Solvent-dependent lithium bridging in allenyl-propargyllithium reagents

Article abstract of DOI:10.1021/ol990412g

(figure presented) The solution structures of two allenyl-propargyllithium reagents (6-Li and 7-Li) which give ¹³C NMR chemical shifts intermediate between those expected for the allenyl and propargyl isomers have been studied by the Saunders i

Full text of DOI:10.1021/ol990412g

ORGANIC
LETTERS
2000
Vol. 2, No. 6
783-786
Solvent-Dependent Lithium Bridging in
Allenyl−Propargyllithium Reagents
Hans J. Reich* and Jennifer L. Thompson
Department of Chemistry, UniVersity of Wisconsin, Madison, Wisconsin 53706
hjreich@facstaff.wisc.edu
Received December 30, 1999
ABSTRACT
The solution structures of two allenyl−propargyllithium reagents (6-Li and 7-Li) which give 13C NMR chemical shifts intermediate between
those expected for the allenyl and propargyl isomers have been studied by the Saunders isotope perturbation technique. Variable equilibrium
isotope shifts were detected, showing that these reagents adopt structures ranging from equilibrating localized allenyl- and propargyllithium
reagents (1A h 1B) to equilibrating unsymmetrically bridged structures (11) and symmetrically bridged structures (2), depending mostly on
the details of solvation.
Important aspects of the chemistry of allenyl-propargyl (A-
P) lithium reagents are likely to depend on the association
between carbon and lithium, which can have the limiting
forms 1-3. Reagents with innocuous substituents such as
structures of compounds such as 6-Li1b and 7-Li which show
13C shifts intermediate between the allenyl and propargyl
compounds that we address here.
H and alkyl at either CP or CA, as well as most of those with
anion stabilizing groups such as PhSe, PhS, RO, R3Si, and
Ph at CA (e.g., 4A-Li), have the allenyl structure 1A, on the
basis of the strongly deshielded central carbon atom CC in
the 13C NMR spectra (170-190 ppm) and well-defined C-Li
coupling at the lithium-bearing carbon.1a,b,2
Reagents having anion stabilizing groups at CP show more
diverse structures. Those with alkoxy substituents at CP are
allenyllithiums; those with PhSe and PhS (e.g., 5P-Li) are
localized propargyllithiums (CC at 90-100 ppm).1a It is the
Several explanations for the 13C chemical shifts of 6-Li,
7-Li, and their analogues can be offered: (a) the compounds
have the bridged structure 2 as suggested by computational
studies3a and single-crystal X-ray structures,4 (b) they are
equilibrating mixtures of 1A and 1P isomers (with Keq near
10.1021/ol990412g CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/01/2000

1
1) or equilibrating structures intermediate between 1 and 2,
(c) they are separated ion pairs (SIPs) 3,1b,5a or (d) they are
equilibrating mixtures of contact ion pairs (CIPs6), either 1
or 2, and SIPs. A careful study of the solvent and temperature
dependence of the 13C NMR chemical shifts has provided
circumstantial evidence that 6-Li and its analogues are
equilibrating A-P isomers with a barrier to isomerization
of less than 4 kcal/mol.1b
observation that JCH is larger in SIPs of this type than in
CIPs10a and larger in CP of allenyllithiums than in propargyl-
lithiums.10b We assume that JCH at least qualitatively
correlates with bond strength and bond stiffness.
1
The required deuterated precursors 8 and 10 were prepared
as shown in Scheme 1. The small amount of deuterated allene
The Saunders isotope perturbation technique7 can un-
ambiguously distinguish between static and rapidly equili-
brating structures. Static structures will show only intrinsic
H/D isotope effects on chemical shifts (∆δD-H). The ∆δD-H
values have little or no temperature dependence, and they
are rapidly attenuated as the distance from the site of isotopic
substitution increases. Equilibrating structures will show
temperature-dependent ∆δD-H values, provided that two
criteria are met: there must be a difference in the C-H(D)
bond strength between the two isomers, and the nucleus
observed must have a chemical shift difference (∆δ) in the
two isomers. In fact, unlike the intrinsic shifts, the equilib-
rium isotope shifts are independent of the number of bonds
separating the site of isotopic substitution and the observed
nucleus but are directly proportional to ∆δ.
