Synthesis and chemistry of 1,3,5,7-tetranitrocubane including measurement of its acidity, formation of o-nitro anions, and the first preparations of pentanitrocubane and hexanitrocubane

Article abstract of DOI:10.1021/ja970552q

Nitro groups on alternate corners of cubane enhance the acidity of cubyl hydrogen (pK(a) ~21) and provide sufficient activation for ready anion formation. The sodium salt of 1,3,5,7-tetranitrocubane reacts easily with electrophiles and leads thereby to yet more highly substituted cubanes, like carbomethoxy- and (trimethylsilyl)tetranitrocubane. Anions from these species are also easily formed and are useful for further substitution on the cubane nucleus. Dinitrogen tetraoxide reacts with the anion of tetranitrocubane to give 1,2,3,5,7-pentanitrocubane, the first cubane to contain vicinal nitro groups. The reaction probably involves oxidation of the anion to the corresponding radical. Similarly, the anion of pentanitrocubane is used to prepare 1,2,3,4,5,7-hexanitrocubane, the most highly nitrated cubane made to date. Single-crystal X-ray structural information is given for both penta- and hexanitrocubane.

Full text of DOI:10.1021/ja970552q

VOLUME 119, NUMBER 41
OCTOBER 15, 1997
© Copyright 1997 by the
American Chemical Society
Synthesis and Chemistry of 1,3,5,7-Tetranitrocubane Including
Measurement of Its Acidity, Formation of o-Nitro Anions, and
the First Preparations of Pentanitrocubane and Hexanitrocubane1
Kirill A. Lukin,† Jianchang Li,† Philip E. Eaton,*,† Nobuhiro Kanomata,†
Ju1rgen Hain,† Eric Punzalan,† and Richard Gilardi‡
Contribution from the Department of Chemistry, The UniVersity of Chicago,
5735 S. Ellis AVenue, Chicago, Illinois 60637, and Laboratory for the Structure of Matter,
The NaVal Research Laboratory, Washington, D.C. 20375
ReceiVed February 20, 1997X
Abstract: Nitro groups on alternate corners of cubane enhance the acidity of cubyl hydrogen (pKa ∼21) and provide
sufficient activation for ready anion formation. The sodium salt of 1,3,5,7-tetranitrocubane reacts easily with
electrophiles and leads thereby to yet more highly substituted cubanes, like carbomethoxy- and (trimethylsilyl)tet-
ranitrocubane. Anions from these species are also easily formed and are useful for further substitution on the cubane
nucleus. Dinitrogen tetraoxide reacts with the anion of tetranitrocubane to give 1,2,3,5,7-pentanitrocubane, the first
cubane to contain vicinal nitro groups. The reaction probably involves oxidation of the anion to the corresponding
radical. Similarly, the anion of pentanitrocubane is used to prepare 1,2,3,4,5,7-hexanitrocubane, the most highly
nitrated cubane made to date. Single-crystal X-ray structural information is given for both penta- and hexanitrocubane.
Introduction
above 220 °C), but a thermodynamic powerhouse (∆Hf ∼150
kcal/mol).3 It is one of the most dense hydrocarbons known
(1.29 g/cm3).4 As molecular volume is (very roughly)5 a group
additive property,6 highly nitrated cubanes can be predicted to
be very dense,6 and hence they are expected to be very powerful
explosives. Octanitrocubane (1) is calculated to be 15-30%
The “effectiveness” of an explosive is dependent on many
things. The energetics of the decomposition reaction and the
number of moles and molecular weight of the gaseous products
are among the critical factors.2 Density is also crucial. The
more moles of an explosive that can be packed into the limited
volume of a shell the better. Less obvious, but more important,
the velocity of the propagation wave through an explosive, i.e.,
the rate of energy release, is proportional to its density squared.
Cubane is a kinetically stable compound (decomposition only
(3) (a) Kybett, B. D.; Carroll, S.; Natalis, P.; Bonnell, D. W.; Margrave,
J. L.; Franklin, J. L. J. Am. Chem. Soc. 1966, 88, 626. (b) Kirklin, D. R.;
Churney, K. L.; Domalski, E. S. J. Chem. Thermodynam. 1989, 21, 1105.
(4) Fleischer, E. B. J. Am. Chem. Soc. 1964, 86, 3889. For comparison:
the density of adamantane is approximately 1.09 g cm-3
.
(5) A good discussion can be found in: Marchand, A. P.; Zope, A.;
Zaragoza, F.; Bott, S. G.; Ammon, H. L.; Z. Du, Tetrahedron 1994, 50,
1687.
(6) For example: Stine, J. R. Prediction of Crystal Densities by Group
AdditiVity, Report LA-8920; Los Alamos National Laboratotry: Albuquer-
que, NM, 1981.
(7) (a) Gilbert, E. E. Private communication, 1979. (b) Alster, J.; Sandus,
O.; Genter, R.; Slagg, N.; Ritchie, J. P.; Dewar, M. J. S. Working Group
Meeting on High-Energy Molecules, Hilton Head, SC, 28-29 April, 1981.
† The University of Chicago.
‡ The Naval Research Laboratory.
X Abstract published in AdVance ACS Abstracts, September 1, 1997.
(1) Part of this work was communicated previously: Lukin, K. A.; Li,
J.; Gilardi, R.; Eaton, P. E. Angew. Chem., Int. Ed. Engl. 1996, 35, 864.
(2) (a) Kamlet, M. J.; Jacobs, S. J. J. Chem. Phys. 1968, 48, 23. See
also the overviews in: (b) Marchand, A. P. Tetrahedron 1988, 44, 2377
and (c) Bottaro, J. C. Chem. Ind. 1996, 7, 249.
S0002-7863(97)00552-0 CCC: $14.00 © 1997 American Chemical Society

9592 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
Lukin et al.
Scheme 1
necessarily have nitro groups on adjacent carbons none can be
made without development of a new methodology. We present
such methodology in this paper.
The Precursors
In 1993 we reported the first synthesis of cubane-1,3,5,7-
tetracarboxylic acid.9 For this we utilized techniques for ortho
metalation of amide-activated cubanes developed specially for
the purpose.10 The approach was novel, but long and complex.
Once 1,3,5,7-tetranitrocubane (4) was obtained from the tetraacid
(cf., Scheme 1) and was found to be very dense,9 shock
insensitive, and a powerful explosive exceeding expectation,11
it was clear that a better synthesis would be required. In 1993
Bashir-Hashemi provided just that with his brilliant application12
of photochemically induced chlorocarbonylation (the Kharasch-
Brown reaction13) to functionalization of the cubane system.
He showed that irradiation of cubane carboxylic acid chloride
5 in excess oxalyl chloride led to substantial formation of
cubane-1,3,5,7-tetracarboxylic acid chloride (12)sthe key in-
termediate in our original synthesis of tetranitrocubane 4spre-
sumably by one or more of the paths in Scheme 3.
Scheme 2
better than HMX,7 an important explosive produced and used
in large quantity.
The product distribution (ratio of positional isomers) from
this photochlorocarbonylation is exceedingly dependent on
reaction conditions, some evident and others less so. Bashir-
Hashemi initially reported the use of tungsten lamp radiation
for the chlorocarbonylation. In our hands this proved useful
for introduction of one or two chlorocarbonyl groups but
impractical for production of the desired tetraacid chloride due
to long reaction times (several days). Use of a Rayonet
photoreactor (254 nm, quartz cell) at its ambient temperature
(40 °C), later recommended,12b gave us a mixture of tetraacid
chlorides 12, 13, and 14 in about a 53:12:35 ratio,12c but still
required long periods of irradiation. We switched to a higher
flux systemsa Hanovia, 450-watt, medium pressure Hg arcsand
followed the extent of chlorocarbonylation and its stereospeci-
ficity by 1H NMR analysis of aliquots. Our initial experiments
were carried out using a Pyrex photoreactor. Under our
conditions (see Experimental Section) the first chlorocarbony-
lation was not as selective as initially described;12a,14 in addition
to the 1,3- and 1,4-diacid chlorides 6 and 7, cubane-1,2-
dicarboxylic acid chloride (8) was produced in substantial
amount. The ratio of ortho to meta to para disubstituted
products,15 8 to 6 to 7, determined before much triacid chloride
had formed, was about 6:10:1. This ratio is approximate as
the NMR signal for the 1,4-diacid overlaps one of those of the
1,2. Still, it is substantially different from the simple statistical
expectation of 3:3:1. Hrovat and Borden’s ab initio calculations
The introduction of a few nitro groups onto the cubane
nucleus is straightforward in certain cases. It can be ac-
complished by functional group transformation of cubanecar-
boxylic acids to the corresponding amines via Curtius rear-
rangement followed by oxidation, as shown, for example, in
Scheme 1 for 1,4-dinitrocubane (2).8 The method fails however
for cubanes with adjacent carboxylic acid groups. Although
the corresponding vicinal diamines can be made and are
relatively stable compounds, ring cleavage reactions intervene
during attempted oxidation to a 1,2-dinitrocubane. The cubane
system is destroyed entirely, probably by a push-pull mecha-
nism something like that shown in Scheme 2. This sort of
skeletal cleavage in the cubane systems always happens when
attempts are made to place an electron-donating group on a
cubane carbon adjacent to one bearing an electron-withdrawing
substituent. For example, all attempts at making 2-aminocu-
banecarboxylic acid have failed.
(10) (a) Eaton, P. E.; Castaldi, G. J. Am. Chem. Soc. 1985, 107, 724-
726. (b) For a review, see: Eaton, P. E. Angew. Chem., Int. Ed. Engl. 1992,
31, 1421.
(11) Army Research, Development and Engineering Center, Picatinny,
NJ. Private communication.
(12) (a) Bashir-Hashemi, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 612.
(b) Bashir-Hashemi, A.; Li, J.; Gelber, N.; Ammon, H. J. Org. Chem. 1995,
60, 698. (c) We and Bashir-Hashemi (private communication) have noted
that the reaction is somewhat cleaner at 0 °C.
Cage cleavage via the push-pull mechanism limits the
methodology of Scheme 1 to the preparation of nitrocubanes
without vicinal nitro groups. It has been applied successfully
to the preparation of 1,3,5-trinitrocubane (3) and 1,3,5,7-
tetranitrocubane (4),9 but as all more highly nitrated cubanes
(8) (a) Eaton, P. E.; Shankar, B. K. R.; Price, G. D.; Pluth, J. J.; Gilbert,
E. E.; Alster, J; Sandus, O. J. Org. Chem. 1984, 49, 185. (b) Eaton, P. E.;
Wicks, E. J. Org. Chem. 1988, 53, 5353.
(9) Eaton, P. E.; Xiong, Y.; Gilardi, R. J. Am. Chem. Soc. 1993, 115,
10195.
(13) (a) Kharasch, M. S.; Brown, H. C. J. Am. Chem. Soc. 1940, 62,
454. (b) Ibid. 1942, 64, 329.
(14) A later report of Bashir-Hashemi et al.12b includes the the formation
of cubane-1,2-dicarboxylic acid chloride (see footnote 9 therein).

