On the mechanism of internal ortho-lithiation in a mixed complex between BuLi and a chiral lithium amide

Article abstract of DOI:10.1002/1522-2675(200211)85:11<3814::AID-HLCA3814>3.0.CO;2-O

The chiral lithium amide Li-1, prepared from (S)-N-[(R)-2-methoxy-1-phenylethyl]-α-methylbenzylamine 1, is known to form a mixed complex with BuLi, i.e., Li-1/BuLi. When this complex is kept at room temperature for 3-5 h, a regioselective lithiation of one of the ortho-C-atoms in the amide occurs. Kinetic studies indicate that this reaction proceeds within the mixed aggregate with a first-order rate constant for the rate-limiting step k₂ = 9.4 × 10^-4 s^-1 (ΔG_273K^≠ = 21 kcal mol^-1). Molecular modeling (semi-empirical PM3) suggests that the observed selectivity and reactivity are a consequence of the solvation and structure of the mixed complex, i.e., an example of a complex induced proximity effect (CIPE).

Full text of DOI:10.1002/1522-2675(200211)85:11<3814::AID-HLCA3814>3.0.CO;2-O

3814
Helvetica Chimica Acta Vol. 85 (2002)
On the Mechanism of Internal ortho-Lithiation in a Mixed Complex Between
BuLi and a Chiral Lithium Amide
by Per I. Arvidsson*
Department of Organic Chemistry, Institute of Chemistry, Uppsala University, Box 531, SE-751 21 Uppsala
and
Gˆran Hilmersson* and ÷jvind Davidsson1)
Organic Chemistry, Department of Chemistry, Gˆteborg University, SE-412 96 Gˆteborg
Dedicated to Professor Dr. Dieter Seebach, a pioneer in the field of lithium organic chemistry and our role
model as scientist, on the occasion of his 65th birthday
The chiral lithium amide Li-1, prepared from (S)-N-[(R)-2-methoxy-1-phenylethyl]-a-methylbenzylamine
1, is known to form a mixed complex with BuLi, i.e., Li-1/BuLi. When this complex is kept at room temperature
for 3 5 h, a regioselective lithiation of one of the ortho-C-atoms in the amide occurs. Kinetic studies indicate
that this reaction proceeds within the mixed aggregate with a first-order rate constant for the rate-limiting step
k2 ˆ 9.4  10À4 sÀ1 (DG2=73K ˆ 21 kcal molÀ1). Molecular modeling (semi-empirical PM3) suggests that the
observed selectivity and reactivity are a consequence of the solvation and structure of the mixed complex, i.e., an
example of a complex induced proximity effect (CIPE).
Introduction. D. Seebach wrote the following in this journal in 1983 [1]: −Die noch
immer steigende Bedeutung von Lithiumorganylen in der organischen Synthese man
denke z. B. an die −neue× Aromatenchemie weckt den Wunsch nach mehr Wissen ¸ber
die Zusammenh‰nge zwischen Struktur und Reaktivit‰t, das heisst nach einem weniger
empirischen Vorgehen bei der Suche nach weiteren Anwendungen dieser Verbindungs-
klasse×2). Detailed structural and mechanistic studies, too many to be properly cited
here, have increased our knowledge about the structure and reactivity of organo-
lithiums tremendously over these twenty years (for selected reviews, see [2]).
Nonetheless, the unprecedented structural complexity of the aggregates formed by
organolithiums makes the reactivity of these substances highly intricate. Thus, it is not
uncommon that we still find new, and unforeseen, reactivity where these substances are
involved. As this contribution illustrates, unexpected reactions may still be encoun-
tered in the, by now, quite −alten Aromatenchemie×3).
Among the many reactions involving lithium organic species, few have attracted the
attention of organic chemists more than this −Aromatenchemie×, i.e., the ortho-directed
lithiation of (especially) heteroatom-containing aromatic rings [3] Scheme 1.
