Selective side-chain oxidation of alkyl aromatic compounds catalyzed by cerium modified silver catalysts

Article abstract of DOI:10.1016/j.molcata.2010.08.001

Silver supported on silica effectively catalyzes the aerobic side-chain oxidation of alkyl aromatic compounds under solvent-free conditions. Toluene, p-xylene, ethylbenzene and cumene were investigated as model substrates. Typically, the reaction was performed at ambient pressure; only for toluene an elevated pressure was required. Carboxylic acids, such as benzoic acid or p-toluic acid, additionally increased the reaction rate while CeO₂ could act both as a promoter and an inhibitor depending on the substrate and the reaction conditions. Silver catalysts were prepared both by standard impregnation and flame spray pyrolysis. Addition of a Ce precursor to the FSP catalyst resulted in significantly smaller silver particles. Ce-doped FSP catalysts in general exhibited a superior catalytic performance with TONs up to 2000 except for cumene oxidation that appeared to proceed mainly by homogeneous catalysis. In addition, flame-made catalysts were more stable against silver leaching compared to the impregnated catalysts. The structure of the silver catalysts was studied in detail both by X-ray absorption spectroscopy and transmission electron microscopy suggesting metallic silver to be required for catalytic activity. Catalytic studies point to a radical mechanism which differs depending on the type of substrate.

Full text of DOI:10.1016/j.molcata.2010.08.001

Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Selective side-chain oxidation of alkyl aromatic compounds catalyzed by cerium
modified silver catalysts
a
b
c
d
b
Matthias J. Beier , Bjoern Schimmoeller , Thomas W. Hansen , Jens E.T. Andersen , Sotiris E. Pratsinis ,
a,e,∗
Jan-Dierk Grunwaldt
a
Department of Chemical and Biochemical Engineering, Technical University of Denmark, Søltofts Plads Building 229, DK-2800 Kgs. Lyngby, Denmark
Particle Technology Laboratory, Institute of Process Engineering, Department of Mechanical and Process Engineering, ETH Zürich, Sonneggstrasse 3, CH-8092 Zürich, Switzerland
Center for Electron Nanoscopy, Technical University of Denmark, Fysikvej Building 307, DK-2800 Kgs. Lyngby, Denmark
Department of Chemistry, Technical University of Denmark, Kemitorvet Building 207, DK-2800 Kgs. Lyngby, Denmark
b
c
d
e
Institute for Chemical Technology and Polymer Chemistry, Karlsruhe Institute of Technology (KIT), Engesserstr. 20, D-76131 Karlsruhe, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 May 2010
Received in revised form 8 July 2010
Accepted 2 August 2010
Available online 11 August 2010
Silver supported on silica effectively catalyzes the aerobic side-chain oxidation of alkyl aromatic com-
pounds under solvent-free conditions. Toluene, p-xylene, ethylbenzene and cumene were investigated as
model substrates. Typically, the reaction was performed at ambient pressure; only for toluene an elevated
pressure was required. Carboxylic acids, such as benzoic acid or p-toluic acid, additionally increased the
reaction rate while CeO2 could act both as a promoter and an inhibitor depending on the substrate and
the reaction conditions. Silver catalysts were prepared both by standard impregnation and flame spray
pyrolysis. Addition of a Ce precursor to the FSP catalyst resulted in significantly smaller silver parti-
cles. Ce-doped FSP catalysts in general exhibited a superior catalytic performance with TONs up to 2000
except for cumene oxidation that appeared to proceed mainly by homogeneous catalysis. In addition,
flame-made catalysts were more stable against silver leaching compared to the impregnated catalysts.
The structure of the silver catalysts was studied in detail both by X-ray absorption spectroscopy and
transmission electron microscopy suggesting metallic silver to be required for catalytic activity. Catalytic
studies point to a radical mechanism which differs depending on the type of substrate.
Keywords:
Selective aerobic oxidation
Silver catalysis
Flame spray pyrolysis
p-Xylene
Homogeneous vs. heterogeneous catalysis
©
2010 Elsevier B.V. All rights reserved.
1
. Introduction
metal promoting the chain initiation and the decomposition of
peroxo species [4,9].
Selective catalytic oxidation by inexpensive reagents such as
Active catalysts for the selective side-chain oxidation of alkyl
aromatic compounds are reported mainly for liquid phase oxida-
tions. Only a few examples are reported in the gas phase [10–12]
or in supercritical media [13,14]. The selective liquid phase oxida-
tion can be conducted both homogeneously and heterogeneously
involving mostly metals with two or more stable oxidation states
typically under acidic conditions. Most catalytic systems are based
on Co often modified with Mn and a Br source [9,15,16] but other
examples based on e.g. Pt [17] and Cu [18,19] are also known.
Typical oxidizing agents are organic peroxides, hydrogen peroxide,
and oxygen offering a cost-effective route to important industrial
intermediates such as benzoic acid, acetophenone, benzophenone,
toluic acid, etc. [5]. Interestingly, ring and side-chain oxidation can
be controlled by the choice of oxidant [20,21]. Due to the relative
inertness of the corresponding C–H bonds, the reactions are often
performed at high pressure and in an additional solvent. Avoiding
elevated pressure and working under neat conditions would be the
next step to render the oxidation process more cost-efficient.
Based on the homogeneous Co/Mn/Br system used in the Amoco
process many heterogeneous catalysts employ Co as a catalyti-
oxygen, hydrogen peroxide or t-butyl hydroperoxide is a viable way
of introducing functional groups and thereby upgrading saturated
and unsaturated hydrocarbons [1–3]. Finding catalytic systems for
these oxidation reactions is of significant industrial importance [4],
e.g. for p-xylene and cyclohexane oxidation to terephthalic acid
and adipinic acid, respectively, which are monomers for commod-
ity polymers [5]. The high industrial impact is also reflected by the
large amount of related patent literature (e.g. [6–8]). Nevertheless,
these catalytic oxidation reactions still require harsh conditions in
industry [4]. As an example, a very effective but highly corrosive
and environmentally harmful homogeneous Co/Mn/Br system is
employed for p-xylene oxidation [9] in liquid phase using acetic
acid as solvent which is partly decomposed during reaction. The
oxidation follows a radical chain mechanism with the transition
∗ Corresponding author at: Karlsruhe Institute of Technology (KIT), Institute for
Chemical Technology and Polymer Chemistry, Engesserstr. 20, 76131 Karlsruhe,
Germany. Tel.: +49 721 608 2121; fax: +49 721 608 4820.
E-mail address: grunwaldt@kit.edu (J.-D. Grunwaldt).
1
381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.08.001

M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
41
Table 1
Overview over the silver catalysts prepared by flame spray pyrolysis and impregnation.
2
a
CeO2 crystal sizeb (nm)
Ag crystal sizeb (nm)
n.d.c
Catalyst
Ag (wt.%)
CeO2 (wt.%)
SiO2 (wt.%)
SSA (m /g)
Entry
FSP
1
1
1
1
%Ag/SiO2
1
1
1
1
10
0
10
30
50
0
100
90
70
50
100
293
273
194
152
128
–
1
2
3
4
5
FSP
FSP
FSP
c
%Ag/10%CeO2–SiO2
%Ag/30%CeO2–SiO2
%Ag/50%CeO2–SiO2
3.3
5.8
8.3
–
n.d.
n.d.c
n.d.c
43
imp
1
0%Ag/SiO2-900
a
Determined by nitrogen adsorption.
