A general approach to the size- and shape-controlled synthesis of platinum nanoparticles and their catalytic reduction of oxygen

Article abstract of DOI:10.1002/anie.200800073

Catalytic cobblestones: Monodisperse platinum nanoparticles were prepared with controlled sizes (3-7 nm) and shapes (polyhedron, truncated cube, or cube). The cubic nanoparticles are a much more active cathode catalyst for the oxygen reduction reaction: the current density J from 7 nm cubes is four times that of the other shapes (see picture), indicating great potential for fuel cell applications. (Graph Presented)

Full text of DOI:10.1002/anie.200800073

Communications
DOI: 10.1002/anie.200800073
O Reduction
2
A General Approach to the Size- and Shape-Controlled Synthesis of
Platinum Nanoparticles and Their Catalytic Reduction of Oxygen**
Chao Wang, Hideo Daimon, Taigo Onodera, Tetsunori Koda, and Shouheng Sun*
The synthesis of platinum nanoparticles (Pt NPs) with
controlled sizes and shapes is an attractive goal in developing
highly active platinum catalysts for the synthesis of fine
ditions. A rotating disk electrode (RDE) was used to
eliminate O2 diffusion near the electrode and to analyze
intrinsic catalytic ORR activity of the platinum catalysts in
0.5m H SO . The current density from the ORR for platinum
nanocubes is four times that of polyhedral platinum or the
truncated cubic Pt NP catalyst, indicating that the ORR
activity is indeed dependent on the shape, not on the size, of
the Pt NPs.
[
1]
chemicals. Well-dispersed Pt NPs are also an important
catalyst for fuel-cell reactions: in commonly studied polymer
electrolyte membrane fuel cells (PEMFCs), they catalyze
hydrogen oxidation at the anode and oxygen reduction at the
2
4
[
2]
cathode. In the oxygen reduction reaction (ORR), the
surface of the platinum catalyst tends to be shielded by the
In the synthesis (see the Experimental Section), the size
and the shape of the Pt NPs were controlled by reaction
temperature at which Fe(CO) was injected into the Pt(acac)
[3]
electrolyte or a hydroxy layer. As a result, the active sites on
the platinum surface are reduced in number, and the
efficiency of the catalyst is limited. To improve the ORR
activity, platinum catalysts should be either alloyed with
5
2
solution. For example, injecting Fe(CO) at 1808C gave 3 nm
5
polyhedral platinum, at 1608C yielded 5 nm truncated cubic
platinum, whereas at 1208C followed by controlled heating at
[
4]
various transition metals, or made with controlled shapes to
reduce the binding strength between platinum atoms and the
adsorbed species. Previous research has revealed that, in a
ꢀ
1
3–58Cmin to 2008C, 7 nm cubic platinum was produced.
Without Fe(CO) , Pt NPs were still prepared, but both the
5
H SO4 medium commonly used for PEMFCs, the ORR
size and shape of the particles were much less controlled.
Figure 1 shows representative transmission electron micro-
2
activity on Pt(100) is higher than that on Pt(111) owing to the
[5]
different adsorption rate of sulfates on these facets.
Although various methods have been developed to make
[
6]
Pt NPs, controlled synthesis of Pt NPs with a narrow shape
[7]
distribution has been rarely reported.
We recently reported that monodisperse platinum nano-
cubes could be made by reduction of platinum acetylaceto-
nate, Pt(acac) , in the presence of oleic acid, oleylamine, and a
2
[
8]
trace amount of iron pentacarbonyl, Fe(CO)5. Further
synthesis revealed that, by controlling the reaction temper-
ature, monodisperse Pt NPs were readily produced as poly-
hedrons, truncated cubes, or cubes. This work allowed
detailed studies in shape-dependent catalysis for the ORR
and for verification of a recent observation that fuel-cell
activity of the Pt NP catalyst is independent of the size of the
catalyst. Herein we report a general route to monodisperse
Pt NPs with sizes tunable from 3 nm to 7 nm, and controlled
polyhedral, truncated cubic, or cubic shapes, and the study of
their catalysis for the ORR under PEMFC reaction con-
Figure 1. Representative TEM images of a) the 3 nm polyhedral b) the
5
nm truncated cubic and c) the 7 nm cubic Pt NPs. The insets are the
representative HRTEM images of corresponding single particles, show-
ing a) Pt(111), b) Pt(100), and c) Pt(100) lattice fringes. All scale bars
in the insets correspond to 1 nm.
