Synthesis, characterization, and catalytic properties of zeolites IM-5 and NU-88

Article abstract of DOI:10.1016/S0021-9517(02)00178-1

The synthesis, characterization, and catalytic properties of two high-silica zeolites IM-5 and NU-88, whose structures still remain unresolved, were presented. The strength of both Bronsted and Lewis acid sites was stronger in H-IM-5 than in H-NU-88. However, despite its nanocrystallinity, NU-88 exhibited a very high thermal stability. The catalytic properties of these two zeolites were evaluated for the skeletal isomerization of 1-butene to isobutene and the cracking of n-octane and compared to those from various zeolites with known structures. H-IM-5 and H-NU-88 both exhibited a very high 1-butene conversion compared to H-ZSM-35 over the period of time studied, whereas the opposite holds for the formation of isobutene. Their pore topologies were large enough to allow undesired side reactions, e.g., 1-butene dimerization followed by cracking to light olefins. They also showed the initial n-octane cracking activity comparable to that on H-ZSM-5. However, a notable decrease in n-octane conversion on these two zeolites with increasing lime on stream was observed. When the isomerization and cracking activities of all zeolites employed were correlated with their coke-forming propensities, both materials presented a shape-selective character falling within the category of multidimensional, large-pore zeolites.

Full text of DOI:10.1016/S0021-9517(02)00178-1

Journal of Catalysis 215 (2003) 151–170
www.elsevier.com/locate/jcat
Synthesis, characterization, and catalytic properties
of zeolites IM-5 and NU-88
Song-Ho Lee,a Dong-Koo Lee,a Chae-Ho Shin,a,∗ Yong-Ki Park,b Paul A. Wright,c
Won Mook Lee,d and Suk Bong Hong d,∗
a
Department of Chemical Engineering, Chungbuk National University, Chungbuk 361-763, South Korea
b
Division of Chemical Technology, Korea Research Institute of Chemical Technology, Taejon 305-606, South Korea
c
School of Chemistry, University of St. Andrews, St. Andrews, KY16 9ST, UK
d
Division of Chemical Engineering, Hanbat National University, Taejon 305-719, South Korea
Received 16 September 2002; accepted 26 November 2002
Abstract
The synthesis, characterization, and catalytic properties of two high-silica zeolites IM-5 and NU-88, whose structures still remain
unresolved, are presented. When the 1,5-bis(N-methylpyrrolidinium)pentane and 1,6-bis(N-methylpyrrolidinium)hexane cations are used
as organic structure-directing agents, respectively, crystallization of pure IM-5 and NU-88 was possible only from synthesis mixtures with
a narrow range of SiO /Al O and NaOH/SiO ratios. The overall characterization results of this study strongly suggest that IM-5 is a
2
2
3
2
new multidimensional large-pore zeolite, whereas NU-88 is a nanocrystalline material that could be an intergrowth of several hypothetical
polymorphs in the beta family of zeolites. Despite its nanocrystalline nature, however, no detectable extraframework Al species were found
to exist in H–NU-88, revealing excellent thermal stability. H–IM-5 and H–NU-88 both exhibit a very high 1-butene conversion compared
to H–ZSM-35 over the period of time studied here, whereas the opposite holds for the formation of isobutene. This reveals that their pore
topologies are large enough to allow undesired side reactions such as 1-butene dimerization followed by cracking to light olefins. They also
show the initial n-octane cracking activity comparable to that on H–ZSM-5. However, a notable decrease in n-octane conversion on these
two zeolites with increasing time on stream is observed. When the isomerization and cracking activities of all zeolites employed in this study
are correlated with their coke-forming propensities, it can be concluded that both materials present a shape-selective character falling within
the category of multidimensional, large-pore zeolites.
2003 Elsevier Science (USA). All rights reserved.
Keywords: Zeolites IM-5 and NU-88; Pore topologies; Synthesis; Characterization; Catalytic evaluation
1. Introduction
chemical and catalytic properties of newly discovered zeo-
lites, together with their structure elucidation, is of funda-
mental importance in finding the successful industrial appli-
cation of these materials.
Since the first introduction of zeolites as solid acid cat-
alysts by Mobil in the early 1960s, the use of this im-
portant class of microporous materials has continued to
grow remarkably in the petrochemical and refining indus-
tries [1,2]. Indeed, the success of zeolites as commercial cat-
alysts is largely due to the steady discovery of materials with
novel pore topologies and thus new shape-selective proper-
ties that have offered valuable opportunities in developing
new process technologies, as well as in improving the exist-
ing processes. Therefore, detailed knowledge of the physico-
IM-5 and NU-88 are two new, high-silica zeolites that
have recently been reported by Benazzi et al. [3,4] and
Casci et al. [5], respectively. Although the framework struc-
tures of these synthetic zeolites still remain undetermined,
recent results from various catalytic test reactions have
suggested that both materials may either contain the two-
dimensional pore system with two intersecting 10-ring chan-
nels or the one-dimensional pore system consisting of 10-
ring channels with large internal cavities [4–8]. If such
is the case, they should be potentially useful as shape-
selective catalysts in many hydrocarbon conversions, es-
pecially in isomerization, alkylation, and/or cracking of
*
Corresponding author.
E-mail addresses: chshin@cbucc.chungbuk.ac.kr (C.-H. Shin),
sbhong@hanbat.ac.kr (S.B. Hong).
0021-9517/03/$ – see front matter
doi:10.1016/S0021-9517(02)00178-1
2003 Elsevier Science (USA). All rights reserved.

152
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
hydrocarbons catalyzed by medium-pore zeolites. Accord-
ing to the initial patents for IM-5 and NU-88, on the
other hand, the synthesis of both materials include the
use of diquarternary alkylammonium cations with formula
(C5H11)N+(CH2)nN+(C5H11) with n = 5 and 6, respec-
tively [3,5], which are formed of two 1-methylpyrrolidinium
groups connected by the polymethylene bridging unit. It
has been repeatedly shown that the phase selectivity of ze-
olite syntheses in the presence of such flexible, linear or-
ganic cations as structure-directing agents (SDAs) is sen-
sitive not only to the length of the central alkyl chain
and the nature of the groups on the ammonium ion em-
ployed but also to the oxide composition of synthesis mix-
tures [9–14]. This has stimulated us to investigate the syn-
thetic details on how to prepare zeolites IM-5 and NU-88
using the 1,5-bis(N-methylpyrrolidinium)pentane and 1,6-
bis(N-methylpyrrolidinium)hexane cations from synthesis
mixtures having different oxide compositions, respectively.
In this paper we describe our attempts to synthesize
pure IM-5 and NU-88 and their physicochemical proper-
ties that have been extensively characterized by using a
number of analytical techniques including powder X-ray
diffraction, elemental and thermal analyses, transmission
electron microscopy, N2 and Ar adsorption, 1H, 13C, 27Al,
and 29Si MAS NMR, IR, Raman, temperature-programmed
desorption of ammonia, and IR measurements of adsorbed
pyridine. In addition, we report the catalytic properties
of H–IM-5 and H–NU-88 for the skeletal isomerization
of 1-butene to isobutene and the cracking of n-octane.
The catalytic results are compared to those obtained from
H–ZSM-5, H–ZSM-35, H–EU-1, H–mordenite, and H–beta
zeolites, which possess structural features considerably dif-
ferent from one another, to gain a better understanding of
their pore structures.
solution, Aldrich), Al(NO3)3 · 9H2O (98%, Junsei), fumed
silica (Aerosil 200, Degussa), and deionized water. The final
composition of the synthesis mixture was 3.0R · xNa2O ·
yAl2O3 · 30SiO2 · 1200H2O, where R is the organic SDA
prepared here, x is varied between 7.0 ꢀ x ꢀ 15.0, and y is
varied between 0 ꢀ y ꢀ 2.0. When necessary, NaOH was
replaced by the equivalent amount of LiOH · H2O (98%,
Aldrich) or KOH (45% aqueous solution, Aldrich). After
being stirred at room temperature for 24 h, the synthesis
mixture was charged into Teflon-lined 45-mL autoclaves and
heated to 160 ◦C under rotation (100 rpm) for 14 days. The
solid products were recovered by filtration, washed repeat-
edly with water, and then dried overnight at room temper-
ature. As-synthesized zeolites were calcined under flowing
air at 550 ◦C for 8 h to remove the organic SDA occluded.
The calcined samples were then refluxed twice in NH4NO3
solutions for 6 h followed by calcination at 550 ◦C for 8 h to
ensure that they were completely in their proton form. If re-
quired, the resulting solids were subsequently heated under
flowing air at 600–800◦C for 4 h.
For comparison, ZSM-35 (SiO2/Al2O3 = 27) and EU-1
(SiO2/Al2O3 = 40) were prepared and converted into their
proton form according to the procedures described else-
where [15,16]. In addition to these zeolites, NH4–ZSM-5
(SiO2/Al2O3 = 27), H–mordenite (SiO2/Al2O3 = 33), and
NH4–beta (SiO2/Al2O3 = 25) were obtained from ALSI-
PENTA Zeolithe, Tosoh, and PQ, respectively.
2.2. Characterization
Powder X-ray diffraction (XRD) patterns were collected
on a Rigaku Miniflex or a Rigaku 2500H diffractometer with
Cu-Kα radiation. Elemental analysis was carried out by the
Analytical Laboratory of the Korea Institute of Science and
Technology. Thermogravimetric analyses (TGA) were per-
formed in air on a TA Instruments SDT 2960 thermal an-
alyzer, where weight loss related to the combustion of or-
ganic SDA was determined from differential thermal analy-
ses (DTA) using the same analyzer. Approximately 15 mg
of sample was used at a heating rate of 10 ◦C min−1. Crys-
tal morphology and size were determined by a JEOL JSM-
6300 scanning electron microscope (SEM) or a Carl Zeiss
EM912 OMEGA transmission electron microscope (TEM).
The Ar and N2 sorption experiments were performed on a
Micromeritics ASAP 2010 analyzer.
2. Experimental
2.1. Synthesis
The divalent 1,5-bis(N-methylpyrrolidinium)pentane
(MPP) cation was prepared by reacting 1,5-dibromopentane
(97%, Aldrich) with an excess of 1-methylpyrrolidine (97%,
Aldrich) in acetone as a solvent with rapid stirring at room
temperature overnight. The excess amine was removed by
extraction with diethyl ether and recrystallizations were per-
formed in methanol–diethyl ether mixtures. The 1,6-bis(N-
methylpyrrolidinium)hexane(MPH) dibromide salt was pre-
pared by using 1,6-dibromohexane(96%, Aldrich) instead of
1,5-dibromopentane,respectively, with procedures similar to
the MPP preparation. Due to the hygroscopic nature, after
purification, these two diquaternary ammonium salts were
stored in a desiccator prior to their use as organic SDAs.
The synthesis of zeolites IM-5 and NN-88 in the pres-
ence of MPP and MPH cations was carried out using alumi-
nosilicate gels prepared by combining NaOH (50% aqueous
The 29Si MAS NMR spectra were measured on a Bruker
DSX 400 spectrometer at a spinning rate of 12.0 kHz. The
operating 29Si frequency was 79.492 MHz, and the spectra
were obtained with an acquisition of ca. 800 pulse transients,
which were repeated with a π/5 rad pulse length of 2.0 µs
and a recycle delay of 60 s. The 29Si chemical shifts are
referenced to TMS. The 27Al MAS NMR spectra were
recorded on the same spectrometer with a spinning rate of
13.0 kHz at a 27Al frequency of 104.269 MHz. The spectra
were obtained with an acquisition of 2048 pulse transients,
which were repeated with a π/20 rad pulse length of 0.5 µs

