A general approach to fabricate fe3O4 nanoparticles decorated with Pd, Au, and Rh: Magnetically recoverable and reusable catalysts for Suzuki C-C cross-coupling reactions, hydrogenation, and sequential reactions

Article abstract of DOI:10.1002/chem.201301769

A facile strategy has been explored for loading noble metals onto the surface of ferrite nanoparticles with the assistance of phosphine-functionalized linkers. Palladium loading is shown to occur with participation of both the phosphine function and the surface hydroxyl groups. Hybrid nanoparticles containing simultaneously Pd and Au (or Rh) are obtained by successive loading of metals. Similarly, ferrite nanoparticles decorated with Pd, Au, and Rh have also been formed by using the same strategy. The catalytic properties of the new nanoparticles are evidenced in processes such as reduction of 4-nitrophenol or hydrogenation of styrene. Besides, the sequential process involving a cross-coupling reaction followed by reduction of 1-nitrobiphenyl has been successfully achieved by employing Pd/Au decorated nanoferrite particles. Copyright

Full text of DOI:10.1002/chem.201301769

FULL PAPER
Nanoparticles
F. Gonzꢀlez de Rivera, I. Angurell,*
M. D. Rossell, R. Erni, J. Llorca,
N. J. Divins, G. Muller, M. Seco,
O. Rossell* .................... &&&& — &&&&
Towards a universal catalyst? One,
two, and up to three noble metals have
been loaded onto the surface of ferrite
nanoparticles peripherally functional-
ized with phosphine groups (see
scheme). The catalytic properties of
the new nanoparticles are evidenced in
processes such as reduction of 4-nitro-
phenol or hydrogenation of styrene.
A General Approach To Fabricate
Fe3O4 Nanoparticles Decorated with
Pd, Au, and Rh: Magnetically Recov-
erable and Reusable Catalysts for
À
Suzuki C C Cross-Coupling Reac-
tions, Hydrogenation, and Sequential
Reactions
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

DOI: 10.1002/chem.201301769
A General Approach To Fabricate Fe3O4 Nanoparticles Decorated with Pd,
Au, and Rh: Magnetically Recoverable and Reusable Catalysts for Suzuki
À
C C Cross-Coupling Reactions, Hydrogenation, and Sequential Reactions
Ferran Gonzꢀlez de Rivera,[a] Inmaculada Angurell,*[a] Marta D. Rossell,[b] Rolf Erni,[b]
Jordi Llorca,[c] Nfflria J. Divins,[c] Guillermo Muller,[a] Miquel Seco,[a] and Oriol Rossell*[a]
Abstract: A facile strategy has been ex-
plored for loading noble metals onto
the surface of ferrite nanoparticles with
the assistance of phosphine-functional-
ized linkers. Palladium loading is
shown to occur with participation of
both the phosphine function and the
surface hydroxyl groups. Hybrid nano-
particles containing simultaneously Pd
and Au (or Rh) are obtained by succes-
sive loading of metals. Similarly, ferrite
nanoparticles decorated with Pd, Au,
and Rh have also been formed by
using the same strategy. The catalytic
properties of the new nanoparticles are
evidenced in processes such as reduc-
tion of 4-nitrophenol or hydrogenation
of styrene. Besides, the sequential proc-
ess involving a cross-coupling reaction
followed by reduction of 1-nitrobiphen-
yl has been successfully achieved by
employing Pd/Au decorated nanofer-
rite particles.
Keywords: ferrite · gold · nanopar-
ticles · palladium · rhodium
Introduction
vantage of easy separation. But this is not the common sce-
nario. Attempts to improve separation include the use of bi-
phasic aqueous/organic systems,[4] ionic liquid biphasic con-
ditions,[5] and metal nanoparticles supported on large surface
area substrates.[6] Recently, magnetic nanoparticles-support-
ed catalysts have been efficiently employed as heterogene-
ous catalysts. They have the added advantage of being mag-
netically separable, thereby eliminating the requirement of
catalyst filtration after completion of the reaction.[7] Conse-
quently, the use of metal-supported magnetic nanoparticles
prevents loss of the catalyst.[8] To immobilize the metal cata-
lyst on the magnetic nanoparticle, two approaches have
been designed. The first consists of direct reaction of a
metal salt with bare magnetic nanoparticle followed by re-
duction. The second involves the previous capping of the
nanoparticle with a linker containing a terminal coordinat-
ing group. Then, the metal ions are added and deposited on
the surface after reduction. The latter is the most common
procedure reported in the literature since the influence of
the ligands on the nucleation and grow process of the nano-
particles has been shown to be crucial.[9] In particular, the
size and distribution of Pd nanoparticles are strongly de-
pendent on the affinity ligands used.[10]
In the vast majority of examples the stabilizers appear to
be amine ligands because they are known to stabilize nano-
particles against aggregation without disturbing their desira-
ble properties and are also recognized to increase their cata-
lytic activity.[11] In particular, dopamine is the most widely
employed linker.[12] In contrast, phosphine-functionalized
capping agents are far less used.[13] This is somewhat surpris-
ing considering that phosphine groups are softer donors and
would be better stabilizers for palladium catalytic active spe-
cies than amine groups.
Metal nanoparticles exhibit much higher catalytic efficiency
as compared with the corresponding bulk material due to
their large surface to volume ratio.[1] However, the nanopar-
ticles must be stabilized by protective agents to prevent ag-
glomeration to the thermodynamically favored bulk metal.
In addition, the incorporation of adequate capping-agents
can also offer better dispersion of the catalyst in appropriate
solvents as well as new potential catalytic centers.[2] Yet, the
main problem in using such nanoparticles for catalysis is as-
sociated with the separation of the catalyst after the reac-
tion.[3] Obviously, those catalysts that are not soluble in the
same phase as the organic reagents have the inherent ad-
[a] F. Gonzꢁlez de Rivera, Dr. I. Angurell, Prof. G. Muller, Dr. M. Seco,
Prof. O. Rossell
Departament de Quꢂmica Inorgꢁnica
Universitat de Barcelona
Martꢂ i Franquꢃs 1-11, 08028 Barcelona (Spain)
Fax : (+34)934907725
E-mail: inmaculada.angurell@qi.ub.es
oriol.rossell@qi.ub.es
[b] Dr. M. D. Rossell, Dr. R. Erni
Electron Microscopy Center
Empa, Swiss Federal Laboratories for
Materials Science and Technology
ꢄberlandstrasse 129, 8600 Dꢅbendorf (Switzerland)
[c] Dr. J. Llorca, N. J. Divins
Institut de Tꢃcniques Energꢃtiques i
Centre de Recerca en Nanoenginyeria
Universitat Politꢃcnica de Catalunya
Diagonal 647, 08028 Barcelona (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301769.
&
2
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Magnetically Recoverable and Reusable Catalysts
FULL PAPER
À
Up to now, a significant number of reports of monometal-
lic-supported nanoparticles have been published,[7a] but ex-
amples of magnetic nanoparticles containing two or more
different metals are very scarce. To our knowledge, only
magnetite nanoparticles containing core/shell Au/Ag,[14] Pt/
Au,[9a] Pd/Au alloy particles,[15] and Ag@Pd satellites[16] have
been described. The intrinsic catalytic interest of this kind of
nanocomposites, derived from the presence of a second ele-
ment, prompted us to explore the possibility of extending
our knowledge to hybrid systems containing two or even
three different metal atoms on the surface of magnetic
nanoparticles.
transformed infrared spectroscopy (FTIR) (n
(N C=O):
À1
1637 cmÀ1; n
A
À
dopPPh2 (3.88%) was deduced from microanalysis. The
crystalline structure of the Fe3O4dopPPh2 was characterized
by XRD as shown in Figure S2 in the Supporting Informa-
tion.
The (220), (311), (400), (422), (511), and (440) diffraction
peaks reflect the magnetite crystal with a cubic spinel struc-
ture.[18] An average crystallite size of 16 nm was estimated
by applying the Scherrer formula[19] on the full width at half
maximum (0.498) of the strongest (100%) reflection, whose
2q value is 35.608.
