Preparation and structure of novel hexaazaisowurtzitane cages

Article abstract of DOI:10.1002/chem.200501032

Hexaazaisowurtzitane or cage molecules have attracted attention concerning their synthesis because hexanitrohexaazaisowurtzitane (HNIW or CL20) is presently the most powerful energetic compound. The synthesis of hexaazaisowurtzitanes was considered to be

Full text of DOI:10.1002/chem.200501032

FULL PAPER
DOI: 10.1002/chem.200501032
Preparation and Structure of Novel Hexaazaisowurtzitane Cages
GrØgoire Herve,*[a] Guy Jacob,[a] and Roger Gallo[b]
Abstract: Hexaazaisowurtzitane or
cage molecules have attracted attention
concerning their synthesis because hex-
anitrohexaazaisowurtzitane (HNIW or
CL20) is presently the most powerful
energetic compound. The synthesis of
hexaazaisowurtzitanes was considered
to be limited solely to the condensation
of certain benzylamines with glyoxal.
Here, we present the synthesis and
characterization of seven novel non-
benzylic hexaazaisowurtzitanes, such as
hexapropargylhexaazaisowurtzitane.
The substituents on the six nitrogen
atoms are different to those of the
benzyl or substituted benzyl groups to
which previous syntheses were limited.
X-ray structures are given for the hex-
apropargyl and hexa-2-thienylmethy-
lene derivatives. Steric strains limit the
synthesis with a-substituted benzyl and
allyl derivatives. The reaction mecha-
nism and the role of the intermediate
diimines are discussed. Some of the
novel hexaazaisowurtzitanes are poten-
tial precursors of hexanitrohexaazaiso-
wurtzitane.
Keywords: cage
CL20 condensation reactions
glyoxal · hexaazaisowurtzitane
compounds
·
·
·
Introduction
addition, this elegant molecule proved to be a precursor of
one of the best modern energetic materials: hexanitrohexa-
azaisowurtzitane (HNIW or CL20).[4a–d] However, the re-
placement of the six benzyl groups by nitro functions, to
give the six nitramides in HNIW, could not be carried out
directly, but required three steps involving debenzylation–
acylation, nitrosation, and nitration.[5] A more efficient one-
stepsynthesis would be valuable for the industrial develop-
ment of the reaction; however, no direct displacement of a
benzyl moiety by a nitro function on HBIW has been re-
ported,[6] despite the potential interest of the reaction. On
the other hand, several functional groups are suitable for ni-
Organic chemists have always been attracted by fascinating
molecules, many of which are cyclic and striking because of
their geometry and symmetry.[1] The wurtzitane (iceane),
from the series of cage compounds having 12 atoms, was
identified in 1974.[2] Later, Nielsen et al. reported the syn-
thesis of
a new polyazapolycyclic ring system: the
2,6,8,10,12-hexaazatetracyclo[5.5.0.05,9.03,11]dodecane,[3] com-
monly named hexaazaisowurtzitane, owing to its similarity
with the isowurtzitane structure. The synthesis of hexaben-
zylhexaazaisowurtzitane (HBIW) is remarkable because it
builds a 12-atom polyclic cage compound in a one-pot cas-
cade reaction with three glyoxal and six benzylamines.[3] In
ꢀ
trolysis of secondary amines (transformation R2N Y!
[7a–f]
ꢀ
R2N NO2),
which could hopefully give the HNIW from
the corresponding hexafunctional-hexaazaisowurtzitane. The
major difficulty, however, is that the synthesis described by
Nielsen and co-workers,[3] despite several attempts, seems
limited to benzylamines, as stated in reference [3]: “apparent
uniqueness of benzylamines in the condensation with glyox-
al, which leads to hexabenzylhexaazaisowurtzitane”. To the
best of our knowledge, since the work of Nielsen, no amine
other than benzylamine (or substituted benzylamine) has
been reported to give this reaction. As confirmation, a
recent publication about the synthesis of new substituted
hexabenzylhexaazaisowutzitane from Klapçtke et al.[8] recal-
led that “the formation of hexaazaisowurtzitane cage system
is limited to certain benzylamines that condense in an acid-
catalyzed reaction with glyoxal”. We took these statements
as a challenge and describe here the preparation of novel
[a] G. Herve, G. Jacob
Energetic Molecule Synthesis Laboratory
SNPE MatØriaux EnergØtiques (SME), SNPE Group
Centre de recherches du Bouchet
9, rue Lavoisier, 91710 Vert-le-Petit (France)
Fax : (+33)164-991-175
E-mail: g.herve@snpe.com
[b] R. Gallo
FacultØ des Sciences de Saint JØrôme
UniversitØ Aix-Marseille III Paul CØzanne
UMR6180CNRS-case A62, Avenue Normandie-Niemen
13397 Marseille cedex 20 (France)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: 1H, 13C NMR
and DEPT 135 spectra of cage compounds 2a–g and elemental analy-
ses of 2a–g are available.
Chem. Eur. J. 2006, 12, 3339 – 3344
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3339

Table 1. Synthesis of hexasubstituted-hexaazaisowurtzitanes 2a–g.
hexaazaisowurtzitane molecules, made from primary amines
other than benzylamines (or substituted benzylamines), to-
gether with the study of their structure and of the mecha-
nism of synthesis of these cage compounds.
