Production of synthesis gas by direct catalytic oxidation of methane on Pt{110} (1 × 2) using supersonic molecular beams

Article abstract of DOI:10.1021/jp0004112

The absolute rate of the methane oxidation reaction over Pt{110} was investigated over the temperature range 400-900 K using molecular beams under ultrahigh vacuum (UHV) conditions. It was found that the surface reaction is biphasic, with CO₂ b

Full text of DOI:10.1021/jp0004112

6462
J. Phys. Chem. B 2000, 104, 6462-6467
Production of Synthesis Gas by Direct Catalytic Oxidation of Methane on Pt{110} (1 × 2)
Using Supersonic Molecular Beams
A. V. Walker and D. A. King*
Department of Chemistry, UniVersity of Cambridge, Lensfield Road, Cambridge, United Kingdom CB2 1EW
ReceiVed: February 2, 2000; In Final Form: April 14, 2000
The absolute rate of the methane oxidation reaction over Pt{110} was investigated over the temperature
range 400-900 K using molecular beams under ultrahigh vacuum (UHV) conditions. It was found that the
surface reaction is biphasic, with CO2 being the major product at <550 K and CO the major product at >600
K. Under our reaction conditions, the reaction products H2, CO, and CO2 were detected. The H2O production
was below the limit of detectability. The product selectivity to CO and CO2 can be controlled by varying the
beam composition as well as the surface temperature. A mean field coupled differential equation model of
the process provides a reasonable description of all the experimental observations, although it does not include
a proper description of prompt CO desorption at high surface temperatures.
Introduction
proceeds by conversion to the full oxidation products, H2O and
CO2, followed by steam reforming, WGS, and CO2 reforming
reactions to the partial oxidation products, CO and H2. Schmidt
et al.5,6 observed an almost complete conversion of CH4 to
syngas with a CO selectivity >90% using Pt- and Rh-coated
monoliths at short contact times (10-3 s) and high spatial
velocities (106 h-1). They proposed a second mechanism, a
pyrolysis reaction, in which CO and H2 form first and are
subsequently oxidized to H2O and CO2.
In recent years, there has been increased interest in the
conversion of methane to synthesis gas (syngas, a mixture of
CO and H2, ideally in a 1:2 ratio). Many industrial processes
use syngas as a feedstock; for examples, the Haber process (NH3
production),1 methanol synthesis,2 and the Fischer-Tropsch
process3 (reaction to higher hydrocarbons).
Synthesis gas is usually produced by steam reforming:
CH4 + H2O f CO + 3H2
∆Hr ) 206.1 kJ mol-1 (1)
To elucidate the overall reaction mechanism of the partial
oxidation of methane, we conducted a systematic series of
studies on Pt{110} (1 × 2) prior to the present investigation
that we summarize here. The dissociative sticking probability
of O2 on Pt{110} (1 × 2) is ∼0.47 at a surface and thermal
beam temperature of 300 K, which is much larger probability
than that for CH4, which is only ∼10-6 8 at the same temper-
ature. Thus, the rate-determining step is believed to be the
dissociative adsorption of methane.
