Microsolvation and 13C-Li NMR coupling

Article abstract of DOI:10.1021/ja8026828

The empirical expression ¹J_CLi = L[n(a + d)] ^-1 is proposed; it claims a reciprocal dependence of the NMR coupling constant ¹J(¹³C, Li) in a C-Li compound on two factors: (i) the number n of lithium nuclei in bonding contact with the observed carbanion center and (ii) the sum (a + d) of the numbers a of anions and dof donor ligands coordinated at the Li nucleus that generates the observed ¹J_CLi value. The expression was derived from integrations of separate NMR resonances of coordinated and free monodentate donor ligands (t-BuOMe, Et₂O, or THF) in toluene solutions of dimeric and monomeric 2-(α-aryl-α-lithiomethylidene)-1,1,3,3-tetramethylindan at moderately low temperatures. This unusually slow ligand interchange is ascribed to steric congestion in these compounds, which is further characterized by measurements of nuclear Overhauser correlations and by solid-state structures of the dimers bearing only one donor per lithium atom (d = 1). Increasing microsolvation numbers dare also accompanied by typical changes of the NMR chemical shifts δ (positive for the carbanionic ¹³C ^α, negative for C^para and p-H). The aforementioned empirical expression for ¹J_CLi appears to be applicable to other cases of solvated monomeric, dimeric, or tetrameric C-Li compounds (alkyl, alkenyl, alkynyl, and aryl) and even to unsolvated (d ≈ 0) trimeric, tetrameric, or hexameric organolithium aggregates, indicating that ¹J_CLi might serve as a tool for assessing unknown microsolvation numbers. The importance of obtaining evidence about the ¹³C NMR C-Li multiplet splitting of both the nonfluxional and fluxional aggregates is emphasized.

Full text of DOI:10.1021/ja8026828

Published on Web 10/02/2008
Microsolvation and 13C-Li NMR Coupling†
Rudolf Knorr,* Thomas Menke, Kathrin Ferchland, Johann Mehlsta¨ubl, and
David S. Stephenson
Department of Chemistry and Biochemistry, Ludwig-Maximilians-UniVersita¨t,
Butenandtstrasse 5-13, 81377 Mu¨nchen, Germany
Received April 17, 2008; E-mail: rhk@cup.uni-muenchen.de
1
Abstract: The empirical expression JCLi ) L[n(a + d)]-1 is proposed; it claims a reciprocal dependence
of the NMR coupling constant 1J(13C, Li) in a C-Li compound on two factors: (i) the number n of lithium
nuclei in bonding contact with the observed carbanion center and (ii) the sum (a + d) of the numbers a of
1
anions and d of donor ligands coordinated at the Li nucleus that generates the observed JCLi value. The
expression was derived from integrations of separate NMR resonances of coordinated and free monodentate
donor ligands (t-BuOMe, Et2O, or THF) in toluene solutions of dimeric and monomeric 2-(R-aryl-R-
lithiomethylidene)-1,1,3,3-tetramethylindan at moderately low temperatures. This unusually slow ligand
interchange is ascribed to steric congestion in these compounds, which is further characterized by
measurements of nuclear Overhauser correlations and by solid-state structures of the dimers bearing only
one donor per lithium atom (d ) 1). Increasing microsolvation numbers d are also accompanied by typical
changes of the NMR chemical shifts δ (positive for the carbanionic 13CR, negative for Cpara and p-H). The
aforementioned empirical expression for 1JCLi appears to be applicable to other cases of solvated monomeric,
dimeric, or tetrameric C-Li compounds (alkyl, alkenyl, alkynyl, and aryl) and even to unsolvated (d ≈ 0)
1
trimeric, tetrameric, or hexameric organolithium aggregates, indicating that JCLi might serve as a tool for
assessing unknown microsolvation numbers. The importance of obtaining evidence about the 13C NMR
C-Li multiplet splitting of both the nonfluxional and fluxional aggregates is emphasized.
Chart 1
Introduction
Microsolvation of organolithium compounds (RmLim) by
coordinating electron-pair-donating ligands (“explicit Do”) is
essential1 for the interpretation of reactivity. As shown in Chart
1, C-Li compounds may be monomeric (1) or aggregated in
dimers (2 and 3), tetramers (4), trimers (5), hexamers (6), or
other types2 of clusters.3 Because three and four are the usual
coordination numbers of lithium cations in RLi, it may be
guessed how many donor ligands (Do) would probably be
required to satisfy the coordination sphere of lithium: usually
three (and only occasionally two4 or four5,6) Do per Li in
monomers R1Li1 (1) and not more than one Do per Li in
tetramers R4Li4 (4). On the other hand, hexamers R6Li6 (6) and
one7 of the very rare trimers R3Li3 of type 5, namely, the trimer
of 8 in Chart 2, are the stable forms in noncoordinating solvents.
Predictions4 of one or two Do per Li in the dimers R2Li2Do2
(2) and R2Li2Do4 (3) are less straightforward. Unfortunately,
all of the Li-Do bonds in solution can be labile, with ligand-
interchange processes that are fast on the NMR time scales.
Therefore, it is usually not possible to differentiate between
coordinated (nonchelating) and noncoordinated (free) mono-
dentate ligands by NMR spectroscopy, because only averaged
spectra can be observed at all accessible temperatures. The very
strong donor ligand hexamethylphosphoramide (HMPA) is an
exception8 to this rule, as are monodentate ethereal ligands such
as tetrahydrofuran (THF) when coordinating to the more polar
† Sterically Congested Molecules, 20. For Part 19, see ref 64.
(1) Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1997, 119, 5573–
5582 (specifically p 5581).
(2) (a) Setzer, W. N.; Schleyer, P. von R. AdV. Organomet. Chem. 1985,
24, 353–451. (b) Weiss, E. Angew. Chem. 1993, 105, 1565–1587 and
Table 2 therein. (c) Weiss, E. Angew. Chem., Int. Ed. Engl. 1993, 32,
1501–1523.
(3) It should be noted that the lines drawn between Li and R in Chart 1
symbolize connectivities rather than the usual two-electron bonds.
(4) Pratt, L. M.; Truhlar, D. G.; Cramer, C. J.; Kass, S. R.; Thompson,
J. D.; Xidos, J. D. J. Org. Chem. 2007, 72, 2962–2966.
(5) Lucht, B. L.; Collum, D. B. J. Am. Chem. Soc. 1995, 117, 9863–9874
and references therein.
(6) Jantzi, K. L.; Guzei, I. A.; Reich, H. J. Organometallics 2006, 25,
5390–5395.
(7) Knorr, R.; Hoang, T.-P.; No¨th, H.; Linti, G. Organometallics 1992,
11, 2669–2673.
(8) Reich, H. J.; Kulicke, K. J. J. Am. Chem. Soc. 1996, 118, 273–274
and references therein.
9
10.1021/ja8026828 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 14179–14188 14179

A R T I C L E S
Knorr et al.
Chart 2
Chart 3
1
relation does not agree with the JCLi values observed in this
work (Table 1) or with the 1JCLi value of 11.9 Hz measured in
ref 13 for monomeric tert-butyllithium (t-BuLi) in THF, as was
already noticed by those authors.13,14,16 Further cases of
disagreement have been reported,17 pointing to an omission of
at least one important factor in the 17/n relation. Subsequent
computational studies confirmed18 the inverse dependence of
1JCLi on n but also revealed that coordination of electron-pair
donors at monomeric methyllithium18,19 (MeLi) and at mono-
lithium amides5,9-11 or to the endocyclic Li of dimeric
Me2CuLi·LiCN.12 By use of 31P NMR integrations at temper-
atures below -140 °C, the first8 detection and quantification of
monodentate ether solvates in C-Li compounds became pos-
sible using a monomeric alkyllithium compound bearing one
HMPA ligand, an intramolecularly chelating tertiary amine, and
dimethyl ether or oxetane as the third ligand. The present article
reports on alkenyllithium compounds (8 and 12) for which the
interchange processes of some monodentate ligands are slowed
down to an unusual degree, probably as a result of steric
congestion, enabling microsolvation numbers d to be measured
1
meric t-BuLi18 led to strongly reduced JCLi values, whereas a
continuum solvation model19 produced smaller reductions. Thus,
microsolvation by individual (“explicit”) particles appears to
be the most important missing factor, while the computed 1JCLi
values depend less strongly on the nature of the ligands18,19
and carbanions.18 We have developed an experimental verifica-
tion and quantification of these computational hints in a novel
1
at moderately low temperatures by direct H and 13C NMR
integrations of the separated resonances of free and coordinated
Do. Despite steric congestion, the monomers of 8 and 12 can
aggregate to disolvated dimers of type 2 in both liquid and solid
phases.
1
expression for JCLi that appears to provide a new tool for
assessing unknown microsolvation numbers.
