Pd catalysts supported on MnCeOx mixed oxides and their catalytic application in solvent-free aerobic oxidation of benzyl alcohol: Support composition and structure sensitivity

Article abstract of DOI:10.1016/j.jcat.2011.06.021

Both crystalline and amorphous MnCeO_x supports were synthesized by co-precipitation and redox precipitation methods, respectively. Pd was subsequently deposited by an easy microwave-assisted polyol reduction procedure, leading to the formation

Full text of DOI:10.1016/j.jcat.2011.06.021

Journal of Catalysis 283 (2011) 34–44
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Pd catalysts supported on MnCeOx mixed oxides and their catalytic application
in solvent-free aerobic oxidation of benzyl alcohol: Support composition
and structure sensitivity
Yuanting Chen a, Huijian Zheng a, Zhen Guo a, Chunmei Zhou a, Chuan Wang b, Armando Borgna b,
,
⇑
Yanhui Yang a,
⇑
a School of Chemical and Biomedical Engineering, Nanyang Technological University, Singapore 637459, Singapore
b Institute of Chemical and Engineering Sciences (ICES), A⁄STAR, 1 Pesek Road, Jurong Island, Singapore 627833, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
Both crystalline and amorphous MnCeOx supports were synthesized by co-precipitation and redox pre-
cipitation methods, respectively. Pd was subsequently deposited by an easy microwave-assisted polyol
reduction procedure, leading to the formation of highly dispersed Pd nanoclusters. MnCeOx supports
were remarkably enhanced in both catalytic activity and selectivity in the aerobic oxidation of benzyl
alcohol, compared with pure MnOx and CeO2. The highest qTOFs (quasi-turnover frequencies) were
achieved over Pd/7Mn3Ce-C (15,235 hꢀ1) and Pd/7Mn3Ce-A (14,438 hꢀ1), and the activity could be main-
tained over five consecutive reaction runs. Pd acts as a single active component, and the synergetic inter-
actions among Pd, MnOx, and CeO2 result in enhanced catalytic activity. Good accessibility of the Pd active
sites and a high surface concentration of Pd0 contribute to the high initial reaction rate over crystalline
MnCeOx-supported Pd catalyst. Amorphous MnCeOx-supported Pd catalyst exhibits enhanced catalyst
stability due to mutual promotion between redox properties and oxygen mobility.
Received 10 May 2011
Revised 17 June 2011
Accepted 21 June 2011
Available online 31 August 2011
Keywords:
Manganese–cerium mixed oxides
Palladium
Benzyl alcohol oxidation
Composition and structure sensitivity
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
compositions and the textural, structural, and redox properties of
these materials [1,10,11]. Several preparation routes have been
Among various transition metal and rare earth mixed oxide cat-
alysts, MnCeOx has attracted great attention because of its poten-
tial industrial applications, due to its high activity and versatility
in many catalytic reactions, such as the catalytic wet oxidation of
waste waters containing phenol [1], ammonia [2], and pyridine
[3], the combustion of volatile organic compounds [4], and the con-
version/decomposition of NOx [5,6]. The superior activity of
MnCeOx compared with pure MnOx and CeO2 oxides is attributed
to the synergistic interactions between manganese and cerium
oxides. The MnOx domain, with high dispersion and multiple
oxidation states, serves as the main active site, while CeO2 acts
as a promoter to enhance the oxygen mobility, due to its unique
capacity of storing oxygen under oxidizing conditions and releas-
ing oxygen under reducing conditions. This is achieved through
the transformation between Ce4+ and Ce3+ oxidation states, form-
ing abundant surface and bulk vacancies to improve the redox
capabilities [7–9].
developed to obtain MnCeOx with controlled physicochemical
properties. Co-precipitation [12] and redox precipitation [7] are
commonly employed to prepare crystalline and amorphous mate-
rials, respectively. The phase segregation of oxide domains corre-
lates with the Mn/Ce molar ratio and the preparation method
[13]. Due to the different precipitation kinetics of metal precursors,
co-precipitation methods result in segregated manganese and cer-
ium oxide crystallites with surface-enriched Ce. More specifically,
either a thin layer of Ce-rich oxide covering the Mn crystallites
or coexisting large MnOx crystallites and small aggregates of Mn/
Ce solid solution seems to be formed [14]. In contrast, redox pre-
cipitation routes favor the simultaneous precipitation of MnOx
and CeO2, leading to a homogeneous mixture of Mn and Ce oxide
species at the atomic level [15]. Our group has synthesized a series
of amorphous MnCeOx catalysts by redox precipitation, which have
been tested in the catalytic aerobic oxidation of benzyl alcohol
[16]. This work demonstrated that MnOx is the active component
and the addition of highly dispersed CeO2 promotes the catalytic
performance by enhancing the redox properties of MnCeOx.
Nonetheless, a rapid loss of catalytic activity was found over
MnCeOx materials in most cases, which is mainly attributed to
either the deposition of carbonaceous by-products on the catalyst
surface or the leaching of the active metals during the reaction.
A large number of previous studies have demonstrated that the
catalytic activity of MnCeOx composites strongly depends on the
⇑
Corresponding authors. Fax: +65 6794 7553.
E-mail addresses: armando_borgna@ices.a-star.edu.sg (A. Borgna), yhyang@
ntu.edu.sg (Y. Yang).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.06.021

Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
35
This activity loss can be suppressed by adding a promoter. Santiago
et al. reported that doping MnCeOx mixed oxides with a certain
amount of potassium showed improved performance and stability
of the catalyst, since K+ could act as a modifier or chemical pro-
moter [17]. Hamoudi et al. suggested that the catalyst deactivation
could be diminished by a three-component catalyst using noble
metals promoters, e.g., Ag and Pt, to significantly reduce the
amount of carbonaceous deposits, prolonging the catalyst life
[18]. Tang et al. studied the effect of Ag addition to MnCeOx on
the complete oxidation of formaldehyde and suggested that the
enhanced activity was due to improved oxygen activation on the
catalyst surfaces and to oxygen mobility among Ag, Mn, and Ce
[12]. Ru catalysts supported on MnCeOx were also reported to be
highly active for alcohol oxidation reactions [19,20]. Recently,
Zou et al. have shown that a Pd/MnOx–CeO2 catalyst synthesized
by a surfactant-assisted precipitation method exhibits excellent
activity in CO oxidation due to accelerated oxygen activation in
the gas phase and oxygen transfer via spillover [21].
In this study, we prepared both crystalline and amorphous
MnCeOx supports. Then, Pd particles were supported on these
mixed oxides by an easy microwave-assisted polyol reduction
(MAPR) method. The catalytic performance in the selective oxida-
tion of benzyl alcohol using molecular oxygen under solvent-free
conditions was evaluated. Benzyl alcohol oxidation was selected
as a probe reaction due not only to its relatively high reactivity
but also to the limited number of possible by-products. Various
physicochemical characterization techniques were employed to
investigate the textural, structural, and redox properties as well
as the surface composition of these Pd/MnCeOx catalysts. The cat-
alytic performance was correlated with these properties in a com-
parative manner between amorphous and crystalline support
materials.
The obtained solid was added to ethylene glycol (40 mL) and son-
icated for 30 min to produce a homogenous suspension. The sus-
pension was transferred into
a three-necked flask with a
condenser and placed in a microwave reactor (Sineo, MAS-II).
The slurry was agitated by a magnetic stirrer, and the temperature
inside the flask was monitored by an infrared sensor. The suspen-
sion was heated to 160 °C in 0.5 min and kept at the same temper-
ature for another 1.5 min. After cooling to room temperature, the
solid was filtered, washed with deionized water, and dried at
100 °C in vacuum. The as-synthesized catalyst was denoted as
Pd/xMnyCe-M (M = C or A for crystalline or amorphous structure,
respectively).
2.2. Catalyst characterizations
Nitrogen physisorption isotherms were measured at ꢀ196 °C on
a static volumetric instrument (Autosorb-6b, Quanta Chrome).
Prior to each measurement, the sample was degassed at 250 °C
for 12 h under high vacuum. The specific surface area was calcu-
lated by the Brunauer–Emmett–Teller method [23]; the pore size
distribution was calculated by the Barrett–Joyner–Halenda method
using the desorption branch [24].
Powder X-ray diffraction (XRD) patterns were recorded on a
Bruker AXS D8 diffractometer under ambient conditions using fil-
tered Cu
Ka radiation (k = 1.5406 Å) operated at 40 kV and
40 mA. Diffraction data were collected from 20° to 70° with a res-
olution of 0.02° (2h). Prior to the test, samples were dried at 100 °C
overnight. The average crystallite sizes of MnCeOx grains were cal-
culated from the full width at half maximum of the ceria (1 1 1) or
Mn2O3 (Bixbyite-C) (2 2 2) peak according to Scherrer’s equation
and using a Scherrer constant of 0.89. The lattice constants were
derived following the Cohen procedures [25].