To drive this endothermic process, the reactor is heated either
externally or internally. In the latter case, heating is achieved
by adding O2 to the feed to provide the energy via highly
exothermic combustion reactions. A typical industrial steam
reformer operates at 1120-1170 K and 15-30 atm over a Ni/
Al2O3 catalyst. The CO/H2 is further adjusted by the water gas
shift (WGS) reaction:
On Pt{110} (1 × 2),8 we observed that as the incident
translational energy, Et, of the incident methane beam was
increased, the initial dissociative sticking probability, s0, passes
through a minimum at ∼100 meV. At Et > 100 meV, an acti-
vated adsorption process becomes dominant, with an activation
barrier of ∼130 meV. At translational energies >230 meV, s0
attains a limiting value, which is strongly enhanced by the ex-
citation of the CsH stretch modes. We also observed that an
increase in surface temperature enhances s0. Two other reaction
steps were studied: the interaction of methane with preadsorbed
oxygen adatoms, Oa,9 and the oxidation of adsorbed carbon.10
On all surfaces studied to date, Ni{100},11 Ni{111},12,13 Pt-
{111},14 and Pt{110} (1 × 2),9 the initial (zero coverage)
dissociative sticking probability of methane, s0, falls with
increasing oxygen coverage. We also observed that CO gas was
evolved during exposure of the O-precovered surface to the
incident CH4 beam. As the surface temperature is raised to >550
K in the carbon oxidation reaction, there is a switch in selectivity
from CO2 to CO production;10 the proportion of CO2 decreases
from 95% at Ts ) 450 K to <1% at Ts ) 750 K. Using angle-
resolved product distribution measurements, we also observed
that CO desorbs in a sharp lobe, with an excess translational
CO + H2O T CO2 + H2
∆Hr ) -43.2 kJ mol-1 (2)
Typical WGS reactors operate either at ∼470 K using a Cu-
based catalyst or at ∼670 K using an iron oxide/chromia
catalyst. Hence, the conversion of methane to syngas is both
expensive and complicated.
An alternative reaction is the direct catalytic partial oxidation
of methane to syngas:
1
2
CH4 + O2 f CO + 2H2
∆Hr ) -35.7 kJ mol-1 (3)
This exothermic reaction could not only produce syngas with
the ideal ratio for CH3OH production (if a good catalyst can be
found for this process, rather than the steam reforming process
using CO2) or the Fischer-Tropsch process, but it could also
lead to the use of smaller and simpler reactors. However, this
direct oxidation process is not without its technical difficulties
because the reactions to form the secondary oxidation products,
H2O and CO2, are extremely fast.
Two main proposed reaction mechanisms have been reported
in the literature. Prettre et al.4 postulated that the reaction
energy estimated at 135 ( 5 kJ mol-1
.
10.1021/jp0004112 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/17/2000

Synthesis Gas Production
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6463
Experimental Section
The experiments were carried out in an ultrahigh vacuum
(UHV) chamber15 equipped with a differentially pumped rotating
quadrupole mass spectrometer (QMS), and a stagnation tube
and spinning rotor gauge, to measure the beam flux in the
molecular beam scattering plane. In an upper level, there is a
second QMS, an Ar+ sputter gun, and low-energy electron
diffraction (LEED) optics.
The sample was a clean single crystal Pt{110} surface
orientated to within 1° of the plane and mechanically polished.
Initially the surface was cleaned using sputter-anneal cycles and
oxygen treatments. Afterward, the surface was cleaned by
annealing until a temperature of 1240 K was reached, treating
with oxygen while cooling from 1100 to 950 K, and annealing
again at 950 K. This procedure yields a clean Pt{110}(1 × 2)
surface that exhibits excellent LEED patterns and reproducible
temperature-programmed desorption (TPD) spectra for CO16 and
O2.7,17 The sample temperature was monitored using a K-type
thermocouple and controlled by a Eurotherm 904P program-
mable temperature controller.
The CH4 and O2 used were >99.9995% and 99.995% pure,
respectively, as quoted by the suppliers (Messer UK Ltd.). To
increase the translational energy of the CH4/O2 mixed beam
without increasing the internal energy, it was seeded in He. The
translational and internal energy of the CH4/O2 mixed beam
was also increased by increasing the nozzle temperature. The
highest nozzle temperature used was 850 K; we have evidence
for methane cracking above 900 K. The translational energy
was determined using time-of-flight techniques.
Figure 1. The variation of the rates of production of CO and CO2
with time at a surface temperature of 900 K. The methane-to-oxygen
ratio is 2:1 (effective pCH4 ) 2 × 10-8 mbar). To aid viewing, the CO
and CO2 traces have been vertically displaced. The dotted lines represent
the baseline pressures: we note that these are based on more extended
ranges of background levels, both before and after beaming, than those
shown here.
s0 for oxygen decreases under the same experimental condi-
tions.7 Thus, at a translational energy of 500 meV and nozzle
temperature of 800 K, the methane initial sticking probability
is ∼0.01, whereas the oxygen s0 decreases to 0.23. Under these
reaction conditions, if the CH4 partial pressure at the sample is
10-9 mbar and if all the Ca produced from the dissociative
adsorption of methane reacts to form CO and CO2, it is expected
that the total partial pressure rise for the products would be 10-11
mbar, which can be easily detected.
