“Conducted Tour” Migration of Li+ during the cis/trans Stereoinversion of α-Arylvinyllithiums

Article abstract of DOI:10.1002/chem.201702353

A “conducted tour” migration keeps a mobile client on a profitable route even though an occasional side-step may seem attractive. A stereochemical manifestation of such a migration had been suggested by Donald J. Cram (1964), and we present now a different example that concerns the cis/trans stereoinversion of monomeric H₂C=C(Li)-aryl compounds: Upon tetrahydrofuran (THF)-assisted heterolysis of the Li?C bond with formation of a solvent-separated ion pair (SSIP), the unchained “mobile client” Li⁺(THF)₄ is proposed to surmount the rim of the electronically fixed aryl group and to disdain the less encumbered pathways across the H₂C=C region. This interpretation is based on knowledge from a previously published series of monomeric α-arylalkenyllithiums in combination with two new members: 4-(α-lithiovinyl)-2,2-dimethylbenz[f]indane (1) revealed both a barrier against α-aryl rotation and a route-distinguishing retardation as compared with the corresponding migration-dependent cis/trans stereoinversion rate constant of 1-(α-lithiovinyl)naphthalene (2). Monomeric and dimeric ground states of 1 and 2 and their microsolvation numbers were determined by using the recently developed primary and secondary NMR criteria.

Full text of DOI:10.1002/chem.201702353

A Journal of
Accepted Article
Title: A "Conducted Tour" Migration of Li+ during the cis/trans
Stereoinversion of alpha-Arylvinyllithiums
Authors: Rudolf H. Knorr, Claudia Behringer, Jakob Ruhdorfer, Ulrich
von Roman, Ernst Lattke, and Petra Böhrer
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201702353
Link to VoR: http://dx.doi.org/10.1002/chem.201702353
Supported by

10.1002/chem.201702353
Chemistry - A European Journal
A “Conducted Tour” Migration of Li+ during the cis/trans Stereoinversion
of -Arylvinyllithiums
Rudolf Knorr,* Claudia Behringer, Jakob Ruhdorfer, Ulrich von Roman, Ernst Lattke, and
Petra Böhrer[a]
Dedicated to the memory of Professor Donald J. Cram
[a] Prof. Dr. R. Knorr, Dr. C. Behringer, Dr. J. Ruhdorfer, Dr. U. von Roman, P. Böhrer
Department Chemie, Ludwig-Maximilians-Universität München, Butenandtstrasse 5-13
(Haus F), 81377 München, Germany
E-mail: rhk@cup.uni-muenchen.de
Supporting information for this article includes lithiation shifts , preparation of 15,
tabulated primary NMR data, and diastereotopomerization rate constants; it is given via a link
at the end of the document.
Key words. conformational analysis • cis,trans configurational lability • reaction mechanisms
• sp2-carbanion inversion • structural elucidation
Abstract. A “conducted tour” migration keeps a mobile client on a profitable route even
though an occasional side-step may seem attractive. A stereochemical manifestation of such a
migration had been suggested by Donald J. Cram (1964), and we present now a different
example that concerns the cis/trans stereoinversion of monomeric H2C=C(Li)–aryl
compounds: Upon THF-assisted heterolysis of the Li–C bond with formation of a solvent-
separated ion pair (SSIP), the unchained “mobile client” Li+(THF)4 is proposed to surmount
the rim of the electronically fixed aryl group and to disdain the less encumbered pathways
across the H2C=C region (see Scheme 1). This interpretation is based on knowledge from a
previously published series of monomeric -arylalkenyllithiums in combination with two new
members: 4-(-lithiovinyl)-2,2-dimethylbenz[f]indane (1) revealed both a barrier against -
aryl rotation and a route-distinguishing retardation as compared with the correspondingly
migration-dependent cis/trans stereoinversion rate constant of 1-(-lithiovinyl)naphthalene
(2). Monomeric and dimeric ground states of 1 and 2 and their microsolvation numbers were
determined by means of recently developed primary and secondary NMR criteria.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
Introduction
The pseudomonomolecular, ionic mechanism[1,2] of cis/trans interconversion
(configurational lability) of -arylvinyllithiums in tetrahydrofuran (THF) as a solvent depends
on three kinds of motion: Heterolytic cleavage of the Li–C() bond, followed by migration of
the Li cation, and a transitory rehybridization of the charge-bearing, quasi-sp2 hybridized
center C() that inverts its configuration. The whole process is illustrated in Scheme 1 for
the imagined 4-(-lithiovinyl)-2,2-dimetrhylbenz[f]indane (1), whose stereoinversion
(“diastereotopomerization”) would interconvert the NMR chemical shifts  of the two
diastereotopic -H atoms. The trisolvated ground state of 1 (to be disclosed in this work) is a
contact ion pair 1-CIP and is expected to demand a high concentration of THF for the
immobilization of a fourth THF ligand at lithium with formation of the THF-separated ion
pair 1-SSIP as a highly reactive intermediate. The increased distance of Li+(THF)4 from the
carbanionic center C() in 1-SSIP would enable the charge-carrying, quasi-sp2 orbital at C()
to change into a transitory px orbital (Scheme 1) in the proposed[1] transition state of
stereoinversion. The ensuing descent from the transition state would transform px into the
quasi-sp2 orbital of the stereoinverted intermediate 1´-SSIP. However, 1´-SSIP cannot be
formed without a suitable migration of Li+(THF)4: Since ionic dissociation of an SSIP
intermediate into the free ions had been excluded for related -arylalkenyllithiums through
the observation of concentration-independent rate constants k,[1–5] such a migration must
occur within the THF-solvated ion pairs on the way from 1-SSIP to 1´-SSIP. The subsequent
loss of one THF ligand from 1´-SSIP would form the final product 1´-CIP, so that the
transitory immobilization of one additional THF ligand at Li implies the whole process to be
catalyzed by THF and hence to run with a pseudomonomolecular rate constant k that
depends on the THF concentration.
A seemingly facile migration pathway (-route in Scheme 1) would conduct the cation
Li+(THF)4 across the olefinic CH2- region. Favorable electrostatic interactions might make
it more attractive for Li+(THF)4 to migrate across the still charge-bearing, temporarily sp-
hybridized C() center (-route). The -route had been proposed previously[1–5] to proceed
as a “conducted tour”[6] along the charge gradients that resulted through delocalization of
negative electric charge from the Li–C() bond into the -aryl -electron system: Such a
delocalization is at its maximum in the depicted transition state with a lateral overlap
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
(Scheme 1) of the two px orbitals at C() and C(4) which are situated within the C()=C()
double-bond plane; this confers a partial double-bond character to the C(4)–C() bond
(quasi-benzyl-anion resonance[7]) and creates an electronic barrier against deviation from the
close to perpendicular orientation of the -aryl plane with respect to the double-bond plane.
Such a rotational barrier would disfavor the possibility of a transport of Li+(THF)4 riding on a
 face of a rotating -aryl group. The energetic stabilization against -aryl rotation and the
concomitant accumulation of delocalized negative -charge in the para position (at C-9) are
stronger in the linear transition state than in the non-linear ground state 1-CIP; this had been
established[1] through the Hammett reaction constant  = +5.2 for the stereoinversion rate
constants as determined for a related familiy of -arylalkenyllithiums (para-substituted
derivatives of 7a in Scheme 2).
The new -arylvinyllithium 1 was designed to provide more direct evidence for Li+
migration on the -route across the -aryl part: Steric or electronic hindrance by its 2,2-
dimethylpropan-1,3-diyl side-chain (Scheme 1) might raise the cis/trans stereoinversion
barrier compared with that of the less encumbered 1-(-lithiovinyl)naphthalene (2); the
envisioned barriers would be obtained from the cis/trans interconversion rates of the olefinic
CH2- protons in 1 and 2. In order to minimize steric differences, 1 was also designed to
resemble the known[2] -(2,6-dimethylphenyl)vinyllithium (4) sterically in the
neighbourhood of the Li–C() bond: The 2-/6-CH3 substituents of 4 correspond to CH2-3
and CH-5 of 1 wherein they are fixed as members of a cyclopentene and a phenyl ring,
respectively. Since the single unsubstituted ortho-position might influence the barrier of 2,
we searched for a more general answer to this question through rate comparisons of the -(2-
alkylphenyl)vinyllithiums[7] 3 and 5 with the published data of -(2,6-
dialkylphenyl)alkenyllithiums 4[2] and 6.[4] As a basic requirement, 1 and 2 should be
monomeric in THF (as are 3–7). The design of 1 was also thought to reveal the rapidity of -
aryl rotation in the ground state of 1 by means of NMR coalescences (averaging of the 
values) within the signal pairs of geminal protons in CH2-1 and in CH2-3 and of the 2-CH3
substituents in 1. The resulting rotational barriers were expected to keep the aromatic ring
plane in a perpendicular orientation with respect to the C()=C() double-bond plane in 1-
CIP.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
Our first task will consist in identifying the monomeric and dimeric species of 1 and 2
(ground states in Section 1), followed by Section 2 with measurements of the cis/trans
stereointerconversion rates and evaluation of the corresponding stereoinversion barriers in
search of evidence for the “conducted tour” migration of Li+(THF)4 on the -route.
