Methane reaction with NO over alumina-supported Ru nanoparticles

Article abstract of DOI:10.1006/jcat.2001.3489

The reaction between NO and CH₄ over alumina-supported Ru nanoparticles was analyzed from NO abatement, CH₄ partial oxidation, and structural sensitivity points of view. The NO was selectively converted to N₂ starting from

Full text of DOI:10.1006/jcat.2001.3489

Journal of Catalysis 207, 66–75 (2002)
doi:10.1006/jcat.2001.3489, available online at http://www.idealibrary.com on
Methane Reaction with NO over Alumina-Supported Ru Nanoparticles
Ioan Balint,∗,1
Akane Miyazaki,† and Ken-ichi Aika†
∗
Institute of Physical Chemistry, Romanian Academy, Spl. Independentei 202, 77208 Bucharest, Romania; and †Department of Environmental
Chemistry and Engineering, Interdisciplinary Graduate School of Science and Technology, Tokyo Institute of Technology,
4259 Nagatsuta, Midori-ku, Yokohama 226-8502, Japan
Received October 1, 2001; revised December 3, 2001; accepted December 5, 2001
selectively oxidized to syngas (CO and H2). The study of
The reaction between NO (1%) and CH (0.55%) over alumina NO/CH4 reaction from a structure sensitivity perspective is
4
supported Ru nanoparticles is analyzed from NO abatement, interesting as well because it provides deeper insight into
methane partial oxidation, and structural sensitivity points of view.
the reaction mechanism.
The 6 and 12 wt% Ru/Al O catalysts were prepared by deposi-
2
3
Tostudytogethertheaforementionedphenomena, asuit-
able catalytic system should be selected and prepared ac-
cordingly. The catalyst chosen in this study was Ru/Al2O3
because (i) ruthenium proved to be active and selective for
the partial oxidation of methane (POM) (1), and (ii) well-
structured metallic Ru nanoparticles can be obtained via
the colloid method (2) (control of particle morphology is
tion of well-defined colloidal Ru nanoparticles on alumina. The
◦
NO was selectively converted to N starting from 450 C. Methane
2
conversion to COx, H , and H O proved to be structure sensitive at
2
2
◦
◦
4
50 C and structure insensitive at T > 450 C. Methane was selec-
tively converted to CO at 450 C over the 12% Ru/Al O catalyst
◦
2
2 3
havinglargeraverageRuparticlesize(averageparticlesize, 5.8nm).
Methane was oxidized to CO and CO over the 6% Ru/Al O cata-
2
2 3
lyst with smaller Ru nanoparticles (average particle size, 4.8 nm). important in structure-sensitive reactions).
◦
It is concluded that, at low temperature (450 C), methane is pref-
In this article the state of the art with respect to NO/CH4
erentially converted to CO and H over small Ru nanoparticles. reaction is analyzed in more detail from environmental,
2
The highest reaction selectivity to CO (≈80%) and H (≈78%) was
2
POM, and structural sensitivity points of view.
◦
observed at 600 C for both catalysts investigated. Ru/Al O cata-
2
3
The direct decomposition of NO, although thermody-
namically favored, encounters serious difficulties because
most of the catalysts are self-poisoned by the oxygen re-
sulting from NO decomposition. Therefore, hydrocarbons
lysts showed high and stable catalytic activity in time (i.e., 96 h),
◦
even at high reaction temperature (600 C). A selectivity to CO and
◦
H of almost 100% was observed at 600 C for both a stoichiomet-
2
ric mixture of NO/CH and one with an excess of methane. The
4
(
i.e., methane, ethylene, propylene) are used to remove the
catalyst with smaller Ru nanoparticles (6% Ru/Al O ) was rapidly
2
3
poisoning oxygen from the surface of the catalyst. Methane
is one of them, which from economical point of view is very
deactivated either by O or by NO, whereas the catalyst with larger
2
Ru particles (12% Ru/Al O ) proved to be more resistant to oxygen
2
3
poisoning. Both catalysts were rapidly reactivated in pure CH4. convenient because it is relatively inexpensive.
A reaction mechanism is proposed in light of the experimental
results.
Key Words: Ru catalyst; methane partial oxidation; methane
totaloxidation;structure-sensitivereactions;NOselectivereduction
by methane.
The oxidative conversion of methane to syngas is also fa-
vored from a thermodynamic point of view, but the problem
lies in the deep oxidation of methane to CO2 and water. Ru,
Ni-based catalysts—perovskites, pyrochlores—have been
reported to have a good activity and selectivity in POM
reaction (3–6). The main problem is the high temperature
ꢀc 2002 Elsevier Science (USA)
required to achieve high conversion and selectivity (i.e.,
1
. INTRODUCTION
◦
850 C). The oxidizing agents typically used are oxygen and
carbondioxide(methanecarbondioxidereforming). Toour
knowledge, there are no published reports on POM using
NO as oxidant.
Presumably, NO/CH4 should be a structure sensitive re-
actionbecauseitinvolvestheformationofN–Nbonds(7). It
iswellknownthat, inastructure-sensitivereaction, catalytic
activity and selectivity are strongly related to the morpho-
logy of the supported metal particles (i.e., size and shape)
The reaction of NO with CH4 is a particularly interest-
ing one and worthy of investigation from several points
of view. From an environmental point of view it is clear
that nitrogen oxides are extremely harmful pollutants and
their removal is imperative. Methane reaction with NO can
be also a subject of interest, especially if methane can be
1
To whom correspondence should be addressed at the temporary ad-
(
8). The conventional method of catalyst preparation,
dress at the Tokyo Institute of Technology. Fax: +81-045-924-5441. E-mail:
balint@chemenv.titech.ac.jp (or ibalint@chimfiz.icf.ro).
impregnation of the porous support (e.g., γ -Al O ) with
2
3
66
0
ꢀ
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.

