Incarceration of (PdO)n and Pdn clusters by cage-templated synthesis of hollow silica nanoparticles

Article abstract of DOI:10.1002/anie.201201288

Imprisoned palladium: A unique approach is developed to incarcerate metal clusters with strictly controlled n values within hollow silica nanoparticles. A Pd₁₂L₂₄ spherical complex is used as a template for the hollow silica synthesis. The incarcerated Pd₁₂L₂₄ core is calcinated to give (PdO)_n oxide clusters and subsequently reduced to Pd_n metal clusters within the protective hollow silica. Copyright

Full text of DOI:10.1002/anie.201201288

Angewandte
Chemie
DOI: 10.1002/anie.201201288
Metal Clusters
Incarceration of (PdO)n and Pdn Clusters by Cage-Templated Synthesis
of Hollow Silica Nanoparticles**
Kiyotaka Takao, Kosuke Suzuki, Tatsuya Ichijo, Sota Sato, Hiroyuki Asakura,
Kentaro Teramura, Kazuo Kato, Tomonori Ohba, Takeshi Morita, and Makoto Fujita*
Metal nanoclusters have recently attracted considerable
interest because of their distinct chemical and physical
properties, such as catalytic activities,[1] optical properties,[2]
and magnetic behavior,[3] which are different from those of
both bulk materials and mononuclear metal species. When
the number of metal centers (n) is small, the properties of the
metal clusters (Mn) change dramatically with increasing
n value. Therefore, intense efforts have been made to
synthesize small-numbered, well-defined clusters with non-
distributed n values. Gas-phase preparation by metal vapor-
ization is a promising physical method to prepare small-
numbered Mn clusters (typically in the range of n = 10–
10000).[4] Solution synthesis through the reduction of metal
ion precursors within templates is also an efficient method to
give Mn clusters of similar size.[5] With all of these methods it is
still difficult to strictly control the number of metal centers
particularly when n is around 10 or less. Furthermore, the
resulting synthesized clusters are often chemically and physi-
cally labile and are quite prone to oxidation or fusing into
larger clusters. The most controllable methods to synthesize
metal clusters with small n values (n = 4–60) have been
reported utilizing the voids of dendrimers,[1a,e,6] but this
method is hampered by the tedious dendrimer synthesis.
Herein, we report a unique approach to incarcerated metal
clusters with strictly controlled n values (ca. 12) within hollow
silica nanoparticles. Our method utilizes a Pd12L24 spherical
complex[7] as a template for the hollow silica synthesis. After
sol-gel condensation around the sphere, the incarcerated
Pd12L24 core is calcinated to give (PdO)n oxide clusters and
subsequently reduced to Pdn metal clusters (Figure 1). X-ray
absorption fine structure (XAFS) analysis indicated that the
Figure 1. Preparation of (PdO)n and Pdn cluster-incarcerated silica
nanoparticles (n=ca. 12).
[*] K. Takao, Dr. K. Suzuki, T. Ichijo, Dr. S. Sato, Prof. Dr. M. Fujita
Department of Applied Chemistry, School of Engineering
The University of Tokyo
7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656 (Japan)
E-mail: mfujita@appchem.t.u-tokyo.ac.jp
(PdO)n and Pdn clusters prepared in situ are strictly controlled
with n = ca. 12, because the Pd12L24 precursors are completely
isolated from each other by the hollow silica and have no
opportunity to fuse into larger clusters during calcination and
reduction. All of the metal-incarcerated hollow silica nano-
particles have been well characterized by transmission
electron microscopy (TEM), X-ray photoelectron spectros-
copy (XPS), and/or XAFS to demonstrate that our strategy
provides a unique and highly effective method for cluster
synthesis with a limited number of metal atoms.
For the spherical Pd12L24 precursor 2, we designed bent
ligand 1 with a triethoxysilyl group at the convex (Figure 2a).