Determined by XRD.
Not detectable.
b
c
cally active metal. We were interested in the potential of silver
as an active material as we showed recently that silver is a potent
oxidation catalyst for the selective liquid phase oxidation of alco-
hols when used in combination with CeO2 nanoparticles [22]. The
impregnation route for catalyst synthesis afforded a broad silver
particle size distribution with particles ranging from a few nm up
to 50 nm. Flame spray pyrolysis (FSP) has extensively been studied
for the preparation of different metal oxide catalysts [23,24] as well
as supported noble metal catalysts often affording high metal dis-
persions with particle sizes of less than 5 nm on the support [25].
In this way, a good mixing of the oxide particles can be achieved.
Therefore, we prepared in the present study the Ag/SiO –CeO cat-
on a glass microfiber filter (Whatman GF/D, 257 mm in diameter).
FSP
FSP
The catalysts are denoted as
1%Ag/SiO2
1%Ag/10%CeO –SiO ,
2 2
FSP
1%Ag/30%CeO –SiO2 and FSP1%Ag/50%CeO –SiO , respectively.
2
2
2
Details are given in Table 1.
2.2. Catalyst testing
Catalyst tests were performed in
a three-necked flask
equipped with a gas inlet and a reflux condenser with p-xylene
(Sigma–Aldrich >99%), ethylbenzene (Fluka, >99%) or cumene
(Fluka, 98%) as substrates with and without CeO2 nanoparticles
(Aldrich, <25 nm), and benzoic acid (Sigma–Aldrich, 99%) or p-toluic
2
2
alyst system not only by impregnation but also by single-step flame
spray pyrolysis and investigated its potential for synthesis of cat-
alysts for the oxidation of aryl aromatic compounds. Comparing
acid (Aldrich, 98%). In a typical experiment the flask was charged
with 122 mmol of alkyl aromatic compound and 100 mg biphenyl
(Sigma–Aldrich, 99.5%) as internal standard. Biphenyl was unreac-
Ag/CeO –SiO catalysts made by these two routes, the FSP-made
2
2
tive under the applied reaction conditions, i.e. its presence did not
give any additional peaks in the gas chromatogram. The flask was
equipped with a magnetic stirring bar and placed in an oil bath
catalysts exhibited an improved performance even at significantly
lower silver loadings for the oxidation of toluene, p-xylene, ethyl-
benzene and cumene. The structure was further elucidated by
X-ray absorption spectroscopy (XAS) and transmission electron
microscopy (TEM).
◦
◦
heated to 140 C which led to a temperature of the reaction mix-
ture of ca. 135 C. The mixture was saturated with oxygen (Linde,
9
9.95%) via the gas inlet and equilibrated for 1.5 h to ensure a
uniform standard and substrate distribution. Benzoic acid (0.45 g,
3.7 mmol, 3.0 mol% with respect to the hydrocarbon; for ethylben-
zene and cumene oxidation) and p-toluic acid (0.50 g, 3.7 mmol,
3.0 mol% with respect to the hydrocarbon; for p-xylene oxidation),
respectively, were added. After 5 min a sample was taken for GC
analysis and the reaction was started by adding 100 mg catalyst and
optionally 50 mg CeO2 (Aldrich, <25 nm) for catalysts not already
containing CeO2.
2
. Experimental
2.1. Catalyst synthesis
The impregnated catalyst was prepared by adding 25.0 mL of
an aqueous AgNO3 (Fluka, >99.5%) solution (31 mg/mL) to 5.00 g of
SiO2 (Aerosil 200, Degussa, 200 m /g) resulting in a Ag:SiO2 ratio
2
of 1:10. The mixture was stirred with a glass rod and sonicated for
Analysis was carried out after 1, 2 and 3 h, respectively, by
means of an Agilent 6890 N gas chromatograph equipped with a
7683 B series autosampler, a flame ionization detector and a HP-5
column using biphenyl as internal standard. The following temper-
ature programs were used for product analysis: inlet temperature
◦
1
h. After drying at 80 C for 24 h the catalyst was calcined in air
◦
at 900 C for 1 h in a muffle oven controlled by a Eurotherm 904
controller. The catalyst is denoted as
imp
10%Ag/SiO -900.
2
For catalysts made by flame spray pyrolysis, silver
nitrate (Fluka, >99.5%) dissolved in ethanol (EtOH) and
diethyleneglycol–monobutylether (DEGME, Fluka, >98%; 1:1,
Ag 0.5 M), hexamethylsiloxane (HMDSO, Fluka >98%) and cerium-
◦
◦
◦
◦
250 C; oven: 110 C, hold for 10 min; ramp to 120 C (5 C/min);
◦
◦
◦
ramp to 300 C (40 C/min); 300 C, 2 min; the analysis of hydroper-
oxides was performed at lower temperatures to ensure minimal
◦
2
-ethylhexanoate (40% Ce w/v, Strem) were used as silver, silica
decomposition during analysis: inlet temperature 100 C; oven:
◦
◦
◦
◦
and ceria precursors, respectively. The appropriate amounts of the
above mentioned precursor solutions were mixed with solvents,
i.e. DEGME (Fluka, >98%) and 2-ethylhexanoic acid (Aldrich, >98%)
while keeping the solvent composition constant at 1:1 by volume.
The total support metal (Si + Ce) concentration was kept constant
80 C, hold for 10 min; ramp to 120 C (5 C/min); ramp to 300 C
◦
◦
(40 C/min); 300 C, 2 min. Products were identified by GC–MS.
Individual response factors were obtained from commercial com-
pounds, i.e. cumene hydroperoxide (Aldrich, 80% in cumene),
p-methyl benzyl alcohol (Aldrich, 98%), p-methyl benzaldehyde
(Aldrich, 97%), benzaldehyde (Aldrich, >99%), benzyl alcohol (SAFC,
>99%), 2-phenyl-2-propanol (Aldrich, 97%), acetophenone (Aldrich,
99%) and 1-phenyl ethanol (Aldrich, 98%). Ethylbenzene hydroper-
oxide was synthesized according to Ref. [27] and the response factor
determined from a mixture of 1-phenylethanol and ethylbenzene
hydroperoxide.
at 0.6 M in these solutions. The nominal CeO weight fraction in the
2
product powder ranged from 0 to 50 wt.%. The silver precursor was
added accordingly to reach a nominal 1 wt.% Ag loading in the final
powder product. The Ag/CeO –SiO2 powders were produced in a
2
laboratory scale FSP reactor as described elsewhere [26]. The liquid
and dispersion gas feed rate were kept constant at 5 mL/min and
of 5 L/min, respectively. The pressure drop above the nozzle was
adjusted to 1.7 bar and a sheath gas flow of 5 L/min O2 was applied
through a metal ring (11 mm i.d., 18 mm o.d.) with 32 holes (0.8 mm
i.d.) to ensure complete combustion of the precursor. The powders
were collected with the aid of a vacuum pump (Busch SV 1025 B)
Since starting material was lost during the reaction despite using
reflux condensers, yields rather than conversions and selectivities
are presented from which TONs are calculated on the basis of the
total amount of silver. In general, results were reproducible within
an error margin of 5–10%. Toluene required higher temperatures

4
2
M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
than its boiling point, thus, toluene oxidation was done in an auto-
clave view cell (NWA GmbH, Lörrach). The autoclave was charged
with 13.0 mL toluene (122 mmol), 100 mg biphenyl, 0.45 g benzoic
acid, 50 mg CeO2 and 100 mg imp10%Ag/SiO -900. A sample for GC
2
analysis was taken, the autoclave was sealed and the 50 mL resid-
ual volume charged with air (10 bar, 4.2 mmol O ). The autoclave
2
was heated up while stirring. After 1 h at the desired temperature
the autoclave was cooled and opened after carefully releasing the
overpressure followed by sampling for GC analysis.
