Synthesis of a Pd on Ni-B nanoparticle catalyst by the replacement reaction method for hydrodechlorination

Article abstract of DOI:10.1016/j.jcat.2008.03.026

A Pd on Ni-B (Pd/Ni-B) bimetallic nanoparticle catalyst was prepared through a replacement reaction of amorphous Ni-B nanoparticles with Pd²⁺ ions. In the hydrodechlorination of chlorobenzene, the Pd/Ni-B bimetallic catalyst exhibited better activity than a Pd catalyst supported on polyvinylpolypyrrolidone, a mixture of Pd and Ni-B nanoparticles, and a Pd on Ni catalyst obtained by the replacement reaction of nanocrystalline Ni with Pd²⁺ ions. The properties of the Pd/Ni-B catalyst were studied in detail by XRD, TEM, XPS, and H₂ chemisorptions. The characterizations demonstrated that a Pd-Ni-B surface alloy occurred on the Pd/Ni-B catalyst with heating treatment. The resulting surface alloy promoted resistance to chlorine poisoning, adsorption of the reactants (chlorobenzene and hydrogen), and desorption of products, explaining the better catalytic activity and stability of the Pd/Ni-B bimetallic catalyst.

Full text of DOI:10.1016/j.jcat.2008.03.026

Journal of Catalysis 256 (2008) 323–330
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Synthesis of a Pd on Ni–B nanoparticle catalyst by the replacement reaction
method for hydrodechlorination
∗
Zhijie Wu, Minghui Zhang , Zongfang Zhao, Wei Li, Keyi Tao
Institute of New Catalytic Materials Science, College of Chemistry and Engineering Research Center of High-Energy Storage and Conversion (Ministry of Education), Nankai University,
Tianjin 300071, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
A Pd on Ni–B (Pd/Ni–B) bimetallic nanoparticle catalyst was prepared through a replacement reaction of
amorphous Ni–B nanoparticles with Pd2+ ions. In the hydrodechlorination of chlorobenzene, the Pd/Ni–B
bimetallic catalyst exhibited better activity than a Pd catalyst supported on polyvinylpolypyrrolidone, a
mixture of Pd and Ni–B nanoparticles, and a Pd on Ni catalyst obtained by the replacement reaction of
nanocrystalline Ni with Pd2+ ions. The properties of the Pd/Ni–B catalyst were studied in detail by XRD,
TEM, XPS, and H2 chemisorptions. The characterizations demonstrated that a Pd–Ni–B surface alloy oc-
curred on the Pd/Ni–B catalyst with heating treatment. The resulting surface alloy promoted resistance
to chlorine poisoning, adsorption of the reactants (chlorobenzene and hydrogen), and desorption of prod-
ucts, explaining the better catalytic activity and stability of the Pd/Ni–B bimetallic catalyst.
© 2008 Elsevier Inc. All rights reserved.
Received 19 November 2007
Revised 21 March 2008
Accepted 30 March 2008
Available online 28 April 2008
Keywords:
Bimetallic catalyst
Palladium
Ni–B
Hydrodechlorination
Chlorobenzene
1. Introduction
reaction also is used to prepare bimetallic nanoparticles [13–17],
and some bimetallic nanoparticles have shown improved catalytic
The dehalogenation of the carcinogenic and mutagenic aromatic
halides has attracted much attention [1–3]. Aromatic chlorides are
much less reactive than aromatic bromides and iodides [4], and
few effective methods are available for the dechlorination (DC) of
aromatic chlorides [5]. The most attractive and efficient DC method
is catalytic hydrodechlorination (HDC), in which a variety of cata-
lysts (e.g., Rh, Pt, Pd, Ni, Cu) and gaseous hydrogen or alcohols (as
hydrogen donor) are used [6]. HDC is emerging as a nondestructive
alternate technology in which chlorinated waste can be converted
into products of commercial value [1,7]. Among the HDC catalysts,
Pd-based catalysts have been widely investigated because of their
superior catalytic properties [8,9].
Bimetallic nanoparticle catalysts have received increasing atten-
tion over the last 50 years. The addition of the second metal plays
a key role in controlling the activity, selectivity, and stability of
catalysts in certain reactions [10]. Recently, a Pd on Au bimetallic
nanoparticle catalyst for the HDC of trichloroethene was prepared
by adding controlled amounts of Pd salt and a reducing agent to an
Au sol [11,12]. The resulting catalyst demonstrated excellent cat-
alytic properties by distributing Pd atoms homogeneously over the
Au surface and producing a synergistic effect of Pd with Au. The
Pd on Au bimetallic catalyst structure provides a new synthesis
approach in improving the catalytic properties of monometallic Pd
materials [11]. Other than the chemical reduction, the replacement
properties [18–21]. The replacement reaction is defined as a re-
action of a metal ion with a more reactive metal solid. Amor-
phous Ni–B nanoparticle is characteristic of a unique short-range
ordering structure but a long-range disordering structure, along
with a high concentration of coordinatively unsaturated sites. It
demonstrates better catalytic activity and selectivity than its cor-
responding crystalline counterparts [22,23]. In the present work,
a Pd on Ni–B (Pd/Ni–B) bimetallic nanoparticle catalyst was syn-
thesized by reacting solutions of Na2[PdCl4] with amorphous Ni–B
nanoparticle instead of nanocrystalline Ni. Its catalytic performance
in the HDC of chlorobenzene was examined and compared with
that of Pd supported on polyvinylpolypyrrolidone (PVP) and Pd on
Ni (Pd/Ni) bimetallic nanoparticles from the replacement of crys-
talline Ni nanoparticles by Pd. The correlation with the structural
characteristics was investigated.
2. Experiments and methods
2.1. Catalyst preparations
Amorphous Ni–B nanoparticles were prepared by a modified
chemical reaction. In a typical synthesis, 11.1 ml of ethylenedi-
amine was dissolved in 1800.0 ml of deionized water, and then
21.8 g of NiCl2·6H2O and 9.0 g of KBH4 were added to the
ethylenediamine aqueous solution under stirring. The pH of the
solution was controlled at 12.5–13.0 by the addition of NaOH. To
prepare amorphous Ni–B nanoparticles, the solution was heated to
Corresponding author. Fax: +86 22 2350 7730.
