An NMR study of the nitration of toluene over zeolites by HNO3-Ac2O

Article abstract of DOI:10.1039/b103821b

Liquid and solid state NMR studies of the nitration of toluene with a mixture of nitric acid and acetic anhydride were performed in homogeneous solution and in a heterogeneous system using zeolite beta, mordenite, and ZSM-5 as catalysts. Instantaneous formation of acetyl nitrate occurred when nitric acid and acetic anhydride were mixed together but the reaction was not complete. The equilibrium constant for the reaction was ～ 5. Hydrolysis of acetic anhydride to acetic acid shifted the equilibrium back to the reactants, leading to nitric acid formation. The presence of this free nitric acid in the nitrating mixture acted as an acid catalyst during the nitration of toluene. The acetyl nitrate species was the effective nitrating species while isolated nitronium ions were not observed. Interaction of nitric acid with the zeolite led to strong bonding of the nitrate species to the framework aluminum, leading to a surface-bonded aluminum-nitrate complex. Heterogeneous zeolite-catalyzed nitration of toluene by acetyl nitrate improved the para- to ortho-nitrotoluene ratio compared to the homogeneous liquid phase nitration. The best para-selectivity was obtained with Al-beta and to a lesser extent with dealuminated mordenite, suggesting that the selectivity might be related to the presence of flexible lattice aluminum in these large pore zeolites.

Full text of DOI:10.1039/b103821b

View Article Online / Journal Homepage / Table of Contents for this issue
An NMR study of the nitration of toluene over zeolites by
HNO –Ac O
3
2
Mohamed Haouas,a Samantha Bernasconi,a Andreas Kogelbauerb and Roel Prins*a
a L aboratory of T echnical Chemistry, Swiss Federal Institute of T echnology (ET H),
CH-8092 Zurich, Switzerland. E-mail: prins=tech.chem.ethz.ch
b Department of Chemical Engineering and Chemical T echnology, Imperial College of Science,
T echnology and Medicine, L ondon, UK SW 7 2BY
Received 27th April 2001, Accepted 11th September 2001
First published as an Advance Article on the web 29th October 2001
Liquid and solid state NMR studies of the nitration of toluene with a mixture of nitric acid and acetic
anhydride have been performed in homogeneous solution and in a heterogeneous system using zeolite beta
(
small and large crystalline Al-beta with Si/Al \ 13 and pure silicon beta), mordenite (Si/Al \ 74 and 5) and
ZSM-5 (Si/Al \ 20) as catalysts. Acetyl nitrate proved to be the reactive species, although it decomposed
slowly to tetranitromethane which was not active in the nitration. No NMR evidence for the presence of
isolated nitronium ions, often thought to be the nitrating species, was found in the liquid phase. In the absence
of acetic anhydride, nitric acid interacted strongly with the framework aluminium of the zeolite leading to a
complex of one nitrate per aluminium. In contrast to free nitric acid, this species did not seem to take part in
the nitration of toluene. Introduction of acetic anhydride converted the strongly bonded nitrate to the more
reactive acetyl nitrate and acetic acid which were coordinated to the framework aluminium of the zeolite.
Heterogeneous zeolite-catalysed nitration of toluene by acetyl nitrate improved the para- to ortho-nitrotoluene
ratio compared to the homogeneous liquid phase nitration. The best para-selectivity was obtained with Al-beta
and to a lesser extent with dealuminated mordenite, suggesting that the selectivity may be related to the
presence of Ñexible lattice aluminium in these large pore zeolites.
nucleus has a low sensitivity, a low natural abundance and a
Introduction
very long relaxation time, and the 14N nucleus has a quadru-
pole moment. 15N NMR has been successfully applied in the
study of surface-active sites and adsorption complexes,21h24
yet, to the best of our knowledge, neither the adsorption of
nitrating agents on solid acid catalysts, nor heterogeneous
nitration reactions have ever been studied by solid state
NMR. Liquid state 15N NMR studies, on the other hand,
have been extensively undertaken to follow the nitration reac-
tion, using the chemical induced dynamic nuclear polarisation
(CIDNP) e†ects,25h29 and to investigate some homogeneous
nitrating systems.30h33
Nitration of aromatics is one of the most important chemical
reactions in organic chemistry because it constitutes a conve-
nient way of synthesising nitro compounds1 that lead to a
large number of derived substituted aromatics. Acetyl nitrate
is a commonly nitrating agent in homogeneous nitration
because it can be easily produced in situ under mild conditions
by reacting nitric acid with acetic anhydride, which also acts
as a solvent.2 The active nitrating agent in the nitration with
nitric acid in acetic anhydride is supposed to be acetyl nitrate
acting as the carrier for the nitronium ion (NO `).
Nitration is one among a wide variety of organic reactions
2
As part of our on-going programme to understand the
origin of selective nitration of toluene over solid acid catalysts
we have recently reported initial results of solid state NMR on
zeolite beta using nitric acid and acetic anhydride as nitrating
agent.34 In this paper we extend these studies to a variety of
zeolite structures. Since there seemed to be disagreement
between our and other results regarding the activity of nitric
acidÈacetyl nitrate mixtures for the nitration of toluene we
also undertook a 14N liquid state NMR study on the
H OÈHNO ÈAc O nitrating system at various concentrations
that can be catalysed by homogeneous and heterogeneous
catalysts.3,4 Examples of solid acid catalysts used in the nitra-
tion of aromatics include supported clays,5 zeolites,6h8 mixed
metal oxides,9 sulfated zirconia10 and supported silica.11 Many
of these systems are especially attractive since they permit one
to shift the product distribution to commercially more desir-
able compounds. Nitration of simple aromatic compounds
using a nitric acidÈacetic anhydride mixture over solid acid
catalysts shows high regioselectivity.10,12 In order to under-
stand the speciÐc para-selectivity of zeolite beta in toluene
nitration, which is difficult to rationalise by the traditional
shape-selectivity concepts in this wide-pore zeolite,6,13h17 we
carried out a solid state NMR study on some selected zeolites
2
3
2
and nitric acid to acetic anhydride ratios in order to charac-
terise the nitration system. Particular attention was paid to
kinetic and thermodynamic considerations. Regarding the
similarities and di†erences between homogeneous and heter-
ogeneous systems, solid state 15N MAS NMR of the impreg-
nated zeolite with nitric acid and acetic anhydride was also
performed. The nitration of toluene with these homogeneous
and heterogeneous nitrating systems was characterised in situ
by means of liquid and solid state NMR.
(
beta, mordenite and ZSM-5) with di†erent properties.