Scheme 1. Oxidation of p-xylene with molecular oxygen catalyzed by Ag/SiO2 in
the presence of ceria and a carboxylic acid.
3. Results and discussion
2.3. Catalyst characterization
3.1. Catalytic side-chain oxidation of p-xylene by Ag/SiO2
X-ray absorption near edge structure (XANES)/extended X-ray
prepared by impregnation
absorption fine structure (EXAFS): X-ray absorption spectroscopy
XAS) was measured at beamline X1 at the Hamburger Syn-
(
The oxidation of p-xylene with an impregnated silica sup-
ported Ag catalyst ( 10%Ag/SiO2-900) in the absence of any
imp
chrotronstrahlungslabor (HASYLAB) at the Deutsche Elektronen-
Synchrotron (DESY, Hamburg). XAS spectra for silver were recorded
in step-scanning mode around the Ag–K-edge (E = 25.514 keV) with
a Si(3 1 1) double crystal monochromator in transmission mode and
a silver foil for energy calibration. EXAFS spectra were processed
by energy calibration, background subtraction, deglitching, nor-
solvent resulted in the formation of p-methylbenzyl alcohol, p-
◦
methylbenzaldehyde and p-toluic acid at about 3% yield at 140 C
after 3 h (Scheme 1 and Fig. 1) which corresponds to a TON of
43 based on the overall amount of silver. Note, that at this point
TONs only serve as a performance comparison being microscopi-
cally seen not relevant as the reaction is presumably based on a
radical autoxidation mechanism. The addition of CeO2 nanoparti-
cles and a carboxylic acid more than doubled the overall catalytic
performance resulting in a combined yield of 7% after 3 h (TON
100). No difference between benzoic acid and p-toluic acid could
be found with respect to the catalytic performance (Table 2, entry
1 and 2). As underlined by reaction Scheme 1 the second methyl
group remained untouched due to the deactivating effect of the
electron withdrawing substituent. Some p-xylene hydroperoxide
as an intermediate reaction product was also observed (less than
0.5%) but no pertoluic acid. Note that the reaction proceeded in
the absence of light in contrast to other examples found in litera-
ture [31,32]. Fig. 2 depicts the typical development of the products
during the reaction. After a short induction period p-methylbenzyl
alcohol and p-methylbenzaldeyhde were formed at approximately
the same rate. Hence, both compounds are presumably formed
simultaneously and hardly by successive oxidation. The formation
of p-toluic acid was first detected after an induction period. Thus,
p-toluic acid was formed via one (or more) of the intermediates,
e.g. by reaction between aldehyde and peroxide.
3
malization and Fourier-transformed after k -weigthing using the
WinXAS 3.1 software [28]. EXAFS data fitting was performed in R-
space taking only single scattering paths within the first Ag–Ag or
Ag–O shells into account. EXAFS data were fitted with a damping
factor obtained from silver foil of 0.85. Scattering amplitudes and
phase shifts were calculated with the FEFF 7.0 code [29].
X-ray diffraction (XRD): The XRD patterns were collected with an
X’Pert PRO Diffractometer (PANalytical) with a Cu K␣ X-ray source
operated at 45 kV and 40 mA equipped with a Ni filter and a slit.
◦
Diffractograms were recorded between 2ꢀ (Cu K␣) = 20–90 with
◦
a step width of 0.00164 . The crystallite size was estimated from
Rietveld refinement using a Windows-adapted version of the LHMP
program [30].
Inductively coupled plasma mass spectrometry (ICP-MS): The con-
centrations of cerium and silver in solution were measured by
ICP-MS (Perkin-Elmer Elan 6000) equipped with a cross-flow neb-
ulizer. The plasma Ar flow was set to 13 L/min at 800 W RF power
and the nebulizer Ar flow was set to 0.80 L/min. The sample solu-
tion was injected at a flow rate of 3.0 mL/min. Quantification of
Ce and Ag was performed by using multielement ICP-MS standard
solutions (Fluka). Samples for ICP-MS were prepared via filtering of
a hot solution obtained after 3 h reaction time through preheated
celite. An aliquot of 1.0 mL of the filtrate was taken. Volatile com-
Silver has a complex interaction with oxygen influencing the
catalytic properties [33,34]. Different silver-oxygen species can be
formed in an oxygen containing atmosphere depending on the cal-
cination temperature [35,36]. Therefore, we studied the effect of
◦
pounds were removed at 120 C in vacuo overnight. The residual
was treated with 3.5 mL concentrated HNO3 (Merck, p.a.), 0.5 mL
HCl (Merck, p.a.) followed by 2 mL double distilled water and diges-
tion in an Anton Paar Microwave Digestion System Multiwave
3
000. The solution was neutralized with ammonium hydroxide
solution (Sigma–Aldrich, 28–30%), filtered, and diluted with dis-
tilled water. The overall dilution ratio was 1:500.
Gas adsorption: The specific surface area (SSA) of the cata-
lyst particles was determined by N2 adsorption at 77 K using
the BET method (Micromeritics Tristar 3000, 5-point isotherm,
0
.05 < p/p < 0.25).
0
Transmission electron microscopy (TEM): TEM images were
acquired on an FEI Tecnai T20 microscope equipped with a W-
filament as electron source operated at 200 kV. HRTEM images
were taken on an FEI Titan 80–300 microscope operated at 300 kV.
Catalyst samples were dispersed in dry form on lacey carbon film
supported on Cu grids. Sample agglomerates overlapping the holes
in the support film were imaged using a Gatan US1000 CCD camera.
Crystal lattice spacings and particle sizes were analyzed using the
Digital Micrograph version GMS 1.8 software package from Gatan
Inc.
Fig. 1. Influence of carboxylic acid (p-toluic acid) and CeO2 as additives on the
oxidation of p-xylene. Reaction conditions: 122 mmol p-xylene, optionally 3 mol% p-
imp
toluic acid and/or 50 mg CeO2, 100 mg biphenyl, 100 mg
10%Ag/SiO2-900, oxygen
◦
atmosphere, 140 C, 3 h reaction time.

M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
43
Table 2
Oxidation of p-xylene with silver catalysts in the presence of CeO2 and carboxylic acid. Reaction conditions: 122 mmol p-xylene, 3 mol% p-toluic acid, 100 mg biphenyl,
◦
1
00 mg catalyst, oxygen atmosphere, 140 C, 3 h reaction time.
