Pseudomonomolecular, ionic sp2-stereoinversion mechanism of 1-aryl-1-alkenyllithiums

Article abstract of DOI:10.1021/om4000852

The trans/cis stereoinversion of the trigonal carbanion centers C-α in a series of monomeric 2-(α-aryl-α-lithiomethylidene)-1,1,3,3- tetramethylindanes (known to be trisolvated at Li) is rapid on the NMR time scales (400 and 100.6 MHz) in THF solution. The far-reaching redistribution of electric charge in the ground-state molecules caused by lithiation (formal replacement of α-H by α-Li) is illustrated through NMR shifts, Δδ. The transition states for stereoinversion are significantly more polar and charge-delocalized than the ground states (Hammett ρ = +5.2), pointing to a mechanism that involves heterolysis of the C-Li bond via a solvent-separated ion pair (SSIP). This requires immobilization of only one additional (the fourth) THF molecule at Li⁺, which accounts for part of the apparent activation entropies of ca. -23 cal mol^-1 K ^-1 and constitutes a kinetic privilege of THF depending on microsolvation at Li. Thus, the sp²-stereoinversion process is "catalyzed" by the solvent THF; its mechanism is monomolecular with respect to the ground-state species because the pseudo-first-order rate constants, measured through NMR line shape analyses, are independent of the concentrations (inclusive of decomposition) of the dissolved species (hence no associations and no dissociation to give free carbanion intermediates). In the deduced pseudomonomolecular mechanism (bimolecular through solvent participation), the angular C-α of the SSIP undergoes rehybridization (approximately in-plane inversion) through a close-to-linear transition state; this motion occurs with a concomitant "conducted tour" migration of Li⁺(THF)₄ and is unimpaired by additional ortho-methylations at α-aryl. The synthetic route started with preparations of three α-chloro congeners through the carbenoid chain reaction, followed by vinylic substitution of α-Cl by α-SnMe ₃ (most efficient in THF despite steric congestion). The final Sn/Li interchange reaction afforded the new 1-aryl-1-alkenyllithium samples, initially uncontaminated by free Li⁺.

Full text of DOI:10.1021/om4000852

Article
pubs.acs.org/Organometallics
Pseudomonomolecular, Ionic sp2‑Stereoinversion Mechanism of
1‑Aryl-1-alkenyllithiums†
Rudolf Knorr,* Thomas Menke, Claudia Behringer, Kathrin Ferchland, Johann Mehlstaubl,
̈
and Ernst Lattke
Department Chemie, Ludwig-Maximilians-Universitat, Butenandtstrasse 5-13, 81377 Munchen, Germany
̈
̈
S
* Supporting Information
ABSTRACT: The trans/cis stereoinversion of the trigonal
carbanion centers C-α in a series of monomeric 2-(α-aryl-α-
lithiomethylidene)-1,1,3,3-tetramethylindanes (known to be
trisolvated at Li) is rapid on the NMR time scales (400 and
100.6 MHz) in THF solution. The far-reaching redistribution
of electric charge in the ground-state molecules caused by
lithiation (formal replacement of α-H by α-Li) is illustrated
through NMR shifts, Δδ. The transition states for stereo-
inversion are significantly more polar and charge-delocalized
than the ground states (Hammett ρ = +5.2), pointing to a mechanism that involves heterolysis of the C−Li bond via a solvent-
separated ion pair (SSIP). This requires immobilization of only one additional (the fourth) THF molecule at Li+, which accounts
for part of the apparent activation entropies of ca. −23 cal mol−1 K−1 and constitutes a kinetic privilege of THF depending on
microsolvation at Li. Thus, the sp2-stereoinversion process is “catalyzed” by the solvent THF; its mechanism is monomolecular
with respect to the ground-state species because the pseudo-first-order rate constants, measured through NMR line shape
analyses, are independent of the concentrations (inclusive of decomposition) of the dissolved species (hence no associations and
no dissociation to give free carbanion intermediates). In the deduced pseudomonomolecular mechanism (bimolecular through
solvent participation), the angular C-α of the SSIP undergoes rehybridization (approximately in-plane inversion) through a close-
to-linear transition state; this motion occurs with a concomitant “conducted tour” migration of Li+(THF)4 and is unimpaired by
additional ortho-methylations at α-aryl. The synthetic route started with preparations of three α-chloro congeners through the
carbenoid chain reaction, followed by vinylic substitution of α-Cl by α-SnMe3 (most efficient in THF despite steric congestion).
The final Sn/Li interchange reaction afforded the new 1-aryl-1-alkenyllithium samples, initially uncontaminated by free Li+.
−78 °C on the laboratory time scale: The pseudo-first-order
rate constants became enhanced with increasing concentrations
of 2 (and also of LiI) or upon dilution with the aggregation-
promoting solvent Et2O, whereas a 20-fold retardation was
measured for monomeric 2 when the dimerization process was
inhibited.1
INTRODUCTION
■
Knowledge of the temperature-dependent configurational
stability of C−Li centers can be helpful for the use of
organolithium reagents in asymmetric syntheses. The reaction
rates of such configurational inversion may depend on
structural properties, on molecular aggregation, on the solvent,
and on the normally uncertain microsolvation of the C−Li
centers by donor ligands; Reich et al.1 have published an
instructive short survey of the primary literature.2 For a further
example, the monomolecular3 racemization of chiral benzyl-
lithium derivatives such as 1 (R = CH2Ph3a or Ph3b in Scheme
1) in tetrahydrofuran (THF) solution was thought3a to occur
through heterolytic breaking of the bond between Li and the
quasi-sp3-hybridized carbon atom of a polar ground-state CIP-1
(CIP = contact ion pair carrying d donor ligands “Don” at Li).
The assumed ensuing immobilization of free THF molecules
would generate the more polar solvent-separated ion pair SSIP-
1 with a flattened3b,4 (but not planar) C-α center, followed by
an achiral transition state on the way to the enantiomers ent-
SSIP-1 and finally ent-CIP-1.
The stereoinversion mechanism at the quasi-sp2-hybridized
carbanionic center of lithioalkenes such as α-(1,1,3,3-
tetramethylindan-2-ylidene)benzyllithium (3) is uncertain, but
3 has the most welcome property of being monomeric with
known microsolvation numbers d at Li in preponderantly
ethereal solutions and also in Me2N−CH2CH2−NMe2
(TMEDA),5 whereas it is trimeric (with d = 0) in toluene6
and dimeric (with d = 1) in the presence of small amounts of
ethereal donor ligands in both toluene solution and the
crystals.6 The depicted angular (bent) configuration of
monomeric 3 is stable at sufficiently low temperatures
(exhibiting two sets of six-proton NMR singlet signals for the
four methyl groups, for example). Upon warming in THF
solution, 3 was averaged with the configuration 3′ by trans/cis
In contrast, the bimolecular axial/equatorial diastereoisome-
rization1 of 3eq,5eq-diphenylcyclohexyllithium (2) in THF
requires aggregation (presumably1 dimerization) to occur at
Received: January 30, 2013
Published: July 29, 2013
© 2013 American Chemical Society
4070
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
(5c), an excess of 4 furnished only 29% of the almost pure
chloroalkene 6c. Halodesilylation was the method of choice for
transforming 6a with N-chlorosuccinimide into 6d as the only
product or with bromine in acetic acid into 92% of pure 6e.
Despite steric shielding, the chlorine atoms at C-α of 6a (in t-
BuOMe), 6b (in THF), and 6c (in Et2O) are sufficiently
reactive for a quick Cl/Li interchange reaction with n-
butyllithium (n-BuLi). However, the emerging α-lithioalkenes
8a, 8b, and 8c, respectively (Scheme 3), deteriorated quickly
Scheme 1
Scheme 3
diastereotopomerization7 with a roughly estimated8 preinver-
sion half-life of ca. 0.014 s at −45 °C, as recognized by the
NMR coalescences (one 12-proton Me4 singlet, for instance) of
the signal pairs of diastereotopic protons and 13C nuclei,
because the 3/3′ positional interchange became fast on the
NMR time scale. Having synthesized further α-aryl analogues of
3 with similar quasi-benzyllithium properties, we are now able
to present evidence for the pseudomonomolecular, ionic
inversion mechanism of the quasi-sp2 carbanion derivative 3
in THF, including an elucidation of the transient increase of
microsolvation at lithium.
RESULTS AND DISCUSSION
A. Syntheses. The preparation of 2-(α-chloro-4′-trimethyl-
silylbenzylidene)-1,1,3,3-tetramethylindane (6a in Scheme 2)
■
Scheme 2
through reactions presumably involving the byproduct 1-
chlorobutane; for example, the transient presence of 8c was
established through carboxylation to give the acid 10 in poor
yield. This situation required a detour via the α-(trimethyl-
stannyl) derivatives 7a−e, which were obtained through
treatment of 6a−e with Me3SnLi in THF by the procedure
described previously12 for 7a. The stannanes 7a−d were usually
accompanied by the respective “parent” alkenes 9a−d (yield up
to 38% of 9c from 7c). The above-mentioned enhanced α-Cl
reactivity was again encountered with the α,4′-dichloride 6d,
which furnished 7d with the intact 4′-Cl substituent. However,
the 4′-bromo-α-chloroalkene 6e and Me3SnLi (ca. 3 equiv)
provided 85% of the pure α,4′-bis(trimethylstannyl) product
7e. A NOESY experiment with 7c suggested that the small
119Sn couplings (J ≈ 2.9 Hz) to 3-CH3 of 7a12−e (formally over
five bonds, but not observed to 1-CH3) may be interpreted as
through-space spin−spin coupling.
from the sterically shielded dichloroalkene 4 with p-
(trimethylsilyl)phenyllithium (5a) was published9 as an
example of the carbenoid chain reaction mechanism10 of
unactivated vinylic substitution in the absence of a transition
metal catalyst. Under similar conditions, the addition of solid 4
to a THF solution of 5b, as prepared from p-bromobiphenyl
with tert-butyllithium (t-BuLi, 2 equiv), afforded 66% of the
purified α-chloroalkene 6b. Using the alternative variant11 with
tert-butyl methyl ether (t-BuOMe) replacing THF in
consideration of the higher basicity of p-methylphenyllithium
The α-lithioalkenes 8a−f were preferentially generated from
7a−e through Sn/Li interchange reactions (Scheme 3) with
cyclopentane solutions of labeled n-Bu6Li (ca. 1.07−2.5 equiv)
or with unlabeled n-BuLi in hexanes. This method avoids the
formation of disturbing byproducts such as LiHal or alkyl
4071
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
halides. The various excess amounts of n-BuLi served to
compensate for losses caused by the conversion of byproduct
Me3SnBu with n-BuLi into Me2SnBu2. Purification of 8a−c was
easy because crystalline samples were obtained in NMR tubes
at room temperature through the very slow Sn/Li interchange
in the alkane solutions containing the donor ligands t-BuOMe
(2.5 equiv for 8a or 8b) or, better, Et2O (2−3 equiv for 8a−c).
The dimeric constitution of one of these crystals (8a) had been
established by a single-crystal X-ray diffraction analysis,6 and
the dimeric nature of 8b and 8c followed from the similarity of
their 13C NMR data with those of the Et2O solvate 8a in
[D8]toluene.6 The crystals were washed with dry pentane, dried
by blowing with dry argon gas for five seconds, and dissolved in
anhydrous THF with a little [D8]toluene or [D6]benzene,
forming solutions of monomeric 8a−c. Whereas 8a was stable
up to at least 52 °C in THF, 8c isomerized within ca. one hour
at 53 °C to generate the benzyllithium derivative 11 with its
typical NMR chemical shifts (Tables S24 and S25).13 The 4′-
chloro-α-lithioalkene 8d could not be crystallized and had to be
prepared through Sn/Li interchange directly in THF at −78
°C, because it decomposed within five minutes at room
temperature. The Sn/Li interchange reaction of the α,4′-
distannyl compound 7e in THF was selective: Only the α-
SnMe3 group was removed by one equivalent of n-BuLi to
furnish the α-lithio-4′-SnMe3 product 8e. Generated with
additional n-BuLi (1.5 equiv), the α,4′-dilithio product 8f
exhibited chemical shift values (Table S23)13 very similar to
those of 3 (Tables S11 and S12),13 with four exceptions: The
C-4′ resonance of 3 at δ = 113.8 was missing, and C-3′/C-5′
had changed from δ = 126.3 (in 3) to 142 ppm in the direction
expected for a phenyllithium derivative.14 The resonance of 4′-
H (δ = 6.18 ppm) in 3 had been replaced in 8f by a quasi-AB
spectral system for 2′-/3′-H (and 6′-/5′-H), establishing the
substitution at C-4′. Carboxylation of 8f furnished mainly the
α,4′-dicarboxy derivative in very low yield.
