Catalytic flash vacuum pyrolysis (CFVP): A new way to afford 7H-dibenzo[b,d]azepin-7-one

Article abstract of DOI:10.1016/j.apcata.2009.11.003

Catalytic flash vacuum pyrolysis (cfvp) of 2-(1H-1,2,3-benzotriazol-1-yl)-phenylethanone (1) was performed over different mesoporous solids of the MCM-41 type: pure Si-MCM-41, Al-MCM-41 and Ti-MCM-41 with Si/metal ratio equal to 20 and 120. While in the absence of the catalyst (conventional fvp) the thermal reaction afforded 7H-dibenzo[b,d]azepin-7-one (5) as the main product between 400 and 500 °C, in the cfvp systems different products were obtained between 350 and 450 °C depending on the type of catalysts. Al-MCM-41 showed the best catalytic performance with azepinone selectivity of 87% at 450 °C. The mechanisms of formation of these different products as well as the effect of different solids are discussed. Mesoporous materials used in this study have been prepared by direct hydrothermal synthesis. Structural regularity was determined by XRD and highly ordered structures were obtained.

Full text of DOI:10.1016/j.apcata.2009.11.003

Applied Catalysis A: General 373 (2010) 98–103
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Catalytic flash vacuum pyrolysis (CFVP): A new way to afford
7H-dibenzo[b,d]azepin-7-one
a
b
Elizabeth L. Moyano a, Griselda A. Eimer b, , Paola L. Lucero , Corina M. Chanquia ,
*
Eduardo R. Herrero b, Gloria I. Yranzo a
a
´
´
´
´
INFIQC, Departamento de Quı´mica Organica, Facultad de Ciencias Quımicas, Universidad Nacional de Cordoba, Ciudad Universitaria, 5016 Cordoba, Argentina
b
´
´
´
CITeQ, Universidad Tecnologica Nacional, Facultad Regional Cordoba, 5016 Cordoba, Argentina
A R T I C L E I N F O
A B S T R A C T
Article history:
Catalytic flash vacuum pyrolysis (cfvp) of 2-(1H-1,2,3-benzotriazol-1-yl)-phenylethanone (1) was
performed over different mesoporous solids of the MCM-41 type: pure Si-MCM-41, Al-MCM-41 and Ti-
MCM-41 with Si/metal ratio equal to 20 and 120. While in the absence of the catalyst (conventional fvp)
the thermal reaction afforded 7H-dibenzo[b,d]azepin-7-one (5) as the main product between 400 and
500 8C, in the cfvp systems different products were obtained between 350 and 450 8C depending on the
type of catalysts. Al-MCM-41 showed the best catalytic performance with azepinone selectivity of 87% at
450 8C. The mechanisms of formation of these different products as well as the effect of different solids
are discussed. Mesoporous materials used in this study have been prepared by direct hydrothermal
synthesis. Structural regularity was determined by XRD and highly ordered structures were obtained.
ß 2009 Elsevier B.V. All rights reserved.
Received 18 September 2009
Accepted 2 November 2009
Available online 18 November 2009
Keywords:
MCM-41
Flash vacuum pyrolysis
Azepinone
Mesoporous catalysts
Nitrogen extrusion
1. Introduction
elements into the silica framework is studied in order to increase
the acidity, redox properties or specific catalytic activity of these
MCM-41, a member of the M41S family, has a well-defined pore
structure [1,2] that makes it a potentially useful selective catalyst
or a catalyst support for the production of both commodities and
also fine chemicals. Other interesting physical properties of these
materials include a high specific surface of up to 1000 m2/g, a pore
volume of up to 1.3 ml/g, and a high thermal stability. This material
consists of a hexagonal array of unidimensional, hexagonally
shaped pores whose diameters vary systematically between 1.5
and 10 nm according to the type of surfactant as the template,
auxiliary chemicals and reaction conditions. During the last
decades, the development of these ordered mesoporous materials
has paved the way for the processing of large organic molecules
using heterogeneous catalysts, since their pores are larger than
those of microporous molecular sieves, allowing a faster diffusion
of bulky reactants into the active sites inside the pores. Thus, the
use of mesoporous molecular sieves for reactions involving larger
molecules is of marked interest to fine chemical and pharmaceu-
tical industries. On the other hand, the pore shape, chemical
properties and surface areas of these materials can also be
modified according to the desired catalytic properties [3–9].
