Reaction pathways and site requirements for the activation and chemical conversion of methane on Ru-based catalysts

Article abstract of DOI:10.1021/jp030783l

Kinetic and isotopic tracer and exchange measurements were used to determine the identity and reversibility of elementary steps required for CH₄ reforming reactions on Ru-based catalyst. CH₄ reactions were limited by C-H bond activat

Full text of DOI:10.1021/jp030783l

J. Phys. Chem. B 2004, 108, 7253-7262
7253
Reaction Pathways and Site Requirements for the Activation and Chemical Conversion of
Methane on Ru-Based Catalysts
Junmei Wei and Enrique Iglesia*
Department of Chemical Engineering, UniVersity of California at Berkeley, Berkeley, California 94720
ReceiVed: June 24, 2003; In Final Form: September 29, 2003
Kinetic and isotopic tracer and exchange measurements were used to determine the identity and reversibility
of elementary steps required for CH4 reforming reactions on Ru-based catalyst. These studies provide a simple
mechanistic picture and a unifying kinetic treatment for CH4/CO2 and CH4/H2O reforming reactions and CH4
decomposition. Forward kinetic rates were measured from net rates by correcting for the approach to
equilibrium, after ruling out transport artifacts using pellet and bed dilution tests. The kinetic processes involved
are exclusively limited by C-H bond activation, and CH4 reaction rates are unaffected by the identity or the
concentration of co-reactants (H2O, CO2). Similar normal kinetic isotopic effects (kC-H/kC-D ) 1.40-1.51)
were measured for CO2 reforming, H2O reforming, and CH4 decomposition, consistent with kinetically relevant
C-H bond activation steps. The ratio of CH4/CD4 cross-exchange to methane chemical conversion rates
during the reaction of CO2 reforming with CH4-CD4 mixtures was 0.05, suggesting that steps involving
C-H bond activation are essentially irreversible. Binomial D-atom distributions in dihydrogen and water
were obtained during reactions of CH4/CO2/D2 mixtures, and their D-contents were identical to those expected
from complete equilibration between D2 and H-atoms from reacted CH4, indicating that H-OH and H-H
recombination steps are quasi-equilibrated. Reactions of 12CH4/12CO2/13CO mixtures gave identical 13C contents
in CO and CO2, even far away from the CO2 reforming equilibrium; thus, CO2 activation is reversible and
quasi-equilibrated during CO2 reforming on Ru-based catalysts, as expected from the kinetic irrelevance of
co-reactant activation steps. These conclusions suggest that water-gas shift reactions are also equilibrated, as
confirmed by chemical analyses of reaction products. Forward CH4 turnover rates increased with increasing
Ru dispersion, but they were essentially unaffected by the identity of the support. This behavior reflects the
higher reactivity of coordinatively unsaturated surface atoms, prevalent in small Ru clusters, for C-H bond
activation reactions, as previously inferred from the effect of crystal orientation on CH4 activation rates.
2
k(PCH - (PH PCO2/RPCO ))
Introduction
4
2
2
r )
(1)
1+(PCO2/âPCO
)
CH4 reactions with CO2 or H2O can be used to produce
synthesis gas mixtures for ultimate conversion to desired fuels
and chemicals. Fischer and Tropsch1 first showed that group
VIII metals (Ni, Ru, Rh, Pt, Pd, and Ir) catalyzed CO2-CH4
reactions to form these H2-CO mixtures. Specifically, Ru
clusters supported on Al2O3,2-13 TiO2,2,14 MgO,15,16 La2O3,7-9
SiO2,12,17 NaY,18 and carbon2,19 effectively catalyze these
reactions. The relevant elementary steps and the effects of metal
dispersion and support on reaction rates have not been un-
equivocally established on supported Ru catalysts.
2
Studies of the stoichiometric activation of CO2 and CH4 on Ru/
Al2O3 led to a proposal that CH4 dissociation is aided by
chemisorbed oxygen formed via CO2 dissociation; the latter was
in turn promoted by chemisorbed H-atoms formed in C-H bond
activation.3 CO2 reforming turnover rates on Ru/TiO2, Ru/C,
and Ru/Al2O3 were influenced by conversion, because reverse
steps contributed to measured rates as reactions approached
equilibrium.2 After corrections for reverse reactions, forward
rates were accurately described by a simple rate equation:
Bifunctional CO2 reforming pathways on Ru/Al2O3 were
proposed to involve CH4 decomposition on Ru and CO2
activation on OH groups in the Al2O3 support.11 Matsui et al.9
proposed a redox mechanism on Ru supported on La2O3, ZrO2,
and Y2O3, in which CH4 forms Ru-CHx species and CO2
dissociates to form CO and chemisorbed oxygen atoms; Ru-
CHx then reacts with chemisorbed oxygen atoms to form CO
and Ru. Mark and Maier10 proposed rate-determining CH4
decomposition steps to form chemisorbed carbon and H2, and
the reaction of carbon with CO2 in a fast process; this proposal
led to a rate equation consistent with kinetic data:
a
b
rf ) kfPCH PCO
(2)
4
2
where a and b are given by 0.52((0.36) and 0.21((0.40),
respectively. This expression was shown to be consistent with
a sequence involving slow and reversible CH4 dissociation to
form CHx species and irreversible slow decomposition of CHxO
species to form CO and hydrogen, but these conclusions
remained speculative because of large uncertainties in reported
reaction orders.
Rostrup-Nielsen and Hansen16 reported the only parallel study
of CO2 and H2O reforming reactions on Ru catalysts. They
proposed that CO2 and H2O reforming mechanisms are similar
* Author to whom correspondence should be addressed. Tel: (510) 642-
9673. Fax: (510) 642-4778. E-mail: iglesia@cchem.berkeley.edu.
10.1021/jp030783l CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/29/2004

7254 J. Phys. Chem. B, Vol. 108, No. 22, 2004
Wei and Iglesia
on Ru/MgO, but found significantly higher reaction rates for
steps and with the kinetic irrelevance of CO2 and H2O activation.
These conclusions resemble those reached in our recent studies
of CH4 reforming and decomposition reactions on Rh,25 Pt,26
Ir,27 and Ni28 catalysts, the evidence for which is presented
elsewhere.
H2O reforming (80 kPa H2O, 20 kPa CH4) (8.9 mol site-1 s-1
)
than for CO2 reforming (80 kPa CO2, 20 kPa CH4) (2.9 mol
site-1 s-1) at 823 K. These authors proposed that the replacement
of H2O co-reactants with CO2 introduced a kinetic bottleneck
associated with CO2 activation, which became the rate-
determining step in CO2 reforming reactions
Experimental Methods
Several studies on model metal surfaces have concluded that
C-H bond activation is probably the kinetically relevant step
in CH4 conversion reactions, but these studies have measured
only rates of stoichiometric CH4 decomposition reactions,
typically at temperatures much lower than those required to
overcome kinetic and thermodynamic hurdles during CH4
reforming catalysis.20(and refs therein) For these systems and reac-
tions, C-H bond dissociation occurs more rapidly on step and
kink sites than on terrace sites, apparently because of the higher
reactivity of coordinatively unsaturated surface metal atoms.22-24
Stoichiometric activation of CH4 appears to depend sensitively
on surface structure; this structure sensitivity, by the definition
of Boudart,21 should lead to strong effects of metal cluster size
on catalytic turnover rates. Yet, we have not found systematic
studies of dispersion effects on CH4 reactions catalyzed by Ru
or of the effects of Ru surface structure on catalytic reforming
reactions.
