Kinetics and mechanism of the ( - )-Sparteine-mediated deprotonation of (E)-N-Boc-N-(p-methoxyphenyl)-3-cyclohexylallylamine

Article abstract of DOI:10.1021/ja001955k

The ( - )-sparteine-mediated asymmetric lithiation-substitution of (E)-N-Boc-N-(p-methoxyphenyl)3-cyclohexylallylamine ((E)-5) to afford γ-substituted enantiomerically enriched products 6 is reported. The solution structure for the lithiated intermediate

Full text of DOI:10.1021/ja001955k

J. Am. Chem. Soc. 2001, 123, 4919-4927
4919
Kinetics and Mechanism of the (-)-Sparteine-Mediated Deprotonation
of (E)-N-Boc-N-(p-methoxyphenyl)-3-cyclohexylallylamine
Daniel J. Pippel, Gerald A. Weisenburger, Neil C. Faibish, and Peter Beak*
Contribution from the Department of Chemistry, UniVersity of Illinois at UrbanasChampaign,
Urbana, Illinois 61801
ReceiVed June 2, 2000
Abstract: The (-)-sparteine-mediated asymmetric lithiation-substitution of (E)-N-Boc-N-(p-methoxyphenyl)-
3-cyclohexylallylamine ((E)-5) to afford γ-substituted enantiomerically enriched products 6 is reported. The
solution structure for the lithiated intermediate 8‚1 in these reactions was determined by heteronuclear NMR
to be a configurationally stable, R-lithio, η1-coordinated monomer. This intermediate is proposed to exist as
two rotamers that are rapidly equilibrating on the NMR time scale; competitive electrophilic substitution of
each conformation results in the formation of Z or E products. Kinetic measurements of the lithiation by in
situ infrared spectroscopy provide pseudo-first-order rate constants for reactions with a variety of concentrations
of amine, (-)-sparteine, and n-BuLi. The reaction is first order in amine and zero order in 1:1 base-ligand
complex. When the concentration of n-BuLi is varied independently of (-)-sparteine concentration, the reaction
rate exhibits an inverse dependence on n-BuLi concentration. The deuterium isotope effect for the reaction
was determined to be 86 at -75 °C, a result consistent with C-H bond breaking in the rate-determining step
and indicative of tunneling. A reaction pathway involving a prelithiation complex is supported by kinetic
simulations.
Introduction
of the asymmetric lithiation-substitution sequence of an N-Boc
allylamine derivative for a (-)-sparteine-mediated deprotonation
reaction.
Organolithium compounds are essential reactants or reagents
in a majority of contemporary organic syntheses.1,2 Yet, reaction
sequences that proceed via organolithium intermediates have
been among the more difficult to characterize. Uncertainties
related to aggregation states and the structures of the reactive
species are problems that have to be resolved to unravel the
course of most organolithium reactions.3-8
In recent years, the applications of organolithium reagents
have been extended by their applications to asymmetric
syntheses. For these newly developed reactions, the stereo-
chemical-determining steps must be delineated, and questions
of absolute configuration, configurational stability, and stereo-
chemical course of reaction become important.9-14 In this report,
we provide an investigation into the kinetics and mechanism
Asymmetric lithiation-substitution sequences that provide
homoenolate synthetic equivalents via dipole-stabilized carban-
ions have been reported for an oxygen-based system in seminal
work by Hoppe, and for a nitrogen system from our labora-
tories.15-17 In our work, we have characterized the reaction
pathways for the highly enantioselective (-)-sparteine (1)-
mediated lithiation-substitution reactions of the cinnamylamine
2 to give enecarbamates 4. The enantiodetermining step of
(1) Sapse, A.-M.; Schleyer, P. v. R. Lithium Chemistry: A Theoretical
and Experimental OVerView; John Wiley & Sons: New York, 1995.
(2) Wakefield, B. J. The Chemistry of Organolithium Compounds;
Pergamon Press: New York, 1974.
(3) Recent leading work that focuses on the solution4 and solid-state5
characterizations of intermediates, determination of thermodynamic equi-
libria,6 the establishment of kinetic parameters for reactions,7 and the
computational determination of structures and transition states8 is exemplary.
(4) Thompson, A.; Corley, E. G.; Huntington, M. F.; Grabowski, E. J.
J.; Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1998, 120, 2028-
2038.
(5) Hoppe, I.; Marsch, M.; Harms, K.; Boche, G.; Hoppe, D. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2158-2160.
(6) Streitwieser, A.; Wang, D. Z.-R. J. Am. Chem. Soc. 1999, 1999,
6213-6219.
asymmetric deprotonation, the stereochemical course of sub-
stitution reactions, and the solid-state and solution structures
(7) Sun, X.; Collum, D. B. J. Am. Chem. Soc. 2000, 122, 2452-2458.
(8) Wu¨rthwein, E.-U.; Behrens, K.; Hoppe, D. Chem. Eur. J. 1999, 5,
3459-3463.
(9) Hoppe, D.; Hense, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 2282-
2316.