Results and Discussion
1. Ground States of the Alkenyllithiums
The carboxylic acid 10 was prepared from 4-bromo-2,2-dimethylbenz[f]indane[8] (8) via
the lithio derivative 9 and its carboxylation (Scheme 3); it was important to realize that 10
was weakly soluble in both Et2O and aqueous NaOH. The ensuing formation of the 4-acetyl
derivative 11 from 10 with methyllithium (3 equiv) was terminated through quenching with
ClSiMe3 before aqueous workup, so that remnant methyllithium was destroyed before the
ketone 11 was liberated through hydrolysis.[9] Brominative deoxygenation[10] with the
reagent 12 converted 11 into 4-(-bromovinyl)-2,2-dimethylbenz[f]indane (13). The
temperature-dependent NMR spectra of some diastereotopic nuclei in 13 (2-CH3 and CH2-3
in Scheme 3 and Tables[11] S1) disclosed a barrier of G‡(35 and 48 °C) = 15.9(2) kcal mol–1
against -aryl (half-)rotation with passage through a possibly coplanar conformation of the -
aryl and the C()=C() double-bond planes: The responsible obstacle was the -Br atom,
1
since the corresponding -aryl rotations could not be “frozen” on the H and 13C NMR time
scales with 4-CO2H in 10, 4-acetyl in 11 (down to –103 °C), and 4-vinyl in the alkene 14.
1
The assignments of all H and 13C NMR signals of 14 agreed satisfactorily with the
known[12] assignments for 1-vinylnaphthalene (15) which were needed further below.
Preparation of the polymerization-prone 15 through dehydration of 1-(1-naphthyl)ethanol by a
reagent generated from N,N-dimethylformamide with oxalyl chloride can be recommended.
The Br/Li interchange reaction (Scheme 3) of the 4-(-bromovinyl) derivative 13 with n-
butyllithium (nBuLi) in cyclopentane or hexane occurred smoothly at room temperature (rt) in
the absence of electron-pair donor substances (such as Et2O). This caused an initial confusion
because the quickly emerging BrCH2 1H NMR triplet of the co-product 1-bromobutane
(nBuBr) was not accompanied by the expected signals of the =CH2 part of 1 in the olefinic
spectral region that displayed only a very broad yet shallow baseline elevation. But when this
clear solution was treated with tert-butyl methyl ether (tBuOMe) in a slight excess over 13,
the missing =CH2 part of 1 appeared with a the expected intensity of the two doublets and
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
2
with a mutual two-bond coupling constant of JH,H = 6.0 Hz; as a constitutional proof,
protolysis with a little tert-butyl alcohol generated the “parent” alkene 14. This behavior
made it clear that we would not be able to investigate donor-free 1, perhaps due to the
formation of a spectroscopically unresolved mixture of highly aggregated species of 1 in the
absence of donor ligands. Unfortunately, this unsolvated 1 did not precipitate from these
hydrocarbon solvents, so that it could not be purified by the usually applied washing with
pentane. Unlike 4,[13] 1 did also not precipitate during its preparation in tBuOMe as the
solvent; thus, we had to remove nBuBr and all other volatiles by evaporation at rt in a high
vacuum. The solid residue, containing some tBuOMe that was coordinated to 1 (and possibly
to some LiBr), dissolved readily in hydrocarbons or ethereal solvents; these solutions had to
be stored at –20 °C under argon gas cover till the ensuing NMR determinations of the
(non)aggregation and microsolvation states could be carried out. Thus, we investigated
always samples with at least one equivalent of tBuOMe.
An orange-colored solution of isotope-labeled [6Li]1 in [D8]toluene displayed a 13C-
NMR quintet at –55 °C. The intensity ratio of 1:2:3:2:1 of the five quintet components is
characteristic of a 13C6Li2 motif since it results through magnetic coupling of the observed
6
13C nucleus with n = 2 Li nuclei;[14] it can be detected only if the intermolecular Li
scrambling is sufficiently slow to preserve the intramolecular 13C,6Li spin coherence on the
NMR time scale. This primary NMR criterion of aggregation pointed to the four-membered
ring core (C2Li2) of a dimeric structure, as known for the established[13] dimers of 4. The
frequency intervals of the quintet components furnished the one-bond NMR coupling constant
1J(13C,6Li) = 7.5 Hz; by means of the empirical[15] Equation (1), this value revealed
microsolvation at 1 by d = 1 tBuOMe ligand per Li, since a = 2 13C- centers have contacts
6
6
with each Li nucleus in the proposed 13C2 Li2 core of 1. Thus, the [D8]toluene solution
contained the disolvated dimer, namely, (1&tBuOMe)2, in analogy with the known[13] solid-
state structure of (4&tBuOMe)2. This dimer remained the only species of 1 at all
temperatures from 25 to –90 °C, as shown by the practically constant 13C chemical shifts 
(Table[11] S2a). As a secondary NMR criterion, the two-bond interproton coupling constant
2JH,H = 5.8 Hz of the =CH2 moiety remained also unchanged from –55 up to 25 °C
(Table[11] S2b), which provided two important pieces of evidence: By means of the
empirical[16] Equation (2), it confirmed that the microsolvation number of d = 1 in
(1&tBuOMe)2 remained valid at least up to 25 °C and not only at –55 °C as deduced above
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
through 13C splitting. Second, it established that the rapid stereoinversion process to be
described in Section 2 was less rapid (namely, below coalescence) here in toluene at 25 °C.
