Entrapped Single Tungstate Site in Zeolite for Cooperative Catalysis of Olefin Metathesis with Br?nsted Acid Site

Article abstract of DOI:10.1021/jacs.8b03012

Industrial olefin metathesis catalysts generally suffer from low reaction rates and require harsh reaction conditions for moderate activities. This is due to their inability to prevent metathesis active sites (MASs) from aggregation and their intrinsic poor adsorption and activation of olefin molecules. Here, isolated tungstate species as single molecular MASs are immobilized inside zeolite pores by Br?nsted acid sites (BASs) on the inner surface. It is demonstrated that unoccupied BASs in atomic proximity to MASs enhance olefin adsorption and facilitate the formation of metallocycle intermediates in a stereospecific manner. Thus, effective cooperative catalysis takes place over the BAS-MAS pair inside the zeolite cavity. In consequence, for the cross-metathesis of ethene and trans-2-butene to propene, under mild reaction conditions, the propene production rate over WO_x/USY is ca. 7300 times that over the industrial WO₃/SiO₂-based catalyst. A propene yield up to 79% (80% selectivity) without observable deactivation was obtained over WO_x/USY for a wide range of reaction conditions.

Full text of DOI:10.1021/jacs.8b03012

Subscriber access provided by UNIV OF DURHAM
Entrapped Single Tungstate Site in Zeolite for Cooperative
Catalysis of Olefin Metathesis with Brønsted Acid Site
Pu Zhao, Lin Ye, Zhenyu Sun, Benedict T.W. Lo, Harry Woodcock, Chen
Huang, Chiu Tang, Angus I. Kirkland, Donghai Mei, and Shik Chi Edman Tsang
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b03012 • Publication Date (Web): 16 Apr 2018
Downloaded from http://pubs.acs.org on April 16, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Entrapped Single Tungstate Site in Zeolite for Cooperative Catalysis
of Olefin Metathesis with Brønsted Acid Site
Pu Zhao1†, Lin Ye1†, Zhenyu Sun1‡, Benedict T. W. Lo1, Harry Woodcock1, Chen Huang2, Chiu
Tang3, Angus I. Kirkland2,3, Donghai Mei4 and Shik Chi Edman Tsang1*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1 Wolfson Catalysis Centre, Department of Chemistry, University of Oxford, Oxford, OX1 3QR, UK.
2 Department of Materials, University of Oxford, Oxford, OX1 3PH, UK.
3 Diamond Light Source Ltd, Harwell Science and Innovation Campus, Didcot, OX11 0DE, UK.
4 Physical and Computational Sciences Directorate & Institute for Integrated Catalysis, Pacific Northwest National
Laboratory, Richland, WA, PO Box 999, USA.
ABSTRACT: Industrial olefin metathesis catalysts generally suffer from low reaction rates and require harsh reaction
conditions for moderate activities. This is due to their inability to prevent metathesis active sites (MAS) from aggregation
and their intrinsic poor adsorption and activation of olefin molecules. Here, isolated tungstate species as single molecular
MAS is immobilized inside zeolite pores by Brønsted acid sites (BAS) on the inner surface. It is demonstrated that
unoccupied BAS in atomic proximity to MAS enhance olefin adsorption and facilitate the formation of metallocycle in-
termediates in a stereospecific manner. Thus, effective cooperative catalysis takes place over the BAS-MAS pair inside the
zeolite cavity. In consequence, for the cross-metathesis of ethene and trans-2-butene to propene, under mild reaction
conditions, the propene production rate over WOx/USY is ca. 7,300 times that over the industrial WO3/SiO2 based cata-
lyst. A propene yield up to 79% (80% selectivity) without observable deactivation was obtained over WOx/USY for a wide
range of reaction conditions.
1. INTRODUCTION
becomes clear when one studies the fundamental olefin
metathesis reaction mechanism. It is generally accepted
that olefin metathesis proceeds via metallocycle interme-
diates in the rate-determining steps (Scheme 1).1,7,8 The
efficient formation of metallocycle intermediates requires
fast adsorption of olefin molecules and appropriate
alignment between the C=C bond of adsorbed olefin mol-
ecules and the W=O/W=C bond of metathesis active sites
(MAS) for [2+2] cycloaddition.7,9,10 This may be difficult to
achieve with a catalyst that contains only isolated MAS.
There is an indication from previous studies that the in-
troduction of acidic sites to olefin metathesis catalysts can
increase the rate of olefin metathesis reaction.11 Thus, en-
hanced activity is obtained when supported WO3 or MoO3
is blended with zeolites.12,13 However, the ill-defined inter-
face between bulk acidic zeolite and MAS do not lead to
significant rate enhancement, indicating the cooperative
catalysis is not yet optimized.
