Colloids and Surfaces A: Physicochem. Eng.

Aspects 276 (2006) 116–121

Synthesis of palladium-coated magnetic nanoparticle and

its application in Heck reaction
Zhifei Wang a , Pengfen Xiao a , Bin Shen b , Nongyue He a,∗
a State Key Laboratory of Bioelectronics, Southeast University, 2 Sipailou, Nanjing 210096, PR China
b Department of Chemistry, Southeast University, 2 Sipailou, Nanjing 210096, PR China

Received 11 March 2005; received in revised form 14 October 2005; accepted 25 October 2005
Available online 1 December 2005

Abstract
In this paper, we synthesized the new palladium catalyst based on magnetic nanoparticles by the “bottom-up” approach, and then its catalytic
behavior in the cross coupling of acrylic acid with iodobenzene was investigated. When compared with the traditional palladium catalysts supported
on carbon or the others, such catalyst has a good performance in the first run (i.e. TOF of 3749 h−1 for NaOAc base). However, the activity in
re-use is poor and especially influenced by the base in the reaction. The results of palladium leaching analysis and TEM image indicate that one
of the causes for the above activity drop is the leaching of Pd nanoparticle (or palladium ion) from the surface of magnetic nanoparticle into the
solution. Finally, the reaction of styrene with iodobenzene (or bromobenzene) was also employed to investigate its catalytic behavior.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Palladium catalyst; Magnetic body; Nanoparticle; Heck reaction; Bottom-up

1. Introduction bodies as well as the transport of them within porous catalyst

bodies can be rate limiting because the transport rate to the
In recent years, there are growing interests on the catalytic surface of the catalyst bodies is proportional to 1/D (D is the
properties of transition metal nanoparticles because of their large diameter of catalyst body) [8]. To raise the rate of the reaction, it
surface area and a great ratio of atoms remaining at the surface is therefore attractive to utilize small catalyst bodies. Recently,
[1–4]. Although many researches have been done, there is still there are some significant advances in this field, such as the appli-
the paramount challenge for the wide application of transition cation of dendritic catalyst [3,4], but the main problem related
metal nanoparticle as catalysts in the industry, i.e. how to sepa- to this approach is the separation and leaching of heterogeneous
rate and recycle them completely from the products. At present, dendritic catalyst by the nanofiltration membrane, which does
the main idea presented for recyclable systems may be built not meet the large-scale application in the industry.
in liquid–liquid and solid–liquid modes [1]. For liquid–liquid Thus, it is necessary to develop new catalysts with the mag-
system, due to the high interfacial tension between water and netic cores and their shell consisting of the catalytic species
low-polar organic liquids, the area of the interface is small even for the ease separation by the external magnetic field. So for,
with vigorous stirring. Therefore, separation and catalytic effi- the reports on the magnetic nanoparticles as catalyst supports
ciency come to a contradiction. Another approach to separate are mainly focused on nickel–iron alloy nanoparticle with its
and recycle metal nanoparticles is to immobilize them on solid surface coated by carbon layer [9,10], and other paramagnetic
supports such as organic polymeric (resin) [5], inorganic micro- nanoparticles are less studied. In comparison with other super-
sphere (Pd/C, Pd/SiO2 , Pd/Al2 O3 , etc.) [1,6,7], making it really paramagnetic nanoparticle, Fe3 O4 nanoparticle could be easily
easy and simple to separate the catalyst from the reaction prod- produced by the co-deposition of ferrite with the basic solu-
ucts mixture. However, with solid catalysts suspended within tion. In addition, many investigations have been done on the
liquids the transport of reactants within the liquid to the catalyst surface modification due to its application in biological technol-
ogy [11–13], which laid good foundation on further research as
the catalyst support.
∗ Corresponding author. Tel.: +86 25 83792245; fax: +86 25 83619983. Herein, we synthesized the new palladium catalyst based on
E-mail address: nyhe1958@163.com (N. He). magnetic body of silica coated Fe3 O4 by “bottom-up” approach,

