The Construction of Au–Fe–TS-1 Interface Coupling Structure for Improving Catalytic Performance of Propylene Epoxidation with H2 and O2

Article abstract of DOI:10.1007/s10562-020-03222-y

Abstract: The gas-phase propylene epoxidation for noble metal system as a green process has confronted a great challenging task in metal utilization and stability. In this work, we realize the enhancement of Au capture efficiency and high-dispersed Au nanoparticles by supports modification and interfacial regulation to construct Au–Fe–TS-1 interface coupling structure. The Au capture efficiency of the optimum catalyst arrived at 69percent and it displayed a 7.3percent C₃H₆ conversion with a PO selectivity of 89.9percent, giving a H₂ efficiency of about 25.7percent. Then some typical characterizations systematically elucidate that Fe is not only incorporated into TS-1 framework and captures Au species with high dispersion and stability, but also inhibits the losing of Ti species. Furthermore, Fe-modified catalysts change the property of Au active site and maintain lower apparent activation energy. This report may provide a new strategy to design advanced catalysts for enhancing metal utilization efficiency by the construction of interface coupling structure. Graphic Abstract: [Figure not available: see fulltext.].

Full text of DOI:10.1007/s10562-020-03222-y

Catalysis Letters
https://doi.org/10.1007/s10562-020-03222-y
The Construction of Au–Fe–TS‑1 Interface Coupling Structure
for Improving Catalytic Performance of Propylene Epoxidation with H2
and O2
Xiaohan Li1,2 · Along Gao1,2 · Zhehong Wan1,2 · Qingming Huang3 · Xiaohui Chen1,2
Received: 16 January 2020 / Accepted: 16 April 2020
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
The gas-phase propylene epoxidation for noble metal system as a green process has confronted a great challenging task in
metal utilization and stability. In this work, we realize the enhancement of Au capture efciency and high-dispersed Au
nanoparticles by supports modifcation and interfacial regulation to construct Au–Fe–TS-1 interface coupling structure. The
Au capture efciency of the optimum catalyst arrived at 69% and it displayed a 7.3% C3H6 conversion with a PO selectivity
of 89.9%, giving a H2 efciency of about 25.7%. Then some typical characterizations systematically elucidate that Fe is not
only incorporated into TS-1 framework and captures Au species with high dispersion and stability, but also inhibits the losing
of Ti species. Furthermore, Fe-modifed catalysts change the property of Au active site and maintain lower apparent activa-
tion energy. This report may provide a new strategy to design advanced catalysts for enhancing metal utilization efciency
by the construction of interface coupling structure.
Graphic Abstract
Keywords Propylene epoxidation · Gold catalysts · Fe modifed · Interface coupling structure · Capture efciency
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10562-020-03222-y) contains
supplementary material, which is available to authorized users.
2
* Qingming Huang
National Engineering Research Center of Chemical Fertilizer
Catalyst, Fuzhou University, Fuzhou 350002, Fujian, China
qmhuang@fzu.edu.cn
3
* Xiaohui Chen
Testing Center, Fuzhou University, Fuzhou 350116, Fujian,
China
chenxhfzu@fzu.edu.cn
1
College of Chemical Engineering, Fuzhou University,
Fuzhou 350116, Fujian, China
Vol.:(0123456789)
1 3

X. Li et al.
be controlled by metal-support interaction [20, 21], which
is mainly involved the interfacial efect by electron transfer
and interfacial vacancies. Therefore, constructing defects
by supports modifcation and regulation may be one of
efective ways to enhance the deposition and dispersion
of Au NPs.
1 Introduction
As an important bulk chemical feedstock, propylene oxide
(PO) is a high-value-added commodity and has attracted
great attention to petrochemical industries, aiming at
manufacturing a variety of intermediate materials such
as polyurethane, polyester and propylene glycol used for
foams, textile, and additives [1–3]. The tradition routes for
PO formation using propylene as raw material were chlo-
rohydrin and hydroperoxide processes [4]. However, the
former is considered as toxic, corrosive and environmental
unfriendly process. And the latter as oxidants seems to be
a green process with high PO selectivity, but is still sub-
jected to high cost and difcult to transport. Both of them
require complex steps in the liquid phase and produce
amounts of waste byproducts. Consequently, it is neces-
sary to design a productive, energy-saving and green pro-
cess to produce PO. Considering the above factors, direct
gas-phase epoxidation of propylene using molecular H2
and O2 has been put forward and prevalent investigated
for industrial application. Gas-phase catalysts have been
mainly gold- [5–7], silver- [8, 9] or copper-based [10] sub-
stances and other photocatalysts [11]. However, the Au
catalysts have become a popular subject in recent years
for the mild reaction conditions and high PO selectivity.
It was frst reported for propylene epoxidation in 1998
by Haruta [12] that gold particles deposited on TiO2
using deposition precipitation (DP) method showed a PO
selectivity higher than 90% with a 1–2% propylene con-
version. This route is simple and green at low tempera-
ture and attracts great attention for further investigation.
