Impact of Zeolite Framework Composition and Flexibility on Methanol-To-Olefins Selectivity: Confinement or Diffusion?

Article abstract of DOI:10.1002/anie.202007609

The methanol-to-olefins reaction catalyzed by small-pore cage-based acid zeolites and zeotypes produces a mixture of short chain olefins, whose selectivity to ethene, propene and butene varies with the cavity architecture and with the framework composition. The product distribution of aluminosilicates and silicoaluminophosphates with the CHA and AEI structures (H-SSZ-13, H-SAPO-34, H-SSZ-39 and H-SAPO-18) has been experimentally determined, and the impact of acidity and framework flexibility on the stability of the key cationic intermediates involved in the mechanism and on the diffusion of the olefin products through the 8r windows of the catalysts has been evaluated by means of periodic DFT calculations and ab initio molecular dynamics simulations. The preferential stabilization by confinement of fully methylated hydrocarbon pool intermediates favoring the paring pathway is the main factor controlling the final olefin product distribution.

Full text of DOI:10.1002/anie.202007609

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: Impact of zeolite framework composition and flexibility on MTO
selectivity: confinement or diffusion?
Authors: Avelino Corma, Pau Ferri, Chengeng Li, Reisel Millán,
Joaquín Martínez-Triguero, Manuel Moliner, and Mercedes
Boronat
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202007609
Link to VoR: https://doi.org/10.1002/anie.202007609

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
Impact of zeolite framework composition and flexibility on MTO
selectivity: confinement or diffusion?
Pau Ferri, Chengeng Li, Reisel Millán, Joaquín Martínez-Triguero, Manuel Moliner, Mercedes
Boronat,* Avelino Corma*
P. Ferri, C. Li, R. Millán, Dr. J. Martínez-Triguero, Dr. M. Moliner, Dr. M. Boronat, Dr. A. Corma
Instituto de Tecnología Química
Universitat Politècnica de València - Consejo Superior de Investigaciones Cientificas (UPV-CSIC)
Av. de los Naranjos, s/n, 46022 Valencia, Spain
E-mail: boronat@itq.upv.es, acorma@itq.upv.es
Supporting information for this article is given via a link at the end of the document.
Abstract: The methanol-to-olefins reaction catalyzed by small-pore
cage-based acid zeolites and zeotypes produces a mixture of short
chain olefins, whose selectivity to ethene, propene and butene varies
with the cavity architecture and with the framework composition. The
product distribution of aluminosilicates and silicoaluminophosphates
with the CHA and AEI structures (H-SSZ-13, H-SAPO-34, H-SSZ-39
and H-SAPO-18) has been experimentally determined, and the
impact of acidity and framework flexibility on the stability of the key
cationic intermediates involved in the mechanism and on the diffusion
of the olefin products through the 8r windows of the catalysts has been
evaluated by means of periodic DFT calculations and ab initio
molecular dynamics simulations. The preferential stabilization by
confinement of fully methylated hydrocarbon pool intermediates
favoring the paring pathway is the main factor controlling the final
olefin product distribution.
mechanism, the total amount of ethene, propene and butene
finally obtained depends on the relative contribution of each of the
two alternative pathways proposed to the total conversion of
methanol.
Introduction
The methanol-to-olefins (MTO) reaction is an effective process to
Scheme 1. Paring and side-chain pathways of the aromatics-based
obtain short chain alkenes such as ethene, propene and butene
at an industrial scale.[1,2] The catalysts employed for the MTO
reaction are acid zeolites (aluminosilicates) and SAPOs
hydrocarbon pool mechanism proposed for the MTO reaction. The key gem-
methylated polymethylbenzenium intermediate is framed in red. Based on
reference [
.
12]
(
silicoaluminophosphates) which, besides containing Brönsted
acid centers, are able to host within their inorganic structure
organic species that co-catalyze the reaction.[3–6] In particular,
framework structures containing large internal cavities or cages
connected by small 8-membered rings (8r), such as H-SAPO-34,
are successfully used in commercial plants since 2010.[2,7]
Previous kinetic and isotopic labeling studies proposed that the
selectivity to ethene and propene depends on the degree of
methylation of the aromatic intermediates,[17–19] and a clear
relationship between the dimensions and topology of the zeolite
cavities and the short olefin product distribution in the MTO
reactions was also demonstrated.[4,15,16,20–22] Davis group
introduced recently a structural parameter, the cage-defining ring-
size, to classify fourteen different zeolite structures into four
categories that differ in their product distribution.[22] Going one
step further, we combined DFT calculations with catalyst
synthesis and testing to show that the MTO olefin product
distribution is directly related to the degree of methylation of the
entrapped poly-methyl-benzenium cations, which in turn is
determined by the zeolite cavity architecture.[23,24] Thus, the
paring route is energetically accessible in cavities able to host and
The accepted dual-cycle mechanism for the MTO reaction
assumes the formation, during an initial induction period, of
organic hydrocarbon-pool (HP) species, either alkenes or
aromatics, that remain trapped inside the cavities and participate
in the formation of olefins by successive methylation and cracking
steps.[8–12] The HP species in small-pore cage-based zeolites with
the CHA, AEI, DDR, ITE or RTH structures are aromatic
[
13–
polymethylbenzenes and their corresponding carbenium ions.
