646 J. CHEM. RESEARCH (S), 1998
J. Chem. Research (S),
1998, 646±647$
Amidoethylation of Anthracene Hydride by
N-Aroylaziridines: Inner-sphere Single Electron
Transfer (SET) and Radical Coupling confirmed$
P.-Y. Lin and H. Stamm*
Faculty of Pharmacy, University of Heidelberg, Neuenheimer Feld 346, D-69120 Heidelberg,
Germany
Regioselectivity (near 1:1) of substitutive ring opening of 1-benzoyl-2-methylaziridine by anthracene hydride is
incompatible with common nucleophilic attack and thus confirms the radical coupling path.
Reactions of N-aroylaziridines with excess anthracene
hydride (AH ) may be exempli®ed by means of 1a (Scheme
1). Aziridino ketyl 4a is an essential intermediate3 generated
by benzylic fragmentation (BFR)4 of the rapidly formed3
carbonyl adduct 2a. Homolytic ring cleavage of 4a aords
the amidatoalkyl radical 5a, a precursor of the main product
6. The second product is 7.
When the aziridine ring of 1a carries substituents, ana-
logues of 7 are obtained3 unless they arise from 2-phenyl-
aziridines and are unstable under usual conditions.3,5 The
assumption3 that 7 and its analogues are formed by coup-
ling of amidatoalkyl radicals with anthracenide Aꢀ was
supported by a regioselectivity of ring opening that seemed
to exclude a direct SN2-like path to analogues of 7 and
hence also to 7. Subsequently it was found6 from a study
of 1-acyl-2,2-dimethylaziridines that SN2-like ring opening
may require planarization of the nitrogen pyramid thereby
shifting the mechanism to a borderline type whose regio-
selectivity is compatible with the AH results. This reopened
the mechanistic question since the very fast initial carbonyl
attack is reversible.7
Ring opening of 1-acyl-2-methylaziridines by strong
nucleophiles was recently8 shown to strongly prefer cleavage
of the N±CH2 bond. AH and 2-methyl-1-pivaloylaziridine
provided a mixture of products (total 94%) with an overall
regioselectivity isopropylamides:n-propylamides of 35:1.
Scheme 1
The reaction of xanthenyl anion (oxa analogue of AH
devoid of the BFR path) with 1b yielded 82% of benzoyl-
xanthene and 14.5% of amidoethylated xanthenes with an
able mixture of isomeric 10 by characteristic 1H NMR
iso to normal regioselectivity of 28:1. Thus, one may expect
a ratio of about 30:1 if i-10 and n-10 (Scheme 2) are formed
from 1b and AH only, or mainly, by nucleophilic ring
opening.
signals for the non-aromatic double bond. A doublet (J 10.4)
for H-4 at 6.70 ppm shows ®ne splitting (ca. 1.1 Hz) of
the lines from coupling with H-10. A doublet (J 10.4) of
approximated triplets (J ca. 5) at 6.08 ppm comes from H-3,
the triplets indicating attachment of the amidatopropyl
chain to position 1. Ole®nic and additional aromatic signals
are in accord with those of 2-vinylnaphthalene.10
There were at least four methyl doublets (J ca. 6.8 Hz,
at 1.02, 1.11, 1.31 and 1.42 ppm) in addition to those of
both 10. This is compatible with a mixture of i-11 and n-11
when one considers diastereoisomerism. However, part of
these signals may come from structural isomers of 11, e.g.
Y carried in position 2. Weak signals in the range of 5.9±6.7
ppm point to isomerism in the non-aromatic ring.
Two three-day runs of 10 mmol of 1a with 16 mmol
of AH Li in 200 ml of THF provided 58% (47%) of iso-
propylamide i-9, 14% (18%) of n-propylamide n-9, 9%
(4%) of i-10 and 9% (5%) of n-10 (values in parentheses
are the yields of the second run). The yields of both 10 are
crude yields in the sense that they were estimated by
1H NMR from fractions containing minor amounts of
unknown products, probably isomers of 10, one of them
being 11 (see below). But the yield ratios i:n 1 (0.8), deter-
mined from the methyl doublets at 1.21 and 0.94 ppm, are
suciently reliable. These ratios of isomeric 10 are far from
the 30:1 ratio expected for an SN2 mechanism. Both 10
arise consequently only or nearly so from coupling of
anthracenide Aꢀ (generated by BFR) with amidatoalkyl
radicals i-8 and n-8. Moreover, coupling with position 1 of
Aꢀ obviously formed traces of 11 (one or two products
with isomeric side chains). 11 was identi®ed in the insepar-
Cleavage 4b 4 i-8 will be kinetically controlled (cf. ref. 9)
and reduction of the amidatoalkyl radicals by a second
4b forms probably the primary carbanion faster than the
secondary one. It is therefore not surprising to ®nd much
more i-9 than n-9.
Experimental
The reactions were performed as described in ref. 4 starting with
17 mmol of dihydroanthracene AH2 and 16 mmol BuLi (hexane).
The reactions were quenched with acetic acid. The residue obtained
after the usual workup was chromatographed (silica gel Merck,
0.063±0.200 mm, 40 cm  4 cm, toluene±ethyl acetate 9:1); compo-
*To receive any correspondence.
$This is a Short Paper as de®ned in the Instructions for Authors,
Section 5.0 [see J. Chem. Research (S), 1998, Issue 1]; there is there-
fore no corresponding material in J. Chem. Research (M).
Lin