The efficiency of formation of each of the reaction products
is quoted in terms of a reaction probability, ê () product flux/
CH4 flux), which gives the probability that a CH4 molecule
incident on the surface reacts to form that product. The
selectivity of the reaction products is also quoted for CO [sCO
) p(CO)/[p(CO) + p(CO2)]].
Variation of the Steady-State Products with Surface
Temperature and Beam Composition. As the CH4/O2/He
beam impinges on the clean surface maintained at 900 K (Figure
1), the rate of CO production rises rapidly, followed by a slower
rise to a maximum before decreasing to a steady-state value.
Thus, the steady-state CO production level is reached after ∼10
s, whereas the instantaneous level is ∼5% higher. Although the
noise level on the signal is quite high ((2.5%), the initial higher
increase was reproducibly observed. There is a much slower
rise in the CO2 production rate. Because the CO2 production
rate at a crystal temperature of 900 K is very low, the signal-
to-noise ratio at steady state is small (∼2). Nevertheless, an
abrupt fall in CO2 pressure is observed when the beam is
terminated (at 193 s in Figure 1), but no abrupt rise is observed
when the beam is allowed to impinge on the crystal (at 10 s in
Figure 1). Quite clearly, the rate of CO2 formation rises much
more slowly than that for CO.
Isothermal data such as that shown in Figure 1 were obtained
at surface temperatures ranging from 900 down to 400 K. (A
further example of the data obtained will be presented in Figure
6b.) Steady-state rates of CO and CO2 production were always
achieved within 50 s of allowing the beam to impinge on the
crystal. Two factors change as the substrate temperature is
lowered. First, the rise in CO production below ∼600 K is no
longer instantaneous, and much less CO is produced. Second,
the CO2 production rate at steady state is significantly increased
All the reaction products were detected using the fixed QMS.
To increase the data collection rate without loss of signal-to-
noise, sequential experiments were performed; in the first
experiment, only the reaction products CO and CO2 were
followed, and in the second, under the same conditions, the
reaction products H2 and H2O were recorded. The O2 and CH4
beam fluxes were calibrated using a stagnation tube and spinning
rotor gauge to quantify the partial pressure rise measured by
the QMS. The absolute rate of CO and H2 produced was also
calibrated in a similar way using CO and H2 beams. To calibrate
the CO2 produced, CO was titrated with a known amount of Oa
on the surface and the amount of CO2 produced was measured.
All coverages are quoted as fractional monolayers relative to
the number density in the Pt{110} (1 × 1) surface, 9.2 × 1014
Pt atoms cm-2
.
After the reaction, the sample was oxygen treated at 650 K
and then heated to 1200 K at 3 K s-1 to remove any residual
Ca present on the surface.
Results and Discussion
Previous studies of CH4 partial oxidation5,6,18 have demon-
strated that a number of reaction parameters are important,
including the methane-to-oxygen feed ratio and the surface
temperature. In the present study, to gain an overall view of
the reaction profile, measurements were made of the variation
of the reaction products with surface temperature and beam
composition. These reactions were then modeled to gain further
insights into the reaction mechanism.
At a surface temperature, Ts, of 400 K and a beam energy of
95 meV, the initial (zero coverage) dissociative sticking
probability, s0, of O2 is 0.34,7 which is 5 orders of magnitude
larger than that for CH4 (s0 ) 3.4 × 10-6 8) under the same
reaction conditions. However, we note that the value of s0 for
methane increases sharply with increasing surface temperature,
incident translational energy, and vibrational energy,8 whereas

6464 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Walker and King
Figure 4. Selectivity to CO formation from a CH4 + O2 beam at
normal incidence on clean Pt{110}(1 × 2) at various beam compositions
at a surface temperature of 700 K.
