Preparation and Catalytic Application of Transition Metal (Fe, V, or Cu) Oxides Homogeneously Dispersed in the Wall of Mesoporous Nb2O 5

Article abstract of DOI:10.1246/cl.2003.1034

Fe, V, or Cu oxide was mixed in mesoporous Nb₂O₅, at a few atom percent, and their extremely high dispersion was confirmed by elemental analysis in about 5-nm spots using TEM-EDS. These mesoporous Nb ₂O₅ showed characteristic selectivity for cyclohexene oxidation depending on the kind of the doped metal oxide.

Full text of DOI:10.1246/cl.2003.1034

1
034
ChemistryLetters Vol.32, No.11 (2003)
Preparation and Catalytic Application of Transition Metal (Fe, V, or Cu) Oxides Homogeneously
Dispersed in the Wall of Mesoporous Nb O
2
5
y
ꢀy
Tomohiro Yamashita, Daling Lu, Junko N. Kondo, Michikazu Hara, and Kazunari Domen
Chemical Resources Laboratory, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8503
y
CREST (Core Research for Evolutional Science and Technology) of JST (Japan Science and Technology)
(Received July10, 2003; CL-030624)
Fe, V, or Cu oxide was mixed in mesoporous Nb2O5, at a
few atom percent, and their extremelyhigh dispersion was con-
firmed byelemental anal ys is in about 5-nm spots using TEM–
EDS. These mesoporous Nb2O5 showed characteristic selectiv-
ityfor cyclohexene oxidation depending on the kind of the doped
metal oxide.
Table 1. Physical properties of doped mesoporous Nb2O5 ma-
terials
a
b
Content
/mol % /m g
S
Pore Size Pore Volume d(100)
ꢁ1
Doped Element
2
ꢁ1
/nm
/mLg
/nm
Pure Nb2O5
—
1.0
3.0
5.0
3.0
3.0
202
4:1 ꢂ 0:5
5:7 ꢂ 0:8
5:0 ꢂ 0:7
4:9 ꢂ 0:6
4:0 ꢂ 0:3
5:0 ꢂ 0:8
0.29
0.37
0.43
0.37
0.29
0.38
6.4
6.3
6.6
6.3
6.4
7.1
Fe
Fe
Fe
V
208
254
253
200
233
Mesoporous silica and metal oxides are regarded as advan-
tageous matrix for supporting metal and metal oxides due to
their high surface area. Extremelyhigh dispersion is also ex-
Cu
1
a
BET surface area. b Errors were determined to involve 80% meso-
pected when the active homogeneous catalysts are anchored
byreactions with surface h yd roxy(OH) groups.
pores.
1
Although we
have synthesized various mesoporous transition metal oxides,
those of late transition metals, which are characteristic catalysts,
could not be prepared. Accordingly, we attempted to highly dis-
perse second transition metal oxide at a few atom percent in the
inorganic network of mesoporous Nb2O5 materials, which have
was aged at 313 K for 4–14 days until it was converted to dry
films, and the surfactant was removed bycalcination of the re-
sulting transparent samples at 723 K for 5 h. Mesoporous
Nb2O5 mixed with a transition metal oxide (Fe2O3, V2O5, or
CuO) is hereafter referred to as ‘‘M-doped mesoporous Nb2O5
(M = Fe, V, or Cu)’’ for brevity.
3
2
been recentlyprepared. Sintering of the incorporated transition
metal oxides are expected to be avoided at temperatures below
crystallization temperature of Nb2O5 (773 K), where mesopo-
rous structure is collapsed. Moreover, distinct catalytic behavior
depending on the incorporated metal oxides, which would not be
observed in the corresponding bulk materials, are expected be-
cause each transition metal atom is surrounded byNb 2O5. In this
studyFe, V, and Cu oxides were homogeneouslydispersed in
mesoporous Nb2O5, and their catalytic activities were evaluated
byc yc lohexene oxidation.
N2 adsorption isotherms of Fe-doped mesoporous Nb2O5
samples are compared in Figure 1. The amount of Fe2O3 was
roughlyadjusted to 1, 3, or 5 mol% of Nb 2O5. All the Fe-doped
Nb2O5 samples indicated the type IV isotherm pattern typical to
mesoporous materials, and the estimated BET surface areas and
pore volumes are summarized in Table 1 together with d(100)
values in their XRD patterns.
As observed in the shift of the P=P0 value corresponding to
capillarycondensation, pore size of the samples increased from
ca. 4.0 to 5.0 nm byFe-involvement. BET surface area and pore
volume also increased bymixing a small amount of Fe 2O3 in the
mesoporous Nb2O5. Because d(100) values were almost the
same as that of pure Nb2O5, the wall of mesoporous samples
is considered to become thinner than the pure Nb2O5. Because
3% Fe-doped mesoporous Nb2O5 sample possessed the largest
pore volume and BET surface area, its pore arrangement was di-
rectlyobserved byTEM. In Figure 2, TEM images of Fe(3%)-
doped mesoporous Nb2O5 are shown, which are viewed from
perpendicular and vertical to the direction of straight pore chan-
nels. Considering the wormhole-like pore structure of pure
mesoporous Nb2O5 (not shown), doping of a small amount of
4
A block co-polymer template (P85, PEO26PPO39, PEO26,
g) was dissolved in 1-propanol (10 g) at ambient temperature
1
followed bydispersion of a transition metal oxide (either Fe 2O3,
CuO, or V2O5, 0.07–0.35 mmol) bysonification for 5–15 min.
Then, 7 mmol NbCl5 was added with stirring for 5 min, and NaCl
aqueous solution (0.05 M, 1 mL) was subsequentlyinvolved.
The mixture was further stirred until the dispersed metal oxide
powders were completelydissolved accompanied byevolution
of HCl and the decrease in pH after NbCl5 addition. The sol
2
2
1
1
50
00
50
00
(
c)
(d)
b)
(
(a)
5
0
0
0
0.2
0.4
0.6
0.8
1.0
P / P0
Figure 1. N2 adsorption isotherms of Fe-doped mesoporous
Nb2O5, (a) 0 mol%, (b) 1 mol%, (c) 3 mol%, and (d) 5 mol%
Fe-doped mesoporous Nb2O5.
Figure 2. TEM images of Fe(3%)-doped mesoporous Nb2O5.
Copyright Ó 2003 The Chemical Societyof Japan

