Partial oxidation of methane to synthesis gas over Ru/TiO2 catalysts: Effects of modification of the support on oxidation state and catalytic performance

Article abstract of DOI:10.1006/jcat.2000.3120

The effects of modification of the support on the oxidation state of Ru and the catalytic performance of Ru/TiO₂ catalysts under conditions of partial oxidation of methane to synthesis gas have been investigated employing XPS and FTIR technique

Full text of DOI:10.1006/jcat.2000.3120

Journal of Catalysis 198, 195–207 (2001)
doi:10.1006/jcat.2000.3120, available online at http://www.idealibrary.com on
Partial Oxidation of Methane to Synthesis Gas over Ru/TiO Catalysts:
2
Effects of Modification of the Support on Oxidation State
and Catalytic Performance
�
�
�
,1
C. Elmasides, D. I. Kondarides, S. G. Neophytides,† and X. E. Verykios
�
Department of Chemical Engineering, University of Patras, GR-26500 Patras, Greece; and †Institute of Chemical Engineering and High Temperature
Chemical Processes (ICE/HT-FORTH), P.O. Box 1414, GR-26500 Patras, Greece
Received June 5, 2000; revised November 23, 2000; accepted November 23, 2000; published online February 13, 2001
within catalyst beds, which can irreversibly damage the ac-
The effects of modification of the support on the oxidation state of tive ingredient of the catalyst or even the reactor. Hence,
Ru and the catalytic performance of Ru/TiO2 catalysts under condi- knowledge of the reaction mechanism and, especially, un-
tions of partial oxidation of methane to synthesis gas have been in-
vestigated employing XPS and FTIR techniques. It has been found
that the oxidation state of Ru depends on the reaction temperature
and on the supporting material and that, under conditions where
the direct reaction route is operable, the catalytic performance (con-
methane, two alternative routes have been proposed: (a) a
version of methane and selectivity toward synthesis gas formation)
derstanding of the nature of active surface sites that par-
ticipate in reaction steps can lead to the development of
appropriate catalysts for industrial usage.
Concerning the reaction pathway of partial oxidation of
sequential scheme according to which the initial total ox-
idation of methane is followed by reformation of the un-
converted methane with CO2 and H2O (indirect scheme)
is improved over samples that mainly contain Ru in its metallic
6
+
form. Doping of TiO2 with small amounts of W cations favors
oxygen adsorption on Ru under reaction conditions, thus resulting
in stabilization of a fraction of the catalyst in its oxide forms. This is and (b) the direct conversion of CH4 to synthesis gas (H2
reflected in lower activity and selectivity of this catalyst compared and CO), which does not involve the formation of CO2 and
to those of the undoped one. In contrast, Ru supported on undoped H O as reaction intermediates (direct scheme).
2
and Ca2 -doped TiO2 exists mainly in the metallic form and, as
a result, it promotes the direct conversion of methane to synthesis
gas. In situ FTIR spectroscopy showed that an adsorbed CO species
exists on the surface of the latter two catalysts under reaction con-
ditions, even at temperatures as high as 1073 K. This species, which
+
The catalytic partial oxidation of methane has been stud-
ied on noble metals, such as Rh, Ru, Pd, and Pt in unsup-
ported (2–5) or supported form (6–8) and on supported Ni
catalysts (9, 10). It is generally agreed that the type of cat-
alyst employed may strongly influence the reaction steps
of the catalytic partial oxidation of methane. Mallens et al.
6
+
is not observable over the W -doped catalyst, is related to a Ru
surface site, which is responsible for the dissociation of methane
and formation of CO.
(
2) have studied the partial oxidation of methane to syn-
�c 2001 Academic Press
thesis gas over Rh and Pt catalysts and found differences in
the selectivity toward CO and H2. These differences were
attributed to the lower activation energy of methane de-
composition on Rh than on Pt. The authors support the
view that the ability of the catalyst to activate methane de-
termines the concentration of active surface species such as
oxygen, carbon, and hydrogen and, consequently, product
distribution. Fathi et al. (3) have also studied Pt catalysts for
the partial oxidation of methane and proposed that product
distribution is determined by the kind and concentration of
surface oxygen species. For supported metal catalysts, the
choice of the support may also influence the concentration
of adsorbed oxygen and, consequently, the activation of
methane and product distribution (6).
1
. INTRODUCTION
Several attempts have been made in recent years to
achieve synthesis gas formation via catalytic partial oxida-
tion of methane (CPO) (1). Compared to steam reform-
ing, which is the dominant commercial method employed
to produce synthesis gas, the direct oxidation of methane
is more energy efficient and can produce a more desirable
H2/CO ratio, required for methanol or Fischer–Tropsch syn-
thesis (1). However, the direct oxidation reaction has not
been developed at industrial scale because it involves co-
feeding of CH4 and O2 under flammable or even explo-
sive conditions. Moreover, local hot spots may be formed
Previous investigations in this laboratory have shown
that Ru supported on TiO2 exhibits unique performance by
catalyzing the direct reaction scheme, to an appreciable ex-
tent, at relatively low reaction temperatures (11, 12). More
1
To whom correspondence should be addressed. Fax: +30-61-991 527.
E-mail: verykios@chemeng.upatras.gr.
195
0021-9517/01 $35.00
Copyright �c 2001 by Academic Press
All rights of reproduction in any form reserved.