Subsequently, many researchers take measures to regulate
physical and chemical properties of Au catalysts with Ti-
containing supports and simultaneously achieve higher
performance [5, 7, 13–15]. In another study, heteroatoms
doping is introduced as a novel modifcation strategy to
tune size and electronic efect [16, 17]. It is an efcient
way to introduce hetero-metal as modifer for Au catalysts
by DP method to change the nature and amount of gold
active sites [6, 18, 19]. Au catalysts modifed by alkali and
alkaline earth metal compounds have been investigated
and the results show that alkaline earth metals are better
than alkali metals by the efect of the isoelectric point of
the support, increasing Au capture efciency and disper-
sion [6]. In recent years, the addition of Ag(NO3)3 as a
promoter for Au/uncalcined TS-1 catalysts has a decrease
in Au nanoparticle and enhances oxygen adsorption from
Au to O2 [19]. Although DP method has become an efec-
tive way to prepare Au-based catalysts, more than 50% of
Au ions are lost with the process of washing [6]. It may be
explained by the poor anchor efect between supports and
Au ion since the catalysis of supported metal catalysts can
Hierarchical TS-1 with secondary macropores has been
synthesized by mild acid etching and inevitably extracted
Si in the TS-1 sample [22]. Similarly, alkaline treated TS-1
possessing a large specifc surface area has the property
of rough exterior surfaces and presents a large number of
defects, where metal species could be highly anchored and
dispersed [5, 23]. Overall, no matter acid or alkaline treat-
ment gives rise to slightly etch on the surface of TS-1. More-
over, ion exchange by introducing a heterovalent metal has
emerged as an important strategy for heterogeneous metal
catalysts to capture active metal and control the size and
distribution of dispersed metal [24–26]. Accoding to the lit-
eratures, iron with weak electronegativity can accomplish
the electrons transfer with gold [27, 28] and also has ability
to increase the number of vacancy sites to form peroxide
or superoxide [29, 30]. Therefore, tetrachloroauric acid and
[Fe(III), NH4]-EDTA as Au and Fe precursor, respectively,
are selected to synchronous satisfy supports etching efect
and capture active components via electron transfer. And it is
not only shed light on the role of Fe chelate based on struc-
tural- physicochemical features but also further revealed the
mechanism for improving catalytic performance.
2 Experimental
TS-1 was supplied with a Si/Ti ratio of 50 by Pengcheng
Chemical Co., Ltd. [Fe(III), NH4]-EDTA was purchased by
Yuancheng Chemical Co., Ltd, tetrachloroauric acid (47.8%
Au) and cesium carbonate (AR) by Sinopharm Chemical
Reagent Co., Ltd. All were used without further purifcation.
2.1 Catalyst Preparation
2.1.1 Deposition of Gold
Au/TS-1 catalyst was prepared by the DP method with
Cs2CO3 as precipitant in the absence of light. Then 1 mL
of HAuCl4 (0.0485 M) and 1 g of TS-1 were dispersed in
20 mL of deionized water at 343 K for 0.5 h with ultrasonic
treatment. The pH was kept by adding 0.5 M Cs2CO3 at 7–8.
And the suspension was placed in a water bath at 343 K stir-
ring for another 5.5 h. The solution was not centrifuged and
washed once until it was cooled down to room temperature.
Finally, the solid was dried overnight at 333 K in a vacuum
oven without any calcination.
1 3

The Construction of Au–Fe–TS-1 Interface Coupling Structure for Improving Catalytic Performance…
2.1.2 Synthesis of Bimetallic Catalysts
Moles of PO
H2 efficiency =
× 100%
Moles of H2 converted
The bimetallic Au–Fe/TS-1 catalysts were prepared by the DP
method. Typically, 1 mL of HAuCl4 (0.0485 M) was dissolved
in 15 mL of deionized water, following by adding 1 g of TS-1
support. The mixture was treated at 343 K under supersonic.
Then a desired amount of Fe(III) chelate precursor, namely
[Fe(III), NH4]-EDTA, was dispersed in 5 mL of water and
added dropwise into the above mixture for 0.5 h. The mix-
ture was neutralized to a pH of 7 by slowly addition of 0.5 M
Cs2CO3. The slurry was transferred into a water bath and con-
tinued to vigorous stir for another 5.5 h at the same tempera-
ture. Finally, it was centrifuged, washed once, and dried over-
night under vacuum at 333 K. There was no further calcination
for all samples. The theoretical Au loading of each catalyst is
kept at 1 wt%, while Fe content is adjusted with diferent Fe/
Au molar ratios of 0.5, 0.6, 0.7, 0.8 and 0.9 according to theo-
retical Au loading. All the prepared catalysts were named as
Au–Fex/TS-1, where x is represented as the Fe/Au molar ratio.
Cpo
PO formation rate =
× 100%
mAu
where mAu stands for the mass of Au, and CPO means the
molar fow rate of PO.
2.3 Characterizations
The X-ray difraction patterns (XRD) of prepared cata-
lysts and support were collected on a PANalytical X’pert
Pro difractometer using Cu Kα radiation (45 kV, 40 mA,
λ = 0.1542 nm) at a scanning speed of 10 K min−1 with a
range of 2θ from 5° to 80°. Lattice parameters are calcu-
lated by crystal structure Rietveld refnement method. The
ultraviolet–visible spectroscopy (UV–Vis) was carried out
on a Perkin-Elmer Lambda 950 spectrophotometer using
BaSO4 as the reference. The Fourier transform infrared
spectroscopy (FTIR) was recorded on Nicolet 6700 FTIR
spectrometer with the frequency range of 4000–400 cm−1.
