Design and Synthesis of the Active Site Environment in Zeolite Catalysts for Selectively Manipulating Mechanistic Pathways

Article abstract of DOI:10.1021/jacs.1c04818

By combining kinetics and theoretical calculations, we show here the benefits of going beyond the concept of static localized and defined active sites on solid catalysts, into a system that globally and dynamically considers the active site located in an

Full text of DOI:10.1021/jacs.1c04818

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
pubs.acs.org/JACS
Article
Design and Synthesis of the Active Site Environment in Zeolite
Catalysts for Selectively Manipulating Mechanistic Pathways
Chengeng Li,† Pau Ferri,† Cecilia Paris, Manuel Moliner, Mercedes Boronat,* and Avelino Corma*
Cite This: J. Am. Chem. Soc. 2021, 143, 10718−10726
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: By combining kinetics and theoretical calculations,
we show here the benefits of going beyond the concept of static
localized and defined active sites on solid catalysts, into a system
that globally and dynamically considers the active site located in an
environment that involves a scaffold structure particularly suited
for a target reaction. We demonstrate that such a system is able to
direct the reaction through a preferred mechanism when two of
them are competing. This is illustrated here for an industrially
relevant reaction, the diethylbenzene−benzene transalkylation. The
zeolite catalyst (ITQ-27) optimizes location, density, and environ-
ment of acid sites to drive the reaction through the preselected and
preferred diaryl-mediated mechanism, instead of the alkyl transfer
pathway. This is achieved by minimizing the activation energy of
the selected pathway through weak interactions, much in the way
that it occurs in enzymatic catalysts. We show that ITQ-27 outperforms previously reported zeolites for the DEB-Bz transalkylation
and, more specifically, industrially relevant zeolites such as faujasite, beta, and mordenite.
alkylation steps that proceed through unstable penta-
coordinated carbonium ion intermediates.16,26 The first
dealkylation of DEB produces an EB molecule and a surface
ethoxy group that, in a second step, reacts with benzene and
yields EB. However, the ethoxy group can also react with EB or
DEB molecules present in the zeolite channels to form
undesired overethylated byproducts such as DEB and
triethylbenzene (TEB).27 In addition, the surface ethoxy can
decompose into ethene and regenerate the Brønsted acid site,
further decreasing the yield of EB.
1. INTRODUCTION
Production of alkylaromatics is a crucial process in the modern
chemical industry since they are well-established precursors for
many important intermediates and chemicals.1−3 Ethylbenzene
(EB) is one of the industrial alkylaromatics with a higher
production capacity worldwide, which is mostly consumed in
the manufacture of polystyrene.4 Alkylaromatics are mainly
produced by alkylation of benzene (Bz) with alkyl reagents
such as light olefins and alcohols using acid zeolites as
catalysts.5−7 However, the formation of undesired polyalky-
lated byproducts with lower added value is inevitably occurring
during the industrial alkylation process.8 To improve
selectivity, the current production of EB combines two
reaction processes: alkylation of benzene with ethene, followed
by EB separation and transalkylation of the polyethylated
byproducts with benzene to increase the global yield of EB
(Figure S1 in the Supporting Information).9−11
Some of the catalysts employed in the industrial trans-
alkylation of diethylbenzene (DEB), which is the main
byproduct of benzene ethylation, are large pore zeolites such
as faujasite, beta, MCM-22, and mordenite.9,10,12−14 The
presence of wide channels and/or cavities or hemicavities
within these materials allows the diffusion of the polyethylated
aromatic molecules involved in the transalkylation and the
formation of the bulky diaryl intermediates. Diethylbenzene
transalkylation can proceed through two main reaction
pathways (Scheme 1).15,16,25,17−24 The alkyl transfer mecha-
nism (Scheme 1a) involves consecutive dealkylation and
The second pathway proceeds through cationic diaryl
intermediates formed by reaction of benzene with a
diethylbenzenium carbenium ion (DEB+) (Scheme 1b).16,28
The initial hydride abstraction step necessary to form DEB+
̈
from DEB is difficult on Bronsted acid sites, with high
activation energy barriers of ∼200 kJ/mol,22,25,29 but it has
been proposed that this step is energetically affordable on
Lewis acid sites like La3+ cations15,20 or extra-framework Al
species.16,17 Once the first DEB+ cation is generated, it enters
the true catalytic cycle, at the end of which it is regenerated by
Received: May 11, 2021
Published: July 9, 2021
https://doi.org/10.1021/jacs.1c04818
J. Am. Chem. Soc. 2021, 143, 10718−10726
© 2021 American Chemical Society
10718

