Reductive debenzylation of hexabenzylhexaazaisowurtzitane - The key step of the synthesis of polycyclic nitramine hexanitrohexaazaisowurtzitane

Article abstract of DOI:10.1007/s11172-007-0377-5

Main features of the reductive debenzylation of 2,4,6,8,10,12-hexabenzyl-2, 4,6,8,10,12-hexaazaisowurtzitane were studied. This process is the key step of the synthesis of 2,4,6,8,10,12-hexanitro-2,4,6,8,10,12-hexaazatetracyclo[5.5.0. 0^3,11.0

Full text of DOI:10.1007/s11172-007-0377-5

Russian Chemical Bulletin, International Edition, Vol. 56, No. 12, pp. 2370—2375, December, 2007
2370
Reductive debenzylation of hexabenzylhexaazaisowurtzitane — the key step
of the synthesis of polycyclic nitramine hexanitrohexaazaisowurtzitane

A. P. Koskin, I. L. Simakova, and V. N. Parmon
G. K. Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences,
5 prosp. Akad. Lavrent´eva, 630090 Novosibirsk, Russian Federation.
Fax: +7 (383) 330 8056. Eꢀmail: simakova@catalysis.ru
Main features of the reductive debenzylation of 2,4,6,8,10,12ꢀhexabenzylꢀ2,4,6,8,10,12ꢀ
hexaazaisowurtzitane were studied. This process is the key step of the synthesis of 2,4,6,8,10,12ꢀ
hexanitroꢀ2,4,6,8,10,12ꢀhexaazatetracyclo[5.5.0.03,11.05,9]dodecane (hexanitrohexaazaꢀ
isowurtzitane, CLꢀ20), a compound with unique energetic and explosive characteristics. The use
of the latter is restricted so far by the high cost of the twoꢀstep process of debenzylation during
which the compound is rapidly deactivated. The expensive Pd/C catalyst is deactivated in the
first step of the process, which limits the use of this polycyclic nitramine. The influence of the
solvent nature; loadings of the reactants, catalyst, and cocatalyst; the hydrogen pressure and
reaction temperature on the general features of the process and the yield of the target precursor
of CLꢀ20 was studied.
Key words: hexanitrohexaazaisowurtzitane, 2,4,6,8,10,12ꢀhexabenzylꢀ2,4,6,8,10,12ꢀ
hexaazatetracyclo[5.5.0.03,11.05,9]dodecane, debenzylation, palladiumꢀcontaining catalyst,
deactivation of Pd/C.
Scheme 1
In the recent decades, much attention in the area of
highꢀenergy materials chemistry has been given to invesꢀ
tigation of the properties of 2,4,6,8,10,12ꢀhexanitroꢀ
2,4,6,8,10,12ꢀhexaazatetracyclo[5.5.0.03,11.05,9]dodecane
(hexanitrohexaazaisowurtzitane, CLꢀ20) and search for
efficient methods for the synthesis of this polycyclic nitꢀ
ramine.1,2 The studies of the explosive and detonation
properties of CLꢀ20 and related compositions showed that
presently this compound is the most powerful stable exꢀ
plosive.3,4 However, due to the absence of an economical
industrial method of the synthesis of CLꢀ20 its wide use is
limited.1
The best known method for the synthesis of CLꢀ20 is
presented in Scheme 1. Based on this protocol, in 1990
the Thiokol corporation (USA) made a pilot unit for the
synthesis of CLꢀ20.5 The first step of the synthesis is the
construction of the hexaazaisowurtzitane framework usꢀ
ing the condensation of glyoxal 1 with benzylamine, and
the second step is the substitution of the benzyl groups at
the nitrogen atom for the nitro group. The direct nitration
of 2,4,6,8,10,12ꢀhexabenzylꢀ2,4,6,8,10,12ꢀhexaazatetraꢀ
cyclo[5.5.0.03,11.05,9]dodecane (2) does not afford CLꢀ20,
because the hexaazaisowurtzitane framework desintegrates
by nitrozative debenzylation. Therefore, the intermediate
step of the synthesis of precursors of nitramine CLꢀ20 is
necessary. This step is effected by means of the simultaꢀ
neous hydrodebenzylation—acetylation process during
which the C—N bonds are hydrogenolyzed on the pallaꢀ
dium catalysts followed by the acetylation of the forming
amines with acetic anhydride (2 → 3). The benzyl groups
Reagents, conditions, and yield: i. PhCH2NH2, MeCN, H+, 80%;
ii. Pd/C, H2, Ac2O; iii. Pd/C, H2, HCOOH; iv. HNO3,
NH4NO3.