1
)
)
Present address: Medicinal Chemistry, AstraZeneca R&D Mˆlndal, SE-43183 Mˆlndal.
Freely translated: −The constantly growing importance of organolithiums in one thinks especially on the
−new× aromatic substitution chemistry encourage further knowledge about the connection between structure
and reactivity, that means a less empirical approach for finding further use of this class of molecules×.
−Old× aromatic substitution chemistry.
2
3
)

Helvetica Chimica Acta Vol. 85 (2002)
3815
Scheme 1. ortho-Lithiation of an Aromatic Ring by an Alkyllithium or Lithium Amide Reagent. X ˆ Heter-
oatom, TMEDA ˆ N,N,N',N'-tetramethylethylenediamine.
The complementary of this method for aromatic substitution, as compared to
traditional methods based on electrophilic substitution, has triggered an enormous
collective effort of both academic and industrial chemists to exploit ortho-lithiations for
the regiocontrolled synthesis of polysubstituted arenes (for a few recent examples, see
[4]). The origin of the ortho-selectivity observed and rate enhancements found for
heteroatom-substituted aromatics has also been under intense evaluation [5 7]. The
most generally accepted explanation relies on the suggestion that an initial coordina-
tion of the metal to the lone pair on the substituent increases the inductive effect of the
substituent as well as positions the metallating reagent in the vicinity of the ortho-H-
atom and, thus, facilitates its removal of the carbanion. Although the details and
relative importance of these factors are under discussion, the importance of complex-
ation for controlling reactivity, stereo-, and regioselectivity, have been recognized and
designated the −complex induced proximity effect× (CIPE) [8].
Background. Previous investigations by our group have outlined the solution-state
structure of the chiral lithium amide Li-1, prepared from (S)-N-[(R)-2-methoxy-1-
phenylethyl]-a-methylbenzylamine 1. Detailed studies by NMR spectroscopy, X-ray
diffraction, and computational methods have shown that Li-1 forms a C2-symmetric
dimer, i.e., (Li-1)2 ¥ S (S ˆ DEE) in Scheme 2, in diethyl ether ((D10)DEE) solution [9].
The dimer is solvated by one solvent molecule, exhibiting slow exchange on the NMR
time-scale, and is capable of stereoselective solvation, when a racemic solvent mixture,
i.e., 2,3,4,5-tetrahydro-2-methylfuran, is added [10]. Application of the chiral lithium
amide in the enantioselective rearrangement of meso-epoxides to chiral allylic alcohols
has also been demonstrated [9a].
Scheme 2. Preparation and Solution-Phase Aggregation of Li-1. NMR-Spectroscopic studies have shown that
the chiral lithium amide Li-1 is present as a dimer (Li-1)2 in DEE solution. In THF, both dimers and monomers
are present.
Addition of small amounts of THF to a (D10)DEE solution of (Li-1)2 ¥ S (S ˆ DEE),
results in replacement of the solvating DEE molecule by one molecule of THF. In pure
(D8)THF, Li-1 persists as an equilibrium between the dimer (Li-1)2 ¥ S (S ˆ THF) and
the monomer Li-1 ¥ 2 S (S ˆ THF). The disolvated monomer is the major aggregate in
THF.

3816
Helvetica Chimica Acta Vol. 85 (2002)
Addition of BuLi to the dimer (Li-1)2 in DEE leads to formation of a mixed
complex between the chiral lithium amide and BuLi, i.e., Li-1/BuLi [11] (Scheme 3).
The equilibrium between homo dimeric lithium amide, tetrameric BuLi, and the mixed
lithium amide/BuLi complex depends on the reagents involved as well as on the
solvent. Detailed knowledge about the factors influencing this equilibrium is pivotal,
since mixed complexes of this kind are efficient reagents for the enantioselective
addition of alkyllithiums to aldehydes, providing selectivities up to 99% enantiomeric
excess (ee) [12].
