New one-pot synthesis of a benzonorcaradiene derivative by reduction of naphthalic anhydride with LiAlH4

Article abstract of DOI:10.1039/a700605e

Reaction of naphthalic anhydride (naphthalene-1,8-dicarboxylic anhydride) (1) or of 1,8-bis(hydroxymethyl)naphthalene (3) with LiAlH₄ in refluxing THF yields the benzonorcaradiene derivatives 2 in yields of up to 50%. It is proposed that in the formation of this (strained) product an anti-hydroalumination reaction is involved, which is facilitated by the assistance of a neighbouring alkoxide function.

Full text of DOI:10.1039/a700605e

View Article Online / Journal Homepage / Table of Contents for this issue
New one-pot synthesis of a benzonorcaradiene derivative by reduction
of naphthalic anhydride with LiAlH4
Geerlig W. Wijsman, Lars A. van der Veen, Willem H. de Wolf and
Friedrich Bickelhaupt *
Scheikundig Laboratorium, Vrije Universiteit, De Boelelaan 1083, NL-1081 HV Amsterdam,
The Netherlands
Reaction of naphthalic anhydride (naphthalene-1,8-dicarboxylic anhydride) (1) or of
1,8-bis(hydroxymethyl)naphthalene (3) with LiAlH 4 in refluxing TH F yields the benzonorcaradiene
derivatives 2 in yields of up to 50%. It is proposed that in the formation of this (strained) product an
anti-hydroalumination reaction is involved, which is facilitated by the assistance of a neighbouring
alkoxide function.
HO
OH
Results
In our efforts to synthesize strained derivatives of [5]metacyclo-
phanes,1 we were interested in using 1,8-disubstituted naphth-
alene systems as bridging groups. As the starting material, we
used the readily available naphthalic anhydride (systematic
name: naphthalene-1,8-dicarboxylic anhydride) (1). However,
in the reduction reaction of 1 with LiAlH4, we encountered an
unexpected product which was identified as the benzonor-
caradiene derivative 2 (systematic name: 1a,7b-dihydro-7-
hydroxymethyl-1H-cyclopropa[a]naphthalene) [reaction (1)].
3
(2)
THF, ∆
LiAlH4
HO
HO
O
O
O
+
+
5
2
6
3J = 7.5 Hz) and 7.04 (dd, 3J = 7.5 Hz, 4J = 1.2 Hz)] indi-
cating a trisubstituted benzene ring; (ii) of the (conjugated)
(1)
1
3
3
olefin [AB-system at δ 6.26, δA = 6.27 (dd, Jcis = 9.4 Hz, J =
3Jcis = 9.4 Hz)]; (iii) of the prochiral
LiAlH4
THF, ∆
4.7 Hz), δB = 6.24 (d,
hydroxymethyl group [δ 4.83 (AB-system: δA = 4.85, δB =
2
4.81, JAB = 12.4 Hz)]; and (iv) of the multiplets at δ 2.55
HO
OH
HO
O
(benzylic), 1.96 (allylic), 1.59 and Ϫ0.21. The latter two
high-field signals are attached to one carbon (CH-COSY),
with a geminal coupling of 3.5 Hz, which is characteristic
8
1
7b
7
7a
6
5
1a
2
+
+
for
a cyclopropane CH2-group. Also, the CH-couplings
3a
of the three cyclopropane carbons fall in the expected
range of 160–170 Hz.
4
3
3
2
4
It was found that the direct precursor of 2 was the diol 3.
