Metal-induced reductive cleavage reactions: An experimental and theoretical (MNDO) study on the stereochemical puzzle of birch and vinylogous birch processes

Article abstract of DOI:10.1021/jo951834g

The stereochemical puzzle posed by the lithium-promoted Birch and vinylogous Birch reductive cleavage of unsaturated benzyl ethers (BICLE; takes place with retention of configuration of the sensitive ²Δ double bond) and the corresponding cinnamyl analogs (VIBICLE; gives rise to ca. 2.5:1 E:Z mixtures) has been approached by experimental and theoretical means. NMR experiments indicate that the π-type organolithium compounds resulting from these reactions do not form observable mixed aggregates with the lithium silyloxide species generated alongside in the reaction and do not undergo observable isomerization at the temperature of operation. A simplified model for contact, solvent-separated, and isolated ion pairs has allowed us to evaluate these complex reactions in great detail from a theoretical viewpoint, using the MNDO semiempirical method. Relevant features that come out from these comprehensive studies, for which we have employed lithium naphthalenide (LiNaph) or lithium benzenide (LiBenz) as promoters, are as follows: (1) the lowest energy routes for cleavage are those involving contact ion pairs (CIPs) in which the lithium counterion plays a key role by acting as a handle (Lewis acid) to which the leaving group ^-OR adheres prior to detachment; (2) the different haptomeric structures which reside (local minima) in the potential hypersurface of either the so-called radical anion or the dianion routes show that haptomeric activation is key to understanding cleavage of the C-O bond which, eventually, takes place as a syn β elimination of LiOR; and (3) reductive cleavage of unsaturated benzyl ethers (BICLE) involves transient cation/anion radicals which undergo cleavage and subsequent reduction to the final organolithium with retention of configuration, in accordance with experiment, whereas that of vinylogous cinnamyl ethers (VIBICLE) involve transient dianion/dication species resulting from long-lived cation/anion radicals. In good qualitative agreement with experiment, MNDO finds two diastereomeric routes (ΔΔG* = 0.2 kcal/mol) for cleavage of (appropriately substituted) cinnamyl ethers, but only one for cleavage of the unsaturated benzyl analogs.

Full text of DOI:10.1021/jo951834g

J . Org. Chem. 1996, 61, 1035-1046
1035
Meta l-In d u ced Red u ctive Clea va ge Rea ction s: An Exp er im en ta l
a n d Th eor etica l (MNDO) Stu d y on th e Ster eoch em ica l P u zzle of
Bir ch a n d Vin ylogou s Bir ch P r ocesses
J ose´ M. Saa´,* Pablo Ballester, Pere M. Deya´,* Magdalena Capo´, and Xavier Garc´ıas
Departament de Quı´mica, Universitat de les Illes Balears, 07071 Palma de Mallorca, Spain
Received October 11, 1995X
The stereochemical puzzle posed by the lithium-promoted Birch and vinylogous Birch reductive
cleavage of unsaturated benzyl ethers (BICLE; takes place with retention of configuration of the
2
sensitive ∆ double bond) and the corresponding cinnamyl analogs (VIBICLE; gives rise to ca. 2.5:1
E:Z mixtures) has been approached by experimental and theoretical means. NMR experiments
indicate that the π-type organolithium compounds resulting from these reactions do not form
observable mixed aggregates with the lithium silyloxide species generated alongside in the reaction
and do not undergo observable isomerization at the temperature of operation. A simplified model
for contact, solvent-separated, and isolated ion pairs has allowed us to evaluate these complex
reactions in great detail from a theoretical viewpoint, using the MNDO semiempirical method.
Relevant features that come out from these comprehensive studies, for which we have employed
lithium naphthalenide (LiNaph) or lithium benzenide (LiBenz) as promoters, are as follows: (1)
the lowest energy routes for cleavage are those involving contact ion pairs (CIPs) in which the
lithium counterion plays a key role by acting as a handle (Lewis acid) to which the leaving group
-OR adheres prior to detachment; (2) the different haptomeric structures which reside (local minima)
in the potential hypersurface of either the so-called radical anion or the dianion routes show that
haptomeric activation is key to understanding cleavage of the C-O bond which, eventually, takes
place as a syn â elimination of LiOR; and (3) reductive cleavage of unsaturated benzyl ethers
(BICLE) involves transient cation/anion radicals which undergo cleavage and subsequent reduction
to the final organolithium with retention of configuration, in accordance with experiment, whereas
that of vinylogous cinnamyl ethers (VIBICLE) involve transient dianion/dication species resulting
from long-lived cation/anion radicals. In good qualitative agreement with experiment, MNDO finds
two diastereomeric routes (∆∆G* ) 0.2 kcal/mol) for cleavage of (appropriately substituted) cinnamyl
ethers, but only one for cleavage of the unsaturated benzyl analogs.
In tr od u ction
at the stereocontrolled synthesis of prenylated quinones
and aromatics, we used the so-called vinylogous Birch
hydrogenolysis (VIBIHY) of cinnamyl silyl ethers as well
as the Birch hydrogenolysis (BIHY) of the corresponding
unsaturated benzyl silyl ethers as the key steps for the
2∆ stereocontrolled introduction of prenyl chains into
aromatics.5 In this study, we uncovered a previously
unnoticed observation, namely that vinylogous Birch
hydrogenolysis of cinnamyl alcohols (as silyl ethers)
rendered the expected isoprenyl chains as 2∆ E,Z mix-
tures (ca. 2.5:1), whereas Birch hydrogenolysis of the
corresponding E or Z unsaturated benzyl alcohols (also
The metal-induced reductive cleavage of benzylic ethers
and related compounds1 is commonly used in preparative
synthetic organic chemistry.2 Nonetheless, both the
detailed mechanism3 and the actual structure of the final
species thereby generated, i.e., in the case of a lithium-
promoted reaction, an organolithium compound together
with a lithium salt (alkoxide, sulfide, halide, or other),
are largely ignored at present.4 In a recent study aimed
X Abstract published in Advance ACS Abstracts, J anuary 1, 1996.
(1) (a) Maercker, A. Angew. Chem., Int. Ed. Engl. 1987, 26, 972. (b)
Tiecco, M. Synthesis 1988, 749. (c) Bhatt, M. V.; Kulkarni, S. U.
Synthesis 1983, 249. (d) Cohen, T. Bhupathy M. Acc. Chem. Res. 1989,
22, 152. For earlier literature on the subject, see the references cited
in the above review articles.
(2) (a) Hook, I. M.; Mander, L. N. Nat. Prod. Rep. 1986, 3, 35. (b)
Birch, A. J . J . Chem. Soc. 1945, 89. (c) Birch, A. J .; Maung, M.; Pelter,
A. Aust. J . Chem. 1969, 22, 1923. (d) Birch, A. J .; Maung, M.
Tetrahedron Lett. 1967, 8, 3275. (e) Flisak, J . R.; Hall, S. S. J . Am.
Chem. Soc. 1990, 112, 7299. (f) McEnroe, F. J .; Sha, C.-K; Hall, S. S.
J . Org. Chem. 1976, 41, 3465. (g) Hall, S. S.; McEnroe, F. J . J . Org.
Chem. 1975, 40, 271. (h) Zilenovsky, J . S. R.; Hall, S. S. J . Org. Chem.
1981, 46, 4139. (i) Hall, S. S.; Lipsky, S. D.; McEnroe, F. J .; Bartels,
A. P. J . Org. Chem. 1971, 36, 2588. (j) Pak, C. S.; Lee, E.; Lee, G. H.
J . Org. Chem. 1993, 58, 1523.
(3) For an excellent review on the role of electrons as activating
“messengers”, see: Chanon, M.; Rajzmann, M.; Chanon, F. Tetrahedron
1990, 46, 6193.
(4) The crystal structure of several clusters have been reported.
See: (a) Marsch, M.; Harms, K.; Lochmann, L.; Boche, G. Angew.
Chem., Int. Ed. Engl. 1990, 29, 308. (b) Harder, S.; Streitwieser, A.
Angew. Chem., Int. Ed. Engl. 1993, 32, 1066. For solution structures,
see ref 17 and: (c) Lochmann, L; Pospisil, J .; Vodnansky, J .; Trekoval,
J .; Lim, D. Collect. Czech. Chem. Commun. 1965, 30, 2187. See also
ref 11.
(5) The stereochemical results observed with Li/THF were identical
to those obtained with lithium in liquid ammonia. See: (a) Ballester,
P.; Capo´, M.; Garc´ıas, X.; Saa´, J . M. J . Org. Chem. 1993, 58, 328. (b)
Ballester, P.; Capo´, M.; Saa´, J . M.Tetrahedron Lett. 1990, 31, 1339.
(c) Garc´ıas, X.; Ballester, P.; Capo´, M.; Saa´, J . M. J . Org. Chem. 1994,
59, 5093.
(6) For the lithium-promoted clevage of in situ generated benzyl
alcohols, see: (a) Flisak, J . R.; Hall, S. S. J . Am. Chem. Soc. 1990,
112, 7299. (b) Hall, S. S.; Lipsky, S. D. J . Org. Chem. 1973, 38, 1735.
(c) Hall, S. S. J . Org. Chem. 1973, 38, 1738. (d) Zilenovski, J . S. R.;
Hall, S. S. J . Org. Chem. 1981, 46, 4139.
(7) The electrochemical reduction of both cinnamyl alcohols and
ethers has been reported to yield 1-phenyl-1-propene and 1-phenyl
propane. See: (a) Mairanovsky, V. G. Angew. Chem., Int. Ed. Engl.
1976, 15, 281 and references therein. (b) Santiago, E.; Simonet, J .
Electrochim. Acta 1975, 20, 853. (c) Lund, H.; Doupeux, H.; Michel,
M. A.; Mousset, G.; Simonet, J . Electrochim. Acta, 1974, 19, 629. (d)
Horner, L.; Ro¨der, H. Ann. Chem. 1969, 723, 11. Nuntnarumit, C.;
Hawley, M. D. J . Electroanal. Chem. 1982, 133, 57.
(8) The radical anion/cation and dianion/dication nomenclature is
used throughout for clarity. By dianion/dication we mean to describe
not only plain ion pairs but also triple ion species (i.e., cation/dianion/
cation) presumably resulting from reactions that employ alkali metals
(with or without a carrier) as reductants.
0022-3263/96/1961-1035$12.00/0 © 1996 American Chemical Society

