Amine- and ether-chelated aryllithium reagents - Structure and dynamics

Article abstract of DOI:10.1021/ja028301r

Chelation and aggregation in phenyllithium reagents with potential 6- and 7-ring chelating amine (2, 3) and 5-, 6-, and 7-ring chelating ether (4, 5, 6) ortho substituents have been examined utilizing variable temperature ⁶Li and ¹³C NMR spectroscopy, ⁶Li and ¹⁵N isotope labeling, and the effects of solvent additives. The 5- and 6-ring ether chelates (4, 5) compete well with THF, but the 6-ring amine chelate (2) barely does, and 7-ring amine chelate (3) does not. Compared to model compounds (e.g., 2-ethylphenyllithium 7), which are largely monomeric in THF, the chelated compounds all show enhanced dimerization (as measured by K = [D]/[M]²) by factors ranging from 40 (for 6) to more than 200 000 (for 4 and 5). Chelation isomers are seen for the dimers of 5 and 6, but a chelate structure could be assigned only for 2-(2-dimethylaminoethyl)phenyllithium (2), which has an A-type structure (both amino groups chelated to the same lithium in the dimer) based on NMR coupling in the ¹⁵N, ⁶Li labeled compound. Unlike the dimer, the monomer of 2 is not detectably chelated. With the exception of 2-(methoxymethyl)phenyllithium (4), which forms an open dimer (12) and a pentacoordinate monomer (13), the lithium reagents all form monomeric nonchelated adducts with PMDTA.

Full text of DOI:10.1021/ja028301r

Published on Web 03/01/2003
Amine- and Ether-Chelated Aryllithium ReagentssStructure
and Dynamics
Hans J. Reich,* Wayne S. Goldenberg, Aaron W. Sanders, Kevin L. Jantzi, and
C. Christoph Tzschucke
Contribution from the Department of Chemistry, UniVersity of Wisconsin,
Madison, Wisconsin 53706
Received August 27, 2002 ; E-mail: reich@chem.wisc.edu
Abstract: Chelation and aggregation in phenyllithium reagents with potential 6- and 7-ring chelating amine
(2, 3) and 5-, 6-, and 7-ring chelating ether (4, 5, 6) ortho substituents have been examined utilizing variable
temperature 6Li and 13C NMR spectroscopy, 6Li and 15N isotope labeling, and the effects of solvent additives.
The 5- and 6-ring ether chelates (4, 5) compete well with THF, but the 6-ring amine chelate (2) barely
does, and 7-ring amine chelate (3) does not. Compared to model compounds (e.g., 2-ethylphenyllithium
7), which are largely monomeric in THF, the chelated compounds all show enhanced dimerization (as
measured by K ) [D]/[M]2) by factors ranging from 40 (for 6) to more than 200 000 (for 4 and 5). Chelation
isomers are seen for the dimers of 5 and 6, but a chelate structure could be assigned only for
2-(2-dimethylaminoethyl)phenyllithium (2), which has an A-type structure (both amino groups chelated to
the same lithium in the dimer) based on NMR coupling in the 15N, 6Li labeled compound. Unlike the dimer,
the monomer of 2 is not detectably chelated. With the exception of 2-(methoxymethyl)phenyllithium (4),
which forms an open dimer (12) and a pentacoordinate monomer (13), the lithium reagents all form
monomeric nonchelated adducts with PMDTA.
Chelation effects are ubiquitous in the design of organolithium
reagents and interpretation of their chemistry. We have under-
taken systematic studies of several classes of such reagents to
establish the existence of chelation (i.e., determine the winner
in the competition between external and internal solvation) and
to determine the effects of the chelate ring size and chelating
atom (N or O1b,c) on the structural,1d stereochemical,1f and
chemical properties of such lithium reagents. In a previous paper,
the solution structure and dynamics of a series of five-membered
ring ortho amine-chelated phenyllithium derivatives (e.g., 1)
were analyzed.1a,1h We report here full details on the solution
behavior of aryllithiums 2 and 3 with potentially chelated six-
and seven-membered ring amines, as well as the analogous
series of reagents with pendant ether linkages at the ortho
position which could form 5, 6, and seven-membered ether
chelates (4-6).1g
There is a substantial literature on the chemistry of chelated
organolithium reagents, beginning with early reports by Hauser
on the chelation-assisted ortho metalation of N,N-dimethylben-
zylamine to form 12 and followed by numerous studies of
assisted metalation of aryl, vinyl and alkyl protons.3a,c,43,4
Chelation effects are widely used in organolithium chemistry
as design elements to facilitate metalations and increase
structural rigidity,3a,4-6 as well as to provide rationales for
stereochemical, regiochemical and reactivity effects. Although
(1) (a) Reich, H. J.; Gudmundsson, B. O¨ . J. Am. Chem. Soc. 1996, 118, 6074-
6075. (b) Reich, H. J.; Kulicke, K. J. J. Am. Chem. Soc. 1995, 117, 6621-
6622. (c) Reich, H. J.; Kulicke, K. J. J. Am. Chem. Soc. 1996, 118, 273-
274. (d) Reich, H. J.; Sikorski, W. H.; Gudmundsson, B. O¨ .; Dykstra, R.
R. J. Am. Chem. Soc. 1998, 120, 4035-4036. (e) Reich, H. J.; Holladay,
J. E. J. Am. Chem. Soc. 1995, 117, 8470-8471. (f) Reich, H. J.; Dykstra,
R. R. Angew. Chem., Int. Ed. Engl. 1993, 32, 1469-1470. (g) Some of
these results were reported in a preliminary communication: Reich, H, J.;
Goldenberg, W. S.; Sanders, A. W.; Tzschucke, C. C. Org. Lett. 2001, 3,
33-36. (h) Reich, H. J.; Goldenberg, W. S.; Gudmundsson, B. O¨ .; Sanders,
A. W.; Kulicke, K. J.; Simon, K.; Guzei, I. A. J. Am. Chem. Soc. 2001,
123, 8067-8079. (i) Reich, H. J.; Green, D. P.; Medina, M. A.; Goldenberg,
W. S.; Gudmundsson, B. O¨ .; Dykstra, R. R.; Phillips, N. H. J. Am. Chem.
Soc. 1998, 120, 7201-7210. (j) The sample temperature during the NMR
(2) Jones, F. N.; Zinn, M. F.; Hauser, C. R. J. Org. Chem. 1963, 28, 663-
665. Klein, K. P.; Hauser, C. R. J. Org. Chem. 1967, 32, 1479-1483.
(3) (a) Beak, P.; Snieckus, V. Acc. Chem. Res. 1982, 15, 306-312. Snieckus,
V. Chem. ReV. 1990, 90, 879-933. Beak, P.; Meyers, A. I. Acc. Chem.
Res. 1986, 19, 356-363. (b) Beak, P.; Kerrick, S. T.; Gallagher, D. J. J.
Am. Chem. Soc. 1993, 115, 10 628-10 636. (c) Gschwend, H. W.;
Roderiquez, H. R. Org. React. 1979, 26, 1-360.
experiment was measured using the tris(trimethylsilyl)methane internal 13
C
NMR chemical shift thermometer: Sikorski, W. H.; Sanders, A. W.; Reich,
H. J. Magn. Reson. Chem. 1998, 36, S118-S124. (k) DNMR simulations
and integrations of overlapping peaks were performed with a version of
the computer program WINDNMR (Reich, H. J. J. Chem. Educ. Software,
1996, 3D, 2; http://www.chem.wisc.edu/areas/reich/plt/windnmr.htm).
9
10.1021/ja028301r CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 3509-3521
3509

