Effect of treatment temperature on structures and properties of zirconia-supported ruthenium oxide catalysts for selective oxidation of methanol to methyl formate

Article abstract of DOI:10.1016/j.cattod.2011.05.021

The structures of the RuO_x species supported on monoclinic ZrO₂ (m-ZrO₂) after treatment in a wide temperature range 673-1323 K were characterized by X-ray diffraction, Raman, infrared and X-ray photoelectron spectroscopies, and temperature-programmed reduction in H ₂. Their catalytic properties were examined in the selective oxidation of methanol to methyl formate at 373 K. The RuO_x species were present as highly dispersed RuO₄^2- on m-ZrO ₂ with Ru surface densities of 0.2-2.2 Ru/nm², irrespective of the treatment temperatures. Their molecular structures evolved from umbrella-like dioxo (O)₂Ru(OH)-O-Zr (containing a Ru-OH bond and a Ru-O-Zr bond to m-ZrO₂ surface) into tetrahedral dioxo (O) ₂Ru-(O-Zr)₂ with increasing the treatment temperatures from 673 to 773 K, and into pyramidic mono-oxo ORu-(O-Zr)₄ above 1173 K, induced by the stronger interaction between the RuO_x species and underlying m-ZrO₂ surface at the higher temperatures. Following such structural changes, the turnover rates of the methanol oxidation increased markedly, and reached the greatest value (39.0 mol/mol Ru-h) on the catalyst treated at 1223 K, as a result of the parallel increase in the reducibility of the RuO_x species, consistent with the known Mars-van Krevelen redox mechanism using lattice oxygen atoms on RuO_x. These understandings may be useful for improving the reactivity of the RuO_x-based catalysts for the selective oxidation of methanol as well as of other alcohols and probably light hydrocarbons, for instance, by synthesis and tuning of the more reducible di-oxo RuO₄^2- structures.

Full text of DOI:10.1016/j.cattod.2011.05.021

Catalysis Today 183 (2012) 58–64
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Effect of treatment temperature on structures and properties of
zirconia-supported ruthenium oxide catalysts for selective oxidation of methanol
to methyl formate
Hua Huang1, Weizhen Li1, Haichao Liu∗
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Stable and Unstable Species, College of Chemistry and Molecular Engineering,
Green Chemistry Center, Peking University, Beijing 100871, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 May 2011
Received in revised form 21 May 2011
Accepted 23 May 2011
Available online 23 June 2011
The structures of the RuOx species supported on monoclinic ZrO2 (m-ZrO2) after treatment in a wide
temperature range 673–1323 K were characterized by X-ray diffraction, Raman, infrared and X-ray pho-
toelectron spectroscopies, and temperature-programmed reduction in H2. Their catalytic properties
were examined in the selective oxidation of methanol to methyl formate at 373 K. The RuOx species
were present as highly dispersed RuO4 on m-ZrO2 with Ru surface densities of 0.2–2.2 Ru/nm2, irre-
2−
spective of the treatment temperatures. Their molecular structures evolved from umbrella-like dioxo
(O )2Ru(OH)–O–Zr (containing a Ru–OH bond and a Ru–O–Zr bond to m-ZrO2 surface) into tetrahe-
dral dioxo (O )2Ru–(O–Zr)2 with increasing the treatment temperatures from 673 to 773 K, and into
pyramidic mono-oxo O Ru–(O–Zr)4 above 1173 K, induced by the stronger interaction between the
RuOx species and underlying m-ZrO2 surface at the higher temperatures. Following such structural
changes, the turnover rates of the methanol oxidation increased markedly, and reached the greatest
value (39.0 mol/mol Ru-h) on the catalyst treated at 1223 K, as a result of the parallel increase in the
reducibility of the RuOx species, consistent with the known Mars–van Krevelen redox mechanism using
lattice oxygen atoms on RuOx. These understandings may be useful for improving the reactivity of the
RuOx-based catalysts for the selective oxidation of methanol as well as of other alcohols and probably
light hydrocarbons, for instance, by synthesis and tuning of the more reducible di-oxo RuO42− structures.
Keywords:
Methanol oxidation
Methyl formate
Ruthenium oxide
Treatment temperature effect
Structures
Catalytic property
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
HCHO intermediate, and subsequent reactions of HCHO, which
also competitively lead to other products, especially undesired COx
Methyl formate (HCOOCH3, MF) is a versatile intermediate in
the synthesis of important formamides, carboxylic acids, and esters
[1–6]. It can be synthesized by selective oxidation of methanol,
which appears to be a greener and thermodynamically favor-
able alternative to the current industrial processes based on
homogenous carbonylation and non-oxidative dehydrogenation of
methanol generally using liquid sodium methoxide and Cu cata-
lysts, respectively [1,4–6].
