Lithium Diisopropylamide-Mediated Lithiations of Imines: Insights into Highly Structure-Dependent Rates and Selectivities

Article abstract of DOI:10.1021/ja030409z

Lithium diisopropylamide-mediated lithiations of N-alkyl ketimines derived from cyclohexanones reveal that simple substitutions on the N-alkyl side chain and the 2-position of the cyclohexyl moiety afford a 60,000-fold range of rates. Detailed rate studie

Full text of DOI:10.1021/ja030409z

Published on Web 11/13/2003
Lithium Diisopropylamide-Mediated Lithiations of Imines:
Insights into Highly Structure-Dependent Rates and
Selectivities
Songping Liao and David B. Collum*
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853-1301
Received July 8, 2003; E-mail: dbc6@cornell.edu
Abstract: Lithium diisopropylamide-mediated lithiations of N-alkyl ketimines derived from cyclohexanones
reveal that simple substitutions on the N-alkyl side chain and the 2-position of the cyclohexyl moiety afford
a 60,000-fold range of rates. Detailed rate studies implicate monosolvated monomers at the rate-limiting
transition structures in all instances. Comparisons of experimentally derived regioselectivities and rates,
taken in conjunction with density functional theory computational studies, reveal a number of factors that
influence reactivities including: (a) axial versus equatorial disposition of the proton on the cyclohexane
ring, (b) syn versus anti orientation of the lithiation relative to the N-alkyl group, (c) the presence or absence
of a potentially chelating methoxy moiety on the N-alkyl group, (d) the presence of a 2-methyl substituent
at the geminal or distal R-carbon, and (e) branching in the N-alkyl group. The isolated contributions are not
large, yet they display a strong and predictable additivity leading to a kinetic resolution of imines derived
from racemic 2-methylcyclohexanone.
Introduction
reactivity and offer a vehicle to control reactivity and selectivity
unavailable to ketones and ketone enolates.1,2 The favorable
The organolithium chemistry of imines has played a central
role in synthetic organic chemistry.1-3 Although imines and
lithioimines appear to be functionally equivalent to ketones and
ketone enolates, their reactivities are quite different. Imine
lithiations are orders of magnitude slower than the analogous
ketone enolizations,4 and the resulting lithioimines are substan-
tially more reactive toward electrophiles than are their enolate
counterparts.2,3 Self-condensations during the lithiations of
imines are almost nonexistent due to the markedly lower
electrophilicity of the imine CdN moiety as compared to the
ketone CdO moiety.5 Of particular importance, the N-alkyl
substituents on imines and lithioimines markedly influence their
properties of imines and lithioimines that have captured the
imagination of synthetic chemists also make them excellent
templates for mechanistic studies.6 We illustrate this point by
foreshadowing results summarized in Chart 1 and Scheme 1.
Chart 1 contains structurally related imines along with relative
rate constants (krel; in parentheses) for the lithium diisopropyl-
amide (LDA)-mediated lithiations (eq 1). Seemingly minor
structural variations afford a striking 60 000-fold range of
relative reactivities. In addition, results emerged that were not
predicted based on previous studies:6 (1) LDA/THF-mediated
lithiation of a 1:1 mixture of diastereomeric imines shown in
Scheme 1 proceeds with a 75:1 preference for the lithiation of
one diastereomer, and (2) lithiation occurs predominantly at the
more substituted R-carbon. Further investigation of LDA-
mediated lithiations of the imines in Chart 1 reveals structure-
reactivity relationships that display surprising additivity.
(1) Recent reviews describing some organolithium chemistry of imines:
Denmark, S. E.; Nicaise, O. J.-C. In ComprehensiVe Asymmetric Catalysis;
Jacobsen, E. N., Pfaltz, A., Yamamoto, Y., Eds; Springer-Verlag: Heidel-
berg, 1999; Chapter 26.2. Kobayashi, S.; Ishitani, H. Chem. ReV. 1999,
99, 1069. Enders, D.; Reinhold, U. Tetrahedron: Asymmetry 1997, 8, 1895.
Volkmann, R. A. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Eds; Pergamon: Oxford, 1991; Chapter 1.12. Bloch, R. Chem.
ReV. 1998, 98, 1407.
(2) For general reviews on the chemistry of lithiated imines, see: Bergbreiter,
D. E.; Momongan, M. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Pergamon: New York, 1989; Vol. 2, p 503. Martin, S.
F. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: New York, 1989; Vol. 1, p 475. Hickmott, P. W. Tetrahedron
1982, 38, 1975. Enders, D. In Current Trends in Organic Synthesis; Nozaki,
H., Ed.; Pergamon: New York, 1983. Whitesell, J. K.; Whitesell, M. A.
Synthesis 1983, 517. Also, see ref 3.
(6) (a) Wanat, R. A.; Collum, D. B.; Van Duyne, G.; Clardy, J.; DePue, R. T.
J. Am. Chem. Soc. 1986, 108, 3415. (b) Kallman, N.; Collum, D. B. J. Am.
Chem. Soc. 1987, 109, 7466. (c) Bernstein, M. P.; Romesberg, F. E.; Fuller,
D. J.; Harrison, A. T.; Williard, P. G.; Liu, Q. Y.; Collum, D. B. J. Am.
Chem. Soc. 1992, 114, 5100. (d) Bernstein, M. P.; Collum, D. B. J. Am.
Chem. Soc. 1993, 115, 789. (e) Bernstein, M. P.; Collum, D. B. J. Am.
Chem. Soc. 1993, 115, 8008. (f) Romesberg, F. E.; Collum, D. B. J. Am.
Chem. Soc. 1995, 117, 2166. (g) Ma, Y.; Collum, D. B., unpublished.
(3) Fraser, R. R. In ComprehensiVe Carbanion Chemistry; Buncel, E., Durst,
T., Eds.; Elsevier: New York, 1980.
(4) By comparison, whereas LDA/THF lithiates simple imines of cyclohexanone
slowly below 0 °C, enolizations of cyclohexanone are too fast to monitor
at -78 °C.
(5) Huryn, D. M. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming,
I., Eds.; Pergamon: New York, 1991; Vol. 1, Chapter 1.2.
9
15114
J. AM. CHEM. SOC. 2003, 125, 15114-15127
10.1021/ja030409z CCC: $25.00 © 2003 American Chemical Society

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
Chart 1. Relative Rate Constants for LDA/THF-Mediated Lithiations at -40 °C
Scheme 1
Literature Background
It is instructive to preface the results with a brief survey of
the literature germane to understanding the mechanism of LDA-
mediated imine lithiations. Because the chemistry of ketimines
differs considerably from that of their aldimine, hydrazone,
oxime, and oximino ether counterparts,2,3 we focus solely on
the N-alkyl ketimines.
Syn-Anti Stereochemistry. The consequences of the ster-
eochemistry about the CdN bond of imines and lithioimines
reviewed by Fraser3 can be quite confusing due to a large
number of related issues. In particular, syn and anti are used to
refer to the stereochemical relationship of the N-alkyl moiety
relative to (1) a substituent on an unsymmetrical substrate, (2)
the site of deprotonation, (3) location of the substituent
introduced through alkylation, and (4) placement of the car-
banionic carbon of the resulting lithioimine. Past debates and
lingering issues are summarized as follows.
half-life for syn-anti isomerization is typically g0.5 h at 25
°C,3,7-9 the stereochemistry of the CdN bond can influence
the reactions of imines.10
(2) After some early confusion about whether lithiations occur
syn or anti to the N-alkyl moiety of imines and related Schiff’s
bases, the groups of Bergbreiter, Newcomb, and Meyers
meticulously investigated lithiations of acyclic ketimines and
found that the regiochemistry depended on a number of factors
including the steric demands of the N-alkyl group as well as
temperature.10 Subsequent semiempirical calculations revealed
(7) Knorr, R.; Hintermeyer-Hilpert, M.; Bo¨hrer, P. Chem. Ber. 1990, 123, 1137.
Fraser, R. R.; Banville, J.; Akiyama, F.; Chuaqui-Offermanns, N. Can. J.
Chem. 1981, 59, 705.
(8) Leading references: Jennings, W. B.; Boyd, D. R. J. Am. Chem. Soc. 1972,
94, 7187. Boyd, D. R.; Jennings, W. B.; Waring, L. C. J. Org. Chem. 1986,
51, 992.
(9) Steinig, A. G.; Spero, D. M. Org. Prep. Proced. Int. 2000, 32, 205.
Willoughby, C. A.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116, 11703.
(10) (a) Smith, J. K.; Bergbreiter, D. E.; Newcomb, M. J. Am. Chem. Soc. 1983,
105, 4396. (b) Smith, J. K.; Newcomb, M.; Bergbreiter, D. E.; Williams,
D. R.; Meyers, A. I. Tetrahedron Lett. 1983, 24, 3559. (c) Smith, J. K.;
Bergbreiter, D. E.; Newcomb, M. J. Org. Chem. 1981, 46, 3158.