The alkoxysilyl group is essential in initiating the sol-gel
condensation of the Si(OMe)4 monomer at the periphery of
the sphere. Ligand 1 was synthesized in high yield in three
steps from 3,5-dibromophenol: 1) Williamson reaction with
4-bromo-1-butene, 2) Suzuki–Miyaura cross coupling with
4-pyridylboronic acid pinacol ester, and 3) hydrosilylation
with HSi(OEt)3. When ligand 1 (5.0 mmol) was treated with
H. Asakura, Dr. K. Teramura
Department of Molecular Engineering, Graduate School of
Engineering, Kyoto University
Kyotodaigaku Katsura, Nishikyo-ku, Kyoto 615-8510 (Japan)
Dr. K. Kato
Japan Synchrotron Radiation Research Institute
1-1 Kouto, Sayo-cyo, Sayo-gun, Hyogo 679-5198 (Japan)
Dr. T. Ohba
Graduate School of Science, Chiba University
1-33 Yayoi-cho, Inage-ku, Chiba 263-8522 (Japan)
Dr. T. Morita
Graduate School of Advanced Integration Science, Chiba University
1-33 Yayoi-cho, Inage-ku, Chiba 263-8522 (Japan)
[**] This research was supported by the JST-CREST project, for which
M.F. is the principal investigator, and also in part by KAKENHI and
MEXT. Synchrotron XAFS studies were performed at SPring-8, and
TEM studies were performed at Chiba University and at the
University of Tokyo.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201201288.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Figure 2. Preparation of Pd12L24 precursor 2. a) Synthesis of 2 from
ligand 1 and PdII ions. b,c) 1H NMR spectra (500 MHz, 278C,
[D6]DMSO) of b) ligand 1 and c) complex 2. d) CSI mass spectrum of
2 (DMSO/CH3CN=1:24).
Pd(BF4)2 (2.6 mmol) in DMSO (0.52 mL) at 508C for 1 h,
a Pd12L24 sphere with 24 alkoxysilyl groups at the periphery of
the sphere was quantitatively formed. In the 1H NMR
spectrum, the a and b protons on the pyridyl groups (PyHa
and PyHb) are shifted downfield owing to coordination to Pd
ions (Dd = 0.72 and 0.51 ppm for PyHa and PyHb, respectively,
Figure 2b,c). Diffusion-ordered NMR spectroscopy (DOSY)
showed a diffusion coefficient of D = 4.0 ꢀ 10À11 m2 sÀ1 in
[D6]DMSO, which indicates the formation of a huge mole-
cule. This result is comparable to those of related Pd12L24
spheres, but considerably smaller than that of the component
ligand (D = 1.8 ꢀ 10À10 m2 sÀ1 in [D6]DMSO). The visible
signal broadening is also indicative of the formation of
a huge molecule (Figure 2b,c). The Pd12L24 composition was
determined by cold-spray ionization (CSI) mass spectrometry
Figure 3. Template synthesis of a Pd12L24 complex incarcerated in
a hollow silica nanoparticle 2. a) Synthesis of 2ꢀ(SiO2)m by the sol-gel
condensation of Si(OMe)4 around the template 2. b) 1H NMR
(500 MHz, 278C, [D6]DMSO/CD3OD/D2O=4:36:1) observation of the
conversion of 2 into 2ꢀ(SiO2)m; the spectrum at the top is before
addition of Si(OMe)4; the middle and bottom spectra are 48 h and
120 h after addition of Si(OMe)4 (2000 equiv), respectively. c,d) TEM
images of hollow silica nanoparticles c) 2ꢀ(SiO2)2000 and d) 2ꢀ
(SiO2)7200. The inset in (c) is the magnified image of a hollow silica
nanoparticle. e,f) Size distribution determined from the TEM images
of e) 2ꢀ(SiO2)2000 and f) 2ꢀ(SiO2)7200
.
with prominent peaks for [2À(BF4 )a]a+ (a = 13–6; Fig-
À
ure 2d).
