Performance and characterization of supported rhenium oxide catalysts for selective oxidation of methanol to methylal

Article abstract of DOI:10.1021/jp013770l

The unique performances of supported rhenium oxide catalysts for the selective oxidation of methanol to methylal (dimethoxymethane) [3CH₃OH + 1/2O₂ a?? CH₂(OCH₃)₂ + 2H₂O] were examined in a fixed-bed flow reactor and in a pulse reactor. Rhenium oxides supported on ?±-Fe₂O₃, ?3-Fe₂O₃, and V₂O₅ among many oxide supports showed high activities (15-49% conversion) and selectivities (90-94% selectivity) for catalytic methylal synthesis at 513 K. These Re catalysts are the first to show good performances for one-stage methylal synthesis from methanol. The catalysts were characterized by XRD, XPS, TPR, TPD, and FTIR spectroscopy to provide insight into the active Re species relevant to the selective oxidation of methanol to methylal. Rhenium oxides on ?±-Fe₂O₃ were dispersed as tetrahedral ReO₄ species up to one monolayer at 0.7 wt % Re loading. It was suggested that active Re-Fe-O mixed-oxide clusters were formed above 1 wt % Re. ReO₂ crystallites were also observed for the Re/?±-Fe₂O₃ samples with Re loadings above 6 wt %. On ?3-Fe₂O₃, which has a higher surface area, no ReO₂ species were observed. It is suggested that the redox capability of rhenium oxides, Re^6-7+ a?? Re⁴⁺, is responsible for the catalytic performance, although appropriate Lewis acidity of the rhenium oxides is also necessary for the acetalization of formaldehyde with methanol to methylal. The oxide supports prevent rhenium oxides from being sublimated and reduced to ReO₂ particles.

Full text of DOI:10.1021/jp013770l

J. Phys. Chem. B 2002, 106, 4441-4449
4441
Performance and Characterization of Supported Rhenium Oxide Catalysts for Selective
Oxidation of Methanol to Methylal
†
Youzhu Yuan and Yasuhiro Iwasawa*
Department of Chemistry, Graduate School of Science, The UniVersity of Tokyo, Hongo, Bunkyo-ku,
Tokyo 113-0033, Japan
ReceiVed: October 10, 2001; In Final Form: January 4, 2002
The unique performances of supported rhenium oxide catalysts for the selective oxidation of methanol to
1
methylal (dimethoxymethane) [3CH
3
OH + /
2
O
2
f CH
2
(OCH )
3 2
+ 2H
2
O] were examined in a fixed-bed
, γ-Fe , and V among many
flow reactor and in a pulse reactor. Rhenium oxides supported on R-Fe
2
O
3
O
2 3
2 5
O
oxide supports showed high activities (15-49% conversion) and selectivities (90-94% selectivity) for catalytic
methylal synthesis at 513 K. These Re catalysts are the first to show good performances for one-stage methylal
synthesis from methanol. The catalysts were characterized by XRD, XPS, TPR, TPD, and FTIR spectroscopy
to provide insight into the active Re species relevant to the selective oxidation of methanol to methylal.
Rhenium oxides on R-Fe
loading. It was suggested that active Re-Fe-O mixed-oxide clusters were formed above 1 wt % Re. ReO
crystallites were also observed for the Re/R-Fe samples with Re loadings above 6 wt %. On γ-Fe
which has a higher surface area, no ReO species were observed. It is suggested that the redox capability of
a Re , is responsible for the catalytic performance, although appropriate Lewis
acidity of the rhenium oxides is also necessary for the acetalization of formaldehyde with methanol to methylal.
The oxide supports prevent rhenium oxides from being sublimated and reduced to ReO particles.
2 3 4
O were dispersed as tetrahedral ReO species up to one monolayer at 0.7 wt % Re
2
2
O
3
2 3
O ,
2
rhenium oxides, Re6
-7+
4+
2
1
. Introduction
We have found that crystalline SbRe2O6 is active for the
catalytic selective oxidation of methanol to form methylal
The ability of rhenium oxides to take a variety of oxidation
(dimethoxymethane), CH2(OCH3)2, where three methanol mol-
1
-5
states that are illustrated in both binary and ternary oxides,
ecules are incorporated into a methyal molecule [3CH3OH +
might allow for the exploitation of new catalytic chemistry and
applications. Supported Re catalysts are highly active and
selective in metathesis reactions.6,7 Rhenium is also an important
1
22,23
/
2O2 f CH2(OCH3)2 + 2H2O].
The selectivity to methylal
reached 92.5% at a conversion of 6.5% at 573 K. Methylal is
used as a gasoline additive, a solvent in the perfume industry,
a key intermediate for preparing high-concentration formalde-
hyde, and a reagent in organic syntheses. The activity and
selectivity of the SbRe2O6 catalyst for the methylal synthesis
were much higher than those of other catalysts reported to date,
8
component in a rheniforming catalytic process. Further, sulfided
Re catalysts might be more active than conventional Co-Mo-S
catalysts in hydrodesulfurization.9 Nevertheless, rhenium
oxides have not been employed as a key species in catalytic
selective oxidation reactions because of sublimation under
reaction conditions.11
,10
24
25
26,27
such as V/TiO2, V-Mo-O, PMoH-5.75/SiO2,
Mo/
28
29
30
MCM-41, MoO3(100) plane, and so on. Industrially,
methylal is produced by two-stage processes: methanol oxida-
tion to formaldehyde on silver and ferric molybdate catalysts
and dehydrative condensation of the formaldehyde with metha-
nol with sulfuric acid. Thus, the one-stage heterogeneous
reaction process on the SbRe O catalyst has economical and
Few reports have been devoted to this topic, and they have
described the failure of the use of inorganic Re compounds for
catalytic oxidation.12 These negative contributions of rhenium
oxides to oxidation processes is somewhat surprising because
the oxides of tungsten and osmium, which are neighbors of
rhenium in the periodic table, work as efficient epoxidation and
2
6
environmental benefits in the production of methylal. It was
1
3-16
cis-hydroxylation catalysts, respectively.
It was reported
suggested that the performance of the crystalline SbRe O6
2
that methyltrioxorhenium (CH3ReO3) was active for alkene
catalyst was due to the rhenium oxide connecting with Sb-O
epoxidation at or below room temperature.1
7,18
In contrast to
chains. The surface area of the crystalline SbRe O is as low
2
6
2
-1
other metal-containing epoxidation catalysts, almost no decom-
position of H2O2 occurred with CH3ReO3. The unique oxidation
property of CH3ReO3 has generated much interest in high-
valence oxo-organometallic chemistry.1
as 1 m g , which might be a disadvantage in applications to
catalytic processes.
Recently, we revealed the ability of supported rhenium oxides
9-21
31
to selective catalyze the oxidation of methanol to methylal.
The present study reports, in detail, the performances of sup-
ported rhenium oxide catalysts for the selective oxidation of
methanol of methylal, the effects of the support and the Re load-
ing on the performance, and the characterization of the active
Re species by XRD, XPS, TPR, TPD, and FTIR spectroscopy.
*
To whom correspondence should be addressed. E-mail: iwasawa@
chem.s.u-tokyo.ac.jp. Fax: 81-3-5800-6892.