1
96
ELMASIDES ET AL.
specifically, it has been found that, in the absence of sig- mixtures consisting of CH4/O2/N2 (6/3/91 vol% ). The reac-
nificant mass and heat transfer resistances, high selectivity tor employed consists of a quartz tube (6-mm O.D.), with
to synthesis gas (>60% ) is obtained over Ru/TiO2 catalysts an enlargement (8 mm O.D.) at its central part where the
in the low methane conversion range (oxygen conversion catalyst is placed. The catalyst bed is supported by means
<
100% ) at temperatures as low as 873–973 K. On the other of quartz wool. The quantity of catalyst used in these exper-
hand, very low selectivity to synthesis gas in the presence of iments varied between 10 and 30 mg and the particle size
oxygen is obtained, over other metal catalysts (Ni, Rh, Pd was in the range of 0.12–0.18 mm. A K-type thermocouple
and Ir) as well as over Ru catalysts supported on carriers enclosed in a quartz well (3-mm o.d.) was positioned inside
other than TiO2 (12). This has been attributed to the abil- the catalyst bed for accurate measurement of the catalyst
ity of the TiO2 carrier to stabilize Ru in its reduced state, bed temperature. The total feed flow rate was in the range
3
thus inhibiting oxidation of the catalyst under conditions of 400–1000 cm (STP)/min.
of partial oxidation of methane (13). In the present study,
Two gas chomatographs were used for the analysis of re-
2+
the effects of modification of TiO2 by doping with Ca or actants and products. CH4 and CO2 were separated in a
6+
W
cations on the oxidation state of supported Ru has Carboxen 1000 column and O2, N2, and CO in a molecu-
been investigated by XPS and FTIR. The results obtained lar sieve 5A column with He as the carrier gas. Hydrogen
are related to the catalytic performance of Ru supported on was analyzed in the second chromatograph, employing N2
modified TiO2 carriers under conditions of partial oxidation as the carrier gas. H2O produced during the reaction was
of methane to synthesis gas.
condensed downstream of the reactor.
2
.3. XPS Experiments
2
. EXPERIMENTAL PROCEDURES
XPS experiments were carried out with a SPECS LHS
0 spectrometer, employing the Mg K� exciting radiation
TiO2 carriers doped with 0.95 at.% Ca2+ or 0.68 at.% W6+ (1253.6 eV). The X-ray source was operated at 15 kV–20
2
.1. Catalyst Preparation
1
were prepared bythe method ofhigh-temperature diffusion mA and the kinetic energy of the photoelectrons was mea-
of the dopant cation into the crystal matrix of titania. The sured by the hemispherical electron analyzer in the pass-
parent TiO2 carrier used in the present study was Degussa energy mode (36 eV). The sample, which was mounted on
2
P25 (specific surface area: 50 � 2 m /g) and the precursors a transfer rod probe, could be easily transferred from the
used for the preparation of the W6 - and Ca -doped car- atmospheric pressure chamber, where it could be treated
riers were WO3 and CaO, respectively, obtained from Alfa with gases at atmospheric pressure and elevated tempera-
Products. In the final step of preparation of the doped car- tures (up to 973 K), to the UHV chamber. During pretreat-
riers, materials were calcined in air at 1173 K for 5 h. For ment of the catalyst sample under reaction gas mixture,
comparison purposes, and to produce materials with similar on-line analysis of the effluent gas was possible by means
physical characteristics, undoped TiO2 was calcined at 973 of a mass spectrometer (Omnistar) connected at the exit of
K for 5 h. The specific surface area of the doped and un- the pretreatment chamber.
+
2+
doped carriers was measured with the BET technique, em-
ploying nitrogen physisorption at the temperature of liquid gions of Ti(2p), Ru(3d), C(1s), and O(1s) lines. Since the
nitrogen.
C(1s) signal is superimposed with the Ru(3d) signal, the
XP spectra were obtained in the binding energy (BE) re-
Ru catalysts, with a metal loading of 0.5 wt% , were pre- main line of the support (Ti(2p) at 458.7 eV) was taken
pared by employing the incipient wetness impregnation as a secondary reference. Signal shape analysis for the
method, using Ru(NO)(NO3)3 (Alfa Products) as the start- Ru(3d)/C(1s) region wasperformed with the program pack-
ing material. After preparation, all catalysts were reduced age Xpspeak. Binding energies and full width at half maxi-
in flowing H2 at 573 K for 2 h and then stored in sealed mum (FWHM) were kept constant in a narrow range of ad-
vials until further use. Catalysts were characterized in terms mitted deviations. The Ru(3d) spin–orbit coupling energy
of metal dispersion, employing hydrogen chemisorption at (difference between the BE of Ru(3d3/2) and Ru(3d5/2)) was
room temperature. Monolayer coverages were estimated kept at 4.10 � 0.05, while the area ratio of these components
by extrapolation of the linear part of the H2 chemisorption was kept at 0.65 � 0.05 (13). Atomic ratios of elements were
isotherms to zero pressure.
calculated by integration of the XPS signals in preselected
BE windows, employing a Shirley-type background.
2
.2. Catalytic Performance Tests
2
.4. FTIR Spectroscopy
The catalytic performance of supported Ru catalysts for
the partial oxidation of methane to synthesis gas was stud-
A Nicolet 740 FTIR spectrometer equipped with a
ied in a continuous flow reactor in the temperature range of DRIFT cell, an MCT detector, and a KBr beamsplitter was
73–1073 K. Experiments were conducted using dilute feed used in the present study. For all spectra, a 64-scan data
9

METHANE TO SYNTHESIS GAS OVER Ru/TiO2 CATALYSTS
197
acquisition was carried out at a resolution of 4.0 cm 1. The
apparatusand proceduresfor obtainingin situ FTIR spectra
have been described in detail elsewhere (13).
�
CO adsorption was studied at room temperature (RT)
over samples pretreated at selected conditions, i.e., after
treatment with oxygen, hydrogen, or a CH4/O2/Ar mixture.
In situ FTIR spectra were also recorded at reaction condi-
tions in the temperature range of 773 to 1073 K. In these
experiments, each sample was first reduced in flowing hy-
drogen (20% H2 in Ar) at 773 K for 60 min and then the
gas phase was purged by switching to Ar flow, at 773 K, for
1
0 min. After the purging process, the flow was switched
to the reaction mixture and spectra were recorded after
ca. 30 min on stream when steady state conditions were
reached. Subsequently, the sample was heated at a higher
reaction temperature under flow of the reaction mixture,
and the next spectrum was recorded when the new reaction
temperature had stabilized.
3
. RESULTS
3
.1. Catalysts Characterization
The morphological characteristics of the catalysts exam-
FIG. 1. Selectivities toward CO (SCO) and H2 (SH ) formation, ob-
2
tained over the Ru catalysts, as functions of methane conversion (XCH ).
ined are summarized in Table 1. It is observed that the high-
temperature pretreatment of the doped and undoped car-
riers results in reduction of the specific surface area of the
4
T = 1073 K; feed composition: 6% CH4–3% O2 (in N2).
ties toward both products increase with increasing methane
conversion. It is important to note that, in the first region
2
parent TiO2 material from 50 to 11–15 m /g and in complete
transformation of the oxide to its rutile form. According to
the H2 chemisorption results, the dispersion of Ru is some-
what higher over the doped samples than over the undoped
one, and the average crystallite size is ca. 2 nm.
(
XCH < 35% ), analysis of the gas at the effluent of the reac-
4
tor shows that oxygen is not fully consumed (XO < 100% )
2
and it varies between 10 and 90% . In this region, selectiv-
ities toward both CO and H2 depend strongly on the na-
ture of the support (Fig. 1): Doping TiO2 with Ca2 cations
+
3
.2. Kinetic Experiments
results in increased selectivities while the opposite is true
upon doping TiO2 with W6 cations. The activity of the cat-
+
Several sets of experiments have been carried out in the
alysts was also found to depend on the carrier and to follow
temperature range of 973 to 1073 K to evaluate the kinetic
behavior of the three Ru catalysts. Typical results, obtained
2+
a pattern similar to that of selectivity; i.e., the Ca -doped
catalyst exhibits highest activity while the W6 -doped cat-
alyst the lowest (Table 2). It is also apparent from the re-
sults reported in Table 2 that selectivity toward CO and H2
formation in the presence of oxygen increases significantly
with increasing temperature.
+
over Ru/TiO2, Ru/TiO2(Ca2 ), and Ru/TiO2(W ) catalysts
+
6+
at 1073 K, are presented in Fig. 1, in which selectivities to-
ward CO (SCO) and H2 (SH ) are shown as functions of
2
CH4 conversion (XCH ). In all cases, two distinct regions
4
can be observed: one for XCH < 35% , where SCO and SH
4
2
It should be noted that, in all experiments where methane
conversion is less than 100% , SCO is essentially indepen-
dent of space velocity and that selectivity toward CO for-
remain constant, and one for XCH > 35% , where selectivi-
4
TABLE 1
mation is improved upon doping TiO2 with Ca2 while it is
+
Morphological Characteristics of the Ru Catalysts
6+
significantly reduced upon doping TiO2 with W cations.
A qualitatively similar behavior was observed over a wide
temperature range (973–1073 K).
Ru/TiO2 Ru/TiO2(Ca2 ) Ru/TiO2(W
+
6+
)
2
� 1
)
Specific surface area (m g
11
34
12
50
15
56
Ru dispersion (% )
3.3. XPS Spectra of Supported Ru Catalysts
at Reaction Conditions
Ru particle size ( A˚ )a
26
18
16
Crystalline mode
of the support
Rutile
Rutile
Rutile
XPS spectra obtained in the Ru(3d) region from the
unmodified Ru/TiO2 catalyst are shown in Fig. 2A. It is
a
Calculated from metal dispersion, assuming spherical particles.

