The stereochemical course of the generation and interception of a six-membered cyclic allene: 3δ2-1H-naphthalene (2,3-didehydro-1,2-dihydronaphthalene)

Article abstract of DOI:10.1002/ejoc.200600443

The bromofluorocarbene adduct rac-1 of indene, possessing the fluorine atom at the endo position, is a useful substrate for the generation of the isonaphthalene 4 by the Doering-Moore-Skattebol reaction. By resolution of rac-1, an enantiomerically pure precursor of a six-membered cyclic allene was obtained for the first time. The treatment of (+)- or (-)-1, dissolved in 2,5-dimethyl-, 2-tert-butyl-5-methyl-, or 2,5-bis-(tert-butyl)furan, with methyllithium gave rise to the [4+2] cycloadducts 6 of 4 to the furans. It was shown by means of HPLC on Chiralcel OD that the formation of 6 proceeded with about 40 % ee, and that this value was independent of the type and concentration of the furan, and the reaction temperature. The absolute configurations of the enantiomers 1, as well as those of the enantiomers 6, were determined by simulation of the CD spectra by quantum chemical methods and by the comparison of them with the experimental spectra. In the case of (+)-1, the reliability of this procedure was checked by X-ray crystal structure analysis. On the basis of these results, a model is proposed for the steric course of the reaction sequence, leading from a pure enantiomer 1 to the 70:30 mixtures of the product enantiomers 6. The use of indene as trapping reagent for 4 furnished the [2+2] cycloadduct 15. For the preparation of rac-15, the particularly simple one-pot procedure was employed, in which indene, tetra-bromomethane, and methyllithium were combined and in which the dibromocarbene adduct 2 of indene serves as precursor of 4. Compound 2 could not be isolated, but was characterised by low-temperature ¹H NMR spectra. The conversion of (+)- or (-)-1 into 4 in the presence of indene afforded 15 only with a very low enantioselectivity. The constitutions and the relative configurations of 15 as well as the compounds 8 and 16, which resulted from thermolysis of the cycloadducts 6b and 15, respectively, were elucidated by X-ray crystal diffraction. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejoc.200600443

FULL PAPER
DOI: 10.1002/ejoc.200600443
The Stereochemical Course of the Generation and Interception of a
Six-Membered Cyclic Allene: 3δ2-1H-Naphthalene (2,3-Didehydro-1,2-
dihydronaphthalene)[‡]
Manfred Christl,*[a] Martin Braun,[a] Hartmut Fischer,[a] Stefan Groetsch,[a]
Germar Müller,[a] Dirk Leusser,[b] Stephan Deuerlein,[b] Dietmar Stalke,*[b] Mario Arnone,[a]
and Bernd Engels*[a]
Dedicated to Professor Gerhard Erker on the occasion of his 60th birthday
Keywords: CD spectra / Cycloadditions / Cyclic allenes / Enantioselectivity / Quantum chemistry calculations
The bromofluorocarbene adduct rac-1 of indene, possessing
the fluorine atom at the endo position, is a useful substrate
for the generation of the isonaphthalene 4 by the Doering–
these results, a model is proposed for the steric course of the
reaction sequence, leading from a pure enantiomer 1 to the
70:30 mixtures of the product enantiomers 6. The use of in-
Moore–Skattebøl reaction. By resolution of rac-1, an enantio- dene as trapping reagent for 4 furnished the [2+2] cycload-
merically pure precursor of a six-membered cyclic allene was
obtained for the first time. The treatment of (+)- or (–)-1, dis-
solved in 2,5-dimethyl-, 2-tert-butyl-5-methyl-, or 2,5-bis-
(tert-butyl)furan, with methyllithium gave rise to the [4+2]
cycloadducts 6 of 4 to the furans. It was shown by means of
HPLC on Chiralcel OD that the formation of 6 proceeded
with about 40% ee, and that this value was independent of
the type and concentration of the furan, and the reaction
temperature. The absolute configurations of the enantiomers
1, as well as those of the enantiomers 6, were determined by
simulation of the CD spectra by quantum chemical methods
and by the comparison of them with the experimental spec-
tra. In the case of (+)-1, the reliability of this procedure was
checked by X-ray crystal structure analysis. On the basis of
duct 15. For the preparation of rac-15, the particularly simple
one-pot procedure was employed, in which indene, tetra-
bromomethane, and methyllithium were combined and in
which the dibromocarbene adduct 2 of indene serves as pre-
cursor of 4. Compound 2 could not be isolated, but was char-
acterised by low-temperature 1H NMR spectra. The conver-
sion of (+)- or (–)-1 into 4 in the presence of indene afforded
15 only with a very low enantioselectivity. The constitutions
and the relative configurations of 15 as well as the com-
pounds 8 and 16, which resulted from thermolysis of the cy-
cloadducts 6b and 15, respectively, were elucidated by X-ray
crystal diffraction.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
tum chemical calculations corroborated that the ground
state of 1,2-cyclohexadiene has the allene structure, whereas
Introduction
By elegant experiments, Balci and Jones[1,2] established the corresponding allyl diradical, being 15–18 kcalmol–1
the chirality of the six-membered cyclic allenes[3] 1,2-cyclo- higher in energy, serves as the transition state for the enanti-
hexadiene and bicyclo[3.2.1]octa-2,3,6-triene. Later, quan- omerisation.[4,5] The barrier to enantiomerisation of the ti-
tle allene 4 (Scheme 1) was calculated to be significantly
lower, namely 11 kcalmol–1.[5] If a nitrogen or an oxygen
atom is directly attached to the allene subunit within a six-
membered ring, this barrier is predicted to have a consider-
ably smaller value or may even vanish.[3,5–8] We now report
that, in spite of the rather low barrier to enantiomerisation,
[‡] Cycloallenes, Part 20. Ref.[3] is counted as Part 19. For Part 18,
see ref.[7]
[a] Institut für Organische Chemie, Universität Würzburg,
Am Hubland, 97074 Würzburg, Germany
Fax: +49-931-8884606
.
E-Mail: christl@chemie.uni-wuerzburg.de, engels@chemie.uni- enantioselectively generated 4 can be trapped by activated
wuerzburg.de
olefins before racemisation occurs.
[b] Institut für Anorganische Chemie, Universität Göttingen,
Our approach to the problem consists in the utilisation
of a chiral precursor to the isonaphthalene 4. Previously, 4
was generated from the bromofluorocarbene adduct rac-
1[9,10] of indene, the dibromocarbene adduct 2[10] of indene,
Tammannstraße 4, 37077 Göttingen, Germany
Fax: +49-551-393373
E-Mail: dstalke@chemie.uni-goettingen.de
Supporting information for this article is available on the
WWW under http:/www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5045

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
designed for the observation of 2-fluoronaphthalene, but
previous papers describe its preparation from indene and
halofluorocarbenes.[12]
Compared to 5, the greater stability of rac-1 is explained
by the stereoelectronic requirements of the ring expan-
sion,[13] which demand the ejection of a fluoride ion. How-
ever, the strength of the C–F bond, whose bond dissoci-
ation energy is ca. 40 kcalmol–1[14] larger than that of a C–
Br bond, inhibits this step at 20 °C and permits it only at
130 °C, at which temperature 2-bromonaphthalene emerges
from rac-1 (Scheme 2).
The attainment of the pure enantiomers from rac-1 was
achieved by HPLC on Chiralcel OD. An X-ray crystal
structure analysis [Flack parameter = –0.003(8)] established
the 1S,1aR,6bR-configuration of (+)-1 (Figure 1).
Scheme 1. Precursors to the isonaphthalene 4.
and
from
3-bromo-1,2-dihydronaphthalene
(3)[11]
(Scheme 1). As 3 is achiral and 2 cannot be isolated (see
section 4), only rac-1 suited our goal.
Results and Discussion
1. Preparation of the Precursor rac-1 of the Cyclic Allene 4,
Resolution of rac-1, Generation of 4 from the Pure
Enantiomers (+)-1 and (–)-1 and Trapping of 4 by Furans
The preparation of rac-1 is described here in detail. Pre-
viously, it has only briefly been mentioned in a conference
report[9] and a communication.[10] Exposure of indene to a
mixture of dibromofluoromethane and sodium hydroxide in
dichloromethane containing benzyltriethylammonium chlo-
ride gave rise to a 15% yield of rac-1 (Scheme 2). The rather
modest yield is a consequence of the simultaneous forma-
tion of the diastereomers 5 of rac-1, which cannot be iso-
lated due to their rapid conversion into 2-fluoronaphtha-
lene by ejection of the bromide ion with concomitant ring
expansion and final loss of a proton. Our workup was not
Figure 1. Molecular structure of (+)-1. The anisotropic displace-
ment parameters are depicted at the 50% probability level.
In order to trap the isonaphthalene 4, generated from a
pure enantiomer 1, we chose 2,5-dimethylfuran as the first
reagent. This diene furnished mainly one cycloadduct,
namely rac-6a in 8% yield, and its regioisomer in less than
1% yield, when 4 had been liberated from 3-bromo-1,2-di-
hydronaphthalene (3).[11] The other activated olefins used
to trap 4, except styrene and α-methylstyrene, gave rise to
products consisting of several components in similar quan-
tities.[3] Indeed, the reaction of rac-1, dissolved in pure 2,5-
dimethylfuran, with methyllithium at –30 °C brought about
a 40% yield of rac-6a. Whether or not rac-6a was ac-
companied by a small amount of its regioisomer, was not
investigated.
The utilisation of the pure enantiomer (–)-1 did not lead
to the formation of rac-6a, but to 6a and ent-6a in the ratio
of 73:27 (Scheme 3). Due to the rather ready rearrangement
of rac-6a to its regioisomer,[11] we did not carry out the
analysis of the enantiomeric ratio directly, but after the de-
hydrogenation of 6a/ent-6a by DDQ yielding the ep-
oxyphenanthrenes 7/ent-7 (Scheme 4), the mixture of which
was subjected to HPLC on Chiralcel OD. To establish
which signal belonged to which enantiomer, that is, to de-
termine the absolute configuration, we compared the exper-
imental CD spectra with the calculated ones. The procedure
is described in section 2. The same treatment of (+)-1 as
above gave rise to 6a and ent-6a in the ratio of 28:72.
The fact that some of the stereochemical information of
a pure enantiomer 1 is transferred to the cycloadducts 6a/
ent-6a, proves the chirality of the intermediate 4. Simulta-
Scheme 2. The addition of bromofluorocarbene to indene and the
conversion of the adducts rac-1 and -5 into 2-bromo- and 2-fluoro-
naphthalene, respectively.
5046
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058