Therefore, the incorporation of aluminum and transition metal
mesoporous silica molecular sieves. The use of flash vacuum
pyrolysis (fvp) for synthesis and mechanistic studies is well
documented [10–12]; however, the information on reactions
performed in heterogeneous media employing thermostable solid
catalysts is scarce, which makes the study of catalytic flash vacuum
pyrolysis (cfvp) reactions an interesting challenge. Heterogeneous
fvp reactions may be classified into reactions with solid catalysts
and reactions using solid reagents. In the former, examples include
the preparation of azaxylylene from aminobenzyl alcohol or 1-
phenylbenzoxazinone [13], stilbene isomerization using silica–
alumina and clay as catalysts [14],
a-pinene transformation over
faujasite-like zeolites and amorphous silicoaluminates [15],
formation of oxapentacycloundecanones in fvp over mineral solid
acids [16], fvp reactions of NH-pyrazoles over zeolites [17] and
hydrotalcites [18], and synthesis of azolopyrimidin-7-ones in fvp
of azolyl-malonamates over acid zeolites [19]. On the other hand,
there are some references on heterogeneous fvp with solid
reagents such as dehalogenation reactions over magnesium [20]
and over solid bases such as sodium carbonate [21]. Such fvp
reactions with the use of different solid reagents are also called
vacuum gas–solid reactions (VGSR) [22].
Previous homogeneous fvp studies performed in our group
showed that acylbenzotriazole 1 underwent nitrogen elimination to
give the azatropone 7H-dibenzo[b,d]azepin-7-one 5 as the main
product between 400 and 500 8C [23] (Scheme 1). This methodology
* Corresponding author. Tel.: +54 351 4690585.
E-mail address: geimer@scdt.frc.utn.edu.ar (G.A. Eimer).
0926-860X/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2009.11.003

E.L. Moyano et al. / Applied Catalysis A: General 373 (2010) 98–103
99
Scheme 1. Nitrogen extrusion of acylbenzotriazole 1.
constitutes an interesting alternative to prepare bioactive azatro-
pones without several complications of dimerization reactions or
hydrolysis described in the conventional synthesis [24,25]. How-
ever, the pyrolytic process was not selective; many products were
obtainedandyieldsto5werenotgoodenough.Takinginaccountthe
resultsfoundinthe thermalreactionsofNH-pyrazoles usingzeolites
as catalysts, where it was possible to lower the reaction temperature
for nitrogen extrusion about 300 8C improving the selectivity [17],
we decided to apply solid catalysts to the fvp reactions of
benzotriazole 1. Due to their excellent characteristics, the mesopor-
ous MCM-41 materials mentioned above were chosen for the
development of heterogeneous systems. Thus, in this work we have
studied cfvp systems using pure Si-MCM-41, Ti-MCM-41 (Si/Ti = 20
and 120) and Al-MCM-41 (Si/Al = 20) as solid catalysts in the
reactions of acylbenzotriazole 1.
water/Si = 60. For comparison, a pure siliceous MCM-41 sample
was also prepared by the same synthesis method and at the same
reaction conditions. To remove the template, the samples were
heated (heating rate of 2 8C/min) under N2 flow up to 500 8C
maintaining this temperature for 6 h and subsequently calcined at
500 8C under air flow for 6 h.
2.2. Physicochemical characterization of catalysts
X-ray powder diffraction patterns were collected in air at room
temperature on a Rigaku diffractometer using Cu K
a radiation of
wavelength 0.15418 nm. Diffraction data were recorded at an
interval of 0.018 and a scanning speed of 28/min was used. The
repetition distance of the pores was obtained by the Bragg law
using the position of the first X-ray diffraction line. The lattice
parameter a0 of the hexagonal unit cell can be calculated by a0 = (2/
H3)d1 0 0. Diffuse reflectance spectra of the Ti-MCM-41 samples
were recorded using an Optronicss OL 750-427 spectrometer in the
wavelength range 200–500 nm [8]. Solid-state NMR spectra of the
Al-MCM-41 sample were taken on a BRUKER MSL300 spectro-
meter operating at 78.2 MHz for 27Al [7]. The specific surface area
of all the samples was determined using Pulse Chemisorb
equipment. The samples were previously dried at 400 8C under
N2 flow for 2 h.
2. Experimental
2.1. Catalysts preparation
The mesoporous materials MCM-41 were prepared by hydro-
thermal synthesis using dodecyltrimethyl ammonium bromide
(DTMABr) as a template, following the procedures reported
elsewhere [8,9]. Tetraethoxysilane (Fluka, 98%), titanium isoprop-
oxide (Fluka, 98%) and sodium aluminate (Johnson Matthey) were
used as the Si, Ti and Al sources, respectively. In a typical synthesis,
the reactants were vigorously mixed for 30 min. Then, 25 wt.%
solution of DTMABr (Fluka, 98%) in ethanol and 70% of the
tetraethylammonium hydroxide 20 wt.% aqueous solution
(TEAOH) (Fluka) were added dropwise under stirring during 3 h.
Finally, the remaining TEAOH and the water were further added
dropwise to the milky solution which was then heated at 80 8C for
30 min to remove ethanol used in solution and produced in the
hydrolysis of TEOS. The pH of the resultant gel was 11.5. This gel
was transferred into Teflon-lined stainless-steel autoclave and
kept in an oven at 100 8C for 7 days under autogenic pressure. The
final solid was filtered, washed and dried at 60 8C overnight (as-
synthesized MCM-41). The molar composition of the gels
subjected to hydrothermal synthesis was as follows: Si/
metal = 120 and 20, TEAOH/Si = 0.30, surfactant/Si = 0.4, and
2.3. Flash vacuum pyrolysis reactions
Cfpv and fvp reactions were carried out in a Thermolyne 21100
furnace using a vycor glass of 30 cm long and a 1.2 cm i.d. reactor.