A R T I C L E S
Reich et al.
Scheme 1
recent work has provided substantial insights into the conse-
quences of chelation,,1a-d,3b,4b,c,7a,8-10 these applications are still
hampered by a lack of knowledge about the strength and even
the occurrence of chelation. There is no question from single-
crystal X-ray structures that organolithium reagents with pendant
alkoxy and amino groups are chelated under poorly solvated or
solvent-free conditions.11 It is, however, less clear how well
such chelating groups compete with commonly used ethereal
solvents such as THF.1e,14
Table 1. Thermodynamic and Kinetic Data for Chelated
Aryllithiums and Model Compounds
Results and Discussion
Syntheses. The lithium reagents were prepared by Li/Sn
exchange of the appropriate o-trimethylstannyl compounds,
typically in diethyl or dimethyl ether solution, from which most
of the lithium reagents could be purified by crystallization. The
bromo precursors of the tin derivatives were not used directly
because of interference from lithium bromide (if t-BuLi was
used) or 1- and 2-bromobutane (if n- or sec-BuLi was used).
‡
compd
R
KMD/M-1
∆G°a (T /°C)
∆G DMa (T/°C)
8
-H
210
1.7
0.19
<0.23
-1.5 (-128)b
-0.2 (-135)e
0.4 (-140)b
8.3c (-101)d
-CH3
7
9
1
2
3
4
5
6
-CH2CH3
6.8b (-139)b
-(CH2)2-iPrf
g0.6 (-125)b
6
To facilitate NMR studies, the Li enriched isotopomers7d of
e-2.6 (-131)b >12.5g,h (-36)b
f
-CH2NMe2
>17 000
226
<0.23
>35 400
11 000
12
all compounds were prepared and the amines were synthesized
in 15N enriched form. The synthesis of the amines is illustrated
for 2 in Scheme 1. Compound 3 was prepared analogously from
3-(2-bromophenyl)propanoic acid. The ethers 4 and 5 were
prepared by straightforward methods.15
-(CH2)2NMe2
-(CH2)3NMe2
-CH2OMe
-1.5 (-137)b
g0.6 (-130)i
9.4k (-107)i
e-3.0 (-127)b g9.5g (-80)j
-(CH2)2OMe
-(CH2)3OMe
-2.8 (-121)b
-0.7 (-141)b
10.7k (-83)b
8.8j(-107)i
a Free energies in kcal/mol for kDM and KDM. ∆G° is the free energy
difference between a dimer and two molecules of monomer. b 3:2 THF/
Et2O. c Ref 1i. d THF. e 4:1 THF/Et2O. f Ref 1h. g Coalescence of the 1:2:
3:2:1 quintet of the C-Li carbon. h The rate is bimolecular. i 3:2:1 THF/
Me2O/Et2O. j 3:2:1 Me2O/THF/Et2O. k Exchange of monomer and dimer
(4) (a) Klumpp, G. W. Recl. TraV. Chim. Pays-Bas 1986, 105, 1-21. (b)
Luitjes, H.; Schakel, M.; Schmitz, R. F.; Klumpp, G. W. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2152-2153. (c) Schmitz, R. F.; Schakel, M.; Vos,
M.; Klumpp, G. W. Chem. Commun. 1998, 1099-1100. (d) Klumpp, G.
W.; Geurink, P. J. A.; Spek, A. L.; Duisenberg, A. J. M. J. Chem. Soc.,
Chem. Commun. 1983, 814-816. (e) Moene, W.; Schakel, M.; Hoogland,
G. J. M.; de Kanter, F. J. J.; Klumpp, G. W.; Spek, A. L. Tetrahedron
Lett. 1990, 31, 2641-2642. (f) Moene, W.; Vos, M.; de Kanter, F. J. J.;
Klumpp, G. W. J. Am. Chem. Soc. 1989, 111, 3463-3465. (g) Klumpp,
G. W.; Vos M.; de Kanter, F. J. J.; Slob, C.; Krabbendam, H.; Spek, A. L.
J. Am. Chem. Soc. 1985, 107, 8292-8294. (h) Vos, M.; de Kanter, F. J. J.;
Schakel, M.; van Eikema Hommes, N. J. R.; Klumpp, G. W. J. Am. Chem.
Soc. 1987, 109, 2187-2188. (i) Geurink, P. J. A.; Klumpp, G. W. J. Am.
Chem. Soc. 1986, 108, 538-539.
6
signals in Li NMR spectrum.
Solution Structures. All of the lithium reagents were studied
6
7
by low temperature 13C, Li, Li, and, where appropriate, 31P
NMR spectroscopy in mixed solvents containing THF with
dimethyl and/or diethyl ether as cosolvents to allow operation
at temperatures below the freezing point of THF (-108.5 °C).
The effects of cosolvents N,N,N′,N′-tetramethylethylenediamine
(TMEDA), N,N,N′,N′′,N′′-pentamethyldiethylenetriamine (PM-
DTA) and hexamethylphosphoric triamide (HMPA15) were also
studied, since these provided a window into the strength and
nature of chelation and aggregation.
(5) Paquette, L. A.; Kuo, L. H.; Tae, J. J. Org. Chem. 1998, 63, 2010-2021.
Smyj, R. P.; Chong, M. J. Org. Lett. 2000, 3, 2903-2906.
(6) Lamothe, S.; Chan, T. H. Tetrahedron Lett. 1991, 32, 1847-1850. Hartley,
R. C.; Lamothe, S.; Chan, T. H. Tetrahedron Lett. 1993, 34, 1449-1452.
(7) (a) Fraenkel G.; Cabral J.; Lanter C.; Wang, J. J. Org. Chem. 1999, 64,
1302-1310. Fraenkel G.; Duncan J. H.; Wang J. J. Am. Chem. Soc. 1999,
121, 432-443. (b) For an insightful discussion of dynamic processes
involving C-Li multiplets in monomeric aryllithium reagents, see: Fraen-
kel, G.; Subramanian, S.; Chow, A. J. Am. Chem. Soc. 1995, 117, 6300-
6307. (c) Fraenkel, G.; Chow, A.; Winchester, W. R. J. Am. Chem. Soc.
1990, 112, 6190-6198. (d) For many organolithium reagents, especially
aggregated ones, quadrupolar broadening is severe enough with 7Li (92.6%
natural abundance) that C-Li coupling cannot be resolved. The 6Li
analogues show little or no quadrupolar broadening and thus couplings are
more easily seen. Fraenkel, G.; Fraenkel, A. M.; Geckle, M. J.; Schloss, F.
J. Am. Chem. Soc. 1979, 101, 4745-4747.
Table 1 summarizes the thermodynamic and kinetic data we
have gathered for compounds 1-9. Table 2 presents chemical
shift and coupling information. In particular, the 13C chemical
shifts of the C-Li carbon provided a clear definition of the
aggregation state of the lithium reagents,15 the monomers have
C-1 at ca. δ 195 and dimers at ca. δ 187. These assignments
were supported by the C-Li coupling observed, a 1:1:1 triplet
(8) Sato, D.; Kawasaki, H.; Shimada, I.; Arata, Y.; Okamura, K.; Date, T.;
Koga, K. J. Am. Chem. Soc. 1992, 114, 761-763.
(9) Arvidsson, P. I.; Hilmersson, G.; Ahlberg, P. J. Am. Chem. Soc. 1999,
121, 1883-1887.
6
(10) Intramolecular comparison of a 5- and 6-ring ether chelated vinyllithium
reagent suggested that the 5-ring chelate is stronger. Mitchell, T. N.;
Reimann, W. J. Organomet. Chem. 1987, 322, 141-150.
for the monomeric Li enriched compounds, and a 1:2:3:2:1
quintet for the dimers. The observation of a quintet does not
rule out cyclic trimers. In cases where two aggregates were
present (2, 5, 6), we determined their concentration dependence
to establish the molecularity.
(11) (a) 5-ring ether-chelated structures have been found for 1-lithio-1-(2-
methoxyphenyl)-3,3-diphenylallene (Dem′yanov, P.; Boche, G.; Marsch,
M.; Harms, K.; Fyodorova, G.; Petrosyan, V. Liebigs Ann. 1995, 457-
460), 3-lithio-1-methoxybutane[4d] and 2,2-bis(methoxymethyl)-1-lithio-
propane.[4e] (b) 5- and 6-ring amine-chelated structures have been found
for 1,1-bis(dimethylaminomethyl)-2-lithiopropane,[4f] N-(2-lithiocyclohex-
enyl)-N,N′,N′-trimethyl-1,3-propanediamine,[12a] and o-dialkylaminophenyl-
lithium.[13a]
Lithium NMR Spectra. The 6Li/7Li NMR spectra were also
very informative in determining the solution structure. Figure
6
(12) (a) Polt, R. L.; Stork, G.; Carpenter, G. B.; Williard, P. G. J. Am. Chem.
Soc. 1984, 106, 4276-4277. (b) Waldmu¨ller, D.; Kotsatos, B.; Nichols,
M. A.; Williard, P. G. J. Am. Chem. Soc. 1997, 119, 5479-5480. (c)
Williard, P. G.; Liu, Q.-Y. J. Am. Chem. Soc. 1993, 115, 3380-3381.
(13) (a) Rietveld, M. H. P.; Wehman-Ooyevaar, I. C. M.; Kapteijn, G. M.; Grove,
D. M.; Smeets, W. J. J.; Kooijman, H.; Spek, A. L.; van Koten, G.
Organometallics 1994, 13, 3782-3787. (b) Jastrzebski, J. T. B. H.; van
Koten, G.; Konijn, M.; Stam, C. H. J. Am. Chem. Soc. 1982, 104, 5490-
5492. Wehman, E.; Jastrzebski, J. T. B. H.; Ernsting, J.-M.; Grove, D. M.;
van Koten, G. J. Organomet. Chem. 1988, 353, 145-155.
1 summarizes the Li spectra obtained. These spectra will be
discussed in more detail in the individual sections below.
Model Systemss2-Ethylphenyllithium (7). A principal
focus of our studies is the relationship between chelation affects
and the level of aggregation. To allow a distinction to be made
(15) See the Supporting Information for NMR spectra, simulations, and exchange
matrixes.
(14) McDougal, P. G.; Rico, J. G. J. Org. Chem. 1987, 52, 4817-4819.
9
3510 J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003