When 3 was allowed to react with LiAlH4 in tetrahydrofuran
(THF ) for 140 h at reflux temperature, 50% of unreacted diol 3
was recovered after work-up; about 30% of the consumed
diol was converted to 2, 20% to 5, and about 3% to 1,8-
dimethylnaphthalene (6) [reaction (2)]. The analogously per-
formed reduction reaction of 4 (see reaction 1), of methyl
1-naphthoate (7) or of 1-hydroxymethylnaphthalene (8)
The reduction of 1 has been reported, but besides the expected
diol 3, only the cyclic ether 2-oxaperinaphthane (4) (systematic
name 2-oxa-2,3-dihydrophenalene) was mentioned as
a
byproduct; apparently the formation of 2 has not been
observed.2 The unexpected occurrence of 2 in our experiments
tempted us to investigate this reaction in more detail.
First of all, the formation of 2 turned out to be strongly
dependent on the quality of the LiAlH4 used; yields varied
from traces to 16%. Sometimes 1-hydroxymethyl-8-methyl-
naphthalene [5, reaction (2)] was also observed as a product.
Purification of 2 was achieved as follows: after crystallization
of most of the diol 3, an oily residue remained which was
separated by careful column chromatography or preferably by
TLC. An analytical sample of 2 was obtained by preparative
GLC (mp = 111 ЊC). Identification was achieved by NMR
spectroscopy; essential features of its spectrum were almost
identical with those reported in the literature for analogous
compounds.3,4 Very characteristic are the signals (i) of the
aromatic ring [δ 7.21 (dd, 3J = 7.5 Hz, 4J = 1.2 Hz), 7.12 (t,
MeO
O
HO
7
8
yielded no benzonorcaradiene derivatives; only starting material
was recovered, in addition to a small amount of 1-methyl-
naphthalene (from 7 to 8). It follows that both hydroxymethyl
groups are required for the formation of 2.
J. Chem. Soc., Perkin Trans. 2, 1997
2095

View Article Online
The reaction was also performed with 3 and LiAlD4 which
gave, after hydrolysis, [2H]2, mono-deuterated at position 7b
was proposed to give the conformer 12aЈ which, with sub-
sequent intramolecular nucleophilic substitution of the com-
plexed alkoxide function by the carbanion equivalents, leads to
the formation of the cyclopropane ring (Scheme 1).6,11
2
(1H NMR: the signal at δ 2.55 was missing; H NMR: one
signal at δ = 2.57), together with [2H]5, [2H2]6 and [2H2]9, a
dideuterated hydrogenation product of [2H]2 [reaction (3)]. The
Mechanistic studies have been performed in order to deter-
mine the stereochemistry of the reduction and of the cyclo-
propane ring formation. These studies indicate that the reaction
occurs with loss of stereochemical integrity. Thus, both (Z)-
and (E)-α,β-dimethylcinnamyl alcohol yielded only cis-1,2-
dimethylphenylcyclopropane,11 and reduction of methyl cin-
namate with LiAlD4 followed by deuterolysis yielded equal
amounts of erythro- and threo-PhCHD᎐CHD᎐CD2OH.5a
There are several mechanisms which would explain this lack of
specificity. Snyder favoured a rapid equilibrium of the initially
formed intermediate 11a with its diastereomer 11b by Al᎐C
bond fission (ki in Scheme 1). This should occur via the carban-
ions 12a and 12b (or their conformers 12aЈ or 12bЈ).5a Other
mechanisms [such as equilibration of the initially formed car-
banion (12a to 12b) prior to formation of 11, or non-
stereospecific solvolysis of the aluminium᎐carbon bond] can-
not be excluded, but seem less plausible. As only derivatives of
cinnamyl alcohol undergo this remarkable reaction, it seems6
that the phenyl substituent at C3 is necessary, obviously for the
benzylic stabilization of the carbanion (cf. 12). Whether 12 is
best described as a ‘free’ carbanion or as a carbanion having a
partially negatively charged carbon atom in close contact to
aluminium in the five-membered cycle remains unclear,
although the latter seems more logical.10
HO
HO
D
D
+
HO
OH
H
D
D
[2H2]9
[2H]2
LiAlD4 H2O
(3)
THF, ∆
HO
D
D
3
+
[2H]5
[2H2]6
formation of the last presumably occurs by reduction of the
styrene type double bond of [2H]2.5 This reduction and also the
high yield of [2H]2 (about 50% isolated yield) is possibly due to
the ‘good quality’ and high activity of the LiAlD4. We cannot
exclude the alternative explanation suggested by a referee that
H/D isotope effects are involved. However, a systematic investi-
gation of this aspect has not been performed.