The actually Au content in each sample was determined by
inductively coupled plasma-optical emission spectroscopy
(ICP-OES) on OPTIMA 8000 (PerkinElmer). Transmis-
sion electron microscopy (TEM) was obtained on a Tecnai
G2F20 microscope. The X-ray photoelectron spectroscopy
(XPS) spectra were taken on an ESCALAB 250xi spectrom-
eter (Thermo Scientifc) with Al Kα radiation source. The
binding energies were calibrated using C1s peak (284.8 eV)
as a reference.
2.2 Catalyst Testing
The activity of propylene epoxidation was carried out using
0.15 g of catalysts in a quartz tubular reactor with an inner
diameter of 5 mm. The catalysts were tested under atmos-
pheric pressure and heated from room temperature to 473 K
for about 0.5 h, then reduced with a mixture gas H2/N2 of
3.5/12.5 by volume at 473 K for an hour. After the reduced
pretreatment, it was continued to heat until arrived reaction
temperature. In this work, the reactant gas mixture was kept
at propylene (C3H6, 99.99%):oxygen (O2, 99.99%): hydrogen
(H2, 99.99%):nitrogen (N2, 99.99%)=3.5:3.5:3.5:24.5 mL m
in−1, with a total fow rate of 35 ml min−1 and a total space
velocity of 14,000 mL h−1 gcat−1. The reactants and products
were analyzed using an on-line gas chromatograph (GC-2060)
equipped with a TCD detector. All lines and valves between
the exit of the reactor and the gas chromatograph were kept
at 130 ℃ to prevent condensation and ensure efcient separa-
tion of organic products. The oxygenated products contain PO,
propionaldehyde, CO2 and acetone. The propylene conversion,
PO selectivity, H2 efciency, and PO formation rate were cal-
culated as follows:
3 Results and Discussion
3.1 Catalyst Activity Measurements
Diferent Fe content catalysts are analyzed with H2 and
O2 for propylene epoxidation and the results are shown
in Table 1 and Table S1. Au/TS-1 catalyst has the lowest
C3H6 conversion of 0.75% under the test conditions. It is
signifcantly increased as the Fe is incorporated into Au/
TS-1 catalyst. Moreover, when the molar ratio of Fe/Au is
up to 0.7, the propylene conversion is sharply increased to
7.3%, about nine times than that of Au/TS-1. According to
the literature, the enhancement of Au capture efciency has
a positive efect on the activity of Au/TS-1 for propylene
epoxidation [31]. In this work, the amount of Au species is
sharply increased according to ICP results ranging from 10
to 69%. It is revealed that the addition of Fe has improved
the loading efciency of Au. However, the catalytic activ-
ity shows a volcano-trend and is gradually decreased as the
Propylene conversion
∑
oxygenated products converted to C3
=
× 100%
initial C3H6
PO selectivity
Moles of PO
=
× 100%
∑
Moles of oxygenated products converted to C3
1 3

X. Li et al.
Table 1 Propylene epoxidation over Au–Fex/TS-1 catalysts with diferent Fe loadings
Au catalysts
Au capture
Conversion (%)
Selectivity (%)
Efciency of PO formation rate
efciency (%)a
H2 (~%)
−1
−1
PO
CO2
Othersb
(gPO h−1 gAu
)
(gPO h−1 kgcat
)
Au/TS-1
10
59
64
69
65
57
0.75
4.8
5.9
7.3
6.9
6.5
94.3
88.1
90.4
89.9
88.8
89.2
5.7
9.0
7.7
8.2
9.1
8.6
Trace
2.9
1.9
2.2
2.1
37.2
18.4
20.1
25.7
26.1
26.5
25.7
26.1
30.3
34.6
34.2
36.9
25.7
153.6
193.6
238.8
222.5
210.5
Au–Fe0.5/TS-1
Au–Fe0.6/TS-1
Au–Fe0.7/TS-1
Au–Fe0.8/TS-1
Au–Fe0.9/TS-1
2.2
Reaction conditions: 0.15
GHSV=14,000 mL gcat−1 h−1
g
catalyst mixed; C3H6/O2/H2/N2 =3.5/3.5/3.5/24.5 mL min−1
;
Reaction temperature 503 K;
aDetermined by ICP-OES
bOthers contain propionaldehyde and acetone
molar ratio is more than 0.7. In order to demonstrate the role
of Au in propylene epoxidation, Fe/TS-1 was tested in the
same condition, which is implied only the presence of Fe
is inactive for propylene epoxidation. It is possible because
H2O2, the intermediate species, may not be generated by
H2 and O2 on Fe surface for the epoxidation step; however,
Au nanoparticles are more active in PO formation [19, 32].
In addition, PO formation rate has increased about 10 gPO
h−1 gAu−1 on per gram of Au with more Fe loading. It can be
explained by the changed property of Au active site. Interest-
ingly, the rate is sharply increased for per kilogram catalysts.
