Solid-state deuterium NMR studies of organic molecules in the tectosilicate nonasil

Article abstract of DOI:10.1021/jp9731462

Solid-state deuterium NMR spectroscopy is used to study the dynamics of organic molecules occluded in the as-synthesized high-silica tectosilicate nonasil. The nonasil samples are synthesized using trimethylalkyl-ammonium structure-directing agents to det

Full text of DOI:10.1021/jp9731462

J. Phys. Chem. B 1998, 102, 2339-2349
2339
Solid-State Deuterium NMR Studies of Organic Molecules in the Tectosilicate Nonasil
Daniel F. Shantz and Raul F. Lobo*
Center for Catalytic Science and Technology, Department of Chemical Engineering, UniVersity of Delaware,
Newark, Delaware 19716
ReceiVed: September 29, 1997; In Final Form: January 14, 1998
Solid-state deuterium NMR spectroscopy is used to study the dynamics of organic molecules occluded in the
as-synthesized high-silica tectosilicate nonasil. The nonasil samples are synthesized using trimethylalkyl-
ammonium structure-directing agents to determine the role of electrostatic interactions. Size effects are
2
quantified by performing H NMR spin-lattice (T
1
) relaxation experiments, and the mobility of the substituent
2
alkyl groups is studied using H MAS NMR. The charge-compensating defect sites are characterized using
29
1
Si and H NMR. The motion of the trimethylammonium group of the structure-directing agent is a composite
motion of methyl group rotations and rotation about the nitrogen C axis in all samples down to 190 K. The
and H MAS NMR results illustrate the steric confinement the nonasil cage exerts on the larger (C
) substituent alkyl groups. Isotropic motion is not observed for any of the structure-directing agents at 370
3
2
2
H T
1
n
g
C
5
K, indicating strong organic-inorganic interactions. This is in sharp contrast to nonasil samples made with
electrically neutral amines where rapid isotropic reorientation is observed at room temperature. These results
have interesting implications in zeolite synthesis. For aluminosilicates synthesized with charged structure-
directing agents, similar organic-inorganic interactions may allow for aluminum preferentially occupying
specific framework sites. This could lead to tailoring the distribution of catalytic sites in zeolites based on
the charge distribution of the structure-directing agent.
directing agent.7 It is also believed that electrostatic interac-
-9
Introduction
tions play an important role in this molecular recognition
process.
Before a high-silica zeolite can be used in catalytic, adsorp-
tion, or ion-exchange applications, the structure-directing agent
needs to be removed by heating at high temperatures in the
presence of oxygen (calcination). There is an extensive body
of work on the characterization of zeolites and related materials
Electrostatic forces play a fundamental role in many physical
and chemical systems, including protein structure and stability,
ionic conduction, the properties of many colloidal systems, self-
1
assembly, etc. The important role of electrostatic forces in
the determination of the catalytic, adsorptive, and other proper-
ties of zeolites and other microporous systems has long been
2
recognized. Zeolites are crystalline microporous materials used
(
after calcination) using diffraction, microscopy, and spectro-
industrially as heterogeneous catalysts and as adsorbents in gas
10
scopic techniques. On the other hand, relatively little char-
acterization work has been done of the organic-inorganic
composite material obtained directly from synthesis. The
objective of this work is to investigate as-made zeolites and to
characterize the interactions between the organic and inorganic
components. Electrostatic interactions are of particular interest
as they are not well understood, yet have important implications
in not only the physical properties of zeolites, but also in the
synthesis of zeolites and the process of structure-direction.
3
separations and purifications. The motion of organic molecules
in zeolites is complex because the pores of a zeolite are nearly
4
the same size as the organic molecules occluded in them. The
mobility of an organic molecule is largely determined by the
forces between an organic molecule and the zeolite framework,
and strongly influences macroscopic properties such as the
diffusivity and catalytic activity. These interactions have
received considerable attention in recent years and remain a
subject of active research.5
The tectosilicate nonasil (NON), originally synthesized by
11
Synthetic high-silica zeolites are prepared using an organic
molecule, usually a quaternary ammonium cation, in the
synthesis mixture. The organic molecule is often called a
structure-directing agent (SDA) because the shape and size of
the molecule strongly influence the structure of the final
synthesis product. Usually a close relationship is observed
between the size and shape of the organic molecule, and the
Marler et al., is used as a model system for this work. Nonasil
is a clathrasil, a class of natural and synthetic materials with
cages, in contrast to zeolites which have one- or multidimen-
sional open pore structures. The framework of nonasil can be
constructed from 9-hedra bounded by one four-membered ring
1
8
and eight five-membered rings [4 5 unit]. The nonasil material
1
8
has three cages formed by (1) the 4 5 unit, (2) an 8-hedra
formed by four five-membered rings and four six-membered
6
size and shape of the zeolite pores and pore intersections. ZSM-
4
4
rings [5 6 unit], and (3) a 20-hedra formed by eight five-
5
, ZSM-18, and SSZ-26 are three examples where the zeolite
8
12
membered rings and twelve six-membered rings [5 6 unit].
The three cages have free volumes of approximately 30, 25,
structure clearly reflects the geometry and size of the structure-
3
and 290 Å , respectively, and the organic structure-directing
*
Corresponding author: Center for Catalytic Science and Technology,
8
12 11
agents are only found in the largest cage [5 6 ]. There are
Department of Chemical Engineering, University of Delaware, Newark, DE.
Phone: (302) 831-1261. Fax: (302) 831-2085. E-mail: lobo@che.Udel.edu.
8
12
four 5 6 cages per unit cell (see Figure 1) and the free
S1089-5647(97)03146-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/11/1998

2340 J. Phys. Chem. B, Vol. 102, No. 13, 1998
Shantz and Lobo
Figure 2. Organic molecules used as structure-directing agents for
the synthesis of nonasil.
2
3
geometry of molecular motion. Rotations that are fast on the
2
H NMR time scale result in a narrowing of the signal (Pake
doublet) as compared to the static value of approximately 120
kHz. For motion with axial symmetry (i.e., η ) 0) (the
asymmetry parameter η is defined as η ) (Vxx - Vyy)/Vzz, where
Vxx, Vyy, and Vzz are the diagonal elements of the electric field
gradient (EFG) tensor, and for motions with a 3-fold or higher
2
4
axis of symmetry, typically η ) 0), the frequencies of the
horns of the Pake doublet are determined by
8
12
Figure 1. Large [5 6 ] cage of nonasil.
2
3
2πe qQ
2
ν ) ( /
(3 cos θ - 1)
8
(
)
h
dimensions of the large cage are approximately 9.8 × 11.6 ×
9
.8 Å. Marler et al.11 determined the structure using single-
2
where (2πe qQ/h) is the static quadrupole coupling constant
crystal X-ray diffraction refining the structure in the ortho-
rhombic space group Fmmm with unit cell lattice constants ao
(
∼2π × 170 kHz), h is Planck’s constant, Q is the nuclear
quadrupole moment, eq ) Vzz is the principal component of
the EFG tensor, and θ is the angle between the rotation axis
and the magnetic field in the laboratory frame. Motions that
)
22.232 Å, bo ) 15.058 Å, and co ) 13.627 Å. The space
group Fmmm is the highest topological symmetry for the NON
topology. Gies and Marler used electrically neutral molecules
such as 2-methylpyrrolidine, 2-aminopentane, and 2-methyl-
piperidine to synthesize the nonasil samples and found that the
structure-directing agents were positionally disordered in the
as-made material. If boron is present in the synthesis mixture,
2
are fast on the H NMR time scale can be quantified using spin-
lattice (T1) relaxation experiments in conjunction with NMR
Several groups have used H NMR
to study the interactions between aromatic molecules and
extraframework cations in faujasite-type zeolites.
studies showed a strong interaction between the extraframework
cations and the π-system of the aromatic molecule (i.e., energies
2
4-26
2
line shape simulations.
1
2
27
nonasil crystallizes in the orthorhombic space group Cmca.
