Functional Analysis of Catalysts for Lower Alkane Oxidation

Article abstract of DOI:10.1002/cctc.201601194

The catalytic performance of 1) crystalline MoVTeNb oxide that exhibits the electronic properties of a n-type semiconductor, 2) submonolayer vanadium oxide supported on meso-structured silica (SBA-15) as an insulating support, and 3) surface-functionalized carbon nanotubes that contain neither a redox active metal nor bulk oxygen, but only surface oxygen species have been compared in the oxidative dehydrogenation of ethane and propane under equal reaction conditions. The catalytic results indicate similarities in the reaction network over all three catalysts within the range of the studied reaction conditions implying that differences in selectivity are a consequence of differences in the rate constants. Higher activity and selectivity to acrylic acid over MoVTeNb oxide as compared to the other two catalysts are attributed to the higher density of potential alkane adsorption sites on M1 and the specific electronic structure of the semiconducting bulk catalyst. Microcalorimetry has been used to determine and quantify different adsorption sites revealing a low V_surface/C₃H_{8 ads} ratio of 4 on M1 and a much higher ratio of 150 on silica-supported vanadium oxide. On the latter catalyst less than one per cent of the vanadium atoms adsorb propane. Barriers of propane activation increase in the order P/oCNT (139 kJ mol^?1)≤M1 (143 kJ mol^?1)<6V/SBA-15 (162 kJ mol^?1), which is in agreement with trends predicted by theory.

Full text of DOI:10.1002/cctc.201601194

Accepted Article
Title: Functional Analysis of Catalysts for Lower Alkane Oxidation
Authors: Annette Trunschke, Pierre Kube, Benjamin Frank, Sabine
Wrabetz, Jutta Kröhnert, Michael Hävecker, Juan Valesco-
Vélez, Johannes Noack, and Robert Schlögl
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.201601194
Link to VoR: http://dx.doi.org/10.1002/cctc.201601194
A Journal of
www.chemcatchem.org

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Functional Analysis of Catalysts for Lower Alkane Oxidation
Pierre Kube,[a] Benjamin Frank,[a] Sabine Wrabetz,[a] Jutta Kröhnert,[a] Michael Hävecker,[a,b]
Juan Velasco-Vélez,[a] Johannes Noack,[a,c] Robert Schlögl,[a] and Annette Trunschke*[a]
Abstract: The catalytic performance of (i) crystalline MoVTeNb
oxide that exhibits the electronic properties of n-type
dispersed monolayers at the surface of supports.[3]
a
Complex mixed MoV-based oxides composed of the bronze-like
structure M1 (ICSD no. 55097)[4] have been reported to show
high activity in the oxidation of propane[5] and ethane.
Depending on feed composition and reaction conditions the
catalyst primarily produces either olefins or oxygenates.[5b]
Under reaction conditions the surface of M1 is enriched in Te4+
and V5+ forming a thin self-supported active layer.[5b, 6]
In contrast, vanadium oxide monolayer catalysts are active in
oxidative dehydrogenation, but valuable oxygenates are formed
at most in traces irrespective of the reaction conditions.
Supported vanadium oxide species VxOy exist in different degree
of oligomerization including monomeric species without V-O-V
bonds and oligomers with different chain length.[3e] The effect of
cluster size on reactivity is still unclear.[3g, 7]
semiconductor, (ii) sub-monolayer vanadium oxide supported on
meso-structured silica (SBA-15) as an insulating support, and (iii)
surface-functionalized carbon nanotubes that contain neither a redox
active metal nor bulk oxygen, but only surface oxygen species have
been compared in the oxidative dehydrogenation of ethane and
propane under equal reaction conditions. The catalytic results
indicate similarities in the reaction network over all three catalysts
within the range of the studied reaction conditions implying that
differences in selectivity are a consequence of differences in the rate
constants. Higher activity and selectivity to acrylic acid over
MoVTeNb oxide as compared to the other two catalysts are
attributed to the higher density of potential alkane adsorption sites
on M1 and the specific electronic structure of the semiconducting
bulk catalyst. Microcalorimetry has been used to determine and
quantify different adsorption sites revealing a low Vsurface/C3H8ads
ratio of 4 on M1 and a much higher ratio of 150 on silica-supported
vanadium oxide. On the latter catalyst less than one per cent of
surface vanadium atoms adsorb propane. Barriers of propane
activation increase in the order P/oCNT (139 kJ mol-1) ≤ M1 (143 kJ
mol-1) < 6V/SBA-15 (162 kJ mol-1), which is in agreement with trends
predicted by theory.
MoVTeNbOx (M1)
VOx/SBA-15
P2O5/CNTs
Introduction
Oxidative dehydrogenation (ODH) is an interesting alternative to
current endothermic processes, like steam cracking, fluid
catalytic cracking (FCC), or catalytic dehydrogenation (DH), for
the manufacture of olefins from light alkanes.[1] However,
avoiding total combustion of the alkane to carbon oxides
remains the major challenge that impedes application.[2]
Vanadium oxide has been regarded as promising catalyst
component either in bulk mixed oxides or in the form of highly
O
O
V
V
O
O
O
S
O
O
Scheme 1. Simplified illustration of potential surface features of vanadium-
containing bulk and monolayer catalysts on the one hand and metal-free
nanostructured carbon catalysts on the other hand.
[a]
P. Kube, Dr. B. Frank, Dr. S. Wrabetz, J. Kröhnert, Dr. M. Hävecker,
Dr .J. Velasco-Vélez, Dr. J. Noack, Prof. Dr. R. Schlögl, Dr. A.
Trunschke
Department of Inorganic Chemistry
Another class of catalysts is based on elemental carbon, like
carbon nanotubes (CNT).[8] Edges and defects at the carbon
surface are decorated with a variety of oxygen-containing
functional groups, which are structurally similar to the functional
groups on vanadia catalysts (e.g., V-OH/C-OH, V=O/C=O, V-
O-V/C-O-C).[8b] These oxygen functionalities, in particular the
redox couple of quinone and hydroquinone species, are
regarded as active sites in the ODH reaction.[8b] A simplified
illustration of potential surface features of the three catalyst
types is presented in Scheme 1.
Fritz-Haber-Institut der Max-Planck-Gesellschaft
Faradayweg 4-6, 14195 Berlin (Germany)
E-mail: trunschke@fhi-berlin.mpg.de
[b]
[c]
Dr. M. Hävecker
Max Planck Institute for Chemical Energy Conversion
Stiftstr. 34-36, D-45470 Muelheim an der Ruhr (Germany)
Dr. J. Noack, and current address of Dr. B. Frank
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.
Catalytic properties of these catalysts in oxidation of ethane
1
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
and propane are not directly comparable due to very different
reaction temperatures, contact times, and feed compositions
applied so far (Tab. 1). In our work we compare bulk MoVTeNb
M1 oxide with vanadium oxide supported on mesoporous silica
SBA-15 and vanadium-free carbon nanotube catalysts stabilized
by phosphorous oxide (referred to as M1, 6V/SBA-15, and
P/oCNT, respectively) in the oxidative dehydrogenation of
propane and ethane in the same temperature range and feed.
The objective is to identify differences and similarities in reaction
network and apparent rate parameters in order to ascertain if
vanadium oxide supported on an apparently inert support, such
as silica, and V-free carbon catalysts may serve as models for
V-containing bulk catalysts in oxidative dehydrogenation
reactions of alkanes.
Table 1. Reaction conditions and catalytic properties reported in the literature for the three types of catalysts in oxidative dehydrogenation of ethane and propane.
T
Feed
τ
ralkane
Salkene
[%]
Soxygenates
[%]
Ref.
[°C]
[alkane/O2/inert]
[g s ml-1]
[mmol g-1 h-1]
Oxidative dehydrogenation of ethane
M1
MoVTeNbOx[a]
380
30/20/50
1.8
14.2
~90
~10
[9]
VOx/SiO2
6.7 wt% V/Aerosil
7.1 wt% V/MCM-41
480
600
7.3/0.93/91.77
4/8/88
0.12
0.51
0.39
45.3
~100
60.2
[10]
[11]
39.8
P/oCNT
400
20/10/70
0.33
0.24
~66.4
~33.6
[12]
Oxidative dehydrogenation of propane
M1
MoVTeNbOx
MoVTeNbOx
380
380
8/10/37/45[b]
1.5
107
8.7
89
91.3
11
[3i]
3/6/91
0.06
1.7
[13]
VOx/SiO2
4.5 wt% V/SBA-15
2.7 wt% V/SBA-15
2.8 wt% V/MCM-41
7.1 wt% V/MCM-41
0.5 wt% V/MCM-48
0.5 wt% V/Aerosil
600
500
500
550
550
550
1/1/4
2.15[c]
0.125
0.04
0.8
75[d]
21.6
412
33
52
48
[14]
[15]
[16]
[11]
[17]
[17]
16.9/16.9/67.5
30/20/50
4/8/88
~50
42.3
45.5
65.3
60.2
~50
41.4
54.5
34.7
39.8
4/8/88
11.2
9.6
1.13
1.73
4/8/88
P/oCNT
400
1/1/4
3
0.32
55
45
[18]
[a] as synthesized catalyst without high-temperature treatment; [b] alkane/O2/inert/H2O [c] τ = s [d] ralkane [mmol ml-1 h-1]
cat
compared to 6V/SBA-15 (Tab. 2). The M1 surface concentration
of V corresponds to 1.7 V atoms/nm2 (Tab. 2), calculated based
on XPS measurements (Fig. S4).