1
6]
The transformation of gem-methylated polymethylbenzenium
intermediates, framed in red in Scheme 1, can proceed through
two main competitive pathways. The side-chain pathway includes
exo-methylation, methyl shift, and side-chain elimination steps
and yields predominantly ethene, while the paring route starts with
a ring-contraction step that forms cyclopentenyl cations after
splitting off propene or butene. According to this aromatics-based
stabilize the fully methylated heptamethylbenzenium cation
+
(
7MB ), and consequently the production of propene is enhanced,
while the preferential stabilization of the less substituted
+
pentamethylbenzenium cation (5MB ) leads to the side-chain
1
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
pathway and a higher production of ethene. The ability of each
One factor that could be invoked to explain this behavior is the
lower acid strength of SAPOs as compared to zeolites, that leads
to a lower optimal reaction temperature for the more acidic
zeolites and might also have an impact on the reaction
mechanism.[29,30] The different framework flexibility of SAPOs and
zeolites could also be claimed as the origin of these observations,
+
+
cage to host 7MB or 5MB can be quantitatively described by the
interaction energies between the cations and pure silica models
of different cavities obtained from DFT calculations, and an
interaction energy ratio or Eint(7/5) parameter was established as
descriptor of this relative stabilization. A linear relationship was
found between the C =
/C
2
=
ratios measured for different zeolites
since the diffusion of the larger olefins might be enhanced in the
3
more flexible SAPO catalysts.[
32–35]
However, the fact the MTO
under different reaction conditions and the Eint(7/5) parameter
corresponding to each structure, thus confirming the confinement
effect associated to cage topology as the factor governing the
MTO product selectivity.[24] However, the pure silica cluster
models used in that study did not take into account the possible
effect of acid site concentration or distribution, nor the influence
of framework composition on the MTO reactivity. In this sense, it
is interesting to compare the catalytic behavior of pairs of zeolites
product selectivity varies with framework composition in CHA
structures but remains similar in AEI-type catalysts indicates that
the subject is not simple and requires a detailed study of each
factor, i.e., acidity, framework flexibility and olefin diffusion,
separately. In this work, we use periodic density functional theory
(DFT) to analyze the impact of acid site location and framework
flexibility on the stabilization of the key MTO reaction
intermediates determining the mechanism, and ab initio molecular
dynamics (AIMD) simulations to study the diffusion of ethene and
propene through the 8r windows of zeolites and SAPOs. Catalyst
testing combined with the theoretical information allow to connect
the higher flexibility of SAPO-34 with a better stabilization of 7MB+
and a higher propene production, and lead to a more generalized
(
aluminosilicates) and SAPOs (silicoaluminophosphates) with the
same framework crystallographic structure but different
framework composition, such as for instance H-SSZ-13 and H-
SAPO-34 with the CHA structure, or H-SSZ-39 and H-SAPO-18
with the AEI structure (Figure 1).
correlation between olefin product distribution and
E
int(7/5)
parameter in small-pore cage-based zeolites and zeotypes.
Results and Discussion
Catalyst synthesis, characterization and catalytic activity
tests
First, a systematic evaluation of the catalytic performance of
zeolites and SAPOs with the CHA and AEI structures with
different physico-chemical properties was performed to
unambiguously establish the trends in product distribution. H-
SSZ-13 (CHA) and H-SSZ-39 (AEI) aluminosilicate zeolites, as
well as H-SAPO-34 (CHA) and H-SAPO-18 (AEI)
silicoaluminophosphate zeotypes were synthesized following the
procedures described in detail in the Supporting Information. The
as-synthesized materials show the characteristic PXRD patterns
of CHA and AEI structures (Figure S1), and the textural properties
of the calcined materials as determined by N
experiments are analogous to those reported in the literature
Table S1). The catalyst samples were prepared with different
2
adsorption
Figure 1. Structures of a) CHA and b) AEI frameworks.
(
chemical composition and crystal size (Table S1 and Figures S2-
S3). H-SSZ-13 zeolite was prepared with Si/Al molar ratios of ⁓
Both crystallographic structures are composed by double six-
membered ring units (d6r) that link to form cavities connected by
eight-membered rings (8r) with a similar pore opening of 3.8 Å in
diameter, but the shape of the resulting cavities is clearly different.
The CHA cages are cylindrical whereas AEI presents a basket-
like topology with a wider part (see Figure 1). This particular
16, either as nano-crystals of 60 nm (SSZ-13_2) or as micron-
sized particles of 1 µm (SSZ-13_1), whereas the isostructural H-
SAPO-34 was also obtained as micron-sized crystals of 1 µm
with a Si/TO molar ratio of 0.09 (SAPO-34). H-SSZ-39 was
2
synthesized with a Si/Al molar ratio close to 9 and different crystal
sizes, from nanocrystallites of 60 nm (SSZ-39_2) to larger
crystals of 700 nm (SSZ-39_1). Finally, two samples of H-SAPO-
+
shape of the AEI cavity allows a better stabilization of the 7MB
intermediate that leads to a large production of propene when
either H-SSZ-39 or H-SAPO-18 catalysts are tested in the MTO
reaction. Thus, a product distribution consisting of ⁓50% propene
and similar amounts of ethene and butene, around 20% each,
have been reported for the two AEI-based catalysts
independently of their framework composition.[21,22,24–28] In
contrast, CHA-type catalysts tend to produce more ethene and
much less propene and butene, but there are differences in
product distribution associated to framework composition. H-SSZ-
2
18 were synthesized with Si/TO molar ratios of 0.08 and with
small (150 nm, SAPO-18_1) and large (800 nm, SAPO-18_2)
crystal sizes. The 29Si and Al MAS NMR spectra of the calcined
samples reveal that most of the Si and Al species remain in
tetrahedral coordination within the zeotype and zeolite
frameworks, respectively (See Figure S4).
