Influence of the mobility of oxygen in a complex oxide carrier on the mechanism of partial oxidation of methane

Article abstract of DOI:10.1134/S1070363208110406

The catalytic partial oxidation of methane was studied over single channels of monolith catalysts Pt/PrCeZrO/α-Al₂O₃ and Pt/GdCeZrO/α-Al₂O₃ using the temporal analysis of products (TAP) and kinetic transients. E

Full text of DOI:10.1134/S1070363208110406

ISSN 1070-3632, Russian Journal of General Chemistry, 2008, Vol. 78, No. 11, pp. 2191–2202. © Pleiades Publishing, Ltd., 2008.
Original Russian Text © E.L. Gubanova, A. Van Veen, C. Mirodatos, V.A. Sadykov, N.N. Sazonova, 2008, published in Rossiiskii Khimicheskii Zhurnal,
2008, Vol. 52, No. 1, pp. 21–31.
Influence of the Mobility of Oxygen in a Complex Oxide Carrier
on the Mechanism of Partial Oxidation of Methane
E. L. Gubanovaa, A. Van Veenb, C. Mirodatosb, V. A. Sadykova, and N. N. Sazonovaa
aBoreskov Institute of Catalysis, Siberian Branch, Russian Academy of Sciences,
pr. Lavrent’eva 5, Novosibirsk, 630090 Russia
e-mail: sadykov@catalysis.ru
b Institute for Research in Catalysis and Environment, Lyon, France
Received October 1, 2007
Abstract—The catalytic partial oxidation of methane was studied over single channels of monolith catalysts
Pt/PrCeZrO/α-Al2O3 and Pt/GdCeZrO/α-Al2O3 using the temporal analysis of products (TAP) and kinetic
transients. Effects of catalyst composition, oxidation state, time offset between O2 and CH4 pulses on activity,
selectivity and dynamics of product formation were elucidated. Realization of the direct pyrolysis–СH4 partial
oxidation route was reliably established. This route is favored by optimum lattice/surface oxygen mobility and
reactivity controlled by the dopant type (Gd, Pr) and oxidation state of the complex cerium/zirconium oxide.
DOI: 10.1134/S1070363208110406
catalytic layer, after which the temperature falls down
sharply in the main part of the catalytic layer due to
endothermic steam and carbon dioxide methane
reforming reactions. Probably, high temperatures at the
inlet and efficient heat transfer along the catalyst layer
in the endothermic zone pose serious problems in
terms of catalyst stability and reactor design. In this
connection of interest is to develop reactors operating
at short contact times, in which syngas is manufactured
by the pyrolysis–direct oxidation scheme [8–10].
Recent research show that the use such active
components as precious metals (platinum, rhodium)
applied to complex oxides of the fluorite structure
(Ce–Zr–Me–O, Me = Pr, Gd, La) can favor direct
oxidation of methane into syngas [11, 12]. The
question of whether the reaction mechanism depends
on such factors as mobility of oxygen in the lattice of
oxide carriers or short-lived surface compounds is still
under discussion. One of the reasons for the limited
knowledge of the active component structure–catalyst
activity interrelationship is that short-contact-time and
high-temperature reaction mechanisms are difficult to
study. In conventional reactors, reliable results are
hardly obtainable because of heat and mass transfer
which take place even in dilute mixtures and at high
feed rates. On the other hand, use of spectral methods
(IR spectroscopy, etc.) is prevented by strong radiation
INTRODUCTION
In the medium term the greater accessibility of
natural gas compared to oil resources has stimulated
development of technologies for its processing into
valuable products [1]. In industry natural gas is used
for manufacture of syngas (a mixture of СО and Н2)
which is processes into high-cost products (alcohol,
synthetic fuels, ammonia).
The manufacture of syngas by endothermic
methane steam reforming features a high capital
capacity and
a
low energy efficiency [2, 3].
Exothermic partial oxidation of natural gas is
energetically more efficient [4]. Moreover, with
oxygen as oxidizer, the Н2/СО ratio in the resulting
syngas is close to two, which is quite suitable for the
manufacture of methanol or liquid fuels [3].
Two reaction routes have been suggested for the
catalytic partial oxidation of methane into syngas [5–7]:
(1) complete oxidation of part of methane into СО2
and water, followed by steam and carbon dioxide
reforming of the remaining methane molecules;
(2) direct pyrolysis–selective oxidation of natural gas.
In the first route, all oxygen is produced, with a
strong heat release, in a narrow inlet part of the
2191