0927-7757/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.colsurfa.2005.10.028
 Z. Wang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 276 (2006) 116–121 117

and its catalytic behavior in Heck reaction was also investigated. 42 ml of H2 O, 28 ml of ethanol and 0.067 g of PVP was refluxed
Considering the fact that Fe3 O4 is fairly reactive, especially with for 3 h.
acid encountered in liquid-phase catalytic reaction frequently, After isolating APTS coated SiO2 /Fe3 O4 nanoparticle from
the silica was employed to shield the inner magnetic core from the reactant, the as-synthesized colloid solution of Pd nanopar-
the external environment. In comparison with the other shells ticles was added and the resulting solution was stirred for 3 h.
such as carbon layer, the silica layer around magnetite pro- Finally, the catalyst was separated from the reaction medium
vides the surface silanol groups, which can react with silane under the magnetic field, and then washed with DI water for
coupling agents to produce stable dispersions in non-aqueous three times.
solvent without the risk of aggregation. Also, its isoelectric point
is lowed to pH ∼3 [13]. So, the utilization of the silica shell can 2.5. General procedure for the catalytic tests
improve the stability of the catalyst in both aqueous solvent and
non-aqueous solvent. The cross coupling of acrylic acid with iodobenzene was per-
formed under different reaction conditions. Typically, 1 equiv
2. Experimental (4 mmol) of iodobenzene, 2 equiv (8 mmol) of acrylic acid,
15 mg of catalyst, 20 ml of a CH3 CN solution and 2.5 equiv
2.1. Material (10 mmol) of NaOAc were added to 60 ml of water, and the
mixture was heated at reflux for 12 h. After cooling to room tem-
All chemicals used were of analytical grade from Shanghai perature, the catalyst was separated from the reaction medium
Chemical Reagent Corporation except iodobenzene, which is by the magnetic field. Then, H2 O and 5% HCl were added to the
chemical grade. Water used in the experiments was deionized residual mixture, from which white solid was formed. Finally,
(DI), doubly distilled and deoxygenated prior to use. the solid was filtered, washed several times with fresh water,
and then recrystallized by water/ethanol (3:1) solution. For the
2.2. Preparation of Fe3 O4 nanoparticle reaction of iodobenzene and styrene, the residue was extracted
with ether and chromatographed over silica gel after CH3 CN
Fe3 O4 nanoparticles were produced by the chemical co- was removed by rotary evaporation.
precipitation. Typically, a solution of mixture of 0.85 ml of
12 mol/l HCl, 25 ml of deoxygenated DI water, 5.2 g of FeCl3 2.6. Characterization
and 2.0 g of FeCl2 was prepared under N2 protection. Then the
resulting solution was added drop-wise into 250 ml of 1.5 mol/l The particle size and morphology of the samples were deter-
NaOH solution under vigorous stirring. The obtained precipitate mined by transmission electronic microscopy (TEM) with JEM-
was diluted to 7 g/l by tetramethylammonium hydroxide (TMA) 200CX operating at 200 kV. Magnetization measurements of
and DI water after the washing by DI water for three times. both Fe3 O4 nanoparticle and SiO2 coated Fe3 O4 nanoparticle
were performed at room temperature using vibration sample
2.3. Preparation of silica coated Fe3 O4 nanoparticle magnetometer (VSM).
(SiO2 /Fe3 O4 )
3. Results and discussion
A 31 ml of the silicate solution (the silicate solution was
obtained by 0.58 wt% sodium silicate stock solution pass an acid Catalyst Pd/(SiO2 /Fe3 O4 ) is prepared by ligand-mediated
exchange resin column) was mixed with 169 ml of ferrofluid immobilization of Pd0 nanoparticle on functionalized oxide sur-
diluted with DI water to the concentration of 1.4 g/l. After that, faces as shown in Scheme 1. It can be found that several Fe3 O4
pH was adjusted to 10 by slow titration with 0.5 mol/l HCl. Then,
the obtained solution was stirred for 2 h and allowed to stand for
2 days. Further silica growth on the obtained particles was per-
formed according to Stöber method. To 500 ml of the sol in 1:4
water/ethanol containing the obtained particle was added 0.3 ml
of TEOS and 0.6 ml of ammonia. The solution was allowed to
stand for 12 h under mild magnetic stirring. The surface modi-
fication of SiO2 coated Fe3 O4 nanoparticle with 3-aminopropyl
triethoxysilane (APTS) was performed by the addition of 0.4 ml
of APTS solution and rapid stirring for 8 h.