In situ solid state NMR studies of chemical reactions over
zeolites have been successfully realised under static batch con-
ditions.18,19 Nitrogen NMR has found very wide application
in liquid and solid state chemistry,20 even though the 15N
DOI: 10.1039/b103821b
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075
This journal is ( The Owner Societies 2001
5067

View Article Online
Table 1 Characteristics of the zeolites used
BET
surface area
/m2 g~1
External
surface
area/m2 g~1
Pore
volume
/cm3 g~1
Ave.
mesopore diam.
/AŽ
Micropore
volume
/cm3 g~1
Si/Al
Al-beta
MC-Al-beta
Si-beta
Mor
Dea-mor
ZSM-5
13
13
O
5
74
20
634
620
294
544
512
433
182
4
34
6
38
64
1.06
0.27
0.35
0.27
0.31
0.51
136
27
133
64
57
98
0.19
0.26
0.11
0.22
0.20
0.16
Table 2 Chemical composition of the solutions studied
Experimental
Sample name
wt.%a
Ac O/HNO b
2
3
Materials
Binary system
9
8
7
6
0N0Ac
0N0Ac
0N0Ac
0N0Ac
90
79
69
60
È
È
È
È
1
5N-enriched (]98%) nitric acid as a 40 wt.% aqueous solu-
tion was obtained from Cambridge Isotope Laboratories, Inc.
MA, USA). This solution was mixed with natural abundance
(
fuming nitric acid to obtain a solution of 90 wt.% nitric acid,
resulting in a dilution of the 15N isotope to ca. 7.6%. All
NMR samples were prepared from this mother solution.
Acetic anhydride and toluene were purchased from Fluka as
high purity grade materials (]99.5%). The high purity of
acetic anhydride was checked by gas chromatography analysis
and 1H NMR spectroscopy. The acetic acid content was
found to be negligible (ca. \0.1%). Toluene was stored over a
molecular sieve. These reagents were used as received and no
further puriÐcation was performed.
Ternary system
HNO
90
77
66
59
3 excess
9
8
0N0.5Ac
0N0.5Ac
0.40
0.51
0.28
0.26
70N0.5Ac
60N0.5Ac
Equimolar
9
8
7
6
0N1Ac
0N1Ac
0N1Ac
0N1Ac
90
75
73
58
0.99
0.77
0.84
1.02
Zeolite beta was received from PQ Corporation in its
ammonium form. Zeolites ZSM-5 and mordenite were
received from CU Chemie Uetikon in the sodium form. They
were Ðrst dried for 5 h at 120 ¡C and then calcined at 500 ¡C
for 8 h. The proton form was prepared by 1 h ion exchange
with a 1 M aqueous NH NO solution (10 ml for 1 g zeolite)
Ac O excess
90
76
65
65
2
90N2Ac
1.88
2.00
1.78
2.00
8
7
6
0N2Ac
0N2Ac
0N2Ac
a Concentration of nitric acid in water before addition of Ac O.
b Excess of Ac O needed to react with water is not taken into
4
3
2
under reÑux, followed by drying and calcination as before. A
dealuminated mordenite (dea-mor) was prepared in the fol-
lowing way: the Na-form was Ðrst ion exchanged with 1 M
HCl and then with 1 M NH NO . Subsequently it was cal-
2
account.
For solid state NMR, the zeolite was treated overnight at
130 ¡C, to remove adsorbed species such as water and volatile
organic compounds, and then each zeolite was impregnated,
Ðrst with nitric acid and then with acetic anhydride. The
loading was kept particularly low to keep the sample in the
powder form, necessary for high speed spinning of the NMR
rotor. Thus, typically 6È9 mmol nitric acid (g zeolite)~1 was
used. Acetic anhydride was varied from 0 to 12 mmol (g
zeolite)~1. To study the nitration reaction, toluene was
adsorbed by impregnation on the zeoliteÈHNO ÈAc O
system, at a molar composition of Ac O/HNO /toluene \
/1/1. All samples were prepared at room temperature under
atmospheric conditions. After preparation, the samples were
transferred directly to the NMR rotor, and the rotor was
sealed.
4
3
cined for 2 h at 750 ¡C and then extracted with 6 M HNO to
3
remove the extra-framework aluminium.8 Macrocrystalline
beta (MC-Al-beta) and pure silica beta were synthesised
according to the procedures of Kunkeler et al.35 and Camblor
et al.36 respectively. The procedures for template removal and
ion-exchange were the same as for the commercial zeolites. All
zeolites were fully characterised before use by means of BET,
TGA, AAS, XRD, IR, SEM and solid state NMR. The charac-
teristics of the zeolites are reported in Table 1.
3
2
2 3
1
Sample preparation
For the liquid state NMR study, a HNO ÈH O binary system
3
2
and a HNO ÈH OÈAc O ternary system were prepared by
3
2
2
mixing Ðrst 90 wt.% nitric acid (7.6% 15N) and distilled water
and then adding acetic anhydride. The concentration of the
aqueous nitric acid was varied from 60 to 90 wt.%. The
amount of acetic anhydride added was controlled to give a
nitric acid to acetic anhydride molar ratio between 0.3 and 2.0
NMR
Liquid state. Liquid state NMR spectra were recorded at
room temperature on a Bruker DPX 300 spectrometer oper-
ating at a static Ðeld of 7.05 T using standard 5 mm sample
tubes. The 14N NMR spectra were acquired quantitatively at
a Larmor frequency of 21.69 MHz. Some selected 15N NMR
experiments were carried out to check whether all signals were
visible in the 14N NMR spectra because of the quadrupolar
broadening e†ect. 14N NMR experiments were found more
convenient especially for quantitative purposes. The longest
14N spinÈlattice relaxation time was less than 0.3 s, much
shorter than that of 15N which was longer than 5 min. To
illustrate the feasibility of 14N NMR experiments in the
present study, Fig. 1 compares the 15N and 14N spectra of 62
(
the amount of acetic anhydride that reacted with the water
contained in the initial mixture to give acetic acid has been
subtracted from these reported values). The chemical composi-
tions of these solutions are summarised in Table 2. These
samples were equilibrated for 3 days before NMR measure-
ments. Other samples were prepared in the same way and
analysed by NMR at di†erent stages of mixing from 5 min up
to 1 week. For the nitration studies, toluene was added to an
aged or fresh 90È75 wt.% HNO ÈAc O nitrating reagent. The
3
2
HNO /Ac O ratio was in the range 0.8È1.0 and the
3
2
HNO /toluene ratio was ca. 1. The NMR analyses were run 5
wt.% HNO to which one equivalent of acetic anhydride had
been added. The quality of the 14N spectrum in terms of
3
3
min after mixing.
5068
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075

View Article Online
Fig. 2 Dependence of the 14N NMR chemical shift of aqueous nitric
acid solution on the concentration.