Scheme 4
temperatures by the triplet (1:1:1) splitting of their 13C-α
NMR resonances caused by scalar magnetic coupling with only
one 6Li nucleus (spin quantum number I = 1). The magnitudes
of the one-bond coupling constants 1J(13C,6Li) are all similar, as
shown in the last two columns of Table 1; they indicate
microsolvation by d = 3 THF donor ligands at Li according to
the empirical relationship6 of eq 1, where n = a = 1 are the
numbers of CLi contacts at 13C and 6Li, respectively, for
monomers, and L = 41 ( 3) Hz is a sensitivity parameter valid
for the 1-aryl-1-alkenyllithiums in entries 2−8 of Table 1. Such
trisolvation by THF is most important for the present
1
Although the known11 α-chloro-2′,6′-dimethylalkene 12 did
not react with elemental Li0, n-BuLi, or Me3SnLi in Et2O, it
reacted readily with Me3SnLi (2 or 4 equiv) in THF at room
temperature to afford a 68:32 mixture of the α-stannyl
derivative 13 and its “parent” olefin 15 (Scheme 4). Similarly,
the Sn/Li interchange reaction of 13 with n-BuLi to give 14 was
not possible in Et2O solution at room temperature but occurred
instantly in THF at 3 °C and was performed with n-Bu6Li (1.1
equiv) at −30 °C for preparing the sample of [6Li]14 to be
used in the kinetic experiments in THF. This precaution was
necessary because 14 disappeared within 20 min at room
temperature, forming some parent olefin 15 along with the
benzyllithium derivative 17, which was stable for up to three
days in this milieu. The constitution of 17 followed from the
diminished (upfield) 13C NMR shifts (Tables S28 and S29)13
of C-3′and C-5′ (the o- and p-positions of the benzyllithium
part of 17), in accord with its deuteriolysis to give 16 with the
purposes: Ground states with these JC,Li values and micro-
solvation numbers d = 3 cannot be solvent-separated ion pairs
whose Li+ would be tetrasolvated (coordinatively saturated)
and thus not in a direct contact with a carbanionic 13C nucleus
(so that JC,Li ≈ 0).
d = L × (n × 1JC,Li
)
− a
−1
(1)
The bent (angular) structures of the above 1-aryl-1-
alkenyllithiums (see 3) in solution followed from their NMR
spectra, which established a molecular Cs symmetry (the C-α/
C-2 double-bond plane) of their ground states.6 The α-aryl
groups of 3 and 8a−f (and more so of 14) are rotationally
impeded by the two methyl groups at C-1, so that they adopt a
close to orthogonal conformation with respect to the C-α/C-2
double bond, as confirmed6 by several single-crystal X-ray
diffraction analyses. This conformation ensures an almost
maximum possible overlap of the pz orbital at C-1′ with the
charge-carrying, quasi-sp2 orbital of the Li−C(α) bond (ionic
with a modest covalent contribution), as depicted in formulas 3
and 3′. This overlap leads to delocalization of negative electric
charge into the α-aryl π-system, so that fractions of negative π-
charge are expected to emerge at the 2′/6′ and 4′ positions
(ortho and para, respectively), in analogy with π-conjugation in
1
triplet (1:1:1) H and 13C NMR splittings of its CH2D group.
Rotation by ca. 180° about the C-α/C-1′ single bond in 17 is
slow on the 13C NMR time scale up to at least 25 °C, as
recognized through the detection of five methyl signals instead
of the three resonances (intensities 2:2:1) expected for a rapidly
reversible half-rotation. Therefore, a 6′-methyl group, as the
“smallest” of the 2′-/6′-substituents, suffices to retard the half-
rotation in 17 and hence also in 12−16.
B. Ground States of the 1-Aryl-1-alkenyllithiums of
Type 3. The ground states of 3, 8a−d, 8f, and 14 in THF as
the solvent are monomeric contact ion pairs, as revealed6 at low
−
the benzyl anion (PhCH2 ), whose predominant mesomeric
form, 18alpha, is drawn in Scheme 5 (part a) with positional
numbers corresponding to those of 3: If the pz orbital axes on
the C-α and C-1′ atoms are parallel to each other
(perpendicular to the molecular plane in part b), this provides
4072
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
Table 1. Lithiation NMR Shifts Δδ = δ(α-Li Compound) − δ(α-H Parent) and Coupling Constants 1J(13C,6Li) of Benzyllithium
(Li+&18) and of the Monomeric 1-Aryl-1-alkenyllithiums 3, 8a−d, 8f, and 14 in THF Solution
Δδ (ppm)
1J(13C,6Li)
entry cpd no. aryl substituent temp, °C 2′-/6′-H 3′-/5′-H
4′-H
C-2′/-6′ C-3′/-5′
C-4′
C-1′
C-2
C-α
Hz
T, °C
1
2
3
4
5
6
7
8
Li+&18
(a)
+25
−102
−72
−89
−90
−88
−90
−95
−1.03
−0.74
−0.75
−0.75
−0.69
−0.74
−0.67
−0.86
−0.57
−0.58
−0.38
−0.52
−0.55
−0.42
−0.46
−1.69
−0.98
−12.7
−8.2
−8.3
−9.0
−8.2
−6.4
−8.0
−11.5
−0.4
−1.5
−1.0
+1.4
−21.8
−13.2
−18.1
−14.1
−14.5
−14.5
+22.0
+23.2
+23.0
+22.9
+23.2
+22.8
+15.4
+66.5
+66.1
+66.2
+66.4
+65.6
+65.9
+66.0
3
4′-H
−15.2
−16.1
−16.2
−14.6
−15.6
−15.8
−17.3
10.7
10.0
9.5
≤−37
−72
≤−89
−90
≤−75
−90
−61
8a
4′-SiMe3
4′-C6H5
4′-CH3
4′-Cl
4′-Li
2′-/6′-CH3
8b
8c
−2.4
−1.6
+14.9
−1.3
11.0
10.0
10.2
10.7
8d
8f
14
−0.88
−12.4
+21.0
temperatures (no thermal loss of THF, therefore), because the
corresponding disolvated monomers are known6 to have
significantly smaller Δδ values, which are practically equal for
monomeric 3 in Et2O (d = 2), t-BuOMe (d = 2), and Et2O/
[D8]toluene as the solvents. The remaining Δδ data for the
1,1,3,3-tetramethyl-2-indanylidene skeleton (Table S8)13 verify
the previously20 described conspicuous π-polarization at the
positions C-8 ≠ C-9 (positive) and C-5 ≠ C-6 (negative), as
also illustrated in Chart S2 of the Supporting Information of ref
6. It will now be established that the polar character of these
CIP ground states (“tight ion pairs”) increases on the way to
the transition states of trans/cis interconversion.
Scheme 5. Benzyl Anion Seen (a) from above and (b) from
Close to within the Molecular Plane (with pz Orbitals)
C. Cis/trans Diastereotopomerization of the 1-Aryl-1-
alkenyllithiums of Type 3. The trans/cis sp2-stereoinversion
(such as monomeric 3 → 3′, or 21 → 21′ in Scheme 6) leads to
for the best possible charge delocalization into the phenyl ring,
as visualized by the mesomeric forms 18ortho and 18para.
Effects of such a charge distribution can be observed as
“upfield” (less positive) NMR shifts δ for the ortho and para
positions in the real ion pair benzyllithium (Li+&18) and will
be utilized as follows.
Scheme 6. Proposed sp2-Inversion Mechanism of 21
Looking at benzyllithium (Li+&18) as being derived from
toluene through α-lithiation, we calculated its lithiation NMR
15
1
shifts Δδ = δ(Li+&18) − δ(toluene ) in THF from its H (at
25 °C)16 and 13C NMR15 shifts, as depicted in entry 1 of Table
1. With due reservations, the Δδ values of at least the remote
4′-H and C-4′ positions might be considered to be π-charge
indicators with the approximate conversion factors of −10.7
ppm (1H)17 and roughly −165 ppm (13C)18 per π-electron.
Comparisons of these Δδ values with those for 4′-H and C-4′
of 3 in entry 2 of Table 1 suggest ca. 60% of the benzyllithium
charge delocalization for trisolvated monomeric 3 in THF
solution (“quasi-benzyllithium”). The total set of Δδ data for 3
(Table S8)13 was computed from the δ values (Tables S11 and
S12)13 of 3 at −102 °C with respect to those of the parent
olefin 9f, the temperature-dependent δ values19 of which were
extrapolated to −102 °C in THF. In the same manner,
lithiation shifts Δδ of 8a−d, 8f, and 14 (Table S8)13 were
calculated from δ values in THF (Tables S13−S16, S19−S23,
S26, S27)13 minus those of the corresponding parent olefins
9a−d, 9f, and 15 at the corresponding temperatures in THF.
The results are listed in entries 3−8 of Table 1 and show that
Δδ values are not always suitable indicators of π-charges: As the
most obviously deviating example, Δδ(C-α) is strongly positive
despite the negative charge expected for a carbanionic center.
Nevertheless, the Δδ data of C-1′, C-2, and C-α are particularly
suitable for assigning the alkenyllithiums in entries 2−8 to the
family of trisolvated monomers. In addition, the weak
temperature dependencies of δ (and Δδ) values for C-α, C-2,
and C-1′ (Tables S11−S23, S26, S27)13 establish that 3, 8, and
14 remain the same trisolvated ground states at all investigated
the positional interchange (diastereotopomerization)7 within
pairs of diastereotopic nuclei (1-CH3/3-CH3, 1-CH3/3-CH3, C-
1/C-3, C-8/C-9, etc.). This process may be monomolecular or
bimolecular with respect to the ground-state species, as
exemplified in the Introduction (Scheme 1) for sp3-inversions;
if monomolecular, it might be dissociative (via free carbanion
and free Li+) or occur by intramolecular migration of Li+ in an
ion pair with a temporary abolition of the C−Li contact that
defines the CIP ground state. Rate measurements can provide
4073
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
the keys for eliminating the incorrect ones of such possibilities
as follows.
with R = Ph) and in Figures S1−S6 for the others;13 this
establishes the monomolecular, kinetic first-order role that the
trisolvated monomers 21 play in the stereoinversion rate. Such
insensitivity to concentrations of 21 includes losses of 21
through decomposition, possibly with a concomitant gen-
eration of free Li+(THF)4 that could have reduced the kψ values
through a mass-law suppression of ionic dissociation of 21;
however, such a deceleration could not be inferred from Figures
1 and S1−S6. The conventional extraction of pseudoactivation
The temperature-dependent pseudo-first-order (explained
below) rate constants kψ of the 21/21′ interconversion were
1
measured by H and 13C NMR line shape analyses through
their simulation by a computer program21 for the interchange
of noncoupled nuclear spins. These kψ did not depend on the
concentrations of 21, as exemplified in Figure 1 for 8b (= 21
⧧
⧧
enthalpies ΔHψ and entropies ΔSψ from such kψ diagrams
furnished the results collected in Table 2, where the free
⧧
⧧
pseudoactivation enthalpies ΔGψ (273 K) = ΔHψ − (273 K)
⧧
× ΔSψ are included for comparisons at the common
temperature of 273 K. Evidence for an increasing charge
transfer from C-α into the α-aryl groups on the way from 21 to
a transition state was obtained through the Hammett
correlation (Figure S7)13 of these ΔGψ (273 K) values versus
⧧
−
the p-substituent constants σp , which quantify the interaction
of negative π-charge with the 4′-substituents. (Our choice of
the σp− values is documented in Section S3.)13 The slope of the
regression line in Figure S713 equals −2.303RTρ and leads to
the rather large Hammett reaction constant ρ(273 K) = +5.2
( 0.2); this positive value indicates a transition state that is
energetically stabilized more than ground state 21 (or 21′) by
electron-withdrawing 4′-substituents (SiMe3, Cl, phenyl) or
destabilized by 4′-CH3, so that the diastereotopomerization7
rate constants increase or decrease, respectively. This
establishes the temporary development of considerably more
negative π-charge density in the 4′-position during stereo-
inversion. That property is thought to apply also to 8f (4′-Li)
1
with its similar J(13C,6Li) and pseudoactivation parameters in
entry 6 of Table 2, although the modest acceleration as
compared with 3 remains unexplained in default of a reliable
−
σp value for 4′-Li. The addition of the tridentate amine
PMDTA (pentamethyl diethylenetriamine) to 8f caused only a
roughly 2-fold decrease of the rate constant.
D. Mechanistic Deductions. With an increasing transfer of
electron density from C-α to the α-aryl group on the way to the
transition state experimentally settled (Hammett ρ = +5.2) for
3 and 8a−d, an increasing separation of positive from negative
electric charges appears established as an essential feature of the
Figure 1. Arrhenius diagram of the natural logarithms of pseudo-first
order sp2-stereoinversion rate constants kψ [s−1] versus 1000/T [K−1]
for 8b in THF solution. Concentrations: open symbols, 0.3 M;
hatched, 0.1 M; filled, 0.02 M.