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
pubs.acs.org/JACS
Article
directing agents (OSDAs) that mimic intermediates or
transition states of the target reactions. This methodology
includes the following working steps: (i) geometry optimiza-
tion of the transition state or key intermediate of the reaction
considered, (ii) synthesis of an OSDA that mimics the size,
shape, and charge distribution of the transition state or
intermediate, (iii) synthesis of zeolites using this OSDA, and
(iv) catalyst testing of the zeolite or zeolites obtained to check
if they efficiently catalyze the target reaction. Notice that in the
third step of the process, that is, the zeolite synthesis, it is not
possible to predict whether amorphous materials, an already
existing zeolite, or a new zeolite structure will be obtained.
Within our proposal, success does not necessarily imply that
we obtained a new zeolite structure. Success implies that the
zeolite obtained, regardless of if it is new or already known, is
optimal for the specific reaction under study. In the case that
the synthesized zeolite structure already exists, two possibilities
arise: (i) the zeolite has already been applied in academia or
industrially for the target reaction, in which case the validity of
the transition state that mimics methodology to directly
synthesize the optimal catalyst is confirmed, or (ii) the zeolite
structure has never been claimed for the reaction under study,
in which case the impact of the result is important from both
fundamental and industrial points of view.
Scheme 1. Proposed Mechanisms for the Diethylenzene−
Benzene Transalkylation: (a) Alkyl-Transfer and (b) Diaryl-
Mediated Pathways and the Proposed Mimicking OSDA
Here, following the working steps described above, we have
selected an OSDA, diphenyldimethylphosphonium
(DPDMP+), that mimics the size, shape, and charge local-
ization of the diaryl cation intermediate involved in the
transalkylation of DEB and benzene (see Scheme 1). With this
OSDA, a zeolite (ITQ-27) has been synthesized which, as far
as we know, was never reported for this reaction and
outperforms other zeolites currently reported for the DEB-Bz
transalkylation. We show here by means of kinetic and
computational studies that not only the “imprinted” zeolite
structure preferentially stabilizes the transition state involved in
the diaryl-mediated pathway but also it favors, more than the
other zeolites, this pathway with respect to the less desired
alkyl transfer mechanism. All this results in a higher activity
and selectivity and a lower rate of deactivation than the other
zeolites that are used for DEB-Bz transalkylation.
hydrogen transfer between reactants and products. For the
transalkylation to take place, a proton in the cationic I1+
intermediate must be transferred from one aromatic ring to the
other, generating a different I4+ intermediate that, by cracking,
produces EB. The proton transfer converting I1+ into I4+ can
occur intramolecularly through four- or six-membered cationic
transition states (red and blue paths in Scheme 1b) or via
consecutive deprotonation and protonation steps with the
participation of the zeolite acid sites (gray path in Scheme 1b).
The formation of undesired higher-ethylated byproducts
according to the diaryl-mediated pathway is less probable
than in the alkyl transfer mechanism due to steric constraints.
Therefore, in order to increase the yield of EB while decreasing
catalyst deactivation during the transalkylation process, the
participation of the diaryl-mediated mechanism must be
maximized. To do that, we thought that the microporous
structure of an optimized zeolite catalyst for EB transalkylation
should be able to accommodate and stabilize the diaryl cationic
intermediates and transition states.
Starting with this hypothesis, and in analogy with enzymatic
catalysts, we should look for a zeolite structure with an
adequate scaffold for the reaction to take place, while
introducing the active site, a proton in this case, in the
adequate position.30−32 Contrary to what happens in the case
of enzymes, the zeolite structure is not so flexible during the
reaction. Therefore, we must find a zeolite structure whose
pores and cavities match the transition state for the reaction
pathway selected. Indeed, electrostatic and, specifically, weak
interactions between the walls and the transition state will
decrease the activation energy of the transalkylation reaction
through the preferred reaction mechanism, i.e., the diaryl-
mediated pathway.
2. RESULTS AND DISCUSSION
2.1. Synthesis and Characterization. Based on the
above exposed premises, the DPDMP+ mimicking the diaryl
intermediates involved in the DEB-Bz transalkylation mecha-
nism was used as OSDA, and the synthesis resulted in the
selective crystallization of the ITQ-27 zeolite with the IWV
framework (see details in the Supporting Information,
Experimental Section). The achieved material, designated as
IWV-M, shows the characteristic PXRD pattern of the IWV
structure without impurities (Figure S2). Particles between 1
and 2 μm with plate-like morphology were observed by
FESEM (Figure S3), and no extra-framework Al was observed
by solid 27Al MAS NMR (Figure S4). The DEB-Bz
transalkylation activity of IWV-M was tested in a fixed bed
continuous reactor in gas as well as in the liquid phase and
compared with that of catalysts currently employed in industry
such as faujasite (FAU), beta (*BEA), MCM-22 (MWW), and
mordenite (MOR).9,10,12−14
The reference catalysts used include two samples of FAU
with different Si/Al ratio and a crystal size of 400−500 nm
(CBV760 and CBV720), as well as a sample of beta zeolite
(Zeolyst CP814C), with a Si/Al ratio of 16.4 composed by
In order to achieve this, we applied a methodology for the
design of zeolites involving the use of organic structure-
10719
https://doi.org/10.1021/jacs.1c04818
J. Am. Chem. Soc. 2021, 143, 10718−10726