6
0 to 90 wt.% (Fig. 2), expressed as: d(ppm) \ [0.58(HNO
3
wt.%) ] 14.3. It is well known that the nitrogen chemical shift
of nitric acid is a function of its concentration in water.20,37
This is the result of a fast chemical exchange on the NMR
timescale between at least two species, the unprotonated
nitrate ion NO ~ and the nitric acid species HNO . Using
Fig. 1 Comparison between the 15N and 14N NMR spectra of a
mixture of 62 wt.% HNO ÈAc O (Ac O/HNO \ 1; partially 15N
3
2
2
3
enriched D7.6%). The acquisition time is about 14 h (1690 scans) for
1
5N and 30 min (512 scans) for 14N.
3
3
the relationship of Fig. 2 the chemical shift of pure nitric acid
signal to noise ratio is remarkably good since the three sharp
peaks observed in 15N NMR at [68.6, [47.8 and [35.2
ppm, corresponding to acetyl nitrate, tetranitromethane and
trinitromethane respectively, do not su†er much from quadru-
polar broadening e†ects, except for the signal at the highest
Ðeld. All peaks remained detectable, however. The broadness
of the signal of free nitric acid around [42 ppm is not due to
an intrinsic property of the nucleus since it is equally broad in
the 15N spectrum, but could be due to a chemical exchange
process between nitric acid and water. Discussions about
signal assignments will be presented in the results section.
The 14N NMR spectra were recorded with a pulse width of
(
100%) is calculated to be [43.7 ppm, in quite good agree-
ment with the reported value of [42.5 ppm for neat liquid
HNO .37 Similarly, a linear relationship was observed
3
between the chemical shift of the single proton NMR signal
and the nitric acid concentration. In this case, a dynamic
process between water and nitric acid species takes place, the
higher the nitric acid concentration the less shielded the signal
in proton NMR.
1
4N NMR of the ternary HNO ÈH OÈAc O system equili-
3
2
2
brated for 3 days showed di†erent features as a function of the
Ac O/HNO ratio. Hardly any e†ect of the nitric acid concen-
tration was observed on the 14N NMR spectrum in the range
5
are shown in Fig. 3; they correspond to the example of the 90
wt.% HNO ÈAc O system (90NxAc samples). At least three
peaks were observed in these solutions: a very sharp signal at
2
3
21.0 ls (Ñip angle 90¡), an acquisition time of 2.12 s and a
8 to 90 wt.% for a given Ac O/HNO ratio. Typical spectra
relaxation delay of 0.1 s. 128 scans were accumulated with 16
k data points for a spectral width of 4000 Hz. Deuterated
nitromethane, placed in a capillary tube, was used as external
lock and also as reference for the chemical shift scaling. The
chemical shifts had an accuracy better than 0.1 ppm.
2
3
3
2
[
[
47.8 ppm, a broad one at [42.1 ppm and a small peak at
35.1 ppm. Increasing the amount of acetic anhydride led
Ðrst to the disappearance of the broad signal at [42.1 ppm
and then to the disappearance of the small peak at [35.1
ppm. Since these signals behaved independently, they must
belong to di†erent species. The broad signal is attributed to
Solid state. Solid state NMR measurements were carried
out on a Bruker AMX 400 spectrometer at 9.40 T, which cor-
responds to a resonance frequency of 40.54 MHz for 15N.
Samples were sealed in a 7 mm ZrO rotor in air. The rotor
free HNO without any interaction with water since all water
is consumed by the acetic anhydride, as conÐrmed by proton
2
3
was spun up to 5 kHz using a double bearing system of a dry
nitrogen Ñux. Magic-angle spinning (MAS) spectra were
obtained without cross-polarisation or dipolar decoupling by
Fourier transformation of a simple Bloch decay. 15N spectra
were collected using a 6.3 ls pulse (Ñip angle approximately
NMR. Its chemical shift around [42.1 ppm also indicates
that the medium is free of water, in contrast to the binary
4
2
5¡), a sample window of ^38 kHz and recycle delays of 30 s.
880 transients were averaged for each spectrum.
2
7Al NMR spectra were recorded with a short 0.8 ls pulse
(
corresponding to 15¡ solid Ñip angle); 2160 scans were accu-
mulated. All experiments were recorded in lock-Ðeld mode
using deuterated water as an external lock. Chemical shifts
were referenced to 15N-labelled glycine (Cambridge Isotope
Laboratories, Inc.) as a secondary reference for 15N, in which
the main peak was Ðxed at [347.6 ppm with respect to neat
liquid nitromethane,37 and to a standard 1 M Al(NO ) solu-
3
3
tion for 27Al.
Results and discussion
Nitrating systems
The 14N NMR spectrum of the binary HNO ÈH O system
3
2
Fig. 3 14N NMR spectra of 90 wt.% HNO ÈAc O mixtures
showed a single signal whose chemical shift was a function of
the concentration. A linear correlation exists between the
chemical shift and the nitric acid concentration in the range
3
2
(
samples 90NxAc, x \ 0.5, 1, 2 in Table 2) as a function of the
Ac O/HNO ratio. The signal at 0.0 ppm corresponds to the external
2
3
reference (CD NO ).
3
2
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075
5069

View Article Online
system as seen previously. Further addition of 90 wt.% nitric
acid to a mixture in which nitric acid was in excess relative to
acetic anhydride, a†ected only this peak by increasing its
intensity and shifting its position toward low Ðeld (Fig. 4).
These observations conÐrm that the broad peak in Fig. 3 is
due to nitric acid species. The increase in its intensity is due to
the presence of additional nitric acid and the shift to low Ðeld
is due to the interaction of nitric acid with the water from the
aqueous solution that was in excess to that required for com-
plete conversion of the acetic anhydride to acetic acid.
Since the expected acetyl nitrate species was not observed in
1
4N, or in 1H NMR (spectra not shown) in these equilibrated
solutions, experiments were performed on freshly prepared
solutions and the e†ect of time was investigated. The spectral
evolution with time of the 14N NMR spectra of a fresh
mixture of 90 wt.% HNO
(H O)Èacetic anhydride
Fig. 5 14N NMR spectra of a 90 wt.% HNO ÈAc O mixture
3
2
3
2
(
Ac O/HNO \ 0.95) can be seen in Fig. 5. Within the Ðrst
(Ac O/HNO \ 0.95) as a function of time.
2
3
2
3
minutes after mixing, mainly two broad signals were
observed: the Ðrst at [42.3 ppm due to free nitric acid; the
second at [68.3 ppm, a new signal that was not observed in
the aged solution. These two signals decreased slowly in inten-
sity with time and the one at [68.3 ppm disappeared com-
pletely after approximately 3 days. Two other sharp peaks
appeared later: one at [35.4 ppm followed by another at
at 166.6 ppm and the other in the methyl range at 17.3 ppm.
Proton NMR also conÐrmed the presence of acetyl nitrate
through a signal at 1.64 ppm.