Table 2. Pseudoactivation Parameters ΔGψ⧧ (kcal mol−1 at 0 °C), ΔHψ⧧ (kcal mol−1), and ΔSψ⧧ (cal mol−1 K−1) of Cis/Trans
Diastereotopomerization Rates of the Monomeric 1-Aryl-1-alkenyllithiums 21 (= 3, 8a−d, 8f, or 14) in THF, Depending on
a
−
Available Hammett Parameters σp for Conjugating 4′-Substituents
1J(13C,6Li), Hz
σp
⧧
⧧
⧧
−
entry
1
cpd no.
aryl substituent
ΔGψ (0 °C)
ΔHψ
ΔSψ
3
4′-H
13.35
0.03
6.63
0.24
5.83
0.06
5.81
0.05
8.56
0.16
5.90
0.09
6.09
0.19
6.77
0.18
−24.6
1.0
10.7
0.0
2
3
4
5
6
7
8a
8b
8c
8d
8f
4′-SiMe3
4′-C6H5
4′-CH3
4′-Cl
12.40
0.01
−23.9
0.3
10.0
9.5
+0.18
0.01
11.41
0.01
−20.5
0.2
+0.28
0.02
14.44
0.01
−21.5
0.5
11.0
10.0
10.2
10.7
−0.17
12.26
0.02
−23.3
0.4
+0.19
4′-Li
13.02
0.01
−25.4
0.7
14
2′-/6′-CH3
12.47
0.01
−20.8
0.7
a
The averaged “true” activation entropy would be ΔS⧧(av) ≈ ΔSψ⧧(av) − R ln[free THF] ≈ − 22.9 − 5.1 = −28.0 cal mol−1 K−1.
4074
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
sp2-carbanionic stereoinversion mechanism. A straightforward
interpretation infers Li−C bond heterolysis as the causal
process, which is usually induced and supported by the
immobilization of THF molecules at Li+ in a solvent-separated
ion pair such as 19 (SSIP in Scheme 6). As 19 maintains a short
distance of cation and anion (without direct contact of Li to C-
α), which precludes the stereoinversion, that distance must be
modified within the ion pair 19 in a subsequent step toward a
more charge-delocalized transition state, because the con-
version of 19 into 19′ requires lithium migration. A sufficiently
fast C-α/C-1′ half-rotation with Li riding on one of the two π-
faces of α-aryl in 19 appears unlikely, recalling that such a half-
rotation in the ground state 14 had been concluded in Section
A (see 17) to occur much more slowly than would be visible on
the NMR time scales. One might be tempted to assume more
rotational freedom (less hindrance by the 1-CH3 groups) for α-
aryl on rehybridization of C-α to a roughly linear (sp-
hybridized) arrangement, as shown in the ion pair 20, which
is attained through a lateral α-aryl motion that appears
reasonable on account of steric congestion and energetic
resistance against torsion of the C-2/C-α double bond.
However, this arrangement implies a more developed quasi-
benzyl carbanion π-system in 20 with parallel pz orbital axes
(perpendicular to the α-aryl plane) and a correspondingly
enhanced charge delocalization away from C-α toward C-4′ as
compared with the less charge-delocalized (because bent)
quasi-benzyl carbanion parts in 19 and 21 (hence the positive
Hammett ρ value). Such an increased benzyl-type carbanion
resonance and its energetic lowering of 20 would be sacrificed
on C-2/C-α half-rotation, so that a substantial energetic barrier
against that half-rotation should be expected. Of course, the
latter consideration would also argue against 1-aryl rotation in
less congested transition states of 1-aryl-1-alkenyllithiums
without bulky substituents at C-2.
The inferred participation of THF shown in 19 (SSIP) and
20 (SSIP) means that we have been measuring pseudo-first-
order rate constants kψ that contain the concentration of Don =
THF in eq 2 in an experimentally unconfirmed kinetic order
(m) of reaction, but the obvious experimental tests could not be
carried out, because significant (at least 2-fold) diminutions of
[free THF] would be accompanied by serious changes of the
solvent polarity and hence would furnish questionable evidence.
Therefore, we leave the constant THF concentration latent in
kψ, conceding that the “true” ΔS⧧ values (as derived from
unconfirmed “true” rate constants) would be more negative
than our ΔSψ (average ca. −22.9 cal mol−1 K−1 in Table 2
⧧
from kψ data) by a mathematical correction of roughly23 R ln
[THF]m = ca. 5 × m cal mol−1 K−1 (where R = 1.986 cal mol−1
K−1), so that ΔS⧧(average) = ca. −28 cal mol−1 K−1 if m = 1. If
desired, transformations of the pseudoactivation parameters of
3, 8a−f, and 14 in Table 2 into the “true” values ΔG⧧, ΔH⧧,
and ΔS⧧ would be simple: In the case of m = 1 (eq 2) as the
kinetic order of Don = THF as the solvent, the primary kψ
values13 may be divided by the practically constant solvent
concentration of [free THF] (up to 13 M); this affords the
second-order k0 data to be used for the extraction of “true”
ΔH⧧ and ΔS⧧. For the example of 8b, this would provide ΔS⧧
= −25.6 (in place of ΔSψ = −20.5) cal mol−1 K−1 and an
⧧
unchanged ΔH⧧ = 5.81 kcal mol−1 = ΔHψ , so that the
⧧
expression R ln[free THF] = R ln(kψ/k0) = −(ΔHψ⧧ − ΔH⧧)/
T + ΔSψ − ΔS⧧ (as derived from eq 2) simplifies to ΔS⧧ =
⧧
⧧
ΔSψ − R ln[free THF]. In fact, there can be little doubt that
the pseudo-first-order stereoinversion rate constants kψ (eq 2),
as measured in Section C, must contain the concentration of
Don = THF in a first-order (m = 1) of reaction: As the ground
states 21 are already trisolvated by THF,6 only one further
THF ligand can be expected to become immobilized in 19
(SSIP), because a fifth THF molecule would bind only very
weakly. Consequently, one independent particle (19) is formed
from two (21 plus THF), which should contribute23 roughly
What role may be expected for Li+(THF)4 during the
stereoinversion? Ionic dissociation of 19 would leave the free
carbanion 22 (Scheme 6) behind, whose mechanistic
participation would be recognized through a faster sp2-
stereoinversion in a more diluted solution because of an
increased dissociation; however, the dilution-independent kψ
data in Section C (Figure 1) excluded this possibility. Entries 1
and 7 of Table 2 show that 3 and 14 have almost the same
−11 cal mol−1 K−1 in ΔSψ , whereas the experimental ΔSψ
⧧
⧧
values in Table 2 are clustered at about −22.9 cal mol−1 K−1;
this difference confirms the proposal in Scheme 6 that the
stereoinversion rate is not determined by the formation of 19
alone but is co-determined by a subsequent transition state
(20). Thus, this process represents a pseudomonomolecular
mechanism because it is monomolecular with respect to the
ground-state species 21 but also “catalyzed” by THF as the
solvent.
⧧
pseudoactivation enthalpies ΔHψ , which suggests that the 2′-/
6′-methyls in 14 do not significantly aggravate the circum-
navigation of Li+(THF)4 about the anion and that the
increasing negative π-charge in position 4′ is more important
for an energetic lowering of the transition state than the π-
charges in the 2′- and 6′-positions are. This may be so if
Li+(THF)4 migrates along the charge gradient (C-3′/-5′ → C-
4′) toward closer to the 4′-position, as proposed in formula 20
(SSIP), where the indicated motion of Li+(THF)4 may occur
via a Cs-symmetric structure that contains the α-aryl ring in the
plane of symmetry. The term “conducted tour” mechanism had
been coined by D. J. Cram22 for a corresponding cation
migration about a chiral carbanion, because that process
occurred with isotope-reported cation retention during the
enantiomerization (that is, sp3-stereoinversion) of the carban-
ion. In our system, the conducted tourist Li+(THF)4 in 20 may
be envisioned as being steered across the α-aryl edge by the
charge gradient as a guide.
A final look at the above entropy difference of ca. −22.9 −
(−11) ≈ −12 cal mol−1 K−1 is instructive: That entropic
penalty may have to do more with Li+(THF)4 migration than
with the lateral motion (19 → 20) of α-aryl in Scheme 6. This
expectation is based on observations24,25 of practically zero
activation entropies13 for the sp2-stereoinversion of the lone
electron pair at nitrogen in the Schiff base families 23 and 24
(Scheme 7). If the ground states and diastereotopomerizations7
of 23 and 23′ can be taken as isoelectronic models for the
carbanion parts of 19 and 19′ in the absence of lithium, it may
be concluded that the lateral motion of α-aryl alone toward sp-
hybridization as in 20 would not necessarily create a strong
contribution to the activation entropy. With regard to ΔH⧧,
that lateral motion may be expected to release the steric strain13
inherent in 19 (where α-aryl is repelled by two 1-CH3) to a
similar degree to that determined24 for 23 (namely, by −4.7
kcal/mol).
kψ = k0[free Don]m
(2)
4075
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
CLi−Ph was surprisingly soluble in hexane/pentane solution
and rearranged in this milieu quantitatively to the insoluble E
isomer within ca. four hours at room temperature. A
mechanistic stereoinversion study with such less congested
examples will meet the problem that their ground states are
certainly aggregated rather than monomeric, judging from the
observations6,23 of one trimeric and several dimeric forms of
our more congested model compound 3.
Scheme 7
The spatial limitations at Li in monomeric 3 had been
proposed6 to explain why trisolvation is unattained in Et2O, t-
BuOMe, or TMEDA as the solvents. Because this micro-
solvation privilege (d = 3) of THF turned up here for the rates
of stereoinversion, we sought independent confirmations of
that proposal6 through examination of the mobility about the
C(O) carbon center within the ketone adducts 25−27 (Scheme
8) of 3. While t-Bu2CO did not form a covalently bound
E. The Privilege of THF. In accord with the inferred
formation (Scheme 6) of NMR-invisible amounts of 19 (SSIP)
from 21 (CIP), the high concentration of [free Don] of the
solvent THF is essential for the high configurational lability of
our monomeric 1-aryl-1-alkenyllithiums of type 3: The same
trisolvated ground states 21 did not show any indication of
stereoinversion on the NMR time scales in hot toluene, where
the concentration of [free THF] was diminished by ca. 2 orders
of magnitude. (We cannot separate this effect from the
supposed retardation caused by the nonpolar character of
toluene.) Likewise, a mixture (0.12 M) of monomeric 3 and its
dimer in [D8]toluene with THF (0.57 M) showed no hint of
cis/trans diastereotopomerization7 up to 90 °C: from the
smallest preserved NMR frequency difference (16 Hz for the
Scheme 8
⧧
diastereotopic pair C-4/C-7), we estimated ΔGψ (90 °C) ≫
19.0 kcal mol−1 in this solution, to be compared with ΔGψ (90
⧧
°C) = 15.6 kcal mol−1 for 3 in THF as the solvent (calculated
from entry 1 of Table 2).
The kinetic privilege of THF is evident on comparison with
other donor solvents: The cis/trans diastereotopomerization7
was also slower (albeit not quantified) for the disolvated6
monomers 3&2Don in the solvent Et2O, t-BuOMe, or TMEDA
adduct,6 the more compact adamantan-2-one added readily to
give 25. NMR spectra of 25 at 25 °C showed apparent Cs
symmetry (the C-α/C-2 double-bond plane), but hindered
rotation about the C(α)−C(O) single bond became evident at
−63 °C in CDCl3 solution. The resultant asymmetric
conformation was recognized through unequal chemical shifts
for the two 3-CH3 groups and all diastereotopic 13C nuclei: C-
2′ ≠ C-6′, C-3′ ≠ C-5′, C-1″ ≠ C-3″, C-4″ ≠ C-9″, C-8″ ≠ C-
10″, two 1-CH3, and two 3-CH3. This asymmetry indicates that
the volume of 2-adamantyl is sufficiently small to fit in one half-
space (below or above the C-α/C-2 double-bond plane) that
would be available for donor coordination at Li+ in 3. No such
rotational deceleration was observed for the less congested
adducts of dicyclopropyl ketone (27b, the smaller cousin of
25), fluorenone (26 at −60 °C), and benzophenone (27a).
(The attempted addition of 3 to diisopropyl ketone was slow
enough to be suppressed by a faster proton transfer to 3 with
enolate formation.) Thus, the hindered rotation in front of the
plane C-1′/C-2′/C-3′ in 25 is thought to confirm the sufficient
but not excessive availability of space for the corresponding
fragments O−Li−O of two Et2O or two t-BuOMe ligands in
the two half-spaces of monomeric 3, whereas trisolvation would
be a privilege of the more compact donor THF.
⧧
(experimental estimate ΔGψ > 17.1 kcal mol−1 at +52 °C in
Et2O). The ionic mechanism in Scheme 6 would now require
the immobilization of m = 2 additional Et2O or t-BuOMe
donor ligands for arriving at a possibly tetrasolvated transition
state [20 with (Don)4 in place of (THF)4]. This implies a more
negative activation entropy (by another ca. −11 cal mol−1 K−1
relative to Don = THF with m = 1)23 whose decelerating effect
appears to be incompletely balanced by a presumably (because
of m = 2) lowered activation enthalpy. Such an entropic
handicap (for m = 2) would get exacerbated in toluene as the
solvent by the squared concentration factor [free Don]2 in eq 2,
where a ca. 100-fold diminished concentration of [free Don]
might entail a 104-fold rate reduction for the disolvated
monomer 3&2Don. Under such strongly adverse conditions,
the pseudomonomolecular ionic mechanism (Scheme 6) might
perhaps be outrun on a route involving more highly aggregated
intermediates, as supposed1 for 2. However, heating our
disolvated dimers (d = 1)6 in toluene provided no evidence of
such a stereolability.