Catalyst
Yields (%)
TONa
Entry
Combined
imp
0%Ag/SiO2-900b
1
1
2.1
2.0
0.4
3.0
2.3
2.8
2.4
0.2
2.1
2.0
0.6
2.6
2.1
2.5
2.2
0.3
2.8
2.8
0.2
7.8
3.3
5.2
3.8
0.3
7.0
6.8
1.2
13.4
7.7
10.5
8.4
100
100
160
1800
1000
1400
1100
1
2
3
4
5
6
7
8
imp
FSP
c
0%Ag/SiO2-900
1
%Ag/SiO2 b
FSP
FSP
FSP
FSP
1
1
1
1
%Ag/10%CeO2–SiO2
%Ag/10%CeO2–SiO2 d
%Ag/30%CeO2–SiO2
%Ag/50%CeO2–SiO2
b
n.d.c
SiO2
0.8
a
Based on the number of product molecules with respect to the total amount of silver during one experiment.
b
c
5
3
1
0 mg CeO2 added.
mol% benzoic acid used instead of p-toluic acid.
st re-use.
d
the calcination temperature on the catalytic performance (Fig. 3).
Within the temperature range investigated, the highest calcination
◦
temperature, i.e. 900 C resulted in the most active silver catalyst
with a significant increase in catalytic activity between 600 and
◦
7
00 C. After the high temperature treatment rather large silver
particles with an average crystal size of 43 nm were observed from
XRD analysis (not shown). These most active catalysts were used
during further parametric studies. Extended X-ray absorption spec-
troscopy (EXAFS) revealed that silver was metallic (Fig. 4) with no
indication of oxidized silver suggesting metallic silver to be the
active species. The 1st shell silver–silver coordination number of
1
2 obtained from EXAFS data fitting is in line with metallic sil-
ver particles >5 nm and corroborates the XRD data. Calcination at
◦
5
00 C was previously investigated [22] and resulted in a lower
coordination number of 7.9 despite several large silver particles
observed.
As described before, CeO2 nanoparticles and carboxylic acid
addition generally had a beneficial effect on the catalytic per-
formance yielding more of the product acid. Also initial addition
of the carboxylic acid only had a beneficial effect. Adding, how-
Fig. 3. Yields of p-methylbenzyl alcohol, p-methylbenzaldehyde and p-toluic acid
imp
obtained during the oxidation of p-xylene with
at temperatures between 300 and 900 C in the presence of ceria and carboxylic acid.
10%Ag/SiO2 calcined for 1 h in air
◦
Reaction conditions: 122 mmol p-xylene, 3 mol% p-toluenic acid, 100 mg biphenyl,
imp
1
1
00 mg
10%Ag/SiO2 calcined in air as indicated, 50 mg CeO2, oxygen atmosphere,
◦
40 C, 3 h reaction time.
Fig. 4. EXAFS spectrum of imp10%Ag/SiO2 calcined at 500 C (a), at 900 C (b) and
silver foil (c) recorded at the Ag K-edge. Note the smaller EXAFS oscillation amplitude
in (a) compared to (b). Inset: Experimental (solid line) and 1st shell EXAFS data fitted
◦
◦
Fig. 2. Formation of p-methylbenzyl alcohol (᭹), p-methylbenzaldehyde (ꢀ) and
FSP
p-toluic acid (ꢁ) during the oxidation of p-xylene with 1%Ag/10%CeO2–SiO2. Reac-
tion conditions: 122 mmol p-xylene, 3 mol% benzoic acid, 100 mg biphenyl, 100 mg
imp
˚ −1
.
(
dashed line) FT EXAFS function of
10%Ag/SiO2-900 from k = 2–15 A
FSP
◦
%Ag/10%CeO2–SiO2, oxygen atmosphere, 140 C.
1

4
4
M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
ever, CeO2 nanoparticles alone to imp10%Ag/SiO -900 resulted in
the formation of hardly any reaction products (Fig. 1). This pin-
points to a radical autoxidation mechanism which is suppressed
by CeO2 nanoparticles known to have radical scavenger properties
tend to form bimodal Ag distributions. Large Ag particles are dis-
cussed to form by droplet-to-particle conversion while fine Ag
2
particles come from gas-to-particle conversion [26]. TEM images
FSP
of
1%Ag/10%CeO2–SiO2 (Fig. 5c) showed distinct CeO2 nanopar-
[
37,38]. This is an intriguing feature of ceria since it both promotes
ticles with the size in the range as indicated by XRD (3 nm, Table 1).
Similar morphologies of segregated crystal domains have been
observed previously for FSP-made Ta2O5/SiO2 [43] and TiO2/SiO2
made in diffusion flames [44]. The presence of ceria allowed no
clear determination of the silver particle size by STEM analysis
non-radical oxidations [22,39,40] and is capable of increasing
the selectivity by suppressing reactive radical species. Further
evidence for a radical mechanism was found by suppression of
p-xylene oxidation by either 2,6-di(tert-butyl)phenol or 2,2,6,6-
tetramethylpiperodine-1-oxyl (TEMPO) which can act as a radical
scavenger [41]. When investigating the Ag/SiO –CeO –carboxylic
but no large Ag particles with a size even close to that found
FSP
for
1%Ag/SiO2 were observed (Fig. 5d) indicating that the pres-
2
2
acid system with the impregnated catalyst for leaching, the cerium
concentration was below the limit of detection of 0.1 mg/L. In addi-
tion, a reaction mixture treated with CeO2 and carboxylic acid
for several hours did not show the promoting effect after filtra-
tion of CeO2 and addition of the silver catalyst. Hence there is no
indication for soluble cerium species to be responsible for the pro-
moting effect. It must be added, however, that soluble Ce3 (from
Ce(III)ethylhexanoate) indeed promoted the reaction at a high con-
centration of 10 mg/L also in the absence of carboxylic acid to result
in an overall product yield of 9.5% or a TON of 140 with respect to
silver. In contrast to the Ce concentration being below the limit
of detection, a high silver concentration of 83 mg/L was found and
raises the question whether silver promotes the oxidation reac-
tion homogeneously or heterogeneously. Further tests showed that
under non-reactive conditions, i.e. in the absence of carboxylic
ence of ceria might have hindered the formation of such large Ag
nanoparticles. The stabilizing effect of CeO2 on the silver nanopar-
ticles during synthesis might also be one cause for the pronounced
difference in catalytic activity between doped and undoped cata-
lysts.
The Ag K-edge EXAFS spectrum of the impregnated catalyst
showed mainly Ag–Ag backscattering contributions due to the
metallic state of silver. In addition to a Ag–Ag shell, the FSP
+
FSP
FSP
catalysts
1%Ag/SiO2 and
1%Ag/10%CeO2–SiO2 also featured
a Ag–O shell within the first coordination sphere (Fig. 6). The
FSP
as-prepared
1%Ag/SiO2 catalyst showed both contributions of
oxidized and reduced silver. The fitted silver–silver distance of
2.84
A˚ is close to that for bulk metallic silver (2.89 A˚ ) while the
silver–oxygen distance was fitted with 2.29
A˚ (Table 3, entry 2).