These
Harris and Zones13 synthesized nonasil using several trimeth-
ylalkylammonium compounds as structure-directing agents and
performed energy minimizations using molecular mechanics
calculations. Balkus and Shepelev14 have synthesized nonasil
2
of activation of ∼25 kJ/mol were observed). H NMR has been
used extensively to study the dynamics of organics in inclusion
+
-
28
using bis(cyclopentadienyl)cobalt(III) hydroxide ([Cp2Co] OH ),
and Behrens et al.15 synthesized nonasil with [Cp2Co] but
prepared by the fluoride method.14 The structure of fluoride
Cp2Co(III)-NON was refined in the orthorhombic space group
Pccn.16 Their single-crystal X-ray work shows that the met-
allocene complex is ordered in the nonasil cage and forms
hydrogen bonds between the Cp hydrogens and the oxygen
atoms in the nonasil framework. In contrast, Gies17 has shown
that electrically neutral amines used to synthesize nonasil
isotropically reorient on the microsecond time scale.
compounds, although little work has been done with as-
+
29
synthesized zeolites.
This work investigates the dynamics of charged structure-
directing agents occluded inside the large [5 6 ] cage of nonasil
Figure 1). Electrically neutral amines used to synthesize
nonasil and other clathrasils isotropically reorient and exhibit
only weak interactions between the organic molecules and the
zeolite framework. The results for the electrically neutral
8
12
(
17
2
amines are similar to many H NMR studies of organic
molecules in inclusion compounds where steric constraints and
van der Waals interactions dominate molecular motion.28 In
contrast, we are employing cationic structure-directing agents
in the synthesis of nonasil and our results should point out the
effects of electrostatic forces. There are four advantages to using
nonasil as a model system: (1) interactions between structure-
directing agents are negligible, (2) nonasil is an all-silica
material so there are no extraframework cations, (3) the same
phase can be prepared with organic molecules of different size,
and (4) since the structure of nonasil is related to the structure
of other high-silica zeolites, our results can be qualitatively
extended to other microporous silicates.
Solid-state NMR is a powerful tool for probing the local
structure of zeolites
1
0d,18
18,19
and solid materials in general.
2
9
27
Si and Al NMR are useful to investigate the local structure
of the framework atoms,1
0d,20
and Na, Li, and Cs NMR
23
7
133
compliment diffraction techniques for characterizing the loca-
tion, site populations, and mobility of extraframework cations
21 2
in zeolites.
H NMR can be used to obtain detailed informa-
3
0
tion about the dynamics of organic molecules occluded inside
the zeolite micropores.22 Deuterium has a nuclear spin I ) 1,
a static quadrupole coupling constant between 150 and 200 kHz,
and it is an experimentally convenient quadrupolar nuclei
because the quadrupolar interaction dominates dipole-dipole
coupling and chemical shift effects. Another advantage is that
the quadrupole interaction is small enough that the whole
powder spectrum can be observed with high-power pulsed NMR
Samples of nonasil synthesized with molecules I-IV (see
Figure 2) have been characterized using X-ray diffraction
(XRD), scanning electron microscopy (SEM), and nuclear
magnetic resonance spectroscopy (NMR). Solid-state deuterium
NMR is used to study the dynamics of the organic molecules
occluded in the nonasil samples. Different parts of the organic
molecules are selectively labeled with deuterium. Our results
2
3
2
methods. The main limitation of H NMR spectroscopy is
that it requires the use of isotopically enriched samples. The
value of the quadrupole coupling constant (QCC), an experi-
mentally measurable quantity, is determined by the rate and

Organic Molecules in the Tectosilicate Nonasil
J. Phys. Chem. B, Vol. 102, No. 13, 1998 2341
TABLE 1: Nonasil Synthesis Conditions
molar ratios
indicate that, for all structure-directing agents used, the trimeth-
ylammonium segment of the molecule is rapidly rotating along
the R-N and N-C bonds and that the large substituent alkyl
groups (Cn g C5) are nearly immobile in the nonasil cage. The
lack of isotropic reorientation in all samples up to 370 K
indicates that electrostatic forces between the framework and
the structure-directing agent prevent the molecule from isotropic
motion, in contrast to nonasil made with electrically neutral
synthesis
a
sample
in the synthesis gel
temperature, K
+
-
TMC3-NON
TMC5-NON
TMC6-NON
TMC7-NON
1 SiO
2
:0.15 R OH :0.125
433
433
443
433
NaOH:45 H
1 SiO :0.15 R OH :0.10
2
O
+
-
2
2
NaOH:45 H O
+
-
1 SiO
2
:0.15R OH :0.125
1
7
NaOH:45 H O
2
amines.
+
-
1 SiO
2
:0.15 R OH :0.15
NaOH:45 H
2
O
Experimental Section
a
TMC3 (I): N,N,N-trimethylisopropylammonium hydroxide. TMC5
(II): N,N,N-trimethylcyclopentylammonium hydroxide. TMC6 (III):
N,N,N-trimethylcyclohexylammonium hydroxide. TMC7 (IV): N,N,N-
trimethylnorbornylammonium hydroxide.
Sample Preparation. Nonasil samples are synthesized using
a modification of the method reported by Harris and Zones
1
3
with the four trimethylalkylammonium compounds I-IV (Fig-
ure 2) as structure-directing agents. The four structure-directing
agents used are N,N,N-trimethylisopropylammonium (TMC3),
N,N,N-trimethylcyclopentylammonium (TMC5), N,N,N-trimeth-
ylcyclohexylammonium (TMC6), and N,N,N-trimethylnorborn-
ylammonium (TMC7).
Synthesis of N,N,N-trimethylcyclopentylammonium Io-
dide. TMC5 is prepared by adding 100 mmol of cyclopentyl-
amine (99% Aldrich) and 20 g of anhydrous K2CO3 (85+%
Fisher) to 100 mL of methanol (99% ACS reagent, Fisher). The
amine is alkylated with 50% excess methyl iodide (99%
Aldrich), and the solution is allowed to react for 24 h in the
absence of light at room temperature. The liquid phase is
separated from the solids by filtration. Solvent and other volatile
components are removed under vacuum at 50-60 °C using a
rotavapor yielding the crude TMC5 salt (∼85% yield) as a white
powder. The crude template is recrystallized from a boiling
a syringe. The BF3OEt2 is added dropwise to the flask while
immersed in the ice bath. The mixture is allowed to react for
3 h, after which 14.4 g of hydroxylamine-o-sulfonic acid (99%,
Aldrich) dissolved in 50 mL diglyme are added to the mixture
dropwise. The mixture is heated to 100 °C and allowed to react
for 3 h. The resulting solution is cooled and made acidic (pH
<
2) with concentrated HCl. The resultant mixture is extracted
three times with ether to remove any residual boric acid, and
the organic phase is discarded. The aqueous phase is made
strongly basic (pH > 12) using NaOH and is again extracted
three times with ether. The ether is removed at reduced
pressure, and the crude solid recovered is alkylated as described
previously. The yield from the reaction is approximately 50%.
The selective deuterium-labeling of the rings was verified by
1
3
C solution NMR (see Supporting Information).
7
5
0:30 ethyl acetate/methanol mixture (V/V) which is cooled to
°C and left overnight (∼40% yield). Three grams of the
Nonasil Synthesis. The nonasil synthesis gel stoichiometry
varies slightly depending on the structure-directing agent used
(see Table 1). The synthesis gel molar ratios are 1 SiO2:0.15
recrystallized TMC5 salt are ion-exchanged to the hydroxide
form using a 5-fold excess of resin (IONAC NA-38, OH form)
-
+
-
+
-
R OH :0.125 NaOH:45 H2O, where R OH is the structure-
directing agent exchanged to the hydroxide form. The main
difference in the synthesis conditions is the adjustment of the
amount of sodium.
for 24 h. The conversion of the iodide to hydroxide is greater
than 95% based on titration of the resultant solution. The other
structure-directing agents can be made using the same procedure
with similar yields. The 13C NMR results for the recrystallized
iodide salts in D2O are TMC3-iodide δ ) 67.5, 51.2, 17.1,
For example, the TMC5-nonasil sample is prepared by
mixing 2.4 g of Cab-O-Sil (M5 grade), 20 mL of 0.3 N N,N,N-
trimethylcyclopentylammonium hydroxide, and 13.4 mL of
deionized water. After a homogeneous mixture is obtained, 0.16
g of NaOH (pellets 95+%, Aldrich) are ground into a fine
powder and added to the solution. The solution is mixed for
approximately 1 h, until a liquidlike gel is obtained. The
solution is sealed in two 23 mL Teflon-lined Parr autoclaves
which are heated at 160 °C for 7 days under autogenous pressure
and rotated at 30 rpm. The autoclaves are cooled to room
temperature after the synthesis period. The solids are collected
by filtration, washed several times with deionized water, and
dried overnight at 80 °C. Seeding was used only for the
synthesis of TMC3-NON which was seeded with 1 wt %
TMC5-NON crystals. None of the samples used in the
deuterium NMR studies were seeded.