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
153
and a recycle delay of 1 s. The 27Al chemical shifts are
(2.0 kPa) was passed over the same microreactor with 0.1 g
of catalyst at 310–500 ◦C. In the absence of a catalyst, no
conversion of n-octane was observed, even at 500 ◦C. He
was used as a carrier gas with n-octane as a feed, and
the total gas flow was fixed to 50 cm3 min−1. The cracked
products were analyzed online in a Chrompack CP 9001 gas
chromatograph. Gaseous products with < C4 were separated
in a packed Porapak Q column (1/8 in. × 1.8 m) and
analyzed in a thermal conductivity detector (TCD). Gaseous
and liquid products with ꢁ C4 were separated in a capillary
PONA column (0.25 mm × 100 m) and analyzed in an FID.
Conversion and selectivity to each product were calculated
following the methods described elsewhere [13].
referenced to an Al(H2O)6 solution. The 1H MAS NMR
3+
spectra at a spinning rate of 13.0 kHz were recorded at
a proton Larmor frequency of 400.146 MHz with a π/5
rad pulse length of 2.0 µs, a recycle delay of 3 s, and an
acquisition of 32 pulse transients. The 1H–13C CP MAS
NMR spectra at a spinning rate of 4.5 kHz were recorded
at a 13C frequency of 100.623 MHz with a π/2 rad pulse
length of 5.0 µs, a contact time of 1 ms, and a recycle delay
of 3 s. Approximately 6500 scans were accumulated. Both
1H and 13C chemical shifts are reported relative to TMS.
The IR spectra in the structural region were measured
on a Nicolet Magna 550 FT-IR spectrometer using the KBr
pellet technique. The IR spectra in the OH region were
measured on the same spectrometer using self-supporting
zeolite wafers of approximately 12 mg (1.3-cm diameter).
Prior to IR measurements, the zeolite wafers were pretreated
under vacuum at 500 ◦C for 2 h inside a home-built IR
cell with CaF2 windows. For IR spectroscopy with adsorbed
pyridine, the activated self-supporting wafer was contacted
with a pyridine-loaded flow of dry He at room temperature,
evacuated (10−3 Torr) to remove physisorbed pyridine,
and then heated at different temperatures. The Raman
spectra were recorded on a Bruker RFA 106/S FT-Raman
spectrometer equipped with a Nd:YAG laser operating at
1064 nm. The samples were exposed to a laser power of
200 mW at the spectral resolution of 2 cm−1. Typically,
1024 scans were accumulated for obtaining the Raman
spectra and 256 scans for the IR spectra. Temperature-
programmed desorption (TPD) of ammonia was recorded on
a fixed-bed, flow-type apparatus attached to a Balzers QMS
200 quadruple mass spectrometer, following the procedure
described in our previous work [13].
3. Results and discussion
3.1. Synthesis
Table 1 lists the results from syntheses using MPP as
an organic SDA and aluminosilicate gels with different
oxide compositions under the conditions described above.
In each case, the zeolitic products listed were the only
ones obtained in repeated trials. The powder XRD patterns
of three IM-5 zeolites with different SiO2/Al2O3 ratios in
the as-synthesized form, which were obtained from runs
3, 1, and 5 in Table 1, respectively, are shown in Fig. 1.
The positions and relative intensities of all the X-ray peaks
from these IM-5 samples are in good agreement with those
reported in the literature [3,6].
Table 1
Representative products obtained using 1,5-bis(N-methylpyrrolidinium)-
a
pentane (MPP) as an organic SDA
2.3. Catalysis
b
Run
Gel composition
Product
SiO /Al O
NaOH/SiO
2
2
2
3
All the catalytic experiments were conducted under
atmospheric pressure in a continuous-flow apparatus with a
fixed-bed microreactor. Prior to the experiments, the catalyst
1
2
3
4
5
6
7
8
9
10
11
12
13
14
60
15
30
40
120
240
∞
60
60
60
60
60
0.73
0.73
0.73
0.73
0.73
0.73
0.73
1.00
0.87
0.60
0.47
0.33
IM-5
Analcime + IM-5
IM-5
was routinely activated under flowing He (50 cm3 min−1
)
IM-5
IM-5
at 500 ◦C for 1 h and kept at the desired temperature
to establish a standard operating procedure, allowing time
for the product distribution to stabilize. In the skeletal
isomerization of 1-butene, a reactant stream with a He/
1-butene molar ratio of 9.0 was fed into a quartz reactor
containing 0.1 g of zeolite catalyst at 400 ◦C. The total gas
Mordenite
c
L
Analcime
Mordenite + Analcime
Analcime + IM-5
ZSM-12
ZSM-12
ZSM-12
flow at the reactor inlet was kept constant at 50 cm3 min−1
.
d
60
60
0.73
The reaction products were analyzed online in a Chrompack
CP 9001 gas chromatograph equipped with an Al2O3/KCl
Plot capillary column (0.53 mm × 50 m) and a flame
ionization detector (FID), with the first analysis carried
out after 10 min on stream. Conversion was calculated in
terms of the mole percent of 1-butene, and selectivity to
isobutene was calculated by dividing the isobutene yield
by the 1-butene conversion. In the cracking of n-octane,
a reaction stream with a fixed partial pressure of n-octane
e
0.73
Mordenite
a
The oxide composition of the synthesis mixture is 4.5R:xNa O:
2
yAl O :30SiO :1200H O, where R is organic SDA, x is varied between
2
3
2
2
5.0 ꢀ x ꢀ 15.0, and y is varied between 0 ꢀ y ꢀ 2.0. All the syntheses
were carried out under rotation (100 rpm) at 160 C for 14 days.
◦
b
The phase appearing first is the major phase.
Unknown, probably layered material.
c
d
LiOH/SiO ratio.
2
e
KOH/SiO ratio.
2

154
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
be prepared without using any organic SDA. This is not
unexpected because the strong influence of the type of alkali
metal cations on the product selectivity in zeolite syntheses
is frequently observed, even in the presence of organic
additives [12,13,17].
According to the original patent on the synthesis of
IM-5 [3], on the other hand, this zeolite is claimed to
crystallize in the presence of MPH, as well as MPP. Thus,
we have carried out the synthesis using MPH, instead
of MPP, as an organic SDA, under the conditions where
the optimized gel (SiO2/Al2O3 = 60 and NaOH/SiO2 =
0.73) was employed. The powder XRD pattern of the solid
obtained from this synthesis run is given in Fig. 1d. Notice
that the positions and relative intensities of the observed
X-ray peaks are notably different from those of any of
the three IM-5 zeolites prepared here. For example, several
peaks of medium intensity in the region 12◦ ꢀ 2θ ꢀ 18◦
are present in IM-5 but absent in product synthesized with
MPH. In particular, all the X-ray peaks of this solid are
not well resolved compared to those of IM-5, despite the
fact that their XRD patterns were measured with identical
scan parameters. A comparison with the XRD patterns of
zeolitic materials available in the literature reveals that the
relative peak intensities and positions of the solid from the
MPH-mediated synthesis is exactly the same as those of
zeolite NU-88 recently patented by Casci et al. [5]. It should
also be noted that the XRD pattern of NU-88 is similar
to that of ITQ-10 with some differences, one of recently
reported intergrowths of several hypothetical polymorphs in
the beta family of zeolites [18]. This led us to speculate
that NU-88 may fall into the structural regime of large-
pore materials, although additional information regarding
its pore topology cannot be readily obtained from its XRD
pattern due to the very broad nature of X-ray peaks. Like
those to IM-5, the reported synthesis routes to NU-88 also
include the use of MPP and MPH as organic SDAs [5].
When MPP is used as a SDA, however, our attempts to
reproduce the synthesis of this zeolite under the reported
conditions were not successful: a mixture of IM-5 and
analcime, instead of NU-88, was produced. The same result
was observed from syntheses using aluminosilicate gels with
a wide range of SiO2/Al2O3 and NaOH/SiO2 ratios (see
Table 1). There is a large volume of work on the use of
flexible diquaternary alkylammonium cations with aliphatic
and/or cyclic moieties in the synthesis of zeolites and related
materials, and it has been repeatedly shown that the phase
selectivity of the crystallization depends highly on the length
of the spacing alkyl chain and on the size and shape of the
alkylammonium moieties [9–14]. We believe that this is the
case for the synthesis of IM-5 and NU-88.
Fig. 1. Powder XRD patterns of IM-5 and NU-88 zeolites prepared here:
(a) IM-5(I), (b) IM-5(II), (c) IM-5(III), and (d) NU-88 in the as-synthesized
form.
It can be seen from Table 1 that the oxide composition
range yielding pure IM-5 in the presence of MPP is rather
narrow. When the NaOH/SiO2 ratio in the synthesis mixture
was fixed to 0.73, for example, the SiO2/Al2O3 ratio leading
to the successful IM-5 formation was found to be in the
range 30–120. When using sodium aluminosilicate gels with
SiO2/Al2O3 < 30, however, we always obtained a mixture
of analcime and IM-5, where the former material was the
major phase from the gel with a lower SiO2/Al2O3 ratio.
In addition, mordenite was the phase that crystallized from
the synthesis mixture with SiO2/Al2O3 = 240 under the
conditions described above. When the initial SiO2/Al2O3
ratio in the synthesis mixture was fixed to 60, on the
other hand, ZSM-12 with the one-dimensional 12-ring pore
system was the phase formed from synthesis mixtures
with NaOH/SiO2 ꢀ 0.5. In contrast, the synthesis using
aluminosilicate gels with NaOH/SiO2 ꢁ 0.8 again yielded
mordenite or analcime rather than IM-5. These results
indicate that the presence of Na+ ions with a certain level
of concentration in the synthesis mixture, together with the
organic MPP, is the major factor determining the phase
selectivity of the crystallization. Therefore, it appears that
the structure-directing ability of MPP itself is not strong
enough to govern the crystallization of IM-5. We also note
that the replacement of NaOH with the equivalent amount
of LiOH under the conditions where the crystallization of
IM-5 proved to be highly reproducible yielded ZSM-12,
whereas the use of KOH gave mordenite that could also
A summary of products from syntheses in the presence of
MPH as an organic SDA is listed in Table 2. These data re-
veal that the SiO2/Al2O3 ratio range of synthesis mixtures
leading to pure NU-88 formation is much narrower than that
found in the crystallization of IM-5. However, the general
trend drawn from Table 2 is quite similar to that observed