Similar distances of 2.32
A˚ (Ag–O) and 2.81 A˚ (Ag–Ag) were also
FSP
acid but with CeO , no silver leaching was observed (Ag concen-
found for the ceria-doped catalyst
1%Ag/10%CeO2–SiO2 (entry
2
tration <1 mg/L). Hence, leaching is caused by the carboxylic acid
and/or the reaction products. Removal of the catalyst by hot filtra-
tion almost completely stopped the catalytic reaction. Furthermore,
the reaction was performed with silver benzoate as a soluble sil-
3). The low Ag–Ag distance especially in the latter case might be
explained by the lattice contraction of metallic Ag nanoparticles
in the low nm-range [45]. The lower Ag–Ag coordination number
FSP
FSP
of
1%Ag/10%CeO2–SiO2 compared to
1%Ag/SiO2 is qualita-
ver source instead of imp10%Ag/SiO -900 (Ag silver concentration
tively in accordance with the smaller particle size found for the
doped catalyst with TEM. However, due to the Ag–O contribu-
tion no further quantitative conclusions on the particle size can
2
of 300 mg/L). Despite the higher silver concentration compared to
the leaching experiment, an overall yield of only 1.6% was obtained
imp
˚
compared to 7.0% with
10%Ag/SiO -900.
be drawn. The Ag–O distance in the range of 2.3 A significantly
2
differs from that found in Ag O being 2.047 A˚ from XRD [46] or
2
2
.07 A˚ from EXAFS analysis (entry 5). Ag–O bond lengths between
3.2. Comparison of impregnated catalysts with flame spray
2.2 and 2.6 A˚ are typical for silver silicates [47–50]. The high mixing
pyrolysis (FSP) derived catalysts
and fast cooling in the flame could facilitate the atomic mixing of
Ag and Si and so amorphous silver silicate formed under FSP con-
Ag/SiO has been prepared previously by FSP resulting in rather
ditions might indeed account for the observed Ag–O shell. Spent
2
FSP
large silver particles [42]. We prepared a number of CeO –SiO sup-
1%Ag/10%CeO –SiO2 (entry 4) has a lower but still appreciable
2
2
2
ported silver catalysts without and with 10, 30 and 50 wt.% CeO2
Table 1). An increasing fraction of CeO decreased the surface area
from 293 m /g without CeO2 to 152 m /g with 50% CeO2 and addi-
tionally the ceria particle size increased with the ceria loading from
catalytic activity compared to fresh catalyst. The Ag–Ag backscat-
tering contribution is higher pointing to partial reduction of silver
under reaction conditions indicating that oxidized silver species
are at least partly present on the surface and not in the bulk silica.
Possibly also sintering of silver nanoparticles <10 nm might occur
as these are already very mobile at temperatures close to those
applied in the catalytic reaction [51].
(
2
2
2
3
to 8 nm (obtained from XRD). High silver loadings of 5–10 wt.%
gave silver particles around 10 nm achieving only low overall yields
of less than 2% in the oxidation of p-xylene (not shown). No silver
diffractions were found in the XRD patterns for CeO –SiO2 materi-
2
als with 1 wt.% silver loading due to the low concentration. When
tested for catalytic activity in the oxidation of p-xylene, FSP cata-
lysts containing CeO2 both gave higher TONs and higher product
3.3. Extension to further alkyl aromatic compounds
yields than the impregnated catalyst (Table 2, cf. entries 4, 6, 7
Extending the study to other alkyl aromatic compounds, the
oxidation of toluene, ethylbenzene and cumene was investigated.
None of the catalysts was able to convert toluene to benzyl alcohol,
benzaldehyde or benzoic acid under atmospheric pressure which is
likely to be connected to the boiling point of toluene being ca. 30 C
lower than that of p-xylene and the absence of a second presumably
FSP
with entry 1) with
1%Ag/10%CeO –SiO2 as the most active cat-
2
alyst (entry 4). Reusing the catalyst indicated catalyst deactivation
FSP
(
entry 5).
1%Ag/SiO2 hardly gave higher yields (entry 3) than
◦
those obtained in the absence of silver (entry 8). Leaching studies
on the mixtures after reaction with fresh FSP1%Ag/10%CeO –SiO2
2
showed that the silver concentration was below the limit of detec-
tion (<1 mg/L) in contrast to the impregnated catalyst (83 mg/L,
vide supra) even in the presence of carboxylic acid. In addition, the
activating methyl group. Thus, toluene oxidation was attempted in
◦
an autoclave both at 140 and 170 C with the impregnated cata-
◦
◦
lyst. Hardly any conversion was found at 140 C but at 170 C the
reaction was completed after 1 h (Table 4) the limiting factor here
being the amount of oxygen present in the reactor. Note that at
this temperature thermal toluene autoxidation becomes relevant
but proceeds at lower reaction rates [52].
cerium concentration was lower than 0.1 mg/L.
FSP
TEM analysis of
1%Ag/SiO2 indicated large silver particles of
up to 50 nm which might be a cause for the observed inferior cat-
alytic activity (Fig. 5a and b). Silica precursors with silver nitrate

M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
45
Fig. 5. TEM images of fresh FSP1%Ag/SiO2 (a and b) and fresh 1%Ag/10%CeO2–SiO2 (c and d). Dark spots represent silver particles. Circles indicate CeO2 particles.
Table 3
Overview over structural parameters obtained from EXAFS data fitting.
Catalyst
Shell
N (± 0.5)
R (Å)
ꢁ (Å−2)
Residual
Entry
imp
1
0%Ag/SiO2-900
Ag–Ag
Ag–O
Ag–Ag
Ag–O
Ag–Ag
Ag–O
Ag–Ag
Ag–O
12
1.7
3.1
3.8
0.5
3.1
3.8
2
2.86
2.29
2.84
2.32
2.81
2.28
2.85
2.07
3.37
2.86
0.010
0.016
0.016
0.022
0.013
0.030
0.010
0.003
0.027
0.009
6.7
1.5
1
2
FSP1%Ag/SiO2
FSP1%Ag/10%CeO2–SiO2
5.1
1.7
9.4
2.3
3
4
5
6
FSP1%Ag/10%CeO2–SiO2a
Ag2O
Ag–Ag
Ag–Ag
6
12
Ag
a
Used catalyst.
Table 4
Oxidation of toluene with imp10%Ag/SiO2-900 in the presence of ceria and carboxylic acid. Reaction conditions: 122 mmol toluene, 3 mol% benzoic acid, 100 mg biphenyl,
imp
1
00 mg
10%Ag/SiO2-900, 10 bar air (4.2 mmol O2), temperature as indicated, 1 h reaction time.
Yields (%)
Reaction temperature
Combined
TON
◦
◦
1
1
40
70
C
C
<0.1
1.1
<0.1
1.0
<0.1
0.5
<0.1
2.6
–
34

4
6
M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
Table 5
Oxidation of ethylbenzene with silver catalysts in the presence of CeO2 and carboxylic acid. Reaction conditions: 122 mmol ethylbenzene, 3 mol% benzoic acid, 100 mg
biphenyl, 100 mg catalyst, oxygen atmosphere, reflux, 3 h reaction time.
