Metalation of pyridines with nbuli-li-aminoalkoxide mixed aggregates: The origin of chemoselectivity

Article abstract of DOI:10.1021/ja910350q

The reactivity of alkyllithium-lithium-aminoalkoxide unimetallic superbases has been investigated. These systems are used for deprotonative lithiation of pyridine derivatives in apolar non-coordinating media with excellent regio- and chemoselectivity, in deep contrast with alkyllithium. With the aim of getting a better understanding of the chemistry behind these promising reagents, we have carried out a joint experimental and theoretical study of the metalation of 2-chloropyridine with combinations of nBuLi and (S)-(-)-N-methyl-2- pyrrolidinylmethoxide (LiPM). Nucleophilic addition or α-lithiation has been observed, depending on conditions (solvent, temperature, stoichiometry), while ortho-metalation was not detected. Theoretical calculations using Density Functional Theory (B3LYP/6-31 G(d) method) have then been carried out in gas phase at 195 K to characterize the relevant chemical species (reactive aggregates, transition structures) and estimate free energies of activation and relative reaction rates. Solvent effects in hexane have been neglected according to previous calculations. The effect of coordinating solvents such as THF has been qualitatively discussed. A major achievement of the present work has been to demonstrate that chemoselectivity crucially depends on aggregate type: dimers systematically lead to nucleophilic addition, while tetramers lead to a-lithiation. Besides, the calculations predict dimers to be more reactive than tetramers, yet they are much less stable, so that the observed selectivity results from the combination of both properties. A simple procedure to evaluate the basicity of an organlithium compound has been proposed. It has allowed us to show that the nBuLi-LiPM tetramer has a significantly larger basicity than its corresponding dimer, which is not at all the case for nBuLi aggregates, thus explaining differences in selectivity. Solvent and temperature effects on nBuLi-LiPM reactivity have been analyzed. By increasing the temperature in hexane, or changing the solvent from hexane to THF, dimer concentration is expected to rise, and likewise the weight of nucleophilic addition rises, in agreement with the experimental findings.

Full text of DOI:10.1021/ja910350q

Published on Web 02/01/2010
Metalation of Pyridines with nBuLi-Li-Aminoalkoxide Mixed
Aggregates: The Origin of Chemoselectivity
Hassan K. Khartabil,† Philippe C. Gros,‡ Yves Fort,‡ and Manuel F. Ruiz-Lo´pez*,†
Equipe Chimie et Biochimie The´oriques and Equipe Synthe`se Organome´tallique et Re´actiVite´,
SRSMC, Nancy-UniVersity, CNRS, BP 70239, 54506 VandoeuVre-le`s-Nancy, France
Received December 8, 2009; E-mail: Manuel.Ruiz@cbt.uhp-nancy.fr
Abstract: The reactivity of alkyllithium-lithium-aminoalkoxide unimetallic superbases has been investigated.
These systems are used for deprotonative lithiation of pyridine derivatives in apolar non-coordinating media
with excellent regio- and chemoselectivity, in deep contrast with alkyllithium. With the aim of getting a
better understanding of the chemistry behind these promising reagents, we have carried out a joint
experimental and theoretical study of the metalation of 2-chloropyridine with combinations of nBuLi and
(S)-(-)-N-methyl-2-pyrrolidinylmethoxide (LiPM). Nucleophilic addition or R-lithiation has been observed,
depending on conditions (solvent, temperature, stoichiometry), while ortho-metalation was not detected.
Theoretical calculations using Density Functional Theory (B3LYP/6-31G(d) method) have then been carried
out in gas phase at 195 K to characterize the relevant chemical species (reactive aggregates, transition
structures) and estimate free energies of activation and relative reaction rates. Solvent effects in hexane
have been neglected according to previous calculations. The effect of coordinating solvents such as THF
has been qualitatively discussed. A major achievement of the present work has been to demonstrate that
chemoselectivity crucially depends on aggregate type: dimers systematically lead to nucleophilic addition,
while tetramers lead to R-lithiation. Besides, the calculations predict dimers to be more reactive than
tetramers, yet they are much less stable, so that the observed selectivity results from the combination of
both properties. A simple procedure to evaluate the basicity of an organlithium compound has been
proposed. It has allowed us to show that the nBuLi-LiPM tetramer has a significantly larger basicity than
its corresponding dimer, which is not at all the case for nBuLi aggregates, thus explaining differences in
selectivity. Solvent and temperature effects on nBuLi-LiPM reactivity have been analyzed. By increasing
the temperature in hexane, or changing the solvent from hexane to THF, dimer concentration is expected
to rise, and likewise the weight of nucleophilic addition rises, in agreement with the experimental findings.
2,2,6,6-tetramethylpiperidide).4 These reagents have successfully
effected the metalation of several pyridine derivatives but
equilibrated reactions were observed in some cases and implied
in situ trapping of lithio intermediates.5 Other alternatives have
been reported. A well-known example is the combination of
nBuLi with complexing agents such as TMEDA, known to
disrupt the organolithium aggregates into much more reactive
dimeric complexes.6 Another example is the mixture of nBuLi
with potassium tert-butoxide to form the so-called LICKOR
1. Introduction
Organolithium compounds constitute a widely used class of
reagents in organic synthesis.1 A variety of n-butyllithium-based
(nBuLi) reagents has been employed to achieve different
functionalizations of aromatic compounds by metalation. How-
ever, the deprotonative lithiation of heteroaromatic systems
remains problematic regarding chemo- and regioselectivity. In
fact, though the presence of the heteroatom might potentially
be useful to direct metalation by coordination to the reactive
lithium aggregate, these systems often display π-deficiency,
which makes them highly electrophilic and sensitive toward
nucleophilic addition of the alkyllithium compound.