Generation and Interception of a Six-Membered Cyclic Allene
FULL PAPER
Under these conditions, the intermediate 4 should have a
longer lifetime and hence a higher probability to undergo
enantiomerisation than in the above experiments. However,
the same enantiomeric ratios were observed, that is, 6a/ent-
6a = 72.5:27.5 from (–)-1 in 2,5-dimethylfuran/diethyl ether,
3:1, and 27:73 from (+)-1 in 2,5-dimethylfuran/diethyl ether,
1:1. These results exclude the conversion of (P)-4 into (M)-
4 and vice versa under the reaction conditions and suggest
an activation barrier to the cycloaddition of 4 that lies well
below 11 kcalmol–1.
A further experiment was conducted at room tempera-
ture, that is about 50 °C higher than in the above cases.
Again, the ratio of 6a and ent-6a determined after the reac-
tion of (–)-1 was 72:28. Like the results obtained with dif-
ferent concentrations of 2,5-dimethylfuran, this finding sug-
gests that the allene 4 has a stable configuration in trapping
experiments, which is inconsistent with the rationalisation
of the observations by Balci and Jones in the cases of 1,2-
cyclohexadiene and bicyclo[3.2.1]octa-2,3,6-triene.[1,2] Gen-
erating these allenes as nonracemic mixtures, the authors
observed optically active cycloadducts of 1,3-diphenylisob-
enzofuran, if the reactions were carried out at 53 °C, but
racemic products at 100 °C. They suggested that the racem-
isation of the allenes is faster than their interception at
100 °C, which does not apply at 53 °C. In view of the above
independence of the 6a/ent-6a ratio on the temperature, this
explanation is likely to miss the point, as 1,2-cyclohexa-
diene has a higher barrier to enantiomerisation than the
Scheme 3. Generation of the isonaphthalene 4 from the pure en-
antiomer (1R,1aS,6bS)-1 [(–)-1] and interception of 4 by 2,5-disub-
stituted furans.
isonaphthalene
4
(15–18 kcalmol–1 for 1,2-cyclohexa-
diene[4,5] vs. 11 kcalmol–1 for 4[5]). Bicyclo[3.2.1]octa-2,3,6-
triene might have a barrier similar to that of 4, because of
the enhanced strain energy due to the rigidity of the bicyclic
skeleton. Most probably, the effect noted by Balci and Jones
reflects the temperature dependence of kH/kD in the elimi-
nation step of nonracemic 1-bromo-6-deuteriocyclohex-
ene[1] and the enantioselectivity in the elimination of hydro-
gen bromide from 1-bromocyclohexene or 3-bromobicy-
clo[3.2.1]octa-2,6-diene by enantiomerically pure potassium
menthoxide.[2]
Next, we turned to tests to see if the cycloaddition step
of 4 is responsible for the partial loss of the stereochemical
information introduced with (–)- or (+)-1 and, thus, in-
creased the steric demand of the trapping reagent. Our first
choice was 2-tert-butyl-5-methylfuran. When utilised as the
solvent for the reaction of rac-1 with methyllithium, this
compound gave rise to a 25% yield of rac-6b, rac-8, and 9
Scheme 4. Further products or consecutive products of the inter-
ception of 4 by 2,5-disubstituted furans.
neously, the following question arises: at which stage is in the ratio of 5:1:2. A competition experiment showed that
about 30% of this stereochemical information lost? Does 2,5-dimethyl- and 2-tert-butyl-5-methylfuran trapped the
that leakage occur during the generation of 4 from a pure intermediate 4 equally fast.
enantiomer 1 or by partial equilibration of (M)- and (P)-4
With regard to the analysis of the 6b/ent-6b ratio with
via the corresponding diradical as transition state[5] or by a the utilisation of a pure enantiomer 1, the procedure as in
low selectivity in the cycloaddition of an enantiomer 4 onto the case of 6a/ent-6a was to no avail, as the treatment of
2,5-dimethylfuran?
rac-6b, employed as a mixture with rac-8 and 9, with DDQ
We first addressed this problem by dilution of 2,5-di- did not afford the racemic epoxyphenanthrene analogous
methylfuran. In the above experiments, the enantiomer 1 to rac-7. We then tried to simplify the 5:1:2 mixture of rac-
was dissolved in the pure trapping reagent for 4 and treated 6b, rac-8, and 9 by thermolysis. Akin to the rearrangement
with methyllithium, but now mixtures of 2,5-dimethylfuran of rac-6a to the respective tetrahydroepoxyanthracene,[11]
and diethyl ether in the ratios 3:1 and 1:1 were employed. rac-6b was converted into rac-8 with 68% yield on refluxing
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5047

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
in toluene, whereas 9 remained largely unchanged. When the precursor to all three products. This diradical is believed
the product was subjected to HPLC on Chiralcel OD, a to be first formed by attack of 4 with its central allene car-
baseline separation of the peaks of 8 and ent-8 was bon atom at C-2 of the furan, and collapses to rac-6 and
achieved.
rac-10 under kinetic control at low temperature. It would
The reaction of pure (+)-1 with methyllithium in the be regenerated on heating of rac-6 and rac-10 and finally
presence of 2-tert-butyl-5-methylfuran and the subsequent form a four-membered ring en route to 11. In view of the
thermolysis of the product furnished 8 and ent-8 in the ratio rather modest yield and the insufficient purity of 11, we
of 34:66, which is also valid for 6b/ent-6b. The fractions of decided to try the resolution directly with the 5:1 mixture
pure 8 and ent-8 were collected, and, after evaporation of of rac-6c and rac-10 by HPLC on Chiralcel OD. Two sig-
most of the solvent, crystals grew in one of the samples nals of equal intensity were observed. According to the CD
on storage at –30 °C. An X-ray crystal structure analysis spectra, these originated from the minor components (10
established the constitution and the relative configuration and ent-10), although the ratio of 6c or ent-6c and 10 or
of 8/ent-8 (Figure 2), but not the absolute configuration, ent-10 was still about 5:1 for the material collected for each
1
which was determined by the comparison of the experimen- signal, as shown by the H NMR spectra. It is detailed in
tal and calculated CD spectra (see section 2).
the Exp. Sect. that these signals can also be used to quantify
the relative amounts of 6c and ent-6c. Thereupon, pure (+)-
1, dissolved in 2,5-di-tert-butylfuran, was subjected to
methyllithium, and the isolated product was analysed to
contain 6c and ent-6c in the ratio of 26:74.
In spite of the substantial difference in their steric de-
mand, the allenophiles 2,5-dimethyl-, 2-tert-butyl-5-methyl-
and 2,5-di-tert-butylfuran furnished enantiomeric mixtures
of 6 and ent-6 with the very similar ratios of 28:72, 34:66,
and 26:74, respectively, on use of (+)-1 as the substrate. This
is why we conclude that the addition of an enantiomer 4 to
each of the furans is highly stereoselective. Taking into ac-
count the above finding that the interconversion of the
enantiomers 4 does not take place under the reaction condi-
tions, we are left with the insight that the partial loss of the
stereochemical information introduced with a pure enanti-
omer 1 already occurs during the generation of 4 from 1.
The small variation of the above ratios may be caused by
the error limits of the analysis and/or by solvent effects of
Figure 2. Molecular structure of 8/ent-8. The anisotropic displace-
ment parameters are depicted at the 50% probability level.
The loss of the stereochemical information introduced by the different furans on the first step of the reaction se-
pure enantiomer 1 was not very different for 2,5-dimethyl- quences.
and 2-tert-butyl-5-methylfuran as reaction partners of 4.
Previously, it has been shown that the liberation of 1-
Therefore, we increased the steric demand of the trapping phenyl-1,2-cyclooctadiene from the enantiopure dibro-
reagent further by utilising 2,5-di-tert-butylfuran. This het- mocarbene adducts of 1-phenylcycloheptene proceeds with
erocycle being the solvent in the treatment of rac-1 with a high stereoselectivity.[16] In contrast, the Doering–Moore–
methyllithium gave rise to a 5:1 mixture of rac-6c Skattebøl reaction of an enantiomer 1 in the above furans
(Scheme 3) and rac-10 (Scheme 4). The similarity of their brings about both enantiomers 4 in a ratio of about 70:30.
NMR spectroscopic data with those of rac-6a,b and rac-8, Most likely, this divergence has its origin in the substitution
respectively, leaves no doubt as to the structure of rac-6c pattern of the bridgehead positions of the cycloallene pre-
and rac-10. A yield of only 10% already indicated a rela- cursors. In the case of 1, these positions (C-1a, C-6b) carry
tively low reactivity of 2,5-di-tert-butylfuran toward the cy- a hydrogen atom each, and with regards to the activation
cloallene 4. Indeed, a competition experiment revealed that barriers, it does not make a big difference whether 1a-H or
the relative rate constant is only one tenth of that of 2,5- 6b-H of (–)-1 moves to the other side of the ring system en
dimethylfuran. In order to convert rac-6c into rac-10 and route to (P)- or (M)-4, respectively (Scheme 3). However,
hence to arrive at a uniform product, whose analysis of the dibromocarbene adduct of 1-phenylcycloheptene has a
enantiomers would be simpler, we thermolysed the 5:1 mix- hydrogen atom and a phenyl group at the bridgehead posi-
ture in refluxing benzene. Surprisingly, no rac-10 at all was tions, and because of its steric demand the phenyl group
found in the product but only the formal [2+2] cycloadduct might be strongly hindered in changing the ring side by the
11 of 4 onto 2,5-di-tert-butylfuran in 23% yield. Appar- adjacent methylene group, whereas the hydrogen atom at
ently, 11 is thermodynamically more stable than the formal the other bridgehead could smoothly undergo this move-
[4+2] cycloadducts rac-6c and rac-10. As discussed pre- ment.
viously, in the cases of the furan adducts of 4[11] and 3δ2-
For a more detailed consideration of the stereochemical
chromene (2,3-didehydro-2H-1-benzopyran),[15] a common course of the pathway from (+)- or (–)-1 to 6 and ent-6,
racemic diradical (see 13, Scheme 5 in section 3) should be knowledge of the absolute configurations of the products is
5048
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058

Generation and Interception of a Six-Membered Cyclic Allene
FULL PAPER
necessary, which has been presented in advance in Scheme 3 and its time-dependent (TD) variant by applying the TUR-
and the corresponding discussion above. The elucidation of BOMOLE program package.[19] The geometries of the in-
these absolute configurations is the subject of section 2.
vestigated compounds were optimised at the BLYP/SVP
level[20–22] by applying the resolution of identity approxi-
mation[23,24] together with the matching auxiliary basis
sets.[24,25] The electronic excitations were calculated at the
TD-B3LYP[26–28] level of theory in combination with the
TZVP basis sets.[29] For the molecules (–)-1 and 6c, the first
20 excitations were calculated, and for 7, 8, and 15, the first
40 excitations. The UV spectra were simulated by superim-
posing Gauss functions, weighted according to the oscil-
lator strength,[18] and then compared with the experimental
spectra. With respect to the position of the strongest ab-
sorption, the calculated spectra of (–)-1, 7, and 15 showed
a hypsochromic shift of 15, 14, and 13 nm, respectively. In
2. Determination of Absolute Configurations by Calculation
of the CD Spectra and Comparison with the Experimental
Spectra
Circular dichroism (CD) is a chiroptical phenomenon,
observed in nontransparent regions of the spectrum of non-
racemic chiral compounds.[17] By resolving rac-1, rac-7, rac-
8, and rac-15 (see section 4), we obtained sufficient amounts
of the enantiomers to measure CD spectra in the UV re-
gion. The fact that in all cases the spectrum of one enanti-
omer was almost the exact mirror image of that of the other
is an indication of the purity of the substances, the struc-
tural identity of which was checked by NMR spectroscopy.
For the elucidation of the absolute configuration of a
chiral compound, the simulation of its CD spectrum by
quantum chemical methods and the comparison with the
experimental spectrum is an established method.[17,18] The
present calculations utilised the density functional theory
Figure 4. Comparison of the experimental and calculated (B3LYP)
CD spectra of 7.
Figure 3. Comparison of the experimental and calculated (B3LYP)
CD spectra of (–)-1 (top) and (+)-1 (bottom).
Figure 5. Comparison of the experimental and calculated (B3LYP)
CD spectra of ent-8.
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5049

