Stereochemistry of solvation of benzylic lithium compounds: Structure and dynamic behavior

Article abstract of DOI:10.1021/ja990485v

Several sec-benzylic lithium compounds, both externally coordinated, [α-(trimethylsilyl)benzyl]-lithium·PMDTA (12) and p-tert-butyl-α-(dimethylethylsilyl)benzyllithium·TMEDA (13), and internally coordinated, [α-[[[cis-2,5-bis(methoxymethyl)-1-pyrrolidinyl]methyl]dimethylsilyl]-p- tert-butylbenzyl]lithium (14) and [α-[[[(S)-2-(methoxymethyl)-1-pyrrolidinyl]methyl]dimethylsilyl]benzyl] lithium (15), have been prepared. Ring ¹³C NMR shifts indicate that 12-15 have partially delocalized structures. Externally solvated allylic lithium compounds are found to be delocalized, and only some internally coordinated species are partially delocalized. Compound 15 exists as > 95% of one stereoisomer of the two invertomers at C_α. This is in accord with a published ee of > 98% in products of the reactions of 15 with aldehydes. All four compounds show evidence of one-bond ¹³C-⁶Li spin coupling, ca. 3 Hz, which indicates a small detectable C-Li covalence. Averaging of the ¹³C-⁶Li coupling of 12 with increasing temperature provides the dynamics of intermolecular C-Li bond exchange, with ΔH?_ex = 9 ± 0.5 kcal mol^-1. Carbon-13 NMR line shape changes due to geminal methyls, and ligand carbons gave similar rates of inversion at C_α in 13 (externally solvated) and 14 (internally solvated), ΔH?_inv ≈ 4.9 ± 0.5 kcal mol^-1. By contrast, barriers to rotation around the ring-C_α bonds vary widely, depending on the mode of lithium coordination, ΔH?_rot ≈ 8 ± 0.5 to 19 ± 1.0 kcal mol^-1. Some mechanisms for these processes are proposed.

Full text of DOI:10.1021/ja990485v

10538
J. Am. Chem. Soc. 1999, 121, 10538-10544
Stereochemistry of Solvation of Benzylic Lithium Compounds:
Structure and Dynamic Behavior
Gideon Fraenkel,* Joseph H. Duncan, Kevin Martin, and Jinhai Wang
Contribution from the Department of Chemistry, The Ohio State UniVersity, Columbus, Ohio 43210
ReceiVed February 16, 1999
Abstract: Several sec-benzylic lithium compounds, both externally coordinated, [R-(trimethylsilyl)benzyl]-
lithium‚PMDTA (12) and p-tert-butyl-R-(dimethylethylsilyl)benzyllithium‚TMEDA (13), and internally
coordinated, [R-[[[cis-2,5-bis(methoxymethyl)-1-pyrrolidinyl]methyl]dimethylsilyl]-p-tert-butylbenzyl]lithium
(
14) and [R-[[[(S)-2-(methoxymethyl)-1-pyrrolidinyl]methyl]dimethylsilyl]benzyl]lithium (15), have been
13
prepared. Ring C NMR shifts indicate that 12-15 have partially delocalized structures. Externally solvated
allylic lithium compounds are found to be delocalized, and only some internally coordinated species are partially
delocalized. Compound 15 exists as >95% of one stereoisomer of the two invertomers at CR. This is in accord
with a published ee of >98% in products of the reactions of 15 with aldehydes. All four compounds show
evidence of one-bond C- Li spin coupling, ca. 3 Hz, which indicates a small detectable C-Li covalence.
Averaging of the 13C- Li coupling of 12 with increasing temperature provides the dynamics of intermolecular
13
6
6
q
ex
-1
C-Li bond exchange, with ∆H ) 9 ( 0.5 kcal mol . Carbon-13 NMR line shape changes due to geminal
methyls, and ligand carbons gave similar rates of inversion at CR in 13 (externally solvated) and 14 (internally
q
inv
-1
solvated), ∆H ≈ 4.9 ( 0.5 kcal mol . By contrast, barriers to rotation around the ring-CR bonds vary
q
rot
-1
widely, depending on the mode of lithium coordination, ∆H ≈ 8 ( 0.5 to 19 ( 1.0 kcal mol . Some
mechanisms for these processes are proposed.
1
1
compounds show 13C NMR shifts that are to be expected of
ion-pairs containing delocalized allylic anions, see 1-1. By
Main group allylic and benzylic organometallic compounds
4
have always been regarded as among the simplest delocalized
carbanionic species, a concept which is largely qualitatively
supported by the results of X-ray crystallographic, spectro-
scopic,4 and calculational studies. Solvated allylic lithium
2,3
,5
6,7
(1) (a)Wardell, J. L. In ComprehensiVe Organometallic Chemistry;
Wilkinson, G., Stone, F. G. H., Abel, E. W., Eds.; Pergammon Press:
Oxford, U.K. 1982; Vol. 7, pp 83-91, 97-106. (b) Seyferth, D.; Julia, T.
F. J. Organomet. Chem. 1967, 8, C13. (c) Schlosser, M.; Bosshardt, H.;
Walde, A.; St a¨ hle, N. Angew. Chem. 1980, 92, 302. (d) Brownstein, S.;
Bywater, S.; Worsfold, D. J. Organomet. Chem. 1980, 199, 1. (e) Hsieh,
H. L. J. Polym. Sci. A 1965, 3, 153-161; 163-172; 173-180.
contrast, unsolvated alkane-soluble allylic lithium compounds,
13
most likely aggregated as, for example, 1-2, show C NMR
8
shifts similar to those of alkenes. Therefore, they should be
(
2) (a) K o¨ ster, M.; Weiss, E. Chem. Ber. 1982, 115, 3422. (b) Sch u¨ mann,
regarded as largely localized compounds. Species with structures
between 1-1 and 1-2, i.e., partially delocalized, 1-3, with some
detectable C,Li covalency have not been recognized until
recently, and then only for compounds with attached pendant
U.; Weiss, E.; Dietrich, H.; Mahdi, W. J. Organomet. Chem. 1987, 322,
2
1
99. (c) Sebastian, J. F.; Grumwell, J. R.; Hsu, B. J. Organomet. Chem.
