In situ Raman study on the partial oxidation of methane to synthesis gas over Rh/Al2O3 and Ru/Al2O3 catalysts

Article abstract of DOI:10.1016/j.jcat.2008.03.009

In situ microprobe Raman and XRD techniques were used to follow the oxidation state of Rh/Al₂O₃ and Ru/Al₂O₃ catalysts during the catalytic ignition process of the partial oxidation of methane (POM) to synthesis gas. It was found that the catalyst was in the fully oxidized state before ignition of the POM reaction, and abruptly changed its oxidation state at the temperature at which the POM reaction started. After the POM reaction was ignited, the amount of Rh₂O₃ or RuO₂ in the catalyst at the entrance of the catalyst bed was below the detection level of Raman spectroscopy. Due to the greater M{single bond}O bond strength of Ru{single bond}O compared with Rh{single bond}O, Ru/Al₂O₃ demonstrated a greater tendency to oxidize than Rh/Al₂O₃ under the POM conditions. This factor affects the oxygen coverage on the two catalysts under reaction conditions and consequently affects the pathways of synthesis gas formation.

Full text of DOI:10.1016/j.jcat.2008.03.009

Journal of Catalysis 256 (2008) 192–203
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
In situ Raman study on the partial oxidation of methane to synthesis gas over
Rh/Al2O3 and Ru/Al2O3 catalysts
∗
Ying Liu, Fei-Yang Huang, Jian-Mei Li, Wei-Zheng Weng , Chun-Rong Luo, Mei-Liu Wang, Wen-Sheng Xia,
∗
Chuan-Jing Huang, Hui-Lin Wan
State Key Laboratory of Physical Chemistry of Solid Surfaces and Department of Chemistry, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen, 361005, China
a r t i c l e i n f o
a b s t r a c t
Article history:
In situ microprobe Raman and XRD techniques were used to follow the oxidation state of Rh/Al2O3 and
Ru/Al2O3 catalysts during the catalytic ignition process of the partial oxidation of methane (POM) to
synthesis gas. It was found that the catalyst was in the fully oxidized state before ignition of the POM
reaction, and abruptly changed its oxidation state at the temperature at which the POM reaction started.
After the POM reaction was ignited, the amount of Rh2O3 or RuO2 in the catalyst at the entrance of the
catalyst bed was below the detection level of Raman spectroscopy. Due to the greater M–O bond strength
of Ru–O compared with Rh–O, Ru/Al2O3 demonstrated a greater tendency to oxidize than Rh/Al2O3 un-
der the POM conditions. This factor affects the oxygen coverage on the two catalysts under reaction
conditions and consequently affects the pathways of synthesis gas formation.
Received 25 December 2007
Revised 11 March 2008
Accepted 13 March 2008
Available online 25 April 2008
Keywords:
Methane partial oxidation
Synthesis gas
Rh/Al2O3
Ru/Al2O3
© 2008 Elsevier Inc. All rights reserved.
In situ Raman
In situ X-ray powder diffraction
Reaction mechanism
1. Introduction
pleted within a very narrow zone near the entrance of the catalyst
bed [20]. As a result, the catalyst in various parts of the reactor
Catalytic partial oxidation of methane (POM) over supported
metal catalysts is one of the most promising alternatives to the
conventional stream-reforming process [1] for synthesis gas pro-
duction, particularly in chemical plants of medium size [2–4]. The
elucidation of the reaction pathways in POM to synthesis gas over
supported transition metal catalysts is a major challenge in the
study of this reaction [5]. Although numerous attempts have been
made to better understand the mechanisms of synthesis gas forma-
tion, the reaction pathway remains under debate [6–18]. Despite
the differences in the reaction mechanisms, it is generally agreed
that the products of methane oxidation depend on the type and
amount of oxygen species on the catalyst [11–18]. Combustion of
CH4 to CO2 and H2O is preferred on a metal oxide or metal surface
with high oxygen concentrations [11,12,14–18], whereas reduced
metal sites are active for dissociative activation of CH4 to surface
carbon species (CHx, x = 0–3) [12–14,16,19], the first step in the
CH4 partial oxidation to synthesis gas by the direct mechanism.
Thus, the reaction pathways of POM to synthesis gas are closely
related to the oxidation state of the catalyst under the reaction
conditions.
may experience different chemical states, and the oxidized metal
site is found only at the top of the catalyst bed. By monitoring the
structure of the catalysts under reaction conditions (especially at
the front end of the catalyst bed, where O2 is still available), we
should be able to distinguish between the two proposed mecha-
nisms for the POM reaction.
Alumina is widely used as a catalyst support in the POM re-
action. An earlier investigation of rhodium-on-alumina catalysts
◦
calcined at 600 and 900 C indicated that calcination of Rh/Al2O3
◦
at 900 C may bring about reaction of the rhodium with the alu-
mina support, leading to an appreciable drop in the reducibility
of the rhodium species that had reacted with the alumina and
consequently affecting the catalyst’s performance in the POM re-
action [21].
As a continuation of the previous research, here we report a
more detailed investigation of the Rh/Al2O3 and Ru/Al2O3 catalysts
under a simulated POM reaction feed, using both in situ micro-
probe Raman and in situ XRD techniques. The study focused on
the oxidation state of the noble metal species in the catalysts
during the catalytic ignition process of the POM reaction and its
relationship with the reaction pathways of methane. Some paral-
lel measurements on the SiO2-supported Rh and Ru catalysts, as
well as on pure Rh and Ru metal powders, were performed for
comparison. Because this study did not concentrate on the reac-
tion of the active metal (oxide) species with the Al2O3, the focus
For the POM reaction carried out in a fixed-bed reactor using
CH4 and O2 in a stoichiometric ratio, O2 in the feed usually is de-
Corresponding authors. Fax: +86 592 2183047.
E-mail addresses: wzweng@xmu.edu.cn (W.-Z. Weng), hlwan@xmu.edu.cn
*
◦
(H.-L. Wan).
was on catalysts thermally pretreated at 110 and 600 C. Hopefully,
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.03.009

Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
193
CH4(in) − CH4(out)
Table 1
XCH
=
× 100%,
× 100%,
4
CH4(in)
BET surface area and metal dispersion of the catalysts
*
H2
Sample
BET surface
Metal dispersion
Mred(600)/Mtotal
× 100%
SH
=
2
−1
)
area (m2
g
2(CH4(in) − CH4(out))
O/Mred(600) CO/Mred(600)
CO
Al2O3
Al2O3-600
249
182
231
219
184
186
231
/
/
/
SCO
SCO
=
× 100%,
/
/
/
CO + CO2
1 wt% Rh/Al2O3-110
3 wt% Rh/Al2O3-110
1 wt% Rh/Al2O3-600
3 wt% Rh/Al2O3-600
1 wt% Ru/Al2O3-110
3.7 wt% Ru/Al2O3-110 215
1 wt% Ru/Al2O3-600 186
3.7 wt% Ru/Al2O3-600 164
87.4a
97.2a
59.8a
80.9a
89.2a
>99.0a
85.6b
96.3b
/
0.91
0.52
1.1
0.69
0.75
0.31
0.05
0.02
/
1.2
0.72
1.3
0.93
0.70
0.32
0.02
0.01
/
CO2
=
× 100%.
2
CO + CO2
◦
The pulse reaction of CH4 was carried out at 600 C in a fixed-
bed quartz tube reactor with 50.0 mg of the supported catalyst
or 10 mg of the noble metal oxide powder (made by calcinating
◦
RhCl3•nH2O or RuCl3•nH2O powder in air at 600 C for 4 h) us-
SiO2-600
1 wt% Rh/SiO2-600
1 wt% Ru/SiO2-600
491
488
495
ing He containing a few ppm of O2 as carrier gas. After reduction
∼100a
0.35
0.04
0.23
0.02
◦
∼100b
with H2 at 600 C for 30 min, the catalyst was purged with carrier
gas (30 mL/min) at the same temperature until the baseline of
the mass spectrometer was flat. The CH4 pulses (99.9%, ∼80 μL, at
room temperature) were then admitted to the reactor every 6 min.
In between the pulses of CH4, the catalyst was reoxidized by the
oxygen present in the carrier gas that was passed continuously
through the catalyst bed. The reaction products were analyzed by
an online Balzers OmniStar quadrupole mass spectrometer (model
QMS 200).
*
Mred(600)/Mtotal: Fraction of the metal species reducible by H2 at temperature
◦
below 600 C. The data were calculated based on the H2-TPR profiles of:
a
re-oxidized, and
fresh catalysts.
b
this investigation will provide insight into the mechanisms of POM
for synthesizing gas over supported noble metal catalysts.
2. Experimental
2.3. Catalyst characterization
2.1. Catalyst preparation
X-ray powder diffraction (XRD) analysis was carried out by
◦
a Panalytical X’pert PRO diffractometer scanning 2θ from 10 to
The Al2O3- and SiO2-supported noble metal catalysts with the
metal loadings specified in Table 1 were prepared by the im-
pregnation method using an aqueous solution of RhCl3•nH2O or
RuCl3•nH2O (Sino-Platinum Metals Co. Ltd.). The concentration of
the noble metal in the solution was determined by the inductively
coupled plasma (ICP) method. To ensure that all of the metal salt
was taken up by the support during the impregnation, the no-
ble metal solution containing the required amount of Rh or Ru
was diluted with distilled water to a volume beyond which the
impregnated catalyst began to appear wet. The Al2O3 (TL-02, 60–
80 mesh, ∼250 m2/g, Guizhou Alumina Factory, China) and SiO2
(Aldrich 60A, 60–80 mesh, ∼490 m2/g) supports were calcined at
◦
80 . Cu-Kα radiation obtained at 40 kV and 30 mA was used
as the X-ray source. The in situ XRD experiment was also car-
ried out with the same instrument using a XRK-900 reactor. In
each in situ XRD experiment, the catalyst sample was first reduced
◦
with H2/Ar = 5/95 (volume ratio) at 400 C (for Ru/Al2O3-110) or
◦
600 C (for Rh/Al2O3-110) before switching to the reaction feed
(CH4/O2/Ar = 2/1/197) at room temperature. The temperature of
◦
the reactor was then raised from 25 to 750 C. During this process,
the XRD patterns were collected at prespecified temperatures.
The BET surface area of the catalyst was measured by N2 ad-
◦
sorption at −196 C using a Micromeritics Tristar 3000 instrument.
◦
Before the measurements, the samples were degassed at 300 C
◦
400 C for 4 h before use. After water was evaporated, the Al2O3
for 2 h. The dispersion of the metal was determined by O2 and
supported catalyst was divided into two parts. The first part of the
◦
CO chemisorption at 35 C by assuming O/M and CO/M = 1 sto-
◦
sample was dried at 110 C (designated M/Al2O3-110, M = Rh or
ichiometry. In the experiments, performed using a Micromeritics
ASAP 2010 instrument, the catalyst sample (0.4–1.2 g) was first
◦
Ru), and the second part was calcined at 600 C in air for 4 h (des-
ignated M/Al2O3-600, M = Rh or Ru). All of the SiO2 supported
◦
◦
reduced with H2/Ar = 5/95 at 600 C (or 400 C for M/Al2O3-110
◦
catalysts were calcined at 600 C in air for 4 h (designated M/SiO2-
◦
◦
samples) for 30 min, followed by evacuation at 450 C (or 350 C
for M/Al2O3-110 samples) for 30 min. The sample was then cooled
to 35 C under vacuum for O2 (99.999%) or CO (99.999%) adsorp-
600, M = Rh and Ru).
◦
2.2. Catalytic performance
tion.
The H2 temperature-programmed reduction (H2-TPR) experi-
ments were performed with a GC-TPR apparatus. For each sample,
two TPR experiments were performed. The fresh catalyst (0.1 g)
was first heated in a flow of H2/Ar = 5/95 mixture (50 mL/min)
The catalytic performance of the catalysts was studied using a
CH4/O2/Ar = 2/1/45 mixture (volume ratio) as a reaction feed. The
reaction was carried out in a fixed-bed quartz tube reactor (5 mm
i.d.) at atmospheric pressure. The catalyst (15.0 mg) was reduced
◦
◦
from 5 to 400 C (for Ru/Al2O3) or 600 C (for Rh/Al2O3, Rh/SiO2,
◦
◦
◦
with H2 at 400 C (for Ru/Al2O3-110) or 600 C (for Ru/Al2O3-
600, Rh/Al2O3-110, Rh/Al2O3-600, and SiO2-supported catalysts)
for 30 min before switching to the reaction feed at room tem-
perature. The temperature of the furnace was then raised from
Ru/SiO2) at the rate of 10 C/min to obtain the TPR profile of the
◦
fresh catalyst. After reduction with H2/Ar at 400 or 600 C for
30 min, the reduced catalyst was reoxidized with an O2/Ar = 5/95
◦
mixture (volume ratio, 20 mL/min) at 400 C for 30 min, then
◦
◦
◦
room temperature to 600 C at a rate of 10 C/min. During this
process, the performance of the catalyst was evaluated at selected
temperatures. The products were analyzed by an online gas chro-
matograph equipped with a thermal conductivity detector (TCD)
using Ar (99.999%) as the carrier gas and a carbon sieve column
(1.5 m) for the separation of CH4, O2, H2, CO, and CO2. The conver-
cooled to 5 C under O2/Ar atmosphere. The sample was then
◦
purged with a H2/Ar = 5/95 mixture at 5 C for about 10 min until
the baseline was flat. The TPR profile of the reoxidized catalyst was
◦
obtained by heating the treated sample from 5 to 900 C at a rate
◦
of 10 C/min in a flow of a H2/Ar = 5/95 mixture (50 mL/min). The
◦
effluent gas mixture was passed through a cold trap at ∼−60 C
sion of CH4 (XCH ) and the selectivities of CO (SCO) and H2 (SH
)
to remove water. The hydrogen consumption was monitored by
a TCD. The results of the second H2-TPR experiments were also
4
2
were calculated based on the following equations:

