Mechanism and Site Requirements for Activation and Chemical Conversion of Methane on Supported Pt Clusters and Turnover Rate Comparisons among Noble Metals

Article abstract of DOI:10.1021/jp036985z

The mechanism and site requirements for activation and chemical conversion of methane on supported Pt clusters and turnover rate comparisons among noble metals were presented. Isotopic trace and kinetic measurements led to a simple mechanistic picture and a unifying kinetic treatment of CH₄-CO₂, CH₄-H₂O, and CH₄ decomposition reactions, as well as water-gas shift, on Pt-based catalysts. Reforming and decomposition rates were first-order on CH₄ concentration and independent of the concentration or identity of the co-reactants. The normal CH₄/CD₄ kinetic isotope effects measured were similar for all three CH₄ reactions and thus also independent of co-reactant identity. Forward CH₄ turnover rates increased monotonically with increasing Pt dispersion for CO₂ reforming, H₂O reforming, and CH₄ decomposition reactions. The rates of structure-insensitive CO oxidation reactions were similar before and after CH₄ reforming.

Full text of DOI:10.1021/jp036985z

4094
J. Phys. Chem. B 2004, 108, 4094-4103
Mechanism and Site Requirements for Activation and Chemical Conversion of Methane on
Supported Pt Clusters and Turnover Rate Comparisons among Noble Metals
Junmei Wei and Enrique Iglesia*
Department of Chemical Engineering, UniVersity of California at Berkeley, Berkeley, California 94720
ReceiVed: October 5, 2003; In Final Form: January 21, 2004
Isotopic tracer and kinetic studies are used to probe the identity and reversibility of elementary steps required
for H
mechanistic equivalence for CO
Reforming rates are exclusively limited by C-H bond activation on essentially uncovered Pt crystallite surfaces
and unaffected by the concentration or reactivity of co-reactants (H O, CO ). Kinetic isotopic effects are
consistent with the sole kinetic relevance of C-H bond activation (k /k ) 1.58-1.77 at 873 K); these
isotope effects and measured activation energies are similar for H O reforming, CO reforming, and CH
decomposition reactions. CH /CD cross exchange rates are much smaller than the rate of methane chemical
conversion in CO and H O reforming reactions; thus, C-H bond activation steps are irreversible, except as
required by the approach to equilibrium for the overall reforming reaction. Reactions of CH
mixtures led to identical 13C fractions in CO and CO
, indicating that CO activation is quasi-equilibrated
and kinetically irrelevant. Binomial water and dihydrogen isotopomer distributions during reactions of CH
2
O and CO
2
reforming of CH
4
on supported Pt clusters and to demonstrate a rigorous kinetic and
2
and H
2
O reforming, CH decomposition, and water-gas shift reactions.
4
2
2
H
D
2
2
4
4
4
2
2
1
2
12
13
4
2
/ CO / CO
2
2
4
/
CO
CO
2
/D
2
mixtures indicate that these products form in quasi-equilibrated steps. Turnover rates for H
2
O and
2
reforming and CH
4
decomposition increased with increasing Pt dispersion, suggesting that coordinative
unsaturated surface Pt atoms, prevalent in small crystallites, are more reactive than Pt atoms in low-index
surfaces for C-H bond activation. Pt dispersion but not turnover rates were influenced by the identity of the
2 2 3 2 2 4
support (ZrO , γ-Al O , ZrO -CeO ). Similar CO oxidation rates were measured before and after CH reactions,
indicating that Pt dispersion is not affected by unreactive deposits or sintering during catalysis. These
mechanistic conclusions and metal dispersion effects appear to apply generally to CH reactions on Group
4
VIII metals, but the surface reactivity of Pt clusters in C-H bond activation reactions is greater than for
similar size clusters of other metals. Turnover rates are compared here, for the first time, for most catalytically
important noble metals (Rh, Ir, Pt, Ru) as a function of their metal dispersion on several supports. These
turnover rates rigorously exclude transport and thermodynamic artifacts and provide a direct comparison of
the reactivity of noble metal clusters for catalytic reactions of CH
industrial practice.
4
on materials and at conditions relevant to
1
. Introduction
and chemisorbed carbon reacts with carbonates at metal-support
1
9-22
interfaces.
CO2-CH4 reforming at interfaces on Ni cata-
CH4 reactions with CO2 or H2O on supported Ni, Rh,3,4
1
,2
33
9,10
lysts were later proposed also for Pt-based catalysts.
This
Ru,5 Ir, and Pt
precursors to fuels and petrochemicals. Pt appears to be one of
the most active and stable metals for these reactions.
ZrO2, for instance, has been used in CH4-CO2 reactions for
5
sons among different metals are difficult because of ubiquitous
transport and thermodynamic artifacts and of incomplete as-
sessments of the number and cleanliness of exposed metal
atoms.
,6
7
2,3,8-32
lead to H2/CO mixtures useful as
proposal claimed that CH4 activation reversibly formed adsorbed
CHx and H species on Pt and then reacted with CO2 dissociation
fragments in equilibrated steps that form CHxO, which irrevers-
ibly decomposed to CO and H2 and led to a complex rate
equation:
1
3,19-23
Pt/
00 h without detectable deactivation.2
1-23
Rigorous compari-
aPCH4PCO2
rCH4
)
(1)
(4-x)/2
+ (1 + cPCH4)PCO2
H2
bPCOP
Redox cycles involving parallel CH4 activation on Pt clusters
and CO2 dissociation on reduced support sites were proposed
where x is the number of H-atoms in CHxO.
11
2
for CH4-CO2 reactions on Pt/ZrO2. Such cycles would turn
over with rates dependent on the relative abundance and
reactivity of metal and support sites, which must act in concert
to activate CH4 and remove carbon atoms via reactions with
CO2 (or H2O). Infrared spectra of CO2 adsorbed on Pt/Al2O3,
Pt/TiO2, and Pt/ZrO2 at 775 K indicated that carbonates formed
on supports, which led to the proposal that Pt activates CH4
Rostrup-Nielsen and Hansen reported similar activation
energies for CO2 and H2O reforming with CH4 on Pt/MgO,
suggesting that the two reactions share kinetically relevant steps.
Yet, measured reforming rates were about five times greater
with H2O than with CO2 co-reactants, a finding attributed to
rate-determining CO2 dissociation steps during CH4-CO2
reactions, despite the similar measured activation energies.
The effects of surface structure on reaction rates remain
unresolved for CH4 reactions on metal catalysts. Several studies
showed that C-H bond activation in alkanes is structure-
*
Author to whom correspondence should be addressed. Tel: (510)-642-
9
673. Fax: (510)-642-4778. E-mail: iglesia@cchem.berkeley.edu.
1
0.1021/jp036985z CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/10/2004

Activation and Chemical Conversion of Methane on Pt
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4095
TABLE 1: Pt Dispersion (D), Crystallite Diameter (d), and Forward CH
4
Turnover Rates on Pt-Based Catalysts (873 K, 20 kPa
CH
4
, 25 kPa CO
2
or H
2
O, 100 kPa Total Pressure, Balance Ar)
forward CH
4
turnover rate (s-1)
crystallite
Pt
diameter
CH
4
decomposition
(nm)b
CH
-CO
c
catalyst
dispersion
4
2
4 2
CH -H O
a
1
.6 wt % Pt/ZrO
.8 wt % Pt/ZrO
.4 wt % Pt/ZrO
.2 wt % Pt/ZrO
.6 wt % Pt/ZrO
.8 wt % Pt/ZrO
.4 wt % Pt/ZrO
2
2
2
2
2
2
2
-CeO
-CeO
-CeO
-CeO
2
2
2
2
0.256
0.345
0.435
0.572
0.159
0.276
0.383
0.219
3.9
2.9
2.3
1.7
6.3
3.6
2.6
4.4
15.0
18.6
21.1
23.8
12.8
16.4
14.8
19.5
22.5
24.7
13.1
16.9
19.2
14.7
0
0
0
1
0
0
1
18.6
12.4
15.8
.6 wt % Pt/Al
2
O
3
11.2
a
Zr/Ce(mole) ) 4. b Estimated from fraction Pt dispersion using equation D ) 1/d.48 Initial CH
c
4
decomposition turnover rate on Pt surface.
sensitive4,34,35 by the definition of Boudart,36 a conclusion
Al2O3, ZrO2, and ZrO2-CeO2, but they did not depend on the
identity of the support, as also found for Rh, Ir, and Ru catalysts.
