Impact of the Spatial Organization of Bifunctional Metal–Zeolite Catalysts on the Hydroisomerization of Light Alkanes

Article abstract of DOI:10.1002/anie.201915080

Improving product selectivity by controlling the spatial organization of functional sites at the nanoscale is a critical challenge in bifunctional catalysis. We present a series of composite bifunctional catalysts consisting of one-dimensional zeolites (ZSM-22 and mordenite) and a γ-alumina binder, with platinum particles controllably deposited either on the alumina binder or inside the zeolite crystals. The hydroisomerization of n-heptane demonstrates that the catalysts with platinum particles on the binder, which separates platinum and acid sites at the nanoscale, leads to a higher yield of desired isomers than catalysts with platinum particles inside the zeolite crystals. Platinum particles within the zeolite crystals impose pronounced diffusion limitations on reaction intermediates, which leads to secondary cracking reactions, especially for catalysts with narrow micropores or large zeolite crystals. These findings extend the understanding of the ??intimacy criterion” for the rational design of bifunctional catalysts for the conversion of low-molecular-weight reactants.

Full text of DOI:10.1002/anie.201915080

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Impact of the Spatial Organization of Bifunctional Metal-Zeolite
Catalysts for Hydroisomerization of Light Alkanes
Authors: Kang Cheng, Lars I. van der Wal, Hideto Yoshida, Jogchum
Oenema, Zhaorong Zhang, Glenn Sunley, Jovana Zečević,
Krijn P. de Jong, and Justine Harmel
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.201915080
Angew. Chem. 10.1002/ange.201915080
Link to VoR: http://dx.doi.org/10.1002/anie.201915080
http://dx.doi.org/10.1002/ange.201915080

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
Impact of the Spatial Organization of Bifunctional Metal-Zeolite
Catalysts for Hydroisomerization of Light Alkanes
Kang Cheng,[a] Lars I. van der Wal,[a] Hideto Yoshida,[b] Jogchum Oenema,[a] Justine Harmel,[a]
Zhaorong Zhang,[c] Glenn Sunley,[d] Jovana Zečević,[a] and Krijn P. de Jong*[a]
Abstract: To improve the product selectivity by controlling the
spatial organization of functional sites at nanoscale is a critical
challenge in bifunctional catalysis. Herein, we design a series of
composite bifunctional catalysts consisting of one-dimensional
zeolites (ZSM-22 and mordenite) and γ-alumina binder, with
platinum particles controllably deposited either on the alumina binder
or inside the zeolite crystals. The hydroisomerization of n-heptane
demonstrates that the bifunctional catalysts with platinum particles
on the binder, which separates platinum and acid sites at nanoscale
distance, surprisingly offers a higher yield of desired isomers than
catalysts with platinum particles inside the zeolite crystals. Locating
platinum particles within the zeolite crystals imposes pronounced
diffusion limitations to reaction intermediates, giving rise to
secondary cracking reactions, especially for catalysts with narrow
micropores or large zeolite crystals. These findings extend the
understanding of the “intimacy criterion” for rational design of
bifunctional catalysts for the conversion of low molecular weight
reactants.
and the diffusion of molecules involved in the reaction.[4]
Representative catalysts for hydroisomerization are
composed of platinum and acidic zeolites. The metal sites
catalyze the dehydrogenation of alkanes to alkenes and
subsequent hydrogenation of iso-alkenes to desired iso-alkanes
while the Brønsted acid sites (BAS), provided by zeolites,
catalyze the skeletal rearrangement or cracking reactions via
carbenium ions as intermediates.[5] The catalytic performance of
bifunctional catalysts is determined by many parameters such as
zeolite topology, acid strength and density, metal properties, and
the ratio of metal sites relative to acid sites.[3b,6] Furthermore, the
distance between the two functional sites plays an important role
in catalysis.[6c,7] To date, research has particularly focused on
the aforementioned parameters, while Weisz’s intimacy criterion
for decades was often applied as “the closer the better” for
optimizing spatial arrangement of metal and acid sites.[8]
However, one should note that Weisz’s intimacy criterion for
bifunctional catalysts, which was verified at micron-scale, mainly
correlates to catalytic activity. The impact of intimacy between
metal and acid sites at nanoscale and on product selectivites is
still not described well.
Introduction
Regarding the nanoscale intimacy, Höchtl et al. reported that
at nanoscale, the distance between acid and metal sites has
only a minor influence for n-heptane hydroisomerization, as long
as the metal and zeolite are in direct contact.[9] In contrast, our
group recently studied the hydroconversion of C10–C19 paraffins
over bifunctional catalysts, containing zeolite Y and alumina
binder, and with platinum metal controllably deposited either
within the zeolite crystals or on the binder.[10] It was found that
the platinum-on-binder catalyst offered higher isomerization
selectivity and lower cracking selectivity than platinum-in-zeolite
catalyst. Subsequently, Samad, Moussa and their colleagues
reported that an atomic-scale proximity between metal and acid
sites was not required to reach a great bifunctionality in
isomerization reactions, while catalyst with a nanoscale distance
between two functional sites performed better.[11] Nevertheless,
up till now few studies have touched upon the impact of
nanoscale proximity between functional sites due to the limited
synthesis protocols and challenges in characterization. In
particular the impact of platinum location on intracrystalline
diffusion of linear and branched alkanes with low molecular
weight has not been demonstrated.