Results
6V/SBA-15 features
a sub-monolayer of two-dimensional
vanadium oxide species on the surface of mesoporous silica
SBA-15 as support. The regular pore structure of SBA-15 is
essentially maintained after deposition of vanadia (Fig. S3), but
the support (Tab. S1) loses roughly half of its specific surface
area (Tab. 2, Tab. S1), which is mainly due to blockage of
micropores by vanadium oxide species (Tab. S1). Ninety
percent of the micropores are not accessible anymore for
nitrogen adsorption after V deposition. Therefore, it is uncertain
whether all the supported vanadium oxide species are
accessible by molecules reacting from the gas phase. Therefore,
the surface density of 1.9 V atoms/nm2 (Tab. 2) calculated
based on the V content is rather an apparent value even though
Structure and chemical properties of the catalysts
High crystallinity and phase-purity of the studied bulk MoVTeNb
M1 oxide catalyst (M1 structure: ICSD No. 55097) have been
confirmed by electron microscopy (Fig. S1) and X-ray diffraction
(Fig. S2). Structural parameters of the applied M1 catalyst are
summarized in the Experimental Section. The material is
composed of dense, rod-like particles (Fig. S1) yielding a
polycrystalline, macroporous solid that is characterized by a
comparatively low specific surface area (Tab. 2, Tab. S1, Fig.
S3). The macropores are formed by voids enclosed within
aggregates and agglomerates of the primary catalyst particles.
The total vanadium content of M1 is only slightly higher
2
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Raman spectroscopy (Fig. S5) confirms high vanadium oxide
dispersion under reaction conditions.
defective/amorphous (D1, D2) as well as crystalline/graphitic (G)
carbon domains. In addition to carbon, the catalyst contains 4.9
wt% oxygen and 1.2 wt% phosphorus, which was introduced via
impregnation using ammonium hydrogen phosphate in order to
increase the thermal stability and resistance against oxidation.[20]
The density of surface oxygen less the amount associated with
phosphorous is 2.8 O atoms/nm², calculated based on XPS
measurement (Fig. S8).
The nanostructured carbon catalyst is
a
composite of
agglomerates of closely packed multi-wall carbon nanotubes
(MWCNTs) (Fig. S6). The catalyst possesses a comparatively
small quantity of meso- and micropores (Fig. S3, Tab. S1). This
indicates that stabilization of the CNTs with phosphoric acid
leads to loss of microporosity. The Raman spectrum of P/oCNTs
(Fig. S7) shows bands at 1330, 1590 and 1610 cm-1, denoted by
D1, G, and D2 bands, respectively, in the literature,[19] indicating
Table 2. Specific surface area of the catalysts, concentration of the presumed active catalyst component VOx or O, respectively, reaction rates and kinetic
parameters in the oxidative dehydrogenation of propane and ethane.
M1
6V/SBA-15
P/oCNT
SBET (m² g-1)
10.6
355
229
Total content of V (M1 and 6V/SBA-15) and O (P/oCNT)[a] (wt-%)
Number of surface V (M1 and 6V/SBA-15) and O (P/oCNT) atoms[b] (atoms g-1)
Surface concentration of V (M1 and 6V/SBA-15) and O (P/oCNT)[b] atoms (µmol g-1)
Surface density of V (M1 and 6V/SBA-15) and O (P/oCNT) atoms[b] (atoms nm-2)
Density of adsorbed propane (at constant heat of adsorption) (nm-2)[c]
7.2
1.8.1019
5.6
6.7.1020
2.5
6.4.1020
30
1113
1063
1.7
1.9
2.8
0.44 (63 kJ mol-1)
0.0014 (52 kJ mol-1)
0.0114 (44 kJ mol-1)
52-44
0.013 (36 kJ mol-1)
Differential heat of adsorption of propane (kJ mol-1)
Differential heat of adsorption of ethane (kJ mol-1)
63
36
34
14
14
Ethane
rethane consumption ,0[d] (mmol g-1 h-1)
rethane consumption ,0[d] (10-22 mmol nm-2 h-1)
rethylene formation,0[d] (mmol g-1 h-1)
rethylene formation,0[d] (10-22 mmol nm-2 h-1)
Ea, ethane consumption,0[e] (kJ mol-1)
n[f] (O2)
5.4±0.2
5094
0.055±0.001
1.5
0.030±0.001
1.3
5.2±0.2
4906
0.043±0.001
1.2
0.022±0.001
1.0
90±2
121±2
110±7
0.1±0.01
0.9±0.01
5.3±0.2
5000
0.2±0.01
0.8±0.02
0.45±0.006
12.7
0.4±0.02
0.6±0.01
0.21±0.004
9.2
n[g] (C2H6)
Propane
rpropane consumption,0[h] (mmol g-1 h-1)
rpropane consumption,0[h] (10-22 mmol nm-2 h-1)
rpropylene formation,0[h] (mmol g-1 h-1)
rpropylene formation,0[h] (10-22 mmol nm-2 h-1)
Ea, propane consumption,0[e] (kJ mol-1)
n[f] (O2)
5.1±0.04
4811
0.36±0.002
10.1
0.19±0.01
8.3
80±3
110±2
103±7
0.1±0.01
0.8±0.02
0.3±0.01
0.8±0.02
0.4±0.01
0.7±0.01
n[g] (C3H8)
[a] V in case of M1 and 6V/SBA-15 as determined by XRF,O in case of P/oCNT as determined by EDX (less oxygen associated with phosphate); [b] Surface
analysis of M1 and P/oCNT was performed by XPS (less oxygen associated with phosphate) (Figs. S4, S8), 6V/SBA-15 contains highly dispersed vanadium
oxide (Fig. S5), therefore, it was assumed that all V determined by XRF corresponds to surface V; [c] determined by propane adsorption using microcalorimetry;
[d] T=380°C, W/F=0.06 to 0.37 g s ml-1 for M1, 1.33 to 2.40 g s ml-1 for 6V/SBA-15 and P/oCNT; [e] T=380 - 420°C for M1 and 6V/SBA-15 and 360 - 380°C for
P/oCNT, feed of CnH2n+2/O2/N2 = 10/5/balance [f] T=400°C for M1 and 6V/SBA-15 and 360°C for P/oCNT, feed of CnH2n+2/O2/N2 = 10/3 to 7/balance; [g] T=400°C
for M1 and 6V/SBA-15 and 360°C for P/oCNT, feed of CnH2n+2/O2/N2 = 6 to 14/5/balance; [h] T=380°C, W/F=0.06 to 0.09 g s ml-1 for M1, 1.34 to 2.00 g s ml-1 for
6V/SBA-15 and P/oCNT.
temperatures from 360 to 420°C are compared in Fig. 1 for all
Oxidative dehydrogenation of ethane and propane
catalysts as a function of alkane conversion.
Pure silica SBA-15 shows measurable conversion in the
studied temperature range (Fig. 1). The increasing olefin
selectivity (12 - 16% ethylene, 39 – 41% propylene) with
Oxidative dehydrogenation has been performed in the same
feed and temperature range over all three catalysts, but at
different contact times to compensate differences in the
activity. Low conversion was adjusted to avoid consecutive
reactions that prevail at higher conversion. The selectivity
towards the alkene in ODH of (a) ethane and (b) propane at
increasing alkane conversion is
a hint that gas-phase
reactions or surface-initiated gas-phase reactions dominate
over pure SBA-15. The opposing trends visible in Fig. 1 for
SBA-15 and 6V/SBA-15 indicate that contributions of the free
3
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
silica surface to the reactivity of the sub-monolayer catalyst
6V/SBA-15 are negligible.
The three catalysts show significant differences in the
selectivity towards ethylene (Fig. 1a), while the selectivity to
propylene seems to follow a similar trend with increasing
conversion for all three catalysts (Fig. 1b).
Despite its low reactivity, ethylene is further oxidized in
particular over 6V/SBA-15. The selectivity of all products as a
function of conversion measured in the entire range of
applied temperatures and contact times is presented in Fig.
2-4.
be rather explained by fast post-combustion of oxygen
containing intermediates than by the absence of catalytic
sites that accomplish oxygen insertion. Indeed, it has been
shown that carbon nanotubes can catalyse the oxidation of
acrolein to acrylic acid.[21]
Due to the low activity of P/oCNTs and restrictions in the
reactor dimensions, the data shown in Fig. 4a have been
measured at extremely low conversions between 0.1% and
0.3%. In this range, the selectivity to ethylene is constant at
almost 80%. At such low conversions, minor contributions
due to the combustion of the catalyst itself may become
noticeable and may contribute to the selectivity to carbon
oxides. Using a kinetic model that describes the combustion
of CNTs in oxygen atmosphere,[20] the rate of carbon
combustion rox under the present reaction conditions (but
assuming that the feed contains only 5% O2 in N2) was
calculated resulting in a ratio of hydrocarbon conversion r to
rox of 2. The impact of catalyst degeneration is certainly lower
in presence of the hydrocarbon in the feed and is negligible at
higher conversion as also demonstrated by the high stability
of the catalyst over a term of 13-20 days on stream (Fig. S9).
100
75
100
80
a
b
M1 #11040
50
60
6V/SBA-15 #11713
5P/O-CNT #12129
SBA-15 #9179
25
40
0
1.6
2.0
1.4
1.2
1.0
0.8
1.6
1.2
0.8
0
1
2
3
4
5
0
3
6
9
12
Xethane(%)
Xpropane (%)
100
80
60
40
20
0
100
80
60
40
20
0
a
b
propylene
CO
CO2
Figure 1. Selectivity as a function of alkane conversion in ODH of (a)
ethane and (b) propane over the model catalysts at T= 360-420°C, W/F =
0.034-0.72 g s ml-1 (M1) 1.1-2.4 g s ml-1 (6V/SBA-15 and P/oCNT) and 2.0
g s ml-1 (SBA-15) in a feed composed of CnH2n+2/O2/N2 = 10/5/85.
acetic acid
acrylic acid
ethylene
CO
CO2
MoVTeNb M1 oxide shows an outstanding ethylene
selectivity of 95% up to 50% conversion of ethane (Fig. 2a).