Herein we describe the formation of phosphine-function-
alized ferrite nanoparticles, Fe3O4dopPPh2 (dop=dop-
ACHTUNGTRENNUNGamine), and their capacity to load different metal atoms on
Metals loading onto Fe3O4dopPPh2: As described above,
two strategies for loading metals on the magnetite nanopar-
ticles surface are generally applied: 1) addition of metal ions
to functionalized nanoparticles with appropriate coordinat-
ing ligands and 2) incorporation of metal ions by reaction
with the hydroxyl groups of the surface of the magnetite
nanoparticles. The latter will be named “direct route”
throughout this paper. The final step in both processes con-
sists of the reduction of the metallic ions. It is surprising to
observe that in all publications dealing with functionalized
nanoparticles, the total content of the deposited metal is
only attributed to the metal ions coming from their previous
coordination to the linker. That is, the component due to
direct immobilization on the surface without functionalized-
linker mediation is not considered. We conjecture that both
routes are available and operate simultaneously so that the
total content of the immobilized metal on the nanoparticle
surface results from the addition of both contributions.
their surface. The goal was to synthesize a catalyst contain-
ing up to three different transition metals (Pd, Au and Rh),
which were expected to be excellent candidates for a
number of catalytic processes. In addition, the comparison
of the Pd loading among Fe3O4dopPPh2@Pd, naked Fe3O4
and new expressly designed Fe3O4catechol nanoparticles has
permitted to establish that the incorporation of the metal
onto Fe3O4dopPPh2 surface occurs simultaneously and with
similar participation through two routes: 1) Coordination to
phosphine ligands and 2) bonding to hydroxyl surface
groups. The catalytic behavior of the new magnetic catalysts
is demonstrated in various processes (Suzuki–Miyaura reac-
tions, styrene hydrogenation, and nitrophenol reduction)
À
and also in the sequential tandem reaction involving the C
C cross-coupling reaction of 1-bromo-4-nitrophenyl with
phenylboronic acid, followed by reduction of the nitro to
amino group.
The process followed in this work to load Pd metal onto
Fe3O4dopPPh2 nanoparticles involved the following steps:
1) Addition of K2ACHTNUTRGNEUNG[PdCl4] to a previously sonicated suspen-
Results and Discussion
sion of Fe3O4dopPPh2 nanoparticles in water, and stirring
for two hours; 2) isolation, washing, and drying of the nano-
particles and redispersion in water; 3) reduction with
NaBH4, and extraction of the new species with an external
magnet. The Pd salt was added in slight excess to saturate
both the phosphine ligands and the active peripherals hy-
droxyl groups. A ratio of Pd/magnetite nanoparticles of
5:100 mg was inferred from a set of experiments controlled
by inductively coupled plasma optical emission spectrometry
(ICPoes). The average weight percentage of Pd in the result-
ing Fe3O4dopPPh2@Pd nanoparticles was found to be about
1.31 after at least three separate experiments.
Magnetite nanoparticles were synthesized according to a lit-
erature procedure.[17] The ligand dopPPh2 was obtained by
treatment of dopamine hydrochloride with (diphenylphos-
phino)benzoic acid (Scheme 1).
The 31P{1H} NMR spectrum showed a signal at d=
À5.8 ppm revealing the presence of free phosphine (see Fig-
À
ure S1 in the Supporting Information) and the nACTHNUGRTNEUNG(N C=O)
band at 1637 cmÀ1 evidenced the formation of the amide
function. The ligand was anchored easily onto the surface of
the as-synthesized Fe3O4 nanoparticles by sonication in
methanol during two hours. The new starting Fe3O4dopPPh2
nanoparticles were isolated with an external magnet. An-
choring of the dopPPh2 molecules was confirmed by Fourier
To better understand the role of the ligands grafted on
the magnetite surface and also for comparison purposes, we
prepared Fe3O4@Pd nanoparticles by treatment of Fe3O4
nanoparticles (without any
modification of the surface)
with K2ACHTNUGTRNEUNG[PdCl4] in water. This
procedure rendered a smaller
content of Pd (about 1.1%).
At this point it seemed im-
portant to evaluate the relative
weight of each one of the
Scheme 1. Synthetic route to obtain dopPPh2: Coupling between dopamine and (diphenylphosphino)benzoic
acid. EDC=1-Ethyl-3-(3-dimethyl
ACHTUNGTRENaNUGN minopropyl)carbodiimide, NHS=N-hydroxysuccinimide.
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

I. Angurell, O. Rossell et al.
routes for palladium incorporation. For the first approach,
one could propose that Pd loading by means of functional-
ized linkers could be simply deduced by subtracting 1.10%
(direct reaction) from the average value of 1.31% obtained
using Fe3O4dopPPh2 nanoparticles, that is, 0.21%. However,
this assumption does not take into consideration that an-
chorage of the linkers inhibits the activity of a great number
of hydroxyl groups on the surface of the nanoparticles for
further Pd loading through direct reaction. To correct this
effect we designed the synthesis of new magnetite nanopar-
ticles functionalized with catechol (o-C6H4(OH)2),
Fe3O4catechol. Clearly, this ligand is unable to coordinate
metal ions. The weight percentage of deposited Pd on
Fe3O4catechol nanoparticles was found to be around 0.65%.
It is reasonable to assume that this value should not be very
different from that achieved through hydroxyl groups using
Fe3O4dopPPh2. Thus, the percentage of Pd immobilized on
Fe3O4dopPPh2 nanoparticles until 1.31%, about 0.65%,
would be attributed to the phosphine linkers. However, to
be more precise, and considering that catechol molecules
are less voluminous than dopPPh2, we thought that the
number of catechol ligands incorporated on the surface
should be higher than that of dopPPh2 for nanoparticles of
the same size. Accordingly, the Fe3O4dopPPh2 nanoparticles
would offer more available hydroxyl groups for the direct
reaction. Indeed, using bare Fe3O4 nanoparticles obtained in
the same preparation, the weight percentage of the ligands
was found to be 1.14% for catechol and 3.88% for
dopPPh2. Therefore, 1 mg of nanoparticles incorporates
1.04ꢆ10À1 mmol of catechol molecules or 8.79ꢆ10À2 mmol of
dopPPh2. Consequently, the assumed Pd loading (0.65%) on
the Fe3O4dopPPh2 through hydroxyl should be increased
about 15% up to 0.75%. Thus, the results obtained for the
Pd deposition through hydroxyl (0.70–0.80%) and through
phosphine ligands (0.50–0.60%) roughly indicate that both
routes contribute similarly, although with a slightly higher
participation of the former.
ited on the magnetite surface. The Au content in
Fe3O4dopPPh2@Au was estimated to be 1.89%.
Rh was immobilized onto Fe3O4dopPPh2 nanoparticles by
reaction with RhCl3 in water. In this experiment, the RhIII
ion oxidizes the phosphine to phosphine oxide, inhibiting its
coordination ability. Thus, once neutralized this metal load-
ing route it is reasonable to obtain a smaller Rh content,
0.53%, than that found for Pd and Au in the same reaction
conditions.
In line with the mechanism suggested by Rossi and co-
workers[20] we propose that PdII and AuI complexes react
with the phosphine groups of the linkers grafted on the sur-
face of the magnetite to give PdCl2(phosphine)2 or AuCl(-
phosphine) terminal units. The chelate palladium coordina-
tion is supported by the ratio Pd/dopPPh2 obtained from an-
alytical data. Besides, a fraction of metal ions are in turn
grafted onto the surface through hydroxyl groups. Reduction
of metal ions produces deposition of small metal nanoparti-
cles, whereas the phosphine ligands as well as the hydroxyl
groups are left free to again accept a new load of metals
(Scheme 2).
To support this mechanism, two new loads of Pd were car-
ried out successively on Fe3O4dopPPh2@Pd nanoparticles.
Indeed, the analytical data showed that the total content of
Pd was almost doubled (2.46%) (Fe3O4dopPPh2@2ꢆPd)
and tripled (3.80%) (Fe3O4dopPPh2@3ꢆPd), respectively, in
relation to the weight percentage obtained with only one
load (1.31%), confirming the effectiveness of our predic-
tion.
Characterization of Fe3O4@M and Fe3O4dopPPh2@M (M=
Pd, Au, Rh) nanoparticles: Magnetic characterization of the
new nanoparticles was carried out by magnetometry at
300 K by using a superconducting quantum interference
device that revealed the superparamagnetic behavior of all
the metal-supported nanoparticles. No hysteresis loops and
no remnant magnetization were detected at 300 K in any
case. The first relevant feature relies on its very high satura-
tion magnetization MS at room temperature of these species
(Figure 1).
We extended our study to Au and Rh metals following
the same methodology. The starting gold compound was
[AuCl
ACHTUNGTRENNUNG(tht)] (tht=tetrahydrothiophene) in dichloromethane.
After reduction with NaBH4 gold nanoparticles were depos-
Scheme 2. Formation of Fe3O4dopPPh2@Pd. Incorporation of K2PdCl4 occurs 1) through coordination to phosphine ligands and 2) through bonding to hy-
droxyl surface groups. The final step consists in the reduction of palladium with NaBH4 and formation of Pd nanoparticles.
&
4
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Magnetically Recoverable and Reusable Catalysts
FULL PAPER
dopPPh2 is oxidized by the RhIII ions preventing its coordi-
nating capability.