d = 46 Hz  (n  1JC,Li)–1 – a
d = (2JH,H – 3.9 Hz)/1.67
(1)
(2)
Quite similar NMR spectral features were observed for 1 in its orange-colored tBuOMe
1
solutions: Most of the 13C (Table[11] S3a) and H (Table S3b) chemical shifts  were
numerically close to the above data in toluene and did not depend on the temperature between
25 and –103 °C, so that the same disolvated dimer (1&tBuOMe)2 as above appeared to be
the only species of 1 also in tBuOMe. In spite of the abundant solvent, the well-resolved
2JH,H = 6.0 Hz coupling confirmed [by Eq. (2)] microsolvation by only d = 1 tBuOMe ligand
1
per Li. All 13C and H chemical shifts of this species were assigned through selective
1
decoupling experiments and the two-dimensional H,6Li (HOESY) correlations (short
6
distances) of Li with both 5-H and the upfield olefinic -H (trans to -aryl). These data and
the above assignments for the “parent“ alkene 14 were required for calculations of lithiation
NMR shifts  = (RLi) – (RH) which can purge (RLi = 1) from special effects that are
also present in (RH = 14), making the effects of lithiation more clearly visible. We use them
as another secondary NMR criterion of (non)aggregation: Despite substantially different
chemical shifts ,[11] the characteristic  data of constitutionally comparable positions in
1&tBuOMe and 4&tBuOMe were sufficiently similar (Table 1) to conclude that these two
systems responded similarly to lithiation: These dimers of 1 and 4 have almost equal (C-
para) values in tBuOMe (entries 2 and 5) or in toluene (entries 1 and 4), including also
cyclopentane (entry 3 and Table[11] S4). Such upfield lithiation NMR shifts are partially[7]
caused by the delocalization of electric charge from the Li–C() bond to the para position (as
detailed in references [2] and [7]), so that their similar magnitudes indicated similar
interplanar angles between the -aryl and the C()=C() double-bond planes in dimeric 1 and
4. These magnitudes revealed close to perpendicular conformations that provide for a close to
maximum overlap of the charge-carrying Li–C() bond (quasi-sp2) orbital and the px orbital
at C-ipso in the C()=C() double-bond plane (quasi-benzyl-anion resonance). The small
deviation of (C-para) of monomethyl-bearing, dimeric 3 (entry 6) may be due to a slightly
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
different conformational preference. Complete sets of  values are collected in Figure
S1.[11]
Comparison of the 13C NMR spectrum of dimeric 1 in tBuOMe (Figure 1a) with that in
Figure 1b established that 1 cannot be dimeric in THF solution. The dark-green color of THF
solutions of 1 seemed to be due to traces of contaminants that did not disturb the NMR
1
properties of 1; this followed from the well-resolved 13C and H NMR spectra at and below
–30 °C, whose chemical shift values did not depend on the temperature (Tables[11] S5a and
S5b). Accordingly, protolytic quenching led to decolorization and afforded the parent alkene
14. Adventitiously generated 14 in small amounts vanished soon through polymerization (in
contrast to the stable solutions of 14 with 1 in the non-THF solvents). Since 13C- splitting
by 6Li (as mentioned above) could not be detected down to –103 °C, we took again recourse
to the secondary criterion of lithiation shifts : The strongly downfield-shifted, practically
equal (C-) = +76.3(1) ppm of both 1 and 4 (entries 7 and 8 in Table 1) suggested that 1
was also a trisolvated monomer in THF. In fact, microsolvation by d = 3 THF ligands was
2
confirmed through Equation (2) by the interproton coupling constant of JH,H = 8.9 Hz (entry
7). The almost equal magnitudes of both (C-para) = –11.3(3) ppm and (para-H) = –
0.74(2) ppm for 1&3THF (entry 7) and 4&3THF (entry 8) were typically greater than those of
dimeric 1 in entries 1–3 and bear witness to the characteristically increased  delocalization in
the ground states of our monomeric, trisolvated alkenyllithiums. As a consequence of this
strengthened quasi-benzyl-anion resonance, the barrier against rotation about the C()–
C(ipso) bond must be higher for the monomers than for the dimers. This expectation was
experimentally verified through observation of separate NMR signals of the diastereotopic
protons (Scheme 3) in CH2-1, CH2-3, and 2,2-(CH3)2 at 400 MHz between –85 and –103 °C
(Table[11] S5b) for monomeric but not for dimeric 1. The pairwise averaging of these signals
occurred through sufficiently rapid -aryl rotation in monomeric 1 at individual coalescence
temperatures and furnished the rotational barrier of G‡(–60 and –66 °C) = 9.9(2) kcal mol–1,
as detailed for 1 in the Experimental Section. Since deviations from the close to
perpendicular conformations against such barriers would generally diminish[7] the 
delocalization and hence sacrifice part of the appertaining energetic stabilization, even the
monomers of the two -(2-alkylphenyl)vinyllithiums 3 and 5 tended to preserve such a
conformation in spite of their unsubstituted 6-positions. This may be visualized by the
known[17] lithiation shifts (C-para) = –10.5(1) ppm of 3 and 5 in THF (entries 10 and 11)
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
whose slightly diminished magnitudes may be due to a low tendency to populate a dimeric
2
equilibrium component, in accord with their slightly reduced coupling of JH,H = 8.0 Hz.
1-(-Lithiovinyl)naphthalene (2) precipitated after its generation through a Br/Li
interchange reaction in pentane and could be purified through washing with pentane. Its
constitution was established through protolysis which gave the parent alkene 15 with the same
1H NMR spectrum as that obtained with the initially mentioned, independently synthesized[11]
2
sample. In THF as the solvent, 2 displayed interproton CH2- coupling constants of JH,H
=
9.0 Hz at and below rt (Table S6[11] and entry 9 of Table 1), which was translated by means
of eq (2) into the microsolvation number d = 3 THF ligands per Li. This trisolvated ground
state of 2&3THF is monomeric, as shown by its magnitude of (C-) = +76.2 ppm which
equals that of the trisolvated monomer of 1&3THF (entries 9 and 7). In spite of its single
unsubstituted ortho position, 2&3THF appeared to maintain the close to perpendicular -aryl
conformation in analogy with 1&3THF, judged by (C-para) = –11.3 ppm. Many of the
further  values (Figure[11] S1) were also similar for 1&3THF and 2&3THF, which
suggested that  delocalization toward the para position was hardly influenced by the 2,2-
dimethylpropan-1,3-diyl side-chain in 1.
This section demonstrated that the ground states of 1 and 2 are trisolvated monomers in
THF and that close to perpendicular conformations are common for their monomers and for
the dimer of 1. The -aryl group of monomeric 1-CIP rotates across a barrier of G‡(–63 °C)
= 9.9(2) kcal mol–1 that includes the transitory sacrification of electron delocalization from
the Li–C() bond into the naphthyl  system. The corresponding rotational barrier of dimeric
1 is lower (albeit unknown) due to the generally weaker -electron delocalization in dimers;
quantitative information on such an effect was reported[4] for the ground states of monomeric
1
and dimeric 6. The dimeric species of 1, 3, and 4 showed no H NMR evidence for the
cis/trans stereoinversion that is described for the monomers in the followin section.
2. A “Conducted Tour” Migration of Li+(THF)4 on the -Route
Due to severe obstruction of the - and -routes, the -route across an -phenyl group was
the most reasonable possibility of Li+ migration in the derivatives of 2-(-lithiobenzylidene)-
1,1,3,3-tetramethylindane (7a in Scheme 2), which was the original system that provided the
discovery[15] of Equation (1) and the deduction[1] of the ionic stereoinversion mechanism as
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
illustrated in Scheme 1. For the planned search for migration-impeding substituents,
however, this system was unsuitable because two ortho-methyl substituents in 7b accelerated
the stereoinversion process instead of impeding it: The free pseudoactivation enthalpy barrier
–1
‡
at 0 °C, G (0 °C), of 7b (entry 2 of Table 2) was lower by 0.9 kcal mol than that of 7a
(entry 1). This resulted through intramolecular repulsive interactions[1] by the 1,1,3,3-
tetramethyl substituents (energetic elevation of the ground state) and was also perceptible by
the bigger barrier reduction from 16.3 kcal mol–1 (entry 8) in -(2,6-
dimethylphenyl)vinyllithium (4) to 13.35 kcal mol–1 for 7a (entry 1). A comparable
inversion-accelerating effect (3 – 5 kcal mol–1) of the bulky 1,1,3,3-tetramethyl armament had
previously[18] been noted with N-phenyl imines that are sterically equivalent and isoelectronic
models of the inverting carbanionic moieties in 4 and 7a. Therefore, it became necessary to
get rid of the accelerating front strain in 7a,b and to use barriers of the less encumbered
vinyllithiums such as 4 as references for the wanted substituent effects at the -aryl groups.
This choice restricted our NMR-based rate measurements to the cis and trans -H atoms that
are constitutionally equivalent but stereochemically different (namely, “diastereotopic”):
With increasingly rapid cis/trans interconversion (“diastereotopomerization”), the four
1
absorption lines of this H2C=C H NMR spectrum will first suffer increasing line broadening
and then “coalesce” into a singlet at the averaged resonance position of (cis + trans)/2.[19]
Computer-aided simulations[20] of these rate-dependent line shapes can furnish the pseudo-
first-order rate constants k of the pseudomonomolecular (namely, THF-catalyzed) cis/trans
stereoinversion. These k values were used for obtaining linear plots of lnk versus 1000/T,
‡
as exemplified for 2&3THF in Figure 2, and furnished the pseudoactivation parameters G
‡
‡
= H – TS of this stereoinversion process as listed in Table 2.