1
98
ELMASIDES ET AL.
TABLE 2
Results of Activity and Selectivities toward CO (SCO) and H2 (SH ) Formation at Different Reaction
2
Temperatures under Conditions of Low Oxygen Conversion
Ru/TiO2(Ca2
+
)
Ru/TiO2(W
6+
)
Ru/TiO2
a
a
a
Reaction
temp. (K)
RCH
SCO
(% )
SH2
(% )
RCH
SCO
(% )
SH2
(% )
RCH
SCO
(% )
SH2
(% )
4
4
4
�
1
� 1
� 1 � 1
� 1 � 1
(mol g
s
)
(mol g
s
)
(mol g
s
)
5
� 5
� 5
� 5
� 6
� 6
� 5
9
023
073
73
8.0 � 10�
61
64
71
27
31
34
8.9 � 10
11.6 � 10
14.9 � 10
70
75
81
35
37
42
6.8 � 10
8.1 � 10
1.4 � 10
30
37
45
6
9
11
�
5
5
1
1
10.1 � 10
11.9 � 10�
a
The rate of CH4 consumption refers to the region where XCH4 < 10% .
observed that the spectrum obtained following reduction In addition, small amounts of Ru(II) are detectable fol-
of the fresh catalyst with 20% H2/Ar at 773 K (trace a) lowing treatment at 973 K (trace d). Under these condi-
consists of a doublet with a Ru(3d5/2) BE of 280.0 eV, tions, the Ca2 -doped catalyst is more active (XCH = 59% )
+
4
characteristic of metallic ruthenium (Ru(0)) (13, 14). The and more selective (SCO = 81% ) than the undoped one
constraints used in the spectral analysis regarding the area (Table 3).
ratio of the Ru(3d5/2) and Ru(3d3/2) peaks and their BE
In contrast, XPS spectra obtained from the W6 -doped
+
difference resulted in a satisfactory fitting of the signal catalyst are significantly different (Fig. 2C). While after H2
shape only when a second doublet, with a Ru(3d5/2) BE pretreatment of the W6 -doped catalyst the largest por-
+
of 281.1 eV, and a contribution from C(1s) at 284.6 eV tion of Ru is in the metallic state (trace a), relatively large
(
13) were also taken into account. The second doublet is amounts of Ru exist in oxidized forms at reaction conditions
attributed to Ru(II) species (15). (traces b–d). In these spectra, the contribution of Ru(IV)
The sample was then heated under He flow to 673 K is clearly necessary for the satisfactory fitting of the signal
and was subsequently exposed to a 10% CH4–5% O2 (in Ar) shapes, even at the temperature of 823 K. The differences in
mixture for 15 min (Fig. 2A, trace b). It is observed that the the oxidation state of ruthenium under reaction conditions
contributions of Ru(II) and C(1s) signals increase in inten- are also reflected in the catalytic behavior of the W6 -doped
+
sity compared to those of the H2-treated sample and that catalyst, which is less active and less selective compared to
an additional ruthenium doublet with a BE of 283.4 eV the undoped and the Ca2 -doped catalysts (Table 3). In
is required for the satisfactory fitting of the signal shape. addition, the relative intensity of the C(1s) peak is much
This doublet is attributable to Ru(IV) species (15). Results lower over the present catalyst. This is better observed in
of on-line mass spectrometry obtained at the exit of the Fig. 3 in which the C/Ru atomic ratios calculated from the
pretreatment chamber are summarized in Table 3. It is ob- XPS spectra of the examined catalysts are plotted as func-
served that, at 673 K, conversion of CH4 is low (8% ), with tions of reaction temperature. It is observed that significant
+
CO2 and H2O being the only reaction products.
quantities of carbon are deposited on the undoped cata-
Subsequent treatment of the Ru/TiO2 catalyst with the lyst, the amount of which increases with increasing reac-
CH4–O2 mixture at 823 K (trace c) results in an increase of tion temperature, especially at temperatures above 823 K.
the relative intensity of the Ru(0) signal at the expense of The C/Ru atomic ratio is small and nearly independent of
Ru(II) and Ru(IV) signals. Under these conditions, the con- temperature for the W6 -doped catalyst, while it takes in-
+
version of CH4 increases to 33% and CO appears in the gas termediate values for the Ca2 -doped sample. It is worth
+
phase (Table 3). At the reaction temperature of 973 K, Ru noticing that under reaction conditions the carbon signal
is fully reduced (trace d) and conversion of methane and on the Ca2 -doped sample appears at a higher binding
+
selectivity toward CO further increase to 42% and 74% , energy (ca. 0.7 eV) compared to the carbon deposits on the
6+
respectively (Table 3). In addition to the progressive re-
duction of the catalyst, increasing the reaction temperature will be further discussed, the carbon detected on the Ca
results in a significant increase of the intensity of the C(1s) doped sample may be attributed to stable CHx species (16,
W
-doped and undoped catalysts (Fig. 2, traces c, d). As
2+
-
signal (Fig. 2A, b,c).
17) created on the surface during the dissociative adsorp-
The corresponding spectra obtained from the Ca2+-do- tion of methane on the Ru/TiO2 (Ca ) catalyst.
2+
ped catalyst are shown in Fig. 2B. It is observed that
the spectra are qualitatively similar to those obtained
from the undoped catalyst (compare with Fig. 2A). How-
3
.4. FTIR Experiments
ever, the relative intensity of the C(1s) peak is lower in the
3.4.1. CO adsorption on fresh and spent catalysts. The
present case under all experimental conditions employed. oxidation state of the Ru catalysts was also investigated by

METHANE TO SYNTHESIS GAS OVER Ru/TiO2 CATALYSTS
199
FIG. 2—Continued
monitoringCO adsorption at 298K over samplespreviously
treated with a 10% CH4–5% O2 (in Ar) mixture at 773 and
9
73 K. For comparison, similar spectra were also obtained
over H2-reduced and partially oxidized samples. In the lat-
ter case, the samples were treated with oxygen (7% O2
in Ar) under mild conditions, i.e., at 473 K for 10 min.
2+
Results obtained from the Ru/TiO2, Ru/TiO2(Ca ), and
Ru/TiO2(W6 ) catalysts are presented in Figs. 4A–4C, re-
+
spectively.
The FTIR spectrum obtained following exposure of the
H2-reduced Ru/TiO2 catalyst to 7% CO for 15 min at RT is
shown in Fig. 4A, trace a. Several bands due to the adsorbed
CO species can be observed in the �(C–O) region, located
�
1
at 2135, 2070(sh), and 2043 cm while a broad spectral
feature may also be distinguished as a low-frequency “tail”
�
1
� 1
of the 2043-cm band. The band located at 2043 cm
may be safely attributed to CO linearly adsorbed on
0
reduced metal crystallites (Ru –CO) (18). This has also
been confirmed in our previous study (13) based on the
frequency dependence of this peak to surface coverage.
�
1
Peaks at 2135 and 2070 cm appear at frequencies where
n+
multicarbonyls on oxidized Ru sites [Ru (CO)x, n = 2 or
3
, x = 2 or 3] are expected to absorb (19–21). These peaks
most probably originate from two different multicarbonyl
2
+
FIG. 2. XPS spectra obtained from (A) Ru/TiO2, (B) Ru/TiO2(Ca ),
n+
6
+
species adsorbed on isolated oxidized Ru
It is, however, possible that part of the 2070-cm band
corresponds to Ru –CO monocarbonyl. For example,
sites (13).
and (C) Ru/TiO2(W ) catalysts after reduction with H2 at 773 K for 20 min
�
1
(
(
a) following treatment with a 10% CH4–5% O2 (in Ar) mixture at 673 K
b), 823 K (c), and 973 K (d).
n+