Ru/Al2O3 with 1.6 and 3.2 wt % Ru and Ru/ZrO2 with 3.2
wt % Ru were prepared by incipient wetness impregnation of
Al2O3 or ZrO2 with an aqueous solution of Ru(NO)(NH3)3 (Alfa,
CAS#34513-98-9). Impregnated samples were dried at 393 K
in ambient air and treated in flowing dry air (Airgas, UHP, 1.2
cm3/g-s) by increasing the temperature to 873 K at 0.167 K s-1
and holding at 873 K for 5 h. Samples were then treated in H2
(Airgas, UHP, 50 cm3/g-s) by heating to 873 K at 0.167 K s-1
and holding at 873 K for 2 h. The 3.2 wt % Ru/Al2O3 sample
was also treated in H2 (Airgas, UHP, 50 cm3/g-s) by increasing
the temperature to 1023 K at 0.167 K s-1 and holding at 1023
K for 2 h in order to vary the size of Ru clusters. Al2O3 (160
m2/g) was prepared by treating Al(OH)3 (Aldrich, 21645-51-2)
in flowing dry air (Airgas, UHP, 1.2 cm3/g-s) while increasing
the temperature to 923 K at 0.167 K s-1 and holding at 923 K
for 5 h, a procedure that leads to γ-Al2O3.29 ZrO2 (45 m2/g)
was prepared by hydrolysis of a 0.5 M aqueous solution of
ZrOCl2‚8H2O (Aldrich, >98 wt %) at a constant pH of 10,
maintained by addition of controlled amounts of a 14.8 M NH4-
OH solution.30 The precipitates were immediately filtered and
washed repeatedly by redispersing it in a warm NH4OH solution
(pH 10, ∼333 K) to remove residual Cl ions, until no Cl ions
were detected by a AgNO3 test (Cl- < 10 ppm). The samples
were then dried at 393 K overnight in ambient air and treated
in flowing dry air (Airgas, UHP, 1.2 cm3/g-s) by heating to
923 K at 0.167 K s-1 and holding at 923 K for 5 h. X-ray
diffraction showed the predominant presence of monoclinic
ZrO2.
Ru dispersion was measured by volumetric H2 chemisorption
at 373 K31 using a Quantasorb chemisorption analyzer (Quan-
tachrome Corp.). Catalysts were reduced in H2 at 873 K for 2
h, and then evacuated at 873 K for 0.5 h. After cooling to 373
K, a H2 chemisorption isotherm was measured between 3 and
50 kPa. A backsorption isotherm was measured by repeating
this procedure after evacuating samples at 373 K for 0.5 h. Both
isotherms were extrapolated to zero H2 pressure and their
difference used as a measure of the uptake of strongly
chemisorbed hydrogen. Ru dispersions were calculated by
assuming that one hydrogen atom chemisorbed on each surface
Ru;31 dispersion values are shown in Table 1 for each of the
catalysts used in this study.
The supports used to disperse Ru crystallites often influence
CO2 reforming rates, but concurrent effects of supports on Ru
dispersion, on transport artifacts, or on approach to equilibrium
are seldom independently considered. Matsui9 found that CH4
conversions in CO2 reforming were higher on Ru/ZrO2 and Ru/
La2O3 than on Ru/Al2O3, and proposed, without direct evidence,
that such effects arose from the different reactivities of the
various supports in CO2 activation. Ferreira-Aparicio et al.11
suggested that OH groups on supports catalyzed rate-determining
CO2 activation steps. Bradford and Vannice2 found higher
turnover rates when Ru was dispersed on TiO2 than on Al2O3
or carbon, even though Ru dispersions (from H2 chemisorption)
were lower on TiO2 (51%) supports than on Al2O3 (78%) or
carbon (100%) supports. This study provided infrared evidence
for the decoration of Ru crystallites with TiOx species during
catalyst reduction and proposed that such sites exhibit unique
catalytic activity because of the resulting intimate metal-support
contacts. The nature of these interactions, their specific role in
kinetically relevant steps, and even the survival of these
decoration effects in contact with CO2 and H2O at high
temperatures remain unclear.
Here, we probe the identity and reversibility of elementary
steps required for H2O and CO2 reforming of CH4 on supported
Ru catalysts. We provide evidence for a catalytic sequence that
rigorously combines the kinetics and pathways for water-gas
shift, CH4 decomposition, and CO2 and H2O reforming reac-
tions. Kinetic and isotopic experiments confirmed this se-
quence and established the sole kinetic relevance of C-H
bond activation and the essentially uncovered nature of Ru
surfaces during steady-state catalysis. Reaction rates were
measured in the absence of transport artifacts and rigorously
corrected for the approach to equilibrium of reforming react-
ions. Similar rate constants determined for C-H bond activation
in H2O reforming, CO2 reforming, and CH4 decomposition
reactions are compared on samples with varying Ru dispersion
on Al2O3 and ZrO2 supports. Turnover rates were strongly
influenced by Ru dispersion but essentially insensitive to the
support used and to the identity or concentration of the
co-reactant, consistent with CH4 activation rate-determining
Catalytic rates were measured by placing samples (5 mg,
250-425 µm) within a quartz or steel tube (8 mm inner
diameter) with a type K thermocouple enclosed within a
sheath in contact with the catalyst bed. Samples were diluted
with ground, acid-washed quartz powder (500 mg, 250-425
µm) to avoid temperature gradients. Transport artifacts were
ruled out using pellet and bed dilution with the pure support
without detectable changes in rates or selectivities, as shown
in the Appendix. The effects of CH4, H2O, and CO2 pres-
sures on CH4 reaction rates were measured at 823-1023 K and
0.1-0.5 MPa total pressure over a wide range of reactant
concentrations. Reactant mixtures were prepared using 50%
CH4/Ar (Matheson) and 50% CO2/Ar (Matheson) certified
mixtures and He (Airgas, UHP) as balance. For H2O reforming
reactions, H2O was introduced using a syringe pump (Cole-
Parmer, 74900 series). All transfer lines after H2O introduction

Chemical Conversion of CH4 on Ru-Based Catalysts
TABLE 1: Forward CH4 Turnover Rate on Supported Ru Catalysts (873 K, 20 kPa CH4, 20 kPa CO2 or H2O)
Forward CH4 Turnover Rate (s-1
J. Phys. Chem. B, Vol. 108, No. 22, 2004 7255
)
reduction
Ru
CH4
catalyst
temperature (K)
dispersion (%)
CH4-CO2
CH4-H2O
decomposition
reference
1.6 wt % Ru/γ-Al2O3
3.2 wt % Ru/γ-Al2O3
3.2 wt % Ru/γ-Al2O3
3.2 wt % Ru/ZrO2
873
873
1023
873
55.5
44.2
33.1
29.8
4.8
3.1
2.7
2.5
4.9
3.3
2.6
2.3
this study
this study
this study
this study
3.1
2.2
1.6 wt % Ru/η-Al2O3
4.8 wt % Ru/C
1.0 wt % Ru/γ-Al2O3
0.64 wt % Ru/γ-Al2O3
1.0 wt % Ru/NaY
78.0
100
5.5
51.0
28.0
15.3a
1.1a
1.2a
5.1a
2.6a
2
2
3
12
18
a
Net rates were corrected to forward rates by approach to equilibrium using eq 5, then extrapolated to our reaction conditions (873 K, 20 kPa
CH4) using r ) A exp(-Ea/RT) PCH .