1
5.1 ppm; TMC5-iodide δ ) 71.2, 51.9, 28.0, 24.2 ppm;
TMC6-iodide δ ) 74.4, 51.6, 26.6, 24.9, 24.5 ppm; TMC7-
iodide δ ) 78.9, 51.8, 37.8, 37.0, 36.1, 34.7, 28.9, 26.4 ppm.
Preparation of Deuterium-Labeled Structure-Directing
Agents. Synthesis of structure-directing agents selectively
labeled with deuterium in the methyl groups is carried out as
described above except d3-CD3I (99%+d Aldrich) is used
instead of CH3I in the alkylation step. A procedure described
3
1
by Brown et al. is used to synthesize ring-deuterated primary
amines (II-IV) with deuterium â to the amine group. The
overall reaction is
Samples are also prepared using D2O (99.9%-d, Aldrich) as
the solvent. For these samples, ion-exchange, titration of the
exchanged structure-directing agent, and the nonasil synthesis
mixture preparation are performed in a glovebag under a dry
argon atmosphere.
Analytical. Powder X-ray diffraction (XRD) is performed
on samples using a Phillips 3000 X’Pert System with Cu KR
radiation. The patterns used for indexing are collected from
4°-50° 2θ using the step scan mode, a 0.02° step size, and 15
s count time per step. The peak deconvolution was carried out
using the Phillips package ProFit and refinement of the unit
Synthesis of d1-Cyclopentylamine. A flame-dried 250 mL
round-bottom flask is fitted with a dried condensor and dropping
funnel and purged using dry argon. 60 mL of diglyme (99%,
Fisher) are added to the flask and 2 g of NaBD4 (99%-d,
Aldrich) are added to the diglyme. After 15 min of mixing,
1
0.2 mL of cyclopentene (99+%, Aldrich) are added via a
syringe. The flask is then set in an ice bath and 8.1 mL of
BF3OEt2 (99%, Aldrich) are added to the dropping funnel using

2342 J. Phys. Chem. B, Vol. 102, No. 13, 1998
Shantz and Lobo
cell parameters was done using the program Lapod.32 Scanning
electron microscopy was done using a JEOL 850 at 10 kV.
1
3
Solution C NMR spectra were measured on a Bruker AC 250
spectrometer at 62.26 MHz with the chemical shifts referenced
2
9
13
1
2
to tetramethylsilane. Si, C, H, and H solid-state NMR
spectra were measured on a Bruker MSL 300 spectrometer at
5
9.63, 75.47, 300.13, and 46.07 MHz, respectively. The
29
13
1
chemical shifts for the Si, C, and H spectra were referenced
to tetramethylsilane. The 2 Si MAS NMR spectra were acquired
using a 7 mm probe with ZrO2 rotors, a spin rate of 3 kHz, and
a 100 s recycle delay to avoid relaxation effects in the signal
intensities. The structural integrity of the structure-directing
9
1
3
agents inside the as-made nonasil samples was verified by
C
CP-MAS NMR using a contact time of 5 ms, a spin rate of 3
1
kHz, a 90° H pulse of 6 µs, and high-power proton decoupling.
1
H MAS spectra were acquired using a 4 mm probe with ZrO2
rotors, a spin rate of 9 kHz, a 30 s recycle delay, and a 4 µs
6
0° pulse.
The static H NMR experiments were performed using the
usual quadrupole echo pulse sequence33 with a 90° pulse of
.25 µs unless otherwise noted, a pulse spacing of 30 µs, and
2
3
recycle delays between 1 and 3 s. Variable pulse spacing
experiments were performed at room temperature to check for
distortions in the quadrupole echo.34 Variable-temperature
experiments were performed from the low to the high-temper-
ature range of the probe (190 to 370 K). The temperature was
controlled with a Bruker B-1000 temperature controller to within
Figure 3. X-ray powder diffraction patterns of nonasil samples.
TABLE 2: Unit Cell Lattice Constants for Nonasil Samples
(1K with an accuracy of (2K. Temperature calibration for
structure-directing
the probe was done by placing a thermocouple inside the sample
chamber and comparing the results to a calibrated thermocouple
located inside the probe, just outside the sample chamber, over
the temperature range of 190-370 K. Spin-lattice (T1)
agent
a (Å) B (Å) c (Å)
22.08 15.03 13.74 90.7° 90.1° 90.2°
22.15 15.14 13.69 90° 90° 90°
23.26 15.04 13.55 90.71° 90.11° 90.17°
R
â
γ
TMC3-NON
TMC5-NON
TMC6-NON
TMC7-NON
relaxation experiments were performed using a 6.5 µs 180° pulse
22.00 15.27 13.67 90°
90°
90°
90°
90°
2-aminopentane11 22.23 15.05 13.62 90°
and the phase cycling suggested by Vold.25 2H magic-angle
3
5
spinning (MAS) NMR, as described by Jakobsen et al., was
used for the analysis of samples selectively labeled with
deuterium on the rings â to the trimethylammonium group due
for all samples show a nearly spherical morphology. The crystal
size depends on which structure-directing agent is used, but in
all samples, smaller crystals (∼1 µm) with the same morphology
are observed on the surface of the larger crystals. The TMC5
structure-directing agent forms the largest crystals (30-40 µm),
and the TMC3 structure-directing agent forms the smallest
crystals (1-8 µm). As it can be observed in the XRD patterns
(Figure 3) some of the samples show very sharp peaks (TMC5-
NON) while others show broad peaks (TMC3-NON). These
differences are not due to crystal size effects as the crystals in
all cases are larger than 1 µm. Inspection of the TMC6-NON
XRD pattern shows the splitting of the 111 reflection into 4
distinct peaks, (111, 1h 11, 1 1h 1, and 11 1h ) suggesting that the
actual symmetry of the material is triclinic and not orthorhombic.
For this reason the TMC3-NON and TMC6-NON samples
have been indexed using a triclinic unit cell. The TMC7-NON
is also of lower symmetry, but due to extensive peak overlap,
it has been indexed using a pseudo-orthorhombic unit cell. Note
that due to the strong peak overlap the unit cell parameters of
TMC3-NON and TMC7-NON are only approximate. The
refined unit cell lattice constants for the samples are tabulated
in Table 2 and are in good agreement with previously reported
2
to signal-to-noise improvements. The H single-pulse MAS
NMR spectra were obtained using a 4 mm MAS probe, a 45°
pulse of 4 µs, a recycle delay of 4 s, and a spin rate (νr) of 6
2
9
2
kHz, unless noted. Si NMR line shape simulations and H
NMR line shape fitting were done using software written in-
2
house. H NMR line shape simulations using different motion
models were done using TurboPowder and FastPowder.
2
6
Elemental analysis for H, C, N, Si, and Na was done by
Galbraith Laboratories (Knoxville, TN) and Chevron R & D
(Richmond, CA). Molecular dynamics simulations and energy
minimizations were performed using the Biosym-MSI package
Cerius 3.2. The Burchart-Universal force field was used in all
simulations. The molecular dynamics simulations were ran for
a 10 ns long trajectory with a 0.01 ps time step at 300 K, using
the NVT ensemble and a Nos e´ thermostat. All simulations used
a rigid, defect-free framework, and the initial conformation of
the structure-directing agent used was found using the energy
minimization routine.
Results and Discussion
11
values. The changes in symmetry are not surprising, as several
zeolitic systems exhibit changes in symmetry either due to a
change of temperature or the presence of adsorbates.1 As
noted in the Introduction, the symmetry of nonasil samples
depends heavily on the structure-directing agent.
Material Characterization. The X-ray diffraction patterns
for the nonasil samples prepared with molecules I, II, III, and
IV are shown in Figure 3. The as-made materials are highly
crystalline nonasil, no impurities or diffuse scattering back-
ground are detected by either XRD or SEM. Scanning electron
micrographs (see Supporting Information) indicate that most
of the crystallites are greater than 1 µm in size. The micrographs
0a
Nonasil has not been successfully synthesized in the absence
of a structure-directing agent. The large [5 6 ] cage of nonasil
must be occupied by a molecule to help stabilize the structure.