V2O5–TiO2 and MoO3–SnO2 mixed oxides have been studied
for methanol oxidation to MF [4,7,8], but significant improve-
ments in their catalytic activities or selectivities are required for
industrial applications. It is known that methanol oxidation to
MF involves the kinetically relevant C–H activation step to form
(CO + CO2) [3,9]. In line with the mechanism, Liu and Iglesia [10]
explored more reducible RuOx domains for methanol oxidation,
and discovered their unprecedentedly superior activities relative
to VOx and MoOx-based catalysts at lower reaction temperatures
(330–400 K). Recently, we reported that the RuOx domains present
as isolated RuO42− (containing Ru6+) structures on monoclinic ZrO2
(m-ZrO2) surface are efficient for the MF synthesis with a high selec-
tivity of ∼96% (at ca. 20% methanol conversion) at temperatures of
2−
as low as 373 K [11]. These RuO4
structures are more reactive
than RuO2 domains [11], reflecting the structure-sensitivity of the
methanol oxidation reaction [3].
2−
The isolated RuO4 structures on m-ZrO2 are formed from the
strong interaction of the RuOx species with m-ZrO2 surface during
thermal treatment [11]. This promotes us to study the effects of the
treatment temperatures on the m-ZrO2 supported RuOx (RuOx/m-
ZrO2) catalysts, as part of our ongoing efforts to rationally control
∗
Corresponding author. Tel.: +86 10 62754031; fax: +86 10 62754031.
E-mail address: hcliu@pku.edu.cn (H. Liu).
These authors contributed equally to this work.
2−
the active RuO4
structures and their reactivity in the selective
1
methanol oxidation. In this work, we characterize the RuOx/m-ZrO2
0920-5861/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.05.021

H. Huang et al. / Catalysis Today 183 (2012) 58–64
59
catalysts treated at different temperatures (673–1323 K) by X-ray
diffraction (XRD), X-ray photoelectron (XPS) and Raman spectro-
scopies. We probe their reducibility by temperature-programmed
reduction (TPR) in H2. Their catalytic properties in the oxidation
of methanol to MF are examined and then related to the proposed
2−
evolution of the RuO4
structures on m-ZrO2 with the different
treatment temperatures.
2. Experimental methods
2.1. Sample synthesis
RuOx/m-ZrO2 catalysts with a 0.5 wt% Ru loading were prepared
by incipient wetness impregnation of m-ZrO2 with an aqueous
solution of Ru(NO)(NO3)3 (Alfa Aesar, 31.3% Ru) at 298 K for
3 h. The samples were dried in ambient air at 398 K overnight
after impregnation, and then treated in ambient air at different
temperatures between 673 and 1323 K for 4 h. m-ZrO2 was syn-
thesized hydrothermally by hydrolysis of aqueous zirconyl nitrate
(ZrO(NO3)2·2H2O; >45.0% ZrO2, Beijing Chemicals) solutions with
urea (CO(NH2)2; >99.9%, Beijing Chemicals) at 433 K for 20 h in a
Teflon-lined stainless-steel autoclave, followed by filtration of the
precipitates and treatment in ambient air at 398 K overnight and
subsequently at 673 K for 4 h [12].
Fig. 1. X-ray diffraction patterns for RuOx/m-ZrO2 catalysts after treatment in flow-
ing dry air at different temperatures between 673 and 1323 K.
2.2. Sample characterization
2.3. Catalytic methanol oxidation reactions
BET surface areas for the RuOx/m-ZrO2 samples were mea-
sured by N2 physisorption at 78.3 K using an ASAP 2010 analyzer
(Micromeritics) after they were evacuated at 393 K for 4 h under
dynamic vacuum (<2.66 Pa) conditions.
Methanol oxidation reactions were carried out in a packed-bed
quartz microreactor at 373 K and atmospheric pressure. The cata-
lysts (60–80 mesh) were diluted with quartz powders to prevent
temperature gradients and hot spots. The amounts of the catalysts
varied with their activities to keep similar methanol conversions
of about 20%. The catalysts were treated in 10% O2/N2 (Beijing
Huayuan, 99.999%) flow (30 cm3 min−1) at 673 K for 1 h before
they were exposed to the reactant mixtures consisting of 3.5 kPa
CH3OH (Beijing Chemical, 99.99%), 30 kPa O2 and balance N2 (Bei-
jing Huayuan, 99.999%). Methanol was introduced by bubbling N2
through a glass saturator filled with liquid methanol (Beijing Chem-
ical, 99.99%). All stream lines from the reactor to the detector were
kept above 393 K to avoid condensation of the products. Reactants
and products were analyzed by on-line gas chromatography (Shi-
madzu 2010 GC) using two packed columns (5A molecular sieve
for O2/N2 and Porapak Q for other components) and TCD detec-
tors. All data in this study were collected at steady state after 2 h
on-stream. Selectivities are reported on a carbon basis and rates as
molar CH3OH conversion rates per mole of Ru per hour (mol/mol
Ru-h). Blank experiments were conducted on quartz and pure m-
ZrO2 powders, showing no detectable CH3OH conversions under
all reaction conditions employed in this study.
XRD patterns were collected in the 2ꢀ range of 10–80◦ using
a Rigaku D/MAX-2000 diffractometer with a Cu K␣ radiation
˚
source (ꢁ = 1.5406 A) operated at 30 kV and 100 mA. The average
particle sizes (D) were estimated by the Scherrer equation [13],
D = 0.90ꢁ/ˇ cos ꢀ, where ꢀ is the diffraction angle and ˇ is the full
width at half-maximum (fwhm).