4
Figure 1. Effects of CH4 (a) and CO2 (b) partial pressure on forward CH4 reaction rate for CO2 reforming of CH4 on 3.2 wt % Ru/Al2O3 reduced
at 873 K (5 mg catalyst, 873 K, total flow rate 100 cm3/min, balance He, average pressure is the average of inlet and outlet pressures of the reactor).
were kept above 373 K to avoid condensation. Reactant and
product concentrations were measured with a Hewlett-Packard
6890 gas chromatograph using a Carboxen 1000 packed column
(3.2 mm × 2 m) and thermal conductivity detection. Unless
otherwise noted, catalysts were reduced at 873 K before CH4
reforming reactions. No products were detected at 823-1023
K in empty reactors.
Ru/Al2O3 and Ru/ZrO2 catalysts (20 mg, treated in H2 at 873
K; 3.2 wt %) diluted with 500 mg quartz powder were used for
CH4 and CD4 decomposition reactions at 873 K. Chemical
compositions 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 CD4 (Isotec, chemical purity
> 99.0%) with Ar (Airgas, UHP) as an inert internal standard
used for accurate CH4 conversion measurements. Initial CH4
decomposition rates were used to estimate rate constants for
CH4 decomposition using the observed linear dependence of
rates on CH4 concentration.
Isotopic tracer studies were carried out on 3.2 wt % Ru/Al2O3
reduced at 873 K with a 44.2% Ru dispersion using a transient
flow apparatus with short hydrodynamic delays (<5 s). Chemi-
cal and isotopic compositions were measured using on-line mass
spectrometry (Leybold Inficon, Transpector Series). CD4 (Isotec,
chemical purity > 99.0%), D2O (Isotec, chemical purity >
99.0%), and 5% D2/Ar and 13CO (Isotec, chemical purity >
99.0%) were used as reactants without further purification.
Intensities at 15 and 17-20 amu were used to measure methane
isotopomer concentrations. CH4 and CD4 standard fragmentation
patterns were measured, and those for CHD3, CH2D2, and CH3D
were calculated using reported methods.32 Intensities at 18, 19,
and 20 amu were used to determine water isotopomers and those
at 28, 29, 44, and 45 amu to measure 12CO, 13CO, 12CO2, and
13CO2 concentrations, respectively. Detailed experimental condi-
tions are shown together with the corresponding data in the
Results section.
Carbon formation rates were measured during reforming
reactions at 873 K using a tapered element quartz oscillating
microbalance (Rupprecht & Patashnick, Series 1500). Catalyst
treatment procedures and reaction conditions were similar to
those used in kinetic measurements.
Results and Discussion
Kinetic Dependence of Reforming Rates on CH4, CO2, and
H2O Partial Pressures. The kinetic dependence of CH4 reform-
ing rates on CH4, CO2, and H2O concentrations was measured
on 3.2 wt % Ru/Al2O3 treated at 873 K in H2 (44.2% Ru
dispersion) at conditions leading to stable rates and undetectable
carbon formation. Filament carbon formation was not detected
in parallel microbalance experiments or by transmission electron
microscopy analyses of catalyst samples after use.
Figure 1 shows the effects of CH4 and CO2 pressures on
forward CH4 turnover rate (rf, normalized by the number of
exposed surface Ru atoms) at 873 K and 100-500 kPa total
pressure. Net rate measurements far from equilibrium require
very low CH4 conversions at low temperatures, because of
unfavorable thermodynamics, and they become impractical at
high temperatures, because very fast reaction rates lead to
ubiquitous temperature and concentration gradients. Measured
reaction rates were corrected for approach to equilibrium (η)
using thermodynamic data33 and prevalent pressures of reactants

7256 J. Phys. Chem. B, Vol. 108, No. 22, 2004
Wei and Iglesia
Forward CH4-CO2 reaction rates increased linearly with
increasing CH4 partial pressure (5-125 kPa) at 873 K and were
independent of CO2 partial pressure (5-125 kPa) (Figure 1(a,b)).
Forward rates were also insensitive to CO, H2, and H2O
pressures, whether these pressures were varied by adding these
species to the inlet stream or by changing residence times and
CH4 conversions. Measured concentrations during CH4 reform-
ing corresponded to equilibrated water-gas-shift (WGS) reactions
at all temperatures between 823 and 1023 K (Figure 2). CH4-
CO2 reaction rates are simply described by a first-order
dependence in CH4 and a zero-order dependence in CO2,
rf ) kCO PCH
(6)
2
4
Once reverse reaction rates are considered using eqs 3-5, this
expression describes CO2 reforming rates at all temperatures
(823-1023 K).
Figure 2. Extent of water-gas-shift equilibrium at different reaction
temperatures as a function of space velocity on 3.2 wt % Ru/Al2O3
catalysts reduced at 873 K (44.2% Ru dispersion). (Reaction condi-
tions: CO2/CH4/Ar ) 1:1:2, 100 kPa total pressure, ηWGS ) ([PCO]-
[PH O])/([PH ][PCO ]KWGS)).
More complex rate expressions reported in previous studies3,10
may reflect transport artifacts or nonrigorous accounts of reverse
reactions. Equation 6 is consistent with CH4 activation on Ru
surfaces as the sole kinetically relevant elementary step and with
fast steps involving recombinative hydrogen desorption to form
H2 and reactions of CO2 with CH4-derived chemisorbed species
to form CO. These fast steps maintain Ru surfaces essentially
uncovered by reactive intermediates during CH4-CO2 reactions.
Otherwise, higher CO2 pressure would increase the rate of
removal of adsorbed intermediates and lead to positive effects
on CH4 reforming rates. These data do not preclude the presence
of completely unreactive residues during catalysis, an issue that
we address below.
2
2
2
and products to give forward rates for CH4-CO2 and CH4-
H2O reactions:
[PCO]2[PH ]2
2
1
KEQ1
η1 )
η2 )
×
×
(3)
(4)
[PCH ][PCO
]
4
2
[PCO][PH ]3
2
1
KEQ2
The kinetic irrelevance of carbon removal by co-reactants
and the mechanistic equivalence of H2O and CO2 reforming
reactions were confirmed by CH4-H2O reaction rates measured
on 3.2 wt % Ru/Al2O3. These rates are shown together with
those for CH4-CO2 reactions in Figure 3 as a function of CH4
and co-reactant pressures. Forward CH4-H2O reaction rates are
proportional to CH4 partial pressures (5-25 kPa) and indepen-
dent of H2O partial pressure (5-25 kPa). As in CH4-CO2
reactions, rates are simply described by
[PCH ][PH O
]
4
2
In these equations, [Pj] is the average partial pressure of species
j (in units of atm) in the reactor. Average pressures were used
in order to correct for minor depletion of reactants along the
catalyst bed. KEQ1 and KEQ2 are equilibrium constants for each
reforming reaction.33 The (1 - η) values were larger than 0.8
for all catalytic measurements reported here. Net reaction rates
(rn) are used to obtain forward reaction rates using
rf ) kH O PCH
(7)
2
4
rn ) (rf - rr) ) rf (1 - η)
(5)
Rate constants for H2O (kH O) and CO2 (kCO ) reforming are
2
2
where rr is the reverse reaction rate.34 This equation accurately
described the observed effects of reactor residence time and CH4
conversion level on reforming rates.
similar to each other at each reaction temperature (Figure 4)
and show similar activation energies. The corresponding pre-
exponential factors for these rate constants are shown in Table
Figure 3. Effects of CH4 (a) and CO2 or H2O (b) partial pressure on forward CH4 reaction rate for CH4-CO2 and CH4-H2O reactions on 3.2 wt
% Ru/Al2O3 reduced at 873 K (44.2% Ru dispersion) (5 mg of catalyst, 873 K, total flow rate 100 cm3/min 20 kPa CO2 or H2O in (a) and 10 kPa
CH4 in (b), balance He).