974, 78, C1. (d) Boche, G.; Etzrodt, H.; Marsch, M.; Massa, W.; Baum,
G.; Dietrich, H.; Mahdi, W. Angew. Chem. 1986, 98, 84. (e) Boche, G.;
Fraenkel, G.; Cabral, J.; Harms, K.; Van Eikema-Hommes, N. J. R.; Lorenz,
J.; Marsch, M.; Schleyer, P. v. R. J. Am. Chem. Soc. 1992, 114, 1562-
9
ligands, such as 1-4. For such cases, we proposed that restricted
1
565.
(
3) (a) Patterman, S. P.; Karle, I. L.; Stucky, G. D. J. Am. Chem. Soc.
1
970, 96, 1150. (b) Beno, M. A.; Hope, H.; Olmsted, M. M.; Power, P. P.
Organometallics 1985, 4, 2117. (c) Zarges, W.; Marsch, M.; Harms, K.;
Boche, G. Chem. Ber. 1989, 122, 2303. (d) Boche, G.; Marsch, M.; Harbach,
J.; Harms, K.; Ledig, B.; Schubert, F.; Lohrenz, J. C. W.; Ahlbrecht, H.
Chem.Ber. 1993, 126, 1887-1894. (e) Zarges, W.; Marsch, M.; Harms,
K.; Koch, W.; Frenking, G.; Boche, G. Chem. Ber. 1991, 124, 543-549.
(
4) (a) West, P.; Purmort, J. I.; McKinley, S. V. J. Am. Chem. Soc. 1968,
stereochemistry of internal coordination of lithium favored a
9
1
2
1
9
5
0, 797. (b) O’Brian, D. H.; Hart, A. J.; Russell, C. R. J. Am. Chem. Soc.
975, 97, 4410. (c) Benn, R.; Rufinska, A. J. Organomet. Chem. 1982,
39, C19. (d) Fraenkel, G.; Winchester, W. R. J. Am. Chem. Soc. 1989,
11, 3794-3797. (e) Bates, R. B.; Beavers, W. A. J. Am. Chem. Soc. 1974,
6, 5001. (f) Thompson, T. B.; Ford, W. T. J. Am. Chem. Soc. 1979, 101,
459.
(6) (a) Van Eikema-Hommes, N. J. R.; B u¨ hl, M.; Schleyer, P. v. R. J.
Organomet. Chem. 1991, 409, 307-320. (b) Erusalimski, C. B.; Kormer,
V. M. Zh. Org. Khim. 1984, 20, 2028. (c) Boche, G.; Decher, G. J.
Organomet. Chem. 1983, 259, 31. (d) Clark, T.; Rohde, C.; Schleyer, P. v.
R. Organometallics 1983, 2, 1344. (e) Pratt, L. M.; Khan, I. M. J. Comput.
Chem. 1995, 16, 1070.
(5) (a) Takahashi, K.; Kondo, Y.; Asami, R. Org. Magn. Reson. 1974,
6
1
1
, 580-582. (b) Takahashi, K.; Takaki, M.; Asami, R. Org. Magn. Reson.
971, 3, 539-543. (c) McKeever, L. D.; Waack, R. J. Organomet. Chem.
971, 28, 145-151. (d) Waack, R.; Baker, E. U.; Doran, M. A. Chem.
(7) (a) See: Lipkowitz, K. B.; Uhegbu, C.; Naylor, A. M.; Vance, R. J.
Comput. Chem. 1985, 6, 662. (b) Setzer, W. N.; Schleyer, P. v. R. AdV.
Organomet. Chem. 1985, 24, 353. (c) Vanermen, G.; Toppet, S.; Beylen,
M. Van. J. Chem. Soc., Perkin Trans. 2 1986, 707.
Commun. 1967, 1291-1293.
1
0.1021/ja990485v CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/30/1999

SolVation of Benzylic Lithium Compounds
J. Am. Chem. Soc., Vol. 121, No. 45, 1999 10539
9
structure in which lithium lies closer to the allyl plane when
compared to the externally solvated species, in which lithium
lies normal to this plane.2 Compounds such as 1-4 are the only
of external ligands. These interconversions are shown in
reactions 11 f 12, 4 f 13, 7 f 14, and 10 f 15.
,4
13
6
allylic lithium compounds which exhibit one-bond C- Li spin
9
coupling constants. Their small values of 3-4 Hz contrast with
the 15 Hz that was observed for many monomeric organolithium
10
compounds. This argues for a small detectable C-Li covalency
in 1-4 and similar compounds.9
Analysis of the temperature-dependent 13C NMR line shapes
of several ion-paired allylic lithium compounds11 showed a rich
diversity of dynamic behavior that included (1) transfer of
coordinated lithium between faces of the allyl planes, (2)
reorientation of coordinated lithium on one side of the allyl
plane, (3) rotation about the allyl bonds, and (4) local reversible
Li,N (of ligand) dissociation with inversion at nitrogen. Ad-
ditionally in the case of the internally coordinated allylic lithium
compounds, averaging of the 13C- Li coupling constants was
used to monitor the dynamics of bimolecular C-Li bond
7
9
exchange.
Effects similar to those described above have been observed
for some benzylic lithium compounds.12 However, as will be
shown below, only the R-disubstituted benzylic lithium com-
pounds exhibit properties expected of delocalized carban-
1
3,14
ions;
only the R-monosubstituted species appear to be
partially delocalized. Their behavior will be described below.
In particular, we compare both the structures and dynamic
behavior of externally and internally solvated compounds.