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 12, pp. 2290—2295, December, 2007.
1066ꢀ5285/07/5612ꢀ2370 © 2007 Springer Science+Business Media, Inc.

Debenzylation of hexabenzylhexaazaisowurtzitane Russ.Chem.Bull., Int.Ed., Vol. 56, No. 12, December, 2007 2371
The relative rate of hydrogen absorption (v) was expressed as
the amount of hydrogen absorbed within the first hour of the
process referred to the amount of loaded compound 2.
at the nitrogen atoms of the sixꢀmembered ring (N(4) and
N(10)) are more stable in this process, and the second
step of hydrodebenzylation is carried out to substitute
them under more severe conditions (3 → 4). The subseꢀ
quent nitration of the synthesized 4,10ꢀdiformylꢀ2,6,8,12ꢀ
t e t r a a c e t y l ꢀ 2 , 4 , 6 , 8 , 1 0 , 1 2 ꢀ h e x a a z a t e t r a c y c ꢀ
lo[5.5.0.03,11.05,9]dodecane (4) affords CLꢀ20 in high
yield.6 At present the methods for the synthesis of the
polycyclic framework7,8 and nitration of the polyacetyl
hexaazaisowurtzitane derivatives9 have been studied rather
well and it is not difficult to use them. At the same time,
two catalytic steps of the process 2 → 3 → 4, which are
most complicated and most expensive in the synthesis of
CLꢀ20, are studied to considerably less extent. The main
problem of practical realization of these steps is the fast
deactivation of the Pd/C catalyst.
The composition of the components in the liquid reaction
mixture was determined by GC/MS on a VGꢀ7070 GC/MS
instrument (VG Analytical) on a quartz capillary column
(30 m × 0.2 mm) (Silicone SEꢀ30). The structures of the reacꢀ
tion products were determined by 1H and 13C NMR spectroscoꢀ
py. The spectra were recorded on a Bruker ACꢀ200 instrument
(1H NMR, 200.13 MHz; 13C NMR, 50.32 MHz) in CDCl3 and
DMSOꢀd6. The chemical shifts are presented in the δ scale
relative to Me4Si.
2,4,6,8,10,12ꢀHexabenzylꢀ2,4,6,8,10,12ꢀhexaazatetraꢀ
cyclo[5.5.0.03,11.05,9]dodecane (2) was presented by the Instiꢀ
tute of Problems of Chemical Energetic Technologies, Siberian
Branch of the Russian Academy of Sciences (Biisk, Russia),
m.p. 152—153 °C. 1H NMR (CDCl3), δ: 3.62 (s, 2 H, CH); 4.13
(br.d, 8 H, PhCH2); 4.08 (s, 4 H, PhCH2); 4.20 (s, 4 H, CH);
7.20—7.30 (m, 30 H, Ph). 13С NMR (CDCl3), δ: 56.3, 57.0
(both t, PhCH2 each); 77.0, 80.7 (both d, СН); 126.8, 126.9
(both d, pꢀCPh); 128.2, 128.3 (both d, mꢀCPh); 128.7, 129.3
(both d, oꢀCPh); 140.9 (s, ipsoꢀCPh).
The purpose of this work is to study the catalytic steps
of the debenzylation process and to search for a possibility
to enhance the yield of compound 4, viz., target precursor
of CLꢀ20, and improve the stability of the Pd/C catalyst.
2,6,8,12ꢀTetraacetylꢀ4,10ꢀdibenzylꢀ2,4,6,8,10,12ꢀhexaazaꢀ
tetracyclo[5.5.0.03,11.05,9]dodecane (3) was synthesized
by the reaction of compound 2 (3.4 g) and HCOOH (5 mL) in
DMF (10 mL) and PhBr (0.06 mL) in the presence of the
Pd(10%)/C catalyst (0.34 g) and molecular hydrogen (4 bar).
The yield was 2.12 g (86%), m.p. 318—321 °C. 1H NMR
(DMSOꢀd6), δ: 2.1 (m, 12 H, Me); 4.12 (s, 4 H, CH2); 5.51
(br.s, 4 H, CH); 6.58 (br.s, 2 H, CH); 7.34—7.44 (m, Ph, 10 H).
