MeLi + LiCl in THF: One Heterodimer and no tetramers

Article abstract of DOI:10.1021/jo101282m

The structure of the aggregates formed when mixing methyllithium and lithium chloride in THF has been studied by multinuclear magnetic resonance at 170 K. The data suggest that only one new entity is observed, that is the dimer [(MeLi)(LiCl)], in equilibrium (K ≈ 0.6) with [MeLi]₄ and [LiCl]₂. NMR diffusion measurements lead to the conclusion that this dimer is trisolvated in THF at 170 K, a solvation scheme in agreement with DFT computations.

Full text of DOI:10.1021/jo101282m

pubs.acs.org/joc
MeLi þ LiCl in THF: One Heterodimer and No Tetramers†
Baptiste Lecachey,‡ Hassan Oulyadi,‡ Pedro Lameiras,‡ Anne Harrison-Marchand,‡
§
Helene Gerard, and Jacques Maddaluno*
,‡
ꢀ ꢁ
ꢀ
‡
ꢀ
CNRS UMR 6014 & FR 3038, Universite de Rouen and INSA de Rouen, 76821 Mont St. Aignan Cedex,
France, and UPMC Univ Paris 6 and CNRS, UMR 7616, Laboratoire de Chimie Theꢀorique,
4 place Jussieu, 75252 Paris Cedex 05, France
§
jmaddalu@crihan.fr
Received July 6, 2010
The structure of the aggregates formed when mixing methyllithium and lithium chloride in THF has
been studied by multinuclear magnetic resonance at 170 K. The data suggest that only one new entity
is observed, that is the dimer [(MeLi)(LiCl)], in equilibrium (K ≈ 0.6) with [MeLi]4 and [LiCl]2. NMR
diffusion measurements lead to the conclusion that this dimer is trisolvated in THF at 170 K, a
solvation scheme in agreement with DFT computations.
Introduction
the selectivity.2 The enhanced reactivity of Grignard re-
agents in the presence of LiCl recently reported by Knochel3
(Turbo-Grignard RMgX-LiCl) provides a spectacular re-
cent example of this effect. Even if the interactions between
lithium chloride or bromide and alkyllithiums,4 lithium
amides,5,6 or lithium enolates7 have been studied on chemical,
spectroscopic, or theoretical grounds, the precise phenomena
Because they easily form mixed aggregates, lithium halides
are known to exert a significant influence on the behavior of
organometallic reagents, affecting both the reactivity1 and
† This paper is dedicated to Prof. Saverio Florio on the occasion of his 70th
birthday.
(1) See for instance: (a) Huisgen, R.; Mack, W. Chem. Ber. 1960, 93, 332–
340. (b) Glaze, W.; West, R. J. Am. Chem. Soc. 1960, 82, 4437–4437. (c)
Winkler, H. J. S.; Winkler, H. J. Am. Chem. Soc. 1966, 88, 969–974. (d)
Rathman, T. L.; Bailey, W. F. Org. Process Res. Dev. 2009, 13, 144–151.
(2) See for instance in enantioselective protonation: (a) Yasukata, T.;
Koga, K. Tetrahedron: Asymmetry 1993, 4, 35–38. (b) Yanagisawa, A.;
Kikuchi, T.; Kuribayashi, T.; Yamamoto, H. Tetrahedron 1998, 54, 10253–
10264. (c) Yanagisawa, A.; Kikuchi, T.; Yamamoto, H. Synlett 1998, 174–
(5) (a) Sugasawa, K.; Shindo, M.; Noguchi, H.; Koga, K. Tetrahedron
Lett. 1996, 37, 7377–7380. (b) Toriyama, M.; Sugasawa, K.; Shindo, M.;
Tokutake, N.; Koga, K. Tetrahedron Lett. 1997, 38, 567–570.
(6) (a) DePue, J. S.; Collum, D. B. J. Am. Chem. Soc. 1988, 110, 5518–
5524. (b) DePue, J. S.; Collum, D. B. J. Am. Chem. Soc. 1988, 110, 5524–
5533. (c) Mair, F. S.; Clegg, W.; O’Neil, P. A. J. Am. Chem. Soc. 1993, 115,
3388–3389. (d) 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, 5053–5055.
(e) Hall, P. L.; Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum, D. B.
J. Am. Chem. Soc. 1991, 113, 9575–9585. (f) Romesberg, F. E.; Collum, D. B.
J. Am. Chem. Soc. 1994, 116, 9187–9197. (g) Romesberg, F. E.; Collum, D. B.
J. Am. Chem. Soc. 1994, 116, 9198–9202. (h) Henderson, K. W.; Dorigo,
A. E.; Liu, Q.-Y.; Williard, P. G.; Schleyer, P. v. R.; Bernstein, P. R. J. Am.
Chem. Soc. 1996, 118, 1339–1347. (i) Clegg, W.; Greer, J. C.; Hayes, J. M.;
Mair, F. S.; Nolan, P. M.; O’Neil, P. A. Inorg. Chim. Acta 1997, 258, 1–9.
(j) Pratt, L. M. Bull. Chem. Soc. Jpn. 2005, 78, 890–898.
ꢀ
176. (d) Asensio, G.; Aleman, P. A.; Domingo, L. R.; Medio-Simon, M.
Tetrahedron Lett. 1998, 39, 3277–3280. (e) Asensio, G.; Aleman, P. A.; Gil, J.;
ꢀ
Domingo, L. R.; Medio-Simon, M. J. Org. Chem. 1998, 63, 9342–9347. See
also in enantioselective deprotonation: (f) Bunn, B. J.; Simpkins, N. S. J. Org.
Chem. 1993, 58, 533–534. (g) Majewski, M.; Lazny, R.; Nowak, P. Tetra-
hedron Lett. 1995, 36, 5465–5468. (h) Sugasawa, K.; Shindo, M.; Noguchi,
H.; Koga, K. Tetrahedron Lett. 1996, 37, 7377–7380. (i) Toriyama, M.;
Sugasawa, K.; Shindo, M.; Tokutake, N.; Koga, K. Tetrahedron Lett. 1997,
38, 567–570.