Amine- and Ether-Chelated Aryllithium Reagents
A R T I C L E S
Table 2. 13C NMR Chemical Shifts of 2-Alkylphenyllithium
Reagents.
a
RLi
solv.
T/°C
C-1
C-2
C-6
δLi
JCLi (6Li)
Monomers
(p ) PMDTA):
b
(PhLi)1
(PhLi)1‚pb
(7)1
(2)1
(2)1‚p
(3)1
(3)1‚p
(4)1‚p
(5)1‚p
(6)1
THF -111 196.4 143.2 143.2 1.15
15.3c
15.6
14.6g
12c
d
-125 196.5 143.8 143.8 1.95
f
-141 195.2 155.3 142.5 1.24
h
-130 195.9 154.3 142.6 1.23
h
-145 196.7 150.5 142.3 2.06
15.5c
15c
f
-127 195.5 153.2 142.4 1.26
f
-127 195.8 152.5 142.2 2.05
13.3c
17c
h
-127 195.8 148.4 142.6 2.1
h
-147 197.0 148.0 142.1 2.05
14.8c
f
-125 195.6 152.9 142.5 1.28
15
f
k
(6)1‚p
-125 195.7 152.2 142.3 2.1
Dimers
7.9j
b
(PhLi)2
(7)2
THF -111 188.2 144.5 144.5 1.59
f
k
-141 186.6 157.3 144.0 1.57
(1)2 (A)l
-132 188.8 152.2 143.5 1.81,
7.0j
h
3.84
(1)2 (B)l
(1)2 (C)l
(2)2 (A)
-132 188.4 151.9 143.2 2.61
7.0j
h
h
h
[i]
i
-132
151.2 143.0 2.90
-130 184.3 152.4 144.1 1.87,
1.96
7j
f
f
(4)2 (B/C)
(5)2 (B/C)
(6)2 (X)m
-127 186.1 149.7 143.1 2.38
8.1j
-127 188.2 152.4 143.2 1.74
8.1j
k
Et2O -126 188.0 152.8 144.0 1.67
Et2O -126 186.8 153.7 144.2 1.50
k
(6)2 (Y)m
Tetramers:
b
(PhLi)4
Et2O -106 174.0 143.8 143.8 2.39
5.1
b
c
d
a Reference: 0.3 M LiCl in methanol. Ref 1i. 1:1:1 Triplet. 2:1
THF/Me2O. e 45:25:30 THF/Me2O/Et2O. f 3:2 THF/Et2O. g Calculated from
a measured JC-7Li coupling of 38.5 Hz. 3:2:1 THF/Me2O/Et2O. Signal
not found. 1:2:3:2:1 quintet. Splitting not resolved. Ref 1h. X is the
minor and Y the major isomer (35:65).
h
i
j
k
l
m
6
Figure 1. Low temperature Li NMR spectra of lithium reagents 1-8 in
3:2 THF/Et2O (M ) monomer, D ) dimer, for A/B/C see Scheme 2).
Chemical shifts are referenced to external 0.3 M LiCl in methanol.
between the steric effects of ortho substituents and their
chelating effects, we have used several model systems. Phe-
nyllithium (8) forms both monomers and dimers in THF (KMD
) 30 M-1 at -118 °C).1i,16,17 In our mixed solvent system 3:2
THF/Et2O KMD ) 208 M-1 at -128 °C. To mimic the steric
effect of the chelating ortho substituents we have examined
2-ethylphenyllithium (7), which can be crystallized, unlike our
previous model 9.1h The reagent was mostly monomer in 3:2
THF/Et2O, but a higher aggregate was easily detected, and
shown to be a dimer by the concentration dependence of the
6Li NMR signals (KMD ) 0.19 ( 0.04 M-1 at -140 °C, ca. 5%
dimer at 0.14 M total lithium concentration). In this and previous
papers1c,h we have made extensive use of two solvent mixtures
which qualitatively show similar behavior: 3:2 THF/Et2O,
which can be used down to -140 °C, and 3:2:1 THF/Me2O/
Et2O, which is fluid enough for high-resolution NMR work
down to -155 °C. We find KMD ) 0.19 ( 0.03 M-1 at -139
°C for 7 in 3:2:1 THF/Me2O/Et2O. Thus, the two solvent
mixtures are identical within experimental error in their ability
to cause deaggregation, at least for this specific lithium reagent.
Interaction of 7 with PMDTA. The behavior of the chelated
lithium reagents with PMDTA has in several cases provided
an estimate of the strength of chelation, since the tridentate
ligand is in competition with the chelating group. The PMDTA
titration of 7 is very similar to other model systems (8, 9) we
have examined. The association constant is almost too large to
measure. A solution of 7 at -127 °C with 1.06 equiv of PMDTA
had 96% of ArLi complexed, giving KPMDTA ([7‚PMDTA]/[7M]-
[PMDTA]) ) 1880 M-1
.
The dynamics of ligand exchange were readily measured by
6
DNMR studies, either using Li NMR spectra at 0.5 equiv of
PMDTA, or the 13C PMDTA signals with excess ligand present.
Two ligand exchange processes were identified. At low tem-
perature 9 signals were seen in the 13C NMR spectra of the
bound ligand, in addition to the 4 signals of free PMDTA (see
Figure 2 for representative spectra).15 Around -100 °C the
bound ligand symmetrizes, with 9 signals going to 4. The central
methyl signal is unaffected. We did not detect the two-phase
process, first C-Li bond rotation then Me2N-Li decoordination/
inversion, that was seen for neopentyllithium‚PMDTA.7c How-
ever, the free PMDTA NMe2 signal interferes with detailed
observation of the methyl coalescences in our study. At ca. -80
°C, bound and free PMDTA coalesce. This appears to be a
simple dissociative process because the rates were identical for
solutions containing 0.6 and 1.9 equiv of PMDTA,15 where the
concentration of free PMDTA differs by several orders of
magnitude. Figure 3 shows an Eyring plot of the rate constants
that were determined by a 6-spin DNMR simulation of the
PMDTA CH2 signals,15 together with corresponding data for
ligand dynamics in 4‚PMDTA and 5‚PMDTA.
Five-Membered Ring Amine Chelations2-(Dimethylami-
nomethyl)phenyllithium (1). Previous work on 1 and its
congeners1h has shown that these are mixtures of dimeric
chelation isomers (A, B, and C, X ) NMe2, NEt2, or NMeiPr,
Scheme 2),18 with much lower rates of dimer to monomer
(16) (a) Bauer, W.; Seebach, D. HelV. Chim. Acta 1984, 67, 1972-1988. (b)
Heinzer, J.; Oth, J. F. M.; Seebach, D. HelV. Chim. Acta 1985, 68, 1848-
1862. (c) Seebach, D.; Ha¨ssig, R.; Gabriel, J. HelV. Chim. Acta, 1983, 66,
308-337.
(17) Bauer, W.; Winchester, W. R.; Schleyer, P. v. R. Organometallics 1987,
6, 2371-2379.
9
J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003 3511

A R T I C L E S
Reich et al.
6
Figure 4. Selected Li NMR spectra from a variable temperature experi-
ment of 0.10 M [6Li, 15N]2 in 3:2 THF/Et2O. The upper trace of each pair
is a simulated spectrum using the rate constants shown and the exchange
matrix of Fig. S-2.15
Figure 2. Variable temperature NMR study of [6Li]7 in 3:2 THF/Et2O
with 1.9 equiv of PMDTA. Line shape analysis15 gave ∆H‡ ) 4.4 ( 0.1
kcal/mol, ∆S‡ ) -18 ( 3 eu for ligand symmetrization (ksymm) and ∆H‡
) 7.1 ( 0.1 kcal/mol, ∆S‡ ) -10 ( 1 eu for ligand dissociation (k diss).
concentration and were assigned to the two lithiums of a type
A dimer (Scheme 2). The 6Li signal at δ 1.87 is slightly broader
6
than the one at δ 1.96, presumably due to residual Li-14N
coupling. The minor component (6Li δ 1.23) was assigned as a
monomer. Because of the relatively low concentration of the
monomer we were able to get only a poorly resolved 1:1:1 triplet
signal for the C-Li carbon in the 13C NMR spectra at δ 195.9,
but all evidence is consistent with such an assignment. These
6
assignments were confirmed by the variable temperature Li
NMR spectra of the 15N-6Li doubly labeled material (Figure
4).
Selected spectra from a variable temperature NMR study of
the 15N-6Li doubly labeled compound are shown in Figure 4.19
Below -128 °C the upfield 6Li signal of the dimer is split into
a 1:2:1 triplet (JLiN ) 2.7 Hz), thus confirming the A structure.
It is unusual for N-coordination to result in an upfield 6Li shift.
Normally, both intramolecular1b,1h and intermolecular (with
TMEDA and PMDTA1i) N-coordination results in downfield
shifts. The monomer shows no hint of Li-N coupling in the
6LiNMR spectra. The C-Li signal in the 13C NMR spectra was
not well resolved but showed signs of C-Li coupling, so
intermolecular Li-Li exchange is not the reason that Li-N
coupling is absent. We conclude that the monomer, unlike the
dimer, is not chelated.
Figure 3. Eyring plot for rates of ligand symmetrization (ksymm)and ligand
dissociation (kdiss) of PMDTA complexes of 4, 5, and 7 (∆G‡-100 for ksymm
was 6.21, 7.14 and 7.64 and for kdiss, 7.69, 8.74 and 8.86 for 4, 5, and 7,
respectively).
Scheme 2. Dimeric Chelation Isomers.
Scheme 3 presents a scenario for the effect of chelation on
the monomer-dimer equilibrium, and our best estimates for
the fractions of the two species in brackets that are not
observable.15 We estimate Kchel ([chelated]/[unchelated]) as g
25 for the dimer and e 0.18 for the monomer. The hypothetical
monomer-dimer equilibrium constant KMD ([D]/[M]2) with both
dissociation and much higher aggregation constants than the
largely monomeric model systems. Of the several cosolvents
tried (TMEDA, PMDTA, HMPA), only HMPA was capable
of causing detectable deaggregation.
monomer and dimer chelated is estimated at g 10 000 M-1
and with both monomer and dimer unchelated, e 0.24 M-1
,
.
Six-Membered Ring Amine Chelations2-(2-Dimethylami-
noethyl)phenyllithium (2). Compound 2 showed two sets of
signals in the 13C and 6Li NMR spectra (Figure 1).15 The major
component showed the characteristic C-Li quintet of a dimer
at δ 184.3 in the 13C NMR spectrum. Two closely spaced signals
The latter value is quite consistent with the measured KMD of
0.19 M-1 for the model system 2-ethylphenyllithium. These
numbers illustrate the dramatic influence of chelation on the
monomer-dimer equilibrium, a switch of a factor of at least
40 000 in KMD and almost 3 kcal/mol in ∆∆G°.
6
in the Li NMR spectra maintained a 1:1 ratio independent of
(18) (a) Chelation isomers of a similar type have been identified for tetrameric
3-dimethylaminopropyllithium,[4g] and dimeric 1,1-bis(dimethylaminom-
ethyl)-2-lithiopropane[4f] and 3-lithio-1,5-dimethoxypentane.[4h]
(19) Other examples of 6Li-15N coupling in amine chelates have been
reported.8,12b,20,21,22a
9
3512 J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003