The disappearance of the signals of acetyl nitrate with time
demonstrates that acetyl nitrate is unstable under the
described conditions and decomposes slowly at room tem-
perature in an acetic anhydrideÈacetic acid medium. Many
years ago it was reported that tetranitromethane is formed
quantitatively in an equimolar mixture of nitric acid and
acetic anhydride after several days of reaction,38,39 and that
dilution of acetyl nitrate leads to smooth thermolysis into
carbon dioxide and nitromethane.40 Nitromethane was not
observed in our nitrating solutions, neither by 14N nor by 13C
NMR. However, tetra-, tri- and to a lesser extent dinitro-
methane were identiÐed by the NMR coupling of 13C with
[
47.8 ppm. These peaks grew continuously in intensity with
time, with the latter becoming dominant after 1 day. After
approximately 7 days no further change was observed and the
spectrum was similar to that of the equilibrated solution in
Fig. 3. Further addition of one equivalent of acetic anhydride
led to a lowering of the signal of free nitric acid and a regener-
ation of the signal at [68.3 ppm. This latter signal was there-
fore attributed to acetyl nitrate. ConÐrmation of this
assignment was obtained from 13C NMR which showed that
only two new signals were present, besides those of acetic acid
and unreacted acetic anhydride, when the signal at [68.3
ppm was present in 14N NMR. One was in the carbonyl range
1
4N in these compounds.41 The proton-decoupled 13C peak of
tetranitromethane splits into a multiplet due to coupling with
the 14N nuclei (I \ 1) because of the negligible quadrupolar
broadening. The NMR results for poly-nitromethanes
observed in our nitrating solutions are compared with liter-
ature data in Table 3. The slight discrepancies in the chemical
shifts, especially in 14N NMR, may be due to solvent and
chemical environment e†ects. The sharpness of the 14N signal,
especially of tetranitromethane, is an indication of the weak
quadrupolar interaction in these compounds. The very low
amount of dinitromethane could explain why it was not
detected in 13C NMR. The very low signal-to-noise ratio,
which is also due to its very low amount, prevents the precise
determination of the scalar coupling constant in the case of
trinitromethane.
Based on our NMR results, we present in Scheme 1 a
general description of the main chemical reactions occurring
in the ternary system HNO ÈH OÈAc O. Although it is
3
2
2
known that acetyl nitrate decomposes to tetranitromethane,40
the details of the decomposition process were not determined,
most probably because the nitro-containing intermediates had
not been identiÐed. As shown by proton and 13C NMR spec-
troscopy, acetic anhydride Ðrst undergoes hydrolysis. All
Fig. 4 14N NMR spectrum of (a) 80 wt.% HNO ÈAc O mixture
3 2
sample 80N0.5Ac in Table 2) with Ac O/HNO ratio of 0.51, (b) the
(
2
3
same solution 2 months later and just before further addition of nitric
acid, (c) after addition of one more equivalent of 90 wt.% HNO
3
leading to an Ac O/HNO ratio of 0.22.
2
3
Table 3 NMR characteristics of poly-nitromethane
1
4N chemical shifts/ppma
13C chemical shifts/ppma
(J(CN)/Hz)
(
line width/Hz)
Compound
H CNO
Ref. 41c
Present work
Ref. 41b
Present work
[3.8(14)
[25.1(7)
[38.2(2)
[48.1(1)
È
63.4([)
98.6([)
È
È
3
2
H C(NO )
[30.8(6)
[35.4(4)
[47.8(1)
2
2 2
HC(NO )
114.4(10)
118.6(10)
116.0([)
2
3
C(NO )4
118.9(10)
2
a 14N and 13C shifts are relative to internal neat nitromethane and upÐeld of external tetramethylsilane. b Determined as a 10% solution in
chloroform-d. c After correction of chemical shifts from nitromethane in chloroform solution to neat nitromethane.
5070
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075

View Article Online
Fig. 6 VanÏt Ho† plot of acetyl nitrate formation according to: (…)
the present 14N NMR work, (L) SchoÐeld data from ref. 1.
water reacted with acetic anhydride, as indicated by the 14N
chemical shift of free nitric acid (ca. [42 ppm). An excess of
acetic anhydride reacts instantaneously with nitric acid to give
acetyl nitrate. The 14N NMR signal of this latter molecule is
observed within the Ðrst minutes of mixing and then decreases
in intensity with time as acetyl nitrate decomposes slowly into
nitromethane and carbon dioxide. This reaction could
proceed by a radical mechanism.40 Due to the high reactivity
of acetyl nitrate, successive nitration of nitromethane may
occur quickly leading to the Ðnal stable compound tetra-
nitromethane. The observation of di- and trinitromethane in
very low amounts is consistent with this decomposition
mechanism.
To study the equilibrium constant of acetyl nitrate forma-
tion (reaction (2) in Scheme 1), quantitative 14N and 1H NMR
experiments were performed on a 75 wt.% nitric acid and
equimolar acetic anhydride mixture. The experiments were
run within a few minutes after mixing, just before the decom-
position reaction. 14N NMR provided the acetyl nitrate to
nitric acid ratio and thus the degree of reaction progress,
whereas 1H NMR gave the acetic acid to acetic anhydride
ratio. Unfortunately, the proton NMR signals of acetic anhy-
dride and acetyl nitrate overlap and deconvolution is difficult.
Nevertheless, based on the degree of reaction determined from
Scheme 1
signal for nitric acid. Only the microcrystalline beta, ZSM-5
and parent mordenite zeolites showed clearly one distinct
environment for nitric acid. However, macrocrystalline beta
and dealuminated mordenite, and to a lesser extent pure
silicon beta, showed at least two peaks for nitric acid inter-
acting with di†erent amounts of water. The assignment of
these two signals to di†erent environments of nitric acid was
checked twice. First, the proton MAS NMR also showed two
signals for nitric acid interacting with di†erent amounts of
water when two distinct signals in the nitric acid resonance
range were observed in the 15N NMR spectra. Secondly, when
additional experiments were performed with di†erent initial
concentrations of impregnated nitric acid, and therefore di†er-
ent water contents, the observed variation of the relative peak
areas and of the position of these peaks conÐrmed the exis-
tence of di†erent domains of nitric acid within these zeolites.
Besides the signals of free nitric acid, an additional signal was
observed in the case of micro- and macrocrystalline beta at
[
4.2 ppm that was not observed in the liquid state systems.
This signal was only observed with aluminium-containing
beta zeolite and appears to be speciÐc for this zeolite type.
