(Ni,Mg)3Si2O5(OH)4 solid-solution nanotubes supported by sub-0.06 wt % palladium as a robust high-efficiency catalyst for Suzuki-Miyaura cross-coupling reactions

Article abstract of DOI:10.1021/ic2024378

(Ni_1-x,Mg_x)₃Si₂O ₅(OH)₄ solid-solution nanotubes (NTs) with tunable compositions were hydrothermally synthesized by altering the molar ratio of Mg²⁺ to Ni²⁺. The as-synthesized NTs were loaded with sub-0.06 wt % palladium (Pd; ～0.045 wt %) for Suzuki-Miyaura (SM) coupling reactions between iodobenzene or 4-iodotoluene and phenylboronic acid. The (Ni,Mg)₃Si₂O₅(OH)₄ (Mg ²⁺:Ni²⁺ = 1.0:1.0) NTs supported by 0.045 wt % Pd promoted the iodobenzene-participated coupling reaction with a high yield of >99%, an excellent recycling catalytic performance during 10 cycles of catalysis with yields of ～99%, and also an extremely low Pd releasing level of ～0.02 ppm. High-activity Pd and PdO clusters, multitudes of dislocations, and defects and terraces contained within the NTs should contribute to the (Ni,Mg) ₃Si₂O₅(OH)₄ (Mg²⁺:Ni ²⁺ = 1.0:1.0) NTs supported by 0.045 wt % Pd as a robust, reusable, and high-efficiency catalyst for SM coupling reactions with an extremely low Pd releasing level. The present hydrothermally stable (Ni,Mg)₃Si ₂O₅(OH)₄ (Mg²⁺:Ni²⁺ = 1.0:1.0) solid-solution silicate NTs provided an ideal alternative tubular-structured support for noble- or transition-metal catalysts with low Pd loading, good recycling, and extremely low ppb levels of Pd release, which could also be extended to some other SM coupling reactions.

Full text of DOI:10.1021/ic2024378

Article
pubs.acs.org/IC
(Ni,Mg)3Si2O5(OH)4 Solid-Solution Nanotubes Supported by Sub-0.06
wt % Palladium as a Robust High-Efficiency Catalyst for Suzuki−
Miyaura Cross-Coupling Reactions
Wancheng Zhu, Yan Yang, Shi Hu, Guolei Xiang, Biao Xu, Jing Zhuang, and Xun Wang*
Department of Chemistry, Tsinghua University, Beijing, 100084, People's Republic of China
S
* Supporting Information
ABSTRACT: (Ni1−x,Mgx)3Si2O5(OH)4 solid-solution nano-
tubes (NTs) with tunable compositions were hydrothermally
synthesized by altering the molar ratio of Mg2+ to Ni2+. The as-
synthesized NTs were loaded with sub-0.06 wt % palladium
(Pd; ∼0.045 wt %) for Suzuki−Miyaura (SM) coupling
reactions between iodobenzene or 4-iodotoluene and phenyl-
boronic acid. The (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0)
NTs supported by 0.045 wt % Pd promoted the iodobenzene-
participated coupling reaction with a high yield of >99%, an
excellent recycling catalytic performance during 10 cycles of
catalysis with yields of ∼99%, and also an extremely low Pd
releasing level of ∼0.02 ppm. High-activity Pd and PdO clusters, multitudes of dislocations, and defects and terraces contained
within the NTs should contribute to the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported by 0.045 wt % Pd as a
robust, reusable, and high-efficiency catalyst for SM coupling reactions with an extremely low Pd releasing level. The present
hydrothermally stable (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) solid-solution silicate NTs provided an ideal alternative
tubular-structured support for noble- or transition-metal catalysts with low Pd loading, good recycling, and extremely low ppb
levels of Pd release, which could also be extended to some other SM coupling reactions.
pharmaceutical industry.10 To cope with these burning issues,
INTRODUCTION
■
heterogeneous Pd catalysis has been a promising option,8,10 and
In the past few decades, palladium (Pd)-catalyzed carbon−
carbon (C−C) coupling reactions including Heck,1−3 Ne-
recyclable/reusable Pd catalysts have emerged as an alternative
competitive solution.15 For heterogeneous Pd catalysis, Pd is
generally immobilized on a solid support, and the particle size,
surface area, pore structure, and acid−base properties of the
support usually have an impact on the activity of the catalytic
system.8 Up to now, great efforts have been made to develop
new catalyst carriers and high-activity Pd species for the
referred SM reactions. For example, activated carbon
(charcoal),16 zeolites, molecular sieves, metal oxides (silica,
alumina, MgO, ZnO, etc.), layered clays,17 alkaline-earth
carbonates (CaCO3, BaCO3, SrCO3, etc.), and organic
polymers,18 etc., have been found as effective Pd supports. In
many cases, however, the hydrothermal stability of the support,
e.g., mesoporous silica with amorphous walls, still remained a
great challenge. The amorphous walls of the support could
easily be destroyed by the base used for the reaction, resulting
in a collapse of the mesoporous structure. In addition, a 5 wt %
or even higher Pd loading is generally used for most of the Pd
catalysts.8 Pd on charcoal (Pd/C) is by far the most often used
heterogeneous Pd catalyst for coupling reactions, and the
commercially available Pd/C catalysts contain a Pd content of
1−20%. Nonetheless, the Pd/C catalyst suffers from relatively
gishi,4,5 Sonogashira,6,7 etc., have been paid worldwide
attention and employed versatilely for the distinguished
cornerstone position within modern organic synthesis.
Among the various C−C couplings, Suzuki−Miyaura (SM)
couplings have attracted particular concerns, have been
considered as one of the most efficient synthetic routes to
biaryl compounds, and have thus been extensively employed in
academic laboratories as well as pharmaceutical and fine-
chemical industries.8−11 Compared with other C−C bond
construction reactions, the mainstay and popularity of SM
couplings was based on several unique advantages,12,13 such as
the required relatively mild reaction conditions, commercial
availability of the starting diverse boronic acids, tolerance of a
variety of functional groups as starting partners, ease of
handling/removal of boron-bearing byproducts, and in some
cases environmental benign character for accessibility in
aqueous media.10,14
Homogeneous Pd-catalyzed C−C coupling reactions have
been extensively studied for their excellent activity and
selectivity, which, however, suffers from a number of draw-
backs, especially in the removal and reuse of the catalyst. This
leads to the unavoidable loss of expensive metal and ligands and
also resultant product contamination, which has emerged as an
acute issue for large-scale synthesis, especially in the
Received: November 13, 2011
Published: May 14, 2012
© 2012 American Chemical Society
6020
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
previously collected NTs supported by sub-0.06 wt % Pd, 15.0 mL of
ethanol, and 1.0 mmol of iodobenzene were mixed within a three-neck
flask, then heated to 70 °C, and kept in an isothermal state for 3.0 h,
within a preheated oil bath. After the reaction, the liquid product phase
was separated from the solid phase by centrifugation (10000 rpm, 20.0
min) and set aside for further characterization. For the specific
(Ni,Mg)3Si2O5(OH)4 (molar ratio of Mg2+:Ni2+ = 1.0:1.0) NTs, to
evaluate the recycling catalytic performance, the solid catalyst phase
after the previous centrifugation was washed with 40.0 mL of DI water
to eliminate K2CO3, separated by centrifugation, then washed with
20.0 mL of ethanol to substitute remanent DI water, separated by
centrifugation, and finally transferred to the next loop of catalysis by
adding 15.0 mL of ethanol, followed by introduction of the same
amount of fresh phenylboronic acid, iodobenzene, and K2CO3, under
the same reactant conditions. In addition, for the NTs with a molar
ratio of Mg2+:Ni2+ = 1.0:1.0, a second SM coupling reaction between
4-iodotoluene (1.0 mmol) and phenylboronic acid (1.5 mmol) was
also carried out under the same conditions as those described
previously, using the NTs supported by sub-0.06 wt % Pd as the
catalyst. For comparison, a blank test corresponding to the first SM
coupling reaction was performed using the same amount of bare
(Ni,Mg)3Si2O5(OH)4 (molar ratio of Mg2+:Ni2+ = 1.0:1.0) NTs
without Pd loading as the catalyst, with other conditions unchanged.
To investigate the possible migration and mechanism of the trace Pd
during the SM coupling reaction, the reaction system was interrupted
when proceeded at 70 °C for ca. 0.5 h and then quenched by room
temperature water or cooled naturally to room temperature by raising
the flask out of the oil bath in due course, followed by centrifugation as
previously.
