Chemical Imaging of the Binder-Dependent Coke Formation in Zeolite-Based Catalyst Bodies During the Transalkylation of Aromatics

Article abstract of DOI:10.1002/cctc.201900777

The choice of binder material, added to a zeolite-based catalyst body, can significantly influence the catalyst performance during a reaction, i. e. its deactivation and selectivity. In this work the influence of the binder in catalyst extrudates on the formation of hydrocarbon deposits was explored during the transalkylation of toluene with 1,2,4-trimethylbenzene (1,2,4-TMB). Using in situ UV-vis micro-spectroscopy and ex situ confocal fluorescence microscopy approach, coke species were revealed to predominantly form on the rim of zeolite crystals within Al₂O₃-bound extrudates. It was found that this was due to Al migration between the zeolite crystals and the Al₂O₃-binder creating additional acid sites near the zeolite external surface. In contrast, minimal isomerization of 1,2,4-TMB in the SiO₂-bound extrudate allowed greater access to the zeolite internal pore network, creating a more homogeneous coke distribution throughout the zeolite crystals.

Full text of DOI:10.1002/cctc.201900777

Accepted Article
Title: Chemical Imaging of the Binder-dependent Coke Formation
in Zeolite-based Catalyst Bodies during the Transalkylation of
Aromatics
Authors: Suzanna P. Verkleij, Gareth T. Whiting, Denise Pieper, Sonia
Parres Esclapez, Shiwen Li, Machteld M. Mertens, Marcel
Janssen, Anton-Jan Bons, Martijn Burgers, and Bert Marc
Weckhuysen
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: ChemCatChem 10.1002/cctc.201900777
Link to VoR: http://dx.doi.org/10.1002/cctc.201900777
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
Chemical Imaging of the Binder-dependent Coke Formation in
Zeolite-based Catalyst Bodies during the Transalkylation of
Aromatics
[
A]
[A]
[A]
[B]
[B]
Suzanna P. Verkleij , Gareth T. Whiting* , Denise Pieper , Sonia Parres Esclapez , Shiwen Li , Machteld M.
[
B]
[B]
[B]
[B]
[A]
Mertens , Marcel Janssen , Anton-Jan Bons , Martijn Burgers , Bert M. Weckhuysen*
The choice of binder material, added to a zeolite-based catalyst
body, can significantlyinfluence the catalyst performance during
a reaction, i.e. its deactivation and selectivity. In this work, the
influence of the binder in catalyst extrudates on the formation of
hydrocarbon deposits was explored during the transalkylation of
toluene with 1,2,4-trimethylbenzene. Using in situ UV-vis micro-
spectroscopy and ex situ confocal fluorescence microscopy
approach, coke species were revealed to predominantly form on
reactions, for example 1,2,4-TMB can isomerize into the more
bulky 1,2,3-TMB and 1,3,5-TMB.
For the performance of the zeolite catalyst during the
transalkylation reaction, pore structure and acidity are two
important factors. Firstly, the pores need to be large enough for
the diffusion of the bulky TMB species. Previous studies have
shown that 12-membered rings have a higher catalytic activity
than 10-membered rings due to their enhanced diffusion capacity.
the rim of zeolite crystals within Al
found that this was due to Al migration between the zeolite
crystals and the Al -binder creating additional acid sites near
the zeolite external surface. In contrast, minimal isomerization of
,2,4-TMB in the SiO -bound extrudate allowed greater access to
2 3
O -bound extrudates. It was
[
13–15]
However, it was found that medium-pore zeolites, like ZSM-
O
2 3
5, have a higher selectivity towards p-xylene.[16,17] Secondly, the
reaction pathway taking place depends on the acid site strength.
Previously, it has been shown that transalkylation and
disproportionation mainly occur on strong acid sites, while
isomerization can take place on weak acid sites.[18] Due to this
isomerization, the selectivity towards p-xylene decreases.[
Finally, it has been reported that the more bulky 1,3,5-TMB and
1
2
the zeolite internal pore network, creating a more homogeneous
coke distribution throughout the zeolite crystals.
19,20]
1
,2,3-TMB, formed by the isomerization of 1,2,4-TMB, diffuse
Introduction
slower through medium-pore zeolite due to steric constrains and
will mainly react at the surface of the zeolite.[
21–23]
However, direct
Xylene is an important petrochemical compound that has many
applications. p-Xylene is for example used in the production of
terephthalic acid, which is a main component in polyethylene
terephthalate (PET). Moreover, xylene (in particular, the isomer
p-xylene and o-xylene) is used to produce plastics, polyester
fibers, resins and plasticizers.[1–4] Xylene is mainly produced by
steam cracking of heavy fractions and catalytic reforming of
naphtha. Another possible route to make xylene is by the
transalkylation of the less valuable molecules toluene and
visual evidence of this phenomena is still missing.
The importance of the zeolite acid site strength for the
transalkylation reaction has mainly been investigated on pure
zeolite powder. However, when zeolites are employed in an
industrial reactor, they are usually combined with a binder
material and shaped into catalyst bodies, such as extrudates. The
shaping is done to improve the handling, attrition resistance and
to lower pressure-drop (compared to a powder).[24] From literature
it is known that the binder can have a large effect (both beneficial
or detrimental) on the catalytic activity.[25,26] The addition of a
trimethylbenzene (TMB), which has attracted substantial research
interests.[
4–8]
Simplified, the transalkylation of one mole of toluene
and one mole of TMB forms two moles of xylene. However, the
reaction of toluene and TMB on zeolites is rather complex, as
multiple chemical reactions do occur.[9–12] Besides the
transalkylation reaction, dealkylation, disproportionation and
isomerization reactions can take place. In such instances, not only
does the transalkylation pathway yield xylene, but also the
disproportionation of two molecules of toluene or TMB can form
xylene. Furthermore, the dealkylation of toluene forms benzene
and the dealkylation of TMB results in the formation of xylene.