Olefins are essential feedstocks for the production of
many petrochemicals, oleochemicals, polymers and spe-
cialty chemicals. Catalytic olefin metathesis is an im-
portant reaction in the chemical industry. For example,
the cross-metathesis of ethene and 2-butene to propene
and the reverse reaction play a significant role in buffer-
ing the global supply of ethene and propene. Over the
past fifty years, a number of metal oxides including Re2O7,
MoO3 and WO3 supported on SiO2 and/or Al2O3 have
been found to be effective as olefin metathesis catalysts
under continuous flow conditions.1 For various reasons,
an inexpensive and robust WO3/SiO2 based material is the
only widely-employed catalyst in the industry for the
cross-metathesis of ethene and 2-butene to propene and
the reverse reaction. High temperature (300−550oC) and
high pressure (> 20 bar) are required for moderate activi-
ties to meet the industrial requirement.1,2 Active site isola-
tion shifting from WO3 nanocrystals to surface polymeric
WOx and ultimately to isolated WOx was found to be
promising in improving catalytic activity.3 Considerable
efforts have been made on immobilizing isolated organ-
ometallic complexes on SiO2 and/or Al2O3 for high activi-
ties despite the fact that they tend to aggregate on open
surfaces.4-6 Nonetheless, the industrial application of the-
se catalysts requires further development. Another way to
improve the activity of industrial WO3/SiO2 based catalyst
Thus we designed a new type of bifunctional olefin me-
tathesis catalysts by placing an isolated tungstate as WOx
MAS in a zeolite cavity where unoccupied Brønsted acid
sites (BAS) are in atomic proximity for efficient coopera-
tive catalysis. As a result, we show for the first time that,
at 200oC and 1 bar, the propene production rate over
WOx/USY is ca. 7,300 times that over the industrial
WO3/SiO2 based catalyst. A propene yield up to 79% with
80% selectivity without observable deactivation can be
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
achieved over WOx/USY for a wide range of reaction con-
length) was loaded into the reactor with quartz wool
packed at both ends. The catalyst was pretreated with
nitrogen (N2, 16 ml·min-1) at 550oC for 1 h. After the reac-
tor cooled down to the reaction temperature, the reac-
1
2
3
4
5
6
7
8
ditions. The structure of WOx/USY catalyst was elucidat-
ed by synchrotron X-ray powder diffraction (SXRD) and
Rietveld refinement, supported by aberration-corrected
scanning transmission electron microscopy (STEM), ex-
tended X-ray absorption fine structure (EXAFS) spectros-
copy, and so on. These characterization results and densi-
ty functional theory (DFT) calculation suggest that olefin
adsorption can be enhanced by unoccupied BAS and the
formation of metallocycle intermediates over isolated
WOx MAS can be greatly facilitated by the nearby unoc-
cupied BAS through the appropriate alignment between
the C=C bond of adsorbed olefin molecules and the
W=O/W=C bond of WOx MAS for [2+2] cycloaddition.
This accounts for the exceptional olefin metathesis rate
observed.
=
=
tants, 4% C2 /N2 and 2% trans-2-C4 /N2, were introduced
(105 ml·min-1 each, weight hourly space velocity (WHSV)
= 0.27 h-1, volume hourly space velocity (VHSV) = 26−35 h-
1). The flow rates were controlled by Brooks mass flow
controllers. The reaction temperature was controlled by a
custom-made furnace. The reaction pressure was con-
trolled by a Swagelok back pressure regulator. The prod-
ucts were analyzed online by a Shimadzu GC-2014 gas
chromatograph equipped with an SGE BP-1 column (100%
dimethylpolysiloxane, 30 m length, 0.53 mm inner diame-
ter, 5.0 µm film thickness) and a flame ionization detec-
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
=
tor. The trans-2-C4 conversion, propene (C3=) selectivity
and carbon balance were calculated as follows:
Conversion(trans-2-C=4ꢀ
ꢂꢃtrans-2-C=4 in feedꢄ ꢅ ꢂꢆtrans-2-C=4 remainingꢀ
ꢁ
ꢂꢃtrans-2-C=4 in feedꢄ
ꢂꢆC=3 ꢀ
Selectivity(C=3 ꢀ ꢁ
ꢂꢆall productsꢀ
ꢆ ꢀ
ꢂ C in products
Scheme 1. The reaction mechanism of the cross-
metathesis of ethene and trans-2-butene to propene
over isolated WOx MAS.1,7,8 (1, 2) Pseudo-Wittig initia-
tion mechanism for W=CHCH3 formation and (3−6) the
following metallacyclobutane catalytic cycle to form
propene (Chauvin reaction mechanism).
Carbon balance ꢁ
ꢆ ꢀ
ꢂ C in feed
=
Yield(C3=) = Conversion(trans-2-C4 ) × Selectivity(C3=)
× Carbon balance
High-resolution STEM. High-resolution electron mi-
croscopy was carried out using an aberration-corrected
JEOL ARM200F. The microscope was operated at 200 kV
in high-angle annular dark-field (HAADF)-STEM mode
with HAADF detector collection semi-angles from 73 to
236 mrad. The convergence semi-angle was 31.48 mrad
with probe array sizes of 1024 by 1024 and a dwell time of
0.8 μs per pixel. The probe current was 11 pA.
2. EXPERIMENTS
Catalyst preparation. WO3/SiO2 was commercially
available. WOx/zeolite catalysts (WOx/USY, WOx/ZSM-5,
WOx/SAPO-11, WOx/SAPO-34 and WOx/β) were prepared
by wet impregnation. USY was obtained from Tosoh.
ZSM-5 and β were provided by Sinopec. SAPO-11 and
SAPO-34 were purchased from ACS Material. The chemi-
cal composition of zeolite supports was analyzed by in-
ductively coupled plasma atomic emission spectroscopy
(ICP-AES) or obtained from the literature (Table S1). All
zeolites have crystal sizes less than 1 μm. Other chemicals
were purchased from Sigma-Aldrich. Typically, 583 mg
ammonia metatungstate hydrate ((NH4)6H2W12O40·xH2O)
was dissolved in 20 mL deionized water before 5 g zeolite
was added to the solution. After 1 h magnetic stirring, the
dispersion was dried in an 80oC oven overnight. The re-
sulting powder was ground and calcinated at 550oC for 2 h
in air. The heating rate was 1oC·min-1. The W loading of
WOx/zeolite catalysts was confirmed to be the same as
that of WO3/SiO2 by ICP-AES (Table S2). Na+-poisoned
WOx/USY (WOx/Na-USY) was prepared by a similar wet
impregnation method. Typically, 177 mg sodium nitrate
(NaNO3) was used for 5 g WOx/USY.