Scheme 3. Mixed Complex Formation between Li-1 and BuLi. The dimer (Li-1)2 is in equilibrium with a mixed
dimer Li-1/BuLi when excess BuLi is present. The mixed dimer undergoes a regioselective ortho-lithiation upon
standing at room temperature for 3 5 h. The dilithiated species thus formed dimerizes to the structure shown in
the insert.
Interestingly, when a DEE solution of the mixed complex Li-1/BuLi is kept at room
temperature for 3 5 h, one of the ortho-H-atoms on the methylbenzylamine moiety of
the molecule is regiospecifically abstracted. Remarkably, this ortho-lithiation proceeds
without addition of any specific Li-coordinating ligand, such as, for example, N,N,N',N'-
tetramethylethylenediamine (TMEDA). Based on multinuclear and multidimensional
NMR experiments, including 6Li,1H heteronuclear Overhauser effect spectroscopy, we
proposed that the resulting dilithiated product form a homo dimer, i.e., (Li2-1)2 in the
insert of Scheme 3.
The objective of the present study is to shed some light on this intriguing ortho-
lithiation reaction with a combination of kinetic and computational methods.
Kinetic Study. Our previous studies have established a fast exchange between
homo complexed BuLi (in the form of (BuLi)4), the dimer (Li-1)2, and the mixed Li-1/
BuLi complex. This exchange is fast also at low temperatures (À 908). Thus, the ortho-
lithiation reaction must consist of at least two consecutive reaction steps. The first step
involves a pre-equilibrium between (Li-1)2, BuLi, and the mixed complex Li-1/BuLi.
This pre-equilibrium is then followed by the rate-determining ortho-lithiation reaction,
suggested to proceed within the aggregate (Scheme 4). Subsequently, the ortho-

Helvetica Chimica Acta Vol. 85 (2002)
3817
Scheme 4. Schematic Representation of the ortho-Lithiation Reaction. The reaction is proposed to proceed
through a pre-equilibrium between the lithium amide and BuLi, followed by an intra-aggregate rate-limiting
proton transfer.
lithiated lithium amide (Li2-1) undergoes a rapid dimerization to yield the dimer (Li2-
1)2 (Scheme 3), previously studied by NMR.
Knowing that the pre-equilibrium is much faster than the lithiation reaction, i.e.,
kÀ1 ꢀ k2, and by keeping [BuLi] ꢀ [Li2-1] we may formulate the pseudo-first-order
rate law in Eqn. 1.
À
Á
d‰Li2-1Š
ˆ kobs ‰Li2-1Š1À‰Li2 À 1Š
(1)
dt
Where kobs is given by Eqn. 2:
k2K‰BuLiŠ
kobs
ˆ
(2)
ꢁ1 ‡ K‰BuLiŠ†
Kinetic data were obtained by monitoring the ortho-lithiation reaction of the amine
([Li-1] ˆ 50 mm) with BuLi (1200 mm) in (D10)DEE at 208 directly in the NMR tube,
as described in the Exper. Part. A straight line is obtained by plotting ln([Li21]1 À
[Li21]) vs. time (Fig. 1). Thus, apparently the reaction may be described as a pseudo-
first-order process with kobs ˆ 5.6  10À4 sÀ1.
The pre-equilibrium constant for this reaction has previously been determined as
1.22m [12a]. Thus, the first-order rate constant for the rate-determining step k2 may be
derived from Eqn. 2, k2 ˆ 9.4  10À4 sÀ1. This, in turn, corresponds to a free energy of
activation of DG=273K ˆ 21 kcal molÀ1.
The ortho-lithiation was also performed with different amounts of BuLi, while the
concentration of lithium amide was kept constant at 50 mm. The pseudo-first-order rate
constants kobs thus obtained was plotted against [BuLi] (80, 700, and 1200 mm, Fig. 2).
The plot shows only a weak dependence on kobs vs. [BuLi], indicating that the reaction
may follow saturation kinetics; otherwise, we would expect a linear dependence of kobs
vs. [BuLi] with the intercept close to zero, and not around 2.4 as observed. At even
higher BuLi concentrations, we expect the observed rate constant to reach a maximum
value for the situation, when all lithium amide is complexed with BuLi.
