1-phenyl-1,2-cyclohexadiene: Astoundingly high enantioselectivities on generation in a doering-moore-skattebel reaction and interception by activated olefins

Article abstract of DOI:10.1002/chem.200900718

The resolution of (1α,5α,6α)-6-bromo-6-fluoro-l- phenylbicyclo[3.1.0]hexane (rac-5) provided the enantiomerically pure precursors (-)-5 and (+)-5 of 1-phenyl-1,2-cyclohexadiene. On treatment of (-)-5 with methyllithium in the presence of 2,5-dimethylfuran

Full text of DOI:10.1002/chem.200900718

DOI: 10.1002/chem.200900718
1-Phenyl-1,2-cyclohexadiene: Astoundingly High Enantioselectivities on
Generation in a Doering–Moore–Skattebøl Reaction and Interception by
Activated Olefins**
Manfred Christl,* Hartmut Fischer, Mario Arnone, and Bernd Engels*[a]
Dedicated to Professor Helmut Quast on the occasion of his 75th birthday
Abstract: The resolution of (1a,5a,6a)-
6-bromo-6-fluoro-1-phenylbicyclo-
ACHTUNGTRENNUNG[3.1.0]hexane (rac-5) provided the
enantiomerically pure precursors (ꢀ)-5
and (+)-5 of 1-phenyl-1,2-cyclohexa-
diene. On treatment of (ꢀ)-5 with
methyllithium in the presence of 2,5-di-
methylfuran, the pure (ꢀ)-enantiomer
of the [4+2] cycloadduct of 2,5-di-
enantiospecifically intercepted to give
the product. In the case of indene as
trap for (M)-7, the (ꢀ)- and the (+)-
enantiomer of the [2+2] cycloadduct
were formed in the ratio of 95:5.
Highly surprising, remarkable enantio-
selectivities were also observed, when
(M)-7 was trapped with styrene to fur-
nish two diastereomeric [2+2] cycload-
ducts. Hence, the achiral conformation
of the diradical conceivable as inter-
mediate cannot play a decisive part.
The enantioselective generation of
(M)- and (P)-7 by the b-elimination
route was tested as well. Accordingly,
1-bromo-2-phenylcyclohexene was ex-
posed to the potassium salt of (ꢀ)-
menthol in the presence of 2,5-dime-
thylfuran, and the enantiomeric [4+2]
cycloadducts of the latter onto (M)-
and (P)-7 were produced in the ratio of
55:45.
ACHTUNGTRENNUNGmethylfuran onto 1-phenyl-1,2-cyclo-
hexadiene was obtained exclusively.
From this result, it is concluded that
pure (M)-1-phenyl-1,2-cyclohexadiene
((M)-7) emerged from (ꢀ)-5 and was
Keywords: asymmetric induction ·
cyclic allenes
·
cycloaddition
·
diradicals · enantioselectivity
Introduction
The steric course of the generation of a cyclic allene and its
interception by an activated olefin raise highly interesting
questions as to the mechanisms involved therein. In a first
attempt to give answers, we recently generated the isonaph-
thalenes (P)-2 and its enantiomer (M)-2 from the enantio-
merically pure isolatable bromofluorcarbene adducts of
indene, for example, 1, in Doering–Moore–Skattebøl reac-
tions in the presence of 2,5-disubstituted furans (3) and ob-
tained the [4+2] cycloadducts 4 and their enantiomers ent-4
[a] Prof. Dr. M. Christl, Dr. H. Fischer, Dr. M. Arnone,
Prof. Dr. B. Engels
Institut fꢀr Organische Chemie
Universitꢁt Wꢀrzburg
Am Hubland, 79074 Wꢀrzburg (Germany)
Fax : (+49)931-8884606
E-mail: christl@chemie.uni-wuerzburg.de
engels@chemie.uni-wuerzburg.de
[**] Cycloallenes, Part 22. For Part 21, see reference [5], preceding paper
in this issue.
Scheme 1. Generation of the isonaphthalenes (P)- and (M)-2 from the
pure precursor enantiomer 1 and interception of 2 by 2,5-disubstituted
furans 3, which gives rise to the [4+2] cycloadducts 4 and ent-4 in the
ratio of approximately 70:30.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900718.
11266
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11266 – 11272

FULL PAPER
in a ratio of about 70:30 (Scheme 1).[1] The results of care-
fully designed experiments led to the conclusion that (P)-
and (M)-2 emerge from 1 in a ratio of about 70:30 and,
without enantiomerisation, enantiospecifically add onto 3.
This was unexpected, since the barrier to enantiomerisation
of 2 had been estimated by quantum chemical methods to
be only 11 kcalmolꢀ1,[2,3] which is why the reaction of 2 with
3 must be very fast.
