Magnetic Core–Shell to Yolk–Shell Structures in Palladium-Catalyzed Suzuki–Miyaura Reactions: Heterogeneous versus Homogeneous Nature

Article abstract of DOI:10.1002/cplu.201600094

This study describes a comparative investigation on the heterogeneous versus homogeneous nature of the Pd-catalyzed Suzuki–Miyaura cross-coupling reaction mechanism with specific magnetic hierarchical core–shell and yolk–shell structures. The hierarchical core–shell Fe₃O₄@SiO₂-Pd@mCeO₂ (m=mesoporous) catalyst contains a core of nonporous silica-sheltered magnetite (Fe₃O₄) nanoparticles (NPs), a transition layer of active palladium (Pd) NPs, and an outer shell of porous ceria (CeO₂). The magnetic yolk–shell Fe₃O₄@h-Pd@mCeO₂ (h=hollow) catalyst was prepared by selectively etching the nonporous silica interlayers. Notably, the results of the hot-filtration heterogeneity test, the effect of Pd concentration, and solid-phase poisoning, indicate that the two kinds of catalysts function in Pd-catalyzed Suzuki–Miyaura cross-coupling reactions through different catalytic mechanisms. Moreover, both catalysts demonstrated better catalytic activity than the Fe₃O₄@SiO₂-Pd catalyst. This finding can be ascribed to the outermost CeO₂ shell having a high concentration of trivalent cerium and oxygen vacancies, which gives rise to the increased electron density of Pd NPs, and a faster rate-determining step in the oxidative addition reaction for the Suzuki reaction. In addition, we propose a feasible mechanism elucidating the synergistic effect between the supporting CeO₂ and active species.

Full text of DOI:10.1002/cplu.201600094

DOI: 10.1002/cplu.201600094
Full Papers
Magnetic Core–Shell to Yolk–Shell Structures in Palladium-
Catalyzed Suzuki–Miyaura Reactions: Heterogeneous
versus Homogeneous Nature
Yinle Li,[a] Zhuqing Zhang,[a] Tao Fan,[a] Xiaoguang Li,[a] Jin Ji,[a] Pei Dong,[b] Robert Baines,[b]
Jianfeng Shen,*[a, b] and Mingxin Ye*[a]
This study describes a comparative investigation on the heter-
ogeneous versus homogeneous nature of the Pd-catalyzed
Suzuki–Miyaura cross-coupling reaction mechanism with spe-
cific magnetic hierarchical core–shell and yolk–shell structures.
The hierarchical core–shell Fe3O4@SiO2-Pd@mCeO2 (m=meso-
porous) catalyst contains a core of nonporous silica-sheltered
magnetite (Fe3O4) nanoparticles (NPs), a transition layer of
active palladium (Pd) NPs, and an outer shell of porous ceria
(CeO2). The magnetic yolk–shell Fe3O4@h-Pd@mCeO2 (h=
hollow) catalyst was prepared by selectively etching the non-
porous silica interlayers. Notably, the results of the hot-filtra-
tion heterogeneity test, the effect of Pd concentration, and
solid-phase poisoning, indicate that the two kinds of catalysts
function in Pd-catalyzed Suzuki–Miyaura cross-coupling reac-
tions through different catalytic mechanisms. Moreover, both
catalysts demonstrated better catalytic activity than the
Fe3O4@SiO2-Pd catalyst. This finding can be ascribed to the
outermost CeO2 shell having a high concentration of trivalent
cerium and oxygen vacancies, which gives rise to the increased
electron density of Pd NPs, and a faster rate-determining step
in the oxidative addition reaction for the Suzuki reaction. In ad-
dition, we propose a feasible mechanism elucidating the syner-
gistic effect between the supporting CeO2 and active species.
Introduction
The Pd-catalyzed Suzuki–Miyaura cross-coupling reaction is
one of the most useful tools for the formation of carbon–
carbon (CÀC) bonds in organic synthesis.[1] This reaction has
become a main branch in modern organic synthetic chemistry
for the synthesis of bioactive compounds,[2] polymers,[3] phar-
maceutical intermediates,[4] and functional materials.[5] General-
ly, the homogeneous metal Pd complexes are known to dis-
play better catalytic activities than the heterogeneous Pd cata-
lyst systems. However, when it comes to large-scale application
in liquid-phase reactions, more difficulties arise, such as the
barriers of recycling of the catalyst, obstacles in further purifi-
cation of the final product, and difficulties in preventing the
Pd catalyst from aggregating. Meanwhile, the removal of Pd
complexes from reaction systems needs further improvement,
considering their high cost and potential toxicity. Therefore,
much attention has been directed towards the development
of heterogeneous Pd catalyst systems with the metal Pd NPs
loaded on inorganic or organic supports.[6]
In recent years, a broad array of highly effective heterogene-
ous Pd catalysts has been developed for Suzuki reactions.[7,8]
These Pd catalysts can be used repeatedly and consequently
offer an efficient and economical way to carry out Suzuki reac-
tions. Moreover, they do not require phosphine ligands, are
not susceptible to moisture or air, and provide products in
high yields with short reaction times. As such, heterogeneous
Pd catalysts are considered to be promising alternatives to ho-
mogeneous metal Pd complexes. However, an inherent prob-
lem with heterogeneous Pd-catalyzed Suzuki reactions is their
mechanistic nature.[6,9–14] Viewpoints about this in the literature
differ as to whether heterogeneous Pd-catalyzed cross-cou-
pling reactions result from the leaching of active Pd NPs or the
supporting Pd NPs themselves.[9,10,15–24] A direct comparison of
the results of each case proves futile, since the reaction condi-
tions, the nature of the employed Pd NPs and their stabilizers
vary from experiment to experiment. Therefore, the design of
strategies to avoid the variety of factors for comprehensive
mechanistic study is a considerable challenge.
[a] Dr. Y. Li, Z. Zhang, T. Fan, X. Li, J. Ji, Prof. J. Shen, Prof. M. Ye
Institute of Special Materials and Technology
Fudan University, Shanghai 200433 (P. R. China)
E-mail: mxye@fudan.edu.cn
Generally, nano-sized noble-metal catalysts have more ex-
posed surface area that enhances catalytic activity.[25] However,
owing to their high surface energies, naked noble metal NPs
tend to leach and sinter during catalytic reactions and lead to
a sharp drop in catalytic activity and selectivity in reactions.
Another considerable factor underlying the utilization of Pd
NPs is the capacity to extract them from an intricate heteroge-
neous system.[26–28] Therefore, the design of structures for so-
jfshen@fudan.edu.cn
[b] P. Dong, R. Baines, Prof. J. Shen
Department of Materials Science and Nanoengineering
Rice University, 6100 Main Street
Houston, TX 77005 (USA)
Supporting information for this article (including EDX, XPS, and FTIR
analyses, magnetization curves, N2 adsorption–desorption isotherms,
SEM and TEM images after recycling, and Tables S1–S6) can be found
under http://dx.doi.org/10.1002/cplu.201600094.
ChemPlusChem 2016, 81, 564 – 573
564
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
phisticated mechanistic study should also avoid these factors.
More recently, novel hierarchical core–shell structures com-
posed of an inner protected Fe3O4 core–shell nanostructure,
a porous outer shell, and noble-metal NPs active species
clamped by the inside and outside shell, which can stabilize
the active metal NPs to a higher degree and reduce the
amount of leaching.[29–35] The magnetic yolk–shell structures,
which have a characteristic hollow component, an extraordina-
ry configuration of a movable magnetic interior core, void
space, and a penetrable outer shell renders them viable candi-
dates.[36–42] As such, they have also been a hot spot of scientific
research of late motivating application research in areas like
photocatalysis,[43] nanoscale reactors,[44] and drug carriers.[45]
Zhao et al. synthesized multicomponent and multifunctional
Fe3O4@SiO2-Au@mSiO2 hierarchical core–shell structures for the
reduction of 4-nitrophenol and styrene epoxidation.[46] Our
Scheme 1. Synthesis of Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-Pd@mCeO2
NPs.
