Propene epoxidation with O2 and H2: Identification of the most active gold clusters

Article abstract of DOI:10.1016/j.jcat.2010.11.012

Gold was deposited on alkaline-treated TS-1 (TS-1-Na1) to prepare Au/TS-1-Na1(DP) by deposition-precipitation (DP) and Au/TS-1-Na1(SG) by solid grinding (SG). ¹²⁹Xe NMR technique has detected that tiny Au clusters have been incorporated into th

Full text of DOI:10.1016/j.jcat.2010.11.012

Journal of Catalysis 278 (2011) 8–15
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Propene epoxidation with O and H : Identification of the most active gold clusters
2
2
Jiahui Huang a,b, Enrique Lima , Tomoki Akita b,d, Ariel Guzmán , Caixia Qi , Takashi Takei a,b,
Masatake Haruta
a
c,
⇑
e
f
a,b,
⇑
Graduate School of Urban Environmental Sciences, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji, Tokyo 192-0397, Japan
Japan Science and Technology Agency (JST), CREST, 4-1-8 Hon-cho, Kawaguchi, Saitama 332-0012, Japan
Instituto de Investigaciones en Materiales, Universidad Nacional Autónoma de México, Circuito exterior s/n, Cd. Universitaria, Del. Coyoacán, CP 04510, México D.F., Mexico
Research Institute for Ubiquitous Energy Devices, National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan
Instituto Politécnico Nacional – ESIQIE, Avenida IPN UPALM Edificio 7, Zacatenco, 07738 México D.F., Mexico
b
c
d
e
f
Institute of Applied Catalysis, Yantai University, Yantai 264005, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 July 2010
Revised 11 November 2010
Accepted 11 November 2010
Available online 15 January 2011
Gold was deposited on alkaline-treated TS-1 (TS-1-Na1) to prepare Au/TS-1-Na1(DP) by deposition–pre-
129
cipitation (DP) and Au/TS-1-Na1(SG) by solid grinding (SG).
Xe NMR technique has detected that tiny
Au clusters have been incorporated into the microporous channels of Au/TS-1-Na1(DP), while they were
absent inside the microporous channels of Au/TS-1-Na1(SG). On the other hand, HAADF-STEM observa-
tion showed that the amount of Au clusters (1.0–2.0 nm) over the exterior surfaces was much larger in
Au/TS-1-Na1(SG) than in Au/TS-1-Na1(DP). In propene epoxidation with O
2
and H
2
, Au/TS-1-Na1(SG)
Keywords:
Propene epoxidation
O /H mixture
2 2
ꢀ1 ꢀ1
ꢀ1 ꢀ1
exhibited much higher PO formation rate (127 gPO kgcat
h ) than Au/TS-1-Na1(DP) (74 gPO kgcat h ),
indicating that Au clusters with diameters of 1.0–2.0 nm are more active for PO synthesis than tiny Au
clusters incorporated inside the microporous channels.
Alkaline-treated TS-1
Gold clusters
Ó 2010 Elsevier Inc. All rights reserved.
129
Xe NMR
1
. Introduction
as the only by-product). Since Hayashi et al. [4] first reported in
1
998 that propene epoxidation with O
by gold nanoparticles (Au NPs, 2.0–5.0 nm) deposited on anatase
TiO with PO selectivity above 90%, many efforts have been de-
voted to this reaction [5–19]. So far, great progresses have been
made, and C conversion (5.0–9.8% [5]), PO selectivity (90–96%
[5]) and H efficiency (30–47% [8]) have approached the estimated
2 2
and H could be catalyzed
Propene epoxide (PO) is a very important feed stock in chemical
industry and widely used to produce a variety of derivatives such
as polyether polyols and propylene glycol. In Industry, the chloro-
hydrin process and several organic peroxide processes are used to
produce PO [1]. However, these processes suffer from the environ-
mental unfriendliness and market mismatch of co-products,
respectively. The chlorohydrin process is accompanied by the by-
production of toxic chlorinated organic compounds as well as salts
such as CaCl . The organic peroxide processes usually require mul-
2
tiple reaction steps in liquid phase and often suffer from the mis-
match in market demand of co-products such as tert-butanol.
2
3 6
H
2
industrial targets (10%, 90% and 50%, respectively).
For Au catalysts, there are two key factors that define their cat-
alytic performance: the type of supports and the size of gold parti-
cles. For supports, usually materials containing Ti cations are
required, and until now, three materials have been widely used
as supports, anatase TiO
[5–7,9,10] and microporous titanosilicalite-1 (TS-1) [8,11,15–19].
Although Au/TiO could present high PO selectivity (>90%), C
2 2
[4,12], mesoporous Ti–silicate (Ti–SiO )
The latest PO production plants are based on H
2] or on cumene recycling by the reduction of co-product, cumyl
alcohol with H [3].
Direct epoxidation of propene (C
hydrogen (H ) is a promising process for PO synthesis, because it
is simple (one-step reaction in gas phase) and green (with H
2 2
O as an oxidant
[
2
3 6
H
2
conversion is usually very low, less than 1.0%, due to the low reac-
tion temperature (usually lower than 373 K). In contrast, Au/mes-
3 6 2
H ) with oxygen (O ) and
2
oporous Ti–SiO
temperature, and both high C
ity could be achieved. Promoted by Ba(NO
porous Ti–SiO displayed a high C conversion of 9.8% with a
2
and Au/TS-1 could be used at higher reaction
conversion and high PO selectiv-
, 0.30 wt% Au/meso-
2
O
3 6
H
3 2
)
2
3 6
H
⇑
Corresponding authors. Address: Graduate School of Urban Environmental
Sciences, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji, Tokyo 192-
high PO selectivity of about 90% at 433 K [5], while 0.05 wt% Au/
TS-1(Si/Ti = 36) without promoters could also present a high
0
397, Japan. Fax: +81 42 677 2852 (M. Haruta), Fax: +52 55 5616 1371 (E. Lima).
C H conversion of 8.8% with a PO selectivity of 81% at 473 K,
3 6
E-mail addresses: lima@iim.unam.mx (E. Lima), haruta-masatake@center.t-
mu.ac.jp (M. Haruta).