194
Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
used to calculate the fraction of the metal species reducible by H2
below 600 C (Mred(600)/Mtotal) in the Rh/Al2O3, Ru/Al2O3-110, and
Table 2
◦
Dependence of catalytic performance on the reaction temperature for the Rh/Al2O3
catalysts
Rh/SiO2-600 catalysts. To calculate the fraction of the metal species
◦
Catlalysts
Furnace temp.
( C)
Conv. (%)
Sel. (%)
H2/CO
reducible by H2 at temperature below 600 C in the ruthenium cat-
◦
◦
alysts calcined at 600 C, three additional H2-TPR experiments (at
5–900 C) were performed on the fresh Ru/Al2O3-600 and Ru/SiO2-
CH4
O2
CO
H2
CO2
◦
1 wt% Rh/Al2O3-600 350
0.8
2.0
8.0
2.0
6.2
33.3
0
0
0
0
0
0
100
100
100
/
/
/
400
460
470
500
550
600
600 catalysts.
The O2 temperature-programmed desorption (O2-TPD) experi-
ments were performed with a Micromeritics AutoChem II 2920
instrument. The catalyst (∼0.2 g) was first reduced with a H2/Ar =
51.9
58.9
71.8
82.6
100
49.3 77.3
58.4 79.8
76.0 82.1
87.7 83.4
50.7 3.1
100
100
100
41.6 2.7
24.0 2.2
12.3 1.9
◦
5/95 mixture at 600 C for 30 min, followed by oxidation with
◦
O2/He = 1/4 (volume ratio) at 500 C for 30 min and cooling to
1 wt% Rh/Al2O3-110 350
0.46
1.5
5.7
2.0
0
0
0
0
0
0
0
0
100
/
/
/
/
◦
20 C under the same atmosphere. The sample was then purged
400
450
460
470
480
500
550
600
6.9
23.2
32.6
100
100
100
◦
with He (99.999%) at 20 C for 30 min, then heated from 20 to
◦
◦
8.1
1100 C at a rate of 10 C/min in a flow of He (50 mL/min). The
O2 desorption was analyzed with an online ThermoStar quadrupole
mass spectrometer (model GSD301T2).
52.7
54.5
58.8
71.7
82.5
100
49.4 75.8
53.0 75.8
60.1 78.5
76.1 78.2
88.4 82.1
50.6 3.1
47.0 2.9
39.9 2.6
23.9 2.1
11.6 1.9
100
100
100
100
The in situ Raman characterization was performed with Ren-
ishaw R1000 and Dilor LabRam
I microprobe Raman systems
equipped with CCD detectors using a home-built high-temperature
in situ Raman cell designed for the microprobe Raman spectrom-
eter. A diagram of this Raman cell is available elsewhere [21].
The samples were placed in a sample holder equipped with a
thermocouple placed below the holder for temperature measure-
ment. The excitation wavelength was a 325-nm He–Cd laser (for
Rh/Al2O3 catalyst) with a power of ∼3 mW or a 632.8-nm He–Ne
laser (for Ru/Al2O3 catalyst) with a power of ∼5 mW measured
at the analysis spot. The microscope attachment for the Renishaw
R1000 system is based on a Leica DMLM system using an OFR
LMU-15×-NUV objective (for a 325-nm laser); that for the Dilor
LabRam I is based on an Olympus BX40 system using an Olym-
pus 50× objective (for a 632.8-nm laser). The Raman spectra were
3 wt% Rh/Al2O3-600 350
1.0
2.4
9.7
48.0
49.9
61.0
73.3
84.1
3.5
11.6
41.0
0
0
0
0
0
0
100
100
100
60.9 3.8
56.8 3.5
37.7 2.6
22.5 2.1
11.1 1.9
/
/
/
400
430
440
450
500
550
600
100
39.1 74.3
43.2 76.1
62.3 81.8
77.5 82.2
88.9 82.4
100
100
100
100
3 wt% Rh/Al2O3-110 350
1.2
4.7
8.7
45.7
47.9
49.4
60.6
73.0
84.0
3.6
18.0
35.2
0
0
0
0
0
0
100
100
100
65.5 4.3
61.9 3.9
59.0 3.7
39.7 2.7
23.8 2.1
11.9 1.9
/
/
/
400
420
430
440
450
500
550
600
100
34.5 73.8
38.1 73.9
41.0 76.0
60.3 80.9
76.2 81.3
88.1 82.4
100
100
100
100
100
−1
−1
(632.8 nm) spectral
measured with 6 cm
(325 nm) and 4 cm
resolutions. In each experiment, the catalyst sample was first re-
◦
◦
duced with H2/Ar (5/95) at 400 C (for Ru/Al2O3-110) or 600 C (for
Ru/Al2O3-600 and Rh/Al2O3) before switching to the reaction feed
(CH4/O2/Ar = 2/1/45) at room temperature. The temperature of the
−1
Reaction condition: CH4/O2/Ar = 2/1/45, mcat. = 15 mg, SV = 2.0 × 105 mL h
g
.
−1
◦
Catalysts were reduced with H2 at 600 C before testing. Data were collected after
20 min on stream.
◦
Raman cell was then raised from room temperature to 600 C. The
Raman spectra of the catalyst were recorded at selected tempera-
tures during this process.
the fractions of the noble metal species in the catalysts reducible
by H2 at temperatures below 600 C (Mred(600)/Mtotal), based on
◦
the results of H2-TPR experiments. Compared with the metal dis-
persion data obtained by O2 adsorption, CO adsorption was sig-
nificantly greater over Rh/Al2O3 catalysts, with CO/Rh ratios > 1
in some samples, possibly due to the adsorption of CO in the
gem-dicarbonyl form [22–25]. The metal dispersion of Ru/Al2O3
decreased significantly after the catalysts were calcined in air at
600 C. Comparatively speaking, the dispersion of Rh on Al2O3 was
not significantly affected by calcination. These findings are consis-
tent with the results of XRD characterization given earlier.
3. Results
3.1. Bulk structure, surface area, and metal dispersion of the catalysts
Fig.
1
shows the XRD patterns of the fresh catalysts. For
◦
◦
the Al2O3-supported catalysts dried at 110 C (Rh/Al2O3-110 and
Ru/Al2O3-110) and the Rh/Al2O3 catalysts calcined at 600 C
(Rh/Al2O3-600), the XRD patterns show only the diffraction peaks
◦
due to boehmite (AlO(OH)) (PDF No. 00-021-1307) and/or γ -
Al2O3 (PDF No. 00-010-0425). For the Ru/Al2O3 calcined at 600 C
◦
3.2. Catalytic performance of the catalysts at different temperatures
(Ru/Al2O3-600), however, the diffraction pattern of RuO2 (PDF
No. 00-040-1290) also can be observed. For the 1 wt% Rh/SiO2
Tables 2 and 3 characterize the catalytic performance of the
Rh/Al2O3 and Ru/Al2O3 catalysts at various temperatures dur-
ing stepwise heating of the previously reduced catalysts under
◦
catalyst calcined at 600 C, a broad diffraction peak of high inten-
◦
sity with 2θ at 22.5 and a broad peak of weak intensity with
◦
◦
2θ at about 34 can be seen. The former can be assigned to the
CH4/O2/Ar = 2/1/45 flow from 30 to 600 C. For comparison, the
diffraction of SiO2, and the latter may be diffraction lines from the
(104) and (110) faces of the Rh2O3 (PDF No. 00-042-0541) micro-
catalytic performance of 1 wt% Rh/SiO2-600 and 1 wt% Ru/SiO2-
600 catalysts is presented in Table 4. For all of the catalysts,
three temperature regions can be distinguished from the evolu-
tion of CH4 conversion products (CO2, CO, and H2). In the low-
temperature region, O2 in the reactant was not completely con-
sumed. Oxidation of CH4 over the catalysts started at between 300
◦
crystals. For the 1 wt% Ru/SiO2 sample calcined at 600 C, both
diffraction patterns of SiO2 and RuO2 (PDF No. 00-040-1290) are
clearly visible.
The BET surface areas and metal dispersion data (measured
◦
◦
by O2 and CO chemisorption at 35 C) of the catalysts are given
and 350 C, depending on the amount of metal loading, but no
in Table 1. The surface areas of the catalysts are very close to
those of the Al2O3 and SiO2 supports. The metal dispersion data
(O/Mred(600) and CO/Mred(600)) given in Table 1 were calibrated with
CO or H2 was detected in the product. The CH4 conversion in this
temperature region was low and increased slowly with increasing
temperature. In the mid-temperature region (370–530 C), the POM
◦

Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
195
Fig. 1. XRD patterns of the fresh catalysts. a. 1 wt% Rh/Al2O3-600, b. 3 wt% Rh/Al2O3-600, c. 1 wt% Rh/Al2O3-110, d. 3 wt% Rh/Al2O3-110, e. 1 wt% Ru/Al2O3-600, f. 3.7 wt%
Ru/Al2O3-600, g. 1 wt% Ru/Al2O3-110, h. 3.7 wt% Ru/Al2O3-110, i. 1 wt% Rh/SiO2-600, j. 1 wt% Ru/SiO2-600.
reaction over the catalysts (which can be determined experimen-
tally by the rapid increase in CH4 and O2 conversion accompanied
by H2 and CO formation) was seen to ignite. On both Rh/Al2O3
and Ru/Al2O3, the temperatures required to ignite the POM reac-
tion decreased with increasing metal loading. Similar results have
been reported by Wang and Ruckenstein for the POM reaction on
Rh/Al2O3 catalysts [26]. In the high-temperature region after the
POM reaction was ignited, O2 conversion was close to 100%. The
amounts of CO and H2 produced showed an increasing trend with
temperature, whereas the formation of CO2 exhibited an opposite
trend. Moreover, even though the reaction mixture (at CH4/O2 = 2)
used in the experiment had been diluted with a large amount of
Ar, the observed CH4 conversions and CO selectivities for the POM
reaction were still higher than the thermodynamic equilibrium val-
ues for the reaction reported previously [27] at the temperatures
specified in Table 2, indicating that a hot-spot layer still existed in
the catalyst bed under the experimental conditions.
As shown in Tables 2 and 3, the temperatures required to ig-
nite the POM reaction on Rh/Al2O3 were always lower than those
on Ru/Al2O3 at comparable metal loadings. A similar phenomenon
was observed on the SiO2-supported catalysts (Table 4). Moreover,
for the Rh and Ru catalysts with comparable metal loadings, the
temperature required to ignite the POM reaction was much lower
on the SiO2-supported catalyst than that on the Al2O3-supported
catalyst, indicating a significant influence of the support on the
performance of the metal species. The stronger interaction of Rh
and Ru with Al2O3 compared with SiO2 should be responsible for
the higher POM ignition temperatures on the Al2O3-supported cat-
alysts. The much lower interaction of rhodium with SiO2 is evident
from the XRD patterns shown in Fig. 1. The weak diffraction max-
imum due to Rh2O3 seen on the silica support indicates that the
Rh/SiO2-600 catalyst contained larger rhodium particles.
3.3. H2-TPR and O2-TPD characterization
H2-TPR and O2-TPD were used to investigate the reducibility
and the metal–oxygen bond strength of the metal oxide species
on the supported metal catalysts. The corresponding results are
shown in Figs. 2 and 3. The quantified amounts from the TPR pro-
files for the catalysts are specified in Table 1. The TPR profiles of
Rh/Al2O3 are composed of two major reduction bands with tem-
peratures at peak maxima (Tm) of below 110 and above 750 C,
respectively. These bands can be attributed to the reduction of
rhodium oxide (mainly amorphous Rh2O3) species (Tm < 110 C)
◦
◦
with different extent of interaction with Al2O3 and the Rh(AlO2)y
◦
species (Tm > 750 C) formed by diffusion of rhodium oxides into
◦
sublayers of Al2O3 structure after high-temperature (>500 C) oxi-
dation [28]. Due to the interaction between Rh and Al2O3 [28,29],
the fractions of the oxidized Rh species in the Rh/Al2O3 catalysts
irreducible at temperatures below 600 C increased significantly
in the samples with low (1 wt%) Rh loading as well as in sam-
◦
◦
ples calcined at high temperature (600 C). The position of the
◦
low-temperature TPR peak (Tm = 98–105 C) shifted toward the
low-temperature direction in the high-Rh loading samples.
The TPR profiles of Ru/Al2O3 are composed of one or two rel-
atively sharp low-temperature reduction peaks with Tm of 81–96
◦
and 157–170 C, respectively, and a broad high temperature reduc-
◦
tion band with Tm above 700 C. According to the literature [30–
32], the peaks with Tm around 90 C can be assigned to the reduc-
◦
tion of well-dispersed RuOx species containing mainly RuO2, and
◦
the peaks with Tm around 160 C can be attributed to the reduc-
◦
tion of bulk RuO2 species. The broad TPR band with Tm > 700 C
may be assigned to the reduction of the oxidized Ru species that
interact strongly with Al2O3. Similar to the Rh/Al2O3 catalysts, the
fraction of the oxidized Ru species in the Ru/Al2O3 catalysts irre-
◦
ducible by H2 at temperatures below 600 C also increased with

196
Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
◦
◦
Fig. 2. H2-TPR profiles of the catalysts. Before the experiment, the catalysts were reduced with H2/Ar = 5/95 mixture at 400 C (for Ru/Al2O3) or 600 C (for Rh/Al2O3) for
◦
30 min followed by oxidation with O2/Ar = 5/95 mixture at 400 C for 30 min.
◦
increasing calcination temperature and decreased with increasing
Ru loading (Table 1). Compared with the TPR profiles of Ru/Al2O3-
110, the intensities of the TPR signals for Ru/Al2O3-600 were signif-
icantly lower, possibly because of the significantly smaller particle
size of the ruthenium species in the former catalyst. As demon-
strated by the metal dispersion data in Table 1, the particle size
of the ruthenium species in Ru/Al2O3 increased significantly after
low 600 C. Due to the strong interaction between rhodium oxide
species and Al2O3, the O2 desorption temperatures on Rh/Al2O3-
◦
600 (880, 985 C) were significantly higher than those on Rh/SiO2-
600 (820, 860 C), in agreement with the results of H2-TPR ex-
periments presented earlier. The O2-TPD profile of Ru/Al2O3-600
or Ru/SiO2-600 comprises only a single desorption peak with Tm
◦
◦
above 1024 C. Compared with the Rh catalysts, the O2 desorption
◦
temperatures on Ru/Al2O3-600 and Ru/SiO2-600 are very close,
possibly due to the weaker interaction between RuO2 and Al2O3
compared with that between Rh2O3 and Al2O3.
calcination at 600 C. Because larger metallic Ru particles better
preserve their metallic state compared with well-dispersed species
[33], only the surface of large metallic Ru particles in Ru/Al2O3-600
are oxidized during the process of reoxidation treatment before
the TPR experiment, whereas most of the small metallic Ru par-
ticles in Ru/Al2O3-110 can be converted to RuO2 by reoxidation
3.4. High-temperature in situ Raman and XRD characterizations
◦
at 400 C. For the SiO2-supported catalysts, TPR peaks with Tm at
In situ Raman and in situ XRD measurements were performed to
gain insight into the chemical states of the noble metal species in
the catalyst during the catalytic ignition process of POM reaction
and its relationship with the performance of the catalysts char-
acterized in Tables 2 and 3. In these experiments, in situ Raman
spectroscopy was used to monitor the oxidation state of the noble
metal species in the catalysts located at the entrance of catalyst
bed, whereas in situ XRD was used to analyze the bulk structure
of the catalysts during the catalytic ignition process of the POM
reaction.
◦
49 and 113 C were observed on 1 wt% Rh/SiO2-600 and 1 wt%
Ru/SiO2-600, respectively, and no reduction peak was observed in
the high-temperature region. Compared with the Al2O3-supported
Rh and Ru catalysts, reduction of the Rh2O3 and RuO2 species on
SiO2 occurred at lower temperatures, also due to the weaker inter-
action of Rh and Ru with SiO2.
Fig. 3 shows the O2-TPD profiles of the Al2O3- and SiO2-
supported Rh and Ru catalysts. The O2-TPD profiles of Rh/Al2O3-
600 and Rh/SiO2-600 consist of two O2 desorption peaks with
◦
◦
maxima at 880, 985 C and 820, 860 C, respectively, attributable
to two kinds of rhodium oxide species in the samples. Based on
our previous O2-TPD and H2-TPR characterizations on Rh/Al2O3
and Rh/SiO2 catalysts [21], the O2 desorption peaks for the Rh cat-
alysts shown in Fig. 3 can be assigned to the decomposition of
the rhodium oxide species reducible by H2 at temperatures be-
Fig. 4 shows the Raman spectra of 3 wt% Rh/Al2O3-600 and
3.7 wt% Ru/Al2O3-600 catalysts recorded at 600 C under O2,
◦
H2/Ar = 5/95, and a simulated POM feed with CH4/O2/Ar = 2/1/45.
In the spectra recorded under O2 (Fig. 4a), a broad band with
−1
maximum at ∼550 cm
corresponding to amorphous Rh2O3 [34]
was observed on 3 wt% Rh/Al2O3-600, and two relatively sharp

Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
197
Table 3
Table 4
Dependence of catalytic performance on the reaction temperature for the Ru/Al2O3
catalysts
Dependence of catalytic performance on the reaction temperature for the Rh/SiO2
and Ru/SiO2 catalysts
Catlalysts
Furnace temp.
( C)
Conv. (%)
CH4 O2
0.44
0.65
1.9
Sel. (%)
H2/CO
Catlalysts
Furnace temp.
( C)
Conv. (%)
Sel. (%)
H2/CO
◦
◦
CO
H2
CO2
CH4
O2
CO
H2
CO2
1 wt% Ru/Al2O3-600
300
350
400
450
500
520
530
550
600
1.8
2.8
7.4
0
0
0
0
0
0
0
0
0
0
0
0
100
100
100
100
100
100
/
/
/
/
/
/
1 wt% Rh/SiO2-600 300
2.5
11.6
22.2
25.0
29.2
39.0
53.1
68.0
81.6
8.3
45.9
89.2
97.5
0
0
0
0
0
2.3
7.6
100
100
100
/
/
/
350
360
370
400
450
500
550
600
4.8
20.0
46.1
65.2
100
100
100
4.2
95.8 3.7
11.2
16.2
62.0
68.9
80.9
100
16.7 26.4
41.3 53.4
59.0 72.1
75.5 82.9
87.7 84.6
83.3 3.2
58.7 2.6
41.0 2.4
24.5 2.2
12.3 1.9
100
100
100
100
64.3 80.7
73.1 82.0
85.9 82.5
35.7 2.5
26.0 2.2
14.1 1.9
1 wt% Ru/Al2O3-110
300
350
400
450
500
510
550
600
0
0.83
2.3
0
3.2
8.2
28.5
71.8
100
100
100
0
0
0
0
0
0
0
0
0
0
0
100
100
100
100
43.1 2.8
24.8 2.2
12.5 1.9
/
/
/
/
/
1 wt% Ru/SiO2-600 300
0
0.77
0
2.4
10.5
39.7
100
100
100
100
0
0
0
0
0
0
0
0
100
100
/
/
350
400
440
450
500
550
600
3.4
9.9
37.0
47.7
60.7
72.1
/
7.7
0.23 100
/
17.7
58.2
71.1
82.5
20.2 56.8
41.7 71.9
64.4 78.3
79.4 79.7
79.8 5.6
58.3 3.4
35.6 2.4
20.6 2.0
56.8 79.9
75.2 83.0
87.5 83.0
−1
−1
.
3.7 wt% Ru/Al2O3-600 300
0.74
1.6
4.1
10.4
16.1
50.2
56.9
71.3
82.8
2.5
0
0
0
0
0
0
0
0
0
0
100
/
/
/
/
/
Reaction condition: CH4/O2/Ar = 2/1/45, mcat. = 15 mg, SV = 2.0 × 105 mL h
g
◦
350
400
450
470
480
500
550
600
6.4
16.5
43.8
64.9
100
100
100
100
56.1 3.5
44.9 2.9
24.9 2.2
12.2 1.9
Catalysts were reduced with H2 at 600 C before testing. Data were collected after
20 min on stream.
100
43.9 76.6
55.1 80.5
75.1 82.3
87.8 82.4
100
100
100
3.7wt% Ru/Al2O3-110
300
350
400
450
460
500
550
600
1.0
2.6
7.3
22.3
44.9
59.
72.4
83.9
3.8
0
0
0
0
0
0
0
0
100
/
/
/
/
11.3
29.2
89.0
100
100
100
64.7 4.07
42.7 2.84
23.1 2.14
11.3 1.85
100
35.3 71.9
57.3 81.3
76.9 82.4
88.7 82.1
100
100
100
−1
−1
.
(for
Reaction condition: CH4/O2/Ar = 2/1/45, mcat. = 15 mg, SV = 1.2 × 105 mL h
g
◦
◦
Catalysts were reduced with H2 at 400
C
(for Ru/Al2O3-110) and 600
C
Ru/Al2O3-600), respectively, before testing. Data were collected after 20 min on
stream.
−1
bands at 504 and 618 cm
corresponding to RuO2 [35] were ob-
served on 3.7 wt% Ru/Al2O3-600. These bands disappeared within
2 min when the O2-pretreated samples were switched to a flow of
◦
H2/Ar = 5/95 mixture at 600 C (Fig. 4b). It is interesting that the
◦
Raman spectra of the catalysts recorded at 600 C under a simu-
lated POM feed (Fig. 4c) were very similar to those recorded under
the H2/Ar atmosphere. No Raman bands belonging to Rh2O3 or
RuO2 species were detected. These findings indicate that most of
the noble metal species in the catalysts located at the entrance
of the catalyst bed were in the metallic state under the condition
when the POM reaction occurred with CH4 conversion >82% and
CO selectivity >87% (Tables 2 and 3).
Fig. 3. O2-TPD profiles of the catalysts. Before the experiment, the catalysts were
◦
reduced with H2/Ar = 5/95 mixture at 600 C for 30 min followed by oxidation
◦
Figs. 5 and 6 show the Raman spectra of the Rh/Al2O3-110 and
Ru/Al2O3-110 catalysts recorded at specified temperatures under a
CH4/O2/Ar = 2/1/45 atmosphere as the temperature of the Raman
with O2/He = 1/4 mixture at 500 C for 30 min.
◦
◦
3.7 wt% Ru/Al2O3-110 disappeared at 500 and 470 C, respectively.
cell was raised from 30 to 550 or 600 C. As soon as the H2-
Interestingly, the temperatures at which the noble metal oxide
species in the catalysts vanished were in good agreement with the
temperatures at which the POM reaction over the catalysts ignited
(see Tables 2 and 3). Similar phenomena were observed on the
Rh/Al2O3-600 and Ru/Al2O3-600 catalysts (spectra not shown). No
other species were detected by Raman spectroscopy on either the
Rh/Al2O3 or Ru/Al2O3 samples.
Figs. 7 and 8 show XRD patterns of the Rh/Al2O3-110 and
Ru/Al2O3-110 catalysts recorded under a flow of a CH4/O2/Ar =
2/1/197 mixture (80 mL/min) during the stepwise heating of the
prereduced Rh/Al2O3-110 or Ru/Al2O3-110 catalyst was switched to
CH4/O2/Ar at room temperature, Raman bands of the correspond-
ing noble metal oxides (Rh2O3 or RuO2) species appeared, indicat-
ing that metallic rhodium or ruthenium species in the catalyst (at
least those on the surface of the metal particles) can be oxidized
by O2 in the reactant even at room temperature. As the temper-
ature of the samples increased, the Raman band of Rh2O3 species
(553 cm−1) on 1 wt% Rh/Al2O3-110 and 3 wt% Rh/Al2O3-110 disap-
◦
peared at 470 and 440 C, respectively, whereas the Raman bands
of RuO2 species (504 and 618 cm−1) on 1 wt% Ru/Al2O3-110 and

198
Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
◦
Fig. 4. Raman spectra of 3 wt% Rh/Al2O3-600 and 3.7 wt% Ru/Al2O3-600 catalysts recorded at 600 C under a. O2, b. H2/Ar = 5/95, and c. CH4/O2/Ar = 2/1/45.
Fig. 5. Raman spectra of 1 wt% Rh/Al2O3-110 and 3 wt% Rh/Al2O3-110 catalysts recorded at the indicated temperature during stepwise heating of the previously reduced
◦
samples under CH4/O2/Ar = 2/1/45 atmosphere from room temperature to 600 C. The Raman spectra of Al2O3 support were subtracted. Before switching to CH4/O2/Ar
◦
mixture, the catalysts were reduced with H2/Ar = 5/95 mixture at 600 C for 30 min.
◦
H2-prereduced samples from 25 to 750 C in the XRK-900 reactor.
Similar to the XRD patterns of fresh Rh/Al2O3 catalysts, no char-
acteristic diffraction peaks due to the rhodium oxide or metallic
Rh species were detected on the Rh/Al2O3-110 catalysts over the
entire temperature range, indicating high dispersion of rhodium
species on the Al2O3 support.
O2 in the reactant in this temperature range. In addition, the ag-
gregation of the oxidized RuO2 species to larger RuO2 particles also
may have occurred when the Ru/Al2O3-110 samples were heated
◦
under the reaction mixture to above 300 C, as evidenced by the
increased intensity of the XRD peaks of RuO2 with increasing tem-
perature. With a further rise in the temperature, the XRD peaks of
RuO2 over 1 wt% Ru/Al2O3-110 and 3.7 wt% Ru/Al2O3-110 vanished
For the Ru/Al2O3-110 catalysts under a CH4/O2/Ar mixture, no
diffraction patterns due to RuO2 or Ru species were detected at
temperatures below 200 C. When the temperature of the reac-
◦
at 500 and 450 C, respectively, and the characteristic diffraction
◦
peaks of metallic Ru phase were detected. Similar to the results of
in situ Raman characterization presented earlier, the temperatures
at which the RuO2 species in the catalysts disappeared also were
in good agreement with the temperatures at which the POM reac-
tion over the catalysts ignited.
◦
tor was raised to 300–450 C, diffraction peaks of crystalline RuO2
◦
phase (2θ = 28.0, 35.1 and 54.3 ) were observed, with intensities
increasing with temperature. These results suggest that the metal-
lic ruthenium species in the catalysts were extensively oxidized by

Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
199
Fig. 6. Raman spectra of 1 wt% Ru/Al2O3-110 and 3.7 wt% Ru/Al2O3-110 recorded at the indicated temperature during stepwise heating of the previously reduced samples
◦
under CH4/O2/Ar = 2/1/45 atmosphere from room temperature to 600 C. Before switching to CH4/O2/Ar mixture, the catalysts were reduced with H2/Ar = 5/95 mixture at
◦
400 C for 30 min.
Fig. 7. XRD patterns of 1 wt% Rh/Al2O3-110 and 3 wt% Rh/Al2O3-110 catalysts recorded at the indicated temperature during stepwise heating of the previously reduced
◦
samples under a flow of CH4/O2/Ar = 2/1/197 (80 mL/min) from 30 to 750 C. Before switching to CH4/O2/Ar mixture, the catalysts were reduced with H2/Ar = 5/95 mixture
◦
at 600 C for 30 min.