According to its chemical shift this signal could be either
attributed to the nitrosonium ion (NO`)30,31,43 or to nitrate
species.22,37 The former possibility is ruled out since the cor-
responding 27Al NMR spectrum showed a very sharp signal
in the range of octahedral aluminium as an indication of the
formation of a highly symmetric framework aluminium
species upon interaction of nitrate ions with the zeolite
surface.34 Since the species characteristic of the 15N NMR
signal at [4.2 ppm was absent in pure silicon beta it is con-
cluded that these species interact with the aluminium of the
zeolite lattice and form octahedral co-ordination complexes.
14N NMR and the assumption that all water from the initial
nitric acid solution reacts quickly with acetic anhydride to
give acetic acid (which suppresses the formation of acetyl
nitrate), one can calculate the amount of each component. The
last assumption should be valid according to the 14N chemical
shift of free nitric acid (ca. [42 ppm) that is consistent with
pure nitric acid in its fully protonated form. Moreover, that
assumption was also checked by 1H NMR by comparing the
(
AcOH)/(Ac O ] AcONO ) ratio obtained from the integra-
2
2
tion of the proton NMR signal of the acetic acid signal and
the non-resolved signals of acetic anhydride and acetyl nitrate,
with the same ratio calculated from 14N NMR, making the
assumption that all water had reacted. Good agreement was
found between the experimental (1H NMR) and the calculated
values of the (AcOH)/(Ac O ] AcONO ) ratio which were
2 2
.55 and 0.56 respectively. In Fig. 6 we report the VanÏt Ho†
0
curve of acetyl nitrate formation (ln(K) vs. T ~1). The ther-
modynamic parameters extracted were *H¡ \ [3.1 ^ 0.2 kJ
mol~1 and *S¡ \ 3.3 ^ 0.8 J mol~1 K~1, di†erent from those
given in the literature: [24.4 ^ 0.5 kJ mol~1 and
[
83.7 ^ 4.2 J mol~1 K~1 respectively.1 The equilibrium con-
stant varies very little between 0 and 25 ¡C (5.6 ^ 0.4 and
.1 ^ 0.1 respectively), and was in all cases much higher than
reported in the literature (Fig. 6) based on a kinetic study.1,42
5N MAS NMR spectra of 75È90 wt.% aqueous nitric acid
5
1
solutions impregnated on microcrystalline beta, macro-
crystalline beta, pure silicon beta, mordenite, dealuminated
mordenite and ZSM-5 zeolites are shown in Fig. 7. Free nitric
acid species are characterised by signals in the range [25.4 to
Fig. 7 15N MAS NMR spectra of (a) microcrystalline beta, (b)
macrocrystalline beta, (c) pure silicon beta, (d) mordenite, (e) dealu-
minated mordenite and (f) ZSM-5 impregnated with a 75È90 wt.%
HNO aqueous solution (7È8 mmol HNO (g zeolite)~1).
[
41.0 ppm depending on the degree of hydration (water
content). The spectra of some zeolites showed more than one
3
3
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075
5071

View Article Online
Table 4 Amount and loss of tetrahedral framework Al as measured
The expected nitronium species, which is known to give a
signal in the range [124 to [133 ppm,30h33,43h45 was not
observed. In agreement with this assignment, Raman spectros-
copy measurements of the macrocrystalline beta loaded with
by NMR after impregnation of the zeolite with 75È90 wt.% HNO
3
Tetrahedral
Loss of tetrahedral
framework Al (%)b
Zeolite
framework Al (%)a
90 wt.% nitric acid did not show any evidence of the presence
of either nitrosonium or nitronium species. Nitrosonium and
nitronium ions are characterised by bands at 2291 and 1400
Al-beta
MC-Al-beta
Mor
Dea-mor
ZSM-5
2
8
56
20
56
97
91
34
80
38
cm~1,46 as checked in test experiments of NaNO and
2
CsNO dissolved in a 95 wt.% H SO solution, respectively.
3
2
4
The nitric acid loaded macrocrystalline beta, however, showed
bands of nitric acidÈnitrate species47 at ca. 1305, 1050 and 956
cm~1 very similar to those obtained with the nitric acid solu-
tion used for impregnation. Moreover, an additional band at
a Relative to all observed aluminium. b Relative to parent material.
5
19 cm~1 was also observed with the zeolite loaded sample
tion of framework aluminium underwent this transformation
in the case of the parent mordenite and ZSM-5. The fact that
we did not observe the corresponding surface-bonded species
in the 15N MAS spectra of these zeolites may be due to a
NMR detection limitation since the amount of aluminium
involved in the formation of these species is very low com-
pared to that in zeolite beta. The aluminium sites in zeolite
beta seem particularly accessible whereas nitric acid does not
seem to reach all the aluminium sites in mordenite and
ZSM-5.
Addition of acetic anhydride to the nitric acidÈzeolite
system led to the disappearance of the low Ðeld peak at [4.2
ppm in Al-beta and MC-Al-beta and to the appearance of the
signal characteristic of acetyl nitrate at [67.7 ppm. This
means that the surface-bonded nitrate was transformed into
adsorbed acetyl nitrate upon addition of acetic anhydride.
Also, as an indication of the consumption of all water present
in the medium upon addition of excess acetic anhydride, the
remaining nitric acid signal showed a chemical shift of ca.
[42 ppm for all zeolites. Nitric acid is therefore in a uniform
environment in all these systems, although di†erent domains
existed before acetic anhydride addition. In order to check the
stability of the nitrating agent with time, an additional spec-
trum was recorded of an aged sample. Fig. 9 compares spectra
of the same acetic anhydrideÈnitric acidÈAl-beta system mea-
sured just after mixture preparation and 3 months later. The
signal of acetyl nitrate had disappeared and the signal of tetra-
nitromethane appeared at [47.3 ppm. As in solution, acetyl
nitrate adsorbed on the zeolite seems to be unstable and to
undergo decomposition to tetranitromethane.
which was not conclusively identiÐed and was tentatively
attributed to the surface-bonded nitrate species. In order to
gain more information about the new 15N NMR signal at
[
4.2 ppm, static NMR measurements and 15NM1HN CP-MAS
experiments were carried out. Fig. 8 shows the spectra
obtained on microcrystalline beta together with the MAS
spectrum for comparison. While the MAS spectrum showed
two peaks (Fig. 8(a)), static 15N solid state NMR (Fig. 8(b))
showed only the nitric acid signal. This indicates that the
nitric acid species are in a liquid-like state. Spinning at 5 kHz
does not a†ect the line shape much; the width is only moder-
ately reduced by approximately 30 Hz. The low Ðeld signal,
on the other hand, apparently has a very strong chemical shift
anisotropy (CSA); it broadens and merges with the base line
under static conditions. In line with this, the MAS spectrum
showed spinning side bands only for this peak. It is therefore
concluded that this signal belongs to species that interact
strongly with the zeolite surface, such as ionic entities. Such
charged species could be nitrate moieties bonded to the zeolite
wall. CP-MAS experiments failed to show any signals, indicat-
ing that no efficient cross polarisation from proton to nitrogen
occurred. The reason may be the high mobility of the physi-
sorbed free nitric acid species and, for the case of strongly
surface-bonded nitrate species, the absence of protons close
enough to the nitrogen nucleus.