†
Permanent address: Department of Chemistry, State Key Laboratory
for Physical Chemistry of Solid Surface, Xiamen University, Xiamen
61005, China.
3
1
0.1021/jp013770l CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/21/2002

4
2
442 J. Phys. Chem. B, Vol. 106, No. 17, 2002
Yuan and Iwasawa
-
1
. Experimental Section
recorded at a ramping rate of 10 K min . The effluent gas
from the reactor was analyzed by an on-line gas chromatograph.
TPR spectra were recorded in the same flow reactor system.
The catalyst (30 mg) was pretreated at a given temperature under
a He flow for 1 h and then cooled to 298 K under a He flow.
The He flow was switched to a 7% H2/Ar flow, and then TPR
2.1. Catalyst Preparation. Supported rhenium oxide catalysts
were prepared by an incipient wetness impregnation method
using an aqueous solution of ammonium perrhenate (NH4ReO4;
Soekawa Chemical Co., 99.99% purity), followed by evapora-
tion to remove the solvent on a water bath and drying at 383 K
for 12 h. Commercially available oxides such as TiO2 (rutile),
TiO2 (anatase), V2O5, ZrO2, R-Fe2O3, γ-Fe2O3, SiO2, R-Al2O3,
Sb2O3, Bi2O3, and MoO3 were used as supports. The Re loadings
were varied in the range of 0-10 wt % as Re/support. The
samples were put into a glass fixed-bed reactor in a flow system
and heated to 673 K at a heating rate of 4 K min- in a He
flow and held at 673 K for 6 h. The samples thus obtained were
further treated in situ in the fixed-bed flow reactor under He
flow at 573 K for 1 h before use as catalysts. Unsupported
rhenium oxides such as ReO3 (Soekawa) and ReO2 (Soekawa)
were also pretreated at 673 K in a similar way.
-1
measurements were made at a ramping rate of 10 K min . The
effluent gas from the reactor was analyzed by an on-line gas
chromatograph.
FTIR spectra were measured on a Perkin-Elmer Spectrum
-1
2
000 instrument with a MCT detector at a resolution of 4 cm .
ReOx/R-Fe2O3 was pressed into a self-supporting thin disk,
which was placed in a quartz IR cell with CaF2 windows
combined in a vacuum system. The sample in the IR cell was
heated at a rate of 4 K min to 673 K and held at 673 K for
0 min under vacuum. Then, the sample was cooled to 298 K
and exposed to dry pyridine (3 × 10 Pa), followed by
evacuation at 373 K for 30 min. The FTIR spectra were
measured after evacuation at given temperatures from 373 to
1
-1
3
2
2
.2. Catalytic Methanol Oxidation. The catalytic reactions
on the supported and unsupported rhenium oxide catalysts were
carried out in a conventional fixed-bed flow reactor. The
catalytic performances were examined in the temperature range
6
23 K and recorded as difference spectra from the spectrum
for the sample before pyridine adsorption. For comparison,
spectra of the R-Fe2O3 support were also measured.
4
h
73-553 K under the reaction conditions GHSV ) 40 000 mL
-
1
-1
gcat and He/O2/MeOH ) 86.3/9.7/4.0 (mol %) at 1 atm.
3
. Results
.1. Methanol Oxidation on Supported Re Catalysts.
The mixture was regulated by mass flow controllers. Methanol
Wako Chemical Co., purity 99.8%) was introduced into the
3
(
Selective methanol oxidation to methylal on several oxide-
supported ReOx catalysts was performed at 513 K under the
reaction conditions of GHSV ) 40 000 mL h gcat and He/
O2/MeOH ) 86.3/9.7/4.0 (mol %). Table 1 shows the perfor-
mances of the supported rhenium oxide catalysts, where the
flow reactor by bubbling He gas through a glass saturator filled
with methanol. The outlet stream line from the reactor to a gas
chromatograph was heated at about 423 K to avoid condensation
of reaction products. The products were analyzed by an on-
line gas chromatograph using Porapak N (3 m) and Unibeads
-
1
-1
1
6
00% conversion of methanol corresponds to a reaction rate of
5.4 × 10 mol h gcat . The performances of bulk rhenium
(3 m) C columns.
-
3
-1
-1
2.3. Catalyst Characterization. Surface areas of the samples
oxides (ReO3 and ReO2) and two typical supports (R-Fe2O3 and
V2O5) are also listed in Table 1 for comparison. It was found
that rhenium oxides supported on TiO2 (rutile and anatase),
V2O5, ZrO2 (monoclinic), Fe2O3 (R and γ), and R-Al2O3 were
active in order of the reaction rates per gram of Re for the
supports TiO2(anatase) > TiO2 (rutile) > γ-Fe2O3 > ZrO2 >
V2O5 > R-Al2O3 g R-Fe2O3. Of the samples, Re/V2O5, Re/γ-
Fe2O3, Re/R-Fe2O3, Re/ZrO2, and Re/R-Al2O3 showed selectivi-
ties of 88-94% to methylal at 513 K. Rhenium oxides supported
on SiO2 produced a significant amount of COx (CO2 and CO),
and those supported on Sb2O3 and Bi2O3 showed no activity at
were measured by a BET nitrogen adsorption method at 77 K
using a BELSORP36 machine.
XRD patterns were measured on a Rigaku Miniflex goni-
ometer. The analysis was carried out in a continuous θ/2θ scan
reflection mode using Cu KR radiation (λ ) 0.154 18). The
anode was operated at 30 kV and 15 mA. The 2θ angles were
scanned from 5 to 80° at a rate of 2°/min.
XPS spectra were measured on a Rigaku XPS-7000 spec-
trometer using Mg KR radiation (1253.6 eV) with X-ray power
of 200 W (accelerating voltage 20 kV, emission current 10 mA).
Samples were pressed into thin disks, placed on holders,
5
13 K (Table 1). R-Fe2O3 and V2O5 themselves were almost
-
5
outgassed to less than 2.6 × 10 Pa in a prechamber, and
transferred to an analysis chamber. The procedures for sample
preparation for the XPS measurement were conducted in a
glovebox filled with high-purity N2 to avoid the samples contact
with air as much as possible. The binding energy was referred
to 284.6 eV for C 1s. The peak deconvolution and fitting were
performed using SpXzeigR2.1 software running with Igor Pro
and a Gaussian-Lorentzian line shape, with both the spin-
orbit splitting and the relative intensity of the spin-orbit
components fixed.
inactive for methylal formation, although V2O5 produced
formaldehyde. When these oxides were supported by rhenium
oxides, the obtained Re/R-Fe2O3 and Re/V2O5 samples showed
selectivities to methylal as high as 90.5-93.7% at 513 K. ReO3
was most selective (99% selectivity), while the selectivity of
ReO2 was lower (64.6% for commercial ReO2; 84.2% for ReO2
prepared by decomposition of NH4ReO4 at 673 K). When the
mixture of He/MeOH/O2 was admitted to Re2O7 at 513 K, the
color changed from yellow (Re2O7) to red (ReO3) in less than
3 min, accompanied by sublimation of some of the Re O , and
2
7
Pyridine TPD spectra were recorded in a flow reactor
equipped with an on-line gas chromatograph that was monitored
by a GC workstation. The catalyst (200 mg) was pretreated at
a given temperature under a He flow (Linder Co., purity
the catalyst showed a selectivity to methylal of >90% after 10
min of time on stream. This indicates that bulk Re2O7 is not a
stable phase under catalytic reaction conditions.