Synthesis and Chemistry of 1,3,5,7-Tetranitrocubane
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9593
Scheme 3
attributable to protons at C7 and C4,6, respectively. A control
chlorocarbonylation of pure cubane-1,2-dicarboxylic acid chlo-
ride (8) gave 10 and 11 in a 2:1 ratio.
Further chlorocarbonylation of compounds 9-11 in a Pyrex
photoreactor was slow; we switched to a quartz apparatus.
Optimum results were obtained when the temperature of the
reaction mixture was kept below 15 °C (see Experimental
Section). Under these conditions the first and second chloro-
carbonylations were remarkably quick and were over in less
than 15 min. Further chlorocarbonylation was quite selective,
controlled apparently by the fact that many potential sites are
deactivated by two ortho chlorocarbonyl groups (positions 2,
4, 6, 8 in 9; 3, 5 in 10; 8 in 11). Thus, the favored products are
the tetraacid chlorides 12, 13 and 14.19 Further reaction of 12
and 13 is made unlikely by the combined deactivation effects,
but an additional chlorocarbonylation was observed on 14 (f
15); position 6 thereof being still sufficiently reactive.
On a preparative scale (10 mmol) irradiation of a 0.1 M
solution of cubanecarboxylic acid chloride 5 in oxalyl chloride
gave after 3 h roughly a 70:8:22 mixture of the tetrasubstituted
cubanes 12-14, from which the desired 1,3,5,7 isomer 12 could
be precipitated in 95% purity by addition of diethyl ether. From
most runs we were usually able to obtain about a 30% yield of
12 oVerall from cubane monoacid. However, we do not wish
to leave the impression that the chlorocarbonylation reaction is
truly “under control”. The variables are many in this complex
free-radical reaction and even subtle changes (such as whether
the light source is new or aged) affect the yield significantly.
In spite of this caveat, the reaction is very useful and provides
a convenient preparation of an extraordinarily versatile inter-
mediate.20
In previous work the tetraacid chloride 12 had been converted
via quadruple Curtius rearrangement of the tetraacyl azide 16
to the corresponding tetraisocyanate 17 (Scheme 4).19 Acyl
Scheme 4
of the charge distribution over the cubane nucleus in the
transition state pertaining to cubyl radical formation shows that
the least positive charge develops at the positions meta to the
site of radical formation.16 Thus, preference for meta over ortho/
para substitution is in agreement with the electron-withdrawing
properties of a chlorocarbonyl group and the radical mechanism
of the reaction.17 On the other hand, it is not clear why the
observed ratio of ortho to para substitution is ∼6:1. Although
statistics favors ortho substitution (3:1), the inductive effect must
work against this.18
When the irradiation was carried further the major products
were the 1,3,5-, 1,2,4-, and 1,2,3-cubanetricarboxylic acid
chlorides 9, 10, and 11 in approximately a 6:4:2 ratio. The
first two were identified by comparison to samples prepared
from known cubane diacids.19 The identification of 11, derived
from the previously unknown 1,2,3-triacid, was based on
azides are often shock sensitive, dangerous compounds. We
have learned (fortunately without serious incident) that cubane
acyl azides are primary explosives, worthy of the greatest
respect. Such compounds cannot be worked with safely unless
always in dilute solution. This presents experimental difficulties
for the uncatalyzed reaction of cubyl acid chlorides with
trimethylsilyl azide (TMSN3) is very slow in solutions dilute
enough to keep the product acyl azide in solution. If excess
TMSN3 be used to enhance the rate, then before thermal
rearrangement of the acyl azide can be accomplished in good
yield: (1) the excess must be removed (a treacherous procedure);
or (2) the solution must be diluted substantially to reduce the
1
multiplets in its H NMR spectrum at δ 4.08 and 4.45 ppm
(15) The formal Baeyer system numbering of the cubane skeleton is
shown for compound 5 in Scheme 3. Ortho, meta, para designations, as
commonly understood, are more easily visualized.
(16) Hrovat, D. A; Borden, W. T. J. Am. Chem. Soc. 1994, 116, 6459.
(17) Pryor, W. A.; Davis, W. H, Jr.; Stanley, J. P. J. Am. Chem. Soc.
1973, 95, 4754.
(18) Steric effects between ortho substituents on a cubane nucleus are
very small.
(19) Eaton, P. E.; Lee, C. H.; Xiong, Y. J. Am. Chem. Soc. 1989, 111,
8016 and references therein.
(20) For example, Carell, T.; Wintner, E. A.; Bashir-Hashemi, A.; Rebek,
J., Jr. Angew. Chem., Int. Ed. Engl. 1994, 33, 2059.

9594 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
Lukin et al.
reaction of TMSN3 with the isocyanate.21 The problem is better
resolved by using a catalyst to accelerate acyl azide formation.
Tertiary amines and pyridines are familiar catalysts for
transformations of acid chlorides. However, these have not been
used as catalysts in “one-pot” azidation/Curtius rearrange-
ments,22 probably because they can also catalyze oligomerization
of the isocyanate.23 Hertzler,24 working here, found that
triethylamine or pyridine nicely accelerated the conversion of
1,2,4,7-cubanetetraacid chloride (13) to the tetraacyl azide, but
its rearrangement to isocyanate in the presence of these amines
was accompanied by the formation of higher molecular weight
material. The much more hindered amine 2,6-di-tert-butylpyr-
idine (10 equiv %) catalyzed azidation as well, but did not
interfere with the isocyanate. This was mildly puzzling: if the
hindered amine were capable of “classic” catalysis of the
reactions of the acid chloride, it should also have initiated
isocyanate oligomerization.23 Ultimately it became clear that
the real catalyst in the Hertzler work was the pyridinium
hydrochloride, formed adventitiously. When a small amount
(0.11 equiv %) of 2,6-di-tert-butylpyridine dihydrochloride25 was
purposely added to a dilute solution of acid chloride and TMSN3
formation of the tetraacyl azide was complete within an hour.
The catalytic action of the salt is probably simple, providing
free chloride ion to displace azide anion from TMSN3. The
resulting pyridinum azide is highly ionized, and thus azide ion
becomes available at a good concentration. Hydrogen chloride
by itself proved useless as a catalyst for it reacts with TMSN3
to give TMSCl and hydrazoic acid, which itself is very little
ionized (pKa ∼ 4.7). The rearrangement of the tetraacyl azide
was carried out in reasonably dilute solution to suppress
secondary reactions. The tetraisocyanate 17 was isolated in
>80% yield, pure enough to be used directly in the next
conversions. In passing, we note that this tetraisocyanate, which
has tetrahedral symmetry, is an excellent candidate as a core
molecule for polyurethane dendrimers.
tetranitrocubane have been presented previously.9 Here we are
concerned with the wet chemistry of the compound. The
symmetry of the molecule (pseudo-tetrahedral) is such that each
of the four identical hydrogens is surrounded by three nitro
groups, one on each â carbon. The inductive/field27 electron-
withdrawing effect of the nitro group is stronger than that of
any other uncharged functionality.28,29 This, added to the
already enhanced acidity of a cubyl C-H group over a typical
hydrocarbon,30 might make it relatively easy to generate the
anion of 1,3,5,7-tetranitrocubane. Coordination of the electron-
rich oxygen of one or more of the nitro groups with the gegenion
could also offer particular stabilization (illustrated in the
extreme) to a salt such as 18. One can also speculate that the
highly polarized nitro group(s) might aid the approach of a base
and the kinetic deprotonation of the CH group.31
Regardless of the intimate details, the acidity of the cubane
hydrogens is indeed so enhanced by the nitro groups in 1,3,5,7-
tetranitrocubane that their exchange can be effected rapidly even
with bases as mild as sodium methoxide in methanol. Reaction
of 4 with 0.125 molar CD3ONa in CD3OD for 45 min at room
temperature followed by quenching with hydrochloric acid and
extraction with EtOAc gave in 87% yield 4-d4. The reverse
experiment, reaction of 4-d4 with CH3ONa in CH3OH, returned
4.
To our knowledge, there is still8b no way to effect oxidative
cleavage of the carbon-nitrogen double bond of an isocyanate.
We have, however, demonstrated that dimethyldioxirane (DMDO)
in wet acetone efficiently converts isocyanates into nitro
compounds via hydration of the isocyanate groups to the
carbamic acids, their immediate decarboxylation, and rapid
DMDO oxidation of the amines so generated.8b,26 Pure 1,3,5,7-
tetranitrocubane (4) was isolated in 45% yield overall from
tetraacid chloride 12sa remarkable feat considering that at least
five distinct transformations must occur at each of four sites.
(24) Hertzler, R. L.; Eaton, P. E. Abstracts of Papers; 205th National
Meeting of the American Chemical Society, Denver, CO, Spring 1993;
American Chemical Society: Washington, DC, 1993; Paper 128.
(25) (a) Hopkins, H. P., Jr.; Ali, S. Z. J. Am. Chem. Soc. 1977, 99, 2069.
(b) Arnett, E. M.; Chawla, B. J. Am. Chem. Soc. 1978, 100, 214.
(26) Murray, R. W.; Jeyaraman, R.; Mohan, L. Tetrahedron Lett. 1986,
27, 2335.
(27) Sacher, E. Tetrahedron Lett. 1986, 27, 4683.
(28) Ceppi, E.; Eckhardt, W.; Grob, C. A. Tetrahedron Lett. 1973, 3627.
(29) Marriott. S.; Reynolds, W. F.; Taft, R. W.; Topsom, R. D. J. Org.
Chem.1984, 49, 959.
(30) Cubane is approximately 60 000 times more acidic than cyclohexane
(Dixon, R. E.; Streitwieser, A.; Williams, P. G.; Eaton, P. E. J. Am. Chem.
Soc. 1991, 113, 357), a function of the increased s character in the exocyclic
orbitals. Yet this is still quite insufficient to permit significant equilibrium
metalation with even powerful amide bases like LiTMP.10
The Acidity of 1,3,5,7-Tetranitrocubane. The physical
properties, crystal structure, and thermal behavior of 1,3,5,7-
(31) We can find no definitive evidence in the literature for metalation
e.g., CH f CLi) assisted by an ortho nitro group. Ortho metalation of
aromatic nitro compounds is usually unsuccessful; electron transfer is the
dominant process instead (Seebach, D.; Colvin, E. W.; Lehr, F.; Weller, T.
Chimia 1979, 33, 8). However, ortho-lithiated nitrobenzenes (thermally
unstable) have been generated by metal-halogen exchange at temperatures
below -100 °C (Ko¨brich, G.; Buck, P. Chem. Ber. 1970, 103, 1412; 1970,
103, 1420). When deprotonation â to a nitro group does occur (e.g., in
â-nitro ketones because of the carbonyl group) elimination of nitrite and
olefin formation follows immediately (Ono, N. In Nitro Compounds. Recent
AdVances in Synthesis and Chemistry, Feuer, H., Nielsen, A. T., Eds.; VCH
Publishers Inc., 1990, p 86).
(21) See for example: Gilardi, R.; George, C.; Karle, J.; Eaton, P. E.;
Rao, M. J. Heterocyc. Chem. 1993, 1389.
(22) Potassium azide complex with 18-crown-6 in benzene (Warren, J.
D.; Press, J. B. Synth. Communs. 1980, 10, 107) and zinc iodide (Farooq,
O.; Wang, Q.; Wu, A.; Olah, G. A. J. Org. Chem. 1990, 55, 4282) have
been reported to be good catalysts for azidation of acid chlorides with
TMSN3. The latter proved unsatisfactory for the case at hand. We did not
investigate the other.
(23) (a) Frish, K. C.; Rumao, L. P. J. Macromol. Sci. 1970, C5, 103. (b)
However, hindered amines do not react with isocyanates: Kno¨lker, H.-J.;
Braxmeier, T. Tetrahedron Lett. 1996, 37, 5861.