200
Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
Fig. 8. XRD patterns of 1 wt% Ru/Al2O3-110 and 3.7 wt% Ru/Al2O3-110 catalysts recorded at the indicated temperature during stepwise heating of the previously reduced
◦
samples under a flow of CH4/O2/Ar = 2/1/197 (80 mL/min) from 30 to 750 C. Before switching to CH4/O2/Ar mixture, the catalysts were reduced with H2/Ar = 5/95 mixture
◦
at 400 C for 30 min.
3.5. Pulse reaction of CH4 over the catalysts
strated a greater tendency toward oxidation than metallic Rh under
the experimental conditions.
For both SiO2- and Al2O3-supported catalysts with comparable
amounts of noble metal loading, the temperatures required to ig-
nite the POM reaction were always lower on the Rh catalysts than
on the Ru catalysts. The results of in situ Raman and XRD charac-
terizations suggest that reduction of the RuO2 species was more
difficult than reduction of the Rh2O3 species under the POM reac-
tion atmosphere. To further elucidate the relationship between the
redox properties of the noble metal species and performance in
the POM reaction, CH4-pulse experiments were performed over Ru
and Rh metal powders, as well as over the Al2O3-supported cata-
lysts. The experiments were performed using He containing a few
ppm of O2 as the carrier gas; the results are shown in Fig. 9. When
pulses of CH4 were introduced to the H2-prereduced Ru metal or
the supported Ru catalysts at 6-min intervals, the completely oxi-
dized product, CO2, could be detected in every pulse. In contrast,
CO2 was formed only in the first two pulses (and never thereafter)
when CH4 was pulsed over the supported Rh catalysts, and no CO2
was detected in the product for the pulse reaction over the H2-
prereduced Rh metal. Combustion of CH4 to CO2 and H2O was
favored on the catalysts with high surface oxygen concentrations
[14–18]. On the other hand, at a sufficiently low surface oxygen
concentration on the catalyst, CH4 can be selectively converted
to CO [15,17]. Thus, the results of CH4 pulse reaction over the
catalysts can be plausibly related to the significantly higher sur-
face oxygen concentrations on the ruthenium catalysts than on the
rhodium catalysts. These results indicate that metallic Ru demon-
4. Discussion
4.1. Chemical state of the noble metal species in the Rh/Al2O3 and
Ru/Al2O3 catalysts before ignition of the POM reaction
When a H2-prereduced sample is heated under a CH4/O2 mix-
ture at the POM stoichiometric ratio from 25 to 600 or 750 C, the
◦
oxidation state of the metal species in the catalyst is determined
mainly by the relative rates of two competitive reactions at the dif-
ferent temperatures, that is, the oxidation of metal species by O2
to the corresponding metal oxide [Eq. (1)] and the reduction of the
metal oxide species by CH4 to the corresponding metal [Eq. (2)],
y
xM + O2 → MxOy
(1)
2
and
y
y
y
MxOy + CH4 → CO2 + H2O + xM (M = Rh, Ru).
(2)
4
4
2
The results of in situ Raman characterization (Figs. 5 and 6) clearly
indicate that when a H2-prereduced catalyst was switched to the
CH4/O2/Ar = 2/1/45 mixture at room temperature, the surface of
the Rh and Ru particles located at the front end of the catalyst
bed were readily oxidized by O2 in the feed, as demonstrated
by the bands of Rh2O3 (∼550 cm−1) and RuO2 (503, 618 cm
)
−1
species in the corresponding Raman spectra. Considering the very

Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
201
◦
Fig. 9. CO and CO2 formations by pulsing CH4 over Rh and Ru metals as well as over 5 wt% Rh/Al2O3-600 and 3.7 wt% Ru/Al2O3-600 catalysts at 600 C using He containing
a trace amount of O2 as a carrier gas. The catalysts were pre-reduced with H2 at 600 C.
◦
low catalyst activity and O2 conversion at room temperature, it is
reasonable to conclude that O2 in the reactant should be capable
of oxidizing the surface of the noble metal particles in the entire
catalyst bed.
Rh/Al2O3 catalyst under CH4/O2/He (6/3/91) reaction mixture re-
ported by Grunwaldt et al. Both of these results indicate that the
rhodium species in the catalyst were in the oxidized state before
ignition of the POM reaction. A weight increase due to the oxida-
tion of metallic rhodium to rhodium oxide was observed when a
The corresponding in situ XRD patterns recorded at tempera-
◦
◦
tures below 200 C exhibited only the diffraction peaks of γ -Al2O3
prereduced Rh/γ -Al2O3 catalyst was heated at 350 C in a flow of
CH4/O2/Ar (2/1/20.5) [38].
and AlOOH. The absence of any detectable noble metal or ox-
ide peaks in the XRD patterns of Rh/Al2O3-110 and Ru/Al2O3-110
On heating the Ru/Al2O3-110 catalysts under the CH4/O2/Ar re-
◦
◦
recorded at temperatures below 200 C suggests that both metallic
action mixture from 200 C to the temperature of ignition of the
and oxidized Rh and Ru species were well dispersed on the sup-
port and that the average crystallite size of the noble metal species
in the catalysts was below the limit of detection by XRD.
POM reaction, the XRD pattern of bulk RuO2 was detected (Fig. 8),
indicating that metallic ruthenium species in the catalysts were
extensively oxidized to RuO2 by O2 in the reaction mixture in
this temperature range. Compared with the Rh/Al2O3 system, the
interaction of ruthenium dioxide with the alumina surface appar-
ently was much weaker, because the ruthenium dioxide was able
to sinter to a particle size providing a measurable intensity in
the XRD profile. Sintering may have occurred due to volatiliza-
tion from well-dispersed ruthenium oxide species, or migration of
tiny ruthenium oxide particles over the surface of the alumina sup-
port.
The foregoing results of in situ Raman and XRD characteriza-
tions indicate that in the temperature range before ignition of the
POM reaction, the reduction rates of Rh2O3 and RuO2 were lower
than the oxidation rates of metallic Rh and Ru, resulting in an ox-
idized surface. Because CO2 is the only carbon-containing product
for the reaction in this temperature range, it can be concluded that
noble metal oxide species (e.g., Rh2O3 or RuO2) are active only for
the combustion of CH4 to CO2 and H2O. This is also in agreement
with the results reported in the literature.
Sintering of the rhodium species in the Rh/Al2O3 catalyst at
high temperature was not significant, because of the strong inter-
action of rhodium oxide with the surface of the alumina support
[21,28,29]. Consequently, the in situ XRD analysis provided no infor-
mation on the structure of the rhodium species (oxide or metal) in
the catalyst during catalytic ignition of the POM reaction. However,
combined with the results of in situ Raman characterization shown
in Fig. 5 (indicating that metallic rhodium species in Rh/Al2O3
could be oxidized by O2 in the reactant even at room temperature)
and the results of the catalytic performance test given in Table 2
(indicating <40% O2 conversion in the reaction feed even when the
◦
furnace temperature was only 10 C below the POM ignition tem-
perature), it is reasonable to conclude that the rhodium species
in the catalyst bed (at least those on the surface of rhodium par-
ticles) are also in the oxidized state before ignition of the POM
reaction. This conclusion is also in agreement with the results of
in situ X-ray absorption fine structure (XAFS) [36] and in situ X-
ray absorption near-edge structure (XANES) [37] studies on the