Plots of selectivity versus conversion per unit surface area
for the supported ReOx catalysts are shown in Figure 1. It was
found that Re/R-Fe2O3, Re/γ-Fe2O3, Re/V2O5, and Re/ZrO2
exhibited high catalytic performances for the selective methylal
synthesis. No deactivation of the supported rhenium oxides
occurred for 6 h of time on stream at 513 K. There was no
significant change in the surface area of the oxide samples with
99.995%) for 1 h and then cooled to 298 K under a He flow.
Pyridine was introduced into the flow reactor by bubbling He
gas through a glass saturator filled with pyridine. Pyridine
adsorption was conducted for 30 min at 298 K. The flow
involving pyridine was switched to pure He for at least 1 h
until a straight GC baseline was obtained. The TPD spectra were

Rhenium Oxide Catalysts for Methanol Oxidation
J. Phys. Chem. B, Vol. 106, No. 17, 2002 4443
TABLE 1: Catalytic Methanol Oxidation on Supported Rhenium Oxide Catalysts at 513 Ka
MeOH conversion
selectivity(mol %)
S
BET
(m g-1)
rate
(mmol h
2
-1
Re-1)
b
catalyst
mol %
g
CH
2
(OCH
3
)
2
HCHO
3
(CH )
2
O
HCOOCH
3
CO
x
1
1
1
1
1
1
1
1
1
1
1
1
R-Fe
V
ReO
ReO
ReO
0 wt % Re/TiO
0 wt % Re/TiO
0 wt % Re/V
0 wt % Re/ZrO
0 wt % Re/R-Fe
0 wt % Re/γ-Fe
2
-rutile (1)
5
50
6
9
3
16
23
36
10
1
1
5
53.7
59.5
21.5
35.8
15.5
48.4
51.6
15.1
16.3
0.0
0.0
9.1
0.0
9.3
351.2
389.1
140.6
234.1
101.4
319.2
337.5
98.8
83.1
78.5
93.7
89.4
90.5
91.0
82.7
60.7
88.3
0.0
1.9
4.1
0.0
2.0
2.0
2.4
2.6
1.3
2.8
0.0
0.0
0.0
0.0
91.5
0.0
6.4
0.0
0.7
1.1
4.3
trace
1.0
1.0
0.6
trace
trace
0.0
0.0
19.0
0.0
9.1
11.7
0.0
7.6
6.0
4.6
9.5
11.9
5.9
0.0
0.0
0.0
0.0
0.0
0.5
10.2
5.9
5.2
4.6
2.0
1.0
0.5
1.0
4.6
26.1
2.9
0.0
0.0
1.0
0.0
2
-anatase (2)
2
O
5
(3)
(4)
2
O
2 3
(5)
(6)
2
O
3
3
c
0 wt % Re/Fe
0 wt % Re/SiO
0 wt % Re/R-Al
0 wt %Re/Sb
0 wt % Re/Bi
0 wt % Re/MoO
(Soekawa)
2
O
2
(7)
O
2 3
(8)
106.6
0.0
0.0
2 3
O
2
O
3
0.0
80.0
0.0
3
(9)
59.5
2
O
3
3
6
1
7
0.0
O
5
(Soekawa)
(Soekawa)
e(Soekawa)
10.8d
10.2
50.0
7.7
1.0
7.4
0.5
2.0
8.9
trace
0.0
16.8
1.0
2
3
2
x
12.4
65.3
10.0
99.0
64.6
84.2
1
a
GHSV ) 40 000 mL h-1
4
cat-1, He/O
/MeOH ) 86.3/9.7/4.0 (mol %). b CO
, followed by filtration, washing with deionized water, drying at 393 K, and
was obtained by temperature-programmed calcination (4 K/min) of NH ReO at
73 K for 6 h in He, and the crystalline phase thus obtained is mainly due to monoclinic ReO by XRD. The numbers in parentheses are used for
numbering the supported ReO catalysts in Figure 1.
+ CO. c Fe
g
2
2
2 3
O was prepared by adding an aqueous solution of
NH
calcining at 673 K for 4 h in air. Units of mol h
OH (Wako, purity 99.9%) to an aqueous solution of Fe(NO )
3 3
d
-1
-1 e
V
g
.
ReO
x
4
4
6
2
x
Figure 2. Variations of the conversion and selectivity with Re loading
in the Re/R-Fe catalysts. Reaction temperature ) 513 K, GHSV )
2
1
O
3
-
-1
4
0 000 mL h
2
gcat , He/O /MeOH ) 86.3/9.7/4.0 (mol %). O, MeOH
conversion; ], methylal selectivity.
Figure 1. Dependency of the selectivity to methylal on the conversion
per unit surface area (0) and dependency of the reaction rate per gram
with increasing Re loading and reached maximum values of
18.1 and 93.7%, respectively, at 3 wt % Re. Above this loading
of Re on the surface area (O) of the supported ReO
wt % Re: 1, Re/TiO (rutile); 2, Re/TiO (anatase); 3, Re/V
Re/ZrO ; 5, Re/R-Fe ; 6, Re/γ-Fe ; 7, Re/SiO ; 8, Re/R-Al
Re/MoO
x
catalysts with 10
; 4,
; 9,
the conversion and selectivity decreased slightly to 15.5 and
90.5%, respectively, at 10 wt % Re.
2
2
2 5
O
O
2 3
2
2
O
3
O
2 3
2
3
.
Figure 3 shows the methanol conversion and the selectivities
to several products including methylal, formaldehyde, and
methylformate as functions of reaction temperature. The metha-
nol conversion increased with increasing temperature, whereas
the methylal selectivity passed through a maximum at 513 K
and then decreased with increasing temperature. The decrease
in selectivity was due to an increase in formaldehyde formation,
and the formation of undesired products such as methylformate
and COx (CO2 and CO) was still low. These results confirm
that supported Re catalysts are promising catalysts for selective
methanol oxidation.
3.3. Pulse Reaction. Pulses of ca. 1 mL of a mixture of 96.0/
4.0 (mol %) He/MeOH in the absence of O2 were introduced
onto the 10 wt % Re/R-Fe2O3 catalyst at 513 K at an interval
of about 15 min. Figure 4 presents plots of the methanol
conversion and the selectivities to several products versus the
number of the methanol pulses. Methylal was formed in the
first to fourth methanol pulses, but the selectivity decreased to
zero at the fifth pulse. The formation of formaldehyde also
Re loading. The reaction rates per gram of Re are also plotted
versus surface area in Figure 1, which provides information on
the efficiency per 1 wt % Re quantity for each catalyst. The
reaction rate increased with the surface area, with four straight
lines for four oxide groups indicating the different reactivities
of the rhenium oxide species. This is consistent with the results
of the plots of selectivity vs conversion/surface area. For
example, Re/R-Fe2O3, Re/V2O5, Re/ZrO2, and Re/γ-Fe2O3 are
classified in the same group, but Re/TiO2 (anatase), Re/R-Al2O3
and Re/MoO3 are classified in another group. The rhenium oxide
species of the catalysts in the same group are regarded to have
similar activities for the catalytic selective oxidation of methanol
to methylal.