In this work, we particularly focus on the nanoscale intimacy
for the hydroisomerization of light alkanes over industrially
relevant bifunctional catalysts, which consist of one-dimensional
(1D) zeolites, γ-alumina binder and platinum particles. ZSM-22
(TON-type) and Mordenite (MOR-type) with different micropore
openings are employed as acid functions due to their wide
applications in industry.[12] Next to industrial application these 1D
zeolites have been studied for their intriguing characteristics for
diffusion of molecules, sometimes referred to as single–file
diffusion.[13] We selectively locate platinum particles either on the
Bifunctional catalysis is involved in a wide range of extremely
important catalytic processes, such as biomass upgrading,
hydrogenation
of
CO/CO2
to
hydrocarbons
and
hydroisomerization of n-paraffins to high-quality fuels.[1] In many
cases, bifunctional coupling of two different reactions can
provide enhanced catalytic activities and unexpected product
selectivities.[2] One of most applied bifunctional catalysts
consists of a transition metal, for example, platinum, palladium,
or molybdenum (sulfide), supported on solid acids such as
alumina, alumina–silica or zeolites.[3] Therefore, the bifunctional
catalysis naturally brings out the issue how to organize the two
components rationally to maximize the synergy, without
disturbing the communication between two types of active sites
[a]
Dr. K. Cheng, L. I. van der Wal, J. Oenema, Dr. J. Harmel, Dr. J.
Zečević and Prof. K. P. de Jong
Inorganic Chemistry and Catalysis
Debye Institute for Nanomaterials Science, Utrecht University
3584 CG, Utrecht (The Netherlands)
E-mail: k.p.dejong@uu.nl
[b]
[c]
Dr. H. Yoshida
The Institute of Scientific and Industrial Research, Osaka University,
8-1 Mihogaoka, Ibaraki, Osaka 567-0047 (Japan)
Dr. Z. Zhang
Applied Chemistry and Physics Centre of Expertise
BP Group Research
150 West Warenville Road, Naperville, IL 60563, (U.S.A.)
Dr. G. Sunley
[d]
Hull Research and Technology Center
BP plc, c/o BP Chemicals, Saltend, Hull HU 12 8DS, (U.K.)
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.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
alumina binder or within the zeolite crystals, providing distinct
intimacy levels of metal and acid sites at nanoscale. Moreover,
two mordenite zeolites with different crystal sizes are used to
investigate the effect of diffusion length on catalytic
performances. To unequivocally determine the platinum location
and sizes, we use high-angle annular dark-field scanning
transmission electron microscopy (HAADF-STEM) to image the
70-nm-thick ultramicrotomed sections of these composite
catalysts. Bifunctional catalysts are evaluated under industrially
relevant conditions for the hydroisomerization of n-heptane.
Interestingly, locating platinum out of the zeolite crystals is
beneficial for the selectivity of C7 isomers, while embedding
platinum inside the zeolite crystals favors cracking reactions.
the micropore channels can be attributed to the collapse of
neighboring micropores during the growth of Pt particles, which
was confirmed by Zečević et al. using electron tomography.[15]
The loading of Pt particles in and on H-ZSM-22 crystals had no
significant impact on zeolite porosity (Table S1). Although Pt-H-
ZSM-22 shows a heterogeneous distribution of Pt particles, this
catalyst still provides the closest proximity between Pt and acid
sites.
Results and Discussion
H-ZSM-22 based catalysts with 10MR micropores
SEM imaging depicts the typical agglomerates of rod-like H-
ZSM-22 crystals with an average crystal length of 1.6 µm (Figure
1a). The micropore channels with 10-membered rings (10MR)
openings normally run parallel to the crystal length, and there is
no communication between these channels (Figure 1b). By
using an ion exchange method with Pt(NH3)4(NO3)2 and parent
2+
H-ZSM-22 zeolite, Pt(NH3)4
ions can be introduced by
interacting with cation exchange sites. It is notable that the
2+
kinetic diameter of Pt(NH3)4 ion (0.48 nm) is close to the H-
ZSM-22 pore openings (0.45×0.55 nm),[14] so the diffusion of
2+
Pt(NH3)4 ions into H-ZSM-22 micropore channels might be
impeded. Figure 1c shows that the parallel micropore channels
run along with the length of crystals (c-axis), and Pt particles
were well-dispersed on or in H-ZSM-22 crystals with average
size of 1.6 nm. However, the spatial distribution of Pt particles
along H-ZSM-22 crystals cannot be clarified by only one
orientation. Up to now, there is no study on the spatial
distribution of Pt particles within individual H-ZSM-22 crystals.
To clarify the distribution of Pt particles in H-ZSM-22, HAADF-
STEM was used to investigate the 70-nm-thick ultramicrotomed
sections of reduced Pt-H-ZSM-22 catalyst. Cutting the resin-
embedded catalysts with an ultramicrotome to produce thin
slices exposes the cross-sections of catalyst granules, thus
providing information of spatial arrangements of functional sites
inside the shaped catalyst bodies (Figure 1d and S1). Figure 1e
displays the cross sections (a-b plane) of H-ZSM-22 crystals,
which are perpendicular to the c-axis direction. These sections
suggest that a fraction of Pt particles was located inside the H-
ZSM-22 crystals, but more Pt particles were deposited on the
external surface of H-ZSM-22 crystals (Figure S2). Pt particles
may deposit in and on four possible positions of H-ZSM-22:
micropores, mesopores, external surface and zeolite crystals
(Figure S3). Considering there are no mesopores in the H-ZSM-
22 crystals used and almost no Pt particles smaller than 0.5 nm
were present (Figure 1c), we can exclude the first two options.
By counting more than 200 platinum particles from the cross
sections of Pt-H-ZSM-22, around 25% of platinum particles are
estimated to be present inside H-ZSM-22 crystals (Figure S2).
