Reactions of aliphatic ketones R2CO (R = Me, Et, iPr, and tBu) with the MCl4/Li(Hg) system (M = U or Ti): Mechanistic analogies between the McMurry, Wittig, and Clemmensen reactions

Article abstract of DOI:10.1002/1521-3765(20010716)7:14<3043::AID-CHEM3043>3.0.CO;2-D

Analysis of the products of the reactions of ketones R₂CO (R = Me, Et, iPr, tBu) with the MCl₄/Li(Hg) system (M = U, Ti) at 20°C revealed significant differences. For R = Me, the reaction proceeded exclusively (M = U) or preferentially (M = Ti) via a metallopinacol intermediate resulting from dimerization of ketyl radicals. Pinacol was liberated by hydrolysis, and tetramethylethylene was obtained after further reduction at 65°C. For R = iPr, formation of iPr₂C=CiPr₂ as the only coupling product, the nonproduction of this alkene by reduction of the uranium pinacolate [U]-OCR₂CR₂O-[U] (R = iPr) at 20°C, and the instability of the corresponding titanium pinacolate towards rupture of the pinacolic C-C bond indicated that reductive coupling of iPr₂CO did not proceed by dimerization of ketyl radicals. Formation of 2,4-dimethyl-2-pentene was in favor of a carbenoid intermediate resulting from deoxygenative reduction of the ketyl. These results revealed that for sterically hindered ketones, McMurry reactions can be viewed as Wittig-like olefination reactions. For R = tBu, no coupling product was obtained and the alkane tBu₂CH₂ was the major product. The involvement of the carbenoid species [M]=CtBu₂ was confirmed by its trapping with H₂O, leading to tBu₂CH₂, and with the aldehydes RCHO, giving the cross-coupling products tBu₂C=C(R)H (R = Me, tBu). Therefore, in the case of severely congested ketones, McMurry reactions present strong similarities to the Clemmensen reduction of ketones, owing to the involvement in both reactions of carbenoid species which exhibit similar reactivity. Wiley-VCH Verlag GmbH, 2001.

Full text of DOI:10.1002/1521-3765(20010716)7:14<3043::AID-CHEM3043>3.0.CO;2-D

FULL PAPER
Reactions of Aliphatic Ketones R CO (R ˆ Me, Et, iPr, and tBu) with the
2
MCl /Li(Hg) System (M ˆ U or Ti): Mechanistic Analogies between the
4
McMurry, Wittig, and Clemmensen Reactions
Claude Villiers* and Michel Ephritikhine*[a]
Abstract: Analysis of the products of
can be viewed as Wittig-like olefination
reactions. For R ˆ tBu, no coupling
product was obtained and the alkane
iPr) at 208C, and the instability of the
corresponding titanium pinacolate to-
wards rupture of the pinacolic C�C
the reactions of ketones R CO (R ˆ Me,
2
Et, iPr, tBu) with the MCl /Li(Hg)
4
system (M ˆ U, Ti) at 20 8C revealed
significant differences. For R ˆ Me, the
reaction proceeded exclusively (M ˆ U)
or preferentially (M ˆ Ti) via a metal-
lopinacol intermediate resulting from
dimerization of ketyl radicals. Pinacol
was liberated by hydrolysis, and tetra-
methylethylene was obtained after fur-
ther reduction at 658C. For R ˆ iPr,
formation of iPr CˆCiPr as the only
tBu CH2 was the major product. The
2
bond indicated that reductive coupling
involvement of the carbenoid species
of iPr CO did not proceed by dimeriza-
2
[M]ˆCtBu was confirmed by its trap-
2
tion of ketyl radicals. Formation of 2,4-
dimethyl-2-pentene was in favor of a
carbenoid intermediate resulting from
deoxygenative reduction of the ketyl.
These results revealed that for sterically
hindered ketones, McMurry reactions
ping with H O, leading to tBu CH , and
2
2
2
with the aldehydes RCHO, giving the
cross-coupling products tBu CˆC(R)H
2
(R ˆ Me, tBu). Therefore, in the case of
severely congested ketones, McMurry
reactions present strong similarities to
the Clemmensen reduction of ketones,
owing to the involvement in both reac-
tions of carbenoid species which exhibit
similar reactivity.
2
2
Keywords: carbenoids ´ C�C cou-
pling ´ McMurry reactions ´ titani-
um ´ uranium
coupling product, the nonproduction of
this alkene by reduction of the uranium
pinacolate [U]�OCR CR O�[U] (R ˆ
2
2
Introduction
Me
Me
UCl4, Li(Hg)
.
O
[U]
2
O
2
Recently we reported on the McMurry type reactions of
acetone with uranium tetrachloride and lithium amalgam,
which gave pinacol and tetramethylethylene.[1] For the first
time, pinacolate intermediates in the reductive coupling of an
aliphatic ketone could be isolated and characterized, and the
reactions shown in Scheme 1 confirmed the mechanism that
had so far been postulated and generally accepted. The two
major steps were clearly defined: the coupling itself, which
occurred at 208C by dimerization of ketyl radicals and led to
the metallopinacol and pinacol after hydrolysis, whereas
deoxygenation of the metallopinacol into tetramethylethy-
lene, the rate-determining step, was observed at higher
temperature. By monitoring these studies, we were surprised
to find that the reductive coupling of diisopropyl ketone did
not afford the expected a-diol at room temperature, but gave
tetraiisopropylethylene. Moreover, 2,4-dimethyl-2-pentene
THF, 20 °C
Me
Me
Me
O
[U]
[U]
Me
Me
Me
[
2]
O
H2O
Li(Hg), 65 °C
Me Me
Me
Me
Me
Me
Me
Me
HO
OH
Scheme 1. Reductive coupling of Me
2
4
CO with the UCl /Li(Hg) system
�
(
[U] ˆ UCl
4
).
[
a] Dr. C. Villiers, Dr. M. Ephritikhine
Service de Chimie Mol e culaire, DSM, DRECAM, CNRS URA 331
CEA Saclay, 91191 Gif-sur-Yvette (France)
Fax : (‡33)169086640
was the major product of the reaction.[3] This distinct behavior
of iPr CO led us to examine in more detail the reactions of the
2
E-mail: villiers@drecam.cea.fr, ephri@drecam.cea.fr
aliphatic ketones R CO (R ˆ Me, Et, iPr, tBu) with the UCl /
2
4
Chem. Eur. J. 2001, 7, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3043 $ 17.50+.50/0
3043

C. Villiers and M. Ephritikhine
FULL PAPER
Li(Hg) system, and also the TiCl /Li(Hg) system, which is a
typical McMurry reagent. The results presented here give new
insights into the mechanism of the McMurry reaction.
Table 1. Product distribution [%] for the reaction of ketone R
2
CO (R ˆ
4
[
a]
Me, Et, iPr) with MCl
4
(M ˆ U, Ti) and Li(Hg).