Catalyst
Yields (%)
TON
Entry
Combined
imp
1
0%Ag/SiO2-900
4.4
4.0
2.0
1.7
<0.1
11.8
6.8
3.8
3.2
<0.1
4.1
4.2
6.0
1.8
0.8
20.3
15.0
11.8
6.7
290
2000
1600
890
–
1
2
3
4
5
FSP
FSP
FSP
1
1
1
%Ag/10%Ce/SiO2
%Ag/30%Ce/SiO2
%Ag/50%Ce/SiO2
SiO2
0.9
The oxidation of ethylbenzene gave both 1-phenylethanol
and acetophenone and additionally ethylbenzene hydroperoxide
catalyst was most active in the presence of ceria and carboxylic
acid, ceria decreased the catalyst activity in the case of ethylben-
zene (Fig. 7a). Flame synthesized catalysts gave higher TONs of up
to 2000 (Table 5, entry 2) compared to the impregnated catalyst
but lower overall product yields.
(
Table 5), which are typical products for the radical (aut)oxidation
of ethylbenzene [32,53]. Compared to the oxidation of p-xylene, the
effect of ceria was different. While for p-xylene, the impregnated
Cumene oxidation gave mainly cumene hydroperoxide, ace-
tophenone and 2-phenyl-2-propanol (Table 6) with minor side
products being benzaldehyde and ␣-methyl styrene. The hydroper-
oxide exhibited transient behavior, i.e. its yield exhibited a
maximum after 1–2 h in the presence of carboxylic acid and
CeO . Interestingly, FSP catalysts supported on CeO –SiO gener-
2
2
2
ally resulted in lower yields (entries 2–4). Even in the presence
of added ceria nanoparticles the obtained yield was higher for
Fig. 6. EXAFS spectra of flame synthesized FSP1%Ag/SiO2 (a), FSP1%Ag/10%CeO2–SiO2
Fig. 7. Influence of the carboxylic acid and CeO2 on the oxidation of ethylbenzene (a)
as-prepared (b) and after the reaction (c) the Ag K-edge. Insets: Experimental (solid
and cumene (b). Reaction conditions: 122 mmol ethylbenzene or cumene, optionally
3
3 mol% benzoic acid and 50 mg CeO , 100 mg biphenyl, 100 mg imp10%Ag/SiO -900,
line) and fitted (dashed line, EXAFS data fitting in R-space) k -weighted Fourier-
2
2
◦
transformed EXAFS function.
oxygen atmosphere, 140 C, 3 h reaction time.

M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
47
Table 6
Oxidation of cumene with silver catalysts in the presence of CeO2 and carboxylic acid. Reaction conditions: 122 mmol cumene, 3 mol% benzoic acid, 100 mg biphenyl, 100 mg
◦
catalyst, 140 C, 3 h reaction time.
Catalyst
Yields (%)
TON
Entry
Combined
imp
1
0%Ag/SiO2-900
4.2
1.6
1.7
2.4
<0.1
9.1
0.1
0.1
0.1
0.1
9.5
0.6
0.5
0.4
1.7
22.8
2.3
2.3
2.9
1.8
330
310
310
390
–
1
2
3
4
5
FSP
FSP
FSP
1
1
1
%Ag/10%Ce–SiO2
%Ag/30%Ce–SiO2
%Ag/50%Ce–SiO2
SiO2
imp10%Ag/SiO -900 (entry 1). One of the reasons for this discrep-
ancy might be the close proximity between ceria and silver being
tion and termination (Scheme 2). The ratio of rate constants for
these steps determines the overall catalytic performance and is
dependent on the catalyst. Conversion of starting material was
already observed when only the impregnated silver catalyst was
used. The addition of peroxides to facilitate the start of the reaction
as reported in other cases [56] was not necessary. Hence, silver facil-
itates the formation of initial radicals (Scheme 2, step 1). The chain
reaction is likely not to be supported by the silver catalyst, firstly,
because the product distribution in the absence of carboxylic acid
is typical for uncatalyzed autoxidations [57,58]. In addition, chain
propagation is fast compared to initiation. This would also require
a catalyst in the same phase as e.g. in case of the homogeneous
Co/Mn/Br catalyst system [9] but a correlation between leaching
and catalyst activity could not be established for p-xylene. Diffusion
of the radicals to Ag particles and CeO2 would be likely to occur in
the same time regime so that the rate constants for chain propaga-
tion and termination would be similar if chain propagation relies
on silver. Silver may, however, play a role in the decomposition of
hydroperoxides influencing the radical concentration.
2
the radical initiator for CeO –SiO2 supported FSP catalysts and the
2
smaller CeO2 particle size (3–8 nm, Table 1, compared to 25 nm for
the commercial ceria). This may also be the origin for the deac-
imp
tivating effect when adding ceria to
10%Ag/SiO -900 even in
2
the presence of benzoic acid (Fig. 7b) as observed for ethylben-
zene oxidation. Interestingly, flame synthesized silver catalysts
with SiO –CeO support gave not only lower overall yields but also
2
2
an altered product distribution. Only small amounts of alcohol and
peroxide were formed (which are the main products in the reac-
tions with the impregnated Ag catalysts) the main product being
acetophenone. These alterations indicate a different reaction mech-
anism and are probably due to a changing ratio of the reaction rates
of chain termination vs. propagation and the suppression of some
radical reaction pathways.
imp
When
10%Ag/SiO -900 was used as a catalyst, the oxida-
2
tion of ethylbenzene and cumene continued after removal of the
solid phase. In case of cumene a silver concentration of 9 mg/L was
found by ICP-MS while for FSP1%Ag/10%CeO –SiO2 the silver con-
FSP synthesized catalysts featured both metallic silver and Ag–O
2
centration was below the detection limit. This indicates a shift from
heterogeneous to homogeneous catalysis and explains also the
inferior performance and the different product distribution over
the FSP catalysts that tend to leach less. Note that peroxo species
formed during the presence of the catalyst and their higher stabil-
ity compared to p-xylene oxidation may also maintain the reaction
to a certain extent.
species which might be reduced under reaction conditions since
FSP
the used
1%Ag/10%CeO2–SiO2 catalyst showed a stronger Ag–Ag
backscattering in the EXAFS spectrum. Silver in the active impreg-
nated catalyst was found to be only metallic. Thus, metallic silver
appears to be required for an active catalyst. The ambivalent influ-
ence of ceria suggests that ceria is interacting with two parts of the
overall radical reaction one of which is obviously the chain termi-
nation (Scheme 2, step 3), since the reaction is suppressed when
3.4. Discussion and mechanistic considerations
The oxidation of different alkyl aromatic compounds can be pro-
moted by heterogeneous silver catalysts prepared via impregnation
or flame spray pyrolysis. Yields are similar to other heterogeneous
catalyst systems [19] but there is still a large gap in performance
compared to the industrial homogeneous Amoco process. In the
case of p-xylene and ethylbenzene the FSP catalysts were con-
siderably more active with respect to the Ag loading. In addition,
the FSP catalysts exhibited a significantly lower leaching tendency
which might originate from a stronger metal-support interaction
established during synthesis by combustion of the liquid precur-
sor. In contrast, the impregnated catalyst performed superior in
the cumene oxidation and gave another product distribution than
the FSP catalysts thus indicating a different mechanism changing
from heterogeneous to homogeneous catalysis.
The suppression of the oxidation reaction by radical scavengers
and the formation of hydroperoxides strongly suggest that the
investigated p-xylene oxidation reaction is based on a radical reac-
tion mechanism, which is supported by many examples from the
literature for analogous cases (e.g. [9,54,55]). Radical reactions
comprise mainly three steps, namely initiation, radical chain reac-
Scheme 2. Schematic reaction sequence in the radical autoxidation of p-xylene and
potential roles of Ag, CeO2 and p-toluic acid promoters.