ꢀ1
ꢀ1
which corresponded to a PO formation rate of 116 gPO kg
h
cat
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.11.012

J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
9
[
16]. As for the catalytic stability, among these three types of Au
washed five times with 2000 mL (total amount) of H
2
O to remove
catalysts, Au/TS-1 is superior.
Beside supports, the size of gold particles is also critically
2
important. For example, for Au/TiO only hemispherical Au NPs
residual NaOH and then dried at 373 K in air. The obtained alka-
line-treated TS-1 was named as TS-1-Na1, where 1 stands for
1.0 h of alkaline treatment period.
with size of 2.0–5.0 nm are active for PO synthesis, whereas spher-
ical Au NPs larger than 5.0 nm favor the complete combustion of
2.2. Deposition of Au by DP
propene to CO
pene hydrogenation [4]. Although Au/mesoporous Ti-SiO
Au/TS-1 are more active for PO synthesis than Au/TiO , the exact
2
and Au clusters smaller than 2.0 nm lead to pro-
2
and
The procedure of Au deposition on TS-1 and TS-1-Na1 by the DP
method in aqueous solution at room temperature is as follows.
2
size or size range of Au particles most active for PO synthesis is
not clear yet. Lu et al. [9] have observed that in Au/mesoporous
4
First, 1.0 g of support was added to 50 mL of 9.1 mM HAuCl aque-
ous solution (Au/Support = 9.0 wt.%), and then the aqueous solu-
tion of NaOH was added dropwise until the pH reached 9.0. The
suspension was stirred vigorously at room temperature for 3.0 h,
and meanwhile the pH of 9.0 was kept constant by adding a small
amount of NaOH aqueous solution. Finally, the solid was centri-
fuged out, washed in 50 mL of water, centrifuged again and then
dried under vacuum at room temperature overnight. The Au cata-
lysts thus prepared were denoted as Au/TS-1(DP) and Au/TS-1-
Na1(DP).
Ti-SiO
much higher catalytic performance than both large Au clusters
2.0 nm from TEM or 1.4 nm from EXAFS) and Au NPs (2.5 nm from
TEM). Yap et al. [15] speculated that in Au/TS-1, Au clusters
<2.0 nm) which were hardly visible by TEM would be responsible
2
catalysts, small Au clusters (0.9 nm from EXAFS) showed
(
(
for PO synthesis. Joshi et al. [20,21] proposed based on theoretical
calculations that in Au/TS-1 tiny Au clusters incorporated in micro-
porous channels (ꢁ0.55 nm) of TS-1, such as Au
3
clusters, should
be active for PO synthesis.
Our recent results showed that Au/alkaline-treated TS-1 pre-
pared by solid grinding (SG) could display a high PO formation rate
2.3. Deposition of Au by SG
ꢀ1
ꢀ1
of 137 gPO kg
h
[8], comparable to the best results reported by
The preparation of Au/TS-1 and Au/TS-1-Na1 by the SG method
was done according to the procedure previously reported [8,29].
Since TS-1 contains 0.75 wt% K [8], before SG, TS-1 was treated in
cat
Cumaranatunge and Delgass [17] over Au/TS-1 prepared by depo-
sition–precipitation (DP). HAADF-STEM (high-angle annular dark-
field scanning transmission electron microscopy) observations,
which could provide stronger contrast for Au against the support,
showed that Au clusters with diameters of 1.0–2.0 nm were situ-
ated on the exterior surfaces of alkaline-treated TS-1. It is likely
that those Au clusters are responsible for the excellent catalytic
performance. On the other hand, it could not be denied that tiny
Au clusters incorporated inside microporous channels of alkaline-
treated TS-1 exhibited high catalytic activity for PO synthesis.
Accordingly, a key question is which is more active for PO synthe-
sis, Au clusters (1.0–2.0 nm) on the exterior surfaces or tiny Au
clusters incorporated inside microporous channels.
4
a similar manner to the DP method by replacing HAuCl with HCl
to eliminate the potential effect caused by K element and then
dried at 373 K in air overnight. First, the powder of TS-1 or TS-1-
Na1 and the required amount of dimethyl Au(III) acetylacetonate
3 2
[(CH ) Au(acac)] having a vapor pressure of 1.1 Pa at 298 K were
ground in an agate mortar in air for 15 min at room temperature
and then reduced by 10 vol.% H
1.0 h. The temperature was raised from room temperature to
423 K at a heating rate of 1.0 K min . The Au catalysts thus pre-
pared were denoted as Au/TS-1(SG) and Au/TS-1-Na1(SG).
ꢀ1
2
in Ar (20 mL min ) at 423 K for
ꢀ
1
In this context, xenon is a monoatomic probe to sense free
spaces such as pores. 1 Xe is chemically inert, small with a diam-
eter of 0.44 nm and very sensitive to the chemical surroundings of
microporous channels of zeolites [22]. Therefore, 1 Xe NMR spec-
troscopy is a powerful technique to detect metal clusters incorpo-
rated into the channels of zeolites [23]. Previous research showed
that once metal clusters such as Ir, Pt, Rh, Os and Au were incorpo-
2.4. Characterization of catalysts
29
HAADF-STEM was utilized to observe gold particles on a JEOL
JEM-3000F transmission electron microscope equipped with a dig-
itally processed STEM imaging system. The operating voltage was
300 kV, and the resolution was about 0.20 nm. Here, not fresh
29
2 2
but used gold catalysts in propene epoxidation with O and H
1
29
rated into microporous channels of zeolites, much greater
Xe
mixture were selected and directly dispersed on a micro-grid sup-
ported on a copper mesh without solvent. Elemental compositions
were analyzed by inductively coupled plasma-atomic emission
spectroscopy (ICP-AES).
chemical shifts at low Xe loading could be detected than those
on pure zeolites due to the strong interaction of Xe atoms with me-
tal clusters [24–28]. Herein, 1 Xe NMR spectroscopy was utilized
to characterize Au/alkaline-treated TS-1 prepared by SG and DP
to examine whether tiny clusters could be incorporated into the
microporous channels or not. For comparison, Au/TS-1 prepared
by DP and supports (TS-1 and alkaline-treated TS-1) were also
29
Xenon gas (Praxair, 99.999%) was used for the 129Xe NMR mea-
surements. Two sets of samples were selected. One set of samples
consist of TS-1 and 0.11 wt% Au/TS-1(DP) and the former was also
treated in a similar manner to the DP method by replacing HAuCl
with HCl to eliminate the potential effect caused by different Na
content. The other set of samples consist of TS-1-Na1, 0.20 wt%
4
1
29
characterized by
Xe NMR spectroscopy.