Temperatures were 350–450 8C, pressures of 10ꢀ2 Torr and
contact times of 10ꢀ2 s. Sample amounts were 18–20 mg and
catalyst:substrate ratios were equal to 25. After the experiments
were completed, the pyrolyzate was extracted with organic
solvents and submitted to purification and conventional analysis
(FTIR, 1H NMR, 13C NMR and GC/MS). All the catalytic materials
were pressed to wafers, crushed and sieved to obtain catalysts
particles of 10–20 mesh size and placed in the middle region of the
reactor using ceramic fibber as an inert support. Before the
reaction the catalysts were activated in the presence of air flow at
500 8C for 2–3 h.

100
E.L. Moyano et al. / Applied Catalysis A: General 373 (2010) 98–103
Table 1
Infrared spectra were recorded on
a Nicolet FTIR-55XC
Textural properties of MCM-41 catalysts.
spectrometer. 1H and 13C NMR spectra were recorded on a Bruker
Advance II 400 (1H at 400 MHz and 13C at 100 MHz) and Bruker FT-
200 (1H at 200 MHz) spectrometer using Cl3CD. Chemical shifts are
reported in parts per million (ppm) downfield from TMS. Gas
chromatography/mass spectrometry (GC/MS) analyses were per-
formed in a Shimadzu GC–MS-QP 5050 spectrometer equipped
with an VF column, using Helium as eluent at a flow rate of 1 ml/
min and a heating rate of 40 8C for 5 min and 40–280 8C for 35–
40 min.
Sample
Si/Ma
2
u
d1 0 0
a0 (nm)
Surface
area (m2/g)
(1 0 0) (8)
(nm)
Si-MCM-41
1
2.827
2.823
2.651
2.700
3.125
3.13
3.609
3.614
3.848
3.780
1790
1650
1440
1445
Ti-MCM-41 (120)
Ti-MCM-41 (20)
Al-MCM-41 (20)
120
20
20
3.333
3.270
a
Si/M molar ratio in the synthesis gel, M = Ti, Al.
7H-Dibenzo[b,d]azepin-7-one (5), phenanthridine (6) and 1,2-
Dihydro-dibenzo[b,d]azepin-7-one (7) products were purified by
chromatographic column using silica gel 60 (Merck) and hex-
ane:ethyl acetate (90:10) as eluents. The identity of compound 6
was established by comparison with authentic sample and the 1H
NMR spectrum.
1-Phenyl-2-(phenylimino)ethanone (4) compound was only
found in the reaction crude, it could not be isolated probably
due to hydrolysis of the imino group during workup, so its
structure was postulated by comparison of its mass spectrum with
the spectrum found in the equipment database with 90% match.
2-Phenyl-4H-benzo[d][1,3]oxazin-4-one (16) compound was
isolated from the reaction crude together with the minor isomer
2-phenyl-4H-benzo[d][3,1]oxazin-4-one (17). The GC/EM analysis
of 16 and the 1H NMR of the mixture of both isomers is included.
2-Phenyl-benzoxazole (11) compound was not isolated and the
structure was established by comparison with authentic sample by
GC/MS analysis.
materials. The structural regularity and phase purity were
confirmed by XRD. Fig. 1 shows the XRD patterns of the samples,
which are typical of MCM-41 structures. All the catalysts exhibit a
main (1 0 0) reflection peak and weak two ascribed to (1 1 0) and
(2 0 0) reflections. However, when the metal cation is introduced
into the material, a decrease in the intensity of the first diffraction
peak besides an evident broadening for all peaks can be observed,
indicating a reduction in the long-range order of the structure.
Thus, the presence of the metal cation in the synthesis gel changes
in some way the good organization of the silicate species with the
surfactant molecules, being possible to suggest that there is a
maximum amount of metal which can be incorporated into the
mesostructure. Above this value the formation of ordered MCM-41
structures could not occur. On the other hand, comparing the Ti-
MCM-41 and Al-MCM-41 samples which possess the same molar
relation Si/metal in the synthesis gel, it can be observed that the
aluminum cation can be incorporated into the structure causing
minor structural disorder than the titanium cation. This feature
could be explained taking into account the differences in the
atomic radio of titanium and aluminum. The substitution of the
larger Ti atom in place of Si atom would distort the geometry of the
silica framework in a higher degree, leading to a large structure
deformation. Therefore, the size of the metal ions can affect the
pore formation during synthesis and consequently the pores
structure quality. Finally, no diffraction peaks in the region of
higher angles (10–508) could be observed suggesting that MCM-41
samples are pure phases [8,9].
