Diels–Alder adduct 14 must not have been formed from 9
during the reaction; the precursor to 16 has to be 17, the basic
skeleton of which is identical with 14 but having the alkenic F
and R groups interchanged. The formation of the unexpected
intermediate tricyclic compound 17 can be rationalised most
simply on the basis of a retro-Diels–Alder reaction of 12 to give
6
the cyclohexa-2,4-dienylmethyl fluoroketene 18—a very rare
reaction type—followed by the alternative intramolecular
Diels–Alder cyclisation as shown in Scheme 2. Intermolecular
(
4 + 2) p reactions of ketenes to form six-membered carbocyclic
7
rings are uncommon, but an intramolecular process of this type
has been recorded.8
The present work begs the question: why do the complex
molecular dynamics involved in the rearrangement of 12 to 17
take place in preference to the direct formation of 14 having the
same basic carbon skeleton? We have no real answer to this
4
question but models show that the formation of structures 6 and
1
4 from 2 and 9 respectively are sterically more demanding than
for the formation of compounds 3 and 12, which were isolated
under milder conditions. Consequently, even the formation of 7
is likely to proceed via this new molecular rearrangement
reaction.
Fig. 2 Molecular structure of 16 (50% displacement ellipsoids; double
bonds shown in black).
Notes and references
†
Crystal data for 12: C15
5 9
H F O, M = 372.2, monoclinic, space group C2/c
compound 15 which would have been formed if the Diels–Alder
adduct 14 had been produced directly from the Claisen
intermediate 9 by analogy with the mechanism proposed earlier
for the formation of 7 from 1 via 2 and then 6. The material was
shown by X-ray crystallography† to have the structure 16 (Fig.
(No. 15), a = 21.242(3), b = 6.219(2), c = 20.254(2) Å, b = 93.05(1)°,
23
21
c
U = 2671.8(8) Å, Z = 8, D = 1.851 g cm , m = 1.84 mm , T = 150
K, 3080 reflections (2394 unique) with 2q @ 150°, 247 variables, R
1
=
0.037 and wR
2
= 0.098 on 1926 data with I ! 2s(I), max. residual Dr =
0.25 e Å . For 16: C15 O, M = 372.2, monoclinic, space group P2/c
No. 13), a = 13.446(2), b = 11.033(1), c = 19.249(1) Å, b = 108.41(1)°,
2
3
5 9
H F
(
2
) which enabled all the NMR data to be rationalised.
The formation of compound 16 (a racemate, but having the
2
3
21
c
U = 2709.4(5) Å, Z = 8, D = 1.825 g cm , m = 1.81 mm , T = 150
K, 5009 reflections (4172 unique) with 2q @ 135°, 492 variables, R
1
=
enantiomeric structure shown in Scheme 2 when formed from
2 with the configuration given), isomeric with the starting
0
0
4
2
.046 and wR = 0.100 on 3238 data with I ! 2s(I), max. residual Dr =
1
2
3
.23 e Å . X-Ray experiments were performed on a Rigaku AFC6S
material 8, poses an intriguing mechanistic problem since the
¯
-circle diffractometer (Cu-Ka radiation, l = 1.54184 Å, 2q/w scan mode);
structure solution (direct methods) and least-squares refinement (non-H
2
atoms anisotropic, all H refined isotropically, against F of all data) with
SHELX-97 software (G. M. Sheldrick, University of G o¨ ttingen, Germany,
1
997); CCDC 182/1326. See http://www.rsc.org/suppdata/cc/1999/1549/
for crystallographic data in .cif format.
1
2
3
G. M. Brooke and D. H. Hall, J. Fluorine Chem., 1982, 20, 163.
G. M. Brooke, Tetrahedron Lett., 1971, 2377.
G. M. Brooke and D. H. Hall, J. Chem. Soc., Perkin Trans. 1, 1976,
1
463.
4
5
G. M. Brooke, J. Chem. Soc., Perkin Trans. 1, 1974, 233.
M. W. Buxton, R. H. Mobbs and D. E. M. Wotton, J. Fluorine Chem.,
1
972/73, 2, 231.
6
D. J. Pollart and H. W. Moore, J. Org. Chem., 1989, 54, 5444; T. T.
Tidwell, Ketenes, Wiley, New York, 1995.
7
8
J. A. Hyatt and P. W. Raynolds, Org. React., 1994, 45, 159.
T. Miyashi, H. Kawamoto, T. Nakajo and T. Mukai, Tetrahedron Lett.,
1
979, 155.
Scheme 2
Communication 9/03209F
1550
Chem. Commun., 1999, 1549–1550