The microsolvation privilege of THF has not yet been
cleared up for less congested 1-aryl-1-alkenyllithiums whose
cis/trans sp2-stereoinversion was observed on the NMR time
scale8 at 60 MHz in TMEDA solution and on the laboratory
time scale with Et2O in hydrocarbons,26,27 while the
appertaining mechanisms have remained uncertain. Even the
absence of any donor ligands (which precludes the ionic SSIP
mechanism of Scheme 6) did not prevent the slow Z/E
stereoinversion of Ph−CHCLi−Ph in neat benzene26 at 27
°C. Likewise, the unsolvated, pure Z isomer28 of Ph−CMe
CONCLUSION
■
Our deduction of the ionic mechanism of stereoinversion is
based on the following considerations: (i) The pseudo-first-
order rate constants kψ are independent of the concentrations
and decomposition products of the monomeric substrates
(hence no dimeric transition states, no free carbanions as
intermediates); the stereoinversion is monomolecular with
4076
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
constant. Natural logarithms (ln) are understood to be calculated with
the dimensionless magnitudes of the employed quantities.
respect to the ground-state species and neither inhibited nor
catalyzed by contaminating Li+. (ii) The positive Hammett
reaction constant ρ = +5.2 establishes a significantly higher
polarity of the transition states than of the ground states. This
points to Li−C bond heterolysis with an increasing charge
separation in a THF-separated ion pair (SSIP) and in an
ensuing transition state. The required immobilization of THF
will contribute to the strongly negative pseudoactivation
2-(α-[6Li]Lithiobenzylidene)-1,1,3,3-tetramethylindane (3).
The donor-free, crystalline cyclotrimer of 3 (compound 8 in ref 6)
was prepared as described on pp 14187 and S21 of ref 6. Dissolved in
anhydrous THF, 3 is purely monomeric (<0.1% of dimer between
−102 and +100 °C)30 and configurationally labile on our NMR time
scales above −37 °C (Tables S11 and S12).13 Because the fast
diastereotopomerization7 does not affect the NMR parameters of the
negatively charged α-phenyl substituent, a [D8]THF solution of 3 at
+27 °C was used for determining accurate scalar interproton coupling
constants in the α-phenyl group through computer simulation of the
⧧
entropies ΔSψ (clustered about −22.9 cal mol−1 K−1 for all
seven compounds). (iii) An essential piece of evidence became
available through the empirical relationship6 (eq 1 in Section B)
of NMR coupling constants 1J(13C,Li) with microsolvation
numbers d at Li, which revealed d = 3 THF ligands solvating
the monomers of 3, 8a−d, 8f, and 14. With this microsolvation
privilege of THF (as distinguished from d = 2 for t-BuOMe,
Et2O, and TMEDA), the trisolvated ground states 21 need
immobilize only one more (the fourth) equivalent of THF at Li
for generating the SSIP 19 (or 19′ in the reverse direction), so
that the stereoinversion rate should be of first order in both
THF and the substrate. Besides the negative portion (perhaps
ca. 40%) of the mathematically corrected ΔS⧧ that is due to
immobilization of that fourth THF ligand in the SSIP 19, an
additional entropic price of undetailed composition has to be
paid for arrival at transition state 20. (iv) α-Aryl rotation as a
means for the cis/trans migration of Li+(THF)4 during
stereoinversion is unlikely in the neighborhood of the transition
states 20. Blessed with this series (i−iv) of favorable properties,
our family of model compounds 14 and 21 turned out to be
unusually suitable for eliminating incorrect alternative mecha-
nistic possibilities.
The presently unveiled roles of both nonaggregation and
microsolvation could not be explored in the earlier,
prophetic27,29 proposals of a linear (sp-hybridized) transition
state in the spirit of 20. At those times, when the
microsolvation numbers of d = 3 THF ligands for the
monomeric ground states 21 were unknown, an erroneously
guessed d = 2 might have led us8 here to interpret the measured
ΔSψ⧧ values as containing the entropic effect of immobilization
of two THF ligands instead of only one. Consequently, a major
3
3
higher-order spectrum measured at 60 MHz: J(2′/3′) = J(6′/5′) =
7.90(3) Hz, 3J(3′/4′) = 3J(5′/4′) = 7.30(2) Hz, 4J(2′/4′) = 4J(6′/4′) =
1.26(2) Hz, 4J(2′/6′) = 1.77(4) Hz, 4J(3′/5′) = 1.51(4) Hz, 5J(2′/5′)
= 5J(6′/3′) = 0.47(3) Hz. Comparison of these two 3J values with 3J =
7.5 Hz for benzene suggests a substantial contribution of a para-
quinonoid mesomeric formulation (analogous to 18para in Scheme 5)
to the ground state of 3. As already reported for 3 in t-BuOMe
solution,6 3 is stable toward di-tert-butyl ketone at 25 °C also in THF
and in Et2O without any changes of the δ values of 3.
2-(α-Chloro-4′-trimethylsilylbenzylidene)-1,1,3,3-tetrame-
thylindane (6a). Published as compound 4m in ref 10.31
2-(α-Chloro-4′-phenylbenzylidene)-1,1,3,3-tetramethylin-
dane (6b). See ref 32. Prepared in analogy with 6a9 from commercial
p-bromobiphenyl (1.83 g, 7.85 mmol) in THF (15 mL) with t-BuLi
(16.3 mmol) in pentane (9.60 mL) and subsequent addition of
dichloride 4 (1.53 g, 6.00 mmol) to yield moderately contaminated,
solid 6b (1.49 g, 67%). The cottony needles of pure 6b had mp 197−
198.5 °C (ethanol); 1H NMR (400 MHz, CDCl3) δ 1.23 (s, 6 H, 2 ×
3
1-CH3), 1.78 (s, 6 H, 2 × 3-CH3), 7.05 (dm, J = ca. 7.5 Hz, 1 H, 7-
H), 7.21 (dm, 1 H, 4-H), ca. 7.25 (m, 2 H, 5-/6-H), 7.37 (tt, 3J = 7.4
4
3
Hz, J = ca. 1.3 Hz, 1 H, p′-H), 7.43 (d, J = 8.2 Hz, 2 H, 2′-/6′-H),
7.46 (tm, 3J = 7.5 Hz, 2 H, 2 × m′-H), 7.62 (d, 3J = 8.2 Hz, 2 H, 2 ×
o′-H), 7.64 (tm, 3J = ca. 8 Hz, 2 H, 3′-/5′-H), assigned through
comparison with compound 4k in Scheme 731 of ref 10 and on p S8 of
the Supporting Information therein; 13C NMR (100.6 MHz, CDCl3) δ
28.0 (qq, 1J = 127 Hz, 3J = 4.5 Hz, 2 × 3-CH3), 31.3 (qq, 1J = 127 Hz,
3J = 4.5 Hz, 2 × 1-CH3), 49.1 (unresolved, C-1), 50.0 (unresolved, C-
3), 122.2 (dd, 1J = 157 Hz, 3J = 7.7 Hz, C-7), 122.5 (dd, 1J = 157 Hz,
1
1
3J = 7.7 Hz, C-4), 126.7 (dd, J = 161 Hz, C-3′/-5′), 127.1 (dm, J =
ca. 160 Hz, 2 × C-o′), 127.20 (overlaid dm, C-6), 127.28 (overlaid dm,
1J = ca. 159 Hz, C-5), 127.32 (partially hidden t, 3J = 5 Hz, C-α), 127.6
(dt, 1J = 161 Hz, 3J = 7.5 Hz, C-p′), 128.8 (dd, 1J = ca. 160 Hz, 2 × C-
m′), 130.5 (dd, 1J = 164 Hz, C-2′/6′), 139.5 (t, 3J = 8 Hz, C-1′), 140.5
and 140.9 (2 broad m, C-4′ and C-i′), 149.1 (unresolved, C-8), 149.7
(unresolved, C-9), 155.5 (unresolved, C-2); IR (KBr) ν 2961, 2865,
1595, 1485, 1458, 1362, and 754 cm−1. Anal. Calcd for C26H25Cl
(372.9): C, 83.74; H, 6.76; Cl, 9.51. Found: C, 84.03; H, 6.71; Cl 9.22.
2-(α-Chloro-4′-methylbenzylidene)-1,1,3,3-tetramethylin-
dane (6c). Applying the more basic LiC6H4-p-CH3 (5c) at first in
excess over the dichloride 4 (as usual in the carbenoid chain
protocol)11 furnished 6c together with large amounts of the parent
olefin 9c, because the excess portion of 5c dechlorinated 6c too rapidly
with formation of ClC6H4-p-CH3 and the alkenyllithium 8c, which
finally became hydrolyzed to give 9c. Therefore, the 5c/4 ratio had to
be inverted, and the slower variant11 in the solvent t-BuOMe was
chosen: Generated from p-bromotoluene (0.97 mL, 7.84 mmol) in t-
BuOMe (10.0 mL) at −70 °C with t-BuLi (15.7 mmol) in pentane
(9.22 mL), 5c exhibited δH = 2.19 (CH3), 6.93 and 7.96 ppm (2 + 2
H, pseudo-AB system with 3J = 7 Hz) and was stable up to at least 40
°C. Dichloride 4 (3.000 g, 11.76 mmol) in t-BuOMe (15 mL) was
added, and the mixture warmed to 35−40 °C for four hours (or at
least two hours). The crude sample obtained from the usual workup11
contained no olefin 9c; its chromatography over a column (diameter 2
cm) of silica gel (31 g, 63−200 μm) with dry petroleum ether (550
mL) afforded (after foreruns with 4) almost pure 6c (695 mg, 29%).
23
⧧
fraction (ca. 2 × −11 cal mol−1 K−1) of ΔSψ ≈ −22.9 cal
mol−1 K−1 (and also of ΔS⧧) would have been wrongly ascribed
to THF immobilization, and the experimental ΔHψ = ΔH⧧
⧧
might thus have been interpreted as being composed of a 2-fold
energetic lowering of 19 (SSIP) by THF and a correspondingly
increased enthalpic ascent from 19 to 20. The lesson is that
⧧
ΔSψ values alone should not be used unconditionally for
deriving solvation changes toward the transition state.
EXPERIMENTAL SECTION
■
Rate Measurements. The NMR tube (5 mm) containing an
alkenyllithium compound in a nondeuteriated solvent (with TMS and
ca. 0.06 mL of [D8]toluene or [D6]benzene added) was tightly sealed
and stored at −18 °C in a large Schlenk tube filled with dry argon gas.
The actual spectrometer temperatures before and after measurement
were determined with the customary methanol or ethylene glycol
NMR tubes in place of the sample. The experimental rate constants kψ
and their error limits were obtained from the H and 13C NMR line
1
shapes in strongly expanded regions of the spectra through visual
comparison with computed line shapes, using the theory21 of simple
A/B positional interconversion for noncoupled singlets A and B (1-/3-
CH3 and several pairs of diastereotopic, proton-decoupled 13C nuclei).
The most reliable kψ values obtained at a certain temperature from the
various diastereotopic pairs were averaged to give the final rate
1
The pure needles 6c had mp 102−103.5 °C; H NMR (400 MHz,
CDCl3) δ 1.18 (s, 6 H, 2 × 1-CH3), 1.75 (s, 6 H, 2 × 3-CH3), 2.39 (s,
3 H, 4′-CH3), 7.03 (dm, 3J = 7.6 Hz, 1 H, 7-H), 7.18 (overlaid d, 3′-/
4077
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
1
3
5′-H), 7.20−7.26 (m, 5 H, 4-/5-/6-H and 2′-/6′-H); 13C NMR (100.6
MHz, CDCl3) δ 21.3 (qt, 1J = 127 Hz, 3J = 4.3 Hz, 4′-CH3), 28.0 (qq,
satellites, J = 441 Hz; C-α), 140.9 (m, apparent J = ca. 3 Hz, C-i′),
3 2
144.7 (sharp t, J = 7.5 Hz; 119Sn satellites, J = 20.5 Hz; C-1′), 149.9
(unresolved, C-9), 150.3 (unresolved, C-8), 167.7 (m; 119Sn satellites,
2J = 16.0 Hz; C-2), assigned through the 119Sn satellites, comparisons
1
1J = 127 Hz, 3J = 4.3 Hz, 2 × 3-CH3), 31.2 (qq, J = 127 Hz, 3J = 4.3
1
Hz, 2 × 1-CH3), 49.1 (m, C-1), 49.9 (m, C-3), 122.2 (dm, J = 157
Hz, C-7), 122.5 (dm, 1J = 157 Hz, C-4), 127.1 (dd, 1J = 160 Hz, C-6),
with 7c and the olefin 9b, and selective {1H} decoupling as follows:
{m′-H} → C-m′ simplified and C-i′ as a broadened s, {3′-/5′-H} → C-
1′ as a s, C-i′ and C-3′/5′ simplified, {o′-H} → C-4′ as a t 7.2 Hz; IR
(KBr) ν 2986, 2960, 2922, 1608 (w), 1484, 762 (s), 735, 698, and 528
cm−1. Anal. Calcd for C29H34Sn (501.3): C, 69.48; H, 6.84. Found: C,
69.64; H, 6.77.
