Carbenoid chain reactions through proton, deuteron, or bromine transfer from unactivated 1-bromo-1-alkenes to organolithium compounds

Article abstract of DOI:10.1021/jo070623w

(Chemical Equation Presented) The deceptively simple vinylic substitution reactions Alk₂C=CA-Br + RLi → Alk₂C=CA-R + LiBr (A = H, D, or Br) occur via an alkylidenecarbenoid chain mechanism (three steps) without transition metal catal

Full text of DOI:10.1021/jo070623w

Carbenoid Chain Reactions through Proton, Deuteron, or Bromine
Transfer from Unactivated 1-Bromo-1-alkenes to Organolithium
Compounds#
Rudolf Knorr,* Claudio Pires, and Johannes Freudenreich
Department of Chemistry and Biochemistry, Ludwig-Maximilians-UniVersita¨t,
Butenandtstr. 5-13, 81377 Mu¨nchen, Germany
rhk@cup.uni-muenchen.de
ReceiVed March 26, 2007
The deceptively simple vinylic substitution reactions Alk2CdCAsBr + RLi f Alk2CdCAsR + LiBr
(A ) H, D, or Br) occur via an alkylidenecarbenoid chain mechanism (three steps) without transition
metal catalysis. 2-(Bromomethylidene)-1,1,3,3-tetramethylindan (Alk2CdCH-Br, 2a) is deprotonated
(step 1) by phenyllithium (PhLi) to give the Br,Li-alkylidenecarbenoid Alk2CdCLi-Br (3). In the ensuing
chain cycle, 3 and PhLi (step 2) form the observable alkenyllithium intermediate Alk2CdCLisPh that
characterizes the carbenoid mechanism in Et2O and is able to propagate the chain (step 3) through
deprotonation of 2a, furnishing carbenoid 3 and the product Alk2CdCHsPh. The related 2-(dibromo-
methylidene)-1,1,3,3-tetramethylindan (Alk2CdCBr2, 2c) and methyllithium (MeLi) generate carbenoid
3 (step 1), which incorporates MeLi (step 2) to give Alk2CdCLisCH3, which reacts with 2c by bromine
transfer producing Alk2CdCBrsCH3 and carbenoid 3 (step 3). PhCtCLi cannot carry out step 1, but
MeLi can initiate (step 1) the carbenoid chain cycle (steps 2 and 3) of 2c with PhCtCLi leading to
Alk2CdCBrsCtC-Ph. Reagent 2a may perform both proton and bromine transfer toward Alk2Cd
CLisCH3, feeding two coupled carbenoid chain processes in a ratio that depends on the solvent and on
a primary kinetic H/D isotope effect.
Introduction
investigations. The deceptively simple relationship between 6
and 2a may, for example, suggest an addition-rotation-
elimination (ARE)5 mechanism involving a nucleophilic attack6,7
of RLi at CR of 2a. However, it will be reported here that RLi
) CH3Li and 2a can form 4a and 5c also (as depicted in Scheme
1) along with the “normal” product 5a of type 6. The formation
of 5c, in particular, may appear surprising because the possible
byproduct RBr ) CH3Br (Scheme 1) cannot be expected to
produce major portions of 5c (from 3?) or 5a (from 1a?) under
the reaction conditions, as will be explained in section 3.
Previous investigations6 of the related chloromethylidene
reagents 7 and 8 (Scheme 1) have refuted the ARE5 and several
other7 mechanisms, revealing instead carbenoid chain processes
An organolithium compound RLi confronted with 2-(bro-
momethylidene)-1,1,3,3-tetramethylindan (2a)1 may choose
between several reaction modes (Scheme 1): (i) the Br/Li
interchange reaction leading to the formation of 1a, as observed2
with n-butyllithium, (ii) the R-deprotonation mode generating
the short-lived Br,Li-alkylidenecarbenoid3,4 3 and its descen-
dants, and (iii) other modes which might be thought to generate
the “normal” products 6 obtained from 2a in the present
# Sterically Congested Molecules, 19. Part 18: ref 6.
(1) Syntheses of the reagents 2a, 2b, and 2c are outlined in section 4 of
Results and Discussion.
(2) Knorr, R.; Freudenreich, J.; Polborn, K.; No¨th, H.; Linti, G.
Tetrahedron 1994, 50, 5845-5860.
(3) Carbenoids carry both a metal cation and a nucleofugal group at the
same carbon atom: Ko¨brich, G. Angew. Chem. 1972, 84, 557-570 and
first footnote therein; Angew. Chem., Int. Ed. Engl. 1972, 11, 473-485.
(4) Nomenclature: Stang, P. J. Acc. Chem. Res. 1982, 15, 348-354.
(5) Rappoport, Z. Acc. Chem. Res. 1992, 25, 474-479, and cited
literature.
(6) Knorr, R.; Pires, C.; Behringer, C.; Menke, T.; Freudenreich, J.;
Rossmann, E. C.; Bo¨hrer, P. J. Am. Chem. Soc. 2006, 128, 14845-14853.
(7) Depicted in Scheme 1 of ref 6.
10.1021/jo070623w CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/11/2007
6084
J. Org. Chem. 2007, 72, 6084-6090

Carbenoid Chain Reactions
SCHEME 1
SCHEME 2
product 11 of type 6. This proton transfer would generate
carbenoid 3 without participation of PhLi, so that further
repetitions of the step 2/step 3 cycle might convert reagent 2a
to product 11 with a total consumption of only (a little more
than) 1 equiv of PhLi. Thus the characterizing stoichiometry
of PhLi/2a should normally be found between 1:1 (chain) and
2:1 (nonchain carbenoid) depending on chain interruptions due
to an eventual depletion of reagent 2a or due to a “leakage”
(Scheme 2) from 12 to 11 through protonation by sources other
than 2a (for example, the solvent). Note that a perfect chain,
consisting of steps 2 and 3, would leave essentially no residual
chain carrier 12 from which to obtain the acid 9d.
In practice, a run with PhLi (2.1 equiv, 0.24 M) in Et2O
solution at room temperature was almost completed in 20 min.