(2) Fraser, R. R.; Baigne´e, A.; Bresse, M.; Hata, K. Tetrahedron Lett.
1982, 23, 4195. Hosomi, A.; Ando, M.; Sakurai, H. Chem. Lett. 1984,
8, 1385. Comins, D. L.; LaMunyon, D. H. Tetrahedron Lett. 1988,
29, 773. Galiano-Roth, A. S.; Kim, Y.-J.; Gilchrist, J. H.; Harrison,
A. T.; Fuller, D. J.; Collum, D. B. J. Am. Chem. Soc. 1991, 113, 5053.
Romesberg, F. E.; Collum, D. B. J. Am. Chem. Soc. 1994, 116, 9187.
(3) Marsais, F.; Queguiner, G. Tetrahedron 1983, 39, 2009.
(4) Olofson, R. A.; Dougherty, C. M. J. Am. Chem. Soc. 1973, 95, 582.
Taylor, S. L.; Lee, D. Y.; Martin, J. C. J. Org. Chem. 1983, 48, 4156.
Fraser, R. R.; Bresse, M.; Mansour, T. S. J. Chem. Soc., Chem.
Commun. 1983, 620. Hall, P. L.; Gilchrist, J. H.; Collum, D. B. J. Am.
Chem. Soc. 1991, 113, 9571. Ple´, N.; Turck, A.; Martin, P.; Barbey,
S.; Que´guiner, G. Tetrahedron Lett. 1993, 34, 1605.
To overcome this side-reaction and favor lithiation, one
alternative has consisted in turning alkyllithium species into
sterically hindered non-nucleophilic lithium dialkylamides such
as LDA (Lithium diisopropylamide)2,3 or LTMP (Lithium
† Equipe Chimie et Biochimie The´oriques.
‡ Equipe Synthe`se Organome´tallique et Re´activite´.
(1) Clayden, J. Organolithiums: SelectiVity for Synthesis; Pergamon:
Oxford, 2002; Hodgson, D. M. Organolithiums in EnantioselectiVe
Synthesis; Springer: New York, 2003; Rappoport, Z.; Marek, I. The
Chemistry of Organolithium Compounds. John Wiley & Sons: New
York, 2004; Majewski, M.; Snieckus, V., Organometallics: Com-
pounds of Group 1 (Li. . Cs)., Thieme: Stuttgart, 2006.
(5) Tre´court, F.; Mallet, M.; Marsais, F.; Que´guiner, G. J. Org. Chem.
1988, 53, 1367.
(6) Collum, D. B. Acc. Chem. Res. 1992, 25, 448. Chadwick, S. T.;
Rennels, R. A.; Rutherford, J. L.; Collum, D. B. J. Am. Chem. Soc.
2000, 122, 8640.
9
2410 J. AM. CHEM. SOC. 2010, 132, 2410–2416
10.1021/ja910350q  2010 American Chemical Society

Chemoselectivity in Pyridines with Mixed Aggregates
A R T I C L E S
Scheme 1
Scheme 2
superbases, which exhibit enhanced basicity but rather poor
regioselectivity in pyridine series.7
These results have encouraged the research of new reagents
that would allow a fine-tuning of chemo and regioselectivity
by an appropriate choice of ligands that associate a coordinating
atom and an electron-rich atom. Application of this concept has
led to the development of a new class of unimetallic superbases
combining an alkyllithium and a lithium aminoalkoxide.8 The
parent reagent was BuLi-LiDMAE (LiDMAE
)
Me2N(CH2)2OLi). This superbase promoted the first R-metala-
tion of 2-methoxy pyridine and lately of other substrates instead
of the usual nucleophilic addition encountered with nBuLi9 or
ortho-directed lithiation promoted by dialkylamides2,3 (Scheme
1).
The metalation occurred with tolerance of sensitive func-
tionalities and numerous lithio-pyridine intermediates were
generated (Scheme 2).10-12
Scheme 3. Regio- and Enantiocontrol using nBuLi-LiPM with
2-Chloropyridine
An important scope extension was the switching of the
aminoalkoxide from LiDMAE to the chiral LiPM (LiPM) (S)-
(-)-N-methyl-2-pyrrolidinylmethoxide).TheobtainedBuLi-LiPM
reagent performed the first chemo-, regio- and enantioselective
functionalization of halogenopyridines leading to chiral py-
ridylcarbinols in a one-pot fashion, an example of which is given
in Scheme 3.13
While very promising, this latter process still suffers from
moderate enantiomeric excesses and conversions. Unfortunately,
there are significant difficulties to further progress in this direction
due to the imprecise knowledge of the underlying reaction
mechanisms and substrate-superbase interactions. This is explained
by the marked trend of organolithium compounds to form ag-
gregates and the intricate role of solvation effects on aggregation.