EXAFS and XANES. The local environment of W atoms
in WOx/USY and WOx/ZSM-5 was probed by analyzing
their k3-weighted W L3-edge EXAFS spectra collected at
beamline B18, Diamond Light Source, UK. The energy
resolution (ΔE/E) was 2×10-4. Fluorescence mode was
used. To ascertain data reproducibility, at least two data
sets were collected for each sample. W metal was used as
a reference to avoid energy shift of the measured EXAFS
spectra. Data analysis was performed using IFEFFIT 1 with
Horae packages 2 (Athena and Artemis). The amplitude
parameter was set as a fixed input parameter to allow the
refinement of the coordination number of W. This pa-
rameter was obtained from the analysis of the EXAFS
spectrum of W metal. First-shell data analysis was per-
formed under the assumption of single scattering with the
errors estimated by an R-factor. The oxidation states of W
in WOx/USY samples were evaluated by analyzing their
k3-weighted W L3-edge X-ray absorption near edge struc-
ture (XANES) spectra. These XANES spectra were com-
pared with that of (NH4)6H2W12O40·xH2O in which the
oxidation state of W is +6.
Catalytic activity evaluation. The reaction between
=
=
ethene (C2 ) and trans-2-butene (trans-2-C4 ) (n(C2=)/
=
n(trans-2-C4 ) = 2/1) was carried out in a stainless steel
continuous flow reactor (3/8 in inner diameter and 15 in
length). In a typical experiment, 2.1 g catalyst (15−20 cm
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Catalytic activities of WOx/zeolite catalysts. (A) Trans-2-butene conversion, propene selectivity, propene yield and
carbon balance over WOx/zeolite catalysts compared to those over WO3/SiO2 at T = 400oC, p = 1 and (B) 20 bar. Time on steam
(TOS) = 1 h. (C) Propene yield over WO3/SiO2, WOx/USY and WOx/Na-USY at T = 100−550oC, p = 1 bar and (D) at T = 400oC, p =
1 and 20 bar. TOS = 0.25 h. Carbon balance is above 95% in all experiments. (E) Propene yield over WO3/SiO2, [WO3/SiO2]+USY,
WOx/[SiO2+USY] and WOx/USY at T = 400oC, p = 1 bar. SiO2 to USY mass ratio = 19/1. TOS = 1 h. Carbon balance is above 95% in
all experiments. (F) Arrhenius plots for WO3/SiO2 and WOx/USY at T =100−500oC, p = 1 bar. (G) Dependencies of propene pro-
duction rate on the partial pressure of trans-2-butene over WO3/SiO2 and WOx/USY at T = 450oC, p = 1 bar. α: order of reaction.
SXRD. High-resolution SXRD data were collected at
beamline I11, Diamond Light Source, UK. The energy of
the incident X-ray was 15 keV. The wavelength and the 2θ
zero point were determined by fitting the diffraction data
of high-quality silicon powder (SRM640c). SXRD data
were collected for a WOx/USY sample, a WOx/ZSM-5
sample, a WOx/Na-USY sample, and a trans-2-butene-
adsorbed WOx/USY (t2be-WOx/USY) sample. Before data
collection, WOx/USY and WOx/ZSM-5 was treated under
vacuum at 300oC for 2 h. Adsorbed water should have
been removed completely. The adsorption of trans-2-
butene (2% in N2) on WOx/USY was performed at room
temperature using a Schlenk line. The sample was treated
under vacuum at 300oC for 2 h beforehand. Adsorbed wa-
ter should have been removed completely. Liquid N2 was
used to quench the adsorption of trans-2-butene.
method17,18 with a kinetic cutoff energy of 500 eV. The
exchange-correlation functional was the Perdew-Burke-
Ernzerhof (PBE) functional.19 Brillouin zone sampling was
restricted to Γ point. The energy and force convergence
criteria were 1.0 × 10-6 eV and 0.02 eV·Å-1 in all structure
optimizations. The van der Waals dispersion interaction
between the adsorbate and the zeolite framework was
included using DFT-D2 scheme.20 Transition states of
elementary steps in the reaction pathways were located
using the climbing image nudged elastic band (CI-NEB)
method21,22 and the dimer method23.
For BAS-catalyzed reaction, the confinement and steric
hindrance strongly affect the stabilities of reaction inter-
mediates and transition states.24 To account for important
entropic contribution (ꢀS) and zero-point vibrational
energy (ꢀZPVE) corrections, both Gibbs free energy (ꢀG)
and enthalpy (ꢀH) changes along reaction pathways were
calculated as follows:25,26
All samples were loaded into borosilicate glass capillar-
ies (0.7 mm inner diameter) in a glove box. Glass wool
was packed on top of the sample. SXRD data were collect-
ed in a Debye-Scherrer geometry using multi-analyzer
crystals (MAC) detectors in the 2θ range of 0–150o with
0.001o data binning. Each data set was collected for 1 h for
good statistics.
