is dominated not by steric effects but by substrate-based
electronic effects. On the basis of this newly recognized con-
trol element, the extension of the McMurry reaction to the
preparation of a range of tetrakis(ortho-substituted) tetraphen-
ylethylenes and related compounds has been realized.
reduction products. Replacement of DME with THF as sol-
vent promotes simple reduction over coupling (entry 3), a
predictable consequence of the increased propensity for hy-
drogen atom donation from tetrahydrofuran over an acyclic
ether.
To suppress the formation of reactive surface hydrides,7
LiAlH4 was replaced by Zn(Cu), another standard McMurry
reductant.2 The use of Zn(Cu) proved transforming, dramati-
cally shifting the product distribution to provide, for the first
time, reasonable selectivity for the formation of the desired
tetraarylethylene product (entry 4).
Using bis(5-tert-butyl-2-methoxyphenyl)methanone 1a as
the test substrate, we initially undertook an empirical
optimization of the relevant reaction parameters (Table 1).6
Among the operational parameters, the impact of reagent
preparation time is highly significant, despite previous reports
that no reduction of TiCl3 by Zn(Cu) occurs in the absence
of substrate.8 Omitting the reagent preparation step entirely
produces several unidentified minor compounds at the
expense of the desired alkene (entry 5), but intermediate
reagent preparation times inexplicably produce less selective
reactions (entries 6 and 7). Neither the reductant stoichiom-
etry (entries 8 and 9) nor the reaction temperature (entries
10 and 11) exerts more than an incremental effect on yields
and product distributions.9
The reaction is reasonably tolerant of the titanium source;
both bulk unsolvated TiCl3 and first quality TiCl3‚1.5DME
function very similarly.10 The latter reagent, however, is
relatively insensitive to the reagent preparation time, provid-
ing nearly identical results under both standard conditions
and the operationally advantageous “instant” method8b
(entries 12 and 13). Finally, the reaction provides greater
selectivity at larger scale (entry 14): this substrate, clearly,
suffers no inhibitory steric effect.
Table 1. Reagents, Conditions,a and Product Distribution in
the McMurry Olefination Reaction of Benzophenone 1a6
yield (%)d
time (h)b (°C)c 2a 3a 4a
reductant
(equiv)
activation Treact
entry
Ti source
TiCl3
1
2
LiAlH4 (0.5)
LiAlH4 (0.25)
LiAlH4 (0.25)
Zn(Cu) (0.5)
Zn(Cu) (0.5)
Zn(Cu) (0.5)
Zn(Cu) (0.5)
Zn(Cu) (4.0)
Zn(Cu) (1.0)
Zn(Cu) (0.5)
Zn(Cu) (0.5)
14
14
14
14
0
rt
rt
rt
rt
rt
rt
rt
rt
rt
14
9 61
TiCl3
38 27 16
3e TiCl3
30
9
30
5
4
5
TiCl3
TiCl3
TiCl3
TiCl3
TiCl3
TiCl3
TiCl3
TiCl3
74 20
60 21
51 24
48 26
The dramatic shift in product distribution observed by
using the TiCl3/Zn(Cu) reagent rather than TiCl3/LiAlH4
suggests the incorporation of strongly Lewis acidic [Zn(II)?]
coordination sites into the reagent, promoting substrate
binding and reduction on the reagent surface, strongly
favoring the bimolecular olefination. Substrate binding may
arise via chelation of the proximal ether and carbonyl
functionality or via unidentate coordination of the electron-
rich carbonyl. To address this question and distinguish steric
from electronic effects, a series of differentially substituted
benzophenones have been investigated (Table 2).
1
6
2
4
2
7
6
8
14
14
14
14
0
63 13 22
74 13 14
9
10
11
reflux 65 19
0
3
37
rt
rt
rt
68 20
12f TiCl3‚1.5DME Zn(Cu) (0.5)
79 4i 7i
78 7i 4i
84 5i 7i
13g TiCl3‚1.5DME Zn(Cu) (0.5)
14
14
14h TiCl3
Zn(Cu) (0.5)
a In DME solvent, unless otherwise noted. b Reaction time to generate
the low-valent titanium reagent (LiAlH4, rt; Zn(Cu), 89 °C). c Temperature
of the reductive coupling reaction. d Isolated yields after careful purification
by flash column chromatography. e THF used as solvent. f Large-scale
experiment using 6.0 g of 1a. g Large-scale experiment using 13.7 g of 1a.
h Large-scale experiment using 9.6 g of 1a. i For these experiments, the
1
yields of the minor products were obtained by H NMR integration of the
(6) Complete experimental procedures and characterization data are
provided as Supporting Information.
crude product mixture, normalized to the isolated yield of 2a.
(7) The TiCl3/LiAlH4/THF reagent consists principally of the hydride-
rich cluster [HTiCl(thf)0.5]x, a potential source of H•, H2, and reactive
hydride: Aleandri, L. E.; Becke, S.; Bogdanivic, B.; Jones, D. J.; Rozie`re,
J. J. Organomet. Chem. 1994, 472, 97-112.
(8) (a) Fu¨rstner, A.; Jumbam, D. N. Tetrahedron 1992, 48, 5991-6010.
(b) Fu¨rstner, A.; Hupperts, A.; Ptock, A.; Janssen, E. J. Org. Chem. 1994,
59, 5215-5229. (c) Bogdanivic, B.; Bolte, A. J. Organomet. Chem. 1995,
501, 109-121. (d) Stahl, M.; Pidun, U.; Frenking, G. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2234-2237.
(9) Control experiments conducted by adding tetrakis(5-tert-butyl-2-
methoxyphenyl)ethene 2a to an active McMurry reduction of 2,2′-
dimethoxybenzophenone 1b clearly indicate that overreduction product 3a
does not arise from direct reduction of 2a.
(10) The unsolvated TiCl3 powder has recently become difficult to obtain
commerically. TiCl3‚1.5DME is conveniently prepared from TiCl4 and Ti
powder in DME: Sullivan, J. M. U.S. Patent 6307063, 2001.
In nearly all cases, the McMurry reaction proceeds to give
three principal reduction products, accompanied by traces
of other, unidentified, byproducts (eq 1). The low-valent ti-
tanium species generated from titanium trichloride and Li-
AlH4 is a generally inferior reagent, with the highest percen-
tage of undesired reduction products 3a and 4a obtained from
using the literature-optimized3a 1:0.5 molar ratio of TiCl3/
LiAlH4 (entry 1). Lowering the ratio of TiCl3/LiAlH4 to
1:0.25 dramatically improves the isolated yield of olefination
product 2a (entry 2), but the reaction remains dominated by
1492
Org. Lett., Vol. 8, No. 7, 2006