Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and its components in oxidative coupling of methane

Article abstract of DOI:10.1016/j.cattod.2016.04.021

Relationships between catalytic performance in oxidative coupling of methane (OCM) and properties of lattice oxygen of the mixed NaWMn/SiO₂ oxide and its components are studied. It is demonstrated that in hydrogen and methane flows tungsten and manganese can be practically completely reduced from W⁶⁺ to W⁰ and from Mn⁴⁺/Mn³⁺ to Mn²⁺ states, respectively. Reduction in hydrogen proceeds at substantially lower temperatures. If the system is reduced in methane, the formation of methane oxidation products (C₂ hydrocarbons, CO, CO₂, water, H₂) is observed, and the product distribution substantially changes at increasing degree of reduction. In addition to strongly-bonded oxygen which can be removed from the NaWMn/SiO₂ system by reduction, a more weakly-bonded form can be detected using temperature programmed desorption (TPD). This form of oxygen can be reversibly removed at temperatures above 640 °C and replenished at much lower temperatures. Its amount (～16 μmol O₂/g) is about 4.5% of that removable by reduction in hydrogen and methane, or about 10% of the formal oxygen surface monolayer. Using the sequential O₂/CH₄ pulse technique, it is also shown that this weakly-bonded form of oxygen can participate in the steady-state catalytic OCM reaction because its presence substantially increases the rate of selective formation of C₂ hydrocarbons and its lifetime at typical OCM temperatures (around 800 °C) exceeds by far the characteristic time of the catalytic reaction.

Full text of DOI:10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Oxygen availability and catalytic performance of NaWMn/SiO2 mixed
oxide and its components in oxidative coupling of methane
Yury Gordienko, Timur Usmanov, Viktor Bychkov, Vladimir Lomonosov,
Zukhra Fattakhova, Yury Tulenin, Dmitry Shashkin, Mikhail Sinev∗
Semenov Institute of Chemical Physics, R.A.S., 4 Kosygin Street, Moscow 119991, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Relationships between catalytic performance in oxidative coupling of methane (OCM) and properties of
lattice oxygen of the mixed NaWMn/SiO2 oxide and its components are studied. It is demonstrated that in
hydrogen and methane flows tungsten and manganese can be practically completely reduced from W6+
to W0 and from Mn4+/Mn3+ to Mn2+ states, respectively. Reduction in hydrogen proceeds at substantially
lower temperatures. If the system is reduced in methane, the formation of methane oxidation products
(C2 hydrocarbons, CO, CO2, water, H2) is observed, and the product distribution substantially changes
at increasing degree of reduction. In addition to strongly-bonded oxygen which can be removed from
the NaWMn/SiO2 system by reduction, a more weakly-bonded form can be detected using temperature
programmed desorption (TPD). This form of oxygen can be reversibly removed at temperatures above
640 ◦C and replenished at much lower temperatures. Its amount (∼16 ␮mol O2/g) is about 4.5% of that
removable by reduction in hydrogen and methane, or about 10% of the formal oxygen surface monolayer.
Using the sequential O2/CH4 pulse technique, it is also shown that this weakly-bonded form of oxygen
can participate in the steady-state catalytic OCM reaction because its presence substantially increases
the rate of selective formation of C2 hydrocarbons and its lifetime at typical OCM temperatures (around
800 ◦C) exceeds by far the characteristic time of the catalytic reaction.
Received 30 December 2015
Received in revised form 24 March 2016
Accepted 25 April 2016
Available online xxx
Keywords:
Methane
Oxidative coupling
Catalyst
Mixed oxide
Lattice oxygen
Kinetics
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
and Pb/Al2O3) [7,8]. This implies a complex and multistep charac-
ter of the catalytic cycle over NaWMn/SiO2, and also that water is
Supported mixed NaWMn/SiO2 oxide is a promising catalyst for
ethylene production from methane via its oxidative coupling [1,2].
Unlike many other systems that efficiently catalyze this reaction
and do not contain any transition metal capable of changing their
oxidation state in the course of catalytic oxidation (see, e.g. Refs.
[3–5]), this mixed oxide contains two redox cations (W and Mn)
which are able to widely vary their oxidation state. Its structural
features and catalytic performance have been extensively studied
(see, e.g., Ref. [6]). In spite of that, some important issues are still
not well understood, including the nature of its catalytic activity.
In some respect this system displays rather unusual behavior. In
particular, recently it was demonstrated that the rates of methane
and ethane oxidation tend to increase at increasing conversion and
water has a strong positive effect on both activity and selectiv-
ity of this catalytic system (unlike some others, such as La/MgO
somehow involved in it. On the other hand, it was shown that the
rates of light alkane oxidation over this catalyst can be described in
the framework of the classical Mars-van Krevelen (redox) kinetic
scheme [9]. The latter agrees with the idea that lattice oxygen,
or oxygen associated with W and/or Mn on the surface of the
NaWMn/SiO2 oxide is responsible for the C H bond activation in
methane [10–13].