∆G = ∆H - T∆S
∆H = ∆Utrans + ∆Urot + ∆(ZPVE) + ∆Helec
∆S = ∆Strans + ∆Srot + ∆Svib
where translational, rotational, and vibrational contribu-
tions to internal energy (ꢀU) and entropy (ꢀS) were cal-
culated using standard statistical thermodynamic meth-
od.27 The electronic term (ꢀHelec) was derived from DFT
calculation. The vibrational frequencies of the molecules
of interest were calculated in the framework of harmonic
oscillator approximation with a displacement of 0.01 Å.
DFT calculation. All periodic calculations were per-
formed using spin-polarized DFT within the generalized
gradient approximation (GGA) as implemented in the
Vienna ab initio simulation package (VASP).14-16 The core
and valence electrons were represented by the plane-wave
basis set using the projector augmented wave (PAW)
ACS Paragon Plus Environment

Journal of the American Chemical Society
For the calculation of vibrational frequencies, only mobile
Page 4 of 9
Isolated WOx MAS. The isolation of WOx MAS in USY
pores was confirmed by high-resolution STEM. Isolated
W-containing species can be clearly identified in the
HAADF-STEM image of WOx/USY (Figs. 2A, 2B and S11).
They appear to locate inside USY pores. USY framework
underwent significant radiation damage during the image
acquisition process due to the well-known electron beam
sensitivity of zeolites.31,32 Thus it is acceptable that part of
the W-containing species aggregated and formed small
clusters on the particle surface with increased electron
beam exposure and consequent damage. Isolated WOx
MAS was subsequently revealed to be a discrete molecular
dioxo tungstate (O=)2W(−O)2 (WO4 MAS) by the best
fitting of the k3-weighted W L3-edge EXAFS spectra of
WOx/USY (Figs. 2C, S12 and S13, Tables 1 and S4).
1
2
3
4
5
6
7
8
reaction intermediates over WO4 MAS as well as first- and
second-shell neighboring atoms at BAS were considered.
All other atoms of USY framework were fixed. In the es-
timation of vibrational entropy contribution to Gibbs free
energy, the spurious imaginary and low-lying vibrational
frequencies (<50 cm-1) that were inevitably obtained in
the vibrational frequency calculation were replaced by a
normal mode of 50 cm-1.28-30
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. RESULTS AND DISCUSSION
Catalytic activities. Laboratory characterization re-
sults indicate that WOx MAS are well dispersed over all
WOx/zeolite catalysts except for WOx/SAPO-34 (Figs.
S1−4, Table S3). Figure 1A shows that the catalytic activi-
ties of all WOx/zeolite catalysts are superior to that of
WO3/SiO2 at 1 bar, which can be indicative of a synergy
between WOx MAS and the acidic sites in zeolites. Figure
1B shows that the catalytic activity of WO3/SiO2 can be
significantly improved at 20 bar. This could be due to the
increased rate of olefin adsorption over WOx MAS under
higher applied pressure. It is interesting that the catalytic
properties of all WOx/zeolite catalysts remain similar at
the two different applied pressures, indicating that olefin
adsorption could have readily taken place at a lower pres-
sure in the presence of acidic sites. The difference in the
catalyst acidity and zeolite structure can be seen as the
primary cause of the variation in the catalytic activities of
WOx/zeolite catalysts. WOx/ZSM-5, WOx/SAPO-11 and
WOx/β with high acidic strength and/or large acid quanti-
ty (Fig. S5) can give high trans-2-butene conversions but
low propene selectivities and poor carbon balances. These
materials favor the formation of heavy hydrocarbons and
carbonaceous deposition due to their strong acidity and
characteristic porous structures (Figs. S6−9). Less effec-
tive WOx dispersion is found in WOx/SAPO-34, as the
smaller SAPO-34 pores inhibit the immigration of WOx
during synthesis. This accounts for the lower catalytic
activity of WOx/SAPO-34.
Figure 2. Isolated WOx MAS. (A) HAADF-STEM image of
WOx/USY. The typical USY pores along zone axes [1 1 0] are
shown to be filled with atomic W-containing species (red
arrows). The lower left inset shows an enlarged area where
the W-containing species is clearly visible. (B) Crystallo-
graphic model of WOx/USY from Rietveld refinement (also
see Figure 3). The locations of W are in good agreement with
those found in the STEM image. (C) k3-weighted W L3-edge
EXAFS spectra of WOx/USY. k-range for fitting is 3−12 Å-1. R-
range for fitting is 1−2 Å.
Among all the catalysts studied, WOx/USY exhibits the
highest catalytic activity. As shown in Figs. 1C and 1D,
with WO3/SiO2 as a reference, WOx/USY achieves signifi-
cantly and consistently higher trans-2-butene conversions
(85−92%), propene selectivities (78−88%) and propene
yields (70−79%) over a wide range of reaction conditions
(T = 200−400oC, p = 1−20 bar). WOx/USY also shows good
stability for at least 20 hours in our academic laboratory
(Fig. S10). Under mild conditions (T = 200oC, p = 1 bar),
the propene production rate over WOx/USY (3.5
mmol·gcat.-1·h-1) is ca. 7,300 times that over the industrial
WO3/SiO2 based catalyst (4.8×10-4 mmol·gcat.-1·h-1) (See SI
for rate calculation and number of active sites). We at-
tribute this unprecedented high reaction rate over
WOx/USY to the isolation of WOx MAS in zeolite pores
and the placement of WOx MAS in atomic proximity to
BAS for highly efficient cooperative catalysis of olefin me-
tathesis.
Table 1. Structural data for WOx MAS in WOx/USY from
its best fitted k3-weighted W L3-edge EXAFS spectra.