Finally, the aromatic reactants can undergo isomerization
binder material can for example, influence the overall acidity of
the catalyst. It has been shown that for shaped catalysts with an
Al
2
O
3
binder, Al migration can occur from the binder to the zeolite,
[27–30]
creating extra acid sites.
It is however not known where such
additional acid sites are located and how these acid sites affect
the performance. Opposite to the Al binder, the addition of a
SiO binder can decrease the acidity of a zeolite catalyst. This
was linked to the migration of silicon into the zeolite.
2
O
3
2
[26,31,32]
The
results above show that the binder can have a significant
influence (beneficial or detrimental) on the acidity of the catalyst
and thus the catalytic performance.[24,25] Considering this and the
limited application of high spatiotemporal chemical imaging
techniques to study such intriguing features, it is vital to
investigate the shaped catalyst body in more detail.
[
a]
b]
Inorganic Chemistry and Catalysis, Debye Institute for
Nanomaterials Science, Utrecht University, Utrecht, The
Netherlands
g.t.whiting@uu.nl, b.m.weckhuysen@uu.nl
ExxonMobil Chemical Europe, Inc., European Technology Centre,
Machelen, Belgium
[
In this work, we investigate and visualize the influence of the
binder material on the formation and location of coke molecules
on the catalyst during the transalkylation of toluene and 1,2,4-
TMB. Using in situ UV-vis micro-spectroscopy, we show there is
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
a substantial difference in coke behavior between zeolite H-ZSM-
content, due to the larger pores present in the binder. These
trends in surface area and pore volume also apply to the silicalite
samples. Furthermore, if the catalyst extrudates with the two
binders are compared, it indicates that the pore volume of 20:80
5-containing mm-sized extrudates bound with either SiO
2 2 3
or Al O .
Furthermore, the distinct location of the hydrocarbon species
formed within each of the catalyst bodies, particularly coke, is
imaged in 3-D with ex situ confocal fluorescence microscopy and
the difference in product formation between the catalyst
extrudates was investigated. As a proof of concept, we employ
3
3
Al
2
O
3
(0.49 cm /g) or sil20:80 Al
2
O
3
(0.52 cm /g) is higher than
3
the pore volume of 20:80 SiO
2
(0.28 cm /g). The predicted surface
areas and pore volumes are calculated using the ratio of pure
components and are in good agreement with the experimental
data.
Al
2
O
3
-bound silicalite-based extrudates to demonstrate that the
-bound extrudates are indeed
migration of Al in the Al
2
O
3
responsible for the formation of additional acid sites.
NH
The NH
0:80 Al
than the 20:80 SiO
addition of the acid sites in the Al
As expected, the SiO binder has no acid sites. Furthermore, the
values are similar to the expected values, except for the sil20:80
Al extrudate that has a higher quantity of acid sites than
expected. Furthermore, elemental analysis was used to
determine the purity of the samples. As shown in Table 1, the H-
ZSM-5 crystals do not contain any Na impurities. However, the
binders themselves do contain a small amount of Na ions, which
introduces a small amount of Na atoms into the extrudate samples.
3
TPD was used to study the acidity of the different samples.
TPD results in Table 1 (and Figure S3) show that the
extrudate contains more acid sites (0.46 mmol/g NH
extrudate (0.12 mmol/g NH ) due to the
binder to those of the zeolite.
3
2
2
O
3
3
)
Results and Discussion
2
3
2
O
3
1. Physicochemical characterization of the catalyst materials
2
The different samples (zeolite-based extrudates and pure
components) were characterized with scanning electron
microscopy (SEM), Ar physisorption, NH
2
O
3
3
temperature
programmed desorption (NH TPD) and elemental analysis. The
3
external surfaces of the cylindrical extrudates were imaged with
SEM and the images are displayed in Figure S1 in the supporting
information. The SEM images of the catalyst extrudates show that
the zeolites are almost completely embedded in the extrudate and
covered with the binder. This is the case for all extrudates,
whatever the binder, and demonstrates that the different materials
are comparable. Furthermore, the distribution of the H-ZSM-5
crystals within the extrudate was analyzed with SEM-EDX (Figure
S2). The Si distribution within the extrudate is ascribed to the H-
2. Coke formation during transalkylation at 450 °C
In situ diffuse reflectance UV-vis micro-spectroscopy was used to
follow the coke formation during the transalkylation of toluene and
1,2,4-trimethylbenzene (1,2,4-TMB) on the extrudates. The
(
charged) aromatic species formed during reaction are known to
*
absorb light due to their π-π transitions, leading to specific
absorption regions in the UV-vis spectra (Figure 1a). Absorption
in the region above 33000 cm-1 (< 300 nm) is attributed to the
reactants and products. During the reaction, in the region around
2 3
ZSM-5 crystals imbedded in the Al O -binder, which is
homogeneously distributed throughout the cross section.
