Isotope Studies in Oxidation of Propane over Vanadium Oxide

Article abstract of DOI:10.1002/cctc.201700847

The oxidation of propane has been studied over silica-supported vanadium oxide and polycrystalline, bulk MoVTeNb oxide with M1 structure. Temperature-programmed reaction experiments were performed, and the reactivity of propane molecules labeled with deuterium and ¹³C, respectively, was analyzed under steady-state conditions. The measurement of kinetic isotope effects reveals fundamental differences in the activation of propane over the two catalysts. The reaction network of consecutive and parallel reactions of the formed propylene is comparable. However, oxygen insertion into the CHO group of acrolein under formation of acrylic acid is faster over M1 than oxidation at the CH₂ group and decarbonylation to acetaldehyde. In contrast, the latter process is preferred over silica-supported vanadium oxide resulting in lower selectivity to unsaturated oxygenates.

Full text of DOI:10.1002/cctc.201700847

Accepted Article
Title: Isotope Studies in Oxidation of Propane over Vanadium Oxide
Authors: Pierre Kube, Benjamin Frank, Robert Schlögl, and Annette
Trunschke
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201700847
Link to VoR: http://dx.doi.org/10.1002/cctc.201700847
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
Isotope Studies in Oxidation of Propane over Vanadium Oxide
[
a]
[a, b]
[a]
[a]
Pierre Kube, Benjamin Frank,
Robert Schlögl, and Annette Trunschke*
Abstract: The oxidation of propane has been studied over silica-
supported vanadium oxide and polycrystalline, bulk MoVTeNb oxide
with M1 structure. Temperature-programmed reaction experiments
were performed, and the reactivity of propane molecules labeled with
selectivity. But so far, research was mainly focused on the first
supposedly rate determining step (RDS) that comprises hydrogen
abstraction under formation of a propyl radical (Scheme 1, red)
and a hydroxyl group. Analyses of kinetic isotope effects (KIE)
over supported vanadium oxide revealed that hydrogen in
methylene position is abstracted first affecting the rate of the
1
3
deuterium and C, respectively, was analyzed under steady-state
conditions. The measurement of kinetic isotope effects reveals
fundamental differences in the activation of propane over the two
catalysts. The reaction network of consecutive and parallel reactions
of the formed propylene is comparable. However, oxygen insertion
into the CHO group of acrolein under formation of acrylic acid is faster
[
6]
overall reaction. Consequently, the barrier of the first step
controls activity, although quantum chemical calculations of the
(010) surface of V
2 5
O indicated that subsequent steps such as
the regeneration of the active site by dehydroxylation might
[
7]
over M1 than oxidation at the CH
2
group and decarbonylation to
exhibit high barriers as well.
acetaldehyde. In contrast, the latter process is preferred over silica-
supported vanadium oxide resulting in lower selectivity to unsaturated
oxygenates.
In general, selectivity is affected by favouring specific
pathways compared to others. High barriers in the multistep
pathways to carbon dioxide, i.e., deceleration of the
corresponding reaction(s), may then result in increased selectivity
to propene. Herein we investigate the reaction network in
propane oxidation over bulk mixed MoVTeNb oxide (M1) and a
silica-supported (mesoporous silica SBA-15) vanadium oxide
monolayer model catalyst (6V/SBA-15) by temperature-
programmed reaction and steady-state experiments using isotope
labelled reactants. The reaction conditions are limited to fat, dry
Introduction
The activation of C-H bonds in small alkane molecules has
been widely investigated in chemistry in general and specifically
in heterogeneous catalysis with the intent to use the large natural
gas reserves of our planet economically in the synthesis of
3 8 2
feed (C H /O /He = 10/5/85) and low propane conversion. Under
[
1]
chemicals. Theoretical and kinetic studies of selective oxidation
and oxidative dehydrogenation of alkanes with molecular oxygen
over metal oxide catalysts disclosed an intricate network of
parallel and consecutive reactions that limits the selectivity
these reaction conditions the selectivity to undesired carbon
oxides is higher over silica-supported vanadium oxide compared
[4]
to M1. Factors that determine selectivity in the oxidation of
short-chain alkanes over vanadium oxide catalysts are discussed
based on the comparative analysis of the reaction network over
the two catalysts.
[
2]
towards desired olefins or unsaturated oxygenates. In kinetic
studies of oxidative dehydrogenation (ODH) of propane over
supported vanadium oxide catalysts the reaction network is,
however, usually condensed and presented as
a
triangle
[
3]
(
Scheme 1, black symbols and lines).
Albeit in traces,
oxygenates have been detected in the gas phase, revealing that
a more complex underlying network (Scheme 1, blue symbols)
has implications on the selectivity towards the main products
[
4]
propene, CO, and CO
2
.
The selectivity to propene is highly
dependent on catalyst composition (nature of support, promoters
and modifiers, vanadium oxide loading), catalyst structure
(
molecular structure of vanadium oxide surface species, degree
of oligomerization, degree of hydroxylation, texture of catalyst or
support) and reaction conditions (partial pressures, temperature,
Scheme 1. Simplified reaction network in oxidative dehydrogenation of propane
[
5]
steam content). Detailed knowledge of the reaction network and
its dependence on catalyst structure is the key to control
over vanadium oxide catalysts (black symbols) and minimum extension based
[
4, 8]
on detected gas phase intermediates
(blue symbols) and the supposed
Ÿ
C H surface / gas phase intermediate (red).
3 7
[
[
a]
b]
P. Kube, Dr. B. Frank, Prof. Dr. R. Schlögl, Dr. A. Trunschke
Department of Inorganic Chemistry
Fritz-Haber-Institut der Max-Planck-Gesellschaft
Faradayweg 4-6, 14195 Berlin (Germany)
E-mail: trunschke@fhi-berlin.mpg.de
Current address of Dr. B. Frank
Results and Discussion
Oxidative dehydrogenation of propane was performed in
temperature-programmed experiments by increasing the
temperature stepwise (Fig. 1). Each individual data point
presented in Figs.1a-b is the average of 3-5 measurement points
in steady state at the corresponding reaction temperature. Steady
state confirms that the corresponding products are formed
catalytically. The reaction products in the gas phase were
1
BasCat - UniCat BASF Joint Lab
TU Berlin, Sekr. EW K 01
Hardenbergstr. 36, D-10623 Berlin (Germany)
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
analysed by gas chromatography. Conversion and selectivity as a
function of temperature are shown in Fig. S1. A multitude of
products was observed over both catalysts. The types of products
are the same over M1 and 6V/SBA-15, with the exception of
propionic acid that was only detected over 6V/SBA-15. However,
the relative concentrations and the temperatures, at which the
individual products appear first, differ.