bulk silica gels did not precipitate and the solution remained
clear even after 5 days. Thus, the sol-gel condensation seems
to proceed only around sphere 2. The formation of the silica
nanoparticle incarcerating 2 (which is hereafter presented as
2ꢀ(SiO2)2000) is strongly indicated by DOSY, which shows
a single product with a diffusion coefficient of D = 5.6 ꢀ
10À11 m2 sÀ1 in the mixed solvent, a value which is considerably
smaller than that of sphere 2 (D = 1.3 ꢀ 10À10 m2 sÀ1) in the
same solvent mixture and consistent with the increased
molecular size caused by the silica coating. Even when the
amount of TMOS was increased (7200 equiv), the condensa-
To incarcerate sphere 2 within a hollow silica nano-
particle, the sol-gel condensation of tetramethoxysilane
(TMOS) at the periphery of 2 was carried out (Figure 3a).
A DMSO solution of sphere 2 (0.30 mm, 80 mL) was diluted
with methanol (0.72 mL) and then TMOS (2000 equiv), water
(50000 equiv), and nitric acid (4.2 equiv) were added. The
reaction was monitored by NMR spectroscopy (Figure 3b).
The signals derived from the core of 2 gradually broadened,
indicating the formation of a silica layer around the periphery
of 2. Despite the use of a large excess of TMOS (2000 equiv),
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
tion proceeded without silica-gel precipitation to form a much
larger silica-coated sphere (2ꢀ(SiO2)7200). As the 1H NMR
signals were heavily broadened, clear signals were not
observed in DOSY and the diffusion coefficient was no
longer measurable. The presence of the Si(OEt)3 group on the
ligand is essential to the initiation of the sol-gel condensation
at the periphery of the sphere: when the Si(OEt)3 groups were
absent, the condensation occurred in bulk solution and silica
gel precipitated while the spherical complex remained intact.
TEM observation showed highly dispersed nanoparticles
with a very narrow size distribution: the diameters of
2ꢀ(SiO2)2000 and 2ꢀ(SiO2)7200 were estimated to be [7.8 Æ
0.3] and [8.8 Æ 0.8] nm, respectively (Figure 3c,d). More
importantly, the TEM showed a clear contrast in the hollows
at the center of the silica nanoparticles (inset in Figure 3c).
From the TEM image, the diameter of the hollows is ca. 3 nm,
a value consistent with the diameter of the Pd12L24 spherical
shell (3.5 nm). Presumably, the highly accumulated Si(OEt)3
groups at the periphery of the sphere initiated the condensa-
tion predominantly at the surface of the sphere, forming a thin
layer of silica on the surface. Once
with a Xe beam at 3 kV followed by XPS experiments at
every 0.1 min. The peak intensities of the Pd 3d XPS
increased as Xe sputtering proceeded, and the binding
energy shifted to lower energy as the duration of X-ray
irradiation increased, which indicates that some of the (PdO)n
clusters were reduced by X-ray irradiation. As shown in
Figure 4b, the Pd/Si molar ratio, as evaluated by comparison
of Pd 3d and Si 2p peaks, increased with increasing sputtering
duration and reached a maximum of 0.35% after 0.4 min.
Further sputtering had no effect on the Pd/Si ratio. This
behavior indicates that the silica shell was destroyed by the
heavy Xe ion for the first 0.4 min of the sputtering and that
the Pd species was then exposed on the surface, at which time
the Pd/Si ratio on the surface became constant. This
observation supports the existence of (PdO)n clusters inside
the silica capsule.
Structural details of the (PdO)nꢀ(SiO2)7200 were eluci-
dated by XAFS at the K-edge of Pd. The X-ray absorption
near edge structure (XANES) spectra of the (PdO)n cluster
and those of Pd foil and PdO as reference samples showed
the layer was formed, the TMOS
monomers could no longer enter
the spherical shell and further
condensation occurred only on
the surface, giving the hollow
structure.
It is noteworthy that the
hollow silica incarcerates a single
3.5 nm-sized Pd12L24 spherical
complex. Because of this fact, we
expected that a Pd12 cluster could
be formed by removal of the
ligands without breaking the
hollow structure of the silica.