NPs were modified with (3-aminopropyl)triethoxysilane
(APTES). This process imbued the surface with amino groups,
which increased the coordination bond interaction of Pd. Next,
Pd NPs were introduced to the surface of the Fe3O4@SiO2 NPs
by using a PdCl2 precursor with NaBH4 reductant. The prepared
Pd-immobilized Fe3O4@SiO2 NPs are denoted Fe3O4@SiO2-Pd.
Finally, using Ce(NO3)3·6H2O as the cerium source and hexam-
ethylene tetramine (HMT) to incrementally produce OHÀ, we
created a homogeneous CeO2 coating around the Fe3O4@SiO2-
Pd NPs by means of a sol–gel process.[49] Subsequently, we
team also synthesized
a
similar hierarchical Fe3O4@C-
Pd@mCeO2 catalyst for the Suzuki–Miyaura cross-coupling re-
actions and the reduction of 4-nitrophenol.[47] However, with
the Fe3O4@C-Pd@mCeO2 catalyst it is difficult to form the yolk–
shell structures. Most of the CeO2 layers crack after high-tem-
perature-calcination removal of the carbon shell.
In this study, we report the design and fabrication of a well-
defined magnetic hierarchical Fe3O4@SiO2-Pd@mCeO2 catalyst
and the corresponding magnetic yolk–shell Fe3O4@h-
eliminated HMT with
a
calcination process, creating
Fe3O4@SiO2-Pd@mCeO2 NPs with sandwich-like structures. Fi-
nally, we selectively etched the nonporous silica interlayers
using an alkaline (NaOH, 10 wt%) solution, giving rise to the
formation of a magnetic yolk–shell Fe3O4@h-Pd@mCeO2 cata-
lyst.
Pd@mCeO2 catalyst. The mesoporous CeO2 shell had
a
Ce3+/Ce4+ redox cycle. In tandem with the active noble-metal
NPs, the mesoporous shell enhanced catalytic and electro-
chemical performances because of the synergetic effects be-
tween CeO2 NPs and noble-metal NPs.[48] Both the hierarchical
Fe3O4@SiO2-Pd@mCeO2 core–shell and Fe3O4@h-Pd@mCeO2
yolk–shell structures were highly mesoporous for the permea-
tion of reactants and products, and magnetic for reusability,
and had good dispersibility in water and ethanol. The catalytic
activity was systematically evaluated by assessing each struc-
ture’s application in the Suzuki–Miyaura cross-coupling reac-
tions. These specific core–shell and yolk–shell structures pro-
vided a platform from which we studied the aforementioned
“mechanistic nature” issue witnessed in Pd-catalyzed cross-cou-
pling reactions. We also propose a feasible mechanism explain-
ing the electron-donating effect between the CeO2 support
and the Pd NP active species and describe the catalytic reac-
tion mechanism with the characterization of catalyst and the
experimental results.
The morphology and structure of the obtained composites
were characterized by scanning electron microscopy (SEM),
transmission electron microscopy (TEM), high-resolution (HR)-
TEM, high-angle annular dark field (HAADF) scanning transmis-
sion electron microscopy (STEM), and energy-dispersive X-ray
spectroscopy (EDS) mapping. Figure 1a,b shows the SEM
images of the Fe3O4 NPs and Fe3O4@SiO2 NPs. Clearly, the
Fe3O4 NPs are 200–250 nm and spherical, while the later-ob-
tained Fe3O4@SiO2 NPs (Figure 1b) are a larger 270–320 nm.
Results and Discussion
The procedure for the synthesis of the magnetic hierarchical
core–shell and yolk–shell catalysts is shown in Scheme 1. First,
the Fe3O4 NPs were modified by trisodium citrate (Na3Cit)
through a solvothermal reaction.[46] The modified Fe3O4 NPs ex-
hibited remarkable dispersibility in the polar solvents (water
and ethanol). Second, we coated a uniform SiO2 shell on the
surface of Fe3O4 NPs by using a sol–gel process consisting of
hydrolysis and condensation of tetraethyl orthosilicate (TEOS)
in an ethanol/ammonia mixture. Thus we created a Fe3O4@SiO2
core–shell structure. To deposit Pd NPs efficiently, Fe3O4@SiO2
Figure 1. SEM images of (a) Fe3O4, (b) Fe3O4@SiO2 (inset is the corresponding
TEM image), (c) Fe3O4@SiO2-Pd@mCeO2, (d) Fe3O4@h-Pd@mCeO2.
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
565
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
The uniform SiO2 shell, about 35 nm thick, formed on individu-
al Fe3O4 NPs seeds (inset in Figure 1b), creating a core–shell
Fe3O4@SiO2 structure. The SEM images in Figure 1c illustrate
that the Fe3O4@SiO2-Pd@mCeO2 catalyst retains a clean spheri-
cal shape after loading onto the CeO2 shell. However, the
yellow coil in Figure 1d indicates that parts of the spherical
Fe3O4@h-Pd@mCeO2 catalyst were broken down in the etching
process. Figures S1a and b in the Supporting Information
show the EDS spectra of Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-
Pd@mCeO2 catalysts. We estimated the loading amounts of Pd
to be 2.51 and 2.83 wt%, respectively.
mental mapping depicted the elemental distribution of these
novel structures. The HADDF-STEM image in Figure 3 shows
the “shell-in-shell” structure of the Fe3O4@SiO2-Pd@mCeO2 cata-
lyst. The corresponding EDS mapping images indicate that Fe,
Si, Pd, Ce, and O were dispersed evenly with differently sized
spherical shapes. The HADDF-SETM image of the Fe3O4@h-
Pd@mCeO2 catalyst (Figure S1c in the Supporting Information)
and the corresponding EDS mapping images also affirmed the
differently sized spherical shape and the distribution of ele-
ments Fe, Pd, Ce, and O but without element Si.
The TEM image of Fe3O4@SiO2-Pd (Figure 2a) indicates the
numerous Pd NPs deposited uniformly on the SiO2 shell. After
further deposition of CeO2, a CeO2 shell on Fe3O4@SiO2-
Pd@mCeO2 with a thickness of 25 nm was obtained (Figure 2c).
The HR-TEM images (Figure 2b,d) elucidate that the Pd NPs
and the CeO2 nanocrystallites had sizes of approximately 5–10
and 5 nm, respectively. The lattice fringes are visible, with d
Figure 3. HADDF-STEM images of Fe3O4@SiO2-Pd@mCeO2 and the corre-
sponding EDS mapping images.
To detail the surface chemical composition and further con-
firm the formation of the Pd NPs and the CeO2 shell, X-ray
photoelectron spectroscopy (XPS) was conducted (Figure 4
and Figure S2 in the Supporting Information). As noticeable in
Figure 4d, it is hard to detect the peak for Fe from the wide
scan survey of Fe3O4@SiO2-Pd (Figure S2a). This suggests that
the majority of the Fe3O4 was in fact coated by a silica layer
Figure 2. TEM and HR-TEM images of (a) Fe3O4@SiO2-Pd, (b) Pd,
(c) Fe3O4@SiO2-Pd@mCeO2, (d) CeO2, (e,f) Fe3O4@h-Pd@mCeO2.
spacings of 0.23 and 0.32 nm corresponding to the lattice
spacing of the Pd NPs (111) lattice planes and the cubic uorite
phase CeO2 (111) lattice planes.[50] Figure 2e,f show TEM
images of Fe3O4@h-Pd@mCeO2 with yolk–shell structures after
the SiO2 shell etching reaction. These consist of Fe3O4 cores, an
apparent chamber, and the outer CeO2 layers. Unfortunately,
the rupture of outer CeO2 layers resulted in the formation of
individual Fe3O4 balls and dispersive CeO2 NPs (the arrows in
Figure 2f).