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
155
Table 2
the surrounding inorganic species at the crystallization tem-
perature in these two synthesis mixtures. The synthesis re-
sults in Tables 1 and 2 demonstrate that there is a reasonable
amount of lattice charge in the synthesis mixture, together
with a certain level of Na+ concentration, in the crystalliza-
tion of IM-5 and NU-88 using MPP and MPH, respectively.
In our view this may make it possible to considerably differ
the conformation of MPP or MPH from that in the synthe-
sis mixture with a particular range of oxide compositions
yielding the other zeolitic phase because the number of con-
formations available to such flexible organic SDAs and their
geometric differences must be much larger than those for
ethylene glycol or rigid polycyclic molecules with similar
C/N+ values. In this regard, the synthesis results presented
so far support our recent proposal [14,19] that the control of
conformations of flexible, hydrophilic organic SDAs such
as linear diquaternary alkylammonium ions with aliphatic
and/or cyclic moieties by varying the concentrations of in-
organic components in zeolite synthesis mixtures may be an
area of considerable possibility for finding new materials.
Representative products from syntheses using 1,6-bis(N-methyl-pyrrolidi-
nium)hexane (MPH)
a
b
Run
Gel composition
Product
SiO /Al O
NaOH/SiO
2
2
2
3
15
16
17
18
19
20
21
22
23
24
25
26
27
28
60
15
30
40
0.73
0.73
0.73
0.73
0.73
0.73
0.73
1.00
0.87
0.60
0.47
0.33
NU-88
Mordenite
Mordenite
Mordenite
120
240
∞
60
Mordenite + NU-88
c
D
D
c
Mordenite
Mordenite
ZSM-12
ZSM-12
ZSM-12
60
60
60
60
60
60
d
c
0.73
D
D
e
c
0.73
a
The synthesis conditions and oxide composition of the synthesis
mixtures used are the same as those in Table 1.
b
The phase appearing first is the major phase.
Unknown, probably dense material.
LiOH/SiO ratio.
KOH/SiO ratio.
c
d
2
3.2. Characterization
e
2
The powder XRD patterns (not shown) of the proton
form of three IM-5 zeolites with different SiO2/Al2O3
ratios and one NU-88 zeolite (i.e., the materials obtained
from runs 3, 1, and 5 in Table 1 and run 15 in Table 2,
respectively) agree well with those of the as-synthesized
form of the corresponding materials given in Fig. 1, except
the minor change in relative X-ray peak intensities and
positions. This reveals that all IM-5 and NU-88 zeolites
prepared here maintain their structures during the initial
calcination at 550 ◦C to re+move the organic SDAs occluded
and the subsequent NH4 ion exchange and calcination
steps, which can be further supported by the N2 adsorption
data (vide infra). The powder XRD pattern of the H–IM-
5(II) sample was found to be successfully indexed on an
orthorhombic cell, with a = 19.972(2) Å, b = 19.073(3) Å,
and c = 14.172(2) Å, as given by Le Bail [20] profile
matching in the space group Pnnm, which satisfactorily
accounts for the systematically absent reflections. It is
interesting to note here that these values are quite similar
to those of ZSM-5 with the b and c parameters a little
shorter and longer, respectively [21], although this does
not mean that the framework topology of IM-5 is closely
related to that of ZSM-5. We are carrying out the structure
determination of IM-5 using synchrotron powder XRD
and high-resolution TEM data, and the results will be
described elsewhere. Due to the very broad feature of the
X-ray peaks of NU-88, on the other hand, we were not
able to reasonably index its XRD pattern. The chemical
compositions of IM-5 and NU-88 zeolites prepared here,
which were determined from a combination of elemental
and thermal analyses, are given in Table 3. A considerable
dissimilarity between the amount of Al and the sum of
organic and alkali cations compensating for framework
for the MPP-mediated synthesis of IM-5 in that MPH can
also produce more than one zeolite structure (i.e., mordenite,
ZSM-12, and NU-88), depending on the oxide composition
of synthesis mixtures. The TGA/DTA curves (not shown)
of as-synthesized mordenite and ZSM-12 clearly show the
existence of MPH in their pores, where the decomposition
pattern of the occluded organic SDA differs significantly ac-
cording to the zeolite structure. A similar result was also ob-
tained from IM-5, mordenite, and ZSM-12 prepared with
MPP. This implies that the flexible organic cations (i.e.,
MPP and MPH) occluded within the voids of these three
different products may adopt conformations distinctly dif-
ferent from one another, due to the difference in geometri-
cal restrictions imposed by the particular framework topol-
ogy of each zeolite. Very recently, we performed variable-
temperature 1H combined rotation and multiple-pulse spec-
troscopy (CRAMPS) and 2H NMR measurement on the eth-
ylene glycol molecules in two sodalites with SiO2/Al2O3
ratios of ∞ and ca. 10 [19]. The conformations of the en-
capsulated ethylene glycol molecules in these two materials
were maintained even at 200 ◦C, which is higher than the
crystallization temperature (175 ◦C) of the zeolite host, be-
cause the intermolecular hydrogen bonds between the frame-
work oxygens and the OH groups of the encapsulated ethyl-
ene glycol, which appear only in the Al-substituted zeolite,
are strong enough to overcome the thermal energy avail-
able at 200 ◦C. This strongly suggests that both conforma-
tions of ethylene glycol are actually relevant ones acting as
SDAs at the nucleation stage in aluminosilicate and pure-
silica gels, respectively, which cannot happen without no-
table differences in the hydrogen-bonding interactions we
proved, and more likely in the electrostatic interactions with

156
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
Table 3
a
Chemical compositions of IM-5 and NU-88 zeolites prepared in this study
b
c
c
Zeolite
Run
SiO /
R /
Na O/
(R + Na O)/
2
2
2
Al O
2
Al O
2
Al O
2
Al O
3
3
3
2 3
IM-5(I)
IM-5(II)
IM-5(III)
NU-88
3
1
5
18.8
25.3
32.2
25.6
0.84
1.24
1.28
1.65
0.41
0.28
0.17
0.15
1.25
1.52
1.45
1.80
15
a
Determined from a combination of elemental and thermal analyses.
The same as the run numbers in Tables 1 and 2.
Doubly charged MPP and MPH cation.
b
c
charges is observed for all four zeolites. It thus appears
that part of the N-methylpyrrolidinium moieties of the
organic SDAs occluded within these materials exists in
the form of bromide or hydroxide to serve as space-filling
species rather than as charge-compensating cations. The
data in Table 3 also show an enrichment of Al in all
products with respect to their synthesis mixtures. This trend
is more apparent to IM-5 zeolites obtained from synthesis
mixtures with higher SiO2/Al2O3 ratios, which allowed
us to obtain IM-5 as a pure phase only at a very narrow
range [19–32] of SiO2/Al2O3 ratios under the synthesis
conditions described above. Because the SiO2/Al2O3 ratio
(25.3) of the IM-5(II) sample is essentially the same as
that (25.6) of NU-88, on the other hand, we have mainly
used the characterization results from this IM-5 sample in
comparison with the physicochemical properties of NU-88.
For convenience sake, therefore, we will refer to the IM-
5(II) sample simply as IM-5 from now on (unless otherwise
stated).
Fig. 2 shows the TGA/DTA curves for as-synthesized
IM-5 and NU-88. Zeolite IM-5 gives four stages of weight
loss: 25–250, 250–440, 440–550, and 550–700◦C. The first
loss is endothermic and can be assigned to the desorption of
water. The other losses are accompanied by three exotherms
around 420, 490, and 610 ◦C, respectively. Thus, they must
be from the thermal decomposition of MPP. As seen in
Fig. 2, by contrast, the decomposition of MPH in NU-88
is characterized by four weight losses at 250–340, 340–
440, 440–530, and 530–700 ◦C that have distinct exothermic
peaks in the DTA. We note here that the overall organic
Fig. 3. Pore size distribution curves for (a) H–IM-5 and (b) H–ZSM-5
calculated from their Ar adsorption isotherms using the Horvath–Kawazoe
formalism. The inset shows the adsorption branch isotherms of the
corresponding zeolites.
content (21.0 wt%) in NU-88 is approximately one and
half times as large as that (14.7 wt%) in IM-5. The
TGA/DTA curves (not shown) of both H–IM-5 and H–NU-
88 gave no detectable exothermic peaks in the temperature
region higher than 550 ◦C. This indicates that their high-
temperature weight loss is not due to dehydroxylation of
Brønsted acid sites, but is probably due to oxidation of
products arising from incomplete combustion of organic
SDAs below 550 ◦C.
The pore size distribution (PSD) curves of H–IM-5 and
H–NU-88 calculated from their Ar adsorption isotherms
using the Horvath–Kawazoe (HK) formalism [22] are shown
in Figs. 3 and 4, respectively. Included for comparison
are the curves for H–ZSM-5 and H–beta with a similar
SiO2/Al2O3 ratio (∼ 25). The PSD curve for H–ZSM-5
with two intersecting 10-ring channels is characterized by
one sharp maximum at 5.2 Å. As seen in Fig. 3, however,
the curve for H–IM-5 shows signs of two components in
the micropore size region: one feature around 5.3 Å and
the other component around 5.9 Å. Notice that these values
are quite similar to the effective sizes observed for the
Fig. 2. TGA/DTA curves for as-synthesized (a) IM-5 and (b) NU-88.