Possible mechanisms responsible for these dynamic effects are
suggested.
Results and Discussion
Synthesis. Starting materials required for subsequent meta-
lation experiments were prepared as described by 1-10, below.
Two different silylations (chlorodimethylethylsilane and chlo-
romethylchlorodimethylsilane) of benzylic Grignard 3 produced
4
and 5, respectively. Amination of chloro compound 5 with
cis-2,5-bis(methoxymethyl)pyrrolidine gave 7, and a similar
reaction using 9 (from Grignard 8) with (S)-2-methoxymeth-
1
5
ylpyrrolidine provided enantiomerically pure 10 with ee >
8%. Compounds 4, 7, 10, and 11 all smoothly underwent
benzylic metalation at -78 °C in THF or diethyl ether using
9
Volatile components were removed under high vacuum from
preparations 12-15 and replaced by THF-d8 or diethyl ether-
d10. These samples, largely ca. 0.25 M, were retained for
subsequent NMR study.
6
7
n-butyllithium- Li for 7, 10, and 11 and its Li isotopomer for
. The metalations of 4 and 11 were carried out in the presence
4
NMR Parameters. Cursory inspection of the NMR data for
12-15 reveals that each of the spectra obtained at low-
temperature represents a single molecular species, although rapid
interconversion among several species giving rise to a single
averaged spectrum cannot be neglected. The manner in which
signal averaging takes place at higher temperatures implies that
the same species or species distribution prevails throughout the
temperature range investigated.
(
8) (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. D 1969, 1434. (c) Glaze, W. M.; Hanicak, J. E.; Moore, M. L.;
Chaudhuri, J. J. Organomet. Chem. 1972, 44, 39. (d) Glaze, W. H.; Hanicac,
J. E.; Chaudhuri, J.; Moore, M. L.; Duncan, D. P. J. Organomet. Chem.
973, 51, 13. (e) Bywater, S.; Lachance, P.; Worsfold, D. J. J. Phys. Chem.
975, 79, 2142.
1
1
(
9) (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.
10) (a) Bauer, W.; Winchester, W. R.; Schleyer, P. v. R. Organometallics
(
(
Table 1 lists the 13C NMR chemical shifts obtained at <180
K for compounds 12-15. These shifts are drawn around the
proposed structures. Compound 13 is drawn by analogy to the
structure proposed for R-(trimethylsilyl)benzyllithium‚TMEDA
1
987, 6, 2371-2379. (b) Fraenkel, G.; Chow, A.; Subramanian, S. J. Am.
Chem. Soc. 1995, 117, 6300-6307. (c) Fraenkel, G.; Chow, A.; Winchester,
W. R. J. Am. Chem. Soc. 1990, 112, 6190-6198.
(11) (a) Fraenkel, G.; Winchester, W. R. J. Am. Chem. Soc. 1990, 112,
1
382-1386. (b) Fraenkel, G.; Chow, A.; Winchester, W. R. J. Am. Chem.
Soc. 1990, 112, 2582-2585. (c) Cabral, J.; Fraenkel G. J. Am. Chem. Soc.
6
1
16
based on HOESY ( Li{ H}) measurements. These results are
1
1
993, 115, 1551-1557. (d) Fraenkel, G.; Cabral, J. J. Am. Chem. Soc. 1992,
3e
also consistent with Boche’s X-ray crystallographic structure.
The C NMR shifts for the R-secondary benzylic lithium
compounds reported here and previously are remarkably
14, 9067-9075.
1
3
(
12) Fraenkel, G.; Martin, K. J. Am. Chem. Soc. 1995, 117, 10336-
1
0344.
12
(
13) (a) Fraenkel, G.; Geckle, J. M. J. Am. Chem. Soc. 1980, 102, 2869-
similar. Each spectrum shows alternating shifts that are quali-
tatively indicative of a conjugated anion. However, 12-15
cannot be regarded as incorporating benzylic anions. Based on
2
9
880. (b) Fraenkel, G.; Russell, J. G.; Chen, Y. H. J. Am. Chem. Soc. 1973,
5, 3208-3215.
(14) (a) Spiesecke, H., Schneider, W. G. Tetrahedron Lett. 1961, 468.
(
b) Tokuhiro, T.; Fraenkel, G. J. Am. Chem. Soc. 1969, 91, 5005.
(15) Kippart, H.; Enders, D.; Fey, P. Org. Synth. 1987, 65, 173.
(16) Berger, S.; M u¨ ller, F. Chem. Ber. 1995, 128, 799-802.

10540 J. Am. Chem. Soc., Vol. 121, No. 45, 1999
Fraenkel et al.
Table 1. 13C NMR Chemical Shifts with Structures of 12-15
the well-known linearity of 13C shift with π charge,14 the model
K. One-bond 13C- Li coupling constants were also observed
6
1
3
system which exhibits shifts to be expected for a delocalized
for 14 (3 Hz) and 15 (2.5 Hz). That the CR resonance of 13
is selectively broadened at low temperature compared to the
other resonances and narrows with increasing temperature
1
3
benzylic anion is R-methyl-R-neopentylbenzyllithium‚TMEDA.