XPS measurements were preformed on an Axis Ultra spectrom-
eter using monochromatic Al K␣ (1486.71 eV) radiation at a source
power of 225 W (15 mA and 15 kV). The binding energies were cal-
ibrated by referring to the C 1s peak at 284.8 eV for adventitious
carbon.
Raman spectra were measured in the range of 100–1500 cm−1
on a Bruker Senterra confocal microscope spectrometer equipped
with a Nd:YAG laser (532 nm) and a CCD camera. The RuOx/m-
ZrO2 samples were placed in a quartz cell for in situ treatment and
measurement without exposure to ambient air.
H2 temperature-programmed reduction (H2-TPR) experiments
were carried out on a flow unit (TP5000, Tianjin Xianquan) in a
∼5% H2/N2 flow (30 cm3 min−1; Beijing Huayuan). The RuOx/m-
ZrO2 samples (∼0.1 g) were treated in a quartz cell in flowing air for
1 h to remove surface contaminants, and then they were switched
to the H2/N2 flow at 298 K and heated linearly to 873 K at a ramp-
ing rate of 10 K min−1. The H2 consumption was measured using
an on-line quadruple mass spectrometer (Hiden HPR 20), and the
response was calibrated by the reduction of CuO powders (Beijing
Chemicals, >99.0%) in H2 (∼5% H2/N2).
3. Results and discussion
3.1. Catalyst characterization
Table 1 shows BET surface areas, support particle sizes and nom-
inal surface densities for RuOx/m-ZrO2 samples treated from 673 to
1323 K. The surface areas decreased from 129.6 to 90.5 m2/g with
increasing the treatment temperatures from 673 to 773 K, and then
sharply to 52.5 and 13.3 m2/g at 873 and 1323 K, respectively. Such
decrease in the surface areas indicates the thermal sintering of the
ZrO2 support, as also evidenced from the concurrent increase in its
particle sizes from 7.8 to 36.0 nm (estimated from the width of the
diffraction line at 28.1◦ shown in Fig. 1). Accordingly, the normal
RuOx surface density, calculated from the Ru content and BET sur-
Diffuse reflectance infrared Fourier transform (DRIFT) spectra
were recorded in reflectance mode with 4 cm−1 resolution on a
Bruker Tensor 27 FT-IR spectrometer equipped with a Harrick
diffuse-reflectance attachment and a reaction chamber. The sam-
ples were treated in the reaction chamber in a N2 flow at 573 K for
2 h before measuring the spectra in the range of 1000–4000 cm−1 at
ambient temperature. All the spectra were obtained by subtracting
the spectrum of KBr treated under the identical conditions.

60
H. Huang et al. / Catalysis Today 183 (2012) 58–64
Table 1
BET surface areas, average ZrO2 crystallite sizes, nominal Ru surface densities and XPS Ru 3d5/2 and Ru 3d3/2 binding energies for monoclinic ZrO2-supported RuOx catalysts
treated at different temperatures.
Treatment temperature (K)
Surface area (m2/g)
Support particle sizea (nm)
Ru surface density (Ru/nm)
Binding energy (eV)
Ru 3d5/2
Ru 3d3/2
673
723
773
873
973
1073
1173
1223
1273
1323
129.6
110.9
90.5
52.5
38.4
34.8
26.6
19.2
15.1
13.3
7.8
8.3
8.3
14.9
16.1
18.6
22.7
27.0
33.3
36.0
0.2
0.3
0.3
0.6
0.8
0.9
1.1
1.6
2.0
2.2
282.7
282.6
282.6
282.7
282.7
282.7
282.7
282.7
282.8
282.7
286.9
286.8
286.8
286.9
286.9
286.9
286.9
286.9
287.0
286.9
a
Calculated from X-ray diffraction patterns for these samples by the Scherrer equation.
face area, increased from 0.2 to 2.2 Ru/nm2. In this surface density
range, the RuOx domains were reported to be highly dispersed on
the m-ZrO2 surface with nearly 100% dispersion [11].
Fig. 1 shows XRD patterns for the RuOx/m-ZrO2 catalysts follow-
ing treatment at different temperatures in the range 673–1323 K,
where all diffraction peaks at 2ꢀ = 24.2◦, 28.2◦, 31.5◦, and 34.3◦ are
assigned to monoclinic ZrO2. No tetragonal ZrO2 and crystalline
RuO2 phases were detected, because of the absence of their char-
acteristic peaks at 30.2◦ and 35.2◦, respectively [11,12].
progressive blue-shift may reflect the stronger interaction of the
RuOx species with m-ZrO2 at the higher treatment temperatures,
and the consequent increase in its Ru O bond strength and stretch-
ing vibration frequencies. Along with such blue-shift, a shoulder
appeared at 970 cm−1 at 873 K, and became more intense at
higher temperatures. Above 1173 K, a weak Raman band emerged
at 1020 cm−1. It also became more intense currently with the
attenuation of the 1003 cm−1 band with increasing the treatment
temperatures, and ultimately reached the equivalent intensity of
the 970 cm−1 band at 1323 K. To assign the two new bands, pure
m-ZrO2 was characterized by Raman for comparison after treat-
ment at the same temperatures (Supplementary data, Fig. S1). The
two bands were also observed on m-ZrO2 after treatment above
873 K, but essentially with equivalent intensities. These features
together suggest that the two new bands in the RuOx/m-ZrO2 sam-
ples are attributed not only to m-ZrO2 support itself, but also to the
RuOx species, which is indeed confirmed by the combined results
of H2-TPR and Raman characterization, as discussed below.