Transmission electron microscopy (TEM) was performed on a
JEOL JEM-2100F operated at 200 kV. The samples were suspended
in ethanol and dispersed on a holey carbon-coated Cu grid. The me-
tal content was measured by inductively coupled plasma (ICP)
analysis using a dual-view Optima 5300 DV ICP-OES system. Nitric
acid was used to dissolve the sample.
2. Experimental
2.1. Catalyst synthesis
Crystalline MnCeOx supports with different Mn/Ce molar ratios
(1:3, 7:3, and 10:1) were synthesized following a well-established
Hydrogen temperature-programmed reduction (H2 TPR) and
oxygen temperature-programmed desorption (O2 TPD) measure-
ments were carried out using an Autosorb-1C system (Quanta
Chrome), equipped with a thermal conductivity detector. Prior to
each H2 TPR run, 100 mg of the sample was loaded in the quartz
tube and pretreated with ultra-pure air at 100 °C for 3 h. After cool-
ing to room temperature, the gas flow was switched to 5 vol.% H2/
Ar and the baseline was monitored until it became stable. The sam-
ple cell was heated at a ramping rate of 5 °C/min and held at 600 °C
for 1 h to ensure complete reduction. For the O2 TPD test, 100 mg of
the sample was pretreated in pure O2 at 550 °C for 1 h. After cool-
ing to room temperature, the sample was flushed with pure helium
until the baseline became stable. The desorption was conducted by
heating the sample to 900 °C with a heating rate of 10 °C/min.
X-ray photoelectron spectroscopy (XPS) was performed on a VG
co-precipitation procedure [8]: 100 mL of
a
solution of
Mn(NO3)2ꢁ4H2O and Ce(NO3)3ꢁ6H2O with a total concentration of
150 g/L and a given nominal Mn/Ce mole ratio was prepared. A 2-
M NaOH aqueous solution was added dropwise at 50 °C under vigor-
ous stirring until pH 10.5. The precipitate was further aged at 50 °C
for 2 h. The obtained solid was filtered, washed with deionized
water, dried at 100 °C for 12 h, and calcined at 500 °C for 6 h in air.
The as-calcined MnCeOx support was designed as xMnyCe-C, x/y
being the nominal molar ratio of Mn/Ce. The amorphous MnCeOx
supports were prepared by redox precipitation [15]: A certain
amount of KMnO4 precursor dissolved in deionized water was ti-
trated with an aqueous solution containing both Mn(NO3)2ꢁ4H2O
and Ce(NO3)3ꢁ6H2O at 70 °C under vigorous stirring, where KMnO4
was in 10% stoichiometric excess. A 2-M NaOH aqueous solution
was added dropwise to maintain a constant pH value at 8.0. After
further stirring for 1 h, the obtained solid was filtered, repeatedly
washed with hot deionized water, dried at 100 °C for 12 h, and cal-
cined in air at 400 °C for 6 h. This support was denoted as xMnyCe-
A. For comparison, pure manganese oxide (MnOx) and cerium oxide
(CeO2) were also prepared by co-precipitation using Mn(NO3)2ꢁ
4H2O and Ce(NO3)3ꢁ6H2O as precursors.
Escalab 250 spectrometer equipped with an Al anode (Al Ka =
1486.6 eV). The background pressure in the analysis chamber
was lower than 1 ꢂ 10ꢀ7 Pa. Measurements were performed using
20-eV pass energy, 0.1-eV step, and 0.15-min dwelling time. The
correction of the binding energies (BE) was carried out using the
C1s peak of adventitious C at 284.6 eV. The background contribu-
tion B (E) caused by inelastic process was obtained by the Shirley
method and subtracted. The curve fitting was performed with a
Gaussian–Lorentzian shape function to deconvolute the overlap-
ping peaks. The atomic ratios of the different elements were
estimated using the ratio of the normalized peak intensities cor-
rected with the atomic sensitivity factors (7.399 for Ce 3d, 2.420
for Mn 2p, 0.711 for O 1s, and 4.462 for Pd 3d).
Palladium nanoparticles supported on these calcined MnCeOx
mixed oxides were successfully prepared by an easy MAPR proce-
dure as reported by Guo et al. [22]. To deposit Pd precursors onto
MnCeOx mixed oxides, 0.1 g of support was immersed in 94.0
lL
of PdCl2 aqueous solution (0.05 M) and dried at 100 °C in vacuum.

36
Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
2.3. Catalytic activity
undesired by-product. Pd/MnOx exhibits a poor catalytic conver-
sion of 17.9% along with 99.4% benzaldehyde selectivity. Regard-
less of the support structure and composition, either benzyl
alcohol conversion or selectivity toward benzaldehyde can be en-
hanced using MnCeOx mixed oxide supports. The benzaldehyde
selectivity is well maintained above 98%, whereas the catalytic
activity depends on the nature and composition of support materi-
als. For Pd catalysts supported on both MnCeOx-C and MnCeOx-A
materials, the catalytic activity (qTOF) first increases to a maxi-
mum value after the addition of a small amount of Mn to Ce, and
then slightly decreases with further addition. The highest qTOFs
are achieved over Pd/7Mn3Ce-C (15,325 hꢀ1) and Pd/7Mn3Ce-A
(14,438 hꢀ1) for MnCeOx crystalline and amorphous support mate-
rials, respectively. A control reaction was also carried out over a
physical mixture of Pd/CeO2 and Pd/MnOx catalysts. A poor conver-
sion of 9.6% is obtained, indicating that only chemically mixed Mn
and Ce species can facilitate the catalytic reaction. In addition, bare
supports show negligible catalytic activity (entries 10 and 11),
implying that Pd acts as the sole active species. Therefore, the im-
proved catalytic activity on Mn–Ce mixed oxides could be corre-
lated with the synergetic interactions among Pd, CeO2, and MnOx
domains. As a structural and electronic promoter, CeO2 improves
the oxygen storage and release ability, enhances the oxygen mobil-
ity and availability, and tunes the redox properties of the mixed
oxides when associated with MnOx [16].
The evolution of the catalytic performance as a function of time
was studied over Pd/7Mn3Ce-C and Pd/7Mn3Ce-A catalysts. The re-
sults are depicted in Fig. 1. The conversion of benzyl alcohol in-
creases monotonically with the reaction time, whereas the qTOF
decreases. The selectivity toward benzaldehyde is high (P97%)
over both catalysts during the first 0.5 h of reaction. Nevertheless,
it gradually deteriorates at longer reaction times, which is ascribed
to the formation of benzoic acid by further oxidation of benzalde-
hyde. The benzyl alcohol conversion over Pd/7Mn3Ce-C reaches a
plateau after 4 h, while that of Pd/7Mn3Ce-A increases continu-
ously, implying faster deactivation for Pd/7Mn3Ce-C, which may
be attributed to the pronounced deposition of carbonaceous species
[19]. After 12 h, 63.1% and 81.9% conversion of benzyl alcohol is ob-
tained over Pd/7Mn3Ce-C and Pd/7Mn3Ce-A catalysts, respectively.