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
157
Table 4
N
sorption data for various zeolites studied here
2
a
Zeolite
Run
SiO /
BET
Micropore Mesopore
2
b
c
Al O
2
surface area
volume
volume
3
2
−1
3
−1
3
−1
ratio
(m g
)
(cm g
)
(cm g )
H–IM-5(I)
H–IM-5(II)
H–IM-5(III)
As-syn. NU-88
H–NU-88
3
1
5
15
15
18.8
25.3
32.2
25.6
25.6
27.0
25.0
375
460
559
97
622
398
571
0.14
0.18
0.22
0.04
0.24
0.14
0.22
0.06
0.11
0.26
0.34
0.50
0.02
0.29
H–ZSM-5
H–beta
a
The same as the run numbers in Tables 1 and 2.
In the diameter range ꢀ 20 Å.
In the 20- to 500-Å diameter range. Calculated using the BJH
b
c
formalism.
Fig. 4. Pore size distribution curves for (a) H–NU-88 and (b) H–beta
calculated from their Ar adsorption isotherms using the Horvath–Kawazoe
formalism. The inset shows the adsorption branch isotherms of the
corresponding zeolites.
Table 4 shows the N2 adsorption data for the proton form
of all IM-5 and NU-88 zeolites prepared here. The H–IM-
5(I) sample, which has the highest Al content among the
three IM-5 zeolites prepared in this work, exhibits a BET
surface area of 375 m2 g−1. This value agrees well with that
reported for this zeolite [7]. As seen in Table 4, however,
the measured BET surface area increases remarkably with
decreasing Al content in the zeolite. As a result, the BET
surface area (559 m2 g−1) of the H–IM-5(III) with the lowest
Al content was found to be approximately one and half times
as large as that of the H–IM-5(I) sample. The same trend can
be observed from the micropore volumes of these three IM-5
zeolites. Even in zeolites with the same framework structure
and composition, on the other hand, their micropore volumes
can differ considerably according to the procedure employed
in the calculation [23]. Because the N2 adsorption isotherms
of all IM-5 and NU-88 prepared here is not flat at high
relative pressures, we calculated their micropore volumes
from the volumes of the adsorbed N2 at P/P0 = 0.1,
corresponding to the pore diameter of ca. 20 Å. Table 4
shows that the−H1–IM-5(I) sample has a micropore volume
of 0.14 cm3 g , which is quite similar to the reported
value (0.13 cm3 g−1) [7], whereas the H–IM-5(II) and H–
IM-5(III) samples exhibit considerably higher micropore
volumes (i.e., 0.18 and 0.22 cm3 g−1, respectively). The
27Al MAS NMR spectra of all H–IM-5 zeolites prepared
here show a resonance around 0 ppm (vide infra), indicating
the formation of extraframework Al species during the
calcination and exchange steps. Because this 27Al resonance
did not appear in the spectra of their as-synthesized form,
however, the observed increase in micropore volume for
lower Al contents cannot be due to differences in the
crystallinity. Furthermore, no noticeable differences in the
intensity ratio of the two 27Al resonances due to tetrahedral
and octahedral Al species for the three H–IM-5 zeolites
were found. Then, it can be expected that the amount of
extraframework Al species formed should be smaller for the
zeolite with a lower Al content, leading to a less serious
pore blockage and thus to an increase in micropore volume.
This seems to be a sensible hypothesis, if the IM-5 structure
medium-pore H–ZSM-5 and the large-pore H–beta zeolites,
respectively. This led us to first postulate that IM-5 could
contain a dual-pore system of two different sizes, one type of
micropores with an effective size virtually identical to either
the straight 10-ring or the sinusoidal 10-ring channels in
ZSM-5 and the other type of pores whose size may rank this
new zeolite in the structural regime of large-pore (i.e., 12-
ring) materials. If such is not the case, the other possibility
could be that the IM-5 structure has two types of 12-ring
pores in which one of them is very elliptical. The presence
of two types of pores with different effective sizes in H–
IM-5 can also be achieved by comparing the Ar adsorption
isotherm of this zeolite with that of H–ZSM-5. As seen in
Fig. 3, the pores in both materials start to fill up at a relative
pressure of P/P0 ∼ 5 × 10−6. Unlike H–ZSM-5, however,
H–IM-5 gives a second inflection point in the vicinity of
P/P0 = 5 × 10−4, which can be assigned to the filling
of large 12-ring pores. On the other hand, H–NU-88 has
essentially the same PSD curve as that of H–beta (Fig. 4).
Also, no detectable differences in the Ar adsorption isotherm
are observed for these two zeolites. Recent studies on the
catalytic properties of IM-5 and NU-88 have suggested that
both materials may contain either the two-dimensional pore
system with two intersecting 10-ring channels or the one-
dimensional pore system consisting of 10-ring channels with
large internal cavities [4–8]. Taking into account the PSD
curves in Figs. 3 and 4, however, it is clear that these
two zeolites with unknown framework structures cannot be
regarded as medium-pore 10-ring materials. In our view,
IM-5 may contain a multidimensional large-pore system
with intersecting 12- and 10-ring windows, although the
presence of strongly elliptical 12-ring pores, instead of 10-
ring pores, in this zeolite cannot be completely ruled out.
In addition, it appears that NU-88 is a large pore material
that could belong to the beta family of zeolites, as suggested
above.

158
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
Fig. 5. TEM photographs of as-synthesized (a) IM-5 and (b) NU-88.
is not a three-dimensional 12-ring pore system but a dual-
pore system of intersecting 12- and 10-ring windows. On
the other hand, the micropore volume of H–NU-88 was
calculated to be 0.24 cm3 g−1 which, in principle, makes
this zeolite to fall into the microporosity range of large-
pore zeolites. Table 4 also lists the micropore volumes of
H–ZSM-5 and H–beta obtained from the adsorbed volumes
at P/P0 = 0.1. The values obtained in this work are in
good agreement with those typically reported for these two
zeolites [6,24], supporting the reliability of the micropore
volumes of the H–IM-5 and H–NU-88 zeolites measured in
this study. In parallel with the PDS data from Ar adsorption
experiments, therefore, it is clear that both materials are
large-pore materials rather than medium-pore ones.
Fig. 5 shows the TEM pictures of IM-5 and NU-88
zeolites synthesized in this study. IM-5 consists of needle-
like crystallites approximately 0.6–1.0 µm in length and
less than 0.1 µm in diameter. It should be noted here that
the aspect ratio of IM-5 crystallites considerably decreases
with decreasing Al content in the zeolite, although IM-5
can be obtained as a pure phase only in a narrow range
of SiO2/Al2O3 ratios (vide ante). As a result, most of the
H–IM-5(III) crystallites with the lowest Al content among
the three IM-5 zeolites prepared here are not elongated
but appear to be cuboids of 0.2 × 0.2 × 0.2 µm3. Of
particular interest is the extremely small crystal size (20–
30 nm) of NU-88, although we were unable to determine
its accurate crystal morphology. This strongly suggests that
the very broad nature of its X-ray peaks (Fig. 1d) may
originate from particle size effects rather than from poor
crystallinity. In fact, we have found that no detectable line
narrowing in the XRD pattern of NU-88 is caused by an
additional 2 weeks of heating in the crystallization medium.
When the Scherrer equation with K = 1 is applied to
six different reflections in the 2θ region 7–45◦ [25], the
average crystal size was calculated to be about 30 nm,
which is in good agreement with the value estimated
from TEM measurements. There is growing interest in
the direct synthesis of uniform, nanocrystalline zeolites
because of potential crystal size effects on many important
properties of these highly porous materials, most notably
their adsorption and catalytic properties. However, only
a handful of fullycrystalline but nanometer-sized zeolites
(sodalite, A, X, ZSM-5, L, and beta) have thus far been
reported [26–29].
To further investigate the nanocrystalline nature of NU-
88, we have performed N2 adsorption measurements on its
as-synthesized and proton forms. Here, the as-synthesized
NU-88 material was outgassed at 100 ◦C under vacuum to
a residual pressure of 10−5 Torr for 4 h before measure-
ment. This outgassing treatment, selected upon considering
the TGA/DTA results in Fig. 2, yielded no noticeable re-
moval of the occluded organic SDAs from the pores of NU-
88 because a negligible microporous adsorption capacity
(0.04 cm3 g−1) was observed from the as-synthesized NU-
88 crystallites. The N2 adsorption–desorption isotherms for
the as-synthesized and proton forms of NU-88 are shown in
Fig. 6. The isotherms of both materials possess a steep step
Fig. 6. Pore size distribution curves for the (a) as-synthesized and (b) proton
forms of NU-88 calculated using the BJH formalism from the N desorption
2
branch isotherm. The inset shows the N adsorption–desorption isotherms
2
of the corresponding zeolites.

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
159
at high relative pressures, together with a clear hysteresis
loop, typical of mesoporosity. The corresponding parame-
ters of the pore structure are summarized in Table 4. Note
that the as-synthesized NU-88 exhibits a mesopore volume
of 0.34 cm3 g−1, suggesting significant interparticle adsorp-
tion. Barrett–Joyner–Halenda (BJH) analyses [30] show that
this uncalcined material has an average pore size of ∼ 110 Å.
An important result is that the mesoporous capacity of NU-
88 increases to a very large extent during the initial calci-
nation at 550 ◦C fo+r 8 h to remove the MPH cation and
the successive NH4 ion exchange and calcinations steps:
H–NU-88 has a mesopore volume of 0.50 cm3 g−1, which
is approximately twice as large as its micropore volume.
Comparison of the mesopore distribution of H–NU-88 with
that of its as-synthesized form reveals a large increase in
mesopore size together with a broadening of the distribu-
tion, which can be attributed to sintering effects during the
calcination of NU-88 and subsequent treatments. We note
here that several other H–NU-88 samples obtained from dif-
ferent synthesis batches gave a similar mesopore volume
(0.50 0.03 cm3 g−1). Because H–NU-88 presents both a
large interparticle mesoporosity and a relatively large ze-
olitic microporosity, and an excellent thermal stability (vide
infra), therefore, this nanocrystalline material could have in-
teresting applications in some shape-selective reactions such
as gas oil cracking where the reactant diffusivity depends
highly on both the crystal size and mesoporosity of the zeo-
lite catalyst employed [31]. In addition, we believe that the
nanocrystalline nature of NU-88 would be of considerable
advantage for other important applications, for example, the
synthesis of ultra-thin zeolite films [32] and the use as pho-
tochemical hosts [33].
1
Fig. 7 shows the H–13C CP MAS NMR spectra of as-
synthesized IM-5 and NU-88. For the sake of compari-
son, the liquid 13C NMR spectra of D2O solutions of the
MPP and MPH bromide salts are also given in Fig. 7.
These data clearly show that both the MPP and MPH
cations are occluded intact in IM-5 and NU-88, respectively.
When the MPP cation is trapped in the IM-5 pores, how-
ever, the splitting of the methyl carbon resonance of N-
methylpyrrolidinium moieties into two peaks around 48 and
53 ppm is observed. It thus appears that the considerable
change in conformation is imposed on MPP upon occlusion
into the IM-5 pores, which must be closely related to the
structural aspects of the zeolite host. By contrast, the MPH in
NU-88 gives only one methyl carbon resonance at 49.0 ppm
that is shifted slightly downfield relative to that of free MPH.
The same trend is also observed for the other 13C NMR res-
onances of the occluded MPH. It thus appears that unlike
the organic in IM-5, the MPH occluded in NU-88 does not
experience severe geometric constraints and van der Waals
interactions with the zeolite framework. This can be further
supported by the fact that all the carbon resonances observed
for NU-88 are considerably narrower than those from IM-5.
Additional evidence for the conclusion given above can be
obtained from the 1H MAS NMR spectra of as-synthesized
13
13
Fig. 7. C NMR spectra of organic SDAs, (a) MPP and (b) MPH, showing the assignment of each resonance. The bottom traces are the Bloch decay
C
1
13
NMR spectra of D O solutions of the bromide form of MPP and MPH, and the top traces are the H– C CP MAS NMR spectra of as-synthesized IM-5 with
2
MPP and MPH occluded in their pores, respectively.