Run Cpd Catalyst
Solvent
T [8C] t [h] Yield[a]
1
2
3
4
5
6
7
8
2a
2b
2c
2d
2e
2 f
2g
2g
HCO2H
HCO2H
HCO2H
HCO2H
HCO2H
HCO2H
HCO2H
MeCN/H2O 10:1
MeCN/H2O 10:1
MeCN/H2O 10:1
MeCN
MeCN/H2O 10:1
MeCN
20
20
20
0
20
0
18 40
18 18
18 19
1
20
18 40
17
Results and Discussion
1
MeCN/H2O 10:1
20
20
36 25
36 62
[Yb(OTf)3] MeCN/H2O 30:1
The seven novel hexaazaisowurtzitane cage compounds 2a–
g were prepared by reacting glyoxal with seven primary
amines 1a–g (Scheme 1).
[a] Yield of isolated, purified product (%).
notably the yield from 25 to 62%. Compounds 2a–g were
fully characterized by spectroscopic methods; compounds
2a and 2 f by X-ray crystallographic analyses.
The cage compounds 2a–g were detected in their respec-
tive crude mixtures by performing NMR analyses. Results
of previous NMR analyses showed that cage compounds,
such as hexabenzyl- and hexa(substituted-benzyl)hexaazai-
sowurtzitane[3] have characteristic NMR proton and carbon
signals. The cage skeleton can be seen as a six-membered
boat ring and two near-planar five-membered rings linked
Scheme 1. Condensation of primary amines 1a–g on glyoxal; a: R=2-
ꢀ
ꢀ
ꢀ
thienyl CH2, b: R=PhCH=CH CH2, c: R=3-pyridyl CH2, d: R=CH2=
ꢀ
ꢀ
ꢀ
CH CH2, e: R=2-furfuryl, f: R=CH=C CH2, g: R=1-naphtyl CH2, h:
ꢀ
ꢀ
ꢀ
R=CH2=CH CHMe, i: R=PhCHMe, j: R=CH2=CH CMe2, k: R=
PhCMe2.
by a C C bridge; indeed, in the carbon and proton spectra,
two different peaks for each signal appear in a 1:2 ratio, cor-
responding to a six-membered ring and two five-membered
rings, due to the symmetry of the cage skeleton. Characteris-
tics of the proton spectra are, for instance, two singlets for
the two types of cage ring methine protons observed in a
4:2 ratio at around d=4.0 and 3.5 ppm (see Table 2). The
compounds are also recognizable by the signals for the adja-
cent methylene protons at d close to 4.0 ppm in an 8:4 ratio.
The reaction conditions and yields are given in Table 1.
Optimum yields were obtained by controlling kinetic pa-
rameters, such as temperature and reaction time. We ob-
served that a longer reaction time lowered the yield and in-
creased the ratio of the impurities as a result of reversible
reactions between the cage for-
mation and the intermediate
Table 2. 1H NMR spectral data of cage compounds 2 (208C) [d in ppm].
diimine. Despite applying the
best kinetic parameters for
each run, significant amounts of
impurities remained in the
crude products and made the
further isolation and purifica-
tion of the cage compounds dif-
ficult. The main impurities were
polymers, formed from the
major side reaction, which was
also mentioned by Nielsen,[3]
and which prevented crystalli-
zation of the cage compounds.
To purify the products, chroma-
tographic methods specific for
such compounds with low sta-
bility on acidic materials like
silica gel, were required. The purification procedures are de-
scribed in the Experimental Section, but no optimization of
reaction parameters by experimental design[9] to increase
the yields was attempted. The use of formic acid as an acidic
catalyst (runs 1–7) gave yields[11] in the range of 17–40%;
with runs 7 and 8, use of a water-tolerant Lewis acid catalyst
[Yb(OTf)3], recently proposed by Kobayashi,[10] increased
Cpd
Solvent
CDCl3
Cage ring
CH2
Aromatic
(Ac)Ethylenic
6.3–6.8 (12H)
2a
3.83 (s, 2H)
4.24 (s, 4H)
4.20 (s, 2H)
4.46 (s, 4H)
3.89 (s, 2H)
4.24 (s, 4H)[a]
3.85 (s, 2H)
4.16 (s, 4H)
3.57 (s, 2H)
4.24 (s, 4H)[a]
4.15 (s, 2H)
4.47 (s, 4H)
3.45 (s, 2H)
4.60 (s, 4H)[a]
4.24 (s, 4H)
4.35 (AB, 8H)
3.9–4.1 (m, 12H)
6.6–7.3 (18H)
2b
2c
2d
2e
2 f
2g
CDCl3
CDCl3
CDCl3
CDCl3
CDCl3
CDCl3
7.3–7.6 (30H)
7.3–7.8 (24H)
4.34 (AB, 8H)
4.45 (s, 4H)[a]
3.5–3.7 (m, 12H)
5.0–5.3 (12H)
2.21 (t, 4H)
5.7–6.0 (6H)
5.9–6.3 (12H)
7.35 (6H)
4.04 (s, 4H)[a]
4.08 (AB, 8H)
3.78 (m, 12H)
2.28 (t, 2H)
5.9 (4H)
6.9 (4H)
4.33 (AB, 8H)
4.69 (s, 4H)[a]
7.3–8.2 (34H)
[a] Singlet assignments may refer to either cage-ring protons or methylene protons.