8
12

Organic Molecules in the Tectosilicate Nonasil
J. Phys. Chem. B, Vol. 102, No. 13, 1998 2343
Figure 5. 1H NMR spectrum of TMC5-NON. (Asterisks denote
spinning sidebands).
for samples made with cationic structure-directing agents than
samples made with neutral amines. Moreover, additional silanol
groups are present in excess of the siloxy and silanol groups
needed to balance the cationic charge on the structure-directing
Figure 4. 29Si MAS NMR spectra of as-made nonasil samples.
TABLE 3: Summary of 29Si MAS NMR Spectra
Simulations
3
4
agent. The integrated intensities of the spectra result in Q /Q
values of 0.23, 0.23, 0.20, and 0.22 for the TMC3-NON,
TMC5-NON, TMC6-NON, and TMC7-NON samples, re-
Q3
Q4
3
structure- chemical
chemical
no. of
spectively, corresponding to approximately four Q sites per
3
8 12
directing shift, δ
relative
shift, δ
relative
Q per
cage (there are 22 Si atoms per [5 6 ] cage). The results of
fitting the spectra are given in Table 3. Note that number of
3
4
agent
(ppm) intensity, I (ppm) intensity, I Q /Q cage
n
TMC3 -102.7
107.7
TMC5 -104.6
TMC6 -104.7
15.6
2.8
18.6
14.9
2.1
13.6
4.5
-116.3
81.6
0.23
4.1
lines used to fit each Q line (1 or 2) was chosen to minimize
-
the error between the experimental and simulated line shapes
-117.3
-116.9
-120.3
-116.4
-119.9
81.4
35.8
47.2
26.1
55.8
0.23
0.20
4.1
3.7
(i.e., they do not represent crystallographically inequivalent
sites). These results are in good agreement with the work of
-
108.6
TMC7 -102.4
3
7
3
+
Koller et al. which reports a Q /N ∼ 4 for several all-silica
0.22
4.0
2
1
materials. Our nonasil samples have no Q or Q groups, so
the structure-directing agents are confined to one cage. The
structure-directing agents cannot move through the six-
membered rings of the nonasil framework; they need larger
apertures in order to move between cages. Only framework
sites with more than one of their Si-O-Si bonds broken could
facilitate intercage mobility.
-
106.6
The large cage could be occupied by the structure-directing
agent, byproducts of structure-directing agent degradation, or
13
sodium. The C CP-MAS spectra unambiguously confirm that
the structure-directing agents are all intact inside the nonasil
cages (see Supporting Information). The elemental analyses
indicate that at least 80-90% of the cages are occupied by an
organic structure-directing agent.36 There is also less than 1
wt % sodium in the as-made materials. The small amount of
sodium present is consistent with previous results for as-made
To investigate the effect of the small amount of sodium (<1wt
%) occluded in all of the nonasil samples, we prepared a sample
of TMC5-NON without sodium in the synthesis mixture. The
3
4
29
Q /Q value calculated from the Si NMR spectrum of that
+
3
high-silica zeolites, where also a small amount of Na was found
sample also corresponds to four Q per cage, therefore the
5
c
occluded in the final product.
sodium ions do not appreciably contribute to the quantity of
defect sites in our samples.
The 29Si MAS NMR spectra of the as-made materials are
3
shown in Figure 4. We have found that consistent with previous
reports,37 nonasil samples prepared using cationic structure-
directing agents contain a large number of silanol groups. Each
spectrum contains two broad lines, one line is in agreement with
The spatial proximity of the Q groups cannot be determined
29
from the Si NMR results, however, this can be studied using
1
1
H NMR. The H NMR spectrum of the TMC5-NON sample
is shown in Figure 5. Two lines are present in the spectra for
all the as-made materials, a broad line at 3 ppm assigned to the
protons on the organic structure-directing agents and a line at
10.3 ppm. The broad line at 3 ppm is asymmetric and also has
a shoulder at 4.5 ppm due to a small amount of water present
3
silicon atoms adjacent to one siloxy or silanol group (Q , where
n
- 10d
Q stands for X4-nSi[OSi]n, X ) OH or O ) and the second
line is consistent with silicon atoms bonded to four silicon atoms
via an oxygen atom (Q ). Note that contributions to the Q
4
3
37
intensity from surface silanols are negligible due the to crystal
sizes; we have calculated that less than 0.2% of the silicon atoms
would be surface silanol groups for a defect-free 5 × 5 × 5
in the sample. Koller et al. observed a line at 10.2 ppm ((0.2
ppm) for a series of as-made high-silica zeolites, including a
nonasil sample made with the TMC7 structure-directing agent.
Their work concluded that the line at 10.2 ppm was due to
silanol groups engaged in hydrogen bonding with siloxy groups
3
4
-5
µm crystal. This results in a Q /Q of 2 × 10 .
3
4
The ratio of integrated intensity of the Q /Q lines is directly
related to the number of silanol groups in the as-made material
-
with an average Si-OH‚‚‚ O-Si distance of ∼2.6 Å. Using
4
0
(see Table 3). Most all-silica zeolites are made with cationic
known chemical shift correlations our results suggest an
-
structure-directing agents and charge neutrality requirements
result in the formation of a charge-compensating defect site.
The quantity and nature of these defects, which can strongly
average OH‚‚‚ O-Si distance of ∼2.7 Å, in agreement with
3
8
37
1
Koller et al. within experimental error. The H NMR spectra
3
verify that some of the Q sites in the framework are in close
3
9
affect the physical properties of zeolites, are as yet unclear.
proximity, either as isolated pairs or in clusters.
3
7
Koller et al. have shown that a siloxy group is the charge
compensator for the cationic structure-directing agent and not
Deuterium NMR. The deuterium NMR spectroscopy work
2
consists of four parts. First, the H NMR spectra of the neat
6d
the hydroxide anion. Lobo et al. have observed that the
number of defects in the as-made nonasil material is much larger
solid salts with deuterium at the methyl groups are described.
2
Second, H NMR spectroscopy results of the nonasil samples

2344 J. Phys. Chem. B, Vol. 102, No. 13, 1998
Shantz and Lobo
TABLE 4: Quadrupole Coupling Constant, Asymmetry
Parameter, and Spin-Lattice Relaxation Time (T
-Labeled TMC7-, TMC6-, TMC5-, and
TMC3-Nonasils, and Neat Salts at 298 K
1
) for
d
9
quadrupole coupling
constant (kHz)
asymmetry
parameter, η
sample
T
1
(ms)
TMC7-NON
neat TMC7
TMC6-NON
neat TMC6
TMC5-NON
neat TMC5
TMC3-NON
neat TMC3
15.5a
∼15.6
15.2
13.9
15.1
15.0
13.5
∼0
0.05a
N/A
0.06
0.06
0.05
0.05
0.14
N/A
18.9
21.7
24.7
45.2
a
The motion of the TMC7 solid is in the intermediate motion regime
so the QCC can only be estimated.
2
Figure 6. Static H NMR spectra for d
9
-nonasil samples at 298 K.
There are several possibilities that could explain the decrease
in the QCC value (∼15 kHz) from the expected value (∼19
kHz) if only two rotations are present. One possible reason
for this reduction in the averaged QCC could be the presence
of additional motions. Specfically, the reduced QCC values
can be accounted for by allowing an outward bending of the
D-C-N bond angles on average of 4°. The breathing motion
also introduces asymmetry into the line shape as the breathing
motion has a pseudo 2-fold symmetry axis. The simulated NMR
spectrum assuming an average bending of four degrees results
in an averaged QCC of 15.1 kHz, consistent with the d9-TMC5
and d9-TMC6 nonasil results (see Supporting Information). The
d9-TMC7 spectrum can be simluated accurately assuming a
bending motion on average of 3°.
Another possible explanation of the observed reductions in
the QCCs is the presence of librational motion. The decrease
of the QCCs as the SDA gets smaller are in agreement with
this conclusion. Clearly, for the TMC3-NON, the breathing
motion is not sufficient to account for the obtained QCC.