Chemical Conversion of CH4 on Ru-Based Catalysts
J. Phys. Chem. B, Vol. 108, No. 22, 2004 7257
TABLE 2: Forward CH4 Reaction Rate, Rate Constants, and Kinetic Isotope Effects for CH4 Reforming Reactions on 3.2 wt %
Ru/Al2O3 Reduced at 873 K (44.2% metal dispersion) (873 K, 25 kPa CH4 or CD4, 25 KPa CO2 or H2O, balance Ar, total flow
rate 100 cm3/min)
Pre-Exponential Factor
(s-1 kPa-1
)
turnover
rate constant
kinetic isotope
effectb
activation energy
(kJ/mol)
a
co-reactant
rate (s-1
)
(s-1 kPa-1
)
measured
estimatedc
CO2
H2O
none
3.9
4.2
3.8
0.16
0.18
0.15
1.42
1.40
1.51
96
91
99
8.7 × 104
4.7 × 104
8.4 × 104
5.5 × 103
5.5 × 103
5.5 × 103
a Initial CH4 turnover rate on Ru surface. bkCH /kCD . cCalculated on the basis of transition-state theory treatments of CH4 activation steps proceeding
4
4
via an immobile activated complex.32
activation of a C-H bond catalyzed by interactions with Ru
surface atoms.
Figure 4 shows Arrhenius plots for CH4 reforming and
decomposition rate constants. Activation energies for CO2 (96
kJ/mol), H2O reforming (91 kJ/mol) and CH4 decomposition
(99 kJ/mol) are similar (Table 2), consistent with similar
kinetically relevant steps. These activation energies resemble
those reported previously for CO2 reforming on 1.0 wt % Ru/
Al2O3 (92.4 kJ/mol)2 and 1.6 wt % Ru/Al2O3 and 4.8 wt %
Ru/C (106 kJ/mol)3, but they are much larger than reported for
CH4 activation on Ru single crystals36,37 and on Ru/SiO2 at
38
lower temperatures. An activation energy of 51 ( 6 kJ/mol was
reported for CH4 activation on Ru (0001) from the amount of
carbon deposited after various elapsed times.36 Even lower
values (36.1 kJ/mol) were measured by others on similar Ru
(0001) surfaces from electron energy loss measurements of
chemisorbed carbon37 and on Ru/SiO2 (29 kJ/mol) using a pulse
microreactor.38
Figure 4. Arrhenius plots for CO2 reforming (b), H2O reforming (]),
and CH4 decomposition (4) rate constants on 3.2 wt % Ru/Al2O3
reduced at 873 K (44.2% Ru dispersion).
Density functional theory (DFT) led to 85 kJ/mol39 and 78
kJ/mol40 estimates of activation energies for CH4 activation on
Ru (0001) surfaces. These estimates lie between values mea-
sured in the present study for catalytic and stoichiometric CH4
reaction (91-107 kJ/mol) and those reported for stoichiometric
reactions on single crystals and supported clusters at lower
temperatures (29-51 kJ/mol) studies. These differences remain
puzzling and may well reflect the contribution of minority and
catalytically irrelevant coordinatively unsaturated defects, which
do not turn over because of their strong interactions with
chemisorbed carbon formed in C-H activation steps. It appears,
however, that rates and kinetic parameters measured during
steady-state catalysis, and reflecting exclusively CH4 activation
steps, are most relevant to descriptions of catalytic surfaces at
reaction conditions. It is possible that unreactive carbon deposits
form at edge or kink steps in small Ru clusters and Ru single
crystals; these sites would be most effective in stabilizing
transition states required for C-H bond activation. If so, the
resulting carbon species must be entirely unreactive during CH4
reforming reactions, because their surface density (and conse-
quently reaction rates) would otherwise depend on the concen-
tration and identity of co-reactants. They must also form very
rapidly during initial contact with CH4 reactants, in view of the
lack of detectable deactivation at our reaction conditions.
Differences among activation energies measured on catalysts
and on model surfaces were also observed on Rh,25 Pt,26 Ir,27
and Ni28 catalysts. CO oxidation rates measured on Rh,25 Pt,26
and Ir27 before and after reforming reactions were identical,
indicating that the density and type of exposed metal atoms were
unchanged during catalytic reforming reactions. We cannot
exclude that a very small fraction of surface atoms, with
remarkable reactivity in stoichiometric CH4 activation but unable
to turn over, are initially exposed on fresh samples but become
unavailable during initial contact with CH4 reactants. We
conclude, however, that such sites, if present, do not turn over;
Figure 5. CH4 reaction rate for CH4 decomposition and reforming
reactions on 3.2 wt % Ru/Al2O3 catalyst reduced at 873 K (44.2% Ru
dispersion) (873 K, 20 kPa CH4, 100 kPa total pressure, total flow rate
100 cm3/min).
2. Preexponential factors predicted from transition-state theory
treatments of CH4 activation steps proceeding via immobile
activated complexes35 are also shown in Table 1. The measured
values are larger than theoretical estimates, but become similar
if limited mobility is assumed for activated complexes.
CH4-CO2 and CH4-H2O reaction rate constants are also
similar to those measured during the early stages of CH4
decomposition in the absence of either H2O or CO2 co-reactants
on 3.2 wt % Ru/Al2O3 (Figure 5, Table 2). It appears that the
sole kinetically relevant step in catalytic CH4 reactions with
H2O or CO2 to form H2-CO mixtures and in stoichiometric
CH4 decomposition to form C* and H2 on Ru is the initial

7258 J. Phys. Chem. B, Vol. 108, No. 22, 2004
Wei and Iglesia
on supported Ru catalysts. Kinetic isotope effects were measured
from the relative forward rates of CH4-CO2 and CD4-CO2
reactant mixtures at 873 K on 3.2 wt % Ru/Al2O3 (reduced at
873 K; 44.2% dispersion). Kinetic isotope effects for H2O
reforming reactions were obtained from forward reaction rates
with CH4-H2O, CD4-H2O, or CD4-D2O reactant mixtures also
at 873 K.