The in situ 1H NMR spectrum showed residual PhLi, the
product9 11, the known6,10 intermediate 12, and residual reagent
2a in the molar ratio 64:15:67:18. The observation of 12 (67%)
establishes a predominant course via carbenoid 3 with efficient
trapping of 3 by PhLi (fast step 2), because an alternative
formation of 12 by deprotonation of 11 was never detected.
The increasing PhLi/2a ratio (from 2.1:1 to 64:18) is compatible
with the formation of 11 from 12 through the chain propagation
step 3 (which consumes only reagent 2a). The observed
coexistence of PhLi, 2a, and 12 implies that both steps 1 and 3
occur slowly. In view of the 15:67 ratio of 11/12, the steps can
then be ranked11 with respect to their rates (rapidity of the flux
of material) as 2 . 1 J 3. The known9 allenic “dimer” 10
emerged very slowly through an unknown process after the
consumption of reagent 2a. The final carboxylative workup with
solid CO2 after 44 h furnished the acids 9d9 (7% yield),12 9a2
(e1% yield), and benzoic acid (from 0.1 equiv of surviving
PhLi), showing that not all of intermediate 12 was destroyed
which involved either R-deprotonation (of 7) or Cl/Li inter-
change reactions (of 8). By use of the developed6 criteria
(stoichiometry, observability of intermediates or their detection
through trapping with CO2 to give acids 9, and kinetic H/D
isotope effects, as applicable), it will be inquired here about
the feasibility of the carbenoid chain pathway (2a-c + RLi f
3 f 6) for RLi ) phenyllithium, CH3Li, and PhCtCLi.
Because most of the results are presented within the running
text for the sake of instantaneous deductions, these results are
reiterated for reference in the same order in Table S1 of the
Supporting Information. The 1,1,3,3-tetramethyl-2-indanylidene
model system facilitated these investigations for several rea-
sons: Some of the conceivable8 side-reactions of the alkylidene-
carbenoid 3 or its precursors are suppressed, and the six-proton
NMR singlets of 1,1-Me2 and 3,3-Me2 provide distinctive and
easily spotted marks for detecting the products in situ even on
a rather small scale (≈0.1 mmol).
Results and Discussions
1. Phenyllithium (RLi ) PhLi) with Reagent 2a: Chain
Propagation through Proton Transfer. Scheme 2 presents the
alkylidenecarbenoid chain mechanism6 of PhLi with 2a as an
example. The initiating step 1 employs PhLi as a base to
deprotonate reagent 2a1 with formation of the carbenoid 3, which
is then substituted8 in step 2 by PhLi acting as a nucleophile to
form the alkenyllithium intermediate 12. At this point, the
carbenoid process may halt, with 2 equiv of PhLi consumed
per 2a (2:1 stoichiometry), and might be recognized by
carboxylation which would convert 12 to 9d. However, the
carbenoid chain mechanism may append step 3 in which
intermediate 12 is protonated by reagent 2a to give the final
(9) Knorr, R.; Lattke, E.; Ra¨pple, E. Liebigs Ann. Chem. 1980, 1207-
1215.
(10) Knorr, R.; Hoang, T. P.; No¨th, H.; Linti, G. Organometallics 1992,
11, 2669-2673.
(11) With the assumption that the substitution step 2 with LiBr
elimination is practically irreversible. To be successful, step 2 must occur
more rapidly than the decomposition of the alkylidenecarbenoid which is
usually8 fast at room temperature.
(12) Most of the yields were determined in small-scale runs (0.1-1
mmol) under nonoptimized conditions.
(8) Knorr, R. Chem. ReV. 2004, 104, 3795-3849.
J. Org. Chem, Vol. 72, No. 16, 2007 6085

Knorr et al.
through its easy conversion to the known6 alkyne 13b with
MeLi. This run established that at least some portion of 2c and
MeLi had generated an intermediate (3) that could choose to
react either with MeLi (steps 2A and 3A) to give 5c or
(preferentially) with PhCtCLi to produce 13a-c via steps 2B
and 3B (in analogy with the behavior6 of the dichloro reagent
8). A complete exclusion of non-carbenoid routes (5c r 2c)7
would have to be based on evidence for a decelerated13
consumption of 2c in the presence of PhCtCLi, but such
evidence was not obtained here because 2c was always
consumed too rapidly.
SCHEME 3
In THF solution, the interaction of MeLi with 2c (started at
-70 °C) furnished the “butatriene” 16 as a major product in
less than 7 min at room temperature. Because 16 results from
self-substitution at CR of carbenoid 3 (“dimerization”),14 this
indicated that the Br/Li interchange step 1 of Scheme 3 was
sufficiently fast to build up a relatively high concentration of 3
and that the subsequent “dimerization” could keep pace with
step 2A. The amount of MeBr (step 1) corresponding to 16
(including some ethane formed from MeBr with MeLi) and no
trace of residual 2c or MeLi were detected through 1H NMR in
situ, which displayed the products MeBr + 16 + 5c + 5a in
the molar ratio 67:36:26:2. The leakage product 5a together with
a trace of the acid 9c (isolated after carboxylation) revealed
that at least some of the chain carrier 14 had survived in the
presence of MeBr up to 3 h. Employment of an excess of MeLi
led to almost the same molar ratio (0:35:26:4), except that MeBr
was converted completely to ethane within 14 min, so that close
to 2 equiv of MeLi were consumed. The quantities of MeBr
(or ethane) and 5c, resulting from fast Br/Li interchange
reactions, suggest the rankings11 of the step rates to be 1 J 3A
and 2A ≈ carbenoid dimerization, which amounts to a disturbed
chain process. Clearly, it was preferable to use the undisturbed
chain process in Et2O solution where 5c was formed exclusively
because step 1 was not obtrusive.
by “leakage” during that long period of time. The nonacidic
fraction contained mainly 11, some allenic “dimer” 10, and a
trace of olefin 4a. The tiny portions of 4a and 9a observed
demonstrate that PhLi hesitated to perform a Br/Li interchange
with 2a to give 1a.
