The Impact of Reaction Pressure on the Catalytic Performance of the Pd-Sb/TiO2 Catalyst in the Acetoxylation of Toluene into Benzyl Acetate

Article abstract of DOI:10.1002/cctc.201200522

The acetoxylation of toluene in the presence of acetic acid and oxygen was performed over a Pd-Sb/TiO₂ catalyst at 210°C and at various reaction pressures (1-10bar). A remarkable improvement in the catalytic performance and a significant shortening of the induction period were found with an increase in pressure. At a pressure of 6bar, the highest toluene conversion of 75% and 100% selectivity for benzyl acetate were observed. This result could be due to both the restructuring of the catalyst surface and to the formation of active Pd-Sb particles of desired size, composition, and shape during the course of the induction period. Samples of the most-active and spent catalysts were studied by ex situ and insitu XRD, XPS, and TEM.

Full text of DOI:10.1002/cctc.201200522

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201200522
The Impact of Reaction Pressure on the Catalytic
Performance of the PdÀSb/TiO Catalyst in the
2
Acetoxylation of Toluene into Benzyl Acetate
Neetika Madaan, Suresh Gatla, Venkata Narayana Kalevaru, Jçrg Radnik,
Marga-Martina Pohl, Bernhard Lꢀcke, Angelika Brꢀckner, and Andreas Martin*
[
a]
The acetoxylation of toluene in the presence of acetic acid and
of 75% and 100% selectivity for benzyl acetate were observed.
This result could be due to both the restructuring of the cata-
lyst surface and to the formation of active PdÀSb particles of
desired size, composition, and shape during the course of the
induction period. Samples of the most-active and spent cata-
lysts were studied by ex situ and in situ XRD, XPS, and TEM.
oxygen was performed over a PdÀSb/TiO catalyst at 2108C
2
and at various reaction pressures (1–10 bar). A remarkable im-
provement in the catalytic performance and a significant short-
ening of the induction period were found with an increase in
pressure. At a pressure of 6 bar, the highest toluene conversion
1
. Introduction
Over the past two decades, the Pd-catalyzed oxidation of a-
olefins and alkyl–aromatic compounds has become an increas-
ingly attractive field and, in some instances, it is the only route
for the selective synthesis of vinylic, allylic, and benzylic alco-
hols and esters. In this context, the development of more-ef-
fective and more-economic syntheses of benzylic esters/alco-
hols and even aldehydes as intermediates will be of great in-
terest for the synthesis of various industrially important
chemicals.
food industry, owing to its fruity aroma. BA is also notably
used in the chemical industry as a solvent for the manufacture
[7]
of cellulose acetate. The global demand for BA ranges from
5–10 kt per annum. In addition, in the future, it could serve as
an intermediate for the manufacture of benzyl alcohol that is
presently obtained by the rather harmful chlorination of tolu-
ene into benzyl chloride and subsequent hydrolysis.
[8–10]
From previous investigations,
it has been observed that
the performance of the acetoxylation reaction is strongly de-
pendent on different parameters, such as Pd-particle size, the
nature of the support, the type of the promoters (Sb seems to
be the most beneficial), the preparation method, the calcina-
tion atmosphere, etc. Of these parameters, the size of the Pd
particles has shown a strong influence on the catalytic perfor-
mance, that is, 50–100 nm particles were found to exhibit
better performance than smaller ones. It looks that this critical
size, that is, a defined ensemble of Pd atoms, is needed to
reach sufficient conversion and product selectivity. The catalyst
only exhibits maximum activity at this critical size, which is
reached after an induction period of 7–50 h, depending on the
catalyst precursor that is used and its conditioning prior to the
The acetoxylation of toluene (Tol) is a one-step process for
the synthesis of benzyl acetate (BA) in the presence of acetic
acid and oxygen (Scheme 1). Palladium-containing catalysts are
Scheme 1. The synthesis of benzylacetate.
[11,12]
widely used in various acetoxylation reactions, such as the
manufacture of vinyl acetate monomer (VAM) from ethyl-
reaction.
In addition, various investigations have led to the
conclusion that a defined Pd/promoter surface-atom ratio
seems to be advantageous (in the case of Sb, the Pd/Sb ratio
[
1,2]
ene.