Recyclability is a critical factor in evaluating a heterogeneous
catalyst. When the metal component is not properly anchored onto
the support, it may be leached out during the course of the reac-
tion, resulting in an active homogeneous catalyst and a loss of
catalytic activity on subsequent runs. The recyclabilities of both
The solvent-free aerobic oxidation of benzyl alcohol using
molecular O2 was carried out in a batch-type reactor operated un-
der atmospheric conditions. Experiments were conducted using a
three-necked glass flask (capacity 25 mL) precharged with 5.4 g
of benzyl alcohol and 20 mg of catalyst. The mixture was stirred
using a magnetic stirrer and heated in a silicon oil bath. The system
was equipped with a thermocouple to control the temperature and
a reflux condenser. In each reaction run, the mixture was heated to
160 °C under vigorous stirring (stirring rate 1000 rpm). Oxygen
was bubbled into the mixture at a constant flow rate of 20 mL/
min using a mass flow controller to initiate the reaction. After
the allowed reaction time, the catalyst powder was filtered off
and the liquid organic products were analyzed using an Agilent
6890 gas chromatograph equipped with an HP-5 capillary column
(30 m long and 0.32 mm in diameter, packed with silica-based
Supel-cosil). Dodecane was used as the internal standard to calcu-
late benzyl alcohol conversion and benzaldehyde selectivity. The
conversion, selectivity, and quasi-turnover frequency (qTOF) are
defined as follows:
moles of reactant converted
moles of reactant in feed
moles of product formed
conversion=% ¼
selectivity ð%Þ ¼
qTOF ðhꢀ1Þ ¼
ꢂ 100%
ꢂ 100%
moles of reactant converted
moles of reactant converted
moles of total active sites ꢂ reaction time
:
3. Results and discussion
3.1. Catalytic results
Both crystalline and amorphous MnCeOx-supported Pd catalysts
have been tested in the selective oxidation of benzyl alcohol in the
presence of molecular oxygen. The results are listed in Table 1. It is
possible to deposit 0.5 wt.% Pd on MnOx support, but only
0.36 wt.% on CeO2 In addition, the Pd loading increases with the
addition of Mn to Ce. Therefore, mixing Mn with Ce to form
MnCeOx mixed oxides leads to an enhanced Pd deposition, result-
ing in more catalytic active sites. Despite a 23.4% conversion of
benzyl alcohol, Pd/CeO2 shows 88.9% selectivity toward benzalde-
hyde due to the formation of a large amount of toluene as an
Table 1
Catalytic performance of 0.5 wt.% Pd-containing catalysts supported on MnCeOx.a
Entry
Catalyst
Pd contentb (wt.%)
Conversion (%)
Selectivity (%)
Benzaldehyde
qTOFc (hꢀ1
)
Toluene
Benzoic acid
1
2
3
4
5
6
7
8
9
10
11
12
13
Pd/CeO2
0.36
0.28
0.32
0.41
0.27
0.38
0.51
0.5
0.41
–
–
0.32
0.38
23.4
15.6
18.4
23.2
12.3
20.6
15.6
17.9
9.6
1.0
0.9
0.9
0.9
88.9
99.1
98.6
98.7
98.3
99.3
99.0
99.4
100
100
100
97.9
99
11.1
0.9
0.4
0.8
1.2
0.7
1.0
0.6
0
0
0
2.1
1
0
0
0
0.5
0.4
0
0
0
0
0
0
0
17,572
14,850
15,325
15,066
12,085
14,438
8688
9526
6246
–
Pd/1Mn3Ce-C
Pd/7Mn3Ce-C
Pd/10Mn1Ce-C
Pd/1Mn3Ce-A
Pd/7Mn3Ce-A
Pd/10Mn1Ce-A
Pd/MnOx
d
Pd/CeO2 + Pd/MnOx
10Mn1Ce-C
7Mn3Ce-A
–
777
605
Pd/7Mn3Ce-Ce
Pd/7Mn3Ce-Ae
0
a
b
c
Reaction conditions: catalyst, 20 mg; benzyl alcohol, 50 mmol; O2, 20 mL/min; temperature, 160 °C; reaction time, 1 h; carbon balance, 98.5%.
Pd content was tested by ICP.
qTOF is calculated using the Pd content obtained from ICP.
Physical mixture of Pd/CeO2 and Pd/MnOx catalyst (1:1 mass ratio).
N2 used instead of O2.
d
e

Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
37
100
80
60
40
20
0
20000
15000
10000
5000
100
80
60
40
20
0
20000
15000
10000
5000
0
benzaldehyde
selectivity
benzaldehyde
selectivity
Pd/7Mn3Ce-C
benzyl alcohol
conversion
Pd/7Mn3Ce-C
TOF
q
benzyl alcohol
conversion
q
TOF
dx/dt=0.23
r0=11.5 mmol/h
rs0=151316 mmol/gPd⋅h
1
2
3
4
5
0
2
4
6
8
10
number of batches
time, h
100
80
60
40
20
0
20000
15000
10000
5000
0
100
80
60
40
20
0
20000
15000
10000
5000
benzaldehyde
selectivity
benzaldehyde
selectivity
Pd/7Mn3Ce-A
benzayl alcohol
conversion
Pd/7Mn3Ce-A
TOF
q
benzyl alcohol
conversion
qTOF
dx/dt=0.17
r
=8.4 mmol/h
r0s0=110526 mmol/gPd⋅h
1
2
3
4
5
0
2
4
6
8
10
number of batches
time, h
Fig. 2. Recyclability of Pd/7Mn3Ce-C and Pd/7Mn3Ce-A for benzyl alcohol oxida-
tion. Reaction conditions: catalyst, 20 mg; benzyl alcohol, 50 mmol; O2, 20 mL/min;
temperature, 160 °C; reaction time, 1 h.
Fig. 1. Time courses of Pd/7Mn3Ce-C and Pd/7Mn3Ce-A for benzyl alcohol
oxidation. Reaction conditions: catalyst, 20 mg; benzyl alcohol, 50 mmol; O2,
20 mL/min; temperature, 160 °C; reaction time, 1 h.
Pd/7Mn3Ce-C and Pd/7Mn3Ce-A were examined for five consecu-
tive cycles, and the results are depicted in Fig. 2. The catalyst
was recovered after each reaction run, thoroughly washed with
acetone, dried at 60 °C, and reused in the next reaction. Both cata-
lysts maintain the high benzaldehyde selectivity well and undergo
a moderate decrease in the catalytic activities, which could be
attributed to the loss of catalyst during the recovery. The Pd con-
tent after five consecutive reaction cycles was almost the same
as that of the fresh catalysts. This clearly suggests that the interac-
tion between Mn and Ce, forming either crystalline or amorphous
MnCeOx mixed oxides, not only improves the catalytic perfor-
mance but also enhances the catalyst stability.
samples, implying that the crystal growth of MnCeOx-C can be sup-
pressed more significantly during the redox reaction among Mn2+
,
Ce3+, and MnOꢀ4 [26].
Fig. 3 shows the XRD patterns of MnCeOx supports along with the
catalysts containing 0.5 wt.% Pd. The calculated average grain sizes
of CeO2 and MnOx and the lattice constants (only for crystalline sam-
ples) are listed in Table 2. Pure MnOx exhibits the characteristic dif-
fraction peaks ascribed to Mn2O3 (Bixbyite-C, PDF# 41-1442) and
Mn3O4 (Hausmannite, PDF# 24-0734), suggesting the coexistence
of +2 and +3 oxidation states of Mn. The XRD pattern of pure cerium
oxide shows intense and sharp peaks at 2h = 28.5°, 33.1°, 47.5°, and
56.3°, which can be assigned to CeO2 (PDF# 43-1002). For the
1Mn3Ce-C sample, no other crystalline phases, but broad reflections
due to the cubic fluorite structure of CeO2, are observed, which could
be attributed to the formation of a solid solution between manga-
nese and cerium oxides by incorporation of Mn cations into CeO2
lattices. For 7Mn3Ce-C and 10Mn1Ce-C samples, in addition to the
broad reflection peaks of CeO2, diffraction peaks ascribed to
Mn2O3 and Mn3O4 appear, suggesting the coexistence of segregated
Mn and Ce oxide phases. The remarkable weakening and broadening
of these diffraction peaks with increasing Ce content suggest that
the incorporated cerium enhances the dispersion for both Ce and
Mn species [27]. Furthermore, several weak diffraction peaks that
fit the d-spacing of MnO2 (Ramsdellite, PDF# 39-0375) and Mn5O8
(PDF# 39-1218) are also observed, implying that Ce addition might
increase the oxidation state of Mn and stabilize Mn oxide species in
higher oxidation states. These results are consistent with a recent
report indicating that the phase composition of crystalline MnCeOx
strongly depends on the molar ratio of Mn/Ce [28].
3.2. Catalyst characterizations
Table 2 summarizes the textural properties of all synthesized
MnCeOx support materials, as well as MnOx and CeO2. Crystalline
MnOx has a rather low surface area of 36.3 m2/g. The specific sur-
face area increases to 48.0 m2/g upon addition of Ce and monoton-
ically increases with further Ce addition. 1Mn3Ce-C shows a
specific surface area of 116.0 m2/g, which is higher than that of
CeO2. The increased surface area after addition of Ce to MnOx im-
plies close interaction between MnOx and CeO2. Similarly, an in-
crease in Ce content also gradually increases the surface area for
MnCeOx-A samples, the 1Mn3Ce-A sample being the one that pos-
sesses the largest surface area (217.6 m2/g). In general, the pore
volumes of all MnCeOx samples are larger than those of pure MnOx
and CeO2. Amorphous MnCeOx support materials exhibit larger
surface areas but smaller pore diameters than crystalline MnCeOx

38
Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
Table 2
Textural and structural properties of MnCeOx support materials with different Mn/Ce molar ratios.