The addition of Fe makes more Au particles deposited on
support surface and enhances the whole catalysis for PO
formation rate, which could be attributed to the changed
number of Au active site. Therefore, an appropriate ratio
is conducive to increase the gold content and further efec-
tively improves catalytic performance.
number of surface defects in acid or alkaline treatments,
according to literatures [35, 36]. Figure 1 shows the XRD
patterns of supports, single metal-supported catalysts and
standard orthorhombic TS-1 (JCPDS#70-4276) with MFI
topological structure. In comparison, the crystal structure
of the MFI framework is not changed when supported
acidity Au or alkalescent Fe precursor solution. In addi-
tion, lattice parameters are analyzed from Table S3 in
order to explore the change of TS-1 microstructure. It
shows that no matter Au or Fe supported makes unit cell
volume of TS-1 decrease. We also carry out the character-
istic of Si, O and Ti atom on three samples and the results
are shown in Fig. 2 and Fig. S1. The high-resolution XPS
spectra of Si2p displays a shift to low binding energy
(BE) with Au or Fe loading and have the same shift com-
pared to raw TS-1. Similarly, O1s has the same kind shift
of Au or Fe catalysts. It can be implied that Si-rich in
electron is present attributed to the formation of unsatu-
rated coordination structure by etching and also brings
Various metal-supported Au/Ti-containing system cata-
lysts applied for directly propylene epoxidation are presented
in Table S2. According to other reported articles [6, 18],
magnesium, calcium, barium, and germanium modifed cata-
lysts had also enhanced catalytic performance compared to
bulk Au/TS-1. In addition, gold-copper based catalysts no
matter on TiO2 or SiO2 had an infuence on PO formation at
high temperature [33, 34]. All of the above put forward the
comprehension of structure-performance relationships and
further testifed the pivotal rule in various modifed metal
between Au NPs and support for the epoxidation of propyl-
ene. However, all these catalysts have room for the improve-
ment on Au capture efciency and C3H6 conversion. Obvi-
ously, the addition of Fe chelate shows a superior catalytic
activity because more Au sites are present.
3.2 Efect of Precursor Properties on TS‑1 Structure
Considering that unique property of TS-1 structure, it is
reflected that roughening the TS-1 surface increases the
Fig. 1 XRD patterns of TS-1 support, Au/TS-1, and Fe/TS-1
1 3

The Construction of Au–Fe–TS-1 Interface Coupling Structure for Improving Catalytic Performance…
Fig. 2 High-resolution of XPS analysis showing a Si2p and b O1s of TS-1, Au/TS-1 and Fe/TS-1
Table 2 Ti content on TS-1, Au/TS-1 and Fe/TS-1 catalysts
Ti contenta
TS-1
Au/TS-1
8.025
Fe/TS-1
8.151
Ti (ppm)
8.273
aMeasured by ICP-OES
about oxygen vacancies to maintain the charge neutral-
ity [37, 38]. And the Si etching effect of acid and alkali
on TS-1, which is HAuCl4 and Fe chelate precursor, is
similar. The Ti2p on XPS characterization in Fig. S1 is
demonstrated that some of Ti is present as non-framework
structure on acid condition, the peak at 458.6 eV, more
than the other two catalysts. ICP characterization on Ti
content clearly shows the loss of Ti at acid or alkaline
condition compared to bulk TS-1 from Table 2. However,
alkalescent Fe treatment TS-1 makes less etching effect
on Ti atom. It may be speculated that part of Ti atoms
may become non-framework TiO2 via etching effect, a
small number of Ti are in solution. All of them are appre-
ciably clarified that the contraction of TS-1 unit cell is
caused by etching effect on acid and alkaline atmosphere,
directly producing Si, O and Ti vacant defects. And it also
reflects that the loss of framework Ti can be reduced with
Fe treatment.
Fig. 3 XRD patterns of Au–Fex/TS-1 catalysts with diferent Fe/Au
molar ratios: (a) Au/TS-1, (b) Au–Fe0.5/TS-1, (c) Au–Fe0.6/TS-1, (d)
Au–Fe0.7/TS-1, (e) Au–Fe0.8/TS-1, and (f) Au–Fe0.9/TS-1 catalysts.
The Au loading of each catalyst is 1 wt%
are formed for Au–Fex/TS-1. Besides, there are no peaks
attributed to the characteristic of Fe species from XRD
and XPS in Fig. S1. While Fe component is detected from
ICP results, which implied that Fe is contained in Au/
TS-1 catalyst and well dispersion, but it is not on the sur-
face of catalyst, it may be covered by Au nanoparticles.
From Table S3, we can see that the unit cell volume of
Au–Fex/TS-1 decreases at frst and then increases with
more Fe doping, which is all smaller than bulk TS-1 and
Au–Fe0.7/TS-1 has the lowest cell volume. The results can
be explained that Si is etched at the condition of the acid
solution and then Fe3+ is dissociated slowly by H+ to locate
the Si vacant defects on TS-1 structure. The radius of the
Fe3+ ion (64.5 pm) is larger than that of Si4+ (40 pm), the
lattice parameters and unit cell volume may become larger
3.3 Structural and Physicochemical Features of Au–
Fe Catalysts
Figure 3 shows XRD patterns of Au–Fex/TS-1 catalysts
with varying Fe loadings to study the structure composi-
tion and physical properties. No Au difraction peaks are
observed for Au/TS-1 catalyst. However, the (111) and
(200) difraction peaks of Au species at 38.1° and 44.4°
1 3

X. Li et al.
when the Si4+ is replaced by Fe3+ incorporation into the
framework. However, considering that charge neutrality,
oxygen vacancies are produced by charge compensation
[38] and as a result, the volume is eventually decreased.
