C. Strohmann et al.
tically favoured due to the transition state trans-S-TS2 and
trans-S-TS3.
Experimental Section
All experiments were carried out under a dry, oxygen-free argon atmos-
phere by using standard Schlenk techniques. All reagents were manipu-
lated under a dry argon atmosphere. Solvents were dried over sodium
and distilled prior to use. TMEDA, phenyllithium and n-butyllithium
were obtained from Aldrich. The lithium reagents were titrated prior use
against diphenylacetic acid. (E)-benzyl(2-propenylphenyl)amine was pre-
pared by high-yield Suzuki–Miyaura cross-coupling of benzyl-(2-bromo-
phenyl)amine with (E)-propenyl boronic acid.[9a]
It is important to recognise that in such fixed systems
small variations in the bond length between the calculated
structures and the actual structures of the intermediates re-
sults in inaccuracies of the description of the repulsion ef-
fects and thus also in the energy of the system. Additional
entropy effects may also influence the energies of the sys-
tems. However, for polar compounds entropy is crucially in-
fluenced by solvent effects. A remarkable feature of all tran-
sition states is that these structures are rather crowded. All
moieties (benzyl, phenyl, methyl and ethyl) can only arrange
in one specific position and orientation to prevent repulsion
with adjacent groups. This is particularly interesting with re-
spect to the decrease in selectivity observed when the corre-
sponding methyl amide was utilised instead of benzyl amide
9 in experiment.[9c] As is evident from the amide structures 9
and the corresponding transition states, the benzyl group
seems to be necessary to fix the coordinating (À)-sparteine
in one specific arrangement relative to the amide group.
This arrangement then controls the approach of the alkyl
lithium and thus the selectivity of the reaction. Overall the
proposed mechanism via an analogous intermediate to the
isolated amide 8 seems also reasonable from the theoretical
point of view. With low reaction barriers and the energetic
preference for the S enantiomer, the calculations are in
good agreement with experimental observations.
Preparation of crystals of 4, 7 and 8: (E)-Benzyl(2-propenylphenyl)amine
(4): (E)-benzyl(2-propenylphenyl)amine (4, 100 mg, 0.45 mmol) was dis-
solved in 1 mL of n-pentane and cooled to À788C, giving the amine as
colourless needles. Lithium amide 7: 4 (100 mg, 0.45 mmol) was dissolved
in 1 mL of n-pentane and toluene. The mixture was cooled to À808C and
PhLi (0.27 mL, 0.50 mmol, 1.85m in di-n-butyl ether) was carefully
added. After warming to room temperature the solution was again
cooled to À788C, giving yellowish crystals of the dimeric compound after
3 days. Lithium amide 8: 4 (100 mg, 0.45 mmol) and TMEDA (0.60 mg,
0.52 mmol) were added to 1 mL of n-pentane and toluene until complete
dissolution. The mixture was cooled to À708C and nBuLi (0.33 mL,
0.48 mmol, 1.45m in hexane) was carefully added. The formed precipitate
was dissolved by warming to room temperature and the solution again
cooled to À788C, giving yellowish needles of the monomeric compound
after 12 h.[19]
X-ray crystal structure determination: Data collection of all compounds
was conducted with a Bruker APEX-CCD (D8 three-circle goniometer,
Bruker AXS), cell determination and refinement with Smart version
5.622 (Bruker AXS, 2001), integration with SaintPlus version 6.02
(Bruker AXS, 1999) and empirical absorption correction with SADABS
version 2.01 (Bruker AXS, 1999). The crystals of all three compounds
were mounted in an inert oil (perfluoro polyalkyl ether) at À608C (N2
stream), by using the X-TEMP 2 device.[20] Data were collected at
À1008C with MoKa radiation (l=0.71073 ꢃ). The structures were solved
by applying direct and Fourier methods with SHELXS-90 (G. M. Shel-
drick, University of Gçttingen, 1990) and SHELXL-97 (G. M. Sheldrick,
University of Gçttingen, 1997). CCDC 742502 (4), 742503 (7) and 742504
(8) contain the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystallograph-
further information about the data collection and structure refinement of
compounds 4 and 8. Further information about the X-ray diffraction
analysis of the dimeric lithium amide 7 is available in the Supporting In-
formation.
Conclusion
We have reported mechanistic studies on the asymmetric
carbolithiation of simple b-methylstyrene and an amino-
functionalised derivative. Both addition reactions show suffi-
ciently low reaction barriers to be overcome at low tempera-
tures. The calculations of the non-functionalised system
show that the high selectivities are due to repulsion effects
in the relevant transition states upon approximation of the
alkyl lithium·(À)-sparteine adduct. In this respect, the inter-
action between the phenyl ring and the chiral ligand com-
petes with interactions between the alkyl lithium and the
methyl group. As a result smaller energetic differences be-
tween the diastereomeric transition states are observed for
isopropyllithium compared to ethyllithium.
Structural analyses of intermediate lithium amides of
ortho-amino-functionalised styrene 4 show molecular struc-
tures with one side of the double bond shielded by the
ligand-coordinated amide group. This indicates a mechanism
of carbolithiation in which the alkyl lithium·(À)-sparteine
adduct approaches from the more open side of the double
bond. This process is supported by computational studies,
which also point to the crucial role of the substituents in de-
termining the specific arrangement of the transition states.
Computational details: All calculations were performed without symme-
try restrictions. Starting coordinates were obtained with Chem3DUltra
10.0. Optimisation and additional harmonic vibrational frequency analy-
ses (to establish the nature of stationary points on the potential energy
surface) were performed with the software package Gaussian 03 (Revi-
sion D.01 and Revision E.01) at the same level.[11] The total (SCF) and
zero-point energies (ZPE) and the coordinates of all systems are avail-
able in the Supporting Information. The global minima and vibrational
frequency analyses were performed at the B3LYP/6-31+G(d), M052X/6-
31+G(d) and M052X/6-31G levels. The vibrational frequency analyses
showed imaginary frequencies for the transition states representing the
corresponding vibration for the deprotonation. For the reactants no
imaginary frequencies were obtained. For polar compounds entropy is
crucially influenced by solvent effects. Additionally, calculated Gibbs free
energies seem in such large systems to be less reliable due to very low
frequencies, for which the harmonic oscillator model produces significant
deviations.[21] Thus, enthalpy values are discussed. Corrections for basis
set superposition errors (BSSE) are not included.
To validate the encountered theoretical methods the aggregation of
methyllithium, tert-butyllithium, lithium chloride and the a-lithiation of
TMEDA were studied by applying the MP2, G2 and G3 method as refer-
ence, respectively. Furthermore, calculations on the deaggregation of
methyllithium with different ligands were performed to explore the effect
of changes in the coordination number of lithium. These calculations (see
3002
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2996 – 3004