Efficient electrocatalytic oxidation of formic acid using Au@Pt dendrimer-encapsulated nanoparticles

Article abstract of DOI:10.1021/ja4010305

We report electrocatalytic oxidation of formic acid using monometallic and bimetallic dendrimer-encapsulated nanoparticles (DENs). The results indicate that the Au₁₄₇@Pt DENs exhibit better electrocatalytic activity and low CO formation. Theore

Full text of DOI:10.1021/ja4010305

Communication
pubs.acs.org/JACS
Efficient Electrocatalytic Oxidation of Formic Acid Using Au@Pt
Dendrimer-Encapsulated Nanoparticles
Ravikumar Iyyamperumal,†,‡ Liang Zhang,†,‡,§ Graeme Henkelman,*,†,‡,§ and Richard M. Crooks*,†,‡
†Department of Chemistry and Biochemistry, ‡Texas Materials Institute, and §Institute for Computational and Engineering Sciences,
The University of Texas at Austin, 105 East 24th Street, Stop A5300, Austin, Texas 78712-1224, United States
S
* Supporting Information
The DENs in this study were synthesized, and their structure
characterized, using methods we previously reported.11−13
ABSTRACT: We report electrocatalytic oxidation of
formic acid using monometallic and bimetallic den-
drimer-encapsulated nanoparticles (DENs). The results
indicate that the Au147@Pt DENs exhibit better electro-
catalytic activity and low CO formation. Theoretical
calculations attribute the observed activity to the
deformation of nanoparticle structure, slow dehydration
of formic acid, and weak binding of CO on Au147@Pt
surface. Subsequent experiments confirmed the theoretical
predictions.
Briefly, Pt147 DENs were synthesized by adding 147 equiv of
K2PtCl4 to a 2.0 μM aqueous solution of a sixth-generation,
hydroxyl-terminated poly(amidoamine) dendrimer (G6-OH).
The Pt2+ was allowed to complex with the dendrimer for 3
days, and then a 10-fold excess of NaBH4 was added to reduce
the G6-OH(Pt2+)147 precursor to G6-OH(Pt147) DENs. A
similar procedure was used for the synthesis of Au147 DENs,
except the reduction was initiated almost immediately after
addition of HAuCl4 to avoid competitive reduction by the
hydroxyl groups of the dendrimers. Pt and Au DEN-modified
electrodes were prepared by sonicating an aqueous solution,
containing 20% isopropyl alcohol, with 2 mg mL−1 of Vulcan
EC-72R carbon for 5 min, and then drop-casting 4.0 μL of the
resulting ink onto a freshly polished glassy carbon electrode
(GCE). The electrode was then dried at room temperature.
Following a method pioneered by Adzic and co-workers,14 and
our previous experimental procedure,12 Au147@Pt DENs were
synthesized by electrochemical underpotential deposition
(UPD) of Pb onto Au147 DENs, followed by galvanic exchange
with PtCl42−. Specifically, Au147 DENs were immobilized on a
GCE, and then a monolayer of Pb was added to the Au core by
UPD from a N2-saturated aqueous solution containing 1.0 mM
Pb(NO3)2 and 0.10 M HClO4. Next, 10.0 mL of a 1.0 mM
K2PtCl4 solution was added to initiate galvanic exchange of Pb
for Pt. The resulting DENs were then characterized by UV−vis
spectroscopy, transmission electron microscopy, and electro-
chemical methods (Supporting Information) and found to be
consistent with our previous findings.12
n this Communication we show that electrochemically
Isynthesized nanoparticles consisting of a 147-atom Au core
and a Pt shell (Au147@Pt) facilitate the direct oxidation of
formic acid to carbon dioxide. Specifically, experiments and
first-principles calculations show that ∼2-nm-diameter Au147
@
Pt dendrimer-encapsulated nanoparticles (DENs)1 undergo a
structural distortion that suppresses formation of adsorbed CO
(COads) compared to Pt-only DENs.2 This finding is significant,
because it represents an unusual case in which a well-defined
structural deformation of a very small nanoparticle results in a
dramatic improvement in catalytic performance.