In THF solution, product 11 was almost the only tetrameth-
ylindan derivative generated from 2a with PhLi (1 equiv) within
less than 15 min at room temperature (no acids, no residual 2a,
no allenic “dimer” 10, not more than a trace of 4a). Thus, chain
propagation by proton transfer to 12 (step 3) was presumably
more efficient in THF than in Et2O, judging from the clean 1:1
stoichiometry and other parallel results observed earlier6 for the
carbenoid chain processes of PhLi with the dichloro reagent 8.
2. Methyllithium (RLi ) MeLi) with Reagent 2c: Chain
Propagation through Bromine Transfer. The reaction of MeLi
(0.34 M) in diethyl ether with reagent 2c1 (0.083 M) was
complete in 4 min at room temperature and required 1.0 equiv
of MeLi to produce 5c exclusively (Scheme 3). The in situ 1H
NMR spectra showed no trace of the leakage product 5a, and
carboxylation afforded no acid 9c (to be expected from chain
carrier 14). All this accords with an efficient chain process (step
rates ranking11 as 2A and 3A . 1), but could it not equally
well be explained with a direct conversion (5c r 2c) through
some other7 mechanism, considering that 2c is obviously much
more reactive than the dichloro reagent6 8 toward MeLi? Such
direct (non-carbenoid) pathways to 5c should not be influenced
by the presence of PhCtCLi (which did not interact with 2c
for over more than 4 h). But PhCtCLi (2.0 equiv) was
incorporated (steps 2B and 3B of the carbenoid chain B) as
soon as MeLi (0.1 equiv) was added at -70 °C, which induced
the consumption of ≈70% of 2c in <5 min at room temperature,
at which point the chain reaction stopped because the chain
carrier 15 had been converted to the leakage product 13a. One
restart with a second small batch of MeLi (0.1 equiv) sufficed
for the total conversion of 2c, and carboxylative workup afforded
5c (4% yield with respect to 2c), 13c (71%), 13b (2%), and
13a (6%) but no 5a and no acids other than PhCtCCO2H (from
residual PhCtCLi). The constitution of 13c was confirmed
3. MeLi with Reagents 2a and 2b: Two Chains Sharing
Step 2. The monobromide 2a (0.10 M) in diethyl ether
consumed roughly 1 equiv of MeLi (1.0 M) with a first t1/2
≈
20 min at room temperature.15 Carboxylative workup after 3.3
h furnished the acid 9c (Scheme 4, 16% yield, seemingly
uncontaminated by the acid 9a2,16), along with the proton transfer
product 5a (85% yield) which was not contaminated by the
product 5c (<2%) of bromine transfer.12 The exclusive forma-
tion of substitution products (5a and 9c), obtained in a
quantitative yield, implies that step 2 of Scheme 4 occurred
much more rapidly than step 1A. Step 1A was not much slower
than step 3A, however, because 16% of the intermediate 14
(trapped as 9c) had accumulated owing to a shortage of reagent
2a caused by step 1A. Hence the ranking11 of step rates can be
inferred to be 2 . 3A > 1A. The predominance of the
alkylidenecarbenoid chain process followed from the ≈1:1
stoichiometry in combination with a distinct (at least 7-fold)
deuterium-induced rate-depression15 for reagent 2b1 (0.10 M)
reacting with MeLi (1.36 M). Accordingly, the almost-isotope-
independent step 2 cannot be rate-limiting.11 In consequence
(13) See the comments to Scheme 5 of ref 6.
(14) pp 3817-3818 and 3843 of ref 8.
(15) All t1/2 values in this work are meant to convey semiquantitative
rate information only and to point at primary kinetic isotope effects in
reactions of 2b. They cannot be used to quantitate kH/kD ratios because t1/2
increases with decreasing initial concentrations [RLi].
(16) Knorr, R.; von Roman, T.; Freudenreich, J.; Hoang, T. P.;
Mehlsta¨ubl, J.; Bo¨hrer, P.; Stephenson, D. S.; Huber, H.; Schubert, B. Magn.
Reson. Chem. 1993, 31, 557-565.
6086 J. Org. Chem., Vol. 72, No. 16, 2007

Carbenoid Chain Reactions
SCHEME 4
SCHEME 5
product 17a17 in a 45:27:21:7 ratio, whereas the corresponding
product ratio 5b/5c/4c/17b ) 13:44:40:3 was detected in the
run with 2b. Hence the retardation of D transfer to 14 in step
3A had increased the portion of Br transfer in step 3B (affording
more 5c and 4c). The leakage products 5a and 4b were not
present immediately after the consumption of 2b, but they were
found after a total of 65 min at room temperature when
carboxylative workup afforded a nonacidic fraction with the
product ratio 5a/5b/5c/4c/17b/4b ) 6:14:39:32:4:5 accompanied
by an acid fraction containing the fully (>99.5%) CR-deuterated
acid 9b (11% yield) but no 9c. This delayed formation of 5a in
the run with 2b disclosed a transitory minor accumulation of
intermediate 14 and its subsequent leakage reaction, while the
absence of acid 9c testified to a shorter lifetime of 14 in THF
than in Et2O solution (where 16% of 9c had been found). The
more stable secondary chain carrier 1b (trapped as 4b + 9b)
was accumulated because of the deuterium-induced deceleration
of step 4B, as compared to the run with 2a whose carboxylation
after 81 min had furnished only a trace of the acid 9a and no
9c. A part of 1b may have been generated through the almost-
isotope-independent step 1B (in analogy with 1a r 2a in
Scheme 1), but the byproduct MeBr cannot have formed major
portions of both 5c (from 3) and 5a,b (from 1a,b) because such
isotope-independent processes would not explain the observed
variations of the product pattern correctly.
of this decelerated consumption of 2b (steps 1A and 3A), the
bromine transfer from 2b to 14 (in step 3B, without C-D bond
scission and, hence, almost isotope-independent) became detect-
able through the appearance of small amounts of both 5c and
the R,R-dideuterated olefin 4c, the latter being formed via step
4B and identified through its 13C NMR quintet for CR (1JCD
)
23.8 Hz). Both chains A and B share step 2 which connects the
chain carriers 3 and 14. Notice that chain B employs three
(instead of two) steps (2, 3B, and 4B) and requires only 0.5
equiv of MeLi per mol of 2a or 2b because the latter two
reagents are consumed in both step 3B and step 4B, while MeLi
is incorporated only through step 2. According to the NMR
analyses in situ, the H/D ratio in products 5a/5b increased slowly
from 19:72 in the early stage to 43:48 in the end (>3 h), because
of the leakage reaction of intermediate 14, affording the final
product ratio 5a/5b/5c/4c/4b ) 43:48:5:3:1. Strictly speaking,
a slow process with an alternative mechanism,7 confined to
transforming 2b to 5b directly, might have participated in this
retarded run, but such an alternative pathway would then be
unable to contribute significantly to the much faster conversion
of 2a to 5a observed above.
