How the by-products hint at mechanisms and suggest new synthetic routes in some organolithium reactions

Article abstract of DOI:10.1002/poc.575

The recent proliferation of structural studies of organolithium reagents has led to an extended knowledge of the aggregation state of these pervasive synthetic intermediates. Nevertheless, detailed investigations of organolithium reaction mechanisms are still very limited. The most popular techniques in the relatively few reported studies are UV, EPR and NMR spectroscopy, the use of radical clocks and investigation of the stereochemical course of the reaction when applied. This paper describes how the search for minor products can be an additional mechanistic tool. Metal-halogen exchange was studied with a new suitable fast radical clock bearing a phenyl group at the alkene C-terminus, even when the probe was able to trap very short-lived radical intermediates, the results showed that the reaction proceeds through an open lithiated intermediate. In another study, radicals of benzil were generated on the lithium surface and their reactions studied in THF. Characterization of the minor products, the rates of decay of the reagent and formation of products, and also periodical EPR of the reaction mixture, allowed the proposal of the whole reaction mechanism. In both reactions studied, the detailed understanding of the reaction mechanisms based on the side-products provided new more economical and environmentally friendly alternatives for the synthesis of substituted 2,3-dihydrobenzo[b]furans and aromatic esters. Copyright

Full text of DOI:10.1002/poc.575

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2002; 15: 903–910
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.575
How the by-products hint at mechanisms and suggest new
synthetic routes in some organolithium reactions†
Norma Sbarbati Nudelman,1* Graciela V. GarcõÂa1 and Susana Velurtas2
1Depto. QuõÂmica OrgaÂnica, Facultad de Ciencias Exactas, Universidad de Buenos Aires, Ciudad Universitaria, 1428 Buenos Aires, Argentina
2Depto. de QuõÂmica, Facultad de Ciencias Exactas, Universidad Nacional de Mar del Plata, Ciudad Universitaria, 1428 Buenos Aires,
Argentina
Received 19 November 2001; revised 17 December 2001; accepted 24 January 2002
ABSTRACT: The recent proliferation of structural studies of organolithium reagents has led to an extended
knowledge of the aggregation state of these pervasive synthetic intermediates. Nevertheless, detailed investigations of
organolithium reaction mechanisms are still very limited. The most popular techniques in the relatively few reported
studies are UV, EPR and NMR spectroscopy, the use of radical clocks and investigation of the stereochemical course
of the reaction when applied. This paper describes how the search for minor products can be an additional mechanistic
tool. Metal–halogen exchange was studied with a new suitable fast radical clock bearing a phenyl group at the alkene
C-terminus, even when the probe was able to trap very short-lived radical intermediates, the results showed that the
reaction proceeds through an open lithiated intermediate. In another study, radicals of benzil were generated on the
lithium surface and their reactions studied in THF. Characterization of the minor products, the rates of decay of the
reagent and formation of products, and also periodical EPR of the reaction mixture, allowed the proposal of the whole
reaction mechanism. In both reactions studied, the detailed understanding of the reaction mechanisms based on the
side-products provided new more economical and environmentally friendly alternatives for the synthesis of
substituted 2,3-dihydrobenzo[b]furans and aromatic esters. Copyright  2002 John Wiley & Sons, Ltd.
KEYWORDS: organolithium; halogen–lithium exchange; lithium surface reactions; anionic cyclization; ate
complex; Li/oxygen carbenoid.
INTRODUCTION
mechanisms for organolithium reactions. It has been
shown recently that many addition reactions of organo-
lithium reagents to carbonyl compounds proceed via an
ET mechanism11–13 Previous studies in our laboratory on
the reaction of phenyllithium with carbon monoxide by
EPR and 13C NMR spectroscopy have provided evidence
that the reaction occurs through the intermediacy of
paramagnetic species.14 The UV and/or EPR spectro-
scopic observation of a radical intermediate does not
necessarily mean that a radical is involved in the major
reaction pathway, because both techniques detect inter-
mediates at very low concentrations,8 but a developed
13C NMR methodology allowed quantitative determina-
tions of the radical concentrations up to nearly 1 M,14b
and elucidation of the reaction mechanism.15 The top end
of the alkyl radical kinetic scale has been recently
adjusted by laser flash photolysis calibrations of fast
radical clocks.16
Although organolithium compounds are powerful re-
agents in organic synthesis,1,2 in only a few cases have
the reaction intermediates been isolated and fully
characterized;3–5 for most of the reactions the mechan-
isms by which they proceed remain unclear or poorly
understood. As an example, the mechanism of one of the
most powerful methods for the preparation of organo-
lithium compounds, halogen–metal exchange, is still
controversial.6 Bailey and Patricia reviewed the data and
they noted, as do most recent workers in the field, that the
pathway followed may be dependent on the reactants and
reaction conditions.7
Reactions with radical probes8,9 and the trapping of
carbanionic intermediates by deuteration10,11 have been
used as tools for deciding between the two most preferred
On the other hand, it has been shown recently that the
first step in the addition of aryllithiums to a,b-unsaturated
aldehydes is electron transfer from the organolithium
reagent to the carbonyl moiety.17 The main addition
product of PhLi, to (E)-cinnamaldehyde is the alcohol
(E)-1,3-diphenyl-2-propen-1-ol, as expected, but a care-
ful investigation of by-products showed the presence of
*Correspondence to: N. S. Nudelman, Depto. Qu´ımica Orga´nica,
Facultad de Ciencias Exactas, Universidad de Buenos Aires, Ciudad
Universitaria, 1428 Buenos Aires, Argentina.
E-mail: nudelman@qo.fcen.uba.ar
†Presented in part at the Sixth Latin American Conference on Physical
Organic Chemistry (CLAFQO-6), held at Isla Margarita, Venezuela,
during December 2001.