2192
GUBANOVA et al.
of samples at operating temperatures and extremely
short lifetimes of intermediate surface compounds.
In the present communication we report the first
results of TAP and relaxation research on the
mechanism of partial oxidation of methane on a
complex active component Pt/Ce–Zr–Me–O (Me =
Gd, Pr) incorporated in a block catalyst. Physico-
chemical and catalytic properties of powdered catalysts
have been studied in detail and described in [11, 18–22].
The pulse method of studying the dynamics of
product formation on pulse feeding the reagents in a
vacuum (Temporal Analysis of Products, TAP) seems
to provide one of the most promising approaches for
mechanistic studies [13, 14]. This method allows one
to assess the effect of the oxidation level of a catalyst
on its activity in direct methane oxidation, which is a
challenging problem in the case of fast reactions at
atmospheric pressure. The TAP technique makes it
possible to fix a certain oxidation level of a catalyst
and then to test the reactivity of the latter by its
exposure to small methane pulses which scarcely
affect the surface. Since the quantities of reactants in a
pulse are very small, the TAP technique allows the
effect of the oxidation level of the surface to be
explored without complicating heat transfer effects.
TAP data should be analyzed together with steady-
state characteristics of catalyst activity at atmospheric
pressure and with results of non-steady-state relaxation
experiments on changing one mixture to another [15].
First we studied the oxidized state of catalyst
surface. This is important for understanding processes
that occur at the inlet of block catalysts, where the
highest oxygen concentration and temperatures are
observed, and the catalyst works with maximum
performance. At a low primary selectivity of the partial
oxidation of methane, the final conversion of methane
will be determined by the endothermic steam and
carbon dioxide conversions of methane unreacted with
oxygen. Under adiabatic conditions, these processes
are controlled by a fairly slow heat transfer from the
inlet part of the layer, which requires large-size
reactors. Since oxygen adsorption occurs much faster
than methane conversion on the surface [23], then,
obviously, the active component in the inlet part of the
layer will be mostly present in the oxidized state. Thus,
the first criterion to rely on in choosing an active
component for partial methane oxidation at short
contact times is high primary CO selectivity of the
oxidized surface.
Until now TAP has been applied for research on
powered catalysts placed as thin layers between layers
of inert particles [13, 14]. For partial oxidation of
natural gas at short contact times, block catalysts have
been suggested, in which active components are
applied as a thin layer onto walls of a ceramic or
metallic carrier. The way in which active components
(for example, platinum metals) are applied on even
such relatively inert carriers as corundum or zirconium
dioxide much affects their dispersity and reactivity.
Active component–carrier interactions take even
greater significance when these materials have a
complex chemical composition [12]. The kinetics of
the partial oxidation of methane into syngas at short
contact times on block catalysts was proposed to be
described using small structural elements of the blocks,
such as short triangular channels (hollow trihedral
prisms) [16]. This allows, due to the small amount of
applied catalyst, to minimize and control temperature
gradients and to provide required gas flow over the
channel. This structural element is similar in design to
Forzatti’s reactor [17] in which the catalyst layer has
the shape of a ring on the surface of inner inert ceramic
tube, and narrow intertube space is bounded by gas
flow. However, Forzatti’s reactor was used exclusively
for steady-state kinetic studies on hydrocarbon
oxidation at atmospheric pressure.
EXPERIMENTAL
The preparation of complex fluorite-type
Pr0.3Ce0.35Zr0.35Ox and Gd0.3Ce0.35Zr0.35Ox oxides by a
modified Pekini technique is described in [18–22]. The
choice of these compositions was motivated by the fact
that the mobility of oxygen in their lattices, as
estimated by means of dynamic isotope exchange [11,
21, 22], is much different. After decomposing in air the
complexes of metals with esters derived from
polyethylene glycol and citric acid were calcined in air
at 500°C for 4 h. The resulting powders were used for
applying on corundum carrier channels (see below)
and also as carriers for disperse Pt catalysts. Platinum
was applied by impregnation (by moisture capacity)
from an H2PtCl6 solution followed by drying and
calcination at 500°C for 2 h.
Separate structural elements of block catalysts were
cut out of a honeycomb block carrier (α-Al2O3)
calcined at 1300°C (specific surface area 3 m2 g–1) as
triangular channels (length 10 mm, wall thickness
0.2 mm, and triangle side length 2.33 mm) [16].
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