Not surprisingly, a large number of theoretical investigations have
been devoted to it,14-17 although only a few have focused on mixed
alkyllithium-lithium alkoxide aggregates.17–20
(7) Schlosser, M. J. Organomet. Chem. 1967, 8, 9. Verbeek, J.; George,
A. V. E.; De Jong, R. L. P.; Brandsma, L. J. Chem. Soc., Chem.
Commun. 1984, 257. Verbeek, J.; Brandsma, L. J. Org. Chem. 1984,
49, 3857. Schlosser, M.; Faigl, F.; Franzini, L.; Geneste, H.; Katsoulos,
G.; Zhong, G.-F. Pure Appl. Chem. 1994, 66, 1439.
(8) Gros, P.; Fort, Y.; Que´guiner, G.; Caube`re, P. Tetrahedron Lett. 1995,
36, 4791. Gros, P.; Fort, Y.; Caube`re, P. J. Chem. Soc., Perkin Trans.
1 1997, 3071.
(14) Bauer, W.; Clark, T.; Schleyer, P. v. R. J. Am. Chem. Soc. 1987, 109,
970. Ro¨thlisberger, U.; Klein, M. L. J. Am. Chem. Soc. 1995, 117,
42. Pratt, L. M.; Khan, I. M. J. Comput. Chem. 1995, 16, 1067.
Ro¨thlisberger, U.; Sprik, M.; Klein, M. L. J. Chem. Soc., Faraday
Trans. 1998, 94, 501. Hilmersson, G.; Arvidsson, P. I.; Davidsson,
(9) Ziegler, K.; Zeiser, H. Ber. Dtsch. Chem. Ges. 1930, 63, 1847. Giam,
C. S.; Stout, J. L. J. Chem. Soc. D., Chem. Commun. 1969, 142. Giam,
C. S.; Stout, J. L. J. Chem. Soc. D., Chem. Commun. 1970, 478. Giam,
C. S.; Knaus, E. E.; Pasutto, F. M. J. Org. Chem. 1974, 39, 3565.
Scriven, E. F. V., ComprehensiVe Heterocyclic Chemistry; Pergamon:
New York, 1984.
¨
O.; Ha˚kansson, M. Organometallics 1997, 16, 3352. Armstrong, D. R.;
Henderson, K. W.; Kennedy, A. R.; Kerr, W. J.; Mair, F. S.; Moir,
J. H.; Moran, P. H.; Snaith, R. J. Chem. Soc., Dalton Trans. 1999,
4063. Pratt, L. M.; Streitwieser, A. J. Org. Chem. 2000, 65, 290.
Gohaud, N.; Begue, D.; Darrigan, C.; Pouchan, C. J. Comput. Chem.
2005, 26, 743. Gohaud, N.; Begue, D.; Pouchan, C. Chem. Phys. 2005,
310, 85. Ge´rard, H.; De La Lande, A.; Maddaluno, J.; Parisel, O.;
Tuckerman, M. E. J. Phys. Chem. A 2006, 110, 4787. Tacke, M.;
Leyden, R.; Cuffe, L. P. Int. J. Quantum Chem. 2006, 106, 692. de la
Lande, A.; Fressigne´, C.; Ge´rard, H.; Maddaluno, J.; Parisel, O.
Chem.sEur. J. 2007, 13, 3459. Pratt, L. M.; Truhlar, D. G.; Cramer,
C. J.; Kass, S. R.; Thompson, J. D.; Xidos, J. D. J. Org. Chem. 2007,
72, 2962. Khartabil, H. K.; Martins-Costa, M. T. C.; Gros, P. C.; Fort,
Y.; Ruiz-Lo´pez, M. F. Theor. Chem. Acc. 2008, 121, 321.
(10) Gros, P.; Fort, Y. Eur. J. Org. Chem. 2002, 3375.
(11) Martineau, D.; Gros, P.; Fort, Y. J. Org. Chem. 2004, 69, 7914.
Loue¨rat, F.; Gros, P.; Fort, Y. Tetrahedron 2005, 61, 4761. Parmentier,
M.; Gros, P.; Fort, Y. Tetrahedron 2005, 61, 3261. Martineau, D.;
Beley, M.; Gros, P. C. J. Org. Chem. 2006, 71, 566. Loue¨rat, F.; Gros,
P. C.; Fort, Y. Synlett 2006, 1379. Comoy, C.; Banaszak, E.; Fort, Y.
Tetrahedron 2006, 62, 6036.
(12) Gros, P. C.; Fort, Y. Eur. J. Org. Chem. 2009, 4199.
(13) Fort, Y.; Gros, P.; Rodriguez, A. L. Tetrahedron Asymmetry 2001,
12, 2631.
9
J. AM. CHEM. SOC. VOL. 132, NO. 7, 2010 2411

A R T I C L E S
Khartabil et al.
Scheme 4
Using quantum chemistry calculations, we have recently
analyzed20 the thermodynamics of nBuLi-LiPM aggregation
processes in different solvents (main results will be outlined
below). This has allowed us to clarify the type of aggregates
that are present in apolar media and in coordinating solvents
such as THF, thereby opening the route to the investigation of
reaction mechanisms.
In the present work, we have carried out a combined
experimental and theoretical study with the aim of elucidating
the mechanism of reactions between nBuLi-LiPM unimetallic
superbases and pyridine derivatives, as well as the main factors
that determine their particular chemo and regioselectivities. The
superbase reactivity is compared to that of nBuLi. 2-Chloro-
pyridine has been chosen as prototypical reagent since it is
particularly sensitive to nucleophilic addition of BuLi21 (and is
thus a good probe for aminoalkoxide effect on selectivity). As
mentioned, experimental data on nBuLi-LiPM reactions has
already been reported by some of us.10,12,13 In that work, we
focused on the addition step of lithiopyridines to aldehydes
without examining in detail the metalation step. To complete
those results, it has been necessary to carry out a series of
systematic experiments to examine the effect of stoichiometries,
solvents and temperature on the selectivity. The corresponding
data are presented and discussed in the first part of this work.