Normal kinetic isotopic effects were measured for both CH4-
CO2 and CH4-H2O reactions and for CH4 decomposition (Table
2), and their values were identical within experimental accuracy
(1.40-1.51). These similar values are consistent with equivalent
kinetically relevant C-H bond activation steps in all three
reactions. Forward reaction rates were identical for CD4-H2O
and CD4-D2O reactant mixtures, indicating that activation of
co-reactants and any reactions between adsorbed species formed
from co-reactants and CH4 do not influence overall reaction
rates.
Figure 6. Arrhenius plots for CH4 decomposition rate constants on
3.2 wt % Ru/Al2O3 (this study) (4), 2.75 wt % Ru/SiO238 ([), and Ru
(0001)36 (2).
Elmasides and Verykios45 measured a kinetic isotope value
of 1.6 for partial oxidation of CH4-O2 reactant mixtures at 903
K on Ru/TiO2. This value is very similar to those reported here
for H2O and CO2 reforming reactions, suggesting that partial
oxidation, probably occurring via sequential combustion and
reforming reactions, may also be limited by C-H bond
activation. Kinetic isotope effects for H2O reforming, CO2
reforming, or CH4 decomposition reactions on Ru have not been
previously reported. The values reported here are similar to those
reported previously for CH4 decomposition on Ni/SiO2 (1.60)
at 773 K46 and for CO2 reforming on Ni/Al2O3 (1.45) at 873
K,47 as well as to those we recently have reported on Rh (1.54-
1.60),25 Pt (1.58-1.77),26 Ir (1.68-1.75),27 and Ni (1.62-
1.71).28 Another study,48 however, failed to detect a kinetic
isotope effect for CO2 reforming on Ni at near-equilibrium
conversions, which led to the proposal that H2O or CO2 co-
reactant activation and chemisorbed oxygen atoms were in-
volved in rate-determining steps. These latter data are incon-
sistent with the kinetically relevant step and the kinetic isotope
effects reported here on Ru-based catalysts; they appear to reflect
thermodynamic instead of kinetic isotope effects, because these
measurements were made at CH4 conversions near thermody-
namic equilibrium.
Bradford and Vannice2 proposed that CH4-CO2 reactions
proceed on Ru surfaces via reversible CH4 dissociation to form
adsorbed CHx and H species, followed by quasi-equilibrated
steps, in which CO2 adsorbs, dissociates, reacts with CHx and
hydroxyl groups to form CHxO species and H. In this proposal,
CHxO ultimately dissociates to form adsorbed CO and H, which
then desorb to form CO and H2. This mechanism provides
plausible elementary steps for CO2-CH4 reactions, but intro-
duces a level of detail that cannot be experimentally tested,
because these steps become kinetically irrelevant and the
corresponding reactive intermediates are spectroscopically inac-
cessible when their concentrations are low during steady-state
catalysis, as inferred from our kinetic and isotopic studies. We
propose instead a set of elementary steps involving simpler
intermediates, including chemisorbed carbon, because of its
likely formation reactions that lead sequentially from CH4 to
C* with increasing rate as H-atoms are sequentially abstracted
from CH4.49
thus, they are not relevant to the analysis and prediction of
catalytic rates of CH4-H2O and CH4-CO2 reactions. Indeed,
Wang et al.41 examined CH4 decomposition on stepped Pd (679)
surfaces and showed that C* preferentially forms at steps and
kinks, while (111) terraces remain largely uncovered. On Pt
surfaces, carbon also forms preferentially at step sites during
n-hexane reactions, while terrace sites remain uncovered and
active for catalytic reactions.42 It is not certain that these findings
are relevant to Ru surfaces and to the steady-state behavior of
catalytic Ru clusters.
Figure 6 shows Arrhenius plots for CH4 decomposition data
on 3.2 wt % Ru/Al2O3, from the present study, together with
CH4 decomposition rates on Ru (0001)36 and 2.75 wt % Ru/
38
SiO2 at lower temperatures. Measured CH4 decomposition
rates on Ru (0001)36 are ∼100 times larger than on Ru/Al2O3,
while CH4 decomposition rates are similar on 3.2 wt % Ru/
Al2O3 and 2.75 wt % Ru/SiO2.38 Thus, it appears that the type
of defect sites responsible for CH4 activation on large single
crystals are not available on small Ru metal clusters, even
though surfaces of small clusters are often described as quite
rough and densely populated by coordinatively unsaturated
sites. One possibility is that this description becomes inap-
propriate at high temperatures, because of the tendency of small
metal clusters to melt, at least in near-surface regions, well below
the melting temperature of the corresponding bulk metal.43,44
This process leads to a liquidlike layer a few atoms thick
stabilized by a crystalline metal core, which becomes appar-
ent above 800 K for Pt clusters about 8 nm in diameter, even
though bulk Pt melts at 2042 K.43,44 Smaller clusters and the
presence of a support to stabilize molten structures would favor
these phenomena and lead to significant loss of coordinative
unsaturation for the supported Ru clusters of this study (bulk
Ru melts at 2607 K). Thus, catalytic reactions at high temper-
atures, such as CH4 reforming, on small metal clusters may be
unable to benefit from coordinative unsaturation, even when
such unsaturated sites were able to undergo a catalytic turnover.
Although this explanation remains speculative at this point, it
deserves additional examination in view of its marked con-
sequences on the choice of models systems that describe
faithfully the relevant features of small clusters of catalytic
relevance.
These elementary steps are shown as Scheme 1. CH4
decomposes to C* in a series of elementary H-abstraction steps,
with the first abstraction as the kinetically relevant step because
of the low prevalent concentration of all CHx* intermediates.
This step is followed by the removal of the fragments formed
using CO2 or H2O co-reactants.
Mechanistic Evidence from Kinetic Isotope Effects. Several
steady-state isotopic tracer studies and kinetic isotope effect
(KIE) measurements were used to probe the role and revers-
ibility of specific elementary steps involved in CH4-H2O and
CH4-CO2, as well as the mechanistic relevance of co-reactants

Chemical Conversion of CH4 on Ru-Based Catalysts
J. Phys. Chem. B, Vol. 108, No. 22, 2004 7259
SCHEME 1: Sequence of Elementary Steps for CH4
Reforming Reactions on Ru-Based Catalysts
Figure 7. Methane reaction rate and CH4/CD4 cross-exchange rates
during the reaction of CH4/CD4/CO2 mixture on 3.2 wt % Ru/Al2O3
catalyst (5 mg of catalyst, 873 K, 12.5 kPa CH4 and CD4, 25 kPa CO2,
balance Ar, total flow rate 80 cm3/min).
for the overall reaction, which rigorously requires that the rate-
determining step become exactly as reversible as the overall
reaction. C-H bond activation steps on Ru crystallites at 873
K are irreversible, except as required by the approach of the
overall reaction to thermodynamic equilibrium.
The H/D ratio in the dihydrogen formed from equimolar
CH4-CD4 mixtures was greater than one (1.50), and dihydrogen
molecules show a binomial isotopomer distribution. This H/D
ratio reflects the higher reactivity of CH4 relative to CD4, as
shown from independent reaction rates for these two methane
isotopomers (1.42). The binomial distribution of dihydrogen
isotopomers indicates that recombinative hydrogen desorption
steps are quasi-equilibrated during CH4-CO2 reactions on Ru
at 873 K.