The activation of the co-reactant (CO2 or H2O) is fast relative
to C-H bond activation steps and therefore not kinetically
relevant. These conclusions are consistent with those recently
consistent with experimental and theoretical studies on well-
defined surfaces;3
7-45
these latter studies indicate that coordi-
natively unsaturated surface atoms lead to higher CH4 sticking
and dissociation rates than atoms on close-packed surfaces. For
instance, CH4 dissociation rates using molecular beams are
higher on Pt(533) at 600 K than on close-packed Pt(111)
surfaces at even higher temperatures (800 K),44 and alkane
dissociation did not occur, except possibly at minority defect
4
6
reported for Rh and Ru clusters; they appear to apply generally
to chemical conversion of CH4 on Group VIII metals.
2
. Experimental Methods
2.1. Synthesis and Characterization. Pt/Al O , Pt/ZrO , and
45
sites, on Pt(111). CO2-CH4 and H2O-CH4 catalytic turnover
2
3
2
rates increased with increasing Rh3,4 and Ru6 dispersion,
Pt/ZrO -CeO with varying Pt content (0.2-1.6 wt %) were
2
2
prepared by incipient wetness impregnation of Al2O3, ZrO2 and
ZrO2-CeO2 with aqueous H2PtCl6‚6H2O solutions (Aldrich,
Lot# 10013LO, 99%). Samples were dried in ambient air at
consistent with higher rates on coordinatively unsaturated atoms.
We have not found similar systematic studies of Pt dispersion
effects on CH4 reforming turnover rates.
3
93 K and then treated in flowing dry air (Airgas, UHP, 1.2
Oxides used to support Pt clusters and metal-support
interfaces are often implicated in bifunctional CH4 reforming
pathways, but they also determine metal dispersion and transport
properties, and in doing so, can introduce thermodynamic and
transport corruptions into kinetic measurements. CH4 conver-
sions during CH4-CO2 reactions were higher on Pt/ZrO2 than
on Pt/SiO2 or Pt/Al2O3, a finding attributed to CO2 dissociation
on ZrO2, but neither Pt dispersion nor turnover rates were
3
-1
cm /g-s) at 873 K (0.167 K s ) for 5 h. Samples were reduced
3
-1
in H2 (Airgas, UHP, 50 cm /g-s) at 873 K (0.167 K s ) for 2
2
h. Al2O3 (160 m /g) was prepared by heating Al(OH)3 (Aldrich,
3
2
1645-51-2) in flowing dry air (Airgas, UHP, 1.2 cm /g-s)
-
1
46
at 923 K (0.167 K s ) for 5 h to form γ-Al2O3. ZrO2 (45
m /g) was prepared as previously described. ZrO2-CeO2 (Zr/
Ce ) 4) was synthesized by hydrolysis of 0.5 M ZrOCl2‚8H2O
(Aldrich, >98 wt %) and Ce(NO ) (Alfa, CAS # 15336-18-
2, 99.5%) at a constant pH of 10, maintained by controlled
2
4
1
1,12
measured.
Support effects would be detected when bifunc-
3 3
22
tional pathways require Pt-support interfaces, but they cannot
influence reaction rates directly when C-H bond activation on
metal surfaces is the only kinetically relevant step, as inferred
from relative bond energies in CH4 and CO2 (or H2O) and from
studies on well-defined metal surfaces.
addition of 14.8 M NH4OH. The precipitate was filtered and
washed by re-dispersing powders in a NH4OH solution (pH 10,
-
∼
333 K) until Cl ions were no longer detected by AgNO3 (Cl
<
10 ppm). ZrO2-CeO2 powders were then dried in ambient
air at 393 K overnight and treated in flowing dry air (Airgas,
UHP, 1.2 cm /g-s) at 923 K (0.167 K s ) for 5 h. Supported
Rh, Ru, and Ir catalysts were also prepared by incipient wetness
and detailed synthesis and characterization procedures are
Detailed kinetic and isotopic studies of CH4-H2O, CH4-
CO2, and CH4 decomposition reactions on Pt are unavailable;
reported reaction rates have not been, for the most part, corrected
for approach to equilibrium, measured under kinetic-controlled
conditions, or normalized by the number of exposed Pt atoms.
We report here, for the first time, a comparison of CH4 reaction
turnover rates on catalytically relevant noble metals (Pt, Ir, Rh,
Ru) for a wide range of metal dispersion (0.15-0.70) and
supports. We also report kinetic and isotopic evidence for the
identity and reversibility of elementary steps required for H2O
and CO2 reforming of CH4 on supported Pt clusters and
demonstrate a rigorous kinetic and mechanistic equivalence for
CO2 and H2O reforming, CH4 decomposition, and water-gas
shift reactions. These studies provide conclusive kinetic and
isotopic evidence for the sole kinetic relevance of C-H bond
activation and for the essentially uncovered state of Pt clusters
during steady-state catalysis and for the general relevance of
these conclusions to other noble metals. We also find that Pt
surfaces are the most active metal among Group VIII metals
for C-H bond activation and for CH4 reforming with H2O or
CO2. Turnover rates increased with increasing Pt dispersion on
3
-1
4
,6,47
reported elsewhere.
Pt dispersion was measured using volumetric methods from
uptakes of strongly chemisorbed hydrogen at 373 K.4 These
dispersions (shown in Table 1) were used to estimate average
crystallite diameters by assuming hemispherical geometry using:
,6
D ) 1/d
(2)
where D is the fractional dispersion and d is the crystallite
4
8
diameter (in nm).
2.2. Catalytic and Stoichiometric Reactions of Methane.
The microreactor system was described previously (quartz or
steel tube with 8 mm inner diameter and with a type K
thermocouple enclosed within a sheath in contact with the
4
,6
catalyst bed). Reaction rates were measured on catalysts (5
mg; 250-425 µm particles) diluted with inert support within
the pellets (25 mg) and then diluted the catalyst bed with acid-

4096 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Wei and Iglesia
washed quartz powder (500 mg) of similar size. Reactant
mixtures consisted of 50% CH4/Ar (Matheson) and 50% CO2/
Ar (Matheson) certified mixtures with He (Airgas, UHP) as a
diluent. H2O was introduced using a syringe pump (Cole-
Parmer, 74900 series), and transfer lines were kept above 373
K after H2O introduction to avoid condensation. The effects of
CH4 and CO2 pressures on CH4 reaction rates were measured
at 823-1023 K and 100-1000 kPa total pressure over a wide
range of reactant concentrations. The effect of H2O pressure
on CH4 reaction rates was measured at 823-1023 K and 100
kPa total pressure. Reactant and product concentrations were
measured with a Hewlett-Packard model 6890 gas chromato-
graph equipped with a Carboxen 1000 packed column (3.2 mm
×
2 m) and a thermal conductivity detector.