Synthesis and Chemistry of 1,3,5,7-Tetranitrocubane
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9595
None of the salts which must intervene in these exchanges
could be detected in CH3OD/CH3ONa by NMR. Their con-
centration must be quite low. We could, however, approximate
the pKa of tetranitrocubane 4 from low-temperature proton NMR
measurements on equilibrium mixtures with various bases.
Reaction of 4 with sodium bis(trimethylsilyl)amide in THF-d8
at -75 °C showed clearly the formation of tetranitrocubyl
sodium 19. The NMR resonance of three residual protons is a
sharp singlet at δ 5.54 ppm, 0.53 ppm higher field than that of
the neutral parent under the same conditions.
The Chemistry of 1,3,5,7-Tetranitrocubylsodium. Tet-
ranitrocubylsodium (19), formed directly on treatment of 4 with
sodium bis(trimethylsilyl)amide in THF at -75 °C, is stable
for hours at that temperature, but it decomposes rapidly above
-50 °C. Its reactions with electrophiles provide a useful and
convenient way to achieve further functionalization of the
cubane nucleus. For example, reaction of 19 in THF at -75
°C with carbon dioxide followed by neutralization gave the
tetranitrocubane carboxylic acid 20 in 85% isolated yield. This
was characterized as its methyl ester 20a. Similarly, iodination
of 19 using 1,2-diiodoethane as the iodinating agent gave iodo-
tetranitrocubane 21 in 70% yield. These compounds are
colorless, stable crystalline materials.
Almost complete (g99%) salt formation occurs in THF on
treatment of 4 with one equivalent of sodium bis(trimethylsi-
lyl)amide, but none (e1%) with R-lithiobenzyl cyanide. As
the pKa in THF of bis(trimethylsilyl)amine is 25.832 and that of
benzylcyanide is 18.7,33 the pKa of 4 must be between about
20 and 24. We have been able to narrow this range to 20.5-
22.5 by noting that 4 is partially converted to its anion by 1.5
equiv of potassium tert-butoxide (the pKa of tert-butyl alcohol
in THF is given as 21.6).34 However, this cannot be used to
pinpoint the pKa of 4 as the alcohol and alcoholate give only
one averaged NMR signal due to rapid proton exchange.
How many nitro groups on cubane are necessary to achieve
a measurable and/or synthetically useful acidity? Early experi-
ments with mono- and 1,4-dinitrocubane designed to look for
substituent-enhanced acidity were frustrating; complete destruc-
tion with strong bases and lack of reactivity with weaker bases
were all that could be found.35 Not even the simplest exchange
reaction could be achieved. Clearly, activation by one nitro
group is insufficient for â-anion formation, if it occurs at all,
instead of what are probably single-electron transfer reactions
that somehow unravel the cubane nucleus.
A brief study of 1,3,5-trinitrocubane (3),9 which possesses
three hydrogens each with two adjacent nitro groups and one
hydrogen surrounded by three nitro groups, showed that only
the latter hydrogen underwent D-for-H exchange with CD3OD
in the presence of CD3ONa. Clearly this demonstrates that the
remote fourth nitro group in 4 is not essential for anion
formation. But, all three nitro groups about a single CH do
seem necessary for successful ortho anion formation with
methoxide as the base. We have not yet looked to see if two
nitro groups ortho to the same cubyl CH would suffice were
stronger bases used, but we suspect this will prove to be the
case.
1,3,5,7-Tetranitrocubylsodium can also be used conveniently
for the preparation of various more covalently bound and more
stable metal/metalloid derivatives of tetranitrocubane. (Trimeth-
ylsilyl)tetranitrocubane (22) was obtained in 80% yield by
quenching 19 with chlorotrimethylsilane. Although the com-
pound is air stable, the TMS group is lost on silica gel
chromatography, returning tetranitrocubane 4. Such sensitivity
is unusual for carbon-silicon bonds. (Trimethylsilyl)cubane,36
without nitro groups, is completely stable under such conditions.
The easy loss of the TMS group from 22 must trace to the
stabilization available to a tetranitrocubyl anion. Nonetheless,
the much more hindered silyl derivative (triisopropylsilyl)tet-
ranitrocubane 23 produced in 80% yield when 19 was quenched
with triisopropylsilyl chloride (TIPSCl) fully survived purifica-
tion by chromatography.
Heavy metal derivatives of 4 could be formed with similar
ease. Quenching 19 with ethylmercury chloride gave the
mercuriated tetranitrocubane 24 in 65% isolated yield. Reaction
of 19 with triethyllead chloride gave the plumbylnitrocubane
25 in 67% yield. All these metallic derivatives of tetranitrocu-
bane are air-stable solids, soluble in common organic solvents
like chloroform or acetone. We anticipate numerous synthetic
applications.37
1
The H NMR spectrum of each of the cubanes prepared in
this series of derivatives of 1,3,5,7-tetranitrocubane shows only
a single sharp line, reflecting the high molecular symmetry. The
13C NMR is more revealing, identifying the three identical
C-NO2 groups, the three identical C-H groups, the singular
C-NO2 group remote from the newly introduced substituent,
and the carbon bearing this substituent. The introduction of
electron-withdrawing substituents on 4 caused only slight
(32) Fraser, R. R.; Mansour, T. S.; Savard, S. J. Org. Chem. 1985, 50,
3232.
(33) Fraser, R. R.; Mansour, T. S.; Savard, S. Can. J. Chem. 1985, 63,
3505.
(36) From cubyllithium and TMSCl: Emrick, T., this Laboratory.
(37) For example: (a) Lukin, K. A.; Li, J.; Gilardi, R.; Eaton, P. E.
Angew. Chem., Int. Ed. Engl. 1996, 35, 866. (b) Lukin, K. A.; Li, J.; Gilardi,
R.; Eaton, P. E. J. Org. Chem., in press.
(34) Brown, C. A. J. Chem. Soc., Chem. Commun. 1974, 680.
(35) We are grateful to Drs. Ravi Shankar and Gene Wicks for these
experiments.

9596 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
Lukin et al.
downfield shifts of the remaining cubyl protons: compare 4, δ
) 6.085; 20, 6.19; 21, 6.19 ppm. Electron-donating substituents
have an even smaller effect on the proton chemical shift:
∼-0.01 ppm for 22 and 24. Electron-withdrawing and electron-
donating substituents exhibited opposite (but minor) effects on
cubyl carbons in other than the ipso position. For example,
the iodine substituent in 21 resulted in a 2.0 ppm downfield
shift of the signal from the ortho (â) carbon, a 1.0 ppm upfield
shift of the para carbon and did not affect the corresponding
ortho/para/meta shifts for the TMS-substituted compound 22
are +1.3, -1.4, and 0 ppm. Interestingly, the effects “cancel”
at the meta position.
Scheme 5
We often observed the formation of small amounts of doubly
functionalized products from the reaction: 4 + (TMS)2NNa (f
19) followed by quenching with electrophiles. When an excess
of the base was used, this increased noticeably. Treatment of
4 with 5 equiv of (TMS)2NNa followed by quenching with
carbon dioxide and neutralization gave a 70:30 mixture of 20
with 26, the tetranitro diacid, characterized as its dimethyl ester
26a. Similarly, use of 3 equiv of base followed by reaction
with TMSCl or TIPSCl gave approximately 20:80 mixtures of
mono- and bis-silylated materials, from which the bis-TMS or
bis-TIPS tetranitrocubane 27 or 28, respectively, could be
isolated in about 70% yield. Probably 26-28 result from
sequential anion formation, the second formed after the first is
quenched, but we are looking to see if dianion formation might
be involved.
tentatively assigned from NMR considerations (Vide infra) as
pentanitrocubane 29 and chlorotetranitrocubane 30. When the
reaction was run at a much lower temperaturesan excess of
nitryl chloride was condensed onto a frozen solution of 19 in
THF at -196 °C, and then the mixture allowed to warmsthe
yield of 30 was increased to 60%. However, the yield of
pentanitrocubane remained low. A substantial amount of
tetranitrocubane was always present in the product mixture,
although clearly its sodium salt could not have survived
treatment with NO2Cl. This strongly suggests that NO2Cl
oxidized 19 to the radical 31 which then (1) abstracted H-atom
from the solvent (THF), generating tetranitrocubane, (2) reacted
with chlorine from NO2Cl, generating 30, or (3) reacted with
NO2 from NO2Cl, forming pentanitrocubane 29 (Scheme 5).
If the mechanism in Scheme 5 be correct then dinitrogen
tetraoxide (N2O4) should be a better choice than NO2Cl for the
preparation of pentanitrocubane. Indeed, condensation of excess
N2O4 onto the surface of a frozen solution of 19 [from 4 and
(TMS)2NNa] in THF at -196 °C followed by warming
produced a 60:40 mixture of 1,2,3,5,7-pentanitrocubane (29)
and tetranitrocubane 4. The reaction was otherwise exception-
ally clean. Careful NMR analysis revealed the presence of ∼3-
5% of (trimethylsilyl)tetranitrocubane (22), probably resulting
from combination of 19 (or maybe 31) with a product derived
from some reaction between (TMS)2NNa or (TMS)2NH with
N2O4. A similar amount of hexanitrocubane (Vide infra) was
also formed.
More Highly Nitrated Cubanes. We hoped that reaction
of electrophilic nitrating reagents with o-nitro carbanions (e.g.,
19) derived from 1,3,5,7-tetranitrocubane would provide access
to more highly nitrated cubanes. The literature is unsupportive
of this optimistic view. Nitro group transfers from nitrate esters
to enolates and similar reactions are well known, but as best
we could determine no successful nitration of a localized
(nonresonance stabilized) group 1A organometallic had been
confirmed at the time of this work.38 Nitrodemetalations of
group IA alkyl or aryl organometallics have been observed, but
these cases involve highly delocalized anions such as cyclo-
pentadienyl,39 and diphenylmethyl.40 Thus, it is not surprising
that we were almost entirely unsuccessful in nitrating the sodium
salt of 1,3,5,7-tetranitrocubane with NO2BF4, acetyl nitrate,
tetranitromethane, amyl nitrate, or similar reagents; only traces
of the desired compounds could be found, and then only
sometimes. Poor results were also obtained with nitrosating
reagents.
As tetranitrocubane 4 has very low solubility in chloroform
it could be separated and recovered from the reaction product
mixture simply. Final purification of pentanitrocubane 29 was
achieved by column chromatography on silica gel. Pure 29 was
isolated in 40% yield. Its structure was confirmed by NMR
spectroscopy and single-crystal X-ray analysis (see later).
1,2,3,5,7-Pentanitrocubane is colorless and highly crystalline.
It is the first nitrated cubane to contain adjacent nitro groups.
Despite the predictions of naysayers, pentanitrocubane is stable
under ordinary conditions, showing no obvious shock sensitivity
or special thermal sensitivity. Differential scanning calorimetry
(20 °C/min) showed that decomposition sets in above 250 °C,
rather like other cubane derivatives.9
The reaction of nitryl chloride (NO2Cl) with 19 in THF at
-75 °C provided a crucial lead. Two new compounds were
formed reproducibly in 10-15% yields. Their structures were
(38) Very recently successful nitrations of metalated aromatics have been
demonstarted: see ref 41 and Stagliano, K. W.; Malinakova, H. C.;
Takayama, A. Synth. Commun. 1997, 27, 2413.
We sought to improve the yield of pentanitrocubane from
low-temperature N2O4 nitration of 19 by simple variations in
reaction conditions, but were unrewarded. Nonetheless, ap-
plication of the method to simpler systems revealed that we
had found a fairly general and quite useful reaction for nitrating
group IA organometallics. Its success, interestingly, is appar-
(39) Thiele, J. Chem. Ber. 1900, 33, 666.
(40) (a) Feuer, H. In The Chemistry Of Amino, Nitroso and Nitro
Compounds and Their DeriVatiVes; Supplement F. Patai, S., Ed.; Wiley:
New York, 1982; Supplement F, p 805. (b) For an interesting special
example, see: Sitzmann, M. E.; Kaplan, L. A.; Angres, I. J. Org. Chem.
1977, 42, 563.