Results and Discussion
At low temperatures, where the C-Li exchange is slow on
the 13C NMR time scale, the resonance of a carbanionic center
in R of 1-6 may split into a (2nI + 1) multiplet caused by
direct bonding contact between the carbanion and n equivalent
1. Characterization of the Donor-Free Cyclotrimer (8a)3
in Solution. The unsolvated 2-(R-lithiobenzylidene)-1,1,3,3-
tetramethylindan (8a, shown in Chart 3) was prepared from
2-(R-bromobenzylidene)-1,1,3,3-tetramethylindan (7) with n-
butyllithium (n-BuLi) in pentane (Chart 2) by a slightly
improved version20 of the published7 procedure. This Br/Li
interchange reaction is remarkable because it proceeds in the
absence of donor ligands; it is also convenient because of its
reliable tendency to produce single crystals that are free from
side products when properly rinsed with pentane. According to
the published7 X-ray diffraction analysis, the crystals consist
of trimeric aggregates of type 5, but the constitution of (8a)3 in
solution had not yet been established and will now be deduced.
When dissolved in [D8]toluene at -84 °C, (8a)3 displays a single
set of quaternary 13C NMR resonances, among them a quintet
6
lithium nuclei. When the Li isotope (nuclear spin quantum
number I ) 1) is chosen for the preparation of RLi, as was
done in the majority of our experiments, one expects (2n + 1)
multiplet components: a monomer 1 (n ) 1) should exhibit a
1:1:1 triplet (see Figure 1 in section 2 of the Results and
Discussion), while a dimer 2 or 3 and also a trimer 5 should
show quintet splitting with the intensity ratio 1:2:3:2:1 (because
6
of n ) 2 equivalent neighboring Li nuclei). For tetramers 4
and hexamers 6, splitting into a septet by n ) 3 nearest Li
neighbors is to be expected. The spacings (frequency intervals)
of these multiplets equal the one-bond coupling constants 1J(13C,
6Li), whose magnitudes were proposed13-15 to obey the very
1
6
absorption of CR (at δ ) 172.2 ppm) with J(13C, Li) ) 8.5
Hz (Table 1, entry 15). This quintet establishes the presence of
two C-Li connections (n ) 2), which is compatible with a
dimer but also with a “nonfluxional” (static) trimer 5, the latter
1
simple “17/n” relation JCLi ) (17 ( 2 Hz)/n irrespective of
the structure and hybridization of the carbanion. However, this
6
without contact of 13CR to the third, more remote Li nucleus
(9) Lucht, B. L.; Collum, D. B. J. Am. Chem. Soc. 1994, 116, 6009–
(3J ≈ 0 over three bonds). Positive evidence of the trimeric
aggregation was achieved simply by warming the sample to a
temperature (-21 °C) at which intraaggregate scrambling of
the three Li cations became sufficiently fast for observation of
6010.
¨
(10) Hilmersson, G.; Davidsson, O. J. Org. Chem. 1995, 60, 7660–7669.
¨
(11) Hilmersson, G.; Ahlberg, P.; Davidsson, O. J. Am. Chem. Soc. 1996,
118, 3539–3540.
(12) Henze, W.; Vyater, A.; Krause, N.; Gschwind, R. M. J. Am. Chem.
Soc. 2005, 127, 17335–17342 and Figure 8 therein.
(13) Bauer, W.; Winchester, W. R.; Schleyer, P. von R. Organometallics
1987, 6, 2371–2379.
(16) Bauer, W.; Feigel, M.; Mu¨ller, G.; Schleyer, P. von R. J. Am. Chem.
Soc. 1988, 110, 6033–6046 (specifically p 6036).
(14) Bauer, W.; Schleyer, P. von R. AdV. Carbanion Chem. 1992, 1, 89–
175 (specifically pp 94 and 98).
(17) For instance, see note 21 in the following: Reich, H. J.; Holladay,
J. E.; Mason, J. D.; Sikorski, W. H. J. Am. Chem. Soc. 1995, 117,
12137–12150.
(15) (a) Lambert, C.; Schleyer, P. von R. Angew. Chem. 1994, 106, 1187–
1199. (b) Lambert, C.; Schleyer, P. von R. Angew. Chem., Int. Ed.
Engl. 1994, 33, 1129–1140.
(18) Koizumi, T.; Kikuchi, O. Organometallics 1995, 14, 987–991.
9
14180 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Microsolvation and 13C-Li NMR Coupling
A R T I C L E S
1
Table 1. Aggregation States, NMR Parameters (∆δ, JCLi) with Pertinent Conditions, Experimental Contact Numbers (n, a, d), and Sensitivity
Factors (L) for Alkenyllithium Compounds 8 and 12
∆δ(13C)a
contact numbersb
c
entry
RLi
solvent
Do (equiv/anion)
T (°C)
CR
Cpara
1JCLi (Hz)
n
a
d d
Tcoll (°C)e
L (Hz)
1
2
3
4
5
6
7
8
(8b)1
(8b)1
(8b)1
(8b)2
(12b)1
(8c)1
t-BuOMe
toluene
toluene
toluene
toluene
Et2O
toluene
THF
toluene
toluene
toluene
THF
t-BuOMe
t-BuOMe (5.0)
t-BuOMe (∼1.5)
t-BuOMe (∼1.5)
t-BuOMe (3.0)
Et2O
Et2O (1.3)
THF
THF (∼50.0)
THF (4.0)
THF (1.0)
THF
-70
-57
-82
-82
-51
-105
-32
-37
-66
-90
-63
-72
-60
-42
-84
+61.3
+61.7
+61.6
+53.3
+61.6
+61.7
+52.4
+66.5
+65.6
+65.8
+52.2
+66.0
+51.8
+59.7
+48.8
-11.8
-11.3
-11.5
-7.3
-16.1
-11.8
-7.4
-13.0
-12.4
-12.6
-7.6
-18.5
-
13.5
13.5
13.5
7.0 g
12.5
14.0
7.3
10.7
10.8
10.8
7.3
1
1
1
2
1
1
2
1
1
1
2
1
2
1
2
1
1
1
2
1
1
2
1
1
1
2
1
2
1
2
(2)
2
2
1
2
(2)
1
(3)
(3)
(3)
1
-60(5)
-35(8)
-18(3)
-
-43(5)
-80(5)
-1(2)
-23(5)
-30(5)
-25(5)
-40(5)
-28(5)
-40(10)
-15(10)
-10(5)
40.5
40.5
40.5
42.0
37.5
42.0
43.8
42.8
43.6
43.2
43.8
40.0
∼41.4
42.6
42.5
f
f
(8c)2
(8d)1
(8d)1
(8d)1
(8d)2
(12d)1
(12d)2
(8e)1
9
10
11
12
13
14
15
10.0
∼6.9
14.2
8.5
(3)
1
toluene
TMEDA
toluene
THF (1.8)
TMEDA
none
-11.1
-5.5
(2)
(8a)3
0.5h
a ∆δ ) δ(RLi) - δ(RH), i.e., it is the δ value relative to that for the protolysis product RH (9 or 13) present in situ (Tables S6 and S7).20 b n ) no.
of Li cations bound directly to 13CR; a ) no. of anionic centers bound directly to Li; d ) no. of Do ligands in contact with Li, determined through 1H
and/or 13C NMR integration as illustrated in Figures 2-10. c 13C-6Li NMR coupling constants. d The d values in parentheses could not be measured for
the monomers but were inferred from the characteristic lithiation shifts ∆δ of CR and Cpara e Tcoll is the minimum temperature of collapse of the 13CR
.
multiplet splitting; values in parentheses are the errors in the measured values. f Monomer/dimer ratio ) 55:45. g Measured at -52 °C. h d ≈ “0.5” for π
plus C-H coordination?20
an averaged 1JCLi value while the interaggregate Li-interchange
reaction was still slow on the 13C NMR time scale;21 the
experimentally determined splitting of 5.9 Hz agrees well with
the expected average (21J + 3J)/3 ) (2 × 8.5 + 0)/3 ) 5.7 Hz
calculated for the “fluxional” trimer. This agreement excludes
a dimer, which would not exhibit such a change, and all higher
oligomers RmLim, because these would have significantly smaller
averaged coupling constants, (2 × 8.5 + 0)/m, through
°C. The considerably broadened, coalesced o-H resonance at δ
) 5.94 ppm is characteristic of (8a)3 at room temperature; it
exhibits the expected doublet splitting (3JHH ) 7.5 Hz) only at
g25 and 105 °C when measured at 80 and 400 MHz,
respectively. During prolonged measurements at 105 °C, (8a)3
in [D8]toluene slowly decomposed with formation of [R-D]-9
(X ) D for 9 in Chart 2), which suggests that (8a)3 was
deuterated by the solvent and may be a stronger base than
benzyllithium in toluene.26
6
scrambling of m g 4 Li cations.
The following observations indicate that (8a)3 in [D8]toluene
solution retains essential traits of the pseudothreefold axial
symmetry of its solid-state structure,7 in which each of the three
8 units is devoid of symmetry because of single Li-Cortho and
Li-CH3 interactions: The only nonequivalences observed for
(8a)3 in toluene from -60 to -96 °C were diastereotopic
splittings (1:1) of the twin types of nuclei, namely, o-H, m-H,
3-CH3, 1-CH3, Cortho, Cmeta, 3-CH3, and 1-CH3, whose chemical
shifts are listed in Tables S10 and S11.20 The large shift
nonequivalences20 of Cortho, 3-CH3, and o-H are compatible with
an accumulation of negative electric charge at the ortho and 3
positions, which are close to the Li cation (as in the crystal7).