It was difficult to identify the interface effect between Pd
NPs and CeO2 because of the concentrated electron density of
polycrystalline CeO2 and the specific core–shell architecture.[51]
Therefore, the encapsulated Pd NPs could not be explicitly rec-
ognized from the TEM images of the Fe3O4@SiO2-Pd@mCeO2
and Fe3O4@h-Pd@mCeO2 structures. HADDF-STEM and EDS ele-
Figure 4. XPS spectra of (a) Ce 3d of Fe3O4@SiO2-Pd@mCeO2, (b) Ce 3d of
Fe3O4@h-Pd@mCeO2, (c) Pd 3d, (d) Fe 2p.
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
566
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and formed a core–shell structure.[52] The high-resolution XPS
spectrum of Pd (Figure 4c) depicts its binding energies: 341.2
and 335.8 eV, associated with the Pd 3d3/2 and Pd 3d5/2 of Pd0,
respectively.[53] These values confirm the formation of metallic
Pd on the surface of Fe3O4@ SiO2 NPs. The XPS spectra of Fig-
ure S2d, Figure S2b and c indicate that C, O, and Ce existed
on the surface of Fe3O4@SiO2@mCeO2, Fe3O4@SiO2-Pd@mCeO2,
and Fe3O4@h-Pd@mCeO2. With regard to the XPS signals of the
Ce 3d spectra (Figure S2e, Figure 4a and b): the five character-
istic peaks of Ce 3d3/2, situated at 916.2, 907.2, 903.9, 900.7,
and 899.3 eV correspond to u’’’, u’’, u’, u, and u0 units,[54,55]
whereas the other five characteristic peaks of Ce 3d5/2, situated
at 898.6, 888.6, 884.4, 882.2, and 880.6 eV match well with
components v’’’, v’’, v’, v, and v0 units.[54,55] The signals v0, v’, u0,
and u’ were characteristic peaks of Ce3+, and the other peaks
corresponded to the characteristic peaks of Ce4+. The presence
of Ce3+ strongly suggests the synergistic effect between CeO2
and Pd NPs. The concentration of Ce3+ in CeO2 can be deter-
mined with the equations in the Supporting Information. The
Ce3+ to Ce4+ ratio in the Fe3O4@SiO2@mCeO2, Fe3O4@SiO2-
Pd@mCeO2, and Fe3O4@h-Pd@mCeO2 catalysts was calculated
to be 16.8, 18.5, and 23.3%, respectively.
NPs in Figure 5e, which is a testament to their low content
loading and their high dispersity.[56]
The Fourier transform infrared spectroscopy (FTIR) spectra of
Fe3O4,
Fe3O4@SiO2,
Fe3O4@SiO2-NH2,
Fe3O4@SiO2-Pd,
Fe3O4@SiO2-Pd@mCeO2, and Fe3O4@h-Pd@mCeO2 NPs are
shown in Figure S3 in the Supporting Information. As depicted
in Figure S3a, the bands at 580 and 1650 cmÀ1 were assigned
to the stretching vibration of FeÀO and the C=O of Na3Cit.[57]
The broad peak around 3400 cmÀ1 is attributed to the vibra-
tion of OH. Figure S3b displays the absorption band at
1088 cmÀ1 and corresponds to an antisymmetric stretching vi-
bration of the SiÀOÀSi bond at the oxygen–silica tetrahedron.
This indicates that the surfaces of Fe3O4 NPs were successfully
coated by the silica shells. After functionalization with APTES,
an obvious band appears at 2926 cmÀ1 (Figure S3c) and is
a factor of the CÀH stretching vibrations of APTES. From the
FTIR spectra of Figure S3d and e, we note that the absorption
peaks of Fe3O4@SiO2-Pd and Fe3O4@SiO2-Pd@mCeO2 are similar
to those of the Fe3O4@SiO2-NH2 NPs. However, the absorption
peak of SiÀOÀSi at 1088 cmÀ1 evidently disappears in the spec-
trum of yolk–shell Fe3O4@h-Pd@mCeO2 (Figure S3f), which indi-
cates that the middle SiO2 shell was thoroughly etched.
The phase and composition of the as-synthesized samples
were thoroughly examined using X-ray diffraction (XRD). All
the peaks at 29.9, 35.2, 43.0, 56.8, and 62.58 are in agreement
with the Fe3O4 face-centered cubic structure in Figure 5a–f.[47]
From Figure 5b it can be seen that there is no distinct change
in the XRD pattern of Fe3O4@SiO2 NPs relative to the XRD pat-
tern of Fe3O4 NPs. In the XRD pattern of Fe3O4@SiO2-Pd in Fig-
ure 5c, except for the peaks of the Fe3O4 core, another peak lo-
cated at 40.08 is associated with face-centered Pd (111) peaks
(JCPDS no. 05-0681).[50] Figure 5d–f show the XRD patterns of
Fe3O4@SiO2@mCeO2, Fe3O4@SiO2-Pd@mCeO2, and Fe3O4@h-
Pd@mCeO2 NPs, respectively. These figures index the peaks at
28.7, 33.1, 47.8, and 56.28, corresponding to the (111), (200),
(220), and (311) Bragg diffraction planes of CeO2 (JCPDS No.34-
0394).[56] There were no diffraction peaks characteristic of Pd
Magnetic hysteresis loops of the samples were measured
using a superconducting quantum interference device (SQUID)
magnetometer at room temperature from À20000 to
20000 Oe. Figure S4 depicts the magnetization curves of Fe3O4
NPs (black), Fe3O4@SiO2 NPs (red), Fe3O4@SiO2-Pd NPs (blue),
Fe3O4@SiO2-Pd@mCeO2 NPs (gray), and Fe3O4@h-Pd@mCeO2
NPs, respectively. At room temperature, the saturation magnet-
ization value is 79.5, 22.8, 22.1, 18.2, and 38.5 emugÀ1, respec-
tively.
Brunauer–Emmett–Teller (BET) measurements were carried
out to further analyze the pore-size distribution and surface
area. Figure S5 (Supporting Information) shows the N2 adsorp-
tion–desorption isotherms of the Fe3O4@SiO2-Pd@mCeO2 and
Fe3O4@h-Pd@mCeO2 NPs, which exhibit characteristic type-IV
curves. The BET surface area was calculated to be 95 and
90 m2 gÀ1, respectively. As mentioned above, a fraction of the
CeO2 layers was broken down in the etching process. The frag-
mented CeO2 nanoparticles and Fe3O4 nanospheres can reunite
to form a stabilized group, or they may attach themselves to
an independent Fe3O4@h-Pd@mCeO2 nanosphere. From the
SEM images of Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-
Pd@mCeO2, we also found that the reunion phenomenon of
the Fe3O4@h-Pd@mCeO2 catalyst is more obvious, which may
be the reason for the diminished BET surface areas of the yolk–
shell Fe3O4@h-Pd@mCeO2 catalyst. The pore-size distribution of
the Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-Pd@mCeO2 NPs (inset
in Figures S5a and b) indicates that the pore sizes range pri-
marily from 2.0 to 6.0 nm, with a characteristic peak appearing
at around 4.2 and 4.8 nm, respectively; this indicates a mesopo-
rous CeO2 shell. This result is consistent with the literature.[49]
The successful preparation of the Fe3O4@SiO2-Pd@mCeO2
and Fe3O4@h-Pd@mCeO2 catalysts could be ascribed to the fol-
lowing essentials. Firstly, the Fe3O4 NPs were decorated with
citrate groups by a solvothermal method, which makes them
more soluble in water and ethanol for the further coating by
Figure 5. XRD patterns of (a) Fe3O4 NPs, (b) Fe3O4@SiO2 NPs, (c) Fe3O4@SiO2-
Pd NPs, (d) Fe3O4@SiO2@mCeO2 NPs, (e) Fe3O4@SiO2-Pd@mCeO2 NPs,
(f) Fe3O4@h-Pd@mCeO2 NPs.