Besides olefins, aromatic compounds, such as toluene,
[10]
can be converted over Pd-containing catalysts in the same
is 5:1). Furthermore, catalyst deactivation is observed after
a steady-state operating period, owing to surface reduction of
[
3–6]
way.
BA is naturally found in the odorous substances of var-
dÀ
[11,12]
ious plants; therefore, it is mainly used in the perfume and
PdO into Pd and Pd species and coking, respectively.
Re-
generation at 3008C approximately restores the concentration
of PdO but, at 3508C, the PdO portion on the surface remains
very low. In addition, Pd , PdO, and Sb synergy is needed for
[
a] Dr. N. Madaan, Dr. S. Gatla, Dr. V. N. Kalevaru, Dr. J. Radnik,
Dr. M.-M. Pohl, Prof. B. Lꢀcke, Prof. A. Brꢀckner, Dr. A. Martin
Leibniz-Institut fꢀr Katalyse e. V. an
der Universitꢁt Rostock
Albert-Einstein-Str. 29a, 18059 Rostock (Germany)
Fax: (+49)0381-1281-51246
0
a stable and high performance of the catalyst.
In general, the reaction is performed at a mild reaction tem-
perature of about 2108C and at a modest pressure of 2 bar.
Pressure effects in catalysis are widely known from detailed
E-mail: andreas.martin@catalysis.de
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 185

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
studies, for example, for changing the product distribution in
the Fischer–Tropsch synthesis over Nb-supported Co-contain-
[
13]
ing catalysts. Other examples might be the effect of pressure
[
14]
on the conversion of MeOH over zeolites or the impact of
reaction pressure on the thermal and catalytic cracking of in-
[15]
dustrial feedstock and coke formation in catalytic cracking.
Besides these effects, it was expected that, in the acetoxylation
of toluene, the reaction pressure might also influence the Pd-
particle size owing to increased surface coverage.
Moreover, it was speculated by ourselves that larger Pd par-
ticles, which were demonstrably present after the synthesis
and precursor calcination under an inert-gas (He) atmosphere
at high temperature, more or less completely erode and that
the more-beneficial smaller Pd particles are formed in an ad-
vantageous Pd/Sb ratio. The rationale for this assumption was
the fact that 1) pure Pd particles were found in the as-synthe-
sized solids and 2) regardless of the calcination procedure,
whether it led to small (2–5 nm) or large (1–2 mm) Pd particles,
medium-sized particles (50–100 nm) were present in the
above-described Pd/Sb ratio after passing the induction
period. In the most-beneficial case, a shortening of the induc-
tion period could occur by increasing the reaction pressure.
Herein, we report, for the first time, the marked effect of the
reaction pressure (up to 10 bar) on the geometric and elec-
tronic structure and the related catalytic performance of a PdÀ
Figure 1. XRD pattern of the parent PdÀSb/TiO
2
catalyst after calcination
under helium.
Sb/TiO catalyst. The aim of this work is to gain deeper insight
2
into the dynamic changes in the surface structure, size, and
composition of the Pd particles that occur during the catalytic
reaction as a consequence of pressure. Special emphasis is de-
voted to enhancing the activity and selectivity, along with
a shorter induction period.
Figure 2. TEM image of big Pd particles (about 1–2 mm) on the surface of
the fresh calcined PdÀSb/TiO catalyst.
2
surface area with increasing reaction pressure, probably
caused by progressive coking (Table 1).
2
. Results and Discussion
2.1. Composition, BET surface area, and phase behavior of
the most-active and spent catalysts
Table 1. BET-surface areas (BET-SAs) of the most-active and spent cata-
lysts at different reaction pressures.
[
a]
The fresh calcined PdÀSb/TiO2 solid showed the following
composition, as obtained by ICP-OES: 9 wt.% Pd, 12 wt.% Sb.