129
Au/TS-1-Na1(SG) and 0.23 wt% Au/TS-1-Na1(DP). Before
NMR analysis, Au catalysts were used to catalyze propene epoxida-
tion with O and H mixture at 473 K for 2.0 h according to the pro-
cedures described in the following Part 2.5. Catalytic tests. During
Xe
2
. Experimental
2
2
2.1. Alkaline treatment of TS-1 with NaOH aqueous solution
1
29
Xe NMR measurements, supports or Au catalysts were first
Microporous TS-1 with molar ratio of silicon (Si) to titanium (Ti)
placed in an NMR tube equipped with J. Young valves, where the
samples were dehydrated by slowly raising temperature to 473 K
of 48 was synthesized according to the previous report [8]. The
alkaline treatment of TS-1 was as follows. First, 1.0 g of calcined
TS-1 was added to the water (100 mL). Second, to the suspension,
ꢀ4
in a vacuum (1.33 ꢂ 10 kPa) and keeping the temperature at
473 K for 2.0 h. Then, dehydrated samples were contacted with dif-
1
29
1
.0 M NaOH aqueous solution was added dropwise under vigorous
ferent pressures of xenon at 291 K. Finally,
Xe NMR spectra were
stirring at 303 K until the pH reached 12. The pH was maintained
at this value for 1.0 h by continuing to add a small amount of NaOH
aqueous solution. Finally, the solid was collected by filtration,
recorded between 210 and 291 K in a Bruker DMX-500 spectrom-
eter operating at 138.34 MHz. Single excitation pulses were used,
and at least 1000 scans were collected with a delay time of 2 s.

1
0
J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
The chemical shift was referenced to xenon gas extrapolated to
gated (Fig. 1). All the catalysts displayed good catalytic stability at
reaction temperature of 473 K, especially C conversion and PO
1
29
zero pressure. The
Xe NMR experiments can be repeated very
3 6
H
well. The peaks were sufficiently narrow, and the error associated
with chemical shift value is roughly 0.10 ppm.
selectivity. These results are consistent with previous reports
[8,16]. Interestingly, the steady state over 0.10 wt% Au/TS-1-
Na1(SG) and 0.20 wt% Au/TS-1-Na1(SG) could be rapidly reached
when the temperature was elevated to 473 K. However, that over
2.5. Catalytic tests
0
.23 wt% Au/TS-1-Na1(DP) could be reached only after time on
stream of 2.0 h at 473 K due to the necessity of in situ activation
of Au species [17].
Gold catalysts prepared by SG were first reduced at 423 K by H
10 vol.% H in Ar) and then cooled down to room temperature in
2
2
(
It is interesting that the increase of Au loading from 0.10 wt% to
the same stream. Gold catalysts prepared by DP were first dried
under vacuum at room temperature overnight and then used di-
rectly without treatment. In both cases, 0.15 g of gold catalyst
was weighed and then loaded into a vertical fixed-bed U-shaped
quartz reactor with an inner diameter of 10 mm. The reactant feed
0
.20 wt% in Au/TS-1-Na1(SG) can only lead to a relatively small ef-
fect on the PO formation rate per gram of catalyst. This result
might be explained as follows. A key step in propylene epoxidation
with O
7,10]. Therefore, the formation of H
effective transfer to isolated Ti sites are very important. Higher
density of Au clusters will produce larger amount of H but cov-
ers larger amount of isolated Ti sites. Thus, the amount of effective
(free, uncovered by Au clusters) isolated Ti sites in capturing H
will be reduced. In addition, higher density of Au clusters will en-
hance the direct decomposition of H over the Au surfaces to
O. Finally, the increase of Au loading can only lead to a relatively
2
and H
2
is assumed to be the formation of Ti–OOH species
[
2 2
O
on the Au surfaces and its
gas of C
3 6 2 2
H , O , H and Ar with volume ratio of 10/10/10/70 was
ꢀ1
passed through at a flow rate of 20 mL min , which corresponded
to a space velocity of 8000 mL g
ꢀ
1
ꢀ1
2 2
O
h
. Reaction temperature was
cat
raised slowly from room temperature to 473 K with a heating rate
ꢀ1
2 2
O
of 1.0 K min and then kept at 473 K. Reactants and products were
analyzed by on-line GCs equipped with TCD (Porapak Q column)
and FID (HR-20M column) detectors.
2 2
O
H
2
small increase in the amount of Ti–OOH species and thus a rela-
tively small increase in PO formation rate per gram of catalyst. This
3
. Results
interpretation might be supported by reduced H
Au loadings in our previous research [8].
2
efficiency with
3.1. Catalytic activity
In our previous work, we found that Au loading could be easily
1
29
adjusted by SG [8]. Thus, herein we first prepared Au/TS-1 by the
DP method, and then after analyzing the real Au loading, we pre-
pared Au/TS-1 with the same Au loading by the SG method. As
shown in Table 1, Au/TS-1(DP) with a Au loading of 0.11 wt% was
3.2.
Xe NMR characterization
Fig. 2 and Fig. 3 show the effect of the amount of adsorbed xe-
non (Xe atoms per g of catalyst) on the
should be noted that in our experiments, all the
1
29
Xe NMR chemical shift. It
1
29
prepared by DP and gave only a low C
H
3 6
conversion of 1.4% with
Xe NMR spectra
a PO selectivity of 82% and with a H
2
efficiency of 31%, presenting a
showed only a single and symmetrical peak. Therefore, in Figs. 2
and 3, only one curve could be plotted for each sample. Fig. 2
ꢀ
1
ꢀ1
low PO formation rate of 22 gPO kg
h
. Then, similar Au content
cat
1
29
was loaded by SG on TS-1, which has been treated in a similar
manner to the DP method by replacing HAuCl with HCl to remove
K element and did not contain mesopores based on N adsorption
measurements, but the catalytic performance was inferior. In con-
trast, over 0.10 wt% Au/TS-1-Na1(SG), a greatly enhanced C
conversion of 7.4% was achieved with a PO selectivity of 85% and
with a H efficiency of 30%, corresponding to a high PO formation
rate of 119 gPO kg
shows that the
Xe NMR chemical shift over TS-1 increases al-
4
most linearly with Xe concentration, which is the common behav-
ior of xenon adsorbed in micropores. On the contrary, over
2
1
29
0.11 wt% Au/TS-1(DP), the
Xe NMR chemical shift first decreases
3
H
6
gradually and then increases linearly with Xe concentration. Actu-
ally, at high xenon loading, both curves are parallel. At the lowest
Xe loadings with Xe pressure of about 30 Torr, the chemical shift
over 0.11 wt% Au/TS-1(DP) is 110.5 ppm, 12.4 ppm higher than
that of 98.1 ppm over TS-1.