Only small amounts of acetic acid are formed. Further
products are CO and CO2, which are not formed in a common
path, since the SCO/SCO2 ratio is only constant at low
conversion (Fig. 1a), but increases with increasing ethane
conversion (Fig. 2a). Acrylic acid is formed as a stable
oxidation product of propylene over M1 (Fig. 2b), but only
traces of acids are observed over 6V/SBA-15 (Fig. 3a and b).
0
1
2
3
0
3
6
9
12
Xethane(%)
Xpropane(%)
Figure 3. Selectivity of reaction products (labelled in the legend) as a
function of alkane conversion in the oxidation of ethane (a) and propane
(b) over 6V/SBA-15 under the following reaction conditions: T = 350 –
420°C, Feed CnH2n+2/O2/N2 = 10/5/85, W/F = 0.82 – 2.40 g s ml-1.
100
80
60
40
20
0
100
80
60
40
20
0
a
b
100
96
100
80
60
40
20
0
a
b
ethylene
CO
CO2
ethylene
acetic acid
CO
propylene
CO CO2
92
propylene
CO
CO2
CO2
acetic acid
acrylic acid
4
2
0
0.0
0.1
0.2
0.3
0.4
0.5
0.0
0.5
1.0
Xpropane(%)
1.5
2.0
Xethane(%)
0
10
20
30
40
50
0
1
2
3
4
5
Xethane(%)
Xpropane(%)
Figure 4. Selectivity of reaction products (labelled in the legend) as a
function of alkane conversion in the oxidation of ethane (a) and propane
(b) over P/oCNT under the following reaction conditions: T = 360 – 480°C,
Feed CnH2n+2/O2/N2 = 10/5/85, W/F = 0.67 – 2.00 g s ml-1.
Figure 2. Selectivity of reaction products (labelled in the legend) as a
function of alkane conversion in the oxidation of ethane (a) and propane
(b) over MoVTeNb M1 oxide under the following reaction conditions: T =
350 – 420°C, Feed CnH2n+2/O2/N2 = 10/5/85, W/F = 0.03 – 0.72 g s ml-1.
The consumption rate of the alkanes extrapolated to zero
conversion (r0) at 380°C decreases in the order
Since oxygenated products have been detected at all, the low
selectivity to oxygenates over 6V/SBA-15 and P/oCNTs could
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
M1>>6V/SBA-15>P/oCNTs (Tab. 2), irrespective of the
nature of the hydrocarbon substrate. The M1 catalyst is by an
order of magnitude more active in propane oxidation than
sub-monolayer vanadium oxide and the transition metal-free
catalyst. In case of oxidative dehydrogenation of ethane this
difference comprises even two orders of magnitude.
Interestingly, over M1 the rates of propane and ethane
consumption, respectively, are practically the same. In
contrast, ethane is more difficult to activate compared to
propane over 6V/SBA-15 and P/oCNT.
functional theory.[23] The oxidative dehydrogenation of
methane, ethane and propane was studied over two-
dimensional graphene-like cluster models terminated with
oxygen atoms in different configurations. The calculations
reveal that the barrier for the abstraction of the first hydrogen
atom (59-104 kJmol-1) by ketone groups located at the zigzag
edge of the two-dimensional graphene model is generally
lower than the barrier for the abstraction of the second H
atom (82-106 kJmol-1) suggesting that the latter step is rate-
limiting.
The differences in activity between the three different types of
catalysts can be caused by differences in the number and/or
nature of active sites. Assuming that vanadium oxide species
are the only active sites for C-H activation of alkanes in both,
MoVTeNb M1 oxide and 6V/SBA-15 catalysts, the vanadium
species located at the surface of MoVTeNb M1 oxide are
much more active compared to vanadium oxide supported on
SiO2. This results from a rough estimation of the surface
density of vanadium (Tab. 2). The number of surface V atoms
present in 1 g M1 catalyst corresponds to approximately
1.8.1019 atoms. The number of V atoms at the surface of 1 g
6V/SBA-15 is significantly higher (6.7.1020 atoms) and in the
same order of magnitude as the surface oxygen atoms in 1g
P/oCNT (6.4.1020 atoms). Contributions of the other elements
at the surface of M1 cannot be excluded since, for example,
supported molybdenum and niobium oxides are also active in
the oxidative dehydrogenation of propane and ethane.[22] But
still, due to the low specific surface area of the MoVTeNb M1
oxide, the total number of active sites in the reactor would be
smaller compared to 6V/SBA-15 even if a part of the
vanadium on SBA-15 should be trapped in micropores. In
summary, the apparent difference in the number of active
sites suggests a significant higher intrinsic activity of the
active sites at the surface of M1.
The stable catalytic performance of all three catalysts at all
reaction temperatures (exemplarily shown in Fig. S9 for
selected reaction temperatures in the ODH of propane and
ethane) allowed the analysis of kinetic parameters (Figs. S10-
S19, Tab. 2). Apparent activation energies (Ea) for alkane
consumption were calculated based on the initial alkane
consumption rates by extrapolating to zero conversion (Tab.
2). Mass transport limitations as reason for differences in the
activation energy between the three catalysts can be
excluded (Fig. S10). The low apparent activation energy
measured over M1 is in agreement with its high intrinsic
activity (Tab. 2). 6V/SBA-15 shows the strongest temperature
dependence, followed by P/oCNT. The apparent activation
energy calculated from the alkane consumption rate is higher
for ethane compared to propane. Interestingly, the difference
with reference to all three catalysts is nearly the same
(approximately 10 kJ mol-1 (Tab. 2)).
Oxygen activation seems to be fast over M1 under the
applied reaction conditions as reflected in the low reaction
order with respect to O2, but more demanding over the
vanadia monolayer catalyst and, in particular, over P/oCNT
(Tab. 2). Fast oxygen activation over M1 is in agreement with
the experimental observations that molecular O2 can be
activated at low reaction temperature for oxidation reactions
in the liquid phase,[24] and with the absence of electrophilic
oxygen species as intermediates of oxygen reduction under
conditions of propane oxidation.[5b, 25]
In summary, the main differences observed in oxidative
dehydrogenation over the three different types of catalysts
are
(1) an order of magnitude higher intrinsic activity of M1
compared to 6V/SBA-15 and P/oCNT,
(2) acids are formed over M1, but occur only in traces
over 6V/SBA-15 and P/oCNT,
(3) Facilitated oxygen activation over M1 compared to
6V/SBA-15 and P/oCNT.
The oxidative dehydrogenation of ethane and propane over
vanadium oxide-based catalysts has been frequently
analyzed in kinetic investigations implementing for example
Eley-Rideal,[26] Langmuir-Hinshelwood,[27] and Mars-van
Krevelen[22a] formalisms. In the present work, the
experimental data of propane oxidation have been fitted
applying a simple model-free time-law (Eq. 1) based on
reaction networks presented in Scheme
2
assuming
irreversibility, first order with respect to the hydrocarbon, and
zero order with respect to oxygen.
m
n
p
p
O2
⎛
⎞
⎛
⎞
C3H8
⎜
⎜
⎟
⎟
⎜
⎜
⎟
⎟
r = k ⋅
⋅
p0
p0
⎝
⎠
⎝
⎠
Equation (1)
All products detected in the gas phase have been included
into the simulation with the aim to identify the most relevant
reaction intermediates that may be responsible for the
observed selectivity patterns. Simulations were not performed
for ethane oxidation due to the lower conversions and,
consequently, larger errors. The fitting results are compared
with the experimental data in the Supporting Information (Figs.
S20-S22). The measured data points can be fitted with the
same model for all three catalysts implying that the deviating
selectivity patterns are a result of differences in the rate
constants, but the reaction network is basically the same
within the range of the studied reaction conditions. The
calculated rate constants for all three catalysts are
summarized in Tab. S2. For better comparison of the
catalysts, the rate constants were normalized with respect to
The apparent reaction order with respect to the alkane is
slightly different over the three catalysts (Tab. 2) indicating
different mechanisms of C-H bond activation in the
hydrocarbon molecule over the transition metal free catalyst
P/oCNT compared to the V-containing catalysts, which is in
line with the results of model calculations applying density-
5
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
k1. A graphical representation of the normalized values is
presented in Fig. 5, the respective numerical values are
summarized in Tab. S3.
increasing conductivity and a decreasing work function of
MoVTeNb M1 oxide under operation in alkane oxidation have
been measured by changing the organic substrate from
ethane to propane to n-butane.[28] This indicates progressive
reduction of M1 under reaction conditions with increasing
chain length of the substrate hydrocarbon. In other words, the
surface of MoVTeNb M1 oxide is more oxidized in ethane
containing feed compared to propane containing feed. These
differences are apparently reflected in the product
distribution,e.g., via consecutive reactions of the acids formed.
The olefin, especially propylene, is quite selectively converted
into acrylic acid by oxygen insertion over M1, while over
6V/SBA-15 and P/oCNT consecutive combustion of
propylene contributes significantly to the product distribution.
Based on rate constants resulting from our simulation
(Scheme 2 right, Scheme S2, rate constants summarized
exemplarily for one reaction temperature in Tab. S2),
apparent activation energies of the most significant reaction
steps in oxidative dehydrogenation of propane (Fig. 5) have
been calculated (Tab. 3). Reaction steps that exhibit rate
constants in the dimension of 10-2 or smaller (Tab. S2) have
not been taken into account.
k3b
k1c
k3a
k2c
Acetic acid
k2
Acetic acid
k2b
k1
k1
k1a
k1b
k2
k2a
C2H6
C2H4
k2a
CO
C3H8
C3H6
Acrylic acid
k4a
k1a
k3a
k3b
CO
k4b
k1b
k2b
CO2
CO2
Scheme 2. Assumed reaction networks in oxidative dehydrogenation of
ethane (left) and propane (right). The system of equations used for the
simulation is presented in Scheme S1 for ethane, and Scheme S2 for
propane, respectively, in the Supporting Information.