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
the case of 8, the absorption maximum at lowest energy why (P)-4 and (M)-4 add to a furan to give exclusively 6
was used for comparison and found to be bathochromically and ent-6, respectively, and not vice versa. Quantum chemi-
shifted by 27 nm. The CD spectra were simulated by super- cal calculations[30] as well as the experimental result of the
imposing Gauss functions, weighted according to the ro- addition of (Z,Z)-1,4-dideuterio-1,3-butadiene to 1,2-cyclo-
tator strength,[18] and then shifted by the number of wave- hexadiene[3] support a two-step mechanism via diradical in-
length units that corresponded to the deviation of the ex- termediates not only for [2+2] but also for [4+2] cycload-
perimental and calculated UV spectra, given above. In Fig- ditions of allenes. In addition, the interception of rac-4 by
ures 3, 4, 5, and 6 the simulated CD spectra of (–)- and 2-tert-butyl-5-methylfuran to give rac-6b, rac-8, and 9 in the
(+)-1, 7, ent-8, and ent-15 are compared with the matching ratio of 5:1:2 is inconsistent with a concerted process. Such
experimental spectra, whereby the absolute configurations a mechanism would demand bond making between the cen-
of the resolved samples were determined.
tral allene carbon atom of 4, which is the sterically least
encumbered one, and the tert-butyl-substituted carbon
atom of the furan, with formation of 9 as the major prod-
uct. However, in the case of a diradical intermediate (see
13, Scheme 5), this should preferentially emerge from the
attack of the central allene carbon atom of 4 at the methyl-
bearing carbon atom of the furan, which is the easier ac-
cessible of the two reaction centres.
Scheme 5. Possible transition states (12, 14) for the addition of the
central allene carbon atom of (P)- and (M)-4 to one of the enantio-
topic faces of the CR1 group of a furan of Scheme 3, with forma-
tion of the diradical 13, being the intermediate en route to 6. The
enantiomers (P)-4 (12) and (M)-4 (14) are represented by the
CH=C=CH subunits only and approach the furan from above the
drawing plane.
Figure 6. Comparison of the experimental and calculated (B3LYP)
CD spectra of ent-15.
3. Proposal for the Stereochemical Course of the
Generation of the Isonaphthalene 4 from 1 and Its
Cycloaddition to Furans
Scheme 5 shows the diastereomeric transition states 12
Notwithstanding the knowledge of the absolute configu- and 14 for the formation of the diradicals 13 by addition of
ration of the precursor 1 for the intermediate 4 and the the central allene carbon atom of (P)-4 (12) and (M)-4 (14)
products resulting from 4, the stereochemical courses of the to one of the enantiotopic faces of the CR1 group of the
formation of 4 and its interception remain a matter of furans (Si-face for R1 = Me, Re-face for R1 = tBu). We
speculation, as to date 4 is not amenable to direct observa- consider 12 for the most favourable arrangement leading to
tion. The first question refers to the conversion of 1 into 4: 13 and eventually to 6. The attack of (M)-4 on the same
Which enantiomer of 4 preferentially emerges from (–)-1, face (14) would also give rise to 13, but 14 is believed to
(P)-4 or (M)-4? In other words, which of the two hydrogen suffer from overcrowding more than 12. In both 12 and 14,
atoms, 1a-H or 6b-H, of (–)-1 changes sides of the ring sys- one hydrogen atom of the allene subunit interacts with R1
tem more readily than the other? As illustrated in Scheme 3, about the same extent. However, the second hydrogen atom
we prefer 1a-H, because it has to only move past a hydrogen approaches the oxygen atom of the furan in 12 and the CH
atom of the methylene group (C-2). On the other hand, 6b- group of the 3-position in 14. As the spatial demand of an
H faces considerably more steric hindrance, as the adjacent oxygen atom is substantially smaller than that of a CH
carbon atom (C-6a) does not carry a hydrogen atom but a group, 12 should be preferred over 14. Because our experi-
CH group. Consequently, (P)-4 should be the major (ca. ments suggest that (P)-4 and (M)-4 react most selectively
70%) and (M)-4 the minor species (ca. 30%) formed on to furnish 6 and ent-6, respectively (see Scheme 3), 14 seems
treatment of (–)-1, dissolved in furans, with methyllithium. to have no importance at all, and (M)-4 adds to the other
Given this assignment and the experimentally based con- face of the CR1 group of a furan via a transition state that
clusion (see section 1) that (P)- and (M)-4 do not intercon- is the mirror image of 12 (= ent-12) and leads to ent-13 and
vert under the reaction conditions, it has to be rationalised eventually ent-6.
5050
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058

Generation and Interception of a Six-Membered Cyclic Allene
FULL PAPER
4. Trapping of Isonaphthalene 4 by Indene
illustrated the decomposition of 2, which most probably
gave rise to 2-bromonaphthalene, although we did not
prove that.
The above furans undergo a [4+2] cycloaddition with
isonaphthalene 4, but [2+2] cycloadditions of 4 are not less
common,[3,10,11] which is why we tested the stereochemical
course of such a reaction as well. We tried out indene as a
trapping reagent for 4, which has been utilised to intercept
only two six-membered cyclic allenes heretofore.[31] Indeed,
the pentacyclic hydrocarbon rac-15 (Scheme 6) was formed
smoothly as the sole product.
In order to synthesise rac-15, we treated 2, which was
obtained as detailed above at –60 °C, with methyllithium in
the presence of indene at –30 °C and isolated the product
in 30% yield. The configuration as drawn in Scheme 6 was
determined by X-ray crystal structure analysis (Figure 7).
As expected on the basis of experience with the styrene ad-
duct of 4,[10] the thermal stability of rac-15 turned out to
be limited. Heating at 80 °C gave rise to the isomer rac-16,
the configuration of which was also established by X-ray
crystal structure analysis (Figure 8). The difference in the
thermodynamic stability of rac-15 and rac-16 is probably
caused by the position of the ethylene subunit, which is iso-
lated in the former but conjugated with a benzene group in
the latter. In analogy with the rearrangements discussed
above (6bǞ8 and 6c,10Ǟ11), the pathway from 15 to 16
should involve the diradical 19 (see Scheme 7).
Figure 7. Molecular structure of 15/ent-15 as determined by X-ray
crystal diffraction. The anisotropic displacement parameters are
depicted at the 50% probability level.
Scheme 6. Preparation of the dibromocarbene adduct 2 of indene,
its use as precursor of the isonaphthalene 4 within the one-pot
procedure for the synthesis of the indene adduct rac-15 of 4, and
the thermal rearrangement of rac-15 to rac-16.
Being particularly simple, the one-pot procedure was em-
ployed for the preparation of rac-15. This method was brief-
ly mentioned previously,[3,10] and takes advantage of the di-
bromocarbene adduct 2 of indene. An earlier attempt to
obtain 2 under the routine conditions for the addition of
dibromocarbene onto olefins (HCBr3, KOtBu) resulted in
the formation of 2-bromonaphthalene.[32] The reasons for
the instability of 2 are the same as those suggested for the
conversion of 5 into 2-fluoronaphthalene (Scheme 2). In
contrast to that of 5, the existence of 2 could be proved
directly by NMR spectroscopy. With this end in view, we
treated a mixture of indene and tetrabromomethane with
methyllithium[33] at –60 °C, then removed most of the sol-
Figure 8. Molecular structure of 16/ent-16 as determined by X-ray
crystal diffraction. The anisotropic displacement parameters are
depicted at the 50% probability level.
Although we did not provide a large excess of indene for
vent (diethyl ether) under reduced pressure at –60 °C, and the interception of 4, the 30% yield of rac-15 indicated a
recorded a 1H NMR spectrum at –60 °C, which clearly indi- relatively high reactivity of this allenophile towards 4. By
cated the presence of 2 (Scheme 6) by comparison with the means of a competition experiment, indene and 2,5-dimeth-
1
1H NMR spectroscopic data of 1. H NMR spectra taken ylfuran were shown to react equally fast.
at –30 °C and 0 °C were unchanged relative to that obtained
By HPLC on Chiralcel OD, rac-15 could be resolved.
at –60 °C. However, the NMR sample darkened quickly at The absolute configuration was again established by com-
temperatures above 0 °C, and a spectrum recorded at 25 °C parison of the experimental and the calculated CD spectra
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5051

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
The other possible origin for the low stereoselectivity re-
sides in the similar size of the 1-CH2 group and the 3-CH
group of indene. Therefore, the transition states 17 and 18
(Scheme 7) for the attack of both enantiomers 4 to the Si-
face of indene could be very close in energy. The result
would be the diradical 19 and eventually 15, whereas their
enantiomers ent-19 and ent-15 would emerge from the ad-
dition of the enantiomers 4 to the Re-face of indene.
Conclusions
Firstly, the results of this work prove that the six-mem-
bered cyclic allene 4 is a chiral molecule. In addition, the
experiments and the calculated CD spectra of several com-
pounds enabled us to propose a model for the detailed ste-
reochemical course of the generation of the enantiomers 4
from a pure enantiomer 1 in a Doering–Moore–Skattebøl
reaction and the interception of 4 by 2,5-disubstituted fu-
rans. Accordingly, the partial loss of the stereochemical in-
formation introduced by a pure enantiomer 1 occurs in the
first step, which provides a 70:30 mixture of (P)- and (M)-
4 from (–)-1. The interconversion of (P)- and (M)-4, pre-
viously calculated to have a barrier of only 11 kcalmol–1,[5]
does not take place under the reaction conditions, because
trapping by the furans proceeds much faster. As the product
ratios 6:ent-6 are virtually independent of the spatial de-
mand of the furans, we believe that the addition of an en-
antiomer 4 is highly stereoselective. More speculations are
necessary to explain the facts in the case of indene as an
important component of the solvent and as a reaction part-
ner of 4. The very low stereoselectivity, i.e. the almost equal
amounts of 15 and ent-15, has its origin either in similar
reaction rates for the formation of both enantiomers 4 from
one enantiomer 1 due to a different solvent effect, as com-
pared to the furans as solvent, or in the insignificant prefer-
ence of one enantiomer 4 for one of the enantiotopic faces
of indene.
Scheme 7. Possible transition states (17, 18) for the addition of the
central allene carbon atom of (P)-4 (17) and (M)-4 (18) to the Si-
face of indene en route to the diradical 19, which is the precursor
to 15. The enantiomers 4 are represented by the CH=C=CH sub-
units only and approach indene from above the drawing plane.
(Figure 6), according to which the faster eluting compound
is the enantiomer ent-15. When the pure enantiomers 1 were
exposed to methyllithium in the presence of indene, the ster-
eoselectivity was found to be lower than in the case of the
above furans. The results are collated in Table 1.
Table 1. Ratios of enantiomers 15 and ent-15 on trapping of 4 by
indene, when 4 was generated from pure (+)- or (–)-1 with meth-
yllithium at –30 °C.
Substrate
Solvent
Ratio of 15:ent-15
(+)-1
(–)-1
(–)-1
indene/diethyl ether, 1:1
indene/diethyl ether, 2:1
indene/THF, 2:2.8
52:48
48:52
42:58
Two reasons can be suggested to explain the lower stere-
oselectivity compared to that of the reactions of the furans
above. One is the change of the solvent polarity and the
other the small difference in the steric interaction of one
enantiomer 4 with the enantiotopic faces of indene. Because
of the solidification of indene at –2 °C, it could not be uti-
lised in pure form as a reaction medium, but had to be
diluted. Diethyl ether served this purpose, as some of it was
anyway introduced as the solvent of methyllithium. As the
first two entries of Table 1 show, the variation of the pro-
portion of diethyl ether had no effect. Presumably, the mix-
tures of indene and diethyl ether employed were less polar
than the pure furans in the above experiments. This might
have influenced the conversion of an enantiomer 1 into (P)-
and (M)-4 in such a way so that almost equal amounts
emerged. It is conceivable that this reaction is subject to a
solvent effect, as the intermediate carbenoid, resulting from
1 and transforming to the carbene being the precursor to
(P)- and (M)-4, is a highly polar molecule. Consistent with
this explanation, the replacement of diethyl ether by THF,
To test these hypotheses and, if applicable, develop them
further, the enantioselective generation of a 1-substituted
1,2-cyclohexadiene will be helpful, which is why we are cur-
rently investigating 1-phenyl-1,2-cyclohexadiene.
Experimental Section
General Remarks: NMR: Bruker Avance 400 and DMX 600 spec-
trometers; chemical shifts δ in ppm relative to Me4Si (= 0 ppm) by
using solvent signals as internal reference [CDCl3: δ = 7.26 ppm
(1H NMR) and 77.0 ppm (13C NMR); C6D6: δ = 7.16 ppm (1H
NMR of C6D5H) and 128.0 ppm (13C NMR)]. MS: Finnigan MAT
8200 and MAT 90. Elemental analyses: Carlo Erba Strumenta-
tione Elemental Analyser 1106. Melting points: Kofler hot stage
apparatus from C. Reichert, Optische Werke AG, Vienna, Austria.
UV: JASCO V-570 UV/Vis/NIR Spectrophotometer. CD: JASCO
which in general solvates organolithium compounds better J-715 Spectropolarimeter. Specific rotations: JASCO P-1020 Polari-
meter. Frequently used solvents: DEE = diethyl ether, PE = light
petroleum ether (b.p. 40–65 °C), MTBE = tert-butyl methyl ether.
than diethyl ether does, resulted in an increase in the
enantioselectivity (third entry in Table 1), and this may be
traced back to an increased ratio of the enantiomers 4 on
conversion of (–)-1.
Resolution of Racemates and Analyses of Nonracemic Mixtures of
Enantiomers: An HPLC system was used consisting of a Knauer
5052
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058