propoxide surface intermediate or whether iso-propoxide as
probable precursor of acetone is formed after re-adsorption of
propene. It has to be noted at this point that the sensitivity of the
thermal conductivity detector (TCD) used for analysis of the
carbon oxides is lower compared to the sensitivity of the flame
ionization detector (FID) applied for analysis of hydrocarbons and
oxygenates. Consequently, the fact that no carbon oxides are
observed below 200°C when acetone is already formed does not
Propene is the first product detectable over both catalysts at a
temperature as low as 120°C. The low temperature clearly
indicates that C-H activation is not the major problem on
vanadium oxide catalysts. At higher temperature (160-180°C)
acetone appears as second product, whereas the concentration
of propene and acetone is higher over M1 compared to 6V/SBA-
[9]
necessarily mean that acetone is no precursor for CO
CO
the other hand, it has been shown that acetone is quite stable
2
,
because
2
is only detectable at a minimum concentration of 1 ppm. On
[8]
under oxidizing conditions over MoVTeNb oxide.
detected first over M1 at 250°C. In contrast, CO appears on
V/SBA-15 with a temperature offset of only 20 K with respect to
acetone already at 200°C and might consequently stem more
likely from total oxidation of acetone as well.
CO
2
is
2
6
1
5. It is not possible to conclude from the present experiments
whether acetone and propene are formed via the same iso-
Figure 1. a) Temperature-programmed oxidation of propane (heating rate 1K/min, measurement isothermal every 20 K) over a) M1 (C
3
H
8
/O
2
/He = 10/5/85, W/F =
-1
-
1
-1
-1
0
.06 g s ml , mcat = 10 mg, Ftotal = 10 ml min ), and b) 6V/SBA-15 (C H /O /He = 10/5/85, W/F = 1.33 g s ml , mcat = 222 mg, Ftotal = 10 ml min ); Arrhenius plots
3 8 2
measured over M1 (c) and 6V/SBA-15 (d) taking into account the overall rate of propane consumption calculated based on sum of the reaction products.
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
[
12]
[13]
The primary formation of carbon oxides (without desorption of
reaction intermediates) cannot be excluded based on the present
experiments. Indication for the latter process has been found on
found recently by experiment, and theory. Such a change in
the shape of the acrolein formation curve is not observed over M1,
which is in agreement with a low concentration of electrophilic
oxygen species on the surface of M1 also at low temperatures
[
8]
[10]
M1, and supported vanadium oxide catalysts.
Here, the
[
14]
propylene selectivity increases with increasing vanadia loading,
which has been attributed to coverage of unselective support
resulting in the dominance of the allylic mechanism.
[
10]
[10b]
It is interesting to note that the shapes of the CO evolution
2
sites,
or to the higher activity,
and/or
a
modified
[
10a]
curves observed over M1 and 6V/SBA-15, respectively, are very
similar. The slopes of the curves change at 300°C indicating that
reducibility
of supported polyvanadate species.
at higher temperatures additional processes contribute to CO
formation.
2
The next group of products comprises acetaldehyde, ethylene,
and CO arising at 220°C (with an offset for CO of + 20 K perhaps
due to the detection limit of the TCD detector) over 6V/SBA-15
and at 300°C over M1 revealing that also the disintegration of the
carbon backbone occurs already at much lower temperature over
6
V/SBA-15 compared to M1. Decomposition pathways are not
directly retrievable from the present experiment. It is interesting to
note that over M1 the appearance of acetaldehyde, ethylene and
CO coincide with acrylic acid formation, whereas a temperature
difference of about 200 K is observed over 6V/SBA-15 between
first detection of these products and the formation of acrylic acid.
The reason might be that acetaldehyde is a common intermediate
[
15]
that may be formed starting from acetone, acrolein, as well as
acrylic acid (Scheme 3, Scheme S1). In particular, the
decomposition pathways of acrolein to CO
ethylene and CO may share a joint C2 surface intermediate
Scheme 3, Scheme S1).
x
and acrylic acid to
x
(
Scheme 2. Low-temperature (a) and high-temperature (b) formation of acrolein,
acetone and propionaldehyde; Low-temperature oxidation via propylene oxide
occurs mainly on silica-supported vanadium oxide; High-temperature allylic
oxidation occurs on both M1 and silica-supported vanadium oxide and requires
The shapes of the acrylic acid evolution curves (steep slope)
are comparable over the two catalysts. Formation starts at 400°C
over 6V/SBA-15. The curve is shifted by 100 K to lower
temperatures over M1 indicating that in contrast to other
oxygenates formation of acrylic acid is particularly facilitated over
M1.
x y
at least V O trimers.
Acrolein also starts to form already at 200°C over 6V/SBA-15.
Acrolein formation occurs here together with CO and with a
2
Acetic acid occurs only above 400°C over both catalysts and
the curves progress in parallel to the acrylic acid formation curves
confirming previous results that acetic acid might be formed by
temperature offset of + 20 K with respect to acetone. Propylene
oxide has been proposed as a common intermediate in the
formation of acrolein and acetone in density functional
[8]
decomposition of acrylic acid. But it cannot be excluded that
[
2i]
calculations applying silsesquioxane models, considering either
isolated vanadyl surface groups or peroxovanadate species as
active sites. Although propene oxide has not been detected in the
gas phase in the present experiment, the nearly simultaneous
appearance of acetone and acrolein over 6V/SBA-15 is not in
contradiction with such pathways. In contrast, over M1 the two
products arise at very different temperatures. While acetone
appears first at 160°C on M1, acrolein is observed only at 270°C
with an offset of + 70 K compared to 6V/SBA-15. The slope of the
acrolein formation curve measured for 6V/SBA-15 changes in the
temperature range in which acrolein formation sets in over M1. It
might be, therefore, possible that different mechanisms contribute
to acrolein formation over 6V/SBA-15, such as a low-temperature
acetic acid is also formed by oxidation of adsorbed C2 precursors
formed in acrolein decomposition (Scheme 3, Scheme S1). C2
and C3 precursors are then apparently oxidized to the
corresponding acids in the same temperature range. Acetic acid
could also be formed by oxidation of re-adsorbed ethylene, which
is formed in C-C bond splitting reactions that occur already at
lower temperatures (see ethylene evolution curves in Fig. 1).
However, this is less likely, since the slope of the ethylene
evolution curves are not affected by the onset of acetic acid
formation (Fig. 1). Acrylic acid is either more slowly formed or
faster decomposed to acetic acid over 6V/SBA-15 compared to
M1, because over the former catalyst the concentration of acetic
acid is higher than the concentration of acrylic acid whereas the
reverse situation is observed over M1.