(10) Beak, P.; Basu, A.; Gallagher, D. J.; Park, Y. S.; Thayumanavan,
S. Acc. Chem. Res. 1996, 29, 552-560.
(11) Zschage, O.; Schwark, J.-R.; Kramer, T.; Hoppe, D. Tetrahedron
1992, 48, 8377-8388.
(12) Pearson, W. H.; Lindbeck, A. C.; Kampf, J. W. J. Am. Chem. Soc.
1993, 115, 2622-2636.
(13) O’Brien, P.; Powell, H. R.; Raithby, P. R.; Warren, S. J. Chem.
Soc., Perkin Trans. 1 1997, 1031-1039.
(14) Hoffmann, R. W.; Koverstein, R.; Harms, K. J. Chem Soc., Perkin
Trans. 2 1999, 183-191.
10.1021/ja001955k CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/04/2001

4920 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Pippel et al.
of the lithiated intermediate have been reported.17-19 The
sequence via 3‚1 is of particular synthetic value as the
employment of Michael acceptors as electrophiles provides 1,4-
addition adducts with control over two newly generated ste-
reogenic centers.20,21 Despite the well-developed synthetic,
mechanistic, and structural aspects of this transformation, our
efforts to obtain a complete kinetic profile and rate law for the
metalation of 2 have been hampered by the high forward
velocity of the reaction.22
To determine whether the asymmetric induction for these
reactions occurred in the substitution portion of the sequence,
the racemic stannane (E)-rac-7 was transmetalated with n-BuLi/
1, and the resulting organolithium species was allowed to react
with methyl iodide. This experiment provided the enecarbamates
(Z)-6 and (E)-6 in 46% overall yield with no enantiomeric
We now report that the lithiation-substitution of (E)-N-Boc-
N-(p-methoxyphenyl)-3-cyclohexylallylamine ((E)-5) with n-BuLi/
(-)-sparteine (1) provides products with significant enantiomeric
enrichments. For this reaction, we have established the enan-
tiodetermining step and the diastereodetermining step, character-
ized the structure of the lithiated intermediate in solution, and
determined the kinetics of the lithiation step.
Results and Discussion
Deprotonation of (E)-5 with n-BuLi/1 at -78 °C in toluene
followed by addition of 3 equiv of methyl iodide resulted in a
2:1 ratio of the γ-substituted products (Z,R)-6 and (E,S)-6. The
enrichment. The result shows that asymmetric substitution is
not enantiodetermining and suggests that the enantiodetermining
step in the formation of both the Z and E enecarbamates from
(E)-5 is an asymmetric deprotonation that gives rise to con-
figurationally stable intermediates.
The results from transmetalation-electrophilic substitution
of enantiomerically enriched (E,S)-7 also support the assignment
of configurational stability to the intermediates in this sequence.
In this case, the enecarbamates 6 are formed in 52% overall
yield with enantiomeric enrichment of 84:16. The configuration
at the newly formed stereogenic center of both products is
opposite to that obtained from the normal lithiation-substitution
sequence. This result in conjunction with literature precedent
suggests a stereochemical course for the reaction. Reactions of
8‚1 with methyl iodide and trimethyltin chloride proceed
antarafacially, while tin-lithium exchange occurs with retention
of configuration.18
diastereomers are separable chromatographically, and the se-
quence thus provides access to either configuration at the
γ-position. The absolute configurations of the products (Z,R)-6
and (E,S)-6 are assigned by analogy to those determined
definitively for products arising from lithiation-substitution of
(E)-2.17
NMR Structure of the Lithiated Intermediate. Structure
determination for the key intermediate(s) in this sequence was
carried out and 8‚1 obtained by lithiation of (E)-5, which was
The high γ-selectivity observed upon substitution with methyl
iodide was not observed for the analogous reaction with
trimethyltin chloride as the electrophile. In that case, the
R-substituted product (E,S)-7 was obtained in 79% yield with
an enantiomeric ratio of 94:6.
13C-labeled at both allylic termini, with 0.9 equiv of n-Bu6Li/1
6
in toluene-d8 at -78 °C. The Li NMR and 13C NMR spectra
of (E)-8‚1 are shown in Figure 1a and b.
The 6Li NMR has three signals centered at 1.17 ppm,
consistent with two partially overlapping doublets in an 87:13
ratio with J(6Li,13C) ) 4.2 Hz.23 The spectrum is interpreted
1
(15) For a general review on chiral homoenolate synthetic equivalents,
see:Ahlbrecht, H.; Beyer, U. Synthesis 1999, 365-390.
(16) Behrens, K.; Fro¨hlich, R.; Meyer, O.; Hoppe, D. Eur. J. Org. Chem.
1998, 1998, 2397-2403.
to indicate the presence of two monomeric η1 complexes, (E,R)-
6
8‚1 and (E,S)-8‚1, in which each Li is bound to only one 13C
atom. The absolute configurations for the major and minor
(17) Weisenburger, G. A.; Faibish, N. C.; Pippel, D. J.; Beak, P. J. Am.
Chem. Soc. 1999, 121, 9522-9530.
(21) Park, Y. S.; Weisenburger, G. A.; Beak, P. J. Am. Chem. Soc. 1997,
119, 10537-10538.
