Catalytic activity and regeneration property of a Pd nanoparticle encapsulated in a hollow porous carbon sphere for aerobic alcohol oxidation

Article abstract of DOI:10.1021/la102824s

A core-shell composite consisting of a palladium (Pd) nanoparticle and a hollow carbon shell (Pd@hmC) was employed as a catalyst for aerobic oxidation of various alcohols. The core-shell structure was synthesized by consecutive coatings of Pd nanoparticle

Full text of DOI:10.1021/la102824s

pubs.acs.org/Langmuir
©
2010 American Chemical Society
Catalytic Activity and Regeneration Property of a Pd Nanoparticle
Encapsulated in a Hollow Porous Carbon Sphere for Aerobic
Alcohol Oxidation
†
,†
†
‡
^
Takashi Harada, Shigeru Ikeda,* Fumihiro Hashimoto, Takao Sakata, Keita Ikeue,
§
†
Tsukasa Torimoto, and Michio Matsumura
†
Research Center for Solar Energy Chemistry, Osaka University, 1-3 Machikaneyama, Toyonaka 560-8531,
Japan, Research Center for Ultra-High Voltage Electron Microscopy, Osaka University, 7-1, Mihogaoka,
‡
^
Ibaraki 567-0047, Japan, Graduate School of Science and Technology, Kumamoto University, 2-39-1
Kurokami, Kumamoto 860-8555, Japan, and Department of Crystalline Materials Science, Graduate School of
Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8603, Japan
§
Received July 15, 2010. Revised Manuscript Received September 7, 2010
A core-shell composite consisting of a palladium (Pd) nanoparticle and a hollow carbon shell (Pd@hmC) was
employed as a catalyst for aerobic oxidation of various alcohols. The core-shell structure was synthesized by consecutive
coatings of Pd nanoparticles with siliceous and carbon layers followed by removal of the intermediate siliceous layer. Structural
characterizations using TEM and N adsorption-desorption measurements revealed that Pd@hmC thus-obtained was
2
composed of a Pd nanoparticle core of 3-6 nm in diameter and a hollow carbon shell with well-developed mesopore (ca. 2.5
nm in diameter) and micropore (ca. 0.4-0.8 nm in diameter) systems. When compared to some Pd-supported carbons,
Pd@hmC showed a high level of catalytic activity for oxidation of benzyl alcohol into benzaldehyde using atmospheric
pressure of O as an oxidant. The Pd@hmC composite also exhibited a high level of catalytic activity for aerobic oxidations of
2
other primary benzylic and allylic alcohols into corresponding aldehydes. The presence of a well-developed pore system in the
lateral carbon shell enabled efficient diffusion of both substrates and products to reach the central Pd nanoparticles, leading to
such high catalytic activities. This core-shell structure also provided high thermal stability of Pd nanoparticles toward
coalescence and/or aggregation due to the physical isolation of each Pd nanoparticle from neighboring particles by the carbon
shell: this specific property of Pd@hmC resulted in possible regeneration of catalytic activity for these aerobic oxidations by a
high-temperature heat treatment of the sample recovered after catalytic reactions.
Introduction
catalytic system, the use of heterogeneous catalysts is advantageous
because of their facile handling and reusability. Therefore, many
kinds of heterogeneous catalysts based on metal catalysts loaded
on various supports have been reported for the oxidation of
Liquid-phase oxidation of alcohols into the corresponding
carbonyl compounds and carboxylic acids is one of the funda-
mental and important reactions in organic synthesis. Stoichio-
metric inorganic oxidants, such as potassium permanganate
3
alcohols.
Among the various heterogeneous catalysts, Pd-supported cat-
alysts have attracted much attention due to their availability for
(
KMnO ) and potassium dichromate (K Cr O ), have been tradi-
4 2 2 7
1
tionally employed for this reaction. However, the stoichiometric
reagents have low atomic efficiency and usually produce some toxic
compounds. From economical and environmental points of view,
the system based on these stoichiometric inorganic oxidants should
be replaced by harmless and recyclable systems. Many researchers
have investigated the catalytic systems for this strategy. Metal salts
and complexes are often employed as catalysts in homogeneous
liquid-phase oxidation of alcohols with clean and inexpensive
4
oxidations and various other organic transformation reactions,
though the catalytic performance in most organic transformations
has been lower than that of homogeneous catalysts. In order to
achieve further improvement of catalytic activity for the oxidation
of alcohols, the synthesis of heterogeneous catalysts composed of
highly dispersed Pd nanoparticles stabilized on the support is
desirable. In the past several years, many researchers have reported
Pd nanoparticles supported on Al O , SiO , hydroxyapatite, acti-
2
2
3
2
oxidants (O and H O ). However, these homogeneous catalytic
2
2 2
vated carbon, and carbon nanotubes and have shown that these Pd-
nanoparticle-based heterogeneous catalysts were highly effective for
systems have difficulties in catalyst separation from the reaction
solution and repeated utilization. As another counterpart of the
5-9
alcohol oxidation reaction in liquid phase.