In addition, more Fe loading may present as clusters and
reduce the substitution of Si with Fe, giving rise to the unit
cell volume increasing a bit.
are shown in Fig. 4 to investigate the morphology and
particle size distribution of catalysts. From Fig. 4a and b,
it can be observed that Au/TS-1 has only several large Au
particles with a mean particle size of 3.7 nm. However,
high and well-dispersed Au particles of 3.1 nm diameter
are displayed over Au–Fe0.7/TS-1. TEM results show that
the introduction of iron as a modifer of Au/TS-1 catalyst
has efectively decreased Au nanoparticle size and pro-
moted the dispersion of Au particles.
As is well-known, the preparation of Au catalysts is
sensitive to the atmosphere and gives rise to the change
of particle size and Au morphology [39, 40], which is
of paramount importance to reaction activity. However,
there is no accurately consensus yet regarding the location
and size of Au NPs, it has been accepted that Au particle
size that desired PO synthesis is smaller than 5 nm for
direct propylene epoxidation. In our work, TEM images
The Ti coordination environment of TS-1 and contain-
ing Fe catalysts are recorded by UV–Vis spectroscopy
(Fig. 5). All the samples have strong adsorption at 220 nm,
indicating the structure of isolated tetrahedral coordinated
Ti in the TS-1 framework is almost not damaged under
acid treatment [41]. It also shows a broad band at 520 nm
Fig. 4 TEM images and diameter distribution of Au particles for a, b Au/TS-1 and c, d Au–Fe0.7/TS-1
1 3

The Construction of Au–Fe–TS-1 Interface Coupling Structure for Improving Catalytic Performance…
Fig. 5 UV–Vis spectra of TS-1 support and Au–Fex/TS-1 catalysts
Fig. 6 FTIR spectra of (a) TS-1, (b) Au/TS-1, (c) Au–Fe0.5/TS-1 (d)
with diferent Fe/Au molar ratio under normalization correction. The
Au–Fe0.6/TS-1, (e) Au–Fe0.7/TS-1, (f) Au–Fe0.8/TS-1, and (g) Au–
Fe0.9/TS-1 catalysts. The Au loading of each catalyst is 1 wt%
Au loading of each catalyst is 1 wt%
for Au–Fex/TS-1 catalyst attributed to the characteristic
surface plasmon resonance (SPR) adsorption of solid Au
NPs [42], and the band area continues to increase as the
Fe/Au on a molar basis increases up to 0.7. However, they
tend to diminish at elevated Fe loading. We infer that it
may form Fe clusters and cannot coordinate with more Au
species, making Au contents decreasing. It is illustrated
that appropriate Fe loading can efectively enhance Au
contents.
under acidic condition with the formation of Fe–O–Si,
which is directly associated with the oxygen vacancies to
keep charge balance [47]. Therefore, the unit cell volume
of TS-1 support shrinks after Fe modifed catalysts, just
as the XRD results.
Both of Au–Fe0.7/TS-1 and Au/TS-1 catalysts have
two typical peaks for Au4f around 83.4 eV and 87.1 eV
in Fig. 7b; however, the former catalysts have stronger
and more distinct peaks with the presence of Fe. They are
attributed to the characteristic of the spin–orbit splitting of
Au core-level spectra for Au4f7/2 and Au4f5/2, respectively,
in line with Au0 chemical state [48]. The sharp strength
of the Au4f band for containing Fe catalyst efectively
illustrates more Au contents on catalyst surface than Au/
TS-1. Moreover, it is observed that the Au4f7/2 peak is
shifted to lower BE for Au–Fe0.7/TS-1 catalysts from 83.6
to 83.4 eV, compared to Au/TS-1. This shift may prove the
existence of coordination relationship between Au and Fe
by electronic efect, giving rise to an increase in Au ions
capture. In conclusion, Fe is inserted into framework Si of
TS-1 and simultaneously realizes the capture of active Au
species. This interface coupling structure of Au–Fe–TS-1
has an enhancement in the Au capture and promotes the
dispersion and stability of Au particles.
In order to further explore vibration modes of Ti and Si
with the presence of Fe, FTIR spectroscopy is analyzed on
TS-1 support and diferent Au–Fex/TS-1 catalysts. Similar
to UV–Vis characterization, the band at 980 cm−1 is fur-
ther clarifed titanium into the zeolite framework in Fig. 6,
which plays a pivotal role in having a reduced rate for H2O2
decomposition [43]. However, the typical difraction band
of Ti–O–Si entities is shown at 970 cm−1. It is probably
due to iron substituting part of silicon, which gives rise to
the vibration form of Fe–O–Si and infuences the absorb-
ance of framework Ti [44, 45]. And the electronegativity of
iron is lower than that of silicon, making electron density
of Ti in Ti–O–Si structure increasing and shifting to higher
wavenumber [46]. All these results are proved that the MFI
framework is not destroyed according to UV–Vis and FTIR
characterizations.