R
M
R
1
2
CO (R
2
COH)
2
R
3
2
CˆCR
2
R
4
2
CHOH MeCR'ˆCHCHR'Me
2
5b or 5c
Me
Et
iPr
Me
Et
U
U
U
U
U
U
0
100
14
0
100
28
0
27
6
0
0
3
4
0
0
0
3
2
1
0
0
23
30
0
10
5
0
0
1
0
51
56
0
57
93
9
10
0
4
1
3
5
3
2
Results and Discussion
[
b]
b]
b]
[
[
For R ˆ Me, Et, and iPr, the ketones R CO 1 were effectively
2
iPr
coupled by the MCl /Li(Hg) system (M ˆ U, Ti) to give a
4
Me Ti 67
Et Ti 87
iPr Ti 95
Me Ti 25
Et Ti 42
mixture containing HOCR CR OH 2 and R CˆCR 3. Not
2
2
2
2
surprisingly, the total yield of 2 and 3 varied with R in the
order Me > Et > iPr, reflecting the more difficult coupling of
[
c]
61
22
0
12
18
14
[
c]
15
11
3
8
the more sterically hindered ketones. Coupling of tBu CO did
2
[
c]
[
4]
iPr Ti 63
not occur, as invariably noted so far with a variety of
[
2, 5]
McMurry reagents.
For a given ketone, the yield of
[a] Molar ratio 1:1:1 when not specified. [b] Stepwise addition of 1 to the
mixture of UCl and Li(Hg). [c] Molar ratio 1:1:2.
coupling products was higher with M ˆ U; the remarkable
4
efficiency of the uranium system was previously observed in
[
6]
the reductive coupling of aromatic ketones. Most instructive Table 2. Product distribution [%] for the reaction of ketone R CO (R ˆ Me, Et, iPr)
2
[
a]
was the variation with R of the relative proportions of 2 and 3, with MCl
4
(M ˆ U, Ti) and Li(Hg).
and also the nature and yield of the other products.
C CR
ˆ
ˆ
R
M
R
1
2
CO (R
2
COH)
2
R
3
2
2
R
4
2 2
CHOH MeCR' CHCHR'Me RCH R
2
5b or 5c
6
Reactions of the ketones R CO (R ˆ Me, Et, iPr) with the
[b]
[b]
2
Me
U
U
U
0
87
56
0
50
17
0
0
8
MCl /Li(Hg) system (M ˆ U or Ti): In addition to the Et
0
10
0
5
27
26
32
23
11
23
13
29
38
21
29
0
4
iPr
Me Ti
Et Ti
1
coupling products 2 and 3, reactions of the ketones R CO 1
2
[
b]
[b]
(
a, R ˆ Me; b, R ˆ Et; and c, R ˆ iPr) with MCl (M ˆ U, Ti)
4
0
7
10
4
4
and Li(Hg) gave the products resulting from the reduction
and deoxygenation of 1Ðthe alcohol 4, the alkene 5 (b, R' ˆ
H; c R' ˆ Me), and the alkane 6 (Scheme 2, Tables 1 and 2).
iPr Ti 10
[
a] Molar ratio 1:1:2 (M ˆ U) or 1:1:3 (M ˆ Ti). [b] Propene and propane were not
detected.
R
MCl4, Li(Hg)
THF, 20 °C
O
formed (1c). The presence of the other products 3 ± 6 revealed
R
that formation of 2b or 2c by dimerization of the ketyl radical
1
.
Li[Cl UOCR ] (R ˆ Et, iPr) was in competition with other
a, R = Me; b, R = Et; c, R = iPr
4
2
side reactions. The occurrence of competing reduction
processes in the reactions of 1b and 1c was also indicated
by the greater yield of 2b and the lower yields of 5b or 5c
when the reactions were carried out in two steps, by addition
R
R
R
R
R
R
R
R
OH
H
H
H
R
R
Me
Me
+
+
+
+
of the ketone to the mixture of UCl and Li(Hg). With this
HO
OH
R
R
4
R'
R'
procedure, no change was observed with 1a, confirming that
2
3
4
5
6
Li(Hg) served only for the reduction of UCl to LiUCl , which
4
4
b, R' = H
reacted with 1a to give 2a (Scheme 1). Evidently, some of the
reducing agent Li(Hg) was used for reducing 1b and 1c into
c, R' = Me
Scheme 2. Reaction of R
system (M ˆ U, Ti).
2
CO (R ˆ Me, Et, iPr) with the MCl
4
/Li(Hg)
5
b and 5cÐwhich were quite unexpected productsÐand one
could think that 3b and 3c would result from reduction of the
pinacolate intermediate.
The hydrocarbons 3, 5, and 6 were formed before hydrolysis.
Other products, which represent at most 15% of the total
yield, have not been identified; they probably resulted from
aldol reactions.
Conversion of 1 with UCl and Li(Hg) was total (1a, 1b) or
4
near 90% (1c) when the reagents were used in the molar ratio
of 1:1:2 (Table 2). With these proportions, reaction of 1a again
gave 2a as the only coupling product. The color change of the
solution, from green to red, indicated that the pinacolate
Treatment of 1a with UCl and Li(Hg) in the molar ratio of
4
III
1
:1:1 at 208C gave the metallopinacol Li Cl UOCMe C-
intermediate was reduced to the U analogue without
2
4
2
Me OUCl , and the pinacol 2a was obtained in quantitative
suffering deoxygenation; 3a was liberated only after heating
the solution under reflux. This result is in accordance with the
generally accepted mechanism of the McMurry reaction.[
Therefore, formation of 3c as the sole coupling product from
1c at 208C was most striking, since the McMurry alkene
synthesis supposedly requires higher temperature for the
deoxygenation step. It is then difficult to explain why the
2
4
yield after hydrolysis (Scheme 1). It was demonstrated that
the metallopinacol resulted from dimerization of the ketyl
2]
.
radical Li[Cl UOCMe ] , formed by electron transfer from
4
2
[
1]
LiUCl to the carbonyl. In contrast, conversion of 1b and 1c
4
was not complete under the same conditions (Table 1) and the
a-diol was either not the only organic product (1b) or was not
3044
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3044 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

McMurry Reactions
3043±3051
uranium pinacolates issued from 1b and 1c would be more
easily reduced and deoxygenated at 208C than the pinacolate
issued from 1a.
more convincing was the reaction of 7c with TiCl to give 1c
4
and TiCl , indicating that the titanium pinacolate was not
3
stable towards rupture of its C�C bond. Such reversible
cleavage of metallopinacols was previously assessed in the
reaction of benzophenone and acetophenone with titanium
When UCl was replaced with TiCl , three equivalents of
4
4
Li(Hg) were required for complete conversion of the ketone
II
[7]
(
Tables 1 and 2); this is in agreement with the fact that Ti
compounds and in the production of the unsymmetric
III
rathern than U species were the active species in the
alkene Ph CˆCTol (Tol ˆ p-Me-C H ) by treating a mixture
2
2
6
4
reductive process. Treatment of the less reactive ketone 1c
of pinacols HOCPh CPh OH and HOCTol CTol OH with
2 2 2 2
uranium powder.[ Reaction of 7b with TiCl4 was less
conclusive since 3b was obtained in 52% yield; however, 5b
was also formed, showing that the pinacolic C�C bond was
8]
with 1 equivalent each of TiCl and Li(Hg) gave essentially
4
TiCl , which was inert towards the carbonyl substrate. In the
3
reactions of 1 with TiCl and Li(Hg) in the molar ratio of 1:1:3,
4
it was again surprising that the alkenes 3b and 3c were the
major or only coupling products. Moreover, in contrast to that
observed with the uranium system, reaction of 1a gave 3a.
Although the ratio 3b:2b, but not 3a:2a, was found to
increase after further addition of Li(Hg), changing from 32:17
to 48:1, it is again unlikely that the titanium pinacolate issued
from 1b was more easily reduced than the pinacolate issued
from 1a. Also, if 3 was formed by reduction and deoxygena-
tion of a pinacolate intermediate, it is then difficult to explain
why a titanium pinacolate would be more easily reduced and
deoxygenated than its uranium counterpart. In fact, it became
very doubtful that formation of this alkene occured by this
route since the yield of 3a was not significantly increased after
further addition of Li(Hg) to the reaction mixture.
cleaved under these conditions.