M1
CO2
CO2
k1
k3b
k1b
C2H6
k1a
k2b
Acrylic acid
Acetic acid
k2
k1
k2
C2H4
Acetic acid
k3a
C3H8
k1a
C3H6
k2a
k2c
k2a
Table 3. Calculated apparent activation energies of the most significant
reaction steps (Fig. 5) in oxidative dehydrogenation of propane over
MoVTeNb M1 oxide, and 6V/SBA-15 (T=380 - 420°C) and over P/oCNT
(T=360 - 380°C).
CO
CO
6V/SBA-15
CO2
CO2
k1
k1b
k2b
M1
6V/SBA-15
(kJ mol-1)
P/oCNT
k2b
Acrylic acid
Acetic acid
k2
(kJ mol-1)
(kJ mol-1)
k1
C2H6
C2H4
C3H8
C3H6
k2a
k2c
k3a
k2a
k1a
P/oCNT
k1b
Ea,k1
Ea,k2
Ea,k2a
Ea,k2b
86
65
90
70
110
-
100
-
CO
CO
59
58
50
47
CO2
CO2
k1
k2b
C2H4
k1b
k2b
C3H6
k2a
k1
C2H6
k1a
C3H8
k1a
k2a
The calculated Ea values of the first step, which is the
oxidative dehydrogenation of propane to propylene (Ea,k1), are
in good accordance with the apparent activation energies
calculated based on the experimentally determined initial
propane consumption rates rpropane consumption,0 (Tab. 2).
The agreement indicates that our simplified model describes
the reaction network satisfactorily at least with respect to the
first reaction step. Calculated activation energies of the
consecutive unselective pathways of propylene to CO2 and
CO, Ea,k2a and Ea,k2b, respectively, are about 50 kJ mol-1 lower
than Ea,k1 as far as the catalysts 6V/SBA-15 and P/oCNT are
concerned. Such a difference has also been found by Chen
et al. in their study of zirconia-supported vanadia catalysts in
oxidative dehydrogenation of propane.[22a] In contrast, deep
oxidation pathways of propylene to carbon oxides have
higher (CO formation) or only slightly lower (CO2 formation)
apparent activation energies relative to Ea,k1 over M1. As
regards M1, the calculated activation energy of the route from
propylene to acrylic acid (Ea,k2) is lower than the Ea of the
CO
CO
Figure 5. Graphical representation of the simulated rate constants in the
oxidative dehydrogenation of ethane (left) and propane (right) over the
three catalysts MoVTeNb M1 oxide, 5V/SBA-15, and P/oCNT normalized
to k1 based on the models presented in Scheme 2.
The normalized rate constants k1a, and k1b, that represent
deep oxidation of the alkane without desorption of an
intermediate, as well as the rate constants k2a, and k2b, that
correspond to deep oxidation of the desired olefin, are one or
more orders of magnitude higher for the sub-monolayer
vanadium oxide catalyst and the carbon catalyst compared to
M1.
Interestingly, acetic acid formed over M1 under reaction
conditions of oxidative dehydrogenation of ethane seems to
be much more prone to total oxidation compared to acetic
acid formed under conditions of ODH of propane. An
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
deep oxidation pathways to CO and CO2. This implies that
the energy of the highest transition state in the course of the
acrylic acid formation route k2 might be less than the energy
of the highest transition states in deep oxidation routes of
propylene over M1. Generally, Ea calculated for all
consecutive reactions of propylene are lower than Ea for
formation of propylene in the first reaction step, except
reaction of propylene to CO over M1. The varying difference
between Ea of propane and propylene consumption reactions
observed for the three catalysts reflects the differences in
their selectivity patterns.
surface alcoholate species.[29] Due to the strong interaction of
propylene with M1, irreversible adsorbed intermediates of
selective oxidation products are already formed at 40°C
providing an explanation for partial irreversible adsorption.
80
70
60
50
40
30
20
10
0
80
70
60
50
40
30
20
10
0
M1
M1
ads. C2H6
ads. C2H4
reads. C2H4
ads. C3H8
reads. C3H8
ads. C3H6
reads. C3H6
0
10
20
30
40
0
10
20
30
40
Adsorption of alkanes and alkenes
ammount of adsorbed C2H6/C2H4
ammount of adsorbed C3H8/C3H6
at 40°C (µmol g-1
)
at 40°C (µmol g-1
)
Differential heats of adsorption of the reactants ethane and
propane and the reaction products ethylene and propylene
measured at 40°C on all catalysts pretreated in the
calorimeter cell in the reaction feed under steady-state
conditions are shown in Figure 6.
80
70
60
50
40
30
20
10
0
80
6V/SBA-15
6V/SBA-15
ads. C2H6
ads. C2H4
ads. C3H8
reads. C3H8
ads. C3H6
reads. C3H6
70
60
50
40
30
20
10
0
Few sites (0.3 µmol g-1) at the surface of M1 pretreated in the
ODH feed adsorb propane strongly, which is reflected in high
heats of adsorption of 79 kJ mol-1 (Fig. 2, top right, dark blue
line). These sites may be attributed to tellurium oxide surface
species.[6] Other sites (7.7 µmol g-1) show slightly weaker
interaction with propane (63 kJ mol-1) and are more
characteristic for adsorption at vanadium oxide.[6] With
increasing coverage the heat of adsorption decreases strictly
and approaches a value of 20 kJ mol-1, which is close to the
heat of condensation. Apparently, the surface of M1 exposed
to ODH conditions (dry and rich in alkane) differs drastically
from the surface of M1 that has been exposed to a more lean
feed that contains 40% steam as usually applied when acrylic
acid is the target product in propane oxidation. Here, propane
is adsorbed with constant heat of adsorption of 40 kJ mol-1
until the monolayer is reached.[6] The time constants of the
heat signals measured in the present experiment at low
coverage are slightly increased indicating C-H activation of
propane already at room temperature. The product of this
interaction is not strongly adsorbed at the high-energy sites,
since re-adsorption after evacuation at 40°C (light blue line in
Fig. 2, top right) can be considered as reversible within the
measurement accuracy.
0
10
20
30
40
0
10
20
30
40
ammount of adsorbed C2H6/C2H4
ammount of adsorbed C3H8/C3H6
at 40°C (µmol g-1
)
at 40°C (µmol g-1
)
80
70
60
50
40
30
20
10
0
80
P/oCNT
P/oCNT
ads. C2H6
ads. C2H4
ads. C3H8
reads. C3H8
ads. C3H6
70
60
50
40
30
20
10
0
0
10
20
30
40
0
10
20
30
40
ammount of adsorbed C2H6/C2H4
ammount of adsorbed C3H8/C3H6
at 40°C (µmol g-1
)
at 40°C (µmol g-1
)
Figure 6. Differential heats of adsorption of the reactants ethane and
propane and the products ethylene and propylene adsorbed at the
catalysts M1, 6V/SBA-15 and P/oCNT petreated under steady-state
conditions in the oxidation of propane and ethane, respectively.
Ethane shows a weaker interaction with the surface of M1
compared to propane (Fig. 6, top, left, blue line), reflected in a
heat of adsorption of 34 kJ mol-1 (1.8 µmol g-1). With
increasing coverage, weaker sites are occupied (30-20 kJ
mol-1). Ethane adsorption is completely reversible.
Subsequent ethylene adsorption reveals that ethylene
interacts stronger with M1 than ethane (Fig. 6, top left, red
line) (41 kJ mol-1), but still, the adsorption is reversible (Fig. 6
top left, orange line). Due to the smaller size, adsorption
capacity of C2 hydrocarbons on M1 is higher compared to C3
hydrocarbons.
At 6V/SBA-15 (Fig. 6 middle right, dark bue line), a few sites
(0.8 µmol g-1) are characterized by a heat of propane
adsorption of 52 kJ mol-1, which exhibit strong interaction with
propane. At higher coverage from 0.8 to 10 µmol g-1 the heat
of adsorption decreases to 44 kJ mol-1 on average.
Adsorption of propylene on M1 after two propane adsorption
experiments shows lower heats of adsorption compared to
propane, and reduced adsorption capacity. Moreover,
propylene adsorption is partially irreversible. FTIR
spectroscopy after 20 hours interaction of propylene with M1
performed to simulate the microcalorimetry experiment (Fig. 7,
blue spectrum) confirms that adsorbed oxygenates are
formed already at room temperature. All bands visible in Fig.
7 are due to adsorbed and gas-phase species, since the
spectrum of the pretreated catalyst has been used as
reference. The intense bands between 1900-1800, 1700-
1600, and 1500-1300 cm-1 originate from gas-phase
propylene. The peaks due to adsorbed species are marked in
Fig. 7. The band at 1556 cm-1 is attributed to the asymmetric
COO stretching vibration of adsorbed carboxylate groups,
while the peaks below 1300 cm-1 originate from various
7
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Adsorption of propane on 6V/SBA-15 changes the nature of
the adsorption sites and also slightly the amount since
readsorption of propane after evacuation at 40°C follows a
different curve compared to the first adsorption experiment
(Fig. 2, middle right, light blue line) showing a plateau at 39
kJ mol-1 (8 µmol g-1) and an overall reduced adsorption
capacity. Interestingly, the amount of propane adsorbed at
6V/SBA-15 with a heat higher than the heat of condensation
is similar compared to M1 although the number of surface
vanadium atoms on 6V/SBA-15 is much higher (Tab. 2).
condensation. There is no specific interaction with the oxygen
containing functional groups at the surface.
In summary, all hydrocarbon molecules are stronger
adsorbed and reach a much higher surface concentration on
M1 (Tab. 2) as compared to 6V/SBA-15 and P/oCNT.