Formic acid is used as a substitute fuel for hydrogen in
proton exchange membrane fuel cells (PEMFCs), because
oxidation of hydrogen and formic acid occurs at similar
thermodynamic potentials.3 Electro-oxidation of HCOOH to
CO2 on Pt electrocatalysts occurs via two mechanisms: (1)
through formation of a reactive intermediate (direct oxidation
pathway), and (2) through formation of COads and subsequent
oxidation of COads to CO2 (indirect oxidation pathway).4 The
second pathway is problematic, because COads poisons the Pt
surface. However, it has been shown that the presence of Au
can improve the CO tolerance of Pt, and hence its catalytic
performance.5−7 For example, Xu and co-workers found that
Au@Pt nanoparticles having submonolayer Pt shells exhibit
improved electrocatalytic activity for formic acid oxidation due
to specifically engineered electronic interactions between the
two metals.8 An alternative to improving CO tolerance is to
simply avoid CO altogether. For example, Masel and co-
workers found that carbon-supported Pd is highly active for
formic acid oxidation, because it favors the direct oxidation
pathway.9 However, Pd/C suffers significant loss in activity
during formic acid oxidation due to slow adsorption of CO-like
intermediates, bridge-bonded formate, perchlorate, and other
anions.10
Figure 1 shows cyclic voltammograms (CVs) for formic acid
oxidation using Au147, Pt147, and Au147@Pt DEN-modified
GCEs in 0.10 M HClO4 containing 0.10 M HCOOH. The
Au147 DENs do not exhibit detectable activity for formic acid
oxidation. In contrast, the Pt147 DEN-modified electrode reveals
two peaks at −0.12 and 0.21 V (vs Hg/Hg2SO4) in the scan
toward more positive potentials. These are attributable to the
direct oxidation of HCOOH to CO2 and oxidation of COads,
generated by dehydration of HCOOH, respectively.15 The low
d
ratio (2.8) of the current densities (jp /jpind, Table 1) for these
two peaks indicates a substantial degree of indirect electro-
catalytic oxidation of formic acid on Pt147 DENs.16 Upon scan
reversal, a single peak is observed at −0.24 V, which
corresponds to direct oxidation of HCOOH to CO2. With
Received: January 29, 2013
Published: April 8, 2013
© 2013 American Chemical Society
5521
dx.doi.org/10.1021/ja4010305 | J. Am. Chem. Soc. 2013, 135, 5521−5524

Journal of the American Chemical Society
Communication
Table 1 and Figure 1 are normalized to the electrochemically
active surface areas of Pt). This level of catalytic performance is
also substantially better than that observed using Pd black on
carbon (Table 1).18 The increased activity of the Au147@Pt
DENs results from the desirable electronic properties of the
core@shell structure. Fourth, the peak current for the forward
(positive-going) scan for the Au147@Pt DENs is higher than for
the reverse scan, which is the opposite of what is observed for
the Pt147 DENs and for a bulk polycrystalline electrode (Figure
S5). This is a consequence of a decrease in CO poisoning at the
more negative potentials and differences in the rate of oxide
formation and subsequent reduction at more positive
potentials.
As mentioned previously, one of the most distinctive features
of formic acid oxidation on Au147@Pt DENs is near-elimination
of the indirect formic acid oxidation pathway. We hypothesized
that the reduction in CO poisoning could arise from either of
the following phenomena: (1) slowed dehydration of
HCOOH, which eliminates formation of CO, or (2) weakening
of the CO binding energy. With regard to the first hypothesis,
Yancey et al. proposed a deformed partial shell (dps) model for
Au147@Pt DENs.12 These materials, which we denote as Au@
Ptdps, are the same as those used in the present study (Figure
2).12 In this model, a surface distortion was observed during
Figure 1. CVs for formic acid oxidation of Au147, Pt147, and Au147@Pt
DEN-modified GCEs in a N2-saturated aqueous electrolyte solution
containing 0.10 M HClO4 and 0.10 M HCOOH. The scan rate was 50
mV/s, and the geometric area of the GCE was 0.071 cm2. For the Pt147
and Au147@Pt DENs, the current axis is normalized to the
electrochemically active surface area of Pt as measured by H UPD.
For the Au147 DENs, the current is normalized to the electrochemically
active surface area of Au as measured by reduction of the Au oxide
peak.
the exception of the peak potential on the reverse scan, which is
shifted positive by about 60 mV, these peak potentials are in
nearly the same positions as those obtained using a macro-
scopic polycrystalline Pt electrode under the same experimental
conditions (Supporting Information).15
The Au147@Pt DEN electrocatalysts display superior catalytic
performance compared to the Pt147 DENs and a bulk,
polycrystalline Pt electrode. This is demonstrated by the red
CV shown in Figure 1. There are four striking differences
between this CV and the Pt-only CVs (Pt147 DENs and the
bulk, polycrystalline Pt electrode, Figure S5). First, the peak at
∼0.2 V, which corresponds to oxidation of COads, is nearly
absent on the Au147@Pt DENs. The fact that very little COads
on the bimetallic DENs indicates that formic acid is oxidized
almost exclusively via the direct pathway on this electrocatalyst.