4. Syntheses of the Reagents 2b and 2c. The published18
method of preparing 2a (Scheme 5) via epoxidation of 4a and
“nucleophilic bromination” was utilized to obtain the deuterated
monobromide 2b from 4c. The dibromide 2c cannot be made
by electrophilic bromination19 of 4a (because of methyl migra-
tion)18 or of 2a, and several other approaches19 were also futile.
A successful route to 2c started from the dichloro reagent 8
which was inert toward LiSn(n-Bu)3 but reacted readily when
20-22
added to a sufficiently concentrated solution of LiSnMe3
in THF to give 19, whose constitution was supported by the
In THF solution, reagent 2a (0.26 M) vanished within the
warm up period of 12 min (or within 1 h at -70 °C) and
consumed 0.9 equiv of MeLi (0.65 M, added at -70 °C).
Reagent 2b (0.26 M) reacted with a longer first t1/2 ≈ 5 min15
(establishing the carbenoid route) and required only 0.7 equiv
of MeLi (0.58 M initially), which points to the carbenoid chain
B (as anticipated above). This primary kinetic H/D isotope effect
excludes step 2 as rate-limiting,11 so that the step rates can be
ranked as 2 . 3B ≈ 3A > 1A in consideration of the following
(17) The side product 17a and its formation (which was suppressed in a
run at -70 °C) from the reaction of 1a with 2a or with the chloroalkene 7
will be analyzed (together with 17b) in a later publication.
(18) Knorr, R.; Freudenreich, J.; von Roman, T.; Mehlsta¨ubl, J.; Bo¨hrer,
P. Tetrahedron 1993, 49, 8837-8854.
(19) Details are given in the Supporting Information.
(20) Tamborski, C.; Ford, F. E.; Soloski, J. E. J. Org. Chem. 1963, 28,
237-239.
(21) Kitching, W.; Olszowy, H.; Waugh, J.; Doddrell, D. J. Org. Chem.
1978, 43, 898-906.
1
evidence. The final in situ H NMR spectra observed with 2a
(22) Reich, H. J.; Reich, I. L.; Yelm, K. E.; Holladay, J. E.; Gschneidner,
G. J. Am. Chem. Soc. 1993, 115, 6625-6635.
showed the presence of 5a, 5c, 4a, and the butadiene-type side-
J. Org. Chem, Vol. 72, No. 16, 2007 6087

Knorr et al.
characteristic magnitude23 of the 119Sn/117Sn coupling constant
2J ) 653 Hz. This vinylic substitution reaction is thought to
begin with the generation of carbenoid 18 and to proceed via
20; indeed, the carbenoid “dimerization” product 16 was formed
as a troublesome side product if the addition of 8 to LiSnMe3
was not carefully controlled or if LiSnMe3 was added to 8. In
contrast, the treatment of acyclic 1,1-dichloroalkenes with
LiSnMe3 had been reported24 to produce alkynes instead of
trimethylstannylated alkenes, presumably via acyclic Cl,Li-
alkylidenecarbenoids (corresponding to 18) which rearranged8
too fast. In the case of 18, this rearrangement would expand
the five-membered ring with little driving force and is therefore
retarded (which was one of the reasons for us to utilize indan
derivatives). Careful titration of pure 19 with elemental bromine
in CCl4 solution furnished pure 2c without any rearrangement
products.
onto the corresponding alkenyllithium intermediates 12 (observ-
able by NMR in Et2O) or 14. On the other hand, 2a and n-BuLi
performed only2 the Br/Li interchange reaction to give 1a in
both Et2O (at +25 °C) and THF solutions (at -70 °C). The
propensity of 14 for proton transfer was less marked in THF
than in Et2O and sufficiently small to permit the bromine
abstraction from 2b in competition with deuteron transfer,
because the latter was handicapped by a sizable primary kinetic
H/D isotope effect (which established the carbenoid pathway).
Thus the product pattern became isotope-dependent in both Et2O
and THF, confirming the proposed coupling of two carbenoid
chain processes (Scheme 4). These peculiar Et2O/THF solvent
effects are thought to depend on the unknown transition state
solvation in steps 1 and 3; hence, they cannot be explained at
this time. More examples with ramifications and some limita-
tions of the carbenoid chain mechanism will be reported
separately.
Conclusions
Experimental Section
Fully developed carbenoid chain processes (with step rate
rankings11 “2 and 3 . 1” and a 1:1 stoichiometry) occur less
frequently in the bromoalkene (2a-c) system than with the less
reactive, analogous 1-chloro-1-alkenes6 7 and 8. This is so
because the chains of 2a-c may become disturbed by an overly
fast step 1 (up to carbenoid “dimerization”) or by a sluggish
propagation step 3. In the latter (nonchain carbenoid) situation
which requires up to 2 equiv of RLi per reagent 2, the
accumulated intermediate Alk2CdCLi-R may, for preparative
purposes, be quenched by protonation or with a Br transfer
source, as applicable, immediately after the total conversion of
a reagent 2. In the former case (fast step 1), the 1-bromoalkene
reagent should be added slowly to a concentrated solution of
RLi kept at -70 °C in order to increase the rate of RLi with
the carbenoid 3 (substitution step 2) and to avoid the imminent
carbenoid decomposition. The clean course of these processes
depends also on the inability of intermediate Alk2CdCLi-R
(unless R ) H, compare 17 in Scheme 4) to attack its parent
alkylidenecarbenoid (3) at CR. For a contrasting example, the
corresponding intermediates R-(CH2)j-M (M ) Li25,26 or Mg
halide27), generated from R-(CH2)j-1-M (j g 1) through
substitution at a saturated carbenoid BrCH2M, not only have to
propagate their (supposed) chain reactions by Br transfer from
CH2Br2 to re-create BrCH2M but also have to compete for
substitution at BrCH2M with the residual portions of all of their
ancestors R-(CH2)i-M (0 e i < j), thus creating homologated
intermediates and eventually affording (by Br transfer) unwel-
come product mixtures R-(CH2)k-Br. Earlier reported side
products such as Me2CdCBr-CH3 (obtained28 from Me2Cd
CBr2 and MeLi) or (CH2)3CdCBr-Ph (from (CH2)3CdCBr2
and PhLi29) suggest that the alkylidenecarbenoid chain mech-
anism with Br transfer may not be confined to the sterically
shielded reagents 2a-c.