*
Corresponding author. Associate Professor Shigeru Ikeda. Telephone/
facsimile: þ81-6-6850-6696/6699, E-mail address: sikeda@chem.es.osaka-u.ac.jp.
1) Cainelli, G.; Cardillo, G. Chromium Oxidations in Organic Chemistry;
Springer-Verlag: Berlin, 1984.
2) (a) Nishimura, T.; Onoue, T.; Ohe, K.; Uemura, S. J. Org. Chem. 1999, 64,
750. (b) Sheldon, R. A.; Arends, I. W. C. E.; Dijksman, A. Catal. Today 2000, 57, 157.
c) Hallman, K.; Moberg, C. Adv. Synth. Catal. 2001, 3, 343. (d) Schultz, M. J.;
Sigman, M. S. Tetrahedron 2006, 62, 8227.
3) (a) Mallat, T.; Bodnar, Z.; Hug, P.; Baiker, A. J. Catal. 1995, 153, 131. (b)
(4) (a) Milne, J. E.; Buchwald, S. L. J. Am. Chem. Soc. 2004, 126, 13028. (b)
Shaughnessy, K. H.; DeVasher, R. B. Curr. Org. Chem. 2005, 9, 585. (c) Felpin, F.-X.; Ayad,
T.; Mitra, S. Eur. J. Org. Chem. 2006, 12, 2679. (d) Yin, L.; Liebscher, J. Chem. Rev. 2007,
107, 133.
(5) (a) Keresszegi, C.; Grunwaldt, J. D.; Mallat, T.; Baiker, A. J. Catal. 2004,
222, 268. (b) Wu, H. L.; Zhang, Q. H.; Wang, Y. Adv. Synth. Catal. 2005, 347, 1356.
(6) Jiao, L.; Regalbuto, J. R. J. Catal. 2008, 260, 329.
(7) (a) Mori, K.; Yamaguchi, K.; Hara, T.; Mizugaki, T.; Ebitani, K.; Kaneda,
K. J. Am. Chem. Soc. 2002, 124, 11572. (b) Mori, K.; Hara, T.; Mizugaki, T.; Ebitani,
K.; Kaneda, K. J. Am. Chem. Soc. 2004, 126, 10657.
(
(
6
(
(
Choudary, B. M.; Kantum, M. L.; Rahman, A.; Reddy, Ch. V.; Rao, K. K. Angew.
Chem., Int. Ed. 2001, 40, 763. (c) De Vos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A.
Chem. Rev. 2002, 102, 3615. (d) Yamaguchi, K.; Mizuno, N. Angew. Chem., Int. Ed.
(8) Harada, T.; Ikeda, S.; Miyazaki, M.; Sakata, T.; Mori, H.; Matsumura, M.
2
002, 41, 4538. (e) Abad, A.; Concepci ꢀo n, C.; Corma, A.; García, H. Angew. Chem.,
Int. Ed. 2005, 44, 4066. (f) Ng, Y. H.; Ikeda, S.; Harada, T.; Morita, Y.; Matsumura, M.
Chem. Commun. 2008, 3181.
J. Mol. Catal., A 2007, 268, 59.
(9) Villa, A.; Wang, D.; Dimitratos, N.; Su, D.; Trevisan, V.; Prati, L. Catal.
Today 2010, 150, 8.
1
7720 DOI: 10.1021/la102824s
Published on Web 10/12/2010
Langmuir 2010, 26(22), 17720–17725

Harada et al.
Article
Recently, core-shell materials have attracted attention as a
mesoporous SiO
2
with 10% aqueous HF generated Pd@hmC
novel catalytic structure, and several researchers have reported on
(caution: HF solution is very hazardous and corrosive and should
be handled according to the Material Safety Data Sheet guidelines).
As reference catalyst samples, we employed activated carbon
loaded with Pd nanoparticles (Pd/AC) and Pd nanoparticles
supported on hollow mesoporous carbon (Pd/hmC). The former
Pd/AC sample was purchased from N. E. Chemcat and consisted
of highly dispersed Pd nanoparticles with a narrow size distribu-
tion centered at 2.5 nm (Supporting Information Figure S1a) and
1
0
the fabrication and their catalytic functions. For example, Li
et al. reported the fabrication of the palladium core-porous silica
shell particles and their catalytic activity for the hydrogenation of
1
0b
4
-carboxybenzaldehyde.
However, due to the induction of
surface coverage with shell components, surface active site(s) of
central core particles should not be utilized efficiently.
