8
78
Chemistry Letters Vol.35, No.8 (2006)
Ti–Si Mixed Oxides by Non-hydrolytic Sol–Gel Synthesis as Potential Gold Catalyst Supports
for Gas-phase Epoxidation of Propylene in H and O
2
2
ꢀ
Maohua Dai, Dingliang Tang, Zhijie Lin, Hongwei Yang, and Youzhu Yuan
State Key Laboratory of Physical Chemistry of Solid Surfaces, Department of Chemistry,
College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, P. R. China
(Received May 11, 2006; CL-060562; E-mail: yzyuan@xmu.edu.cn)
The amorphous Ti–Si mixed oxides prepared from chloride
solution of HAuCl4 (4 wt % Au loading based on the support)
ꢃ1
precursors using a one-step nonhydrolytic sol–gel route could
serve as efficient gold catalyst supports for gas-phase epoxida-
tion of propylene in O2 and H2. A propylene conversion of
and a pH of 7:5 ꢁ 0:1 adjusted by 0.05 molꢂL aqueous solution
of NaOH. The solids were collected by filtration followed by
washing for three times. The final samples were calcined in air
at 673 K for 4 h. Catalytic reactions were carried out at 393 K
in a fixed-bed flow reactor operated at atmospheric pressure with
a feed containing 10 vol % each of C3H6, H2 and O2 in nitrogen
6
.8% at the initial 60 min and 4.0% after 4 h of time-on-steam,
keeping a selectivity to propylene oxide as high as 95%, was
obtained over the gold supported on a Ti–Si mixed oxide
containing 10 mol % of Ti.
3
ꢃ1 ꢃ1
4
(space velocity 4000 cm ꢂh ꢂg cat.). The products were
analyzed by two on-line gas chromatographs equipped with
TCD detectors (5A, 3 m and Porapak Q, 3m) and FID detector
(Porapak T, 2 m).
Nonhydrolytic sol–gel process which is based on the forma-
0
tion of M–O–M or M–O–M bridges by thermal condensation
The XRD measurements clarified that the amorphous Ti–Si–
N oxides were formed when x ꢄ 12. They possessed mesopores
between metal chlorides and metal alkoxides at moderate tem-
perature, dividing as halide–ether route and halide–alkoxide
route, may offer an alternative way to prepare mixed oxides with
x
and higher surface areas by the results of N -adsorption/desorp-
2
tion (Table 1). No Au crystalline diffraction lines were observed
with the samples of Au/Ti–Si–N (Figure 1a). High-resolution
TEM images (not shown) revealed that the size of Au nanopar-
1
good homogeneity and high surface area. The Ti–Si xerogels by
the nonhydrolytic sol–gel have been reported as promising cata-
lysts for epoxidation of alkenes and might be potential for the
x
ticles was about 2–4 nm. The UV–vis spectra (Figure 1b) showed
that the samples with Ti contents up to 12.0 mol % exhibited
absorbance peaks at 220–245 nm, which might result from tetra-
hedrally and octahedrally coordinated Ti centers. The broad
band and that at longer wavelength region than 250 nm suggest-
ed that the presence of polymeric octahedral coordination was
possible and also might be due to the Ti in a distorted tetrahedral
environment, although no peak at ca. 330 nm was observed
2
preparation of supported catalysts. On the other hand, much at-
tention has been paid on the epoxidation of propylene using O2
with gold nanoparticles deposited on Ti-containing supports
3
since the pioneer work of Haruta and co-workers. The propyl-
ene oxide (PO) selectivity is larger than 90% over the catalysts,
but it is necessary to remove several hurdles such as the lower
conversion of propylene, rapid catalyst deactivation, and regen-
eration problems before making the process commercially via-
ble. Three factors are believed as the key issues for overcoming
these barriers: a molecular scale dispersion of the Ti atoms
which are in synergism of the reduced dimensions of the metal
particles (2–4 nm), supports with mesopores for effective disper-
sion of Au nanoparticles inside the pores, and support hydropho-
bicity for better PO desorption. Indeed, selected supports such as
TS-1, Ti-MCM-41, Ti-MCM-48, Ti-containing hydrophobic sil-
sesquioxanes, and 3D mesoporous titanosilicates have been
claimed as being adequate for preparing active supported Au cat-
4
for the segregated TiO phase when x < 14. In addition, the
2
typical plasmon band at 500–600 nm for small Au particles at
the catalyst surfaces is evident as shown in Figure 1b.3
Table 1 shows the catalytic performance of Au deposited on
different supports for the gas-phase epoxidation of propylene. As
b,6
for Au/Ti–Si–N , the propylene conversion and PO selectivity
x
4
alysts. Herein, we present the first supported Au catalysts with
the Ti–Si mixed oxides prepared by a one-step nonhydrolytic
sol–gel synthesis for the gas-phase epoxidation of propylene in
O2 and H2.
We chose the halide–ether route to prepare Ti–Si mixed
1
c
oxides according to the procedure in the literature. The gels
were formed in a well-sealed autoclave at 393 K for 96 h without
contacting air. Calcination of the xerogels was performed in a
furnace under flowing dry air for 5 h at 773 K with a heating rate
of 1 K/min. The samples were labeled as Ti–Si–Nx (where x
stands for the Ti mol % in the initial gel). The preparation hy-
drolysis sol–gel was based on the reported procedure using TiCl4
Figure 1. (a) XRD patterns and (b) diffuse-reflectance UV–vis
spectra of Au/Ti–Si–Nx catalysts (the short dashs in UV–vis
stands for the supports of Ti–Si–Nx). Ti contents: 1) 6.0
mol %; 2) 8.0 mol %; 3) 10.0 mol %; 4) 12.0 mol %; 5) 14.0
mol %; 6) 20.0 mol %.
5
and SiCl4 as precursors and labeled as Ti–Si–Hx. The deposi-
tion of gold on all the supports was carried out by a deposi-
4
tion–precipitation (DP) method at 333 K, using an aqueous
Copyright Ó 2006 The Chemical Society of Japan
Dai, Maohua
