Highly alkyl-selective addition to ketones with magnesium ate complexes derived from Gignard reagents

Article abstract of DOI:10.1021/ol047685i

(Chemical Equation Presented) A highly efficient alkyl-selective addition to ketones with magnesium ate complexes derived from Grignard reagents and alkyllithiums is described. The nucleophilicity of R in R₃MgLi is remarkably increased compared to that of the original RLi or RMgX, while the basicity of R₃MgLi is decreased. Furthermore, a highly R-selective addition to ketones is demonstrated using RMe₂MgLi in place of R ₃MgLi.

Full text of DOI:10.1021/ol047685i

ORGANIC
LETTERS
2005
Vol. 7, No. 4
573-576
Highly Alkyl-Selective Addition to
Ketones with Magnesium Ate
Complexes Derived from Grignard
Reagents
Manabu Hatano, Tokihiko Matsumura, and Kazuaki Ishihara*
Graduate School of Engineering, Nagoya UniVersity, Furo-cho, Chikusa,
Nagoya 464-8603, Japan
ishihara@cc.nagoya-u.ac.jp
Received November 10, 2004
ABSTRACT
A highly efficient alkyl-selective addition to ketones with magnesium ate complexes derived from Grignard reagents and alkyllithiums is
described. The nucleophilicity of R in R3MgLi is remarkably increased compared to that of the original RLi or RMgX, while the basicity of
R3MgLi is decreased. Furthermore, a highly R-selective addition to ketones is demonstrated using RMe2MgLi in place of R3MgLi.
While Grignard and organolithium reagents continue to
remain popular choices for C-C bond formation in organic
chemistry,1-3 their reactions with ketones can, on occasion,
give rise to reduction products or aldol addition adducts/
starting ketone due to competing enolization problems.4 In
light of this, we decided to examine the reactions of a range
of “typical” ketones with various metal “ate” complexes,
formed from Grignard reagents and organolithiums5-7 to
see if this rise to improved results; we report here our
findings.
First, n-Bu-addition to acetophenone 1 was examined with
various common lithium and/or magnesium reagents (1.0
equiv in n-hexane or n-heptane solution) in THF at -78 °C
for 5 h (Table 1). n-BuLi gave the corresponding alcohol 2
in 62% yield along with 7% of an undesired aldol product
3, since it is sufficiently basic to generate enolate from 1
(entry 1). Grignard reagent n-BuMgCl was similarly in-
effective and gave 2 in 50% yield, 3 in 9% yield, and the
reduced product 4 in 8% yield (entry 2). n-Bu2Mg gave rise
to a better conversion than traditional n-BuLi and n-BuMgCl,
although the selectivity of 2 decreased, along with an increase
(1) (a) Wakefield, B. J. Organomagnesium Methods in Organic Chem-
istry; Academic Press: San Diego, 1995. (b) Silverman, G. S.; Rakita, P.
E. Handbook of Grignard Reagents; Marcel Dekker: New York, 1996.
(c) Richey, H. G., Jr. Grignard Reagents: New DeVelopment; Wiley:
Chichester, 2000.
(2) Reviews: (a) Lai, Y.-H. Synthesis 1981, 585-604. (b) Eisch, J. J.
Organometallics 2002, 21, 5439-5463. (c) Knochel, P.; Dohle, W.;
Gommermann, N.; Kneisel, F. F.; Kopp, F.; Korn, T.; Sapountzis, I.; Vu,
V. A. Angew. Chem., Int. Ed. 2003, 42, 4302-4320.
(3) Inoue, A.; Shinokubo, H.; Oshima, K. J. Synth. Org. Chem. Jpn.
2003, 61, 25-36.
(4) Grignard reagent with CeCl3: (a) Imamoto, T.; Sugiura, Y.; Tak-
iyama, N. Tetrahedron Lett. 1984, 38, 4233-4236. (b) Imamoto, T.;
Takiyama, N.; Nakamura, K. Tetrahedron Lett. 1985, 39, 4763-4766. (c)
Imamoto, T.; Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya, Y. J.
Am. Chem. Soc. 1989, 111, 4392-4398.
(5) (a) Uchiyama, M.; Kondo, Y.; Sakamoto, T. J. Synth. Org. Chem.
Jpn. 1999, 57, 29-41. (b) Uchiyama, M. J. Pharm. Soc. Jpn. 2002, 122,
29-46. (c) Shinokubo, H.; Oshima, K. Eur. J. Org. Chem. 2004, 2081-
2091.
(6) (a) Schulze, V.; Hoffmann, R. W. Chem. Eur. J. 1999, 5, 337-344.
(b) Kitagawa, K.; Inoue, A.; Shinokubo, H.; Oshima, K. Angew. Chem.,
Int. Ed. 2000, 39, 2481-2483. (c) Iida, T.; Wada, T.; Tomimoto, K.; Mase,
T. Tetrahedron Lett. 2001, 42, 4841-4844. (d) Inoue, A.; Kitagawa, K.;
Shinokubo, H.; Oxhima, K. J. Org. Chem. 2001, 66, 4333-4339. (e) Inoue,
A.; Kondo, J.; Shinokubo, H.; Oshima, K. Chem. Eur. J. 2002, 8, 1730-
1740. (f) Dumouchel, S.; Mongin, F.; Tre´court, F.; Que´guiner, G.
Tetrahedron Lett. 2003, 44, 2033-2035.
(7) (a) Ashby, E. C.; Chao, L.-C.; Laemmle, J. J. Org. Chem. 1974, 39,
3258-3263. (b) Richery, H. G., Jr.; DeStephano, J. P. J. Org. Chem. 1990,
55, 3281-3286.
10.1021/ol047685i CCC: $30.25
© 2005 American Chemical Society
Published on Web 01/15/2005

