Perturbation of Conjugation in Allylic Lithium Compounds Due to Stereochemical Control of Internal Lithium Coordination: Crystallography, NMR, and Calculational Studies

Article abstract of DOI:10.1021/ja030370y

Several allylic lithium compounds have been prepared with ligands tethered at C₂. These are with (CH₃OCH₂CH ₂)₂NCH₂-, 6, 1-TMS 5, 1,3-bis(TMS) 8, and 1,1,3-tris(TMS) 9. Allylic lithiums with (CH₃-OCH₂CH ₂)₂NCH₂C(CH₃)₂-, are 10, 1-TMS 11, and 1,3-bis(TMS), 12 compounds with -C(CH₃) ₂CH₂N-((S)-(2-methoxymethyl)-pyrrolidino) at C ₂ 13, 1-TMS 14, and 1,3-bis(TMS) 15. In the solid state, 8-10 and 12 are monomers, 6 and 13 are Li-bridged dimers, and 5 and 7 are polymers. In solution (NMR data), 5, 7-12, 14, and 15 are monmeric, and 6 is a dimer. All samples show lithium to be closest to one of the terminal allyl carbons in the crystal structures and to exhibit one-bond ¹³C-⁷Li or ¹³C₁-⁷Li spin coupling, for the former typically ca. 3 Hz and for the latter 6-8 Hz. In every structure, the C ₁-C₂ allyl bond is longer than the C₂-C ₃ bond, and both lie between those for solvated delocalized and unsolvated localized allylic lithium compounds, respectively, as is also the case for the terminal allyl ¹³C NMR shifts. Lithium lies 40-70° off the axis perpendicular to the allyl plane at C₁. These effects are variable, so the trend is that the differences between the C ₁-C₂ and C₂-C₃ bond lengths, ¹³δ₃-¹³δ₁ values, and the ¹³C₁-⁷Li or ¹³C-⁶Li coupling constants all increase with decreasing values of the torsional angle that C₁-Li makes with respect to the allyl plane.

Full text of DOI:10.1021/ja030370y

Published on Web 03/04/2004
Perturbation of Conjugation in Allylic Lithium Compounds
Due to Stereochemical Control of Internal Lithium
Coordination: Crystallography, NMR, and Calculational
Studies
Gideon Fraenkel,* Albert Chow, Roland Fleischer, and Hua Liu
Contribution from the Department of Chemistry, The Ohio State UniVersity,
100 West 18th AVenue, Columbus, Ohio 43210-1185
Received June 20, 2003; E-mail: fraenkel@chemistry.ohio-state.edu
Abstract: Several allylic lithium compounds have been prepared with ligands tethered at C2. These are
with (CH3OCH2CH2)2NCH2-, 6, 1-TMS 5, 1,3-bis(TMS) 8, and 1,1,3-tris(TMS) 9. Allylic lithiums with (CH3-
OCH2CH2)2NCH2C(CH3)2-, are 10, 1-TMS 11, and 1,3-bis(TMS), 12 compounds with -C(CH3)2CH2N-
((S)-(2-methoxymethyl)-pyrrolidino) at C2 13, 1-TMS 14, and 1,3-bis(TMS) 15. In the solid state, 8-10 and
12 are monomers, 6 and 13 are Li-bridged dimers, and 5 and 7 are polymers. In solution (NMR data), 5,
7-12, 14, and 15 are monmeric, and 6 is a dimer. All samples show lithium to be closest to one of the
terminal allyl carbons in the crystal structures and to exhibit one-bond 13C-7Li or 13C1-7Li spin coupling,
for the former typically ca. 3 Hz and for the latter 6-8 Hz. In every structure, the C1-C2 allyl bond is longer
than the C2-C3 bond, and both lie between those for solvated delocalized and unsolvated localized allylic
lithium compounds, respectively, as is also the case for the terminal allyl 13C NMR shifts. Lithium lies 40-
70° off the axis perpendicular to the allyl plane at C1. These effects are variable, so the trend is that the
differences between the C1-C2 and C2-C3 bond lengths, 13δ3-13δ1 values, and the 13C1-7Li or 13C-6Li
coupling constants all increase with decreasing values of the torsional angle that C1-Li makes with respect
to the allyl plane. In the solid state, all internal ligands are fully coordinated to lithium and none of the X-ray
structures incorporate solvent. Also, in the solid-state monomers 8-10 and 12, Li lies near the center of
one face of a trigonal pyramid with two methoxy oxygens and nitrogen at its corners. Nitrogen occupies
the apex. The above results support a variable degree of delocalization in the allylic moiety due to the
proximity of lithium to one of the terminal allyl carbons. The relative contribution of ionic and covalent
interactions to the C-Li bond have yet to be determined. The ligand tether constrains the stereochemistry
of lithium coordination, thus stabilizing these otherwise abberant structures. Dynamics of inversion at lithium-
bound carbon obtained from changes in the 13C1 NMR line shapes due to the ligands gave ∆Hq values
(kcal/mol) of 6.5 ( 0.4 and 15 ( 1.3 for ligand tethers of one and two carbons, respectively. Averaging of
the 13C1-7Li or 13C1-6Li coupling constants observed with increasing temperature, which is due to
bimolecular C1-Li exchange, typically involves ∆Hq values of 12 ( 0.5 kcal/mol. In the case of symmetrically
substituted compounds 6, 8, 10, 12, and 15 13C NMR line shape changes provide the dynamics of lithium
1,3-sigmatropic shifts. The rates vary widely among the compounds studied, most likely due to steric effects.
Allylic lithium compounds are among the simplest potentially
conjugated carbanionic species. As a result of extensive chemi-
cal,1 spectroscopic,2 crystallographic,3 and calculational4 studies,
two principal types of structures have been identified. Solvated
allylic compounds consist of contact ion-pairs. The anion is
delocalized,2,3 and coordinated Li+ is sited on the axis perpen-
dicular with and close to the center of the allyl plane, 13. By
contrast, unsolvated allylic lithium compounds are regarded as
localized species since their NMR spectra so closely resemble
those of alkenes;5 see 2a and 2b5. Until recently, structures that
lie between 1 and 2 have not been recognized.
(3) (a) Koster, H.; Weiss, E. Chem. Ber. 1982, 115, 3422. (b) Schumann, U.;
Weiss, E.; Dietrich, H.; Mahdi, W. J. Organomet. Chem. 1987, 322, 299.
(c) Sebastian, J. F.; Grunwell, J. R.; Hsu, B. J. Organomet. Chem. 1974,
78, C1. (d) Boche, G.; Etzrodt, H.; Marsch, M.; Massa, H.; Baum, G.;
Dietrich, H.; Mahdi, W. Angew. Chem. 1986, 98, 84. (e) Boche, G.;
Fraenkel, G.; Cabral, J.; Harms, K.; Eikema-Hommes, N. J. P. Van.; Lorenz,
J.; Marsch, M.; Schleyer, P. v. R. J. Am. Chem. Soc. 1992, 114, 1562-1565.
(4) (a) Erusalimski, C. B.; Kormer, V. H. Zh. Org. Khim. 1984, 20, 208. (b)
Tidwell, E. R.; Russell, B. R. J. Organomet. Chem. 1974, 80, 175. (c)
Boche, G.; Decher, G. J. Organomet. Chem. 1983, 259, 31. (d) Clarke, T.;
Jemmis, E. D.; Schleyer, P. v. R.; Binkley, J. S.; Pople, J. A. J. Organomet.
Chem. 1978, 150, 1. (e) Clarke, T.; Rhode, C.; Schleyer, P. v. R.
Organometallics 1983, 2, 1344. (f) Bushby, R. J.; Tytho, M. P. J.
Organomet. Chem. 1984, 270, 265. (g) Pratt, L. M.; Khan, I. M. J. Comput.
Chem. 1995, 16, 1070. (h) Eikemma-Hommes, N. J. R. Van; Bu¨hl, M.;
Schleyer, P. v. R. J. Organomet. Chem. 1991, 409, 307-320.
(1) (a) Wardell, J. L. ComprehensiVe Organometallic Chemistry; Wilkinson,
G.; Stone, F. G. H.; Abel, E. W.; Eds.; Pergammon Press: Oxford, 1982;
Vol. 7, p 97. (b) Seyferth, D.; Julia, T. F. J. Organomet. Chem. 1967, 8,
C13. (c) Schlosser, M.; Stahle, N. Angew. Chem. 1980, 92, 477. (d) Stahle,
M.; Schlosser, M. J. Organomet. Chem. 1981, 220, 277. (e) Neugebauer,
W.; Schleyer, P. v. R. J. Organomet. Chem. 1980, 198, C1. (f) Brownstein,
S.; Bywater, S.; Worsfold, D. J. J. Organomet. Chem. 1980, 199, 1. (g)
Gu¨nther, H. Encyclopedia of NMR 1996, 5, 2807.
(2) (a) West, P.; Purmort, J. I.; McKinley, S. V. J. Am. Chem. Soc. 1968, 90,
797. (b) O’Brian, D. H.; Hart, A. J.; Russell, C. R. J. Am. Chem. Soc.
1975, 97, 4410. (c) Benn, R.; Rufinska, A. J. Organomet. Chem. 1982,
239, C19. (d) Bates, R. B.; Beavers, W. J. Am. Chem. Soc. 1974, 96, 5001.
(e) Dolinskaya, E. R.; Poddabnyi, I. Ya.; Tseretech, I. Yu. Dokl. Akad.
Nauk. SSSR 1970, 191, 802. (f) Thompson, T. B.; Ford, W. T. J. Am. Chem.
Soc. 1974, 101, 5459. (g) Gu¨nther, H. J. Brazil Chem. Soc. 1999, 10, 241.
9
10.1021/ja030370y CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 3983-13
3983

A R T I C L E S
Fraenkel et al.
compounds continuously variable? How does this change with
the length of the ligand tether and the nature of substitution on
the allyl moiety? Clearly, quantitative structural information is
required.