The information on the chemical and electronic state of
active sites on catalyst surface can be investigated by XPS
via using the characteristic binding energy. High-resolu-
tion XPS spectra of Si2p are shown in Fig. 7a, Si2p shifts
to lower BE for Fe modifed catalyst, located from 103.8 to
103.7 eV. This may be caused by the formation of unsatu-
rated coordination of Si with the substitution by Fe. That
is, the dissociated Fe3+ is inserted into Si vacant defects
3.4 Catalyst Stability
It has been reported that Au catalysts are easily subjected to
aggregation and deactivation under reaction conditions as a
function of time [6, 49–51]. In this work, the stability of the
optimum catalyst for Au–Fe0.7/TS-1 at reaction condition
was tested and the results are displayed in Fig. 8. It can be
1 3

X. Li et al.
Fig. 7 High-resolution of XPS analysis showing a Si2p and b Au4f of Au–Fe0.7/TS-1 and Au/TS-1
3.5 The Role of Fe on Structure Performance
Relationship
The catalytic performance in Table 1 shows a volcano-
trend with diferent molar ratios of Fe/Au as a function
of time. The activity of optimum catalyst has achieved
high conversion increasing from 0.75 to 7.3% and has
enhanced PO formation rate on per kilogram of catalysts to
238.8 gPO h−1 kgcat−1. This clearly shows the promotional
efect of Fe. In addition, PO selectivity and H2 efciency
of Au/TS-1 are slightly higher than that of Au–Fe0.7/TS-1
catalyst. As reported in literature, impact factors of both H2
efciency and PO selectivity are consistent in many aspects,
which are Au particle size and the property of support [3].
The decrease for those might be due to the changing prop-
erty of TS-1 support for the Au–Fex/TS-1catalysts prepara-
tion period, giving rise to the etching efect on TS-1 struc-
ture. Moreover, the stability test was carried out and found
no obvious deactivation after reaction for 40 h. It could
be explained that the addition of Fe makes Au nanoparti-
cles well-dispersed, suppressing the deactivation of gold
aggregation.
Fig. 8 Stability tests result for the Au–Fe0.7/TS-1 over 40 h.
(flled triangle) PO formation rate; (flled square) C3H6 conver-
sion; (flled circle) PO selectivity. Reaction conditions: 503 K,
atmosphere pressure, 0.15 g of catalyst, 35 ml min−1 total gas fow
(C3H6:O2:H2:N2 =1:1:1:7)
seen that the catalytic activity can quickly achieve stable at
below 1 h, and the activity keeps at a conversion of 7% and
PO selectivity of 90%. The optimum catalyst is stable for
long-term performance of about 40 h with no signifcant
decline for the propylene conversion and selectivity. The
spent catalyst was further investigated by TEM with a wide
distribution of Au metal particles in the range of 2–6 nm
from Fig. S2, indicating gold NPs migrate gradually and
form larger gold clusters with time on stream after reaction
for 40 h. And the mean diameter of Au particles is 4 nm,
larger than that in a fresh catalyst. However, it has no appar-
ent deactivation which is proved that Fe has a role in resist-
ance to sintering.
From XPS and XRD on lattice parameters results for
diferent precursor solution, etching efect occurs on TS-1
aiming at Si atoms during the preparation. Those two char-
acterizations have no peaks attributed to Fe species, which
are contrary to ICP results. It is suggested that Fe species are
present, which may be covered by Au nanoparticles. Then
XPS shifting results of O and Si atom confrm the existence
of Si–O–Fe coordination bonds, and oxygen vacancy may
be produced by the diferent valence substitution of Fe to
Si to maintain charge neutrality. It can be further confrmed
that Fe ions occupy Si defects and generate the formation
of Fe–O–Si structure, combined with FTIR results. The
XPS spectrum of Au4f shifting to lower energy after the
1 3

The Construction of Au–Fe–TS-1 Interface Coupling Structure for Improving Catalytic Performance…
introduction of Fe reveals that Au in this compound is in a
state of more electrons. It demonstrates the coupling rela-
tionship between Au and Fe, and achieves the enhancement
of Au capture efciency through electronic efect. Moreover,
ICP characterization directly demonstrates more Au contents
on Au/TS-1 when Fe is introduced. This is also certifed by
the peak area of Au species increased from UV–Vis and
XPS spectra. Furthermore, the tendency of PO formation
rate reveals that the amount and property of gold active sites
have changed. It has been reported that reaction pathways
have remarkably attributed to the size of Au particles. Both
experimental and theoretical analyses demonstrate that pro-
pylene epoxidation predominantly takes place with Au size
of 2–5 nm [32, 52]. The smaller Au nanoparticles are more
active toward PO formation because of more active Au sites
present [53]. Feng et al. [19] proposed the catalytic activity
is greatly infuenced via the decrease in the Au nanoparticle
size to 2.7 nm with the synergy between Au and Ag. All of
these articles illustrate the fact that the reactivity efects pro-
duce by the smaller particle size of gold. In this work, TEM
results show that the presence of Fe makes Au particles size
decreased and highly dispersed, preventing from agglomera-
tion to an extent. The role of Au–Fe coordination captures
large number of small Au particles and promotes its disper-
sion and stability by coupling efect. Meanwhile, the nature
of Au has also changed from the shifting of Au4f in XPS. All
of the factors might explain why catalytic activity improves
for Au–Fex/TS-1 catalysts. Surprisingly, we also found that
the loss of framework Ti on catalyst surface is inhibited after
Fe loading from Table S4 and the XPS spectrum of Ti2p in
Fig. S1 further reveals that the content of non-framework Ti
is reduced by Fe modifed. It is concluded that the doping
of Fe replaces parts of Si atoms and produces vacancy sites
on TS-1 supports by ion exchange. It also captures more Au
ions deposited and highly dispersed on support surface by
interfacial regulation, promoting the stability and PO forma-
tion rate over the whole catalysts. In addition, the presence
of Fe is conducive to inhibiting Ti losing at acid atmosphere
and reducing the content of non-framework Ti. All these
results are verifed that the interface coupling structure of
Au–Fe–TS-1 has been constructed by support modifcation
and interfacial regulation.