Contract/grant sponsor: Universidad de Buenos Aires.
Contract/grant sponsor: CONICET.
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

´
N. S. NUDELMAN, G.-V. GARCIA AND S. VELURTAS
904
(E)-chalcone and 1,3-diphenyl-1-propanone in minor
amounts. The remarkable feature in that reaction is the
coordination of a second lithium atom to the b-carbon in
the 1,2-adduct, leading to a b-lithiated cyclic adduct
whose structure determined by 13C NMR showed an Z3-
coordination of the lithium atom,18 as has been shown
recently by other organolithiums in solution.19
The knowledge of the reaction mechanism suggested a
new tandem strategy for the synthesis of interesting b-
alkyl-substituted dihydrochalcones, in good to excellent
yields.18,20 Tandem reactions have recently been of
interest in organic synthesis because they constitute a
convenient, economical and environmentally friendly
synthetic alternative when compared with conventional
stepwise reactions.21
In our recent research involving organolithium reac-
tions, we have observed that a careful search of the minor
by-products formed in the different steps under varying
reactions conditions could be an additional useful tool for
the elucidation of the mechanisms of their reaction and
the suggestion of original convenient synthetic routes. In
this paper, the reactions of organolithiums with aryl
halides and the generation of radicals on lithium surface
are discussed.
before.25 Benzil was prepared by reported methods and
crystallized from ethanol, m.p. 94–95°C (lit.24 94–95°C).
Alkyl halides were commercial or prepared by reported
methods and purified by distillation immediately prior to
use. All glassware, syringes and needles were dried in a
vacuum oven and cooled in a desiccator. 2-Bromophenyl
3-phenyl-2-propenyl ether, 4, was prepared by reaction of
2-bromophenolate with cinnamyl bromide in acetone as
described previously (yield 82%), m.p. 55–56°C (etha-
nol).26 An authentic sample of phenyl (E)-3-phenyl-2-
propenyl ether was prepared according to the literature.27
General procedure for the reaction of 4 with n-BuLi. A
solution of 0.05 mmol of 4 in 10 ml of THF (0.05 M) was
cooled to À85°C under a blanket of dry nitrogen and 1.5
equiv. of n-BuLi as a solution in hexane was added
dropwise via syringe over a 1–2 min period. The
temperature was maintained at À80°C during a chosen
time and the reaction mixture was quenched with 0.2 ml
of MeOH. The reaction mixture was washed with
aqueous NH4Cl solution, extracted with Et2O and dried
(MgSO4). The organic layer was analysed by GC and 1H
NMR. GC–MS of the reaction mixture was carried out
after 5 min of reaction.
[2-2H]Phenyl (E)-3-phenyl-2-propenyl ether,5-d 1. M.p.
1
67.5–68.5°C. H NMR (200 MHz, CDCl3), ꢀ 4.70 (dd,
EXPERIMENTAL
2H, J = 5.7 and 1.3 Hz), 6.42 (dt, 1H, J = 15.8 and
5.7 Hz), 6.74 (d, 1H, J = 15.8 Hz), 6.96 (m, 2H), 7.33 (m,
7H). MS, m/z (rel. abs.) 211 (5), 118 (45), 117 (100), 116
(25), 115 (70), 91 (27), 78 (6), 77 (5), 66 (8), 65 (7), 51
(8).
General methods. All reactions involving organolithium
reagents were carried out by using standard techniques
for the manipulation of air- and water-sensitive com-
pounds.22 All compounds reported here were fully
characterized by melting point (when applicable), mass
spectrometry (using a QP 5050A gas chromatograph–
mass spectrometer, Shimadzu) and nuclear magnetic
resonance spectroscopy (determined on a Bruker Model
200, 300 or 500 spectrometer operating at 200, 300 or
500 MHz for 1H and 50, 75 or 125 MHz for 13C). The 1H
chemical shifts are referenced relative to TMS and the
13C chemical shifts are referenced relative to CDCl3 at
ꢀ = 77.0 ppm. EPR spectra were determined on a Bruker
EPR 200D (X band) TE102 cavity using procedures
similar to those described for the reactions of phenyl-
lithium with CO.14a High-resolution mass spectra were
measured on a ZAB-SEQ4F mass spectrometer. GC
analyses were carried out on a Hewlett-Packard Model
5890 gas chromatograph using an HP-5 column.
3-(1-Phenylpentyl)-2,3-dihydrobenzo[b]furan, 6. Com-
pound 6 is obtained as a mixture of two diastereoisomers,
diastereomeric ratio (dr) = 80:20. The diastereomeric
1
ratio was determined by H NMR and GC analysis.
1
Main diastereoisomer: H NMR (200 MHz, CDCl3), ꢀ
0.80 (t, 3H, J = 7.1 Hz), 1.08 (m, 2H), 1.25 (m, 2H), 1.67
(m, 2H), 2.73 (m, 1H), 3.68 (dd, 1H, J = 6.2 and 8.8 Hz),
4.44 (dd, 1H, J = 6.2 and 9.1 Hz), 4.61 (dd, 1H, J = 8.8
and 9.1 Hz), 6.26 (d, 1H, J = 7.7 Hz), 6.59 (dt, 1H, J = 1.0
and 7.5 Hz), 6.71 (d, 1H, J = 8.0 Hz), 7.09 (m, 3H), 7.26
(m, 3H). 13C NMR (50 MHz, CDCl3), ꢀ 13.85, 22.56,
29.48, 33.04, 48.05, 50.37, 75.06, 109.29, 119.80,
125.41, 126.57, 128.10, 128.34, 128.53, 129.07, 142.98,
160.34. Minor diastereoisomer: 1H NMR (200 MHz,
CDCl3), ꢀ 0.79 (t, 3H, J = 7.1 Hz), 1.08 (m, 2H), 1.25 (m,
2H), 1.67 (m, 2H), 2.73 (m, 1H), 3.55 (ddd, 1H, J = 4.0,
8.8 and 8.8 Hz), 4.17 (dd, 1H, J = 4.4 and 9.1 Hz), 4.28
(dd, 1H, J = 8.8 and 9.1 Hz), 6.76 (d, 1H, J = 8.8 Hz),
6.85 (dt, 1H, J = 1.0 and 7.5 Hz), 7.10 (m, 4H), 7.26 (m,
3H). 13C NMR (50 MHz, CDCl3), ꢀ 13.85, 22.66, 29.40,
32.48, 47.92, 49.94, 75.52, 109.72, 119.91, 125.97,
126.40, 128.34, 128.49, 128.53, 129.50, 143.06, 160.60.