It is well-known that C2+ hydrocarbons can form from methane
during the interaction of the latter with certain metal oxides as
active oxygen suppliers. According to the classical works [14–17],
manganese oxides display the superior efficiency in such mode of
operation, especially while supported on silica and doped with
alkali and phosphorus [17]. Unfortunately, the OCM selectivity
of the same systems at steady-state catalytic operation was not
satisfactory, so it was difficult to settle the question about the
relationships between redox properties and catalytic performance.
One may assume that the replacement of phosphorous in Mn-silica
system by one more redox cation (W) that dramatically improves
the OCM selectivity at steady-state operation would also positively
∗
Corresponding author.
E-mail addresses: sinev@chph.ras.ru, mysinev@rambler.ru (M. Sinev).
http://dx.doi.org/10.1016/j.cattod.2016.04.021
0920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
2
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
affect the ability to produce the OCM products in sequential redox
regime. If so, there is a chance to reveal some factors that control
the formation of the OCM products in both modes of operation.
In this work we attempted to shed more light onto the proper-
ties of the reactive lattice oxygen of this mixed oxide system, and
on the relationships between its composition, structure, oxygen
availability and catalytic performance.
Productivity Pi with respect to a certain product was calculated
as the number of methane molecules converted into this product
per unit time (second) per gram of catalyst, i.e.
Pi , mol/(s × g) = ni × Ci,out × Wout/(22400 × m)
where m–mass of the catalyst in the reactor.
2.2.2. Oxygen TPD and pulse redox studies
2. Materials and methods
These types of experiments were performed using a setup that
included a flow quartz cell placed into the thermostatic block of the
differential scanning calorimeter (Setaram “DSC-111”) that allowed
us to run experiments in both isothermal and temperature pro-
grammed regimes. This experimental setup and the on-line GC
analytic system are described in detail elsewhere [18,19].
Typically, 50 mg of fresh mixed oxides was taken for each mea-
surement. In oxygen TPD experiments, the sample was pretreated
in oxidizing gas flow (5 vol.% O2 in He) at 500 ◦C and cooled in the
same flow to 200 ◦C; then the gas flow was switched to helium,
and after purging for 30 min the temperature was linearly risen to
800 ◦C (10 K/min). Oxygen release was measured by the on-line TC-
detector. When temperature reached 800 ◦C, the recording of the
TPD curve was continued until the concentration of oxygen reaches
the detection limit.
In pulse redox experiments, sequential pulses of oxygen and
methane were supplied at 800 ◦C onto the sample in a perma-
nent pure helium flow passing through the on-line analytic system
(“Porapak Q” and “5A” molecular sieve columns, TC-detectors)
using two independent sampling valves (∼0.2 ml dosing each). In
each methane pulse concentrations of ethane, ethylene, CO, and
CO2 were measured. The time delay between O2 and CH4 pulses
supplied by two valves was varied in a wide range (from 0 to
∼1000 s). The sample was purged by helium for 30 min after each
pair of O2/CH4 pulses in order to exclude the interference from the
previously chemisorbed weakly-bonded oxygen.
2.1. Catalyst preparation
Mixed NaWMn/SiO2 oxide and samples containing its compo-
nents (Na, W, Mn) and their combinations (Na-W, Na-Mn, W-Mn)
supported on silica were obtained using two techniques: incipient
wetness impregnation (IWI) and sol-gel (SG) synthesis. In all sam-
ples used in this study, the content of metals was fixed (in wt.%): 0.8
Na, 3.2 W, 0.8 Mn (atomic ratio Na:W:Mn ≈2:1:0.835). Their phase
composition was studied using powder X-ray diffraction (DRON-2
diffractometer, Cu K␣ radiation).
2.1.1. IWI method
Appropriate amounts of inorganic salts containing Na and W
(sodium tungstate or carbonate, ammonium tungstate) were dis-
solved in distilled water and the resulted solution was added to
silica gel (Aldrich Davisil grade 646). After drying at 130 ◦C for 8 h,
if necessary, the sample was impregnated with the appropriate
amount of manganese nitrate water solution. After repeated drying,
the samples were finally calcined at 900 ◦C.
2.1.2. SG synthesis
Appropriate amounts of tetraethoxysilane, Na2WO4 and Mn
nitrate were thoroughly mixed, and after gelation and drying (48 h
at room temperature) the resulted solid precursor was finally cal-
cined at 900 ◦C.
2.2.3. Combined TGA-MS measurements
2.2. Catalyst testing and redox studies
Combined TGA-MS measurements were performed using
a high-temperature thermobalance (Setaram “Setsys Evolution
16/18”) equipped with on-line MS analyzer (Pfeiffer “OmniStar GSD
301”) [20]. The operation unit of the thermogravimetric set-up con-
sists of a vertical flow tubular furnace with an inner alumina tube.