3
.2. Effects of Re Loading and Reaction Temperature.
Figure 2 shows the methanol conversion and the selectivity to
methylal as a function of Re loading in the Re/R-Fe2O3 catalyst.
Both methanol conversion and methylal selectivity increased

4444 J. Phys. Chem. B, Vol. 106, No. 17, 2002
Yuan and Iwasawa
Figure 3. Variations of the conversion and selectivity with reaction
temperature in methanol conversion to methylal on the 10 wt % Re/
-
1
-1
R-Fe
O
2 3
catalyst with GHSV ) 40 000 mL h
2
gcat and He/O /MeOH
)
86.3/9.7/4.0 (mol %). [, MeOH conversion; ], methylal selectivity;
b, methyl formate selectivity; 2, formaldehyde selectivity.
Figure 5. XRD patterns for the Re/R-Fe
available ReO
Fe , (d) 10 wt % Re/R-Fe
h, (e) 6 wt % Re/R-Fe , (f) 10 wt % Re/Fe
the decomposition of NH ReO at 673 K), (g) 10 wt % Re/γ-Fe
h) commercially available R-Fe , (i) commercially available γ-Fe
2
O
3
catalysts: (a) commercially
3
, (b) commercially available ReO
after methanol oxidation at 513 K for
(Fe prepared by
2
, (c) 10 wt % Re/R-
O
2 3
2 3
O
2
2
O
4
3
2
O
3
2 3
O
4
2 3
O ,
(
2
O
3
2 3
O .
Figure 4. Pulse experiments: conversion and selectivity vs number
of methanol pules [He/MeOH ) 96/4 (mol %)] on the 10 wt % ReO
x
/
2 3
R-Fe O catalyst at 513 K. [, MeOH conversion; ], methylal
selectivity; 2, formaldehyde selectivity; b, CO
x
selectivity; X, CH
4
selectivity.
decreased but more gently with increasing number of methanol
pulses and continued throughout the 10 pulses. Methylal
formation might be favorable on the catalyst surfaces with more
lattice oxygen as compared to formaldehyde formation. The
formation of dimethyl ether (not shown) also decreased with
the number of pulses, in agreement with the decrease in methylal
formation because both reactions involve intermolecular dehy-
dration processes. Methylformate formation (not shown) de-
creased with the number of methanol pulses, whereas the
formations of H2, CH4, and CO2 increased throughout the 10
pulses. The molecular ratio of H2, CH4, and CO2 was ap-
proximately 2:1:1, which agrees with the assumptions that
methylformate is produced by a dehydrogenative condensation
of formaldehyde with methanol (CH3OH f HCHO + H2,
HCHO + CH3OH f HCOOCH3 + H2) and CH4 and CO2 are
produced by decomposition of the methylformate (HCOOCH3
Figure 6. Re 4f and Fe 3p XPS spectra for iron-oxide supported ReO
catalysts with 10 wt % Re: (a) NH ReO /R-Fe
Fe catalyst, (c) Re/γ-Fe catalyst, (d) Re/Fe
x
4
4
2
O
3
catalyst, (b) Re/R-
2 3
O catalyst.
O
2 3
O
2 3
crystallization to ReO . In the case of the 10 wt % Re/R-Fe O
2
2
3
2
-1
catalyst with a low surface area (3 m g ), small peaks due to
ReO were observed, as shown in Figure 5c. The state of the
2
rhenium oxides did not change after catalytic methanol oxidation
at 513 K for 2 h, as shown in Figure 5d.
Figure 6 shows the Re 4f and Fe 3p XPS spectra for Re/R-
Fe O , Re/γ-Fe O , and Re/Fe O with 10 wt % Re. For
2
3
f CH4 + CO2).
.4. Characterization by XRD and XPS. Figure 5 shows
2
3
2
3
2
3
3
comparison, the XPS spectrum for as-prepared NH4ReO4/R-
Fe2O3 is also shown in Figure 6. The XPS spectra for the Re/
R-Fe2O3 catalyst showed a peak at 42.4 eV, which is assigned
the XRD patterns for various Fe2O3-supported ReOx catalysts,
bulk rhenium oxides, and iron oxides. Calcination of NH4ReO4
at 673 K in He produced monoclinic crystalline ReO2 as
evidenced by XRD,32 whereas the 6 wt % Re/R-Fe2O3, 10 wt
4
+
to the Re 4f7/2 level for Re species, and a peak at 47.8 eV,
6-7+
which is assigned to the Re 4f5/2 level for the Re
species.
%
Re/γ-Fe2O3, and 10 wt % Re/Fe2O3 [prepared from Fe(NO3)3]
The most intense peak around 45 eV is then considered as the
6
-7+
catalysts exhibited no diffraction pattern except for the XRD
patterns for the R-Fe2O3 or γ-Fe2O3 support. The results imply
that rhenium oxides are dispersed on the supports without
sum of the Re 4f7/2 peak for the Re
species and the Re
4
+
4f5/2 peak for the Re species. The peak around 45 eV was
deconvoluted into two peaks at 45.4 and 44.8 eV for Re6
-7+

Rhenium Oxide Catalysts for Methanol Oxidation
J. Phys. Chem. B, Vol. 106, No. 17, 2002 4445
x 2 3
Figure 9. Re 4f and Fe 3p XPS spectra for the 10 wt % ReO /Fe O
catalysts: (a) after pretreatment at 673 K for 6 h in He, (b) after 10
methnaol pulses (He/MeOH) at 513 K, and (c) after catalytic methanol
oxidation at 513 K for 2 h.
2 3
Figure 7. Re 4f and Fe 3p XPS spectra for the Re/R-Fe O catalysts.
Re loading (wt %): (a) 0.0, (b) 0.5, (c) 1.0, (d) 2.0, (e) 3.0, (f) 6.0, (g)
1
0.0.
Figure 10. TPR profiles for (a) commercially available ReO
2
, (b)
commercially available ReO , (c) commercially available NH ReO ,
3
4
4
Figure 8. Dependencies of the intensities of Re 4f and Fe 3p XPS
peaks normalized to the intensity of O 1s peaks on Re loading in the
(
(
(
d) commercially available R-Fe
2
O
3
, (e) 1.0 wt % Re/R-Fe
catalyst, (g) 6.0 wt % Re/R-Fe
catalyst, (i) 10.0 wt % NH ReO
2
O
3
catalyst,
f) 2.0 wt % Re/R-Fe
2
O
3
O catalyst,
2 3
Re/R-Fe
2
O
3
catalysts.
h) 10.0 wt % Re/R-Fe
O
2 3
4
4 2 3
/R-Fe O .
+
4
f7/2 and Re4 4f5/2, respectively. The binding energy of 45.4
7
+
Figure 9 shows the Re 4f XPS spectra for the 10 wt % Re/
eV is lower by 1.0-1.4 eV than that reported for Re2O7 (Re )
6+
33,34
R-Fe2O3 catalysts (a) after treatment at 673 K for 6 h in He, (b)
after the 10 methanol pulses at 513 K, and (c) after catalytic
methanol oxidation at 513 K for 2 h. After the 10 methanol
pulses at 513 K, the Re 4f XPS spectrum exhibited a significant
but higher by 1.0-1.2 eV than that for ReO3 (Re ).