E-mail address: zhangmh@nankai.edu.cn (M. Zhang).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.03.026

324
Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
◦
70 1.0 C. The reaction lasted until no bubbles were observed,
and the resulting black solid was washed free of inorganic ions
and organic substances with deionized water and ethanol. Crys-
talline Ni nanoparticles were prepared by the reduction of Ni2+ by
hydrazine as described previously to [24]. First, 1.0 g of NiCl2·6H2O
(0.0042 mol) was dissolved in 7.5 ml of ethanol in a beaker; then
4.0 g of 50% N2H4·H2O (0.04 mol) was mixed with 2.5 g of NaOH
by stirring, and the resulting slurry was added to the beaker. The
reaction was initiated by adding a small amount of inducing agent,
KBH4, which reduced Ni2+ to metal nanoparticles that catalyzed
◦
the reaction between Ni2+ and hydrazine at 25 C. The black pre-
cipitate was washed with ammonia, deionized water, and finally
acetone.
A PVP-supported Pd catalyst was synthesized as described pre-
viously [25,26]. To a solution of 0.055 g of PdCl2 in 1.0 ml of
5 mol/L NaCl, 50.0 ml of absolute ethanol and 0.7 g of PVP
were added. The resulting red-brown solution of PVP-PdCl2 (6.3 ×
−3
10
μmol/L Pd, N/Pd = 20 [molar ratio]) was stirred for 48 h at
room temperature and then reduced by a 1 mol/L KBH4 solution
under a 40-ml/min N2 flow. The product was washed with water
and ethanol and then collected by centrifugation. For the prepara-
tion of Pd/Ni–B or Pd/Ni bimetallic catalysts, 0.055 g of PdCl2 was
dissolved in 50.0 ml of 0.1 mol/L NaCl solution under the N2 at-
mosphere, after which the Ni–B or crystalline Ni was added to the
Na2[PdCl4] solution and stirred for 4 h. The Pd loading was se-
lected at a weight ratio of 1/99 (Pd/Ni).
Fig. 1. XRD patterns of Pd/Ni–B catalysts treated in a 40-ml/min N2 flow at different
temperatures for 1 h.
The supported catalysts were prepared as follows. The Ni–
B/TiO2 was prepared as described previously [27,28]. Then 0.02 g
of silver nitrate, 0.15 g of ammonium hydroxide, 0.005 g of sodium
hydroxide, and 0.002 g of formaldehyde were dissolved in 425.0 ml
of distilled water, after which 6.39 g of titania (anatase, 27 m2/g)
HDC/DC capability of ethanol were <0.1%, suggesting that ethanol
served as a solvent and did not promote HDC activity of the cata-
lyst.
2.3. Characterization
◦
was added to this solution. After stirring at 40 C for 4 h, the
resulting Ag/TiO2 was washed with distilled water and dried at
80 C for 4 h. The electroless Ni–B plating solution comprised nick-
Powder X-ray diffraction (XRD) patterns of the catalysts were
obtained on a Rigaku D/max 2500 X-ray diffractometer (Cu Kα,
λ = 1.54178 Å). The compositions of the samples were measured
by inductively coupled plasma atomic emission spectrometry (ICP-
AES) on an IRIS Intrepid spectrometer. Transmission electron mi-
croscope (TEM) images were obtained with an FEI Tecnai G2 high-
resolution transmission electron microscope. X-ray photoelectron
spectroscopy (XPS) analysis of the catalysts was carried out us-
ing a Kratos Axis Ultra DLD spectrometer using a monochromatic
Al Kα source, hybrid magnetic/electrostatic optics, and a multi-
channel plate and delay line detector (DLD). The surface of the
◦
elous sulfate hexahydrate (12 g/L), ethylenediamine (12 g/L), and
potassium borohydride (5.5 g/L). The pH value of the solution was
adjusted to 13.5 by NaOH. To prepare the supported Ni–B/TiO2
catalyst, 2.0 g of Ag/TiO2 was added to 150.0 ml of as-prepared
◦
Ni–B plating solution under stirring at 50 C. The reaction lasted
until no significant bubbles were observed. The resulting samples
were washed thoroughly with distilled water until a pH of 7 was
achieved, then washed with absolute alcohol. For the preparation
of supported Pd/Ni–B/TiO2 catalyst, 0.055 g of PdCl2 was dissolved
in 50.0 ml of 0.1 mol/L NaCl solution under the N2 atmosphere.
Then the Ni–B/TiO2 was added to the Na2[PdCl4] solution, stirred
for 4 h, and washed as described above. The ICP results revealed
the following weight ratio of the components in the Pd/Ni–B/TiO2:
0.1% Pd: 1.5% B: 10.4% Ni: 87.5% TiO2.
+
samples was etched with Ar ions for 15 min to remove the
oxidated layer. Temperature-programmed desorption (TPD) of hy-
drogen, benzene, or chlorobenzene was performed as follows: The
surface of the sample (20 mg) was purged by a flow of ultra-pure
◦
argon (99.999%) for 2 h at 200 C. (The temperature was kept rel-
atively low to avoid crystallization of the amorphous catalysts.)
Once the catalyst was cooled to 25 C in argon, H2, chlorobenzene,
2.2. Reaction
◦
or benzene was introduced to the catalyst for adsorption. After
12 h, desorption of the adsorbate was performed under an argon
flow of 30 ml/min. All gases were of ultra-high purity and were
further purified by a Chrompack clean-oxygen filter and molecular
The HDC of chlorobenzene was carried out in a 100-ml auto-
clave in which the catalysts containing 0.1 mmol Pd, 50.0 mmol of
chlorobenzene, 0.5 g NaOH, and 50.0 ml of ethanol were mixed.