To overcome this drawback, we proposed the fabrication of
2
-1
AC with a large surface area (ca. 1300 m g ) derived from its
well-developed micropore system. The latter Pd/hmC composite
was prepared by a conventional impregnation technique summar-
core-shell composites consisted of a Pt or Rh nanoparticle core
1
1
and a hollow porous carbon shell (i.e., Pt@hmC or Rh@hmC). In
these composites, the core metal nanoparticles are entrapped in a
medial void space surrounded by the carbon shell. These structural
features lead to significant tolerances to aggregation/coalescence
with neighboring nanoparticles. Moreover, since the void space
provides specific microenvironments suitable for catalytic reactions,
these composites were shown to act as a highly active catalyst for
hydrogenation of olefins and aromatic rings and for other reductive
reactions of functional groups in comparison with the activities of
3
ized as follows. To an aqueous solution (4 cm ) containing 9 μmol
of H PdCl was added 299 mg hollow mesoporous carbon
(
2
4
1
3
hmC). Then, the mixture was stirred at 80 ꢀC to dryness, and
the resulting powder sample was heated at 300 ꢀC in air for 1 h.
Finally, the powder containing a PdO component was treated at
3
00 ꢀC under H for 1 h. The amount of Pd loading in the Pd/hmC
2
sample was fixed at the same amount as that in the Pd@hmC
sample (see below).
Characterization of Materials. Transmission electron mi-
croscopy (TEM) images were taken on a Hitachi H-800 instru-
ment operated at 200 kV. X-ray absorption fine structures
(XAFS) of the Pd K-edge were recorded at the BL14B2 stations
with the approval of the Japan Synchrotron Radiation Research
Institute (JASRI) (Proposal No. 2008B1881) in Hyogo, Japan,
with ring energy of 8 GeV. The incident X-rays were monochro-
matized by a Si(111) double crystal monochromator. Pd K-edge
spectra were recorded at 25 ꢀC in the transmission mode. X-ray
absorption near-edge structure (XANES) spectra were normal-
ized for atomic absorption on the basis of the average absorption
coefficient. Analyses of surface area and pore structure were
carried out using a Quantachrome AUTOSORB-1 automated
11,12
several Pt and Rh supported catalysts.
On the basis of these facts and considerations, the core-shell
structure based on Pd nanoparticles (Pd@hmC) is highly promis-
ing for oxidation of alcohols in the liquid phase. In this report,
therefore, we describe the fabrication of Pd@hmC and its catalytic
activity for oxidation reaction of various alcohols. Regeneration
ability for the oxidation reaction derived from the specific struc-
tural feature of the Pd@hmC sample is also discussed.
Experimental Section
Fabrication of a Pd Nanoparticle Encapsulated in a
Hollow Porous Carbon Sphere. To a mixed solution of water
2
gas-sorption system, employing N as the adsorbate, after pre-
3
3
treatment of the sample at 200 ꢀC for 2 h under reduced pressure.
Amounts of CO irreversibly held on the surface of Pd particles
were measured by a pulse CO chemisorption experiment. Prior to
(
H
(
30 cm ) and ethylene glycol (70 cm ) containing 0.06 mmol of
PdCl was added 266 mg of poly(N-vinyl-2-pyrrolidone)
PVP). The solution was refluxed at 383 K for 1 h to yield a
2
4
3
2
measurement, the sample was flushed with H flow (30 cm
homogeneous dispersion of PVP-stabilized Pd nanoparticles (Pd-
PVPs). Then, 30 cm of the Pd-PVP solution was combined with
-1
3
min ) at 300 ꢀC for 1 h. After the temperature had been reduced
to room temperature, CO pulses were injected at constant inter-
vals. The amount of CO was determined by using a Shimadzu
GC-14B gas chromatograph equipped with an active carbon
column and a flame ionization detector (FID). Prior to introduc-
tion in an FID detector, CO molecules were converted into
methane using a methanizer. Quantitative analysis of the amount
of Pd was carried out by using a Perkin-Elmer OPTIMA 3000-
XL inductively coupled plasma (ICP) emission spectrometer. For
ICP analysis, Pd@hmC samples were immersed in aqua regia at
room temperature to dissolve the Pd particles. The undissolved
carbon particles were filtered by using a Millipore syringe-driven
membrane filter. The clear solution was then diluted to an
appropriate concentration before elemental analysis.
3
an ethanolic solution (270 cm ) containing 8.8 mmol tetraethyl
3
orthosilicate (TEOS) and 20 cm aqueous ammonia (28%), and the
mixture was stirred at room temperature for 6 h. After addition of
3
00 cm toluene, the mixture was centrifuged (4500 rpm, 15 min) to
6
recover a precipitate of silica-covered Pd nanoparticles (Pd@SiO
2
).