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
pubs.acs.org/JACS
Article
small crystallites with diameter below 100 nm (Table S1 and
Figure S3). MOR and MWW samples with Si/Al ratios of 12.7
and 20.4, respectively, and crystallite sizes of 100 and 200 nm,
respectively, were synthesized following the literature.33,34
2.2. Catalytic Test. The materials were first tested in the
liquid-phase DEB-Bz transalkylation at 250 °C and 3.5 MPa,
with a DEB weight hourly space velocity (WHSVDEB) of 10 h−1
using a feedstock of reactant mixture (Bz/DEB ratio of 3:1 wt/
wt), similar to the industrially employed conditions. With the
catalyst particle size and flows selected, no control by either
external or internal diffusion was observed (see details in the
Supporting Information, Experimental Section). The results in
Figure 1 show that IWV provides the highest activity and
in large pore zeolites. During this induction period, higher
alkylated aromatics are formed inside the zeolite pore, resulting
in a higher proportion of benzene within the products and
therefore in a higher selectivity when reaching the steady
state.28 Notably, the MWW material exhibits a clear decrease
in selectivity along TOS because the formation of the bulkier
products during the induction period probably blocks the
smaller 10-ring channels, leaving only residual activity on the
external surface.
After discovering that IWV is an excellent zeolite structure
for DEB-Bz transalkylation reaction, we attempted to
synthesize the zeolite with a higher Al content to further
increase the catalyst activity. However, we were not successful
in doing that under our synthesis conditions using DPDMP+ as
OSDA. Nevertheless, IWV could be synthesized with a higher
Al content with an imidazolium-based dicationic OSDA
following the procedure described by Davis et al.35 In this
way, an IWV sample with a Si/Al ratio of 13.4 was obtained,
and it was named as IWV-D. For comparison purposes,
CBV720 was employed to compare with IWV-D since its Si/Al
ratio was similar (Si/Al ∼ 13.4, Table S1). The results in
Figure S5 show that, at a WHSV of 10 h−1, the catalysts with a
higher amount of Al reach higher steady-state DEB conversion,
83% and 74% for IWV-D and CBV720, respectively, and a
selectivity toward EB above 95% in both cases. Since the
thermodynamic equilibrium DEB conversion of the DEB-Bz
transalkylation under these conditions approaches 90%, the
samples were further evaluated at a much higher space velocity
(WHSVDEB = 40 h−1) to obtain lower conversion to better
observe catalytic differences. Under these conditions, it is
clearly observed that the zeolites with the IWV structure are
more active and selective that the corresponding FAU samples
with similar Al content and also more active than the
corresponding MWW and MOR zeolites (Figure S5).
2.3. Kinetic Study. To have kinetic insight into the
catalytic performance of the different zeolites, the catalysts
were tested in gas-phase reaction conditions (see Experimental
Section in the Supporting Information) in order to avoid
surface saturation of the zeolite with reactants. Initial reaction
rates were obtained from catalytic tests performed at different
contact times (w/F, where w is the weight of catalysts and F is
the feeding flow of DEB). The results plotted in Figure 2 and
summarized in Table S2 clearly show again that the highest
initial reaction rates are obtained with the catalysts with the
IWV structure (0.36 and 0.41 molEB/(gcat h) for IWV-M and
IWV-D, respectively), closely followed by the FAU sample
with low Si/Al ratio (FAU-CBV720, 0.30 molEB/gcat h). The
MOR and FAU-CBV760 samples exhibit similar activities
(0.11 and 0.09 molEB/(gcat h), respectively). For the sake of
comparison, the initial reaction rate was normalized with
respect to the Al content leading to the following order: IWV-
M > IWV-D > FAU-CBV720 > FAU-CBV760 > MOR (Table
S2), which indicates a strong correlation between the
framework structure and the catalytic activity.
Regarding selectivity, the mechanism depicted in Scheme 1
indicates that, besides the desired transalkylation between DEB
and Bz yielding two molecules of EB, two other primary
processes can take place: DEB dealkylation producing EB and
ethene and disproportionation of two DEB molecules forming
EB and TEB. The plots of the yield to different reaction
products versus DEB conversion in Figure 2 show that the
prevailing reaction is the DEB-Bz transalkylation in all zeolites,
with the yield of EB approaching the stoichiometric value.
Figure 1. DEB conversion (a) and EB selectivity (b) with time on
stream (TOS) for the transalkylation of DEB over zeolite catalysts
with different framework topologies. Reaction conditions: T = 250
°C, P = 3.5 MPa, feeding composition: Bz:DEB = 3:1 (wt). WHSVDEB
= 10 h−1.
selectivity toward ethylbenzene (EB) under these reaction
conditions compared with the different zeolite structures that
have been reported for the transalkylation reaction. Thus, the
steady-state DEB conversion with IWV-M approaches 75%
with an EB selectivity higher than 95%, while the CBV760
catalyst, with the FAU structure and similar framework Si/Al
ratio, achieved a lower steady-state DEB conversion of ∼26−
28% and an EB selectivity of ∼90%.
The other materials tested, MOR and beta containing 12-
ring channels and MWW with 10-ring channels and 12-ring
cavities connected by 10-ring windows and hemicavities at the
external surface, showed lower steady-state DEB conversion
and EB selectivity. The increase in conversion at short time on
stream (TOS) in all catalysts is due to the existence of a well-
known induction period for transalkylation of alkyl aromatics
10720
https://doi.org/10.1021/jacs.1c04818
J. Am. Chem. Soc. 2021, 143, 10718−10726