(1) Previous workers investigated the relative preference for
the syn versus anti orientation in unsymmetrical imines as well
as the mechanism of the syn-anti isomerization.3 Because the
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15115

A R T I C L E S
Liao and Collum
little inherent syn-anti preferences in the deprotonation,6f yet
they focused on the lithiations anti to the N-alkyl moiety.
Lithiations of imines derived from unsymmetrically substituted
ketones are complex due to the superposition of syn-anti
preferences and steric effects. Lithiations often occur at the less
substituted site10-12 as illustrated by the conversion of 19 to 20
(eq 2).13 The regioselectivity seems logical in light of highly
regioselective LDA-mediated ketone enolizations.14 Nonetheless,
lithiations at the more substituted positions of imines have been
reported, the most compelling cases being those reported by
Sakurai and co-workers.15 The asymmetric alkylations reported
by Meyers and co-workers (eq 2) are especially pertinent.13
Although it was unclear to Meyers whether the isomerization
of syn imine 19 to the anti isomer occurred before or after
lithiation, the methyl moiety in 19 appeared to retard lithiation.
unclear. Glazer and Streitwieser suggested that syn alkylations
may derive from aggregation effects.16
Stereoelectronic Effects. The relative reactivities of axial
and equatorial positions on cyclohexane ring systemssso-called
stereoelectronic effectsshave received considerable attention.3,19
Although we have not uncovered support in the literature for
axially selective deprotonation of imines (21),20 the resulting
lithioimines display a high propensity to alkylate axially (22).3,21
Mechanisms of Lithiation. Most studies of the lithiation and
alkylation of imines offer no insight into the underlying
organolithium chemistry. In contrast, we became interested in
exploiting the imines as vehicles to investigate organolithium
chemistry without becoming embroiled in debates about the syn
effect and other substrate-centric issues. A series of rate studies
of LDA-mediated imine lithiations uncovered three general
mechanisms loosely depicted by transition structures 23-
25.6d,e,22 Although LDA is a disolvated dimer (26) in a wide
range of solvents at all experimentally accessible concentra-
tions,23 rate studies showed that simple N-isopropyl ketimines
lithiate via transiently formed monosolvated monomers as
depicted generically by 23. Semiempirical computational studies
provided some evidence that a syn lithiation might be
competitive.6f LDA/THF-mediated lithiations of an imine bear-
ing a potentially chelating N,N-dimethylamino appendage
revealed little influence of the Me2N moiety on the relative rates
or rate law for the lithiation,6d,e suggesting that the dimethyl-
amino moiety is unable to compete with THF for coordination
to lithium (24). (This is not completely correct, as shown below.)
By contrast, lithiations of potentially chelating imines using
LDA solvated by poorly coordinating trialkylamines proceed
via a dimer-based transition structure suggested to be a solvent-
free open dimer 25.6d,e We will demonstrate below that chelation
by the potentially chelating N-alkyl moiety depends on other
substituents within the substrate.
(3) There appears to be consensus that lithioimines alkylate
syn to the N-alkyl moiety as illustrated in eq 2.3,10,16,17
Nonetheless, the origins of the effect have been a topic of
considerable debate.
(4) The syn alkylations have been attributed to a high
preference for the lithioimine to orient the N-alkyl moiety syn
to the carbanion.3,17 Subsequent solid-state and solution-phase
structural studies of an N-phenyl lithioimine confused the issue
by showing that the N-phenyl moiety was neither syn nor anti
to the carbanionic carbon but rather skewed out of the CdC-N
plane in a dimeric structure.6a,b Moreover, some concern arose
that the dimers and monomers characterized by NMR spectros-
copy6a,b might have been misassigned as conformational iso-
mers.18 We hasten to add, however, that the relationship of
N-phenyl lithioimines to the N-alkyl lithioimines remains
(11) Fustero, S.; de la Torre, M. G.; Jofre, V.; Carlon, R. Q.; Navarro, A.;
Fuentes, A. S.; Cario, J. S. J. Org. Chem. 1998, 63, 8825. Larcheveque,
M.; Cuvigny, T.; Normant, H. Synthesis 1975, 256. Ahlbrecht, H.; Von
Daacke, A. Synthesis 1984, 610. Bunnelle, W. H.; Singam, P. R.;
Narayanan, B. A.; Bradshaw, C. W.; Liou, J. S. Synthesis 1997, 439.
Katritzky, A. R. M.; Fang, Y.; Donkor, A.; Xu, J. Synthesis 2000, 2029.
(12) A report of an anti lithiation may derive from selective destruction of the
lithioimine derived from the predominant syn lithiation: Salgado, A.;
Boeykens, M.; Gauthier, C.; Declerq, J.-P.; De Kimpe, N. Tetrahedron
2002, 58, 2763.
(13) Meyers, A. I.; Williams, D. R.; Erickson, G. W.; White, S.; Druelinger,
M. J. Am. Chem. Soc. 1981, 103, 3081.
(14) Heathcock, C. H. In ComprehensiVe Carbanion Chemistry; Buncel, E.,
Durst, T., Eds.; Elsevier: New York, 1980; Vol. B, Chapter 4. Evans, D.
A. In Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New
York, 1983; Vol. 3, Chapter 1. Evans, D. A. Aldrichimica Acta 1982, 15,
23. d’Angelo, J. Tetrahedron 1976, 32, 2979.
Results
Reagents and Starting Materials. The LDA generated from
diisopropylamine and n-BuLi was recrystallized.24 Condensa-
tions of amines and ketones to form imines followed standard
literature procedures.25 The 2,6,6-trideuterated and 2,2,6,6-
tetradeuterated analogues were prepared from 2,6,6-trideuterio-
(19) Fraser, R. R.; Banville, J.; Dhawan, K. L. J. Am. Chem. Soc. 1978, 100,
7999.
(20) For computational studies of axial versus equatorial deprotonation of
cyclohexanones, see: Behnam, S. M.; Behnam, S. E.; Ando, K.; Green,
N. S.; Houk, K. N. J. Org. Chem. 2000, 65, 8970. Bordwell, F. G.;
Scamehorn, R. G. J. Am. Chem. Soc. 1968, 90, 6749. Abou Rachid, H.;
Larrieu, C.; Chaillet, M.; Elguero, J. Tetrahedron 1983, 39, 1307.
(21) Huff, B. J. L.; Tuller, F. N.; Caine, D. J. Org. Chem. 1969, 34, 3070.
House, H. O.; Umen, M. J. J. Org. Chem. 1973, 38, 1000. Kahne, D.; Gut,
S.; DePue, R.; Mohamadi, F.; Wanat, R. A.; Collum, D. B.; Clardy, J.;
Van Duyne, G. J. Am. Chem. Soc. 1984, 106, 4685.
(15) Hosomi, A.; Araki, Y.; Sakurai, H. J. Am. Chem. Soc. 1982, 104, 2081.
Welch, J. T.; Seper, K. W. J. Org. Chem. 1986, 51, 120. Welch, J. T.;
Seper, K. W. J. Org. Chem. 1988, 53, 2991. Hayes, J. F.; Shipman, M.;
Twin, H. J. Org. Chem. 2002, 67, 935. Evans, D. A. J. Am. Chem. Soc.
1970, 92, 7593. Hart, T. W.; Guillochon, D.; Perrier, G. Sharp, B. W.;
Toft, M. P.; Vacher, B.; Walsh, R. J. A. Tetrahedron Lett. 1992, 33, 7211.
(16) For computational studies as well as a detailed discussion of syn alkylations
of imines and syn-anti isomerizations of lithioimines, see: Glaser, R.;
Streitwieser, A. J. Org. Chem. 1991, 56, 6612. Glaser, R.; Streitwieser, A.
J. Org. Chem. 1991, 56, 6625. For other computational studies, see: Pratt,
L. E.; Hogen-Esch, T. H.; Khan, I. M. Tetrahedron 1995, 21, 5955. Stork,
G.; Polt, R. L.; Li, Y.; Houk, K. N. J. Am. Chem. Soc. 1988, 110, 8360.
(17) Houk, K. N.; Strozier, R. W.; Rondan, N. G.; Fraser, R. R.; Chuaqui-
Offermanns, N. J. Am. Chem. Soc. 1980, 102, 1426.
(22) Rutherford, J. L.; Hoffmann, D.; Collum, D. B. 2002, 124, 264. Qu, B.;
Collum, D. B., unpublished.
(23) Galiano-Roth, A. S.; Collum, D. B. J. Am. Chem. Soc. 1989, 111, 6772.
Gilchrist, J. H.; Collum, D. B. J. Am. Chem. Soc. 1992, 114, 794. Collum,
D. B. Acc. Chem. Res. 1993, 26, 227.
(24) Kim, Y.-J.; Bernstein, M. P.; Galiano-Roth, A. S.; Romesberg, F. E.; Fuller,
D. J.; Harrison, A. T.; Collum, D. B.; Williard, P. G. J. Org. Chem. 1991,
56, 4435.