Amine- and Ether-Chelated Aryllithium Reagents
A R T I C L E S
Scheme 3. Estimated Dimerization and Chelation Equilibrium
Relations for 0.10 M (total ArLi) 2 in 3:2 THF/Et2O at -137 °Ca
molecules), from about 14% monomer at -137 °C in the sample
of Figure 4 (KMD ) [D]/[M]2 ) 226 M-1) to about 9% at -110
°C (KMD ) 560 M-1). Simulation of the spectra also gave values
for the rates of dimer-monomer equilibration. Figure 5 shows
a ∆G‡ vs T plot of the data obtained. The negative entropy of
activation implies that solvent assists in the dimer to monomer
process, as expected for a dissociative process. Note that 1 has
an opposite temperature dependence of ∆G‡ (positive ∆S‡),
consistent with the known associative nature of that process.1h
Interaction of 2 with TMEDA. The incremental addition
of TMEDA to a sample of 2 produced changes in the NMR
spectra consistent with complexation with both the monomer
and dimer. A new set of triplet and singlet signals appeared in
6
a The items in square brackets were not detected, and the Keqs involving
the Li NMR spectra downfield of the original ones, assigned
these species are based on estimates of their maximum likely concentration.15
to 2A‚TMEDA (2A‚t in Figure 6). The monomer signal broad-
ens and moves steadily downfield. Apparently the complexed
and uncomplexed monomers are in rapid equilibrium at -134
°C (as they are for PhLi‚TMEDA1i and 2-isopentylphenyllithium‚
TMEDA1h). Figure 6 also shows spectra of a similar sample
with 2 equiv of TMEDA at lower temperatures. The signals
decoalesced at -140 °C and are well resolved at -153 °C. The
TMEDA-complexed monomer (2M‚TMEDA) is not chelated.
The complexation is sub-stoichiometric. With two equiv of
TMEDA 52% of the lithium species are coordinated to TMEDA.
Figure 5. Rates of intra- and intermolecular exchanges of 2 in 3:2 THF/
Et2O compared with a related processes in 1.1h kAA′
: intramolecular
We have determined by line-shape simulation the relative
proportions of the four species (M, Mt, D, Dt) in the spectrum
at -153 °C at 2 equiv of TMEDA, and have estimated
monomer-dimer association constants for the TMEDA com-
plexed species (KMD ) [Dt]/[Mt] [M] ) 64 M-1) and the THF
complexed species (KMD ) 170 M-1). Similarly, we can define
the TMEDA affinity of the monomer (K ) [Mt]/[M] [t] ) 11
M-1) and for the dimer (K ) [Dt]/[D] [t] ) 4 M-1). Thus,
TMEDA shows approximately a 3-fold higher affinity for the
monomer than for the dimer.
Interaction of 2 with PMDTA. Figure 7a shows the effect
of the addition of PMDTA to a sample of [6Li,15N]2. The new
sharp singlet at δ 2.1 was assigned to the PMDTA complex.
There is no detectable Li-N coupling, so the complexation
forms monomeric unchelated 2‚PMDTA.
exchange between two Li signals of 2A; kAB: intramolecular exchange of
the A and B chelation isomers of the dimer of 1; kDM: dimer to monomer
exchange of 2; kD: loss of coupling of C-Li carbon in the dimer of 1
(associative through trimer or tetramer).[1h]
Dynamic Processes for 2. Two dynamic processes are seen
for 2 (Figure 4). First, the singlet and triplet in the 15N labeled
A-dimer coalesce to form a new triplet (in the unlabeled
compound the two singlets coalesce to one). This is the result
of intramolecular exchange of the chelating ligands between
the two sites, so that each lithium becomes coupled equally to
both nitrogens, now with half the coupling constant. The second
process is the coalescence of the dimer signals with that of the
monomer, the result of intermolecular exchange. We have
simulated the spectra by treating the A-dimer singlet as a 1:2:1
triplet with JLiN ) 0,1k and allowing pairwise exchange of the
equivalent signals in the two “triplets.”15 This simulation
provided an excellent fit between calculated and experimental
spectra confirming the assignment of the complex changes in
line shapes. The free energies of activation for the degenerate
chelation isomerization of 2 are similar to those found for the
intramolecular interconversion of the chelation isomers of 1,
although this compound shows a substantial negative ∆S‡,
whereas 2 has ∆S‡ of essentially 0 (Figure 5). This suggests
that for 1 the dechelation requires association of one or more
additional solvent molecules at the transition state, whereas for
2 this is a unimolecular process.
The PMDTA titration experiment also provided an op-
portunity to determine the molecularity of the two species we
(21) Eriksson, J.; Arvidsson, P. I.; Davidsson, O¨ . J. Am. Chem. Soc. 2000, 122,
9310-9311.
(22) (a) Aubrecht, K. B.; Lucht, B. L.; Collum, D. B. Organometallics 1999,
18, 2981-2987. (b) Romesberg, F. E.; Gilchrist, J. H.; Harrison, A. T.;
Fuller, D. J.; Collum, D. B. J. Am. Chem. Soc. 1991, 113, 5751-5757. (c)
Remenar, J. F.; Lucht, B. L.; Kruglyak, D.; Romesberg, F. E.; Gilchrist, J.
H.; Collum, D. B. J. Org. Chem. 1997, 62, 5748-5754. (d) In other systems,
TMEDA competes less well: Collum, D. B. Acc. Chem. Res. 1992, 25,
448-454. (e) Lucht, B. L.; Bernstein, M. P.; Remenar, J. F.; Collum, D.
B. J. Am. Chem. Soc. 1996, 118, 10 707-10 718. (f) Kim, Y- J.; Bernstein,
M. P.; Roth, A. S. G.; Romesberg, F. E.; Williard, P. G.; Fuller, D. J.;
Harrison, A. T.; Collum D. B. J. Org. Chem. 1991, 56, 4435-4439.
(23) Knorr, R.; Freudenreich, J.; Polborn, K.; No¨th, H.; Linti, G. Tetrahedron
1994, 50, 5845-5860.
As is often the case,1d,7c,16b,c,23-25 the higher aggregate of 2
was favored at higher temperature (aggregation releases solvent
(24) Vinyllithium also complexes with PMDTA as a dimer. It was not established
whether the complexation of the PMDTA was bidentate or tridentate. Bauer,
W.; Griesinger, C. J. Am. Chem. Soc. 1993, 115, 10 871-10 882.
(25) Harder, S.; Boersma, J.; Brandsma, L.; Kanters, J. A.; Bauer, W.; Schleyer,
P.v. R. Organometallics 1989, 8, 1696-1700.
(20) Huls, D.; Gu¨nther, H.; van Koten, G.; Wijkens, P.; Jastrzebski, J. T. B. H.
Angew. Chem., Int. Ed. Engl. 1997, 36, 2629-2631.
9
J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003 3513

A R T I C L E S
Reich et al.
Figure 6. 6Li NMR spectra. Left panel: TMEDA titration of a 0.16 M
solution of [6Li, 15N]2 in 3:2 THF/Et2O at -134 to -137 °C. Right panel:
a variable temperature study of a similar sample (0.13 M in 3:2:1 THF/
Me2O/Et2O) with 2 equiv of TMEDA, showing decoalescence of the
TMEDA complexed and uncomplexed monomers (t ) TMEDA, M )
monomer).
Figure 8. 6Li NMR spectra showing the effect of PMDTA on solution in
3:2:1 THF/Me2O/Et2O of (a) [6Li, 15N]3 (the indicated equiv of PMDTA
are low by ca. 20%); (b) [6Li]4 (c) [6Li]5.
Five-Membered Ring Ether Chelations2-(Methoxymeth-
yl)phenyllithium (4). Only a single species is detectable in
solution of 4 in THF mixed solvents. The compound is
aggregated as judged by the 13C NMR signal of the C-Li carbon
at δ 186.1, a 1:2:3:2:1 quintet (JC-6Li ) 7.6 Hz). Both the
chemical shift and the coupling pattern are consistent with a
cyclic dimer, but higher cyclic oligomers have not been ruled
out. No signal that could be identified as a monomer was found
6
in the 13C and Li NMR studies carried out on 4. The tallest
peak/noise signal between δ 0.5 and δ 2.0 in the spectrum of 4
in Figure 1 has 0.5% of the peak height of the major dimer
signal. If we conservatively estimate that no more than 1% of
monomer could be present, we calculate KMD g 35 400 M-1
.
It is much harder to get direct evidence for or against chelation
in the ethers than for the amine chelates, since there is no
suitable oxygen isotope for high-resolution NMR studies which
would detect J coupling between Li and oxygen. The single
6Li chemical shift of δ 2.4 is not directly useful for structure
assignment, since it is possible that conformational isomers of
the A/B/C type are in rapid equilibrium. Such equilibration is
supported by the proton NMR signal of the benzylic CH2 group,
which appeared as a singlet at δ 4.70 even at -151 °C.15
Dynamic Processes for 4. We have estimated the barrier to
aggregate equilibration of 4 from collapse of the C-Li coupling
(1:2:3:2:1 quintet to singlet). Sample spectra, simulations and
exchange rate constant data in the form of a ∆G‡ vs T plot are
shown in Figure 9. The OMe chelated compound 4 is substan-
tially more labile than 1 (barriers ca. 3 kcal/mol lower), and
shows the opposite sign for ∆S‡. This may indicate a different
mechanism for the interaggregate exchange of 1 and 4. The
exchange of 1 is faster at higher concentrations, consistent with
an associative process going through higher aggregates. We have
not measured the concentration dependence for 4, but the large
negative entropy is consistent with coordination of additional
solvent in the transition state for dimer dissociation. However,
we do not want to over-interpret the ∆S‡ values. The changes
in line shape are relatively small, and obtaining accurate rates
is hampered by the long acquisition times required, typically
30-100 min, resulting in some temperature drift during the
experiment. The internal 13C chemical shift thermometer tris-
(trimethylsilyl)methane1j effectively corrects for small temper-
Figure 7. (a) 6Li NMR spectra of a PMDTA titration of a 0.13 M solution
of doubly labeled [6Li,15N]2 in 3:2:1 THF/Me2O/Et2O at -145 °C. (b)
Concentration dependence of the uncomplexed monomer and dimer signals
during the PMDTA titration.
have assigned as the monomer and dimer of 2. A log-log plot
of the concentrations of the non-PMDTA complexed species
(signals at δ 1.2 vs δ 1.7-2.0) gave a slope of 2.1, confirming
the assignment of aggregation states (Figure 7b).
Seven-Membered Ring Amine Chelations2-(3-Dimethy-
laminopropyl)phenyllithium (3). 13C and 6Li NMR studies of
3 show chemical shifts and couplings characteristic of a
monomer (C-Li δ 195.5, 1:1:1 triplet, JC-Li ) 15 Hz). There
6
is no detectable Li-15N coupling in either THF/Et2O media
or solutions with TMEDA, PMDTA, or HMPA added, so the
compound exists largely in an unchelated form. A small broad
6
signal attributable to a dimer appears at δ 2.1 in the Li NMR
spectra amounting to 6-8% of the area of the monomer, which
would give a monomer-dimer association constant KMD ) 0.27
M-1. Concentration-dependent studies were not done to support
such an assignment. If this signal corresponds to a dimer, the
substantial downfield shift compared to model compounds,
where the dimer signal appears at δ 1.5-1.6 (see Figure 1),
suggests that it is chelated.
PMDTA complexed stoichiometrically to 3 (Figure 8a).
HMPA also formed a stoichiometric mono-complex. Past one
+
equiv a substantial amount of triple ion Ar2Li- Li(HMPA)4
was formed, with a singlet at δ 3.8 and quintet at δ -0.4 in 6Li
NMR spectra.
9
3514 J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003