MS, m/z (rel. abs.) 266 (16), 209 (22), 120 (32), 119
(100), 115 (36), 92 (27), 91 (50), 77 (19), 65 (30). Anal.
Materials. Tetrahydrofuran and hexane were purified as
described previously;23 they were distilled from dark-
blue solutions of sodium benzophenone ketyl under
nitrogen immediately prior to use. Lithium wire (Merck,
>99%, 3 mm diameter) was weighed under ligroin,
washed with hexane and cut into small portions over a
reaction flask containing warm solvent. n-BuLi was
prepared as described previously;24 its concentration was
determined by the double titration method described
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

MECHANISMS OF ORGANOLITHIUM REACTIONS
905
Calcd for C19H22O: C, 85.67; H, 8.32. Found: C, 85.60;
H, 8.34%.
47.81, 52.28, 75.03, 109.29, 119.83, 125.41, 126.59,
128.10, 128.34, 128.59, 129.10, 142.60, 160.31. Minor
diastereoisomer: H NMR (200 MHz, CDCl3), ꢀ 0.71 (t,
1
1
Phenyl (Z)-3-phenyl-1-heptenyl ether,7. Oil. H NMR
3H, J = 7.3 Hz), 1.67 (m, 2H), 2.65 (dt, 1H, J = 4.4 and
9.1 Hz), 3.56 (ddd, 1H, J = 4.0, 4.4 and 8.8 Hz), 4.17 (dd,
1H, J = 4.0 and 9.1 Hz), 4.28 (dd, 1H, J = 8.8 and 9.1 Hz),
6.76 (d, 1H, J = 7.3 Hz), 6.85 (t, 1H, J = 7.3 Hz), 7.12 (m,
4H), 7.26 (m, 3H). 13C NMR (50 MHz, CDCl3), ꢀ 11.75,
25.68, 47.65, 51.66, 75.60, 109.72, 119.91, 126.00,
126.46, 128.34, 128.40, 128.51, 129.52, 142.74, 160.52.
MS, m/z (rel. abs.) 238 (25), 120 (33), 119 (100), 91 (51),
65 (33).
(200 MHz, CDCl3), ꢀ 0.86 (t, 3H, J = 6.6 Hz), 1.32 (m,
4H), 1.72 (m, 2H), 3.94 (m, 1H), 4.96 (dd, 1H, J = 6.2 and
9.9 Hz), 6.39 (d, 1H, J = 6.2 Hz), 7.00 (m, 3H), 7.27 (m,
7H). 13C NMR (50 MHz, CDCl3), ꢀ 14.02, 22.58, 29.70,
36.21, 40.34, 116.38, 117.05, 122.49, 125.90, 127.30,
128.38, 129.52, 139.74, 145.53, 157.49. MS, m/z (rel.
abs.) 266 (4), 209 (50), 131 (13), 115 (100), 91 (12), 77
(18). Anal. Calcd for C19H22O: C, 85.67; H, 8.32. Found:
C, 85.02; H, 8.16%.
3-(1-Phenylheptyl)-2,3-dihydrobenzo[b]furan.
M.p.
1
13C NMR spectrum of the reaction mixture of 4 with
RLi in THF-d8. Taking into account that the n-BuLi should
be prepared in hexane and that the butyl moiety is
incorporated in intermediate 10, the reaction to investi-
gate likely intermediates by 13C NMR was carried out
using PhLi as a more suitable RLi reagent for this
purpose. To a 5 mm NMR tube capped with a septum at
À90°C under a nitrogen atmosphere containing 96.3 mg
(0.3 mmol) of 2-bromophenyl 3-phenyl-2-propenyl ether,
4, in 0.5 ml of THF-d8, a solution of 84 mg (1 mmol) of
PhLi in 0.5 ml of THF-d8 was added. The mixture was
shaken, to ensure complete mixing, for 5 min at À90°C.
decoupled 13C NMR spectrum was determined. The
center peak of the downfield quintet of the THF-d8 was
used as the reference peak and was set at ꢀ 67.50 ppm. A
signal at ꢀ 148.1 ppm, that was not present at the
beginning or at the end of the reaction, was observed.
That signal is easily assigned to the ipso-carbon of the
aromatic ring bonded to the lithium-bearing carbon atom.