(3) See for instance: (a) Krasovskiy, A.; Straub, B.; Knochel, P. Angew.
Chem., Int. Ed. 2006, 45, 159–162. (b) Clososki, G. C.; Rohbogner, C. J.;
Knochel, P. Angew. Chem., Int. Ed. 2007, 46, 7861–7684. (c) Piller, F. M.;
Appukkuttan, P.; Gavryushin, A.; Helm, M.; Knochel, P. Angew. Chem., Int.
Ed. 2008, 47, 6802–6806. (d) Mosrin, M.; Knochel, P. Org. Lett. 2008, 10,
2497–2500.
(4) (a) Waak, R.; Doran, M. A.; Baker, E. B. Chem. Commun. 1967, 1291–
1293. (b) Novak, P. D.; Brown, T. L. J. Am. Chem. Soc. 1972, 94, 3793–3798.
(c) Kieft, R. L.; Novak, D. P.; Brown, T. L. J. Organomet. Chem. 1974, 77,
(7) (a) Seebach, D. Angew. Chem., Int. Ed. Engl. 1988, 27, 1624–1654. (b)
Juaristi, E.; Beck, A. K.; Hansen, J.; Matt, T.; Mukhopadhyay, T; Simson,
M.; Seebach, D. Synthesis 1993, 1271–1290.
(8) For lithium halide salt effects on the reactivity of organometallics, see:
(a) Chastrette, M.; Amouroux, R. Bull. Soc. Chim. Fr. 1970, 4348–4353. (b)
Loupy, A.; Tchoubar, B. Salt effects in organic and organometallic chemistry;
VCH Publishers: New York, 1992; pp 1-322. (c) Cahiez, G.; Razafintsala-
ma, L.; Laboue, B.; Chau, F. Tetrahedron Lett. 1998, 39, 849–852. (d)
Krasovskiy, A.; Malakohov, V.; Gavryushin, A.; Knochel, P. Angew. Chem.,
Int. Ed. 2006, 45, 6040–6044. (e) Metzger, A.; Schade, M. A.; Knochel, P.
Org. Lett. 2008, 10, 1107–1110. For lithium halide effects on the stereo-
selectivities of organometallics, see: (f) Ye, M.; Logaraij, S.; Jackman, L. M.;
Hillegass, K.; Hirsh, K. A.; Bolliger, A. M.; Grosz, A. L. Tetrahedron 1994,
50, 6109–6116. (g) Asensio, G.; Aleman, P. A.; Domingo, L. R.; Medio-
Simon, M. Tetrahedron Lett. 1998, 39, 3277–3280.
€
299–305. (d) Eppers, O.; Gunther, H. Helv. Chim. Acta 1990, 73, 2071–2082.
(e) Williard, P. G. Compr. Org. Chem. 1991, 1, 1–47. (f) Gunther, H. J. Braz.
€
Chem. Soc. 1999, 10, 241–262. (g) Desjardins, S.; Flinois, K.; Oulyadi, H.;
Davoust, D.; Giessner-Prettre, C.; Parisel, O.; Maddaluno, J. Organometal-
€
lics 2003, 22, 4090–4097. (h) Fox, T.; Hausmann, H.; Gunther, H. Magn.
Reson. Chem. 2004, 42, 788–794.
5976 J. Org. Chem. 2010, 75, 5976–5983
Published on Web 08/02/2010
DOI: 10.1021/jo101282m
r
2010 American Chemical Society

Lecachey et al.
JOCArticle
FIGURE 1. 1H (left), 6Li (middle), and 13C (right) NMR spectra of “pure” Me6Li (top) and Me6Li þ 0.25 equiv of 6LiCl (bottom) in THF-d8
(0.3 M, 170 K).
triggered by the salts are often difficult to assign at the
molecular level and therefore a “black-box type” description
is frequently proposed.5b,8 Expectedly, the prototypical
methyllithium has been the object of particular attention.
Early works by Brown,9 West,10 and others4 afforded an
important corpus of results on the organization of MeLi-LiBr
and MeLi-LiI in ethereal solvents, in particular THF that is
commonly employed with these reagents. The mixtures invol-
ving the two remaining lithium halides, viz. LiF and LiCl, have
never been described to our knowledge.11 If LiF is strictly
insoluble in THF, this void is much more surprising for LiCl,
which is a well-known contaminant of all alkyllithiums pre-
pared by reacting metallic lithium with R-Cl. A better know-
ledge of the species in solution would be especially useful when
dealing with the preparation of “salt-free” methyllithium,12,13
a reagent used commonly for transmetalations and generation
of elaborated organometallics.14 Our previous experience in
the characterization of the MeLi-LiBr aggregates, as well as
those formed between LiCl and LiNR2,15 prompted us to
undertake an in-depth study on the MeLi-LiCl couple thanks
to a set of 6Li and 6Li/13C simple and double-labeling experi-
ments and subsequent NMR and theoretical studies. The
results are reported below.
Results and discussion
Spectroscopic Identification of the [(MeLi)m(LiCl)n] Aggre-
gate(s) in THF. A sample of “salt-free”16 singly labeled
Me6Li in THF-d8 was first prepared13 and the effect of the
6
1
addition of known amounts of LiCl was followed by H,
6Li, and 13C NMR at 170 K (-103 °C).17 The starting sample
of “pure” MeLi is characterized by the expected 1H and 6Li
singlets at δ -2.07 ppm ([CH3Li]4 in Figure 1, top left) and
1.87 ([CH3Li]4 in Figure 1, top middle),18 respectively, plus a
13C heptet at δ -15.05 ppm ([CH3Li]4 in Figure 1, top right;
6
J = 5.8 Hz, in line with a 13C surrounded by three Li for
which I = 1).4c,e,f Minor singlets were also observed at
δ -2.16, -2.17, -2.20, and -2.22 ppm on the 1H spectrum
and at δ 0.95, 0.90, and 0.85 on the 6Li spectrum (Figure 1,
enlargements top line).