Internal repulsion was indeed much less of a problem in the ground states of our -
arylvinyllithiums 1 – 6, as shown by the hardly lower barrier (entry 5: 15.79 kcal mol–1)[4] of
-(2,6-diisopropylphenyl)vinyllithium (6) in comparison with that of 4 (entry 8). Internal
repulsion was not totally absent, however, as revealed by a substantially lower barrier[5] of -
(2,6-di-tert-butylphenyl)vinyllithium. In spite of the significant differences of 1 – 7b in their
structural details and barriers, the pseudoactivation entropies in entries 1 – 3 and 5 – 8 fall into
the narrow range of S = –23  3 cal mol–1 K–1, which is characteristic of the ionic
‡
mechanism and includes a contribution from the transitory immobilization of the fourth THF
ligand that is required for the two THF-separated ion pair intermediates (SSIP) and the
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
‡
transition state in analogy with Scheme 1. The S value of 1 was also strongly negative,
but it was not included in entry 4 because it was numerically less reliable due to the
‡
competing decomposition of 1 in THF. Nevertheless, the magnitude of G (0 °C) of 1 was
reliable since not extrapolated but calculated directly from the k value[11] measured at 0 °C
(Table S5c). Comparison of 4 with its 2-methyl congener 3 (entries 8 and 7, Table[11] S7)
established that the presence of the second ortho-methyl substituent in 4 hardly changed the
barrier. This appeared to apply also to the presence of the second ortho-isopropyl substituent
in 6 (entries 5 and 6) since it must be admitted that the mono-isopropyl congener 5 apparently
began to dimerize in THF at the exceptionally high temperatures of up to 92 °C (a common
trend of the usually[21] endothermic dimerizations). This became visible for 5 in THF through
the published[22] concentration dependence of the rate constants k at 88 °C and the
‡
extrapolated, apparent G (0 °C) values that increased with increasing concentrations of 5
(footnotes “b” in Tables[11] S8–S10). Although measured with the most dilute (most
deaggregated) THF solution of 5 (0.3 M), the barrier denoted in entry 6 is probably still a little
too high, so that the second ortho-isopropyl group in 6 hardly changed the barrier observed
for 5. On the basis of such an insignificant influence of additional ortho-alkyl substituents,
the 0.93 kcal mol–1 higher barrier of 1 (entry 4) relative to that of 1-(-
lithiovinyl)naphthalene (2 in entry 3) cannot be due to the ortho-positioned CH2-3 group in 1.
This left the meta-positioned CH2-1 group as the responsible obstacle that impeded Li+
migration; analogously, a preferred passage across the meta-position around ortho-positioned
tert-butyl groups was already suspected previously.[23] As outlined in Section 1 and shown by
the (para) values in entries 7 – 9 of Table 1, the 2,2-dimethylpropan-1,3-diyl side chain
appeared to have almost no electronic influence on the relevant charge gradient toward the
para position. Therefore, this barrier difference between 1 and 2 is ascribed to a steric effect
of the CH2-1 part of that side chain. Such a retardation by a meta-positioned substituent
would not be expected on the - or -routes, which leaves the -route as an unrefuted
possibility.[24]
This section demonstrated the possibility of Li+(THF)4 migration on the -route in 1-(-
lithiovinyl)naphthalene (2) during cis/trans stereoinversion by the pseudomonomolecular,
ionic mechanism: Even without the -route-obstructing 1,1,3,3-tetramethyl substituents of
7a,b and despite an inviting unobstructed -route in 2, the migrating cation preferred to be
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
conducted along the charge gradients on the -route, presumably with the aim of surmounting
the aryl rim in the meta/para region.
Conclusion
The thermal ground state of 4-(-lithiovinyl)-2,2-dimethylbenz[f]indane (1) is a disolvated
dimer in tBuOMe and other non-THF solvents but a trisolvated monomer in THF. The
delocalization of electric charge from the Li–C() bond into the -aryl part of the monomer
provides for a close to perpendicular orientation of the aryl and the C()=C() double-bond
planes with a 9.9(2) kcal mol–1 barrier against aryl rotation about the C()–C(4) bond. These
properties qualified 1 to serve for the following investigations.
–1
‡
The stereoinversion barriers of G (0 °C)  16 kcal mol in THF are almost equal (Table
2) for -arylvinyllithiums with 2,6-dimethylphenyl (4), 2-methylphenyl (3), 2,6-
diisopropylphenyl (6), and 2-diisopropylphenyl (5) groups. Thus, one unsubstituted ortho
‡
position (in 3 and 5) did not lower the G (0 °C) barrier as far as the ground states were
truly monomeric. (5 tended perceptively[22] to populate its dimeric species at the higher
temperatures and concentrations.) At first sight, this might be taken as evidence for the - or
-route of Li+(THF)4 migration and would then imply that these routes were hardly sensitive
to a second small ortho-alkyl substituent and certainly even less so to small meta-alkyl
–1
groups. If so, the increase from G (0 °C) = 12.94(1) kcal mol for 2 to 13.87(1) kcal mol–
‡
1
for 1 would be unexpected. On the other hand, a -route may also be hardly sensitive to a
second small ortho-alkyl group but can of course be expected to sense obstruction by a small
meta-positioned alkyl substituent, as observed for 1. We conclude that the -route is an
appropriate stereoinversion pathway for 2 (as it was for 7a and 7b); there appears to be no
evidence against the -route along the similar -charge gradients (“conducted tour”) in other
-arylvinyllithiums.
However, a corresponding deduction for 1 would expect that an even more migration-
impeding derivative of 1 should exhibit a still higher stereoinversion barrier; if not, both 1
and its prospective derivative might be suspected to have changed over to Li+(THF)4
migration across the - or -regions.
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
Experimental Section
General. The utilized techniques and spectrometers were desribed previously.[15] -
Arylvinyllithiums 1 – 5 were studied either in commercial [D8]toluene or in nondeuteriated,
1
anhydrous ethereal solvents containing up to ca. 15% (v/v) of dry [D12]cyclohexane. H and
13C NMR shifts were referenced to internal SiMe4. NMR abbreviations were d = doublet, m =
multiplet, q = quartet, qi = quintet, quat = quaternary, sept = septet, t = triplet.
4-(-Lithiovinyl)-2,2-dimethylbenz[f]indane (1&tBuOMe). A dry NMR tube (5 mm)
was charged with the bromoalkene 13 (50 mg, 0.17 mmol) and anhydrous tBuOMe (0.5 mL).
This solution was cooled under argon gas cover to –30 °C and treated with nBu6Li (0.18
mmol) in cyclopentane (0.122 mL) or with unlabelled nBuLi (0.18 mmol) in hexane (0.079
mL). The open NMR tube with the yellow solution was warmed up to rt and immediately
placed in a large Schlenk flask (still and always unter argon gas) and fixed in a tilted position
for the cautious, very slow evaporation of the solvent and all volatiles at rt down to a pressure
of ca. 0.01 mbar. After the slow admission of dry argon gas, the non-volatile residue was
cooled and dissolved in the chosen solvent with a trace of SiMe4 (and [D12]cyclohexane, if
1
necessary), then stored at –20 °C and kept only briefly at higher temperatures. H NMR of
dimeric 1 ([D8]toluene, 400 MHz, 25 °C)  1.21 (s, 6H, 2  2-CH3), 2.82 (s, 2H, CH2-3), 2.85
(s, 2H, CH2-1), 6.20 (broadened d, 2J = 5.8 Hz, 1H, -H trans to -aryl), 6.50 (d, 2J = 5.8 Hz,
1H, -H cis), 7.24 (s, 1H, 9-H), 7.24 (obscured t, 1H, 6-H), 7.29 (t, 3J = 7.3 Hz, 1H, 7-H),
7.71 (d, 3J = 7.8 Hz, 1H, 8-H), 8.34 (d, 3J = 8.0 Hz, 1H, 5-H) ppm, assigned through
1
decoupling (see 13C) and comparison with 1 in tBuOMe; H NMR of dimeric 1 (tBuOMe,
400 MHz, 25 °C)  ca. 1.14 (obscured s, 2  2-CH3), 2.70 (s, 2H, CH2-3), 2.77 (d, 4J  1.0
Hz, CH2-1), 5.89 (broadened d, 2J = 6.0 Hz, 1H, -H trans to -aryl), 6.14 (d, 2J = 6.0 Hz, 1H,
-H cis), 7.06 (s, 1H, 9-H), 7.08 (td, 3J = 8.1 Hz, 4J = 1.1 Hz, 1H, 6-H), 7.14 (td, 3J = 8.1 Hz,
4J = 1.1 Hz, 1H, 7-H), 7.52 (d, 3J = 7.5 Hz, 1H, 8-H), 8.07 (d, 3J = 8.1 Hz, 1H, 5-H) ppm,
assigned through HOESY cross peaks of 6Li with 5-H and trans-H (and also with both solvent
signals) and confirmed through selective {1H} decoupling as follows: {5-H}  6-H as a d ca.