The above results suggest that reactions of ketones 1 with
the MCl /Li(Hg) system followed two distinct routes, leading
4
to the pinacol 2 or the alkene 3, respectively. In particular, the
absence of 2c in the coupling products, the nonproduction of
3c by reduction of the uranium pinacolate, and the instability
of the titanium pinacolate indicate that reductive coupling of
1c did not proceed by dimerization of ketyl radicals.
So what is the mechanism of formation of 3? The presence
of significant amounts of 5b and 5c among the organic
products was informative. These alkenes are in fact the major
products of the reactions of 1b and 1c with the UCl /Li(Hg)
4
system and it is possible that propene, which would be formed
from 1a, was not detected by GC because it escaped into the
vapor phase. Formation of 5b and 5c, which did not result
from Li(Hg) reduction of the alkoxide Cl TiOCHR (R ˆ Et,
Synthesis and reduction of the uranium and titanium
pinacolates by successive treatment of LiOCR CR OLi 7
2
2
3
2
with MCl and Li(Hg) confirmed that metallopinacols were
iPr), is reminiscent of the synthesis of alkenes by deoxygena-
tion of alicyclic ketones with zinc and chlorotrimethylsilane.[
These reactions were shown to proceed by the intermediacy of
zinc carbenoid species, which were formed by reduction and
4
9]
not likely intermediates in the formation of alkenes 3. The
lithium pinacolates 7 were prepared in situ by treating the a-
diol 2 with two equivalents of LitBu in THF. As previously
[1b]
described for the synthesis of Li Cl UOCMe CMe OUCl ,
deoxygenation of the ketyl radical. The Zn�Me SiCl combi-
2
4
2
2
4
3
7
was treated with two equivalents of MCl4 to give the
nation was also effective in the reductive coupling of aryl and
a,b-unsaturated carbonyl compounds; here again, these
reactions were shown to involve carbenoid species and no
pinacol-type intermediates.[ Such involvement of carbenoid
corresponding uranium or titanium pinacolate. This latter was
not isolated but directly reduced with Li(Hg) (3 or 4
equivalents for M ˆ U or Ti, respectively) and after 24 h at
10]
2
08C, the reaction mixture was hydrolyzed. The yields of
species in the reactions of 1 with the MCl /Li(Hg) system
4
compounds 2 ± 6 are listed in Table 3.
would thus explain the formation of 5, after a-H migration,
and also that of 3. The mechanism would be similar to that of
the reductive coupling of ketones by means of low-valent
Table 3. Product distribution [%] for the successive treatment of LiOCR
2
CR
2
OLi 7
[11]
niobium,
molybdenum,
or
tungsten
compounds
with MCl
4
and Li(Hg).[
a]
Scheme 3).[12±14] Cleavage of a ketone CˆO bond with
(
R
M
R
1
2
CO (R
2
COH)
2
R
3
2
CˆCR
2
R
4
2
CHOH MeCR'ˆCHCHR'Me RCH
2
R
2
5b or 5c
6
[
b]
[b]
Me U
0
0
0
0
0
0
98
90
90
92
35
3
0
0
1
5
52
21
0
0
5
0
0
R
R
Et
iPr
U
U
5
0
0
R
R
R
R
O
O
R
R
R
R
R
R
0
2
[M]
[b]
[b]
[M]
[M]
Me Ti
Et Ti
iPr Ti
O
-[M]=O
-
[M]=O
R
5
1
7
R
[
c]
39
15
Scheme 3. Reductive coupling of ketones via a carbenoid species.
[
a] Molar ratio 1:2:3 (M ˆ U) or 1:2:4 (M ˆ Ti). [b] Propene and propane were not
detected. [c] Treatment of 7c with TiCl gave 1c quantitatively (see text).
4
formation of alkylidene and oxo ± metal compounds was
[13]
indeed observed with [WCl (PMePh ) ]
and dinuclear
2
2 4
The uranium pinacolates Li Cl UOCR CR OUCl (R ˆ
niobium and tungsten complexes, Na Nb [p-tBu-calix[4]-
(O)4]2 and [W (OCH tBu) (pyridine) ]. The alkylidene
2 2 6 2
2
4
2
2
4
III
2
2
[11]
[14]
Me, Et, iPr) were reduced to the corresponding U com-
plexes, which were stable and in no case could be further
compound resulting from this four-electron addition was
found to react with another molecule of ketone or aldehyde to
give the coupling alkene, presumably via a metallaoxetane
intermediate. Formation of carbenoid species by reduction of
deoxygenated at 208C. Reaction of 7a with TiCl clearly
4
showed that 3a could not result from reduction of a titanium
pinacolate, since only 5% of the alkene was obtained. Even
Chem. Eur. J. 2001, 7, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3045 $ 17.50+.50/0
3045

C. Villiers and M. Ephritikhine
FULL PAPER
ketones with low-valent titanium complexes has not been
reported so far but such a reaction is quite conceivable in view
of the highly oxophilic and reducing character of titanium. It is
important to recall here that in 1973, one year before the first
publication of McMurry and Fleming, Tyrlik and Wolocho-
in agreement with the above considerations, 2c was never
detected among the coupling products. However, it is
surprising that the formation of 5c was overlooked thus far.
For example, it was reported that 1c reacted with TiCl and K
3
to give 3c in 37% yield;[ we confirmed this result but also
found that the major product of the reaction was 5c (45%
yield). Interestingly the best route to the diol 2c is the reaction
16]
wicz, who studied the reaction of 1a with the TiCl �Mg
3
system, suggested that 3a was obtained via the carbene
[15]
[17]
Me C:, which resulted from deoxygenation of the ketone.
of 1c with lithium metal in THF. When using UCl or TiCl ,
2
4
4
In the reductive coupling of 1 by the MCl /Li(Hg) system, the
reductive coupling of 1b certainly proceeded by the two
possible mechanistic pathways, via the metallopinacol and the
carbenoid intermediate, to give both 2b and 3b. These two
routes would be well separated, at least for M ˆ U, since the
metallopinacol was not the precursor of 3b. It is noteworthy
4
putative carbenoid species were likely obtained by deoxyge-
native reduction of the ketyl radical rather than oxidative
addition of 1; the presence of the ketyl was evidenced by the
formation of the alcohol 4, which was not observed in
reactions with the tungsten complexes. The carbenoid species
were not characterized, they could not be detected by NMR
spectroscopy and their involvement was inferred only from
analysis of the reaction products.
that successive treatment of 7b with TiCl and Li(Hg) gave a
4
small quantity of 5b, confirming that the metallopinacol could
be transformed back into the ketyl radical by reversible
cleavage of the C�C bond and was not necessarily the unique
The two distinct routes followed by the reductive coupling
of ketones 1 depicted in Scheme 4 were dependent on the
reactivity of the first and common intermediate, the ketyl
precursor of 3b.
In addition to being dimerized or deoxygenatively reduced
to give the metallopinacol or the carbenoid species respec-
tively, the ketyl radical was transformed into the alkoxide
Cl M�OCHR , the precursor of 4 by hydrolysis. This trans-
3
2
R
R
R
R
2
MCl4
formation proceeded either by H abstraction from the solvent
or by disproportionation of the ketyl radical. Such dispropor-
tionation of ketyls, which was first observed during the
reduction of saturated ketones by means of alkali metals in
O
2
.