The BET surface area and the pore volume of this sample were
2
À1
Pressure [bar]
BET-SA [m g
]
Most-active catalyst
Spent catalyst
2
À1
3
À1
found to be 28.6 m g and 0.1 cm g , respectively. XRD anal-
ysis revealed the presence of only metallic Pd in the fresh cata-
lyst (Figure 1). A TEM image of the fresh catalyst showed quite-
large Pd particles with sizes that ranged from 1–2 mm
1
2
4
6
8
39.4
40.6
27.7
24.8
26.1
28.1
25.1
23.8
21.4
12.8
(
Figure 2). EDX analysis revealed that these particles exclusively
2
À1
[
a] The fresh catalyst has a BET-SA of 28.6 m g .
contained Pd.
The composition of samples of the most-active and spent
catalysts was also studied. However, their composition was
close to that of the parent sample, that is, neither a loss of Pd
nor of Sb was observed. The BET surface area of the most-
active samples revealed a significant increase for those sam-
ples at pressures of 1 or 2 bar; this result point to a significant
surface-restructuring that occurs during the induction period.
At higher reaction pressures, the most-active solids showed
surface areas that were close to that of the parent sample.
However, the spent samples revealed a significant drop in BET
XRD patterns of the most-active and spent catalysts at differ-
ent pressures are shown in Figure 3a–d. Reflections that were
related to different phases, such as metallic Pd, NaCl (still pres-
ent from the synthesis), and TiO (support, anatase), were pres-
2
ent in all of the samples, irrespective of the applied reaction
pressure. However, some changes were observed in the Pd re-
flections with respect to peak intensity and peak position. In
general, all of the patterns of the most-active and spent sam-
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 186

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 4. Effect of the reaction pressure on the catalytic performance of the
PdÀSb/TiO
2
catalyst after the induction period, that is, at maximum activity;
À1
T=2108C, molar ratio of Tol/AcOH/O
2
/N
2
=1:4:3:16, GHSV=2688 h (con-
version of toluene: X-Tol, selectivity for benzyl acetate: S-BA).
Figure 3. XRD patterns of samples that were treated at different reaction
pressures (the parent sample is included for comparison): a) catalysts at their
maximum performance; b) most-active samples in the range of the Pd[111]
reflection (2q=38–418); c) spent, deactivated catalysts; and d) these deacti-
vated samples in the range 2q=38–418.
ples revealed a decrease in intensity and/or a shift in the posi-
tion of the Pd[111] reflection towards lower 2q values, owing
[
10]
to the incorporation of Sb.
In addition, a shoulder or
a second peak at 2q=39.58 could be noticed at higher reac-
tion pressures, owing to the growth of a Pd–carbide phase, as
[
10]
observed previously.
Figure 3b shows the diffractograms of the most-active sam-
ples in the range 2q=38–428; there is a clear shift in the Pd
reflections compared to the parent sample. Careful examina-
tion of the reflections showed that, with an increase in pres-
sure, the shift in the peak position is decreased; that is, a shift
in the Pd reflection is much more prominent at 1, 2 and 4 bar
compared to the samples at 6 and 8 bar. At higher pressures,
Pd tends to exist in its metallic form. XRD patterns of the
spent (deactivated) catalysts (Figure 3c) showed no difference
in their phase composition compared to the patterns of the
most-active samples. However, a shift in the Pd reflection (Fig-
ure 3d) was still observed in all of the samples, but with a mar-
ginal shift in the solids that were treated at higher pressures.
Figure 5. Dependence of the catalyst-induction period and the time of
stable operation, including the induction period, on reaction pressure;
T=2108C, p=1–10 bar, molar ratio of Tol/AcOH/O
2
/N
2
=1:4:3:16,
À1
GHSV=2688 h
.
some adverse effects on long-term stability, particularly when
the pressure was raised to 6 bar and above. In fact, the long-
term stability decreased from 40 h at 1 bar to 12 h at 10 bar.
However, one should keep in mind that the catalyst was ex-
posed to a tenfold proportion of reactant molecules and its
productivity, that is, space-time-yield, increased simultaneously.
Figure 6 and Figure 7 show the time-on-stream changes
over up to 12 h with respect to X-Tol and S-BA in more detail.