However, we found that 13C NMR analyses were more rele-
vant and convenient for detecting the cage compounds in
the crude mixture. The 13C NMR spectrum shows two sig-
nals for the carbon atoms of the cage ring in an average 2:1
ratio in the 75–83 ppm range. The same observation was
made for the carbon atoms of the adjacent methylene
groups within the 50–60 ppm range, depending on the
3340
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3339 – 3344

Hexaazaisowurtzitane Cages
FULL PAPER
nature of the unsaturated group, and at around 41–42 ppm
for the propargyl derivative 2 f (see Table 3).
mediate diimines 3h–k in good yield (see Table 4), but did
not cyclize further to the cage compounds (Scheme 2).
Neither extension of the re-
action time for runs 9–12 to 6
or 11 days, nor the use triflate
ytterbium ([Yb(OTf)3]), which
had proved highly efficient as a
catalyst (Table 1, run 8), gave
the expected cage compounds.
Only diimines[13] and larger
amounts of polymers were
Table 3. 13C NMR spectral data of cage compounds 2 (208C) [d in ppm].
Cpd Solvent Cage ring CH2
Aromatic
(Ac)Ethylenic
2a
2b
2c
CDCl3
76.4, 81.6
51.3, 52.8 125.3, 125.4, 125.5, 126.0, 126.9, 127.0,
145.0, 146.1
55.7, 56.4 127.5, 127.6, 128.3, 128.4, 129.7, 129.8,
130.5, 130.7, 132.4, 132.6, 138.7, 138.8
54.3, 55.1 124.1, 124.3, 135.4, 135.6, 136.4, 137.1,
149.3, 149.4, 150.2, 151.1
acetone 78.2, 82.2
CDCl3
77.1, 81.8
2d
2e
CDCl3
CDCl3
77.7, 80.8
77.2, 80.3
56.4, 56.7
116.6, 117.6, 137.9, 138.4
formed. The formation of poly-
mers was evidence for diimine
49.2, 50.3 108.1, 108.4, 110.7, 142.3, 142.4, 154.0,
154.8
intermolecular additions rather
than intramolecular cycloaddi-
tions.
Condensation of glyoxal with
primary amines usually gives di-
2 f
2g
CDCl3
CDCl3
75.6, 81.6
77.7, 78.3
41.3, 42.2
71.4, 73.3, 80.8, 80.9
53.2, 55.2 124.8, 125.2, 125.3, 125.4, 125.6, 125.7,
125.8, 125.9, 126.6, 127.36, 127.39, 127.8,
128.5, 128.7, 132.4, 132.6, 134.0, 134.2,
135.6, 136.0
The isolated pure products showed expected assignments,
but for the methylene protons, slight differences to refer-
ence [3] were noted. For each cage compound, except for
propargylic 2 f, and allylic 2b and 2d, a 4H singlet and an
8H AB quartet were observed for the methylene com-
pounds, due to a complex multiplicity with the sp and sp2
protons, respectively. This differs from Nielsenꢁs observa-
tions because of higher resolution;[12] the phenomenon
might be explained by the steric effects of heterocyclic rings.
The free rotation around the methylene carbon atoms
linked to the two near-planar five-membered rings may be
quite limited, due to the size of the heterocyclic or aromatic
ring inducing a differentiation between the two methylene
protons. In contrast, a free rotation around the methylene
carbon atoms linked to the six-membered boat ring seems
to be easier because of the greater space available; the
proton spectra indeed show a sole singlet for the two meth-
ylenes.
Table 4. Synthesis of diimines 3h–k.
Run
Diimine
Yield[a]
9
3h
3i
3j
87
80
76
85
10
11
12
3k
[a] Yield calculated from crude product without further purification (%).
Scheme 2. Condensation of primary amines 1h–k on glyoxal.
carbinol amines and further diimines or tetrakisamino-
ethane, depending on the reaction conditions and reagent
structures (Scheme 3).[14]
Nielsen has proposed a mechanism for the formation of
HBIW via the dicarbinolamine, the diimine, and further cyc-
lization to the cage compound.[3] He checked that the di-
imine was the reaction intermediate. As a confirmation of
this result, we observed that performing the experiment of
run 7 (Table 1) over 12 h afforded the corresponding diimine
3 f, which, after isolation and introduction into the reaction
In addition, we reasoned that during the next anticipated
nitrolysis of the bridgehead benzylamine in HBIW, the
benzyl moiety should be displaced as a cationic leaving
groupby NO +. Therefore, beside the benzyl substituted on
2
the phenyl group, reported in previous studies,[3,8] the suc-
cessive replacement of the hydrogen atoms by methyl
groups in the a position of the benzyl should give secondary
and tertiary cations of increased stability and ease of dis-
placement as electrophilic leav-
ing groups. This should also be
expected for the a position of
the allyl. Therefore, we pre-
pared the amines 1h–k by
known synthetic procedures and
carried out the syntheses shown
in Scheme 1. Unfortunately,
with amines 1h–k, the reactions
failed and no cage compounds
were observed; more precisely,
the reactions reached the inter- Scheme 3. Formation of diimines.