The inner diameter of the alumina tube is 20 mm and the length
of the controlled temperature zone is ∼30 mm. The standard TGA
and DSC thermogravimetric rods could not be applied because they
contained platinum parts in their constructions, which are not inert
towards the gas reactants used in this study.
The sample under study was placed in the centre of a heating
zone and blown by the appropriate gas downward flow. The initial
gas mixture was pre-mixed in a stainless steel cylinder and its flow
rate was maintained by the flow controller of the SETSYS set-up.
In this series of experiments we stopped the heating program at
850 ◦C in order to prevent possible collapse of the reduced structure
that could take place at overheating and distort the behavior of
the system in sequential redox cycles. Since the reduction in the
case of methane was not completed during the linear heating at
temperatures below 850 ◦C, we continued the process at constant
temperature until the measurable changes in the sample weight
completed.
2.2.1. Catalytic tests
Catalytic tests were carried out using a conventional quartz
microreactor with on-line GC analysis of the reaction mixture
described elsewhere [8,9]. Typically, 50 ml/min of the mixture con-
taining 50% methane and 50% air was supplied at 860 ◦C onto
50–80 mg (0.2–0.3 mm particle size) of the catalyst (depending on
its bulk density) placed into the reactor. Free space in the reactor
before and after the catalyst bed was filled with quartz beads in
order to diminish the contribution of non-catalytic reaction in the
void volume.
Parameters of the process – methane conversion X(CH4), yields
Yi, and selectivities Si – were calculated based on the measured
values of concentrations of components and total flow of reaction
mixture entering and leaving the reactor, as follows:
X(CH4), % = 100 × (CCH4,in × Win − CCH4,out × Wout)/CCH4,in × Win
Yi, % = 100 × ni × Ci,out × Wout/CCH4,in × Win
Si, % = 100 × Yi/X(CH4)
A small part of the gas below the sample was directed via
the stainless capillary into the mass-spectrometer. The time delay
between the mass-spectrometer response and the current gas com-
position in the reaction zone was about 8 s. Ion currents at 2 (H2+),
15 (CH3+), 16 (O+, CH4+), 17 (OH+, CH5+), 18 (H2O+), 28 (CO+, C2H4+),
29 (C2H5+), 30 (C2H6+), 32 (O2+), 44 (CO2+) m/z values, as well
as ion currents at 4 or 40 m/z in the experiments when He or Ar
where: CCH4,in and CCH4,out – methane concentrations in the inlet
and outlet mixtures, respectively (vol.%); Win and Wout – total flows
of the reaction mixture entering and leaving the reactor, respec-
tively (ml/s); ni–stoichiometric coefficient (equal to the number of
methane molecules required to form one molecule of this products:
ni = 1 and 2 for carbon oxides and C2 hydrocarbons, respectively).
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
3
Fig. 1. Catalytic performance (OCM productivity and selectivity) of NaWMn/SiO2 mixed oxide obtained by IWI and SG methods and its components.
were used as balance gases (carrier gases) in the initial gas mix-
ture were detected in a semi-permanent mode (duty cycle < 1 s) in
online measurements. The composition of the reaction mixture was
determined based on the corresponding ion current values.
3. Results and discussion
3.1. Phase composition and catalytic performance
The data on the phase compositions of the fresh samples are
in a good agreement with corresponding literature data (see, e.g.,
Refs. [11,13,21]). The samples of the same composition obtained
by impregnation and SG synthesis have similar phase composition:
all Na-free samples are amorphous, whereas in the presence of Na
the formation of ␣-cristobalite as the main phase is observed; also
Na2WO4 and Mn-oxides as minor components are detected.
Also in agreement with existing literature data, the samples con-
taining all three dopants (Na, W, Mn) demonstrate high selectivity
Fig. 2. TG (1, 3) and DTG (2, 4) curves obtained during the reduction of NaWMn/SiO2-
IWI sample in hydrogen (1, 2) and its reoxidation (3, 4).
(above 70%) to C2-hydrocarbons at methane conversions exceed-
ing 15% (see Fig. 1). At the same chemical composition, the catalyst
obtained by the IWI method is substantially more active compared
to the one obtained via SG route.
The performance of the one- and two-component systems is
substantially poorer. Na-free (amorphous) samples are much less
selective, whereas the absence of redox cations (W and especially
Mn) makes the samples much less active compared to the ones
containing them (see Fig. 1). In other words, the presence of all
three dopants in the system is crucial for the formation of catalytic
systems with optimal catalytic properties.
between Mn4+ and Mn3+. In this case the ratio of the amounts of
oxygen originated from tungsten and manganese reduction is 1:
(0.14–0.28).