The
NH4ReO4/R-Fe2O3 catalyst with Re7 exhibited the binding
+
energies of 48.1 and 45.7 eV for Re 4f5/2 and 4f7/2, respectively
(
Figure 6a). In contrast, the Re/γ-Fe2O3 and Re/Fe2O3 catalysts
6
-7+
4+
6-7+
decrease in the peak intensity for the Re
species, whereas
showed no Re XPS peaks, and only Re
present at the catalyst surfaces (Figure 6c and d). Fe 2p and 3p
species were
4
+
the intensity of the Re peaks increased. After the catalytic
methanol oxidation at 513 K, the XPS spectrum had features
similar to those in the spectrum for the fresh catalyst.
bands appeared at binding energies of 711.0 and 55.4 eV,
respectively, which are the values typical of Fe3+ for R-Fe2O3.
Figure 7 shows the Re 4f and Fe 3p XPS spectra for the
3.5. H2 TPR and Pyridine TPD and FTIR Spectroscopy.
To examine the redox and acidic properties of the Re/R-Fe2O3
catalysts (as selective methanol conversion to methylal involves
both oxidation and dehydration steps), H2 TPR and pyridine
TPD and FTIR data were collected for the catalysts and
reference compounds. The TPR profiles for ReO2, ReO3, NH4-
ReO4, R-Fe2O3, 10 wt % NH4ReO4/R-Fe2O3, and Re/R-Fe2O3
catalysts with different Re loadings are shown in Figure 10.
The TPR peak for ReO2 was as high as 730 K, whereas the
TPR peak for ReO3 was observed at 675 K, which was slightly
higher than that for NH4ReO4. Reduction of R-Fe2O3 occurred
around 700 K, which is ascribed to the reduction of Fe2O3 to
Re/R-Fe2O3 catalysts with different Re loadings. The catalysts
with Re loadings of 0.5 and 1.0 wt % contained only Re
6
-7+
4+
species. For Re loadings of 2.0-10.0 wt %, Re peaks appeared
in addition to the Re6
-7+
peaks. The Fe 3p binding energies
were almost the same as those for R-Fe2O3. The XPS intensities
of the Re (all 4f5/2 and 4f7/2) and Fe 3p peaks are plotted against
the Re loading in Figure 8. The XPS intensities are normalized
by the O (1s) intensity in each catalyst. There are clear breaks
at 0.7-0.8 wt % in both plots, although the data are somewhat
scattered. The features of Figure 8 can be explained by the
Stranski-Krastanov growth mode for rhenium oxides on
3
5
36
R-Fe2O3.
Fe3O4. The 10 wt % NH4ReO4/R-Fe2O3 catalyst showed TPR

4446 J. Phys. Chem. B, Vol. 106, No. 17, 2002
Yuan and Iwasawa
4
. Discussion
.1. Catalytic Properties of Supported Rhenium Oxides
4
for the Selective Oxidation of Methanol to Methylal. Many
catalysts have been screened to improve the performance in the
selective methanol conversion to methylal, but no supported
3
1
24
rhenium oxides have been reported to date. V/TiO2, V-Mo-
25
26,27
28
29,30
O, PMoH-5.75/SiO2,
Mo/MCM-41, and others
have
been claimed to be active for catalytic methylal synthesis from
methanol. The selectivities toward methylal on V/TiO2 and
PMoH-5.75/SiO2 were as low as 40-56%. The Mo/MCM-41
catalyst showed a selectivity of 76.2% at a low conversion of
0
.7% at 543 K, but rapid deactivation was a serious problem
because of the a significant leaching of Mo species from MCM-
1. Recently, we found that crystalline SbRe2O6 exhibited
excellent performances in the catalytic selective oxidation of
4
2
2,23
Figure 11. Pyridine TPD spectra for (a) commercially available ReO
2
,
methanol to methylal.
A high selectivity of 92.5% to
(
(
(
(
b) commercially available ReO
3
, (c) commercially available R-Fe
2
O
3
,
methylal at 6.5% conversion at 573 K was achieved on the new
crystalline catalyst, where the high performance of SbRe2O6
was ascribed to the rhenium oxide species stabilized by specific
connections with to antimony oxides at the crystal surface.
Extending our knowledge about and achievements with the
crystalline catalyst, we have found a new catalytic property of
supported rhenium oxides that provides a promising one-stage
process for catalytic methylal synthesis from methanol that is
preferable economically and environmentally to the two-stage
industrial processes involving methanol oxidation to formalde-
hyde on silver or ferric molybdate catalysts and dehydrative
condensation (acetalization) of the formaldehyde with methanol
d) 1.0 wt % Re/R-Fe
2
O
3
catalyst, (e) 2.0 wt % Re/R-Fe
2
O
O
3
catalyst,
catalyst,
f) 3.0 wt % Re/R-Fe
h) 10.0 wt % Re/R-Fe
2 3
O catalyst, (g) 6.0 wt % Re/R-Fe
2
3
2
O
3
catalyst. ReO
2
and ReO
3
were pretreated at
5
73 K for 1 h in He, whereas the other catalysts were pretreated at 673
K for 1 h in He.
a
TABLE 2: Number of Desorbed Pyridine Molecules in the
Low- and High-Temperature Peaks in the TPD Spectra for
the Re/r-Fe
2 3
O Catalysts with Different Re Loadings
Re loading (wt %)
(
µmol/mg)
0.0
1.0
0.0148 0.0503 0.0575 0.1052 0.0919
0.0239 0.0566 0.1381 0.1408
2.0
3.0
6.0
10.0
peak R
peak â
0
0
0
31
using sulfuric acid as a catalyst. The selectivities of Re/R-
a
Fe2O3, Re/γ-Fe2O3, and Re/V2O5 were 90.5-93.7% at conver-
sions of 15.5-48.4% at 513 K, as shown in Table 1. The
selectivities of these catalysts are similar to that of SbRe2O3,
but their catalytic activities are much larger than the activity of
SbRe2O6. It is to be noted that R-Fe2O3 is inactive for methylal
synthesis and V2O5 is active only for formaldehyde formation.
Rhenium oxide species supported on the oxide supports showed
remarkable performances for catalytic methylal synthesis by one-
stage methanol conversion.
Number of Lewis acid sites.
peaks at 620-630 K, which is lower than the NH4ReO4 TPR
peak because of dispersion of NH4ReO4 on the support. For
the Re/R-Fe2O3 catalysts TPR peaks at 713 and 696 K developed
with increasing Re loading. The TPR peak features for the 1
wt % Re/R-Fe2O3 catalyst were different from those for the other
Re/R-Fe2O3 catalysts. Further increase of Re loading to 10 wt
%
6
%
gave a shoulder peak at 680 K in addition to the peaks at
96 and 713 K (Figure 10h). The TPR spectrum for the 10 wt
ReOx/γ-Fe2O3 catalyst exhibited two reduction peaks at 693
The plots of the reaction rate versus the surface area in Figure
1
, which correspond to the efficiencies per 1 wt % Re quantity
of the catalysts, indicate that the supported ReOx catalysts
employed in this study can be classified into four groups from
the viewpoint of the reactivity of the ReOx species. The Re/
TiO2 (rutile) catalyst shows a high activity even though the
and 742 K. γ-Fe2O3 had a reduction peak at 719 K. The
separation of the two peaks was more definite than for Fe/R-
Fe2O3, but the origins of the TPR peaks should be similar.