Synthesis and Chemistry of 1,3,5,7-Tetranitrocubane
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9597
ently dependent on the reaction occurring at the melting interface
between N2O4 and the frozen solution of the anion.41
a 30:70 mixture of 1,2,3,4,5,7-hexanitrocubane (34) and pen-
tanitrocubane 29 (Scheme 6). The yield was not good. The
separation of the two polynitrocubanes was quite difficult,
aggravated by the fact that hexanitrocubane, although stable in
acetone and acetonitrile, is not stable in alcohols.
The relatively high ratio of hydrogen abstraction (f 4) to
nitration (f 29) in the reaction of tetranitrocubylsodium (19)
with N2O4 (cf., Scheme 5) might be accounted for by a mismatch
between tetranitrocubyl radical (31) and NO2 radical, both of
which are quite electrophilic. Therefore, placing an electron-
donating substituent on the cubane might be helpful. Earlier
work in this Laboratory on cubyl cation42 suggested σ-donors
like trimethylsilyl. Attempted metalation/nitration of the tri-
methylsilyl-substituted tetranitrocubane 22 gave, unexpectedly,
a mixture of penta- (29) and tetranitrocubanes (4), both without
a TMS substituent. Control experiments showed that metalation
Scheme 6
Metalation/nitration of the TIPS-substituted pentanitrocubane
32savailable either by silylation of 33 or, better, by N2O4
nitration of the anion from 23sprovided for a much improved
preparation of hexanitrocubane. Formation of an anion from
32 was slow when sodium bis(trimethylsilyl)amide was em-
ployed; the analogous potassium base, K(TMSN)2, proved to
be much more reactive and generated the necessary anion easily
at -105 °C in THF. Interfacial nitration with N2O4 gave a
mixture of (triisopropyl)hexa- and (triisopropyl)pentanitrocu-
banes, 35 and 32, in a 60:40 ratio by 1H NMR analysis.
Surprising perhaps, but clearly related to our earlier observation
on the behavior of (trimethylsilyl)tetranitrocubane 22, column
chromatography on silica gel of the mixture of 35 and 32
selectively removed the TIPS group from 35. The liberated
hexanitrocubane 34 was then easily separated on the column
from the still silylated (triisopropylsilyl)pentanitrocubane, itself
quite stable under these conditions. A 30% isolated yield of
each compound pure was obtained (Scheme 7).
of 22 with (TMS)2NNa is more difficult than the corresponding
metalation of 4 (reasonable considering the electron donating
effects of the silyl substituent); desilylation (f 19) occurs
instead. This probably proceeds by nucleophilic substitution
at silicon, a process that would be hindered by bulkier
substituents on the silicon atom. Metalation of (triisopropyl)-
silyltetranitrocubane 23 with (TMS)2NNa proceeded without
competition from desilylation. Nitration of the resulting anion
with N2O4 gave a mixture of the TIPS-substituted penta- and
tetranitrocubanes 32 and 23. 1H NMR analysis of the crude
gave a 82:18 ratio for nitration vs hydrogen abstraction, much
improved over the 60:40 ratio obtained from nitration of the
unsilylated tetranitrocubane. The TIPS-substituted pentani-
trocubane 32 was isolated in 70% yield by column chromatog-
raphy. This material turned out to be very useful in the
preparation of hexanitrocubane.
Scheme 7
Hexanitrocubane. It was not at all surprising, knowing the
reactions of the tetranitrocubane, to find that pentanitrocubane
29 was metalated readily by (TMS)2NNa in THF to give the
sodium salt 33. However, it soon became clear that such salts
are less stable than those of tetranitrocubane. The recovery of
skeletally intact cubanes from sequences involving formation
of 33 and quenching it with various electrophiles (e.g., TIPSCl
f 32) was significantly lower than that from comparable
reactions involving 19. Nonetheless, the methodology for
nitration of 4 via its sodium salt could be extended to nitration
of pentanitrocubane 29 via its salt 33. Reaction of 29 with
(TMS)2NNa in THF at -78 °C followed by reaction with N2O4
at a melting interface in the manner described previously gave
The structure of 1,2,3,4,5,7-hexanitrocubane (34) was proven
by NMR and X-ray analysis (Vide infra). The compound forms
nicely crystalline, stable solvates with some nonprotic polar
solvents. The bis-acetonitrile solvate was used for the X-ray
analysis.
Spectral Comparisons. The introduction of additional nitro
groups onto tetranitrocubane 4 causes substantial downfield
shifts of the remaining cubyl protons: 4 (tetranitro), δ ) 6.085;
29 (pentanitro), 6.43; 34 (hexanitro), 6.72 ppm. The effects of
variously positioned nitro groups on the 13C chemical shift of
a cubyl carbon are roughly constant, as noted previously for
lesser nitrated cubanes.9 At that time, when no vicinal dini-
trocubanes were available, the best approximation for predicting
(41) Tani, K.; Lukin, K. A.; Eaton, P. E. J. Am. Chem. Soc. 1997, 119,
1476.