2
ꢀ
1
cat
ꢀ1
h
. In order to eliminate the effect of different
supports (TS-1 vs. TS-1-Na1) on the very different catalytic activity
over 0.11 wt% Au/TS-1(DP) and 0.10 wt% Au/TS-1-Na1(SG), the
same support (TS-1-Na1) was selected to deposit Au at about
Compared with TS-1 (Fig. 2), TS-1-Na1 displayed a very differ-
ent 1 Xe NMR chemical shift curve, which first decreases and then
increases almost linearly with Xe concentration (Fig. 3). At the low-
est Xe concentration with Xe pressure of about 30 Torr, a chemical
shift of 74.1 ppm was obtained over TS-1-Na1, much lower than
that of 98.1 ppm over TS-1. Over 0.23 wt% Au/TS-1-Na1(DP) and
0.20 wt% Au/TS-1-Na1(SG), similar chemical shift curves to that
over TS-1-Na1 were obtained (Fig. 3). At the lowest Xe loadings
29
0
.20 wt% by the SG and the DP method (Table 1). The PO formation
rate over 0.20 wt% Au/TS-1-Na1(SG) prepared by SG is
ꢀ
1
cat
ꢀ1
ꢀ1 ꢀ1
1
0
27 gPO kg
h
, much higher than that of 74 gPO kgcat
.23 wt% Au/TS-1-Na1(DP) prepared by DP.
The catalytic stability of 0.10 wt% Au/TS-1-Na1(SG), 0.20 wt%
h
over
Au/TS-1-Na1(SG) and 0.23 wt% Au/TS-1-Na1(DP) was also investi-
Table 1
Propene epoxidation with O
2 2
and H mixture over Au catalysts prepared by DP and SG.
Au catalystsa
Mean Au diameter (nm)
Na (wt.%)
K (wt.%)
C
H
conv. (%)
H
2
effici. (%)
PO sel. (%)
ꢀ1
ꢀ1
3
6
PO formation rate (gPO kg
h )
cat
0
0
0
0
0
.11 wt% Au/TS-1(DP)
.10 wt% Au/TS-1(SG)
.10 wt% Au/TS-1-Na1(SG)
.20 wt% Au/TS-1-Na1(SG)
.23 wt% Au/TS-1-Na1(DP)
4.3
4.0
1.6
1.8
1.7
0.95
0.95
0.93
0.93
0.96
<0.01
<0.01
<0.01
<0.01
<0.01
1.4
0.9
7.4
8.3
4.8
31
32
30
24
27
82
90
85
84
81
22
16
119
127
74
b
ꢀ
h ; temperature, 473 K. All the data were taken under a steady state after
cat
1
ꢀ1
Reaction conditions: feed gas, C
.0 h duration.
PO: propene epoxide; conv.: conversion; sel.: selectivity; effici.: efficiency.
3
H
6
/O
2
/H
2
/Ar = 10/10/10/70 in vol.%; space velocity, 8000 mL g
2
a
wt.% shows the real gold loadings obtained by ICP-AES analyses. Ti content is 1.8 wt% in all the catalysts.
Before solid grinding, TS-1 was treated in a similar manner to the DP method by replacing HAuCl with HCl.
4
b

J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
11
Fig. 1. Catalytic performance of 0.10 wt% Au/TS-1-Na1(SG) (}), 0.20 wt% Au/TS-1-Na1(SG) (N) and 0.23 wt% Au/TS-1-Na1(DP) (.) in C
3
H
6
epoxidation with an O
2
and H
ꢀ
h . When the reaction
cat
2
1
ꢀ1
mixture. Reaction conditions: catalyst, 0.15 g; temperature, 473 K; feed gas, C
temperature just reached 473 K, time on stream began.
H
3 6
/O
/H
2 2
/Ar = 10/10/10/70 in vol.%; space velocity, 8000 mL g
with Xe pressure of about 30 Torr, the chemical shift over 0.23 wt%
Au/TS-1-Na1(DP) is 86.6 ppm, 12.5 ppm higher than that of
from white (the color of TS-1) to slight pink (the color of Au parti-
cles). This color change over Au/TS-1(SG) during SG clearly indi-
7
0
4.1 ppm over TS-1-Na1; however, the chemical shift over
.20 wt% Au/TS-1-Na1(SG) is 77.6 ppm, only ꢁ3.5 ppm higher than
cates that (CH
of surface Ti-OH species and/or Si–OH species to form –O–
Au(CH [30]. The formed –O–Au(CH species adsorbed on Au/
3 2
) Au(acac) can easily decompose in air by virtue
that over TS-1-Na1.
3
)
2
3 2
)
TS-1(SG) will be reduced and coagulate to form Au particles during
solid grinding in air.
3.3. HAADF-STEM observations
Fig. 4a shows one of the typical HAADF-STEM pictures of
.11 wt% Au/TS-1(DP) taken in many different regions. In this pic-
4. Discussion
0
ture, only three Au particles can be observed with diameters of 1.8,
.7 and 7.5 nm. Counting all the Au particles revealed a very low
population density of 6.4 Au particles per 100 nm ꢂ 100 nm. In
addition, the diameter distribution of Au particles was also very
broad (1.0–10.5 nm), and only 18% of Au particles existed as clus-
ters (<2.0 nm), finally leading to a relative large mean diameter of
In order to obtain the high catalytic performance for PO synthe-
sis not TS-1 but TS-1 treated with alkaline solution is indispensable
as a support for Au clusters (Table 1). Second, the SG method is
superior to the DP method. The results of the characterizations
indicated that support materials (TS-1 and TS-1-Na1) in all the five
Au catalysts (shown in Table 1) possess very similar crystalline MFI
structure (based on XRD and Si MAS NMR analysis), similar coor-
dination circumstance of Ti sites (based on UV–vis and XANES) and
similar particle morphology (based on TEM). In addition, these Au
catalysts also possess similar Na content and the same Ti content
(Table 1). Therefore, the very different catalytic activity among
them should be caused by the different size and different size dis-
tribution of Au particles.