These 13C shifts are, in δ units; R, 71; ipso, 131; o, 105; m,
28; and p, 87. Values similar to those just listed have been
1
3
7
1
clearly implicates CR- Li coupling. Its value should be 9.8
13a
1
13
6
reported for cumylpotassium and 7-phenyl-7-metallonorbor-
nane (M ) K, Cs), all in THF.1 Compared to these latter
compounds the R-secondary benzylic lithium compounds 12-
Hz based on the 3.7 Hz value for J( C, Li) of 12. Both
8
compounds should have the same electronic structures, as
1
3
evidenced by their C ring shifts. They should therefore have
5 and those reported previously12 must all be regarded as
13
1
similar C-Li lithium coupling constants when corrected for
partially delocalized, to a similar extent. Consistent with this
proposal is our observation of one bond 13CR- Li spin coupling
of 3 to 4 Hz, as shown below. Such spin coupling implies a
small detectable degree of CR-Li covalence.
the isotope. Lithium-7 quadrupole-induced relaxation rates
increase with decreasing temperature and thus would be
6
1
3
responsible for narrowing of the C resonance on cooling the
R
7
sample of 13- Li‚TMEDA. Narrowing of this resonance with
increasing temperature can only be due to intermolecular
Below, the structure and dynamic behavior of 12-15 will
be considered separately. We will show how the dynamics of
inversion at CR, of carbon lithium bond exchange, and of rotation
around the ring benzyl carbon have been determined separately.
Down to 170 K, 13C NMR of R-TMS-benzyllithium‚TMEDA
carbon-lithium bond exchange. That compound 12 exhibits
13
6
CR- Li coupling whereas 13 does not suggests that tridentate
coordination of PMDTA to 12 reduces the rate of carbon-
lithium exchange when compared to that of 13‚TMEDA.
gave no evidence for 13CR- Li or CR- Li spin coupling, nor
did the TMEDA resonances show the type of fine structure
indicative of slow ion-ion reorientation within an unsym-
6
13
7
13
6
The one-bond C- Li coupling constants of 3-4 Hz reported
herein and previously for secondary benzylic lithium compounds
and for internally solvated monomeric allylic lithium com-
metrical ion-paired environment.1 By contrast, the CR NMR
1a,b
13
9
pounds are much smaller than the ca. 16 Hz values reported
6
spectrum of R-TMS-benzyllithium- Li in the presence of an
for most monomeric organolithium compounds. Among a wide
variety of solvated and unsolvated (RLi)n species (R ) alkyl,
vinyl, aryl, alkynyl) studied under conditions of slow intermo-
equivalent of pentamethyldiethylene triamine, 12‚PMDTA, in
THF at 230 K consists of an equally spaced 1:1:1 triplet with
13
6
1
13
6
splitting of 3.7 Hz due to one-bond CR- Li spin coupling.
Note that the spin of Li is 1. This pattern clearly indicates that
lecular C-Li exchange, the coupling constants J( C, Li) fall
into a remarkably simple and unexpected common pattern. The
values depend only on the state of aggregation, n (1), not on
6
compound 12 is a monomer, and intermolecular carbon-lithium
bond exchange is slow relative to the NMR time scale at 230
1
0
the nature of the organic moiety. While this relationship is
not well understood, it may be rationalized as follows:
(
17) Peoples, P. R.; Grutzner, J. B. J. Am. Chem. Soc. 1980, 102, 4709-
715.
18) Wehrli, F. W. J. Magn. Reson. 1978, 30, 193.
4
1
13
6
(
J( C, Li) ) 16/n
(1)

SolVation of Benzylic Lithium Compounds
J. Am. Chem. Soc., Vol. 121, No. 45, 1999 10541
We assume that these one-bond coupling constants are largely
19
due to the Fermi contact interaction and use the average energy
and normalization approximations proposed by Karplus, Grant,
2
1
and Lichtman in their treatment of one bond coupling con-
2
0,21
1
13
6
stants.
For a given state of aggregation, the J( C, Li) values
1
3
6
2
3
2
would be proportional to γ Cγ Lis Z , where s is the “s” char-
acter associated with the C-Li bond, Z is the effective nuclear
charge on carbon, and Z is a function of the C-Li covalence.
For this coupling constant to be independent of the organic
3
2
3
moiety, s Z would have to be constant. Such an exact inverse
relationship between “s” character and C-Li covalence is unex-
pected but not unreasonable from a qualitative point of
2
2-24
view.
Externally solvated allylic and tertiary benzylic lithium
1
3,18
1
3
6
13
7
compounds do not show CR- Li or C- Li spin coup-
ling. Their C shifts are consistent with delocalized carb-
anions within solvated ion-pairs. When compared to the 16 Hz
1
3
1
13
6
J( C- Li) values observed for the many “common pattern”
monomers, the much smaller coupling constants for secondary
benzylic and internally solvated allylic lithium compounds
suggest that their C-Li ionicities are somewhere between those
of the “common pattern” species and the solvated ion-pairs. This
13
is also supported by the C shifts reported in this paper and
12
previously.
Free PMDTA, at 180 K, shows 13C shifts, all in δ units, at
6.20 (N(CH3)2), 44.60 (NCH3), 58.50 (CH2), and 57.30 (CH2).
4
By comparison at 190 K, all carbons in contained PMDTA in
the sample of 12 are magnetically nonequivalent; all the
PMDTA resonances are of equal intensity, see Figure 1. This
13
observation, together with the 1:1:1 triplet for the CR resonance
13
6
due to C- Li coupling, implies that the PMDTA is tridentately
coordinated to lithium and that the spectrum represents a single
13
monomeric species. The pattern of the PMDTA C NMR bears
a close resemblance to the shifts of PMDTA complexed to
1
0b
monomeric mesityllithium‚PMDTA
and to those of neo-
Figure 1. 13C NMR spectra of 12‚PMDTA in diethyl ether-d ,
1
0
pentyllithium‚PMDTA.1 In the latter case, PMDTA reso-
0c
PMDTA portion, different temperatures, assignments from left (190 K
spectrum), in δ units: 57.32, 57.2, 54.71, 51.39 (CH ); 50.52, 47.01
(N(CH ) ); 45.34 (NCH ); 43.94, 42.02 (N(CH ) ).
nances, in δ units, at 43.00, 44.60, 47.50 and 49.80 (N(CH3)2),
2
1
0c
4
5.00 (NCH3), 51.50, 54.50 (CH2), and 57.10 (both CH2’s)
3
2
3
3 2
compare closely to 42.02, 43.94, 47.01, 50.52 (N(CH3)2, 45.34
1
were prepared incorporating geminal silyl methyls in each. We
hoped these geminal methyls would be magnetically nonequiva-
lent due to the expected chiral environment within these species.