4. Mercury Poisoning Experiment. The mercury poisoning test
was carried out in the process of catalysis, with (Ni,Mg)3Si2O5(OH)4
(molar ratio of Mg2+:Ni2+ = 1.0:1.0) NTs supported by sub-0.06 wt %
Pd as the catalyst. All reactants, base, solvent, and NTs supported by a
catalyst were reduced by half of those employed in the precedent
typical SM coupling reaction. When the catalysis proceeded at 70 °C
for 15.0 min with an approximate conversion of ca. 15−50%, the flask
was raised out of the oil bath and quenched by room temperature
water (∼22 °C) rapidly. After centrifugation, the supernatant was
sampled ∼500 μL for gas chromatography (GC)-mass spectrometry
(MS), then the residual catalytic system was shaked up and thoroughly
transferred to the flask and added by a mercury bead with a diameter
of ∼2.6 mm. The flask was sealed and vigorously stirred magnetically
within a working fume hood for 12.0 h, to ensure excellent mixing.
Then the flask was recovered to 70 °C for another 2.0 h and 45.0 min.
Finally, the flask was cooled naturally to room temperature, and the
liquid product phase was collected after centrifugation, with the
remanent solid phase including mercury beads deposited on the
bottom of the flask post-treated with sulfur for security.
5. Characterization. The structure of the sample was identified by
a powder X-ray diffractometer (D8-Advance, Bruker, Germany) using
Cu Kα radiation (λ = 1.54178 Å) and a fixed power source (40.0 kV;
40.0 mA). The morphology and microstructure of the samples were
examined by a high-resolution transmission electron microscope
(JEM-2010, JEOL, Japan, at 120.0 kV). High-angle annular dark-field
(HAADF) scanning transmission electron microscopy (STEM) and
elemental mapping were determined by a high-resolution transmission
electron microscope (Tecnai G2 F20 S-Twin, FEI, USA, at 200.0 kV)
equipped with a Fischione accessory. Optical properties were
examined by a UV−vis spectrophotometer (UV-3600 230VCE,
Shimadzu, Japan) with an integrating sphere accessory. The actual
amount of Pd loaded on the NTs and also that of the liquid product
phase derived from the first SM coupling reaction were determined by
inductively coupled plasma mass spectrometry (ICP-MS) analysis
(XSeries 2, Thermo Fisher, USA). N2 adsorption−desorption
isotherms were measured using a chemisorption−physisorption
analyzer (Autosorb-1, Quantachrome, USA) at 77 K after the samples
had been outgassed at 200 °C for 120 min. The specific surface area
was calculated from the adsorption branches in the relative pressure
range of 0.10−0.31 using the multipoint Brunauer−Emmett−Teller
(BET) method, and the pore-size distribution was evaluated from the
severe Pd leaching; e.g., leaching up to 14% Pd was observed in
Heck reactions.8 Recently, novel Pd species have also been
reported, such as Pd nanoparticles (NPs),9,19 colloidal Pd,20,21
and Pd clusters.22,23 Reducing the Pd employed by minimizing
Pd leaching and residual species within the product is of great
significance especially for pharmaceutical and fine-chemical
synthesis. To date, it is still a challenge to develop novel
support for Pd as a high-efficiency catalyst for heterogeneous
C−C cross-coupling reactions with low loading, high activity,
and low leaching.
In this work, (Ni1−x,Mgx)3Si2O5(OH)4 solid-solution nano-
tubes (NTs) with tunable compositions were hydrothermally
synthesized, among which the (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs were loaded with sub-0.06 wt %
Pd (∼0.045 wt %) for the SM coupling reactions between
iodobenzene or 4-iodotoluene and phenylboronic acid. The low
level of Pd loading, the low ppb level of Pd release, and the
excellent hydrothermal stability of the nanotubular support
without destruction after 10 cycles of high-efficiency catalysis
definitely indicated that, the as-synthesized solid-solution NTs
supported by 0.045 wt % Pd were an ideal robust catalyst for
the selected SM coupling reactions, which could also be
extended as a competitive reusable high-efficiency catalyst for
some other SM coupling reactions.
EXPERIMENTAL SECTION
■
1. Synthesis of (Ni1−x,Mgx)3Si2O5(OH)4 NTs. All reagents were of
analytical grade and were used directly without further purification.
(Ni,Mg)3Si2O5(OH)4 NTs (Mg2+:Ni2+ = 1.0:1.0) were synthesized by
a facile hydrothermal route. In a typical procedure, 1.0 mmol of
Mg(NO3)2·6H2O and 1.0 mmol of NiCl2·6H2O were dissolved into 30
mL of deionized (DI) water under constant magnetic stirring at room
temperature. A total of 15.39 mL of liquid Na2O·nSiO3 (n = 3.1−3.4)
was dissolved and adjusted to 100 mL of mixed solution. Then 4.0 mL
of the previous solution was added to the preprepared magnetically
stirred system, resulting in a gray-green slurry. After ca. 10 min of
stirring, 5.8 g of NaOH was poured into the resultant slurry, with the
color changing to a grayer green. Stirred for another 10 min, the as-
obtained slurry was transferred to a Teflon-lined stainless steel
autoclave with a capacity of 45.0 mL. The autoclave was sealed, heated
to 210 °C (heating rate: 5 °C min−1), and kept in an isothermal state
for 24.0 h. After hydrothermal synthesis, the autoclave was cooled to
room temperature naturally, and the as-synthesized precipitate was
washed with DI water and separated by centrifugation (10000 rpm,
20.0 min) three times and then dried at 80 °C for 24.0 h for further
characterization and evaluation. To evaluate the effect of the relative
amount of Mg2+ and Ni2+ on the composition and morphology of the
hydrothermal product, the molar ratios of Mg2+ and Ni2+ were altered
within the range of 0:2.0 to 2.0:0, keeping the whole amount of Mg2+
and Ni2+ as 2.0 moles and other conditions unchanged.
2. Loading of sub-0.06 wt % Pd. Typically, 0.2 g of the as-
synthesized NTs was added into the solution containing 20.0 mL of
ethanol and 30.0 mL of DI water within a three-neck flask equipped
with a condenser under constant magnetic stirring. The flask was
heated within a preheated oil bath to the boiling state (ca. 110 °C),
followed by a rapid injection of 22.54 μL of a PdCl2 solution (0.05 mol
L−1), and refluxed for 0.5 h in the absence of protection of inert gas,
with a theoretical 0.06 wt % Pd versus the NTs to be loaded. Then the
system was stopped and separated by centrifugation (10000 rpm, 20.0
min), to the liquid phase was added ca. 0.4 g of NaBH4 to monitor
whether all of the Pd2+ ions have been reduced, and the solid phase
(i.e., NTs supported by the sub-0.06 wt % Pd catalyst used hereafter)
was washed with ethanol for a second time and finally collected for
catalytic evaluation.
3. Catalytic Performance of the (Ni1−x,Mgx)3Si2O5(OH)4 NTs
Supported by Sub-0.06 wt % Pd. In a first typical SM coupling
reaction, 1.5 mmol of phenylboronic acid, 2.0 mmol of K2CO3,
6021
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
Figure 1. XRD patterns (a), TEM images (b−g), digital pictures (h), and UV−vis diffuse-reflectance spectra (i) of the hydrothermally synthesized
(Ni1−x,Mgx)3Si2O5(OH)4 solid-solution NTs, with the molar ratios of Mg2+:Ni2+ = 0:2.0 (a1, b, c, h1, i1); 0.5:1.5 (a2, d, h2, i2); 1.0:1.0 (a3, e, e1, h3, i3);
1.5:0.5 (a4, f, h4, i4); 2.0:0 (a5, g, h5, i5). Composition: (a1) Ni3Si2O5(OH)4 (JCPDS No. 49-1859); (a2−a4) (Ni,Mg)3Si2O5(OH)4 (JCPDS No. 25-
0524); (a5) Mg3Si2O5(OH)4 (JCPDS No. 52-1562). Mass of the samples within each vial (g): 0.900.