Static structures such as 2 and 3 (explanations a and c)
should show only intrinsic isotope effects. Rapidly equili-
brating structures such as 1 or a mixture of CIPs6 and SIPs
(explanations b and d) should show isotopic perturbation
effects since the hybridization at CP in 1A or 3 is expected
to be near sp2, whereas in 1P this carbon is pyramidalized.10
Deuterium substitution at CP would move the equilibrium
away from 1P and toward 1A and 3, on the basis of the
Scheme 1. Preparation of the Deuterated Precursors for 6-Li
and 7-Li
9 formed during deuteration of the allenyl-titanium reagent
was removed by a partial metalation and carboxylation (9 is
kinetically more acidic than 8). Solutions of the lithium
reagents were prepared in the appropriate solvents by in situ
metalations of 1-(phenyldimethylsilyl)-2-butyne and 1-phen-
yl-2-butyne (and deuterated analogues) with t-BuLi and
n-BuLi, respectively. These solutions had to be kept cold to
avoid prototropic isomerization and other decompositions.
We have examined the 13C NMR spectra of mixtures of
the H and D isotopomers at CP of 6-Li and 7-Li in THF,
2,5-dimethyltetrahydrofuran, diethyl ether, and/or dimethyl
ether at several temperatures. Both compounds show sub-
stantial equilibrium isotope shifts (∆δD-H, defined as δD -
δH)11 under almost all conditions for CA, CC, and the CH3
group. The resonance for CC is the most sensitive because
of the large chemical shift difference (approximately 80 ppm)
between the A and P isomers but is sometimes rather broad.
The isotope shifts for CP are hard to measure because of the
low sensitivity of the deuterated carbons (splitting, no
Overhauser enhancement) and because at low temperatures
deuterium quadrupolar relaxation is sufficiently fast that a
well-resolved peak is not seen for the CD carbon. Some
partial 13C NMR spectra and representative isotope shift data
(1) (a) Reich, H. J.; Holladay, J. E. J. Am. Chem. Soc. 1995, 117, 8470-
8471. Reich, H. J.; Holladay, J. E.; Mason, J. D.; Sikorski, W. H. J. Am.
Chem. Soc. 1995, 117, 12137-12150. (b) Reich, H. J.; Holladay, J. E.;
Walker, T. G.; Thompson, J. L. J. Am. Chem. Soc. 1999, 121, 9769-9780.
Reich, H. J.; Holladay, J. E. Angew. Chem., Int. Ed., Engl. 1996, 35, 2365-
2367. (c) Reich, H. J.; Borst, J. P.; Dykstra, R. R.; Green, D. P. J. Am.
Chem. Soc. 1993, 115, 8728-8741.
(2) Van Dongen, J. P. C. M.; van Dijkman, H. W. D.; de Bie, M. J. A.
Recl. TraV. Chim. Pays-Bas 1974, 93, 29-32.
(3) Lambert, C.; Schleyer, P. v. R.; Wu¨rthwein, E.-U. J. Org. Chem.
1993, 58, 6377-6389.
(4) Several X-ray crystal structures of allenyl-propargyllithium reagents
have shown dimeric structures, with lithium coordination to both CA and
CP: (a) Schleyer, P. v. R. Pure Appl. Chem. 1984, 56, 151-162. (b) Setzer,
W. N.; Schleyer, P. v. R. AdV. Organomet. Chem. 1985, 24, 353-451. (c)
Dem’yanov, P.; Boche, G.; Marsch, M.; Harms, K.; Fyodorova, G.;
Petrosyan, V. Liebigs Ann. 1995, 457-460.
(5) (a) Dem’yanov, P. I.; Styrkov, I. M.; Krut’ko, D. P.; Vener, M. V.;
Petrosyan, V. S. J. Organomet. Chem. 1992, 438, 265-88. Dem’yanov, P.
I.; Krut’ko, D. P.; Borzov, M. V.; Luk’yanov, E. V.; Petrosyan, V. S. Russ.
