Computational and experimental approach to the role of structure-directing agents in the synthesis of zeolites: The case of cyclohexyl alkyl pyrrolidinium salts in the synthesis of β, EU-1, ZSM-11, and ZSM-12 zeolites

Article abstract of DOI:10.1021/jp027506j

The role of structure-directing agents (SDA) in the synthesis of zeolites is investigated, and the structures obtained in the synthesis are rationalized in terms of the energetic stabilization between the SDA and the microporous zeolite structure. An explanation is provided for the synthesis outcome in terms of a balance between kinetic and thermodynamic factors throughout the nucleation and crystallization stages. The stability of β, EU-1, ZSM-11, and ZSM-12 zeolites is calculated over a wide range of Si/Al ratios when cyclohexyl alkyl pyrrolidinium salts are used as the SDA. The role of the SDA allows us to explain the final stability and the Si/Al range in which each structure can be synthesized. The stabilization of intermediate species during the nucleation is proposed to orient the final result of the synthesis. A simple kinetic model is proposed to explain the synthesis process.

Full text of DOI:10.1021/jp027506j

5432
J. Phys. Chem. B 2003, 107, 5432-5440
Computational and Experimental Approach to the Role of Structure-Directing Agents in the
Synthesis of Zeolites: The Case of Cyclohexyl Alkyl Pyrrolidinium Salts in the Synthesis of
â, EU-1, ZSM-11, and ZSM-12 Zeolites
German Sastre,* Sandra Leiva, Maria J. Sabater, Ignacio Gimenez, Fernando Rey,
Susana Valencia, and Avelino Corma
Instituto de Tecnologia Quimica U.P.V.-C.S.I.C., UniVersidad Politecnica de Valencia,
AVenida Los Naranjos s/n, 46022 Valencia, Spain
ReceiVed: NoVember 20, 2002; In Final Form: April 8, 2003
The role of structure-directing agents (SDA) in the synthesis of zeolites is investigated, and the structures
obtained in the synthesis are rationalized in terms of the energetic stabilization between the SDA and the
microporous zeolite structure. An explanation is provided for the synthesis outcome in terms of a balance
between kinetic and thermodynamic factors throughout the nucleation and crystallization stages. The stability
of â, EU-1, ZSM-11, and ZSM-12 zeolites is calculated over a wide range of Si/Al ratios when cyclohexyl
alkyl pyrrolidinium salts are used as the SDA. The role of the SDA allows us to explain the final stability and
the Si/Al range in which each structure can be synthesized. The stabilization of intermediate species during
the nucleation is proposed to orient the final result of the synthesis. A simple kinetic model is proposed to
explain the synthesis process.
1
. Introduction
energy minimization techniques to calculate the stability and
location of tetraalkylammonium cations in ZSM-5, ZSM-11,
and â zeolites and bis-quaternary amines in EU-1 and ZSM-
The ability of certain organic and inorganic species to direct
the synthesis of crystalline aluminosilicates toward a particular
structure is determined by kinetic and thermodynamic factors,
the latter being determined by the stabilization gained by the
system SDA-zeolite with respect to the separate components.
The use of suitable organic SDA molecules in the synthesis
allows us to select between possible phases with a priori similar
2
3. Their results allow us to achieve a de novo synthesis by
selecting the appropriate template for a particular structure. A
similar technique was used to explain the influence of the SDA
molecule and cobalt concentration in the competitive formation
of Co-AlPO-5 and Co-AlPO-34 starting from similar synthesis
13
gel compositions. Monte Carlo docking algorithms based on
1,2
thermodynamic stability and allows us to make more energeti-
cally favorable the synthesis of a given structure. Framework
inorganic cations can also have an influence in the first stages
of the synthesis during the nucleation process by orienting the
formation of Si-O-T (T ) tetrahedral atom) oligomers toward
certain secondary building units that influence the final crystal-
a force field approach combined with crystallographic data have
14
been used by Toby et al. to study the effect of structure-
directing agents in the inhibition or formation of stacking faults
in CIT-1. A number of other studies based on computer
simulation of the rationalization of zeolite synthesis are available
15-18
in the literature,
and they predict the stability of organic
lized structure(s).3 Other factors, such as OH/Si and T /T
ratios, are also crucial in directing the synthesis toward a
particular zeolite.
-5
IV III
species occluded in microporous materials. Recent studies by
19-21
Shantz et al.
explain how the distribution of framework
aluminum in ZSM-12 can be controlled by the template location,
The organic can sometimes act as a pore-filling agent, and
when this occurs, the structure-directing effect of the organic
is limited. In other syntheses, an important interaction between
the SDA and the framework occurs, and then the selectivity of
the organic toward the formation of a given structure is higher
than in the previous case. Finally, there are very few structures
in which a full match between the organic and the inorganic
counterpart exists and the SDA can be considered to be a
template for that particular structure.6-11 Quaternary ammonium
salts have become widely used as SDAs because they allow
strict control of the hydrophobic/hydrophilic properties by
22
and a similar effect has been observed in ZSM-18. The purpose
of this work is to test two SDAs with related molecular
+
structures and C/N ratios that should behave similarly from
the point of view of pH, solubility, and hydrophobicity.
However, the geometrical differences in the two molecules
should influence the energies of zeolite-SDA interactions and
the stability of the pores and topologies formed. The experi-
mental results obtained by using these SDA molecules in the
synthesis of zeolites will be rationalized on the basis of
molecular simulations.
+
controlling the C/N ratio of molecules within a large range of
sizes and shapes.
2
. Experimental Section
The SDAs were prepared by following a well-known
Computational methods help us to understand better the
physicochemical interactions between the SDA and the zeolite
framework. Lewis et al. have combined Monte Carlo and
methodology, namely, the reductive amination of carbonylic
23
12
compounds. In this case, the reaction proceeds via a condensa-
tion between an amine and a ketone at pH e 8, thus forming
an imine or enamine (depending on whether the amine is
primary or secondary). Under these experimental conditions,
*
To whom correspondence should be addressed. E-mail: gsastre@
itq.upv.es.