1
8 Hz, and {8-H}  7-H as a d ca. 7 Hz; H NMR of monomeric 1 (THF, 400 MHz, –85 °C)
 1.01 and 1.18 (2  br s, 3+3H, 2  2-CH3), 2.44 and 2.68 (AB system, 2  d, 2J = 15 Hz,
1+1H, CH2-3), 2.62 and 2.76 (AB system, 2  d, 2J = 14 Hz, 1+1H, CH2-1), 5.32 (d, 2J = 8.9
Hz, 1H, -H trans to aryl), 5.75 (d, 2J = 8.9 Hz, 1H, -H cis), 6.75 (s, 1H, 9-H), 6.93 (t, 3J =
7.4 Hz, 1H, 6-H), 7.04 (t, 3J = 7.4 Hz, 1H, 7-H), 7.40 (d, 3J = 7.5 Hz, 1H, 8-H), 8.00 (d, 3J =
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
7.6 Hz, 1H, 5-H) ppm, compare Table 5b for the temperature dependence; 13C NMR of
dimeric 1 ([D8]toluene, 100.6 MHz, 25 °C)  29.4 (qm, 1J = 125 Hz, 2  2-CH3), 39.8
1
(unresolved, C-2), 48.5 (t, 1J = 128 Hz, CH2-3), 48.6 (t, 1J = 128 Hz, CH2-1), 116.2 (ddm, J
= 156 Hz, C-9), 120.5 (broadened t, 1J = 147 Hz, CH2-), 122.7 (dd, 1J = 158 Hz, 3J = 8 Hz,
C-6), 124.1 (obscured dd, 3J = 8 Hz, C-7), 127.1 (obscured dd, 3J = 6 Hz, C-5), 128.1
(obscured dt, 3J = 6 Hz, C-8), 129.3 (q, 3J = 6 Hz, C-4a), 130.6 (dm, 3J = 6 Hz, C-3a), 134.8
(t, 3J = ca. 6 Hz, C-8a), 142.4 (broad, C-9a), 151.6 (dm, 3J = 15 Hz, C-4 coupling with -H
trans), 201.1 (unresolved, C-, at –55 °C as a qi 1:2:3:2:1 with 1JC,Li = 7.5 Hz) ppm, assigned
through the JC,H coupling patterns and selective {1H} decoupling as follows: {2-CH3}  C-
2
3
3
2 as a qi with J = ca. 2 Hz; {CH2-1/-3}  C-3a as a d with J = 6 Hz, C-9 as a dd with J =
3
5 Hz, C-9a as a s; {-H trans}  C-4 and C- narrowed; {-H cis}  C-4 as a d with J =
3
14 Hz; {6-/7-/9-H}  C-4a as a s, C-8a as a d with J = 4.5 Hz; {8-H}  C-4a and C-8a
1
1
3
narrowed, C-6 as a J d; {5-H}  C-7 as a J d, C-8a as a d with J = 6 Hz; 13C NMR of
dimeric 1 (tBuOMe, 100.6 MHz, 25 °C)  29.5 (qm, 1J = 125 Hz, > qi with 3J = 4.5 Hz, 2 
2-CH3), 40.0 (m, 2J = 3 Hz, C-2), 48.5 (obscured, CH2-3), 48.8 (obscured, CH2-1), 115.7 (dm,
1J = 156 Hz, C-9), 119.6 (t, 1J = 148 Hz, CH2-), 122.4 (dd, 1J = 158 Hz, 3J = 8.5 Hz, C-6),
124.0 (dd, 1J = 158 Hz, 3J = 8.8 Hz, C-7), 127.2 (dd, 1J = 157 Hz, 3J = 7.0 Hz, C-5), 127.9 (dt,
1J = 156 Hz, 3J = 6.0 Hz, C-8), 129.3 (q, 3J = 6.1 Hz, C-4a), 130.2 (unresolved, C-3a), 134.8
(dd, 3J = 7.5 Hz and ca. 5 Hz, C-8a), 142.3 (m, apparent J = 4.3 Hz, C-9a), 152.0 (tm, 3J =
12.8 Hz, C-4), 201.3 (unresolved, C-) ppm, assigned through the JC,H coupling patterns and
selective {1H} decoupling as follows: {CH2-1/-3}  C-3a as a 3J d, C-9a as a sharp s; {-H
trans}  C-4 and C- narrowed; {-H cis}  C- narrowed, C-4 less narrowed; {6-/7-/9-
1
3
H}  C-4a narrowed, C-5 as a sharp J d, C-8a as a J d; {8-H}  C-4a narrowed, C-6 as
1
a sharp J d; {5-H}  C-7 as a sharp 1J d, C-8a as a broad 3J d; 13C NMR of monomeric 1
(THF, 100.6 MHz, –55 °C)  29.5 (q, 2  2-CH3), 39.9 (quat, C-2), 48.0 (t, CH2-3), 48.9 (t,
CH2-1), 111.2 (d, CH-9), 111.4 (t, CH2-), 120.6 (d, C-6), 123.4 (d, CH-7), 126.1 (quat, C-
4a), 127.2 (d, CH-5), 128.4 (d, CH-8), 128.6 (quat, C-3a), 134.6 (quat, C-8a), 142.2 (quat, C-
9a), 157.3 (quat,C-4), 210.9 (quat, C-) ppm, assigned through DEPT and comparisons with 1
in tBuOMe, compare Table 5a for the temperature dependence; -aryl rotation of monomeric
1 in THF: The pairwise geminal protons of CH2-1 and CH2-3 and the diastereotopic 2-CH3
groups displayed separate chemical shifts at –85, –94, and –103 °C (Table S5b). These 400
13
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
MHz spectra of CH2-1 with  = 60 Hz and 2J = 14 Hz, of CH2-3 with  = 96 Hz and 2J =
15 Hz, and of 2-CH3 with  = 72 Hz and 2J = 0 Hz coalesced at estimated temperatures of
–66(3), –60(5), and –66(3) °C, respectively. The rate constants kc at these temperatures were
calculated from kc = (2 + 6J2)0.5/20.5, yielding the rotational G barriers of 9.9(2),
9.9(2), and 10.0(3) kcal mol–1, respectively.
1-(-Lithiovinyl)naphthalene (2). 1-(-Bromovinyl)naphthalene[25] (4.45 g, 19.1 mmol)
and pentane (5 mL) were stirred in a Schlenk flask at –78 °C under argon gas cover during
the slow introduction of nBu6Li (22.9 mmol) in pentane (18.4 mL). The Br/Li interchange
reaction was almost complete after 25 min at rt, whereupon the product 2 began to precipitate
completely within less than 4 hours. The supernatant was withdrawn by pipette, and the
precipitate was washed with dry pentane (2), then evaporated to dryness at rt in vacuo:
1
yield, 2.0 g (66%). H NMR (THF, 80 MHz, –62 °C)  5.20 (d, 2J = 9.0 Hz, 1H, -H trans
to -aryl), 5.75 (d, 2J = 9.0 Hz, 1H, -H cis), 6.37 (dd, 3J = 6 Hz, 4J = 1.2 Hz, 1H, 2-H), ca.
7.0 and ca. 7.3 (2  obscured m, 3-/4-/5-/6-/7-H), 8.09 (m, 1H, 8-H) ppm, assigned through
comparison with 1, 3 and 5, see Table S6 for the temperature dependence; 13C NMR (THF,
25.15 MHz, –47 °C)  112.0 (t, CH2-), 113.3 (d, C-2), 116.7 (d, C-4), 121.6 (dd, C-7), 124.3
(d, C-6), ca. 126.2 (d, C-3), 126.7 (d, C-8), 127.2 (d, C-5), 130.2 (quat, C-8a), 134.2 (quat, C-
4a), 163.2 (quat, C-1), 210.5 (quat, C-) ppm, assigned through gated and off-resonance
decoupling and comparison with 1; no temperature and concentration dependence detected
between –38 and –96 °C.