The surface area and pore volume of the different studied
samples were analyzed with Ar physisorption and the results are
shown in Table 1. The results show that the BET surface area
decreases when a binder material is added, due to the dilution of
-1
-1
2
8000 cm (350 nm) and around 23000 cm (430 nm), absorption
increases due to the formation of charged alkylated aromatics
which are the reaction intermediates), as shown in Figure 1a. The
smaller the wavenumber, the more conjugated the molecules
(
2
the zeolite in the sample (i.e. the BET surface area of 20:80 SiO ,
are.[ In the region around 16000 cm (625 nm) absorption is
due to the formation of poly-aromatic molecules that block the
pores of the zeolite and hence cause deactivation. [3,34,35]
c
33]
-1
2
2
2
21 m /g, is lower than that of H-ZSM-5 powder, 374 m /g). In
contrast, the total pore volume increases with increasing binder
Table 1. BET surface areas and bpore volumes of the different samples under investigation, measured with Ar physisorption. Quantity of NH
a
per g of catalyst,
3
3 3
measured with NH TPD. Predicted surface areas, pore volumes and quantity of NH are calculated with the ratio of the pure zeolite and the pure binder, and are
d
e
shown between parentheses. The Na content in ppm of the different samples. wt.% of coke after 1 h of transalkylation reaction at 450 °C on the catalyst extrudate,
measured with thermogravimetrical analysis (TGA).
H-ZSM-5:binder
wt.%)
Binder
BET surface
area (m /g)a
Pore volume
(cm /g)b
NH
(mmol/g)c
3
TPD
Na content
(ppm)d
wt.% cokee
2
3
(
H-ZSM-5
100:0
0:100
0:100
100:0
20:80
20:80
20:80
-
374
0.19
0.48
0
5
7
1
1
2
7
4
Al
O
2 3
Al
2
O
3
209
0.47
0.45
174
290
13
SiO
2
SiO
2
184
0.32
0.02
silicalite
-
413
0.19
0
20:80 SiO
2
SiO
2
221 (222)
241 (242)
261 (250)
0.28 (0.29)
0.49 (0.41)
0.52 (0.41)
0.12 (0.11)
0.46 (0.46)
0.43 (0.36)
174
83
20:80 Al
2
O
3
Al
Al
2
O
3
3
sil20:80 Al
2
O
3
2
O
203
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
Figure 1. (a) Schematic of different regions that increase in absorption in the diffuse reflectance UV-vis spectra during the reaction of toluene and 1,2,4-TMB. In
the different spectral regions, an example of the type of molecule that absorbs the light is presented. The laser line wavelengths used to excite such molecules
within confocal fluorescent microscopy are also highlighted. Time-resolved in situ diffuse reflectance UV-vis spectra during the transalkylation of toluene and 1,2,4-
TMB at 450 °C, over (b) 20:80 Al
spectra.
2 3 2 2 3
O , (c) 20:80 SiO , (d) Al O binder and (e) H-ZSM-5 powder for 1 h. The dashed lines indicate an absorption region in the UV-vis
Figure 1b and c show the time resolved in situ diffuse reflectance
UV-vis spectra during the transalkylation reaction of toluene and
charged diphenylmethane (reaction intermediates) and large
poly-aromatic molecules, respectively. For the 20:80 Al
extrudates, an extra absorption band is formed in the region
around 28000 cm-1 (350 nm) in comparison to 20:80 SiO
, which
can be assigned to charged methylated benzene carbocations.
2
O
3
1,2,4-TMB on 20:80 Al
2
O
3
and 20:80 SiO
2
extrudates,
respectively. In the UV-vis diffuse reflectance spectra of 20:80
2
-1
2
SiO , the tail-end of an absorption band at around 40000 cm (250
nm) is visible, due to the absorption of the reactants and products.
During reaction, in the region around 23000 cm-1 (430 nm) and
Furthermore, in the 20:80 Al
2
O
3
spectra, the absorption around
23000 cm-1 (430 nm) and 16000 cm (625 nm) are significantly
-1
around 16000 cm-1 (625 nm), the absorption increases in intensity, higher in intensity than in the 20:80 SiO
spectra, indicating that
2 3
the Al O extrudates form a higher concentration of more large
2
with these regions assigned to alkylated benzene carbocations,
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
conjugated species. Thermogravimetric analysis (TGA) (Table 1)
was used to measure the amount of residue in the samples and
to verify the in situ UV-vis results. The weight % of residue in
spectra between 20:80 Al
explained by the absorption band formed on the Al
does not explain the entire difference. For the ZSM-5-bound
catalyst extrudates, the amount of poly-aromatic compounds
formed is higher than can be expected from the combination of
2
O
3
and 20:80 SiO
2
can largely be
binder, it
2
O
3
2 3 2
20:80 Al O (7 wt.%) is higher than that of 20:80 SiO (2 wt.%),
which corresponds well with the diffuse reflectance UV-vis
spectra, showing more poly-aromatic species are formed for the
the H-ZSM-5 spectra (Figure 1e) and the binders alone (Al
2 3
O
20:80 Al
2
O
3
sample.
binder in Figure 1d and SiO binder in Figure S4). Previously, it
2
has been shown that species (such as Si and/or Al) can migrate
from the binder to the zeolite and vice versa and for extrudates
The differences between the diffuse reflectance UV-vis spectra of
0:80 Al and 20:80 SiO extrudates can be largely explained
by the presence of Al binder. The in situ diffuse reflectance
UV-vis spectra of pure Al
absorption around 28000 cm- (350 nm), which becomes more
intense and shifts to lower wavenumbers due to the formation of
charged alkylated benzenes. The weak acid sites present in the
2
2
O
3
2
with Al
within the extrudate.