Chem. Eur. J. 2006, 12, 3339 – 3344
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3341

G. Herve et al.
medium, gave the cage compound 2 f. If diimines are indeed
reaction intermediates for all isowurtzitane cage com-
pounds,[3,8] the lack of formation of 2h–k may be due to
steric effects that have already been advocated to prevent
the 1,2-dipole self-reaction of bulky imines to produce hexa-
hydrotriazines.[3,15] The formation of polymers is proof of
the inability of diimines to participate in a parallel reaction
toward cyclization. To evaluate the steric strains caused by
the methyl groupin the a position of benzyl- or allylamines
(1h–k), we calculated the strain energy of the hexabenzyl-
and hexaallyhexaazaisowurtzitane and their corresponding
a-methyl and a,a’-dimethyl derivatives by using Molecular
Mechanics[16,17](Table 5).
Table 5. Strain energy of the hexasubstituted-hexaazaisowurtzitane.
Substituent
Strain energy [kcalmolꢀ1
]
hexabenzyl
59.7
72.0
160.0
67.1
hexa(a-methylbenzyl)
hexa(a,a’-dimethylbenzyl)
hexaallyl
Figure 1. Stick plot of 2a. Hydrogen atoms are omitted for clarity.
hexa(a-methylallyl)
hexa(a,a’-dimethylallyl)
84.5
100.8
This provides only a rough evaluation of the steric strains
involved in the reaction because estimation of the activation
energy of cyclization from the diimines to the cage com-
pound should be made at the transition state of the rate-de-
termining step. This would involve laborious theoretical
work beyond the scope of this study, though some computa-
tional works have been performed and will be published
later. They also highlight significantly different levels of
energy at the transition state of the rate-determining steps
between cage compounds. Data from Table 3 show a large
increase in strain with one a-methyl, for both allyl and
benzyl groups, and an even larger strain (more specifically
with the benzyl) with the a,a’-dimethyl derivatives. Such
strain energies for a-methyl derivatives clearly indicate that
the activation energy to related cage compounds should be
very high. This seems a reasonable explanation for the lack
of formation of 2h–k.
The stick plots of the X-ray structures of hexa(2-thienyl-
methylene)hexaazaisowurtzitane 2a and hexapropargylhexa-
azaisowurtzitane 2 f are displayed in Figures 1 and 2, respec-
tively.
Both cages have the same structure: a bottom six-mem-
bered boat ring, two five-membered near-planar rings, and
two seven-membered rings having a chair conformation. As
observed by Nielsen,[3] in the six-membered ring, the two
benzyl methylene–nitrogen bonds are exocyclic, whereas in
the two five-membered rings, the four benzyl methylene–ni-
trogen bonds are alternatively exo- and endocyclic. Lengths
Figure 2. Stick plot of 2 f.
ꢀ
C C bond (from 1.58 to 1.60 Š for the hexathienylmethylene
compound 2a). Strains inside the five-membered rings are
also highlighted by torsion angles (1028 between the bridge
ꢀ
carbon and the two adjacent nitrogen atoms). The C N
bond lengths range on a larger scale, from 1.442 to 1.497 Š,
flanking the usual mean standard value (1.47 Š).[18] This
may be due to the difference in structure around the nitro-
gen atoms, exemplified by C-N-C angles in the range 123–
1508.
Conclusion
ꢀ
of the C C single bonds for the six-membered ring are
around 1.56 Š, a value slightly larger than a mean standard
Seven novel hexaazaisowurtzitanes were prepared. This syn-
thesis of hexaazaisowurtzitanes is not limited solely to the
condensation of certain benzylamines with glyoxal, as stated
previously.[3,8] A water-tolerant Lewis acid catalyst is effi-
cient for the synthesis. a-Substituted benzyl- and allylamines
[18]
ꢀ
C C value (1.54 Š),
indicating small internal strains in
the cage structure. However, the highest strain in the cage
ꢀ
structures seems to be located on the C C bridge bond link-
ing the two five-membered rings because of a fairly long
3342
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3339 – 3344

Hexaazaisowurtzitane Cages
FULL PAPER
which the temperature was maintained at between 0 and 58C. Reaction
details and work-upare given below.
afford diimines in good yield, but fail to give cage com-
pounds, owing to steric effects, as reasonably explained by
Molecular Mechanics. The direct displacement by nitronium
of the substituents on the nitrogen atoms of cage com-
pounds 2a–g to yield HNIW has given favourable results[19]
and was studied further here.
Hexa(2-thienylmethylene)hexaazaisowurtzitane (2a): The mixture was
stirred for 18 h at ambient temperature and then evaporated to dryness
under vacuum. The residue was dissolved in chloroform (50 mL), dried
over magnesium sulfate, filtrated, and concentrated under vacuum. The
crude product as a yellow resin was purified by flash chromatography on
silica gel (hexane/Et2O, 5:1). A white solid was obtained (40%). M.p.
1148C; 1H NMR (400 MHz, CDCl3): d=3.83 (s, 2H), 4.24 (s, 8H), 4.35
(AB, J=15 Hz, 8H), 6.6–7.3 ppm (m, 18H); 13C NMR (200 MHz,
CDCl3): d=51.3, 52.8, 76.4, 81.6, 125.3, 125.4, 125.5, 126.0, 126.9, 127.0,
145.0, 146.1 ppm; IR (KBr): n˜ =3105, 2916, 2853, 1685, 1654, 1439, 1348,
1314, 1289, 1274, 1211, 1176, 1130, 1036, 986, 923, 832, 695 cmꢀ1; MS
(CI+): m/z (%): 745 (91) [M+H]+, 650 (25) [M+H2ꢀCH2C4H3S]+; ele-
mental analysis calcd (%): C 58.0, H 4.8, N 11.3, S 25.9; found: C 57.4, H
4.8, N 11.8, S 26.4.