Interestingly, the mean ratio of surface vanadium atoms
relative to adsorbed propane molecules Vsurface/C3H8ads (Tab.
2) corresponds to approximately 4 in case of M1 and 150 in
case of 6V/SBA-15 revealing that only a part of the vanadium
oxide species in 6V/SBA-15 (approximately 1%) adsorb
propane. In this respect it is interesting to note that tetramers
of vanadium oxide species have been postulated to be the
required ensemble size for oxidative dehydrogenation of
propane.[2] The heat of adsorption increases with the chain
length of the hydrocarbon molecule on all catalysts.
Propylene reacts slowly with the surface of M1 already at
room temperature under formation of carboxylates.
Carboxylates are not formed on the other two catalysts.
M1
0.12
6V/SBA-15
0.10
0.08
0.06
0.04
0.02
0.00
Concluding Discussion
The present study highlights three model catalysts with
substantial structural diversity, which all are capable to
catalyse the oxidative dehydrogenation of ethane and
propane. The two V-containing catalysts represent sub-
monolayer catalysts. On the one hand vanadium oxide
species are supported on silica as an insulating support
(6V/SBA-15). On the other hand a monolayer enriched in V5+
and Te4+ is self-supported on a crystalline multi-metal mixed
oxide (MoVTeNb M1 oxide) that exhibits the electronic
properties of a n-type semiconductor.[28, 31] The third type of
catalyst is composed of surface-functionalized carbon
nanotubes (P/oCNT) that contain neither vanadium oxide nor
bulk oxygen, but only surface oxygen species of which
quinone-hydroquinone groups are assumed to be the active
redox couple.
The catalytic performance of the three catalysts has been
compared under identical reaction conditions in terms of
temperature and feed composition. The results reveal general
similarities such as the alkenes being the main reaction
products and CO is preferentially formed compared to CO2 as
product of deeper oxidation. However, the overall by-product
spectrum and the macrokinetic parameters show the
following remarkable differences:
1. The rate of alkene formation is 1-2 orders of magnitude
higher over M1 compared to 6V/SBA-15 and P/oCNT.
2. Interestingly, the rate of olefin formation over M1 is
approximately the same in oxidative dehydrogenation of
ethane and propane. On the other two catalysts, bigger
differences in the reactivity of the two alkane molecules are
observed.
3. The three catalysts show different selectivity to the olefin at
similar conversion. The spectrum of by-products differs as
well.
2000
1800
1600
1400
1200
1000
Wavenumber / cm-1
Figure 7. FTIR spectra of adsorbed species at the surface of M1 and
6V/SBA-15 after adsorption of propylene at 40°C for 20 hours in presence
of 10 mbar propylene (equilibrium pressure). The peaks labelled with band
positions arise for adsorbed species.
Propylene gives low heats of 30 kJ.mol-1 at coverage of 10
.
µmol g-1. The formation of carboxylate species at the surface
of 6V/SBA-15 can be excluded due to the absence of bands
in the window between 1600 and 1500 cm-1 in the infrared
spectrum of propylene adsorbed on 6V/SBA-15 (Fig. 7). This
suggests that the intermediate product propylene interacts
differently
with
6V/SBA-15
and
M1,
respectively.
Unfortunately, the region below 1300 cm-1 is not accessible
due to complete absorption of the infrared light by the silica
support (Fig. 7) and possibly adsorbed alkoxide species
cannot be probed.
Adsorption measurement of ethane and ethylene on 6V/SBA-
15 yield a low heat of 14 kJ mol-1, which is equal to the heat
of condensation of the molecules. There is no specific
interaction with the vanadium oxide species at the surface.
Propane adsorption on P/oCNT is reversible with a heat of 36
kJ mol-1 and a coverage of 5 µmol g-1, which corresponds to
the interaction with surface functional groups.[30] The amount
of adsorbed propane is comparable with the other two
catalysts.
Propylene adsorption on P/oCNT is reversible exhibiting
.
.
coverages of 7 µmol g-1 and an adsorption heat of 36 kJ mol-1.
Adsorption measurements of ethane and ethylene on P/oCNT
yield a low heat of 14 kJ mol-1 that corresponds to the heat of
8
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Differences in activity
reaction pathways, which are basically not known in great
detail due to the complexity of the reaction network. Batiot
and Hodnett analysed a number of oxidation reactions and
observed a correlation between selectivity at fixed conversion
and the difference between the bond dissociation enthalpy of
the weakest C-H bond in the reactant and the bond
dissociation enthalpy of the weakest bond in the selective
oxidation product irrespective of the type of oxide-based
catalyst that has been used.[34] The concept implies that
energetic constraints limit selectivity in oxidation catalysis. It
provides a rule of thumb to estimate the threshold value of
selectivity that depends on the gas-phase stability of reactant
and product. The selectivity to ethylene is expected to be high
if the reaction stops at the formation of ethylene since the
We propose that the different density of alkane adsorption
sites (Tab. 2) is responsible for differences in the activity of
M1, 6V/SBA-15 and P/oCNT. Despite the fact that more
potential V or oxygen adsorption sites per g catalyst are
available at the surface of 6V/SBA-15 and P/oCNT,
respectively, (Tab. 2), M1 adsorbs more alkane molecules
per nm2 surface (Tab. 2, density of propane adsorption sites)
and overall a similar amount per g catalyst (Fig. 6). In addition,
M1 shows a higher intrinsic activity, which could be attributed
to a cluster size effect (size of VxOy oligomers) based on the
low overall Vsurf /propane ratio of 4. It has been shown by DFT
calculations that the barrier for abstraction of the first
hydrogen atom under formation of a propyl radical decreases
with increasing cluster size.[32] Values of 160, 148, 139 and
132 kJ mol-1 have been calculated for monomers, dimers,
tetramers and octamers, respectively. With the assumption
that all intermediates occur in pseudo-steady-state, and
adsorption as well as regeneration steps are in quasi-
equilibrium,[22a] the overall barrier for propylene formation can
be estimated based on measured adsorption enthalpies (Fig.
6, Tab. 2) and apparent activation energies in propane
oxidation (Tab. 2). The estimation is based on the assumption
that the heat of adsorption is nearly independent of the
temperature.[33] The thus calculated intrinsic barrier for
propane activation increases in the order P/oCNT (139 kJ
mol-1) < M1 (143 kJ mol-1) < 6V/SBA-15 (162 kJ mol-1). The
similar values for P/oCNT and M1 suggest similar C-H
activation mechanisms over these two catalysts. The high
value for 6V/SBA-15[32] is in agreement with the high
dispersion of vanadium in 6V/SBA-15. The latter barrier has
been calculated based on the heat of propane adsorption of
52 kJ mol-1 (Fig. 6, Tab. 2) that correspond to a density of
adsorption sites of 0.0014 nm-2 on 6V/SBA-15 (Tab. 2).
According to the DFT calculations by Rozanska et al.,[32a]
these sites could be monomeric species that comprise
approximately 0.1% of all V atoms. At higher coverage
propane adsorbs with a higher density of 0.0114 nm-2 on a
different type of sites with a heat of adsorption of 44 kJ mol-1.
For these sites a slightly lower barrier of 154 kJ mol-1 can be
estimated, which is close to the value calculated for silica-
supported vanadium oxide dimers (0.6% of all V atoms).[32a]
Hence, microcalorimetry apparently allows the determination
and quantification of propane adsorption sites in vanadium
oxide monolayer catalysts and resolves different degrees of
VxOy oligomerization.
weakest C-H bond in ethane (bond dissociation energy D0
C-
H=419.5 kJ.mol-1) can be activated more readily than the
weakest C-H bond in ethylene (D0C-H=444 kJ.mol-1).[34] In the
current work, however, a much higher selectivity to ethylene
is observed over M1 compared to 6V/SBA-15. The difference
that comprises up to 30% is not obvious within the concept
mentioned above, in particular because M1 catalyses the
consecutive oxidation of ethylene to acetic acid as well and
should, therefore, show a lower selectivity to ethylene. In
contrast, a smaller difference between M1 and 6V/SBA-15 is
observed in the selectivity to propylene in ODH of propane
(10%) under the applied reaction conditions. The result is
again not straightforward, because the product distribution of
the two catalysts in propane oxidation is very different. The
current work illustrates that in addition to the thermodynamic
stability of a substrate or product molecule the interaction of
reactant or product with the catalyst surface has an impact on
the selectivity.
We observed only weak interaction between propylene and
the V-containing catalysts, which is an unexpected result.
Propylene adsorption on M1 yields adsorbed acrylates
already at room temperature in a slow reaction. The high
density of adsorption sites on M1 apparently renders possible
concerted reactions, which involve oxygen insertion. Such an
ensemble effect together with the balanced oxygen
activation[28] result in improved selectivity to acrylic acid over
MoVTeNb M1 oxide.
Vanadium oxide supported on an insulating support behaves
electronically more like a single-site catalyst, whereas the
selectivity over a semiconducting catalyst, such as M1, can
be controlled by the redox-level of the (dynamic) surface layer
that has an impact on the surface potential barrier, which
again determines charge transfer between adsorbed
molecules and the catalyst as will be explained in detail
below.[35] Comparable dynamic properties have not been
proven so far on monolayer catalysts under reaction
conditions. The presented results imply that the validity is
limited of vanadium oxide monolayers supported on insulating
supports to serve as a model for the surface layer of V-
containing bulk catalysts. Studies using isotope labeled
reactants are underway to further elucidate the differences in
the reaction mechanism.
Intrinsic barriers for ethane activation are (124 kJ mol-1) for
M1, and (135 kJ mol-1) for 6V/SBA-15, which is also in
agreement with the activity difference.