Pt/ZrO2 (0.8, 1.6 wt %, 20 mg) and Pt/ZrO2-CeO2 (0.4 wt
, 20 mg) samples diluted with 500 mg of quartz powder were
%
used to measure rates of CH4 and CD4 decomposition (without
co-reactants) at 873-973 K. CH4 conversions were measured
by on-line mass spectrometry (Leybold Inficon, Transpector
Series). Reactant mixtures with 20% CH4/Ar or 20% CD4/Ar
were prepared using 50% CH4/Ar (Matheson, certified mixture)
or pure CD4 (Isotec, chemical purity > 99.0%) and Ar (Airgas,
UHP). Intensities at 15 and 18 amu were used to measure CH4
and CD4 effluent concentrations, respectively. Ar was used as
an internal standard. Initial CH4 decomposition rates were used
to estimate rate constants for CH4 decomposition, based on the
observed linear dependence of decomposition rates on CH4
partial pressures.
Figure 1. Net CH
CO reaction on 1.6 wt % Pt/ZrO
with 50 mg of Al within pellets, then diluted with 500 mg of ground
quartz, pellet size 250-425 µm; (4) 5 mg of catalyst diluted with 25
mg of Al within pellets, then diluted with 500 mg of ground quartz,
pellet size 250-425 µm; (2) 5 mg of catalyst diluted with 25 mg of
Al within pellets, then diluted with 500 mg of ground quartz, pellet
4
turnover rates versus residence time for CH
4
-
2
2
(873 K, (b) 5 mg of catalyst diluted
2
O
3
2 3
O
2
O
3
size 63-106 µm).
measured on 1.6 wt % Pt/ZrO2 at conditions leading to stable
rates, without detectable carbon formation, sintering of metal
particles, or transport artifacts. Figure 1 shows net CH4 turnover
rates as a function of residence time at 873 K on 1.6 wt %
Pt/ZrO2 with different pellet sizes and different diluent/catalyst
ratios within pellets. Varying the diameter of catalyst pellets
2
.3. Isotopic Tracer Studies and Kinetic Isotope Effects.
Isotopic tracer studies of CH4 reforming reactions were carried
out on 0.4 wt % Pt/ZrO2-CeO2 using a transient flow apparatus
with short hydrodynamic delays (<5 s). Chemical and isotopic
compositions were measured using on-line mass spectrometry
(
250-425 vs 63-106 µm) or the extent of dilution within the
pellets (5:1 to 10:1) did not influence CH4 turnover rates (Figure
), indicating that measured net rates are unaffected by transport
(
Leybold Inficon, Transpector Series). CD4, D2O (Isotec,
13
chemical purity > 99.0%), 5% D2/Ar, and CO (Isotec,
chemical purity > 99.0%) were used as reactants without further
purification. Intensities at 15 and 17-20 amu were used to
measure the concentration of methane isotopomers. CH4 and
CD4 fragmentation patterns were measured and those for CHD3,
1
artifacts. Undiluted pellets and beds, often used in previous
studies, led to lower reaction rates as a result of mass and heat
transfer restrictions. Forward CH4 turnover rates were obtained
from measured rates by rigorously correcting them for reactant
depletion and approach to equilibrium in CO2 (eq 3) or H2O
4
9
CH2D2, and CH3D were calculated using reported methods.
Intensities at 18, 19, and 20 amu were used to determine water
isotopomers and those at 28, 29, 44, and 45 amu to measure
(eq 4) reforming reactions:
1
2
13
12
13
CO, CO, CO2, and CO2 concentrations, respectively.
2
2
[
PCO] [P ]
H
2
Kinetic isotope effects were measured from the relative reaction
rates with CH4/CO2 and CD4/CO2 reactants for CO2 reforming
and with CH4/H2O, CD4/H2O, and CD4/D2O reactants for steam
reforming. Detailed experimental conditions are shown with the
corresponding data in the results section.
1
η1 )
η2 )
×
(3)
(4)
[PCH4][PCO2] KEQ1
3
[
PCO][P ]
H
2
1
KEQ2
×
2
.4. CO Oxidation Rates. CO oxidation was used as a
[PCH4][P
]
H O
2
structure-insensitive reaction to detect any changes in the number
of exposed Pt atoms as a result of catalytic CH4 reforming
reactions. Pt/ZrO2 (0.4, 0.8, and 1.6 wt %, 10 mg) catalysts
diluted with 500 mg of quartz powder were used to measure
CO oxidation reaction rates at 413 K, 0.19 kPa CO, and 0.19
kPa O2. These rates were measured before and after reforming
reactions by measuring reactant and product concentrations using
gas chromatography and the same protocols described above
for CH4 reactions. Mixtures of 25% O2/He (Matheson, certified
mixture) and 81.5% CO/N2 (Matheson, certified mixture) were
used as reactants.
where [P ] is the average partial pressure of species j (in atm)
j
within the catalyst bed and K
EQ1
and KEQ2 are equilibrium
5
0
constants for each reaction at a given temperature. The
fractional distance from equilibrium (1 - η) ranges from 0.70
to 0.97 for the experiments reported here. Net turnover rates
(r ) are used to obtain forward turnover rate (r ) using
n
f
r
) r
(1 - η)
(5)
n
f
This equation accurately described all observed effects of reactor
residence time on rates, as indicated by the curve in Figure 1.
All rates reported from this point forward represent the rate of
forward reforming reactions.
3
. Results and Discussion
.1. Kinetic Dependence of Forward CH4 Reaction Rate
3
on Partial Pressure of Reactants. The kinetic response of CH4
reforming rates to CH4, CO2, and H2O concentrations was
Figure 2 shows the effects of average CH4, CO2, and H2O
pressures, defined as the linear average of their inlet and outlet

Activation and Chemical Conversion of Methane on Pt
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4097
Figure 2. Effects of CH
4
(a) and CO
2
or H
2
O (b) partial pressure on forward CH
4
turnover rate for CH and CH
4
-CO
2
4
2
-H O reactions on 1.6 wt
3
%
Pt/ZrO (5 mg of catalyst, 873 K, total flow rate 100 cm /min, 20 kPa CO
2
2
or H O in (a) and 20 kPa CH in (b), balance He).
2
4
Figure 3. Effects of CH
4
(a) and CO
2 4 2 4 2
(b) average partial pressures on forward CH turnover rate for CO reforming of CH on 1.6 wt % Pt/ZrO
3
(5 mg of catalyst, 873 K, total flow rate 100 cm /min, balance He, average pressure is the average of inlet and outlet pressures of the reactor).
values, on forward CH4 turnover rates at 873 K on 1.6 wt %
Pt/ZrO2. This linear average is an accurate representation of
the kinetic driving force for the small axial concentration
gradients prevalent in this study, and becomes rigorous if
reaction rates are first-order in CH4, as shown by the data in
Figure 2. Turnover rates for forward CO2 and H2O reforming
reactions increased linearly with increasing CH4 partial pressure
on all Pt-based catalysts examined here. Figure 3 shows the
effects of CH and CO pressures on forward CH turnover rates
at 873 K over a much larger pressure range (100-1000 kPa).
4
2
4
Forward CH turnover rates increased linearly with increasing
4
CH partial pressure throughout this wider pressure range, which
4
includes conditions relevant to industrial practice. Rates were
not influenced by CO2 pressure or by the pressure of H2, CO,
or H2O formed during reaction.
(Figure 2a); they were not influenced by CO2 or H2O partial
pressures (Figure 2b). Turnover rates were very similar for these
two reactions, indicating that co-reactants are required by
stoichiometry, but their activation or subsequent reactions are
not kinetically relevant. Forward turnover rates were not
influenced by CO and H2 pressures (0-10 kPa), whether varied
by changing residence time or by addition of H2 or CO to CH4-
CO2 or CH4-H2O reactants. These reaction products influence
the approach to equilibrium (η) and thus net rates, but not
forward rates (from eqs 3-5).