3
2
9
4
.3 nm (Fig. 4b and Table 1). In addition, Au species were deposited
by SG on TS-1 which was treated in a similar manner to the DP
method by replacing HAuCl with HCl (see Part 2.2. Deposition of
4
Au by DP). Only Au NPs (>2.0 nm) but not Au clusters (<2.0 nm)
could be formed. In contrast, a large number of Au particles could
be clearly observed over 0.10 wt% Au/TS-1-Na1(SG) (Fig. 4c), and a
high population density of 49.7 Au particles per 100 nm ꢂ 100 nm
was achieved. Furthermore, the diameter distribution of Au parti-
cles was very narrow (1.0–5.0 nm), giving a mean diameter of
In the preparation of Au/TS-1 and Au/TS-1-Na1 by DP or SG,
there are mainly two possible locations for Au deposition. One
location is in the microporous channels of the supports, where
1
.6 nm (Fig. 4d and Table 1), and about 90% of Au particles were
clusters (<2.0 nm).
3
Au should exist as tiny clusters such as Au with diameters smaller
In addition, Au with a higher loading of about 0.20 wt% was also
deposited on TS-1-Na1 by both SG and DP method (Table 1).
HAADF-STEM observations showed that in 0.20 wt% Au/TS-1-
Na1(SG), the population density of Au particles per
than the size of channels, ꢁ0.55 nm. The other is on the exterior
surfaces of TS-1 and TS-1-Na1, where Au particles could exist as
1
29
clusters (<2.0 nm) and NPs (>2.0 nm).
Xe NMR characterizations
and HAADF-STEM have been utilized to detect Au particles inside
microporous channels and over the exterior surfaces, respectively.
1
00 nm ꢂ 100 nm was 51.4 (Fig. 5a), higher than that of 17.4 Au
particles per 100 nm ꢂ 100 nm in 0.23 wt% Au/TS-1-Na1(DP)
Fig. 5c). In addition, in 0.20 wt% Au/TS-1-Na1(SG), the diameter
1
29
Before analyzing the
lysts, we first focus on the
and TS-1-Na1). As described in Part 3.2.,
Xe NMR results of supported Au cata-
1
29
(
Xe NMR results of supports (TS-1
1
29
distribution is very narrow and 87% of Au particles existed as clus-
ters smaller than 2.0 nm (Fig. 5b), while in 0.23 wt% Au/TS-1-
Na1(DP), only 66% of Au particles existed as clusters (Fig. 5d).
It should be noted that after the deposition of Au on TS-1-Na1
by SG, the color of Au/TS-1-Na1(SG) was white, the color of TS-1-
Na1 itself. However, during the deposition of Au on TS-1 without
alkaline pretreatment by SG, the color of Au/TS-1(SG) changed
Xe NMR characteriza-
tion, TS-1 (Fig. 2) and TS-1-Na1 (Fig. 3) displayed very different
chemical shift value (98.1 vs. 74.1 ppm) at low xenon loadings
and very different shapes of the curves (linear vs. nonlinear). Since
both supports possess similar elemental contents, similar crystal-
line structures and similar morphologies, the differences above
can be ascribed to their different pore structures. Although TS-1

1
2
J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
Fig. 2. Effect of adsorbed Xe atoms per gram of samples on 129Xe NMR chemical
shifts (T = 291 K) on TS-1 (s) and 0.11 wt% Au/TS-1(DP) (d).
Fig. 3. Effect of adsorbed Xe atoms per gram of samples on 129Xe NMR chemical
shifts (T = 291 K) on TS-1-Na1 (j), 0.20 wt% Au/TS-1-Na1(SG) (N) and 0.23 wt% Au/
TS-1-Na1(DP) (.).
(
treated with aqueous NaOH solution at pH of 9.0) possesses only
micropores, TS-1-Na1 possesses both mesopores (created during
alkaline treatment at pH of 12) and micropores based on N
ical shift of reference and has been fixed to zero, dXe–W is the
parameter due to collisions between xenon atoms and zeolite
walls, dXe–Xe is caused by xenon–xenon collisions. At high loadings
of xenon, dXe–Xe determines the value of d, while at low loadings of
xenon, dXe–W plays a crucial role. Related investigation showed that
the interaction of adsorbed Xe atoms with microporous wall was
2
129
adsorption measurements and
10 K (described in the following).
For TS-1 and TS-1-Na1, the 129Xe NMR chemical shift could be
expressed as follows: d = d + dXe–W + dXe–Xe, where d is the chem-
Xe NMR characterization at
2
0
0
Fig. 4. HAADF-STEM images (a and c) and diameter distribution of Au particles (b and d) for 0.11 wt% Au/TS-1(DP) (a and b) and 0.10 wt% Au/TS-1-Na1(SG) (c and d). HAADF-
STEM observation of Au catalysts was carried out after catalytic tests at 473 K for 2.0 h. DAu denotes diameters of Au particles.

J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
13
Fig. 5. HAADF-STEM images (a and c) and diameter distribution of Au particles (b and d) for 0.20 wt% Au/TS-1-Na1(SG) (a and b) and 0.23 wt% Au/TS-1-Na1(DP) (c and d).
HAADF-STEM observation of Au catalysts was carried out after catalytic tests at 473 K for 2.0 h. DAu denotes diameters of Au particles.
much stronger than that with mesoporous wall and thus gives a
higher chemical shift [31,32]. Since 1 Xe NMR spectroscopy at
low temperature is very sensitive to the pore structure of supports,
the parameter due to collisions between xenon atoms and metal
clusters incorporated into the channels of zeolite. Previous investi-
29
gation showed that the incorporation of metal clusters into micro-
1
29
129
the
Xe NMR characterizations of TS-1-Na1 at low temperature
porous channels of zeolites led to a great increase in
Xe
were carried out. As shown in Fig. 6, when the operation tempera-
ture decreased to 210 K, two peaks could be clearly observed. The
peak at high d is due to xenon adsorbed in micropores, and the
peak with low d is due to xenon adsorbed on mesopores. The suc-
cessful detection of mesopores in TS-1-Na1 by 1 Xe NMR spec-
chemical shifts at low Xe loading range, indicating that the interac-
tion of adsorbed Xe atoms with incorporated metal clusters is
much stronger than that with microporous wall of zeolites [24–
28]. Therefore, at low Xe loadings, the observed chemical shifts d
29
(d = d
0
+ dXe–M + dXe–W + dXe–Xe) can be ascribed mainly to dXe–M
.