Differences in selectivity
Hydrogen abstraction is the rate-determining step in ODH of
alkanes. Therefore, the discussion above refers only to the
differences in the activity. Selectivity, however, is determined
by the relative values of the rate constants of the various
consecutive and parallel reactions or differences in the
9
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Materials science of model catalysts
know from combustion kinetics of carbon and hence provides
the opportunity for facile combustion. This mobile active
oxygen counteracts the low density of active di-ketonic sites
that should be suitable for selective operation and leads to
the poor performance both in respect of activity and
selectivity.
The investigated catalysts represent three classes of
materials with differing electronic properties. The presented
results imply that counting the local density of specific
structural elements (VxOy oligomers, quinone-hydroquinone
couples) is a necessary but insufficient description of the
material science required for generating an active and
selective catalyst. The electronic properties with respect to
charge carriers controlling the overall potential of the active
structural elements needs also to be put into consideration.
This is often done implicitly by choosing the “support” of an
active structure. Figure 8 schematically indicates this by
illustrating how the charge carriers of an assumed redox
reaction are transported from the energy level of the
hydrocarbon substrate to the energy level of oxygen. This
downhill in energy provides an elementary driving force for
the reaction to occur. It may thus be expected that the ease
of charge carrier transport will affect the observed kinetics of
the reaction. In agreement with the well-known concept of
selectivity being the consequence of reactant abundance it
can be expected that the ease of charge carrier transport
limits selectivity. Resultant is the observation that fast
reactions are not so selective and selective reactions are not
so fast.
The semiconductor M1 with a termination layer containing
VxOy is characterized by a situation of band bending caused
by formation of surface states in the feed of the reactant (Fig.
8a).[22b] This allows connecting the electron donor level of the
hydrocarbon with the electron acceptor level of the adsorbed
oxygen sitting on the VxOy d-states of the active clusters. In
this way a self-limiting charge carrier transport situation is
created. A facile execution of the elementary steps of
hydrocarbon oxidation and oxygen reduction occurs through
charge carriers from the catalyst and the steady state is
maintained by provision of charge carriers from the reactants.
In Figure 8b the situation is shown for VxOy supported on the
insulator silica. Here no nearly free charge carriers exist in
the material, no band bending occurs and no support of the
charge carrier exchange between reactants can be expected
from the bulk. Only local exchange between co-adsorbed
reactants can cause catalytic reaction. This requires highly
special geometric situations of VxOy clusters that can store 4
redox equivalents and more than one hydrocarbon molecule
plus an oxygen molecule in close proximity. It is conceivable
that only few sites are structurally suitable and that
combustion is facile due to the proximity of electrophilic
oxygen occurring during partial charge exchange between
hydrocarbon and molecular oxygen.
In Figure 8c we see the situation of the semi-metal sp2 carbon
on which di-ketonic groups capable of undergoing charge
carrier exchange with the sp2 carbon skeleton are the active
sites. Here the lack of an energy gap in the carbon backbone
allows facile charge carrier exchange. The localization of the
electrons in the quinoidic groups provides the energy barrier
necessary for selective operation. The very low numbered
density of active sites for oxygen activation requires the
movement of partially activated electrophilic oxygen as we
Figure 8: Schematic representation of the charge carrier transport
situation in the model catalysts (a) M1,[22b] (b) VxOy/SBA-15,[40] and (c)
quinone on oCNTs.
10
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
calcined at 550°C in static air for 2 h. The catalyst is called 6V/SBA-
15 (sample ID 11713).
In summary, the present study shows that fine-tuning of static
and dynamical aspects of catalytic surfaces are necessary.
The conventional approach focusing on a single property of a
catalyst (such as the nuclearity of VxOy clusters or their static
oxidation state for example) is insufficient to capture the
complexity of the dynamical situation of a redox catalyst. We
note that the elementary processes of handling the protons
and the formation of water being a kinetically most relevant
aspect have not been considered here. This is due to the
absence of macro-kinetic effects of these reactions in the
chosen regime of low conversion. If catalysts are operated
under higher loads these aspects become relevant and
complicate the picture further as then acid-base properties
are additionally relevant.
Synthesis of P/oCNT
Multi-walled CNTs from Nanocyl (NC3100) have been used as
starting material for the synthesis of phosphorous-modified CNTs.
The CNTs (sample ID 5664) were refluxed in concentrated HNO3 for
2 h, filtrated under vacuum, washed with deionized H2O to pH 6-7 and
dried at 110°C in air (oCNTs, sample ID 11450). Phosphorous
modification was done by incipient wetness impregnation, using 30.66
ml of an aqueous (NH4)2HPO4 solution (0.247 mol l-1 P) mixed with
10.22 g oCNTs and 29.34 ml H2O to achieve a nominal loading of 5
wt% P2O5. The resulting paste was thoroughly kneaded in a mortar
and dried at 110°C for 4 days. The resulting catalyst is called P/oCNT
(sample ID 12129).
Catalyst characterization
Experimental Section
Catalyst preparation
The morphology of the primary MoVTeNb M1 oxide particles was
studied by scanning electron microscopy (SEM) using a Hitachi S-
4800 electron microscope operating at 2 kV in secondary electron
(SE) mode.
Synthesis of MoVTeNb M1 oxide
Phase-pure M1 was synthesized according to
purification Briefly, 22.95
method.[36]
a
precipitation-
Phase analysis was performed by X-ray diffraction (XRD) using a
Bruker D8 ADVANCE diffractometer (Cu Kα radiation, secondary
graphite monochromator, scintillation counter). The unit cell
parameters of M1 were refined by least-square fitting of the diffraction
peak positions using the M1 structure (orthorhombic, space group
Pba2 [ICSD 55097]) utilizing the program package TOPAS (v 4.2,
Bruker AXS).
g
(18.57 mmol)
(NH4)6Mo7O24·4H2O, 4.56 g (38.98 mmol) NH4VO3, and 6.87 g (29.91
mmol) Te(OH)6 were subsequently dissolved in 100 ml H2O at 80 °C.
A solution of 7.16 g (23.63 mmol) NH4[NbO(C2O4)2]·xH2O dissolved in
30 ml H2O at 40°C was finally added to the Mo/V/Te solution at room
temperature. The obtained slurry with a nominal ratio of Mo:V:Te:Nb =
1:0.3:0.23:0.18 was spray-dried. The resulting powder was calcined at
275°C in
a 1 h and
flow of synthetic air (100 ml min-1) for
Nitrogen adsorption was performed at -196°C using the Autosorb-6B
analyser (Quantachrome) after outgassing the catalysts in vacuum
(M1 for 2 h at 120°C, 6V/SBA-15 for 16 h at 120°C, P/oCNT for 2 h at
200°C). All data treatments were performed using the Quantachrome
Autosorb software package. The specific surface area SBET was
calculated according to the multipoint Brunauer-Emmett-Teller
method (BET) in the p/p0 = 0.05-0.15 pressure range assuming the N2
cross sectional area of 16.2 Å2. The pore size distribution of SBA-15
and 6V/SBA-15 was determined by NLDFT method using a model
based on equilibrated adsorption of N2 on silica assuming cylindrical
pores at -196°C. The micropore surface area Sµ was estimated using
the t-plot method in the statistical thickness t = 4.5-6.5 Å range. The
total pore volume VP was determined by using the amount of
physisorbed nitrogen at a relative pressure p/p0 = 0.95.
subsequently annealed at 600°C in a flow of Ar (100 ml min-1) for 2 h.
The M2 phase present in the as-prepared biphasic MoVTeNb oxide
was dissolved in a 15% H2O2 solution at ambient temperature under
continuous stirring for 24 h. The washed sample was vacuum filtrated,
washed with H2O and dried at 95°C for 3 h. Finally, the obtained
powder was treated at 600°C for 2 h in a flow of Ar (100 ml min-1)
resulting in crystalline, phase-pure MoVTeNb M1 oxide (sample ID
11040). The metal composition normalized to molybdenum
corresponds to Mo1.0V0.29Te0.1Nb0.15Ox. Results of the basic catalyst
characterization are summarized in the Supporting Information. The
SEM image (Fig. S1) reveals the typical rod-shaped microstructure of
the primary catalyst particles that are composed of the M1 phase only.
The Rietveld refinement results of the experimental powder X-ray
diffraction patterns that confirm phase purity of the catalyst are
presented in Fig. S2 (lattice parameters (Å): a=21.1210(13);
b=26.5999(16); c=4.01689(21)).
X-ray fluorescence spectroscopy (XRF) was used for elemental
analysis of M1 and 6V/SBA-15 applying a Bruker S4 Pioneer X-ray
spectrometer. For sample preparation, the mixture of 0.1 g of the
catalyst and 8.9 g of lithium tetraborate ( > 99.995 %, Aldrich) was
fused into a disk using an automated fusion machine (Vulcan 2 MA,
Fluxana).
Synthesis of 6V/SBA-15
Mesoporous silica SBA-15 (sample ID 9179) was used as support.
The synthesis of SBA-15 has been described in detail elsewhere.[37]
Dispersed vanadia species were deposited by a grafting procedure.[37]
Briefly, 11.24 g support material (dried for 16 h at 130°C in air) was
suspended in 284.4 g toluene. After 30 min, 25.58 g of a solution of
10.01 g (41.0 mmol) vanadium (V) triisopropoxide in 50.23 g toluene
was added to the SBA-15 suspension and stirred for 2 h at ambient
temperature. The evaporation of isopropanol was performed in a
rotary evaporator at 50°C at a residual pressure of 50 mbar. The
resulting light orange powder with a nominal V content of 6 wt% was
Raman spectroscopic investigation of the catalyst samples was done
at 532 nm excitation wavelength using a confocal TriVista microscope
setup TR557 (S&I GmbH, Warstein, Germany) equipped with a liquid
nitrogen cooled spectroscopy CCD system PyLoN:2kBUV and 750
mm focal length of the monochromator (Princeton Instruments).