However, while the breathing motion is straightforward to
simulate, the librational motion is not. The NMR results
obtained cannot rule out one possibility over the other, and both
could in fact be present. This is almost certainly the case for
the TMC3-NON sample.
made with structure-directing agents with labeled methyl groups
are discussed. Third, we discuss H spin-lattice (T1) relaxation
experiments done in conjunction with line shape simulations
2
for the d9-labeled samples to complement the variable temper-
2
ature experiments. Finally, the H MAS NMR results for
samples with labeled alkyl groups will be discussed. The
characterization work has verified that the samples are all highly
crystalline nonasil with the organic structure-directing agents
8
12
intact inside the [5 6 ] cage. These results do not however,
imply long-range ordering of the structure-directing agents or
the defect sites. Disorder of the defect sites or the structure-
directing agents could result in molecules having a distribution
of mobilities. Our results, however, seem to indicate that at
2
least for the methyl groups the motion is uniform. All H NMR
line shape simulations were performed using the programs
FastPowder and TurboPowder.26
The motion of the structure-directing agents in the as-made
2
nonasil samples can be well characterized using H NMR.
2
Recall that deuterons whose motions are slow on the H NMR
-
7
time scale (∼10 s) result in a splitting between the horns of
the Pake doublet of roughly 120 kHz. Rapid internal reorienta-
tion of the methyl groups but no rotations along the nitrogen
19
C3 axis reduces the splitting to ∼40 kHz. Rapid rotations
along both the nitrogen C3 axis and the methyl group C3 axes
In addition to molecular motion, the reduction in the splitting
of the Pake doublet could be due to large deviations from
tetrahedral geometry in either the R-N-C bond angle, the
N-C-D bond angle, or both. Distortions of the electric field
gradient around the deuteron could also result in a lowering of
the QCC value. Large deviations from tetrahedral geometry
seem unlikely as both the methyl group rotations and the R-N
bond rotation are fast on the NMR time scale. Distortions in
the electric field gradient could be possible due to the cationic
nature of the structure-directing agents, but would be very
difficult to quantify. The case of a large nonzero asymmetry
(η ) 0.2) for an axially symmetric motion was recently reported
22b
reduce the splitting to about 14 kHz (QCC ) 18.8 kHz). Also,
since all of the above motions should be axially symmetric, the
asymmetry parameter η should be zero.
2
H NMR Spectra of the Structure-Directing Agents in
2
Nonasil. The H NMR spectra of the d9-nonasil samples at
2
-
1
98 K are shown in Figure 6. Simulations of the d9-TMC5,
TMC6, and -TMC7 nonasil spectra at 298 K yield QCCs of
5.1, 15.2, and 15.5 kHz and asymmetry values η of 0.05, 0.06,
and 0.06, respectively. These values are similar to the neat
solids (summarized in Table 4), with the exception of the TMC3
salt whose spectrum is a Lorenztian line with a full-width at
half-maximum of 456 Hz (see Supporting Information). The
simulated line shape of the d9-TMC3 nonasil sample yields a
QCC of 13.5 kHz and an asymmetry value of 0.14, which are
different from the other samples and are discussed in detail
below. The motions of the occluded organics consist at least
of fast rotations along both the nitrogen and methyl group C3
axes. The values of the QCCs are lower than would be expected
42
by Griffin and co-workers on [W(Cp*Me4)][PF6]. Their work
suggests that deviations from tetrahedral geometry were pri-
marily responsible for the observed asymmetry, and distortions
of the electric field gradient (EFG) could be present, but were
less significant. This conclusion was drawn based on the
structure of the compound determined from crystallographic
work.
2
2b,41
for only two rotations, even with the nonzero asymmetry.
If an additional motion is present, it is likely a breathing
motion of the methyl groups, a librational motion of the SDA,
or both. Either of the motions would also introduce asymmetry
into the line shape which is observed for all samples. The
nonzero asymmetry is unusual since motions that can be
modeled as rotations about a 3-fold axis (i.e., fast rotation of
The static quadrupole coupling constant could not be obtained
for any of the samples because neither of the rotations were in
the slow motion regime at 190 K; consequently, the following
analysis will assume the static QCC is approximately 170
kHz.1
9,24,41

Organic Molecules in the Tectosilicate Nonasil
J. Phys. Chem. B, Vol. 102, No. 13, 1998 2345
(a)
(b)
(c)
(d)
(e)
2
2
Figure 7. (a) Variable temperature H NMR spectra for d
9
-TMC5 nonasil. (b) Variable temperature H NMR spectra for d
9
-TMC6 nonasil. (c)
2
2
2
Variable temperature H NMR spectra for d
NMR spectra for d -TMC3 nonasil.
9 9
-TMC7 nonasil. (d) Low-temperature H NMR spectra for d -TMC3 nonasil. (e) High-temperature H
9
deuterons about the C3 axis of a methyl group) usually exhibit
no asymmetry.24 Asymmetry in the H NMR line shape can
arise from motions with a 2-fold symmetry axis, such as flipping
of a D2O molecule about its C2 axis.34 We believe that the
reduction in the observed QCCs and the presence of asymmetry
in of our samples are most likely due to an additional breathing
motion of the methyl groups, a librational motion of the SDA,
ments allow observation of the freezing out of motions at low
temperatures, or the onset of liquidlike mobility at elevated
temperatures. These changes in molecular mobilities will help
to quantify the organic-inorganic interactions in the as-made
nonasil samples. Molecular dynamics simulations predict that
steric effects alone should prevent the TMC5, TMC6, and TMC7
structure-directing agents from isotropic rotation. For the TMC3
structure-directing agent, however, the nonasil cage does not
exhibit sufficient steric confinement to prevent isotropic rotation.
2
2
or both. The H NMR results obtained do not let us distingush
between these possiblities. However, the two rotations account
for about 95% of the observed reduction in the QCCs and they
dominate the narrowing of the NMR spectra.
2
The variable temperature H NMR spectra of d9-TMC5
nonasil are shown in Figure 7a. The line shape has lost the
sharp features of the Pake doublet at 190 K. The results at this
temperature suggest the rotation rates are reaching the inter-
2
2
Variable Temperature H NMR. Variable temperature H
NMR experiments were performed to study the strength of the
organic-inorganic interactions. Variable temperature experi-
3
4
mediate motion regime, where the rate of the motion is the

2346 J. Phys. Chem. B, Vol. 102, No. 13, 1998
Shantz and Lobo
same order of magnitude as the NMR time scale. Most likely
the rotation about the R-N bond is the slower motion, consistent
with work done by Vega41 on d9-trimethylamine (this will be
further addressed using T1 experiments, vide infra).
2
The variable temperature H NMR spectra for d9-TMC6
nonasil are shown in Figure 7b. The spectrum at 190 K also
shows a loss of the Pake doublet similar to the d9-TMC5 nonasil
sample.
2
Variable temperature H NMR spectra for the d9-TMC7
nonasil sample are shown in Figure 7c. The shoulders of the
doublet are substantially broader at 190 K than any of the other
samples, possibly indicating that the rotation about the nitrogen
C3 axis is slower in this sample. This is the largest structure-
directing agent and the most sterically confined.
2
Figure 8. Model used for H NMR line shape simulations.
TABLE 5: Variable Temperature Spin-Lattice Relaxation
1 1 2 9
Times (T ) and Jumping Frequencies k and k for d -TMC6
Nonasil Using the 9-site Jump Model
2
k1 (s-1)
k2 (s-1)
The trends observed in the variable temperature H NMR
temperature (K)
T1 ms
spectra for the TMC5, TMC6, and TMC7 nonasil systems follow
the relative sizes of the molecules and are similar to the results
for the neat solid salts. In all three samples two major rotations
are observed, the rotation along the nitrogen C3 axis and along
the methyl group C3 axes. The rotation along the nitrogen C3
axis appears to slow to the order of the NMR time scale at 190
8
9
9
7
8
8
2
2
340
40
98
5.7
21.7
28.2
7.8 × 10
7.0 × 10
8.3 × 10
9.0 × 10
5.8 × 10
7.0 × 10
2H NMR Relaxation Experiments. When molecular mo-
tions are fast on the NMR time scale, only a lower bound on
the rate of motion can be estimated. Spin-lattice relaxation
2
K. The rotation rates will be quantified using H T1 experiments
2
described below. Variable temperature H NMR was also
(T ) experiments in conjunction with line shape simulations can
1
performed on the d9-TMC5, -TMC6, and -TMC7 nonasil
samples between 298 and 370 K (not shown). For all samples
the spectra did not change appreciably between 298-370 K.
be used to quantify the rates of rotations in the limit of fast
24-26
motion.