2
00
ELMASIDES ET AL.
TABLE 3
Conversion of Methane (XCH4 ) and Selectivity toward CO (SCO) Measured at the Exit of the XPS
Pretreatment Chamber under the Conditions Described in Fig. 2
+
6+
)
Ru/TiO2
XCH4 (% )
Ru/TiO2(Ca2
)
Ru/TiO2(W
SCO (% )
Reaction
temp. (K)
SCO (% )
SCO (% )
XCH4 (% )
XCH4 (% )
6
8
9
73
23
73
0
50
74
8
33
42
0
50
81
11
35
59
0
35
54
13
24
28
Yokomizo et al. (19), who studied CO adsorption on �(C–O) region unless they are too weak to be observed
�
1
Ru/SiO2, attributed a band at 2140 cm and a portion due to the orientation of the molecular axis of CO with re-
�
1
n +
of a band at 2080 cm to tricarbonyl species bonded on spect to the surface normal (22). For example, Ru (CO)2
partially oxidized particles, while the remainder of the dicarbonyls are characterized by a pair of bands at 2092–
band at 2080 cm was attributed to CO linearly bonded 2045 and 2038–1970 cm (13), the latter of which lies in the
�
1
� 1
on Ru sites perturbed by adsorbed oxygen atoms.
The FTIR spectra obtained from the Ru/TiO2 catalyst
spectral region where the “tail” under discussion is located.
The FTIR spectra obtained from the Ru/TiO2(Ca2
+
)
previously treated with the reaction mixture at 773 and catalyst are shown in Fig. 4B. As in the case of the undoped
73 K are presented in Fig. 4A, traces b and c, respectively. catalyst, the spectra obtained from the sample previously
9
These spectra are qualitatively similar to the one obtained treated with the CH4–O2 mixture at 773 and 973 K (traces
over the reduced catalyst (compare with trace a), indicating b and c, respectively) are characterized by an intense band
0
� 1
that Ru remains in its reduced state following treatment at due to Ru –CO, which is now shifted to ca. 2050 cm . Com-
these reaction conditions. This is also evidenced by compar- parison with the spectra obtained following treatment with
ison with the (very different) spectrum obtained from the H2 (trace a) or O2 (trace d) shows that, as in the case of
preoxidized sample, which is characterized by an intense the undoped catalyst, under reaction conditions Ru mainly
�
1
band at 2070 cm (trace d).
exists in its metallic state.
+
)
In addition to the clearly distinguishable bands discussed
The correspondingspectra obtained from Ru/TiO2 (W6
above, one or more bands might be present in the IR spec- are presented in Fig. 4C. In contrast to the undoped and the
2+
tra of Fig. 4A, which are responsible for the low-frequency Ca -doped catalysts, spectra obtained following treatment
�
1
shoulder of the 2043-cm band. A portion of this “tail” at reaction conditions (traces b and c) are dominated by
�
1
may be attributed to the low-wavenumber counterpart of intense bands at 2142 and 2080 cm , which, as discussed
the species responsible for the observation of the 2135- and above, are characteristic of CO species adsorbed on oxi-
070-cm peaks. Multicarbonyl species are normally ex- dized sites.
�
1
2
pected to exhibit more than one vibrational bands in the
3
.4.2. In situ FTIR spectra under reaction conditions. In
2+
situ FTIR spectra obtained from Ru/TiO2, Ru/TiO2(Ca ),
and Ru/TiO2(W6 ) catalysts under reaction conditions, in
the temperature range of 773–1073 K, are presented in
Fig. 5A–5C, respectively. In these experiments, the efflu-
ent from the IR cell was analyzed by gas chromatography
and the results are summarized in Table 4. It is observed
+
TABLE 4
Conversion of Methane (XCH4 ) and Selectivity toward CO (SCO)
Measured at the Exit of the FTIR Cell under the Conditions
Described in Fig. 5
Ru/TiO2(Ca2
+
)
Ru/TiO2(W
6+
)
Ru/TiO2
TR (K) SCO (% ) XCH4 (% ) SCO (% ) XCH4 (% ) SCO (% ) XCH4 (% )
7
8
9
73
23
73
11
33
59
67
18
20
28
31
8
36
64
75
16
26
36
41
0
7
14
32
6
14
17
24
FIG. 3. Dependence of the C/Ru atomic ratios, determined from XPS
measurements, on reaction temperature.
1073

METHANE TO SYNTHESIS GAS OVER Ru/TiO2 CATALYSTS
201
phase CO (2180/2100 cm 1) are not present at this reaction
temperature, analysis of the effluent gas at the exit of the
FTIR cell shows that there was CO production (Table 4).
The absence of the corresponding bands may be attributed
to the relatively low concentration of CO in the gas phase.
�
FIG. 4. Spectra of adsorbed CO obtained from (A) Ru/TiO2, (B)
2
+
6+
Ru/TiO2(Ca ), and (C) Ru/TiO2(W ) catalysts obtained at 300 K af-
ter (a) reduction with 20% H2 (in Ar) at 773 K for 60 min, (b) treatment
with 10% CH4–5% O2 (in Ar) at 773 K, (c) treatment with the same mixture
at 973 K, and (d) oxidation with 7% O2 at 473 K for 10 min.
that, in the case of Ru/TiO2, interaction of the 10% CH4–
5
% O2 (in Ar) mixture with the reduced catalyst at 773 K
FIG. 5. In situ FTIR spectra obtained from (A) Ru/TiO2, (B) Ru/
TiO2(Ca ), and (C) Ru/TiO2(W ) catalysts during interaction with flow-
ing 10% CH4–5% O2 (in Ar) mixture in the temperature range of 773–
gives rise to FTIR bands due to gas-phase CH4 (3250–2750
cm ), to gas-phase CO2 (2360/2340 cm ), and to a broad
band at ca. 1985 cm (Fig. 5A). Although bands due to gas 1073 K.
2+
6+
�
1
� 1
�
1