To test the model of the stereochemical course proposed
on the basis of the results with 2, we performed a study with
a 1-substituted 1,2-cyclohexadiene. For such a case, a higher
retention of the stereochemical information of an enantio-
merically pure precursor was anticipated than with 2.[1] We
chose 1-phenyl-1,2-cyclohexadiene ((M)- and (P)-7), for
which the barrier to enantiomerisation is expected to be
close to that of 2, since the phenyl group should have a simi-
lar effect on the allyl-diradical structure of the transition
state as the benzo group of 2. After all, a chiral ground state
of 7 is supported by the result of an AM1 calculation.[4] The
racemic intermediate (rac-7) was examined by two groups in
addition to the Wꢀrzburg team previously. The results are
comprehensively reported in the preceding paper.[5]
Figure 1. Comparison of the experimental (c) and calculated (B3LYP;
a) CD spectra of 5, which is thus proved to be the (ꢀ)-enantiomer.
Results and Discussion
Generation of the cyclic allenes (M)- and (P)-7 from the
pure enantiomers 5 and ent-5 and trapping of the intermedi-
ates 7: As shown in the preceding paper,[5] the bromofluoro-
ACHTUNGTRENNUNGcarbene adduct rac-5 of 1-phenylcyclopentene is a useful
precursor of rac-7 with regard to the Doering–Moore–Skat-
tebøl reaction. For the enantioselective generation of (M)-
and (P)-7, rac-5 had to be separated into the pure enantio-
mers 5 and ent-5. Indeed, the resolution of rac-5 was
ACHTUNGTRENNUNGachieved by HPLC using a Chiralcel OJ-H column. To de-
termine the absolute configuration, we compared the experi-
mental CD spectra with the calculated ones, which were ob-
tained by the quantum chemical procedure described at the
end of the Experimental Section. Such a comparison is illus-
trated by Figure 1, which proves that the (ꢀ)-enantiomer
has the 1R,5S,6R configuration (5).
We used 5 as well as its enantiomer (ent-5) for the libera-
tion of the cyclic allene 7, which was trapped by 2,5-di-
ACHTUNGTRENNUNGmethylfuran, indene and styrene. The results for the case of
Scheme 2. The reaction of the pure enantiomer 5 with methyllithium
most probably led to the pure M enantiomer of 1-phenyl-1,2-cyclohexa-
diene ((M)-7), which was trapped by 2,5-dimethylfuran, indene and sty-
rene to give pure 6, 8 and ent-8 in the ratio of 95:5, 9 and ent-9 in the
ratio of 94:6, and 10 and ent-10 in the ratio of 85:15, respectively.
5 are collected in Scheme 2. The formulas represent abso-
lute configurations. As with 5 and ent-5, those of the prod-
ucts (6, 8–10) were established by resolution of the racemic
products[5] by HPLC using Chiralcel columns and compari-
son of the experimental with the calculated CD spectra (see
the end of the Experimental Section). Figures 2, 3, 4 and 5
display these comparisons. The analysis of the products as to
their enantiomeric ratios was carried out by HPLC using a
Chiralcel OJ-H column.
As is evident from Scheme 2, the treatment of 5 with
methyllithium in the presence of 2,5-dimethylfuran fur-
nished the pure [4+2] cycloadduct 6 of the intermediate
cyclic allene in 31 % yield. Since no ent-6 could be detected,
the stereochemical information of 5 was completely trans-
ferred to the product, which is why the intermediate must
have been also a pure enantiomer, and we propose that this
was (M)-7 in the next subsection. Indene as a trap afforded
a 22 % yield of the [2+2] cycloadduct enantiomers 8 and
ent-8 in the ratio of 95:5. In the case of styrene, a mixture of
the two possible diastereomeric [2+2] cycloadducts in the
ratio of 53:47 was obtained in 59 % yield. The endo diaste-
reomer consisted of the enantiomers 9 and ent-9 in the ratio
of 94:6, whereas the exo diastereomer comprised the enan-
Chem. Eur. J. 2009, 15, 11266 – 11272
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11267

M. Christl, B. Engels et al.
Figure 2. Comparison of the experimental (c) and calculated (B3LYP;
a) CD spectra of 6, which is thus proved to be the (ꢀ)-enantiomer.
Figure 4. Comparison of the experimental (c) and calculated (B3LYP;
a) CD spectra of ent-9, which is thus proved to be the (ꢀ)-enantiomer.
Figure 3. Comparison of the experimental (c) and calculated (B3LYP;
a) CD spectra of ent-8, which is thus proved to be the (+)-enantiomer.