1036 J . Org. Chem., Vol. 61, No. 3, 1996
Saa´ et al.
Sch em e 1
cleavage (VIBICLE). By so doing, we aimed at reaching
the goal of providing a full picture of metal-promoted
reductive cleavages in which the role of counterion and
solvent effects would be, at least in part, examined.
Central to the first objective was the question of
whether or not a π-type organolithium compound would
form a “unimetal superbase”10 by interacting with the
lithium alkoxide being generated alongside in the reac-
tion.4,11 Were this the case, we conceived the idea that
the above stereochemical issue could be the consequence
of some poorly understood equilibrating process acces-
sible for the mixed species derived from cinnamyl ethers
but not for those resulting from the isomeric unsaturated
benzyl ethers.12 Accordingly, we decided to examine the
solution structure4a,13 and stereochemical properties of
the organolithium species resulting from BICLE and
VIBICLE processes by means of NMR spectroscopy of the
magnetic nuclei (1H, 13C, Li, and/or Li) available.14 In
addition, we planned to carry out a quantum (MNDO)
chemical study which would allow us to deal with
hapticity, a structural property of π organolithium com-
pounds that we deemed of considerable significance for
understanding the stereochemical issues posed by metal-
promoted reductive cleavage (see below). It is worth
mentioning in this regard that benzyllithium actually
crystallizes out as different haptomeric structures de-
pending on the solvent and ligands present during its
generation,15 thereby showing that lithium haptomers
(and lithium tropomers16 ) might be separated by just a
few kcal/mol and, accordingly, that the equilibria involv-
ing these species may be displaced toward one side or
another by modifying such operational variables as
solvent, temperature, or concentration.17
6
7
as silyl ethers) takes place instead in a highly stereo-
controlled manner, i.e. with complete retention of con-
2
5
figuration of the sensitive ∆ double bond,2e-i, thereby
providing a unique entry to prenylated systems (Scheme
1).5-7
This puzzling stereochemical behavior could be ther-
modynamic in nature and operate either after or before
cleavage (Figure 1). The former possibility (i.e., after
cleavage) required that either the radicals (if cleavage
were to occur from a radical anion/cation species8 ) or the
organolithium compounds (if cleavage were to take place
at the dianion/dication stage8) resulting from BICLE and
VIBICLE were structurally different species so that only
one could undergo equilibration. Assuming, in principle,
that a common mechanism would operate for both
reactions, the above condition would certainly not be
attended for simple radicals or organolithium species, but
it could be met if complexation with the coexisting
lithium salts in solution were possible, a chance that we
could envisage for the final organolithium compounds
(see below). The later possibility (i.e., thermodynamic
equilibration prior to cleavage) required BICLE and
VIBICLE to give rise to intermediate species having well-
differentiated lifetimes so as to allow for equilibration
in one case but not in the other.9 On the other hand,
the observed results might well be entirely kinetic in
nature, thereby implying that a bias should be available
for the vinylogous series only. Alternatively, BICLE and
VIBICLE reactions might just involve different reaction
mechanisms. In trying to find appropriate explanations
to the above intriguing problems, we decided to address,
from experimental and theoretical viewpoints, the fol-
lowing two main objectives: (1) the structure and (ster-
eochemical) properties of the final organometallic species
present in solution and (2) the mechanistic details of the
lithium-promoted Birch (BICLE) and vinylogous Birch
Our second objective called for carrying out a theoreti-
cal study of BICLE and VIBICLE processes in order to
assess the likely role of the counterion18 and solvent both
(10) Caube`re, P. Chem Rev. 1993, 93, 2317 and references therein.
For some recent monographs on “complex bases” or “super bases”, see
also: Schlosser, M. Pure Appl. Chem. 1988, 60, 1627. Mordini, A. In
Advances in Carbanion Chemistry; Snieckus, V., Ed.; J AI Press;
Greenwich, CT, 1992. Multimetal superbases are also the matter of
current interest and controversy. See, for example: Bauer, W.;
Lochman, L. J . Am. Chem. Soc. 1992, 114, 7482.
(11) (a) Lochmann, L.; Pospisil, J .; Lim, D. Tetrahedron Lett. 1966,
257. (b) Schlosser, M. J . Organomet. Chem. 1967, 8, 9. (c) McGarrity,
J . F.; Ogle, C. A. J . Am. Chem. Soc. 1985, 107, 1805. (d) McGarrity, J .
F.; Ogle, C. A.; Brich, Z.; Loosli, H. R. J . Am. Chem. Soc. 1985, 107,
1810. (e) Al-Aseer, M. A.; Allison, B. D.; Smith, S. G. J . Org. Chem.
1985, 50, 2715. (f) Baryshnikov, Yu. N.; Kaloshina, N. N.; Vernovskaya,
G. I. J . Gen. Chem. USSR 1977, 47, 2535.
(12) For a recent review on polar allyl-type organometallics, see:
Schlosser, M.; Desponds, O.; Lehmann, R.; Moret, E.; Rauchswalbe,
G. Tetrahedron 1993, 49, 10203.
(13) Setzer, W.; Schleyer, P. v. R. Adv. Organomet. Chem. 1985, 24,
353.
(14) For a recent review on the NMR structural investigation of
organolithium compounds, see: Bauer, W.; Schleyer, P. v. R. In
Advances in Carbanion Chemistry; Snieckus, V., Ed.; J AI Press;
Greenwich, CT, 1992; Vol. 1.
(15) Benzyllithium X-ray structures: (a) Patterman, S. P.; Karle, I.
L.; Shicky, G. D. J . Am. Chem. Soc. 1970, 92, 1150. (b) Beno, M. A.;
Hope, H.; Olmstead, M. M.; Power, P. P. Organometallics 1985, 4, 2117.
(c) Zarges, W.; Marsch, M.; Harms, K.; Boche, G. Chem. Ber. 1989,
122, 2303. For some theoretical studies on the structure of benzyl-
lithium, see: (d) Lipkowitz, K. B.; Uhegbu, C.; Naylor, A. M.; Vance,
R. J . Comput. Chem. 1985, 6, 62. (e) Sygula, A.; Rabideau, P. W. J .
Am. Chem. Soc. 1992, 114, 821.
(9) For a recent incisive analysis into the stepwise versus concerted
nature of the reductive eliminations induced by electron transfer
(mostly electrochemically mediated), see: (a) Andrieux, C. P.; Le
Gorande, A.; Save´ant, J .-M. J . Am. Chem. Soc. 1992, 114, 6892. (b)
Bertra´n, J .; Gallardo, I.; Moreno, M.; Save´ant J .-M. J . Am. Chem. Soc.
1992, 114, 9576. (c) Lexa, D.; Save´ant, J .-M.; Su, K. B.; Wang, D. L. J .
Am. Chem. Soc. 1988, 110, 7617. (d) Andrieux, C. P.; Blochman, C.;
Dumas-Bouchiat, J .-M.; Save´ant, J .-M. J . Am. Chem. Soc. 1979, 101,
3431. (e) Andrieux, C. P.; Gallardo, I.; Save´ant, J .-M.; Su, K. B. J . Am.
Chem. Soc. 1986, 108, 638. (f) Andrieux, C. P.; Differding, E.; Robert,
M.; Save´ant, J .-M. J . Am. Chem. Soc. 1993, 115, 6592. (g) Andrieux,
C. P.; Save´ant, J .-M. J . Am. Chem. Soc. 1993, 115, 8044. (f) Save´ant,
J .-M. Tetrahedron 1994, 50, 10117 and references therein.
(16) For an example in which lithiotropicity has been observed in
the solid state, see: Boche, G.; Etzrodt, H.; Massa, W. Baum, G. Angew.
Chem., Int. Ed. Engl. 1985, 24, 863.
(17) (a) Schleyer, P. v. R. Pure Appl. Chem. 1983, 55, 355. (b)
Schleyer, P. v. R. Pure Appl. Chem. 1984, 56, 151. (c) Streitwieser, A.,
J r. Acc. Chem. Res. 1984, 17, 353.
(18) Counterion-free “carbanions” are more a pedagogic aid than
really abundant chemical species. See: Lambert, C.; Schleyer, P. v. R.
Angew. Chem., Int. Ed. Engl. 1994, 33, 1129.