Chem. Bull. (Engl. Transl.) 1997, 46, 1939-1947. (b) Maercker, A.;
Fischenich, J. Tetrahedron 1995, 51, 10209-10218.
(6) We make no distinction between contact ion pairs (CIPs) and
“covalent” organolithium reagents and use the term simply to identify
lithium reagents with a C-Li contact and distinguish such species from
separated ion pairs (SIPs), where there is no C-Li contact.
(9) Allyllithiums can become substantially localized in nonpolar solvents
or when chelation effects favor such structures: Fraenkel, G.; Martin, K.
V. J. Am. Chem. Soc. 1995, 117, 10336-10344. Fraenkel, G.; Qiu, F. J.
Am. Chem. Soc. 1997, 119, 3571-3579 and references therein.
1
(7) Siehl, H.-U. AdV. Phys. Org. Chem. 1987, 23, 63-163. This technique
has been occasionally applied to problems in carbanion chemistry,8
especially to the structures of allylmetals.8a-d
(10) (a) For 7-Li JC-H is 155 Hz in 3:2 THF-ether and 161 Hz in
THF-ether-HMPA, where the compound is fully ion separated. For 5P-
Li JC-H at CP is 155 Hz in the CIP and 175 Hz in the SIP.1a (b) Comparison
1
(8) (a) Schlosser, M.; Sta¨hle, M. Angew. Chem., Int. Ed. Engl. 1980,
19, 487-489. (b) Faller, J. W.; Murray, H. H.; Saunders: M. J. Am. Chem.
Soc. 1980, 102, 2306-2309. (c) Equilibrating bridged structures have been
proposed for allyllithium dimer: Winchester, W. R.; Bauer, W.; Schleyer,
P. v. R. J. Chem. Soc., Chem. Commun. 1987, 177-179.9 (d) Fraenkel, G.;
Chow, A.; Winchester, W. R. J. Am. Chem. Soc. 1990, 112, 1382-1386.
(e) Ahlberg, P.; Davidsson, O¨ .; Lo¨wendahl, M.; Hilmersson, G.; Karlsson,
A.; Håkansson, M. J. Am. Chem. Soc. 1997, 119, 1745-1750. (f) Jonsa¨ll,
G.; Ahlberg, P. J. Chem. Soc., Perkin Trans. 2 1987, 461-467.
of JC-H values in allenyl and propargyl structures is difficult, since
substituents have to be different, but 1JC-H at CP in allenyllithium itself is
162 Hz,2 somewhat larger than the value for 5P-Li (155 Hz). Sulfur
1
substitution increases JC-H (CH4, 125 Hz; CH3-S-CH3, 137.9 Hz).
(11) Isotope shifts were typically measured at six to eight temperatures
between -135 and 0 °C. As required for an isotope perturbation shift, all
showed a substantial temperature dependence (e.g., ∆δD-H for CC in 6-Li
varied between 1376 ppb at -107 °C and 716 ppb at -1 °C in 4.8:3.2:1
THF-ether-pentane).
784
Org. Lett., Vol. 2, No. 6, 2000

for the lithium reagents and the precursor 6P-H are reported
in Figures 1 and 2. As expected from the observation that
of SIP) by its NMR spectroscopic behavior during a low-
temperature HMPA titration.1c,13
An extensive spectroscopic study of a series of phenylated
allenyl-propargyllithiums showed that many were at least
partially SIPs in THF solution;4c,5 therefore, this is also a
real possibility for 7-Li. Although the HMPA titrations of
7-Li in ether solvents were significantly different from those
of fully separated ions, they lack some of the most distinct
signatures of CIPs.14 Nevertheless, a CIP/SIP explanation
for the observed isotope shifts can also be ruled out. It is
implausible that 7-Li would be partially ion separated over
the range of solvent polarities (from pure ether to 60% THF/
40% ether) and temperatures (from -151 to 14 °C) over
which isotope perturbations were observed. A more direct
argument is that the direction of the isotope shift is wrong
for an CIP/SIP explanation. The separated ion has the
stronger C-H bond (as indicated by JC-H10); thus, deuterium
substitution should favor the SIP over the CIP. Since the
13C chemical shift of CC of the SIP (formed with excess
HMPA) is upfield of that in the CIP (Figure 2), the signal
for the deuterated isotopomer should be upfield of that for
the protonated one, instead of downfield as is actually
obserVed.