Figure 2. Reaction probability for CO, CO2, and H2O formation from
a CH4 + O2 beam at normal incidence on clean Pt{110}(1 × 2) at
various surface temperatures. The methane-to-oxygen ratio is 2:1 (the
stoichiometric ratio for CO production).
O2.10 Hence, we conclude that the same reaction is dominant.
It was demonstrated that CO formed in the surface reaction at
high temperatures desorbs with excess energy of 135 ( 5 kJ
mol-1; a detailed discussion of the mechanism is given
elsewhere.10 There is a sharp rise in CO selectivity at ∼700 K.
At low coverages, O2 desorbs at ∼700 K, and this loss of surface
oxygen adatoms is associated with the fall in CO2 production
at these temperatures. We note that at a surface temperature of
400 K, the selectivity to CO (0.05) is higher than that observed
for the oxidation of preadsorbed C [s(CO) ) 0.02].14 Very
recently, Watson et al.21 have shown that at temperatures below
∼500 K, the stable product of methane dissociation is methyne
(CH). The reaction with O2 at these temperatures proceeds
through a new pathway involving CH, and not adsorbed C. The
rate of the reaction of O2 with C at 400 K is negligible compared
with this new reaction with CH.20
Figure 4 shows the selectivity to CO for various beam
compositions at a surface temperature of 700 K. As the CH4-
to-O2 ratio changes from 1:2 to 10:1, the CO selectivity increases
somewhat from ∼79% at a ratio of 1:5 to 88% at a ratio of
10:1. These observations are also reminiscent of the variation
of CO selectivity with incident O2 beam flux measured for Ca
oxidation.10
Temporal Evolution of the Products. The formation rate
of CO was always observed to rise very rapidly, before the CO2
production rate (Figure 1), and we conclude from the temporal
evolution of CO and CO2 that the reaction mechanism is similar
to that observed for Ca oxidation.10 At surface temperatures
below ∼650 K, the rate of CO2 formation also passes through
a maximum, whereas at higher temperatures, the CO2 formation
rate rises slowly. Oxygen slows down the dissociative adsorption
of methane;9 as the reaction progresses, the methane dissociative
sticking probability, and hence the rate of Ca formation,
decreases to a steady-state value. Thus, the rates of CO and
CO2 formation are observed to pass through maxima. At higher
surface temperatures, the net dissociative sticking probability
is lower and O2 desorption occurs, producing a lower Oa
coverage. The dissociative sticking probability of methane
therefore remains higher and the maximum in the rate of CO
and CO2 production is less pronounced (Figure 1).
Figure 3. Selectivity to CO formation from a CH4 + O2 beam at
normal incidence on clean Pt{110}(1 × 2) at various surface temper-
atures. The methane-to-oxygen ratio is 2:1 (the stoichiometric ratio).
and, as a function of time after switching on the reactants, it
shows a pronounced maximum before decaying to the steady-
state rate of CO2 production. Figure 2 displays ê for the
formation of the products CO, CO2, and H2 at steady state for
a methane-to-oxygen ratio of 2:1 (the stoichiometric ratio for
the partial reaction to CO) at various surface temperatures, Ts.
Under the reaction conditions used, no water formation was
detected at any surface temperature (or beam composition)
measured, which is surprising because the H2O formation
reaction from H and O adatoms is fast at room temperature and
above.19 However, after the work described in the present paper
Watson, van Dijk, and King20 were able to study the reaction
at incident methane fluxes ∼10 times higher than those
described in the present work. In this way, H2O production was
observed, with the selectivity being dependent on both the
surface temperature and the CH4:O2 ratio in the beam. We
therefore conclude that the inability to detect H2O production
in the present work arose from the relative insensitivity to H2O
detection.
Mathematical Model
As the surface temperature is raised from 400 to 900 K, the
selectivity switches from CO2 to CO (Figure 3). At 400 K, the
CO selectivity is 4%, but at 900 K it is 95%. This behavior is
similar to that observed for the reaction of pre-adsorbed C and
To obtain further insights into the experimental observations
and to test the hypothesis that the dominant mechanism to form
CO and CO2 is Ca oxidation, the reaction was modeled using a

Synthesis Gas Production
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6465
TABLE 1: Reaction Parameters Used in the Model
mean-field approximation based on the following elementary
processes:
rate
reaction
(s-1
)
Ea (kJ mol-1
)
ν (s-1
)
ref
k1
CH4 impingement rate k1 4.155 × 105 ML mbar s-1
calc.
calc.