Herein we report crystal structures for several of these
internally solvated allylic lithium compounds together with some
new NMR data that provide unusual insight into dynamic
behavior. These X-ray crystallographic results confirm in
quantitative fashion the qualitative conclusions we have made,
which are based on the results of our NMR studies. We show
below how changes among the structural parameters correlate
with trends among the NMR data and how the degree of
delocalization in these internally solvated allylic lithium com-
pounds is, in fact, continuously variable.
Our low-temperature NMR line shape analysis of solvated
allylic lithium compounds has revealed the dynamics of a rich
variety of fast molecular reorganization processes at equilibri-
um.6 These include rotation around the C-C allyl bonds, transfer
of coordinated lithium between faces of the plane (inversion),
rotation of coordinated lithium on one side of the allyl plane,
first-order reversible lithium ligand dissociation, and bimolecular
lithium ligand exchange.
Results and Discussion
New crystallographic and NMR data are herein reported on
5 (see above) and on 6-15 (listed in Table 1 with 13C NMR
shifts).
The preparations of starting materials, not previously pub-
lished, are outlined in the reactions 16 to 17, 18 to 19, 18 to
20, 21 to 22 to 23, 23 to 24, and 24 to 25 (Scheme 1).
NMR data are reported herein for compounds 6, 9-12, 14,
and 15 (see Table 1). The samples were prepared by metalating
compounds 16-20, 24, and 25, respectively, with n-butyllithium
in ether/hexane or THF/hexane. Then, solvents were replaced
by THF-d8 or diethyl-ether-d10.
For the X-ray structure determinations, crystals of the allylic
lithium compounds 5-10, 12, and 13 developed after metalation
of their precursors using n-butyllithium in hexanes.
Salient Features of the Crystallographic Data. The crystal-
lographic structures of compounds 5-10, 12, and 13 correlate
with the NMR data. These compounds are monomers in the
solid state except for 6 and 13, which are lithium-bridged dimers,
and 5 and 7 are polymers in the solid state. A Hirschfeld rigid
bond analysis of our data,9a part of the PLATON software
package,9b revealed nothing out of the ordinary among the
thermal parameters. Selected structural and NMR parameters
are listed in Table 2.
All the crystallographic structures exhibit full coordination
of the pendant ligand to lithium and localized C-Li bonding.
The crystal structure of a typical internally solvated monomer,
compound 8, is shown in Figure 1. Specific details of structure
and bonding are outlined in the following sections.
Bond Lengths and Substitution. In all the crystal structures
reported herein, C1 C2, C3, and C4 are coplanar. The numbering
is indicated around structure 5. Where a terminal allyl carbon
carries a single silyl substituent, the C-Si bond lies close to
the allyl plane. In 9 with geminal silyl substituents at C1, these
lie well separated on opposite sides of the allyl plane (see Table
2).
In experiments to slow these fast reorientational processes
and thus facilitate the study of ion-pair structure, we encapsu-
lated Li+ in an allylic lithium compound by coordinating it to
an attached ligand.7 Using such an approach, we determined
the first structure of an internally solvated ion-pair, 4, using
NMR methods.7 Subsequently, we prepared several allylic
lithium compounds with the pendant ligand attached to the
center carbon of the allylic moiety.8a-c By contrast to 4, these
new compounds (see for example 5) were not ion-paired salts.
Their terminal allyl 13C NMR shifts all lie between those for 1
and 2.8 They all displayed hitherto unprecedented spin coupling
in an allylic lithium compound between one of the 13C termini
and 6Li, indicating a small detectable degree of C-Li covalence.
We proposed that the ligand tether in these compounds is too
short to place lithium on the axis normal to the center of the
allyl plane, as is the case in the delocalized structures, 1. Instead,
the ligand tether sites lithium off the axis perpendicular to the
allyl plane at C1. This restricts the stereochemistry of lithium
coordination. The proximity of lithium to C1 induces partial
localization of the allylic moiety.8b,c
Compounds such as 5 also displayed interesting fast reorga-
nization processes at equilibrium such as inversion at lithium-
bound carbon, bimolecular C1-Li exchange, and 1,3 lithium
sigmatropic shifts.8b,c
Given that internal coordination of an otherwise delocalized
organolithium compound can perturb its electronic structure, it
is of interest to investigate the extent of the above-described
effects: Is the degree of delocalization of our allylic lithium
(6) (a) Fraenkel, G.; Chow, A.; Winchester, W. R. J. Am. Chem. Soc. 1990,
112, 1382-1386. (b) Fraenkel, G.; Winchester, W. A.; Chow, A. J. Am.
Chem. Soc. 1990, 112, 2582-2585. (c) Fraenkel, G.; Cabral, J. A. J. Am.
Chem. Soc. 1993, 115, 1551-1557. (d) Fraenkel, G.; Cabral, J. J. Am.
Chem. Soc. 1992, 114, 9007-9015. (e) Fraenkel, G.; Cabral, J.; Lanter,
C.; Wang, J. J. Am. Chem. Soc. 1999, 64, 1302-1310.
(7) Fraenkel, G.; Cabral, J. A. J. Am. Chem. Soc. 1993, 115, 1551.
(8) (a) Fraenkel, G.; Qiu, F. J. Am. Chem. Soc. 1997, 119, 3571-3579. (b)
Fraenkel, G.; Qiu, F. J. Am. Chem. Soc. 1996, 118, 5828-5829. (c)
Fraenkel, G.; Duncan, J. H.; Wang, J. J. Am. Chem. Soc. 1999, 121, 432-
443. (d) Fraenkel, G.; Martin, K. J. Am. Chem. Soc. 1995, 117, 10336-
10344.
(5) (a) Fraenkel, G.; Halasa, A. F.; Mochel, V.; Stumpe, R.; Tate, D. J. Org.
Chem. 1985, 50, 4563-4565. (b) Glaze, W. H.; Jones. P. C. J. Chem. Soc.
1969, 1434. (c) Glaze, W. H.; Hanicac, J. E.; Moore, M. L.; Chandpuri, J.
J. Organomet. Chem. 1972, 44, 39. (d) Glaze, W. H.; Hanicac, J. E.;
Chandpuri, J.; Moore, M. L.; Duncan, D. P. J. Organomet. Chem. 1973,
51, 13.
(9) (a) Hirschfeld, F. L. Acta Crystallogr. 1976, A32, 239. (b) Spek, A. L. J.
Appl. Crystallogr. 2003, 36, 7.
9
3984 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Conjugation in Allylic Lithium Compounds
A R T I C L E S
Table 1. Structures of Internally Solvated Allylic Lithium Compounds with 13C NMR Shifts
a NMR previously reported. b Subject to fast sigmatropic shift; see text. c Did not crystallize; d Two stereoisomers in a 1:1 ratio. e Insoluble. f Two
stereoisomers in a 2:1 ratio; shift for major species is indicated by top value of each stack.
The two allyl C-C bond lengths in simple externally solvated
allylic lithium compounds are very similar,3 as are also the
terminal allyl 13C NMR shifts.2 By contrast, within each of the
internally solvated allylic lithium compounds, both the allyl bond
lengths and the terminal allyl 13C shifts are significantly different
(see Table 2). The values of these bond lengths and shifts lie
between those of delocalized 1 and values for simple substituted
alkenes, which we assume resemble the bond lengths of 2. The
localized allyl 13C shifts of 2 are very similar to those of alkenes.
The data for the compounds reported in this article indicate a
trend for the chemical shift differences 13δ3-13δ1 to increase
with differences between the C1-C2 and C2-C3 bond lengths.
These differences are largest for 9, the allyl 13C shifts and bond
distances of which closely resemble those of an alkene. This is
due to steric interactions associated with the geminal silyl
substituents at C1 (see specific details below).
The crystallographic structures of 5 and 7 are exceptions to
the trend desribed above. These compounds are polymers in
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3985

A R T I C L E S
Scheme 1 a-i
Fraenkel et al.
a n-BuLi 3 equiv, hexane/Et2O, 0 °C; bTMSCl 3 equiv, -78 °C; cn-BuLi 1 equiv, hexane/Et2O, 0 °C; dTMSCl 1 equiv, -78 °C; en-BuLi 2 equiv,
f
g
h
i
hexane/Et2O, 0 °C; TMSCl, 2 equiv, -78 °C; SOCl2, 85 °C; (S)-(+)-2-methoxymethylpyrrolidine; LAH, THF, reflux.
the solid state due to a fifth coordination between lithium and
the terminal methylene carbon of a neighboring allylic lithium
unit. As a consequence, the two allyl bond lengths C1-C2 and
C2-C3 in these compounds are quite similar. Figure 2 illustrates
this bonding arrangement in the crystal structure of compound
5. However, the results of NMR studies show that, in solution,
5 and 7 appear to be monomers.8a-c
compound is quite close to the average allylic C-C bond lengths
in the corresponding externally solvated delocalized allylic
lithium compound. For example, the average bond lengths for
the unsubstituted internally solvated compounds 6 and 10, 1.392
(2) and 1.396 (2) Å, respectively (see Table 2), compare closely
to the 1.370 Å value reported for the allyllithium pentamethyl-
diethylenetriamine3b (PMDTA) complex, despite the fact that
6 is a dimer while allyllithium‚PMDTA and 10 are monomers.