The shape of the sample reduction curve obtained with methane
is more complex. A significant weight loss starts at temperature
above 600 ◦C; simultaneously the formation of methane oxidation
products is observed–correspondingMS signals increase along with
weight loss at rising temperature. As can be seen at Fig. 3, the
two-step character of the reduction in this case is much more pro-
nounced. The weight loss measured in the first step corresponds to
∼105 ␮mol O2 per gram, i.e. about one third of the total amount
removed by hydrogen.
It is worth noticing that in two steps of sample weight loss in
methane different gaseous products are forming. In the first step of
reduction that proceeds between 600 and 820 ◦C the formation of
water (m/z = 18), ethane (m/z = 29 and 30), ethylene (m/z = 26, 27),
and CO2 (m/z = 44) is taking place. As to the ions with m/z = 28, in
this temperature region they can be attributed to both ethylene and
CO.
3.2. Reduction and reoxidation
Our TGA/MS experiments demonstrate that the optimal
NaWMn/SiO2 catalyst obtained by IWI method can be reduced
in hydrogen and methane flows and re-oxidized in O2-containing
gas flow in the conditions of linear heating. As can be seen on
Figs. 2 and 3, hydrogen is a much more active reducing agent for this
system compared to methane: in H2 flow NaWMn/SiO2-IWI sam-
ple starts losing weight above 300 ◦C; the reduction becomes more
intense above 500 ◦C (see Fig. 2). It is accompanied by the forma-
tion of water. The measured weight loss corresponds to ∼345 ␮mol
O2/g, which is close to the equivalent of W and Mn cations reduc-
tion to W0 and Mn2+ states assuming that all tungsten was initially
in the W6+ state and the average oxidation state of manganese was
The evolving gas composition in the second reduction step is
substantially different. First, no significant amount of water is form-
ing: in spite of the weight loss comparable to that observed in
the first step of reduction, the relative intensity of the signal with
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
4
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 3. TG curve and MS-signals (ion currents) obtained during the reduction of NaWMn/SiO2-IWI sample in methane and their assignments: (1) m/z = 2 (H2), (2) m/z = 18
(H2O), (3) m/z = 26 (ethylene fragmentation), (4) m/z = 27 (ethylene fragmentation), (5) m/z = 28 (ethylene + CO), (6) m/z = 29 (ethane fragmentation), (7) m/z = 30 (ethane), (8)
m/z = 44 (CO2); for curves (2) and (5) the scale is reduced by factor of 5.
m/z = 18 is almost negligible. At the same time, the formation of
hydrogen (m/z = 2) is clearly seen. It is interesting to analyze the
behavior of the signal with m/z = 28. As can be concluded from Fig. 3,
at temperatures above 830 ◦C there is no correlation between the
signals with m/z = 28, that can be assigned to molecular ions of both
ethylene and CO, and with m/z = 26 and 27, which definitely orig-
inate from C2 molecule (likely ethylene) fragmentation. Thus, we
can conclude that the major part of the ion current at m/z = 28 at
830–850 ◦C is due to the presence of carbon monoxide in the gas
mixture. Also, the formation of carbon dioxide (m/z = 44) in the sec-
ond step of oxidation is more pronounced than in the first one. In
other words, selectivity with respect to the formation of different
gaseous products from methane during its oxidation with the lat-
tice oxygen of the NaWMn/SiO2 mixed oxide substantially differs
in two stages of the oxide reduction.
cisely evaluated based on this number due to the sample coking
(see below).
After the reduction in both hydrogen and methane, the sample
can be reoxidized in O2-containing gas flow. The major features of
the reoxidation processes are seen at Fig. 4A, B. In both cases the
sample reoxidation starts at temperatures above 200 ◦C. The differ-
ence in the shape of reoxidation curves obtained after reduction in
hydrogen (Fig. 4A) and methane (Fig. 4B) is as follows: in the first
case (H2) reoxidation curve is smoother. If the sample is reduced
in methane, after the period of slowly growing oxidation rate, its
sharp increase at about 440 ◦C is observed. However, in both cases
the rate of reoxidation reaches its maximum at 480–510 ◦C and
then falls. Also very similarly in both cases the second peak of oxy-
gen uptake is observed between 600 and 700 ◦C, after which the
reoxidation completes.
It can be also mentioned that the kinetic features of C2 hydrocar-
bons formation (m/z = 26, 27, 29, and 30) differ from those observed
in the second reduction stage for carbon oxides: whereas the sig-
nals with m/z = 28 and 44 start to decline right after the heating
program stops (i.e., when the process becomes isothermal), the for-
mation of C2 hydrocarbons, particularly ethylene, keeps growing at
increasing degree of reduction. There are also more tiny effects in
the behavior of different MS signals (for instance, a more delayed
maximal hydrogen evolution compared to CO), but the elucidation
of their origin requires additional oriented experiments.