O
[M]
Cl
2
Li(Hg)
1
b
Li(Hg)
[18]
a
c
aprotic solvents, was also found to occur in the reaction of
III
the U compound [(C H ) U(thf)] with ketones RCOCH R',
5
5
3
2
leading to an equimolar mixture of [(C H ) U�OCHR-
5
5 3
[
19]
(
CH R')] and [(C H ) U�OCR(ˆCHR')].
In that case,
2
5
5 3
R
R
R
.
O
[M] Cl
Me
Me
dimerization of the ketyl radical [(C H ) U�OCR(CH R')]
5
5
3
2
R
[
M]
was impeded by the steric hindrance of the (C H ) U frag-
5
5 3
R
R'
R'
O
[M] Cl
R
ment. In the reactions of 1 with the MCl /Li(Hg) system, the
4
5
enolate derivative resulting from disproportionation of the
ketyl radical would react with 1 to afford aldolization
products.
H2O
d
Li(Hg)
b
1
If we now consider the differences between the MCl4/
Li(Hg) systems (M ˆ U or Ti), interestingly in the reactions of
R
R
R
R
R
R
R
1
a and 1b, the ratio [2]:[3] is much greater for M ˆ U, while
HO
OH
R
the yield of 4 is higher for M ˆ Ti. These results, which
indicate that the pinacolization route was more highly favored
with the uranium system, can again be explained by steric
factors: the uranium ketyls, being less sterically encumbered
than the titanium counterparts because of the greater size of
the metal center, would dimerize more easily. Since the
alkenes 3b and 5b result respectively from the two possible
reactions of the carbenoid species, intermolecular coupling
with another molecule of 1b or intramolecular a-H migration,
the ratio [3b]/[5b] is expected to increase when the carbenoid
species is in the presence of higher concentrations of 1b. This
was indeed observed with the titanium system ([3b]/[5b] ˆ
32:7), by comparison with the uranium system ([3b]/[5b] ˆ
5:21), in which a major part of 1b was found in the form of the
pinacolate intermediate. In the reductive coupling of 1c, the
ketyl radical was transformed either into the alkoxide or the
carbenoid species and the ratio ([3c]‡[5c])/[4c], which is
equal to 2.4 and 0.9 for M ˆ U and Ti respectively, shows that
deoxygenative reduction of the ketyl radical is more effective
2
3
Scheme 4. Reductive coupling of R
2
CO (R ˆ Me, Et, iPr) with the MCl
4
/
Li(Hg) system at 208C. ([M] ˆ UCl or TiCl). a) M ˆ Ti or U and R ˆ Me or
2
Et; b) M ˆ Ti and R ˆ Me, Et, iPr; M ˆ U and R ˆ Et, iPr; c) M ˆ Ti or U
and R ˆ Et, iPr; d) M ˆ Ti and R ˆ Et.
radical, and this reactivity was determined by steric factors:
the more sterically hindered the ketyl radicals are, the more
difficult they are to dimerize. For R ˆ Me, the ketyl radical
was easily dimerized and 1a was transformed exclusively
(
M ˆ U) or preferentially (M ˆ Ti) into 2a. In contrast, with
M ˆ U or Ti, coupling of the ketyl radical issued from 1c did
not occur, having been impeded by the steric hindrance of the
iPr groups and the instability of the resulting pinacolate (for
M ˆ Ti). In that case, reaction of the ketyl radical followed a
different pathway, the deoxygenative reduction to the carbe-
noid species, which provided 3c after further reaction with 1c,
or 5c after H migration. Such McMurry reactions of 1c were
previously performed with a variety of titanium reagents and
with M ˆ U. With the UCl
/Li(Hg) system, the ratio [3c]:[5c]
4
3046
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3046 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

McMurry Reactions
3043±3051
was equal to 27:29 at 208C and to 15:50 at 658C, suggesting
that intramolecular H migration within the carbenoid species
was favored at higher temperature.
previously reported in a few cases, especially with sterically
hindered ketones, without any comment on the mechanism of
their formation.[
5, 22, 23]
It is likely, from the hydrolysis and
deuterolysis experiments, that 6d was formed by successive
addition of H(D) atoms to a carbenoid species [M]ˆCtBu , via
2
Reaction of tBu CO with the MCl /Li(Hg) system (M ˆ U or
2
4
an alkyl intermediate [M]�CH(D)tBu2 (Scheme 6). Treat-
Ti): Although the McMurry reaction proved to be very
efficient for the synthesis of sterically hindered and strained
olefins, it failed in the reductive coupling of ditertiobutyl
tBu
tBu
tBu
tBu
tBu
MCl4, Li(Hg)
H(D)
H(D)
[
2, 5]
ketone 1d,
and the long-awaited diol 2d and alkene 3d
O
[
M]
tBu [M]
H(D)
could never be prepared, despite many synthetic ap-
[
20, 21]
1d
proaches.
However, it was interesting to examine in
detail the reactivity of 1d towards the MCl /Li(Hg) system,
4
RCHO
H migration
which permitted us to get further insights into the species that
are actually formed under the conditions of the McMurry
reaction.
By comparison with 1c, total conversion of 1d was
observed after 24 h at 208C when one more equivalent of
the reducing agent was used, that is with the components 1d,
tBu
tBu
R
H
Me
Me
tBu
tBu H(D)
tBu
H(D)
12, R = Me
3, R = tBu
11
6d
1
MCl , and Li(Hg) in a molar ratio of 1:1:3 for M ˆ U, and
4
Scheme 6. Involvement of carbenoid species in the reaction of tBu
with the MCl
/Li(Hg) system (M ˆ U, Ti).
2
CO
1
6
:1:4 for M ˆ Ti. In addition to the alcohol 4d and the alkane
4
d, smaller amounts or traces of the alkenes 8, 9, 10, and the
cyclopropyl derivative 11 were also obtained (Scheme 5). The
yields of compounds 4d, 6d, and 8 ± 11 are indicated in
Table 4.
ment of 1d with the TiCl /Li(Hg) system in [D ]THF, which
4
8
afforded partially deuteriated 6d ([D ]-6d:[D ]-6d ˆ 27:73) in
1
0
the first step of the reaction, revealed that H(D) atoms were
provided essentially by the ketone (H) and, to a lesser extent,
by the solvent (D). It is quite plausible that formation of H
atoms has to be related to that of isobutylene 8, which would
itself result from H abstraction from a tertiobutyl radical. In
agreement with this hypothesis, much more deuterium was
incorporated into 6d ([D ]-6d:[D ]-6d ˆ 75:25), while 8 was
tBu
MCl4, Li(Hg)
O
THF, 20 °C
tBu
1
d
1
0
formed in trace amounts when TiCl was replaced with UCl .
4
4
tBu
tBu
tBu
tBu
tBu
H
Me
Me
Me
Me
OH
H
H
H
Me
Me
tBu
tBu
tBu
The lack of C5 hydrocarbons among the products suggests
+
+
+
+
+
.
that a tBu radical could not be produced by cleavage of the
carbenoid or alkyl species. On the other hand, dissociation of
.
4
d
6d
8
9
10
11
tBu from the ketyl radical would give an acyl derivative
Scheme 5. Reaction of tBu
2
CO with the MCl
4
/Li(Hg) system (M ˆ U, Ti).
[Ti]COtBu, which would be unstable and decompose by
decarbonylation into a titanium carbonyl [Ti]CO and another
.
tBu radical; however, CO was not detected in the reaction
mixture.