2
02
ELMASIDES ET AL.
Increasing reaction temperature to 823 K results in the conversions has recently been strengthened by a kinetic
appearance of bands due to gas phase CO, located at model, which satisfies very nicely the experimental results
�
1
around 2180 and 2100 cm , which continuously increase (25).
in intensity with increasing temperature, up to 1073 K
It may be concluded that, under conditions where the
(
Fig. 5A). At the same time, the bands due to gas phase direct route is operable (XCH < 35% ), doping of TiO2
4
CO2 and CH4 decrease in intensity. This implies increased with Ca2 results in enhancement of the reactions that
+
methane conversions and increased selectivities toward lead toward CO and H2 formation, while the opposite
CO, which is confirmed by GC analysis (Table 4). The is true upon doping with W6 cations. The fact that the
1
+
�
1
985-cm band decreases in intensity and shifts toward values of SCO and SH2 remain constant as long as there
lower frequencies with increasing reaction temperature and is oxygen in the gas phase (XCH < 35% , Fig. 1) implies
4
can be hardly observed at temperatures above 973 K.
that the Ru surface remains unchanged with respect to the
2+
The spectra obtained from the Ca -doped sample percentage of the active sites that are responsible for the
Fig. 5B) are qualitatively similar to the corresponding ones direct route of partial oxidation of methane for each one
(
obtained over the undoped catalyst and are also character- of the catalysts examined.
ized by a band in the spectral region where adsorbed CO
Regarding the nature of the catalytically active sites, re-
species are expected to absorb. This band, which is now sults of published work indicate that the reduced metallic
�
1
located at ca. 1995 cm , is of much higher intensity com- sites are mainly involved in CH4 activation and synthesis
pared to the one observed over the undoped catalyst (note gas formation (9, 26–28). Theoretical calculations (29) and
the difference in the scale of the y axis). In addition, the cor- experimental results (30) have shown that the rate of CH4
responding species can be clearly observed at reaction tem- adsorption increases over catalyst surfaces containing very
peratures as high as 1073 K (Fig. 5B). The peak maximum little surface oxygen or when oxygen is located at metal “on
gradually shifts toward lower frequencies upon increasing top” sites. In all cases, the presence of reduced sites seems
reaction temperature above 923 K.
In contrast to the undoped and Ca -doped catalysts, no
to be necessary for the reaction to proceed.
Since all experiments presented in Fig. 1 have been
2+
�
1
bands are observable in the 2000- to 1950-cm region over conducted under similar experimental conditions, the
the W6 -doped catalyst, even at low reaction temperatures differences observed in the catalytic performance may be
+
(
Fig. 5C). Again, this catalyst is much less active and selec- attributed to the effect of modification of the supporting
2+
tive compared to the undoped and Ca -doped ones, the material on one or more critical properties of the Ru crys-
latter exhibiting the best catalytic behavior (Table 4).
tallites.
4
. DISCUSSION
4.2. The Oxidation State of Supported Ru Crystallites
under CPO Reaction Conditions
4
.1. Effect of Carrier Doping on the Catalytic
Performance of Ru
4
.2.1. Effect of reaction temperature. XPS results of the
present study clearly show that the oxidation state of ruthe-
As observed in Fig. 1 and Table 2, under experi- nium depends on reaction temperature; i.e., the percentage
mental conditions where oxygen is not fully consumed of metallic Ru in the ruthenium crystallites increases with
XCH < 35% ), selectivities of Ru-containing catalysts increasing reaction temperature, as is clearly demonstrated
(
4
toward CO and H2 formation strongly depend on the in Figs. 2A–2C for all three catalysts examined. The en-
nature of the supporting material. It must be noticed that, hancement of the reduced fraction of the Ru with temper-
as long as there is oxygen present in the gas phase, the ature is accompanied by enhancement of selectivity of CO
contribution of CH4-reforming reactions (steam and CO2 and H2 production in the presence of oxygen (Table 2) and
reforming) is very small and, therefore, formation of CO by an increase of the amount of surface carbon that accu-
and H2 takes place mainly via the direct route of partial mulates under reaction conditions, as indicated by the C/Ru
oxidation of CH4 (12).
atomic ratios, which are shown in Fig. 3. These observations
Additional indication that the direct route is operable maybe explained asfollows:At the lower reaction tempera-
comes from the experimental conditions employed, of ture of 673 K where the catalyst is partially oxidized (Fig. 2),
which space velocity is of great importance. When the the Ru surface is almost fully covered with oxygen because
partial oxidation of methane proceeds via the indirect the competitive adsorption between oxygen and a saturated
scheme, it is expected that selectivities toward CO and H2 hydrocarbon like CH4 stronglyfavorsoxygen under the par-
will decrease with increasing space velocity (9, 23, 24). The ticular experimental conditions. This is due to the fact that,
constant selectivities observed in the present study with methane, which is a thermodynamically stable molecule,
variation of space velocity (results not shown) imply that has a low sticking coefficient at low temperatures (31), espe-
CO and H2 are the primary products of the reaction. The cially when a portion of the surface is covered by adsorbed
theory of primary formation of CO and H2 at low methane oxygen species (32, 33). This is the case for all three catalysts

METHANE TO SYNTHESIS GAS OVER Ru/TiO2 CATALYSTS
203
examined. Thus, considering CH4 dissociative adsorption as carbon at high reaction temperatures and CH4 conversions.
the rate-limiting step, the production of CO2, with selectiv- In contrast, the existence of carbon in the form of CHx on
ity approaching 100% at temperatures lower than 673 K, the Ca2 -doped sample implies that the dissociation rate
can be attributed to the significantly higher activity of the of CH4 is lower than the rate of carbon consumption and
oxidized catalysts toward combustion of adsorbed species, that the surface is covered mainly with CHx species. In the
+
which result from the dissociative adsorption of CH4. Al- case of the W6 -doped catalyst, a considerable part of the
ternatively stated, at low reaction temperatures the com- Ru crystallites remains in its oxidized form (Fig. 2C). This
petition between CH4 and oxygen for interaction with Ru is the reason for the low CH4 conversions that have been
favors the latter, resulting in an oxidized Ru surface, which observed (Table 3) and the lower activity, by 1 order of
+
leads toward CO2 and H2O formation.
magnitude, of this catalyst (Table 2).
When the reaction temperature is increased, the disso-
ciation probability of CH4 increases (30, 31) and so does
4.2.2. Effect of carrier doping. XPS (Fig. 2) and FTIR
the reaction probability, which is evidenced by the higher (Fig. 4) resultsshowthat under the same reaction conditions
CH4 conversions observed (Table 3). At 823 K, this is ac- the oxidation state of Ru depends on the supporting ma-
companied by reduction of the Ru surface (Fig. 2) and the terial. In particular, it is observed that a significant portion
+6
appearance of CO in the gas phase (Table 3). It must be of the surface of Ru crystallites dispersed on W -doped
noticed that the contribution of CH4-reforming reactions TiO2 becomes oxidized, in contrast to the other two cat-
(
CO2 and H2O) to the total CO production is negligible. alysts in which the Ru mainly exists in its metallic state.
The fact that reforming reactions contribute negligiblly to Correlation of this observation with the results of catalytic
CO and H2 production has been established in our detailed performance, which are listed in Tables 3 and 4, provides
kinetic study (25) and is supported by the observations of significant evidence that catalytic activity and selectivity
low CH4 conversions at 823 K (Table 3) and the fact that the may be related to the oxidation state of ruthenium, which
C/Ru atomic ratios do not change significantly with temper- in turn, is influenced by altervalent cation doping of the
ature in this temperature range (Fig. 3).
A definite correlation exists between the oxidation state
support.
The influence of the support on the electronic state and,
of the Ru crystallites, as observed by XPS (Fig. 2b–d), and consequently, on catalytic properties of supported metal
selectivity toward CO and H2 formation at low CH4 and particles has been the subject of extensive research work
O2 conversions (Fig. 1). The high CO and H2 selectivi- (34–37) and it is widely accepted that the oxidation state of
ties observed over the undoped and Ca2 -doped catalysts a supported metal catalyst can be strongly influenced by the
may be attributed to the ability of the corresponding carri- interaction of the metal with the carrier. Charge transfer
ers to stabilize the Ru crystallites mainly in their metallic in metal catalysts supported on a TiO2 carrier has been
state in the presence of oxygen in the gas phase. In con- proposed by many investigators to account for the observed
+
trast, on the W6 -doped catalyst, a large fraction of the Ru differences in the adsorptive and catalytic properties of the
crystallites exist in the oxidized state, which results in re- supported metal particles (34–36). Electron transfer from
duced selectivity toward CO and H2 formation via the direct the support to the metal has been proposed in the case of
+
route.
Ru catalysts supported on 19 different kinds of oxides (37).
The dissociation probability of CH4 becomes comparable The adsorptive and catalytic behavior of Ru was correlated
to the reaction rate of the oxidation processes with increas- with the electronegativity of the support, which is a measure
ing catalyst temperature. This is accompanied by further of its electron-donating properties.
reduction of the Ru surface (Fig. 2), the depletion of sur-
Regarding the present catalysts, the effect of altervalent
face oxygen, and the enhanced production of CO and H2. cation doping of TiO2 on the adsorptive and catalytic prop-
As shown in Table 3, at temperatures above 900 K CH4 erties of supported metal catalysts has also been investi-
conversion exceeds 35% and the C/Ru ratio is increased in gated (38–41). It has been shown that doping TiO2 with
2+
the case of the undoped and Ca -doped catalysts (Fig. 3). higher valence cations than that of the host cation (e.g.,
6+
These observations clearly indicate the onset of CO2- and
W
) leads to an increase of the n-type semiconductiv-
H2O-reforming reactions. It is quite remarkable that, in ity of the metal oxide, while the opposite is expected for
the case of the Ca2 -doped catalyst, under the same con- the lower valence cation (e.g., Ca ) doping. In the case
ditions of temperature and CH4 conversion, the C/Ru ra- of doped TiO2, such a metal–support interaction has been
tio is significantly lower compared to that of the Ru/TiO2 established by theoretical analysis (42) and experimental
catalyst, while the XP spectra (Fig. 2B, c, d) indicate the ex- observations (38–40). It has been proposed that a signifi-
istence of CHx species on the catalyst surface, evidenced by cant charge transfer can occur at the interface of a Group
the binding energy of the C(1s) peak at 285.2 � 0.2 eV (16, VIII metal and TiO2 when TiO2 is either reduced or doped
+
2+
1
7). Based on these considerations, it may be concluded that with higher valence cations. The magnitude of charge trans-
the undoped Ru/TiO2 catalyst is very active toward CH4 dis- fer depends on the metal particle size and can be as high as
sociative adsorption, resulting in the formation of graphitic 0.5 electrons per metal atom for crystallites smaller than 2