160
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
Al are observed for these materials. However, the 27Al
MAS NMR spectrum of H–IM-5 gives a small resonance
around 0 ppm, indicating that a minor part of framework
Al atoms has been extracted from the IM-5 framework
during the calcination and exchange steps. The same result
was also observed for the other two IM-5 zeolites prepared
here, i.e., the IM-5(I) and IM-5(III) samples. An interesting
result obtained from Fig. 9 is that, unlike H–IM-5, H–NU-
88 exhibits no additional line around 0 ppm. In fact, we
found that H–NU-88 gives only one 27Al resonance around
55 ppm, even after heating at 800 ◦C for 4 h, revealing its
remarkable thermal stability. This result is important in that
the nanocrystallinity has little effect on the thermal stability
of NU-88, despite its relatively high Al content. As shown
in Fig. 10, on the other hand, the 29Si MAS NMR spectra
of as-synthesized IM-5 and NU-88 are characterized by a
broad resonance around −110 ppm, which is indicative of
the existence of multiple T-sites in these zeolite structures.
Fig. 10 also shows that the subsequent calcination and
exchange treatments caused no significant changes in the
29Si MAS NMR spectra of IM-5 and NU-88, whereas some
dealumination in IM-5 may have occurred according to the
27Al MAS NMR results.
1
Fig. 8. H MAS NMR spectra of as-synthesized (a) IM-5 and (b) NU-88.
IM-5 and NU-88 in Fig. 8. The intensities between 1.0 and
3.5 ppm are due to the resonance of the protons of organic
SDAs; the signals are relatively sharp for NU-88, whereas
they reveal one broad peak for IM-5. This suggests that the
mobility of the occluded organic SDAs is higher in NU-88
than in IM-5.
Fig. 11 shows the IR spectra in the structural region for
the as-synthesized and proton forms of IM-5 and NU-88
prepared here. One significant difference between these two
zeolites appears in the 500- to 650-cm−1 region. According
to the IR investigations of zeolite frameworks reported thus
far, the bands appearing in this region are attributed to
external linkage vibrations [34–36]. As seen in Fig. 11,
H–IM-5 exhibits one single band at 555 cm−1, typical of
5-ring subunits for pentasil zeolites [36]. Unlike H–IM-5,
Fig. 9 shows the 27Al MAS NMR spectra of the as-
synthesized and proton forms of IM-5 and NU-88. The
spectra of the as-synthesized IM-5 and NU-88 exhibit one
resonance at 51–53 ppm, typical of tetrahedral Al in the
zeolite framework. No signals corresponding to octahedral
however, H–NU-88 gives two bands at 536 and 570 cm−1
,
27
Fig. 9. Al MAS NMR spectra of (a) IM-5 and (b) NU-88 in the as-synthesized (bottom) and proton (top) forms. Spinning side bands are marked by asterisks.

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
161
29
Fig. 10. Si MAS NMR spectra of (a) IM-5 and (b) NU-88 zeolites in the as-synthesized (bottom) and proton (top) forms.
Fig. 11. IR spectra in the structural region of (a) IM-5 and (b) NU-88 zeolites in the as-synthesized (bottom) and proton (top) forms. IR bands from the
occluded organic SDAs are marked by asterisks.
suggesting that IR spectroscopy, as well as powder XRD, can
be regarded as a means to distinguish between both zeolites.
Fig. 12 shows the Raman spectra of the as-synthesized
and proton forms of IM-5 and NU-88 in the 200- to
1600-cm−1 region. The Raman spectra of MPP and MPH in
aqueous solution are also compared in Fig. 12. In agreement
of most of these changes cannot be done without normal-
coordinate calculations. However, some of them provide
clear evidence for the existence of host–guest interactions
occurring within zeolites prepared here. When compared
with the Raman spectra of various tetraalkylammonium salts
and pyrrolidine derivatives in the literature [37,38], for ex-
ample, the band appearing at 731 cm−1 in the Raman spec-
trum of MPP in aqueous solution can be tentatively assigned
to the C–N stretching mode. As seen in Fig. 12, however,
as-synthesized IM-5 gives this vibration at 726 cm−1. The
same trend can be observed for the C–N stretching mode
of MPH (733 cm−1 for free MPHBr2; 728 cm−1 for oc-
1
with the H–13C CP MAS NMR data given above, Raman
spectroscopy reveals that both organic SDAs remain intact
upon their occlusion into the pores of IM-5 and NU-88, re-
spectively. It can also be seen from Fig. 12 that many spec-
tral changes take place as the organic cation is trapped inside
the zeolite pores. The exact assignment and interpretation

162
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
−1
Fig. 12. Raman spectra in the 200- to 1600-cm region of (a) IM-5 and (b) NU-88 in the as-synthesized (middle) and proton (top) forms. The bottom traces
are 0.5 M aqueous solutions of MPPBr and MPHBr , respectively.
2
2
cluded MPH). Fig. 12 also shows that a strong band ap-
pearing at 558 cm−1 in the Raman spectra of free MPP and
MPH, assignable to the C–N bending mode, is nearly miss-
ing in the Raman spectra of as-synthesized IM-5 and NU-88.
In addition to perturbations on the C–N bond, the Raman
signature of occluded organic molecules includes notable
differences in the band position and intensity of the C–C
stretching mode of the polymethylene bridging unit at 1040–
1080 cm−1. These observations indicate that the mirror sym-
metry of flexible organic SDAs has been removed upon their
occlusion into the resultant zeolite pores and thus the two
halves of the molecules occluded are no longer equal to each
other. It is thus clear that organic SDAs in zeolites prepared
here adopt conformations distinctly different from those of
the corresponding, free molecules. However, we note that
the ring breathing vibration band of N-methylpyrrolidinium
groups, which appears as a strong band at 900 cm−1 in the
Raman spectra of both IM-5 and NU-88, is essentially un-
changed relative to that of free MPP or MPH. This led us
to speculate that changes in the conformation imposed on
these doubly charged cations upon occlusion into the zeo-
lite pores occur mainly in the flexible central alkyl chain
rather than in the puckered pyrrolidinium rings, which might
be sufficient to cause the considerable reduction of geomet-
ric constraints and van der Waals interactions with the par-
ticular zeolite structure formed. When the Raman spectrum
of H–IM-5 is compared with that of H–NU-88, finally, no-
table differences in the number and position of structural Ra-
man bands were found. H–IM-5 exhibits four well-resolved
bands at 301, 362, 425, and 487 cm−1. As seen in Fig. 12,
however, the Raman spectrum of H–NU-88 is characterized
by three main bands around 340, 390, and 467 cm−1 with
relatively poor resolution. Although the unequivocal assign-
ment of the structural Raman bands of H–IM-5 and H–NU-
88 cannot be made at this time, due to the lack of the re-
liable theoretical correlation between the specific structural
units in zeolites and their Raman band positions, it is in-
teresting to note that the general feature of the spectrum in
the structural region of H–NU-88 is similar to that reported
for H–beta [39], whereas there are notable differences in the
number and position of structural Raman bands of H–IM-5
and H–ZSM-5 [40].
Fig. 13 shows the NH3 TPD profiles obtained from H–
IM-5 (i.e., H–IM-5(II) sample) and H–NU-88. The TPD
profiles of the proton form of both zeolites are characterized
by two desorption peaks with maxima in the temperature
Fig. 13. NH TPD profiles from (a) H–IM-5 and (b) H–NU-88 zeolites.
3

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
163
regions 1700–190 and 360–410◦C. Despite the similarity in
their SiO2/Al2O3 ratios, the total area of NH3 desorption
(i.e., the density of acid sites), especially the area of the
high-temperature desorption peak, was found to be much
higher for H–IM-5 than for H–NU-88. Also, the temperature
maximum of the high-temperature desorption peak from
H–IM-5 is slightly higher than that from H–NU-88. This
suggests that the number and strength of strong acid sites
are higher in H–IM-5 than in H–NU-88. However, we cannot
exclude the effects of the nanocrystallinity of H–NU-88 on
its TPD profile because the extent of NH3 re-adsorptionfrom
one acid site to another acid site during the TPD experiment
is greatly influenced by the crystal size of zeolites [41].
Fig. 14 shows the IR spectra of H–IM-5 before and
after pyridine adsorption followed by desorption at different
temperatures. Three well-resolved bands at 3745, 3666, and
3608 cm−1 in the OH stretching region can be observed
from the IR spectrum of dehydrated H–IM-5. The first band
is attributed to silanol groups on the external surface of H–
IM-5 crystallites, and the other two bands can be assigned to
hydroxyl groups belonging to extraframework Al and those
of the Brønsted acid sites, respectively [42]. The presence
of a weak band around 3666 cm−1 in the IR spectrum of
dehydrated H–IM-5 indicates that partial dealuminatio+n has
occurred during the SDA removal and successive NH4 ion
exchange and recalcination steps, as evidenced by 27Al MAS
NMR measurements. After pyridine adsorption the broad
band at 3608 cm−1 disappears completely and two bands
associated with the pyridinium ion adsorbed at Brønsted acid
sites and pyridine coordinated to Lewis acid sites appear
at 1545 and 1455 cm−1, respectively. With the increase
of desorption temperature the intensity of the two bands
at 1545 and 1455 cm−1 is reduced, but some fraction of
both pyridinium ion and pyridine still remains adsorbed at
500 ◦C. This indicates that the IM-5 framework contains
very strong acid sites. The IR spectra of H–NU-88 before
and after pyridine adsorption followed by desorption at
different temperatures are shown in Fig. 15. The spectrum
of dehydrated H–NU-88 in the OH stretching region is
characterized by two weak and broad bands around 3655 and
3570 cm−1, as well as by one strong band at 3745 cm−1
,
typical of surface SiOH groups not involved in hydrogen
bonding. When compared to the spectrum of dehydrated H–
IM-5, therefore, there are notable differences in the position
and intensity of OH bands appearing below 3700 cm−1
.
Fig. 15 also shows that pyridine adsorption gives rise not
only to the notable decrease in intensity of the weak band
around 3655 cm−1 but also to the disappearance of the
very broad band around 3570 cm−1, which is indicative
of interaction of these OH groups with pyridine. Because
the 27Al MAS NMR spectrum of H–NU-88 shows no
noticeable band around 0 ppm, even after heating at 800 ◦C
for 4 h, as described above, the weak OH band around
3655 cm−1 can be attributed to bridging OH groups in Si–
(OH)+–Al Brønsted acid sites rather than to hydroxylated
extraframework Al species. In contrast, assigning the broad
band around 3570 cm−1 simply to bridging Si–(OH)+–
Al species appears to be unreasonable because its position
does not fall into the known range (3660–3600 cm−1) of
stretching for bridging OH groups in zeolites [42]. Although
the exact origin of this band is not clear at this time, one
possible explanation is that it could be due to hydrogen-
bonded Si–(OH)+–Al bridges with SiOH groups. It should
be noted here that a similar OH band appearing below the
typical stretching region (< 3600 cm−1), which disappears
upon adsorption of pyridine, has also been observed for
other zeolites such as SSZ-24 and ITQ-4 [43,44]. As seen
in Fig. 15, on the other hand, the two bands at 1545 and
−1
−1
Fig. 14. IR spectra of H–IM-5 in the (A) 3300- to 3800-cm
and (B) 1400- to 1700-cm
regions (a) before and after pyridine adsorption followed by
◦
desorption at (b) 200, (c) 300, (d) 400, and (e) 500 C for 2 h.