76–78°C, dr = 83:17. Main diastereoisomer: H NMR
(200 MHz, CDCl3), ꢀ 0.83 (t, 3H, J = 5.9 Hz), 1.18 (m,
8H), 1.66 (m, 2H), 2.72 (m, 1H), 3.67 (dd, 1H, J = 6.6 and
8.8 Hz), 4.43 (dd, 1H, J = 6.6 and 9.1 Hz), 4.61 (dd, 1H,
J = 8.8 and 9.1 Hz), 6.25 (d, 1H, J = 7.3 Hz), 6.59 (dt, 1H,
J = 1.0 and 7.3 Hz), 6.71 (d, 1H, J = 8.0 Hz), 7.12 (m,
3H), 7.26 (m, 3H). 13C NMR (50 MHz, CDCl3), ꢀ 13.96,
22.56, 27.22, 29.19, 31.64, 33.34, 48.03, 50.37, 75.06,
109.29, 119.80, 125.41, 126.57, 128.10, 128.32, 128.51,
129.07, 142.98, 160.31. Minor diastereoisomer: 1H NMR
(200 MHz, CDCl3), ꢀ 0.82 (t, 3H, J = 5.1 Hz), 1.18 (m,
8H), 1.66 (m, 2H), 2.72 (m, 1H), 3.54 (ddd, 1H, J = 4.0,
8.4 and 8.8 Hz), 4.17 (dd, 1H, J = 4.0 and 9.1 Hz), 4.29
(dd, 1H, J = 8.4 and 9.1 Hz), 6.76 (d, 1H, J = 8.8 Hz),
6.86 (dt, 1H, J = 1.0 and 7.3 Hz), 7.01 (m, 3H), 7.26 (m,
3H). 13C NMR (50 MHz, CDCl3), ꢀ 13.96, 22.56, 27.22,
29.29, 31.64, 32.82, 47.89, 49.94, 75.52, 109.72, 119.91,
125.97, 126.40, 128.32, 128.43, 128.51, 129.50, 143.06,
160.55. MS, m/z (rel. abs.) 294 (14), 120 (28), 119 (100),
118 (72), 92 (30), 91 (80), 65 (20).
1
Then it was allowed to warm to À50°C, and the H
General procedure for the synthesis of 2,3-dihydro-
benzo[b]furans. The organolithium derived from 4 was
generated as described above. After 5 min at À80°C, the
electrophile, pure or as a solution in THF, was added
rapidly via a syringe. The reaction mixture was
immediately allowed to reach 0°C and to stand for
5 min before quenching with MeOH. After work-up, the
products were isolated by TLC and identified by melting
point, 1H and 13C NMR. The diastereomers were
separated by chromatography over silica gel with a
mixture of 2% ethyl acetate, 70% cyclohexane and 28%
hexane as eluent. The compounds were isolated and fully
characterized.
General procedure for the reaction of benzil with
lithium wire in THF. Lithium wire was weighed under
ligroin, washed with THF and cut into small pieces in a
reaction flask containing a small portion of warm THF,
under a stream of dry argon. Preliminary runs cutting the
wire in very tiny pieces, or even working with lithium
suspension, showed that these changes of the lithium
surface seemed not to have a significant influence. The
reaction flask was capped with an ‘air-tight’ stopper and
alternatively evacuated and flushed with nitrogen several
times. This is a common technique in our laboratory and
it has been found that by carrying out the procedure fairly
fast, no complication with any lithium reaction with
dinitrogen giving the gray lithium nitride is observed. A
solution of benzil of the desired concentration in THF
was transferred by syringe, and allowed to reach the
working temperature (no significant temperature effect
was observed in the range 0–50°C, therefore most of the
reactions were carried out at room temperature). Aliquots
of 0.2 ml of the reaction mixture were taken at time
intervals, quenched with MeOH and analysed by GC,
using benzophenone as internal standard. Since this is a
3-(1-Phenylpropyl)-2,3-dihydrobenzo[b]furan. Oil, dr =
80:20. Main diastereoisomer: 1H NMR (200 MHz,
CDCl3), ꢀ 0.75 (t, 3H, J = 7.1 Hz), 1.70 (m, 2H), 2.65
(dt, 1H, J = 5.5 and 9.1 Hz), 3.69 (dt, 1H, J = 6.2 and
9.1 Hz), 4.43 (dd, 1H, J = 6.2 and 9.1 Hz), 4.61 (dd, 1H,
J = 9.1 Hz), 6.28 (d, 1H, J = 7.3 Hz), 6.59 (t, 1H,
J = 7.3 Hz), 6.71 (d, 1H, J = 8.0 Hz), 7.12 (m, 3H), 7.26
(m, 3H). 13C NMR (50 MHz, CDCl3), ꢀ 11.99, 26.22,
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

´
N. S. NUDELMAN, G.-V. GARCIA AND S. VELURTAS
906
heterogeneous reaction, runs were made at least in
triplicate and the reproducibility was satisfactory.
probe reactions.39 The finding of cyclized products, 2, is
usually interpreted as evidence for a radical mechanism,
while open-chain products, 3, indicate that carbanionic
intermediates are involved [Eqn (1)].
EPR spectra. The independently prepared radical anion
solutions or the reaction mixtures in preparative con-
centrations (0.2–1.5 M) were diluted to nearly 10À5
M
with THF. Optimization of the final concentration was
adjusted in each case. The temperature was kept constant
using a Dewar cavity with a thermostated stream of air;
calibration was carried out at different modulation
amplitudes using a THF solution of galvinoxyl.
It should also be noted, however, that the finding of
open-chain products from radical probes does not
definitely rule out a radical intermediate during the
formation of the aryl carbanions. Indeed, the radical
probe could not be efficient enough to trap a very short-
lived aryl radical through a fast radical cyclization. In the
case of X = Br and Z = O the intramolecular cyclization
of the radical derived from the 2-bromophenyl 3-
phenylprop-2-enyl ether, 4, was recently found to be kC
>8.1 Â 109 sÀ1, 30°C40 (Ref. 40a corrects the value
initially given by Johnston et al.40b). This radical probe
has been used recently to establish the participation of
radicals during the formation of aryl Grignard reagents.41
The main product of the reaction of 4 with n-BuLi in THF
at À80°C is an open-chain product, 5, which has a
structure similar to 3 with Z = O, plus two unexpected by-
products, 6 and 7.