[
9]
scopy (TEM) images of the 3 nm, 5 nm, and 7 nm Pt NPs
(more TEM images are given in the Supporting Information,
Figure S1 and S2). Selected area electron diffraction of the
nanocube assembly (see the Supporting Information, Fig-
ure S1a) gives a pattern of bright spots arranged with four-
fold symmetry (Supporting Information, Figure S1b), indicat-
ing the formation of a (100) texture in the assembly.
Quantitative elemental analyses with a scanning electron
microscope (SEM) equipped with spatially resolved energy-
dispersive X-ray spectroscopy show that iron cannot be
detected on the surface of the Pt NPs, indicating that the trace
amount of iron added during the synthesis is only involved in
the formation of nuclei.
[
*] C. Wang, T. Koda, Prof. S. Sun
Department of Chemistry
Brown University
Providence, RI 02912 (USA)
Fax: (+1)401-863-9046
E-mail: ssun@brown.edu
H. Daimon, T. Onodera, T. Koda
Technology & Development Division
Hitachi Maxell Ltd.
6-20-1 Kinunodai, Tsukubamirai, Ibaraki 300-2496 (Japan)
[**] The work was supported by NSF/DMR 0606264, a Brown University
Research Seed Fund and a scholarship from Hitachi Maxell, Ltd.
The crystal structure of the Pt NPs was obtained by X-ray
diffraction (XRD). The diffraction patterns from different
NPs (Figure 2) indicate that the particles have the face-
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3
588
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3588 –3591

Angewandte
Chemie
Information, Figure S3). Pt NPs were obtained by taking
aliquots of the reaction solution at 1708C and 2008C during
the synthesis of 7 nm platinum nanocubes. The particles were
precipitated quickly by adding isopropanol to the octadecene
solution, and the product was separated by centrifugation and
redispersed in hexane. At 1708C, the size of the particles
reached 4–5 nm, but the shape was polyhedral/truncated
cubic. At 2008C, the NPs grew into 7 nm cubes owing to the
preferred growth on {111} over {100} facets.
To remove the organic coating around the NPs for ORR
studies, the Pt NPs were dispersed in hexane, then separated
by adding ethanol followed by centrifugation. The NPs were
washed by ethanol a further three times and irradiated by UV
[11]
light (185 nm and 254 nm) for 24 h in air. SEM images of
the 5 nm truncated cubic and 7 nm cubic Pt NPs after the
irradiation (see the Supporting Information, Figure S4) show
that the treatment does not lead to any obvious NP
morphology change.
The Pt NPs were prepared for electrochemical measure-
ments (see the Experimental Section). CVs were measured at
Figure 2. XRD patterns of a) the 3 nm polyhedral, b) the 5 nm trun-
cated cubic, c) the 3D randomly oriented 7 nm cubic and d) the (100)
textured 7 nm cubic Pt NP assemblies.
3
08 K with potential scanned from 0.03 V to 0.5 V versus the
normal hydrogen electrode, NHE, as shown in Figure 3. The
centered cubic (fcc) structure, and the NP sizes estimated
from Scherrerꢀs formula are consistent with that measured in
the respective TEM images, indicating the presence of single
crystals. For the 3 nm and 5 nm Pt NPs, the (111) peak is much
stronger than the (200) peak. This is expected for a three-
dimensional randomly oriented assembly. For the 7 nm
nanocubes, fast evaporation of the NP dispersion led to an
almost three-dimensionally randomly oriented assembly
(
Figure 2c), but slow evaporation of the solvent resulted in
a (100) textured assembly, as seen by the dominant (200) peak
in Figure 2d. This result confirms that the 7 nm nanocubes are
surrounded by {100} planes, and have a very narrow shape
distribution.
The synthesis reported herein indicates that the presence
of trace Fe(CO)5 facilitates fast platinum nucleation and
growth of Pt NPs with controlled size and shape. The
formation of platinum nuclei was seen by a color change of
the solution to black immediately after the injection of
Figure 3. CVs of the Pt NPs (* polyhedron, ~ truncated cube,
&
cube).
ꢀ1
The potentialwas app il ed with a scanning rate of 10 mVs
.
Fe(CO) at high temperatures (1608C and 1808C). If Fe(CO)5
5
was injected at 1208C, the solution was dark yellow, but
turned to black when it was heated to 1608C. The shape of the
nuclei is likely to be polyhedral, as it is known that in the
growth of fcc-structured nanocrystals, the nuclei formed in the
early stage of the synthesis tend to adopt a polyhedral shape
peak at about 0.22 V originates from the hydrogen adsorp-
tion/desorption on the Pt(100) face, and the increased peak
intensity reflects the fact that more (100) planes are present
[
5]
on the NP surface. This result also shows that the surfactants
have been successfully removed by the UV treatment;
without this treatment, only featureless CV curves were
obtained. It can be seen from Figure 3 that the 7 nm
nanocubes have more (100) planes exposed to the solution.