INFLUENCE OF THE MOBILITY OF OXYGEN
2193
Powdered Pr0.3Ce0.35Zr0.35Ox and Gd0.3Ce0.35Zr0.35Ox
mixed oxides were peptized in a weak nitric acid (рН
~3) doped with 3 wt % of polyethylene glycol as
surfactant for ultrasonication (T-25 disperser), until a
stable zol with the solid phase concentration of 10 wt %
formed. This zol was used for impregnating channels.
After every impregnation stage the samples were
calcined at 900°C in air, and the whole procedure was
repeated until the applied oxide concentration reached
8–10 wt %. Platinum (1.4 wt %) was applied by
impregnation (by moisture capacity) from an H2PtCl6
solution followed by drying and calcination in air 900°C.
quartz reactors by the procedures described in detail in
[12, 16]. The component concentrations were deter-
mined by chromatography and by means of infrared
absorption sensors for СО, СО2, and CH4, a
conductometric sensor for oxygen, and a polarographic
sensor for hydrogen. The samples were preconditioned
under О2 at 700°C for 1 h.
Relaxation experiments at atmospheric pressure
were performed in the same device in small quartz
reactors. After preconditioning oxygen was displaced
with helium (control by the zero point of the oxygen
sensor) and then helium was displaced with a required
reaction mixture. The composition of the mixture was
continuously analyzed. Control experiments with a
single-channel fragment of the corundum carrier
showed that, at gas flow rates of about 30 l h–1, the
purge time of the system is no longer than 2–4 s.
Oxygen–surface bond strength was characterized
by means of temperature-programmed desorption
(TPD). Powdered active component, 20 mg, was
placed in an isothermal zone of a high-temperature
TAP reactor between layers of ground quartz (fraction
0.25–0.5 mm).
RESULTS AND DISCUSSION
The same reactor was used to study the catalytic
activity of separate structural elements (10 mm in
length) under dynamic vacuum conditions. The
reference gas in all experiments was argon. Before
experiments, samples were heated to 800°C and
treated with O2/Ar pulses to remove residual carbon.
The catalytic properties were assessed at 700°C by
feeding pulses of O2/Ar and CH4/Ar mixtures
(component ratio 1:1 in each mixture). First an oxygen
pulse was fed and then, after a short delay (so-called
flushing–testing mode), a methane pulse. The ratio
between oxygen and methane pulses was such that the
quantity of methane was double that of oxygen at a
pulse size of 4.1×1015–6.8×1015 molecules. Before
TAP experiments with oxygen the sample was treated
with O2/Ar pulses (pulse size 3.6×1014–2.4×1015
molecules) until dynamic saturation at every
temperature. After that, not interrupting pulse feeding,
the temperature was lowered to 250°C. Before starting
temperature-programmed heating at a rate of 20°C/min,
the sample was exposed to high vacuum for 10 min to
remove weakly bound oxygen. Analysis of responses
was performed on a quadrupole mass spectrometer
with continuous control of m/e 32 (O2), 16 (CH4), 44
(CO2), 28 (CO), 2 (H2), and 40 (Ar). The CO responses
were obtained from an analysis of m/e 28 ion currents
minus the contribution of ionized fragments of CO2
molecules, after the corresponding expansion of the
ion currents of these molecules according to their
fragmentation scheme.
Steady-State Catalytic Characteristics
The catalysts studied in the range 650–800°C at a
stoichiometric methane/oxygen ratio in the starting
mixture (СН4 4.5–20%) and contact times 4–15 μs
ensured complete oxygen conversion. The conversion
of methane and the selectivity in syngas varied over a
wide range depending on the conditions of the process
and the nature of the catalyst. Table 1 lists activity and
selectivity data for the catalysts studied (contact time
15 ms, starting mixture 7% СН4 + 3.5% О2 in N2.
Effect of temperature on the conversion and selectivity of
partial oxidation of methane in the pulse mode. Pulse com-
position: 7% СН4 and 3.5% О2 in argon; contact time 15 ms
Conversion, CO selectivity, H2 selectivity,
T, °C
H2/CO
%
%
%
Pt/PrCeZrO
650
700
750
800
52.9
60.6
72.9
87.6
43.5
60.6
72.5
81.8
89.6
94.1
2.5
2.1
2.0
2.0
74.9
81.8
Pt/GdCeZrO
9.6
650
700
750
800
35.7
57.8
83.9
94.5
27.6
70.4
90.8
95.3
4.3
2.6
2.2
2.1
41.5
73.7
Steady-state catalytic activity data at atmospheric
pressure and short contact times were obtained in flow
83.5
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