Metal-Induced Reductive Cleavage Reactions
J . Org. Chem., Vol. 61, No. 3, 1996 1037
F igu r e 1. Alternative routes for BICLE and VIBICLE cleavage. Common intermediates and/or products in boxes.
in cleavage and stereochemistry. To this end, we have
carried out calculations for radical anion/cation and
dianion/dication species in the following three limiting
situations: (a) with discrete counterion solvation as a
model for reactions involving contact ion pairs;19 (b) in a
solvent continuum (nonintervening counterions are con-
sidered apart for thermodynamic purposes) as a very
simplified model for solvent-separated ion pairs;20 and
(c) counterion-free species in vacuo as model for cleavage
reactions promoted by electron transfer (from a metal or
otherwise) which, presumably, involve counterion-free
species.9,21 By so doing, we expected to be able to assess
the relative importance of some of the myriad equilibria
involved in reductive cleavage reactions and determine
the role of the counterion that we envisaged, at the
outset, responsible for cleavage.
Reductive cleavage reactions promoted by electron
transfer encompass two fundamental operations irre-
spective of the number of electrons involved, according
to the unifying mechanistic pattern generally accepted.3
First, the unsaturated group (if any) of the substrate acts
as a (temporary) captor during the primary act of electron
transfer. Presumably, electrons act as “messengers” by
migrating intramolecularly to a σ moiety of the substrate
which is thus activated toward its eventual dissociation
because the σ* state becomes partially populated.22 Both
operations (electron transfer and cleavage) can either be
concerted or stepwise, thus giving rise to the standard
set of alternative mechanistic routes.3,21 Likely, as
pointed out by Bock and co-workers,23 the mechanistic
picture is that of a multidimensional network equilibria
actually interconnecting electron transfer, ion pair sol-
vation, and aggregation processes in which the metal
intervenes decisively. However, until recently, the role
assigned to the metal in metal-promoted processes was
merely that of acting as the electron source but none in
the rate-determining cleavage step. In particular, coun-
terion-free radical anions have been usually invoked as
key intermediates in the alkali metal-promoted reductive
cleavage of ethers1a,3 (diaryl and alkylaryl ethers), alkyl
halides,24 and sulfides,1d,25 mostly on the basis of product
studies.26 Moreover, according to the principle of regio-
conservation of spin density due to Maslak and Guthrie,27
the regioselectivity of the R cleavage undergone by anion
radicals (whether long or short lived) derived from
alkylaryl and diaryl ethers should occur so that the
unpaired electron remains on the same side of the
scissible bond. However, even the regiochemical issue
is far from being definitely settled as new, controversial
results, stereochemical or otherwise, continue to appear.28
Remarkably, no attention has been paid in these studies
to the possible role played by the metal center in the
cleavage of radical anions, in spite of the well-known
effect of the metal on the hyperfine coupling observed in
(24) Garst, J . F. Acc. Chem. Res. 1971, 4, 400. See also ref 34.
(25) (a) Screttas, C. G.; Screttas, M. M. J . Org. Chem. 1978, 43, 1064.
Cohen, T.; Bupathy, M. Acc. Chem. Res. 1989, 22, 152. For electro-
chemical studies see the following: (b) Severin, M. G.; Farnia, G.;
Vianello, E.; Arevalo, M. C. J . Electroanal. Chem. 1988, 251, 369. (c)
Severin, Arevalo, M. C.; M. G.; Farnia, G.; Vianello, E. J . Phys. Chem.
1987, 91, 466. (d) Arevalo, M. C; Farnia, G.; Severin, M. G.; Vianello,
E. J . Electroanal. Chem. 1987, 220, 201. (e) Griggio, L.; Severin, M.
G. J . Electroanl. Chem. 1987, 223, 185. (f) Griggio, L. J . Electroanal.
Chem. 1982, 140, 155.
(26) (a) Fish, R.; Dupon, J . W. J . Org. Chem. 1988, 53, 5230. (b)
Koppang, M. D.; Woolsey, N. F.; Bartak, D. E. J . Am. Chem. Soc. 1984,
106, 2799. (c) Thornton, T. A.; Woolsey, N. F.; Bartak, D. E. J . Am.
Chem. Soc. 1986, 108, 6497. (d) Masnovi, J .; Maticic, J . J . Am. Chem.
Soc. 1988, 110, 5189.
(27) Maslak, P.; Guthrie, R. D. J . Am. Chem. Soc. 1986, 108, 2628.
Maslak, P.; Guthrie, R. D. J . Am. Chem. Soc. 1986, 108, 2637.
(28) (a) Lexa, D.; Savea´nt, J .-M.; Scha¨fer, H. J .; Su, K-B.; Vering,
B.; Wang, D. L. J . Am. Chem. Soc. 1990, 112, 6162 and references
therein. (b) Dorigo, A. E.; Houk, K. N.; Cohen, T. J . Am. Chem. Soc.
1989, 111, 8976.
(29) (a) Hobey, W. D.; McLachlan, A. D. J . Chem. Phys. 1960, 33,
1695. (b) McConnell, H. M.; McLachlan, A. D. J . Chem Phys. 1961,
34, 1. (c) McConnell, H. M. J . Chem. Phys. 1961, 34, 13. (d) Gerson,
F.; Huber, W.; Mu¨llen, K. Helv. Chim. Acta 1981, 64, 2766. (e) Tuttle,
T. R., J r.; Danner, J . C.; Graceffa, P. J . Chem. Phys. 1972, 76, 2866.
(f) Hirota, N. J . Am. Chem. Soc. 1968, 90, 3603. (g) Hirota, N.;
Carraway, R.; Schook, W. J . Am. Chem. Soc. 1968, 90, 3611. (h) Hirota
, N.; Kreilick, R. J . Am. Chem. Soc. 1966, 88, 615.
(19) Szwarc, M.Ions and Ion-Pairs in Organic Reactions; Wiley-
Interscience: New York, 1972; Vol. 1.
(20) The structure of solvent-separated ion pairs is much more
dynamic than that of contact ion pairs, the distance between ions being
no less than 7 Å. See: Weller, A. Z. Phys. Chem. 1982, 130, 129. Our
choice of a model for SSIP’s, though static, mets two of these criteria,
namely: (1) there is no interaction between ions, and (2) they are
strongly solvated by the medium.
(21) Electrochemically-induced reductive cleavage reactions are
ususally written as involving counterion-free species. See, for ex-
ample: (a) Save´ant, J .-M. Tetrahedron 1994, 34, 10117. (b) Save´ant,
J .-M. Adv. Phys. Org. Chem. 1990, 26, 1. (c) Bunnett, J . F. Acc. Chem.
Res. 1978, 11, 413.
(22) Maslak, P.; Narvaez, J . N.; Parvez, M. J . Org. Chem. 1991, 56,
602.
(23) Bock, H.; Na¨ther, C.; Havlas, Z.; J ohn, A.; Arad, C. Angew.
Chem., Int. Ed. Engl. 1994, 33, 875. See also ref 37.