The spin-lattice relaxation time T is a V-shaped
1
function of the correlation time of the motion. The T minimum
1
The variable temperature results for the d9-TMC3 nonasil
sample are shown in Figure 7d,e. Simulations of the 298 K
spectrum for this sample gives a smaller QCC of 13.5 kHz and
a larger asymmetry of η ) 0.14 than for the other three nonasil
samples. The differences suggest that in this case there is an
additional motion that introduces significant asymmetry into the
is approximately 1 ms, and this corresponds to a correlation
time (τcMIN) that is the inverse of the Larmor frequency (ωL) of
2
H (∼46 MHz). Motions that are faster or slower than τcMIN ∼
-
9
1/ωL ∼ 3 × 10 s result in a longer T1.
The line shapes are simulated with the program Fastpowder
using two motions: the rotation along the nitrogen C3 axis and
the rotations of the methyl groups. The rate constant for the
rotation along the nitrogen C3 axis is k2 and the rate constant
for the rotation of the methyl groups is k1. These rotations are
simulated using a nine site jump modeled as shown in Figure
8 and for simplicity, the two rotations were not coupled (i.e.,
jumps from site 1 to sites 2, 3, 4, and 7 are allowed, but jumps
from site 1 to site 5 are not). The bending motion discussed
above was not accounted for in the simulation, because the two
rotations capture the essential features of the NMR line shape
and the molecular motion. The motion of the structure-directing
agent is a composite motion, as a result, the line shape
simulations can be done with nonunique values of the rate
constants for the different motions. Work by Vega41 on solid
d9-trimethylamine indicates that the rotation along the nitrogen
C3 axis is the slower of the motions and this constraint is used
in the line shape simulations.
2
line shape of the H NMR spectrum, likely librational in nature.
The spectra for this sample also include a Lorentzian line which
is approximately 8% of the total integrated intensity. The
Lorentzian line is assigned to molecules undergoing isotropic
reorientation on the NMR time scale, and the peak intensity is
invariant to temperature within the temperature range investi-
gated (see Figure 7d,e). One plausible explanation for the
8
12
presence of the Lorentzian line is a small fraction of the [5 6 ]
cages containing two organic molecules which rotate isotropi-
cally between two sites in the cage. Energy minimization
calculations show that the TMC3 structure-directing agent is
8
12
small enough for the [5 6 ] cage to accommodate two TMC3
molecules. Another possibility is that a fraction of the cages
contain both a structure-directing agent and a sodium ion. The
elemental analysis results indicate that 90% of the cages contain
structure-directing agents, and nearly 20% of the cages in the
TMC3-NON sample contain sodium (on a weight basis). At
this point the available data do not allow us to distinguish
between the two alternatives.
Variable temperature T experiments were performed on the
1
d9-TMC6 nonasil sample to verify that the rotation about the
nitrogen C axis is the slower of the two motions (Table 5). On
3
The main point is that at all temperatures up to 370 K more
than 90% of the structure-directing agents are not isotropically
reorienting. The TMC3 structure-directing agent is small and
is not sterically prevented from isotropic reorientation; as a result
we can conclude there are forces between the structure-directing
agent and the framework that restrict the mobility of the TMC3
molecule. The lack of rapid reorientation in the TMC3-NON
sample is in sharp contrast to nonasil samples made with
electrically neutral amines. Gies17 has shown that the 2-ami-
nopentane used to synthesize nonasil reorients isotropically in
the basis of the known temperature dependence of T1 we can
predict how T1 should change with temperature (motion rate).
Specifically, the rate of motion should increase as the temper-
ature increases, resulting in a longer T1 if the motion rate
constants at 298 K are larger than 1/τcMIN. The variable
temperature results are listed in Table 5, and T1 decreases as
the temperature decreases between 240-340 K. At the lowest
temperature of 240 K the simulated k2 is less than τcMIN, which
means the motion rate likely crossed the minima of the T1 curve
between 298 and 240 K. Nevertheless, these results show that
the rotation rates are faster than 1/τcMIN at 298 K.
8
12
the 5 6 cage. The difference in mobilities observed suggest
that the forces between the structure-directing agent and the
nonasil framework are electrostatic in nature.
Room-temperature T1 experiments were performed on the four
samples and are summarized in Table 4. The values of the

Organic Molecules in the Tectosilicate Nonasil
J. Phys. Chem. B, Vol. 102, No. 13, 1998 2347
spin-lattice (T1) relaxation time are consistent with size effects.
On the basis of the variable temperature T1 results for the d9-
TMC6 sample, we can conclude that for all samples the methyl
group rotation rates are faster than 1/τcMIN at 298 K. The rate
constants used for the 298 K line shape simulations (not shown)
also verify that the rotation about the R-N bond axis is slowest
in the d9-TMC7 sample.
(a)
2
2
H MAS NMR. One of the major limitations of static H
NMR experiments is the need for relatively high concentrations
2
of H in the sample. For our samples with deuterium labeled
alkyl groups, there is approximately 0.07 wt % deuterium in
the sample. The combination of a small amount of deuterium
present in the sample, the small amount of sample used (∼50
2
mg in the static probe), and the low receptivity of H (about
3
1
00 times weaker than 1 C) results in very long acquisition
35
times. Single-pulse MAS experiments can be used to improve
the signal-to-noise ratio. The signal-to-noise ratio improves
because of an increase in sample volume (4 mm MAS rotors)
and because all the spectral information is contained in sharp
spinning sidebands. The MAS experiments can also be used
effectively with soft pulses compared to the traditional static
(b)
2
experiments. For these reasons we have used H MAS NMR
rather than the traditional static quadrupole echo experiment to
study the dynamics of structure-directing agents with deuterium-
labeled substituent alkyl groups. This experimental technique
is limited in our case because spectrometer CPU memory
constraints limit the size of the FID. This makes phasing the
spectra problematic, and as a result the 2H MAS results
presented here are only qualitative.
(c)
As a comparison to the static experiments, Figure 9a shows
the static and MAS spectra for the d9-TMC5 nonasil sample
(spin rate νr ) 2.00 kHz). The envelope of the spinning
sidebands of the MAS spectrum is in good qualitative agreement
with the results from the static experiments. Figure 9b contains
2
the H MAS spectrum of the d1-TMC5 nonasil sample spun at
2
2
6
.0 kHz. The theoretical QCC value for a rapidly puckering
Figure 9. (a) Comparison of static H and H MAS NMR spectra of
2
d
9
-TMC5 nonasil. Spin rate ν
NMR spectra of (b) d -TMC5 nonasil, and (c) d
rate ν ) 6.0 kHz.
r
) 2.0 kHz for MAS spectrum. H MAS
but otherwise immobile cyclopentyl ring is approximately 110
kHz.43 The line shape at 298 K has a QCC of approximately
1
1
-TMC7 nonasil. Spin
r
9
5 kHz. This QCC value is consistent with a rapid puckering
of the cyclopentyl ring, with some additional librational motion.
The additional reduction in the QCC from 110 kHz could also
be due to small angle rotations about the R-N axis. Free
rotation of the cyclopentyl ring can be ruled out as this motion
would result in a QCC at least a factor of 2 smaller than the
observed value. Molecular dynamics simulations predict that
the cyclopentyl ring will undergo small angle rotations about
the R-N bond on the nanosecond time scale. The small angle
rotations would lead to a reduction in the QCC from the value
of 110 kHz for a nonrotating, rapidly puckering cyclopentyl
ring.
quantitative analysis of the line shapes is not possible. The
lack of mobility of the rings are, however, reflected in the values
of the QCCs obtained.