2,6,8,12ꢀTetraacetylꢀ4,10ꢀdiformylꢀ2,4,6,8,10,12ꢀhexaazaꢀ
tetracyclo[5.5.0.03,11.05,9]dodecane (4) was synthesized by the
reaction of compound 3 (2.12 g) and НСООН (10.6 mL) in the
presence of the Pd(10%)/C catalyst (0.34 g) and molecular hyꢀ
drogen (4 bar). The yield was 1.29 g (84%), m.p. 291—295 °C.
1H NMR (DMSOꢀd6), δ: 2.1 (m, 12 H, Me); 6.4 (br.d, 4 H,
CH); 6.49 (s, 2 H, CH); 8.34 (s, 2 H, CHO).
Experimental
The palladium catalysts for debenzylation were synthesized
by the method of palladium ion adsorption from an aqueous
solution of H2[PdCl4] (1 mol L–1) on the Sibunit carbon supꢀ
port (granular size 50—100 µm, specific BET surface Ssp(N2) =
320 m2 g–1, pore volume Vs = 0.86 mL g–1, maximum in the size
pore distribution curve at 4 nm) at room temperature followed
by the reduction to metallic palladium with a solution of sodium
formate. The content of the noble metal in the catalyst was ∼4,
∼6, and ∼10 wt.%. The percentage content of the metal on the
carbon support was determined by the Xꢀray spectral method on
a VRAꢀ30 fluorescence analyzer with the Cr anode of the Xꢀray
tube. The palladium catalysts were studied by transmission elecꢀ
tron microscopy (TEM) on a JEMꢀ2010 instrument (JEOL,
Japan) with an accelerating voltage of 200 kV and a limiting
resolution by the lattice of 0.14 nm. The average size of the
palladium particles was estimated from the histograms of the
palladium particle size distribution constructed by processing
TEM microphotographs.
Dimethylacetamide (DMA), DMF, and Nꢀmethylpyrroliꢀ
done were dried over CaH2 and distilled in vacuo. Acetic anꢀ
hydride was purified collecting the fraction with b.p. 139—140 °C.
The catalytic debenzylation of compound 2 was carried out
in a temperatureꢀcontrolled (0—100 °C) stainless steel autoꢀ
clave (volume 150 mL) equipped with an electromagnetic stirrer
(800—1000 rpm) and a system of gaseous hydrogen supply. The
volume of the absorbed gas was quantitatively measured using a
Sapfirꢀ22 instrument (ООО "Neotekhnologiya," Russia) and
computation of the experimental data. Catalytic experiments
were performed according to a described procedure.10 The staꢀ
bility of the catalyst in deactivation processes was studied loadꢀ
ing the spent Pd/C catalyst after two debenzylation steps
(2 → 3 → 4) into the next cycle. For this purpose, the spent
catalyst separated from the reaction mixture was successively
washed with formic acid, a 1% solution of Na2CO3, and distilled
water and dried in vacuo, and the catalytic activity was studied
using a standard procedure.
Results and Discussion
Identification of the step of the debenzylation process
making the main contribution to the deactivation of the
Pd/C catalyst. The initial substrate of debenzylation (2) is
unstable and decomposes easily at elevated temperatures
in acids and upon the interaction with acetic anhydride.
When crystalline compound 2 is dissolved in acetic anhydride
(or is heated in DMF), the solution darkens rapidly. The
1H NMR analysis of the obtained mixture indicates a
decrease in the intensity of signals at δ 3.5—4.5. This
interval contains the signals of the CH groups characterꢀ
istic of the hexaazaisowurtzitane framework. This process
is accelerated upon the addition of the Pd/C catalyst to
the solution; however, the mixture remains colorless, beꢀ
cause the colored products of the destruction of comꢀ
pound 2 are sorbed by the catalyst. In this case, the cataꢀ
lyst stored in a mixture of compound 2 with acetic anꢀ
hydride and DMF for 1 day was completely deactivated.
