2, and O1s is readily assigned to
the formation of RuO2 (Figure S4). Such a strong oxidation
(
with ꢀ70% of Ru atoms in oxidative states) can be ascribed
[18]
to the poor protection of PVP on Ru NPs, as well as its small
[19]
particle size. However, for alloyed compositions, employing
Pd Ru alloy as a showcase, the XPS spectrum indicates the
60
40
Figure 1. a) Low-magnification bright-field TEM and b) HAADF–STM images
of Pd60Ru40 NPs, the inset gives the HRTEM image of an optional NP; c) high-
magnification HAADF–STEM image and the corresponding EDX mapping of
d) Pd and e) Ru; f) EDX line-scan profile of optional two neighbored NPs.
0
presence of solely metallic Ru located at 279.9 eV (Ru3d5/2)
and 461.2 eV (Ru3p3/2) without any clue to ruthenium oxides
[19b,20]
(
Figure 2a, b).
In Figure 2c, the two peaks at 336.7 and
3
41.9 eV can be attributed to metallic Pd3d5/2 and 3d , re-
3/2
spectively. Compared with those of the monometallic Ru and
Pd in solid bulk, the binding energies of Ru3p3/2 and Pd3d5/2
in Pd Ru alloy NPs have a negative shift of approximately
(
HAADF–STEM) images of the well-dispersed Pd Ru NP alloys
60 40
(
3.80Æ0.35 nm) are shown, respectively. Under similar condi-
60
40
tions, 4.19Æ1.39 nm Pd NPs, 3.10Æ0.72 nm Pd Ru NPs,
À0.4 eV and a positive shift of approximately +1.5 eV, respec-
90
10
3
0
.30Æ0.61 nm Pd Ru , 3.88Æ0.57 nm Pd Ru , and 3.51Æ
tively. Such a strong shift is also reported in many Pd-based
75
25
40
60
[21]
[22]
.70 nm Pd Ru were also formed (Figure S1, 2). All NP alloys
alloys, such as Pd–Ag, Pd–Al, and others. The O1s XPS
20
80
[23]
were ranging from 3 to 5 nm, demonstrating the homogeneity
of this method. In the inset in Figure 1a, the corresponding
high-resolution (HR) TEM image of an optional Pd Ru NP is
spectra originate solely from carbonyl in PVP molecules, fur-
ther verifying that there is no sign of metal oxides (Figure 2d).
Therefore, both Pd and Ru maintain in their metallic state in
the alloy. Notably, such a significant change in the electronic
state of Pd atoms caused by Ru incorporation further confirms
60
40
shown, the shape of which is pseudo-spherical, without long-
range order. As the size of Pd–Ru alloy NPs is extremely small,
their growth into well-defined low-index planes of its surface is
too difficult. On the contrary, small NPs are bounded by vari-
ous low- and high-index planes, generating a large lattice
stress on their surfaces. As a result, these small NPs usually
provide more favorable terms for catalytic reactions than the
[9b,12c]
the formation of an alloy phase.
Formation mechanism of Pd–Ru alloy NPs
To figure out the formation process of Pd–Ru alloys, aliquots of
the reaction solution were withdrawn at different periods of
time for TEM analysis. In the initial stage of the reaction (t=
40 s, Figure 3a), Ru NPs with diameter of approximately 1.2 nm
(size-distribution histogram in Figure S5a) were formed, which
can also be testified by the EDX elemental analysis in Fig-
ure S6. Analysis by HRTEM (Figure 3a inset) revealed that the
[
2b,16]
bigger ones.
In Figure 1c–e, the HAADF–STEM image of
several representative Pd Ru NPs and the corresponding
60
40
energy-dispersive X-ray (EDX) mapping with Pd (red) and Ru
green) are shown, respectively. The color distribution within
(
each NP suggests that the alloy has been formed. EDX line
scan of two optionally neighbored Pd Ru NPs further con-
60
40
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
2
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[25]
“
crystalline” NPs.
Following,
we began to add the K PdCl so-
2
4
lution (3 mL) drop by drop, and
this process took approximately
4–5 min. At t=3 min, we had
dropwise added 1.5 mL of
K PdCl solution. EDX elemental
2
4
analysis reveals the coexistence
of both Ru and Pd (Figure S7).