(22) The lithiation of (E)-2 was demonstrated to be a very rapid process
requiring less than 30 s. See Supporting Information section XIX for details.
(23) Two additional weak absorptions present at 0.71 and 0.72 ppm are
unassigned.
(18) Pippel, D. J.; Weisenburger, G. A.; Wilson, S. R.; Beak, P. Angew.
Chem., Int. Ed. Engl. 1998, 37, 2522-2524.
(19) Weisenburger, G. A.; Beak, P. J. Am. Chem. Soc. 1996, 118, 12218-
12219.
(20) Curtis, M. D.; Beak, P. J. Org. Chem. 1999, 64, 2996-2997.

Deprotonation of Organolithium Compounds
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4921
Figure 2. Reaction profile for dynamic, kinetically controlled reaction
of syn-anti-(E,R)-8‚1 and syn-syn-(E,R)-8‚1 at -78 °C.
Bu6Li/1 and the unlabeled R-stannane (E)-rac-7. The 6Li NMR
spectrum contains two singlets at 1.22 and 1.11 ppm in a 47:53
ratio as shown in Figure 1c.27 Thus, the diastereomeric ratio
(E,R)-8‚1 to (E,S)-8‚1 and ultimately the enantiomeric ratios
of products 6 are not controlled by an equilibration process,
but rather result from the mode of formation, consistent with
the assignment of configurational stability to these species.
Origin of the Z/E Ratio. To explain the Z/E ratio in the
reaction with methyl iodide, we hypothesize that the two
rotamers syn-anti-(E,R)-8‚1 and syn-syn-(E,R)-8‚1 give rise to
the diastereomeric products (Z,R)-6 and (E,S)-6, respectively.
Because only a single major diastereomeric intermediate is
detectable in the NMR, syn-anti-(E,R)-8‚1 and syn-syn-(E,R)-
8‚1 must rapidly equilibrate at -78 °C on the NMR time scale.
To compare the rate of equilibration by rotation to the rate of
reaction with electrophile, a modified version of Hoffmann’s
test for configurational stability was performed.10,28,29 Depro-
tonation of (E)-5 with n-BuLi/1 at -78 °C in toluene followed
6
Figure 1. (a) 73.6-MHz Li and (b) 125.7-MHz 13C NMR spectra of
(E,R)-8‚1 (6Li and 13C-labeled) in toluene-d8 (0.10 M) at -78 °C. (c)
6
73.6-MHz Li spectrum of (E)-rac-8‚1 (6Li labeled, no 13C labels) in
toluene (0.10 M) at -78 °C.
diastereomers are based on analogy to the absolute configuration
of 3‚1, which was determined through X-ray crystallography.18
In the corresponding 13C NMR spectrum four total resonances
are present, two at 97.3 and 96.5 ppm in a 15:85 ratio and two
at 69.6 and 69.4 ppm in a 13:87 ratio. These signals are ascribed
to the four allylic terminal carbons of (E,R)-8‚1 and (E,S)-8‚1.
The major absorptions at 96.5 and 69.4 ppm have a half-height
line width of 10.6 and 16.3 Hz, respectively. The broadened
signal at 69.4 ppm is ascribed to 6Li,13C coupling and is assigned
6
to the carbon bearing the Li atom.24 On the basis of similar
by addition of 0.10 equiv of methyl iodide resulted in a 2:1
Z/E product ratio. This ratio is identical to the result obtained
with an excess of methyl iodide. These results are consistent
with a dynamic, kinetically controlled process being the pathway
that leads to the observed Z/E product ratio, as illustrated by
the energy diagram shown in Figure 2.30,31
reasoning, the resonance at 69.6 ppm for the minor diastereo-
meric pair is assigned to the carbon bearing the 6Li atom. Since
the 6Li NMR spectrum indicates that the organolithium species
are monomeric, the 13C absorptions at 69.6 and 69.4 ppm should
be 1:1:1 triplets; however, the 6Li-13C coupling was not
resolved, presumably due to the greater natural line width.25,26
To corroborate these assignments, an equimolar mixture of
(E,R)-8‚1 and (E,S)-8‚1 was obtained from 1.0 equiv of n-
For reactions of the η3 lithiated intermediate 3‚1, Z/E product
ratios were typically found to be >40:1. The ratio was
determined to be dependent on conversion, indicating that the
intermediates leading to Z and E products were not rapidly
equilibrating.17 It is not surprising that rotation about the C-1,
C-2 bond of η1 lithiated intermediate 8‚1 is more facile than in
the η3 lithiated intermediate 3‚1.
(24) To establish that the resonance at 69.4 ppm was properly assigned
to the carbon R to nitrogen, a monolabeled variant of (E)-5, 13C-labeled R
to nitrogen, was prepared, deprotonated, and observed spectroscopically.
See Supporting Information section V for details.

4922 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Pippel et al.