We carried out separately the reactions 2 → 3 and
3 → 4 on two samples of the same Pd/C catalyst. The

2372
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 12, December, 2007
Koskin et al.
hexaazaisowurtzitane products obtained after hydrogenolꢀ
ysis were isolated from the reaction mixture, their yield
was calculated, and the samples of the Pd/C catalyst were
washed, dried, and loaded into the reactor to undergo the
repeated cycle of hydrogenolysis. It turned out that the
hydrogenation of compound 3 to compound 4 on the
same catalyst can be carried out at least three times, whereꢀ
as the hydrogenation of compound 2 to compound 3 can´t
be repeated (Table 1). Therefore, the main contribution
to the catalyst deactivation is made by the first debenzylaꢀ
tion step. In the second debenzylation step, the destrucꢀ
tion of the hexaazaisowurtzitane framework occurs much
less intensively, because compound 3 is substantially more
stable than compound 2. In fact, the higher stability of the
framework of compound 3 makes it possible to use more
drastic reaction conditions for the replacement of the
benzyl groups by the formyl groups at the N(4) and N(10)
atoms and even to carry out the synthesis of CLꢀ20 by the
nitrozative debenzylation of compound 3 as shown in
Ref. 11.
We studied the composition of the liquid phase of the
reaction mixture after the first hydrodebenzylation step
by GC/MS. Benzylacetamide and several nonidentified
nitrogenꢀcontaining admixtures were found in addition to
the residual peaks of the reactants and the group of peaks
of the expected products of debenzylation (toluene, benꢀ
zene, and acetic acid). Benzylacetamide is formed due to
the hexaazaisowurtzitane framework opening followed by
the acylation of the destruction products.12 It is known
that benzylacetamide can form stable complexes with palꢀ
ladium, and the leading role of this compound in the
deactivation of the palladium catalyst was repeatedly pointꢀ
ed out in the literature.6 The preliminary treatment of the
Pd/C catalyst with benzylacetamide (synthesized by the
counter synthesis) results, in fact, in the complete deactiꢀ
vation of the catalyst (0.34 g of the palladium catalyst was
treated with 0.05 g of benzylacetamide). Nevertheless,
when performing the hydrodebenzylation 2 → 3, we
observed no correlation between the amount of benzylacꢀ
etamide in the reaction mixture (determined by GC/MS)
and the degree of deactivation of the catalyst (estimated
by the depth of the debenzylation process). It is most
likely that benzylacetamide is not the single and strongest
catalytic poison for the Pd/C catalyst formed in the reacꢀ
tion mixture. Earlier12 more than five different nitrous
residues of the hexaazaisowurtzitane framework formed by
hydrogenolysis have been identified. Each of the residues
can deactivate the Pd/C catalyst to this or another extent.
Effect of the solvent nature and metal loadings of the
reactants, cocatalyst, and catalyst on the course of the
process and the yield of the target precursor of CLꢀ20. The
catalytic debenzylation of compound 2 proceeds stepꢀ
wise, which was shown13 by the analysis of the reaction
mixture using HPLC and GC. Bromobenzene (PhBr) is
added to the reaction mixture to accelerate the first deꢀ
benzylation step. Bromobenzene is the source of hydroꢀ
gen bromide, which is the cocatalyst of the hydrogenolyꢀ
sis of the C—N bond. The examples of the kinetic curves
of hydrogen absorption by the reaction mixture for the
Pd/C catalysts with different amounts of the active metal
are shown in Fig. 1. The acceleration of the process in the
initial debenzylation step is presumably related to the acꢀ
cumulation of acids in the reaction mixture. As known,
the acids, namely, HBr and AcOH, promote the hydroꢀ
genolysis of the C—heteroatom bond.14 Then the rate of
hydrogen absorption decreases, because solubility of H2
in DMF decreases with the substitution of the benzyl
groups for the acetyl groups in the hexaazaisowurtzitane
substrate and, hence, the hydrogenolysis rate decreases.
In addition the Pd/C catalyst is gradually deactivated.
[H2]/mmol
25
Table 1. Determination of the contribution of the debenzylation
2
steps to the deactivation of the palladium catalyst
20
1
Reaction
Yield of target product (%)
15
10
5
Fresh
First
Second
2 → 3 a
82
85
69
0
77
0
—
75
—
3 → 4 b
2 → 3 → 4 c
a Loadings of the reactants: 3.4 g of compound 2, 0.34 g of
Pd(6%)/Sibunit, 5 mL of Ac2O, 0.06 mL of PhBr, and 10 mL of
DMF. The hydrogen pressure of 4 bar, and the temperature was
25 °C.