2.4. Incorporation of Pd nanoparticle into SiO2 coated

Fe3 O4 nanoparticle
Scheme 1. Procedure of the preparation of catalyst Pd/(SiO2 /Fe3 O4 ). (A) Fe3 O4
Pd nanoparticles were prepared according to the literature nanoparticle; (B) SiO2 /Fe3 O4 ; (C) APTS coated SiO2 /Fe3 O4 nanoparticle; (D)
[14]. Typically, a mixture of 30 ml of a 2 mM H2 PdCl4 solution, Pd/(SiO2 /Fe3 O4 ).
118 Z. Wang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 276 (2006) 116–121

Fig. 1. TEM image of the nanoparticles. (A) Fe3 O4 nanoparticle; (B) SiO2 coated Fe3 O4 nanoparticle prepared at 0.08 g/l Fe4 O3 nanoparticle, 1:4 H2 O/ethanol,
50 ␮l of TEOS per 500 ml of reaction mixture, 0.2 ml of NH3 OH; (C) SiO2 coated Fe3 O4 nanoparticle prepared at 0.17 g/l Fe4 O3 nanoparticle, 1:4 H2 O/ethanol,
0.5 ml of TEOS per 500 ml of reaction mixture, 1 ml of NH3 OH; (D) SiO2 /Fe3 O4 coated with Pd nanoparticle.

nanoparticles were firstly embedded in silica coating by the From Fig. 1B, it can be found that its average diameter is about
Stöber method, and then the surfaces of the obtained nanopar- 100 nm.
ticles were modified with APTS through the reaction between After the functionalization of SiO2 coated Fe3 O4 nanopar-
–OH on the surface and APTS. Finally, the as-synthesized Pd ticles by APTS, the as-synthesized Pd colloid, which size is
colloid was bound to the catalyst carrier through the pendent about 3 nm according to the literature [14], was bound to func-
amine group. So far, there have been several reports on the tionalized oxide surfaces by the strong coordination of –NH2
preparation of SiO2 coated Fe3 O4 nanoparticle [11–13]. Herein, group with the palladium surface. In order to assure that the
the thin silica layer was firstly deposited on the magnetic par- surface of SiO2 coated Fe3 O4 nanoparticles was covered by
ticle by mixing the silicate solution with aqueous ferrofluid, palladium nanoparticles, the concentration ratio of Pd colloid
and the further silica growth on the above-obtained nanoparticle to SiO2 coated Fe3 O4 nanoparticles was kept at 104 :1. As can
was performed by the Stöber method [15]. The representative be seen from the TEM image (Fig. 1D), palladium particles are
TEM image of Fe3 O4 nanoparticle prepared by the chemical nearly uniformly distributed on the surface of the SiO2 coated
co-precipitation is shown in Fig. 1A, from which it can be seen Fe3 O4 nanoparticle and its coverage is high, which means that
that most of the particles are quasi-spherical with an average more surface palladium atoms are available for catalysis as com-
diameter of 8 nm. By adjusting experimental conditions such as pared to using the bulk palladium as catalyst, resulting in high
the amount of TEOS and concentration of Fe3 O4 nanoparticle, activity per gram of supported catalysis. The element analysis
we were able to vary the shell thickness and control the number of catalyst by the flame atomic absorption spectroscopy (AAS)
of Fe3 O4 nanoparticle in core. In this work, two kinds of typical indicates that the content of palladium is about 0.8 wt%. Fig. 2
SiO2 coated Fe3 O4 nanoparticles were prepared. Fig. 1B and shows the magnetization curves of the samples. It can be seen
C give the TEM images of those nanoparticles containing one that after the reaction, the saturation of Fe3 O4 particle decreases
and several embedded magnetite particles, respectively. Con- from 31.4 to 9.8 emu/g for the SiO2 coated Fe3 O4 nanoparticle.
sidering the requirement for SiO2 coated Fe3 O4 nanoparticle There is no hysteresis, and both remanence and coercivity are
as catalyst support, it must have relatively high saturation and zero, suggesting that such nanospheres are superparamagnetic.
appropriate surface area. SiO2 coated Fe3 O4 nanoparticles with In order to investigate its catalytic behavior, Heck reaction of
some embedded iron oxide were chosen as the catalyst support. the cross coupling of acrylic acid with iodobenzene was studied.
 Z. Wang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 276 (2006) 116–121 119

Fig. 2. Magnetization curves obtained by VSM at room temperature for Fe3 O4

Fig. 4. Product yield as a function of the times of cycle under the difference
nanoparticle (a) and SiO2 coated Fe3 O4 nanoparticle (b).
reaction condition.