The surface composition of the metal-supported nanopar-
ticles has been analyzed by X-ray photoelectron spectrosco-
py (XPS). Peaks corresponding to carbon, phosphorous,
iron, palladium, or rhodium are clearly observed. In all the
spectra, two oxidation states are shown by Fe, as expected
for Fe3O4.
Figure 2 shows the Pd 3d spectrum of the nanoparticles
obtained after one load of Pd to Fe3O4dopPPh2 nanoparti-
cles. It can be seen that the spectrum showed the peak bind-
ing energies of 335.3 eV (Pd 3d5/2) and 340.5 (Pd 3d3/2),
Figure 1. Mass magnetization (M) as a function of applied field (H) for
different magnetic nanocomposites.
As expected, the MS of the naked Fe3O4 and
Fe3O4dopPPh2 nanoparticles are almost identical and very
similar to the species containing coordinated PdII ions. How-
ever, the metal immobilization on the surface reduces the
MS. This effect is more evident after loading the nanoparti-
cle three times. After the third load the MS measured
44.6 emugÀ1. Even with this reduction, the solid can be effi-
ciently separated from solution with a small neodymium
permanent magnet. The MS values and magnetic behavior
of the Rh- and Au-supported nanoparticles are similar to
the Pd ones.
The TEM images of the Fe3O4dopPPh2@M (M=Pd, Au)
show approximately spherical nanoparticles with an average
particle size around 12 nm (see Figure S3a and b in the Sup-
porting Information). In addition, smaller metal Pd or Au
nanoparticles appear to be immobilized on the magnetite
surface. The mean diameter of the Pd nanoparticles was es-
timated to be about 2 nm (see Figure S3a in the Supporting
Information,), whereas that of Au was larger (5–30 nm; Fig-
ure S3b in the Supporting Information). They are clearly
distinguishable from the magnetic nanoparticles by their
stronger contrast. On the contrary, the sample obtained by
the direct method using bare Fe3O4 nanoparticles contains
larger Pd0 nanoparticles (of about 2–6 nm; Figure S4 in the
Supporting Information).
A comparison between the high resolution (HR)TEM
images obtained from the Fe3O4dopPPh2@Pd samples after
one, two, and three Pd loads evidenced that the increased
Pd load translates to a larger number of Pd nanoparticles
displaying similar size. It is clear that the Pd nanoparticles
formed during the first load do not act as seeds to favor
their growth in consecutive Pd additions (see Figure S5 in
the Supporting Information).
The TEM image of the Fe3O4dopPPh2@Rh sample (see
Figure S6 in the Supporting Information) shows a group of
nanoparticles exhibiting strong size dispersion (from about 2
to 8 nm), in good agreement with those nanoparticles descri-
bed in the literature obtained without the use of functional-
ized protecting linkers. Note that, as mentioned above, the
Figure 2. XPS spectra of Fe3O4dopPPh2@Pd nanocomposites.
which confirmed the presence of Pd0. Moreover, it can be
also seen the Pd 3d5/2 and 3d3/2 peaks corresponding to the
PdII ions with photoelectron energies of 337.2 and 342.4 eV,
which indicates the presence of PdII ions, probably resulting
from the oxidation of the Pd nanoparticles upon exposure
to the air. The atomic ratio Pd/Fe found was 0.062. For com-
parison purposes, we took also the XPS spectrum of the
Fe3O4@Pd obtained through direct reaction with the bare
magnetite nanoparticles. The spectrum is similar and shows
peak binding energies at 335.5 and 340.8 eV along with
those resulting of the presence of PdII. However, the Pd/Fe
ratio is clearly smaller (0.015). Taking into consideration
that the content of Pd in both types of nanoparticles is simi-
lar (about 1.20–1.30%), the smaller ratio Pd/Fe found in
this sample indicates that the Pd nanoparticles should be
larger since the internal Pd atoms in the nanoparticles are
not seen in the XPS experiment. The XPS spectra of the Pd-
supported nanoparticles after one and three loads is also rel-
evant in that the Pd/Fe atomic ratio increases three times
from 0.062 (one load) to 0.183 (three loads). These data
permit to rule out growing of the Pd nanoparticles, in ac-
cordance to the HRTEM findings (see the Supporting Infor-
mation). The XPS spectrum of the Fe3O4dopPPh2@Rh evi-
dences also two oxidation states of the Rh, Rh0 [307.5 (Rh
3d5/2) and 312.3 (Rh 3d3/2)] and RhIII [309.0 (Rh 3d5/2) and
313.7 (Rh 3d3/2)] (see Figure S7 in the Supporting Informa-
tion). In this sample the smaller Rh content agrees with the
metal content obtained by ICPoes analysis and with the
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

I. Angurell, O. Rossell et al.
poorer dispersion of the particles as compared with those of
Pd.
The system containing both Pd and Rh nanoparticles was
formed similarly, by using K2PdCl4 and RhCl3 as reagents
and water as the solvent. It is interesting to underline that
the weight percentage of the deposited metals is strongly af-
fected by the order in which the reagents are added. As
commented above, the RhIII ions oxidize the dopPPh2
ligand, inhibiting it for further metal coordination. Thus, the
Pd content estimated by ICPoes was 1.28% when the Pd
metal was the first to be added and only 0.91% when it was
added second, whereas the percentage of Rh was similar in
both samples (0.50–0.60%), as expected. Due to the small
size of the Pd and Rh nanoparticles it is very difficult to
detect them in the overview TEM image of Figure 4a.
Nonetheless, this becomes possible at higher magnifications.
The HRTEM image of Figure 4b shows the presence of two
Pd/Rh nanoparticles highlighted by arrowheads. Their diam-
eters are typically found in the 1–4 nm size range. The
HAADF-STEM image of Figure 4c shows the presence of
small, bright Pd/Rh nanoparticles on the ferrite surfaces.
However, due to the small atomic-number difference be-
tween Pd (Z=46) and Rh (Z=45) it is not possible to dis-
criminate between them by using HAADF-STEM imaging.
The heterotrimetallic sample containing Pd, Au, and Rh
was obtained by the same procedure. In this experiment, the
first metal to be deposited was Pd, followed by Au and Rh.
Magnetic nanoparticles containing smaller nanoparticles of
the three metals with an adequate size to be observed and
analyzed by TEM were formed after three impregnations of
Pd, one of Au, and another three of Rh. The weight percent-
age of the metals was 3.45, 1.83, and 1.36 for Pd, Au, and
Rh, respectively; that is, 1 mg of magnetic nanoparticles
contains on average 0.32, 0.09, and 0.13 mmol of Pd, Au, and
Rh, respectively.
Magnetic nanoparticles containing two or three different
metal atoms deposited by successive metal loadings: To
date, multicomponent hybrid materials composed of binary
noble metals have become a hot topic in the area of materi-
als science due to the synergism interaction between their
compositions, which it is expected to have superior proper-
ties to their individual counterparts.
The procedure followed to load metals on the
Fe3O4dopPPh2 nanoparticles opens the possibility of deposit-
ing two or more different metals on their surface making
use of the free phosphine ligands and the hydroxyl groups li-
berated after reduction of the first metal. In fact, this strat-
egy is an extension of that followed for the three successive
Pd loads on the Fe3O4dopPPh2 nanoparticles. Our target
hybrid material was nanoparticles containing separated Pd
and Au nanoparticles immobilized on the surface of the
magnetic nanoparticles. The procedure involved the previ-
ous deposition of Pd, according to the method described
before, and, once the particles were isolated, further reac-
tion with [AuClACHTUNGTRENNUNG(tht)] in dichloromethane, and reduction.
The metal content found was 1.25 (Pd) and 1.81% (Au). An
overview of the Fe3O4dopPPh2@Pd/Au sample is shown in
the TEM image of Figure 3a. The Au nanoparticles are
clearly distinguishable from the ferrite nanoparticles by
their darker contrast. They are often twinned and present
diameters in the 5–30 nm size range. On the other hand, the
Pd nanoparticles are much smaller; their mean diameter
was determined to be about 2 nm. One of these particles is
highlighted with an arrowhead in Figure 3b. The bimetallic
Pd/Au ferrite sample was also investigated by high-angle an-
nular dark field scanning transmission electron microscopy
(HAADF-STEM) image. The intensity of HAADF-STEM
images approximately scales with the square of the atomic
number Z and with the thickness of the sample. Thus, be-
cause of its favorable atomic-number contrast it is possible
to distinguish the Au (Z=79) nanoparticles from the Pd
(Z=46) nanoparticles: the brighter particle consists of Au,
whereas the smaller ones (indicated with white arrowheads
in Figure 3c) consist of Pd.
