A priori phase prediction of zeolites: Case study of the structure-directing effects in the synthesis of MTT-type zeolites

Article abstract of DOI:10.1021/ja070303u

This study first uses molecular modeling to examine the structure-directing effects of small amines that are selective for the crystallization of MTT-type zeolite phases. The optimized van der Waals interactions of these small amines are compared within the one-dimensional pore zeolites with the MTT, TON, and MTW frameworks. From these results and our previous molecular modeling studies of structure-directing agents (SDA) for MTT-type zeolites, a large number of amines or quaternary ammonium molecules are successfully predicted to be selective for MTT phases. These molecules were chosen by matching the crystallographic periodicity of the pore structure with the distances between the centers of branched groups in these molecules. These molecules vary in length and in the number of branched moieties, and a few of these molecules are polymeric or oligomeric. In test cases where the distances between the branched groups are not multiples of the pore periodicity, with few exceptions these molecules usually do not produce MTT phases. Finally, we discuss the inorganic conditions necessary for crystallization of MTT phases in borosilicate preparations with some of the diamines in this investigation.

Full text of DOI:10.1021/ja070303u

Published on Web 05/25/2007
A Priori Phase Prediction of Zeolites: Case Study of the
Structure-Directing Effects in the Synthesis of MTT-Type
Zeolites
Allen W. Burton
Contribution from the CheVron Energy Technology Company, Richmond, California 94802
Received January 23, 2007; E-mail: buaw@chevron.com
Abstract: This study first uses molecular modeling to examine the structure-directing effects of small amines
that are selective for the crystallization of MTT-type zeolite phases. The optimized van der Waals interactions
of these small amines are compared within the one-dimensional pore zeolites with the MTT, TON, and
MTW frameworks. From these results and our previous molecular modeling studies of structure-directing
agents (SDA) for MTT-type zeolites, a large number of amines or quaternary ammonium molecules are
successfully predicted to be selective for MTT phases. These molecules were chosen by matching the
crystallographic periodicity of the pore structure with the distances between the centers of branched groups
in these molecules. These molecules vary in length and in the number of branched moieties, and a few of
these molecules are polymeric or oligomeric. In test cases where the distances between the branched
groups are not multiples of the pore periodicity, with few exceptions these molecules usually do not produce
MTT phases. Finally, we discuss the inorganic conditions necessary for crystallization of MTT phases in
borosilicate preparations with some of the diamines in this investigation.
I. Introduction
mechanisms of self-assembly that determine which zeolite
phases are promoted under a given set of synthesis conditions.
Both the inorganic and organic (typically an amine or quaternary
ammonium compound) components of a synthesis gel influence
the zeolite phases that may crystallize. Since the organic species
are a strong determinant of the phases that are crystallized, they
are often referred to as “structure-directing agents” (SDAs).
It is important to appreciate that zeolites are metastable
phases. For identical gel compositions with the same SDA,
changes in temperature, hydroxide concentration, or sources of
inorganic reactants or even the presence of seeds may influence
the phase that is crystallized. Furthermore, Ostwald ripening
may occur in which a zeolite phase converts to either a more
thermodynamically favored zeolite phase or a dense silica phase
such as quartz. Because crystallization products from zeolite
syntheses are kinetically driven, a hypothetical SDA/zeolite
composite that represents a thermodynamic minimum (when
compared to the same SDA within other zeolite frameworks)
may not necessarily crystallize in a given system.
However, Sastre et al. reasoned that the relative stabilities of
potentially competing nuclei may “give us an estimation of
which zeolite is preferentially forming during the early stages
of the synthesis.”14 In this simplified treatment, by assuming
that the structures of competing nuclei resemble the structures
of their respective SDA/framework composites, we can ap-
proximate the relative energies of those competing nuclei. This
reasoning is similar to that presented by the late Barrie Lowe,
who suggested that nuclei with greater free energy are more
Zeolites are an important class of crystalline microporous
materials with industrial applications in ion exchange, adsorp-
tion, and catalysis. High-silica zeolites are particularly vital in
the synthesis of petrochemicals and refining of petroleum.1-6
They are used as solid acid catalysts in the alkylation of benzene
to cumene and ethylbenzene,7,8 the isomerization of xylene,8
and the transalkylation of aromatics (e.g., toluene dispropor-
tionation).8,9 Siliceous zeolites are essential additives in modern
catalytic cracking units,10,11 and they are also used to reduce
the pour point in motor lubricant stocks either by selective
cracking or by selective isomerization of long-chain paraf-
fins.3,12,13
For over half a century, zeolites have been synthesized in
industrial and academic laboratories. Despite their widespread
use as catalysts, we still do not fundamentally understand the
(1) Zeolites for Cleaner Technologies; Guisnet, M., Gilson, J.-P., Eds.; Imperial
College: London 2002.
(2) Weitkamp, J.; Puppe, L. Catalysis and Zeolites: Fundamentals and
Applications; Springer: New York 1999.
(3) Degnan, T. Top. Catal. 2000, 13, 349-356.
(4) Chen, N. Y. Shape SelectiVe Catalysis in Industrial Applications; Marcel
Dekker: New York 1996.
(5) Corma, A. Catal. Lett. 1993, 22, 33-52.
(6) Bhatia, S. Zeolite Catalysis: Principles and Applications; CRC Press: Boca
Raton, FL, 1990.
(7) Perego, C.; Ingallina, P. Catal. Today 2002, 23, 3-22.
(8) Cejka, J.; Wichterlova, B. Catal. ReV. 2002, 44, 375-421.
(9) Tsai, T.-C.; Liu, S.-B.; Wang, I. Appl. Catal. A: Gen. 1999, 181, 355-
398.
(10) Scherzer, J. Octane-Enhancing Zeolite FCC Catalysts: Scientific and
Technical Aspects; Marcel Dekker: New York, 1990.
(11) Fluid Cracking Catalysts; Occelli, M., O’Connor, P., Eds.; Marcel
Dekker: New York, 1998.
(12) Sivasanker, S. Bull. Catal. Soc. India 2003, 2, 100-106.
(13) Miller, S. U.S. Patent 5135638, 1992.
(14) Sastre, G.; Leiva, S.; Sabater, M. J.; Gimenez, I.; Rey, F.; Valencia, S.;
Corma, A. J. Phys. Chem. B 2003, 107 (23), 5432-5440.
9
10.1021/ja070303u CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 7627-7637
7627

A R T I C L E S
Burton
unique zeolite framework topology assigned by the Structure
Commission of the International Zeolite Association. The
Appendix in the Supporting Information contains images of all
the frameworks discussed in this paper. Although there are
several examples of zeotype frameworks that may be prepared
in both aluminosilicate and aluminophosphate forms, they are
rarely prepared using SDAs with identical backbone structures.
The case of MCM-61 and Mu-13 is therefore a very notable
example of structure direction.
After we determine the zeolite phases that are directed by a
particular organic, we frequently perform molecular modeling
to show, a posteriori, that there is a good energetic fit of the
molecule within the host framework. Despite the existence of
powerful computing algorithms22-24 for building SDA molecules
within zeolite void spaces, rarely in the literature do we find
examples in which a previously unstudied molecule is accurately
predicted to crystallize a particular zeolite. Such predictions
often require that an unusual or dominant feature in the zeolite
be consistent with the size and shape of the proposed molecule.