ChemistryLetters Vol.32, No.11 (2003)
1035
7
different element improved the ordering of the mesoporous
structure. The pore structure of the other Fe-doped mesoporous
Nb2O5 samples (1 and 5 mol %), were less ordered than those
observed in Figure 2.
those over pure CuO and CuO-loaded mesoporous Nb2O5.
The high selectivityto c yc lohexen-1-one was found to be
originated from CuO. Considerable selectivityto c yc lohexene
oxide was observed on pure CuO, while decrease of it was ac-
companied bythe formation of 1,2-c yc lohexanediol over CuO-
doped and CuO-loaded mesoporous Nb2O5 due to combination
of CuO with the acidic Nb2O5. However, only3 mol % involve-
ment of Cu species in Nb2O5 effectivelysuppressed the acid cat-
alyzed reaction over pure Nb2O5 owing to the high dispersion.
Although the selectivityto cyclohexene oxide was not improved,
the advantages of the presentlyprepared Cu-doped mesoporous
Nb2O5 among the three types of CuO catalysts are; (1) it im-
proved the original conversion more than twice, (2) nearly
90% selectivitywas attained for production of ketones, and (3)
the acid catalysis of Nb2O5 was suppressed byonly3 mol %
doping. Therefore, present strategy, preparation of a homogene-
ouslydispersed second transition metal element in mesoporous
metal oxide, is considered to be advantageous to the convention-
al loading method in regard with the preparation of unique cata-
lysts. Moreover, selections of the doped element and the original
mesoporous metal oxide would lead to new types of characteris-
tic catalysts.
V- and Cu-doped mesoporous Nb2O5 were also prepared at
mol % addition, and their properties are shown in Table 1. In
3
both cases, high BET surface area and pore volume were main-
tained. Pore size was not affected byV-addition, while Cu-doped
mesoporous Nb2O5 had a pore size distribution centered at about
5.2 nm similarlyto the case of Fe-doping. Pore arrangements of
the V- and Cu-doped mesoporous Nb2O5 samples were worm-
hole structure similarlyto the non-doped mesoporous Nb 2O5
(
see TEM image of Cu-doped sample in graphical abstract).
Next, the dispersion of the Fe and V species in 3 mol % dop-
ed Fe and V were evaluated in 5.0-nm diameter ranges bythe el-
5
emental analysis (EDS) using TEM apparatus. 2.2 and 3.4%
were given as averaged doping amount of Fe and V, while
ICP analysis indicated 2.7 and 3.8% for the same samples. Al-
though ICP measurements showed slightlylarger value, the high
dispersions of Fe and V in both samples were confirmed byEDS.
Therefore, homogeneous mixing of a different transition metal
oxide in Nb2O5 was successful bydissolving the metal oxide
in acidic solution of NbCl5, even after calcinations at 723 K.
Catalytic properties of the M-doped (M = Fe, V, and Cu)
mesoporous Nb2O5 samples were evaluated byoxidation of cy-
References and Notes
1
T. J. Barton, L. M. Bull, W. G. Klemperer, D. A. Loy, B.
McEnaney, M. Misono, P. A. Monson, G. Pez, G. W.
Scherer, J. C. Vartuli, and O. M. Yaghi, Chem. Mater., 11,
2633 (1999), and the references therein.
B. Lee, J. N. Kondo, D. Lu, and K. Domen, Chem. Lett.,
2002, 1058; B. Lee, J. N. Kondo, D. Lu, and K. Domen, J.
Am. Chem. Soc., 124, 11256 (2002).
X-raydiffraction (XRD) was measured on Rigaku RINT200
diffractometor with Cu Kꢀ radiation, and N2 adsorption iso-