N2 desorption isotherm using the Barrett−Joyner−Halenda (BJH)
method. Qualitative analysis of the catalytic liquid phase was carried
out on a gas chromatograph−mass spectrometer (GCMS-QP 2010 SE,
Shimadzu, Japan), and the yield of the product biphenyl or 4-
methylbiphenyl was determined with GC−MS measurements
(Thermo Scientific DSQ, USA) by the selected ion monitoring
method with iodobenzene included, using n-dodecane as the internal
standard and ethanol as the solvent. The valence state of the element
was determined by an X-ray photoelectron spectroscopy (XPS)
microscope (ESCALAB250, Thermofisher, U.K., at 15 kV and 150
W). The intermediate liquid phase quenched during the first SM
coupling reaction was concentrated by a rotary evaporator and
monitored by an electron-spin resonance spectrometer (JES-FA200,
JEOL, Japan), using the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0)
NTs supported with or without sub-0.06 wt % Pd as the catalyst.
for 24.0 h can also be tuned, as shown in Figure 1. When
Mg2+:Ni2+ = 0:2.0, the product was composed of the pure phase
of Ni3Si2O5(OH)4 (JCPDS No. 49-1859) with high crystallinity
(Figure 1a1) and distinct tubular structure (Figure 1b,c),
containing an inner diameter of 10−15 nm and an outer
diameter of 20−30 nm, similar to the literature results.24 The
interplanar spacing of 0.753 nm detected from the legible lattice
fringes along the axis direction of the as-obtained NTs was
quite similar to that of the standard value for the (002) planes
of monoclinic Ni3Si2O5(OH)4, indicating the axial direction of
the Ni3Si2O5(OH)4 NTs parallel to the (002) planes (Figure
1c). As the molar ratios of Mg2+:Ni2+ increased to 0.5:1.5 and
1.0:1.0, the product was constituted of the orthorhombic solid-
solution phase of (Ni,Mg)3Si2O5(OH)4 (JCPDS No. 25-0524)
with gradually decreased crystallinity (Figure 1a2−a3) and also a
tubular structure (Figure 1d,e), bearing an inner diameter of
10−15 nm and an outer diameter of 20−30 nm. The
interplanar spacings of 0.741 nm also confirmed the axial
direction of the NTs along the (002) planes.
RESULTS AND DISCUSSION
■
1. Composition, Microstructure, and Optical Absorb-
ance Property of the As-Synthesized NTs. Keeping the
whole amount of Mg2+ and Ni2+ as 2.0 moles and tuning the
molar ratios of Mg2+:Ni2+ within the range of 0:2.0−2.0:0, the
composition, morphology, color, and optical absorbance
property of the hydrothermal products obtained at 210 °C
Compared with the previous NTs, when the molar ratio of
Mg2+:Ni2+ was 1.0:1.0, the as-obtained NTs (Figure 1e) were
unstable under electron-beam irradiation even with a
6022
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
magnification as 40 K for a short time, resulting in porous
structures (Figure 1e1) and further failure to acquire the high-
resolution transmission electron microscopy (HRTEM) image
with legible lattice fringes. When the molar ratio of Mg2+:Ni2+
was 1.5:0.5, the hydrothermal product was also the
(Ni,Mg)3Si2O5(OH)4 phase with ordinary crystallinity (Figure
1a4) and a tubular structure (Figure 1f), containing inner and
outer diameters similar to the former ones. However, the
detected interplanar spacings between lattice fringes along the
axis of the NTs were slightly broadened to be 0.756 nm because
of the successive increase of the molar ratio of Mg2+:Ni2+. With
the molar ratio of Mg2+:Ni2+ further increased to 2.0:0, the
product was proven to be the pure phase of monoclinic
Mg3Si2O5(OH)4 (JCPDS No. 52-1562) with relatively low
crystallinity (Figure 1a5) and also a tubular structure (Figure
1g).25 The as-obtained NTs were also not stable under
electron-beam irradiation and tended to be porous.
On the other hand, with the molar ratios of Mg2+:Ni2+
increasing from 0:2.0 to 2.0:0, the color of the hydrothermally
synthesized NTs changed from the original bottle green to
bright green to light green and finally to white, as shown in
Figure 1h. Meanwhile, different samples of the same mass
(0.900 g) within the same sized vials demonstrated different
heights because of the difference of the bulk density caused by
the various porous structures of the NTs. The UV−vis diffuse-
reflectance spectra (Figure 1i) showed that when the molar
ratio of Mg2+:Ni2+ = 0:2.0, 0.5:1.0, 1.0:1.0, and 1.5:0.5, the as-
synthesized NTs exhibited distinct absorption within the near-
UV region concentrated on the wavelength of 380 nm and also
wider absorption within the visible and near-IR regions from
550 to 800 nm (Figure 1i1−i4). In accordance with the color
changes of the NTs (Figure 1h1−h4), the UV−vis spectra also
revealed some slight differences. When Mg2+:Ni2+ was
increased to 1.5:0.5, the absorption within the two regions
Figure 2. HAADF-STEM (a−c) images and corresponding elemental
mappings (a1−a4, b1−b4, c1−c4) of the hydrothermally synthesized
(Ni1−x,Mgx)3Si2O5(OH)4 solid-solution NTs loaded without (a and c)
and with (b) sub-0.06 wt % Pd before catalysis, with the molar ratios
of Mg2+:Ni2+ = 0.5:1.5 (a and a1−a4), 1.0:1.0 (b and b1−b4), and
1.5:0.5 (c and c1−c4). Elemental mapping: O−K (red, a1, b1, c1); Mg−
K (orange, a2, b2, c2); Si−K (yellow, a3, b3, c3); Ni−K (green, a4, b4,
c4).
tended to be weaker (Figure 1i4). Notably, when Mg2+:Ni2+
=
2.0:0, i.e., pure Mg3Si2O5(OH)4 NTs (Figure 1i5), no distinct
absorption was observed within the whole wavelength range
tested, indicating the quasi-transparent characteristic of the
Mg3Si2O5(OH)4 NTs within the UV−vis region.26 Confirmed
by the successive conversion of the X-ray diffraction (XRD)
patterns, microstructures, colors, and UV−vis spectra, the
compositions of the as-synthesized (Ni1−x,Mgx)3Si2O5(OH)4
solid-solution NTs could thus be facilely tuned by altering the
molar ratio of Mg2+:Ni2+ from 0:2.0 to 2.0:0.
Figure 2 reveals the HAADF-STEM images and correspond-
ing elemental mappings of the hydrothermally synthesized
solid-solution NTs, with molar ratios of Mg2+:Ni2+ = 0.5:1.5
(Figure 2a,a1−a4), 1.0:1.0 (Figure 2b,b1−b4), and 1.5:0.5
(Figure 2c,c1−c4). The dark-field STEM images clearly
reconfirmed the tubular structures of the as-synthesized NTs,
with different resolutions due to various crystallinities (Figure
2a−c). The higher the crystallinity of the NTs, the higher the
resolution of the STEM image, in accordance with the XRD
and TEM characterizations to some extent (Figure 1a2−a4,d−
f). For all referred NTs, constitutional elements such as O, Mg,
Si, and Ni were homogeneously dispersed within the bulk phase
of the NTs. With the molar ratio of Mg2+:Ni2+ increasing from
0.5:1.5 to 1.5:0.5, the orange mapping signals of Mg−K tended
to be stronger (Figure 2a2,b2,c2) and the green mapping signals
of Ni−K became weaker (Figure 2a4,b4,c4). However, although
the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs have been
loaded with sub-0.06 wt % Pd, neither the remarkable mapping
signals of Pd−K nor those of Pd−L were detected for such a
low loading (Figure 2b), which was below the detection limit of
the apparatus. Apparently, the HAADF-STEM images and
especially elemental mappings reconfirmed the successive
conversion and facile tuning of the present
(Ni1−x,Mgx)3Si2O5(OH)4 solid-solution NTs, creating a tunable
microenvironment containing different amounts of Ni and Mg.
2. Porous Structures of the (Ni1−x,Mgx)3Si2O5(OH)4
NTs. Figure 3 shows the N2 adsorption−desorption isotherms
and corresponding pore diameter distribution of the NTs, and
Table 1 further summarizes the porous structures of the NTs.
The N2 adsorption−desorption isotherms displayed a very
similar type IV with an H3-type hysteresis loop at a high
relative pressure (Figure 3a−e). The present hysteresis loops
observed at high relative pressures were probably due to the
textural mesoporosity among the randomly aggregated NTs.