General Remarks. Organolithium compounds were handled
under a stream of dry argon cover gas. Experiments in NMR tubes
(5 mm) were performed with nondeuterated solvents (≈0.7 mL,
containing ≈0.04 mL of C6D12 if required as a “lock substance”),
allowing product analyses to be carried out in situ before workup.
1
Concentrations were estimated by comparison with the H NMR
integral of a sealed capillary filled with pure ClCH2CtN (δH
≈
3.9) or of the low-field 13C satellites of the solvents. Hydrogen
versus deuterium distributions were determined by pairwise integra-
tions of the baseline separated 13C NMR absorptions having
n
sufficiently large isotope-induced shift differences ∆ (n > 1, so
as to obviate NOE differences), with the machine parameters set
for maximum resolution (for example, number of points np )
160000 (1H at 400 MHz) or 524288 (13C), and acquisition times at
≈ 13 s). Commercially available solutions of methyllithium (δH ≈
-2) in Et2O, containing LiBr, of n-butyllithium in hexanes (δH ≈
-0.75 in benzene) and of tert-butyllithium (t-BuLi) in pentane, were
used. The Br/Li interchange reaction with the ensuing â-elimination
of HBr (t-BuLi + ArBr f t-BuBr + ArLi, then t-BuLi + t-BuBr
f t-BuH + LiBr + Me2CdCH2) was employed in ethereal solvents
to prepare solutions of phenyllithium (δH ≈ 8.0 for two o-H) from
bromobenzene.
2-(Bromomethylidene)-1,1,3,3-tetramethylindans (2a and 2b).
These were prepared along the lines described in ref 18. Residual
dCH NMR absorptions (1H, s δ 6.16; 13C, δ 100.0) could not be
detected for the R-D derivative 2b.
2-(Dibromomethylidene)-1,1,3,3-tetramethylindan (2c). (a) A
solution of pure 2-[bis(trimethylstannyl)methylidene]-1,1,3,3-tet-
ramethylindan (19, 35 mg, 0.068 mmol) in CCl4 (0.40 mL),
contained in an NMR tube, was titrated at -25 °C with a CCl4
solution (1 M) of elemental bromine. 1H NMR control spectra
revealed the intermediate formation of 2-[bromo(trimethylstannyl)-
methylidene]-1,1,3,3-tetramethylindan with δ (200 MHz, CCl4)
+0.45 (s, 2J(119Sn) ) 54 Hz, 1SnMe3) and 1.43 and 1.70 (2 s, 2 +
2 1-/3-CH3). The final byproduct BrSnMe3 had δ 0.75 (s) with 2J(119
-
Sn) ) 57 Hz (SnMe3); it diminished on overtitration to give CH3-
Reagent 2a preferred to transfer its R-proton (rather than its
Br) onto PhLi or MeLi in both Et2O and THF solutions and
Br (δ 2.63) and Br2SnMe2 (δ 1.35 with J(119Sn) ) 66 Hz). The
2
mixture was diluted with Et2O (10 mL) and 2 M NaOH, and the
aqueous layer was extracted with Et2O (2×). The combined ethereal
phases were washed until neutral, dried over Na2SO4, and concen-
trated to give practically pure 2c as a colorless powder (21 mg,
89%). The purification and spectra of 2c are described in the sequel.
(b) The imminent destannylation of 19 by recrystallization from
hot ethanol suggested to employ incompletely purified specimens
of 19, most often contaminated with small amounts of the
“butatriene” 16. A CCl4 solution (10 mL) of 19 (912 mg, 1.78
mmol) was stirred in an ice bath during the slow titration with a
(23) Mitchell, T. N.; Amamria, A. J. J. Organomet. Chem. 1983, 252,
47-56.
(24) Mitchell, T. N.; Reimann, W. Organometallics 1986, 5, 1991-1997.
(25) Kirmse, W.; von Wedel, B. Liebigs Ann. 1963, 666, 1-9.
(26) Huisgen, R.; Burger, U. Tetrahedron Lett. 1970, 3053-3056 and
references therein.
(27) Villie´ras, J. Bull. Soc. Chim. Fr. 1967, 1511-1520, on p 1513.
(28) Hartzler, H. D. J. Am. Chem. Soc. 1964, 86, 526-527.
(29) Fitjer, L.; Kliebisch, U.; Wehle, D.; Modaressi, S. Tetrahedron Lett.
1982, 23, 1661-1664.
6088 J. Org. Chem., Vol. 72, No. 16, 2007

Carbenoid Chain Reactions
CCl4 solution (0.97 M) of elemental bromine. The end point was
reached when the red color of bromine persisted for more than
60 s. Overtitration consumed first an eventual admixture of the
“butatriene” 16 (recommended) and then the byproduct BrSnMe3
(see above) with formation of MeBr and Br2SnMe2 (not desirable).
After we confirmed by 1H NMR the complete formation of 2c, the
whole mixture was concentrated in vacuo (with due attention to
the toxicity of volatile tin compounds) and then dissolved in boiling
ethanol (slow dissolution). Weakly soluble contaminants (such as
16) were removed by filtration, and part of the ethanol was distilled
off until a first fraction of pure 2c (233 mg, 38%) crystallized slowly
olefin 4a (10%). Thus the lowered temperature had the effects of
suppressing the formation of the butadiene-type side-product 17a
and of accumulating the intermediate 1a which was trapped as 9a
(of which only a trace had been produced in the run conducted at
room temperature).