The total amount of oxygen (weight gain) consumed in re-
oxidation step practically coincides with the weight loss observed
in the case of reduction with hydrogen (see Fig. 2). So we can con-
clude that this very deep reduction of the sample (almost complete)
is reversible in the sense of the amount of oxygen that can be
removed and replenished in the redox cycle.
In the case of reduction with methane such balance is compli-
cated because of the sample coking during the reduction. As we
already mentioned, the total measured weight change in experi-
ments with two reducing agents is different: in the case of methane
it is somewhat less, and certain difference is observed in the parallel
runs. Also, during the reoxidation run we observed the formation of
carbon oxides (m/z = 28 and 44, see Fig. 4C) in the same temperature
region where the most intense oxygen uptake is taking place. It is
worth noticing that no significant formation of water is observed.
Since the heating program was terminated at 850 ◦C, we only
can say that the second step of reduction starts at about 830 ◦C and
proceeds in isothermal conditions at 850 ◦C until its rate becomes
negligible. In the case of reduction in methane flow, the final weight
loss in less, but the real amount of oxygen removed cannot be pre-
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
5
dehydrogenation [26] accompanied with the formation of molec-
ular hydrogen and carbonaceous deposits–the source of carbon
oxides during the reoxidation. The data on the gas composition (see
Fig. 3), namely the formation of significant amounts of molecular
hydrogen simultaneously with C2 hydrocarbons and in compara-
ble amounts with them, indicate that the formation of coke via the
decomposition of CH3 radicals is taking place in the second step of
reduction (above 830 ◦C), although the direct methane decomposi-
tion also cannot be ruled out.
Also, a high exothermicity of oxidation of such deposits (coke)
can explain the difference in the shape of reoxidation curves for
two different samples, i.e. reduced in hydrogen and in methane.
One may assume that the sharp increase of reoxidation rate in the
second case is due to additional “in situ” heat evolution due to the
coke oxidation that leads to the sample overheating and process
acceleration.
Despite some differences in the reoxidation patterns for the
samples reduced in hydrogen and methane, the general features
of the process are very similar, so it is very likely that the state of
the solid reduced with these two reducing agents is the same. As
we mentioned above, this reduction is reversible in the sense of
the amount of oxygen that can be removed and replenished in the
redox cycle. This does not mean that in the course of re-oxidation
the sample returns exactly to the initial state, i.e. that not only its
stoichiometry, but also its initial structure fully restores. However,
according to the obtained data, the rate of re-oxidation at typi-
cal OCM temperatures is much higher than the rate of reduction
in methane, so such deep reduction is not relevant for the steady-
state operation and we believe that the bulk structure of this system
should not suffer from it.
3.3. Oxygen TPD
The multi-step character of NaWMn/SiO2 reduction and reox-
idation described in the previous section suggests non-uniform
properties of strongly-bonded (’lattice’) oxygen that can be
removed by reduction; also its properties may change in the course
of the redox treatments.
Additional information about the forms of oxygen bound with
this system was obtained in our oxygen TPD experiments. During
the heating of the pre-oxidized NaWMn/SiO2 (IWI) sample in a He
flow no oxygen evolution is observed at temperatures below 600 ◦C.
Thus, one may assume that there is no significant amount of oxygen
in weakly adsorbed forms present on the surface of NaWMn/SiO2
mixed oxide.
A detectable release of oxygen to the He flow starts at tem-
perature about 640 ◦C (see Fig. 5). During the linear heating and
isothermal exposure at 800 ◦C the sample releases about 16 ␮mol
O2 per gram, i.e. about 4.5% of the amount that can be removed by
the reduction in hydrogen. Taking into account the specific surface
area of the IWI NaWMn/SiO2 sample (4.3 m2 g−1), it corresponds to
about 10% of the formal surface monolayer.
The characteristics of this TPD peak (temperature range, shape,
amount of O2 evolved) are very well reproduced in the repeated
cycles of reoxidation-desorption. The latter suggests that the
removal of this oxygen from the sample if fully reversible and does
not lead to any significant changes in its structure and reactivity.
The shape of the TPD curve requires some additional discussion.
As can be seen, the shape of the curve on Fig. 5 can be represented
as superposition of the “plain” oxygen release that starts around
640 ◦C and the “hump” that is always observed on it. This “hump”
may be associated with the presence of different phases contain-
ing W and Mn in the NaWMn/SiO2 system (see above Section 3.1).
They may release oxygen at different temperatures with differ-
ent intensity. Also, since one of the phases present in the sample,
namely sodium tungstate, melts at 698 ◦C [27], and this tempera-
Fig. 4. DTG curves obtained during the reoxidation of NaWMn/SiO2-IWI sample
after reduction in hydrogen (A) and in methane (B) and MS-signals (ion currents) of
CO (m/z = 28) and CO2 (m/z = 44) obtained during the reoxidation after reduction in
methane (C).