The Pd@SiO precipitate was stirred in a mixed solution containing
2
3
3
00 cm ethanol, 25.5 cm aqueous ammonia (28%), 17.7 mmol
3
TEOS, and 3.8 mmol n-octadecyltrimethoxysilane (ODTMS) at
room temperature for 2 h. The resulting suspension containing
octadecyl group-incorporated silica shell-coated Pd@SiO2 was
3
retrieved by centrifugation (4500 rpm, 15 min) with 600 cm of
toluene and calcined at 550 ꢀC for 5.5 h in air to produce Pd@SiO
coated with a mesoporous silica shell (Pd@m-SiO ). After pretreat-
ment at 300 ꢀC under H (5% in Ar) flow for 3 h to reduce the Pd
2
2
General Reaction Method for Alcohol Oxidation. Cataly-
tic oxidation of alcohols was performed in a reaction tube
equipped with a full-filled oxygen balloon. Prior to the reaction,
all of the catalyst powders were pretreated by heating at 300 ꢀC
2
nanoparticle core, the Pd@m-SiO
2
sample (210 mg) was stirred in
3
cm methanol containing 22.8 mg of resol-type phenol-formalde-
7
hyde resin (PF, supplied by Sumitomo Bakelite Co.) at room
temperature for 12 h to induce adsorption of PF into the porous
under H
2
flow for 1 h. A mixture of a substrate (0.25 mmol) and a
CO aqueous
solution (5 cm ) was put into the reaction tube. After purging
catalyst (5 mg, 0.15 μmol of Pd) in a 50 mM K
2
3
shell in the Pd@m-SiO
0 ꢀC, the remaining solid component was heated at 700 ꢀC under
evacuation for 5 h to carbonize PF. Finally, dissolution of SiO and
2
composite. After evaporation of solvents at
3
5
inside the reaction tube with O , the mixture was heated at 80 ꢀC
2
2
under the condition of vigorous stirring (3000 rpm) for the
prescribed time. Then, the mixture was quickly cooled to room
temperature and transferred into a sample tube. After separation
of the catalyst powder from the reaction solution by centrifuga-
tion (3500 rpm), a portion of the solution part was withdrawn
with a syringe and subjected to gas chromatographic analysis
(
10) (a) Mei, Y.; Lu, Y.; Polzer, F.; Ballauff, M. Chem. Mater. 2007, 19, 1062. (b)
Li, K. T.; Hsu, M. H.; Wang, I. Catal. Commun. 2008, 9, 2257. (c) Cargnello, M.;
Wieder, N. L.; Montini, T.; Gorte, R. J.; Fornasiero, P. J. Am. Chem. Soc. 2010, 132,
1
402.
11) (a) Ikeda, S.; Ishino, S.; Harada, T.; Okamoto, N.; Sakata, T.; Mori, H.;
Kuwabata, S.; Torimoto, T.; Matsumura, M. Angew. Chem., Int. Ed. 2006, 45,
063. (b) Harada, T.; Ikeda, S.; Ng, Y. H.; Sakata, T.; Mori, H.; Torimoto, T.; Matsumura,
M. Adv. Funct. Mater. 2008, 18, 2190.
12) Harada, T.; Ikeda, S.; Okamoto, N.; Ng, Y. H.; Higashida, S.; Torimoto,
(
7
2
(13) The hmC sample was prepared from SiO particles coated with a mesopor-
ous silica layer by the same procedure for fabrication of the carbon shell in the
(
T.; Matsumura, M. Chem. Lett. 2008, 37, 948.
Pd@hmC sample.
Langmuir 2010, 26(22), 17720–17725
DOI: 10.1021/la102824s 17721

Article
Harada et al.
(
inset of Figure 1a). The monodispersibility of Pd-PVP should be
significant to obtain the final Pd@hmC sample with a narrow size
distribution of core Pd particles. Surface coverage of Pd nano-
particles with a silica layer performed by polycondensation of
tetraethylorthosilicate (TEOS) in the presence of Pd-PVP yielded
a desired core-shell structure, i.e., silica sphere of ca. 60 nm in
diameter with centrally located Pd nanoparticles (Pd@SiO ), as
2
shown in Figure 1b. Subsequent polycondensation of TEOS
and n-octadecyltrimethoxysilane (ODTMS) in the presence of
Pd@SiO2 in ethanol solution followed by heat treatment at
5
50 ꢀC in air resulted in the formation of an irregularly shaped
layer entirely covering the surface of Pd@SiO (Figure 1c). The
2
fact that analysis of the N adsorption-desorption isotherm of
2
this sample showed the existence of mesopores with an average
2
pore size of 3.0 nm and relatively large surface area (530 m /g)
2
compared with that of Pd@SiO (78 m /g) indicates successful
2
formation of the desired mesoporous silica-covered Pd@SiO (i.
2
e., Pd@m-SiO2).