As the sample is warmed, the interchange within both the Cortho
and 3-CH3 diastereotopic pairs becomes fast on the 13C NMR
time scale at comparable coalescence temperatures (Tc ≈ -40
°C; Table S11)20 and hence involves similar energy barriers at
these positions. This rules out a simple phenyl rotation but is
compatible with switching of Li between the diastereotopic ortho
and 3-methyl groups. Both 3-CH3 and o-H remain sufficiently
close to 6Li to produce 1H-6Li HOESY22-25 cross peaks at 25
Because (8a)3 is only sparingly soluble in cyclopentane, the
6Li coupling with 13CR in natural abundance could not be
measured in this solvent. Most of the 13C chemical shifts of
(8a)3 in cyclopentane (except for C2 in Table S9)20 are
sufficiently similar to those in [D8]toluene,27 including the
diastereotopic splittings at -84 °C, to allow the conclusion that
only (8a)3 exists and significant amounts of other aggregates
are excluded in cyclopentane solution.
2. Dimeric Aggregates (8b-d)2 and (12b-d)2 are Disolvated.
The Br/Li interchange reaction (Chart 2) of the bromoalkene 7
with n-BuLi in cyclopentane20 and diethyl ether (Et2O, 3 equiv)
furnished another kind of single crystals of 8. X-ray diffraction
analysis (Figure S1)20 revealed a dimeric structure (8c)2 of type
2, disolvated by Do ) Et2O (one per Li) as sketched in Chart
4. The corresponding THF disolvate (8d)2 crystallized with a
similar structure (Figure S2)20 from a saturated toluene solution
of 8 in the presence of Do ) THF (1.5 equiv per Li). The
p-trimethylsilyl derivatives 10-13 were prepared by a different
(19) Parisel, O.; Fressigne´, C.; Maddaluno, J.; Giessner-Prettre, C. J. Org.
Chem. 2003, 68, 1290–1294.
(24) Bauer, W.; Schleyer, P. von R. Magn. Reson. Chem. 1988, 26, 827–
833.
(20) Details are reported in the Supporting Information.
(21) For similar identifications of trimers, see the following: (a) Bauer,
W. J. Am. Chem. Soc. 1996, 118, 5450–5455. (b) Harder, S.; Ekhart,
P. F.; Brandsma, L.; Kanters, J. A.; Duisenberg, A. J. M.; Schleyer,
P. von R. Organometallics 1992, 11, 2623–2627.
(25) Bauer, W. In Lithium Chemistry; Sapse, A.-M., Schleyer, P. von R.,
Eds.; Wiley: New York, 1995; pp 125-172.
(26) Acidity of toluene in THF: (a) pKa ) 40.9: Streitwieser, A.; Ciula,
J. C.; Krom, J. A.; Thiele, G. J. Org. Chem. 1991, 56, 1074–1076.
(b) pKa ) 40.7: Ahlbrecht, H.; Schneider, G. Tetrahedron 1986, 42,
4729–4741.
(22) (a) Bauer, W.; Mu¨ller, G.; Pi, R.; Schleyer, P. von R. Angew. Chem.
1986, 98, 1130–1132. (b) Bauer, W.; Mu¨ller, G.; Pi, R.; Schleyer, P.
von R. Angew. Chem., Int. Ed. Engl. 1986, 25, 1103.
(23) Bauer, W.; Clark, T.; Schleyer, P. von R. J. Am. Chem. Soc. 1987,
109, 970–977.
(27) This contrasts with the interaction of toluene with monomeric
LiN(SiMe3)2 as reported by: Lucht, B. L.; Collum, D. J. Am. Chem.
Soc. 1996, 118, 2217–2225 (specifically p 2224) and references therein.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14181

A R T I C L E S
Knorr et al.
Et2O ligands in the same double-gauche conformation as in (8c)2
(Figure S1).20 In accord with Chart 4, the architecture of (12d)2
(shown in Figure S4)20 is similar to that of (8d)2 in Figure S2,
inclusive of the coordinated THF.
In view of the different crystal systems, the amassed evidence
of architectural similarity (Table S4)20 suggests that this
arrangement is not dominated by crystal effects but should be
characteristic of molecular energetic minima and thus that this
structure may be expected to carry over to the solutions. The
close packing in those dimers, as quantified in Chart 4 through
the short carbon-carbon distances of the R-aryl group from
both 1-Me (320 pm) and 3-Me (380 pm), is confirmed by
NOESY experiments in solution.20 Moreover, the contacts of
Li to the CR-Cipso bond (the C14-C15 bond in Figures
S1-S4),20 one of the two Cortho-H bonds (C16-H16A), and
two of the six C-H bonds of 3-Me are in accord with the results
of 6Li-1H HOESY measurements20 on (8d)2 in toluene solution.
Furthermore, the almost orthogonal orientation of the R-aryl
group with respect to the double bond C2dCR (common to all
four of these crystalline dimers) allows an efficient delocalization
of negative charge from the CR-Li bonds into the π-system of
the R-aryl group (“quasi-benzyllithium” conjugation), with the
consequence of significant upfield NMR shifts ∆δ (see below)
for Cpara (and of R ) H in 8a-e) in solution.
The following 13C NMR parameters were found to provide
reliable evidence for the dimeric aggregation of the alkenyl-
lithium compounds 8b-d and 12b-d in those solutions that
frustrate molecular mass determinations: (i) When labeled with
6Li, these compounds present the 13C resonances of their
carbanionic centers CR as 1:2:3:2:1 quintets with 1JCLi ≈ 7 Hz
at low temperatures, as exemplified in Table 1 (entries 4, 7, 11,
and 13) and displayed in Figure 1a-d. These quintets establish
Figure 1. CR quintets (of [6Li] dimers) and triplets (of [6Li] monomers).
(a) (8b)2 with t-BuOMe (1.5 equiv) in [D8]toluene at -52 °C: 1JCLi ) 6.9
1
Hz. (b) (8c)2 with Et2O (1.3 equiv) in [D8]toluene at -32 °C: JCLi ) 7.3
Hz. (c) (8d)2 with THF (0.85 equiv) in [D8]toluene at -72 °C: 1JCLi ) 7.3
Hz. (d) (12d)2 with THF (1.8 equiv) in [D8]toluene at -60 °C: 1JCLi ) 7.0
1
Hz. (e) (8b)1 with t-BuOMe (5 equiv) in [D8]toluene at -57 °C: JCLi
)
13.5 Hz. (f) (8b)1 in t-BuOMe at -70 °C: 1JCLi ) 13.5 Hz. (g) (8d)1 with
THF (4 equiv) in [D8]toluene at -90 °C: JCLi ) 10.8 Hz. (h) (8d)1 in
1
6
contacts of a CR nucleus with two Li nuclei (i.e., n ) 2) and
1
THF at -37 °C: JCLi ) 10.7 Hz.
persist on warming of the sample [up to at least -1 °C in the
case of (8c)2 (entry 7)] until they collapse through the accelerated
interaggregate Li interchange. Therefore, at least (8c)2 and
presumably also the other dimers do not (and should not) exhibit
Chart 4. Generalized Solid-State Structures of the Disolvated
Dimers (8c)2 and (12c)2, Both with Do ) Et2O, and of (8d)2 and
(12d)2, Both with Do ) THF
1
the JCLi averaging expected from fluxional higher oligomers,
as described above for (8a)3. (ii) In the presence of donor ligands
in amounts suitably limited to avoid causing deaggregation, the
chemical shifts of CR and Cpara are almost independent of
temperature from the domain of frozen Li scrambling (CR
multiplets) up to the range above collapse of the multiplets.
The latter two shift criteria for an unchanged species may be
visualized more clearly (Chart S2)20 in the form of lithiation
shifts ∆δ ) δ(RLi) - δ(RH), where RH (Tables S6 and S7)20
is the protolysis product (9 or 13) inevitably present in situ.
The dimers listed in Table 1 (entries 4, 7, 11, and 13) have
∆δ(CR) ≈ 52 ppm (downfield) and ∆δ(Cpara) ≈ -7.4 ppm
(upfield because of the delocalized negative charge), the
magnitudes of which are significantly smaller than those of the
monomers (treated in section 3) and larger than those of (8a)3
(entry 15). With one or two of these criteria satisfied for
(8b-d)2, (12c)2, and (12d)2, a most unusual property was
discovered close to the lower temperature limit of the liquidity
range of the donor/toluene mixtures: separate (decoalesced)
NMR resonances (1H and 13C) were observed for coordinated
and free Do. This allowed the determination via NMR integra-
tions that 1 equiv of Do per carbanion is coordinated in each
of these cases, corresponding to the four crystal structures
summarized in Chart 4.
route, as indicated in Chart 2: the known28 R-chloroalkene 10
reacted with LiSnMe3 in THF to yield the R-(tri-
methylstannyl)alkene 11 (62%), which was used in place of 10
for preparing the alkenyllithium compound 12 uncontaminated
by LiCl. This Sn/6Li interchange reaction of 11 in pentane was
carried out with n-BuLi in the presence of either tert-butyl
methyl ether (t-BuOMe, 2.5 equiv, very slow) or Et2O (2.1
equiv) and caused crystals of the dimers (12b)2 or (12c)2,
respectively, to precipitate as formed. The X-ray diffraction
structure of (12c)2 (Figure S3)20 reveals disolvation with two
(28) Knorr, R.; Pires, C.; Behringer, C.; Menke, T.; Freudenreich, J.;
Rossmann, E. C.; Bo¨hrer, P. J. Am. Chem. Soc. 2006, 128, 14845–
14853.