Characterization data for the identified products are the
following:
7H-Dibenzo[b,d]azepin-7-one (5). IR (KBr): 1762, 1707 cmꢀ1; 1H
NMR (400 MHz):
13C (100 Hz):
d 8.38 (dd, J = 1.6 and 7.8 Hz, 1H), 7.37 (m, 8H);
= 121.9, 123.2, 124.7, 128.3, 128.8, 129.1, 129.3,
130.0, 132.1, 136.8, 159.8, 161.2, 193.5; GC/MS: tr: 5.606 min, M+
d
207 (39), 179 (100), 152 (12), 76 (38), 50 (27).
Phenanthridine (6). 1H NMR (400 MHz):
d 9.33 (s, 1H), 8.62 (m,
2H), 8.22 (d, J = 7 Hz, 1H), 8.08 (d, J = 8 Hz, 1H), 7.89 (t, J = 7 Hz, 1H),
7.65–7.82 (m, 2H), 7.48–7.59 (m, 1H). GC/MS: tr: 5.206 min, M+
179 (100), 151 (15), 152 (10), 76 (14), 63 (12).
In a previous work [9] we have been demonstrated by UV–vis
diffuse reflectance (DRUV–vis) spectroscopy that the titanium, in
1,2-Dihydro-dibenzo[b,d]azepin-7-one (7). 1H NMR (200 MHz,
acetone-d6):
d 7.56–7.44 (m, 3H), 7.34–7.23 (m, 3H), 6.95 (d,
J = 8 Hz, 1H, 2H), 6.79 (t, J = 7 Hz, 1H), 6.10 (s, 1H), 3.22 (dd,
J = 17 Hz, 1H). GC/MS: tr: 7.595 min, M+ 209 (100), 180 (100), 152
(14), 90 (11).
1-Phenyl-2-(phenylimino)ethanone (4). GC/MS: tr: 5.899 min, M+
209 (14), 105 (50), 104 (100), 77 (60).
2-Phenyl-4H-benzo[d][1,3]oxazin-4-one (16). GC/MS: tr:
7.455 min, M+ 223 (100), 179 (60), 146 (20), 105 (71), 77 (61).
2-Phenyl-4H-benzo[d][3,1]oxazin-4-one (17). GC/MS: tr:
7.945 min, M+ 223 (42), 195 (100), 167 (23), 63 (52). Mixture
of 16 and 17: 1H NMR (400 MHz):
d 8.58 (dd, J = 2 and 9 Hz, 1H),
8.32 (dd, J = 2 and 8 Hz, 2H), 8.25 (dd, J = 2 and 8 Hz, 2H), 7.98–7.48
(m, 13H).
2-Phenyl-benzoxazole (11). GC/MS: tr: 4.673 min, M+ 195 (100),
167 (17), 63 (18).
3. Results and discussion
3.1. Synthesis and characterization of the MCM-41 catalysts
Table 1 summarizes the textural properties of the different
MCM-41 samples prepared in this study. As it can be seen,
although the incorporation of a metal cation into the structure
decreases the specific surface area to some extent compared with
pure silica MCM-41, almost all the catalysts prepared possess large
surface areas above 1400 m2/g, being typical of M41S group
Fig. 1. XRD patterns of the samples: (a) Si-MCM-41, (b) Ti-MCM-41 (120), (c) Ti-
MCM-41 (20) and (d) Al-MCM-41 (20).

E.L. Moyano et al. / Applied Catalysis A: General 373 (2010) 98–103
101
Table 2
the Ti-MCM-41 catalysts, is incorporated inside the silica frame-
Fvp and cfvp experiments of acylbenzotriazole 1.
work. An intense UV–vis band at 210 nm present in all the samples
clearly indicate that most of Ti species are isolated and in
tetrahedral coordination. Moreover, compared to the bulk anatase,
the lack of an absorption band characteristic of octahedral extra-
framework titanium at about 300–330 nm in all the samples
suggests that no separated titania phase (anatase) is formed during
the synthesis process. The 27Al solid-state MAS NMR spectrum of
the Al-MCM-41 sample reported elsewhere [8] reveals that the
aluminum was also incorporated mainly with tetrahedral coordi-
nation in the framework of MCM-41 prepared by us.
Catalyst
None
T (8C)
Conversion (%)
Selectivity (%)
5
6
Others
400
450
500
52.0
76.0
55.7
47.4
30.0
3.8
6.6
40.4a
46.1a
55.0b
100.0
15.0
Ti-MCM-41 (120)
Ti-MCM-41 (20)
350
64.0
60.9
3.1
35.9b
300
350
400
77.0
80.0
7.8
16.3
37.0
3.9
16.3
9.0
88.3a
67.5b
54.0b
100.0
Al-MCM-41 (20)
350
400
450
83.0
98.0
7.2
56.1
87.0
8.4
8.2
84.3a
35.7b
3.0b
3.2. Flash vacuum pyrolysis of 2-(1H-1,2,3-benzotriazol-1-yl)-
phenylethanone
100.0
10.0
Si-MCM-41
350
93.0
15.1
11.8
73.1b
The results of the cfvp experiments carried out on Al-MCM-41
(Si/Al = 20), Ti-MCM-41 (Si/Ti = 20 and 120) and pure Si-MCM-41
as well as the ones obtained in the homogeneous system [23], are
shown in the Table 2. Schemes 1 and 2 depict the formation of main
and minor products, respectively.
a
Minor products: 4, 7, 11, 16, 17, 18, 20 (see Schemes 1 and 2) and benzonitrile.