Figure 1. 90.56 MHz 13C NMR spectra of a 1.2:1 mixture of 6-Li
and 6-Li-d at -58 °C in 3:2 THF-ether: chemical shifts (for the
protio compound) and 13C isotope shifts of H/D labeled 6-Li and
7-Li and the precursor 6P-H.
the chemical shifts of each of the four carbons in the allenyl
isomer are downfield of those in the propargyl isomer (see
4A-Li and 5P-Li), the deuterated compound gives the
downfield signal for each carbon, in the opposite direction
to the intrinsic isotope shifts.
This isotope perturbation experiment is more complicated
than in most previous applications of the technique in
organometallic chemistry8 because the equilibrium constant
between the two species is infinitely variable. In another
sense the interpretation is simpler since the most useful
carbon is â to deuterium, where intrinsic isotope shifts are
almost negligible. The magnitude of the isotope perturbation
shift depends principally on three factors: (1) the chemical
shift difference (∆δ) between the two equilibrating species
(ca 80 ppm for 1A/1P), (2) the ratio of the two species at
equilibrium (∆δD-H is maximum at a 1:1 ratio of isomers,
corresponding to a δ value of ca. 140 ppm for CC), and (3)
the isotope effect on the free energy difference between the
two species (∆∆GD-H14). Figure 3 shows in graphical form
Figure 2. Temperature-dependent 13C chemical shifts and 90.56
MHz 13C NMR spectra illustrating the isotope perturbation shifts
observed for CC of 7-Li in several solvents.
The detection of substantial isotope shifts rules out the
static bridged structure 2 for 6-Li under all conditions and
for 7-Li under all conditions except in dimethyl ether at low
temperature. Before we can conclude that these are equili-
brating mixtures of the 1A and 1P isomers, however, we
have to consider the alternative of a CIP/SIP equilibrium.
The most distinctive feature of a CIP6 is the observation of
C-Li J coupling; unfortunately, no such coupling could be
observed for either 6-Li or 7-Li12 even at temperatures down
to -155 °C. However, 6-Li can be firmly identified as a
CIP in THF-ether mixtures (without a significant fraction
Figure 3. Isotope perturbation shifts for CC in 6-Li and 7-Li in
various solvents, T ) -90 to -100 °C. The dotted lines are
calculated values for the ∆∆GD-H values indicated, assuming a 1A/
1P equilibrium with δ1P 100 ppm and δ1A 180 ppm.
Org. Lett., Vol. 2, No. 6, 2000
785

some of the observed shifts and the theoretical curves for
∆∆GD-H ) 0.2, 6, 18, and 29 cal/mol.
structures,8c to fully bridged structures such as 2.9 We propose
a static bridged structure for 7-Li in dimethyl ether, since
under these conditions there is no detectable isotopic
perturbation (<18 ppb), although δ is near the centroid of
the chemical shift range where the largest ∆δD-H values are
expected. Other features of the NMR data are also supportive
of a different structure for 7-Li in dimethyl ether compared
to either THF or diethyl ether. In dimethyl ether the chemical
shifts of CC have the opposite temperature dependence and
are substantially more allenic in character than in either ether
or THF, although dimethyl ether is between the two solvents
in donor ability (Figure 2). Similarly, the isotope shifts for
6-Li in 2,5-dimethyltetrahydrofuran are very small, suggest-
ing equilibration between structures (11) with a substantial
degree of bridging in this solvent. This reagent shows
essentially temperature-independent shifts in 2,5-dimeth-
yltetrahydrofuran but strongly temperature-dependent chemi-
cal shifts in dimethyl ether and THF, where the isotope shifts
are also much larger (CC changes by 30.7 ppm between -35
and -155 °C in dimethyl ether).1b
We can make an estimate for ∆∆GD-H as follows. The
observed ∆δD-H value of 1.9 ppm for allylmagnesium
bromide at 24 °C corresponds to a ∆∆GD-H ) 43 cal/mol
(RT ln(K)), using an estimated difference of 70 ppm for the
chemical shifts of C-1 and C-3.8a Similarly, a ∆∆GD-H value
of 24 cal/mol can be calculated for the (trimethylstannyl)-
cyclopentadiene system studied by Faller, Murray, and
Saunders.8b The bonding in both of these structures is likely
to be localized.15 The highest ∆δD-H value observed with
the allenyl-propargyllithium reagents studied (1.4 ppm for
6-Li in 3:2 THF/ether) corresponds to a ∆∆GD-H value of
29 cal/mol, which is in the same range as the examples above
and thus consistent with a largely localized pair of structures.