CH4/NO REACTION OVER Al-SUPPORTED Ru NANOPARTICLES
67
an aqueous solution of a metal salt, yields structures con- (Philips CM 20 transmission electron microscope). Addi-
sisting of metal particles that are nonuniform in size and tionally, the average size of the Ru nanoparticles was cal-
shape and often too small to be characterized precisely. culated by H2 and CO chemisorption methods by assuming
Additionally, the support has great influence on the activ- that each exposed Ru atom can adsorb one hydrogen atom
ity of the metal when classic methods are used for catalyst or one CO molecule.
preparation (9). We have shown in a series of articles that
The concentration of Ru, both on the support and in
the catalytically active phase (metal) contains, depending the liquid phase after solid decantation, was determined by
on the impregnation conditions (temperature and pH) and inductively coupled plasma (ICP) MS (Seiko SPS-4000) to
the metal precursor used, variable amounts of aluminum as evaluate the efficiency of Ru deposition on alumina.
a result of support dissolution (10, 11).
X-Ray power diffraction analysis (CuKα radiation) was
It is somewhat surprising that information on the struc- carriedoutwithaRigakuMultiflexdiffractometerprovided
tural effects on the NO/CH4 reaction is relatively scarce, with peak assignment software.
although the reaction has been intensively investigated. In
Temperature-programmed reduction (TPR) experi-
structure-sensitive reactions, morphological control of the ments were performed in a flow system with a quadrupole
supported metal particles is essential to obtain high cata- mass spectrometer (Anelva M-QA200TS) using a mixture
lytic activity and selectivity to the desired products (7). of 5% H2 in Ar and 0.05 g of catalyst. The heating rate and
Therefore, a necessary step to a better understanding of the the total flow of the H2–Ar mixture in TPR measurements
◦
−1
−1
NO/CH4 reaction mechanism is the preparation of catalysts were 10 C min and 10 ml min , respectively.
with well-defined metal particles (size and/or crystal struc- Activity tests of Ru/γ -Al2O3 catalysts for NO/CH4 reac-
ture). A practical way to prepare well-defined metal parti- tion were performed at atmospheric pressure with 0.05 g of
cles is the colloid route. The major advantage of this method catalyst loaded in a quartz microreactor (i.d. = 5 mm). The
is that it provides relatively monodispersed metal particles. reactor was heated by a furnace connected to a temperature
Additionally, the support effect on the active metallic phase controller (Shimaden, Model SR 25). Electronic flow con-
is minimized because the metallic nanoparticles (i.e., Ru) trollers (Kofloc, Model 3660) were used to control the flow
are deposited onto the support after their formation is com- of each reactant gas. The total flow rate of reactant mix-
3
−1
plete (2).
ture (NO in Ar and CH4) was 50 cm min STP (standard
The aims of our research are (i) to support uniform col- temperature and pressure). The corresponding GHSV (gas
−
1
loidal Ru nanoparticles on alumina and then to test the hour space velocity) was 60,000 h . The typical composi-
specific catalytic activity for NO/CH4 reaction; (ii) to gain tion of the reactant mixture, using Ar as diluent, was 1%
insight into the reaction mechanism, and (iii) to prove a NO and 0.55% CH4.
relationship exists between the structure of the supported
Effluent gases from the reactor were analyzed with a GL
Ru nanoparticles (size) and the catalytic activity as well as Sciences (Model 320) gas chromatograph equipped with
selectivity to specific reaction products.
TCD detectors. The typical reaction products were COx ,
N2, and H2O. H2, N2, NO, CH4, and CO were separated and
analyzed by using a molecular sieve 13X column, and CO2
was analyzed using a Porapaq-Q column. Peak areas were
measured using a Shimadzu (Model C-R 5A) electronic
integrator.
2
. EXPERIMENTAL
The Ru/Al2O3 catalyst was prepared by reduction of
RuCl3 with ethylene glycol in the presence of γ -alumina
The heating rate of the catalyst up to the reaction temper-
2
−1
(
Aerosil, 73 m g ). The details of preparation are de-
◦
−1
ature was 10 C min . At each temperature of investigation
the catalyst was held at least 30 min in the reactant mixture
to reach steady-state activity.
scribed elsewhere (2). Briefly, alumina was added under
stirring to the ethylene glycol solution containing the dis-
solved RuCl3 · 3H2O (Wako Chemicals, purity >99%) to
form a suspension. Then, the temperature of the suspen-
The activity of the catalyst is defined as the conversion
[
X(%)] of the reactant (NO or CH4) to all products,
◦
sion was slowly increased to 180 C to allow the reduction
of RuCl3 by the ethylene glycol on the surface of alumina.
After ethylene glycol was removed by diluting the suspen-
sion with 0.3 M NaNO3, the remaining solid was calcined
[
[
NO]in − [NO]out
XNO(%) =
× 100,
[1]
[2]
[NO]in
◦
at 500 C in air for 8 h to decompose the remaining organic
CH4]in − [CH4]out
phase (ethylene glycol). Two types of catalysts, containing 6
and 12% Ru on Al2O3, were prepared. The surface area of
the final catalytic materials was similar to that of the initial
support (alumina Aerosil).
XCH (%) =
× 100,
4
[CH4]in
where the subscripts “in” and “out” indicate the inlet and
The size distributions of the Ru nanoparticles supported outlet concentrations of the reactant gases (NO and CH4),
on alumina for both catalysts were determined by TEM respectively.

6
8
BALINT, MIYAZAKI, AND AIKA
Reaction selectivity to CO [SCO(%)] and CO2 [SCO (%)]
TABLE 1
2
is defined as
Average Size of Ru Nanoparticles, as Determined from
Transmission Electron Micrographs and by H
sorption
2
and CO Chemi-
[
CO]
CO2] + [CO]
CO2]
CO2] + [CO]
SCO(%) =
× 100,
× 100,
[3]
[4]
[
[
Ru particle
size (nm)
[
SCO (%) =
2
Catalyst
TEM
H2
CO
Reaction selectivity to H2 (SH ) was deduced from the
amount of CH4 converted and from the resulting amount
of H2:
2
6% Ru/Al2O2 (fresh)
4.8
7.6
4.8
—
5.4
—
6% Ru/Al O (after more than 80 h
2
2
time-on-stream in the temperature
◦
range 400–600 C)
[
H2]
12% Ru/Al2O3 (fresh)
5.8
8.9
5.7
—
5.9
—
SH (%) =
× 100.