10
CH4,g
9
k2
9
8 Ca + 4Ha
(1)
(2)
(3)
(4)
(5)
(6)
(7)
O2 impingement rate
Ca + Oa f COa
k2 2.938 × 105 ML mbar s-1
k3
0
24
2 × 109 25, 26
2 × 1016 16, 26
1 × 1013 23, 24
3 × 1011 26
O2,g
8 2Oa
COa + Oa f CO2 (g)
k4 58.6
k5 158.8
k6 54.4
COa f CO (g)
2Ha + Oa f H2O (g)
2Ha f H2 (g)
k3
Ca + Oa
98 COa
k7 E (θ ) 0) ) 54.8
R ) 10.5 kJ mol-1 mL-1
k4
COa + Oa
98 CO2,g
k5
9
COa
8 COg
k6
2Ha + Oa
98 H2Og
k1
9
2Ha
8 H2,g
where the subscripts a and g denote an adsorbed surface and a
gaseous species, respectively. Equations 1 and 2 represent the
dissociative adsorption of methane and oxygen. The dissociative
adsorption of methane to form carbon and hydrogen adatoms
is treated as a direct activated process.8 The presence of the
reaction products on the surface are only considered for H2 and
CO because CO2 and H2O have a very short lifetime in this
temperature range. We note that this simple Langmuir-
Hinshelwood formulation, in which all surface species and
gaseous products are assumed to be equilibrated to the surface
temperature, does not describe the reaction to form hot CO via
a [C-O-O]* intermediate.10 However, hot CO is only formed
at the highest surface temperatures used. At intermediate tem-
peratures, equilibrated CO is formed, and this is appropriately
described by the Langmuir-Hinselwood formulation. Similar
considerations and reservations apply to our mean-field model-
ing of the reaction of O2 with preadsorbed carbon.10 Nevertheless
the formalism provides a relatively simple way forward.
All reactions take place on the (1 × 2) phase because the
adsorption of oxygen has been shown not to induce the lifting
of the (1 × 2) surface reconstruction to the (1 × 1) phase;7 the
CO coverage on the surface under the conditions described is
always extremely small, ,0.01 mL, and is therefore too low to
induce the lifting of the (1 × 2) reconstruction.16 Even at
temperatures below the desorption temperature of CO, a low
coverage is maintained because of the reaction with adsorbed
oxygen to form CO2.
Figure 5. Selectivity to CO formation from the model simulations
for the reaction of CH4 + O2 in a 2:1 ratio (effective pCH4 ) 2 × 10-8
mbar) at various surface temperatures. The experimental observations
are also shown for comparison.
dissociative sticking probability, sCH with oxygen adatom
4
coverage has been measured under the same experimental
conditions.9 The sticking probability also varies with carbon
coverage,8 but this variation is ignored because the carbon
coverage is expected to remain very low. The rate constant k3
used is taken from ref 10.
Equation B describes the variation in oxygen coverage due
to adsorption (k2) and the reactions to form CO (k3), CO2 (k4),
and water (k5). We note that the sticking probability sO is a net
2
sticking probability. The same rate constants, k4 and k5, are used
as in earlier modeling of NO + H2,22,23 CO + O2,24 and NO +
CO.25 Similarly, eq C describes the variation of CO coverage
due to formation (k3) and removal by either reaction to CO2
(k4) or desorption (k5). During the experiment, θCO is expected
to remain extremely low and the desorption energy was therefore
assumed to remain constant.
A set of four coupled differential equations is used to describe
the variations in the adsorbate coverages (θC, θO, θCO, and θH)
on the surface. All coverages are quoted relative to the Pt atom
density in the ideal (1 × 1) surface. Sticking probabilities for
The variation of θH is described by eq D. The processes
involved are methane dissociation adsorption (k1), reaction to
H2O (k6), and desorption (k7). The data were fitted with a
desorption energy of the form Ed ) Ed (θ ) 0 mL) - Rθ, where
R is an interaction energy between neighboring H adatoms.