The first definitive example of this occurred when Schmitt
and Kennedy used molecular modeling to identify candidate
molecules to replace the triquat shown in Figure 1b as an SDA
for ZSM-18.25 In this example the investigators sought a
molecule that possessed the same charge as well as a similar
size and symmetry as the first SDA found to make ZSM-18.
This led them to consider the trisquaternary molecule shown in
Figure 1c. Using this molecule they were able to prepare samples
of ZSM-18 that could be calcined to remove the organic without
loss of crystallinity (as they had observed with the first triquat).
Recently Casci and co-workers used molecular modeling to
identify the C5-N,N′-bis(1-methylpyrrolidinium) diquat as a
candidate for the synthesis of NES-type zeolites.26 Sastre and
co-workers also recently successfully applied molecular model-
ing to find an SDA that was selective for ITQ-7 (ISV) rather
than the structurally similar ITQ-17 (BEC).27 To that point, only
one other SDA had been reported to crystallize ITQ-7. ITQ-
17, on the other hand, crystallizes from germanosilicate
compositions in the presence of several different quaternary
ammonium molecules. Their study was motivated by the fact
that the precursor amine used to prepare the previously
discovered SDA for ITQ-7 is no longer commercially available.
Another example of a priori zeolite phase prediction involves
the synthesis of ITQ-24. In their studies using cyclohexylpyr-
rolidine derivatives, Rey and co-workers found that one of their
diquaternary SDAs promoted the synthesis of ITQ-23 in gels
with Si/Al ) 50 to θ.28 ITQ-23 belongs to the SSZ-26/33
intergrowth family of zeolites described by Lobo et al.29 In the
structure elucidation of these zeolites, Lobo described the 26/
Figure 1. (a) 18-Crown-6 molecule enclathrated within the 18 ring cages
of MCM-61 and Mu-13 (only oxygen and carbon atoms shown (from ref
20). (b) Triquat 2 within the MEI cage. (c) Triquat 2 within the MEI cage.
C, N, O, and H atoms are represented by gray, blue, red, and white spheres,
respectively.
likely to redissolve than to participate in crystallization.15
Although a phase predicted to be thermodynamically favored
does not always crystallize, we can say that it is more probable
that a phase will form if it is energetically favored. This is borne
out by the fact that there is a “good” calculated fit for many
observed SDA/framework composites, especially in cases where
a particular siliceous framework is rarely or never observed to
crystallize with other SDA molecules (vide infra). In previous
work we identified three major contributions to the energy of
an SDA/framework composite: the inherent energy of the
defect-free framework, the interactions between the SDA and
the framework (in which we include the conformational energy
of the occluded SDA relative to the minimum energy conforma-
tion), and the concentration of silanol/siloxy defects in the final
product.16 In general, defect-free zeolite structures with similar
framework density tend to have similar framework energies.17,18
If the concentration of silanol/siloxy defects (due to a greater
concentration of extraframework charge relative to the concen-
tration of framework aluminum) is similar in two zeolites with
similar framework densities, then the relative energies of the
two phases will largely be determined by the interactions
between the SDA and the respective frameworks. In this respect,
molecular modeling is useful for examining the structure-
directing effects of known molecules as well as predicting
candidate SDA molecules for a given zeolite structure.
Excellent examples of structure direction are the 18-Crown-6
molecules in the synthesis of MCM-61 (and its aluminophos-
phate analogue Mu-13) and the triquaternary ammonium (tri-
quat) molecule used in the synthesis of ZSM-18.19 In both cases
the size and symmetry of the molecules are reflected in the cages
within which they are occluded (Figure 1a-c). The alumino-
silicate MCM-6120 and the aluminophosphate Mu-1321 both
possess the MSO framework. The three-letter code indicates a
(22) Lewis, D. W.; Willock, D. J.; Catlow, C. R. A.; Thomas, J. M.; Hutchings,
G. J. Nature 1996, 382, 604.
(23) Lewis, D. W.; Sankar, G.; Wyles, J.; Thomas, J. M.; Catlow, C. R. A.;
Willock, D. J. Angew. Chem. 1997, 36, 2675.
(24) Willock, D. J.; Lewis, D. W.; Catlow, C. R. A.; Hutchings, G. J.; Thomas,
J. M. J. Mol. Catal. A 1997, 119, 415-424.
(25) Schmitt, K.; Kennedy, G. Zeolites 1994, 14, 635.
(15) Lowe, B. M. Zeolites 1983, 3 (4), 300-305.
(16) Burton, A. W.; Lee, G. S.; Zones, S. I. Microporous Mesoporous Mater.
2006, 90 (1-3), 129-144.
(26) Casci, J. L.; Cox, P. A.; Henney, R. P. G.; Maberly, S.; Shannon, M. D.
Book of Abstracts, Proceedings of the 14th International Zeolite Conference,
Cape Town, South Africa; The Catalysis Society of South Africa: 2004;
p 109.
(27) Sastre, G.; Cantin, A.; Diaz-Cabanas, M. J.; Corma, A. Chem. Mater. 2005,
17 (3), 545-552.
(28) Leiva, S.; Gimenez, I.; Corma, A.; Rey, F.; Sabater, M.; Sastre, J. Book of
Abstracts, Proceedings of the 14th International Zeolite Conference, Cape
Town, South Africa; The Catalysis Society of South Africa: 2004; p 265.
(29) Lobo, R. F.; Pan, M.; Chan, I.; Medrud, R. C.; Zones, S. I.; Crozier, P. A.;
Davis, M. E. J. Phys. Chem. 1994, 98 (46), 12040-12052.
(17) Piccione, P.; Yang S.; Navrotsky, A.; Davis, M. J. Phys. Chem. B 2002,
106 (14), 3629-3638.
(18) Henson, N. J.; Cheetham, A. K.; Gale, J. D Chem. Mater. 1998, 10, 3966.
(19) Ciric, J. U.S. Patent 3,950,496, 1976.
(20) Shantz, D. F.; Burton, A.; Lobo, R. F. Microporous Mesoporous Mater.
1999, 31, 61-73.
(21) Paillaud, J.-L.; Caullet, P.; Schreyeck, L.; Marler, B. Microporous
Mesoporous Mater. 2001, 42, 177-189.
9
7628 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Synthesis of MTT-Type Zeolites
A R T I C L E S
33 family as a random intergrowth of two polytypes whose
relationship is similar to that of the pure polytypes A and B of
zeolite beta. Lobo also described a third hypothetical polytype
C that possesses double four-ring units in its framework. From
their previous work with germanosilicate gels, Corma et al. knew
that the presence of germanium favors formation of double four-
ring units. A previous publication showed that germanium
promoted formation of beta type C (a beta polytype with a high
density of double four rings) in systems that otherwise formed
conventional zeolite beta when germanium is absent.30 Since
the pore dimensions and distances between channel intersections
are similar in polytypes A, B, and C (and therefore the organic
should have similar interactions with the framework), Rey et
al. speculated that inclusion of germanium should promote the
synthesis of the polytype C in this system. Their experiments
with germanium indeed led to the synthesis of ITQ-24,31 which
possesses the framework of the previously unobserved
polytype C.32
In the present study we examine the design of SDA molecules
for the syntheses of zeolites with the MTT framework topology.