Further details on the reactions can be found in the original
references.10,12,13 Afterward, theoretical predictions for the
reaction rates of R-metalation, ortho-metalation, and nucleophilic
addition are reported and compared to experiment. From this
comparison, we derive conclusions concerning the mechanism
of the reactions, the role of solvent/temperature and the origin
of chemoselectivity.
Table 1. Metalation of 1 with BuLi-LiPM under Various Conditionsa
BuLi LiPM
(equiv.) (equiv.)
entry
solvent
T°C
conv.%b 2 (%)b 3 (%)
4 (%)b
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
2
2.5
3
0
0
1
1
1
1
2
2.5
3
hexane -78
THF -78
hexane -78
hexane -40
37
70
72
55
35
65
80
95
>99
30
57
10
28
-
52
1
tr.
62
27
hexane
THF
0
-78
hexane -78
hexane -78
hexane -78
75
93
98 (90)c
a Reaction performed on 2 mmol of 1. b GC yields. c Isolated yield
after column chromatography.
exclusively with LTMP or LDA.3,22 A common feature to all
metalations was the strong solvent dependence of the chemose-
lectivity. In noncoordinating hexane, BuLi alone gave 223
resulting from nucleophilic addition of BuLi to the azomethine
bond in moderate yield which was increased in THF (entries 1
and 2). In agreement with the investigations by Que´guiner and
co-workers,23 the butylpyridone 2 was the only addition product
detected and isolated; other aromatized addition products such
as 2-butylpyridine or 2-butyl-5-trimethylsilylpyridine were found
absent of the reaction mixtures. The effect of LiPM on the
reaction pathway was clearly observed. In agreement with
BuLi-LiDMAE,10 BuLi-LiPM gave exclusively the clean
R-metalation product 4 when used in hexane at -78 °C (entries
3,7-9) while only nucleophilic addition was observed in THF
(entry 6) signing a profound change in the reagent structure in
the latter solvent. The product resulting from the ortho-
metalation at C-3 was never obtained whatever the conditions
used clearly indicating the high degree of regiospecificity at
the R position. Interestingly, the increase of base amount not
only improved the conversion, but also decreased the amount
of nucleophilic addition product which became negligible
providing that 2 eq. of BuLi-LiPM were used. The reaction
was completed using 3 equiv. of BuLi-LiPM producing 4 in
excellent yield (entry 9).
2. Results and Discussion
2.1. Experimental Data. Details on experimental methods are
provided as Supporting Information. The various BuLi-LiPM
combinations were prepared by adding nBuLi (x equiv.) to a
solution of LiPM (y equiv.) in hexane at 0 °C. The reaction
medium was then stabilized at the desired temperature and
2-chloropyridine was added as a solution in THF or hexane.
The lithiation products were subsequently reacted with chlo-
rotrimethylsilane and the reaction contents analyzed by GC. The
GC was standardized using freshly prepared authentic samples
of the known formed products 2 and 4. The results of the
reaction of 2-chloropyridine with BuLi or nBuLi-LiPM under
various conditions are given in Scheme 4 and Table 1. Since
we were mainly interested on chemo- and regio-selectivities, a
detailed investigation on reaction kinetics was not carried out.
However, some data are included in the Supporting Information
for comparison with previous works on related reactions.
At first, whatever the conditions used, the ortho-metalation
product 3 was never detected, while it is known to be obtained
The metalation temperature effect was also examined using
1 equiv. of BuLi-LiPM. An increase of the temperature from
-78 to 0 °C was found to be deleterious for conversion and
chemoselectivity.
2.2. Computational Study. Reaction rate computations have
been carried out in the framework of Density Functional Theory
using the hybrid method B3LYP/6-31G(d),24 which represents
a good compromise between accuracy and computational cost,
(15) Kwon, O.; Sevin, F.; McKee, M. L. J. Phys. Chem. A 2001, 105, 913.
Pratt, L. M.; Streitwieser, A. J. Org. Chem. 2003, 68, 2830.
(16) Catak, S.; Aviyente, V.; Nugay, T. J. Polym. Sci., Part A: Polym.
Chem. 2005, 43, 455.
(17) Pratt, L. M.; Kwon, O.; Ho, T. C.; Van Nguyen, N. Tetrahedron 2008,
64, 5314.
(18) Sorger, K.; Schleyer, P. v. R.; Fleischer, R.; Stalke, D. J. Am. Chem.
Soc. 1996, 118, 6924.
(19) Goldfuss, B.; Schleyer, P. V. R.; Hampel, F. J. Am. Chem. Soc. 1997,
119, 1072. Khartabil, H. K.; Martins-Costa, M. T. C.; Gros, P. C.;
Fort, Y.; Ruiz-Lo´pez, M. F. J. Phys. Chem. B 2009, 113, 6459.
(20) Khartabil, H. K.; Gros, P. C.; Fort, Y.; Ruiz-Lo´pez, M. F. J. Org.
Chem. 2008, 73, 9393.
(22) Gribble, G. W.; Saulnier, M. G. Tetrahedron Lett. 1980, 21, 4137.