The differential equations were integrated numerically using
the fourth-order Runge Kutta formula27 with time steps between
5 × 10-7 and 1 × 10-4 s. The parameters used are listed in
Table 1.
Variation of the Product Distribution with Surface Tem-
perature. Without further fitting or alteration of any of the
literature parameters, the model was found to have a very high
selectivity to H2O, which increased from 69% at 400 K to 77%
at 900 K. Under our present low CH4 incident beam flux
conditions, we were unable to detect the H2O formed, and this
prediction could not be confirmed. Excellent agreement with
the observed experimental CO selectivities is, however, ob-
tained for surface temperatures between 400 and 900 K (Figure
5).
each gas are denoted by sgas
:
dθ
C ) k1sCH PCH - k3θCθO
(A)
4
4
dt
dθ
dt
O ) k2sO pO - k3θCθO - k4θCOθO - k6θHθO (B)
2
2
dθ
CO ) k3θCθO - k4θCOθO - k5θCO
(C)
(D)
dt
dθ
H ) 4k1sCH pCH - 2k6θHθO - k7(θH)2
4
4
dt
Variations in θc are described in eq A, which contains adsorption
(k1) and reaction to form COa (k3). The variation of the

6466 J. Phys. Chem. B, Vol. 104, No. 27, 2000
Walker and King
the competition between CO desorption and reaction to CO2 as
a function of surface temperature and surface oxygen coverage.
The proposed mechanism is a direct oxidation process and
has some similarities to the mechanism proposed by Schmidt
and co-workers.5,6 It can be broken down into a series of steps.
Steps 1 and 2 describe methane and oxygen dissociative
adsorption.
k1
CH4,g
9
k2
9
8 Ca + 4Ha
(1)
(2)
O2,g
8 2Oa
At surface temperatures above ∼550 K, methyl, formed from
the dissociative adsorption of methane, rapidly dehydrogenates
to Ca and Ha.28,29 Thus, all reactions occur on the surface
between carbon, oxygen, and hydrogen adatoms.
The experimental observations indicate that the oxygen
adatom coverage is important. As θO increases, there is a
decrease in the dissociative sticking probability of methane13
and hence in the rate of C production.
The CO formation rate rises almost instantaneously as the
mixed CH4 + O2 beam impinges on the surface, followed by a
slower rise. After a measurable delay, the rate of CO2 production
increases. Thus, the dominant reaction mechanism for the
production of CO and CO2 is similar to that observed for Ca
oxidation;18 we conclude that a significant proportion of CO
desorbs from the surface translationally and vibrationally hot
at high surface temperatures.
As the surface temperature is increased, the selectivity to CO
and the reaction probability, ê (CO), increases. The increase in
ê reflects an increase in the dissociative sticking probability of
methane with surface temperature, Ts, whereas the concomitant
decrease in ê (CO2) is associated with a decrease in steady-
state Oa coverage with increasing Ts and the increasing propen-
sity for prompt (hot) CO to form. At a surface temperature of
900 K and a methane-to-oxygen ratio of 2:1 (the stoichiometric
ratio for CO production), the CO selectivity is 95%.
Figure 6. The variation of simulated rates of production of CO and
CO2 with time at a surface temperature of 500 K. The methane-to-
oxygen ratio is 2:1 (effective pCH ) 2 × 10-8 mbar): (a) model; (b)
4
experimental results. The dotted lines indicate the baseline pressures.
Temporal Evolution of the Reaction Products. Using the
model, we have computed the temporal evolution of the
products. In qualitative agreement with experiment, as soon as
the mixed CH4/O2/He beam impinges upon the surface (t ) 0
s), there is an increase in the rate of production of CO that rises
to a maximum before decreasing to a steady-state value (Figure
6). However, the increase in production rate is not as fast as
that observed experimentally. This initial increase is followed
by a slower rise in the CO2 production rate. At surface
temperatures <650 K, the computed CO2 production rate goes
through a maximum. The model does predict maxima at longer
times than is experimentally observed, particularly for CO2
production. However, the experimentally observed trends are
reproduced. Also in agreement with experiment, at temperatures
<650 K, there is a slow build up of the CO2 production rate.