Generation and Interception of a Six-Membered Cyclic Allene
FULL PAPER
HPLC Pump 64, a Gynkotek UV-Detektor UVD, operated at bined organic phases were dried with MgSO4 and concentrated in
260 nm in most cases, and a Chiralcel OD column (250×21 mm), vacuo. The residue (1.2 g of a brown oil) was purified by flash
protected by a guard column (50×21 mm) with the same stationary
phase. The flow rate of the eluants was maintained at 12 mL min–1.
Before use as eluants, the solvents were distilled from a Vigreux
column (20 cm). During the chromatography, a constant stream of
helium was bubbled through the solvents. The quantitative analysis
of the UV signals was achieved with a Shimadzu C-R3A Chro-
matopac Integrator.
chromatography (SiO2; PE/MTBE, 20:1) to give rac-6a (390 mg,
40%) as a yellow oil, the spectroscopic data of which agreed with
those reported previously.[11]
1,4-Dihydro-1,4-dimethyl-1,4-epoxyphenanthrene (rac-7): A mixture
of rac-6a (390 mg, 1.74 mmol) and 2,3-dichloro-5,6-dicyano-p-
benzoquinone (500 mg, 2.20 mmol) in benzene (20 mL) was stirred
at room temperature for 12 h. Then, pentane (20 mL) was added,
the resulting precipitate was filtered off, the filtrate was concen-
trated in vacuo, and the residue, an orange oil, was purified by
flash chromatography (SiO2; pentane/DEE, 15:1) to give rac-7
(242 mg, 63%) as a yellow oil, whose 1H NMR spectrum agreed
with the literature data.[35] UV (methanol): λmax (logε) = 336 (3.21),
308 (3.43), 264 (sh, 3.60), 254 (sh, 3.92), 246 (sh, 4.05), 229 (4.56)
nm.
(1α,1aα,6bα)-1-Bromo-1-fluoro-1,1a,2,6b-tetrahydrocyclopropa[a]in-
dene (rac-1): A solution of indene (23.2 g, 200 mmol), benzyltrieth-
ylammonium chloride (800 mg, 3.51 mmol), and dibromofluo-
romethane[34] (38.2 g, 199 mmol) in dichloromethane (100 mL) was
cooled in an ice bath. After the addition of aqueous sodium hy-
droxide (32 g, 800 mmol in 32 mL of water), which was precooled
to 5 °C, the mixture was intensively stirred for 3 d. During that
period, it reached room temperature after several hours. Water
(50 mL) was admixed, and the layers were separated. The aqueous
phase was extracted with DEE (3×30 mL), the combined organic
layers were dried with MgSO4, and the solvents were evaporated
in vacuo. Unreacted indene and 2-fluoronaphthalene, presumably
formed via 5, were removed by distillation at 65 °C (kugelrohr)/
0.05 mbar. The remaining dark brown oil (9.6 g) was purified by
flash chromatography (SiO2; PE/MTBE, 40:1) to give rac-1 (6.90 g,
15%) as a slightly brown solid, m.p. 71–72 °C (from methanol),
which sublimed slowly at 20 °C/1 bar. 1H NMR (CDCl3): see ref.[10]
1H NMR (400 MHz, C6D6): δ = 2.01 (tt, average of J1a,2α and J1a,6b
= 7.5, average of J1a,2β and J1a,F = 1.5 Hz, 1 H, 1a-H), 2.63 (br.
Resolution of rac-7: The equipment described above was used, and
the elution was carried out with hexane/2-propanol, 99:1. Samples
(100 µL) of a 0.2  solution of rac-7 in 2-propanol were injected.
The retention times were 13.5 min for the (–)-enantiomer (7) and
17.5 min for the (+)-enantiomer (ent-7). Specific rotation (meth-
anol, c = 0.1, T = 21 °C): [α]D = +143 and –138. CD (methanol):
see Figure 4.
Reaction of the Pure Enantiomers (+)- and (–)-1 with Methyllithium
in the Presence of 2,5-Dimethylfuran and Subsequent Aromatisation
of the Product: The reaction of (+)-1 (100 mg, 0.440 mmol) with
methyllithium (2.0 mmol, 2.0 mL of 1.0  in DEE) in the presence
of 2,5-dimethylfuran (4 mL) was conducted as described above for
rac-1. Without purification, the crude product (160 mg) was dis-
solved in benzene (6 mL) and then dehydrogenated with 2,3-
dichloro-5,6-dicyano-p-benzoquinone (140 mg, 0.616 mmol) as de-
scribed above for rac-6a. The product was purified by chromatog-
raphy (SiO2; pentane/DEE, 10:1) to give a yellow oil (26 mg, 27%),
which was shown to consist of 7 and ent-7 in the ratio of 28:72. The
dd, J2α,2β = 17.6, J1a,2α = 7.3 Hz, 1 H, 2-Hα), 2.84 (br. d, J2α,2β
=
17.6 Hz, 1 H, 2-Hβ), 2.85 (dd, J1a,6b = 7.7, J6b,F = 1.4 Hz, 1 H, 6b-
H), 6.82 (m, 1 H, 3-H or 6-H), 6.91–6.99 (m, 2 H, 4-H, 5-H), 7.01
(m, 1 H, 6-H or 3-H) ppm. 13C NMR: see ref.[10] UV (methanol):
λmax (logε) = 276 (3.22), 270 (3.09), 262 (sh, 2.94), 230 (3.74), 215
(sh, 3.92) nm. MS (70 eV, EI): m/z (%) = 228, 226 (0.3%, 0.3%)
[M]+, 148 (11), 147 (100), 146 (44), 127 (18). C10H8BrF (227.1):
calcd. C 52.89, H 3.55; found: C 52.73, H 3.66.
1
structure of the product was confirmed by H NMR spectroscopy,
whereas the enantiomeric ratio was determined by HPLC analysis
by applying the conditions as for the resolution of rac-7.
Thermolysis of rac-1: Without a solvent, rac-1 (100 mg, 0.44 mmol)
was heated at 130 °C for 3.5 h. On cooling to room temperature,
the product became a brown solid, which was recrystallised from
ethanol to give 86 mg (94%) of 2-bromonaphthalene. Its identity
When (–)-1 was used instead of (+)-1, the ratio of 7 and ent-7
turned out to be 73:27.
1
was established by comparison of its H NMR spectrum with that
In two further experiments, the enantiomers of 1 were not dissolved
in pure 2,5-dimethylfuran prior to treatment with methyllithium,
but in mixtures of 2,5-dimethylfuran with DEE. These mixtures
were prepared by combining 3 and 2 mL of 2,5-dimethylfuran with
1 and 2 mL, respectively, of DEE to give a total volume of about
4 mL in each case. However, the ratios of 7 and ent-7 were virtually
the same as above, namely 72.5:27.5 from (–)-1 and 27:73 from (+)-
1.
of an authentic sample.
Resolution of rac-1 and Determination of the Absolute Configuration
of the Enantiomers: The equipment described above was used, and
the elution was carried out with hexane/2-propanol, 199:1. Samples
(100 µL) of a saturated solution of rac-1 in 2-propanol were in-
jected. The retention times were 9.5 min for (+)-1 and 12 min for
(–)-1. Per injection, 4.3 mg of each enantiomer were obtained. Spe-
cific rotation (methanol, c = 0.2, T = 20.8 °C): [α]D = +62.1 and
–61.5. CD (methanol): see Figure 3. The pure enantiomers are
colourless solids with m.p. 70–71 °C. A single-crystal X-ray crystal
structure determination established the 1S,1aR,6bR-configuration
of (+)-1 (Figure 1).
In addition, one experiment was conducted at room temperature,
whereas the ones above proceeded at –30 °C. Again, the ratio of 7
and ent-7 turned out to be 72:28 from (–)-1.
(1α,4α,4aα)-4-tert-Butyl-1,4,4a,9-tetrahydro-1-methyl-1,4-epoxy-
phenanthrene (rac-6b), (1α,4α,4aα)-4-tert-Butyl-1,4,4a,10-tetra-
hydro-1-methyl-1,4-epoxyanthracene (rac-8), and (1α,4α,4aα)-1-tert-
Butyl-1,4,4a,9-tetrahydro-4-methyl-1,4-epoxyphenanthrene (9): Un-
der nitrogen, a stirred solution of rac-1 (800 mg, 3.52 mmol) in 2-
tert-butyl-5-methylfuran[36] (5 mL), cooled to –30 °C, was treated
dropwise with methyllithium (5.50 mmol, 5 mL of 1.1  in DEE)
in a manner so that the temperature remained at –30 °C. After
removal of the cooling bath, the temperature was allowed to rise
(1α,4α,4aα)-1,4,4a,9-Tetrahydro-1,4-dimethyl-1,4-epoxyphen-
anthrene (rac-6a): Under nitrogen, a stirred solution of rac-1
(1.00 g, 4.40 mmol) in 2,5-dimethylfuran (20 mL) was cooled to
–30 °C and was treated dropwise with methyllithium (16.5 mmol,
15 mL of 1.1  in DEE) in a manner so that the temperature re-
mained at –30 °C (30 min). After removal of the cooling bath, the
temperature was allowed to rise to 0 °C, and the mixture was then
cautiously hydrolysed (20 mL). The layers were separated, the to 0 °C, and the mixture was then cautiously hydrolysed (6 mL).
aqueous layer was extracted with DEE (3×10 mL), and the com- The layers were separated, the aqueous layer was extracted with
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5053