Thus, the silica nanoparticle 2ꢀ
(SiO2)7200 was precipitated by
removing the MeOH solvent and
adding poor solvents (ethyl ace-
tate and ether). The precipitated
2ꢀ(SiO2)7200 powder was calci-
nated at 4008C for 2 h in air.
For the calcinated powder,
XPS analyses were performed to
determine the valence state of the
Pd atoms. The core-level Pd 3d3/2
and 3d5/2 peaks of (PdO)nꢀ
(SiO2)7200 were observed at 341.3
and 336.2 eV, respectively. The
comparison of the values to
those of Pd foil and PdO as
reference samples showed the
Figure 4. XPS analyses of (PdO)nꢀ(SiO2)7200 and Pdnꢀ(SiO2)7200. a) Pd 3d XPS spectra of PdO (c),
(PdO)nꢀ(SiO2)7200 (c), Pd foil (c), and Pdnꢀ(SiO2)7200 (c). b) Time course of the Pd/Si ratio of
(PdO)nꢀ(SiO2)7200 during surface sputtering by a Xe beam for a duration of 0.6 min. c) Pd K-edge
XANES spectra of PdO (c), (PdO)nꢀ(SiO2)7200 (c), Pd foil (c), and Pdnꢀ(SiO2)7200 (c).
oxidized state of the Pd atoms,
a result consistent with the for-
mation of (PdO)nꢀ(SiO2)7200 (Fig-
ure 4a). To determine if (PdO)n
clusters were incarcerated in the
silica capsule, the surface of
d) Curve fitting result of the Fourier transformation of the EXAFS of (PdO)nꢀ(SiO2)7200; Dk=2–15 ꢀÀ1
,
DR=1.4–3.4 ꢀ. Curve fitting was performed by Artemis with back scattering and phase shift parameters
calculated by FEFF ver. 8.40, and fell within the Nyquist criteria. Curve fitting was done in R space
(PdO)nꢀ(SiO2)7200 was sputtered (R-factor=0.01078). CN=coordination number, DW=Debye-Waller factor.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
2ꢀ(SiO2)m was analyzed. 1H NMR (500 MHz, [D6]DMSO, 278C):
d = 9.28 (br, 96H), 8.15 (br, 96H), 7.95 (br, 24H), 7.48 (br, 48H), 4.09
(br, 48H), 1.80 (br, 48H), 1.55 (br, 48H), 0.63 ppm (br, 48H).
Synthesis of (PdO)nꢀ(SiO2)m and Pdnꢀ(SiO2)m: 2ꢀ(SiO2)m. was
calcinated at 4008C for 2 h in air to afford (PdO)nꢀ(SiO2)m. The
resulting brown solids were reduced at 4008C for 4 h under hydrogen
atmosphere to afford Pdnꢀ(SiO2)m. The reduction was conducted
that the valence state of the Pd atoms in (PdO)n is divalent,
a result which is consistent with the XPS experiments
(Figure 4c). Extended X-ray absorption fine structure
(EXAFS) experiments were performed to estimate the
value of n for (PdO)n. The coordination number (CN) and
the atomic distance (d) were determined by curve fitting
analyses for adjacent X-ray absorbing and scattering atom
under a continuous flow of H2/Ar (10:90) at a rate of 100 mLminÀ1
.
After reduction, the resulting solids were kept under Ar atmosphere
to avoid oxidation by air, and analyzed.
À
À
À
pairs (Ab Sc) of Pd O and Pd Pd (Figure 4d). The exper-
imental CN and d values were reasonable for the structure of
a (PdO)n cluster where n is ca. 12. The (PdO)n clusters derived
from Pd12L24 complexes were isolated within the silica
capsule, as indicated by the XPS experiments, and we
consequently believe we have obtained nano-sized (PdO)n
clusters (n = ca. 12).
Received: February 16, 2012
Revised: March 30, 2012
Published online: && &&, &&&&
Keywords: cage-templated synthesis ·
.
We further examined the conversion of the (PdO)n oxide
cluster to a Pdn metal cluster by reduction with hydrogen.