Table 4. Product distribution [%] for the reaction of tBu
2
CO with MCl
4
(M ˆ U, Ti)
CCH CHR
and Li(Hg).[
a]
Another indication of the involvement of carbenoid species
in the reaction of 1d with the MCl /Li(Hg) system was
provided by the formation of the cyclopropyl compound 11
since this latter was the major hydrocarbon resulting from
thermal or photochemical decomposition of the carbene
4
M
R
2
CHOH RCH
2
R
Me
8
2
CˆCH
2
Me
9
2
CˆCHR
R
10
2
CˆCH
2
Me
11
2
2
6d[
b]
4
d
U
Ti 25
35
55 (20)
40 (24) 20
1
1
5
8
8
0
2
[
21, 24]
precursors tBu CN PPh or tBu CN .
tane, [M][CH(tBu)CMe CH ], formed by intramolecular
A metallacyclobu-
2
2
3
2
2
[
a] Molar ratio 1:1:3 (M ˆ U) or 1:1:4 (M ˆ Ti). [b] Total yield (yield before
2
2
hydrolysis).
insertion of the carbene moiety into a b C�H bond could be
an intermediate in the formation of 11. However, 11 was
detected in trace amounts, indicating that rearrangement of
the carbenoid species by b-H migration is not a preferred
pathway.
While the hydrocarbon products were observed before
hydrolysis of the reaction mixture, importantly the alkane 6d
was also formed, together with the alcohol 4d, upon addition
of water. Formation of the alkane 6d, which is the major
product, and the other hydrocarbons 8 ± 11 was never noticed
in such McMurry reactions of 1d, and the alcohol 4d was the
The presence of carbenoid species in the reactions of 1d
with the MCl /Li(Hg) system was further strongly evidenced
4
by their trapping with the aldehydes RCHO (R ˆ Me, tBu),
which readily afforded the cross-coupling products
[
5]
only product characterized. However, reduction of ketones
into the corresponding alkanes during McMurry reactions was
tBu CˆC(R)H (R ˆ Me, 12 or tBu, 13) (Scheme 6). This
2
coupling reaction would require the approach of the carbonyl
Chem. Eur. J. 2001, 7, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3047 $ 17.50+.50/0
3047

C. Villiers and M. Ephritikhine
FULL PAPER
[
29]
molecule to the sterically hindered carbenoid species and it
was observed only with aldehydes; even the less encumbered
ketone 1a was not able to couple.
Formation of the alkenes 9 and 10 apparently resulted from
CH2 migration between two ditertiobutylene fragments
coupling of cyclohexanone with the TiCl �K system and the
3
alkenes RCHˆCH (R ˆ Me, Ph) were formed during the
2
coupling of acetone and acetophenone on reduced alumina.[
These alkenes would indicate the occurrence of carbenes as
intermediates, even if such species could not be trapped with
the usual reagents. It was also observed that during the
23]
(
C H ). This rearrangement would proceed by C�H and
9
18
C�C bond activation, possibly within the carbenoid species.
coupling of PhCOMe on reduced TiO surfaces, the pinacol
product was evolved at much higher temperature than
2
The metallacyclobutane [M][CH(tBu)CMe CH ] would un-
2
2
dergo C�C bond cleavage, leading to 9 and the carbenoid
PhCHˆCH ; this result was explained by the formation of
2
species [M]ˆCH , which would react with 1d to give 10; such
pinacolate species at protected sites that are difficult to reduce
at low temperature,[ but it is possible that pinacols and
alkenes were produced in parallel rather than sequential
processes.[
2
23]
rearrangement constitutes the key step in alkene metathesis
[
25]
reactions.
However, it is difficult to explain why, in the
30]
reaction of 1d with the UCl /Li(Hg) system, the alkene 10
4
(
C H ) was obtained in 8% yield, whereas 9 (C H ) was
1
0
2
0
8
1
6
formed only in a trace amount. Notably, reductive coupling of
benzaldehyde on reduced TiO surfaces afforded, in addition
2
to stilbene, small amounts of styrene and benzene. It was
suggested that the latter were formed by cracking of
stilbene.
The above results confirm that the ketyls formed by
reduction of 1d do not undergo dimerization; this is not
surprising in view of the steric hindrance of the tBu groups.
Interestingly, in contrast to the reaction with 1c, treatment of
Analogies between the McMurry and the Wittig and Clem-
mensen reactions: Reactions of ketones 1 with the MCl4/
Li(Hg) system revealed that the McMurry reactions of
sterically hindered ketones (1c) can be viewed in fact as
Wittig-like olefination reactions. Carbonyl olefinations by
means of titanium carbenoid species are well documented;[
these are performed with a variety of reagents designed by
[
26]
31]
Tebbe,[ Grubbs, Petasis, and Takai and co-workers.
32]
[33]
[34]
[35]
1
d with sodium metal in diethyl ether did not afford the
[
36]
[11]
corresponding pinacol but gave the 1,4 diol tBu C(OH)-
Alkylidene complexes of zirconium, niobium, and Group
2
6 metals[
12±14]
are also effective in such Wittig-like olefination
CH CH C(OH)tBu , which resulted from reaction of the
2
2
2
[
27]
dianion of 1d with the solvent. As in the case of 1c, the
ketyl radical of 1d was reduced to a carbenoid species,
reactions, which are supposed to proceed via metallaoxetane
intermediates.
[
M]ˆCtBu , which exhibited a distinct behavior since it did
When the ketone is so sterically hindered that it cannot be
2
not react with another ketone molecule, because of steric
factors and was transformed preferentially into the alkane
coupled (1d), its reaction with the MCl /Li(Hg) system
4
leading to the alkane 6 resembles the Clemmensen reaction,
reduction of a carbonyl to a methylene group by means of zinc
and hydrochloric acid.[ On the other hand, it is well known
that Clemmensen reactions of ketones RCOR' also produce a
number of side products, in particular the pinacol RR'C(OH)-
C(OH)RR', the coupling alkene RR'CˆCRR', and alkenes
6
d. Interestingly, in the reactions of 1c or 1d with the
37]
MCl /Li(Hg) system, the total yields of compounds issued
4
from the carbenoid species (3c‡5c‡6c and 6d) are
almost the same, about 55% for M ˆ U and about 40% for
M ˆ Ti; this result indicates that the same amount of
carbenoid species was formed in both reactions. The
resulting from deoxygenation of the ketone, such as cyclo-
hexene from cyclohexanone. The involvement of a common
intermediate in the McMurry and Clemmensen reactions is
then questionable. In fact, although the mechanism of the
Clemmensen reduction is not perfectly understood, it is
species [M]ˆCtBu appeared to be much more stable
2
than [M]ˆCiPr . They were present in the reaction mixture
2
before hydrolysis and could be characterized more directly
by deuterolysis and reactions with aldehydes. Even if
they were more stable, the actual carbenoid and alkyl species
could not be separated from the alkoxide [M]OCHtBu2
generally accepted that
a bis(chlorozinc)alkyl species
[37, 38]
(ClZn) CRR' or a carbenoid species ZnˆCRR',
the likely
2
(
M ˆ U, Ti, or Li) and their exact composition was not
precursors of the alkane RCH R', could react with the ketone
2
determined. It is possible that such species adopt a dinuclear
structure with bridged oxo and alkylidene or alkyl ligands,
to give the coupling alkene. It has also been demonstrated that
pinacol coupling is a competing reaction that does not have a
common intermediate.[ Moreover, it has already been
outlined that the mechanistic features of the conversions of
the cyclohexanone-cyclohexene type by means of zinc and
chlorotrimethylsilane bear a strong resemblance to those of
the Clemmensen reduction, with the involvement of organo-
zinc carbenoid intermediates.[ Also, the reductive coupling
of aryl and a,b-unsaturated carbonyl compounds by this
38]
[
M] (m�O)(m�CtBu ) or [M] (m-O)(m-CHtBu ), similar to
2
2
2
2
those obtained by reaction of ketones with lanthanide metals
[
14, 28]
or electron-rich metal complexes.