(23) Tre´court, F.; Marsais, F.; Gu¨ngo¨r, T.; Que´guiner, G. J. Chem. Soc.,
Perkin Trans. 1 1990, 2409.
(24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. Lee, C.; Yang, W.; Parr,
R. G. Phys. ReV. B 1988, 37, 785.
(21) Choppin, S.; Gros, P.; Fort, Y. Org. Lett. 2000, 2, 803.
9
2412 J. AM. CHEM. SOC. VOL. 132, NO. 7, 2010

Chemoselectivity in Pyridines with Mixed Aggregates
A R T I C L E S
as shown in many related studies.15,18,25 Considering the large
size of the systems, the use of higher-level methods was not
possible (the role of diffuse functions and the magnitude of basis
set superposition error has been discussed before20). Full
geometry optimization in the gas phase has been done for both
reactants and transition structures followed by standard vibra-
tional frequency to verify the nature of the stationary point
(minimum or saddle point) and obtain thermal contributions to
free energy at 1 atm and 195 K. Intrinsic reaction coordinate
analyses have been done to check the species connected by a
transition state. Thermodynamic contributions and final struc-
tures in the intrinsic reaction coordinate calculations are provided
as Supporting Information. Reaction rates have been estimated
in gas phase using Transition State Theory. Solvent effects in
hexane can be safely neglected according to previous calcula-
tions20 using the multipole-expansion continuum model.26 For
instance, free energy changes in going from gas phase to hexane
were estimated to less than 1% for the processes of interest in
the present work (dimer/tetramer equilibria). On the contrary,
solvent effects in THF were found to be very large for aggregate
equilibria20 and their role on chemoselectivity will be qualita-
tively discussed below. All calculations have been done using
the Gaussian 03 program.27
Figure 1. Optimized geometries for complexes of nBuLi and nBuLi-LiPM
aggregates with 2-Cl-pyridine.
Table 2. Energies of Complexation with 2-Cl-Pyridine and Li · · ·N
Distances for Systems in Figure 1 (in kcal/mol and Å, Free Energy
at 1 atm and 195 K)
aggregate
∆E
∆G
dLi· · ·N
nBuLi
(nBuLi)2
(nBuLi)4
(nBuLi)(LiPM)
(nBuLi)2(LiPM)2
-21.7
-13.2
-10.8
-10.7
-8.9
-14.4
-6.8
-4.3
-2.8
-2.7
2.049
2.112
2.159
2.125
2.118
Scheme 5. Schematized Transition Structures for the Reactions
Investigateda
In principle, several conformations and aggregate dimensions
can be considered for each system. Conformations appearing
to be the most appropriate ones from the chemical point of view
have been selected, although computations for different initial
geometries were done in some cases (only lowest energy
structures are considered below). The size of the reactive
aggregates is an intriguing question and deserves some com-
ments. It is widely accepted that reactivity decreases with
increasing aggregate size28 so that reactive aggregates do not
necessarily correspond to major aggregates in a given solvent
(exceptions are possible29 however). Most theoretical works in
the literature have dealt with reactions in coordinating solvents
and have focused on monomer or dimer reactions. However,
for reactions in apolar media, the role of monomers is probably
negligible (due to insignificant concentrations) while that of
larger aggregates is potentially important. Therefore, we focus
below on the reactivity of dimers and tetramers. Computations
for the monomers have been done for comparison. The role of
hexamers (the major species in apolar media) has been neglected
since complexation with a Lewis base promotes their dissocia-
tion into smaller aggregates.20 A further approximation in the
present work concerns the type of nBuLi-LiPM superbase
a For simplicity, a single BuLi unit is drawn instead of the whole
aggregate (dimers ou tetramers shown in Figure 1).
tetramers considered: only the 2:2 aggregate has been studied
since it is a little more stable than asymmetric 1:3 and 3:1
tetramers and is predominant in equimolar mixtures.20
2.3. Prereactive Complexes. Organolithium aggregates and
2-Cl-pyridine may form stable complexes whose structure is
interesting to analyze before considering the associated transition
structures. The optimized geometries of the complexes are
presented in Figure 1 and some properties are collected in Table
2. All complexation energies are negative and large so that free
energies remain negative in spite of the unfavorable entropic
contribution characteristic of this type of processes. The
complexation energy for the nBuLi monomer is close to that
reported by Catak et al16 for nBuLi and pyridine at the B3LYP/
6-31+G(d) level (∆E ) -19.8 kcal/mol with zero-point energy
corrections). Values for all of the other aggregates (dimers or
tetramers) are significantly smaller. Interestingly, complexes
with homoaggregates display lower formation energies than
heteroaggregates of the same size and similarly dimers exhibit
lower energies than tetramers of the same type.
2.4. Transition Structures. Optimized geometries for the
transition structures (TS) of the studied reactions (see Scheme
5) are drawn in Figure 2. Notation is as follows: T or D holds
for a TS involving a tetramer or a dimer, respectively; NA, R,
or o subscripts hold for nucleophilic addition, R-metalation and
ortho-metalation, respectively. Values of the forming C-C or
C-H distances, energies ∆E‡ and free energies ∆G‡ of
activation are summarized in Table 3 (values are given relative
to separated aggregate and 2-Cl-pyridine reactant). The distances
allow comparison of the reaction advancement between the
processes. In general, the superbase nBuLi-LiPM leads to
shorter distances (except for the TNA TS) and the difference
with respect to nBuLi is especially important for TR. As far as
activation energies are concerned, the most striking result is
the large difference predicted between the two bases for the
(25) Abbotto, A.; Streitwieser, A.; Schleyer, P. V. R. J. Am. Chem. Soc.