164
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
−1
−1
regions (a) before and after pyridine adsorption followed by
Fig. 15. IR spectra of H–NU-88 in the (A) 3300- to 3800-cm
desorption at (b) 200, (c) 300, (d) 400, and (e) 500 C for 2 h.
and (B) 1400- to 1700-cm
◦
Table 5
be further confirmed by the data in Table 5 where the
amounts of Brønsted and Lewis acid sites in H–IM-5
and H–NU-88 zeolites as a function of temperature were
determined from the intensities of the IR bands at 1545
and 1455 cm−1, respectively, by using the molar extinction
coefficients reported by Emeis [45].
Amounts of pyridine still retained on Brønsted and Lewis acid sites in H–
IM-5 and H–NU-88 after desorption at different temperatures
a
−1
Desorption
temperature ( C)
Amount (µmol g ) of pyridine retained
◦
H–IM-5
Brønsted
H–NU-88
Brønsted
Lewis
Lewis
200
300
400
500
160
133
97
47
33
28
3
121
57
25
7
16
5
4
3.3. Catalysis
31
1
Table 6 lists the physical properties of all zeolites em-
ployed in the skeletal isomerization of 1-butene to isobutene
and the cracking of n-octane. Fig. 16 shows 1-butene con-
version and selectivity to isobutene as a function of time
on stream in the skeletal isomerization of 1-butene on
H–IM-5, H–NU-88, H–ZSM-35, H–ZSM-5, H–EU-1, H–
mordenite, and H–beta zeolites measured at 400 ◦C and
10.1 kPa 1-butene in the feed. Here, we have considered the
results obtained at 5 min on stream as the intrinsic activi-
ties of zeolite catalysts, to minimize the well-known effect
of coke deposition within the zeolite pores on the selectiv-
a
Determined from the intensities of the IR bands at 1545 and
−1
1455 cm , respectively, by using the extinction coefficients given by
Emeis [45].
1455 cm−1 due to pyridine species adsorbed on Brønsted
and Lewis acid sites in H–NU-88, respectively, become
considerably weak after desorption, even at 300 ◦C, unlike
the trend observed for H–IM-5. Therefore, it is clear that
the number and strength of both Brønsted and Lewis acid
sites are higher in H–IM-5 than in H–NU-88. This can
Table 6
Physical properties of zeolite catalysts used in this study
Zeolite
IZA
code
Pore topology
SiO /Al O
2 3
Crystal shape and size (µm)
BET surface
area (m g )
2
a
2
−1
ratio
H–ZSM-5
H–ZSM-35
H–EU-1
MFI
FER
EUO
BEA
MOR
–
–
–
–
2D, 10-rings
2D, 10- and 8-rings
27
27
40
25
33
19
25
32
26
Spherulites, 1–3
398
435
352
667
532
375
460
559
571
Oval plates, 1 × 3
1D, 10-rings + side pockets
Rice-grains, 0.5–1.0
Spherical agglomerates, 3–4
Small rods, 0.1 × 0.3
Rods, 0.1 × 0.5
∗
H–beta
3D, 12-rings
H–mordenite
H–IM-5(I)
H–IM-5(II)
H–IM-5(III)
H–NU-88
2D, 12- and 8-rings
–
–
–
–
Rods, 0.2 × 0.7
Cuboids, 0.1–0.2
Very small spheres, < 0.05
a
BET surface areas calculated from nitrogen adsorption data.

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
165
Fig. 16. (a) 1-Butene conversion and (b) selectivity to isobutene as a function of time on stream in skeletal isomerization of 1-butene on H–IM-5 (2),
◦
H–NU-88 (Q), H–ZSM-35 ("), H–ZSM-5 (F), H–EU-1 (ꢂ), H–mordenite (ꢃ), and H–beta (a) at 400 C and 10.1 kPa 1-butene pressure.
ity to isobutene [13,46–51]. On the other hand, the acidity
and pore size of zeolites and related materials are the two
important factors determining their activity and selectivity
during this isomerization [46–51]. Because there are no sig-
nificant differences in the SiO2/Al2O3 ratio of zeolite cata-
lysts employed in this study (see Table 6), however, the cat-
alytic results in Fig. 16, particularly those obtained at 5 min
on stream, could be mainly regarded as a reflex of geomet-
rical constraints imposed by the particular pore structure of
each zeolite.
size exhibits an excellent stability in time for the skeletal
isomerization of 1-butene, regardless of its SiO2/Al2O3
ratio [13,52]. The isomerization activities of the three H–
IM-5 zeolites with different SiO2/Al2O3 ratios [19–32]
prepared here are compared in Fig. 17. A trend of decreasing
1-butene conversion with increasing time on stream is
clearly observed for all the H–IM-5 zeolites. This led us to
speculate that the pore topology of H–IM-5 is substantially
different than that of H–ZSM-5. Fig. 17 also shows that
H–IM-5 with a lower Al content gives a higher selectivity
to isobutene, although the change in isobutene formation
on H–IM-5 with its Al content is not so high. This is
not unexpected because the extent of successive reactions
leading to by-products such as propene and pentenes should
decrease with decreasing the density of acid sites in the
zeolite framework [48,49].
It is well-established that due to the inherently shape-
selective nature of zeolites and molecular sieves, coke for-
mation on these microporous materials is also a shape-
selective process that depends highly on their pore architec-
ture [31,53]. After the skeletal isomerization of 1-butene at
400 ◦C for 8 h, therefore, the amounts of coke deposited on
all zeolites employed in this study have been determined by
TGA/DTA and are given in Table 7. No significant differ-
ences in the amount (9–10%) of coke formed are observed
for H–ZSM-5 and H–IM-5(I) with the highest Al content (or
the lowest SiO2/Al2O3 ratio) among the H–IM-5 zeolites
prepared here. As seen in Table 7, however, the amount of
coke deposited on H–IM-5 increases significantly with de-
creasing Al content in the zeolite. As a result, the amount
(14.9%) of coke formed on H–IM-5(III) with the lowest Al
content was found to be approximately one and half times
larger than that (9.8%) on H–IM-5(I). Notice that this coke
level is much higher than the amount (10.6%) of coke on
H–mordenite, a two-dimensional pore system consisting of
ZSM-35 is the high-silica analog of ferrierite, which
contains a two-dimensional pore system consisting of 10-
ring (4.2 × 5.4 Å) channels intersected by 8-ring (3.5 ×
4.8 Å) channels [21] and is known as the best catalyst for this
reaction [46–51]. When H–IM-5 is compared with H–ZSM-
35 under the same reaction conditions, the former zeolite
gives a much higher initial 1-butene conversion (85% vs
42%). However, H–IM-5 produces a much larger amount of
by-products (mainly propene and pentenes) than H–ZSM-35
from the beginning of the reaction, revealing its nonselective
behavior for isobutene formation. This indicates that the free
diameter of pores in H–IM-5 is considerably larger than
that at the H–ZSM-35 intersection, giving rise to a more
efficient accommodation of two n-butenes for side reactions
such as their dimerization followed by cracking to light
hydrocarbons. Fig. 16 also shows that 1-butene conversion
on H–IM-5 decreases continuously with increasing time
on stream, whereas the opposite holds for selectivity to
isobutene. This trend is somewhat different from that found
in H–ZSM-5 with an intersecting 10-ring pore system
between the straight (5.3 × 5.6 Å) and the sinusoidal (5.1 ×
5.5 Å) channels because the extent of changes in 1-butene
conversion or selectivity to isobutene with time on stream is
higher on H–IM-5 than on H–ZSM-5. In fact, it has been
repeatedly reported that H–ZSM-5 with the uniform pore

166
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
Fig. 17. (a) 1-Butene conversion and (b) selectivity to isobutene as a function of time on stream in skeletal isomerization of 1-butene on H–IM-5 zeolites with
different SiO /Al O ratios: 19 (2), 25 (Q), and 32 ("). The reaction conditions are the same as those given in Fig. 16.
2
2 3
intersecting 12-ring (6.5 × 7.0 Å) and 8-ring (3.4 × 4.8 Å)
channels and is even higher than that (10.8%) observed
for H–beta with mutually perpendicular intersecting 12-ring
(5.6 × 5.6 and 6.6 × 6.7 Å) channels [21], despite the sim-
ilarity in their Al contents. As described earlier, the amount
of Al atoms removed from IM-5 framework sites during the
calcination step is smaller for the zeolite with a lower Al
content. Thus, if IM-5 is not a three-dimensional large-pore
material like beta but a zeolite with intersecting 12- and 10-
ring pores, the possibility of blockage of the inner pores by
extraframework Al species should be lower for the H–IM-
5(III) zeolite. This appears to be reasonable because there
is a notable increase in the amount of coke formed on H–
IM-5, as well as in its N2 BET surface area, with increas-
ing the SiO2/Al2O3 ratio (see Table 7). It is well known
that the coke-forming propensity of medium-pore zeolites
such as H–ZSM-5 and H–ZSM-11 is lower than that of zeo-
lites with large 12-ring windows, due to the geometric con-
straints imposed by 10-ring pores [31,53]. Although coke
formation during the skeletal isomerization of 1-butene oc-
curs on the exterior surface of zeolite catalysts, as well as
within their inner pores, therefore, the TGA/DTA results in
Table 7 strongly suggests that IM-5 may belong to the struc-
tural regime of large-pore materials rather than of medium-
pore zeolites.
Table 7
Amounts of coke deposits formed on various zeolite catalysts during the
1-butene skeletal isomerization and n-octane cracking
a
b
Catalyst
SiO /Al O
2 3
Amount (wt%) of coke deposited
2
ratio
1-Butene skeletal
isomerization
n-Octane cracking
H–ZSM-5
H–ZSM-35
H–EU-1
27
27
40
25
33
19
25
32
26
9.0
4.1
4.4
10.8
10.6
9.8
11.1
14.9
12.2
3.2
–
2.2
7.8
11.3
2.8
3.5
7.0
10.0
c
H–beta
H–mordenite
H–IM-5(I)
H–IM-5(II)
H–IM-5(III)
H–NU-88
a
After 8 h on stream. The reaction temperatures for 1-butene skeletal
◦
isomerization and n-octane cracking are 400 and 500 C, respectively. The
other conditions for both reactions are the same as those stated in the text.
b
Determined from TGA/DTA.
Not reacted.
c
sion and selectivity to isobutene were found to be 67 and
20% for this one-dimensional medium-pore zeolite, respec-
tively, which are in good agreement with the results reported
by Guisnet et al. [54]. Thus, the initial 1-butene conversion
on H–EU-1 is considerably lower than that on H–IM-5, and
the opposite is observed for the selectivity to isobutene. As
seen in Fig. 16, this trend remains unchanged during the pe-
riod of 8 h on stream. Further support for the speculation
that the IM-5 structure is not formed by unidirectional 10-
ring pores with large lobes can be obtained from the fact
that the amount (4.4%) of coke formed on H–EU-1 is much
smaller than that observed for any of the three H–IM-5 ze-
olites studied here (see Table 7). Like H–IM-5, on the other
hand, H–NU-88 exhibits the initial 1-butene conversion sim-
ilar to that of H–ZSM-5, revealing its nonselective behav-
ior for isobutene formation. However, this zeolite shows a
Based on their results from several test reactions and
adsorption of hydrocarbons with different sizes, Corma
et al. [6] considered the possibility that IM-5 could possess
a pore system of unidirectional 10-rings with large internal
cavities. However, this can be ruled out by comparing the
1-butene isomerization activity of H–IM-5 with that of H–
EU-1, whose structure consists of one-dimensional 10-ring
(4.1 × 5.4 Å) pores in the main channel bounded by 12-ring
pores leading to large side pockets (6.8 × 5.8 Å in cross sec-
tion and 8.1 Å in depth) [21]. The initial 1-butene conver-