1
0.1021/jp027506j CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/13/2003

Structure-Directing Agents in Zeolite Synthesis
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5433
SCHEME 1: Mechanism of the Reductive Amination
Methodology
SCHEME 2: Representation of the Synthesis of
Organocations 1 and 2
both the imine and enamine are reduced as imminium salts to
form the new amines (Scheme 1) that are finally quaternized
by reaction with an alkylhalide, namely, iodomethane or
bromobutane, to yield to the quaternary tetraalkylammonium.
The initial amine that was chosen was pyrrolidine, which had
already been used in the form of the N,N-dimethyl-pyrrolidinium
cation for the synthesis of other zeolites. This cation showed,
under certain conditions, a low structure selectivity, and a wide
mixture of medium- and large-pore zeolites as well as clathrasils
was obtained. It was thought that when using the pyrrolidine in
the reductive amination of cyclohexanone a series of larger
molecules that could be transformed into quaternary ammonium
salts would be of better selectivity and at the same time their
use may avoid the formation of clathrasils. Scheme 2 shows
Anal. Calcd for C11H23NO: C, 44.7; H, 7.8; N, 4.75. Found:
C, 43.5; H, 8.05; N, 4.83.
Characterization Data for Compound 2. 13C NMR (200
MHz; D2O): δ 77.28, 66.83, 62.84, 33.68, 32.06, 31.58, 30.63,
15.48. Anal. Calcd for C14H28BrN (+2H2O): C, 51.54; H, 9.81;
N, 4.30. Found: C, 51.51; H, 9.45; N, 4.23.
The above synthesized halide tetraalkylammonium salts were
transformed to the corresponding hydroxide forms by anionic
exchange using a DOWEX resin in a water solution. The degree
of exchange always leaves the halide concentration below the
detection limit of the corresponding selective electrode. There-
fore, the total concentration of SDAOH was achieved by acid
titration using phenolphthaleine as an indicator.
+
the cations obtained and their corresponding C/N ratios.
2
.1. Synthesis of Structure-Directing Agents. Quaternary
2.2. Synthesis of Zeolites. The influence of the different
structure-directing agents on the synthesis of zeolites was tested
in gels having the following general molar composition: SiO2/
xAl2O3/0.54SDA(OH)/0.54HF/7.25H2O. SDA(OH) is the struc-
ture-directing agent employed in each synthesis. SiO2 was
incorporated into the synthesis gel as tetraethylortosilicate
(TEOS from Aldrich), and Al2O3, as aluminum triisopropoxide
(AIP from Aldrich Co.). The influence of the aluminum content
in the synthesis gel was studied by varying x ratios between 0
and 0.0333.
In a typical synthesis, TEOS and AIP were added to a SDA-
(OH) solution. The initial clear solution was vigorously stirred
until the alcohols produced during the hydrolysis of TEOS and
AIP were completely removed and the required water content
was reached by slow evaporation at 25 °C. Then, a 50 wt %
aqueous solution of HF was added to the above reaction mixture,
resulting in the immediate formation of a very thick gel that
was manually homogenized. The final gel was loaded in Teflon-
lined stainless steel autoclaves at different temperatures (135,
150, and 175 °C), and samples were withdrawn at different
crystallization times. The resulting zeolitic materials were
collected by filtration, exhaustively washed with distilled water
and acetone, and finally dried at 100 °C overnight. The different
zeolite phases formed during the crystallization process were
identified by an X-ray diffraction technique. The different
ammonium salts were prepared from a series of amines,
synthesized by the reductive amination strategy mentioned
above. The following procedure was generally used: 50 mL of
5
N HCl-methanol was added to a solution of 44.6 g (638
mmol) of pyrrolidine in 200 mL of methanol, followed by the
addition of 10.0 g (102 mmol) of cyclohexanone and 5.1 g (81
mmol) of NaBH3CN. The resulting solution was stirred for 72
h and then concentrated, and 5 N HCl was added until pH < 2
was reached. Then, the methanol was removed under vacuum
at room temperature. The residue was washed and purified with
two 200-mL portions of ether. The aqueous solution was brought
to pH > 12 with KOH, saturated with NaCl, and extracted with
three 100-mL portions of ether. The organic extract was dried
(
MgSO4) and evaporated in vacuum. The resulting tertiary
amines were quaternized with different alkyl halides as fol-
lows: a 250-mL round-bottom flask was charged with 10.0 g
(
65.4 mmol) of the cyclohexylpyrrolidine amine and 100 mL
of CHCl3. A solution of 126.7 mmol of alkyl iodide (methyl or
butyl bromide) and 50 mL of CHCl3 was added dropwise. A
white solid was formed almost immediately. The reaction was
stirred at room temperature for 3 days. The solid was collected
by filtration, washed exhaustively with ether, and dried.
Characterization Data for Compound 1. 13C NMR (200
MHz; D2O): δ 81.75, 70.63, 58.5, 34.27, 31.89, 31.59, 27.91.

5434 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Sastre et al.
TABLE 1: Synthesis Conditions and Detected Phases
purely siliceous ones. In fact, EU-1 was the unique phase
detected when the Si/Al ratio of the synthesis was 50, and â
becomes the most stable product by further increasing the Al
content until the Si/Al ratio is 25. Finally, lamellar phases start
to be formed in the synthesis gel of a Si/Al ratio of 15. This
indicates that there is a clear directing effect toward more open
structures as the Al content in the gel (and also in the final
solids) increases. This result can be rationalized by considering
that the amount of compensating positive charged cations
occluded within the zeolitic micropores must increase as the
aluminum content increases; therefore, more open structures,
such as â or EU-1, will be favored.