As reported before, 27Al MAS NMR of the zeolite after
impregnation with 75È90 wt.% HNO showed the appearance
of a very sharp signal at 0.4 ppm for all zeolites. Quantitative
3
measurements indicated that some framework aluminium was
converted into an octahedral environment. Nitrate ions seem
to co-ordinate strongly to framework aluminium atoms of the
zeolite and transform them to highly symmetric octahedral
aluminium complexes. The extent of this transformation of the
aluminium environment depends on the zeolite type, as shown
in Table 4. The tetrahedral to octahedral conversion was
almost complete with zeolite beta and to a lesser extent with
the dealuminated mordenite, but only a relatively small frac-
The 27Al NMR spectrum showed that the characteristic
very sharp peak due to the octahedral aluminium nitrate
complex species disappeared when acetic anhydride was
added to the mixture, consistent with the disappearance of the
15N NMR peak at [4.2 ppm. At the same time, some tetra-
Fig. 8 15N solid state NMR spectra of microcrystalline beta zeolite
impregnated with 90 wt.% HNO (8 mmol HNO (g zeolite)~1) at
Fig. 9 15N MAS NMR spectra of microcrystalline beta impregnated
Ðrst with 90 wt.% HNO (6 mmol HNO (g zeolite)~1) and then with
3
3
di†erent NMR experimental conditions (a) 5 kHz MAS, (b) static and
3
3
(
c) proton CP-MAS using a 10 s recycling delay, a 5 ms CP contact
time and a 5 kHz spinning rate. Stars indicate side bands.
acetic anhydride (Ac O/HNO \ 2): (a) just after acetic anhydride
2
3
impregnation, (b) 3 months later.
5072
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075

View Article Online
hedral aluminium was regenerated although a dominant
broad signal persisted in the octahedral range. This broad
signal in the octahedral range could be attributed to alu-
minium co-ordinated with the carboxylic functionality of
acetyl nitrate, acetic acid and/or acetic anhydride. The partial
recovery of tetrahedral framework aluminium is strong evi-
dence that the sharp signal obtained when nitric acid interacts
with the zeolite, is due to nitrate bonded to framework alu-
minium. The transformation of tetrahedral to octahedral alu-
minium and vice versa in the framework of zeolite beta has
been demonstrated by Bourgeat-Lami et al.48 The Ñexibility
of the lattice in such a large pore zeolite may allow such a
change of the framework aluminium co-ordination sym-
metry.34
Nitration of toluene
The nitration of toluene in homogeneous solution was fol-
Fig. 10 15N MAS NMR spectra of microcrystalline beta; (a) impreg-
nated with 90 wt.% HNO (8 mmol HNO (g zeolite)~1) and then
lowed by 14N NMR using
a fresh 90 wt.% HNO
H O)ÈAc O solution. The 14N NMR spectrum of this nitrat-
3
3
3
(
with one equivalent of toluene, (b) Ðrst impregnated with 90 wt.%
2
2
ing agent just before addition of toluene showed the presence
of acetyl nitrate species. When one equivalent of substrate was
added to the nitrating agent mixture, acetyl nitrate disap-
peared completely and two broad signals appeared at [5.7
and [10.3 ppm. These signals were identiÐed as those of
ortho- and para-nitrotoluene, respectively by comparison with
the 14N spectra of the authentic compounds. According to
proton and 13C NMR, they constitute the main reaction pro-
ducts. The ortho-to-para ratio was 2, indicating that there was
no regio-selectivity in the nitration in agreement with catalytic
studies reported earlier.11 Based on the fact that the acetyl
nitrate 14N NMR signal disappeared instantaneously when
toluene was added, it is concluded that the e†ective nitrating
species is acetyl nitrate. Because the reaction is fast, the kinetic
product (the ortho-isomer) is favoured, which may explain the
poor para-to-ortho ratio in homogeneous nitration. In an
additional nitration experiment, using an old nitrating solu-
tion in which the dominant species was tetranitromethane and
acetyl nitrate was absent, no nitration reaction occurred.
HNO (8 mmol HNO (g zeolite)~1) and acetic anhydride
3
3
(
Ac O/HNO \ 0.9), and then with one equivalent of toluene. The
2
3
arrow shows the signal of ortho-nitrotoluene overlapping with that of
the aluminium nitrate species.
indicating that they do not interfere with the octahedrally co-
ordinated aluminium nitrate species. However, the 15N signal
of nitric acid decreased substantially in intensity and also
shifted to lower Ðeld. This indicates that nitric acid was con-
sumed and that some water was formed as by-product. In the
case of nitration in the presence of acetic anhydride, the signal
of free nitric acid was shifted to higher Ðeld as a consequence
of water consumption and its conversion to acetic acid. The
acetyl nitrate signal disappeared after the reaction, conÐrming
that it is also the e†ective reactive species in the heterogeneous
nitration.
Table 5 summarises the activity and selectivity of the
toluene nitration with di†erent zeolites as measured in situ by
solid state NMR. The general tendency is that the activity is
higher when acetic anhydride is present. Presumably the
reason for this is the formation of water as by-product, in the
absence of acetic anhydride, which could act as a poison for
the catalytic sites in the zeolite. On the other hand, the heter-
ogeneous nitration gave a higher para-selectivity than the
homogeneous nitration in the absence of a zeolite in which the
para-to-ortho ratio was only 0.6. Nevertheless, the para-
selectivity varied signiÐcantly with the zeolite structure and its
crystal size and aluminium content. The best results were
obtained with microcrystalline beta, while pure silicon beta
gave the lowest para-to-ortho ratio. Aluminium sites seem to
play an important role in the selectivity of the catalyst. The
zeolite topology of the structure also appears to be critical
since mordenite and ZSM-5 were less selective than beta
zeolite.15 Macrocrystalline beta showed a lower selectivity
than microcrystalline beta, presumably due to the lack of
accessibility for the reactants to surface active sites in large
crystal zeolites.
The nitration of toluene by either acetyl nitrate or nitric
acid (in the absence of acetic anhydride) over the zeolite cata-
lysts was also followed by solid state 15N MAS NMR. Typical
spectra are shown in Fig. 10 for the case of microcrystalline
beta. Nitration occurred in both cases for all the zeolites, in
the presence as well as the absence of acetic anhydride, as
indicated by the presence of the signals of ortho- and para-
nitrotoluene at ca. [5.4 and [10.9 ppm respectively. In the
absence of acetic anhydride and in case of Al-beta, the signal
of surface bonded nitrate ([4.2 ppm) was still observed after
the nitration. This nitrate species does not seem to participate
directly in the reaction since the intensity of its NMR signal
did not change signiÐcantly during the course of the nitration.