Raman spectra (Fig. 2) confirmed the XRD results. In the range
250–650 cm−1, Raman bands appeared only at 618, 558, 504, 477,
381 and 335 cm−1, which are characteristic of m-ZrO2 with no
detectable bands corresponding to tetragonal ZrO2 and crystalline
RuO2 phases (e.g. at 271 and 521 cm−1, respectively) [14,15]. In
the range 900–1100 cm−1, a broad band at 978 cm−1 was detected
on the RuOx/m-ZrO2 sample treated at 673 K, which is attributed
2−
to terminal Ru O stretching vibration in the tetrahedral RuO4
structures [11]. This band shifted from 978 to 1003 cm−1 with
increasing the treatment temperatures from 673 K to 1173 K, and
then remained around 1003 cm−1 between 1173 and 1323 K. The
Fig. 3 presents H2-TPR profiles for the treated RuOx/m-ZrO2
samples. For comparison, pure m-ZrO2 was also examined, and it
shows a reduction peak at 730 K. On the RuOx/ZrO2 sample treated
1003
429
590
1020
970
357
1323 K
1323 K
1273 K
1223 K
1173 K
1273 K
1223 K
1003
1173 K
1073 K
1073 K
973 K
973 K
873 K
873 K
398
985
773 K
773 K
723 K
723 K
435
978
673 K
m-ZrO2
673 K
730
273
423
573
723
873
1100
1100
1000
1000
900 650
450
250
900 650 800450
Temperature (K)
Raman shift (cm-1)
Fig. 3. H2 temperature-programmed reduction profiles for RuOx/m-ZrO2 catalysts
after treatment in flowing dry air at different temperatures between 673 K and
1323 K.
Fig. 2. Raman spectra for dehydrated RuOx/m-ZrO2 catalysts after treatment in
flowing dry air at different temperatures between 673 and 1323 K.

H. Huang et al. / Catalysis Today 183 (2012) 58–64
61
at 673 K, there were two reduction peaks around 435 and 590 K. The
2−
low-temperature peak corresponds to the reduction of the RuO4
A
373 K
429 K
species to Ru0 [11], consistent with its reduction stoichiometry of
approximately 3 H2/Ru. The peak around 590 K can be assigned to
m-ZrO2 reduction. The observed shift to lower reduction temper-
ature relative to pure m-ZrO2 may be due to the promoting effect
of RuOx on the mobility of the lattice oxygen atoms of m-ZrO2 or
to the effect of hydrogen spillover from metallic Ru (derived from
the RuOx reduction) to m-ZrO2, as generally found with other sup-
ported metal catalysts [16–19]. The reduction peak of the RuOx
species shifted to lower temperatures from 435 to 357 K with
increasing the treatment temperatures from 673 up to 1323 K; the
peak intensity decreased sharply above 1173 K. Meanwhile, a new
reduction peak was noticed around a higher temperature of 429 K
on the samples treated at 1173 K, which became more intense at
higher treatment temperatures, apparently at the expense of the
lower-temperature peak. The corresponding amounts of H2 con-
sumption for the two reduction peaks were estimated, and the
molar fraction derived from the new peak increased from 27.7
and 51.5% to 89.0 and 89.9% with increasing the treatment tem-
peratures from 1173 and 1223 K to 1273 and 1323 K, respectively.
However, the reduction stoichiometry of the RuOx species for the
two peaks was always ca. 2.9–3.2 H2/Ru, irrespective of the treat-
ment temperatures, indicating the presence of only Ru6+ species, i.e.
RuO42−, as further confirmed by XPS. The XPS result exhibited only
peaks at binding energies of 282.7 and 286.9 eV (Table 1), corre-
sponding to Ru 3d5/2 and Ru 3d3/2 for the Ru6+ species, respectively
[11,20].
(c)
(b)
(a)
403 K
300
350
400
450
500
Temperature (K)
νs(Ru(=O)2)
νs(Ru=O)
B
νas(Ru(=O)2)
(d)
To more clearly assign the Raman bands (Fig. 2) and H2-TPR
reduction peaks (Fig. 3), the RuOx/m-ZrO2 samples were character-
ized upon controlled reduction in H2 by in situ Raman and TPR. Fig. 4
shows the representative results for the sample treated at 1223 K.