In this scheme, f denotes an irreversible step, and § a quasi-
equilibrated step, and ki is the rate coefficient and Ki the
equilibrium constant for step i. CH4 irreversibly decomposes
in a sequence of elementary steps to form chemisorbed carbon
and hydrogen atoms. When (*) is the most abundant surface
intermediate, only the rate constant for step (8) appears in the
rate expression and the rate becomes proportional to CH4 and
independent of the presence or concentration of CO2 or H2O
co-reactants. Steps (12), and (14)-(19) are assumed to be
reversible and quasi-equilibrated. We note that these elementary
steps provide pathways for reactions of CH4 with either CO2
or H2O and also for water-gas shift reactions, which have been
frequently, but inappropriately and nonrigorously, treated as a
separate independent kinetic process in many previous studies
of CH4 reforming reactions.
Isotopic Tracer and Exchange Evidence for the Revers-
ibility of Specific Elementary Steps. The reversibility of some
of the elementary steps shown in Scheme 1 was probed using
isotopic tracer and exchange methods. The reversibility of CH4
activation steps was determined from measurements of the rate
of formation of CHxD4-x (0 < x < 4) isotopomers during
chemical conversion of CH4/CD4/CO2 mixtures. A CH4/CD4/
CO2 (1:1:2) mixture was allowed to react at 873 K on 3.2 wt
% Ru/Al2O3 (treated at 873 K; 44.2% dispersion). Chemical
conversion and isotopic scrambling rates were measured using
on-line mass spectrometry after removing HxD2-xO in a trap
held at 218 K (to avoid interference between fragments for H2O,
HDO, and D2O and for CHxD4-x). The rates of CHxD4-x (0 <
x < 4) formation and of methane chemical conversion are shown
in Figure 7. The CH4/CD4 cross-exchange turnover rate, defined
as the sum of the rates of formation of CHD3, CH2D2 (taken
twice), and CH3D, is 0.18 s-1. The turnover rate for CH4
chemical conversion (3.3 s-1) was about 19 times greater than
for isotopic cross-exchange. The approach to equilibrium for
this reaction, η, estimated from the prevalent concentrations of
all reactants and products is 0.04, corresponding to the ratio of
the forward overall reaction rate to the reverse reaction rate of
25, indicating that formation of traces of CHD3, CH2D2, and
CH3D isotopomers is merely due to some slight reversibility
Reactions of CH4/CO2/D2 (1:1:0.2) mixtures at 873 K on 3.2
wt % Ru/Al2O3 were used to probe the reversibility of
elementary steps leading to the formation of water and dihy-
drogen. Here, water was not removed before mass spectrometric
analysis, and all transfer lines were kept above 373 K to prevent
water condensation. No deuterated methane isotopomers were
detected, as expected from the irreversible nature of C-H bond
activation steps; as a result, water isotopomer measurements
were unaffected by mass fragments from deuterated methane
molecules. The H/D fraction expected if all H-atoms in the
converted CH4 molecules and all D-atoms in the inlet D2 stream
contributed to surface intermediates is 0.76. The water molecules
formed during reaction and the dihydrogen molecules in the
effluent stream both contained identical H/D ratios of 0.74. Thus,
dihydrogen and water molecules and their corresponding
precursors in the chemisorbed phase are in quasi-equilibrium.
Table 3 shows the isotopomer distribution in water molecules
formed from CH4/CO2/D2 reactant mixtures. The isotopomer
distribution is binomial with a D-content identical to that in
the available reactant pool; this is consistent with fast and quasi-
equilibrated recombination of H* and OH* in step (18) in
Scheme 1. Binomial distributions were also observed for
dihydrogen isotopomers, as expected from reversible and quasi-
equilibrated recombinative hydrogen desorption steps (step (15)
and (16) in Scheme 1) during CH4/CO2 reactions on Ru-based
catalysts. In view of the kinetic equivalence of elementary steps
involved in CH4 reactions with CO2 and H2O, we consider these
conclusions to be also valid for CH4-H2O reactions on Ru-
based catalysts.
The reversibility of CO2 activation steps (step (12) in Scheme
1) was probed using 12CH4/12CO2/13CO (1:1:0.4) reactant

7260 J. Phys. Chem. B, Vol. 108, No. 22, 2004
Wei and Iglesia
TABLE 3: Distribution of Water Isotopomers during
Reactions of CH4/CO2/D2 Mixtures on 3.2 wt % Ru/Al2O3
Reduced at 873 K (873 K, 16.7 kPa CH4, 16.7 kPa CO2, 3.3
kPa D2, balance Ar, total flow rate 150 cm3/min)
Distribution (%)
measured
(H/D ) 0.74)
binomial
isotopomer
(H/Da ) 0.76)
H2O
HDO
D2O
0.19
0.48
0.33
0.19
0.49
0.32
a
(H/D) ratio predicted from H in reacted methane and D2 in ambient
stream if complete mixing between the two isotopes occurred during
reaction.
mixtures at 873 K on 3.2 wt % Ru/Al2O3. The 13C fraction in
CO (0.275) and CO2 (0.256) molecules in the effluent are similar
to each other at all reaction temperatures, even though the inlet
reactant mixture contained isotopically pure 12CO2 and 13CO.
The 13C content corresponds to complete chemical and isotopic
equilibration between CO and CO2, even at CH4 chemical
conversion levels (6.4%) far from equilibrium (η ) 0.06). These
data indicate that CO2 activation steps are much faster than
kinetically relevant CH4 dissociation steps, and that steps (12)
and (14) occur in both directions many times in the time required
for a CH4 chemical conversion turnover. Thus, CO2 activation
steps are reversible and quasi-equilibrated during CH4-CO2
reactions and by kinetic analogy also during CH4-H2O reactions
at similar reaction conditions. Only trace amounts of 13CH4 were
detected during reactions of 12CH4/12CO2/13CO mixtures, as
expected from the low value of η (0.06) during these experi-
ments, which lead to essentially irreversible conversion of CH4
to H2 and CO.
Steps (12) and (14)-(19) in Scheme 1 describe the reverse
water-gas shift reaction, which must, in view of the equilibrated
nature of its component steps, also be quasi-equilibrated during
CH4 reforming reactions on Ru-based catalyst (Figure 2). Indeed,
we find that the approach to equilibrium for this reaction,
estimated from the prevalent concentrations of all reactants and
products, is near unity at all conditions (Figure 2).
These isotopic studies were carried out on a 3.2 wt % Ru/
Al2O3 and predominately at 873 K, but the similar rate
expressions obtained at all temperatures and on all catalysts do
not indicate any mechanistic shifts with changes in reaction
temperature or catalyst composition and thus support the general
relevance of the proposed elementary steps. We note that these
elementary steps also provide a rigorous basis for kinetic
treatments of carbon filament formation during CH4 reforming
reactions by defining a concentration or thermodynamic activity
for chemisorbed carbon as a function of prevalent pressures and
of rate constants and equilibrium constants for elementary steps,
as we discuss in detail elsewhere.25-28
Figure 8. Forward CH4 reaction rate for CO2 reforming of CH4 vs
metal dispersion on different Ru-based catalysts (873 K, 20 kPa CH4,
([)1.0 wt % Ru/γ-Al2O3 (ref 3), (C) 1.0 wt % Ru/NaY (ref 18), (4)3.2
wt % Ru/ZrO2 (this study), (b) 3.2 wt % Ru/γ-Al2O3 reduced at 1023
K (this study), (O) 1.6, 3.2 wt % Ru/γ-Al2O3 reduced at 873 K, (*)
0.64 wt % Ru/γ-Al2O3 (ref (12)).
al.18 and Ferreria-Aparicio et al.12 did not report activation
energies; therefore, we have used our value of the activation
energy (96 kJ/mol) in extrapolating their data to our reaction
conditions. Forward CH4 reaction rates were reported by
Bradford et al.2 and Portugal et al.,18 but the net rates reported
by Solymosi et al.3 and Ferreria-Aparicio et al.12 were converted
to forward rates, using η values of 0.05 and 0.1, respectively,
based on their respective CH4 conversion levels.