The HAADF–STEM image clearly
indicates that several brighter
NPs (Pd) adhere to each darker
particle (Ru), suggesting the for-
mation of Pd-on-Ru heterostruc-
ture by heterogeneous nuclea-
tion (Figure 3b). The average
size of the heterostructure was
2.40 nm (Figure S5b), indicating
that the adhesive Pd cluster was
approximately 0.6 nm in size.
When we finished adding
K PdCl at t=5 min, homogene-
2
4
ous NPs of approximately 3.2 nm
were formed (Figure S5c and
S8a). The STEM mapping scan
further confirmed the alloy
nature (Figure 3c). With exten-
sion of the reaction time to 1 h,
the size increased to approxi-
mately 3.8 nm (Figure 1a and b).
As the precursors were com-
pletely depleted at high temper-
ature near 1968C, the growth of
the Pd–Ru NP alloy was mainly
attributed to Ostwald ripen-
Figure 2. High resolution XPS spectrum of Pd60Ru40 alloy NPs and its simulated peak fittings of a) Ru3d+ C1s,
b) Ru3p3/2, c) Pd3d, d) O1s.
[11]
ing. Further extension of the
reaction time to 10 h did not
result in any obvious morpho-
logical/size change (Figure S5d
and S8b–c), indicating that the
Pd–Ru NP alloys were highly
stable in EG at such high
temperature.
From the above results, we
can deduce
a
three-stepped
atomic diffusion mechanism for
the formation of sub-5 nm Pd–
Ru alloy NPs. As shown in Fig-
ure 3d, first, high temperature is
an essential condition for the al-
loying. Experimental observation
Figure 3. Time-dependent TEM/HAADF–STEM images in the formation of Pd60Ru40 NPs: a) 40 s, the presence of Ru
NPs; the inset gives the HRTEM image; b) 3 min, half of the volume of K
2 4
PdCl solution was added dropwise to
form heterostructures; c) 5 min, finished adding of K PdCl , the insets in 5d shows the corresponding EDX map-
2
4
ping of the squared area; d) proposed formation mechanism.
Ru NPs exhibit poorly defined lattice fringes and its surface,
suggesting an aggregation of Ru atoms rather than crystalline
state. Such phenomenon has also been observed to be domi-
nant in 0.8–1.5 nm noble-metal NPs synthesized by other
indicated that the reduction of RuCl processes much more
3
slowly than that of Pd below 1708C. On the contrary, at 1968C
the added RuCl bursts into nucleation and forms the initial Ru
3
NPs with very small sizes. Then the newly reduced Pd atoms
are prone to confine to the vicinity of the Ru NPs surface, be-
cause such heterogeneous nucleation can greatly reduce the
[
19b,24]
methods.
Notably, the coarse surface on these “disor-
dered” Ru is more accessible to deposit Pd atoms than the
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
3
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
global Gibbs free energy if compared to the homogeneous nu-
[
4c,26]
cleation.
Notably, the speed with which we added K PdCl4
2
solution was also important in this case. If we added too fast
at first, the concentration of Pd salts in EG would have been
too high for heterogeneous nucleation. On the contrary, homo-
geneous nucleation would happen, leading to phase separa-
tion. Once the concentration of Pd atoms has reached a critical
value, these Pd atoms will nucleate and grow into small clus-
ters around the Ru NPs, and form a Pd-on-Ru heterostructure.
At the end, Pd–Ru NP alloys formed by means of atomic diffu-
sion under conditions of reflux at high temperature. The nucle-
ation and growth processes of Pd are believed to initiate on
the facets of Ru NPs with the highest surface energy and then
[
27]
proceed to those with lower energies. If the reduction rate is
slow, such as at a lower temperature, the nucleation and diffu-
sion processes will be under equilibrium conditions, finally
forming core-shell NPs or Pd-on-Ru heterostructures. However,
under the high-temperature condition, the fast reduction rate
will break down near-equilibrium conditions and favor the for-
[
25a,26]
mation of alloy NPs by quick diffusion.