2-(1,1,3,3-Tetramethyl-2-indanylidene)acetic Acid (9a). De-
scribed in ref 2.
2-(1,1,3,3-Tetramethyl-2-indanylidene)propanoic Acid (9c).
The preparation of the pure acid 9c by carboxylation of the short-
lived alkenyllithium 14 was difficult, because 14 reacted faster with
proton sources (leakage) than it could be generated from bromoalk-
ene 5c with either MeLi in THF or n-BuLi in Et2O at room
temperature. The successful preparation consisted in adding 5c
(≈0.5 mmol as a product mixture with the olefins 4a and 5a) in
Et2O (5 mL) slowly over 2 min to a solution of tert-butyllithium
(t-BuLi, 4.30 mmol) in pentane (5.00 mL) stirred at -70 °C, so
that 14 was generated in the presence of t-BuLi in excess which
scavenged the coproduct t-BuBr of the Br/Li interchange reaction.
This mixture was stirred for 5 min more at -70 °C, then poured
onto solid CO2, warmed up, and dissolved in Et2O and pure water.
The alkaline aqueous layer was acidified and extracted with Et2O.
The latter Et2O extract was washed until neutral, dried over Na2-
SO4, and concentrated to provide a mixture of t-BuCO2H and the
acid 9c (97:3). Pure 9c (weakly soluble, colorless needles, 42 mg,
≈34%) was isolated through two crystallizations from boiling
1
as transparent, glimmer-like leaflets: mp 177-179 °C; H NMR
(400 MHz, CDCl3) δ 1.65 (s, 4 1-/3-CH3),16 7.13 (m, 4-/7-H), 7.25
(m, 5-/6-H); 13C NMR (100.6 MHz, CDCl3) δ 27.5 (qq, 1J ) 127.5
Hz, 3J ) 4.5 Hz, 4 1-/3-CH3), 52.3 (unresolved m, C1,3), 82.2 (sharp
s, CR), 122.4 (dm, 1J ) 156 Hz, C4,7), 127.5 (ddd, 1J ) 159 Hz, 3J
) 7 Hz, C5,6), 148.8 (blurred t, 3J ≈ 7 Hz, C8,9), 160.4 (unresolved
m, C2), assigned16 by comparison with (dibromomethylidene)-
cyclobutane;29 IR (KBr): 2990, 2961, 2926, 2864, 1576, 1487,
1455, 1362, 802, 756, and 737 cm-1. Anal. Calcd for C14H16Br2
(344.1): C, 48.87; H, 4.69. Found: C, 48.72; H, 4.62. Brominated
“butatriene” (16‚Br2) accumulated in the mother liquors: 1H NMR
(200 MHz, CDCl3) δ 1.63, 1.65, 1.72, and 1.82 (4 s, 4 1-/3-CH3).
2-Methylidene-1,1,3,3-tetramethylindan (4a). See compound
6 in ref 18.
1
petroleum ether (80-110 °C): mp 220-221.5 °C; H NMR (400
2-Ethylidene-1,1,3,3-tetramethylindans (5a and 5b). Described
in ref 6.
MHz, CDCl3) δ 1.55 (s, 2 1-CH3), 1.57 (s, 2 3-CH3), 2.19 (s,
R-CH3), 7.13 (m, 7-H), 7.16 (m, 4-H), 7.23 (m, 5-H), 7.25 (m,
6-H), assigned through the NOESY correlations R-CH3 T 3-CH3
T 4-H and 1-CH3 T 7-H T 6-H; 13C NMR (100.6 MHz, CDCl3)
2-(1-Bromoethylidene)-1,1,3,3-tetramethylindan (5c). (a) From
2c: The dibromide 2c (137 mg, 0.40 mmol) was added under argon
cover gas to the contents of an NMR tube containing MeLi (1.22
mmol) in Et2O (0.88 mL) at -70 °C. The stoppered tube was shaken
vigorously at room temperature for rapid mixing of the reactants.
1
1
δ 18.35 (sharp q, J ) 128.9 Hz, R-CH3), 29.11 (qq, J ) 127.5
Hz, 3J ) 4.5 Hz, 2 3-CH3), 30.23 (qq, 1J ) 127.5 Hz, 3J ) 4.5 Hz,
2 1-CH3), 48.08 (m, C1), 48.31 (m, C3), 122.18 (dm, qq, 1J ) 156
Hz, C4), 122.19 (dm, 1J ) 156 Hz, C7), 122.83 (sharp q, 3J ) 6.5
1
The first H NMR spectrum, recorded after 3 min, showed that
reagent 2c had vanished with consumption of 1 equiv of MeLi,
generating 5c as the only product. The tube was emptied into 0.2
M HCl (5 mL) and rinsed with Et2O and water. The Et2O extracts
were washed with distilled water until neutral, dried over Na2SO4,
and concentrated to provide crude 5c (104 mg, 93%), the purifica-
tion of which was remarkably difficult: Colorless 5c crystallized
very slowly from methanol (0.5 mL); the first crop was washed
with cold methanol (-18 °C, 2×), then recrystallized from pentane
(0.3 mL) at -18 °C, separated from the supernatant, and dried first
in a stream of N2 gas and then in vacuo over P4O10 and chipped
Hz, CR), 127.16 (ddd, J ) 159.7 Hz, J ) 7.4 Hz, J ) 1.2 Hz,
1
3
4
1
3
4
C5), 127.29 (ddd, J ) 159.7 Hz, J ) 7.4 Hz, J ) 1.2 Hz, C6),
149.09 (m, C9), 149.71 (m, C8), 159.79 (m, C2), 177.2 (sharp q, 3J
) 4.4 Hz, CO2H), assigned in accord with the C-H couplings
through 1H/13C heterocorrelations and through the following hetero-
multiple-bond correlations (HMBC): R-CH3 T CR (2J), 1-CH3 T
C1 (2J), and 3-CH3 T C3 (2J), in addition to the 3J connections C8
T 1-CH3 T C2 T 3-CH3 T C9, 1-CH3 T 1-CH3, 3-CH3 T 3-CH3,
C2 T R-CH3 T CO2H, 4-H T C6, 5-H T C7, 6-H T C4, and 7-H
T C5; FT-IR (diamond, ATR) 3200-2500 (vbr O-H), 1682 (s,
CdO), 1281, 1252, and 759 (vs) cm-1. Anal. Calcd for C16H20O2
(244.3): C, 78.65; H, 8.25. Found: C, 78.26; H, 8.24.