Scattering path
R = 0.3%, Enot = 4.1
Bond
Coordination Debye-
length (Å)
number
Waller
factor
W−O1
W−O2
W−O3
1.75(1)
1.90(2)
2.12(1)
2.2(1)
0.8(1)
1.1(2)
0.003(1)
0.003(2)
0.006(2)
It is anticipated that WO4 MAS inside USY pores may
interact with BAS nearby. This is illustrated by comparing
the catalytic activities of WOx/USY and WOx/Na-USY
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
(Figs. 1C and 1D). In the absence of BAS (H+ replaced by
lite cage due to the polarization effect of EFAl on O4.38
These strong BAS (sBAS) are fully taken up by WO4 MAS
for immobilization according to their resolved site occu-
pancies (sBAS: 2.1 per unit cell (u.c.), WO4 MAS: 2.8 per
u.c.). There are remaining weak BAS (wBAS: 5.1 per u.c.)
in close proximity to the immobilized WO4 MAS in USY
cavity. The unoccupied wBAS in WOx/USY can be selec-
tively replaced by Na+. This is confirmed by the
crystallographic position and resolved site occupancy (5.0
per u.c.) of Na+ in WOx/Na-USY (Fig. 3E, see SI for de-
tailed Rietveld refinement procedure and Fig. S15 for data
fitting). As described previously, the inhibition of wBAS
results in a dramatic decrease in catalytic activity. This
reflects the important role of unoccupied BAS in atomic
proximity to WO4 MAS for enhanced reaction rates.
Na+), WOx/Na-USY clearly delivers inferior catalytic activ-
ity to WOx/USY despite its catalytic activity is still superi-
or to that of WO3/SiO2. This supports our hypothesis that
BAS in zeolite work in synergy with WO4 MAS to enhance
the overall catalytic activity. As shown in Fig. 1E, the cata-
lytic activity of our bifunctional catalyst is critically de-
pendent on the interface between WO4 MAS and BAS.
The physical mixture of WO3/SiO2 and USY ([WO3/SiO2]+
USY) and the WOx-deposited mixture of SiO2 and USY
(WOx/[SiO2+USY]) have an ill-defined interface between
WO4 MAS and BAS. Thus, they are far less active than
WOx/USY in the cross-metathesis of ethene and trans-2-
butene to propene. The high catalytic activity of
WOx/USY appears to be due to the effective interaction(s)
between WO4 MAS and BAS inside zeolite.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
WO4 MAS and BAS in atomic proximity. To confirm
that WO4 MAS and BAS are in atomic proximity in USY
pores for high catalytic activity, SXRD combined with
Rietveld refinement was employed to elucidate the spatial
relationships between WO4 MAS and BAS in WOx/USY. It
is difficult to visualize local 3-dimensional atomic interac-
tion(s) by traditional spectroscopic techniques, e.g. UV-
Vis, Raman, infrared, electron paramagnetic resonance
and X-ray absorption spectroscopy, which are commonly
used for olefin metathesis catalyst characterization.3,11,33,34
This information is also essential when developing an
understanding of the cooperative catalysis within a con-
fined zeolite cavity.35-37
The structural details of unmodified USY along with
the strength, quantity, and location of BAS have been
reported in our previous work.38 BAS (H+) locates on O4
of USY framework with various strength. We determined
the location of WO4 MAS in USY by refining the crystal
structure of WOx/USY using Rietveld method in TOPAS-
Academic 539 based on its high-resolution SXRD data. See
SI for general Rietveld refinement procedure. The rigid
bodies describing WO4 MAS were built based on EXAFS
results and were added into USY framework one by one,
following the indication from difference Fourier map.
Weighted-profile R-factor (Rwp) and goodness-of-fit
(GOF) were used to gauge the quality of Rietveld refine-
ment. After the first WO4 rigid body was added, Rwp and
GOF were found to decrease, indicating the addition of
WO4 rigid body was appropriate. The site occupancy of
the second WO4 rigid body added turned to be almost
zero. Thus, the existence of the second WO4 rigid body
was not considered. As a result, one WO4 rigid body was
used to fit the SXRD data (Figs. 3A and S14). The resulting
crystallographic model of WOx/USY (Fig. 2B) matches
with its STEM images (Figs. 2A and S11).
Figure 3. Structure of WOx/USY. SXRD data of (A)
WOx/USY and (B) trans-2-butene-adsorbed WOx/USY, fitted
by Rietveld refinement. Data in the 2θ range of 20–60o are
zoomed in to illustrate the quality of Rietveld refinement.
(C) Crystallographic model of WOx/USY from Rietveld re-
finement. Although H is invisible in SXRD, O4 of USY, where
BAS (H+) locate, are shown in red. (D) WO4 MAS immobi-
lized through hydrogen bonding to (H)O4 of USY. Dummy
H+ is added for illustration. (E) The crystallographic position
of Na+ in WOx/Na-USY (replacing H+). (F) Trans-2-butene
adsorbed on WOx/USY showing significant interaction(s)
between its double bond, unoccupied BAS, and W=O of WO4
MAS to form the stereospecific adsorption geometry shown.