9598 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
Lukin et al.
Table 1. Bond Lengths (Å) and Angles (deg) for
1,2,3,5,7-Pentanitrocubane (29)
C(1)-C(2)
C(1)-C(6)
C(2)-C(5)
C(3)-C(4)
C(4)-C(7)
C(6)-C(7)
C(1)-N(1)
C(3)-N(3)
C(7)-N(7)
N(1)-O(1B)
N(2)-O(2A)
N(3)-O(3A)
N(5)-O(5B)
N(7)-O(7B)
1.535(13)
1.574(13)
1.563(12)
1.558(12)
1.549(14)
1.535(14)
1.482(13)
1.510(12)
1.452(13)
1.217(11)
1.217(11)
1.198(10)
1.209(10)
1.217(11)
C(1)-C(8)
C(2)-C93)
C(3)-C(8)
C(4)-C(5)
C(5)-C(6)
C(7)-C(8)
C(2)-N(2)
C(5)-N(5)
N(1)-O(1A)
N(2)-O(2B)
N(3)-O(3B)
N(5)-O(5A)
N(7)-O(7A)
1.553(14)
1.558(13)
1.523(14)
1.531(14)
1.526(14)
1.569(13)
1.457(12)
1.490(13)
1.198(11)
1.202(11)
1.183(10)
1.199(11)
1.199(11)
C(2)-C(1)-C(8)
C(8)-C(1)-C(6)
C(1)-C(2)-C(5)
C(8)-C(3)-C(4)
C(4)-C(3)-C(2)
C(5)-C(4)-C(3)
C(6)-C(5)-C(4)
C(4)-C(5)-C(2)
C(5)-C(6)-C(1)
C(6)-C(7)-C(4)
C(4)-C(7)-C(8)
C(3)-C(8)-C(7)
N(1)-C(1)-C(2)
N(1)-C(1)-C(6)
N(2)-C(2)-C(3)
N(3)-C(3)-C(8)
N(3)-C(3)-C(2)
N(5)-C(5)-C(4)
N(7)-C(7)-C(6)
N(7)-C(7)-C(8)
O(1B)-N(1)-C(1)
O(2A)-N(2)-C(2)
O(3A)-N(3)-C(3)
O(5B)-N(5)-C(5)
O(7B)-N(7)-C(7)
92.7(8) C(2)-C(1)-C(6)
90.8(7) C(1)-C(2)-C(3)
89.2(6) C(3)-C(2)-C(5)
92.0(7) C(8)-C(3)-C(2)
91.0(7) C(5)-C(4)-C(7)
89.2(7) C(7)-C(4)-C(3)
91.3(8) C(6)-C(5)-C(2)
91.8(7) C(5)-C(6)-C(7)
89.1(7) C(7)-C(6)-C(1)
90.3(8) C(6)-C(7)-C(8)
90.5(7) C(3)-C(8)-C(1)
89.0(7) C(1)-C(8)-C(7)
125.0(8) N(1)-C(1)-C(8)
121.2(8) N(2)-C(2)-C(1)
125.9(8) N(2)-C(2)-C(5)
121.7(8) N(3)-C(3)-C(4)
127.5(8) N(5)-C(5)-C(6)
129.2(8) N(5)-C(5)-C(2)
123.8(9) N(7)-C(7)-C(4)
126.5(8) O(1A)-N(1)-C(1)
115.8(10) O(2B)-N(2)-C(2)
116.2(10) O(3B)-N(3)-C(3)
118.0(9) O(5A)-N(5)-C(5)
116.6(10) O(7A)-N(7)-C(7)
90.4(7)
86.8(7)
88.0(7)
93.0(7)
88.8(8)
88.5(7)
91.2(7)
89.5(8)
89.0(7)
Figure 1. The molecular structure and numbering scheme for 1,2,3,5,7-
pentanitrocubane (29). Anisotropic thermal envelopes are shown at the
25% population level.
91.7(7)
87.4(8)
88.5(7)
126.6(9)
125.8(8)
128.2(9)
122.2(8)
121.2(8)
122.1(8)
123.7(8)
119.1(10)
119.1(9)
115.5(9)
118.0(10)
119.9(11)
117.3(10) O(1A)-N(1)-O(1B) 125.0(10)
O(2B)-N(2)-O(2A) 124.7(10) O(3B)-N(3)-O(3A) 126.5(9)
O(5A)-N(5)-O(5B) 125.4(11) O(7A)-N(7)-O(7B) 122.8(10)
tetranitrocubane (4).9 The average interior CCC angle is 88.6°
at the even-numbered carbons in 29 and 91.4° at the odd-
numbered, nitro-bearing carbons; this pattern and magnitude of
angular distortion is found as well in 4. Although the connection
may be tenuous, pentanitrocubane, as noted earlier, has thermal
and shock stability similar to those of lesser nitrated cubanes.9
Molecules of 29 are arranged in a interesting way within the
crystal polymorph examined. The packing consists of close-
packed layers of molecules in the ac plane, perpendicular to
the 2-fold b axis; these layers are seen edge-on in Figure 2.
Each of these layers is paired with another that is related by a
glide plane (e.g., x, 0.5 - y, 0.5 + z), with which it closely
interlocks. A single molecule of 29 has 29 nonbonded O‚‚‚O
and O‚‚‚C contacts at less than van der Waals contact
distances;42 the shortest are 2.748Å (O‚‚‚O) and 2.801Å (O‚‚‚C).
Only three of these contacts are across the boundary between
the interlocked layers.
Figure 2. A schematic drawing of the full contents of the unit cell of
1,2,3,5,7-pentanitrocubane (29) in projection down the a axis. Four
vertical layers of molecules are shown; each represents a two-
dimensional layer of molecules in the crystal. The two layers on the
left, related to one another by the glide plane, interlock to form a closely
packed double layer. The same applies to the two layers on the right.
There are 26 sub-van der Waals contacts across the two layers on the
left and, separately, the two on the right. All of the close contacts occur
within symmetry-related pairs. The shortest of these is C6‚‚‚O2A (2.80
Å), more than 0.4 Å less than the 3.22 Å vdW contact. There are only
three sub-van der Waals contacts across the boundary between these
pairs of layers.
the chemical shift of any particular cubyl carbon in a nitrocubane
was δ (ppm) ) 49.5 + 38i + 5.9o + 0.7p - 4.9m, where i, o,
p, m are, respectively, the number of ipso, ortho, para, and meta
nitrated carbons. Now, with data for more highly nitrated
cubanes available, the approximation can be refined to δ (ppm)
) 49.5 + (38 - 4o)i + 5.9o + 0.7p - (4.9 - o)m. This should
prove useful identifying the numerous nitrocubanes still sought,
but yet to be made, particularly those of reduced symmetry.
Structural Features. The structure of pentanitrocubane
(Figure 1) was obtained by single-crystal X-ray analysis. As
only small crystals were available, the precision of the atom
positions is low. Nonetheless, it is abundantly clear from the
X-ray analysis that the fifth nitro group, although surrounded
by three others, enters without the need for notable distortion
in the cubane framework. The bond angles and lengths (Table
1) are essentially the same as those corresponding in 1,3,5,7-
The distances seen between molecules of 29 are surprisingly
short in the interlocked double layers shown in Figure 2. A
(42) Eaton, P. E.; Zhou, J. P. J. Am. Chem. Soc. 1992, 114, 3118.
(43) We examined the literature, briefly, to see if more recent results
might have rendered obsolete traditional values for van der Waals radii,
particularly the value for aliphatic carbon. A recent, massive review
(Rowland, R. S.; Taylor, R. J. Phys. Chem. 1996, 100, 7384) examined
close contacts in 28 000 suitable crystal structures. The analysis pointed
out inaccuracies in a few of the still-used 1948 Pauling vdW values (Pauling,
L. The Nature of the Chemical Bond, 2nd ed.; Cornell University Press:
Ithaca, NY, 1948; pp 187-193) but generally supported the 1964 tabulation
of Bondi (Bondi, A. J. Phys. Chem. 1964, 68, 441). This gives a carbon
van der Waals radius of 1.7 Å and an oxygen radius of 1.52 Å).

Synthesis and Chemistry of 1,3,5,7-Tetranitrocubane
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9599
Table 2. Bond Lengths (Å) and Angles (deg) for
1,2,3,4,5,7-Hexanitrocubane (34) Bis-acetonitrile Solvatea
C(1)-N(1)
1.466(3)
1.475(3)
1.471(3)
1.213(3)
1.214(3)
1.203(3)
1.220(3)
1.212(3)
1.220(3)
1.444(4)
C(1)-C(8)
C(1)-C(2)
C(1)-C(8)*
C(2)-C(3)*
C(2)-C(3)
C(3)-C(8)
C(3)-C(2)*
C(8)-C(1)*
C(1S)-N(1S)
1.550(3)
1.561(3)
1.572(3)
1.565(3)
1.559(3)
1.548(3)
1.565(3)
1.572(3)
1.137(3)
C(2)-N(2)
C(3)-N(3)
N(1)-O(1A)
N(1)-O(1B)
N(2)-O(2A)
N(2)-O(2B)
N(3)-O(3A)
N(3)-O(3B)
C(1S)-C(2S)
C(8)-C(1)-C(2)
C(8)-C(1)-C(8)*
C(2)-C(1)-C(8)*
C(8)-C(3)-C(2)
C(8)-C(3)-C(2)*
C(2)-C(3)-C(2)*
N(1)-C(1)-C(8)
N(1)-C(1)-C(2)
N(1)-C(1)-C(8)*
N(2)-C(2)-C(3)
N(2)-C(2)-C(1)
N(2)-C(2)-C(3)*
N(3)-C(3)-C(8)
N(3)-C(3)-C(2)
N(3)-C(3)-C(2)*
91.5(2) C(3)-C(2)-C(1)
91.1(2) C(3)-C(2)-C(3)*
90.7(2) C(1)-C(2)-C(3)*
91.6(2) C(3)-C(8)-C(1)
91.4(2) C(3)-C(8)-C(1)*
89.8(2) C(1)-C(8)-C(1)*
124.6(2) O(1A)-N(1)-O(1B) 126.7(2)
125.3(2) O(2A)-N(2)-O(2B) 126.6(2)
123.5(2) O(3A)-N(3)-O(3B) 126.8(2)
130.0(2) O(1A)-N(1)-C(1)
125.1(2) O(1B)-N(1)-C(1)
122.6(2) O(2A)-N(2)-C(2)
126.4(2) O(2B)-N(2)-C(2)
125.8(2) O(3A)-N(3)-C(3)
121.4(2) O(3B)-N(3)-C(3)
88.0(2)
90.2(2)
88.9(2)
88.8(2)
89.1(2)
88.8(2)
Figure 3. A view of the unit cell and observed packing assembly in
a crystal of the bis-acetonitrile solvate of 1,2,3,4,5,7-hexanitrocubane
(34). The acetonitrile methyl group is disordered. All molecules are
more or less in the plane of the paper. The next layer is shifted laterally.
Solvent coordination is indicated (see Figure 4).
117.4(2)
115.9(2)
118.5(2)
114.9(2)
117.8(2)
115.3(2)
N(1S)-C(1S)-C(2S) 179.1(3)
a The following symmetry transformations are used to generate
equivalent atoms: 1 - x, y, 1.5 - z. (Entries are indicated with
asterisks.)
the supramolecular assembly. We suggest that similarly strong
dipolar forces are acting in the polynitrocubanes. Obviously,
there are large dipoles in these molecules, for nitro groups are
known for their electron-withdrawing nature. The negative,
oxygen poles of the nitro groups are exposed and, as seen in
29, can approach closely the hindered, electron-deficient faces
of the cubane skeleton. In the double layers seen in Figure 2,
the nitro group on C5 manages remarkably to approach the
trinitrosubstituted face (C2C3C4C5) of a neighboring molecule
to less than 3 Å; similarly, the nitro group on C2 approaches
the dinitrosubstituted face (C4C5C6C7) to 2.8 Å.
As mentioned earlier, density is of critical importance to the
performance of an explosive. The overall packing in 29
described above leads to a density of 1.96 g/cm3 at 21 °C, a
value exceeded by very few other CHNO compounds. Even
so, it is not idle to speculate on the existence of a yet more
dense arrangement. All of the aforementioned close contacts
involve just two of the five nitro groups of 29, those on C2 and
C5; the other nitro groups are relatively uncrowded. Thus, an
all-around closer (higher density) packing assembly might be
achieved in some other polymorph of 1,2,3,5,7-pentanitrocubane
or in one of its two positional isomers.
We were unable to obtain a single crystal of 1,2,3,4,5,7-
hexanitrocubane free of solvent. The packing in the multimol-
ecule unit cell of the bis-acetonitrile solvate is shown in Figure
3. Note the “channels” of solvent molecules which themselves
approach the hexanitrocubane molecules very closely (indeed
much less than van der Waals contact distances; see Figure 4).
In 34, each acetonitrile “guest” approaches a trinitrosubstituted
face of the cubane frame. The shortest CN‚‚‚C distance, 2.87
Å, is well below the Bondi43 van der Waals close contact
distance of 3.25 Å and closer than any intermolecular CN‚‚‚C
seen in the Cambridge Structural Database, other than the
CN‚‚‚C contact of 2.76 Å seen between acetonitrile molecules
in the solvent cavity of a cyclophane clathrate.45
Figure 4. The molecular structure and numbering scheme for the bis-
acetonitrile solvate of 1,2,3,4,5,7-hexanitrocubane (34). Anisotropic
thermal envelopes are shown at the 25% population level. The left and
right sides of the molecule are related by a vertical 2-fold axis of
symmetry. There are numerous sub-van der Waals close contacts
between the solvent nitrogen atom (N1S) and components of the
cubane: N1S‚‚‚N1, 2.963 Å; N1S‚‚‚C8, 2.867 Å; N1S‚‚‚C1, 2.976 Å;
M1S‚‚‚C3 3.031 Å.
few distances less than accepted van der Waals contacts43 are
seen in most crystal structures. These are closely scrutinized
by crystallographers for sometimes they indicate a local
electrostatic or covalent bonding attractive enough to overcome
the “Lennard-Jones repulsion” energy that is rapidly rising at
the van der Waals contact distance. Otherwise, close (com-
pressed) contacts are the result of a global packing-energy
compromise made to produce a better fit overall. The non-
bonded O‚‚‚O contacts in 29 fit this latter pattern, but the nitro
oxygen to cubane carbon contacts near the N2 and N5 nitro
groups are so short that an unusual attraction must be present
if “accepted” van der Waals radii values are correct.43
Eight pairs of C‚‚‚O approaches seen in 29 in the range
(2.80-3.06 Å) are considerably below the 3.22 Å Bondi vdW
contact distance. Such short distances are not unprecedented,
but have heretofore been seen only between polar moieties and
trigonal carbons. For example, in chloranil there is a 2.85 Å
CdO‚‚‚C close contact between near neighbors in the crystal.44
This indicates a strong electrostatic interaction that dominates
(44) Chu, S. C.; Jeffrey, G. A.; Sakurai, T. Acta Crystallogr. 1962, 15,
661.
(45) Hirotsu, K.; Kamitori, S.; Higuchi, T.; Tabushi, I.; Yamamura, K.;
Nonoguchi, H. J. Inclusion Phenom. 1984, 2, 215.
The numerical results for bond lengths and angles in 34 (Table
2) are quite precise; the average distance esd is only 0.003 Å.