1038 J . Org. Chem., Vol. 61, No. 3, 1996
Saa´ et al.
the ESR spectra of radical anion/cation species.29 Only
a recently published investigation claimed, on the basis
of experimental (ESR, ENDOR, and triple studies) and
low-level theoretical (Hu¨ckel and INDO) calculations,
that the counterion actually determines the lowest energy
conformation of the radical anion/cation pair.30 Further-
more, even though solvation, concentration, and temper-
ature are known to affect the structure of the radical
anion salts (ketyls) resulting from metal reduction of
aromatic ketones,31 counterion and solvation effects in
redox reactions remain “sunk at unknown depths”,32
especially for those processes which involve two electron
transfer steps with concomitant cleavage of a C-X bond
in a position R or â to an existing π system.33
Much more information is available for electrochemi-
cally-induced reductive cleavage reactions which are
known to take place in a thin reaction layer on the
electrode surface.9g Save´ant’s cyclic voltammetric studies
call for two mechanistic scenarios according to the
concerted or stepwise nature of single electron transfer
and cleavage.21 In the inner sphere process, both are
concerted whereas in the outer sphere process electron
transfer and cleavage are consecutive steps involving an
intermediate radical anion. The electrochemical reduc-
tion of alkyl halides,9b benzyl halides,9a aryl halides9d and
N-halosultams9f appears to conform to this dual mechanim,
claims to solvent9a and counterion effects being only made
to account for the strikingly different kinetic behavior of
the otherwise identical radical anion species resulting
from electrochemical or radiolysis processes.9f
Irrespective of whether cleavage involves a radical
anion/cation or dianion/dication species, we hypothesized
that lithium ought to play a key role in promoting
cleavage in BICLE and VIBICLE processes by acting as
an internal Lewis acid in assisting the departure of the
leaving group and thus perhaps provide a bias for
stereochemical discrimination. In addition, we contended
that the metal atom should also be responsible for the
“messenger” role in activating positions far away from
the electron sink because of the likely shallow nature of
the potential energy surface of π organolithium species.39
Herein, we show that BICLE and VIBICLE are both
stepwise processes, each giving rise to π-type organo-
lithium species which do not undergo appreciable equili-
bration at T < 0 °C. However, as shown below by
calculations, a fundamental difference exists between
them: BICLE processes involve transient radical anion/
cation intermediates whereas VIBICLE reactions give
rise to radical anion/cation intermediates that can ac-
cumulate as they are more stable (either as contact ion
pairs, CIPs, or solvent separated ion pairs, SSIPs19) than
starting materials and, hence, undergo conformational
(but not configurational, see below) equilibration prior
to cleavage. Actually, according to MNDO, the stereo-
chemical outcome of VIBICLE reactions appears to be
due to the combination of two factors: (a) cleavage in the
vinylogous series involves the dianion/dication species,
and (b) the styryl unit can (if appropriately substituted)
function as a chiral plane, thereby giving rise to two
competitive, diastereoisomeric routes for cleavage.
The scenario for our studies was still somewhat
broader as, in principle, the rate-determining cleavage
step of BICLE and VIBICLE processes (Figure 1) might
involve not only the so-called radical anion but also the
dianion species (actually radical anion/cation and/or
dianion/dication pairs).34 Walsh has provided kinetic
evidence for the simultaneous occurrence of cleavage (of
a carbon-carbon bond) mechanisms from the radical
anion and dianion intermediates,35 the later species
resulting from disproportionation36 or reduction of the
corresponding radical anion.34 Other examples from the
old and recent literature37 claim that cleavage occurs
from dianions, though to the best of our knowledge, only
Kiesele38 has been able to demonstrate that the counte-
rion plays a key role, however unespecified, in the
rearrangement reaction following cleavage. Curiously
enough, despite the fact that dianions been have also
invoked by numerous workers in the past, this mecha-
nism is not unfrequently quoted as “unnecessary”.1a,34
Com p u ta tion a l Meth od s
For this study we have used the MNDO semiempirical
treatment for which lithium has been parametrized.40
MNDO, in spite of its well-known shortcomings,41 has
proven to be reliable for studying the very large struc-
tures of organolithium compounds as contact ion pairs,
(37) (a) Birch, A. J . J . Chem. Soc. 1944, 430. (b) Birch, A. J . J . Chem.
Soc. 1947, 102. (c) Eisch, J . J . J . Org. Chem. 1963, 28, 707. (d)
Normant, H.; Cuvigny, T. Bull. Soc. Chim. Fr. 1966, 3344. (e)
Testaferri, L.; Tiecco, M.; Tingoli, M.; Chianelli, D.; Montanucci, M.
Tetrahedron 1982, 38, 3687. (f) Itoh, M.; Yoshida, S.; Ando, T.; Miyaura,
N. Chem Lett. 1976, 271. (g) Eargle, D. H., J r. J . Org. Chem. 1963, 28,
1703. (h) Grovenstein, E. J r.; Bhatti, A. M.; Quest, D. E.; Sengupta,
D.; VanDerveer, D. J . Am. Chem. Soc. 1983, 105, 6290. (i) Gerdil R.
Helv. Chim. Acta 1973, 56, 196. (j) Cox, J . A.; Ozment, C. L.
Electroanal. Chem. Interfac. Electrochem. 1974, 51, 75. (k) Oku, A.;
Yoshida, M.; Matsumoto, K. Bull. Chem. Soc. J pn. 1979, 52, 524. (l)
Kawase, T.; Fujino, S.; Oda, M. Tetrahedron Lett. 1990, 31, 545. (m)
Kawase, T.; Fujino, S.; Oda, M. Chem. Lett. 1990, 1683. Kawase, T.;
Kurata, H.; Fujino, S.; Ekinaka, T.; Oda, M. Tetrahedron Lett. 1992,
33, 7201. (n) Farnia, G.; Marcuzzi, F.; Melloni, G.; Sandona`, G. J . Am.
Chem. Soc. 1984, 106, 6503. (o) Farnia, G.; Marcuzzi, F.; Melloni, G.;
Sandona`, G. Zucca, M. V. J . Am. Chem. Soc. 1989, 111, 918.
(38) Kiesele, H. Angew. Chem., Int. Ed. Engl. 1987, 109, 1511. (b)
Kiesele, H. Tetrahedron Lett. 1981, 22, 1097.
(30) Lazana, M. C. R. L. R.; Franco, M. L. T. M. B.; Herold, B. J . J .
Am. Chem. Soc. 1989, 111, 8640.
(31) (a) Goldberg, I. B.; Bolton, J . R. J . Chem. Phys. 1970, 74, 1965.
(b) Mao, S. W.; Nakamura, K.; Hirota, N. J . Am. Chem. Soc. 1974, 96,
5341. (c) Chen, K. S.; Mao, S. W.; Nakamura, K.; Hirota, N. J . Am.
Chem. Soc. 1971, 93, 6004.
(32) Bock, H.; Ruppert, K.; Na¨ther, C.; Havlas, Z.; Herrmann, H.-
F.; Arad, C.; Go¨bel, I.; J ohn, A.; Meuret, J .; Nick, S.; Rauschenbach,
A.; Seitz, W.; Vaupel, T.; Solouki, B. Angew. Chem., Int. Ed. Engl. 1992,
31, 550.
(39) According to calculations, delocalized “carbanions” have a very
flat potential hypersurface. See: (a) Schleyer, P. v. R.; Kos, A. J .;
Wilhelm, D.; Clark, T.; Boche, G.; Decher, H.; Etzrodt, H.; Dietrich,
H.; Mahdi, W. J . Chem. Soc., Chem. Commun. 1984, 22, 1495. (b)
Bushby, R. J .; Tytko, M. P. J . Organomet. Chem. 1984, 270, 265. (c)
Morton-Blake, D. A.; Corish, J .; Be´niere, F. Theor. Chim. Acta 1985,
68, 389. See also ref 23.
(33) The formation of multiply charged ions resulting from the two
successive one-electron reductions of conjugated cyclic hydrocarbons
can be monitored by appropriately combining ESR and NMR measure-
ments. See, for example: (a) Mu¨llen, K. Angew. Chem., Int. Ed. Engl.
1987, 26, 204. (b) Mu¨llen, K. Chem Rev. 1984, 84, 603. (c) K. Mu¨llen,
Huber, W.; Neumann, G.; Schmieders, C.; Unterberg, H. J . Am. Chem.
Soc. 1985, 107, 801. Mu¨llen, K. Pure Appl. Chem. 1986, 58, 177.
(34) Holy, N. L. Chem Rev. 1974, 74, 243.
(35) Walsh, T. D. J . Am. Chem. Soc. 1987, 109, 1511.
(36) Disproportionation has been shown to be dependent on the
counterion. See: J ensen, B. S.; Parker, V. D. J . Am. Chem. Soc. 1974,
97, 5211. See also ref 19.
(40) Stewart, J . J . P. MOPAC 93, Fujitsu Limited, Tokyo, J apan,
1993. Available from Quantum Chemistry Program Exchange, Uni-
versity of Indiana, Bloomington, IN.
(41) A detailed account of pros and cons is provided by: Clark, T. A
Handbook of Computational Chemistry; Wiley Interscience: New York,
1985; Chapter 4. MNDO overestimates C-Li bond covalency and its
energy. See: Schleyer, P. v. R. Pure Appl. Chem. 1983, 55, 355. Li-H
interactions are overestimated by MNDO. See: Kauffmann, E.; Ragha-
vachari, K.; Reed, A.; Schleyer, P. v. R. Organometallics 1988, 7, 1597.
Streitweiser has stressed the inadequacy of MNDO to deal with
dilithioacetaldoxime derivatives. See: Glaser, R.; Streitweiser, A. J .
Org. Chem. 1989, 54, 5491.