27
The catalytic performance of these materials in the MTO reaction
was tested at 623 K and 673 K, with a WHSV of 0.8 h-1 (see Table
13 always produces more ethene than propene, 45% and 35%
1
and Figures 2, S5-S15). In agreement with previous work, when
respectively, while the opposite relationship is always found for H-
SAPO-34.[21,22,24,29–31]
zeolites with the same structure and composition are compared
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
=
at 623 K (SSZ-13_1 and SSZ-13_2, SSZ-39_1 and SSZ-39_2),
larger catalysts lifetimes are observed for the samples with
smaller particle size. In contrast, the selectivity to the different
olefins formed, i.e., ethene, propene, and butene, is better
correlated with the framework topology. The two H-SSZ-13
catalysts with different Si/Al ratio and crystal size produce
preferentially ethene, with 46% selectivity at 95% methanol
samples (see Table 1 and Figures 2 and S8-S11). The C =
3 2
/C
olefin ratios follow the order SSZ-13 < SAPO-34 < SSZ-39 ⁓
SAPO-18, and this trend is maintained when the catalysts are
tested at 673 K (Table 1 and Figures S12-S15).
SSZ-13_1 (CHA)
SAPO-34 (CHA)
100
60
50
100
90
80
70
60
50
40
30
20
10
0
60
50
40
30
20
10
0
90
=
=
=
=
conversion, and the C
3
/C
2
and C
4
/C
2
ratios are similar for both
80
70
40
materials, 0.8 and 0.3 (see Table 1). In contrast, the two H-
SSZ-39 catalysts preferentially produce propene, with 46%
selectivity at 95% methanol conversion, and nearly equivalent
60
50
30
40
20
30
20
10
0
10
=
=
amounts of ethene and butene, 20% each, resulting in C
and C /C
4 2
3
/C
2
0
0
100
200
300
400
500
0
200
400
600
800
=
=
Time (min)
Time (min)
ratios of ⁓2.2 and ⁓1.0, respectively (see Table 1).
SSZ-39_2 (AEI)
SAPO-18_2 (AEI)
1
00
60
100
90
80
70
60
50
40
30
20
10
0
60
50
40
30
20
10
0
Moreover, these product distributions and ratios mostly remain
constant during the MTO reaction under the studied conditions
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
5
0
0
4
(
see Figures 2, S5, S6, S8 and S9), even during catalyst
30
20
deactivation when methanol conversion values are below 100%.
10
0
0
200
400
600
800
50
150
250
350
450
550
Time (min)
Time (min)
Table 1. Product selectivity at the same methanol conversion level (X=95%) for
X (%) C1+Paraffins C2= C3= C4= C5+
the different small pore zeolites and zeotypes and catalyst lifetime. Reaction
-
1
conditions: T=623 or 673 K, WHSV=0.8 h , wcat=50 mg.
Figure 2. Methanol conversion and product selectivities (%wt) with TOS using
Time
min)
Selectivity
(%wt)
Ratios
CHA-type (SSZ-13_1 and SAPO-34) and AEI-type (SSZ-39_2 and SAPO-18_2)
(
-1
catalysts (Reaction conditions: T=623 K, WHSV=0.8 h , wcat=50 mg).
=
=
=
4
=/C =
C
4 2
=/C =
Sample
T (K)
X
95
C
2
C
3
C
C
3
2
Periodic DFT study of the stability of reaction intermediates
Framework flexibility. The three-dimensional pore systems of
CHA and AEI crystallographic structures contain small d6r units
that link to form larger cavities accessible through 8r windows
SSZ-13_1
SSZ-13_2
SAPO-34
623
623
623
623
623
623
623
673
673
673
673
260
1085
447
267
480
138
246
670
298
446
471
45.1
47.1
33.6
20.9
22.6
22.9
22.8
56.4
37.9
33.6
34.5
37.0
34.2
45.9
44.4
47.9
47.8
48.4
30.4
41.8
44.5
46.4
12.4
12.1
13.7
19.6
22.0
21.0
18.7
9.2
0.82
0.73
1.40
2.12
2.20
2.09
2.12
0.54
1.10
1.32
1.34
0.27
0.26
0.41
0.94
0.98
0.92
0.82
0.16
0.35
0.42
0.38
(
see Figure 1). The global dimensions or total volume of the
SSZ-39_1
SSZ-39_2
SAPO-18_1
SAPO-18_2
SSZ-13_1
SAPO-34
internal cavities in zeolites SSZ-13 and SSZ-39 are similar (see
Table S2 in the Supporting Information), but their topology is
clearly different. According to IZA database,[ changing the
framework composition from silicate to aluminophosphate results
in an increase of the unit cell volume of 2.1% for CHA and 1.6%
for AEI structures. To confirm the ability of our computational
approach to reproduce these trends the unit cell parameters and
atomic positions of the SSZ-13, AlPO-34, SSZ-39 and AlPO-18
catalyst models were fully relaxed without restrictions and, in
agreement with experiment, the AlPO-34 and AlPO-18 unit cell
volumes obtained from the periodic DFT calculations are,
respectively, 3.4% and 3.1% larger than the corresponding silica
counterparts (see Table S2).