Figure 5. Comparison of the experimental (c) and calculated (B3LYP;
a) CD spectra of 10, which is thus proved to be the (+)-enantiomer.
tiomers 10 and ent-10 in the ratio of 85:15. The HPLC trace
is depicted in Figure 6 (see the Experimental Section) and
provided the precise ratio of 9/ent-9/10/ent-10 to be
50:3:40:7.
of the two bridgehead hydrogen atoms, H1a or H6b, has to
change sides of the ring system, that is, H6b en route to
(M)-2 and H1a en route to (P)-2. Since on that move H6b
experiences a stronger steric hindrance by the neighbouring
aromatic CH group than H1a by a hydrogen atom of the
CH2 group, we suggested the preferential formation of (P)-2
(ca. 70 %).[1] The application of this reasoning to the reac-
tion 5!7 leads to an unambiguous conclusion, as this pro-
cess requires the change of the ring sides either from the
bridgehead hydrogen atom or the phenyl group. Owing to
its size, the phenyl group should encounter a much larger re-
straint of its move by the proximal CH2 group than the
bridgehead hydrogen atom. Therefore we propose the exclu-
The stereochemical courses of the Doering–Moore–Skatte-
bøl reaction of 5 and the cycloadditions of (M)-7: The find-
ing that only one enantiomer of the [4+2] cycloadduct of
2,5-dimethylfuran onto the intermediate cyclic allene,
namely, 6, emerged from the pure enantiomer 5 demands
the stereospecific conversion of 5 into the intermediate. The
question is only whether (M)- or (P)-7 results from 5. On
generation of the isonaphthalenes 2 from 1 (Scheme 1), one
11268
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11266 – 11272

1-Phenyl-1,2-cyclohexadiene
FULL PAPER
sive liberation of (M)-7 from 5. As we had anticipated,[1] the
stereoselectivity of the generation of a cyclic allene is im-
pressively improved by a phenyl group at one bridgehead
position of a substrate for the Doering–Moore–Skattebøl re-
action.
ApACHTNGUTERNNUpG arently, the subtle deviation in the size of these two
groups causes the preference of the attack of (M)-7 at the
Re face of indene (14) over that at the Si face (16) by a
factor of 19 (Scheme 4).
Given this assignment and the fact that (M)-7 does not at
all undergo the inversion to (P)-7 under the reaction condi-
tions, it has to be rationalised why (M)-7 adds onto 2,5-di-
methylfuran to give 6 exclusively. Quantum chemical calcu-
lations[6] as well as the experimental result of the addition of
(Z,Z)-1,4-dideuterio-1,3-butadiene
onto
1,2-cyclohexa-
diene[3] support a two-step mechanism through diradical in-
termediates even for the [4+2] cycloadditions of six-mem-
bered cyclic allenes. Scheme 3 shows the diastereomeric
Scheme 4. Transition states (14, 16) for the addition of the central allene
carbon atom of (M)-7 to both enantiotopic faces of indene. The major
product 8 should be formed through 14 and the diradical 15, the minor
product ent-8 through 16 and ent-15. In 14 and 16, (M)-7 is represented
by the PhC=C=CH subunit only and approaches indene from above the
drawing plane.
Scheme 3. Possible transition states (11, 13) for the addition of the central
allene carbon atom of (M)-7 to both enantiotropic faces of the CMe
group of 2,5-dimethylfuran. Only 11 leads to diradical 12, which should
be the intermediate en route to 6. In 11 and 13, the cyclic allene (M)-7 is
represented by the PhC=C=CH subunit only and approaches the furan
from above the drawing plane.
Previously, we had examined the addition of indene to the
isonaphthalenes 2 by using the pure enantiomer 1 as allene
precursor (Scheme 1) and found only very little transfer of
the stereochemical information carried by 1 to the prod-
uct.[1] Possibly, (M)- and (P)-2 are unable to clearly distin-
guish between the Re and the Si face of indene because of
two hydrogen atoms at the allene termini. Clearly, the
phenyl group of (M)- and (P)-7 makes the difference.
transition states 11 and 13 for the formation of the diradi-
cals by addition of the central allene carbon atom of (M)-7
to both enantiotopic faces of a CMe group of 2,5-dimethyl-
furan (Re: 11; Si: 13). Only 11 can give rise to the diradical
12, which is the precursor of 6, whereas 13 would produce
the enantiomer of 12 (ent-12) and in consequence the enan-
tiomer of 6 (ent-6). The phenyl group, which requires a
great deal of space, is located as far away as possible from
the core of the allenophile in both transition states. The cru-
cial difference between them has to be seen in the proximity
of the phenyl group to the oxygen atom in 11 and to a CH
group in 13. As the spatial demand of an oxygen atom is
substantially smaller than that of a CH group, 11 is pre-
ferred over 13 to the extent that 13 has no importance at all.