2194
GUBANOVA et al.
Temperature, °C
Temperature, °C
Fig. 1. Oxygen thermodesorption curves for samples doped with (a) gadolinium and (b) praseodymium, after their dynamic
saturation with oxygen at (1) 550, (2), 600, (3) 650, and (4) 700°C.
In the lower temperature region, the methane
conversion and CO selectivity of the praseodymium-
doped sample are higher. At temperatures of about
800°C, the methane conversion and CO and H2
selectivity of both samples are close to equilibrium
values. The Н2/СО ratio at lower temperatures is
higher than the equilibrium value 2 and approaches it
as the temperature increases. This is probably
associated with the preferential oxidation of СО to
СО2, on account of the fact that СО is retained on the
catalyst surface, whereas hydrogen is readily desorbed
into the gas phase. At higher temperatures, the Н2/СО
ratio gets close to equilibrium [16].
pulses at thin coatings. At sufficiently thick coatings,
oxygen slowly desorbs, which gives rise to the so-
called “background effect:” Desorption appears as a
continuous slowly descending tail above the
background. Such a slow vacuum desorption affects
the oxidation level of the surface and prevents
complete saturation of the surface with oxygen, which
can be realized when the gas phase is equilibrium with
oxygen at a certain pressure. At the same time,
dynamic saturation of the catalyst surface with oxygen,
depending on pulse size and feeding rate. Under
equilibrium conditions the desorption rate of oxygen is
equal to the rate of its adsorption at pulse feeding,
defined as the ratio of the quantity of oxygen adsorbed
from a pulse to the time delay between two pulses.
The gadolinium-doped catalyst in steady-state
conditions is more active at lower temperatures, which
is manifested in the prevalence of complete oxidation
and lower СО and Н2 selectivity. Dynamic isotope
exchange data point to a higher mobility of oxygen in
the praseodymium-doped complex oxide lattice [21].
These results are slightly unexpected and suggest a
fairly intricate reaction mechanisms and presence of
surface oxygen species differing from each other by
bond strength, mobility, and reactivity. Obviously,
insight in these aspects can be gained by means of
nonsteady-state kinetic techniques, primarily TAP.
To assess the oxygen–surface bond strength and
oxygen mobility, we performed TPD experiments. The
corresponding TPD patterns of oxygen after dynamic
surface saturation at varied temperatures are shown in
Fig. 1.
The gadolinium-doped sample features low-tem-
perature oxygen desorption, whereas the praseo-
dymium-doped sample begins to evolve oxygen above
350°C. Thus, doping the cerium/zirconium mixed
oxide with gadolinium gives rise to more reactive
weakly bound oxygen species which should be the first
to take part in oxygen transfer to platinum over the
carrier surface. At the same time, the total quantity of
desorbed oxygen and the desorption intensity at high
temperatures are higher for praseodymium-doped
sample. This correlates with a greater mobility of
oxygen in this sample, as follows from estimates by
dynamic isotope oxygen heteroexchange [11, 21].
TAP Study of Partial Methane Oxidation
Oxygen Desorption
under Dynamic Vacuum Conditions
The implementation of TAP for catalysts with a
high oxygen capacity involves a number of problems.
Platinum catalysts strongly adsorb oxygen, which
prevents obtaining well-defined responses on oxygen
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