When 0.25 equiv of 6LiCl was added to this sample, both the
1H δ-2.22 ppm signal ([CH3Li]complex in Figure 1, bottom left)
6
and the Li δ 0.85 ppm peak ([Li]0.85 in Figure 1, bottom
middle) were amplified. For this later nucleus, a new singlet
also appeared at δ 0.00 ppm ([Li]0.00 in Figure 1, bottom
middle). In the 13C dimension, a small quintet at δ -14.43 ppm
was observed ([CH3Li]complex in Figure 1, bottom right), in
contrast to the heptet reported for LiBr or LiI mixed aggre-
gates. This multiplicity is characterized by a coupling constant
(J ≈ 9.5 Hz, Figure 1, bottom right) significantly larger than in
the LiBr (J ≈ 5.9 Hz)4f-h and LiI (J ≈ 5.8 Hz)4d,f,h analogues.
The multiplicity and the empirical Bauer-Winchester-
Schleyer rule for coupling constants (JLi-C = (17 ( 2)/nC,
nC being the number of lithium cations directly connected to
(9) Novak, D. P.; Brown, T. L. J. Am. Chem. Soc. 1972, 94, 3793–3798.
(10) Glaze, W.; West, R. J. Am. Chem. Soc. 1960, 82, 4437–4437.
(11) The formation of a [Me3ClLi4(THF)4] cubic complex in the solid
state that would not evolve toward [Me2Cl2Li4(THF)4] upon addition of LiCl
has been proposed. However, they could not be observed: Ogle, C. A.;
Huckabee, B. K.; Johnson, H. C., IV; Sims, P. F.; Winslow, S. D.; Pinkerton,
A. A. Organometallics 1993, 12, 1960–1963.
(12) (a) Kamienski, C. W.; Esmay, D. L. J. Org. Chem. 1960, 25, 1807–
1809. (b) Talalaeva, T. V.; Rodinov, A. N.; Kocheshkov, K. A. Dokl. Akad.
Nauk SSSR 1961, 140, 847–850. (c) Lusch, M. J.; Phillips, W. V.; Sieloff,
R. F.; Nomura, G. S.; House, H. O. Org. Synth. 1984, 62, 101–110.
ꢀ
(13) Pate, F.; Oulyadi, H.; Harrison-Marchand, A.; Maddaluno, J.
Organometallics 2008, 27, 3564–3569.
(14) See for instance: (a) Stork, G.; Hudlrik, P. F. J. Am. Chem. Soc.
1968, 90, 4464–4465. (b) House, H. O.; Czuba, L. J.; Gall, M.; Olmstead,
H. D. J. Org. Chem. 1969, 34, 2324–2336. (c) House, H. O.; Gall, M.;
Olmstead, H. D. J. Org. Chem. 1971, 36, 2361–2371. (d) Lautens, M.;
Kumanovic, S. J. Am. Chem. Soc. 1995, 117, 1954–1964. (e) Clayden, J.;
Kenworthy, M. N. Org. Lett. 2002, 4, 787–790.
ꢀ
(15) (a) Corruble, A.; Davoust, D.; Desjardins, S.; Fressigne, C.; Giess-
ner-Prettre, C.; Harrison-Marchand, A.; Houte, H.; Lasne, M.-C.; Madda-
luno, J.; Oulyadi, H.; Valnot, J.-Y. J. Am. Chem. Soc. 2002, 124, 15267–
(16) A total amount of 5-10 mol % chloride was measured by potentio-
metry after hydrolysis of a sample.
(17) Above 170 K, the high-field 6Li signals collapse in a broad signal
centered at 0.4 ppm (see Figure 1S of the Supporting Information).
(18) 6Li chemical shifts are given with respect to an external 0.3 M LiCl
solution in THF.
ꢀ
15279. (b) Pate, F.; Duguet, N.; Oulyadi, H.; Harrison-Marchand, A.;
Fressigne, C.; Valnot, J.-Y.; Lasne, M.-C.; Maddaluno, J. J. Org. Chem.
2007, 72, 6982–6991.
(19) (a) Bauer, W.; Winchester, W. R.; Schleyer, P. v. R. Organometallics
1987, 6, 2371–2379. (b) Bauer, W.; Schleyer, P. v. R. Adv. Carbanion Chem.
1992, 1, 89-175 and references cited therein.
ꢀ
J. Org. Chem. Vol. 75, No. 17, 2010 5977

Lecachey et al.
JOCArticle
FIGURE 3. 1H (left) and 6Li (right) NMR spectra of “pure”
13CH36Li (lines 1 and 2) and 13CH36Li þ 1.0 equiv of LiCl (lines
6
3 and 4) in THF-d8 (0.3 M, 170 K).
The information brought by these experiments implies the
existence of a single mixed aggregate [(MeLi)m(LiCl)n] char-
acterized by only one type of Li cation and only one type of
methyl appendage directly connected to two Li. The com-
plementary information regarding the surrounding lithiums
was furnished by the NMR analysis of doubly labeled
FIGURE 2. 1H (left) and 6Li (right) NMR spectra of “pure” Me6Li
(top) and Me6Li þ x equiv of 6LiCl (x = 0.5 to 2.5) in THF-d8 (170 K).
the observed carbon)19,20 suggest that the only complex formed
exhibits one methyl surrounded by two lithium cations, and is
thus not a mixed cubic tetramer (contaminated by small
amounts of putative mixed dimers) as for LiBr or LiI.
6
13CH3 Li. This compound, never described before to our
knowledge, was synthesized from 13CH3I and n-Bu6Li fol-
lowing Waack and West’s procedure.23 Note that the draw-
back of this method is that residual amounts of LiI are
almost impossible to eliminate, and despite all our efforts to
The simplicity of these spectra, with respect to those
obtained upon addition of LiBr or LiI to MeLi, led us to
refine the study by recording a full set of spectra after the
addition of known quantities of LiCl (between 3 and 71 mol %).