Meanwhile, the BJH desorption pore diameter distribution
curves revealed a relatively broad distribution characteristic
(Figure 3a1−e1), as confirmed in Table 1. All of the NTs had
relatively large multipoint BET specific surface areas within the
range of 100−210 m2 g− 1
. Particularly, the
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs exhibited
6023
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
Figure 3. N2 adsorption−desorption isotherms (a−e) and pore diameter distribution (a1−e1) of the hydrothermally synthesized
(Ni1−x,Mgx)3Si2O5(OH)4 solid-solution NTs, with molar ratios of Mg2+:Ni2+ = 0:2.0 (a and a1), 0.5:1.5 (b and b1), 1.0:1.0 (c and c1), 1.5:0.5 (d
and d1), and 2.0:0 (e and e1).
Table 1. Porous Structures of the Hydrothermally Synthesized (Ni1−x,Mgx)3Si2O5(OH)4 Solid-solution NTs, with Molar Ratios
of Mg2+:Ni2+ = 0:2.0 (a), 0.5:1.5 (b), 1.0:1.0 (c), 1.5:0.5 (d), and 2.0:0 (e)
a
b
c
d
e
specific surface area (m2 g−1
total pore volume (cm3 g−1
average pore diameter (nm)
pore diameter distribution (nm)
)
101.5
0.523
20.60
113.8
0.754
26.52
152.3
1.107
29.08
165.47
0.757
14.36
209.1
1.217
23.29
)
3−7, 10−50
3−8, 10−100
5−7, 10−20, 20−100
4−6, 7.5−12, 14−70
4−6, 7−12, 15−90
not only large specific surface areas but also large total pore
volumes and average pore diameters. The relatively large
variation in the total pore volume between samples was
probably associated with the difference of the stability or
crystallinity of the NTs (Table 1). Mesopores with diameters
within the range of 10−20 nm were due to the inner spaces of
the tubular structures. In contrast, bigger mesopores larger than
20 nm or even macropores larger than 50 nm corresponded to
those existing among the randomly stacked NTs. Obviously,
the as-synthesized NTs exhibited some distinct characteristics,
such as tunable compositions, relatively large specific surface
area, total pore volumes, average pore diameters, and relatively
broad pore diameter distribution. This indicated the as-
synthesized (Ni1−x,Mgx)3Si2O5(OH)4 NTs as potential ideal
supports for the catalysts, especially nanocatalysts.
3. Single-Pass and Recycling Catalytic Performances
of the (Ni1−x,Mgx)3Si2O5(OH)4 NTs Supported by Sub-
0.06 wt % Pd. The as-synthesized different types of NTs were
loaded with sub-0.06 wt % Pd as the catalyst for the selected
SM coupling reaction between iodobenzene and phenylboronic
acid:
After the reduction of Pd2+ by ethanol, the liquid phase was
separated by centrifugation and then enough NaBH4 was
added. As a result, no black precipitate was found, indicating
the previously complete reduction of Pd2+ and also successful
loading of Pd onto the NTs. Variation of the yield of biphenyl
with different molar ratios of Mg2+:Ni2+ for the hydrothermally
6024
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
Figure 4. Single-pass (a) and recycling (b) catalytic performances of the as-synthesized (Ni1−x,Mgx)3Si2O5(OH)4 NTs supported by sub-0.06 wt %
Pd for the SM coupling reaction between iodobenzene (a1 and b) or 4-iodotoluene (a2) and phenylboronic acid, with molar ratios of Mg2+:Ni2+
=
0:2.0−2.0:0 (a) and 1.0:1.0 (b).
synthesized NTs is shown in Figure 4a1, with the same amount
of NTs supported by Pd as the catalyst (denoted as the left Y
axis). When Mg2+:Ni2+ = 0:2.0, 0.5:1.5, 1.0:1.0, 1.5:0.5, and
finally 2.0:0, the yield changed from 88.4% to 87.4% to 99.1%
(peak value) and then significantly decreased to 81.2% and
ultimately to 74.1%. Apparently, (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs supported by sub-0.06 wt % Pd
performed best as the catalyst for the selected coupling reaction
between iodobenzene and phenylboronic acid.
To further evaluate the catalytic activity of the hydro-
thermally synthesized NTs supported by sub-0.06 wt % Pd, a
second SM coupling reaction between 4-iodotoluene and
phenylboronic acid was also performed as follows:
centrifugation, washed with DI water, separated again, then
washed with ethanol, and finally participated in the next loop of
catalysis, the yield of biphenyl did not drop significantly. As a
matter of fact, even though the original 0.2 g of
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported
by sub-0.06 wt % Pd was collected as 0.10 g after the 9th cycle
and 0.09 g after 10th cycle of catalysis, the yield of biphenyl
fluctuated within the range of 98.5−99.4%, during the 10 cycles
of catalysis.
The specific actual amount of Pd loading on the
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs was further
determined by ICP-MS, and the details of the sample
preparation as well as the analytical results were also given, as
shown in Table S1 (Supporting Information, SI). The actual Pd
amounts loaded on the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+
=
1.0:1.0) NTs before the 1st cycle and after the 10th cycle of
catalysis were 0.0449 and 0.0447 wt %, respectively. Thus, on
the one hand, the specific Pd loading of the
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported
by sub-0.06 wt % Pd with high-efficiency single-pass and
recycling catalytic performances was determined as 0.045 wt %.
On the other hand, the Pd loading on the NTs almost did not
change after 10 cycles of catalysis, implying no or neglectable
Pd leaching in the selected catalysis using the present NTs
supported by a trace amount of Pd as the catalyst. Typical GC−
MS spectra of the liquid phase derived from the 1st, 4th, 7th,
and 10th cycles of catalysis are shown in Figures S1−S4 (SI),
respectively. It was notable that, although the collected solid
catalyst gradually became less and less during the past 10 cycles
of catalysis predominately owing to the loss originating from
the separation and transformation, the yield still remained
relatively stable or, strictly speaking, fluctuated slightly. In
combination with the ICP-MS results (Table S1 in the SI), it
was believed that there was just very slight leaching of Pd
during the catalytic process or, in other words, almost all of the
Pd species have redeposited onto the solid
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs after catalysis.
The unexpected good recycling performance revealed that
the SM reaction between iodobenzene and phenylboronic acid
might have reached or passed the terminal point within the
designated time, or put another way, there might also exist
some space for reducing the Pd loading to achieve the present
yield within the present time. Thus, specific control experi-
ments were designed to further ascertain the robust high-
efficiency catalyst. Figure S5 (SI) shows the GC−MS spectra of
Figure 4a2 indicates the yield of 4-methylbiphenyl (denoted
as the right Y axis). When the molar ratio of Mg2+:Ni2+
increased from 0:2.0 to 0.5:1.5, 1.0:1.0, 1.5:0.5, and finally
2.0:0, the yield gradually decreased from 99.6% to 99.5%,
98.4%, 95.2%, and finally 76.6%, respectively. Using the same
amounts of the NTs supported by sub-0.06 wt % Pd as the
catalyst, the monotonic alteration of the yield with the molar
ratios of Mg2+:Ni2+ for the second SM coupling reaction was
slightly different from that for the first one, probably because of
the various molar ratios of Mg2+:Ni2+, microstructures of the
NTs, and also the intrinsic characteristics of the selected
reactions. Expectedly, (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+
=
1.0:1.0) supported by sub-0.06 wt % Pd could be utilized as
a high-efficiency catalyst for the selected SM coupling reactions,
something to do with the high specific surface area and total
pore volume, which, however, seemed not to be the crucial
factor, taking the ill catalytic performances of the
Mg3Si2O5(OH)4 NTs supported by sub-0.06 wt % Pd into
consideration (Table 1).