202
Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
4.2. Chemical state of the noble metal species in the Rh/Al2O3 and
Ru/Al2O3 catalysts under the POM reaction condition
most of the Rh species located at the front of the Rh-coated α-
Al2O3 foam catalyst bed also were in the reduced state under the
experimental conditions.
As the temperature of the sample was increased, the reduction
rate of metal oxide by CH4 [Eq. (2)] increased. When the rate of
Eq. (2) exceeded that of Eq. (1), the noble metal oxide species in
the catalysts was reduced to the metallic state. With an increas-
ing amount of reduced metal sites, the POM reaction over the
catalyst was ignited. The combustion-reforming mechanism (CRR)
postulates two reaction zones in the catalyst bed and suggests that
synthesis gas is formed via the initially exothermic combustion of
CH4 to CO2 and H2O, followed by the endothermic reformation of
unconverted CH4 with CO2 and H2O. This reaction scheme requires
that most of the metal species in the catalyst near the entrance of
the catalyst bed be in oxidized form or that the oxygen coverage
on the surface of the metal particles be high enough to catalyze
the combustion of CH4 to CO2 and H2O. Previously obtained two-
dimensional X-ray absorption spectroscopy (XAS) mapping images
on the distribution of Rh3+/Rh0 species along the Rh/Al2O3 [37,
39] and Pt–Rh/Al2O3 [40] catalyst beds under a simulated POM
On the reduced metal sites, CH4 can be readily activated by dis-
sociation to surface carbon species (CHx, x = 0–3), which can then
convert to CO through a series of surface reactions [6]. In our pre-
vious work, Raman bands of surface carbon species (∼1330 and
∼1570 cm−1) formed by CH4 dissociation were clearly seen at the
entrance of an Ir/SiO2 catalyst bed under working conditions [41].
For the POM reaction over Rh/Al2O3 and Ru/Al2O3 catalysts, how-
ever, no Raman bands of surface carbon species were detected on
the catalysts under similar experimental conditions. This may be
due to the higher oxygen coverage on the Rh/Al2O3 and Ru/Al2O3
catalysts compared with the Ir/SiO2 catalyst under working con-
ditions. Another possible explanation is that the reactivity of the
surface carbon species was higher on Rh/Al2O3 and Ru/Al2O3 than
on the supported Ir catalysts. The Raman bands of the carbon
species on Ir/SiO2 decreased in intensity as the temperature was
◦
◦
increased from 500 to 700 C, and vanished at 750 C. Further ex-
periments are needed to clarify these questions.
◦
feed (CH4/O2/He = 6/3/91) at ca. 320 C are consistent with the
CRR scheme.
4.3. Comparison of Rh/Al2O3 and Ru/Al2O3
In contrast to the CRR mechanism, the direct oxidation mecha-
nism (DPO) assumes that H2 and CO are primary reaction products
formed in the zone at the catalyst entrance, where O2 in the re-
action mixture has not been completely consumed. After methane
pyrolysis (CH4 → CHxs +4-xHs), surface CHx species react with sur-
face oxygen to CO, and surface hydrogen atoms combine to H2.
This mechanism requires that most of the metal species in the cat-
alyst at the entrance of the catalyst bed be in the reduced form so
that methane dissociation to CHx can occur [12–14,16,19].
The results of in situ Raman and XRD characterizations shown
in Figs. 5–8 clearly indicate that after ignition of the POM reaction,
no Rh2O3 or RuO2 species were detected even in the catalyst lo-
cated at the top of the catalyst bed, where O2 was still available
in the reaction feed. This means that the noble metal species in
the entire catalyst bed were in a highly reduced form under the
POM reaction conditions. A sharp axial transition of the catalyst
(or noble metal species) from the oxidized state (e.g., Rh3+) to the
reduced state (Rh0), as observed by Grunwaldt et al. [37] and Han-
nemann et al. [39] on Rh/Al2O3 and by Grunwaldt and Baiker [40]
on Pt–Rh/Al2O3 at temperatures slightly above the ignition point
of the POM reaction, were not seen under our experimental condi-
tions. This may be because the ignition temperatures for the POM
reactions over the Rh/Al2O3 catalysts of the present study were
The results of catalytic performance testing for the reactions
over Al2O3- and SiO2-supported catalysts given in Tables 2–4 in-
dicate that the temperatures required to ignite the POM reaction
are significantly higher over supported Ru catalysts than over sup-
ported Rh catalysts with comparable metal loadings. A similar phe-
nomenon also has been reported for γ -Al2O3-supported [38] and
MgO-supported [42] Rh and Ru catalysts. On the other hand, the
ignition temperatures of the POM reaction over Ru/Al2O3-110 and
Ru/Al2O3-600 catalysts with similar Ru loadings were very close,
even though the dispersion of Ru species was much greater in the
former (Table 1). These observations suggest that for catalysts with
comparable metal loadings, the ignition temperature of the POM
reaction is related more to the nature of the noble metal itself
than to the dispersion or particle size of the metal species on the
support.
The foregoing results of in situ Raman and XRD characteriza-
tions clearly indicate that the noble metal species in the catalysts
were in the fully oxidized state before ignition of the POM reac-
tion. On the oxidized metal surface, the reaction between CH4 and
O2 occurs via the Eley-Rideal mechanism; that is, CH4 in the gas
phase (or weakly adsorbed) reacts with strongly adsorbed or lat-
tice oxygen [43]. This suggests that the ignition temperature of
the POM reaction should be closely related to the metal–oxygen
(M–O) bond strength of the metal oxides. M–O species with higher
bond strength are less reactive, and thus higher temperatures are
needed to initiate their reaction with CH4. The temperature gap
between the ignition temperatures of supported Rh and Ru cata-
lysts can be rationally explained by the significant differences in
the Rh–O and Ru–O bond strengths. The bond strength of Ru–O
(481 63 kJ/mol) is much higher than that of Rh–O (377 63
kJ/mol) [44]. This should result in a more difficult reduction of
ruthenium oxide than of rhodium oxide under the POM reaction
atmosphere. A linear relationship between the ignition tempera-
tures and M–O bond strengths also has been reported for reactions
of a C2H6/air mixture over Pt, Pd, Rh, and Ir foils [45].
◦
◦
between ∼430 and ∼470 C, compared with the 330 C reported
by Grunwaldt et al. on their Rh/Al2O3 catalyst (prepared by flame
spray pyrolysis). The reaction to metallic rhodium, which calls for
a relatively high temperature, would proceed preferentially under
our experimental conditions, but the amount of oxidized rhodium
◦
species in the catalyst at temperatures above 430 C might be too
small to be detected by Raman spectroscopy. To the best of our
knowledge, all of the spatially resolved in situ XAS characteriza-
tions reported to date on catalysts for the POM reaction in which
a distinct change in the oxidation state of the active metal species
along the catalyst bed was observed were performed at tempera-
◦
tures below 400 C.
The results of in situ Raman characterization on the Rh/Al2O3
catalyst are in line with the results of spatially resolved gas species
analysis along the catalyst bed reported by Horn et al. [20] for
the catalytic partial oxidation of CH4 on autothermally operated
Rh-coated α-Al2O3 foams under steady-state conditions. In their
experiment, both CO and H2 were detected in the region near the
entrance of the catalyst bed, where O2 in the reactant had not
been completely consumed. Because metallic Rh sites are required
for the formation of synthesis gas, it is reasonable to conclude that
Once the POM reaction is ignited, most of the metal species
in the Rh/Al2O3 and Ru/Al2O3 catalysts are in the reduced state.
The reaction of CH4 with O2 occurs through the Langmuir–
Hinshelwood mechanism, in which adsorbed CH4 and O2 species
react with one another [43]. Normal (>1) deuterium isotope ef-
fects on CH4 conversion (or CH4 consumption rate) and CO yield
(or CO formation rate) have been reported for the POM reaction
on supported Rh catalysts [26,46] and Ru catalysts [47] at 550–