Entry
Sample
Molar ratio (Mn/Ce)
Surface area (m2/g)
Pore volume (cm3/g)
Pore diameter (nm)
Grain size (nm)
Lattice constant (nm)
1
2
3
4
5
6
7
8
CeO2
0:1
1:3
7:3
10:1
1:3
7:3
10:1
1:0
79.4
116.0
67.6
0.19
0.32
0.38
0.36
0.46
0.47
0.32
0.33
9.1
13.4
15.7
24.9
9.0
1.8
3.6
36.4
13.3
5.0
5.3
8.4
–
–
–
38.3
0.5416a
0.5279a
0.5318a
0.5376a
–
1Mn3Ce-C
7Mn3Ce-C
10Mn1Ce-C
1Mn3Ce-A
7Mn3Ce-A
10Mn1Ce-A
MnOx
48.0
217.6
190.6
101.5
36.3
–
–
0.9362b
a
Grain size and lattice constant were calculated according to the CeO2 (1 1 1) peak.
Grain size and lattice constant were calculated according to the Mn2O3 (Bixbyite-C) (2 2 2) peak.
b
contraction. This suggests the formation of manganese–cerium
oxides solid solutions. Furthermore, the diffraction peaks of CeO2
broaden after Mn addition, indicating that crystal growth and con-
comitant sintering are suppressed. Thus, small crystallites with a
more defective lattice and less crystalline character are formed.
Amorphous MnCeOx-A samples exhibit significantly different
XRD patterns than MnCeOx-C. A broad peak spanned in the 2h
range of 20–38°, along with other less intense peaks in the range
of 40–60°, suggests the lack of long-range order, which may be
due to the hindered growth of large crystalline domains during re-
dox precipitation, since Mn and Ce are intimately contacted [26]. A
decrease of peak intensity with increasing Mn/Ce ratio implies that
these diffractions are mainly associated with ceria species [15]. No
reflections due to Pd can be observed after Pd deposition on either
7Mn3Ce-C and 7Mn3Ce-A supports, indicating the highly dis-
persed nature of these supported Pd nanoparticles.
(111)
MnO2
• Mn5O8
∗ Mn2O3
♦ Mn3O4
Δ CeO2
Δ
(a)
Δ
Δ
Δ
Δ
CeO2
Δ
Δ
Δ
Δ
Δ
1Mn3Ce-C
7Mn3Ce-C
Δ
Δ
Δ
∗
∗
∗
♦
10Mn1Ce-C
♦
♦
∗
♦
♦
•
•
∗
∗
∗
(222)
∗
MnOx
♦
♦
∗
♦
♦
∗
♦
♦
∗
♦
♦
♦
∗
♦
∗
∗
∗
•
•
♦
Pd/7Mn3Ce-C
♦
♦
•
∗
♦
∗
20
30
40
50
60
70
Fig. 4 shows TEM images of Pd/7Mn3Ce-C and Pd/7Mn3Ce-A, as
well as Pd/CeO2 and Pd/MnOx. All these catalysts, except amor-
phous Pd/7Mn3Ce-A, consist of a large number of nanocrystallites
of metal oxide grains. According to size and morphology, two kinds
of particles can be distinguished for Pd/MnOx catalyst: round large
and nonfaceted crystals with a size of ca. 40 nm and small aggre-
gates (<7 nm), which are attributed to Mn2O3 and Mn3O4. For Pd/
7Mn3Ce-C catalyst, a new crystal phase besides Mn2O3 and
Mn3O4 can be discerned, which exhibits a rod-like shape with a
length ranging from several nanometers to several micrometers.
These nanorods were also observed by Perez-Omil et al. and attrib-
uted to MnO2 [29]. Pd/7Mn3Ce-A exhibits only a densely packed
amorphous matrix. The CeO2 nanocrystallites are of ca. 15 and
8 nm for Pd/CeO2 and Pd/7Mn3Ce-C, respectively, in good agree-
ment with the grain sizes derived from XRD (listed in Table 2). This
further confirms that adding a small amount of Ce not only in-
creases the surface area of MnOx but also enhances the dispersion
for both Ce and Mn species [27]. Therefore, the Mn and Ce oxide
phases in Pd/MnCeOx-A are homogeneously mixed at the quasi-
molecular level, forming a prevalently amorphous structure [13],
whereas in the Pd/MnCeOx-C sample, they mostly retain the typical
features of MnOx crystallites. Pd nanoclusters with small sizes and
high dispersion can be seen on the support surfaces. On Pd/
7Mn3Ce-A, only the sporadic presence of Pd nanoparticles is ob-
served, implying that Pd species may be embedded in the irregular
and amorphous matrix.
2θ, degree
(b)
1Mn3Ce-A
7Mn3Ce-A
10Mn1Ce-A
Pd/7Mn3Ce-A
20
30
40
50
60
70
2θ, degree
Fig. 3. XRD patterns of MnCeOx supports and of catalysts containing 0.5 wt.% Pd: (a)
crystalline; (b) amorphous.
Notably, the diffraction peaks attributed to the fluorite-like
structure of CeO2 shift toward higher Bragg angles in the presence
of more Mn, implying the contraction of the CeO2 unit cell. This is
further confirmed by the changes in lattice constants of MnCeOx-C
samples, as depicted in Table 2. All the calculated lattice constants
monotonically increase with increasing Mn content. Nevertheless,
all constants are smaller than those of both pure CeO2 (0.5146 nm)
and MnOx (0.9362 nm). Considering the smaller ionic radius of
Mn2+ (0.065 nm) and Mn3+ (0.066 nm) than of Ce4+ (0.094 nm),
the small lattice constants for MnCeOx-C suggest that Mn ions
may be incorporated into the ceria lattice, provoking the unit cell
H2 TPR profiles of representative MnCeOx support materials are
illustrated in Fig. 5. Pure MnOx shows two main reduction peaks
(327 and 452 °C), which are ascribed to a two-step reduction:
reduction of MnO2 and/or Mn2O3 to Mn3O4 and subsequent reduc-
tion of Mn3O4 to MnO [30]. Pure CeO2 shows a broad and indistinct
reduction peak starting from 300 °C with a maximum at 380 °C,
which is assigned to the easily reducible Ce4+ species [1,16]. When
Mn is mixed with Ce to form crystalline 7Mn3Ce-C, the two-step
Mn reduction is retained. However, the two characteristic reduc-
tion peaks shift to lower temperatures, 287 and 393 °C,

Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
39
MnO2
Pd/CeO2
Pd/7Mn3Ce-C
Mn3O4
Mn2O3
Pd/MnOx
Pd/7Mn3Ce-A
Fig. 4. TEM images of representative MnCeOx-supported Pd catalysts.
oxide species having higher oxidation states. To study the possible
effect of bulk Mn/Ce composition, the reducibility of 10Mn1Ce-C
was also examined, as depicted in Fig. 5. The 10Mn1Ce-C sample
gives a TPR profile similar to that of 7Mn3Ce-C, though both reduc-
tion peaks of Mn slightly shift to higher temperatures, further con-
firming the enhanced reducibility provided by incorporating Ce.
The reduction of amorphous 7Mn3Ce-A shows a different pat-
tern, exhibiting two poorly resolved peaks at 327 and 363 °C,
respectively. In fact, these two overlapped peaks can be assigned
to the reduction of isolated Mn4+ [15] and Mn3+ species due to
the absence of reduction at around 450 °C. The remarkable shift to-
ward lower temperature is attributed to three factors: molecular-
level Mn–Ce interdispersion, higher average oxidation states, and
strong Mn–O–Ce interaction [13]. The reduction is promoted by
highly coordinatively unsaturated Mn species and, perhaps, the
presence of neighboring Ce4+ ions, which undergo reduction to
Ce3+ species. The TPR results reveal that the mobility and availabil-
ity of lattice oxygen are enhanced by the incorporation of Ce and
the strong synergistic interaction between Mn and Ce oxides.
CeO2, well known for its oxygen storage capacity, can provide oxy-
gen to Mn, thereby increasing its oxidation state. Manganese spe-
cies possess higher oxidation states (+3 and +4) and enhanced
reducibility in amorphous MnCeOx than those in crystalline coun-
terpart (+2, +3, and +4).