Minor products: 4, 11, 16, 17, 18, 20 and benzonitrile (see Schemes 1 and 2).
b
When the cfvp experiments were carried out at 350 8C using
pure Si-MCM-41 the conversion of acylbenzotriazole 1 was almost
total (93%). Thus, the reaction temperature was lowered in about
150 8C regarding the homogeneous systems. However, the
450 8C. According to these results, it is possible suggest that the
metal cation nature in the MCM-41 framework influences strongly
on the azepinone 5 formation.
The formation of compound 5 can be explained as shown in
Scheme 1 by initial extrusion of N2 to give intermediates 2 and 3.
Intramolecular cyclization of these intermediates followed by
hydrogen loss (path a), direct cyclization reaction of the
intermediate 1-phenyl-2-(phenylimino)ethanone (4) detected in
the crude by GC/MS (path b) and/or hydrogen loss from
dibenzoazepinone 7 (path c) could be the possible ways to form
5. The reaction proposed in the path b was also observed in flash
pyrolysis of 1-benzylbenzotriazole at 500–600 8C, where the
benzaldehyde N-phenylimine intermediate was converted to
phenanthridine (6) [26]. As the temperature was raised, the
decarbonylation reaction of azepinone 5 afforded the aromatic ring
6 (Table 1). It should be mentioned that the elimination of CO in
conventional fvp systems usually takes place at higher tempera-
tures (T > 600 8C) [27]; however in this case, the reaction could be
selectivity towards 7H-dibenzo[b,d]azepin-7-one
5 was not
improved (ꢁ15%) and many products (volatile and non-volatile
compounds) were detected.
When the reactions were performed over the Ti-MCM-41(20)
material, the total conversion of 1 was achieved at 400 8C with
dibenzoazepinone 5 and phenanthridine 6 as main products.
Although the selectivity to 5 (37%) was comparable to that of
homogeneous systems, in these cfvp reactions, the total conversion
value could be reached at 100 8C below. If we compare these results
with those achieved in the thermal reactions performed over Al-
MCM-41, this catalyst showed the best catalytic performance,
reaching an almost total conversion of starting material at 400 8C
with a selectivity towards azepinone 5 of 56%. Moreover, the
highest selectivity to 5 (87%) was obtained on this same catalyst at
Scheme 2. Formation of compounds 11, 16, 17 and 20 from acylbenzotriazole 1.

102
E.L. Moyano et al. / Applied Catalysis A: General 373 (2010) 98–103
favored by the assistance of metal sites and the formation of the
stable aromatic phenanthridine.
When Ti-MCM-41 catalysts were used, it was found that at 350–
400 8C, the selectivity to 5 was closely similar to those seen in the
homogeneous fvp system. Nevertheless, the best performance was
observed for the Al-MCM-41 catalyst with selectivity to 5 of 87% at
450 8C, as it was previously described.
In all the pyrolyzates, other minor products were also detected:
1-phenyl-2-(phenylimino)ethanone (4), 1,2-dihydro-dibenzo[b,-
d]azepin-7-one (7), 2-phenyl-benzoxazole (11), 2-phenyl-4H-3,1-
benzoxazin-4-one (16), 2-phenyl-4H-1,3-benzoxazin-4-one (17),
biphenylene (18), benzoxazole (20) and benzonitrile (Scheme 2).
Compound 7 could be formed from diradical 2 by cyclization
followed by 1,5-hydrogen shift, from carbene 3 by C–H insertion
followed by 1,3-H migration, or from both processes. It was
observed that this compound afforded ring fragmentation at
higher temperatures to give benzonitrile.
The catalytic activity of MCM-41 solids can be attributed to
different interactions between the solid and the substrate in the
gas phase. First, silanol groups can interact with the carbonyl group
and even with the nitrogen atom in benzotriazole 1, as it is known
that carbonyl and amino compounds can effectively interact with
mesoporous molecular sieves [44]. Second, metal sites can
facilitate the coordination environment increasing the reactivity
of substrates. If we compare the decarbonilation reaction from
azepinone 5 to give phenanthridine 6, it can be seen that Ti-
materials show a better selectivity than aluminum materials for
this transformation. A tentative explanation for this behavior could
be the high affinity of titanium species for carbonyl compounds
[44]. Then, Al site appears to be more selective to the formation of
azepinone 5 inhibiting competing reactions. Third, acid sites can
catalyze nitrogen extrusion reaction, thus lowering the reaction
temperature, as described for other denitrogenation processes
[17,45,46]. In this manner, the better performance of Al-MCM-41
for the formation of azepinone would be in agreement with the
enhanced acid properties of this material.