As determined by XANES measurement, the oxidation state of
Pd component in the Pd@SiO sample was divalent, namely, the
2
central Pd core transformed PdO during the formation of the
Pd@SiO structure (Supporting Information Figure S2b). Since
2
the PdO core could be reduced to the metallic state by the 300 ꢀC
heating under H (5% in Ar) as shown in Supporting Information
2
Figure S2c, the Pd@m-SiO2 sample was pre-reduced by the
treatment (see below); after that, the mesopore system of the
lateral silica shell was filled with a PF resin. The sample was then
heated at 700 ꢀC under evacuation to carbonize the resin
component, followed by immersion in an aqueous HF to dissolve
siliceous components. Figure 2a shows a typical TEM image of
the Pd@hmC sample thus-obtained. As we expected, the sample
seems to be a core-hollow shell structure, i.e., Pd nanoparticles
are encapsulated in a hollow carbon cage of ca. 100 nm in
diameter and ca. 25 nm in shell thickness. Successful encapsula-
tion of Pd nanoparticles in the shell was also confirmed by
comparison of the TEM image of Pd nanoparticles supported
on the surface of hmC (i.e., Pd/hmC) as shown in Figure 2b: in the
Pd/hmC sample, Pd nanoparticles with average particle size of ca.
6
.7 nm in diameter were mainly observed just above the carbon
wall.
It should be noted that Pd nanoparticles in the Pd@hmC
sample exhibited a narrow particle size distribution with an
average particle size of 4.6 nm (inset of Figure 2a), though the
size and size distribution are larger and broader than those of
original Pd-PVP (see Figure 1a). On the basis of the fact that the
melting point (MP) of PdO (870 ꢀC) is lower than that of Pd
(
1555 ꢀC) and that significant drops in MP are known to occur on
14
particles with nanosized regions, the appreciable growth of Pd
nanoparticles during the formation of the Pd@hmC structure was
likely due to partial melting of these nanoparticles in the carbo-
nization step (i.e., heating at 700 ꢀC under evacuation), leading to
coalescence between neighboring Pd particles in a hollow space.
Indeed, when carbonization was performed on the PF resin-
Figure 1. TEM images of (a) Pd-PVP, (b) Pd@SiO
2
Pd@m-SiO . Scale bars correspond to 30 nm. Inset of part a shows
the size distribution of Pd particles determined from TEM images
by measuring sizes of more than 200 particles.
2
, and (c)
using a Shimadzu GC-2010 gas chromatograph equipped with a
TC-FFAP capillary column and an FID or a mass-spectrometer.
adsorbed Pd@m-SiO sample without prereduction treatment at
2
3
00 ꢀC in H , a significant growth of Pd nanoparticles with sizes of
2
more than 7.2 nm in diameter occurred after the 700 ꢀC carbo-
nization (data not shown). Hence, the prereduction step seems to
beindispensable for inhibitingthecoalescence ofPd nanoparticles
as much as possible.
Results and Discussion
Fabrication of a Pd Nanoparticle Encapsulated in a
Hollow Mesoporous Carbon Sphere. Figure 1a shows a
TEM image of Pd nanoparticles protected with Poly(N-vinyl-
pyrrolidone) (Pd-PVP) used as a source material of the Pd core in
Pd@hmC. It is clear from the TEM image that the Pd nanopar-
ticles in Pd-PVP were highly monodispersed: there is a narrow
particle size distribution with an average diameter of 2.7 nm
The N adsorption-desorption isotherm of Pd@hmC showed
2
a typical type IV isotherm with a characteristic steep increase in
(
14) (a) Qi, W. H.; Wang, M. P. Mater. Chem. Phys. 2004, 88, 280. (b) Dick, K.;
Dhanasekaran, T.; Zhang, Z.; Meisel, D. J. Am. Chem. Soc. 2002, 124, 2312.
1
7722 DOI: 10.1021/la102824s
Langmuir 2010, 26(22), 17720–17725

Harada et al.
Article
Table 1. Oxidation of Benzyl Alcohol into Benzaldehyde by Various
Pd Catalysts
a
b
conv. /%
c
yield /%
d -1
TOF /h
entry
catalyst
1
2
3
4
a
Pd@hmC
Pd/AC
Pd/hmC
48
57
37
50
2940
2090
1528
-
e
e
26
<1
22
-
e
e
Pd@SiO
2
Reaction conditions: 0.15 μmol of Pd, 0.25 mmol of benzyl alcohol,
3
cm of 50 mM K
5
2
CO
3
aqueous solution, 80 ꢀC, 1 h, O atmosphere.
2
b
c
Conversion of benzyl alcohol. Yield of benzaldehyde. Turnover
d
frequency defined as the amount of benzaldehyde formed for 1 h per
total amount of surface Pd atoms. Reaction was carried out for 2 h.
e
the above TEM results. Moreover, it should be noted that
estimated Pd particle sizes in all the samples from the CO
chemisorption results were consistent with those determined from
the TEM observations, indicating that all of the Pd nanoparticles
in these catalysts should expose available surface sites for CO
sorption uniformly, irrespective of differences in their locations
and nature of support materials.