Second, the Au147@Pt electrocatalyst exhibits an onset potential
(defined here as the potential at which the current density =
0.10 mA/cm2)17 for formic acid oxidation of −0.54 V, which is
0.14 and 0.20 V lower than for Pt147 DENs and a polycrystalline
Pt electrode, respectively. Third, the maximum current density
Figure 2. Optimized structures for (a) the Pt-only DEN model (Pt147),
(b) the complete shell model (Au@Ptcs), which is equivalent to Au55@
Pt92, and (c) the deformed partial-shell model (Au@Ptdps), which is
equivalent to Au147@Pt102. The superimposed numbers represent
possible CO binding sites for each model.
molecular dynamics simulations in which the Pt-covered (100)
facets sheared into a diamond (111) structure. This distortion
may play an important role in the rate of dehydration of
HCOOH to CO on the DENs, because formation of CO
during formic acid oxidation on Pt(111) single-crystal electro-
des is not favorable.19,20
The second hypothesis was tested using DFT to calculate the
CO binding energies on Pt147, the Au@Ptdps model, and a
complete shell model of Au@Pt (Au@Ptcs). These three
structures, showing possible CO binding sites, are provided in
d
(jp ) for direct formic acid oxidation is substantially higher for
the Au147@Pt DENs compared to either Pt147 DENs or the
bulk, polycrystalline Pt electrode (note that the currents in
Table 1. Electrochemical Parameters for Formic Acid Oxidation Derived from Figure 1
b
c
peak potential (V)
current density (mA/cm2)
a
d
ind
d
ind
d
ind
onset potential (V)
Ep
Ep
jp
jp
ratio jp /jp
Pt (polycrystalline)
G6-OH(Pt147
−0.34
−0.40
−0.54
−0.44
−0.13
−0.12
−0.20
0.06
0.15
0.21
0.20
0.90
0.36
1.91
1.35
0.52
0.13
0.07
1.7
2.8
)
G6-OH(Au147@Pt)
Pd black18
27.3
a
b
The onset potentials are defined as the potential corresponding to formic acid oxidation at a current density of 0.10 mA/cm2.17 Epd and Epind are
c
d
ind
the potentials corresponding to the maximum currents for the positive-going scans in Figure 1. jp and jp are the peak current densities
corresponding to Epd and Epind, respectively. These values are normalized to the true area of Pt determined by hydrogen UPD. The current density is
also normalized for the true surface area of Pd.
5522
dx.doi.org/10.1021/ja4010305 | J. Am. Chem. Soc. 2013, 135, 5521−5524

Journal of the American Chemical Society
Communication
Table 2. CO Binding Energies for Different Sites on the Pt147, Au55@Pt92, and Au147@Pt102 Model
Pt-only model (Pt147
)
complete shell (Au@Ptcs) model (Au55@Pt92)
deformed partial shell (Au@Ptdps) model (Au147@Pt102)
a
binding site
BE_CO (eV)
binding site
BE_CO (eV)
binding site
BE_CO (eV)
1
2
3
4
(100)
(111)
corner
edge
−3.12
−2.65
−3.30
−3.15
1
2
3
4
(100)
(111)
corner
edge
−3.34
−3.07
−3.36
−3.40
1
2
3
4
5
6
D(111) center
D(111) corner
D(111) edge
T(111) corner
T(111) edge
Au
−1.74
−1.96
−1.80
−1.44
−1.37
−0.47
a
D(111) and T(111) denote the diamond (111) and triangle (111) facets, respectively.
Figure 2, and the corresponding CO binding energies are listed
in Table 2. The results indicate that Au@Ptcs binds CO more
strongly than Pt147, which is in agreement with results from
previous DFT calculations on bulk model surfaces and with
temperature-programmed desorption experiments.21 Impor-
tantly, however, the binding energy calculations reveal
significant weakening of CO binding to Au@Ptdps. These CO
binding trends can be interpreted in terms of the average Pt−Pt
experiment, the DEN-modified electrodes were each immersed
in a solution containing 0.10 M HClO4 and 0.10 M HCOOH
for 5 min at open-circuit potential (OCP). This results in
chemisorption of CO arising from dehydration of
HCOOH.19,22 The electrode was then removed from the
formic acid solution, washed, and placed in a fresh solution
containing only N2-purged 0.10 M HClO4, and then CVs were
obtained. In the case of Pt147 DENs (Figure 3a), a broad peak is
observed at ∼0.10 V during the first scan toward positive
potentials. We attribute this peak to oxidation of COads.