1997, 119, 11255. Verstraete, P.; Deffieux, A.; Fritsch, A.; Rayez,
J. C.; Rayez, M. T. J. Mol. Struct. (THEOCHEM) 2003, 631, 53. Sott,
R.; Granander, J.; Dine´r, P.; Hilmersson, G. Tetrahedron Asymmetry
2004, 15, 267. Zuend, S. J.; Ramirez, A.; Lobkovsky, E.; Collum,
D. B. J. Am. Chem. Soc. 2006, 128, 5939.
(26) Rinaldi, D.; Bouchy, A.; Rivail, J.-L.; Dillet, V. J. Chem. Phys. 2004,
120, 2343. Rinaldi, D.; Bouchy, A.; Rivail, J.-L. Theor. Chem. Acc.
2006, 116, 664. Ruiz-Lo´pez, M. F., The multipole moment expansion
solvent continuum model: a brief review. In Challenges and AdVances
in Computational Chemistry and Physics; Canuto, S., Ed.; Springer:
New York, 2008; Vol. 6. For a comparison with other continuum
models, see: Curutchet, C.; Cramer, C. J.; Truhlar, D. G.; Ruiz-Lo´pez,
M. F.; Rinaldi, D.; Orozco, M.; Luque, F. J. J. Comput. Chem. 2003,
24, 284.
(27) Frisch, M. J., et al. Gaussian 03, Gaussian, Inc.: Wallingford, CT,
2004.
(28) McGarrity, J. F.; Ogle, C. A. J. Am. Chem. Soc. 1985, 107, 1805.
McGarrity, J. F.; Ogle, C. A.; Brich, Z.; Loosli, H.-R. J. Am. Chem.
Soc. 1985, 107, 1810.
(29) Sun, X.; Winemiller, M. D.; Xiang, B.; Collum, D. B. J. Am. Chem.
Soc. 2001, 123, 8039.
9
J. AM. CHEM. SOC. VOL. 132, NO. 7, 2010 2413

A R T I C L E S
Khartabil et al.
Figure 2. Transition structures for nBuLi and nBuLi-LiPM aggregates. Notation is as follows: T or D holds for a transition structure involving a tetramer
or a dimer, respectively; NA, R or o indices hold for nucleophilic addition R-metalation and ortho-metalation, respectively. Hydrogen atoms are not shown
for simplicity (except for the hydrogen atom being transferred in metalation processes).
Table 3. C-C and C-H Lengths (Å) for the Bonds Being Formed in the Transition Structures of Figure 2 and Energy and Free Energy of
Activation (kcal/mol)a
TS
C-Y length (Y ) C or H)
nBuLi-LiPM
∆E‡
∆G‡
nBuLi
∆
nBuLi
nBuLi-LiPM
∆
nBuLi
nBuLi-LiPM
∆
DNA
DR
Do
TNA
TR
CC
CH
CH
CC
CH
CH
2.439
1.481
1.586
2.234
1.682
1.640
2.381
1.461
1.533
2.340
1.455
1.513
-0.058
-0.020
-0.053
0.106
-0.227
-0.127
0.6
9.6
17.7
18.8
19.5
24.9
1.3
9.5
14.2
12.1
11.5
21.3
0.7
-0.1
-3.5
-6.7
-8.0
-3.6
9.0
13.2
21.6
27.3
24.7
29.7
10.2
14.4
19.4
21.3
16.2
26.7
1.2
1.2
-2.2
-6.0
-8.5
-3.0
To
a The latter are given with respect to the separated aggregate and 2-Cl-pyridine reactant.
reactions through TNA and TR, the superbase nBuLi-LiPM
leading to more stable transitions structures. Another notable
result is that free energies of activation involving dimers are
considerably smaller than those involving tetramers, when
comparing equivalent processes, thus confirming the higher
reactivity of the smaller aggregates. It is worth noting that free
energies of activation for processes involving dimers vary in
the order nucleophilic addition < R-metalation < ortho-meta-
lation, while in the case of tetramers, the order is R-metalation
< nucleophilic addition < ortho-metalation. Since, experimen-
tally, nucleophilic addition is observed with nBuLi and R-meta-
lation is mainly obtained with mixed nBuLi-LiPM, our
calculations suggest that dimers would be involved in the first
case, while tetramers would be involved in the second. Further
arguments supporting this statement will be presented in the
next section. Ortho-metalation is clearly disfavored with re-
spected to other processes because of the switch of strong Li-N
interactions, not possible in ortho-metalation TSs owing to the
long distances involved, to weak Li-Cl interactions.
be larger than that of the dimer for the same process. In this
case, R-metalation and nucleophilic addition are almost equally
favorable suggesting that monomers should not be involved in
the observed reactions.