380
363
CeO2
327
7Mn3Ce-A
7Mn3Ce-C
393
287
447
320
10Mn1Ce-C
452
327
MnOx
MnO2/Mn2O3 → Mn3O4 → MnO
100
200
300
400
500
temperature, oC
TPD tests, an effective method to examine the mobility of oxy-
gen species, were performed on three representative catalysts:
Pd/7Mn3Ce-A, Pd/7Mn3Ce-C, and Pd/10Mn1Ce-C. The results are
shown in Fig. 6. The indistinct and board peak at 260 °C is assigned
to the surface-active oxygen species, probably chemisorbed oxygen
Oꢀ2 , which are originated from the oxygen vacancies on the interface
of MnOx and CeO2 during the oxidation pretreatment before the
TPD tests [21]. This peak is more intense in Pd/7Mn3Ce-A than in
Pd/7Mn3Ce-C, indicating higher abundance of Oꢀ2 component over
Fig. 5. H2 TPR profiles of representative MnCeOx support materials.
respectively, indicating improved reducibility and oxygen mobility
caused by the incorporation of Ce as well as the interaction be-
tween manganese and cerium oxides [1,12]. The increased ratio
of first to second peak area for 10Mn1Ce-C compared with MnOx
qualitatively reveals the existence of a larger fraction of manganese

40
Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
amorphous MnCeOx. Pd/7Mn3Ce-C shows a slightly higher surface
692
Pd concentration (0.48%) than Pd/7Mn3Ce-A (0.31%), suggesting
that more exposed Pd active sites are available on the Pd/
7Mn3Ce-C sample, while Pd may be partially embedded or covered
in Pd/7Mn3Ce-A. Trovarelli et al. also reported that precious metals
such as Pt and Pd interact strongly with CeO2 and tend to penetrate
inside it [9].
O-
633
2-260
O
The Ce3d core-level spectra in Fig. 7a were deconvoluted using
×5
two spin–orbit components (u and
v refer to the 3d3/2 and 3d5/2
(a)
(b)
spin–orbit components, respectively) to distinguish Ce3+ and Ce4+
contributions. The spectra of Ce4+ can be resolved into six features
O2-
750
(u, u00, u000, , v00, and v000), while Ce3+ can be fitted with two features
v
(u0 and v0) [21]. The high-BE doublet u000/v000 at 916.9 and 898.5 eV is
assigned to the final state of Ce4+3d94f 0 O2p6. The doublet u00/v00 at
907.3 and 888.5 eV is attributed to the hybridization state of
Ce4+3d94f1 O2p5 and the doublets u/v at 901.6 and 883.0 eV corre-
spond to the state of Ce4+3d94f2 O2p4. The doublets uo/ o, which
v
correspond to the Ce3+3d04f1 O2p6 final state, are absent in all sam-
ples, indicating that +4 is the dominant oxidation state. The
7Mn3Ce-C support exhibits higher intensities of the doublets v000/
u000 and v00/u00 than the 7Mn3Ce-A sample even after Pd is added,
suggesting the presence of more Ce4+ species in crystalline MnCeOx
samples. The relative abundances of Ce4+ and Ce3+ are estimated
based on the fitted peak areas of the XPS Ce3d spectra and are
listed in Table 3. Despite the predominance of Ce4+, Ce3+ can be de-
tected in these samples. A higher abundance of Ce3+ exists in
7Mn3Ce-A than in 7Mn3Ce-C. Anchoring Pd onto the support
may result in more Ce3+ species at the expense of Ce4+, which is
clearly suggested by the stronger u0 and v0 peaks. The existence
of Ce3+ implies the formation of oxygen vacancies, since the crea-
tion of each Ce3+ generates two oxygen vacancies, while Ce3+ is as-
sumed to retain the normal coordination number of 8, as in the
case of Ce4+. It was suggested that the perturbation of Ce–O and/
or M–O bonds in CeO2–MOx mixed oxides would form highly dis-
ordered oxygen in the lattice, promoting oxygen transfer from
the bulk to the support surface. This enhanced oxygen mobility re-
sults in more Ce3+ species and oxygen vacancies [31]. In a similar
Ca-doped CeO2 oxide, Ce3+ surface enrichment was obtained by
both Ca M O M Ce interactions and O vacancies [32]. Recently,
the enrichment of Ce3+ by Pd doping has also been reported for a
Pd/MnCeOx sample, contributing to the interaction between Pd
and CeO2 [21]. Thus, the higher abundance of Ce3+ in amorphous
samples indicates more oxygen vacancies and enhanced oxygen
mobility.
Fig. 7b shows the experimental and fitted Mn2p3/2 spectra of
these four samples. Previous XRD, TEM, and TPR results indicated
the coexistence of different manganese oxides. The peaks around
642.2 and 640.9 eV are attributed to the presence of Mn4+ and
Mn3+ species, respectively. Surprisingly, Mn2+ is undetectable in
the fitted spectra, probably because Mn2+ may be partially incorpo-
rated into the fluorite structure of CeO2 by isomorphous substitu-
tion [33]. Since only the outmost layers can be detected by XPS,
this results in a higher oxidation state of Mn on the surface than
in the bulk. The unobvious peak at around 645.0 eV originates from
charge transfer between the outer electron shell of the ligand and
the unfilled 3d shell of Mn during the creation of the core–hole in
the photoelectron process [26]. Mn4+ is the dominant species in
7Mn3Ce-A, whereas the amount of Mn3+ becomes considerable
in 7Mn3Ce-C. Moreover, the main Mn2p3/2 component of
7Mn3Ce-C possesses a lower BE than 7Mn3Ce-A, indicating that
more electrons are transferred from Ce to Mn during co-precipita-
tion. Deposition of Pd slightly decreases the oxidation state of sur-
face manganese species, and the feature assigned to Mn3+ is
discernable in Pd/7Mn3Ce-A, which may be due to the partial
reduction of Mn species during MAPR synthesis. Pd/7Mn3Ce-A
exhibits a higher Mn BE than Pd/7Mn3Ce-C, implying that a large
(c)
100 200 300 400 500 600 700 800 900
o
Temperature, C
Fig. 6. O2 TPD profiles of representative Pd/MnCeOx catalysts: (a) Pd/7Mn3Ce-A; (b)
Pd/7Mn3Ce-C; (c) Pd/10Mn1Ce-C.
Pd/7Mn3Ce-A. Nonetheless, this peak nearly disappears in Pd/
10Mn1Ce-C, implying the lack of such interface oxygen species,
probably due to the decrease of the Mn–Ce interface as a result of
discrete contact between large MnOx and CeO2 crystallites. The
peak at ca. 630 °C, which is attributed to the chemisorbed atomic
oxygen Oꢀ, is significantly intense in Pd/7Mn3Ce-C and Pd/
10Mn1Ce-C, while it becomes a shoulder in the Pd/7Mn3Ce-A.
The board peak centered at 750 °C in both Pd/7Mn3Ce-C and Pd/
10Mn1Ce-C samples is assigned to the lattice oxygen O2ꢀ. However,
this peak drastically shifts to 692 °C in Pd/7Mn3Ce-A sample,
indicating remarkably enhanced oxygen mobility.
XPS is a useful technique for analyzing phenomena occurring at
the outermost layers of MnCeOx surfaces and evaluating the sur-
face species (Ce, Mn, O, and Pd) in nonstoichiometric surface oxi-
des of both crystalline and amorphous MnCeOx supports and
their corresponding Pd-containing catalysts. To understand the ef-
fect of support structure and Pd deposition on the surface chemis-
try, we have examined 7Mn3Ce-C and 7Mn3Ce-A supports and
their corresponding Pd-containing catalysts. The evolution of XPS
spectra of Ce3d, Mn2p, O1s, and Pd3d core levels are shown in
Fig. 7, and the surface compositions are summarized in Table 3.
7Mn3Ce-A exhibits a Mn/Ce ratio (2.13) that is close to the theoret-
ical bulk composition, indicating homogeneous dispersion of Mn
and Ce species due to simultaneous co-precipitation, whereas
7Mn3Ce-C gives a slightly lower Mn/Ce atomic ratio (1.96), sug-
gesting that the crystalline support is surface-enriched in Ce. This
Ce surface enrichment of co-precipitated MnCeOx can be attributed
either to partial coverage of Mn crystals by a thin layer of Ce-rich
oxide or to the presence of large MnOx crystals, which contribute
less significantly to the total surface area than small Ce-rich Mn–
Ce solid solution crystallites [14]. After Pd is added, the Mn/Ce ratio
remains almost unchanged for Pd/7Mn3Ce-A (2.19), while it de-
creases for Pd/7Mn3Ce-C (1.73), implying that Pd is uniformly dis-
persed in Pd/7Mn3Ce-A while it exhibits preferential anchoring on
Mn species in Pd/7Mn3Ce-C. In amorphous samples, the relative
abundance of both Mn and Ce is higher than in crystalline samples,
further verifying the good interdispersion of Mn and Ce species in

Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
41
Ce 3d
(a)
(b)
Mn 2p
3/2
u′′′
v′′′
u
u′′
u′
4+
7Mn3Ce-A
Mn
v
v
v′′
v′′
v′
7Mn3Ce-A
u′′′
v′′′
v′′′
u
u′′ u′
7Mn3Ce-C
4+
Mn
v′
v′
3+
Mn
7Mn3Ce-C
u′′′
u′′
u
u′
Pd/7Mn3Ce-A
Pd/7Mn3Ce-C
4+
Mn
v
v
v′′
v′′
3+
Pd/7Mn3Ce-A
Mn
u′′′
v′′′
u
u′′
u′
4+
Mn
v′
3+
Pd/7Mn3Ce-C
Mn
920
910
900
890
880
650
645
640
635
BE, eV
BE, eV
Pd 3d
O 1s
Pd 3d3/2
(c)
(d)
Pd 3d
5/2
Pd/7Mn3Ce-A
7Mn3Ce-A
OΙ
O
Π
Pd2+
Pd0
7Mn3Ce-C
OΙ
OΙ
OΙ
OΠ
OΠ
Pd 3d3/2
Pd 3d
5/2
Pd/7Mn3Ce-C
Pd/7Mn3Ce-A
Pd2+ Pd0
Pd/7Mn3Ce-C
OΠ
348 346 344 342 340 338 336 334 332
536
534
532
530
528
526
BE, eV
BE, eV
Fig. 7. XPS results of representative MnCeOx samples: (a) Ce3d; (b) Mn2p; (c) O1s; (d) Pd3d.