The diastereotopicity could then be exploited to measure the
rate of overall inversion in 13 and 14, which is the result of
transfer of coordinated lithium between two faces of the benzyl
planes. Note that diastereotopic methylene hydrogens were
observed by Hoffmann et al. in low-temperature NMR of
R-hetero benzylic lithium compounds and were utilized to
(NCH3), 51.39, 54.71 (CH2), 57.2, and 57.32 (C H2) that were
obtained for compound 12, see Figure 1. Apparently, the
magnetic environment of PMDTA tridentately complexed to
lithium in these three RLi monomers appears to be largely
independent of the organic moiety.
Previous studies of ion-ion reorientation within allylic and
benzylic lithium compounds did not resolve the dynamics of
the different reorientational processes. Compounds 13 and 14
(
(
(
19) Ramsey, N. F. Phys. ReV. 1953, 91, 303.
25
measure the rate of inversion at CR.
20) Karplus, M.; Grant, D. M. Proc. Natl. Acad. U.S.A. 1959, 45, 1269.
21) (a) Grant, D. M.; Litchman, W. M. J. Am. Chem. Soc. 1965, 87,
At low temperature, the geminal silyl methyls of 13 and 14
1
3
3
2
994. (b) Litchman, W. M.; Grant, D. M. J. Am. Chem. Soc. 1967, 89,
228.
are, as we predicted, nonequivalent in C NMR. Each gives
rise to an equal doublet (Table 1), indicating that at this
temperature overall inversion is slow relative to the NMR time
scale. The NMR of attached ligand in 14 also supports slow
inversion at CR, as all ligand carbons are magnetically non-
equivalent. However, although benzylic inversion is slow in 13
at 180 K, as indicated by nonequivalent geminal silyl methyls,
(
22) (a) Lambert, C.; Schleyer, P. v. R. Angew. Chem., Int. Ed. Engl.
1
994, 33, 1129. (b) Van Eikema-Hommes, N. J. R.; Schleyer, P. v. R.,
unpublished results.
23) Consistent with and in support of these results is the observation of
(
13
7
one-bond CR- Li spin coupling of 7.0 Hz in a benzylic lithium compound
with the 2,6 positions bridged by CH2(OCH2CH2)4OCH2, wherein internal
encapsulation of lithium reduces the rate of intermolecular C-Li ex-
change: Ruhland, T.; Hoffmann, R. W.; Schade, S.; Boche, G. Chem. Ber.
coordinated TMEDA must be undergoing other reorientational
1
995, 128, 551-556.
13
processes faster than benzyl inversion since the TMEDA
C
(
24) A number of R-hetero and R, R-dihetero organolithium compounds
1
3
7
NMR consists of just two peaks down to 160 K. NMR line
shape analysis of the dynamics of inversion processes is treated
further below.
(hetero ) Te, Si, S, Se) show one bond C- Li spin coupling of 8-21
Hz, less than the 42 Hz values observed for the common pattern
organolithium compounds. (a) Reich, H. J.; Kulicke, K. J. J. Am. Chem.
Soc. 1995, 117, 6621-6622. (b) Reich, H. J.; Kulicke, K. J. J. Am. Chem.
Soc. 1996, 118, 273-274. (c) Reich, H. J.; Dykstra, R. R. J. Am. Chem.
Soc. 1993, 115, 7041-7042. (d) Reich, H. J.; Dykstra, R. R. Angew. Chem.,
Int. Ed. Engl. 1993, 32, 1649-1650.
(25) (a) Ahlbrecht, H.; Harbach, J.; Hoffmann, R. W.; Ruhland, T. Liebigs
Ann. 1995, 211-216. (b) Dress, R. K.; R o¨ lle, T.; Hoffmann, R. W. Chem.
Ber. 1995, 128, 673-677.

10542 J. Am. Chem. Soc., Vol. 121, No. 45, 1999
Fraenkel et al.
Compound 1526 is a special case due to the chirality at C2 of
the pyrrolidino ring. The compound was prepared starting from
S)-(+)-2-methoxymethylpyrrolidine, and it is unlikely that the
Table 2. Activation Parameters for Inversion, Benzylic Rotation,
and C,Li Bond Exchange for Benzylic Lithium Compounds in
THF-d
8
(
13C
resonance
q
q
∆H ((0.5
∆S
stereochemistry at C2 of pyrrolidine changed during transforma-
tion of 9 to 15. Carbon-13 NMR of 15 in THF-d8 solution was
monitored down to 180 K. At this temperature, rotation around
the Si-CR bonds in similar solvated benzylic lithium compounds
-
1
compd
process
kcal mol
)
((3 eu)
12
C,Li exchange
inversion
benzyl rotation
inversion
C
Si(CH
o C’s
R
9.4
5.1
-15
-21
5.6
-24
-26
-27
-24
-14
23
1
3
3
)
)
2
3
14.2a
4.8
1
2
has been shown to be slow relative to the NMR time scale.
1
4
Si(CH
3
The two rotamers around the silicon benzyl bond, with lithium
on opposite sides of the benzyl plane, would be diastereomeric
and would be expected to give rise to different NMR spectra.
OCH
NCH
OCH
o C’s
o C’s
3
4.3
3.9
4.6
2
However, at 180 K, 15 exhibited a single 13C NMR spectrum
benzyl rotation
benzyl rotation
8.5a
1
5
19.2
in which all carbons are nonequivalent and all resonances are
of equal intensity. A second species was not detected. If present,
its concentration would be less than 4% of the predominant
stereoisomer. By 240 K and above, interconversion of the two
possible stereoisomers would be fast relative to the NMR time
scale. The observed NMR spectrum would be the weighted
average of the two stereoisomeric spectra. The ratio of the two
stereoisomers might be expected to vary with temperature. Yet
a
(1.0 kcal.mol-1
The behavior of the 13C resonances of 14 and 15 qualitatively
R
resembles that of 12. However, the small coupling constants
and reduced resolution do not permit acceptable analysis of the
line shape changes.