Identical dispersion effects and turnover rate values were
obtained for CH4-H2O, CH4-CO2, and CH4 decomposition
reaction rates (Table 1), as expected from their rigorous
mechanistic equivalence. Turnover rates increased monoton-
ically with increasing Ru dispersion for both reactions, sug-
gesting that coordinatively unsaturated Ru surface atoms,
likely to be prevalent in small crystallites even if they become
mobile at high temperatures, are more active than those in low-
index planes predominately exposed on large Ru crystallites.
Edge and corner atoms, with fewer Ru neighbors than those on
terraces, bind CHx and H more strongly and apparently decrease
the energy required to form the relevant transition state for C-H
bond activation.36 No previous literature reported systematic Ru
dispersion effects on CH4 reaction rates, but similar effects of
coordinative unsaturation were previously reported on other
metal surfaces.24-28,50,51. Klier et al.24 found that CH4 dis-
sociation rates on a Pd single crystal increased with increas-
ing density of steps and kinks. These coordinative unsaturated
surface atoms showed reaction rates 10-100 times larger than
on hexagonal closed-packed Pd(111) surfaces. Johnson and
Weinberg50 reported that defects sites on Ir surfaces were much
more active than terrace sites for alkane dissociation react-
ions. Molecular beam studies by Weaver et al.51 showed that
surface defects in Pt(111) markedly increased alkane dissociation
rates.
Dispersion and Support Effects on H2O and CO2 Reform-
ing on Ru. Ru clusters with a range of dispersion were prepared
by varying the metal content, the reduction temperature, and
the identity of the support. Turnover rates were calculated from
forward rates normalized by the number of exposed Ru atoms;
they are shown for CH4-CO2, CH4-H2O, and CH4 decomposi-
tion reactions in Table 1. CH4-CO2 turnover rates reported by
Bradford and Vannice,2 Ferreria-Aparicio et al.,12 Portugal et
al.,18 and Solymosi et al.3 on Ru-based catalysts are also shown
in Table 1. We have corrected these literature rates for approach
to equilibrium and extrapolated to our reaction conditions (873
K, 20 kPa CH4) using a first-order CH4 dependence and the
activation energies reported in each literature report (92.5 kJ/
mol, Solymosi et al.3; 106 kJ/mol, Bradford et al.2). Portugal et
The identity of the support did not directly influence turnover
rates (Table 1, Figure 8), but it can influence Ru dispersion
and, in this manner, also turnover rates. Matsui et al.9 have
reported much higher CH4 conversions on Ru/Y2O3 and Ru/
ZrO2 (25-29%) than on Ru/SiO2 (12%) and attributed these
effects to CO2 activation on the supports, but neither Ru
dispersions nor turnover rates were reported. These support
effects are inconsistent with our findings on Al2O3 and ZrO2
supports and with the kinetic irrelevance of co-reactant activation
during CH4 reforming reactions on Ru-based catalysts.

Chemical Conversion of CH4 on Ru-Based Catalysts
J. Phys. Chem. B, Vol. 108, No. 22, 2004 7261
Figure 8 shows that literature turnover rates show a consistent
effect of dispersion, irrespective of the identity of the support,
except for the data Bradford et al.2 on Ru/η-Al2O3, which
showed higher turnover rates and Ru/C, which showed lower
turnover rate. Once Ru dispersions are used to normalize
reaction rates, turnover rates increased monotonically with
increasing Ru dispersion. These dispersion effects, and the
measured turnover rates, are identical for CH4 reactions with
either H2O or CO2 co-reactants and also for CH4 decomposition
on supported Ru catalysts (Table 1).
Conclusions
Isotopic studies and forward reaction rate measurements led
to a simple mechanistic picture and to a unifying kinetic
treatment of CH4-CO2, CH4-H2O, and CH4 decomposition
reactions and of water-gas shift on Ru-based catalysts. CH4
reactions are limited by C-H bond activation and unaffected
by the identity or concentration of co-reactants or of the presence
of reaction products. Turnover rates were identical for CH4
decomposition, CO2 reforming, and H2O reforming reactions,
and activation energies were similar for the latter two reactions.
The kinetic relevance of C-H bond activation was confirmed
by kinetic isotope effect measurements; isotope effects were
identical for CH4-CO2 and CH4-H2O reactions and for CH4
decomposition. Cross-exchange rates are much smaller than
chemical conversion rates for CH4/CD4/CO2 mixtures and
indicate that C-H bond activation is exactly as reversible as
the overall chemical reaction. Reactions of the CH4/CO2/D2
mixture led to binomial isotopomer distributions of water and
dihydrogen and to D-contents identical to those expected from
quasi-equilibrated water and dihydrogen desorption steps.
12CH4/12CO2/13CO mixtures led to identical 13C contents in CO
and CO2, consistent with equilibrated CO2 dissociation steps.
These results demand that the water-gas shift reaction be at
thermodynamic equilibrium during CO2 and H2O reforming
reactions on Ru-based catalysts, as indeed found from the
chemical composition of the reactor effluent at all reaction
conditions.
Figure 9. Net CH4 turnover rates versus residence time for CH4-
H2O reaction on 3.2 wt % Ru/Al2O3 at 873 K((b) 10 mg of catalyst
diluted with 100 mg of Al2O3 within pellets, then diluted with 500 mg
of ground quartz, pellet size 250-425 µm; (∆) 10 mg of catalyst diluted
with 50 mg of Al2O3 within pellets, then diluted with 500 mg of ground
quartz, pellet size 250-425 µm; (2) 10 mg of catalyst diluted with 50
mg of Al2O3 within pellets, then diluted with 500 mg of ground quartz,
pellet size 63-106 µm).
transport artifacts. Extrapolating the net reaction rate to zero
residence time gives forward CH4 reaction rates.
References and Notes
(1) Fischer, V. F.; Tropsch, H. Brennstoff-Chemie 1928, 3, 39.
(2) Bradford, M. C. J.; Vannice, M. A. J. Catal. 1999, 183, 69.
(3) Solymosi, F.; Kustan, G.; Erohelyi, A. Catal. Lett. 1991, 11, 149.
(4) Claridge, J. B.; Green, M. L. H.; Tsing S. C. Catal. Today 1994,
12, 455.
(5) Vernon, P. D. F.; Green, M. L. H.; Cheetham, A. K.; Ashcroft, A.
T. Catal. Today 1992, 13, 417.
(6) Ashcroft, A. T.; Cheetham, A. K.; Green, M. L. H.; Vernon, P. D.