(18) Fraser, R. R.; Chuaqui-Offermanns, N.; Houk, K. N.; Rondan, N. G. J.
Organomet. Chem. 1981, 206, 131.
9
15116 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
1 h. The neat oil and solutions in THF contain 9 and 10 as a
1:10 mixture. Curiously, samples in CDCl3 equilibrate substan-
tially faster and contain only a 1:3 mixture of 9 and 10 as
noted.30 Syn and anti isomers 11 and 12 were prepared and
characterized using analogous methods.31 Imines 7 and 8 were
prepared from 4-tert-butylcyclohexanone as a 1:1 mixture.
Unfortunately, the anticipated NOE between the syn-axial proton
at the 2-position of the cyclohexane and the methyl moiety on
the N-alkyl side chain of imine 8 was too small to detect. It is
presumed, using analogy to computations described herein and
computations of the analogous ketone enolizations,20 that the
more reactive isomer is the one requiring axial deprotonation.
(The preference for axial deprotonation will be discussed below.)
2-methylcyclohexanone and 2,2,6,6-tetradeuteriocyclohex-
anone26 (respectively) using deuterated ammonium salts (RND3Cl)
to ensure >97% isotopic enrichment.6d,e
Structures of Imines. We must first describe how the imines
in Chart 1 were prepared and characterized using representative
examples and relying heavily on the Supporting Information.
Syn and anti orientations of the N-alkyl moieties in imines have
been assigned on the basis of 13C chemical shifts and simple
NOE studies.3,27 We used a more contemporary approach that
included combinations of 1H,1H-NOESY, 1H,1H-COSY, and
1H,1H-J-resolved spectroscopies.28 The conformational assign-
ments are consistent with long-standing conformational prin-
ciples and are supported by density function theory (DFT)
computations (see the Supporting Information).29
Several conformational issues are germane to the forthcoming
mechanistic discussion. First, the imines substituted by a
secondary N-alkyl group are suggested to reside in the rotamers
that place the methine hydrogen atom proximate to the equatorial
hydrogen on the cyclohexane-depicted generically in 27 and
28-so as to minimize A1,3 strain.29 Second, the imines derived
from 2-methylcyclohexanone are either syn-axial (27) or anti-
equatorial (28). The syn versus anti orientations are readily
ascertained by the NOEs indicated. The distinction of axial and
equatorial protons follows directly from the magnitudes of the
coupling constants.
Imines 13-16, corresponding to syn-anti pairs of two
diastereomers, were initially prepared as a four-component
mixture by treating racemic 2-methylcyclohexanone with (S)-
2-amino-1-methoxypropane. A 1:10:1:10 mixture was observed
by 13C NMR spectroscopy in either THF solution or as the neat
oil consistent with equilibrium populations containing low
concentrations of syn isomers (13 and 15) and high concentra-
tions of anti isomers (14 and 16).32 CDCl3 solutions contain a
1:3:1:3 mixture, consistent with the reduced anti selectivity noted
in the simpler cases above. The 13/15 and 14/16 pairs were
distinguished as follows. Sequential lithiation and alkylation of
imine 4 affords what was shown by the bevy of NMR
spectroscopies to be a 6:1 mixture of syn-axial imines 13 and
15. Hydrolysis without allowing intervening syn-anti isomer-
ization afforded predominantly (S)-2-methylcyclohexanone (4:1
er),13 confirming the relative stereochemistry of the stereogenic
center on the cyclohexanone moiety and the N-alkyl side chain
of 13 and 15. On standing at 35 °C, the 6:1 mixture of 13 and
15 isomerized (t1/2 ) 6 h) predominantly to the anti isomer (14/
16 ) 4:1). Hydrolysis of the isomerized material afforded (S)-
2-methylcyclohexanone with limited epimerization (3.4:1 er).
Overall, 13 and 15 can be prepared stereoisomerically enriched.
As shown below, the lithiation of 16 can be easily monitored
in a 13-16 mixture. The most elusive imine, 15, can only be
formed as a minor component of a four-component mixture.
Rates of Syn-Anti Isomerization. To summarize the details
above, isomerization of the syn and anti isomers routinely occurs
with half-lives of 6-10 h at room temperature consistent with
literature reports.3,8,31 The isomerization is most readily observed
by monitoring the equilibration of the syn isomers available by
alkylation.3,10 Under all circumstances studied, the lithiations
proved to be fast as compared to the isomerizations, allowing
lithiation rates to be ascertained for syn and anti isomers
independently.33
The assignments of the syn-anti stereochemistry require
elaboration. Five pairs of isomers-7/8, 9/10, 11/12, 13/14, and
15/16-are interconverted by syn-anti isomerization (eq 3). Syn
imine 9, prepared >95% stereochemically pure by lithiation
(LDA/THF) and syn-selective alkylation (MeI),10,16 was shown
to be of general structure 27 using the 2-D NMR spectroscopies
described above. On standing in THF at 35 °C, 9 isomerizes to
predominantly anti imine 10 (general structure 28) with t1/2
≈
(30) Knorr, R. Chem. Ber. 1980, 113, 2441.
(25) Larock, R. C. ComprehensiVe Organic Transformations; VCH: New York,
1989; p 758.
(26) Peet, N. P. J. Labelled Compd. 1973, 9, 721.
(27) Fraser, R. R.; Akiyama, F.; Banville, J. Tetrahedron Lett. 1979, 3929.
Oliveros, E.; Riviere, M.; Lattes, A. Org. Magn. Reson. 1976, 8, 601. Also,
see refs 10a,c and 19.
(28) NMR Spectroscopy Techniques; Bruch, M. D., Ed.; Dekker: New York,
1996.
(29) Johnson, F. Chem. ReV. 1968, 68, 375.
(31) The equilibration of 11 and 12 is measurably faster than that of 9 and 10,
for example, as evidenced by a t1/2 of approximately 0.5 h at 20 °C and
1.0 h at 35 °C, respectively.
(32) A similar mixture was observed for an imine derived from 2-methylcy-
clohexanone: Kosmrlj, J.; Weigel, L. O.; Evans, D. A.; Downey, C. W.;
Wu, J. J. Am. Chem. Soc. 2003, 125, 3208.
(33) Although, in principle, some ill-defined species in the reaction could
facilitate syn-anti exchange, this would manifest equivalent rates and
regioselectivities for the lithiations of the syn and anti isomers.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15117

A R T I C L E S
Liao and Collum
Table 1. Regioselectivity of Imine Lithiations by LDA/THF (eq 5)a
Table 2. Rate Data for Imine Lithiations
compound
temp (°C)
THF ordera
LDA order
isotope effect
1
2
-40
0
0
0.54 ( 0.02
0.58 ( 0.04
0.59 ( 0.03
10 ( 1
9 ( 1
3
4
5
6
7
8
9
-78
-65
-40
-78
-78
-78
0
0b
0
14 ( 4
12 ( 2
8 ( 1
0
0
starting material
compd
syn/antib
product
temp, time
(°C, h)
0b
0b
0
8 ( 1
c
entry
29:30
reference
23 ( 4
9 ( 1
1
2
3
4
5
6
7
8
9
9
9:10
10
>20:1
1:10
1:>20
>20:1
1:10
1:>20
>20:1
1:10:1
1:>20
>100:1 0, 1.5
0.49 ( 0.04
0.49 ( 0.02
0.57 ( 0.05
0.51 ( 0.02
0.55 ( 0.05
0.50 ( 0.05
10c
11
12
13
14
15
16
31c
32
20
0
8 ( 2
1:4.2
15, 13
1:8.0d
calcd (entries 1,2)
calcd (entries 4,5)
-65
-40
-40
-20
-20
15
0b
0
1.5 ( 0.2
2.6 ( 0.1
11 ( 2
12 ( 2
11
>100:1 -65, 1.0
0b
0
11:12
12
1:1.4
-40, 0.8
1:1.8d
13:15e
13:14:15
14
>100:1 -40, 1.0
0b
0
0
0.56 ( 0.01
0.55 ( 0.03
0.51 ( 0.03
5.3 ( 0.4
>5
1:2.4
-20, 0.8
1:5.5d
calcd (entries 7,8)
0
-40
10 13:14:15:16 1:10:1:10 1:2.0
11 16 1:>20
15, 7.0
1:1.6d
calcd (entries 7-10)
a The zeroth orders were determined by visual inspection rather than by
numerical fit (Figure 3). b Reaction rates increase at low THF concentration.
c From a previous study.6e
a Reactions were carried out under pseudo-first-order conditions as
described in the text. b Numerical ratios are obtained from 13C NMR spectra.
c Ratios obtained from gas chromatography. d Calculated from the regiose-
lectivity of the corresponding syn isomer and the mixture of syn/anti isomers.
e 13:15 ) 6:1.
Inspection of the syn-anti ratios and the regioselectivities
(Table 1) indicates that the regioselectivities are determined
largely, but not entirely, by the orientation of the N-alkyl groups.