In similar fashion, the average allylic C-C bond length of 1.391
It is interesting that the average of the C1-C2 and C2-C3
bond lengths within each internally solvated allylic lithium
9
3986 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Conjugation in Allylic Lithium Compounds
A R T I C L E S
Table 2. X-ray Structural and NMR Parameters of Internally Solvated Allylic Lithium Compounds (Note: M ) Monomer, D ) Dimer, P )
Polymer; for Numbering See Structure 5)
bond lengths (Å)
13C shifts
torsional anglesa
Si1 C1 C2
C0
solid
solution
J 13
C−Li (Hz)
C1−
C2 C2
−C3
C1
C3
Li−C1
−C2
−
C0
−
−
−
Si3−C3−C2−
C0
4
4
4
6
5
7
8
9
D
P
P
M
M
D
b
1.436(4)c
1.397(4)
1.415(7)
1.431(3)
1.494(7)
1.349(1)c
1.361(4)
1.351(8)
1.351(3)
1.306(8)
58.55b
41.10
42.81
54.12
37.01
58.55b
76.30
72.95
78.10
118.88
-50.12(0.03)c
-55.27(0.25)
47.8(5)
52.97(0.2)
-30.3(0.61)
M
M
M
M
3.0 6Li
7.0 7Li
6.1 7Li
15.9 7Li
173.59(0.19)
179.6(4)
176.7(0.16)
-5.32(0.34)
-9.13(1.12)
82.03(0.58)d
-142.37(0.48)d
10
11
12
M
e
M
M
M
M
b
1.426(2)
e
1.415(8)
1.366(2)
e
1.368(20)
51.62b
37.50
51.62b
72.52
73.46(0.14)
e
-90.9(3)
8.1 7Li
4.6 7Lif
4.1 7Lif
g
e
67.33b
69.11b
171.3(3)
15.1(7)
13
14
15
D
e
e
g
1.415(8)c
e
e
1.368(20)c
e
e
g
g
-103.58(3)c
e
e
M
8.7 7Li
5.1 7Li
37.73
53.64h
65.47h
74.24
80.15h
80.15h
e
e
Mh
e
a Torsional angle with respect to allyl plane at C2-C4. b Exchange averaged. c Weighted average over two dimers. d Refers to 1,1-bis(TMS). e Did not
crystallize. f Averaged values; see text. g Insoluble. h Two stereoisomers.
Figure 2. ORTEP of two units of 5 showing 50% thermal ellipsoids and
selected atom labels. Hydrogen atoms have been omitted for clarity.
though the full range is between 70 and 100° (see Table 2).
Figure 1. ORTEP diagram of compound 8 showing 50% thermal ellipsoids,
The dihedral angle that C1-Li makes with respect to the allyl
plane on the ligand side, Li-C1-C2-C4, depends on the ligand
tether as well as on substitution at the allyl termini. In
compounds 5-8, this torsional angle, Li-C1-C2-C4, is close
to 50°. All four compounds, 5-8, incorporate a one-carbon
ligand tether. Compound 6 is unsubstituted at the allyl termini,
whereas in 5, 7, and 8 there is a single substituent at C1 and the
C1-Si bond lies close to the allyl plane. By contrast, in 9, also
with a one-carbon ligand tether, there are two silyl substituents
at C1, which are well separated by 135° on opposite sides of
the allyl plane. The lower Li-C1-C2-C4 torsional angle of
30.03° minimizes potential steric interactions between the silyl
substituents and the ligand. By contrast to compounds 5-8, the
longer two-carbon ligand tether in 10, 12, and 13 allows the
Li-C1-C2-C4 torsional angle to increase, respectively, to
73.46, 90.01, and 101° (see Table 2).
Lithium Coordination. The coordination geometry around
lithium is unusual among these internally solvated allylic lithium
compounds. In compounds that are monomers in the solid state,
8-10 and 12, lithium lies close to the center of one face of a
trigonal pyramid with C1 and the two methoxy oxygens at its
corners. Nitrogen occupies the apex (see Figure 3a). A similar
scheme applies to 5 and 7 with the exception that these com-
pounds are polymerized in the solid state due to a fifth coor-
dination from lithium to the allyl methylene carbon of a neigh-
boring allylic lithium unit (see Figure 3b). Lithium lies 0.4 Å
and selected atoms are labeled. Hydrogen atoms have been omitted for
clarity.
Å in internally solvated 1,3-disilyl 8 (see Table 2) is close to
the corresponding value of 1.403 Å in externally solvated 26.3e
Site of Lithium. In delocalized externally solvated allylic
lithium compounds, lithium lies perpendicularly on the axis to
be close to the center of the allyl plane.3 These compounds have
not been reported to exhibit 13C-7Li or 13C-6Li spin coupling.2
By contrast, X-ray crystallography of 6-10, 12, and 13 shows
lithium to be closer to the terminus, defined as C1, of the longer
C-C allyl bond than to the other two allyl carbons. Note that
almost all the solution 13C NMR data for these internally sol-
vated allylic lithium compounds show the more shielded of the
allyl 13C termini to be spin coupled to 6Li or 7Li. This matches
the latter carbon to C1 in the X-ray crystallographic data. Thus,
X-ray crystallography and NMR together strongly support our
proposal that these compounds are partially localized.8a-c
Further insight into these structures is provided by angular
relationships. The Li-C1-C2 angle is typically around 90°,
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3987

A R T I C L E S
Fraenkel et al.
compound are diastereotopic at 170 K and average with
increasing temperature above 170 K. Line shape analysis of the
latter methyl resonance as well as the ligand resonances gave
rise to the same rate constants.8c This also establishes that
configurational inversion at C1 in these compounds is a much
faster process than inversion at nitrogen alone.
Figure 3. (a) Coordination around lithium in monomeric 8-10 and 12.
(b) Coordination around lithium in polymeric 5 and 7.
Table 3. Internally Solvated Allylic Lithium Compounds: Bond
Distances around Lithium, Å
N−Li
Li
−C1
Li−O1
Li−O2
Li−C3*
Typical ∆Hq values for inversion among internally solvated
allylic lithium compounds are 6.5 ( 0.4 kcal/mol for a one-
carbon tether and 15 ( 1.3 kcal/mol when the tether is two
carbons long, as in compound 11 below (see Table 5).
Fast configurational inversion may also result from sigma-
tropic shifts in symmetrical compounds, which is discussed next.
1,3 Lithium Sigmatropic Shifts. We previously reported that
on warming compound 8 between 260 and 330 K, the two
trimethylsilyl 13C resonances progressively average as do also
the terminal allyl 13C resonances.8b This is due to a fast 1,3
lithium sigmatropic shift (see eq 2). These shifts are a potential
feature of the dynamic behavior of our symmetrically substituted
and unsubstituted internally solvated allylic lithium compounds.
In such latter cases, where the pendant ligand is -N(CH2CH2-
OCH3)2, the sigmatropic shift also brings about configurational
inversion (see eq 2).
6a
2.147(7)
2.138(5)
2.182(10)
2.041(4)
2.009(11)
2.065(3)
2.081(7)
2.317(8)
2.316(5)
2.515(11)
2.186(4)
2.171(12)
2.158(3)
2.220(8)
2.087(7)
2.145(5)
2.138(10)
1.968(4)
2.047(12)
2.038(3)
2.025(8)
2.15(7)
2.382(7)
2.442(5)
2.481(11)
5b
2.375(5)
2.117(10)
2.035(4)
2.012(11)
2.002(3)
2.025(7)
7b
8c
9c
10c
12c
a Dimer; see structure 6 weighted averages over two dimers. b Polymers,
C3* is C3 of nearest neighbor unit; for numbering, see Figure 3b.
c Monomers; see Figure 3a.
on the opposite side of nitrogen with respect to the C1-O1-O2
plane within the resulting distorted trigonal bipyramid (see Figure
3b). Of the two C-Li bonds associated with 5, C*-Li (external)
at 2.442(5) Å is significantly longer than C-Li (internal) at
2.316(5) Å, whereas in 7 the corresponding bonds are of similar
lengths, C1-Li 2.515(11) Å (internal) and C*-Li 2.48 (11) Å
(external). The two external angles N-Li-C* in 5 and 7 are,
respectively 147.5 and 168.2° (see Figure 3b and Tables 2-4).
Influence of Reorganization Dynamics on NMR Spectra.
Comparisons of X-ray crystallographic and NMR data for our
internally solvated allylic lithium compounds indicate that in
most cases the solid- and solution-state structures are very
similar. On the basis of our previous NMR studies of 5, 7, and
8 and the crystallographic results reported herein, one would
expect compounds 6 and 9-15 to all exhibit 13C-6Li or 13C-
7Li spin coupling and that, in the 13C NMR of each compound,
all the carbons should be magnetically nonequivalent. That this
is not always the case is largely due to the operation of one or
more dynamic effects. Thus, as seen below, NMR alone is
inadequate to reveal structures and unravel dynamic behavior
for several compounds described herein. Below, we first
summarize the influence of different dynamic phenomena on
the NMR of our internally solvated allylic lithium compounds.
Then, by combining NMR and crystallographic data we will
individually discuss and interpret the results for compounds 6,
10, 13, 8, 12, 11, and 14, in the preceding order, in terms of
their structures and the nature of their reorganization processes.
Inversion Dynamics. At low temperature (170 K), all carbons
of the N(CH2CH2OCH3)2 pendant ligand in 5 are magnetically
nonequivalent and the 13C resonance for each of CH3O, CH2O,
and NCH2C consists of an equal doublet. When the temperature
is increased above 170 K, these doublets progressively average
to single lines at their respective centers.8b NMR line shape
analysis of each of these averaging doublets gives rise to the
same rate constants. This effect was ascribed to fast transfer of
coordinated Li+ between faces of the allyl plane, essentially
inversion at C1. This proposal was confirmed by similar 13C
NMR data obtained with an analogue of 5 using 27 (X )
(CH3)2SiCH2CH3). The geminal silyl methyls in the latter
Details regarding the dynamics of sigmatropic shifts in
compounds 6, 10, and 13 (unsubstituted) followed by 8, 12,
and 15 (1,3-disubstituted) are outlined and discussed separately
below. The rates of these sigmatropic shifts vary widely among
the latter compounds.
13C-6Li and 13C- 7Li Spin Coupling and its Perturbation
7
by the Dynamics of C-Li Exchange and Li Relaxation.