The kinetic data presented indicate that the stoichiometry of the TS is composed of
only the mixed aggregate BuLi/(Li-1); this suggests an intra-aggregate mechanism for
the ortho-lithiation. To gain further insight into the details of this mechanism, we
turned to a computational approach.
Computational Study. Several high-level computational studies have addressed
the mechanism and origin of selectivity in the ortho-lithiation of heteroatom-containing
aromatic rings [7]. Despite these efforts, a detailed understanding of the different
factors, e.g., precomplexation and inductive effects, and their relative importance for

3818
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 1. Plot of ln([Li2-1]1-[Li2-1]) vs. time. The straight line in this plot shows that the ortho-lithiation follows a
pseudo-first-order process with kobs ˆ 5.6  10À4 sÀ1
.
reactivity and selectivity remains undisclosed. Thus, instead of adding to this debate
with our quite special system, we chose to settle for a simpler theoretical model, i.e., the
semi-empirical PM3 method. This method is known to give reliable structural data for
lithium-organyl complexes [9b]; however, the energies obtained with this method
should be treated with care.
The result of our calculations is shown in Fig. 3. The initial complex between the
lithium amide Li-1 and EtLi (used as a model for BuLi) has a dimeric core where the
two Li-atoms are coordinated to the amide N-atom and to the carbanionic C-atom. One
of the Li-atoms is internally coordinated to the MeO group of the amide. In the initial
state, the distance between the carbanionic C-atom and the ortho-H-atom of the top Ph
ring (hypothetical reaction) is somewhat shorter than the distance to the H-atom being
abstracted experimentally (bottom Ph ring). Two transition states (TS), corresponding
to lithiation of the top and the bottom Ph rings, respectively, could be located4). The
4
)
Although the alkyllithium is the thermodynamically strongest base, another TS, involving initial proton
transfer to the N-atom, was also investigated. Proton transfers to more-electronegative atoms, e.g., N and
O, are known to be kinetically preferred to those involving less-electronegative atoms, i.e., C; thus a fast
intermediate −shuffling× of the abstracted proton via the N-atom could not be excluded from the kinetic
data. However, this TS was found to be significantly higher in energy (>20 kcal molÀ1 above the observed
TS); consequently, this possible pathway was excluded.

Helvetica Chimica Acta Vol. 85 (2002)
3819
Fig. 2. Plot of kobs vs. [BuLi]. The weak dependence of kobs on [BuLi] suggests that the reaction follows
saturation kinetics. ([Li-1] ˆ 50 mm and [BuLi] ˆ 80, 700, and 1200 mm).
energy for the TS leading to the experimentally observed product is somewhat lower
than that for the hypothetical reaction. The H-atom is transferred to the same extent,
somewhat more than halfway (CaÀH 1.46 ä and CoÀH 1.53 ä), in both transition
states. The stabilization of the negative charge developing on the ortho-C-atoms are
also similarly compensated for in the two TS as seen from the similar LiÀCo distances
(2.21 and 2.19 ä, resp.). The two ortho-lithiated products obtained from the two routes
are also shown in Fig. 3. Here, it can be seen that both Li-atoms stabilize the newly
formed carbanion in the experimentally observed product, while only one Li-atom is
coordinating to the carbanion in the product that would result from ortho-lithiation of
the top Ph ring. These structures are of minor importance, since the dilithiated amide is
known to undergo rapid dimerization.
Discussion. The small energy differences found between the two transition states,
and the similarities observed in their structures, do not account for the complete
regioselectivity observed. An alternative explanation for the high selectivity must be
invoked.
When solvation of the Li-atoms is included in the analysis, one initially expects two
DEE molecules to coordinate to the Li-atom lacking internal coordination (Li(2)), and
one DEE molecule to coordinate to the internally (MeO) coordinated lithium (Li(1)).