It is shown that the isolated WO4 MAS is trapped in
USY cavity by W−O···(H)O4 hydrogen bonding
(OWO4···O4 = 2.73(1) Å) on the inner surface (Figs. 3C and
3D). DFT calculation indicates the interaction energy be-
tween WO4 MAS and (H)O4 (BAS) is -141 kJ·mol-1 (See
DFT calculation section and SI for details). The BAS di-
rectly bound to WO4 MAS has been demonstrated to be
reinforced by the extra-framework Al3+ (EFAl) in the soda-
We also found a similar entrapment and site isolation
of WO4 MAS in WOx/ZSM-5 (see SI for detailed explana-
tion, Figs. S16-18 and Table S5) and unoccupied BAS were
also shown to be present. We expect that acid quantity,
distribution, and strength will significantly affect the ad-
ACS Paragon Plus Environment

Journal of the American Chemical Society
sorption kinetics of olefin molecules, hence influencing
Page 6 of 9
and 2 extra-framework Al3+ were added in the sodalite
cage. The resulting negatively charged USY framework
was compensated by adding 7 H+ at the O2- directly con-
nected to Al3+ (framework) and adding 2[(O=)2W(−O)2]2-
(WO4 MAS) at random positions. These added H+ are
active BAS of USY. Structure optimization shows that
WO4 MAS inside USY are most likely bound to H−O4 of
the framework (Fig. S20). This is in line with the results
from Rietveld refinement.
1
2
3
4
5
6
7
8
the overall olefin metathesis rate over nearby WO4 MAS.
Strong unoccupied BAS in zeolites with stereospecific
arrangements may also catalyze side reactions such as
isomerization, undesirable cracking, heavy hydrocarbon
formation and/or carbon deposition, leading to poor pro-
pene selectivity and rapid catalyst deactivation, as in the
case of WOx/ZSM-5. The weak unoccupied BAS in
WOx/USY guarantees high olefin metathesis catalytic
activity and stability of the catalyst without significant
side products and extensive carbon formation. In addi-
tion, the weak unoccupied BAS in WOx/USY appear to be
capable of catalyzing 1-butene (present in industrial
reformate gases) to form 2-butene which participates in
the cross-metathesis reaction with ethene to form
propene. As a result, ca. 55% propene yield with ca. 95%
1-butene conversion can be achieved over WOx/USY at
400oC and 1 bar. This reflects the highly cooperative ad-
sorption-metathesis mechanism in this well-designed
catalyst.
9
We first calculated the adsorption and conversion of
ethene and trans-2-butene over WO4 MAS, in order to
determine which of the two, W=CH2 and W=CHCH3, is
the intermediate to initiate the metallacyclobutane cata-
lytic cycle. As shown in Fig. S21, the formation of the
W=CHCH3 from trans-2-butene is thermodynamically
more favored due to its lower Gibbs reaction energies. In
particular, the formation of W=CHCH3 in the presence of
an unoccupied neighboring BAS is kinetically feasible.
Therefore W=CHCH3 with unoccupied BAS nearby was
used as the intermediate to initiate the metallacyclobu-
tane catalytic cycle.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cooperative catalysis over the BAS-MAS pair.
Effective cooperative catalysis over the BAS-MAS pair is
envisaged to be due to the appropriate alignment be-
tween the C=C bond of olefin molecules adsorbed over
BAS and the W=O bond of WO4 MAS. To confirm this
hypothesis, trans-2-butene was loaded in WO4/USY for
the study of its adsorption geometry by SXRD and
Rietveld refinement (Fig. 3B, see SI for detailed Rietveld
refinement procedure and Fig. S19 for data fitting). The
trans-2-butene molecule was found to be adsorbed over
the unoccupied BAS with an O4(H)···C=
distance of
t2be
4.14(2) Å (t2be: trans-2-butene) (Fig. 3F). This distance is
expected for a significant interaction between the elec-
tron-rich C=C bond of trans-2-butene and the unoccupied
BAS (H+). This could easily lead to the protonation of
trans-2-butene at elevated temperature.40
Perhaps, the most intriguing observation is that the
C=C bond of trans-2-butene appears to have aligned with
the W=O bond of nearby WO4 MAS at room temperature
Figure 4. Energy diagram of the metallacyclobutane
catalytic cycle over W carbene (W=CHCH3) in WOx/USY.
Numbers in black and blue are DFT calculated Gibbs free
energies in the presence and absence of an unoccupied BAS
near W=CHCH3, respectively. The optimized structures in
the presence of an unoccupied neighboring BAS are shown.
with a noticeable interaction (WWO4···C=
= 2.86(1) Å).
t2be
The C=C bond of trans-2-butene and the W=O bond of
WO4 MAS show a distorted square arrangement, which
resembles the anticipated oxametallocycle intermediate
formed in the initiation step (Pseudo-Wittig initiation
mechanism, Scheme 1). The oxametallocycle intermediate
will lead to the formation of a reactive W carbene
(W=CHCH3) to start the metallacyclobutane catalytic
cycle for catalytic olefin metathesis (Chauvin reaction
mechanism, Scheme 1). In the metallacyclobutane catalyt-
ic cycle, it is envisaged that BAS facilitates the adsorption
and alignment of ethene to the nearby W carbene with
Figure 4 shows the energy diagram of the metallacyclo-
butane catalytic cycle over W=CHCH3 in the pres-
ence/absence of an unoccupied BAS nearby. The metal-
lacyclobutane catalytic cycle is initialized by ethene being
adsorbed at W=CHCH3 to form a W cyclobutane inter-
mediate. This is followed by the concerted C−C and W−C
bond scissions leading to the formation of the first
propene. This step is exothermic with a reaction energy
(ꢀG) of -21 kJ·mol-1 in the absence of BAS. The calculated
Gibbs energy of activation (ꢀGǂ) is 94 kJ·mol-1. This is
comparable to the experimental apparent activation ener-
gy (Ea) obtained for WO3/SiO2 which does not have BAS
close to its MAS (108.6 kJ·mol-1, Fig. 1F). This kinetically
relevant step in the first half of the metallacyclobutane
catalytic cycle is significantly promoted when an unoccu-
stereospecificity in
a
similar manner to form
a
metallocyclobutane intermediate before propene is pro-
duced.