2 3
O binder, the migration of Al creates extra acid sites
[
27–30]
2
O
3
2
O
3
(Figure 1d) displays an large
1
In correlation with in situ diffuse reflectance UV-vis spectroscopy
results, whereby the formation of poly-aromatic coke molecules
was monitored, 3-D confocal fluorescence microscopy has been
used to chemically image the location of such species within
catalyst extrudates. The samples were illuminated with four lasers
of different wavelengths, namely, 404 nm, 488 nm, 561 nm and
642 nm, as shown in Figure 1a. Certain molecules are excited by
Al
these charged alkylated aromatics. However, the catalytic data
vide infra) show that pure Al extrudates do not form any
products. Although the difference in diffuse reflectance UV-vis
2 3
O binder can bind with the toluene and 1,2,4-TMB, and form
(
2 3
O
Figure 2. Ex situ 3-D confocal fluorescence microscopy images taken after 1 h of transalkylation of toluene and 1,2,4-TMB at 450 °C, for (a) 20:80 Al
2 3
O , (b) 20:80
SiO and (c) H-ZSM-5 powder. Excitation with 404 nm, 488 nm, 562 nm and 642 nm laser causes respectively the blue, green and red (both 562 and 642 nm laser)
2
fluorescence. For all samples an enlargement of the marked area is shown in the center panel. On the right side of the image the ratio of different types of coke
formation in the corresponding sample is plotted. Type A has mainly coke molecules at the rim of the zeolite crystals; and type B has coke molecules throughout
the crystal.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
a laser with a wavelength in the absorption band of these
molecules and can, subsequently, fluoresce. In the 3-D chemical
image, the fluorescent molecules have a different color, namely:
blue for 404 nm laser excitation; green for the 488 nm laser; and
red for both the 561 nm and 642 nm laser.
Not only are the type of coke molecules different for the 20:80
SiO and 20:80 Al O extrudates, but also the location of the coke
2 2 3
molecules in the zeolite crystals. Two different types of coking
profiles in the embedded zeolite crystals can be distinguished:
type A, has mainly fluorescence (coke) at the edge of the zeolite
crystals, with a thickness of <1 µm; and type B, where zeolite
crystals are completely fluorescent (coked) throughout. To
distinguish between Type A and Type B coke characteristics, first
the center of the zeolite is identified by scrolling through the 3-D
image. To have a Type A coke profile, the center must clearly
have a lower fluorescence intensity then the edge, otherwise it is
marked as a Type B coke profile. The right column in Figure 2
shows the ratio of type A to type B zeolite crystals displayed on
Figure 2 shows the ex situ 3-D confocal fluorescence microscopy
images of 20:80 Al O , 20:80 SiO and H-ZSM-5 imaged after 1 h
2 3 2
of reaction. In all three samples, only the molecules present in/on
the zeolite crystals are displaying fluorescence, whereas the
binders themselves do not have any fluorescence and are
displayed in Figure S5. Molecules within the zeolite crystals in the
2
0:80 SiO
excited with all lasers, except 404 nm. Comparing this with 20:80
Al (Figure 2a) shows that the molecules inside the extrudate
2
extrudate (Figure 2b) exhibit fluorescence when
average in each extrudate. It is clear that the SiO
influence the location of the fluorescent coke molecules, as the
zeolites within 20:80 SiO mainly display a type B coking profile,
which is similar to the H-ZSM-5 powder. In contrast to the SiO
extrudate, the 20:80 Al extrudate displays a substantially
2
binder does not
2
O
3
fluoresce when excited at longer wavelengths. This is due to more
large conjugated poly-aromatic molecules inside the extrudate, as
observed with diffuse reflectance UV-vis spectroscopy (Figure 1b)
and TGA (Table 1). Furthermore, the H-ZSM-5 powder displays
fluorescence when excited at shorter wavelengths, compared to
the binder-bound extrudates, confirming the less conjugated
nature of the coke species, as observed by diffuse reflectance
UV-vis spectroscopy.
2
2
2
O
3
higher quantity of type A than type B crystals, when compared to
the H-ZSM-5 powder. Previously, it has been shown that Al
migration between the binder and the zeolite can occur for
catalyst shaped with an Al
2
O
3
binder, creating extra acid sites.[27–
These acid sites could contribute to the unusual spatial
3
0]
distribution of the coke species on the edge region of the ZSM-5
crystals.
Figure 3. Time-resolved in situ diffuse reflectance UV-vis spectra during the transalkylation of toluene and 1,2,4-TMB at 450 °C, over (a) sil20:80 Al
silicalite for 1 h. The dashed lines indicate an absorption region in the UV-vis spectra. (c) Ex situ 3-D confocal fluorescence microscopy images taken after 1 h of
transalkylation of toluene and 1,2,4-TMB at 450 °C, for sil20:80 Al . Excitation with 404 nm, 488 nm, 562 nm and 642 nm laser causes the blue, green and red
both 562 and 642 nm laser) fluorescence, respectively. An enlargement of the marked area is shown in the center of the image. The fluorescence images were
2 3
O and (b) 100
2 3
O
(
measured with a higher laser power than the previous images. On the right side of the image the ratio of different types of coke formation in the corresponding
sample is plotted. Type A has mainly coke molecules at the rim of the zeolite crystals; and type B has coke molecules throughout the crystal.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
Figure 4. (a) Total feed conversion, (b) m/p-xylene and (c) 1,3,5-TMB yield as a function of time on stream, during the transalkylation reaction of toluene and 1,2,4-
-
1
TMB over 20:80 Al
O
2 3
(red), 20:80 SiO
2
(black), ZSM-5 (blue) and 100 Al
O
2 3
(green) at 450 °C and 2 h WHSV. The results are normalized per g of catalyst.
3
. Linking binder-effects to the spatial coke distribution
Whether or not the location of these acid sites can influence the
spatial distribution of coke species was confirmed using ex situ 3-
D confocal fluorescence microscopy on the silicalite-based
extrudates after reaction. As the silicalite-containing extrudates
To investigate whether the creation of additional acid sites from
Al migration during catalyst preparation, can contribute to the
specific location of coke species, extrudates containing 20 wt.%
have less acid sites and consequently contain
a lower
silicalite crystals bound with 80 wt.% Al
prepared and evaluated. Silicalite crystals do not contain
aluminum and consequently have no acid sites. However, if Al
2
O
3
(sil20:80 Al
2
O
3
) were
concentration of coke molecules, the CFM images were
measured at a higher laser power for visibility. The 3-D chemical
image of sil20:80 Al O , in Figure 3c, again indicates the formation
2 3
migrates from the Al
can be created.[27–30] The in situ diffuse reflectance UV-vis spectra
of the sil20:80 Al extrudate (Figure 3) display the formation of
several absorption bands during the transalkylation reaction. The
2
O
3
binder to the silicalite crystals, acid sites
of poly-aromatic species, corroborating the diffuse reflectance
UV-vis spectra (Figure 3a). As the fluorescent poly-aromatic coke
molecules are located only on the previously inert silicalite
crystals (Figure S5), the creation of additional acid sites is
2
O
3
-1
absorption in the regions around 28000 cm (350 nm) and 23000
cm-1 (430 nm) can be assigned to charged alkylated aromatics.
The large absorption band around 28000 cm-1 (350 nm) can also
2 3
unequivocal. Remarkably, the sil20:80 Al O sample also displays
[35]
both type A and type B coking profiles, even though it is more
difficult to distinguish between type A and type B coking profiles
than for the ZSM-5-containing extrudates, due to the lower
fluorescence intensity in the silicalite-containing extrudates. In
particular, the presence of type A coking, indicates that indeed,
acid sites created at the zeolite crystal ‘rim’ do contribute to
corresponding ‘rim’ formation of coke species.
be found in the UV-vis diffuse reflectance spectra of the Al
binder (Figure 1d). Additionally, absorption around 16000 cm
2
O
3
1
-
(
625 nm) appears due to the formation of large poly-aromatic
molecules. The latter molecules cannot be formed by either of the
two separate components (Al binder in Figure 1d and silicalite
2
O
3
powder in Figure 3b) and confirms that acid sites are created upon
mixing/binding of the individual components and can indeed
participate in this reaction.
This creation of extra acid sites influences the acidic properties of
the catalyst and can consequently influence the transalkylation
reaction. The differences in catalyst performance during the
reaction were therefore tested in a fixed bed reactor. Figure 4
2
Figure 5. Ex situ 3-D confocal fluorescence microscopy images taken of 20:80 SiO extrudate after 1 h of reaction with 1,3,5-TMB. Excitation with 404 nm, 488 nm,
562 nm and 642 nm laser causes respectively the blue, green and red (both 562 and 642 nm laser) fluorescence. For all samples, an enlargement of the marked
area is shown in the center panel. On the right side of the image the ratio of different types of coke formation in the corresponding sample is plotted. Type A has
mainly coke molecules at the rim of the zeolite crystals; and type B has coke molecules throughout the crystal.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
shows the total conversion, m/p-xylene yield and 1,3,5-TMB yield Conclusions
per g of catalyst, as a function of time on stream. Comparing the
total conversion for the Al
SiO -bound catalyst extrudate shows a 2-fold increase for the
Al -bound extrudate, even though both extrudates contain 20
wt.% of H-ZSM-5 zeolite crystals. This increase in conversion is
not due to the Al binder itself as the 100 Al extrudate is
unreactive and therefore must be due to a binder effect in the
0:80 Al extrudate, for example the increase in acid sites at
2
O
3
-bound catalyst extrudate with the
Using a combination of in situ diffuse reflectance UV-vis micro-
spectroscopy and ex situ confocal fluorescence microscopy, we
demonstrate that the choice of the binder material has a large
effect on the deactivation of catalyst extrudates during the
transalkylation of toluene with 1,2,4-trimethylbenzene. The
deactivation process during the transalkylation reaction was
investigated for millimeter-sized zeolite H-ZSM-5-containing
2
2
O
3
2
O
3
2 3
O
2
2 3
O
the surface of the zeolite crystals within the extrudates. These
results are consistent with the UV-vis results, as the formation rate
catalyst extrudates with either SiO
UV-vis micro-spectroscopy combined with TGA, showed that the
catalyst extrudate with 20:80 Al binder forms more large poly-
aromatic species than the catalyst extrudate with SiO binder. The
difference cannot be explained by the two separate components,
and must be due to binder effects occurring upon the combination
of the individual components. Furthermore, using complementary
confocal fluorescence microscopy, it was demonstrated that in the
2 2 3
or Al O as binder material.
for the absorption bands is higher for the 20:80 Al
2
O
3
catalyst
2 3
O
extrudate than for the 20:80 SiO
higher conversion.