SDA
cation Si/Algel
temperature time
(°C) (days)
phase
sample
1
2
4
1
4
6
1
â
1
2
3
4
5
1
35
ZSM-12 + â
ZSM-12 + â
ZSM-12
∞
1
50
75
0
1
ZSM-12
1
5
0
135
1
1
17
14
5
EU-1
EU-1
EU-1
6
7
8
50
75
2
5
5
175
175
16
15
â
9
1
â + laminar phase
10
3
. Computational Methods
1
4
4
2
â + ZSM-11
ZSM-11
11
12
150
The calculations have been performed using lattice-energy
24
minimization techniques and the GULP code. The interatomic
potentials used to model the interactions between the atoms in
the zeolite structure included the following terms: Coulombic
interaction, short-range pair potentials (described by a Buck-
ingham function), and a three-body bond-bending term. The
shell model was used to simulate the polarizability of the oxygen
atoms. A cutoff distance of 12 Å was applied to the short-range
interactions (Buckingham- and Lennard-Jones-type interactions,
see eqs 3 and 6 below). The Ewald summation technique has
∞
5
1
â + ZSM-11
13
14
2
175
2
ZSM-11
5
2
0
175
175
3
ZSM-11
ZSM-11
15
16
5
18
been used for the summation of the long-range Coulombic
interactions. The potentials used for the zeolite2
5,26
were
parametrized to reproduce the structure of the R-quartz and have
been demonstrated to model a number of zeotype structures
27-30
31
successfully.
The force field by Kiselev et al. was selected
for the intermolecular SDA-zeolite and SDA-SDA interac-
3
2
tions, and the force field by Oie et al. was selected for
intramolecular interactions between the atoms of the SDA. For
the organic SDA, the charge distribution has been obtained by
means of the quantum chemistry Hartree-Fock method by using
33
a 6-31G** basis set, and the calculations have been performed
3
4
by means of the NWCHEM package. More details of the
3
5,36
computational methods can be found elsewhere.
The total
potential energy function and the respective terms are as follows:
Figure 1. XRD patterns of some representative samples.
Vtotal ) Vzeo + VSDA + VSDA-SDA + Vzeo-SDA
(1)
synthesis gel compositions and the zeolite phases observed in
this work are listed in Table 1. Also, representative XRD
patterns of the different phases that were detected are shown in
Figure 1. XRD patterns were acquired using Cu KR radiation
on a Phillips X’Pert MPD diffractometer equipped with a
PW3050 goniometer using secondary monochromated Cu KR
radiation and a step size and time per step of 0.02° and 2 s,
respectively.
It was observed that the use of cation 1 in the synthesis of
purely siliceous zeolites yields the formation of zeolite â when
low crystallization temperatures or times are used (samples 1-3
in Table 1), whereas the ZSM-12 structure is the unique phase
detected after crystallization at high temperatures or relatively
long synthesis times (samples 4 and 5). Similar results were
observed in the syntheses carried out in the presence of cation
Vzeo ) VBuckingham + VCoulombic + Vthree body + Vcore-shell (2)
VSDA ) Vtwo body + Vthree body + Vfour body + VCoulombic
(3)
(4)
(5)
VSDA-SDA ) VLennard-Jones + VCoulombic
Vzeo-SDA ) VLennard-Jones + VCoulombic
rij
Cij
V (Buckingham) ) A ‚exp
-
(
-
(6)
(7)
ij
ij
F)
6
r
qi‚qj
rij
Vij(Coulombic) )
1
2
0
2
V (three body) ) k ‚(θ - θ ijk)
ijk
ijk
ijk
with θ ) O-T-O (8)
2
, but in this case, the end product was the ZSM-11 structure
1
2
0
2
V (core-shell) ) k ‚(r - r )
(9)
instead of the ZSM-12 zeolite. These results indicate that zeolite
â seems to be a kinetic product rather than the thermodynamic
solid using either cation 1 or 2. Therefore, most of the study
presented here was performed at relatively high temperature (175
ij
ij
ij
ij
1
2
0
2
V (two body) ) k ‚(r - r )
(10)
ij
ij
ij
ij
°
C) and prolonged crystalization times.
It is well known that by introducing Al into the synthesis
V (four body) ) A ‚[1 + cos(n‚φ - δ )]
(11)
(12)
ijkl
ijkl
ijkl
ijkl
Bij Cij
gels in the presence of cation 1 the SDA gives way to the
formation of different phases of those obtained by the analogous
Vij(Lennard-Jones) )
-
1
2
6
r
r

Structure-Directing Agents in Zeolite Synthesis
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5435
Figure 2. Energy scheme (top) for the synthesis of two zeolite structures (Z) in the presence of a SDA starting from the same gel (G) and going
through different nuclei (N). Path 1 requires less activation energy for the nucleation stage, and path 2 gives the most stable final structure. A
mechanism (middle) in two stages is proposed with a reversible nucleation followed by the crystallization. Nuclei form (bottom) when silicoalumina
oligomers surround the SDA, giving an aggregate resembling a significant part of a structure.
A final consideration regarding the methodology that was used
concerns the possible presence of water in these systems and
whether water should be included in the calculations. It is known
that zeolites synthesized as high Si/Al ratio materials and in
the presence of fluoride anions are obtained as nondefective
materials and therefore are highly hydrophobic, this being the
case for the syntheses reported here, which makes the influence
of water unimportant.
Also, it is generally accepted that the synthesis of zeolites
occurs through consecutives steps. The most important step
consists of the organization of the hydrophobic organocation
with silicate species in solution. This interaction proceeds via
an overlap of the hydration spheres, and then there is a strong
interaction between the hydrophobic organocation and the
silicate oligomers. It has been proven that this second step leads
to phase selectivity toward a precise zeolite structure, and the
presence of water molecules does not seem to be of importance;
in fact, it is thought that water is mostly excluded in the
formation of these prenuclation composites. This concept is
fulfilled by the empirical observation that the largest organo-
cation is the most selective toward a single phase. In our
calculations, we have tried to model these composites as the
starting point during zeolite formation; therefore, water has not
been included in our calculations.
4. Theory
Microporous materials provide a host-type environment for
stabilizing charged molecular species occluded in their cavities.
The stabilization of the “zeolite-guest” system with respect to
the separate components, “zeolite + guest”, is due to the
interactions between the guest and the zeolite. The energy
change in the synthesis process starts from the initial reactants
(
a silica-alumina gel plus the organic SDAs) and goes toward
the final zeolite-SDA system.
