Dynamics of Bulk Oxygen in the Selective Oxidation of Acrolein

Article abstract of DOI:10.1002/cctc.201700124

The gas-phase oxidation of acrolein to acrylic acid on a hydrothermally prepared mixed oxide catalyst was investigated by steady-state isotopic transient kinetic analysis (SSIKTA) as well as different types of concentration-programmed techniques (CPR-pulse, CPO) under in situ or process-relevant conditions. Balancing the amounts of active oxygen gives an overview of the quantities of participating bulk oxygen species. The dynamics of bulk oxygen lead to re-oxidation processes on the catalyst surface and thus influence the selectivity pattern of the network of acrolein oxidation. Furthermore, the bulk dynamics are activated by temperature.

Full text of DOI:10.1002/cctc.201700124

Accepted Article
Title: Dynamics of Bulk Oxygen in the Selective Oxidation of Acrolein
Authors: Maurice Heid, Stefan F. Knoche, Niklas S. A. Gora, Dominik
Ohlig, Alfons Drochner, Bastian J. M. Etzold, and Herbert G.
Vogel
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.201700124
Link to VoR: http://dx.doi.org/10.1002/cctc.201700124
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
Dynamics of Bulk Oxygen in the Selective Oxidation of Acrolein
Maurice Heid, Stefan Knoche, Niklas Gora, Dominik Ohlig, Alfons Drochner*, Bastian J.M. Etzold
[
a]
Herbert Vogel
[
3d, 3e]
Abstract: Gas phase oxidation of acrolein to acrylic acid on a
hydrothermal prepared mixed oxide catalyst was investigated by
steady state isotopic transient kinetic analysis (SSIKTA) as well as
others,
containing the metals molybdenum, vanadium and
tungsten. However, it is known that the active site needs the
interplay of molybdenum and vanadium, while the addition of
different types of concentration programmed techniques (CPR-Pulse, tungsten leads to more stable catalysts. Tungsten also has an
[5]
CPO) under in situ or process relevant conditions. Balancing the
amounts of active oxygen gives an overview of the quantities of
participating bulk oxygen species. The dynamics of bulk oxygen
leads to re-oxidation processes on the catalyst surface and thus
influences the selectivity pattern of the network of acrolein oxidation.
Furthermore, the bulk-dynamics is activated by temperature.
impact on the oxygen-mobility through the catalyst but does not
take part in the redox-chemistry.
[
3g, 6]
From a global perspective, the mechanism of the partial
oxidation of ACR on mixed oxides follows the Mars-
[
7]
van Krevelen concept,
which means that the reactant is
oxidised by lattice oxygen and the reduced catalyst is re-
oxidised by gas-phase oxygen.
For oxidations on mixed oxides it is generally shown, that
the pathways of re-oxidation can be quite different. On the one
hand, the reduced active-site can be re-oxidised directly by gas-
phase oxygen. On the other hand, it can be re-oxidised by bulk
oxygen while the catalyst itself is re-oxidised at a separate
Introduction
[
1]
Acrylic acid (AA) is an important chemical intermediate. It is
produced industrially by the partial oxidation of propene by a
two-step heterogeneously catalysed process. The acrolein
[8]
active-site. In the latter case, the oxygen has to be transported
[9]
from one site to another. A measurement of this oxygen
(
ACR) formed in the first step is converted into AA on modified
transport under process relevant conditions is not trivial. It is well
known that under steady state conditions, the selective catalyst
Mo/V/W mixed oxide catalysts in the second step. Currently, a
selectivity of over 90 % can be achieved with industrial
optimised catalyst systems. Carbon oxides, such as CO and
surface is partially reduced and so the number of surface- and
[10]
bulk-vacancies is different to the initial ones.
Thus, it is
CO
AA.
2
are formed as by-products from the combustion of ACR and
important to measure the oxygen transport under process
relevant conditions.
[
2]
Generally, the oxygen transport can be measured by
different methods, e.g. by secondary ion mass spectrometry
(
SIMS) on probes which are labelled with isotopic oxygen
[
11]
[12]
before,
programmed methods.
via x-ray imaging,
or different temperature
[
13]
However, only few experiments
performed in situ are mentioned in literature. One example is the
oxygen diffusion in MgO at different degrees of surface
reduction, where MgO was used as model catalyst for the
[
14]
oxidative methane dimerization.
Another example is the work
of Sokolovskii et al., who studied the oxygen transport of various
mixed oxides at different degrees of reduction, with model
reductants like CO or H .
2
Scheme 1. Reaction network of ACR oxidation.
[
15]
In previous works we showed some oxygen pathways
within the mechanism of ACR oxidation on Mo/V/W mixed
In the past a lot of work has been done to identify and
[
3]
characterise the active catalytic phases. Often discussed are
[
2, 6b, 13b, 16]
[
4]
[3a, 3b]
oxides.
Predominantly results from transient response
structure types like Mo
5
O
14
,
Mo
4
O
11
,
6
Mo O21 among
methods (TRM), especially in combination with isotopic
assessment, enabled the observation of a fast oxygen exchange
[
2,
between ACR as reactant and oxygen from the catalyst lattice.