36]
13.4
14.2
13.2
SSZ-39_1
SAPO-18_2
The larger unit cell volume of AlPO materials as compared to
zeolites could help stabilizing the bulkier 7MB intermediate, thus
When catalysts with the same framework topology but different
composition are compared, i.e., H-SSZ-13 with H-SAPO-34 and
H-SSZ-39 with H-SAPO-18, two different situations appear. For
CHA-type catalysts framework composition affects the product
distribution, and the relative concentrations of ethene and
propene obtained with H-SSZ-13 zeolite are reversed in H-SAPO-
+
favoring the paring route and enhancing the formation of propene
and butene, a fact that was experimentally observed in the case
of H-SAPO-34 (see Table 1). To clarify this point, the interaction
+
+
energies of 5MB and 7MB cationic intermediates with neutral
frameworks of SSZ-13, AlPO-34, SSZ-39 and AlPO-18 catalysts
were obtained from periodic DFT calculations. In a first set of
calculations, the atoms of the zeolite or AlPO lattices were kept
fixed to simulate rigid materials, and only the cationic
intermediates were fully optimized without restrictions. In a
second step, all the framework atoms were also allowed to relax
to simulate flexible materials. In all cases, and as expected from
the different acid strength of AlPOs and zeolites (see Table S3
and discussion in the Supporting Information), the calculated
interaction energies are larger in the silicate models than in the
aluminophosphates. But this effect is quite similar for both cations
34 (see Figures 2 and S7). H-SSZ-13 shows higher selectivity to
ethene and less to propene at 623 K (46 and 35%, respectively,
=
=
=
=
see Table 1) with C
3
/C
2
and C
4
/C
2
ratios of 0.8 and 0.26,
whereas, in the case of H-SAPO-34, propene is the most
abundant olefin with 46% selectivity at 623 K, resulting in an
=
=
=
=
increase of the C
3
/C
2
and C
4
/C
2
ratios to 1.4 and 0.4,
respectively (see Table 1). In contrast, in catalysts with the AEI
structure, the composition of the framework does not alter the
product distribution and very similar selectivity values and C
3 2
=/C
=
=
=
4 2
and C /C ratios are obtained for H-SSZ-39 and H-SAPO-18
3
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
and therefore their relative stabilization, measured by the Eint(7/5)
parameter, is not affected by the catalyst composition when we
consider rigid materials (see Table 2). Under these geometry
deformations that the entrapped hydrocarbon pool intermediates
provoke in the four cavities considered have been analyzed in
detail (see Figure 3 and Table 3). According to Davis [22] the cage-
defining ring encircles the center of the cages in cage-based
zeolites, that is, it is the ring enclosing the cationic intermediates
of the MTO reaction, and its size can be defined as the number of
tetrahedral atoms of the ring, 12 for CHA and 16 for AEI. The
cage-defining ring can also be considered as an ellipse whose
area can be more accurately estimated from the a and b distances
obtained from the geometry optimizations (see Figure 3). Empty
cavities in AlPO-34 are initially larger than in SSZ-13 (see Table
+
+
constraints, 5MB is always better stabilized than 7MB because
of its smaller size, and the Eint(7/5) parameters calculated with fixed
framework indicate, on one side, that the two AEI catalysts would
produce more propene than the two CHA materials and, on the
other side, that changing catalyst composition in CHA would have
no effect on product distribution, while SAPO-18 would produce
more propene than SSZ-39. This result clearly contradicts the
experimental observations described above, and suggests that
rigid models are not adequate to describe these systems.
+
+
3), and the presence of entrapped 5MB and 7MB cations leads
to a systematic enlarging of the ellipse area in the flexible
aluminophosphate. In contrast, the dimensions of the cages in the
more rigid SSZ-13 framework can only increase slightly in the
presence of entrapped cations, resulting in a low stabilization of
Table 2. Interaction energies between the key reaction intermediates 5MB+ and
+
7
MB and neutral catalyst models and Eint(7/5) parameters calculated with fixed
and relaxed frameworks.
Fixed framework atoms
int(5MB+) int(7MB+)
+
the bulkier 7MB . Notice that the ellipse area in empty AlPO-34 is
+
Relaxed framework atoms
int(5MB+) int(7MB+)
E
int(7/5)
larger than in SSZ-13 hosting 7MB , which would explain the
different stabilization of this intermediate in the two CHA catalysts.
E
E
E
E
Catalyst
E
int(7/5)
(kJ/mol)
(kJ/mol)
(kJ/mol)
(kJ/mol)
Table 3. Cavity geometry deformation in neutral catalysts due to the presence
of entrapped 5MB+ and 7MB+ cations. The ellipses used to define each cavity
are shown in Figure 3.
SSZ-13
AlPO-34
SSZ-39
AlPO-18
-598
-590
-632
-619
-548
-536
-590
-602
0.91
0.91
0.93
0.97
-607
-598
-649
-623
-569
-565
-640
-619
0.93
0.95
0.99
1.00
Area of empty
Area with
5MB+ (Å )
Area with
7MB+ (Å )
Catalyst
2
2
2
cavity (Å )
SSZ-13
AlPO-34
SSZ-39
AlPO-18
40.6
41.9
47.9
48.2
41.9
42.1
45.6
42.5
41.8
43.2
45.9
43.9
The situation is completely different in the catalysts with the AEI
structure. The area of the ellipse in the empty SSZ-39 and AlPO-
1
8 cavities is much larger than in the empty cavities of SSZ-13
and AlPO-34, and the presence of entrapped 5MB and 7MB+
cations leads to a contraction of the ring to try to increase the
stabilizing interactions between the organic cations and the
framework oxygens. Again, the geometry deformation is easier in
the more flexible AlPO-18 than in the more rigid zeolite SSZ-39,
and the final optimized area of the ellipse containing the cationic
intermediates in the less restricted AlPO-34 and AlPO-18
materials are quite similar. The larger cage-defining ring size of
AEI (16T atoms) as compared to CHA (12T atoms) could explain
the easier deformation of the AEI cavity to stabilize entrapped
organic cations, irrespectively of the framework composition.