2.5. Chemo- and Regioselectivity. The selectivity of the
organolithium aggregates can be discussed by considering the
relative reaction rates for nucleophilic addition, R-metalation
and ortho-metalation involving either dimers (Dim) or tetramers
(Tet):
kdim
Dim + Pyr
Tet + Pyr
8 Prod
(1)
(2)
ktet
9
8 Prod
(Pyr ) 2-Cl-pyridine, Prod ) reaction product). We shall
assume first order reactions with respect to aggregate concentra-
tion. In addition, in order to compare the reaction rates between
dimers and tetramers, we further assume the hypothesis of
pre-equilibrium:
Computations of TSs involving nBuLi monomers lead to ∆G‡
values of 6.0 kcal/mol, 5.5 and 10.0 kcal/mol for the nucleophilic
addition, R-metalation, and ortho-metalation processes, respec-
tively. The energetics for nucleophilic addition is comparable
to the study reported by Catak et al16 using the B3LYP/6-
31+G(d) method (estimated activation energy with respect to
separated reactants at 0 K was 1.9 kcal/mol). One should note
that, as expected, the reactivity of the monomer is predicted to
Tet { k } 2 Dim
k-1 > > kdim
(3)
1
k-1
The computed free energy for this equilibrium, ∆Geq, amounts
31.7 and 31.0 kcal/mol for nBuLi and nBuLi-LiPM, respectively.
9
2414 J. AM. CHEM. SOC. VOL. 132, NO. 7, 2010

Chemoselectivity in Pyridines with Mixed Aggregates
A R T I C L E S
Table 4. Predicted Relative Reaction Rates at 195 Ka
mechanism
nBuLi
nBuLi-LiPM
DNA
DR
Do
TNA
TR
1
6.2 × 10-9
1.5 × 10-13
3.3 × 10-19
2.3 × 10-6
1
2.0 × 10-5
7.9 × 10-15
5.3 × 10-6
5.3 × 10-3
1.2 × 10-8
To
1.7 × 10-12
a We assume [Tet] ) 10-5 M.
The estimated relative reaction rates are summarized in Table
4 (some details on equations are given in the Supporting
Information). The predicted chemo- and regioselectivities (cor-
responding to main reaction rates) are in perfect agreement with
the experimental observations: nBuLi leads to nucleophilic
addition, and the nBuLi-LiPM superbase leads to R-metalation.
The computed rates do confirm the close relationship existing
between selectivity and aggregate size that we have suggested
above. The main reaction with nBuLi proceeds through a dimer
(DNA mechanism) while the main reaction with nBuLi-LiPM
proceeds through a tetramer (TR mechanism). This conclusion
means that special attention should be paid to the aggregate
size, considering quantum mechanical calculations on organo-
lithium reactions. It is a fundamental parameter and theoretical
studies based on too simple models (monomers or dimers) might
lead to wrong conclusions.
Figure 3. Protonation energies for (nBuLi)n(LiPM)m aggregates in gas phase
according to eq 4.
The metalation ability of a given organolithium system may
be associated to its intrinsic basicity and therefore it is interesting
to discuss this property for the species considered in our work.
Superbases are defined30 as stronger bases than a proton sponge
(1,8-bis(dimethylamino)naphthalene: DMAN) with absolute
proton affinity larger than 245.3 kcal. In the context of
organolithium chemistry, however, there is not a clear-cut
definition. Superbases are considered in general as the combina-
tion of an organolithium compound and a heavy alkali metal
alkoxide exhibiting an enhanced reactivity with respect to that
of the parent compounds.31 Combination of organolithium
compounds and lithium alkoxides does not lead in general to
superbases31 and therefore, the unusual properties observed for
the mixed nBuLi-LiPM aggregates deserves some explanation.
The basicity of B is defined from the free energy released in
the process: B + H+ f BH+. Similarly, for an organolithium
compound RLi, the basicity may be gauged by the energy
released in the process: RLi + H+ f Li+ + RH. For nBuLi
and nBuLi-LiPM aggregates, this equation may be written:
Values in Table 4 assume [Tet] ) 10-5 M. Owing to the
equilibrium represented by eq 3, changing the concentration of
[Tet] will modify the relative rates between D and T mechanisms
so that decreasing [Tet] should favor D mechanisms. Since the
predicted relative rates for secondary processes are very (or even
negligibly) small, chemoselectivity is not expected to change
for reasonable changes of concentration. However, it must be
noticed that the nature of aggregates evolves along the experi-
ments, concomitantly with nBuLi concentration diminution. The
relative amount of LiPM/nBuLi changes, so that for a suf-
ficiently advanced reaction 2:2 tetramers are not necessarily the
main reactive species. In addition, the formed products (lithiated
pyridine) may participate to the formation of aggregates. The
observation of significant amounts of secondary products under
certain conditions is probably connected to that.
2.6. Origin of Chemoselectivity: Superbasicity, Solvent
and Temperature Effects. Differences in reactivity between
nBuli and nBuLi-LiPM are now interpreted in terms of their
different chemical properties. To this aim, two main questions
must be answered: why the nBuLi-LiPM superbase reaction
involves the tetramer rather than the dimer (in contrast to
nBuLi), and why the nBuLi-LiPM tetramer leads to R-meta-
lation rather than to nucleophilic addition (in contrast to the
nBuLi-LiPM dimer).
(nBuLi)n(LiPM)m+H+ f (nBuLi)n-1(LiPM)mLi+ + nBuH
(4)
The corresponding energies in gas phase are represented in
Figure 3 (we do not take into account geometry relaxation in
the charged aggregates). Some interesting remarks can be made:
• For m ) 0, i.e., for nBuLi aggregates, proton affinity is
slightly below that of the proton sponge, 245.3 kcal/mol that,
as said above, can be considered as the limit for superbasicity.
• For m > 0, i.e., for nBuLi-LiPM aggregates, proton affinity
is beyond this threshold and remarkably, it increases linearly
with the number of LiPM monomers.