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
567
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
other oxides or polymers. Thus, a uniform SiO2 layer can be
lightly coated on the surface of Fe3O4 NPs by a sol–gel process.
To deposit Pd NPs efficiently, Fe3O4@SiO2 was modified with
APTES to provide plentiful amino groups on the surface of
SiO2, which boosted the interaction with Pd through the coor-
dination bonds. Since the obtained Fe3O4@SiO2-Pd was hydro-
philic and highly soluble in ethanol and water, the succeeding
coating of CeO2 was accomplished in the presence of HMT.
The as-prepared catalyst was able to be readily dispersed in
water, ethanol, dimethylformamide (DMF), and tetrahydrofuran
(THF) and formed a steady dispersion owing to the electrostat-
ic repulsion effect, and could be rapidly isolated from the
water dispersion by an external magnetic field. Thus, the Fe3O4
core–shell and yolk–shell catalyst could be recycled from the
reaction system by an external magnetic field.
After detailed characterization of the synthesized material,
we selected the reaction of 4-iodoanisole and phenylboronic
acid as a benchmark reaction to evaluate the catalytic efficien-
cy of our catalysts for the Suzuki reaction. We chose water and
ethanol as the green solvent to carry out this reaction without
any external agents like phosphine ligands or biphasic media.
As depicted in Table S1 in the Supporting Information, the re-
action did not proceed at all in water (Table S1, entry 1). A low-
level yield of the desired product was acquired by
Fe3O4@SiO2@mCeO2 without Pd NPs, with ethanol as solvent
(Table S1, entry 2). No obvious enhancement in yield was ob-
served by prolonging the reaction time (Table S1, entry 3) or
by using a mixed solvent of water and ethanol (Table S1,
entry 4). This illustrates that CeO2 has catalytic activity, but not
as the primary catalytic active site required for the Suzuki–
Miyaura reaction.[58] Next, we conducted the reaction using
Fe3O4@SiO2-Pd catalyst with the same solvent, and obtained
the desired result, with excellent yield, within three hours
(Table S1, entry 5–7). Based on these results, it is clear that the
cross-coupling reaction was primarily catalyzed by Pd NPs.
When Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-Pd@mCeO2 were
used as catalysts (Table S1, entry 8–13) the yield improved rela-
tive to the corresponding Fe3O4@SiO2-Pd catalyst. Interestingly,
it could be found that the Fe3O4@h-Pd@mCeO2 catalyst had
a better catalytic activity than the Fe3O4@SiO2-Pd@mCeO2 cata-
lyst from the yields.
Figure 6. Plots of Suzuki–Miyaura coupling reactions; reaction conditions are
given in Table 1: (a) conversion; (b) catalyst removed after 30 min; (c) varying
Pd concentrations for Fe3O4@SiO2-Pd; (d) varying Pd concentrations for
Fe3O4@SiO2-Pd@mCeO2; (e) varying Pd concentrations for Fe3O4@h-
Pd@mCeO2. (f) PVPy added as a poison at start of reaction; GC conversion,
conversion were averages of two independent runs.
pared the catalytic performance of different magnetic Pd-
based catalysts by studying the coupling reaction of iodoben-
zene and phenylboronic acids. The results are summarized in
Table S3 in the Supporting Information. For the sake of com-
paring the performances of magnetic core–shell immobilized
Pd nanocatalyst systems for Suzuki reactions, we have collect-
ed some typical results from the literature, as shown in the
Table S4. All these catalytic systems with low mol% Pd gave
rise to excellent coupling product yields with at least two to
ten stable reaction cycles. Some of them, including
Fe3O4@SiO2-Pd@mCeO2, were reported to leach trace amounts
of Pd after stable reaction cycles. Accordingly, Fe3O4@SiO2-
Pd@mCeO2 represents one of the good-performance magnetic
core–shell immobilized Pd nanocatalyst systems for Suzuki re-
actions to date. Nevertheless, even though some of them ac-
quired high yields, toxic solvents (such as THF, CH2Cl2, or DMF)
were used or the reaction was given a long time to occur. The
better catalytic activity can be attributed to the high disper-
sion of the catalyst in the solvent, the outer CeO2 shell effec-
tively prevent the leaching and agglomeration of Pd NPs
during the reactions, and the synergistic effect between the
supporting CeO2 and active Pd NPs species.
To compare catalytic activity, the reaction kinetics of the
template reaction was studied with the reaction rate as the
measurement criteria with 0.5 mmol% Pd loading. The kinetic
profiles (Figure 6a) clearly show the hierarchy of catalytic
activity:
Fe3O4@h-Pd@mCeO2 >Fe3O4@SiO2-Pd@mCeO2 >
Fe3O4@SiO2-Pd. The initial reaction rates were calculated to be
1.8, 1.2, and 1.0 mmolmLÀ1 minÀ1 for the catalysts Fe3O4@h-
Pd@mCeO2, Fe3O4@SiO2-Pd@mCeO2, and Fe3O4@SiO2-Pd, re-
spectively. After that, the effect of the solvent was investigated
(Table S2 in the Supporting Information). The use of isopropyl
alcohol (IPA), THF, DMF, 1,4-dioxane, dimethyl sulfoxide
(DMSO), dimethylacetamide (DMA), toluene, or dimethylben-
zene in combination with potassium carbonate as the base
each give high coupling yields. Significantly, addition of 10%
H2O to a particular selected solvent could effectively raise the
yield of the product (Table S2, entries 1–16). We further com-
Although various metal oxides were used to elevate the cat-
alytic activity and stability of the Pd NPs’, the interfacial effect
between those oxides and the active Pd NPs was ambiguous
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
568
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
due to multiple factors (such as morphology, surface area, and
accessible active sites).[59] The selective CeO2 shell, which had
a Ce4+/Ce3+ redox cycle, is a promising carrier owing to its
high oxygen mobility, high oxygen storage capability, and low
cost.[60] CeO2 has a fluorite structure, in which the center of the
Ce4+ cation is encircled by each O2À anion in a tetrahedron ar-
rangement. The non-stoichiometric oxides CeO2Àx, which were
achieved through wet-chemical synthesis, had a high density
of oxygen vacancies because of the partial O atom elimination
from O2À of the CeO2 structure.[61] Therefore, the surplus elec-
trons, after the elimination process, were either delocalized in
the conductor band, distributed to Ce3+ cations, or localized in
the center Ce4+ to form Ce3+. Simultaneously, the adsorption
and dissociation of water occurring on the O vacancy of
CeO2Àx produced the reactive OHdÀ groups.[62] Both Ce3+ spe-
cies and OHdÀ groups were active reactive dÀ sites, and they
imbued the Pd NPs with electrons.[62] The Ce3+ cations and O
vacancy therefore impart a synergistic effect that enhances the
catalytic activity.