A simple inspection of the models suggest that the gain in enthalphy by coordination of
a DEE molecule to Li(1), will probably be low due to congestion; when also taking the
negative contribution from the entropy into account, it is highly likely that this process
will be unfavorable. This notion is supported by our previous NMR spectroscopic
studies, which suggested the presence of one tricoordinated- and one tetracoordinated

Fig. 3. PM3-Optimized Geometries. The structure of the initial mixed complex is shown to the left, the TS leading to the experimentally observed
product is shown in the center (bottom), while the TS for the hypothetical reaction that would lead to the alternative regioisomer is shown in the
center (top). The optimized geometries of the final states of the two paths are also shown; however, these complexes are known to rapidly form
heterodimers. Selected distances [ä] are shown in blue, and key distances in the TS are shown in red. Also shown are the relative energies [kcal
molÀ1] for the different species, and the calculated reaction enthalphies.

Helvetica Chimica Acta Vol. 85 (2002)
3821
Li-atom in the Li-1/BuLi complex in DEE [12a]5). Thus, Li(1) is expected to be
tricoordinated in the mixed complex.
Consequently, the TS leading to observed product is probably reasonably well
described in this calculation. On the other hand, the energy for TS involving lithiation
of the top Ph ring is expected to increase dramatically, if solvating ligands were included
in the calculation since the two DEE molecules at this Li-atom would make Li(2) less
capable of stabilizing the developing negative charge. This explanation is supported by
the good (almost too good) agreement between the experimentally observed activation
energy (DG=273K ˆ 21 kcal molÀ1) and the calculated activation barrier (20 kcal molÀ1).
The important role of the solvent in this reaction is further evidenced by the lack of
observable ortho-lithiation in THF, where both Li-atoms are known to be tetracoordi-
nated.
Conclusions. Kinetic investigations of the ortho-lithiation reaction in the mixed
Li-1/BuLi complex in DEE solution showed that the reaction proceeds within the
mixed complex. The rate constant for the lithiation step at 208 was found to be k2 ˆ
9.4  10À4 sÀ1, corresponding to DG273K ˆ 21 kcal molÀ1. With knowledge about the
≥
transition-state composition, we investigated the reaction computationally with the
semi-empirical PM3 method. Calculations in the absence of solvent molecules did not
give a completely satisfactory explanation for the regioselectivity observed. However,
when the solvation around the Li-atoms is included in the analysis, a more adequate
explanation emerges. Thus, in conclusion, it seems likely that the observed reactivity is
a direct consequence of the complex formation (no mixed complex formation no
observable ortho-lithiation); further, the presence of one coordinatively unsaturated
Li-atom in DEE solution appears to be a prerequisite for the observed reactivity and
selectivity.
Financial support from Vetenskapsradet The Swedish Research Council, is gratefully acknowledged.
Experimental Part
General. All glassware were dried overnight in an oven at 1208 (syringes were dried at 508 in a vacuum
oven) before transfer into a glove box (Mecaplex GB 80 equipped with a gas-purification system that removes
O2 and moisture) containing a N2 atmosphere. Typical moisture content was less than 2 ppm. All manipulations
of the Li compounds were carried out in the glove box with gas-tight syringes. Ethereal solvents were distilled
and stored over Deporex (Fluka) prior to use. All NMR spectra were recorded with a Varian Unity 400
spectrometer equipped with an inverse detection probe. Measuring frequency was 400 MHz (1H). The 1H-NMR
spectra were referenced to the solvent (D10)DEE signal at d 1.06 (1HÀMe). Probe temp. were measured after
more than 30 min of temp. equilibrium with the standard MeOH thermometer supplied by Varian instruments.
Computational Methods. Geometry and transition-state optimizations were done on a Silicon Graphics
Indy workstation by the semi-empirical PM3 method as implemented in the Spartan 5.0 program package [13].