1
29
troscopy confirms the analyses based on
measurements.
N
2
adsorption
Accordingly, we will focus on the
Xe NMR chemical shift curve
at low Xe loading. As shown in Fig. 2, 0.11 wt% Au/TS-1(DP) dis-
plays at a low xenon loading a chemical shift of 110.5 ppm,
12.4 ppm higher than that (98.1 ppm) of TS-1, indicating that tiny
Au clusters have been incorporated into microporous channels of
TS-1 by DP. In addition, it is also found that after the deposition
of Au on TS-1 by DP, the slope of the curve d versus xenon amount
increases, meaning the decrease in the volume accessible to xenon
due to the presence of Au species. This further confirms the incor-
poration of tiny Au clusters into microporous channels of TS-1 by
DP. The successful detection of Au clusters incorporated in the
microporous channels of TS-1 by Xe NMR technique indicates that
Xe atoms can freely diffuse into the microporous channels.
Although the pore structure of TS-1 is 3D interconnected and Xe
atoms can diffuse into the microporous channels from anyone of
the surface pore openings, the exterior pore openings are also
the preferential adsorption sites for Au species during DP process.
Therefore, tiny Au clusters incorporated might be located at defect
sites inside microporous channels because the size of channels of
TS-1 was only ꢁ0.55 nm and the diameter of Au atoms is about
0.29 nm. This finding is consistent with the prediction by Joshi
et al. [21] based on theoretical calculations that in Au/TS-1
Now, it appears reasonable that at a low Xe loading, the chem-
ical shift over TS-1 at 291 K (Fig. 2) is higher than that over TS-1-
Na1 at 291 K (Fig. 3), because the former corresponds to only
strong dXe–W (microporous wall), while the latter corresponds to
an average dXe–W of both strong dXe–W (microporous wall) and
weak dXe–W (mesoporous wall) due to the fast exchange of Xe
atoms in micropores and in mesopores. With an increase in Xe
loading, over TS-1-Na1 the contribution from dXe–W (mesoporous
wall) will increase, and thus the chemical shift decreases corre-
spondingly (Fig. 3). After the chemical shift reaches the minimum,
it will increase with Xe loadings due to the contribution of dXe–Xe
contribution. At high xenon loadings, the slopes of chemical shift
versus xenon amount over both TS-1 (Fig. 2) and TS-1-Na1
(Fig. 3) are similar, which suggests that at high xenon loadings, xe-
non–xenon collisions contribute more significantly to give smaller
difference in chemical shift. However, the chemical shift values are
not exactly the same probably due to the differences in the porous
network of these samples.
For metal/zeolites, the 129Xe NMR chemical shift could be ex-
0
pressed as follows: d = d + dXe–M + dXe–W + dXe–Xe, where dXe–M is

1
4
J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
Fig. 7. Effect of adsorbed Xe atoms per gram of samples on 129Xe NMR chemical
shifts (T = 210 K) of xenon adsorbed in micropores of both TS-1-Na1 (j) and
Fig. 6. Temperature dependence of the 129Xe NMR spectrum in TS-1-Na1 sample
equilibrated at 300 Torr of xenon.
0.20 wt% Au/TS-1-Na1(SG) (N) and in mesopores of both TS-1-Na1 (h) and 0.20 wt%
Au/TS-1-Na1(SG) (4).
Au/TS-1(DP) and Au/TS-1-Na1(DP), tiny Au clusters incorporated
into microporous channels should be the strong adsorption sites.
However, it is not the case for TS-1-Na1 and Au/TS-1-Na1(SG). As
shown in Fig. 7, TS-1-Na1 displays one minimum in chemical shift
curve (at low Xe loadings and at 210 K) attributed to Xe atoms ad-
sorbed into microporous channels, while Au/TS-1-Na1(SG) dis-
plays two minimum (at low Xe loadings and at 210 K) in
chemical shift curves of Xe atoms adsorbed both into microporous
channels and into mesoporous channels, respectively. For TS-1-
Na1, the only strong adsorption sites might correspond to the de-
fects created during the alkaline treatment of TS-1. However, for
Au/TS-1-Na1(SG), the two strong adsorption sites might corre-
spond to both the defects and Au species incorporated into meso-
porous channels, respectively.
HAADF-STEM observation shows that over the exterior surfaces,
the population density of Au particles over 0.23 wt% Au/TS-1-
Na1(DP) is 17.4 Au particles per 100 nm ꢂ 100 nm (Fig. 5c), lower
than that of 51.4 Au particles per 100 nm ꢂ 100 nm over 0.20 wt%
Au/TS-1-Na1(SG) (Fig. 5a), although 0.23 wt% Au/TS-1-Na1(DP)
possesses a slightly higher Au loading and a similar mean diameter
of Au particles in comparison with 0.20 wt% Au/TS-1-Na1(SG) (Au
loading, 0.23 vs. 0.20 wt%; mean diameter, 1.7 vs. 1.8 nm; respec-
tively). Interestingly, even Au/TS-1-Na1(SG) with a lower Au load-
ing of 0.10 wt% and with a mean diameter of 1.6 nm also displays
much higher population density of 49.7 Au particles per
100 nm ꢂ 100 nm than 0.23 wt% Au/TS-1-Na1(DP). Therefore, it is
likely that some part of Au atoms in 0.23 wt% Au/TS-1-Na1(DP)
prepared by DP, tiny Au clusters are preferentially adsorbed on de-
fect sites of microporous channels.
As shown in Fig. 3 the chemical shift over 0.23 wt% Au/TS-1-
Na1(DP) is 12.5 ppm higher than that over TS-1-Na1 (86.6 vs.
