First examples of arylazo derivatives of cyclooctatetraene†
Zori V. Todres and Ghevork Ts. Hovsepyan
Research Institute, The Cleveland Clinic Foundation 4075 Monticello Boulevard #205A,
Cleveland, Ohio 44121, USA
D isodium and dipotassium derivatives of 1,6-dinitrocyclooctatetraene (substrates) react with
benzenediazo compounds (reagents) in TH F to yield bis(azo)- or nitro(azo)-cyclooctatetraenes, the first
cyclooctatetraene arylazo derivatives. Structures of the azo compounds are established by conventional
methods and confirmed by X-ray crystallography. Together with the main products of electrophilic
substitution, products of substrate-to-reagent electron transfer are formed. The electron-transfer
products predominate with D M SO instead of TH F as a solvent. The difference is apparently a result of
ion-pair disintegration in the case of the substrate reactions when D M SO, a strong dissociating solvent, is
used.
M+
Introduction
NO2
NO2
The present work describes the study of alkali metal com-
pounds of nitro and azoxy derivatives which exhibit distinct
organometallic character and differ fundamentally in reactivity
from their respective anion-radicals and dianions which have
separate alkali-metal counterions.2,3 Here we focus on the
disodium and dipotassium derivatives of 1,6-dinitrocyclo-
octa-1,3,5,7-tetraene 1.
O2N
+2e, +2M+
=
O2N
M+
2
1
M = Na, K
The neutral molecule of 1 has a boat conformation with the
two nitro groups located above the boat.4
+2e, +2M+
For the di(Na/K) derivative of 1, planar and non-planar
structures are theoretically possible. The planar form 2 is char-
acterized by delocalization of the two excess electrons in the
eight-membered ring. The same manner of delocalization is
characteristic for the di(Na/K) derivative of unsubstituted
cyclooctatetraene (COT) and dimethyl-COT.5 The non-planar
form 3 requires fixation of the two excess electrons at the two
nitro groups blocked by two alkali metals (see Scheme 1).
NOOM
MOON
3
Scheme 1
Conditions for existence of forms 2 and 3
The di(Na/K) derivatives of 1 were prepared by means of inter-
change between 1 and di(Na/K) derivatives of COT in tetra-
hydrofuran (THF). An alternative approach consisted of
treatment of 1,4-dinitrocycloocta-2,5,7-triene with (Na/K)
methoxides in methanol.
The aim of our present study was to verify the prediction for
the case of azo coupling and, if successful, to obtain new
arylazo derivatives of cyclooctatetraene.
By means of spectral [IR, NMR (1H, 13C)] methods, we
earlier found that form 3 became stable in non-dissociating
solvents.6 In other words, the salts of the dianion of 1 exist in
the form of ion pairs, and the excess negative charges are local-
ized primarily on the nitro groups. In dissociating solvents, the
ion pairs are broken up, and free ions are formed. This leads to
a considerable shift of electron density in the eight-membered
ring, see structure 2.6 We had concluded in our earlier work 6
that the difference between forms 2 and 3 must be reflected in
their reactivity. The increase in delocalization of the excess elec-
trons, which is a consequence of dissociation of the ion pairs,
must lead to an increase in the level of the donor properties of
the substrates. Thus, when the free dianion 2 is treated with
electrophiles, it will more readily give up electrons than it will
enter into electrophilic substitution. The characteristic reaction
of electrophilic substitution of an aci-nitro group (NOOϪM+),
in contrast, should proceed more readily in non-dissociating
solvents, when the substrates act as the ion pairs 3.6
Interaction between di(Na/K) derivatives of 1 and benzenediazo
cations
In THF or 1,2-dimethoxyethane (DME) the disodium or di-
potassium derivative of 1 reacts with benzenediazo cations in
two ways. One of them, the minor, is one-electron reduction of
the cations. It results in N2 evolution and formation of 1 and
benzene derivatives RC6H5 (see Scheme 2 and Table 1).
This reaction proceeds to 10–20% only, whereas the main
route is the formation of bis(azo) and nitro(azo) derivatives of
COT. All appearances suggest that the key stage of the reaction
is ipso-coupling, with intermediates 4 and 5 probably being
formed (see Scheme 3).
Alkali nitrite is cleaved from the intermediate 4, while di-
nitrogen tetroxide is liberated from the intermediate 5. At the
same time, the cyclooctatriene ring is transformed into a
cyclooctatetraene ring to give the final products 6 and 7. Trans-
formation of 5 into 6 is also possible (with elimination of
p-RC6H4N2NO2).
Mono 6a–c and bis(azo) derivatives 7a–c are obtained
together, irrespective of the stoichiometric relationship between
the reagents and the substrate. Under similar conditions, yields
† This work was initially presented in the VI International Conference
on Organic Synthesis.1
J. Chem. Soc., Perkin Trans. 1, 1997
747