We next consider NMR data indicative of inversion phe-
nomena. Equal doublets observed at low temperature for the
13
6
aside from averaging of the CR- Li spin coupling and of the
two ortho carbon resonances, the 13C NMR spectrum of 15 is
unchanged between 180 and 300 K. This signifies the predomi-
nance of a single stereoisomer. Changes seen for CR and the
ortho carbon NMR reflect the dynamics of bimolecular CR,Li
bond exchange and rotation around the ring-CR respectively.
This will be further described below.
13
geminal methyl silyl groups in C NMR of 13 and 14 were
ascribed to the chiral environment in these complexes. We
initially proposed that averaging of the diastereotopicity would
be the result of inversion alone and would thus separate and
reveal the dynamics of inversion from all the other ion-ion
reorientation processes. In fact, warming these two samples
above 180 K does average the geminal methyl doublets to single
lines at their respective centers. Comparison of observed and
calculated line shapes yields the rate constants for inversion.
The resulting activation parameters are listed in Table 2. It is
interesting that, although the ligands for 13 (external) and 14
Of the two structures that can be proposed for internally
solvated monomeric 15, the structure shown would minimize
the steric interaction between phenyl and methoxymethyl
compared to the species obtained by inverting the configuration
26
at CR in 15, as was proposed by Chan.
The results described for 15 also explain the 98% ee reported
(
internal) are different, the activation parameters for inversion
2
6
by Chan for the reaction product of 15 with benzaldehyde.
Clearly, this is largely the result of the presence of one
stereoisomer of 15 under their reaction conditions. The chemistry
used to prepare 15 was also carried out with racemic 2-meth-
oxymethylpyrrolidine, resulting in 15‚RAC. As expected, the
25
at CR are remarkably similar. Hoffmann and Ahlbrecht have
used a similar technique based on diastereotopicity of methylene
hydrogens to monitor the dynamics of carbanionic inversion in
several R-hetero, including R-SCH2Ph and R-SO2Ph, benzylic
lithium compounds. Where temperature-dependent data were
used, their ∆H values were larger, ca. 7 kcal mol , compared
to those reported herein.
25
1
3
C NMR spectra of 15 and 15‚RAC are identical.
q
-1
Dynamic Behavior. With increasing temperature, compounds
2-15 exhibited interesting changes in the line shapes of their
C NMR spectra. This is indicative of the operation of different
1
The effects described above with regard to the geminal methyl
1
3
13
silyl C resonances of 13 and 14 do not apply to 15 since this
dynamic effects. These effects are (1) bimolecular C,Li bond
exchange, (2) inversion at CR, which is equivalent to transfer
of coordinated lithium between opposite faces of the benzyl
plane, and (3) rotation around the ring benzyl bond. The
dynamics of each of these processes have been examined
separately.
compound largely exists as one enantiomer, >95%. Intercon-
version between the latter and a minor (<5%) stereoisomer
13
would not be detected among the C NMR data.
Compound 14 was designed to incorporate a pendant bis-
(
2,5-methoxymethyl)pyrrolidino group, a more constrained
12
version of the bis(2-methoxyethyl)amino moiety used before.
We previously showed that signal averaging effects seen for
pendant ligands on allylic lithium compounds could be ascribed
to several different effects, including inversion. At low tem-
perature, 180 K, all carbons of the pendant ligand on 14 are
magnetically nonequivalent. With increasing temperature, each
The dynamics of carbon-lithium bond exchange in 12 were
13
investigated using the CR NMR line shapes. With increasing
temperature above 230 K, the 1:1:1 13CR triplet due to C- Li
scalar coupling progressively averages to a single line by 290
K. This is the result of increasingly faster bimolecular CR-Li
13
6
13
bond exchange. In terms of CR NMR, the process is modeled
13
set of 1:1 C doublets for the CH3O, CH2O, ring CH2, and
for a system with 1 C in natural abundance as shown in eq 2.
3
pyrollidino CH carbons, respectively, progressively signal
average to single lines at their respective centers. NMR line
shape analysis of the collapsing doublets provides the rate
constants and ultimately the activation parameters listed in Table
2. All the ligand line shapes for 14 give rise to closely similar
activation parameters (Table 2). The dynamic process uncovered
1
3
Calculation of the CR line shape as a function of the specific
1
3
6
12
6
13
6
12 6
C Li* + C Li a C Li + C Li*
(2)
2
7
rate of exchange is carried out as described previously.
Comparison of observed and calculated 13CR NMR line shapes
provided the k1 values, which are the pseudo-first-order rate
constants for bimolecular C,Li exchange. Activation parameters
obtained from the Eyring plot are listed in Table 2.
(
27) (a) Kaplan, J. I.; Fraenkel, G. NMR of Chemically Exchanging
Systems; Academic Press: New York, 1980; Chapters 5 and 6. (b) Fraenkel,
G. In Techniques in Chemistry, InVestigation of Rates and Mechanisms of
Reactions, 4th ed.; Bernasconi, D. F., Ed; Wiley-Interscience: New York,
1986; Part 2, pp 357-604.
(26) Lamothe, S.; Chan, T. H. Tetrahedron Lett. 1991, 32, 1847-1850.