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
DEE (3×5 mL), and the combined organic phases were dried with
MgSO4 and concentrated in vacuo (15 mbar). The residue still con-
tained 2-tert-butyl-5-methylfuran, which was distilled off in a kug-
elrohr (40 °C/0.05 mbar) and thus recovered. The remaining prod-
uct, 859 mg of an orange oil, was shown by NMR spectroscopy to
contain rac-6b, rac-8, and 9 in the ratio of 5:1:2. Flash chromatog-
raphy (SiO2; PE/MTBE, 30:1) furnished an 8:1:2 mixture of rac-6,
rac-8, and 9 as a light yellow oil (238 mg, 25%). MS (70 eV, EI):
m/z (%) = 266 (8) [M]+, 210 (13), 209 (20), 182 (15), 181 (100), 167
(18), 166 (30), 165 (32), 148 (28), 147 (38), 146 (20), 142 (11), 141
(13), 128 (20), 127 (18), 115 (25), 57 (21), 43 (23), 41 (10). HRMS
(70 eV, EI): calcd. for C19H22O [M]+ 266.1665; found 266.1666.
tance 1.1 Hz, 1 H, 7-H) ppm; the assignments are based on a
NOESY spectrum. 13C NMR (151 MHz, C6D6): δ = 14.8 (1-Me),
26.6 (tBu-Me), 33.2 (tBu-Cq), 35.7 (C-10), 43.4 (C-4a), 86.0 (C-1),
97.6 (C-4), 114.7 (C-9), 126.5 (C-8), 126.9 (C-6), 127.2 (C-7), 128.4
(C-5), 134.8 (C-10a), 135.2 (C-3), 135.6 (C-8a), 142.1 (C-2), 154.4
(C-9a) ppm; the assignments are based on HMQC and HMBC
spectra.
Resolution of rac-8: The equipment described above was used, and
the elution was carried out with hexane/2-propanol, 1000:1. Sam-
ples (100 µL) of an 0.1  solution of rac-8 in 2-propanol were in-
jected. The retention times were 12.0 min for the (+)-enantiomer 8
and 13.5 min for the (–)-enantiomer (ent-8). Specific rotation
(methanol, c = 0.5, T = 21 °C): [α]D = +157 and –196. CD (meth-
anol): see Figure 5.
1
rac-6b: H NMR (600 MHz, CDCl3): δ = 1.32 (s, 9 H, tBu), 1.68
(s, 3 H, 1-Me), 3.31–3.37 (m, 3 H, 9-H2, 4a-H), 5.79 (m, 1 H, 10-
H), 5.91 (d, J2,3 = 5.6 Hz, 1 H, 2-H), 6.51 (d, J2,3 = 5.6 Hz, 1 H,
3-H), 7.11–7.23 (m, 3 H, 6-H, 7-H, 8-H), 7.36 (dm, J5,6 Ϸ 7 Hz, 1
H, 5-H) ppm; the position of the substituents and the configuration
is supported by a NOESY spectrum. 13C NMR (151 MHz,
CDCl3): δ = 15.6 (1-Me), 28.0 (tBu-Me), 33.2 (C-9), 33.5 (tBu-Cq),
49.2 (C-4a), 83.9 (C-1), 97.4 (C-4), 112.0 (C-10), 124.6 (C-5), 124.9,
125.0, 126.6 (C-6, C-7, C-8), 134.6 (C-3), 136.7 (C-2), 140.8, 140.9
(C-4b, C-8a), 150.4 (C-10a) ppm; as far as specified, the assign-
ments are based on HMQC and HMBC spectra.
Reaction of the Pure Enantiomer (+)-1 with Methyllithium in the
Presence of 2-tert-Butyl-5-methylfuran and Subsequent Thermal Re-
arrangement of 6b/ent-6b to 8/ent-8: Under nitrogen, a stirred solu-
tion of (+)-1 (130 mg, 0.572 mmol) in 2-tert-butyl-5-methylfuran
(2 g), cooled to –30 °C, was treated dropwise with methyllithium
(4.0 mmol, 5.0 mL of 0.8  in DEE) in a manner so that the tem-
perature remained at –30 °C. After removal of the cooling bath,
the temperature was allowed to rise to 0 °C, and the mixture was
then cautiously hydrolysed (5 mL). The layers were separated, the
aqueous layer was extracted with DEE (3×5 mL), and the com-
bined organic phases were dried with MgSO4 and concentrated in
vacuo. From the residue, the excess of 2-tert-butyl-5-methylfuran
was distilled off in a kugelrohr (50 °C/15 mbar). The remaining oil
(76 mg) was dissolved in toluene (10 mL), and the solution was
refluxed for 4 h. The cold solution was concentrated in vacuo, and
the residue was purified by flash chromatography (SiO2; PE/ethyl
acetate, 100:1) to furnish 8.4 mg (6%) of a colourless oil, which
was shown to consist of 8 and ent-8 in the ratio of 34:66 by HPLC
on Chiralcel OD under the conditions described for the resolution
of rac-8. The mixture was separated completely, and the solutions
of 8 and ent-8, as obtained from HPLC, were concentrated in vacuo
to a volume of 1 mL each. In one of the samples, crystals (m.p. 91–
93 °C) formed on storage at –30 °C, which were subjected to X-
ray crystal structure determination furnishing constitution and
relative configuration, but not the absolute configuration of 8/ent-
8 (Figure 2).
rac-8: See next experiment.
9: 1H NMR (600 MHz, C6D6; values in brackets refer to CDCl3 as
solvent): δ = 1.24 [1.21] (s, 9 H, tBu), 1.83 [2.03] (s, 3 H, 4-Me),
3.02 [3.18] (dd, J9α,9β = 17.6, J9β,10 = 6.6 Hz, 1 H, 9-Hβ), 3.08 [3.14]
(dm, J4a,9α = 6.2 Hz, 1 H, 4a-H), 3.14 [3.32] (br. d, J9α,9β = 17.6,
J4a,9α = 6.2 Hz, 1 H, 9-Hα), 5.74 [5.96] (dt, J9β,10 = 6.6, J4a,10
J9α,10 = 2.3 Hz, 1 H, 10-H), 5.93 [6.16] (d, J2,3 = 5.6 Hz, 1 H, 3-
H), 5.99 [6.26] (br. d, J2,3 = 5.6 Hz, 1 H, 2-H), [7.28] (br. d, J5,6
=
Ϸ
7.7 Hz, 1 H, 5-H) ppm; the signals of the other aromatic protons
are superimposed by signals of rac-6b and rac-8. 13C NMR
(151 MHz, CDCl3): δ = 20.3 (4-Me), 26.5 (tBu-Me), 32.5 (tBu-Cq),
32.6 (C-9), 52.7 (C-4a), 86.1 (C-4), 95.5 (C-1), 114.0 (C-10), 122.9
(C-5), 125.6, 125.7, 126.8 (C-6, C-7, C-8), 135.5 (C-3), 136.1 (C-2),
138.3, 139.1, 145.8 (C-4b, C-8a, C-10a) ppm.
(1α,4α,4aα)-4-tert-Butyl-1,4,4a,10-tetrahydro-1-methyl-1,4-epoxy-
anthracene (rac-8) by Thermal Rearrangement of rac-6b: A solution
of an 8:1:2 mixture of rac-6b, rac-8, and 9 (200 mg) in toluene
(20 mL) was refluxed for 3 h. After evaporation of the solvent in
vacuo (0.1 mbar), an NMR spectrum of the residue (193 mg) indi-
cated that rac-6b had been converted completely into rac-8,
whereas 9 had remained largely unchanged. Flash chromatography
(SiO2; PE/ethyl acetate, 33:1) furnished a 10:1 mixture of rac-8 and
9 (136 mg, 68%) as a colourless oil. UV (methanol): λmax (logε) =
297 (sh, 3.67), 278 (3.95), 227 (sh, 4.19), 220 (4.32), 215 (4.32), 211
(sh, 4.29) nm. MS (70 eV, EI): m/z (%) = 266 (17) [M]+, 223 (20),
209 (19), 194 (10), 182 (17), 181 (100), 167 (34), 166 (24), 165 (27),
57 (24), 43 (17). HRMS (70 eV, EI): calcd. for C19H22O [M]+
266.1665; found 266.1666.
rac-8: 1H NMR (600 MHz, C6D6; the values in brackets refer to
CDCl3 as solvent): δ = 1.10 [1.19] (s, 9 H, tBu), 1.52 [1.66] (s, 3 H,
1-Me), 2.33 [2.43] (br. dd, J4a,10β = 15.9, J10α,10β = 14.5 Hz, 1 H,
10-Hβ), 2.75 [2.94] (dd, J10α,10β = 14.5, J4a,10α = 5.8 Hz, 1 H, 10-
Hα), 3.21 [3.20] (ddd, J4a,10β = 15.9, J4a,10α = 5.8, J4a,9 = 2.4 Hz, 1
H, 9-H), 5.90 [6.21] (d, J2,3 = 5.6 Hz, 1 H, 3-H), 6.05 [6.19] (d, J4a,9
= 2.4 Hz, 1 H, 9-H), 6.21 [6.47] (dd, J2,3 = 5.6, J = 0.5 Hz, 1 H, 2-
H), 6.89 [7.02] (dtq, J5,6 = 7.3, further line distances 1.4, 0.7 Hz, 1
H, 5-H), 6.96 [7.02] (br. dd, J7,8 = 7.3, J6,8 = 1.4 Hz, 1 H, 8-H),
Relative Reactivity of 2,5-Dimethylfuran and 2-tert-Butyl-5-meth-
ylfuran Towards the Intermediate 4: Under nitrogen, a stirred solu-
tion of rac-1 (50 mg, 0.22 mmol) in a mixture of 2,5-dimethylfuran
(700 mg, 7.29 mmol) and 2-tert-butyl-5-methylfuran (1.00 g,
7.25 mmol) was treated with methyllithium (1.60 mmol, 2.0 mL of
0.8  in DEE). The conditions and the workup were as described
above for the analogous reaction of rac-1 in one of the two furans
as solvent. After the removal of the unreacted 2-tert-butyl-5-meth-
ylfuran by distillation in a kugelrohr (30 °C/0.02 mbar), the crude
product (61 mg) was shown to contain rac-6a, rac-6b, rac-8, and 9
in the ratio of about 8:5:1:2 by NMR spectroscopy, indicating that
2,5-dimethylfuran and 2-tert-butyl-5-methylfuran trap 4 equally
fast.
(1α,4α,4aα)-1,4-Bis(tert-butyl)-1,4,4a,9-tetrahydro-1,4-epoxyphen-
anthrene (rac-6c) and (1α,4α,4aα)-1,4-Bis(tert-butyl)-1,4,4a,10-
tetrahydro-1,4-epoxyanthracene (rac-10): Under nitrogen, a stirred
solution of rac-1 (500 mg, 2.20 mmol) in 2,5-di-tert-butylfuran[37]
(8 mL), cooled to –20 °C, was treated dropwise with methyllithium
(10 mmol, 10 mL of 1.0  in DEE) in a manner so that the tem-
perature did not exceed –20 °C. After removal of the cooling bath,
7.00 [7.07] (td, average of J5,6 and J6,7 = 7.4, J6,8 = 1.4 Hz, 1 H, 6- the temperature was allowed to rise to 0 °C, and the mixture was
H), 7.07 [7.13] (tt, average of J6,7 and J7,8 = 7.4, further line dis- then cautiously hydrolysed (15 mL). The layers were separated, the
5054
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058