• As a consequence, the basicity of (nBuLi)2(LiPM)2 is much
larger than that of (nBuLi)(LiPM), while those of (nBuLi)4 and
(nBuLi)2 are quite close.
Actually, these two questions are interrelated. On one hand,
one has to keep in mind that the relative dimer /tetramer stability
is similar for the base and the superbase so that differences in
base/superbase selectivity must come from differences in relative
dimer/tetramer reactivity. As already said and confirmed by the
calculations above, one generally expects aggregate reactivity
to decrease with increasing aggregate size. However, inspection
of activation free energies in Table 3 shows that R-metalation
with nBuLi-LiPM displays a singular behavior. For this
reaction, ∆G‡ values for DR and TR differ by only 1.8 kcal/mol
(in all other reactions, the equivalent difference varies between
7 and 18 kcal/mol). In other words, the metalation ability of
the (nBuLi)2(LiPM)2 tetramer seems to be abnormally high.
The higher basicity of mixed aggregates come from the
stabilizing effect of the LiPM amino groups in the charged
aggregate through interaction with the Li atoms. One argument
supporting this interpretation is the fact that mixed aggregates
nBuLi-ROLi (R ) alkyl) do not exhibit enhanced basicity with
respect to nBuLi.31
When the experimental conditions are such that the concen-
tration of (nBuLi)2(LiPM)2 aggregates is disfavored with respect
(30) Ishikawa, T., Superbases for Organic Synthesis: Guanidines, Amidines,
Phosphazenes and Related Organocatalysts; John Wiley & Sons:
Chichester, West Sussex, U.K., 2009.
(31) Lochmann, L. Eur. J. Inorg. Chem. 2000, 1115.
9
J. AM. CHEM. SOC. VOL. 132, NO. 7, 2010 2415

A R T I C L E S
Khartabil et al.
to that of the dimer (nBuLi)(LiPM), the metalation ability of
the superbase should be reduced. The observed solvent and
temperature effects can be explained in this way. Although, as
said above, the dimer/tetramer equilibrium for nBuLi-LiPM
is not quite sensitive to solvent effects in apolar media such as
hexane, it undergoes a remarkable modification in THF. Thus,
∆Geq changes20 from 31.0 kcal/mol in gas phase to 30.8 kcal/
mol in hexane and 17.2 kcal/mol in THF (after correction by
solvent concentration, see Table 6 and eq 7 in ref 20). In such
a case, the concentration of the more reactive dimers is large
enough for this aggregate to become the reactive species and,
consequently, chemoselectivity changes from metalation in
hexane to nucleophilic addition in THF. Likewise, when the
temperature rises in hexane, the dimer/tetramer equilibrium is
shifted toward the dimer because of its largest entropy contribu-
tion (note that in coordinating solvents such as THF, the opposite
would be true).20 If the dimer concentration increase is large
enough, then it can contribute to the whole reactivity of the
superbase so that some addition product is also obtained, as
actually observed (compare reactions at -78 °C and -40 °C
in Table 1).
the experimental findings. Both, the nBuLi base and
nBuLi-LiPM superbase display the following common trends:
dimers are more reactive than tetramers, dimers favor nucleo-
philic addition, and tetramers favor R-metalation. Overall, the
observed R-metalation ability of equimolar mixtures of
nBuLi-LiPM compounds in hexane can be explained by the
simultaneous fulfillment of two requirements: a very low
stability of the (nBuLi)(LiPM) dimer and a sufficiently high
reactivity of the (nBuLi)2(LiPM)2 tetramer. Factors favoring the
[Tet]/[Dim] ratio should therefore favor R-metalation over
nucleophilic addition, thus explaining the dependence of
experimental data on solvent, temperature and reagent/substrate
ratio. The superbase behavior of nBuLi-LiPM aggregates in
hexane has been explained by the stabilizing effect of intramo-
lecular Li · · ·N interactions.
Although further research efforts will be necessary to better
understand chemoselectivity in organolithium processes, the
experimental and theoretical results discussed in the present
work have allowed us to identify some key features. Their
knowledge will be especially helpful to (1) develop new
computational procedures with predictive power and (2) design
new agents to prepare metalated heterocycles with improved
enantioselectivities in the addition to prochiral electrophiles.
3. Conclusions
Acknowledgment. The authors thank the CINES facility,
Montpellier, France, for allocation of computer time (Project
lct2550).
Thepresentstudyhasshownthat,liketheparentBuLi-LiDMAE
reagent, mixed nBuLi-LiPM compounds are very useful
metalating agents to synthesize R-functionalized pyridine de-
rivatives. The chemoselectivity is strongly dependent on the
metalation solvent used. While in apolar hexane, metalation was
favored, in THF, nucleophilic addition occurred. The reagent/
substrate ratio was found to dramatically modify the reaction
pathway. The metalation occurred exclusively providing that
at least 2 equiv. of nBuLi-LiPM were used.
Supporting Information Available: Details on experimental
procedures, kinetics study, complete ref 27, Cartesian coordi-
nates of optimized structures and total energies (reagents,
complexes, transitions structures, IRC calculations), thermo-
chemistry, equations used to compute relative rates. This
information is available free of charge via the Internet at http://
pubs.acs.org.
Quantum chemical computations on transition structures
through dimers and tetramers for different reactions have
allowed us to describe the mechanisms involved and interpret
JA910350Q
9
2416 J. AM. CHEM. SOC. VOL. 132, NO. 7, 2010