Figure 2b shows that CH4 turnover rates for H2O and CO2
reforming are similar for a given CH4 pressure. These rates are
also similar to those measured during the initial stages of CH4
decomposition on 1.6 wt % Pt/ZrO2 (without H2O or CO2 co-
reactants) (Figure 4, Table 2). First-order rate constants for H2O
(
kH O) and CO2 (kCO ) reforming and for CH4 decomposition
2
2
(kdecomp) are similar within experimental error at each temper-
ature (Figure 5); this leads to activation energies for CO2 (83
-
1
-1
kJ mol ) and H2O (75 kJ mol ) reforming and for CH4
CH4 reforming reactions can be accurately described by a
first-order dependence on CH4 and a zero-order dependence on
CO2, H2O, H2, or CO:
-
1
decomposition reactions (78 kJ mol ) that are also similar
within our ability to measure them.
These similarities in turnover rates, rate constants, and
activation energies for the three reactions indicate a common
kinetically relevant step, which cannot involve species derived
from co-reactants. The first-order CH4 dependence and the
measured kinetic insensitivity to co-reactants indicate that CH4
rf ) kPCH4
(6)
This first-order rate equation is identical for CH4-CO2 and
CH4-H2O reactions at all reaction temperatures (823-1023 K)

4098 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Wei and Iglesia
TABLE 2: Forward CH Reaction Rates, Rate Constants, Activation Energies, and Pre-exponential Factors for CH Reactions
4
4
on 1.6 wt % Pt/ZrO
2
(873 K, 20 kPa CH
4
, 25 kPa CO
2
2
or H O, 100 kPa Total Pressure, Balance Ar)
pre-exponential factor
-
1
-1
(s
kPa )
turnover rate
rate constant
activation energy
(kJ/mol)
-
1
a
-1
-1
estimatedb
co-reactant
(s
)
(s kPa
)
experimental
4
3
CO
2
12.8
13.1
12.4
0.64
0.66
0.62
83
75
78
5.9 × 10
5.5 × 10
4
3
H
2
O
2.0 × 10
5.5 × 10
4
3
None
2.9 × 10
5.5 × 10
a
b
4 4
Initial CH turnover rate on Pt surface. Calculatedon the basis of transition-state theory treatments of CH activation steps proceeding via an
5
1
immobile activated complex.
TABLE 3: Kinetic Isotope Effects for CH
4
Reactions on 1.6
, 25 kPa CO
wt % Pt/ZrO
2
(873 K, 25 kPa CH
4
or CD
4
2
,
2 2
H O, or D O, 100 kPa total pressure, balance Ar)
kinetic isotope effect
8
73 K
923 K
973 K
r
r
r
r
CD /H O:rCD /D O
1.07
1.69
1.77
1.58
1.12
1.49
1.44
1.05
1.31
1.34
4
2
4
2
CH /H O:rCD /H O
4
2
4
2
CH /CO2:rCD /CO2
4
4
CH /none:rCD /none
4
4
factors are 3-10 times larger than experimental values, a level
of agreement that is adequate in view of incompleteness of
transition state treatments, and which can be improved by
allowing modest two-dimensional mobility of activated com-
plexes.
3
.2. Kinetic Isotope Effects. The kinetic relevance of C-H
bond activation steps was confirmed by measuring isotope
effects from forward turnover rates for CH4-CO2 and CD4-
CO2 reactant mixtures at 873-973 K on 1.6 wt % Pt/ZrO2.
Kinetic isotope effects for H2O reforming were similarly
measured from forward turnover rates for CH4-H2O and CD4-
H2O reactants. The irrelevance of co-reactant activation and of
subsequent reactions of intermediates formed from water co-
reactants were confirmed from the similar turnover rates
measured with CD4-H2O and CD4-D2O mixtures. Kinetic
isotope effects for CH4 decomposition were measured from
initial rates of CH4 and CD4 decomposition.
Figure 4. CH
of time (873 K, 20 kPa CH
4
decomposition rates on 1.6 wt % Pt/ZrO
2
as a function
4
, 100 kPa total pressure, balance Ar).
Normal kinetic isotopic effects were obtained for CH4-CO2
and CH4-H2O reactions and for CH4 decomposition (Table 3);
their values were identical within experimental accuracy for all
three reactions at each reaction temperature. Kinetic isotope
effect values at 873 K were 1.69, 1.77, and 1.58 for CO2 re-
forming, H2O reforming, and CH4 decomposition, respectively.
These results confirm the involvement of similar kinetically rel-
evant C-H bond activation steps in these three reactions. CD4-
H2O and CD4-D2O reactant mixtures gave similar turnover rates
(kH/kD ) 1.05-1.12), indicating that activation of co-reactants
and any reactions involving adsorbed species formed from co-
reactants are not involved in kinetically relevant steps.
Kinetic isotope effects for CH4 activation on Pt(111) and
Figure 5. Arrhenius plots for CO
2
reforming ([), H
2
O reforming (0),
and CH
ZrO2.
4
decomposition (4) first-order rate constants on 1.6 wt % Pt/
34
52,53
Pt(110)(1 × 2)
surfaces have been measured using molecular
5
2
activation is the common kinetically relevant step. C-H bond
activation on essentially uncovered Pt surfaces is the sole
kinetically relevant step, and reactions of CO2 or H2O with CH4-
derived chemisorbed carbon species are fast, essentially quasi-
equilibrated, and kinetically irrelevant.
beam methods. Walker et al. reported a kinetic isotope effect
of 2 at 1000 K for CH4 sticking probabilities on Pt(110) (1 ×
2). We have not found previous kinetic isotope studies for
catalytic H2O and CO2 reforming or for stoichiometric decom-
position of CH4 on Pt. The values reported here for Pt resemble
those we have recently obtained in parallel studies on Ru (1.40-
1.51), Rh (1.54-1.60), and Ni (1.62-1.71), on which isotope
effects were also similar for H2O reforming, CO2 reforming,
and CH4 decomposition; these values also resemble those
reported by others for CH4 decomposition on Ni (1.60) at 773
If C-H bond activation is the sole kinetically relevant step
in these three reactions, measured rate constants, preexponential
factors can be estimated from transition state theory by merely
assuming a degree of mobility for the relevant activated
complex. Measured preexponential factors for these three
reactions and those predicted by assuming immobile activated
6
4
54
5
5
56,57
K. Several studies,
however, detected no kinetic isotope
51
complexes are compared in Table 2. Predicted preexponential
effects for CO2 reforming on Ni, apparently because near-

Activation and Chemical Conversion of Methane on Pt
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4099
SCHEME 1: Sequence of Elementary Steps for CH4
Reactions with H2O or CO2 on Pt Surfaces (w
irreversible step; S quasi-equilibrated step)
Figure 6. Water-gas-shift approach to equilibrium as a function of
space velocity at 873 K on 1.6 wt % Pt/ZrO
kPa CO , 25 kPa CH , 100 kPa total pressure, balance Ar, ηWGS
[PCO][PH2O])/([P 2][PCO2]KWGS)).
2
(Reaction conditions: 25
2
4
)
(
H
reaction temperatures (800-1200 K), as we conclude also from
the lack of kinetic inhibition by CO added to reactant streams.
equilibrium methane conversions led to measurements of little
kinetic relevance.
3
.4. Isotopic Tracer Studies of C-H Bond Activation
Steps. The reversibility of C-H bond activation during CO2
and H2O reforming on Pt catalysts was probed by measuring
rates of formation of deuterated isotopomers during reactions
of CH4/CD4/CO2 and CH4/CD4/H2O mixtures on 0.4 wt % Pt/
ZrO2-CeO2 at 873-973 K. Reversible C-H bond activation
would lead to similar chemical conversion and cross-exchange
rates, because the latter merely reflects the reverse of C-H
dissociation. Cross-exchange would occur more slowly than
chemical conversion when C-H bond activation is irreversible.