After heating the cluster under a hydrogen/argon atmosphere
at 4008C for 4 h, XPS analysis of the product showed the core-
level Pd 3d3/2 and 3d5/2 peaks of Pdn at 340.7 and 335.1 eV,
respectively (Figure 4a). Thus, we confirmed the reduction of
the PdO cluster to a Pd0 species, which was also confirmed by
XANES analysis (Figure 4c). The n value of the Pdn cluster
should be comparable to that of (PdO)n (n = ca. 12), because
hydrogen reduction is moderate enough to keep the capsular
structure of silica and the reduction therefore occurs within
the isolated hollows of the silica. Solid-state 29Si NMR spectra
supported that the structures of silica shell were retained after
the reduction (Supporting Information, Figure S8).
In summary, we have succeeded in preparing hollow silica
nanoparticles templated by a Pd12L24 spherical complex. The
incarcerated Pd12L24 complex is a suitable precursor for the
(PdO)n and Pdn clusters (n = ca. 12), which can be prepared in
the hollows of the silica. As related spherical complexes (for
example, M6L12,[8] M18L24,[9] M24L24,[10] and M24L48;[11] M = Pd
or Pt) have been efficiently synthesized by self-assembly, the
present method could also be applicable to the incarceration
of various metal or metal oxide clusters with predetermined,
exact n values in the range of 10–20. For previous nanocluster
preparation methods, values in this range have been shown to
be the most difficult to control. New catalytic and physical
properties for the incarcerated (PdO)n/Pdn clusters are
currently under investigation.
hollow silica nanoparticles · nanoclusters · palladium ·
self-assembly
[1] a) L. S. Ott, R. G. Finke, Coord. Chem. Rev. 2007, 251, 1075 –
1100; b) K. Okamoto, R. Akiyama, H. Yoshida, T. Yoshida, S.
Kobayashi, J. Am. Chem. Soc. 2005, 127, 2125 – 2135; c) T.
Mitsudome, K. Nose, K. Mori, T. Mizugaki, K. Ebitani, K.
Jitsukawa, K. Kaneda, Angew. Chem. 2007, 119, 3352 – 3354;
Angew. Chem. Int. Ed. 2007, 46, 3288 – 3290; d) W. E. Kaden, T.
Wu, W. A. Kunkel, S. L. Anderson, Science 2009, 326, 826 – 829;
e) K. Yamamoto, T. Imaoka, W.-J. Chun, O. Enoki, H. Katoh, M.
Takenaga, A. Sonoi, Nat. Chem. 2009, 1, 397 – 402.
[2] a) U. Kreibig, M. Vollmer, Optical Properties of Metal Clusters,
Springer, Berlin, 1995; b) J. P. Wilcoxon, P. P. Provencio, J. Am.
Chem. Soc. 2004, 126, 6402 – 6408; c) S. W. Han, Y. Kim, K. Kim,
J. Colloid Interface Sci. 1998, 208, 272 – 278.
[3] a) J. P. Bucher, D. C. Douglass, L. A. Bloomfield, Phys. Rev. Lett.
1991, 66, 3052 – 3055; b) I. M. L. Billas, A. Chꢁtelain, W. A.
de Heer, Science 1994, 265, 1682 – 1684; c) J. Bansmann, et al.,
Surf. Sci. Rep. 2005, 56, 189 – 275.
[4] a) W. A. de Heer, Rev. Mod. Phys. 1993, 65, 611 – 676; b) A.
Piednoir, E. Perrot, S. Granjeaud, A. Humbert, C. Chapon, C. R.
Henry, Surf. Sci. 1997, 391, 19 – 26; c) H. Yasumatsu, T. Hay-
akawa, S. Koizumi, T. Kondow, J. Chem. Phys. 2005, 123,
124709 – 12717; d) R. Alayan, L. Arnaud, M. Broyer, E. Cottan-
cin, J. Lermꢂ, J. L. Vialle, M. Pellarin, Phys. Rev. B 2006, 73,
125444 – 125457; e) T. Wu, W. E. Kaden, W. A. Kunkel, S. L.
Anderson, Surf. Sci. 2009, 603, 2764 – 2770.