This framework is shared by zeolites ZSM-23, KZ-1, EU-13,
ISI-6, and SSZ-32. ZSM-23 was first discovered in 1976 by
researchers at Mobil using pyrrolidine as an SDA.33 Later Moini
reported that ZSM-23 could be prepared using C7-, C8-, C11-,
and C12-trimethylammonium (TMA) diquats.34 In this
context the number after C signifies the carbon length of the
methylene chain between the trimethylammonium centers.
In 1982 Parker and Bibby reported the crystallization of
KZ-1 from aluminosilicate gels (Si/Al > 55) in the presence of
pyrrolidine, isopropylamine, or dimethylamine.35 EU-13 was
reported to crystallize from gels containing tetramethylammo-
nium and rubidium.36 SSZ-32 was synthesized by workers at
Chevron using the N,N′-diisopropylimidazolium (DIPI) and
the N-isopropyl-N′-methylimidazolium cations as SDAs.37
Nakagawa subsequently discovered that SSZ-32 could be
prepared using isobutylamine, trimethylamine, and diisobuty-
lamine.38 SSZ-32 can be made with lower Si/Al ratios than the
other MTT phases. What features do most of these molecules
have in common that lend themselves to the synthesis of MTT-
type zeolites?
Figure 2. (a) Comparison of MTT and TON frameworks showing large
troughs shaded (green in MTT, red and green in TON). (b) Model of the
one-dimensional 12-ring pore framework for ZSM-12 (MTW).
commercially available) and examined their structure-directing
effects in zeolite syntheses.
This work is a sequel to a study we recently published on
the use of diquaternary SDAs based upon imidazolium deriva-
tives.39 In that work molecular modeling was used to examine
the structure-directing effects of imidazolium monomer deriva-
tives as well as their diquaternary analogues in which two
imidazole rings were connected by methylene chains of varying
carbon lengths. A key point in that investigation was that the
isopropyl groups of the DIPI molecule are separated by a
distance of 5 Å, equivalent to the periodicity along the pore
axis of the MTT framework. The isopropyl groups were
determined to reside above the corrugations, or troughs, centered
within the wider ends of the teardrop-shaped pores (Figure 2).
This optimized configuration is verified by a single-crystal XRD
study of an as-made sample prepared in fluoride-containing
gels.40 The space above these corrugations allows optimal van
der Waals contacts with isopropyl-type groups. Further support
for this explanation is that the trimethylammonium groups in
the C7 and C11 diquats investigated by Moini41 are also
separated by 10 and 15 Å, distances that are multiples of the
periodicity between consecutive troughs. This led us to consider
diquaternary molecules in which N-isopropylimidazolium groups
were connected by methylene chain lengths of five and nine C
atoms to give distances between the isopropyl groups of 15 and
20 Å, respectively. However, to this author’s disappointment,
none of the various conditions of synthesis yielded an MTT-
type phase. Both diquaternary molecules yielded ZSM-5 in
aluminosilicate compositions and Nu-86 in all-silica composi-
tions in concentrated fluoride media. From these results we
concluded that although these SDA molecules have good
energetic fits within MTT pore structure, kinetic rather than
thermodynamic effects may be the overriding factor in the phase
We first address this question by performing a series of
molecular modeling calculations to examine the optimal packing
arrangements of some of the aforementioned molecules in the
MTT framework. From these energy minimization calculations
and from crystallographic considerations of the zeolite frame-
work host we are able to understand the determinant features
of the SDA molecules that specify MTT-type phases. On the
basis of these considerations, we can predict a large family of
previously unexamined molecules of varying carbon lengths and
shapes that should also yield MTT-type phases. For this study
we synthesized many of these novel molecules (others are
(30) Corma, A.; Navarro, M. T.; Rey, F.; Valencia, S. Chem. Commun. 2001,
1720-21.
(31) Rey, F.; Corma, A. Personal correspondence.
(32) Castaneda, R.; Corma, A.; Fornes, V.; Rey, F.; Rius, J. J. Am. Chem. Soc.
2003, 125 (26), 7820-1.
(33) Plank, C. J.; Rosinski, E. J.; Rubin, M. K. U.S. Patent 4,076,842, 1976.
(34) Valyocsik, E. W. U.S. Patent 4,490,342, 1984.
(35) Parker, L. M.; Bibby, D. M. Zeolites 1983, 3, 8.
(36) Araya, A.; Lowe, B. M. U.S. Patent 4,705,674, 1987.
(37) Zones, S. I. U.S. Patent 5,053,373, 1991.
(39) Zones, S. I.; Burton, A. W. J. Mater. Chem. 2005, 15, 4215-4223.
(40) Zones, S. I.; Darton, R. J.; Morris, R.; Hwang, S.-J. J. Phys. Chem. B 2005,
109 (1), 652-661.
(41) Moini, A.; Schmitt, K. D.; Valyocsik, E. W.; Polomski, R. F. Zeolites 1994,
14, 564.
(38) Nakagawa, Y. U.S. Patent 5,707,601, 1998.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7629

A R T I C L E S
Burton
Table 1. Energy Minimization Calculations for SDA Host
Complexes Involving Small Molecules
Figure 3. Comparison of two conformations of 1,9-bis(isopropylimida-
zolium)nonane. The top conformation possesses slightly lower energy than
the bottom one.
selectivity. Another consideration is that both sets of imidazole
rings and isopropyl groups must be perfectly aligned to allow
an optimal fit within the one-dimensional MTT pore. There
are several conformations of the diquaternary molecules that
have nearly the same or even less internal energy than the
straight chain conformation required for occlusion of the
molecule within the pore space (see Figure 3). At typical
synthesis conditions it is probably entropically difficult for the
molecule to have the necessary conformation for occlusion
within the MTT structure. This suggested that molecules without
the rigid rings may be more suitable candidates.
a Two unit cell along pore direction.
Molecular Modeling. Molecular modeling simulations were per-
formed with the Cerius242 software. The reported stabilization energies
are differences in energy of the optimized free molecule and the
molecule occluded within the zeolite framework. Contributions from
van der Waals interactions, valence bond, angle, and torsion energies
were determined with the combination of the Burchart and Universal
force fields.43 The positions of framework atoms were fixed during
the energy minimizations, and Coulombic interactions between the
cation and the zeolite framework were neglected. At least 15 different
configurations of the SDA were sampled for each framework by (1)
placing the molecule in random locations within the void space of the
zeolite and (2) changing the initial torsional angles of appropriate
functional groups or chains within the SDA molecule. Since in a few
cases the length of the SDA is greater than or nearly equal to the length
of the unit cell dimension along the pore axis, a supercell was created
(as necessary) from multiple unit cells along the one-dimensional pore
axis in order to allow adequate space for each molecule. The energy
reported for each SDA/framework pair is the minimum found among
the configurations sampled.
II. Experimental Section
Zeolite Syntheses. Unless otherwise stated with the text, zeolite
syntheses using neutral amines were performed with gels of the
following composition: 0.10 K2O:0.20 SDA:1.0 SiO2:0.008, 0.015,
0.033 Al2O3:42 H2O. Cabosil M-5 was used as the silica source, and
Reheis F-2000 aluminum hydroxide was used as the aluminum source.