As can be seen from Fig. 4A, variations of the MS signal attributed
to water (m/z = 18) does not exceed substantially the noise level.
Low evolution of water during reoxidation after reduction with
methane strongly suggests that carbon oxides are forming mainly
not from any adsorbed fragments of methane molecule (CHX, where
X = 1–3), but from carbonaceous deposits with very low hydrogen
content.
The most evident explanation of the simultaneous formation
of H2 and surface carbon deposits would be a decomposition of
methane on the reduced metal (tungsten) surface. Supported W
and Mo metals and their carbides are known to catalyze a non-
oxidative methane conversion to higher hydrocarbons [22–24].
Alternatively, on non-supported metals methane decomposes to
H2 and carbon [22]. However, the latter process is highly activated.
But, as said above, during the reduction with methane substan-
tial amounts of OCM products (C2 hydrocarbons) are forming, and
methyl radicals are well known to be reactive intermediates in
the formation of OCM products from methane. According to the
previous studies [25,26], metal molybdenum (which is relative
to tungsten and behave in a similar way in analogous circum-
stances, particularly–in the course of interaction with methane in
the absence of oxygen) can actively capture methyl radicals in a
wide range of temperatures. Whereas such capturing at low tem-
perature leads to the formation of methane, H2 and CO [25], at
elevated temperatures adsorbed radicals undergo rapid sequential
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
6
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 5. Oxygen desorption from pre-oxidized NaWMn/SiO2-IWI sample.
ture is very close to the position of the ‘hump’ maximum on the TPD
curve. The latter forces us to assume that the more intense oxygen
release is stimulated by this phase transition in one of the compo-
nents present in the system. Such behavior known as the Hedvall
effect is often observed in the course of chemical transformations in
condensed phases, especially conducted in the programmed heat-
ing regime. A general explanation of this effect is associated with
spontaneously increasing mobility of structural units of solids reac-
tants in the area of parameters (first of all, temperature) close to
the phase transition point. A more thorough examination of the
observed behavior would provide with a deeper understanding.
The presence of free sodium tungstate as a separate phase and its
volatility also may cause some change in the chemical composition
in the course of experiments described below. However, since we
did not observe any substantial change in the redox behavior in
repeated cycles, we can conclude that either the loss of tungsten in
the course of these experiments is negligible, or its loss is within
the range where it has no effect onto the behavior under study.
The TPD experiments of same type with other systems showed
that (see Fig. 6):
On the one hand, at this temperature the OCM reaction efficiently
proceeds over this catalyst in a steady-state regime. On the other
hand, according to our oxygen TPD data, lifetime of the removable
oxygen form may be within the range of time delays between pulses
of different gases accessible in our experiments. Also, according
to the TG/MS data, at this temperature methane can interact with
strongly-bonded lattice oxygen.
The results presented at Fig. 7 demonstrate that indeed methane
oxidation products are formed in such regime. Their concentra-
tions sharply decrease at increasing delay between oxygen and
methane pulses (␶) in a He flow. The rate of the formation of C2-
hydrocarbons decreases by the factor of ∼ 3.7, whereas that of CO2
and CO decreases 4.2 times and more than 6 times, respectively.
Correspondingly, selectivity to C2-hydrocarbons increases from 89
to 91%, which may reflect the decreasing contribution of the consec-
utive oxidation of OCM products to carbon oxide at their decreasing
concentration.
Taking into account the results of TG/MS and oxygen TPD exper-
iments, we can assign the rates of methane oxidation in alternating
O2/CH4 pulse regime to its interaction with two different forms of
oxygen. Indeed, in two extreme cases we have opposite situations
from the standpoint of the presence of different forms of oxygen
and their interaction with methane:
– the amount of oxygen desorbed from the sample of the same
composition, but obtained via the SG synthesis is substantially
less, and also decreases in the second cycle, unlike for the sample
obtained by the IWI method;
– oxygen release from the samples of different compositions (con-
taining one or two dopants) is by far less pronounced, less
reversible in the repeated cycles, and in some cases is negligible.
(1) ␶ → ∞–corresponds to the state in which the sample has
released the whole amount of oxygen that can be removed in
the TPD regime; as we see from Fig. 7, this state can be char-
acterized by a relatively low rate of interaction with methane
and high OCM selectivity;
(2) ␶ = 0–in this case two reactants are present in the gas phase
simultaneously, i.e. we mimic a catalyst steady-state operation;
Fig. 7 shows that in this state the rate of methane conver-
sion is the highest, although the OCM selectivity is lower than
at increasing ␶; in this state the content of the form of reac-
tive oxygen that leaves the sample in the TPD regime reaches
the maximum that can be achieved in the steady-state regime
at given temperature; correspondingly, its contribution to the
overall rate of methane conversion in pulse regime is maximal
and likely predominant.