DFT calculation. DFT calculation was carried out to
study the metallacyclobutane catalytic cycle. USY model
(T192O384, T = Si or Al) was built based on our previous
work.38 It has a lattice parameter of 24.2680 Å and 7 BAS.
9 Si4+ at specific framework T sites were replaced by Al3+
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
pied neighboring BAS is present. A dramatically low ꢀGǂ
(62 kJ·mol-1) was obtained and the concerted C−C and
W−C bond scissions are more exothermic (ꢀG = -103
kJ·mol-1). The cooperation between the active W=CHCH3
and the unoccupied neighboring BAS is shown to play a
key role in increasing olefin metathesis reaction rate, de-
spite that the experimental Ea for WOx/USY (29.9 kJ·mol-1,
Fig. 1F) cannot allow a direct comparison with the activa-
tion barrier obtained from DFT calculation. The value of
experimental Ea is indicative of diffusion limitation which
leads to a complex rate-temperature relationship.41 Upon
releasing the first propene, a reactive W=CH2 is formed
and is ready for the adsorption of trans-2-butene to form
a W cyclobutane intermediate subsequently. It is found
that the adsorption of trans-2-butene and the formation
of the W cyclobutane intermediate are affected by the
presence of the unoccupied neighboring BAS. More im-
portantly, the calculated ꢀGǂ of the second concerted C−C
and W−C bond scissions in the presence of the unoccu-
pied neighboring BAS (53 kJ·mol-1) is much lower than
that without BAS present (86 kJ·mol-1).
tion, and suggested future work supplied as Supporting In-
formation. This material is available free of charge via the
Internet at http://pubs.acs.org.
1
2
3
4
5
6
7
8
AUTHOR INFORMATION
Corresponding Author
* edman.tsang@chem.ox.ac.uk
Author Contributions
† contributed equally as co-first author.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Present Addresses
‡ College of Chemical Engineering, Beijing University of
Chemical Technology, China.
ACKNOWLEDGMENT
The STEM, EXAFS and SXRD facilities provided by Diamond
Light Source (UK) are acknowledged.
REFERENCES
(1) Lwin, S.; Wachs, I. E., ACS Catal. 2014, 4 (8), 2505-2520.
(2) Mol, J. C., J. Mol. Catal. A: Chem. 2004, 213 (1), 39-45.
(3) Lwin, S.; Li, Y.; Frenkel, A. I.; Wachs, I. E., ACS Catal. 2016,
6 (5), 3061-3071.
(4) Copéret, C.; Comas-Vives, A.; Conley, M. P.; Estes, D. P.;
Fedorov, A.; Mougel, V.; Nagae, H.; Núñez-Zarur, F.; Zhizhko, P.
A., Chem. Rev. 2016, 116 (2), 323-421.
(5) Mazoyer, E.; Szeto, K. C.; Merle, N.; Norsic, S.; Boyron, O.;
Basset, J.-M.; Taoufik, M.; Nicholas, C. P., J. Catal. 2013, 301, 1-7.
(6) Balcar, H.; Čejka, J., Coord. Chem. Rev. 2013, 257 (21–22),
3107-3124.
(7) Grubbs, R. H.; Wenzel, A. G.; O'Leary, D. J.; Khosravi, E.
Handbook of Metathesis, Wiley‐VCH, Weinheim, 2015.
(8) Hérisson, J.-L.; Chauvin, Y., Makromol. Chem. 1971, 141 (1),
161-176.
The order of reaction with respect to trans-2-butene in
excessive ethene was evaluated for WO3/SiO2 (1.8 ± 0.1)
and WOx/USY (1.3 ± 0.1), respectively (Fig. 1G). The ap-
parent higher order of reaction with respect to trans-2-
butene in the case of WO3/SiO2 again suggests WO3/SiO2
has a stronger dependence on the availability of adsorbed
olefin. This reflects the higher activation energy process
of surface trans-2-butene adsorption is rather difficult to
proceed in WO3/SiO2. In contrast, as demonstrated, the
presence of unoccupied BAS in USY makes a significant
contribution to increasing the availability of trans-2-
butene (with a lower Ea and a lower order of reaction, Fig.
1F and Fig. 1G) and ethene to WO4 MAS in the atomic
vicinity. Thus both experimental and theoretical evalua-
tions suggest that the cooperative catalysis between BAS
and MAS lead to an overall enhanced catalytic activity.
(9) Cheng, Z.; Lo, C. S., ACS Catal. 2015, 5 (1), 59-72.
(10) Cheng, Z.; Lo, C. S., ACS Catal. 2012, 2 (3), 341-349.
(11) Amakawa, K.; Wrabetz, S.; Kröhnert, J.; Tzolova-Müller,
G.; Schlögl, R.; Trunschke, A., J. Am. Chem. Soc. 2012, 134 (28),
11462-11473.
(12) Huang, S.; Liu, S.; Xin, W.; Bai, J.; Xie, S.; Wang, Q.; Xu, L.,
J. Mol. Catal. A: Chem. 2005, 226 (1), 61-68.
(13) Li, X.; Zhang, W.; Liu, S.; Xu, L.; Han, X.; Bao, X., J. Catal.
2007, 250 (1), 55-66.