Such an exclusiveness does not hold for the cycloaddition
of indene, in which the enantiomeric [2+2] cycloadducts 8
and ent-8 resulted in the ratio of 95:5. Assuming that the
generation of pure (M)-7 from 5 also occurs, we have to
take into account the transition states 14 and 16, from which
the major product 8 and the minor one ent-8 emerge
through the diradicals 15 and ent-15, respectively. Exhibiting
the common feature that the phenyl group as the only big
substituent is located as far away as possible from the core
of indene, 14 and 16 differ in the arrangement of the five-
membered ring relative to the phenyl group as the latter is
next to the 3-CH group in 14 and to the CH2 group in 16.
Even assuming that the reaction of 5 with methyllithium
in the presence of styrene liberated pure (M)-7, we were
highly surprised about the relatively high enantiomeric
ratios of 94:6 and 85:15 for the [2+2] cycloadducts 9/ent-9
and 10/ent-10, respectively. Based on previous studies,[7] we
had anticipated a diradical intermediate with the achiral
conformation 17 (Scheme 5) for the addition of (M)- or (P)-
7 onto styrene and hence the complete loss of the stereo-
chemical information introduced by 5. However, the mecha-
nism of the styrene addition onto six-membered cyclic al-
lenes had already proven to be interesting earlier, since on
use of (Z)-b-deuteriostyrene the stereochemical information
is entirely lost in the trapping of 1,2-cyclohexadiene,[3,7,8] but
partially retained in the case of 1-oxa-2,3-cyclohexadiene[3]
and completely retained with a 1-thia-5-aza-2,3-cyclohexa-
diene derivative.[3,9] Based on their result, Elliott et al.[9] sug-
gested the cycloaddition to be concerted and supported
their view by quantum chemical calculations of a model
system. A one-step course was also proposed for several
[2+2] cycloadditions of 1-oxa-2,3-cyclohexadiene.[10] Proba-
bly, conclusive insight into the mechanisms will be provided
only by quantum chemical calculations, which can treat
equally well closed-shell transition states on one side and
Chem. Eur. J. 2009, 15, 11266 – 11272
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11269

M. Christl, B. Engels et al.
ergetically favourable transition states 18a and 18c are
passed and that the resulting diradicals 17a and 17b under-
go a conformational change to give ent-17b and ent-17a to a
minor extent. The low barriers to the collapse of 17a,b and
ent-17a,b are reminiscent of the behaviour of the diradicals
generated on thermolysis of 3-vinylmethylenecyclobutanes
and converted into 4-methylenecyclohexenes without com-
plete equilibration.[11]
Generation of cyclic allenes (M)- and (P)-7 from 1-bromo-
2-phenylcyclohexene (19) by enantiomerically pure potassi-
um menthoxide and trapping of the intermediates: In 1981,
Balci and Jones[12] treated 3-bromobicyclo
ACTHNUTRGNE[NUG 3.2.1]octa-2,6-
diene, 1-bromocyclohexene and 1-bromocycloheptene with
enantiomerically pure potassium menthoxide in the pres-
ence of 1,3-diphenylisobenzofuran, which intercepted the
cyclic allene intermediates bicycloACTHUNTGRNEUNG[3.2.1]octa-2,3,6-triene,
1,2-cyclohexadiene and 1,2-cycloheptadiene in [4+2] cyclo-
additions. Since the products were optically active, the chir-
ality of the intermediates was thus demonstrated, although
the specific rotations were tiny ([a]2D5 ꢁ1.9). An analysis of
the ratio of the product enantiomers was not carried out
and would have been very difficult at that time.
Because we now had available the analytical means, we
quantitatively tested the capability of enantiomerically pure
potassium menthoxide for asymmetric induction in the b
elimination en route to 1-phenyl-1,2-cyclohexadiene (7). Ac-
cordingly, we exposed 1-bromo-2-phenylcyclohexene (19) to
the potassium salt of (ꢀ)-menthol in the presence of 2,5-di-
methylfuran. Previously, we had obtained the racemic [4+2]
cycloadduct rac-6 of rac-7 with this furan on treatment of 19
with potassium tert-butoxide in 9% yield.[5] The chiral and
enantiomerically pure base now brought about a 26% yield
of 6 and ent-6, the ratio of which was determined from the
specific rotation ([a]D22 =++34) to be 45:55 (Scheme 6).