To investigate the recycling catalytic performance, the solid
phase, i.e., (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs
supported by sub-0.06 wt % Pd catalyst, was recovered after the
single-pass first SM coupling reaction and employed for the
next new catalysis. As shown in Figure 4b, although after each
cycle of catalysis the solid catalyst was separated by
6025
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
Table 2. Comparison of the Pd Catalyst Supported on the Present (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs and Other
Supports, for the Selected SM Coupling Reaction between Iodobenzene and Phenylboronic Acid
Pd
loading
Pd/iodobenzene
(mol %)
reaction conditions solvent/base/
temperature/time
no. of
cycles
average isolated yield
a
b
entry
catalyst support
silica-APTS
(wt %)
(%)
ref
c
1
2.3
10.53
7.26
3
1.0
0.5
DMA/K2CO3/100 °C/2 h
H2O/K2CO3/RT/5−6 h
15
5
95.5 (99−93)
(95−92)
95 (97−94)
95
27
55
56
28
29
30
31
32
57
33
2
MgO
d
3
DOPA-pine- Fe2O3
MWCNT
PVC-DETA
PDrum
3.4
DMF−H2O/K2CO3/100 °C/10 min
MeOH/NaOAc/reflux/2 h
5
4
0.28
0.25
nr
nr
8
e
5
6.4
DMF−H2O/NaHCO3/90 °C/4−10 h
toluene/K2CO3/reflux/4 h
96 (100−93)
95 (95)
6
6.04
5.53
1.0
10
10
5
7
MCM41
TGAP
0.6
dioxane/K2CO3/80 °C/2 h
96 (96)
e
8
0.94
1.0
dioxane−H2O/K2CO3/90 °C/24 h
DMF−H2O/Na2CO3/50 °C/1 h
DMA−heptane−H2O/TEA/95 °C/20 h
EtOH/K2CO3/70 °C/3 h
98.5 (99−98)
94 (95−92)
65 (92−32)
9
NHC
1.17
5.50
10
5
f
10
11
PNIPAM
0.2
g
e
(Ni,Mg)3Si2O5(OH)4
NTs
0.045
0.11
10
99 (99.4−98.4)
this work
a
b
Percentage data showing wt % Pd loading when available; data in parentheses indicate mol % Pd used. Data in parentheses showing the range of
c
d
e
f
g
yields. Under a nitrogen atmosphere. Under microwave irradiation. GC yield. Molar ratio of Pd to support. Actual amount of Pd loaded on the
NTs, with 0.022 wt % Pd leading to a yield of 96.9% in single-pass catalysis.
Figure 5. Digital picture (a), XRD patterns (b), and TEM images (c and d) of the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs before (a1 and
b1) and after (a2) 0.045 wt % Pd loading and also those after 10 cycles of catalysis (a3, b2, c, and d), with a molar ratio of Mg2+:Ni2+ = 1.0:1.0. Vertical
line in (b): standard pattern for (Ni,Mg)3Si2O5(OH)4 (JCPDS No. 25-0524).
the liquid product phase derived from the specifically designed
reaction system using (Ni,Mg)3Si2O5(OH)4 (molar ratio of
Mg2+:Ni2+= 1.0:1.0) NTs loaded with 0.022 wt % Pd as the
catalyst. Suprisingly, even with half of the Pd loading as
employed previously, no significant amount of residual
iodobenzene was detected, and the yield was quite satisfactory
as 96.9%. Nevertheless, the blank test indicated that when bare
(Ni,Mg)3Si2O5(OH)4 (molar ratio of Mg2+:Ni2+ = 1.0:1.0) NTs
without Pd loading as the catalyst were utilized, no remarkable
biphenyl was detected and the yield was neglectable as less than
0.15% (Figure S6 in the SI). Consequently, on the one hand,
the presence of an appropriate amount of Pd was absolutely
necessary, but, on the other hand, such a low loading as 0.022
wt % Pd on the present (Ni,Mg)3Si2O5(OH)4 (molar ratio of
Mg2+:Ni2+ = 1.0:1.0) NTs exhibited a satisfactory high yield for
the selected single-pass SM coupling reaction.
The catalytic performance of the (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs supported by 0.045 wt % Pd is
further summarized in Table 2, based on a comparison with the
reported results using supported Pd for the same SM coupling
reaction. As can be seen, compared with those immobilized on
inorganic materials,27,28 the necessary Pd loading (correspond-
6026
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
Figure 6. Core-level Pd 3d XPS spectra taken from the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported by 0.045 wt % Pd before the 1st
cycle (a) and after the 10th cycle (b) of catalysis, with a molar ratio of Mg2+:Ni2+ = 1.0:1.0. BEs of 335.2 and 339.7 eV: Pd 3d5/2 and Pd 3d3/2 for
metallic Pd, respectively. BEs of 336.7 and 336.9 eV: Pd 3d5/2 for Pd2+ in PdO.
ing to the support) and required Pd (corresponding to
iodobenzene) were within the ranges of 2.3−10.53 wt % and
0.28−3.4 mol %, respectively. The referred inorganics
supported by Pd NPs were generally recycled 5−15 times,
with an average isolated yield of ca. 95%. For polymer-
supported Pd NPs,29,30 the amounts of Pd loaded and
employed were 6.04−6.4 wt % and 0.25 mol %, respectively,
with the recycles of 8−10 times and an average yield of 95−
96%. In addition to Pd NPs, Pd complexes were also
immobilized on a mesoporous silicate molecular sieve MCM-
4131 and so on.32,33 In contrast with the Pd NPs or complexes
immobilized on the above-described supports, the as-synthe-
sized (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs in this
work only contained 0.045 wt % Pd (corresponding to the
supports) and 0.11 mol % (corresponding to iodobenzene) but
could be recycled at least 10 times with an average GC yield of
99% (Table 2, entry 11). It was clear that both the amount of
Pd loaded and that employed versus iodobenzene were much
lower than those reported in the literature. Moreover, although
each referenced case could reach higher yield, when the time
was prolonged long enough, the present work indicated the
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported by
0.045 wt % Pd as a robust high-efficiency catalyst for the
traditional SM coupling reaction with excellent recycling
catalytic performance, within the present short time.
4. Further Characterization of the Trace Pd Supported
on (Ni,Mg)3Si2O5(OH)4 NTs. To probe the robust catalytic
performance of the present supported trace Pd catalyst, the
trace amount of Pd supported on the (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs was further characterized. Figure 5
shows the digital picture, XRD patterns, and TEM images of
the (Ni,Mg)3Si2O5(OH)4 NTs before and after 0.045 wt % Pd
loading and also those after 10 cycles of catalysis. The color of
the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs changed
from bright green (Figure 5a1, bare NTs) to dark green (Figure
5a2, after 0.045 wt % Pd loading) and remained almost
unchanged after 10 cycles of catalysis (Figure 5a3). Compared
with those original NTs before Pd loading (Figure 5b1), the
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported by
0.045 wt % Pd exhibited much higher crystallinity after 10
cycles of catalysis (Figure 5b2), due to a series of SM coupling
reactions at 70 °C for 3.0 h each time. As can be seen, the trace
loading of 0.045 wt % Pd did not change the crystal phase of
the hydrothermally synthesized NTs. Notably, however, no
diffraction peak of Pd was detected, owing to the low loading
amount, which was far below the detection limit of XRD. This
was quite different from the XRD pattern recorded from 7.0 wt
% Pd/CNT.34
On the other hand, the TEM (Figure 5c) and HRTEM
(Figure 5d) images reconfirmed the well-preserved tubular
structure and greatly improved crystallinity, without significant
mesopores within the walls observed after long-time electron-
beam irradiation, compared with the unstable ones before Pd
loading (Figure 1e). The legible lattice fringes bearing
interplanar spacings of 0.736 nm also confirmed the axial
direction of the NTs parallel to the (002) planes. This revealed
the excellent hydrothermal stability of the as-synthesized
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs without
destruction after 10 cycles of high-efficiency catalysis under
the base conditions, which were initially derived from an
environmentally rich base during hydrothermal synthesis.
Unfortunately, an attempt to seek the Pd NPs by TEM and
HRTEM failed, although tiny Pd NPs could grow bigger during
the series of catalysis because of the well-known Ostwald
ripening mechanism.35 As reported, Pd NPs also failed to be
found within even 0.4 wt % Pd(NH3)4Cl2/USY zeolite by TEM
and HRTEM,23 not to mention the present 0.045 wt % Pd
loading.
XPS spectra were recorded on the (Ni,Mg)3Si2O5(OH)4
NTs (Mg2+:Ni2+ = 1.0:1.0) supported by 0.045 wt % Pd
collected before the 1st cycle and after the 10th cycle of
catalysis, as shown in Figures S7 (SI) and 6. Spectra of Ni 2p
(Figure S7a), Mg 2p (Figure S7b), Si 2p (Figure S7c), and O 1s
(Figure S7d) reconfirmed the constitutional elements within
the as-synthesized (Ni,Mg)3Si2O5(OH)4 NTs, in accordance
the elemental mapping (Figure 3) and XRD (Figure 5b)
results. The XPS spectra from the Pd 3d core-level region were
also captured on the NTs supported by 0.045 wt % Pd in both
cases, and the asymmetric experimental photoelectron lines
could also be deconvoluted (Figure 6). The binding energies
(BEs) of 335.2 and 339.7 eV corresponded with the Pd 3d5/2
and Pd 3d3/2 peaks of the metallic Pd, respectively, in good
agreement with the literature results.36 The BEs of 336.7 and
336.9 eV were more likely assigned to the Pd 3d5/2 peak for
Pd2+ in PdO.36,37 Takasu and co-workers reported that Pd NPs
of 1.6 nm diameter gave rise to a BE of 336.6 eV,38 very similar
to 336.7 and 336.9 eV in the present case. Thus, both phases of
PdO and metallic Pd NPs, or clusters, might have contributions
to the BE of 336.7 and 336.9 eV. Probably, the formation of the
PdO phase was attributed to oxidation of the metallic Pd cluster
phase during separation of the product and transfer of the
recovered solid-solution NT phase between the various cycles
6027
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
of catalysis, owing to the intrinsic high activity of the clusters
themselves.