RESULTS AND DISCUSSION
The halogen–lithium exchange reaction is of one of the
most powerful methods for the preparation of organo-
lithium compounds, but although widely used, its mech-
anism is still controversial. Up to now, four different
mechanisms have been suggested, each of which has its
adherents and all possibilities appear in the recent
literature.6,7 The earliest suggestions of a concerted SN2
displacement and an ‘ate complex’ intermediate were
made by Sunthankar and Gilman28 and Wittig and
Scho¨llkopf,29 respectively. A ‘four-center’ mechanism1
was initially suggested to explain retention of configura-
tion when the reaction was carried out with asymmetric
halides; nevertheless, more recent studies established that
the percentage retention of configuration observed can be
easily accommodated within mechanisms involving other
intermediates.30 Intermediates such as ate complexes
have been firmly established in the aryllithium–aryl
iodide exchange31 and recently correlated with theoret-
ical studies.32 Beak and co-workers6,33 have shown that
evolution of the transition structure geometry for reaction
of aryl bromide with alkyllithium is consistent with a
trigonal bipyramid structure of a ‘10-Br-2-ate complex’
or an SN2 transition state; recent kinetic studies of
organolithium–aryl bromide exchange have been also
interpreted in terms of an SN2 mechanism.34 On the other
hand, evidence for a single electron transfer mechanism
has been afforded by running ESR35 and NMR36
spectroscopy, and more recently it has been strongly
supported by the observation of coupling11,24 and rear-
ranged products,37,38 especially from radical probes. The
UV and/or EPR spectroscopic observation of a radical
intermediate does not necessarily mean that a radical is
involved in the major reaction pathway leading to
products, because both techniques detect intermediates
at very low concentrations.8
The products were isolated and fully characterized by
spectroscopic determinations and independent synthesis.
By quenching the reaction mixture after 5 min at À80°C
with MeOD, high deuterium incorporation (>95%) was
observed in the 2-position of the open product 5,
indicating the structure of the open lithiated intermediate
8 (Scheme 1). Efforts were then concentrated on
improving the yields of the by-products 6 and 7, to
obtain insights into the mechanisms of their formation.
By allowing longer reaction times, the yields of 6 and 7
increase at the expense of 5; the yield of 6 especially
increases (up to 47% after 60 min of reaction), suggesting
that 8 could be the precursor of 6 (Table 1, entries 1 and
2). To confirm this assumption, the effect of increasing
the temperature of the reaction was examined. Each
experiment was performed at least three times, and
reproducible yields of products were obtained. Entries 3–
5 in Table 1 summarize the drastic effect of temperature
upon the yield of 5. Warming the solution from À80°C to
various temperatures leads to complete isomerization of
the intermediate, yielding 5 in only 5 mins at tempera-
tures ꢀÀ50°C.
To afford some new elements for the alkyllithium–ary
bromide exchange reaction, we synthesized a new
suitable fast radical clock, 1, bearing a phenyl group at
the alkene C-terminus, able to trap very short-lived
radical intermediates. The kinetic accelerating effect of
aryl group substitution in the opening and closing of
carbinyl radicals has been used recently in intramolecular
Scheme 1 shows the whole reaction mechanism,
outlined on the basis of the experimental findings. The
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

MECHANISMS OF ORGANOLITHIUM REACTIONS
907
Scheme 1
À50°C (conditions 2 in Table 1), the [2-2H]phenyl (E)-3-
phenyl-2-propenyl ether 5-d1 was obtained.
Table 1. Reaction of 2-bromophenyl 3-phenylprop-2-enyl
ether, 4, with n-BuLi in THFa
To gain an insight into the likely involvement of the
‘ate complex,’ the reaction with 1 (X = O, Z = O) was
also studied under the same conditions; no evidence of
any lithium–chlorine exchange was observed. As was
shown by recent calculations for X = I, Br, the formation
of the ‘ate complex’ is highly exothermic while the MP2/
II energy for Cl is ‡0.4.32
Relative yield (%)b
Temperature
Time
(min)
Entry
(°C)
5
6
7
1
2
3
4
5
À80
À80
À70
À50
0
5
60
5
5
5
90
39
37
0
5
5
47
53
87
84
14
10
14
16
0
The phenyl allyl ether structure allows stabilization of
the carbanion through coordination of the lithium atom to
the delocalized p-electrons of the partially isomerized
double bond, giving 9. Quenching the reaction at longer
times showed the slow cyclization of the first formed
lithium intermediate to the cyclic lithium intermediate
10, which yields the substituted 2,3-dihydrobenzo[b]-
furan 6, by further reaction with n-BuBr previously
generated in the halogen–metal exchange reaction.
Evidence for intermediate 10 was found in the 13C
NMR of the reaction mixture of 4 with PhLi in THF-d8 at
À80°C. In the NMR spectrum measured immediately
after mixing, a signal at ꢀ 148.1 ppm was observed,
which could be unambiguously assigned to the ipso-
carbon of the aromatic ring bonded to the lithium-bearing
carbon atom.
a
[4] = 0.05 M; [n-BuLi] = 0.75 M.