Syn imines 9, 11, and 13 show a strong preference for lithiation
anti to the N-alkyl moiety because anti lithiation also corre-
sponds to lithiation at the less substituted (methylene) carbon.
Anti imines 10, 12, 14, or 16 lithiate predominantly anti to the
N-alkyl moiety despite the requisite unfavorable lithiations at
the more substituted 2-position to provide 2,2-dimethylcyclo-
hexanone. (The impact of lithiation geminal to the methyl is
discussed below.) However, the 10-40% lithiation at the less
substituted sites cannot be derived solely from the 10%
contamination by the syn isomers. It would appear, therefore,
that some syn lithiation must be occurring. Logically, the
regioselectivities derived from syn-anti mixtures depend on
the extent of lithiation attributable to the rapid lithiation of the
minor syn isomers followed by the substantially slower lithiation
of the major anti imines.
Regioselectivity of Lithiation. The regioselectivities of the
lithiations were measured by quenching the lithioimines with
MeI, hydrolyzing the imine with pH 4.5 buffer,13 and measuring
the ratio of 2,6-dimethylcyclohexanone (29) and 2,2-dimethyl-
cyclohexanone (30) by GC (eq 5, Table 1).34 The alkylations
were fast at -78 °C as shown by in situ IR spectroscopy.35
Moreover, the regioselectivities arose from kinetically controlled
lithiation. For example, solutions of 6-lithio-2-methylcyclohex-
anone imines derived from the syn-oriented imines 9 or 13 do
not equilibrate to provide the 2-lithio derivatives on standing
in the absence or presence of unlithiated imine at 25 °C. Limited
(1%) proton transfer was observed by letting the anion derived
from 6 and unlithiated 12 stand at 20 °C for 2.0 h. The
regioselectivities for lithiations of imines with a simple N-alkyl
group syn to the 2-methyl moiety are measured directly using
substrates available from the highly syn-selective alkylation. In
contrast, the regioselectivities for the anti imines (entries 3 and
6) are extracted from results derived from 1:10 syn-anti
mixtures (entries 2 and 5) by factoring out the contributions
from the minor syn isomers (entries 1 and 4). Regioselectivities
for 13-16 included an additional layer of complexity. The
regioselectivity for the lithiation of 13 is measured using a highly
enriched sample of 13 (with 15% 15) available by syn alkylation
(Table 1, entry 7). The regioselectivity for the three imines that
lithiate quickly (13-15; entry 8), in conjunction with an
exhaustive lithiation (13-16; entry 10), provides the regiose-
lectivity for 16 (entry 11). The regioselectivity derived from
14 and 15 cannot be extracted. However, if one assumes by
analogy to the other syn imines 9 and 11 (supported by the
computations) that 15 lithiates with total anti selectivity, then
the selectivity for 14 (entry 9) can be extracted.
Kinetics: General Methods. Rate studies of the LDA/THF-
mediated lithiations provide a foundation to understand the
substituent effects. Pseudo-first-order conditions were achieved
by setting the initial concentrations of imines at 0.005 M. The
LDA and THF were maintained at high, yet adjustable,
concentrations using hexane as the cosolvent.36 The lithiations
were monitored using in situ IR spectroscopy by following the
loss of the imine absorbance at 1652-1667 cm-1 as described.6e
The resulting pseudo-first-order rate constants (kobsd) are
independent of the initial concentration of imine, confirming
the first-order dependence.37 Isotope effects measured using
2,2,6-trideuterated and 2,2,6,6-tetradeuterated substrates (Table
2), despite the considerable range (kH/kD ) 1.5-23), point to a
rate-limiting proton transfer.6d,e,38
Kinetics: Relative Rate Constants. The relative rate con-
stants (krel) listed in Chart 1 were determined from the values
of kobsd measured under pseudo-first-order conditions as de-
scribed above. Lithiation rates for imines 1, 2, 5, and 12-14
(34) Alkylations of 11 and 12 both show 40-60% 2-methylcyclohexanone after
hydrolysis that could derive from N-alkylation. This could cause some
distortion in the reported regioselectivities.
(35) (a) Sun, X.; Collum, D. B. J. Am. Chem. Soc. 2000, 122, 2452. (b) Rein,
A. J.; Donahue, S. M.; Pavlosky, M. A. Curr. Opin. Drug DiscoVery DeV.
2000, 3, 734.
(36) The concentration of the lithium amide, although expressed in units of
molarity, refers to the concentration of the monomer unit (normality).
(37) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms; McGraw-
Hill: New York, 1995; p 15.
9
15118 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
Figure 1. Biphasic decay of the lithiation of a 1:1 mixture of imines 7
and 8 (total concentration: 0.005 M) by LDA (0.13 M) in THF (12.2 M)
at -78 °C. The decay was fit to (1) y ) a[exp(-kobsdt)] (- - - line), kobsd
) (6.0 ( 0.1) × 10-4; (2) y ) a[exp(-k7t) + exp(-k8t)] (solid line), k7 )
(1.53 ( 0.08) × 10-3, k8 ) (3.4 ( 0.2) × 10-4
.
were measured directly at -40 °C. Especially reactive substrates
(3, 4, 6-8, and 11) and relatively unreactive substrates (9, 10,
and 16) were monitored at two temperatures and extrapolated
to -40 °C. Rate studies described below reveal that all lithiation
rates are independent of the THF concentrations at >5.0 M THF
(see the Supporting Information). Imines 1-6 presented no
technical challenges beyond the temperature extrapolations.
Measuring the relative rate constants of the other imines
demanded varying degrees of innovation as follows.
In principle, a 1:1 mixture of 7 and 8 might display biphasic
kinetics (see below) due to differential lithiation rates. Indeed,
the loss of the IR absorbance at 1663 cm-1 corresponding to
the 7/8 mixture displayed a deviation from a clean first-order
decay in which the initially high rates in the first half-life give
way to attenuated rates in the remaining half-lives as shown by
the dotted line in Figure 1. Alternatively, the data were fit to
the expression
Figure 2. Biphasic decay of the lithiation of a 1:10:1:10 mixture of imines
13, 14, 15, and 16 (total concentration: 0.005 M) by LDA (0.13 M) in
THF (12.2 M). (A) -20 °C; the decay was fit to y ) exp(-kobsdt) + c,
kobsd ) (1.3 ( 0.5) × 10-3, used as kobsd of 14; (B) 15 °C; the decay
(excluding points from the first 600 s) was fit to y ) exp(-kobsdt) + c,
kobsd ) (1.3 ( 0.2) × 10-4
.
Consequently, concentrated solutions of 9 generated in situ were
injected without workup or warming into solutions of LDA.
The small amounts of LiI or other contaminants could, in
principle, influence the rate of lithiation. Nonetheless, an
analogous sample of imine 10 generated by alkylation followed
by warming to 25 °C (to effect syn-to-anti isomerization)
displays the same reactivity as samples of 10 injected as a neat
oil. The relative rate constants for 11 and 12 were determined
using analogous methods.
[imine]total ) [7]0 exp(-k7t) + [8]0 exp(-k8t)
(k7 and k8 correspond to the rate constants for lithiation of 7
and 8), affording an excellent fit. (The curve is obscured by
the data points in Figure 1.) Because 7 and 8 could not be
distinguished spectroscopically, we rely on computational
evidence described herein and elsewhere20 to attribute the
approximate 5-fold difference to the preferred axial lithiation
of 7 as compared to the equatorial lithiation of 8 (vide infra).
Lithiation of anti imine 10, contaminated by 10% syn isomer
9, was monitored directly by IR spectroscopy. The relatively
rapid lithiation of 9 at the outset allowed 10 to be monitored
independently. By contrast, the syn imine 9 was prepared
isomerically pure by the syn-selective alkylation protocol but
could not be isolated without intervening isomerization to 10.
Measuring the relative rate constants for the lithiation of 13-
16 proved challenging. The lithiations of a 1:1 mixture of 14
and 16 (contaminated by low concentrations of syn imines 13
and 15 discussed above) showed striking biphasic kinetics. The
lithiation followed a first-order decay to 50-60% conversion
at -20 °C due to rapid lithiation of 14 (Figure 2A) as well as
the low concentrations of 13 and 15. Alternatively, lithiation at
15 °C proceeded immediately to 50-60% conversion due to
the nearly instantaneous lithiation of 13-15 followed by a first-
order loss of the remaining imine corresponding to the lithiation
of 16 (Figure 2B). Therefore, the lithiation rates of 14 and 16
(38) The small isotope effects for 11 and 12 could be interpreted as evidence
of a rate-limiting syn-anti isomerism as suggested previously.10 The LDA-
concentration-dependent rate constants argue against such a mechanism
due to the dependence of regioselectivity on syn/anti orientation in the
substrate. Moreover, we have noted surprisingly low isotope effects in the
LDA-mediated lithiation of the dimethylhydrazone of cyclohexanone, in
which symmetry renders syn-anti isomerism irrelevant. Collum, D. B.;
Romesberg, F. E., unpublished.
were measured independently. Monitoring the lithiations by 13
C
NMR spectroscopy confirmed that the 13CdN resonances of
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15119

A R T I C L E S
Liao and Collum
Figure 3. Plot of kobsd versus free [THF] in hexane for the lithiation of 4
(0.005 M) by LDA (0.13 M) at -65 °C. The line depicts an unweighted
least-squares fit to kobsd ) k[THF] + k′ (k ) (3 ( 2) × 10-5, k′ ) (2.5 (
0.1) × 10-3).