[5]
2
12% Ru/Al2O3 (after more than 80 h
time-on-stream in the temperature
{
[CH4]in − [CH4]out} × 2
◦
range 400–600 C)
3
. RESULTS
around 1%) Ru nanoparticles can be identified, as well.
The average size of the Ru nanoparticles, determined statis-
tically from transmission electron micrographs, was 4.8 nm.
Incontrastto6%Ru/Al2O3, theinitialsizedistributionof
the Ru nanoparticles for 12% Ru/Al2O3 is broader, ranging
between 3 and 10 nm. Most of the Ru nanoparticles are 5 nm
The preparation variables, such as the concentration of
Ru precursor in the organic phase (ethylene glycol) and
reduction temperature, have great influence on the final
morphology of the supported Ru nanoparticles. Generally,
a high concentration of RuCl3 in ethylene glycol and low
reduction temperature favor the formation of large metal
clusters followed by their aggregation. The optimal reduc-
tion temperature and concentration of RuCl3 in ethylene
glycol to obtain uniform Ru nanoparticles were found to be
(
≈26%) and 6 nm (≈28%) but there is also an important
fraction of larger Ru particles ranging between 7 and 10 nm
≈26%). The average diameter of the Ru nanoparticles was
5.8 nm.
In addition to the TEM measurements, the average size
(
◦
−3
1
80 C and 10 M, respectively.
Figure 1 presents comparatively the size distribution of
of the alumina-supported Ru nanoparticles was determined
by H2 and CO adsorption (Table 1). It can be seen that there
is a good agreement between the average size of Ru parti-
cles determined statistically from transmission electron mi-
crographs and the average size estimated from chemisorp-
tion data.
Table 1 also shows that the supported Ru nanoparticles
are subject to thermal coarsening with time. The average
size of the alumina-supported Ru nanoparticles increased
from 4.8 to 7.6 nm for 6% Ru/Al2O3 and from 5.8 to 8.9 nm
the alumina-supported Ru nanoparticles for low (6%
Ru/Al2O3) and high (12% Ru/Al2O3) metal loading as
statistically determined from transmission electron micro-
graphs.
The initial size distribution of the Ru nanoparticles for
6
6
% Ru/Al2O3 is remarkably narrow, ranging between 3 and
nm. Around 78% of the Ru nanoparticles are either 4
or 5 nm. Larger (6 nm, around 21%) and smaller (3 nm,
for 12% Ru/Al2O3 after cycling up and down for more than
◦
8
0 h in the temperature range 400–600 C.
3.1. Catalytic Activity for NO/CH4 Reaction
The dependence of catalytic activity on reaction temper-
ature for 6% Ru/Al2O3 is illustrated in Fig. 2. Conversion
◦
of NO at 400 C is close to zero (only traces of NO decom-
posed, for a short time, to N2). An increase in temperature
◦
from 400 to 450 C led to a sudden increase in NO conver-
sion, up to 100%. At any of the reaction temperatures, NO
was selectively converted to N2 over both Ru/Al2O3 cata-
lysts investigated. On the other hand, methane was con-
verted to COx , H2, and H2O. As expected, the amount of
methane transformed tended to increase with increasing
temperature. As can be seen in Fig. 2, the concentration of
CO increased progressively with increasing temperature,
FIG. 1. Initial size distribution of the alumina supported Ru nanopar-
ticlesfor6%and12%Ru/Al2O3 catalystsasdeterminedfromtransmission
electron micrographs.

CH4/NO REACTION OVER Al-SUPPORTED Ru NANOPARTICLES
69
FIG. 2. NO reduction by CH4 over 6% Ru/Al2O3 in the temperature
FIG. 4. Comparison between the catalytic behavior of 6 and 12%
Ru/Al2O3 at 450 C for NO/CH4 reaction (reactant mixture: 1% NO,
.55% CH4, and balance Ar).
◦
range 400–600 C. (ꢀ) NO, (×) CH4; (+) N2; (ꢁ) CO2; (ꢂ) CO; (ꢃ) H2
◦
(
reaction mixture 1% NO, 0.55% CH4, and balance Ar).
0
◦
◦
from 0.16% at 450 C to 0.42% at 600 C. In contrast to CO,
the concentration of CO2 decreased with increasing reac-
◦
At 450 C, the two catalytic materials investigated show
significant differences in their activity and selectivity
Fig. 4). The catalyst (6%Ru/Al2O3) with smaller Ru
nanoparticles has higher activity to methane conversion
and higher selectivity to CO as compared with the cata-
lyst (12% Ru/Al2O3) with larger Ru nanoparticles. Inter-
◦
◦
tion temperature, from 0.23% at 450 C to 0.1% at 600 C.
The evolution of hydrogen closely follows that of CO. The
concentration of H2 increases with increasing reaction tem-
perature from around 0.2% at 450 C to 0.86% at 600 C.
As a consequence, water production (estimated from the
chromatographic peak area) decreases as the reaction tem-
perature is raised.
(
◦
◦
◦
estingly, at 450 C CH4 is selectively converted to CO2 and
H2O over 12% Ru/Al2O3 catalyst. This feature can be use-
ful from the perspective of NO abatement since both CH4
and NO are converted to nonharmful compounds, CO2 and
N2, respectively.
Similar to 6% Ru/Al2O3, the 12% Ru/Al2O3 catalyst ex-
hibits a dependence on reaction temperature (Fig. 3). CO
and H2 production increased with increasing reaction tem-
perature. The relatively poor carbon balance of ≈84% ob-
Relatively similar catalytic behavior of the two catalysts
◦
◦
investigated has been observed for T > 450 C. At 600 C,
the conversion of NO and CH4 is 100% and the selectivities
to N2, CO, and CO2 for both catalysts are around 100, 80,
and 20 respectively (Fig. 5). The only small difference is that
the 6% Ru/Al2O3 catalyst has a slightly higher selectivity
to H2 (around 80%) as compared with the 12% Ru/Al2O3
catalyst (76%).