Before the HDC reaction, the catalysts were treated with a 40-
◦
◦
sieve. The temperature was raised from 25 to 600 C at a rate of
ml/min N2 flow at 25, 200, 300, and 400 C for 2 h. The HDC
◦
◦
10 C/min.
reaction was carried out at 90 C. The hydrogen pressure in all ex-
periments was 1.0 MPa. The products were analyzed with a gas
chromatograph equipped with a flame ionization detector. The re-
sults demonstrated formation of only benzene and HCl; thus, ei-
ther benzene yield or chlorobenzene conversion was used to eval-
uate the HDC activity.
3. Results
3.1. Characterization
HDC was performed without the use of catalyst (blank exper-
iment), and ethanol’s HDC/DC capability was evaluated by carry-
ing out the HDC reaction without hydrogen or catalyst. Both the
conversion of chlorobenzene without the use of catalyst and the
The atom composition of amorphous Ni–B nanoparticle was
Ni67.4B32.6, and the Pd loading on Pd/Ni–B catalyst was 0.8 wt%.
It was well known that many metal oxides or salts can be spon-
taneously dispersed on the surface of oxide supports to form

Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
325
◦
◦
◦
◦
Fig. 2. TEM images of (a) Ni–B, (b) Pd/Ni–B treated at 25 C, (c) Pd/Ni–B treated at 200 C, (d) Pd/Ni–B treated at 300 C, and (e) Pd/Ni–B treated at 350 C in a flow of
40 ml/min N2.
a monolayer due to the formation of interfacial chemical bonds
[18–20]. Wang et al. found that a metal can disperse on the sur-
face of another metal to form a highly dispersed bimetallic system
due to the formation of bonds between the two metals. They noted
that a highly dispersed Pd–Ni system occurred when <2.0 wt% Pd
was loaded on a Pd/Ni catalyst [19]. In the present work, XRD
was used to characterize the structure of the Pd/Ni–B bimetal-
lic nanoparticles. No diffraction peak corresponding to crystalline
Pd was found, indicating a high dispersion of Pd. Figs. 1a and 1b
◦
show a broad, featureless peak at around 2θ = 45 indicating the
amorphous structure of Pd/Ni–B bimetallic nanoparticles treated at
◦
◦
25 and 200 C. Increasing the treatment temperature to 300 C led
to the appearance of peaks corresponding to crystalline Ni; on a
◦
further increase to 350 C, the Ni3B phase appeared. Peaks due to
crystalline Pd or to Pd compounds still were not seen, however.
Fig. 2 presents a series of TEM micrographs of samples heated
at different temperatures. Fig. 2a shows spherical, flower-shaped
Ni–B particles of ∼30 nm. The SAED pattern (insert in Fig. 2a)
indicates the amorphous nature of these particles. After the re-
placement of Ni by Pd, the shape and size of the particles changed
(Fig. 2b); the SAED pattern shows the amorphous structure of the
Pd/Ni–B bimetallic particles as well, in agreement with the results
shown in Fig. 1a. As shown in Fig. 2c, the particle shape changed
from flower-like to a solid sphere on treating the Pd/Ni–B at
◦
200 C, but the sample retained an amorphous structure (Fig. 1b).
◦
Compared with the Pd/Ni–B treated at 25 C, the Pd/Ni–B treated
at 200 C had a somewhat smaller particle size and exhibited sev-
◦
Fig. 3. XPS spectra of Ni 2p3/2 for Pd/Ni and Pd/Ni–B samples treated at different
temperatures in a 40-ml/min N2 flow.
eral bright dots in its corresponding SAED pattern [29], suggesting
◦
the formation of small crystalline metal clusters at 200 C. When
◦
◦
the treatment temperature was increased to 300 C and then to
B treated at 25 and 350 C) in Supplementary material. These find-
◦
350 C, more bright dots were detected in the SAED patterns (in-
ings indicate that some polycrystalline compounds were formed at
◦
serts in Figs. 2d and 2e), indicating that the sintering of Ni crys-
talline occurred. Sintering also was observed in the nano Ni and
Ni–B with heating treatment, as shown in Figs. S1 (TEM images of
treatment temperatures above 300 C.
Figs. 3–5 show the XPS spectra of the catalysts. The XPS in-
tensity ratios of the metal and the support provide information
on the degree of dispersion of the metal on the support [30,31].
◦
nano Ni treated at 25, 250, and 350 C) and S2 (TEM images of Ni–

326
Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
trations resulted from two mechanisms: (i) The Pd atoms diffused
into the subsurface or the bulk of Ni–B to form an alloy with in-
creasing treatment temperature, and (ii) the Pd formed the larger
Pd particles aggregated on top of the Ni–B. Wang et al. found no
peaks due to crystalline Pd or Pd compounds in the XRD patterns
of Pd/Ni bimetallic particles prepared via a replacement reaction
◦
and treated at 120 C [19]. The same result also was found for the
Pd/Ni bimetallic catalyst in the present study (Fig. S3 in Supple-
mentary material), suggesting that Pd atoms were highly dispersed
on the Ni surface, rather than formed into larger Pd particles [19,
20]. As shown in Fig. 1, no crystalline Pd atoms or Pd compounds
were found in the Pd/Ni–B catalyst treated at different tempera-
◦
tures from 25 to 350 C, indicating that the Pd atoms in Pd/Ni–B
sample should be dispersed on heating treatment. On the other
hand, Sra and Schaak reported the formation of AuCu and AuCu3
bimetallic nanoparticle aggregates by a mixture of crystalline Au
and amorphous Cu nanoparticles on aging at room temperature
[15]. With increasing treatment temperature, Cu diffused into the
Au bulk, forming the disordered solid solution CuxAu1−x, and an
◦
atomically ordered AuCu was formed by heat treatment at 200 C;
thus, it was concluded that the Pd diffused into the Ni–B particles
somewhat, and the Pd atoms or compounds were dispersed over
the Ni–B particles.