Catalytic Activity of Pd@hmC for Alcohol Oxidation
Reaction in Liquid Phase. For evaluation of the catalytic
activity of the Pd@hmC sample, oxidation of benzyl alcohol into
benzaldehyde was carried out in an aqueous K CO solution at
2
3
8
0 ꢀC under atmospheric pressure of O as a model reaction. For
2
comparison, the same reactionwas alsoperformed onPd/AC, Pd/
hmC, and Pd@m-SiO samples. The results are summarized in
2
Table 1. The Pd@hmC sample showed 48% conversion of benzyl
alcohol to yield 37% of benzaldehyde for 1 h reaction (entry 1).
Differences in these values are due to overoxidation to form
benzoic acid, a byproduct. Similar to these results, certain
amounts of benzoic acid were also observed in addition to the
formation of benzaldehyde when Pd/AC and Pd/hmC were used
as catalysts (entries 2 and 3). While the Pd@hmC sample
apparently shows lower conversion and yield than those of the
Pd/AC sample, superior catalytic ability of Pd@hmC compared
to Pd/AC can be revealed by using turnover frequency (TOF),
defined as the amount of benzaldehyde formed for 1 h per total
amount of surface Pd atoms, as a catalytic parameter that
represents efficiency of utilization of the surface reaction sites of
Pd nanoparticles. Efficient utilization of Pd surface sites in the
Pd@hmC sample was also proven when the TOF value of
Pd@hmC was compared to that of Pd/hmC. In the Pd@hmC
structure, the presence of a well-developed pore system in the
lateral carbon shell enables efficient diffusion of both the sub-
strate and products. Moreover, the Pd nanoparticle core in
Pd@hmC lies at a spacious void space with its naked surface
sites exposed. These structural characteristics of Pd@hmC re-
sulted in a large TOF in this reaction. In Pd/AC and Pd/hmC
samples, some of Pd nanoparticles are likely to be anchored at an
inner part of intricate micropore systems of AC and hmC
supports, leading to difficulties for efficient access/adsorption of
benzyl alcohol to these surface sites, whereas these micropores
barely allow penetration of relatively small CO molecules to be
accessed at the surface of Pd nanoparticles as discussed above. In
Figure 2. TEM images of (a) Pd@hmC and (b) Pd/hmC. Scale
bars correspond to 30 nm. Inset shows the size distribution of Pd
particles determined from TEM images by measuring sizes of more
than 200 particles.
N uptake at relative pressure (P/P ) below 0.1, indicating the
2
0
presence of mesopore and micropore systems (Supporting In-
formation Figure S3). Corresponding size distributions revealed
uniform pores centered at ca. 2.5 nm in diameter and a broad
micropore system of ca. 0.4-0.8 nm in diameter. Moreover, the
2
-1
specific surface area was calculated to be 1800 m g . These
results indicate the formation of a well-developed pore system in
the lateral carbon shell. In addition, as confirmed by FT-IR
analysis, central Pd nanoparticles in the Pd@hmC sample were
free from any organics (data not shown), suggesting complete
removal of PVP, i.e., exposure of the “naked” surface. All of the
results described above, therefore, seem consistent with our
expected Pd@hmC structure.
As estimated by the results for CO adsorption of the Pd@hmC
sample at room temperature, the amount of CO molecules
irreversibly held (CO ) on the surface of the Pd nanoparticle is
irr
6
.3 μmol per 1 gram of sample. ICP analysis of the sample
indicated the presence of Pd components with 0.32 wt % as Pd
metal. On the basis of these results, dispersion, defined as the ratio
addition, Pd@m-SiO , the intermediate of Pd@hmC, showed no
2
of COirr to the total amount of Pd atoms (CO /Pd), and particle
catalytic activity for this reaction (entry 4). As discussed above,
the Pd core surface in this sample was coated with the medial
nonporous silica layer, resulting in significant reduction of the
surface active site(s).
irr
sizes of Pd on the sample could beestimated to be 0.21 and 4.7 nm,
1
5
respectively. These values are moderate compared to those
obtained on Pd/AC (CO /Pd = 0.39, particle size: 2.5 nm) and
irr
Pd/hmC (CO /Pd = 0.12, particle size: 8.2 nm), as expected from
irr
The scope for aerobic oxidation using Pd@hmC is summarized
in Table 2 together with the results obtained for the Pt/AC
catalyst. Similar to the results for benzyl alcohol oxidation shown
in Table 1, the Pd@hmC catalyst showed high levels of catalytic
(
15) Chary, K. V. R.; Naresh, D.; Vishwanathan, V.; Sadakane, M.; Ueda, W.
Catal. Commun. 2007, 8, 471.
Langmuir 2010, 26(22), 17720–17725
DOI: 10.1021/la102824s 17723

Article
Harada et al.