Scheme 5. Possible transition states (18a–d) and intermediates (17a,b
and ent-17a,b) for the addition of the central allene carbon atom of (M)-
7 to the methylene group of styrene. In 18a–d, (M)-7 is represented by
the PhC=C=CH subunit only and approaches styrene from above the
drawing plane.
diradical intermediates on the other. Since even [4+2] cyclo-
additions of 1,2-cyclohexadiene are considered to proceed
stepwise through diradicals,[3,6] we adhere to a general two-
step course of the styrene additions onto cyclic allenes all
the more, as the postulated diradical intermediates are per-
fectly stabilised by resonance. Thus, instead of 17, its chiral
conformers 17a, ent-17a, 17b and ent-17b (Scheme 5)
should be intermediates en route from (M)-7 to the prod-
ucts. ApACHTUNGTRENNUNGparently, these conformers have to cyclise faster
than they equilibrate. This conclusion can be drawn even
from the finding that 9+ent-9 and 10+ent-10 are formed in a
ratio of close to 1:1, although the equilibrium ratio is 1:10 or
even smaller.[5]
Although the temperature of this experiment was higher
by 808C than in the above reaction of 5 with methyllithium
As illustrated by Scheme 5, the four transition states 18a–
d are conceivable for the attack of the central allene carbon
atom of (M)-7 at the methylene group of styrene. Because
of the mutual proximity of the two phenyl groups, 18b
should be the least favourable transition state. This is the
most probable reason why ent-9, the product emerging from
18b through the diradical ent-17a, resulted as the least
abundant stereoisomer. On the other hand, 18a and 18c
suffer from only one 1,3- and 1,4-interaction, respectively,
between a phenyl group and a hydrogen atom. This would
explain why 9 and 10 arise as major stereoisomers with 17a
and 17b, respectively, being the intermediate diradicals. In
18d, steric overcrowding due to both a 1,3- and a 1,4-inter-
action between a phenyl group and a hydrogen atom leads
to a considerable strain energy; this accounts for the small
quantity of ent-10 observed, which should originate from
18d through ent-17b. A possible alternative rationalisation
for the small amounts of ent-9 and ent-10 is that only the en-
Scheme 6. Reaction of 1-bromo-2-phenylcyclohexene (19) with the potas-
sium salt of (ꢀ)-menthol in the presence of 2,5-dimethylfuran.
11270
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11266 – 11272

1-Phenyl-1,2-cyclohexadiene
FULL PAPER
(ent-6) using a Chiralcel OD column and 6 min for 6 and 17 min for ent-6
on a Chiralcel OJ-H column. Specific rotations (methanol, c=1, T=
208C): [a]D =ꢀ328 and +321. The absolute configuration was deter-
mined by comparison of the experimental with the calculated CD spec-
trum (Figure 2).
in the presence of 2,5-dimethylfuran with formation of pure
6, we believe that the equilibration of (M)- and (P)-7 did
not proceed here, either. Given this prerequisite, the asym-
metric induction of the enantiomerically pure potassium
menthoxide is low, that is, the base recognised only a very
small difference between the enantiotopic protons of the 6-
methylene group of 19. However, we predict that other
chiral bases may be much more effective in this distinction.
Reaction of the pure enantiomer 5 with methyllithium in the presence of
2,5-dimethylfuran: The reaction of 5 (60.1 mg, 0.236 mmol), dissolved in
2,5-dimethylfuran (3 mL), with methyllithium (2.0 mmol, 2.0 mL of 1.0m
in diethyl ether) was conducted as described for rac-5[5] and furnished the
purified product as a colourless oil (18.2 mg, 31 %). A solution of this
product in 10 mL of heptane was analysed for the enantiomeric purity by
using the Chiralcel OJ-H column (see above). Only the signal of 6 ap-
peared, whereas that of ent-6 was completely missing (>98%ee). Specific
rotation (methanol, c=0.6, T=208C): [a]D =ꢀ352. The deviation from
the above value (ꢀ328) may have its origin in the error limits of the
weighing.
Conclusion
With 1-phenyl-1,2-cyclohexadiene ((M)- and (P)-7) as an ex-
ample, we have demonstrated that 1-substituted 1,2-cyclo-
hexadienes can be generated with a high enantioselectivity
from enantiomerically pure precursors in Doering–Moore–
Skattebøl reactions and intercepted by activated olefins in
[4+2] and [2+2] cycloadditions with complete or substantial
retention of the stereochemical information introduced by
the precursors. The mechanistic details proposed for the for-
mation and the trapping of (M)- and (P)-7 are a challenge
for computational chemists and, if verified, will considerably
extend the insight into the chemistry of strained cyclic al-
lenes and diradical intermediates.
Resolution of rac-8 and determination of the absolute configuration of
the enantiomers: The equipment described above was used. A Chiralcel
OJ-H column was utilised and heptane/ethanol (220:1) was the eluant.
Samples (500 mL) of an 0.02m solution of rac-8 in heptane were injected.