SolVation of Benzylic Lithium Compounds
J. Am. Chem. Soc., Vol. 121, No. 45, 1999 10543
here is clearly almost entirely inversion at CR, since the rates
derived from the ligand NMR line shape changes are so similar
to those from the geminal methyls on silicon. These latter
changes can only be due to inversion. It is not unexpected that
changes in the ligand 13C NMR line shapes of 14 result mainly
from inversion. The other reorientational motions, which include
rotation of ligand-coordinated lithium on one side of the benzyl
plane and fast reversible ligand lithium dissociation accompanied
by inversion at lithium, are not easily available to the ligand in
Scheme 1
1
4 due to its pendant and constrained nature. These other
motions are expected to be faster than inversion in the case of
lithium coordinated to an external ligand, such as TMEDA
complexed to 13.
As described above, 12 forms a 1:1 complex with PMDTA.
This is evidenced by the nonequivalence of all PMDTA carbons
at 190 K, Figure 1, and by the similarity of this spectrum to
It is interesting that, although inversion and rotation in these
compounds must involve very different mechanisms, the rate
ratios, inversion/rotation, at 300 K are similar. Their values for
1
2
1
3, 14, and 16 are 6.6, 5.0, and 10.9, respectively.
1
0b
those of PMDTA complexed to mesityllithium
and to
neopentyllithium.1 The spectra of the latter samples above 190
K revealed the operation of two dynamic processes: (1) the
rotation of the entire coordinated ligand with respect to the
carbanionic moiety, and (2) fast reversible N,Li dissociation
accompanied by inversion at nitrogen. These effects can be
detected in Figure 1. However, the quality of the data do not
allow an acceptable line shape analysis.
0c
Conclusions. What follows will summarize what has been
established in this work. Given that tertiary benzylic lithium
and potassium compounds are suitable models for delocalized
benzylic anions, we have found that a variety of R-secondary
benzylic lithium compounds, 12-15, have similar partially
The slowest of the dynamic processes observed in NMR of
2-15 is rotation around the ring-CR bond, as evidenced by
signal averaging of the ortho carbon resonances in these
1
compounds. All four compounds, 12-15, exhibit nonequivalent
13
13
delocalized structures. This is evidenced by their ring C NMR
ortho aromatic carbons at 220 K. There are also small C shifts
between the meta carbons. With increasing temperature, these
1
3
6
shifts and by the existence of a small, one-bond CR- Li spin
coupling constant observed under conditions of slow intermo-
lecular carbon,lithium bond exchange. By contrast, externally
solvated allylic lithium compounds appear to be delocalized
independently as to the extent of alkyl substitution. Only selected
internally coordinated allylic lithium compounds displayed
partial delocalization, due to the restricted stereochemistry of
1
:1 doublets signal-average to single lines at their respective
centers, as do the meta carbon resonances. Due to the magni-
tudes of these internal shifts, acceptable line shape analysis could
be carried out for 13, 14, and 15, but not for 12. Thus, the
collapsing ortho carbon doublets were treated as equally
populated half-spin-exchanging systems. Comparison of ob-
served with calculated line shapes yields the rates of rotation.
The resulting activation parameters for rotation around the ring
benzyl bonds in 13, 14, and 15 are listed in Table 2.
9
lithium coordination.
Carbon-13 NMR line shape changes have been used to
determine the dynamics of inversion at CR, C-Li bond
exchange, and rotation around the ring-CR bonds. Line shape
changes due to inversion and rotation are independent of
concentration of the organolithium compounds; hence, these
must be considered to be first-order processes. In contrast,
bimolecular C-Li bond exchange necessarily requires a dimeric
transition state that is most likely preceded by a dimeric
Though more localized than the tertiary benzylic lithium
13
compounds, the secondary species reported above and previ-
1
2
ously show substantial barriers around the ring-CR bonds of
-
1
1
1
4-18 kcal mol , respectively. The exception is compound
-
1
4, with a barrier of only 8 kcal mol .
These ring-CR barriers to rotation depend significantly on the
intermediate. This is analogous to the many solvated RLi dimers
13
nature of ligand coordination to lithium. It has been suggested
that development of the transition state for rotation involved
an increase in CR-Li covalence accompanied by some change
in the nature of lithium ligand coordination,1 see eq 3. With
2-5
described in the literature
and shown in 17. An abbreviated
mechanism is outlined in eq 4.
3a
(3)
similar ground states, differences among the ring-CR barriers
of these secondary benzylic lithium compounds must be due
more to energy changes of the transition states for rotation. The
significantly lower barrier for 14 might be rationalized as being
due to the constrained nature of the pendant 2,5-bis(methoxy-
methyl)pyrrolidino ligand, which renders its coordination to
lithium particularly energetically favorable in the transition state
for rotation.
A unified mechanism is proposed in Scheme 1 to account
for both inversion and rotation phenomena in 12-15, taking
account of the first-order character of both of these processes.
The circles diagram the projection along the ring-CR bonds, to
show the bonding arrangement around CR. The dotted line
represents the plane of the aromatic ring end-on. To account
for inversion, we propose a very fast equilibrium between the

10544 J. Am. Chem. Soc., Vol. 121, No. 45, 1999
Fraenkel et al.