Thus, the present work shows that at -78 °C (E)-5 undergoes
an asymmetric deprotonation to afford a kinetic ratio of the
configurationally stable diastereomeric complexes (E,R)-8‚1 and
(E,S)-8‚1. Highly stereoselective electrophilic substitution pro-
vides the enecarbamate product as a mixture of the Z and E
isomers in a ratio that is established by a dynamic, kinetically
controlled reaction of the rotamers syn-anti-(E,R)-8‚1 and syn-
syn-(E,R)-8‚1.
In Situ Infrared Spectroscopy. The course of the lithiation
of (E)-5 with n-BuLi/1 can be followed by in situ infrared
spectroscopy.7,32-35 The spectra for a sample reaction in which
1.2 equiv of n-BuLi was added to a premixed solution of 1
equiv of (E)-5 and 1.2 equiv of (-)-sparteine in toluene at -73
°C are shown in Figure 3. The spectral baseline was reset to
zero immediately prior to addition of n-BuLi, and 570 spectra
were collected over a period of 600 s.
As shown in Figure 3, the absorbance associated with (E)-5
at 1695 cm-1 rapidly decreases upon addition of n-BuLi as (E)-5
is consumed. At a slower rate, an absorbance associated with
8‚1 at 1640 cm-1 becomes increasingly positive.36 A third
absorbance at 1675 cm-1 is of particular interest. As shown in
Figure 3a, this peak grows initially upon addition of base, but
then slowly disappears as the reaction progresses. This absor-
bance is attributed to transient formation of a prelithiation
complex, or complexes, 9.37,38 Our long-standing interest in this
Figure 3. Progression of the (-)-sparteine-mediated stoichiometric
reaction between (E)-5 and n-BuLi in toluene as monitored by in situ
infrared spectroscopy. (a) Three-dimensional plot of absorbance versus
wavenumber versus time. (b) Several sequential two-dimensional
infrared spectra taken at various times of the reaction.
tions of n-BuLi and 1 (0.66 M), were maintained for reactions
executed over a 5-fold difference in initial allylic amine (E)-5
concentrations (0.0236-0.005 06 M). For these reactions, (E)-5
(30) It is also conceivable that the rotamers do not equilibrate on the
time scale of reaction with electrophile and that the activation energies for
reactions with syn-anti-(E,R)-8‚1 and syn-syn-(E,R)-8‚1 are experimentally
indistinguishable. In this case, the relative thermodynamic stabilities of the
two rotamers would establish the ultimate Z/E ratio. Because the rotation
is believed to be rapid on the NMR time scale and the reaction with
electrophile is known to take in excess of 30 min, we assume that this
possibility is unlikely.
(31) The ground-state energies of syn-anti-(E,R)-8‚1 and syn-syn-(E,R)-
8‚1 are arbitrarily set as equal. Their relative values are, in fact, unknown.
(32) Al-Aseer, M.; Beak, P.; Hay, D.; Kempf, D. J.; Mills, S.; Smith, S.
G. J. Am. Chem. Soc. 1983, 105, 2080-2082.
(33) Hay, D. R.; Song, Z.; Smith, S. G.; Beak, P. J. Am. Chem. Soc.
1988, 110, 8145-8153.
(34) Sun, X.; Kenkre, S. L.; Remenar, J. F.; Gilchrist, J. H.; Collum, D.
B. J. Am. Chem. Soc. 1997, 119, 4765-4766.
(35) All in situ infrared spectra were collected by means of an ASI
ReactIR 1000 equipped with a 1.0-in. DiComp ATR probe (range 4400-
2150, 1950-650 cm-1). ASI Applied Systems, Inc., Millersville, MD 21108.
See Supporting Information section I.E. for additional details.
(36) The lithiation of N, N-dimethyl-2,4,6-triisopropylbenzamide resulted
in a shift from 16501 to 1588 cm-1 for the carbonyl stretching frequency.33
(37) Contrary to expectation, the species responsible for the absorbance
at 1675 cm-1 does not form instantly. Instead, after addition of base, the
peak can be seen growing to maximum intensity over a period of several
seconds. Because base-carbonyl complexation in the case of N,N-dimethyl-
2,4,6-triisopropylbenzamide has previously been demonstrated by stopped-
flow infrared spectroscopy to require less than 3 ms, we attribute the
observation in our case to poor mixing.33 For data that confirms the
sluggishness of macromixing for the React-IR instrument as configured,
see Supporting Information section XII.
(38) Addition of electrophile to the reaction mixture while the concentra-
tion of species 9 remained high resulted in the recovery of a significant
amount of starting material. We interpret this result to indicate that 9 is, in
fact, a prelithiation complex or complexes and not an isomer of a lithiated
intermediate. See Supporting Information section XIV for details.
(39) Beak, P.; Meyers, A. I. Acc. Chem. Res. 1986, 19, 356-363.
type of species prompted further investigation in attempt to
determine whether the spectroscopically visible complex(es) 9
lie along the reaction pathway.39
To determine reaction order in allylic amine, pseudo-first-
order conditions were employed. High and constant concentra-
(25) In an attempt to increase the likelihood of observing coupling in
the 13C NMR spectrum, an additional experiment was performed using
7
unlabeled n-BuLi/1. However, Li,13C coupling was not resolved.