Metal-Induced Reductive Cleavage Reactions
J . Org. Chem., Vol. 61, No. 3, 1996 1039
their predictions being, in many cases, in close agreement
with experimental data resulting from X ray and/or
spectroscopic studies.42
For CIPs, calculations have been performed using
discrete solvation43 of lithium with two NH3 molecules
(per lithium) as working model.44 Calculations “in vacuo”
or in a “dielectric continuum” (COSMO)45 have been used
to modelize isolated, nonsolvated ion pairs (IIPs)46 and
the corresponding solvent separated ion pairs (SSIPs),
respectively.19 No geometrical constraints were imposed
except for the case of cinnamyl derivatives for which the
π system was kept planar throughout. Due to MNDO’s
overestimation of steric repulsions, the corresponding
nonplanar structures were found to be somewhat lower
in energy. All structures were optimized and further
refined by minimization of the gradient norm at least
below 0.1 kcal/Å‚deg. All stationary points on the
potential energy surfaces were characterized by calculat-
ing and diagonalizing the Hessian matrix and by check-
ing the number of negative eigenvalues.47 All ∆G values
refer to 273 K.
F igu r e 2.
6Li in dried THF-d8,5 at 0 °C for 1 h, could be gleaned
from its NMR spectra.5 In both BIHY and VIBIHY cases,
6
the Li spectra showed two different signals at 2.18 and
3.03 ppm, presumably corresponding to a single organo-
lithium compound concurrent with lithium trimethylsilyl
oxide. The H and 13C NMR spectra which were found
1
to be similar to those already published for other phe-
nylallyllithium compounds also confirmed this initial
assignment.48 Relevant to our first objective, we found
that both the organolithium compound and the lithium
silyloxide species in solution did not interact with each
other as evidenced by the 6Li-1H HOESY49 spectrum
(supporting information) for which the phase cycle re-
ported by Bauer50 gave best results. Thus, the lithium
at 3.03 ppm correlated with the methyl groups of the Me3-
Si group only, whereas that at 2.18 ppm gave a large
cross peak with the benzylic hydrogen (3.66 ppm), two
small ones with the vinylic hydrogen (6.06 ppm), and an
aromatic ortho hydrogen (5.86 ppm) and a very small one
with the methyl group (1.58 ppm) syn to the benzylic
hydrogen. Altogether, these features clearly define the
following structural picture: (1) the solution structure
of the organolithium compound 3ia (Figure 2) generated
by both BICLE and VIBICLE processes is that of a vinyl
substituted η3 benzyl lithium,51 thereby suggesting a
reasonable explanation for the regioselectivity of the
protonation step;52 and (2) the organolithium compound
and the lithium (trimethylsilyl) oxide (or derivative
thereof) present in solution do not interact with each
other as shown by the absence of cross peaks in the
6Li-1H HOESY spectra and lithium-lithium interactions
in the bidimensional 6Li-6Li INADEQUATE53 spectra.54
This structural analysis is, at least, valid for the -50 to
25 °C range for which no significant modifications on the
above spectra were observed, except for some sharpening
of the lithium silyloxide signal at 3.03 ppm. At temper-
atures above 40 °C, quenching of the organolithium
species took place. The possible existence of mixed
species in very low equilibrium concentration cannot,
however, be discarded from the available data (see below
for quantum chemical results).
Discu ssion
Exp er im en ta l Stu d ies. Experiments were set so as
to learn if the final π organolithium compounds would
form mixed aggregates with the coexisting lithium salts.
The structure of the organolithium species 3ia (Figure
2) generated by treatment of silyl ethers 1a and 2a with
(42) (a) Wilhelm, D.; Clark, T.; Schleyer, P. v. R.; Dietrich, H.; Mahdi,
W J . Organomet. Chem. 1985, 280, C6. (b) Streitwieser, A., J r.;
Swanson, J . T. J . Am. Chem. Soc. 1983, 105, 2502. (c) Boche, G.;
Decher, H.; Etzrodt, H.; Dietrich, H.; Mahdi, W.; Kos, A. J .; Schleyer,
P. v. R. J . Chem. Soc., Chem. Commun. 1984, 22, 1493. (d) Boche, G.;
Fraenkel, G.; Cabral, J .; Harms, K.; Hommes, N. J . R. v. E. Lohrenz,
J .; Marsch, M. Schleyer, P. v. R. J . Am. Chem. Soc. 1992, 114, 1562.
Dietrich, H.; Mahdi, W.; Wilhelm, D.; Clark, T.; Schleyer, P. v. R.
Angew. Chem., Int. Ed. Engl. 1984, 23, 621. Schleyer, P. v. R.; Hacker,
R.; Dietrich, H.; Mahdi, W. Chem. Commun. 1985, 622. Neugebauer,
W.; Geiger, G. A..P.; Kos, A. J .; Stezowski, J . J .; Schleyer, P. v. R. Chem.
Ber. 1985, 118, 1504. Stezowski, J . J .; Hoier, H.; Wilhelm, D.; Clark,
T.; Schleyer, P.v.R. Chem. Comm. 1985, 1263. Hacker, R.; Schleyer,
P. v. R.; Reber, G.; Mu¨ller, G.; Brandsma, L. J . Organomet. Chem. 1986,
316, C4. McKee, M. L. J . Am. Chem. Soc. 1987, 109, 559. Hacker, R.;
Kauffmann, E.; Schleyer, P. v. R.; Mahdi, W.; Dietrich, H. Chem. Ber.
1987, 120, 1533. Bausch, J . W.; Gregory, P. S.; Olah, G. A.; Prakasch,
G. K.; Schleyer, P. v. R.; Segal, G. A. J . Am. Chem. Soc. 1989, 111,
3633. Paquette, L. A.; Bauer, W.; Sivik, M. R.; Bu¨hl, M.; Feigel, M.;
Schleyer, P. v. R. J . Am. Chem. Soc. 1990, 112, 8776. Bauer, W.;
O’Doherty, G. A.; Schleyer, P. v. R.; Paquette, L. A. J . Am. Chem. Soc.
1991, 113, 7093. Romesberg, F. E.; Colum, D. B. J . Am. Chem. Soc.
1992, 114, 2112.
(43) (a) Sun˜er, G. A.; Deya´, P. M.; Saa´, J . M. J . Am. Chem. Soc.
1990, 112, 1467. (b) Bauer, W.; Feigel, M.; Mu¨ller, G.; Schleyer, P. v.
R. J . Am. Chem. Soc. 1988, 110, 6033. (c) Bauer, W.; Clark, T.;
Schleyer, P. v. R. J . Am. Chem. Soc. 1987, 109, 970. (d) Saa´, J . M.;
Deya´, P. M.; Sun˜er, G. A.; Frontera, A. J . Am. Chem. Soc. 1992, 114,
9093. (e) Saa´, J . M.; Morey, J .; Frontera, A.; Deya´, P. M. J . Am. Chem.
Soc. 1995, 117, 1105.
(44) Bauer, W.; Lochmann, L. J . Am. Chem. Soc. 1992, 114, 7482.
Best fit with experiment was achieved when real (dimethyl ether)
ligands were employed in MNDO calculations.
(45) COSMO (conductor-like screening model) evaluates the solvent
screening energy for a cavity based on the solvent-accesible surfaces
and for a charge distribution derived from a distributed multipole
analysis: Klamt, A.; Schu¨u¨rmann, G. J . Chem. Soc., Perkin Trans. 2
1993, 799. For a number of other “dielectric continuum” solvation
models, see: (a) Katritzky, A. R.; Karelson, M. J . J . Am. Chem. Soc.
1991, 113, 1561. (b) Cramer, C. J .; Truhlar, D. G. Science 1992, 256,
231. Cramer, C. J .; Truhlar, D. G. J . Am. Chem. Soc. 1991, 113, 8552
and 8305. (c) Miertus, S.; Scrocco, E.; Tomasi, J . Chem. Phys. 1981,
55, 117. (d) Ford, G. P.; Wang, B. J . Am. Chem. Soc. 1992, 114, 10563.
(46) For prior theoretical studies on cation-free radical anions, see:
(a) Casado, J .; Gallardo, I.; Moreno, M. J . Electroanal. Chem. 1987,
219, 197. (b) Andrieux, C. P.; Save´ant, J .-M.; Zann, D. Nouv. J . Chim.
1984, 8, 107. (c) Beland, F. A.; Farwell, S. O.; Callis, P. R.; Geer R. D.
J . Electroanal. Chem. 1977, 78, 145.
(48) (a) Freedman, H. H.; Sandel, V. R.; Thill, B. P. J . Am. Chem.
Soc. 1967, 89, 1762. (b) Sandel, V. R.; Freedman, H. H.; McKinley, S.
V. J . Am. Chem. Soc. 1968, 90, 495.
(49) (a) Bauer, W.; Mu¨ller, G.; Pi, R.; Schleyer, P. v. R. Angew.
Chem., Int. Ed. Engl. 1986, 25, 1103. (b) Bauer, W.; Feigel, M.; Mu¨ller,
G.; Schleyer, P. v. R. J . Am. Chem. Soc. 1988, 110, 6033. (c) Kaufmann,
E.; Raghavachari, K.; Reed, A. E.; Schleyer, P. v. R. Organometallics
1988, 7, 1597. (d) Bauer, W.; Clark, T.; Schleyer, P. v. R. J . Am. Chem.
Soc. 1987, 109, 970. (e) Gregory, K.; Bremer, M.; Bauer, W.; Schleyer,
P. v. R.; Lorenzen, N. P.; Kopf, J .; Weiss, E. (f) Bauer, W.; Schleyer, P.
v. R. J . Am. Chem. Soc. 1989, 111, 7191.
(50) Bauer, W.; Schleyer, P. v. R. Magn. Reson. Chem. 1988, 26,
827.
(47) McIver, J . W.; Komornicki, A. J . Am. Chem. Soc. 1972, 94, 2625.