Experimental Section
Melting points were determined by using an Electrothermal 9100 capilla-
ry apparatus and are uncorrected. 1H and 13C NMR spectra were record-
ed by using Bruker AC 200 and Bruker Avance 400 NMR spectrometers
at 208C. Chemical shifts were referred to tetramethylsilane (TMS) for
13C and 1H (d values) and are reported in ppm. Coupling constants (J
values) are given in Hz. Multiplicity is indicated by the following abbrevi-
ations: s for singlet, d for doublet, t for triplet, m for multiplet, AB for
AB quartet. Infrared spectra were recorded by using an Avatar 320
FTIR Nicolet apparatus. Elemental analyses were performed in-house
for compounds 2a, 2b, 2d, and 2e, and by using a Thermo Finnigan
Flash EA 1112 at the University of Aix-Marseille III for the other cage
structures. Mass spectra (MS) were recorded by using a Nermag R10–
10H mass spectrometer by electronic impact (IE, 70 eV) and positive
chemical ionization (IC+, NH3), except for cage compound 1g, which
was characterized by positive ion spray (IS) by using an Api 165 Applied
Bio-System, with samples dissolved in methanol containing 0.1% tri-
fluoroacetic acid. Cage compounds were purified by flash chromatogra-
phy on basic aluminium oxide of 63–200 mm particle size, except for cage
compound 2a, which was purified on silica gel (Kieselgel of 70–200 mm
particle size). Reagent-grade acetonitrile (solvent) was used without fur-
ther purification. Commercially available chemicals, such as glyoxal
(40% in water), formic acid or ytterbium catalyst, and primary amines
were used as received. Cinnamylamine was prepared from commercial
cinnamyl chloride by using the Gabriel method.[20] As described in the lit-
erature, a-methylallylamine,[21] a,a’-dimethylallylamine,[22] and a,a’-dime-
thylbenzylamine[23, 24] were prepared by following known procedures.
Hexacinnamylhexaazaisowurtzitane (2b): The mixture was stirred for
18 h at ambient temperature. A very viscous resin was decanted off. The
mixture was stored for 1 h at 08C. The liquid layer was removed and the
remaining resin was dissolved in ether (150 mL). The organic layer was
washed three times with distilled water (3”20 mL), dried over magnesi-
um sulfate, filtrated, and concentrated under vacuum at 258C. A yellow
solid was obtained and was purified further by flash chromatography on
deactivated (6% H2O) basic alumina gel (hexane/Et2O, 5:2) giving a
white solid (18%). M.p. 608C; 1H NMR (400 MHz,CDCl3): d=3.97–4.04
(m, 12H), 4.20 (s, 2H), 4.46 (s, 4H), 6.3–6.8 (m, 12H), 7.3–7.6 ppm (m,
30H); 13C NMR (200 MHz, [D6]acetone): d=55.7, 56.4, 78.2, 82.2, 127.5,
127.6, 128.3, 128.4, 129.7, 129.8, 130.5, 130.7, 132.4, 132.6, 138.7,
138.8 ppm; IR (KBr): n˜ =3080, 3057, 3024, 2922, 2841, 1674, 1654, 1637,
1598, 1494, 1448, 1385, 1352, 1160, 1125, 1071, 912, 742, 692 cmꢀ1; MS
(CI+): m/z (%): 865 (5) [M+H]+.
Hexa(3-pyridinylmethylene)hexaazaisowurtzitane (2c): The mixture was
stirred for 18 h at ambient temperature and then concentrated under
vacuum (T<308C). Diethyl ether (150 mL) was added and the organic
layer was washed twice with distilled water (2”20 mL), dried over
sodium sulfate, filtrated, and concentrated under vacuum at 258C. A
clear orange oil was obtained and was purified further by flash chroma-
tography on deactivated (6% H2O) basic alumina gel. Some impurities
were removed by ether/chloroform (1:1) elution. A white gummy product
containing a small amount of impurities was obtained by chloroform/
NEt3 (10:0.1) elution (19%); 1H NMR (400 MHz, CDCl3): d=3.89 (s,
2H), 4.24 (s, 4H), 4.33 (AB, J=13 Hz, 8H), 4.45 (s, 4H), 7.3–8.7 ppm (m,
24H); 13C NMR (200 MHz, CDCl3): d=54.3, 55.1, 77.1, 81.8, 124.1, 124.3,
135.4, 135.6, 136.4, 137.1, 149.3, 149.4, 150.2, 151.1 ppm; IR (KBr): n˜ =
3030, 2928, 2853, 1698, 1592, 1577, 1478, 1423, 1359, 1334, 1319, 1219,
1170, 1126, 1028, 994, 913, 795, 714 cmꢀ1; MS (CI+): m/z (%): 715 (100)
[M+H]+.