Amine- and Ether-Chelated Aryllithium Reagents
A R T I C L E S
Figure 9. (a) Selected variable temperature 13C NMR spectra of the C-Li
carbon of 0.15 M [6Li]4 in 3:2:1 Me2O/THF/Et2O. The simulated spectra
(upper trace of each pair) were obtained using the 6 × 6 exchange matrix
reported previously.1h (b) Eyring plots and activation parameters for the
intermolecular exchange rate (∆G‡
) 9.9 kcal/mol). Related processes
-65
for 1 in 3:2 THF/Et2O are also shown in the graph for comparison.
ature drifts to provide time-averaged temperatures, but it was
not available for this experiment.
Interaction of 4 with TMEDA. In contrast to the amine
analogue 1, which complexes modestly with TMEDA, spectra
of 4 gave no indication of interaction with TMEDA.15 The
failure to form a TMEDA complex is unlikely to have steric
origins since other ortho-substituted lithium reagents such as 1
or 2 complex reasonably well and are at least as crowded. We
conclude that 4 must be chelated and have a strong preference
for the B and/or C isomers (Scheme 2), which cannot form
bidentate complexes with TMEDA, over the A structure.
Interaction of 4 with PMDTA. Given the absence of
significant interaction with TMEDA, it is surprising that 4
complexed strongly with PMDTA. The spectroscopic details
of the interaction were apparent only at very low temperatures
(-150 °C). Two species are formed, one gives a 1:1 ratio of
two peaks in the 6Li NMR spectra (Figures 8b and 10), and the
second gives a singlet δ 2.04. The ratio of the two sets of peaks
changes to favor the singlet at higher concentration of PMDTA,
in a way consistent with formulating it as a monomer, and the
other species as a dimer. Both species have unexpected
spectroscopic features.
The first species to be formed gives a 1:1 ratio of two peaks
at δ 1.9 and 2.8 at -150 °C. An unusual bonding situation is
indicated by the very low temperatures (< -140 °C) required
to decoalesce the various solvated species. In contrast, the
PMDTA complexes of 1,1h 2, 3, 5, 6, 7, PhLi1i and other lithium
reagents,7b,c are much less labile. We have considered three
precedented structures for 4D‚PMDTA: a triple ion 10,1d an
A-type dimer 11, and an open dimer structure 12 (Figure 10c).
The triple ion structure 10 can be ruled out. The 6Li chemical
shift of SIP-PMDTA complexes (e.g., Li‚PMDTA+ // -C-
(SiMe3)3) is close to δ 0, whereas the species here has shifts of
δ 2-3. Furthermore, all of the 13C NMR signals in the less
polar solvent 2:1 Me2O/Et2O, where a higher fraction of the
4D‚PMDTA complex formed and the rates are slower, are
doubled. The ipso carbon signals are especially revealing: at δ
193.2 (distorted 1:1:1 triplet) and δ 188.0 (coupling unresolved),
quite close to the PMDTA-monomer signal at δ 195.8 and the
non-PMDTA-dimer signal at δ 185.3 (Figure 10a). The sub-
stantial chemical shift difference of 5.1 ppm between the ipso
carbons argues against assignment of an A-type dimer 11 in
Figure 10. NMR studies of a 0.33 M solution of [6Li]4 in 2:1 Et2O/Me2O.
(a) 13C spectra of a PMDTA titration at -135 °C. (b) Variable temperature
6Li NMR study of the sample containing 0.75 equiv of PMDTA. (c) Possible
structures for the 4-PMDTA species.
which a restricted conformation of the PMDTA ligand makes
the two rings diastereotopic. We conclude that the 4D‚PMDTA
complex has an open dimer structure 12 where one of the C-Li
signals mimics that of the monomer, whereas the other
resembles the dimer. Open dimers have been detected for lithium
diisopropylamide and other lithium amides,22b,c,12c but not for
organolithium reagents.24
Both the 6Li and the 13C spectra of the PMDTA-dimer showed
distortions in peak shape at -140 °C. The origin of these
distortions became apparent when the samples were cooled
further. Each of the 6Li peaks decoalesced to an approximately
2:1 ratio of singlets (Figure 10b). We have not established the
origin of this dynamic process, with ∆G‡
) 7.1 kcal/mol.
-141
Slowing of some conformational change within either the
PMDTA ligand or the MeO groups, or restricted rotation around
the C-Li(PMDTA) bond7c are possible explanations.
The second species which forms on treatment of 4 with
6
PMDTA gives a singlet at δ 2.04 in the Li NMR spectrum
and is formed to the extent of 70% at one equiv and 91% at 3
equiv in 3:2:1 THF/Me2O/Et2O solvent. The C-Li carbon
chemical shift (δ 195.8) and coupling (1:1:1 triplet, JC-Li ) 17
Hz for 6Li) are consistent with a monomeric PMDTA complex.
The PMDTA must be coordinating lithium in a tridentate
fashion, otherwise TMEDA would have given a similar effect.
We also conclude that the lithium in 4M‚PMDTA is still
chelated to the CH2OMe group (structure 13), based on the
peculiar properties of the PMDTA ligand signals in the 13C
NMR spectra. In complexes of 2, 3, 5, 6 and 7 with PMDTA
the complexed ligand shows 9 resolved signals at temperatures
between -130 and -145 °C (Figures 2 and 11).26 In contrast,
it is necessary to cool a sample of 13 to -150 °C to decoalesce
(26) Similar spectra have been reported for mesityllithium‚PMDTA[7b] and
neopentyllithium‚PMDTA.[7c]
9
J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003 3515

A R T I C L E S
Reich et al.
Figure 11. Ligand 13C signals of the monomeric PMDTA complexes of
3, 4, 6, and 7. The spectra of 4 were measured in 3:2:1 THF/Me2O/Et2O,
the others in 3:2 THF/Et2O.
Figure 12. (a) Variable temperature NMR study of a solution containing
0.27 M [6Li]4 and 0.27 M [6Li]PhLi in 3:2 THF/Et2O. The upper trace of
each set was a 3-spin DNMR simulation1k performed using the rate constants
shown. The k values are the NMR rate constants for interchange of nuclei;
the physical rate constants for interchange of dimer molecules would be
half these values. (b) Eyring plots for the intermolecular exchanges of 4.
the signals, and then only six of nine peaks, still broadened by
exchange, are resolved for the bound PMDTA carbons. Like
7‚PMDTA (Figure 2), 4‚PMDTA shows two dynamic processes
in the PMDTA ligand, symmetrization and intermolecular
exchange (dissociation), except that the barriers are at least 1
kcal/mol lower than for 7‚PMDTA and other models (Figure
3). In addition, at least one of the signals (arrow in Figure 11)
is shifted by ca. 1 ppm from that of the other complexes, which
all have nearly identical ligand shifts. The methoxy group must
be chelated in 13, giving a rare structure with pentacoordinate
lithium. Previously reported RLi‚PMDTA complexes have been
tetracoordinate at Li.7c,27
species (most likely monomers, but mixed trimers and tetramers
are also plausible) which are not detectable and so cannot be
incorporated into the simulation. However, the values obtained
for k4-14 should approximate the values for aggregate exchange
of 4D. In fact, they are within a factor of 2 of those measured
from collapse of the C-Li multiplet in a slightly different
solvent mixture (3:2:1 Me2O/THF/Et2O vs 3:2 THF/Et2O for
the mixed dimer experiment, Figure 12b). Thus, this experiment
provides a way of estimating the kinetic stability of dimers for
which other techniques are not available. A similar mixed dimer
experiment for 5 described below was more informative since
here the monomeric species can be detected and explicitly
incorporated into the simulations.
Mixed Dimer of 4 with Phenyllithium. The paucity of
information that deals directly with the issue of chelation in 4
encouraged us to examine the mixed dimer of 4 and PhLi. Like
6
1,1h solutions of 4 containing PhLi (8) show signals in the Li
and 13C NMR spectra consistent with the formation of a mixed
dimer 14 (Figure 12a) in higher than statistical amounts. The
comproportionation equilibrium constant ([14]2/[4D][8D]) is 17;
statistical is 4. The mixed dimer would be expected to show
two signals in the lithium NMR spectra (as does the mixed dimer
of 1). Since only one new signal for 14 is seen, ligand
reorganization must be fast on the NMR time scale at -137
°C. An accidental equivalence is unlikely since the chemical
shifts of the homodimers are separated by 0.7 ppm.
At higher temperatures the signals of 4D, 14 and 8D coalesce.
A 3-spin DNMR simulation gave an excellent fit using only
two rate constants, conversion of 4D to 14 (k4-14), and
conversion of 14 to 8D (k14-8). A study of the concentration
dependence gave ambiguous results (a 4-fold dilution led to a
reduction in exchange rates by a factor of 1.2 to 1.6, suggesting
at least some involvement of associative processes (e.g., mixed
trimers) in the exchange. Unfortunately, there is not a direct
connection between these rate constants and the physically real
rate constants of aggregate exchange, because these must involve
Six-Membered Ring Ether Chelations2-(2-Methoxyeth-
yl)phenyllithium (5). The homologue of 4, compound 5, shows
none of the anomalous behavior exhibited by 4. Two sets of
6
signals in a 4:1 ratio were observed in the Li and 13C NMR
spectra (Figure 13), and this ratio was concentration independent.
We were unable to resolve the C-Li coupling, but the line shape
and 13C chemical shift (δ 188.2) of the C-Li carbon are very
similar to those observed for other dimers (Table 2). We
conclude that 5 is a mixture of two dimers, probably of the
B/C type. A small signal at δ 1.2 was assigned to the monomer,
but this could not be confirmed by observing the key 13C signal
for the C-Li carbon. Its intensity varied as expected for a
monomer.15 Our best value for KMD is 11 000 M-1 at -121 °C
in 3:2 THF/Et2O, almost 5 orders of magnitude higher than the
model systems (KMD < 0.2 M-1), and almost 2 orders of
magnitude higher than the amine analogue 2 (KMD ) 240 M-1).
Dynamic Processes for 5. Excerpts from a variable temper-
ature experiment are shown in Figure 13. The signals for the
two dimers 5D and 5D′ coalesced at ca -110 °C, the monomer
5M coalesced with the dimers above -90 °C. We did not
perform the DNMR experiment at other concentrations to
determine the molecularity of the exchange processes.
(27) Ru¨ffer, T.; Bruhn, C.; Maulitz, A. H.; Stro¨hl, D.; Steinborn, D. Organo-
metallics 2000, 19, 2829-2831. Schu¨mann, U.; Kopf, J.; Weiss, E.; Angew.
Chem. 1985, 97, 222; Angew. Chem., Int. Ed. Engl. 1985, 24, 215-216.
Lappert, M. F.; Engelhardt, L. M.; Raston, C. L.; White, A. H. J. Chem.
Soc., Chem. Commun. 1982, 1323-1324. Karsch, H. H.; Zellner, K.;
Mikulcik, P.; Lachmann, J.; Mu¨ller, G. Organometallics 1990, 9, 190-
194.
Interaction of 5 with TMEDA. Treatment of solutions of 5
with up to 8 equiv of TMEDA gave no significant change in
9
3516 J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003