The lowest reaction pressure (1 bar) displayed relatively low in-
itial activity (X-Tolꢀ1%, S-BAꢀ5%), which increased with
time-on-stream and reached its maximum (X-Tol=33%, S-BA=
63%) after an induction period of 11 h. This period seems to
be the time required for the Pd particles to transform from 1–
2 mm (observed in the fresh solid) into the crucial size of 50–
100 nm, which is necessary for improved performance. This in-
crease in size is evident from the TEM measurements (see
below). At a pressure of 2 bar, the catalyst also showed a low
initial activity (X-Tolꢀ3%, S-BAꢀ10%) that increased steadily
with time-on-stream. However, some changes in the induction
period, toluene conversion, selectivity for BA, and long-term
stability were evident compared to the sample at 1 bar reac-
tion pressure. In this case, the maximum conversion of toluene
2.2. Pressure dependence of the catalytic performance
Figure 4 shows the catalytic data after reaching maximum ac-
tivity, that is, after passing the induction period, and, thus,
clearly demonstrates the effect of pressure on the conversion
of toluene (X-Tol) and the selectivity for benzyl acetate (S-BA).
X-Tol increased from 34 to 75% on increasing the reaction
pressure from 1 to 6 bar and then remained constant with fur-
ther increases in pressure. In a similar way, S-BA also increased
to 100% when the reaction pressure was raised to 6 bar and
then decreased slightly to 86% with a further increase in pres-
sure. Simultaneously, the pressure showed an additional bene-
ficial effect on the induction period, which was observed to
decrease from 11 h (1 bar) to only 3 h (6 bar), as shown in
Figure 5. On the other hand, the reaction pressure had shown
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 187

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
2.3. In situ XRD analysis of particle restructuring
In previous investigations, we found out that catalysts that
contained mixed-metal particles with a Pd/Sb atomic ratio of 5
and a particle size of between 50 and 100 nm showed the
[10]
highest catalytic performance. However, these characteristics
were only reached after a certain conditioning period, fresh (in-
[10]
active) catalysts contained pure Pd particles of 1–2 mm.
Clearly, this restructuring of the particles requires a certain
time, which depends on the reaction pressure, that is, the con-
centration of the reactants. Such dynamic behavior of the sur-
face, which includes the dissolution of the parent Pd particles
and the formation of mixed PdÀSb particles with optimum
composition and size, was confirmed by an in situ XRD experi-
ment that showed a decrease in the Pd[111] reflection and the
growth of a new peak that was assigned to a PdÀSb phase.
Figure 8 shows two patterns from a set of in situ X-ray diffrac-
Figure 6. Dependence of a plot of toluene conversion (X-Tol) versus time-
on-stream of the PdÀSb/TiO
2
catalyst on reaction pressure; T=2108C, p=1–
À1
10 bar, molar ratio of Tol/AcOH/O
2
/N
2
=1:4:3:16, GHSV=2688 h
.
Figure 8. In situ-XRD patterns after the start of the in situ reaction and after
12h on-stream.
Figure 7. Dependence of a plot of benzyl acetate selectivity (S-BA) versus
time-on-stream of the PdÀSb/TiO
2
catalyst on reaction pressure; T=2108C,
À1
p=1–10 bar, molar ratio of Tol/AcOH/O
2
/N
2
=1:4:3:16, GHSV=2688 h
.
tograms that were obtained under acetoxylation-reaction con-
ditions at 2108C and ambient pressure in an in situ cell (Anton
Paar, XRK900). These patterns reveal a loss of the intensity of
both of the Pd reflections (the [111] reflection at 2q=408 and
the [200] reflection at 2q=46.58) and the growth of new re-
flections at slightly lower 2q values. These reflections can be
correlated with an expansion of the Pd lattice, presumably
owing to the incorporation of Sb and carbon. Recently, a similar
phenomenon of sintering and reactivation of Pt and Rh in au-
rose to 54% and the selectivity for BA reached 89%. In addi-
tion, the induction period was halved from 11 h (at 1 bar) to
6
h (at 2 bar). This result indicates that the rate of Pd restruc-
turing is somehow faster at higher pressures. In addition, no
considerable deactivation took place up to this pressure. A
similar behavior (i.e., low initial activity) was also observed at
higher pressures of up to 6 bar, at which the catalyst reached
its highest performance. After starting with low activity (X-
Tol =8%, S-BA=63%), a remarkable conversion of 75% and
a selectivity to BA of 100% were obtained within only 3 h on-
stream of the induction period. Obviously, the induction
period has been shortened significantly from 11 h at 1 bar to
only 3 h at 6 bar. It appears that these beneficial effects are
due to an acceleration of the changes in the surface, size, and
composition of the Pd particles under the influence of pres-
sure, that is, increased surface-reactant concentrations. Such
[16]
tomotive-exhaust catalysts was described. In our case, the
beneficial influence of pressure on the restructuring could be
observed at pressures of up to 6 bar. Higher pressures (above
6 bar) showed a detrimental effect (Figure 5). Obviously, such
reaction pressures, that is, higher surface concentrations of the
reactants, hampered an effective desorption of the product. In
addition, this increased surface-molecule traffic led to the for-
mation of enhanced oligomers and coke deposition.