Spectra resemble an average of multiple measurements at different
spots of the sample. Calcination of 6V/SBA-15 and ODP reaction
have been done in
a
CCR1000 reactor cell (Linkam Scientific,
11
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Tadworth, UK) at 550°C in synthetic air and in ODP feed (C3H8/O2/N2
= 10/5/85) at 380°C, respectively.
Conversion of propane X and product selectivity Sk were calculated
based on the sum of products as follows:
In situ X-ray photoelectron spectroscopy (XPS) was performed at the
synchrotron radiation facility BESSY II in Berlin. At the ISISS
(Innovative Station for In Situ Spectroscopy) beamline monochromatic
X-ray light was used to obtain high pressure XP spectra in the
n
ni (product)
∑
ν i
i=1
X =
k
nj (C -compounds)
∑
presence of reactive gases (alkane
+ oxygen) at elevated
ν j
j=1
temperatures. Details of the vacuum system and the electronanalyser
were reported before.[38] For the XPS studies 10 mg of M1 powder
were pressed into a self-supporting pellet (1 ton pressing pressure,
diameter of pellet: 8 mm). In the experiment, M1 was heated to 400°C
with a heating rate of 5 K/min in the presence of alkane/oxygen feed
(volume flow ratio 1 sccm/2 sccm). The total pressure was 25 Pa
during the experiment. Alkanes, olefins (ethylene, propylene), CO,
and CO2 were analyzed with a micro gas chromatograph (micro-GC,
Varian) after compressing the gas to atmospheric pressure. In parallel,
the oxygenates (acetic acid, acrylic acid, maleic anhydride) were
detected with a proton transfer reaction mass spectrometer (PTR-MS,
IONICON). Core level spectra of O1s, V2p, Mo3d, Nb3d, Te3d, and
C1s were obtained with a constant kinetic electron energy of 150 eV
corresponding to an inelastic mean free path (IMFP) of 0.6 nm. To
calculate the elemental composition at the surface of M1, normalized
core level intensities were evaluated after subtraction of a Shirley type
background taking into account the photon energy dependence of the
atomic subshell photoionization cross sections,[39] using CASA data
analysis software (Neil Farley, www.casaxps.com). Atomic
abundance for PoCNT was performed following the same procedure
and same KE (150 eV). The background was subtracted using a
Shirley type. The analyzed peaks were C 1s, O1s and P 2p.
Equation (2)
Nk (C - atoms)× nk ( product)
Sk
=
, k = 1…n
n
N (C - atoms)× n ( product)
∑
i
i
i=1
Equation (3)
Reaction rates for propane consumption and propylene formation
were determined using the following equation.
d(ni )
= vir
i
W
F
⎛
⎜
⎝
⎞
⎟
⎠
d
Equation (4)
Reaction rates at zero contact time are obtained by linear
extrapolation of the calculated reaction rates to W/F=0.
Absence of external and internal transport limitations were verified by
calculating the dimensionless Mears and Weisz-Prater criteria and
measuring the catalyst performance for different catalyst amounts in
two different gas flows (10 and 15 ml/min) (see Figure S10, data at
contact time 0.06 g s ml-1). The very active M1 catalyst shows the
highest Mears-modulus of 3.3·10-6 and a Weisz-Prater modulus of
1.76·10-3 indicating that neither external nor internal mass transport
limitations play a role.
Catalytic tests
The catalytic tests were carried out using a setup for partial oxidation
(Integrated Lab Solutions) with 8 fixed bed quartz reactors (6 mm
inner diameter) in parallel. Each reactor was equipped with
a
thermocouple for measuring the temperature inside the catalyst bed.
The catalytic performance was determined at atmospheric pressure
under steady state conditions. The reactant feed comprised the
hydrocarbon (C3H8 or C2H6), O2, and N2 as diluent. The reaction
conditions have been varied and are indicated in the results part.
Starting and reference point for all variations was 360°C (P/oCNT) or
Microcalorimetry
Differential heats of propane and propylene adsorption on used
catalysts were determined at 313 K using a MS70 Calvet Calorimeter
(SETRAM). The catalysts were pretreated in the calorimeter cell in a
feed of 10% hydrocarbon (C3H8 or C2H6) and 5% oxygen in helium
with a total flow rate of 20 ml min-1. The reaction temperature was
400°C for 6V/SBA-15, 360°C for P/oCNT and 350°C for M1. The
reaction was performed at steady state for 20 hours, subsequently,
the cell was cooled down to room temperature in pure helium. The
cell was then sealed and transferred to the calorimeter. The
calorimeter is equipped with a custom-designed high vacuum and gas
dosing apparatus. Hydrocarbons were stepwise introduced into the
evacuated cell (p < 3·10-8 mbar), and the pressure evolution and the
heat signal were recorded for each dosing step.
400°C (M1, 6V/SBA-15, and SBA-15) and
a dry feed of 10%
hydrocarbon, 5% O2, and 85% N2. The catalysts were pressed under
~55 MPa, crushed and sieved to a particle size of 250-355 µm. Then,
different amounts of catalyst were loaded into the reactor to realise
different contact times at the same gas flow. In the case of the low
amount of M1 samples, the catalyst was diluted with ~3 g of SiC
(sieve fraction 250-355 µm). The calculated pressure drop was below
0.5 mbar for all loadings. A total flow of 10 ml/min was used in all
experiments.
An online gas chromatograph (Agilent 7890A) is used for gas analysis.
A combination of Plot-Q (length 30 m, 0.53 mm internal diameter, 40
µm film thickness) and Plot-MoleSieve 5A columns (30 m length, 0.53
mm internal diameter, 50 µm film thickness), connected to a thermal
conductivity detector (TCD), was used to analyse the permanent
gases CO, CO2, N2, O2, and CH4. A system of a FFAP (length 30 m,
0.53 mm internal diameter, 1 µm film thickness) and a Plot-Q column
(length 30 m, 0.53 mm internal diameter, 40 µm film thickness),
connected to a flame ionization detector (FID), was used to analyse
C2-C3 hydrocarbons and oxygenates.
In-situ FTIR spectroscopy
FTIR spectra of adsorbed species at the surface were measured
using a PE100 spectrometer (Perkin Elmer) equipped with a MCT
detector and a homemade quartz cell, which is connected to a
vacuum and gas dosing system. The spectra were recorded with 64
scans and a resolution of 4 cm-1. Catalyst powders were pressed with
260 MPa into self-supported wafers and the samples were pre-treated
as follows. In case of M1, the sample was heated up in vacuum to
400°C with 10 K min-1. After a hold time of 30 min the sample was
cooled to 40°C and the final starting pressure for the experiment was
12
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
[9]
T. T. Nguyen, M. Aouine, J. M. M. Millet, Catalysis Communications
2012, 21, 22-26.
Keywords: Silica-supported vanadium oxide • M1 phase •
CNTs • C-H activation • reaction mechanism
[10] J. Le Bars, J. C. Vedrine, A. Auroux, S. Trautmann, M. Baerns,
Applied Catalysis A: General 1992, 88, 179-195.
[11] B. Solsona, T. Blasco, J. M. Lopez Nieto, M. L. Pena, F. Rey, A.
Vidal-Moya, Journal of Catalysis 2001, 203, 443-452.
1.3.10-6 mbar. 6V/SBA-15 was heated up in 200 mbar oxygen to
550°C with a heating rate of 10 K min-1. After 30 min hold time the
sample was cooled to 100°C. At this temperature the cell was
evacuated and cooled to 40°C. Final starting pressure for the
experiment was 9.1.10-6 mbar. Propylene was stepwise introduced at
40°C up to 10 mbar and left for 20 h under these conditions. Then the
measurement was performed using the spectra of the pre-treated
catalysts as reference.
[12] B. Frank, M. Morassutto, R. Schomäcker, R. Schlögl, D. S. Su,
ChemCatChem 2010, 2, 644-648.
[13] R. Naumann d’Alnoncourt, Y. V. Kolen’ko, R. Schlögl, A. Trunschke,
Combinatorial Chemistry and High Throughput Screening 2012, 15,
161-169.
[14] Y.-M. Liu, Y. Cao, N. Yi, W.-L. Feng, W.-L. Dai, S.-R. Yan, H.-Y. He,
K.-N. Fan, Journal of Catalysis 2004, 224, 417-428.
[15] A. Dinse, S. Khennache, B. Frank, C. Hess, R. Herbert, S. Wrabetz,
R. Schlögl, R. Schomäcker, Journal of Molecular Catalysis A:
Chemical 2009, 307, 43-50.
Acknowledgements
[16] O. V. Buyevskaya, A. Brückner, E. V. Kondratenko, D. Wolf, M.
Baerns, Catalysis Today 2001, 67, 369-378.
We thank Dr. M. Sanchez-Sanchez and Dr. T. Wolfram for
the synthesis of the catalysts, Dr. F. Girgsdies and E.
Kitzelmann for XRD experiments, M. Hashagen and G.
Lorenz for BET measurements, G. Weinberg for SEM
measurements and Dr. C. Heine and Dr. R. Naumann
d’Alnoncourt for the experimental assistance.
[17] M. L. Pena, A. Dejoz, V. Fornes, F. Rey, M. I. Vazquez, J. M. Lopez
Nieto, Applied Catalysis, A: General 2001, 209, 155-164.
[18] B. Frank, J. Zhang, R. Blume, R. Schlögl, D. S. Su, Angewandte
Chemie International Edition 2009, 48, 6913-6917.
[19] A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner, U. Pöschl,
Carbon 2005, 43, 1731-1742.
[20] B. Frank, A. Rinaldi, R. Blume, R. Schlögl, D. S. Su, Chemistry of
Materials 2010, 22, 4462-4470.
[1]
aF. Cavani, N. Ballarini, A. Cericola, Catalysis Today 2007, 127,
113-131; bF. Cavani, Catalysis Today 2010, 157, 8-15.