Amine- and Ether-Chelated Aryllithium Reagents
A R T I C L E S
Figure 14. 6Li NMR spectra in 3:2 THF/Et2O. (a) 0.026 M 5 at -121 °C.
(b) 0.01 M PhLi (8) and 0.01 M 5 at -124 °C. (c) 0.045 M PhLi at -128
°C.
) 5.4 ( 0.3 kcal/mol, ∆S‡ ) -10 ( 2 eu) was essentially
identical to that of the model system, whereas the symmetriza-
tion of the ligand (ksymm, ∆H‡ ) 6.7 ( 0.2 kcal/mol, ∆S‡ )
-12 ( 1 eu) was faster by a factor of 2-4.
Figure 13. Variable temperature 6Li NMR study of [6Li]5. The upper trace
for the -118 and -90 °C spectra are 3-spin DNMR simulations which
give ∆G‡-118 (D-D′) ) 8.50 kcal/mol and ∆G‡-90 (D-M) ) 10.51 kcal/
mol.
Scheme 4. Thermodynamic Relations in the PMDTA Complexes
of 5 and 7a
Mixed Dimer of 5 with Phenyllithium. Some additional
insights into the effects of chelation were obtained from a study
of the mixed dimer of PhLi with 5. A spectrum from such an
experiment, together with spectra of the constituents, is shown
in Figure 14. Signals for both monomer and dimer of 5, the
mixed dimer 15, and monomer and dimer of PhLi can be seen.
Line shape simulation of the signals in Figure 14b allowed
calculation of the following monomer-dimer KMD values ([D]/
[M]2): for 5: 14 000 M-1; for 15: 3800 M-1; for 8 (PhLi):
240 M-1. Thus the monomer-dimer association constant goes
up by about an order of magnitude for each chelating meth-
oxyethyl group.
Dynamic NMR studies of solutions of 5/PhLi (Figure 15)
were complicated by the small area of the monomer signals,
the superposition of the signals for 8D (PhLi) and 5D′, the
a The dotted arrows are not true equilibria, they indicate predicted free
energy relations between 7 and the unchelated model for 7 (5).
6
chronic difficulty in assigning inherent line widths to the Li
signals, and the large number of adjustable parameters inherent
in a 6-spin simulation (15 rate constants, 6 line widths, 6
chemical shifts and 6 populations). We made the following
simplifying assumptions. Because of the low-temperature coa-
lescences of 8D and 8M, and the two dimers 5D and 5D′, their
relative chemical shifts, line widths, and populations are poorly
defined through most of the temperature range. These were set
from the low-temperature spectra and changed minimally during
the simulation. Only four rate processes of the 15 possible were
allowed: the interconversion of 5D and 5D′, the dissociation
of 5D to 5M (both isomers of the dimer at the same rate), the
dissociation of 8D to 8M and the dissociation of mixed dimer
15 to the two monomers (i.e., no bimolecular exchanges between
the three dimers). It was possible to achieve satisfactory
simulations throughout the temperature range (Figure 15) despite
the drastically reduced set of adjustable parameters.
the spectra.15 Because TMEDA complexes well to similarly
hindered ArLi reagents such as 2 and 9,1h the failure to form
significant amounts of a type A‚TMEDA complex suggests
strongly that the dimers must be of the B/C type (Scheme 2),
as also observed for 4.
Interaction of 5 with PMDTA. Figure 8c shows a 6Li NMR
study of a PMDTA titration of 5. A monomeric complex is
formed which has nearly identical ligand chemical shifts to the
model systems 3 and 7 (see Figure 11), and thus is not chelated.
However, complexation is no longer stoichiometric. In a solution
of 5 containing 2.1 equiv of PMDTA at -126 °C 88% of ArLi
was complexed, giving a KPMDTA (Scheme 4) of 9.5 M-1
,
whereas solutions of 7 had KPMDTA ) 1880 M-1. These data,
combined with the measured KMD values for 5 (11 000) and 7
(0.19) can be used to construct the free energy diagrams shown
in Scheme 4. If we assume that 5M and 5D, were they not
chelated, would behave identically to the model 7 aggregates,
we can conclude that 5D is stabilized to the extent of 1.7 kcal/
mol by chelation, whereas 5M is negligibly stabilized (0.1 kcal/
mol in the diagram, but the errors are at least ( 0.1 kcal/mol).
The dynamics of the PMDTA ligand are affected in only a
minor way by chelation (Figure 3). In contrast to 4‚PMDTA,
the intermolecular exchange of PMDTA (kdiss in Figure 3, ∆H‡
The variable temperature spectra showed a series of four
coalescences: first the monomer and dimer of PhLi, and in the
same temperature regime (-100 °C), the two dimers of 5
coalesced with each other. Then, between -93 and -74 °C the
mixed dimer 15 coalesced with the averaged PhLi signals (and
presumably also with 5M, but this cannot be directly observed).
Finally, between -52 and -41 °C, the averaged signals of 5
coalesced with the rest. The rates measured in this experiment
9
J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003 3517

A R T I C L E S
Reich et al.
Figure 16. Carbon-carbon couplings and chemical shifts in 13CH3O 4, 5
and model compounds in 3:2 THF/Et2O at 11 °C.
6
Figure 15. (a) Variable temperature Li NMR study of a solution 0.011
M in 5 and 0.011 M in PhLi (8) in 3:2 THF/Et2O. The upper trace at each
temperature is the 6-spin DNMR simulation. The rate constants shown are
the NMR rates. The physical rate constants for dissociation of 5D and 8D
(kBE and kDF) are 0.5 times the values shown. (b) Eyring plot of the four
independent rate constants used in the simulation: dissociation of 5D (kBE
) kCE, ∆H‡ ) 6.34 ( 0.31 kcal/mol, ∆S‡ ) -23.0 ( 1.5 eu); dissociation
of 8D (kDF, ∆H‡ ) 7.21 ( 0.15 kcal/mol, ∆S‡ ) -7.2 ( 0.9 eu);
dissociation of the mixed dimer 15 (kAE ) kAF, ∆H‡ ) 7.74 ( 0.33 kcal/
mol, ∆S‡ ) -11.3 ( 1.6 eu); interconversion of the two isomeric dimers
of 5 (kCD, ∆H‡ ) 8.31 ( 0.56 kcal/mol, ∆S‡ ) -1.9 ( 3.3 eu). (c) Eyring
plot comparing the 0.011 M mixed dimer experiment of part (a) and (b)
(filled circles and squares) with data from a similar variable temperature
study of a solution 0.16 M in 5 and 0.16 M in 8 (PhLi) (open circles and
squares), showing that the rates are experimentally indistinguishable.
6
Figure 17. Variable-temperature 13C and Li NMR spectra of 6. Shown
are the C-Li carbons (left), the methoxy signals (center) and the 6Li signals
(M ) monomer, D ) dimer). The peaks marked “H” and “Sn” in the -119
°C spectrum are the OMe signals of the protonated product (3-methoxy-
propyl)benzene and trimethylstannyl precursor.
labeled 4 and 5 and measured the 3-bond and 4-bond coupling
in the lithium reagent and model compounds (Figure 16). The
couplings in the various nonchelated (or largely nonchelated)
models are essentially identical, and in each case there is
significant change in the coupling on going to the organolithium
reagent. There are also substantial chemical shift changes,
especially of the benzylic carbon, which are also in part due to
the conformational changes resulting from chelation. However,
the strongly polarizing C-Li bond is the principal source of
these changes, as shown by the fact that the methyl group of
o-tolyllithium monomer (δ 30.1) moves downfield by 8.7 ppm
from its position in toluene (δ 21.4). Thus the coupling constants
also may reflect changes in electronic structure as much as
changes in conformation. Although the changes in couplings
support the conclusions reached from other experiments that 4
and 5 are chelated, they are probably not large enough to provide
firm evidence on their own.
are within experimental error of the corresponding rates
measured for pure 5 (Figure 13).
The quality of the simulations validates the assumptions that
were made, i.e., that the exchanges of the three dimers largely
involve unimolecular dissociations, without the participation of
trimers or tetramers. This conclusion was also supported by a
parallel experiment run at a 16-fold higher concentration.15
Although the Eyring plots for this rate study showed more scatter
than the one shown in Figure 15b, the principal rates (dissocia-
tion of 5D and 15) were within experimental error of each other
(Figure 15c), ruling out bimolecular exchange processes ob-
served for the dimer-monomer exchange for 1.
The dimer dissociation rates follow the trends found in other
chelated ArLi reagents. For example, the relative rates of
dissociation of 8D, 15, and 5D at -83 °C are calculated to be
540:17:1.9 from the activation parameters (∆G‡-83 ) 8.6, 9.9,
and 10.7 kcal/mol). There is thus approximately 1 order of
magnitude increase in kinetic stability of the dimer for each
chelating CH2CH2OMe group, similar to the increase in
monomer-dimer association (KMD) values.
Seven-Membered Ring Ether Chelations2-(3-Methoxy-
propyl)phenyllithium (6). We have also examined the poten-
tially 7-ring chelated ether 6. The 6Li and 13C NMR spectra at
-141 °C show a monomer as a prominent species (ca. 34% of
total at 0.24 M in 3:2 THF/Et2O, KMD ) 12 M-1). In the Li
6
NMR spectra three additional signals were present (Figure 17).
These signals were assigned to dimers from analysis of variable
concentration experiments,15 and from the characteristic 13C line
shape and chemical shifts of the C-Li carbons (δ 187.5 and
188.0). Unfortunately, one variable-concentration experiment
gave aggregation as high as 3.5, so more work would need to
be done to conclusively rule out, for example, cyclic trimer
structures for one or more of the three signals.
Carbon-Carbon Coupling in the Side Chain of 4 and 5.
The information provided above provides strong circumstantial
evidence that 4 and 5 are chelated. We have attempted to provide
more direct evidence by measuring 2JCOC and 3JCCOC in the side
chain, in anticipation that the conformational changes of a
chelated vs a nonchelated structure might result in diagnostic
trends in these coupling constants. We prepared O-13CH3
9
3518 J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003