In a few examples zeolite Y (LZY-52) was instead used as the
aluminum source. For the borosilicate syntheses with neutral amines,
the typical gel composition was 0.05 K2O:0.40 SDA:1.0 SiO2:0.017
B2O3:42 H2O. Potassium tetraborate decahydrate was the boron source
in these syntheses.
For the quaternary ammonium molecules most syntheses were
performed with the SDA in the halide form. The all-silica syntheses
typically were performed with OH/SiO2 ratios between 0.15 and 0.20.
The aluminous gels required higher hydroxide concentrations
(0.25-0.30). In one instance (for Molecule O), the high alkali
concentration in the aluminosilicate gels resulted in a layered phase.
For this example the SDA was ion-exchanged into the hydroxide
form by dissolving it in water and adding it to a 5-fold excess of
Biorad AG 1-X8 exchange resin. After 24 h, the exchange resin was
removed by filtration, and the hydroxide concentration of the SDA
solution was determined by titration with a standard solution of 0.1 N
HCl. The gel components for the zeolite syntheses were mixed
together in 23 mL Teflon liners that were then sealed and placed inside
steel Parr autoclaves. These were then inserted into a rotating spit (43
rpm) inside an oven at a fixed temperature (150-170 °C) for 5-30
days. Most syntheses were performed at 160 °C. In general, the all-
silica and borosilicate syntheses required 5-8 days of synthesis
time, the syntheses with SAR (silica to alumina ratio) ) 66 required
about 7-10 days, and the syntheses with SAR ) 33 required 14-17
days. The syntheses were monitored by periodically performing XRD
analyses on small amounts of gel removed from the reactors. The
powder XRD measurements were performed with a Siemens D-500
instrument. For each SDA that produced an MTT phase, at least two
syntheses were performed to check for reproducibility of its structure-
directing effects.
III. Results
Modeling/Synthesis Results with Small Molecules. The first
part of this study examines small molecules that fit within a
single unit cell along the pore axis. Table 1 shows the results
of the energy minimization calculations for these molecules in
three different one-dimensional zeolites that are commonly
observed phases in syntheses with molecules of similar dimen-
sions: MTT, TON, and MTW. MTT and TON are zeolite
frameworks with one-dimensional medium pores. The MTW
framework has one-dimensional large pores. Note that the entries
in the left half of the table are the absolute stabilization energies
of each molecule, while the entries in the right half normalize
those values according to the number of tetrahedral (T) atoms
in the unit cell or supercell. The latter numbers are the better
basis of comparison since they provide relative quantitative
evaluations of how the silica is stabilized for each of the
frameworks.
The first entry for trimethylamine shows a very favorable
stabilization of -6.9 kJ/mol T atom in the MTT framework.
(42) Cerius2 Modeling EnVironment, Release 4.8; Accelrys Software Inc.: San
Diego, 2005.
(43) de Vos Burchart, E. Ph.D. Thesis, Technical University of Delft, The
Netherlands, 1992; Table I, Chapter XII.
9
7630 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Synthesis of MTT-Type Zeolites
A R T I C L E S
Table 2. Phase Selectivity For Syntheses with Small Aminesa
Figure 4. Cross-sectional views of the energy-optimized positions of (a)
DIPI, (b) trimethylamine, (c) dimethylamine, and (d) pyrrolidine.
Figure 4a,b compares cross-sectional views of the positions of
the DIPI molecule and the trimethylamine molecule within the
MTT pore. The size and shape of both molecules are nicely
reflected in the dimensions and shape of the MTT pore.
Pyrrolidine is another molecule that possesses a very favorable
energetic fit within the MTT pore (Figure 4d). Removal of a
methyl group (Figure 4c, dimethylamine) reduces the stabiliza-
tion in MTT to -5.6 kJ/mol T atom, a still very favorable
energy. However, note that the TON framework has a nearly
equivalent energy of -5.3 kJ/mol T atom. The researchers at
Mobil also showed a nearly equivalent stabilization in MTT
and TON for the ethylamine molecule, which has a backbone
structure identical to dimethylamine.44 We need to keep this in
mind as we consider the synthesis results for this molecule. The
lesson here is that each methyl moiety in the trimethylamine
molecule provides significant stabilization in the MTT pore.
However, when tert-butylamine is considered, the larger pore
dimensions of MTW are calculated to give a better fit. In this
case, steric repulsions between neighboring molecules slightly
reduce the stabilization within the MTT pore. In its energy-
optimized position a slight adjustment in the position of the
molecule or exchanging the position of the amino group for a
methyl group results in substantial decreases in the stabilization.
An interesting difference is observed with the tetramethylam-
monium molecule. Although this molecule has the same
backbone structure as tert-butylamine, its stabilization is
calculated to be less than one-half that of tert-butylamine. The
space required by the extra proton further increases the steric
repulsions between adjacent molecules located within the same
pore.
When the backbone of the molecule is again extended to give
neopentylamine, the molecule is now too large to fit within one
unit cell of both the MTT and TON frameworks without
suffering severe steric repulsions from neighboring molecules.
In this case a supercell two unit cells in length must be chosen
for those frameworks, and only one-half of the large troughs in
the MTT framework are stabilized by van der Waals contacts
with either an isopropyl- or a tert-butyl-type group. Here the
12-ring pore of MTW is calculated to give a more favorable
fit.
Table 2 gives the results of our syntheses with some of these
small molecules at SAR gel ratios of 33 and 66. Note that the
first two entries, trimethylamine and isopropylamine, have
identical backbone structures and that they each yield the same
results. At SAR ratios of 33 and 66, the experimental results
are MFI and MTT, respectively. Although trimethylamine has
a very good energetic fit within the MTT pore structure (greater
than more selective molecules as we will discuss later), it is
only selective for MTT at the higher SAR ratio. Nonetheless,
the calculations successfully indicate that MTT should be a
preferred phase using these molecules as SDAs. It is worth
a SAR ) SiO2/Al2O3. b Intergrowth of MTT and TON.
pointing out that although trimethylamine has been previously
reported to promote the synthesis of MTT, the person who
performed these calculations was then unaware of that result
and therefore was able to provide an unbiased assessment of
the potential structure-directing ability of that molecule.
In the experiments of the current work dimethylamine yielded
only amorphous products at the low SAR, and in most cases
an amorphous product was also obtained at the higher SAR ratio.
However, in one trial SSZ-54 an MTT/TON intergrowth (vide
infra) was observed. It is interesting to note that the calculated
stabilizations of dimethylamine are similar in the MTT and
TON frameworks. Parker and Bibby prepared an MTT phase
using dimethylamine in gels where the SAR ratio was about
110. We are able to reproduce their synthesis using Al2(SO4)3
as the aluminum source, although our experiments require 7-9
days rather than the previously reported 40 h. While dimethy-
lamine does have a favorable fit in the MTT pore structure, its
weaker selectivity for MTT (over a wide range of SAR values)
may be a reflection of the relative stabilizations of trimethy-
lamine and dimethylamine.
tert-Butylamine was slow to crystallize any products at SAR
) 33. Only a minor amount of MTT crystallized after 23 days,
and the rest of the product was mostly cristobalite. Seeding
changed neither the crystallization rate nor the phase selectivity.