However, no corresponding peak for oxidation of COads is
observed for the Au147@Pt DENs. These results indicate that
dehydration of HCOOH to COads is slower on Au147@Pt
DENs than Pt147 DENs surfaces, which is consistent with the
calculations alluded to earlier.
bond lengths in the different models: Au@Ptcs, 2.79 Å; Pt147
,
2.76 Å; and Au@Ptdps, 2.65 Å. The compression of the Pt−Pt
bond in Au@Ptdps leads to weaker binding of CO on the Pt
surface compared to either Pt147 or Au@Ptcs DENs.21
On the basis of the foregoing discussion, we conclude that
weakened CO binding and slow dehydration of HCOOH on
the Au147@Pt surface account for reduced CO poisoning and
the enhanced activity of this catalyst for formic acid oxidation.
To test these conclusions experimentally, two different CO
oxidation experiments were performed using GCEs modified
with Pt147 and Au147@Pt DENs (Figure 3). For the first
The second experiment was carried out by immersing the
DEN-modified GCEs in a CO-saturated aqueous 0.10 M
HClO4 electrolyte solution for 10 min at −0.60 V (rather than
at the OCP). These conditions result in formation of a strongly
adsorbed monolayer of CO on the surface of the DENs. The
electrolyte solution was then purged with N2 for 10 min, while
maintaining the electrode potential at −0.60 V, to remove
dissolved CO. Finally, CVs were obtained. For the Pt147
-
modified electrode, a large oxidation wave attributable to CO
oxidation is observed at ∼0.20 V (Figure 3c). A similar peak,
but shifted negative by about 100 mV, is observed for the
Au147@Pt DENs (Figure 3d). This shift indicates that CO
adsorbs more weakly to the bimetallic DENs, which, again, is
consistent with the binding energies obtained from DFT
calculations. Taken together, the experiments summarized in
Figure 3 and the DFT calculations strongly suggest that the
high catalytic activity and small extent of CO poisoning,
revealed by the CVs in Figure 1, arise from both slow
dehydration of formic acid and weak binding of CO on the
Au147@Pt DENs.
In conclusion, we have shown that Au147@Pt DENs exhibit
better electrocatalytic activity for HCOOH oxidation reaction
than Pt147 DENs, a bulk, polycrystalline Pt electrode, and
carbon-supported Pd black. DFT calculations provide an
explanation for these findings that is fully consistent with the
experiments. Specifically, the Au147@Pt DENs undergo a
structural transition (Figure 2) that leads to both slower
dehydration of formic acid and weaker binding of COads.
Perhaps the most significant outcome of this work, however, is
that there appear to be major structural differences between
particles in the 1−2 nm size range of DENs, and the more
commonly studied (and more experimentally tractable) nano-
particles having sizes >3 nm. This is likely due to the lower
stability of the smaller particles, and thus their enhanced ability
to deform in the presence of adsorbates and during catalysis. In
forthcoming publications we will show that this is a rather
Figure 3. The red CVs show oxidation of COads on Pt147- and Au147
@
Pt-modified GCEs (peaks indicated by “1”). (a,b) The DEN-modified
electrodes were exposed to an aqueous solution containing 0.10 M
HClO4 and 0.10 M HCOOH for 5 min at the open circuit potential
(OCP), rinsed in water, and then placed in a HCOOH-free, 0.10 M
HClO4 solution for electrochemical analysis. (c,d) The DEN-modified
electrodes were immersed in a CO-saturated, aqueous 0.10 M HClO4
electrolyte solution and held at a potential of −0.60 V for 10 min.
Without changing the electrode potential, the solution was then
purged with N2, and the CVs were recorded. The black curves are
control experiments in which the electrodes were not exposed to CO
or HCOOH. The scan rate was 50 mV/s, and the geometric area of
the GCE was 0.071 cm2.
5523
dx.doi.org/10.1021/ja4010305 | J. Am. Chem. Soc. 2013, 135, 5521−5524

Journal of the American Chemical Society
Communication
(20) Rhee, C. K.; Kim, B.-J.; Ham, C.; Kim, Y.-J.; Song, K.; Kwon, K.