(
200-300°C) oxidation pathway catalysed by activated oxygen
species at the surface, such as peroxovanadate, via propylene
oxide as intermediate followed by oxidative dehydrogenation
Remarkably, allyl alcohol, which is an intermediate of propene
allylic oxidation, is still detectable over both catalysts at
temperatures above 400°C when the molecule reaches sufficient
high concentration in the gas phase revealing that allylic oxidation
is a common pathway over both catalysts.
(
Scheme 2a), and the allylic oxidation of propylene at higher
temperatures (>300°C) (Scheme 2b). In the low-temperature
range acetone and propionaldehyde may be formed as
[
11]
isomerization products of propylene oxide.
Indications for the
participation of peroxide species in oxidative dehydrogenation of
propane over supported vanadium oxide catalysts have been
Finally, a pathway via adsorbed n-propoxide resulting in
propionaldehyde (occurrence above 300°C) and propionic acid
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
(
detectable only over 6V/SBA-15 above 400°C) is operative over
changes in the reaction network do not occur in the studied
temperature range at points where new products appear in the
gas phase meaning that all detected products might belong to a
single complex reaction network branching out from propane.
Slight deviation occurs at low temperatures over 6V/SBA-15 for
the reasons discussed above (Scheme 2). The apparent
activation energies (Figs. 1c-d) agree well with published values
both catalysts as well. Formation of propionaldehyde at low
temperature by isomerization of propylene oxide cannot be
excluded.
The rate of overall propane consumption calculated based on
the sum of products has been plotted as a function of 1/T (Figs.
1
c-d). The function is overall linear for both catalysts (R² = 0.9985
[4, 8]
[16]
for M1
and SBA-15-supported vanadium oxide, respectively.
for M1, and R² = 0.9949 for 6V/SBA-15) indicating that significant
Scheme 3. Reaction network in propane oxidation outlined based on intermediate products detected in the gas phase (blue) over M1 and 6V/SBA-15 in dry feed
/O /He = 10/5/85) in the temperature range 100-450°C during temperature-programmed propane oxidation; Furthermore, the results of the experiments with D-
(
C
3
H
8
2
1
3
13
and C-labelled propane (the red colour indicates the C-labelled carbon atom) have been integrated into the Scheme; A version of the Scheme that highlights
major principal similarities and differences in the reaction networks over 6V/SBA-15 and M1 is presented in the Supporting Information ( Scheme S1).
In summary, the temperature-programmed experiments
reveal that a multitude of reaction products (11 over M1 and 12
over 6V/SBA-15) are detectable by gas chromatography in the
temperature range 100-450°C and consequently involved in the
reaction network of propane oxidation over both, M1 and 6V/SBA-
presented in Scheme 3. The experimentally proven reaction
intermediates are highlighted in blue. Surface intermediates have
been proposed based on spectroscopic studies of selective and
[
17]
total oxidation catalysts. The type of products detected and the
shape of the product evolution curves are strikingly similar at
normal reaction temperatures of propane oxidation (300-400°C)
suggesting that the same global reaction network predominates
over the two catalyst systems, but the rate constants of individual
1
5 catalysts. A hypothetical reaction network that includes all
intermediates detected so far in the gas phase under the specific
reaction conditions applied in the present experiments is
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
reaction steps differ. However, at temperatures below 300°C
differences are observed in particular with respect to acrolein
formation. Hydrogen-abstraction under formation of propylene
occurs already at very low temperature over both catalysts
followed by oxyhydration/oxidation to acetone or allylic oxidation
to acrolein. Acrolein formation and cleavage of C-C bonds are
reactions particularly favoured over 6V/SBA-15, while acrylic acid
formation is favoured over M1.
further decreased yielding an overall KIE of 2.31. This means that
the hydrogen atoms in methylene as well as methyl position are
simultaneously involved in the consumption of propane by rate
determining process(es). Parallel formation of iso-propoxide that
requires C-H activation in methylene position and n-propoxide
that requires C-H activation in methyl position should not be
responsible for the observed effect, since propionaldehyde is also
formed over 6V/SBA-15 to a similar extent (compare Figs. 1a-b,
and S1), but here the deuteration of the methyl group has no
impact.
Based on the assumption that the entire reaction network in
propane oxidation over M1 and 6V/SBA-15 is similar, differences
in the selectivity of the overall reaction result from differences in
the activation energy of individual reaction steps. In the following,
we have analysed kinetic isotope effects (KIE) based on rates of
overall propane consumption and propylene formation. If the
abstraction of the first hydrogen atom in adsorbed propane is rate
limiting, the rate of propane consumption must be affected by
deuteration. To verify whether hydrogen abstraction in methylene
position is the rate-limiting step of the overall propane oxidation,
Table 2. Kinetic isotope effects (KIE) in propane consumption and propylene
formation calculated based on the rates provided in Table 1.
Catalyst
Isotopes
KIE (C3)
1.49
KIE (C3=)
1.72
CH
CH
3
3
CH
CD
2
2
CH
CH
3
3
/
/
/
/
CH
3
CD
2
CH
3
3
1.01
1.50
1.49
1.55
2.31
2.84
4.87
2.13
1.10
2.34
6
V/SBA-15
CD CD CD
propane consumption rates were measured using C
6, respectively. Similarly, the impact of splitting a C-H bond
3 8
in methyl position was studied comparing 2,2-C and C D ,
3
H
8
and 2,2-
3
2
3 2
C D H
CH
CD
3
3
CH
CD
2
2
CH
CD
3
3
3
D
2
H
6
respectively. Rates extrapolated to zero conversion determined in
contact time variations at T=400°C (Figs. S2-S3) have been used
CH
CH
3
3
CH
CD
2
2
CH
CH
3
3
(
Tab. 1) to calculate the KIE (Tab. 2).
In agreement with isotope studies of propane ODH over
CH CD CH /
3
2
3
[
6]
M1
CD CD CD
3
2
3
zirconia supported vanadium oxide catalysts, the substitution of
H by D in methylene position affects the overall consumption rate
of the substrate over 6V/SBA-15 (Tab. 1) resulting in a significant
KIE of 1.45 (Tab. 2) and revealing that C-H activation of
methylene hydrogen in propane affects the propane consumption
rate also over silica-supported vanadium oxide. Deuteration in
methyl position has no effect resulting in a KIE of 1 again in
agreement with the literature.
