COMMUNICATION
sis. The authors concluded that the reaction proceeds
through two concerted cycloadditions, the first of which was
proposed to be reversible. The syn adduct was reported as
the exclusive product in both cases, a result which was erro-
neously attributed to a second dienophile addition to the
“outside” face of the “cup-shaped” monoadduct. (The syn-
bisadduct must arise from both dienophiles approaching
from the same face of the dienyne.) In this report, neither
reaction conditions nor yields were disclosed.
The rather obvious synthetic potential of this transforma-
tion coupled with its neglected status—in both an experi-
mental and theoretical sense—led us to initiate a research
programme in this area to further investigate this intriguing
reaction. At the outset, it was our intention to: a) improve
upon the very low yields in all previous reports of this trans-
formation; b) examine the stereoselectivity of the reaction;
c) investigate the scope and limitations of the process; d)
develop an understanding of both the mechanism and the
stereoselectivity of the reaction, by way of high level com-
putational analysis.
of toluene and xylene as solvents did not lead to improved
yields and again, prompted much longer reaction times.
Under these optimised conditions only about 67% of the
dienyne was accounted for in the isolated products and re-
covered dienyne; a complex mixture of polar degradation
products made up the mass balance. The bisadduct was
shown to survive the reaction conditions with little degrada-
tion. Re-subjecting ene product 3 to standard reaction con-
ditions, however, afforded only polar degradation products.
These results lead us to suspect that the ene reaction is a
significant competing pathway to that leading to the double
dehydro-Diels–Alder product, and the polar products
making up the mass balance from the reaction depicted in
Scheme 3 result from the ene product.
Keeping with NMM as dienophile, the optimised reaction
conditions were applied to a range of 1,5-dien-3-ynes
(Table 1, entries a–d). In each case, only the endo,endo,syn-
isomer of the bisadducts was observed, as confirmed by
single-crystal X-ray structure analyses.[6] The yields of the
double Diels–Alder products were in the 40–60% range. In-
terestingly, the reaction of 2,5-dimethylhexa-1,5-dien-3-yne 8
(Table 1, entry d) was considerably slower than the ꢁsemicy-
clicꢂ dienynes 1, 4 and 6 (Table 1, entries a, b and c). Phenyl
acetylenes have been successfully used in single (intramolec-
ular)[4] dehydro-Diels–Alder reactions reported by Baddar
and co-workers,[7] Saꢃ and co-workers,[8] and others.[9] Inter-
estingly, phenyl enyne 10 (Table 1, entry e) did not furnish
the desired double dehydro-Diels–Alder product, but in-
stead the major product 11 was that derived from the ene
pathway.[10]
Next, the scope of the reaction with respect to the dieno-
phile was examined (Table 1, entries f–j). Both Z and E
electron poor dienophiles participate in the reaction, with
isolated yields in the 36–61% range. Whilst only the expect-
ed endo,endo,syn-bisadduct was observed in the reaction
with maleic anhydride (Table 1, entry f) and N-phenylmalei-
mide (Table 1, entry g), N-tert-butylmaleimide afforded two
diastereoisomeric bisadducts 14 and 15 in a ca. 2:1 ratio
(Table 1, entry h). As elucidated by single-crystal X-ray
structure analysis,[6] the major product was once again the
endo,endo,syn-isomer 14, but a new isomer, namely the en-
do,exo,syn-adduct 15, was also formed.
The reactions with E-dienophiles were particularly inter-
esting. Fumaronitrile afforded two stereoisomeric bisadducts
16 and 17 in a 72:28 ratio (Table 1, entry i).[11] In contrast,
dimethyl fumarate gave four products (18–21) in a high
overall yield (Table 1, entry j). This is at odds with previous
reports.[12] The two minor products, 20 and 21, are analogous
to those obtained from the fumaronitrile reaction. The two
major products 18 and 19, however, are the result of an anti-
double addition pathway followed by a formal 1,3-hydride
shift.
Upon analysis of prior art,[3] some general features of this
unique double Diels–Alder reaction are evident: it proceeds
only at high temperature; the dienophile must be highly ac-
tivated and yields are always low. N-Methylmaleimide
(NMM) and 1,2-dicyclohexenyl acetylene 1 were chosen as
representative reaction partners for our optimisation study.
During the course of these investigations the following pa-
rameters were varied: a) temperature (110–1908C); b)
method of heating (conventional versus microwave); c) re-
action time (5 min–45 h); and d) reaction stoichiometry
(2.2–10.0 equiv of dienophile). We also explored the effects
of various radical inhibitors, solvents and treated glassware.
Initially, and consistent with literature reports,[3] we obtained
poor isolated yields of products (11–22%). After much ex-
perimentation, we were delighted to find that under opti-
mised reaction conditions (Scheme 3) the double dehydro-
Diels–Alder adduct 2 was obtained in a much improved
61% yield along with ene product 3 in a 4% yield. The
structure and stereochemistry of both products were con-
firmed by single-crystal X-ray analysis.[6] Notably, only one
stereoisomer of both products could be detected. Lowering
the reaction temperature failed to improve the yield of the
bisadduct and below about 1308C, the solvent-free transfor-
mation required prohibitively long reaction times. The use
These results prompt several questions relating to the
mechanism and origin of stereoselectivity in the double de-
hydro-Diels–Alder sequence: 1) Do the two dienophiles add
through concerted Diels–Alder processes? 2) What is the
nature of the intermediate and is it formed reversibly?
Scheme 3. Double dehydro-Diels–Alder reaction of 1,2-dicyclohexenyl
acetylene (1) with NMM. Reagents and conditions: a) dienyne
1
(1.0 equiv), NMM (5.0 equiv), hydroquinone (0.1 equiv), mwave, 300 W,
1908C, 30 min, 2: 61%, 3: 4%.
Chem. Eur. J. 2010, 16, 760 – 765
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
761