6
b, 16b-e]
[
a]
M. Heid, S. Knoche, N. Gora, D. Ohlig, Prof. Dr.-Ing. A. Drochner,
Prof. Dr.-Ing. B. J.M. Etzold, Prof. Dr.-Ing. H. Vogel
Ernst-Berl-Institut für Technische und Makromolekulare Chemie
Technische Universität Darmstadt
This exchange reaction takes place via an equilibrium
forming intermediately a surface acetal (ACR + 2 M-OH
surface acetal + H O). From this key-intermediate the oxidation
2
of ACR to an acrylate species or AA takes place subsequently,
leaving an oxygen vacancy at the catalyst surface, which is re-
oxidised by gas phase or bulk oxygen.
Alarich-Weiss-Straße 8, 64287 Darmstadt (Germany)
Fax: (+49)6151-16-23703
E-mail: h.vogel@tc1.tu-darmstadt.de
Furthermore, results form SSTIKA and CPR led to the
hypothesis that the degree of surface reduction significantly
influences the selectivity pattern of the network of ACR
[
2, 13b, 16a, 16d]
oxidation.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
At this point it should be emphasised, that the previously
mentioned results are all based on measurements on spray-
The stoichiometry of the resulting catalyst measured via
XPS and ICP shows a slightly deviation from the target
stoichiometry Mo V W O , which belongs to the stoichiometric
8 2 0.5 x
[
6b, 13b, 16]
dried mixed oxides.
Recently it was found, that the
performance of hydrothermal prepared catalysts is significantly
higher, which will be introduced within this contribution. The aim
of this study is to investigate the participation of bulk oxygen in
the ACR oxidation on such hydrothermal prepared catalysts in
more detail. Both, SSITKA and CPR, especially as pulse-
technique, are used in combination for balancing the amount of
active oxygen under technical relevant or in situ conditions,
respectively. Furthermore, temperature depended activation of
the dynamics or participation of bulk oxygen as well as its
influence on the catalytic performance is focussed.
composition of the initial precursor mixture (see experimental
section).
Via the preparation route, nanoscale non-porous needle-
like particles (Figure 1) are received. The particles exist of
smaller rods. The rods have diameters between 40 and 100 nm
and the needles are between 200 nm and 2 µm in length. The
specific surface area (BET, see experimental section) of the
2
-1
catalyst is 34 m g .
Results and Discussion
As mentioned before, a lot of work has been done concerning
[3]
the catalyst solid-science or structure. However, this study
focuses on the dynamics of bulk oxygen and its participation in
the catalytic cycle. In order to get more insight into the reaction-
mechanism of the partial oxidation of acrolein, especially the
[
17]
dynamics of the bulk, a model catalyst is needed.
requirements a model catalyst must meet are:
Some

simple composition (compared to the industrial
catalyst)



well-defined surface to bulk ratio
thermal and redox stability
high conversion and selectivity at industrial relevant
conditions
One catalyst that fits all of the above criteria is the hydrothermal
prepared Mo (21.5 < x < 30.5) mixed oxide. It was
8
2 0.5 x
V W O
characterised to be non porous, nano-cristalline and has shown
a high performance in the partial oxidation of acrolein as it is
shown below.
Model catalyst and its steady state performance
The knowledge of the exact stoichiometry of the catalyst is
needed for further investigation (oxygen balances, see below).
Therefore, XPS (surface) and ICP (overall composition) analysis
were made (Table 1).
Figure 1. SEM-pictures of the hydrothermal prepared mixed-oxide catalyst
with the formal stoichiometry Mo V W O .
8 2 0.5 x
Based on the observed particle shape, the catalyst fulfils
the prerequisite of a well-defined surface to bulk ratio. This
allows a more detailed investigation on the influence of the bulk
for the selective oxidation of ACR.
Table 1. Composition of the catalytic material, measured with XPS and ICP.
Normalised to the stoichiometric index of molybdenum of eight.
XRD measurements after preparation, before and after
Measurement
XPS
Mo
8
V
W
certain treatments (pre-oxidised with O
2
and reduced with ACR),
1.8
1.5
1.1
0.6
verified the redox stability of the catalyst (Figure 2).
ICP-OES
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
[16d]
prepared mixed oxides (Mo
8
2
V W
1.5
x
O ).
However, the spray-
dried oxide has two major disadvantages: On the one hand, its
2
-1
relative low BET-surface (4 m g ), and on the other hand, the
habitus of the particles. They are partially broken hollow spheres,
and so do not have a well-defined surface to bulk ratio, which
makes conclusions about the bulk participation quite difficult.
Nevertheless, the main findings for the spray-dried oxide are
[2, 6b, 16d]
summarised as follows:


The sum of all components remains constant when
1
6
18
O
2
is replaced by
2
O , i.e. no kinetic isotope effect
takes place.
ACR exchanges its carbonyl oxygen with oxygen
from the catalyst. Consequently, all isotopologues
are formed.
Figure 2. XRD of the hydrothermal prepared mixed oxide catalyst
(
8 2 0.5 x
Mo V W O ). As prepared, after pre-oxidation and after reduction with ACR.


Oxygen exchange takes place via a surface-acetal
(
ACR).
The catalytic performance in terms of conversion (X) and
AA also exchanges its oxygen with the catalysts
surface.
selectivity (S) from steady state experiments is given in Figure 3.
1
8
It should be emphasised that the
2
O -SSITKA experiment
in Figure 4 shows the result of the hydrothermal model catalyst
and that the main findings are in accordance with those from the
spray-dried catalyst.