An overview of the heterotrimetallic sample is shown in
the HAADF-STEM image of Figure 5a. Because of its fa-
vorable atomic-number contrast it is possible to distinguish
the Au (Z=79) nanoparticles from the Rh/Pd (Z=45:46)
nanoparticles: the brighter particles (indicated with white ar-
rowheads in Figure 5a) consist of Au, whereas the smaller
ones consist of Rh and/or Pd. The majority of the brighter
Au particles are in the 5–20 size range, whereas the small
Rh/Pd particles are in the 1–3 nm size range.
Figure 3. a) TEM, b) HRTEM, and c) HAADF-STEM images of Fe3O4dopPPh2@Pd/Au. White arrowheads indicate the presence of Pd nanoparticles.
&
6
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Magnetically Recoverable and Reusable Catalysts
FULL PAPER
Figure 4. a) TEM, b) HRTEM, and c) HAADF-STEM images of Fe3O4dopPPh2@Pd/Rh. White arrowheads indicate the presence of Pd and Rh nanopar-
ticles.
range, two spectral peaks are characteristic for Rh: a large
peak at 2.696 KeV (Rh La1) and a smaller one at 2.834
KeV (Rh Lb1), as observed in the spectrum II. Similarly,
the Pd spectral peaks are found at 2.838 KeV (Pd La1) and
at 2.990 KeV (Pd Lb1). Therefore, there is a strong overlap
between the Rh Lb1 and Pd La1 peaks. Nevertheless, from
the other two spectral peaks it is still possible to distinguish
between Rh and Pd. As an example, spectrum I exhibits
both Pd peaks and an additional Rh La1 peak at 2.696 KeV.
This proves that particle I consists mainly of Pd with a rela-
tively smaller content of Rh. Spectra obtained from several
tenths of particles revealed that the small nanoparticles
either consist of pure Rh, pure Pd, or Pd with different
amounts of Rh (see Figure 5c and Figure S8 in the Support-
ing Information). Additionally, the EDX spectra obtained
from the larger (and brighter) nanoparticles evidence that
they not only consist of Au, but contain small amounts of
Pd and Rh (see spectrum III of Figure 5c and Figure S8 in
the Supporting Information). To determine the spatial distri-
bution of Pd and Rh present in the Au particles, we ac-
quired EDX spectrum images of several bright particles.
Figure 6 shows the Au, Pd/Rh, and Fe elemental maps ex-
tracted from an EDX spectrum image obtained from the
area marked with a white rectangle in the accompanying
Figure 5. a) HAADF-STEM image of Fe3O4dopPPh2@Pd/Au/Rh. White
arrowheads indicate the presence of Au nanoparticles. The region high-
lighted with a square is magnified in panel b; b) EDX spectra were ob-
tained from the nanoparticles marked as I–III; c) Corresponding EDX
spectra.
The composition of these nanoparticles was assessed by
Figure 6. HAADF-STEM image of Fe3O4dopPPh2@Pd/Au/Rh and ele-
energy dispersive X-ray (EDX) spectroscopy. Figure 5c ex-
mental maps calculated from the EDX image obtained from the area
outlined with a white rectangle. The RGB map is generated using red for
Pd/Rh, green for Au, and blue for Fe.
hibits three EDX spectra obtained from the particles
marked with circles in Figure 5b. In the displayed energy
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

I. Angurell, O. Rossell et al.
HAADF-STEM image. The EDX data reveal that the
bright nanoparticle located at the top of the maps consists
of an Au core and a Pd/Rh shell. This becomes clear when
observing the red, green, blue (RGB) color map. The three
small nanoparticles at the bottom of the image consist of Pd
and Rh. Due to the strong overlap of Rh Lb1 and Pd La1
peaks it was not possible to spatially separate the Rh from
the Pd signal with our experimental setup.
Additional information from the Pd/Rh nanoparticles was
obtained by HRTEM. Figure 7a is a low magnification
TEM image of the Fe3O4dopPPh2@Pd/Au/Rh sample show-
ing the presence of several small Pd/Rh nanoparticles on the
surface of the magnetite particles. A closer look at these
nanoparticles reveals that they are fcc single-crystalline par-
ticles with well-defined truncated octahedral morphologies
(Figure 7b). On the other hand, the larger Au (Pd/Rh) parti-
cles present roundish morphologies and are often twinned
(not shown here).
Figure 7. a) TEM image of Fe3O4dopPPh2@Pd/Au/Rh showing the pres-
ence of Pd/Rh nanoparticles on the surface of the magnetite nanoparti-
cles. b) HRTEM images of three Pd/Rh particles viewed along their [110]
zone axis.
À
Catalytic studies: Suzuki–Miyaura C C coupling reaction:
The palladium-catalyzed Suzuki–Miyaura cross-coupling of
aryl halides and arylboronic acids to give biaryl structures
have found applications in fine chemical and pharmaceutical
synthesis.[21]
and absence of the phosphine ligand, but no coupling reac-
tion was observed in any case.
The recyclability of Fe3O4dopPPh2@Pd nanoparticles was
investigated by treating phenylboronic acid with 4-bromoa-
nisole. After completion of the reaction the magnetic cata-
lyst could easily be recovered from the reaction mixture
with an external magnet and reused directly in a subsequent
The catalytic activity of the Fe3O4dopPPh2@Pd nanoparti-
cles was tested in Suzuki–Miyaura cross-coupling reactions
of the aryl halides listed in Table 1 and phenylboronic acid
(Scheme 3).
In a typical experiment, phenylboronic
Table 1. Substrate scope of the Suzuki–Miyaura coupling with phenylboronic acid.[a]
acid (1.2 equiv), potassium hydroxide
(2 equiv) and the catalyst were added to
aryl bromide (1 equiv) in toluene. After 6 h
of heating, the reaction mixture was cooled
down to room temperature and then an ex-
ternal magnet was used to separate the cata-
lyst from the reaction mixture.
Entry Substrate
Product
Yield[b] Biphenyl[c] Selectivity[d]
[%]
[%]
[%]
1
2
3
4
100
8
92
100
100
58
0
0
0
100
100
100
The most relevant results in terms of con-
version and selectivity were obtained by
using 4-substituted bromoarenes bearing
electron-withdrawing groups (entries 1–3,
Table 1), but for the substrates bearing elec-
tron-donating groups the yield dropped sig-
nificantly, as expected.[22] The scope of the
reaction was expanded to other challenging
chloride derivatives (entries 9 and 10,
Table 1). Unfortunately, they showed much
less reactivity than the bromide counter-
parts in good accord with the literature.[23]
An important feature of our catalytic
system is that it does not need any addition-
al phosphine ligand to carry out the reac-
tions, due to the pendant phosphine ligands.
To exclude any activity related to the cat-
alyst support, the reaction of phenylboronic
acid and 4-bromoanisole was carried out in
the same conditions of the model reaction,
without metal catalyst and in the presence
5
6
7
32
25
11
7
0
0
78
100
100
8
7
7
0
9
5
4
0
2
100
50
10
[a] Reaction conditions: substrate (1 equiv), PhB(OH)2 (1.2 equiv), KOH (2 equiv), Pd
(0.5 mol%), toluene, 1008C, 6 h. [b] Based on the substrate. [c] Based on the initial amount
of phenylboronic acid. [d] Selectivity towards heterocoupling.
&
8
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Magnetically Recoverable and Reusable Catalysts
Scheme 3. Suzuki–Miyaura cross-coupling reaction.
FULL PAPER
(M=Au, Pd) nanoparticles
with that containing a mixture
of
both
metals,
Fe3O4dopPPh2@Pd/Au nano-
particles.
reaction. The nanoparticles kept similar conversion and se-
lectivity after 10 successive runs.
The rapid reduction of 4-nitrophenol to 4-aminophenol is
monitored by time-resolved UV/Vis spectra as shown in
The heterogeneous nature of this catalyst system was
demonstrated by the negligible loss of palladium around of
0.20% of the total metal present in the catalyst, in the recy-
cling process after ten runs. Moreover, a hot-magnetic sepa-
ration experiment was also performed. Thus, the reaction of
4-bromoanisole and phenylboronic acid was carried out at
1008C. After 30 minutes the solid catalyst was separated
magnetically and the filtrate was transferred to another
Schlenk flask. Without the presence of nanoparticles the re-
action progressed only an additional 5% after 5 h (Figure 8)
thus evidencing the fact that leaching of palladium scarcely
contributes to the catalytic activity of the nanoparticles.
Figure 9. The addition of
a
trace amount of
Figure 9. UV/Vis absorption spectra of the catalytic reduction of 4-nitro-
phenol over Fe3O4dopPPh2@Au/Pd nanoparticles.