4
8
M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
only ceria and silver catalyst is used. Initiation by silver might be
influenced by ceria (Scheme 2, step 1). Ceria is known to activate
oxygen [39], which could then support breaking of C-H bonds in p-
xylene which are stronger compared to ethylbenzene and cumene.
Thus, CeO2 has no beneficial effect in the latter cases. Additionally,
the latter more stable peroxo radicals [59] might have a longer half
life which makes quenching by scavengers, which CeO2 is, more
likely. Molecular oxygen does not play a dominant role in the radi-
cal chain reaction since the reaction of alkyl radicals with molecular
oxygen is fast [57]. Addition of carboxylic acid did not only increase
the overall yield but also influenced the product distribution and
hence changed the overall mechanism either directly via the prop-
agation mechanism [54] (Scheme 2, step 2) or via destabilization
of the termination product, i.e. the hydroperoxide, in the presence
of radicals [60]. Indeed, in the case of p-xylene, ethylbenzene and
cumene oxidation the hydroperoxide exhibited transient behavior.
A higher degree of hydroperoxide dissociation would also result in
a higher radical concentration which explains why not only the
product distribution was shifted but also a higher overall yield
was obtained. The higher O–H bond dissociation energies (BDE) of
benzoic acid and toluic acid compared to the C–H BDE of alkyl aro-
matic compounds and the O–H BDE of the corresponding peroxides
cantly less prone to leaching compared to the impregnated catalyst.
This may be exploited in the future to prepare more stable catalysts
for liquid phase reactions where leaching should be circumvented.
In order to be active, silver is required to be metallic as suggested
by EXAFS analysis. The oxidation reaction is most likely based on
a radical mechanism where CeO2 possibly assists silver as a chain
initiator, while the carboxylic acid influences the dissociation of
peroxides to radical species to initiate radical chains. Ceria as a
radical scavenger additionally induces chain termination.
Acknowledgements
Financial support of the study by the Technical University of
Denmark (DTU) and the Danish Research Council (FTP-project) as
well the support of the PhD grant by DTU, the Graduate School
MP T and Haldor Topsøe A/S are gratefully acknowledged. Finan-
2
cial support by ETH Research Grant TH-41 06-1 and discussions
with Georgios Sotiriou (ETH Zürich) are gratefully acknowledged.
We thank Helge Rasmussen for assistance during the XRD measure-
ments, Evonik Degussa GmbH for donation of support materials and
HASYLAB at DESY, Hamburg, for beamtime and the EU for finan-
cial support of the beamtime (Contract RII13-CT-2004-506008).
Birgit Elkjær Ascanius (DTU KT) is acknowledged for providing
GC-MS analysis. The A.P. Møller and Chastine Mc-Kinney Møller
Foundation is gratefully acknowledged for its contribution towards
establishment of the Center for Electron Nanoscopy.
[
59] makes the presence of carboxyloxo radical species generated
from the added carboxylic acid unlikely. Thus, the carboxylic acids
presumably promote the peroxide dissociation. The capability of
•
carboxylic acids to act as an H-bond acceptor can stabilize OH
radicals [61]. This lowers the activation energy for e.g. peroxide
dissociation thereby generating a higher radical concentration and
promoting the chain initiation.
References
Cumene oxidation with FSP synthesized CeO –SiO2 sup-
ported catalysts proceeded with a low reaction rate and gave
mainly acetophenone as a product. This was not the case when
2
[1] J.M. Thomas, R. Raja, Solid State Sci. (2001) 675–687.
[2] J.M. Bregeault, Dalton Trans. (2003) 3289–3302.
[
[
3] T. Punniyamurthy, S. Velusamy, J. Iqbal, Chem. Rev. 105 (2005) 2329–2363.
4] A.K. Suresh, M.M. Sharma, T. Sridhar, Ind. Eng. Chem. Res. 39 (2000) 3958–3997.
imp
1
0%Ag/SiO -900 was used in the presence of ceria nanopar-
2
[5] M. Bohnet, Ullmann’s Encyclopedia of Industrial Chemistry, seventh ed., Wiley-
VCH, Weinheim, 2003.
ticles, but here leaching of silver is important. In both cases the
product distribution differs from typical cumene autoxidations for
which the main product is cumene hydroperoxide. Thermal decom-
position of the hydroperoxide alone cannot be the reason since
in the absence of ceria and benzoic acid cumene hydroperoxide
was by far the dominating product. Ceria was used before as a
catalyst for alkyl aromatics oxidation with bromate affording the
carbonyl compound as the main product [58]. Acetophenone can
be formed from cumene hydroperoxide either by reaction with a
reducing agent [62] or generation of a hydroxyl radical via cleavage
of the oxygen–oxygen bond [57] resulting an intermediate ␣,␣-
dimethylbenzyloxy radical. According to Bhattacharya [57] this
radical preferably reacts to form acetophenone. If either one of
these reactions is promoted by CeO2 the close proximity between
[
6] H.J. Rozie, M.L.C. Dsinter, J.B. Dakka, A. Zoran, Y. Sasson, PCT Int. Appl., 9,520,560
1995).
(
[
7] M.L. Kantam, B.M. Choudary, P. Sreekanth, K.K. Rao, K. Naik, T.P. Kumar, A.A.
Khan, Eur. Pat. Appl., 1,088,810 (2001).
[
[
8] S. Park, J.S. Yoo, K. Jun, D.B. Raju, Y. Kim, U.S. Patent, 6,476,257 (2002).
9] W. Partenheimer, Catal. Today 23 (1995) 69–158.
[
10] V. Kumar, P.D. Grover, Ind. Eng. Chem. Res. 30 (1991) 1139–1141.
[11] G. Busca, J. Chem. Soc., Faraday Trans. 89 (1993) 753–755.
[12] J.S. Yoo, J.A. Donohue, M.S. Kleefisch, P.S. Lin, S.D. Elfline, Appl. Catal. A 105
(
1993) 83–105.
[
13] P.A. Hamley, T. Ilkenhans, J.M. Webster, E. Garcia-Verdugo, E. Venardou, M.J.
Clarke, R. Auerbach, W.B. Thomas, K. Whiston, M. Poliakoff, Green Chem. 4
(2002) 235–238.
14] T. Seki, M. Baiker, Chem. Rev. 109 (2009) 2409–2454.
15] H.V. Borgaonkar, S.R. Raverkar, S.B. Chandalia, Ind. Eng. Chem. Prod. Res. Dev.
[
[
23 (1984) 455–458.
[16] M. Hronec, Z. Hrabe, Ind. Eng. Chem. Prod. Res. Dev. 25 (1986) 257–261.
17] M. Ilyas, M. Sadiq, Catal. Lett. 128 (2009) 337–342.
18] B. Chou, J. Tsai, S. Cheng, Micropor. Mespor. Mater. 48 (2001) 309–317.
[19] F. Wang, J. Xu, X. Li, J. Gao, L. Zhou, R. Ohnishi, Adv. Synth. Catal. 347 (2005)
1987–1992.
[
[
CeO particles and silver stable against leaching on the FSP catalysts
2
in the investigated reaction media might be the reason why CeO2
already present on the support showed a more deactivating effect
[
[
[
[
[
20] R. Raja, J.M. Thomas, V. Dreyer, Catal. Lett. 110 (2006) 179–183.