Figure 3. Conversion and selectivity from steady state measurements for the
oxidation of ACR to AA using a hydrothermal prepared Mo/V/W mixed oxide.
Composition of the model catalyst was Mo
8 2 0.5 x
V W O . Feed consists of 5 vol. %
Acr and 10 vol. % O (balance He). The catalyst mass was 50 mg and the total
2
-
1
volume flow was 20 ml min (STP).
Under the chosen conditions (5 vol. % ACR and 10 vol. %
O
2
), the catalyst shows a performance of 83 % conversion and
89 % selectivity at 300 °C. Thus, the performance of the model
catalyst is in the range of an industrial one.
All in all, this catalyst allows a more detailed investigation
on the role of bulk oxygen in the partial oxidation of ACR.
1
8
2
O -SSITKA
SSITKA (steady state isotopic transient kinetic analysis) give
insights into the mechanism of catalytic reactions under real
[
16d, 16e]
18
steady state conditions.
were performed (exchange of
Here,
2
O -SSITKA experiments
1
6
18
O
2
against
2
O under global
steady state) to get information about the participation of bulk
oxygen in the selective oxidation of ACR.
18
In previous works
2
O -SSITKA experiments within ACR
oxidation have already been presented with spray-dried
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
which is the only remaining oxygen-source. One possible
mechanistic explanation is illustrated in Scheme 2.
Scheme 2. Illustration of the bulk dynamics and the re-oxidation by gas phase
18
oxygen during a SSITKA-experiment with
2
O .
1
8
After adsorption and dissociation, the O can change its
1
6
[18]
place with a O from the bulk in a hopping-like mechanism,
16
resulting in a transport of O to the surface. So, more and more
O can be removed from the catalyst by substitution with O.
This is proofed by the O-balance, which is positive and in the
same quantity like the missing O.
1
6
18
1
8
1
6
To get a concrete number of the bulk oxygen participation,
16
the removed amount of O-oxygen (nO,rem) is related to the
amount of oxygen from the catalyst (nO,cat). In this context, there
are two interesting values to look at. The first is the potentially
available oxygen (nO,cat) and the second is the removable
amount via redox processes (nO,redox). Both can be calculated
from the stoichiometry of the catalytic material and its mass. For
the used model catalyst the high oxidised stoichiometry is
Mo
8 2 0.5 8 2 0.5
V W O30.5 and the reduced one is Mo V W O21.5, this is
+
VI
+IV
related to the oxidation states of the metals (Mo Mo
,
+
V
+IV [3g]
V
V ). So, the available oxygen in the highest estimated
-3
oxidation state (nO,cat) is about 1.05 ∙ 10 mole and for the lowest
estimated oxidation state 0.742 ∙ 10 mole (related to 50 mg of
1
8
Figure 4. Concentration profiles of an
O
2
-SSITKA-experiment with 5 vol. %
-3
-
1
ACR and 10 vol. % O
Mo ). After 600 s
replacement the formation of all isotopologues of the products can be
2
at 300 °C. Total volume flow 20 ml min (STP). (50 mg
1
6
1
8
the model catalyst). The difference between these two values is
8
V
2
W
0.5
O
x
O
2
is replaced by
2
O . Soon after this
-3
the amount of oxygen (0.3 ∙ 10 mole) accessible via redox
processes (nO,redox). The latter is used for the CPR-Pulse
experiments later, while the first is of interest for SSITKA.
The experimental identified amount (Figure 4) of removed
18
observed. ACR is formed due an exchange reaction with oxygen from the
surface.
1
8
1
6
-3
Additionally, the
2
O -SSITKA reveals an opportunity to
O-oxygen (nO,rem) was 0.12∙10 mole. This was removed from
balance the quantities of elemental oxygen. Therefore, one has
the catalyst within ten minutes at 300 °C. At a first glance, this
does not seem very much but with respect to the overall amount
of oxygen in the catalyst, this is about 11 % of the oxygen in the
16
18
to calculate the elemental balances for O as well as for O.
More precisely, here the elemental balance is defined as
difference between the amount of substance entering the reactor
and leaving it.
-1
catalytic material (nO,rem ∙ nO,cat ) and 38 % of the oxygen which
-1
can take part in the redox processes (nO,rem ∙ nO,redox ). The
results for SSITKA experiments at different temperatures are
listed in Table 2.
16
By balancing the amount of substance for O during the
2
phase, it can be observed that this balance is negative. This
1
8
O
1
6
means more O leaves the system than enters it. It follows that
16
the missing amount of O originates from the catalyst bulk,
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
1
6
Table 2. Amount of removed O from 50 mg Mo
8
V
2
W
0.5
O
30.5 in the selective
-SSITKA within ten minutes at different
temperatures. nO,cat is the whole amount of oxygen in the catalyst. nO,redox
belongs to the oxidation states of the different metals
Mo Mo ;V  V ; W does not take part in the redox processes).
1
8
oxidation of ACR during an
O
2
+VI
+IV +V
+IV
[3g]
(
270 °C
300 °C
330 °C
-
6
n
O,rem / 10 mol
58.5
0.06
0.19
118
0.11
0.38
140
0.13
0.45
-
1
n
O,rem ∙ nO,cat
-1
n
O,rem ∙ nO,redox
16
With an increasing temperature the amount of
O
removed from the bulk within ten minutes increases. So, this is a
kinetic effect where the bulk oxygen transport is activated by
temperature. This shows the dynamics of the catalyst itself and
the participation of bulk oxygen under steady state conditions.