In our specific case, however, the situation is more complex.
Here the double bond is part of a naphthalene ring which has
steric consequences, and, more importantly, the initial attack of
a hydride at C1 must be accompanied by loss of the resonance
energy of one of the two rings of naphthalene! First we discuss
the two possible reaction pathways which will lead to the
five-membered oxa–alumina cycles 13a and 13b, which, in
analogy to 11,10 are believed to be intermediates in the reaction
(Scheme 2).
Starting from the diol 3, the bisaluminate 14 is a logical
primary product. In the normal syn-hydroalumination, 14
would be transformed to 13a via the transition state 15. How-
ever, this syn-hydroalumination is unlikely for several reasons.
In the first place, the initially formed intermediate 13a is rela-
tively strained. As the dihydronaphthalene has a fairly rigid
structure, the trans-five-membered ring in 13a is more strained
than the cis-five-membered ring in 13b; this follows con-
vincingly from the inspection of models. Furthermore, if the
sequence 14→15→13a really is the reaction pathway, it would
be difficult to understand why the monofunctional derivatives
4, 7 and 8 do not follow the same course, which in reality is not
the case: neither reduction of the double bond nor cyclo-
propane formation was observed.
The two cyclic intermediates 13a and 13b differ in the pos-
ition of the five-membered ring; they can interconvert by Al᎐C
bond fission in which the (planar?) allylic carbanion 16 is an
intermediate. However, in contrast to Scheme 1 where a ‘simple’
rotation of the benzylic C᎐C bonds in 12 interconverts both
stereoisomers, this kind of rearrangement requires a more
complex conformational movement to transfer the aluminium
from one side of the ring to the other; it would occur with more
extensive, unfavourable separation of carbanionic charge and
aluminium center. Hence, the transformation of 13a to 13b is
thermodynamically favourable, but might be kinetically more
difficult than in rotationally ‘free’ systems like 12.
Discussion
Formation of a cyclopropane ring during the reduction of an
allylic alcohol with LiAlH4 has precedents. In the mid 1960s,
Jorgenson and Friend reported the formation of a cyclo-
propane derivative in the reduction of cinnamyl alcohol.6
Independently, Uyeda and Cram discovered the formation of a
cyclopropane system in the reduction of a cinnamic acid deriva-
tive.7 Another example was found in the reduction of substi-
tuted 3-hydroxymethyl-3,4-dihydro-2-naphthoic acid lactones
to yield substituted tetrahydro-1H-cyclopropa[a]naphthalenes.8
The mechanism postulated for the (aluminium) hydride
reduction of cinnamyl alcohol involves a syn-hydroalumination
of the double bond as outlined in Scheme 1. In the first step, the
H
Ph
H
D
Ph
H
H
D
Ph
H
H
2
3
ki
D
Al
Al
-
-
R
R
O
O
O
R
R
Al
R
-
11a
11b
R
10
H
H
Ph
H
H
D
H
O
Ph
Ph
D
-
-
-
H
D
O
O
Al
R
R
AlR2
AlR2
12a'
12b'
12a
-AlR2O–
-AlR2O– 'slow'
'slow'
H
H
Ph
H
H
Ph
D
D
Scheme 1
aluminium species is complexed to the alkoxide group to give
10. It has been suggested that the following transformation
proceeds via a hydride transfer from the complexed aluminium
species in 10 to C2, with simultaneous formation of a (benzylic)
carbanion at C3 (12a). This is, in a fast step, transformed to the
five-membered aluminium cycle 11a, which is believed to be an
intermediate in the reaction.9,10 As the last step, ring opening
The intermediates 13a and 13b are formal products of a
syn- and anti-hydroalumination, respectively (Scheme 2). The
normal syn-addition, resulting in 13a, is highly unlikely for the
reasons discussed above. The alternative anti-addition to yield
13b directly normally does not occur as it would require
approach of the hydride and of the aluminium of the Al᎐H
bond to take place from opposite sides of the ring.