Most of the monomeric compounds described herein exhibit
one-bond 13C-7Li or 13C-6Li spin coupling; typical coupling
constants are 7.5 ( 0.5 and 2.8 ( 0.2 Hz,8 respectively (see
Table 2). These values are similar to those we have reported
for one-bond 13C-lithium coupling constants in internally
solvated benzylic lithium compounds8d but rather smaller than
those observed for one-bond 13C-7Li or 13C-6Li coupling con-
stants in many different monomeric organolithium compounds,
41 Hz for the former and 15.5 Hz for the latter. It has been
pointed out that one-bond 13C-6Li (or 7Li) coupling constants
should decrease with increasing δ-C‚‚‚∂+Li dipole moments.10a
Further, a careful reading of the pioneering theoretical treatment
(10) (a) Bartlett, R. J.; Del Bene, J. E.; Perera, S. A. B. Structures and
Mechanisms: from Ashes to Enzymes; ACS Symposium Series 827;
American Chemical Society: Washington, DC, 2002; pp 150-164 and
references therein. (b) Ramsay, N. F. Phys. ReV. 1953, 91, 303. (c) Grant,
D. M.; Litchman, W. M. J. Am. Chem. Soc. 1965, 87, 3994. (d) Litchman,
W. M.; Grant, D. M. J. Am. Chem. Soc. 1967, 89, 2228. (e) Karplus, M.;
Grant, D. M. Proc. Nat. Acad. Sci. U.S.A. 1959, 45, 1269.
9
3988 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Conjugation in Allylic Lithium Compounds
A R T I C L E S
Table 4. Internally Solvated Alllylic Lithium Compounds Lithium Coordination Angles (deg), Monomers unless Otherwise Noted
O1LiO2
O1LiC1
O2L1C1
N1LiC1
NLiO1
NLiO2
C1*LiC1
C1*LiO1
C1*LiO2
6a
5b
7b
8
114.2(2)
145.2(2)
130.3(3)
117.3(2)
102.8(5)
119.78(12)
100.5(3)
106.6(11)
104.7(2)
112.4(4)
129.8(2)
123.4(5)
122.85(13)
107.8(3)
131.2(11)
95.59
83.3(2)
85.2(2)
81.1(3)
88.7(2)
89.9(4)
112.26(12)
108.5(3)
79.3(4)
80.2(2)
77.6(3)
87.5(2)
85.1(4)
84.38(9)
84.8(3)
79.0(3)
76.1(17)
81.4(4)
84.1(2)
83.2(4)
86.45(10)
84.5(3)
6a
109.0(5)
C3*LiC1
127.3(2)
110.7(4)
97.6(1)
C3*LiO1
86.6(2)
97.1(4)
91.7(0)
C3*LiO2
100.1(2)
94.4(4)
108.0(4)
112.1(2)
132.5(5)
115.9(12)
149.6(4)
5b
7b
9c
10c
12c
a Dimer; see structure 6. b Polymer; C3* is terminal allyl carbon of nearest neighbor; see Figure 3b. c Monomer; see Figure 3a.
Table 5. Quantitative and (Qualitative) Dynamic Behavior of
Internally Solvated Allylic Lithium Compounds, Diethyl Ether-d10
Solution
1,3 Li sigmatropic
C
−
Li exchange
inversion
shift 250 K
1
compd
∆
Hq (kcal/mol)
∆S (eu)
∆
Hq
∆
Sq
k
1 (s-
)
5
6
7
8
11( 0.5a
(fast)b
12e
-6( 2a
8( 0.5a -10( 3a
(fast)b
8.8 × 106 c,d
1.0
-5( 2e
9( 0.5e -7( 2e
9
(slow)f
(fast)b
11 ( 0.5
(slow)f
6 ( 0.3
(slow)f
10
11
12
14
15
(fast)b
1.4 × 105 c,g
>108
-15 ( 4 15 ( 1
+2 ( 0.5
(slow)e
-27 ( 5
<10-2
a Ref 8b. b Too fast to measure. c Estimated from line broadening, see
text. d SD ) ( 2.6 × 105 s-1
.
e Ref 8c. f Too slow to measure. g SD ) (
6.1 × 104 s-1
.
of one-bond Fermi contact coupling by Ramsey, Karplus, and
Grant10b-d leads to the conclusion that a small contribution of
covalence to scalar coupling cannot be eliminated.
Two dynamic effects are responsible for averaging the 13C-
7Li coupling in our internally solvated allylic lithium com-
pounds. These are bimolecular C-Li exchanges, whose rate
increases with temperature8 and 7Li electric quadrupole-induced
relaxation.11 The latter rate increases with decreasing temper-
ature. In the case of 13C-6Li coupling, only bimolecular C-Li
exchange averages this coupling since the quadrupole moment
Figure 4. 13C NMR (75.45 MHz) line shapes for 11, 7Li-bound 13C (left)
observed at different temperatures, and (right) calculated with a first-order
rate constant. Dissymmetry is due to an impurity resonance below the starred
peak. It was taken into account in the line shape calculation by addition of
a fifth density matrix equation.12
similar activation parameters for C-Li exchange with ∆Hq )
11 to 12 kcal/mol (see Table 5). This may imply that decoor-
dination of lithium to the pendant ligand is rate determining
for C-Li exchange among these compounds. However, with a
different ligand, as in 14, ∆Hq for C-Li exchange is much lower
at 6 ( 0.3 kcal/mol. By contrast, activation parameters for the
other reorganization processes that these compounds undergo
vary greatly among the different species we have investigated.8a-c
In several cases, for example 8, progressive averaging of the
13C-7Li coupling constant is also observed on cooling the
sample below the temperature at which this coupling is clearly
resolved, 260 K. The character of these line shape changes is
6
6
of Li is too small to perturb the 13C NMR of Li bound 13C.
The dynamics of bimolecular C-Li exchange can be deter-
mined if there is a temperature at which 13C-7Li or 13C-6Li
spin coupling is clearly resolVed. If when the sample is warmed
the coupling is progressively averaged, this effect is exclusively
due to bimolecular C-Li exchange.8d,12,13 As far as the 13C
NMR of 7Li-bound carbon is concerned, the monomeric
exchanging system with 13C in natural abundance is simulated
as in eq 3.
7
uniquely diagnostic of the operation of Li nuclear electric
quadrupole-induced relaxation, whose rate increases with de-
13C7Li + 12C7Li* a 13C7Li* + 12C7Li
(3)
creasing temperature.11 Note that depending on the relative rates
7
of bimolecular C-Li exchange and Li nuclear electric quad-
Comparison of the experimental 13C1 NMR spectrum with line
shapes calculated to take account of the above exchange process,
eq 3,12,13 provides the second-order rate constants and the
accompanying activation parameters for C-Li exchange. For
example, in the case of 11 in diethyl ether-d10, comparison of
the 13C line shapes for lithium-bound carbon shown in Figure
4 results in ∆Hq and ∆Sq for bimolecular C-Li exchange of,
respectively, 11 ( 0.5 kcal/mol and -15 ( 4 eu (Table 5). In
fact, all the allylic lithium compounds internally coordinated
to the N(CH2CH2OCH3)2 ligand that we have studied show very
rupole-induced relaxation, there may not be a temperature at
which 13C-7Li coupling is well resolved.11
Compound 6. Carbon-13 NMR of 6 in diethyl ether-d10
solution at room temperature gave a single narrow resonance
for the terminal allyl carbons at 58.55 δ. The two 2-methoxy-
ethyl groups were magnetically equivalent, and 13C-6Li spin
coupling was not observed. Similar spectra were obtained for
solutions of 6 in benzene-d6 and in toluene-d8. At first, these
results appeared to be consistent with a delocalized structure.
(12) Kaplan, J. A.; Fraenkel, G. In NMR of Chemically Exchanging Systems;
Academic Press: New York, 1980; Chapters 5 and 6.
(13) Ref 9, p 10341.
(11) Fraenkel, G.; Subramanian, S.; Chow, A. J. Am. Chem. Soc. 1995, 117,
10336-10344.
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3989

A R T I C L E S
Fraenkel et al.
interconversion (eq 4) would average the 13C-7Li or 13C-6Li
coupling constants. In addition, the fast 1,3 lithium sigmatropic
shift would invert the configuration of 28 and overall average
the shifts between the two arms of each pendant ligand in 6
(see eq 5).
Figure 5. ORTEP of 6 showing 50% thermal ellipsoids. Selected atoms
are labeled. Hydrogen atoms have been omitted for clarity.
Evidently, the 1,3 lithium sigmatropic shift (see eq 5) is still
too fast at 150 K to allow observation of well resolved 13C
resonances for the terminal allyl carbons of 6. However, a value
for δ3-δ1 for 6 could be estimated via ab initio calculations
using the Gaussian 98 programs.14 Starting with its crystal-
lographic parameters, compound 6 was geometry optimized at
the RHF level of theory with basis set 6-31 G*. Frequency
calculations were used to confirm that the optimized geometry
corresponded to the minimum-energy configuration. The latter
geometry and the same basis set and Hamiltonian were used to
calculate 13C shift tensors using GIAO.15 This procedure resulted
in a value of 60 ppm for δ3-δ1 of 6.
At 250 K, the terminal allyl 13C resonance of 6 consists of a
broad line at 58.55 δ with a half width at half-height of 37.5
Hz. We assume that this resonance is the result of dynamic
averaging of a 1:1 half spin uncoupled doublet of separation
60 ppm (4500 Hz at 70 054 G). Then, NMR line shape
analysis12,13 yields k1 at 8.8 × 106 s-1 for the 1,3-sigmatropic
shift of 6 at 250 K (Table 5). Recognizing an uncertainty of
(l0% in this calculated shift difference leads to an uncertainty
of (20% in the latter rate constant.
Compound 10. The organolithium moieties in 6 (dimer) and
10 (monomer) are identical with the exception that the ligand
tether in 6 has one carbon, whereas in 10, the tether is two
carbons long. Despite their different states of aggregation the
NMR spectroscopic behaviors of 6 and 10 are very similar. At
292 K, the 13C NMR of the terminal allyl carbons of 10 in
diethyl ether solution consists of a single sharp peak at 51.62
δ. On cooling the sample, this resonance progressively broad-
ened and disappeared into the baseline at 170 K. On further
cooling, the compound precipitated from the solution. Through-
Figure 6. 13C NMR (75.45 MHz) of 6 over the range 57-61 δ showing
the temperature dependence of the 13C resonance at 58.5 δ due to terminal
allyl carbon; CH3O resonance at 59.4 δ; star marks an impurity.