INFLUENCE OF THE MOBILITY OF OXYGEN
2195
Under dynamic saturation conditions, the higher
mobility of lattice oxygen during treatment provides a
deeper penetration of oxygen into fluorite-type carrier
particles, which reveals itself in its higher effective
oxygen capacity. It should be noted that the dynamic
isotope exchange technique [11, 20–22] failed to detect
a weakly bound oxygen species on the surface of the
gadolinium-doped sample. This species appears as a
strong shoulder at ~250°C on the principal peak at
~280–290°C in the spectrum of the temperature-
programmed reduction (TPR) of platinum-promoted
samples of the given series with hydrogen [21]. In the
spectrum of the praseodymium-doped sample, this
shoulder is almost lacking, and the intensity of the
principal TPR peak is several times higher than in the
case of gadolinium.
“working state” for partial methane oxidation at
atmospheric pressure. The “working state” (WS) is not
identical to the traditional steady state which takes
place on prolong contact of catalyst with a flow of the
reaction mixture of constant composition at certain
temperature and pressure.
Research into the effect of small deviations of the
surface from the WS gives information on the reaction
mechanism. The state of the catalyst surface can be
changed by exposure of the catalyst in the WS to
oxygen pulses. This experimental strategy was realized
in studying structural elements of block catalysts
containing active Pt layer on the cerium/zirconium
complex oxide doped with Pr or Gd. The WS was
attained by feeding oxygen (1 part) and methane
(2 parts) pulses at 0.5 s intervals (flashing–sample
series). The oxidation level of the catalysts was varied
by exposure to varied number of oxygen pulses. For
reactivity assessment of the catalytic surface we used
characteristics of the first 24 pulses in the flushing–
sample mode after the oxidative series. All
experiments on methane oxidation were performed at
700°C.
Both catalysts are sufficiently active and selective
in the partial oxidation of methane at temperatures
above 600°C [18–20]. Obviously, under these condi-
tions, oxygen desorption from the catalyst surface
takes place, which makes impossible continuous
coating of the surface with oxygen or its saturation. On
the other hand, it is the mobility of oxygen in the
lattice of fluorite-type oxides that is one of the most
important factors that enhances activity of applied
platinum in the partial methane oxidation to СО,
compared to inert carriers.
Praseodymium-doped Catalyst
The response of catalyst surface on methane and its
oxidation products (H2, СО, and СО2) as a function of
the deviation from the WS is shown in Fig. 2.
To develop an approach to assessing the reactivity
of a catalyst as a function of its oxidation level, let us
identify three different states of a surface.
The response of unconverted methane (Fig. 2a)
slightly decreases with increasing number of oxygen
pulses. The decreased responses of H2 and СО
(Figs. 2b and 2d) and increased response of СО2
(Fig. 2c) suggest that the СО selectivity decreases with
increasing number of oxygen pulses, whereas the
selectivity with respect to complete oxidation products
increases. It should be noted that СО forms fairly
slowly, as follows from the extended response for the
WS. A fairly strong CO response is already observed
at a “zero” time. This СО formation is associated with
conversion of tightly retained CHx fragments
remaining after the preceding methane pulse in the
“flushing–sample” sequence. Even though it is not
quite clear, whether such a too slow CO formation and,
as a result, insufficiently complete removal of CHx
fragments from the catalyst surface under conditions of
TAP experiments, means that carbon deposits are also
accumulated at atmospheric pressure; however, the
very fact that CHx fragments are present in the Pr-
doped sample in the WS casts no doubts.
(1) “Evacuated” (gas-free) surface is fairly readily
formed by vacuum treatment at high temperature,
when the maximum possible quantity of oxygen is
removed without a reducer. Such surface can simulate
the state of a catalyst in the absence of oxygen.
(2) “Oxidized” state in TAP conditions is attained
on dynamic saturation. This state can be simulated by
any frontal layer at the catalyst inlet, where the oxygen
concentration in the mixture is maximal. Probably, the
low selectivity of methane conversion into СО and Н2
at the oxidized state of the catalyst will result in strong
overheats. However, since the real reaction mixture
contains methane, and the inlet temperature can be
higher than 1000°C [23], it is obvious that the
completely oxidized state is rarely realized in practice.
(3) As the catalyst is exposed to a big series of
oxygen and methane pulses fed in succession in a 1:2
ratio, steady-state selectivity is attained. Such
treatment brings the catalyst surface closest to the
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

2196
GUBANOVA et al.
Fig. 2. Dynamics in the yields of (a) methane, (b) hydrogen, (c) СО2, and (d) СО on exposure to methane pulses of the catalyst
surface preliminarily treated with oxygen pulses. Number of oxygen pulses: (1) 5, (2) 20, (3) 50, and (4) 100. Catalyst: 1.4 wt %
Pt/Pr0.3Ce0.35Zr0.35Ox/Al2O3; 700°C; (WS) working state.
The results obtained at a great number of oxygen
pulses (50 and 100) show that a stable hydrogen yield
can be obtained even if the catalyst is highly oxidized.
It should be noted that CO is always slowly firmed
with this catalyst, which is likely to be explained by
stabilization of СНх fragments or СО molecules on the
surface due to strong binding with high-charge Pr4+
cations. The fact that, as the oxidation level of the
surface increases, СО formation is suppressed in
preference to hydrogen formation suggests that such
fragments scarcely contain hydrogen atoms, i.e. consist
of carbon atoms or СО molecules.
comparison of the shapes of the СО and СО2 responses
shows that СО is formed as a primary product, since
its signal sharply increases already during methane
pulse. By contrast, СО2 is a secondary product: It is
formed slower. Noteworthy, Н2 and СО are formed
almost concurrently, and no CO signal is observed at a
zero time. Obviously, with the gadolinium-doped
catalyst, methane transformations are not accompanied
by accumulation of carbon particles on the surface.
Comparison of these results with the oxygen TPD
curves leads us to suggest that the higher rate of
conversion of СНх fragments into СО for this sample
is associated with the presence of weakly bound
surface oxygen species. Actually, as the oxidation
level of the surface increases, the CO and H2 responses
increases after 20 oxygen pulses. It should be noted
that in this case, too, the CO and H2 responses have
“tails,” implying decomposition of oxygen-containing,
probably formate, complexes or oxygen diffusion from
the bulk of the catalyst to its surface. With the
praseodymium-doped sample, such CO and hydrogen
Gadolinium-doped Catalyst
The effect of the oxidation level of the catalyst
on its activity and selectivity is shown in Fig. 3.
Considerable responses of all the products suggest
that СО formation occurs much faster than in the case
of the praseodymium-doped sample. Moreover,
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