CH
CD
3
3
CH
CD
2
2
CH
CD
3
3
/
The results may be interpreted in terms of a different structure
of the intermediate in propylene formation on M1 compared to
6
V/SBA-15 that leads to similar barriers for the abstraction of the
first and the second hydrogen atom, respectively. For better
illustration, the part of the reaction network (Scheme 3) that
seems to predominate propylene formation is enlarged in
Scheme 4 and the differences between M1 and 6V/SBA-15 are
highlighted in blue and green, respectively. Over M1 the limiting
case might be a simultaneous mechanism, which seems to be not
Table 1. Propane consumption rates, propylene formation rates, and
selectivity to propylene over the two catalysts obtained in contact time
-
1
-1
variation experiments (W/F = 0.02-0.06 g s ml for M1, and 0.30-1.34 g s ml
for 6V/SBA-15) in the feed C /O /He = 10/5/85 at T=400°C with different
3
H
8
2
isotopes after extrapolation to zero propane conversion (raw data shown in
the Supporting Information, Figs. S2-S3).
unlikely over Mo oxide containing catalysts according to DFT
-
7
-7
r
C3,0*10
r
C3=,0*10
[18]
catalyst
Reactant
S
C3=,0
calculations using Mo
3
O
9
model clusters.
Our experimental
-
1
-1
-1 -1
[mol g s ]
[
mol g s ]
numbers agree well with calculated KIEs from a 2 + 4 pathway in
which both methyl and methylene C-H bonds are simultaneously
involved. However, the calculated barrier for the corresponding
transition state (48.3 kcal/mol) is higher compared to the
CH
3
CH
2
CH
3
3
3
3
3
3
1.86
1.25
1.24
20.8
14.0
9.0
1.61
87
75
27
89
62
88
6
V/SBA-15
CH CD CH
0.94
3
3
3
2
2
2
[
4]
experimentally determined value of 34 kcal/mol. The 2 + 4
mechanism corresponds to a concerted heterolytic C-H activation
CD
CH
CD
CH
CD
CH
0.33
(
two-electron process) in which all in all 6 electrons are involved
18.5
simultaneously including both metal atom and oxygen atom of a
M=O species as active site. The one-electron hydrogen
M1
CH CD CH
8.7
3
3
2
2
abstraction under formation of
a radical is, however, the
CD
CD
CD
7.9
energetically most feasible pathway in propane activation over
[
18]
[19]
Mo
3
O
9
model clusters,
and also over vanadium oxide.
Therefore, quasi-simultaneous H-abstraction on neighbouring
M=O sites (Scheme 4) might be taken into consideration as an
explanation for the KIEs measured over M1. Such a mechanism
has not been studied by theory so far. The high density of
In contrast, both, deuteration of the methylene as well as the
methyl group affects the rate of propane consumption over the
M1 catalyst resulting in a KIE for methylene of 1.49 and for
methyl of 1.55. In case of fully deuterated propane the rate is
[4]
propane adsorption sites on the surface of M1 might render
such a mechanism possible. Interestingly, a high density of
5
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
oxidising sites has been reported to favour propylene formation
concentration of electrophilic oxygen species and adsorbed
oxygen on the surface of M1. The equal selectivity in the
experiments with native and fully deuterated propane,
respectively, also support that consecutive reactions of propylene
instead of C-C bond cleavage reactions on vanadium-based
[
20]
oxides.
Furthermore, we analysed the rates of propylene formation
albeit interpretation of the corresponding data is less
straightforward, since formation of propylene requires the
abstraction of two hydrogen atoms that must occur via several
elementary steps. Interestingly, the selectivity to propylene is
affected in a different way by isotope labelling over 6V/SBA-15
and M1 (Tab. 1). Consecutive reactions of the formed propylene
are neglected in the following discussion of the difference, since
the selectivity to propylene SC3=,0 was determined based on
extrapolation to zero propane conversion. Still, parallel reactions
of propane may affect propylene formation rates at zero
conversion, but the selectivity of propylene approaches
approximately 90% at X=0% in the reaction of native propane
over both catalysts due to parallel formation of n-propoxide and
its further oxidation (Tab. 1). By using fully deuterated propane
the selectivity to propylene drops to 27% compared to the
reaction of native propane (87% propene selectivity) over
(
at least reactions that involve C-(H,D) splitting, such as allylic
oxidation) are indeed negligible under the applied boundary
conditions of zero propane conversion. In contrast, selectivity
decreases by 27% over M1 when propane is deuterated in 2-
position. The difference between M1 and 6V/SBA-15 is easily
plausible when
a simultaneous H-abstraction mechanism is
assumed over M1 (Scheme 4), which is in agreement with the
KIEs measured over M1 and the much higher number of propane
[4]
adsorption sites on the surface of M1 compared to 6V/SBA-15.
Upon exchange of
H by D in 2-position a simultaneous
mechanism is no longer possible leading to faster abstraction of
methyl hydrogen that results in the formation n-propyl species as
precursor of saturated oxygenated products and carbon oxides
(
Schemes 3-4).
6
V/SBA-15. After abstraction of the first D atom in the reaction of
, a C intermediate is formed on 6V/SBA-15 that reacts in
the second step either to fully deuterated propene by another D
abstraction at the neighbouring methyl group (k in Scheme 1) or
to undesired by-products via an unspecified pathway (k in
C
3
D
8
3 7
D
1
2
Scheme 1). The ratio of the corresponding two rate constants
determines the selectivity to propene in the isotope exchange
experiments in the limiting case of zero per cent conversion of
propane. The decrease in propylene selectivity in the conversion
of fully deuterated propane over 6V/SBA-15 points at a reaction
3 7
pathway competitive to selective C-D activation in the C D
intermediate, which is not decelerated by isotope exchange, i.e.,
which does not involve D-C bond splitting. Competitively,
reactions with gas-phase oxygen via alkylperoxy and
[
21]
hydroperoxyalkyl radicals, or fast attack of adsorbed oxygen at
the carbon backbone of C in the adsorbed state may occur
Scheme 4). By using CH CD CH instead of native propane the
selectivity to propylene is only slightly affected over 6V/SBA-15
Tab. 1), because only the first step from CH CD CH to
CH CDCH , which determines activity, but not selectivity, involves
D abstraction. Propylene is formed in the subsequent step by
Scheme 4. Differences in the formation mechanism of propylene over silica-
supported vanadium oxide (green) and M1 (blue) analysed based on KIE
measurements.
3
D
7
(
3
2
3
(
3
2
3
Table 3. Propane conversion and product selectivity in dependence of steam
-
1
3
3
addition (10 vol%) to the feed at 400°C, W/F = 0.06 g s ml (M1), and 1.34 g
-
1
3 8 2
s ml (6V/SBA-15) and C H /O /He = 10/5/85.
abstraction of H from the methyl group of the CH
3
CDCH
/k
3
M1
6V/SBA-15
intermediate, like in case of native propane. However, the k
1
2
ratio may be influenced by secondary isotope effects resulting in
slightly diminished selectivity of 75% compared to 87% for native
propane (Tab. 1). Trends in the KIE’s determined based on
propylene formation rates (Tab. 2) are in agreement with this view.