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
167
large decrease in 1-butene conversion during the first 1 h on
stream. As seen in Table 7, in addition, the amount (12.2%)
of coke on H–NU-88 is comparable to that formed on H–
mordenite or H–beta. Therefore, it appears that H–NU-88
may meet the catalytic behavior of large-pore zeolites, which
is in contrast to the proposal of Lacombe et al. [8].
different from those for the other four zeolite catalysts with
known structures in that H–IM-5 and H–NU-88 give a higher
initial selectivity to propane, although the amount of propane
produced becomes lower at higher times on stream. We also
note that the production of aromatic hydrocarbons such as
benzene, xylenes, and toluene is negligible for all zeolite
catalysts studied here. To investigate the effect of reaction
temperature on product selectivities, on the other hand, we
have carried out the n-octane cracking reaction on H–IM-
5 and H–NU-88 in the temperature region 310–500◦C.
Selectivities to ethylene and propylene on both zeolites were
enhanced at higher temperature, whereas the opposite was
observed for the selectivities to longer hydrocarbons,mainly,
n-butane and butenes. This is not unexpected because the
extent of secondary cracking activity of zeolites normally
becomes larger at a higher temperature.
Fig. 18 shows n-octane conversion and selectivities to
ethylene and propylene as a function of time on stream in n-
octane cracking on H–IM-5, H–NU-88, H–ZSM-5, H–EU-1,
H–mordenite, and H–beta zeolites at 500 ◦C and 2.0 kPa n-
octane pressure in the feed. It can be seen that all the zeolite
catalysts except H–ZSM-5 exhibit a continuous decrease in
n-octane conversion during the period of 8 h on stream,
although their initial conversions are different from one
another. A similar result was observed when selectivities
to ethylene on these five zeolites are plotted with time
on stream. As shown in Fig. 18, however, differences in
Table 8 lists the conversions and product distributions
from the catalytic cracking of n-octane on H–IM-5 and H–
NU-88 with a similar SiO2/Al2O3 ratio (∼ 25) measured
at 500 ◦C, 2.0 kPa n-octane pressure in the feed and 500
s on stream. For comparison, the catalytic results from
H–ZSM-5, H–EU-1, H–mordenite, and H–beta are also
given in Table 8. These data reveal that the n-octane
conversion (89%) on H–IM-5 is almost the same as that
(90%) on H–ZSM-5, serving as a prime example of the
uniqueness of medium-pore zeolites in the cracking of linear
paraffins [31,55,56]. Thus, no severe spatial constraints
for the free diffusion of n-octane molecules appear to
exist within the pores of IM-5. This may also be the
case of H–NU-88 because it yields an n-octane conversion
of 91% that is very close to that on H–ZSM-5 but is
considerably higher than the value observed for H–EU-1, H–
mordenite, or H–beta. The catalytic data in Table 8 also show
that ethylene, propane, and propylene are the three most
dominant products observed for H–IM-5 and H–NU-88. The
other major products include ethane, n-butane, isobutane,
and butenes. These product distributions are somewhat
Table 8
◦
a
Conversion and product distribution from catalytic cracking of n-octane on various zeolite catalysts at 500 C
H–IM-5
H–NU-88
H–ZSM-5
H–EU-1
H–mordenite
H–beta
Conversion (mol%)
Selectivity (mol%) to
Methane
Ethane
Ethylene
Propane
Propylene
n-Butane
Isobutane
89.3
91.0
89.8
83.1
64.3
76.1
2.6
4.4
21.6
21.6
23.0
6.5
6.5
3.9
2.1
1.4
0.1
0.6
0.2
1.3
0.1
0.5
0.1
–
3.6
3.2
18.3
18.4
23.9
7.7
8.8
5.0
2.2
1.5
0.2
0.9
0.2
1.5
0.1
0.4
–
0.1
0.1
0.9
1.9
–
0.2
–
0.6
0.3
2.7
6.9
27.3
14.0
27.7
5.8
1.5
4.3
1.9
1.3
0.1
0.2
0.2
2.0
0.1
0.3
0.1
0.2
0.2
0.7
1.4
0.2
0.1
0.2
0.5
0.1
2.4
4.3
20.1
17.8
28.0
6.0
4.5
6.3
2.7
1.9
0.2
0.3
0.2
1.7
0.1
0.5
–
0.2
0.1
0.5
1.1
0.2
0.2
0.2
0.5
–
2.8
–
6.0
4.1
15.2
10.8
25.4
7.3
6.2
8.1
3.5
2.6
0.3
0.8
0.4
3.2
0.2
0.9
–
0.4
0.1
0.9
2.1
–
0.3
0.1
1.0
0.1
21.0
16.3
26.0
6.6
8.3
6.9
3.1
2.2
0.3
0.9
0.3
1.6
0.2
0.8
–
0.2
0.1
0.4
0.7
0.4
0.2
0.1
0.4
0.2
1-Butene
cis-2-Butene
trans-2-Butene
n-Pentane
2-Methylbutene
1-Pentene
2-Pentene
cis-2-Pentene
trans-2-Pentene
n-Hexane
2-Hexene
3-Methylpentane
Benzene
Toluene
2-Methyheptane
o-Xylene
m-Xylene
p-Xylene
0.1
0.8
1.4
0.1
0.1
–
0.4
0.6
C
9
>
a
3
−1
Data are reported as the values of 500 s on stream at a total gas flow of 50 cm min .

168
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
Fig. 18. (a) n-Octane conversion and selectivities to (b) ethylene and (c) propylene as a function of time on stream in n-octane cracking on H–IM-5 (2),
◦
H–NU-88 (Q), H–ZSM-5 ("), H–EU-1 (F), H–mordenite (ꢂ), and H–beta (ꢃ) at 500 C and 2.0 kPa n-octane pressure.
Fig. 19. (a) n-Octane conversion and selectivities to (b) ethylene and (c) propylene as a function of time on stream in n-octane cracking on H–IM-5 zeolites
with different SiO /Al O ratios: 19 (2), 25 (Q), and 32 ("). The reaction conditions are the same as those given in Fig. 18.
2
2 3
the initial selectivities to propylene on these zeolites and
changes with time on stream are not so large as those
observed for the selectivity to ethylene. Therefore, it is again
clear that both IM-5 and NU-88 are not multidimensional
medium-pore zeolites because they are much less stable than
H–ZSM-5 with a similar SiO2/Al2O3 ratio (∼ 25). It should
also be noted here that the deactivation pattern and product
distribution of H–NU-88 with time on stream are quite
similar to those observed for H–beta. The cracking activities
of three H–IM-5 zeolites with different SiO2/Al2O3 ratios
are shown in Fig. 19. When the IM-5(I) zeolite with
the lowest SiO2/Al2O3 ratio is compared with H–ZSM-
5, n-octane conversions and selectivities to ethylene and
propylene on both zeolites remain almost constant over the
period of time studied here. However, the other two IM-
5 zeolites with higher SiO2/Al2O3 ratios (and hence with
higher N2 BET surface areas) deactivate continuously during
the n-octane cracking. This is significantly different than the
trend found in four H–ZSM-5 zeolites with a SiO2/Al2O3
ratio ranging from 27 to 170 in that there is no noticeable
decrease in n-octane conversion on these H–ZSM-5 zeolites
with time on stream up to 8 h, although the level of
conversion depends highly on their SiO2/Al2O3 ratio [57].
The TGA/DTA results in Table 7 reveal that the amount
of coke formed on H–IM-5 after the n-octane cracking
at 500 ◦C for 8 h increases with increasing SiO2/Al2O3
ratio in the zeolite, which is also opposite to the trend
found in a series of used H–ZSM-5 zeolites with different
SiO2/Al2O3 ratios [57]. These arguments taken in total led
us to believe that the cracking behavior of IM-5 cannot
be represented by a multidimensional, 10-ring system with
the uniform pore size like ZSM-5. Due to the similarity