While the 1H and 6Li signals of methyllithium decreased, the
1H signal at δ -2.22 ppm increased continuously, as well as
6
diminish its concentration, we had to work with 13CH3 Li
samples contaminated with 7-20% of this salt. This results
in spectra displaying major signals (in black, Figure 3) that
correspond to the 13 Me6Li/6LiCl complex we are interested
in, and minor peaks (in gray) that represent the undesired 13
Me6Li/6LiI aggregates. Those later could, however, easily be
6
the two Li peaks at δ 0.85 and 0.00 ppm, albeit these two
latter did not evolve at the same rate and their ratio was not
constant. Note that the δ 0.00 ppm signal corresponds to the
“free” LiCl oligomer(s), 0.3 M LiCl in THF being taken as
the external reference. Since no new signal was observed on
the 1H/6Li/13C spectra (Figure 2), whatever the quantity of
LiCl added and in contrast to what happens with LiBr or
LiI,4 it seems that one single new species is formed, char-
assigned thanks to Gunther’s previous results4h (see Figure
€
4S in the Supporting Information).
Let us first describe the spectra related to the “pure”
doubly labeled methyllithium (Figure 3, lines 1 and 2). The
1H spectrum (top line, left) displays the expected doublet at
δ -2.07 ppm (1J = 97 Hz), which becomes a singlet at the
same chemical shift upon 13C irradiation (line 2, left). The 6Li
spectrum (top line, right) consists mainly of a quartet (1J =
5.8 Hz) at δ 1.87 ppm. This multiplicity is characteristic of a
lithium cation surrounded by three 13C as expected for the
well-known cubic arrangement of methyllithium. The quar-
tet turns into the likely singlet at δ 1.87 ppm upon 13C
irradiation (line 2, middle). Complementarily, the 13C-{1H}
spectrum (see Figure 5S in the Supporting Information) is
not the simple heptet observed for 6Li monolabeled methyl-
lithium. An overmultiplicity is observed that is not related to
1
6
acterized by one H and one Li singlet. A bidimensional
HOESY21 run on a 1:2 MeLi/LiCl sample (Figure 2S of the
Supporting Information) supports this hypothesis, a corre-
lation being observed between the new signals.22
(20) This rule applies not only to homogeneous hydrocarbons lithium
aggregates but also to various mixed lithium aggregates (see ref 4).
(21) Bauer, W.; Schleyer, P. v. R. Magn. Reson. Chem. 1988, 26, 827–833.
(22) It is worth noting that the HOESY spectrum shows a meaningless
correlation between the new signal on the 1H dimension and that of LiCl on
the 6Li dimension. This extra correlation is due to the strong exchange that
takes place between LiCl and the new complex, even at 170 K (see the EXSY
spectrum, Figure 3S of the Supporting Information). We warmly thank Prof.
David Collum (Cornell University) for pointing out the risk of misinterpret-
ing these results leading to erroneous conclusions.
(23) McKeever, L. D.; Waack, R.; Doran, M. A.; Baker, E. B. J. Am.
Chem. Soc. 1969, 91, 1057–1061.
5978 J. Org. Chem. Vol. 75, No. 17, 2010

Lecachey et al.
JOCArticle
FIGURE 4. [(MeLi)m(LiCl)n] aggregates accounting for the preliminary NMR observations.
the fluxional character of the aggregate24 but to an addi-
Table 1 show that, for x = 1, the relative proportions
[MeLi]4/[(MeLi)(LiCl)]x calculated from the 1H integrations
are in fine agreement with those obtained from the 6Li
spectra and correlate a regular increase of the proportion
of the complex with the added LiCl. Obviously, exactly the
same correlation would be obtained for x = 2 with different
proportions between [MeLi]4 and [(MeLi)(LiCl)]2.
2
tional JC-C coupling arising between the magnetically
nonequivalent 13C of the cubic arrangement. It transforms
the heptet in an AXX0X00 A0XX0X000 A00XX00X000 A000X0X00X000
2
system. A spectrum simulation25 using a JC-C coupling in
the 2.5-3.0 Hz range gives a very good account for this
observation (Figure 4S, Supporting Information). Note that
this value is in line with those computed for the unsolvated
optimized methyllithium 13C4 H126Li4 tetramer (2JC-C
Discriminating between two highly symmetrical aggre-
gates such as A and F by NMR is not easy. NMR diffusion
experiments (DOSY), a technique that has recently been
applied with success to organolithium species,27 can even-
tually give access to useful information about the hydro-
dynamic radius of solvated species and therefore to their
molecular weights. We thus ran a 1H DOSY experiment on a
mixture of Me7Li and 7LiCl in the presence of three internal
references (trimethyl-, tri-tert-butyl-, and triphenylbenzene
chosen for their solubility, distinct resonance frequencies,
and lack of reactivity toward the lithiated species considered
here) in THF at 170 K.28 The fine linear correlation (R2 =
0.9995, Figure 5 and Supporting Information, S13) observed
between the logarithm of the diffusion coefficients and the
logarithm of the formula weights of the references provides
an equation that can be used to interpolate the formula
weight of the solvated mixed aggregate on the basis of the
diffusion coefficients. Note that THF was not taken as an
internal reference since the signal we observe results from the
rapid exchange between molecules coordinated to the
lithiated species in the solution and molecules in the bulk.
The result obtained (FW = 277) suggests that we are
dealing either with [A] [THF]2.7 or [F] [THF]1.9. These
1
=
3.3-3.6 Hz, see Computational Details in the Supporting
Information, S13). Addition of 1 equiv of 6LiCl to the “pure”
1
13Me6Li sample led to a H spectrum showing two main
doublets (Figure 3, line 3, left): one corresponds to the signal
of the free methyllithium at δ -2.07 ppm while the second
matches with the [(MeLi)m(LiCl)n] peak at δ -2.22 ppm. The
6
13C-6Li coupled Li spectrum (line 3, right) exhibits three
signals that correspond to (i) the free methyllithium quartet
at δ 1.87 ppm,26 (ii) a doublet at δ 0.85 ppm (J = 9.6 Hz)
consistent with a lithium coupled to one carbon only, and
(iii) a singlet at δ 0.00 ppm (LiCl).
Knowing that the [(MeLi)m(LiCl)n] complex under inves-
tigation contains a single type of Me surrounded by two
equivalent lithium cations and a single type of Li bounded to
only one methyl group allows us to sort among all possible
aggregates exhibiting dicoordinated lithium cations (Figure 4).