F. Nature 1991, 352, 225.
(7) Basini, L.; Sanfilippo, D. J. Catal. 1995, 157, 162.
(8) Matsui, N.; Nakagawa, K.; Ikenaga, N.; Suzuki, T. J. Catal. 2000,
194, 115.
(9) Matsui, N.; Anzai, K.; Akamatsu, N.; Nakagawa, K.; Ikenaga, N.;
Suzuki, T. Appl. Catal. A: General 1999, 179, 247.
(10) Mark, F. M.; Maier, W. F. J. Catal. 1996, 164, 122.
(11) Ferreira-Aparicio, P.; Rodriguez-Ramos, I.; Anderson, J. A.;
Guerrero-Ruiz, A. Appl. Catal. A 2000, 202, 183.
(12) Ferreira-Aparicio, P.; Marquez-Alvarez, C.; Rodriguez-Ramos, I.;
Schuurman, Y.; Guerrero-Ruiz, A.; Mirodatos, C. J. Catal. 1999, 184, 202.
(13) Sutton, D.; Parle, S. M.; Ross, J. R. H. Fuel Process. Technol.
2002, 75, 45.
Forward turnover rates for both CO2 and H2O reforming
increased monotonically with increasing Ru dispersion, sug-
gesting that the coordinatively unsaturated surface atoms
prevalent in small crystallites are significantly more active than
those in the low-index planes predominately exposed on large
crystallites. No effects of supports, beyond their influence on
Ru dispersion, were detected, as expected from the kinetic
irrelevance of co-reactant activation steps, which have been
previously and nonrigorously claimed to occur on support sites.
(14) Bradford, M. C. J.; Vannice, M. A. Catal. Today 1999, 50, 87.
(15) Qin, D.; Lapszewicz, J. Catal. Today 1994, 21, 551.
(16) Rostrup-Nielsen, J. R.; Hansen, J. H. B. J. Catal. 1993, 144, 38.
(17) Guerrero-Ruiz, A.; Ferreira-Aparicio, P.; Bachiller-Baeza, M. B.;
Rodriguez-Ramos, I. Catal. Today 1998, 46, 99.
(18) Portugal, U. L.; Marques, C. M. P.; Araujo, E. C. C.; Morales, E.
V.; Giotto, M. V.; Bueno, J. M. C. Appl. Catal. A 2000, 193, 173.
(19) Schuurman, Y.; Mirodatos, C.; Ferreira-Aparicio, P.; Rodriguez-
Ramos, I.; Guerrero-Ruiz, A. Catal. Lett. 2000, 66, 33.
(20) Pitchai, R.; Klier, K. Catal. ReV.-Sci. Eng. 1986, 28, 13.
(21) Boudart, M. AdV. Catal. 1969, 20, 153.
(22) McMaster, M. C.; Madix, R. J. J. Chem. Phys. 1993, 98, 15.
(23) Gee, A. T.; Hayden, B. E.; Mormiche, C.; Kleyn, A. W.; Riedmuller,
B. J. Chem. Phys. 2003, 118, 3334.
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) are
gratefully acknowledged.
(24) Klier, K.; Hess, J. S.; Herman, R. G. J. Chem. Phys. 1997, 107,
4033.
Appendix
(25) Wei, J.; Iglesia, E. J. Catal., in press.
The effects of the catalyst pellet size and the extent of dilution
within the catalyst particles were studied on 3.2 wt % Ru/Al2O3
(reduced at 873 K, Ru dispersion 44.2%) for H2O/CH4 reactions
at 873 K. The results are shown in Figure 9. Varying the
diameter of catalyst pellets (250-425 vs 63-106 µm) or the
extent of dilution within the pellets (5:1 to 10:1) did not
influence net CH4 reaction rates, indicating that measured net
rates are not affected in any way by intrapellet or interpellet
(26) Wei, J.; Iglesia, E., J. Phys. Chem. B 2004, 108, 4094.
(27) Wei, J.; Iglesia, E. Angew Chem. Int. Ed., in press.
(28) Wei, J.; Iglesia, E., J. Catal., in press.
(29) Bahlawane, N.; Watanabe, T. J. Am. Ceram. Soc. 2000, 83, 2324.
(30) Barton, D. G. Ph.D. dissertation, University of California at
Berkeley, 1998.
(31) Mazzieri, V.; Coloma-Pascual, F.; Arcoya, A.; Largentie`re, P. C.;
F´ıgoli, N. S. Appl. Surf. Sci. 2003, 210, 222.
(32) Price, G. L.; Iglesia, E. Ind. Eng. Chem. 1989, 28, 839.

7262 J. Phys. Chem. B, Vol. 108, No. 22, 2004
Wei and Iglesia
(33) Stull, D. R.; Edgar, F.; Westrum, J.; Sinke, G. C. The Thermody-
namics of Organic Compounds; Robert, E., Ed.; Krieger Publishing Co.:
Malabar, FL, 1987.
(42) Davis, S. M.; Zaera, F.; Somorjai, G. A. J. Catal. 1982, 77, 439.
(43) Wang, Z. L.; Ahmad, T. S.; El-Sayed, M. A. Surf. Sci. 1997, 380,
302.
(34) Boudart, M.; Djega-Mariadassou, G. The Kinetics of Heterogeneous
Catalytic Reactions; Princeton University Press: Princeton, NJ, 1984.
(35) Dumesic, J. A.; Rudd, D. F.; Rekoske, J. E.; Trevino, A. A. The
Microkinetics of Heterogeneous Catalysis; American Chemical Society:
Washington, DC, 1993.
(36) Egeberg, R. C.; Ullmann, S.; Alstrup, I.; Mullins, C. B.; Chork-
endorff, I. Surf. Sci. 2002, 497, 183.
(37) Wu, M. C.; Goodman, D. W. Surf. Sci. 1994, 306, L529.
(38) Carstens, J. N.; Bell, A. T. J. Catal. 1996, 161, 423.
(39) Ciobica, I. M.; Frechard, F.; van Santen, R. A.; Kleyn, A. W.;
Hafner, J. J. Phys. Chem. B 2000, 104, 3364.
(40) Liu, J. P.; Hu, P. J. Am. Chem. Soc. 2003, 125, 1958.
(41) Wang, Y.; Herman, R. G.; Klier, K. Surf. Sci. 1992, 279, 33.
(44) Wang, Z. L.; Petroski, J. M.; Green, T. C.; El-Sayed, M. A. J.
Phys. Chem. B 1998, 102, 6145.
(45) Elmasides, C.; Verykios, X. E. J. Catal. 2001, 203, 477.
(46) Otsuka, K.; Kobayashi, S.; Takenaka, S. J. Catal. 2001, 200, 4.
(47) Osaki, T.; Horiuchi, T.; Suzuki, K.; Mori, T. Catal. Lett. 1997, 44,
19.
(48) Schuurman, Y.; Kroll, V. C. H.; Ferreira-Aparicio, P.; Mirodatos,
C. Catal. Today 1997, 38, 129.
(49) Au, C.; Ng, C.; Liao, M. J. Catal. 1999, 185, 12.
(50) Johnson, D. F.; Weinberg, W. H. J. Chem. Phys. 1994, 101, 6289.
(51) Weaver, J. F.; Krzyzowski, M. A.; Madix, R. J. Surf. Sci. 1997,
391, 150.