1040 J . Org. Chem., Vol. 61, No. 3, 1996
Saa´ et al.
The results of an analogous NMR study on the BICLE
and VIBICLE reactions undergone by (E)-1b, (Z)-1b and
(E)-2b clearly demonstrated that isomerization of the
final organolithium species (E)-3ib and (Z)-3ib (Figure
2) is extremely slow at the temperature of operation (-20
°C), thus proving that the puzzling stereochemistry
observed must be settled in previous steps. Actually,
ical questions posed by BICLE and VIBICLE processes.
Accordingly, we headed for a detailed theoretical study46
of the above ether cleavage reactions as applied to
R-vinylbenzyl methyl ether 4 and the corresponding
cinnamyl methyl ether 11, our models for BICLE (Scheme
2) and VIBICLE processes (Scheme 3), respectively. As
mentioned above, the most relevant parts of this broad
plan included the study of the following: (1) reactions
involving contact ion pairs (CIP’s are labeled throughout
with the appropriate numerals and a subscript (i or ii)
indicating the number of lithium atoms) which we
modeled as organolithium species having discrete solva-
tion (two ammonia ligands per lithium) around the
counterion; (2) reactions involving solvent-separated ion
pairs (SSIP’s; labeled with a numeral with the appropri-
ate (i) or (ii) subscripts) which we modeled by using
noninteracting ions solvated by the “dielectric continuum”
provided by COSMO;47 and (3) reactions involving iso-
lated, nonsolvated ion pairs (IIP’s; labeled throughout the
work by numerals, only).
analysis of the H and 13C NMR spectra as well as Li-
1H HOESY of the red solutions of organolithium com-
pounds (E)-3ib and (Z)-3ib resulting from (E)-1b (95%
stereochemically pure) and (Z)-1b (92% stereochemically
pure), respectively, attested to the fact that BICLE
processes had occurred with retention of configuration
1
6
2
of the sensitive ∆ double bond, the observed E:Z ratio
for (E)-3ib (95:5) and (Z)-3ib (8:92) not being modified
by standing at -20 °C for 24 h. Isomerization is fast (a
ca. 2.5:1 E:Z mixture is obtained after 30 min) at a higher
temperature (27 °C).48b The VIBICLE process carried out
with pure (E)-2b yielded,5 instead, a ca. 2.5:1 mixture of
E and Z organolithium species as determined by integra-
tion of benzylic (δ 3.63 and 3.76 ppm) signals, respec-
tively. This ratio did not suffer alteration after being
allowed to stand for 24 h at 25 °C.
MNDO Stu d ies. As neither the final organolithium
nor the corresponding protonated material isomerized at
the temperature of operation,55 we were left with the
“dark” part of the mechanism to explain the stereochem-
As processes involving electron transfer from metals
to organic substrates are not amenable to semiempirical
calculations because neither kinetic nor thermodynamic
reliable data can be easily obtained,56 we focused, instead,
our attention on lithium naphthalenide-promoted reac-
tions (either as LiNaph•, Li+(solv)||Naph-•(solv) or Li+ Naph-•)
to circumvent the problem. Closely related lithium
benzenide has also been employed in our calculations as
a model for the well-known mediator lithium di-tert-butyl
biphenylide (LiDBB).57 The only kinetic stages that
escaped from this analysis were, obviously, those of
electron transfer since the TS for these processes cannot
be found.56 For the sake of discussion a clear-cut distinc-
tion is made between the reduction and cleavage steps.
(51) The MNDO-determined structure for 3ia accounts reasonably
well for the spectroscopic data obtained as only those hydrogens located
at <3.5 Å should give cross peaks with lithium in 6Li-1H HOESY
spectra. A rapid equilibrium between the haptomeric structures shown
below (geometric parameters relevant to 7Li-1H HOESY experiments
in boxes). cannot be discarded with the available NMR data. MNDO
calculations reveal that these haptomeric structures are separated by
only 1 kcal/mol.
Th e Red u ction Step : Activa tion for Clea va ge.
The role of the media and that played by the metal during
reduction of different substrates can be scrutinized from
the following thermodynamic data. One-electron reduc-
tion of 4 (BIHY) or 11 (VIBIHY) gives rise to radical
anions 5 or 12, respectively, but to radical anion/cation
pairs either as CIP’s (5i and 12i) or SSIP’s (5(i) and
12(i)). Particularly meaningful to our stereochemical
questions is the finding that reduction of 4 with lithium
naphthalenide is a clear-cut thermodynamically uphill
process (4 f 5i, ∆G ) +2.9 kcal/mol; 4 f 5(i), ∆G ) +7.6
kcal/mol; 4 f 5 ∆G ) +9.2 kcal/mol), whereas that of
the vinylogous system 11 is slightly downhill for CIP’s
only (11 f 12i, ∆G ) -2.8 kcal/mol; 11 f 12(i), ∆G )
+0.4 kcal/mol; 11 f 12, ∆G ) +0.3 kcal/mol). As
expected, subsequent reduction to the corresponding
dianion/dication species was found to be endothermic in
all cases for both BICLE (4 f 9ii, ∆G ) 20.4 kcal/mol; 4
f 9(ii), ∆G ) 36 kcal/mol; 4 f 9, ∆G ) 141.3 kcal/mol))
and VIBICLE (11 f 16ii, ∆G ) 3.2 kcal/mol; 11 f 16(ii),
∆G ) 10.6 kcal/mol; 11 f 16, ∆G ) 108.9 kcal/mol)
processes.58 Calculations using LiBenz in place of LiNaph
(supporting information) head in the same direction as
(52) Tanaka, J .; Nojima, M.; Kusabayashi, S. J . Chem. Soc., Perkin
Trans. 2 1987, 673.
(53) Eppers, O.; Fox, T.; Gu¨nther, H. Helv. Chim. Acta 1992, 75,
883.
(54) Curiously enough, benzyl methyl ether itself when submitted
to the typical reaction conditions (6Li, THF-d8, 0 °C, sonication) showed
only one lithium atom in the 6Li NMR spectra, namely that corre-
sponding to benzyllithium. Lithium methoxide was apparently absent
(no hydrogen or lithium signals in the 1H and 6Li NMR spectra), thus
proving that lithium methoxide prefers to aggregate rather than form
a mixed complex with a π-type organolithium species. Actually, lithium
methoxide is seen as a suspension in the NMR tube.
(55) The final isoprenylated aromatics did not suffer isomerization
under the usual5 reaction conditions (lithium in liquid ammonia),55a
as demonstrated by quenching the deep blue solution resulting from
lithium/liquid ammonia cleavage of (E)-1b or (E)-2b with Fe(III) and
stirring the colorless solution (containing lithium amide) for several
hours at room temperature. The standard workup furnished the same
results as for the regular case.5 In other words, lithium amide did not
induce significant isomerization of the prenylated aromatics at the
temperature of operation.55b (a) Rabideau, P. W. Tetrahedron 1989,
45, 1579. Rabideau, P. W.; Marcinow, Z. In Organic Reactions; J ohn
Wiley & Sons: New York, 1992; Vol. 42. (b) Rabideau, P. W.; Huser,
D. L. J . Org. Chem. 1983, 48, 4266.
(56) AM1 has been reported to behave well in a simple test case,
namely the self-exchange electron transfer reactions between a neutral
aromatic and the corresponding cation radical. See: Rauhut, G.; Clark,
T. J . Am. Chem. Soc. 1993, 115, 9127.
(57) Freeman, P. K.; Hutchinson, L. L. J . Org. Chem. 1980, 45, 1924.
Choi, H.; Pinkerton, A. A.; Fry, J . L. J . Chem. Soc., Chem. Commun.
1987, 225. Yus, M.; Ramo´n D. J . J . Chem. Soc., Chem. Commun. 1991,
398.
(58) Ab initio calculations revealed that the formal reaction C6H6
+ 2Li f C6H6Li2 is endothermic by 13.4 kcal/mol. See: Sygula, A.;
Rabideau, P. W. J . Am. Chem. Soc. 1991, 113, 7797.

Metal-Induced Reductive Cleavage Reactions
J . Org. Chem., Vol. 61, No. 3, 1996 1041
Sch em e 2. Alter n a tives Rou tes for Bir ch Red u ctive Clea va ge of Ben zyl Eth er 4 P r om oted by Lith iu m
Na p h th a len id e
Sch em e 3. Alter n a tives Rou tes for Vin ylogou s Bir ch Red u ctive Clea va ge of Cin n a m yl Eth er 11 P r om oted
by Lith iu m Na p h th a len id e
illustrated by the finding that ∆G ) +3.5 kcal/mol for
the BICLE process (4 + LiBenz f 5i) but -1.5 kcal/mol
for the VIBICLE (11 + LiBenz f 12i) tranformation. In
principle, a plausible explanation for the stereochemical
dilemma posed by BICLE and VIBICLE could be built
on these grounds as only VIBICLE processes can lead to

1042 J . Org. Chem., Vol. 61, No. 3, 1996
Saa´ et al.
anion/cation 5i is partially restricted as illustrated by the
fact that this bond is somewhat shortened in going from
4 (1.538 Å) to 5i (1.527 Å). Interestingly, this is irrespec-
tive of the presence or absence of counterion (according
to our modelization for SSIPs, their geometrical features
are almost indistinguishable to those of the corresponding
IIPs). Actually, the counterion-free radical anion 5 shows
an even shorter Cipso-C bond (1.507 Å), and activation
of the C-O bond (1.426 Å) is stronger than in radical
anion/cation pair 5i (see below), no doubt due to much
extensive delocalization in the former species. Curiously
enough, the vinylogous 12i system also becomes planar59
and shows some restricted rotation both around the
corresponding Cipso-C bond (1.453 Å) and the Cvinyl-CH2-
OMe, thereby suggesting that overlapping is crucial for
understanding activation and cleavage in positions far
away from the electron sink. In other words, by intro-
ducing an extra electron in an aromatic system such as
4 or the vinylogous species 11, the system replies by
activating the C-O bond (benzylic or vinylogous) for
dissociation; i.e., the σ* orbital is partially populated as
demonstrated by its increased bond length and reduced
bond order, regardless of whether the counterion is taken
into consideration [5i (1.419 Å), [0.930]; 12i (1.408 Å),
[0.946]] or not [5 (1.426 Å), [0.902]; 12 (1.416 Å), [0.918]].
These features (Figure 3) lend credit to the general
assertion that stereoelectronic factors are extremely
important in determining the rate of these type of
cleavages60 or, in terms of the mixed valence approach,61
that a low energy barrier exists for reaching the σ* state
from the π* state.
Clea va ge: Ha p tom er ic Activa tion . Subtle differ-
ences exist, however, in considering activation of the C-O
bond toward dissociation for the different ion pair species,
as evidenced by its bond length and bond order. Activa-
tion in SSIP’s and IIP’s is obviously related to the size of
the conjugated π system and the stereoelectronic require-
ments for overlap, as shown by the fact that C-O
activation in 5 is significantly more advanced [1.426 Å,
[0.902] than in the vinylogous system 12 [(1.416 Å),
[0.918]]. On the other hand, since reductive cleavage is
in actuality a â elimination reaction, activation for
dissociation in CIP species must be a function of the
location of the counterion in regard with the leaving
group heteroatom, itself a function of hapticity.62 Herein,
we use the term hapticity to indirectly define the site of
the metal on the π system and propose the term hapto-
meric activation to define the extent to which the C-O
bond is prepared for cleavage in the different haptomeric
species along the reaction coordinate. As for benzyl-
F igu r e 3. MNDO-determined structures of intermediates and
transition states for the cleavage of benzyl and cinnamyl ethers
4 and 11 through anion-radical routes involving CIP’s and IIP’s
(SSIP’s species not shown). Geometrical features relevant to
haptomeric activation are shown: (1) C-OMe bond length and
bond order (in brackets); 2) Li-OMe distance; and (3) enthalpy
of formation.
an intermediate radical anion/cation (as a CIP species)
that can accumulate and, therefore, undergo isomeriza-
tion. However, the considerable double bond character
(1.359 Å, [1.718]) retained by C2-C3 in 12i (or haptomers
thereof; see below) clearly militates against E/Z isomer-
ization at this point of the reaction coordinate, as it would
be too energy costly. Thus, we decided to study cleavage
from the so-called radical anion and dianion species
(either as IIP’s, CIP’s, and SSIP’s) derived from Birch
and vinylogous Birch reactions.
A number of mechanistically meaningful features
result after one-electron reduction of 4 or 11 (Figure 3),
the most relevant being that the C-O bond becomes fixed
in a partially locked conformation. For the most part,
this is a consequence of the fact that the key C-O bond
tries to become parallel (syn or anti) to the π system
(dihedral angle ) 59.8°, for the syn species) to facilitate
overlap and also marginally due to the interaction
between lithium and the oxygen atom. Accordingly,
rotation around the Cipso-CHROMe bond in radical
(59) This is remarkable in light of the well-known tendency of
MNDO to overestimate repulsive steric repulsions.
(60) (a) Gerson, F.; Huber, W.; Mullen, K. Angew. Chem., Int. Ed.
Engl. 1978, 17, 208. (b) Gerson, F.; Moshuk, G.; Schwyzer, M. Helv.
Chim. Acta 1971, 54, 361. (c) Bauld, N. L.; Hudson, C. E. Tetrahedron
Lett. 1974, 3174. (d) Dodd, J . R.; Pagni, R. M.; Watson, C. R. J . Org.
Chem. 1981, 46, 1688. (e) Dodd, J . R.; Winton R. F.; Pagni, R. M.;
Watson, C. R.; Bloor, J . J . Am. Chem. Soc. 1974, 96, 7846. (f) Rieke,
R.; Ogliaruso, M.; McClung, R.; Winstein, S. J . Am. Chem. Soc. 1966,
88, 4729. (g) Nelson, S. F.; Gillespie, J . P. J . Org. Chem. 1973, 38,
3592. (h) Hartman, R. F.; Van Camp, J . R.; Rose, S. D. J . Org. Chem.
1987, 52, 2684. (i) Bauld, N. L.; Chang, C. S.; Farr, F. R. J . Am. Chem.
Soc. 1972, 94, 7164.
(61) (a) Ratner, M. A. Int. J . Quant. Chem. 1978, 14, 675. (b) Hale,
P. D.; Ratner, M. A. Int. J . Quant. Chem. 1984, 18, 195. (c) Wong, K.
Y.; Schatz, P. N. Prog. Inorg. Chem. 1981, 28, 369.
(62) At least in one case calculations have been used to determine
the position of the counterion. See: Bolton, J . R.; Goldberg, I. B. J .
Chem. Phys. 1970, 74, 1965. For controversy with the electrostatic
model, see: Brooles, J . J .; Rhine, W.; Stuky, G. J . Am. Chem. Soc. 1972,
94, 7346.