After reduction at 403 K, i.e. only selective removal of the lower-
temperature peak at 373 K, the corresponding Raman bands at 1003
and 970 cm−1 disappeared and sharply attenuated, respectively,
whereas the 1020 cm−1 band remained essentially unchanged. Fur-
ther reduction of the sample at 503 K led to a decrease of the Raman
band at 1020 cm−1 in intensity, and no detectable further change in
the 970 cm−1 band, which can be more clearly seen from the differ-
ence spectrum between b and c (Fig. 4B, curve d). Thus, we conclude
that the Raman bands at 1003 and 970 cm−1 correspond to the
(c)
(b)
1003
970
2−
1020
RuO4 species with the lower reduction temperature, which can
(a)
be tentatively assigned to the symmetric (ꢂs) and asymmetric (ꢂas
)
2−
stretching vibrations of the terminal Ru O bonds for the RuO4
species, respectively. The band at 1020 cm−1 is attributed to the
2−
RuO4
species formed at treatment temperatures above 1173 K,
1100
1050
1000
950
900
which showed the higher reduction temperature. After such reduc-
tion treatment and complete removal of the RuOx species, the two
bands at 1020 and 970 cm−1 were still observed, however, with
Raman shift (cm-1)
Fig. 4. H2 temperature-programmed reduction profiles (A) and in situ Raman spec-
tra (B) for RuOx/m-ZrO2 after treatment in flowing dry air at 1223 K before (a) and
after pre-reduction in H2 at 403 K (b) and 503 K (c). (d) In (B) is the difference Raman
spectrum between spectra (b) and (c).
similar intensity, reflecting the features of pure m-ZrO2 support
2−
(Fig. S1). These results confirm that both RuO4
species and m-
ZrO2 support for RuOx/m-ZrO2 coincidently contribute to the two
Raman bands.
Clearly, the RuOx species are present as isolated RuO42− species
in the RuOx/m-ZrO2 samples, but the observed changes in their H2
reduction and Raman features (Fig. 4A and B) indicate the evolu-
conditions. Isolated VOx species are present as the distorted tetra-
hedral mono-oxo structures containing one terminal V O bond and
three V–O bonds, while isolated MoOx and WOx species form both
mono-oxo (O Mo(–O-support)4 and O W(–O-support)4) and di-
oxo ((O )2Mo(–O-support)2 and (O )2W(–O-support)2) structures
on the support surfaces. These structures exhibit distinct spec-
troscopic features. The mono-oxo M O bonds under dehydrated
conditions show only symmetric (ꢂs) vibration Raman bands at
2−
tion of the molecular structures of the RuO4 species on m-ZrO2
induced by the thermal treatment at different temperatures. Such
structural evolution of the RuOx species has been scarcely reported
to date, which, however, can be understood by referring to the
extensive studies on the molecular structures of other isolated
metal oxides (MOx), e.g. MoOx, VOx, and WOx [21–23]. These sur-
face MOx species on different supports (e.g. SiO2, Al2O3, TiO2, and
ZrO2) generally possess M O and M–O bonds, and the number of
these bonds depends on the oxidation state of the metal oxides,
the surface structures of the supports, and the thermal treatment
higher frequencies relative to the bands for the di-oxo O
M O
bonds that exhibit both symmetric (ꢂs) and asymmetric (ꢂas) vibra-
tions separated by ca.10–30 cm−1 [22,24,25]. The mono-oxo and
di-oxo structures are known to be inter-convertible by chang-

62
H. Huang et al. / Catalysis Today 183 (2012) 58–64
100
Ru-OH Zr-OH
40
30
20
10
0
80
60
40
20
0
Methylformate
Methylformate
673 K
Rate
Rate
R
723 K
773 K
873 K
HCHO
HCHO
CO2
CO2
673 773 873 973 1073 1173 1273
Treatment temperature (K)
3900
3700
3500
3300
3400 3200
4000
3800
3600
-1
-1
Wave number (cm )
Wave number (cm )
Fig. 6. CH3OH oxidation rates normalized per Ru atom and selectivities at 373 K
as a function of treatment temperatures (673–1323 K) on RuOx/m-ZrO2 catalysts
at methanol conversions of ∼20% (0.2–2.2 Ru/nm2, 3.5 kPa CH3OH, 30 kPa of O2,
balance N2).
Fig. 5. Diffuse reflectance infrared Fourier transform spectra for RuOx/m-ZrO2 cat-
alysts after treatment in flowing dry air at different temperatures between 673 and
873 K.
molecule to form a new Ru–O–Zr bond at the higher treatment
temperatures.
2−
In summary, the molecular structures of the isolated RuO4
ing the treatment temperatures, water vapor pressures, etc. For
example, dioxo (O )2Mo(–O-support)2 can convert to mono-oxo
species on m-ZrO2 most likely evolve from umbrella-like di-oxo
(O )2Ru(OH)–Ob-Zr to tetrahedral di-oxo (O )2Ru–(O–Zr)2 with
increasing the treatment temperatures from 673 to 773 K, and then
to pyramidic mono-oxo O Ru–(O–Zr)4 above 1173 K, as depicted in
Scheme 1. Such changes in the structures of the RuO42− species led
to the parallel changes in their reducibility and catalytic reactivity
in the selective oxidation of methanol, as discussed below.