However, X-ray crystallography showed 6 to be a lithium-
bridged dimer in the solid state with localized C1-Li bonding
(see Figure 5). Determination of the freezing point of a solution
of 6 in benzene showed 6 to be a dimer in benzene solution as
well as in the solid state. At this point, 13C NMR of 6 in diethyl
ether was reinvestigated as a function of temperature. On cooling
the sample below room temperature, the resonance we assigned
to the terminal allyl carbons at 58.55 δ progressively broadened
and disappeared into the baseline at 200 K (see Figure 6). In
all other respects the spectrum remained unchanged throughout
the entire temperature range 300-150 K. Resonances that could
be assigned to the terminal allyl carbons were not detected
between 200 and 150 K. Similar behavior was also observed
using toluene-d8 solutions of 6. These results strongly imply
that dimer 6 is the principal species present in benzene, toluene,
and diethyl ether as well as in the solid state.
To account for the results described, above we would like to
propose that dimer 6 is in rapid equilibrium with a low and
undetectable concentration of monomer 28 (eq 4) and that the
terminal allyl 13C shifts in 6 are averaged by a fast 1,3 lithium
sigmatropic shift that takes place in monomer 28 (see eq 5).
Varying the concentration of 6 between 0.05 and 0.6 M did not
change the appearance of this spectrum. Fast monomer-dimer
(14) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.1; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(15) (a) Wolinski, K.; Hilton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112,
8251. (b) Dodds, J. L.; McWeeny, R.; Sadlej, A. J. Mol. Phys. 1980, 41,
1419. (c) McWeeny, R. Phys. ReV. 1962, 126, 1028. (d) London, F. J.
Phys. Radium, Paris 1937, 8, 397. (e) Ditchfield, R. Mol. Phys. 1974, 27,
789-807.
9
3990 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Conjugation in Allylic Lithium Compounds
A R T I C L E S
Figure 8. 13C NMR (75.45 MHz) of 12 in diethyl ether-d10, between 64.4
and 68.8 δ, 292 K.
(C2-C3), despite the different states of aggregation. Both are
unsubstituted at the allyl termini, and both have the same ligand
tether. However, the pendant ligands are different.
Figure 7. Relationship of lithiums with respect to the allyl carbons and
coordinating ligand atoms in 13. The remainder of the ligands and all
hydrogen atoms have been omitted for clarity.
Compound 8. The solid-state structure of 8 confirms the 13
C
shift assignments for the sample in diethyl ether-d10. Well
resolved spin coupling of 6.3 Hz between 13C1 and 7Li observed
at 240 K is averaged on warming the sample due to bimolecular
C-Li exchange. Averaging of 1J(7Li, 13C) is also observed on
cooling the sample below 240 K. The latter changes are
out the entire temperature range 292-170 K, the two arms of
the pendant ligand remained magnetically equivalent. In addi-
tion, there was no evidence for 13C-7Li spin coupling.
The above results are consistent with a fast 1,3-lithium
sigmatropic shift in the monomer coupled with a fast equilibrium
interconversion between monomer and a low, undetectable
concentration of dimer. This system resembles 6 with the
exception that the main component of 6 is the dimer and only
a small fraction consists of monomer.
Chemical shift calculations for 10 were carried out in a
fashion similar to those of 6. The smaller size of monomeric
10 compared to the dimer 6 allowed use of the B3LYP and
MPW1PW91 levels of theory in addition to RHF, all in
conjunction with 6-311G* and 6-31+G* basis sets.14 These
methods gave very similar optimized geometries, which were
confirmed to be stable energy-minimized structures with the
usual freqeuency calculations. Using the latter geometry in
conjunction with MPW1PW95 and basis set 6-31+G*, 13C
NMR shift tensors were calculated with GIA0,15 giving a value
for δ3-δ1 of 34 ( 3 ppm, similar to corresponding shift
differences for monomeric 5, 7, 8, 11, 14, and 15. Assuming
that the observed terminal allyl 13C half width of 7.5 Hz at 250
K is the result of averging δ3 with δ1, we calculate k1 for this
shift to be 1.4 × 105 s-1. An uncertainty in the above shift
difference of (9% results in an excursion of (20% in the latter
calculated rate constant (see Table 5).
Compound 13. This is the first of three allylic lithium
compounds that incorporate a chiral auxiliary pendant ligand
N[(S)-(2-methoxymethyl)-pyrolidino]-. These are intended to
be used for future study of potentially enantioselective carbanion
chemistry. Compound 13 was the starting material for preparing
14 and 15. Though 13 is insoluble in all ethers tested, it did
crystallize into two lithium-bridged dimers. Each of the (S)-
(+)-2-methoxymethylpyrrolidino ligands is fully coordinated to
lithium. Each lithium bears a η3 relationship to one of the allyl
moieties, as well as an interaction with one terminal carbon of
the other allyl moiety even though the C-C allyl bond lengths
are significantly different. Figure 7 shows the relationship of
the two lithiums with respect to the allyl carbons and ligand
sites. The remaining ligand structures and all the hydrogen atoms
have been omitted for clarity. Interestingly, the two allyl bond
lengths of 13 (weighted averages over two dimers), which are
1.415 (8) Å (C1-C2) and 1.368 (20) Å (C2-C3), are remarkably
similar to those of 10, 1.426 (2) Å (C1-C2) and 1.366 (2) (Å)
7
diagnostic of Li nuclear electric quadrupole-induced relax-
ation.11
As reported previously8b between 270 and 330 K, there is
progressive signal averaging of the two silyl methyl 13C
resonances due to a fast lithium 1,3 sigmatropic shift. NMR
line shape analysis yields ∆Hq ) 18.3 ( 1.5 kcal/mol and ∆Sq
) 7 ( 2 eu for this process. That a single structure has been
observed for this compound implies that the Li sigmatropic shift
is accompanied by rotation around both allylic C-C bonds (see
eq 2).
Preparation 12. In the solid state, the product of metalation
of 20 by butyllithium as determined by X-ray crystallography
is a single, partially localized, internally solvated monomer,
described as structure 12. The C-Si bonds of 12 lie close to
the allyl plane. However, as shown below, in solution the above
metalation product consists of a mixture of isomers. This mixture
will be temporarily referred to as “12”.
Carbon-13 NMR of “12” in diethyl ether-d10 between 170
and 292 K shows a clean equal doublet for each of OCH2,
OCH3, and NCH2C. Also, within this temperature range, the
terminal allyl carbons exhibit resonances at 64.11 and 67.33 δ
of approximately equal intensity, both of which consist of 1:1:
1:1 equally spaced quartets with splittings of 4.6 and 4.1 Hz.
These splittings are due to the effects of 13C-7Li spin coupling
(see Figure 8).
With one exception, the 13C NMR spectrum of “12” did not
change between 170 and 292 K. On cooling the sample below
292 K the quartet structure at 64.11 and 67.33 δ due to 13C1-
7Li coupling progressively averaged. This is most likely due to
7Li electric quadrupole-induced relaxation.11
Above 292 K with increasing temperature, 13C NMR of the
three equal doublets due to OCH2, OCH3, and NCH2C progres-
sively average to single lines at their respective centers.
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3991

A R T I C L E S
Fraenkel et al.
13C NMR results similar to those obtained for “12” in diethyl
ether-d10 were observed using solutions of “12” in toluene-d8
and benzene-d6.
in Table 1 under “major” and “minor” components, respectively,
since the shifts cannot be matched to the specific proposed
structures.
The similarity of the NMR parameters of 15 and 30 to those
of 10-12 and 14 supports partially delocalized structures for
15 and 30. One terminal 13C resonance each of 15 and 30
displays splitting due to 13C-7Li spin coupling of 4.3 Hz at
65.47 δ for the minor component and 7.3 Hz at 56.34 δ for the
major component. The second terminal 13C shifts of 13 and 30
are both at 80.26 δ. The terminal allyl shifts of the major
component are similar to those of 8 at 54.1 and 78.1 δ, implying
that the major components of preparation 15 and 8 have similar
electronic structures.
Measurement of the freezing point depression of a solution
of “12” in benzene established that the average molecular weight
of “12” is identical to that of the crystallographic structure drawn
above as 12, i.e., the material is monomeric. Combining this
result with the NMR data, it seems that “12” assumes the same
structure or distribution of structures in all three media used
for the NMR studies.
On the basis of the understood behavior of the other internally
solvated allylic lithium compounds both previously published8a-c
and described herein, we would like to propose that in solution,
“12” consists of two stereoisomers, 12 and 29, in almost equal
concentrations that interconvert rapidly by means of a 1,3
lithium sigmatropic shift (see eq 6). The rate at temperatures
down to 170 K is so fast as to preclude resolution of the terminal
allyl 13C resonances of 12 and 29. By contrast, throughout the
temperature range 170-292 K, rotation around the C1-C2 and
C2-C3 bonds of 12 and 29, bimolecular C-Li bond exchange,
and transfer of coordinated lithium between faces of the allyl
planes are all slow relative to the NMR time scale.
Compound 11. This compound did not crystallize. However,
its NMR parameters are so similar to those of compounds whose
structures were obtained via X-ray crystallography that one can
assume that the structure of 11 resembles those of 8, 10, and
12.
NMR of the ligand carbons was used to investigate the
dynamics of face to face lithium transfer giving ∆Hq ) 15 (
1 kcal/mol and ∆Sq ) +2 ( 0.5 eu, which is considerably
slower than for compounds with a one-carbon tether for which
∆Hq is ca. 7 ( 0.5 kcal/mol. On the other hand, the dynamics
of bimolecular C-Li exchange obtained from averaging of the
13C-7Li spin coupling constant gives ∆Hq ) 11 ( 0.5 kcal/
mol and ∆Sq ) -15 ( 5 eu (see Table 5).