Metal-Induced Reductive Cleavage Reactions
J . Org. Chem., Vol. 61, No. 3, 1996 1043
lithium itself,15e MNDO is capable of detecting a number
of local minima in the potential hypersurface of the
radical anion/cation species resulting from 4 or 11,
namely η6 structures 5i and 12i and η3 structures 6i and
13i (found to lay 9.3 and 4.2 kcal/mol higher in energy
(∆H) than precursors 5i and 12i, respectively). Of course,
only one counterion-free species (5 and 12 or 5(i) and
12(i)) was found. The increasing activation of the C-O
bond in moving from 5i (1.419 Å, [0.930]) to 6i (1.440 Å,
[0.880]) or from 12i (1.408 Å, [0.946]) to 13i (1.412 Å,
[0.945]) can be assigned to the approaching of lithium to
the OMe leaving group which induces an increase in
strength of ionic and covalent forces. Eventually, the
haptomer having a lithium atom â to the leaving group
undergoes syn â elimination, as expected.63 Transition
state structures TS6i (1.963, [0.279]) and TS13i (1.995,
[0.252]) clearly illustrate this assertion (Figure 3). On
the other hand, â elimination in the case of counterion-
free radical anions (either solvent separated (SSIP’s) or
the so-called isolated (IIP’s)) should be of the E1cb type.
For these cases C-O bond cleavage is not assisted by the
counterion, and in all cases, ∆G* is larger than for the
analogous CIP.
In accordance with the expected higher population of
the σ* orbital at the dianion stage, calculations clearly
show that C-O activation for dissociation is significantly
more advanced for the dianion species (Figure 4), as
previously suggested by Walsh.35 As for the above radical
anion/cations, activation is induced during the reduction
stage and as a consequence of haptomeric equilibria (for
CIP’s, only). Evidence for the former is found when the
same indicators (enlarged C-O bond; reduced bond
order) as above are examined for the CIP dilithio deriva-
tives 9ii [(1.420 Å), [0.923]] and 16ii [(1.415 Å), [0.922]]
resulting from BIHY and VIBIHY processes. Haptomeric
activation for dissociation is evident when C-O bond
parameters are examined for 10ii [(1.423 Å), [0.915]],
17ii [(1.415 Å), [0.917]], and 18ii [(1.448 Å), [0.846]], the
former in equilibrium with 9ii and the last two with 16ii
(Figure 4). On the other hand, activation of the C-O
bond for cleavage in IIP’s 9 [(1.480 Å), [0.729]] and 16
[(1.460 Å), [0.771]] or SSIP’s 9(ii) [(1.428 Å, [0.882]] and
16(ii) [(1.420 Å), [0.892]] is much more advanced as, in
these cases, counterions do not serve the purpose of
sharing electron density as in the case of CIP’s.
Th e Ra d ica l An ion /Ca tion vs. Dia n ion /Dica tion
Rou tes for Clea va ge. According to our MNDO calcula-
tions, the question of whether reductive cleavage of
benzyl ethers and the vinylogous cinnamyl systems
involve radical anion/cation or dianion/dication species
is also related to the nature of ion pairing as ∆G* values
(measured, in all cases, from the lowest energy precusors
on the reaction coordinate) vary considerably with it.
BIHY reactions, in particular, prefer to take place
through transient radical anion/cation pairs (either CIP’s
or SSIP’s or IIP’s), the routes involving the corresponding
dianion/dication pairs being more energy costly. Thus,
for example, Birch reductive cleavage of 4 through
transition structure TS6i was found to lie at ∆G* ) 36.8
kcal/mol (43.0 kcal/mol in the case of TS6(i) and 45.6
kcal/mol for TS6, for the cleavage of SSIP and IIP,
respectively), whereas transition structure TS10ii in-
volving cleavage of the CIP dianion/dication species was
found to lie at ∆G* ) 38.1 kcal/mol (55.9 kcal/mol for
F igu r e 4. MNDO-determined structures of intermediates and
transition states for the cleavage of benzyl and cinnamyl ethers
4 and 11 through dianion routes involving CIP’s and IIP’s
(SSIP’s species not shown). Geometrical features relevant to
haptomeric activation are shown: (1) C-OMe bond length and
bond order (in brackets); (2) Li-OMe distance; (3) enthalpy
of formation.
TS10(ii) and 143.6 kcal/mol for TS10), as illustrated in
Figure 5. When LiBenz was considered, instead of Li-
Naph, close values were obtained: ∆G* ) 38.1 kcal/mol
for TS6i and ∆G* ) 40.6 kcal/mol for TS10ii. These
results clearly suggest that (1) the lowest energy route
for BICLE reactions is that involving cleavage of radical
anion/ cation species, likely as CIP’s, and (2) routes
involving cleavage of CIP radical anion/ cation species
and dianion/ dication species may actually compete in
some particular cases (see below), as experimentally
demonstrated by Walsh.35
In striking contrast with the above clear-cut analysis
for BICLE processes, calculations show that VIBICLE
reactions are a bit more complex as, now, the routes
involving dianion/dication species (as CIP’s or SSIP’s,
only) are kinetically favorable over those involving the
(63) Wakefield, B. J . The Chemistry of Organolithium Compounds;
Pergamon Press: New York, 1974.

1044 J . Org. Chem., Vol. 61, No. 3, 1996
Saa´ et al.
F igu r e 5. Energy (∆G273, kcal/mol) profiles for cleavage of benzyl ether 4 promoted by lithium naphthalenide through BICLE
routes involving CIP’s [4 + 2Li(NH3)2Naph + 4NH3), SSIP’s (4 + 2(Li+(NH3)4 ) + 2Naph-•), and IIP’s [4 + 2(Li+(NH3)4 + 2Naph-•].
Stoichiometry is mantained throughout in all calculations.
corresponding radical anion/cation species, except when
the rather unrealistic IIP’s are considered. The following
data for VIBICLE processes concerning cinnamyl ether
analog 11 illustrate this analysis: ∆G* ) 37.2 kcal/mol
for TS13i but only 28.9 kcal/mol for TS18ii for cleavage
of CIPs, and ∆G* ) 42.0 kcal/mol for TS13(i) but only
36.8 kcal/mol for TS18(ii)) for cleavage through SSIP
(Figure 6). Calculations using LiBenz instead of LiNaph
afforded closely related values for cleavage of CIP: ∆G*
) 37.5 kcal/mol for transition structure 13iTS and 30.7
kcal/mol for 18iiTS. Thus, in accordance with MNDO
calculations, reductive cleavage of cinnamyl derivatives
(and, presumably of related systems as well) is clear-cut
biased toward the dianion/ dication routes when either
CIP’s or SSIP’s are involved.
represented by steps 11 f 12i f 16ii f 17ii f 18ii f
15i + MeOLi.64 That for nonconjugated systems,64
however, is even more difficult to predict on the basis of
the above MNDO results, as both routes compete and
either solvent, substrate, reductive agent, leaving group,
carrier, or metal effects may, in a given case, divert the
reaction toward one side or another.
An important consequence of kinetic significance which
derives from the above analysis is that, irrespective of
whether radical anion/ cation or dianion/ dication species
are involved in cleavage, the route of lowest energy barrier
is that involving CIP’s, those for SSIP’s or IIP’s being, in
all cases, more energy costly. This result might shed some
light into some unsolved problems such as the experi-
mental observation on radical anion species which are
Calculations on the 2-naphthyl analogs as CIP (sup-
porting information) further insist on the generality of
this phenomenon: whereas both BICLE processes (the
so-called radical anion/cation and dianion/dication routes)
for 2-(1′-methoxy-2-propen-1′-yl)naphthalene (19) pro-
moted by LiNaph are energetically very close to one
another (∆G* ) 37.1 and 36.6 kcal/mol, respectively), the
corresponding VIBICLE routes for the conjugated analog
(E)-2-(3′-methoxy-1′-propen-1′-yl)naphthalene (20) were
found to be quite far apart in energy (∆G* ) 35.9 for the
radical anion/cation route and 27.4 kcal/mol for the
dianion/dication route).
(64) MNDO calculations predict that complexation between either
radical 7 or the final π organolithium compound 15i () 8i) and lithium
methoxide is possible. Nevertheless, the equilibrium (shown below)
So, according to these results, the question of whether
reductive cleavage reactions involve radical anion/cation
or dianion/dication pairs is difficult to predict a priori.
Some trends, however, appear to be general. For exten-
sively conjugated systems such as cinnamyl ether 11 and
its naphthyl analog 20 the lowest energy route is that
of the resulting complex with the isolated species is clearly displaced
to the right due to the tendency of lithium methoxide to dimerize, in
line with the experiments described in this work.