Chemical conversion and isotopic scrambling rates were
measured with CH4/CD4/CO2 (1:1:2) or CH4/CD4/H2O (1:1:2)
mixtures using on-line mass spectrometry, after removing
HxD2-xO at 218 K (to avoid interference between its fragments
and those for CHxD4-x). CHxD4-x (0e x e 4) formation and
chemical conversion rates are shown in Figure 7 at 873 K. The
CH4/CD4 cross exchange turnover rate, defined as the sum of
3
.3. Elementary Steps in Chemical Conversion of Meth-
ane. The similar kinetic rate expressions for H2O and CO2
reforming and decomposition reactions, taken together with
isotopic tracer studies and measured kinetic isotope effects, led
us to propose a common sequence of elementary steps for CH4
reactions on Ru, Rh, and Ni catalysts. Here, we extend this
mechanism, which contains all steps required for reforming,
decomposition, and water-gas shift reactions, to Pt surfaces.
Relevant elementary steps are shown in Scheme 1, where f
denotes an irreversible step, and § represents a quasi-
equilibrated step; ki is the rate constant and Ki the equilibrium
constant for step i.
6
4
54
CH4 reacts in a sequence of elementary steps to form
chemisorbed carbon and hydrogen. If unoccupied sites (*),
corresponding to free Pt ensembles, are the most abundant
reactive species, only the rate constant for step 1 (k1) influences
reaction rates; in this case, rates are proportional to CH4 pressure
and independent of the presence, identity, or specific concentra-
tion of CO2 or H2O co-reactants. CH4 decomposition is followed
by removal of CH4-derived fragments using intermediates
derived from CO2 or H2O co-reactants. Measured reactant and
product concentrations during CH4-CO2 and CH4-H2O reac-
tions indicate that water-gas-shift (WGS) reactions are equili-
brated on Pt at all reaction conditions (Figure 6). Thus,
elementary steps 2-5 in Scheme 1, which include all steps
required for water-gas shift, must also be quasi-equilibrated
during CH4 reforming reactions.
Hickman and Schmidt58 used a similar sequence and rate
parameters reported on model surfaces to describe nonisothermal
partial oxidation rates and selectivities on monolith-type Pt
catalysts. O2 dissociation and CO2, CO, and H2 formation and
desorption were assumed to be reversible, while CH4 dissocia-
tion was irreversible and assumed to have a activation energy
of 43 kJ/mol, as reported on Pt(111).59 The CO adsorption
enthalpy of 128 kJ/mol used in their model,58 would lead to
CO coverages below 0.1% at 873 K, even at equilibrium CH4
conversions; thus, no chemisorbed CO is expected during CH4
reforming or partial oxidation on Pt-based catalysts at typical
-1
CHD3, CH3D, and CH2D2 (twice) formation rates, is 1.9 s ,
while the forward turnover rate for methane chemical conversion
-1
is 19.0 s . The approach to equilibrium, η, during reaction was
0.10, while the ratio of the reverse (cross-exchange) to the
forward (chemical conversion) CH activation rates is also 0.10.
4
This indicates that CHD , CH D , and CH D isotopomers form
3
2
2
3
merely because the overall reaction itself approaches equilibrium
and becomes reversible, which requires that the rate-determining
step, when one exists, be exactly as reversible as the overall
catalytic sequence. In effect, C-H bond activation on Pt at 873
K is irreversible, except for the extent to which reversibility is
thermodynamically required by the approach to equilibrium for
the overall chemical reaction.
CH4/CD4 cross exchange and chemical conversion rates are
shown in Table 4 during reactions of CH4/CD4/H2O mixtures
at several reaction temperatures. At all temperatures, methane
chemical conversion rates are much higher than CH4/CD4 cross
exchange rates; thus, kinetically relevant C-H bond activation
steps are irreversible also during H2O reforming reaction, except
as required by overall thermodynamics.
3.5. Isotopic Tracer Studies of Co-Reactant Activation
Steps. The reversibility of CO2 activation steps (step 2 in
12
12
13
Scheme 1) was determined from reactions of CH4/ CO2/ CO

4100 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Wei and Iglesia
TABLE 4: Methane Chemical Conversion and CH
4
/CD
4
Cross Exchange Rates during Reaction of CH
4
/CD
4
/H
2
O Mixture on
0
.4 wt % Pt/ZrO
2
-CeO
2
Catalyst (12.5 kPa CH
4
, 12.5 kPa CD
4
, 25 kPa H O, 100 kPa total pressure, balance Ar)
2
methane
reaction
temperature(K)
chemical conversion
cross exchange
-
1
a
-1
b
r
exch/rreactionc
ηd
turnover rate (s
)
rate (s
)
8
9
9
73
23
73
20.2
44.1
72.7
1.8
4.8
6.7
0.09
0.11
0.09
0.10
0.09
0.12
a
Forward methane chemical conversion rate. Total methane isotopomers formation rate. c Ratio of total methane isotopomers formation rate to
b
d
3
the methane chemical conversion rate. η ) ([PCO][P 2
H
2
] )/([PH O][PCH4] K); approach to equilibrium for H O reforming reaction.
2
TABLE 5: Distribution of Water Isotopomers during
Reaction of CH /CO /D Mixtures on 0.4 wt % Pt/
ZrO -CeO (873 K, 16.7 kPa CH , 16.7 kPa CO , 3.3 kPa
, 100 kPa Total Pressure, Balance Ar)
4
2
2
2
2
4
2
D
2
distribution (%)
measured
(H/D ) 1.43)
binomial
(H/D ) 1.41)
a
isotopomer
H
HDO
D O
2
2
O
0.33
0.48
0.16
0.34
0.49
0.17
a
2
(H/D) ratio predicted from H in reacted methane and D in ambient
stream if complete mixing between the two isotopes occurred during
reaction.
(Figure 6); thus, water and dihydrogen formation must involve
quasi-equilibrated elementary steps.
The isotopic composition of products formed from CH4/CO2/
D2 (1:1:0.2) mixtures on 0.4 wt % Pt/ZrO2-CeO2 at 873-973
K was measured by on-line mass spectrometry. The H/D fraction
expected if H-atoms in any CH4 molecules converted to
synthesis gas and D-atoms in the added D2 contributed to surface
intermediates is 1.41. The measured H/D ratios were 1.43 and
Figure 7. Methane turnover rates and CH
during the reaction of CH /CD /CO mixture on 0.4% Pt/ZrO
catalyst (5 mg of catalyst, 873 K, 12.5 kPa CH and CD , 25 kPa CO
00 kPa total pressure, balance Ar).
4
/CD
4
cross exchange rates
4
4
2
2
-CeO
2
4
4
2
,
1
1
.44 in water and dihydrogen, respectively, indicating complete
reactants. Reversible CO2 dissociation would lead to rapid
equilibration between gas-phase dihydrogen and water molecules
and their corresponding chemisorbed precursors. Water isoto-
pomers formed from CH4/CO2/D2 mixtures obey a binomial
distribution (Table 5), as expected from quasi-equilibrated H*
and OH* recombination (step 5 in Scheme 1). The quasi-
equilibrated nature of the steps leading to water and dihydrogen,
taken together with the expected equilibration of nondissociative
CO chemisorption, would lead to thermodynamic equilibrium
for water-gas shift reactions, as found experimentally. Binomial
distributions were also observed for dihydrogen isotopomers at
1
3
formation of CO2 via its microscopic reverse, while an
12
irreversible step would preserve the isotopic purity of the CO2
reactants.