9600 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
Lukin et al.
the reaction mixture after 15 min irradiation showed that nearly all of
5 had been converted to the diacid chlorides 6, 7, and 8 in very
approximately a 6:10:1 ratio.
Cubane-1,3-dicarboxylic acid chloride (6): δ 4.09 (m, 2 H), 4.43
(m, 2 H), 4.81 ppm (m, 2 H).
As this is far better than we were able to do for pentanitrocu-
bane, it is reassuring that detailed comparisons with other cubane
and nitrocubane structures reveal again that there are no
particular distortions due to steric forces engendered by the
numerous adjacent nitro groups. The six (NO2)C-C-C(NO2)
cube edge lengths average 1.562 Å, not significantly longer than
the average of 1.557 Å for the six HC-C-C(NO2) edges. The
longest C-C bonds in the molecule, 1.572 Å between C(1) and
C(8) and the same between the symmetry mates, are of the latter
group. Their elongation might be due to an interaction with
the solvent. The two solvent molecules closest to 34 in the
crystal are nearest the disubstituted face of the molecule and
the electronegative ends of the strong acetonitrile dipoles
approach most closely (2.867 Å) the two CH corners thereof
(Figure 4). It should be mentioned, however, that this effect,
although interesting and probably real, is still quite small and
probably does not indicate a weakening in the carbon bonding
within the cubane framework. Cubane carbon-carbon bond
distances have been observed in stable cubanes all the way from
1.522 Å to 1.607 Å.
Cubane-1,4-dicarboxylic acid chloride (7): δ 4.47 ppm (s).
Cubane-1,2-dicarboxylic acid chloride (8): δ 4.05 (m, 2 H), 4.47
ppm (pseudo t, J ) 3.5 Hz, 4 H).
1
After 50 min of irradiation, a similar H NMR analysis (CDCl3)
indicated that most of the cubane diacid chlorides had been converted
into a mixture of the 1,3,5-, 1,2,4- and 1,2,3-triacid chlorides 9-11 in
roughly a 6:4:2 ratio.
Cubane-1,3,5-tricarboxylic acid chloride (9): δ 4.15 (dq, J ) 0.8,
5.2 Hz, 1 H), 4.80 (dd, J ) 2.8, 5.2 Hz, 3 H), 5.16 ppm (dq, J ) 0.8,
2.8 Hz, 1 H).
Cubane-1,2,4-tricarboxylic acid chloride (10): δ 4.47 (m, 1 H),
4.51 (m, 2 H), 4.86 ppm (m, 2 H).
Cubane-1,2,3-tricarboxylic acid chloride (11): δ 4.08 (m, 1 H),
4.40-4.60 ppm (m, 4 H).
In a separate experiment the solution of cubanecarboxylic acid
chloride 5 in oxalyl chloride was irradiated for 2.75 h using the same
lamp but in a quartz photoreactor. During this time the temperature
of the reaction mixture rose slowly to about +12 °C. Evaporation of
the excess oxalyl chloride left a mixture composed mainly of the
Conclusion
1
In summary, we have demonstrated for the first time direct
ortho metalation of saturated nitro compounds, and we have
shown that the resulting o-nitro carbanions are valuable synthetic
intermediates useful for introduction of various substituents. We
have developed direct nitration of o-nitro carbanions with
dinitrogen tetraoxide and have used this method for preparation
of the first vicinal nitrocubanes, pentanitrocubane and hexani-
trocubane, both of which have been characterized.
tetraacid chlorides 12-14 in 70:8:22 ratio as determined by H NMR
(CDCl3).
Cubane-1,3,5,7-tetracarboxylic acid chloride (12): δ 5.14 ppm
(s).
Cubane-1,2,4,7-tetracarboxylic acid chloride (13): δ 4.88 ppm
(s).
Cubane-1,2,3,7-tetracarboxylic acid chloride (14): δ 4.51 (dt, J
) 0.8, 5.2 Hz, 1 H), 4.80 (dd, J ) 2.4, 5.2 Hz, 2 H), 5.17 ppm (dt, J
) 0.8, 2.4 Hz, 1 H).
The mixture was treated with dry ether (10 mL) and put in a freezer
overnight. The resulting precipitate was collected and washed with a
small amount of cold ether to give 12 as a white solid (1.1 g, 31%,
>95% purity). This material was used without further purification.
Cubane-1,3,5,7-tetraisocyanate (17). As this procedure inVolVes
as an intermediate a highly explosiVe acyl azide (16) it must be done
behind a safety shield. A face mask and heaVy gloVes should be worn.
The acyl azide must be kept in solution at all times.
Procedure A. Cubanetetracarboxylic acid chloride 14 (0.7 g, 2
mmol) and 2,6-di-tert-butylpyridine (0.2 g, 1 mmol, 15 equiv %) were
mixed in dry, alcohol-free chloroform (60 mL) and stirred under
nitrogen until the mixture became homogeneous (∼25 min). A solution
of trimethylsilyl azide (0.97 g, 8 mmol) in chloroform (alcohol free, 5
mL) was added dropwise over 5 min at room temperature. The reaction
mixture was stirred for 1.5 h at room temperature (f tetraacyl azide
16) taking care that none of the solution splashed up into the necks.
An additional 200 mL of dry, alcohol-free chloroform was added and
the solution heated at full reflux for 1 h. The solution was cooled and
the solvent removed at room temperature in Vacuo leaving crude
tetraisocyanate 17 containing di-tert-butylpyridine, etc. This material
was used in the next step without further purification.
Procedure B (Recommended). Anhydrous hydrogen chloride,
bubbled through concentrated H2SO4, was introduced slowly for 5 min
into a solution of 2,6-di-tert-butylpyridine (0.100 g, 0.523 mmol) in
10 mL of dry diethyl ether. A white precipitate formed. The solvent
was removed in Vacuo at room temperature. The residue was pumped
at <1 mm for 15 min (no longer!). The white salt so obtained, assumed
to be 2,6-di-tert-butylpyridine‚2HCl,25 was dissolved in dry, alcohol-
free chloroform (4.00 mL). Separately, cubanetetracarboxylic acid
chloride 14 (79.6 mg, 0.225 mmol) was refluxed in oxalyl chloride (5
mL) for about 2 h in oxalyl chloride to ensure that any that had been
partially hydrolyzed during weighing or storage was reconverted to
14. After the solution was cooled to room temperature it was clarified
by filtration through a fine frit (Schlenk technique), and the oxalyl
chloride was removed in Vacuo at room temperature. The residue was
dissolved in dry, alcohol-free chloroform (5.6 mL). A clear solution
was obtained. A catalytic amount of the pyridinium salt solution (191
mL, 4.78 mg, 0.0250 mmol) was added. A solution of trimethylsilyl
Experimental Section
General. Unless otherwise specified, NMR spectra were run in
acetone-d6: 1H NMR spectra at 400 MHz and referenced to the central
line of acetone-d5 (δ 2.05 ppm); 13C NMR spectra at 100.6 MHz and
referenced to the central line of acetone-d6 (δ 29.8 ppm). Proton
chemical shifts are (0.01 ppm. Carbon chemical shifts are (0.1 ppm.
Observed 13C-207Pb couplings are given as JC-Pb. Merck silica gel 60
(230-400 mesh) was used for column chromatography. The eluting
solvent is given parenthetically in the individual procedures. THF was
distilled from sodium benzophenone ketyl. THF-d8 was dried over
activated 4 Å molecular sieves. All metalations were carried out under
argon in oven-dried glassware. “Removal of solvent in Vacuo” and
similar phrases generally refer to rotary evaporation at ∼50 Torr. The
evaporator bath was not heated above room temperature. Oxalyl
chloride and trimethylsilyl azide were purchased from Aldrich and
distilled before use. Dimethyldioxirane was prepared as described
earlier.8b Dinitrogen tetraoxide was purchased from Matheson.
CAUTION. Most cubane compounds are quite stable kinetically.
Nonetheless, as they are high-energy materials it is prudent to run all
reactions thereof behind safety shields. Great care should be taken to
assure that crude reaction products are not concentrated at eleVated
temperature, particularly in the presence of acidic or metallic
contaminants. Nitrating agents were always handled in a good hood,
behind safety shields.
Photochemical Chlorocarbonylation of Cubanecarboxylic Acid
Chloride with Oxalyl Chloride. Cubanecarboxylic acid46 (1.48 g, 10
mmol) was added in four portions to oxalyl chloride (100 mL) well
stirred at room temperature. Immediate, vigorous gas evolution was
observed. The resulting solution of cubanecarboxylic acid chloride [5,
1H NMR (CDCl3) δ 4.02 (m, 1 H), 4.06 (m, 3 H), 4.44 ppm (m, 3 H)]
was transferred into a Pyrex photoreactor and cooled to -5 °C by
pumping through the reactor jacket a 40:60 ethylene glycol/water
mixture precooled to -12 °C in an external cooling bath. The cold
solution was irradiated using a Hanovia, medium-pressure, 450-watt,
mercury arc lamp. 1H NMR analysis (CDCl3) of a sample taken from
(46) Eaton, P. E.; Nordari, N.; Tsanaktsidis, J.; Upadhyaya, S. P. Synthesis
1995, 501.