2096
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
HO
OH
AlR2
O
D
R2Al
O
-
H
3
18
+ 2 [AlR2D2]–
- 2 HD
O
HO
R2Al
O
D
O
O
–
D
R2A–l
AlR2
-
H
AlR2
H
D
D
R
Al
[2H]2
13b
14
O
O
syn-hydroalumination
anti-hydroalumination
D
-
O
O
O
O
D
H
-
-
AlR2
-
R2Al
-
R2Al
D
R2Al
H
D
D
19
Scheme 3
17
15
three-membered ring. This is an important difference compared
with the previous cases which involve attack of the double bond
of an allylic alcohol.6–8 Due to the relative ease of the reaction
(refluxing THF ), we were encouraged to investigate whether,
aided by a similarly favourable substitution pattern, a mono-
cyclic benzene ring would also be cyclopropanated. The
thermodynamic odds were against us, because they are clearly
less favourable for benzene [∆∆fHЊ(benzene Ϫ 1,3-cyclo-
hexadiene) = ϩ1.4 kcal molϪ1] than for naphthalene [∆∆fHЊ-
(naphthalene Ϫ 1,2-dihydronaphthalene) = Ϫ6.2 kcal molϪ1].12
Nevertheless we tried to synthesize norcaradiene 20 from diol
21, but without success [reaction (4)]. An additional unfavour-
- D–
- D–
-
R2Al
O
R2Al
O
O
O
D
D
-
-
AlR2
AlR2
H
H
13a
13b
R2Al
O
O
D
AlR2
H
–
HO
HO
OH
HO
OH
*
LiAlH4
16
X
(4)
THF, ∆
Scheme 2
20
22
21
However, in this specific case, an anti-hydroalumination by a
different mechanism is feasible: attack of a hydride of one alu-
minate group may occur on the ipso-carbon of the neighbour-
ing group from the top as shown in 17. The double bond is
simultaneously activated by electrophilic attack by the other
aluminate function from underneath during formation of the
five-membered ring intermediate 13b. In this anti-
hydroalumination, both aluminate functions are necessary, in
contrast to the syn-hydroalumination of cinnamyl alcohol.
Examination of models suggests that the neighbouring group
participation is favoured by the formation of a relatively strain-
free transition state. This neighbouring group participation also
explains the non-reactivity of monofunctional derivatives such
as 4, 7 and 8.
After formation of the cyclic intermediate 13b, the reaction
may proceed as postulated for the cinnamyl reductions. Ring
opening of 13b with subsequent nucleophilic displacement of
the aluminate group yields the cyclopropane ring product [2H]2
(Scheme 3). This may occur either via the ‘normal’ ring opening
in which the same aluminate is the leaving group (pathway a,
transition state 18) or, again, neighbouring group assistance
may occur; the ‘free’ aluminium replaces that in the five-
membered ring (pathway b, transition state 19) followed by sub-
stitution of the aluminium alkoxide function through attack by
the (quasi) carbanion. Both reaction pathways yield the same
product by an analogous SN 2-type process. Finally, work up
and decomplexation yield [2H]2.
able factor in this case may be that the aluminated hydroxyethyl
group, which must serve as neighbouring group, is too flexible
and points away from the ipso-carbon atom (marked by a
star *) which has to be attacked by the hydride ion. Therefore,
a more rigid structure (such as 22) might prove to be a better
candidate for this interesting reaction.