Compound 14. This compound failed to crystallize in the
pure state, but it did provide clean, well-resolved NMR spectra,
which indicate a single molecular species. Impurities such as
starting materials or oxidation products could not be detected
among these NMR spectra. Clearly, the pendant chiral auxiliary
is responsible for over 97% enantioselectivity in the formation
of 14.
Compound 14 exhibits one-bond 13C-7Li coupling of 8.7 Hz
and terminal allyl shifts of 37.22 δ (C1) and 74.24 δ (C3). These
values are so similar to corresponding values for 5, 7, and 11
that one can assume that the silyl allyl moieties in all four
species have similar electronic structures. The single silyl
substituent at one allyl terminus appears to be responsible for
the similar structural and NMR parameters among all four
compounds. Meanwhile, ligand tethers and pendant ligands have
far less influence on these parameters.
Averaging of the 13C1-7Li coupling with increasing temper-
ature above 215 K yielded ∆Hq and ∆Sq for bimolecular C-Li
exchange of 6 ( 0.3 kcal/mol and -27 ( 5 eu, respectively
(see Table 5).
First, note that preparation “12” has 13C in natural abundance.
For convenience, the terminal allyl carbons are labeled as a
and b. Consider the two equilibrating isotopomers 29-a-13C with
12-a-13C and 29-b-13C with 12-b-13C (see eq 6). Thus, of the
two 7Li bound 13C resonances at 64.11 and 67.33 δ, one is the
average shift of 13Ca in 29-a-13C with 13Ca in 12-a-13C. The ca.
4 Hz splitting of this resonance is the average of the one-bond
13Ca-7Li coupling in 29-a-13C, ca. 8 Hz, with the very small
or zero long-range 13Ca-7Li coupling in 12-a-13C. Similar
considerations apply to the averaged 13C resonance of 13Cb in
29-b-13C with that of 13C in 12-b-13C. These two averaged 13
C
resonances at 64.11 and 67.33 δ remain different because C-Li
bimolecular exchange and rotation around the C1-C2 and C1-
C3 bonds in 12 and 29 are both slow relative to the NMR time
scale.
The magnitudes of the averaged terminal allyl 13C shifts and
of the splittings due to 13C-7Li spin coupling in 12 and 29 are
consistent with an equilibrium between two stereoisomers with
closely similar concentrations. Thus, compounds 8, 12, and 29
have similar structures but with different ligand tethers. The
average 13C terminal allyl shift in 8 is 66.18 δ. This compares
favorably with the corresponding proposed averaged values of
61.11 and 67.33 δ for the equilibrium between 12 and 29.
Preparation 15. This material did not crystallize, but it did
give rise to well-defined solution 13C NMR spectra. The NMR
data are consistent with an equilibrium between internally
solvated monomeric allylic lithium compounds, proposed as 15
and 30, in ratio of 2/1 (see eq 7), which interconvert by means
of a 1,3 lithium sigmatropic shift. This process is slow relative
to the NMR time scale up to 300 K. NMR parameters obtained
from the solution of 15 with 30 are listed around the structure
Compound 9. Of all the compounds studied in this work, 9
is the most localized, with allyl bond lengths of 1.494 (7) Å
(C1-C2) and 1.306 (8) Å (C2-C3) and 13C shifts for C1, C2,
and C3 of 36.84, 171.2 and 118.8 δ, respectively. These
parameters closely resemble those for a simple acyclic alkene.
With two silyl substituents at C1 in 9 that are well separated
by 135° on opposite sides of the allyl plane, the torsional angle
Li-C1-C2-C4 is just 30.03°. This arrangement minimizes
9
3992 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Conjugation in Allylic Lithium Compounds
A R T I C L E S
potential steric interactions between the ligand and the TMS
groups at C1. It also places the C-Li bond closer to the allyl
plane than in any of the other structures reported in this article.
In this structure, the TMS groups at C1 lie in very similar
environments. In fact, just one sharp 13C NMR peak is observed
for each of CH3O, CH2O, and NCH2C at temperatures down
to 150 K.
A second consequence of the proximity of the C-Li bond
to the allyl plane is that 1J(13C, 7Li) is 15.7 Hz, that is, ca. twice
that found for all other compounds described herein. Due to
the small Li-C1-C2-C4 angle in 9, the C2s contribution to the
C1-Li bond orbital would be larger than in the other internally
solvated allylic lithium compounds we have studied.
General Conclusions. Comparison of the X-ray crystal-
lographic and solution data show that the structures of our
internally solvated allylic lithium compounds in the solid state
and in solution must be very similar. In every case, the pendant
ligand is fully coordinated to lithium and no solvent molecules
are incorporated in the crystal structures. Most of these species
are monomers in solution. Compound 6 is a dimer in both the
solid state and in solution; 13 is a dimer in the solid state and
insoluble in diethyl ether and in THF.
and 13C-7Li coupling constants, respectively, cannot be elimi-
nated. It is still uncertain. Although major advances have been
made in the theory of scalar NMR coupling and in calculation
of coupling constants, application of these methods to calculating
7
13C-6Li (or Li) coupling constants has not been successful.
From the data in Table 2, it is evident that the degree of
delocalization among 5-15 must be continuously variable. For
example inspection of the data in Table 2 reveals in a qualitative
way that as the torsional angle Li-C1-C2-C4 decreases and
the C-Li bond approaches the allyl plane, the difference
between the allylic C1-C2 and C2-C3 bond lengths increases,
as does the allylic 13C chemical shift difference δ3-δ1. The
limiting localized species is 9 with a Li-C1-C2-C4 angle of
-30.3°. Its allylic 13C shifts and bond distances closely resemble
those of a simple alkene. Compound 9 also exhibits the largest
13C-7Li coupling constant reported herein, 15.9 Hz. This is most
likely due to the increased “S” character associated with the
C-Li bond in 9 when compared to those in the other
compounds, 5, 7, 8, 11, 14, and 15.
Averaging of the 13C-7Li or 13C-6Li coupling constants with
increasing temperature is necessarily due to bimolecular C-Li
exchange. This requires that the mechanism for the exchange
process proceeds via a dimeric transition state that is most likely
preceded by a dimeric intermediate. In fact, two of the
compounds in Table 1, 6 and 13, are dimers. Thus, it is
reasonable to assume that dimeric intermediates are energetically
accessible to our internally solvated allylic lithium compounds,
which are monomeric in their ground states.
Lithium is localized near one of the allylic termini, and most
of the compounds exhibit one-bond 13C-7Li or 13C-6Li spin
coupling at low temperature. The existence of such coupling is
consistent with the observed proximity of lithium to one of the
allyl termini carbons.
Carbon-13 chemical shifts and bond lengths of the allylic
moieties of the internally solvated allylic lithium compounds
lie between those of the model species, the delocalized externally
solvated ion-paired 1 and the unsolvated localized 2.5 Thus, the
combined NMR and X-ray crystallographic data strongly support
the proposition that these compounds are partially localized.
Because the ligand tether is too short to place lithium normal
to the center of the allyl plane, as in 1, lithium is sited off the
axis normal to the allyl plane at one of the carbon termini. This
restricts the stereochemistry of coordination of lithium to the
aberrant arrangements observed in the crystal structures and
thereby stabilizes a partly delocalized allylic lithium compound,
a new molecular species. It is tempting to consider that the
proximity of Li+ to a terminal allyl carbon is responsible for
polarizing what would ordinarily be a delocalized anion into a
partially localized one.
Considering the wide variety of structures we have investi-
gated, the activation enthalpies we have determined for bi-
molecular C-Li exchange are remarkably similar. They all fall
within the range 11-12 kcal/mol. A species as different as the
internally solvated benzylic lithium compound, 31, also exhib-
ited a ∆Hq for C-Li exchange of 11 kcal/mol for bimolecular
C-Li exchange.8d These results suggest that the rate-determining
step for bimolecular C-Li exchange involves dechelation of
lithium from the pendant ligand.
Prior to our report on 5,8a,b spin coupling had not been
6
7
observed between 13C and Li or Li in any allylic lithium
compound. The values reported herein for one-bond 13C-7Li
spin coupling are much smaller than the common value of ca.
42 Hz observed for a wide variety of monomeric organolithium
compounds (RLi, where R ) alkyl, vinyl, aryl, alkynyl, and
heterocyclic).16a The smaller and variable values of 1J(13C, 7Li)
for compounds 5-15 compared to the “common pattern” value
of 42 Hz16 may be due to the increased ionic character associated
with the C-Li bonds in 5-15 compared to those for C-Li
bonds in the above “common pattern” compounds, as has been
recently proposed in a discussion on 13C-Li spin coupling.10a
While the carbon lithium bonds in the species described in
this article are most likely mainly ionic, the role of a small
degree of C-Li covalency to these bonds and to the 13C-6Li
Dynamics of configurational inversion are obtained from
NMR line shape analysis of the pendant ligand resonance. It is
a first-order process and results overall from transfer of
coordinated lithium between faces of the allyl plane. The rates
depend on the ligand tether with ca. ∆Hq of 6.5 ( 0.5 kcal/mol
for a one-carbon tether and 15 ( 1 kcal/mol for a two-carbon
tether, in the case of compound 11. Configurational inversion
also accompanies fast lithium sigmatropic shifts in symmetrical
internally solvated allylic lithium compounds.
By contrast to the fairly uniform activation parameters for
inversion and bimolecular C-Li exchange, 1,3 lithium sigma-
tropic shifts take place at remarkably different rates. The
dynamics of this sigmatropic shift can only be investigated in
the cases of compounds 6, 8, 10, 12, and 15. These are
(16) (a) Bauer, W.; Winchester, W. R.; Schleyer, P. v. R. Organometallics 1997,
6, 2371-2379. (b) Bauer, W.; Griesinger, C. J. Am. Chem. Soc. 1993, 115,
10871. (c) Lambert, C.; Schleyer, P. v. R. Angew. Chem., Int. Ed. Engl.