2
catalyst extrudate, indicating a
2
Furthermore, the yields of both p/m-xylene and 1,3,5-TMB can
be compared for the 20:80 Al and 20:80 SiO catalyst
2
O
3
2
extrudate. The xylene yield is similar for both the extrudate
samples, which is expected as both samples contain 20 wt.% of
H-ZSM-5 crystals. However, the 1,3,5-TMB yield is higher for the
catalyst extrudate containing SiO
predominantly throughout the zeolite crystals, similar to the
zeolite H-ZSM-5 without binder. In the Al extrudate, however,
2
binder, the coke is formed
2
O
3
2
0:80 Al
2
O
3
extrudate compared to the 20:80 SiO
2
extrudate.
the coke is mainly formed on the rim of the zeolite crystals. Both
discrepancies between the two different extrudates (more coke
Taking into account that CFM showed that the 20:80 Al
2
O
3
catalyst extrudate forms coke on the rim of the zeolite crystals
imbedded in the extrudate, we suggest that extra acid sites are
created between the zeolite and the binder. These extra acid sites
can isomerize the 1,2,4-TMB into 1,3,5-TMB. Dumitriu et al.[
2 3
formation and coke at the rim of the zeolite for the Al O -bound
extrudate) were explained by the creation of extra acid sites on
the zeolite/binder interface. The migration of species was
confirmed by performing the transalkylation reaction on
18]
showed that trimethylbenzene can isomerize on weaker acid sites, extrudates containing silicalite crystals (i.e., in the absence of
whereas the transalkylation and disproportionation occur on
stronger acid sites. The 1,3,5-TMB, which is isomerized at the
edge of the zeolite, is larger and diffuses slower through the
2
aluminium) bound with Al O3. It was shown that these catalyst
extrudates contain acid sites created by Al migration and form
poly-aromatic species during reaction, that were not observed for
zeolite pores, causing coke to form on the rim of the ZSM-5 zeolite. the corresponding individual components of the extrudate. These
As the conversion includes the formation of TMB isomers, the
acid sites enhance 1,2,4-TMB isomerization into more bulky
1,3,5-TMB and 1,2,3-TMB. The 1,3,5-TMB reacts mainly at the
zeolite external surface to from coke, which was confirmed by the
increased conversion for the 20:80 Al
increased formation of 1,3,5-TMB.
2 3
O extrudates is due to the
2
reaction of 1,3,5-TMB on the SiO -bound catalyst extrudate. Such
To explore if the type A coking of the zeolite crystals is indeed
caused by the reaction of sterically hindered 1,3,5-TMB at the
zeolite external surface, the 20:80 SiO extrudate was reacted
2
with only 1,3,5-TMB under the same reaction conditions as the
previous toluene and 1,2,4-TMB reaction. Figure 5 shows the
fluorescence microscopy images of the extrudate after reaction
with 1,3,5-TMB and the corresponding in situ diffuse reflectance
UV-vis spectra are shown in Figure S6. The results demonstrate
there is a 2.5 times increase in type A coke formation compared
visual confirmation of both binder effects in zeolite catalyst bodies
could shed light into how to limit or enhance binder effects by
catalyst design, in order to prolong catalyst lifetime, or to minimize
the formation of side products during a catalytic reaction.
Experimental Section
to the reaction with toluene and 1,2,4-TMB on 20:80 SiO
2
(Figure
2b). This confirms that 1,3,5-TMB reacts at the outside of the
H-ZSM-5 extrudates bound with Al O or SiO2 as binder material were
2
3
zeolite crystals, and provides the first direct visual confirmation
that a rim of coke species can be formed near the external surface
of zeolite crystals in industrial-grade catalysts.
used for the transalkylation reaction of toluene and 1,2,4-trimethylbenzene
(1,2,4-TMB). The SiO - or Al -bound extrudates were prepared by
mixing zeolite ZSM-5 crystals with H O, extrusion aid and binder,
according to patent US6,039,864.[36] The zeolite H-ZSM-5 crystals had a
Si/Al ratio of 32 and an average crystal size of 3 µm. For the SiO -bound
extrudates, SiO gel (AEROSIL 300) and silica sol (NALCOAG 1034A)
were used as binder and, for the Al bound extrudates, pseudoboehmite
alumina (Versal 300) was used as binder. The prepared mixture was
extruded into 1 mm extrudates, dried overnight at 130 °C and calcined at
500 °C for 18 h. Ion-exchange was performed by suspending the
extrudates in a 1 M ammonium nitrate solution for 4 h, followed by washing
and drying and a final calcination at 500 °C for 18 h.
2
2 3
O
2
2
2 3
The consequence of the creation of extra acid sites in Al O -
2
bound extrudates, is that more coke species, formed at the edge
of the zeolite crystals, dramatically influences the selectivity of the
catalyst, particularly yielding more isomerization products. This
should be considered when selecting specific binders for
alkylation-type reactions.
2
O
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
The formed extrudates contained 20 wt.% H-ZSM-5 and 80 wt.% Al
SiO binder. The small zeolite weight ratio of 20:80 wt.% was chosen, so
2
as to observe the effects of the binder more clearly. The different samples
were named as follows: A:B binder, in which “A” corresponds to the wt.%
H-ZSM-5 content and “B” corresponds to the wt.% binder, for example,
2
O
3
or
equipped with a 15x 0.28 NA reflective lens and a 30W halogen lamp.
During reaction, the diffuse reflectance spectra were collected from a
region (82x82 µm) on top of the extrudate external surface and were
always taken on the first extrudate that this flow of reactants reaches. The
spectra were recorded every 10 s for 1 h.
‘
5
20:80 SiO
crystals. Furthermore, reference extrudates, containing only binder, and
zeolite H-ZSM-5 powder were used. Additionally, silicalite extrudates were
used and the extrudates containing 20 wt.% silicalite and 80 wt.% Al
binder were named, ‘sil20:80 Al ’.