oxidative addition reaction is the rate-determining step in our
catalytic system.[14]
By using the optimized reaction conditions, cross-coupling
reactions of several aryl iodides, aryl bromides, and aryl chlor-
ides with arylboronic acid were used with a Pd loading of
0.5 mmol% (Table 1). Aryl iodides and aryl bromides could be
Table 1. Results for cross-coupling of aryl halides with arylboronic acid.[a]
Entry
R1
X
R2
Yield[b]
Yield[c]
1
2
3
4
5
6
7
8
OCH3
OCF3
NH2
2-CH3
COCH3
OCH3
COCH3
OCH3
COCH3
H
I
I
I
I
I
Br
H
H
H
H
H
H
H
H
94
96
90
78
93
85
87
68
88
75
95
96
90
>99
>99
93
84
>99
92
93
72
92
82
Br
Cl[d]
A proposed mechanism for this synergistic effect is illustrat-
ed in Scheme 2. During the synthesis of CeO2, Ce3+ cations
(the presence of which was confirmed by X-ray photoelectron
9
Cl[d]
H
10
11
12
13
14[e]
I
I
I
I
I
2-CH3
4-CH3
4-OCH3
4-F
H
H
H
H
97
>99
96
H
>99
>99
[a] Reaction conditions: aryl halide (1.0 mmol), arylboronic acid
(1.2 mmol), potassium carbonate (2.5 mmol), 5 mL ethanol/water (9:1),
catalyst (0.5 mmol% Pd), at 808C for 3 h. [b] GC yield: 20.62 mg with
Fe3O4@SiO2-Pd@mCeO2 as catalyst. [c] GC yield: 17.4 mg with Fe3O4@h-
Pd@mCeO2 as catalyst. [d] DMF/H2O 9:1, 1008C. [e] Catalyst (0.1 mmol%
Pd), 1 h, TOF=990 mol product molÀ1 PdhÀ1
.
coupled efficiently with phenylboronic acid. The system at-
tained high conversion to the cross-coupled products within
three hours, regardless of the electron-withdrawing substituent
(ÀCOCH3), electron-donating group (ÀOCH3), or coordinating
group (ÀNH2) (entries 1–3 and 5–7). ortho-Methyl aryl iodide
gave inferior yields of the 2-arylation product (entry 4) due to
steric hindrance. Aryl chlorides could also be coupled with
phenylboronic acid using the mixed solvent of DMF and water
under a higher temperature (entries 8 and 9). Aryl chlorides
containing electron-withdrawing substituents gave an excel-
lent yield within three hours (entry 9). Aryl chlorides with elec-
tron-donating substituents gave inferior product yields
(entry 8).
Scheme 2. The proposed mechanism for the synergistic effect between the
CeO2 and Pd NPs.
spectroscopy (XPS) as shown in Figure 4) and the O vacancy
formed. The active OHdÀ groups were produced by means of
the adsorption and dissociation of H2O on the O-vacancy sites.
Afterwards, the electron pair of the Ce3+ cations and OHdÀ
groups were transferred to the Pd NPs as a result of the elec-
tron-donating effect. This increased the electron density of the
Pd NPs; this had the effect of accelerating the oxidative addi-
tion reaction, which was regarded as the rate-determining step
in the CÀC coupling reactions,[1] to form Ar–PdII–X. Our kinetic
investigations using Fe3O4@SiO2-Pd@mCeO2 as catalyst showed
a strong positive influence between the concentration of the
aryl halide and the initial rates, whereas no influence of the
phenylboronic acid (Figure S6a–d in the Supporting Informa-
tion) was detected. These observations demonstrate that the
Furthermore, various substituted arylboronic acid substrates
underwent the cross-coupling reaction with iodobenzene in
our catalytic system (entries 10–13). ortho-Methyl phenylboron-
ic acid gave inferior yields of the 2-arylation product (entry 10)
due to steric hindrance. Neither the electron-donating group
(4-OCH3) nor electron-withdrawing substituent (4-F) had this
inhibition effect under the same conditions (entries 12 and 13).
The reaction of iodobenzene and phenylboronic acid reached
full conversion in just one hour with a 0.1 mmol% Pd loading
(entry 14), and the TOF was 990 mol product molÀ1 PdhÀ1. The
results in Table 1 illustrate that our catalysts have excellent cat-
alytic activities and substrate compatibility with the same reac-
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
569
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
tion conditions, thus, the factors of reaction conditions and re-
actants for studying the reaction mechanism can be eliminat-
ed.
or Fe3O4@SiO2-Pd. This assertion is based on the following ob-
servations.
During the experiment, we removed the catalyst with
a magnet, and permitted the reaction to continue. The yield
stopped increasing immediately for the Fe3O4@SiO2-Pd@mCeO2
catalyst but not for the Fe3O4@h-Pd@mCeO2 or Fe3O4@SiO2-Pd
catalysts (see Figure 6b). If a trace amount of soluble leaching
Pd species was the active species, the catalytic yield would
have increased, even after the removal of catalyst. The hot-fil-
tration tests indicated that the catalysis is from the leached Pd
over Fe3O4@h-Pd@mCeO2 and over Fe3O4@SiO2-Pd, and unclear
In addition to prominent catalytic activity, isolation and recy-
clability of a catalyst are considerable requirements for any
practical catalytic reaction. In our systems, the catalysts were
readily recovered from their dispersion using an external
magnet. To test the recyclability of the catalyst, a benchmark
reaction was chosen as a comparison. As illustrated in Fig-
ure S7a, the Fe3O4@SiO2-Pd@mCeO2 catalyst could be recycled
by using selective magnetic separation up to ten times with-
out a notable decrease in yield. However, as revealed by Figur-
es S7b and c, the recycling test for Fe3O4@h-Pd@mCeO2 and
Fe3O4@SiO2-Pd indicated a decrease in catalytic efficacy after
the reaction was repeated five and three times, respectively.
Assessing the leaching of active Pd NPs into the reaction
mixture is another key factor in testing the stability of the Pd
catalyst. To verify the possibility of leaching, the three kinds of
catalyst were detected by inductively coupled plasma atomic
emission spectrometry (ICP-AES) after ten cycles, five cycles,
and three cycles. The results are listed in Table S5 in the Sup-
porting Information. The benchmark reaction shows a determi-
nable Pd loss for the Fe3O4@SiO2-Pd@mCeO2 catalyst. This
result might, to some extent, explain the slight yield decreases
following increased recycling. Figures S8a and S9b show the
SEM and TEM images of the Fe3O4@SiO2-Pd@mCeO2 catalyst
for the Suzuki reactions after ten runs of recycling. These
images indicate that the catalyst retained its spherical shape
after recycling. For Fe3O4@h-Pd@mCeO2 and Fe3O4@SiO2-Pd
catalysts, the leaching of Pd NPs is notable (Table S5). The SEM
image of the Fe3O4@SiO2-Pd catalyst (Figure S9a), compared
with the TEM image (Figure 2a), reveals that most of the Pd
NPs leached into the solvent. The SEM image (Figure S8b) and
TEM images (Figures S9c and d) of the Fe3O4@h-Pd@mCeO2
catalyst display that most of the outer CeO2 shell was dam-
aged, while only parts of the catalyst maintained a yolk–shell
structure. Prominently, the recyclability of the Fe3O4@SiO2-
Pd@mCeO2 catalyst was the best among the three. One could
easily ratiocinate that the hierarchical core–shell structures ef-
fectively hinder the aggregation and leaching of Pd NPs owing
to the robust protection provided by porous CeO2 shells. The
easy isolation and good reusability of the catalyst illustrate
that our catalysts have high separability and stability for study-
ing a reaction mechanism.
over Fe3O4@SiO2-Pd@mCeO2.