[21] B. Frank, R. Blume, A. Rinaldi, A. Trunschke, R. Schlögl,
Angewandte Chemie - International Edition 2011, 50, 10226-10230.
[22] aA. Celaya Sanfiz, T. W. Hansen, D. Teschner, P. Schnörch, F.
Girgsdies, A. Trunschke, R. Schlögl, M. H. Looi, S. B. A. Hamid,
The Journal of Physical Chemistry C 2010, 114, 1912-1921; bC.
Heine, M. Hävecker, M. Sanchez-Sanchez, A. Trunschke, R.
Schlögl, M. Eichelbaum, The Journal of Physical Chemistry C 2013,
117, 26988-26997; cC. Heine, M. Havecker, A. Trunschke, R.
Schlögl, M. Eichelbaum, Physical Chemistry Chemical Physics 2015,
17, 8983-8993.
[2]
[3]
R. Schlögl, Topics in Catalysis 2011, 54, 627-638.
aG. C. Bond, S. F. Tahir, Applied Catalysis 1991, 71, 1-31; bE. A.
Mamedov, V. Cortés Corberán, Applied Catalysis A: General 1995,
127, 1-40; cT. Blasco, J. M. Lopez Nieto, Applied Catalysis, A:
General 1997, 157, 117-142; dB. M. Weckhuysen, D. E. Keller,
Catalysis Today 2003, 78, 25-46; eC. Hess, ChemPhysChem 2009,
10, 319-326; fN. F. Dummer, J. K. Bartley, G. J. Hutchings, in
Advances in Catalysis, Vol 54, Vol. 54 (Eds.: B. C. Gates, H.
Knozinger), Elsevier Academic Press Inc, San Diego, 2011, pp.
189-247; gI. E. Wachs, Dalton Transactions 2013, 42, 11762-
11769; hA. Trunschke, in Nanostructured Catalysts: Selective
[23] aK. Chen, A. T. Bell, E. Iglesia, The Journal of Physical Chemistry B
2000, 104, 1292-1299; bE. Heracleous, M. Machli, A. A. Lemonidou,
I. A. Vasalos, Journal of Molecular Catalysis A: Chemical 2005, 232,
29-39; cP. Viparelli, P. Ciambelli, L. Lisi, G. Ruoppolo, G. Russo, J.
C. Volta, Applied Catalysis A: General 1999, 184, 291-301; dC. S.
Guo, K. Hermann, M. Hävecker, J. P. Thielemann, P. Kube, L. J.
Gregoriades, A. Trunschke, J. Sauer, R. Schlögl, The Journal of
Physical Chemistry C 2011, 115, 15449-15458; eK. Amakawa, L.
Sun, C. Guo, M. Hävecker, P. Kube, I. E. Wachs, S. Lwin, A. I.
Frenkel, A. Patlolla, K. Hermann, R. Schlögl, A. Trunschke,
Angewandte Chemie International Edition 2013, 52, 13553-13557.
[24] O. V. Khavryuchenko, B. Frank, A. Trunschke, K. Hermann, R.
Schlögl, The Journal of Physical Chemistry C 2013, 117, 6225-6234.
[25] 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, R. Schlögl, A. Trunschke, ACS Catalysis 2013, 3, 1103-1113.
[26] J. Kubo, N. Watanabe, W. Ueda, Chemical Engineering Science
2008, 63, 1648-1653.
Oxidation Reactions,
1 ed. (Eds.: C. Hess, R. Schlögl), RSC
Nanoscience & Nanotechnology, Cambridge, 2011, pp. 56-95; iW.
Ueda, D. Vitry, T. Katou, Catalysis Today 2005, 99, 43-49; jH.
Launay, S. Loridant, D. L. Nguyen, A. M. Volodin, J. L. Dubois, J. M.
M. Millet, Catalysis Today 2007, 128, 176-182; kA. Chieregato, J. M.
López Nieto, F. Cavani, Coordination Chemistry Reviews 2015,
301–302, 3-23.
[4]
[5]
aP. DeSanto, Jr., D. J. Buttrey, R. K. Grasselli, C. G. Lugmair, A. F.
Volpe, Jr., B. H. Toby, T. Vogt, Zeitschrift fuer Kristallographie 2004,
219, 152-165; bX. Li, D. Buttrey, D. Blom, T. Vogt, Topics in
Catalysis 2011, 54, 614-626.
aD. Vitry, Y. Morikawa, J. L. Dubois, W. Ueda, Topics in Catalysis
2003, 23, 47-53; bR. Naumann D'Alnoncourt, L. I. Csepei, M.
Hävecker, F. Girgsdies, M. E. Schuster, R. Schlögl, A. Trunschke,
Journal of Catalysis 2014, 311, 369-385.
[6]
[7]
[8]
M. Hävecker, S. Wrabetz, J. Kröhnert, L.-I. Csepei, R. Naumann
d'Alnoncourt, Y. V. Kolen'ko, F. Girgsdies, R. Schlögl, A. Trunschke,
Journal of Catalysis 2012, 285, 48-60.
[27] R. Grabowski, J. Słoczyński, Chemical Engineering and Processing:
Process Intensification 2005, 44, 1082-1093.
[28] R. Quintana-Solórzano, G. Barragán-Rodríguez, H. Armendáriz-
Herrera, J. M. López-Nieto, J. S. Valente, Fuel 2014, 138, 15-26.
[29] aV. Sanchez Escribano, G. Busca, V. Lorenzelli, The Journal of
Physical Chemistry 1990, 94, 8939-8945; bM. McEntee, W. Tang, M.
Neurock, J. T. Yates, Journal of the American Chemical Society
2014, 136, 5116-5120; cP. G. Harrison, B. Maunders, Journal of the
Chemical Society, Faraday Transactions 1: Physical Chemistry in
Condensed Phases 1985, 81, 1345-1355; dE. Finocchio, G. Busca,
V. Lorenzelli, V. S. Escribano, Journal of the Chemical Society,
aP. Gruene, T. Wolfram, K. Pelzer, R. Schlögl, A. Trunschke,
Catalysis Today 2010, 157, 137-142; bM. D. Argyle, K. Chen, A. T.
Bell, E. Iglesia, Journal of Catalysis 2002, 208, 139-149.
aJ. Zhang, X. Liu, R. Blume, A. Zhang, R. Schlögl, D. S. Su,
Science 2008, 322, 73-77; bD. S. Suꢀ, J. Zhang, B. Frank, A.
Thomas, X. Wang, J. Paraknowitsch, R. Schlögl, ChemSusChem
2010, 3, 169-180.
13
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Faraday Transactions 1996, 92, 1587-1593; eP. Concepción, P.
Botella, J. M. L. Nieto, Applied Catalysis A: General 2004, 278, 45-
56; fR. Coast, M. Pikus, P. N. Henriksen, G. A. Nitowski, The
Journal of Physical Chemistry 1996, 100, 15011-15014.
[30] B. Frank, S. Wrabetz, O. V. Khavryuchenko, R. Blume, A.
Trunschke, R. Schlögl, ChemPhysChem 2011, 12, 2709-2713.
[31] J.-M. M. Millet, Topics in Catalysis 2006, 38, 83-92.
[32] aX. Rozanska, R. Fortrie, J. Sauer, Journal of the American
Chemical Society 2014, 136, 7751-7761; bX. Rozanska, R. Fortrie,
J. Sauer, The Journal of Physical Chemistry C 2007, 111, 6041-
6050.
[33] aD. C. Tranca, N. Hansen, J. A. Swisher, B. Smit, F. J. Keil, The
Journal of Physical Chemistry C 2012, 116, 23408-23417; bE. J.
Maginn, A. T. Bell, D. N. Theodorou, The Journal of Physical
Chemistry 1995, 99, 2057-2079.
[34] C. Batiot, B. K. Hodnett, Applied Catalysis A: General 1996, 137,
179-191.
[35] M. Eichelbaum, M. Hävecker, C. Heine, A. M. Wernbacher, F.
Rosowski, A. Trunschke, R. Schlögl, Angewandte Chemie
International Edition 2015, 54, 2922-2926.
[36] Y. V. Kolen'ko, W. Zhang, R. N. d'Alnoncourt, F. Girgsdies, T. W.
Hansen, T. Wolfram, R. Schlögl, A. Trunschke, ChemCatChem
2011, 3, 1597-1606.
[37] N. Hamilton, T. Wolfram, G. Tzolova Müller, M. Hävecker, J.
Kröhnert, C. Carrero, R. Schomäcker, A. Trunschke, R. Schlögl,
Catalysis Science & Technology 2012, 2, 1346-1359.
[38] H. Bluhm, M. Hävecker, A. Knop-Gericke, M. Kiskinova, R. Schlögl,
M. Salmeron, MRS Bulletin 2007, 32, 1022-1030.
[39] J. J. Yeh, I. Lindau, Atomic Data and Nuclear Data Tables 1985, 32,
1-155.
[40] D. Maganas, A. Trunschke, R. Schlögl, F. Neese, Faraday
Discussions, 2016, 188, 181-197.
14
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601194
ChemCatChem
FULL PAPER
Entry for the Table of Contents
Layout 1:
FULL PAPER
Molecular aspects in terms oft the
number of adsorption sites determined
by microcalorimetry, and collective
electronic properties of the bulk are
integrated in a concept that explains
differences in the performance of
single-site and bulk catalysts for
oxidation of ethane and propane.
Pierre Kube, Benjamin Frank,
Sabine Wrabetz, Jutta Kröhnert,
Michael Hävecker, Juan Velasco-Vélez,
Johannes Noack, Robert Schlögl, and
Annette Trunschke*
Page No. – Page No.
Functional Analysis of Catalysts for
Lower Alkane Oxidation
15
This article is protected by copyright. All rights reserved.