The formation of benzoxazole 11 and benzoxazinones 16 and 17
is assumed to take place through the intermediate generated by
diazomethane elimination from zwitterionic benzotriazole
8
(Scheme 2). The presence of diazomethane was confirmed by
direct GC/MS analysis of the volatile fraction in the reaction
mixtures. Like other zwitterionic species achieved under flash
vacuum pyrolysis conditions [28], intermediate 8 can be formed
from 1 by migration of the acyl group to N3 in the benzotriazole
ring. This type of rearrangement was also found in the thermolysis
of acyloxy-benzotriazoles [29]. 2-Phenylbenzoxazole (11) can be
directly generated from diradical 9 and/or carbene 10 in a way
similar to thermolysis of N-acylbenzotriazoles [30–32].
The formation of 16 and 17 seems more complicated. It is
proposed that diradical 9 can rearrange to diradical 12 which can
lose benzyne to give 13, a hybrid of oxazirene, acylnitrene and nitrile
oxide [33,34] (Scheme 2). The oxazirene might undergo different
types of reactions: (a) rearrangement to form benzisoxazole (19),
which rearranges to the more stable benzoxazole (20) [35]; (b)
elimination of oxygen to yield benzonitrile [36]; and (c) inter-
molecular radical coupling with diradical species 14 to form stable
adductslike16and17[37]. Compound 16couldalsobegeneratedby
a [2 + 3] cycloaddition of 1H-cyclopropabenzen-1-one (15) with the
nitrile oxide in a preparation similar to other benzoxazinones [38].
Although reactive species 13–15 could not be characterized
under experimental analysis conditions, there was some experi-
mental evidence of their formation. The presence of benzoxazole
(20) (detected by GC/MS in very small amounts) and benzonitrile
confirms the intermediacy of oxazirene and/or nitrile oxide.
Moreover, it is well known that these species and the nitrene
intermediate can be formed under thermal or photochemical
conditions [39]. Biphenylene (18) was also identified in the
pyrolyzate as a dimerization product of benzyne.
On the other hand, the existence of diradical 14 could be
verified by the appearance of dimerization products of MW = 208
as in other similar thermal transformations [40]. Additional
support for the course of this dimerization was found when
azepinone 5 was pyrolyzed under homogeneous conditions giving
products with MW = 208 and benzonitrile. Azepinone 7 could also
be a possible precursor of species 14 and 15.
The coupling reaction to afford benzoxazinones 16 (major
isomer) and 17 (minor isomer) is assumed to occur near the furnace
exit or in the cold trap in a way similar to the occurrence of other
dimerization reactions described in fvp systems [41,42]. In addition,
if the couplingreactiontookplaceinthe middleregionofthe furnace
where the catalyst was placed, hydrolysis of lactones 16 and 17
should be obtained. It is well established that benzoxazinones could
be attacked by OH groups similar to the OH groups available on the
surface of mesoporous materials [43]. However, hydrolysis products
were not detected in the reaction mixtures.
The influence of the surface area and Si/metal ratio (Table 1)
was also analyzed in the different solids used. When the Si-MCM-
41 catalyst with a surface area of 1790 m2/g was used, a high
degradation of 1 with a low selectivity to 5 was found. When cfvp
experiments with Ti-MCM-41 with different Si/Ti ratios (Si/Ti = 20
and 120) are compared, it can be seen that the catalyst with Si/
Ti = 120 and larger surface area showed a conversion of 1 lower but
with a selectivity towards 5 similar to the homogeneous fvp,
indicating that the amount of Ti is more significant than the surface
area for selectivity. This is more manifest when Ti-MCM-41 and Al-
MCM-41 with the same Si/Ti ratio (20) and similar surface areas
are compared. In this case the best selectivity for the formation of 5
was achieved with the second one, which shows that metal, rather
than the surface area, is crucial for selectivity.
It is also important to examine the mass balance; the results
found in different experimental conditions are compared in
Table 3. The amount observed of non-volatile products in the
pyrolyzate can result from severe cracking of the starting material
as well as from adsorption in the catalyst surface. When cfvp
reactions were carried out at lower temperatures, low yields of
non-volatile material were obtained in the cold trap due to
adsorption in the catalyst surface. As temperature was raised, the
desorption process favored the reaction and increased the yield of
the products in the cold trap. This effect is observed when we
compared mass balance values (Table 3) which are indicative of the
amount of the volatile products (homogeneous and heterogeneous
systems) and adsorbed material (heterogeneous systems) present
in the fvp reactions. In general, the mass balance percentages in the
homogeneous fvp decreases as temperature increases, confirming
the formation of volatile products by ring fragmentation reactions.