INFLUENCE OF THE MOBILITY OF OXYGEN
Intensity, rel. units
2197
Intensity, rel. units
WS
WS
Intensity, rel. units
Intensity, rel. units
WS
WS
Fig. 3. Responses of (a) methane, (b) hydrogen, (c) СО, and (d) СО2 to methane pulses for the “working” state of the catalyst after
a series of 20–100 oxygen pulses. Catalyst: 1.4 wt % Pt/Gd0.3Ce0.35Zr0.35Ox/Al2O3; 700°C. Number of oxygen pulses: (1) 20, (2) 50,
and (3) 100.
liberation “tails” on a partially oxidized surface were
lacking (Figs. 2b and 2d).
Effect of Delay between Oxygen
and Methane Pulses
Further increase of the number of oxygen pulses
decreases the intensity of СО and Н2 responses,
whereas methane oxidation is much accelerated. Pre-
liminary feeding of 100 oxygen pulses provides almost
complete profound oxidation of methane. It is
interesting to note that the normalized СО2 response
curves point to its much faster formation, implying that
a different reaction mechanism is possible. Probably,
this is associated with enhanced coating of platinum
with oxygen, since saturation of the fluorite-like carrier
surface with oxygen and the low mobility of oxygen in
this oxide lattice hinders oxygen spillover.
In the experiments with varied delay between
oxygen and methane pulses we studied the degree of
oxygen spillover. Obviously, spillover occurs very fast
on the “working” catalyst surface. To suppress
spillover, the catalyst surface was enriched with
oxygen to attenuate СО selectivity. With the Pr-doped
catalyst, 5 oxygen pulses proved to be sufficient for
this purpose, and with the Gd-doped system, 75 pulses
were fed.
The effects of 0.1-, 0.5-, and 1-s delays between O2
and CH4 pulses were studied for both the catalysts.
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

2198
GUBANOVA et al.
Fig. 4. Effect of pulse delay on the responses of (a) methane, (b) hydrogen, (c) СО2, and (d) СО on exposure of the oxidized
catalyst to methane pulses. Catalyst 1.4 wt % Pt/Gd0.3Ce0.35Zr0.35Ox/Al2O3; after treatment with 75 oxygen pulses at 700°C.
Figures 4 and 5 compare product selectivities at varied
pulse delays.
СО2 peaks characteristic of secondary products of
methane conversion.
As seen, the H2 selectivity of the Gd-doped catalyst
is much dependent on pulse delay and decreases with
decreasing pulse delay. The selectivity of the Pr-doped
catalyst is only slightly dependent on pulse delay.
Probably, the latter catalyst provides a more stable
partial oxidation route due to a high mobility of lattice
oxygen and accumulation of excess oxygen in bulk
complex oxide particles, which compensates for its
increased concentration in the gas phase. The
sensitivity of the Gd-doped catalyst to pulse delay is
probably explained by the occurrence on the oxide
carrier surface of short-lived weakly bound oxygen
species which, once an oxygen pulse has been fed, fast
oxidize platinum-activated methane to СО2. The
limited mobility of oxygen in the gadolinium-
containing complex oxide lattice preserves reactive
oxygen on the surface. Moreover, as the pulse delay
with the Pr-doped sample is decreased, sharp CO
peaks characteristic of primary products and diffuse
The above information shows that the catalyst
efficiency in partial oxidation of methane is deter-
mined by the relative contributions of the mobility of
lattice oxygen and oxygen spillover between platinum
and carrier. With the Gd-doped catalyst, the presence
of a moderate quantity of oxygen on the platinum
surface favors fast СО formation. Since hydrogen is
mostly fast released (except for the Gd-doped sample,
when its surface oxidation level slightly deviates from
the WS), we can suggest that hydrogen is primarily
formed due to fast dissociation of methane on the
platinum surface, which requires no oxygen to occur.
Even though one cannot a priory exclude the
possibility that methane dissociates due to reaction
with high-charge Pr4+ cations on the carrier surface,
this issue should be further explored using a platinum-
free carrier. On the other hand, hydrogen and СО are
readily oxidized when there is much oxygen on the
surface. In terms of the selectivity of methane oxida-
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