The sizes for Pt particles inside zeolite crystals being larger than
Figure 1. Structure of H-ZSM-22 and spatial distribution of Pt particles along
H-ZSM-22 crystals. a) SEM image of H-ZSM-22 crystals. b) Pore orientation
(c-axis) and pore openings of H-ZSM-22. c) HAADF-STEM image of Pt-H-
ZSM-22 along the c-axis direction without ultramicrotomy. d) Schematic
diagram of sectioning of resin-embedded catalyst particles using a diamond
knife. e) HAADF-STEM image of a-b plane of ultramicrotomed Pt-H-ZSM-22.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 2. HAADF-STEM images with ultramicrotomy of a) Pt-H-ZSM-22/Al2O3 with Pt in and on the zeolite crystals and b) Pt-Al2O3/H-ZSM-22 with Pt on the
alumina binder. Impact of Pt location on activities and yields of n-heptane hydroconversion over Pt-H-ZSM-22/Al2O3 (black) and Pt-Al2O3/H-ZSM-22 (red): c) The
n-heptane conversion against the reaction temperature; d) The yields of total C7 isomers (solid lines with solid circles), dibranched C7 isomers (solid lines with
open triangles) and cracking products (dashed lines with open squares).
The Pt-H-ZSM-22 was subsequently mixed with boehmite to
a composite bifunctional catalyst, with an intimate mixture of
zeolite and -alumina binder (wt/wt = 1). HAADF-STEM image of
the ultramicrotomed sections of Pt-H-ZSM-22/Al2O3 confirms
that Pt particles with an average size of 1.6 nm were still
exclusively located on or inside the zeolite crystals (Figure 2a).
While Pt-Al2O3/H-ZSM-22 with Pt particles (1.9 nm) exclusively
located on the alumina binder provided a nanoscale distance
between Pt and acid sites (Figure 2b). Alumina binder has a
mesoporous structure with a pore diameter of ~14 nm (Table S1
and Figure S4), which is readily accessible for linear and
branched C7 hydrocarbon molecules. The inductively coupled
plasma (ICP) analysis shows that the Pt loading of Pt-H-ZSM-
22/Al2O3 and Pt-Al2O3/H-ZSM-22 were 0.6 wt% and 0.5 wt%,
respectively (Table S2). For the acid function, the NH3-TPD
suggests that these two catalysts have similar acid densities
(Figure S5). The ratios of accessible platinum atoms relative to
acid sites (nPt/nA, mol/mol) for Pt-H-ZSM-22/Al2O3 and Pt-
Al2O3/H-ZSM-22 are 0.17 and 0.12 respectively (Table S2),
which are sufficient to maintain the metal–acid balance needed
for so-called ideal hydrocracking.[6,7] In this work, we have
confirmed that even a loading of 0.05 wt% Pt is sufficient for
catalysis (Figure S6), i.e., the dehydrogenation and
hydrogenation reactions can reach quasi-equilibrium.[16] Overall,
the most significant difference of these two catalysts is the Pt
location. Therefore, we can evaluate the impact of nanoscale
proximity on the hydroisomerization of n-heptane.
The Pt-Al2O3/H-ZSM-22 offered slightly higher activity than
Pt-H-ZSM-22/Al2O3 in hydroconversion of n-heptane (Figure 2c).
The Arrhenius plots (n-heptane conversion <20%) show that the
activation energy of isomerization reaction for Pt-Al2O3/H-ZSM-
22 is 164 kJ/mol, which is higher than 147 kJ/mol for Pt-H-ZSM-
22/Al2O3 (Figure S7). It is acknowledged that isomerization
reaction on acid sites is the rate-determining step for
hydroisomerization of linear alkanes.[17] The difference in
activation energies over Pt-H-ZSM-22/Al2O3 and Pt-Al2O3/H-
ZSM-22 is probably due to the diffusion effects inferred by the
platinum location. The yield of C7 isomers is the most important
performance indicator. The Pt-Al2O3/H-ZSM-22 with a nanoscale
distance between Pt and acid sites offers higher yield of C7
isomers and lower yield of cracking products than the Pt-H-ZSM-
22/Al2O3 with closest proximity (Figure 2d). The main byproducts
are C3 and C4 alkanes from acid cracking. For both catalysts, the
main C7 isomers are monobranched, and the yield of dibranched
isomers is lower than 2% (n-heptane conversion < 80%). From
thermodynamics, the equilibrium composition of dibranched C7
isomers is comparable to that of monobranched within our
reaction temperature range (Figure S8). The formation of
dibranched C7 isomers needs consecutive isomerization of
monobranched C7 isomers (Figure S9a).[18] One should note that
due to the narrow channel structure of H-ZSM-22, isomerization
of n-alkanes over H-ZSM-22 mainly takes place in the pore
mouth region, which is described as pore mouth catalysis.[19]
Therefore, the formation of C7 dibranched isomers is strongly
impeded over H-ZSM-22 catalyst.[20] Overall, control of the
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
proximity between Pt and acid sites in H-ZSM-22 with a
nanoscale distance is beneficial for both activity and product
selectivity.
n-heptane hydroisomerization may include acid cracking,
hydrogenolysis and dimerization cracking.[6c,21] The main
cracking product for both catalysts are propane, butane and iso-
butane. One feature of H-ZSM-22-based catalyst is that it favors
the formation of n-alkanes in cracking reactions, such as n-
butane, n-pentane and n-hexane. For example, the fraction of n-
butane in C4 is higher than 50% over Pt-H-ZSM-22/Al2O3, while
the value over H-ZSM-5 and H-beta based catalysts is lower
than 10%.[22] Furthermore, Pt-H-ZSM-22/Al2O3 catalyst produced
much more cracking product than Pt-Al2O3/H-ZSM-22 catalyst
(Figure 3). Besides, it is noteworthy that Pt-H-ZSM-22/Al2O3
catalyst with closest proximity also produced more methane and
ethane than Pt-Al2O3/H-ZSM-22 catalyst with nanoscale
proximity. The formation of methane and ethane could be an
indicative of hydrogenolysis or the formation of primary
carbenium ions.[5] Choudhury et al. attributed the formation of
methane and ethane to a (s, p) β-scission reaction, which was
described as the acid cracking of secondary alkylcarbenium ions
into primary inside the H-ZSM-22 micropore channels (Figure
S9b).[23] It is only under the condition with severe confinement
inside narrow micropores with high density of acid sites, that the
energetically unfavorable (s, p) β-scission contributes
significantly to the overall product distribution. The formation of
smaller n-alkanes is a proof that cracking reactions over H-ZSM-
22 catalysts take place inside the narrow micropore channels to
a large extent,[23] otherwise the cracking product with ≥4 carbon
numbers should be mostly branched. Therefore, locating Pt sites
(even a fraction of them) inside the H-ZSM-22 crystals imposes
more undesired cracking reactions than Pt-on-binder.
Figure 3. Product selectivity over Pt-H-ZSM-22/Al2O3 (black) at 330 oC with
the n-heptane conversion of 81.6%, and Pt-Al2O3/H-ZSM-22 (red) at 320 oC
with the n-heptane conversion of 78.1%. C8+ denotes hydrocarbon product
with 8 or more carbon numbers.
Figure 3 exhibits the detailed product selectivities over Pt-H-
ZSM-22/Al2O3 and Pt-Al2O3/H-ZSM-22, at the temperature their
yield of C7 isomers reaches the maximum. The side reactions in
Figure 4. SEM images of a) H-MOR-S and b) H-MOR-L. HAADF-STEM images with ultramicrotomy of d) Pt-H-MOR-S/Al2O3 and d) Pt-H-MOR-L/Al2O3 with
closest proximity, e) Pt-Al2O3/H-MOR-S and f) Pt-Al2O3/H-MOR-L with a nanoscale proximity.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 5. Impact of platinum location on n-heptane conversion over Pt-H-MOR-S/Al2O3, Pt-Al2O3/H-MOR-S, Pt-H-MOR-L/Al2O3 and Pt-Al2O3/H-MOR-L. a) and b)
The n-heptane conversion against the reaction temperature. c) and d) The yields of total C7 isomers (solid lines with solid circles), dibranched C7 isomers (solid
lines with open triangles) and cracked products (dashed lines with open squares).
H-MOR based catalysts with 12MR micropores with different
micropore length
S1), the crucial difference is their average diffusion length
through the micropore channels.
Different form Pt-H-ZSM-22/Al2O3 with
a non-uniform
distribution of platinum particles, the HAADF-STEM imaging of
the catalyst sections confirms that Pt particles with diameters of
1.5–1.6 nm were well-distributed across zeolite crystals for Pt-H-
MOR-S/Al2O3 and Pt-H-MOR-L/Al2O3 (Figure 4c, 4d and Table
S2). One could expect that Pt(NH3)42+(aq) ions (0.48 nm) can
easily diffuse into mordenite micropores (0.65 × 0.70 nm) in the
ion-exchange step. The Pt particle size being larger than the
MOR micropores points to a local collapse of micropores upon
particle growth,[15] while platinum particles are still surrounded by
micropores as shown in position 4 of Figure S3. Additionally,
platinum particles were also exclusively deposited on alumina
binder for Pt-Al2O3/H-MOR-S and Pt-Al2O3/H-MOR-L (Figure 4e
and 4f). The average Pt particle sizes of these four samples are
in the range of 1.5–1.8 nm (Table S2). The Pt loadings for these
bifunctional catalysts are also in the range of 0.5–0.6 wt% (Table
S2). NH3-TPD confirms very similar acidities for the four
catalysts (Figure S12). Therefore, these four catalysts with
identical metal function and acid density allows us to evaluate
the impact of nanoscale intimacy between metal and acid sites,
and the effect of crystal size on n-heptane hydroisomerization.
The Pt-H-MOR-S/Al2O3 with closest intimacy between Pt and
acid sites provided similar activity as Pt-Al2O3/H-MOR-S with a
nanoscale distance (Figure 5a). Regarding the yield of desired
C7 isomers, similar with H-ZSM-22-based catalysts, the Pt-
The mordenite structure contains 12MR straight channels of
0.65 × 0.70 nm interconnected by 8MR pores of 0.34 × 0.48 nm.
Because the 8MR channels are inaccessible to molecules larger
than methane, mordenite structure is generally considered as
one-dimensional.[24] H-MOR-S with small crystals has an
average crystallite size of ~200 nm as evidenced from
measuring the long sides of scattered crystals (Figure 4a). The
1D micropore channels of mordenite also run along with c-axis
as H-ZSM-22, but the quantification of micropore length for H-
MOR-S is difficult due to the irregular morphology (Figure 4a).
H-MOR-L with plate-like morphology was synthesized by a
seed-induced method using H-MOR-S as starting materials,
which can also be regarded as an epitaxial growth process on
parent H-MOR-S planes. Thus, the diffusion length of H-MOR-L
(Figure 4b) should be longer than parent H-MOR-S. The fast
Fourier transform (FFT) of HRTEM suggests that the surface of
plate-like crystals is essentially the “a-b” plane, which is
perpendicular to the 12MR channels (Figure S10). The average
micropore length for H-MOR-L is ~300 nm, which is significantly
larger than the average crystallite size of H-MOR-S (~200 nm).
In fact, the actual difference of diffusion length between H-MOR-
S and H-MOR-L should be much more than 100 nm due to the
irregular morphology of H-MOR-S. These two mordenites have
very similar crystallinities and Si/Al ratios (Figure S11 and Table
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
Al2O3/H-MOR-S exhibited higher yield of C7 isomers and lower
yield of cracking products than Pt-H-MOR-S/Al2O3 (Figure 5c).
The yield of dibranched isomers over H-MOR-S-based catalysts
is much higher than H-ZSM-22-based catalysts (Figure 2d and
5c), because the consecutive skeletal isomerization can take
place inside the 12MR micropore channels.[25] Specifically, the
maximum yield of dibranched isomers over Pt-Al2O3/H-MOR-S
was 9.4% at n-heptane conversion lever of ~80%, while this
number is 1.3% over Pt-Al2O3/H-ZSM-22. However, the
dibranched isomers are susceptible to cracking into C3 and C4
hydrocarbons (Figure S9b). The maximum yield of C7 isomers
over Pt-Al2O3/H-MOR-S is 57%, which is significantly lower than
68% over Pt-Al2O3/H-ZSM-22. Locating Pt particles inside the H-
MOR-S crystals (Pt-H-MOR-S/Al2O3) further decreases the yield
to 51%. Therefore, locating Pt particles inside H-MOR-S crystals
is also detrimental for the yield of C7 isomers.
and acid sites are inside mordenite crystals. The long diffusion
channels would bring strong diffusion limitations especially for
branched alkene intermediates, causing severe acid cracking
reaction. The selectivity of cracking product in the range of C3–
C6 over Pt-H-MOR-L/Al2O3 catalyst was two times as high as
that over Pt-Al2O3/H-MOR-L. Interestingly, the maximum yield of
C7 isomers over Pt-Al2O3/H-MOR-L was 55%, which is close to
57% over Pt-Al2O3/H-MOR-S with short diffusion length. Besides,
the product selectivity was almost the same for these two Pt-on-
binder catalysts (Figure S15). Based on these findings, we
believe that in the case of Pt-on-binder catalysts, most
isomerization reactions take place in pore mouth region of
zeolite, even for mordenite with long 12MR micropore channels.
Compared with two H-MOR-S-based catalysts, the activities
of two H-MOR-L-based catalysts are lower (Figure 5a and 5b).
Considering the density of Pt and acid sites are similar for these
four catalysts (Table S2 and Figure S12), the difference in
activity between H-MOR-S and H-MOR-L should be caused by
accessibility of active sites: the acid sites deeply inside the H-
MOR-L crystals do not significantly participate in the
isomerization reaction.[26] It was revealed that the adsorption of
linear alkanes in zeolite is dominated by Van der Waals
interactions, but linear alkenes are additionally stabilized by
formation of a π-complex or alkoxide with BAS with extra
adsorption energies.[27] This phenomenon was experimentally
confirmed by propene-TPD and propane-TPD (Figure S13),
which proved the enrichment of immobile olefinic species in acid
zeolites. The activation energies for H-MOR-L based catalysts
were systematically lower than H-MOR-S based catalysts
(Figure S7), which suggests mass transfer limitations over
zeolites with large crystals.[9,28] The apparent turnover frequency
(TOF) based on BAS was also lower over H-MOR-L based
catalysts than H-MOR-S based catalysts (Figure S14), indicating
that the effective utilization of acid sites was influenced by the
diffusion length of zeolite crystals. The lower activity of Pt-
Al2O3/H-MOR-L in comparison to Pt-H-MOR-L/Al2O3 is not
understood well, but is probably due to the long diffusion path
length between platinum and acid active sites for the former
catalyst in line with the conventional Weisz’s criterion. It is noted
that at low reaction rates, i.e. at low conversion levels, locating
platinum inside H-MOR-L crystals was found to be beneficial for
TOFBAS (Figure S14).
Figure 6. The impact of platinum location on product distribution. Product
selectivity over Pt-H-MOR-L/Al2O3 (black) at 290 oC with the n-heptane
conversion of 69%, and Pt-Al2O3/H-MOR-L (red) at 305 oC with the n-heptane
conversion of 74%. C8+ denotes hydrocarbon product with 8 or more carbon
numbers. Catalytic behaviors at identical temperatures are listed in Table S3.
Reaction mechanisms
We have investigated the nanoscale intimacy in 1D zeolite-
based bifunctional catalysts for n-heptane hydroisomerization.
Our findings demonstrate that the intimacy between Pt and acid
sites, especially at the nanoscale, has a significant influence in
catalytic activity and selectivity. Specifically, when using H-ZSM-
22 with 10MR channels, controllably depositing Pt particles
exclusively on alumina binder is beneficial for both catalytic
activity and yield of C7 isomers. Most of C7 isomers are
monobranched due to the mechanism of pore-mouth catalysis.
The diffusion of both n-heptane and iso-alkenes to platinum sites
is fast when platinum is located on the alumina binder (Figure
7a). Once the n-heptenes are isomerized in the pore mouths of
the zeolite, it is faster for iso-heptenes to diffuse out of the
channels and back to platinum sites for hydrogenation than to
diffuse into the H-ZSM-22 channels for further isomerization and
cracking. For Pt-H-ZSM-22/Al2O3 catalyst, the spatial distribution
of platinum over Pt-H-ZSM-22 is non-uniform with around 25%
of platinum particles inside the crystals. The Pt particles on the
The mordenite-based catalysts produced almost no methane
o
and ethane (<0.2%) at even 320 C, and most C4–C6 products
were iso-alkanes from acid cracking (Figure 6). This distribution
suggests that β-scission reactions with tertiary-to-secondary and
secondary-to-tertiary modes are favored over mordenite
catalysts (Figure S9b). In addition, the formation of C5, C6 and
together with C8+ (without formation of methane and ethane)
indicates that dimerization cracking also contributes to a low
extent. Regarding the impact of proximity, Pt-H-MOR-L/Al2O3
catalyst with Pt sites inside the zeolite crystals displays a
maximum yield of C7 isomers as low as 38% (Figure 5d). In the
case of bifunctional catalyst with Pt sites inside H-MOR-L
crystals, conversion of n-heptane has to take place inside the
narrow micropores due to the fact that the majority of Pt sites
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 7. A Schematic presentation of the impact of spatial organization of platinum in or out zeolites on n-heptane hydroisomerization. a) H-ZSM-22-based
catalysts with different proximity between Pt and acid sites. b) Mordenite-based catalysts with different proximity between Pt and acid sites.
external surface of H-ZSM-22 crystals can still catalyze the
(de)hydrogenation, acting the same as Pt particles on the
alumina binder. We propose that platinum particles inside H-
ZSM-22 crystals brings down the yield of C7 isomers, because n-
heptane feedstock can diffuse relatively rapidly inside its
channels and is converted, giving rise to secondary cracking
reactions due to diffusion constraints (Figure 7a). The formation
of linear butane and pentane over H-ZSM-22 catalysts,
especially with platinum particles inside zeolite crystals, strongly
suggests that a significant fraction of cracking reactions take
place inside zeolite channels.
In the case of mordenite-based bifunctional catalysts, the
yield of dibranched isomers for mordenite-based catalysts is
generally higher than that for H-ZSM-22-based catalysts. The
activity is not affected whether the Pt particles reside on alumina
binder or inside the 12MR channels when using small mordenite
crystals. Different with Pt-H-ZSM-22, the vast majority of
platinum particles for Pt-H-MOR-S was located inside zeolite
crystals, because the micropore size of mordenite is larger than
Pt(NH3)42+(aq) ions. The n-heptane feedstock and alkene
intermediates can fast diffuse to active sites for
(de)hydrogenation and isomerization (Figure 7b). However,
locating Pt particles inside the zeolite crystals with closest
intimacy produces more acid cracking products than catalyst
with Pt-on-binder. This detrimental influence is intensified when
using mordenite with long diffusion length. The similarity in
product selectivity for the two Pt-on-binder catalysts with distinct
crystal sizes suggests that the isomerization reaction over
mordenite also follows the general pore mouth catalysis
mechanism.
cracking reaction rather than locating platinum on alumina
binder in hydroconversion of C10–C19 paraffins.[10] Recently Ben
Moussa et al. also reported a nanosized beta-zeolite system that
locating platinum on alumina binder outperformed embedding
platinum particles inside crystals of beta-zeolite in n-heptane
isomerization at atmospheric pressure.[11b] Therefore, the
influence of nanoscale intimacy still exists when using 3D
crystals albeit in interplay with the reactant molecular size.
To further investigate the scope of nanoscale proximity on n-
heptane hydroisomerization, we evaluated the effect of platinum
location over H-ZSM-5 based catalysts with a 3D micropore
network. H-ZSM-5 consisting of aggregates of ~40 nm crystals
was used to construct bifunctional catalysts to minimize the
diffusion limitations (Figure S16). We have controlled the
platinum location either inside H-ZSM-5 crystals or on the
alumina binder (Figure S17). The catalytic performances show
that the location of platinum particles did not affect the activity
and iso-C7 yield for n-heptane hydroisomerization (Figure S18).
However, it is noteworthy that the iso-butane to n-butane ratio
over Pt-H-ZSM-5/Al2O3 was significantly lower than that over Pt-
Al2O3/H-ZSM-5 (Figure S18c), indicating that the energetically
unfavorable secondary-to-secondary cracking reactions (Figure
S9b) tend to take place more extensively inside H-ZSM-5
micropores. Therefore, still the platinum location has an
influence on product selectivities for H-ZSM-5 with a 3D pore
structure. Vandegehuchte et al. reported that in n‑hexane
hydroconversion, branched isomers would suffer from strong
intracrystalline diffusion limitations in H-ZSM-5 crystallites, but
the limitations could be mitigated by decreasing the size of H-
ZSM-5 crystals.[29] In fact, to decrease the effective crystal sizes
of zeolites and build hierarchical structure has often been
employed to increase the isomer selectivity over H-MOR and H-
beta catalysts without considering platinum location.[30] We think
Our previous study based on Y zeolite (crystal size 0.5–1.0
μm) with a 3D micropore network and 12MR channels showed
that locating platinum particles inside zeolite crystals favored
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
[3]
[4]
a) J. H. Sinfelt, Adv. Chem. Eng. 1964, 5, 37–74; b) L. Liu, U. Díaz, R.
Arenal, G. Agostini, P. Concepción, A. Corma, Nat. Mater. 2016, 16,
132–138.
that the detrimental effect on product selectivity imposed by
locating platinum particles inside the zeolite crystals in
hydroisomerization reactions is generalizable for zeolites with
1D pore systems and 3D zeolites with long diffusion channels.
a) M. Andersen, A. J. Medford, J. K. Nørskov, K. Reuter, Angew. Chem.
Int. Ed. 2016, 55, 5210–5214; M. Andersen, A. J. Medford, J. K.
Nørskov, K. Reuter, Angew. Chem. 2016, 128, 5296–5300; b) E. L.
Margelefsky, A. Bendjériou, R. K. Zeidan, V. Dufaud, M. E. Davis, J.
Am. Chem. Soc. 2008, 130, 13442–13449; c) E. Iglesia, D. G. Barton, J.