Table 1. Addition to Acetophenone 1 with Li- or Mg-alkyl
Table 2. Addition to Ketones 5 with n-Bu3MgLia,b
Reagents
yield (%)
entry
n-BunMXa
2
3
4
1
2
3
4b
n-BuLi
62
50
48
82
7
0
n-BuMgCl
n-Bu2Mg
n-Bu3MgLi
9
27
0
8
20
0
a n-BunMgX (1.0 equiv) was used. b n-Bu3MgLi was prepared in situ
from 1 equiv each of n-BuLi and n-Bu2Mg.
in byproducts 3 and 4. These results indicate that the careful
control of alkyl reagents between the minimum basicity and
maximum nucleophilicity is necessary to achieve the desired
addition to a ketone.8 When we examined the ate complex
n-Bu3MgLi, which is easily generated in situ from a 1:1
mixture of n-BuLi and n-Bu2Mg, we found that it exhibited
high reactivity and gave 2 in 82% yield, with no side
reactions.
Encouraged by these results with n-Bu3MgLi, we next
examined the scope of its reaction with ketones of general
structure 5 with n-Bu3MgLi in comparison with n-BuLi
(Table 2). However, even 3.0 equiv of n-BuLi could not bring
about a full conversion to product 6 in several cases (see
data in parentheses in entries 1, 4, and 7). Interestingly, we
found that the addition of equimolar 2,2′-bipyridyl (Bpy) to
n-Bu3MgLi was quite effective at accelerating the butylation
of acetophenone; this giving the corresponding alcohol in
96% yield as the sole product without a self-aldol reaction
(entry 1).9 The other alkylphenyl ketones reacted readily with
n-Bu3MgLi/Bpy, as evidenced by the reaction of propio-
phenone which proceeded readily to give the alcohol product
6 in 97% yield (entry 2); isobutyrophenone also showed an
enhancement in yield with n-Bu3MgLi as did acetophenone
(entry 3 vs entry 1). n-Bu3MgLi/Bpy was suitable for other
aromatic ketones such as R- and â-acetonaphthone which
gave the corresponding alcohols in improved respective
yields of 71%10 and 93% compared to n-BuLi (entries 4 and
5). For the cyclic ketone, R-tetralone, the corresponding
alcohol was again formed in higher (77%) yield with
n-Bu3MgLi/Bpy (entry 7). When n-Bu3MgLi/Bpy was re-
acted with dicyclohexyl ketone, no significant advantages
were noted and the products were obtained in almost the
a Reactions were carried out in THF at -78 °C for 5 h with ketone and
n-BuLi (1.0 equiv) or n-Bu3MgLi (1.0 equiv)/Bpy (1.1 equiv). n-Bu3MgLi
was prepared in situ from 1 equiv each of n-BuLi and n-Bu2Mg, except for
d in entry 1. b No reduced products were obtained in any of the runs.
c Isolated yield when 3.0 equiv of n-BuLi were used. d n-Bu3MgLi prepared
with n-BuLi (2.0 equiv) and n-BuMgCl (1.0 equiv) was used without Bpy.
same yield (95%) as with n-BuLi (entry 8). Significantly, a
troublesome aldol and/or reduction reaction did not occur
under any of the conditions tested in n-Bu3MgLi systems.
Thus, the n-Bu3MgLi reagent rather than n-BuLi or n-
BuMgCl has an ideal nucleophilicity and a negligible basicity
that makes it ideal for addition to ketones.
With regard to the effect of Bpy, n-Bu3MgLi prepared with
n-BuLi (2.0 equiv) and n-BuMgCl (1.0 equiv) was found to
be useful for this butylation without additives and gave 6 in
99% yield (data in brackets in entry 1, Table 2).11
A
magnesium ate complex with a homo component [e.g.,
R3MgLi, prepared from RLi (2.0 equiv) and RMgX (1.0
equiv)] was examined in its reaction with benzophenone 8
(8) (a) Sassian, M.; Panov, D.; Tuulmets, A. Appl. Organomet. Chem.
2002, 16, 525-529. (b) Tuulmets, A.; Panov, D.; Sassian, M. Tetrahedron
Lett. 2003, 3943-3945. (c) Sussian, M.; Tuulmets, A. HelV. Chim. Acta
2003, 86, 82-90.
(9) 12-Crown-4 instead of Bpy was also an effective activator, and 2
was obtained in 99% yield under the same reaction conditions.
(10) Interestingly, with R-acetonaphthone, 2-(R-naphthyl)-2-butanol (Et-
adduct) was obtained in 16% yield in entry 4. The ethyl-moiety in the butyl-
unit may migrate to the carbonyl, with carbon-carbon bond cleavage and
the release of ethylene. Detailed mechanistic studies are now underway.
(11) These additives, such as Bpy and 12-crown-4, generally did not
have any effects with other R3MgLi prepared from RLi and RMgX, and
such effects were only observed with n-Bu2Mg. These additives may affect
the effective ate complexation of n-Bu3MgLi from n-Bu2Mg as a magnesium
source, although the details are still unclear.
574
Org. Lett., Vol. 7, No. 4, 2005