therms were measured byCOULTER OMNISORP 100CX
and SA3100 systems. BET surface area was obtained at rel-
ative pressure (P=P0) in 0.005–0.2 range, and pore size dis-
tribution was obtained from the analysis of the adsorption
branch of the isotherms using the BJH (Barrett–Joyner–
Halenda) method. Transmission electron microscope
(TEM) images were obtained on a JEOL 2010F electron mi-
croscope operated at 200 keV. Inductivelycoupled plasma
(ICP) atomic emission analysis was performed on a Shima-
dzu ICPS-8100 after dissolving the sample byHF solution
6
clohexene using tert-butylhydroperoxide (TBHP). As listed in
Table 2, various oxidized products were obtained besides cyclo-
hexene oxide; 2-cyclohexen-1-one, 2-cyclohexen-1-ol, 1,2-cy-
clohexanediol, and 1,2-cyclohexanedione. 1,2-cyclohexanediol
was obtained as a main product in the case of mesoporous Nb2O5
due to the hydrolysis of the cyclohexene oxide produced by ox-
idation of cyclohexene. Both oxidation ability and acidity of
Nb2O5 resulted in the high selectivityto 1,2-cyclohexanediol be-
cause no such oxidation products were produced over Nafion, a
representative acid catalyst, under the same condition nor over a
non-acidic CuO (Table 1). The selectivityof 2-c yc lohexen-1-ol
decreased and 1,2-cyclohexanedione was produced over Fe-dop-
ed mesoporous Nb2O5. In the case of V-doped mesoporous
Nb2O5, the increases in conversion and selectivityto c yc lohex-
ene oxide were observed. Considering the slight increase in se-
lectivityto 1,2-c cy lohexanediol over V-doped mesoporous
Nb2O5, the oxidation to cyclohexene oxide is encouraged by
V-doping, followed byconsequent increase in formation of
2
3
ꢁ
1
1
,2-cyclohexanediol through the acid catalyzed hydrolysis by
Nb2O5.
Cu-doped mesoporous Nb2O5 showed evident effects of
to adjust to 5–10 mgꢃL
.
4
5
The ratio was determined bythe number of metal in metal
oxide.
Cu was not quantified owing to the use of Cu grid for TEM
observations. 10 particles were selected for each sample, and
the elemental analyses were performed on 60 and 64 spots
for Fe- and V-doped mesoporous Nb2O5 samples.
doping; more than twice of conversion of pure Nb2O5 was ob-
tained with a high selectivityto 2-c yx lohexen-1-one. The result
over Cu-doped mesoporous Nb2O5 should be compared with
Table 2. Oxidation of cyclohexene over doped mesoporous
Nb2O5 materials
6
6
7
0.20 g catalyst and 36 mmol cyclohexene were mixed in
1
5
mL acetonitrole, followed byaddition of 6.6 mmol tert-
Selectivity /%
butylhydroperoxide (TBHP) at 333 K. The products were an-
alyzed by a gas–chromatograph (Shimadzu, GC-18A) with a
capillarycolumn (J&W capillarycolumn, DB-FFAP).
Acetonitrile was used as an internal standard.
CuO-loading (3%) was performed bydispersing mesoporous
Nb2O5 in 0.01 M CuSO4 aqueous solution, followed by
calcination at 573 K in air. The amount of loaded CuO was
measured as 3.4% byICP anal ys is.
O
OH
Conversion
OH
OH
O
O
Catalyst
O
alcohols ketones
/%
Pure Nb2O5
Fe (3%)-doped
V (3%)-doped
Cu (3%)-doped
CuO
6.5
6.5
8.5
6.8
12.1 14.4 64.8
0.0
7.8
0.0
9.7
0.0
0.0
89.2
73.6
74.5
8.1
12.1
19.5
12.4
89.8
71.5
72.8
11.7
12.4
80.1
4.8 68.8
3.6 70.9
10.6
14.0
2.1
13.0
1.8
2.2
5.9
0.0
7.4
11.0
4.2
71.5 17.0
72.8 15.5
17.0
22.9
Cu (3%)-loaded
6.3
Published on the web (Advance View) October 13, 2003; DOI 10.1246/cl.2003.1034