Hierarchical high-silica zeolites as superior base catalysts

Article abstract of DOI:10.1039/c3sc51937f

For more than four decades, the design of zeolite base catalysts has relied on the application of aluminium-rich frameworks exchanged with alkali metal cations (preferably Cs⁺). However, moderate activity associated with access and diffusion limitations, and high manufacturing costs associated with high caesium content (typically over 30%) have hampered their industrial implementation so far. Herein, we have discovered that high-silica USY zeolites outperform their Al-rich counterparts in a variety of base-catalysed reactions of relevance in the fine chemical industry, as well as in the upgrading of biofuels. The benefits of this class of materials are amplified upon the alleviation of diffusion constraints through the introduction of a network of intracrystalline mesopores by post-synthetic modification. For example, the resulting cation-free hierarchical USY provides an up to 30-fold Knoevenagel condensation activity compared to the benchmark Cs-X, and similar observations were made upon application in liquid-phase (nitro)aldol reactions. Moreover, in the gas-phase aldol condensation of propanal, high-silica zeolites provide superior activity, selectivity, and lifetime compared to caesium-containing zeolites and even a strong solid base such as MgO. We decouple the complex interplay between mesoporosity and intrinsic zeolitic properties such as crystallinity, and quantify the increase in catalyst effectiveness upon hierarchical structuring as a function of reactant size. The obtained results are a major step to resolve the drawbacks of zeolites catalysis and thereby revitalise their potential for industrial application.

Full text of DOI:10.1039/c3sc51937f

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Hierarchical high-silica zeolites as superior base
catalysts†
Cite this: Chem. Sci., 2014, 5, 677
Tobias C. Keller, Stephane Isabettini, Danny Verboekend, Elodie G. Rodrigues
´
and Javier Perez-Ramırez*
´
´
For more than four decades, the design of zeolite base catalysts has relied on the application of aluminium-
+
rich frameworks exchanged with alkali metal cations (preferably Cs ). However, moderate activity
associated with access and diffusion limitations, and high manufacturing costs associated with high
caesium content (typically over 30%) have hampered their industrial implementation so far. Herein, we
have discovered that high-silica USY zeolites outperform their Al-rich counterparts in a variety of base-
catalysed reactions of relevance in the fine chemical industry, as well as in the upgrading of biofuels. The
benefits of this class of materials are amplified upon the alleviation of diffusion constraints through the
introduction of a network of intracrystalline mesopores by post-synthetic modification. For example, the
resulting cation-free hierarchical USY provides an up to 30-fold Knoevenagel condensation activity
compared to the benchmark Cs–X, and similar observations were made upon application in liquid-phase
(nitro)aldol reactions. Moreover, in the gas-phase aldol condensation of propanal, high-silica zeolites
provide superior activity, selectivity, and lifetime compared to caesium-containing zeolites and even a
strong solid base such as MgO. We decouple the complex interplay between mesoporosity and intrinsic
zeolitic properties such as crystallinity, and quantify the increase in catalyst effectiveness upon
hierarchical structuring as a function of reactant size. The obtained results are a major step to resolve
the drawbacks of zeolites catalysis and thereby revitalise their potential for industrial application.
Received 10th July 2013
Accepted 15th October 2013
DOI: 10.1039/c3sc51937f
www.rsc.org/chemicalscience
of this costly metal cation. Due to a severe underutilisation of
the active centers in the micropores resulting from diffusion
Introduction
Due to their outstanding selectivity in ne-chemistry related and access constraints, the catalysts showed only modest
reactions, the application of zeolites as base catalysts has been a activity, making their industrial application unfeasible. To
1
topic of research for over forty years. Their mild basicity resolve this undesirable situation, alternative pathways to
rendered them attractive alternatives in aldol-type condensa- prepare basic zeolites were devised, targeting either an
2
tion reactions, where stronger bases promote the formation of enhancement of the basicity or independence from alkali
3
undesired side-products. However, despite this promising metals.
perspective, zeolites were unable to compete with industrially
Through the deposition of alkali metal oxide clusters by
applied homogeneous bases, as their synthesis either involves impregnation and thermal decomposition of precursor salts,
costly materials or results in catalysts of poor stability.
nanocomposites with strong basic sites suitable for demanding
4
7
Traditionally, basic zeolites were obtained through an ion- gas-phase applications were prepared. However, the latter
exchange of framework charge balancing cations with alkali materials featured an even higher alkali metal loading, and,
5
5c
metal cations. The basicity of these zeolites was correlated to moreover, they proved unstable under ambient conditions.
the Sanderson electronegativity of the framework oxygen atoms, Through the substitution of framework oxygen atoms by more
increasing with the aluminium content of the framework and basic nitrogen in a high-temperature treatment with ammonia,
6
the size of the cations. Accordingly, the most basic zeolites an inexpensive pathway to alkali metal independent basic
8
combined high aluminium loadings with large alkali metals zeolites was discovered. Although active, the tendency of
such as caesium, resulting in catalysts containing over 30 wt% nitrided zeolites to undergo hydrolysis inhibited their repeated
application in condensation reactions, where a stoichiometric
8c
amount of water is formed. Other approaches, i.e. surface
derivatisation with organic bases or incorporation of germa-
nium into the zeolite resulted in good catalytic performance,
Institute for Chemical and Bioengineering, Department of Chemistry and Applied
Biosciences, ETH Zurich, Wolfgang-Pauli-Strasse 10, CH-8093 Zurich, Switzerland.
E-mail: jpr@chem.ethz.ch; Fax: +41 44 6331405; Tel: +41 44 6337120
†
Electronic supplementary information (ESI) available: Suppliers and extended but questionable industrial appeal due to prohibitive costs and
characterisation of parent and hierarchical zeolites, experimental conditions,
9
problematic stability.
and methods. See DOI: 10.1039/c3sc51937f
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 677–684 | 677

View Article Online
Chemical Science
Edge Article
Recent efforts focused on the preparation of basic zeolites contains less than one-hundreth of the aluminium (and
10
with an auxiliary mesopore network. These hierarchical cata- cations) of zeolite X. While for all other tested zeolites, the
+
+
+
lysts were evaluated in model reactions such as Knoevenagel activity increased with the size of the cation (H < Na < K <
+
condensations, with the aim of quantifying and alleviating Cs ), the high-silica faujasite performed equally well indepen-
diffusion limitations. For both Cs-exchanged and nitrided dent of the cation studied. Even in the protonic form, which was
zeolites, the secondary porosity was identied to be the domi- completely inactive for zeolites with higher aluminium
nant performance descriptor, enabling a more efficient uti- contents, H-USY385 displayed only a slight reduction in activity
lisation of the incorporated alkali metals. However, with the compared to its alkali metal exchanged equivalents.
major disadvantages of high cost and insufficient stability
unresolved, the perspective for zeolite base catalysis remained
grim.
The importance of hierarchical porosity
Through alkaline treatment with tetrapropylammonium
Herein, we describe for the rst time the remarkable
performance of high-silica zeolites in base-catalysed reactions.
Through an in-depth investigation of the importance of zeolitic
properties and mesoporosity as a function of reactant size,
design guidelines are developed. The resulting hierarchical
cation-free catalysts are proven to outperform their alkali
metal–containing counterparts in terms of activity, selectivity,
and lifetime in a broad range of gas- and liquid-phase reactions,
including aldol, nitroaldol and Knoevenagel condensations.
11a
bromide (TPABr) as a pore-directing agent, we introduced a
secondary mesopore network into the USY385 zeolite while
preserving its crystalline framework. Aiming at the identica-
tion of an optimal degree of hierarchy, the NaOH concentration
was tuned in a range of 0.05–0.3 M. Samples subjected to
alkaline treatment were coded “-ATx”, where x represents the
applied concentration of sodium hydroxide. While the external
surface area (Smeso) and the total pore volume (Vpore) increased
with the alkalinity of the solution, the micropore volume (Vmicro
)
was reduced for NaOH concentrations exceeding 0.2 M
(Fig. S1†). This decrease can be attributed to a gradual
amorphisation of the zeolite framework, which is in agreement
with the observations made by X-ray diffraction. For the latter,
Results and discussion
High-silica zeolite base catalysts
Aiming for an optimal utilisation of caesium in the zeolite, we
studied a set of faujasite-type zeolites of a broad compositional
range (Si/Al ¼ 1.4–385) with selected monovalent counter-
the drop in crystallinity can already be observed for treatments
11b
at lower alkalinity due to the smaller crystalline domains.
A
TEM micrograph of a representative sample is given in Fig. 2,
where the coexistence of the crystalline framework and a
secondary mesopore network is clearly visible. Illustrative
examples of nitrogen isotherms and X-ray diffractograms are
given in Fig. S2.† In order to quantify the degree of hierarchical
porosity as combination of micro- and mesopores in a single
descriptor, the hierarchy factor (HF) was calculated according to
+
+
+
+
cations (H , Na , K and Cs ). The samples were coded C-Zx,
where C denotes the cation identity, Z the zeolite name, and x
the Si/Al ratio (e.g. K-USY30 is a USY zeolite with molar Si/Al
ratio 30 in the potassium form). These catalysts were rst
evaluated in the base-catalysed Knoevenagel condensation of
benzaldehyde with malononitrile. In contrast to the expected
importance of utilising zeolites with a high aluminium
11c
eqn (1).
2,5
content, activity was observed to increase with the Si/Al ratio,
levelling out in a plateau at high silicon contents (Fig. 1). The
sample with the lowest aluminium content was identied as the
most active catalyst, a USY zeolite with a Si/Al ratio of 385 that
S
meso
BET
V
micro
HF ¼
(1)
S
V
pore
The HF values, plotted versus the NaOH concentrations in
Fig. 3, achieved a maximum at cNaOH ¼ 0.15 M, before the loss in
micropore volume outweighed the generation of additional
Fig.
1 Conversion of benzaldehyde (XBA) in the Knoevenagel
condensation with malononitrile over alkali metal exchanged faujasite
zeolites with different Si/Al ratios.
Fig. 2 TEM micrograph of the hierarchical Na-USY385-AT0.15.
678 | Chem. Sci., 2014, 5, 677–684
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
Fig.
3 Conversion of benzaldehyde (XBA) in the Knoevenagel
Fig. 4 Kinetic profiles for the conversion of benzaldehyde (XBA) in the
Knoevenagel condensation with malononitrile over different catalysts.
condensation with malononitrile and hierarchy factor (HF) for alkaline-
treated USY385 zeolites versus the NaOH concentration of the
treatment.
especially when the amphoteric nature of zeolites is considered.
external surface. The catalytic performance of the samples in Knoevenagel condensations can be run via a base-catalysed
1
2a
12b
the condensation of benzaldehyde with malononitrile was in enolate mechanism, but also an acid-catalysed enol pathway
12c
good agreement with the hierarchy factor, particularly for the and cooperative acid–base mechanisms have been reported.
samples with high crystallinity (Fig. 3), clearly conrming the Accordingly, it is essential to determine the impact of acidity and
merit of the auxiliary porosity for addressing diffusion limita- basicity on the catalytic activity. A purely acid-mediated mecha-
tions. For the partially amorphised samples however, the nism can be excluded, as the protonic form remains inactive for
observed activity was higher than what would be expected on all zeolites except USY385. Upon alkali metal ion-exchange, the
the basis of the HF. This indicates that a crystalline zeolite acidity of the samples decreases as evidenced by the temperature
framework is benecial, but not essential for having an active programmed desorption (TPD) of ammonia (Fig. S3†), while the
silica-based catalyst. This hypothesis was tested by catalytic catalytic activity is increased. This trend evidences that the
evaluation of sodium exchanged SBA-15, an ordered meso- presence of acid sites has a negative effect on the catalytic
2
ꢀ1
porous silica with an external surface area of ca. 540 m g
activity. As no optimum level of activity for zeolites of interme-
(Table 1). Despite this excess in mesoporosity, the sample per- diate Si/Al ratio was observed, a cooperative acid–base pathway is
formed similarly compared to Na-USY385, (Fig. 4) conrming unlikely and the catalytic performance of the zeolites is expected
the strongly benecial nature of the crystalline structure. A toresult solely from their basicity. On this basis, the performance
proof of concept for other zeolite frameworks was attained by of the H-USY385 zeolite can also be explained. Due to the low
testing hierarchical silicalite-1 (coded Na-Z1000-AT0.2). amount of aluminium in the sample, the detrimental effect of
Although the MFI-framework type of this zeolite was previously the acid sites cannot outweigh the basicity anymore and there-
5
described as poor for base catalysis, it outperformed the fore, the catalytic activity is only slightly reduced.
reference Cs–X catalyst by a factor of 8.
According to the consensus in the literature, the presence of
aluminium increases the electron density on the framework
oxygen atoms, particularly in the presence of so cations, hence
Acidic and basic mechanisms
6
increasing the basicity of the material. As high-silica frameworks
Although the USY zeolites have a higher external surface area do not experience this charge redistribution, their basicity is
5
2 2
than their Al-rich counterparts (Table S1†), their unexpected considered to be minor. The CO uptake measured by CO -TPD
activity raises questions on the type of catalysis observed, conrms this traditional viewpoint, as it could be correlated
Table 1 Physico-chemical properties of selected base catalysts
a
b
2
ꢀ1
c
2
ꢀ1
c
3
ꢀ1
Sample
Si/Al
S
BET (m g
)
S
meso (m g
)
V
micro (cm g )
Cs-X1.4
Cs-Y2.6
Na-USY385
Na-USY385-AT0.2
Na-Z1000-AT0.2
Na-SBA-15
1.4
2.6
385
—
1060
N
365
468
700
890
456
939
14
8
109
267
222
542
0.18
0.25
0.29
0.32
0.12
0.20
a
b
c
Molar Si/Al ratio determined by ICP-OES. BET-method. t-plot method.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 677–684 | 679

View Article Online
Chemical Science
Edge Article
2
linearly with the caesium content (Fig. S4†). If the CO -based the basicity requirements by reacting aldehydes of a distinct
descriptor is, however, plotted against the conversion of benzal- size with the identical activated methylene compound. Two
dehyde in the Knoevenagel condensations with malononitrile, a catalysts were tested, the parent Na-USY385 zeolite and the
linear correlation with a negative slope is obtained, highlighting Na-USY385-AT0.2. The initial rate of the hierarchical sample
that the traditional understanding of zeolite base catalysts fails to (robs,hierarchical) was then normalised by the initial rate
identify the potential of high-silica zeolites. Considering the high observed for the conventional zeolite (robs,parent), yielding the
performance of the USY385 zeolites also in more demanding descriptor a (eqn (2)). This parameter can be regarded as an
reactions (vide infra), it is unlikely that their basicity originates activity amplication factor due to the presence of hierar-
from the bridging oxygen atoms of silica materials. This is rein- chical porosity. Assuming that no deactivation occurs, the
forced by the comparison of the Cs-USY15 zeolite with the nitrided reaction rates can be expressed as the product of the intrinsic
9
b
ni
i
USY15-sample reported in an earlier contribution. Despite the rate coefficient (k ), the weighted reactant concentrations ðc Þ
r
incorporationofnitrogenintotheframework, resultinginstronger and the effectiveness factor (h ).
i
Q
basicity, the performance of the two samples is comparable at
n
i
k
r
c h
i
hierarchical
X
BA ¼ 50%. Therefore, a complementary source of basicity is
r
obs;hierarchical
i
h
hierarchical
parent
a ¼
¼
Q
¼
(2)
ni
expected to exist in zeolites that is predominantly present in high-
silica materials and likely not connected to the framework oxygen
atoms.
r
obs;parent
k
r
c hparent
h
i
i
As the basic character of the zeolite remains preserved and
the reaction conditions remain constant over both catalysts, the
expression for a simplies to the ratio of the effectiveness
factors and quanties the relative increase in catalyst effec-
tiveness through the introduction of a secondary mesopore
network. The size of the molecules was estimated through a
kinetic diameter equivalent, the detailed determination proce-
dure is given in Fig. S6.† A broad range of aldehydes were tested
in the Knoevenagel condensation with malononitrile (Table
S2†), ranging from butanal (dkin ¼ 0.31 nm) to 9-phenan-
threnealdehyde (dkin ¼ 0.56 nm). According to the kinetic
diameter, all reactants and products t the FAU micropores
Nature of the active site
13a
In 1985, Chester et al. observed a cation-exchange capacity of
high-silica zeolites that exceeded their aluminium content,
suggesting the presence of anionic sites in zeolites. These
negatively charged species were postulated to be siloxy groups,
and their occurrence was linked to defects in the zeolite
framework. Through the low-pressure adsorption of alkenes,
13b
Huon et al. proved their basic character, and their struc-
1
ture was suggested on the basis of H-NMR to be deprotonated
silanol groups stabilised by tetrahedrally coordinated neigh-
15
13c
(0.74 nm), so the reactions are expected to run under diffusion
control rather than access control. Even for very small
substrates like butanal, a was greater than unity and increased
with the kinetic diameter up to an impressive maximum of 4.5
bouring silanol groups.
These sites are apparently also
formed in amorphous silica materials and were reported in as-
synthesised mesostructured silica in the presence of cationic
13d,e
14
templates.
Indeed, Kubota et al. reported the high activity
(
Fig. 5). The observed trend highlights the key impact of diffu-
of uncalcined, template-containing molecular sieves in
comparable reactions. The performance could, however, not be
conrmed for calcined materials, which could imply a co-
catalytic role of the incorporated ammonium species. The
superior activity of zeolites is likely the result of their sharply
dened tetrahedral coordination of silicon atoms, compared to
the broad distribution observed for amorphous silicas such as
SBA-15 (Fig. S5†), that could lead to a pronounced presence of
siloxy groups in these crystalline materials. The combined
effects of a low amount of active sites (ca. 1–1.5 sites per unit
sion and enables us for the rst time to quantify the importance
of hierarchical porosity in zeolites as a function of the reactant
size. Although the kinetic diameter is a simple, readily obtain-
able descriptor, it succeeds in estimating the impact of hierar-
chical porosity even for molecules of fundamentally different
10a
cell) and strong access or diffusion constraints explain why
conventional (bulk) techniques fail to identify the catalytic
potential of silica-based materials in base-catalysed reactions
(Fig. S4†). Accordingly, defect-sensitive characterisation
methods are required to conrm the identity of the active site,
such as the spectroscopic investigation of adsorbed acidic
probe molecules by IR or NMR, supported by computational
methods.
Impact of reactant size on diffusivity
With the purpose of quantifying the impact of diffusion
limitations, the importance of hierarchical porosity as a
function of reactant size was studied. To achieve this, the
Fig. 5 Activity amplification through the introduction of hierarchy (a)
for Knoevenagel condensations with malononitrile as a function of the
steric demand of the reaction was altered without changing kinetic diameter (dkin) of the aldehyde.
680 | Chem. Sci., 2014, 5, 677–684
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
nature such as benzaldehyde and cyclohexylcarbaldehyde.
However, if a very accurate estimation is desired, a more
detailed descriptor accounting for exibility and sphericity
needs to be developed.
Versatility in aldol, nitroaldol, and Knoevenagel reactions
Liquid-phase condensations. To assess the strength of the
basic sites, Knoevenagel condensations with less acidic
substrates than malononitrile (pKa,DMSO
¼
11.1) were
1
6
investigated. As all of the reactions can be carried out
with high selectivity over zeolites (>95%), the strength of
the active sites can be determined by comparison of the
3
Fig. 7 Conversion and yield in the aldol condensation of benzalde-
hyde and heptanal to jasminaldehyde and the nitroaldol condensation
of nitroethane with benzaldehyde over the conventional Cs–X and the
hierarchical Na-USY385-AT0.2 catalysts.
reaction rates. In the investigated reactions, we aimed to
benchmark the catalytic performance of a hierarchical high-
silica sample with the conventional Cs–X catalyst. While for
ethyl cyanoacetate (ECA), the high-silica sample out-
performed its Cs-containing counterpart (Fig. 6), the
conversion levels were strongly decreased over both cata-
lysts for ethyl acetoacetate (EAA) and diethyl malonate
self-condensation of heptanal. For the nitroaldol conden-
sation, the major side-product resulted from the subse-
quent addition of nitroethane to the reaction product,
which was formed to a lower extent over the high silica
zeolite catalyst compared to the caesium exchanged zeolite
X. These results conrm that the application of high-silica
zeolites in condensation reactions is feasible, however,
differences in selectivity can occur compared to the tradi-
tional zeolites. This altered selectivity, which could origi-
nate from differences in basicity, acidity, and micropore
size, needs further investigation if an accurate under-
standing is desired.
(
DEM) despite an increase in reaction temperature. As a
consequence, diffusional constraints became less important
and the two catalysts performed at a comparable level.
These results highlight that the presence of aluminium is
not necessary also for more challenging reactions. Accord-
ingly, the application of high-silica zeolites in ne chem-
istry related reactions such as aldol condensations and
nitroaldol condensations appears promising. In this
contribution, two reactions were investigated as a proof of
concept: the aldol condensation of benzaldehyde and hep-
tanal to form jasminaldehyde, which is used as a fragrance
3
Gas-phase aldol condensation of propanal. To demonstrate
the potential of the high-silica zeolites in gas-phase applica-
tions, the hierarchical Na-USY385 samples were tested in the
self-condensation of propanal. This reaction is a model
system for the catalytic deoxygenation of pyrolysis product
streams, where a reduction in the oxygen content of the
obtained bio-oil through catalytic condensation reactions of
the aldehyde fraction (3–18 wt%) is desired. The tests were
in the perfume industry, as well as the nitroaldol
condensation of nitroethane with benzaldehyde. In both
reactions, the hierarchical high-silica USY catalyst clearly
outperformed the caesium-exchanged zeolite X (Fig. 7),
conrming the viability of the approach. The selectivity for
jasminaldehyde was slightly reduced over the Na-USY385-
AT0.2 catalyst, the major side-product resulted from the
ꢁ
carried out at 400 C, which is an operating temperature in
1
7a
the range of a pyrolysis outlet stream. The predominant
reaction pathways observed in the system are summarised in
Scheme 1. The aldol condensation of propanal, as well as the
Fig. 6 Conversion of benzaldehyde in the Knoevenagel condensation
with a) ethyl cyanoacetate, b) ethyl acetoacetate, and c) diethyl mal-
onate over conventional Cs–X and hierarchical Na-USY385-AT0.2 Scheme
1
Reaction pathways for the gas-phase conversion of
catalysts.
propanal.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 677–684 | 681

View Article Online
Chemical Science
Edge Article
samples, where coking is expected to occur because of acid-
catalysed cracking reactions. The main reaction product for
the high-silica zeolites was the aldol condensation dimer,
which was formed with high selectivity (up to 80%), while
trimer formation was more pronounced at higher conversion
levels (Fig. 8). While for USY385 zeolites almost no 3-penta-
none or propene formation was observed, the USY15 zeolites
comprising more aluminium showed pronounced formation
of these products. Accordingly, acid sites favour these reaction
1
7c–e
pathways, which is in line with ndings in the literature.
The absence of these products, combined with the excellent
stability of the USY385 zeolite, supports the theory that their
activity does not result from acid–base bifunctional mecha-
nisms, but only from basic sites of moderate strength. In
contrast, for materials with stronger basicity, deactivation
occurs again due to coking, as observed for MgO. The
combination of high activity, selectivity and stability marks
hierarchical high-silica zeolites as outstanding catalysts for
gas-phase condensation reactions, rendering them a prom-
ising candidate for catalytic bio-oil upgrading, as well as
similar condensation reactions.
Fig. 8 Product yields in the gas-phase self-condensation of propanal
over alkaline-treated Na-USY385 and alkali metal ion-exchanged
ꢁ
USY15 zeolites at 400 C after 2 hours on stream.
subsequent cross-condensation of the reaction product with
propanal were the major products detected for the basic
catalysts. In the presence of acid sites, the formation of
3
-pentanone via the Tishchenko reaction and subsequent
1
7b,c
ester ketonisation could be observed,
accompanied by
Conclusions
propene production resulting either from the hydrolysis and
dehydration of the Tishchenko product or from cracking The identication of high-silica zeolites as an alternative to
1
7c–e
reactions.
The parent Na-USY385 zeolite displayed a very traditional Al-rich counterparts revitalizes the industrial
poor activity (Xpropanal ¼ 2%, Fig. 8). Through alkaline treat- potential of these materials in base catalysis. Their efficient
ment however, the catalytic activity aer 2 h on stream utilisation, enhanced through the post-synthetic introduction
increased with the hierarchy factor up to a maximum of of hierarchical porosity, spurs their application in liquid-phase
Xpropanal ¼ 36% at c
¼ 0.2 M. This remarkable perfor- condensation reactions, where hierarchical cation-free zeolites
NaOH
mance is higher than the initial activity of an investigated display up to a 30-fold higher activity than the benchmark Cs–X.
2
ꢀ1
commercial MgO sample (SBET ¼ 153 m g ), and also higher Moreover, their superior activity, selectivity, and lifetime in the
than the activity of alkali metal ion-exchanged USY15 zeolites. gas-phase aldol condensation of propanal compared to
Additionally, the USY385 catalyst showed a high stability, aluminium-containing zeolites and even strong bases such as
whereas the Al-containing zeolites, and also MgO, displayed a MgO highlight this new class of catalysts as promising candi-
continuous deactivation (Fig. 9). This deactivation resulted dates for the upgrading of bio-oil. These results open up a wide
from coking of the catalysts, which was investigated by ther- eld of potential applications.
mogravimetric analysis (TGA, Fig. S7†). The USY385 samples
showed a 5-fold lower amount of coke compared to the USY15
Experimental
Zeolite base catalysts were prepared by the alkaline treatment
and ion exchange of commercial samples. For ion exchange, the
catalysts were subjected to three subsequent treatments in 0.2 M
4
chloride salt solutions (CsCl, KCl, NaCl, NH Cl) of the desired
ꢁ
ꢀ1
cation at 25 C for 8 h with 20 g L of zeolite. Alkaline treatments
were carried out in a solution of 0.05–0.3 M of sodium hydroxide
ꢁ
and 0.2 M of tetrapropylammonium bromide at 65 C for 0.5 h
ꢀ
1
with a zeolite-to-liquid ratio of 33 g L . Aer alkaline treatment,
or to obtain the protonic form of the zeolites from the ammo-
ꢁ
nium ion-exchanged form, zeolites were calcined at 550 C for 5 h
ꢁ
ꢀ1
with a heating rate of 5 C min . Suppliers, compositions,
porous properties, and coding of parent zeolites are listed in
Table S1.† SBA-15 was obtained in calcined form from Zeochem,
MgO (99.5%) was purchased from Strem Chemicals. Before the
ꢁ
catalytic tests, all samples were dried at 100 C in a vacuum oven
Fig. 9 Conversion of propanal (Xpropanal) versus time-on-stream (t) in
ꢁ
ꢁ
the self-condensation reaction over different base catalysts at 400 C. at 25 mbar. Activation at 550 C did not result in an increased
682 | Chem. Sci., 2014, 5, 677–684
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
activity in the Knoevenagel condensation of benzaldehyde and
malononitrile.
Acknowledgements
Powder X-ray diffraction (XRD) patterns were acquired with This work was supported by ETH Zurich (research grant ETH-31
a PANalytical X’Pert PRO-MPD diffractometer using Ni-ltered 13-1) and the Swiss National Science Foundation (project
Cu Ka radiation (l ¼ 0.1541 nm). Data were recorded in the number 200021-134572).
ꢁ ꢁ
q range of 5–70 with an angular step size of 0.05 and a
2
counting time of 8 s per step. The relative crystallinity of the
faujasite samples was determined according to ASTM stan-
dard D3906. Si and Al concentrations in the solids were
determined by inductively coupled plasma optical emission
spectroscopy (ICP-OES) by using a Horiba Ultima 2 instru-
ment equipped with photomultiplier tube detection. Nitrogen
Notes and references
1 Y. N. Sidroenko, P. N. Galich, V. S. Gutyrya, V. G. Il’in and
I. E. Neimark, Dokl. Akad. Nauk SSSR, 1967, 173,
132.
2 (a) A. Corma, V. Forn ´e s, R. M. Mart ´ı n-Aranda, H. Garc ´ı a and
J. Primo, Appl. Catal., 1990, 59, 237; (b) M. J. Climent,
A. Corma, S. Iborra and J. Primo, J. Catal., 1995, 151, 60; (c)
M. J. Climent, H. Garcia and J. Primo, Catal. Lett., 1990, 4, 85.
3 M. J. Climent, A. Corma, S. Iborra and A. Velty, J. Mol. Catal.
A: Chem., 2002, 182–183, 327.
ꢁ
sorption at ꢀ196 C was carried out in a Micromeritics
TriStar II instrument. Prior to the measurement, the samples
ꢁ
were evacuated at 300 C for 3 h. The total surface area (SBET
)
was determined by the BET method and the t-plot method
was used to discriminate between micro- and mesoporosity.
The protocols for other applied characterisations are detailed
in the ESI.†
4 (a) K. Tanabe and W. F. H ¨o lderich, Appl. Catal., A, 1999, 181,
399; (b) H. Hattori, Appl. Catal., A, 2001, 222, 247; (c)
R. J. Davis, J. Catal., 2003, 216, 396.
Knoevenagel condensations with malononitrile were
3
carried out in a Radleys Carousel 6+ equipped with 50 cm
5 (a) Y. Ono, J. Catal., 2003, 216, 406; (b) H. Hattori, Chem. Rev.,
1995, 95, 537; (c) A. Corma and S. Iborra, in Fine Chemicals
through Heterogeneous Catalysis, ed. R. A. Sheldon and H.
van Bekkum, Wiley-VCH, New York, 2002, vol. 1, p. 309.
6 D. Barthomeuf, Catal. Rev. Sci. Eng., 1996, 38, 521.
7 (a) P. E. Hathaway and M. E. Davis, J. Catal., 1989, 116, 263;
(b) H. Tsuji, F. Yagi, H. Hattori and H. Kita, Stud. Surf. Sci.
Catal., 1993, 75, 1171.
8 (a) M. Srasra, S. Delsarte and E. M. Gaigneaux, J. Phys. Chem.
C, 2010, 114, 4527; (b) K. D. Hammond, et al., Chem. Mater.,
2010, 22, 130; (c) S. Ernst, M. Hartmann, T. Hecht,
P. Cremades Ja ´e n and S. Sauerbeck, Stud. Surf. Sci. Catal.,
2002, 142, 549.
two-necked round-bottom asks and reux cooling under a
nitrogen atmosphere. In a typical experiment, equimolar
amounts (4 mmol) of aldehyde and malononitrile, decane as
the internal standard (6 mmol), and pre-dried catalyst (50 mg)
3
ꢁ
were reacted in 30 cm of toluene at 80 C for 4 h under time-
resolved sampling. For the screening of alkaline-treated
USY385 zeolites (Fig. 3), the reaction temperature was lowered
ꢁ
to 60 C to reduce the conversion levels. The catalytic tests of
aldehydes with different kinetic diameters were carried out at
room temperature, the reaction time was adapted for each
substrate according to Table S2.† Solvent-free liquid-phase
3
reactions were carried out in Ace pressure tubes (15 cm , front
seal). In
a
typical experiment, an activated methylene
9 (a) X. Wang, K. S. K. Lin, J. C. C. Chan and S. Cheng, J. Phys.
Chem. B, 2005, 109, 1763; (b) A. Corma, R. M. Martin-Aranda
and F. Sanchez, J. Catal., 1990, 126, 192.
compound, aldehyde, internal standard and pre-dried catalyst
were heated to reaction temperature and le to react for the
desired reaction time. Then, the reaction was quenched by 10 (a) G. V. Shanbhag, M. Choi, J. Kim and R. Ryoo, J. Catal.,
cooling and diluted in acetone. The exact reaction conditions
are listed in Table S3.† Samples were analysed using a HP
2009, 264, 88; (b) D. Verboekend, T. C. Keller, S. Mitchell
and J. P ´e rez-Ram ´ı rez, Adv. Funct. Mater., 2013, 23, 1923.
Agilent 6890 gas chromatograph equipped with an HP-5 11 (a) D. Verboekend, G. Vil ´e and J. P ´e rez-Ram ´ı rez, Adv. Funct.
column and an FID detector. Due to the high boiling point,
the conversion of 9-phenanthrenealdehyde was determined by
Mater., 2012, 22, 916; (b) D. Verboekend, S. Mitchell,
M. Milina, J. C. Groen and J. P ´e rez-Ram ´ı rez, J. Phys.
Chem. C, 2011, 115, 14193; (c) J. P ´e rez-Ram ´ı rez,
D. Verboekend, S. Abell ´o and A. Bonilla, Adv. Funct.
Mater., 2009, 19, 3972.
ꢀ
1
ATR investigation of the C]O stretching band at 1695 cm
on a Bruker Vertex 70 spectrometer by coaddition of 96 scans
ꢀ1
with a resolution of 2 cm . Conversion and selectivity for all
reactions are reported with respect to the limiting aldehyde. 12 (a) A. C. Cope, J. Am. Chem. Soc., 1937, 59, 2327; (b)
The gas-phase aldol condensation of propanal was carried out
at atmospheric pressure in a continuous-ow xed-bed reactor
D. T. Mowry, J. Am. Chem. Soc., 1945, 67, 1050; (c)
S. L. Hruby and B. H. Shanks, J. Catal., 2009, 263, 181.
(internal diameter of 12 mm) using 0.3 g of catalyst (particle 13 (a) A. W. Chester, Y. F. Chu, R. M. Dessau, G. T. Kerr and
size ¼ 0.2–0.4 mm). Prior to the reaction, the catalyst was
C. T. Kresge, J. Chem. Soc., Chem. Commun., 1985, 289; (b)
J. R. Huon, D. M. Ruthven and R. P. Danner, Microporous
Mater., 1995, 5, 39; (c) D. F. Shantz, J. S. auf der G u¨ nne,
H. Koller and R. F. Lobo, J. Am. Chem. Soc., 2000, 122,
6659; (d) S. B. Hong and M. A. Camblor, Chem. Mater.,
1997, 9, 1999; (e) D. H. Olson, W. O. Haag and
W. S. Borghard, Microporous Mesoporous Mater., 2000, 35–
36, 435.
ꢁ
ꢁ
ꢀ1
activated at 450 C (heating rate ¼ 10 C min ) for 1 h under
3
ꢀ1
a 50 cm min He ow. Propanal was fed into the reactor
3
ꢀ1
using a syringe pump at 1.2 cm h via a vaporisation line.
Reactions were conducted at 400 C in a 50 cm min He
ow. The reaction mixture was analysed by an on-line HP
ꢁ
3
ꢀ1

Agilent 6890 gas chromatograph equipped with an HP-5
column and an FID detector.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 677–684 | 683

View Article Online
Chemical Science
Edge Article
1
1
1
4 Y. Kubota, et al., Microporous Mesoporous Mater., 2004, 70, 17 (a) M. St ¨o cker, Angew. Chem., Int. Ed., 2008, 47, 9200; (b)
135.
W. Shen, G. A. Tompsett, R. Xing, W. C. Conner Jr and
G. W. Huber, J. Catal., 2012, 286, 248; (c) C. A. Gaertner,
J. C. Serrano-Ruiz, D. J. Braden and J. A. Dumesic, J. Catal.,
2009, 266, 71; (d) X. Zhu, L. L. Lobban, R. G. Mallinson
and D. E. Resasco, J. Catal., 2010, 271, 88; (e) T. Q. Hoang,
X. Zhu, T. Sooknoi, D. E. Resasco and R. G. Mallinson, J.
Catal., 2010, 271, 201.
5 C. Baerlocher and L. B. McCusker, Database of Zeolite
Structures, http://www.iza-structure.org/databases/.
6 (a) W. S. Matthews, et al., J. Am. Chem. Soc., 1975, 97, 7006; (b)
F. G. Bordwell and H. E. Fried, J. Org. Chem., 1981, 46, 4327;
(c) F. G. Bordwell, Acc. Chem. Res., 1988, 21, 456; (d)
W. N. Olmstead and F. G. Bordwell, J. Org. Chem., 1980, 45, 3299.
684 | Chem. Sci., 2014, 5, 677–684
This journal is © The Royal Society of Chemistry 2014