and gave the desired product in high yield without reduction
products (Table 3).
Table 5. Ethylation to Ketones 5 with EtMe2MgLi Reagentsa,b
Table 3. Addition to Benzophenone 5 with R3MgLi
entry
R3MgLi
yield (%) of 9
1a
2b
3c
4d
n-Bu3MgLi
Ph3MgLi
Me3MgLi
Et3MgLi
95
87
>99
64
a n-Bu3MgLi: 2 equiv of n-BuLi and 1 equiv of n-BuMgCl. b Ph3MgLi:
2 equiv of PhLi and 1 equiv of PhMgBr. c Me3MgLi: 2 equiv of MeLi
and 1 equiv of MeMgBr. d Et3MgLi: 2 equiv of EtLi and 1 equiv of
EtMgBr.
n-Bu3MgLi, Ph3MgLi, and Me3MgLi gave 9 in respective
yields of 95%, 87%, and >99% (entries 1, 2, and 3). A
dramatic enhancement in reactivity was observed with
Et3MgLi to afford 9 in 64% yield, since ethylation with 1.2
equiv of EtLi showed 37% yield12 while that with 1.2 equiv
of EtMgBr showed 14% yield (entry 4).
Next, we turned our attention to hetero magnesium ate
complexes, i.e., R1R2 MgLi (R1 * R2), instead of homo
2
magnesium ate complexes such as R3MgLi. Considering the
atom economy of alkylation, the alkyl source (R) in
magnesium ate complexes of R3MgLi is not consumed
completely during the reaction and the remaining R source
would be sacrificed. This prompted us to examine R1R22MgLi
in the alkylation to ketones. The key to success here is likely
the choice of a second alkyl source (R2) in R1R2 MgLi
2
complexes, which can be replaced as an unreactive “dummy”.
Me and Ph were found to be good inactive components in
n-Bu2RMgLi-type complexes and gave the Me-adduct (8%)
and Ph-adduct (8%) as minor products 10 in the alkylation
of acetophenone 1 (entries 1 and 2, Table 4). In both cases,
the n-Bu-adducts 2 were the major products (77-78% yield)
Table 4. Addition to Acetophenone 1 with n-BunR3-nMgLia,b
a Reactions of 5 with EtMgBr (1.2 equiv) or EtMe2MgLi (1.2 equiv)
were carried out in THF at -78 °C for 5 h. b No reduced products 12 were
obtained in any of the runs under EtMe2MgLi conditions. c Reaction time
was 1 h.
yield (%)
(cf. entry 4, Table 1). In sharp contrast to the results with
n-Bu2RMgLi, a clear difference was observed between Me
and Ph components in the n-BuR2MgLi-type complexes,
although it is not clear why a Me-moiety is more favorable
than other “dummies”. Despite the decrease in Bu groups
present, the n-Bu-adduct 2 could be obtained in 72% yield
as the major product (entry 3), while n-BuPh2MgLi did not
lead to good alkyl selectivity (entry 4). While the yield of
the n-Bu-adduct 2 with n-BuMe2MgLi was the same as that
with n-Bu2MeMgLi, the efficiency of n-Bu addition with
entry
n-BunR3-nMgLi
2
10
1
2
3
4
5
n-Bu2MeMgLi
n-Bu2PhMgLi
n-BuMe2MgLi
n-BuPh2MgLi
n-Bu(t-Bu)2MgLi
78
77
72
57
61
8
8
26
41
23
a Reactions of 1 with n-BunR3-nMgLi (1.0 equiv) were carried out in
THF at - 78 °C for 5 h. b No aldol and/or reduced products were obtained
in any of the runs.
Org. Lett., Vol. 7, No. 4, 2005
575