Coordinated Et2O displays separate resonances of the dias-
tereotopic geminal OCH2 protons that are shifted slightly
9
14182 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Microsolvation and 13C-Li NMR Coupling
A R T I C L E S
Figure 5. Partial 13C NMR spectrum of (8c)2 with 5.5 equiv of Et2O in
[D8]toluene at -66 °C; the monomeric component (9%) is not shown.
Figure 2. Partial 1H NMR spectrum of (12c)2 with 2.1 equiv of Et2O in
[D8]toluene at -79 °C. o ) alkene 13, cp ) cyclopentane, pt ) pentane,
tol ) [D7]toluene.
Figure 6. Partial 1H NMR spectrum of (8d)2 with 1.0 equiv of THF in
[D8]toluene at -63 °C. tol ) [D7]toluene.
when >1 equiv of THF was present. The CCH2C protons of
THF (not depicted) are shifted slightly upfield to δ ≈ 1.0 ppm.
Neither C2,5 nor C3,4 of THF show diastereotopic splitting,
indicating that THF rotation about the O-Li bond is probably
not frozen on the 13C NMR time scale at this temperature. The
NOESY30 measurement on (8d)2 at 25 °C reveals cross-
relaxation of o-H of the R-phenyl group with both 1-CH3 and
3-CH3 (the latter being weaker because the interaction occurs
across the Li2C2 ring), confirming the closeness of the two half-
dimers depicted in Chart 4 and hence the rather compact dimeric
structure. The broadened proton resonances of both OCH2 and
CCH2C of coordinated THF exhibit NOESY correlations with
o-H, establishing the existence of a close confrontation of Do
) THF with the R-phenyl group that accounts for the afore-
mentioned upfield shift of the CCH2C protons of THF and
moreso of the methyl protons of Et2O in (8c)2 in Figure 3. Only
the OCH2 protons of (8d)2 correlate with 3-CH3 (very broad)
and less strongly with 1-CH3 (broadened). The 3-CH3 resonance
was assigned through the HOESY22-25 correlations of 6Li with
o-H, OCH2, and 3-CH3 (but not 1-CH3) at 25 °C.
Figure 3. Partial 1H NMR spectrum of (8c)2 with 1.3 equiv of Et2O in
[D8]toluene at -80 °C. o ) alkene 9, cp ) cyclopentane, pt ) pentane, tol
) [D7]toluene.
Figure 4. Partial 13C NMR spectrum of (8c)2 with 1.3 equiv of Et2O in
[D8]toluene at -80 °C. o ) alkene 9.
downfield in the 1H NMR spectra of (12c)2 (Figure 2) and (8c)2
(Figure 3), whereas the broadened triplets of the neighboring
methyl protons appear upfield at δ ≈ 0.65 and 0.60 ppm,
respectively. None of these phenomena are artifacts attributable
to decay products or contaminants because they vanished on
decomposition of (8c)2 in the same sample. The diastereotopicity
does not result from frozen Do rotation about the O-Li bond20
because the 13C nuclei of coordinated Et2O remain equivalent
in the presence of smaller (Figure 4) or greater (Figure 5)
amounts of Et2O.
Coordinated Do ) t-BuOMe provides simpler 1H NMR
spectra because its OMe protons are not prone to diastereotopic
splitting. A toluene solution containing (8b)2 and (8b)1 in a 45:
55 molar ratio (62:38 in monomeric units) at -82 °C20 together
with only ∼1.5 equiv of Do per carbanion (entry 4) displayed
1
no H or 13C NMR absorptions of free Do (Figure 7); instead,
well-separated OCH3 resonances were observed in a 43:57 ratio
(45:55 expected) for the Do fractions coordinated at dimeric
and monomeric 8b, respectively. Small amounts of THF are
able to displace Et2O from (8c)2 and t-BuOMe from (8b)2,
forming (8d)2 and generating the well-resolved resonances of
free Et2O and t-BuOMe, respectively.
Coordinated THF in (8d)2 at -63 °C also exhibits diaste-
reotopic OCH2 protons (Figure 6) whose resonances integrate
for 2 + 2 protons with respect to the six-proton singlet of the
two 1-CH3 groups, indicating that this dimer is also disolvated.29
1
The H or 13C resonances of free THF could be detected only
(29) A minor component, believed to be the monosolvated dimer
(8a)2 ·THF, was observed20 when a substoichiometric amount of THF
was added to the cyclotrimer (8a)3 in [D8]toluene.
(30) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys.
1979, 71, 4546–4553.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14183

A R T I C L E S
Knorr et al.
Figure 7. Partial 13C NMR spectrum of (8b)2 and (8b)1 (62:38 in
monomeric units) with 1.5 equiv of t-BuOMe in [D8]toluene at -82 °C. D
) dimer, M ) monomer, o ) alkene 9, cp ) cyclopentane, pt ) pentane.
Figure 10. Partial 1H NMR spectrum of (12b)1 with 3 equiv of t-BuOMe
in [D8]toluene at -96 °C. o ) alkene 13, cp ) cyclopentane, pt ) pentane,
tol ) [D7]toluene.
microsolvated by only two t-BuOMe donors per carbanion
(entries 2, 3, and 5), as illustrated in Figures 8-10. These
disolvated monomers can also be characterized by their lithiation
shifts ∆δ (Table 1): +62 ppm for CR of (8b)1 and (12b)1, -11
ppm for Cpara of (8b)1, and -16 ppm for Cpara of (12b)1. On
the basis of these ∆δ values, a value of 2 for the microsolvation
number d (in parentheses in Table 1) was inferred for those
cases where NMR resonance separation was not achieved: (8b)1
in neat t-BuOMe, (8c)1 with Do ) Et2O, and (8e)1 with Do )
N,N,N′,N′-tetramethylethylendiamine (TMEDA) (entries 1, 6,
and 14, respectively). This tricoordination (C,O,O or C,N,N)
of the lithium cations in the disolvated monomers (8b)1, (8c)1,
and (8e)1 can be ascribed to steric congestion, which limits the
space available for the donors t-BuOMe, Et2O, and TMEDA.
In accord with such overcrowding, (8b)1 does not react with
di-tert-butyl ketone at ambient temperature to produce the
expected alcoholate. But what about the remaining monomers
in Table 1, (8d)1 and (12d)1, which are coordinated to the
smaller donor THF?
Figure 8. Partial 1H NMR spectrum of (8b)1 with 5 equiv of t-BuOMe in
[D8]toluene at -94 °C. o ) alkene 9, cp ) cyclopentane, tol ) [D7]toluene.
The 1:1:1 triplets of (8d)1 with the surprisingly diminished
1JCLi magnitude of 10.8 Hz (Figure 1g,h) are also independent
of changes of the solvent from neat THF to toluene (entries
Figure 9. Partial 13C NMR spectrum of (8b)1 with 5 equiv of t-BuOMe
in [D8]toluene at -94 °C.
1
8-10 of Table 1). However, the quotient of the JCLi values
for (8d)1 and (8d)2 (entries 10 and 11, respectively) is only 10.8/
7.3 ) 1.48, in clear violation of the expected 2:1 relationship.
3. Monomers (8b-e)1, (12b)1, and (12d)1 Provide a Novel
View of JCLi. The ligand Do ) t-BuOMe promotes formation
1
1
A correspondingly low JCLi quotient of 10.0/6.9 ) 1.45 for
of the monomers (8b)1 and (12b)1, presumably because the
dimers are less efficiently stabilized by this bulkier donor. In
the presence of only 3 equiv of t-BuOMe, (12b)1 coexists with
a small portion of the dimer (entry 5 of Table 1). The toluene
solution of 8b with merely ∼1.5 equiv of Do per carbanion at
-82 °C (entries 3 and 4) contained (8b)2 and (8b)1 in the above-
mentioned 45:55 molar ratio (Figure 7), whereas only (8b)1
could be detected with 5 Do equiv per carbanion at -57 °C
(entry 2) and -94 °C (Figures 8 and 9). The 1:1:1 triplets of
the CR resonances shown in Figure 1e,f indicate one contact of
(12d)1 and (12d)2 (entries 12 and 13, respectively) is obtained.