The retention times were 7.7 min for the (ꢀ)-enantiomer (8) and
10.3 min for the (+)-enantiomer (ent-8). Specific rotations (hexane, c=
0.4, T=208C): [a]D =ꢀ258 and +260. The absolute configuration was de-
termined by comparison of the experimental with the calculated CD
spectrum (Figure 3).
Reaction of the pure enantiomer ent-5 with methyllithium in the pres-
ence of indene: The reaction of ent-5 (59.8 mg, 0.234 mmol), dissolved in
indene (5 mL), with methyllithium (2.0 mmol, 2.0 mL of 1.0m in diethyl
ether) was conducted as described for 6,6-dibromo-1-phenylbicyclo-
AHCTUNGTRENNUNG
[3.1.0]hexane[5] and furnished the purified product as a colourless oil
(14 mg, 22 %). A solution of this oil in heptane was analysed for the
ratio of the enantiomers by HPLC with a Chiralcel OJ-H column under
the conditions given above for rac-8. The ratio of 8/ent-8 turned out to be
5:95. Specific rotation of this mixture (hexane, c=0.3, T=218C): [a]D =
+178. The enantiomeric ratio calculated from the specific rotation
(16:84) is considered to be less reliable than that determined by
HPLC.[13]
Experimental Section
General: See ref. [5]. CD: JASCO J-715 spectropolarimeter. Specific ro-
tations: JASCO P-1020 polarimeter; the units of the [a]D values are
deg10ꢀ1 cm2 gꢀ1; the concentrations (c) are given in grams per 100 mL so-
lution.
Resolution of rac-9 and rac-10 and determination of the absolute config-
uration of the enantiomers: The equipment described above was used. A
Chiralcel OJ-H column was utilised and heptane/ethanol (250:1) was the
eluant. Samples (500 mL) of an 0.04m solution of the 1:1 mixture of rac-9
and rac-10 in heptane were injected. The diastereomers as well as the en-
antiomers were cleanly separated, with the retention times being 9.5 and
19.5 min for the (ꢀ)-enantiomers ent-9 and ent-10, respectively, and 23.6
and 38.4 min for the (+)-enantiomers 10 and 9, respectively (Figure 6).
Specific rotations (heptane, c=1, T=218C): [a]D =+224 (9), ꢀ229 (ent-
9), +35 (10) and ꢀ34 (ent-10). The absolute configurations were deter-
mined by comparison of the experimental with the calculated CD spectra
(Figures 4 and 5).
The racemates of all compounds described below have been fully charac-
terised in the preceding paper.[5] The pure enantiomers and the non-race-
mic mixtures were identified as to their constitution and, where applica-
ble, their diastereomeric nature by NMR spectroscopy.
Resolution of racemates and analysis of non-racemic mixtures of enantio-
mers: An HPLC system consisting of a Knauer HPLC pump 64, a Gyn-
kotek UV detector (UVD), operated at 260 nm, and a Chiralcel OJ-H or
OD column (each 250ꢃ21 mm) was used, with each protected by a guard
column (50ꢃ21 mm) with the same stationary phase. The flow rate was
maintained at a value between 12 and 18 mLminꢀ1. Before their use as
eluants, the solvents were distilled through 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
Reaction of the pure enantiomer 5 with methyllithium in the presence of
styrene: The reaction of 5 (65.0 mg, 0.255 mmol), dissolved in styrene
(5 mL), with methyllithium (2.4 mmol, 1.5 mL of 1.6m in diethyl ether)
was conducted as described for 6,6-dibromo-1-phenylbicyclo-
ACHTUNGTRENNUNGachieved by using a Shimadzu C-R3A Chromatopac integrator.
Resolution of rac-5 and determination of the absolute configuration of
the enantiomers: The equipment described above was used. A Chiralcel
OJ-H column was utilised and heptane/ethanol (250:1) was the eluant.
Samples (100 mL) of an 0.09m solution of rac-5 in heptane were injected.
The retention times were 16.5 min for the (+)-enantiomer (ent-5) and
18 min for the (ꢀ)-enantiomer (5). Specific rotations (heptane, c=0.1
and 0.25, T=208C): [a]D =+4 and ꢀ5. The absolute configuration was
determined by comparison of the experimental with the calculated CD
spectrum (Figure 1).
AHCTUNGTRENNUNG
[3.1.0]hexane[5] and furnished the purified product as a colourless oil
(38.2 mg, 59%). A solution of this oil in heptane was analysed for the
ratio of the enantiomers by HPLC using a Chiralcel OJ-H column under
the conditions given above for rac-9 and rac-10. The ratio of 9/ent-9/10/
ent-10 turned out to be 50:3:40:7 (Figure 6). Specific rotation of this mix-
ture (pentane, c=1.0, T=218C): [a]D =+195.