Summary
The dynamics of several organic structure-directing agents
in the as-made zeolite nonasil have been characterized using
2
H NMR spectroscopy. All structure-directing agents with
deuterated methyl groups show fast methyl group internal
rotations and rotation about the nitrogen C3 axis down to 190
K. Three of the four structure-directing agents are sterically
prevented from isotropic reorientation, but the smallest one,
TMC3, is not. Nevertheless, approximately 90% of the
structure-directing agents in the TMC3-NON sample are not
isotropically reorienting at 370 K. The results clearly show
strong electrostatic interactions between the organic molecules
and the zeolite framework. This in contrast to nonasil samples
made with electrically neutral amines, where van der Waal
forces dominate. The steric factors imposed on the structure-
directing agents by the nonasil structure are clearly illustrated
2
Figure 9c shows the H MAS spectrum for the d1-TMC7
nonasil sample spun at 6.0 kHz. Simulating this spectrum
results in a QCC of 110 kHz. The QCC value can be attributed
to a fixed norbornyl ring with librational motion. The static
2
H NMR spectrum for the d1-TMC7 structure-directing agent
indicates the ring is fixed (not shown). The synthesis by Brown
31
et al. to make the ring deuterated amine yields the exo-isomer
of norbornylamine. In our sample, however, the deuterium may
be cis-, trans-, or a mixture of both to the trimethylammonium
2
group. For those reasons, the H MAS spectrum may be the
2
2
superposition of more than one signal. Nonasil can be made
by the H NMR T1 experiments as well as by the H MAS NMR
1
3
with both the endo- and exo-forms of norbornylamine, but
this may lead to a distribution of mobilities. Due to equipment
limitations, phasing the MAS spectra was problematic and a
spectra of samples with deuterated substituent alkyl groups.
2
9
1
The defect sites have been characterized using Si and H
2
9
NMR. The Si results indicate the as-made materials contain

2348 J. Phys. Chem. B, Vol. 102, No. 13, 1998
Shantz and Lobo
3
1
approximately four Q per cage. The H NMR results indicate
that the silanol groups participate in hydrogen bonding with
M. E. J. Am. Chem. Soc. 1997, 119, 76. (c) Hong, S. B.; Cho, H. C.; Davis,
M. E. J. Phys. Chem. 1993, 97, 1622.
(6) (a) Burkett, S. L.; Davis, M. E. J. Phys. Chem. 1994, 98, 4647. (b)
3
7
the siloxy groups, consistent with the work of Koller et al.
Zones, S. I. Zeolites 1989, 9, 458. (c) Gies, H.; Marler, B. Zeolites 1992,
12, 42. (d) Lobo, R. F.; Zones, S. I.; Davis, M. E. J. Inclusion Phenom.
Mol. Recognit. Chem. 1995, 21, 47. (e) Zones, S. I.; Nakagawa, Y.; Yuen,
L. T.; Harris, T. V. J. Am. Chem. Soc. 1996, 118, 7558. (f) Burkett, S. L.;
Davis, M. E. Chem. Mater. 1995, 7, 1453.
On the basis of these results we can conclude that the defect
sites are in close proximity either as isolated pairs or larger
clusters. One unanswered question is whether the orientation
of the trimethylammonium group is correlated toward the defect
site. We are pursuing this issue using two-dimensional Si-
H correlation NMR spectroscopy.
(7) Flanigen, E. M.; Bennet, J. M.; Grose, R. W.; Cohen, J. P.; Patton,
2
9
R. L.; Kirchner, R. M.; Smith, J. V. Nature (London) 1978, 271, 512.
(8) Lawton, S. L.; Rohrbaugh, W. J. Science 1990, 247, 1319.
1
(9) Lobo, R. F.; Pan, M.; Chan, I. Y.; Li, H. X.; Medrud, R. C.; Zones,
Though nonasil is a model system, these results should hold
qualitatively for other all-silica materials. The lack of mobility
of the TMC3 molecule can only be attributed to electrostatic
interactions, which should be present in any all-silica zeolite
made with a cationic structure-directing agent. Extrapolation
to synthesis conditions must be done with some hesitation, but
there are clearly forces between the nonasil framework and the
structure-directing agent that cannot be overcome by the thermal
energy available at 370 K. These interactions could also be
present at synthesis conditions (433 K). Systems where a small
amount of boron or aluminum are used in the synthesis are more
complex. In addition to the cationic structure-directing agents
used, extraframework cations compete to balance the charge of
the aluminum/boron. If these organic-inorganic interactions
are stronger than inorganic (Al/B)-inorganic (Na, K, etc.)
interactions it may lead to selective incorporation of the
aluminum/boron in the framework in close proximity to the
charge on the structure-directing agent during the synthesis. This
would have important implications about the possibility of
tailoring the distribution of catalytically active sites in zeolites.
S. I.; Crozier, P.; Davis, M. E. Science 1993, 262, 1563.
(10) (a) Gies, H.; Marler, B.; Fyfe, C. A.; Kokotailo, G.; Feng, Y.; Cox,
D. E. J. Phys. Chem. Solids 1991, 52, 1235. (b) Lobo, R. F.; Pan, M.;
Chan, I.; Medrud, R. C.; Zones, S. I.; Crozier, P. A.; Davis, M. E. J. Phys.
Chem. 1994, 98, 12040. (c) Knops-Gerrits, P. P.; De Vos, D. E.; Peijen, E.
J. P.; Jacobs, P. A. Microporous Mater. 1997, 8, 3. (d) Engelhardt, G.;
Michel, D. High-Resolution Solid State NMR of Silicates and Zeolites;
Wiley: Chichester, 1987. (e) Davidson, A.; Weigel, S. J.; Bull, L. M.;
Cheetham, A. K. J. Phys. Chem. B 1997, 101, 3065. (f) Pan, M. Micron
1
996, 27, 319. (g) Terasaki, O.; Oshuna, T.; Ohnishi, N.; Hiraga, K. Curr.
Opin. Solid State Mater. Sci. 1997, 2, 14. (h) Thomas, J. M.; Vaughan, D.
E. W. J. Phys. Chem. Solids 1989, 50, 449.
(11) Marler, B.; Dehnbostel, N.; Eulert, H. H.; Gies, H.; Liebau, F. J.
Inclusion Phenom. 1986, 4, 339.
(12) Marler, B.; Gies, H. Zeolites 1995, 15, 517.
(
13) Harris, T. V.; Zones, S. I. Stud. Surf. Sci. Catal. 1994, 64, 29.
(14) Balkus, K. J. Jr.; Shepelev, S. Microporous Mater. 1993, 1, 383.
15) Behrens, P.; van de Goor, G.; Freyhardt, C. C. Angew. Chem., Int.
44
(
Ed. Engl. 1995, 34, 2680.
(16) van de Goor, G.; Freyhardt, C. C.; Behrens, P. Z. Anorg. Allg. Chem.
995, 621, 311.
1
(17) Gies, H. In Inclusion Compounds; Atwoord, J. L., Davies, J. E.
D., MacNicol D. D., Eds.; Oxford: New York, 1991; Vol. 5 (Inorganic and
Physical Aspects of Inclusion).
(18) Fyfe, C. A.; Mueller, K. T.; Kokotailo, G. T. In NMR Techniques
in Catalysis; Bell, A. T., Pines, A. Eds.; Marcel Dekker: New York, 1994.
(
19) Rohr-Schmidt, K.; Spiess, H. W. Multidimensional Solid-State NMR
Acknowledgment. We are grateful to the regents of the
ACS-PRF fund for financial support. The authors thank the
Department of Chemistry and Biochemistry at the University
of Delaware for use of the NMR spectrometer, C. Dybowski
for useful discussions about the NMR results, and R. Griffin
for use of the deuterium NMR line shape simulation programs.
The authors also thank S. Zones (Chevron) for help with the
nonasil synthesis and elemental analysis on the samples and E.
Gaffney and M. Feuerstein for help with the solid-state NMR
spectrometer.
and Polymers; Academic: London, 1994.
(20) (a) Klinowski, J.; Carpenter, T. A.; Thomas, J. M. J. Chem. Soc.,
Chem. Commun. 1986, 956. (b) Barron, P. F.; Frost, R. L.; Skjemstad, J.
O. J. Chem. Soc., Chem. Commun. 1983, 581. (c) Fyfe, C. A.; Wong-Moon,
K. C.; Huang, Y.; Grondey, H.; Mueller, K. T. J. Phys. Chem. 1995, 99,
8707. (d) Engelhardt, G.; Lohse, U.; Samoson, A.; Magi, M.; Tarmak, M.;
Lippma, E. Zeolites 1982, 2, 59. (e) Fyfe, C. A.; Grondey, H.; Feng, Y.;
Kokotailo, G. T.; Ernst, S.; Weitkamp, J. Zeolites 1992, 12, 50.