(14) Kresse, G.; Furthmuller, J., Comput. Mater. Sci. 1996, 6 (1),
15-50.
(15) Kresse, G.; Furthmuller, J., Phys. Rev. B 1996, 54 (16), 11169-
11186.
(16) Kresse, G.; Hafner, J., Phys. Rev. B 1994, 49 (20), 14251-
14269.
(17) Blochl, P. E., Phys. Rev. B 1994, 50 (24), 17953-17979.
(18) Kresse, G.; Joubert, D., Phys. Rev. B 1999, 59 (3), 1758-1775.
(19) Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.;
Pederson, M. R.; Singh, D. J.; Fiolhais, C., Phys. Rev. B 1992, 46
(11), 6671-6687.
(20) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H., J. Chem.
Phys. 2010, 132 (15), 154104.
(21) Henkelman, G.; Uberuaga, B. P.; Jonsson, H., J. Chem.
Phys. 2000, 113 (22), 9901-9904.
4. CONCLUSIONS
We have designed a new type of bifunctional olefin
metathesis catalysts by immobilizing isolated single WO4
MAS in a zeolite cavity where unoccupied BAS are in
atomic proximity. Within the confined zeolite cavity, we
demonstrate that BAS facilitate the adsorption and activa-
tion of olefin molecules for the metathesis reaction over
WO4 MAS. This leads to a dramatically enhanced olefin
metathesis reaction rate. We report exceptional propene
production rate and thermal stability of WOx/USY under
industrially applicable conditions. In general, the concept
of cooperative catalysis presented here, using two active
sites of different functionalities within atomic proximity
in a molecular cavity, can provide new opportunities for
designing synergistically adsorptive and catalytic materi-
als.
(22) Mills, G.; Jonsson, H.; Schenter, G. K., Surf. Sci. 1995, 324
(2-3), 305-337.
(23) Henkelman, G.; Jónsson, H., J. Chem. Phys. 1999, 111 (15),
7010-7022. 35.
ASSOCIATED CONTENT
Supporting Information. Methods and additional results of
catalyst testing, structure refinement, physical characteriza-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
(24) Gounder, R.; Iglesia, E., Acc. Chem. Res. 2012, 45 (2), 229-
238.
1
(25) Mei, D.; Lercher, J. A., AIChE J. 2017, 63 (1), 172-184.
(26) Zhi, Y.; Shi, H.; Mu, L.; Liu, Y.; Mei, D.; Camaioni, D. M.;
Lercher, J. A., J. Am. Chem. Soc. 2015, 137 (50), 15781-15794.
(27) Psofogiannakis, G.; St-Amant, A.; Ternan, M., J. Phys.
Chem. B 2006, 110 (48), 24593-24605.
(28) Alexopoulos, K.; Lee, M. S.; Liu, Y.; Zhi, Y. C.; Liu, Y. S.;
Reyniers, M. F.; Marin, G. B.; Glezakou, V. A.; Rousseau, R.;
Lercher, J. A., J. Phys. Chem. C 2016, 120 (13), 7172-7182.
(29) Piccini, G.; Alessio, M.; Sauer, J.; Zhi, Y.; Liu, Y.; Kolven-
bach, R.; Jentys, A.; Lercher, J. A., J. Phys. Chem. C 2015, 119 (11),
6128-6137.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(30) Piccini, G.; Sauer, J., J. Chem. Theory Comput. 2014, 10 (6),
2479-2487.
(31) Pan, M.; Crozier, P. A., Ultramicroscopy 1993, 48 (3), 332-
340.
(32) Terasaki, O.; Ohsuna, T.; Ohnishi, N.; Hiraga, K., Curr.
Opin. Solid State Mater. Sci. 1997, 2 (1), 94-100.
(33) Ross-Medgaarden, E. I.; Wachs, I. E., J. Phys. Chem. C
2007, 111 (41), 15089-15099.
(34) Basrur, A. G.; Patwardhan, S. R.; Was, S. N., J. Catal. 1991,
127 (1), 86-95.
(35) Ye, L.; Song, Q.; Lo, B. T. W.; Zheng, J.; Kong, D.; Murray,
C. A.; Tang, C. C.; Tsang, S. C. E., Angew. Chem. Int. Ed. 2017, 56
(36), 10711-10716.
(36) Teixeira, I. F.; Lo, B. T. W.; Kostetskyy, P.; Stamatakis, M.;
Ye, L.; Tang, C. C.; Mpourmpakis, G.; Tsang, S. C. E., Angew.
Chem. Int. Ed. 2016, 55 (42), 13061-13066.
(37) Lo, B. T. W.; Ye, L.; Qu, J.; Sun, J.; Zheng, J.; Kong, D.;
Murray, C. A.; Tang, C. C.; Tsang, S. C. E., Angew. Chem. In. Ed.
2016, 55 (20), 5981-5984.
(38) Ye, L.; Teixeira, I.; Lo, B. T. W.; Zhao, P.; Tsang, S. C. E.,
Chem. Commun. 2017, 53 (70), 9725-9728.
(39) Coelho, A. A. TOPAS-Academic, 4.1; Coelho Software,
Brisbane.: 2007.
(40) Nguyen, C. M.; De Moor, B. A.; Reyniers, M.-F.; Marin, G.
B., J. Phys. Chem. C 2011, 115 (48), 23831-23847.
(41) Bond, G. C., Heterogeneous Catalysis: Principles and Ap-
plications. 2nd ed.; Oxford University Presss: 1987.
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
1
2
3
Table of Contents artwork
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment