Efficient synthesis of β-hydroxy ketones from allylic alcohols by catalytic formation of ruthenium enolates

Article abstract of DOI:10.1002/chem.200801690

A study was conducted to demonstrate the synthesis of β-hydroxy ketones from allylic alcohols by catalytic formation of ruthenium enolates. KOtBu was added to a mixture of complex Ru, containing a cyclopentadienyl ligand bearing five phenyl groups. It was demonstrated that ruthenium chloride reacts with KOtBu, forming a ruthenium tert-butoxide complex. It was also demonstrated that the reaction of the tert-butoxide complex with allylic alcohol produced a new alkoxide, as detected by H NMR spectroscopy.

Full text of DOI:10.1002/chem.200801690

COMMUNICATION
DOI: 10.1002/chem.200801690
Efficient Synthesis of b-Hydroxy Ketones from Allylic Alcohols by Catalytic
Formation of Ruthenium Enolates
[
a]
Agnieszka Bartoszewicz, Madeleine Livendahl, and Belꢀn Martꢁn-Matute*
Transition-metal complexes 1 catalyze the transformation
of allylic alcohols 2 into enols (enolates) (Scheme 1).
This internal redox process avoids the use of stoichiometric
amounts of oxidizing and reducing agents. In the presence
of aldehydes, the in situ generated enolates can be trapped
and form important b-hydroxy ketones (aldols) (Scheme 1a).
major problem has been the efficiency of the reaction due
to the formation of unwanted ketone by-products (3,
Scheme 2b) and a low syn/anti diastereoselectivity. Li et al.
[1a–-c]
used [Ru ACHTUNGRTENN(NUG PPh ) Cl ] (1a) in toluene/H O mixtures at 1008C
3 3 2 2
[2a]
to obtain aldols in moderate yields (27–72%). When they
used aromatic allylic alcohols, such as a-vinylbenzyl alcohol
[2b]
(
2a), propiophenone (3a) became the major product. The
yield of the aldol products dramatically increased in the
[
2b]
presence of In
liquid, the reactions proceeded well at 908C.
employed [Ru(PPh ) HCl] to produce aldols in up to 72%
A
H
U
G
R
N
U
G
They also found that in an ionic
3
[2c–d]
Grꢀe et al.
AHCTUNGTRENNUNG
[3a]
yield and in short reaction times (<2 h). However, under
such conditions, ketones 3 were formed (5–52%). They have
also used Fe and Ni complexes.
[3b–e]
In the former case, small
amounts of regioisomeric aldols were produced. Better re-
sults were obtained with Ni complexes and Mg salts as co-
catalysts; aldol products were formed in high yields together
with small amounts of ketones 3 (2–15%).
Scheme 1. a) Formation of aldols from allylic alcohols using a catalytic
amount of a transition metal complex. b) Unwanted isomerization path-
way.
As part of our ongoing research, we decided to search for
a ruthenium complex that would perform the isomerization–
aldol domino process with the highest possible atom econo-
my. Thus, we aimed to find a ruthenium catalyst that could
completely suppress the formation of unwanted ketones 3
while yielding aldol products in quantitative yields, and ide-
ally, under very mild reaction conditions. We report here the
most efficient ruthenium-catalyzed transformation of allylic
alcohols into aldols, where the formation of unwanted by-
products is completely suppressed and aldol products are
formed in up to 99% yield at ambient temperature. We also
provide evidence for a mechanism via coordinated alkoxide
and coordinated a,b-unsaturated ketone that accounts for
the diastereoselectivity obtained.
This transformation is important not only because of the
new CꢀC bond which is formed, but also because two new
stereogenic centers are created. Furthermore, the formation
of metal enolates via isomerization of allylic alcohols over-
comes some of the limitations of the classical approaches.
For example, stoichiometric amounts of strong bases or stoi-
chiometric formation of enol derivatives are not necessary,
self-condensation products are not produced, and the regio-
selectivity can be controlled.
The coupling of allylic alcohols with aldehydes has al-
[1b,2–4]
ready been performed with some success.
However, a
Ru–halide complexes are the catalysts of choice in many
[1]
transformations involving hydrogen transfer. In the early
90s, a break-through came with the discovery by Bꢁckvall
et al. of the dramatic acceleration effect (10 –10 fold) in-
duced by the addition of a base (KOH) to a transfer hydro-
[
a] A. Bartoszewicz, M. Livendahl, Dr. B. Martꢂn-Matute
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University (Sweden)
Fax : (+46)8-15-4908
3
4
genation reaction of ketones catalyzed by [Ru ACHTUGNTRNEUNG( PPh ) Cl ]
3 3 2
E-mail: belen@organ.su.se
[5]
[2a]
(
1a). Other bases (K CO , RLi, ROK)
to activate
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801690.
2
3
metal–halide complexes can be used. Changing the base
Chem. Eur. J. 2008, 14, 10547 – 10550
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
10547

used to activate the complex can result in a complete
change of the reactivity during catalysis. In particular,
KOtBu has given excellent results, and in one case, a com-
zation of 2a to ketone 3a was not detected. A variety of al-
lylic alcohols (2a–d) could be coupled with a number of al-
dehydes (4a–g) affording aldols 5–14a in excellent yields
(Scheme 2, Table 1), and in most cases, under very mild re-
action conditions. Furthermore,
[6]
plex with structure of L RuOtBu was characterized as a cat-
n
[6d–e]
alytic intermediate.
We reasoned that the activity of a variety of Ru–Cl com-
plexes in the coupling reaction of allylic alcohols with alde-
hydes could be enhanced by the use of KOtBu. Thus, we
studied the Ru-catalyzed couplings between a-vinylbenzyl
alcohol (2a) and p-chlorobenzaldehyde (4a) using commer-
the domino transformation
takes place with complete re-
gioselectivity, as shown in
entry 10, since regioisomeric
[8]
aldol 14b was not formed.
cially available ruthenium complexes [Ru
A
H
U
G
R
N
U
G
Unfortunately, primary allylic
alcohols failed to yield the coupling product.
In an effort to understand the diastereoselectivity
3
3
2
5
and [h -CpRu
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(PPh ) Cl] (1b) activated by
3
2
amount of KOtBu (Scheme 2). Unfortunately, neither cata-
lyst 1a nor 1b gave any coupling product at ambient tem-
achieved, we followed the coupling of 2a and 4a in
1
perature. Instead, isomerization of a-vinylbenzyl alcohol
[D ]toluene by H NMR spectroscopy. We observed that at
8
1
(
2a) to propiophenone (3a, Scheme 2b, R =H, R=Ph) was
the beginning of the reaction, the syn-aldol was the major
diastereomer (syn/anti 94:6). The ratio slowly changed as
the reaction proceeded yielding higher amounts of the anti
diastereomer (Figure 1).
observed in only 3–5 h. Upon heating at 508C, 1a and 1b af-
forded aldol 5 in 22–34% together with ketone 3a in about
40% yield. We decided to turn our attention to other Ru
complexes. We had used Ru complex 1c [h -
Ph MeCp)Ru(CO) Cl] before in the coupling of allylic al-
(
4
2
[7]
cohol 2a with benzaldehyde (4b). Unlike complexes 1a or
b, complex 1c afforded aldol product 6 in high yield (72%,
1
syn/anti 52:48) at 508C after only 2 h. However, 1c did not
suppress the formation of ketone 3a, which was obtained in
28% yield.
Figure 1. Cross-coupling of 2b and 4a in [D
2
8
]toluene catalyzed by 1d at
08C. t=0 min corresponds to the first H NMR spectrum recorded [D:
a; : 5 (syn+anti); ꢂ: % of syn in 5].