At 850–865 eV, the Ni 2p3/2 spectra of the Pd/Ni and Pd/Ni–
◦
B samples treated at 25 C (Fig. 3) were characterized by a main
Fig. 4. XPS spectra of Pd 3d for Pd/Ni and Pd/Ni–B treated at different temperatures
in a 40-ml/min N2 flow.
peak with a satellite at higher binding energy (BE). The main
peak corresponded to the final state 2p53d9L
(L, ligand hole);
−1
the satellite peak, to a 2p53d8 final state [30,31]. The BE of Ni
in Pd/Ni of 854.5 eV was close to that of the Ni species in NiPd
bulk bimetallic nanocluster [32]. Its satellite peak due to 860.0 eV
(curve-fitted value) also was observed in the XPS spectra shown
in Fig. 3. Unlike in the Pd/Ni bimetallic particles, the main peak
of Ni in Pd/Ni–B was decreased to 853.2 eV, a value similar to
that of amorphous Ni–B particles [33], and its satellite peak also
decreased, to 856.5 eV. As shown in Fig. 3, the main and satel-
lite Ni 2p3/2 peaks for Pd/Ni appeared at higher BEs than those
for Pd/Ni–B (1.3 and 3.5 eV, respectively). However, as shown in
Fig. 4, the Pd 3d spectra of Pd/Ni and Pd/Ni–B exhibited similar
BEs as bulk Pd (3d5/2, 335.8 eV and 3d3/2, 341.1 eV, [32]). These
results demonstrate that the difference between Pd/Ni and Pd/Ni–
◦
B treated at 25 C in Fig. 3 can be attributed to the alloying of Ni
and B, rather than to an interaction between Pd and Ni [33]. An in-
creased peak was observed with an increase in treatment tempera-
◦
ture from 25 to 200 C (Fig. 3), demonstrating the different valence
state of nickel due to the alloying of Ni and Pd on heat treatment.
The BE of the main peak increased from 854.5 to 854.7 eV for
Pd/Ni and from 853.2 to 855.7 eV for Pd/Ni–B (curve-fitted values).
This increase in BE can be attributed to the presence of Pd in the
nanoparticles. For NiPd alloys, an increased BE of the main peak
for Ni was observed, related to the reduced nickel d hole density
due to a charge transfer from nickel to adjacent electronegative
palladium [34,35]. The Pd 3d spectra of Pd/Ni and Pd/Ni–B exhib-
ited shoulder peaks (Fig. 4). The Ni 2p3/2 spectra for the Pd/Ni
and Pd/Ni–B samples showed three and four curve-fitted values,
respectively (Fig. 3), and the Pd 3d spectra for Pd/Ni and Pd/Ni–
Fig. 5. XPS spectra of B 1s for Pd/Ni–B treated at different temperatures in a 40-ml/
min N2 flow.
Table 1
Surface contents of B and Pd analyzed from XPS characterization in Pd/Ni–B catalyst
treated at different temperatures
◦
Treatment temperature ( C)
Content of B (wt%)
Content of Pd (wt%)
25
200
300
350
6.4
5.2
5.1
4.9
4.3
2.1
1.7
1.0
◦
B exhibited lower curve-fitted values than those treated at 25 C
(Fig. 4). These findings indicate a stronger alloying effect of Pd with
Ni in the Pd/Ni–B sample, due to the structure in the amorphous
nanoparticles [15,36,37] and the presence of boron [22,33]. Accu-
rately distinguishing the valence states of the Ni species was still
difficult, however.
A well-dispersed supported catalyst would show a low concentra-
tion and relative XPS intensity of active metal. Table 1 and Table S1
in Supplementary material (surface contents of Pd in Pd/Ni catalyst
treated at different temperatures) present the surface concentra-
tions of the catalysts, suggesting a decreasing surface concentration
of Pd in the Pd/Ni–B and Pd/Ni particles with heating treatment.
For the NiPd bulk bimetallic nanoparticles, the Pd atoms tended to
aggregate on the surface of particle [32]. This decreased Pd concen-
A strong reaction and interdiffusion between the deposited Pd
and the amorphous B substrate even at room temperature has
been reported [38]. The BEs of the Pd 3d5/2 of the Pd/Ni–B catalyst
◦
treated at 25 C at 335.8 and 341.1 eV (Fig. 4) were equal to those
◦
of metal Pd [32]. For the Pd/Ni catalyst treated at 200 C, the peak

Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
327
Fig. 6. TPD spectra of various catalysts. (a) Hydrogen, (b) cholorobenzene, and (c) benzene. Pure Pd was prepared by reducing PdCl2 by H2 gas.
of Pd 3d5/2 was split into two peaks, one at 335.4 eV and the other
at 333.5 eV; however, the peaks of Pd 3d5/2 in the Pd/Ni–B catalyst
treated at 200 C were at 335.6 and 334.1 eV. For the Pd/Ni cata-
ing to 188.4 eV is due to B0 [33]. The Pd/Ni–B catalyst treated at
200 C showed a lower peak at 187.1 eV. The decreased BE results
◦
◦
from B’s acceptance of electrons from other metal atoms. Such a
shift was not observed in the treated Ni–B sample [33]. On the
other hand, the reaction of crystalline Ni and B species in the solid
state to produce amorphous material (known as a solid-sate amor-
phization reaction [SSAR]), in which B transfers electrons to Ni, has
been reported [39]. Here the change in BE should be due to the in-
teractions among Ni, Pd, and B. In short, the amorphous structure
of Ni–B and the presence of B promotes the surface-alloying ef-
fect of Pd and Ni; then the surface alloy species on the Pd/Ni–B
nanoparticles should be Pd–Ni–B rather than Pd–Ni.