◦
served at 450 C is probably due to the support coking taking
place in the first stage of reaction. However, the coking did
not affect the activity of Ru/Al2O3 catalysts.
Comparisons between the catalytic behavior of 6%
and 12% Ru/Al2O3 catalysts for NO/CH4 reaction at the
extreme temperatures of our study, 450 and 600 C, are
presented in Figs. 4 and 5.
◦
FIG. 3. NO reduction by CH4 over 12% Ru/Al2O3 in the temperature
FIG. 5. Comparison between the catalytic behavior of 6 and 12%
Ru/Al2O3 at 600 C for NO/CH4 reaction (reactant mixture: 1% NO,
0.55% CH4, and balance Ar).
◦
◦
range 400–600 C. (ꢀ) NO, (×) CH4, (+) N2, (ꢁ) CO2, (ꢂ) CO, (ꢃ) H2
(
reaction mixture: 1% NO, 0.55% CH4, and balance Ar).

7
0
BALINT, MIYAZAKI, AND AIKA
◦
FIG. 6. Activity dependence on the time-on-stream at 550 C for 12%
4 2 3
FIG. 7. Effect of CH concentration on the 12% Ru/Al O catalyst
◦
Ru/Al2O3 catalyst. (ꢀ) NO, (×) CH4, (+) N2, (ꢁ) CO2, (ꢂ) CO, (ꢃ) H2
activity selectivity for NO/CH4 reaction at 600 C. (ꢀ) NO conversion,
(ꢄ) CH4 conversion, (ꢅ) selectivity to N , (ꢁ) selectivity to CO , (ꢂ)
(
reaction mixture: 1% NO, 0.55% CH4, and balance Ar).
2
2
selectivity to CO, (ꢃ) selectivity to H2 (the concentration ofNOin reaction
mixture was 0.8%).
It should be observed that the 6% Ru/Al2O3 catalyst
is relatively sensitive to oxygen poisoning. In the absence
of methane, the catalyst is rapidly deactivated either by
air or by NO. The oxygen-deactivated catalyst can easily
be reactivated by keeping it in pure methane for a few
minutes at any of the reaction temperatures. On the other
hand, the catalyst working in reaction mixture shows stable
activity in time. Catalytic activity is not affected by carbon
deposition (the catalyst shows approximately the same
activity for NO/CH4 reaction even after a long contact
time with pure CH4).
for catalyst deactivation. XRD patterns of the deactivated
(
spectrum a) and active (spectrum b) 6% Ru/Al2O3 catalyst
are comparatively presented in Fig. 8. It can be observed
that for the active catalyst (spectrum b), the XRD reflec-
tions of metallic Ru are better resolved (better separation
andhigherpeakintensity)ascomparedwiththoseofdeacti-
vated catalyst (spectrum a). On the other hand, an increase
◦
in the intensity of the RuO2 XRD peak at around 2θ = 28
can be clearly observed for the deactivated catalyst. From
this it can be concluded that (i) the catalytically active phase
is related to metallic Ru state, and (ii) the RuO2 form is in-
Both 6% Ru and 12% Ru/Al2O3 catalysts exhibit sta-
ble catalytic activity in time. Figure 6 illustrates, as an ex-
◦
active for NO/CH reaction.
4
ample, the catalytic activity of 12% Ru/Al2O3 at 550 C
Figure 9 presents the TPR profiles of 6% Ru/Al2O3 cata-
as a function of time-on-stream. Only a slight decrease in
hydrogenproductionwasobservedforlongerreactiontime.
The NO/CH4 reaction over 6% and 12% Ru/Al2O3 cata-
lystswasperformedunderoxidationconditionsincetheNO
was in excess with respect to the stoichiometry of POM re-
action. The concentration of CH4 was increased to check
if a stoichiometric mixture of oxidant (NO) and reductant
can be selectively converted to syngas over Ru/Al2O3 catal-
lysts, prepared by the deposition of Ru nanoparticles on
alumina. Trace a represents the fresh catalyst, as it was af-
◦
ter calcination at 500 C in air for 8 h. Trace b is the TPR
profile of an active catalyst that worked for several hours
in NO/CH4 reaction mixture.
◦
yst. The reaction temperature selected was 600 C because
at this temperature the selectivity to CO and H2 for 12%
Ru/Al2O3 catalyst is maximal (i.e., Fig. 3). From Fig. 7 it can
◦
be observed that at 600 C, reaction selectivity to CO and H2
increases with increasing concentration of CH4 in the reac-
tion mixture. For a composition close to the stoichiometric
one (0.8% NO–1.1% CH4), reaction selectivities to CO and
H2 are 99 and 93%, respectively. For higher methane con-
centration, reaction selectivities to CO and H2 are 100%.
3
.2. Characterization of Ru/Al2O3 Catalysts
by XRD and TPR
The 6% Ru/Al2O3 catalyst proved to deactivate rapidly
under oxidation conditions, either in air (O2) or in NO.
FIG. 8. XRD pattern of the 6% Ru/Al2O3 catalyst. (a) Catalyst de-
activated by NO in the temperature range 400–600 C. (b) Active catalyst
◦
XRD and TPR methods were used to elucidate the reasons (working in NO/CH4 reaction mixture). (ꢂ) Ru, (ꢃ) RuO2, (ꢀ) Al2O3.