Generation and Interception of a Six-Membered Cyclic Allene
FULL PAPER
aqueous layer was extracted with DEE (3×10 mL), and the com-
bined organic phases were dried with MgSO4 and concentrated in
vacuo. The unreacted 2,5-di-tert-butylfuran was distilled off usinf
a kugelrohr (40 °C/0.05 mbar) and thus recovered. The residue
(390 mg) was purified by flash chromatography (SiO2; PE/ethyl ace-
tate, 70:1) to give rac-6c and rac-10 (69 mg, 10%) in the ratio of
5:1 as a colourless oil. MS (70 eV, EI): m/z (%) = 308 (4) [M]+, 252
(36), 251 (28), 237 (16), 235 (10), 195 (19), 181 (11), 167 (27), 165
(12), 143 (11), 142 (13), 141 (10), 128 (14), 84 (29), 57 (100), 41
(17). HRMS (70 eV, EI): calcd. for C22H28O [M]+ 308.2135; found
308.2137. The UV spectrum virtually coincided with that of rac-8
when superimposed, indicating that rac-6c has much lower molar
extinction coefficients than rac-10. This should be particularly valid
for the long-wave part of the spectrum, as rac-6c has only an iso-
lated benzene chromophore, whereas rac-10 is a styrene derivative.
Indeed, the calculation of the UV spectrum of 6c with the methods
described in section 2 gave a result that significantly deviated from
the experimental spectrum of the 5:1 mixture of rac-6c and rac-10.
signal with a retention time of 8.3 min was assigned to ent-10 and
that at 9.3 min to 10. Whether 6c or ent-6c had the same retention
time as 10 or ent-10 could not be decided.
Reaction of the Pure Enantiomer (+)-1 with Methyllithium in the
Presence of 2,5-Bis(tert-butyl)furan: Under nitrogen, a stirred solu-
tion of (+)-1 (300 mg, 1.32 mmol) in 2,5-di-tert-butylfuran (6 mL),
cooled to –25 °C, was treated dropwise with methyllithium
(4.5 mmol, 5.0 mL of 0.9  in DEE) in a manner so that the tem-
perature remained at –25 °C. After removal of the cooling bath,
the temperature was allowed to rise to 0 °C, and the mixture was
then cautiously hydrolysed (5 mL). The layers were separated, the
aqueous layer was extracted with DEE (3×5 mL), and the com-
bined organic phases were dried with MgSO4 and concentrated in
vacuo (15 mbar). From the remaining oil, the excess of 2,5-di-tert-
butylfuran was distilled off in a kugelrohr (40 °C/0.05 mbar). The
residue was purified by flash chromatography (SiO2; PE/ethyl ace-
tate, 80:1) to furnish a mixture of 6c, ent-6c, 10, and ent-10 (24 mg,
6%) as a colourless oil. The ratio of 10:ent-10 was determined to
be 26:74 according to the procedure described for the resolution of
rac-6c/rac-10. This ratio is also valid for 6c and ent-6c, because 6c
and 10 emerge from a common diradical precursor (see 13 in
Scheme 5) and ent-6c and ent-10 from its enantiomer (ent-13).
1
rac-6c: H NMR (600 MHz, C6D6): δ = 1.20 (s, 9 H, 1-tBu), 1.25
(s, 9 H, 4-tBu), 3.05 (br. dd, J9α,9β = 16.8, J9β,10 = 6.4 Hz, 1 H, 9-
Hβ), 3.10 (ddd, J9α,9β = 16.8, J4a,9α = 4.7, J9α,10 = 2.3 Hz, 1 H, 9-
Hα), 3.30 (m, 1 H, 4a-H), 5.77 (dt, J9β,10 = 6.4, J4a,10 = J9α,10
=
2.3 Hz, 1 H, 10-H), 5.83 (dd, J2,3 = 5.7, J = 0.6 Hz, 1 H, 2-H), 6.29
(d, J2,3 = 5.7 Hz, 1 H, 3-H), 7.03–7.07 (m, 2 H, 7-H, 8-H), 7.10 (m,
1 H, 6-H), 7.32 (br. d, J5,6 = 7.6 Hz, 1 H, 5-H) ppm; the assign-
ments are based on COSY and NOESY experiments. 13C NMR
(151 MHz, C6D6): δ = 26.5 (1-tBu-Me), 28.0 (4-tBu-Me), 32.8 (1-
tBu-Cq), 33.7 (C-9), 33.9 (4-tBu-Cq), 51.3 (C-4a), 93.5 (C-1), 96.0
(C-4), 114.3 (C-10), 125.0 (C-5), 125.1 (double intensity, C-6, C-7),
126.5 (C-8), 133.5 (C-2), 135.1 (C-3), 140.8 (C-8a), 141.5 (C-4b),
147.7 (C-10a) ppm; the assignments are based on HSQC and
HMBC spectra.
Relative Reactivity of 2,5-Dimethyl- and 2,5-Bis(tert-butyl)furan
Towards the Isonaphthalene 4: Under nitrogen, a stirred solution of
rac-1 (50 mg, 0.22 mmol) in a mixture of 2,5-dimethylfuran (1.35 g,
14.1 mmol) and 2,5-di-tert-butylfuran (2.53 g, 14.1 mmol) was
treated with methyllithium (1.60 mmol, 2.0 mL of 0.8  in DEE).
The conditions and the workup were as described above for the
analogous reactions of rac-1 in one of the two furans as solvent.
After removal of the unreacted 2,5-di-tert-butylfuran by distillation
in a kugelrohr (30 °C/0.02 mbar), the crude product (48 mg) was
shown to contain rac-6a, rac-6c, and 10 in a ratio of about 60:5:1
by NMR spectroscopy, indicating that 2,5-dimethylfuran traps 4
ten times as fast as 2,5-di-tert-butylfuran.
1
rac-10: H NMR (600 MHz, C6D6): δ = 1.07 (s, 9 H, 4-tBu), 1.21
(s, 9 H, 1-tBu), 2.32 (br. dd, J10α,10β = 14.4, J4a,10β = 15.8 Hz, 1 H,
10-Hβ), 2.73 (dd, J10α,10β = 14.4, J4a,10α = 5.7 Hz, 1 H, 10-Hα), 3.14
(ddd, J4a,10β = 15.8, J4a,10α = 5.7, J4a,9 = 2.4 Hz, 1 H, 4a-H), 5.91
(d, J2,3 = 5.7 Hz, 1 H, 3-H), 6.41 (br. d, J4a,9 = 2.4 Hz, 1 H, 9-H),
6.49 (dd, J2,3 = 5.7, J = 0.6 Hz, 1 H, 2-H), 6.88 (br. d, J5,6 = 7.3 Hz,
1 H, 5-H), 6.91 (br. dd, J7,8 = 7.4, J6,8 = 1.4 Hz, 1 H, 8-H), 7.00
(td, J5,6 = J6,7 = 7.4 Hz, J6,8 = 1.4 Hz, 1 H, 6-H), ca. 7.50 (superim-
posed by signals of rac-6c, 7-H) ppm; the assignments are based
on COSY and NOESY experiments. 13C NMR (151 MHz, C6D6):
δ = 26.6 (4-tBu-Me), 26.7 (1-tBu-Me), 32.8 (1-tBu-Cq), 33.3 (4-tBu-
Cq), 35.4 (C-10), 44.9 (C-4a), 95.4 (C-1), 96.2 (C-4), 117.1 (C-9),
126.6 (C-8), 126.9 (C-6), 127.1 (C-7), 128.1 (C-5), 134.4 (C-10a),
135.5 (C-3), 135.8 (C-8a), 138.8 (C-2), 152.3 (C-9a) ppm; the as-
signments are based on HSQC and HMBC spectra.
2,9b-Di-tert-butyl-3a,3b,4,9b-tetrahydronaphtho[2Ј,3Ј:3,4]cyclobuta-
[1,2-b]furan (11): A solution of rac-6c and rac-10 in the ratio of 5:1
(69 mg) in benzene (5 mL) was refluxed for 6 h. After evaporation
of the solvent in vacuo, the residue was purified by flash
chromatography (SiO2; PE/ethyl acetate, 114:1) to give a colourless
oil (16 mg, 23%), which was shown by NMR spectroscopy to con-
sist mainly of 11. 1H NMR (400 MHz, C6D6): δ = 1.06 (s, 9 H, 9b-
tBu), 1.12 (s, 9 H, 2-tBu), 2.31 (br. dd, J4,4 = 14.6, J3b,4 = 7.7 Hz,
1 H) and 3.11 (tt, average of J3b,4 and J4,4 = 14.5, J = 1.0 Hz, 1 H)
(4-H2), 2.83 (dtd, J3b,4 = 14.4, average of J3b,4 and J3a,3b = 7.7, J3b,9
= 2.7 Hz, 1 H, 3b-H), 3.31 (dd, J3a,3b = 7.8, J3,3a = 2.8 Hz, 1 H,
3a-H), 4.50 (d, J3,3a = 2.8 Hz, 1 H, 3-H), 6.49 (br. d, J3b,9 = 2.7 Hz,
1 H, 9-H), 6.94 (m, 1 H), 6.95–7.05 (m, 3 H) ppm; as far as speci-
fied, the assignment is based on a HMBC spectrum. 13C NMR
(101 MHz, C6D6): δ = 24.5 (9b-tBu-Me), 28.1 (2-tBu-Me), 28.3 (C-
4), 32.6 (2-tBu-Cq), 34.2 (9b-tBu-Cq), 40.2 (C-3b), 46.3 (C-3a), 90.9
(C-3), 99.7 (C-9b), 120.6 (C-9), 126.75, 126.84, 126.85, 128.7 (C-5,
C-6, C-7, C-8), 135.4, 136.3 (C-4a, C-8a), 147.6 (C-9a), 169.3 (C-
2) ppm; as far as specified, the assignment is based on HMBC and
HMQC spectra.
Resolution of the Mixture of rac-6c and rac-10: The equipment de-
scribed above was used, and the elution was carried out with hex-
ane/2-propanol, 1000:1. As the 5:1 mixture of rac-6c and rac-10
could not be separated by flash chromatography, samples (100 µL)
of its 0.05  solution in hexane were injected. Two signals of equal
intensity were observed. The corresponding substances were col-
lected separately and characterised by 1H NMR spectroscopy
showing that they still contained 6c/ent-6c and 10/ent-10 in about
the ratio of the racemates injected. However, the CD spectra were
mirror images to each other and virtually coincided with those of
8 and ent-8, respectively, when superimposed. As our HPLC system
operates with a UV detector, it shows the strongest response to
compounds with a high molar absorptivity. Although 6c and ent-
6c were in excess, the signals came from 10 and ent-10. Owing to
the similarity of the CD spectra with those of 8 and ent-8, the
(1aα,6bα)-1,1-Dibromo-1,1a,2,6b-tetrahydrocyclopropa[a]indene (2):
Under nitrogen, a stirred solution of indene (500 mg, 4.30 mmol)
and tetrabromomethane (1.57 g, 4.73 mmol) in anhydrous DEE
(7 mL), cooled to –60 °C, was treated dropwise with methyllithium
(4.80 mmol, 3.29 mL of 1.46  in DEE) within 30 min. The re-
sulting yellow solution was stirred for further 15 min at –60 °C and
then concentrated in vacuo (0.01 mbar) at that temperature. To the
residue, whose main component was DEE, CDCl3 was added at
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5055