X-ray structure determination: Compound 2 f was analyzed by M. Giorgi
at the University of Aix-Marseille III. Crystal data: Mw =792.98, triclinic,
colorless crystal (0.4”0.4”0.3 mm3), a=9.3010(4) b=12.8170(2) c=
19.0050(9) Š, a=96.520(3) b=98.121(2) g=92.205(3)8, V=2225.0(2) Š3,
space group P1, Z=2, 1calcd =1.184 gcmꢀ3, mMoKa =0.073 mmꢀ1, 10146 re-
¯
flections measured at 293 K (Bruker-Nonius Kappa CCD diffractome-
ter:[25] 1808 f scan, 28 rotation/frame) in the 2.22–26.318 q range, 8020
unique reflections, 541 parameters refined on F2 by using 6389 [F2 >4sF2]
reflections [Shelxl][26] to final indices R[F2>4sF2]=0.057, wR=0.133 [w=
1/[s2(Fo2)+(0.0459P)2 +0.9094P], in which P=(Fo2 +2F2c)/3]. Refinement
details: compound 2 f crystallized with two independent molecules in the
asymmetric unit. All hydrogen atoms were introduced in theoretical posi-
tions, included in the calculations, but not refined. The last residual Four-
ier positive and negative peaks were equal to 0.19 and ꢀ0.155, respec-
tively. X-ray analysis was performed for compound 2a by G. Pꢂpe at the
University of Aix-Marseille II. Crystal data: Mw =745.07, monoclinic, col-
orless crystal (0.4”0.3”0.3 mm3), a=12.568(2) b=15.827(2) c=
18.923(3) Š, b=106.32 g=92.205(3)8, V=36124(9) Š3, space group P21/
n, Z=4, 1calcd =1.370 gcmꢀ3, mMoKa =0.415 mmꢀ1, 6899 reflections, R[F2 >
4sF2]=0.0683. X-ray data were processed as described above.
Hexaallylhexaazaisowurtzitane (2d): The mixture was stirred for 1.5 h at
28C. The product precipitated in the mixture and was isolated as a white
solid by filtration (25%). M.p. 478C; 1H NMR (400 MHz, CDCl3): d=
3.54–3.63 (m, 12H), 3.85 (s, 2H), 4.16 (s, 4H), 5.0–5.3 (m, 12H), 5.7–
6.0 ppm (m, 6H); 13C NMR (200 MHz, CDCl3): d=56.4, 56.7, 77.7, 80.8,
116.6, 117.6, 137.9, 138.4 ppm; IR (KBr): n˜ =3075, 2998, 2968, 2951, 2921,
2847, 2824, 1640, 1458, 1442, 1416, 1389, 1345, 1317, 1294, 1271, 1242,
1183, 1168, 1154, 1110, 1069, 993, 959, 928, 914, 897, 799, 675, 623, 588,
526, 470, 416 cmꢀ1
;
MS (70 eV, EI): m/z (%): 367 (100)
[MꢀCH2CHCH2]+,
326
(4)
[Mꢀ2(CH2CHCH2)]+,
285
(20)
[Mꢀ3(CH2CHCH2)]+, 244 (15) [Mꢀ4(CH2CHCH2)]+; MS (CI+): m/z
(%): 409 (100) [M+H]+, 368 (10) [M+HꢀCH2CHCH2]+; elemental anal-
ysis calcd (%): C 70.6, H 8.8, N 20.6; found: C 70.8, H 9.0, N 20.7.
CCDC 268281 and 265636 contain the supplementary crystallographic
data for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data
request/cif.
Hexafurfurylhexaazaisowurtzitane (2e): The mixture was stirred for 18 h
at ambient temperature and then concentrated under vacuum (T<
308C). Chloroform (200 mL) was added and the organic layer was
washed twice with distilled water (2”20 mL), dried over magnesium sul-
fate, filtrated, and concentrated under vacuum at 258C. An orange oil
was obtained and was purified further by flash chromatography on deac-
tivated (3% H2O) basic alumina gel (hexane/Et2O, 5:2) giving a white
solid (40%). M.p. 978C; 1H NMR (400 MHz, CDCl3): d=3.57 (s, 2H),
2,4,6,8,10,12-Hexasubstituted-2,4,6,8,10,12-hexaazatetracyclo[5.5.0.0.0]do-
decane (hexasubstituted-hexaazaisowurtzitane): General procedure: over
a 10–15 min period, glyoxal (4.5 g, 40% aqueous solution, 0.031 mol) was
added dropwise to
a solution of primary amine (0.093 mol), water
(10 mL), formic acid (0.4 g, 9.3 mmol) in acetonitrile (100 mL), during
Chem. Eur. J. 2006, 12, 3339 – 3344
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3343

G. Herve et al.
4.04 (s, 4H), 4.08 (AB, J=14 Hz, 8H), 4.24 (s, 4H), 5.9–6.3 (m, 12H),
7.35 ppm (m, 6H); 13C NMR (200 MHz, CDCl3): d=49.2, 50.3, 77.2, 80.3,
108.1, 108.4, 110.7, 142.3, 142.4, 154.0, 154.8 ppm; IR (KBr): n˜ =3144,
3018, 2946, 2917, 2855, 1600, 1507, 1450, 1384, 1352, 1336, 1288, 1273,
1244, 1216, 1174, 1161, 1147, 1124, 1075, 1060, 1018, 1010, 992, 945, 918,
897, 885, 837, 810, 767, 740, 728, 670, 654, 600 cmꢀ1; MS (70 eV, EI): m/z
(%): 567 (4) [MꢀCH2C4H3O]+; MS (CI+): m/z (%): 649 (100) [M+H]+,
567 (10) [MꢀCH2C4H3O]+, 487 (10) [M+Hꢀ2(CH2C4H3O)]+; elemental
analysis calcd (%): C 66.7, H 5.5, N 13.0; found: C 66.4, H 5.6, N 12.9.