1
2
&
The cross-coupling between deuterated allylic alcohol
[
[
D ]2a and aldehyde 4a afforded monodeuterated aldol
D ]5, with the deuterium label exclusively on the methyl
1
1
group (Scheme 3).
Scheme 2. Ru-catalyzed reaction of allylic alcohols 2 with aldehydes 4.
We postulated that if the mechanism of the coupling
occurs via Ru–enolates (Scheme 2), increasing the size of
the ligands on Ru could prevent protonation at the oxygen
atom and thus minimize the formation of ketone 3. Impor-
tantly, the steric environment of the C2 carbon of the Ru–
enolate, where the new CꢀC bond will be formed, would be
Scheme 3. Coupling of a deuterium-labeled allylic alcohol.
affected less. We were pleased to find that the Ru complex
containing a cyclopentadienyl ligand bearing five phenyl
Based on the results shown above, we propose the mecha-
nism shown in Scheme 4. Ruthenium chloride 1d reacts with
5
groups [h -(Ph Cp)Ru(CO) Cl] (1d) activated by KOtBu
5
2
yielded aldol 5 after only 3.5 h at room temperature in
99% yield (Scheme 2 and Table 1, entry 1) from allylic al-
cohol 2a and p-chlorobenzaldehyde (4a, 1.5 equiv). Isomeri-
KOtBu forming
(L RuOtBu, 15).
n
gives a new alkoxide 16,
a
ruthenium tert-butoxide complex
Reaction of 15 with allylic alcohol 2a
[
6d,e–9]
>
[4,7,10]
1
as detected by H NMR spec-
10548
www.chemeurj.org
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 10547 – 10550

b-Hydroxy Ketones from Allylic Alcohols
COMMUNICATION
[
a]
Table 1. Cross-coupling between allylic alcohols and aldehydes catalyzed by Ru complex 1d.
that an s-cis conformation of
[
b]
[c]
[d]
Entry
T [8C]/t [h]
Aldol/Ketone [%]
Aldol product
Yield [%] /syn:anti
the unsaturated ketone, impos-
sible for 2-cyclohexenone to
adopt, is required for the 1,4-
hydride addition to occur.
Therefore, from the s-cis con-
formation of 17, hydride addi-
tion yields a cis-enolate (18).
Formation of the syn-aldol
from cis-enolate-18 and alde-
hyde 4 can be explained by a
Zimmerman–Traxler six-mem-
1
25/3.5
5/3a (>99:<1)
>99 (88)/77:23
[
e]
2
3
4
5
25/4
7/3b (>99:<1)
8/3a (>99:<1)
6/3a (99:1)
>96 /79:21
[f]
25/3.5
35/2.5
35/1
99 (78 )/83:17
97 (84)/69:31
95 (93)/60:40
[11]
bered transition state.
The
anti-aldol may be mainly pro-
duced by Ru-catalyzed epimeri-
zation of the syn-aldol. It is
known that 1d catalyzes the
fast racemization of sec-alco-
9/3a (95:5)
[6c–d]
hols.
In conclusion, we have devel-
oped a Ru-catalyzed coupling
of allylic alcohols and alde-
hydes under mild reaction con-
ditions where the formation of
ketones (3) or other by-prod-
ucts (benzyl alcohols or a,b-un-
saturated ketones) is complete-
ly suppressed, and aldols are
obtained in up to 99% yield in
high syn/anti ratio.