Figure 4. Plot of kobsd versus [LDA] in THF (12.2 M) for the lithiation of
4 (0.005 M) at -65 °C. The curve depicts an unweighted least-squares fit
to kobsd ) k[LDA]n (k ) (9.8 ( 0.4) × 10-3, n ) 0.59 ( 0.03).
13-15 begin to disappear at -20 °C, whereas the 13CdN
resonance of 16 decreases measurably at 15 °C.
The syn imine 13, prepared through a syn alkylation (with a
15% contamination by 15), was further lithiated without
warming or isolation (as noted for 9 above). The loss of 13
proceeds to full conversion by a standard first-order decay. Once
again, a control experiment indicates that samples of imine
generated in situ are well behaved. Thus, if 13 is prepared in
situ, isomerized to anti isomer 14 by warming (with ap-
proximately 10% epimerization of 13 to 16), and then lithiated
by excess LDA, we obtain the same rate constant ((10%) as
that obtained for the lithiation of 14 added to LDA as a neat
oil.
The most difficult relative rate constant to measure was that
of syn imine 15, available only as a minor component in a
mixture of 13-16. By following the loss of the four isomers
by 13C NMR spectroscopy, we could show that syn imine 15
lithiates at approximately the same rate as anti isomer 14. The
NMR spectroscopic experiment also qualitatively confirmed the
relative lithiation rates of all four imines.
Figure 5. Plot of kobsd versus free [THF] in hexane for the metalation of
3 (0.005 M) by LDA (0.13 M) at -78 °C. The curve depicts an unweighted
least-squares fit to kobsd ) k[THF]n + k′ (k ) (1.4 ( 0.5) × 10-2, k′ ) (5.2
( 0.6) × 10-3, n ) -1.6 ( 0.6). The reaction is zeroth order in THF at
high [THF].
Kinetics: Rate Laws. We have argued vehemently that
relative rate constants are of limited value in the absence of
detailed rate studies showing mechanistic homo- or heteroge-
neity.6e The protocol used to determine a number of rate laws
is described below using the lithiation of 4 emblematically. The
results are summarized in Table 2.
A plot of kobsd versus THF concentration with hexane as
cosolvent (Figure 3) reveals a zeroth-order dependence on THF
concentration consistent with a lithiation that involves neither
dissociation nor association of THF from dimer 26 leading to
the rate-limiting transition structure.39 A plot of kobsd versus LDA
concentration shows a clean half-order dependence (Figure 4,
Table 2) consistent with a dimer-monomer preequilibrium.39
The rate data are described by the idealized rate law in eq 6
and are consistent with monomer-based transition structure
[(i-Pr2NLi)(THF)(4)]q.
show increases in the rates at low THF concentrations (Figure
5). These rate spikes are attributable to pathways requiring
dissociation of one or more coordinated THF. Because such
THF concentration-dependencies are confined to only the lowest
THF concentrations, we were unable to probe the details of that
pathway. We suspect, however, that the rate spikes at very low
THF concentrations derive from previously observed open
dimer-based lithiations (analogous to 25). We hasten to add that
the dissociative pathways do not measurably influence the
relative rate constants in Chart 1, nor do they influence the
mechanistic discussions below. Projected investigations using
less strongly coordinating solvents should be revealing.
The biphasic kinetics noted in a previous section for the
lithiation of the 14/16 mixture afforded the two term rate law
described by eq 7. (k14 and k16 correspond to the rate constants
for lithiation of 14 and 16.) The term for lithiation of imine 14
was determined at -20 °C, whereas that for the lithiation of
imine 16 was determined at 15 °C. The concentration-depend-
encies (Table 2) are consistent with monosolvated monomer-
-d[imine]/dt ) k[imine][LDA]1/2[THF]0
(6)
In several of the more reactive methoxy-substituted cases-
3, 7/8, 11, 13, and 16-plots of kobsd versus THF concentration
(39) Edwards, J. O.; Greene, E. F.; Ross, J. J. Chem. Educ. 1968, 45, 381.
9
15120 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
Scheme 2
Table 3. Activation Energies (∆Gq, kcal/mol) for the Lithiation of
Imines A, B, and C (Scheme 2) Calculated with B3LYP/6-31G(d)
based transition structures, [(i-Pr2NLi)(THF)(imine)]q in both
cases. The relative rate constants, however, provide compelling
evidence of substantial mechanistic differences (see below).
-d[imine]/dt ) (k14[14] + k16[16])[LDA]1/2[THF]0 (7)
Kinetics: Me2N-Substituted Imines Revisited. Previous rate
studies of the LDA/THF-mediated lithiations of imine 31/32
(1:1) containing a pendant Me2N moiety afforded rates and a
rate law that were indistinguishable from those of N-isopropyl
analogue 10.6d,e The apparent first-order decay led us to conclude
that diastereomers 31 and 32 lithiate with equal facility.
Moreover, we concluded that the lithiation proceeds via transi-
tion structure 24 with the pendant Me2N moiety unable to
compete with THF for coordination to lithium. On the basis of
the results for the analogous mixture of methoxy-substituted
imines 14 and 16 showing striking biphasic kinetics, we
suspected that we may have missed the lithiation of 32. This
oversight is readily understood by noting that previous rate
studies were carried out using transmission IR spectroscopy
rather than in situ IR spectroscopy. The contents of the reaction
vessel were fixed at either 0 or 25 °C and pumped through the
IR cell using approximately 1.0 m of intramedic tubing,
rendering the first 20-30 s of the reaction spectroscopically
invisible.
We reinvestigated the lithiation of the 31/32 mixture and
indeed found that the diastereomers lithiated at >200-fold
different rates by LDA/THF. The slow-reacting imine 31
behaves indistinguishably from N-isopropyl analogue 10, con-
sistent with transition structure 33. In contrast, the fast-reacting
form lithiates readily at e -40 °C. Detailed rate studies of this
fast-reacting form (Table 2) are consistent with a monomer-
based transition structure [(i-Pr2NLi)(THF)(imine)]q, which we
depict as transition structure 34 bearing a chelating Me2N
moiety. Thus, the rate law previously reported for the lithiation
of the 31/32 mixture is actually the rate law for the lithiation
of the slow-reacting imine 31.
a Structures of B and C with 2-(R)-methyl are not shown in Scheme 2.
b Methoxy group is not coordinated to the lithium in the transition structure.
DFT Computational Studies. We explored the origin of the
structure-dependent regioselectivities and relative rates using
DFT calculations performed with the 6-31G* basis set at the
B3LYP level of theory.40 Me2NLi and Me2O were used as
models for LDA and THF, respectively.41 A range of initial
geometries were sampled for all reactant and transition struc-
tures. Legitimate saddle points were shown by the existence of
a single imaginary frequency. Corrections for the Gibbs free
energy (∆G) are included. The structures and energies for
cyclohexanone- and 2-methylcyclohexanone-derived imines of
general form A, B, and C (Scheme 2) are included in the
Supporting Information. Optimized transition structures of
stoichiometry [(Me2NLi)(Me2O)(imine)]q for the abstraction of
the 10 different protons (labeled H1-H10 on A-C) are also
archived in the Supporting Information. The calculated activation
energies (∆Gq’s) are summarized in Table 3. A select group of
transition structures are depicted in Figures 6 and 7.
The selected geometric details of the calculated structures
i-viii are listed in Table 5. All calculated C-H-N bond angles
(40) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Gill, A.; Nanayakkara, C.; Gonzalez, M.; Challacombe,
P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez,
C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(41) Romesberg, F. E.; Collum, D. B. J. Am. Chem. Soc. 1992, 114, 2112.
Ramirez, A.; Collum, D. B., unpublished.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15121

A R T I C L E S
Liao and Collum
Figure 6. Transition structures and activation energy diagram of Me2NLi-mediated metalation of N-methyl imine of cyclohexanone.
Figure 7. Transition structures and activation energies of the Me2NLi-mediated metalation lithiation of the N-methyloxyethyl imine of cyclohexanone.
fall in a narrow 160-170° range with relatively invariant C-H
and N-H distances. As one might expect, the C-H and CdN
bonds are longer than those found in the imines consistent with
fractional bonding. The lithium in the syn lithiations orients more
nearly perpendicular to the CdN-containing plane than does
the lithium in the anti lithiations as evidenced by the C-C-
N-Li dihedral angles more closely approximating 90°. The syn
structures also show H-C-C-N dihedral angles approaching
90°, allowing the C-H bond to be collinear with the p-orbitals
of the CdN π system. The anti structures show a shorter N-H
bond, possibly emblematic of a more advanced proton transfer.