lyst, the decreased BE of Pd 3d was likely due to electron donation
during the transfer from Ni to Pd. In the presence of amorphous
Ni–B particles, the chemical shift of the B 1s in a metal boride
is influenced by B atom coordination and unoccupied d-electron
states in the metal atom [38]; thus, the strong interaction between
Pd and B atoms resulted in the smaller decrease in the BE of Pd
to 334.1 eV, as shown in Fig. 4. This interaction has been demon-
strated through an electron transfer from Pd to B atoms, resulting
in the formation of palladium boride [38]. Thus, the BE of Pd in
Pd/Ni–B was higher than that of Pd/Ni, as shown in Fig. 4. The B 1s
spectra of the Pd/Ni–B sample are shown in Fig. 5. Fig. S4 in Sup-
plementary material showed the XPS spectra of B 1s spectra before
TPD of HDC reactants and products was performed to obtain
information on the relationship between HDC activity and catalyst
surface properties. Fig. 6 presents the TPD spectra of H2, benzene,
and chlorobenzene. The H2 TPD profile of the nano Ni catalyst ex-
+
◦
Ar etching. The peak of Pd/Ni–B treated at 25 C correspond-

328
Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
Table 2
HDC activities and stability of catalysts
Samples (treatment
temperature)
First
Recycle 1
Recycle 2
Recycle 3
Recycle 4
◦
Pd/Ni (25 C)
56%
53%
60%
91%
90%
95%
7%
11%
56%
87%
91%
93%
–
–
47%
91%
88%
94%
–
–
35%
71%
77%
85%
–
–
6%
32%
45%
54%
◦
Pd/Ni (200 C)
Pd/Ni–B (25 C)
Pd/Ni–B (200 C)
Pd/Ni–B (300 C)
◦
◦
◦
◦
Pd/Ni–B/TiO2 (200 C)
of Cl species, splitting of hydrogen molecules, and desorption of
the reaction product all increased after the surface alloying of Pd
with Ni. TPD of chlorobenzene and benzene for the Pd/Ni–B cata-
lyst apparently is complex, unlike that for nano Ni. The desorption
◦
temperature of chlorobenzene increased from 300 C (shoulder at
◦
◦
◦
256 C) over Ni–B to 395 C (satellite 360 C) over Pd/Ni–B. Two
more peaks at 203 and 510 C were observed in the Pd/Ni–B. The
desorption temperature of benzene TPD for Ni–B was 290 C, much
lower than that of nano Ni (430 C). This was due to the transfer
◦
◦
◦
of electrons from the B to the Ni atoms and the resulting en-
richment of Ni with electrons, which improved the adsorption of
benzene [22,23]. When Pd was deposited on the amorphous B, the
Pd atoms transferred electrons to B, resulting in Pd boride [38] and
a complex electron distribution over the Pd, Ni, and B. The pres-
ence of B led to the slightly higher (1–2 eV) BEs of Ni and Pd
in Pd/Ni–B than in Pd/Ni. Moreover, the benzene TPD of Pd/Ni–B
◦
◦
exhibited two main peaks at 375 (satellite 335 C) and 175 C, sug-
gesting the formation of a new, more active component after Pd
deposition. In our experiment, we found it difficult to distinguish
these two main peaks. The desorption of hydrogen and adsorp-
tion of chlorobenzene were much greater in Pd/Ni–B than in Pd/Ni,
which was ascribed to the surface alloying effect of Ni, Pd, and B,
which also improved the HDC activity. The surface alloy also en-
hanced catalyst stability; the Pd/Ni–B catalyst could be used three
times with no decrease in activity, whereas the Pd/Ni could be
used only once (Table 2).
Fig. 7. Catalytic activities of catalysts treated in a 40-ml/min N2 flow at different
temperatures from 25 to 400 C.
◦
◦
hibited a main peak at 410 C. Compared with nano Ni, the pure Pd
◦
showed a peak at 217 C, indicating that the H was more weakly
bonded to Pd than to Ni. The main peak of the Pd/PVP was higher
◦
(at 232 C) than that for the pure Pd, as shown in Fig. 6a and
Fig. S5 in Supplementary material. The higher temperature and
larger peaks can be attributed to the different particle sizes of
Pd (Fig. S6 in Supplementary material). When depositing Pd on
3.2. HDC activity
◦
nano Ni, the desorption peak shifted to lower temperatures 365 C
Fig. 7 displays the HDC activity of each catalyst studied.
◦
(shoulder at 300 C), demonstrating a weaker hydrogen-bonding
Nanocrystalline Ni (Fig. S1 in Supplementary material) and Ni–B
demonstrated relatively low activity that decreased with increas-
ing treatment temperature. Sintering of Ni was observed in the
nano Ni and Ni–B particles treated by heating (Figs. S1 and S2 in
Supplementary material). The results indicate that the HDC activity
decreased with the sintering of nano Ni or Ni–B. The Pd/PVP cata-
lyst had a ∼20% higher benzene yield than the nano Ni and Ni–B
catalysts. Because the particle size of Pd was stabilized by the PVP
polymer, the HDC activity changed only slightly with changes in
treatment temperature. The HDC activity of Pd + Ni–B (a mixture
of Pd/PVP nanoparticles and Ni–B nanoparticles with weight ra-
tio 1/99) was similar to that of Pd/PVP when both were treated at
effect of Pd. Moreover, the area of the desorption peak increased,
indicating the formation of more active sites for hydrogen split-
ting. As shown in Fig. 7, the HDC activity over Pd/Ni was higher
than the sum of those over Pd and Ni. This suggests that the cat-
alytic activity can be promoted by the surface alloying of Pd and
Ni, in agreement with findings of Wang et al. [18–20]. Compared
with the nano Ni catalyst, Ni–B had a much lower desorption tem-
◦
◦
perature of H2 at 360 C (shoulder at 410 C), due to the increased
number of coordinatively unsaturated Ni sites due to the presence
of B [22,23]. On replacement of Ni by Pd in the Pd/Ni–B sam-
◦
◦
ple, the peak corresponding to 360 C shifted sharply to 260 C,
suggesting that the interaction of Pd with Ni–B promoted the hy-
drogen splitting capacity to a much greater degree than that of the
Pd/Ni catalyst, due to the lower temperature of the main desorp-
tion peak for the Pd/Ni–B catalyst.