6
7
8
9
25/5
50/7
25/7
65/18
10/3a (99:1)
97 (82)/82:18
92 (80)/82:18
>99 (78)/70:30
69 (59)/60:40
11/3a (>99:<1)
12/3a (>99:<1)
13/3c (>99:<1)
Experimental Section
General procedure for the cross-cou-
pling reactions: KOtBu (56 mL; 0.5m
in THF, 7 mol%) was added to a mix-
g]
1
0
35/5
14/3d (>99:<1)
>95 (79)/40:60
ture
.020 mmol,
42 mg, 0.4 mmol) in degassed toluene
of
complex
1d
(13 mg,
0
5
mol%) and Na
2
CO
3
(
1
[
a] See Experimental Section. [b] Determined by H NMR spectroscopy of the crude mixture. [c] Isolated
yield in parentheses. [d] Determined by H NMR spectroscopy. [e] 7 could not be separated from traces of 2b.
f] Only syn-8 was isolated. [g] At RT, 14 was obtained in 75% yield, syn/anti 80:20 after 24 h.
(1 mL) under a nitrogen atmosphere.
The mixture was stirred for 3 min
before a solution of the allylic alcohol
alcohol (2, 0.4 mmol) and aldehyde (4,
1
[
0
.6 mmol) in degassed toluene (1 mL)
was added via syringe. The mixture
was then stirred at the appropriate
temperature (see Table 1). Aliquots
troscopy (see Supporting Information). Next, b-hydride
elimination forms intermediate 17 where the a,b-unsaturat-
ed ketone stays coordinated to the Ru–hydride. This is sup-
ported by the fact that the aldehyde (4) is not reduced to
benzyl alcohol by Ru–H species. 1,4-Addition of the hydride
yields Ru–enolate 18. We believe that cis-enolates are pro-
duced since 2-cyclohexen-1-ol failed to isomerize to cyclo-
hexanone or to yield any aldol in the presence of 4a
1
were taken and analyzed by H NMR spectroscopy. When the analysis
showed that no allylic alcohol (2) was left, the products were isolated by
column chromatography (pentane/AcOEt 100:1!10:1), usually as an in-
separable mixture of syn and anti diastereomers.
(
Scheme 5). Instead, traces of 2-cyclohexenone and p-chlo-
Acknowledgements
ACHTUNGTRENNUNGr obenzyl alcohol were detected. Thus, b-hydride elimination
occurs, but subsequent 1,4-addition of the hydride to the
double bond does not take place. This experiment suggests
Financial support from the wedish Research Council (Vetenskapsrꢄdet)
and the Berzelius Center Exselent is gratefully acknowledged.
Chem. Eur. J. 2008, 14, 10547 – 10550
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10549

B. Martꢂn-Matute et al.
[
[
2] a) M. Wang, C.-J. Li, Tetrahedron Lett. 2002, 43, 3589–3591; b) M.
Wang, X.-F. Yang, C.-J. Li, Eur. J. Org. Chem. 2003, 998–1003;
c) X.-F. Yang, M. Wang, R. S. Varma, C.-J. Li, Org. Lett. 2003, 5,
6
57–660; d) X.-F. Yang, M. Wang, R. S. Varma, C.-J. Li, J. Mol.
Catal. A 2004, 214, 147–154.
3] a) M. Uma, M. Davies, C. Crꢀvisy, R. Grꢀe, Tetrahedron Lett. 2001,
4
2, 3069–3172; b) C. Crꢀvisy, M. Wietrich, V. Le Boulaire, M. Uma,
R. Grꢀe, Tetrahedron Lett. 2001, 42, 395–398; c) V. Branchadell, C.
Crꢀvisy, R. Grꢀe, Chem. Eur. J. 2004, 10, 5795–5803; d) D. Cuperly,
C. Crꢀvisy, R. Grꢀe, Synlett 2004, 93–96; e) D. Cuperly, J. Petrignet,
C. Crꢀvisy, R. Grꢀe, Chem. Eur. J. 2006, 12, 3261–3274.
[
[
4] For a related coupling using Ru–hydrides, see: I. Matsuda, M. Shiba-
ta, S. Sato, J. Organomet. Chem. 1988, 340, C5–C7.
5] a) R. L. Chowdhury, J.-E. Bꢁckvall, J. Chem. Soc. Chem. Commun.