These characteristics eliminate not only the cubic tetramers
(not represented in Figure 4) but also the trimeric complexes
[(MeLi)(LiCl)2] (B) and [(MeLi)2(LiCl)] (C) for which two Li
signals are expected, and the three pseudo-C2v “planar”
tetrameric structures (open cubic tetramers) which are
[(MeLi)3(LiCl)] (D), [(MeLi)2(LiCl)2] (E), and [(MeLi)-
(LiCl)3] (G), presenting two to three different Li. Finally,
only the mixed dimer A and pseudo-D2h tetramer [(MeLi)2-
(LiCl)2] (F) account for all the spectroscopic data. The
unlikely larger oligomers have not been considered.
3
3
fractional solvations can be interpreted as the average num-
ber of THF molecules that remain coordinated to the cations
all along the diffusion delay and are therefore less than or
equal to the “maximum” value of 4 and 8 THF molecules,
respectively. If the value of 2.7 THF associated to A can be
understood as a consequence of the lesser coordination
energy of the second ether molecule to the same lithium
The consistency of the information we had in hand at this
stage was checked by a relative integration of the peaks in 1H
and 6Li dimensions. They both fit the ratios imposed by their
assignment to the [(MeLi)(LiCl)]x complexes A (x = 1) or F
(x = 2) and to the free tetrameric methyllithium. The data in
(27) For a recent review about the determination of molecular size in
solution by DOSY experiments, see: (a) Macchioni, A.; Ciancaleoni, G.;
Zuccaccia, C.; Zuccaccia, D. Chem. Soc. Rev. 2008, 37, 479–489. For
applications to organolithium derivatives, see: (b) Li, D.; Keresztes, I.;
Hobson, R.; Williard, P. G. Acc. Chem. Res. 2009, 42, 270–280. (c) Lecachey,
(24) For leading references on the fluxionality of organolithium deriva-
tives, see: (a) Fraenkel, G.; Fraenkel, A. M.; Geckle, M. J.; Schloss, F. J. Am.
Chem. Soc. 1979, 101, 4745–4747. (b) Fraenkel, G.; Henrichs, M.; Hewitt,
J. M.; Su, B. M.; Geckle, M. J. J. Am. Chem. Soc. 1980, 102, 3345–3350.
(25) Spectra simulations were performed with the Daisy Bruker software.
See the Supporting Information.
ꢀ
B.; Duguet, N.; Oulyadi, H.; Harrison-Marchand, A.; Fressigne, C.; Yamamoto,
Y.; Tomioka, K.; Maddaluno, J. Org. Lett. 2009, 11, 1907–1910. (d) Kagan, G.;
Li, W.; Hopson, R.; Williard, P. G. Org. Lett. 2010, 12, 520–523.
(28) The large tailing on the DOSY signals presented cannot be sup-
pressed at the low temperature of these experiments. However, this phenom-
enon has no effect on the accuracy of the diffusion coefficients since the
figures are generated by using curve fitting of signal attenuation data values
and are not taken directly from the spectra (see the Supporting Information).
(26) In contrast to the 13C spectrum of MeLi, no 2J coupling constant is
observed as 6Li-6Li 2J coupling is very small, as confirmed computationally
(0.07 Hz)
J. Org. Chem. Vol. 75, No. 17, 2010 5979

Lecachey et al.
JOCArticle
TABLE 1. Relative Proportions of the Protons and Lithiums in [MeLi]4 and [(MeLi)(LiCl)] As a Function of LiCl Proportion
1H spectra
6Li spectra
a
c
/
equiv of
LiCl
integrations δ-[MeLi]4
/
rel proportions
(%)
rel integrations δ-[MeLi]4
δ-[(MeLi)(LiCl)]d
rel proportions average
(%)
δ-[(MeLi)(LiCl)]b
(%)
( (%)
0.00
0.25
0.50
0.75
1.00
1.50
2.00
2.50
162.3/8.4
82.8/17.2
56.3/43.7
41.3/58.7
32.0/68.0
24.8/75.2
14.8/85.2
8.5/91.5
1272.9/108.9
879.9/282.9
1122.9/689.3
1080.3/1016.8
872.5/1132.8
356.9/886.5
123.9/575.8
49.5/414. 0
85.4/14.6
60.9/39.1
44.9/55.1
34.7/65.3
27.8/72.2
16.8/83.2
9.7/90.3
84.1/15.9
58.6/41.4
43.1/56.9
33.3/66.7
26.3/73.7
15.8/84.2
9.1/90.9
1.3
2.3
1.8
1.4
1.5
1.0
0.6
0.1
282.1/54.7
296.4/105.4
136.4/72.6
128.7/97.7
76.8/110.2
38.8/104.1
19.2/83.5
5.4/94.6
5.6/94.4
5.5/94.5
a 1H peak at δ -2.07 ppm. b 1H peak at δ -2.22 ppm. c 6Li peak at δ þ1.87 ppm. d 6Li peak at δ þ 0.85 ppm.
FIGURE 5. 1H DOSY NMR experiment (THF, 170 K) of [(MeLi)(LiCl)]x in the presence of three internal references (C6H3Ph3, C6H3t-Bu3,
and C6H3Me3). The bidimensional plot on the left was recorded at 195 K for clarity. TMB: 1,3,5-trimethylbenzene; TTBB: 1,3,5-tri-tert-
butylbenzene; TPB: 1,3,5-triphenylbenzene.
and thus of the third molecule in the trisolvated dimer,29 the
1.9 THF coordination proposed for F is not very realistic as it
requires a full desolvation of two dicoordinated lithium
cations. Nevertheless, and to further discriminate between
these two structures, we have undertaken a DFT geometry
(which is unlikely in view of the theoretical results given in
Table 2S in the Supporting Information, S16), these hypo-
solvated structures exhibit two types of lithium atoms con-
nected to different ligands (“central” lithium connected to
two chloride and a methyl and terminal lithium connected to
one chloride and one methyl, see Figure 6) that would not fit
the NMR data. Finally, no octagonal structure could be
optimized for n = 4, and only a cubic structure was localized.
All this led us to disregard the tetramer F.
We next considered the mixed dimer A and examined its
solvation state, also from a computational point of view.