Table 3
Mass balance percentages for fvp reactions of 1 at different temperatures.
T (8C)
% mass balance
None
Al-MCM-41a
Si-MCM-41
Ti-MCM-41a
Ti-MCM-41b
300
350
400
450
500
100
100
90
63
65
75
79
–
–
76
–
19
25
50
–
–
55
–
An analysis of the effect of the catalyst composition on the
reactions follows. It is evident that both the conversion of 1 and the
products selectivity are strongly dependent on the metal located in
the MCM framework. When cfvp was performed with Si-MCM-41,
high conversion of 1 was achieved, but selectivity to 5 was poor.
89
–
–
78
–
–
–
a
Si/metal = 20.
Si/metal = 120.
b

E.L. Moyano et al. / Applied Catalysis A: General 373 (2010) 98–103
103
References
In the case of cfvp with Al- and Ti-MCM-41, the effect on the mass
balance differs, and this difference can be ascribed to a stronger
interaction of the catalyst with the substrate at lower tempera-
tures. Thus, Al and Ti active sites might enhance coordination
between mesoporous material and substrate through metal–
carbonyl and metal–nitrogen interactions. This is evident when the
mass balance value of 76% at 350 8C with Si-MCM-41 is compared
with the lower values observed for Ti- and Al-catalysts, showing an
effective coordination between metals and substrate. Since more
adsorbed material and more byproducts were obtained for Ti-
MCM-41 (Table 1) by secondary ring fragmentation reactions, it is
also evident that the interaction between Ti and organic
compounds is stronger than that with Al as a result of a longer
residence time.
All the catalysts could be reused approximately 20 times.
Although the catalytic performances of MCM-41 materials are
completely recovered after thermal activation, the Ti catalysts
required longer activation times than the other ones. This could be
attributed to the large amounts of organic compounds adsorbed on
the catalyst surface.
[1] C. Kresge, M. Leonowicz, W. Roth, J. Vartuli, J. Beck, Nature 359 (1992) 710–715.
[2] J. Beck, J. Vartuli, W. Roth, M. Leonowicz, C. Kresge, K. Schmidt, D. Olson, E.
Sheppard, S. McCullen, J. Higgins, J. Schlenker, J. Am. Chem. Soc. 114 (1992)
10834–10843.
[3] A. Corma, Chem. Rev. 97 (1997) 2373–2419.
[4] C. Berlini, M. Guidotti, G. Moretti, R. Psaro, N. Ravasio, Catal. Today 60 (2000) 219–
225.
[5] C. Berlini, G. Ferraris, M. Guidotti, G. Moretti, R. Psaro, N. Ravasio, Micropor.
Mesopor. Mater. 44–45 (2001) 595–602.
[6] J. Gallo, I. Paulino, U. Schuchardt, Appl. Catal. A: Gen. 266 (2004) 223–227.
[7] M. Guidotti, N. Ravasio, R. Psaro, G. Ferraris, G. Moretti, J. Catal. 214 (2003) 242–
250.
[8] G. Eimer, L. Pierella, G. Monti, O. Anunziata, Catal. Lett. 78 (1–4) (2002) 65–75.
[9] G. Eimer, S. Casuscelli, G. Ghione, M. Crivello, E. Herrero, Appl. Catal. A: Gen. 298
(2006) 232–242.
[10] G. Yranzo, J. Elguero, R. Flammang, C. Wentrup, Eur. J. Org. Chem. 12 (2001) 2209–
2220.
[11] H. McNab, Aldrichim. Acta 37 (2004) 19–26.
[12] G. Yranzo, E. Moyano, Curr. Org. Chem. 8 (2004) 1071–1088.
[13] I. Hodgetts, S. Noyce, R. Storr, Tetrahedron Lett. 25 (1984) 5435–5438.
[14] A. van der Waals, A. Klunder, F. van Buren, B. Zwanenburg, J. Mol. Catal. A: Chem.
134 (1998) 179–189.
[15] C. Lo´pez, F. Machado, K. Rodrı´guez, D. Arias, B. Me´ndez, M. Hasegawa, Catal. Lett.
62 (1999) 221–226.
[16] A. Derksen, A. van der Walls, M. van Ham, M. van Dongen, J. Borkent, R. de Gelder,
A. Klunder, B. Zwanenburg, Tetrahedron Lett. 41 (2000) 9189–9193.
[17] E. Moyano, G. Yranzo, J. Org. Chem. 66 (2001) 2943–2947.
[18] E. Moyano, M. del Arco, V. Rives, G. Yranzo, J. Org. Chem. 67 (2002) 8147–8150.
[19] W. Pela´ez, I. Gafarova, G. Yranzo, ARKIVOC 10 (2003) 262–272.
[20] R. Aitken, K. Hogson, A. Oyewale, J. Morrison, Chem. Commun. (1997) 1163–
1164.
[21] R. Aitken, P. Clasper, N. Wilson, Tetrahedron Lett. 40 (1999) 5271–5274.