Amine- and Ether-Chelated Aryllithium Reagents
A R T I C L E S
Dynamic Processes for 6. A variable-temperature 6Li NMR
experiment15 showed that two of the dimers coalesced at -130
°C, the two remaining dimer signals at -120 °C, and finally
the averaged dimers coalesced with the monomer at -100 °C
(∆G‡
) 8.8 kcal/mol).
-107
Detailed assignment of the dimer signals of 6 was hampered
by the low temperatures needed and their peculiar properties.
We have labeled the three dimer signals X, Y, and Z, and the
monomer signal M (Figure 17). At -141 °C the Z-MeO 13C
NMR signal as well as the Z-6Li signals are substantially broader
than the others. By -129 °C the signal for Z-MeO has
sharpened a little, but has now begun to coalesce with the
Y-OMe signal. The Y and Z signals are already coalesced in
the 6Li NMR spectrum because of the much smaller separation
6
(13 Hz for Li vs 120 Hz for 13C) between the signals. This
process causes no detectable change in the C-Li carbons. The
X
°C, and finally the averaged dimer signals coalesce with
monomer between -105 and -96 °C. Presumably, the X, Y,
and Z signals correspond to two or three of the chelation isomers
A, B, and C.
13C signal coalesces with Y/Z between -118 °C and -105
Figure 18. (a) 6Li NMR spectra of a TMEDA titration experiment of 0.2
M [6Li]6 in 3:2 THF/Et2O at -138 °C to -145 °C. The upper trace of
each pair is a 7-spin DNMR simulation[1k] used to integrate the spectra. (b)
Graphical presentation of the concentrations of the components of the
equilibrium.
are presented in Figure 18b. As was observed in the more
qualitative analysis done for 2, TMEDA caused a modest level
of dimer dissociation when complexed to 6. This can be
expressed in terms of two pairs of equilibrium constants. First
the monomer-dimer association constants (t ) TMEDA):
KMD-THF ) [D]/[M]2 ) 8 ( 3 M-1, KMD-t ) [D‚t]/[M] [M‚t]
) 2.4 ( 0.3 M-1. Second, the TMEDA-affinity of monomer
The very low-temperature broadening of the Z signals
suggests additional complications from a slowing of some
conformational process of the methoxypropyl side chain (either
chelate ring inversion, chelation/dechelation or configurational
isomerization at tricoordinate oxygen) analogous to that detected
for 2-diethylaminomethylphenyllithium at very low temperatures.1h
In fact, the simplest explanation for the behavior in THF is that
there are only two dimer structures, and that the complex low-
temperature dynamic behavior is the result of conformational
processes. This would fit well with the spectra seen in Et2O
solution,15 where only two sets of signals (3:2 ratio) are seen
in both the Li (δ 1.5 and 1.7) and 13C NMR spectra (δC-Li 188.2,
186.8), which both appear to be dimers. The signals coalesce
between -110 °C and -90 °C.
Although chemical shift arguments are inherently weak, the
data in Figure 17 provide evidence that, like 2, the dimers of 6
are chelated, whereas the monomer is not. Consider the chemical
shifts of the MeO carbons: the various dimer signals are 1.5-4
ppm upfield of the signal of 3-(methoxypropyl)benzene (the
protonation product) at δ 58.75, suggesting enforced proximity
to the ring system, whereas that of the monomer (δ 58.57) is
within 0.18 ppm, and within 0.34 ppm of the signal of the
trimethylstannyl precursor (δ 58.91).
Interaction of 6 with TMEDA. The TMEDA titration is
shown in Figure 18. As with 2 (Figure 6) it was necessary to
measure the spectra below -140 °C to observe the individual
species. Three new signals appear. The one at δ 1.6 can be
assigned to the monomer-TMEDA complex since it coalesces
with the THF-monomer signal above -140 °C. The downfield
shift of 0.4 ppm is nearly identical to that observed for 2 and
PhLi when they are converted to TMEDA complexes. The other
two signals at δ 1.8 and 2.0 remain in a 1:1 ratio throughout
the titration. We assign these to the two signals of an A-type
dimer-TMEDA complex.
and dimer: KM-t ) [M‚t]/[M] [t] ) 9.5 ( 1.4 M-1; KD-t
)
[D‚t]/[D] [t] ) 3.0 ( 0.6 M-1. The TMEDA-complexed
aryllithium is about one-third as aggregated, and the monomer
has three times the TMEDA-affinity as does the dimer.
Interaction of 6 with PMDTA. The addition of PMDTA
stoichiometrically converted all signals in the 6Li and 13C NMR
spectra into signals of monomeric 6M‚PMDTA.15 This is
consistent with the behavior of all other ArLi compounds we
have investigated that are significantly monomeric in THF.
Chelation and Aggregation. The data reported here dem-
onstrate a broadly based qualitative (and in a few cases,
quantitative) correlation between chelation and the kinetic
stability and thermodynamic preference for dimers in amine and
ether-chelated aryllithium reagents. Table 1 summarizes the
thermodynamic and kinetic data we have collected. For two of
the compounds, the 5-ring chelates 11h and 4, no monomer was
detected by NMR spectroscopy in THF-ether solutions, so only
upper limits of KMD can be obtained. For three compounds (2,
5, and 6), both monomer and dimer were observed and both
thermodynamic and kinetic stability could be measured.
The free energy of dimerization increases by 3.2 kcal/mol
on going from the model system with an ortho ethyl group to
the 6-ring ether chelate 5 (increase in KMD by almost 5 orders
of magnitude), and by more than that for the 5-ring ether and
amine chelates (4 and 1). Similarly, the activation energy for
dissociation of monomers increase by 3 kcal/mol for 4, 4 kcal/
mol for 6 and 5 kcal/mol for 1 compared to the model. The
sequence of dimer stabilities from the available KMD values is:
1 g 4 > 5 > 2 > 6 > 3, 7. A similar order is also found for
the kinetic stability of those dimers for which DNMR techniques
could be used to measure rates (Table 1). Here, we find that
∆G‡ is in the order 1 > 5 g 4 > 2 > 6 > 7.
We have attempted to quantitate the interaction of 6 with
TMEDA by simulation of the spectra of Figure 18a. It was
necessary to do a DNMR simulation because of partial
coalescence between 6M‚TMEDA and 6M as well as between
the Y and Z dimer signals (Figure 17). The results of the analysis
9
J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003 3519