Reactions of 12CH4/ CO2/ CO (1:1:0.4) mixtures on 0.4 wt
12
13
13
%
Pt/ZrO2-CeO2 at 923 K led to similar C fractions in the
CO (0.25) and CO2 (0.23) present in the reactor effluent, even
though reactants were isotopically pure CO2 and CO. This
C content corresponds to complete chemical and isotopic
12
13
1
3
equilibration between CO and CO2, even at the low CH4
conversion (15%) and approach to equilibrium (η ) 0.12)
prevalent in these experiments. These data confirm that CO2
activation steps are much faster than the kinetically relevant
steps for CH4 chemical conversion to synthesis gas; thus, CO2
activation steps must be quasi-equilibrated and kinetically
irrelevant. In effect, steps 2 in Scheme 1 occur many times in
both directions in the time scale required for a CH4 chemical
conversion turnover, a situation that also preserved surfaces
essentially uncovered by CH4-derived reactive intermediates.
By inference from kinetic analogies between CH4-CO2 and
CH4-H2O reactions, we conclude that H2O activation steps are
also likely to be quasi-equilibrated during H2O reforming.
873-973 K, as expected from reversible and quasi-equilibrated
hydrogen recombinative desorption steps (step 4 in Scheme 1)
during CH4 reforming reactions.
3
.7. Effects of Pt Dispersion and of Support on CH4
Reactions. Methane activation studies on model surfaces have
concluded that activation of C-H bonds depends sensitively
on surface structure; specifically, surface steps and kinks are
60-62
much more active than atoms on close-packed surfaces.
Surface roughness and coordinative unsaturation increase as
metal clusters become smaller. To our knowledge, no systematic
studies of Pt cluster size effects on CH4 activation or reforming
rates are available.
3
.6. Isotopic Probes of Hydrogen and Hydroxyl Recom-
The effects of Pt dispersion on CH reforming turnover rates
4
bination Steps. Water forms during CO2 reforming of CH4 via
reverse water-gas shift (RWGS) reactions. More rigorously, H2O
forms because chemisorbed O-atoms from CO2 and chemisorbed
H-atoms from CH4 occasionally react to form water, in a
reaction that is nonrigorously treated as a separate stoichiometric
are shown in Table 1 and Figure 8 for CH -CO , CH -H O,
4
2
4
2
and CH decomposition reactions. CH turnover rates for each
4
4
reaction increased monotonically with increasing Pt dispersion,
suggesting that coordinatively unsaturated Pt surface atoms,
prevalent in small crystallites, are indeed more active than atoms
on the low-index surfaces predominately exposed on larger
crystallites. Surface atoms with fewer Pt neighbors appear to
bind CHx and H species more strongly and to stabilize activated
(water-gas shift) reaction in most previous studies. The expected
reactants and products of water-gas shift reactions exist at their
equilibrium relative concentrations during CH4 reforming

Activation and Chemical Conversion of Methane on Pt
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4101
Figure 8. Forward CH
4
turnover rates for CO
2
(a) and H
2
O (b) reforming of CH
4
as a function of Pt dispersion on various supports (873 K, 20
kPa CH , (2) Pt/ZrO2, (b) Pt/γ-Al
4
2
O , (]) Pt/ZrO -CeO , (O) CO turnover rate for CO oxidation at 413 K on Pt/ZrO ).
3 2 2 2
complexes involved in the formation of these intermediates, thus
decreasing C-H bond activation energies.
CO2, CH4 decomposition reactions on the Pt catalysts of this
study (Table 1, Figure 8), as expected from the rigorous
mechanistic equivalence among these reactions. Similar effects
of metal dispersion, also equivalent for all three reactions, were
measured in parallel studies of CH4 reforming and decomposi-
Thermal desorption and low-energy electron diffraction
studies showed that the activation energy for dissociative
chemisorption of ethane on Pt(110)-(1 × 2) surface is 25 kJ/
mol lower than on close-packed Pt(111) surfaces.6 Activation
barriers for dissociation of CH4 molecular beams on stepped
Pt(533) surfaces were 30 kJ/mol lower than that on close-packed
Pt(111).43 n-Butane and n-pentane adsorbed dissociatively on
3,64
4
6
47
tion reactions on Rh, Ru, and Ir catalysts.
Figure 9 shows similar effects of metal dispersion on turnover
rates for CH4-CO2 and CH4-H2O reactions on other supported
noble metals (Rh, Ir, Ru) in addition to the Pt catalysts of this
study. For each metal, at least two supports were used to disperse
metal clusters and no effects of support on turnover rates were
detected. On all metals, CH4 turnover rates increase with
increasing metal dispersion. Pt surfaces are more active than
those of the other noble metals for any given cluster size. This
appears to be the first rigorous and direct comparison of the
reactivity of supported Group VIII metal clusters for catalytic
reactions of CH4 using materials and conditions relevant to
industrial practice. The exclusion of transport artifacts and
reverse reactions and the measurement of rates normalized by
the relevant exposed surface area of metal clusters make these
rate comparisons a rigorous and sole reflection of the kinetic
reactivity of such metal surfaces.
65
Pt(110)-(1 × 2) but not on Pt(111) at 250 K. Alkane reactions
3
9
were not detected on Pt(111) in one ultrahigh vacuum study.
Molecular beam methods showed that initial ethane trapping
probabilities on open Pt(110)-(1 × 2) surfaces were greater
than on close-packed Pt(111).66 These effects of coordinative
unsaturation appear to be ubiquitous for molecule dissociation
reactions. In fact, theory and experiment showed that N2
9
dissociation on Ru is about 10 times faster on steps than on
terraces;67 activation barriers on steps were about 150 kJ/mol
smaller than on terraces. Molecular beam studies showed that
H2 dissociation is nonactivated on stepped Pt(332), but proceeds
6
8
with an activation barrier of 2-6 kJ/mol on Pt(111).
Figure 8 and Table 1 also show the effects of support on
forward CH4 turnover rates for CH4 reforming and decomposi-
tion reactions on Pt-based catalysts. The identity of the support
did not influence measured turnover rates, but it can influence
Pt dispersion. Once measured rates are rigorously normalized
by the number of exposed Pt surface atoms, CH4 reaction rates
depend only on the size (and structure) of Pt clusters and not
on the identity of the support. In any case, any potential
involvement of the support in activation of CO2 and H2O co-
reactants is inconsequential, because reaction rates depend only
on C-H bond activation rates.
3.8. Characterization of Supported Pt Clusters by CO
Oxidation. Here, we confirm the accuracy of chemisorptive Pt
dispersions measured before reaction and their relevance to the
state of Pt surfaces in working catalysts. These measurements
are critical to ensure that observed dispersion effects and
measured activation energies do not reflect blockage of exposed
Pt atoms by unreactive carbon residues during initial catalytic
turnovers, which could occur to varying extents as metal
dispersion or reaction conditions vary. We note that we have
not detected any transient effects during the first few seconds
of exposure to CH4-CO2 or CH4-H2O reactants, but typical
turnover times are short (<0.1 s), and any such initial transients
may not be detected when they occur concurrently with short
but unavoidable hydrodynamic delays (∼2 s).
1
1
One study reported much higher conversions during CO2
reforming on Pt/ZrO2 than on Pt/SiO2 and suggested that ZrO2
promotes CO2 dissociation, but neither Pt dispersions nor turn-
over rates were reported. Thus, these support effects may reflect
secondary effects of supports on Pt dispersion or on the extent to
which mass or heat transfer influenced reaction rates. Another
study2 suggested that CH4 conversion is controlled by Pt-support
contact perimeter, because of the bifunctional nature of CO2-CH4
reforming pathways on Pt. These support effects are inconsistent
with the kinetic irrelevance of co-reactant activation, shown here
for reaction conditions resembling those in these previous
studies, or with the lack of support effects reported in this study.