◦
25 C, but decreased with increasing treatment temperature. The
◦
HDC activity of the Pd/Ni and Pd/Ni–B catalysts treated at 25 C
was similar to that of the mixtures of Pd/PVP with nano Ni or
of Pd/PVP with Ni–B. The Pd/Ni–B catalyst treated at 200–300 C
◦
For the TPD of chlorobenzene and benzene, a desorption peak
at 420 C was observed for the Pd/PVP catalyst. The pure Pd sam-
demonstrated the highest HDC activity (91%), but its activity de-
creased with increasing treatment temperature. Table 2 displays
the stability of as-prepared catalysts. Fig. 8 shows TEM images of
Pd/Ni–B/TiO2 prepared by the replacement reaction of Ni–B/TiO2.
◦
◦
ple exhibited a peak at 480 C for chlorobenzene TPD and one at
◦
450 C for benzene TPD. The main desorption peak of chloroben-
zene increased from 265 C over nano Ni to 345 C (shoulder at
305 C) over Pd/Ni, with the corresponding temperatures for ben-
zene decreasing from 275 and 430 C to 210 and 380 C (satellite
◦
◦
◦
The Pd/Ni–B in Pd/Ni–B/TiO2 catalyst treated at 25 C also ex-
◦
hibited a flower-like shape with a particle size of ∼35 nm. The
◦
◦
particle size changed little, but the particle shape changed to a
◦
◦
at 340 C). The area of peaks increased after Pd was deposited on
solid sphere, with an increase in treatment temperature to 400 C.
nano Ni. These results suggest that the capacities for adsorption
This finding suggests that the thermal stability of Pd/Ni–B was pro-

Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
329
◦
◦
Fig. 8. TEM images of Pd/Ni–B/TiO2 nanoparticles treated at a 40-ml/min N2 flow. (a) Treated at 25 C and (b) treated at 400 C.
moted by the support. Fig. 7 and Table 2 show that the supported
Pd/Ni–B/TiO2 exhibited the highest HDC activity and stability.
The XPS characterization done to investigate the surface species
formed in Pd/Ni–B catalyst showed that the surface alloy formed
in Pd/Ni–B by heating comprised Pd, Ni, and B atoms and differed
from the NiPd alloy in the Pd/Ni sample. Fig. 7 shows that the HDC
activity for Pd/Ni–B was increased due to the alloying effect of Pd,
Ni, and B and decreased by the NiPd alloy formed in the Pd/Ni cat-
alyst. Moreover, Table 1 and Table S1 in Supplementary material
show that the Pd surface concentrations decreased with increasing
heating temperature. Thus, it can be concluded that the presence
of B in the bimetallic nanoparticles mainly promoted the catalytic
properties. This means that the surface alloy formed on Pd/Ni–B
nanoparticles had a promoting effect on the HDC reactions. On
the other hand, the HDC activity of the Pd/Ni–B catalyst increased
4. Discussion
The superior catalytic activity for hydrogenation of both C=C
and C=O bonds in bimetallic nanoparticles synthesized by the
replacement reaction method is attributed mainly to the highly
dispersed noble metal [18–20]. Generally, the HDC reaction is ex-
plained by hydrogenolysis reactions. The particle size and disper-
sion of active sites can affect the catalytic activity and selectivity
in this reaction [10,40]. There is still no general consensus regard-
ing the structure-sensitivity of the HDC reaction, however [40,41].
For structure-sensitive reactions, the catalytic activity and stabil-
ity is thought to be related to the particle/crystal size of the active
metal, rather than to the dispersion of active metal [10]. For the
structure-sensitive reaction catalyzed by metal catalysts, a popular
hypothesis is that the chemisorbed intermediate forms a number
of bonds with the metal surface, and thus a surface site consisting
of a single active metal is not adequate [42,43]. Finding a suitable
array of surface atoms to accommodate such a chemisorbed inter-
mediate poses no difficulty on a large metal crystal. On very small
crystallites, however, a large fraction of the surface atoms exist at
edges and corners. If the required array comprised a large number
of metal atoms, then the likelihood of identifying such an array in
the surface might be significantly lower than on large crystals, and
the specific hydrogenolysis activity of highly dispersed metal clus-
ters would be lower than that of large crystals [10]. This means
that the reaction requires that the crystal size of the metal pro-
vide sufficient surface metal atoms to form a crystal to adsorb the
intermediate. In a metal particle catalyst, crystal size is related to
particle size [10], and the larger particle size of metal could pro-
vide more active sites for reactions. Some reports have suggested
that the HDC of chlorobenzene over Ni catalyst is a structure-
sensitive reaction [44,45]; HDC activity increases with increasing
particle size, but exhibits little relation to the dispersion of Ni.
As shown in Fig. 7, the HDC activity of Ni–B and Ni nanoparti-
cles decreased with increasing treatment temperature and particle
size (Figs. S1 and S2 in Supplementary material); this can be ex-
plained by the decrease in catalyst surface area. Fig. 2 shows that
the Pd/Ni–B bimetallic nanoparticles grew from ∼30 to ∼50 nm
◦
with an increase in treatment temperature from 25 to 300 C
◦
but decreased at treatment temperatures above 300 C (Fig. 7).