1994, 33, 1129.
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3993

A R T I C L E S
Fraenkel et al.
hydrogen atom positions were taken from the difference electron density
map and refined freely. Other hydrogen atoms were positioned at ideal
geometries, and a riding model was employed in the refinement. The
denoted R values are defined as follows:
symmetrical species that are either unsubstituted at the allyl
termini or possess one TMS group at each terminal allyl carbon.
As seen in Table 5 at 250 K, relative to the NMR time scale,
the 1,3 lithium sigmatropic shift for 12 a 29 is too fast to
measure, while that of 15 a 30 is too slow to measure. Between
the latter two extremes are 8 with a k1 of 1 s-1 and 6 and 10
whose lithium sigmatropic shifts are estimated to be faster than
that of 8 by factors of ca. 107 and 105, respectively.
Modeling of these latter compounds indicates that differences
among steric interactions that accompany the Li sigmatropic
shifts may be responsible for the wide variations among the
rates. Thus, the unsubstituted compounds 6 and 10 offer
minimum constraints to rearrangement. However, the shift in 8
would be hindered by the proximity of the tethered ligand to
the silyl substituents. The longer tether in 12 reduces the latter
steric constraints. By contrast to the 29 a 12 interconversion
(eq 6), the steric interactions that accompany 15 a 30 would
be far more severe due to the proximity of the -Li-N-CH-
CH2O- moiety to the allyl surface.
In summary, it has been established that, among several
internally solvated allylic lithium compounds, the ligand tether
is too short to place coordinated lithium normal to the center
of the allyl plane as in 1. Instead, coordinated lithium is placed
closer to one of the terminal allyl carbons. This polarizes what
would otherwise be a delocalized allyl anion to a partly localized
species. Dynamic behavior associated with inversion, C-Li
exchange, and 1,3 lithium sigmatropic shifts has also been
uncovered.
R1 )
||Fo| - |Fc||
|Fo|
∑
∑
/
2 2
2 2
wR2 ) { w(Fo2 - Fc ) / w(Fo ) }1/2
;
∑
∑
w ) 1/{σ2(Fo ) + (g1*P)2 + g2*P}; P ) (Fo +2Fc )/3
2
2
2
Further details on structure investigation can be obtained from the Dir-
ector of the Cambridge Crystallographic Data Centre, 12 Union Road,
GB-Cambridge C132 1 EZ, UK, by quoting the full journal reference.
Allylic Lithium Compounds, General Procedure for NMR
Samples. Butlyllithium in hexane (3.13 mL, 1.6 M, 5 mmol) was added
slowly by syringe at 0 °C to a solution of the appropriate substituted
propene (5 mmol) in 10 mL of freshly distilled dry, oxygen-free diethyl
ether. After this addition was complete, the mixture was allowed to
warm to room temperature and then stirred for 1 h. An aliquot (1 mL)
of this solution was transferred into an NMR tube with attached joint
and straight bore stopcock. All volatile material was removed in vacuo
and then replaced by deuterated solvent, diethyl ether-d10 or THF-d8.
The latter was distilled from sodium over the vacuum line into the
NMR tube. The sample tube was degassed via three freeze-thaw cycles
and then sealed off frozen with pumping.
Silylation of Allylic Lithium Compounds: General Procedure.
A freshly prepared sample of an allylic lithium compound (5 mmol)
was cooled to -78 °C, and chlorotrimethylsilane (0.54 g, 0.64 mL, 5
mmol) was added slowly by syringe. Upon completion of the addition,
the mixture was allowed to warm to room temperature. The onset of
reaction at -40 °C was indicated by precipitation of LiCl. The mixture
was stirred for an additional 1 h. Saturated aqueous NaHCO3 (10 mL)
was added to remove excess chlorotrimethylsilane. The aqueous phase
was extracted twice with 10 mL diethyl ether. The combined nonaque-
ous phases were dried over Na2SO4 the solvent was removed, and the
residue was distilled.
Experimental Section
General manipulations of organometallic compounds were carried
out under an atmosphere of argon or nitrogen. Solvents were dried
over Na/K alloy. NMR spectra were obtained using Bruker AM250 or
Bruker Avance 300 equipment.
Preparation of Allylic Lithium Samples for Crystallography.
Typical Procedure: 2-Bis(2-methoxyethyl)aminomethyl-1-trimeth-
ylsilylallyllithium (5). Under an argon atmosphere at 0 °C, butyllithium
(1.25 mL, 1.6 M, 2 mmol) in hexane was slowly added to a solution
(0 °C) of 2-bis(2-methoxyethyl)aminomethyl-3-trimethylsilyl-1-propene
(0.52 g, 2 mmol) in 10 mL of hexane. The mixture was allowed to
warm to room temperature and then stirred for 1 h. The volume of the
reaction mixture was reduced to 5 mL under vacuum. Crystals
developed after standing the sample overnight at room temperature.
Toluene has also been used in these crystallization procedures.
2-Bis(2-methoxyethyl)aminomethylallyllithium, Crystals 5. Me-
thyllithium (1.9 mL, 1.4 M, 2.2 mmol) in ether was added to 2-bis(2-
methoxyethyl)aminomethylpropene (0.5 g, 2,7 mmol) in 5 mL of THF
at 0 °C. After addition of the CH3Li, the solution was allowed to warm
to room temperature slowly and then stirred for 15 min. The volume
was reduced to 3 mL. Colorless crystals developed after storing the
sample at room temperature overnight.
X-ray Crystallography: General Procedures. A sample of the
crystalline material was removed from the mother liquor using standard
Schlenk techniques and coated with Paratone N, an inert oil. Crystals
were handled under oil under a microscope. Suitable crystals were then
mounted at the end of a glass fiber in a drop of Paratone N and then
shock cooled on the diffractometer. All data were collected at 173 K
on an Enraf-Nonius diffractometer with a CCD area detector using
graphite monochromatic Mo KR radiation (λ ) 0.71073 Å). Structures
were solved by direct methods with SHELXS-9717 and were refined
by full matrix least squares procedures on F2 using SHELXL-97.18 All
non-hydrogen atoms were refined anisotopically. Allylic and vinylic
Allylic Lithium Samples for NMR Study. At 0 °C under an
atmosphere of argon, n-butyllithium (3.13 mL, 1.6 M, 5 mmol) in
hexane was slowly added to a solution of the appropriate substituted
alkene (5 mmol) in 10 mL of freshly distilled dry, oxygen-free diethyl
ether. Upon complete addition, the reaction mixture was allowed to
warm to room temperature and then stirred for 1 h. A 1 mL aliquot of
this solution was syringed into a 5 mm OD NMR tube equipped with
an attached joint and a 2 mm straight bore stopcock. All volatile material
was removed in vacuo and replaced by diethyl ether-d10 or THF-d10.
The latter solvent was distilled from sodium over the vacuum line into
the NMR tube. The loaded sample was degassed via three freeze-
thaw cycles and then sealed off frozen with pumping.
Freezing Point Depression of 6 in Benzene. A solution of
butyllithium in hexane (1.25 mL, 1.6 M, 2 mmol) at 0 °C under argon
was slowly added to 16 (0.37 g, 2 mmol) in 5 mL of diethyl ether. The
mixture was allowed to warm to room temperature and then stirred for
1 h. All volatile components of this reaction mixture were removed in
vacuo (10-3 Torr), and the residue was dissolved in 12.2 g of dry,
oxygen-free benzene. The resulting solution was syringed into a dry,
oxygen- and moisture-free 50 mL insulated flask equipped with a
Beckman thermometer and mechanical stirrer. The above-described
benzene solution showed a freezing point depression of 0.36 ( 0.03
°C compared to pure benzene. This corresponds to a state of aggregation
of 2.17 compared to benzene. Assuming all of the starting material is
converted to 6 and using the freezing point depression of 49 °C for 1
mol of solute in 100 g of benzene, the above freezing point depression
corresponds to a state of aggregation of 2.23.
2-(Bis(2-methoxyethyl)aminomethyl)-1,3,3-tris(trimethylsilyl)-1-
propene, (17). At 0 °C and under an atmosphere of argon, n-
butyllithium (2.3 mL, 2.5 M, 57 mmol) in hexane was slowly added to
16 (3.63 g, 19 mmol) in 25 mL of diethyl ether. After stirring for 30
(17) Sheldrick, G. M. Acta Crystallogr., Sect A 1990, 481, 467.
(18) Sheldrick, G. M. SHELXL-97, Program for Crystal Structure Refinement;
University of Go¨ttingen: Go¨ttingen, 1997.
9
3994 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Conjugation in Allylic Lithium Compounds
A R T I C L E S
min, the mixture was cooled to -78 °C and chlorotrimethylsilane (6.2
g, 7.5 mL, 57 mmol) was added dropwise. After completion of the
addition, the mixture was allowed to warm to room temperature and
then stirred for 1 h. The mixture was treated with 30 mL of saturated
Na2CO3 (aq); the phases were separated, and the aqueous phase was
extracted with diethyl ether (3 × 20 mL). The combined organic layers
were washed with NaCl (aq) (2 × 20 mL) and dried over Na2CO4.
Removal of solvent and distillation of the residue gave 6.15 g of the
Treatment of the above aqueous solution with KOH (4 g) followed
by extraction into ether recovered 4.8 g of the unreacted amine.
4-N-((S)-(+)-(2-Methoxymethyl)pyrrolidino-2,3,3-trimethyl-1-
butene (23). A solution of amide 22 (9.4 g, 42 mmol) in 20 mL of dry
THF was slowly dropped into a suspension of LiAlH4 (1.3 g, 45 mmol)
in 100 mL of THF. After refluxing this mixture overnight, 15 mL of
15% NaOH (aq) was added, which produced a white precipitate.