Figure 11 shows pyridine TPD spectra for Re/R-Fe2O3
catalysts with different Re loadings, ReO2, ReO3, and R-Fe2O3.
No significant pyridine adsorption occurred on R-Fe2O3, ReO2,
and ReO3. However, two desorption peaks at 417-426 and
4
1
in the two peaks are listed in Table 2. The amounts increased
with increasing Re loading up to 6 wt % Re, where it was nearly
saturated. Two distinct desorption peaks at 450 and 500 K were
also observed for Re/γ-Fe2O3, whereas the desorption of
pyridine from γ-Fe2O3 alone was negligible. There was no
significant difference in the TPD features for the R-Fe2O3- and
γ-Fe2O3-supported samples.
The nature of the acidic sites was examined by FTIR
spectroscopy for adsorbed pyridine on 6.0 wt % ReOx/R-Fe2O3.
The FTIR spectrum exhibited weak peaks at 1450, 1494, and
2
-1
surface area is as small as 5 m g , whereas the Re/SiO2
2
-1
catalyst, which has a relatively high surface area of 36 m g ,
shows a low activity. The Re/TiO2 (anatase) catalyst with the
largest surface area among the employed catalysts shows the
highest activity per gram of Re, but the reactivity of the ReOx
species on TiO2 (anatase) is similar to the reactivities of the
Re/R-Al2O3 and Re/MoO3 catalysts, which have intermediate
selectivity values. The Re/R-Fe2O3, Re/γ-Fe2O3, Re/V2O5, and
Re/ZrO2 catalysts are grouped according to the similar reac-
tivities per unit surface area of the rhenium oxide species. This
group shows high selectivities for the catalytic methylal
synthesis. The classification can be justified by the plots of
selectivity versus conversion per unit surface area in Figure 1.
The selectivity increased with increasing conversion, passed
through a maximum, and then gradually decreased with further
increasing conversion. The results demonstrate that the medium
activity (conversion per unit surface area) leads to the high
selectivity often observed in many catalytic processes. The Re/
R-Fe2O3, Re/γ-Fe2O3, Re/V2O5, and Re/ZrO2 catalysts exhibited
good performances for the catalytic selective oxidation of
70-479 K were observed for the Re/R-Fe2O3 catalysts with
.0-10.0 wt % Re loadings. The amounts of pyridine desorbed
-1
1
588 cm , which are assigned to pyridine coordinated to Lewis
acidic sites. There was no peak due to pyridine adsorbed on
Br o¨ nsted acidic sites. Under identical conditions, almost no
desorption peak was observed with the R-Fe2O3 support alone.

Rhenium Oxide Catalysts for Methanol Oxidation
J. Phys. Chem. B, Vol. 106, No. 17, 2002 4447
methanol to methylal: 90.5% selectivity at 15.5% conversion,
9
2
1.0% selectivity at 48.4% conversion, 93.7% selectivity at
1.5% conversion, and 89.4% selectivity at 35.8% conversion,
respectively, as shown in Table 1.
The conversion increased with reaction temperature, whereas
the selectivity to methylal passed through a maximum at 513
K andt hen decreased with increasing temperature, as shown in
Figure 3. However, the selectivity to methylal (product) +
formaldehyde (intermediate) maintained high values of 92.5%
at 15.5% conversion (513 K), 92.5% at 21.5% conversion (533
K), and 90.0% at 28.8% conversion (553 K). The fact that the
production of undesirable products such as dimethyl ether, CO,
and CO2 does not increase significantly with reaction temper-
ature also provides evidence that the supported rhenium oxides
are promising catalysts for selective methanol conversion to
methylal. Hereinafter, the surface states of Re/R-Fe2O3 and Re/
γ-Fe2O3 as the representative catalysts are discussed in relation
to the active Re species for catalytic methylal synthesis.
Figure 12. Surface models for the Re/R-Fe
2 3
O catalysts with Re
loadings of 0.7, 2.0, 3.0, and 10 wt %. Small circles, ReO monomers;
4
large circles, ReO
x
2
clusters (probably Re-Fe-O phase); squares, ReO .
are intermediate between those for bulk Re2O7 and bulk ReO3.
However, the 0.5-1 wt % Re/R-Fe2O3 catalysts with rhenium
oxides in the monolayer show 48.2 and 45.7 eV for Re 4f5/2
and 4f7/2, respectively. The values are almost the same as the
values (48.1 and 45.7 eV) obtained for NH4ReO4 with tetra-
4
.2. Surface States of Fe2O3-Supported Re Catalysts. A
monolayer coverage of ReOx species on R-Fe2O3 surface is
calculated to be about 0.7 wt % Re by assuming the cross section
-
19
2
of a ReO3 unit (selective bulk) to be 1.39 × 10
m . In the
plots of the Re 4f and Fe 3p XPS peak intensities versus Re
loading in Figure 8 there were definite breaks at 0.7-0.8 wt %
Re loading, followed by gently sloping straight lines, although
the data were somewhat scattered. The break feature in the XPS
plots of Figure 8 can be explained by the Stranski-Ktastanov
7
+
hedral Re dispersed on R-Fe2O3 (Figure 6). The XAFS study
on a monolayer Re/γ-Fe2O3 catalyst has revealed that the
rhenium oxides on γ-Fe2O3 were dispersed as [ReO4] species
37
with a tetrahedral symmetry. The preliminary EXAFS analysis
for the Re/γ-Fe2O3 catalysts also revealed the presence of Re-
O-Fe bonding at high Re loadings above a monolayer.37 Thus,
it is suggested that the Re species are stabilized by making Re-
O-Fe bonds. We assume the growth of a 3D Re-Fe-O mixed-
oxide phase following [ReO4] monolayer formation. The active
3
5
growth mode, in which ReOx species are dispersed up to a
monolayer coverage at the break and then three-dimensional
(3D) rhenium oxide species are formed on the ReOx monolayer.
Both the catalytic conversion and selectivity increased linearly
with increasing Re loading in the range 0-1 wt % and further
increased up to 3 wt % Re loading, which is about 4 times larger
than monolayer coverage (Figure 2). Thus, the monolayer ReOx
species are active for the selective methanol conversion to
methylal, but the 3D rhenium oxide species on the monolayer
species are more active than the monolayer species. Above 3
wt % Re, neither conversion nor selectivity continued to
increase, as shown in Figure 2. The active rhenium oxides at
the catalyst surface were saturated at 3 wt %. On the other hand,
the XPS Re 4f signals further increased with an increase in Re
loading above 3 wt %. The escape depth of the photoelectrons
for Re 4f levels in the XPS spectra is 1.5 nm, which means
that the XPS signals provide information on rhenium oxides in
not only the catalytically active outermost surface layer but also
several surface layers. The XPS spectra in Figure 7 reveal that
3
D rhenium oxides are saturated at four-layer coverage on
average, indicating that the rhenium oxide clusters are small.