Therefore, the syn lithiations show geometries that seem
intuitively more in line with stereoelectronic principles, yet this
is not manifested in the calculated energies.
The calculated activation energies provide insight into the
relative facilities of proton abstraction (1) syn versus anti to
the N-alkyl group, (2) axial versus equatorial on the cyclohexane,
(3) geminal to a methyl moiety versus a proton, and (4) distal
to an axial methyl versus a proton. Of the 55 possible
comparisons of ∆Gq’s in Table 3, a select group of especially
pertinent examples are listed in Table 4. The general structure-
reactivity relationships will be exploited to provide insights into
the structure-dependent rates, regioselectivities, and kinetic
resolutions.42
9
15122 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
Table 4. Selected Comparisons of Calculated Activation Energies
(kcal/mol) for the Lithiations of Imines A, B, and C (Scheme 2)
from Table 3
on whether the substitutionally unsymmetrical imines were
prepared from unsymmetrical ketones or by alkylation of imines
derived from symmetric ketones. These results are consonant
with those from the groups of Bergbreiter, Newcomb, and
Meyers.10
Mechanism of Lithiation. The results from detailed rate
studies are summarized in Table 2. In short, the rate laws are
all of a mathematical form (eq 6) consistent with lithiation via
monosolvated monomers. At risk of stating the obvious, the
rate studies do not offer insights into geometric details such as
whether (1) a pendant methoxy moiety chelates at the transition
structure, (2) deprotonation occurs from an axial or equatorial
site, or (3) lithiation occurs syn or anti relative to the N-alkyl
moiety. The rate laws do, however, serve the critical function
of confining computational studies and mechanistic discussions
to transition structures bearing a single, well-defined stoichi-
ometry-[(i-Pr2NLi)(THF)(imine)]q.39 Of course, it is much
simpler to explore relative reactivities of the four R protons on
a computer than in the lab. The synergies arising from the
structural, rate, and computational studies are central to unravel-
ing the complex structure-reactivity relationships.
Caveats. Before specific structure-reactivity relationships
are discussed, it seems prudent to digress briefly by mentioning
a few misconceptions that often emerge at the murky interface
where prose and thermochemistry meet.
First, it is easy to be fooled by seemingly large relative rate
constants. For example, although the difference in relative rate
constants for 7 (krel ) 44 000) and 8 (krel ) 8100) may seem
vastly different on first inspection, their relative magnitudes are
a modest 5.4:1, corresponding to only 0.8 kcal/mol at -40 °C.
The reactivities of 5 (krel ) 190) and 2 (krel ) 120) are essentially
indistinguishable.
Second, there is an insidious phenomenon that we facetiously
refer to as the “universal ground-state assumption”6e that seems
to arise from confusion over the distinction of “relative transition
state energies” versus “relative activation energies”. In short, it
is surprisingly common for authors to ignore the role of ground-
state energies when discussing reactivity. By doing so, one
implicitly (and quite erroneously) assumes that reactants are of
a common energy. The 7-fold greater rate for the lithiation of
syn imine 9 as compared to anti isomer 10 is illustrative.
Because syn and anti imines generally do not equilibrate at the
temperatures of the lithiations (rendering the Curtin Hammett
principle not at issue), it might be tempting to focus on divergent
steric or electronic effects in putative transition structures 35
and 36 to explain the higher reactivities of syn imine 9 as
compared to those of anti isomer 10. However, if one considers
the relative energies of the reactants (as determined by their
populations at equilibrium), one finds that isomeric transition
structures 35 and 36 are of nearly equal energy. Consequently,
the relative lithiation rates of 9 and 10 derive almost entirely
from the relative stabilities of imines 9 and 10 rather than the
a Nonchelating transition structures are not included.
Discussion
The organolithium chemistry of ketimines and aldimines has
been investigated largely from a substrate-centric perspective
in which the nuances of the underlying organometallic chemistry
are largely ignored.1-3,18 Conversely, our previous studies follow
an organolithium-centric approach in which the imines are
simply exploited as templates to study issues of solvation and
aggregation.6 It appears naive in retrospect to have believed that
an approach philosophically biased in either direction could
adequately explore frequently observed odd stereo- and regio-
selectivities and substrate-dependent rates exemplified in Scheme
1 and Chart 1. By studying a range of relatively simple substrates
(1-12), we sought to understand how structural features within
the imines influence their reactivities. Do these structure-
reactivity relationships then allow us to understand the odd
selectivities in Scheme 1 and relative reactivities of imines 13-
16? Before this point is addressed, some of the more germane
observations are summarized as follows.
Regioselectivities. The LDA-mediated lithiation highlighted
in Scheme 1 exhibited a 3:1 preference for lithiation at the more
substituted R-carbon as well as a 75:1 relative reactivity of the
two diastereomers. A more comprehensive investigation of
regioselectivity (Table 1) provides some useful details. Imines
prepared directly from 2-methylcyclohexanone exist as ap-
proximate 1:10 syn/anti mixtures both in THF solution and as
neat oils (eq 4; Table 1, entries 2 and 5). The syn and anti
isomers both lithiate several orders of magnitude faster than a
potentially competing syn-anti equilibration, allowing the
regioselectivities and lithiation rates to be determined indepen-
dently. As summarized pictorially below, the observed regio-
selectivities were attributed to (1) a totally regioselective
lithiation anti to the N-alkyl moiety of the syn form, and (2) an
attenuated (e9:1) regioselective lithiation anti to the N-alkyl
moiety in the anti form. Although the regioselectivities in Table
1 were determined by indirect methods in some cases, it is clear
that the lithiations display a high, but not total, anti selectivity.
The seemingly sporadic reversals in regioselectivity re-
ported10,13,15 are easily understood as deriving from varying
proportions of the syn and anti forms which, in turn, depend
(42) For excellent references and discussions of kinetic resolutions, see: Acc.
Chem. Res. 2000, issue No. 6.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15123

A R T I C L E S
Liao and Collum
Table 5. Selected Bond Lengths, Bond Angles, and Dihedral Angles for Calculated Transition Structures i-viii (Figures 6 and 7)
parametera
C1-C2
i
ii
iii
iv
v
vi
vii
viii
Bond Lengths (Å)
1.45
1.53
1.33
1.35
1.45
1.52
1.32
1.35
1.53
1.45
1.32
1.53
1.45
1.32
1.45
1.53
1.32
1.33
1.45
1.53
1.32
1.32
1.53
1.45
1.32
1.53
1.45
1.32
C1-C3
C1-N1
C2-H
C3-H
1.34
1.47
2.09
1.96
1.31
1.51
2.08
1.95
1.33
1.47
2.23
1.98
2.02
1.98
1.31
1.51
2.21
1.98
2.03
1.98
N2-H
1.42
2.00
1.96
1.42
1.99
1.96
1.45
2.07
1.98
2.08
1.99
1.45
2.06
2.00
2.09
2.02
N1-Li
N2-Li
O1-Li (v-viii)
O2-Li
1.93
1.93
1.91
1.91
Bond Angles (deg)
C2-H-N1
162.6
165.5
163.8
165.3
C3-H-N1
165.3
166.3
112.3
135.2
112.5
167.0
168.1
N1-Li-N2
115.2
127.4
117.1
116.0
124.2
119.6
112.4
134.9
112.7
110.0
117.7
111.6
83.8
119.4
109.4
107.3
120.4
119.5
82.6
123.3
98.5
107.1
120.8
109.3
83.7
122.2
107.0
106.4
121.4
110.0
83.9
122.2
106.0
N2-Li-O2
O2-Li-N1
O1-Li-N1 (v-viii)
O1-Li-N2 (v-viii)
O1-Li-O2 (v-viii)
Dihedral Angles (deg)
H-C2‚‚‚C1‚‚‚N1
H-C3‚‚‚C1‚‚‚N1
C2‚‚‚C1‚‚‚N1-Li
C3‚‚‚C1‚‚‚N1-Li
62.9
30.9
58.9
30.5
64.9
40.9
61.5
39.8
74.5
79.7
72.2
78.2
72.9
77.2
69.9
76.8
a C2 ) anti-CR, C3 ) syn-CR, N1 ) nitrogen of imine, N2 ) nitrogen of lithium amide, O1 ) oxygen of chelating methoxy group, O2 ) oxygen of
dimethyl ether.