1
paraffin wax: mp 131-132.5 °C; H NMR (400 MHz, CDCl3) δ
1.53 (s, 2 3-CH3), 1.63 (s, 2 1-CH3), 2.61 (s, R-CH3), 7.12 (m,
4-H), 7.14 (m, 7-H), and 7.22 (m, 5-/6-H), assigned by the NOESY
correlations R-CH3 T 3-CH3 T 4-H and 1-CH3 T 7-H; 13C NMR
2-Phenyl-2-(1,1,3,3-tetramethyl-2-indanylidene)acetic Acid
(9d). Described in ref 11.
2-(3-Phenyl-2-propyn-1-ylidene)-1,1,3,3-tetramethylindan (13a).
Described in ref 6.
2-(1-Methyl-3-phenyl-2-propyn-1-ylidene)-1,1,3,3-tetrameth-
ylindan (13b). Described in ref 6.
1
3
(100.6 MHz, CDCl3) δ 28.1 (qq, J ) 127.5 Hz, J ) 4.3 Hz, 2
1-CH3), 29.0 (sharp q, 1J ) 129.1 Hz, R-CH3), 29.5 (qq, 1J ) 127.5
Hz, 3J ) 4.3 Hz, 2 3-CH3), 49.7 (m, C3), 50.1 (m, C1), 117.1 (q, 2J
1
) 7.3 Hz, CR), 122.2 and 122.6 (2 dm, J ) 156 Hz, C4,7), 127.2
1
3
and 127.3 (2 dd, J ) 160 Hz, J ) 8 Hz, C5,6), 149.3 (m, C9),
150.0 (m, C8), 154.2 (m, C2), assigned by 1H/13C heterocorrelation
and the selective {1H} decouplings {3-CH3} f C3 (narrowed) and
C9 (t), {1-CH3} f C1 and C8 (both narrowed), {R-CH3} f CR (s);
IR (KBr): 2981, 2960, 2925, 2863, 1637, 1591, 1486, 1456, 1364,
1086, 1018, 757 cm-1; MS (GC/EI, 70 eV) m/z (%) 280.1 (1.5)
and 278.1 (1.5, M+), 265.1 (100) and 263.1 (98.3, M+ - CH3),
199.2 (57, M+ - Br). Anal. Calcd for C15H19Br (279.2): C, 64.52;
H, 6.86. Found: C, 64.65; H, 6.97.
2-(1-Bromo-3-phenyl-2-propyn-1-ylidene)-1,1,3,3-tetrameth-
ylindan (13c). Addition of MeLi (0.118 mmol) in Et2O (0.085 mL)
to a solution of phenylacetylene (0.013 mL, 0.118 mmol) in Et2O
(0.60 mL) under argon cover gas led to the quantitative formation
of PhCtCLi (1H NMR in situ δ 7.32, dm, 2 o-H). Reagent 2c (20
mg, 0.058 mmol) was added but did not react with PhCtCLi in
the course of 4.5 h. Then the chain process was started at -70 °C
by injection of a tiny batch of MeLi (0.004 mL, 0.006 mmol), which
ceased to consume 2c at ≈70% conversion after <5 min at room
temperature. Injection of a further batch of 0.004 mL of the MeLi
solution completed the consumption of 2c. The mixture remained
unchanged over 8 days at -18 °C and was then poured onto solid
CO2. The usual separation with Et2O/NaOH as above afforded e1
mg of PhCtCCO2H along with 18 mg of a nonacidic product
mixture containing 5c (4%), 13c (71%), 13b6 (2%), and 13a6 (6%).
13c crystallized in colorless tufts with mp 117-118 °C (cooled
(b) From 2a at -70 °C: A solution of monobromide 2a (200
mg, 0.754 mmol) in anhydrous THF (10.0 mL) was stirred at
-70 °C under argon cover gas during the addition of MeLi (1.80
mmol) in Et2O (1.46 mL). The mixture was stirred for 1 h at
-70 °C, then poured onto solid CO2, warmed up, and dissolved in
Et2O and 2 M NaOH. The acidified NaOH layer furnished the
known2 acid 9a (29 mg, 17%). The Et2O phases were washed until
neutral, dried over MgSO4, and concentrated to provide a partly
solidifying oil containing 5a (13% yield), 5c (31%), and the terminal
1
ethanol); H NMR (600 MHz, CDCl3, numbering of Scheme 3) δ
J. Org. Chem, Vol. 72, No. 16, 2007 6089

Knorr et al.
1.68 (s, 2 1-CH3), 1.71 (s, 2 3-CH3), 7.17 (m, 4-/7-H), 7.26 (m,
5-/6-H), 7.37 (m, 2 m-H and 1 p-H), and 7.50 (m, 2 o-H); 13C
NMR (150.8 MHz, CDCl3) δ 27.5 (qq, 1J ) 127 Hz, 3J ) 4.5 Hz,
2 1-CH3), 28.3 (qq, 1J ) 127 Hz, 3J ) 4.5 Hz, 2 3-CH3), 50.9 (m,
C1), 51.2 (m, C3), 88.3 (sharp s, Câ), 95.6 (sharp s, CR), 95.8 (t, 3J
was added from a pressure-equalizing dropping funnel over a period
of no less than 15 min. (The inverse mode of adding LiSnMe3 to
8 promoted the formation of “butatriene” 16 as a troublesome side
product.) Stirring was continued for 30 min at -70 °C and then
for 60 min at room temperature. The reaction mixture was recooled
to -70 °C, and the excess of LiSnMe3 was quenched with methanol
(3 mL). The mixture was diluted with Et2O and distilled water,
and the Et2O layer was washed until neutral, dried over Na2SO4,
and concentrated to deliver a partially solidifying, pale yellow oil
(2.29 g) containing only 19 and nonaromatic methyltin compounds.