Since all of the dimers in Table 1 have practically equal values
of 1JCLi (6.9-7.3 Hz), it seems that (8d)1 and (12d)1 alone (but
not the b and c monomers) are responsible for this apparent
breaking of the 1/n relationship. Why is this the case?
Likewise surprisingly, the magnitudes of lithiation shifts ∆δ
of these two THF-coordinated monomers are distinctly larger
than those of the other monomers (Table 1): +66 ppm for CR
of (8d)1 and (12d)1 versus +61.5 ppm for the others, -13 ppm
for Cpara of (8d)1 versus -11.5 ppm for (8b)1 and (8c)1, and
-18.5 ppm for Cpara of (12d)1 versus -16.1 ppm for (12b)1.
Since the corresponding lithiation shifts are almost equal for
the disolvated dimers (entries 4, 7, 11, and 13) irrespective of
the nature of the donor, it can be concluded that only the smallest
donor, THF, endows the monomers with magnetic properties
6
CR with Li (i.e., n ) 1) in (8b)1, and the concomitant values
1
of JCLi (13.5 Hz) are equal in the solvents toluene (entries 2
1
and 3) and neat t-BuOMe (entry 1). The quotient of the JCLi
values for (8b)1 (entries 1-3) and (8b)2 (entry 4), 13.5/7.0 )
1.93, leads to the relation 1JCLi ) (13.6 Hz)/n, which confirms
the 17/n relation13-15 in its functional form but not numerically.
Similarly, the Et2O complexes (8c)1 and (8c)2 (entries 6 and 7)
can be characterized by 1JCLi ≈ (14.3 Hz)/n, again falling short
of the 17/n relation. Fortunately, it was possible to observe and
1
(∆δ and JCLi) that differ significantly from those induced by
the other electronically but not sterically similar ethereal donors.
On this basis, we reasoned that d, the number of Do ligands
1
1
integrate separate H and 13C NMR resonances of free and
coordinated at Li, might determine the magnitude of JCLi and
coordinated Do in these spectra. Both (8b)1 and (12b)1 are
that a reciprocal relation as expressed in eq 1 should obtain, in
9
14184 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Microsolvation and 13C-Li NMR Coupling
A R T I C L E S
Table 2. Aggregation States of Four Other Organolithium Compounds, with Donor Ligands (Do), NMR Coupling Constants (1JCLi) and
Pertinent Conditions, Contact Numbers (n, a, d), and Sensitivity Factors (L) Computed Using Equation 2
contact numbersa
entry
RLi
agg. stateb
solventc
Doc
T (°C)
1JCLi (Hz)d
refs
n
a
d
L (Hz)
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
t-BuLi
t-BuLi
t-BuLi
t-BuLi
t-BuLi
t-BuLi
t-BuLi
s-BuLi
M
M
M
D
D
T
T
M
T
H
D
D
T
T
T
THF
THF
THF
Et2O
THF
-90
-108
-108
e-64
-80
11.9
14.3
13.0
7.6
7.8
5.4
5.44
14.0
6.1
13
39
39
13
36
36
34, 35
13
40
41
1
1
1
2
2
3
3
1
3
3
2
2
3
3
3
3
2
3
3
1
1
1
2
2
3
3
1
3
3
2
2
3
3
3
3
2
3
3
3
47.6
47.5e
47.5e
45.6
46.8
48.6
49.0
56
55
55
64.8
64
66.0
66.0
63.6
64.8
66.4
70.8
69.6
TMTAN
TDMAEA
Et2O
Et2O
Et2O
none
PMDTA
none
“2.3”
“2.7”
1
1
0
0
3
0
0
2
2
1
1
1
1
2
1
C5H10
C5H10
C5H10
THF
C5H10
C5H10
toluene
THFg
THFg
(MeO)2CH2
Et2O
-80
e-10
-96
s-BuLi
s-BuLi
-41
none
-41
6.1
8.1
n-BuLi
n-BuLi
n-BuLi
n-BuLi
n-BuLi
n-BuLi
H2CdCHLi
H2CdCHLi
H2CdCHLi
TMEDAf
THF
-78
43
e-90
e-96
-73
8.0(2)e
5.5(2)e
5.5
44-46
23, 44, 46
THF
(MeO)2CH2
Et2O
Et2O
THF
47
33
47
49
49
41
-70
-90
-100
-90
-60
5.3
5.4(1)
8.3
5.9
5.8
T
D
T
Et2O
THF
THF
Et2O
THF
Et2O
T
1
a n ) no. of Li cations bound directly to 13CR; a ) no. of anionic centers bound directly to Li; d ) no. of Do ligands in contact with Li, chosen for the
individual cases as explained in section
1,4,7-trimethyl-1,4,7-triazacyclononane; TDMAEA
Me2NCH2CH2NMe2. d 13C-6Li NMR coupling constants.32 e Mean value. f Labeled with 15N. g Or 1:5 THF/pentane.44
4
in the text. b M
(Me2NCH2CH2)3N; C5H10
)
monomer;
D
)
dimer;
T
)
tetramer;
H
)
hexamer. c TMTAN
)
)
)
)
cyclopentane; PMDTA )
MeN(CH2CH2NMe2)2; TMEDA
qualitative accord with the computational hints18,19 mentioned
in the Introduction:
are the sensitivity factor L and the number d of electrically
neutral donors Do coordinating to the lithium cation:
L ) n(a + d)1JCLi
(2)
1JCLi ) L[n(a + d)]-1
(1)
Values of L characterizing the four RLi compounds in Table 2
were unambiguously established through a special property of
one aggregate of each compound.32 This property was either
15N NMR coupling of Do ) TMEDA with 6Li (revealing d )
2 in the case of known n ) a in entry 26 of Table 2) or
unequivocal evidence of a nonfluxional type-4 tetramer (n )
3) carrying d ) 0 or 1 donor per Li (entry 22 or 33,
respectively), as outlined below. Equation 2 would be verified
if it is possible to describe various aggregation states of a certain
organolithium compound by a common, reasonably sized
sensitivity factor L within acceptable limits. The best (averaged)
L value can then be used to calculate d values for other donors
or other RLi compounds of the same type (i.e., with a
comparable substituent pattern at the carbanionic center).
In the first RLi series in Table 2, unsolvated t-Bu6Li is a
tetramer of type 4 (but without Do) in the solvents cyclohex-
ane,33 cyclopentane (entry 22),34,35 cyclopentane with excess
Do ) Et2O (entry 21),36 and THF (by computation)4 as well as
in the solid state.37 The tetrameric structure in cyclopentane
solution follows from the fact that the fluxional nonet-splitting
(4.1 Hz)34 of 13CR above -10 °C equals 3/4 of the nonfluxional
(septet-splitting) coupling constant 1JCLi ) 5.44 Hz (entry 22);
this rules out a hexameric structure (for which a fluxional
In this expression, a is the number of anions connected with
the Li cation under consideration, so a + d is the total number
of atoms in the coordination sphere of that lithium; the sensitivity
1
31
7
factor L (in Hz) controls the magnitude of JCLi
.
For the Li
3
isotope, which has nuclear spin I ) /2 and hence 2nI + 1 )
3n + 1 multiplet components of the 13CR resonance, L would
be larger by a factor of 2.641. The values of L for 6Li are listed
in the last column of Table 1 and depend hardly at all on the
nature of Do: the averaged factor L ) 42 ( 2 Hz is seen to
apply to 8 and 12 (entries 1-4 and 6-14 of Table 1) with Do
) t-BuOMe, Et2O, THF, and TMEDA. A single striking
deviation [by -6% compared with the value for (12d)1 in entry
12] is provided by the combination of the weak donor t-BuOMe
with the para substituent SiMe3 in (12b)1 (entry 5).
4. Other Applications of Equation 1. Although it was
conceived for the solvated monomers and dimers of 8 and 12,
eq 1 is also applicable to the monomers and aggregates of
several other organolithium species whose coupling constants
1JCLi and C-Li contact numbers n have been determined from
their 13CR multiplets but whose microsolvation numbers d in
solution remained unknown. Because the number a of anion
contacts to a Li nucleus under consideration most often equals
n, as may be seen from the model structures 1-6 shown in
Chart 1, the unknowns in eq 2 (obtained by rearranging eq 1)
1
multiplet splitting of JCLi/2 ) 2.72 Hz would have been
(33) McKeever, L. D.; Waack, R. J. Chem. Soc., Chem. Commun. 1969,
750–751.
1
(31) In simple terms, L/4 ) JCLi for a trisolvated (a + d ) 4) monomer
(34) Thomas, R. D.; Clarke, M. T.; Jensen, R. M.; Young, T. C.
Organometallics 1986, 5, 1851–1857.
(n ) 1); L ) 68 Hz would be consistent with the 17/n relation.
(32) Reported coupling constants 1J(13C, 7Li) were converted to 1J(13C,
6Li) values by multiplication with the quotient of the gyromagnetic
ratios, γ(6Li)/γ(7Li) ) 1/2.641. Because of the small nuclear quad-
rupole moment of 6Li, this isotope provides much better conditions
(35) Thomas, R. D.; Jensen, R. M.; Young, T. C. Organometallics 1987,
6, 565–571.