(21) (a) Hunger, M.; Engelhardt, G.; Koller, H.; Weitkamp, J. Solid State
Nucl. Magn. Reson. 1993, 2, 111. (b) Feuerstein, M.; Hunger, M.;
Englehardt, G.; Amoureux, J. P. Solid State Nucl. Magn. Reson. 1996, 7,
9
5. (c) Koller, H.; Burger, B.; Schneider, A. M.; Engelhardt, G.; Weitkamp,
J. Microporous Mater. 1995, 5, 219.
22) (a) Vega, A. J.; Luz, Z. J. Phys. Chem. 1986, 90, 4103. (b)
Kustanovich, I.; Luz, Z.; Vega, S.; Vega, A. J. J. Phys. Chem. 1990, 94,
138. (c) Eckman, R. A.; Vega, A. J. J. Phys. Chem. 1986, 90, 4679. (d)
Silbernagel, B. G.; Garcia, A. R.; Newsam, J. M.; Hulme, R. J. Phys. Chem.
(
Supporting Information Available: 13C liquid NMR spec-
trum of N,N,N-trimethylnorbornylammonium iodide with and
3
13
without deuterium â to the trimethylammonium group, C CP-
MAS NMR spectrum of TMC3-NON, SEM micrographs of
1
989, 93, 6506. (e) Vega, A. J.; Luz, Z. Zeolites 1988, 8, 19.
(23) Vold, R. R. In Nuclear Magnetic Resonance Probes of Molecular
Dynamics, Understanding Chemical ReactiVity; Tycko, R., Ed.; Kluwer:
London, 1994; Vol. 8.
(
(
2
TMC5-NON and TMC3-NON, H NMR spectra of the neat
2
structure-directing agents, H NMR line shape simulations with
24) Torchia, D. A.; Szabo, A. J. Magn. Reson. 1982, 49, 107.
25) Vold, R. R.; Vold, R. L. In AdVances in Magnetic and Optical
2
and without additional breathing motion, and the H NMR T1
results and simulations for the d9-TMC6 nonasil sample at 298
and 340 K (14 pages). Ordering information is given on any
current masthead page.
Resonance; Warren, W. S., Ed.; Academic: San Diego, 1991; Vol. 16.
(
26) Wittebort, R. J.; Olejniczak, E. T.; Griffin, R. G. J. Chem. Phys.
987, 86, 5411.
27) (a) Bull, L. M.; Henson, N. J.; Cheetham, A. K.; Newsam, J. M.;
1
(
Heyes S. J. J. Phys. Chem. 1993, 97, 11776. (b) Hepp, M. A.; Ramamurthy,
V.; Corbin, D. R.; Dybowski, C. R. J. Phys. Chem. 1992, 96, 2629. (c)
Auerbach, S. M.; Bull, L. M.; Henson, N. J.; Metiu, H. I.; Cheetham, A.
K. J. Phys. Chem. 1993, 100, 5923.
References and Notes
(1) (a) Darby, N. J Protein Structure; Oxford: New York, 1993. (b)
(
28) (a) Sidhu P. S.; Bell, J; Penner, G. H.; Jeffrey, K. R. Can. J. Chem.
Everett, D. H. Basic Principles of Colloidal Science; Royal Society of
Chemistry: London, 1988. (c) Philp, D.; Stoddart, J. F. Angew. Chem., Int.
Ed. Engl. 1996, 35, 1155.
1
1
995, 73, 2196. (b) Meirovitch, E.; Belsky, I.; Vega, S. J. Phys. Chem.
984, 88, 1522. (c) Ripmeester, J. A.; Ratcliffe, C. I. In Inclusion
Compounds; Oxford: New York, 1991; Vol. 5 (Inorganic and Physical
Aspects of Inclusion). (d) Meirovitch, E.; Belsky, I. J. Phys. Chem. 1984,
(
(
2) VanSanten, R. A.; Kramer, G. J. Chem. ReV. 1995, 15, 637.
3) (a) Barrer, R. M. Hydrothermal Chemistry of Zeolites; Academic
8
8, 4308. (e) Collins, M. J.; Ratcliffe, C. I.; Ripmeester, J. J. Phys. Chem.
Press: London, 1982. (b) Moscou, L. Stud. Surf. Sci. Catal. 1989, 58, 1.
c) Breck, D. W. Zeolite Molecular SieVes; Wiley: New York, 1974.
4) (a) Wei, J. Ind. Eng. Chem. Res. 1994, 33, 246. (b) Klein, H.; Fuess,
1990, 94, 157.
(
(
29) den Ouden, C. J. J.; Datema, K. P.; Visser, F.; Mackay, M.; Post,
M. F. M. Zeolites 1991, 11, 418.
30) All-silica material: No trivalent cations (B, Al, etc.) have been
(
H.; Schrimpf, G. J. Phys. Chem. 1996, 100, 11101. (c) Snurr, R. Q.; Bell,
A. T.; Theodorou, D. N. J. Phys. Chem. 1994, 98, 11948.
(
purposely added to the synthesis mixture.
(31) Brown, H. C.; Kim, K. W.; Srebnik, M.; Singaram, B. Tetrahedron
1987, 41, 4071.
(
5) (a) Kubota, Y.; Helmkamp, M. M.; Zones, S. I.; Davis, M. E.
Microporous Mater. 1993, 6, 213. (b) Hong, S. B.; Camblor, M. A.; Davis,

Organic Molecules in the Tectosilicate Nonasil
J. Phys. Chem. B, Vol. 102, No. 13, 1998 2349
(
32) Langford, I. J. Lapod; School of Physics and Space Research:
Alemany, L. B. J. Catal. 1984, 104, 484. (c) Woolery, G. L.; Alemany, L.
B.; Dessau, R. M.; Chester, A. W. Zeolites 1986, 6, 14.
(39) (a) Hunger, M.; Karger, J.; Pfeiefer, H.; Caro, J.; Zibrowius, B.;
Bulow, M.; Mostowicz, R. J. Chem. Soc., Faraday Trans. 1 1987, 83, 3459.
(b) Chezeau, J. M.; Delmotte, L.; Guth, J. L.; Soulard, M. Zeolites 1989, 9,
78. (c) Moser, W. R.; Thompson, R. W.; Chiang, C. C.; Tong, H. J. Catal.
1989, 117, 19.
Univerisity of Birmingham, B15 2TT U.K.
(33) Powles, J. G.; Strange, J. H. Proc. Phys. Soc. 1963, 82, 6.
(34) Spiess, H. W.; Sillescu, H. J. Magn. Reson. 1981, 42, 381.
(35) Kristensen, J. H.; Bildsøe, H.; Jakobsen, H. J.; Nielsen, N. C. J.
Magn. Reson. 1991, 92, 443.
36) In general, the chemical analyses overestimate the C/N ratio by
(
approximately 20%. We believe that this difference is due to the difficulty
of releasing the structure-directing agent occluded in the cages of nonasil.
(40) Eckert, H.; Yesinowski, J. P.; Silver, L. A.; Stolper, E. M. J. Phys.
Chem. 1988, 92, 2055.
(41) Luz, Z.; Vega, A. J. J. Chem. Phys. 1987, 86, 1803.
(42) Maus, D. C.; Copie, V.; Sun, B.; Griffiths, J. M.; Griffin, R. G.;
Luo, S. Schrock, R. R.; Liu, A. H.; Seidel, S. W.; Davis, W. M.; Grohmann,
A. J. Am. Chem. Soc. 1996, 118, 5665.
13
C CP.-MAS NMR unambiguously confirms that the structure-directing
agents are intact inside the nonasil cages.
37) Koller, H.; Lobo, R. F.; Burkett, S. L.; Davis, M. E. J. Phys. Chem.
995, 99, 12588.
38) (a) Nagy, J. B.; Gabelica, Z.; Derouane, E. G. Chem. Lett. 1982,
105. (b) Dessau, R. M.; Smith, K. D.; Kerr, G. T.; Woolery, G. L.;
(
1
(
(43) Meirovitch, E. J. Phys. Chem. 1984, 88, 6411.
(44) Zones, S. I.; Van Nordstrand, R. A. Zeolites 1988, 8, 16.
1