A. Biscardi, M. J. L. Gines, S. L. Soled, Catal. Today 1997, 38, 339–
360.
Conclusion
In summary, we succeeded in locating platinum particles either
on the alumina binder or inside 1D zeolite crystals with clear
TEM-based characterization using ultramicrotomy. The Pt-on-
binder catalyst provided a nanoscale proximity between platinum
particles and acid sites, while catalyst with platinum sites inside
zeolite crystals offered the closest proximity. The catalysis
results suggest that embedding the platinum particles imposes
undesired cracking reactions inside the narrow micropore
channels, so the closest proximity between platinum and acid
sites is detrimental for the product selectivity. The extent of
influence of platinum location is strongly dependent on zeolite
dimensionality, size of molecules and reaction conditions. In the
context of numerous efforts being made to locate metal particles
inside the zeolites, our findings suggest that locating metal
particles on alumina binder may bring many benefits for
catalysis. Earlier this was demonstrated for bulky feedstock
molecules while we have now shown this to hold for light
reactant molecules too.
[5]
[6]
J. Weitkamp, ChemCatChem 2012, 4, 292–306.
a) G. Noh, Z. Shi, S. I. Zones, E. Iglesia, J. Catal. 2018, 368, 389–410;
b) M. Guisnet, F. Alvarez, G. Giannetto, G. Perot, Catal. Today 1987, 1,
415–433; c) N. Batalha, L. Pinard, C. Bouchy, E. Guillon, M. Guisnet, J.
Catal. 2013, 307, 122–131.
[7]
[8]
[9]
N. Batalha, L. Pinard, Y. Pouilloux, M. Guisnet, Catal. Lett. 2013, 143,
587–591.
a) P. B. Weisz, E. W. Swegler, Science 1957, 126, 31–32; b) P. B.
Weisz, Adv. Catal. 1962, 13, 137–190.
M. Höchtl, A. Jentys, H. Vinek, J. Catal. 2000, 190, 419–432.
[10] J. Zecevic, G. Vanbutsele, K. P. de Jong, J. A. Martens, Nature 2015,
528, 245–248.
[11] a) J. E. Samad, J. Blanchard, C. Sayag, C. Louis, J. R. Regalbuto, J.
Catal. 2016, 342, 203–212; b) O. Ben Moussa, L. Tinat, X. Jin, W.
Baaziz, O. Durupthy, C. Sayag, J. Blanchard, ACS Catal. 2018, 8,
6071–6078.
[12] a) A. Chica, A. Corma, P. J. Miguel, Catal. Today 2001, 65, 101–110;
b) D. Verboekend, K. Thomas, M. Milina, S. Mitchell, J. Perez-Ramirez,
J. P. Gilson, Catal. Sci. Technol. 2011, 1, 1331–1335; c) J. A. Martens,
D. Verboekend, K. Thomas, G. Vanbutsele, J. P. Gilson, J. Pérez-
Ramírez, ChemSusChem 2013, 6, 421–425; d) S. Liu, J. Ren, H.
Zhang, E. Lv, Y. Yang, Y. W. Li, J. Catal. 2016, 335, 11–23; e) N. Li, F.
Jiao, X. Pan, Y. Chen, J. Feng, G. Li, X. Bao, Angew. Chem. Int. Ed.
2019, 58, 7400–7404; N. Li, F. Jiao, X. Pan, Y. Chen, J. Feng, G. Li, X.
Bao, Angew. Chem. 2019, 131, 7478–7482.
Acknowledgements
This work has been supported by BP and Topsector Chemie
consortia for Knowledge and Innovation (TKIs). K.P.d.J.
acknowledges funding from the European Research Council, EU
FP7 ERC Advanced Grant no.338846. H.Y. thanks the Program
for Advancing Strategic International Networks to Accelerate the
Circulation of Talented Researchers by JSPS. Nikolaos
Nikolopoulos, Remco Dalebout, Miguel Rivera Torrente and
Silvia Zanoni are acknowledged for measuring and analyzing the
N2 physisorption data.
[13] B. Smit, T. L. M. Maesen, Chem. Rev. 2008, 108, 4125–4184.
[14] M. Schreier, S. Teren, L. Belcher, J. R. Regalbuto, J. T. Miller,
Nanotechnology 2005, 16, S582−S591.
[15] J. Zečević, A. M. J. van der Eerden, H. Friedrich, P. E. de Jongh, K. P.
de Jong, ACS Nano 2013, 7, 3698–3705.
[16] E. G. Acebo, C. Leroux, C. Chizallet, Y. Schuurman, C. Bouchy, ACS
Catal. 2018, 8, 6035−6046.
[17] a) Y. Ono, Catal. Today 2003, 81, 3–16; b) A. Holló, J. Hancsók, D.
Kalló, Appl. Catal. A 2002, 229, 93–102; c) G. F. Froment, Catal. Today
1987, 1, 455–473.
[18] V. M. Akhmedov, S. H. Al-Khowaiter, Catal. Rev. 2007, 49, 33–139.
[19] a) J. A. Martens, W. Souverijns, W. Verrelst, R. Parton, G. F. Froment,
P. A. Jacobs, Angew. Chem. Int. Ed. 1995, 34, 2528–2530; J. A.
Martens, W. Souverijns, W. Verrelst, R. Parton, G. F. Froment, P. A.
Jacobs, Angew. Chem. 1995, 107, 2726–2728; b) W. Souverijns, J. A.