Fe3O4dopPPh2@M (M=Pd, Au or Pd/Au), into the 4-nitro-
phenol/NaBH4 alkaline resulted in a rapid release of H2
bubbles accompanied with decrease of the absorption at
400 nm and the appearance of other absorption at 317 nm
with time, indicating the transformation from 4-nitrophenol
to 4-aminophenol.
The catalytic reduction of these nanocomposites was also
examined in the absence of NaBH4, and no decrease in the
absorption can be seen at approximately 400 nm even after
one hour, which suggests this has a negligible effect on the
reduction kinetics data.
Figure 8. Leaching test for Fe3O4dopPPh2@Pd in Suzuki–Miyaura cross-
&
*
coupling reaction: ( ) without catalyst and ( ) with catalyst.
Since the concentration of NaBH4 exceeds that of 4-nitro-
Catalytic reduction of 4-nitrophenol over Fe3O4dopPPh2@M
(M=Pd, Au, Pd/Au): The reduction of nitro compounds to
amines is one of the most important chemical reactions be-
cause organic amines are essential materials for the produc-
tion of agrochemicals, dyes, pharmaceuticals, polymers, and
rubbers.[24]
In several papers, liquid-phase reduction of 4-nitrophenol
to 4-aminophenol by NaBH4 (Scheme 4) has been used as a
model reaction to evaluate the catalytic potential of the
nanoparticles employed.[25] This is because the substrates
and products are easily detected by spectroscopic methods,
and there is no appreciable byproduct formation.
phenol (cACTHGNURETNNU(G NaBH4)/c(4-nitrophenol)=100:1), the reduction
can be considered as a pseudo-first-order reaction with
regard to 4-nitrophenol only. Figure 10 shows the linear rela-
tionship between lnACTHNUTRGNEUNG(ct/c0) and the reaction time t in reduc-
tions catalyzed by three different catalyst systems.
As all these plots match the first-order reaction kinetics
very well, and the rate constant k can be calculated from
the rate equation lnACTHNUTRGNEUNG(ct/c0)=kt (Table 2). The rate constants
evidence higher catalytic efficiency for the Pd/Au nanocata-
lyst.
Since recyclability is crucial for practical applications of
nanocatalysts, we have checked the reuse of our best cata-
lyst, Fe3O4dopPPh2@Pd/Au. The results show that the cata-
lyst can be successfully recycled and reused in five succes-
sive reactions, all with conversions of 100% within 2 min.
Then, the conversion rate started to drop slowly.
On the other hand, we were interested in comparing for
this process the catalytic behavior of the Fe3O4dopPPh2@M
Hydrogenation of styrene catalyzed by Fe3O4dopPPh2@Rh:
Scheme 4. Reduction of 4-nitrophenol to 4-aminophenol.
Rhodium-catalyzed hydrogenation is frequently used not
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

I. Angurell, O. Rossell et al.
ditional experiment was carried out using bare Fe3O4 nano-
particles, that is, in the absence of rhodium. In these condi-
tions, the hydrogenation just reached 11% after a reaction
time of 8 h.
Sequential tandem reaction by using Fe3O4dopPPh2@Pd/Au
nanoparticles: In view of these precedents and taken into ac-
À
count that Au and Pd nanoparticles catalyze C C cross-cou-
pling reactions as well as the reduction of 4-nitrophenol to
4-aminophenol, we designed a single tandem reaction com-
bining both processes (Scheme 6). Fe3O4dopPPh2@Pd/Au
nanoparticles were chosen as catalyst on the basis of its ex-
cellent behavior in the reduction of 4-nitrophenol.
In the first step, 1-bromo-4-nitrobenzene (1 equiv), phenyl
boronic acid (1.2 equiv), potassium hydroxide (2 equiv), and
catalyst (Pd 1 mol%) were dispersed with 4 mL of toluene
and heated at 1008C during 6 h. After this time, 1 mL of an
aqueous solution of NaBH4 (10 equiv) was added to the re-
action mixture. The conversion to the amino compound was
quantified by gas chromatography. After the first cycle, the
catalytic efficiency was very high, around 94% of the target
1-aminobiphenyl. However, the reusability of the catalyst
was rather poor since after the third recycle the conversion
fell to 54%. Partial agglomeration of the nanoparticles is
probably responsible for the loss of their catalytic activity
(see Figure S9 in the Supporting Information). Nonetheless,
to our knowledge this is the first example of a tandem reac-
tion catalyzed by bimetallic-supported nanoparticles in
which both metals take part in the process.
Figure 10. Relationship of lnACTHNUTRGNEUNG(ct/c0) and reaction time for the reduction of
&
( )
4-nitrophenol
catalyzed
by
the
three
nanoparticles;
~
*
Fe3O4dopPPh2@Pd/Au; ( ) Fe3O4dopPPh2@Pd; ( ) Fe3O4dopPPh2@Au.
ct and c0 represent the concentration of 4-nitrophenol at time t and at the
beginning of the reaction, respectively.
Table 2. Comparison of kinetic constant (k) and R2 of 4-NP reduction
using different catalysts.
Catalyst
k [sÀ1
]
R2
Fe3O4dopPPh2@Pd/Au
Fe3O4dopPPh2@Pd
Fe3O4dopPPh2@Au
0.047
0.020
0.005
0.992
0.971
0.976
only in laboratories but also in industry and the possibility
of depositing rhodium on the surface of magnetite nanopar-
ticles has increased the interest
in this area.[26]
In this paper we have inves-
tigated the rhodium-supported
Scheme 6. Sequential catalytic reaction: 1) Suzuki–Miyaura cross-coupling reaction and 2) reduction of 1-nitro-
biphenyl to 1-aminobiphenyl.
Fe3O4dopPPh2@Rh nanoparti-
cles as a catalyst in the hydro-
genation of styrene in di-
chloromethane
(Scheme 5).
Conclusion
The reactor loaded with the catalyst and the substrate was
subjected to 20 bars pressure of hydrogen and held at 508C.
After 5 h, the nanocomposites were easily separated by
using a magnet and reused in the next reaction. Table 3 illus-
trates excellent catalytic activity of the rhodium catalyst that
was reused up to five times without loss of efficiency. An ad-
We have shown that the functionalization of ferrite nanopar-
ticles with phosphine-terminated linkers permits the facile
loading of noble metals (Pd, Au, or Rh) onto their surfaces.
Thus, homometallic ferrite nanoparticles containing Pd, Au,
or Rh have been obtained after addition of the appropriate
metal salt to the ferrite functionalized nanoparticles, fol-
lowed by reduction with NaBH4. Ferrite bimetallic (Pd/Au
or Pd/Rh) species were obtained by successive metal salt ad-
ditions followed by the corresponding reduction process.
The same approach was applied to obtain a trimetallic Pd/
Au/Rh ferrite hybrid. The catalytic properties of some of
these metal systems are evidenced in several processes; in
particular, the Au/Pd ferrite nanoparticles have been proven
to be efficient for the sequential process involving a cross-
coupling reaction followed by reduction of 4-nitrobiphenyl.
Scheme 5. Hydrogenation of styrene.
Table 3. Reusability test for Fe3O4dopPPh2@Rh in the hydrogenation of
styrene[a]
.
Recycle run
Yield [%]
1
2
3
4
5
96.4
97.6
99.2
97.2
97.5
[a] Reaction conditions: Styrene (0.15 mmol), catalyst (0.75 mmol Rh), di-
chloromethane (5 mL), 508C, 20 bar H2, 5 h.
&
10
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Magnetically Recoverable and Reusable Catalysts
FULL PAPER
À
À
À
258C): d=À5.8 ppm (s); IR (KBr): n˜ =3419 (O H+N H), 3068 (arC
Experimental Section
3
À
À
À
À
À
H), 2928 (Csp H), 1637 (N C=O), 1597 (arC C), 1261 (P C), 805 (N
H); HRMS (ESI): m/z calcd for [C54H48N2O6P2ÀH]À: 881.29; found:
General methods: All manipulations were performed under purified ni-
trogen using standard Schlenk techniques. Chloroform, acetone, and
methanol were distilled from appropriate drying agents. Dichlorome-
thane, toluene, and hexane were purchased from Scharlau (Multisolvent
HPLC grade) and purified by Puresolv column from Innovative Technol-
ogy, INC. Deionized water was obtained from Milipore Helix 3 water pu-
À
À
881.29 [2M H] ; elemental analysis calcd (%) for C27H24NO3P: C 73.46,
H 5.48, N 3.17, O 10.87, P 7.02; found: C 73.34, H 5.47, N 3.16.