4.1. Kinetic and Thermodynamic Aspects during the
Zeolite Synthesis. The synthesis of zeolites may be seen (Figure
2) as a series of consecutive reactions that occur starting from
the gel, and then silicoalumina oligomers start to form (pre-
nucleation). Clusters of secondary building units or zeolite
fragments of little significance then appear when small nuclei
grow (nucleation), and this occurs around the SDA cations if
they are present in the synthesis gel. Larger nuclei continue to
grow and lead eventually to the final zeolite structure (crystal-
lization). In a simplified treatment, we consider the nucleation
and crystallization to be the two main steps (Figure 2). It is
seen that there are two competitive consecutive reactions, each
of them leading during the nucleation stage to the formation of
a stable nucleus (N1, N2) that consists of the SDA cations

5436 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Sastre et al.
+
+
-
surrounded by a series of silica oligomers adopting a distribution
resembling that of a significant zeolitic fragment. Two ap-
proximations to such possible nuclei, for the sake of visualiza-
tion, are also depicted in Figure 2. The energy of such possible
aggregates depends mainly on two energetic terms: the energy
of the corresponding SDA conformation and the stabilization
due to interactions between the SDA and the silicoalumina
oligomers forming the nucleus. For a more detailed analysis of
these two contributions and all of the other terms contributing
to the total energy, see Appendix 1. The relative stability of
the different nuclei (N1, N2) will give us an estimation of which
zeolite is preferentially forming during the early stages of the
synthesis. It is clear from our kinetic model that if N1 is formed
preferentially then Z1-SDA will be the synthesis product,
independently of whether it is more or less stable than Z2-
SDA. The thermodynamic product (Z2-SDA in the case
illustrated in Figure 2) will be the synthesis outcome when the
activation energies are low enough to compete with parallel
reactions. Increasing the temperature and time of the synthesis
will favor the thermodynamic product. Although this is a
simplified treatment and aspects such as, for example, the
competitive reactions between nucleus formation and dissolution
are more complex, the basic phenomena of the zeolite synthesis
are undertaken by our model.
fulfilled that the concentration of SDA plus [SDA] [F]
(SDA ) is constant (C).
0
For two structures, eq 14 can be rewritten as
n1
x1
+
+
) C1
) C2
m + n
m + n
1
1
1
1
n2
x2
m + n
m + n
2
2
2
2
The energy per TO2 (T ) Si, Al) associated with eqs 13 can be
written as follows:
E(zeo -SDA) E(geli)
ni
i
+
∆E
)
-
-
‚E(SDA ) -
i
m + ni
m + ni m + ni
i i
i
xi
0
‚
E(SDA )
(i ) 1, 2) (16)
m + ni
i
To address our question of whether zeo1 or zeo2 is preferentially
formed, the following equations have to be considered:
∆
E ) ∆E - ∆E
(17)
F
2
1
4
.2. Graphic Representation of Zeolite Stability at Dif-
ferent Al Contents. Our approach aims to answer the question
of which zeolite structure(s) is(are) thermodynamically favored
from a given set of initial conditions.
∆E > 0 w zeo -SDA is more stable
F 1
∆
E < 0 w zeo -SDA is more stable
F
2
Starting from the question of which zeolite(s) is(are) pref-
erentially formed from a given set of initial conditions, the
following equations can be written:
E(zeo -SDA) E(zeo -SDA) E(gel2)
2
1
∆E )
-
-
+
F
m + n
m + n
m + n2
2
2
1
1
2
E(gel1)
n2
n1
+
+
0
-
-
‚E(SDA ) -
(
)
gel(m Si, n Al) + n SDA + x SDA f zeo -SDA (13)
m + n
m + n
m + n
1
1
1
1
1
1
1
2
2
1
1
x2
x
1
+
0
0
gel(m Si, n Al) + n SDA + x SDA f zeo -SDA
2
2
2
2
2
-
‚E(SDA ) (18)
(
)
m + n
m + n
2
2
1
1
To calculate the zeolite-SDA energy in our system, it is
supposed that the negative charge brought by Al into the zeolite
framework is compensated for by a positively charged SDA
molecule. Also, the microporous space of the zeolite that is
formed is completely filled with organic species. This means
that when the positively charged SDA does not fill the
microporous voids, neutral [SDA] [F] moieties (called SDA )
are located in the corresponding microporous space that is left.
Therefore, the total amount of organic material (SDA + SDA )
is constant in each structure regardless of the Al content
incorporated by the framework, as demonstrated by the ther-
mogravimetric measurements and elemental analyses. Because
The energy of the gel per TO2 unit is the same in both cases,
which means that the corresponding terms vanish. Also, this
process refers to different zeolites formed from the same
synthesis gel, which means that the Al content is the same in
both cases:
+
+
+
-
0
n2
n1
+
0
)
(19)
m + n
m + n
2
2
1
1
+
Therefore, the term containing E(SDA ) also vanishes, and we
+
-
have the following:
it is not possible to model [SDA] [F] , the closest neutral
compound is a neutralized SDA molecule in which a fixed
0
E(zeo -SDA) E(zeo -SDA)
number of atomic charges have been added so as to make the
total charge equal to zero by means of a standard averaging
procedure.
From the experimental conditions outlined, two conditions
follow:
2
1
∆EF )
-
-
m + n
m + n
2
2
1
1
3
7
x2
x1
0
-
‚E(SDA ) (20)
(
)
m + n
m + n
1 1
2
2
Finally, by substituting the values of x1/(m1 + n1) and x2/(m2 +
n2) from eqs 14 and taking into account eq 18, the final
expression for ∆EF is obtained:
n
x
+
) C
(14)
(15)
m + n m + n
nAl w nSDA+
E(zeo -SDA) E(zeo -SDA)
2
1
∆
EF )
-
-
n is the number of Al atoms, and this means that each Al atom
is compensated by a SDA molecule; x is the number of SDA
molecules, and m is the number of Si atoms. The condition is
m + n
m + n
2
2
1
1
+
0
0
(
C - C )‚E(SDA ) (21)
2 1

Structure-Directing Agents in Zeolite Synthesis
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5437
To compare the relative stabilities of zeolites, the following
structure-dependent magnitude Z is defined:
E(zeo -SDA)
ni
i
+
Zi )
-
‚E(SDA ) -
m + n
m + ni
i
i
i
xi
0
‚
E(SDA ) (22)
m + ni
i
From what has been demonstrated above, it can be seen that
∆
∆
∆
Z ) Z - Z ) ∆E
F
(23)
(24)
(25)
2
1
E > 0 w zeo -SDA is more stable w Z < Z
F
1
1
2
E < 0 w zeo -SDA is more stable w Z < Z
F
2
2
1
From this, it follows that plotting Z for each structure versus
Al content gives the relative stabilities of the different structures.