At SAR ) 66, use of tert-butylamine yielded comparable
quantities of MTT and cristobalite as estimated by powder
XRD. These results are interesting when interpreted within the
context of the energy optimizations performed for this molecule.
Recall that although the molecule has a favorable fit within the
pore, the calculations indicate the molecule has little freedom
of rotation or orientation without encountering steric repulsions
from its nearest neighbors. The difficulty in crystallizing MTT
may be a reflection of these observations. Another point worth
consideration is that MTW is not formed with this molecule
although it is calculated to have a better fit. In our group’s work
(44) Rollmann, L. D.; Schlenker, J. L.; Lawton, S. L.; Kennedy, C. L.; Kennedy,
G. J.; Doren, D. J. J. Phys. Chem. B 1999, 103 (34), 7175-7183.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7631

A R T I C L E S
Burton
MTW phases are rarely crystallized using neutral amines as
SDA although quaternary ammonium compounds with similar
size frequently yield MTW phases. It should also be recognized
that this molecule is small enough to pass freely through the
MTW pores with little energetic barrier. This could adversely
affect the ability of the molecule to either nucleate this phase
or prevent its dissolution.
Neopentylamine and N-methyl-tert-butylamine each gave only
amorphous or layered products in our syntheses. These mol-
ecules are unable to provide significant stabilization to the MTT
framework since only one-half of the troughs can be “occupied”
by an isopropyl- or tert-butyl-type group. Again, these molecules
have favorable stabilization energies within the MTW frame-
work, but they also have easy egress from the pores.
The last two molecules in Table 2, 1-methylbutylamine and
3-methylbutylamine, have equivalent backbone structures de-
fined by the connectivity of the carbon and nitrogen atoms.
However, while 1-methylbutylamine yields TON in both SAR
conditions, 3-methylbutylamine yields no crystalline products.
Here the idea fails that molecules that possess equivalent
backbone structures should yield similar phases in their syn-
theses. Several molecules have been reported to crystallize TON
zeolites.45,46 Many of them either are completely linear mol-
ecules or possess a single methyl branch. Outside these
observations we have been unable to find any unifying
characteristics of molecules that promote crystallization of TON
zeolites. As discussed in the Introduction, it is easier to “design”
an SDA for a particular zeolite phase that has salient features
such as a cage or periodically spaced channel intersections. The
TON framework has no such prominent features. The following
sections will discuss several molecules that have been chosen
by design as SDAs for MTT. These considerations take
advantage of large troughs present in the MTT pores and the
periodicity between them.
Synthesis of Amines for This Study. While some of the
molecules examined as SDAs in this study are commercially
available chemicals, many of the amines had to be synthesized
in the laboratory. Two routes were used to prepare the novel
amines. One method simply involves a one-step alkylation of
an amine with an alkyl halide. Figure 5a shows an example of
the synthesis of N-isopropyl-N-isoamylamine by this route.
The syntheses of other amines involved bulky functional
groups that react slowly in alkylation reactions. Also, the
syntheses of some of the diamines in these studies require that
each nitrogen atom in the parent diamine be alkylated only once.
In typical reactions with alkyl halides, it is possible that the
nitrogen atom may be alkylated multiple times if the product
remains in solution. To avoid these problems, the parent amine
or diamine was reacted with an appropriate acid chloride. The
amide products were then reduced with LiAlH4 to obtain the
desired product. Figure 5b shows the synthesis scheme and 1H
NMR of N,N′-diisobutyl-1,3-propanediamine obtained by this
technique. The Supporting Information describes the synthesis
of each novel molecule.
Figure 5. (a) Synthesis scheme and 1H NMR of N-isopropyl, N-
isoamylamine. (b) Synthesis and H NMR of diamine in d-MeOH.
1
to reside above the troughs within the MTT pores. This suggests
that molecules that have terminal isopropyl groups separated
by 5 Å, equivalent to the distance between the centers of these
troughs, may be ideal candidates for the synthesis of MTT
zeolites. Another requirement is that the terminal groups must
be connected by a group that is narrow enough to fit within the
constrictions of the MTT pore. For the DIPI molecule, the
imidazole ring is an ideal fit within these pore constrictions.
Functional groups with larger cross sections will not fit. We
will examine molecules where the branched groups are con-
nected by either linear alkyl chains or linear chains in which
an amino group is substituted for a methylene unit. As discussed
in the previous study on imidazolium diquaternary molecules,39
the distance between every other carbon atom in a straight alkyl
chain is about 2.5 Å. This distance changes very little if a
nitrogen atom is substituted for one of the methylene carbon
atoms. Therefore, if there are three “backbone” atoms connecting
two isopropyl centers, then the distance between the centers
will be 5 Å.
The first four molecules in Table 3 meet the description given
above. Note that the first two molecules are selective for MTT
at both SAR ratios. However, molecules C and D both give
MFI at SAR ) 33 and MTT at SAR ) 66. Surprisingly, when
MTT seeds are used with molecule C (0.02 g of seeds per 0.90
g of Cabosil) for SAR ) 33, the phase selectivity is altered to
MTT. Moreover, when zeolite Y is used as the aluminum source
in gels with SAR ) 33 and without seeds, MTT is again the
preferred phase for molecule D.
Synthesis Results for Molecules with 5 Å Repeat Between
Branched Groups. The results for the molecular modeling of
the DIPI molecule39 indicate that isopropyl-type groups prefer
These are interesting results when considered in the context
of the relative nucleation rates of these phases in gels with high
aluminum concentrations. Zeolite phases with lower framework
(45) Szostak, R. Handbook of Molecular SieVes; Van Nostrand Reinhold: New
York, 1992.
(46) Gies, H.; Marler, B. Zeolites 1992, 12, 42.
9
7632 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Synthesis of MTT-Type Zeolites
A R T I C L E S
Table 3. Phase Selectivity Observed for Molecules with Branch
Table 4. Phase Selectivity for Quaternary Amines with Branched
Groups Separated by 5 Å
Groups Separated by 5 Åa
conditions. F-J yield different results. They all yield MFI at
SAR ) 33. N-Isoamyl-N-ethylamine (I) gives MFI for the
higher SAR. However, it is necessary to point out that
N-isobutyl-N-propylamine (H) and N-isopropyl-N-butylamine
(G) both give a mixture of MFI and MTT at SAR ) 66. We
next repeated syntheses at SAR ) 120 to determine if the phase
selectivity for this group of molecules could be shifted toward
MTT. For this SAR, H and I both gave a pure MTT phase, G
gave an MTT/TON intergrowth, and J gave TON. Within this
collection of molecules it is interesting to note that the lower
stabilization of the SDA (through removal of a single methyl
group) is reflected in the weaker selectivity for MTT. It is also
noteworthy that the two diamines in this group both show good
selectivity for MTT despite removal of the methyl group.
Although there is no current explanation for their better
selectivity, it is tempting to believe it is related to the presence
of the two amino groups.
Molecule K yields SSZ-54,47 an MTT/TON intergrowth. The
synthesis and structure of this material will be discussed in a
future publication.
a SAR ) SiO2/Al2O3.