INFLUENCE OF THE MOBILITY OF OXYGEN
2199
Fig. 5. Responses of (a) methane, (b) hydrogen, (c) СО2, and (d) СО on methane pulses fed with various delay times. Catalyst 1.4 wt %
Pt/Pr0.3Ce0.35Zr0.35Ox/Al2O3 ; after treatment with 5 oxygen pulses at 700°C.
tion to CO, a mobile lattice oxygen is important for
effective transfer of excess oxygen to the bulk of
carrier particles, which decreases the oxygen coating
of the surface. An optimal catalyst should combine a
high mobility of lattice oxygen with a small number of
active surface centers that ensure fast oxidation of CHx
fragments into СО.
long (up to 600 s) relaxation times (Fig. 6). Much
stronger changes are observed with the Gd-doped
sample; therewith, the steady-state concentration of
СО2 (complete oxidation product) is established within
much a shorter time. These results provide direct
evidence for the importance of a certain intermediate
catalyst oxidation level for the formation of partial
methane oxidation products, which is consistent with
TAP data. The higher mobility of lattice oxygen in the
Pr-doped sample can be a factor that stabilizes the CO
selectivity in time.
Relaxation Experiments
Dynamic vacuum TAP data allow us to explain, to
a first approximation, the difference in catalytic
activity and selectivity between the Gd- and Pr-
promoted catalysts. To assess the steady state of the
surface at varied component pressures, we performed
relaxation experiments at atmospheric pressure.
Typical results of such experiments at 650°C and
contact time 15 ms are shown in Fig. 6 and 7.
Decreased contact time had no effect on the observed
relaxations.
Analysis of initial relaxations (Fig. 7) allows
certain conclusions as to the reaction mechanism [11]
and compare them with TAP data. Thus, the large
delay in the appearance of methane in the gas phase for
the Gd-doped sample suggests stronger retention of
methane on its oxidized surface compared to the Pr-
doped sample. The same was observed in the TAP
experiments, after the exposure of the “working”
surface to 100 oxygen pulses (Fig. 3). With the Gd-
doped sample, no СО or Н2 appear in the gas phase
within 10 s, whereas traces of СО2 are detectable. In
Of undeniable interest for understanding the
reaction mechanism is the fact that the concentrations
of hydrogen and CO pass through a maximum at fairly
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

2200
GUBANOVA et al.
Fig. 6. Dynamics in the concentrations of (a) H2, (b) СН4, (c) CO, and (d) CO2 in partial oxidation of methane for catalysts doped
with gadolinium or praseodymium. Mixture 7% СН4 + 3.5% O2 in N2; contact time 15 ms; 650°C. Catalyst: (1) Pt/PrCeZrO and (2)
Pt/GdCeZrO.
view of the TAP data, active oxidation of methane
with weakly bound oxygen to form complexes (for
example, carbonate–carboxylate) that decompose with
liberation of complete oxidation products: СО2 and
Н2О. Hydrogen and СО in the gas phase appear almost
simultaneously at contact times longer than 15 s,
which can be explained both by consumption of
weakly bound oxygen and by decomposition of surface
complexes like formate ones.
CONCLUSION
The effect of the oxidation level of the surface on
the regularities of the oxidation of methane into СО
and Н2 on platinum applied on complex cerium/
circonium fluoride-type carries depends on the nature
of the promoter cation in the carrier. In praseodymium-
doped systems, CO formed in all the conditions
studied. Hydrogen liberated shortly, and its quantity
was only slightly dependent on the oxidation level of
the sample. Moreover, the simultaneous presence of
oxygen and methane introduced in pulses with short
intervals does not result is a loss of СО and Н2
selectivity. Carbon monoxide is fairly slowly formed
on a partially reduced “working” smple surface,
probably, because СО or СНх fragments are fairly
strongly retained on the surface containing high-charge
praseodymium cations. The Gd-doped sample does not
retain СО. When the surfaces of both samples are in
In the case of the Pr-doped sample, hydrogen and
СО appear almost simultaneously with methane, and
СО2 appears after a certain delay. These results show
that even on the oxidized surface of this sample СО is
formed by the primary pyrolysis–direct methane
oxidation process, whereas complete oxidation
products, by subsequent oxidation of СО and
hydrogen. These results are nicely consistent with TAP
data.
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