(36) Bates, T. F.; Clarke, M. T.; Thomas, R. D. J. Am. Chem. Soc. 1988,
110, 5109–5112.
than the isotope 7Li for observing the JCLi splitting of the 13CR
(37) (a) Kottke, T.; Stalke, D. Angew. Chem. 1993, 105, 619–621. (b)
Kottke, T.; Stalke, D. Angew. Chem., Int. Ed. Engl. 1993, 32, 580–
582.
1
resonance, as demonstrated by: Fraenkel, G.; Fraenkel, A. M.; Geckle,
M. J.; Schloss, F. J. Am. Chem. Soc. 1979, 101, 4745–4747.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14185

A R T I C L E S
Knorr et al.
expected). The resulting L value of 49.0 Hz (eq 2) imposes d
) 1 (disolvation by Et2O) on dimeric t-BuLi in Et2O (entry
19)13 or in Et2O/cyclopentane (entry 20)36 in order to reproduce
their common 1JCLi value of 7.7(1) Hz, in accord with the X-ray
analysis of the dietherate.37 With L ) 49.0 Hz, the earlier
assumption36 of a tetrasolvated t-BuLi dimer (d ) 2 Et2O ligands
per Li) would demand 1JCLi ) 6.1 Hz rather than 7.7 Hz. With
the average L value of 47.5 Hz,38 which fits the trisolvation of
monomeric t-BuLi (1JCLi ) 11.9 Hz in entry 16), the coupling
constant value of 14.3 Hz observed39 for monomeric t-BuLi in
the presence of the tridentate donor TMTAN (entry 17) leads
to d ) “2.3”, which might be explained by a mobile equilibrium
between the trisolvated and disolvated monomers. 15N labeling
might be helpful in checking this allegation and also the 2- to
3-fold coordination of the tetradentate donor (Me2NCH2CH2)3N
(TDMAEA), for which d ) “2.7” in entry 18.
The tetrameric type-4 aggregate of vinyllithium
(H2CdCH-Li) in THF solution (d ) 1) is unambiguously
established through its 1JCLi value of 5.9 Hz for the nonfluxional
state (n ) 3, entry 33) in combination with a fluxional multiplet
3
splitting of 4.4 Hz () /4 of 5.9 Hz) at -60 °C and additional
careful investigations.49 The resulting value L ) 70.8 Hz
compares well with that of the Et2O-solvated tetramer (entry
34)41 and implies tetrasolvation (d ) 2, entry 32) for the dimer49
in THF. Results of further applications of eq 2 to methyllithium
(L ) 68 Hz), 1-propyllithium (L ≈ 60 Hz), 2-propyllithium (L
≈ 57 Hz), neopentyllithium (L ) 60 Hz), phenyllithium (L )
61 Hz), lithium acetylides (L ≈ 69 Hz), and other species are
summarized in Table S1,20 and a summarizing overview (Table
S2)20 lists the established or provisionally recommended L
values.
Conclusions
A higher L value of 55 Hz for one of the aggregates of
unsolvated 2-butyllithium (s-BuLi) (entry 24) is obtained as
follows: This species is clearly a nonfluxional tetramer of type
4 (without Do) on account of coupling (1JCLi ) 6.1 Hz) to three
6Li nuclei (n ) 3) in combination with the reduced splitting
of 4.6 Hz () 3/4 of 6.1 Hz) observed40 for the fluxional state
attained at -1 °C. A second aggregate (entry 25) had been
considered to be a fluxional type-6 hexamer mainly on the
basis of the size of its multiplet splitting (3.05 Hz).41 Indeed,
the doubling required in this case (apparent n ) 6) to arrive
2-(R-Lithiobenzylidene)-1,1,3,3-tetramethylindan (8) and its
p-SiMe3 derivative (12) are wholly (>95%) monomeric in
electron-pair-donor solvents at all temperatures investigated,
whereas 8 dissolves with retention of its solid-state cyclotrimer
structure in donor-free toluene.50 The disolvated dimeric ag-
gregates of 8 and 12, as also characterized by X-ray analyses,
become populated in toluene solutions when a shortage of
ethereal donors (Do ) t-BuOMe, Et2O, and THF) disfavors the
monomers. Steric congestion in these monomers and dimers
has the surprising and valuable consequence of retarding the
release of monodentate Do ligands from the Li cation, offering
the unique chance of measuring microsolvation at Li directly
by NMR integrations in toluene solution. This led to the proposal
1
at the nonfluxional hexamer (correct n ) 3) afforded JCLi
) 6.1 Hz with the resulting L value of 55 Hz, in excellent
agreement with the published41 suggestion.42 This L value
demands trisolvation by PMDTA and THF in THF13 for
monomeric s-BuLi (entry 23).
1
of an inverse dependence of JCLi on the sum (a + d) of the
Fluxional and nonfluxional states are spectroscopically identi-
cal for a dimeric aggregate of type 2 or 3, such as (n-
BuLi)2 ·(TMEDA)2 (entry 26). With this chelating Do, the
microsolvation number d ) 2 (from 1 TMEDA) was estab-
lished43 by detecting the 15N-6Li-15N arrangement through
the NMR coupling patterns (6Li, triplet 1:2:1; 15N, triplet 1:1:
1). The resulting L value of 65 Hz sets the stage for the other
n-BuLi aggregates (entries 27-31),23,33,44-47 in agreement with
the corresponding solid-state structures.48 However, it should
be noted that this L value does not conform to the mixed cubic
aggregates44 (n-BuLi)3 ·(ROLi)1 and (n-BuLi)2 ·(ROLi)2, which
numbers of carbanion centers (a) and Do ligands (d) bound to
the Li cation under consideration (eq 1). This dependence could
be discovered here because monomeric 8 and 12 coordinate only
two discrete Et2O or t-BuOMe donor molecules at Li (d ) 2)
rather than the usual d ) 3 (as inferred for Do ) THF), and
thus give rise to changes in 1JCLi even though the values n ) 1
and a ) 1 are common for these monomers. As an empirical
expression, eq 1 may require some refinement.51
The applicability of eq 1 has been shown to extend to other
RLi compounds whose ligand-exchange processes have not as
yet been frozen on the NMR time scale. The possibility of
1
1
have distinctly smaller JCLi values.
determining unknown microsolvation numbers d from JCLi
1
values via eq 2 (a rearranged form of eq 1) endows JCLi with
a new dimension of analytical utility but requires knowledge
of both the aggregation state (contact numbers n and a) and the
sensitivity factor L (as exemplified in section 4). Reliable
evidence for the presence of higher-than-dimeric aggregates can
be obtained by comparing the frequency intervals in the 13CR
multiplets of the nonfluxional and fluxional states of the same
RLi species. Some of the L factors may turn out to be much
smaller than those given in Tables 1 and 2 and Tables S1 and
S2 in the Supporting Information. For example, we did not
(38) The slightly lower L value of 44 Hz is obtained for the mixed type-4
aggregate (t-Bu6Li)3 ·(t-BuO6Li)1 in cyclopentane from its 1JCLi value
of 4.9 Hz (as calculated from the fluxional multiplet splitting of 3.7
Hz).34
(39) Luitjes, H.; Schakel, M.; Aarndts, M. P.; Schmitz, R. F.; de Kanter,
F. J. J.; Klumpp, G. W. Tetrahedron 1997, 53, 9977–9988.
(40) Fraenkel, G.; Henrichs, M.; Hewitt, M.; Su, B. M. J. Am. Chem. Soc.
1984, 106, 255–256.
(41) Fraenkel, G.; Hsu, H.; Su, B. M. In Lithium; Bach, R. O., Ed.; Wiley:
New York, 1985; pp 273-289.
(42) However, the putative dimers40 with multiplet splittings of 4.9 Hz
would need an unreasonably high d value of ∼3.6 for L ≈ 55 Hz and
hence possibly require another assignment.
(43) Waldmu¨ller, D.; Kotsatos, B. J.; Nichols, M. A.; Williard, P. G. J. Am.
Chem. Soc. 1997, 119, 5479–5480.
(49) Bauer, W.; Griesinger, C. J. Am. Chem. Soc. 1993, 115, 10871–10882.
(50) Only ∼3 equiv of t-BuOMe in toluene suffices to deaggregate 8 or
12 almost completely; in contrast, two R-aryl-substituted enolates
remained purely tetrameric in the “poorly coordinating” solvent
t-BuOMe, according to: Streitwieser, A.; Juaristi, E.; Kim, Y.-J.; Pugh,
J. K. Org. Lett. 2000, 2, 3739–3741.
(44) Sun, X.; Winemiller, M. D.; Xiang, B.; Collum, D. B. J. Am. Chem.
Soc. 2001, 123, 8039–8046.
(45) Seebach, D.; Ha¨ssig, R.; Gabriel, J. HelV. Chim. Acta 1983, 66, 308–
337.
1
(46) Eppers, O.; Gu¨nther, H. Tetrahedron Lett. 1989, 30, 6155–6158.
(47) Bergander, K.; He, R.; Chandrakumar, N.; Eppers, O.; Gu¨nther, H.