This is consistent with the TEM and XRD analyses, for which
it was found that the surface of the nanocubes is dominated by
(100) facets.
[
7,10]
to minimize their surface energy.
In the fast nucleation/
growth process at higher injection temperatures, most of the
platinum precursors are consumed, and there is no preference
for the growth in h100i and h111i directions. As a result, small
polyhedral Pt NPs are formed. Injection of Fe(CO)5 at a
lower temperature (1208C) followed by a slow rate of heating
facilitates a slow nucleation/growth process, which leads to
the thermodynamic growth along the h111i over the h100i
direction and the formation of larger cubes.
The catalytic activity of the Pt NPs for the ORR was
measured with an RDE (see the Experimental Section). The
ꢀ
2
current densities (J, mAcm ) were calculated by dividing the
measured electrode currents with the surface areas obtained
from the hydrogen desorption in Figure 3. Figure 4a shows
The growth process can be monitored by quenching the
reaction at different temperatures (see the Supporting
Angew. Chem. Int. Ed. 2008, 47, 3588 –3591ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3589

Communications
ca. 0.92 V for the 7 nm Pt NPs and ca. 0.87 V for the 3 nm and
nm Pt NPs. Around the half-wave potential (ca. 0.75 V), the
5
current density generated from the 7 nm Pt NPs is about four
times that of either 3 nm or 5 nm NPs. This catalytic
enhancement in the ORR for the Pt(100)-dominated nano-
cubes is consistent with what have been observed in Pt(hkl)-
dependent ORR activities from the extended electrode
[
5]
surfaces.
The NP shape-dependent ORR activity can be ascribed to
different adsorptions of sulfate ions on Pt (111) and (100)
facets. Sulfate ions tend to bind to the platinum surface atoms
in solution and impede the ORR activity. The trend of binding
strength is (111) @ (100) owing to the matching threefold
symmetry between three equivalent oxygen atoms in the
sulfate and platinum atoms on the (111) layer (see the
[
13]
Supporting Information, Figure S5). For the ORR, the case
is reversed, and the (100) facet has more active sites than
(
111) in H SO . The nanocubes have dominant (100) facets,
2 4
and therefore the high ORR activity. The similar low
reactivity observed from both 3 nm and 5 nm NPs is
consistent with a recent report on size independence of the
[9]
Pt NPs on fuel-cell activity. The shape-dependent catalysis
further proves that a trace amount of iron in the Pt NPs does
not contribute to the activity increase in the NP catalyst.
Rotation-rate-dependent current density–potential curves
of the 7 nm platinum nanocubes are given in Figure 4b
(similar curves from the 3 nm and 5 nm Pt NPs are given in
the Supporting Information, Figure S6). High rotational
speeds result in the increase in oxygen diffusion to electrode
surface and large currents. The measured disk electrode
current density (J) at different constant potentials can be
[
14]
applied into Levich–Koutecky plots [Equation (1)]:
1
J
1
1
1
1
B w
¼ J þ J ¼ J þ
K
,
=2
ð1Þ
1
K
diff
in which
2
0
=3
0
:62 n F C D
0
B ¼
h1=6
where J is the kinetic current density and J the diffusion-
k
diff
limiting current density, n is the overall number of electrons
transferred, F is the Faraday constant, C is the O concen-
0
2
ꢀ
3
ꢀ1
tration in the electrolyte (1.26 ” 10 molL ), D0 is the
diffusion coeficient of O in the H SO electrolyte (1.93 ”
2
2
4
ꢀ
5
2
1
0
cm s), and h is the viscosity of the electrolyte (1.009 ”
10 cm s). In the J versus w
Figure 4. a) Disk current densities in oxygen saturated 0.5m H SO as
ꢀ2
2
[5]
ꢀ1
ꢀ1/2
2
4
plot, the slope is 1/B. The
a function of potential the platinum catalysts (* polyhedron, ~ trun-
so-called B factor can be applied to obtain the number of
cated cube, cube) at a rotation speed of 1600 rpm. b) Rotation-rate-
&
ꢀ
1
electrons involved in the ORR. Figure 4c shows the J versus
dependent current–potentialcurves for 7 nm p al tinum nanocubes.
c) Koutecky–Levich plots at 0.6 V for the platinum catalysts (~ poly-
ꢀ
1/2
w
plots at 0.6 V for different platinum catalysts. The
hedron, * truncated cube,
scanning rate of 10 mVs
&
cube). The potentialwas app il ed with
number of electrons calculated from the slop is 3.6 for the
7 nm platinum nanocubes, and 0.7 for the 3 nm and 5 nm
Pt NPs. This indicates that there is a nearly complete
ꢀ1
.
reduction of O2 to H O, a four-electron process, on the
2
the electrode current densities of the ORR for the Pt NPs.