H
2
O
D
2
O
H
2
O
2
D O
X [%]
1.8
1.8
2.0
2.1
r
propane
consumption,0
1 -1
-
M1 exhibits
a different behaviour indicating that only
(
mmol g h )
4.81
4.82
0.247
0.255
deuteration of the methylene C-H bonds facilitates formation of
undesired by-products (Tab. 1). As discussed above, the KIE
determined over M1 based on propane consumption rates at zero
conversion is invariant with regard to the methylene or methyl
position. That means that complete deuteration of propane
retards the overall reaction, but will not interfere the structure of
the transition states and intermediates, which is reflected in equal
selectivity to propylene for native propane (89%) and fully
deuterated propane (88%) at zero propane conversion (Table 1).
Other than on 6V/SBA-15, destruction of the carbon skeleton of
S (propene) [%]
S (acrylic acid) [%]
S (acetic acid) [%]
S (C2) [%]
65.8
30.0
1.3
70.4
25.0
1.5
77.8
0.2
5.3
0.2
9.9
6.7
78.1
0.2
5.2
0.2
9.7
6.7
0.1
0.1
S (CO) [%]
2.0
2.2
S (CO
2
) [%]
0.8
0.9
the C
the
3
D
7
intermediate does not occur preferentially. No change in
ratio (Scheme 1) is in agreement with low
k
1
/k
2
a
6
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
1
3
It should be noted that re-establishment of the active catalyst
surface after hydrocarbon oxidation (after first half of the catalytic
cycle) can also influence the KIE. The rate constant for re-
oxidation of an oxygen vacancy on the catalyst surface, which is
does not differ between native and C-2-propane. The propylene
selectivity measured over the two catalysts was only slightly
affected by isotope labelling.
In the experiments with the unlabelled reactant, CO and CO
are formed in almost equal amounts with some excess of CO,
which is reflected in a selectivity ratio CO/(CO+CO ) slightly
greater than 0.5 (Table 4). The overall CO/(CO+CO ) ratio
2
2
formed during formation of H O from two OH groups, should be
the same for (partially) deuterated and native propane. However,
the (formal) rate constants of catalyst dehydration likely differ
among each other depending on the degree of deuteration of the
hydrocarbon substrate, as deuteration of the surface depends on
the deuterium content of the substrate and dehydration of the
surface in the final step of the catalytic cycle involves necessarily
the cleavage of an O-H/O-D bond. The impact of catalyst re-
2
2
changes only slightly when using labelled propane, resulting in
values of 0.54 for 6V/SBA-15 and 0.48 for M1. The observed
deviations are within the error of carbon oxides detection. The
proposed reaction network (Scheme 3) features a multitude of
reaction pathways that finally yield both CO and CO
2
. But when
the C-labelled reactant is pulsed, distinct differences in the
origin of CO and CO over the two catalysts become obvious
Scheme 5).
oxidation on the reaction rates was investigated by adding H
O to the feed of native propane and oxygen under steady state
conditions to modify the concentration of surface OH or OD
groups (Tab. 3). Switching from H O to D O does not change the
2
O or
1
3
D
2
2
(
2
2
activity of both catalysts, indicating that catalyst dehydration does
not affect KIEs obtained with deuterated propane. The selectivity
pattern is not affected over 6V/SBA-15 in a measurable way. For
M1, selectivity of acrylic acid is slightly higher in H
2
O compared to
D
2
O at the expense of all other products. This indicates direct
involvement of water in a kinetically relevant step of acrylic acid
formation or vital H/D exchange in the intermediate product
propene, which results in deceleration of allylic oxidation as a
consequence of the KIE for C-H/C-D bond cleavage. In any case
the observation implies the existence of an additional kinetically
relevant step in the reaction network, which is located within the
path from propylene to acrylic acid.
Scheme 5. Predominant origin of carbon oxides over M1 (blue) or 6V/SBA_15
(green); Line thickness of arrows reflects the selectivity presented in Table 4.
Over M1 the carbon in methylene position contributes slightly
more to CO formation compared to the carbon in methyl position
The formation pathways of carbon oxides during propane
13
13
13
as indicated by the higher CO/( CO+ CO
2
) ratio of 0.75,
1
3
oxidation were studied by means of pulse experiments with C-
labelled/unlabelled propane/oxygen mixtures at T=400°C.
whereas CO preferentially originates from one of the terminal
2
A
carbon atoms (Table 4, Scheme 5). The opposite pattern is
schematic illustration of the experimental procedure is provided in
the Supporting Information (Scheme S2). The applied labelled
observed over 6V/SBA-15, i.e., CO originates mainly from the
2
methyl position and CO from the methylene position, respectively.
1
3
propane contains the C atom in 2-position (labelled in red in
Schemes 3, 5). The products of the pulse experiments were
analysed by mass spectrometry after pre-separation of the
The observation that the majority of CO
2
formed over M1 is
unlabelled indicates that the intermediate to CO
C
3
2
product mixture into permanent gases (CO, CO
2
, and O
2
),
predominantly bears a terminal functional group that contains two
oxygen atoms (e.g., a carboxylic acid group). The C3 acid is
acrylic acid, since propionic acid is not formed over M1 under the
applied reaction conditions (Fig. 1a). The elimination of unlabelled
hydrocarbons (C , C ), and acids by packed columns (for
3
H
6
3 8
H
experiment description see Supporting Information). Propane
conversion and product selectivity are summarized in Table 4.
The conversion of propane has been adjusted for the two
catalysts by choosing appropriate contact times. The conversion
CO
2
(decarboxylation) from acrylic acid gives a residual C2
fragment.
1
3
3 8 2
Table 4. Propane conversion and product selectivity for native and C-substituted propane in propane oxidation in a feed of C H /O /He = 10/5/85 at T=400°C; W/F
-
1
-1
=
0.06 g s ml (M1), and 1.34 g s ml (6V/SBA-15).