plays in achieving high catalytic performance [55]. From
our XPS results, the coupling efect of Au and Fe makes
some electrons transferred to the Au surface. This efect can
also increase the number of Au contents and decrease the
Au particle size from TEM. That is, the electron-rich, high
capture efciency and dispersion of small Au nanoparticles
are formed and facilitate activating more oxygen with the
help of Fe. When the oxygen is adsorbed on Au clusters
surface, it will reduce with the captured electron by forming
reactive peroxide. This phenomenon is conducive to reduce
the apparent activation energy for propylene epoxidation,
which is estimated to be 25.84 kJ mol−1 from 49.09 kJ mol−1
(in Fig. S3) and improves the ability of PO formation. Based
on the above discussion, electron-rich states Au with high
dispersion is appeared, which is benefcial for O2 adsorption
in the form of O2− [56]. In summary, it is concluded that the
structure of interface coupling Au–Fe–TS-1 is benefcial to
improve catalytic performance for propylene epoxidation.
4 Conclusions
To sum up, we have successfully enhanced Au capture ef-
ciency and achieved high-dispersed Au nanoparticles by
both support acidic modifcation and interfacial regulation
with the introduction of Fe. The reactivity shows a volcano-
trend with diferent molar ratios of Fe/Au and optimal activ-
ity occurs at the molar ratio of 0.7. And propylene conver-
sion has greatly improved about 9 times and Au capture
efciency has increased from 10 to 69% compared with the
traditional Au/TS-1 catalyst. The modifed catalyst increases
PO formation rate on per Au and reduces apparent activation
energy. Characterizations revealed that unsaturated coordi-
nated Si, oxygen vacancies and the construction of interface
coupling structure of Au–Fe–TS-1 were present. Meanwhile,
Fe as an electron mediator increases the number of Au active
component by electronic efect and decreases Au binding
energy, which may indirectly efect oxygen activation and
produce hydroperoxide. In addition, the structure of frame-
work Ti is reserved and the content of non-framework Ti is
relatively reduced by Fe loading. This report is signifcantly
important to realize anchor metal and high dispersion by
etching other zeolites.
The process and mechanism of gas-phase with H2 and O2
for C3H6 epoxidation have been widely adopted that both
Au0 and isolated tetrahedral Ti(IV) are the active sites. The
formation of intermediate hydroperoxide considered as the
rate-determining step is related to the activation of O2. Oxy-
gen activation and producing hydroperoxide are dependent
on the property of gold active site. The control of Au nano-
particle size and morphology through catalyst preparation
paves one way to enhance the oxygen activation [54]. The
rate of H2O2 synthesis are maintained to a greater extent
over Au-rich catalysts, which highlights the key role Au
Acknowledgements This work was supported by the National Natural
Science Foundation of China (Grant 21571036). The author would
like to thank Dr. Xinqi Zhang from testing center for helping on the
analysis of TEM results.
Compliance with Ethical Standards
Conflict of interest The authors declare that they have no confict of
interest.
1 3

X. Li et al.
27. Grätzel M (2001) Nature 414:338
References
28. Yang T, Tatsuma T (2005) J Am Chem Soc 127:7632
29. Parida KM, Sahu N, Mohapatra P, Scurrell MS (2010) J Mol Catal
A 319:92
1. Khatib SJ, Oyama ST (2015) Catal Rev 57:306
2. Lambert RM, Williams FJ, Cropley RL, Palermo A (2005) J Mol
Catal A 228:27
30. Carrettin S, Hao YL, Aguilar-Guerrero V, Gates BC, Trasobares
S, Calvino JJ, Corma A (2007) Chemistry 13:7771
31. Cumaranatunge L, Delgass WN (2005) J Catal 232:38
32. Feng X, Duan XZ, Qian G, Zhou XG, Chen D, Yuan WK (2014)
J Catal 317:99
3. Huang JH, Haruta M (2012) Res Chem Intermed 38:1