Other evidence from the literature for the possible involve-
ment of carbenoid species in McMurry reactions: The above
results lead us to think that some McMurry reactions could be
reexamined by considering the possible involvement of
carbenoid intermediates. Formation of large amounts of
alkanes in the McMurry reactions of sterically hindered
9]
Zn�Me SiCl system was shown to proceed via carbenoid
3
species, and not via pinacolic intermediates.[
10]
These observations demonstrate the complexity of the
mechanism of the McMurry reaction; this complexity was not
suspected for a long time, as the analogies with the Wittig and
Clemmensen reactions were overlooked.
[
5, 22, 23]
ketones has already been noted.
Alkenes resulting from
deoxygenation of ketones were also observed. Thus, cyclo-
hexene was detected among the products of the reductive
3048
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3048 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

McMurry Reactions
3043±3051
Conclusion
Dimerization of ketyl radicals issued from aliphatic ketones
is also prohibited by steric factors and in that case, repre-
sented by 1c, the ketyls undergo deoxygenative reduction to
carbenoid species. These further react with the ketone to
afford coupling alkenes RR'CˆCRR' and their rearrangement
From the results reported here and our previous work,
reactions of ketones RCOR' with low-valent titanium species,
the McMurry reaction, follows different pathways depending
on steric factors such as the steric bulk of the R and R'
substituents and the steric saturation of the coordination
sphere of the active metal species; these routes are summar-
ized in Scheme 7.
by H migration give the alkene RCHˆCHR'' (R'ˆCH R''),
2
which may be the major product of the reaction.
Finally, in the case of severely congested ketones, like 1d,
the ketyl radical is also reduced to the carbenoid species but
the latter is not able to react with the ketone and is
In all cases, the first step of the reaction is formation of the
ketyl radical by one-electron transfer from titanium to the
carbonyl. This ketyl can disproportionate or abstract an H
atom from the solvent, leading to the alcohol RR'CHOH and
aldolization products after hydrolysis. When the ketyl is not
sterically hindered, it can dimerize into a titanium pinacolate,
which is readily transformed into the a-diol RR'C(OH)-
C(OH)RR' by hydrolysis, or subsequently deoxygenated to
the corresponding alkene RR'CˆCRR' in the rate determin-
preferentially converted into the alkane RCH R'.
2
Mechanistically, reductive coupling reactions of carbonyl
compounds by means of low-valent titanium reagents are
much more complicated than described in the literature.
Metallopinacols are not the only precursors to the coupling
alkenes, which are alternatively obtained by reaction of a
carbenoid species with the ketone. In that case, the McMurry
reaction can be viewed as a Wittig-like olefination reaction.
The so-called McMurry reactions also present strong anal-
ogies with the Clemmensen reduction of carbonyl compounds,
owing to the involvement in both reactions of carbenoid
intermediates, which exhibit similar reactivity.
ing step. This route corresponds to the mechanism that was
generally accepted for the McMurry reaction. Metallopina-
cols were indeed isolated in the McMurry type reactions of
[
1]
[6]
Me CO and Ph CO with the UCl �Na(Hg) system and
2
2
4
were converted into Me CˆCMe and Ph CˆCPh after
2
2
2
2
further treatment with the reducing agent. However, it was
demonstrated that the symmetric pinacol coupling of an
aromatic ketone PhCOR is impeded by the presence of bulky
R groups and/or coordination of sterically demanding ligands
Experimental Section
General: All reactions were carried out at 208C under argon (<5 ppm
oxygen or water) using standard Schlenk vessel and vacuum-line techni-
ques or in a glove box. THF was dried and deoxygenated over sodium
benzophenone ketyl and distilled immediately before use; [D ]THF was
8
dried over Na-K alloy. The organic products were identified by NMR
spectroscopy, GC, and GC-MS by comparison with authentic samples,
[
39]
on the metal complex. The dissymetric dimerization of the
ketyl radical through para phenyl/carbonyl carbon coupling,
reminiscent of coupling of trityl radicals,[40] was then found to
occur.
R
R
R
MCl4, Li(Hg
OH
dismutation
2
O
2
.
O
[M]
+ aldolization products
H2O
R'
H
R'
R'
R, R' = alkyl or phenyl
symmetric coupling
R, R' = alkyl
deoxygenation
R' = phenyl
para coupling
Li(Hg)
Ph
R
R
R
O
O
[M]
[
M]
O
R'
O
R
[M]
[
M]
R
H
[M]
R'
R'
∆
Li(Hg)
H migration
R' = CH2R''
H abstraction
or H2O
H2O
RCOR'
Li(Hg)
2
H O
Ph
HO
R
R
COR
R
H
H
H
R'
R
R
R
R
R'
OH
R'
H
H
H
R''
HO
R'
R'
Not sterically hindered
ketones
Sterically hindered ketones
Scheme 7. Distinct pathways for the McMurry reactions of ketones RCOR'.
Chem. Eur. J. 2001, 7, No. 14 ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3049 $ 17.50+.50/0
3049

C. Villiers and M. Ephritikhine
FULL PAPER
before and/or after hydrolysis or deuterolysis of the reaction mixture. Most
aldehyde CH
3
CHO (1.6 mL, 0.028 mmol) or tBuCHO (3.1 mL, 0.028 mmol)
was introduced into the tube with a microsyringe. The H NMR spectra
1
1
of the reactions were monitored by H NMR spectroscopy in [D
8
]THF.
1
showed the immediate formation of 12 or 13 (ca. 10% yield).
The H NMR spectra were recorded on a Bruker DPX 200 instrument and
referenced internally using the residual protio solvent resonances relative
to tetramethylsilane (d ˆ 0). The GLC analyses were performed on a
Chrompack CP 9002 apparatus equipped with a capillary CP Wax 57 CB
column. The mass spectra were obtained using a Hewlett-Packard 6890 ±
Acknowledgements
5
2
973 instrument operating in the ionization mode and equipped with an HP
3 (60 m) chromatography column.
Â
Â
The authors thank Dr. Alain Vandais (Service des Molecules Marquees,
CEA Saclay) for recording the GC-MS spectra.
Lithium amalgam (1.05% Li) was prepared by addition of Li to Hg in
boiling p-cymene,[ and was titrated by flame absorption spectroscopy.
41]
4 4
TiCl and tBuLi (Aldrich) were used as received; UCl was prepared as
described in reference [42]. Ketones 1, pinacol 2a, and aldehydes MeCHO
and tBuCHO (Aldrich) were dried on molecular sieves. Compounds 3a,
[
1] a) O. Maury, C. Villiers, M. Ephritikhine, Angew. Chem. 1996, 108,
215 ± 1216; Angew. Chem. Int. Ed. Engl. 1996, 35, 1129 ± 1130; b) M.
1
Ephritikhine, O. Maury, C. Villiers, M. Lance, M. Nierlich, J. Chem.
Soc. Dalton Trans. 1998, 3021 ± 3027.