Thus, these results demonstrate that the oxygen TPD behavior
of the samples under study strongly correlates with their catalytic
performance in the OCM reaction.
3.4. Sequential pulse oxidation/reduction
As our oxygen TPD experiments demonstrate, the most catalyt-
ically efficient system under study, i.e. NaWMn/SiO2-IWI sample,
can reversibly release relatively weakly bonded oxygen at tem-
peratures close to the temperature of the steady-state catalytic
operation. The following experiments were done to find out
whether this oxygen form can interact with methane, in other
words, to what extent catalytic activity can be assigned to the pres-
ence of it in the solid.
In the intermediate state the rate of reaction should reflect
the decreasing concentration of the weakly-bonded oxygen. One
may assume that the rate of methane interaction with strongly-
bonded oxygen (that cannot be removed in the TPD regime) does
not depend on the time between oxygen and methane pulses; con-
sequently, its relative contribution to the formation of products
In this series of experiments we supplied sequential pulses
of oxygen and methane onto the above mentioned sample pre-
oxidized in oxygen and purged for about 1 h in a He flow at 800 ◦C.
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
7
Fig. 6. Amounts of oxygen released from samples of different composition (or preparation) in TPD regime in two first cycles of oxidation-desorption.
Fig. 7. Concentrations of different methane oxidation products formed during the pulse reduction of NaWMn/SiO2-IWI sample as the function of delay between oxygen and
methane pulses.
increases at increasing delay between oxygen and methane pulses.
Assuming that the reactivity of the weakly-bonded oxygen is uni-
form, i.e. the rate constant of methane interaction with it does not
depend on its surface concentration (coverage), one may write:
kinetics curves would give straight lines in different coordinates
according to the integral forms of first-order and second-order
kinetic equations. Taking into account Eq. (1), corresponding equa-
tions should look as follows:
Ln[(C0 − C )/(C − C )] = k␶
(2)
(W − W )/(W − W ) = (C − C )/(C − C ) = ␪ /␪
(1)
∞
␶
∞
␶
∞
∞
␶
∞
∞
␶
0
0
0
in the case of the first-order kinetics and
where: ␶–delay between oxygen and methane pulses; C , C ,
␶
0
C
–current total concentration of methane oxidation products
∞
(C0 − C )/(C − C ) − 1 = k × ␶
(3)
∞
␶
∞
at given ␶, their concentration at ␶ = 0 and ␶ → ∞; W , W ,
␶
0
W
–corresponding rates of methane interaction with the sample;
in the case of the second-order kinetics.
∞
␪ and ␪ –current surface concentration of weakly-bonded oxygen
The validity of the two above descriptions can be evaluated from
Fig. 8. It is clear that the experimental data can be well described in
terms of the second-order kinetics of weakly-bonded oxygen des-
orption, whereas the first-order kinetic equation cannot be applied
to this process. This suggests that this oxygen is likely present on
the surface in the form of atomic species.
According to the obtained data, the content of the weakly-
bonded oxygen in the sample is variable in the time scale of
101–102 s. On the other hand, in a steady-state regime at the same
␶
0
and its initial value (at ␶ = 0).
Eq. (1) allows one to analyze the kinetics of weakly-bonded oxy-
gen desorption based on the data of Fig. 7 and to express some
notions concerning its state. Indeed, if this oxygen is present on
the surface in a molecular form, one may anticipate the first-
order kinetics of its desorption. On the contrary, the desorption of
atomic surface oxygen as O2 molecule more probably should follow
the second-order kinetics with respect to ␪ . Correspondingly, the
␶
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021

G Model
CATTOD-10169; No. of Pages8
ARTICLE IN PRESS
8
Y. Gordienko et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 8. Kinetics of the total product concentration decay in coordinates of Eqs. (2) and (3).
temperature the characteristic time of the catalytic cycle is much
shorter (10−1–1 s). This suggests that this weakly-bonded oxygen
can be responsible for the catalytic activity and in the conditions
of kinetic measurements [9] (relatively low temperatures ≥ 640 ◦C,
short residence times, low conversions) the NaWMn/SiO2 system
behaves like a typical redox catalyst.
The set of observations presented here suggests that two forms
of ‘lattice’ oxygen (strongly-bonded and relatively weakly-bonded)
present in the NaWMn/SiO2 oxide system can be responsible for its
catalytic performance in OCM reaction. Although the contribution
of the reduction of strongly-bonded oxygen to the overall rate of
reaction in the steady-state regime cannot be ruled out, the elucida-
tion of the role of these two forms of active oxygen in the catalytic
performance requires further kinetic analysis.
References
[1] X.P. Fang, S.B. Li, J.Z. Lin, Y.L. Chu, J. Mol. Catal. (China) 6 (1992) 255–261.
[2] D.J. Wang, M.P. Rosynek, J.H.J. Lunsford, Catalysis 155 (1995) 390–402.