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 2. EB yield at different contact times (a) and yield of EB (b)
and ethene (c) versus DEB conversion for different zeolites. Reaction
conditions: T = 200−250 °C, P atmospheric, feeding composition:
N2:Bz: DEB = 30:5:1 (molar), w/F = 0.21−3.36 gcat h/molDEB. The
error bars in (a) stand for the standard deviation of at least three
experiments for each data point. Conversion (a) and EB selectivity
(b) with time on stream.
Dealkylation to produce ethene increases on the catalysts with
higher Al content, and its contribution to DEB conversion is
more important in MOR than in other zeolite structures. In all
cases, the yield of TEB is always below 1% and reaches a
minimum in MOR, probably due to spatial restrictions in the
unidimensional channel.22,36
2.4. How the Scaffold Selects the Reaction Mecha-
nism. Following our initial hypothesis, the fact that the IWV
zeolite gives higher intrinsic activity and selectivity should be
related with a better stabilization of the transition states
involved in the diaryl-mediated reaction pathway by zeolite
confinement. To check this, we performed a detailed
theoretical study of the two possible mechanistic routes
presented in Scheme 1 on the IWV and MOR zeolite
structures by means of periodic DFT calculations. Brønsted
acid sites were placed at the two most stable locations for Al in
the IWV structure, T3 and T6 (see Table S3), and at the T4
position accessible from the 12-ring channel in MOR. The
geometries of all minima and transition states involved in the
alkyl-transfer and diaryl-mediated pathways of DEB-Bz trans-
alkylation and DEB dealkylation were fully optimized without
restrictions at the three sites considered, IWV-T3, IWV-T6,
and MOR-T4, and using the most stable isomer of DEB, para-
DEB. The calculated activation and reaction energies for each
elementary step are summarized in Supporting Information
Table S4, and the corresponding energy profiles are plotted in
Figure 3.
Figure 3. DFT energy profiles for (a) alkyl-transfer pathway; (b)
diaryl-mediated pathway, direct H transfer; (c) diaryl-mediated,
multiple-step H transfer; and (d) diaryl-mediated, neutral inter-
mediate.
reactant being lower than 4 kJ mol−1. The alternative ethyl
transfer to the zeolite framework through TS2 requires
activation energies between 90 and 100 kJ mol−1, which
render this pathway highly improbable (see Figure 3 and Table
S4).
In MOR-T4, the activation barrier for DEB protonation is
higher, 84 kJ mol−1, but the relative stability against backward
decomposition of the DEBH+ carbonium ion formed, with a
barrier of 30 kJ mol−1, allows some competitive dealkylation
through TS2. The surface ethoxy group generated in this step
might react with benzene to form a new EB molecule with an
activation energy of 100 kJ mol−1 or might decompose via
transition state TS5, yielding ethene with a similar activation
barrier of 97 kJ mol−1. Indeed, the experimental observation of
some ethene when performing the transalkylation reaction with
According to Scheme 1, the first step in the alkyl-transfer
pathway is the protonation of DEB through transition state
TS1 to form a DEBH+ carbonium ion with a pentacoordinated
C atom (see optimized geometries in Figures 4, S6, and S7).
The calculated activation energies for this step in IWV are not
too high (around 50 kJ mol−1); however, the process is
endothermic, and the DEBH+ cations formed are unstable,
with the barriers for backward decomposition into DEB
10721
https://doi.org/10.1021/jacs.1c04818
J. Am. Chem. Soc. 2021, 143, 10718−10726