Synthesis of Fe3O4 nanoparticles: The magnetite Fe3O4 nanoparticles
were prepared following the procedure published by Varma and Polshet-
solid mixture of iron(II) sulfate heptahydrate (6.95 g,
25.25 mmol) and iron(III) sulfate (10.0 g, 25.78 mmol) were dissolved in
tiwar.[1c]
A
rification system. H, 13C{1H} and 31P{1H} NMR spectra were obtained on
1
AHCTUNGTRENNUNG
water (250 mL) in a 500 mL Shlenk flask. Then, an ammonium hydroxide
solution (25%) was slowly added to adjust the pH solution to 10. The re-
action mixture was continually stirred for 1 h at 508C. The precipitated
nanoparticles were separated magnetically and washed with water until
the pH reached 7. Finally the nanoparticles were dried under reduced
pressure at 608C for 2 h, and obtained as a bright black powder. IR
Varian Unity 300 and Varian Mercury 400 spectrophotometers. Bidimen-
sional NMR spectra (HSQC 1H/13C{1H}) was recorded on a Varian Mer-
cury 400 spectrophotometer. Chemical shifts are reported in ppm relative
1
to external standards (SiMe4 for H and 13C{1H}), and coupling constants
are given in Hz. MS ESI(À) spectra was recorded in a LC/MSD-TOF
(Agilent Technologies) spectrometer. The high-resolution transmission
electron microscopy (HRTEM) data was obtained at 200 kV either with
a JEOL 2100 microscope, having a point-to-point resolution of 0.21 nm,
or with a JEOL 2200FS TEM/STEM microscope, having a point-to-point
resolution of 0.23 nm. The scanning transmission electron microscopy
(STEM) and the energy-dispersive X-ray (EDX) spectroscopy data was
acquired with the JEOL 2200FS TEM/STEM, which is equipped with a
Gatan Digiscan system. Samples were prepared by placing a drop of solu-
tion on a holey-carbon-coated Cu TEM grid and allowing the solvent to
evaporate in air. In addition, for the STEM-EDX experiments, the sam-
ples were gently plasma cleaned with an Ar (75%)/O2 (25%) plasma for
a few seconds to remove hydrocarbon contamination from the surfaces
of the nanoparticles. Absorption spectra were recorded on a Hewlett–
Packard HP 8453 UV/Vis spectrophotometer. Infrared spectra of the
samples were collected in the transmission mode on a Nicolet Avatar 330
FT-IR spectrophotometer. Magnetic measurements were carried out at
the Unitat de Mesures Magnꢃtiques of the Centres Cientꢂfics i Tecnolꢇg-
ics of the Universitat de Barcelona (CCiTUB) on a Quantum Design
SQUID MPMS-XL susceptometer. ICPoes (inductively coupled plasma
À
À
À
(KBr): n˜ =3412 (O H), 1618 (O H), 573 (Fe O); HRTEM: (12.4Æ
0.9) nm.
Synthesis of Fe3O4dopPPh2 nanoparticles: Fe3O4 nanoparticles (400 mg)
were sonicated in methanol (15 mL) for 30 min. Then, a solution of 1
(400 mg, 0.906 mmol) in methanol (15 mL) was added to the suspension
of nanoparticles. The mixture was sonicated for 2 h. After this time, the
nanoparticles were washed several times with methanol and acetone to
remove the excess of 1. Finally, the nanoparticles were extracted with an
external magnet (neodymium permanent magnet) and dried under re-
À
À
À
duced pressure for 1 h. IR (KBr): n˜ =3442 (O H+N H), 2983 (arC H),
3
À
À
À
À
2927, 2858 (Csp H), 1633 (N C=O), 1489 (arC C), 580 (Fe O); ele-
mental analysis found: C 2.85, H 0.21 (3.88% of dopPPh2).
Synthesis of Fe3O4catechol nanoparticles: The nanoparticles were pre-
pared using the same procedures as for the Fe3O4dopPPh2 nanoparticles,
using catechol (400 mg, 3.633 mmol). Elemental analysis found: C 0.75,
À
À
H 0.06 (1.14% of catechol); IR (KBr): n˜ =1487 (arC C), 1265 (C O),
À
570 (Fe O).
Synthesis of Fe3O4dopPPh2@Pd nanoparticles: Fe3O4dopPPh2 nanoparti-
cles (100 mg) were sonicated in water (12 mL) for 10 min. Then, potassi-
um tetrachloroplatinate(II) (5.0 mg, 12.22 mmol) was added to the sus-
pension and the mixture was stirred (900 rpm) for 2 h. After this time,
the nanoparticles were washed three times with water and acetone and
dried under reduced pressure for 30 min. In a second step, the nanoparti-
cles were dispersed again in water (11 mL) and an aqueous solution of
sodium borohydride (1.0 mL, 0.07m) was added to reduce the palladium.
The suspension was stirred (1200 rpm) for 2 h. Then, the nanoparticles
were washed several times with water and acetone, extracted with an ex-
optical emission spectrometry) measurements were carried out in
a
Perkin–Elmer Optima 3200RL spectrometer. All the reagents were pur-
chased from suppliers (Fluka and Sigma–Aldrich) and used as received.
Synthesis of dopPPh2 (1): N-(3-Dimethylaminopropyl)-N’-ethylcarbodi-
ACHTUNGTRENNUNGimide hydrochloride (220 mg, 1.149 mmol) and N-hydroxisuccinimide
(169 mg, 1.437 mmol), were added to a mixture of 4-(diphenylphosphi-
no)benzoic acid (352 mg, 1.149 mmol) in anhydrous dimethylformamide
(20 mL). The mixture was stirred for 2 h. Then, dopamine hydrochloride
(182 mg, 0.960 mmol) dissolved in an aqueous solution of sodium hydro-
gencarbonate (10 mL, 0.10m) was added to the mixture and stirred over-
night. After this time, chloroform (50 mL), water (50 mL), and a saturat-
ed solution of brine (50 mL) were added to the reaction mixture. The or-
ganic phase was extracted and the aqueous phase was washed twice with
chloroform (50 mL). The whole organic phases were washed with water
(100 mL) and dried with anhydrous magnesium sulfate. The solvent was
eliminated under reduced pressure yielding a bright-green solid. The
product was purified by chromatography on a silica gel column with
ethyl acetate/methanol (20:1) as eluent. The desired compound was ob-
tained as white solid. Yield: 395 mg (76.8%). 1H NMR (400.1 MHz,
ternal magnet and dried under reduced pressure for 1 h. IR (KBr): n˜ =
3
À
À
À
À
À
3431 (O H+N H), 2962 (arC H), 2923, 2852 (Csp H), 1632 (N C=
À
À
O), 1489 (arC C), 579 (Fe O); ICPoes (%): Pd 1.31.
Synthesis of Fe3O4@Pd nanoparticles: The nanoparticles were prepared
using the same procedure as for Fe3O4dopPPh2@Pd nanoparticles, using
Fe3O4 nanoparticles (100 mg) instead of Fe3O4dopPPh2 nanoparticles. IR
À
À
À
(KBr): n˜ =3406 (O H), 1630 (O H), 575 (Fe O); ICPoes (%): Pd 1.10.
Synthesis of Fe3O4catechol@Pd nanoparticles: The nanoparticles were
prepared using the same procedure as for Fe3O4dopPPh2@Pd nanoparti-
cles, using Fe3O4catechol nanoparticles (100 mg) instead of Fe3O4dopPPh2
nanoparticles. ICPoes (%): Pd 0.65.
CDCl3, 258C): d=7.59 (dd, J
7.53–7.26 (m, 14H; PAr-H), 6.74 (d, 3J
4J(H,H)=3 Hz, 1H; CH), 6.56 (dd, 4J
1H; CH), 6.11 (t, 3J
(H,H)=9.2 Hz, 1H; NH), 3.63 (pq, J
2H; CH2), 2.77 ppm (t, 3J(H,H)=9.2 Hz, 2H; 2H); 1H NMR
A
E
AHCTUNGTRENNUNG
Synthesis of Fe3O4dopPPh2@3xPd nanoparticles: The nanoparticles were
prepared using the same procedure as for Fe3O4dopPPh2@Pd nanoparti-
cles. The load and subsequent reduction of palladium was repeated two
more times. ICPoes (%): Pd 3.80.