1
1
127.2 (dd, J = 160 Hz, C-5), 127.8 (overlaid, C-α), 128.6 (dqi, J =
3
1
159 Hz, J = ca. 5 Hz, C-3′/5′), 129.9 (dd, J = 160 Hz, C-2′/6′),
3
3
137.8 (t, J = 7.5 Hz, C-4′), 137.9 (t, apparent J = 6.3 Hz, C-1′),
149.2 (m, C-8), 149.7 (m, C-9), 155.1 (>sept, 3J = ca. 3.4 Hz, C-2); IR
(KBr) ν 2988, 2961, 2924, 2864, 1650 (w), 1591 (w), 1509, 1484,
1456, 1362, 807, and 758 cm−1. Anal. Calcd for C21H23Cl (310.9): C,
81.14; H, 7.46; Cl, 11.40. Found: C, 81.54; H, 7.53; Cl 11.21.
2-(α,4′-Dichlorobenzylidene)-1,1,3,3-tetramethylindane
2-(4′-Methyl-α-trimethylstannylbenzylidene)-1,1,3,3-tetra-
methylindane (7c). This was prepared in analogy with 7a12 from 6c
(250 mg, 0.80 mmol) with Me3SnLi (ca. 2.0 mmol) in anhydrous THF
(5 mL) first at room temperature for five minutes and thereupon in an
ice-bath for 60 minutes. Aqueous workup6 furnished a 3:2 mixture
(358 mg) of 7c and its parent olefin 9c. Crystallization from ethanol
provided thin, cottony needles of 7c (183 mg, 52%) with mp 65−68
3
(6d). p-Chlorophenyllithium (pseudo-AB proton system with J = 7
Hz at δH = 7.07 and 7.96 ppm) is fairly stable in t-BuOMe at 40 °C but
does not work well as the initiating base and nucleophile in the
carbenoid chain reaction. Therefore, the α-chloro-4′-silyl compound
6a (500 mg, 1.35 mmol) and N-chlorosuccinimide (434 mg, 3.25
mmol) in glacial acetic acid (8 mL) were stirred at 80 °C overnight.
The mixture was diluted with distilled water (20 mL) and extracted
with Et2O (4 × 20 mL). The combined extracts were washed with 2 M
NaOH (2 × 20 mL) and then with water (20 mL), dried over MgSO4,
and concentrated to yield a crude material containing almost only 6d.
Crystallizations from EtOH yielded pure 6d (270 mg, 60%), mp
1
°C; recrystallization gave analytically pure 7c, mp 73−74.5 °C; H
NMR (400 MHz, CDCl3) δ 0.04 (s, 9 H; 119Sn satellites, 2J = 51.5 Hz;
SnMe3), 1.20 (s, 6 H, 2 × 1-CH3), 1.51 (s, 6 H; 119Sn satellites, 5J = 3
Hz; 2 × 3-CH3), 2.34 (s, 3 H, 4′-CH3), 6.85 (dm, 3J = 8 Hz, 2 H, 2′-/
6′-H), 7.04 (dm, 1 H, 7-H), 7.06 (broadened d, 3J = 8 Hz, 2 H, 3′-/5′-
H), 7.16 (m, 1 H, 4-H), 7.20 (m, 2 H, 5-/6-H), assigned through
NOESY interactions (see below), 119Sn satellites, and comparison with
7b; 13C NMR (100.6 MHz, CDCl3) δ −5.1 (qm, 1J = 128.6 Hz; 119Sn
satellites, 1J = 337.5 Hz; SnMe3), 21.1 (qt, 1J = 126 Hz, 3J = 4.7 Hz, 4′-
CH3), 31.80 and 31.85 (2 qq, 1J = 127 Hz, 3J = 4.5 Hz, 2 × 1-CH3 + 2
1
139.5−141 °C; H NMR (400 MHz, CDCl3) δ 1.18 (s, 6 H, 2 × 1-
CH3), 1.74 (s, 6 H, 2 × 3-CH3), 7.04 (dm, 3J = 7.5 Hz, 1 H, 7-H), 7.20
(dm, 1 H, 4-H), ca. 7.24 (m, 2 H, 5-/6-H), 7.29 (dm, 3J = 8.5 Hz, 2 H,
3
3
3′-/5′-H), 7.36 (dm, J = 8.5 Hz, 2 H, 2′-/6′-H), assigned through
× 3-CH3), 48.3 (m; 119Sn satellites, J(cis) = 22 Hz; C-3), 50.5 (m;
comparison with chlorobenzene; 13C NMR (100.6 MHz, CDCl3) δ
119Sn satellites, J(trans) = 63 Hz; C-1), 122.23 (dm, J = 156 Hz, C-
3
1
1
3
1
1
1
27.9 (qq, J = 127.5 Hz, J = 4 Hz, 2 × 3-CH3), 31.2 (qq, J = 127.5
Hz, 3J = 4 Hz, 2 × 1-CH3), 49.1 (m, C-1), 50.0 (m, C-3), 122.2 (dd, 1J
= 158 Hz, 3J = 7.5 Hz, C-7), 122.5 (dd, 1J = ca. 156 Hz, 3J = 7.5 Hz, C-
4), 126.3 (t, 3J = 4 Hz, C-α), 127.3 (dm, 1J = 160 Hz, C-6), 127.4 (dm,
7), 122.32 (dm, J = 156 Hz, C-4), 126.73 (dd, J = 159 Hz, C-5),
126.88 (dd, 1J = 159 Hz, C-6), 128.0 (dm, 1J = 156 Hz; 119Sn satellites,
4J = 9.6 Hz; C-3′/-5′), 128.3 (dm, J = 156 Hz; 119Sn satellites, J =
18.8 Hz; C-2′/6′), 133.9 (pseudosextet, apparent J = 6.6 Hz; 119Sn
satellites, 5J = 12 Hz; C-4′), 138.9 (unresolved; 119Sn satellites, 1J = ca.
460 Hz; C-α), 142.3 (sharp t, 3J = 7.8 Hz; 119Sn satellites, 2J = 20.4 Hz;
C-1′), 150.0 (unresolved, C-9), 150.4 (unresolved, C-8), 167.5
1
3
1J = 160 Hz, C-5), 128.3 (dd, J = 166 Hz, J = ca. 5 Hz, C-3′/5′),
1
3
131.5 (dd, J = 163 Hz, 3J = 7.2 Hz, C-2′/6′), 134.1 (tt, J = 10.1 Hz,
2J = ca. 3.0 Hz, C-4′), 139.0 (sharp t, 3J = 7.6 Hz, C-1′), 148.8 (m, C-
8), 149.5 (m, C-9), 156.2 (m, C-2), assigned through comparison of δ
1
3
(unresolved; 119Sn satellites, J = 18 Hz; C-2), assigned as above; IR
2
1
and JCH values with those of chlorobenzene; IR (KBr) ν 2987, 2962,
(KBr) ν 3018, 2983, 2957, 2922, 2862, 1615 (w), 1502, 1485, 1454,
1360, 1110, 1026, 798, 771, and 755 cm−1. Anal. Calcd for C24H32Sn
(439.2): C, 65.63; H, 7.34. Found: C, 65.79; H, 7.67.
2928, 1652 (w), 1594 (w), 1485, 1362, 1090, 814, and 760 cm−1. Anal.
Calcd for C20H20Cl2 (331.3): C, 72.51; H, 6.09; Cl, 21.40. Found: C,
72.08; H, 5.82; Cl 21.64.
The above 119Sn/1H scalar coupling (3 Hz) was observed only for
the 3-CH3 protons in the cis position relative to SnMe3, whereas the 1-
CH3 (trans) protons exhibited no 119Sn satellites. Therefore, an
alternative to the tentative proposal of coupling through five bonds
(5J) may pose this 3 Hz splitting as a case of through-space coupling,
in accord with the close spatial proximity of the SnMe3 and 3-CH3
protons as established in the following NOESY interactions: Sn(CH3)3
↔ 3-CH3 ↔ 4-H, and 7-H ↔ 1-CH3 ↔ 2′-/6′-H ↔ 3′-/5′-H ↔ 4′-
CH3.
2-(4′-Bromo-α-chlorobenzylidene)-1,1,3,3-tetramethylin-
dane (6e). See ref 13.
2-(4′-Trimethylsilyl-α-trimethylstannylbenzylidene)-1,1,3,3-
tetramethylindane (7a). See refs 12 and 13
2-(4′-Phenyl-α-trimethylstannylbenzylidene)-1,1,3,3-tetra-
methylindane (7b). See ref 32. This was prepared in analogy with
7a12 from 6b (672 mg, 1.80 mmol) with Me3SnLi (ca. 3.6 mmol) in
anhydrous THF (6 mL). Aqueous workup6 after 10 min at room
temperature afforded an oil (976 mg), which crystallized from EtOH
2-(4′-Chloro-α-trimethylstannylbenzylidene)-1,1,3,3-tetra-
methylindane (7d). This was prepared in analogy with 7a12 from the
α,4′-dichloride 6d (66 mg, 0.20 mmol) with Me3SnLi (ca. 0.4 mmol)
in anhydrous THF (0.7 mL). Aqueous workup6 after five minutes at
room temperature afforded an oily 4:1 mixture (81 mg) of 7d and its
parent olefin 9d. Crystallization from a little ethanol furnished
spectroscopically pure 7d (61 mg, 66%). The analytic sample had mp
1
to give pure 7b (605 mg, 69%) with mp 127.5−128.5 °C; H NMR
2
(400 MHz, CDCl3) δ 0.08 (s, 9 H; 119Sn satellites, J = 51.8 Hz;
5
SnMe3), 1.24 (s, 6 H, 2 × 1-CH3), 1.53 (s, 6 H; 119Sn satellites, J =
2.8 Hz; 2 × 3-CH3), 7.04 (d, 3J = 8.2 Hz, 2 H, 2′-/6′-H), 7.05 (m, 1 H,
7-H), 7.16−7.23 (m, 2 H, 5-/6-H), 7.18 (dm, 1 H, 4-H), 7.32 (tt, 3J =
7.2 Hz, 4J = 1.2 Hz, 1 H, p′-H), 7.43 (tt, 3J = 7.5 Hz, 2 H, 2 × m′-H),
7.53 (dm, 3J = 8.4 Hz, 2 H, 3′-/5′-H), 7.66 (dm, 3J = ca. 8 Hz, 2 H, 2
× o′-H), assigned through comparison with 6b and 7c and the
following selective {1H} decoupling experiments: {o′-H} → m′-H as a
d, {m′-H} → o′-H as a s, {3′-/5′-H} → 2′-/6′-H as a s, {p′-H} → m′-
H as a dm; 13C NMR (100.6 MHz, CDCl3) δ −5.0 (broadend q, 1J =
128.7 Hz; 119Sn satellites, 1J = 339 Hz; SnMe3), 31.9 (qq, 1J = 127 Hz,
3J = 4.4 Hz, 2 × 1-CH3 + 2 × 3-CH3), 48.4 (unresolved; 119Sn
120.5−122 °C; H NMR (400 MHz, CDCl3) δ 0.07 (s, 9 H; 119Sn
1
satellites, 2J = 51.8 Hz; SnMe3), 1.19 (s, 6 H, 2 × 1-CH3), 1.50 (s, 6 H;
119Sn satellites, J = 3.0 Hz; 2 × 3-CH3), 6.91 (dm, J = 8.5 Hz, 2 H,
2′-/6′-H), 7.05 (dm, 1 H, 7-H), 7.16 (dm, 1 H, 4-H), 7.23 (m, 2 H,
5-/6-H), 7.24 (dm, 3J = 8.5 Hz, 2 H, 3′-/5′-H), assigned through 119Sn
satellites and comparison with 7c; 13C NMR (100.6 MHz, CDCl3) δ
−5.1 (119Sn satellites, 1J = 341 Hz; SnMe3), 31.8 (2 × 1-CH3 + 2 × 3-
5
3
3
satellites, J(cis) = 21.5 Hz; C-3), 50.6 (unresolved; 119Sn satellites,
CH3), 48.45 (119Sn satellites, J(cis) = 20.9 Hz; C-3), 50.59 (119Sn
3
3J(trans) = 61.2 Hz; C-1), 122.25 (dd, 1J = 156 Hz, C-7), 122.32 (dd,
1J = 156 Hz, C-4), 125.9 (dd, 1J = 157 Hz; 119Sn satellites, 4J = 9.9 Hz,
C-3′/-5′), 126.74 (dm, 2 × C-o′), 126.79 (dm, C-5), 126.94 (2 dm, C-
6 and C-p′), 128.70 (dd, 1J = 158 Hz, 2 × C-m′), 128.85 (dd, 1J = 158
satellites, 3J(trans) = 59.6 Hz; C-1), 122.23 (C-7), 122.31 (C-4),
126.88 (C-5), 127.01 (C-6), 127.6 (119Sn satellites, 4J = 9.4 Hz; C-3′/-
5′), 129.7 (119Sn satellites, 3J = 18.4 Hz; C-2′/6′), 130.3 (119Sn
5
1
satellites, J = 14.4 Hz; C-4′), 137.8 (119Sn satellites, J = 438 Hz; C-
Hz; 119Sn satellites, J = 19.0 Hz; C-2′/6′), 137.1 (tt, J = 7.2 and 3.6
α), 144.09 (119Sn satellites, 2J = 21.8 Hz; C-1′), 149.7 (C-9), 150.0 (C-
3
3
Hz; 119Sn satellites, 5J = 12.7 Hz; C-4′), 138.6 (unresolved; 119Sn
8), 168.4 (119Sn satellites, J = 15.0 Hz; C-2), assigned as above; IR
2
4078
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
(KBr) ν 2988, 2958, 2923, 1606 (w), 1482 (s), 1088, 801, 761 (s), and
523 cm−1. Anal. Calcd for C23H29ClSn (459.64): C, 60.10; H, 6.36; Cl,
7.71. Found: C, 59.95; H, 6.15; Cl, 7.94.
°C under argon gas cover, treated with n-Bu6Li (0.16 mmol) in
cyclopentane (0.10 mL), and stored at −70 °C. The solution
contained only monomeric 8d between at least −88 and −2 °C
(Tables S21 and S22)13 and decomposed at room temperature within
five minutes to give the parent olefin 9d.