amount of Mn4+ species exists on the support surface. These
changes are quantitatively reflected in Table 3. The relative surface
concentration of Mn4+ in 7Mn3Ce-A is 100%, while only 67.14% is
detected in 7Mn3Ce-C. After Pd deposition, Mn4+ concentration de-
clines slightly to 93.04% and 64.99% in Pd/7Mn3Ce-A and Pd/
7Mn3Ce-C, respectively. Correspondingly, the AON (average oxida-
tion number) value x in MnOx decreases from 2.05 for 7Mn3Ce-A to
2.00 for Pd/7Mn3Ce-A, while it decreases from 1.84 for 7Mn3Ce-C
to 1.82 for Pd/7Mn3Ce-C.
The quantitative deconvolution of the O1s core-level spectra for
all samples seems to be difficult, due to the large number of oxygen
species related to various cerium and manganese oxides, as well as
the presence of surface oxygen species such as carbonate and hy-
droxyl groups. Particularly, the O1s BE for CeO2 may shift from
529.6 to 530.3 eV, therefore overlapping with those of various
MnOx oxides ranging from 529.3 to 530.0 eV [1]. Hence, no attempt
was made to analyze these components individually, and the O1s
spectra were fitted using only two main features (Fig. 7c): (i) the
low BE component OI assigned to the lattice oxygen (O2ꢀ), and
(ii) the high BE component OII ascribed to other surface oxygen
species (e.g., chemisorbed oxygen Oꢀ2 and Oꢀ) [13]. The amorphous
samples possess a higher BE of the OI component (529.8 eV) than
the crystalline ones (529.5 eV), indicating that an ‘‘Mn O’’ elec-
tron transfer occurs in MnCeOx with high dispersion [15]. The rel-
ative abundance of these two oxygen components is reported in
Table 3. 7Mn3Ce-A has a 55.44% OII component, which is nearly
twice the amount in 7Mn3Ce-C. After Pd deposition, Pd/7Mn3Ce-
A still has more OII component than Pd/7Mn3Ce-C, though the dif-
ference is slightly smaller. The formation of the OII component is
closely related to the nature of the MnCeOx support. Mn–O–Ce
bonds, which are formed at Mn–Ce boundary by partial incorpora-
tion of Mn ions into the cubic CeO2 structure, enhance the

42
Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
Table 3
Surface compositions calculated from XPS results of representative samples.
Entry Sample Surface atoms percentage (%) Surface Mn/Ce ratio Relative
Relative abundance of Mn4+
and Mn3+ (%)
Relative
abundance of
OI and OII/%
Relative
abundance of
abundance of
Ce3+ and Ce4+
(%)
Pd0 and Pd2+/%
Mn
21.80 10.25 67.95
17.61 8.97 73.42
Ce
O
Pd
Ce3+
Ce4+
Mn4+
100
Mn3+
0
AON of Mn OI
OII
Pd0
Pd2+
1
2
3
4
7Mn3Ce-A
7Mn3Ce-C
0
0
2.13
1.96
34.57
19.17
43.46
21.59
65.43
80.83
56.54
78.41
2.05
44.56
76.96
61.69
74.41
55.44
23.04
38.31
25.59
–
–
–
–
67.14 32.86 1.84
93.04 6.96 2.00
64.99 35.01 1.82
Pd/7Mn3Ce-A 17.81
Pd/7Mn3Ce-C 13.97
8.15 73.73 0.31 2.19
8.06 77.49 0.48 1.73
26.49
63.91
73.51
36.09
formation of oxide defects. Therefore, a large number of oxygen
vacancies on the surfaces of CeO2 and MnOx species are formed
due to a high degree of coordinative unsaturation, as suggested
by TPR data. These vacancies act as active sites for molecular oxy-
gen adsorption and activation, forming OII species [34].
The Pd3d XPS spectra in Fig. 7d consists of two asymmetric
peaks assigned to Pd3d5/2 and Pd3d3/2 core levels, which can be
fitted using two doublets. The peaks around 335.4 and 340.7 eV
are attributed to metallic Pd0m, while those around 337.3 and
342.4 eV correspond to Pd2+ species. Despite the existence of both
Pd0 and Pd2+ species in both samples, Pd/7Mn3Ce-A possesses a
higher abundance of Pd2+ than Pd/7Mn3Ce-C (73.51 and 36.09%,
respectively). The results discussed earlier have shown that
the initial reaction rate of benzyl alcohol oxidation can be repre-
sented as r0 = kCA0CM0, where CM0 is the initial concentration of
metal active sites. Since CA0 is constant at the beginning of the
reaction for both catalysts, the initial rate should correlate linearly
with the concentration of available metal active sites [35]. A large
fraction of Pd exposed to the liquid environment results in the high
availability of adsorption sites for the reactants, which may ac-
count for the high activity in the initial stage of the reaction. Here-
in, the presence of more Pd active sites accessible to the reactants
in Pd/7Mn3Ce-C than in Pd/7Mn3Ce-A, which is confirmed by XPS
results, contributes primarily to the higher initial rate. A reason-
able explanation can be derived from textural and structural differ-
ences of the two supports. The 7Mn3Ce-C support possesses a
smaller surface area than the 7Mn3Ce-A support. Nonetheless,
the support surface area may not be the most important factor in
the catalytic activity, since the Pd loading used in this study is
quite low. The enlarged surface area of 7Mn3Ce-A is primarily
due to the existence of small crystals of Mn–Ce solid solutions
rather than large crystallites of Mn oxides [14]. Therefore, the rel-
atively smaller specific surface area but larger pore diameter of
crystalline support than of amorphous support may facilitate the
anchoring and deposition of Pd active species onto the support sur-
faces during the catalyst synthesis, as well as the transport of reac-
tants during the reaction. In contrast, the inferior framework of
amorphous MnCeOx may lead to the Pd nanoparticles embedded
in the matrix, suppressing contact between Pd active sites and
reactants.
7Mn3Ce-A exhibits
a stronger surface OII component than
7Mn3Ce-C, which indicates a stronger oxidizing ability [1]. Thus,
it can be speculated that, during the catalyst synthesis process, this
active OII species may be transferred from the support surface to
the Pd nanoparticles, resulting in a strong interaction between Pd
and MnCeOx mixed oxides and the generation of Pd2+ species. A
similar result was reported by Tang et al., indicating that Ag+ spe-
cies can be more stable after silver deposition onto MnOx–CeO2
support [12]. In addition, Pd2+ in both catalysts exhibits higher
Pd 3d5/2 BEs than PdO (336.8 eV), implying high dispersion and
deficient coordination of Pd in intimate contact with MnCeOx sup-
port, as well as subsequent promoted oxygen transfer from the
support to Pd active sites [31].
Previous results demonstrate that Pd acts as the sole active ele-
ment and both Pd0 and Pd2+ species exist on crystalline and amor-
phous MnCeOx-supported Pd catalysts. To eliminate the effect of
Pd content and investigate the relationship between catalytic
4. Discussion
Our results suggest that, regardless of the structure of MnCeOx
supports (crystalline or amorphous), the catalytic performance
(conversion and selectivity) and stability of Pd catalysts can be im-
proved by combining Mn and Ce to form MnCeOx mixed oxides.