[7] a) J. H. Robson, J. J. Reinhart, J. Am. Chem. Soc. 1955, 77, 2453–
2455; b) “Industrial and Laboratory Nitrations”: E. E. Gilbert, J. R.
Leccacorvi, M. Warnan, ACS Symp. Ser. 1976, 22, 327–340; c) V. T.
Ramakrishnan, M. Vedachaban, J. H. Boyer, Heterocycles 1990, 31,
479–480; d) “Nitration: Recent Laboratory and Industrial Develop-
ments”: R. W. Millar, ACS Symp. Ser. 1996, 623, Chap. 16; e) “Nitra-
tion: Recent Laboratory and Industrial Developments”: P. F. Pago-
ria, A. R. Mitchell, R. D. Schmidt, C. L. Coon, E. S. Jessop, ACS
Symp. Ser. 1996, 623, Chap. 12; f) G. Piacenza, G. Jacob, A. Girard,
H. Graindorge, R. Gallo, Int. Annu. Conf. ICT 1997, 124, 1–14.
[8] T. M. Klapçtke, B. Krumm, H. Piotrowski, K. Polborn, G. Holl,
Chem. Eur. J. 2003, 9, 687–694.
[9] a) G. E. P. Box, W. G. Hunter, J. S. Hunter in Statistics for Experi-
menters, John Wiley, New York, 1978; b) R. Carlson in Design and
Optimization in Organic Synthesis, Elsevier, Amsterdam, 1992;
c) for a recent example, see: H. R. Bjorsvik, L. Ligori, J. A. Vedia
Meriners, J. Org. Chem. 2002, 67, 7493–7500.
[10] a) S. Kobayashi, I. Hachiya, J. Org. Chem. 1994, 59, 3590–3596;
b) S. Kobayashi, K. Manabe, Acc. Chem. Res. 2002, 35, 218–225.
[11] Regarding the yields obtained with benzylamines and derivatives:
our values (17–62%) are comparable to the values obtained by Niel-
sen in MeOH (ref. [3], 11–64%); smaller than the values by the
same author in acetonitrile (ref. [3], 24–80%); larger than the
values reported by Klapçtke (ref. [8], 5–32%).
Hexapropargylhexaazaisowurtzitane (2 f): The mixture was stirred for 1 h
at 08C then concentrated under vacuum (T<308C). Dichloromethane
(150 mL) was added and the organic layer was washed twice with distil-
led water (2”20 mL), dried over magnesium sulfate, filtrated, and con-
centrated under vacuum at 258C. A deep-red oil was obtained and was
purified further by flash chromatography on deactivated (3% H2O) basic
alumina gel (hexane/Et2O, 7:3) giving a white solid (17%). M.p. 1148C;
1H NMR (400 MHz, CDCl3): d=2.21 (t, J=2 Hz, 4H), 2.28 (t, J=2 Hz,
2H), 3.78 (m, 12H), 4.15 (s, 2H), 4.47 ppm (s, 4H); 13C NMR (200 MHz,
CDCl3): d=41.3, 42.2, 71.4, 73.3, 75.6, 80.8, 80.9, 81.6 ppm; IR (KBr): n˜ =
3291, 3213, 2952, 2941, 2925, 2878, 2819, 1637, 1437, 1384, 1354, 1340,
1297, 1187, 1166, 1137, 1074, 1015, 999, 925, 895, 800, 675, 659, 631 cmꢀ1
;
MS (70 eV, EI): m/z (%): 357 (100) [MꢀCH2CCH]+; MS (CI+): m/z
(%): 397 (100) [M+H]+; elemental analysis calcd (%): C 72.7, H 6.1, N
21.2; found: C 72.7, H 6.2, N 19.8.
[12] A 1H NMR analysis of HBIW was performed in CDCl3 by using a
400 MHz spectrometer; the HBIW NMR spectrum indeed shows an
AB quartet as observed for the new hexaazaisowutzitane com-
pounds described: d=3.59 (s, 2H), 4.04 (s, 4H), 4.09 (AB, J=12 Hz,
8H), 4.17 (s, 4H), 7.20–27.24 ppm (m, 30H).
[13] An old US patent describes the condensation of a-methylbenzyla-
mine and a,a’-dimethylbenzylamine with glyoxal to produce dii-
mines: M. D. Hurwitz US-2582128, 1952.
[14] a) G. F. Whitfield, R. Johnson, D. J. Swern, J. Org. Chem. 1972, 37,
95–99; b) J. M. Kliegman, R. K. Barnes, Tetrahedron 1970, 26,
2555–2560; c) M. Chaykowski, W. M. Koppes, T. P. Russel, R. Gilar-
di, C. George, J. L. Flippen-Anderson, J. Org. Chem. 1992, 57,
4295–4297; d) J. M. Kliegman, R. K. Barnes, J. Org. Chem. 1970, 35,
3140–3143; e) M. R. Crampton, H. Javed, M. Ross, G. Fergusson, J.
Chem. Soc. Perkin Trans. 1 1993, 923–929; f) K. Karaghiosoff, T. M.
Klapçtke, A. Michailovski, H. Nçth, M. Suter, Propellants Explos.
Pyrotech. 2003, 28, 1–6.