1
2
12
13
13
S
S
12
13
Catalyst
M1
Reactant
Feed
/O
X (C
[%]
3
H
8
)
S (C
[%]
3
H
6
)
S ( CO)
[%]
S ( CO
2
)
S ( CO)
S ( CO
2
)
CO
CO
(
C
3
)
C
3
2
/He
[%]
[%]
[%]
S
+ S12 CO2
S
+ S13CO2
CO
12
13
CO
C
3
H
8
10 / 5 / 85
10 / 5 / 85
10 / 5 / 85
10 / 5 / 85
4.0
4.0
4.6
4.5
69.4
68.1
65.3
62.8
15.9
7.2
14.6
13.8
15.4
9.1
-
-
0.52
0.34
0.55
0.63
-
1
3
C-2-C
3
3
H
H
8
8
8.2
-
2.8
-
0.75
-
6
V/SBA-15
C
3
H
8
19.2
15.3
1
3
C-2-C
4.7
8.2
0.36
Acetaldehyde, labelled either at the carbonyl group or at the
above, ethylene might be a quite stable end product.
methyl carbon, and ethylene are the C2 products primary formed
either from acrolein or acrylic acid (Scheme 3). As discussed
Acetaldehyde can be either oxidized at the methyl group and
subsequently decarbonylated to give adsorbed methoxy species
7
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
or it is further oxidized to acetic acid by oxygen insertion into the
C-H bond of the CHO group. Acetaldehyde, which is labelled at
the carbonyl group, yields labelled CO by decarbonylation and
an unlabelled surface methoxy group. The latter may be
oxidized to formate species, which finally form CO by
labelled surface methoxy group is further oxidized to labelled CO
2
or CO , respectively (Scheme 2, bottom).
All decomposition reactions of the C3 oxygenates acrolein
and acrylic acid outlined in Scheme 3 may occur in parallel, but
the product distribution observed over M1 and 6V/SBA-15 in the
temperature-programmed experiments (Fig. 1) and the
decarbonylation or CO
Scheme 2, top).
2
by decarboxylation, both unlabelled
(
1
3
distribution of the C isotope in the final combustion products
CO and CO (Table 4, Scheme 3) indicate differences in
1
3
The labelled C atom in acetic acid formed over M1 is
2
primarily located in the methyl group, as indicated by mass
spectrometry (Fig. 2). The mass spectrum of pure native acetic
acid (Fig. 2a) displays two major fragments corresponding to a
reaction rates of the individual steps. The high concentration of
acrylic acid in the product mixture and the preferential formation
2
of unlabelled CO suggest that over M1 oxidation of acrolein to
+
+
carboxylate CO
3), which are also present in the pre-separated acetic acid
fraction of the product mixture observed in oxidation of native
2
H species (m/z 45) and a CH
3
CO species (m/z
acrylic acid via oxygen insertion into the C-H bond of the CHO
group is faster than oxidation of the vinyl group followed by
decarbonylation of the resulting surface species and
4
1
3
+
13
+
propane (Fig. 2b). Peaks due to CH
3
CO or CH
(m/z 46) are observed when
labelled C-2-propane is pulsed. The measured intensity ratio of
these three peaks can be simulated by
ratio of 3:1 indicating preferred
3
CO (m/z 44), acetaldehyde formation. In contrast, decarbonylation of acrolein
+
13
+
CO
2
H
(m/z 45) and CO
2
H
is preferred over 6V/SBA-15 as indicated by the increased
fraction of unlabelled CO. For clarity, the corresponding section
of Scheme 3 is enlarged in Scheme 6 and the major differences
between M1 (blue) and silica-supported vanadium oxide (green)
are highlighted.
1
3
a
1
3
13
3 2 3 2
CH CO H:CH CO H
1
3
formation of acetic acid with the C atom located at the methyl
position. This means that either acrolein or acrylic acid is
preferentially attacked at the terminal methylene group, but
attack at the central carbon is also possible, but less likely.
Scheme 6. Major differences in consecutive reactions of acrolein over M1
1
3
(
blue) and 6V/SBA-15 suggested by the distribution of C in the formed
carbon oxides.
Similarly, over M1 the common intermediate acetaldehyde
seems to be faster oxidized at the CHO group under formation
of acetic acid. In contrast, over 6V/SBA-15 the methyl group of
acetaldehyde is faster oxidized followed by decarbonylation of
the corresponding surface intermediate.
An interpretation of the increased abundance of labelled CO
2
over 6V/SBA-15 is less straightforward. The decomposition of
adsorbed acetone may contribute more significantly to labelled
2
CO over 6V/SBA-15, whereas acetone formed over M1 is more
stable. In addition, differences in the decomposition of the final
surface formate (Scheme 2, bottom right) either via
decarboxylation or decarbonylation may contribute to the
observed pattern.
Figure 2. Mass spectrum of 2.6vol-% acetic acid in He (a), mass spectrum of
the acid fraction of the product mixture in the oxidation of native propane (b),
1
3
-1
and C-2-propane (c) on M1 at T=400 °C, W/F = 0.3 g s ml and C
3 2
/O /He =
2
0/5/75 (blue) and 15/5/80 (red); The propane content in the feed was
increased in the corresponding experiments to increase the reliability of acetic
acid analysis.
Summary and Conclusions
In summary, the measurement of kinetic isotope effects in
propane oxidation revealed essential differences in the
activation of the propane molecule over M1 and silica-supported
Acetic acid that contains a labelled methyl group releases
unlabelled CO
2
in a decarboxylation reaction. The residual
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
vanadium oxide, respectively. Whereas over silica-supported
vanadium oxide the abstraction of the first hydrogen atom at the
This implies that oxygen atoms, which are part of surface oxide
species, are involved in the corresponding reaction.
3 8
methylene group of the C H molecule is rate limiting, activation
Our results confirm that control over occurrence and
distribution of the various oxygen species on the surface of
vanadium oxide catalysts is the key issue in terms of selectivity.
Selective oxidative dehydrogenation of propane to propylene
requires the suppression of any electrophilic oxygen species on
the catalyst surface, which might be solved by pulsed operation
of propane over M1 clearly involves both, methyl and methylene
groups, simultaneously in rate-limiting process(es). The
difference has no significant impact on the network of
consecutive reactions of the formed propylene, but it may
[
4]
explain the lower barrier measured over M1,
and the
corresponding difference between M1 and 6V/SBA-15 in
propane consumption rate that comprises one order of
[
24]
under alternate oxidizing and reducing conditions.
The
[
4]
implementation of selective acrolein formation is more
challenging, because it requires electrophilic oxygen species
and faces the competitive oxidation of propylene in 2-position
under formation of acetone and the consecutive oxidation to
acetaldehyde that occur in the same temperature range. A more
detailed analysis of the impact of reaction parameters on the
rates within the network is required to find solutions. The
formation of acrylic acid may be controlled via appropriate
design of the solid-state chemistry of the catalysts in terms of
oxygen species that selectively insert into the C-H bond of a
terminal CHO group. At this point we refer to the difference
between silica-supported vanadium oxide and M1 in the onset
temperature of acrylic acid formation. However, high selectivity
to acrylic acid requires in addition and primarily the optimization
of acrolein formation as well as the suppression of consecutive
magnitude. At usually applied reaction temperatures (300-
00°C) the overall reaction network in propane oxidation seems
4
to be similar over silica-supported vanadium oxide and M1 with
the allylic oxidation of propylene being the predominant pathway
since the intermediate allylic alcohol has been detected over
both catalysts. The final product of the allylic oxidation over M1
is acrylic acid, while the reaction essentially terminates with
acrolein formation over silica-supported vanadium oxide.