partially delocalized ground states, 18, and a delocalized higher
energy ion-pair, 19. Rate-determining inversion is the result of
transfer of coordinated lithium between two faces of the aromatic
plane of the ion-pair, 19 f 19i. We have already observed
inversion of ion-paired delocalized allylic lithium compounds
1-[[(p-tert-Butylbenzyl)dimethylsilyl]methyl]-cis-2,5-bis(methoxym-
ethyl)pyrrolidine (7). In a 10-mL round-bottom flask equipped with
a reflux condenser, an argon inlet, and a magnetic stir bar, cis-2,5-bis-
(
methoxymethyl)pyrrolidine (6) and p-tert-butyl-R-(chloromethyldim-
ethylsilyl)toluene (5) (2.0 g, 10 mmol) were combined and heated at
0 °C overnight. The mixture was poured into an addition funnel and
8
q
-1 11
with ∆H values of 6-8 kcal mol . In every equilibrium
mixture known to contain similar concentrations of covalent
and ion-paired delocalized forms, for example, neopentylallyl-
lithium, interconversion of the two species was too fast to
the flask rinsed first with water (10 mL) and then diethyl ether (25
mL). These rinsings were also added to the contents of the addition
funnel. The ether layer was washed with aqueous saturated NaCl (10
mL) and then dried over anhydrous Na
2 4
SO . The ether was removed
8
measure down to 160 K relative to the NMR time scale. Below
by rotary evaporation and the residue purified by bulb-to-bulb distil-
this temperature, precipitation of the reagent precluded acquisi-
tion of usable NMR data. With respect to rotation around the
ring CR bonds, the proposed transition state has increased CR-
Li covalence compared to the ground state, 18, and the CR-Li
bond lies in the aromatic plane. This is similar to the transition
1
3
lation to yield the title product, 2.01 g, in 66% yield. H NMR (CDCl ,
δ ) 7.26 ppm): δ 7.23 (d, J ) 12.3 Hz, 2H), 6.95 (d, J ) 12.3 Hz,
2H), 3.33 (m, 2H), 3.32 (s, 6H), 3.13 (m, 2H), 2.78 (s, br, 2H), 2.34
(s, br, 2H), 2.11 (s, 2H), 1.88 (m, 2H), 1.61 (m, 2H), 1.30 (s, 9H),
0.06 (s, 6H). 13C NMR (CDCl
, δ ) 77.0 ppm): δ 146.68, 137.34,
3
state for allyl rotation proposed by Schleyer.6
a
127.95, 125.18, 34.39, 31.70, 7.50, 6.72, -3.83.
Because our internally solvated organolithium compounds are
monomers with well-defined structures, their reactivity should
be more selective and more easily understood compared to that
of the more reactive externally solvated organolithium com-
pounds. These compounds are aggregated to different degrees
with variable solvation.
[r-[[[(cis-2,5-Bis(methoxymethyl)-1-pyrrolidinyl]methyl]dimeth-
ylsilyl]-p-tert-butylbenzyl]lithium (14). A 35-mL Schlenk tube equipped
with a magnetic stir bar and an argon inlet was flame-dried under
vacuum and then flushed with argon. Freshly distilled THF (2 mL)
and 1-[[(p-tert-butylbenzyl)dimethylsilyl]methyl]-cis-2,5-bis(methoxym-
ethyl)pyrrolidine (7) (0.206 g, 0.55 mmol) were introduced into the
tube. After the solution was chilled to -78 °C with a dry ice-acetone
6
bath, a solution of n-butyllithium- Li (0.36 mL, 1.6 M in hexanes, 0.58
Experimental Section
mmol) was introduced by syringe. The mixture was stirred for 2 h at
-78 °C and then allowed to warm to room temperature and stirred for
4 h.
NMR Samples. A dry 7-in., 5-mm-o.d. precision NMR tube with
an attached female S 14/20 joint was fitted with a 2-mm straight bore
stopcock adaptor, the latter with a male S 24/40 joint at each end. The
assembly was gently flame-dried under vacuum and then filled with
argon, and the open end was closed with a serum cap. Then, the RLi
sample was syringed into the NMR tube and solvent carefully
evaporated under moderate (0.1 Torr) vacuum. After removal of
remaining solvent under high vacuum (vacuum line), deuterated solvent
was vacuum-transferred into the RLi sample. The resulting sample was
degassed using two freeze-thaw cycles and then sealed off frozen under
high vacuum.
An NMR sample with a concentration of 0.25 M was prepared as
1
previously described. H NMR (THF-d
8
, δ ) 1.73): δ 0.658 (d, J )
8
3
0
1
.5 Hz, 2H), 6.289 (d, J ) 8.5 Hz, 2H), 3.67 (d, 2H), 3.61 (d, 2H),
.11 (s, 6H), 2.59 (s, br, 2H), 2.18 (s, 2H), 1.78 (m, 4H), 1.36 (s, 1H),
13
.01 (s, 6H). C NMR (THF-d , δ ) 25.3): δ 154.9, 125.56, 124.71,
8
19.19, 73.20, 70.06, 59.18, 51.10, 34.53, 34.10, 32.83, 27.70, 0.95.
Acknowledgment. This research was generously supported
by the National Science Foundation (Grant No. CHE9615116)
as was, in part, acquisition of the NMR equipment used in this
work. We are grateful to Dr. Charles Cottrell, Central Campus
Instrumentation Center, who provided invaluable advice and
assistance with NMR technology. We thank Dr. Kurt Loening
for advice on nomenclature.
NMR Spectroscopy. NMR data were obtained using the Bruker
AC-200 or Avance 300 instrument. The Avance 300 was used for all
variable-temperature work.
6
n-Butyllithium- Li. Within the drybox, under argon atmosphere, a
2
50-mL three-neck round-bottom flask equipped with a condensor, an
addition funnel, and a glass-coated magnetic stir bar was loaded with
finely cut lithium-6 (1.5 g, 0.25 at. wt). The apparatus was closed off
and removed from the box. Pentane (100 mL) was introduced, and
then 1-chlorobutane (7.3 g, 79 mmol) was slowly added from the
addition funnel. The mixture was left to stir overnight, which re-
sulted in a large amount of purple precipitate. Equal volumes of
supernatant were transferred by cannula into two centrifuge tubes (argon
atmosphere) closed with serum caps. The tubes were centrifuged for
Supporting Information Available: Derivation of the NMR
line shape equations for C,Li exchange, calculated and observed
NMR line shapes, Eyring plots, and experimental details
concerning compounds 2, 4, 5, 9, 10, 12, 13, and 15 (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
2
0 min and then cannulated into a Schlenk flask. Titration indicated a
6
solution 0.6 M in n-butyllithium- Li with just a trace (<0.1%) of
alkoxide.
JA990485V