Considering the solubility requirement of the catalyst, the reac- indicating that the active species attached to the surface of the
tant and product, the mixture solvent (CH3 CN/H2 O, 1:3, v/v) catalyst is lost. If we consider the possible mechanism occurred
was used for the reaction. The base was used to neutralize the in Heck reaction, this phenomenon can be explained. During the
acid (HX) ensuing from the formal exchange of a hydrogen atom reaction, the catalysis often performs by the way of the oxidative
with an aryl group. At the end of each reaction, the catalyst was addition of the halide to the surface atoms of Pd particles, which
separate from the reaction medium under the external magnetic leads to the release of palladium in the +2 oxidation state [16].
field of 1.4 T. During the reaction, no obvious agglomeration of So, the good coordinating agents towards palladium (+2) should
the catalyst was observed. enhance the extent of metal leaching.
Fig. 3 gives the effect of the reaction time on the product Palladium leaching after the first cycle was measured to fur-
yield. It can be seen that after the reaction for 10 h, the product ther confirm the above assumption (Table 1). It can be seen that
yield changes little with longer time and it reaches the maximum after the first run palladium leaching in the reaction using NEt3
67% when the reaction time is 12 h. So, 12 h has been chosen is the maximum 151 ppm as compared with other bases, which
as the reaction time for each of the reaction investigated in this is consistent with the coordination ability between the base and
paper. The change of product yield as a function of times of palladium ion. If palladium leaching is only by way of the pal-
cycle under the different reaction conditions is shown in Fig. 4. ladium ion, another question would arise, that is, the catalyst by
It can be seen that with the increase of the times of cycle, the using NEt3 as the base should be more active than that by using
catalyst activity decreases greatly. For example, for the reaction other base. So, we deduce that pre-absorbed Pd nanoparticle is
using NaOAc as the base, the product yield decreases from 67 lost from the surface of the magnetic support during the reaction
to 37% after the sixth run. It can also be seen that the extent of besides its leaching by the ion. Fig. 5 presents the TEM image
the drop in catalyst activity varies with the different base. For of the used catalyst by using NEt3 as the base after six cycles. It
NEt3 base, the catalyst activity is nearly zero after six cycles, can be seen that the catalyst surface looks smooth as compared
with the rough surface before the reaction, which indicates that
most of nanoparticles are lost during the reaction. Several big-
ger palladium nanoparticles directed by the arrow in the image
are also found on the surface of the catalyst, indicating that a

Table 1
Effects of bases on the reaction of idodebenzne and acrylic acid and Pd leaching
of catalysts
Entry Base Conversion (%) TOF (h−1 ) Pd leached into
solvents (ppm)

1 Na2 CO3 58 3247 21

2 NEt3 52 2911 151
3 Na3 PO4 63 3526 35
4 NaOAc 67 3749 23
5a NEt3 78 1337 Blank
a Reaction condition: catalyst: Pd-Y; catalyst loading: 30 mg; reaction temper-
ature: 353 K; reactant: 10 mmol iodobenzene; 15 mmol butylacrylate, 12 mmol
Fig. 3. Effect of the reaction time on product yield. NEt3 , 10 ml solvent dimethylformamide (DMF) or tolunen.
120 Z. Wang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 276 (2006) 116–121

Table 2
Product yields for other Heck reactions in CH3 CN/H2 O (temperature, reflux; time 12 h; base, 2.5 equiv of NaOAc)
Entry ArX Alkene Products Conversion (%)