SM reaction between iodobenzene and phenylboronic acid
were assessed as 1162 and 2347, respectively; the correspond-
ing turnover frequencies (TOFs)44 were also assessed as 0.1076
and 0.2174 s−1, respectively. Owing to the above assumption,
the assessed TONs and TOFs were believed to be somehow
smaller than the authentic values.
To further understand the robust high-efficiency
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported
by trace Pd clusters and a possible catalytic mechanism, the
electron-spin resonance (ESR) spectra were recorded from the
intermediate liquid phase quenched during the first SM
coupling reaction, using the as-synthesized NTs supported by
0.045 wt % Pd as the catalyst and also bare ones without Pd
loading for comparison. As shown in Figure 7, the liquid
5. Possible Catalytic Mechanism of the
(Ni,Mg)3Si2O5(OH)4 NTs Supported by 0.045 wt % Pd.
Variation of the Pd concentration during catalysis was
monitored so as to confirm migration of the Pd species. The
GC yield and ICP-MS results are listed in Table S3 (SI), and
the corresponding GC spectra of some extended control
experiments are shown in Figure S8 (SI). As can be seen, for
the quenched system (Table S3, entry 1, and Figure S8a in the
SI), the naturally cooled system (Table S3, entry 2, and Figure
S8b in the SI) exhibited higher yield owing to the proceeding
reaction during the cooling process. Meanwhile, the Pd
concentration of the cooled system was determined as 43.6
ng mL−1 (i.e., ppb), much higher than that of the quenched one
(14.8 ng mL−1). It implied that the rapid quenching promoted
redeposition of active Pd species to the solid-solution NT
phase. In addition, the coherent reaction without in-between
interruption (Table S3, entry 3, in the SI) exhibited a final high
yield (99.1%) and a low Pd concentration (23.6 ng mL−1,
∼0.02 ppm). Notably, the Pd/C-catalyzed SM coupling
reaction exhibited a significant Pd leaching of 3 ppm in the
organic dimethylacetimide (DMA)/triethylamine (TEA) sol-
vents after 4.0 h of reaction;45 1.0 wt % Pd/TGAP revealed a
relatively low Pd leaching of less than 0.4 ppm in the reaction
system;32 2.3 wt % Pd/silica-APTS demonstrated a lower Pd
leaching of less than 0.17 ppm.27 Crudden and co-workers
reported an excellent Pd catalyst supported on mercaptopropyl-
modified mesoporous silica, which has been recognized as one
of the best catalytic Pd materials with lowest leaching (72−3
ppb in 2.5 mL of solvent; average 242 ng for 4 runs).40,46
Hoshiya and co-workers provided a low-releasing Pd catalyst
supported on sulfur-modified gold via treatment with a piranha
solution (12.7−6.3 ppb in 3 mL of solvent; average 26 ng for 10
runs).47 In contrast, the present NTs supported by 0.045 wt %
Pd exhibited a low Pd concentration of 0.0236 ppm (23.6 ppb)
in 7.5 mL of solvent (177 ng) after the first run. Because Pd
leaching decreased significantly from the second run and
tended lower and lower in the subsequent recycling perform-
ances,32,48 the overall Pd dissociated into the present 10 cycles
of catalysis was estimated at an extremely low level, lower than
that of the Pd material provided by Crudden et al.40,46 and also
comparable to that presented by Hoshiya et al.47
Figure 7. ESR spectra of the liquid phase quenched during the SM
coupling reaction, with (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0)
NTs supported by 0.045 wt % Pd (a) and bare NTs without Pd
loading (b) as the catalyst.
quenched from the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+
=
1.0:1.0) NTs supported by a 0.045 wt % Pd system exhibited
a significant resonance peak in the magnetic field range of 326−
330 mT (Figure 7a). However, bare NTs without Pd loading
(Figure 7b) also demonstrated a significant resonance peak at
the same magnetic field range, denoted as the blue dash-dotted
elliptical region. This implied the absence of a Pd single-
electron resonance, indicating the absence of the odd-valence
Pd species such as Pd+ or Pd3+ within the quenched
intermediate catalytic system.
Thus, the XPS Pd 3d spectra directly and macroscopically
confirmed the existence of Pd species loaded on the NTs, in
combination with the ICP-MS results (Table S1 in the SI).
Nevertheless, owing to the low Pd loading (0.045 wt %) on the
unstable NTs (Mg2+:Ni2+ = 1.0:1.0), the Pd species was neither
successfully detected by XRD nor straightforward captured by
TEM, elemental mapping, and HRTEM. Taking the rapid
reduction during the trace Pd loading within the EtOH−water
phase into account, it was assumed that Pd clusters should have
been formed after reduction and further highly dispersed onto
the (Ni,Mg)3Si2O5(OH)4 NTs (Mg2+:Ni2+ = 1.0:1.0). As a
consequence, the inherent extremely small size and surface
effect contributed to the instability and high activity of the Pd
clusters, which was also found in some other Pd-catalyzed
reactions.22,23,39−41 This was quite similar to the gold particles
highly dispersed on the ZrO2 surface at a low loading level of
0.01−0.08 wt %.42,43
On the other hand, a higher concentration of Pd species
within the liquid phase propelled the formation of product
biphenyl. In addition, the liquid phase derived from either the
quenched (Figure S8c in the SI) or cooled (Figure S8d in the
SI) system after centrifugation could further react continuously
itself when reheated to 70 °C for another 4.0 h, resulting in a
higher yield of biphenyl, without the aid of solid-solution NTs
supported by Pd. This definitely ascertained the presence of Pd
species within the liquid product phase, in good agreement with
the ICP-MS results. Comparatively, the fresh reaction system
was also catalyzed by the sedimental phase of the quenched
system at 70 °C for 4.0 h (Figure S8e in the SI), giving rise to a
quite high yield (ca. 96%), implying more active Pd species
redeposited onto the solid-solution NT phase after quenching.
This test was similar to the classical hot filtration or
centrifugation test,15,49 to some extent. As reported, however,
On the basis of the above analysis, it was rational to speculate
that each Pd atom was a catalytic active site. Thus, as shown in
Table S2 (SI), using (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+
=
1.0:1.0) NTs supported by 0.045 wt % Pd as the catalyst, the
turnover numbers (TONs)44 for the 1st and 10th cycles of the
6028
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
the hot filtration test is probably useful if the bulk metal
catalyst, rather than the nanocluster catalyst, is responsible for
catalysis.49 Similarly, for the nanocluster-participated catalysis,
the centrifugation test could hardly achieve an ideal quantitative
separation of the sediment and supernatant, and, as a
consequence, both of them will probably have some residual
catalytic activity. Especially, the centrifugation test will not work
when the reaction is catalyzed by a low concentration of highly
active nanoclusters because no visible sediment will form.49
To gain deeper insight into the catalytic active species, the
classical mercury poisoning experiment was carried out, which
has been well established for distinguishing heterogeneous from
homogeneous catalysis, with the aid of a requisite control
experiment.49,50 As shown in Figure S9 (SI), the GC spectra of
the liquid phase acquired from the quenched system when
catalysis proceeded for 15.0 min (Figure S9a in the SI) and that
recovered for catalysis for another 2.0 h and 45.0 min after the
mercury poisoning (Figure S9b in the SI) exhibited the same
characteristics with the same yield of the product biphenyl. This
indicated that the addition of excess mercury (125.2 mg; molar
ratio of Hg0 to Pd ∼ 1480) has completely suppressed the SM
cross-coupling reaction between iodobenzene and phenyl-
boronic acid. As reported, the coupling of iodoarenes could
be homogeneously catalyzed by Pd ultratraces at a temperature
higher than 100 °C, by using palladacycles as catalysts.51,52
However, in these nice researches, carefully designed ligands
are needed to ensure high reactivity of these catalysts. In our
work, no special ligands are present during the catalysis process,
which greatly lowers the possibility of homogeneous catalysis
by a trace amount of Pd released to the organic solution. Thus,
taking the present mercury poisoning test and also the previous
SM cross-coupling reaction (Figure 4) without mercury
poisoning performed at 70 °C into consideration, the
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs supported
by 0.045 wt % Pd catalyst exhibited distinct genuine
heterogeneous catalytic nature. Objectively, the Pd concen-
trations of the intermediate and final product liquid phase
detected (Table S3 in the SI) were probably due to the release
of Pd nanoclusters dissociated from the bulk NT phase,
different from those of leached soluble Pd species in other
homogeneous catalyses.