CH4/NO REACTION OVER Al-SUPPORTED Ru NANOPARTICLES
71
FIG. 9. TPR profiles of the 6% Ru/Al2O3 catalyst. (a) Catalyst pre-
pared by Ru colloid deposition on alumina, calcined at 500 C in air for 8 h
◦
FIG. 10. XRD pattern of the 12% Ru/Al2O3 catalyst. (a) Ac-
tive catalyst (working in the NO/CH4 reaction mixture). (b) Cata-
(
fresh catalyst). (b) Catalyst working in reaction mixture (active catalyst)
◦
◦
lyst partially deactivated after treatment with O2 at 600 C for 1 h. (ꢂ)
in the temperature range 400–600 C.
Ru, (ꢃ) RuO2, (ꢀ) Al2O3.
Generally, the supported Ru catalyst exhibits two TPR
peaks (12). The high-temperature peak is attributed to the
reduction of RuO2 that takes place, depending on the par-
on the partially deactivated catalyst was CO2. It is likely
that CO2 formation took place via the reduction of RuO2
by CH4. The XRD pattern of the partially deactivated
catalyst clearly demonstrates that the formation of RuO2
is mainly responsible for catalyst deactivation (Fig. 10,
spectrum b).
◦
ticle size, at T > 220 C. The fresh 6% Ru/Al2O3 catalyst
◦
◦
(
calcined at 500 C for 8 h) exhibits a large TPR peak 326 C
◦
and a smaller one at 182 C (trace a). It seems that some of
the metallicRu was oxidizedtoRuO2 during thecalcination
process. Under reaction conditions (1% NO–0.55% CH4),
the amount of RuO2 decreases as can be seen in Fig. 9,
trace b. The high-temperature peak of the active 6%
◦
The fresh 12% Ru/Al2O3 catalyst calcined at 500 C for
◦
8
h exhibits only the high-temperature TPR peak at 238 C
(
Fig. 11, trace a). Reduction of the RuO2 species of 12%
Ru/Al2O3 takes place more easily as compared with that
of 6% Ru/Al2O3 since the high-temperature TPR peak of
◦
Ru/Al2O3, located at 254 C, is smaller than that of the
calcined catalyst because the initial amount of RuO2 de-
creased under reaction conditions. However, it is clear that
the active 6% Ru/Al2O3 contains, besides the metallic Ru
◦
the former was detected at 326 C (Fig. 9, trace a). In con-
trast to 6% Ru/Al2O3, the calcined 12% Ru/Al2O3 catalyst
shows a relatively high catalytic activity for NO/CH4 reac-
tion from the beginning. Prior to the catalytic reaction, the
(
main phase), small amounts of RuO2 (see also Fig. 8, XRD
pattern of the active catalyst).
6
% Ru/Al2O3 required an activation procedure in reaction
The low-temperature TPR peaks of Fig. 9, located at
◦
mixture (1% NO–0.55% CH4) at 550 C to decrease the
◦
around 170 and 182 C respectively, have been assigned to
initial amount of RuO2.
the reduction of well-dispersed RuOx species (12). Look-
ing at the size distribution presented in Fig. 1 and the TPR
traces in Fig. 9, it becomes clear that the small Ru nanopar-
ticles of the 6% Ru/Al2O3 catalyst were oxidized to RuOx
species during calcination in air or reaction.
The XRD and TPR results strongly suggest that the de-
activation of 6% Ru/Al2O3 catalyst in an oxidizing atmo-
sphere is due to the formation of RuOx species, consisting
mainly of RuO2.
Figure 10 presents the XRD patterns of the active (spec-
trum a) and the partially deactivated (spectrum b) 12%
Ru/Al2O3 catalyst. The catalyst was only partially deac-
◦
tivated during treatment with O2 at 600 C for 1 h. The
XRD spectrum of the active catalyst revealed only the re-
flections of metallic Ru (Fig. 10, spectrum a). The catalyst
subjected to O2 treatment partially lost its catalytic activity
FIG. 11. TPR profiles of the 12% Ru/Al2O3 catalyst. (a) Catalyst cal-
cined in air at 500 C for 8 h (fresh). (b) Catalyst (active form) after working
for NO/CH4 reaction and the conversion of NO to N2 at
◦
◦
6
00 C was only 80%. The only product of CH4 oxidation for several hours in NO/CH4 reaction mixture.

7
2
BALINT, MIYAZAKI, AND AIKA
The TPR pattern of the working (active) 12% Ru/Al2O3 4.2. Considerations on the Formation of Primary
catalyst, presented in Fig. 11, trace b, exhibits both the high-
temperature (223 C) and low-temperature (205 C) reduc-
tion peaks. From this it can be seen that even the active
Reaction Products
◦
◦
There is debate concerning the mechanism of formation
of CO and H2 in POM reaction. Some studies support the
idea that CO2 and H2O are the primary reaction products of
POM, whereas other studies point out that CO and H2 are
the initial reaction products. From isotope tracing experi-
ments, performed in a closed system, it has been concluded
that syngas production over Ru/Al2O3 catalysts takes place
by total oxidation of methane followed by re-formation of
the unconverted methane with carbon dioxide and steam
to syngas (6). In fact, the specific feature of reactions per-
formed in a closed system is the long contact time (duration
ofexperiment)betweenthereactionproductsandreactants
and the catalyst. Therefore, steam reforming of methane
and carbon dioxide reactions, which are known to be slow
processes, may have a significant impact on the final distri-
bution of the reaction products. Under such experimental
conditions (close system) the formation of H2 as well as
that of CO could not be observed because, if formed, these
were readily oxidized to water and carbon dioxide, respec-
tively, during the long contact time with the catalytic bed.
In a FTIR study, Weng et al. (5) observed that on Ru/Al2O3
catalyst the formation of CO2 proceeds much earlier than
that of CO. The conclusion was that CO2 is the primary re-
action product of the O2/CH4 reaction. In reality, the above-
presented results can be interpreted differently, if we take
into consideration that the direct formation of CO and H2
on the surface of the catalyst is a fast process. Direct forma-
tion of CO and H2 sometimes is difficult to prove experi-
mentally because these primary species are readily oxidized
to CO2 and H2O, respectively. Indeed, there are studies that
favor this hypothesis. Hickman and Schmidt (13) observed
that for short contact time (0.01 s), the primary reaction
products are H2 and CO because the secondary reactions,
such as methane steam reforming and water–gas shift reac-
tions, are slow processes. The kinetic behavior of Ru/TiO2
catalyst in the partial oxidation of methane indicates the
direct formation of CO and H2, while the reforming and
water–gas shift reactions are negligible (14).