(26) The apparent presence of a single major intermediate on the NMR
time scale is not inconsistent with the formation of the two products (Z,R)-6
and (E,S)-6 in a 2:1 ratio, as discussed in the next section.
(27) Additional weak absorptions at 0.60 and 1.29 ppm are unassigned.
(28) Hoffmann, R. W.; Ru¨hl, T.; Harbach, J. Liebigs Ann. Chem. 1992,
725-730.
(29) Hirsch, R.; Hoffmann, R. W. Chem. Ber. 1992, 125, 975-982.

Deprotonation of Organolithium Compounds
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4923
Table 2. Pseudo-First-Order Rate Constants and R2 Values for
Reactions Representing a 39:1 to 19:1 Ratio of Base/Ligand
Complex:Carbamate
base/ligand to (E)-5
kobsd (s-1
)
R2 (%)
39:1
34:1
29:1
24:1
19:1
0.0025
0.0030
0.0028
0.0027
0.0021
98.4
98.8
99. 1
97.5
99.5
The results of this experiment imply that the reaction under
pseudo-first-order-conditions gives rise to the same intermediate,
8‚1, as the stoichiometric sequence.44
To test for the dependence on the concentration of the base/
ligand complex, a series of reactions in which the concentration
of (E)-5 was held constant at 0.017 M while the concentration
of 1:1 n-BuLi/(-)-sparteine base/ligand complex was varied
from 0.66 to 0.33 M were executed. From these reactions, it is
apparent that the rate constant, kobsd, changes neither significantly
nor systematically as the ratio of base/ligand complex to (E)-5
changes (Table 2). This apparent zero-order dependence on base/
ligand complex is precedented and may be accounted for by
the kinetic scheme shown in Figure 5.45
In this scheme, a prelithiation complex, designated as C, is
formed and the rate of product formation is directly proportional
to the concentration of C (eq 1). This concentration may be
expressed in terms of (E)-5, shown in Figure 5 as A, the
concentration of base/ligand complex, M, and the equilibrium
constant for the formation of C (eq 2).46 Substituting the new
expression for C into eq 1 results in eq 3. If it is assumed that
the equilibrium represented by Keq should significantly favor
C, then the equation may be simplified such that the rate of
product formation is dependent only on the rate constant kcfp
and the concentration of A (eq 4). Invoking a prelithiation
complex seems appropriate in light of the spectroscopic
observation of such a species.
Determination of Deuterium Isotope Effect and Arrhenius
Parameters. To provide additional evidence for the mechanism
and rate law proposed in Figure 5, kobsd values were determined
for several temperatures with both (E)-5 and (E)-5-d2, which
are isotopically substituted on the allylic carbon R to nitrogen.
For these reactions, pseudo-first-order conditions were main-
tained with n-BuLi (0.36 M) and (-)-sparteine (0.33 M) present
in 20-fold excess relative to (E)-5 (0.016 M).47 All reactions
were performed in cumene as opposed to toluene, as cumene is
less prone to decomposition at higher temperatures in the
presence of high concentrations of n-BuLi and (-)-sparteine.48
The data from these experiments are shown in Table 3 and an
Eyring plot (ln kobsd versus 1/temperature) is shown in Figure
6.
Figure 4. Three-dimensional infrared profile for the pseudo-first-order
reaction between (E)-5 and n-BuLi/1.
Table 1. Pseudo-First-Order Rate Constants and R2 Values for
Reactions Representing a 5-Fold Difference in Initial Carbamate
Concentration
[(E)-5] (M)
kobsd (s-1
)
R2 (%)
0.0236
0.0236
0.0168
0.00506
0.00506
0.0021
0.0022
0.0023
0.0021
0.0019
99.1
99.2
99.1
98.2
97.4
was added last in a toluene solution. As shown in Figure 4 for
the reaction of 1 equiv of (E)-5 with 28 equiv of (-)-sparteine
and n-BuLi at -73 °C, a peak appeared at 1685 cm-1 upon
addition of starting material.40 This peak was gradually replaced
by another peak at 1640 cm-1, attributed to 8‚1. For each run,
plots of ln(abs) versus time proved linear and values for kobsd
remained constant, indicating that the reaction is first order in
(E)-5 with kobsd ) 0.0021 ( 0.0002 s-1 (Table 1).41,42
To ensure that the product 8‚1 acts as neither a catalyst nor
an inhibitor of the reaction, a control experiment was performed.
After completion of a pseudo-first-order reaction, the spectrom-
eter was reset and a second aliquot of (E)-5 was injected into
the reaction mixture. Determination of pseudo-first-order rate
constants for both reactions revealed values of kobsd that were
identical within experimental error.43
In a control experiment aimed at verifying that the reactions
observed under pseudo-first-order conditions are analogous to
stoichiometric variants, 1 equiv of (E)-5 was added to a solution
of n-BuLi (13 equiv) and (-)-sparteine (13 equiv) in toluene
at -78 °C. After 30 min, 15 equiv of methyl iodide was added
and the reaction was allowed to warm to room temperature.