It can be concluded form these data that framework flexibility
+
Figure 3. Cage defining ring in a) CHA and b) AEI frameworks.
+
favors the stabilization of 7MB in AlPO-34 as compared to SSZ-
13, but the effect is almost negligible/less important in catalysts
with the AEI structure. The calculated Eint(7/5) parameters follow
When framework flexibility is included in the second set of
calculations, interaction energies increase systematically due to a
better accommodation of the cations inside the cavities, but the
effect is more pronounced for the bulkier 7MB+ intermediate.
Indeed, while Eint values for 5MB increase from 4 to 17 kJ/mol,
the extra stabilization achieved for 7MB+ ranges from 17 to 50
kJ/mol, leading to significant differences in the Eint(7/5) parameters.
The Eint(7/5) values obtained with relaxed framework atoms for the
two AEI catalysts are similar (see Table 2) and larger than those
calculated for CHA materials, for which catalyst composition has
an influence and the Eint(7/5) parameter calculated for AlPO-34 is
larger than for SSZ-13. To understand this effect, the geometry
the order SSZ-13 < AlPO-34 < SSZ-39 ≤ AlPO-18, in good
=
agreement with the C
(see Table 1).
3
=/C
2
olefin ratio experimentally measured
+
Acid site location. The neutral models employed in previous
section are able to capture the confinement effect associated to
framework architecture or cavity topology, but do not contain the
Brönsted acid sites responsible for the catalytic activity of zeolites
and SAPOs. Al atoms in zeolites and Si atoms in SAPOs can
occupy different framework positions around the confined
intermediates, and the different interactions arising might have an
impact or not on product distribution. To clarify this point Brönsted
4
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
stability is the same for 5MB+ and 7MB , so that the Eint(7/5)
parameters are almost independent of heteroatom location.
Interestingly, the Eint(7/5) values obtained using more realistic
active-site containing models (0.92 for H-SSZ-13, 0.96 for H-
SAPO-34, 0.99 for H-SSZ-39 and 1.00 for H-SAPO-18) are nearly
+
acid sites were created in different positions of CHA and AEI
structures by introducing Al and Si in silicates and
aluminophosphates, respectively, thus generating several H-
SSZ-13, H-SAPO-34, H-SSZ-39 and H-SAPO-18 models (see
Figure 4). The interaction of the cationic intermediates with the
negatively charged heteroatom-containing catalyst models was
evaluated and the results are summarized in Table 4.
2
the same as those obtained considering neutral SiO and AlPO
frameworks (see Table 2), confirming that cavity architecture is
the key factor determining the stability of the entrapped
carbocationic intermediates. Indeed, a linear relationship is found
=/C
=
ratios and Eint(7/5) parameters for the zeolites and
between C
3
2
zeotypes considered, either measured in this work (Figure 5a) or
reported in bibliography using other reaction conditions (Figure
5
b).[22]
Figure 4. Acid site location in a) CHA and b) AEI frameworks.
Table 4. Interaction energies (in kJ/mol) between the key reaction intermediates
and active-site containing catalyst models and Eint(7/5) parameters.
E
int(5MB+)
E
int(7MB+)
(kJ/mol)
-515
-519
-527
E
int(7/5
)
T
1
2
3
4
5
1
3
5
1
2
3
4
2
3
(kJ/mol)
-561
SSZ-13
0.92
0.92
0.92
0.91
0.92
0.96
0.97
0.96
0.99
0.99
0.99
0.99
1.00
1.00
-561
-573
-577
-569
Figure 5. Relationship between measured C =
=
3
/C ratio and Eint(7/5) parameter
2
-527
-519
in small-pore cage-based zeolites and zeotypes. Reaction conditions: a)
-1
WHSV=0.8 h , T=673 K (orange) and T=623 K (blue), data from this work, b)
AlPO-34
SSZ-39
-552
-552
-552
-619
-628
-615
-615
-594
-594
-527
-536
-527
-611
-619
-607
-607
-594
-594
-1
[22]
WHSV=1.3 h , T=673 K, data from reference
.
AIMD study of olefin diffusion through 8r windows in CHA
The larger cavity size and higher flexibility of SAPOs as compared
to zeolites might also have an influence on the relative diffusion
rate of the olefin products, mainly ethene and propene. Previous
experimental studies have shown that ethene diffuses faster than
propene in small-pore cage-based zeolites, and that the relative
diffusion rate depends on multiple factors such as temperature,
window size, catalyst composition or coke content.[32,37–39] It has
been reported that the diffusion of ethene in all silica SSZ-13 at
AlPO-18
As a general rule, the heteroatom locations closer to the positive
charge in the cationic intermediates result in a better stabilization
see Figure 4 and Table 4). But taking into account that the
positive charge in 5MB and 7MB is highly delocalized and the
large number of van der Waals interactions/contacts with
framework oxygen atoms contributing to the stabilization by
confinement, the differences in stability associated to Al or Si
position are relatively small, less than 16 kJ/mol in the zeolites
and almost negligible in the SAPOs. Moreover, the order of
(
3
03 K is 80 times faster than that of propene, but the
+
+
ethene/propene diffusivity ratio is reduced to 28 times at 353 K,[
and might further decrease at the MTO reaction temperature, 623
K. In contrast, the diffusion coefficients of ethene and propene in
SAPO-34 are quite similar even at low temperature, with a
reported diffusivity ratio of 1.7 at 323 K.[38] Therefore, if propene
diffusion is preferentially enhanced in the more flexible SAPOs, a
37]
5
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
=/C
=
ratio would be experimentally observed without the
higher C
3
2
need to invoke mechanistic differences.