Relative yields determined by GC and NMR.
b
organolithium from 4 is assumed to be formed through an
‘ate complex’ intermediate, which is favored by the
presence of Br and O in the substrate.10 Recent
calculations show the lithium atom involved in a T-
shaped hypervalent halogen species for bromine and
iodine;32,42 ate complexes have been observed in the
reaction mixtures involving Li–I exchange in THF.43 The
presence of the open lithiated intermediate 8, was
established by the high deuterium incorporation
(>95%) in the ortho-position observed on quenching
the reaction with MeOD after short reaction times at very
low temperatures (conditions 1 in Table 1). In contrast,
on quenching with MeOD the reaction carried out at
Another observation which supports the polar transi-
tion state rather than the ET is that the formation of o-
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

´
N. S. NUDELMAN, G.-V. GARCIA AND S. VELURTAS
908
bromophenol as by-product is not observed, whereas in
the reaction of magnesium with 4 under the same
conditions (temperature and solvent), where an ET from
magnesium to 4 is generally accepted, o-bromophenol is
regularly formed as a by-product.41 All these observa-
tions converge to suggest the absence of radicals in the
reaction of n-BuLi with 4.
the minor products suggested a new synthetic route for
the synthesis of benzil benzoate, which is a highly
convenient, economic and environmentally friendly
alternative for the quantitative synthesis of the highly
pure ester.46b
Regarding the by-product 7, GC–MS of the crude
mixture revealed no deuterium incorporation at the aryl
C-2 position. Furthermore, results of experiments 4 and 5
in Table 1, and particularly the yield of 4, are identical
when 1.5 or 4 equiv. of n-BuLi are used. These
experimental facts hint that product 7 is formed via
intramolecular rearrangement of the carbanionic inter-
mediate 9 with the allylic protons (Scheme 1), followed
by isomerization to the more stable benzilic carbanion
and then reaction with n-BuBr. The cis structure of the
by-product 7 indicates the resonance 9 $ 11 shown in
Scheme 1; formation of 7 can be visualized through an
intermediate 11, stabilized by Li–oxygen coordination.
The crystal structure of [2-lithio-3-bromobenzofuran
diisopropyl ether]2, an a-lithiated ether, has recently
been determined.44 The structure shows that lithium
bridges C and O. Calculations on simpler lithiated ethers
at a high level are consistent with Li–oxygen carbenoid
structures.45
The finding of the route 4 → 9 → 10 → 6 also suggests
a likely route for the synthesis of other substituted 2,3-
dihydrobenzo[b]furans. The study of a synthetic metho-
dology based on the anionic cyclization of 10 in the
presence of suitable electrophiles, thus providing a
convenient tandem sequence for the synthesis of alkyl-
substituted 2,3-dihydrobenzo[b]furans, is in progress
with successful preliminary results for ethyl and hexyl
bromide (see yields in Scheme 1). Taking into account
the present and previous results, we envisage the lithium–
halogen exchange as a mechanistic spectrum with radical
and carbanion in each extreme, the precise structure/s of
the transition state(s) prevailing in any case depending on
the substrate structure, the nature of the halogen, the
solvent and any other reaction conditions.
Organolithium-derived radicals can also be generated
by the reaction of suitable substrates with lithium metal.
We have recently published a study of the reactions in
THF of benzaldehyde on a lithium surface.46 As
expected, the main reaction product is benzil alcohol,
but determinations of minor products (e.g. benzoin and
benzil) proved to be mechanistically significant. In fact,
they suggested the intermediacy of a benzoyl radical, that
could be trapped by CCl4. The rate of disappearance of
benzaldehyde, a and the rate of formation of the three
products, were measured and the kinetics of all the
reactions pathways involved were determined.46a The
first step was proved to be an ET from lithium to
benzaldehyde producing a radical anion–lithium cation
pair.
We have now studied the reactions of benzil, 12, with
lithium metal. The rates of decay of 12 and of the
formation of products and by-products were followed by
GC as a function of time. Most of the reaction occurs in
the first 60 min but the reaction is complete in 15–24 h
depending on the temperature, the concentration of 12
and the mass of lithium. The main reaction products are:
1,2-diphenylacetophenone, 13, and 1,2-diphenylethane,
14, but careful searching for other products showed the
presence of by-products, 1,2-diphenyl-1-ethanol, 15, and
1,2-diphenyl-1-ethene, 16, in very tiny amounts, that,
together with other evidence, give an insight into the full
reaction mechanism. No reaction was detected by the
quantitative GC determinations in the first 10–15 min,
depending on the temperature and reagent concentration.
The induction period for the decay of 12 was the same as
that for the appearance of 13; almost 90% of 13 is formed
in the first 60 min of reaction; in contrast, 14 appears only
after 20 min (as traces) and reaches its final yield after
nearly 10 h of reaction.
EPR of the reaction mixture was performed periodi-
cally as a probe for the involvement of radicals as
reaction intermediates. Signals appear after 12–15 min of
reaction depending on the reaction conditions. The EPR
spectrum of the reaction mixture in THF ([12] = 0.2 M,
Li = 428 mg) at 298 K after 20 min of reaction is shown
in Fig. 1. The total spectral width is Da = 7.64 G and it
shows hyperfine structure fully consistent with the
dibenzoyl anion [coupling constants: aH = 0.99 G (4),
0.36 G (4) and 1.12 G (2) for the ortho, metha and para-
H, respectively. The reported frequencies for the
dibenzoyl anion chemically generated in THF are 0.90;
0.41 and 0.90, respectively, Da = 7.0 G].12 A close
correspondence was found between the induction period
Figure 1. EPR of the reaction of benzil, 12, with lithium in
THF at 298 K, [12] = 0.2 M; M = 428 mg. Details of the EPR
Li
conditions: n = 9.57 Hz; ®eld control 3436.5 G; span width
DG = 20 G; modulation = 0.16 Gp; t = 200 ns; scanning
rate = 2 ks
The knowledge of the whole reaction scheme based on
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

MECHANISMS OF ORGANOLITHIUM REACTIONS
909
sium surface, induction periods are very well known, and
plausible hypotheses abound,47 but there is no compre-
hensive, documented understanding of the factors that
create the induction period or constitute initiation.48
CONCLUSIONS
This paper has described how the careful investigation of
the by-products formed in tiny amounts in some
organolithiums reactions and the survey of the different
parameters that influence their formation can provide
insights into the specific mechanisms involved. They also
afforded important clues on the likely transition states
and reaction intermediates, indicating that both radical
and carbanionic mechanisms are possible and that the
pathways followed are strongly dependent on the
reactants and reaction conditions. Finally, a thorough
knowledge of the reaction mechanism could suggest new
convenient, economical and environmentally friendly
synthetic methodologies for the preparation of alkyl-
substituted 2,3-dihydrobenzo[b]furan and of aryl esters,
with important advantages over the usual conventional,
stepwise reactions.