Hexa(1-naphtylmethylene)hexaazaisowurtzitane (2g): Over a 10–15 min
period, glyoxal (4.61 g, 40% aqueous solution, 0.032 mol) was added
dropwise to a solution of 1-aminomethylnaphtalene (14.97 g, 0.093 mol),
water (5 mL), ytterbium trifluoromethylsulfonate (2.07 g, 3.2 mmol), one
drop of antifoam A (Fluka supplier) in acetonitrile (150 mL), during
which the temperature was maintained at between 0 and 58C. The mix-
ture was stirred for three days at ambient temperature. A white solid
product (62%) was filtrated and dried over P2O5. M.p. 244–2458C
(decomp); 1H NMR (400 MHz, CDCl3): d=3.45 (s, 2H), 4.33 (AB, J=
14 Hz, 8H), 4.60 (s, 4H), 4.69 (s, 4H), 5.90 (m, 4H), 6.89 (m, 4H), 7.3–
8.2 ppm (m, 34H); 13C NMR (200 MHz, CDCl3): d=53.2, 55.2, 77.7, 78.3,
124.8, 125.2, 125.3, 125.4, 125.6, 125.7, 125.8, 125.9, 126.6, 127.36, 127.39,
127.8, 128.5, 128.7, 132.4, 132.6, 134.0, 134.2, 135.6, 136.0 ppm; MS (IS):
m/z: 1009; elemental analysis calcd (%): C 85.7, H 6.0, N 8.3; found: C
85.2, H 6.0, N 8.3.
[15] T. Nielsen, R. L. Atkins, D. W. Moore, R. Scott, D. Mallory, J. M.
Laberge, J. Org. Chem. 1973, 38, 3288–3295.
[1] F. Vçgtle, Fascinating Molecules in Organic Chemistry, John Wiley,
Chichester, 1992.
[2] C. A. Cupers, L. Hodakowski, J. Am. Chem. Soc. 1974, 96, 4668–
4669.
[3] A. T. Nielsen, R. A. Nissan, D. J. Vanderah, C. L. Coon, R. D. Gilar-
di, C. F. George, J. Flippen-Anderson, J. Org. Chem. 1990, 55, 1459–
1466.
[16] Calculations were carried out with HYPERCHEM.6.03 (ref. [17])
by using the MM+ program with the option Fletcher-Reeves.
[17] HYPERCHEM.6, Hypercube, Gainesville (FL) 2000.
[18] J. Gordon, R. A. Ford in The Chemistꢀs Companion; John Wiley,
New York, 1972.
[19] G. Cagnon, G. Eck, G. HervØ, G. Jacob, US-260086, 2004; EP-
1479683, 2004.
[20] See, for example: J. C. Sheelan, W. A. Bolhofer, J. Am. Chem. Soc.
1950, 72, 2768–2788.
[21] J. D. Roberts, R. H. Mazur, J. Am. Chem. Soc. 1951, 73, 2509–2520.
[22] G. F. Hennion, C. V. Digiovanna, J. Org. Chem. 1965, 30, 2645–
2650.
[23] R. Popielarz, D. R. Arnold, J. Am. Chem. Soc. 1990, 112, 3068–
3082.
[24] P. C. Kamer, J. Noelte, M. Roeland, W. Drenth, J. Am. Chem. Soc.
1988, 110, 6818–6825.
[25] Bruker-Nonius, Kappa CCD Reference Manual, Nonius B. V., P. O.
Box 811, 2600 Av, Delft (The Netherlands), 1998.
[4] a) P. F. Pagoria, G. S. Lee, A. R. Mitchell, R. D. Schmidt, Thermo-
chim. Acta 2002, 384, 187–204; b) H. Bazaki, S. Kawabe, H. Miya,
T. Kodama, Propellants Explos. Pyrotech. 1998, 23, 333–336;
c) A. T. Nielsen in Chemistryof Energetic Materials (Eds.: G. A.
Olah, D. R. Squire), Academic Press, San Diego, 1991, pp. 95–124;
d) “Energetic Materials–Modeling of Phenomena, Experimental
Characterization, Environmental Engineering”: T. M. Klapçtke, B.
Krumm, G. Holl, M. Kaiser, Int. Annu. Conf. ICT 1999, 120–121.
[5] A. T. Nielsen, A. P. Chafin, S. L. Christian, D. W. Moore, M. P.
Nadler, R. A. Nissan, D. J. Vanderah, R. D. Gilardi, C. L. George,
J. L. Flippen-Anderson, Tetrahedron 1998, 54, 11793–11812.
[6] Recent procedures, other than the three-step synthesis described in
ref. [5], all use HBIW as the starting material and involve debenzy-
lation with formation of polyacylazaisowurtzitane intermediates:
a) T. Kodama, M. Tojo, M. Ikeda, US-6472525, 2002; b) A. J. Sand-
erson, K. Warner, R. B. Wardle, US-639130, 2002; c) R. B. Wardle,
J. C. Hinshaw, US-6147209, 2000; d) K. H. Chung, H. S. Kil, I. Y.
Choi, C. K. Chu, I. M. Lee, J. Heterocycl. Chem. 2000, 37, 1647–
1649.
[26] G. M. Sheldrick, SHELXS-97, Program for the Solution of Crystal
Structures, University of Gçttingen, Gçttingen (Germany), 1996.
Received: August 23, 2005
Published online: February 2, 2006
3344
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3339 – 3344