Similar dispersion and supports effects and similar actual
values of turnover rates were measured for CH4-H2O, CH4-
These measurements also address significant differences
between activation energies obtained from sticking coefficients
on model surfaces and from chemical conversion and stoichio-
metric decomposition rates of CH4 on supported Pt catalysts,
which may reflect changes in the state of catalytic surfaces
during reaction. Activation energies reported here (83 and 75
2
kJ/mol for CO2 and H2O reforming, respectively) and in
previous CO2 reforming studies on Pt-based catalysts3
,9,10
(63-
96 kJ/mol) are much larger than those measured or calculated
3
5,59,69
on model Pt surfaces.
For example, atom-superposition

4102 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Wei and Iglesia
Figure 9. Forward CH
4
turnover rates for CO
2
(a) and H
2
O (b) reforming of CH
4
as a function of metal dispersion on various supports (873 K,
2
0 kPa CH , (2) ZrO2, (b) γ-Al
4
2
O
3
, (]) ZrO -CeO
2
2
as support).
electron-delocation molecular orbital methods gave an activation
5
9
barrier of 43 kJ/mol for CH4 activation on Pt(111) surfaces.
Activation energies of 20 and 35 kJ/mol were reported for the
CH4 sticking coefficient using molecular beam methods on
35,69
Pt(110) (1 × 2) and Pt(111), respectively.
Our kinetic study indicates that reactions of H2O or CO2 with
CH4-derived chemisorbed carbon are fast and lead to essentially
uncovered surfaces during CH4 reforming reactions. Yet, we
cannot rule out that unreactive carbonaceous deposits, without
detectable reactivity in reactions with co-reactants, form during
the first few turnovers. For example, sites with the greatest
coordinative unsaturation and reactivity could form irreversibly
chemisorbed carbon, leaving less reactive surface Pt atoms to
catalyze CH4 conversion turnovers. These carbon deposits must
be totally unreactive during CH4 reforming, because their surface
density would otherwise vary with the concentration or reactivity
of the co-reactant, leading to apparent positive orders in the
co-reactant reactant concentration and to different rates with
H2O and CO2.
Figure 10. CO oxidation and forward CH reforming turnover rates
4
on 1.6 wt % Pt/ZrO
0.19 kPa CO, 0.19 kPa O
5 kPa CH , 25 kPa CO ; D, H
kPa CH , 25 kPa H O).
2
(A, C, E, CO oxidation turnover rates, 413 K,
; B, CO reforming turnover rates, 873 K,
O reforming turnover rates, 873 K, 25
Here, we use CO oxidation, a recognized structure-insensitive
70,71
2
2
reaction
(see also insensitivity of turnover rates to dispersion
2
4
2
2
in Figure 8a) to detect any changes in the number of exposed
Pt atoms during CH4 reforming, whether these changes occur
via blockage of surface sites or sintering of small Pt clusters.
We note that chemisorption measurements after reaction are not
possible, because of the very small catalyst amounts required
for isothermal rate measurements (5 mg). CO oxidation rates
on a fresh 1.6 wt % Pt/ZrO2 catalyst and on this catalyst after
CO2 and H2O reforming are identical within experimental
accuracy (Figure 10). This shows unequivocally that the number
of exposed Pt atoms is unchanged during reaction and that
unreactive carbon residues are present at very low coverages
4
2
appropriate for experimental and theoretical studies of CH4
reactions. They also highlight the essential requirement that
experiment and theory rigorously address complete catalytic
cycles, and not merely stoichiometric steps presumably involved
in such cycles.
4
. Conclusions
Isotopic tracer and kinetic measurements led to a simple
(<5%), if at all.
mechanistic picture and a unifying kinetic treatment of CH4-
CO2, CH4-H2O, and CH4 decomposition reactions, as well as
water-gas shift, on Pt-based catalysts. Reforming and decom-
position rates were first-order in CH4 concentration and
independent of the concentration or identity of the co-reactants,
suggesting that reaction rates are exclusively limited by C-H
bond activation on metal cluster surfaces and that co-reactant
activation is not kinetically relevant. These conclusions are
consistent with the similar activation energies measured for CH4
reforming and decomposition reactions. The normal CH4/CD4
kinetic isotope effects measured were similar for all three CH4
reactions and thus also independent of co-reactant identity.
We cannot rule out that some initially exposed surface atoms,
with remarkable reactivity in stoichiometric CH4 activation but
unable to turn over, become unavailable during initial contact
with reforming reactants and do not contribute to catalytic rates.
Their number would have to be extremely small, because CO
oxidation rates would otherwise be detectably influenced by CH4
reforming reactions. We conclude, however, that such sites, if
present, cannot turn over; thus, they are not relevant to the
analysis and prediction of catalytic rates of CH4-H2O and
CH4-CO2 reactions. These results raise important questions
about the nature of model surfaces and methods that are

Activation and Chemical Conversion of Methane on Pt
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4103
Water and dihydrogen formation steps are quasi-equilibrated,
as shown from their binomial isotopomers distributions during
(22) Bitter, J. H.; Seshan, K.; Lercher, J. A. J. Catal. 1998, 176, 93.
(
23) Lercher, J. A.; Bitter, J. H.; Hally, W.; Niessen, W.; Seshan, K.
Stud. Surf. Sci. Catal. 1996, 101, 463.
24) Nagaoka, K.; Seshan, K.; Lercher, J. A.; Aika, K. Catal. Lett. 2000,
70, 109.
13
reactions of CH4/CO2/D2 mixtures. The identical C contents
(
in CO and CO2 molecules during 12CH4/ CO2/ CO reactions
indicate that CO2 activation steps are reversible and quasi-
equilibrated. The quasi-equilibrated nature of these steps in
consistent with the observed chemical equilibrium among
reactants and products of the water-gas shift.
12
13
(25) Mo, L.; Fei, J.; Huang, C.; Zheng, X. J. Mol. Catal. A 2003, 193,
1
77.
(
26) Kim, G. J.; Cho, D. S.; Kim, K. H.; Kim, J. H. Catal. Lett. 1994,
28, 41.
(27) Qin, D.; Lapszewicz, J.; Jiang, X. J. Catal. 1996, 159, 140.
(28) Otsuka, K.; Ushiyama, T.; Yamanaka, I. Chem. Lett. 1993, 1517.
(29) Solymosi, F.; Kustan, G.; Erohelyi, A. Catal. Lett. 1991, 11, 149.
(30) Sakai, Y.; Saito, H.; Sodesawa, T.; Nozaki, F. React. Kinet. Catal.
Forward CH4 turnover rates increased monotonically with
increasing Pt dispersion for CO2 reforming, H2O reforming, and
CH4 decomposition reactions, suggesting that coordinative
unsaturation increases C-H bond activation reactivity. Supports
Lett. 1984, 24, 253.
(
31) Tokunaga, O.; Osada, Y.; Ogasawara, S. Fuel 1989, 68, 990.
(32) Mattos, L. V.; Rodino, E.; Resasco, D. E.; Passos, F. B.; Noronha,
(ZrO2, γ-Al2O3, ZrO2-CeO2) influence Pt dispersion, but not
F. B. Fuel Proc. Technol. 2003, 83, 147.
turnover rates, indicating that co-reactant activation on supports,
if it occurs, is not kinetically relevant. The rates of structure-
insensitive CO oxidation reactions are similar before and after
CH4 reforming; thus, the latter reaction does not influence the
number of exposed Pt atoms via sintering or blockage by
unreactive chemisorbed species.
These metal dispersion effects and the mechanistic conclu-
sions reached from kinetic and isotopic studies are similar to
those found in parallel studies of supported Rh, Ru, Ir, and Ni
catalysts and appear to apply in general to CH4 reactions on
Group VIII metals. Pt surfaces, however, are the most reactive
among the metal clusters examined.
(33) Bradford, M. C. J.; Vannice, M. A. Appl. Catal. A 1996, 142, 97.
(
(
34) Luntz, A. C.; Bethune, D. S. J. Chem. Phys. 1989, 90, 1274.