Scheme 2
shown by the EPR signal appearance and that observed
for the decay of 12 by GC, indicating that formation of
the radical could be the first reaction step.
All the evidence obtained is consistent with the overall
reaction mechanism shown in Scheme 2. Adsorption of
the reagent on the lithium surface was found to be the
initial slow step in the reaction of benzaldehyde with Li,
and it also occurs in the reaction of 12 (all adsorption
phenomena are symbolized by kS in the Scheme 2). These
are followed by a slow one-electron transfer from the
lithium surface to the adsorbed 12, measured by k1,
forming the radical anion (17)–lithium cation pair; 17
partially escapes the surface and can be detected by EPR
of the THF reaction mixture. Further successive reduc-
tions of 17 by lithium metal gives intermediates 18 → 19,
and the latter by loss of lithium oxide gives intermediate
20, precursor of 13, one of the main products (24% yield
in the first 60 min, final yield 35%). Similar reactions
afford intermediates 21, 22 and 23, that are the precursors
of the other isolated products (only 5% of 14 is produced
in the first 60 min of reaction; yields of 14, 15 and 16 after
24 h of reaction are 24, 5 and 6%, respectively).
Treatment of the reaction mixture after 24 h with
deuterated water showed the disappearance of the
Acknowledgements
Financial support from the Universidad de Buenos Aires
and CONICET is gratefully acknowledged.
REFERENCES
1. Wakefield BJ. Organolithium Methods. Academic Press: London,
1988.
2. Jamison TF, Shambayati S, Crowe WE, Schreiber SL. J. Am.
Chem. Soc. 1997; 119: 4353–4363.
¨
˚
3. Hilmersson G, Arvidsson PI, Davidsson O, Hakansson M.
Organometallics 1997; 16: 3352–3358.
4. Adler M, Marsh M, Nudelman NS, Boche G. Angew. Chem., Int.
Ed. Engl. 1999; 38: 1261–1262.
1
aliphatic protons of 13 and 14 in the H NMR spectrum
5. (a) Nudelman NS, Garcı´a Lin˜ares GE. J. Org. Chem. 2000; 65:
1629–1635; (b) Nudelman NS, Schulz HG, Garcı´a Lin˜ares GE,
Bonatti AE, Boche G. Organometallics 1998; 17: 146–150.
6. Beak P, Allen DJ. J. Am. Chem. Soc. 1992; 114: 3420–3425.
7. Bailey WF, Patricia JJ. J. Organomet. Chem. 1988; 352: 1–46.
8. (a) Ashby EC, Park B, Patil GS, Gadru K, Gurumurthy R. J. Org.
Chem. 1993; 58: 424–437; (b) Ashby EC. Acc. Chem. Res. 1988;
21: 414–422.
9. (a) Newcomb M, Williams WG, Crumpacker EL. Tetrahedron
Lett. 1985; 26: 1183; (b) Bailey WF, Patricia JJ, Nurmi TT, Wang
W. Tetrahedron Lett. 1986; 27: 1861–1864; (c) Bailey WF,
Patricia JJ, Nurmi TT, Wang W. Tetrahedron Lett. 1986; 27:
1865–1868.
of the reaction mixture. Traces of pinacol were detected
by periodical HPLC of the reaction mixture: its formation
shows a similar induction period to that of 13; it reaches
the maximum yield after about 40 min of reaction and
then decreases to almost negligible amounts after 60 min.
This is further evidence of the reaction sequence
18 → 19 → 20, which should be faster than the others
shown in Scheme 2 on the basis of the relative yields of
products observed as a function of time.
The reasons for the induction period are not yet clear.
Adsorption of the reagent on the lithium surface was
proved to occur: strong interaction between the adsorbate
and the metal causes the formation of a layer (detected by
GC at the metal surface). Nevertheless, the long
induction period could be related to additional surface
processes. In the Grignard reagent formation on magne-
10. Boche G, Lohrenz JCW, Opel A. In Lithium Chemistry: a
Theoretical and Experimental Overview, Sapse A-M, Schleyer
PvR (eds). Wiley: New York, 1995; Chapt. 7.
11. Ashby EC, Pham T. J. Org. Chem. 1987; 52: 1292–1300.
12. (a) Maruyama K, Katagiri T. J. Phys. Org. Chem. 1991; 4: 381; (b)
Maruyama K, Katagiri T. J. Phys. Org. Chem. 1991; 4: 158, and
references cited therein.
13. (a) Yamataka H, Miyano N, Hanafusa T. J. Org. Chem. 1991; 56:
2573; (b) Yamataka H, Fujimura N, Kawafuji Y, Hanafusa T. J.
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910

´
N. S. NUDELMAN, G.-V. GARCIA AND S. VELURTAS
910
Am. Chem. Soc. 1987; 109: 4305–4308; (c) Matsuyama T,
Yamataka H, Hanafusa T. Chem. Lett. 1988; 1367–1368.
14. (a) Nudelman NS, Doctorovich F, Amorin G. Tetrahedron Lett.