Acetone formation via oxihydration or oxidation of propylene is a
minor parallel pathway. Carbon oxides are mainly formed by
decomposition of the main selective oxidation products acrolein
and acrylic acid, which are preferentially attacked at the terminal
methylene group under formation of the common decomposition
product acetaldehyde. Differences in the preferred degradation
pathways of acetaldehyde have been revealed by pulse
1
3
reactions such as the oxidative attack at the CH
acid.
2
group of acrylic
experiments using
C-2-propane. Over silica-supported
vanadium oxide acetaldehyde seems to be preferentially
oxidized at the methyl group followed by decarbonylation,
whereas oxidation to acetic acid followed by decarboxylation is
the favoured reaction path over M1.
Experimental Section
The most important differences in the reaction network of
propane oxidation under the studied conditions over the bulk
catalyst M1 and the monolayer catalyst 6V/SBA-15 have been
detected in the formation mechanism of propylene (Scheme 4)
and in the consecutive reaction of acrolein (Scheme 6). The
differences in the reaction network over the two catalysts
originate from differences in the surface concentration and
distribution of oxygen species under reaction conditions. Based
on the temperature-programmed experiments the formed
products can be classified into olefins, carbonyl compounds, and
acids. Propylene is formed at very low temperatures indicating
clearly that the strength of the C-H bond is no particular hurdle in
propane activation in general. The result is in contrast to the
popular opinion in oxidation catalysis that the thermodynamic
stability of C-H bonds in substrate molecules and intermediates
Catalysts
Synthesis, and detailed bulk as well as surface analysis of the two
[
4]
catalysts have been described elsewhere.
Propane oxidation experiments
Catalytic measurements were carried out in a self-constructed reactor
setup with plug-flow characteristics using 10 or 50 mg (M1) and 223 mg
(
6V/SBA-15) catalyst. SiC was used to dilute the M1 catalyst for adjusting
the same bed length as for 6V/SBA-15. The reactor (inner diameter 4
mm) was equipped with a thermocouple for measuring the temperature
inside the catalyst bed. C
3
H
8
(Westfalen), C
3
D
8
(Campro Scientific 99
2 6
H (Sigma Aldrich 98
atom% D, Sigma Aldrich 99 atom% D), 2,2-C
3
D
13
13
2
atom% D), C-2-propane (ICON Isotopes 99 atom% C), O (Westfalen)
were used as reactants. The product gas mixtures were analyzed by
online gas chromatography (Agilent 7890) and online mass spectrometry
[
22]
determines the selectivity. Homolytic hydrogen abstraction
has been proposed to be the favoured reaction path in propane
1
3
(QMA 400, Pfeiffer Vacuum). Separation of the gas pulses during C-2-
propane experiments were achieved by using a self packed column
[
19, 23]
activation over vanadium oxide.
nucleophilic oxygen and the vanadyl oxygen has been
The activation requires
TM
TM
(
Porapak Type Q 100-120 Mesh or Carbopack B-DA 80-120 Mesh)
[
19b, 19c]
located between reactor outlet and analytics. More details concerning the
performed experiments are provided in the Supporting Information
considered to be the most active one.
The formation of
acrolein as the major carbonyl compound in the product mixture
occurs also at comparatively low temperatures perhaps via
attack of propylene by electrophilic oxygen species, which seem
to be more abundant on the surface of silica-supported
vanadium oxide compared to M1 due to the different electronic
(
Scheme S2 und corresponding explanation). Starting and reference
point for all measurements were a reaction temperature of 400°C and a
dry feed of 10% native propane, 5% oxygen and 85% He. H O and D O
2
2
(Sigma Aldrich 99.9 atom % D) were introduced into the reactant stream
via vaporizer at 140°C. The calculation of propane conversion and
product selectivity includes a correction for gas impurities in the initial gas
mixture (namely labelled propylene in labelled propanes). Isotopic
scrambling in propane in the gas phase as well as H/D exchange with the
catalyst surface can be neglected (Supporting Information, Fig. S5).
[
4]
structure of the two catalysts as outlined in detail recently.
Finally, formation of acids requires particularly high
temperatures, especially over silica-supported vanadium oxide.
9
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
a
[
[
10]
X. Gao, J.-M. Jehng, I. E. Wachs, Journal of Catalysis 2002, 209, 43-
Acknowledgements
b
5
0; A. Khodakov, B. Olthof, A. T. Bell, E. Iglesia, Journal of Catalysis
999, 181, 205-216.
1
This work was conducted in the framework of the BasCat
collaboration between BASF SE, TU Berlin, FHI, and the cluster
of excellence “Unified Concepts in Catalysis” (UniCat
www.unicat.tu-berlin.de).
11] Z. Zhai, X. Wang, R. Licht, A. T. Bell, Journal of Catalysis 2015, 325,
7-100.
8
[12] A. Held, J. Kowalska-Kuś, K. Nowińska, Journal of Catalysis 2016, 336,
23-32.
[
[
13] V. I. Avdeev, A. F. Bedilo, Research on Chemical Intermediates 2016,
4
a
2, 5237-5252.
14]
K. Amakawa, Y. V. Kolen'ko, A. Villa, M. E. Schuster, L.-I. Csepei, G.
Weinberg, S. Wrabetz, R. Naumann d'Alnoncourt, F. Girgsdies, L. Prati,
Keywords: propane • oxidation • reaction network • kinetic
b
R. Schlögl, A. Trunschke, ACS Catalysis 2013, 3, 1103-1113; E.
isotope effect • vanadium oxide
Sadovskaya, V. Goncharov, G. Popova, E. Ishchenko, D. Frolov, A.
Fedorova, T. Andrushkevich, Journal of Molecular Catalysis A:
c
a
b
Chemical 2014, 392, 61-66; C. Heine, M. Hävecker, M. Sanchez-
[
[
1]
2]
C. Coperet, Chemical Reviews 2010, 110, 656-680; C. A. Gärtner, A.
Sanchez, A. Trunschke, R. Schlögl, M. Eichelbaum, The Journal of
Physical Chemistry C 2013, 117, 26988-26997.
C. van Veen, J. A. Lercher, ChemCatChem 2013, 5, 3196-3217.
a
M. M. Bettahar, G. Costentin, L. Savary, J. C. Lavalley, Applied
b
[15] L. Bui, R. Chakrabarti, A. Bhan, ACS Catalysis 2016, 6, 6567-6580.
[16] A. Dinse, S. Khennache, B. Frank, C. Hess, R. Herbert, S. Wrabetz, R.
Schlögl, R. Schomäcker, Journal of Molecular Catalysis A: Chemical
Catalysis, A: General 1996, 145, 1-48; D. L. Stern, R. K. Grasselli,
c
Journal of Catalysis 1997, 167, 560-569; M. Lin, T. B. Desai, F. W.
d
Kaiser, P. D. Klugherz, Catalysis Today 2000, 61, 223-229; R. Fushimi,
2
009, 307, 43-50.
S. O. Shekhtman, A. Gaffney, S. Han, G. S. Yablonsky, J. T. Gleaves,
a
e
[17]
E. Finocchio, G. Busca, V. Lorenzelli, R. J. Willey, Journal of the
Industrial & Engineering Chemistry Research 2005, 44, 6310-6319; R.
b
Chemical Society, Faraday Transactions 1994, 90, 3347-3356; A. A.
Grabowski, Catalysis Reviews: Science and Engineering 2006, 48, 199
f
Davydov, Molecular spectroscopy of oxide catalyst surfaces, John
-
268; A. Trunschke, in Nanostructured Catalysts: Selective Oxidation
c
Wiley & Sons Ltd., Chinchester, 2003; P. Concepción, P. Botella, J. M.
Reactions, 1 ed. (Eds.: C. Hess, R. Schlögl), RSC Nanoscience &
d
g
L. Nieto, Applied Catalysis A: General 2004, 278, 45-56; Y. Wang, W.
Nanotechnology, Cambridge, 2011, pp. 56-95; K. Alexopoulos, M.-F.
h
Zheng, F. Chen, X. Zhan, Applied Catalysis A: General 2008, 351, 75-
Reyniers, G. B. Marin, Journal of Catalysis 2012, 289, 127-139; K.
e
8
1; D. Shee, G. Deo, Journal of Molecular Catalysis A: Chemical 2009,
Alexopoulos, M.-F. Reyniers, G. B. Marin, Journal of Catalysis 2012,
f
i
308, 46-55; J. M. Lopez Nieto, B. Solsona, P. Concepcion, F. Ivars, A.
2
3
95, 195-206; J. Liu, F. Mohamed, J. Sauer, Journal of Catalysis 2014,
17, 75-82.
g
Dejoz, M. I. Vazquez, Catalysis Today 2010, 157, 291-296; R. Lopez-
Medina, I. Sobczak, H. Golinska-Mazwa, M. Ziolek, M. A. Banares, M.
[
[
3]
4]
K. Chen, A. T. Bell, E. Iglesia, The Journal of Physical Chemistry B
000, 104, 1292-1299.
O. Guerrero-Perez, Catalysis Today 2012, 187, 195-200.
2
[
[
18] G. Fu, X. Xu, X. Lu, H. L. Wan, Journal of Physical Chemistry B 2005,
09, 6416-6421.
P. Kube, B. Frank, S. Wrabetz, J. Kröhnert, M. Hävecker, J. Velasco-
1
Vélez, J. Noack, R. Schlögl, A. Trunschke, ChemCatChem 2017, 9,
a
19]
H. Fu, Z.-P. Liu, Z.-H. Li, W.-N. Wang, K.-N. Fan, Journal of the
5
a
73-585.
b
American Chemical Society 2006, 128, 11114-11123; X. Rozanska, R.
[
5]
C. A. Carrero, R. Schloegl, I. E. Wachs, R. Schomaecker, ACS
b
Fortrie, J. Sauer, The Journal of Physical Chemistry C 2007, 111, 6041-
Catalysis 2014, 4, 3357-3380; I. Muylaert, P. Van Der Voort, Physical
6
050; cX. Rozanska, R. Fortrie, J. Sauer, Journal of the American
Chemical Society 2014, 136, 7751-7761.
[20] E. V. Kondratenko, M. Y. Sinev, Applied Catalysis A: General 2007,
25, 353-361.
Chemistry Chemical Physics 2009, 11, 2826-2832; cC. Hess,
d
ChemPhysChem 2009, 10, 319-326; F. Cavani, N. Ballarini, A.
e
Cericola, Catalysis Today 2007, 127, 113-131; B. M. Weckhuysen, D.
3
E. Keller, Catalysis Today 2003, 78, 25-46.
[
[
[
[
21] J. Zádor, C. A. Taatjes, R. X. Fernandes, Progress in Energy and
Combustion Science 2011, 37, 371-421.
[
[
6]
7]
K. Chen, E. Iglesia, A. T. Bell, Journal of Catalysis 2000, 192, 197-203.
P. C. Redfern, P. Zapol, M. Sternberg, S. P. Adiga, S. A. Zygmunt, L. A.
Curtiss, The Journal of Physical Chemistry B 2006, 110, 8363-8371.
R. Naumann D'Alnoncourt, L. I. Csepei, M. Hävecker, F. Girgsdies, M.
E. Schuster, R. Schlögl, A. Trunschke, Journal of Catalysis 2014, 311,
22] C. Batiot, B. K. Hodnett, Applied Catalysis A: General 1996, 137, 179-
1
91.
[
[
8]
9]
23] F. Gilardoni, A. T. Bell, A. Chakraborty, P. Boulet, Journal of Physical
Chemistry B 2000, 104, 12250-12255.
3
69-385.
24] O. Ovsitser, R. Schomaecker, E. V. Kondratenko, T. Wolfram, A.
Trunschke, Catalysis Today 2012, 192, 16-19.
G.-L. Dai, Z.-H. Li, J. Lu, W.-N. Wang, K.-N. Fan, The Journal of
Physical Chemistry C 2012, 116, 807-817.
1
0
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700847
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
COMMUNICATION
Roadmap to value or waste: The
reaction network of propane oxidation
over vanadium oxide catalysts was
analysed by using temperature-
programmed reactions and
Pierre Kube, Benjamin Frank,
Robert Schlögl, Annette Trunschke*
Page No. – Page No.
Isotope Studies in Oxidation of
Propane over Vanadium Oxide
1
3
experiments with D- and C-labelled
propane molecules. Propylene
formation and oxidation of acrolein
strongly depend on the nature of the
catalyst. Surface density of vanadyl
groups and nature of electrophilic
oxygen species are relevant with
regard to selectivity.
1
1
This article is protected by copyright. All rights reserved.