The number of rhenium oxide clusters increases with increasing
Re loading in the range 0.7-3.0 wt %. As a consequence, the
surface states illustrated in Figure 12 are proposed for the Re/
R-Fe2O3 catalysts.
In the pyridine TPD spectra of the 1.0 wt % Re/R-Fe2O3
catalyst with a Re loading of about 1.4 monolayers in Figure
11, only one peak at 417 K is observed. According to the above
discussion, the peak is assigned to mainly the tetrahedral
7+
[Re O4] species dispersed in the monolayer on the R-Fe2O3
7+
surface. Hence, the acidity of the dispersed Re species is weak.
At Re loadings larger than 1.0 wt %, there are two peaks at
4
%
17-426 and 470-479 K (Figure 12 and Table 2). The 2 wt
Re/R-Fe2O3 catalyst shows two TPD peaks at 420 and 470
4+
Re species appear at 2-3 wt % Re loadings, but they are
K. The TPD data also indicate the formation of different Re
species from the tetrahedral [ReO4] species in the samples with
larger-than-monolayer Re loadings. The intensities of both low-
and high-temperature TPD peaks increased with increasing Re
loading from 3 to 6 wt % and nearly saturated or decreased
slightly at 10 wt % (Table 2). The conversion and selectivity
for catalytic methylal synthesis were saturated at 3 wt % Re
loading, as shown in Figure 2. There is no relationship between
the results of Figures 2 and 11, and hence, the acidity of the
surface is not a decisive factor in the catalytic performance.
The TPR spectra also increased in intensity with increasing
Re loading (Figure 10). The absolute quantities of hydrogen
molecules consumed in the TPR spectra in Figures 10 and 11
could not be determined exactly under the present experimental
conditions given certain experimental errors. However, we can
discuss the relative change in different Re loadings. The TPR
data reflect the redox properties of at least a few surface layers.
The dependency of the catalytic performance on the Re loading
6
-7+
less important than the Re
species in the catalytic perfor-
4
+
mance, as no Re species are observed for the 10 wt % Re/
γ-Fe2O3 catalyst with 91.0% selectivity (Figure 6). Above 3 wt
%
Re loading, the conversion and selectivity decreased slightly
4
+
(Figure 2). The decrease might be due to the increase of Re
species in the Re/R-Fe2O3 catalysts with 6-10 wt % Re, as
shown in Figure 7. ReO2 crystallites were observed for the 10
wt % Re/R-Fe2O3 catalyst (Figure 5c). ReO2 crystallites are also
active for methylal synthesis, but their selectivity is much
smaller than that for ReO3 (Table 1).
To clarify the oxidation state of the active rhenium oxide
species, we measured XPS spectra for the Re/R-Fe2O3 and Re/
γ-Fe2O3 catalysts (Figures 6 and 7). The binding energies of
4
3
1
7.8 eV (Re 4f5/2) and 45.4 eV (Re 4f7/2) observed with the
-10 wt % Re/R-Fe2O3 catalysts in Figure 7 were lower by
7+
.0-1.4 eV than those for Re2O7 (Re ) but higher by 1.0-1.2
6+ 33,34
eV than those for ReO3 (Re ).
The values for the catalysts

4448 J. Phys. Chem. B, Vol. 106, No. 17, 2002
Yuan and Iwasawa
+
(
(
Figure 2) is different from that of the TPR peak intensity
Figure 10), which indicates that the catalytic performance can
(6) On γ-Fe2O3, no Fe4 oxides were observed for the fresh
10 wt % Re catalyst.
be ascribed to the redox property of the outermost surface layer
of rhenium oxide clusters with suitable acidities. In fact, all of
(7) The ReOx species supported on the Fe2O3 surfaces showed
a TPR profile different from those for NH4ReO4, ReO3, and
ReO2.
6
-7+
the Re
species were not reduced by methanol exposure at
5
1
9
13 K, as shown in Figure 9. In the pulse experiments on the
0.0 wt % Re/a-Fe2O3 catalyst at 513 K using a mixture of
6.0/4.0 (mol %) He/MeOH in the absence of O2 (Figure 4),
(8) Lattice oxygen atoms of the ReOx species were active
for the catalytic methylal synthesis, and the consumed oxygen
atoms are replenished by O2 under the reaction conditions used.
(9) ReO2, ReO3, R-Fe2O3, and γ-Fe2O3 showed negligible
pyridine adsorption, whereas the active ReOx/R-Fe2O3 and ReOx/
γ-Fe2O3 catalysts showed two pyridine TPD peaks at 420-
450 and 470-500 K.
methylal formation decreased with an increase in the number
of the pulses, and after the fifth pulse, no methylal was produced.
It is evident that the lattice oxygen atoms of the rhenium oxide
clusters work as active oxygen species for the selective oxidation
of methanol. After 10 methanol pulses at 513 K, the intensities
of the XPS peaks at 47.8 and 45.4 eV decreased, whereas the
6-7+
(10) The redox capability of rhenium oxides between Re
4
+
and Re on the supports might be responsible for the selective
oxidation of methanol to formaldehyde, and the appropriate
acidity of rhenium oxides might also be necessary for the
acetalization of formaldehyde with methanol to form methylal.
4+
Re peak intensities increased (Figure 9b). The consumed
lattice oxygen atoms were replenished by gaseous O2, as verified
by the XPS spectrum of Figure 9c. The results indicate that the
redox capability of the rhenium oxide species, Re6
-7+
/ Re ,
is responsible for the selective oxidation of methanol.
4+
Acknowledgment. This work was supported by Core
Research for Evolutional Science and Technology (CREST) of
Japan Science and Technology Corporation (JST) and partially
supported by the Natural Science Foundation of China (NSFC).
In methanol conversion to methylal, the first reaction step
has been demonstrated to be the formation of methoxy spe-
cies.
38,39
Further transformation of the methoxy species depends
on the redox and acid-base properties of the active rhenium
oxides. Desorption of formaldehyde produced by dehydroge-
nation of the methoxy species is favored on weak acid sites. If
the active sites are strongly acidic, the residence time of the
formaldehyde becomes long enough for dioxymethylene species
References and Notes
(1) Bastide, B.; Enjalbert, R.; Fuess, H.; Galy, J. Solid State Chem.
000, 2, 545.
2
(
2) Watanebe, H.; Imoto, H.; Tanaka, H. J. Solid State Chem. 1998,
1
38, 245.
3
9
to be formed, and the dioxymethylene species can then react
(3) Butz, A.; Miehe, G.; Paulus, H.; Strauss, P.; Fuess, H. J. Solid
with neighboring methoxy species and/or adsorbed methanol
State Chem. 1998, 138, 232.
(4) Liu, H.; Gaigneaux, E. C.; Imoto, H.; Shido, T.; Iwasawa, Y. J.
Phys. Chem. 2000, 104, 2033.
to form methylal.40 If both athe cidic and basic sites are too
strong, the dioxymethylene species are oxidized to formates,
which also react with methanol to form methyl formate or are
(5) Liu, H.; Shido, T.; Iwasawa, Y. Chem. Commun. 2000, 1881.