In addition, below
the melting point, randomly mixed alloy NPs are more stable
than core-shell NPs and heterostructures because of the lower
[
28]
Gibbs free energy.
Catalytic performance of Pd–Ru alloy NPs
Driven by these collective advantages, the Pd–Ru NP alloys
were further characterized by their compositional-dependent
electrochemical behaviors. To perform electrochemical studies,
the as-prepared Pd–Ru NP alloys were loaded onto carbon
black. Figure S9 represents a TEM image of the Pd Ru NPs/C,
À1
Figure 4. a) j/V curve in 0.1m HClO and 0.25m HCOOH solution at 50 mVs
4
on the Pd60Ru40 NPs/C, home-made Pd, and commercial Pd/C catalysts;
b) forward peak current density for formic acid oxidation as a function of
Pd/Ru composition with commercial Pd/C included (for detailed data see
Figure S10).
60
40
demonstrating no observable morphology change or aggrega-
tion. The catalytic activity for each type of the NP catalyst in-
cluding commercial Pd/C (10 wt.%, Aldrich) in 0.1m HClO with
4
0
.25m HCOOH is shown in Figure 4a and Figure S10 (normal-
commercial Pd/C. Pd Ru exhibits an area current density of
60
40
À2
ized by the mass of Pd). In typical cyclic voltammetry curves,
the current density in the forward scan first increases because
higher potential promotes the formic acid oxidation until the
NP’s surface is oxidized, resulting in a drop of current density.
In the backward scan, as the oxidized surface is reduced, the
catalyst regains its activity, leading to a current density
3.223 mAcm
(Figure S11), which is almost 4.9 times and
À2
5.1 times higher than that of Pd NPs (0.657 mAcm ) and com-
À2
mercial Pd/C (0.632 mAcm ), respectively. Such a result is also
consistent with the mass normalized result.
Pd NPs are known to deteriorate quickly in the oxidizing
[4a]
acidic environment. Such poor stability extremely blocks the
commercialization of fuel cells. Bimetallic NPs have been
[
2b]
jump. For a better comparison, Figure 4b illustrates a compo-
sitional dependence of the forward peak current density. The
catalytic activity of Pd–Ru NPs/C catalysts versus the Pd/Ru
composition reveals a “volcano”-type curve, which increases
with increasing Ru atomic percentage up to 40 at.% and then
decreases. However, all Pd–Ru samples demonstrate higher ox-
idation peak current densities (ranging from 0.209 to
[4c,26]
shown to be a possible solution.
Stability of the as-synthe-
sized NPs and commercial Pd/C was examined by the 1500 s
chronoamperometric test at a constant potential of 0.25 V in
0.1m HClO with 0.25m HCOOH. As shown in Figure 5, all Pd–
4
Ru NP alloys have higher start current density (at t=0 s, denot-
ed as j ) and steady current density (at t=1500 s, denoted as
0
À1
À1
0.612 mAmg ) than pure Pd (0.152 mAmg Pd) and commer-
j1500) than the Pd NPs and commercial Pd/C, suggesting
a better stability of bimetallic alloys than monometallic cata-
lysts. For a better comparison, the specific values of j and j
1500
Pd
À1
cial Pd/C (0.149 mAmg Pd), proving the superiority of bimetal-
lic alloy NPs. Pd Ru owns the largest peak current density of
60
40
0
À1
0.612 mAmg
at 0.23 V (Figure 4a), which is almost 4.0 times
are listed in Table S1. The value of j on the Pd–Ru catalysts
Pd
0
and 4.1 times higher than that of Pd NPs and commercial Pd/
C, respectively. In Figure S11, the catalytic activity normalized
by the corresponding electrochemical surface area (ECSA) are
additionally shown. The oxidation peak current densities give
the order of Pd Ru >Pd Ru >Pd Ru >Pd Ru >Pd>
has a similar “volcano”-type changing trend to that of forward
current density in Figure 4b, and Pd Ru owns the largest j0.