1990; 31: 2553; (b) Doctorovich F, Nudelman NS. Magn. Reson.
Chem. 1990; 28: 576–579.
15. Nudelman NS, Doctorovich F. Tetrahedron 1994; 50: 4651–4666.
16. Newcomb M, Choi S-Y, Horner JH. J. Org. Chem. 1999; 64: 1225–
1231.
17. Nudelman NS, Schulz HG, Garc´ıa GV. J. Phys. Org. Chem. 1998;
11: 1–9.
18. Nudelman NS, Garcı´a GV. J. Org. Chem. 2001; 66: 1387–1394.
19. Ahlbrecht H, Harbach J, Kalinowski HO, Lang A, Maier G. Chem.
Ber. 1997; 130: 683–686.
20. Nudelman NS, Garcı´a GV, Schulz HG. J. Org. Chem. 1998; 63:
5730–5731.
21. Tietze LF. Chem. Rev. 1996; 96: 139–148.
22. Shriver DF. The Manipulation of Air Sensitive Compounds. Wiley:
New York, 1989.
23. Pe´rez DG, Nudelman NS. J. Org. Chem. 1988; 53: 408–413.
24. Vitale AA, Doctorovich F, Nudelman NS. J. Organomet. Chem.
1987; 332: 9–18.
25. Trzupek LS, Newirth TL, Kelly EG, Nudelman NS, Whitesides
GM. J. Am. Chem. Soc. 1973; 95: 8118–8133.
26. Moody CJ, Doyle KJ, Elliot MC, Mowlem TJ. J. Chem. Soc.,
Perkin Trans. 1 1997; 2413–2417.
27. White WN, Fife WK. J. Am. Chem. Soc. 1961; 83: 3846–3853.
28. Sunthankar SV, Gilman H. J. Org. Chem. 1951; 16: 8–16.
29. (a) Wittig G, Scho¨llkopf V. Tetrahedron 1958; 3: 91; (b) Wittig G.
Angew. Chem. 1958; 70: 65.
30. Boche G, Walborsky HM. In The Chemistry of the Cyclopropyl
Group, Rappaport Z (ed). Wiley: Chichester, 1987; 701.
31. Reich HJ, Green DP, Phillips NH. J. Am. Chem. Soc. 1991; 113:
1414–1416.
32. Boche G, Schimeczek M, Cioslowski J, Piskorz P. Eur. J. Org.
Chem. 1998; 1851–1860.
Winkler HJS, Winkler H. J. Am. Chem. Soc. 1966; 88: 964–969;
(c) Winkler HJS, Winkler H. J. Am. Chem. Soc. 1966; 88: 969–
974.
35. (a) Fischer H. J. Phys. Chem. 1969; 73: 3834; (b) Russell GA,
Lamson DW. J. Am. Chem. Soc. 1969; 91: 3967–3968.
36. (a) Ward HR. Acc. Chem. Res. 1972; 5: 18; (b) Lawler RG. Acc.
Chem. Res. 1972; 5: 25–32.
37. Ashby EC, Coleman D, Gamasa M. J. Org. Chem. 1987; 52: 4079.
38. Jenneskens LW, de Boer HJR, de Wolf WH, Bickelhaupt F. J. Am.
Chem. Soc. 1990; 112: 8941–8949.
39. Horner JH, Musa OM, Bouvier A, Newcomb M. J. Phys. Org.
Chem. 2000; 13: 688–692.
40. (a) Garden SJ, Avila DV, Beckwith ALJ, Ingold KU, Lusztyk J. J.
Org. Chem. 1996; 61: 805–809; (b) Johnston LJ, Lusztyk J,
Wayner DDM, Abeywickreyma AN, Beckwith ALJ, Scaiano JC,
Ingold KU. J. Am. Chem. Soc. 1985; 107: 4594–4596.
41. (a) Bodineau N, Mattalia J-M, Thimokhin V, Handoo K, Ne´grel J-
C, Chanon M, Org. Lett. 2000; 2: 2301–2304; (b) Bodineau N,
Nudelman NS, Garc´ıa GV, Mattalia J-M, Roge´rio M, Arbelot M,
Chanon M, ARKIVOC 2002; 5: 139–150.
42. Jedlicka B, Crabtree RH, Siegbahn EM. Organometallics 1997;
16: 6021–6025.
43. Farnham WB, Calabrese JC. J. Am. Chem. Soc. 1986; 108: 2449–
2451.
44. (a) Boche G, Bosold F, Zulauf P, Marsch M, Harms K, Lohrenz J.
Angew. Chem. Int. Ed. Engl. 1991; 30: 1455; (b) Boche G, Opel A,
Marsch M, Harms K, Haller F, Lohrenz J, Thummler C, Koch W.
Chem. Ber. 1992; 125: 2265.
45. Viruela-Martin P, Viruela-Martin R, Toma´s F, Nudelman NS. J.
Am. Chem. Soc. 1994; 116: 10110–10116.
46. (a) Nudelman NS, Mendiara S. J. Phys. Org. Chem. 1997; 10: 233–
237; (b) Nudelman NS, Mendiara S. Tetrahedron Lett. 1997; 38:
2245–2248.
47. Silverman GS, Rakita PE. In Handbook of Grignard Reagents,
Heinemann H (ed). Marcel Dekker: New York, 1996.
48. Garst JF, Ungvary F. In Grignard Reagents: New Developments,
Richey HG (ed). Wiley: Chichester, 2000; Chapt. 7, 222–260.
33. Beak P, Allen DJ, Lee WK. J. Am. Chem. Soc. 1990; 112: 1629–
1630.
34. (a) Rogers HR, Houk J. J. Am. Chem. Soc. 1982; 104: 522–525; (b)
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 903–910