2,2-Dimethylbenz[f]indan-4-carboxylic Acid (10). nBuLi (12.2 mmol) in hexane (4.4
mL) was added dropwise to a stirred solution of 4-bromo-2,2-dimethylbenz[f]indane[8] (8,
3.28 g, 11.9 mmol) in anhydrous Et2O (20 mL) at –30 °C under inert gas cover. The mixture
change its color from green to yellow, then orange-red during the addition. After 10 min at rt,
the solution was poured onto solid CO2 and warmed up to rt. The yellow mixture was
completely dissolved in Et2O and a sufficient amount of conc. (6 M!) aqueous NaOH. The
aqueous layer was washed with Et2O (to be discarded), then acidified with conc. HCl, and
extracted with CH2Cl2 (3 ; not Et2O!). The combined CH2Cl2 extracts were washed with
distilled water until neutral, dried over Na2SO4, filtered, and evaporated to yield the crude
acid 10 (2.30 g, 80%). The analytically pure sample was obtained through recrystallization
first from chloroform, then from toluene, and a final sublimation at 150–170 °C (bath
1
temp.)/0.15 mbar; mp 188–190 °C. H NMR (CDCl3, 80 MHz)  1.20 (s, 6H, 2  2-CH3),
14
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
2.88 (s, 2H, CH2-3), 3.20 (s, 2H, CH2-1), 7.60 (m, 2H, 7-/6-H), 7.71 (s, 1H, 9-H), ca. 7.78 (m,
1H, 8-H), 8.60 (m, 1H, 5-H), 11.90 (broad s, 1H, CO2H) ppm, assigned through comparison
with the alkene 14; IR (KBr)  3300–2400 (broad O–H), 3056, 2961, 2899, 2866, 1689 (s),
1631 (w), 1425, 1248, 1217, 746 cm–1. Anal. calcd for C16H16O2 (240.30): C, 79.97; H, 6.71.
Found: C, 80.65; H, 6.62.
4-Acetyl-2,2-dimethylbenz[f]indane (11). Methyllithium (12.5 mmol) in Et2O (8.67 mL)
was added dropwise to a stirred suspension of the acid 10 (1.00 g, 4.16 mmol) in anhydrous
Et2O (7.0 mL) at –20 °C under argon gas cover. As expected, the brisk evolution of CH4
ceased after the addition of the initial third (1 equiv) of methyllithium. The finally orange-
colored solution became red during one hour of refluxing at 40 °C, whereafter it was re-
cooled to –10 °C and treated with ClSiMe3 (0.63 mL, 5.00 mmol) and then stirred at rt for
one hour. The resultant, colorless suspension was cooled in ice during the addition of water
and Et2O (40 mL). The Et2O layer was shaken with aqueous NaOH (2 M, 15 mL), then
distilled water (2  10 mL), and then kept waiting for further workup. The combined aqueous
extracts were shaken with Et2O (15 mL, also kept waiting), acidified with conc. HCl and
shaken with Et2O (3 ). The latter three Et2O phases were combined, washed with distilled
water until neutral, dried over MgSO4, filtered, and evaporated to leave the unconsumed
portion of the acid 10 (130 mg, 13%). The two waiting Et2O layers were combined, dried
over MgSO4, filtered, and concentrated to yield the non-acidic product fraction (830 mg, <
84%). Pure 11 crystallized from low-boiling petroleum ether (2 ) with mp 58–60 °C (486
1
mg, 49%). H NMR (CDCl3 + CH2Cl2 1:2, 400 MHz, 25 °C)  1.16 (s, 6H, 2  2-CH3),
2.60 (s, 3H, acetyl-CH3), 2.85 (s, 2H, CH2-3), 2.86 (d, 2J = 1.3 Hz to 9-H, 2H, CH2-1), 7.42
(m, 2H, 6-/7-H), 7.66 (unresolved, 1H, 9-H), 7.77 (m, 2H, 5-/8-H) ppm, assigned through the
4J splitting of CH2-1 and comparison with the alkene 14; at –103 °C:  1.15, 2.69, 2.88,
2.91, 7.49,7.74, 7.84 (8-H) and 7.88 (d, 1H, 5-H) ppm; 13C NMR (CDCl3 + CH2Cl2 1:2,
100.6 MHz, 25 °C)  28.3 (q of octets, 1J = 125 Hz, 3J = 4.3 Hz, 2  2-CH3), 32.3 (sharp q,
1J = 128 Hz, acetyl-CH3), 41.0 (> sept, 2J = 3.7 Hz, C-2), 47.07 (tm, 1J = 130 Hz, CH2-1),
47.13 (tm, 1J = 130 Hz, CH2-3), 124.6 (dm, 1J = 158 Hz, C-5), 124.9 (ddt, 1J = 159 Hz, 3J =
5.0 and 2.5 Hz, C-9), 125.5 (dd, 1J = 160 Hz, 3J = 8.5 Hz, C-7), 126.1 (dd, 1J = 160 Hz, 3J =
8.5 Hz, C-6), 128.2 (dm, 1J = 158 Hz, C-8), 128.7 (m, 3J = ca. 5 Hz, C-4), 133.4 (m, C-4a),
135.5 (m, C-8a), 139.8 (dtm, 3J = 7.5 and 5.5 Hz, C-3a), 143.1 (tt, 3J = 5.5 Hz, 2J = 3.5 Hz, C-
9a), 205.9 (q, 2J = 5.8 Hz, CO) ppm; at –103 °C:  28.1, 32.7, 41.3, 46.3, 47.0, 124.4,
15
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
125.3, 125.5, 126.2, 128.0, 128.1, 132.7, 134.2, 140.6, 143.1, 207.2 ppm, assigned through
comparison with the alkene 14, the JC,H coupling patterns, and the following selective {1H}
decoupling experiments: {2-CH3}  2-CH3 as a qi with 3J = ca. 4 Hz to both CH2; {CO–
CH3}  CO as a s; {CH2-1/-3}  2-CH3 as a qq with 3J = 4.5 Hz, C-2 as a septet with 2J = 4
Hz, C-3a as a d with 3J = 7.5 Hz, C-9 as a dd with 3J = 5.2 Hz to 8-H, C-9a as a s; {6-/7-H}
 C-8 as a dd (3J to 9-H); {9-H}  C-3a as a t with 3J = 5 Hz, C-4a narrowed; {5-/8-H} 
C-3a as a t, C-4a narrowed, C-5 and C-8 as two narrow 3J t; IR (KBr)  3051 (w), 2954,
2929, 2864, 1683, 1464, 1423, 1364, 1220, 1204, 1154, 774, 753 cm–1. Anal. calcd for
C17H18O (238.33): C, 85.67; H, 7.61. Found: C, 85.84; H, 7.65.