[3] O.V. Krylov, V.S. Arutyunov, Oxidative Transformations of Methane, Nauka,
Moscow, 1998 (in Russian).
[4] R.M. Navarro, M.A. Pen˜a, J.L.G. PeFierro, Metal Oxides, Chemistry and
Applications, in: J.L.G. Fierro (Ed.), CRC Press, 2005, pp. 463–490.
[5] U. Zavyalova, M. Holena, R. Schlögl, M. Baerns, ChemCatChem 3 (2011)
1935–1947.
[6] S. Arndt, T. Otremba, U. Simon, M. Yildiz, H. Schuber, R. Schomäcker, Appl.
Catal. A Gen. 425–426 (2012) 53–61.
[7] K. Takanabe, E. Iglesia, Angew. Chem. Int. Ed. 47 (2008) 7689–7693.
[8] V. Lomonosov, Yu. Gordienko, M. Sinev, Top. Catal. 56 (2013) 1858–1866.
[9] V.I. Lomonosov, Yu.A. Gordienko, M.Yu. Sinev, Kinet. Catal. 54 (2013) 451–462.
[10] Z.C. Jiang, H. Gong, S.B. Li, Stud. Surf. Sci. Catal. 112 (1997) 481–490.
[11] A. Palermo, J.P.H. Vazquez, A.F. Lee, M.S. Tikhov, R.M.J. Lambert, Catalysis 177
(1998) 259–266.
[12] V. Salehoun, A. Khodadadi, Y. Mortazavi, A. Talebizadeh, Chem. Eng. Sci. 63
(2008) 4910–4916.
[13] Z.C. Jiang, C.J. Yu, X.P. Fang, S.B. Li, H.L. Wang, J. Phys. Chem. 97 (1993)
12870–12875.
5. Conclusions
[14] H.L. Mitchell, R.H. Waghorne, Pat. 4205194 USA, 1980.
[15] G.E. Keller, M.M. Bhasin, J. Catal. 73 (1982) 9–19.
[16] C.A. Jones, J.J. Leonard, J.A.J. Sofranko, J. Catal. 103 (1987) 302–310.
[17] C.A. Jones, J.J. Leonard, J.A. Sofranko, J. Catal. 103 (1987) 311–319.
[18] M.Yu. Sinev, V.Yu. Bychkov, V.N. Korchak, O.V. Krylov, Catal. Today 6 (1990)
543–549.
[19] V.Yu. Bychkov, M.Yu. Sinev, Kinet. Catal. 40 (1999) 819–835.
[20] V.N. Bychkov, V.Yu. Tyulenin, Yu.P. Korchak, E.L. Aptekar, Appl. Catal. A Gen.
304 (2006) 21–29.
[21] A.G. Dedov, A.S. Loktev, A.A. Tyunyuaev, V.A. Ketsko, K.V. Parkhomenko, Appl.
Catal. A Gen. 406 (2006) 1–12.
[22] F. Solymosi, J. Cserényi, A. Szöke, T. Bansági, A. Oszkó, J. Catal. 165 (1997)
150–161.
1 The efficiency of the NaWMn/SiO2 mixed oxide system as the
OCM catalyst and its components depends on its composition
and correlates with its structural features and properties of its
lattice oxygen.
2 The most efficient NaWMn/SiO2 catalyst can be reversibly
reduced with hydrogen and with methane at temperatures close
to the typical OCM temperature range (above 640 ◦C).
3 In the TPD regime, ∼16 ␮mol O2/g (or about 10% of the formal
oxygen surface monolayer) can be reversibly removed from the
catalyst. Due to its high reactivity, its contribution to the catalytic
activity in steady-state regime like prevails over the contribution
of strongly-bonded oxygen.
[23] J.L. Zeng, Z.T. Xiong, H.B. Zhang, G.D. Lin, K.R. Tsai, Catal. Lett. 53 (1998)
119–124.
[24] S.I. Galanov, L.N. Kurina, M.Yu. Smirnov, O.I. Sidorova, E.V. Bezrukov, S.S.
Novikov, Pat. 2224734 (Russian Federation), 2004.
[25] B.R. Parker, J.F. Jenkins, P.C. Stair, Surf. Sci. 372 (1997) 185–192.
[26] M.U. Kislyuk, I.I. Tret’yakov, V.I. Savkin, M.Yu. Sinev, Kinet. Catal. (Russ.) 41
(2000) 61–70.
[27] CRC Handbook Chemistry and Physics, in: David R. Lide (Ed.), 85th edition,
CRC Press, 2004.
Acknowledgement
Authors acknowledge the financial support from the Russian
Foundation for Basic Research (Project No. 13-03-00668 A).
Please cite this article in press as: Y. Gordienko, et al., Oxygen availability and catalytic performance of NaWMn/SiO2 mixed oxide and
its components in oxidative coupling of methane, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.04.021