The overpotential for the ORR is reduced by nearly 50 mV by
the 7 nm platinum nanocubes compared with the other
Pt NPs, and can be seen from the onset ORR potentials at
cubic surface of the 7 nm platinum. However, the ORR is
much less efficient on the 3 nm and 5 nm Pt NP surfaces,
which is very likely due to strong sulfate adsorption on their
(111) facets.
3
590
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3588 –3591

Angewandte
Chemie
We have demonstrated a general synthetic approach to
monodisperse Pt NPs with controlled sizes (3–7 nm) and
shapes (polyhedron, truncated cube and cube) by reaction of
HClO or H PO owing to its stability and the superior performance
4 3 4
[
2]
in proton transport through the Nafion membrane in PEMFCs. CVs
were measured with nitrogen purging at 308 K, with the potential
scanned from 0.03 V to 0.5 V versus NHE, and the CVs were used to
calculate the surface area of the catalysts. ORR activities were
measured under oxygen purging in O -saturated 0.5m H SO at 308 K
Pt(acac) with oleic acid and oleylamine in the presence of
2
trace amount of Fe(CO) . The Pt NPs deposited on glassy
5
2
2
4
carbon electrodes are active catalysts for the ORR in H SO4
2
with an RDE (rotation speed 400–4500 rpm).
medium. The current density measured from the RDE for
7
nm platinum nanocubes is four times that of 3 nm poly-
Received: January 7, 2008
Revised: March 4, 2008
Published online: April 9, 2008
hedral (or 5 nm truncated cubic) Pt NPs, indicating a dom-
inant effect of NP shape on the ORR in PEMFC reaction
conditions. Work in preparing a highly efficient platinum
nanocube catalyst for applications in PEMFCs is underway.
Keywords: fuel cells · heterogeneous catalysis · nanoparticles ·
.
oxygen · platinum
Experimental Section
Typical synthesis of 3 nm polyhedral NPs: Pt(acac) (0.1 g), octade-
cene (10 mL), oleic acid (OA) (1 mL), and oleylamine (OAm) (1 mL)
were mixed under N and magnetic stirring. The mixture was then
heated to ca. 658C to dissolve Pt(acac) . The temperature was then
raised to about 1808C. A solution of Fe(CO)5 in hexane (0.1 mL,
prepared by adding 0.1 mL Fe(CO) in 1 mL hexane under argon) was
quickly injected into the hot solution. The solution was further heated
to 2008C and kept at this temperature for 1 hour before it was cooled
down to room temperature. 40 mL of isopropanol was added and then
the suspension was centrifuged (8000 rpm, ca. 10 min) to separate the
NPs. The particles were dispersed in ca. 10 mL hexane and precipi-
tated out by adding ethanol. The process was repeated one more time
to purify the NPs. The final product (ca. 50 mg) was dispersed in
[
1] a) K. Temple, F. Jäkle, J. B. Sheridan, I. Manners, J. Am. Chem.
Soc. 2001, 123, 1355; b) T. Shimada, I. Nakamura, Y. Yamamoto,
J. Am. Chem. Soc. 2004, 126, 10546; c) K. Godula, D. Sames,
Science 2006, 312, 67; d) K. M. Bratlie, H. Lee, K. Komvopoulos,
P. Yang, G. A. Somorjai, Nano Lett. 2007, 7, 3097.
2
2
2
[
2] N. P. Brandon, S. Skinner, B. C. H. Steele, Annu. Rev. Mater. Res.
5
2003, 33, 183.
[
[
3] N. M. Markovic, P. N. Ross, Surf. Sci. Rep. 2002, 45, 117.
4] a) T. Toda, H. Igarashi, H. Uchida, M. Watanabe, J. Electrochem.
Soc. 1999, 146, 3750; b) V. R. Stamenkovic, B. S. Mun, M. Arenz,
K. Mayrhofer, C. A. Lucas, G. Wang, P. N. Ross, N. M. Markovic,
Nat. Mater. 2007, 6, 241; c) H. Yano, M. Kataoka, H. Yamashita,
H. Uchida, M. Watanabe, Langmuir 2007, 23, 6438.
[
5] N. M. Markovic, H. A. Gasteiger, P. N. Ross, J. Phys. Chem.
1
0 mL of hexane for further use.
nm truncated-cubic and 7 nm cubic Pt NPs were synthesized by
1995, 99, 3411.