Table 6. Contributions to Imine Reactivity: Theory versus
Experimenta
theory
sourceb
experiment (−40 °C)
c
entry
contribution
kcal/mol
kcal/mol
0.8
source
7/8
1
2
3
4
5
6
7
axial vs equatorial
syn vs anti
chelation
geminal methyl
branched N-alkyl
1.0-2.4 4 (1, 2)
1.8-6.3 4 (5, 6)
2.2-3.2d 10 and 12d
2.6-6.3 3 (1-10)e 1.8-2.4 1/3, 2/4
1.8-2.9 4 (9)
1.9-2.2 2/10, 3/12
0.4-1.1 1/2, 3/4
0.9-1.3 2/9, 3/11
0.0-2.4 3 (1-4)f
distal (axial) methyl 1.1-2.3 4 (10)
distal (equat) methyl 0.4-0.8 4 (12)
Figure 8. Thermochemistry for lithiation of 9 and 10.
a Comparisons exclude complex imines 13-16. b Table number (entry
numbers). c Compound numbers in Chart 1 used to calculate experimental
contributions (∆G ) -RT ln(k1/k2), T ) -40 °C). d Calculated from
regioselectivities of 10 and 12 (Table 1, entries 3 and 6) and geminal methyl
effect (Table 6, entry 4, 1.9-2.2 kcal/mol). This value does not include
provisions for the distal equatorial methyl. e Comparing Me versus
CH2CH2OCH3 and i-Pr versus CH(CH3)CH2OCH3 in Table 3. f Comparing
Me versus i-Pr and CH2CH2OCH3 versus CH(CH3)CH2OCH3 in Table 3.
energies of their respective transition structures 35 and 36
(Figure 8). Of course, although one can compare relative
activation energies for the reaction of fundamentally different
substrates, it is invalid to discuss the relative energies of two
transition states that are not isomeric or related by a balanced
equation.
interesting patterns that help us understand the more complex
lithiations of imines 13-16.
Contributions to Rates: Theory versus Experiment. With
the aid of both experimentally observed rate constants and
computationally determined activation energies, we can at least
partially dissect the observed rates and selectivities into six
components as summarized in Table 6 and illustrated generically
by 37-39. Thermochemical dissections are often difficult both
scientifically and linguistically because the components are
superimposed and can be correlated. Nevertheless, by focusing
on the lithiations of simple imines 1-12, we discover some
9
15124 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
(1) Axial versus equatorial: Representative anti-axial, anti-
equatorial, syn-axial, and syn-equatorial deprotonations are
illustrated by the calculated transition structures and free energy
diagrams in Figures 6 and 7. A significant (1.0-2.4 kcal/mol)
preference for axial over equatorial deprotonation (cf., entries
1 and 2 in Table 4; see 37) is predicted regardless of whether
the deprotonation occurs syn or anti to the N-alkyl group and
regardless of whether the N-alkyl forms a chelate.
Isotopic labeling studies could, at least in principle, delineate
the preference for axial versus equatorial deprotonation; how-
ever, the challenges confronted in preparing the labeled isomers
due to imine-enamine tautomerization were too daunting.
Nonetheless, conformationally anchored substrates 7 and 8
provide tacit evidence that axial deprotonation is preferred.
Given the strong evidence that chelation is involved (discussed
below), the high conformational control imparted by A-strain
forces the preferred anti lithiations of 7 and 8 to occur from the
axial and equatorial sites, respectively, as illustrated in 40 and
41. Given the 1:1 ratio of 7 and 8,43 the approximate 5-fold
higher rate for the lithiation of 7 reflects a 5-fold preference
for axial proton abstraction (∼0.8 kcal/mol at -40 °C). Thus,
the theory and experiment are in qualitative agreement, although
the calculations using Me2NLi/Me2O as a model appear to
overestimate the stereoelectronic effect.
predicted 2.6-6.3 kcal/mol reduction in activation energy for
the chelated forms. The most pertinent comparisonsthe chelated
versus nonchelated anti-axial deprotonation via transition struc-
tures i and v (Figures 6 and 7)-shows a 3.8 kcal/mol lower
q
∆G1 for the chelated form (Table 3, entry 1). Moreover,
chelation is predicted to accelerate the syn lithiations (vii) more
so than the anti lithiations (cf., R ) Me versus CH2CH2OMe;
Table 4, entry 5), suggesting that chelation might cause a
reduction in regioselectivity.
Although rate studies do not distinguish chelated versus
nonchelated forms in the transition structures, the accelerations
imparted by the potentially chelating side chains, in conjunction
with the computational studies, leave little doubt that chelation
occurs at the transition structure. By example, the 200-fold
higher reactivity of imine 4 when compared to 2 (corresponding
to 2.4 kcal/mol at -40 °C) certainly attests to the importance
of the methoxy moiety. Similarly, the methoxy group on imine
3 causes an approximate 40-fold acceleration by comparison
to the n-butyl group of 1. Moreover, the regioselectivities in
Table 1 suggest that the methoxy moiety of imine 12 erodes
the anti selectivity (entry 6) when compared to the N-isopropyl
group of imine 10 (entry 3), although the loss of regioselectivity
is less than predicted computationally.
(4) Geminal methylation: Introducing a methyl moiety at the
2-position of the cyclohexane provides a regiochemical tag but
also retards the lithiation (see 39). Calculations suggest that a
geminally disposed methyl group in the equatorial position
causes the ∆Gq’s for the preferred anti-axial lithiation to increase
by 1.8-2.9 kcal/mol (Table 4, entry 9). Indeed, the lithiation
of 10, which occurs predominantly geminal to the methyl, is
120 times slower (2.2 kcal/mol at -40 °C) than the lithiation
of its unsubstituted counterpart 2, which matches the calculated
value of 3.1 kcal/mol (cf., ∆G1q and ∆G5q for i-Pr in Table 3).
Similarly, the lithiation of 3 is predicted computationally to be
2.9 kcal/mol more favorable than the lithiation of 12 (Table 4,
entry 9), whereas experimentally an approximate 60-fold greater
rate (1.9 kcal/mol at -40 °C) is observed.
(5) Distal methylation: To understand the regioselectivity,
one must also understand how lithiation at the 6-methylene
group is influenced by a distal 2-methyl substituent (see 37). If
the N-alkyl group is anti to the 2-methyl moiety (as in general
structure B; Scheme 2), syn lithiation at the CH2 position is
calculated to be slower than the syn lithiation in the unsubstituted
case (Table 4, entry 12). However, the 2-methyl group in B is
predicted to retard the anti lithiation (Table 4, entry 9) more so
than the syn lithiation.
Lithiations of the syn-axial imines of general structure C
prefer anti-axial lithiation at the unsubstituted CH2 position
despite the distal methyl moiety (Table 4, entry 4). The ∆Gq’s
for the anti-axial deprotonations are predicted to be inhibited
by only 1.1-2.3 kcal/mol due to the axial distal methyl (Table
4, entry 10).46 Thus, imines with the N-alkyl group syn to the
2-methyl moiety are predicted to lithiate rapidly and with a high
selectivity to the less-substituted side.
(2) Syn versus anti: Transition structures i-viii (Figures 6
and 7) and the affiliated ∆Gq’s (Table 3, ∆G1 -∆G4 ) provide
perspectives of syn versus anti deprotonations (see 38). Com-
paring syn versus anti deprotonation via the preferred axial mode
(Table 4, entry 5) is most germane and shows a distinct anti
selectivity. Although the selectivity is attenuated by both
branching on the N-alkyl group and methyl substitution on the
cyclohexane ring (Table 4, entry 7; vide infra), lithiation anti
to the N-alkyl group is predicted for all cases.
q
q
The experimental results are consistent with the computations.
The regioselectivities listed in Table 1 show a substantial
preference for deprotonation anti to the N-alkyl moiety. Syn
imines 9 and 11 undergo a totally regioselective anti deproto-
nation at the less-substituted site (Table 1, entries 1 and 3). Anti
isomers 10 and 12 undergo predominantly anti lithiation despite
the requisite deprotonation at the more substituted position.
Imines 10 and 12 undergo a small but measurable syn lithiation
as well.
(3) Chelated versus nonchelated: The presence of a poten-
tially chelating methoxy moiety on the N-alkyl group (see 39)
appears to be the largest determinant of reactivity.44 Comparing
calculated ∆Gq’s for simple imines (Table 3, R ) Me)45 and
the chelating imines (Table 3, R ) CH2CH2OMe) reveals a
Comparing the relative rates in Chart 1 for 2 versus 9 (krel
)
17; 1.3 kcal/mol at -40 °C) and 3 versus 11 (krel ) 6.8; 0.9
kcal/mol at -40 °C), we find that syn imines of general structure
C display reactivities that are intermediate between the anti
(43) The calculated relative energies of two isomers corresponding to 7 and 8
with the tert-butyl group omitted show a 0.2 kcal/mol preference for 8.
(Without the tert-butyl group, there are still two isomers.)
(44) For examples and leading references, see: Ram´ırez, A.; Collum, D. B. J.