Several initial geometries for the optimizations were tried in order to find the global minimum. The vibrational
frequencies were also calculated to verify that all optimized structures were either minima (no imaginary
frequency) or transition states (one imaginary frequency) on the potential energy-surface.
5
)
Addition of THF to a DEE solution of Li-1/BuLi initially leads to an increased solvation of one of the Li-
atoms (before all DEE molecules were replaced by THF). At that point, it was not clear which Li-atom
was tri- or tetracoordinated.

3822
Helvetica Chimica Acta Vol. 85 (2002)
(S)-N-[(R)-2-Methoxy-1-phenylethyl]-a-methylbenzylamine (1) and the [6Li]-Salt (Li-1). The preparation
and characterization of this amine, and the in situ generation of the corresponding [6Li]-containing lithium
amide have been reported in [11].
Kinetic Studies of the ortho-Lithiation. Amine 1 (10 mg, ca. 0.035 mmol) was transferred into a dry 5-mm
NMR tube. The NMR tube was fitted with a Wilmad/Omnifit Teflon valve assembly OFV with Teflon/silicon
septum. The solvent (D10)DEE (750 ml) was added via syringe. To the resulting 52 mm soln. of 1 at À208 was
added 10m [6Li]BuLi (ca. 30 ml, 0.37 mmol) with a syringe. The mixture containing Li-1 (96% in 6Li atom) was
then put into the temp.-equilibrated NMR probe. The progress of the reaction was followed by obtaining
1H-NMR spectra every 4 min. The disappearing 1H-NMR signals for the a-H-atom of BuLi and the appearing
signals from the MeO group of the ortho-lithiated amide were integrated and used in the kinetic analysis.
Typically, rate measurements were obtained by following the reactions up to ca. 25% completion.
REFERENCES
[1] D. Seebach, R. H‰ssig, J. Gabriel, Helv. Chim. Acta 1983, 66, 308.
[2] a) G. W. Klumpp, Recl. Trav. Chim. Pays-Bas 1986, 105, 1; b) D. Seebach, Angew. Chem., Int. Ed. 1988, 27,
1624; c) K. Gregory, P. v. R. Schleyer, R. Snaith, Adv. Inorg. Chem. 1991, 37, 47; d) P. G. Williard, in
−Comprehensive Organic Synthesis×, Vol. 1, Eds. B. M. Trost, I. Fleming, Pergamon Press, Oxford, 1991,
p. 1; e) D. B. Collum, Acc. Chem. Res. 1993, 26, 227; f) C. Lambert, P. v. R. Schleyer, Angew. Chem., Int. Ed.
1994, 33, 1129; g) A. Streitwieser, S. M. Bachrach, A. Dorigo, P. v. R. Schleyer, in −Lithium Chemistry A
Theoretical and Experimental Overview×, Eds. A.-M. Sapse, P. v. R. Schleyer, John Wiley & Sons, New
York, 1995, p. 1; h) B. L. Lucht, D. B. Collum, Acc. Chem. Res. 1999, 32, 1035; i) A. R. Katritzky, M. Piffl,
H. Y. Lang, E. Anders, Chem. Rev. 1999, 99, 665; j) G. Boche, J. C. W. Lohrenz, Chem. Rev. 2001, 101, 697;
k) C. Gaul, P. I. Arvidsson, W. Bauer, R. E. Gawley, D. Seebach, Chem. Eur. J. 2001, 7, 4117.
[3] a) H. W. Gschwend, H. R. Rodriguez, Org. React. 1979, 26, 1; b) V. Snieckus, Chem. Rev. 1990, 90, 879; c) F.
Mongin, R. Maggi, M. Schlosser, Chimia 1996, 50, 650; d) P. Beak, A. Basu, D. J. Gallagher, Y. S. Park, S.
Thayumanavan, Acc. Chem. Res. 1996, 29, 552; e) C. S. Salteris, I. D. Kostas, M. Micha-Screttas, G. A.