2
04
ELMASIDES ET AL.
nm or less than 0.01 electrons per metal atom for crystallites lites and bulk Ru metal (45). It is reasonable to assume that,
larger than 10 nm (42). under the high-temperature reaction conditions employed
Based on the theories presented above, since metal crys- here, only low coverages of strongly adsorbed CO will be
tallites may be viewed as electron donors while the ad- present on the catalyst surface and participate in surface
sorbed oxygen molecules behave as surface electron ac- reactions.
ceptors, it is expected that, in the case of the W6 -doped
+
The morphology and crystallite size of Ru may also
catalyst, where charge transfer from the carrier to the sup- determine the exact band position of adsorbed CO (46).
ported Ru is increased, the oxygen chemisorption process Bands in the spectral region where the 1985- (or 1995-)
will be favored. In addition, doping TiO2 with W6 cations cm peak is observed are characteristic of CO adsorbed
+
� 1
will also favor the formation of active oxygen species, such on high-index faces where CO adsorbs more strongly than
�
2�
2
�
as O , O , and O , which are generated by electron accep- that on a flat surface (47). This is supported by the fact
2
tance on the metal surface (43). Therefore, the stabilization that adsorption of CO on well-dispersed catalysts results
of Ru in its oxidized forms observed over the W6 -doped in lower frequency bands compared to the corresponding
catalyst under reaction conditions (Figs. 2C, 4C) may be ones over sintered catalysts (47). As has been discussed
attributed to the stronger interaction of oxygen with the in our previous study (13), spectral features located below
+
�
1
0
Ru surface compared to those of the other two catalysts, 2020 cm may be due to CO adsorption on isolated Ru
which originates from the W6 -induced increase of charge entities of very low nuclearity and/or to Ru –CO species
+
0
transfer from the carrier to the Ru crystallites.
The opposite behavior is expected from the Ca -doped
“diluted” in an oxidized environment.
As mentioned earlier, the various dopants modify
2+
catalysts under similar experimental conditions; i.e., Ru the TiO2 support, altering significantly the electronic
should remain in the metallic state under reaction condi- properties of the Ru crystallites. CO is adsorbed on top of
0
tions and should be more easily reduced compared to the Ru by donating electrons from the 5� orbital to the metal
undoped catalyst. This is because according to the above d band (48) with concomitant backdonation of electrons
discussion, doping TiO2 with Ca2 cations results in reduced to the 2� level, thus affecting the C == O bond. Therefore,
+
�
charge transfer to the Ru crystallites. Thus, it is expected it is expected that the C–Ru bond will be stronger on
that Ru on Ca2 -doped TiO2 will exhibit lower electron the Ca -doped catalyst. However, the C == O bond of the
+
2+
availability at the Fermi level. Under these conditions the adsorbed CO molecules on Ru crystallites will be less
origin and nature of the surface species detected is expected affected on the Ca2 -doped catalyst as compared to that
to be significantly affected, thus leading to the differences on the undoped one. The above conclusions are supported
in reaction rates and selectivities observed over the three by the IR spectra (Figs. 5A and 5B), which show that CO
+
Ru catalysts (Table 2).
coverage under reaction conditions is larger on the Ca2
+
-
doped catalyst while the CO stretching frequency has the
highest value among the three catalysts, denoting stronger
bonding between the carbon and oxygen. Thus, both CO
coverage and C == O stretching frequency increase in the
4
.3. Nature and Origin of Surface Species Detected
under Reaction Conditions
As observed in Fig. 5, the only IR band observed under
order TiO2(W6 ) < TiO2 < TiO2(Ca ), denoting stronger
+
2+
reaction conditions, which is attributable to adsorbed CO
bonding with the Ru surface supported on TiO2(Ca2
+
)
�
1
species, is the one located at 1985 (or 1995) cm over the
and weaker interaction of the C == O bond with the catalyst
undoped (or Ca2 -doped) catalyst. This peak appears at rel-
atively low frequency, which indicates (a) very low surface
coverage of the corresponding adsorbed CO species and/or
+
surface.
The low frequency of the 1985- (1995-) cm 1 band may
also be related to the presence of species like hydrogen
and/or carbon in the vicinity of the adsorption sites. From
various studies concerning interaction of CO + H2 over
supported Ru catalysts (20, 49, 50), IR bands observed at
�
(
(
b) adsorption on very small or isolated Ru sites and/or
c) the presence of an electron donor in the vicinity of the
adsorption sites.
Regarding the effect of surface coverage, it has been re-
ported that adsorption of CO on Ru(001) single crystals
results in the appearance of a single-frequency band due
to linearly bonded CO on reduced sites, shifting from 1984
�
1
around 1990cm were attributed to the formation ofruthe-
nium carbonyl hydrides of the following type:
H
H
�
1
to 2060 cm with increasing CO coverage from 0.003 to
.66 (44). Shifts of similar magnitude were observed by
/
/
Ru
or Ru–CO
0
\
\
Kellner and Bell over Ru/Al2O3 catalysts (45). However,
the highest and the lowest frequencies reported in that
CO
H
study were notably lower than those on single crystals, a re- Formation of ruthenium carbonyl hydrides produced
sult which was partly attributed to differences between the through the dehydrogenation ofmethane could also explain
physical properties of alumina-supported Ru microcrystal- the low vibrational frequencies observed in the present