A positive hot-filtration test
means the occurrence of homogeneous catalysis, whereas
a negative hot-filtration test does not necessarily imply the
emergence of heterogeneous catalysis for a number of reasons
(e.g., fast deactivation or redeposition of soluble active spe-
cies). Consequently, the distinction between heterogeneous
and homogeneous catalytic activity cannot be declared solely
based on the results of our hot-filtration test.[63]
The inverse relationship between the active Pd concentra-
tion and the conversion rate is associated with the homogene-
ous mechanism.[64] The inverse relationship refers to the rela-
tionship between the active Pd concentration and turnover
number (TON) or turnover frequency (TOF). If the TON or TOF
increases with decreasing active Pd concentration, it means
that the homogeneous rather than heterogeneous catalysis
works for the Suzuki reaction.[65] To investigate this inverse rela-
tionship, we assessed the effect of Pd concentration on catalyt-
ic activity. Figure 6c–e depicts the reaction profiles of the cata-
lyst with a Pd concentration of 0.1, 0.5, 1, and 5 mmol%. The
TON and TOF for the different Pd concentrations are shown in
Table S6. It is also clear from Table S6 that the TOF increases
on going from 5 to 0.1 mmol% over Fe3O4@h-Pd@mCeO2 and
over Fe3O4@SiO2-Pd, and is inconspicuous over Fe3O4@SiO2-
Pd@mCeO2. This can be explained by the equilibrium between
the Pd present in the clusters, which formed from the soluble
leaching of Pd, and the Pd NPs involved in the catalytic
cycle.[66] At lower Pd concentrations, this equilibrium can shift
away from the inactive clusters, resulting in a higher percent-
age of active Pd catalyst. The Fe3O4@SiO2-Pd@mCeO2 catalyst
(Figure 6d) displayed an increased conversion rate when the
Pd concentration increased. Interestingly, the Fe3O4@h-
Pd@mCeO2 and Fe3O4@SiO2-Pd catalysts did not show the
same reactivity trend when varying the Pd concentration. As il-
lustrated in Figure 6c,e, both the reactions with a Pd concen-
tration of 1 mmol% were faster than those with higher Pd con-
centrations (5 mmol%). This deactivation of Pd catalyst, the so-
called “homeopathic” mechanism, happened as a result of the
soluble leaching Pd nucleating to form Pd clusters that contin-
ued to grow.[9] As the Pd concentration increased, greater
leaching occurred, and the quenching of the active Pd catalyst
in solution became more efficient.
Another important aspect of a catalyst is the nature of
active noble-metal species (homogeneous or heterogeneous)
in cross-coupling reactions. The literature differs as to whether
the Pd-catalyzed Suzuki–Miyaura reactions result from leaching
Pd NPs or the heterogeneous Pd NPs themselves. This has
been a topic of intense debate and is far from being rid of
controversy. Thus, we carried out a hot-filtration test, solid-
phase poisoning, and assessed Pd concentrations to validate
whether the Pd active sites were encapsulated by the CeO2
shell, or the homogeneous Pd species leached from the CeO2
shell during the reaction process were true catalysts. The re-
sults imply that Fe3O4@SiO2-Pd@mCeO2 completes the reaction
in a true heterogeneous manner, but not Fe3O4@h-Pd@mCeO2
The benchmark reaction was also conducted in the presence
of poly(4-vinylpyridine) (PVPy), which is a well-known solid
poison that restricts homogeneous Pd species in the liquid
phase through chelation.[14] A comparison of conversion clearly
illustrates that the catalytic efficacy of the Fe3O4@SiO2-
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
570
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Pd@mCeO2 catalyst was not affected by the addition of PVPy,
whereas the catalytic efficiency of Fe3O4@h-Pd@mCeO2 and
Fe3O4@SiO2-Pd was notably reduced (see Figure 6a,f). To obtain
more accurate and meaningful results on this issue, we also
carried out a mercury poisoning test, which can form an amal-
gam with leaching Pd nanoparticles.[67] Of note is that, as
pointed by Whitesides,[68] the mercury poisoning test can ac-
count for a homogeneous nature but not a heterogeneous
one. After Hg (300 equiv) addition, an inhibition of catalytic ac-
tivity was observed for Fe3O4@h-Pd@mCeO2 and Fe3O4@SiO2-
Pd but not for Fe3O4@SiO2-Pd@mCeO2 (Figure S10 in the Sup-
porting Information).
CeO2 and the active Pd NPs species, which resulted in the en-
hanced catalytic activity of our catalysts relative to some of the
published literature. The excellent stability of the Fe3O4@SiO2-
Pd@mCeO2 catalyst meant it could be recycled ten times with-
out any obvious decrease of yield.
Fundamentally, the hot-filtration heterogeneity test and Pd
concentration and solid-phase poisoning test in this study
were conducted to assess the heterogeneous or homogeneous
nature of the Pd catalyst. Our results demonstrate that the
active Pd in the Fe3O4@SiO2-Pd@mCeO2 catalyst operates by
a
heterogeneous mechanism, whereas the active Pd in
Fe3O4@h-Pd@mCeO2 and Fe3O4@SiO2-Pd catalysts abide by
a homogeneous mechanism. Our comparative study demon-
strates that understanding of the leaching processes is vital in
designing heterogeneous catalysts for cross-coupling, because
the definitions of heterogeneous/homogeneous are too broad
for diverse reaction systems.
Based on these heterogeneity tests, we ascribe different cat-
alytic mechanisms for the hierarchical core–shell structures and
the yolk–shell structures (Scheme 3). The active Pd of the
Fe3O4@SiO2-Pd@mCeO2 catalyst was supposed to be a hetero-
geneous mechanism for the Suzuki reaction. In this case, both
aryl halide and arylboronic acid come into contact by colliding
directly on the Pd NPs surface rather than through the leach-
ing of Pd NPs within the solution.
Our methodology provides a general way to design and as-
semble core–shell or yolk–shell nanostructures with mutable
morphologies and controllable component performances. It
raises the overall application potential of this type of nano-
composite. Moreover, it provides strong evidence validating
the “mechanistic nature” issue in Pd-catalyzed Suzuki–Miyaura
cross-coupling reactions.
Conclusion
In this study, we successfully prepared a hierarchical core–shell
Fe3O4@SiO2-Pd@mCeO2 catalyst and yolk–shell Fe3O4@h-
Pd@mCeO2 catalyst by integrating the hydrothermal method,
in situ interfacial deposition, and the sol–gel process. Both the
Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-Pd@mCeO2 catalysts ex-
hibited high levels of magnetism, accessible mesoporous CeO2,
and large dispersibility in the solvent. Greater catalytic per-
formance was demonstrated in the Suzuki–Miyaura cross-cou-
pling reaction than when using the Fe3O4@SiO2-Pd catalyst
without an outer CeO2 shell. We proposed a feasible mecha-
nism underlying the synergistic effect between the supporting
Experimental Section
Synthesis of core–shell Fe3O4@SiO2 and APTES-modified
Fe3O4@SiO2
The dispersible magnetic Fe3O4 modified by Na3Cit was synthesized
according to the study reported by Zhao.[46] We synthesized
Fe3O4@SiO2 by using a universal solution sol–gel method as fol-
lows: an aqueous dispersion of Fe3O4 (35 mL, 0.02 gmLÀ1) was
added to a reaction vessel that contained absolute ethanol
Scheme 3. The proposed mechanism for Fe3O4@SiO2-Pd@mCeO2 and Fe3O4@h-Pd@mCeO2.