Synthesis and Chemistry of 1,3,5,7-Tetranitrocubane
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9601
azide (94% by 1H NMR, 140 µL, 0.114 g contained, 0.991 mmol; the
contaminant is hexamethyldisiloxane) in dry, alcohol-free chloroform
(1.2 mL) was added dropwise over 5 min. The funnel was rinsed with
chlorofom (0.5 mL). The mixture was stirred at room temperature for
1,3,5,7-Tetranitro-2-(trimethylsilyl)cubane (22). Chlorotrimeth-
ylsilane (0.3 mL, excess) was added by syringe to the THF solution of
tetranitrocubylsodium at -78 °C. The reaction mixture was allowed
to warm to room temperature, and then the solvent and excess of
chlorotrimethylsilane were evaporated in Vacuo. The residue was
extracted with CHCl3 (3 × 2 mL). Evaporation of the solvent left 22
(14 mg, 80%) as good quality material: 1H NMR δ 0.16 (s, 9 H), 6.07
ppm (s, 3 H); 13C NMR δ -3.3 (3 C, CH3), 67.8 (3 C, CH), 71.7, 74.4
(3 C), 75.5 ppm. Attempted purification of this material by column
chromatography gave tetranitrocubane 4.
1,3,5,7-Tetranitro-2-(triisopropylsilyl)cubane (23). Chloro(triiso-
propyl)silane (21 µL, 0.10 mmol) was added by syringe in one portion
to the solution of tetranitrocubylsodium at -78 °C. The reaction
mixture was stirred at -78 °C for 5 min and then allowed to warm to
room temperature. The solvent was evaporated in Vacuo. The residue
was extracted with CH2Cl2 (3 × 3 mL). Evaporation of the solvent
followed by column chromatography (-78 °C) gave 23 (17 mg,
80%): 1H NMR (δ, CD2Cl2) 1.04-1.08 (m, 21 H), 5.86 ppm (s, 3 H);
13C NMR (CD2Cl2) δ 10.7 (3 C, CH-Si), 19.1 (6 C, CH3), 65.1 (3 C,
CH), 69.7, 74.2 (3 C), 85.5 ppm. Anal. Calcd for C17H24N4O10Si: C,
46.36; H, 5.49. Found: C, 46.09; H, 5.34.
1,3,5,7-Tetranitro-2-(ethylmercuryl)cubane (24). Ethylmercury
chloride (29 mg, 0.11 mmol) was added in one portion to the THF
solution of tetranitrocubylsodium at -78 °C. The reaction mixture
was allowed to warm to room temperature, and then the solvent was
evaporated in Vacuo. The residue was extracted with acetone (3 × 2
mL). Evaporation of the solvent followed by column chromatography
(85:15 CH2Cl2/pentane) gave 24 (14 mg, 55%): 1H NMR δ 1.3 (t, J )
8 Hz, 3 H), 1.43 (q, J ) 8 Hz, 2 H), 6.08 ppm (s, 3 H); 13C NMR δ
13.5 (CH3), 24.9 (CH2), 69.1 (3 C), 73.7, 75.4 (3 C), 107.0 ppm. Anal.
Calcd for C10H8HgN4O8: C, 23.42; H, 1.57. Found: C, 23.07; H, 1.61.
1,3,5,7-Tetranitro-2-(triethylplumbyl)cubane (25). Triethyllead
chloride (33 mg, 0.10 mmol) was added in one portion to the solution
of tetranitrocubylsodium at -78 °C. The reaction mixture was stirred
at -78 °C for 5 min then allowed to warm to room temperature. The
solvent was evaporated in Vacuo. The residue was extracted with
CH2Cl2 (3 × 3 mL). Removal of the solvent followed by column
chromatography (80:20 CH2Cl2/pentane) gave 25 (19 mg, 66%): 1H
NMR (δ, CD2Cl2) 1.49 (t, J ) 8 Hz, 9 H), 1.94 (q, J ) 8 Hz, 6 H),
5.88 ppm (s, 3 H); 13C NMR (δ, CD2Cl2) 13.9 (3 C, CH3, JC-Pb ) 31
Hz), 17.6 (3 C, CH2Pb, JC-Pb ) 175 Hz), 67.5 (3 C, CH), 71.2, 73.6
(3 C), 79.1 ppm (JC-Pb ) 538 Hz). Anal. Calcd for C14H18N4O8Pb:
C, 29.12; H, 3.14. Found: C, 29.27; H, 3.00.
2,4,6,8-Tetranitrocubane-1,3-dicarboxylic Acid Dimethyl Ester
(26a). Tetranitrocubane 3 (14 mg, 0.05 mmol) was treated with excess
(5 equiv) sodium bis(trimethylsilyl)amide in THF at -78 °C. Quench-
ing with CO2 and workup as described above for 20 gave a mixture of
tetranitrocubane mono- and dicarboxylic acids (20 and 26) in a 70:30
ratio. These were converted into the corresponding mono- and dimethyl
esters by treatment with diazomethane. For the diester 26a: 1H NMR
δ 3.90 (s, 6 H), 6.34 ppm (s, 3 H); 13C NMR δ 54.5 (CH3), 66.3 (CH),
72.5, 74.4, 76.5, 159.3 ppm (CdO).
1
about 4 h after which the H NMR spectrum of an aliquot showed no
cubyl signal at higher field than the strong, sharp singlet at 4.69 ppm
corresponding to the tetraacyl azide. Some multiplets at lower field
due to partial rearrangement of acyl azide groups to isocyanates were
just visible. The solution was diluted with dry, alcohol-free chloroform
(28 mL) and heated at reflux for 1 h and then cooled. An aliquot
examined by 1H NMR examination of an aliquot showed a small amount
of material with a multiplet at 4.36 ppm. The desired tetraisocyanate
appeared as a strong, sharp singlet at 4.32 ppm (estimated yield by
integration > 85-95%). Evaporation of the solvent in Vacuo furnished
the crude tetraisocyanate as a slightly yellow solid.
1,3,5,7-Tetranitrocubane (4). A solution of the crude tetraisocy-
anate 17 (procedure A) in carefully dried acetone (20 mL) was added
dropwise over 1.5 h to a stirred solution of dimethyldioxirane (0.06
M) in acetone (660 mL) and water (60 mL) cooled in an iced water
bath. After the addition was completed the bath was removed. The
reaction mixture was stirred at room temperature overnight. Evapora-
tion of solvents in Vacuo left crude 4 as a cream-colored solid. This
was dissolved in methanol (45 mL) and the solution refluxed with
activated charcoal for 15 min. The charcoal was filtered and the
solution concentrated to ∼15 mL. If it were still brown or dark yellow,
the treatment with charcoal was repeated as necessary. Ultimately a
pale yellow solution was obtained. It was concentrated to 5 mL and
then cooled. The precipitate was tetranitrocubane 4 (0.26 g, 47% overall
from 12) identical spectroscopically to that previously obtained:9 1H
NMR δ 6.085 ppm (s); 13C NMR δ 67.9, 73.2 ppm.
Proton-Deuterium Exchange on Tetranitrocubane. Tetrani-
trocubane (4, 10 mg) was dissolved in CD3OD (2.5 mL) at room
temperature. A sodium methoxide solution, prepared earlier by reacting
sodium (10 mg) with CD3OD (1 mL), was added in one portion with
stirring. The reaction mixture was stirred at room temperature for 45
min then added dropwise into stirred aqueous hydrochloric acid (1 M,
25 mL). The mixture was extracted with ethyl acetate (3 × 10 mL).
The combined organic layer was washed with water and dried over
Na2SO4. Removal of the organic solvents left 4-d4 (8.8 mg, 87%):
13C NMR δ 67.6 (t, JC-D ) 28.2 Hz, C-2, C-4, C-6, C-8), 73.0 ppm
(C-1, C-3, C-5, C-7).
2,4,6,8-Tetranitrocubylsodium (19). Sodium bis(trimethylsilyl)-
amide (65 µL, 1 M in THF, 0.065 mmol) was added dropwise to a
solution under argon of tetranitrocubane 4 (14 mg, 0.05 mmol) in THF
(2 mL) cooled to -78 °C. The resulting yellow solution of 19 [1H
NMR δ, (THF-d8, -78 °C) δ 5.54 ppm (s)] was stirred for 5 min at
this temperature. The solution was kept at -78 °C until used. A batch
was made for each of the experiments below.
2,4,6,8-Tetranitrocubanecarboxylic Acid Methyl Ester (20a).
Carbon dioxide was bubbled for 10 min through the solution of
tetranitrocubylsodium at -78 °C. The reaction mixture was allowed
to warm to room temperature and then the stream of CO2 was
discontinued. The solvent was evaporated in Vacuo. The residue was
diluted with water (2 mL), acidified to pH 1 with concentrated
hydrochloric acid, and then extracted with 1:1 ether-THF (3 × 1 mL).
The extract was dried with Na2SO4. Evaporation of the solvent left
crude acid 20 containing 15% of tetranitrocubane 4 by 1H NMR. The
acid was converted into its methyl ester by treatment with ethereal
diazomethane. Crystallization from 95:5 chloroform-acetonitrile gave
the pure ester 20a (13 mg, 75%): 1H NMR δ 3.84 (s, 3 H), 6.22 ppm
(s, 3 H); 13C NMR δ 54.3 (CH3), 67.0 (3 C, CH), 72.0, 73.4, 75.2 (3
C), 159.9 ppm (CdO). Anal. Calcd for C10H6N4O10: C, 35.10; H,
1.76. Found: C, 35.01; H, 1.59.
1,3-Bis(trimethylsilyl)-2,4,6,8-tetranitrocubane (27). Tetranitrocu-
bane 4 (14 mg, 0.05 mmol) was treated with excess (3 equiv) of sodium
bis(trimethylsilyl)amide in THF at -78 °C. Quenching the reaction
mixture with chlorotrimethylsilane and workup as described above gave
a mixture of mono- and bis(trimethylsilyl)tetranitrocubanes 22 and 27
in a 20:80 ratio. The bis(trimethylsilyl) compound 27 (13 mg, 60%)
was obtained pure by crystallization from 60:40 hexane/chloroform:
1H NMR δ 0.19 (s, 18 H), 6.10 ppm (s, 2 H); 13C NMR δ -3.2 (6 C,
CH3), 67.1 (2 C, CH), 72.7 (2 C), 75.4 (2 C), 75.8 ppm (2 C). Anal.
Calcd for C
4.59.
14H20N4O8Si2: C, 39.24; H, 4.70. Found: C, 39.03; H,
1,3,5,7-Tetranitro-2-iodocubane (21). 1,2-Diiodoethane (28 mg,
0.1 mmol) was added to the THF solution of tetranitrocubylsodium at
-78 °C. The reaction mixture was allowed to warm to room
temperature, and then the solvent was evaporated in Vacuo. The residue
was extracted with chloroform (5 × 1 mL). Evaporation of the solvent
followed by column chromatography (CH2Cl2) gave 21 (12 mg, 60%):
1H NMR δ 6.19 ppm (s); 13C NMR δ 43.9, 67.6 (3 C, CH), 72.1, 75.2
ppm (3C). Anal. Calcd for C8H3IN4O8: C, 23.43; H, 0.74. Found:
C, 24.11; H, 0.97.
1,3-Bis(triisopropylsilyl)-2,4,6,8-tetranitrocubane (28). Tetranitro-
cubane 4 (14 mg, 0.05 mmol) was treated with excess (3 equiv) of
sodium bis(trimethylsilyl)amide in THF at -78 °C. Quenching the
reaction mixture with excess chlorotriisopropylsilane (4 equiv) and
workup as described above gave a mixture of mono- and bis(triiso-
propylsilyl)tetranitrocubanes 23 and 28 in a 20:80 ratio. Compound
28 (21 mg, 70%) was isolated by column chromatography (50:50
CH2Cl2/pentane: 1H NMR (δ, CD2Cl2) 1.07 (d, J ) 7.5 Hz, 36 H),
1.25-1.35 (m, 6 H), 5.90 ppm (s, 2 H); 13C NMR (δ, CD2Cl2) 12.0 (6