Y. Liu et al. / Journal of Catalysis 256 (2008) 192–203
203
◦
700 C. This suggests that the dissociation of CH4 to CHx species
or the formation of CO involving cleavage of a C–H bond is a slow
or rate-determining step that controls the overall process. Under
these circumstances, the reaction of surface carbon species with
adsorbed oxygen may be a fast step, and the formation of primary
products (CO or CO2) may depend mainly on the oxygen cover-
age on the metal surface. Higher oxygen coverage is more readily
obtained on metals with greater M–O bond strength or a higher
affinity for oxygen. This is in line with the results of our CH4 pulse
reaction experiments shown in Fig. 9. As mentioned in Section 2
because the catalysts used in the pulse reaction experiments were
Acknowledgments
This project was supported by the Ministry of Science and
Technology of China (2005CB221401), the National Natural Science
Foundation of China (20433030 and 20423002), the Key Scientific
Project of Fujian Province, China (2005HZ01-3), and the Natural
Science Foundation of Fujian Province, China (2007J0168).
References
[1] J.R. Rostrup-Nielsen, Catal. Today 18 (1993) 305.
[2] A.T. Ashcroft, A.K. Cheetham, M.L.H. Green, P.D.F. Vernon, Nature 352 (1991)
225.
◦
prereduced with H2 at 600 C, the source of the oxygen responsible
[3] D.A. Hickman, L.D. Schmidt, Science 259 (1993) 343.
[4] A.P.E. York, T.C. Xiao, M.L.H. Green, Top. Catal. 22 (2003) 345.
[5] Y.H. Hu, E. Ruckenstein, Adv. Catal. 48 (2004) 297.
[6] D.A. Hickman, L.D. Schmidt, J. Catal. 138 (1992) 267.
[7] M. Prettre, Ch. Eichner, M. Perrin, Trans. Faraday Soc. 42 (1946) 335.
[8] I. Tavazzi, A. Beretta, G. Groppi, P. Forzatti, J. Catal. 241 (2006) 1.
[9] F. van Looij, E.R. Stobbe, J.W. Geus, Catal. Lett. 50 (1998) 59.
[10] J.C. Slaa, R.J. Berger, G.B. Marin, Catal. Lett. 43 (1997) 63.
[11] D. Dissanayake, M.P. Rosynek, K.C.C. Kharas, J.H. Lunsford, J. Catal. 132 (1991)
117.
[12] O.V. Buyevskaya, D. Wolf, M. Baerns, Catal. Lett. 29 (1994) 249.
[13] E.P.J. Mallens, J.H.B.J. Hoebink, G.B. Marin, J. Catal. 167 (1997) 43.
[14] K.H. Hofstad, J.H.B.J. Hoebink, A. Holmen, G.B. Marin, Catal. Today 40 (1998)
157.
[15] M. Fathi, F. Monnet, Y. Schuurman, A. Holmen, C. Mirodatos, J. Catal. 190 (2000)
439.
[16] D. Wang, O. Dewaele, A.M. De Groote, G.F. Froment, J. Catal. 159 (1996) 418.
[17] W.Z. Weng, Q.G. Yan, C.R. Luo, Y.Y. Liao, H.L. Wan, Catal. Lett. 74 (2001) 37.
[18] W.Z. Weng, M.S. Chen, H.L. Wan, Chem. Rec. 2 (2002) 102.
[19] Y.H. Hu, E. Ruckenstein, J. Phys. Chem. A 102 (1998) 10568.
[20] R. Horn, K.A. Williams, N.J. Degenstein, L.D. Schmidt, J. Catal. 242 (2006) 92.
[21] W.Z. Weng, X.Q. Pei, J.M. Li, C.R. Luo, Y. Liu, H.Q. Lin, C.J. Huang, H.L. Wan, Catal.
Today 117 (2006) 53.
for the reoxidation of metallic Rh and Ru species in the catalysts
between two CH4 pulses should be mainly the O2 in the He flow.
Due to the greater bond strength of Ru–O compared with Rh–O,
the reduced Ru species will be more easily oxidized by the O2 in
the carrier gas than the reduced Rh species. As a result, the oxygen
coverage on the surface of reduced ruthenium catalysts (Ru pow-
der or Ru/Al2O3) between two CH4 pulses will be sufficiently high
to oxidize CH4 to both CO2 and CO, whereas the oxygen species on
the Rh powder or supported Rh catalysts will favor the formation
of CO only, because its concentration on the catalyst surface under
the same experimental conditions is too low to completely oxidize
the CH4 to CO2. Thus, it can be rationally concluded that for the
POM reaction carried out under similar experimental conditions,
the oxygen coverage will be higher on the surface of Ru/Al2O3
than on the surface of Rh/Al2O3. This suggests that Rh/Al2O3 is a
better catalyst than Ru/Al2O3 for promoting the POM reaction via
the direct partial oxidation pathway.
[22] A.C. Yang, C.W. Garland, J. Phys. Chem. 61 (1957) 1504.
[23] F. Solymosi, M. Pásztor, J. Phys. Chem. 89 (1985) 4789.
[24] M.I. Zaki, G. Kunzmann, B.C. Gates, H. Knözinger, J. Phys. Chem. 91 (1987) 1486.
[25] P. Basu, D. Panayotov, J.T. Yates Jr., J. Phys. Chem. 91 (1987) 3133.
[26] H.Y. Wang, E. Ruckenstein, J. Phys. Chem. B 103 (1999) 11327.
[27] D. Dissanayake, M.P. Rosynek, J.H. Lunsford, J. Phys. Chem. 97 (1993) 3644.
[28] C.P. Hwang, C.T. Yeh, Q. Zhu, Catal. Today 51 (1999) 93.
[29] H.C. Yao, S. Japar, M. Shelef, J. Catal. 50 (1977) 407.
[30] H. Madhavaram, H. Idriss, S. Wendt, Y.D. Kim, M. Knapp, H. Over, J. Aßmann, E.
Löffler, M. Muhler, J. Catal. 202 (2001) 296.
5. Conclusion
Based on the results of our in situ Raman and XRD charac-
terizations and catalytic performance tests on the Rh/Al2O3 and
Ru/Al2O3 catalysts, we can conclude that, at temperatures below
the ignition point of the POM reaction, the noble metal species in
both catalysts (at least those on the surface of the metal particles)
were in the fully oxidized form. The oxidized metal species were
active only for the complete oxidation of CH4 to CO2 and H2O.
Most of the noble metal species in the catalysts changed from the
oxidized form to the metallic state at the start of the POM reac-
tion. At temperatures above the ignition point of the POM reaction,
the catalyst was in a highly reduced state. The amount of Rh2O3
or RuO2 in the catalyst at the entrance of the catalyst bed (where
O2 was still available in the reaction mixture) was below the de-
tection level of Raman spectroscopy. Because of the greater M–O
bond strength of Ru–O compared with Rh–O, the Ru/Al2O3 cata-
lyst showed a greater tendency toward oxidation and was more
difficult to reduce than the Rh/Al2O3 catalyst under the CH4/O2/Ar
mixture at the POM stoichiometric ratio (CH4/O2 = 2). These fac-
tors affect the coverage of oxygen species on the surface of the
Rh/Al2O3 and Ru/Al2O3 catalysts under the POM reaction condi-
tions and, consequently, affect the reaction pathways of synthesis
gas formation over these two catalysts.
[31] P. Betancourt, A. Rives, R. Hubaut, C.E. Scott, J. Goldwasser, Appl. Catal.
Gen. 170 (1998) 307.
A
[32] I. Balint, A. Miyazaki, K. Aika, J. Catal. 207 (2002) 66.
[33] M. Boudart, J. Mol. Catal. 30 (1985) 27.
[34] C.T. Williams, C.G. Takoudis, M.J. Weaver, J. Phys. Chem. B 102 (1998) 406.
[35] S.Y. Mar, C.S. Chen, Y.S. Huang, K.K. Tiong, Appl. Surf. Sci. 90 (1995) 497.
[36] J.D. Grunwaldt, L. Basini, B.S. Clausen, J. Catal. 200 (2001) 321.
[37] J.D. Grunwaldt, S. Hannemann, C.G. Schroer, A. Baiker, J. Phys. Chem. B 110
(2006) 8674.
[38] S. Rabe, T.B. Truong, F. Vogel, Appl. Catal. A Gen. 292 (2005) 177.
[39] S. Hannemann, J.D. Grunwaldt, N. Van Vegten, A. Baiker, P. Boye, C.G. Schroer,
Catal. Today 126 (2007) 54.
[40] J.D. Grunwaldt, A. Baiker, Catal. Lett. 99 (2005) 5.
[41] W.Z. Weng, C.R. Luo, J.M. Li, Y. Liu, H.Q. Lin, C.J. Huang, H.L. Wan, Acta Chim.
Sinica 62 (2004) 1853.
[42] D. Qin, J. Lapszewicz, X. Jiang, J. Catal. 159 (1996) 140.
[43] Y.H. Hu, E. Ruckenstein, J. Phys. Chem. B 102 (1998) 230.
[44] R.C. Weast, M.J. Astle (Eds.), CRC Hand Book of Chemistry and Physics, 62nd
ed., CRC Press Inc., Boca Raton, FL, 1981–1982, F-184.
[45] G. Veser, M. Ziauddin, L.D. Schmidt, Catal. Today 47 (1999) 219.
[46] C.T. Au, H.Y. Wang, J. Catal. 167 (1997) 337.
[47] C. Elmasides, X.E. Verykios, J. Catal. 203 (2001) 477.