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
571
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
(140 mL) and NH3·H2O (2.5 mL, 28 wt%), and was then mechanical-
ly stirred for 15 min at room temperature. Next, TEOS (2.0 mL) was
added drop-by-drop for one minute; the ensuing reaction was
maintained for eight hours. The resulting core–shell Fe3O4@SiO2
were extracted with a magnet and washed with water and ethanol
six times. The obtained products were dispersed in absolute IPA
with a content of about 1 wt% for next uses.
a Schlenk tube containing a magnetic stir bar. Solvent (5 mL) was
added. The mixture was stirred at 808C in air. We monitored the re-
action process by GC at fixed time intervals. After completion of
the reaction, the catalyst was separated from the mixture with
a magnet, washed several times with water and ethanol, and then
dried under vacuum at 608C overnight. The resulting solution was
extracted by EtOAc three times for GC-MS analysis. Our recycling
experiment followed the same procedure.
For surface modifications, we further diluted the previously ob-
tained IPA dispersion (20 mL) with IPA (50 mL) containing APTES
(1 mL) by means of ultrasonication. The ensuing solution was bub-
bled with N2 for thirty minutes, and then mechanically stirred at
708C for six hours. Finally, we collected the APTES-modified
Fe3O4@SiO2 with a magnet, and repeatedly washed them with de-
ionized water and ethanol.
Hot-filtration test
The hot-filtration test—a benchmark Suzuki–Miyaura reaction—
was conducted according to a typical procedure. First, the reaction
was allowed to proceed for about thirty minutes. The solid catalyst
was then separated from the mixture with a magnet while the
mother liquor stayed and reacted for an extended duration under
identical conditions. The conversion was monitored by GC at differ-
ent times.
Synthesis of core–shell Fe3O4@SiO2-Pd
We added APTES-modified Fe3O4@SiO2 (1.0 g) to toluene (40 mL)
containing PdCl2 (0.2 g). After mechanically stirring the solution for
five hours at room temperature, the resulting solid sample was ex-
tracted with a magnet and washed with toluene. Then, NaBH4
(50 mg) was employed to reduce the PdCl2 in a solvent of toluene
and ethanol (v/v 20:1). This suspension was isolated by a magnet,
washed with toluene and ethanol several times, and dried in
vacuum, which yielded the sample denoted Fe3O4@SiO2-Pd. To de-
termine the content of Pd (10.05 wt%), the product was decom-
posed in concentrated nitric acid and analyzed with ICP-AES.
The effect of Pd concentrations
This part of the experiment followed the previously mentioned
procedure, with Pd concentrations of 0.1, 0.5, 1, and 5 mmol%, re-
spectively.
Solid-phase poisoning test
The solid-phase poisoning test as a benchmark Suzuki–Miyaura re-
action was performed in the presence of poly(4-vinylpyridine) (2%
cross-linked) (PVPy) or mercury (1.5 mmol, 300 mg) using
300 equivalents to total palladium content. The reaction and analy-
sis conditions were as mentioned above.
Synthesis of core–shell Fe3O4@SiO2-Pd@mCeO2
Fe3O4@SiO2-Pd (150 mg) was dispersed in ethanol (120 mL) and so-
nicated for thirty minutes. We then added Ce(NO3)3·6H2O (200 mg)
to that solution and subjected it to an ultrasound treatment pro-
cess for fifteen minutes. Subsequently, deionized water (100 mL)
containing HMT (1.0 g) was added to the solution, and was sub-
jected to ultrasound for another ten minutes. We mechanically
stirred the solution for two hours at 708C. The resulting samples
were obtained with a magnet, and washed with water six times to
remove any possible ionic remnants. We vacuum-dried the prod-
ucts at 408C for twenty-four hours. The sample was calcined at
4008C for two hours to gain Fe3O4@SiO2-Pd@mCeO2 nanoparticles.
The product was decomposed in concentrated nitric acid for the
ICP-AES analysis to determine the content of Pd (2.85 wt%).
Acknowledgements
This study was supported by the National Natural Science Foun-
dation of China (51202034).
Keywords: cross-coupling · heterogeneous catalysis · kinetics ·
palladium · reaction mechanisms
[1] A. Balanta, C. Godard, C. Claver, Chem. Soc. Rev. 2011, 40, 4973–4985.
[2] J. Pinto, V. L. M. Silva, A. M. G. Silva, L. M. N. B. F. Santos, A. M. S. Silva, J.
Org. Chem. 2015, 80, 6649–6659.
The synthesis of Fe3O4@SiO2@mCeO2 was the same as that of
Fe3O4@ SiO2-Pd@mCeO2 without Pd.
[3] H. Noguchi, K. Hojo, M. Suginome, J. Am. Chem. Soc. 2007, 129, 758–
759.
[4] J. M. Chalker, C. S. C. Wood, B. G. Davis, J. Am. Chem. Soc. 2009, 131,
16346–16347.
Synthesis of yolk–shell Fe3O4@h-Pd@mCeO2
[5] S. Y. Ding, J. Gao, Q. Wang, Y. Zhang, W. G. Song, C. Y. Su, W. Wang, J.
Am. Chem. Soc. 2011, 133, 19816–19822.
[6] P. Taladriz-Blanco, P. HervØs, J. PØrez-Juste, Top. Catal. 2013, 56, 1154–
1170.
[7] L. Yin, J. Liebscher, Chem. Rev. 2007, 107, 133–173.
[8] . Molnµr, Chem. Rev. 2011, 111, 2251–2320.
[9] C. Deraedt, A. D. Didier, Acc. Chem. Res. 2014, 47, 494–503.
[10] N. Phan, M. Sluys, C. Jones, Adv. Synth. Catal. 2006, 348, 609–679.
[11] D. A. Conlon, B. Pipik, S. Ferdinand, C. R. LeBlond, J. R. Sowa, J. M. Wil-
liams, Y. Shi, Y. Sun, Adv. Synth. Catal. 2003, 345, 931–935.
[12] M. PØrez-Lorenzo, J. Phys. Chem. Lett. 2012, 3, 167–174.
[13] J. P. Huang, W. Wang, H. X. Li, ACS Catal. 2013, 3, 1526–1536.
[14] D. Snelders, G. Koten, R. Klein, J. Am. Chem. Soc. 2009, 131, 11407–
11416.
The annealed Fe3O4@SiO2-Pd@mCeO2 (0.2 g) was added to 10 wt%
NaOH solution (15 mL), which was stirred for four hours at 508C to
remove the SiO2 shell. The products were washed with deionized
water until they attained neutral pH levels, and were then dried
under vacuum at 608C. The Pd content was 3.05 wt% from the
analysis with ICP-AES.
Suzuki–Miyaura cross-coupling reactions
The catalysts with Pd content (0.5 mmol%) (20.60 mg for
Fe3O4@SiO2-Pd@mCeO2, 17.40 mg for Fe3O4@h-Pd@mCeO2, 5.30 mg
for Fe3O4@SiO2-Pd), potassium carbonate (2 mmol), aryl halide
(1.0 mmol), and arylboronic acid (1.2 mmol) were added to
[15] S. S. Ogasawara, S. Kato, J. Am. Chem. Soc. 2010, 132, 4608–4613.
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
572
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[16] R. Greco, W. Goessler, D. Cantillo, C. Kappe, ACS Catal. 2015, 5, 1303–
1312.
[43] L. Chen, L. Li, T. Wang, L. Zhang, S. Xing, C. Wang, Nanoscale 2014, 6,
6603–6608.
[17] C. Crudden, M. Sateesh, R. Lewis, J. Am. Chem. Soc. 2005, 127, 10045–
10050.
[18] B. Cornelio, A. R. Saunders, W. A. Solomonsz, M. Laronze-Cochard, A.
Fontana, J. Sapi, A. N. Khlobystov, G. A. Rance, J. Mater. Chem. A 2015,
3, 3918–3927.
[44] C. Wang, J. Chen, X. Zhou, Y. Deng, Nano Res. 2015, 8, 238–245.
[45] F. Liu, J. Wang, Q. Cao, H. Deng, G. Shao, B. Deng, W. Zhou, Chem.