Explicit THF molecules were used this time to increase the
accuracy of the computations and the model considered was
optimization of the [F] [Me2O]n series of complexes (with
3
n = 0, 2, 3, 4) at a theoretical level described in the Supporting
Information (S14). If F alone (n = 0) could be obtained as a
planar “octagon” (Figure 6, left) relatively similar to the
schematic structure of Figure 4, its partial solvation (n = 2)
triggers distortions at the lithium cations such that the
valence angles get close to the 120° associated to a tricoordi-
nation. A ladder structure is then obtained (Figure 6, center
left). Increasing n to 3 leads to a more stable structure when
the solvent is added on one of the unsolvated lithium cations.
This structure exhibits additional out-of-plane distortions
(Figure 6, center right). Provided they are favored in THF
[A] [THF]n with n = 2-4. In these conditions, going from
3
n = 2 to 3 is associated to -2.6 kcal mol-1 reaction free
3
energy whereas going from n = 3 to 4 leads to ΔrG0 = þ0.7
kcal mol-1. These data are totally coherent with an average
3
trisolvation for A associated to partial decoordination dur-
ing the NMR measurement.
(29) See for instance: (a) Yuan, Y.; Desjardins, S.; Harrison-Marchand,
ꢀ
A.; Oulyadi, H.; Fressigne, C.; Giessner-Prettre, C.; Maddaluno, J. Tetra-
Finally, the consistency of this information leads us to
propose that the mixed aggregate formed when mixing MeLi
and LiCl in THF is [(MeLi)(LiCl)] [THF]3. Note that cases
hedron 2005, 31, 3325–3334. (b) Pratt, L. M.; Johns, D.; Sease, A.; Busch, D.;
Faluade, E.; Nguyen, S. C.; Thanh, B. T. Int. J. Quantum Chem. 2009, 109,
34–42.
3
5980 J. Org. Chem. Vol. 75, No. 17, 2010

Lecachey et al.
JOCArticle
FIGURE 6. Evolution of the optimized geometry of [(MeLi)2(LiCl)2].(OMe2)n. Left: Planar form in the case n = 0. Center left: Ladder form
for n = 2. Center right: Out-of-plane distortion for n = 3. Right: Cubic form and only optimized structure for n = 4. Color code: lithium (blue),
chlorine (yellow), carbon (green), oxygen (red), hydrogen (gray).
TABLE 2. Value of K Calculated at Various Proportions of (LiCl)2
(concentrations in molcdt.L-1
all been evaluated. We found that only for (LiCl)2 is K
reasonably constant (≈0.65, Table 2) on the range of LiCl
concentrations studied here. The equation considered for the
equilibrium is:
)
equiv of LiCl added
[(MeLi)4]
[A]
[(LiCl)2]
K
0
0.1098
0.0761
0.0552
0.0414
0.0317
0.0182
0.0101
0.0057
0.0208
0.0539
0.0730
0.0829
0.0890
0.0974
0.1005
0.0974
0.0072
0.0285
0.0551
0.0804
0.1050
0.1542
0.2070
0.2490
0.42
0.61
0.64
0.65
0.65
0.67
0.70
0.71
0.25
0.5
0.75
1
1.5
2
1
1=4 ðMeLiÞ4 þ =2ðLiClÞ2 / ½ðMeLiÞðLiClÞꢀ
The corresponding K constants have been calculated on the
basis of the concentration of [MeLi]4, A, and [LiCl]2 esti-
mated from the integration of the NMR signals averaged in
Table 1. The fact that K < 1 shows that the equilibrium does
not favor the formation of the mixed dimer.
2.5
of di-, tri-, and tetrasolvation by THF have been reported
before for heterogeneous and homogeneous dimers.27b
Calculation of an Equilibrium Constant. Comparing the
MeLi/LiCl system to the MeLi/LiBr or MeLi/LiI couples
points the finger at an important difference that is the
persistent signal of [MeLi]4 that is observed even in the
presence of 2.5 equiv of LiCl. This signal quickly disappears
in the presence of LiBr. An equilibrium between the methyl-
lithium, the lithium chloride, and now well-identified dimer
A is thus taking place and we decided to evaluate its constant
K. This requires a proper definition of the interacting species.
If methyllithium is known to adopt preferentially a tetra-
meric form in THF, the situation is not as clear-cut for LiCl.
From vapor phase osmometric studies, Wong and Popov
have proposed that LiCl forms a dimer at 310 K (37 °C) in
Theoretical Investigation on the Difference between LiCl
and LiBr Behaviors. The difference between the size of the
oligomers resulting from a lithium chloride or bromide can
be understood within the perspective of a simple “like yields
like” principle:34 tetrameric lithium bromide yields tetra-
meric mixed aggregates whereas dimeric lithium chloride
yields dimeric mixed aggregates. The origin and generality of
this simple and useful empirical rule deserved to be explored.
The thermodynamic cycles described in Figure 7 were de-
signed in that aim. Because this rule applies to both chloride
and bromide in the same solvent, solvation can hardly be a
factor of importance here, and thus the solvent effects were
not taken into account. We decomposed the transformations
tetramers f mixed tetramers (eq 1, Figure 7) and dimers f
mixed dimers (eq 2, Figure 7) into two formal sequences
tetramer f dimers f monomers for LiX (X = Cl or Br).
Along these sequences, the energies of the elementary steps
differ marginally between the bromide and chloride, except
for two of them. The similarity includes the mixing steps, that
is eqs 1 and 2, which are quasi-isoenergetic for chloride and
7
this solvent.30 This result was supported by Li NMR data
published later on by Reich and co-workers.31 On the other
hand, a separated ion pair was proposed by Pregosin and
colleagues on the basis of PGSE diffusion NMR
spectroscopy.32 In the presence of HMPA, a tetrasolvated
tetramer (LiCl)4(HMPA)4 was evidenced in the solid state,33
while a solvated LiCl(HMPA) or LiCl(HMPA)2 monomer
was proposed in THF/HMPA.31
bromide (always less than 2 kcal mol-1 in absolute energy).