Low(high) values of Z will refer to the more(less) stable
structures. Also, for each structure, an interval of Al composition
is defined with a minimum and a maximum value within which
the structure can be synthesized. Equation 14 shows that the
minimum value for Al content is zero (pure silica polimorph)
and that the maximum value is C, which refers to the optimum
+
loading of SDA in the structure. Beyond that, further Al cannot
+
be compensated for by any incoming SDA molecules because
all of the microporous voids are already filled.
5
. Results and Discussion
5
.1. Cation N-Cyclohexyl-N-methyl-pyrrolidinium. This
+
SDA (cation 1 in Scheme 2) gives mainly unidirectional
3+
zeolites. When Al is introduced into the reaction media (Si/
Al ) 50), EU-1 is obtained within a wide range of time and
temperature (Table 1). The calculated final energies (Figure 3)
show that the most stable structure is ZSM-12, but the larger
porosity of EU-1 allows more SDA cations to enter the pores,
and for an Al content, Al/(Al + Si), in the range of [0.036,
Figure 3. Plot of Z (as defined in eq 22) vs Al content for the synthesis
of â, EU-1, ZSM-11, and ZSM-12 zeolites with cation 1 (Scheme 2).
The relative final stability of different zeolites at a given Al content
can be found from the graph as well as the Al composition range that
allows a structure to be synthesized, which depends on the loading
that the micropores can host.
0
.055], EU-1 should be the structure that is obtained. This
corresponds to a Si/Al ratio within about 17 e Si/Al e 27, and
this agrees with the results of the synthesis. The calculations
have been performed with the SDA loadings indicated in Table
(3.15 and 3.35 eV, respectively, from Table 2) allows us to
predict that EU-1 should be the preferentially obtained zeolite,
and this corresponds to the results of the experiments (Table
1). The term E(SDA-zeo) (Table 2) relates to a contribution
that plays an important role during the nucleation stages of the
synthesis. This term is mostly a short-range term between the
organic cation and the zeolite surroundings, and these are
precisely the interactions present when polymeric species start
to form the zeolite micropore around the SDA cation. This can
also explain why ZSM-11 is not the observed product in this
synthesis although its energy is practically the same as that of
EU-1 over the entire Al composition range (Figure 3). According
to the respective contributions, -0.070 and -0.059 eV/SiO2
for EU-1 and ZSM-11, respectively (Table 2), the nucleation
of EU-1 would be more favorable, and the competition during
the first stages of the synthesis precludes the formation of the
equally stable ZSM-11 in favor of EU-1. Finally, the experi-
ments show that â is the only product obtained at larger Al
contents, and this is also predicted from our results in Figure 3,
which show that EU-1 cannot hold more organic material in its
microporous space to compensate for the Al above Al/(Al +
Si) > 0.055 (that is, Si/Al < 17), which is quite close to the
Si/Al ratio at which â appears as the only product (Si/Al <
25).
2
, which correspond closely to the experimental loadings. These
energetic considerations, based on the final energy of the
product, also allow us to predict the obtaining of ZSM-12 when
Al is not present in the synthesis gel, and this is especially
observed when the synthesis is carried out at conditions that
favor the thermodynamically stable products (longer time and
higher temperature). It is clear from Table 1 that in syntheses
carried out in the absence of Al the primary product is the most
stable nucleus (â zeolite), but because of the relatively high
solubility of silica (compared to silica-alumina), the primary
product evolves toward the most thermodynamically stable
material, which is ZSM-12 (footnote a in Table 2). Nevertheless,
kinetic conditions are the most important, and when the
nucleation is the determining step, the products that are formed
will be those that stabilize the intermediates. This energy is
calculated as the short-range interaction energy between the
SDA and its closest neighbors in the microporous structure
(ESDA-zeo in Table 2). It is seen that this stabilization is larger
for the more open structures (-0.070 and -0.072 eV/SiO2 for
EU-1 and â, respectively) than for the monodirectional ZSM-
1
the observed product when Al is introduced into the synthesis
gel. Also, the higher stability of the cation in EU-1 than in â
2 (-0.045 eV/SiO2), and this explains why ZSM-12 is not

5438 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Sastre et al.
TABLE 2: Calculated Energy Terms, as Defined in Appendix 1, Involved in the Synthesis of â, EU-1, ZSM-11, and ZSM-12
Zeolites Obtained with the Use of SDA Cations 1 and 2 (Scheme 2)a
d
exptl loading
org. mol/TO
calcd loadingb
E′(SDA+)c
E(SDA-zeo)
(eV/TO
(
2
)
(org. mol/TO
2
)
(eV)
2
)
â-cation 1
0.074
0.056
not observed.
0.040
0.070
0.054
0.052
0.042
3.35
3.15
3.13
3.02
-0.072
-0.070
-0.059
-0.045
EU-1-cation 1
ZSM-11-cation 1
ZSM-12-cation 1
â-cation 2
not measured
not observed
0.045
0.047
0.036
0.042
0.029
3.46
3.39
3.13
3.45
-0.061
-0.053
-0.062
-0.044
EU-1-cation 2
ZSM-11-cation 2
ZSM-12-cation 2
not observed
a
The energies of the calculated silica zeolite without organic occluded (E
eV/SiO for â, EU-1, ZSM-11, and ZSM-12, respectively. The respective calculated densities are 1.57, 1.80, 1.78, and 1.95 g/cm . After the organic
incorporation, the subsequent zeolite deformation leads to the following energies: -128.55, -128.58, -128.59, and -128.62 eV/SiO for â, EU-1,
ZSM-11, and ZSM-12 respectively (Ezeo′ in Appendix 1). b Organic species present as SDA (x ) 0 in eq 14). Loadings correspond to C in eq 14.