This effect is attributed to the strong synergetic interactions among
MnOx, CeO2, and Pd. However, a significant difference in the cata-
lytic behavior appears when different MnCeOx support materials
are used. Pd/7Mn3Ce-C shows higher activity than Pd7Mn3Ce-A
in the initial stage of the reaction, whereas it undergoes a more ra-
pid loss of both conversion and selectivity over a long reaction run
(Fig. 1). Based on these characterization results, this different cat-
alytic behavior can be explained by three essential factors: texture
and structure, redox properties, and surface composition.
activity and the nature of active species, the specific initial rate is
r0
introduced here: rs0
¼
content. Here, Pd/7Mn3Ce-C possesses a
pd
higher rs0 than Pd/7Mn3Ce-A (151,316 and 110,526 mmol/gPd ꢁ h,
respectively). From XPS results, Pd7Mn3Ce-C contains more sur-
face Pd0 species than Pd/7Mn3Ce-A, indicating that the initial reac-
tion rate is higher when the relative amount of Pd0/Pd2+ is greater.
This is in good agreement with the proposal that metallic Pd0 is
more active than palladium oxide in the aerobic oxidation of
benzyl alcohol [36,37]. In benzyl alcohol oxidation over Pd/Al2O3
catalysts, Pd is in the reduced state (Pd0) even in the presence of
oxygen, which is attributed to the strong reducing character of
benzyl alcohol [38]. Our previous study also showed that a Pd cat-
alyst supported on CNT consists only of metallic Pd0 species after
reaction [39]. Thus, metallic Pd0 is responsible for the high catalytic
activity. When Pd2+ is the primary component, a long induction
period was essential to transform the supported Pd2+ species to
metallic Pd nanoparticles, the true active species [40]. Here, Pd/
7Mn3Ce-A consists of larger amounts of Pd2+ than of Pd0, exhibit-
ing a low initial catalytic activity. Therefore, it can be concluded
that the initial reaction activity is intimately related to the accessi-
bility of exposed Pd active sites and the surface Pd0 concentration.
According to the reaction kinetics, the initial reaction rate is
ꢀ
ꢀ
ꢀ
ꢀ
ð1ꢀxÞꢀ
ꢀ
ꢀ
ꢀ
dCA
dCA0
dx
dt
defined as r0 ¼ ꢀ dt
¼ ꢀ
¼ CA0 t¼0, where CA0 and
dt
t¼0
t¼0
x are the initial concentration and the conversion of benzyl alcohol,
respectively. Thus, the initial rate is directly proportional to the
slope of the tangent of the conversion curves at t = 0. Fig. 1 shows
that Pd/7Mn3Ce-C has a greater slope than Pd/7Mn3Ce-A, i.e., a
higher initial reaction rate. The initial reaction rates r0 are 11.5
and 8.4 mmol/h for Pd/7Mn3Ce-C and Pd/7Mn3Ce-A, respectively.
Recently, Abad et al. have proposed a reaction mechanism in which

Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
43
Despite a low initial reaction activity, Pd/7Mn3Ce-A shows en-
hanced catalyst stability and resistance against deactivation, which
can be observed from the redox properties and the surface compo-
sition of the support. The support does not directly influence the
dehydrogenation step but affects the adsorption, activation, and
transfer of oxygen, which also has a strong effect on the activity.
Previous characterization results have shown that 7Mn3Ce-A sup-
port possesses a higher abundance of oxygen vacancies and surface
OII species (chemisorbed oxygen Oꢀ2 and Oꢀ). OII species have
stronger oxidizing ability and higher mobility, which contribute
to the enhanced oxygen transfer and redox cycle via easy transfor-
mation of the lattice oxygen O2ꢀ due to abundant oxygen vacancies
accompanied by the formation of Ce3+ species [1]. During the oxy-
gen transfer, these short-lived but highly active chemisorbed oxy-
gen species, Oꢀ2 and Oꢀ, can be activated rapidly by diffusing into
the CeO2 lattice vicinity and being converted into lattice oxygen
O2ꢀ [41]. Due to the high oxygen mobility in the 7Mn3Ce-A sup-
port, the vacancies around Pd nanoparticles after the lattice oxygen
O2ꢀ is consumed are rapidly replenished by oxygen spillover from
the bulk of the support, and the oxygen distribution on the surface
is equilibrated. On the other hand, the replenishment of oxygen in
the 7Mn3Ce-C support, which has lower oxygen mobility and few-
er oxygen vacancies, mainly comes from the gas phase. Under such
an insufficient oxygen supply, Pd active sites would be covered
with hydrogen and carbonaceous deposits and deactivated, leaving
only the sites for decarbonylation still active [42]. Thus, the gener-
ation of benzaldehyde would be suppressed and the decomposi-
tion and further oxidation of the carbonyl product would take
place, leading to lower activity and selectivity over the Pd/
7Mn3Ce-C catalyst. The presence of more oxygen vacancies and
surface OII species in the amorphous support, as well as more Mn
ions in a high oxidation state (Mn4+), also contributes significantly
to the oxygen transfer. In contrast to the Ce surface enrichment
and the remarkably low surface dispersion of Mn in crystalline
MnCeOx samples, amorphous MnCeOx support materials exhibit a
homogeneous dispersion of Mn4+ and Ce4+ ions [26]. Mn species
in higher oxidation states can facilitate the redox cycles and im-
prove the oxygen mobility and availability, resulting in enhanced
catalytic activity [1,8,43]. Moreover, the enhanced redox properties
exhibited by MnCeOx can promote the reversible sorption/desorp-
tion cycles of lattice oxygen [28]. Therefore, the mutual promotion
between redox properties and oxygen mobility leads to enhanced
catalyst stability and lower deactivation in the selective oxidation
of benzyl alcohol [16].
Catalytic results show that Pd/MnCeOx catalyst can effectively
suppress the formation of toluene, giving high benzaldehyde selec-
tivity. To clarify the origin of toluene formation and the mechanism
of benzyl alcohol oxidation, catalytic tests were conducted in N2
over Pd/7Mn3Ce-C and Pd/7Mn3Ce-A catalysts. Under such anaer-
obic conditions, both catalysts exhibit catalytic activities of 0.9%
conversion (Table 1, entries 12 and 13), implying the existence of
a dehydrogenation mechanism in which benzaldehyde formation
occurs in the absence of oxygen. Benzyl alcohol is mainly con-
verted to benzaldehyde (toluene selectivities are 2.1% and 1% for
Pd/7Mn3Ce-C and Pd/7Mn3Ce-A, respectively). The formation of
toluene is rather complex and has been extensively investigated.
Besides the hydrogenolysis of benzyl alcohol [38], toluene can also
originate from the disproportionation of benzyl alcohol, yielding an
equimolar mixture of benzaldehyde and toluene [44–46]. More-
over, Jayamani et al. also demonstrated that toluene can be formed
by the decomposition of dibenzyl ether [47]. Since no ether is de-
tected in the present study, this route can be ruled out. Therefore,
the negligible production of toluene under anaerobic conditions
clearly suggests that toluene is formed by hydrogenolysis of benzyl
alcohol rather than by disproportionation, and benzyl alcohol
oxidation follows most likely a dehydrogenation mechanism.
H2O
H
C
H
H
C
O
H
H
C
O
OH
H
H
O2
O2- + O-
H
Pd
Pd
Pd
Ce3+
O2-
Mn4+
Mn3+
Mn-Ce-O
Ce4+
Scheme 1. Overall reaction mechanism for benzyl alcohol oxidation over Pd/
MnCeOx catalysts.
Based on the conventional dehydrogenation mechanism, an
integrated alcohol oxidation mechanism over Pd/MnCeOx is pro-
posed (Scheme 1). In this mechanism, Pd acts as the sole catalytic
center where the reaction takes place, and the oxygen transfer is
completed through the support via a redox cycle. Benzyl alcohol
is adsorbed onto Pd nanoparticles, forming a metal–alkoxide inter-
mediate and adsorbed hydrogen via O–H bond cleavage. Then it is
further converted to benzaldehyde through b-hydrogen elimina-
tion as the rate-limiting step, leaving reduced Pd sites partially cov-
ered by hydrogen, i.e., Pd-hydride. The adsorbed oxygen on the
support surface dissociates into chemisorbed oxygen species O2ꢀ
and Oꢀ, diffuses into the support lattice vicinity, and subsequently
is activated at the oxygen vacancies and converted into lattice
oxygen O2ꢀ, accompanied by the transformation of Ce3+ to Ce4+
.
The lattice oxygen is further transferred through the oxide defects
in MnCeOx supports and migrates to the interface between the sup-
port and Pd active sites. Finally, it reacts with Pd-hydride to regen-
erate Pd active sites, forming H2O as product. The poor catalytic
activity over a mixture of Pd/CeO2 and Pd/MnOx catalysts (entry
11, Table 1) verifies that this oxygen migration can be facilitated
only over MnCeOx mixed oxides rather than over the physical mix-
ture. The dehydrogenation–oxidation of benzyl alcohol is com-
pleted via a synergetic redox cycle and electron transfer involving
Pd, Mn, and Ce: Pd M Mn M Ce. A similar consecutive oxygen trans-
fer mechanism was also proposed in a Ag/MnOx–CeO2 system, in
which oxygen is transferred from the oxygen reservoir of CeO2 to
activate Ag2O sites through MnOx [12]. Consequently, the syner-
getic interactions among Pd, MnOx, and CeO2 are enhanced due to
the mutual promotion of texture and structure, redox properties,
and surface composition, leading to improved catalytic activity
and higher resistance against deactivation.