METHANE TO SYNTHESIS GAS OVER Ru/TiO2 CATALYSTS
205
+
study (7):
sions. Although the W6 -doped catalyst should exhibit even
higher carbon accumulation, the low methane conversions
and the highly oxidizing environment (partially oxidized
catalyst and gas phase oxygen) maintains carbon at a low
level.
CO
H
/
|
OS
CH4 + Ru → CHx � Ru → Ru
or Ru–CO [1]
|
H
\H
4
.4. Mechanistic Implications
The influence of surface carbon on the vibrational fre-
quency of adsorbed CO species was supported by Dalla
Betta and Shelef (50) and Kellner and Bell (45), who as-
signed bands at frequencies lower than 2000 cm to CO
adsorbed in a linear mode at a Ru site adjacent to a nucle-
ophilic adsorbate, such as carbon.
In the present case, surface carbon formed under reac-
tion conditions could be responsible for the observed low
frequency of adsorbed CO species (Figs. 5A and 5B). If
that was the case, one would expect to observe a shift of
the band position toward lower frequencies with increas-
ing concentration of the surface carbon, especially at high
Results of the present study may be related to our recent
work in which the kinetic behavior of the Ru/TiO2(Ca2
+
)
�
1
catalyst in the partial oxidation of methane to synthesis gas
was investigated and a kinetic model was developed (25).
According to this model, CO and H2 are the primary reac-
tion products while CO2 and H2O are secondary oxidation
products. Methane adsorbs dissociatively and irreversibly
on specific surface sites (designated as S1) toward the for-
mation of surface carbon and adsorbed hydrogen atoms:
CH4 + S1 → I → C–S1 + 4H–S1.
[2]
CH4 conversions (XCH > 35% ), where oxygen is fully con-
sumed and, therefore, carbon deposition is favored. This
is clearly shown in the case of the Ca -doped catalyst
4
Results of the present study indicate that “S1” are reduced
Ru sites, which is in accordance with the findings in the
literature (9, 26–28).
Regarding the formation of CO and CO2, it was pro-
posed that the former is a primary product of the reaction
between surface carbon and oxygen and that adsorbed CO
is at equilibrium with gas phase CO (25):
2+
�
1
(
Fig. 5B), where a shift of ca. 10 cm is observed with in-
creasing temperature above 923 K where the CH4 conver-
sion is above 35% . It should be noted that part of this shift
could be attributed to the decrease in the surface coverage
of this species, which takes place upon increasing reaction
temperature.
�
�
C–S1 + O–S → CO–S1 + S
[3]
[4]
2
2
In view of the previous discussion, it may be concluded
that the IR peak located at 1985 (or 1990) cm , which is
�
1
CO + S1 ↔ CO–S1.
observed over the undoped and Ca2 -doped samples at re-
action conditions, is due to CO species linearly bonded on
reduced Ru sites. The low frequency of this peak is due to
the low surface coverage of the corresponding species. As-
suming that there are no significant differences in the mor-
phological characteristics of the three catalysts examined
under reaction conditions, such as structural arrangements
of the Ru crystallites, it is reasonable to suggest that the
presence and population of this adsorbed CO species de-
pends on the oxidation state and the electronic properties
of the catalyst, which in turn, depend on the type of the
support employed to disperse the Ru crystallites, in partic-
+
�
It must be noticed that in the elementary step (3) the S sites
2
originate from the interaction of gas phase oxygen with the
surface sites S2, which are assumed to be different from the
S1 sites. The surface sites S2 constitute the active region for
carbon dioxide production via subsequent oxidation of CO,
following its readsorption on these surface sites (S2):
CO + S2 ↔ CO–S2
CO–S2 + O–S → CO2 + S2 + S .
[5]
[6]
�
�
2
2
It is of interest to notice that the above model predicts
ular, the type of doping cation incorporated into the crystal that CO coverage decreases with increasing temperature,
matrix of TiO2. as expected from adsorption thermodynamics and that, un-
It is generally accepted that the dissociative adsorption der the experimental conditions applied in the in situ FTIR
of methane is significantly favored on metal surfaces with experiments of Fig. 5, CO coverage is ca. 0.05 at 973 K and
a high Fermi level (48). The dissociation of the C–H bond ca. 0.01 at 1073 K (25). This explains the appearance of the
depends on the strong backdonation of electrons into the FTIR band due to adsorbed CO species under such high
antibonding of the stretched C–H bond. Thus, it is reason- reaction temperatures.
able to consider that the CHx species, shown in the XP
In view of the in situ FTIR results of the present study
spectra in Fig. 2, can be more stable on the Ca2 -doped (Fig. 5), it is tempting to correlate the adsorbed CO species
+
catalyst since the electron transfer at Ru/TiO2(Ca2 ) is less of the above kinetic model (namely, CO–S1) with the FTIR
among the three catalyst systems examined. In contrast, in peak observed under reaction conditions. It may be as-
the case of the Ru/TiO2 catalyst, the production of graphitic sumed that this peak corresponds to an adsorption site
carbon is prevailing at high temperatures and CH4 conver- where CO can accumulate and desorb (Eq. [4]). Moreover,
+