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
(400.1 MHz, [D4]MeOH, 258C): d=7.72 (dd,
J
G
JACHTUNGTRENNUNG
ꢀ9 Hz, 2H; CCH), 7.39–7.26 (m, 14H; PAr-H), 6.69 (d, 3J
U
Synthesis of Fe3O4dopPPh2@Rh nanoparticles: Fe3O4dopPPh2 nanoparti-
cles (100 mg) were sonicated in water (12 mL) for 10 min. Then,
4
3
1H; CH), 6.68 (d, J
(H,H)=2 Hz, 1H; CH), 6.55 (dd, J
A
N
rhodiumACTHNGUTERNNU(G III) chloride (2.5 mg, 11.95 mmol) were added to the suspension
G
and the mixture was stirred (900 rpm) for 2 h. After this time, the nano-
particles were washed three times with water and acetone and dried
under reduced pressure for 30 min. In a second step, the nanoparticles
were dispersed again in water (11 mL) and an aqueous solution of
sodium borohydride (1.0 mL, 0.07m) was added to reduce rhodium. The
suspension was stirred (1200 rpm) for 2 h. Then, the nanoparticles were
washed several times with water and acetone, extracted with an external
AHCTUNGTRENNUNG
E
E
G
6.5 Hz; mCHP), 119.8 (s; CH), 115.5 (s; CH), 115.0 (s; CH), 41.6 (s;
CH2N), 34.5 ppm (s; PhCH2); 31P{1H} NMR (300.1 MHz, [D4]MeOH,
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

I. Angurell, O. Rossell et al.
magnet and dried under reduced pressure for 1 h. IR (KBr): n˜ =3423
adding 40 mL of the upper solution to a 3 mL of water in a quartz cuv-
ette.
3
À
À
À
À
À
(O H+N H), 2971 (arC H), 2931, 2892 (Csp H), 1630 (N C=O),
À
À
1429 (arC C), 582 (Fe O). ICPoes (%): Rh 0.53.
Sequential catalytic test: For the sequential catalytic reaction, 1-bromo-4-
nitrobenzene (0.12 mmol), phenylboronic acid (0.14 mmol), potassium
hydroxide (0.23 mmol) and Fe3O4dopPPh2@Pd/Au nanoparticles
(1.20 mmol Pd; 0.95 mmol Au) were added in a Schlenk flask, dissolved in
toluene (4 mL) and held for 6 h at 1008C under N2 atmosphere. After
this time, an aqueous solution of sodium borohydride (1 mL, 1.2m) was
added. After 2 min, an external magnetic field was applied for 15 s to
separate the magnetic catalyst. Then, 40 mL of the upper solution was ex-
amined in a CG. This procedure was repeated every 2 min until the com-
plete reduction.
Synthesis of Fe3O4dopPPh2@Au nanoparticles: Fe3O4dopPPh2 nanoparti-
cles (100 mg) were sonicated in of dichloromethane (12 mL) for 10 min.
Then, chloro(tetrahydrothiophene)gold(I) (4.0 mg, 12.48 mmol) was
added to the suspension. The mixture was stirred (900 rpm) for 2 h. After
this time, the nanoparticles were washed three times with dichlorome-
thane and acetone and dried under reduced pressure for 30 min. In a
second step, the nanoparticles were dispersed again in water (11 mL) and
an aqueous solution of sodium borohydride (1.0 mL, 0.08m) was added
to reduce the gold. The suspension was stirred (1200 rpm) for 2 h. Then,
the nanoparticles were washed times with water and acetone, extracted
with an external magnet and dried under reduced pressure for 1 hour. IR
3
À
À
À
À
(KBr): n˜ =3440 (O H+N H), 2970 (arC H), 2930, 2852 (Csp H), 1630
À
À
À
(N C=O), 1483 (arC C), 585 (Fe O). ICPoes (%): Au 1.89.
Synthesis of Fe3O4dopPPh2@Pd/Au nanoparticles: The nanoparticles
were prepared by using the same procedures as for Fe3O4dopPPh2@Pd
nanoparticles and Fe3O4dopPPh2@Au nanoparticles. Firstly, potassium
tetrachloroplatinate(II) was coordinated and reduced step by step, and
secondly chloro(tetrahydrothiophene)gold(I) was coordinated and re-
Acknowledgements
Financial support from MICINN (projects CTQ2012–31335, ENE2012–
36368 and CTQ2010–15292) is acknowledged. J.L. is grateful to the
ICREA Academia program and F.G. to the MICINN for a scholarship.
M.D.R. and R.E. acknowledge the financial support of the Swiss COST
office under the SBF project number C10.0089.
À
À
À
duced step-by-step. IR (KBr): n˜ =3427 (O H+N H), 2969 (arC H),
3
À
À
À
À
2923, 2854 (Csp H), 1630 (N C=O), 1485 (arC C), 577 (Fe O). ICPoes
(%): Pd 1.25, Au 1.81.
Synthesis of Fe3O4dopPPh2@Pd/Rh nanoparticles: The nanoparticles
were prepared using the same procedures as for Fe3O4dopPPh2@Pd
nanoparticles and Fe3O4dopPPh2@Rh nanoparticles. Firstly, potassium
tetrachloroplatinate(II) was coordinated and reduced step by step, and
secondly rhodiumACHTUNGTRENNUNG(III) chloride was coordinated and reduced step-by-
step. ICPoes (%): Pd 1.28, Rh 0.56.
[1] a) D. Astruc, F. Lu, J. Ruiz, Angew. Chem. 2005, 117, 8062–8083;
Angew. Chem. Int. Ed. 2005, 44, 7852–7872; b) D. Astruc, Inorg.
Chem. 2007, 46, 1884–1894; c) V. Polshettiwar, R. S. Varma, Green
Chem. 2010, 12, 743–754; d) A. Balanta, C. Godard, C. Claver,
Chem. Soc. Rev. 2011, 40, 4973–4985.
[2] a) A. G. Hu, G. T. Yee, W. B. Lin, J. Am. Chem. Soc. 2005, 127,
12486–12487; b) A. G. Hu, S. Liu, W. B. Lin, RSC Adv. 2012, 2,
2576–2580; c) M. J. Jin, D. H. Lee, Angew. Chem. 2010, 122, 1137–
1140; Angew. Chem. Int. Ed. 2010, 49, 1119–1122; d) M. Friederici,
I. Angurell, O. Rossell, M. Seco, N. J. Divins, J. Llorca, Organome-
tallics 2012, 31, 722–728; e) M. Friederici, I, Angurell, O. Rossell,
M. Seco, G. Muller, J. Mol. Catal. A 2013, 376, 7–12.
Synthesis of Fe3O4dopPPh2@Pd/Au/Rh: Fe3O4dopPPh2 nanoparticles
(100 mg) were sonicated in water (12 mL) for 10 min. In a first step, po-
tassium tetrachloroplatinate(II) (5.0 mg, 12.22 mmol) was coordinated
and reduced step-by-step using the same procedure as for
Fe3O4dopPPh2@Pd nanoparticles. The load and subsequent reduction of
palladium was repeated two more times. In a second step, the nanoparti-
cles were sonicated in dichloromethane (12 mL), and chloro(tetrahydro-
thiophene)gold(I) (4.0 mg, 12.48 mmol) was coordinated and reduced
using the same procedure as for Fe3O4dopPPh2@Au nanoparticles. Final-
ly, the nanoparticles were sonicated again in water (12 mL) and rhodium-
[3] J. A. Widegren, R. G. Finke, J. Mol. Catal. A 2003, 191, 187–207.
[4] M. V. Vasylyev, G. Maayan, Y. Hovav, A. Haimov, R. Neumann,
Org. Lett. 2006, 8, 5445–5448.
ACHTUNGTRENNUNG(III) chloride (2.5 mg, 11.95 mmol) was coordinated and reduced step-by-
[5] a) J. Dupont, G. S. Fonseca, A. P. Umpierre, P. F. P. Fichtner, S. R.
Teoxeora, J. Am. Chem. Soc. 2002, 124, 4228–4229; b) T. J. Geld-
bach, D. Zhao, N. C. Castillo, G. Laurenczy, B. Weyershausen, P. J.
Dyson, J. Am. Chem. Soc. 2006, 128, 9773–9790; c) S. Jansat, J.
Durand, I. Favier, F. Malbosc, C. Pradel, E. Teuma, M. Gꢈmez,
ChemCatChem 2009, 1, 244–246; d) L. Rodrꢂguez-Pꢉrez, C. Pradel,
P. Serp, M. Gꢈmez, E. Teuma, ChemCatChem 2011, 3, 749–754.
[6] N. Zheng, G. D. Stucky, J. Am. Chem. Soc. 2006, 128, 14263–14278.
[7] a) N. T. S. Phan, C. S. Gill, J. V. Nguyen, Z. J. Zhang, C. W. Jones,
Angew. Chem. 2006, 118, 2267–2270; Angew. Chem. Int. Ed. 2006,
45, 2209–2212; b) A. H. Latham, M. E. Williams, Acc. Chem. Res.