As expected, the extent of carbon deposition in the spent
À1
samples only marginally increased from 0.043 wt.%h at 1 bar
À1
changes on the PdÀSb/TiO catalysts have also been observed
to 0.063 wt.%h at 4 bar. However, a more-pronounced in-
2
[
10]
À1
in previous studies, yet they took much longer.
crease, up to 0.25 wt.%h , was observed with a further rise in
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 188

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
À1
Figure 9. Plot of carbon deposition (wt.%h ) of the spent samples versus
reaction pressure (1–8 bar); T=2108C, molar ratio of Tol/AcOH/O
2
/
À1
N
2
=1:4:3:16, GHSV=2688 h
.
0
dÀ
Figure 10. Pd 3d spectra (XPS) that shows PdO, Pd , and Pd species in the
pressure to 8 bar (Figure 9). This fact suggests that the higher
concentration of carbonaceous compounds on the catalyst sur-
face at higher reaction pressures promotes coke formation
and, therefore, deactivation. This coke formation leads to 1) a
decrease in BET surface area, 2) a drop in the surface concen-
near-surface region.
pressure. In addition, a third Pd species was found in the cata-
lysts that were tested at 4 bar and above. This species is
dÀ
tration of Pd, and 3) the formation of undesired Pd species
a Pd species with a markedly lower binding energy value
dÀ
(
e.g., Pd ), owing to interactions between deposited surface
(B.E.=333.7 eV) compared to metallic Pd (B.E.=335.1 eV). As
dÀ
coke and surface-Pd species.
has been reported earlier, the formation of a Pd state is
caused by the interactions between surface Pd and deposited
[
11]
dÀ
carbon species. The formation of Pd species is typical in
2.4. Extended XPS studies of the most-active and spent
[11]
catalysts that undergo quick deactivation; it was not ob-
served in the catalysts that were tested at 1 and 2 bar, which
were stable for a longer time (Figure 11). In contrast, the sam-
samples
To gain further insight into the composition and oxidation
states of Pd in its nearest-surface region and the dependence
of these properties on reaction pressure, XPS analysis was per-
formed on the most-active samples and compared with that of
the fresh calcined sample. The Pd 3d XPS spectra revealed the
presence of only oxidized Pd species, with a binding energy
(B.E.) of 335.8 eV on the surface of this sample (Figure 10),
apart from metallic Pd in the bulk sample, as revealed by the
XRD studies (see above). However, these oxidized surface spe-
cies were gradually reduced into metallic Pd during the course
of the reaction. As a result, in all of the most-active samples,
both oxidized and metallic Pd species co-existed in different
proportions, which, again, depended on the reaction pressure.
In fact, the presence of both metallic and oxidized Pd species
is crucial for obtaining high catalytic performance and also for
Figure 11. Distribution of different Pd species in the near-surface region as
a function of reaction pressure.
[
17]
improving long-term stability.
Furthermore, the rate of reduction of oxidized PdO into met-
allic Pd also depends on the reaction pressure, that is, the con-
centration of organic molecules on the surface that act as re-
ducing agents, such as toluene. Our results show that such re-
duction occurred at a slower rate at low pressures (ꢁ2 bar;
Figure 10). Thus, a significant amount of oxidized PdO species
were still found on the surface of the catalysts that were
tested at 1 and 2 bar compared to those that were tested at
higher pressures. Consequently, the peak for metallic Pd (B.E.=
ples that were catalytically tested at higher reaction pressures
dÀ
(4–8 bar) contained these undesired Pd species in different
dÀ
proportions. The concentration of Pd species increased with
pressure, which is in line with the deactivation behavior of the
catalysts.