O
Mo(–O-support)4 on SiO2 by the release of a H2O molecule from
the interaction between one Mo O and two Si–OH bonds, although
the formation of the four Mo–O–Si bonds in the mono-oxo struc-
ture results in substantial strain of the SiO2 surfaces [22]. Similarly,
2−
we envisage the structural transformation of the RuO4
species
from tetrahedral di-oxo (O )2Ru(–O–Zr)2 to pyramidic mono-oxo
Ru(–O–Zr)4 with the observed Ru O stretching frequencies
O
around 1003 (ꢂs)/970 (ꢂas) and 1020 (ꢂs) cm−1, respectively, as a
result of the enhanced interaction between the RuO42− species and
m-ZrO2 surfaces when the treatment temperatures exceed 1173 K.
As mentioned above, the Ru O vibration band and H2 reduc-
tion peak for the supported RuOx species shifted sharply from 978
to 985 cm−1 and from 435 to 398 K, respectively (Figs. 2 and 3),
with increasing the treatment temperature from 673 to 773 K,
3.2. Methanol oxidation rates and selectivities on RuOx/m-ZrO2
catalysts
Fig. 6 shows CH3OH oxidation rates (normalized per Ru atom)
and selectivities at 373 K on RuOx/m-ZrO2 as a function of the
treatment temperatures in the range 673–1323 K. The rates and
selectivities were compared in the kinetic regime at similar CH3OH
conversions (∼20%) to avoid the known effects of CH3OH conver-
sion on the primary and secondary reactions [10]. MF, HCHO and
CO2 were only detected as the products. RuOx/m-ZrO2 treated at
673 K was not selective, and equally formed MF and CO2 (48.1%
vs. 47.2%). By increasing the treatment temperature to 773 K, the
selectivities to MF and CO2 inversely changed sharply to 73.8 and
16.0%, respectively. With further increasing the temperatures in
the range 773–1323 K, the CO2 selectivity gradually decreased to
as low as 4.3%, while the MF selectivity first decreased slightly to
68.8%, then sharply to 39.6% at the treatment temperatures of 1223
and 1323 K, respectively. The HCHO selectivity gradually increased
from 4.7 to 24.1% with increasing the treatment temperatures from
673 to 1223 K, and then abruptly increased to 56.1% at 1323 K, con-
currently with the observed decline in the MF selectivity between
1223 and 1323 K (Fig. 6). These results show that the RuO42− struc-
tures formed at higher temperatures are more selective for the
methanol oxidation to MF and HCHO relative to CO2. The decreased
MF selectivities on the samples treated above 773 K, especially
above 1273 K, appear to be relevant to their decreased ZrO2 sur-
faces (Table 1) that are known to be involved in the conversion of
which then changed more slowly above 773 K. Such progressive
2−
shifts indicate a change in the structures of the RuO4
species
following treatment above 773 K. Fig. 5 shows the DRIFT spec-
tra in the range 3000–4000 cm−1 for the RuOx/m-ZrO2 treated at
673–873 K. All spectra exhibit a band at 3771 cm−1, typical for iso-
lated Zr–OH groups, and a shoulder around 3680 cm−1 for bridging
Zr–OH groups [26–28]. The two bands remained almost unaltered
in the range 673–873 K. In contrast, an additional sharp band was
detected at 3855 cm−1 on the sample treated at 673 K, and it atten-
uated significantly and became nearly invisible with increasing the
treatment temperatures up to 773 K. This band was tentatively
assigned to Ru–OH stretching vibration. Such M–OH bonds have
also been detected spectroscopically, for example, on isolated VOx
species, present as a mono-oxo O V(OH)2–O–M structure on the
different support surfaces (e.g. SiO2, Al2O3, and ZrO2) [29,30]. By
referring to these previous studies, we tentatively propose that
2−
the RuO4 species form umbrella-like di-oxo (O )2Ru(OH)–O–Zr
structure with a Ru–OH bond between 673 and 773 K, which tends
to transform to tetrahedral di-oxo (O )2Ru–(O–Zr)2 above 773 K
most likely by reaction of Ru–OH with Zr–OH and release of a H2O

H. Huang et al. / Catalysis Today 183 (2012) 58–64
63
O
O
O
O
O
Ru
Ru
Ru
>1173 K
>773 K
O
O
O
O
O
O
OH
O
Zr Zr
Zr Zr
Zr
Zr
Zr
a
b
c
Scheme 1. Plausible changes in RuOx structures on m-ZrO2 with treatment temperatures from (a) umbrella-like dioxo (O )2Ru(OH)–O–Zr with two V O bonds, one Ru–OH
bond and one Ru–O–Zr bond to m-ZrO2 surface into (b) tetrahedral dioxo (O )2Ru–(O–Zr)2 with two V O bonds and two Ru–O–Zr bonds to m-ZrO2 surface from 673 to
773 K, and into (c) pyramidic mono-oxo O Ru–(O–Zr)4 with one V O bond and four Ru–O–Zr bonds to m-ZrO2 surface above 1173 K.
HCHO intermediates to MF [10,11], which accordingly led to the
observed increase in the HCHO selectivities (Fig. 6).