At all surface temperatures measured, the computed C coverage
passes through a maximum 10-15 s before sharply dropping
to a steady-state value as the oxygen adatom coverage increases;
as previously noted, oxygen poisons the surface to the dissocia-
tive adsorption of methane.13 The maximum in the carbon
coverage approximately coincides with the maximum in the CO
production rate. Hence, during the first part of the reaction, the
CO formation rate is controlled by the amount of carbon present
on the surface. The carbon coverage at the temporal peak
decreases with increasing Ts, which in turn corresponds to a
less pronounced CO production rate peak.
Summary
In using a molecular beam gas source for a mechanistic study
of catalytic reactions, where a high pumping speed is maintained
in the reaction chamber, the reaction is uniquely examined under
the condition of single collisions of each reacting molecule with
the surface and effectively no collisions of product molecules
with the surface. In the present work we have demonstrated,
for the first time, that under these conditions, the continuous
reaction between CH4 and O2 to the products CO, CO2, and H2
can be studied. The results clearly show no support for the
reaction mechanism suggested many years ago by Prettre et al.,4
invoking complete oxidation to gaseous CO2 and H2O in the
first step and subsequent formation of CO and H2 by reactive
collisions of the initial gaseous products with the surface.
Similarly, our results show the biphasic nature of the surface
reaction, without invoking gaseous CO production in a first step,
followed by CO2 formation when CO is readsorbed. Instead,
we have shown that the product selectivity to gaseous CO or
CO2 is entirely determined by surface processes, and is very
strongly dependent on surface temperature, and to a lesser extent
on the CH4:O2 ratio.
Under the conditions of the reaction, the modeling shows that
the coverage in oxygen adatoms varies over the range 0.2-0.1
ML in the temperature range 500-850 K. Over the same
temperature range, the carbon adatom coverage varies between
0.007 and 0.0045 ML.
Reaction Mechanism
The results are largely consistent with our earlier study10 of
the interaction of gaseous O2 with preadsorbed carbon, formed
by the decomposition of methane at surface temperatures <500
K. At the highest surface temperatures (g650 K), gaseous CO
is formed via a prompt process assigned to the explosive
Product selectivity in the methane oxidation reaction over
Pt{110}(1 × 2) is controlled by the parameters of surface
temperature and surface oxygen coverage. The experimental
observations can be explained within a model that considers

Synthesis Gas Production
J. Phys. Chem. B, Vol. 104, No. 27, 2000 6467
References and Notes
decomposition of a surface intermediate formed between
chemisorbed molecular O2 and adsorbed C: this hot CO, with
an equivalent temperature of ∼11,000 K,10 would need to be
accounted for in the modeling of the reaction in a standard
catalytic reactor. As the surface temperature is lowered, two
new factors come into play. First, the hot CO released onto the
potential energy surface for chemisorbed CO by decomposition
of the O2C reaction intermediate exchanges energy with the
surface, as the surface temperature is lowered to <650 K, and
becomes thermally accommodated to the surface: it can
subsequently desorb as CO or react with adsorbed O to form
CO2. Second, the surface reaction between O and C adatoms
becomes more effective in the production of adsorbed, thermally
equilibrated CO; as the temperature is lowered, the CO lifetime
on the surface increases, the reaction with adsorbed O takes
over, and the dominant product becomes CO2.
One important difference with our previous study of O2
interacting with preadsorbed C is noted. At temperatures
between 350 and 400 K, the stable surface species formed from
methane dissociation is CH.21 Watson et al.20 have shown that
this CH species, at low surface temperatures, is more reactive
to coadsorbed Oa than C. This alternative reaction pathway has
not been accounted for in the present work.
(1) Stacey, M. H. Catalysis (London) 1990, 3, 1998.
(2) Kang, H. H. Catal. ReV. Sci. Eng. 1981, 22, 235.