[5] a) J. P. Wilcoxon, B. L. Abrams, Chem. Soc. Rev. 2006, 35, 1162 –
1194; b) J. D. Aiken, R. G. Finke, J. Mol. Catal. A 1999, 145,
1 – 44.
[6] M. Zhao, L. Sun, R. M. Crooks, J. Am. Chem. Soc. 1998, 120,
4877 – 4878.
Experimental Section
[7] a) M. Tominaga, K. Suzuki, M. Kawano, T. Kusukawa, T. Ozeki,
S. Sakamoto, K. Yamaguchi, M. Fujita, Angew. Chem. 2004, 116,
5739 – 5743; Angew. Chem. Int. Ed. 2004, 43, 5621 – 5625; b) K.
Suzuki, S. Sato, M. Fujita, Nat. Chem. 2010, 2, 25 – 29; c) K.
Suzuki, K. Takao, S. Sato, M. Fujita, Angew. Chem. 2011, 123,
4960 – 4963; Angew. Chem. Int. Ed. 2011, 50, 4858 – 4861.
[8] a) K. Suzuki, M. Tominaga, M. Kawano, M. Fujita, Chem.
Commun. 2009, 1638 – 1640; b) D. Fujita, A. Takahashi, S. Sato,
M. Fujita, J. Am. Chem. Soc. 2011, 133, 13317 – 13319.
[9] Q.-F. Sun, S. Sato, M. Fujita, Nat. Chem. 2012, 4, 330 – 333.
[10] Q.-F. Sun, T. Murase, S. Sato, M. Fujita, Angew. Chem. 2011, 123,
10502 – 10505; Angew. Chem. Int. Ed. 2011, 50, 10318 – 10321.
[11] Q.-F. Sun, J. Iwasa, D. Ogawa, Y. Ishido, S. Sato, T. Ozeki, Y. Sei,
K. Yamaguchi, M. Fujita, Science 2010, 328, 1144 – 1147.
Synthesis of Sphere 2: Ligand 1 (2.23 mg, 5.00 mmol) was treated with
a DMSO solution of Pd(BF4)2(CH3CN)4 (10 mm, 0.52 mL, 5.2 mmol)
at 508C for 1 h. The quantitative formation of sphere 2 was confirmed
by 1H NMR and CSI-MS. 1H NMR (500 MHz, [D6]DMSO, 278C):
d = 9.38 (br, 96H), 8.36 (br, 96H), 7.96 (br,248H), 7.56 (br, 48H), 4.07
(br, 48H), 3.64 (br, 144H), 1.69 (br, 48H), 1.42 (br, 48H), 1.01 (br,
48H), 0.73 ppm (br, 48H).
Synthesis of 2ꢀ(SiO2)m: Sphere 2 (0.024 mmol) in DMSO (80 mL,
0.30 mm) was successively diluted with MeOH (0.72 mL), D2O
(20 mL), and aqueous DNO3 (0.33w/w%, 2.0 mL), and then treated
with tetramethoxysilane (2000 or 7200 equiv vs. 2) at RT for 5 days.
As the reaction progressed, the 1H signals of 2 broadened, which
indicated the formation of silica nanoparticles on the surface of 2
(m = 2000). After all of the Si(OCH3)4 groups were hydrolyzed,
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Metal Clusters
K. Takao, K. Suzuki, T. Ichijo, S. Sato,
H. Asakura, K. Teramura, K. Kato,
T. Ohba, T. Morita,
M. Fujita*
&&&& — &&&&
Incarceration of (PdO)n and Pdn Clusters
by Cage-Templated Synthesis of Hollow
Silica Nanoparticles
Imprisoned palladium: A unique
thesis. The incarcerated Pd12L24 core is
approach is developed to incarcerate
metal clusters with strictly controlled
n values within hollow silica nanoparti-
cles. A Pd12L24 spherical complex is used
as a template for the hollow silica syn-
calcinated to give (PdO)n oxide clusters
and subsequently reduced to Pdn metal
clusters within the protective hollow
silica.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!