(
working) catalyst contains some amount of RuO2 and this
◦
gives the TPR peak at 223 C. The amount of RuO2 in the ac-
tive catalyst is smaller as compared with the calcined (fresh)
one (Fig. 11, trace b). The XRD pattern of the working cata-
lyst (active) did not evidence the presence of RuO2 (Fig. 10)
because the amount of RuO2 was below the sensitivity of
the XRD method.
From the results presented above it can be observed that
the catalyst with larger Ru nanoparticles (12% Ru/Al2O3)
better preserves its catalytic activity in an oxidizing atmo-
sphere because the large Ru particles are more resistant to
oxidation.
4
. DISCUSSION
4
.1. Dependence of the NO/CH4 Reaction Stoichiometry
on Temperature
Variation of the NO/CH4 reaction stoichiometry with
temperature, estimated from the activity tests, for 6 and
2% Ru/Al2O3 catalysts is presented in Table 2. Two re-
1
gions can be distinguished. (i) In the low-temperature re-
gion (450 C) the low- and high-loading Ru catalysts be-
◦
have distinctly. CH4 is selectively oxidized to CO2 over 12%
Ru/Al2O3 catalyst, whereas over 6% Ru/Al2O3, methane
is converted to both CO and CO2. (ii) In the high-
temperature region both catalysts exhibit relatively similar
behavior.
ThedatapresentedinTable2showthatthestoichiometry
of NO/CH4 reaction is strongly dependent on the reaction
temperature. Production of CO and H2 increases gradually,
in parallel with the decrease in CO2, with increasing reac-
tion temperature. A tentative explanation for these results
is given later.
TABLE 2
Thermodynamic calculations for partial oxidation of
methane with oxygen differ significantly from our experi-
mental data (1, 15). For example, at 600 C the theoretically
Variation of the NO/CH
4
Reaction Stoichiometry with
Catalysts
Temperature for 6 and 12% Ru/Al
2
O
3
◦
Temp.
predicted data for H2/CO, CO2/CO, and CH4/CO ratios are
◦
(
C)
Catalyst
Reaction stoichiometry
2.5, 1, and 0.25, respectively, whereas our experimental val-
ues for the same ratios are 1.7, 0.2, and 0 (see Figs. 2 and 3).
The significant deviation of our data from thermodynamic
equilibrium values suggests that, under our experimental
conditions(shortcontacttime, 0.06s), thedistributionofthe
reaction products is insignificantly affected by secondary
equilibrium reactions (i.e., methane steam and carbon diox-
ide reforming, water–gas shift reaction) and therefore it is
likely that both CO and H2 are the primary reaction prod-
ucts. It is clear that the high selectivity to CO and H2 at
4
50
6% Ru/Al2O3
CH4 + 3NO = 0.4CO + 0.6CO2 + 0.6H2
+
1.5N2 + 1.4H2O
1
2% Ru/Al2O3
CH4 + 3.5NO = CO2 + 0.5H2
+
1.75N2 + 1.5H2O
5
00–600
6 and 12%
Ru/Al2O3
CH4 + (1.6 . . . 2.2)NO
=
(0.6 . . . 0.85)CO + (0.4 . . . 0.15)CO2
+
(1.2 . . . 1.6)H2 + (1.1 . . . 0.8)N2
(0.8 . . . 0.4)H2O
+

CH4/NO REACTION OVER Al-SUPPORTED Ru NANOPARTICLES
73
low temperature indicates that the NO/CH4 reaction over Ru clusters (d ≥ 5 nm) are less active at low temperature
Ru/Al2O3 catalysts is occurring under nonequilibrium con- for methane dissociation and therefore methane splits into
ditions.
CHx fragments (0 ≤ x ≤ 3) and hydrogen:
CH4 + (Ru) → (Ru) –CHx + ((4 − x)/2)H2. [6]
4
.3. Methane Conversion
L
L
◦
4
.3.1. Methaneactivationatlowtemperature(450 C). It The small Ru nanoparticles (d < 5 nm) should be more ac-
is a general consensus that methane is activated on the tive for methane pyrolysis (17) and therefore methane is
catalytic surface during POM reaction by a homolytic split- activated as C and H2:
ting mechanism (3). For example, the POM reaction over
Ni/Al2O3 was reported to take place via CH4 pyrolysis (16).
The major surface species are the adsorbed C or CHx and H
CH4 + (Ru) → C–(Ru) + 2H2.
[7]
S
S
The next reaction step consists of the reaction between
CHx species and the Oads, resulting from NO decomposi-
tion. On large Ru nanoparticles the formation of CO2 takes
place by the following reaction:
(
13). After the surface reaction between C and oxygen, the
resulting CO desorbs before further oxidation to CO2 takes
place. In the case of our supported Ru catalysts it is clear
that, in the first stage of reaction, some of the initial RuO2 is
reduced and therefore the initial catalytic activity increases
progressively with time to reach a steady state. TPR (Figs. 9
and 11) as well as XRD (Fig. 10) experiments revealed the
presence of metallic Ru as well as a small amount of RuO2
for the active ruthenium catalyst. The decrease in the initial
amount of RuO2 for the calcined (fresh) catalyst during the
NO/CH4 reaction in parallel with the increase in catalytic
activity indicates that the catalytically active phase is metal-
lic Ru.
CHx + ((4 + x)/2)Oads → CO2 + (x/2)H2O.
[8]
In reality, the reaction depicted by Eq. [8] should be the
result of complex sequences of elementary reactions. The
hydrogen from CHx radicals may react with Oads to give
OH species, which then combine to form H2O (13). How-
ever, the reactions involving CHx and OH should almost
inevitably lead to H2O.