The results are in accordance with those from modelling of
[
16d]
isotopologues pattern of previous SSITKA experiments.
CPR-Pulse experiments
Reduction experiments also give the opportunity for balancing
the removed oxygen from the catalyst. Especially the pulse-
technique with a variation of the pulse-distance provides insight
in the dynamics of this process. So, Concentration Programmed
Reduction-Pulse (CPR-Pulse) experiments with varying pulse-
distances were performed at different temperatures with regard
to selectivities of the individual reactions. These experiments
show the importance of the degree of surface reduction for the
selective oxidation of ACR and the oxygen mobility of the
catalyst’s oxygen.
The raw data of such an experiment are given as volume
fraction for each component at a given time. The resulting
volume fraction was transformed into a molar flow with the ideal
gas law. Subsequent integration over time of each pulse results
in the amount of substance of each component for each pulse.
With these and the fed amount of ACR per pulse (verified by
blank runs) the performance for each pulse is calculated. A
typical response curve for a CPR-Pulse experiment as well as
the integrated amount of substance for each component and the
performance of the catalyst is shown in Figure 5.
-
6
Figure 5. Top: Response of a typical CPR-Pulse experiment. 1.15∙10 mole
ACR per pulse, pulse-distance 60 s, temperature 300 °C and catalyst mass
100 mg Mo
8 2 0.5 x
V W O . Middle: Integrated amount of substance per pulse for
each component. Bottom: Calculated conversion (X), selectivity (S) and yield
(Y) for the CPR-Pulse experiment.
The initial state of the catalyst is non-selective, caused by
its high coverage with oxygen. The first five pulses only show
combustion products. In the region of pulses ten to 18 the
catalyst is very active and high selective. Needless to say that
with an increasing amount of pulses the activity decreases due
to the limited oxygen capacity of the catalyst. At pulse 50 the
conversion decreases to values lower than 22 %. As a result of
this experiment the selectivity towards AA depends on the
[
2]
degree of surface reduction (see also ).
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
In the next step we will follow the question whether the
16
-6
Table 4. Amount of removed O within 50 ACR pulses (each of 1.15∙10
bulk oxygen participates, and if yes in which quantity, during the
reduction experiments or not. In this context some prerequisites
and assumptions will be given for calculating the relation
between the amount of removed oxygen (nO,rem) and oxygen on
the surface of the catalyst (nO,s). The surface oxygen can be
estimated as follows.
mole ACR) from the catalyst. Temperature 300 °C and 100 mg
Mo V W O
8 2 0.5 x.
Pulse-distance
n
O,rem
n
O,rem
∙ nO,redox.s
n
O,rem
-1
-6
-1
/
s
/ 10 mol
∙ n
O,s
6
9
1
0
64.5
65.2
69.2
76.1
79.8
0.54
0.55
0.59
0.64
0.67
1.85
1.86
1.98
2.18
2.28
2
-1
The specific surface area of our model catalyst is 34 m g .
Additionally, we assume that the catalyst exists of ideal crystals
0
with the unit cell of a Mo
this is one of the active phases in the ACR oxidation.
5
O14 phase. As described in literature,
20
[
4]
Furthermore, the smallest length of this unit cell is about 0.4 nm
and can serve as the thickness of one surface-layer. With this
and the BET surface, the volume of a surface-layer can be
180
40
2
[
6b, 19]
calculated.
The ratio of this volume and the volume of the
catalyst (see experimental section) should be the same as the
ratio between the oxygen on the surface (nO,s) and the whole
oxygen of the catalyst (nO,cat). It should be emphasised that this
is only a concept which serves as a model. With these
assumptions and the amount of oxygen in the catalytic material
With an increasing time between two pulses the removed
oxygen increases (Table 4). At higher pulse-distances the bulk is
able to transport more oxygen to the surface and to provide this
for the oxidation of ACR. Additionally, it can be observed that the
selectivity towards AA decreases with an increasing time
between two pulses. This demonstrates the effect of surface-
oxygen on the selectivity. Due to a higher coverage of oxygen
the selectivity decreases with an increasing pulse-distance. The
higher the pulse-distance, the lower is the selectivity towards AA.
Consistent with this, the conversion of acrolein increases with an
increasing pulse-distance. The earliest detection of AA is at
pulse number five. This is due to the fact that the catalyst in its
initial state is covered with “non-selective” oxygen and the
formed acid is oxidised directly towards the combustion products.
This was cross-checked in additional experiments with AA
pulses.
(
described in the SSITKA section) one can calculate,
analogically to the SSITKA experiments, the proportion of
removed oxygen from the total surface-oxygen and the oxygen
available via redox processes by this CPR-Pulse experiment
(
Table 3).
8 2 0.5 x
Table 3. Amount of oxygen removed from 100 mg Mo V W O via CPR-
-6
Pulse experiment within 50 ACR pulses (each of 1.15∙10 mole ACR) at
00 °C and 60 s pulse-distance. nO,s stands for the overall oxygen from one
3
surface-layer (assumed thickness 0.4 nm). nO,redox,s stands for the potentially
available oxygen from one surface-layer via redox processes
+VI
+IV +V
+IV
[3g]
(
Mo Mo ;V  V ; W does not take part in the redox processes).