35) Reeves, C. T.; Seets, D. C.; Mullins, C. B. J. Mol. Catal. A 2001,
1
67, 207.
(
36) Boudart, M. AdV. Catal. 1969, 20, 153.
(37) Weinberg, W. H. SurV. Prog. Chem. 1983, 10, 1.
(38) Szuromi, P. D.; Weinberg, W. H. Surf. Sci. 1984, 145, 407.
(39) Firment, L. E.; Somorjai, G. A. J. Chem. Phys. 1977, 66, 2901.
(40) Salmeron, M.; Somorjai, G. A. J. Phys. Chem. 1982, 86, 341.
(41) Choudhary, T. V.; Goodman, D. W. J. Mol. Catal. A 2000, 163, 8.
42) Burghgraef, H.; Jansen, A. P.; van Santen, R. A. Faraday Discuss.
993, 96, 337.
43) Watwe, R. M.; Bengaard, H. S.; Rosrrup-Nielsen, J. R.; Dumesic,
(
1
(
J. A.; Norskov, J. K. J. Catal. 2000, 189, 16.
(44) Gee, A. T.; Hayden, B. E.; Mormiche, C.; Kleyn, A. W.; Riedmuller,
B. J. Chem. Phys. 2003, 118, 3334.
(
45) Szuromi, P. D.; Engstrom, J. R.; Weinberg, W. H. J. Phys. Chem.
Acknowledgment. This study was supported by BP as part
of the Methane Conversion Cooperative Research Program at
the University of California at Berkeley. Helpful technical
discussions with Drs. John Collins and Theo Fleisch (BP)
throughout these studies are gratefully acknowledged.
1
985, 89, 2497.
(46) Bahlawane, N.; Watanabe, T. J. Am. Ceram. Soc. 2000, 83, 2324.
(
47) Wei, J.; Iglesia, E. Angew. Chem. Int. Ed., in press.
(48) Boudart, M.; Djega-Mariadassou, G. The Kinetics of Heterogeneous
Catalytic Reactions; Princeton University Press: Princeton, NJ, 1984.
(49) Price, G. L.; Iglesia, E. Ind. Eng. Chem. 1989, 28, 839.
(50) Stull, D. R.; Edgar, F.; Westrum, J.; Sinke, G. C. In The
Thermodynamics of Organic Compounds; Robert E. Krieger Publishing
Co.: Malabar, FL, 1987.
(51) Dumesic, J. A.; Rudd, D. F.; Rekoske, J. E.; Trevino, A. A. The
Microkinetics of Heterogeneous Catalysis; American Chemical Society:
Washington, DC, 1993.
References and Notes
(1) Bradford, M. C. J.; Vannice, M. A. Appl. Catal. A 1996, 142, 97.
(2) Rostrup-Nielsen, J. R.; Bak Hansen, J. H. J. Catal. 1993, 144, 38.
(3) Zhang, Z. L.; Tsipouriari, V. A.; Efstathiou, A. M.; Verykios, X.
E. J. Catal. 1996, 158, 51.
(52) Walker, A. V.; King, D. A. J. Chem. Phys. 2000, 112, 4739.
(53) Luntz, A. C.; Winter, W. H. J. Chem. Phys. 1994, 101, 10980.
(54) Wei, J.; Iglesia, E. J. Catal., in press.
(55) Otsuka, K.; Kobayashi, S.; Takenaka, S. J. Catal. 2001, 200, 4.
(56) Schuurman, Y.; Kroll, V. C. H.; Ferreira-Aparicio, P.; Mirodatos,
C. Catal. Today 1997, 38, 129.
(
(
(
(
4) Wei, J.; Iglesia, E. J. Catal., in press.
5) Bradford, M. C. J.; Vannice, M. A. J. Catal. 1999, 183, 69.
6) Wei, J.; Iglesia, E. J. Phys. Chem. B, in press.
7) Erdohelyi, A.; Fodor, K.; Solymosi, F. Stud. Surf. Sci. Catal. 1991,
1
07, 525.
8) Ashcroft, A. T.; Cheetham, A. K.; Green, M. L. H.; Vernon, P. D.
F. Nature 1991, 352, 225.
9) Bradford, M. C. J.; Vannice, M. A. J. Catal. 1998, 173, 157.
(
(57) Kroll, V. C. H.; Swaan, H. M.; Lacombe, S.; Mirodatos, C. J. Catal.
1997, 164, 387.
(
(58) Hickman, D. A.; Schmidt, L. D. AIChE J. 1993, 39, 1164.
(59) Anderson, A. B.; Maloney, J. J. J. Phys. Chem. 1988, 92, 809.
(60) McMaster, M. C.; Madix, R. J. J. Chem. Phys. 1993, 98, 15.
(61) Liu, Z.; Hu, P. J. Am. Chem. Soc. 2003, 125, 1958.
(62) Klier, K.; Hess, J. S.; Herman, R. G. J. Chem. Phys. 1997, 107,
4033.
(63) Johnson, D. F.; Weinberg, H. Science 1993, 261, 5117.
(64) Johnson, D. F.; Weinberg, H. J. Chem. Phys. 1995, 103, 5833.
(65) Seuromi, P. D.; Engstrom, J. R.; Wittrig, T. S.; Weinberg, W. H.
J. Vac. Sci. Technol. A 1985, 3, 1560.
(10) Bradford, M. C. J.; Vannice, M. A. Catal. Lett. 1997, 48, 31.
(11) Stagg, S. M.; Romeo, E.; Resasco, D. E. J. Catal. 1998, 178, 137.
(12) Noronha, F. B.; Fendley, E. C.; Soares, R. R.; Alvarez, W. E.;
Resasco, D. E. Chem. Eng. J. 2001, 82, 21.
13) Stagg, S. M.; Noronha, F. B.; Fendley, E. C.; Resasco, D. E. J.
Catal. 2000, 194, 240.
14) Souza, M. M. V. M.; Aranda, D. A. G.; Schmal, M. J. Catal. 2001,
04, 498.
15) Souza, M. M. V. M.; Aranda, D. A. G.; Schmal, M. Ind. Eng. Chem.
Res. 2002, 41, 4681.
(
(
2
(
(66) Stinnett, J. A.; McMaster, M. C.; Schroeder, S. L. M.; Madix, R.
J. Surf. Sci. 1996, 365, 683.
(16) Chen, Y.; Liaw, B.; Lai, W. Appl. Catal. A: General 2002, 230,
7
2
1
3.
(
17, 23.
(
(67) Dahl, S.; Logadottir, A.; Egeberg, R. C.; Larsen, R. C.; Chorken-
dorff, I. Phys. ReV. Lett. 1999, 83, 1814.
17) Chen, Y.; Liaw, B.; Kao, C.; Kou, J. Appl. Catal. A: General 2001,
(68) Salmeron, M.; Gale, R. J.; Somorjai, G. A. J. Chem. Phys. 1979,
70, 2807.
18) Nagaoka, K.; Seshan, K.; Aika, K.; Lercher, J. A. J. Catal. 2001,
97, 34.
(69) Arumainayagam, C. R.; Mcmaster, M. C.; Schoofs, G. R.; Madix,
R. J. Surf. Sci. 1989, 222, 213.
(70) Boudart, M.; Rump, F. F. React. Kinet. Catal. Lett. 1987, 135, 95.
(71) Engel, G.; Ertl, G. J. Chem. Phys. 1978, 69, 1267.
(
(
(
19) Bitter, J. H.; Seshan, K.; Lercher, J. A. Top. Catal. 2000, 10, 295.
20) Bitter, J. H.; Seshan, K.; Lercher, J. A. J. Catal. 1999, 183, 336.
21) Bitter, J. H.; Seshan, K.; Lercher, J. A. J. Catal. 1997, 171, 279.