Molecular dynamics simulations have demonstrated that olefin
diffusion in small-pore cage-based zeolites is a hindered process,
where olefins remain most of time in one cavity and occasionally
hop to a neighboring cavity crossing the 8r window.[33–35] Recently,
Cnudde et al.[35] used enhanced sampling molecular dynamics
simulations to analyze the influence of Brönsted acid sites and
entrapped hydrocarbon pool intermediates on the diffusion of
ethene and propene in H-SAPO-34. It was confirmed that the
activation barriers for ethene diffusion are lower than for propene,
and it was found that the presence of Brönsted acid sites does not
change this trend. However, a direct comparison of ethene and
propene in SSZ-13 and SAPO-34 materials using the same
methodology is still missing, so that the role of olefin diffusion on
the MTO product distribution obtained with CHA-type catalysts is
not established yet. To clarify this point, we have investigated the
diffusion of ethene and propene through the 8r windows of SSZ-
13 and AlPO-34 catalyst models by means of ab initio molecular
dynamics simulations at the reaction temperature (623 K), using
enhanced umbrella sampling techniques.
Figure 7. Snapshots of the local minima (a, c) and transition states (b, d) on the
free energy surface corresponding to ethene (a, b) and propene (c,d) diffusion
through the 8r windows of SSZ-13 rom umbrella sampling AIMD simulations at
623 K.
Figure 6. Free energy profiles for olefin diffusion through the 8r windows of
CHA-type catalyst from umbrella sampling AIMD simulations at 623 K.
Table 5. Gibbs free energy barriers (in kJ/mol) and geometry changes for short
olefin diffusion through the 8r windows of CHA catalyst models from umbrella
sampling AIMD simulations at 623 K.
Figure 8. Expansion of the 8r window area in CHA-type catalysts due to
diffusion of ethene (blue) and propene (red) through SSZ-13 (solid circles) and
AlPO-34 (empty circles) obtained from umbrella sampling AIMD simulations at
SSZ-13
AlPO-34
623 K.
Catalyst
ethene
propene
ethene
propene
Activation barrier (kJ/mol)
49
72
38
59
This trend could be tentatively associated to the higher flexibility
of the AlPO framework that allows an expansion of the 8r window
when olefins are crossing. However, analysis of the average
geometries obtained from the AIMD simulations (see Figure 8)
shows that the 8r windows are already larger in AlPO-34 than in
SSZ-13 in the structures with ethene or propene adsorbed in the
cavities, and in all cases this area increases about 6-7% to
facilitate the crossing of the olefins from one cavity to another one
2
8
8
8
8
r area in empty cell (Å )
r area at ξ = −4 (Å2)
r area at ξ = 0 (Å2)
33.2
34.0
33.0
34.8
5.4
32.9
35.5
7.9
33.9
35.6
5.0
33.9
36.3
7.8
r area expansion (%)
The free energy profiles in Figure 6 are nearly symmetrical, with
the minima being found at ξ = −4 Å and ξ = 4 Å, that is, in the
center of the initial and final cavities, and with the highest energy
corresponding to the olefin placed in the plane of the 8r, at ξ = 0
Å (see Figure 7 and S16). The calculated free energy activation
barriers in AlPO-34, 38 kJ/mol for ethene and 59 kJ/mol for
propene (see Table 5) are in excellent agreement with previously
reported data using a similar methodology.[35] The AIMD barriers
for ethene and propene diffusion in SSZ-13 are 49 kJ/mol and 72
kJ/mol, respectively, that is, both of them ⁓12 kJ/mol larger than
in AlPO-34.
(
see Table 5). The more flexible AlPO-34 framework allows an
2
2
expansion of 2.0 Å and 2.2 Å for ethene and propene diffusion,
respectively, while the more rigid SSZ-13 is able to expand the
ring area 1.9 Å for ethene and 2.3 Å for propene. These data
suggest that the lower diffusion barriers obtained for AlPO-34 are
not due to a higher flexibility, but to the initially/intrinsically larger
2
2
8r windows in this material.
6
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
The AIMD results presented here confirm that propene diffusion
is hindered compared to ethene, and that diffusion of olefins is
faster/easier in AlPO-34 than in SSZ-13. But the difference in the
Experimental information on synthesis, characterization and
testing of catalysts, and computational details are given in the
Supporting Information.
=
=
diffusion barriers for propene and ethene (Gact
3 2
C – Gact C ) has
a similar value of ⁓12 kJ/mol in both materials, indicating that the
higher flexibility of the AlPO framework does not lead to a
preferential enhancement of propene diffusion rate. Therefore,
Acknowledgements
=
=
the higher C
3 2
/C ratio experimentally obtained when H-SAPO-
This work has been supported by the European Union through
ERC-AdG-2014-671093 (SynCatMatch), Spanish Government
through ‘‘Severo Ochoa” (SEV-2016-0683, MINECO), MAT2017-
34 is used as catalyst for the MTO reaction cannot be attributed
to a faster diffusion of propene in this catalyst.