Furthermore, the Pd/Ti surface-atomic ratios of the most-
active samples were found to decrease considerably with an
increase in pressure. This result is a consequence of the cover-
335.1 eV) was found to increase quite sharply with increasing
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 189

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
ing of the Pd surface by carbon deposits and/or the migration
of Pd to deeper layers on-stream, thereby leading to a loss in
the surface concentration of Pd, which caused a drop in the
catalyst activity with time. However, there was no loss or leach-
ing of Pd from the bulk catalyst, as evidenced by ICP.
lowed by a subsequent growth during the induction period,
thereby reaching the optimum size and composition faster at
higher reaction pressures, up to a maximum of 6 bar. In addi-
tion, the mean composition of the particles showed a Pd/Sb
ratio of about 5 for those samples that were treated at 1, 2, 4
and 6 bar, whereas a Pd/Sb ratio of about 11 was found for the
particles that were treated at 8 bar, which is much higher than
[10]
2.5. TEM investigation of the morphology and composition
the optimum required ratio.
of the PdÀSb particles
Figure 12 shows TEM images of the most-active PdÀSb/TiO
2
3
. Conclusions
catalysts that were tested at different pressures. The most-
active sample from the experiment at 1 bar (11 h on-stream)
These results provide strong evidence that the reaction pres-
sure, that is, the concentration of reactant molecules on the
catalyst surface, has a notable influence on the catalytic activi-
ty, product selectivity, space-time-yield of the desired product,
induction period, coke deposition, and long-term stability of
the catalyst. Catalytic tests were performed at various reaction
pressures from 1–10 bar. We found that the catalyst formation
was strongly influenced by the reaction pressure, that is, the
surface coverage of the active sites by the reactants, intermedi-
ates, and products. At a pressure of 6 bar, the highest toluene
conversion of 75% and 100% selectivity for benzyl acetate was
observed. This result could be due to both a restructuring of
the catalyst surface and to the simultaneous formation of
active PdÀSb particles of desired size, composition, and shape
during an induction period that is significantly shortened with
an increase in pressure. TEM analysis showed an optimum PdÀ
Sb particle size of 50–100 nm. EDX analysis revealed an opti-
mum Pd/Sb ratio of around 5. XPS confirmed the existence of
dÀ
surface PdO and Pd besides the metallic, bulky Pd that was
x
only seen by XRD. Lastly, in situ XRD showed a decrease in the
crystallinity of the large Pd particles in the parent catalyst and
the growth of new phases (e.g., PdÀSb). In summary, these re-
sults can inspire the improvement of catalyst compositions
and reaction conditions for other partial-oxidation reactions in
the near future.
Figure 12. TEM images of the most-active PdÀSb/TiO
2
solids at different re-
action pressures.
only showed particles with sizes of 10–25 nm. An increase in 4. Experimental Section
reaction pressure led to increased particle size (about 30 nm at
4.1. Catalyst synthesis
2
bar, 40–100 nm at 6 bar). This result first was rather strange
The PdÀSb/TiO catalyst was prepared by impregnation over two
and hard to understand. Why would the lowest reaction pres-
sure lead to the smallest particles? This question was answered
by the in situ XRD experiments, which confirmed the restruc-
turing of the catalyst surface. We now can conclude that, at
the start of the treatment, erosion or dissolution of these large
Pd particles (1–2 mm) occurred, depending on the reaction
pressure, that is, the surface concentration of the reactants.
This is a rather slow process at ambient pressure and leads to
the smallest particles, but, increasing the pressure to 2 bar
sped up the restructuring rate and larger Pd particles (>
2
steps and by calcination in a He atmosphere at 6008C for 4 h. The
Pd and Sb content were set to 10 and 16 wt.% (nominal), respec-
tively. The calcined solid showed Pd particles with sizes of about
1
–2 mm. More details on the catalyst preparation have been report-
[10]
ed previously.
4.2. Characterization of the fresh and spent catalysts
The elemental composition (Ti, Pd, Sb, Na) of the catalysts was de-
termined by using inductively coupled plasma-optical emission
spectrometry (ICP-OES) on a Varian 715-ES ICP-emission spectrome-
ter; data analysis was performed with the ICP Expert software.
Carbon, hydrogen, nitrogen, and sulfur analysis was performed on
a CHNS microanalyzer TruSpec (Leco).
30 nm) could be generated after 6 h on-stream. At 6 bar, this
size of Pd particles (40–100 nm) was observed within 3 h and
this result was similar to the observations at pressures of 8 and
10 bar. These results confirm that a fast collapse of the large
initial particles of the fresh catalysts into very small ones
occurs during the very initial period on-stream, which is fol-
The surface area (BET) and pore-size distribution of the catalysts
were determined by using a Nova 4200e device (Quantachrome In-
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 190

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
struments). Before the measurements, the catalyst was evacuated
AcOH) were based on their molar streams at the inlet and the
for 2 h at 2008C to remove any physisorbed water.
outlet of the reactor. The yields of the formed products (Y) and
i
their selectivities (S) were also calculated from their molar streams.
i
Powder X-ray diffraction (XRD) patterns were obtained by using
a Stoe STADI P diffractometer that was equipped with a linear posi-
tion-sensitive detector (PSD) in transmission geometry with CuKa1
radiation (l=1.5406 ꢂ, 40 kV, 35 mA) in the range 2q=10–558
Beside some carbon oxides from total oxidation and BA as the
main product, only minor amounts of benzaldehyde (BAL) were
observed. The carbon balance was evaluated from unconverted re-
actants and from the yields of the identified and calibrated
products.
(
step width: 0.258) under ambient conditions. The time per step
was 25 s for the samples without Si as a standard and 160 s with Si
as a standard. Processing and assignment of the powder patterns
was performed by using the software Win Xpow (Stoe) and the
powder diffraction file (PDF) database of the International Centre
of Diffraction Data (ICDD).
In general, the most-active samples were taken out of the reactor
after reaching their highest activity and spent samples were taken
out after beginning to become deactivated.
In situ-XRD studies on the parent PdÀSb/TiO2 catalyst were per-
formed by using a Stoe STADI P diffractometer in Bragg–Brentano Acknowledgements
geometry with CuKa radiation (l=1.5418 ꢂ, 40 kV, 35 mA) and
a linear PSD by using a XRK 900 reactor chamber (PAAR, Graz, Aus-
S.G. would like to thank the Deutsche Forschungsgemeinschaft
grant no. BR-1380/13-1) for financial funding. N.M. acknowledg-
tria). The alignment was checked by using a silicon standard. The
investigations were performed by heating the sample under an
inert atmosphere (He) from ambient temperature to 2108C, fol-
lowed by the treatment of the sample at this temperature in an
(
es LIKAT for financial support. The authors would like to thank
Dr. M. Schneider for performing the XRD studies.
acetic-acid/toluene mixture (molar ratio: 4:1) under a flow of air
À1
(
30 mLmin ) at ambient pressure. The data were collected in the
Keywords: acetoxylation · antimony · gas-phase reactions ·
palladium · X-ray diffraction
range 2q=35–508 with a step width of 0.258 and a measurement
time of 30 s per step. The phase composition was determined as
described above.
[
1] K. Weissermel, H.-J. Arpe, Industrielle Organische Chemie, 5th ed., Wiley-
VCH, Weinheim, 1998, p. 252–255.
2] F. Calaza, D. Stacchiola, M. Neurock, W. T. Tysoe, J. Am. Chem. Soc. 2010,
TEM measurements were performed at 200 kV by using a JEM-
ARM200F (JEOL) device that was aberration-corrected by a CESCOR
[
(
CEOS) unit for scanning transmission electron microscopy (STEM)
1
32, 2202–2207.
applications. The microscope was equipped with a JED-2300 (JEOL)
energy-dispersive X-ray spectrometer (EDXS) for chemical analysis.