Metal-Induced Reductive Cleavage Reactions
J . Org. Chem., Vol. 61, No. 3, 1996 1045
F igu r e 6. Energy (∆G273, kcal/mol) profiles for cleavage of cinnamyl ether 11 promoted by lithium naphthalenide through BICLE
routes involving CIP’s [4 + 2Li(NH3)2Naph + 4NH3), SSIP’s (4 + 2(Li+(NH3)4 )+ 2Naph-•), and IIP’s [4 + 2(Li+(NH3)4 + 2Naph-•].
Stoichiometry is mantained throughout in all calculations.
longer lived when generated (by pulse radiolysis) in water
than in DMF (electrochemically).9f
The answer to the central stereochemical issue which
originated this study lies, according to MNDO, in the
nature of the species suffering stepwise cleavage. The
lowest energy route found for BICLE processes as applied
to R-vinylbenzyl methyl ether (4) and related compounds
involves cleavage of a radical anion/cation 5i (a CIP
species) which is less stable than starting materials and,
therefore, does not accumulate. As only one chiral center
(it is devoid of chiral plane) exists in 5i, cleavage gives
rise to a vinyl-substituted benzyl radical which retains
the configuration of the vinyl grouping because further
reduction to the final organolithium is much faster than
equilibration. In striking contrast, the vinylogous Birch
cleavage of cinnamyl derivative 11Me (and, presumably,
of related systems as well) leads to the intermediate
formation of a long-lived radical anion/cation intermedi-
ate 12iMe the most stable conformers of which (sepa-
rated by just ∆G ) 0.6 kcal/mol) interconvert through
low energy barrier processes (Figure 7). As shown above,
the lowest energy route for VIBIHY cleavage involves not
the radical anion/cation stage but the transient dianion/
dication CIP species 18iiMe resulting from subsequent
reduction of 12iMe by LiNaph. Since a chiral plane
develops, eventually two diastereoisomeric transition
states involving syn â elimination actually compete,
namely TS18iiMe E and TS18iiMe Z. MNDO calcula-
tions show that ∆∆G* for cleavage of the E (∆G* ) +30.8
kcal/mol) and Z (∆G* ) +31.0 kcal/mol) transition
structures is 0.2 kcal/mol, which give rise to the corre-
sponding E and Z organolithium species 3ic in reason-
F igu r e 7. Stereochemical bias for cleavage of 11Me.
ably close agreement with the experimental facts.

1046 J . Org. Chem., Vol. 61, No. 3, 1996
Saa´ et al.
Lithium-6 spectra were referenced to 1 M LiCl in D2O. 2D
6Li-1H HOESY spectra50 were recorded at -20 °C (mixing
time ) 1.8 s, D1 ) 3 s, 256 increments in t1, 32 scans acquired
in 6Li per increment, 1 K data points collected per scan, total
measuring time ) 12 h). Lithium-6 was purchased from
Aldrich. Commercial THF-d8 was distilled from sodium ben-
zophenone ketyl under argon. Compounds 1a , (E)-1b (con-
taining 5% of Z isomer), (Z)-1b (containing 8% of E isomer)
and (E)-2b were prepared following the described methodology5c
and their stereochemical purity (where appropriate) deter-
mined by NMR.
P r ep a r a tion of NMR Sa m p les. Freshly cut pieces of
lithium wire (3 mol) were placed inside a previously dried
(oven; cooled under argon) 5-mm NMR tube which was then
inmediately fitted with a serum cap. A solution of 0.5 mmol
of precursors 1 or 2 in 0.5 mL of dried THF-d8 was added via
syringe. The tubes were then flamed-sealed under argon. The
mixture was briefly sonicated a 0 °C until reaction starts and
left aside at -25 °C for 48 h. The thus obtained deep orange-
red solutions of 3i were stable for at least 2 months at -20
°C.
In summary, metal-mediated Birch and vinylogous
Birch cleavage reactions has been experimentally dem-
onstrated to give rise to η3 vinyl-substituted benzyllith-
iums which do not undergo isomerization at the temper-
ature (<0 °C) of operation. 6Li-1H HOESY experiments
proved that these species do not form mixed aggregates
with the lithium silyloxide cogenerated in the reaction.
Extensive semiempirical (MNDO) calculations have shed
some light to the stereochemical problem as well as to
the more general subject of the role of counterions and
solvent in metal-promoted (lithium) reductive cleavage
reactions. A simplified model for contact, solvent-
separated, and isolated ion pairs has allowed us to eval-
uate these complex reactions in some detail. In particu-
lar, our studies on model compounds of the Birch and
vinylogous Birch cleavage reactions (BICLE and VIBI-
CLE) suggest the following: (a) the lowest energy routes
for cleavage are, in all cases, those involving contact ion
pairs in which the (lithium) counterion plays a key role
by assisting the leaving group to depart through a syn â
elimination pathway; (b) haptomers of the CIP’s radical
anion/cation or the dianion/dication species are key for
appropriately understanding the cleavage step; (c) routes
involving counterion-free species, either as solvent-
separated or isolated, nonsolvated ion pairs are more
energy costly, in all cases; (d) according to MNDO, the
lithium-promoted cleavage of benzyl derivatives (BICLE)
is a stepwise process involving transient radical anion/
cation species, whereas the lowest route for cleavage of
cinnamyl derivatives (VIBICLE), also stepwise, is that
of the dianion/dication species; and (e) the puzzling
stereochemistry observed for BICLE and VIBICLE reac-
tions is a consequence of the fact that there are two
competing diastereomeric transition states for the latter
due to the existence of a chiral plane, but only one for
the former which does not have chiral plane. According
to this reasoning, one is tempted to predict that dissocia-
tive electron transfer, disregarding the actual stage at
which cleavage occurs, can also take place at more distant
sites of the electron sink provided that haptomeric
equilibria allows it. To the best of our knowledge this is
the first time calculations predict the existence of hap-
tomers in anion radicals.29 In our view, the above results
should be of value for understanding cleavage reactions
promoted by electrochemical means or otherwise.
3ia : 1H NMR (243 K) δ: 6.35 (1H, t, J ) 7.7 Hz), 6.23 (1H,
d, J ) 7.8 Hz), 6.14 (1H, d, J ) 7.8 Hz), 6.06 (1H, d, J ) 11.8
Hz), 5.86 (1H, d, J ) 7.7 Hz), 5.25 (1H, t, J ) 6.7 Hz), 3.40
(1H, d, J ) 11.8 Hz), 1.8 (3H, s), 0.01 (9H, s, LiOSiMe3) ppm;
13C NMR (243 K) δ 149.8, 130.3, 129.7, 128.5, 119.1, 111.6,
6
104.4, 93.5, 29.0, 20.2, 6.5 (LiOSiMe3) ppm; Li NMR δ 3.03,
2.18 ppm.
1
(Z)-3ib: H NMR (253 K) δ 6.38 (1H, t, J ) 7.1 Hz), 6.29
(1H, t, J ) 7.3 Hz), 6.19 (1H, d, J ) 9.6 Hz), 6.08 (1H, d, J )
11.9 Hz), 5.88 (1H, d, J ) 8 Hz), 5.23-5.18 (2H, br m), 3.76
(d,1H, J ) 11.7 Hz), 2.02 (4H,m), 1.79 (3H, s), 1.77 (3H, s),
1.67 (3H, s), 0.00 (9H, s; LiOSiMe3) ppm; 13C NMR (253 K) δ
149.0, 131.1, 130.4, 129.6, 129.0, 128.6, 119.0, 111.2, 103.8,
97.7, 73.5, 36.0, 28.7, 27.7, 27.0, 19.4, 6.5 (LiOSiMe3) ppm;
6Li NMR δ 3.03, 2.18 ppm.
1
(E)-3ib: H NMR (253 K) δ 6.35 (1H, t, J ) 7.2 Hz), 6.23
(1H, t, J ) 7.2 Hz), 6.15-6.10 (2H, m), 5.86 (1H, d, J ) 8Hz),
5.32-5.12 (2H, br m), 3.63 (d, 1H, J ) 11.9 Hz), 2.13-2.08
(4H, m), 1.67 (3H, s), 1.64 (3H, s), 1.63 (3H, s), 0.01 (9H, s,
LiOSiMe3) ppm; 13C NMR (253 K) δ 149.2, 131.1, 130.3, 129.9,
129.0, 128.5, 119.2, 111.5, 104.3, 97.8, 73.5, 44.7, 31.7, 27.7,
6
19.4, 18.5, 6.5 (LiOSiMe3) ppm; Li NMR δ 3.03, 2.18 ppm.
Ack n ow led gm en t. Financial support by the DGI-
CyT (Projects PB90-0040 and PB93-0424) is gratefully
acknowledged. One of us (X.G.) wishes to thank the
M.E.C. (Spain) for a predoctoral fellowship. Thanks are
due to Dr. A. Llobera for some preliminary calculations.
Su p p or tin g In for m a tion Ava ila ble: Tables containing
Z matrices and GNORMs of the optimized molecular struc-
Exp er im en ta l Section
6
tures and the Li-1H HOESY spectra of 3ia , (E)-3ib, and (Z)-
Gen er a l Meth od s. Measuring frequencies were 300.1
MHz (1H), 75.4 MHz (13C), and 44.1 MHz (6Li). Quadrature
detection was carried out in both dimensions for 2D spectra.
Proton, carbon-13, and lithium-6 NMR spectra were obtained
in THF-d8, the former two being referenced to the residual
tetrahydrofuran peaks at 3.58 ppm and 67.4 ppm, respectively.
3ib are provided (46 pages). This material is contained in
libraries on microfiche, inmediately follows this article in the
microfilm version of the journal, and can be ordered from the
ACS; see any current masthead page for ordering information.
J O951834G