Separation of the liquid phase followed by removal of solvent yielded
8.7 g of the title amine in 90% yield. 1H NMR (CDCl3, 250 MHz, δ):
0.985 (s, 3), 1.042 (s, 3), 1.52 (m, 1), 1.66 (m, 2), 1.726 (s, 3), 1.78
(m, 1), 2.180 (m, 1), 2.389 (d, 1, 2J ) 13.1 Hz), 2.57 (m, 1), 2.618 (d,
1, 2J ) 13.1 Hz), 3.150 (m, 2), 3.303 (s, 3), 3.340 (m, 1), 4.706 (m, 2).
13C NMR (CDCl3, 64 MHz, δ): 19.85, 23.78, 25.91, 26.31, 28.35,
40.48, 56.86, 58.90, 65.23, 66.31,77.00, 109.68, 152.45.
1
title compound in 80% yield, bp 99-100 °C at 0.05 Torr. H NMR
(CDCl3, 250 MHz, δ): 0.034 (s, 18), 0.048 (s, 9), 1.728 (s, 1), 2.584
(t, 4, 3J ) 6.2 Hz), 2.988 (s, 2), 3.270 (s, 6), 3.403 (t, 4, 3J ) 6.2 Hz),
4.927 (s, 1). 13C NMR (CDCl3, 75 MHz, δ): 0.37, 0.53, 25.40, 52.95,
58.07, 62.86, 70.97, 124.10, 157.52.
4-Bis(2-methoxyethyl)amino-3,3-dimethyl-2-trimethylsilylmethyl-
1-butene, (19). At 0 °C under an atmosphere of argon, n-butyllithium
(8 mL, 2.5 M, 20 mmol) in hexane was added dropwise to 18 (4.59 g,
20 mmol) in 20 mL of diethyl ether. The resulting reaction mixture
was treated with trimethylchlorosilane (2.54 mL, 20 mmol) at -78
°C. The usual workup afforded 3.48 g of the title compound in 58%
yield, bp 92-97 °C at 0.5 Torr. 1H NMR (CDCl3, 250 MHz, δ): -0.014
(s, 9), 0.935 (s, 6), 1.451 (s, 2), 2.377 (s, 2), 2.644 (t, J ) 6.5 Hz),
3.371 (t, J ) 6.5 Hz), 4.579 (s, 1), 4.730 (s, 1). 13C NMR (CDCl3, 75
MHz, δ): -0.49, 21.35, 25.23, 41.39, 55.81, 58.68, 65.61, 71.60, 108.19,
152.91.
4-N-((S)-(+)-(2-methoxymethyl)pyrrolidino)-3,3-dimethyl-2-tri-
methylsilylmethyl-1-butene (24). At 0 °C under an argon atmosphere,
n-butyllithium in hexane (12 mL, 2.5 M, 30 mmol) was added dropwise
to compound 23 (6.34 g, 30 mmol) in 26 mL of diethyl ether. The
reaction mixture was warmed to room temperature, stirred for 3 h, and
then reacted at -78 °C with trimethylchlorosilane (3.26 g, 3.8 mL, 30
mmol). Workup yielded 6 g of the title compound in 71% yield, bp 91
°C at 0.2 Torr. 1H NMR (CDCl3, 250 MHz, δ): 0.019 (s, 9), 0.952 (s,
2
3), 1.023 (s, 3), 1.55-1.90 (m, 4), 2.2 (m, 1), 2.386 (d, 1, J ) 13.1
Hz), 2.57 (m, 1), 2.597 (d, 1, 2J ) 13.1 Hz), 3.15 (m, 1), 3.313 (s, 3),
2
3.37 (m, 1), 4.769 (d, 1, J ) 1.2 Hz), 4.595 (d, 1, J ) 1.2 Hz). 13C
4-N-Bis(2-methoxyethyl)amino-3,3-dimethyl-1-trimethylsilyl-2-
trimethylsilylmethyl-1-butene (20). At 0 °C under an atmosphere of
argon, n-butyllithium (3.4 mL, 2.5 M, 85 mmol) in hexane was added
dropwise to 18 (6.5 g, 28.3 mmol) in 50 mL of diethyl ether. The
reaction was allowed to warm to room temperature with stirring over
1 h. After the reaction mixture was cooled to -78 °C, trimethylchlo-
rosilane (9.23 g, 10.8 mL, 85 mmol) was slowly added. Reaction was
observed on warming the mixture to -40 °C. The workup described
above gave 8.25 g of the title compound in 78% yield, bp 120 °C at
NMR (CDCl3, 64 MHz, δ): -0.419, 21.37, 23.83, 25.78, 26.05, 28.45,
41.29, 57.14, 58.95, 65.36, 66.40, 77.09, 107.90, 153.86.
4-N-(S)-(+)-(2-Methoxymethyl)pyrolidino)-3,3-dimethyl-1-tri-
methylsilyl-2-trimethyl-silylmethyl-1-butene (25). At 0 °C under an
argon atmosphere, n-butyllithium in hexane (10 mL, 16 mmol, 1.6 M)
was added with stirring to 24 (4.56 g, 16 mmol) in 10 mL of diethyl
ether. The reaction mixture was warmed to room temperature and stirred
for an additional 3 h, and then it was reacted at -78 °C with
trimethylchlorosilane (1.47 g, 2 mL, 16 mmol). Workup yielded 5.01
1
0.2 Torr. H NMR (CDCl3, 250 MHz, δ): -0.044 (s, 9), 0.002 (s, 9),
1
g of the title compound. H NMR (CDCl3, 250 MHz, δ): 0.065 (s,
0.867 (s, 6), 1.701 (s, 2), 2.281 (s, 2), 2.551 (t, 4, J ) 6.7 Hz), 3.176
(s, 6), 3.268 (t, 4, J ) 6.7 Hz), 5.031. 13C NMR (CDCl3, 75 MHz, δ):
0.59, 24.21, 26.29, 43.00, 55.88, 58.67, 67.90, 71.66, 119.47, 163.50.
Freezing Point Depression of 2-(2-Methyl-3-bis(2-methoxyeth-
ylethyl)amino)-2-propyl)-1,3-bis(trimethylsilyl)allyllithium (12). At
0 °C under an atmosphere of argon, n-butyllithium (1.56 mL, 1.6 M,
2.5 mmol) in hexane was added dropwise to 20 (0.93 g, 2.5 mmol) in
5 mL of hexane. The reaction mixture was brought to room temperature
and stirred for 30 min. All volatile material was removed in vacuo
(0.2 Torr), and the residue was dissolved in 8.6 g of benzene. The
resulting solution was syringed into a flame-dried, argon-filled 50 mL
insulated flask equipped with a Beckman thermometer and a mechanical
stirrer. During the experiment, a steady flow of argon was maintained
to ensure an inert atmosphere. This sample showed a freezing point
depression of 1.4 °C versus pure benzene, indicating a state of
aggregation of 1.02.
18), 0.980 (s, 3), 1.028 (s, 3), 1.4-1.7 (b, 4), 1.833 (s, 2), 2.189 (q, 1),
2
2
2.393 (d, 1, J ) 13.3 Hz), 2.53 (m, 1), 2.565 (d, 1, J ) 13.3 Hz),
3.104 (m, 2), 3.314 (s, 3), 3.358 (q, 1), 5.151 (s, 1). 13C NMR (CDCl3,
63 MHz, δ): 0.59, 0.65, 23.90, 24.44, 26.64, 27.52, 28.56, 42.71, 57.33,
58.96, 65.28, 68.89, 77.12, 119.10, 164.14.
Acknowledgment. This research was generously supported
by the National Science Foundation Grant No. CHE 0077605
and by the Newman Chair. R.F. thanks the Deutscher Akademis-
cher Austausch Dienst for Fellowship support. We thank Dr.
Charles Cottrell, Central Campus Instrumentation Center, for
untiring technical assistance. We thank Dr. Judith Gallucci for
clarifying some issues related to the X-ray crystallographic data.
Note Added after ASAP. In the version posted 3/4/04, ref
7 was incorrect. The version posted 3/9/04 and the print version
are correct.
N-((S)-(+)-(2-Methoxymethyl)pyrollidino)-2,2,3-trimethyl-but-3-
enoic Amide (22). To carboxylic acid 21 (6.41 g, 50 mmol) was added
dropwise thionyl chloride (7.16 g, 60 mmol). The mixture was heated
to 85 °C for 1 h. Excess thionyl chloride was removed by distillation
and the remaining mixture was chilled to 0 °C. Free amine (S)-(+)-
2-methoxy-methylpyrollidine (11.4 g, 100 mmol) was added, and the
mixture was stirred for 3 h at room temperature. Then, 15 mL of water
was added, and the layers were separated. The aqueous layer was
extracted with diethyl ether (3 × 15 mL), and the combined nonaqueous
extracts were washed with NaHCO3 (aq) and then dried over MgSO4.
Removal of solvent gave 9.4 g of the title compound in 83% yield.
The product was 99% pure by chromatographic analysis. 1H NMR
(CDCl3, 250 MHz, δ): 1.236 (s, 3), 1.241 (s, 3), 1.645 (s, 3), 1.85
(bm, 4), 3.23 (m, 1), 3.249 (s, 3), 3.35 (m, 1), 3.47 (m, 2), 4.788 (bd,
2). 13C NMR (CDCl3, 63 MHz, δ): 19.70, 25.02, 25.18, 25.30, 26.56,
47.41, 48.43, 58.74, 72.30, 77.00, 109.70, 148.76, 174.02.
Note Added after Print Publication. Due to a production
error, the end of the abstract was formatted as the first four
paragraphs of the main text in the version posted on the Web
March 4, 2004 (ASAP) and published in the March 31, 2004
issue (Vol. 126, No. 12, pp 3983-3995). The correctly
formatted electronic version of the paper was published on
October 26, 2004, and an Addition and Correction appears in
the November 17, 2004 issue (Vol. 126, No. 45).
Supporting Information Available: Crystallographic data and
NMR spectra. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA030370Y
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3995