4
a ± c, 5b, 6b, 6c, 8, and 9 were purchased from Aldrich. The alcohol 4d
was prepared by LiAlH
4
reduction of 1d, and the diol 2b[ was synthesized
43]
[
2] a) J. E. McMurry, Chem. Rev. 1989, 89, 1513 ± 1524; b) G. M. Rob-
ertson in Comprehensive Organic Synthesis, Vol. 3 (Eds.: B. M. Trost,
I. Fleming, G. Pattenden), Pergamon, Oxford, 1991, 563; c) R. G.
Dushin in Comprehensive Organometallic Chemistry II, Vol. 12 (Ed.:
L. S. Hegedus), Pergamon, Oxford, 1995, 1071; d) T. Lectka in Active
Metals: Preparation, Characterization, Applications, (Ed.: A. Fürst-
ner), VCH, Weinheim, 1995, 85; e) A. Fürstner, B. Bogdanovic,
Angew. Chem. 1996, 108, 2582 ± 2609; Angew. Chem. Int. Ed. Engl.
4
by reductive coupling of 1b with the UCl /Li(Hg) system (vide infra).
Compounds 2c,[ 3b, 3c, 5c, 6d, 10,[47] 11, 12, and 13 were
17]
[44]
[16]
[45]
[46]
[21]
[48]
[49]
prepared by published methods.
Reactions of the ketones 1a, 1b, and 1c with MCl
Li(Hg): In a typical experiment, an NMR tube was charged with UCl
4
(M ˆ U or Ti) and
4
(
(
12 mg, 0.031 mmol) or TiCl
20.5 mg for 1 equiv Li) in [D
4
(3.4 mL, 0.031 mmol) and 1.05% Li(Hg)
]THF (0.4 mL). The ketone was introduced
8
into the tube with a microsyringe. The mixture was stirred at 208C by
1
996, 35, 2442 ± 2449; f) T. Wirth, Angew. Chem. 1996, 108, 65 ± 68;
attaching the tube perpendicular to the axis of an electrical rotor. These
Angew. Chem. Int. Ed. Engl. 1996, 35, 61 ± 63; g) M. Ephritikhine,
Chem. Commun. 1998, 2549 ± 2554.
1
reactions were monitored by H NMR spectroscopy. Reactions of 1a were
completed after 2 h, while those of 1c required 24 h to complete. In some
experiments, further Li(Hg) was then added into the reaction mixture and/
or the mixture was heated under reflux by placing the tube in a sand bath at
[
3] C. Villiers, M. Ephritikhine, Angew. Chem. 1997, 109, 2477 ± 2479;
Angew. Chem. Int. Ed. Engl. 1997, 36, 2380 ± 2382.
4] C. Villiers, A. Vandais, M. Ephritikhine, J. Organomet. Chem. 2001,
[
808C. The reaction mixture was hydrolyzed (2 mL D
2
O) and the products
6
17 ± 618, 744 ± 747.
1
± 6 were analyzed by NMR spectroscopy, GC, and GC-MS. The results are
[
[
5] D. Lenoir, H. Burghard, J. Chem. Res. 1980, 4715 ± 4725.
6] C. Villiers, R. Adam, M. Lance, M. Nierlich, J. Vigner, M. Ephriti-
khine, J. Chem. Soc. Chem. Commun. 1991, 1144 ± 1145.
summarized in Tables 1 and 2.
4
Synthesis of 2b: A flask was charged with UCl (440 mg, 1.16 mmol),
1
.05% Li(Hg) (1534 mg, 2.32 mmol Li), and 1b (123 mL, 1.16 mmol) in
[7] a) B. Bogdanovic, A. Bolte, J. Organomet. Chem. 1995, 502, 109 ± 121;
b) R. S. P. Coutts, P. C. Wailes, R. L. Martin, J. Organomet. Chem.
1973, 50, 145 ± 151.
[8] B. E. Kahn, R. D. Rieke, Organometallics 1988, 7, 463 ± 469.
[9] W. B. Motherwell, J. Chem. Soc. Chem. Commun. 1973, 935.
[10] A. K. Banerjee, M. C. Sulbaran de Carrasco, C. S. V. Frydrych-Houge,
W. B. Motherwell, J. Chem. Soc. Chem. Commun. 1986, 1803 ± 1805.
THF (15 mL). After stirring for 24 h at 208C, the reaction mixture was
hydrolyzed (10 mL of brine). The organic layer was evaporated to dryness,
leaving a white powder of 2b (60 mg, 60%). The product was characterized
1
[43]
by its H NMR spectrum, and its mass spectrum. MS (70 eV, EI): m/z
‡
‡
‡
(
(
%): 145 (4) [M � Et], 127 (13) [M � Et � H
2
O], 87 (56) [Et
2
COH ], 69
‡
‡
10) [C H ], 57 (100) [EtCO ].
5 9
[
11] A. Caselli, E. Solari, R. Scopelliti, C. Floriani, J. Am. Chem. Soc. 1999,
21, 8296 ± 8305.
12] Y. Fujiwara, R. Ishikawa, F. Akiyama, S. Teranishi, J. Org. Chem. 1978,
3, 2477 ± 2480.
13] J. C. Bryan, J. M. Mayer, J. Am. Chem. Soc. 1990, 112, 2298 ± 2308.
14] a) M. H. Chisholm, K. Folting, J. A. Klang, Organometallics 1990, 9,
Successive reactions of the pinacolates LiOCR
with MCl
and Li(Hg) (M ˆ U, Ti): In a typical experiment, an NMR tube
was charged with the a-diol 2 (ca. 10 mg) in [D ]THF (0.4 mL) and tBuLi
2 equiv, 1.7m solution in pentane) was added with a microsyringe. After
5 min, MCl
2
CR
2
OLi 7 (R ˆ Me, Et, iPr)
1
4
[
8
4
(
1
[
[
4
(2 equiv) and Li(Hg) (4 or 3 equiv for M ˆ Ti and U,
respectively) were introduced into the tube. The mixture was stirred by
attaching the tube perpendicular to the axis of an electrical rotor. After
6
1
02 ± 606; b) M. H. Chisholm, K. Folting, J. A. Klang, Organometallics
990, 9, 607 ± 613; c) M. H. Chisholm, K. Folting, K. C. Glasgow, E.
24 h at 208C, the mixture was hydrolyzed and the products 1 ± 6 were
Lucas, W. E. Streib, Organometallics 2000, 19, 884 ± 892.
15] S. Tyrlik, I. Wolochowicz, Bull. Soc. Chim. Fr. 1973, 2147 ± 2148.
16] J. E. McMurry, M. P. Fleming, K. L. Keese, L. R. Krepski, J. Org.
Chem. 1978, 43, 3255 ± 3266.
analyzed by NMR spectroscopy and GC. The results are summarized in
Table 3.
[
[
Reactions of tBu
experiment, an NMR tube was charged with TiCl
UCl (22.0 mg, 0.058 mmol) and 1.05% Li(Hg) (153 mg, 0.232 mmol Li for
M ˆ Ti; 115 mg, 0.174 mmol Li for M ˆ U) in THF or [D ]THF (0.4 mL).
2
CO with the MCl
4
/Li(Hg) system (M ˆ U, Ti): In a typical
4
(6.3 mL, 0.058 mmol) or
[
17] I. N. Nasarov, Doklady Akad. Nauk. SSSR 1934, 4, 291; Chem. Abstr.
4
1
935, 2914.