The recrystallization from preheated ethanol (10 mL) should be
performed rapidly, so to avoid the partial ethanolysis which would
form the mono(trimethylstannyl) derivative. The colorless glistening
flakes of 19 (1.62 g, 81%) were obtained analytically pure: mp
1
) 5.5 Hz, Cγ), 122.3 and 122.5 (2 dm, J ) 158 Hz, C4,7), 122.6
(m, Ci), 127.47 and 127.48 (2 dd, 1J ) 160 Hz, C5,6), 128.5 (2 dm,
1
3
1J ≈ 162 Hz, 2 Cm), 128.8 (dt, J ) 161 Hz, J ) 7.6 Hz, Cp),
131.2 (dm, 1J ≈ 164 Hz, 2 Co), 148.6 (m, C9), 149.2 (m, C8), 167.0
(m, C2), assigned by comparison of 1-/3-CH3 with 13a6 and by
2
HMBC (with window at 8 Hz) showing the J correlations 1-CH3
T C1 and 3-CH3 T C3 in addition to the 3J correlations 1-CH3 T
1-CH3 T C8, 3-CH3 T 3-CH3 T C9, 4-/7-H T C5,6, 5-/6-H T
C4,7, m-H T Ci, Cγ T o-H T Cp, p-H T Co; MS (GC/DEI, 70 eV)
m/z (%) 366.2 (7.5) and 364.2 (8.1, M+), 285.2 (83.6, M+ - Br),
199.2 (100). The constitution was confirmed through treatment with
MeLi in THF solution which furnished 13b.
1
123.5-125 °C (ethanol); H NMR (400 MHz, CDCl3) δ +0.32
(s, 119Sn satellites 2J ) 50.1 Hz, 2 SnMe3), 1.41 (s, 119Sn satellites
5
5
| Jcis + Jtrans| ) 3.5 Hz, 4 1-/3-CH3), 7.14 and 7.21 (AA′MM′
1,2-Bis(1,1,3,3-tetramethyl-2-indanylidene)ethene (16). The
separation of this “butatriene” derivative from the product mixture
was straightforward, by leaching with boiling ethanol (5-10 mL/g
of material) which left transparent blocks of 16: mp 238-240 °C
(hexane); 1H NMR (400 MHz, CDCl3) δ 1.49 (s, 8 1-/3-CH3), 7.23
and 7.27 (AA′BB′ system, 2 C6H4); 13C NMR (100.6 MHz, CDCl3)
δ 31.0 (qq, 1J ) 127.4 Hz, 3J ) 4.6 Hz, 8 1-/3-CH3), 48.5
system, C6H4); 13C NMR (100.6 MHz, CDCl3) δ -1.34 (qm, 1J )
3
1
3
128.3 Hz, J ≈ 1 Hz, J(119Sn) ) 323 Hz, J(119Sn) ) 9.2 Hz, 2
SnMe3), 31.3 (qq, 1J ) 126.7 Hz, 3J ) 4.4 Hz, 4J(119Sn, trans?) )
10.8 Hz, 4J(119Sn, cis?) ≈ 3 Hz, 4 1-/3-CH3), 50.7 (unresolved m,
3J(119Sn, trans) ) 92.7 Hz, J(119Sn, cis) ) 43.4 Hz, C1,3), 122.7
3
(dm, 1J ) 155 Hz, C4,7), 126.8 (ddd, 1J ) 159.3 Hz, 3J ) 7.4 Hz,
C5,6), 130.6 (br s, CR), 150.3 (unresolved, C8,9), 183.1 (unresolved,
C2); 119Sn NMR (100.75 MHz, CDCl3, external standard SnMe4)
1
(unresolved, 4 C1,3), 122.7 (dm, J ) 156 Hz, 4 C4,7), 127.2 (ddd,
1
2
1J ) 159 Hz, 3J ) 7 Hz, 4 C5,6), 137.9 (m, 2 C2), 148.9 (m, 4 C8,9),
153.8 (sharp s, 2 CR); IR (KBr): 3019, 2954, 2915, 2852, 1585,
1481, 1454, 1356, 1024, 744 (vs), and 502 cm-1; MS (EI, 70 eV)
m/z (%) 368 (43), 353 (100), 338 (21), 323 (14), 196 (52). Anal.
Calcd for C28H32 (368.6): C, 91.25; H, 8.75. Found: C, 90.90; H,
8.47. 16 is insoluble in methanol.
δ -44.8 (s, J(13CH3) ) 323.5 Hz, J(117Sn) ) 653 Hz, SnMe3);
IR (KBr): 3001, 2963, 2922, 2863, 1596, 1357, and 762 (vs) cm-1
.
Anal. Calcd for C20H34Sn2 (511.9): C, 46.93; H, 6.69. Found: C,
47.17; H, 6.69.
Acknowledgment. This work is dedicated to Professor
Herbert Mayr on the occasion of his sixtieth birthday. Long-
term sponsoring by the Deutsche Forschungsgemeinschaft is
gratefully acknowledged.
2-[Bis(trimethylstannyl)methylidene]-1,1,3,3-tetramethylin-
dan (19). The olive-colored solution of LiSnMe3 in THF (<17 mL),
obtained19 from 11.8 mmol of ClSnMe3, was cooled to -70 °C
under argon cover gas with stirring. (The green color30 may
disappear but will return during the following operation.) A solution
of the 1,1-dichloroalkene 8 (1.00 g, 3.92 mmol) in THF (10 mL)
Supporting Information Available: Tabular survey of reaction
conditions and products; further attempts toward reagent 2c;
preparation of trimethylstannyllithium. This material is available
free of charge via the Internet at http://pubs.acs.org.
(30) Weibel, A. T.; Oliver, J. P. J. Organomet. Chem. 1974, 82, 281-
290.
JO070623W
6090 J. Org. Chem., Vol. 72, No. 16, 2007