60
40
Nevertheless, j1500 increases with the increasing of Ru molar
fraction and Pd Ru has the largest j1500. Additionally, the pro-
20
80
portion of residual activity to initial activity after stability test
60
40
40
60
20
80
90
10
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
4
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Conclusions
Well-dispersed Pd–Ru alloy NPs over the whole composition
range are successfully synthesized by a facile one-pot atomic
diffusion process. For the first time, we can confine their sizes
at approximately 3–5 nm, and carefully discuss the correspond-
ing formation mechanism. The core-level binding energy of Pd
positively shifts owing to electron donation from Pd to Ru, re-
sulting in the downshift of the d-band center. As a conse-
quence, these Pd–Ru alloys are more active and stable than
commercial Pd/C towards formic acid oxidation reaction, and
their activities are composition-dependent, with Ru in
40 atom% exhibiting the optimum catalytic performance. This
Figure 5. j/t curves of as-synthesized NPs and commercial Pd/C at a constant
potential of 0.25 V in 0.1m HClO and 0.25m HCOOH solution.
4
research is of great importance in providing an effective cata-
lyst for fuel cells. Studies on the preparation of other alloy NPs
between two immiscible metals are underway.
(evaluated by j1500/j0) is calculated to give further information
related to stability. All Pd–Ru catalysts displayed smaller j1500/j0
than Pd NPs and commercial Pd black catalyst, suggesting
a better stability of bimetallic alloys than monometallic cata-
lysts. As Pd NPs tend to lose their ECSAs because of their ag-
gregation in an acid environment, Pd Ru with higher Pd con-
Experimental Section
Synthesis of Pd–Ru NP alloys
In a typical procedure for synthesis of Pd Ru NP alloy, aqueous
6
0
40
90
10
RuCl (3m, 0.1 mL) was added to a solution contain EG (15 mL) and
3
tent might also induce faster activity decay. However, with in-
PVP (Aldrich, Mw 55000, 0.1 g), which was preheated to its boiling
point (ꢀ1968C). As soon as the above mixture turned into brown
black (ꢀ40 s), aqueous K PdCl (8 mm, 3 mL) was introduced drop
creasing Ru mole fraction, j1500 increases and Pd Ru has the
20
80
largest j1500, indicating that alloying with Ru can greatly en-
hance the stability of Pd. Taken both activity and stability into
consideration, Pd Ru is the optimum catalysts for FAOR.
2
4
by drop. The dropwise adding of K PdCl solution took approxi-
2
4
60
40
mately 4–5 min. Then, the mixture was heated to reflux for another
1 h. After collection by centrifugation with acetone and washing
three times with ethanol, the collected black product was dis-
persed in ethanol. The Pd–Ru alloys with different composition
XPS analysis provides some evidence to explain the superior
catalytic performance of Pd–Ru alloy NPs. As we mentioned
above, the binding energies of Ru3p3/2 and Pd3d5/2 in Pd Ru40
60
ratios were prepared by varying the volume of RuCl solution and
3
alloy NPs have a negative shift of approximately À0.4 eV and
a positive shift of approximately +1.5 eV, respectively. Such
core-level shifts in binding energy have been correlated to
keeping the volume of K PdCl solution still.
2
4
[
29]
a downshift in the d-band center. And the ability of Pd in
binding adsorbates is directly determined by the d-band
center. That is, lowing the d-band center results in the de-
crease in the bond strength of adsorbates or in their chemi-
sorption energy. For FAOR, the rate of HCOOH directly transfor-
Characterization
XRD patterns were recorded with a Rigaku D/max-2500 X-ray
powder diffractometer with CuKa radiation. TEM and HAADF-STEM
images were recorded by a FEI Tecnai G2 F20 working at 200 kV.
ICP was performed on an Ultima 2 analyzer (Jobin Yyon). XPS was
performed on a ULVAC PHI Quantera microprobe. Electrochemical
measurements were conducted on Epsilon EC-2000 (BASi, America)
analyzer.
mation to CO is the highest at the surface that is free of (COO-
2
[
29c]
H)ads intermediates.