[22] J. Denis, A. Gaumont, in: Y. Valle´e (Ed.), Gas Phase Reactions in Organic Synthesis,
Gordon and Breach Science Publishers, 1997, pp. 195–238.
[23] E. Moyano, P. Lucero, G. Eimer, E. Herrero, G. Yranzo, Org. Lett. 9 (2007) 2179–
2181.
[24] W. Peaston, G. Proctor, J. Chem. Soc. (C) (1968) 2481–2484.
[25] J. Moore, W. Theuer, J. Org. Chem. 30 (1965) 1887–1889.
[26] S. Barker, G. Jones, K. Randles, R. Storr, Tetrahedron Lett. 29 (1988) 953–954.
[27] R. Brown, in: H.H. Wasserman (Ed.), Pyrolytic Methods in Organic Chemistry, vol.
41, Academic Press, New York, 1980, pp. 164–168.
4. Conclusions
All the mesoporous catalysts synthesized in this study showed a
high specific surface area and a well-ordered porous structure.
However, as a metal cation (titanium or aluminum) is incorporated
into the mesostructure, a certain decrease in the structural
regularity can be observed. DRUV–vis and 27Al solid-state MAS
NMR analysis revealed that both titanium and aluminum were
introduced into the MCM-41 framework in tetrahedral positions
without the formation of amorphous phases.
This article showed the advantages of cfvp over fvp of 1 to
obtain azepinone 5 in good yield. The effect of different catalyst
compositions was studied. The Al-MCM-41 material showed the
best selectivity to azepinone (87%) at 450 8C; meanwhile the
conversion was almost total at 400 8C. It was also found that for
these reactions the presence of a metal site is more important than
a large surface area. While pure siliceous material leads to the
degradation of 1 without selectivity to 5, Ti- and Al- materials
display good selectivity to the formation of product 5. In Ti-
catalyzed reactions azepinone 5 was achieved with comparable
selectivities but at temperatures lower than those used in
homogeneous systems. However, this performance is lower than
that of aluminum containing materials. This methodology con-
stitutes a novel alternative to preparing dibenzazepinones like 5 in
a very simple and environmentally friendly way. On the other
hand, this work proposes an interesting strategy in order to modify
the type of active sites depending of the desired reaction.
[28] T. Pinho e Melo, M. Soares, A. Rocha Gonsalves, Tetrahedron Lett. 47 (2006) 791–
794.
[29] S. Nagarajan, S. Wilson, K. Rinehart, J. Org. Chem. 50 (1985) 2174–2178.
[30] A. Katritzky, X. Lan, J. Yang, O. Denisko, Chem. Rev. 98 (1988) 409–548.
[31] J. Druliner, J. Am. Chem. Soc. 90 (1968) 6879–6880.
[32] A. Maquestiau, D. Beugnies, R. Flammang, C. Wentrup, Org. Mass Spectrom. 25
(2005) 197–203.
[33] C. Wentrup, H. Bornemann, Eur. J. Org. Chem. (2005) 4521–4524.
[34] D. Poppinger, L. Radom, J. Pople, J. Am. Chem. Soc. 99 (1977) 7806–7816.
[35] K. Davies, R. Storr, P. Whittle, J. Chem. Soc. Chem. Commun. (1978) 9–10.
[36] H. Dib, A. Al-Awadi, Y. Ibrahim, O. El-Dusouqui, J. Phys. Org. Chem. 17 (2004) 267–
272.
[37] D. Poppinger, L. Radom, J. Am. Chem. Soc. 100 (1978) 3674–3685.
[38] J. Lown, K. Matsumoto, Can. J. Chem. 50 (1972) 584–590.
[39] M. Granier, A. Bacereido, H. Gru¨ tzmacher, H. Pritzkow, G. Bertrand, Angew. Chem.
Int. Ed. Engl. 29 (1990) 659–661.
[40] R. Brown, Pyrolytic Methods in Organic Chemistry, Academic Press, 1980, pp.
168–178.
[41] E. Hedaya, D. McNeil, J. Am. Chem. Soc. 89 (1967) 4213–4214.
[42] E. Duffy, J. Foot, H. McNab, A. Milligan, Org. Biomol. Chem. 2 (2004) 2677.
[43] V. Robinson, R. Spencer, Can. J. Chem. 66 (1988) 416.
[44] L. Yeung, F. Babonneau, US/France REU Program, 2003 www.chem.ufl.edu/REU/
main/participants/summaries/Yeung,%20Laurence%201.doc.
[45] A. Padwa, M. Filipkowski, D. Kline, S. Murpheree, P. Yeskes, J. Org. Chem. 58 (1993)
2061–2067.
Acknowledgements
This work was supported by the CONICET, the UTN-FRC and the
UNC-FCQ of Argentina.
[46] J. Garcı´a Ruano, M. Peromingo, M. Martı´n, A. Tito, Org. Lett. 8 (2006) 3295–3298.