Commun. 2015, 51, 2357–2360.
[46] Y. Deng, Y. Cai, Z. Sun, D. Zhao, J. Am. Chem. Soc. 2010, 132, 8466–
8473.
[19] Y. Wan, H. Y. Wang, M. Klingstedt, O. Terasaki, D. Y. Zhao, J. Am. Chem.
Soc. 2009, 131, 4541–4550.
[20] L. D. Pachón, G. Rothenberg, Appl. Organomet. Chem. 2008, 22, 288–
299.
[21] Z. Niu, Q. Peng, Z. Zhuang, W. He, Y. Li, Chem. Eur. J. 2012, 18, 9813–
9817.
[22] A. V. Gaikwad, A. Holuigue, M. B. Thathagar, J. E. Elshof, G. Rothenberg,
Chem. Eur. J. 2007, 13, 6908–6913.
[47] Y. Li, Z. Zhang, J. Shen, M. Ye, Dalton Trans. 2015, 44, 16592–16601.
[48] L. M. Misch, J. Kurzman, A. R. Derk, R. Seshadri, E. McFarland, G. Stucky,
Chem. Mater. 2011, 23, 5432–5439.
[49] G. Cheng, J. Zhang, Y. Liu, D. Sun, Chem. Commun. 2011, 47, 5732–
5734.
[50] T. J. Yao, T. Y. Cui, X. Fang, J. Wu, Nanoscale 2013, 5, 5896–5904.
[51] Q. Wang, Y. Li, B. Liu, Q. Dong, G. Xu, L. Zhang, J. Zhang, J. Mater. Chem.
A 2015, 3, 139–147.
[23] P. Fang, A. Jutand, Z. Q. Tian, C. Amatore, Angew. Chem. Int. Ed. 2011,
50, 12184–12188; Angew. Chem. 2011, 123, 12392–12396.
[24] A. K. Diallo, C. Ornelas, L. Salmon, J. R. Aranzaes, D. Astruc, Angew.
Chem. Int. Ed. 2007, 46, 8644–8648; Angew. Chem. 2007, 119, 8798–
8802.
[52] M. R. Nabid, Y. Bide, M. Niknezhad, ChemCatChem 2014, 6, 538–546.
[53] M. Shokouhimehr, K. Shin, J. Lee, M. Hackett, T. Hyeon, J. Mater. Chem. A
2014, 2, 7593–7599.
[54] F. Zhang, P. Wang, J. Koberstein, S. Khalid, S. Chan, Surf. Sci. 2004, 563,
74–82.
[25] Z. Kun Sun, J. Yang, J. Wang, W. Li, S. Kaliaguine, X. Hou, Y. Deng, D.
Zhao, J. Mater. Chem. A 2014, 2, 6071–6074.
[55] F. Larachi, J. Pierre, A. Adnot, A. Bernis, Appl. Surf. Sci. 2002, 195, 236–
250.
[26] M. B. Gawande, P. S. Branco, R. S. Varma, Chem. Soc. Rev. 2013, 42,
3371–3393.
[56] Q. Wang, W. Jia, B. Liu, A. Dong, X. Gong, C. Li, P. Jing, Y. Li, G. Xu, J.
Zhang, J. Mater. Chem. A 2013, 1, 12732–12741.
[27] Q. M. Kainz, O. Reiser, Acc. Chem. Res. 2014, 47, 667–677.
[28] D. Wang, D. Astruc, Chem. Rev. 2014, 114, 6949–6985.
[29] R. Ghosh Chaudhuri, S. Paria, Chem. Rev. 2012, 112, 2373–2433.
[30] T. Mitsudome, K. Kaneda, ChemCatChem 2013, 5, 1681–1691.
[31] M. B. Gawande, Y. Monga, R. Zboril, R. K. Sharma, Coord. Chem. Rev.
2015, 288, 118–143.
[32] R. Li, Y. Huang, P. Zhang, H. Zhong, Q. Chen, J. Mater. Chem. 2012, 22,
22750–22755.
[33] Z. Zhang, C. Zhang, J. Sun, T. Kou, Q. Bai, Y. Wang, Y. Ding, J. Mater.
Chem. A 2013, 1, 3620–3628.
[57] J. Liu, Z. Sun, Y. Deng, Y. Gao, F. Li, D. Zhao, Angew. Chem. Int. Ed. 2009,
48, 5875–5879; Angew. Chem. 2009, 121, 5989–5993.
[58] A. S. Diez, M. G. Mayer, M. A. Volpe, Appl. Catal. A 2014, 482, 24–30.
[59] J. L. Shi, Chem. Rev. 2013, 113, 2139–2181.
[60] C. Sun, H. Li, L. Chen, Energy Environ. Sci. 2012, 5, 8475–8505.
[61] H. Y. Kim, H. M. Lee, G. Henkelman, J. Am. Chem. Soc. 2012, 134, 1560–
1570.
[62] S. Zhang, J. Li, W. Gao, Y. Q. Qu, Nanoscale. 2015, 7, 3016–3021.
[63] V. Farina, Adv. Synth. Catal. 2004, 346, 1553–1582.
[64] A. H. M. De Vries, J. M. C. A. Mulders, J. H. M. Mmmers, H. J. W. Hender-
ickx, J. G. de Vries, Org. Lett. 2003, 5, 3285–3288.
[65] L. Huang, P. K. Wong, J. Tan, T. P. Ang, Z. Wang, J. Phys. Chem. C 2009,
113, 10120–10130
[34] H. Yang, G. Li, Z. Ma, J. Mater. Chem. 2012, 22, 6639–6648.
[35] M. Zhu, C. Wang, D. Meng, G. Diao, J. Mater. Chem. A 2013, 1, 2118–
2125.
[36] J. Liu, S. Qiao, J. S. Chen, X. W. Lou, X. Xing, G. Q. Lu, Chem. Commun.
2011, 47, 12578–12591.
[66] G. Collins, M. Schmidt, C. Dwyer, G. McGlacken, J. D. Holmes, ACS Catal.
2014, 4, 3105–3111.
[37] J. Qi, X. Lai, J. Wang, H. Tang, H. Ren, Y. Yang, Q. Jin, L. Zhang, R. Yu, G.
Ma, Z. Su, H. Zhao, D. Wang, Chem. Soc. Rev. 2015, 44, 6749–6773.
[38] Q. Yue, Y. Zhang, C. Wang, X. Wang, Z. Sun, X. Hou, D. Zhao, Y. Deng, J.
Mater. Chem. A 2015, 3, 4586–4594.
[67] C. S. Consorti, F. R. Flores, J. Dupont, J. Am. Chem. Soc. 2005, 127,
12054–12065.
[68] G. M. Whitesides, M. Hackett, R. L. Brainard, A. N. Izumi, S. S. Moore,
D. W. Brown, E. M. Staudt, Organometallics 1985, 4, 1819–1830.
[39] J. Xu, J. Liu, R. Che, Nanoscale 2014, 6, 5782–5790.
[40] W. Liu, Y. Liu, X. Yan, G. Yong, Y. Xu, S. Liu, J. Mater. Chem. A 2014, 2,
9600–9606.
Manuscript received: March 1, 2016
Revised: April 22, 2016
[41] T. Zeng, X. Zhang, S. Wang, Y. Ma, H. Niu, Y. Cai, J. Mater. Chem. A 2013,
1, 11641–11647.
Accepted Article published: April 28, 2016
Final Article published: May 27, 2016
[42] X. Li, F. Gai, B. Guan, Y. Zhang, Y. Liu, Q. Huo, J. Mater. Chem. A 2015, 3,
3988–3994.
ChemPlusChem 2016, 81, 564 – 573
www.chempluschem.org
573
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim