Nanocrystalline ZSM-5: A catalyst with high activity and selectivity for epoxide rearrangement reactions

Article abstract of DOI:10.1016/j.molcata.2009.11.009

Nanocrystalline ZSM-5 (～20-50 nm) has been tested as a catalyst in the liquid phase rearrangement of 1,2-epoxyoctane, 2-methyl-2,3-epoxybutane and isophorone oxide together with a microcrystalline ZSM-5 sample (～5 μm) employed as a reference. Whereas this last material shows a low activity for the three epoxide rearrangement reactions considered, nanocrystalline ZSM-5 leads in all cases to high epoxide conversion. This different catalytic performance is attributed to the differences in the accessibility of the reactant molecules to the catalyst active sites. The smallest crystal size of the nanocrystalline zeolite contributes favourably to both the formation of a high external surface area, with no steric constraints, and to a faster intracrystalline diffusion of the epoxide molecules towards the acid sites located within the zeolite micropores. When sterically hindered epoxides are tested, the catalytic rearrangement occurs mainly on the external surface of the nanocrystalline zeolite. Moreover, it is remarkable that for the three epoxides investigated the high catalytic activity observed over nanocrystalline ZSM-5 is accompanied by the attainment of high selectivities, typically around or over 80%, towards products with commercial applications.

Full text of DOI:10.1016/j.molcata.2009.11.009

Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Nanocrystalline ZSM-5: A catalyst with high activity and selectivity for epoxide
rearrangement reactions
David P. Serranoa,b,∗, Rafael van Griekena, Juan Antonio Meleroa, Alicia Garcíaa, Carolina Vargasa
a Department of Chemical and Environmental Technology, ESCET, Rey Juan Carlos University, c/ Tulipán s/n, 28933, Móstoles, Madrid, Spain
b IMDEA Energía, c/ Tulipán s/n, 28933, Móstoles, Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 June 2009
Received in revised form
12 November 2009
Accepted 12 November 2009
Available online 14 December 2009
Nanocrystalline ZSM-5 (∼20–50 nm) has been tested as a catalyst in the liquid phase rearrangement
of 1,2-epoxyoctane, 2-methyl-2,3-epoxybutane and isophorone oxide together with a microcrystalline
ZSM-5 sample (∼5 ␮m) employed as a reference. Whereas this last material shows a low activity for
the three epoxide rearrangement reactions considered, nanocrystalline ZSM-5 leads in all cases to high
epoxide conversion. This different catalytic performance is attributed to the differences in the accessibility
of the reactant molecules to the catalyst active sites. The smallest crystal size of the nanocrystalline zeolite
contributes favourably to both the formation of a high external surface area, with no steric constraints,
and to a faster intracrystalline diffusion of the epoxide molecules towards the acid sites located within
the zeolite micropores. When sterically hindered epoxides are tested, the catalytic rearrangement occurs
mainly on the external surface of the nanocrystalline zeolite. Moreover, it is remarkable that for the three
epoxides investigated the high catalytic activity observed over nanocrystalline ZSM-5 is accompanied by
the attainment of high selectivities, typically around or over 80%, towards products with commercial
applications.
Keywords:
ZSM-5
Nanocrystalline
Epoxide rearrangement
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Using zeolites as catalysts has the advantage of the existence
of a well-defined pore system, which makes possible the control
Rearrangement of epoxides is a reaction commercially inter-
esting as it allows the synthesis of useful intermediates for the
chemical industry. The activation of epoxides through ring opening
reactions can be catalyzed by Brönsted acid sites, via the addition
of a proton to the epoxide oxygen, as well as by Lewis acid sites,
via the coordination of the oxirane to a multivalent cation. A vari-
ety of products can be formed from those reactions depending on
both the regioselectivity of the ring opening and the presence of
steric constraints [1–3]. Both factors can be also adjusted by the
convenient selection of the catalyst to be employed. A relatively
large number of catalyst types have been used in the literature
for epoxide rearrangement, including both homogeneous and het-
erogeneous systems. Among the latter, zeolites, metal oxides (e.g.
silica and alumina), metal sulfates, and precipitated phosphates are
the preferred ones [4–6]. First works related to aliphatic epoxides
rearrangement over different zeolites were reported by Venuto
and Landis (Na zeolites) [7], Imanaka et al. (HY zeolite) [8], and
Matsumoto et al. (exchanged zeolites) [9,10].
of the product distribution according to shape-selectivity effects.
Likewise, depending on the type of zeolite, the acidic proper-
ties of the catalyst can be adjusted and varied according to the
requirements needed in each reaction. Thus, zeolites in their proton
form (e.g. H-ZSM-5) as well as the weakly Lewis-acidic Ti-silicate
(TS-1) and Ti-Beta have been found to catalyze this type of reac-
tions [11,12]. However, the best results have been obtained with
aluminium-containing zeolites, as it is the case of zeolites X, Y,
ZSM-5, Beta, and mordenite. These materials have been success-
fully tested in the rearrangement of aromatic and cyclic alkane
epoxides, as well as of epoxides with tertiary carbons linked to
the oxirane ring. Thus, high selectivities towards aldehydes have
been obtained in the isomerisation of styrene oxide and derivatives
over various acid zeolites [13,14]. Likewise, the rearrangement of
cyclic ␣,␤-epoxy ketones towards ␤-ketoaldehydes and diketones
has been used for the preparation of pharmaceuticals, synthetic
food flavourings and perfumes [15,16]. Meyer et al. reported the
rearrangement of isophorone oxide with high yields (up to 80%) to
keto aldehyde over zeolitic materials [17]. Likewise, some works
have been published dealing with the use of zeolites as catalysts in
the isomerisation of small size aliphatic epoxides, such as propy-
lene oxide and 3-methyl-2,3-epoxybutane [18,19]. However, few
results have been reported on the rearrangement of long straight-
chain 1,2-alkane epoxides over zeolitic catalysts, probably due to
∗
Corresponding author at: Department of Chemical and Environmental Tech-
nology, ESCET, Rey Juan Carlos University, c/ Tulipán s/n, 28933, Móstoles, Madrid,
Spain. Tel.: +34 91 664 74 50; fax: +34 91 488 70 68.
E-mail address: david.serrano@urjc.es (D.P. Serrano).
1381-1169/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2009.11.009

D.P. Serrano et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
69
the low reactivity of these substrates and their bulky nature. A sig-
nificant amount of different monounsaturated terminal alcohols is
usually obtained in these reactions, along with other condensation
products [20,21].
times with distilled water, dried overnight at 110 ◦C and calcined
in air at 550 ◦C for 5 h.
2.2. Catalyst characterization
Since zeolites are microporous materials with small pore size,
strong diffusional and steric hindrances usually occur when pro-
cessing bulky molecules, and as a consequence, the activity of
zeolites is then mainly originated by the active sites located on
the external surface. Thereby, the external surface is a property of
high relevance when the zeolite is intended to be used as a catalyst
in reactions involving bulky compounds, such as polymer degrada-
tion, cracking of heavy oil fractions and fine chemicals production
[22,23]. Thus, Wang et al. [24] have recently reported the catalytic
performance of both nanoscale and microscale ZSM-5 in differ-
ent reactions: toluene disproportionation, toluene alkylation with
methanol and trimethylbenzene cracking. A superior reactivity was
found over the nanoscale zeolite due its enhanced accessibility to
acid sites.
Crystallinity of zeolite samples was checked through X-ray
diffraction (XRD) patterns acquired with a Philips X’PERT MPD
diffractometer using Cu K␣ radiation. Typically, the data were
collected from 5 to 50◦ (2ꢀ). The Si/Al atomic ratio of the sam-
ples was determined by induced coupled plasma-atomic emission
spectroscopy (ICP-AES) with a Varian VISTA-AX equipment. Mor-
phology and crystal sizes of both ZSM-5 zeolites were determined
by SEM and TEM images. SEM photographs of conventional ZSM-
5 zeolite were obtained using a PHILIPS XL30 ESEM.FEI scanning
microscope operating at 20 kV, whereas TEM images of nanocrys-
talline n-ZSM-5 sample were obtained in a PHILIPS TECNAI 20
electron microscope operating at 200 kV.
Textural properties of catalysts were determined by means
In this context, nanosized zeolites with a considerable amount
of fully accessible acid sites located on the external surface may
be potentially interesting catalysts for rearrangement reactions
involving large and long epoxides. Accordingly, in the present work,
a comparison has been carried out between the catalytic behaviour
of nanocrystalline and microcrystalline ZSM-5 samples for epoxide
rearrangement reactions. In particular, the catalytic behaviour of
these materials has been checked in the rearrangement of the three
following epoxides: 1,2-epoxyoctane, 2-methyl-2,3-epoxybutane
and isophorone oxide.
of nitrogen adsorption–desorption isotherms at 77 K with a
Micromeritics ASAP 2010 porosimeter after outgassing of the cal-
cined samples under vacuum at 200 ◦C for 5 h. The total surface area
was determined according to the BET equation. The total pore vol-
ume was determined from the nitrogen adsorption at p/p0 = 0.99.
Micropore volume and external surface area of the catalysts were
determined using the t-plot method [27]. The BJH model was
employed for assessing the pore size distribution arisen from the
intercrystalline voids of the n-ZSM-5 zeolite.
The coordination of aluminium atoms within the catalyst frame-
work was checked by high resolution 27Al MAS NMR spectra of the
calcined samples. The spectra were recorded at 104.26 MHz using
a VARIAN Infinity 400 spectrometer at spinning frequency of 4 kHz.
Intervals of 30 s between successive accumulations were selected
according to the structural nature of the samples. All the measure-
ments were carried out at room temperature with [Al(H2O)6]3+ as
external standard reference.
The acid properties of the catalysts were determined by ammo-
nia temperature programmed desorption (TPD) in a Micromeritics
2910 (TPD/TPR) equipment. Previously, the samples were out-
gassed under an He flow (50 N ml min−1) by heating with a rate
of 15 ◦C min−1 up to 560 ◦C and remaining at this temperature for
30 min. After cooling to 180 ◦C, an ammonia flow of 35 N ml min−1
was passed through the sample for 30 min. The physisorbed ammo-
nia was removed by flowing He at 180 ◦C for 90 min. The chemically
adsorbed ammonia was determined by increasing the temperature
up to 550 ◦C with a heating rate of 15 ◦C min−1, this temperature
being maintained constant for 30 min. The ammonia concentration
in the effluent He stream was monitored by a thermal conductivity
detector.
2. Experimental
2.1. Catalyst preparation
ZSM-5 zeolite with a micrometer crystal size (␮-ZSM-5) was
prepared in presence of fluoride ions according to the procedure
reported by Guth et al. [25]. Adequate amounts of tetrapropy-
lammonium hydroxide (TPAOH, 40 wt.% aqueous solution from
Alfa), deionized water and aluminium isopropoxide (AIP, 99 wt.%
from Aldrich) were mixed under stirring at 25 ◦C, until a clear
solution was obtained. Then, tetraethylorthosilicate (TEOS, 98 wt.%
from Aldrich) was added to the solution and stirred during 40 h,
the molar composition of the mixture being as follows: Al2O3:60
SiO2:11 TPAOH:900 H2O. Thereafter, hydrofluoric acid (HF, 48 wt.%
aqueous solution from Panreac) was dropwise added with a molar
ratio of HF/TPAOH = 1. The white gel obtained was transferred
into PTFE-lined autoclaves and crystallized in static conditions at
170 ◦C during 5 days. The final solid product was separated from
the remaining liquid phase by centrifugation at 11,000 rpm dur-
ing 30 min, washed several times with distilled water and dried at
110 ◦C overnight. The calcined sample was obtained by treatment
of the as-synthesized material in static air at 550 ◦C for 5 h.
Nanocrystalline ZSM-5 zeolite (n-ZSM-5) was prepared accord-
ing to a procedure previously reported by van Grieken et al.
[26]. Firstly, the aluminium source (aluminium isopropoxide,
AIP, 99 wt.% from Aldrich) was added to an aqueous solution of
tetrapropylammonium hydroxide (TPAOH, 40 wt.% aqueous solu-
tion from Alfa). The mixture was stirred at 0 ◦C to obtain a clear
solution. Then, tetraethylorthosilicate (TEOS, 98 wt.% from Aldrich)
was cooled down and added to the mixture under stirring at 0 ◦C.
In order to achieve a total hydrolysis of the silicon species, the final
solution was left under stirring during 48 h at room temperature.
Thereafter, the ethanol formed was removed by vacuum heating
at 45 ◦C. The concentrated solution was crystallized at 170 ◦C dur-
ing 120 h in teflon-lined stainless-steel autoclaves. Finally, the solid
product obtained was separated by centrifugation, washed several
2.3. Catalytic epoxide rearrangement experiments
The catalytic experiments were carried out in a stirred batch
autoclave (0.1 L), equipped with a temperature controller and a
pressure gauge, under stirring (550 rpm) and autogeneous pres-
sure for 2 h. The reactor was also provided with a device to feed the
epoxide into the teflon-lined reactor once the reaction temperature
is reached. The solvent and the catalyst were initially placed in the
Teflon-lined reactor. The zero time of the reaction was considered
when the temperature reached the set-point value, the epoxide
being then loaded into the reactor. The employed mass ratio epox-
ide/catalyst was 12 (1 g of epoxide and 80 mg of catalyst). Toluene
was stored with zeolite A to minimise its water content (water con-
tent less than 0.04 wt.%). The catalyst, prior to the reaction, was
dried overnight at 140 ◦C to remove adsorbed water, being subse-
quently transferred to the solvent-containing Teflon-lined reactor.
This treatment is aimed to minimise and prevent the presence of

70
D.P. Serrano et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
Table 1
Physicochemical properties of the ZSM-5 catalysts.
b
b
d
Catalyst
Si/Ala
Pore diameter (Å)
SBET (m2 g−1
)
SEXT+MESOPOR (m2 g−1
)
Dc (nm)c
Acidity (meq g−1
)
Tmax (◦C)d
n-ZSM-5
␮-ZSM-5
32
35
5.5
5.5
384
340
70
8
20–50
5000–7000
0.59
0.63
353
342
a
ICP-AES measurements.
Nitrogen adsorption at 77 K.
Mean crystal size.
b
c
d
Ammonia TPD experiments.
water during the reaction as it may favour the formation of diols
from the epoxide.
The reaction products were analyzed with a VARIAN 3800
gas chromatograph equipped with a capillary column (HP-FFAP)
with dimensions 60 m × 0.32 mm, using a flame ionization detec-
tor (FID). Identification of the different reaction products was
performed by mass spectrometry (VARIAN SATURN 2000) using
standard compounds.
3. Results and discussion
3.1. Catalysts properties
The nanocrystalline ZSM-5 zeolite investigated in this work as
a catalyst for the liquid phase rearrangement of different epoxides
is featured by being formed by nanocrystals with a distribution of
sizes in the range 20–50 nm. This fact provides this material with
a high amount of fully accessible acid sites located on the external
surface of the nanocrystals. Accordingly, this catalyst is considered
to be potentially appropriate for the conversion of bulky molecules
such as the epoxides selected in the present work. On the other
hand, a sample of conventional microcrystalline ZSM-5 zeolite has
been also prepared and tested as a catalyst in the liquid phase
rearrangement of the different epoxides in order to be used as a
reference.
As illustrated in Table 1, both ZSM-5 samples possess Si/Al molar
ratios slightly higher than that corresponding to the synthesis gels
(Si/Al = 30). XRD diffractograms in Fig. 1(a) show the typical pattern
corresponding to pure MFI zeolitic structure for both samples. Nev-
ertheless, the intensity of the peaks is lower in the nanocrystalline
ZSM-5 which is consistent with the lower size of the crystals in this
sample.
Fig. 1(b) illustrates the nitrogen adsorption isotherms at 77 K
of both zeolite samples whereas Table 1 summarizes their tex-
tural properties. ␮-ZSM-5 sample exhibits a predominantly type
I isotherm, characteristic of purely microporous solids. In con-
trast, n-ZSM-5 showed a significant slope at intermediate relative
pressures and a sharp increase for p/p0 higher than 0.8 with well-
defined hysteresis loop. This type of isotherm is characteristic of
nanozeolites where adsorption takes place not only in the microp-
ores but also with a significant contribution of the external surface
and of the intercrystalline porosity. These aspects have also a pro-
nounced effect on the textural properties of both ZSM-5 samples.
Thus, owing to its smaller crystal size, n-ZSM-5 catalyst presents
a larger external surface area (70 m2 g−1) and a higher total pore
volume (0.24 cm3 g−1) than ␮-ZSM-5 (8 m2 g−1 and 0.14 cm3 g−1
,
respectively). The presence of intercrystalline porosity in the n-
ZSM-5 sample was confirmed by determining the mesopore size
distribution using the BJH model (curve not shown). A relatively
broad peak in the range 10–70 nm, with maximum at 25 nm, was
so obtained, which agrees well with the size expected for the voids
present between the nanocrystals in sample n-ZSM-5 (20–50 nm
particle size).
As shown in Table 1, the average crystal sizes for ␮-ZSM-5 and
n-ZSM-5 zeolites, determined from SEM and TEM images, present
Fig. 1. Characterization of the ZSM-5 samples: (a) XRD patterns, (b) N2 adsorp-
tion/desorption isotherms, (c) 27Al-MAS NMR spectra.

D.P. Serrano et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
71
Fig. 2. TEM image of n-ZSM-5 sample (a) and SEM image of ␮-ZSM-5 sample (b).
values around 5–7 ␮m and 20–50 nm, respectively (Fig. 2). This
means that the size of the crystallites in n-ZSM-5 sample is about
two orders of magnitude lower than in the ␮-ZSM-5 material. As
described above, the small crystal size of the nanocrystalline ZSM-
5 contributes favourably to the formation of an extended external
surface area and to a consequently higher share of external acid
sites. Furthermore, a small crystal size should also reduce the dif-
fusional hindrances for accessing the internal acid sites. Therefore,
combination of both factors is expected to favour the accessibility
of the molecules to the active sites in nanocrystalline zeolites.
27Al-MAS NMR spectra of the two catalysts before calcination
(not shown) present only a distinct peak at ∼50 ppm, assigned to
tetrahedral coordination, indicating that all the aluminium species
are effectively incorporated into the zeolite framework. However,
in the calcined samples, another small peak is detected, placed at
∼0 ppm and assigned to octahedral (extraframework) aluminium
species (Fig. 1(c)). These results indicate that even in the calcined
samples most of the aluminium is effectively incorporated into the
zeolite structure, while just a small proportion of aluminium is
extracted from the framework during the calcination treatment.
The acidity of both zeolite samples has been investigated by
NH3 TPD experiments (see Table 1). The total amount of ammo-
nia desorbed (meq g−1) may be related to the number of acid sites
whereas the temperature of the maximum in the desorption peak
(Tmax) provides an indication of the acid strength. The acidity val-
ues slightly change from one sample to another according to their
different aluminium content: the Si/Al ratio is 32 for n-ZSM-5 and
35 for ␮-ZSM-5. Likewise, a slight variation is observed in the tem-
perature maximum for ammonia desorption corresponding to both
ZSM-5 zeolite samples.
3.2. Catalytic epoxide rearrangement reactions
In order to compare the catalytic performances of both ZSM-
5 samples in the liquid phase epoxide rearrangement, three
epoxides having different molecular geometry and with com-
mercial applications in fine chemistry have been selected as
model compounds: 1,2-epoxyoctane, 2-methyl-2,3-epoxybutane
and isophorone oxide. As a general rule, it has been observed
that both ZSM-5 catalysts promote the isomerisation of the linear,
branched and cyclic epoxides towards the corresponding ketones
and aldehydes as main products through carbocation-mediated
processes, although other secondary products are also detected.
The effect of the epoxide nature on the catalytic activity of the
reaction, in terms of the epoxide conversion evolution along the
time, is depicted in Fig. 3 for both catalysts. Because some epox-
ides may be isomerised thermally, blank reactions in absence of
catalyst were carried out, yielding a very small conversion. These
results show that the contribution of the non-catalyzed reactions
is negligible, at least under the reaction conditions used in this
work. Preliminary experiments confirmed the results from previ-
ous works [28,29], showing that the catalytic activity of acid solids
Fig. 3. Epoxide rearrangement conversion vs. reaction time: 1,2-epoxyoctane (a), 2-
methyl-2,3-epoxybutane (b) and isophorone oxide (c). Temperatures: 150 ◦C, 80 ◦
C
and 80 ◦C, respectively. Catalysts: (ꢀ) n-ZSM-5, (ꢁ) ␮-ZSM-5.

72
D.P. Serrano et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
for the rearrangement of branched and cyclic epoxides is quite
higher than that obtained with linear epoxides, as a consequence
of the lower reactivity of the latter. Accordingly, the reaction tem-
perature for the rearrangement of 2-methyl-2,3-epoxybutane and
isophorone oxide was set at 80 ◦C, whereas in the case of 1,2-
epoxyoctane the reaction was carried out at 150 ◦C in order to
obtain comparative epoxide conversions.
From the results shown in Fig. 3, it is concluded that in all cases
the epoxide conversion over the ␮-ZSM-5 zeolite reached hardly
10% after 2 h of reaction. In contrast, the epoxide conversion values
obtained over the nanocrystalline ZSM-5 sample are considerable
higher than those corresponding to the former, with values ranging
about 40–80% at the end of the experiments. Both samples possess
micropores with the same size, as well as a comparable amount of
acid sites with a similar strength since their temperature maxima
for ammonia desorption are placed around 350 ◦C. Therefore, the
huge difference in the catalytic activity of both ZSM-5 samples must
be attributed to their crystal size and external surface area.
In the case of the n-ZSM-5 sample the crystal size is in the
nanometer range with a value almost two orders of magnitude
lower than that of the ␮-ZSM-5 material. This fact contributes
favourably to the presence of an enhanced external surface area
and, consequently, to a higher share of external acid sites, which
are not limited by steric constraints as in the case of the acid sites
located within the zeolite micropores.
Moreover, the overall diffusion rate through the micropores is
inversely proportional to the square of the crystal size. Therefore,
a quite higher intracrystalline diffusion rate should be expected
for the n-ZSM-5 sample in regard to the ZSM-5 sample with a
micrometer crystal size. The contributions of both factors causes
an enhanced accessibility of the substrate molecules to the zeolite
active sites, explaining the remarkable catalytic activity exhibited
by the nanocrystalline ZSM-5 zeolite, which leads to high conver-
sion, regardless the nature of the epoxide.
The high activity obtained in the rearrangement of 2-methyl-
2,3-epoxybutane over n-ZSM-5 catalyst is remarkable, mainly
when it is in compared with the conversion values obtained
using a linear epoxide or a cyclic epoxide. After 2 h of reac-
tion, the conversion of 2-methyl-2,3-epoxybutane over n-ZSM-5
is over 80%. It is well established that the acid-catalyzed reac-
tions of epoxide rearrangement involves the initial attack of the
electrophilic reagent to the oxirane ring, followed by the ring open-
ing, yielding intermediate carbocations that subsequently may be
transformed into the corresponding aldehyde and/or ketone [4].
The presence of branching in 2-methyl-2,3-epoxybutane epox-
ide and isophorone oxide provides a reactive tertiary carbon that
favours the initiation step of the rearrangement reaction. Indeed,
as a consequence of the high reactivity of the branched epoxides, 2-
methyl-2,3-epoxybutane and ␣,␤-epoxy ketone isophorone oxide,
the processes were carried out at 80 ◦C, whereas the reaction tem-
perature in the rearrangement of the linear epoxide had to be set
at 150 ◦C in order to obtain significant conversions.
In addition to the different reactivity of the epoxide, and as
it has been above indicated, the catalytic performance over n-
ZSM-5 is also related to the overall accessibility of the substrate
to the active sites, which is influenced by the possible presence
of both steric and intracrystalline diffusion problems. Both fac-
tors depend on the zeolite properties (crystal size, micropore
diameter and proportion of external surface area) but also on
the geometry and effective size and of the epoxide molecules.
In this way, Fig. 4 illustrates the molecular dimensions of the
different epoxide molecules employed in this work, calculated
at the reaction temperature by the Gaussian program [30]. The
geometries of epoxide molecules were optimized using the 6-31G**
basis set, at the ab initio DFT levels. The DFT calculation was car-
ried out mainly with Becke’s three-parameter exchange functional
Fig. 4. Optimized geometry and selected dimensions for the epoxide molecules: (a)
1,2-epoxyoctane, (b) 2-methyl-2,3-epoxybutane and (c) isophorone oxide.
and the gradient-corrected functional of Lee, Yang, and Paar (B3-
LYP hybrid functional) [31–33]. The results obtained show that
the 2-methyl-2,3-epoxybutante molecule present dimensions of
3.4 × 5.5 Å, which are lower or very similar to the diameter of the
ZSM-5 zeolite micropores. Therefore, it may be concluded that this
molecule can enter the zeolite micropores, although having some
limitations for diffusing across the zeolite channels. In the case of
1,2-epoxyoctane molecules, one of the dimensions is lower than
the pore size whereas the other is clearly larger (3.1 × 9.6 Å, respec-
tively). Accordingly, it may be assumed that this molecule may
also enter the zeolite micropores, although suffering significant
diffusion limitations. Finally, in the case of isophorone oxide, both
dimensions (6.1 × 6.9 Å, respectively) are really larger than the pore
size, which avoids its access to the active sites located within the
zeolite micropores. Therefore, for this substrate the active sites are
just those located on the external surface of the zeolite nanocrys-
tals. The high conversion obtained with isophorone oxide should
be related not just with the high reactivity of this epoxide but it
confirms the important role played by the acid sites on the outer
part of the zeolite crystallites.
Fig. 5 summarizes the molar product distribution obtained over
both ZSM-5 samples in the liquid phase rearrangements of the
three epoxides considered in this work. In all cases, the selectivities
towards the main reaction products hardly vary with the course of
the reaction. On the other hand, when analysing the differences
observed between both catalysts, it should be taken into account
that results corresponding to the ZSM-5 zeolite with microme-
ter crystal size may be strongly influenced by the low conversion
obtained over this material.