The last two molecules (L, M) in Table 3 are similar to the
first group except that an extra methyl group has been added to
one of the central atoms of the terminal branched group to create
a “tert-butyl-type” functional group. The synthesis results for
these molecules are very intriguing. We performed energy
minimizations for N,N,N,N′,N′-pentamethyl-1,3-propanediamine
(Molecule O in Table 4) within the MTT framework. The
asymmetry of the molecule allows two possible local configura-
tions of the molecule within the MTT framework: one in which
the “tert-butyl” ends on adjacent molecules directly face each
other, and one in which the “tert-butyl” groups face away. The
second configuration gives a favorable stabilization of -5.7 kJ/
mol T atom, while the first gives only -1.7 kJ/mol T atom.
Molecules L and M possess the same backbone structure as
this molecule, yet the syntheses with these molecules yielded
only amorphous or dense phase products (in one experiment
minor MTT/MFI was detected by XRD). These results are
different from their quaternary analogue (Table 4).
densities are often favored in gels with higher aluminum content.
ZSM-5 can be synthesized in the presence of a large number
of amines, although many of them do not have an exceptional
energetic fit within the MFI framework. Good examples are
the data for isopropylamine and trimethylamine in Table 2.
These results are consistent with ZSM-5 being a more kinetically
favored phase when aluminum is present in high concentration.
By using seeds in the case above we are able to circumvent
this problem by providing nucleation centers from the outset.
Also, by using zeolite Y as an aluminum source we limit the
concentration of reactive aluminum immediately available to
the rest of the gel. However, it is important to point out that
use of seeds or zeolite Y as an aluminum source does not change
the phase selectivity for all of the SDA molecules that produce
MTT only at the low aluminum conditions.
The second group of molecules (E-J) in Table 3 is similar
to the first set with the exception that a single methyl group
has been removed from one of the terminal isopropyl-type
moieties. Removal of a methyl group effectively reduces the
stabilization afforded by an isopropyl group above the pore
corrugations in the MTT pores. We therefore might expect this
group of molecules (collectively) to show less selectivity for
MTT. E is surprisingly selective for MTT at both SAR
Molecule O is selective for MTT at all synthesis conditions
examined including the all-silica synthesis and the borosilicate
synthesis (SBR ) 66). This high selectivity is consistent with
the calculated stabilization mentioned above. The lack of
(47) Zones, S. I.; Burton, A. W. U.S. Patent 6,676,923, 2004.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7633

A R T I C L E S
Burton
Table 5. Phase Selectivity for Molecules with Branched Moieties
Separated by Multiples of 5 Å
Table 6. Other Interesting Candidate Molecules for MTT-type
Zeolites
Synthesis Results For Oligomeric or Polymeric Molecules
with Multiples of 5 Å Between Branched Groups. Table 6
shows molecules that are oligomeric in the sense that there are
multiple pairs of branched dimethyl groups separated by 5 Å.
The phase selectivities of molecules AA and BB mirror those
of P and O, respectively. Both BB and O are selective for MTT
in every condition we examined, but, once again, if a terminal
methyl group is removed (AA), the SDA crystallizes MTT only
in the all-silica composition and MFI in the aluminosilicate and
borosilicate compositions. The triquaternary SDA CC is similar
to the C7-TMA-diquat (X) except that a dimethylammonium
center has replaced the middle carbon atom in the methylene
chain, thereby creating a molecule with three charged nitrogen
centers separated by 5 Å. This molecule has been previously
reported by Mobil as an SDA for ZSM-23.48 Molecule CC
promotes crystallization of MTT phases over a wide range of
Si/Al ratios, but interestingly its phase selectivity switches to
MTW (a one-dimensional 12-ring zeolite) in the borosilicate
gel. Also, unlike the diquat, this molecule gives ZSM-48 in all-
silica syntheses. The charge density of this molecule would
demand a high concentration of silanol defects in an all-silica
zeolite. These defects destabilize zeolite frameworks relative
to their fully intact structures.17,49 It is possible that the
framework of ZSM-48, which is frequently observed in all-
silica compositions with a variety of linear molecules, is better
able to accommodate the required defect density because it
possesses an inherently more stable framework than MTT.
The last two molecules are polycationic polymers with (1)
repeat distances of 5 Å and (2) alternating repeat distances of
5 and 10 Å between the quaternary centers. The polymers are
a significant test of the premise we established because, like
the zeolite structure itself, these SDA molecules possess
periodically spaced units. Polymer EE is selective for MTT over
selectivity observed in its isostructural compounds (L and M)
is not consistent with our predictions. This may be due to the
weak polarity of these molecules since the nitrogen atoms are
unquaternized, the C/N ratio (9) is high, and the nitrogen atoms
are close to tert-butyl groups. If we now consider a compound
in which a single methyl group is removed from one of the
branched groups (P), we should expect a weaker selectivity for
MTT. In fact, this compound yields MTT only in the all-silica
case. Note that addition of another methylene carbon to the
backbone structure (R) creates a compound that does not give
MTT under any synthesis conditions.
Synthesis Results for Molecules with Multiples of 5 Å
Between Branched Groups. Table 5 shows diamines with
isopropyl or tert-butyl groups separated by 10 or 15 Å. The
first two molecules differ only in the positions of the nitrogen
atoms within the connecting chains. In both cases, MFI and
MTT are the preferred phases at SAR ) 33 and 66, respectively.
However, molecule N unexpectedly gives an MTT/TON
intergrowth at SAR ) 66. Intuitively, we expect these molecules
to show weaker selectivity than the first group of molecules in
Table 3 since only every other trough in the MTT pore may be
occupied by a branched moiety. Molecules V and W are
analogous to the first pair except that tert-butyl groups are
substituted for the isopropyl groups at the ends of each molecule.
Although V has the same phase selectivity as T and U, W does
not promote crystallization of MTT under any conditions with
or without seeds. However, we observed that this molecule is
not soluble in water, and after the synthesis mixture was
removed from the oven the organic amine appeared as a separate
layer above the rest of the gel. The organic layer was verified
by NMR to be the original SDA. Its insolubility may explain
its lack of structure-directing ability. The last diamine (Y), with
terminal isopropyl groups separated by 15 Å, also yields MTT
at SAR ) 66.
(48) Moini, A. U.S. Patent 5,332,566, 1994.
(49) Burton, A.; Zones, S.; Elomari, S. Curr. Opin. Colloid Interface Sci. 2005,
10, 211.
9
7634 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Synthesis of MTT-Type Zeolites
A R T I C L E S
a wide range of compositions except at SAR )33. However,
DD (like CC) does not yield MTT in the all-silica compositions,
and a higher concentration of boron (SBR ) 33) is required to
give an MTT phase without significant cocrystallization of
dense phases. Nonetheless, the phase selectivities of the larger
molecules in this section are consistent with the empirical rules
we established for the design of MTT-specific SDA (although
the compositional requisites of the gels may vary for individual
SDA molecules).
with SAR of 33 and 66, respectively. In one case nearly all of
the extraframework charge is accounted for by potassium, and
in the other only one-half of the charge is accounted for by
potassium. By examining the data for the other SDAs it can be
seen that a 1:1 ratio of K/Al is the exception rather than the
rule. How do we account for the remaining charge? Previous
NMR studies by Hwang and Zones indicated that in SSZ-25
samples prepared in the presence of both N,N,N-trimethyl-2-
adamantammonium and isobutylamine isobutylamine was oc-
cluded with the framework as a protonated molecule. The Mobil
study also strongly suggests a large fraction of occluded amines
are protonated.