2-[α,4′-Bis(trimethylstannyl)benzylidene]-1,1,3,3-tetrame-
thylindane (7e). This was prepared in analogy with 7a12 from the 4′-
bromo-α-chloro derivative 6e (800 mg, 2.13 mmol) with Me3SnLi (ca.
6 mmol) in anhydrous THF (12 mL). After 30 min at ice temperature,
aqueous workup6 gave an oily 4:1 mixture (1220 mg) of 7e and the
source 6e along with only a trace of the olefin (α-H in place of the Cl
in 6e). Treatment with ethanol (5 mL) furnished spectroscopically
pure platelets of 7e (1059 mg, 85%) with mp 90−92.5 °C; analytically
pure 7e had mp 98.5−100 °C; 1H NMR (400 MHz, CDCl3) δ 0.04 (s,
2-(α-[6Li]Lithio-4′-trimethylstannylbenzylidene)-1,1,3,3-tet-
ramethylindane (8e). The bis(stannyl) derivative 7e (120 mg, 0.204
mmol) and anhydrous THF (0.50 mL) with [D6]benzene (0.08 mL)
and a trace of TMS in a dry NMR tube under argon gas at −70 °C
were treated with n-Bu6Li (0.207 mmol) in cyclopentane (0.120 mL)
and measured immediately at 25 °C only. 1H NMR (400 MHz, THF)
δ 0.064 (s, 9 H, 4′-SnMe3), 1.34 (s, 12 H, 2 × 1-CH3 and 2 × 3-CH3),
6.57 (d, 3J = 7.8 Hz, 2 H, 2′-/6′-H), 6.92 (d, 3J = 7.8 Hz, 2 H, 3′-/5′-
H); 13C NMR (100.6 MHz, THF) δ −9.54 (SnMe3), 33.7 (broadened,
4 C, 2 × 1-CH3 + 2 × 3-CH3), ca. 49.0 (very broad, C-3/-1), ca. 122.1
(C-2′/6′), 122.8 (C-7/-4), 125.8 (C-5/-6), ca. 127.8 (C-4′), 135.8 (C-
3′/-5′), 147.8 (broad, C-2), 153.9 (broadened, C-8/-9), 162.4 (broad,
C-1′), 188.3 (broad, C-α). All signals were obviously above
coalescence (1-CH3 = 3-CH3, C-3 = C-1, C-4 = C-7, C-5 = C-6, C-
8 = C-9), but C-2′/-6′ and C-3′/5′ showed multiple signals
presumably due to partial conversion of 4′-SnMe3 to 4′-SnMe2Bu
and 4′-SnMeBu2.
2
9 H; 119Sn satellites, J = 51.7 Hz; α-SnMe3), 0.29 (s, 9 H; 119Sn
2
satellites, J = 55 Hz; 4′-SnMe3), 1.20 (s, 6 H, 2 × 1-CH3), 1.51 (s, 6
H; 119Sn satellites, 5J = 2.9 Hz; 2 × 3-CH3), 6.93 (dt, 3J = 7.9 Hz, 2 H,
2′-/6′-H), 7.04 (dm, 1 H, 7-H), 7.17 (dm, 1 H, 4-H), 7.21 (m, 2 H,
5-/6-H), 7.34 (dt, 3J = 7.9 Hz, 2 H, 3′-/5′-H), assigned through
comparison with 7c and 7d; 13C NMR (100.6 MHz, CDCl3) δ −9.4
1
1
(
119Sn satellites, J = 348 Hz; 4′-SnMe3), −5.1 (119Sn satellites, J =
338 Hz; α-SnMe3), 31.82 and 31.84 (2 × 1-CH3 + 2 × 3-CH3), 48.3
(
119Sn satellites, 3J(cis) = 22.0 Hz; C-3), 50.5 (119Sn satellites, 3J(trans)
= 62.3 Hz; C-1), 122.22 (C-7), 122.31 (C-4), 126.74 (C-5), 126.89
(C-6), 128.1 (119Sn satellites, 3J = 18.9 Hz with α-Sn, 3J = 47.5 Hz with
4′-Sn; C-2′/6′), 134.7 (119Sn satellites, 4J = 9.4 Hz with α-Sn, 2J = 37.4
Hz with 4′-Sn; C-3′/-5′), 137.3 (119Sn satellites, 1J = 480 Hz, 5J = 13.2
2-(α,4′-[6Li]Dilithiobenzylidene)-1,1,3,3-tetramethylindane
(8f). See ref 13.
2-(4′-Trimethylsilylbenzylidene)-1,1,3,3-tetramethylindane
(9a). This was published as compound 4l on p S9 in the Supporting
Information of ref 10.
1
Hz; C-4′), 138.9 (119Sn satellites, J = 444 Hz; C-α), 145.3 (119Sn
satellites, 2J = 20.6 Hz with α-Sn, 4J = 10.6 Hz with 4′-Sn; C-1′), 149.9
(C-9), 150.4 (C-8), 167.2 (119Sn satellites, 2J = 16.8 Hz; C-2), assigned
as above; IR (KBr) ν 2985, 2958, 2922, 1609 (w), 1485, 1360, 770,
754 (s), and 526 (s) cm−1. Anal. Calcd for C26H38Sn2 (588.0): C,
53.11; H, 6.51. Found: C, 53.37; H, 6.42.
2-(4′-Phenylbenzylidene)-1,1,3,3-tetramethylindane (9b).
See ref 32.This was isolated from the mother liquors of 8b as
colorless crystals with mp 127.5−129 °C (from very little ethanol); 1H
NMR (400 MHz, CDCl3) δ 1.35 (s, 6 H, 2 × 1-CH3), 1.50 (s, 6 H, 2
× 3-CH3), 6.68 (s, 1 H, α-H), 7.12 (dm, 1 H, 7-H), 7.20−7.24 (m, 3
H, 4-/5-/6-H), 7.337 (overlaid tt, p′-H), 7.343 (d, 3J = ca. 8 Hz, 2 H,
2′-/6′-H), 7.44 (tm, 3J = 7.5 Hz, 2 H, 2 × m′-H), 7.57 (d, 3J = 8 Hz, 2
H, 3′-/5′-H), 7.64 (dm, 3J = 8 Hz, 2 H, 2 × o′-H), assigned through an
NOE difference spectrum ({1-CH3} → only 7-H and 2′-/6′-H
intensified) and selective {1H} decoupling as follows: {2′-/6′-H} →
3′-/5′-H as s, {o′-H} → m′-H as d and p′-H as a sharp t, {m′-H} → o′-
H as s and p′-H as s; 13C NMR (100.6 MHz, CDCl3) δ 31.02 (qq, 1J =
127 Hz, 3J = 4.5 Hz, 2 × 1-CH3), 32.43 (qq, 1J = 127 Hz, 3J = 4.5 Hz, 2
× 3-CH3), 47.23 (broader m, C-1), 47.81 (narrower m, hence C-3),
122.29 (dd, 1J = 156 Hz, 3J = 6.8 Hz, C-7), 122.39 (dt, 1J = 150 Hz, 3J
= 4 Hz, C-α), 122.52 (dm, 1J = 156 Hz, C-4), 126.36 (dd, 1J = 157 Hz,
C-3′/-5′), 126.91 (dd, 1J = 159 Hz, C-5), 126.95 (overlaid dm, 1J = ca.
160 Hz, 2 × C-o′), 127.11 (overlaid dm, 1J = ca. 160 Hz, C-6), 127.15
(overlaid dm, 1J = ca. 160 Hz, C-p′), 128.75 (dm, 1J = 156 Hz, 2 × C-
m′), 129.81 (dm, 1J = 159 Hz, C-2′/6′), 138.0 (sharp t, 3J = 7.5 Hz, C-
1′), 139.0 (tt, 3J = ca. 7 and 3.6 Hz, C-4′), 140.9 (tt, 3J = 7 and 3.2 Hz,
C-i′), 148.6 (unresolved, C-9), 150.6 (unresolved, C-8), 162.3 (m, 3J =
3.5 Hz, C-2), assigned through selective {1H} decoupling as follows:
{1-CH3} → 1-CH3 as s, C-1 narrowed, C-8 as a t; {3-CH3} → 3-CH3
as s, C-3 narrowed, C-9 as t 6 Hz; {7-H} → C-7 as s, C-9 narrowed; IR
(KBr) ν 2959, 2921, 2860, 1484, and 755 cm−1. Anal. Calcd for C26H26
(338.49): C, 92.26; H, 7.74. Found: C, 92.44; H, 7.71.
2-(α-[6Li]Lithio-4′-trimethylsilylbenzylidene)-1,1,3,3-tetra-
methylindane (8a). The crystalline t-BuOMe solvate of 8a was
prepared as reported on pp 14188 and S36 of ref 6 (12b therein), but
finally dissolved in anhydrous THF (1H and 13C NMR at −72 °C
described on p S39 of ref 6), where it was monomeric up to at least 46
°C (Tables S13 and S14).13
2-(α-[6Li]Lithio-4′-phenylbenzylidene)-1,1,3,3-tetramethy-
lindane (8b). The procedure described for 8a was carried out under
argon gas cover with 7b (120 mg, 0.24 mmol) in cyclopentane (0.8
mL) plus Et2O (0.074 mL, 0.71 mmol) and n-Bu6Li (0.36 mmol) in
cyclopentane (0.21 mL), which precipitated slightly yellow rodlets
overnight at room temperature. After washing with dry cyclopentane
and blowing with dry argon gas for five seconds, the rodlets were
dissolved in anhydrous THF or in dry Et2O and measured at variable
temperatures (Tables S15−S18);13 they are almost insoluble in t-
BuOMe or in [D8]toluene. Finally, the addition of D2O (ca. 0.06 mL)
and aqueous workup afforded the parent olefins 9b and [α-D]9b
(42:58, 39 mg, 48%) with mp 122−124.5 °C (see further below).
8b is monomeric in THF between at least −94 and +25 °C (Tables
S15 and S16)13 and in Et2O/[D6]benzene (6:1) between at least −72
and +52 °C (Tables S17 and S18).13 The two 1-/3-CH3 signals in
1
Et2O showed slight broadening but not yet H NMR coalescence at
+52 °C and 400 MHz (where Δν = 18 Hz), so that kψ < 40 s−1 and
ΔG⧧ > 17.1 kcal mol−1 for the cis/trans diastereotopomerization in
Et2O (that is, >5 kcal mol−1 higher than in THF).
2-(4′-Methylbenzylidene)-1,1,3,3-tetramethylindane (9c).
The mother liquors from 7c were stirred in Et2O (5 mL) with
concentrated HCl (5 mL) for 24 h at room temperature. The aqueous
layer was extracted with Et2O (3 × 3 mL), and the combined ethereal
phases were washed with distilled water, dried over MgSO4, and
concentrated. Pure 9c distilled at 118−123 °C (bath temp)/0.001
mbar with mp 50−51.5 °C; 1H NMR (400 MHz, CDCl3) δ 1.30 (s, 6
H, 2 × 1-CH3), 1.47 (s, 6 H, 2 × 3-CH3), 2.36 (s, 3 H, 4′-CH3), 6.62
(s, 1 H, α-H), 7.11 (m, 1 H, 7-H), 7.13 (d, 3J = 8 Hz, 2 H, 3′-/5′-H),
7.15 (d, 3J = 8 Hz, 2 H, 2′-/6′-H), 7.21 (m, 3 H, 4-/5-/6-H), assigned
through comparison with 9b; 13C NMR (100.6 MHz, CDCl3) δ 21.1
2-(α-[6Li]Lithio-4′-methylbenzylidene)-1,1,3,3-tetramethy-
lindane (8c). As described for 8a, the α-SnMe3 compound 7c (60 mg,
0.14 mmol) was partially dissolved in pentane (0.5 mL) with Et2O
(0.043 mL, 0.41 mmol) and then treated with n-Bu6Li (0.34 mmol) in
cyclopentane (0.20 mL). The slow growth of a few rather big cubes (=
dimeric Et2O solvate of 8c) was observed during four days at room
temperature, but the accompanying very small cubes could not be
recrystallized by gentle warming. In a THF solution, 8c was totally
monomeric between −90 and +53 °C (Tables S19 and S20),13 but it
rearranged slowly (one hour at 53 °C) to produce the less basic
benzyllithium isomer 11.