4. Nijhuis TA, Makkee M, Moulijn JA, Weckhuysen BM (2006) Ind
Eng Chem Res 45:3447
5. Huang JH, Takei T, Akita T, Ohashi H, Haruta M (2010) Appl
Catal B 95:430
33. Llorca J, Dominguez M, Ledesma C, Chimentao R, Medina F,
Sueiras J, Angurell I, Seco M, Rossell O (2008) J Catal 258:187
34. Liu XY, Wang AQ, Li L, Zhang T, Mou CY, Lee JF (2011) J Catal
278:288
6. Lu JQ, Zhang XM, Bravo-Suárez JJ, Fujitani T, Oyama ST (2009)
Catal Today 147:186
7. Lee WS, Akatay MC, Stach EA, Ribeiro FH, Delgass NW (2013)
J Catal 308:98
35. Xiong G, Jia QY, Cao YY, Liu LQ, Guo ZD (2017) RSC Adv
7:24046
8. Yu B, Ayvalı T, Raine E, Li T, Li MMJ, Zheng JW, Wu S, Bagabas
AA, Tsang SCE (2019) Appl Catal B 243:304
9. Ghosh S, Acharyya SS, Tiwari R, Sarkar B, Singha RK, Pendem
C, Sasaki T, Bal R (2014) ACS Catal 4:2169
36. Tsai S-T, Chao P-Y, Tsai T-C, Wang I, Liu X, Guo X-W (2009)
Catal Today 148:174
37. Hu Z, Li K, Wu X, Wang N, Li X, Li Q, Li L, Lv K (2019) Appl
Catal B 256:117860
10. Seubsai A, Kahn M, Zohour B, Noon D, Charoenpanich M, Sen-
kan S (2015) Ind Eng Chem Res 54:2638
38. Song P, Qin H, Liu X, Huang S, Zhang R, Hu J, Jiang M (2006)
Sens Actuators B 119:415
11. Li NX, Yang B, Liu M, Chen Y, Zhou JC (2017) Chin J Catal
38:831
39. Zanella R, Louis C, Giorgio S, Touroude R (2004) J Catal 223:328
40. Zanella R, Giorgio S, Henry CR, Louis C (2002) J Phys Chem B
106:7634
12. Hayashi T, Tanaka K, Haruta M (1998) J Catal 178:566
13. Kanungo S, Keshri KS, van Hoof AJF, d’Angelo MFN, Schouten
JC, Nijhuis TA, Hensen EJM, Chowdhury B (2016) J Catal
344:434
41. Alba MD, Luan Z, Klinowski J (1996) J Phys Chem 100:2178
42. Jiang GM, Huang YX, Zhang S, Zhu HY, Wu ZB, Sun SH (2016)
Nanoscale 8:17947
14. Feng X, Sheng N, Liu YB, Chen XB, Chen D, Yang CH, Zhou
XG (2017) ACS Catal 7:2668
43. Notari B (1988) Stud Surf Sci Catal 37:413
44. Mal NK, Ramaswamy V, Ganapathy S, Ramaswamy AV (1995)
Appl Catal A 125:233
15. Li ZS, Wang YN, Zhang JH, Wang DY, Ma WH (2017) Catal
Commun 90:87
45. Vayssilov GN (1997) Catal Rev 39:209
16. Chen D, Zou Y, Wang S (2019) Mater Today Energy 12:250
17. Zhao W, Zhang J, Zhu X, Zhang M, Tang J, Tan M, Wang Y
(2014) Appl Catal B 144:468
46. Wu M, Chou L, Song H (2012) Catal Lett 142:627
47. Wang XH, Li JG, Kamiyama H, Katada M, Ohashi N, Moriyoshi
Y, Ishigaki T (2005) J Am Chem Soc 127:10982
48. Chou J, McFarland EW (2004) Chem Commun 14:1648
49. Sinha AK, Seelan S, Tsubota S, Haruta M (2004) Ang Chem Int
Ed 43:1546
18. Liu T, Hacarlioglu P, Oyama ST, Luo MF, Pan XR, Lu JQ (2009)
J Catal 267:202
19. Feng X, Yang J, Duan XZ, Cao YQ, Chen BX, Chen WY, Lin D,
Qian G, Chen D, Yang CH, Zhou XG (2018) ACS Catal 8:7799
20. Chen GX, Zhao Y, Fu G, Duchesne PN, Gu L, Zheng YP, Weng
XF, Chen MS, Zhang P, Pao CW, Lee JF, Zheng NF (2014) Sci-
ence 344:495
50. Nijhuis TA, Visser T, Weckhuysen BM (2005) Angew Chem Int
Ed 44:1115
51. Uphade BS, Tsubota S, Hayashi T, Haruta M (1998) Chem Lett
27:1277
21. Chen GX, Xu CF, Huang XQ, Ye JY, Gu L, Li G, Tang ZC, Wu
BH, Yang HY, Zhao ZP, Zhou ZY, Fu G, Zheng NF (2016) Nat
Mater 15:564
52. Qi CX, Huang JH, Bao SQ, Su HJ, Akita T, Haruta M (2011) J
Catal 281:12
53. Taylor B, Lauterbach J, Delgass WN (2005) Appl Catal A 291:188
54. Boronat M, Corma A (2010) Dalton Trans 39:8538
55. Santos A, Lewis RJ, Malta G, Howe AG, Morgan DJ (2019) Ind
Eng Chem Res 58:12623
22. Du ST, Chen XX, Sun QM, Wang N, Jia MJ, Valtchevac V, Yu JH
(2016) Chem Commun (Camb) 52:3580
23. Du CL, Cui N, Li LH, Hua ZL, Shi JL (2019) RSC Adv 9:9694
24. Akamatsu K, Ikeda S, Nawafune H, Deki S (2003) Chem Mater
15:2488
56. Linh NH, Nguyen TQ, Diño WA, Kasai H (2015) Surf Sci 633:38
25. Akamatsu K, Shinkai H, Ikeda S, Adachi S, Nawafune H, Tomita
S (2005) J Am Chem Soc 127:7980
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
26. Ikeda S, Yanagimoto H, Akamatsu K, Nawafune H (2007) Adv
Func Mater 17:889
1 3