Borosilicate Syntheses of MTT Phases. One of the surpris-
ing findings in our initial studies was that MTT phases were
rarely observed to crystallize from borosilicate gels in syntheses
with unquaternized amines. The neutral amines gave either dense
or layered phases with KOH/SiO2 ) 0.2. Dense or layered
phases often crystallize when the hydroxide or alkali concentra-
tion is too high. A series of experiments were performed in
which the KOH/SiO2 ratio was incrementally decreased and the
SDA/SiO2 ratio was incrementally increased.
When the OH/SiO2 ratio was decreased to 0.10 with SDA/
SiO2 ) 0.4, pure MTT phases crystallized with several of the
diamine molecules. Compound Y crystallized a borosilicate
MTT phase (gel Si/B ) 33) after 16 days of heating at 160 °C.
A, D, E, and F usually crystallized an MTT phase within 7
days. E was particularly successful in crystallizing MTT phases.
This molecule produced borosilicate MTT phases after 6-7
days of heating. With seeds the crystallization time could be
reduced to 3 days with this SDA. Molecule A crystallized MTT
in 4 days with the use of seeds. However, the monoamine
molecules (including isopropylamine, isobutylamine, or diisobu-
tylamine) were never able to crystallize an MTT phase.
Next a series of experiments was performed in which the
Si/B ratio was systematically varied in gels with molecule E.
Pure MTT phases crystallized from gels with Si/B ratios as
low as 5. However, the concentrations of boron in the products
were significantly lower, an indication of the solubility of boron
in contrast to aluminum. For gels with Si/B ratios of 32, 16,
and 5, the respective product ratios were 49, 34, and 23. It should
be noted that the boron content was increased by adding
potassium tetraborate and that no decreases were made in the
amount of KOH added to the gel. The effective K2O/SiO2 gel
ratio therefore also increased when the boron concentration was
increased.
Molecule Z was not an effective SDA for any zeolite phase.
As in Molecules L and M, the nitrogen atoms in this molecule
are surrounded by bulky groups that may increase the hydro-
phobic character of the molecule. However, in a few of the
preparations a minor MTT phase was detected.
Characterization of MTT Phases and Pore-Filling Effects.
Supporting Information Figure 1 shows characteristic powder
XRD patterns for the MTT phases prepared with each of the
SDA molecules in the study. In a few products cristobalite is
present (peak at 22° 2θ), but the samples are otherwise free of
detectable impurities. Supporting Information Table 1 shows
the Si/Al ratio, K/Al ratio, CHN content, and pore filling for
the MTT zeolites crystallized from selected SDA compared with
theoretical values based upon optimal packing arrangements of
that molecule. The C/N ratios of the as-made zeolites are close
to those expected for their respective SDAs. Note that, in
general, for syntheses involving unquaternized amines the
preparations with high Si/Al ratios give products where the SDA
fills 90-100% of the expected space available within the
micropores. The maximum capacities are based upon the
packing arrangement that gives the best fit of each molecule
within the framework. In contrast, the preparations with higher
aluminum contents yield products with only 80-90% of the
maximum capacity. How do we explain this difference?
We expect the more aluminous zeolites to have higher
concentrations of extraframework potassium for charge com-
pensation. Since there are no isolated cages in the MTT
framework within which potassium cations may reside (except
perhaps within the six rings), many of these potassium cations
will occupy space within the pores that otherwise would be filled
by an SDA molecule. A slightly smaller organic content is
therefore anticipated for zeolites with more extraframework
alkali cations. The only samples that significantly fall short of
their maximum pore-filling capacities are those prepared with
isopropylamine (50%) and trimethylamine (34%). (It should be
noted that the MTT phase produced with trimethylamine
contained cristobalite impurity.) Given the excellent fits of these
two molecules within the MTT pore space, it is difficult to
rationalize why no more than one-half of the internal troughs
are occupied by the SDA. If the amines are occluded primarily
in their protonated forms, then a higher concentration of SDA
molecules requires more framework charge than can be provided
by the available aluminum. This supposition is supported by a
previous Mobil study of small amines in zeolite synthesis.44
Although the stabilization provided by van der Waals contacts
is reduced by the lower SDA loading, the defects required for
charge compensation (of a higher concentration of protonated
SDA) destabilize the framework to a greater extent.
The lower alkali hydroxide content necessary for successful
preparations of borosilicates (with neutral amines) was an initial
surprise. In our group’s experience with quaternary ammonium
molecules as SDA, borosilicate zeolites are often prepared with
alkali hydroxide contents that are similar to those we use for
aluminosilicates. However, examination of the work by Gies
and co-workers50-54 and the group at EniTecnologie55 reveals
that many of their borosilicate syntheses are carried out with
high concentrations of amines and with little or no alkali
(50) Marler, B.; Gies, H. Zeolites 1995, 15, 517.
(51) Vortmann, S.; Marler, B.; Gies, H.; Daniels, P. Microporous Mater. 1995,
4, 111.
(52) Vortmann, S.; Marler, B.; Daniels, P.; Dierdorf, I.; Gies, H. Stud. Surf.
Sci. Catal. 1995, 98, 262.
Another interesting result, however, is the variance of K/Al
ratio observed among the different samples. For example, the
K/Al ratios for diisobutylamine are 0.94 and 0.48 in the preps
(53) Grunewald-Luke, A.; Gies, H. Microporous Mater. 1994, 3, 159.
(54) Gies, H.; Gunawardane, R. P. Zeolites 1987, 7, 442.
(55) Millini, R.; Perego, G.; Bellussi, G. Top. Catal. 1999, 9, 13-34.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7635

A R T I C L E S
Burton
Table 8. Measured C/N Ratio of Zeolite Products for SDA That
Table 7. Synthesis Results for Amines or Quaternary Ammonium
Compounds That Do Not Possess Multiples of 5 Å Between the
Isopropyl, Dimethylamino, or Trimethylammonium Groups
Unexpectedly Yield an MTT Phase
SDA. Therefore, degradation products (such as trimethylamine)
are not responsible for crystallization of the MTT phase.
Although these diquaternary SDA are exceptions to the
empirical rules, it is interesting to compare their phase selectivi-
ties with those for the C7 and C11 analogues. The latter
molecules possess distances of 10 and 15 Å, respectively,
between the trimethylammonium centers. In contrast to the C8
compound, the C7 diquat is able to direct the crystallization of
MTT both for the all-silica condition and for SAR ) 33.
Likewise, the C11 diquat is able to crystallize an MTT phase
at SAR ) 33 with only a minor amount of layered impurity
phase. The C7 and C11 diquats therefore show greater selectivity
for MTT than their C8 and C12 counterparts. The interpretation
of this author is that because the C7 and C11 provide better
fits within the MTT pore structure, they are able to direct the
crystallization of MTT phases over a wider range of conditions.