1
in Appendix 1) are -128.56, -128.59, -128.59, and -128.62
3
2
2
+
c
+
d
Organic species present as SDA (x ) 0 in eq 14). Gas-phase energies: cation 1: 2.93 eV, cation 2: 2.91 eV. Only short-range interatomic
contributions included, without the electrostatic term (E + E in Appendix 1).
4
5
5.2. Cation N-Cyclohexyl-N-butyl-pyrrolidinium. To probe
the effect of the SDA cation conformation on the synthesis of
zeolites, we have carried out experiments using a related organic
cation such as SDA (cation 2 in Scheme 2) in which the butyl
group substitutes for the methyl group in the former SDA cation.
It is expected that the presence of the butyl group could
discourage the formation of monodirectional channel zeolites
such as EU-1 or ZSM-12. Effectively, when this cation is
introduced into the synthesis gel, it gives ZSM-11 (plus â in
some cases) in all of the syntheses that were carried out (Table
1
), indicating that the larger constraint introduced as a conse-
quence of the butyl group makes the small channels of EU-1
no longer favorable to hosting a viable loading that stabilizes
+
its synthesis. The larger size of this SDA causes the organic
conformation to influence the synthesis process. The calculations
allow us to fit loadings as indicated in Figure 4 and Table 2,
again quite close to the experimental results. Although in Figures
3
and 4 the lines representing each zeolite type are very close
to each other, the differences are in agreement with the
experimentally measured data.2 The final geometry of the SDA
cations is very much dependent on the constraining micropore,
and Table 2 shows energies above 3.45 eV for the cations in â
and ZSM-12 zeolites, whereas only 3.13 eV is obtained in ZSM-
,38
1
1 (close to the absolute minimum energy of 2.91 eV). Also,
as previously mentioned, the short-range interaction energy
between the SDA cation and the zeolite is important to the
orientation of the synthesis during the nucleation, and this term
(
Table 2) is more favorable in ZSM-11 and â zeolites (-0.062
and -0.061 eV/TO2, respectively) than in EU-1 and ZSM-12
(
above considerations, ZSM-11 and â should be the synthesis
-0.053 and -0.04 eV/TO2, respectively). Therefore, from the
products, and this is exactly what is found in the experiments
(Table 2). Although the stabilization coming from the short-
range interaction energy points toward both ZSM-11 and â as
synthesis products, we think that the observed preferential
synthesis of ZSM-11 is due to the fact that the SDA cation is
more relaxed inside ZSM-11 than in â, as shown by the
respective energies of 3.13 and 3.46 eV (Table 2). The shape
of the micropores of ZSM-11 seems to be more suited than the
other structures to host this bulkier cation. However, the
possibility of the â zeolite forming does exist despite the higher
energy required for the SDA cation to adopt a conformation
suited to the micropore shape. This high-energy conformation
will require a larger activation energy for the corresponding
nucleus to be formed. Again, as in the previous case with cation
Figure 4. Plot of Z (as defined in eq 22) vs Al content for the synthesis
of â, EU-1, ZSM-11, and ZSM-12 zeolites with cation 2 (Scheme 2).
The relative final stability of different zeolites at a given Al content
can be found from the graph as well as the Al composition range that
allows a structure to be synthesized, which depends on the loading
that the micropores can host.
in Figure 2, middle) evolve toward the most stable structure
between the competing products (â and ZSM-11), which is
ZSM-11 (footnote a in Table 2). Previous studies have proposed
that template configurations occluded in the microporous voids
3
9
are similar to the equilibrium. Our model supports this
hypothesis with the corolary that reaction temperatures always
allow a significant degree of deformation of the SDA organic
1
, the use of pure silica gels allows a higher solubility, and the
primary products in the competitive reactions (kinetic model

Structure-Directing Agents in Zeolite Synthesis
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5439
cations. In many cases, a certain degree of flexibility is crucial
so as to allow the necessary match between SDA and respective
microporous shapes.22 This SDA cation (N-cyclohexyl-N-butyl-
pyrrolidinium) contains a flexible butyl group whose different
possible orientations differ little in energy and allow significant
differences in size and shape, which lead to the formation of at
least two different zeolites, ZSM-11 and â, provided, as seems
to be the case, that the reaction conditions allow it to overcome
the activation energy necessary to reach the adequate SDA
cation conformations.
topology and Al location (ESifAl); (iii) the incorporation of the
SDA into the microporous voids, which produces four energetic
effects: modifies its conformation with respect to the equilib-
+
rium (∆ESDA ), stabilizes the zeolite-SDA interactions through
+
short-range and long-range terms (E
zeo-SDA
), increases the
+
+
framework energy (Ezeo′), and SDA -SDA interactions be-
+
+
+
+
tween the occluded organic appear (E
SDA -SDA
). SDA -SDA
+
interactions depend strongly on the number of SDA molecules
hosted in the zeolite lattice, and the optimum loading is reached
+
when a compromise is reached between the repulsive SDA -
SDA interactions (which preclude excessive loading) and the
attractive zeolite-SDA terms (which avoid loadings that are
+
+
6
. Conclusions
too low). For the calculations of the energy terms, the GULP
code has been used, and the geometry of the organic cations
and the zeolite structures has been optimized with the procedures
used in previous work. The energy decomposition has been
possible by taking the final geometry and calculating the
individual contributions of the corresponding force field acting
on the involved atoms. A more detailed analysis of each
contribution follows.
Short-range energy interactions are the dominating, albeit not
the only, factor that orients the zeolite synthesis. Thermodynamic
factors determine only when the activation energies of the
synthesis steps are low enough for the competitive reactions to
take place, and this occurs at long time, high temperature, and
high solubility. Therefore, from the two energetic schemes
outlined in this study, the analysis of energetic terms seems to
be more appropriate than the plotting of the defined Z function.
The energetic terms corresponding to the short-range interac-
tion between the zeolite and the charge SDA play an important
role during the nucleation stage, and the zeolite synthesis may
be mainly driven by this interaction. In such a case, the results
can be explained in terms of the stability of the appropriate
nuclei during the initial stages of the zeolite formation.