Reaction of 1-bromo-2-phenylcyclohexene (19) with enantiomerically
pure potassium menthoxide in the presence of 2,5-dimethylfuran: A
stirred suspension of potassium hydride (80.2 mg, 2.00 mmol) in tetrahy-
drofuran (10 mL) was treated dropwise with (ꢀ)-menthol (313 mg,
2.00 mmol) at 208C. According to the literature,[14] the mixture was
stirred overnight at 208C and then added dropwise to a vigorously stirred
solution of 19[5] (450 mg, 1.90 mmol) in 2,5-dimethylfuran (4.50 g,
46.8 mmol) at 208C. Stirring was continued for 6 h at 508C. After having
been cooled to room temperature, the mixture was hydrolysed (5 mL)
Resolution of rac-6 and determination of the absolute configuration of
the enantiomers: The equipment described above was used. Both Chiral-
cel OD and OJ-H columns were suitable with hexane/2-propanol (200:1)
and heptane/ethanol (199:1), respectively, as the eluants. Samples
(100 mL) of an 0.2m solution of rac-6 in hexane/2-propanol (1:1) and hep-
tane/ethanol (1:1), respectively, were injected. The retention times were
8 min for the (ꢀ)-enantiomer (6) and 10 min for the (+)-enantiomer
Chem. Eur. J. 2009, 15, 11266 – 11272
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11271

M. Christl, B. Engels et al.
Acknowledgements
This work was supported by the Deutsche Forschungsgemeinschaft. We
are grateful to the group of Prof. Dr. Gerhard Bringmann for help with
the measurements of the CD spectra and the specific rotations.
[1] M. Christl, M. Braun, H. Fischer, S. Groetsch, G. Mꢀller, D. Leusser,
S. Deuerlein, D. Stalke, M. Arnone, B. Engels, Eur. J. Org. Chem.
2006, 5045–5058.
[2] B. Engels, J. C. Schçneboom, A. F. Mꢀnster, S. Groetsch, M. Christl,
J. Am. Chem. Soc. 2002, 124, 287–297.
[3] M. Christl in Modern Allene Chemistry, Vol. 1 (Eds.: N. Krause,
A. S. K. Hashmi), Wiley-VCH, Weinheim, 2004, pp. 243–357.
[4] L. M. Tolbert, Md. N. Islam, R. P. Johnson, P. M. Loiselle, W. C.
Shakespeare, J. Am. Chem. Soc. 1990, 112, 6416–6417.
[5] M. Christl, M. Schreck, T. Fischer, M. Rudolph, D. Moigno, H.
Fischer, S. Deuerlein, D. Stalke, Chem. Eur. J. 2009, DOI: 10.1002/
chem.200900717, preceding paper in this issue.
Figure 6. Diagrams of the HPLC analyses using a Chiralcel OJ-H column
with heptane/ethanol (250:1), of 1:1 mixture of rac-9 and rac-10
(bottom) and of the product obtained from the reaction of the pure enan-
a
tiomer 5 with methyllithium in the presence of styrene (top).
[6] M. Nendel, L. M. Tolbert, L. E. Herring, M. N. Islam, K. N. Houk, J.
Org. Chem. 1999, 64, 976–983.
[7] M. Christl, M. Schreck, Chem. Ber. 1987, 120, 915–920.
[8] S. Harnos, S. Tivakornpannarai, E. E. Waali, Tetrahedron Lett. 1986,
27, 3701–3704.
[9] R. L. Elliott, N. H. Nicholson, F. E. Peaker, A. K. Takle, C. M. Ri-
chardson, J. W. Tyler, J. White, M. J. Pearson, D. S. Eggleston, R. C.
Haltiwanger, J. Org. Chem. 1997, 62, 4998–5016.
[10] R. Ruzziconi, Y. Naruse, M. Schlosser, Tetrahedron 1991, 47, 4603–
4610.
[11] a) C. P. Suhrada, C. SelÅuki, M. Nendel, C. Cannizzaro, K. N. Houk,
P.-J. Rissing, D. Baumann, D. Hasselmann, Angew. Chem. 2005, 117,
3614–3618; Angew. Chem. Int. Ed. 2005, 44, 3548–3552; b) Y.-L.
Zhao, C. P. Suhrada, M. E. Jung, K. N. Houk, J. Am. Chem. Soc.
2006, 128, 11106–11113.
[12] M. Balci, W. M. Jones, J. Am. Chem. Soc. 1981, 103, 2874–2876.
[13] E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Organic
Compounds, Wiley, New York 1994, Chapter 13.
and admixed with dichloromethane (10 mL). The layers were separated;
the aqueous layer was extracted with dichloromethane (2ꢃ5 mL); the
combined organic layers were washed with a saturated aqueous solution
of NH4Cl (2ꢃ5 mL), dried with MgSO4 and concentrated in vacuo
(15 mbar). The menthol was distilled off in a kugelrohr (508C/0.1 mbar).