3
As a consequence, the energy of formation of mixed aggre-
gates by condensation of two homogeneous aggregates of the
same size is not sensitive to the nature of the halide: the
energy of formation of the mixed tetramer (eq 1) and of the
The K constants derived from the various equations
involving LiCl as a monomer, a dimer, or a tetramer have
mixed dimer (eq 2) differ by less than 0.5 kcal mol-1 when
(30) Wong, M. K.; Popov, A. I. J. Inorg. Nucl. Chem. 1972, 34, 3615–
3
3622.
(31) Reich, H. J.; Borst, J. P.; Dykstra, R. R.; Green, D. P. J. Am. Chem.
Soc. 1993, 115, 8128–8741.
changing Cl into Br. The main difference between Cl and Br
lies within the formation of the (Li4MenX4-n) tetramers from
ꢀ
(32) Fernandez, I.; Martınez-Viviente, E.; Breher, F.; Pregosin, P. S.
´
Chem.;Eur. J. 2005, 11, 1495–1506.
(33) Barr, D.; Clegg, W.; Mulvey, R. E.; Snaith, R. J. Chem. Soc., Chem.
Commun. 1984, 79–80.
(34) Liou, L. R.; McNeil, A. J.; Ramirez, A.; Toombes, G. E. S.; Gruver,
J. M.; Collum, D. B. J. Am. Chem. Soc. 2008, 130, 4859–4868.
J. Org. Chem. Vol. 75, No. 17, 2010 5981

Lecachey et al.
JOCArticle
FIGURE 7. Aggregation of MeLi and LiX in homogeneous and heterogeneous dimers and tetramers. Energies are in kcal mol-1, top values
are for X = Cl and values in parentheses for X = Br. Bold figures are used to emphasize steps where chloride and bromide differ by more than
3
5 kcal mol-1
.
3
homogeneous or heterogeneous dimers. In both cases, the
formation of the bromide-containing tetramer is more exoe-
This finding, unexpected for a MeLi/LiX couple, necessi-
tated the combination of multinuclear low-temperature
NMR spectroscopy on monolabeled and doubly labeled
nergetic by nearly 7 kcal mol-1 than those containing
chloride. As a consequence, if the Li4X4 fragmentation into
3
6
13CH3 Li combined with quantum computations despite
two Li2X2 is more endothermic (by 6.9 kcal mol-1) for
the simplicity of the partners involved. Our results also
suggest that (LiCl)2 is the stable oligomer in THF at low
temperature. The occurrence of a single dimer contrasts
sharply with the multiple compact mixed cubic tetramers
observed for the MeLi/LiBr or LiI systems11 and could be at
the molecular origin of the different effects induced by these
salts in organic, inorganic, and organometallic chemistry.
Smaller aggregates and low-coordination complexes in gen-
eral are expected to be more reactive, even if the correlation of
aggregation state and organolithium reactivity is based more
on consensus than on experiment,36 and the effect of LiCl on
the reactivity has been clearly evidenced in many cases.37
3
Li4Br4 than for Li4Cl4 (þ27.1 vs. þ20.2 kcal mol-1), this
3
difference is counterbalanced by the formation of the mixed
tetramer as the combination between two Li2XMe mixed
dimers is more exothermic (by 6.6 kcal mol-1) for Li2BrMe
3
than for Li2ClMe (-45.1 vs -38.5 kcal mol-1). Thus, not
3
only is the dissociation Li4Br4 f 2Li2Br2 disfavored with
respect to Li4Cl4 f 2Li2Cl2 but the recombination 2Li2MeBr
f Me2Br2Li4 is more favorable than the corresponding
2Li2MeCl f Me2Cl2Li4. Such behavior can be generalized
to various compositions of dimer, trimers, and tetramers (see
Figure 6S in the Supporting Information for details).
The tendency for bromide to make higher order aggregates
can be directly correlated to the molecular characteristics of
the lithium halides. The computed dipole moments for LiCl
and LiBr are very similar (7.16 D for Cl and 7.12 D for Br):
the formation of dimers by interaction of two monomers is
thus isoenergtic for Cl and Br. Since the dimers have no
dipole moments, their association into tetramer should be
related to the interactions between quadrupoles.35 Actually,
these latter are significantly larger for LiBr (-45.8, -30.1,
Experimental Section
[6Li]-Methyllithium Salt-Free Solution in Ether.12a,13,15a Finely
cut 6-lithium metal ribbon (0.5 g, 83 mmol), 0.5% (weight) of
sodium (ca. 2,5 mg, 0,11 mmol), and three small pieces of broken
glass were introduced into a two-necked pear-shaped flask
(50 mL) equipped with a glass stopper and a condenser fitted
with a balloon of dry argon. The metallic cuttings were covered
with octadecane (10 mL) and the solution was heated (reflux of
octadecane: 317 °C) with a hot air gun with vigorous stirring.
When a maximum amount of the lithium was melted, the flask
was placed in a cold bath (-40 °C) allowing the lithium to
precipitate as a fine shiny shot. The octadecane was extracted
with freshly distilled heptane (10 mL), using a syringe. After
intensive stirring, the heptane was removed and the metal was
washed twice with this same solvent. Diethyl ether was syringed
and the condenser was quicklyreplaced by a CO2 condenser fitted
-62.0 D A) than for LiCl (-33.3, -12.3, -49.7 D A): this
3
3
leads to larger interaction energies between LiBr dimers.
This can explain why cubic tetrameric arrangements are
always preferred with LiBr; this constraint is not as compul-
sory for LiCl.
Conclusions
This paper shows that a mixture of methyllithium and
lithium chloride in THF is in equilibrium (K ≈ 0.65) with a
single mixed dimer [(MeLi)(LiCl)], never described before.
Our data also suggest that this complex is trisolvated in THF.
(36) Collum, D. B. Acc. Chem. Res. 1992, 25, 448–454.