Regardless the catalysts used, the main reaction products in
the 1,2-epoxyoctane isomerisation were both octaldehyde and
octen-1-ols, although other rearrangement products were also
obtained. The appearance in the reaction product of 1,2-octanediol
indicates the presence of some water, probably contained in the sol-
vent and/or originated from the dehydroxylation of silanol groups

D.P. Serrano et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
73
Fig. 5. Molar product distribution obtained in the epoxide rearrangement over ZSM-5 catalysts: (a, b) 1,2-epoxyoctane (150 ◦C), (c, d) 2-methyl-2,3-epoxybutane (80 ◦C), (e,
f) isophorone oxide (80 ◦C).
located on the surface of the catalysts. The product distribution
obtained over nanocrystalline ZSM-5 zeolite indicates that this
type of catalyst favours the transposition of the positive charge
of carbocations formed by the reaction mechanism [28], leading
to an enhanced formation of octen-1-ols in regard to octaldehyde.
Thus, over this catalyst the selectivity towards octen-1-ols is about
50%, whereas this parameter reaches a value around 40% in the
case of octaldehyde. Nevertheless, taking into account that both of
them are products with commercial interest, as they can be easily
transformed into the corresponding terminal alcohol, the overall
selectivity towards valuable products is over 90%.
Regarding to the molar product distribution obtained in
2-methyl-2,3-epoxybutane rearrangement (Fig. 5(c, d)), both cat-
alysts lead to high selectivities towards methylisopropylketone,
together with a considerable formation of 2,2-dimethylpropanal,
whereas just a small proportion of isoprene is detected. These
results can be explained considering that the ketone formation
from the intermediate carbocation is favoured in comparison to its
transformation to 2,2-dimethylpropanal or isoprene, as the latter
involve methyl transposition and water loss steps, respectively [4].
For the nanocrystalline ZSM-5 zeolite, 2-methyl-2,3-epoxybutane
isomerisation yields the formation of methylisopropylketone

74
D.P. Serrano et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 68–74
with a really high selectivity (around 80%), whereas for 2,2-
dimethylpropanal the selectivity was about 20%.
the rearrangement reaction. Accordingly, whereas for those epox-
ides the catalytic tests were carried out at 80 ◦C, in the case of the
rearrangement of the linear epoxide (1,2-epoxyoctane) the reac-
tion had to be carried out at 150 ◦C in order to obtain comparative
epoxide conversions.
In all cases, just slight changes were observed in the product
distribution along the reaction time. As a consequence, it has been
possible to obtain simultaneously high conversions and selectivi-
ties when using the n-ZSM-5 catalysts. While the conversions after
2 h of reaction ranged between 40 and 80%, the selectivities towards
the desired products were over 80%. These remarkable results show
the excellent catalytic properties exhibited by the nanocrystalline
ZSM-5 in the isomerisation of different types of epoxides.
Finally, in the case of isophorone epoxide rearrangement,
the main reaction product obtained was 2-formyl-2,4,4-
trimethylcyclopentanone with selectivity around 80–85%, whereas
the selectivity of 3,5,5-trimethyl-1,2-cyclohexanedione was set
around 15%. These selectivity values are similar or even higher
than those obtained in a previous work using Al-MCM-41 as
a catalyst [29]. Consequently, these results indicate that over
the ZSM-5 samples the presence of an additional substituent in
the ␤-position of cyclic ␣,␤-epoxy ketones, such as isophorone
oxide, favours the acyl migration, resulting in ring contraction and
consequently in the formation of the aldehyde. In contrast, hydro-
gen migration, leading to the ␣-diketone, takes place in a lesser
extension. Moreover, acid sites in ZSM-5 zeolites do not promote
the deformylation of 2-formyl-2,4,4-trimethylcyclopentanone,
as demonstrated by the almost negligible selectivity towards
2,4,4-trimethylcyclopentanone obtained in this experiment.
In summary, the results obtained in the isomerisation of differ-
ent epoxides over nanocrystalline ZSM-5 can be considered of high
interest, since high catalytic activity and conversions are attained,
whereas the catalyst exhibits also a high selectivity towards prod-
ucts with commercial applications. This can be explained by the
presence of a high accessibility to the active sites in this catalyst,
which favours the extension of the epoxide rearrangement reac-
tions even when using large and/or low reactive epoxides.
References
[1] J. March, Advanced Organic Chemistry, John Wiley, New York, 1992.
[2] J.G. Buchanan, H.Z. Sable, Selective Organic Transformations, Wiley, New York,
1972.
[3] S.M. Waqui, J.P. Horowitz, R. Filler, J. Am. Chem. Soc. 79 (1957) 6283.
[4] W.F. Hölderich, in: R.A. Sheldon, H. van Bekkum (Eds.), Fine Chemicals through
Heterogeneous Catalysis, Wiley/VCH, Weinheim, 2001, p. 217.