The residue, a colourless oil (124 mg, 26 %), was shown by NMR spec-
troscopy to be the [4+2] cycloadduct of 2,5-dimethylfuran onto 7.[5] Spe-
cific rotation (hexane, c=1.0 T=228C): [a]D =+34, which afforded 6/
ent-6=45:55 by using the specific rotations of pure 6 and ent-6 (see
above). However, the different solvents (hexane and methanol, respec-
tively) may cause wide error limits.
Calculation of the CD spectra of the enantiomers of rac-5, rac-6, rac-8,
rac-9 and rac-10: For the elucidation of the absolute configuration of a
chiral compound, the simulation of its CD spectrum by quantum chemi-
cal methods and the comparison with the experimental spectrum is an es-
tablished method.[13, 15]
The present calculations utilised the density functional theory and its
time-dependent (TD) variant by applying the TURBOMOLE pro-
gramme package.[16] The geometries of the investigated compounds were
optimised at the BLYP/SVP level[17–19] by applying the resolution of iden-
tity approximation[20, 21] together with the matching auxiliary basis
sets.[21,22] Absolute energies and Cartesian coordinates of 5, 6, ent-8, ent-9
and 10 are given in the Supporting Information.
The electronic excitations were calculated at the TD-B3LYP[23–25] level of
theory in combination with the TZVP basis sets.[26] The phenyl groups
can rotate almost freely. For the calculation of the spectra, the conforma-
tion of the energy minimum was chosen. It was shown by a test that a de-
viation from the minimum conformation has only a small effect on the
CD spectrum of the respective compound (see the Supporting Informa-
tion).
[14] P. Haquette, S. Dagorne, R. Welter, G. Jaouen, J. Organomet. Chem.
2003, 682, 240–247.
[15] S. Grimme, Rev. Comput. Chem. 2004, 20, 153–218.
[16] TURBOMOLE, v5.8, R. Ahlrichs, M. Bꢁr, H.-P. Baron, R.
Bauernschmitt, S. Bçcker, M. Ehrig, K. Eichkorn, S. Elliott, F.
Haase, M. Hꢁser, H. Horn, C. Huber, U. Huniar, M. Kattannek, C.
Kçlmel, M. Kollwitz, C. Ochsenfeld, H. ꢄhm, A. Schꢁfer, U.
Schneider, O. Treutler, M. von Arnim, F. Weigend, P. Weis, H.
Weiss, University of Karlsruhe, Karlsruhe, 2005.
[17] A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100.
[18] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[19] A. Schꢁfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571–
2577.
For molecule 5 the first 15 excitations, for 6 and 8 the first 30 excitations,
and for 9 and 10 the first 20 excitations were calculated. The UV spectra
were simulated by superimposing Gauss functions, weighted according to
the oscillator strength,[15] and then compared with the experimental spec-
tra. With respect to the position of the strongest absorption, the calculat-
ed spectra of 8, 9 and 10 showed a hypsochromic shift of 15, 16 and
13 nm, respectively. In the case of 5 and 6, the absorption maximum at
lowest energy exhibits only a shoulder in the spectra and no shift could
be determined. The CD spectra were simulated by superimposing Gauss
functions weighted according to the rotator strength[15] and then shifted
by the number of wavelength units that corresponded to the deviation of
the experimental and calculated UV spectra, given above. In Figures 1, 2,
3, 4 and 5 the simulated CD spectra of 5, 6, ent-8, ent-9 and 10, respec-
tively, are compared with the matching experimental spectra, whereby
the absolute configurations of the resolved samples were determined.
[20] O. Vahtras, J. Almlçf, W. M. Feyereisen, Chem. Phys. Lett. 1993,
213, 514–518.
[21] K. Eichkorn, O. Treutler, H. ꢄhm, M. Hꢁser, R. Ahlrichs, Chem.
Phys. Lett. 1995, 242, 652–660.
[22] K. Eichkorn, O. Treutler, H. ꢄhm, M. Hꢁser, R. Ahlrichs, Chem.
Phys. Lett. 1995, 240, 283–290.
[23] R. Bauernschmitt, R. Ahlrichs, F. H. Hennrich, M. M. Kappes, J.
Am. Chem. Soc. 1998, 120, 5052–5059.
[24] R. Bauernschmitt, R. Ahlrichs, Chem. Phys. Lett. 1996, 256, 454–
464.
[25] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[26] A. Schꢁfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829–
5835.
Received: March 19, 2009
Published online: September 11, 2009
11272
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11266 – 11272