3
5
•
A fundamental role is played by the structural type that
crystallizes from the synthesis gel. The size and shape of the
microporous voids determine the framework density, which
correlates with the enthalpy of formation to a large ex-
27,38,40,41
tent.
The comparison between different structural types
has to be performed with equal compositions, hence SiO2
structures are used for this purpose. Optimized silica structures
can be easily calculated within our computational methodology,
and energy difference between the initial gel and the crystallized
zeolite can be calculated by taking only the energy of the
optimized zeolite to estimate relative energies because the
different structures in our experiments crystallize from the same
synthesis gel, which is taken as our zero of energy. Hence, E1
) Ezeo-Si. The corresponding energies appear in Table 2.
• Different abilities of a given structural type to incorporate
aluminum in the framework can be expected. An estimation of
this can be achieved by using the Mott-Littleton methodology35
to calculate defect energies. Although Al incorporation should
not be expected to remain constant as Al content increases, a
simplication can be made when incoming Al atoms are located
far apart and are nearly noninteracting. In that case, an average
over the different T sites can give a reasonable approximation.
Calculating the average defect energy, ESifAl, will give us the
In the case of the synthesis with the N-methyl-N-cyclohexyl-
pyrrolidinium cation, ZSM-12 (the most stable zeolite) is the
product obtained when the synthesis is carried at conditions that
favor the thermodynamically stable products and without Al,
which by making the gel more soluble allows the competitive
reactions to drive the products toward the more thermodynami-
cally stable. As the Al content increases, EU-1 becomes the
+
favored product because of the favorable zeolite-SDA interac-
+
tions and the low SDA energy. At even larger Al content, EU-1
cannot hold the necessary organic to compensate for Al, and â
becomes the observed product because of a favorable zeolite-
+
SDA interaction.
The conformation of the occluded organic cation inside the
zeolite is an important parameter. Zeolite structures that force
the occluded SDA+ to adopt a conformation far from that of
the equilibrium conformation in solution are not favored.
Nevertheless, flexible templates may offer a wide range of
conformations that can be reached under these synthesis
conditions. Butyl cations can adopt a conformation that orien-
tates the synthesis toward â. In most cases, the more favored
conformation and more stable nucleus leading to ZSM-11 is
the observed product.
total energy needed to accommodate Al atoms as E2
)
∆EAl incorporation ) nESifAl, where n is the number of Al atoms
incorporated, which should be small enough to make Al-Al
interaction negligible. The average Mott-Littleton energies (in
eV) related to the Al incorporation for the considered structures
are 38.1 (â), 38.4 (EU-1), 38.3 (ZSM-11), and 38.5 (ZSM-12).
The standard deviation is below 0.1 eV in all cases.
+
Acknowledgment. We thank Generalitat Valenciana (project
4
GV01-492) for financial support and C (Centre de Computacio
•
SDA molecules deform when entering the zeolite micro-
i Comunicacions de Catalunya) and Centro de Calculo de la
Universidad Politecnica de Valencia for the use of their
computational facilities. The High Performance Computational
Chemistry Group from Pacific Northwest National Laboratory
Richland, WA) is acknowledged for making available NWChem
version 4.0, a computational chemistry package for parallel
computers.
cavities, thus increasing the energy of the former with respect
to the equilibrium energy. The difference in energy can be
calculated from the corresponding energies within the zeolite
+
+
cage (E′SDA ) and in the equilibrium free state (ESDA ) as
(
+
+
+
follows: E3 ) ∆ESDA ) E′SDA - ESDA . This term is shown
to be an important influence on the synthesis process because
during the nucleation process SDA cations have to adopt a
+
conformation that matches the channel/pore of the zeolite to be
formed. If this is a high-energy conformation, then only when
the associated activation energy is overcome will this zeolite
synthesis be possible. The flexibility of the SDA is an
important issue because flexible molecules can give several
structures at the expense of losing template ability, that is, the
Appendix 1
We consider three contributions to the energy of the final
+
+
system zeolite-SDA : (i) the stability of the structure (Ezeo);
(ii) Al incorporation into the zeolite framework, which is
expected to take different energetics according to the flexibility,

5440 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Sastre et al.
specificity to direct the synthesis toward a particular structure.
Therefore, templates have to be rigid molecules. Table 2 shows
the energies of SDA cations in the different zeolites considered
(13) Lewis, D. W.; Catlow, C. R. A.; Thomas, J. M. Chem. Mater. 1996,
, 1112.
8
(
14) Toby, B. H.; Khosrovani, N.; Dartt, C. B.; Davis, M. E.; Parise, J.
B. Microporous Mesoporous Mater. 2000, 39, 77.
15) Catlow, C. R. A.; Coombes, D. S.; Lewis, D. W.; Pereira, J. C. G.
Chem. Mater. 1998, 10, 3249.
16) Lewis, D. W.; Sankar, G.; Wyles, J. K.; Thomas, J. M.; Catlow,
C. R. A.; Willock, D. J. Angew. Chem., Int. Ed. Engl. 1997, 36, 2675.
17) van de Graaf, B.; Njo, S. L.; Smirnov, K. S. ReV. Comput. Chem.
000, 14, 137.
18) Wagner, P.; Nakagawa., Y.; Lee, G. S.; Davis, M. E.; Elomari, S.;
+
and in the gas phase.
(
•
Entering into the zeolite also has an effect of stabilization
(
due to the interaction with the zeolite walls, where mainly the
oxygen atoms, which are closer to the guest molecule than to
(
+
the T atoms, stabilize the positively charged SDA (E4 )
2
+
ESDA -zeo). This can be calculated from the corresponding
(
Lennard-Jones plus electrostatic terms between the zeolite and
Medrud, R. C.; Zones, S. I. J. Am. Chem. Soc. 2000, 122, 263.
+
(19) Shantz, D. F.; Lobo, R. F.; Fild, C.; Koller, H. Stud. Surf. Sci. Catal.