0
2
4
6
8
t/h
Fig. 1. Kinetic curves of hydrogen absorption by the reaction
mixture during the process on the Pd(6%)/Sibunit (1) and
Pd(10%)/Sibunit (2) catalysts under a hydrogen pressure of 4
bar and a temperature of 25 °C. Composition of the mixture: 3.4
g of compound 2, 0.34 g of the catalyst, 10 mL of DMF, 5 mL of
Ac2O, and 0.06 mL of PhBr.
b Loadings of the reactants: 1.97 g of compound 3, 0.34 g of
Pd(6%)/Sibunit, and 9.8 mL of HCOOH (88%). The hydrogen
pressure of 4 bar, and the temperature was 25 °C.
c Yield of compound 4 synthesized in the twoꢀstep cycle of hyꢀ
drodebenzylation with respect to compound 2.

Debenzylation of hexabenzylhexaazaisowurtzitane Russ.Chem.Bull., Int.Ed., Vol. 56, No. 12, December, 2007 2373
Scheme 2
P are the destruction products of the polycyclic framework of compound 2 (more than five compounds)
Reagents and conditions: i. Pd/C, H2, PhBr, DMF; ii. Ac2O, DMF; iii. Ac2O, acids, H2O.
anhydride (Ac2O) by acetyl chloride (AcCl) resulted in
the resinification of the reaction mixture, and the use of
benzoic anhydride ((PhCO)2O) revealed no advantages
in the debenzylation processes. The most productive was
the use of solvents containing the amide groups. The solꢀ
vents DMF (on the Pd(6%)/C catalyst, the yield of comꢀ
pound 3 was 82%), DMA (80%), and Nꢀmethylpyrroliꢀ
done (50%) accelerate the acylation reaction due to the
basicity of the amide nitrogen atom. When ethers and
hydrocarbons are used, the yield of the target precursor of
CLꢀ20 was substantially lower. The results obtained for
the process on the Pd/Sibunit catalyst samples prepared
by us confirm the known published data, according to
which PhBr is the optimum source of HBr and its optiꢀ
mum amount is 1/8 molar fraction with respect to the
substrate.11 The absence of bromobenzene substantially
retards the hydrogenolysis of the C—N bond and prevents
this reaction to occur at room temperature.
From the viewpoint of the yield of the final product
and the reaction rate, a hydrogen pressure of 4 bar is
optumum to perform the debenzylation process. The reꢀ
duction of the hydrogen pressure considerably decreases
the v1 rate of debenzylation. An increase in the hydrogen
pressure increases the contribution of the side pathways of
transformation of the components of the reaction mediꢀ
um, which results in the disappearance of compound 3 in
the reaction products.
Hydrogen can be absorbed by the reaction mixture via
several paths: hydrogenation of acetic anhydride and broꢀ
mobenzene, hydrogenolysis of internal and external C—N
bonds of the hexaazaisowurtzitane framework, and hyꢀ
drogenation of residues of the framework. The considerꢀ
ation of Scheme 2 shows that an increase in the rate v1 of
the target hydrogenolysis reaction should increase the v1/v2
ratio and, therefore, should decrease the amount of the
destruction products, whose formation is a reason for catꢀ
alyst deactivation.
We experimentally studied the influence of the conꢀ
centrations of the substrate and acylating agent and the
hydrogen pressure on the course of the process and the
yield of the target product (Table 2). It was found that the
effect of the acetic anhydride concentration is dual: on
the one hand, acylating agent excess accelerates side proꢀ
cesses and, on the other hand, its deficiency results in a
decrease in the v4 acylation rate of unstable amine 6,
which is also unacceptable. The replacement of acetic
Table 2. Influence of the loadings of the reactants and cocatalyst
on the rate of hydrogen absorption and the yield of the target
product in the first step of debenzylation*
Volume
/mL
v
Yield of 3
/mole of H2
(mole of 2)–1 h–1
(%)
Ас2O
DMF
PhBr
We carried out a series of experiments with the variaꢀ
tion of the amount of the Pd(4%)/Sibunit catalyst at the
same concentrations of the reactants and cocatalyst, hyꢀ
drogen pressure, and stirring factor (Fig. 2). The rate of
hydrogen absorption by the reaction mixture increased
with an increase in the catalyst load. The highest rate of
hydrogen absorption was observed when 0.85 g of the
catalyst (per 3.4 g of the substrate, weight ratio catalyst/
substrate = 1/4) was charged; however, the yield of comꢀ
pound 3 decreased, being 71%. For the viewpoint of the
5
5
5
15
15
8
10
10
10
0
0
7
0
0.3
1.5
1.7
0.2
1.6
1.4
0
82
60
0
0
54
0.06
0.12
0
0.06
0.06
* Debenzylation was carried out on the Pd(6%)/Sibunit catalyst
under a hydrogen pressure of 4 bar and at a reaction temperature
of 25 °C; m = 3.4 g.