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
–60 °C, and 1H NMR spectra were recorded at –60, –30, 0, and
+25 °C, which indicated the presence of 2 in the first three cases.
Up to 0 °C, the sample was yellow; it darkened rapidly above 0 °C.
1H NMR at 0 °C: see ref.[10]
indene was distilled off in a kugelrohr (40 °C/0.04 mbar). The resi-
due was purified by flash chromatography (SiO2, PE/ethyl acetate,
51:1) to give a colourless oil (15 mg, 28%), which was shown to be
a 52:48 mixture of 15 and ent-15 by the procedure described for
the resolution of rac-15.
(5aα,11bβ,11cα)-5a,7,11b,11c-Tetrahydro-5H-indeno[2Ј,1Ј:3,4]cy-
clobuta[1,2-a]naphthalene (rac-15): Under nitrogen, a stirred solu-
tion of indene (4.00 g, 34.4 mmol) and tetrabromomethane (12.0 g,
36.2 mmol) in anhydrous DEE (20 mL), cooled to –60 °C, was
treated dropwise with methyllithium (36 mmol, 36 mL of 1.0  in
DEE). The mixture was then allowed to warm to –30 °C within
30 min and treated dropwise with methyllithium (36 mmol, 36 mL
of 1  in DEE) at that temperature. After removal of the cooling
bath, the temperature was allowed to rise to 0 °C, and the mixture
was then cautiously hydrolysed (30 mL). The layers were separated,
the aqueous layer was extracted with DEE (3×30 mL), and the
combined organic phases were dried with MgSO4 and concentrated
in vacuo (15 mbar). From the remaining oil, the unreacted indene
was distilled off in a kugelrohr (40 °C/0.02 mbar). The residue was
purified by flash chromatography (SiO2; pentane/DEE, 20:1) to
give rac-15 (1.26 g) as a light yellow oil. The yield was 30 and 14%
with reference to indene and tetrabromomethane, respectively.
When 2 equiv. of indene and 1 equiv. of tetrabromomethane were
utilised, the yield was about 30 % with reference to tetra-
bromomethane. MS (70 eV, EI): m/z (%) = 244 (34) [M]+, 239 (11),
229 (37), 228 (29), 226 (11), 195 (10), 165 (10), 129 (22), 128 (100),
127 (14), 116 (91), 115 (49), 91 (13), 45 (23), 44 (12), 43 (13).
HRMS (70 eV, EI): calcd. for C19H16 [M]+ 244.1247; found
244.1248. UV (methanol): λmax (logε) = 277 (3.40), 270 (3.39), 263
(sh, 3.29), 257 (sh, 3.21), 226 (sh, 4.14), 216 (sh, 4.33) nm. 1H NMR
When (–)-1 (175 mg, 0.77 mmol), dissolved in a mixture of indene
(3.00 g, 25.8 mmol) and DEE (1.5 mL), was treated with meth-
yllithium as described above for (+)-1, the workup furnished 15
and ent-15 (50 mg, 27%) in the ratio of 48:52.
When (–)-1 (155 mg, 0.68 mmol), dissolved in a mixture of indene
(2.00 g, 17.2 mmol) and anhydrous THF (2.8 mL), was treated with
methyllithium as described above for (+)-1, the workup furnished
15 and ent-15 (10 mg, 6%) in the ratio of 42:58. On storage of this
mixture at 4 °C over several months, colourless needles formed,
m.p. 82–84 °C, which were used for the X-ray crystal structure
analysis (Figure 7).
Relative Reactivity of Indene and 2,5-Dimethylfuran Towards the
Isonaphthalene 4: Under nitrogen, a stirred solution of rac-1
(100 mg, 0.44 mmol) in a mixture of indene (1.21 g, 10.4 mmol)
and 2,5-dimethylfuran (1.00 g, 10.4 mmol) was treated with meth-
yllithium (1.60 mmol, 2.0 mL of 0.8  in DEE). The conditions and
the workup were as described for the analogous reaction of (+)-
and (–)-1 in indene/DEE. The crude product was a yellow oil
(96 mg) and contained the cycloadducts 6a and 15 in the ratio of
1:1. Thus, indene and 2,5-dimethylfuran trap 4 equally fast.
(5aα,11aβ,11bα)-5a,11,11a,11b-Tetrahydro-5H-indeno[2Ј,1Ј:3,4]cy-
clobuta[1,2-b]naphthalene (16): A solution of rac-15 (676 mg,
2.77 mmol) in benzene (30 mL) was refluxed for 11 h. The solvent
of the brown solution was then evaporated in vacuo, and the resi-
due was purified by flash chromatography (SiO2; PE/ethyl acetate,
41:1) to give rac-16 (86 mg, 13 %) as colourless solid, m.p. 80–
82 °C. Crystals for the X-ray crystal structure analysis (Figure 8)
were obtained by dissolution of the solid in hexane (1.5 mL) and
storage of the solution at –30 °C. MS (70 eV, EI): m/z (%) = 244
(39) [M]+, 239 (14), 229 (43), 228 (32), 226 (12), 129 (20), 128 (88),
117 (10), 116 (100), 115 (41), 114 (16), 113 (10), 101 (10), 91 (12).
HRMS (70 eV, EI): calcd. for C19H16 [M]+ 244.1247; found
(600 MHz, C6D6): δ = 3.02 (ddd, J7,7 = 18.3, J6,7 = 6.5, J7,11b
=
2.2 Hz, 1 H) and 3.08 (ddm, J7,7 = 18.3, J7,11b = 7.4 Hz, 1 H) (7-
H2), 3.11 (dd, J5,5 = 16.9, J5,5a = 10.3 Hz, 1 H) and 3.17 (dd, J5,5
= 16.9, J5,5a = 5.1 Hz, 1 H) (5-H2), 3.48 (m, 1 H, 5a-H), 3.68 (m,
1 H, 11b-H), 3.79 (t-like, average of J5a,11c and J11b,11c = 6.2 Hz, 1
H, 11c-H), 5.59 (dddd, J6,7 = 6.5, J5a,6 and J6,7 and J6,11b = 3.2,
2.3, 1.1 Hz, 1 H, 6-H), 7.03 (br. d, J8,9 = 7.5 Hz, 1 H, 8-H), 7.09–
7.16 (m, 4 H, 2-H, 3-H, 4-H, 9-H), 7.20 (br. t, J9,10 = J10,11
=
7.4 Hz, 1 H, 10-H), 7.23 (br. d, J1,2 = 6.9 Hz, 1 H, 1-H), 7.33 (br.
d, J10,11 = 7.4 Hz, 1 H, 11-H) ppm; as far as specified, the assign-
ments are based on a H,H-COSY spectrum. 13C NMR (151 MHz,
C6D6): δ = 31.1 (C-7), 37.6 (C-5), 45.5 (C-5a), 53.3 (C-11b), 53.8
(C-11c), 115.1 (C-6), 123.8 (C-1), 125.5, 126.2, 127.0, 127.1 (C-2,
C-3, C-4, C-9), 125.9 (C-11), 126.3 (C-10), 128.4 (C-8), 137.2,
139.6, 145.5, 145.8, 146.9 (C-4a, C-5b, C-7a, C-11a, C-11d) ppm;
as far as specified, the assignments are based on a HSQC spectrum.
1
244.1248. H NMR (600 MHz, C6D6): δ = 2.55–2.75 (m, 3 H, 11-
H2, 11a-H), 3.07 (br. dd, J5,5 = 17.0, J5β,5a = 5.2 Hz, 1 H, 5-Hβ),
3.11 (dd, J5,5 = 17.0, J5α,5a = 10.2 Hz, 1 H, 5-Hα), 3.30 (m, 1 H,
11b-H), 3.60 (m, 1 H, 5a-H), 6.16 (br. s, 1 H, 6-H), 6.93–7.02 (m,
3 H), 7.04–7.12 (m, 5 H) ppm; as far as specified, the assignments
are based on a H,H COSY spectrum. 13C NMR (151 MHz, C6D6):
δ = 34.3 (C-11), 37.9 (C-5), 45.5 (C-5a), 46.0 (C-11a), 51.1 (C-11b),
116.9 (C-6), 123.8 (C-1), 125.3 (C-4), 126.2 (C-7), 126.3, 126.8,
126.9, 127.0 (C-2, C-3, C-8, C-9), 128.4 (C-10), 135.3, 135.8 (C-6a,
C-10a), 145.4, 146.7 (C-4a, C-11c), 147.1 (C-5b) ppm; as far as
specified, the assignments are based on HSQC and HMBC spectra.
Resolution of rac-15: The equipment described above was used, and
the elution was carried out with hexane/2-propanol, 834:1. Samples
(100 µL) of a 0.1  solution of rac-15 in hexane were injected. The
retention times were 15.4 min for the (–)-enantiomer (ent-15) and
17.8 min for the (+)-enantiomer 15. Specific rotation (hexane, c =
0.2, T = 21 °C): [α]D = –40 and +37. CD (methanol): see Figure 6.
X-ray Crystal Structure Determinations of (+)-1, 8/ent-8, 15/ent-15,
and 16/ent-16: The data for (+)-1, 8/ent-8, and 16/ent-16 were col-
lected from shock-cooled crystals on a Bruker Smart APEX dif-
Reactions of the Pure Enantiomers (+)-1 and (–)-1 with Methylli-
thium in the Presence of Indene: Under nitrogen, a stirred solution of fractometer with a D8 goniometer (graphite-monochromated Mo-
(+)-1 (50 mg, 0.22 mmol) and indene (1.00 g, 8.61 mmol) in anhy-
drous DEE (1.0 mL), cooled to –30 °C, was treated dropwise with
methyllithium (2.4 mmol, 3.0 mL of 0.8  in DEE) in a manner so
that the temperature remained at –30 °C (20 min). After removal
of the cooling bath, the temperature was allowed to rise to 0 °C,
and the mixture was then hydrolysed (3 mL). The layers were sepa-
rated, the aqueous layer was extracted with DEE (3×5 mL), and
the combined organic phases were dried with MgSO4 and concen-
trated in vacuo (15 mbar). From the remaining oil, the excess of
Kα radiation, λ = 71.073 pm) equipped with a low-temperature de-
vice operating at 193(2) K.[38] The data for 15/ent-15 were collected
from shock-cooled crystals on a STOE IPDS II diffractometer
(graphite-monochromated Mo-Kα radiation, λ = 71.073 pm)
equipped with a low-temperature device operating at 133(2) K.[38]
An empirical absorption correction with the program SADABS
2.10 was employed for all structures.[39] They were solved by direct
methods (SHELXS-97[40]) and refined by full-matrix least-squares
methods against F2 (SHELXL-97[41]). R values: R1 = Σ||Fo|–|Fc||/
5056
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058