Relative to the normal sequence, yields of 6 were only slightly
increased and enantiomeric ratios were only slightly eroded.
(40) In reactions where (E)-5 is not present upon establishment of the
spectral baseline, a broad peak at 1685 cm-1 appears upon its addition.
The species contributing to this peak are most likely dependent upon the
exact reaction conditions (vide infra).
(41) In a study aimed at ensuring that the React-IR response was linear
over a range of absorbance values, the infinity absorbance points for the
peaks at 1640 cm-1 were plotted versus expected concentration of 8‚1,
assuming complete lithiation, for several reactions. The results show that
the instrument response is linear; see Supporting Information section XIII
for additional details. See also Supporting Information section XVIII for a
discussion of errors in these measurements.
(42) At lower concentrations of (E)-5, the rate constant appears to still
be valid, but poor signal-to-noise ratios preclude inclusion of these data.
At higher initial concentrations of (E)-5, the rate constant increases in
magnitude, and the process no longer appears to be first order in (E)-5.
The cause of this phenomenon has not yet been determined.
(43) Both reactions exhibited clean first-order kinetics and rate constants
were equivalent within 15%. See Supporting Information section XV for
additional details.
(44) For complete details on these experiments, see section XIV of the
Supporting Information.
(45) Gallagher, D. J.; Beak, P. J. Org. Chem. 1995, 60, 7092-7093.
(46) The n-BuLi/(-)-sparteine complex most likely exists predominantly
as a dimer. However, the active metalating reagent is postulated for the
purpose of the analysis in Figure 5 to be monomeric. The complete kinetic
analysis in Figure 8 employs pseudospecies, as each of the components
“base”, “base‚(-)-sparteine”, and “complex” likely participate in multiple
equilibria with higher order aggregates. The equilibrium constants in Figure
8, K1, K2, and K3, include the contributions from these equilibria.
(47) Reactions executed with an excess of n-BuLi relative to (-)-sparteine
exhibit a decreased rate of reaction (vide infra). This was useful in retarding
reactions of diprotio (E)-5 at higher temperatures, which were too rapid
for accurate determination of kobsd with a 1:1 ratio of base to ligand.
(48) The substitution of cumene for toluene as the reaction solvent results
in a slight increase in rate of reaction but provides the same products with
similar diastereomeric and enantiomeric ratios. See Supporting Information
section XIV for details.

4924 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Pippel et al.
Figure 5. Proposed mechanism for lithiation of (E)-5 via a prelithiation complex.
Table 3. Variation in Pseudo-First-Order Rate Constant for
Reactions of (E)-5 and (E)-5-d2 at Several Temperatures
amine
temperature, °C
kobsd (s-1
)
(E)-5
(E)-5
(E)-5
(E)-5-d2
(E)-5-d2
(E)-5-d2
(E)-5-d2
(E)-5-d2
(E)-5-d2
-55
-65
-75
-56
-60
-66.5
-69.5
-74
-77
0.010 2
0.005 10
0.002 69
0.000 256
0.000 125
0.000 082 0
0.000 050 4
0.000 042 2
0.000 031 1
Thermodynamic parameters for lithiation of both (E)-5 and
(E)-5-d2 may be obtained from the plots in Figure 6 and are
shown in Table 4. It should be emphasized that these values
are rigorously accurate only if kcfp may be closely approximated
by kobsd. This assumption is probably not entirely valid (vide
infra). Nonetheless, we interpret the values for ∆S‡ to be
consistent with the transition state predicted by the mechanism
in Figure 5.
Figure 6. Eyring plot for deprotonation of (E)-5 and (E)-5-d2.
(AH/AD ) 0.39) are consistent with the tunneling phenome-
From these reactions, we further conclude that there exists a
high deuterium isotope effect that is consistent with C-H or
C-D bond breaking in the rate-determining step for the
reaction.49 Both the magnitude of this deuterium isotope effect
of 86 at -75 °C and 53 at -55 °C and the ratio of A values
non.50-53
(50) Kohen, A.; Klinman, J. P. Chem. Biol. 1999, 6, R191-R198.
(51) Melander, L.; Saunders: W. H. Reaction Rates of Isotopic Mol-
ecules; John Wiley & Sons: New York, 1980.
(52) Hoppe, D.; Paetow, M.; Hintze, F. Angew. Chem., Int. Ed. Engl.
1993, 32, 394-396.
(53) Clayden, J.; Pink, J. H.; Westlund, N.; Wilson, F. X. Tetrahedron
Lett. 1998, 39, 8377-8380.
(49) The observed effect is a combination of a primary isotope effect
and a secondary isotope effect.

Deprotonation of Organolithium Compounds
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4925
Table 4. Thermodynamic Parameters for Deprotonation Reactions
of (E)-5 and (E)-5-d2
303 K
(E)-5-d2
200 K
(E)-5-d2
thermodynamic
parameter
(E)-5
(E)-5
Ea (kcal/mol)
6.03
11,835
5.43
-34.9
16.0
8.24
30,424
7.64
-33.0
17.6
6.03
11,835
5.63
-34.0
12.4
8.24
30,424
7.84
-32.2
14.3
A (s-1
)
∆H‡ (kcal/mol)
∆S‡ (cal/mol)
∆G‡ (kcal/mol)
Figure 8. Model for the lithiation of (E)-5 which incorporates the
inverse dependence on n-BuLi concentration.