On the basis of the almost unchanged 0.045 wt % of the
actual Pd loading amount on the NTs before the 1st cycle and
after the 10th cycle of catalysis (Table S1 in the SI),
centrifugation-like (Table S3 and Figure S8 in the SI), and
especially the mercury poisoning (Figure S9 in the SI)
experimental results, the detected Pd species within the
intermediate supernatant derived from centrifugation (Table
S3 and Figure S8 in the SI) were firmly believed to be the solid
Pd nanoclusters dissociated from the bulk NTs during catalysis
performed at elevated temperature under vigorous magnetic
stirring rather than the soluble or leached Pd species. The
dissociated Pd nanoclusters within the supernatant contributed
to further catalysis after centrifugation. Moreover, the SM
cross-coupling reaction took place on the surfaces of the Pd
nanoclusters rather than the homogeneous solution phase,
definitely demonstrating the genuine heterogeneous catalytic
characteristic. The present active Pd nanoclusters were different
from the leached Pd species in the liquid phase, which have
become accepted in Heck and some other coupling
reactions.8,53 Because absolute leaching of soluble Pd species
could not be excluded thoroughly and the Pd ultratraces
released into the organic solution could catalyze the coupling of
iodoarenes in due course,51,52 the hypotheses of the possible
effect of the leached soluble Pd ultratraces on the present
catalysis could not be discarded entirely, which, however, was
somehow believed to be inappreciable and thus neglectable.
Similar to the catalytic cycle for the Pd/C-catalyzed SM
coupling reaction,54 the possible catalytic cycle of the Pd/
(Ni,Mg)3Si2O5(OH)4-catalyzed SM coupling reaction, which
took place on the surfaces of the Pd nanoclusters deposited on
the solid-solution NTs, could thus be depicted based on the
above results, as shown in Scheme 1. The cycle consisted of five
Scheme 1. Possible Catalytic Cycle of the Pd/
(Ni,Mg)3Si2O5(OH)4-Catalyzed SM Coupling Reaction
Taken Place on the Surfaces of the Pd Clusters Deposited on
the Solid-Solution NTs: (i) Oxidative Addition and
Resultant Pd Leaching; (ii) Metathesis due to the Addition
of Hydroxide Ion OH− Originating from Dissociation of the
Base K2CO3; (iii) Transmetalation Owing to the Arylborate
Activation; (iv) Reductive Elimination Leading to Pd
Clusters and Product Containing a Trace Amount of
Leached Pd; (v) Precipitation of Pd Clusters onto the
(Ni,Mg)3Si2O5(OH)4 NTs, Partially Oxidized to PdO due to
High Activity
main steps: oxidative addition and resultant Pd leaching (i),
metathesis due to the addition of hydroxide ion OH−
originating from dissociation of the base K2CO3 (ii), trans-
metalation owing to arylborate activation (iii), reductive
elimination leading to Pd nanoclusters and product containing
very trace amounts of leached Pd (iv), and precipitation of Pd
nanoclusters onto the (Ni,Mg)3Si2O5(OH)4 NTs with probably
and partially oxidized PdO due to high activity (v). It was clear
that the base played a key role in the whole catalytic cycle
within the above metathesis (ii) and transmetalation (iii),
which gave rise to R−Pd−OH and tetrahedral boronate
−
B(OH)4 , respectively. Of note was that, however, beyond
the figured seemingly closed catalytic cycle, the separation of
liquid product containing a very trace amount of leached Pd
and transfer of recovered solid-solution NTs supported by Pd
between various catalytic cycles would lead to Pd loss out of the
cycle to some extent. The physical separation and transfer were
responsible for the overall Pd loss, taking the almost consistent
Pd loading amount on the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+
=
1.0:1.0) NTs before the 1st cycle and after the 10th cycle of
catalysis into account (Table S1 in the SI).
A s f o r t h e p r e s e n t r o b u s t h i g h - e ffi c i e n c y
(Ni,Mg)3Si2O5(OH)4 NTs supported by 0.045 wt % Pd
catalyst, three more aspects should also be taken into
6029
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
consideration. First, high-activity Pd nanoclusters could be
oxidized to PdO (Figure 6) clusters during the separation of
liquid product and transfer of recovered solid-solution NTs
supported by a catalyst. Because PdII is prone to leaching,8 the
PdO clusters might promote the subsequent metathesis,
compared with pure Pd nanocluster phase. Second, the specific
unstable (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs
contained multitudes of dislocations, defaults, and terraces,
which could serve as particularly ideal reservoirs for the
reduction formation and in situ deposition of the newly formed
tiny Pd nanoclusters of different sizes, and this could also
account for the extremely low or even neglectable Pd leaching
level during the 10 cycles of catalysis. Third, extensive
investigation revealed that the present (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs supported by 0.045 wt % Pd did not
perform well for the SM cross-coupling reaction between the
more challenging substrates, e.g., the nonactivated aryl
bromides, with phenylboronic acid. To get the relatively
satisfactory conversion and yield, higher yield of the product,
higher Pd loading, and more NTs supported by the Pd catalyst,
higher temperature and longer time were necessary. Finally,
because low-crystallinity Mg3Si2O5(OH)4 NTs (with also high
specific surface area and analogous tubular structure; Figure 1g)
supported by the same amount of Pd exhibited relatively low
yield (Figure 4a), the possible synergistic effect of Ni−Mg for
the (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs sup-
ported by Pd should also contribute to the corresponding high
yield, excellent recycling catalytic performance, and extremely
low Pd release, which, however, needed further investigation
and deeper understanding.
catalysts with low loading, good recycling, and extremely low
ppb level release, which could also be extended as robust
reusable high-efficiency catalysts for some other SM coupling
reactions.
ASSOCIATED CONTENT
* Supporting Information
■
S
GC, MS, and XPS spectra and tables of ICP-MS and GC yield
results and TON and TOF values. This material is available free
of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: wangxun@mail.tsinghua.edu.cn. Tel: 86-10-62792791.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by NSFC (Grants 91127040,
20971078, and 20921001) and the State Key Project of
Fundamental Research for Nanoscience and Nanotechnology
(Grant 2011CB932402). The authors thank Dr. Yiqing Zhou
and Dr. Guirong Zhang at the Department of Chemistry,
Tsinghua University, for helpful discussion and also the
reviewers for constructive suggestions on the great improve-
ment of this work.
REFERENCES
(1) Heck, R. F. J. Am. Chem. Soc. 1968, 90, 5518.
(2) Heck, R. F. Acc. Chem. Res. 1979, 12, 146.
■
(3) Beletskaya, I. P.; Cheprakov, A. V. Chem. Rev. 2000, 100, 3009.
(4) Negishi, E.; Hu, Q.; Huang, Z.; Qian, M.; Wang, G. Aldrichim.
Acta 2005, 38, 71.
CONCLUSION
■
In summary, (Ni1−x,Mgx)3Si2O5(OH)4 solid-solution NTs with
tunable compositions were synthesized via a facile hydro-
thermal route at 210 °C for 24.0 h, by tuning the molar ratio of
Mg2+ to Ni2+ within the range of 0:2.0 to 2.0:0. The as-
synthesized NTs were loaded with 0.045 wt % Pd for the SM
coupling reactions between iodobenzene or 4-iodotoluene and
(5) Negishi, E. Acc. Chem. Res. 1982, 15, 340.
(6) Rafael, C.; Carmen, N. Chem. Rev. 2007, 107, 874.
(7) Li, P.; Wang, L.; Li, H. Tetrahedron 2005, 61, 8633.
(8) Yin, L.; Liebscher, J. Chem. Rev. 2007, 107, 133.
(9) Fihri, A.; Bouhrara, M.; Nekoueishahraki, B.; Basset, J.-M.;
Polshettiwar, V. Chem. Soc. Rev. 2011, DOI: 10.1039/c1cs15079k.