A R T I C L E S
Reich et al.
We have used the behavior with TMEDA and PMDTA as
probes into the effect of chelation ring size and chelating group11
on the structures of aryllithium reagents. TMEDA disturbs the
monomer-dimer equilibrium in only a minor way (showing,
e.g., about a factor of 3 preference for monomer over dimer in
2 and 6) and so can be used to probe the nature of the chelation
(A, B, or C structures, Scheme 2). Three of the structures which
are known or are likely to have A-type dimers present (1, 2,
and 6) readily complex with TMEDA to form A-type TMEDA
complexed dimers, with a characteristic 1:1 ratio of two signals
in the Li NMR spectra. The 5- and 6-ring oxygen chelates 4
and 5 do not detectably complex with TMEDA, and we
conclude that these must have a strong preference for B and/or
C-type dimers which cannot form bidentate complexes with
TMEDA.
Apparently, the small chelation ring size and the small steric
size of the methoxy chelating group allow higher coordination
than is usually observed for organolithium species.
We have previously discussed possible origins of the strong
correlation between aggregation and chelation in connection with
1.1h
This paper has provided a number of additional and more
specific examples of the effect. One plausible explanation
involves consideration of the smaller C-Li-X (X ) solvent
or chelating group) bite angle in a 5- or 6-ring chelate versus
an externally solvated one, which could lead to less cancellation
of the C-Li dipole by the Li-X dipoles, and thus a higher
propensity for aggregation. The smaller bite angle would also
result in less steric inhibition of dimerization. Some evidence
for a more polar C-Li bond in 5-ring amine-chelated organo-
lithiums than in nonchelated models was provided by the higher
HMPA-affinity of chelated Me2(Me2NCH2)SiCHLiSPh versus
the unchelated model Me3SiCHLiSPh. Unfortunately, no dif-
ference in HMPA affinity was seen between the 5-ring ether-
chelated compound, Me2(MeOCH2)SiCHLiSPh, and the model,1h
so this test does not support the bite-angle hypothesis for the
ethers. We do not currently have a convincing and experimen-
tally testable rationale for the correlation between aggregation
and chelation.
The behavior with PMDTA is particularly instructive. The
observation that excess PMDTA converts 2 through 6 com-
pletely to monomers, whereas TMEDA22d only minimally
increases the fraction of monomers, means that the PMDTA
complexes must be tridentate. (PMDTA is a poorer bidentate
ligand than TMEDA). In addition, it appears that, with the
exception of 4‚PMDTA, chelation is broken in the monomer-
PMDTA complexes. Thus, the fraction of such species formed
provides a qualitative measure of the strength of chelation
because simple steric effects of the ortho-substituent should be
similar through the series. This analysis could be performed
quantitatively for 5 (Scheme 4), which showed that the dimer
was substantially stabilized by chelation, whereas the monomer
was not. It could not be applied to 4, because the PMDTA
complex is chelated (chelating ligands for lithium also show
idiosyncratic behavior in the lithium amides22e). The nonch-
elating model 7 as well as 2, 3, and 6 show >96% conversion
to the monomeric complex with 1 equiv of PMDTA. Chelation
effects are very evident for 5 and 4, which were only partially
converted to monomer at 1 equiv of PMDTA.10 Compound 1
is most strongly affected, because it forms no detectable
monomer on addition of PMDTA.1h We conclude that the
sequence of chelation strength (∆GChel) is: 1 . 4 > 5 > 2, 6,
3, 7, basically in the same sequence as the kinetic and
thermodynamic stability of the dimers.
Summary
For these systems amine chelation is strong in five-membered
rings (1), competitive with THF solvent in six-membered rings
(2), and absent in seven-membered rings (3). Ether chelation is
of intermediate strength in both five- and six-membered rings,
with the five-membered ring chelate 4 slightly more robust by
most criteria.10 The 7-ring ether 6 appears to be weakly chelated.
There is a strong qualitative correlation between the strength
of the chelation, as judged by the effect of PMDTA, and the
strength of the aggregation, as judged by response to polar
cosolvents, magnitude of monomer-dimer association constant
K
MD, and rates of monomer-dimer equilibration.
Experimental Section
General. All reactions requiring a dry atmosphere were performed
in glassware flame-dried or dried overnight in a 110 °C oven, sealed
with septa and flushed with dry N2. Tetrahydrofuran (THF) and Et2O
were freshly distilled from sodium benzophenone ketyl under N2. Me2O
was purified by condensing several mL in a graduated conical flask at
-78 °C from a pressurized gas cylinder, adding a small portion (0.5
mL) of n-BuLi and distilling the dry Me2O via cannula into the desired
vessel at -78 °C. N,N,N′,N′-Tetramethylethylenediamine (TMEDA),
N,N,N′,N′′,N′′-pentamethyldiethylenetriamine (PMDTA) and hexam-
ethylphosphoric triamide (HMPA) were distilled from CaH2 under
reduced pressure (if necessary) and stored over 4 Å molecular sieves
under N2. Common lithium reagents were handled with septum and
syringe-based techniques and titrated against dry n-propanol in THF
using 1,10-phenanthroline as indicator.28 [6Li]BuLi16c and [6Li]EtLi22f
were prepared by literature methods.
Low-Temperature NMR Spectroscopy. All low-temperature NMR
spectra were acquired on Bruker AM-360 or AVANCE spectrometers
using a 10 mm broadband probe at the following frequencies: 360.148
MHz (1H), 90.556 MHz (13C), 52.984 MHz (6Li), 139.905 MHz (7Li)
and 145.785 MHz (31P). All spectra were taken with the spectrometer
unlocked. 13C NMR spectra were referenced internally to the C-O
carbon of THF (δ 67.96), Et2O (δ 66.57) or Me2O (δ 60.25). Lorentzian
multiplication (LB) of 2-3 Hz was applied to 13C spectra. 6Li and 7Li
The thermodynamics and dynamics of 15, the mixed dimer
between 5 and PhLi, also provide a clear-cut example of the
close relationship between chelation and aggregation. The
monomer-dimer association constants (KMD) increase by about
an order of magnitude on going from the PhLi homodimer, to
the mixed dimer 15 (with one chelating group) and then to the
homodimer of 5 (with two chelating groups). This although a
nonchelating ortho alkyl substituent reduces KMD by over 2
orders of magnitude (compare PhLi and 7, Table 1).
The relationship between aggregation and chelation is strik-
ingly illustrated by the behavior of 2, where both monomer and
dimer can be observed. The dimer is unambiguously and fully
chelated, whereas the monomer is unchelated. Although the
evidence is weaker, this also seems to be the case for 5 and 6.
Compound 4 shows several kinds of anomalous behaviors,
most notably the formation of an open dimer PMDTA complex12c
12 not detected for any other aryllithium investigated. Also
interesting is the observation that the PMDTA complex of 4 is
tridentate and chelated, representing a stable pentacoordinate
lithium structure, albeit one with an unusually labile ligand.
(28) Watson, S. C.; Eastham, J. F. J. Organomet. Chem. 1967, 9, 165-168.
9
3520 J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003

Amine- and Ether-Chelated Aryllithium Reagents
A R T I C L E S
spectra were referenced externally to 0.3 M LiCl in MeOH (δ 0.00) or
internally to Li+(HMPA)4 (δ -0.40). 31P NMR spectra were referenced
externally to 1.0 M PPh3 in THF (δ -6.00) or internally to free HMPA
(δ 26.40). Probe temperatures were measured externally by ejecting
the sample and inserting a thermocouple into the probe or internally
with the 13C chemical shift thermometer (Me3Si)3CH.1j
Direct Preparation of Sample in 3:2 THF/Et2O. The aryllithium
precursor aryltrimethylstannane (0.50 mmol), 1.8 mL of THF and 1.2
mL of Et2O were added to a dried and N2-flushed 10 mm NMR tube.
After cooling to -78 °C, 0.63 mL of 0.79 M n-Bu6Li (or Et6Li) in
hexanes (0.50 mmol, 1.00 equiv) was added to give a yellow solution
with an initial aryllithium concentration of 0.14 M. The sample was
stored for 30 min at -20 °C to ensure the lithium-tin exchange was
complete before beginning the NMR experiment.
General Procedure for NMR Spectroscopy of Organolithium
Reagents. Samples of organolithium reagents (0.5 mmol in 3 mL of
solvent) were prepared in thin-walled 10 mm NMR tubes that were
oven-dried overnight, fitted with 9 mm i.d. septa, and flushed with N2.
Silicon grease was applied to the interface between the tube and the
septa before securing with Parafilm for a better seal, as well as to the
top of the septa to seal needle punctures. Samples were stored at -78
°C. After initial adjustment of the shim values the probe was cooled to
-120 °C, the sample was inserted and the sample was shimmed on
the 13C FID of one of the solvent peaks (THF, Et2O or Me2O). Spectra
Quantitation of Arylithium Reagents. The concentrations of
aryllithium solutions were determined by the following methods. (1)
In some cases when the lithium reagent was prepared by Li/Sn
exchange, quantitative transmetalation was assumed. (2) Quenching of
the aryllithium reagent with Me2S2, Me3SiCl or PhCHO followed by
quantitation of the sulfide, silane or benzyl alcohol by 1H NMR
spectroscopy and/or GC analysis were used to determine the yield and
demonstrate the location of the carbanion during the experiment. For
example, the NMR sample was quenched with 1.5 equiv of Me3SiCl
to give the corresponding arylsilane, diluted with 20 mL of 1:1 ether/
hexanes, washed with 100 mL of distilled water or saturated Na2CO3
and 100 mL of brine. The organic fraction was dried over MgSO4,
filtered and concentrated in vacuo. The yield of the arylsilane was
determined by 1H NMR integration against pentachloroethane as a
standard. (3) HMPA has been shown to bind aryllithium reagents in a
stoichiometric fashion in THF and Et2O solutions,1h thus initial addition
of HMPA was assumed to complex aryllithium stoichiometrically.
Integration of the complexed versus uncomplexed aryllithium species
in the 6Li NMR spectrum allowed the solution concentration to be
determined. Stoichiometric HMPA complexation was confirmed by
examining the 31P NMR spectrum for free HMPA.
6
7
of NMR active nuclei which usually included 13C, Li, Li, and 31P
were measured. At this point, a cosolvent titration, variable temperature
(VT) or variable concentration experiment could be performed. In the
case of a titration experiment, for each addition the sample was ejected,
placed in a -78 °C bath, the silicon grease was removed from the top
of the septum, a desired amount of cosolvent was added, silicon grease
was reapplied to the top of the septum and the desired NMR spectra
were measured. Probe temperatures were measured internally with the
13C chemical shift thermometer (Me3Si)3CH1j or externally by ejecting
the sample and inserting a thermocouple into the probe.
The use of 10 mm NMR tubes provided a real advantage over 5
mm tubes because of the approximate 3-fold signal-to-noise enhance-
ment with more sample in the magnetic field, more accurate additions
of cosolvents, and minimal sample warming when transferring tubes
from the -78 °C bath to the NMR probe.
Dynamic NMR Simulation. Many of the variable temperature NMR
spectra were simulated to determine rate constants, activation parameters
and/or peak areas. These simulations were performed with the computer
program WINDNMR,1k using one of the standard exchange matrixes
(random exchange of singlets) or custom setups for more complex
situations.
General Procedures for Preparation of Aryllithium Reagents for
NMR Spectroscopic Study. Crystallization from Diethyl or Dim-
ethyl Ether. The precursor aryltrimethylstannane was added to a dried
and N2-flushed 10 mm NMR tube. The tube was cooled to -78 °C
and 1 mL of Me2O (or Et2O, when noted) was condensed into it. Me2O
was an ideal solvent for this crystallization method, because the lithium/
tin exchange was relatively fast and the aryllithium reagents were
usually not very soluble in pure Me2O. The NMR tube was temporarily
removed from the -78 °C bath, warmed slightly, and shaken to dissolve
the arylstannane. The addition of 1.0 equiv of n-Bu6Li to the NMR
tube gave a homogeneous yellow solution. The NMR tube was stored
overnight at -78 °C, while the aryllithium reagent crystallized from
solution. The yellow supernatant was removed by cannula transfer and
the crystals were washed with Et2O (2 × 0.5 mL) at -78 °C. The
crystals were dissolved in a combination of THF and Et2O, by warming
the solvent mixture, if necessary to room temperature, and vigorously
shaking the NMR tube, to give a clear, colorless solution of aryllithium
reagent with concentrations ranging from 0.08 to 0.16 M. The NMR
tube was returned to the -78 °C bath, and Me2O was added if needed.
The sample was ready for NMR investigation.
Acknowledgment. We thank NSF for financial support
(CHE-0074657) and funding for instrumentation (NSF CHE-
9709065, CHE-9304546). Prof. Gideon Fraenkel provided
assistance with setting up exchange matrixes for the DNMR
simulation of the C-Li exchange of 4.
Supporting Information Available: Experimental procedures
for the preparation of compounds and NMR samples, 6Li, 13C,
and 31P NMR spectra of NMR experiments not presented in
the main body of this paper. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA028301R
9
J. AM. CHEM. SOC. VOL. 125, NO. 12, 2003 3521