2 2
’ is a 80 wt.% SiO bound extrudate containing 20 wt.% H-ZSM-
After 1 h of reaction, the coke formation on the samples was imaged with
ex situ confocal fluorescence microscopy using a Nikon Eclipse 90i
instrument equipped with a 100x 0.73 NA dry objective. The sample was
excited with four different laser lines; 404 nm, 488 nm, 561 nm and 642
nm and the emission was detected with an A1-DU4 4 detector unit in the
range of 425-475 nm, 500-550 nm, 570-620 nm and 662-737 nm,
respectively. Only the in-focus light is collected by the detector due to a
pinhole. By making 2-D microscopy images at different focal depths, a 3-
D chemical image is composed. To determine the different types of coking
of the zeolite crystals in each of the 3-D images, two spots on three
different extrudates were manually counted. Furthermore, after the
reaction, the hydrocarbon deposits inside the extrudate were
characterized with thermogravimetric analysis (TGA). The samples were
2 3
O
2 3
O
The different catalysts were characterized with scanning electron
microscopy (SEM), Ar physisorption, NH temperature programmed
desorption (NH TPD) and elemental analysis. SEM analysis was
3
3
conducted by attaching a whole extrudate to a SEM stub and coated with
Pd/Pt in a Cressington 208HR sputter coater. The surface of the extrudate
was imaged on a JEOL JSM-6340F Field Emission Gun Scanning Electron
Microscope (FEG-SEM) operated at 5 kV using secondary electrons. The
surface area and pore volume of the samples was measured with Ar
physisorption. First the extrudates were dried under vacuum at 300 °C
overnight. All samples were measured with a Micromeritics TriStar 3000
at 77 K. The surface area was calculated with the Brunauer-Emmett-Teller
heated to 700 °C with 5 °C/min under 10 ml/min O
1 TGA instrument.
2
on a PerkinElmer pyris
(
BET) method and the t-plot method was used to determine the pore
volume. NH TPD was used to measure the acidity of the samples and the
measurements were performed on a Mettler Toledo TGA/SDTA 851. Prior
to NH TPD, the samples were dried at 550 °C under He flow with a heating
rate of 10 °C/min for 15 min. Subsequently, the samples were exposed to
2 pulses of 10 % NH in He at 100 °C. By applying a temperature ramp
to 550 °C with 5 °C/min under He flow, desorption data have been
collected. The amount of desorbed ammonia was monitored with a thermal
conductivity detector (TCD). Finally, elemental analysis was done on a
Varian Vista MPX Inductively Coupled Plasma Optical Emission
Spectrometer using multi-element standards provided by Merck and Alfa
Aesar. Before analysis, around 200 mg of the sample was dissolved in a
mixture of acids (30% HCl, 3 mL; 65% HNO3, 1 mL; 40% HF, 2 mL) for 4–
3
Acknowledgements
3
We thank Marjan Versluijs-Helder (Utrecht University, UU) for all
the SEM and TGA measurements and Serguei Matveev (UU) for
the SEM-EDX measurements. This work was financially
supported by ExxonMobil. G.T.W. acknowledges
personal 'Veni' grant (722.015.003).
1
3
a NWO
Keywords: Fluorescence spectroscopy • UV/vis spectroscopy •
zeolites • transalkylation • in-situ microscopy
5
h (the first hour with sonication) and then further diluted to 100 mL milli-
Q water.
[
[
[
1]
2]
3]
W. Vermeiren, J.-P. Gilson, Top. Catal. 2009, 52, 1131–1161.
J. M. Serra, E. Guillon, A. Corma, J. Catal. 2005, 232, 342–354.
S. Al-Khattaf, S. A. Ali, A. M. Aitani, N. Žilková, D. Kubička, J. Čejka,
Catal. Rev. Sci. Eng. 2014, 56, 333–402.
The catalytic tests were performed in a fixed-bed reactor with an internal
diameter of 10 mm, which was operated in up flow mode. The reactor was
loaded with approximately 2 g of crushed and sieved catalyst with a
particle size of 0.2 - 0.4 mm. Prior to the catalytic reaction, the catalyst was
[4]
[5]
[6]
[7]
R. Thakur, S. Barman, R. K. Gupta, Chem. Eng. Commun. 2017, 204,
2
activated under N flow at 450 °C for 1 hour. The reactants were delivered
2
54–264.
J. Hanika, Q. Smejkal, A. Krejci, J. Kolena, D. Kubicka, Petr. Coal. 2003,
5, 78–82.
S. A. Ali, K. E. Ogunronbi, S. S. Al-khattaf, Chem. Eng. Res. Des. 2013,
1, 2601–2616.
via an ISCO pump with a 1:1 molar ratio of toluene:1,2,4-TMB and a
-1
WHSV of 2 h . The products were analyzed with a GC equipped with
DBWAX column. The conversion and product yields were calculated from
the GC results, assuming that the response factor for all the reactants and
products is 1 and the results were calculated per gram of catalyst.
4
9
A. Krejčí, S. Al-Khattaf, M. A. Ali, M. Bejblová, J. Čejka, Appl. Catal. A
Gen. 2010, 377, 99–106.