As Pd have a lower d-band center in the
Pd–Ru alloy, they bind the (COOH)ads intermediates less strong-
ly and reduce the surface (COOH)ads coverage. Therefore,
higher rate of the direct HCOOH decomposition process to
[
30]
CO2 are expected in Pd–Ru alloy. Kibler et al.
and Wang
[
31]
et al. have also demonstrated that a suitable downshift of
the d-band center of Pd increases the electrocatalysis of
HCOOH, yielding higher electrooxidation current at a given po-
tential and/or a negative electrooxidation onset potential.
Hence, the Pd–Ru alloys exhibit much better FAOR catalytic
performance than monometallic catalysts. Meanwhile, with-
drawing electrons from Pd can benefit Ru away from oxidation,
which further favors the promotion of FAOR rates and stability.
Electrochemical measurements
Firstly, an ethanol dispersion of NPs colloid was sonicated with
carbon black for 4 h to give Pd/C catalysts. The catalysts ethanol
suspension was deposited on a modified glassy carbon electrode
(3 mm in diameter). After the solvent was allowed to evaporate
atRT, Nafion (4 mL, 0.2 wt.% in ethanol) was coated on the surface
of the catalysts and dried at RT. Ag/AgCl electrode and Pt wire are
used as a reference electrode and a counter electrode, respectively.
All the electrolyte solutions were deaerated with high-purity nitro-
gen gas for 30 min. Both the mass of Pd (tested by ICP results) and
ECSA were used to normalize the original cyclic voltammograms.
All electrochemical experiments were performed at RT (22Æ18C).
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
5
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[
13] a) H. M. P. W. Hume-Rothery, Z. Kristallogr. 1935, 91, 23–47; b) D. Wu, Z.
Zheng, S. Gao, M. Cao, R. Cao, Phys. Chem. Chem. Phys. 2012, 14, 8051–
Acknowledgements
8
057.
We acknowledge the financial support from 973 Program
[
[
14] S. K. Ryi, A. W. Li, C. J. Lim, J. R. Grace, Int. J. Hydrogen Energy 2011, 36,
9335–9340.
15] K. Kusada, H. Kobayashi, R. Ikeda, Y. Kubota, M. Takata, S. Toh, T. Yama-
moto, S. Matsumura, N. Sumi, K. Sato, K. Nagaoka, H. Kitagawa, J. Am.
Chem. Soc. 2014, 136, 1864–1871.
(2011CB932504, and 2014CB845605), NSFC (21221001, 21203199,
21331006, and 21101155), Fujian Key Laboratory of Nanomateri-
als (2006L2005), China Postdoctoral Science Foundation Funded
Project (2012M521280, 2013T60642).
[
16] X. Chen, G. Wu, J. Chen, X. Chen, Z. Xie, X. Wang, J. Am. Chem. Soc.
2
011, 133, 3693–3695.
[
[
17] K. K. Yu, P. Meric, S. C. Tsang, Catal. Today 2006, 114, 428–433.
18] M. Liu, W. Yu, H. Liu, J. Zheng, J. Colloid Interface Sci. 1999, 214, 231–
237.
Keywords: electrochemistry · fuel cells · nanoparticles ·
palladium · ruthenium
[
19] a) J. Ohyama, T. Sato, Y. Yamamoto, S. Arai, A. Satsuma, J. Am. Chem.
Soc. 2013, 135, 8016–8021; b) N. Chakroune, G. Viau, S. Ammar, L. Poul,
D. Veautier, M. M. Chehimi, C. Mangeney, F. Villain, F. Fievet, Langmuir
[
1] a) X. W. Yu, P. G. Pickup, J. Power Sources 2008, 182, 124–132; b) N.
Yang, Y. Liu, H. Wen, Z. Tang, H. Zhao, Y. Li, D. Wang, ACS Nano 2013, 7,
2
005, 21, 6788–6796.
20] M. Zahmakiran, Y. Tonbul, S. Ozkar, J. Am. Chem. Soc. 2010, 132, 6541–
549.
[
1504–1512; c) J. Du, J. Qi, D. Wang, Z. Tang, Energy Environ. Sci. 2012,
6
5, 6914–6918.
[
[
21] P. Ptꢁ cˇ ek, Z. Bastl, Appl. Surf. Sci. 1990, 45, 319.