Tetrahedron 1994, 50, 5861–5868.
(51) JCLi values are “much more sensitive to the degree of aggregation
and number of solvent molecules in the first solvation shell than to
thermal motions”, according to computational results reported by: de
la Lande, A.; Fressigne´, C.; Ge´rard, H.; Maddaluno, J.; Parisel, O.
Chem.sEur. J. 2007, 13, 3459–3469.
(48) Nichols, M. A.; Williard, P. G. J. Am. Chem. Soc. 1993, 115, 1568–
1572.
9
14186 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008

Microsolvation and 13C-Li NMR Coupling
A R T I C L E S
Information,20 which also provides a short summary (Tables S5a
and S5b) of the typical spectral sequences.
analyze RLi species with more delocalized anionic charges and
1
6
small J(13C, Li) values,32 such as benzyllithium compounds
(2.6-3.8 Hz)52-54 and allyllithium species (2.3-3.0 Hz).55,56
As a tempting extension,57 a roughly inverse relationship
Monomers and Dimers of 8. NMR samples of monomers 8b-e
were obtained from the cyclotrimer7,20 (8a)3 (stored in 5 mm NMR
tubes) by removal of the supernatant cyclopentane with a syringe,
blowing with dry argon through a long pipet for 5 s, and addition
of ∼0.7 mL of one of the following dry undeuterated solvents plus
5-10% of [D12]cyclohexane or [D6]benzene (“lock” substances)
and a trace of tetramethylsilane (TMS) at room temperature:
t-BuOMe, Et2O, THF, or TMEDA. [D8]Toluene solutions of the
monomers (8c)1 and (8d)1 were prepared by the addition of a
suitable amount of the donor ligands Et2O or THF, respectively,
and a trace of TMS.
1
6
between J(15N, Li) and the number of coordinating donor
ligands was noted recently,58 in accord with earlier computa-
tional results;59 however, it may be less straightforward to
formulate a quantitative dependence because of the smaller
1
magnitudes of many JNLi values and occasional overlaps of
their ranges.58 Nevertheless, disolvated monomeric and dimeric
LiN(i-Pr)2 coordinating to a series of donors in toluene/pentane
1
solutions have been characterized60 with JNLi ≈ 9.9 and 5.1
Hz, respectively, which would lead to L ≈ 30 in an analogue
The dimers of 8 could be generated in [D8]toluene from (8a)3
by treatment with smaller amounts20 of the donors. In several cases,
it was necessary to apply ultrasonic irradiation for a few minutes
or heat the sample to 45 °C in order to dissolve the plates of (8a)3
completely. Alternatively, the purified20 crystals of (8c)2 were
prepared from 7 (Chart 2) and dissolved in [D8]toluene.
of eq 1.
Experimental Section
General Remarks. Organolithium compounds were handled
under a stream of dry argon cover gas. [6Li]n-Butyllithium45
(<8.9% 7Li) in cyclopentane20 or commercially available solutions
of unlabeled n-butyllithium in hexanes were used in the prepara-
tion20 of the 2-(R-aryl-R-lithiomethylidene)-1,1,3,3-tetramethylin-
dans 8 and 12. NMR tubes (5 mm) containing 8 or 12 either carried
ground-glass joints with glass stoppers or were sealed with soft
rubber stoppers that were then secured by wrapping with an airtight
film of stretchable plastic foil. The ethereal solutions of 8 or 12
were stable for weeks at -18 °C in the sealed NMR tubes when
kept in an inclined position (not less than 20 degrees above
horizontal) in a tightly closed glass cylinder that was filled with
argon gas and contained only tubes with the same solvent. The
solvent [D8]toluene destroyed 8 or 12 slowly20 at 95 °C and above
by deuteron transfer to CR.
1
Dimer (8b)2. H NMR ([D8]toluene with 1 equiv of t-BuOMe,
400 MHz, -12 °C): δ 0.87 (s, t-Bu), 1.28 (broad s, 2 3-CH3), 1.51
(broad s, 2 1-CH3), 2.89 (s, t-BuOMe), 6.79 (very broad, 2 o-H),
6.82 (broad t, p-H). 13C NMR ([D8]toluene with 1 equiv of
t-BuOMe, 100.6 MHz, -12 °C): δ 26.7 (t-Bu), 32.52 (2 1-CH3),
33.62 (broad, 2 3-CH3), 47.78 (broad, C3), 50.3 (t-BuOMe), 50.41
(C1), 75.8 (Me3COMe), 119.4 (broad, Cpara), 122.7 (C7,4), 122.8
(broad, 2 Cortho), 126.5 (C5), 127.0 (broad, C6), Cmeta hidden, 150.6
(broad, C9), 152.6 (sharp, C8), 156.9 (broad, Cipso), ∼162.4 (very
broad, C2), 175.5 (very broad, CR), assigned in analogy with those
of (8d)2.20
Monomer (8b)1. 1H NMR (t-BuOMe, 400 MHz, -70 °C, 0.07
M): δ 1.11 (s, t-Bu), 1.27 (s, 2 1-CH3), 1.31 (s, 2 3-CH3), 3.10 (s,
3
3
t-BuOMe), 6.29 (t, p-H), 6.54 (d, J ) 7.8 Hz, 2 o-H), 6.83 (t, J
) 7.5 Hz, 2 m-H), 6.98 (m, 7-H), 7.01 (m, 5-/6-H), 7.09 (m, 4-H).
13C NMR (t-BuOMe, 100.6 MHz, -70 °C, 0.07 M): δ 27.1 (t-
Bu), 32.8 (2 1-CH3), 34.8 (2 3-CH3), 46.9 (C3), 49.1 (t-BuOMe),
50.1 (C1), 72.2 (Me3COMe), 115.0 (Cpara), 121.6 (2 Cortho), 122.6
(C7), 122.9 (C4), 125.7 (C5), 126.1 (C6), 127.4 (2 Cmeta), 148.4
NMR spectra were run on a Varian VXR-400S (1H at 400 MHz,
6
13C at 100.6 MHz, Li at 58.9 MHz), a Varian HA-100 (13C at
25.15 MHz), or a Bruker WP-80-DS (13C at 20 MHz) spectrometer.
Working temperatures were measured with the usual calibrated
samples of methanol or ethylene glycol. Coupling constants JCH
were obtained through gated or selective decoupling and 13C-H
multiplicities through the DEPT method. A mixing time of 4 s was
used for the 6Li-1H HOESY22-25 experiments at 25 °C. The actual
formal concentrations in mol/L per carbanion unit of 8, 12, or
1
(C2), 152.1 (C9), 154.0 (C8), 160.1 (Cipso), 184.8 (t, JCLi ) 13.5
Hz, CR). These δH and δC values (assignments established)20 are
almost equal to those at 25 °C,20 indicating that the same species
predominates in this temperature range. (8b)1 is stable in the
presence of di-tert-butyl ketone (2.5 equiv) for hours at ambient
temperature.20
1
n-BuLi were estimated by in situ comparisons with the H NMR
integral of either a sealed, calibrated capillary filled with pure
ClCH2CtN (δH ≈ 3.9 ppm) or the low-field 13C satellites of the
nondeuterated solvents. NMR data of some of the aggregates are
reported below by way of example; they do not duplicate the full
characterization or the preparative details given in the Supporting
Monomer (8c)1. 1H NMR (0.18 M in Et2O, 400 MHz, -70 °C):
δ 1.12 (t, Et2O), 1.30 (2 + 2 1-/3-CH3), 3.36 (q, Et2O), 6.30 (broad
t, 3J ) 7 Hz, p-H), 6.54 (d, 3J ) 7.5 Hz, 2 o-H), 6.85 (t, 3J ) 7.5
Hz, 2 m-H), 6.97 (m, 7-H), 7.02 (m, 5-/6-H), 7.10 (m, 4-H). 13C
NMR (in Et2O, 25.15 MHz, -105 °C): δ 32.8 (2 1-CH3), 34.6 (2
3-CH3), 47.1 (C3), 50.3 (C1), 115.2 (Cpara), 121.1 (2 Cortho), 122.9
(C7), 123.2 (C4), 126.0 (C5), 126.4 (C6), 127.7 (2 Cmeta), 147.8
(52) Ruhland, T.; Hoffmann, R. W.; Schade, S.; Boche, G. Chem. Ber.
1995, 128, 551–556.
(53) Fraenkel, G.; Martin, K. V. J. Am. Chem. Soc. 1995, 117, 10336–
10344.
1
(C2), 152.1 (C9), 153.9 (C8), 160.3 (Cipso), 185.4 (t, JCLi ) 14.0
Hz, CR). Values of δH and δC were assigned in analogy with those
of (8b)1.20
(54) Bo¨hler, B.; Hu¨ls, D.; Gu¨nther, H. Tetrahedron Lett. 1996, 37, 8719–
8722.
1
(55) Fraenkel, G.; Qiu, F. J. Am. Chem. Soc. 1997, 119, 3571–3579.
Addition and Correction: Fraenkel, G.; Qiu, F. J. Am. Chem. Soc. 1998,
120, 6848.