2374
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 12, December, 2007
Koskin et al.
Y (%)
v/(mmole of H2) (mmole of 2)–1 h–1
metal should be used to accelerate the debenzylation by
the Pd/C catalysts. The use of the Pd(4%)/Sibunit cataꢀ
lyst (loading of the catalyst/substrate in the weight ratio
1/10) did not allow us to obtain the target precursor of
CLꢀ20 in the stepwise reaction. In this case, due to the
deactivation processes, the moment when the debenzylaꢀ
tion rate becomes lower than the rate of irreversible side
processes associated with opening of the framework of
compound 2, preceded the step of formation of comꢀ
pound 3. The 1H NMR study of the reaction mixture
indicates that the process ceases at the step of formation
of the diacetyl hexaazaisowurtzitane products. An increase
in the loading of the Pd(4%)/Sibunit sample to the weight
ratio catalyst/substrate = 1/5 was needed for the complete
transformation of compound 2 into 3.
Temperature dependences of the activity and selectivity
of the debenzylation catalysts. Unlike the earlier known
prototypes of the catalysts, namely, Pd(10%)/C (E101
NE/W, Degussa), Pd(20%)(OH)2/C (Pearlman catalyst),
and Pd(5%)/C (ICTꢀ3ꢀ23), the debenzylation process on
the Pd(4—10%)/Sibunit catalyst samples prepared by us
should be carried out at room temperature. Even an insigꢀ
nificant rise of the reaction temperature to 40—50 °C reꢀ
sulted in a substantial decrease in the selectivity of C—N
bond hydrogenolysis and the yield of the target product.
At the same time, the activity of the mentioned protoꢀ
types of catalysts is insufficient for the debenzylation proꢀ
cess to occur at room temperature. Possibly, the differꢀ
ence observed in the properties of the catalysts is caused
by differences in the size of the metallic palladium partiꢀ
cles in the catalyst samples prepared by us and in the
known prototypes. According to the TEM data, metal
particles 3 nm in size predominate in the Pd(6%)/Sibunit
catalyst, whereas the 6ꢀnm particles prevail in Pd(10%)/C
(Degussa) (see Table 3). The activity of the Pd/C cataꢀ
lysts increases with a decrease in the average particle size
and, hence, an increase in the active surface area; howevꢀ
er, their selectivity in C—N bond hydrogenolysis decreasꢀ
es. A possibility to perform the process at lower (room)
temperature has several advantages, the main of which is
the partial suppression of destruction processes of the
hexaazaisowurtzitane framework that occur at elevated
temperatures.
80
60
40
20
3
1
2
2
1
0.50
0.75
1.00 m/g
Fig. 2. Effect of the weight of the loaded Pd(4%)/Sibunit cataꢀ
lyst (m) on the rate of hydrogen absorption v (1) and the yield of
compound 3 (2) under a hydrogen pressure of 4 bar and at
a temperature of 25 °C. The composition of the mixture: 3.4 g of
compound 2, 10 mL of DMF, 5 mL of Ac2O, and 0.06 mL of
PhBr.
yield of the target product, the catalyst/substrate loading
in the weight ratio 1/5 was optimum. With larger loadings
of the Pd/C catalyst (larger than the weight ratio 1/5),
the acceleration of the target reaction of C—N bond hyꢀ
drogenolysis is accompanied by an increase in the rates of
the side reactions.
Testing of the hydrogenation catalysts in the debenzyꢀ
lation of compound 2. We tested a series of the hydrogenaꢀ
tion catalysts of different nature (Table 3) to find a more
stable catalytic system for the debenzylation of comꢀ
pound 2. Unfortunately, the screening performed revealed
no system comparable in catalytic activity with the Pd/C
samples. The samples with a high content of the noble
Table 3. Catalytic systems of hydrogenation in the debenzylaꢀ
tion of compound 2 at 25 (I) and 50 °C (II)
Catalysta
db
v
Yield of 3
/nm
/(mole of H2)
(mole of 2)–1 h–1
(%)
I
II
I
II
Thus, the results of the present study indicate that the
Pd/C catalyst is deactivated during the first step of the
debenzylation process. The dependences of the yield of
the target product on the solvent nature, loadings of the
reactants, catalyst, and cocatalyst, and hydrogen pressure
were found for the reaction on the Pd/Sibunit catalysts.