INFLUENCE OF THE MOBILITY OF OXYGEN
2201
Fig. 7. Dynamics in the concentrations of (a) H2, (b) СН4, (c) CO, and (d) CO2 in the initial period. Feeding the mixture of 7% СН4 +
3.5% O2 in nitrogen on oxidized samples; contact time 15 ms; 650°C. Catalyst: (1) Pt/PrCeZrO and (2) Pt/GdCeZrO.
the “working” state, the intensities and rates of
hydrogen liberation compare with each other, implying
that hydrogen is primarily formed by methane
dissociation on platinum. When the surface of the Gd-
doped sample in the WS was slightly more oxidized,
“delayed” relaxations of hydrogen liberation are
observed, on account of decomposition of formate-
type surface complexes. In a general case, the
increased surface oxidation level for the Gd-doped
sample decreases the selectivity of partial oxidation
until СО and Н2 disappear completely.
The latter circumstance was clearly demonstrated with
the Gd-doped sample.
The results of TAP experiments are nicely
consistent with the steady-state activities and kinetic
relaxation characteristics of the samples at atmospheric
pressure and provide clear evidence for the efficiency
of the employed complex of methods for mechanistic
research on partial oxidation of methane at short
contact times.
ACKNOWLEDGMENTS
The work was performed in the framework of the
Joint Russian–French Catalysis Laboratory (LEA-313)
and supported by the PhD grant of the French Embassy
for E.A. Gubanova and by the RFBR–CNRS grant
no. 05-03-34761.
Comparing the catalytic characteristics of the
samples were should note that under TAP conditions in
dynamic vacuum the the surfaces of both samples in
the WS can have much different oxidation levels, since
a fixed oxidation level is impossible to ensure in view
of the constant oxygen desorption at the working
temperature. In terms of the reaction mechanism, one
should separate the effects of surface oxygen and of
the oxygen introduced together with methane or that
fed in pulses with short delays after methane pulses.
REFERENCES
1. Song, C., Catal. Today, 2002, vol. 77, p. 17.
2. Au, C.T. and Wang, H.Y., J. Catal., 1997, vol. 167,
p. 337.
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008

2202
GUBANOVA et al.
14. Soick, M., Buyevskaya, O., Höhenberger, M., et al.,
3. Rostrup-Nielsen, J.R., Catal. Today, 1993, vol. 18, p. 125.
Ibid., 1996, vol. 32, p. 163.
4. Fox, J.M. III., Catal. Rev. Sci. Eng., 1993, vol. 35, p. 169.
15. Mirodatos, C., Schuurman, Y., van Veen, A.C., et al.,
Stud. Surf. Sci. Catal., 2007, vol. 167, p. 287.
16. Pavlova, S., Sazonova, N., Sadykov, V., et al., Catal.
Today, 2005, vol. 105, p. 367.
5. Horn, R., Williams, K.A., Degenstein, N.J., et al., J. Catal.,
2007, vol. 249, p. 380.
6. Horn, R., Williams, K.A., Degenstein, N.J., et al.,
Chem. Eng. Sci., 2007, vol. 62, p. 1298.
17. Beretta, A. and Forzatti, P., Chem. Eng. J., 2004, vol. 99,
7. Horn, R., Williams, K.A., Degenstein, N.J., et al., J. Catal.,
2006, vol. 242, p. 92.
p. 219.
18. Sadykov, V.A., Frolova-Borchert, Yu.V., Mezentse-
va, N.V., et al., Mater. Res. Soc. Symp. Proc., 2006,
vol. 900E, p. O10.
8. Schmidt, L.D., Klein, E.J., Leclerc, C.A., et al., Chem.
Eng. Sci., 2003, vol. 58, p. 1037.
19. Kuznetsova, T.G., Sadykov, V.A., Moroz, E.V., et al.,
9. Hickman, D.A. and Schmidt, L.D., J. Catal., 1992, vol. 138,
p. 267.
Stud. Surf. Sci. Catal., 2002, vol. 143, p. 659.
20. Sadykov, V.A., Kuznetsova, T.G., Alikina, G.M., et al.,
New Topics in Catalysis Research, David, K., Ed., New
York: Nova, 2006, p. 97.
21. Sadykov, V.A., Mezentseva, N.V., Alikina, G.M., et al.,
Solid State Phenom., 2007, vol. 128, p. 239.
22. Sadykov, V., Mezentseva, N., Alikina, G., et al., Mater.
Res. Soc. Symp. Proc., 2007, vol. 988, p. QQ06-04.
23. Vernikovskaya, N.V., Bobrova, L.N., Pinaeva, L.G., et
al., Chem. Eng. J., 2007, vol. 134, p. 180.
10. Goralski, C.T. Jr.1, O’Connor, R.P., and Schmidt, L.D.,
Chem. Eng. Sci., 2000, vol. 55, p. 1357.
11. Sadykov, V.A., Kuznetsova, T.G., Frolova-Bor-
chert, Yu.V., et al., Catal. Today, 2006, vol. 117, p. 475.
12. Pavlova, S.N., Sazonova, N.N., Ivanova, J.A., et al.,
Ibid., 2004, vols. 91–92, p. 299.
13. Wolf, D., Barré-Chassonnery, M., Höhenberger, M., et
al., Ibid., 1998, vol. 40, p. 147.
RUSSIAN JOURNAL OF GENERAL CHEMISTRY Vol. 78 No. 11 2008