Langmuir 2009, 25, 7140−7147.
general result that deserves more attention than it has thus far
received.
(21) Friedrich, K. A.; Henglein, F.; Stimming, U.; Unkauf, W.
Electrochim. Acta 2000, 45, 3283−3293.
ASSOCIATED CONTENT
■
(22) Sun, S. G.; Clavilier, J. J. Electroanal. Chem. Interfacial
Electrochem. 1987, 236, 95−112.
S
* Supporting Information
Detailed experimental procedures, TEM images and particle-
size distributions, details of the DFT calculations, CVs for Pt147
and Au147@Pt DENs before and after formic acid oxidation,
and CV of a bulk, polycrystalline Pt electrode in the presence
and absence of formic acid. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
crooks@cm.utexas.edu; henkelman@cm.utexas.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully acknowledge support from the Chemical
Sciences, Geosciences, and Biosciences Division, Office of
Basic Energy Sciences, Office of Science, U.S. Department of
Energy (Contract DE-FG02-09ER16090). R.M.C. also thanks
the Robert A. Welch Foundation (Grant F-0032) for sustained
support. The computational work was done primarily at the
National Energy Research Scientific Computing Center and the
Texas Advanced Computing Center. Finally, we thank Mr.
David Yancey (UT-Austin) and Prof. Carol Korzeniewski
(Texas Tech University) for helpful discussions.
REFERENCES
■
(1) Myers, V. S.; Weir, M. G.; Carino, E. V.; Yancey, D. F.; Pande, S.;
Crooks, R. M. Chem. Sci. 2011, 2, 1632−1646.
(2) Ye, H.; Crooks, J. A.; Crooks, R. M. Langmuir 2007, 23, 11901−
11906.
(3) Weber, M.; Wang, J.-T.; Wasmus, S.; Savinell, R. F. J. Electrochem.
Soc. 1996, 143, L158−L160.
(4) Capon, A.; Parson, R. J. Electroanal. Chem. Interfacial Electrochem.
1973, 44, 1−7.
(5) Kristian, N.; Yan, Y.; Wang, X. Chem. Commun. 2008, 353−355.
(6) Park, I.-S.; Lee, K.-S.; Choi, J.-H.; Park, H.-Y.; Sung, Y.-E. J. Phys.
Chem. C 2007, 111, 19126−19133.
(7) Zhang, S.; Shao, Y.; Liao, H.-g.; Liu, J.; Aksay, I. A.; Yin, G.; Lin,
Y. Chem. Mater. 2011, 23, 1079−1081.
(8) Zhang, G.-R.; Zhao, D.; Feng, Y.-Y.; Zhang, B.; Su, D. S.; Liu, G.;
Xu, B.-Q. ACS Nano 2012, 6, 2226−2236.
(9) Rice, C.; Ha, S.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.;
Barnard, T. J. Power Sources 2002, 111, 83−89.
(10) Yu, X.; Pickup, P. G. Electrochem. Commun. 2009, 11, 2012−
2014.
(11) Kim, Y.-G.; Oh, S.-K.; Crooks, R. M. Chem. Mater. 2003, 16,
167−172.
(12) Yancey, D. F.; Zhang, L.; Crooks, R. M.; Henkelman, G. Chem.
Sci. 2012, 3, 1033−1040.
(13) Ye, H.; Crooks, R. M. J. Am. Chem. Soc. 2005, 127, 4930−4934.
(14) Sasaki, K.; Wang, J. X.; Naohara, H.; Marinkovic, N.; More, K.;
Inada, H.; Adzic, R. R. Electrochim. Acta 2010, 55, 2645−2652.
(15) Parsons, R.; VanderNoot, T. J. Electroanal. Chem. Interfacial
Electrochem. 1988, 257, 9−45.
(16) Chen, W.; Xu, L.-P.; Chen, S. J. Electroanal. Chem. 2009, 631,
36−42.
(17) Kanan, M. W.; Nocera, D. G. Science 2008, 321, 1072−1075.
(18) Haan, J. L.; Masel, R. I. Electrochim. Acta 2009, 54, 4073−4078.
(19) Macia,
47, 3653−3661.
́
M. D.; Herrero, E.; Feliu, J. M. Electrochim. Acta 2002,
5524
dx.doi.org/10.1021/ja4010305 | J. Am. Chem. Soc. 2013, 135, 5521−5524