Table 2. Aerobic Oxidation of Various Alcohols over Pd@hmC and
Pd/AC Catalysts
a
Figure 3. Relativeyield of benzaldehydeoverPd@hmC for oxida-
tion of benzyl alcohol after (a) 80 ꢀC desiccation and (b) 350 ꢀC
heat treatment normalized by the yield of benzaldehyde over fresh
catalyst.
Figure 4. Relative yield of benzaldehyde over Pd/AC for oxida-
tion of benzyl alcohol after (a) 80 ꢀC desiccation and (b) 350 ꢀC
heat treatment normalized by the yield of benzaldehyde over fresh
catalyst.
a
Reaction conditions: 0.15 μmol of Pd, 0.25 mmol of benzyl alcohol,
catalysts increased with increasing electron-donating abilities of
these substituents (entries 9-14), suggesting that the formation of
the transition state with a carbocationic character on the benzylic
carbon during the discharge of hydrogen in the rate-determining
step is involved in the oxidation pathway over the present
catalysts. In the same manner, the appreciable activity for the
cinnamyl alcohol oxidation (entries1 and 2), but not for oxidation
of 1-octanol (entries 5 and 6), can be explained by the effect of the
C-C double bond on stabilization of the transition state with the
carbocationic allylic carbon. These results strongly support the
above-mentioned reaction mechanism cycled on the surface of Pd
nanoparticles in Pd@hmC and Pd/AC catalysts.
3
cm of 50 mM K
5
2
CO
3
2
aqueous solution, 80 ꢀC, 1 h, O atmosphere.
b
Turnover frequency defined as the amount of benzaldehyde formed for
h per total amount of surface Pd atoms.
1
activity for oxidation of cinnamyl alcohol, a typical primary
-1
allylic alcohol, to obtain TOF of 560 h (entry 1). Although
oxidation of a secondary alcohol (i.e., 1-phenylethanol) was less
efficient, the activity level of Pd@hmC was still higher than that of
Pd/AC in terms of TOF (entries 3 and 4). In addition, Pd@hmC
did not show catalytic ability for oxidation of an aliphatic alcohol
(
i.e., 1-octanol), as was the case for the Pd/AC sample (entries 5
and 6).
Appreciable catalytic activities for oxidations of primary
Regeneration Ability of Pd@hmC for the Oxidation
Reaction of Benzyl Alcohol. Due to the ease of separation of
the Pd@hmC sample from the reaction mixture by using simple
centrifugation, reusable catalytic function is highly promising.
From this viewpoint, catalytic ability for benzyl alcohol oxidation
over the sample recovered after 1 h reaction at 80 ꢀC was examined.
For this purpose, the Pd@hmC sample was reused after two
different post-treatments: desiccation at 80 ꢀC and heat treatment
at 350 ꢀC in air. Figure 3 shows relative yield of benzaldehyde
formed, which is normalized by the yield of benzaldehyde over a
fresh Pd@hmC sample, on samples reused after these post-treat-
ments for two successive catalytic runs. While an appreciable drop
in catalytic activity was observed on the Pd@hmC sample reused
after desiccation at 80 ꢀC (Figure 3a), decrease in catalytic activity
was significantly suppressed on the Pd@hmC sample reused after
alcohols such as benzyl alcohol and cinnamyl alcohol compared
to those of the secondary alcohol of 1-phenylethanol as men-
tioned above have often been observed on other supported cata-
1
6-18
lysts such as Au/CeO , Ru/Al O , and Pt/porous-carbons.
2
2
3
Hence, the catalytic oxidation over the present Pd-based catalysts
is likely to proceed through a reaction pathway similar to that
over these supported catalysts, i.e., formation of metal-alkoxy
species, β-hydride elimination of them to form corresponding
carbonyl compounds, and oxidation of the remaining hydrides
over metal surfacesto complete catalytic cycles. Whenthis mecha-
nism is assumed, significant decreases in TOFs for the oxidation
of benzyl alcohol induced by deuterium substitution of benzylic
carbon over both catalysts (entries 7 and 8) indicate that the
cleavage of the C-H bond of benzylic carbon, which corresponds
to the β-hydride elimination step in the catalytic cycle, is the rate-
determining step of alcohol oxidation over the present Pd-based
catalysts. In addition, catalytic activities for oxidation of para-
substituent benzyl alcohols over both Pd@hmC and Pd/AC
350 ꢀC heat treatment (Figure 3b). On the other hand, when the
same experiments were performed using Pd/AC, the reused Pd/AC
samples after these post-treatments exhibited significant dec-
reases in their catalytic functions regardless of the post-treatment
(
3
Figure 4). The specific reusability of the Pd@hmC sample after the
50 ꢀC heat treatment is probably due to its characteristic
(
(
(
16) Abad, A.; Corma, A.; Garcia, H. Chem.;Eur. J. 2008, 14, 212.