2
7Al NMR of the tolueneÈHNO Èzeolite system showed the
3
typical narrow peak at 0.4 ppm giving a very similar spectrum
to that of the HNO Èzeolite system before addition of toluene.
3
Neither the presence of the substrate (toluene), nor that of the
nitration product (nitrotoluene) altered the 27Al spectrum,
Table 5
NMR
Activity and selectivity of the zeolite catalysts used in the nitration of toluene with HNO3
and HNO ÈAc O as measured by in situ
3
2
Activity (nitration yield) (%)
HNO3
Selectivity (para/ortho)
Zeolite
HNO ÈAc O
HNO3
HNO ÈAc O
3
2
3
2
Al-beta
MC-Al-beta
Si-beta
Mor
Dea-mor
ZSM-5
27
51
È
17
11
44
50
37
27
È
27
59
0.9
1.3
È
0.9
0.8
1.0
2.5
1.5
0.9
È
1.1
1.0
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075
5073

View Article Online
The para-to-ortho ratios determined by in situ NMR, in
which the products are within the zeolite pores, are quite
similar to those obtained by GC analysis after product extrac-
tion in batch nitration. For example, the para-to-ortho ratios
of 2.5 and 0.9 obtained with microcrystalline beta by NMR on
HNO ÈAc OÈAl-beta and HNO ÈAl-beta were very close to
that could be due to aluminium co-ordinated by acetic acid.
This assignment is based on the observation of a similar signal
at [5.2 ppm when the zeolite was impregnated with acetic
acid only. Aluminium in the zeolite could act as a trap for the
undesirable by-product acetic acid. Irrespective of which of
the two mechanisms actually prevails, both involve the lattice
aluminium as anchoring sites. The particular Ñexibility of the
lattice aluminium of zeolite beta is of special beneÐt in this
respect and has also been correlated with the activity of the
catalyst by others.50
3
2
3
those of 2.8 and 0.8, respectively, found in the batch
reactor.14,15,17 This is conclusive evidence against product
shape selectivity e†ects occurring with this zeolite. On the
other hand, direct steric e†ects do not seem to play a critical
role in the regio-selective nitration of toluene since the para-
selectivity increases with increasing channel size of the zeolite3
and since no substantial enhancement of the regio-selectivity
was observed with pure Si-beta. Therefore, steric hindrance
induced indirectly by a transition-state selectivity e†ect49 or
by adsorption phenomena of the reactive acetyl nitrate species
or of a non-reactive species like acetic acid, are more likely to
govern the selectivity patterns. Two di†erent scenarios are
conceivable in this respect. The regio-selectivity may be
directed by a speciÐc orientation of the activated substrate
toward the adsorbed nitrating species to minimise steric hin-
drance in the microporous channels, as depicted in Scheme 2.
Alternatively, the presence of strongly adsorbed species like
acetic acid, nitrate or even acetyl nitrate could reduce the free
micropore volume. This may subsequently induce classical
transition-state selectivity e†ects, i.e., hindering the formation
of the bulkier transition state for ortho-substitution in the par-
tially blocked micropore volume of the zeolite. The lack of
selectivity enhancement that was observed when the nitration
of toluene was carried out with pure nitric acid suggests that
the scenario depicted in Scheme 2 is more probable. During
the nitration with pure nitric acid all aluminium sites were
occupied with nitrate species, leading to the before-mentioned
reduction of the micropore volume and nitration occurred
through non-bonded nitric acid in the zeolite micropores. It
should also be mentioned that when acetic anhydride was
present during the nitration reaction, the 27Al NMR showed
signiÐcant further recovery of tetrahedral framework alu-
minium as a result of consumption of surface bonded reactive
species.
Conclusion
Nitrogen NMR proved to be a powerful technique for the
characterisation of the nitrating agent and for the nitration
reaction in the liquid and in the solid state. Liquid 14N NMR
was shown to be feasible and very informative due to the rela-
tively small quadrupolar interactions in the species formed in
the HNO ÈAc O system. Instantaneous formation of acetyl
nitrate occurs when nitric acid and acetic anhydride are mixed
together but the reaction is not complete. The equilibrium
constant of this reaction was found to be ca. 5; it was insensi-
tive to temperature between 0 and 25 ¡C. Hydrolysis of acetic
anhydride to acetic acid shifts the equilibrium back to the
reactants, leading to nitric acid formation. The presence of
this free nitric acid in the nitrating mixture may act as an acid
catalyst during the nitration of toluene. Acetyl nitrate,
however, is unstable in the acetic acidÈacetic anhydride
medium and undergoes a spontaneous, slow decomposition to
tetranitromethane within 3 days, resulting in deactivation of
the nitrating agent. The acetyl nitrate species was demon-
strated to be the e†ective nitrating species while isolated nitro-
nium ions were not observed.
Interaction of nitric acid with the zeolite led to strong
bonding of the nitrate species to the framework aluminium
leading to a surface-bonded aluminiumÈnitrate complex. Due
to the strong interaction of these species with the zeolite wall,
they did not take part in the nitration of toluene. However,
acetic anhydride converted them to the reactive nitrating
species, the surface-bonded acetyl nitrate. The beneÐcial
action of acetic anhydride with respect to the improvement of
activity and enhancement of the para-selectivity in toluene
nitration may be related not only to the capture of the water
from the reaction mixture but also to the generation of the
highly active acetyl nitrate species. Among the zeolites used as
catalyst, Al-containing beta gave the best results with regard
to para-selectivity. The accessible aluminium sites in beta
zeolite seem to play an important role for para-selectivity
since pure silicon beta led to the lowest para-to-ortho ratio.
Although the para-selectivity induced by the zeolite is still not
perfectly understood, the present study demonstrates that the
sieve e†ect cannot account for this selectivity since the
product distribution measured in situ by NMR was similar to
that measured by GC ex situ after extraction of the products
from the solid catalyst. Selectivity induced by steric hindrance
of adsorbed activated species, which could also be seen as
transition-state selectivity, is most likely the cause for the high
para-to-ortho ratio obtained with zeolite beta. Other struc-
tures having di†erent void space size or shape, like ZSM-5 or
mordenite, are less selective.
3
2
This, however, cannot be considered conclusive proof since
the nitration with acetyl nitrate involves the presence of
bulkier species such as acetic anhydride and acetic acid which,
bonded to the zeolite surface, might be sufficient to a†ect the
reduction in the free micropore volume. 27Al NMR of the
zeolite after nitration in the presence of acetic anhydride,
showed a broad signal in the octahedral range at [5.1 ppm
Acknowledgements
We are indebted to Mr F. Bangerter for his help in the liquid
state NMR experiments and to Dr G. Pirngruber for his com-
ments on the manuscript. We acknowledge Zeolyst Interna-
tional and CU Chemie Uetikon for providing the commercial
zeolites.