Am. Chem. Soc. 1999, 121, 11114.
q
(45) The calculated ∆G1 for imine 1 bearing an n-butyl group is 21.9 kcal/
(46) In fact, computational studies of the lithiation of 2 and 9 (Table 3) show
q
q
mol.
no difference in ∆G8 versus ∆G1 due to the distal methyl.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15125

A R T I C L E S
Liao and Collum
Table 7. Contributions to Relative Rates for Lithiation of Imines 13-16 Scaled to Imine 3
estimated relative free energies of activation (kcal/mol)
q
q
q
q
q
q
contribution
∆Grel (44)
∆Grel (45)
∆Grel (46)
∆Grel (47)
∆Grel (48)
∆Grel (49)
axial vs equatorial
syn vs anti
chelation
0
0
0
0
0
0
0.8
0
0
0.8
2.2-3.2
0
0
0
0
2.2-3.2
0
1.8-2.4
1.9-2.2
0.4-1.1
0
geminal methyl
N-alkyl branch
distal methyl
chair-chair flip
predicted
0
1.9-2.2
0.4-1.1
0
0
1.9-2.2
0.4-1.1
0
0
0.4-1.1
0.9-1.3
0.4-1.1
0.4-1.1
0.9-1.3
(0.9)d
4.4-6.5
4.7
0b
1.3-2.4
2.1
2.3-3.3
2.8
1.2-1.9
5.3-7.3e
4.7
4.1-5.7
4.7
observeda
(2.8)c
a Scaled to -40 °C (Chart 1). b Distal axial methyl is on the opposite face of the cyclohexane ring. c Approximate. d Calculated energy of placing a
methyl group axial from the preferred equatorial orientation in the ground state for the N-methylimine of cyclohexanone. e Cost of a syn-equatorial lithiation
was estimated from the calculations to be very large (∆G9, Table 3).
forms (B) and unsubstituted derivatives (A). The prohibitively
high ∆Gq’s associated with the syn-equatorial deprotonations
of syn form C are completely borne out by the high regiose-
lectivities for syn imines 9 and 11 (Table 1).
Kinetic Resolution. Having assessed the independent con-
tributions that determine the rates of imine lithiations, we are
now in a position to examine the kinetic resolution depicted in
Scheme 1. The starting imine was shown to be a 1:10:1:10
mixture of imines 13, 14, 15, and 16. The relative rates in Chart
1 reveal the relative reactivities as 13 > 14 ≈ 15 . 16. The
biphasic kinetics affiliated with a kinetic resolution in Scheme
1 arose because 13-15 lithiate readily at -20 °C, whereas 16
lithiates slowly at 15 °C. The individual structure-reactivity
relationships gleaned from Table 6 were compiled for imines
13-15, affording predicted activation energies. The results are
summarized in Table 7. (The energetic contributions listed in
Table 7 are scaled relative to the most reactive imine 3.) Imines
13-15 are predicted to undergo facile lithiation because they
suffer from only minor problems arising from a distal methyl
group (see 44), a geminal methyl group (see 45), or a requisite
equatorial deprotonation (see 46). In contrast, a chelation-
assisted lithiation of 16 is forced to proceed via a syn-equatorial
deprotonation (see 47) predicted by computations to be pro-
hibitively poor. Alternatively, 16 could undergo a chair-chair
flip (costing 0.9 kcal/mol relative to its lowest energy con-
former)47 followed by a syn-axial deprotonation (see 48). The
cost of the conformational adjustment, the syn lithiation, and
the distal methyl renders the relative activation energy quite
large and more than offsets any advantages offered by the
chelation. Consequently, 16 appears to undergo lithiation without
assistance by chelation (see 49) to a significant extent, as
evidenced by a lithiation rate that is analogous to the rate
observed for the N-isopropyl imine 10.
(6) N-Alkyl branching: Calculations predict that branching
on the N-alkyl group (see 39) retards the lithiation. For example,
comparing ∆Gq’s for the lithiation of the unsubstituted imine
A (Table 3, entries 1-4) shows that the methyl group on the
chelating side chain [R ) CH(CH3)CH2OCH3] is predicted to
cause 0.0-2.4 kcal/mol higher activation energy. Similarly, the
preferred anti-axial lithiation of N-isopropylcyclohexanone imine
2 is calculated to be considerably less favorable (∆G1q ) 23.1
kcal/mol; Table 3) than the corresponding anti-axial lithiation
of N-n-butyl-substituted imine 1 (∆Gq ) 21.9 kcal/mol).45
The experimentally measured relative rates show a similar
effect. The lithiation of N-isopropylimine 2 is approximately
10 times slower than that of the N-n-butyl analogue 1. Similarly,
lithiation of 4 is 2-3 times slower than that of the unsubstituted
analogue 3.
Additivities. The six identifiable contributions to the lithiation
rates (listed in Table 6), although of moderate influence in
isolation, display surprising additivities. The additivities are best
illustrated by dissecting the experimentally derived 60 000-fold
relative rates for the lithiation of imines 3 and 10 (5.1 kcal/mol
at -40 °C) into the individual contributions as illustrated by
42 and 43. Whereas both imines can undergo the preferred anti-
axial lithiation, lithiation of 3 (via 43) is facilitated by chelation
(1.8-2.4 kcal/mol from Table 6) and is not inhibited by any
other structural features. In contrast, the lithiation of 10 (via
42) is inhibited by the branching in the N-alkyl moiety (0.4-
1.1 kcal/mol) and by a geminal methyl group (1.9-2.2 kcal/
mol). The predicted relative activation energies of 4.1-5.7 kcal/
mol obtained by summing these contributions compare favorably
to the observed value of 5.1 kcal/mol. This analysis works quite
well for other binary comparisons in Chart 1. It would also work
well using the computational results from Table 6 were it not
for the substantial overestimate of stabilization by the chelate.
Conclusions
A recurring theme emerges in these studies: spectroscopic,
computational, and rate studies function synergistically, offering
views of synthetically important reactivities and selectivities that
are unavailable when the methods are used in isolation. The
structure-dependent rates and selectivities of LDA/THF-medi-
ated imine lithiations highlighted in Scheme 1 and Chart 1 were
traced to a combination of factors including proton abstraction
(1) axial versus equatorial, (2) syn versus anti, (3) geminal to
a methyl moiety, (4) distal to a methyl moiety, (5) with chelation
assistance, and (6) in the presence of a branched N-alkyl
substituent. Although many of these factors have been identified
(47) The 0.9 kcal/mol is an estimate from the value derived for the chair-chair
flip of N,2-dimethylcyclohexanone imine. The cost of the chair-chair flip
must be included because the anti-axial form is not the lowest energy
conformer.
9
15126 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Lithium Diisopropylamide-Mediated Lithiations of Imines
A R T I C L E S
principles described herein translate to the dimer-based lithiation
should be fruitful. In a more general sense, the results offer a
better understanding of how the synthetically important orga-
nolithium chemistry of imines depends critically on controlling
the syn-anti stereochemistry about the CdN bond. This notion
is not new,9 but it may be underappreciated. By example, the
kinetic resolution in Scheme 1, although not yet useful in its
current form due to problems with regiocontrol and partial
epimerization,48 warrants further study.
Experimental Section
Reagents and Solvents. Amines and hydrocarbons were routinely
distilled by vacuum transfer from blue or purple solutions containing
sodium benzophenone ketyl. The hydrocarbon stills contained 1%
tetraglyme to dissolve the ketyl. The LDA was prepared and purified
by recrystallization as described.24 Deuterated imines were prepared
from the deuterated ketones26 and RND3Cl salt (to minimize loss of
deuterium) as described.6e Air- and moisture-sensitive materials were
manipulated under argon or nitrogen using standard glovebox, vacuum
line, and syringe techniques.
NMR Spectroscopic Analyses. Samples were prepared, and the 1H
and 13C NMR spectra were recorded as described in the Supporting
Information.
IR Spectroscopic Analyses. Spectra were recorded with an in situ
IR spectrometer fitted with a 30-bounce, silicon-tipped probe optimized
for sensitivity as described in the Supporting Information and elsewhere.35a
Acknowledgment. We thank the National Institutes of Health
for direct support of this work as well as Dupont Pharmaceu-
ticals, Merck Research Laboratories, Pfizer, Aventis, R. W.
Johnson, Boehringer-Ingelheim, and Schering-Plough for indi-
rect support. We also acknowledge the National Science
Foundation Instrumentation Program (CHE 7904825 and PCM
8018643), the National Institutes of Health (RR02002), and IBM
for support of the Cornell Nuclear Magnetic Resonance Facility.
and studied previously, the selectivities and relative rates, when
placed in the context of kinetic and computational studies,
provides a more unified picture. From this overview emerges a
model in which highly structure-dependent rates-rates that span
almost 5 orders of magnitude-are shown to derive from
relatively modest contributions displaying a surprising additivity.
We must confess that we began this study with the intention
of investigating the previously detected dimer-based pathway
fostered by weakly coordinating donor solvents. However, with
the aid of new technology-in situ IR spectroscopy35-it soon
became apparent that some of our previous observations and
conclusions, although technically not wrong, offered an incom-
plete picture. With a firmer understanding of the chemistry of
a structurally diverse group of imines, a probe of how the
Supporting Information Available: NMR spectra and rate
data (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
JA030409Z
(48) Imine hydrolyses without epimerization have been reported: Nakamura,
M.; Hatakeyama, T.; Hara, K.; Nakamura, E. J. Am. Chem. Soc. 2003,
125, 6362.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15127