2
06
ELMASIDES ET AL.
it could be assumed that the dissociation ofCH4 and the pro- tion site that seems to be necessary for the reaction to
duction of CO take place in the same surface-active sites proceed.
S1. An indication that supports the assumption of CH4 dis-
sociation and the adsorption of CO at the same active site
is the frequency shift of the peak when the conditions are
REFERENCES
favorable for carbon deposition (Fig. 5). The absence of ad-
sorbed oxygen on adjacent sites allows desorption of CO
without further oxidation at these sites.
1
2
3
4
. Pena, M. A., Gomez, J. P., and Fierro, J. L. G., Appl. Catal. A 144, 7
(1996).
. Mallens, E. P. J., Hoebink, J. H. B. J., and Marin, G. B., J. Catal. 167,
43 (1997).
The appearance and relative intensityofthe 1985-(1995-)
. Fathi, M., Monnet, F., Schuurman, Y., Holmen, A., and Mirodatos, C.,
J. Catal. 190, 439 (2000).
. Buyevskaya, O. V., Walter, K., Wolf, D., and Baerns, M., Catal. Lett.
38, 81 (1996).
�
1
cm peak under reaction conditions may be related to the
catalytic performance of the Ru catalysts examined. As ob-
served in Fig. 5, this band is present over the undoped and
2
+
5. Torniainen, P. M., Chu, X., and Schmidt, L. D., J. Catal. 146, 1 (1994).
6. Qin, D., Lapszewicz, J., and Jiang, X., J. Catal. 159, 140 (1996).
7. Tian, Z., Dewaele, O., and Marin, G. B., Catal. Lett. 57, 9 (1999).
8. Soick, M., Buyevskaya, O. V., H o¨ henberger, M., and Wolf, D., Catal.
Today 32, 163 (1996).
Ca -doped catalysts, which are highly active and selective
toward synthesis gas formation and over which the metal
mainly exists in its reduced form. The high selectivity ob-
tained over these catalysts strengthens the view that these
species survive and desorb from the catalyst surfaces with-
9. Dissanayake, D., Rosynek, M. P., Kharas, K. C. C., and Lunsford, J. H.,
J. Catal. 132, 117 (1991).
0. Hu, Y. H., and Ruckenstein, E., J. Catal. 158, 260 (1996).
1. Boucouvalas, Y., Zhang, Z. L., Efstathiou, A. M., and Verykios, X. E.,
Stud. Surf. Sci. Catal. 101, 443 (1996).
6+
out being further oxidized. In contrast, over the W -doped
catalyst, which is much less active and selective and which
under reaction conditions mainly exists in oxidized forms,
such IR bands are absent.
1
1
1
1
1
1
1
2. Boucouvalas, Y., Zhang, Z. L., and Verykios, X. E., Catal. Lett. 40, 189
(
1996).
3. Elmasides, C., Kondarides, D. I., Gr u¨ enert, W., and Verykios, X. E.,
J. Phys. Chem. B 103, 5227 (1999).
4. Fuggle, J. C., Madey, T. E., Steinkilberg, M., and Menzel, D., Surf. Sci.
2, 521 (1975).
5. Chan, H. Y. H., Takoudis, C. G., and Weaver, J., J. Catal. 172, 336
1997).
6. Johnson, E. E., and Ratner, B. D., J. Electron Spectrosc. Related Phe-
nom. 81, 303 (1996).
5
. CONCLUSIONS
The following conclusions may be drawn from the results
of the present study:
5
(
1
. Under conditions where the direct route for the partial
oxidation of methane to synthesis gas formation is operable
(
XCH < 35% ), the catalytic performance of Ru/TiO2 cata- 17. Sabbatini, L., and Zambonin, P. G., J. Electron Spectrosc. Related Phe-
4
nom. 81, 285 (1996).
lysts is improved upon doping of TiO2 with small amounts
of Ca cations, while the opposite is true upon doping with
W
2
+
18. Hadjiivanov, K., Lavalley, J. C., Lamotte, J., Maug e´ , F., Saint-Just, V.,
and Che, M., J. Catal. 176, 415 (1998).
6
+
cations.
1
2
9. Yokomizo, G. H., Louis, C., and Bell, A. T., J. Catal. 120, 1 (1989).
0. Gupta, N. M., Kamble, V. S., Iyer, R. M., Ravindranathan Thampi, K.,
and Gratzel, M., J. Catal. 137, 473 (1992).
2
. The oxidation state of ruthenium depends strongly on
the reaction temperature and, under reaction conditions,
the fraction of metallic Ru increases with increasing tem-
perature. This is accompanied by increased selectivities to-
ward CO and H2 formation, indicating that Ru(0) sites are
necessary for the reaction to proceed.
2
2
2
1. Robbins, J. L., J. Catal. 115, 120 (1989).
2. Chen, H. W., Zhong, Z., and White, J. M., J. Catal. 90, 119 (1984).
3. Van Looij, F., van Giezen, J. C., Stobbe, E. R., and Geus, J. W., Catal.
Today 21, 495 (1994).
2
2
2
2
4. Nakagawa, K., Ikenaga, N., Suzuki, T., Kobayashi, T., and Haruta, M.,
Appl. Catal. A 169, 281 (1998).
5. Elmasides, C., Ioannides, T., and Verykios, X. E., AIChE J. 46, 1260
3
. Altervalent cation doping of TiO2 results in significant
changes in the oxidation state of supported Ru, which is
reflected in its catalytic performance. In particular, doping
(
2000).
6. Wang, D., Dewaele, O., De Groote, A. M., and Froment, G. F., J. Catal.
59, 418 (1996).
7. Au, C. T., and Wang, H. Y., J. Catal. 167, 337 (1997).
with W6 cations results in the stabilization of a portion of
+
1
Ru in its oxide forms under reaction conditions and, con-
comitantly, in lower methane conversions and lower selec- 28. Wang, H. Y., and Ruckenstein, E., Catal. Lett. 59, 121 (1999).
9. Au, C-T., Liao, M-S., and Ng, C-F., J. Phys. Chem. A 102, 3959 (1998).
30. Wang, D., Dewaele, O., and Froment, G. F., J. Mol. Catal. A 136, 301
1998).
1. Koerts, T., Deelen, M. J. A. G., and van Santen, R. A., J. Catal. 138,
01 (1992).
32. Valden, M., Xiang, N., Pere, J., and Pessa, M., Appl. Surf. Sci. 99, 83
1996).
33. Valden, M., Pere, J., Xiang, N., and Pessa, M., Chem Phys. Lett. 257,
89 (1996).
4. Belton, D. N., Sun, Y. M., and White, J. M, J. Phys. Chem. 88, 5172
1984).
2
tivities toward CO and H2 formation compared to the un-
doped catalyst. In contrast, doping of the carrier with Ca2
+
(
cations results in stabilization of Ru in its metallic form
and, thus, to enhanced catalytic performance with respect
to activity and selectivity for CO and H2 formation.
3
1
(
4
. In situ FTIR results show that an adsorbed CO species
2+
exists on the surface of the undoped and Ca -doped cat-
alysts under reaction conditions, even at temperatures as
high as 1073 K. This species, which is not observable over
2
3
(
the W6 -doped catalyst, may be related to an adsorp- 35. Solymosi, F., Tombacz, I., and Kocsis, M., J. Catal. 75, 78 (1982).
+

METHANE TO SYNTHESIS GAS OVER Ru/TiO2 CATALYSTS
207
3
6. Raupp, G. B., and Dumesic, J. A., J. Catal. 97, 85 (1986).
43. Sokolovskii, V. D., Catal. Rev.-Sci. Eng. 32, 1 (1990).
3
7. Ishihara, T., Harada, K., Eguchi, K., and Arai, H., J. Catal. 136, 161
44. Pfnur, H., Menzel, D., Hoffman, F., Ortega A., and Bradshaw, A. M.,
Surf. Sci. 93, 431 (1980).
(
1992).
3
3
8. Akubuiro, E. C., and Verykios, X. E., J. Catal. 103, 320 (1987).
9. Zhang, Z., Kladi, A., and Verykios, X. E., J. Phys. Chem. 98, 6804
45. Kellner, C. S., and Bell, A. T., J. Catal. 71, 296 (1981).
46. Dalla Betta, R. A., J. Phys. Chem. 79, 2519 (1975).
47. Guglielminotti, E., Spoto, G., and Zecchina, A., Surf. Sci. 161, 202
(1985).
48. Van Santen, R. A., and Neurock, M., Catal. Rev.-Sci. Eng. 37, 557
(1995).
(
1994).
0. Solymosi, F., Tombacz, I., and Koszta, J., J. Catal. 95, 578
1985).
1. Zhang, Z., Verykios, X. E., and Baerns, M., Catal. Rev.-Sci. Eng. 36(3),
07 (1994).
2. Ioannides, T., and Verykios, X. E., J. Catal. 161, 560 (1996).
4
4
4
(
5
49. Solymosi, F., Tombacz, I., and Kocsis, M., J. Catal. 75, 78 (1982).
50. Dalla Betta, R. A., and Shelef, M., J. Catal. 48, 111 (1977).