9602 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
Lukin et al.
C, CH-Si), 19.8 (12 C, CH3), 63.7 (2 C, CH), 72.7 (2 C), 77.6 (2 C),
84.5 ppm (2 C). Anal. Calcd for C26H44N4O8Si2: C, 52.32; H, 7.43.
Found: C,52.66; H, 7.59.
dropwise to a solution of pentanitrocubane 29 (10 mg, 0.03 mmol) in
THF (4 mL) cooled to -78 °C under nitrogen. This was frozen under
nitrogen with liquid nitrogen and nitrated with N2O4 (∼0.4 mL, excess)
as described for the preparation of 29. The solvent and excess N2O4
were removed in Vacuo. The residual solid was extracted with acetone
(2 × 1 mL). Evaporation left a mixture of penta- and hexanitrocubanes
1,2,3,5,7-Pentanitrocubane (29). The solution of tetranitrocubyl-
sodium under a nitrogen atmosphere was frozen into a glass by using
a liquid nitrogen bath. Excess N2O4 (∼0.4 mL, condensed into a
separate tube and then evaporated through a cannula into the reaction
flask) was deposited onto this solid at -196 °C. The resulting two-
phase mixture was warmed briefly in a room temperature methanol
bath with shaking until it had just melted and become homogeneous,
then it was stirred and allowed to warm to 10 °C. (The whole warming
process took about 5 min.) The solvent and excess N2O4 were
evaporated in Vacuo. The residual solid was extracted with acetone (2
× 1 mL). Evaporation of the extract gave a mixture of penta- and
tetranitrocubanes 29 and 4 in a 60:40 ratio as detertmined by 1H NMR.
It was triturated with 96:4 chloroform/methanol (6 × 2 mL). Evapora-
tion of the extract left a mixture of penta- and tetranitrocubanes in
85:15 ratio. Pentanitrocubane (29, 7 mg, 40%) was obtained pure by
chromatographic (95:5 dichloromethane-acetonitrile) separation: 1H
NMR δ 6.43 ppm (s, 3 H); 13C NMR δ 66.5 (3 C, CH), 72.8, 78.8 (3
C), 93.0 ppm. X-ray-quality crystals of 29 were prepared by crystal-
lation from 1-propanol.
1
29 and 34 in 70:30 ratio as determined by H NMR.
Method B. Potassium bis(trimethylsilyl)amide (100 µL, 0.5 M in
toluene, 0.05 mmol) was added dropwise over 1 min to a solution under
nitrogen of (triisopropylsilyl)pentanitrocubane 32 (20 mg, 0.04 mmol)
in THF (1.75 mL) cooled to -105 °C (isooctane-N2 bath). The solution
of the potassium salt of 32 so made was then immediately frozen with
liquid nitrogen and nitrated with N2O4 (∼0.4 mL, excess) as described
for the preparation of 29. The solvent and excess N2O4 were evaporated
in Vacuo. The solid remaining was extracted with acetone (2 × 1 mL).
1
Evaporation left a 60:40 mixture (by H NMR) of silylated pentani-
trocubane 32 with the two hexanitrocubanes 34 and 35 (their ratio was
quite variable). Column chromatography (CH2Cl2) gave first 32 (6 mg,
30%). Further elution (96:4 CH2Cl2/acetonitrile) gave hexanitrocubane
34 (5 mg, 30%) as a white solid: 1H NMR (acetone-d6) δ 6.73 ppm
(s, 2 H); 13C NMR δ 65.0 (2 C, CH), 78.9 (2 C), 83.1 (2 C), 91.0 ppm.
Crystallization from 95:5 chloroform/acetonitrile gave X-ray-quality
crystals of the bis-acetonitrile solvate hexanitrocubane.
Sometimes small amounts of (triisopropylsilyl)hexanitrocubane 35
were isolated: 1H NMR (δ, CD2Cl2) 1.13 (d, J ) 7.5 Hz, 18 H), 1.40-
1.50 (m, 3 H), 6.23 ppm (s,1 H).
Single-Crystal X-ray Analysis of Compounds 29 and 34. Details
of the crystal structure investigation are available on request from the
Director of the Cambridge Crystallographic Data Centre, 12 Union
Road, GB-Cambridge CB21EZ (UK).
1,2,3,5,7-Pentanitrocubane (29). Crystals of 29 belong to mono-
clinic space group P21/c with a ) 6.637(3), b ) 23.275(14), c )
7.860(5) Å, â ) 113.21(5)°, Z ) 4, Dx ) 1.959 g/cm3. The thin-plate
crystal was quite small (0.01 × 0.35 × 0.40 mm), leading to a somewhat
weak set of diffraction intensities, and thus low precision in the
numerical results. R ) 0.0668, Rw2 ) 0.1647 for 686 unique reflections
with I > 2σ(I). [R ) 0.1136 for all 1051 data.]
1,2,3,4,5,7-Hexanitrocubane (34). Crystals of 34 are monoclinic,
space group C2/c, with a ) 13.0102(7), b ) 7.7242(6), c ) 18.0758(12)
Å, â )95.404(7)°, Z ) 4, Dx ) 1.676 g/cm3, all at T ) -50 °C. The
asymmetric unit is comprised of 1/2 of 34, and one acetonitrile molecule;
a crystallographic 2-fold axis passes through the centers of the di-and
tetrasubstituted faces of the cube. R ) 0.0407, Rw2 ) 0.1062 for all
1241 unique reflections. The indicated precision in the numerical results
is ∼4 times higher than that for 29; the appearance of bonding-electron
density peaks in the final electron density map indicated above-average
diffraction quality.
Reaction of Tetranitrocubylsodium with Nitryl Chloride. Nitryl
chloride (∼0.2 mL, excess) was evaporated over 1 min into a solution
of tetranitrocubylsodium in THF at -78 °C. The mixture was stirred
at -78 °C for 5 min, then allowed to warm to room temperature. The
solvent and excess nitryl chloride were evaporated in Vacuo. The solid
remaining was identified by 1H NMR as a mixture of tetranitrocubane
4, chlorotetranitrocubane 30, and pentanitrocubane 29 in 75:15:10 ratio.
1-Chloro-2,4,6,8-tetranitrocubane (30). A batch of the solution
of tetranitrocubylsodium under nitrogen was frozen into a glass using
a liquid nitrogen bath. NO2Cl (∼0.25 mL, excess) was deposited onto
the solid at -196 °C. The resulting mixture was warmed briefly with
shaking until it had melted and become homogeneous, and then it was
stirred and allowed to warm to 10 °C (over ∼5 min). The solvent and
excess NO2Cl were evaporated in Vacuo. The remaining solid was
extracted with acetone (2 × 1 mL). Evaporation of the solvent left a
mixture of chlorotetranitrocubane 30, tetranitrocubane 4, and pentani-
1
trocubane 29 in 60:33:7 ratio as determined by H NMR. Chloroni-
trocubane 30 (7 mg, 45%) was obtained by chromatography (95:5
dichloromethane-acetonitrile): 1H NMR δ 6.22 ppm (s, 3 H); 13C
NMR δ 66.3 (3 C, CH), 72.9, 78.7 (3 C), 79.4 ppm. Anal. Calcd for
C8H3ClN4O8: C, 30.16; H, 0.94. Found: C, 30.79; H, 1.00.
1,2,3,5,7-Pentanitro-4-(triisopropylsilyl)cubane (32). Sodium bis(tri-
methylsilyl)amide (100 µL, 1 M in THF, 0.1 mmol) was added dropwise
to a solution under argon of (triisopropylsilyl)tetranitrocubane 23 (33
mg, 0.075 mmol) in THF (3 mL) at -78 °C . The orange solution
was stirred for 5 min at this temperature then frozen and nitrated with
N2O4 (∼0.4 mL, excess) as described for the preparation of 29. The
solvent and excess N2O4 were removed in Vacuo. The solid remaining
was extracted with CH2Cl2 (2 × 2 mL). Evaporation followed by
column chromatography (CH2Cl2) gave 32 (25 mg, 70%, containing
Acknowledgment. This work was supported by the U. S.
Army Research, Development and Engineering Center via Geo-
Centers, Inc. and by the Office of Naval Research. J.L. is
grateful to Geo-Centers for a leave of absence permitting him
to extend his participation in the work at The University of
Chicago. J.H. thanks the Deutsche Forschungsgeminshaft for
fellowship support.
1
8-10% of 23 by H NMR). Crystallization from CHCl3 gave pure
32: 1H NMR (δ, CD2Cl2) 1.11 (d, J ) 7 Hz, 18 H), 1.21-1.32 (m, 6
H) 6.04 ppm (s, 2 H); 13C NMR (δ, CD2Cl2) 11.6 (3 C, CH-Si), 19.6
(6 C, CH3), 63.1 (3 C, CH), 73.3, 74.6, 79.7, 84.1, 90.9 ppm.
1,2,3,4,5,7-Hexanitrocubane (34). Method A. Sodium bis(trimeth-
ylsilyl)amide (38 µL of 1 M solution in THF, 0.038 mmol) was added
JA970552Q