(6) Moulijn, J. A.; Mol, J. C. J. Mol. Catal. 1988, 46, 1. Mol, J. C.;
Moulijn, J. A. In AdVance in Catalysis; Eley, D. D., Pines, H., Weisz, P.
B., Eds.; Academic Press: New York, 1975; Vol. 24, p 131.
39
further oxidized to carbon dioxides. If strong acid sites and
very weak basic sites are present at catalyst surfaces, only
dimethyl ether is formed.41 The present results indicate that the
Re/Fe2O3 catalysts can fit the requirements for both “redox”
and “acid-base” catalyses for selective methanol oxidation to
methylal. The bifunctional catalysis of moderate redox and
(
(
7) Olsthoorn, A. A.; Boelhouwer, C. J. Catal. 1976, 44, 207.
8) Augustine, S. M.; Sachtler, W. M. H. J. Catal. 1989, 116, 184.
(9) Pecoraro, T. A.; Chianelli, R. R. J. Catal. 1981, 67, 430.
10) Thomas, R.; Van Oers, E. M.; De Beer, V. H. J.; Medema, J.;
(
Moulijn, J. A. J. Catal. 1982, 76, 241.
11) Albonetti, S.; Cavani, F.; Trifiro, F. Catal. ReV.-Sci. & Eng. 1996,
38, 413.
(12) Warwel, S.; R u¨ sch, M.; Sojka, M. J. Chem. Soc., Chem. Commun.
1991, 1578.
(
acidic properties at the Re/Fe2O3 catalyst surface might benefit
from the Re6
-7+
Ox clusters (probably a Re-Fe-O phase) grown
7+
following the formation of the [Re O4] monolayer chemically
bound to the Fe2O3 surface.
(13) Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidation of Organic
Compounds; Academic Press: New York, 1981.
(
14) Organic Syntheses by Oxidation with Metal Compounds; Mijis, W.
J., de Jonge, C. R. H. J., Eds.; Plenum Press: New York, 1986.
15) Jacobsen, E. N.; Marc o´ , I.; Mungall, W. S.; Schr o¨ der, G.; Sharpless,
K. B. J. Am. Chem. Soc. 1988, 110, 1968.
(16) Herrmann, W. A.; Eder, S. J.; Scherer, W. Angew. Chem., Int. Ed.
Engl. 1992, 31, 1345.
5
. Conclusion
1) R-Fe2O3-, γ-Fe2O3-, and V2O5-supported ReOx catalysts
(
(
exhibited high activities and selectivities for the catalytic
selective oxidation of methanol to methylal, with 90-94%
selectivities at 15-49% conversions at 513 K.
(
(
(
17) Herrmann, W. A. Angew. Chem., Int. Ed. Engl. 1988, 27, 1297.
18) Sundermeyer, J. Angew. Chem. 1993, 105, 1195.
19) Herrman, W. A.; Fischer, R. W.; Rauch, M. U.; Scherer, W. J.
Mol. Catal. 1994, 86, 243.
(2) The highest reaction rate, in units of moles per hour per
(
20) Espenson, J. H. Chem. Commun. 1999, 479.
mole of Re, on the Re/R-Fe2O3 catalysts was achieved in the
range of Re loadings of 1-3 wt %.
(21) Owens, G. S.; Arias, J.; Abu-Omar, M. M. Catal. Today 2000, 55,
317.
(22) Yuan, Y.; Liu, H.; Imoto, H.; Shido, T.; Iwasawa, Y. Chem. Lett.
(3) Interactions between the rhenium oxides and the Fe2O3
2
000, 674.
(
supports (probably Re-O-Fe bonding) generated by the
treatment of supported NH4ReO4 precursor at 673 K in He
caused stabilization of dispersed tetrahedral Re O4 species
23) Yuan, Y.; Liu, H.; Imoto, H.; Shido, T.; Iwasawa, Y. J. Catal. 2000,
195, 51.
(24) Tatibou e¨ t, J. M. Appl. Catal. A: Gen. 1997, 148, 213.
(25) Tatibou e¨ t, J. M.; Germain, J. E. Bull. Soc. Chim. Fr. 1980, 9-10,
43.
26) Fournier, M.; Aouissi, A.; Rocchiccioli-Deltchff, C. J. Chem. Soc.,
Chem. Commun. 1994, 307.
27) Rocchiccioli-Deltcheff, C.; Aoussi, A.; Launary, S.; Fournier, M.
7+
7+
(e0.7 wt % Re) on the R-Fe2O3 surface and prevented the Re
3
4+
species from being reduced to Re O2.
4) Active Re6-7+Ox clusters (probably Re-Fe-O phase)
(
(
(
were formed on R-Fe2O3 at Re loadings g1 wt % after the
formation of a [Re O4] monolayer.
J. Mol. Catal. A: Chem. 1996, 114, 331.
7
+
(28) Shannon, I. J.; Maschmeyer, T.; Oldroyd, R. D.; Sankar, G.;
Thomas, J. M.; Pernot, H.; Baalikdjian, J. P.; Che, M. J. Chem. Soc.,
Faraday Trans. 1998, 94, 1495.
4+
(
5) On R-Fe2O3, Re O2 species were formed at Re loadings
of g 3 wt %, and ReO2 crystallites were also formed at a Re
(29) Tatibou e¨ t, J. M.; Germain, J. E.; Volta, J. C. J. Catal. 1983, 82,
loading of 10 wt %.
240.

Rhenium Oxide Catalysts for Methanol Oxidation
J. Phys. Chem. B, Vol. 106, No. 17, 2002 4449
(
30) For example, see: Otsuka, K.; Yamanaka, I. Appl. Catal. 1986,
(36) Basiska, A.; Klimkiewicz, R.; Domka, F. Appl. Catal. A: Gen.
2001, 207, 287.
2
6, 401.
(
(
31) Yuan, Y.; Shido, T.; Iwasawa, Y. Chem. Commun. 2000, 1421.
32) Abakumov, A. M.; Shpanchenko, R. V.; Antipov, E. V. Z. Anorg.
(37) Nakagawa, F.; Sasaki, T.; Shido, T.; Iwasawa, Y., to be published.
(38) Chung, J. S.; Miranda, R.; Bennertt, C. O. J. Catal. 1988, 114,
Allg. Chem. 1998, 624, 750.
33) Shpiro, E. S.; Avaev, V. I.; Antoshin, G. V.; Ryashentseva, M. A.;
Minachev, K. M. J. Catal. 1978, 44, 4022.
34) Cimino, A.; De Angelis, B. A.; Gazzoli, D.; Valigi, M. Z. Anorg.
Allg. Chem. 1980, 460, 86.
35) Shido, T.; Okita, G.; Asakura, K.; Iwasawa, Y. J. Phys. Chem. 2000,
04, 12263.
(
354.
(
39) Busca, G. Catal. Today 1996, 27, 457.
40) Yang, T. J.; Lunsford, J. H. J. Catal. 1987, 103, 55.
(
(
(
(41) Chauvin, C.; Saussey, J.; Lavalley, J. C.; Idriss, H.; Hindermann,
J. P.; Kienneman, A.; Chaumette, P.; Courty, P. J. Catal. 1990, 121, 56.
1