Table 5. Pseudo-First-Order Rate Constant for Reactions
Employing Different Ratios of n-BuLi/(-)-Sparteine
kobsd (s-1
)
n-BuLi/(-)-sparteine
experimental
simulated
1.2
1.1
1.0
0.91
0.83
0.00057
0.0010
0.0048
0.0087
0.014
0.00056
0.0011
0.0043
0.011
0.014
Figure 7. Experimental and simulated data displaying inverse ratio
between kobsd and n-BuLi concentration at constant (-)-sparteine
concentration.
Effects of Perturbing n-BuLi/(-)-Sparteine Ratio. The rate
law derived in Figure 5 requires that the lithiation display a
first-order dependence on (E)-5, but be independent of the
concentration of any other species in the reaction mixture. This
holds true as n-BuLi/(-)-sparteine concentrations are altered
so long as the two reagents are maintained in a 1:1 ratio.
However, if the concentration of n-BuLi is changed, while the
concentration of (-)-sparteine is held constant, the rate of the
reaction is impacted. This was demonstrated for the lithiation
of (E)-5 (0.017 M) by n-BuLi (0.34 to 0.49 M) in the presence
of (-)-sparteine (0.41 M) in cumene (Table 5, Figure 7).
Surprisingly, the reaction rate constant shows an inverse
dependence on n-BuLi concentration. As the concentration of
n-BuLi increases from 0.34 to 0.49 M, the value of kobsd
decreases from 0.014 to 0.000 57 s-1. One possible explanation
for this effect considers an additional equilibrium in which
excess n-BuLi may sequester the lithiation substrate (E)-5 as a
species designated as “unreactive complex” (Figure 8).46 The
facility of this mechanism to explain the experimental data was
assessed through simulation and fitting routines performed using
the software packages KINSIM and FITSIM.54-56 This exercise
shows that the mechanism in Figure 8 adequately fits the experi-
Figure 9. Simulated profiles for reactions with (a) excess and (b)
deficient n-BuLi.
mental data if K1 ) 1.3 × 107, K2 ) 2.3 × 104, K3 ) 2.6 ×
105, and k4 ) 2.19 × 10-2. Simulated values for kobsd at various
n-BuLi/(-)-sparteine ratios are shown in Table 5 and Figure 7.
Additional insights into the mechanism may be gained
through consideration of the reaction course for the most
important species. Figure 9 shows the simulated reaction
progress curves for product, carbamate, “complex”, “unreactive
complex”, “base”, and (-)-sparteine at high (1.2:1) and low
(55) Barshop, B. A.; Wrenn, R. F.; Frieden, C. Anal. Biochem. 1983,
130, 134-145.
(56) Zimmerle, C. T.; Frieden, C. Biochem. J. 1989, 258, 381-387.
(54) http://www.biochem.wustl.edu/cflab/message.html.

4926 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Pippel et al.
addition of (-)-sparteine at time 800 s the deprotonation occurs,
as evidenced by the slow increase in absorbance at 1640 cm-1 57
.
The potential for using this to develop a reaction that is catalytic
in (-)-sparteine is recognized.58 However, it would require an
in situ electrophile that would react with the lithiated intermedi-
ate to release (-)-sparteine while remaining inert to free n-BuLi.
Conclusions
The work reported herein establishes that the structure for
the lithiated intermediate in the lithiation-substitution reactions
of (E)-5 in solution is an R-lithio, η1-coordinated monomer. The
results of the heteronuclear NMR study in conjunction with a
Hoffmann-type test for configurational stability establish that
the enantiomeric ratios of the products arise from an initial
asymmetric deprotonation, while the Z/E ratio results from a
dynamic, kinetically controlled reaction of rotamers in the
substitution step. Pseudo-first-order rate constants were obtained
for a variety of concentrations of amine, (-)-sparteine, and
n-BuLi by monitoring the lithiation of (E)-5 through in situ
infrared spectroscopy. The reaction was found to be first order
in (E)-5 and display a zero-order dependence on 1:1 base/ligand
complex, consistent with initial formation of a prelithiation
complex. However, if the concentration of n-BuLi was varied
independently of (-)-sparteine concentration, the reaction
exhibited an inverse dependence on n-BuLi concentration. This
behavior was simulated with a simplified reaction mechanism
and optimized equilibrium and rate constants. If the model
developed for the simulation is correct, then the spectroscopi-
cally observed prelithiation complexes 9 include both unreactive
aggregates containing n-BuLi and carbamate and the reactive
species formulated as (E)-5‚n-BuLi‚(-)-sparteine. The proposed
model is supported by all experimental evidence including
thermodynamic parameters, which were within expected ranges,
and the observed high deuterium isotope effect. The establish-
ment of a mechanism for the deprotonation of (E)-5 is
particularly valuable in that it serves as the best available model
for the synthetically important lithiation of cinnamylamine (E)-
2. This description of the lithiation of (E)-5 by n-BuLi/(-)-
sparteine in conjunction with our earlier analysis of deproto-
nation of (E)-2 by the same base/ligand combination provides
examples of asymmetric deprotonation leading to both localized
and delocalized lithiated N-Boc allylic amines. The synthetic
utility of these species is of continuing interest.