Deactivation of the different samples was followed with in situ diffuse
reflectance UV-vis micro-spectroscopy under similar conditions as the
catalystic tests. The reaction was performed in a TS1500V Linkam
Scientific Instruments cell coupled with TMS94 temperature controller and
PSU1500 unit from Linkam Scientific Instruments. This is a small heating
plate on which the extrudates are placed. An additional calcination was
performed before reaction, by heating the samples to 500 °C with
[8]
[9]
S. H. Cha, S. B. Hong, J. Catal. 2018, 357, 1–11.
J. Čejka, B. Wichterlová, Catal. Rev. 2002, 44, 375–421.
[10] S. Al-Khattaf, M. N. Akhtar, T. Odedairo, A. Aitani, N. M. Tukur, M. Kubů,
Z. Musilová-Pavlačková, J. Čejka, Appl. Catal. A Gen. 2011, 394, 176–
190.
[11] S. A. Ali, A. M. Aitani, C. Ercan, Y. Wang, S. Al-Khattaf, Chem. Eng. Res.
Des. 2011, 89, 2125–2135.
1
0 °C/min under a 1:1 flow ratio of N
reaction was performed at 450 °C for 1 h. For this reaction, toluene and
,2,4-TMB were put in separate saturators and N was bubbled through
the separate saturators, to reach a 1:1 molar ratio and when 1,3,5-TMB
was reacted, an N flow of 16.1 ml/min was used. The reactants are
evaporated by bubbling N through a saturator and this is flown over the
extrudates. Diffuse reflectance UV-vis micro-spectroscopy measurements
were performed on
CRAIC 20/30 PVTM micro-spectrophotometer
2
and O
2
for 1 h. The transalkylation
[12] J. Toda, A. Corma, G. Sastre, J. Phys. Chem. C. 2016, 120, 16668–
16680.
1
2
[13] S. Al-Khattaf, N. M. Tukur, A. Al-Amer, Ind. Eng. Chem. Res. 2007, 46,
4459–4467.
2
[14] I. Wang, T. C. Tsai, S. T. Huang, Ind. Eng. Chem. Res. 1990, 29, 2005–
2
2012.
[15] S. M. Waziri, A. M. Aitani, S. Al-Khattaf, Ind. Eng. Chem. Res. 2010, 49,
a
6376–6387.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
[
16] Y. Li, H. Wang, M. Dong, J. Li, Z. Qin, J. Wang, W. Fan, RSC Adv. 2015,
[27] G. T. Whiting, F. Meirer, M. M. Mertens, A.-J. Bons, B. M. Weiss, P. A.
Stevens, E. de Smit, B. M. Weckhuysen, ChemCatChem. 2015, 7, 1312–
1321.
5, 66301–66310.
[17] F. J. Llopis, G. Sastre, A. Corma, J. Catal. 2004, 227, 227–241.
[
18] E. Dumitriu, C. Guimon, V. Hulea, D. Lutic, I. Fechete, Appl. Catal. A Gen.
[28] Y. Zhang, Y. Zhou, A. Qiu, Y. Wang, Y. Xu, P. Wu, Ind. Eng. Chem. Res.
2006, 45, 2213–2219.
2002, 237, 211–221.
[19] S. Zheng, A. Jentys, J. A. Lercher, J. Catal. 2006, 241, 304–311.
[20] J. Čejka, B. Wichterlová, Catal. Rev. 2002, 44, 375–421.
[21] S. H. Park, H. K. Rhee, Catal. Today. 2000, 63, 267–273.
[22] H. P. Röger, K. P. Möller, C. T. O’Connor, Microporous Mater. 1997, 8,
[29] M. W. Kasture, P. S. Niphadkar, V. V. Bokade, P. N. Joshi, Catal.
Commun. 2007, 8, 1003–1008.
[30] A. Martin, H. Berndt, U. Lohse, U. Wolf, J. Chem. Soc. Faraday Trans.
1993, 89, 1277–1282.
1
51–157.
[31] P. Gélin, T. Des Courières, Appl. Catal. 1991, 72, 179–192.
[32] P. Gelin, C. Gueguen, Appl. Catal. 1988, 38, 225–233.
[33] K. Hemelsoet, Q. Qian, T. De Meyer, K. De Wispelaere, B. De Sterck, B.
M. Weckhuysen, M. Waroquier, V. Van Speybroeck, Chem. Eur. J. 2013,
19, 16595–16606.
[
23] N. M. Tukur, S. Al-Khattaf, Energy and Fuels. 2007, 21, 2499–2508.
24] S. Mitchell, N.-L. Michels, J. Pérez-Ramírez, Chem. Soc. Rev. 2013, 42,
[
6
094–6112.
25] J. S. J. Hargreaves, A. L. Munnoch, Catal. Sci. Technol. 2013, 3, 1165–
171.
26] N.-L. Michels, S. Mitchell, J. Pérez-Ramírez, ACS Catal. 2014, 4, 2409–
417.
[
[
1
[34] M. Guisnet, P. Magnoux, Appl. Catal. A Gen. 2001, 212, 83–96.
[35] F. Chen, G. Coudurier, C. Naccache, Stud. Surf. Sci. Catal. 1989, 49,
1387–1396.
2
[36] G. D. Mohr, J. P. Verduijn, Hydrocarbon Conversion Process Using
Zeolite Bound Zeolite Catalyst, US6,039,864 A, 2000.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900777
FULL PAPER
FULL PAPER
Suzanna P. Verkleij, Gareth T.
Whiting*, Denise Pieper, Sonia Parres
Esclapez, Shiwen Li, Machteld M.
Mertens, Marcel Janssen, Anton-Jan
Bons, Martijn Burgers, Bert M.
Weckhuysen*
Page No. – Page No.
Chemical Imaging of the Binder-
dependent Coke Formation in
Zeolite-based Catalyst Bodies
during the Transalkylation of
Aromatics
This article is protected by copyright. All rights reserved.