4-(-Bromovinyl)-2,2-dimethylbenz[f]indanen (13). A solution of 2-bromo-1,3,2-
benzodioxaphosphole[10] (0.882 mL, 6.97 mmol) in anhydrous chloroform (1.5 mL) was
cooled in an ice bath under inert gas cover and stirred during the dropwise addition of
elemental bromine (0.321 mL, 6.27 mmol). After further stirring at rt for ca. 25 min, the
resultant solution of 2,2,2-tribromo-1,3,2-benzodioxaphosphole (12) was stirred in an ice bath
and treated dropwise with a solution of 4-acetyl-2,2-dimethylbenzo[f]indane (11, 830 mg,
3.48 mmol) in anhydrous chloroform (0.5 mL). The mixture was stirred at rt for at least 15
min for the complete consumption of ketone 11, then diluted with pentane (4 mL) and poured
into ice-cold aqueous NaOH (2 M, 15 mL). After not more than 15 min at rt (so to avoid the
hydrolysis of 13 that would regenerate ketone 11), the aqueous layer was extracted with
pentane (2  5 mL) and then discarded. The three pentane phases were combined and shaken
with aqueous NaOH (3  5 mL, to be discarded) and then washed with distilled water until
neutral, dried over MgSO4, filtered, and concentrated. The crude material (857 mg) was
filtered through a column of silica gel (6 cm, 10 g) with petroleum ether (80 mL) to give
almost pure 13 (411 mg, 39%) that crystallized from low-boiling petroleum ether with mp 71–
1
73 °C. H NMR (CDCl3, 400 MHz, 25 °C, Table S1b)  1.17 (s, 6H, 2  2-CH3), 2.83 and
2.88 (AB system, 2J = 16 Hz, 2H, CH2-3), 2.87 (d, 4J = 1.2 Hz, 2H, CH2-1), 5.83 (d, 2J = 1.4
Hz, 1H, -H cis to -aryl), 6.10 (d, 2J = 1.4 Hz, 1H, -H trans), 7.40 (ddd, 3J = 8.1 and 6.9
Hz, 4J = 1.5 Hz, 1H, 7-H), 7.46 (ddd, 3J = 8.3 and 6.9 Hz, 4J = 1.5 Hz, 1H, 6-H), 7.59
(unresolved, 9-H), 7.75 (d, 3J = 8.1 Hz, 1H, 8-H), 8.04 (d, 3J = 8.3 Hz, 1H, 5-H) ppm,
4
assigned through the J coupling of CH2-1 with 9-H, the coupling of trans-H with C-4 (see
below), and comparison with the alkene 14; 13C NMR (CDCl3, 100.6 MHz, 25 °C, Table
S1a)  28.32 and 28.44 (2  qm, 1J = 125 Hz, 3J = ca. 4 Hz, 2  2-CH3), 40.3 (> nonet, 2J =
16
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
3.6 Hz, C-2), 46.6 (tm, 1J = 130 Hz, CH2-1), 47.6 (t, 1J = 129 Hz, CH2-3), 122.2 (sharp dd,
1
apparent J = 159 and 165 Hz, CH2-), 123.8 (dm, 1J = 158 Hz, C-9), 124.8 (dd, 1J = 159 Hz,
3J = 6.9 Hz, C-5), 125.2 (dd, 1J = 160 Hz, 3J = 8.4 Hz, C-7), 125.4 (dd, 1J = 160 Hz, 3J = 8.2
Hz, C-6), 126.6 (t, 2J = 6.5 Hz, C-), 127.7 (dt, 1J = 158 Hz, 3J = 6.0 Hz, C-8), 129.7 (q, 3J =
6.7 Hz, C-4a), 133.1 (unresolved, C-4 and C-8a, separated at –55 °C), 141.4 (unresolved, C-
3a), 142.5 (tt, 3J = ca. 5 Hz, 2J = ca. 3.5 Hz, C-9a) ppm, assigned through comparison with the
2
alkene 14 and selective {1H} decoupling as follows: {2-CH3}  C-2 as a qi with J = 3.2
Hz, CH2-1 as a tq with 3J = 3.0 Hz but CH2-3 as a tt with 3J = 2.6 Hz; {CH2-1/-3}  C-2 as a
3
3
septet with 2J = 3.7 Hz, C-3a as a d with J = 7.4 Hz, C-9 as a dd with J = ca. 4.8 Hz, C-9a
as a s, CH2-1 as an octet with 3J = 4.0 Hz but CH2-3 as septet with 3J = 4.3 Hz; { -H trans}
 C-4 narrowed, C- as a d wih 2J = 4.8 Hz; {-H cis}  C- as a d with 2J = 5.5 Hz; {6-
3
/7-/8-/9-H}  C-4a as a s, C-8a as a d with J = ca. 5.5 Hz to 5-H; {8-H}  off-resonance
decoupling increased in the sequence of 5-/8-/9-/6-/7-H; {5-H}  C-6 as a dd, C-7 as a sharp
d, C-8a as a d with 3J = ca. 5 Hz, off-resonance decoupling decreased in the sequence of 5-/8-
/9-/6-/7-H; IR (KBr)  3060 (w), 2954, 2927, 1629, 1423, 1363, 1107, 895, 741 (s) cm–1.
Anal. calcd for C17H17Br (301.23): C, 67.78; H, 5.69; Br, 26.53. Found: C, 67.73; H, 5.52;
Br, 27.38. -Aryl rotation in CDCl3: The diastereotopic protons of CH2-3 only displayed the
anticipated two doublets with  = 21.5 Hz and 2J = 16 Hz at rt under 400 MHz; this AB
system coalesced at 48(3) °C. The 2-CH3 groups showed only the expected 13C
nonequivalence ( = 13.4 Hz at 100.6 MHz) that coalesced at 35(3) °C. The rate constants
kc at 48 and 35 °C were calculated from kc = (2 + 6J2)0.5/20.5, yielding the rotational G‡
barriers of 15.9(2) and 16.0(1) kcal mol–1, respectively.
4-Vinyl-2,2-dimethylbenz[f]indane (14). The crude protolysis products of several samples
of 1 were combined and prepurified through distillation at 85–100 °C (bath temp.)/0.01 mbar,
followed by adsorption on a column (5 cm) of silica gel (2 g, 100–200 m) and elution with
low-boiling petroleum ether (forerun of 7 mL discarded). The eluate (67 mg) was crystallized
1
from EtOH at –70 °C to yield pure 14 (53 mg) with mp 27–28 °C. H NMR (CDCl3, 400
MHz, 25 °C)  1.16 (s, 6H, 2  2-CH3), 2.86 (d, 4J = 1.1 Hz, 2H, CH2-1), 2.91 (s, 2H, CH2-
3), 5.51 (dd, 3J = 17.8 Hz, 2J = 1.9 Hz, 1H, -H cis to -aryl), 5.67 (dd, 3J = 11.5 Hz, 2J = 1.9
Hz, 1H, -H trans), 7.19 (dd, 3J = 17.8 Hz, 3J = 11.5 Hz, 1H, -H), 7.38 (m, 1H, 7-H), 7.40
(m, 1H, 6-H), 7.54 (s, 1H, 9-H), 7.74 (m, 1H, 8-H), 8.06 (m, 1H, 5-H) ppm, at –65 °C 
17
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
4
1.15, 2.88, 2.93, 5.53, 5.72, 7.28, 7.46, 7.47, 7.60, 7.81, 8.08, assigned through the J
coupling of CH2-1 with 9-H, DQCOSY, and the following NOESY correlations: 1-H  9-H
 8-H  7-H, 3-H  -H  5-H  6-H; 13C NMR (CDCl3, 100.6 MHz, 25 °C)  28.5
(q of octets, 1J = 125 Hz, 3J = 4.5 Hz, 2  2-CH3), 40.3 (> septet, 2J = 3.6 Hz, C-2), 47.6 (tm,
1J = 129 Hz, CH2-1), 47.9 (tm, 1J = 129 Hz, CH2-3), 119.8 (sharp t, 1J = 158 Hz, CH2-),
122.3 (dm, 1J = 158 Hz, C-9), 124.3 (dm, 1J = 159 Hz, C-5), 124.7 (dd, 1J = 159 Hz, 3J = 8.5
Hz, C-7), 124.9 (dd, 1J = 159 Hz, 3J = 8.5 Hz, C-6), 127.9 (dm, 1J = 157 Hz, C-8), 130.8
(unresoved, C-4a), 130.9 (unresolved, C-4), 133.2 (m, C-8a), 133.7 (sharp d, 1J = 153 Hz, C-
), 140.8 (unresolved, C-3a), 142.9 (sharp t, 3J = 4.5 Hz, C-9a) ppm, at –65 °C  28.2, 40.4,
47.1, 47.6, 120.1, 122.1, 123.9, 124.7, 124.8, 127.7, 130.1, 130.2, 132.5, 133.2, 140.8, 143.0
2
ppm, assigned through HETCOR and the following J and 3J COLOCS (7 Hz) cross-peaks.
2J: C-  -H cis, C-2  2-CH3 and both CH2-1/-3, C-3a  CH2-3, C-5  6-H, C-9a 
3
CH2-1; J: 2-CH3  C-1/-3 and 2-CH3, CH2-1  C-3/-9 and 2-CH3, CH2-3  C-1 and 2-
CH3, -H trans  C-4, 5-H  C-7/-8a, 6-H  C-4a/-8, 7-H C-5 and C-8a, 8-H  C-
4a/-6/-9, 9-H  C-3a/-4a/-8; IR (film)  3067 (w), 2953, 2925, 2865, 2833, 1506, 1465,
1425, 1366, 996, 920, 872, 777, 746 (s), 699 cm–1. Anal. calcd for C17H18 (222.33): C, 91.84;
H, 8.16. Found: C, 91.67; H, 7.97.
[1]
Organometallics 2013, 32, 4070–4081.
[2] R. Knorr, C. Behringer, E. Lattke, U. von Roman, M. Knittl, J. Org Chem. 2015, 80,
6313–6323.
R. Knorr; T. Menke, C. Behringer, K. Ferchland, J. Mehlstäubl, E. Lattke,
[3]
R. Knorr, K.-O. Hennig, P. Böhrer, B. Schubert, J. Organomet. Chem. 2014, 767,
125–135.
[4]
[5]
[6]
R. Knorr, J. Ruhdorfer, P. Böhrer, Organometallics 2015, 34, 1038–1045.
R. Knorr, M. Knittl, E. C. Rossmann, Beilstein J. Org. Chem. 2014, 10, 2521–2530.
This term was applied to non-dissociative cation migrations during enantiomerizations
of chiral, sp3-hybridized carbanion derivatives as studied by: a) D. J. Cram, L. Gosser, J. Am.