It should be mentioned at this point that additional diffusion
limitations due to formation of bulky carbonaceous species such
as polyaromatics and coke cannot be excluded from these
calculations. However, the AIMD simulations by Cnudde et al.[35]
showed that olefins can only diffuse through empty cavities, and
we recently presented an optimized H-SSZ-13 catalyst with a
controlled Al distribution that avoids the formation of HP
intermediates in all the cavities, thus facilitating the product
diffusion through the empty cages.[31] These considerations,
together with the clear differences in product distribution obtained
from the beginning of the reaction for H-SSZ-13 and H-SAPO-34
8
2288-C2-1-P (AEI/FEDER, UE) and RTI2018-101033-B-I00
(
MCIU/AEI/FEDER, UE), and by Generalitat Valenciana through
AICO/2019/060. The Electron Microscopy Service of the UPV is
acknowledged for their help in sample characterization. Red
Española de Supercomputación (RES) and Servei d’Informàtica
de la Universitat de València (SIUV) are acknowledged for
computational resources and technical support. P. F. and R. M.
thank ITQ for their contracts. C.L. acknowledges China
Scholarship Council (CSC) for a Ph.D fellowship..
Keywords: Ab initio calculations • diffusion • MTO • reaction
mechanism • structure-selectivity relationship
(
see Figure 2 and S5-S15) and the fact that they remain constant
during the whole catalyst lifetime, support our proposal that
selectivity is not mainly controlled by diffusion.[40]
References
Conclusion
[
[
[
1]
2]
3]
G. A. Olah, Angew. Chemie Int. Ed. 2005, 44, 2636–2639.
P. Tian, Y. Wei, M. Ye, Z. Liu, ACS Catal. 2015, 5, 1922–1938.
J. F. Haw, W. Song, D. M. Marcus, J. B. Nicholas, Acc. Chem. Res.
Conclusion
2
003, 36, 317–326.
The main factors that have been proposed to control the olefin
product distribution of the MTO reaction catalyzed by small-pore
cage-based zeolites and zeotypes have been analyzed though a
combination of static DFT calculations, AIMD simulations and
catalytic activity tests. Catalyst samples differing in framework
topology (CHA and AEI) and composition (zeolite and SAPO)
exhibit different selectivity to ethene, propene and butene. The
cavity topology and its ability to preferentially stabilize the fully
[
[
4]
5]
U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T. V. W. Janssens, F.
Joensen, S. Bordiga, K. P. Lillerud, Angew. Chemie Int. Ed. 2012,
51, 5810–5831.
V. Van Speybroeck, K. De Wispelaere, J. Van der Mynsbrugge, M.
Vandichel, K. Hemelsoet, M. Waroquier, Chem. Soc. Rev. 2014, 43,
7326–7357.
[6]
7]
I. Yarulina, A. D. Chowdhury, F. Meirer, B. M. Weckhuysen, J.
Gascon, Nat. Catal. 2018, 1, 398–411.
M. Moliner, C. Martínez, A. Corma, Chem. Mater. 2014, 26, 246–
258.
D. M. McCann, D. Lesthaeghe, P. W. Kletnieks, D. R. Guenther, M.
J. Hayman, V. Van Speybroeck, M. Waroquier, J. F. Haw, Angew.
Chemie Int. Ed. 2008, 47, 5179–5182.
[
+
[8]
[9]
methylated 7MB cations involved in the paring mechanism is the
key factor controlling product distribution, as confirmed by the
=
linear relationship between the experimentally determined C
3
=/C
2
C.-M. Wang, Y.-D. Wang, Z.-K. Xie, Z.-P. Liu, J. Phys. Chem. C
2009, 113, 4584–4591.
ratio and the Eint(7/5), parameter. The strength, amount and
location of the Brönsted acid sites in the catalyst structure has a
minor influence, but the larger volume and framework flexibility of
silicoaluminphosphates is key to explain the selectivity
differences between SSZ-13 and SAPO-34. The accommodation
[
10]
S. Ilias, A. Bhan, ACS Catal. 2013, 3, 18–31.
K. De Wispelaere, K. Hemelsoet, M. Waroquier, V. Van
Speybroeck, J. Catal. 2013, 305, 76–80.
K. Hemelsoet, J. Van der Mynsbrugge, K. De Wispelaere, M.
Waroquier, V. Van Speybroeck, ChemPhysChem 2013, 14, 1526–
1545.
J. Li, Y. Wei, J. Chen, P. Tian, X. Su, S. Xu, Y. Qi, Q. Wang, Y.
Zhou, Y. He, Z. Liu, J. Am. Chem. Soc. 2012, 134, 836–839.
S. Xu, A. Zheng, Y. Wei, J. Chen, J. Li, Y. Chu, M. Zhang, Q. Wang,
Y. Zhou, J. Wang, F. Deng, Z. Liu, Angew. Chemie Int. Ed. 2013,
[11]
[
12]
+
[13]
[14]
of the bulky 7MB cations in the cavities of CHA-type catalysts
requires an expansion of the 12T rings enclosing them. This
deformation is more energetically demanding in the more rigid
52, 11564–11568.
+
zeolite, and consequently the final stabilization of 7MB and the
[
[
15]
16]
J. Li, Y. Wei, J. Chen, S. Xu, P. Tian, X. Yang, B. Li, J. Wang, Z.
Liu, ACS Catal. 2015, 5, 661–665.