(10) Lamblin, M.; Nassar-Hardy, L.; Hierso, J.-C.; Fouquet, E.;
Felpin, F.-X. Adv. Synth. Catal. 2010, 352, 33.
phenylboronic acid. (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+
=
1.0:1.0) NTs supported by 0.045 wt % Pd promoted the
iodobenzene-participated coupling reaction with a high yield of
>99%, an excellent recycling catalytic performance during 10
cycles of catalysis with yields of ∼99%, and also an extremely
low Pd releasing level of ∼0.02 ppm. After 10 cycles of catalysis,
(Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) NTs exhibited
well-defined tubular structures and higher crystallinity. The
almost consistent actual Pd loading amount on the NTs before
the 1st cycle and after the 10th cycle of catalysis and Pd 3d XPS
spectra directly confirmed the robust stable existence of Pd and
partially oxidized PdO clusters supported on the NTs. In
addition, (Ni,Mg)3Si2O5(OH)4 NTs supported by 0.022 wt %
Pd could also catalyze the SM coupling of iodobenzene and
phenylboronic acid with a satisfactory yield of ∼97%. High-
activity Pd and PdO clusters, special (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs containing multitudes of
dislocations, defects, and terraces, and also the synergistic
effect of Ni−Mg should contribute to the (Ni,Mg)3Si2O5(OH)4
(Mg2+:Ni2+ = 1.0:1.0) NTs supported by 0.045 wt % Pd as a
robust reusable high-efficiency catalyst for the selected SM
coupling reactions but with extremely low Pd release. The
present (Ni,Mg)3Si2O5(OH)4 (Mg2+:Ni2+ = 1.0:1.0) solid-
solution NTs provided an ideal hydrothermally stable tubular-
structured support for Pd and other noble- or transition-metal
(11) Bellina, F.; Carpita, A.; Rossi, R. Synthesis 2004, 15, 2419.
(12) Kotha, S.; Lahiri, K.; Kashinath, D. Tetrahedron 2002, 58, 9633.
(13) Alonso, F.; Beletskaya, I. P.; Yus, M. Tetrahedron 2008, 64,
3047.
(14) Franzen, R.; Xu, Y. Can. J. Chem. 2005, 83, 266.
́
(15) Molnar, A. Chem. Rev. 2011, 111, 2251.
(16) Seki, M. Synthesis 2006, 18, 2975.
(17) Alonso, F.; Beletskaya, I. P.; Yus, M. Tetrahedron 2005, 61,
11771.
(18) Biffis, A.; Zecca, M.; Basato, M. J. Mol. Catal. A: Chem. 2001,
173, 249.
(19) Cheong, S.; Watt, J. D.; Tilley, R. D. Nanoscale 2010, 2, 2045.
(20) Beller, M.; Fischer, H.; Kuhlein, K.; Reisinger, C.-P.; Herrmann,
̈
W. A. J. Organomet. Chem. 1996, 520, 257.
(21) Zhao, F.; Murakami, K.; Shirai, M.; Arai, M. J. Catal. 2000, 194,
479.
(22) Reetz, M. T.; Breinbauer, R.; Wannlnge, K. Tetrahedron Lett.
1996, 37, 4499.
(23) Okumura, K.; Matsui, H.; Sanada, T.; Arao, M.; Honma, T.;
Hirayama, S.; Niwa, M. J. Catal. 2009, 265, 89.
(24) Yang, Y.; Liang, Q.; Li, J.; Zhuang, Y.; He, Y.; Bai, B.; Wang, X.
Nano Res. 2011, 4, 882.
(25) Wang, X.; Zhuang, J.; Chen, J.; Zhou, K.; Li, Y. Angew. Chem.,
Int. Ed. 2004, 43, 2017.
(26) Li, R.; Tao, X.; Li, X. J. Mater. Chem. 2009, 19, 983.
6030
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031

Inorganic Chemistry
Article
(27) Zhang, L.; Wang, L.; Li, H.; Li, P. Synth. Commun. 2008, 38,
1498.
(28) Pan, H.-B.; Yen, C. H.; Yoon, B.; Sato, M.; Wai, C. M. Synth.
Commun. 2006, 36, 3473.
(29) Cui, Y.-C.; Zhao, X.-W.; Zhang, J.-W.; Zhang, L.; Liu, X.-M. Acta
Chim. Sin. 2006, 64, 42.
(30) Chandrasekhar, V.; Suriya Narayanan, R.; Thilagar, P.
Organometallics 2009, 28, 5883.
(31) Cai, M.; Sha, J.; Xu, Q. J. Mol. Catal A: Chem. 2007, 268, 82.
(32) Artner, J.; Bautz, H.; Fan, F.; Habicht, W.; Walter, O.; Doring,
̈
M.; Arnold, U. J. Catal. 2008, 255, 180.
(33) Bergbreiter, D. E.; Osburn, P. L.; Wilson, A.; Sink, E. M. J. Am.
Chem. Soc. 2000, 122, 9058.
(34) Chen, X.; Hou, Y.; Wang, H.; Cao, Y.; He, J. J. Phys. Chem. C
2008, 112, 8172.
(35) Baldan, A. J. Mater. Sci. 2002, 37, 2171.
(36) Handbook of X-ray Photoelectron Spectroscopy; JF, M., WF, S., PE,
S., KD, B., Eds.; Perkin-Elmer Corp.: Eden Prairie, MN, 1992; Vol. 3.
(37) Semagina, N.; Renken, A.; Laub, D.; Kiwi-Minsker, L. J. Catal.
2007, 246, 308.
(38) Takasu, Y.; Unwin, R.; Tesche, B.; Bradshaw, A. M.; Grunze, M.
Surf. Sci. 1978, 77, 219.
(39) Bulut, H.; Artoka, L.; Yilmaz, S. Tetrahedron Lett. 2003, 44, 289.
(40) Webb, J. D.; MacQuarrie, S.; McEleney, K.; Crudden, C. M. J.
Catal. 2007, 252, 97.
(41) Gulay, D.; Ozge, A.; Levent, A. J. Mol. Catal. A: Chem. 2007,
278, 189.
(42) Zhang, X.; Shi, H.; Xu, B.-Q. Catal. Today 2007, 122, 330.
(43) Zhang, X.; Shi, H.; Xu, B.-Q. Angew. Chem., Int. Ed. 2005, 44,
7132.
(44) Umpierre, A. P.; Jesus, E. d.; Dupont, J. ChemCatChem 2011, 3,
1413.
(45) Chen, J.-S.; Vasiliev, A. N.; Panarello, A. P.; Khinast, J. G. Appl.
Catal., A 2007, 325, 76.
(46) Crudden, C.; Sateesh, M.; Lewis, R. J. Am. Chem. Soc. 2005, 127,
10045.
(47) Hoshiya, N.; Shimoda, M.; Yoshikawa, H.; Yamashita, Y.; Shuto,
S.; Arisawa, M. J. Am. Chem. Soc. 2010, 132, 7270.
(48) Qin, Y.; Wei, W.; Luo, M. Synlett 2007, 15, 2410.
(49) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198,
317.
(50) Weddle, K. S.; Aiken, J. D., III; Finke, R. G. J. Am. Chem. Soc.
1998, 120, 5653.
(51) Wolfe, J. P.; Singer, R. A.; Yang, B. H.; Buchwald, S. L. J. Am.
Chem. Soc. 1999, 121, 9550.
(52) Dupont, J.; Consorti, C. S.; Spencer, J. Chem. Rev. 2005, 105,
2527.
(53) Phan, N. T. S.; Sluys, M. V. D.; Jones, C. W. Adv. Synth. Catal.
2006, 348, 609.
(54) Conlon, D. A.; Pipik, B.; Ferdinand, S.; LeBlond, C. R.; Sowa, J.
R.; Izzo, J. B.; Collins, P.; Ho, G.-J.; Williams, J. M.; Shi, Y.-J.; Sun, Y.
Adv. Synth. Catal. 2003, 345, 931.
(55) Kantam, M. L.; Roy, S.; Roy, M.; Sreedhar, B.; Choudary, B. M.
Adv. Synth. Catal. 2005, 347, 2002.
(56) Polshettiwar, V.; Nadagouda, M. N.; Varma, R. S. Chem.
Commun. 2008, 6318.
(57) Kim, J.-H.; Kim, J.-W.; Shokouhimehr, M.; Lee, Y.-S. J. Org.
Chem. 2005, 70, 6714.
6031
dx.doi.org/10.1021/ic2024378 | Inorg. Chem. 2012, 51, 6020−6031