Chem. Soc. 1964, 86, 2950–2951; b) D. J. Cram, L. Gosser, J. Am. Chem. Soc. 1964, 86,
5457–5465, footnote 11 therein.
18
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
[7]
R. Knorr, E. Lattke, J. Ruhdorfer, U. von Roman, J. Firl, P. Böhrer, J. Organomet.
Chem. 2016, 824, 61–72.
[8]
R. Knorr, J Ruhdorfer, J. Mehlstäubl, P. Böhrer, D. S. Stephenson, Chem. Ber. 1993,
126, 747–754.
[9]
G. M. Rubottom, C. Kim, J. Org. Chem. 1983, 48, 1550–1552.
[10] U. von Roman, J. Ruhdorfer, R. Knorr, Synthesis 1993, 985–992 and literature
quoted therein.
[11] See the Supporting Information.
[12] A. R. Katritzky, G. J. Hitchings, R. W. King, D. W. Zhou, Magn. Reson. Chem.
1991, 29, 2–8, Tables 1–3 therein.
[13] R. Knorr, C. Behringer, H. Nöth, M. Schmidt, E. Lattke, E. Räpple, Chem.
Ber./Recueil 1997, 130, 585–592, compound 12 therein.
[14] 2nI + 1 = 5 components arise through n = 2 6Li nuclei, each with the spin quantum
number I = 1.
[15] R. Knorr, T. Menke, K. Ferchland, J. Mehlstäubl, D. S. Stephenson, J. Am. Chem.
Soc. 2008, 130, 14179–14188.
[16] See p 6318 and Figure 1 in Ref. [2].
[17] Formulas 15 and 16 in Figure 1 of Ref. [7].
[18] R. Knorr, K. Ferchland, J. Mehlstäubl, T. P. Hoang, P. Böhrer, H.-D. Lüdemann, E.
Lang, Chem. Ber. 1992, 125, 2041–2049, showing G‡ values of 15.4 versus 18.1 and 20.3
kcal mol–1 for R = C6H5 in Table 2 on p 2043 therein.
[19] An example of such temperature-dependent line shapes and their simulations may be
viewed in Section C and Figure 5 on pp 1042–1043 of reference [4].
[20] For references to pertinent programs, see p 2526 of reference [5].
[21] R. Knorr, T. Menke, K. Ferchland, Organometallics 2013, 32, 468–472.
[22] R. Knorr, E. Lattke, Tetrahedron Lett. 1977, 18, 3969–3972.
[23] On p 2527 of reference [5].
[24] Further arguments against alternative migration modes were presented in reference [1]
and on pp 6319–6320 of reference [2].
[25] Compound 11s in Ref. [10].
19
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
Scheme 1. Possibilities (-, -, and -route) of Li+(THF)4 migration in the ionic
stereoinversion mechanism of 1 in THF; aryl = 2,2-dimethylbenz[f]indan-4-yl.
Scheme 2. The known -arylalkenyllithiums 3 – 7 to be compared with 1 and 2.
Scheme 3. Preparations of the Precursors and Derivatives of 1.
Figure 1. Partial 13C NMR spectra of 4-(-lithiovinyl)-2,2-dimethylbenz[f]indane (1): a) In
tBuOMe at –70 °C, o = alkene 14; b) in THF at –85 °C (alkene 14 polymerized).
Figure 2. Temperature dependence of the natural logarithms (ln) of the pseudo-first-order
stereoinversion rate constants k (s–1) of 1-(-lithiovinyl)naphthalene (2&3THF) in THF.
20
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
Table 1. Microsolvation numbers d, NMR coupling constants 2JH,H [Hz] of CH2-, and lithiation shifts  = (RLi) – (RH) of the -
arylalkenyllithiums 1–5 in five solvents.
Entry Com-
Solvent
Agg [b]
d
2JH,H
Temp. Ref.
°C [d]
-Aryl
Lithiation shifts  [ppm]
C-ipso C-para para-H
No.
pound Group
(equiv of B)[a]
(at °C) [c]
C-
C-
1
benzindane
toluene (2)
tBuOMe
c-C5H10 (1)
toluene (1.3)
tBuOMe
Et2O
D
1
1
1
1
1
1
3
3
3
3
3
+66.5 [e] +0.9
+20.0
+20.3
+19.9
+19.6
+19.4
+21.8
+25.6
–6.7
–0.12 [f]
–0.40
–0.34
–0.02 [f]
–0.43
–
+25
+25
+25
–84
–40
–70
–30
–89
–47
–38
–60
[g]
[g]
[g]
[2]
[2]
[7]
[g]
[2]
[g]
[7]
[7]
1
1
1
4
4
3
1
4
2
3
5
5.8 ( +25)
6.0 ( +25)
–
2
benzindane
benzindane
2,6-(CH3)2
2,6-(CH3)2
2-CH3
D
+66.8
+66.1
+68.2
+68.1
+67.9
+76.2
+76.4
+76.2
+77.3
+75.1
–0.1
+1.0
–3.6
–3.6
+5.0
–8.2
–7.2
3
D
–6.7
4
D
–6.5
5.7 ( +25)
5.7 ( –7)
–
5
D
–7.0
6
D
–6.1
7
benzindane
2,6-(CH3)2
naphthyl-1
2-CH3
THF
M
M
M
M
M
–11.6
–11.0
–11.3
–10.4
–10.6
–0.72
–0.76
–
8.9 ( –40)
8.8 ( +2)
9.0 ( +28)
8.0 (+25)
8.0 (+25)
8
THF
–10.9 +24.2
9
THF
–5.0
–4.7
–3.9
+27.7
+.8
10
11
THF
–0.61
–0.60
2-CH(CH3)2
THF
+26.7
2
[a] Equivalents of B = tBuOMe present. [b] “M” = monomer, “D” = dimer. [c] Temperature of this JH,H value. [d] Temperature of
1
determinations of  = (RLi) – (RH). [e] Quintet (1:2:3:2:1) at –55 °C with J(13C,6Li) = 7.5 Hz. [f] Special effect of toluene. [g] This
work. …
21
This article is protected by copyright. All rights reserved.

10.1002/chem.201702353
Chemistry - A European Journal
–1
–1
‡
‡
Table 2. Pseudoactivation parameters G (kcal mol at 0 °C), H (kcal mol ), and
–1 –1
‡
S (cal mol K ) of the cis/trans diastereotopomerization rates of monomeric
alkenyllithiums 1 – 7 in THF.
Entry Compd. Aryl group[a] G (0 °C)
H
13.35 (0.03) 6.63 (0.24) –24.6 (1.0) 0.0 [1]
2,6-Me2Ph[d] 12.47 (0.01) 6.77 (0.18) –20.8 (0.7) +5
[1]
naphthyl-1 12.94 (0.01) 6.9 (0.1) –22.3 (0.4) +28 [e]
benzindane 13.87 (0.01) +15 [e]
S
HT[b] Ref.
‡
‡
‡
1
2
3
4
5
6
7
8
7a
7b
2
phenyl[c]
1
–
–
6
2,6-ipr2Ph
2-iprPh
15.79 (0.07) 9.3 (0.4)
16.1 (0.1) 9.0 (0.4)
16.14 (0.05) 10.6 (0.3)
16.3 (0.1) 10.2 (0.7)
–23.6 (1.2) +71 [4]
–25.9 (1.2) +92 [e]
–20.2 (0.8) +74 [e]
–22.2 (2.2) +63 [2]
5
3
2-MePh
4
2,6-Me2Ph
[a] ipr = isopropyl, Me = methyl, Ph = phenyl. [b] HT = highest temperature (°C) of the rate
measurements. [c] 7a = 2-(-Lithiobenzylidene)-1,1,3,3-tetramethylindane. [d] 7b = 2-(-
Lithio-2´,6´-dimethylbenzylidene)-1,1,3,3-tetramethylindane. [e] This work.
[TOC-Graphic:]
Separation and migration; no digressions.
Quasi-benzyl-anion resonance is stronger in the linear transition state (see the Graphic) than
in the two trivolvated, angular groundstates (CIP and CIP´), because immobilization of the
fourth THF ligand at Li increases the charge delocalization from C() to C(4). The excess
negative charge (-) at C(4) appeared to favor the migration of Li+(THF)4 across the
C(3)/C(4) region of the conformationally fixed -aryl group.
Keywords: conformational analysis • E,Z configurational lability• reaction mechanisms •
sp2-carbanion inversion • structural elucidation
22
This article is protected by copyright. All rights reserved.