High-angle annular dark-field (HAADF) and EDXS imaging was op-
erated with a spot size 5c and a 50 mm condenser aperture. For
the TEM measurements, the sample was deposed on a holey
carbon-supported Cu-grid (300 mesh) and transferred onto the mi-
croscope.
[
[
[
3] E. Benazzi, C. J. Cameron, H. Mimoun, J. Mol. Catal. 1991, 69, 299–321.
4] L. Eberson, E. Jonsson, Acta Chem. Scand. Ser. B 1974, 28, 771–778.
5] T. Komatsu, K. Inaba, T. Uezono, A. Onda, T. Yashima, Appl. Catal. A
2003, 251, 315–326.
[6] L. Eberson, L. G. Gonzalez, Acta Chem. Scand. 1973, 27, 1249–1254.
[7] K. Bauer, D. Garbe, H. Surburg, Ullmann’s Encyclopaedia of Industrial
Chemistry, VCH, Weinheim, A 11, 1988, 141.
[
8] V. N. Kalevaru, A. Benhmid, J. Radnik, M. M. Pohl, U. Bentrup, A. Martin,
X-ray photoelectron spectra (XPS) were recorded by using a VG ES-
CALAB 220iXL instrument with AlKa radiation (E=1486.6 eV). The
samples were fixed onto a stainless steel sample holder with
double adhesive carbon tape. The peaks were fitted with Gaussi-
an–Lorentzian curves after Shirley background subtraction. The
electron-binding energy was referenced to the Ti 2p3/2 peak of TiO2
at 458.8 eV. For quantitative analysis of the near-surface region, the
peak areas were determined and divided by the element-specific
Scofield factor and the analyzer-dependent transmission function.
J. Catal. 2007, 246, 399–412.
[9] A. Benhmid, K. V. Narayana, A. Martin, B. Lꢀcke, S. Bischoff, M. M. Pohl, J.
Radnik, M. Schneider, J. Catal. 2005, 230, 420–435.
10] S. Gatla, N. Madaan, J. Radnik, V. N. Kalevaru, M. M. Pohl, B. Lꢀcke, A.
Martin, A. Brꢀckner, Appl. Catal. A 2011, 398, 104–112.
11] J. Radnik, A. Benhmid, V. N. Kalevaru, M.-M. Pohl, A. Martin, B. Lꢀcke, U.
Dingerdissen, Angew. Chem. 2005, 117, 6929–6933; Angew. Chem. Int.
Ed. 2005, 44, 6771–6774.
[
[
[
[
12] V. N. Kalevaru, A. Benhmid, J. Radnik, B. Lꢀcke, A. Martin, J. Catal. 2006,
2
43, 25–35.
13] F. M. T. Mendes, F. B. Noronha, C. D. D. Souza, M. A. P. da Silva, A. B.
Gaspar, M. Schmal, Stud. Surf. Sci. Catal. 2004, 147, 361–366.
4.3. Catalytic tests
[14] C. D. Chang, W. H. Lang, R. L. Smith, J. Catal. 1979, 56, 169–173.
[
[
15] M. Forissier, M. Formenti, J. R. Bernard, Catal. Today 1991, 11, 73–83.
16] H. Hirata, K. Kishita, Y. Nagai, K. Dohmae, H. Shinjoh, S. Matsumoto,
Catal. Today 2011, 164, 467–473.
Catalytic tests were performed in a fixed-bed Hastelloy-C reactor at
[17]
2
108C and a reaction pressure of 1–10 bar. The molar ratio of
À1
Tol/AcOH/O /N was set to 1:4:3:16 (GHSV=2688 h ; GHSV=gas
hourly space velocity). Product analysis was performed by using
[17] N. Madaan, S. Gatla, V. N. Kalevaru, J. Radnik, B. Lꢀcke, A. Brꢀckner, A.
2
2
[17]
Martin, J. Catal. 2011, 282, 103–111.
[17]
off-line GC that was equipped with FID; this method showed
good reproducibility and a relative error of about Æ2.5%. The cal-
culations of the conversion of toluene (X-Tol) and acetic acid (X-
Received: July 31, 2012
Published online on October 26, 2012
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 185 – 191 191