Heropoulos, C. A. Screttas, A. Terzis, J. Org. Chem. 1999, 64, 5589.
[4] a) E. Marzi, A. Bigi, M. Schlosser, Eur. J. Org. Chem. 2001, 7, 1371; b) M. Kauch, D. Hoppe, Can. J. Chem.
2001, 79, 1736; c) P. Garibay, P. Vedso, M. Begtrup, T. Hoeg-Jensen, J. Comb. Chem. 2001, 3, 332; d) G.
Amiet, H. M. Hugel, F. Nurlawis, Synlett 2002, 495; e) G. Karig, N. Thasana, T. Gallagher, Synlett 2002, 808;
f) P. Gros, S. Choppin, J. Mathieu, Y. Fort, J. Org. Chem. 2002, 67, 234.
[5] a) H. Gilman, R. L. Bebb, J. Am. Chem. Soc. 1939, 61, 109; b) G. Wittig, G. Fuhrmann, Chem. Ber. 1940, 73,
1197; c) J. D. Roberts, D. Y. Curtin, J. Am. Chem. Soc. 1946, 68, 1658.
[6] a) C. G. Screttas, J. F. Eastham, J. Am. Chem. Soc. 1965, 87, 3276; b) C. D. Broaddus, J. Org. Chem. 1970, 35,
¬
10; c) C. G. W. Bauer, P. v. R. Schleyer, J. Am. Chem. Soc. 1989, 111, 7191; d) J. M. Saa, G. Martorell, A.
Frontera, J. Org. Chem. 1996, 61, 5194; e) M. Stratakis, J. Org. Chem. 1997, 62, 3024; f) R. A. Rennels, A. J.
Maliakal, D. B. Collum, J. Am. Chem. Soc. 1998, 120, 421; g) M. Schlosser, F. Mongin, J. Porwisiak, W.
Dmowski, H. H. B¸ker, N. M. M. Nibbering, Chem. Eur. J. 1998, 4, 1281; h) R. Maggi, M. Schlosser,
Tetrahedron Lett. 1999, 40, 8797; i) D. W. Slocum, P. Dietzel, Tetrahedron Lett. 1999, 40, 1823; j) D. R.
Anderson, N. C. Faibish, P. Beak, J. Am. Chem. Soc. 1999, 121, 7553.
[7] a) N. J. R. van Eikema Hommes, P. v. R. Schleyer, Tetrahedron 1994, 50, 5903; b) S. T. Chadwick, R. A.
¬
Rennels, J. L. Rutherford, D. B. Collum, J. Am. Chem. Soc. 2000, 122, 8640; c) J. M. Saa, Helv. Chim. Acta
2002, 85, 814.
[8] P. Beak, A. I. Meyers, Acc. Chem. Res. 1986, 19, 356.
[9] a) G. Hilmersson, P. I. Arvidsson, ÷. Davidsson, M. Hakansson, Organometallics 1997, 15, 3352; b) G.
Hilmersson, P. I. Arvidsson, ÷. Davidsson, M. Hakansson, J. Am. Chem. Soc. 1998, 120, 8143.
[10] G. Hilmersson, P. Ahlberg, ÷. Davidsson, J. Am. Chem. Soc. 1996, 118, 3539.
[11] a) G. Hilmersson, ÷. Davidsson, J. Org. Chem. 1995, 60, 7660; b) G. Hilmersson, ÷. Davidsson, J.
Organomet. Chem. 1995, 489, 175; c) G. Hilmersson, ÷. Davidsson, Organometallics 1995, 14, 912.
[12] a) P. I. Arvidsson, G. Hilmersson, ÷. Davidsson, Chem. Eur. J. 1999, 5, 2348; b) P. I. Arvidsson, ÷.
Davidsson, G. Hilmersson, Tetrahedron: Asymmetry 1999, 10, 527; and refs. cit. therein.
[13] SPARTAN, Version 5.0.1, Wavefunction, Inc., Irvine, CA, USA, 1995.
Received June 7, 2002