[5] V.S. Joshi, N.P. Damodaran, S. Dev, Tetrahedron 27 (1971) 459.
[6] V.S. Joshi, S. Dev, Tetrahedron 33 (1977) 2955.
[7] P.B. Venuto, P.S. Landis, Adv. Catal. 18 (1968) 259.
[8] T. Imanaka, Y. Okamoto, S. Teranishi, Bull. Chem. Soc. Jpn. 45 (1972) 3251.
[9] S. Matsumoto, M. Nitta, K. Aomura, Bull. Chem. Soc. Jpn. 47 (1974) 1537.
[10] M. Nitta, S. Matsumoto, K. Aomura, J. Catal. 38 (1975) 498.
[11] G. Paparatto, G. Gregorio, Tetrahedron Lett. 29 (1988) 1471.
[12] W.F. Hölderich, H. van Bekkum, Stud. Surf. Sci. Catal. 137 (2001) 821.
[13] M. Chamoumi, D. Brunel, P. Geneste, P. Moreaus, J. Solofo, Stud. Surf. Sci. Catal.
59 (1991) 573.
4. Conclusions
[14] D.P. Serrano, M.A. Uguina, G. Ovejero, R. van Grieken, M. Camacho, Micropor.
Mater. 4 (1995) 273.
The catalytic performance of nanocrystalline ZSM-5 zeolite in
the liquid phase epoxide rearrangement in comparison to a micro-
crystalline ZSM-5 zeolite, employed as a reference catalyst, showed
that the former presents remarkable catalytic properties, with both
high epoxide conversions and high selectivity towards the desired
products. The strong differences observed in the catalytic activity
of both samples have been attributed to the enhanced accessi-
bility of the epoxide molecules to the active sites in the case of
nanocrystalline ZSM-5. This is a direct consequence of the small
size of the crystals in this sample, which favours the intracrystalline
diffusion process and provides a significant proportion of external
surface area, the acid sites located on the latter being not limited
by steric constraints. Likewise, for the sterically hindered epoxides,
the catalytic rearrangement occurs on the external surface of the
nanocrystalline zeolite.
[15] T.-L. Ho, S.-H. Liu, Synth. Commun. 13 (1983) 685.
[16] M. Asaoka, S. Hayashibe, S. Sonoda, H. Takei, Tetrahedron 47 (1991) 6967.
[17] C. Meyer, W. Laufer, W.F. Hölderich, Catal. Lett. 53 (1998) 131.
[18] M.J. Faraj, US Patent 5 312 995 (1993), to Arco Chemical Technology.
[19] W.F. Hölderich, N. Götz, L. Hupfer, H. Lermer, US Patent 4 980 511 (1987), to
Basf AG.
[20] D. Brunel, M. Chamoumi, P. Geneste, P. Moreau, J. Mol. Catal. 79 (1993) 297.
[21] G.D. Yadav, D.V. Satoskar, J. Chem. Technol. Biotechnol. 69 (1997) 438.
[22] D.P. Serrano, J. Aguado, J.M. Escola, J.M. Rodriguez, J. Anal. Appl. Pyrolysis 74
(2005) 353.
[23] A. Corma, J. Catal. 216 (2003) 298.
[24] K. Wang, X. Wang, Micropor. Mesopor. Mater. 112 (2008) 187.
[25] J.L. Guth, H. Kessler, R. Wey, Stud. Surf. Sci. Catal. 28 (1986) 121.
[26] R. van Grieken, J.L. Sotelo, J.M. Menéndez, J.A. Melero, Micropor. Mesopor.
Mater. 39 (2000) 135.
[27] B.C. Lippens, J.H. De Boer, J. Catal. 4 (1965) 319.
[28] D.P. Serrano, R. van Grieken, J.A. Melero, A. García, Appl. Catal. A: Gen. 269
(2004) 137.
[29] R. van Grieken, D.P. Serrano, J.A. Melero, A. García, J. Catal. 236 (2005) 122.
[30] M.J. Frisch, et al., Gaussian 03, Gaussian, Inc., Wallingford CT, 2004.
[31] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[32] A.D. Becke, J. Phys. Chem. 98 (1993) 5648.
[33] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, J. Phys. Chem. 98 (1994)
11623.
The extent of the epoxide conversion was also dependent on the
configuration of reactant molecule. Thus, the presence of branching
in 2-methyl-2,3-epoxybutane epoxide and isophorone oxide pro-
vides a reactive tertiary carbon that favours the initiation step of