40
30
20
10
0
The CH3OH oxidation rates, as shown in Fig. 6, increased dra-
matically by a factor of 4.7 from 4.9 to 22.9 mol/mol Ru-h with
increasing the treatment temperature from 673 to 773 K, and then
more gradually to 39.0 mol/mol Ru-h as the temperature reached
1223 K. However, further increasing the treatment temperatures
to 1273 and 1323 K led to a sharp decline in the rates to 21.9 and
15.1 mol/mol Ru-h, respectively. From the structural characteriza-
tion of the RuOx species, it is known that most of the Ru atoms,
if not all, are exposed at surfaces and accessible to the reactants
on the RuOx/m-ZrO2 catalysts following treatment throughout the
whole temperature range 673–1323 K. Therefore, the rates per Ru
atom for these catalysts correspond to their turnover rates (TOFs),
and can be used to compare their intrinsic reactivity. Clearly, the
observed changes in the rates (Fig. 6) with the treatment tempera-
2−
tures are not due to the change in the accessibilities of the RuO4
species, but due to the change in their intrinsic reactivity with the
structural evolution induced by thermal treatment at the different
temperatures. Such change in the reactivity for the RuO42− species
is consistent with their differences in reducibility.
0
20
40
60
The reducibility of the RuO42− species was probed by TPR using
H2 as the reductant (Fig. 3). As discussed above, the peak tem-
H2 reduction rate (mol H2/mol Ru-h)
peratures for the reduction of Ru6+ to Ru0 in the di-oxo RuO4
2−
Fig. 7. Dependence of methanol oxidation rates (per Ru atom) at 393 K on initial H2
reduction rates (per Ru atom) at 385 K for RuOx/m-ZrO2 catalysts after treatment in
flowing dry air at different temperatures between 673 and 1223 K (0.2–1.6 Ru/nm2,
3.5 kPa CH3OH, 30 kPa of O2, balance N2).
structures shifted from 435.8 to 357.2 K by 78.6 K with increas-
ing the sample treatment temperatures from 673 to 1323 K, which
shows the more reducible nature of the di-oxo structures formed
at higher treatment temperatures. Our recent kinetics and isotopic
studies [10] have shown that the C–H bond activation steps using
lattice oxygen atoms limit the methanol oxidation rates on the
RuOx domains, involving the incipient RuOx reduction with the for-
mation of a low surface density of oxygen vacancies during the
steady-state catalysis (around 393 K). Thus, the initial stages of the
RuOx reduction are most relevant to the redox cycles required in
the methanol oxidation [10]. The initial H2 reduction rates at 385 K
catalysts [31–36]. Clearly, such correlation and structural under-
standings provide guidance for further tuning and improving the
reactivity of the RuOx-based catalysts via synthesis of the dioxo
2−
RuO4
structures with more reducible nature for the selective
oxidation of methanol as well as of other alcohols and light hydro-
carbons.
2−
for the di-oxo RuO4 structures were extracted from Fig. 3. They
increased with increasing the treatment temperature in the range
673–1223 K (Supporting information, Fig. S2), which are in paral-
lel with the increase in the methanol oxidation rates (i.e. TOFs)
at 373 K, as shown in Fig. 7. After the treatment temperatures
4. Conclusions
Treatment temperatures of m-ZrO2-suppored RuOx catalysts
significantly influence the structures of the RuOx species and their
catalytic performances in the methanol oxidation to methyl for-
2−
exceeded 1223 K, the less reducible mono-oxo RuO4 structures
2−
became the dominant RuOx species, as estimated from the H2 con-
sumption above (over 89%), leading to the observed decline in the
methanol oxidation rates for the two samples treated at 1273 and
1323 K (Fig. 6).
mate at 373 K. RuOx exists as highly dispersed RuO4
species
on m-ZrO2 with Ru surface densities of 0.2–2.2 Ru/nm2, following
treatment at different temperatures in the range 673–1323 K. The
molecular structures of the RuO4
like dioxo (O )2Ru(OH)–O–Zr with a Ru–OH bond into tetrahedral
dioxo (O )2Ru–(O–Zr)2 by increasing the treatment tempera-
tures from 673 to 773 K, and then into pyramidic mono-oxo
2−
species vary from umbrella-
2−
This correlation between the reducibility of the RuO4 species
and their catalytic reactivity reflects the Mars–van Krevelen redox
mechanism involving the C–H bond activation in the chemisorbed
species by lattice oxygen atoms as the kinetically relevant step in
the methanol oxidation on RuOx domains [3,10], as also observed
with the selective oxidation reactions of methanol and other alco-
hols as well as hydrocarbons on MoOx, VOx and other oxide-based
O
Ru–(O–Zr)4 above 1173 K, due to the stronger interaction
between the RuOx species and m-ZrO2 surface at the higher tem-
peratures, partly evident from the blue-shift of the Raman vibration
frequencies for the Ru O bonds. Such structural changes lead to a

64
H. Huang et al. / Catalysis Today 183 (2012) 58–64
parallel increase in the reducibility of the RuO42− species and their
methanol oxidation rates, reflecting the Mars–van Krevelen redox
mechanism using lattice oxygen atoms.
[8] N.G. Valente, L.A. Arrua, L.E. Cadus, Appl. Catal. A: Gen. 205 (2001) 201–214.
[9] G. Busca, Catal. Today 27 (1996) 457–496.
[10] H. Liu, E. Iglesia, J. Phys. Chem. B 109 (2005) 2155–2163.
[11] W. Li, H. Liu, E. Iglesia, J. Phys. Chem. B 110 (2006) 23337–23342.