A smaller value of ∆∆GD-H is expected for a lithium than
for a magnesium or stannyl reagent because of smaller degree
of covalency in the C-Li bond. However, the wide variation
in the ∆δD-H values cannot be explained simply by the
variability in the fraction of A and P isomers (∆GA-P) in
the equilibrium mixture (see Figure 3). Thus, either the
isotope effect ∆∆GD-H and/or the chemical shifts of the two
species in equilibrium must be changing as a function of
solvent.
The isotope perturbation and other NMR data can be
understood in terms of structures which range from es-
sentially localized equilibrating A-P isomers (6-Li in
dimethyl ether), to equilibrating mixtures of partially bridged
It is not the specific solvent but the interplay between anion
structure and lithium solvation that is important. For example,
in dimethyl ether 7-Li shows the smallest and 6-Li one of
the largest isotope shifts we found. It appears that minor
differences in steric effects and solvation have a major
influence on the extent to which the lithium is locally bonded
to a single carbon (where maximum solvation is possible),
bonded unsymmetrically to both carbons, or even bonded
symmetrically.16 These results are consistent with the prefer-
ence for bridged structures found in computational studies3
as well as the very low barrier to interconversion of the A
and P isomers of 6-Li.1b
(12) Although lithium-carbon coupling has only rarely been detected
for benzyl (Fraenkel, G.; Duncan, J. H.; Martin, K.; Wang, J. H. J. Am.
Chem. Soc. 1999, 121, 10538-10544) and allyllithiums,9 we had no
difficulty in observing C-Li coupling in 4A-Li. The coupling constant is
comparable in size to that found for other allenyllithiums.1b
(13) Some of the features that distinguish CIPs from SIPs during an
HMPA titration are as follows. RLi species that are contact and do not
undergo ion separation form at most RLi(HMPA)2 species even with excess
HMPA. Species which are CIPs but undergo ion separation when HMPA
is added will skip an HMPA coordination number on ion separation (e.g.,
they will go from RLi(HMPA) to mostly R-//Li+(HMPA)3). RLi species
which are SIPs will show a characteristic progression from the mono- to
the tetra-HMPA complex as HMPA is added.1c There is an intermediate
class of easily separated CIPs such as 7-Li, where the differences are subtle
and the distinction cannot be made easily.
(14) The value ∆∆GD-H is defined by the amount that the A-P
equilibrium constant is perturbed by replacing H with D at CP. The curves
in Figure 3 are drawn by assuming that this value is independent of the Keq
value between the A and P isomers. Thus: Keq-H ) exp(-∆GA-P/RT),
Keq-D ) exp(-(∆GA-P + ∆∆GD-H)/RT).
(15) Allylmagnesium chloride-TMEDA has a localized σ structure in
the crystal: Marsch, M.; Harms, K.; Massa, W.; Boche, G. Angew. Chem.,
Int. Ed. Engl. 1987, 26, 696-697.
Acknowledgment. We thank the National Science Foun-
dation for financial support of this work.
OL990412G
(16) An interesting example where a lithium bonds in an η3 fashion to
indenyl anion when bis-coordinated to sparteine and in an η1 fashion when
tris-solvated by THF is provided by: Hoppe, I.; Marsch, M.; Harms, K.;
Boche, G.; Hoppe, D. Angew. Chem., Int. Ed. Engl. 1995, 34, 2158-2160.
786
Org. Lett., Vol. 2, No. 6, 2000