1 71

2 58

3 63

4 51

few Pd nanoparticles have grown into bigger nanoparticle by 4. Conclusion

the redeposition of Pd ion leached from the support during the
reaction onto the surface of the support after completion of the In summary, we successfully synthesized the silica coated
reaction. Fe3 O4 nanoparticle as the catalyst support in this work, and
Table 1 also gives the turnover frequency (TOF) values of then palladium catalyst was obtained by the coordination of
the catalyst by using different bases. TOF were calculated as 3 nm palladium nanoparticle with –NH2 group on the support
(moles product) × (moles of catalyst)−1 × (reaction time (h))−1 . surface. TEM image shows that palladium particles are nearly
For the sake of simplicity in the calculations, we consider uniformly dispersed on the SiO2 coated Fe3 O4 particle with high
that the size of the catalyst support is 100 nm and the cov- coverage. Comparing with the traditionally supported catalyst,
erage of palladium nanoparticle on the surface of catalyst is the activity measurement shows this kind of catalyst has good
uniform. It can be seen that the catalyst shows good perfor- activity for Heck reaction in the first run. However, the activity
mance with TOF of 3749 h−1 (for NaOAc base) at the first run in re-use is poor and especially influenced by the base in the
as compared with other palladium catalysts supported on sil- reaction. The results of palladium leaching analysis and TEM
ica, alumina and carbon [1,6,7]. In order to further examine image indicate that the drop of activity in re-use is due to the
the catalyst activity on other Heck reaction, bromo-substrate leaching of palladium ion (or palladium nanoparticle) into the
rather than iodo-substrate and styrene instead of acrylic acid solution. Considering the fact that Pd leaching and redeposition
were used (Table 2). It can be seen that the catalyst has good may depend on the state of Pd dispersion (particle size and the
performance for the cross coupling of styrene with iodoben- interactions of metal/support) and the surface properties of sup-
zene (NaOAc is used as the base). In addition, as is usually the port, we suggest that the bases with the strong coordination with
case, iodoaromatic is found to be more active substrates than Pd should be avoided in future research. In a word, as the fun-
bromoaromtic. damental research, this paper provides the attempt to solve the
recycle problem of metal nanoparticle as the catalyst in reaction
by synthesizing Pd coated magnetic nanoparticle catalyst and
giving its preliminary activity results in Heck reaction.

Acknowledgements

This research was funded by the National Natural Science

Foundation of China 60571032, 20505020 and the Doctoral
Program of Higher Education 20050286014. Thanks to Z.C.
Wang (State Key Lab. Palaeobiology and Stratigraphy, Nanjing
Institute of Geology and Palaeontology, CAS) for help using
transmission electron microscope.

References

[1] A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 102 (2002) 3757–3778.

[2] R. Van Heerbeek, P.C.J. Kamer, P.W.N.M. van Leeuwen, J.N.H. Reek,
Fig. 5. TEM image of the used catalyst after the sixth run by using Et3 N as the Chem. Rev. 102 (2002) 3717–3756.
base. [3] L.K. Yeung, R.M. Crooks, Nano Lett. 1 (2001) 14–17.
 Z. Wang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 276 (2006) 116–121 121

[4] K.R. Gopidas, J.K. Whitesell, M.A. Fox, Nano Lett. 3 (2003) [10] S.C. Tsang, V. Caps, I. Paraskevas, D. Chadwick, D. Thompsett, Angew.
1757–1760. Chem. Int. Ed. 43 (2004) 5645–5649.
[5] M. Králik, A. Biffs, J. Mol. Catal. A: Chem. 177 (2001) 133–138. [11] Y. Lu, Y. Yin, B.T. Mayers, Y. Xia, Nano Lett. 2 (2002) 183–186.
[6] K. Köhler, R.G. Heidenreich, J.G.E. Krauter, J. Pietsch, Chem. -A Eur. [12] P. Tartaj, C.J. Serna, J. Am. Chem. Soc. 125 (2003) 15754–15755.
J. 8 (2002) 622–631. [13] A.P. Philipse, M.P.B. van Bruggen, C. Pathmamanoharan, Langmuir 10
[7] Y. Ji, S. Jain, R.J. Davis, J. Phys. Chem. B 109 (2005) 17232–17238. (1994) 92–99.
[8] W. Teunissen, A.A. Bol, J.W. Geus, Catal. Today 48 (1999) 329–336. [14] T. Teranishi, M. Miyake, Chem. Mater. 10 (1998) 594.
[9] W. Teunissen, F.M.F. de Groot, J. Geus, O. Stephan, M. Tence, C. [15] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1958) 62.
Colliex, J. Catal. 204 (2001) 169–174. [16] A. Biffis, M. Zecca, M. Basato, Eur. J. Inorg. Chem. (2001) 1131–1133.