1
991, 1063–1064; b) A. Attila, G. Csjernyik, K. J. Szabꢅ, J.-E. Bꢁck-
vall, Chem. Commun. 1999, 351–352.
[
6] See, for example: a) J. H. Choi, Y. H. Kim, S. H. Nam, S. T. Shin,
M.-J. Kim, J. Park, Angew. Chem. 2002, 114, 2479–2482; Angew.
Chem. Int. Ed. 2002, 41, 2373–2376; Angew. Chem. Int. Ed. 2002, 41,
Scheme 4. Proposed catalytic cycle.
2
373–2376; b) V. Rautenstrauch, X. Hoang-Cong, R. Churlaud, K.
Abdur-Rashid, R. H. Morris, Chem. Eur. J. 2003, 9, 4954–4967; c) G.
Csjernyik, K. Bogar, J.-E. Bꢁckvall, Tetrahedron Lett. 2004, 45,
6
799–6802; d) B. Martꢂn-Matute, M. Edin, K. Bogꢆr, J.-E. Bꢁckvall,
Angew. Chem. 2004, 116, 6697–6701; Angew. Chem. Int. Ed. 2004,
3, 6535–6539; e) B. Martꢂn-Matute, M. Edin, K. Bogꢆr, F. B.
4
Kaynak, J.-E. Bꢁckvall, J. Am. Chem. Soc. 2005, 127, 8817–8825;
f) P. Crochet, M. A. Fernꢆndez-Zfflmel, J. Gimeno, M. Scheele, Orga-
nometallics 2006, 25, 4846–4849.
[
[
7] B. Martꢂn-Matute, K. Bogꢆr, M. Edin, F. B. Kaynak, J.-E. Bꢁckvall,
Chem. Eur. J. 2005, 11, 5832–5842.
8] Compound 14b could have been formed via isomerization of alco-
hol 2d to the corresponding ketone (3d) followed by aldol reaction
catalyzed by traces of base.
[
9] Ru–alkoxide complexes: a) C. J. A. Daley, S. H. Bergens, J. Am.
Chem. Soc. 2002, 124, 3680–3691; b) S. D. Loren, B. K. Campion,
R. H. Heyn, T. D. Tilley, B. E. Bursten, K. W. Luth, J. Am. Chem.
Soc. 1989, 111, 4712–4718; c) Re–alkoxide complex: I. Saura-
Llamas, J. A. Gladysz, J. Am. Chem. Soc. 1992, 114, 2136–2144;
d) Ru–alcohol complex: C. P. Casey, T. E. Vos, G. A. Bikzhanova,
Organometallics 2003, 22, 901–903; e) W–alcohol and Mo–alcohol
complexes: R. M. Bullock, Chem. Eur. J. 2004, 10, 2366–2374 and
references therein.
Scheme 5. Unsuccessful coupling of 2-cyclohexen-1-ol with 4a.
Keywords: aldol reaction
coupling · isomerization · ruthenium
·
allylic compounds
·
CꢀC
[
10] B. M. Trost, R. J. Kulawiec, J. Am. Chem. Soc. 1993, 115, 2027–
036.
11] H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79, 1920–
923.
[
1] a) Ruthenium in Organic Synthesis (Ed.: S.-I. Murahashi), Wiley-
VCH, Weinheim, 2004; b) R. Uma, C. Crꢀvisy, R. Grꢀe, Chem. Rev.
2
[
2
003, 103, 27–51; c) R. C. van der Drift, E. Bouwman, E. Drent, J.
1
Organomet. Chem. 2002, 650, 1–24; d) S. Gladiali, E. Alberico, in
Transition Metals for Organic Synthesis, Vol. 2 (Eds.: M. Beller C.
Bolm), 2nd ed., Wiley-VCH, Weinheim, 2004, pp. 145–166.
Received: August 14, 2008
Published online: October 22, 2008
10550
www.chemeurj.org
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 10547 – 10550