1
3
1
(qt, J = 126 Hz, J = 4 Hz, 4′-CH3), 31.0 (qq, J = 127 Hz, 2 × 1-
1
CH3), 32.4 (qq, J = 127 Hz, 2 × 3-CH3), 47.2 (broader, C-1), 47.7
(narrower, hence C-3), 122.3 (dm, 1J = 156 Hz, C-7), 122.5 (dm, 1J =
156 Hz, C-4), 122.7 (dt, 1J = 150 Hz, 3J = 3.8 Hz, C-α), 126.8 (dm, 1J
2-(4′-Chloro-α-[6Li]lithiobenzylidene)-1,1,3,3-tetramethylin-
dane (8d). A dry NMR tube was charged with the 4′-chlorostannane
7d (70 mg, 0.15 mmol) in anhydrous THF (0.50 mL), cooled to −70
1
1
= 159 Hz, C-5), 127.0 (ddm, J = 159 Hz, C-6), 128.4 (dqi, J = 156
4079
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
1
3
Hz, C-3′/5′), 129.2 (dm, J = 157.5 Hz, C-2′/6′), 135.7 (tm, J = ca.
3J = 5 Hz, 2′-/6′-CH3), 30.5 (qm, 1J = 126 Hz, 2 × 1-CH3), 33.8 (qm,
1J = 126 Hz, 2 × 3-CH3), 47.3 (unresolved, C-3), 48.6 (unresolved, C-
1), 114.9 (sharp d, 1J = 156 Hz, C-4′), 122.8 (dm, 1J = ca. 154 Hz, C-
7), 122.9 (dm, 1J = ca. 154 Hz, C-4), 125.1 (m, apparent J = 6 Hz, C-
3
6.2 Hz, C-4′), 135.9 (sharp t, J = 7 Hz, C-1′), 148.7 (unresolved, C-
9), 150.7 (unresolved, C-8), 161.6 (unresolved, C-2), assigned through
the 13C/1H coupling patterns and comparison with 9b; IR (KBr) ν
3020, 2962, 2923, 2861, 1512, 1483, 1457, 1360, 1109, 1025, 861, 793,
760, and 661 cm−1. Anal. Calcd for C21H24 (276.42): C, 91.25; H,
8.75. Found: C, 91.53; H, 8.66.
1
1
2′/6′), 125.7 (dd, J = 156 Hz, C-5), 126.0 (dd, J = 156 Hz, C-6),
1
126.3 (dm, J = 150 Hz, C-3′/5′), 142.5 (unresolved, C-2), 153.1
(unresolved, C-9), 154.2 (unresolved, C-8), 158.5 (unresolved, C-1′),
1
186.9 (t, J(13C,6Li) = 10.7 Hz, C-α).
2-(4′-Chlorobenzylidene)-1,1,3,3-tetramethylindane (9d).
See ref 13.
2-(2′-/6′-Dimethylbenzylidene)-1,1,3,3-tetramethylindane
2-(Benzylidene)-1,1,3,3-tetramethylindane (9f). For the sol-
vent and temperature dependencies of 1H and 13C NMR δ values, see
Tables S6 and S7 on pp S19−S20 in the Supporting Information of ref
6.
(15). See ref 13.
2-[2′-(Monodeuteriomethyl)-6′-methylbenzylidene]-1,1,3,3-
tetramethylindane (16). See ref 13.
2-[2′-(Lithiomethyl)-6′-methylbenzylidene)-1,1,3,3-tetrame-
thylindane (17). 17 was formed from 14 through trans-lithiation and
was stable in THF for up to three days at 25 °C. The temperature-
2-(4′-Methylphenyl)-2-(1,1,3,3-tetramethyl-2-indanylidene)-
acetic Acid (10). See ref 13.
1
independent 13C and H NMR nonequivalences of the two 3-CH3
2-(4′-Lithiomethylbenzylidene)-1,1,3,3-tetramethylindane
(11). The α-lithio-4′-methyl compound 8c vanished within one hour
at 53 °C in THF solution, forming the olefin 9c together with 11,
whose later deuteriolysis afforded the deuteriated olefin 9g. The
temperature-independent (Tables S24 and S25)13 NMR data of 11 are
characteristic of a p-substituted benzyllithium derivative in THF
groups (Tables S28 and S29)13 show that (half-)rotation about the C-
1
α/C-1′ bond is slow on our NMR time scales. H NMR (400 MHz,
THF, 25 °C) δ 1.32 (s, 3 H, 1 × 1-CH3; second 1-CH3 not detected),
1.42 and 1.46 (2 s, 2 × 3 H, 2 × 3-CH3), 1.93 (s, 3 H, 6′-CH3), (2′-
CH2Li not detected), 5.44 (d, 3J = 7.0 Hz, 1 H, 5′-H), 6.00 (s, 1 H, α-
15
1
1
inclusive of an enhanced J(13CH2) value (expected 134 Hz): H
3
H), 6.11 (d, 3J = 8.0 Hz, 1 H, 3′-H), 6.26 (t, J = 7.6 Hz, 1 H, 4′-H),
NMR (400 MHz, THF, 25 °C) δ 1.34 (s, 6 H, 2 × 1-CH3), 1.51 (s, 6
7.07 (m, 4 H, 4-/5-/6-/7-H), assigned through comparison of the
lithiation shifts Δδ (Table 1) with those of benzyllithium;16 13C NMR
(100.6 MHz, THF, 25 °C) δ 21.9 (6′-CH3), 29.7 and 30.1 (2 × 1-
CH3), 32.3 and 33.4 (2 × 3-CH3), 38.4 (2′-CH2Li), 47.6 (C-1), 47.9
(C-3), 106.5 (C-5′), 115.1 (C-3′), 121.3 (C-1′), 122.7 (C-7), 122.9
(C-4), 126.0 (C-α), 126.6 (C-5), 126.7 (C-6), 127.1 (C-4′), 133.8 (C-
6′), 150.1 (C-9), 153.1 (C-8), 156.0 (C-2′), 157.9 (C-2), assigned
through comparison with 11 and as above with Δδ of benzyllithium.15
2-[α-(1,1,3,3-Tetramethyl-2-indanylidene)benzyl]-
adamantan-2-ol (25). Prepared from 3 with adamantan-2-one in
3
H, 2 × 3-CH3), 6.06 (d, J = 8.5 Hz, 2 H, 3′-/5′-H), 6.13 (s, 1 H, α-
H), 6.53 (d, 3J = 8.5 Hz, 2 H, 2′-/6′-H); 13C NMR (100.6 MHz, THF,
25 °C) δ 29.8 (qq, 1J = 126 Hz, 2 × 1-CH3), 33.0 (qq, 1J = 126 Hz, 2
× 3-CH3), 41.2 (t, 1J = 136 Hz, 4′-CH2Li), 46.8 (m, C-1), 48.5 (m, C-
3), 113.8 (t, 3J = 7.5 Hz, C-1′), 115.7 (dq, J = 150 Hz, 3J = 7 Hz, C-
1
1
3′/5′), 122.8 (C-7), 122.9 (C-4), 126.4 (dt, J = 141 Hz, C-α), 126.8
(dm, 1J = 159 Hz, C-5), 127.2 (dm, 1J = 158 Hz, C-6), 131.2 (dt, 1J =
3
149 Hz, J = 6 Hz, C-2′/6′), 146.6 (m, C-2), 150.2 (m, C-9), 153.6
3
(m, C-8), 158.4 (t, J = 7.2 Hz, C-4′).
1
Et2O; mp 160.5−162 °C (pentane); H NMR (400 MHz, CDCl3, 25
2-(α-Chloro-2′/6′-dimethylbenzylidene)-1,1,3,3-tetramethy-
lindane (12). See ref 13.
°C) δ 0.98 (broadened s, 6 H, 2 × 1-CH3), 1.48 (d, 2J ≈ 12 Hz, 4 H, 1
× 4″-H, 1 × 8″-H, 1 × 9″-H, and 1 × 10″-H), 1.64 (broad t, 3J ≈ 3 Hz,
2 H, 2 enantiotopic 6″-H), 1.74 (m, 3J ≈ 3 Hz, 1 H, 5″-H), 1.76 (s, 6
H, 2 × 3-CH3), 1.78 (m, 3J ≈ 3 Hz, 1 H, 7″-H), 1.83 (broad d, 2J ≈ 12
Hz, 2 H, 1 × 8″-H and 1 × 10″-H), 2.30 (broad d upon broad m, 2J ≈
2-(2′-/6′-Dimethyl-α-trimethylstannylbenzylidene)-1,1,3,3-
tetramethylindane (13). This was prepared in analogy with 7a12
from 12 (100 mg, 0.308 mmol) with Me3SnLi (ca. 0.62 mmol) in
anhydrous THF (3.4 mL). Aqueous workup6 after 90 min at room
temperature gave a 68:32 mixture (129 mg) of 13 and its parent olefin
15. Analytically pure 13 (72 mg, 52%) had mp 104−105 °C (ethanol);
1H NMR (400 MHz, CDCl3) δ 0.03 (s, 9 H; 119Sn satellites, 2J = 50.9
Hz; SnMe3), 1.17 (s, 6 H, 2 × 1-CH3), 1.57 (s, 6 H, 2 × 3-CH3), 2.17
(s, 6 H, 2′-/6′-CH3), 6.97 (quasi-s, 3 H, 3′-/5′-H and 4′-H), 7.05 (dm,
3J = ca. 7 Hz, 1 H, 7-H), 7.17−7.23 (m, 3 H, 4-/5-/6-H), assigned
through comparisons with 7b and with compound 21 of ref 20; 13C
3
12 Hz, 4 H, 1 × 4″-H and 1 × 9″-H upon 1″-/3″-H), 6.98 (dm, J =
7.5 Hz, 1 H, 7-H), 7.13 (dm, 3J = 7.5 Hz, 1 H, 4-H), 7.16 (td, 3J = 7.2
Hz, 1 H, 6-H), 7.21 (td, 3J = 7.2 Hz, 1 H, 5-H), 7.24 (partially hidden
t, 3′-/5′-H), 7.25 (hidden, 4′-H), 7.35 (dm, 2 H, 2′-/6′-H), assigned
through SCS,20 comparison with 2-tert-butyladamantan-2-ol,33 and the
NOESY correlations 2′-/6′-H ↔ 3′-/5′-H ↔ 1-CH3 (w, no cross-
peaks with adamantyl signals), 6-H ↔ 7-H ↔ 1-CH3 ↔ 2′-/6′-H ↔
8″-/10″-H (at δ = 1.83) ↔ 8″-/10″-H (at δ = 1.48), and 5-H ↔ 4-H
↔ 3-CH3 ↔ 1″-/3″-H ↔ 4″-/9″-/8″-/10″-H (at δ = 1.48) ↔ (5″-/
1
NMR (100.6 MHz, CDCl3) δ −3.78 (sharp q, J = 129 Hz, SnMe3),
21.3 (q, 1J = 126 Hz, 2′-/6′-CH3), 29.3 (qq, 1J = 127 Hz, 3J = 4.6 Hz, 2
1
1
3
7″-H) ↔ 6″-H; H NMR (400 MHz, CDCl3, −63 °C) δ = 0.50
× 1-CH3), 31.8 (q, J = 127 Hz, J = 4.6 Hz, 2 × 3-CH3), 49.2
(unresolved, C-3), 50.3 (unresolved, C-1), 122.1 (dm, 1J = 156 Hz, C-
7), 122.3 (dm, 1J = 156 Hz, C-4), 125.0 (sharp d, 1J = 158.7 Hz, C-4′),
126.8 (dd, 1J = 159 Hz, 3J = 7.5 Hz, C-5), 127.0 (dm, 1J = 159 Hz, C-
6), 127.1 (dm, 1J = 158 Hz, C-3′/-5′), 134.1 (m, C-2′/6′), 137.7 (s, C-
α), 144.3 (m, C-1′), 149.6 (m, C-9), 150.3 (m, C-8), 165.0 (m, C-2),
assigned as above; IR (KBr) ν 2987, 2958, 2921, 1602, 1486, 1460,
1361, 1027, 761 (s), and 521 cm−1. Anal. Calcd for C25H34Sn
(453.25): C, 66.25; H, 7.56. Found: C, 66.44; H, 7.44.