The interaction of NO (or oxygen) with metallic Ru
cluster is also an important factor in NO/CH4 reaction.
We should remember that the 6% Ru/Al2O3 catalyst, with
smaller Ru nanoparticles, is very sensitive to NO or O2 poi-
soning. The explanation is that methane dissociates to CHx
species on a metallic Ru surface but not on the oxidized
Ru surface (3). The 12% Ru/Al2O3 catalyst having larger
Ru particles is less sensitive to oxygen poisoning because
the larger the Ru particles are, the better the metallic char-
acter is preserved. Boudart (8) observed that the behavior
of metal particles larger than 5 nm resembles that of bulk
metal. Our TPR and XRD results (Figs. 8–11) prove that
catalysts rich in RuOx species (especially RuO2) show little
activity for POM reaction. The oxygen-deactivated cata-
lyst can be easily reactivated by keeping it in pure CH4 for
a few minutes. In conclusion, larger Ru clusters are more
suitable for NO/CH4 reaction because they are more resis-
tant to oxygen poisoning.
Our experimental results strongly suggest that the key
step in NO/CH4 reaction is the homolytic splitting of CH4
on the surface of Ru nanoparticles. High-resolution elec-
tron loss spectroscopy (HREELS) studies have shown that
methane can dissociate easily on the surface of Ru, yielding
methylidyne (CH), vinylidene (CCH2), ethylidyne (CCH3),
andgraphiticspecies(17). ThedegreeofCH4 dehydrogena-
tion of Ru on the surface was observed to be dependent on
temperature. Lower temperature favors the formation of
◦
CHx species, whereas at higher temperatures (T > 450 C)
only the graphitic phase is observed.
As we pointed out under Experimental, methane is se-
◦
lectively oxidized to CO2 at 450 C over 12% Ru/Al2O3
catalyst, whereas over 6% Ru/Al2O3 catalyst, methane is
converted to both CO and CO2 (Fig. 4). It is clear that the
mechanism of methane activation and COx formation is
related to the size of Ru particles and to the reaction tem-
◦
perature. The formation of CO2 at 450 C is favored over
The carbon formed on small Ru particles is selectively
oxidized to CO by the Oads resulting from NO:
the catalyst (12% Ru/Al2O3) having larger Ru [(Ru)L] par-
ticles.
On the other hand, the experimental results indicate that
the formation of CO at low temperature (450 C) is related
C–(Ru) + Oads → (Ru) + CO.
[9]
S
S
◦
mainly to the small Ru particles (d < 5 nm). The formation Oxidation of the superficial carbon species selectively gives
of CO was observed only over 6% Ru/Al2O3 catalyst which CO, because the graphitic species shows relatively low
hasanimportantfractionofsmallRunanoparticles[(RuS)]. chemical reactivity (17).
◦
The larger Ru clusters of the same catalyst are likely to be
responsible for the formation of CO2.
In conclusion, the structural effects observed at 450 C are
related to the way in which methane is activated over Ru
In light of the results presented above it is rational to con- nanoparticles. Larger Ru nanoparticles activate methane as
◦
sider that methane dissociation at 450 C over Ru nanopar- CHx and hydrogen, whereas small Ru nanoparticles (more
ticles is dependent on the size of the Ru particles. The large active) activate methane as carbon and hydrogen. The

7
4
BALINT, MIYAZAKI, AND AIKA
surface oxidation of CHx species gives carbon dioxide and
water, whereas the oxidation of bare carbon selectively
gives CO.
◦
As we will see, above 450 C the size of the metal parti-
cles is not an important parameter anymore, and therefore
the NO/CH4 reaction becomes structure insensitive. The
activity and selectivity to reaction products are relatively
the same for both catalysts investigated.
4
.3.2. Methane activation in the temperature range 500–
◦
◦
6
00 C. An increase in temperature (T > 450 C) leads to
an increase in CO production at the expense of CO2 for
both catalysts (selectivity to CO increases with increasing
temperature) (Figs. 2–5). Moreover, the evolution of CO,
CO2, and H2 with temperature is similar for both catalysts.
In the high-temperature domain, the size of the Ru particles
has no influence on the reaction selectivity to CO, CO2, and
H2. Experimental results suggest that the formation of wa-
ter isconnected tothe activation ofmethane as CHx species.
In other words, the presence of hydrogen (as CHx ) on the
surface of the metal leads to the formation of water via reac-
FIG. 12. Simplified model describing the mechanism of NO/CH4 re-
action over Ru nanoparticles supported on alumina.
(
to create surface vacant places for further decomposition
of NO). It was suggested that the relative activities of the
various catalysts reflect the ease of reduction of catalyst by
a reducing agent (20, 21).
It is believed that the reduction of NO by hydrocarbons
takes place by a two-step mechanism (22). The reaction is
initiated by one molecule of hydrocarbon reducing a small
part of the Me–O surface into a metallic surface (i.e., Pt,
Pd, Rh). Once a vacancy is created a surface explosion oc-
curs. Experimental results show that the conversion of NO
increases suddenly from zero to 100% in a narrow temper-
◦
tionbetweenCHx andOads. Ifweassumethatfor T > 450 C
methane decomposition over large Ru clusters gives CHx ,
and over small Ru clusters gives C, the simultaneous forma-
tion of CO (by the oxidation of C) and CO2 and H2O (by
the oxidation of CHx ) can be well explained. It is rational
to consider that the increase in CO and H2 production with
increasing reaction temperature is due to the increase in ac-
tivity of the large Ru clusters for methane decomposition
with increasing temperature. The ratio of hydrogen to car-
bon in CHx radicals decreases (the x coefficient of hydrogen
decreases) with increasing temperature and therefore the
◦
ature range (400–450 C) (Figs. 2 and 3).