17) Yamaguchi, K.; Mizuno, N. Chem.;Eur. J. 2003, 9, 4353.
18) Ng, Y. H.; Ikeda, S.; Morita, Y.; Harada, T.; Ikeue, K.; Matsumura, M.
structural or physicochemical properties that are different from
those of simple supported catalysts such as Pd/AC.
J. Phys. Chem. C 2009, 113, 12799.
1
7724 DOI: 10.1021/la102824s
Langmuir 2010, 26(22), 17720–17725

Harada et al.
Article
CO /Pd value (Figure 5b). Since Pd nanoparticles in Pd/AC
irr
were just attached on the surface of AC, they tend to move on the
surface to be aggregated by heating at a certain temperature
(
Supporting Information Figure S1b). Hence, high-temperature
regeneration was not effective for the Pd/AC sample. On the other
hand, aggregation of Pd particles was inhibited to a large extent in
the case of the Pd@hmC sample, because Pd nanoparticles in the
sample were located inside the carbon shell and were physically
separated from neighboring particles. These results, therefore,
indicate the superiority of the present core-shell structure for
effective prevention of loss of catalytic activity against high-
temperature processing.
Conclusion
We have demonstrated successful formation of a Pd nanopar-
ticle core-hollow porous carbon shell structure. Thus-obtained
Pd@hmC exhibited excellent catalytic activity for oxidation of
various primary benzylic and allylic alcohols. Mechanistic aspects
for alcohol oxidation revealedthatthe catalytic oxidation overthe
Pd@hmC sample proceeds essentially by the same reaction path
as that on conventional Pd-supported carbon such as Pd/AC.
Hence, the primary factor of the high catalytic activity of
Pd@hmC is its specific structure, i.e., naked Pd nanoparticles
encapsulated in a carbon shell with well-developed pore systems,
leading to efficient utilization of surface sites of Pd nanoparticles
compared to that in the case of other supported Pd catalysts.
Moreover, investigation of the durability of Pd@hmC revealed
that regeneration of catalytic activity could be realized by heat
treatment of the Pd@hmC sample at 350 ꢀC in air. Due to the
physical separation of Pd nanoparticles by a porous carbon shell
in the Pd@hmC structure, effective suppression of aggregation of
Pd nanoparticles could be achieved, leading to such specific
thermal stability. Thus, we clarified the significance of the
encapsulation of Pd nanoparticles in the porous carbon shell to
induce better catalytic functions for aerobic oxidation of alcohols.
In addition, further improvement of the activity can be expected if
the particle size of the Pd core is reduced. Studies along this line
are now in progress.
Figure 5. COirr/PdandPdparticlesizeof (a) Pd@hmCand(b) Pd/
AC samples before and after the catalytic run followed by two
different post-treatments.
Figure 5 shows changes in CO /Pd and Pd particle size of
irr
Pd@hmC and Pd/AC samples before and after catalytic runs
followed by two different post-treatments. Note that the former
CO /Pd was determined by CO adsorption and the latter Pd
irr
particle size was estimated from TEM observation. The CO /Pd
irr
value of the recovered Pd@hmC sample after desiccation at 80 ꢀC
was reduced to half that of the fresh sample despite the fact that
there was no change in Pd particle size of the sample; in contrast,
the CO /Pd value of the Pd@hmC sample recovered after the
irr
3
50 ꢀC heat treatment was ca. 90% of that of the fresh sample,
though the treatment induced a slight increase in Pd particle size
Supporting Information Figure S4). These results indicate that
(
there were substantial amounts of organic species strongly ad-
sorbed on the surface of Pd nanoparticles after the reaction and
that 350 ꢀC heat treatment would be required for removal of them
to expose active surfaces for catalytic oxidation. Similar to the
Pd@hmC sample, the Pd/AC sample recovered after desiccation
Acknowledgment. This research was supported by a Grant-in-
Aid for Young Scientists (B) (no. 22760597) from the Ministry of
Education, Culture, Sports, Science and Technology, Japan.
at 80 ꢀC showed lower CO /Pd than that of the fresh sample even
irr
though there was no appreciable change in Pd particle size as
shown in Figure 5b. These results also indicated the adsorption of
organic components on the surface of Pd nanoparticle. When the
Supporting Information Available: Additional TEM
images (Figure S1 and S3) and N adsorption-desorption
2
3
50 ꢀC heat treatment was applied to the recovered Pd/AC sample
isotherms and its corresponding BJH and H-K plots
(Figure S2). This material is available free of charge via the
Internet at http://pubs.acs.org.
in order to remove these surface adsorbates, appreciable growth
of Pd nanoparticles occurred, and this resulted in reduction of the
Langmuir 2010, 26(22), 17720–17725
DOI: 10.1021/la102824s 17725