The optimum conditions were selected for the synthesis
of compound 3. The increase in the dispersity of the metal
particles in the Pd/C catalyst was found to substantially
enhance the catalytic activity, which makes it possible to
perform the process under mild conditions at room temꢀ
perature.
Pd(6%)/C, ICTꢀ3ꢀ23 c
Pd(10%)/C, E101 NE/Wc
Pd(4%)/Sibunit
Pd(6%)/Sibunit (3 nm)
Pd(10%)/Sibunit
5
6
1.0
—
1.8
1.0
1.5
—
1.5
—
3.2
3.3
2.8
3.7
3.8
56 80
61 80
0
82
86
0
0
0
a The catalyst was tested in the hydrodebenzylation of comꢀ
pound 2. Loadings of the reactants: 3.4 g of compound 2, 0.34 g
of Pd/C, 5 mL of Ac2O, 0.06 mL of PhBr, and 10 mL of DMF.
The hydrogen pressure was 4 bar, and the reaction temperature
was 25 °C.
b Average size of the palladium particles.
c Commercially available catalysts.

Debenzylation of hexabenzylhexaazaisowurtzitane Russ.Chem.Bull., Int.Ed., Vol. 56, No. 12, December, 2007 2375
9. R. B. Wardle, J. C. Hinshaw, P. Brathwaite, M. Rose,
G. Johnston, R. Jones, and K. Poush, in Proc. 27th Intern.
Ann. Conf. of ICT, Karlsruhe, 1996, 27/1.
The authors are grateful to A. V. Golovin for help in
recording 1H and 13C NMR spectra and to V. A. Utkin for
GC/MS analysis of compositions of the reaction mixture.
10. A. P. Koskin and I. L. Simakova, Sb. dokl. I Vseros. nauch.ꢀ
tekh. konf. "Perspektivy sozdaniya i primeneniya kondensiroꢀ
vannykh energeticheskikh materialov" [Proceedings of the I Allꢀ
Russia Scientific Technical Conference "Prospects of Creation
and Application of Condensed Energetic Materials"] (Biisk,
September 27—29, 2006), Biisk, 2006, 42 (in Russian).
11. 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. F. George, and J. FlippenꢀAndersen, Tetrahedron, 1998,
54, 11793.
References
1. S. V. Sysolyatin, A. A. Lobanova, Yu. T. Chernikova, and
G. V. Sakovich, Usp. Khim., 2005, 74, 830 [Russ. Chem. Rev.,
2005, 74, 757 (Engl. Transl.)].
2. J. Mathieu and H. Stucki, Chimia, 2004, 58, 383.
3. R. L. Simpson, P. A. Urtiew, D. L. Ornellas, G. L. Moody,
K. J. Scribner, and D. M. Hoffman, Propell. Explos. Pyrot.,
1997, 22, 249.
12. A. J. Bellamy, Tetrahedron, 1995, 51, 4711.
13. W. Han, Y. Ou, and B. Chen, J. Beijing Inst. Tech. (China),
2006, 15, 225; Chem. Abstr., 2000, 132, 78224.
14. S. Nishimura, Handbook of Heterogeneous Catalytic Hydroꢀ
genation for Organic Synthesis, Wiley, New York, 2001, p. 607.
4. U. R. Nair, R. Sivablan, G. M. Gor, M. Giza, Sh. N. Astiꢀ
na, and H. Singh, Fiz. Goren. Vzryva, 2005, 4, 3 [Comb.,
Explos., and Shock Waves, 2005, 41, 121 (Engl. Transl.)].
5. US Pat. 5739325; Chem. Abstr., 1998, 127, 110983.
6. US Pat. 6147209; Chem. Abstr., 2000, 133, 352267.
7. A. T. Nielsen, R. A. Nissan, D. J. Vanderah, C. L. Coon,
R. D. Gilardi, C. F. George, and J. FlippenꢀAndersen, J. Org.
Chem., 1990, 55, 1459.
8. M. R. Crampton, J. Hamid, R. Millar, and G. Ferguson,
J. Chem. Soc., Perkin Trans., 1993, 2, 923.
Received June 7, 2007;
in revised form November 15, 2007