(37) See for instance: (a) Reich, H. J.; Sikorski, W. H.; Gudmundsson,
€
B. O.; Dykstra, R. R. J. Am. Chem. Soc. 1998, 120, 4035–4036. (b) Quirk,
R. P.; Lee, Y. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 145–151.
ꢀ ^
(c) Zune, C.; Archambeau, C.; Dubois, P.; Jerome, R. J. Polym. Sci., Part A:
Polym. Chem. 2001, 39, 1774–1785. (d) Hodgson, D. M.; Miles, T. J.;
Witherington, J. Tetrahedron 2003, 59, 9729–9742. (e) Gupta, L.; Hoepker,
A. C.; Singh, K. J.; Collum, D. B. J. Org. Chem. 2009, 74, 2231–2233.
ꢀ
(f) Ocejo, M.; Carrillo, L.; Badıa, D.; Vicario, J. L.; Fernandez, N.; Reyes, E.
´
J. Org. Chem. 2009, 74, 4404–4407.
(35) We also carried out an analysis in terms of distorsions as described by
ꢀ
ꢀ
Fressigne et al. ( Fressigne, C.; Maddaluno, J.; Giessner-Prettre, C. J. Chem.
Soc., Perkin Trans. 2 1999, 2197–2201.) but did not find any difference
between bromide and chloride.
5982 J. Org. Chem. Vol. 75, No. 17, 2010

Lecachey et al.
JOCArticle
with a balloon of dry argon. Chloromethane (2,3 mL, 41,
5 mmol) was condensed directly from the sealed cylinder to a
graduated trap at -40 °C and then added very slowly (2,3 mL were
added over a period of 45 min) connecting the trap to the top of the
CO2 condenser. The formation of a gray salt corresponding to LiCl
was observed and the disappearance of the 6Li metal was noticed.
After replacing the CO2 condenser with a septum, the resulting
reaction mixture was stirred for 20 h at room temperature under dry
argon. The stirring was stopped allowing LiCl to settle. The ethereal
solution was then pumped off the flask with a syringe and directly
inserted into centrifugation tubes placed under dry argon. The
residual traces of salt were centrifuged and the clear final solution
was collected in a dry flask flushed under dry argon then titrated38
(1.4 M, 55% yield) and kept until further use.
solution (C = 5 ꢁ 10-3 M) with a Mettler DL 21 Titrator fitted
with a silver electrode.
[6Li],[13C]-Methyllithium Salt-Free Solution in Tetrahydrofuran-
d8.23 Labeled [13C]-methyl iodide (0.22 mL) and hexane (2 mL)
were introduced into a centrifugation tube placed under dry argon
and equipped with a magnetic stirrer. An equimolar amount of
[6Li]-n-butyllithium in hexane solution was added at -78 °C. The
temperature of the reaction mixture was raised from an initial -78
to -30 °C (until cloudiness developed, indicating methyllithium
precipitation) and finally to 0 °C over a 20 min period. The
precipitated [6Li],[13C]-methyllithium was centrifuged and the
hydrocarbon solution, containing residual methyl iodide and n-
butyl iodide, was pumped off the tube. The solid residue (labeled
MeLi) was washed twice with hexane (4 mL), then placed under
vacuum (1 mmHg) for 4 h to evaporate a maximum of the
hydrocarbon solvent. The resulting white solid was dissolved in
freshly distilled tetrahydrofuran-d8 and concentrated under
vacuum for another hour. THF-d8 (2.5-3.0 mL) was finally added
and the resulting solution was titrated38 (0.7-1.0 M). Note the
presence of lithium iodide (5-10%) in the solution.
[6Li]-Methyllithium “Salt-Free Solution” in Tetrahydrofuran-
d8.23 A solution of [6Li]-methyllithium in ether prepared above
(2.5 mL) was syringed into a tube fitted with a septum and
flushed under dry argon. The tube was then placed under
vacuum (20 mmHg) for 1 h to evaporate the ether. The resulting
white solid was then dissolved in freshly distilled tetrahydrofur-
an-d8 and concentrated under vacuum for 1 h to evaporate the
last traces of ether. THF-d8 (3 to 3.5 mL) was finally added and
the resulting solution was titrated (0.5 to 0.7 M).
ꢁ
Acknowledgment. B.L. acknowledges the Ministere de La
Recherche et de la Technologie for a Ph.D. grant. We thank
[6Li]-Lithium Chloride Solution in Tetrahydrofuran-d8. Resi-
dual [6Li]-lithium chloride from the [6Li]-methyllithium synth-
esis described above was rinsed three times with diethyl ether
(3 ꢁ 5 mL). Centrifugations were conducted for this operation.
The solvent was finally removed and the tube was placed under
vacuum (20 mmHg) for 4 h to eliminate a maximum of the
diethyl ether. The resulting white solid was dissolved in freshly
distilled tetrahydrofuran-d8 and the solution was concentrated
under vacuum for 1 h to evaporate the last traces of diethyl
ether. THF-d8 (2 mL for about 50 mg of “dry” LiCl) was finally
added and left for a night at room temperature to completely
dissolve the lithium salt. The resulting solution was titrated by a
classical potentiometric technique (0.3 to 0.6 M).
ꢀ
the Region Haute-Normandie for its support to this research
through the CRUNCh interregional network. Calculations
have been performed on supercomputers at CRIHAN
(Saint-Etienne-du-Rouvray, France), CINES (Montpellier,
France), and IDRIS (Orsay, France). The Agence Nationale
pour la Recherche supported this work (grant ANR-07-
BLAN-0294-01). Dr. Olivier Parisel (LCT, Paris VI) is
warmly thanked for his continuous interest in this program
and helpful discussions. We also are extremely appreciative
of Prof. David B. Collum (Cornell University) for his critical
reading of the preliminary version of this paper.
Silvermetric Titration of [6Li]-Lithium Chloride Solution in
Tetrahydrofuran-d8. A solution of [6Li]-lithium chloride in
tetrahydrofuran-d8 (50 μL) was placed in a beaker and diluted
with water (75 mL). The solution was titrated by a silver nitrate
Supporting Information Available: Computational, experi-
mental, and instrumental details, supplementary spectra, as well
as computed binding energies and relative stabilities of the
mixed tetramers. This material is available free of charge via
the Internet at http://pubs.acs.org.
(38) Duhamel, L.; Plaquevent, J. C. J. Organomet. Chem. 1993, 448, 1–3.
J. Org. Chem. Vol. 75, No. 17, 2010 5983