5. Conclusions
Both crystalline and amorphous MnCeOx supports were success-
fully synthesized by co-precipitation and redox precipitation meth-
ods, respectively. Pd was subsequently deposited by an easy MAPR
procedure, leading to the formation of highly dispersed Pd nanoclus-
ters. MnCeOx supports remarkably enhanced both the catalytic
activity and the selectivity in the aerobic oxidation of benzyl alcohol,
compared with the single MnOx and CeO2 oxides. The highest qTOFs
were achieved over Pd/7Mn3Ce-C (15,235 hꢀ1) and Pd/7Mn3Ce-A
(14,438 hꢀ1), and the activity can be maintained over five consecu-
tive reaction runs. Pd acts as the sole active component, and the en-
hanced catalytic activity can be attributed to synergetic interactions
among Pd, MnOx, and CeO2. Good accessibility of Pd active sites and
high concentration of surface Pd0 contribute to the high initial reac-
tion rate over crystalline MnCeOx-supported Pd catalysts. Amor-
phous MnCeOx-supported Pd catalysts exhibit enhanced catalyst

44
Y. Chen et al. / Journal of Catalysis 283 (2011) 34–44
[19] M.P. Checinski, A. Bruckner, J. Radnik, A. Kockritz, Appl. Catal. A 366 (2009)
212.
stability and lower deactivation due to mutual promotion between
redox properties and oxygen mobility.
[20] T. Sato, T. Komanoya, Catal. Commun. 10 (2009) 1095.
[21] Z.Q. Zou, M. Meng, Y.Q. Zha, J. Phys. Chem. C 114 (2010) 468.
[22] Z. Guo, Y. Chen, L. Li, X. Wang, G.L. Haller, Y. Yang, J. Catal. (2010).
[23] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373.
[24] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309.
[25] H.P. Klug, L.E. Alexander, X-Ray Diffraction Procedures for Polycrystalline and
Amorphous Materials, second ed., John Wiley, New York, 1974.
[26] Y.S. Wu, Y.X. Zhang, M. Liu, Z.C.C. Ma, Catal. Today 153 (2010) 170.
[27] L.M. Shi, W. Chu, F.F. Qu, J.Y. Hu, M.M. Li, J. Rare Earth 26 (2008) 836.
[28] M. Machida, M. Uto, D. Kurogi, T. Kijima, Chem. Mater. 12 (2000) 3158.
[29] J.A. Perez-Omil, J.J. Delgado, W. Ouahbi, A.B. Hungria, N. Browning, M.A.
Cauqui, J.M. Rodriguez-Izquierdo, J.J. Calvino, J. Phys. Chem. C 114 (2010) 8981.
[30] F. Kapteijn, L. Singoredjo, A. Andreini, J.A. Moulijn, Appl. Catal. B 3 (1994) 173.
[31] L.H. Xiao, K.P. Sun, X.L. Xu, Acta Phys. Chim. Sin. 24 (2008) 2108.
[32] J.A. Rodriguez, X.Q. Wang, J.C. Hanson, G. Liu, A. Iglesias-Juez, M. Fernandez-
Garcia, J. Chem. Phys. 119 (2003) 5659.
Acknowledgment
Funding from the Singapore Agency for Science, Technology and
Research (A⁄STAR), SERC Grant 102 101 0020, in support of this
project is gratefully acknowledged.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.06.021.
[33] W.-J. Hong, S. Iwamoto, S. Hosokawa, K. Wada, H. Kanai, M. Inoue, J. Catal. 277
(2011) 208.
References
[34] S. Imamura, Y. Okumura, T. Nishio, K. Utani, Y. Matsumura, Ind. Eng. Chem.
Res. 37 (1998) 1136.
[35] A. Abad, A. Corma, H. Garcia, Chem. Eur. J. 14 (2008) 212.
[36] J.D. Grunwaldt, M. Caravati, A. Baiker, J. Phys. Chem. B 110 (2006) 25586.
[37] C. Mondelli, D. Ferri, J.D. Grunwaldt, F. Krumeich, S. Mangold, R. Psaro, A.
Baiker, J. Catal. 252 (2007) 77.
[38] C. Keresszegi, D. Ferri, T. Mallat, A. Baiker, J. Phys. Chem. B 109 (2005) 958.
[39] Y. Chen, L. Bai, C. Zhou, J.-M. Lee, Y. Yang, Chem. Commun. (2011), doi:10.1039/
c1cc11643f.
[40] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, J. Am. Chem. Soc. 126
(2004) 10657.
[1] H. Chen, A. Sayari, A. Adnot, F. Larachi, Appl. Catal. B 32 (2001) 195.
[2] Z.Y. Ding, L.X. Li, D. Wade, E.F. Gloyna, Ind. Eng. Chem. Res. 37 (1998) 1707.
[3] S. Aki, M.A. Abraham, Ind. Eng. Chem. Res. 38 (1999) 358.
[4] D. Delimaris, T. Ioannides, Appl. Catal. B 84 (2008) 303.
[5] H. Li, X.L. Tang, H.H. Yi, L.L. Yu, J. Rare Earth 28 (2010) 64.
[6] G.S. Qi, R.T. Yang, J. Catal. 217 (2003) 434.
[7] F. Arena, G. Trunfio, J. Negro, L. Spadaro, Mater. Res. Bull. 43 (2008) 539.
[8] X.F. Tang, Y.G. Li, X.M. Huang, Y.D. Xu, H.Q. Zhu, J.G. Wang, W.J. Shen, Appl.
Catal. B 62 (2006) 265.
[9] A. Trovarelli, Catal. Rev. Sci. Eng. 38 (1996) 439.
[10] F. Arena, G. Trunfio, J. Negro, L. Spadaro, Appl. Catal. B 85 (2008) 40.
[11] F. Eigenmann, M. Maciejewski, A. Baiker, Appl. Catal. B 62 (2006) 311.
[12] X.F. Tang, J.L. Chen, Y.G. Li, Y. Li, Y.D. Xu, W.J. Shen, Chem. Eng. J. 118 (2006)
119.
[41] C. Bozo, N. Guilhaume, J.-M. Herrmann, J. Catal. 203 (2001) 393.
[42] C. Keresszegi, D. Ferri, T. Mallat, A. Baiker, J. Catal. 234 (2005) 64.
[43] F. Arena, J. Negro, A. Parmaliana, L. Spadaro, G. Trunfio, Ind. Eng. Chem. Res. 46
(2007) 6724.
[13] F. Arena, G. Trunfio, B. Fazio, J. Negro, L. Spadaro, J. Phys. Chem. C 113 (2009)
2822.
[44] S. Meenakshisundaram, E. Nowicka, P.J. Miedziak, G.L. Brett, R.L. Jenkins, N.
Dimitratos, S.H. Taylor, D.W. Knight, D. Bethell, G.J. Hutchings, Faraday
Discuss. 145 (2010) 341.
[45] G. Kovtun, T. Kameneva, S. Hladyi, M. Starchevsky, Y. Pazdersky, I. Stolarov, M.
Vargaftik, I. Moiseev, Adv. Syn. Catal. 344 (2002) 957.
[46] S.S. Hladyi, M.K. Starchevsky, Y.A. Pazdersky, M.N. Vargaftik, I. Moiseev I and II,
Mendeleev Commun. (2002) 45.
[47] M. Jayamani, C.N. Pillai, J. Catal. 82 (1983) 485.
[14] G. Blanco, M.A. Cauqui, J.J. Delgado, A. Galtayries, J.A. Perez-Omil, J.M.
Rodriguez-Izquierdo, Surf. Interface Anal. 36 (2004) 752.
[15] F. Arena, G. Trunfio, J. Negro, B. Fazio, L. Spadaro, Chem. Mater. 19 (2007) 2269.
[16] Q.H. Tang, C.M. Wu, X.O. Huang, Y.H. Yang, Chin. J. Catal. 30 (2009) 207.
[17] A.F.J. Santiago, J.F. Sousa, R.C. Guedes, C.E.M. Jeronimo, M. Benachour, J.
Hazard. Mater. 138 (2006) 325.
[18] S. Hamoudi, A. Sayari, K. Belkacemi, L. Bonneviot, F. Larachi, Catal. Today 62
(2000) 379.