Generation and Interception of a Six-Membered Cyclic Allene
FULL PAPER
Σ|Fo|, wR2 = [Σw(Fo2 – Fc2)2/Σw(Fo2)2]0.5, w = [σ2(Fo2) + (g1P)2 + for assistance in recording the CD spectra and the specific rotations
g2P]–1, P = 1/3[max(Fo2,0)+2Fc ].
and to Prof. Christoph Lambert and his group for recording the
UV spectra.
2
(+)-1: C10H8BrF, M = 227.08, orthorhombic, space group P212121,
Z = 4, a = 505.80(4) pm, b = 983.77(8) pm, c = 1692.01(14) pm,
V = 0.84193(12) nm3, ρc = 1.791 Mg·m–3, 7902 reflections mea-
sured, 1726 unique, R1 [IϾ2σ(I)] = 0.0149, wR2 (all data) = 0.0382,
g1 = 0.0258, g2 = 0.1754 for 141 parameters and 0 restraints, GooF
= 1.058, residual density (max./min.) = 0.248/–0.257 e·Å–3, Flack
parameter = –0.003(8), absolute configuration determined.
[1] M. Balci, W. M. Jones, J. Am. Chem. Soc. 1980, 102, 7607–
7608.
[2] M. Balci, W. M. Jones, J. Am. Chem. Soc. 1981, 103, 2874–
2876.
[3] For a comprehensive review on the chemistry of six-membered
cyclic allenes see: M. Christl, in: Modern Allene Chemistry
(Eds.: N. Krause, A. S. K. Hashmi), Wiley-VCH, Weinheim,
2004, pp. 243–357.
[4] R. O. Angus Jr., M. W. Schmidt, R. P. Johnson, J. Am. Chem.
Soc. 1985, 107, 532–537.
[5] B. Engels, J. C. Schöneboom, A. F. Münster, S. Groetsch, M.
Christl, J. Am. Chem. Soc. 2002, 124, 287–297.
[6] C. J. Emanuel, P. B. Shevlin, J. Am. Chem. Soc. 1994, 116,
5991–5992.
8/ent-8: C19H22O, M = 266.38, monoclinic, space group P21, Z =
2, a = 739.79(7) pm, b = 940.54(8) pm, c = 1090.15(10) pm, β
= 98.5930(10)°, V = 0.75001(12) nm3, ρc = 1.179 Mg·m–3, 16114
reflections measured, 3070 unique, R1 [IϾ2σ(I)] = 0.0329, wR2 (all
data) = 0.0863, g1 = 0.0545, g2 = 0.0934 for 269 parameters and 1
restraint, GooF = 1.072, residual density (max./min.) = 0.230/
–0.176 e·Å–3, Flack parameter = 0.3(11), absolute configuration
could not be determined.
[7] J. C. Schöneboom, S. Groetsch, M. Christl, B. Engels, Chem.
Eur. J. 2003, 9, 4641–4649.
15/ent-15: C19H16, M = 244.34, monoclinic, space group P21/n, Z
= 4, a = 1102.3(2) pm, b = 530.89(11) pm, c = 2227.4(5) pm, β =
97.05(3)°, V = 1.2936(5) nm3, ρc = 1.254 Mg·m–3, 7461 reflections
measured, 2452 unique, R1 [IϾ2σ(I)] = 0.0338, wR2 (all data) =
0.0867, g1 = 0.0417, g2 = 0.0338 for 184 parameters and 0 re-
straints, GooF = 1.039, residual density (max./min.) = 0.160/–0.135
e·Å–3.
[8] P. W. Musch, D. Scheidel, B. Engels, J. Phys. Chem. A 2003,
107, 11223–11230.
[9] M. Christl, M. Braun, in: Strain and Its Implications in Organic
Chemistry (Eds.: A. de Meijere, S. Blechert), Kluwer, Dord-
recht, 1989, pp. 121–131.
[10] M. Christl, M. Braun, G. Müller, Angew. Chem. 1992, 104,
471–473; Angew. Chem. Int. Ed. Engl. 1992, 31, 473–476.
[11] S. Groetsch, J. Spuziak, M. Christl, Tetrahedron 2000, 56,
4163–4171.
16/ent-16: C19H16, M = 244.34, monoclinic, space group P21/c, Z
= 4, a = 2483.0(2) pm, b = 616.29(6) pm, c = 847.64(8) pm, β =
91.530(2)°, V = 1.2966(2) nm3, ρc = 1.252 Mg·m–3, 27579 reflec-
tions measured, 2657 unique, R1 [IϾ2σ(I)] = 0.0746, wR2 (all data)
= 0.1752, g1 = 0.0330, g2 = 2.3117 for 221 parameters and 0 re-
straints, GooF = 1.230, residual density (max./min.) = 0.299/–0.235
e·Å–3.
[12] a) N. V. Volchov, A. V. Zabolotskikh, M. B. Lipkind, O. M. Ne-
fedov, Izv. Akad. Nauk SSSR, Ser. Khim. 1989, 1935; Chem.
Abstr. 1990, 112, 138661d; b) N. V. Volchov, A. V. Zabolots-
kikh, A. V. Ignatenko, O. M. Nefedov, Izv. Akad. Nauk SSSR,
Ser. Khim. 1990, 1609–1613; Chem. Abstr. 1992, 116, 6220g; c)
W. R. Dolbier Jr., J. J. Keaffaber, C. R. Burkholder, H. Koron-
iak, J. Pradhan, Tetrahedron 1992, 48, 9649–9660.
[13] R. B. Woodward, R. Hoffmann, Angew. Chem. 1969, 81, 797–
869; Angew. Chem. Int. Ed. Engl. 1969, 8, 781–853.
[14] S. W. Benson, Thermochemical Kinetics, 2nd ed., Wiley-Inter-
science, New York, 1976, p. 309.
[15] M. Christl, S. Drinkuth, Eur. J. Org. Chem. 1998, 237–241.
[16] M. Christl, S. Groetsch, K. Günther, Angew. Chem. 2000, 112,
3395–3397; Angew. Chem. Int. Ed. 2000, 39, 3261–3263.
[17] E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Or-
ganic Compounds, Wiley, New York, 1994, chapter 13.
[18] S. Grimme, Rev. Comput. Chem. 2004, 20, 153–218.
[19] R. Ahlrichs, M. Bär, H.-P. Baron, R. Bauernschmitt, S. Böcker,
M. Ehrig, K. Eichkorn, S. Elliott, F. Haase, M. Häser, H.
Horn, C. Huber, U. Huniar, M. Kattannek, C. Kölmel, M.
Kollwitz, C. Ochsenfeld, H. Öhm, A. Schäfer, U. Schneider, O.
Treutler, M. von Arnim, F. Weigend, P. Weis, H. Weiss, TUR-
BOMOLE, v5.8, University of Karlsruhe, Germany, 2005.
[20] A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100.
[21] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[22] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97,
2571–2577.
All non-hydrogen atoms were refined with anisotropic displace-
ment parameters. Hydrogen atoms of (+)-1 and 8/ent-8 as well as
the bridgehead hydrogen atoms of 15/ent-15 were located by differ-
ence Fourier syntheses and refined isotropically, whilst those of 16/
ent-16 and the nonbridgehead hydrogen atoms of 15/ent-15 were
assigned ideal positions and refined isotropically by using a riding
model with Uiso constrained to 1.2 times the Ueq of the parent
atom. Compound 16/ent-16 crystallised as a nonmerohedral twin.
Therefore, two matrices for the two twin domaines have been deter-
mined by using the program GEMINI.[42] The data were then inte-
grated by using the integration software SAINT[43] and the orienta-
tion matrix of the main twin domain. The twin rotation matrix was
determined from the integrated data with the program ROTAX
implemented in the program suite WinGX.[44] Based on the 180°
rotation twin axis, a HKLF5 file was received (2 BASF param-
eters), which was used for the subsequent refinements. This pro-
cedure led to a sophisticated structural model with drastically re-
duced R values compared to a standard refinement omitting the
twinning.
CCDC-607798, -607799, -607800, and -607801 contain the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[23] O. Vahtras, J. Almlöf, W. M. Feyereisen, Chem. Phys. Lett.
1993, 213, 514–518.
[24] K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs,
Chem. Phys. Lett. 1995, 242, 652–660.
Supporting Information (see also the footnote on the first page of
this article): Cartesian coordinates and absolute energies of the
compounds (+)-1, (–)-1, 7, ent-8, and ent-15.
[25] K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs,
Chem. Phys. Lett. 1995, 240, 283–290.
[26] R. Bauernschmitt, R. Ahlrichs, F. H. Hennrich, M. M. Kappes,
J. Am. Chem. Soc. 1998, 120, 5052–5059.
[27] R. Bauernschmitt, R. Ahlrichs, Chem. Phys. Lett. 1996, 256,
Acknowledgments
454–464.
[28] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[29] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100,
5829–5835.
This work was supported by the Deutsche Forschungsgemein-
schaft. We are grateful to Prof. Gerhard Bringmann and his group
Eur. J. Org. Chem. 2006, 5045–5058
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5057

M. Christl, D. Stalke, B. Engels et al.
FULL PAPER
[30] M. Nendel, L. M. Tolbert, L. E. Herring, Md. N. Islam, K. N.
Houk, J. Org. Chem. 1999, 64, 976–983.
[31] R. L. Elliott, N. H. Nicholson, F. E. Peaker, A. K. Tackle,
C. M. Richardson, J. W. Tyler, J. White, M. J. Pearson, D. S.
Eggleston, R. C. Haltiwanger, J. Org. Chem. 1997, 62, 4998– [39] G. M. Sheldrick, SADABS 2.10, Progam for Area Detector Ab-
5016.
[32] W. E. Parham, H. E. Reiff, P. Swartzentruber, J. Am. Chem.
Soc. 1956, 78, 1437–1440.
[33] For this method to generate dibromocarbene, see: a) W. T.
Miller Jr., C. S. Y. Kim, J. Am. Chem. Soc. 1959, 81, 5008–
5009; b) L. Skattebøl, Tetrahedron Lett. 1961, 167–172.
[34] M. Schlosser, G. Heinz, Chem. Ber. 1971, 104, 1934–1941.
[35] K.-y. Jung, M. Koreeda, J. Org. Chem. 1989, 54, 5667–5675.
[36] B. Hankinson, H. Heaney, R. P. Sharma, J. Chem. Soc., Perkin
Trans. 1 1972, 2372–2377.
[37] D. N. Munro, D. J. Milner, Chem. Ind. (London) 1982, 603.
[38] a) T. Kottke, D. Stalke, J. Appl. Crystallogr. 1993, 26, 615–619;
b) T. Kottke, R. J. Lagow, D. Stalke, J. Appl. Crystallogr. 1996,
29, 465–468; c) D. Stalke, Chem. Soc. Rev. 1998, 27, 171–178.
sorption Correction, University of Göttingen, Germany, 2002.
[40] G. M. Sheldrick, Acta Crystallogr., Sect. A 1990, 46, 467–473.
[41] G. M. Sheldrick, SHELXL-97, Program for Structure Refine-
ment, University of Göttingen, Germany, 1997.
[42] Bruker Nonius Inc., GEMINI 1.02, Madison, WI, 2000.
[43] Bruker Nonius Inc., SAINT-NT, Madison, WI, 2000.
[44] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837–838.
Received: May 22, 2006
Published Online: September 7, 2006
5058
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5045–5058