[12] W. Li, H. Huang, H. Li, W. Zhang, H. Liu, Langmuir 24 (2008) 8358–8366.
[13] A. Guinier, Theorie et Technique de la Radiocristallographie, third ed., Dunod,
Paris, 1964.
Acknowledgments
[14] L.E. Fernandez, E.V. Sanchez, M. Panizza, M.M. Carnasciali, G. Busca, J. Mater.
Chem. 11 (2001) 1891–1897.
[15] S.Y. Mar, C.S. Chen, Y.S. Huang, K.K. Tiong, Appl. Surf. Sci. 90 (1995) 497–504.
[16] H. Yao, Y. Yao, J. Catal. 86 (1984) 254–265.
[17] A. Trovarelli, G. Dolcetti, C. Leitenburg, J. Kaspar, P. Finetti, A. Santoni, J. Chem.
Soc., Faraday Trans. 88 (1992) 1311–1319.
[18] P. Fornasiero, R. Di Monte, G. Ranga Rao, J. Kaspar, S. Meriani, A. Trovarelli, M.
Graziani, J. Catal. 151 (1995) 168–177.
This work was supported by the National Natural Science Foun-
dation of China (Grants 20825310 and 20973011) and National
Basic Research Project of China (Grants 2011CB201400 and
2011CB808700).
Appendix A. Supplementary data
[19] L. Chen, G. Gonzalez, J. Wang, L. Norena, A. Toledo, S. Castillo, M. Moran-Pineda,
Appl. Surf. Sci. 243 (2005) 319–328.
[20] H.Y.C. Chan, C.G. Takoudis, M.J. Weaver, J. Catal. 172 (1997) 336–345.
[21] E.L. Lee, I.E. Wachs, J. Phys. Chem. C 111 (2007) 14410–14425.
[22] S. Chempath, Y. Zhang, A.T. Bell, J. Phys. Chem. C 111 (2007) 1291–1298.
[23] E.I. Ross-Medgaarden, W.V. Knowles, T. Kim, M.S. Wong, W. Zhou, C.J. Kiely, I.E.
Wachs, J. Catal. 256 (2008) 108–125.
Raman spectra for m-ZrO2 support treated at different temper-
atures between 673 and 1323 K, and Initial H2 reduction rates at
385 K for RuOx/m-ZrO2 catalysts as a function of treatment tem-
peratures (673–1323 K).
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.cattod.2011.05.021.
[24] G. Busca, J. Raman. Spectrosc. 33 (2002) 348–358.
[25] B.M. Weckhuysen, J.M. Jehng, I.E. Wachs, J. Phys. Chem.
7382–7387.
B 104 (2000)
[26] P.A. Agron, E.L. Fuller, H.F. Holmes, J. Colloid Interface Sci. 52 (1975) 553–561.
[27] G.A.H. Mekhemer, Colloids Surf. A 141 (1998) 227–235.
[28] D. Spielbauer, G.A.H. Mekhemer, T. Riemer, M.I. Zaki, H. Knozinger, J. Phys.
Chem. B 101 (1997) 4681–4688.
[29] D.E. Keller, D.C. Koningsberger, B.M. Weckhuysen, J. Phys. Chem. B 110 (2006)
14313–14325.
References
[1] A. Wittstock, V. Zielasek, J. Biener, C.M. Friend, M. Baumer, Science 327 (2010)
319–322.
[2] J. Liu, E. Zhan, W. Cai, J. Li, W. Shen, Catal. Lett. 120 (2008) 274–280.
[3] J.M. Tatibouet, Appl. Catal. A 148 (1997) 213–252.
[4] E. Tronconi, A.S. Elmi, N. Ferlazzo, P. Forzatti, G. Busca, P. Tittarelli, Ind. Eng.
Chem. Res. 26 (1987) 1269–1275.
[5] J.S. Lee, J.C. Kim, Y.G. Kim, Appl. Catal. 57 (1990) 1–30.
[6] A. Guerrero-Ruiz, I. Rodriguez-Ramos, J.L.G. Fierro, Appl. Catal. 72 (1991)
119–137.
[30] D.E. Keller, S.M.K. Airaksinen, A. Outi Krause, B.M. Weckhuysen, D.C. Konings-
berger, J. Am. Chem. Soc. 129 (2007) 3189–3197.
[31] H. Liu, P. Cheung, E. Iglesia, J. Phys. Chem. B 107 (2003) 4118–4127.
[32] J. Dobler, M. Pritzsche, J. Sauer, J. Am. Chem. Soc. 127 (2005) 10861–10868.
[33] W. Zhang, A. Desikan, S.T. Oyama, J. Phys. Chem. 99 (1995) 14468–14476.
[34] B. Kilos, A.T. Bell, E. Iglesia, J. Phys. Chem. B 113 (2009) 2830–2836.
[35] M.C. Abello, M.F. Gomez, L.E. Cadus, Catal. Lett. 83 (1998) 185–192.
[36] K. Chen, A.T. Bell, E. Iglesia, J. Catal. 209 (2002) 35–42.
[7] M. Ai, J. Catal. 77 (1982) 279–288.