7
4.1 ppm) at a low Xe loading, similar to the phenomenon ob-
served over 0.11 wt% Au/TS-1(DP) and TS-1, indicating the incorpo-
ration of tiny Au clusters into the microporous channels of TS-1-
Na1 by DP. On the contrary, 0.20 wt% Au/TS-1-Na1(SG) displays
only a slightly higher chemical shift than that of TS-1-Na1 (77.6
vs. 74.1 ppm). The low values of observed chemical shifts in
0
.20 wt% Au/TS-1-Na1(SG) suggest that Au species were not incor-
porated into microporous channels by the SG method but were lo-
cated over the exterior surfaces as Au clusters and/or
nanoparticles. Actually, the slopes of asymptote of curve Au/TS-
1
-Na1(SG) are very close to those of TS-1-Na1, supporting the
hypothesis that Au species were preferentially located on the
external surfaces. However, the possibility that small amount of
Au species were incorporated into microporous channels and/or
mesoporous channels could not be ruled out due to the slightly
higher chemical shift over 0.20 wt% Au/TS-1-Na1(SG) than that
over TS-1-Na1 at a low Xe loading. Therefore, further 1 Xe NMR
characterizations of TS-1-Na1 and 0.20 wt% Au/TS-1-Na1(SG) at
29
2
10 K were carried out.
As shown in Fig. 7, chemical shift due to xenon adsorbed in
micropores is very close for TS-1-Na1 and 0.20 wt% Au/TS-1-
Na1(SG) samples. However, the chemical shifts due to xenon ad-
sorbed in mesopores for 0.20 wt% Au/TS-1-Na1(SG) are slightly
higher than that for TS-1-Na1, especially at low Xe loadings. This
clearly indicates that during the preparation of Au/TS-1-Na1 by
SG, Au species could not be deposited into microporous channels,
but a small amount of Au species could be deposited into mesopor-
ous channels probably owing to the relatively easier diffusion of
were invisible by HAADF-STEM with
a detection limit of
ꢁ0.6 nm, and they should exist as tiny clusters with atoms less
than 11 (Au11: ꢁ0.8 nm in diameter). Recent experiments by Liu
et al. [34] have showed that even under vacuum conditions, Au11
clusters incorporated into mesoporous channels of mesoporous sil-
ica (SBA-15) and protected by the rough surfaces of mesoporous
wall are not stable at 473 K. Therefore, it is probable that in
3 2
bulky (CH ) Au(acac) into mesoporous channels than into micro-
porous channels.
3
0.23 wt% Au/TS-1-Na1(DP), tiny Au clusters such as Au deposited
It should be noted that 129Xe NMR chemical shift curve over TS-
on the exterior surfaces and in mesoporous channels would sinter
to form Au clusters larger than Au11 and/or NPs under our reaction
1
increases almost linearly with Xe concentration (Fig. 2), while
those over TS-1-Na1, Au/TS-1(DP), Au/TS-1-Na1(DP) and Au/TS-
-Na1(SG) first decrease and then increase almost linearly with
ꢀ1
conditions (temperature, 473 K; feed gas flow, 20 mL min ),
whereas tiny clusters could be stabilized inside the microporous
1
1
29
Xe concentration, giving a minimum of chemical shift at low Xe
loadings (Figs. 2 and 3). The presence of these minimum usually
means the presence of strong adsorption sites (SAS) [33]. For
channels, which is consistent with
Based on the combined analyses by
STEM, a conclusion can be drawn that tiny Au clusters such as
Xe NMR characterizations.
1
29
Xe NMR and HAADF-

J. Huang et al. / Journal of Catalysis 278 (2011) 8–15
15
Au
alkaline-treated TS-1 by DP but not by SG, while Au clusters
1.0–2.0 nm) and NPs (>2.0 nm) are deposited on the exterior sur-
3
are incorporated into microporous channels of both TS-1 and
clusters, and accordingly, Au clusters (1.0–2.0 nm) over the exte-
rior surfaces are more active for PO synthesis than tiny Au clusters
inside the microporous channels.
(
faces of TS-1 and alkaline-treated TS-1 by both DP and SG methods.
In addition, for Au/TS-1-Na1 prepared by both DP and SG, some Au
species might be incorporated into mesoporous channels as Au
clusters (larger than Au11) and/or as NPs. However, the loading of
Au incorporated into mesoporous channels should be much smal-
ler than that over the exterior surfaces because the amount of mes-
opores in TS-1-Na1 is very small [8] and the diffusion and
adsorption of Au species on the exterior surfaces is much easier
than that into mesoporous channels during the preparation of
Au/TS-1-Na1 by both DP and SG, especially by SG.
Acknowledgment
Thanks from E. Lima are due to Conacyt (Consejo Nacional de
Ciencia y Tecnologia, Mexico) for Grant 24676, 101319 and PAP-
IIT-UNAM IN107110.
References
[
[
1] T.A. Nijhuis, M. Makkee, J.A. Moulijn, B.M. Weckhuysen, Ind. Eng. Chem. Res. 45
2006) 3447.
2] A.H. Tullo, P.L. Short, Chem. Eng. News 84 (41) (2006) 22.
(
Joshi et al. [20,21] have predicted based on theoretical calcula-
tions that in Au/TS-1, tiny Au clusters incorporated into the micro-
porous channels of TS-1 could be active for PO synthesis by the
[3] M. Ishino, J. Yamamoto, Shokubai 48 (2006) 511.
[
[
4] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566.
5] A.K. Sinha, S. Seelan, S. Tsubota, M. Haruta, Angew. Chem. Int. Ed. 43 (2004)
2 2
reaction of propene with O and H mixture. However, our exper-
1546.
iments strongly indicated that Au clusters (1.0–2.0 nm) deposited
on the exterior surfaces of alkaline-treated TS-1 were mainly
responsible for PO synthesis [8]. The PO formation rate over
[6] B. Chowdhury, J.J. Bravo-Suárez, M. Daté, S. Tsubota, M. Haruta, Angew. Chem.
Int. Ed. 45 (2006) 412.
[
7] B. Chowdhury, J.J. Bravo-Suárez, N. Mimura, J. Lu, K.K. Bando, S. Tsubota, M.
Haruta, J. Phys. Chem. B 110 (2006) 22995.