8
[
18] a) V. Rautenstrauch, M. Geoffroy, J. Am. Chem. Soc. 1977, 99, 6280 ±
The ketone 1d (10 mL, 0.058 mmol) was introduced into the tube. The
mixture was stirred at 208C for 24 h by attaching the tube perpendicular to
the axis of an electrical rotor. The solvent and the volatile products of the
reaction were transferred under vacuum into another NMR tube cooled in
liquid nitrogen. The mixture of 6d and 8 ± 11 was analyzed by NMR
spectroscopy, GC, and GC-MS. The nonvolatile products of the reaction
6
286; b) Z. Jedlinski, A. Misiolek, W. Glowkowski, Synlett 1990, 213 ±
2
14.
[
[
19] R. Adam, C. Villiers, M. Ephritikhine, Tetrahedron Lett. 1994, 35,
73 ± 574.
5
20] a) A. Krebs, W. Born, B. Kaletta, W. U. Nickel, W. Rüger, Tetrahedron
Lett. 1983, 24, 4821 ± 4824; b) J. Dannheim, W. Gram, H. Hopf, C.
Parrodi, Chem. Ber. 1987, 120, 871Ð872; c) H. M. Sulzbach, E. Bolton,
D. Lenoir, P. v. R. Schleyer, H. F. Schaefer III, J. Am. Chem. Soc. 1996,
2 8
were deuterolyzed (10 mL of D O) in [D ]THF (0.4 mL), leading to the
formation of 4d and 6d. The yields of the reaction products are listed in
Table 4.
1
18, 9908 ± 9914.
Synthesis of the cross coupling alkenes tBu
3): An NMR tube was charged with TiCl (6.3 mL, 0.058 mmol) and 1.05%
Li(Hg) (153 mg, 0.232 mmol Li) in [D ]THF (0.4 mL) and 1d (10 mL,
.058 mmol) was introduced into the tube with a microsyringe. The mixture
2
CˆC(R)H (R ˆ Me, 12 or tBu,
[21] T. G. Back, D. H. R. Barton, M. R. Britten-Kelly, F. S. Guziec, J.
Chem. Soc. Perkin 1 1976, 2079 ± 2089.
[22] T. L. Chen, T. H. Chan, A. Shaver, J. Organomet. Chem. 1984, 268, C1-
C6.
1
4
8
0
was stirred at 208C for 24 h by attaching the tube perpendicular to the axis
of an electrical rotor. The solvent and the volatile products of the reaction
were evaporated off under vacuum and [D ]THF (0.4 mL) was added. The
8
[23] K. G. Pierce, M. A. Barteau, J. Org. Chem. 1995, 60, 2405 ± 2410.
[24] J. E. Gano, R. H. Wettach, M. S. Platz, V. P. Senthilnathan, J. Am.
Chem. Soc. 1982, 104, 2326 ± 2327.
3050
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3050 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

McMurry Reactions
3043±3051
[
25] R. H. Grubbs in Comprehensive Organometallic Chemistry (Eds.: G.
Wilkinson, F. G. A. Stone, E. W. Abel), Pergamon, Oxford, 1982, 499.
26] H. Idriss, K. G. Pierce, M. A. Barteau, J. Am. Chem. Soc. 1994, 116,
[36] M. D. Fryzuk, P. P. Duval, S. S. S. H. Mao, S. J. Rettig, M. J. Zawor-
otko, L. R. McGillivray, J. Am. Chem. Soc. 1999, 121, 1707 ± 1716.
[37] E. Vedejs, Org. React. 1975, 22, 401 ± 422.
[38] T. Nakabayashi, J. Am. Chem. Soc. 1960, 82, 3900 ± 3906, 3906 ± 3908,
3909 ± 3913.
[39] a) K. J. Covert, P. T. Wolczanski, S. A. Hill, P. J. Krusic, Inorg. Chem.
1992, 31, 66 ± 78; b) O. Maury, C. Villiers, M. Ephritikhine, Tetrahe-
dron Lett. 1997, 38, 6591 ± 6592.
[40] W. P. Neumann, W. Uzick, A. K. Zarkadis, J. Am. Chem. Soc. 1986,
108, 3762 ± 3770.
[
[
[
[
3
063 ± 3074.
27] P. D. Bartlett, T. T. Tidwell, W. P. Weber, Tetrahedron Letters 1970, 33,
919 ± 2921.
28] Z. Hou, H. Yamazaki, Y. Fujiwara, H. Taniguchi, Organometallics
992, 11, 2711 ± 2714.
29] R. Dams, M. Malinowski, I. Westdorp, H. Y. Geise, J. Org. Chem.
982, 47, 248 ± 259.
2
1
1
[
[
30] M. A. Barteau, Personal communication.
[41] J. Alexander, G. S. K. Rao, J. Chem. Educ. 1970, 4, 277.
[42] J. A. Hermann, J. F. Suttle, Inorg. Synth. 1975, 5, 143 ± 145.
[43] Ben Redjeb Sadok, Y. L. Pascal, M. L. Bouguerra, Ann. Chim. 1975,
10, 169 ± 176.
31] a) R. Beckaus, Angew. Chem. 1997, 109, 694 ± 722; Angew. Chem. Int.
Ed. Engl. 1997, 36, 686 ± 713; b) S. H. Pine in Carbonyl Methylenation
and Alkylidenation using Titanium-Based Reagents, Vol. 43 (Ed.: L. A.
Paquette), Wiley, 1993, 1.
[44] D. Lenoir, Synthesis 1977, 553 ± 554.
[
32] a) F. N. Tebbe, R. L. Harlow, J. Am. Chem. Soc. 1980, 102, 6149 ± 6151;
b) F. N. Tebbe, G. W. Parshall, D. W. Ovenall, J. Am. Chem. Soc. 1979,
[45] K. Laali, R. J. Gerzina, C. M. Flajnik, C. M. Geric, A. M. Dombroski,
Helv. Chim. Acta 1987, 70, 607 ± 611.
1
01, 5074 ± 5075.
33] a) R. H. Grubbs, R. H. Pine in Comprehensive Organic Synthesis, Vol.
(Eds.: B. M. Trost, I. Flemming, L. A. Paquette), Pergamon, New
[46] a) J. P. Kennedy, J. Org. Chem. 1970, 35, 532Ð536; b) F. C. Whitmore,
H. A. Southgate, J. Am. Chem. Soc. 1938, 60, 2571 ± 2573.
[47] L. Fitjer, U. Quabeck, Synth. Commmun. 1985, 15, 855 ± 864.
[48] S. Hellmann, H. D. Beckhaus, C. Rüchardt, Chem. Ber. 1983, 116,
2219 ± 2237.
[
5
York, 1991, 1115; b) R. H. Grubbs, S. J. Miller, G. C. Fu, Acc. Chem.
Res. 1995, 28, 446 ± 456.
[
34] a) N. A. Petasis, E. I. Bzowej, J. Am. Chem. Soc. 1990, 112, 6392 ±
[49] P. D. Bartlett, T. T. Tidwell, J. Am. Chem. Soc. 1968, 90, 4421 ± 4428.
6394; b) N. A. Petasis, D. K. Fu, Organometallics 1993, 12, 3776 ± 3780.
[
35] a) K. Takai, Y. Kataoka, T. Okazoe, K. Utimoto, Tetrahedron Lett.
1
1
988, 29, 1065 ± 1068; b) K. Utimoto, K. Takai, NATO ASI Ser. Ser. C
989, 269, 379 ± 381.
Received: January 22, 2001 [F3016]
Chem. Eur. J. 2001, 7, No. 14
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3051 $ 17.50+.50/0
3051