Experimental
General
The 1H NMR, H NMR and 13C NMR spectra were recorded
2
on a Bruker MSL-400 spectrometer operating at 400, 61.42 and
100.6 MHz respectively. The 2H NMR samples were measured
in CHCl3, the other samples in CDCl3, with CHCl3 as internal
standard (δ = 7.27). Chemical shifts are given in ppm; J values
in Hz.
HRMS spectra were measured at
spectrometer operating at an ionization potential of 70 eV.
GLC was performed on an Intersmat P120 apparatus using a
glass column (1.5 m × Љ; 15% SE-30 on Chromosorb WAW 60–
a Finnigan Mat-90
1
4
–
80) with H2 as carrier gas.
LiAlH4 was purchased from Baker and Aldrich, LiAlD4 was
purchased from Janssen Chimica. Tetrahydrofuran was freshly
distilled from LiAlH4.
1a,7b-Dihydro-7-hydroxymethyl-1H-cyclopropa[a]naphthalene
(2)
To a suspension of LiAlH4 (27.4 mmol, 1.04 g) in THF (75 ml),
diol 3 (7.17 mmol, 1.35 g) was added in small portions. The
The most surprising aspect of the formation of the cyclo-
propane ring in 2 is that in this process, the aromaticity of one
ring of the naphthalene system is broken to form a strained
J. Chem. Soc., Perkin Trans. 2, 1997
2097

View Article Online
reaction mixture was stirred for 6 days under reflux. Then the
reaction mixture was cooled and the excess hydride carefully
quenched with water. The mixture was poured into a cold 5%
HCl aqueous solution and extracted with diethyl ether. The
combined organic layers were washed with brine and dried over
anhydrous MgSO4. After filtration, the filtrate was concen-
trated and unreacted 3 was crystallized from the solution.
Removal of 3 (0.5 g) and concentration of the mother liquor
yielded a yellow oil (0.47 g). Purification by column chroma-
tography (silica/pentane–diethyl ether) or by TLC yielded 2,
contaminated with small amounts of 4. Further purification
was easily achieved by preparative GLC, yielding 2 as a crystal-
line material, mp 111 ЊC (Found: C, 83.64; H, 7.05. Calc. for
C12H12O: C, 83.67; H, 7.02%); [HRMS (C12H12O) calc.
172.0888. Found 172.0884]; m/z 172 (M ϩ), 154 (M ϩ Ϫ H2O,
[2,7b-2H2]-1a,2,3,7b-Tetrahydro-7-hydroxymethyl-1H-cyclo-
propa[a]naphthalene. [2H2]9: δH (400 MHz; CDCl3) 7.19 [1 H,
dd, J 7.5 and unresolved, C(6)H], 7.06 [1 H, t, J 7.5, C(5)H],
6.99 [1 H, dd, J 7.56 and unresolved, C(4)H], 4.81 [2 H, s,
C(8)H2], 2.64 [1 H, dd, J 15.8 and 2.4, C(3)HAHB], 2.47 [1 H,
dd, J 15.8 and 5.8, C(3)HAHB], 2.06 [1 H, m, C(2)H], 1.99 (1 H,
br s, OH), 1.58 [1 H, m, C(1a)H], 0.96 [1 H, dd, J 8.2 and 4.8,
C(1)Hexo], Ϫ0.21 [1 H, dd, J 5.6 and 4.8, C(1)Hendo]; δD (61.42
MHz; CHCl3) 2.10 [1 D, s, C(7b)D], 1.78 [1 D, s, C(2)D];
δC(100.6 MHz; CDCl3) 8.6 [t, J 161, C(1)], 13.9 [d, J 160, C(1a)],
16.0 [m, C(2)], 18.9 [t, J 20, C(7b)], 25.7 [t, J 129, C(3)], 63.7 [t,
J 143, C(8)], 124.4 [d, J 159, C(5)], 126.1 [d, J 160, C(6)], 128.4
[d, J 163, C(4)], 132.5 [s, C(7a)], 134.6 [s, C(3a)], 138.0 [s, C(7)];
m/z 176 (M ϩ), 158 (M ϩ Ϫ H2O, 100%), 157 (78), 143 (34), 142
᝽
᝽
(34), 130 (39), 129 (37), 116 (25).