Scheme 2
5074
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075

View Article Online
2
5
J. F. Johnston, J. H. Ridd and J. P. B. Sandall, J. Chem. Soc.,
References
Perkin T rans. 2, 1991, 623.
1
2
3
4
5
6
7
8
9
K. SchoÐeld, Aromatic Nitration, Cambridge University Press,
26 J. H. Ridd, S. Trevellick and J. P. B. Sandall, J. Chem. Soc.,
Perkin T rans. 2, 1992, 573.
27 M. Lehnig, J. Chem. Soc., Perkin T rans. 2, 1996, 1943.
28 M. Lehnig, Acta Chem. Scand., 1997, 51, 211.
29 M. Lehnig and K. Schurmann, Eur. J. Org. Chem., 1998, 913.
Ž
Cambridge, 1980.
G. A. Olah, R. M. Malhotra and S. C. Narang, Nitration:
Methods and Mechanisms, VCH, New York, 1989.
L. V. Malysheva, E. A. Paukshtis and K. G. Ione, Catal. Rev.-Sci.
Eng., 1995, 37, 179.
30 G. A. Olah, B. G. B. Gupta and S. C. Narang, J. Am. Chem. Soc.,
1979, 101, 5317.
J. H. Clark and D. J. Macquarrie, Org. Process Res. Dev., 1997, 1,
1
49.
A. Corne
984, 93, 961.
31 G. K. S. Prakash, L. Heiliger and G. A. Olah, Inorg. Chem., 1990,
29, 4965.

lis, P. Laszlo and P. Pennetreau, Bull. Soc. Chim. Belg.,
1
32 G. A. Olah, V. P. Reddy and G. K. S. Prakash, Synthesis, 1992,
1087.
T. J. Kwok, K. Jayasuriya, R. Damavarapu and B. W. Brodman,
J. Org. Chem., 1994, 59, 4939.
33 G. A. Olah, G. Rasul, R. Aniszfeld and G. K. S. Prakash, J. Am.
L. E. Bertea, H. W. Kouwenhoven and R. Prins, Stud. Surf. Sci.
Catal., 1994, 84, 1973.
Chem. Soc., 1992, 114, 5608.
34 M. Haouas, A. Kogelbauer and R. Prins, Catal. L ett., 2000, 70,
61.
L. Bertea, H. W. Kouwenhoven and R. Prins, Appl. Catal., 1995,
1
29, 229.
35 P. J. Kunkeler, D. Moesops and H. vanBekkum, Microporous
H. Sato, K. Hirose, K. Nagai, H. Yoshioka and Y. Nagaoka,
Appl. Catal., 1998, 175, 201.
Mater., 1997, 11, 313.
36 M. A. Camblor, A. Corma and S. Valencia, J. Chem. Soc., Chem.
Commun., 1996, 2365.
1
1
0
1
G. D. Yadav and J. J. Nair, Catal. L ett., 1999, 62, 49.
A. Kogelbauer, D. Vassena, R. Prins and J. N. Armor, Catal.
T oday, 2000, 55, 151.
37 M. Witanowski, L. Stefaniak and G. A. Webb, Annu. Rep. NMR
Spectrosc., 1993, 25, 1.
38 F. D. Chattaway, J. Chem. Soc., 1910, 2099.
39 F. H. Cohen and J. P. Wibaut, Recl. T rav. Pays-Bas, 1935, 54,
409.
1
1
1
1
1
1
1
1
2
3
4
5
6
7
8
9
K. Smith, A. Mudsson and G. A. DeBoos, J. Org. Chem., 1998,
6
3, 8448.
C. E. Webster, R. S. Drago and M. C. Zerner, J. Phys. Chem. B,
999, 103, 1242.
D. Vassena, A. Kogelbauer and R. Prins, Stud. Surf. Sci. Catal.,
1
40 G. R. Bachman and T. F. Biermann, J. Org. Chem., 1970, 35,
4229.
1
999, 125, 501.
41 M. D. Coburn, C. B. Storm, D. W. Moore and T. G. Archibald,
Magn. Reson. Chem., 1990, 28, 16.
D. Vassena, D. Malossa, A. Kogelbauer and R. Prins, Mater. Res.
Soc. Proc., 1999, 2, 1471.
42 A. Fischer, A. J. Read and J. Vaughan, J. Chem. Soc., 1964, 3691.
43 Y. Ko, F. T. Bonner, G. B. Crull and G. S. Harbison, Inorg.
Chem., 1993, 32, 3316.
D. Vassena, A. Kogelbauer and R. Prins, Catal. T oday, 2000, 60,
2
75.
D. Vassena, A. Kogelbauer and R. Prins, Stud. Surf. Sci. Catal.,
000, 130, 515.
T. Xu, J. Zhang and J. F. Haw, J. Am. Chem. Soc., 1995, 117,
171.
44 K. F. Chew, M. A. Healy, M. I. Khalil and N. Logan, J. Chem.
Soc., Dalton T rans., 1975, 1315.
2
45 M. I. Khalil, Inorg. Chem., 1990, 29, 5131.
46 K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, Wiley-Interscience, New York, 1978.
47 H. Lucas and J.-P. Petitet, J. Phys. Chem. A, 1999, 103, 8952.
48 E. Bourgeat-Lami, P. Massiani, F. DiRenzo, P. Espiau, F. Fajula
and T. DesCourieres, Appl. Catal., 1991, 72, 139.
49 J. C. vanderWaal, E. J. Creyghton, P. J. Kunkeler, K. Tan and H.
vanBekkum, T op. Catal., 1997, 4, 261.
3
A. ThursÐeld, M. W. Anderson, J. Dwyer, G. J. Hutchings and D.
Lee, J. Chem. Soc., Faraday T rans., 1998, 94, 1119.
G. W. Buchanan, T etrahedron, 1989, 45, 581.
V. M. Mastikhin, S. V. Filimonova, I. L. Mudrakovsky and V. N.
Romannikov, J. Chem. Soc., Faraday T rans., 1991, 87, 2247.
V. M. Mastikhin and S. V. Filimonova, J. Chem. Soc., Faraday
T rans., 1992, 88, 1473.
W. O. Parker, Magn. Reson. Chem., 1999, 37, 433.
J. Wu and S. C. Larsen, J. Catal., 1999, 182, 244.
2
2
0
1
2
2
50 F. Vaudry, F. DiRenzo, F. Fajula and P. Schultz, J. Chem. Soc.,
Faraday T rans., 1998, 94, 617.
2
2
3
4
Phys. Chem. Chem. Phys., 2001, 3, 5067È5075
5075