(3) Fischer, F.; Tropsch, H. Brennst.-Chem. 1926, 7, 197; Chem. Ber.
1926, 59, 830.
(4) Prettre, M.; Eichner, C.; Perri, M. Trans. Faraday Soc. 1946, 43,
335.
(5) Tornianen, P. M.; Chu, X.; Schmidt, L. D. J. Catal. 1994, 146, 1.
(6) Hickman, D. A.; Schmidt, L. D. ACS Symposium Series 1993, 523,
416; Aiche 1993, 39, 1164; Science 1993, 259, 343.
(7) Walker, A. V.; Klo¨tzer, B.; King, D. A. J. Chem. Phys. 1998, 109,
6879.
(8) Walker, A. V.; King, D. A. Phys. ReV. Lett. 1999, 82, 5156.
(9) Walker, A. V.; King, D. A. Surf. Sci. 2000, 444, 1.
(10) Walker, A. V.; King, D. A. J. Chem. Phys. 2000, 112, 1937.
(11) Quinlan, M. A.; Wood, B. J.; Wise, H. Chem. Phys. Lett. 1985,
118, 478.
(12) Krishnan, G.; Wise, H. Appl. Surf. Sci. 1989, 37, 244.
(13) Alstrup, I.; Chorkendorff, I.; Ullmann, S. Surf. Sci. 1990, 234,
79.
(14) Valden, M.; Xiang, N.; Pere, J.; Pessa, M. Appl. Surf. Sci. 1996,
99, 83.
(15) Hopkinson, A.; Guo, X.-C.; Bradley, J. M. J. Chem. Phys. 1993,
99, 8262.
(16) Hoffmann, P.; Bare, S. R.; King, D. A. Surf. Sci. 1982, 117,
245.
(17) Wilf, M.; Dawson, P. T. Surf. Sci. 1977, 65, 399.
(18) Lee, J. H.; Trimm, D. L. Fuel Proc. Technol. 1995, 42, 339.
(19) Hellsing, B.; Kasemo, B.; Zhdanov, V. P. J. Catal. 1991, 132,
210.
(20) Watson, D. T. P.; van Dijk, J.; Harris, J.; King, D. A., to be
published.
(21) Watson, D. T. P.; Titmuss, S.; King, D. A., submitted for publica-
tion.
(22) Lombardo, J.; Fink, T.; Imbihl, R. J. Chem. Phys. 1993, 98, 5526.
(23) Gruyters, M.; Pasteur, A. T.; King, D. A. J. Chem. Soc., Faraday
Trans. 1996, 92, 2941.
(24) Gruyters, M.; Ali, T.; King, D. A. J. Phys. Chem. 1996, 100, 14417.
(25) Hopkinson, A.; King, D. A. Chem. Phys. 1993, 177, 433.
(26) McCabe, R. W.; Schmidt, L. D. Surf. Sci. 1976, 60, 85.
(27) Numerical Recipes in C: the Art of Scientific Computing, 2nd Ed.;
Press, W. H.; Teukolsky, S. A.; Vetterling, W. T.; Flannery, B. P., Eds.;
Cambridge University Press: New York, 1992; p 710.
(28) Oakes, D. J.; Newell, H. E.; Rutten, F. J. M.; McCoustra, M. R.
S.; Chesters, M. A. J. Vac. Sci. Technol. A 1996, 14, 1439.
(29) Zaera, F.; Hoffmann, H. J. Phys. Chem. 1991, 95, 6397.
We have formulated a Langmuir-Hinshelwood mechanism,
using a mean-field approach, to model the reaction. Although
the model provides a good description of the steady-state
selectivity of the reaction to CO or CO2 and the overall reaction
probability and a reasonably good description of the temporal
evolution of the products, it has its limitations. In particular,
the model does not account for the production of prompt CO,
nor does it account, in its present form, for the presence of CH
on the surface.
Acknowledgment. We acknowledge EPSRC for an equip-
ment grant, Shell Research and Technology Centre, Amsterdam,
for a studentship (A.V.W), and support from and discussions
with Professor Gert Jan Kramer.