᝽
᝽
100%), 153 (66), 141 (52), 128 (40), 115 (30); δH (400 MHz;
CDCl3) 7.21 [1 H, dd, J 7.5 and 1.2, C(6)H], 7.12 [1 H, t, J 7.5,
C(5)H], 7.04 [1 H, dd, J 7.5 and 1.2, C(4)H], 6.27 [1 H, dd, J 9.4
and 4.7, C(2)H], 6.24 [1 H, d, J2,3 9.4, C(3)H], 4.85 [1 H, d, JAB
12.4, C(8)H], 4.81 [1 H, d, JAB 12.4, C(8)H], 2.55 [1 H, ddd, J
9.1, 7.7 and 5.1, C(7b)H], 1.99 (1 H, br s, OH), 1.96 [1 H,
dddd, J 8.9, 7.7, 5.1 and 4.7, C(1a)H], 1.59 [1 H, ddd, J 9.1,
8.9 and 3.5, C(1)Hexo], Ϫ0.21 [1 H, ddd, J 5.1, 5.1 and 3.5,
C(1)Hendo]; δC(100.6 MHz; CDCl3) 8.1 [t, J 165, C(1)], 16.1 [d, J
164, C(1a)], 17.4 [d, J 166, C(7b)], 63.4 [t, J 143, C(8)], 123.6 [d,
J 160, C(3)], 125.4 [d, J 161, C(5)], 127.0 [d, J 156, C(6)], 127.6
[d, J 158, C(4)], 128.7 [d, J 162, C(2)], 130.8 [s, C(7a)], 133.5 [s,
C(3a)], 138.7 [s, C(7)].
Acknowledgements
The authors wish to thank Dr F. J. J. de Kanter for recording
the 400 MHz NMR spectra. This research was supported by the
Netherlands Foundation for Chemical Research (SON) with
financial aid from the Netherlands Organization for Scientific
Research (NWO) (G. W. W.).
References
1 (a) F. Bickelhaupt and W. H. de Wolf, in Advances in Strain
in Organic Chemistry, ed. B. Halton, JAI Press, London, 1993, vol.
3, pp. 185–227 and references cited therein; (b) F. Bickelhaupt and
W. H. de Wolf, Recl. Trav. Chim. Pays-Bas, 1988, 107, 459; (c) G. W.
Wijsman, D. S. van Es, W. H. de Wolf and F. Bickelhaupt, Angew.
Chem., 1993, 105, 739; Angew. Chem., Int. Ed. Engl., 1993, 32, 726.
2 (a) R. E. Beyler and L. H. Sarett, J. Am. Chem. Soc., 1952, 74, 1406;
(b) V. Boekelheide and G. K. Vick, J. Am. Chem. Soc., 1956, 78, 653;
(c) R. H. Mitchell and F. Sondheimer, Tetrahedron, 1968, 24, 1397.
3 (a) M. Pomerantz and G. W. Gruber, J. Am. Chem. Soc., 1967, 89,
6798, 6799; (b) K. A. Burdett, F. L. Shenton, D. H. Yates and
J. S. Swenton, Tetrahedron, 1974, 30, 2057; (c) P. Wan, D. Budac,
M. Earle and D. Shukla, J. Am. Chem. Soc., 1990, 112, 8048.
4 (a) P. Müller, G. Bernardinelli and H. C. Godoy-Nguyen Thi, Helv.
Chim. Acta, 1989, 72, 1627; (b) M. Cristl, R. Stangl and H. Jelinek-
Fink, Chem. Ber., 1992, 125, 485.