000, 130, 845.
the SDA (see eq 5). This term is related to the energetics of
2
+
the nucleation because it takes into account not only the SDA
(20) Shantz, D. F.; Schmedt auf der G u¨ nne, J.; Koller, H.; Lobo, R. F.
conformation around which the zeolite grows but also its major
J. Am. Chem. Soc. 2000, 122, 6659.
(21) Shantz, D. F.; Fild, C.; Koller, H.; Lobo, R. F. J. Phys. Chem. B
1999, 103, 10858.
+
contribution that is made by the silicate units closer to the SDA .
+
+
Similarly, SDA -SDA interactions appear when several
molecules are occluded within channels or cavities (E5 )
(22) Sabater, M. J.; Sastre, G. Chem. Mater. 2001, 12, 4520.
(23) Borch, R. F.; Bernstein, M. D.; Dupont Durst, H. J. Am. Chem.
+
+
ESDA -SDA ). This repulsive energy between positively charged
Soc. 1971, 93, 2897.
+
SDA molecules can be calculated from the Lennard-Jones and
(24) Gale, J. D. J. Chem. Soc., Faraday Trans. 1997, 93, 629.
electrostatics terms as indicated in eq 4. Table 2 shows the
(25) Sanders, M. J.; Leslie, M.; Catlow, C. R. A. J. Chem. Soc., Chem.
+
+
+
Commun. 1984, 1271.
corresponding energies (E4 + E5 ) ESDA -zeo + ESDA -SDA )
for the two organic cations considered in â, EU-1, ZSM-11,
and ZSM-12 structures. The dominant contribution is ∆E4,
which amounts to at least 90% of the total Lennard-Jones short-
range contributions between the zeolite and the organic cations.
(
(
26) Jackson, R. A.; Catlow, C. R. A. Mol. Simul. 1988, 1, 207.
27) Henson, N. J.; Cheetham, A. K.; Gale, J. D. Chem. Mater. 1994,
6
, 1647.
28) Henson, N. H.; Cheetham, A. K.; Gale, J. D. Chem. Mater. 1996,
8, 664.
(
•
The zeolite will be higher in energy than when isolated
(29) Modelling of Structure and ReactiVity in Zeolites; Catlow, C. R.
A., Ed.; Academic Press: London, 1992.
because of the deformation of the framework produced by the
+
(30) Catlow, C. R. A.; Bell, R. G.; Gale, J. D. J. Mater. Chem. 1994, 4,
incorporation of the SDA (E6 ) ∆Ezeo). This term can be
calculated from the final geometry of the system by taking off
the SDA molecules, calculating the energy of the corresponding
7
81.
31) Kiselev, A. V.; Lopatkin, A. A.; Shulga, A. A. Zeolites 1985, 5,
261.
(32) Oie, T.; Maggiora, T. M.; Christoffersen, R. E.; Duchamp, D. J.
Int. J. Quantum Chem., Quantum Biol. Symp. 1981, 8, 1.
33) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54,
24.
34) Harrison, R.; Nichols, J.; Straatsma, T.; Dupuis, M.; Bylaska, E.;
(
“
stressed” zeolite (Ezeo′), and subtracting the energy of the
isolated zeolite (Ezeo). Finally, ∆Ezeo ) Ezeo′ - Ezeo. The
corresponding energies of â, EU-1, ZSM-11, and ZSM-12 are
shown in Table 2.
(
7
(
Fann, G.; Windus, T.; Apra, E.; Anchell, J.; Bernholdt, D.; Borowski, P.;
Clark, T.; Clerc, D.; Dachsel, H.; de Jong, B.; Deegan, M.; Dyall, K.;
Elwood, D.; Fruchtl, H.; Glendenning, E.; Gutowski, M.; Hess, A.; Jaffe,
J.; Johnson, B.; Ju, J.; Kendall, R.; Kobayash, R.; Kutteh, R.; Lin, Z.;
Littlefield, R.; Long, X.; Meng, B.; Nieplocha, J.; Niu, S.; Rosing, M.;
Sandrone, G.; Stave, M.; Taylor, H.; Thomas, G.; van Lenthe, J.; Wolinski,
K.; Wong, A.; Zhang, Z. NWChem, A Computational Chemistry Package
for Parallel Computers, version 4.0; Pacific Northwest National Labora-
tory: Richland, WA, 2000.
References and Notes
(
1) Davis, M. E.; Lobo, R. F. Chem. Mater. 1992, 4, 756.
(2) Petrovic, I.; Navrotsky, A.; Davis, M. E.; Zones, S. I. Chem. Mater.
1
2
(
35) Sastre, G.; Lewis, D. W.; Catlow, C. R. A. J. Phys. Chem. 1996,
1
00, 6722.
(
(
(
(
6) Barrer, R. M. Zeolites 1981, 1, 130.
(
01.
36) Sastre, G.; Fornes, V.; Corma, A. J. Phys. Chem. B 2002, 106,
7) Barrer, R. M. Stud. Surf. Sci. Catal. 1985, 24, 1.
8) Lok, B. M.; Cannan, T. R.; Messina, C. A. Zeolites 1983, 3, 282.
9) Davis, M. E. Acc. Chem. Res. 1993, 26, 111.
7
(
(
37) Rappe, A. K.; Goddard, W. A., III. J. Phys. Chem. 1995, 95, 3358.
38) Piccione, P. M.; Laberty, C.; Yang, S.; Camblor, M. A.; Navrotsky,
(
10) Hong, S. B.; Cho, H. M.; Davis, M. E. J. Phys. Chem. 1993, 97,
622.
11) Corma, A.; Diaz-Caba n˜ as, M. J.; Fornes, V. Angew. Chem., Int.
Ed. 2000, 39, 2346.
12) Lewis, D. W.; Freeman, C. F.; Catlow, C. R. A. J. Phys. Chem. B
995, 99, 11194.
A.; Davis, M. E. J. Phys. Chem. B 2000, 104, 10001.
1
(39) Davis, M. E. CATTECH 1997, 1, 19.
(
(40) Akporiaye, D. E.; Price, G. D. Zeolites 1989, 9, 321.
(41) Kramer, G. J.; de Man, A. J. M.; van Santen, R. A. J. Am. Chem.
Soc. 1991, 113, 6435.
(
1