5
[
6] a) T. S. Ahmadi, Z. L. Wang, T. C. Green, A. Henglein, M. A.
EI-Sayed, Science 1996, 272, 1924; b) M. Yamada, S. Kon, M.
Miyake, Chem. Lett. 2005, 34, 1050; c) T. Herricks, J. Chen, Y.
Xia, Nano Lett. 2004, 4, 2367; d) J. Chen, T. Herricks, Y. Xia,
Angew. Chem. 2005, 117, 2645; Angew. Chem. Int. Ed. 2005, 44,
similar procedures to that described for the synthesis of 3 nm Pt NPs
except that Fe(CO) was injected at 1608C and 1208C, respectively.
Characterization: Samples for TEM analysis were prepared by
depositing one drop of diluted NP dispersion in hexane on amorphous
carbon-coated copper grids. Images were obtained by a Philips EM
5
2589; e) H. Lee, S. E. Habas, S. Kweskin, D. Butcher, G. A.
4
2
20 (120 kV). High-resolution TEM images were obtained on a JEOL
010 TEM instrument (200 kV). XRD patterns were obtained on a
Somorjai, P. Yang, Angew. Chem. 2006, 118, 7988; Angew. Chem.
Int. Ed. 2006, 45, 7824; f) J. Solla-Gullꢁn, V. Montiel, A. Aldaz, J.
Clavilier, J. Electroanal. Chem. 2000, 491, 69; g) X. Teng, H.
Yang, Nano Lett. 2005, 5, 885; h) B. Mayers, X. Jiang, D.
Sunderland, B. Cattle, Y. Xia, J. Am. Chem. Soc. 2003, 125,
Bruker AXS D8-Advanced diffractometer with CuKa radiation (l =
.5418 Š). SEM images were obtained on a Hitachi S-5200 scanning
1
electron microscope. Quantitative elemental analyses were carried
out on a LEO 1560 SEM equipped with spatially resolved energy-
dispersive X-ray spectroscopy (EDS). Samples for EDS were
prepared by depositing the NP dispersion on a silicon substrate and
solvent was allowed to evaporate under ambient conditions.
Electrochemical measurements were performed on an electro-
chemical analyzer, model HR-201 from Hokuto Denko Corp. The
treated NPs were dispersed in water by vigorous stirring and
13364; i) J. Chen, Y. Xiong, Y. Yin, Y. Xia, Small 2006, 2, 1340.
[
[
7] J. Ren, R. D. Tilley, J. Am. Chem. Soc. 2007, 129, 3287.
8] C. Wang, H. Daimon, Y. Lee, J. Kim, S. Sun, J. Am. Chem. Soc.
2007, 129, 6974.
[
9] a) H. Yano, J. Inukai, H. Uchida, M. Watanabe, P. K. Babu, T.
Kobayashi, J. H. Chung, E. Oldfield, A. Wieckowski, Phys.
Chem. Chem. Phys. 2006, 8, 4932; b) H. Yano, E. Higuchi, H.
Uchida, M. Watanabe, J. Phys. Chem. B 2006, 110, 16544.
ꢀ1
sonication. The concentration of NPs was adjusted to 2 mgmL
and 20 mL of the solution was dropped onto a glassy carbon electrode
GCE; 5 mm in diameter from Hokuto Denko Corp., Japan). After
,
[
10] W. Romanowski, Surf. Sci. 1969, 18, 373.
(
[
11] W. Chen, J. Kim, S. Sun, S. Chen, Phys. Chem. Chem. Phys. 2006,
water evaporation under a partial vacuum for 30 h, 10 mL of 0.1 wt%
Nafion solution was dropped on the electrode surface to cover and
stabilize the NP assembly on the electrode surface. The platinum-
loaded GCE was immersed into 0.5m H SO as a working electrode.
The electrochemically active surface area of the Pt NP catalyst was
determined by hydrogen desorption in the cyclic voltammogram
8
, 2779.
12] M. Watanabe, M. Tomikawa, S. Motoo, J. Electroanal. Chem.
985, 182, 193.
[
[
1
2
4
13] M. E. Gamboa-Aldeco, E. Herrero, P. S. Zelenay, A. Wieck-
owski, J. Electroanal. Chem. 1993, 348, 451.
[14] A. J. Bard, L. R. Faulkner, Electrochemical Methods: Funda-
mentals and Applications, Wiley, New York, 2000.
[
12]
(CV). Platinum wire was used as a counter electrode, and Ag/AgCl
as the reference electrode. 0.5m H SO was chosen as electrolyte over
2
4
Angew. Chem. Int. Ed. 2008, 47, 3588 –3591ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3591