(broadened s, 3 H, 1 × 1-CH3), 1.42 (s, 3 H, 1 × 3-CH3), 1.60 (s, 3 H,
1 × 1-CH3), 1.62 (not resolved, 2 H, 2 × 6″-H), 1.76 (1 H, 5″-H),
1.80 (1 H, 7″-H), 1.96 (s, 3 H, 1 × 3-CH3), 3.33 (s, 1 H, OH), 7.05 (1
H, 7-H), 7.44 (very broad, 1 × 2′-/6′-H); 13C NMR (100.6 MHz,
CDCl3, 25 °C) δ 26.1 (dm, 1J = 132 Hz, C-5″), 26.7 (dm, 1J = 132 Hz,
1
3
C-7″), 31.3 (sharp qq, J = 127 Hz, J = 4.5 Hz, 2 × 3-CH3), 32.7
1
1
(broad q, J = 127 Hz, 2 × 1-CH3), 33.4 (t, J = 128 Hz, C-4″/-9″),
35.3 (t, 1J = 128 Hz, C-8″/-10″), 36.2 (d, 1J = 132 Hz, C-1″/-3″), 37.4
1
2-(α-[6Li]Lithio-2′-/6′-dimethylbenzylidene)-1,1,3,3-tetrame-
thylindane (14). A dry NMR tube containing the α-SnMe3
compound 13 (50 mg, 0.11 mmol) in anhydrous THF (0.5 mL)
was cooled to −30 °C, treated with n-Bu6Li (0.12 mmol) in
cyclopentane (0.081 mL), and stored at −70 °C. 14 was totally
monomeric between at least −95 and +25 °C (Tables S26 and S27)13
but suffered rapid trans-lithiation (within 20 min at 25 °C) to give the
benzyllithium derivative 17 along with the parent olefin 15. The final
addition of D2O and aqueous workup yielded 15 containing some 16
(total 31 mg, 97%). 1H NMR of 14 (400 MHz, THF, −95 °C) δ 1.11
(s, 6 H, 2 × 1-CH3), 1.37 (s, 6 H, 2 × 3-CH3), 2.04 (s, 6 H, 2′-/6′-
CH3), 6.19 (t, 3J = 7.3 Hz, 1 H, 4′-H), 6.56 (d, 3J = 7.3 Hz, 2 H, 3′-/
5′-H), 6.95 (m, 1 H, 7-H), 7.00 (m, 2 H, 5-/6-H), 7.09 (m, 1 H, 4-H);
13C NMR of 14 (100.6 MHz, THF, −95 °C) δ 22.7 (qd, 1J = 124 Hz,
(t, J ≈ 130 Hz, C-6″), 48.1 (unresolved, C-3), 51.0 (m, C-1), 79.9
(broad, C-2″), 122.0 (dd, 1J = 156 Hz, 3J = 7.5 Hz, C-7), 122.2 (dd, 1J
3
1
3
= 157 Hz, J = 7.5 Hz, C-4), 125.8 (dd, J = 159 Hz, J ≈ 7.5 Hz, C-
1
3
1
3′/-5′), 126.4 (dt, J = 159.5 Hz, J = 7.5 Hz, C-4′), 126.6 (dd, J =
159.5 Hz, 3J = 7.5 Hz, C-5), 126.9 (dd, 1J = 159 Hz, 3J = 7.5 Hz, C-6),
132.3 (dt, 1J = 159 Hz, 3J = 6.5 Hz, C-2′/-6′), 140.7 (t, 3J ≈ 7 Hz, C-
1′), 145.1 (unresolved, C-α), 148.8 (broad, C-8), 151.6 (broad, C-9),
157.7 (unresolved, C-2), assigned through SCS,20 comparison with 2-
tert-butyladamantan-2-ol,33 and 13C/1H heterocorrelation (HSQC);
13C NMR (100.6 MHz, CDCl3, −63 °C) δ 25.5 (C-5″), 26.0 (C-7″),
30.0 (1 × 3-CH3), 30.2 (1 × 1-CH3), 31.3 (1 × 1-CH3), 32.2 (C-4″ or
C-9″), 33.5 (C-9″ or C-4″), 34.1 (C-8″ or C-10″), 34.4 (1 × 3-CH3),
34.8 (C-1″ or C-3″), 35.4 (C-10″ or C-8″), 36.5 (C-3″ or C-1″), 36.8
(C-6″), 47.8 (C-3), 50.8 (C-1), 79.7 (C-2″), 122.0 (C-7), 122.3 (C-4),
4080
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081

Organometallics
Article
125.4 (broad, C-3′ or C-5′), 126.1 (broad, C-5′ or C-3′), 126.2 (sharp,
C-4′), 126.6 (C-5), 126.9 (C-6), 131.1 (broad, C-2′ or C-6′), 132.4
(broad, C-6′ or C-2′), 140.2 (C-1′), 144.9 (C-α), 148.3 (C-8), 150.9
(C-9), 156.6 (C-2); IR (KBr) ν 3636 (sharp O−H), 2910, 1601 (w),
1488, 762, and 715 cm−1. Anal. Calcd for C30H36O (412.6): C, 87.33;
H, 8.79. Found: C, 88.00; H, 9.02.
9-[α-(1,1,3,3-Tetramethyl-2-indanylidene)benzyl]fluoren-9-
ol (26). See ref 13.
1,1,2-Triphenyl-2-(1,1,3,3-tetramethyl-2-indanylidene)-
ethanol (27a). See ref 13.
(7) Binsch, G.; Eliel, E.; Kessler, H. Angew. Chem. 1971, 83, 618−
619; Angew. Chem., Int. Ed. 1971, 10, 570−572.
(8) Knorr, R.; Lattke, E. Tetrahedron Lett. 1977, 18, 3969−3972.
(9) See compound 4m in ref 10.
(10) Knorr, R.; Pires, C.; Behringer, C.; Menke, T.; Freudenreich, J.;
Rossmann, E. C.; Bohrer, P. J. Am. Chem. Soc. 2006, 128, 14845−
̈
14853.
(11) See compound 4n in ref 10.
(12) See the preparation of 7a in ref 6 (described therein as
compound 11 with reversed positional numbering).
(13) See the Supporting Information.
1,1-Dicyclopropyl-2-phenyl-2-(1,1,3,3-tetramethyl-2-
indanylidene)ethanol (27b). See ref 13.
(14) Reich, H. J.; Green, D. P.; Medina, M. A.; Goldenberg, W. S.;
̈
Gudmundsson, B. O.; Dykstra, R. R.; Phillips, N. C. J. Am. Chem. Soc.
ASSOCIATED CONTENT
1998, 120, 7201−7210 and Figure 13 therein.
■
(15) (a) Takahashi, K.; Kondo, Y.; Asami, R.; Inoue, Y. Org. Magn.
Reson. 1974, 6, 580−582. Notice that the assignments of C-o and C-
m for toluene in THF (Table 1 therein) have to be reversed. (b)
Compare also: 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.
S
* Supporting Information
Preparations of compounds 6e, 7a, 8f, 9d, 10, 12, 15, 16, 26,
27a, 27b, and the α,4′-dicarboxylic acid; cis/trans diaster-
eotopomerization rate constants; Hammett p-substituent
parameters σp− with Figure S7; lithiation NMR shifts; tabulated
primary NMR data; Schiff base 23 as a carbanion model. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(16) Maercker, A.; Stotzel, R. J. Organomet. Chem. 1983, 254, 1−12
̈
and Table 5 therein.
(17) Schaefer, T.; Schneider, W. G. Can. J. Chem. 1963, 41, 966−982.
(18) (a) O’Brien, D. H.; Hart, A. J.; Russel, C. R. J. Am. Chem. Soc.
1975, 97, 4410−4412 and references therein. (b) Hallden-Abberton,
M.; Fraenkel, G. Tetrahedron 1982, 38, 71−74. (c) Bradamante, S.;
Pagani, G. A. J. Org. Chem. 1984, 49, 2863−2870.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: rhk@cup.lmu.de.
(19) Tables S6 and S7 in the Supporting Information of ref 6.
(20) Knorr, R.; von Roman, T.; Freudenreich, J.; Hoang, T. P.;
Mehlstaubl, J.; Bohrer, P.; Stephenson, D. S.; Huber, H.; Schubert, B.
̈
̈
Notes
Magn. Reson. Chem. 1993, 31, 557−565 on p 563.
(21) Binsch, G. Top. Stereochem. 1968, 3, 97−192 on p 178.
(22) (a) Cram, D. J.; Gosser, L. J. Am. Chem. Soc. 1964, 86, 2950−
2951. (b) Cram, D. J.; Gosser, L. J. Am. Chem. Soc. 1964, 86, 5457−
5465 Chart VI therein.
The authors declare no competing financial interest.
†E/Z Equilibria, Part 20. Part 19: Knorr, R.; Behringer, C.;
Noth, H.; Schmidt, M.; Lattke, E.; Rapple, E. Chem. Ber./Recl.
̈
̈
1997, 130, 585−592.
(23) As explained by Knorr, R.; Menke, T.; Ferchland, K.
Organometallics 2013, 32, 468−472.
ACKNOWLEDGMENTS
■
(24) Knorr, R.; Ferchland, K.; Mehlstaubl, J.; Hoang, T. P.; Bohrer,
̈
̈
We gratefully acknowledge financial support by the Deutsche
Forschungsgemeinschaft. This work is dedicated to Professor
Herbert Mayr in recognition of his interest in and support of
the “negative” part of ion kinetics.
P.; Ludemann, H.-D.; Lang, E. Chem. Ber. 1992, 125, 2041−2049
̈
where ΔG⧧ of compound 3b was extrapolated to 126 °C for
comparison with 7b in Table 2 therein.
(25) Knorr, R.; Hoang, T. P.; Mehlstaubl, J.; Hintermeyer-Hilpert,
̈
M.; Ludemann, H.-D.; Lang, E.; Sextl, G.; Rattay, W.; Bohrer, P. Chem.
Ber. 1993, 126, 217−224 on p 220.
̈
̈
REFERENCES
■
(26) Curtin, D. Y.; Koehl, W. J. J. Am. Chem. Soc. 1962, 84, 1967−
1973.
(1) Reich, H. J.; Medina, M. A.; Bowe, M. D. J. Am. Chem. Soc. 1992,
114, 11003−11004.
(27) Panek, E. J.; Neff, B. L.; Chu, H.; Panek, M. L. J. Am. Chem. Soc.
1975, 97, 3996−4000.
(2) The following references contain more recent instructive
collections of contributions to the field of configurational stability of
mainly quasi-sp3 carbanionic compounds in general; they do not cover
the present topic of sp2-carbanion stereoinversion. (a) Basu, A.;
Thuyamanavan, S. Angew. Chem. 2002, 114, 740−763; Angew. Chem.,
Int. Ed. 2002, 41, 716−738. (b) Gawley, R. E. Top. Stereochem. 2010,
(28) ten Hoedt, R. W. M.; van Koten, G.; Noltes, J. G. J. Organomet.
Chem. 1979, 170, 131−149 eq 5 therein.
(29) Curtin, D. Y.; Crump, J. W. J. Am. Chem. Soc. 1958, 80, 1922−
1926 formula X on p 1923.
(30) This estimate rests on the premise that the monomer/dimer
equilibrium parameters measured23 in [D8]toluene with limited
amounts of THF apply also to solutions in preponderant THF.
(31) Notice that we have chosen here an atom-numbering system
that equals that in Scheme 1 of ref 10 but is the reverse of that used
previously in the Supporting Information of ref 10 for compounds 4k
and 4m therein.
(32) The locants i′, o′, m′, and p′ refer to the 4′-phenyl group.
(33) Duddeck, H.; Rosenbaum, D. J. Org. Chem. 1991, 56, 1700−
1707 and Supplementary Material therein.
26, 93−133. (c) Ott, H.; Daschlein, C.; Leusser, D.; Schildbach, D.;
̈
Seibel, T.; Stalke, D.; Strohmann, C. J. Am. Chem. Soc. 2008, 130,
11901−11911. (d) Capriati, V.; Florio, S.; Perna, F. M.; Salomone, A.;
Abbotto, A.; Ahmedjkouh, M.; Nilsson Lill, S. O. Chem.−Eur. J. 2009,
15, 7958−7979. (e) Capriati, V.; Florio, S.; Perna, F. M.; Salomone, A.
Chem.−Eur. J. 2010, 16, 9778−9788. (f) Perna, F. M.; Salomone, A.;
Dammaco, M.; Florio, S.; Capriati, V. Chem.−Eur. J. 2011, 17, 8216−
8225. (g) Seel, S.; Dagousset, G.; Thaler, T.; Frischmuth, A.;
Karaghiosoff, K.; Zipse, H.; Knochel, P. Chem.−Eur. J. 2013, 19,
4614−4622.
(3) (a) Ahlbrecht, H.; Harbach, J.; Hoffmann, R. W.; Ruhland, T.
Liebigs Ann. Chem. 1995, 211−216. (b) Ruhland, T.; Hoffmann, R. W.;
Schade, S.; Boche, G. Chem. Ber. 1995, 128, 551−556.
(4) Reich, H. J.; Dykstra, R. R. Angew. Chem. 1993, 105, 1489−1491;
Angew. Chem., Int. Ed. 1993, 32, 1469−1470.
(5) See Table 1 of ref 6.
(6) Knorr, R.; Menke, T.; Ferchland, K.; Mehlstaubl, J.; Stephenson,
̈
D. S. J. Am. Chem. Soc. 2008, 130, 14179−14188.
4081
dx.doi.org/10.1021/om4000852 | Organometallics 2013, 32, 4070−4081