21] R. Raja, J.M. Thomas, Solid State Sci. 8 (2006) 326–331.
thanCeO addedtothereactionmixture (incase ofthe impregnated
2
catalyst).
22] M.J. Beier, T.W. Hansen, J.-D. Grunwaldt, J. Catal. 266 (2009) 320–330.
23] R. Strobel, A. Baiker, S.E. Pratsinis, Adv. Powder Technol. 17 (2006) 457–480.
24] R. Kydd, W.Y. Teoh, K. Wong, Y. Wang, J. Scott, Q. Zeng, A. Yu, J. Zou, R. Amal,
Adv. Funct. Mater. 19 (2009) 369–377.
25] S. Hannemann, J.-D. Grunwaldt, P. Lienemann, D. Guenther, F. Krumeich, S.E.
Pratsinis, A. Baiker, Appl. Catal. A 316 (2007) 226–239.
4
. Conclusions
[
Silver particles supported on silica effectively promote the side-
[26] H. Schulz, L. Mädler, R. Strobel, R. Jossen, S.E. Pratsinis, T. Johannessen, J. Mater.
Res. 20 (2005) 2568–2577.
chain oxidation of toluene, p-xylene, ethylbenzene and cumene
without any solvent. Except for toluene, no elevated pressures
were needed to facilitate the reaction. Addition of a carboxylic
[
27] J. Yuan, X. Liao, H. Wang, G. Yang, M. Tang, J. Phys. Chem. B 113 (2009)
1418–1422.
[28] T. Ressler, J. Synchrotron Radiat. 5 (1998) 118–122.
[
29] S.I. Zabinsky, J.J. Rehr, A. Ankudinov, R.C. Albers, M.J. Eller, Phys. Rev. B: Condens.
Matter Mater. Phys. 52 (1995) 2995–3009.
acid considerably enhanced the reaction rate. CeO had an ambiva-
2
lent influence on the reaction and could both promote and inhibit
product formation depending on substrate and reaction conditions.
Except for cumene oxidation silver catalysts prepared via impreg-
nation were inferior to flame-made catalysts which gave TONs up
to 2000. In addition, the flame synthesized catalysts were signifi-
[
30] R.J. Hill, C.J. Howard, AAEC (now ANSTO), Australian Atomic Commis-
sion Report No. M112, 2nd ed., Lucas Heights Research Laboratories, New
South Wales, Australia, 1986; winpow was obtained from http://struktur.
kemi.dtu.dk/kenny/powder programs.html (last accessed on April, 16th, 2010).
31] K. Ohkubo, S. Fukuzumi, Org. Lett. 2 (2000) 3647–3650.
[
[32] G. Sereda, V. Rajpara, Tetrahedron Lett. 48 (2007) 3417–3421.

M.J. Beier et al. / Journal of Molecular Catalysis A: Chemical 331 (2010) 40–49
49
[
[
33] C. Rehren, G. Isaac, R. Schlögl, G. Ertl, Catal. Lett. 11 (1991) 253–265.
34] X. Bao, G. Lehmpfuhl, G. Weinberg, R. Schlögl, G. Ertl, J. Chem. Soc., Faraday
Trans. 88 (1992) 865–872.
[46] P. Niggli, Z. Kristallogr., Kristallgeom., Kristallphys. Kristallchem. 57 (1922)
253–299.
[47] F. Liebau, Acta Crystallogr. 14 (1961) 537–538.
[
[
35] Z.P. Qu, M.J. Cheng, W.X. Huang, X.H. Bao, J. Catal. 229 (2005) 446–458.
36] Z.P. Qu, W.X. Huang, M.J. Cheng, X.H. Bao, J. Phys. Chem. B 109 (2005)
[48] M. Jansen, H.L. Keller, Angew. Chem. 91 (1979) 500.
[49] K. Heidebrecht, M. Jansen, Z. Anorg. Allg. Chem. 597 (1991) 79–86.
[50] W. Klein, M. Jansen, Z. Anorg. Allg. Chem. 634 (2008) 1077–1081.
[51] O.A. Yeshchenko, I.M. Dmitruk, A.A. Alexeenko, A.V. Kotko, Nanotechnology 21
(2010) 045203/1–045203/6.
[52] I. Hermans, J. Peeters, L. Vereecken, P.A. Jacobs, ChemPhysChem 8 (2007)
2678–2688.
[53] T.V. Bukharkina, O.S. Grechishkina, N.G. Digurov, N.V. Krukovskaya, Org. Pro-
cess. Res. Dev. 3 (1999) 400–403.
[54] Y. Kamiya, M. Kashima, Bull. Chem. Soc. Jpn. 46 (1973) 905–908.
[55] Q. Jiang, Y. Xiao, Z. Tan, Q. Li, C. Guo, J. Mol. Catal. A: Chem. 285 (2008) 162–168.
[56] P.P. Toribio, A. Gimeno-Gargallo, M.C. Capel-Sanchez, M.P. de Frutos, J.M.
Campos-Martin, J.L.G. Fierro, Appl. Catal. A 363 (2009) 32–39.
[57] A. Bhattacharya, Chem. Eng. J. 137 (2008) 308–319.
1
5842–15848.
37] S. Babu, A. Velez, K. Wozniak, J. Szydlowska, S. Seal, Chem. Phys. Lett. 442 (2007)
05–408.
[
[
[
[
4
38] P. Trogadas, J. Parrondo, V. Ramani, Electrochem. Solid-State Lett. 11 (2008)
B113–B116.
39] A. Trovarelli, G.J. Hutchings (Eds.), Catalysis by Ceria and Related Materials,
Imperial College Press, London, 2002.
40] A. Abad, P. Concepcion, A. Corma, H. Garcia, Angew. Chem., Int. Ed. 44 (2005)
4
066–4069.
[
[
41] T. Vogler, A. Studer, Synthesis (2008) 1979–1993.
42] S. Hannemann, J.-D. Grunwaldt, F. Krumeich, P. Kappen, A. Baiker, Appl. Surf.
Sci. 252 (2006) 7862–7873.
[
[
[
43] H. Schulz, L. Maedler, S.E. Pratsinis, P. Burtscher, N. Moszner, Adv. Funct. Mater.
[58] I. Hermans, J. Peeters, P.A. Jacobs, J. Org. Chem. 72 (2007) 3057–3064.
[59] E.T. Denisov, E.V. Tumanov, Russ. Chem. Rev. 74 (2005) 905–938.
[60] C. Wang, H. Chen, C. Chen, Polym. Adv. Technol. 17 (2006) 579–586.
[61] I. Hermans, J. Peeters, P.A. Jacobs, ChemPhysChem 7 (2006) 1142–1148.
[62] H. Boardman, J. Am. Chem. Soc. 75 (1953) 4268–4271.
1
5 (2005) 830–837.
44] A. Teleki, S.E. Pratsinis, K. Wegner, R. Jossen, F. Krumeich, J. Mater. Res. 20 (2005)
336–1347.
45] B. Medasani, Y.H. Park, I. Vasiliev, Phys. Rev. B: Condens. Matter Mater. Phys.
5 (2007), 235436-1-6.
1
7