NO is selectively converted to N2 over both Ru cata-
lysts investigated. Formation of N2O or NH3 was not ob-
served at any of the reaction temperatures. In contrast to
the Ru catalyst, the supported precious metals (Pt, Pd,
and Rh) give high selectivity to N2O in a lower tem-
perature range or to NH3 in a higher temperature range
◦
formation of CO2 and H2O is drastically reduced. At 600 C,
selectivity to CO and to H2 is around 80% for both catalysts
(
Fig. 5).
(
21, 22).
It is probable that the decomposition of CH4 on the sur-
Burch et al. (20–22) advanced the idea that, on precious
face of Ru nanoparticles is an equilibrium process. If the
concentration of methane in the reaction mixture is in-
creased, the decomposition equilibrium is shifted to the
formation of carbon and hydrogen (CH4 ↔ CHx + H2 ↔
C + H2). Experimental results confirm this assumption.
metals, N2O is formed by the reaction between NOads and
Nads, and NH3 is formed by the reaction between (CHx )ads
and Nads. The absence of N2O for supported ruthenium
catalysts can be attributed to the complete dissociation of
the adsorbed NO, because if there was any molecularly ad-
sorbed NO on the surface of Ru nanoparticles, N2O could
be produced by the reaction between NOads and Nads. Prob-
ably Nads desorbs rapidly as N2 after the decomposition of
NO. This also explains the absence of NH3 in the reaction
products for supported Ru catalysts.
When the concentration of CH4 was increased from 0.5 to
◦
≈
1%, reaction selectivity to CO and H2 at 600 C was close
to 100% and formation of water was not observed.
In summary a simplified mechanism of NO/CH4 reac-
tion over alumina-supported Ru catalyst is proposed in
Fig. 12.
5
. CONCLUSIONS
4
.4. NO Activation
NO decomposes easily after adsorption, to give Nads and
1. NO is selectively reduced by CH4 to N2 over well-
Oads on the surface of reduced metals (18). The decompo- defined Ru nanoparticles supported on γ -Al2O3 starting
◦
sition of NO leaves Oads on the surface of metal and this from 450 C.
blocks the active sites for further NO decomposition (19).
◦
2. At 450 C, CH4 is selectively oxidized by NO to CO2
The reaction is self-poisoned by oxygen. Therefore, a re- and H2O over the catalyst (12% Ru/Al2O3) with larger Ru
ductant should be used to regenerate the metal active sites nanoparticles (d ≥ 5 nm).

CH4/NO REACTION OVER Al-SUPPORTED Ru NANOPARTICLES
75
3
. CH4 is directly converted by NO to CO and H2 over
3. Pe n˜ a, M. A., Gomez, J. P., and Fierro, J. L. G., Appl. Catal. A 144, 7
1996).
4. Roh, H.-S., Jun, K.-W., Dong, W.-S., Park, S.-E., and Joe, Y.-I., Chem.
Lett. 666 (2001).
(
smallRunanoparticles(d < 5nm)inthetemperaturerange
50–600 C.
◦
4
4
. The catalyst with a larger fraction of small Ru nano-
5. Weng, W. Z., Chen, M. S., Yan, Q. G., Wu, T. H., Chao, Z. S., Liao,
particles (6% Ru/Al2O3) is easily deactivated either by air
or by NO, whereas the catalyst with larger Ru nanoparticles
12% Ru/Al2O3) is less sensitive to oxygen poisoning.
Y. Y., and Wan, H. L., Catal. Today 63, 317 (2000).
6. Guerrero-Ruiz, A., Ferreira-Apparicio, P., Bachiller-Baeza, M. B.,
and Rodriguez-Ramos, I., Catal. Today 46, 99 (1998).
(
7
8
9
. Gates, B. C., Chem. Rev. 95, 511 (1995).
. Boudart, M., J. Mol. Catal. 30, 27 (1985).
. Hayek, K., Kramer, R., and Pa a´ l, Z., Appl. Catal. A 162, 1 (1997).
5
. The key point of the reaction mechanism is CH4 acti-
◦
vation over Ru nanoparticles. (i) At 450 C, CH4 molecules
split into CHx fragments and H2 (0 < x ≤ 3) over large Ru
10. Balint, I., Miyazaki, A., and Aika, K., Chem. Mater. 11, 378 (1999).
naparticles. CHx species are further selectively oxidized by 11. Balint, I., Miyazaki, A., and Aika, K., Chem. Mater. 13, 932 (2001).
12. Betancourt, P., Rives, A., Hubaut, R., Scott, C. E., and Goldwasser J.,
Oads to CO2 and H2O. (ii) CH4 splits into C and H2 over
small Ru particles. The C is further oxidized by Oads to CO.
Appl. Catal. A 170, 307 (1998).
3. Hickman, D. A., and Schmidt, L. D., J. Catal. 138, 267 (1992).
4. Elmasides C., Ioannides T., and Verykios X. E., AIChE J. 46, 1260
1
1
(
2000).
ACKNOWLEDGMENT
1
1
5. Zhu, J., Zhang, D., and King, K. D., Fuel 80, 899 (2001).
6. Lu, Y., Xue, J., Yu, C., Liu, Y., and Shen, S., Appl. Catal. A 174, 121
A research grant from the Japanese Society for the Promotion of Science
No. P00136) is greatly appreciated.
(
1998).
(
1
1
7. Choudhary, T. V., and Goodman, D. W., J. Mol. Catal. 163, 9 (2000).
8. Comrie, C. M., Weinberg, W. H., and Lambert, R. M., Surf. Sci. 57, 619
REFERENCES
(1976).
1
9. Walker, A. V., Gruyters, M., and King, D. A., Surf. Chem. 384, L791
1
. Tsang, S. C., Claridge, J. B., and Green, M. L. H., Catal. Today 23, 3
1995).
. Miyazaki, A., Balint, I., Aika, K., and Nakano, Y., J. Catal. 204, 364 21. Burch, R., and Ramli, A., Appl. Catal. B 15, 63 (1998).
2001).
22. Burch, R., and Millington, P. J., Catal. Today 26, 185 (1995).
(1997).
(
20. Burch, R., and Ramli, A., Appl. Catal. B 15, 49 (1998).
2
(