22] a) T. J. Sarapatka, J. Phys. Chem. 1993, 97, 11274; b) F. U. Hillebrecht, J. C.
Fuggle, P. A. Bennett, Z. Zolnierek, Ch. Freiburg, Phys. Rev. B 1983, 27,
[2] a) X. Xia, S. I. Choi, J. A. Herron, N. Lu, J. Scaranto, H. C. Peng, J. Wang,
M. Mavrikakis, M. J. Kim, Y. Xia, J. Am. Chem. Soc. 2013, 135, 15706–
15709; b) X. Wang, J. Yang, H. Yin, R. Song, Z. Tang, Adv. Mater. 2013,
2179.
25, 2728–2732; c) X. Lai, J. E. Halpert, D. Wang, Energy Environ. Sci.
[
[
23] B. D. Beamson G., High Resolution XPS of Organic Polymers: the Scienta
ESCA300 Database, 1992.
24] a) H. Wang, Y. Wang, Z. Zhu, A. Sapi, K. An, G. Kennedy, W. D. Michalak,
G. A. Somorjai, Nano Lett. 2013, 135, 16689–16696; b) S. M. Humphrey,
M. E. Grass, S. E. Habas, K. Niesz, G. A. Somorjai, T. D. Tilley, Nano Lett.
2012, 5, 5604–5620.
[3] a) Y.-J. Deng, N. Tian, Z.-Y. Zhou, R. Huang, Z.-L. Liu, J. Xiao, S.-G. Sun,
Chem. Sci. 2012, 3, 1157; b) Y. J. Tong, Chem. Soc. Rev. 2012, 41, 8195–
8209.
[
4] a) Y. Liu, M. Chi, V. Mazumder, K. L. More, S. Soled, J. D. Henao, S. Sun,
Chem. Mater. 2011, 23, 4199–4203; b) R. Ferrando, J. Jellinek, R. L. John-
ston, Chem. Rev. 2008, 108, 845–910; c) D. Wang, Y. Li, Adv. Mater. 2011,
2
007, 7, 785–790; c) S. Wang, L. Yi, J. Halpert, X. Lai, Y. Liu, H. Cao, R.
Yu, D. Wang, Y. Li, Small 2012, 8, 265–271.
[
25] a) W. Chen, R. Yu, L. Li, A. Wang, Q. Peng, Y. Li, Angew. Chem. Int. Ed.
23, 1044–1060; d) M. Shao, T. Huang, P. Liu, J. Zhang, K. Sasaki, M. B.
2
010, 49, 2917–2921; Angew. Chem. 2010, 122, 2979–2983; b) Y. Liu, C.
Vukmirovic, R. R. Adzic, Langmuir 2006, 22, 10409–10415; e) M. Wakisa-
ka, S. Mitsui, Y. Hirose, K. Kawashima, H. Uchida, M. Watanabe, J. Phys.
Chem. B 2006, 110, 23489–23496.
Wang, Y. Wei, L. Zhu, D. Li, J. S. Jiang, N. M. Markovic, V. R. Stamenkovic,
S. Sun, Nano Lett. 2011, 11, 1614–1617; c) J. Du, X. Lai, N. Yang, J. Zhai,
D. Kisailus, F. Su, D. Wang, L. Jiang, ACS Nano 2011, 5, 590–596.
[
[
[
[
[
5] M. H. Shao, K. Sasaki, R. R. Adzic, J. Am. Chem. Soc. 2006, 128, 3526–
[
[
26] H. You, S. Yang, B. Ding, H. Yang, Chem. Soc. Rev. 2013, 42, 2880–2904.
27] a) Y. Xia, Y. Xiong, B. Lim, S. E. Skrabalak, Angew. Chem. Int. Ed. 2009, 48,
3527.
6] D. L. Wang, H. L. Xin, H. S. Wang, Y. C. Yu, E. Rus, D. A. Muller, F. J. DiSal-
vo, H. D. Abruna, Chem. Mater. 2012, 24, 2274–2281.