n-BuMe2MgLi (72%) was superior to that with n-Bu2MeMgLi
(39% based on total n-Bu). Furthermore, the sterically more
demanding t-Bu was inferior to Me as an inactive moiety
(entry 5 vs entry 3).
Table 6. Addition to Benzophenone 8 with RMe2MgLi
Reagentsa,b
Using RMe2MgLi systems, we examined the ethylation
to ketones. Due to the instability and difficulty of preparing
and handling EtLi,12 there are still limitations with traditional
methods, such as with the use of a common Grignard reagent
(EtMgX). In fact, ethylations to ketones 5 with Grignard
reagent (EtMgBr) either failed or gave only a moderate yield
(Table 5). In particular, the reaction of 2,2-dimethylpro-
piophenone with EtMgBr failed to give any of the desired
product 11; instead the reduced product 12 was obtained in
50% yield. By way of contrast, the corresponding reaction
proceeded quite smoothly with the EtMe2MgLi ate complex
and gave the desired Et-adduct 11 in 89% yield (entry 4).
For other low-yield examples (14∼36%) with EtMgBr such
as isobutyrophenone, R-acetonaphthone, benzophenone and
R-tetralone, EtMe2MgLi showed high selectivities for the Et-
adduct 11 as opposed to the Me-adduct 13, giving the desired
product in extremely high yield (89∼>99%) (entries 3, 5,
7, and 8). Although EtMgBr gave fair results with aceto-
phenone, propiophenone and â-acetonaphthone (66∼77%),
further improvements were observed with EtMe2MgLi; with
Et-adducts 11 were obtained in 93-95% yield (entries 1, 2
and 6). With a heterocyclic compound such as pyridine,
EtMe2MgLi gave the corresponding products in 96-97%
yield (entries 9 and 10). Undesired Me-adducts 13 could be
easily separated by SiO2 flash column chromatography,
although the selectivities between 11 and 13 were in fact
high or perfect.
yield (%) with
RMgBr
yield (%) with
RMe2MgLi
entry
R
9
14
9
15
1
2
3
4c
Et
i-Pr
n-Pr
n-Bu
14
48
15
0
68
36
56
56
>99
87
<1
0
88
92
12
8
a Reactions of 8 with RMgBr (1.2 equiv) or RMe2MgLi (1.2 equiv) were
carried out in THF at -78 °C for 2 h. b No reduced products 14 were
obtained in any of the runs under RMe2MgLi conditions. c n-BuMgCl was
used.
In summary, we have developed a highly efficient alkyl-
selective additions to ketones using magnesium ate com-
plexes derived from Grignard reagents and alkyllithiums.
Remarkable improvements were observed with the R3MgLi
ate complex systems. Furthermore, we have demonstrated
that RMe2MgLi as a R1R2 MgLi ate complex system can be
2
used for highly selective alkylmetal addition to ketones.
These reactions are currently still under investigation in an
effort to further expand the scope of this process, in
particular, to aldehydes, esters, amides, imines, and R,â-
unsaturated carbonyl compounds.
Finally, alkylations with other RMe2MgLi systems were
performed with benzophenone 8 at -78 °C for 2 h (Table
6). As expected, alkylations with i-PrMe2MgLi, n-PrMe2-
MgLi, and n-BuMe2MgLi with high selectivities for the
corresponding R-adducts 9 in high yields of the products
were also apparent. Comparable Grignard reagents were
totally ineffective and gave significant amounts of the
reduced byproducts 14. In particular, dramatic results were
seen with n-BuMe2MgLi where the yield was improved from
0% to 92% (entry 4).
Acknowledgment. Financial support for this project was
provided by the JSPS. KAKENHI (15205021) and the 21st
Century COE Program “Nature-Guided Materials Process-
ing” of MEXT.
Supporting Information Available: Experimental pro-
cedures for addition to ketones with R3MgLi or RMe2MgLi
reagents. This material is available free of charge via the
Internet at http://pubs.acs.org.
(12) EtLi is now commercially available from Aldrich. Direct ethylation
to ketones with EtLi is not suitable at all. As another example, EtLi with
1 gave 34% of 2 along with 7% of 3.
OL047685I
576
Org. Lett., Vol. 7, No. 4, 2005