This should be related to the increased particle size, as shown
in Fig. 2. Aramendía et al. found an association between the HDC
of chlorobenzene and the particle size of Pd, and demonstrated
that higher activity was obtained over catalysts with larger parti-
cle sizes [41]; however, in the present work, the Pd/Ni–B treated
◦
at 350 C, which had the largest particle size, exhibited the lowest
HDC activity (Fig. 2). Aramendía et al. believed that Pd’s larger par-
ticle led to formation of the ρPdH phase, which should be stable
under the reaction conditions [41]. The hydride phase would be an
alternative source of hydrogen. These authors also noted that the
larger particles should show better resistance against passivation
or poisoning. The catalysts with low dispersion (i.e., those with
low metal surface area and large particle size) could promote the
migration of chloride species into large particles, thereby cleaning
the outer surface for further reaction [40,41,44]. But we found that
the hydrogen-splitting ability decreased with increasing treatment
temperature, whereas the absorption ability of benzene increased
(Fig. 6). This suggests that the smaller outer surface (active site) of
larger particles restrains the active sites for splitting hydrogen but
promotes absorption of the HDC product (benzene) (Fig. S5 in Sup-
plementary material and Fig. 6), resulting in a low hydrogenation
rate for the catalyst with larger particle size (54 ml/min hydro-
◦
genation uptake rate in Pd/Ni–B treated at 200 C, and 30 ml/min
◦
treated at 400 C). Here the particle size of the Pd/Ni–B with the
highest HDC activity was about 50 nm, whereas that of the sup-
ported Pd/Ni–B/TiO2 with the highest HDC activity was 35 nm
(72 ml/min hydrogenation uptake rate), as shown in Figs. 7 and 8.
Along with its better HDC activity compared with Pd/Ni, the
Pd/Ni–B catalyst also demonstrated better stability (Table 2). It was
previously reported that Pd has electronegative properties in NiPd
alloy [34,35]. For chlorinated aromatics, the C–Cl bond acquires a
double-bond character as the chlorine atom loses p-electrons due
to resonance [46]. This is followed by the formation of diadsorbed
chloroaromatic species from the adsorption of chloroaromatic com-
pounds on nickel and palladium [47,48], which are attacked by
◦
with an increase in treatment temperature from 25 to 300 C, fol-
lowed by a further increase of 20–150 nm on subsequent heating
◦
to 400 C. The HDC activity increased from 60 to 90% with an in-
◦
crease in the treatment temperature of Pd/Ni–B from 25 to 300 C
(Fig. 7). These findings indicate that the promotion of HDC activity
is ascribed to the surface alloying effect of Pd–Ni–B, rather than
to the increased particle size. Figs. 2 and 7 show a particle size of
ca. 50 nm for Pd/Ni–B nanoparticles with the highest HDC activity
measured.

330
Z. Wu et al. / Journal of Catalysis 256 (2008) 323–330
hydride ions. For the HDC of chlorobenzene, this results in the for-
mation of adsorbed phenyl anions, which can accept protons and
thus convert to benzene [46]. The electron-deficient Pd particles
are more resistant to chlorine attack as the chlorine is weakly ad-
sorbed, due to its electrophilic character [1]. Fig. 6 shows that the
deposition of Pd atoms on nano Ni increased the adsorption of
chlorobenzene and splitting of hydrogen; however, the electron-
enriched Pd atoms in Pd/Ni catalyst conferred considerable insta-
bility to the catalysts. This property is demonstrated by the HDC
over Pd/Ni, in which the hydrogen uptake rate was as high as
67 ml/min (54 ml/min over Pd/Ni–B) initially and then rapidly de-
creased to about 10 ml/min (40 ml/min for Pd/Ni–B) after 10 min
of reaction. This finding demonstrates that the Pd/Ni catalyst de-
activated rapidly in the HDC reaction (Table 2). The Pd/Ni–B/TiO2
catalyst exhibited a less-pronounced decrease in hydrogen uptake
rate, from 72 to 65 ml/min after 10 min of reaction. Besides the
surface alloying effect of Pd–Ni–B, the excellent activity and sta-
bility also can be attributed to the interaction between the active
metal and the support [49].
[3] Y. Monguchi, A. Kume, K. Hattori, T. Maegawa, H. Sajiki, Tetrahedron 62 (2006)
7926.
[4] F.J. Urbano, J.M. Marinas, J. Mol. Catal. A 173 (2001) 329.
[5] R. Hara, K. Sato, W.H. Sun, T. Takahashi, Chem. Commun. 12 (1999) 845.
[6] U.R. Pillai, E.S. Sahle-Demessie, R.S. Varma, Green Chem. 6 (2004) 295.
[7] N. Lingaiah, N.S. Babu, R. Gopinath, P.S.S. Reddy, P.S.S. Prasad, Catal. Surv. Asia
10 (2006) 29.
[8] M.M. Yurii, S. Moshe, Catal. Today 75 (2002) 63.
[9] M.A. Keane, J. Chem. Technol. Biotechnol. 80 (2005) 1211.
[10] J.H. Sinfelt, Bimetallic Catalysts: Discoveries, Concepts, and Application, Wiley,
New York, 1983, p. 5.
[11] M.O. Nutt, J.B. Hughes, M.S. Wong, Environ. Sci. Technol. 39 (2005) 1346.
[12] M.O. Nutt, K.N. Hechk, P. Alvarez, M.S. Wong, Appl. Catal. B 69 (2006) 115.
[13] Y.G. Sun, B.T. Mayers, Y.N. Xia, Nano Lett. 2 (2002) 481.
[14] Y.G. Sun, Z.L. Tao, J. Chen, T. Herricks, Y.N. Xia, J. Am. Chem. Soc. 126 (2004)
5940.
[15] A.K. Sra, R.E. Schaak, J. Am. Chem. Soc. 126 (2004) 6667.
[16] W.R. Lee, M.G. Kim, J.R. Choi, J.I. Park, S.J. Ko, S.J. Oh, J. Cheon, J. Am. Chem.
Soc. 127 (2005) 16090.
[17] D.Y. Wang, C.H. Chen, H.C. Yen, Y.L. Lin, P.Y. Huang, B.J. Hwang, C.C. Chen, J. Am.
Chem. Soc. 129 (2007) 1538.
[18] S.R. Wang, W. Lin, Y.X. Zhu, Y.C. Xie, J.G. Chen, Chin. J. Catal. 27 (2006) 301.
[19] S.R. Wang, Y.X. Zhu, Y.C. Xie, J.G. Chen, Chin. J. Catal. 28 (2007) 676.
[20] S.R. Wang, W. Lin, Y.X. Zhu, Y.C. Xie, J.R. McCormick, W. Huang, J.G. Chen, Catal.