7] C. Du, M. Chen, W. Wang, G. Yin, ACS Appl. Mater. Interfaces 2011, 3,
6
0–103; Angew. Chem. 2009, 121, 62–108; b) X. Xia, S. Xie, M. Liu, H. C.
Peng, N. Lu, J. Wang, M. J. Kim, Y. Xia, Proc. Natl. Acad. Sci. USA 2013,
10, 6669–6673.
28] S. Xiao, W. Hu, W. Luo, Y. Wu, X. Li, H. Deng, Eur. Phys. J. B 2006, 54,
79–484.
1
105–109.
[
[
8] Q. Gao, Y. M. Ju, D. An, M. R. Gao, C. H. Cui, J. W. Liu, H. P. Cong, S. H. Yu,
ChemSusChem 2013, 6, 1878–1882.
4
29] a) P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C. Yu, Z. Liu, S. Kaya,
9] a) J. S. Jirkovsk y´ , I. Panas, E. Ahlberg, M. Halasa, S. Romani, D. J. Schiffrin,
J. Am. Chem. Soc. 2011, 133, 19432–19441; b) F. Gao, D. W. Goodman,
Chem. Soc. Rev. 2012, 41, 8009–8020.
D. Nordlund, H. Ogasawara, M. F. Toney, A. Nilsson, Nat. Chem. 2010, 2,
454–460; b) V. Stamenkovic, B. Fowler, B. Mun, G. Wang, P. Ross, C.
Lucas, N. Markovi c´ , Science 2007, 315, 493–497; c) W. Zhou, A. Lewera,
R. Larsen, R. Masel, P. Bagus, A. Wieckowski, J. Phys. Chem. B 2006, 110,
[
[
10] S. Zhang, O. Metin, D. Su, S. Sun, Angew. Chem. Int. Ed. 2013, 52, 3681–
684; Angew. Chem. 2013, 125, 3769–3772.
11] a) H. Zhang, M. S. Jin, H. Y. Liu, J. G. Wang, M. J. Kim, D. R. Yang, Z. X. Xie,
J. Y. Liu, Y. N. Xia, ACS Nano 2011, 5, 8212–8222; b) Z. Zhu, H. Meng, W.
Liu, X. Liu, J. Gong, X. Qiu, L. Jiang, D. Wang, Z. Tang, Angew. Chem. Int.
Ed. 2011, 50, 1593–1596; Angew. Chem. 2011, 123, 1631–1634.
12] a) N. Alonso-Vante, ChemPhysChem 2010, 11, 2732–2744; b) P. Liu, J. T.
Muckerman, R. R. Adzic, J. Chem. Phys. 2006, 124, 141101; c) H. Ataee-Es-
fahani, J. Liu, M. Hu, N. Miyamoto, S. Tominaka, K. C. Wu, Y. Yamauchi,
Small 2013, 9, 1047–1051; d) A. L. Danberry, B. Du, I. S. Park, Y. E. Sung,
Y. Tong, J. Am. Chem. Soc. 2007, 129, 13806–13807.
3
1
3393–13398.
30] L. A. Kibler, A. M. El-Aziz, R. Hoyer, D. M. Kolb, Angew. Chem. Int. Ed.
005, 44, 2080–2084; Angew. Chem. 2005, 117, 2116–2120.
31] J. Wang, Y. Kang, H. Yang, W. Cai, J. Phys. Chem. C 2009, 113, 8366–
372.
[
[
2
8
[
Received: January 27, 2014
Revised: March 19, 2014
Published online on && &&, 0000
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
6
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
The presence of ruthenium: A simple
preparation method of Pd–Ru nanopar-
ticle alloys over the whole composition
range is presented. Their confined sizes
of 3–5 nm are favorable in direct formic
acid fuel cells (DFAFCs). The alloyed Ru
causes a downshift in d-band center
rendering these Pd–Ru alloys highly
active and stable towards formic acid
oxidation.
D. Wu, M. Cao, M. Shen, R. Cao*
& – &&
&
Sub-5 nm Pd–Ru Nanoparticle Alloys
as Efficient Catalysts for Formic Acid
Electrooxidation
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 7
&
7
&
ÞÞ
These are not the final page numbers!