ꢀ
1
ꢀ1
0
.23 wt% Au/TS-1-Na1(DP) is 74 gPO kg
h
, much lower than
[8] J. Huang, T. Takei, T. Akita, H. Ohashi, M. Haruta, Appl. Catal. B 95 (2010) 430.
[9] J. Lu, X. Zhang, J.J. Bravo-Suárez, K.K. Bando, T. Fujitani, S.T. Oyama, J. Catal. 250
(2007) 350.
10] J.J. Bravo-Suárez, K.K. Bando, J. Lu, M. Haruta, T. Fujitani, S.T. Oyama, J. Phys.
Chem. C 112 (2008) 1115.
cat
ꢀ1
ꢀ1
that of 127 gPO kg
lower than that of 119 gPO kg
h
over 0.20 wt% Au/TS-1-Na1(SG) and even
cat
ꢀ1
ꢀ1
h
over 0.10 wt% Au/TS-1-Na1(SG)
cat
[
(
see Table 1). In Au/TS-1-Na1(DP), the DP method produces both
tiny Au clusters (<0.55 nm) inside microporous channels and Au
clusters (1.0–2.0 nm) on the exterior surfaces, while in Au/TS-1-
Na1(SG), the SG method produces Au clusters (1.0–2.0 nm) selec-
tively on the exterior surfaces. The lower catalytic activity over
Au/TS-1-Na1(DP) than that over Au/TS-1-Na1(SG) suggests that
Au clusters (1.0–2.0 nm) deposited on the exterior surfaces might
be more active than tiny Au clusters (<0.55 nm) incorporated into
microporous channels.
[11] S.T. Oyama, X. Zhang, J. Lu, Y. Gu, T. Fujitani, J. Catal. 257 (2008) 1.
[
[
[
12] T.A. Nijhuis, T. Visser, B.M. Weckhuysen, Angew. Chem. Int. Ed. 44 (2005) 1115.
13] E. Sacaliuc, A.M. Beale, B.M. Weckhuysen, T.A. Nijhuis, J. Catal. 248 (2007) 235.
14] A. Zwijnenburg, M. Makkee, J.A. Moulijn, Appl. Catal. A 270 (2004) 49.
[15] N. Yap, R.P. Andres, W.N. Delgass, J. Catal. 226 (2004) 156.
[
[
[
16] B. Taylor, J. Lauterbach, W.N. Delgass, Appl. Catal. A 291 (2005) 188.
17] L. Cumaranatunge, W.N. Delgass, J. Catal. 232 (2005) 38.
18] B. Taylor, J. Lauterbach, W.N. Delgass, Catal. Today 123 (2007) 50.
[19] J. Lu, X. Zhang, J.J. Bravo-Suárez, T. Fujitani, S.T. Oyama, Catal. Today 147
2009) 186.
(
[
[
20] A.M. Joshi, W.N. Delgass, K.T. Thomson, J. Phys. Chem. B 110 (2006) 2572.
21] A.M. Joshi, W.N. Delgass, K.T. Thomson, J. Phys. Chem. C 111 (2007) 7841.
5
. Conclusions
[22] T. Ito, J. Fraissard, J. Chem. Phys. 76 (11) (1982) 5225.
[
[
23] J. Fraissard, T. Ito, Zeolites 8 (1988) 350.
24] A. Labouriau, G. Panjabi, B. Enderle, T. Pietrass, B.C. Gates, W.L. Earl, K.C. Ott, J.
Am. Chem. Soc. 121 (1999) 7674.
HAADF-STEM combined with 129Xe NMR and elemental analy-
ses indicate that by the SG method, Au is selectively deposited
on the exterior surfaces of alkaline-treated TS-1 as clusters (1.0–
[25] B.A. Enderle, A. Labouriau, K.C. Ott, B.C. Gates, Nano Lett. 2 (2002) 1269.
[
[
26] B. Enderle, A. Labouriau, K.C. Ott, B.C. Gates, J. Phys. Chem. B 106 (2002) 2109.
27] A. Guzmán-Vargas, E. Lima, G. Delahay, B. Coq, V. Lara, Ind. Eng. Chem. Res. 45
2
.0 nm) and partly as NPs (>2.0 nm), while by the DP method, Au
(
2006) 4163.
is deposited not only over the exterior surfaces of TS-1 and alka-
line-treated TS-1 as clusters (1.0–2.0 nm) and partly as NPs
[28] D. Guillemot, V.Yu. Borovkov, V.B. Kazansky, M. Polisset-Thfoin, J. Fraissard, J.
Chem. Soc. Faraday Trans. 93 (19) (1997) 3587.
[
29] J. Huang, T. Akita, J. Faye, T. Fujitani, T. Takei, M. Haruta, Angew. Chem. Int. Ed.
8 (2009) 7862.
[30] J. Guzman, B.C. Gates, Langmuir 19 (2003) 3897.
(
>2.0 nm) but also inside the microporous channels (ꢁ0.55 nm)
4
as tiny clusters. These tiny Au clusters might reside on the defect
sites inside microporous channels, agreeing with the analyses by
Joshi et al. based on theoretical calculations. In propene epoxida-
[
[
[
[
31] C. Jin, G. Li, X. Wang, L. Zhao, L. Liu, H. Liu, Y. Liu, W. Zhang, X. Han, X. Bao,
Chem. Mater. 19 (2007) 1664.
32] Y. Liu, W. Zhang, Z. Liu, S. Xu, Y. Wang, Z. Xie, X. Han, X. Bao, J. Phys. Chem. C
112 (2008) 15375.
33] J.L. Bonardet, J. Fraissard, A. Gedeon, M.A. Springuel-Huet, Catal. Rev. Sci. Eng.
41 (1999) 115.
34] Y. Liu, H. Tsunoyama, T. Akita, T. Tsukuda, Chem. Lett. 39 (2010) 159.
tion with an O
2 2
and H mixture, PO formation rate over 0.23 wt%
Au/TS-1-Na1(DP) with
a
mean diameter of 1.7 nm is
ꢀ1
ꢀ1
ꢀ1 ꢀ1
7
0
4 gPO kg
h
, much lower than that of 127 gPO kgcat
h
over
cat
.20 wt% Au/TS-1-Na1(SG) with a mean diameter of 1.8 nm. The
only difference between the two catalysts is the location of Au