Martens, G. F. Froment, P. A. Jacobs, J. Catal. 1998, 174, 177–184; c)
C. S. Laxmi Narasimhan, J. W. Thybaut, G. B. Marin, J. A. Martens, J.
F. Denayer, G. V. Baron, J. Catal. 2003, 218, 135–147.
Keywords: Bifunctional catalysis • hydroisomerization •
proximity • one-dimensional zeolite • shape selectivity
[1]
a) B. Peng, Y. Yao, C. Zhao, J. A. Lercher, Angew. Chem. Int. Ed. 2012,
51, 2072–2075; B. Peng, Y. Yao, C. Zhao, J. A. Lercher, Angew. Chem.
2012, 124, 2299–2299; b) D. M. Alonso, J. Q. Bond, J. A. Dumesic,
Green Chem. 2010, 12, 1493–1513; c) F. Jiao, J. Li, X. Pan, J. Xiao, H.
Li, H. Ma, M. Wei, Y. Pan, Z. Zhou, M. Li, S. Miao, J. Li, Y. Zhu, D. Xiao,
T. He, J. Yang, F. Qi, Q. Fu, X. Bao, Science 2016, 351, 1065–1068; d)
C. Zhou, J. Shi, W. Zhou, K. Cheng, Q. Zhang, J. Kang, Y. Wang, ACS
Catal. 2020, 10, 302−310; e) J. Kim, S. W. Han, J. C. Kim, R. Ryoo,
ACS Catal. 2018, 8, 10545−10554.
[20] J. A. Martens, G. Vanbutsele, P. A. Jacobs, J. Denayer, R. Ocakoglu, G.
Baron, J. A. Muñoz Arroyo, J. Thybaut, G. B. Marin, Catal. Today 2001,
65, 111–116.
[21] F. J. M. M. de Gauw, J. van Grondelle, R. A. van Santen, J. Catal. 2001,
204, 53–63.
[22] a) Z. B. Wang, A. Kamo, T. Yoneda, T. Komatsu, T. Yashima, Appl.
Catal. A 1997, 159, 119–132; b) J. Kim, W. Kim, Y. Seo, J. C. Kim, R.
Ryoo, J. Catal. 2013, 301, 187–197.
[2]
a) K. P. de Jong, Science 2016, 351, 1030–1031; b) K. Cheng, B. Gu,
X. Liu, J. Kang, Q. Zhang, Y. Wang, Angew. Chem. Int. Ed. 2016, 55,
4725–4728; K. Cheng, B. Gu, X. Liu, J. Kang, Q. Zhang, Y. Wang,
Angew. Chem. 2016, 128, 4803–4806; c) J. Li, Y. He, L. Tan, P. Zhang,
X. Peng, A. Oruganti, G. Yang, H. Abe, Y. Wang, N. Tsubaki, Nat. Catal.
2018, 1, 787–793.
[23] I. R. Choudhury, K. Hayasaka, J. W. Thybaut, C. S. Laxmi Narasimhan,
J. F. Denayer, J. A. Martens, G. B. Marin, J. Catal. 2012, 290, 165–176.
[24] C. E. Webster, A. Cottone, R. S. Drago, J. Am. Chem. Soc. 1999, 121,
12127–12139.
[25] F. Jiménez-Cruz, G. C. Laredo, Fuel 2004, 83, 2183–2188.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
[26] C. Pagis, F. Meunier, Y. Schuurman, A. Tuel, M. Dodin, R. Martinez-
Franco, D. Farrusseng, ChemCatChem 2018, 10, 4525–4529.
[27] a) B. A. De Moor, M. F. Reyniers, G. B. Marin, Phys. Chem. Chem.
Phys. 2009, 11, 2939–2958; b) R. J. Correa, C. J. A. Mota, Phys. Chem.
Chem. Phys. 2002, 4, 375–380.
[28] G. Ye, Y. Sun, Z. Guo, K. Zhu, H. Liu, X. Zhou, M. O. Coppens, J. Catal.
2018, 360, 152–159.
[29] a) B. D. Vandegehuchte, J. W. Thybaut, G. B. Marin, Ind. Eng. Chem.
Res. 2014, 53, 15333–15347; b) B. D. Vandegehuchte, I. R. Choudhury,
J. W. Thybaut, J. A. Martens, G. B. Marin, J. Phys. Chem. C 2014, 118,
22053–22068.
[30] a) J. Pastvova, D. Kaucky, J. Moravkova, J. Rathousky, S. Sklenak, M.
Vorokhta, L. Brabec, R. Pilar, I. Jakubec, E. Tabor, P. Klein, P. Sazama,
ACS Catal. 2017, 7, 5781–5795; b) F. Bauer, K. Ficht, M. Bertmer, W.
D. Einicke, T. Kuchling, R. Gläser, Catal. Sci. Technol. 2014, 4, 4045–
4054; c) M. Tromp, J. A. van Bokhoven, M. T. Garriga Oostenbrink, J.
H. Bitter, K. P. de Jong, D. C. Koningsberger, J. Catal. 2000, 190, 209–
214.
This article is protected by copyright. All rights reserved.

10.1002/anie.201915080
Angewandte Chemie International Edition
RESEARCH ARTICLE
RESEARCH ARTICLE
Bifunctional catalysts with platinum
particles located on the alumina
binder, which separates platinum and
acid sites at a nanoscale distance,
offers higher yield of desired isomers
than catalysts with platinum particles
confined inside the zeolite crystals.
Kang Cheng, Lars I. van der Wal, Hideto
Yoshida, Jogchum Oenema, Justine
Harmel, Zhaorong Zhang, Glenn Sunley,
Jovana Zečević, Krijn P. de Jong*
Page No. – Page No.
Impact of the Spatial Organization of
Bifunctional Metal-Zeolite Catalysts
for Hydroisomerization of Light
Alkanes
This article is protected by copyright. All rights reserved.