Figure 10. Three-dimensional profile showing the progression of
reaction upon addition of n-BuLi at 10 s and addition of (-)-sparteine
at 800 s.
(0.83:1) concentrations of n-BuLi relative to (-)-sparteine. It
is apparent that under this model, at n-BuLi/(-)-sparteine
1.2:1, the dominant substrate species in the reaction mixture is
the “unreactive complex”. The model also predicts that the
concentration of free (-)-sparteine stays fairly constant through-
out the reaction at ∼0.002 M while the concentration of free
“base” is off scale at ∼0.075 M. In the case of deficient base,
n-BuLi/(-)-sparteine 0.83:1, both “complex” and “unreactive
complex” exist at similar concentrations with “complex” being
present in slight excess. Here, “base” is present in a constant
low concentration of ∼0.001 M and (-)-sparteine is off scale
at ∼0.075 M. It is apparent that the higher the concentration of
the “complex” at any given time, the greater the probability of
productive forward reaction. The actual equilibria in solution
are probably much more complex than those presented here with
higher order aggregates also making significant contributions.46
Nonetheless, this model is sufficient to explain the data in a
semiquantitative manner that rationalizes the effect.
The dramatic acceleration of rate in the presence of excess
(-)-sparteine relative to n-BuLi is of interest from a synthetic
perspective. To gauge the generality of the phenomenon, the
reactions were repeated with 1 equiv (E)-5, 1 equiv n-BuLi,
and 1 or 2 equiv of (-)-sparteine.57 In these cases, rate constants
were determined using the method of initial rates. In fact, excess
(-)-sparteine does increase the rate of reaction, even in reactions
that are not under pseudo-first-order conditions; however, the
magnitude of the effect is significantly smaller (kobsd ) 0.0036
s-1 with 1 equiv (-)-sparteine and kobsd ) 0.0056 s-1 with 2
equiv (-)-sparteine). Because this technique for increasing the
rate of lithiation has synthetic potential, we determined the effect
of 2 equiv of (-)-sparteine relative to n-BuLi on the reaction
yield and selectivity. Metalation of 4 under these conditions
and addition of methyl iodide gave rise to products 6 in normal
yields and enantiomeric ratios: ((Z,S)-6, 44% yield, 91:9 er;
(E,S)-6, 26% yield, 96:4 er).
Acknowledgment. We thank both the National Institutes of
Health (GM 18874) and the National Science Foundation (NSF-
95-26355) for financial support of this work. D.J.P. acknowl-
edges the DuPont Pharmaceutical Co. for an ACS Division of
Organic Chemistry Graduate Fellowship. NMR experiments
were performed in the Varian Oxford Instrument Center for
Excellence NMR Laboratory (VOICE NMR Lab), in part funded
by grants from the National Institutes of Health (PHS 1 S10
RR104-01), the National Science Foundation (NSF CHE 96-
01502), and the Keck Foundation. We also gratefully acknowl-
edge the assistance of Professors Stanley Smith, Alex Scheeline,
and Wilfred van der Donk in data interpretation and the guidance
of Professor David B. Collum and Jennifer L. Rutherford in
successful application of the React-IR technology to our
metalation reactions.
(-)-Sparteine as a Reaction “Trigger”. The observed
decrease in reactivity with decreasing concentration of (-)-
sparteine relative to base suggested that the lithiation might not
occur in the absence of ligand. A test of this hypothesis using
the React-IR is shown in Figure 10. Addition of n-BuLi to (E)-5
at time 10 s generates a stable carbamate-n-BuLi complex, as
seen by the rapid appearance of a peak at 1675 cm-1. Upon
Supporting Information Available: General procedures for
(57) In addition to considering the effects of excess (-)-sparteine and
of no ligand, we also briefly explored the differences in reactivity for the
TMEDA- and (-)-sparteine-promoted reactions. The (-)-sparteine reaction
was found to be retarded relative to the TMEDA process by a factor of
∼2. See Supporting Information sections XVI and XVII for details.
reactions, instrumentation, enantiomeric purity analyses, NMR
(58) Carbolithiation reactions catalytic in (-)-sparteine have been
reported: Norsikian, S.; Marek, I.; Klein, S.; Poisson, J. F.; Normant, J. F.
Chem. Eur. J. 1999, 5, 2055-2068.

Deprotonation of Organolithium Compounds
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4927
spectroscopic studies, and React-IR spectroscopic studies, as
well as additional experimental details covering the syntheses
of 5, 6, and 7, syntheses of all isotopically labeled compounds,
determination of mixing time for React-IR setup, determination
of linearity of React-IR, data for lithiation in the presence of
TMEDA, control experiments, and application of React-IR
monitoring to (E)-2 (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA001955K