Lett. 114 (2007) 169.
[21] F. Cheng, H. Ma, Y. Li, J. Chen, Inorg. Chem. 46 (2007) 788.
[22] J.F. Deng, H.X. Li, W.J. Wang, Catal. Today 51 (1999) 113.
[23] Z.J. Wu, W. Li, M.H. Zhang, K.Y. Tao, Front. Chem. Eng. China 1 (2007) 87.
[24] X. Ni, X. Su, Z. Yang, H. Zheng, J. Cryst. Grow. 252 (2003) 612.
[25] Y. Gao, F.D. Wang, S.J. Liao, D.R. Yu, N. Sun, React. Funct. Polym. 44 (2000) 65.
[26] Z.Y. Li, C.H. Han, J.Y. Shen, J. Mater. Sci. 41 (2006) 3473.
[27] Z.J. Wu, M.H. Zhang, S.H. Ge, Z.L. Zhang, W. Li, K.Y. Tao, J. Mater. Chem. 15
(2005) 4928.
[28] Z.J. Wu, S.H. Ge, M.H. Zhang, W. Li, S.C. Mu, K.Y. Tao, J. Phys. Chem. C 111 (2007)
8587.
[29] N. Hedgecock, P. Tung, M. Schlesinger, J. Electrochem. Soc. 122 (1975) 866.
[30] J.P. Espinos, J. Morales, A. Barranco, A. Caballero, J.P. Holgado, A.R. Gonzalez-
Elipe, J. Phys. Chem. B 106 (2002) 6921.
5. Conclusion
In this study, bimetallic Pd/Ni–B nanoparticles were prepared
by the replacement reaction. XPS and TPD characterization demon-
strated the formation of a surface alloy of Pd–Ni–B. The strong
surface alloying of Ni and Pd in the Pd/Ni–B catalyst was at-
tributed to the presence of B. Compared with Pd and Pd/Ni cat-
alysts, the Ni-Pd-B surface alloy of Pd/Ni–B exhibited increased
hydrogen-splitting and chlorobenzene-adsorption capabilities, as
well as increased resistance to chlorine deactivation. Thus, the
Pd/Ni–B catalyst showed better activity and stability in the HDC
of chlorobenzene than the Pd/Ni and Pd/PVP catalysts. An investi-
gation of particle size in chlorobenzene HDC found an association
between the HDC activity of chlorobenzene and the size of the
bimetallic nanoparticles. The most suitable catalyst was the Pd/Ni–
[31] A.R. Naghash, T.H. Etsell, S. Xu, Chem. Mater. 18 (2006) 2480.
[32] P. Lu, T. Teranishi, K. Asakura, M. Miyake, N. Toshima, J. Phys. Chem. B 103
(1999) 9673.
[33] H. Li, H.X. Li, W.L. Dai, W.J. Wang, Z.G. Fang, J.F. Deng, Appl. Surf. Sci. 152 (1999)
25.
◦
B catalyst treated at 200–300 C, with a particle size of ∼50 nm.
[34] G. Treglia, M.C. Desjonqueres, F. Duncastelle, D. Spanjaand, J. Phys. C 14 (1981)
4347.
Acknowledgments
[35] K. Kishi, Y. Motoyoshi, S. Ikeda, Surf. Sci. 105 (1982) 313.
[36] R.E. Schaak, A.K. Sra, B.M. Leonard, R.E. Cable, J.C. Bauer, Y.F. Han, J. Means,
W. Teizer, Y. Vasquez, E.S. Funck, J. Am. Chem. Soc. 127 (2005) 3506.
[37] R.E. Cable, R.E. Schaak, J. Am. Chem. Soc. 128 (2006) 9588.
[38] D.H. Shen, P.A. Kearney, J.M. Slaughter, C.M. Falco, J. Magn. Magn. Mater. 126
(1993) 25.
[39] A.N. Campbell, J.C. Barbour, C.R. Hills, J. Mater. Res. 4 (1989) 1303.
[40] F.H. Ribeiro, C.A. Gerken, G. Rupprechter, G.A. Somorjai, C.S. Kellner, G.W. Coul-
ston, L.E. Manzer, L. Abrams, J. Catal. 176 (1998) 352.
This work was supported by the National Science Foundation
of China (20403009) and the Key Project of the Chinese Ministry
of Education (105045). The authors thank Professor R. Prins for re-
viewing the manuscript.
Supplementary material
[41] M.A. Aramendía, V. Boráu, I.M. García, C. Jiménez, F. Lafont, A. Marinas, J.M.
Marinas, F.J. Urbano, J. Catal. 187 (1999) 392.
[42] J.H. Sinfelt, J.L. Carter, D.J.C. Yates, J. Catal. 24 (1972) 283.
[43] V. Ponec, W.M.H. Sachtler, J. Catal. 24 (1972) 250.
[44] Y.H. Choi, W.Y. Lee, Catal. Lett. 67 (2000) 155.
The online version of this article contains additional supple-
mentary material.
Please visit DOI: 10.1016/j.jcat.2008.03.026.
[45] M.A. Keane, C. Park, C. Menini, Catal. Lett. 88 (2003) 89.
[46] V.A. Yakovlev, V.V. Terskikh, V.I. Simagina, V.A. Likholobov, J. Mol. Catal. A 153
(2000) 231.
References
[47] R.B. LaPierre, D. Wu, W.L. Kranich, J. Catal. 52 (1978) 59.
[48] R.B. LaPierre, L. Guczi, W.L. Kranich, J. Catal. 52 (1978) 230.
[49] S.J. Tauster, S.C. Fung, R.L. Garten, J. Am. Chem. Soc. 100 (1978) 170.
[1] B. Coq, G. Ferrat, F. Figueras, J. Catal. 101 (1986) 434.
[2] M.L. Hitchman, R.A. Spackman, N.C. Ross, C. Agra, Chem. Soc. Rev. 95 (1995)
423.