2008, 41, 411–420; c) C. W. Lim, I. S. Lee, Nano Today 2010, 5, 412–
434; d) V. Polshettiwar, R. Luque, A. Fihri, H. Zhu, M. Bouhrara, J.-
M Basset, Chem. Rev. 2011, 111, 3036–3075; e) R. B. Nasir Baig,
R. J. Varma, Chem. Commun. 2013, 49, 752–770; f) R. B. Nasir Baig,
R. J. Varma, Chem. Commun. 2012, 48, 6220–6222; g) R. B. Nasir
Baig, R. J. Varma, Chem. Commun. 2012, 48, 2582–2584; h) R. B.
Nasir Baig, R. J. Varma, Green Chem. 2012, 14, 625–632; i) R. B.
Nasir Baig, R. J. Varma, Green Chem. 2013, 15, 398–417.
[8] a) D. Guin, B. Baruwati, S. V. Manorama, Org. Lett. 2007, 9, 1419–
1421; b) K. K. Senapati, S. Roy, C. Borgohain, P. Phukan, J. Mol.
Catal. A 2012, 352, 128–134.
[9] a) J. Zang, X. Liu, X. Guo, S. Wu, S. Wang, Chem. Eur. J. 2010, 16,
8108–8116; b) L. M. Rossi, I. M. Nangol, N. J. S. Costa, Inorg.
Chem. 2009, 48, 4640–4642.
[10] D. K. Yi, S. S. Lee, J. Y. Ying, Chem. Mater. 2006, 18, 2461–2463.
[11] F. Zhang, J. Jin, X. Zhong, S. Li, J. Niu, R. Li, J. Ma, Green Chem.
2011, 13, 1238–1243.
step using the same procedure as for Fe3O4dopPPh2@Rh nanoparticles.
The load and subsequent reduction of rhodium was repeated two more
times. ICPoes (%): Pd 3.45, Au 1.83, Rh 1.36.
Catalytic hydrogenation of styrene: The catalytic hydrogenation of styr-
ene was carried out in magnetically stirred (700 rpm) Berghof reactor
vessel containing styrene (0.15 mmol), Fe3O4dopPPh2@Rh nanoparticles
(0.75 mmol Rh), dichloromethane (5 mL), and charged to 20 bars of hy-
drogen. The reaction was stirred at 508C during 5 h. After the comple-
tion of the reaction, the product was isolated by removing the catalyst
magnetically from the reaction mixture.
Suzuki–Miyaura catalytic test: Suzuki–Miyaura reactions were carried
out in magnetically stirred (900 rpm) Schlenk flask containing substrate
(0.48 mmol), phenylboronic acid (0.58 mmol), potassium hydroxide
(0.96 mmol), Fe3O4dopPPh2@Pd nanoparticles (2.40 mmol Pd), toluene
(10 mL), and held at 1008C under N2 atmosphere for 6 h.
Catalytic reduction of 4-nitrophenol: The catalytic reduction of 4-nitro-
phenol was carried out in magnetically stirred (1600 rpm) tube containing
4-nitrophenol (0.06 mmol), sodium borohydride (6.00 mmol) in water
(5.0 mL) at room temperature. After adding Fe3O4dopPPh2@Pd/Au nano-
particles (0.08 mmol Pd; 0.06 mmol Au) or Fe3O4dopPPh2@Pd nanoparti-
cles (0.08 mmol Pd), the bright-yellow solution gradually becomes color-
less as the reaction progressed. Catalytic reduction using
Fe3O4dopPPh2@Au nanoparticles was carried out with the same proce-
dure using 4-nitrophenol (0.06 mmol), sodium borohydride (6.00 mmol),
Fe3O4dopPPh2@Au nanoparticles (0.2 mmol Au), and water (5.0 mL).
During the course of reactions, the products were identified and quanti-
fied by
a
Hewlett–Packard HP 8453 UV/Vis spectrophotometer by
&
12
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Magnetically Recoverable and Reusable Catalysts
FULL PAPER
[12] a) V. Polshettiwar, N. Mallikarjuna, N. Nadagouda, R. S. Varma,
Chem. Commun. 2008, 6318–6320; b) C. J. Xu, K. M. Xu, H. W. Gu,
R. K. Zheng, H. Liu, X. X. Zhang, Z. H. Guo, B. Xu, J. Am. Chem.
Soc. 2004, 126, 9938–9939; c) S. Baruwati, V. Polshettiwar, R. S.
Varma, Tetrahedron Lett. 2009, 50, 1215–1218; d) B. R. Vaddula, A.
Saha, J. Leazer, R. S. Varma, Green Chem. 2012, 14, 2133–2136.
[13] a) Q. W. Du, W. Ahang, H. Ma, J. Zheng, B. Zhou, Y. Q. Li, Tetrahe-
dron 2012, 68, 3577–3584; b) S. Shylesh, L. Wang, W. R. Thiel, Adv.
Synth. Catal. 2010, 352, 425–432; c) P. Li, L. Wang, L. Zhang, G. W.
Wang, Adv. Synth. Catal. 2012, 354, 1307–1318.
Gꢈmez, Eur. J. Inorg. Chem. 2008, 3577–3586; f) V. Chandrasekhar,
R. Suriya Narayanan, P. Thilagar, Organometallics 2009, 28, 5883–
5888.
[22] a) B. Baruwati, D. Guin, S. V. Vanorama, Org. Lett. 2007, 9, 5377–
5380; b) D. Rosario-Amorin, X. Wang, N. Gaboyard, R. Clꢉrac, S.
Nlate, K. Heuzꢉ, Chem. Eur. J. 2009, 15, 12636–12643; c) Q. Zhang,
H. Su, J. Luo, Y. Wei, Catal. Sci. Technol. 2013, 3, 235–243; d) P. D.
Stevens, J. Fan, H. M. R. Gardimalla, M. Yen, Y. Gao, Org. Lett.
2005, 7, 2085–2088; e) G. Lv, W. Mai, R. Jin, L. Gao, Synlett 2008,
9, 1418–1422.
[14] Z. Xu, Y. Hou, S. Sun, J. Am. Chem. Soc. 2007, 129, 8698–8699.
[15] Z. Wu, C. Sun, Y. Chai, M. Zhang, RSC Adv. 2011, 1, 1179–1182.
[16] K. Jiang, H. X. Zhang, Y. Y. Yang, R. Mothes, H. Lang, W. B. Cai,
Chem. Commun. 2011, 47, 11924–11926.
[17] V. Polshettiwar, R. S. Varma, Chem. Eur. J. 2009, 15, 1582–1586.
[18] X. Liu, M. D. Kaminski, Y. Guan, H. Chen, H. Liu, A. J. Rosengart,
J. Magn. Magn. Mater. 2006, 306, 248–253.
[23] F. Lu, J. Ruiz, D. Astruc, Tetrahedron Lett. 2004, 45, 9443–9445.
[24] Y. Jang, S. Kim, S. W. Jun, B. H. Kim, S. Hwans, I. K. Song, B. M.
Kim, T. Hyeon, Chem. Commun. 2011, 47, 3601–3603.
[25] a) S. Kim, E. Kim, M. Kim, Chem. Asian J. 2011, 6, 1921–1925;
b) F. Zhang, N. Liu, P. Zhao, J. Sun, P. Wang, W. Ding, J. Liu, J. Jin,
J. Ma, Appl. Surf. Sci. 2012, 263, 471–475; c) G. Marcelo, A.
MuÇoz-Bonilla, M. Fernꢊndez-Garcia, J. Phys. Chem. C 2012, 116,
24717–24725; d) S. Park, I. S. Lee, J. Park, Org. Biomol. Chem.
2013, 11, 395–399.
[19] H. P. Klug, L. E: Alexander, X-Ray Diffraction Procedures, 2nd ed.,
Wiley, New York, 1974, pp. 687–703.
[20] N. J. S. Costa, P. K. Kiyohara, A. L. Monteiro, Y. Coppel, K. Philip-
pot, L. M. Rossi, J. Catal. 2010, 276, 382–389.
[26] a) M. J. Jacinto, P. K. Kiyohara, S. H. Masunaga, R. F. Jardim, L. M.
Rossi, Appl. Catal. A 2008, 338, 52–57; b) B. Lꢉger, A. Denicourt-
Nowicki, H. Olivier-Bourbigou, A. Roucoux, Inorg. Chem. 2008, 47,
9090–9096.
[21] a) N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457–2483; b) M.
Moreno-MaÇas, R. Pleixats, Acc. Chem. Res. 2003, 36, 638–643;
c) L. Wu, B.-L. Li, Y.-Y. Huang, H.-F. Zhou, Y.-M. He, Q.-H. Fan,
Org. Lett. 2006, 8, 3605–3608; d) N. E. Leadbeater, M. Marco, J.
Org. Chem. 2003, 68, 5660–5667; e) J. Durand, E. Teuma, M.
Received: May 8, 2013
Published online: && &&, 2013
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!