[7b-2H]-1a,7b-Dihydro-7-hydroxymethyl-1H-cyclopropa[a]-
naphthalene ([2H]2)
To a suspension of LiAlD4 (7.78 mmol, 0.35 g) in dry THF (20
ml), diol 3 (2.1 mmol, 0.40 g) was added in small portions. The
reaction mixture was stirred for 6 days at reflux temperature.
Then the mixture was cooled to 0 ЊC and the excess hydride was
carefully quenched with water. The mixture was poured into a
cold 5% HCl aqueous solution and extracted with diethyl ether.
The combined organic layers were washed with brine and dried
over anhydrous MgSO4. After filtration and concentration at
reduced pressure, a yellow oil (0.33 g) remained, consisting
mainly of [2H]2 (1 mmol, 48%) and [2H2]9 (0.5 mmol, 24%).
Purification for analytical purposes by preparative GLC
resulted in a mixture of [2H]2 and [2H2]9 (3:1).
5 (a) E. I. Snyder, J. Org. Chem., 1967, 32, 3531; (b) F. A. Hochstein
and W. G. Brown, J. Am. Chem. Soc., 1948, 70, 3484.
6 M. J. Jorgenson and A. W. Friend, J. Am. Chem. Soc., 1965, 87,
1815.
[2H]2. δH (400 MHz; CDCl3) 7.23 [1 H, dd, J 7.5 and 1.0,
C(6)H], 7.14 [1 H, t, J 7.5, C(5)H] 7.06 [1 H, dd, J 7.5 and 1.0,
C(4)H], 6.29 [1 H, dd, J 9.6 and 4.7, C(2)H], 6.25 [1 H, d, J2,3
9.6, C(3)H], 4.87 [1 H, d, JAB 12.4, C(8)H], 4.83 [1 H, d, JAB
12.4, C(8)H], 1.99 (1 H, br s, OH), 1.97 [1 H, ddd, J 8.8, 4.7 and
4.6, C(1a)H], 1.61 [1 H, dd, J 8.8 and 3.5, C(1)Hexo], Ϫ0.21 [1 H,
dd, J 4.6 and 3.5, C(1)Hendo]; δD (61.42 MHz; CHCl3) 2.57;
δC(100.6 MHz; CDCl3) 8.0 [t, J 166, C(1)], 17.3 [d, J 164, C(1a)],
18.8 [t, J 19, C(7b)], 63.5 [t, J 143, C(8)], 123.5 [d, J 160, C(3)],
125.5 [d, J 161, C(5)], 126.9 [d, J 159, C(6)], 127.6 [d, J 158,
C(4)], 128.6 [d, J 161, C(2)], 130.8 [s, C(7a)], 133.5 [s, C(3a)],
138.0 [s, C(7)]; m/z 173 (M ϩ), 155 (M ϩ Ϫ H2O, 100%), 154
7 R. T. Uyeda and D. J. Cram, J. Org. Chem., 1965, 30, 2083.
8 L. H. Klemm, D. Hsu Lee, P. S. Santhanam, D. V. White, J. Shabtai
and D. R. Olson, J. Org. Chem., 1972, 37, 2043.
9 W. Thatcher Borden, J. Am. Chem. Soc., 1970, 92, 4898.
10 R. B. Gammill, P. M. Gold and S. A. Mizsak, J. Am. Chem. Soc.,
1980, 102, 3095.
11 M. J. Jorgenson and A. F. Thacher, J. Chem. Soc., 1969, 21D, 1290.
12 Obtained from Group Increments, S. W. Benson, Thermochemical
Kinetics, Wiley, New York, 1968.
Paper 7/00605E
Received 27th January 1997
Accepted 30th May 1997
᝽
᝽
(95), 142 (58), 129 (41), 116 (30).
2098
J. Chem. Soc., Perkin Trans. 2, 1997