Monomer (8d)1. H NMR (THF, 400 MHz, -37 °C): δ 1.27
(broad, 2 1-CH3), 1.30 (broad, 2 3-CH3), 6.18 (t, 3J ) 7.2 Hz, p-H),
3
3
6.49 (d, J ) 7.8 Hz, 2 o-H), 6.74 (t, J ) 7.5 Hz, 2 m-H), 6.94
(very broad, 7-H), 6.99 (m, 5-/6-H), 7.08 (very broad, 4-H). 13C
NMR (THF, 100.6 MHz, -37 °C): δ 33.2 (2 1-CH3), 34.2 (2
3-CH3), 47.0 (C3), 50.5 (C1), 114.0 (Cpara), 121.7 (2 Cortho), 122.6
(broad, C7), 123.0 (broad, C4), 125.7 (broad, C5), 126.0 (broad,
C6), 126.9 (2 Cmeta), 147.5 (C2), 153.4 (C9), 154.5 (C8), 162.6
(56) We cannot explain the physical basis of eq 1 at this time. For some
reflections upon the 17/n relationship, see ref 53 and references therein.
(57) The relation 1J(15N, 6Li) ) (7 Hz)/n was proposed on p 106 of ref 14
on the basis of the 1J values of 7.5 Hz for monomeric
PhNLiCHMe2 ·(THF)3 (a + d ) 4) and 3.8 Hz for the disolvated dimer
(PhNLiMe ·Et2O)2 (a + d ) 3) reported on p 5353 in: Jackman, L. M.;
Scarmoutzos, L. M. J. Am. Chem. Soc. 1987, 109, 5348–5355. In the
spirit of eq 1, however, a tetrasolvated dimer (a + d ) 4) would agree
with a 7.5/n relation.
1
(Cipso), 189.7 (t, JCLi ) 10.7 Hz, CR). Values of δH and δC were
assigned in analogy with those of (8b)1 in t-BuOMe.20 This solution
of (8d)1 deteriorated over the course of 3-7 days at ambient
temperature, forming H2CdCH-OLi and ethylene.
(58) Granander, J.; Sott, R.; Hilmersson, G. Chem.sEur. J. 2006, 12, 4191–
4197 and references therein.
(59) Koizumi, T.; Morihashi, K.; Kikuchi, O. Bull. Chem. Soc. Jpn. 1996,
69, 305–309.
2-(4-Trimethylsilyl-r-trimethylstannylbenzylidene)-1,1,3,3-
tetramethylindan (11). Powdered 2-(R-chloro-4-trimethylsilylben-
zylidene)-1,1,3,3-tetramethylindan (10)28 (600 mg, 1.63 mmol) was
(60) Remenar, J. F.; Lucht, B. L.; Collum, D. B. J. Am. Chem. Soc. 1997,
119, 5567–5572 and Table 1 therein.
9
J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008 14187

A R T I C L E S
Knorr et al.
added at -70 °C under dry argon cover gas to a solution of LiSnMe3
(∼3.3 mmol)61-64 in THF (5.4 mL). As soon as 10 had completely
dissolved (∼5 min without cooling), the mixture was placed in an
ice bath for not more than 30 min and then diluted with distilled
water and Et2O. The aqueous layer was extracted (3×) with Et2O,
and the combined Et2O phases were washed until neutral and then
dried over MgSO4, filtered, and concentrated in vacuo. The crude
product (734 mg) containing 11 and the alkene 1328 (70:25) was
crystallized from ethanol (5 mL), yielding spectroscopically pure
11 as colorless needles (504 mg, 62%) with mp 61-65.5 °C. The
Dimers and Monomers of 12. The Sn-6Li interchange reaction
of 11 in pentane (Chart 2) carried out with n-BuLi in the presence
of t-BuOMe (2.5 equiv) proceeded very slowly to afford crystals
of 12b. A solution of these crystals in toluene contained (12b)1
and (12b)2, the latter being recognized through its ∆δ values, which
resemble those of (12d)2 in entry 13 of Table 1. With diethyl ether
(2.1 equiv) as the donor admixture, a faster Sn-6Li interchange
reaction provided single crystals of 12c that were sufficiently stable
for X-ray diffraction at 20 °C (Table S4).20 The analyzed crystal
(Figure S3)20 consisted of dimeric aggregates (12c)2 of type 2
solvated by one Et2O molecule per carbanion, as in (8c)2. Purified
crystals of 12c or 12d were dissolved in [D8]toluene to yield (12c)2
alone (poorly soluble) or (12d)2 alone, respectively. (12d)2 was also
obtained by treating (12c)2 in [D8]toluene with THF, and (12d)1
was generated by dissolving (12b)2 in THF.
1
analytically pure sample had mp 73-74 °C. H NMR (400 MHz,
CDCl3): δ 0.03 (s; 119Sn satellites, 2J ) 51.7 Hz; SnMe3), 0.27 (s,
SiMe3), 1.20 (s, 2 3-CH3), 1.52 (s; 119Sn satellites, 5J ≈ 2.9 Hz; 2
1-CH3), 6.93 (dm, 3J ) 8.2 Hz, 2 o-H), 7.04 (dm, 4-H), 7.15-7.24
3
(m, 5-/6-/7-H), 7.38 (dm, J ) 8.2 Hz, 2 m-H). 13C NMR (100.6
MHz, CDCl3): δ -5.1 (qm, J ) 128.7 Hz; 119Sn satellites, J )
339 Hz; SnMe3), -0.9 (qm, 1J ) 118.7 Hz; 3J ) 2.0 Hz; 29Si
satellites, 1J ) 52.3 Hz; SiMe3), 31.81 and 31.84 (2 qm, 1J ) 127
Hz, 2 + 2 1-/3-CH3), 48.4 (m; 119Sn satellites, 3J(cis) ) 21.9 Hz;
C1), 50.6 (m; 119Sn satellites, 3J(trans) ) 62.2 Hz; C3), 122.2 and
122.3 (2 dm, 1J ) 156 Hz, C7,4), 126.7 and 126.9 (2 dd, 1J ) 159
1
1
Acknowledgment. This work is dedicated to Professor Herbert
Mayr. We thank Prof. H. No¨th and Drs. K. Polborn, M. Schmidt,
and H. Schwenk-Kircher for performing X-ray analyses. Parts of
these studies were sponsored by the Deutsche Forschungsgemein-
schaft.
3
1
Hz, J ) 7.5 Hz, C5,6), 127.7 (dm, J ) 156 Hz; 119Sn satellites,
3J ) 18.7 Hz; 2 Cortho), 132.3 (dm, 1J ) 156 Hz; 119Sn satellites,
4J ) 9.5 Hz; 2 Cmeta), 135.9 (unresolved m; 119Sn satellites, 5J ≈
Supporting Information Available: Further applications of
eq 1 (Tables S1 and S2); solid-state structural data (Tables S3
and S4 and Figures S1-S4); lithiation shifts and spectral
sequences (Tables S5a and S5b); four typical total NMR spectra
11.7 Hz; Cpara), 139.0 (unresolved t, J ≈ 3 Hz; 119Sn satellites
3
not discernible; CR), 145.9 (sharp t, J ) 7.5 Hz; 119Sn satellites,
3
2J ) 20.4 Hz; Cipso), 149.9 (m, C8), 150.4 (m, C9), 167.1 (m, C2).
Values of δH and δC were assigned through SCS65 and in analogy
with those of alkene 13.28 IR (KBr): 2955, 2916, 1612 (w), 1487,
1359, 1248, 1110, 1027, 861, 838, 751, 532, 527 cm-1. Anal. Calcd
for C26H38SiSn (497.4): C, 62.79; H, 7.70. Found: C, 62.95; H,
7.75. 11 could also be formed from (8d)1 in THF with ClSnMe3 at
room temperature within a few minutes.
1
(Figures S5-S8); H and 13C NMR assignments (Tables S6
and S7); and full spectroscopic and preparative details of the
observed monomers and aggregates (Tables S8-S20). This
material is available free of charge via the Internet at http://
pubs.acs.org.
JA8026828
(61) Tamborski, C.; Ford, F. E.; Soloski, J. E. J. Org. Chem. 1963, 28,
237–239.
(64) Knorr, R.; Pires, C.; Freudenreich, J. J. Org. Chem. 2007, 72, 6084–
6090.
(65) Knorr, R.; von Roman, T.; Freudenreich, J.; Hoang, T. P.; Mehlsta¨ubl,
J.; Bo¨hrer, P.; Stephenson, D. S.; Huber, H.; Schubert, B. Magn. Reson.
Chem. 1993, 31, 557–565.
(62) Kitching, W.; Olszowy, H.; Waugh, J.; Doddrell, D. J. Org. Chem.
1978, 43, 898–906.
(63) Reich, H. J.; Reich, I. L.; Yelm, K. E.; Holladay, J. E.; Gschneidner,
D. J. Am. Chem. Soc. 1993, 115, 6625–6635 (specifically p 6632).
9
14188 J. AM. CHEM. SOC. VOL. 130, NO. 43, 2008