W. Zhang, J. Chen, S. Xu, Y. Chu, Y. Wei, Y. Zhi, J. Huang, A.
Zheng, X. Wu, X. Meng, F. Xiao, F. Deng, Z. Liu, ACS Catal. 2018,
8, 10950–10963.
related Eint(7/5) parameter are larger in the more flexible SAPO-34
than in SSZ-13 zeolite. In contrast, the relative diffusion rate of
ethene and propene through 8r windows is not modified by
framework flexibility. Both olefins diffuse faster through the larger
rings of SAPO-34, indicating that selectivity is not controlled by
diffusion.
[
[
17]
18]
W. Song, H. Fu, J. F. Haw, J. Am. Chem. Soc. 2001, 123, 4749–
4754.
S. Svelle, U. Olsbye, F. Joensen, M. Bjørgen, J. Phys. Chem. C
2007, 111, 17981–17984.
A. Hwang, B. A. Johnson, A. Bhan, J. Catal. 2019, 369, 86–94.
Y. Bhawe, M. Moliner-Marin, J. D. Lunn, Y. Liu, A. Malek, M. Davis,
ACS Catal. 2012, 2, 2490–2495.
[
[
19]
20]
[
[
[
21]
22]
23]
J. H. Kang, R. Walter, D. Xie, T. Davis, C.-Y. Chen, M. E. Davis, S.
I. Zones, ChemPhysChem 2018, 19, 412–419.
J. H. Kang, F. H. Alshafei, S. I. Zones, M. E. Davis, ACS Catal.
Experimental Section
2019, 9, 6012–6019.
C. Li, C. Paris, J. Martínez-Triguero, M. Boronat, M. Moliner, A.
Corma, Nat. Catal. 2018, 1, 547–554.
7
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
[
[
[
[
[
[
[
[
24]
25]
26]
27]
28]
29]
30]
31]
P. Ferri, C. Li, C. Paris, A. Vidal-Moya, M. Moliner, M. Boronat, A.
Corma, ACS Catal. 2019, 9, 11542–11551.
J. Chen, J. Li, C. Yuan, S. Xu, Y. Wei, Q. Wang, Y. Zhou, J. Wang,
M. Zhang, Y. He, S. Xu, Z. Liu, Catal. Sci. Technol. 2014, 4, 3268.
M. Dusselier, M. A. Deimund, J. E. Schmidt, M. E. Davis, ACS
Catal. 2015, 5, 6078–6085.
N. Martín, Z. Li, J. Martínez-Triguero, J. Yu, M. Moliner, A. Corma,
Chem. Commun. 2016, 52, 6072–6075.
R. Martínez-Franco, Z. Li, J. Martínez-Triguero, M. Moliner, A.
Corma, Catal. Sci. Technol. 2016, 6, 2796–2806.
F. Bleken, M. Bjørgen, L. Palumbo, S. Bordiga, S. Svelle, K.-P.
Lillerud, U. Olsbye, Top. Catal. 2009, 52, 218–228.
C.-M. Wang, Y.-D. Wang, Y.-J. Du, G. Yang, Z.-K. Xie, Catal. Sci.
Technol. 2015, 5, 4354–4364.
E. M. Gallego, C. Li, C. Paris, N. Martín, J. Martínez-Triguero, M.
Boronat, M. Moliner, A. Corma, Chem. - A Eur. J. 2018, 24, 14631–
1
4635.
D. Chen, K. Moljord, A. Holmen, Microporous Mesoporous Mater.
012, 164, 239–250.
C. Wang, B. Li, Y. Wang, Z. Xie, J. Energy Chem. 2013, 22, 914–
18.
[
[
[
32]
33]
34]
2
9
A. Ghysels, S. L. C. Moors, K. Hemelsoet, K. De Wispelaere, M.
Waroquier, G. Sastre, V. Van Speybroeck, J. Phys. Chem. C 2015,
119, 23721–23734.
[
[
35]
36]
P. Cnudde, R. Demuynck, S. Vandenbrande, M. Waroquier, G.
Sastre, V. Van Speybroeck, J. Am. Chem. Soc. 2020, 142, 6007–
6017.
“Structure Commission of the International Zeolite Association (IZA-
SC), Database of Zeolite structures,” can be found under
http://www.iza-structure.org/databases/, n.d.
[
[
[
[
37]
38]
39]
40]
D. H. Olson, M. A. Camblor, L. A. Villaescusa, G. H. Kuehl,
Microporous Mesoporous Mater. 2004, 67, 27–33.
D. M. Ruthven, S. C. Reyes, Microporous Mesoporous Mater. 2007,
104, 59–66.
N. Hedin, G. J. DeMartin, W. J. Roth, K. G. Strohmaier, S. C.
Reyes, Microporous Mesoporous Mater. 2008, 109, 327–334.
Z. Li, J. Martínez-Triguero, P. Concepción, J. Yu, A. Corma, Phys.
Chem. Chem. Phys. 2013, 15, 14670.
8
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202007609
RESEARCH ARTICLE
Entry for the Table of Contents
The higher flexibility of the SAPO-34 framework as compared to SSZ-13 does not modify the relative diffusion rate of ethene and
propene, but facilitates the accommodation of the bulkier cationic intermediates involved in the paring route of the MTO reaction, thus
increasing the production of propene and butene.
9
This article is protected by copyright. All rights reserved.