As shown in Figure 6, the theory fits the experimental
measured data. In other words, with a variation of the pulse-
distance two theses are verified: On the one hand, the re-
oxidation of the surface by bulk oxygen and, on the other hand,
the connection between the degree of surface reduction and the
selectivity of the catalyst.
n
O,rem
n
O,s
n
O,redox,s
n
O,rem
n
O,rem
∙ nO,redox.s
-
6
-6
-6
-1
-
1
/
10 mol
/ 10 mol
/ 10 mol
∙n
O,s
6
4.6
118
34.9
0.54
1.85
The removed oxygen in the CPR-Pulse experiment with 50
ACR pulses (at 300 °C with a pulse-distance of 60 s) is more
than the available oxygen from one surface-layer accessible via
redox processes. This means that the overall amount of
participating oxygen species of the catalyst is higher than the
available amount of oxygen in the surface-layer. Thus, it is
demonstrated that the bulk transports oxygen to the surface and
provides this oxygen for the ACR oxidation.
Additionally, CPR-Pulse experiments with different pulse-
distances were performed to look at the dynamics of bulk
oxygen and the influence on the selectivity in more detail. Here,
the distances between two pulses were varied between 60 and
240 s. The temperature for the reduction was kept at 300 °C.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
Figure 7. Selectivity of carbon monoxide for each pulse at different pulse-
8 2 0.5 x
distances at 300 °C and 100 mg Mo V W O . Initial amount of ACR per pulse
-
6
1
.15∙10 mole.
These results give hint for at least two active sites – one
for the selective oxidation and another for the combustion
towards CO. However, the second active site is either not in
contact with the bulk or its re-oxidation by bulk oxygen is too
slow at the investigated temperatures.
Additionally, the elemental balances were checked,
defined as the ratio between the overall detected amount of an
element leaving the reactor and entering it. The carbon balance
(
Figure 8) is overall lower than one, which means that more
carbon is entering the reactor than leaving it. So, there must be
some kind of formation of coke or undetected products. However,
the elemental balance shows a minimum in the range of pulse
six and nine. After that, it increases with an increasing number of
pulses and reaches a level in the range of 90 %, decreasing with
higher pulse-distance.
Figure 6. Conversion and selectivity for each pulse at different pulse-
8 2 0.5 x
distances. Temperature 300 °C and 100 mg Mo V W O . Initial amount of
-
6
ACR per pulse 1.15∙10 mole.
By looking at the response curves for the remaining
products (AcOH, CO, CO , H O), one additional phenomenon
2
2
can be observed: The formation of carbon monoxide shows an
interesting trend which is independent from the pulse-distance.
The formation of carbon monoxide can be detected until the first
ten pulses. With higher pulse numbers no considerable amount
of CO can be detected (Figure 7).
Figure 8. Carbon balance for each pulse at different pulse-distances at 300 °C
-
6
and 100 mg Mo V W O . Initial amount of ACR per pulse1.15∙10 mole.
8
2
0.5
x
As mentioned above, the carbon balance increases with a
decreasing pulse-distance. According to the theory of a more
non-selective oxidation with higher pulse-distances, the carbon
balance is lower and so the formation of coke or undetected
products is higher (more non-selective oxidation). However,
there is some carbon missing. To clarify the disposition of
missing carbon, the catalysts were re-oxidised for one hour at
4
(
00 °C with 10 vol. % oxygen (CPO) after the reduction pulses
Figure 9).
The amount of removable carbon from the catalyst during
a CPO is nearly the same for all pulse-distances. Nevertheless,
with higher pulse-distance the formation of undetected
a
products or irreversible coke-deposits increases. By calculating
the ratio between the overall missing carbon (difference between
missing carbon after pulses and the detected carbon after CPO)
and the fed carbon, the carbon-recovery is accessible. The
experiment with a pulse-distance of 60 s has a carbon-recovery
of more than 99 %, while the experiment with a distance of 240 s
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
has a carbon-recovery of 92 %. So, the formation of a reversible
coke-layer is independent from the pulse-distance, while the
formation of undetected products and/or irreversible coke
depends on the distance. However, this also confirms the result
of a more non-selective oxidation at a higher oxidised surface.
be seen that for each temperature an increase of the pulse-
distance leads to more removable oxygen.
Due to the different types of experiments, steady state
reaction (SSITKA) where the oxygen is replaced and not really
removed like it is in the reduction experiments (CPR-Pulse), the
overall amount of removed oxygen from the catalyst is quite
different. Nevertheless, the SSITKA-values can serve as a
reference for the CPR-Pulse experiments and also proof the
result of a kinetic effect in the bulk oxygen transport.
Conclusions
The dynamics of bulk oxygen in the selective oxidation of
acrolein were investigated on a Mo/V/W mixed oxide by using
two different transient response methods under in situ conditions.
For the first time a comparable study of the bulk-dynamics under
steady state and under reductive conditions has been performed.
In summary, the results of SSITKA-experiments and CPR-
Pulse experiments evidence that bulk oxygen plays an important
role in the partial oxidation of ACR. It is shown that the transport
of oxygen from the bulk to the surface of the catalyst can be
measured with these kinds of experiments. Thus, a calculation
of the participating active oxygen is feasible.
Figure 9. Integrated amount of carbon in dependence of pulse-distance.
Square: Missing carbon after 50 ACR reduction pulses. Triangle: Detected
8 2 0.5 x
carbon during re-oxidation (CPO). 100 mg Mo V W O . Initial amount of ACR
-
6
per pulse1.15∙10 mole.
In addition to this, it is shown that a more reduced surface
leads to a more selective but less active oxidation. So, there
exists an optimal degree of surface reduction, low for high
selectivity but high enough for sufficient activity. However, the
degree of surface reduction is influenced by bulk oxygen, so the
bulk oxygen transport is essential for the catalytic performance.
Moreover, the formation of undetected products and/or
irreversible coke-layer also depends on the degree of surface
reduction or the bulk oxygen transport, respectively.
With these experiments the participation of bulk oxygen in
the selective oxidation is evidenced. It is shown that the
transport of oxygen from the bulk to the surface of the catalyst is
measurable.
Additionally, the CPR-Pulse experiments were performed
at different temperatures (Figure 10).
Furthermore, these experiments give hint for at least two
active sites: A selective one, which is in contact with the bulk
and one for the combustion to carbon monoxide, which is (at the
investigated temperatures) not re-oxidised by bulk oxygen.
Thus, the CPR-Pulse experiments can give access to the
kinetics of the bulk-processes under reaction near conditions.
For this reason, CPR-Pulse experiments are a powerful tool to
characterise different oxygen contents and mobility under more
reaction relevant conditions.
Experimental Section
Catalyst preparation and characterisation:
The catalyst with the general formula Mo
8 2 0.5 x
V W O (21.5<x<30.5) was
Figure 10. Removed oxygen from the catalysts by 50 ACR-pulses at different
prepared via hydrothermal synthesis according to the patent specification
temperatures and pulse-distances. 100 mg Mo
ACR per pulse1.15∙10 mole.
8
V
2
W
0.5 x
O . Initial amount of
[20]
DER10 2012 207 811 A1.
An aqueous solution containing ammonium
-
6
heptamolybdate (11.3 g) and ammonium metatungstate (1.1 g) were
dissolved in distilled water (120 ml). Vanadyl sulfate hydrate(4.2 g) was
also dissolved in distilled water (120 ml). These two solutions were mixed
and adjusted to a pH value of 2.0 with sulfuric acid (9 M) and transferred
With an increasing temperature, the removable oxygen
from the catalyst increases. This result is in accordance with the
2
in an autoclave (teflon inner tube). After bubbling with N for ten minutes
-
1
1
8
the stirred reaction mixture was heated up to 448 K with 5 K min and
maintained at this temperature for 24 h under hydrothermal conditions.
one from the
2
O -SSITKA experiments where the bulk oxygen
transport is also activated with temperature. Additionally, it can
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
After cooling the crude product was filtered, washed with water and taken
Kinetic investigations:
-
1
up in 120 mL of a 0.4 mol l oxalic acid solution. The obtained
suspension was stirred for 1 h at 353 K. Subsequent the solid was
filtered, washed with distilled water again and dried overnight at 353 K.
Reactor apparatus:
All experiments were carried out on a self-constructed apparatus for
kinetic investigations, which allows fully automatic changes in
temperature- and concentration-programmed reactions (flow sheet see
Supporting Information). It consists of three main parts: a gas dosing unit,
a reactor and an online analytic. The gas dosing was realised with an
arrangement of several mass flow controllers, multiway-valves and two-
stage gas saturators, allowing a flexible dosage of gaseous and liquid
components. Additionally, a 6-way-valve with a 1 mL sample loop allows
pulse-wise dosing of components. A quartz U-tube (internal diameter
The BET specific surface area of the catalyst was measured from the N
2
sorption isotherms collected on a Quantachrome Quadrasorb at 77 K.
Prior to the measurement, the sample was degassed at 300 °C until a
-
5
stable vacuum of approximately 1.3∙10 bar was reached. The catalyst
was characterised to be non-porous with a specific surface area of
2
-1
3
4 m g (Figure 11).
4
mm) served as the reactor, in which the catalyst was fixed between two
quartz wool stoppers. The reactor was placed in an electrically heated
oven, which was temperature controlled. The reaction gas was analysed
online via mass spectrometry (GAM 400, InProcess Instruments,
Germany). All experiments were performed under continuous flow
-
1
(
overall volume flow 20 mL min (STP)) at ambient pressure. The
common catalyst masses were 50 or 100 mg. Helium was used as inert
carrier gas.
Catalyst pre-treatment:
Before the measurements, the catalyst was pre-treated for 1 h at 400 °C
with oxygen (10 vol. % in He). The purpose of this oxidative pre-
treatment was to achieve a specific oxidation and reproducible initial
state.
Steady State Performance:
After the pre-treatment, the catalyst was cooled down to 100 °C in a
-
1
constant helium flow of 20 mL min (STP). After reaching this
temperature the gas was switched to feed composition (5 vol. % ACR
and 10 vol. % oxygen) and was kept until steady state concentrations
were detected (several hours). Then subsequently the temperature was
raised to the next level. The experiment was performed at 100, 150, 200,
2 8 2 0.5 x
Figure 11. Results of the N -physisorption measurement of Mo V W O .
SEM data were obtained with an FEI XL30 FEG microscope from
Philipps with an acceleration voltage of 10 kV.
225, 250, 270, 300, 330, and 360 °C.
XRD patterns were collected with an Empyrean diffractometer from
PANalytical equipped with a CuK radiation (=0.15 nm), at a step width
of 0.013° in the (2 ) range from 5° to 70°.
SSITKA experiments:
Following the pre-treatment, the catalyst was cooled down to 270 °C in a
-
1
constant helium flow of 20 mL min (STP). Then, the reaction gas (5
vol. % ACR and 10 vol. % oxygen) was lead over the catalyst until steady
concentrations were detected (several hours). Data acquisition was
XPS data were obtained with an SSX 100 ESCA electron
spectrometer from Surface Science Laboratories Inc. using 90 W AlK
-
9
16
18
radiation. The base pressure was about 3 ∙ 10 mbar.
2 2
initialised and O was replaced by O after 10 min and vice versa after
20 min. The measurement was terminated after in total 30 min. The
ICP-OES data were obtained with an ICP Optima 300
spectrometer from Perkin Elmer. Prior to the measurement 20 mg
catalyst were solved in 20 mL 1 M NaOH. The resulting solution was then
temperature was increased to the next level with a heating rate of 2 K
-
1
16
min . After steady state was reached the next SSITKA experiment (
O
2
18
16

O
2

2
O ) was performed. The exchange was carried out at 270,
2
diluted with H O in a ratio of 1:100.
300, 330 °C.
CPR-Pulse experiments:
Following the pre-treatment, the catalyst was cooled down to the desired
-
1
temperature under a constant helium flow of 20 mL min (STP). The
Table 5. Data of the catalytic material needed for the calculation of a
surface-layer.
-
1
overall volume flow is separated into two with 10 mL min (STP) each.
One of these is used for flushing the sample loop of the 6-way-valve
(Figure 12). The feed gas for the sample loop exists of 5 vol. % ACR (or
[
a]
(b]
3
[c]
-1
He-density
BET
/ m g
V
/ 10
catalyst
-9
V
/ 10
surface
-9
V
surf.∙Vcat.
/ %
-
1
3
2
-1
3
2.5 vol. % AA) in He with a total volume flow of 20 mL min (STP). With
an automated plant control (LabView®) it is possible to switch the 6-way-
valve with defined time steps (60, 90, 120, 180, 240 s). The CPR-Pulse-
experiments were carried out at 270, 300, 330 °C.
/
g cm
m
m
4.131
34
0.24
1.36
5.62
[
a] measured by Quantachrome with a Poremaster 60-GT according to
DIN 66133. [b] of 100 mg Mo . [c] with the approximation of a
surface-layer thickness of 0.4 nm.
8 2 0.5 x
V W O
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
[
14] a) O. Y. Ovsitser, K. S. Guljaev, V. D. Sokolovskii, Catal. Lett. 1991, 8,
3
[
4
[
79-384; b) O. Y. Ovsitser, V. D. Sokolovskii, Catal. Lett. 1993, 17, 239-244.
15] V. D. Sokolovskii, O. Y. Ovsitser, React. Kinet. Catal. L. 1992, 46, 415-
19.
16] a) R. Böhling, A. Drochner, M. Fehlings, D. König, H. Vogel, Chem. Ing.
Tech. 1999, 71, 226-230; b) K. Krauß, A. Drochner, M. Fehlings, J. Kunert, H.
Vogel, J. Mol. Catal. A: Chem. 2002, 177, 237-245; c) A. Drochner, P. Kampe,
J. Kunert, J. Ott, H. Vogel, Appl. Catal., A 2005, 289, 74-83; d) A. Drochner, P.
Kampe, N. Menning, N. Blickhan, T. Jekewitz, H. Vogel, Chem. Eng. Technol.
2
014, 37, 398-408; e) T. Petzold, N. Blickhan, A. Drochner, H. Vogel,
ChemCatChem 2014, 6, 2053-2058.
17] a) E. Gross, G. A. Somorjai, J. Catal. 2015, 328, 91-101; b) G. Prieto, F.
Schüth, J. Catal. 2015, 328, 59-71; c) R. Schlögl, Angew. Chem. Int. Ed. 2015,
4, 3465-3520.
[
5
[
[
18] N. Ohler, A. T. Bell, J. Phys. Chem. B. 2005, 109, 23419-23429.
19] a) L. Kihlborg, Acta. Chem. Scand. 1959, 13, 954-962; b) L. Kihlborg,
Figure 12. Scheme of the 6-way-valve with gas flow and two states.
Position A is for loading the sample loop and position B for flushing.
Acta. Chem. Scand. 1969, 23; c) T. Ekström, Acta. Chem. Scand. 1972, 26,
843-1846; d) S. Ishikawa, D. Kobayashi, T. Konya, S. Ohmura, T. Murayama,
N. Yasuda, M. Sadakane, W. Ueda, J. Phys. Chem. C 2015, 119, 7195-7206.
20] BASF SE [DE], BASF Schweiz AG [CH], 2013. Process for the
1
[
Blank runs were made to verify the initial amount of aldehyde per pulse.
They were performed for every pulse-distance and for each temperature.
heterogeneously catalysed gas-phase partial oxidation of (meth)acrolein to
(meth)acrylic acid. C. A. Welker-Niewoudt, A. Karpov, F. Rosowski, K. J.
Mueller-Engel, H. Vogel, A. Drochner, N. Blickhan, N. Duerr, T. Jekewitz, N.
Menning, T. Petzold, S. Schmidt. 2013-11-14. WO2013167405 (A1).
CPO experiments:
Following a CPR-Pulse experiment the catalyst was heated up to 400 °C
-
1
-1
with a rate of 10 K min under 20 mL min (STP) He flow. After reaching
this temperature the catalyst was re-oxidised for one hour by switching to
2
an oxidising atmosphere (10 vol. % O in He).
Acknowledgements
Grateful acknowledgement to the Eduard-Zintl-Institut (TU
Darmstadt) for performing XRD and XPS measurements,
especially to the working group of Prof. Barbara Albert (XRD)
and the working group of Prof. Christian Hess (XPS).
Keywords: acrolein • heterogeneous catalysis • mixed oxides •
oxidation • bulk oxygen • SSIKTA • pulse-technique
[
1]
T. Ohara, T. Sato, N. Shimizu, G. Prescher, H. Schwind, O. Weiberg, K.
Marten, H. Greim, in Ullmann's Encyclopedia of Industrial Chemistry, Wiley-
VCH Verlag GmbH & Co. KGaA, 2000.
[
2]
Petzold, H. Vogel, Top. Catal. 2016, 1-15.
3] a) J. Tichy, J. Machek, Catal. Lett. 1992, 15, 401-404; b) T. V.
A. Drochner, D. Ohlig, S. Knoche, N. Gora, M. Heid, N. Menning, T.
[
Andrushkevich, Catal. Rev. 1993, 35, 213-259; c) O. Ovsitser, Y. Uchida, G.
Mestl, G. Weinberg, A. Blume, J. Jäger, M. Dieterle, H. Hibst, R. Schlögl, J.
Mol. Catal. A: Chem. 2002, 185, 291-303; d) S. Ishikawa, X. Yi, T. Murayama,
W. Ueda, Catal. Today 2014; e) C. Qiu, C. Chen, S. Ishikawa, T. Murayama,
W. Ueda, Top. Catal. 2014, 57, 1163-1170; f) J. Kunert, A. Drochner, J. Ott, H.
Vogel, H. Fueß, Appl. Catal., A 2004, 269, 53-61; g) G. Schimanke, M. Martin,
J. Kunert, H. Vogel, Z. Anorg. Allg. Chem. 2005, 631, 1289-1296.
[
4]
Catal. A: Chem. 2001, 174, 169-185.
5] E. Rödel, O. Timpe, A. Trunschke, G. A. Zenkovets, G. N. Kryukova, R.
Schlögl, T. Ressler, Catal. Today 2007, 126, 112-118.
6] a) S. Endres, P. Kampe, J. Kunert, A. Drochner, H. Vogel, Appl. Catal.,
M. Dieterle, G. Mestl, J. Jäger, Y. Uchida, H. Hibst, R. Schlögl, J. Mol.
[
[
A 2007, 325, 237-243; b) P. Kampe, L. Giebeler, D. Samuelis, J. Kunert, A.
Drochner, F. Haaß, A. H. Adams, J. Ott, S. Endres, G. Schimanke, T.
Buhrmester, M. Martin, H. Fuess, H. Vogel, Phys. Chem. Chem. Phys. 2007, 9,
3577-3589.
[
7] P. Mars, D. W. van Krevelen, Chem. Eng. Sc. 1954, 3, Supplement 1,
41-59.
[
8]
a) I. E. Wachs, K. Routray, ACS Catal. 2012, 2, 1235-1246; b) R.
Grasselli, Top. Catal. 2002, 21, 79-88.
[
[
[
[
9]
G. Mestl, Top. Catal. 2006, 38, 69-82.
10] J. L. Callahan, R. K. Grasselli, AICHE. J. 1963, 9, 755-760.
11] P. Fielitz, G. Borchardt, Solid State Ionics 2001, 144, 71-80.
12] C. L. Pieck, S. del Val, M. López Granados, M. A. Bañares, J. L. G.
Fierro, Langmuir 2002, 18, 2642-2648.
13] a) R. Kapoor, S. T. Oyama, J. Mater. Res. 1996; b) H. Vogel, R.
Böhling, H. Hibst, Catal. Lett. 1999, 62, 71-78.
[
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700124
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Gas phase oxidation of acrolein to
acrylic acid on a hydrothermal
prepared mixed oxide catalyst was
investigated by SSIKTA and
Maurice Heid, Stefan Knoche, Niklas
Gora, Dominik Ohlig, Alfons Drochner*,
Bastian J. M. Etzold, Herbert Vogel
concentration programmed techniques
Page No. – Page No.
(
CPR-Pulse, CPO) with respect to
technical relevant conditions.
Dynamics of Bulk Oxygen in the
Selective Oxidation of Acrolein
Balancing the amounts of active
oxygen gives an overview of the
quantities of participating bulk oxygen
species. The dynamics of bulk oxygen
leads to re-oxidation processes on the
catalyst surface and thus influences
the selectivity pattern of the network of
acrolein oxidation.
This article is protected by copyright. All rights reserved.