S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
169
in the deactivation behavior of all zeolites employed in
4. Conclusions
this study, except H–ZSM-5, however, it is very difficult to
reasonably infer the pore structures of IM-5 and NU-88 from
the catalytic results in Figs. 18 and 19. As stated earlier, the
pore structure of zeolites exerts a very large effect on the rate
of coke formation in a number of acid-catalyzed reactions.
Therefore, we focused on the relationships between the
structural features of H–mordenite, H–EU-1, and H–beta,
and their amounts of coke formed after the n-octane cracking
at 500 ◦C for 8 h. Based on the relationships drawn, we then
attempted to estimate the most probable pore topologies of
IM-5 and NU-88 from their coke-forming propensities.
Mordenite has a two-dimensional pore system consisting
of 12-ring channels intersected by 8-ring channels. However,
this zeolite should be effectively a one-dimensional 12-ring
material as far as the n-octane cracking is concerned because
its small 8-rings may slow down the overall diffusivity.
Fig. 18 shows that H–mordenite exhibits a much lower
n-octane conversion than H–IM-5 or H–NU-88 from the
beginning of the reaction, revealing a much more severe
diffusional limitation in the former zeolite. As seen in
Table 7, in addition, the amount of coke deposited on H–
mordenite is considerably higher than that not only on any
of H–IM-5 zeolites studied here but also on H–NU-88.
When H–EU-1 is compared with H–mordenite, on the other
hand, the former zeolite exhibits a much higher n-octane
conversion over the period of time studied here. Because
the 10-ring (4.1 × 5.3 Å) pores in EU-1 are considerably
smaller than the 12-ing (6.5 × 7.0 Å) channels in mordenite,
however, the diffusion of n-octane molecules should be less
favorable in the former zeolite. Thus, it is not difficult to
speculate that the external acid sites on the H–EU-1 crystals,
on which catalysis is not influenced by shape selectivity, may
contribute mainly to the cracking of n-octane. This can be
further supported the fact that the amount (2.2%) of coke
on H–EU-1 for 8 h on stream is much smaller than that
(11.3%) on H–mordenite, which is also considerably small
compared to H–IM-5 or H–NU-88. Therefore, we can again
rule out the possibility that the IM-5 and NU-88 structures
consist of a pore topology of unidirectional 12-ring channels
or 10-rings with large internal cavities. When H–beta,
a three-dimensional large-pore material, is compared with
H–NU-88, finally, the latter zeolite gives a higher amount
of coke deposited, but its value is still smaller than the
amount formed on H–mordenite (see Table 7). Clearly, it
would be very difficult, or even impossible, to accurately
determine the pore topologies of IM-5 and NU-88 from their
isomerization and cracking activities alone. When correlated
with the coke-forming propensities of these two zeolites
with unknown structures, however, it can be concluded that
both materials are identified as multidimensional, large-pore
zeolites. Therefore, the overall catalytic results do support
the conclusion that IM-5 is a new multidimensional large-
pore zeolite probably formed by intersecting 12- and 10-
ring windows, whereas NU-88 is an intergrowth of several
hypothetical polymorphs in the beta family of zeolites.
Two new zeolites IM-5 and NU-88 were found to crys-
tallize over a narrow range of SiO2/Al2O3 and NaOH/SiO2
ratios when the flexible, doubly charged ammonium cations
1,5-bis(N-methylpyrrolidinium)pentane and 1,6-bis(N-me-
thylpyrrolidinium)hexaneare used as organic SDAs together
with Na+ ions, respectively. The physicochemical proper-
ties of IM-5 and NU-88 prepared here are investigated by
powder XRD, elemental and thermal analyses, TEM, Ar and
N2 adsorption, multinuclear solid-state MAS NMR, IR, and
Raman, NH3 TPD, and IR spectroscopy of adsorbed pyri-
dine. The overall characterization results suggest that IM-5
is a new multidimensional large-pore zeolite that probably
consists of intersecting 12- and 10-ring windows, whereas
NU-88 is a nanocrystalline material that is closely related
to the beta family of zeolites. In addition, the strength of
both Brønsted and Lewis acid sites was found to be stronger
in H–IM-5 than in H–NU-88. Despite its nanocrystallinity,
however, NU-88 exhibits a very high thermal stability. The
catalytic properties of these two zeolites are evaluated for
the skeletal isomerization of 1-butene to isobutene and the
cracking of n-octane and compared to those from various ze-
olites with known structures. A combination of the catalytic
results from all zeolites studied here with their coke-forming
propensities allowed us to provide further evidence that both
IM-5 and NU-88 are multidimensional, large-pore zeolites.
Acknowledgments
Financial support for this work was provided by the Ko-
rea Research Foundation (1999-E00325).We thank Dr. M.A.
Camblor (Industrias Quimicas del Ebro) for helpful discus-
sion.
References
[1] N.Y. Chen, W.E. Garwood, F.G. Dwyer, Shape Selective Catalysis in
Industrial Applications, Marcel Dekker, New York, 1989.
[2] S. Bhatia, Zeolite Catalysis: Principles and Applications, CRC, Boca
Raton, FL, 1990.
[3] E. Benazzi, J.L. Guth, L. Rouleau, PCT WO 98/17581, 1998.
[4] E. Benazzi, J.L. Guth, L. Rouleau, US Patent 6,136,290, 2000.
[5] J.L. Casci, S. Maberly, E. Benazzi, L. Rouleau, R.P. Henney, US Patent
6,027,707, 2000.
[6] A. Corma, A. Chica, J.M. Guil, F.J. Llopis, G. Mabilon, J.A. Perdigon-
Melon, S. Valencia, J. Catal. 189 (2000) 382.
[7] A. Corma, J. Martinez-Triguero, S. Valencia, E. Benazzi, S. Lacombe,
J. Catal. 206 (2002) 125.
[8] S. Lacombe, A. Patrigeon, E. Benazzi, Stud. Surf. Sci. Catal. 135
(2001) 4272.
[9] R. Szostak, Handbook of Molecular Sieves, Van Nostrand Reinhold,
New York, 1992.
[10] A. Moini, K.D. Schmitt, E.W. Valyoscik, R.F. Polomski, Zeolites 14
(1994) 504.
[11] A. Moini, K.D. Schmitt, R.F. Polomski, Zeolites 18 (1997) 2.

170
S.-H. Lee et al. / Journal of Catalysis 215 (2003) 151–170
[12] W.C. Paik, C.-H. Shin, S.B. Hong, Chem. Commun. (2000) 1609.
[13] S.-H. Lee, D.-K. Lee, C.-H. Shin, W.C. Paik, W.M. Lee, S.B. Hong,
J. Catal. 196 (2000) 158.
[14] W.C. Paik, C.-H. Shin, J.M. Lee, B.J. Ahn, S.B. Hong, J. Phys. Chem.
B 105 (2001) 9994.
[15] T.J. Kim, W.S. Ahn, S.B. Hong, Micropor. Mater. 7 (1996) 35.
[16] J.L. Casci, B.M. Lowe, B.M. Whittam, US Patent 4,537,754, 1985.
[17] H.H. Cho, S.H. Kim, Y.G. Kim, Y.C. Kim, H. Koller, M.A. Camblor,
S.B. Hong, Chem. Mater. 12 (2000) 2292.
[18] M.A. Camblor, P.A. Barrett, M.-J. Diaz-Cabanas, L.A. Villaescusa, M.
Puche, T. Boix, E. Perez, H. Koller, Micropor. Mesopor. Mater. 48
(2001) 11.
[19] D.-Y. Han, A.J. Woo, I.-S. Nam, S.B. Hong, J. Phys. Chem. B 106
(2002) 6206.
[20] A. Le Bail, J. Solid State Chem. 83 (1989) 267.
[21] International Zeolite Association, Structure Committee, http://www.
iza-structure.org.
[36] J.C. Jansen, F.J. van der Gaag, H. van Bekkum, Zeolites 4 (1984) 369.
[37] F.R. Dollish, W.G. Fateley, F.F. Bentley, Characteristic Raman Fre-
quencies of Organic Compounds, Wiley, New York, 1974.
[38] D. Lin-Vien, N.B. Colthup, W.G. Fateley, J.G. Grasselli, Handbook of
Infrared and Raman Characteristic Frequencies of Organic Molecules,
Academic Press, San Diego, CA, 1991.
[39] L. Tosheva, B. Mihailova, V. Valtchev, J. Sterte, Micropor. Mesopor.
Mater. 48 (2001) 31.
[40] P.K. Dutta, K.M. Rao, J.Y. Park, J. Phys. Chem. 95 (1991) 6654.
[41] M. Sawa, M. Niwa, Y. Murakami, Zeolites 10 (1990) 307.
[42] H.G. Karge, M. Hunger, H.K. Beyer, in: J. Weitkemp, L. Puppe (Eds.),
Catalysis and Zeolites: Fundamentals and Applications, Springer,
Berlin, 1999, p. 198.
[43] J. Martinez-Triguero, M.J. Diaz-Cabanas, M.A. Camblor, V. Fornes,
Th.L.M. Maesen, A. Corma, J. Catal. 182 (1999) 463.
[44] L.A. Villaescusa, P.A. Barrett, M. Kalwei, H. Koller, M.A. Camblor,
Chem. Mater. 13 (2001) 2332.
[22] G. Horvarth, K.J. Kawazoe, J. Chem. Eng. Jpn. 16 (1983) 470.
[23] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity,
Academic Press, London, 1982.
[45] C.A. Emeis, J. Catal. 141 (1993) 347.
[46] M. Guisnet, P. Andy, N.S. Gnep, E. Benazzi, C. Travers, J. Chem. Soc.
Chem. Commun. (1995) 1685.
[24] M.A. Camblor, A. Corma, S. Valencia, J. Mater. Chem. 8 (1998) 2137.
[25] C. Suryanarayana, M.G. Norton, X-Ray Diffraction: A Practical
Approach, Plenum, New York, 1998.
[26] B.J. Schoeman, J. Sterte, J.E. Otterstedt, J. Chem. Soc. Chem. Com-
mun. (1993) 994.
[47] W.-Q. Xu, Y.-G. Yin, S.L. Suib, J.C. Edwards, C.-L. O’Young, J. Phys.
Chem. 99 (1995) 9443.
[48] G. Seo, H.S. Jeong, S.B. Hong, Y.S. Uh, Catal. Lett. 36 (1996) 249.
[49] G. Seo, H.S. Jeong, D.-L. Jang, D.L. Cho, S.B. Hong, Catal. Lett. 41
(1996) 189.
[27] M. Tsapatsis, M. Lovallo, T. Okubo, M.E. Davis, M. Sadakada, Chem.
Mater. 7 (1995) 1734.
[50] P. Meriaudeau, A. Vu Tuan, N. Le Hung, G. Szabo, Catal. Lett. 47
(1997) 71.
[28] A.E. Persson, B.J. Schoeman, J. Sterte, J.E. Otterstedt, Zeolites 15
(1995) 611.
[51] G. Seo, M.-W. Kim, J.-H. Kim, B.J. Ahn, S.B. Hong, Y.S. Uh, Catal.
Lett. 55 (1998) 105.
[29] M.A. Camblor, A. Corma, S. Valencia, Micropor. Mesopor. Mater. 25
(1998) 59.
[52] J. Houzvicka, R. Klik, L. Kubelkova, V. Ponec, Appl. Catal. A 150
(1997) 101.
[30] E.P. Barrett, L.S. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951)
373.
[31] N.Y. Chen, T.F. Degnan Jr., C.M. Smith, Molecular Transport and
Reaction in Zeolites, VCH, New York, 1994.
[32] J. Hedlund, B.J. Schoeman, J. Sterte, Chem. Commun. (1997) 1193.
[33] N.R. Castagnola, P.K. Dutta, J. Phys. Chem. B 102 (1998) 1696.
[34] E.M. Flanigen, ACS Symp. Ser. 171 (1976) 80.
[35] G. Coudurier, C. Naccache, J.C. Vedrine, J. Chem. Soc. Chem.
Commun. (1982) 1413.
[53] H.G. Karge, Stud. Surf. Sci. Catal. 137 (2001) 707.
[54] M. Guisnet, P. Andy, N.S. Gnep, C. Travers, E. Benazzi, in: M.M.J.
Treacy, B.K. Marcus, M.E. Bisher, J.B. Higgins (Eds.), Proceedings of
the 12th International Zeolite Conference, Materials Research Society,
Warrendale, PA, 1999, p. 1433.
[55] W.O. Haag, R.M. Lago, P.B. Weisz, Nature 309 (1984) 589.
[56] B.A. Watson, M.T. Klein, R.H. Harding, Ind. Eng. Chem. Res. 35
(1996) 1506.
[57] S.-H. Lee, C.-H. Shin, S.B. Hong, unpublished.