In examining our research group’s database of synthesis
results other molecules were found that unexpectedly gave MTT
phases. Table 8 shows examples of such putative SDA
molecules. In the many synthesis trials performed with each
molecule, only a single experiment yielded an MTT phase, and
the product was observed after at least 25 days of synthesis. In
addition, the product included an appreciable quantity of
cristobalite or quartz. Note that these molecules possess trim-
ethylammonium groups. SDA molecules with trimethylammo-
nium groups can degrade through a Hofmann elimination to
form trimethylamine (which is capable of directing the crystal-
lization of MTT). Table 8 shows the C:N ratio of the as-made
MTT product for each of these molecules. Note that the C:N
ratio of the as-made zeolites is very different from the parent
SDA molecules. This strongly suggests that degradation products
are responsible for crystallization of MTT. C:N data are not
available for the fourth molecule, but it should be noted that
many of the synthesis trials with this SDA yielded the dense
clathrate ZSM-39 (MTN). MTN phases are frequently observed
in zeolite syntheses when SDA molecules with trimethylam-
monium groups degrade.
hydroxide. A low alkali hydroxide content may be necessary
to allow a sufficient concentration of protonated amines to exist
in the borosilicate gel mixture. The studies by Zones and Hwang
and by the Mobil group suggest that a significant portion of
occluded amines is in a protonated form. These observations
and the fact that quaternary ammonium molecules successfully
crystallize zeolites at high hydroxide conditions indicate that
the presence of the cationic SDA may be necessary for
nucleation of the zeolite. It was noted earlier that many of the
diamines in this study were successful in crystallizing borosili-
cate MTT while the monoamines were not. Perhaps this is
because the polyamines molecules are more likely to exist as
protonated species.
However, why do amines seem to work well with high
hydroxide conditions in aluminosilicate compositions? Alumi-
nosilicate gels are not as soluble as borosilicate gels in hydroxide
media. The aluminosilicate gels are therefore slower to convert
to dense phases. The extended time for crystallization may allow
a larger number of protonated amines to be “trapped” by silicate
species during the nucleation stages.
Syntheses with Molecules without Multiples of 5 Å
Between Branched Groups and Exceptions to the Rule. To
test the consistency of the empirical rules that have been
established in identifying SDA selective for MTT phases, it is
also important to examine similar SDA molecules with distances
between isopropyl, dimethylamino, or trimethylammonium
centers that are not multiples of 5 Å. The first four molecules
in Table 7 do not crystallize MTT phases under any of the
conditions we examined. However, contrary to our predictions,
the C8 and C12 trimethylammonium diquats succeed in crystal-
lizing MTT phases at SAR ) 66. These results are consistent
with the previous Mobil studies for SAR ) 90. In the all-silica
case, both SDA yield ZSM-48; and for SAR ) 33, the synthesis
with the C12 diquat results in a layered phase, and the C8 diquat
yields a mixture of cristobalite, MTT, MFI, and a layered phase
after more than 1 month of crystallization. CHN combustion
analyses of the as-made MTT products show that the C:N:H
ratios of the occluded organic are consistent with the parent
Interestingly, the C:N ratio for each of the phases does not
equal that for trimethylamine. Although thorough NMR inves-
tigations are required to verify this supposition, this author
believes that trimethylamine may start the nucleation of the
MTT phase and that other degradation products may become
9
7636 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Synthesis of MTT-Type Zeolites
A R T I C L E S
both one methylene unit longer than preferred distances of 10
and 15 Å, respectively. However, it is noteworthy that both the
C7 and C11 diquaternary analogues are selective for MTT over
a wider range of SAR conditions. This stronger selectivity is
probably a manifestation of the better fit these molecules provide
within the pore structure. A few molecules were found that
produce MTT phases although they are too large to fit within
the pore structure. These syntheses required at least 25 days,
and the CHN combustion analyses indicated the as-made
products possess C:N ratios that are very different from the
parent SDA. It is proposed that these SDAs degrade to form
trimethylamine or other SDAs that may nucleate MTT zeolite
phases.
Figure 6. Possible Hofmann elimination reaction for the first molecule
presented in Table 8.
incorporated within the zeolite during crystallization. The first
molecule, however, provides an interesting case where multiple
degradation products are capable of forming an MTT phase.
The measured C:N ratio is 8.0, which is the same as that
expected for the first decomposition product shown in Figure
6. This olefinic decomposition product may further react to form
an alcohol. Note that this compound has a backbone structure
identical to that for molecule P in Table 4. With the exception
of a degradation product formed by elimination of both
trimethylammonium groups there are no degradation products
that have C:N equal to or less than 8. We therefore believe this
molecule (or a structurally related alcohol) is the true SDA for
the observed MTT phase.
Finally, several of the polyamine molecules in this study can
be used to prepare borosilicate MTT samples. This is the first
report of a borosilicate MTT phase. The KOH/SiO2 ratio in
these gel preparations is at most one-half the hydroxide
concentration that we typically use in the crystallization of the
aluminosilicate samples (KOH/SiO2 ) 0.2) with the same SDA.
For the borosilicate preparations, dense or layered phases easily
crystallize in gels with higher hydroxide concentrations.
IV. Conclusions
Molecular modeling has been used to investigate the guest/
framework van der Waals interactions for small SDA molecules
that are known to be specific for MTT-type zeolite phases. On
the basis of these results and crystallographic considerations of
the MTT pore structure, several novel molecules have been
successfully predicted to be selective for MTT. These predic-
tions indicate that amines or quaternary ammonium molecules
with isopropyl, dimethylamino, tert-butyl, and/or trimethylam-
monium groups that are separated by multiples of 5 Å and
connected by methylene spacers should be able to crystallize
MTT phases. The predictions are highlighted by crystallization
of MTT phases using polymeric quaternary ammonium mol-
ecules in which dimethylammonium groups are separated by 5
and 10 Å. When a single methyl group is removed from
molecules with isopropyl or dimethylamino groups separated
by multiples of 5 Å, weaker selectivity is generally observed
for MTT (e.g., molecules O and P and BB and AA). This
reflects a reduced stabilization from the interaction between the
SDA and the MTT framework.
Acknowledgment. This work was supported by the Chevron
Energy Technology Company. I am grateful to G. L. Scheuer-
man and C. R. Wilson for their support of the new materials
research program at Chevron, and I thank S. I. Zones for useful
dialogue throughout the course of this work. I thank Greg Lee
and Yumi Nakagawa for providing some of the zeolite samples
discussed in Table 8 and I would like to acknowledge Steve
Trumbull for his instructions in the LAH reduction reactions. I
also thank Kelly Harvey for her patient assistance in preparing
this manuscript.
Supporting Information Available: Images of all the IZA-
coded frameworks discussed in this paper (Appendix 1); XRD
of MTT with different SDA (Figure 1); pore-filling calculations
for the SDA shown (Table 1); list of the SDA (with letter code
corresponding to Supplementary Figure 1) with detailed infor-
mation on synthesis of the amines. This material is available
free of charge via the Internet at http://pubs.acs.org.
There are a few exceptions to the rule we established. The
C8 and C12 trimethylammonium diquaternary molecules are
JA070303U
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7637