A tandem allylic alcohol isomerization-aldol condensation catalyzed by Rh and Ru complexes

Article abstract of DOI:10.1016/S0040-4039(01)00386-0

Allylic alcohols react with aldehydes, in an atom economy aldol-type reaction, in the presence of catalytic amounts of various rhodium and ruthenium complexes. This reaction occurs with total regiocontrol under mild conditions, but varying amounts of ketone derived from the competing isomerization reaction are also observed.

Full text of DOI:10.1016/S0040-4039(01)00386-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 3069–3072
A tandem allylic alcohol isomerization–aldol condensation
catalyzed by Rh and Ru complexes
Ramalinga Uma, Maxwell Davies, Christophe Cre´visy and Rene´ Gre´e*
ENSCR, Laboratoire de Synthe`ses et Activations de Biomole´cules, associe´ au CNRS, Avenue du Ge´neral Leclerc,
35700 Rennes Beaulieu, France
Received 13 February 2001; revised 14 February 2001; accepted 6 March 2001
Abstract—Allylic alcohols react with aldehydes, in an atom economy aldol-type reaction, in the presence of catalytic amounts of
various rhodium and ruthenium complexes. This reaction occurs with total regiocontrol under mild conditions, but varying
amounts of ketone derived from the competing isomerization reaction are also observed. © 2001 Elsevier Science Ltd. All rights
reserved.
Developing the synthetic potential of enol/enolate inter-
mediates derived from allylic alcohols employing transi-
tion metal catalysts is a promising, yet relatively less
explored approach for CꢀC bond formation.1 Bosnich
et al. have shown that Rh catalyzed isomerization of
allylic alcohols results in the formation of enols which
could be detected by NMR and also trapped in an
ene-type reaction.1 In a different approach, Motherwell
et al. have established that allylic lithium alkoxides can
be isomerized to lithium enolates in the presence of
transition metal catalysts (Rh and Ni) and can be
further engaged in aldol reactions.2 However, this
attractive strategy has some drawbacks due to the
stoichiometric amount of base required to generate the
lithium alcoholates. Therefore, a major improvement
would be to perform such a reaction directly from
allylic alcohols under neutral conditions. We have
recently demonstrated for the first time that a variety of
allylic alcohols of the general structure 1 can be isomer-
ized and trapped in situ by aldehydes 2 to furnish aldol
products 3 in the presence of catalytic amounts of
Fe(CO)5 upon irradiation (Scheme 1).3 Nevertheless,
small amounts of regioisomers 4 as well as ketones 5
were also encountered.
While this novel aldol-type reaction is the first step
towards a complete atom economy process under neu-
tral conditions, the scope of the reaction can be further
improved in terms of regiocontrol and circumventing
the need for photochemical activation. Further, as
Fe(CO)5 is not amenable to extension for asymmetric
catalysis, we explored the possibility of employing other
transition metal complexes for this purpose.
We have screened many different catalytic systems and
present here the preliminary results demonstrating that
several Ru and Rh based catalysts (6–11) are able to
perform the tandem isomerization–aldolization reac-
tion.
While various rhodium and ruthenium catalysts are
known to isomerize allylic alcohols to saturated
R
R1
OH
R
OH
R1
R1
R1
RCHO
2
R2
+
+
Fe(CO)5 cat.
R2
O
R2
O
R2
O
OH
hν, Pentane
3
1
4
5
R = H, alkyl, aryl; R1, R2 = H, CH3
Scheme 1.
Keywords: allylic alcohols; aldols; catalysts; rhodium; ruthenium.
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00386-0

3070
R. Uma et al. / Tetrahedron Letters 42 (2001) 3069–3072
ketones,4 Rh(PPh3)3Cl5 and Ru(PPh3)3Cl2 are them-
selves inactive. However, we have shown that after their
reaction with an organolithium agent (e.g. n-BuLi), an
active catalyst was generated which could bring about
this transformation.6 We were further pleased to find
that in the presence of an aldehyde, such activated
rhodium and ruthenium catalysts could readily effect an
aldol-type reaction on an allylic alcohol (Scheme 2).
have obtained a comparable result (entry 2). According
to the literature,9 the reaction of Wilkinson’s catalyst
with an activator that does not contain any b-hydro-
gen, like MeLi, PhLi, MeMgBr or PhMgBr, affords
(PPh3)3RhPh 7 or (PPh3)3RhMe 8 after the displace-
ment of chlorine. These two catalysts were prepared
from Grignard and organolithium reagents. The result-
ing catalysts were engaged in an aldol-type reaction
with benzaldehyde and octen-3-ol (entries 3–6) to fur-
nish adducts 13 in fair yields (56–68%) and a low
diastereoselectivity (syn/anti 0.6:1). Nevertheless, it has
to be noticed that the reaction rate is far slower when
the catalyst is prepared from Grignard reagent.
We first tried (PPh3)3RhH 6 generated under the condi-
tions reported by Boons.5 Refluxing a THF solution of
allylic alcohol 12 with benzaldehyde in the presence of
catalytic amounts (5 mol%) of catalyst 6 (Table 1, entry
1) afforded diastereoisomeric mixtures7 of aldol 13
(54%) along with ketone 14 (29%). To verify the struc-
ture of the catalytic species involved in the process we
prepared 6 following an alternative procedure8 and
Though known catalysts 6, 7 and 8 have been isolated
in a pure form,9b,9c,10 it proved to be far more conve-
nient to prepare and use them in situ since they are very
6-11, 5mol%
Ph
+
PhCHO
+
OH
O
OH
13
O
THF, reflux
14
12
6 : Rh(PPh3)3H
7 : Rh(PPh3)3Ph
9 : Ru(PPh3)3HCl
10 : Ru(PPh3)3PhCl
11 : Ru(PPh3)3MeCl
8 : Rh(PPh3)3Me
Scheme 2.
Table 1.
Entry
1
Catalyst
Catalyst preparation Additive
Time
Conv. (%)a
100
100
100
80
13 (Yield %)
13 (syn:anti)
40:60
34:66
37:63
41:59
39:61
42:58
43:57
42:58
44:56
55:45
58:42
55:45
51:49
70:30
14 (Yield %)
6
[Rh(PPh3)3Cl–
1 equiv. nBuLi]
[Rh(PPh3)3Cl–
1 equiv. LDA]
[Rh(PPh3)3Cl–
1 equiv. PhLi]
[Rh(PPh3)3Cl–
1 equiv. PhMgBr]
[Rh(PPh3)3Cl–
1 equiv. MeLi]
Rh(PPh3)3Cl–
1 equiv. MeMgBr]
25 min
90 min
35 min
8 h
54
62
29
29
27
19
37
26
23
21
27
17
34
14
13
50
2
6
3
7
63
b
4
7
5
8
35 min
10 h
100
94
56
b
6
8
7
8
[Rh(PPh3)3Cl–
1 equiv. MeLi]
[Rh(PPh3)3Cl–
1 equiv. nBuLi]
[Rh(PPh3)3Cl–
1 equiv. nBuLi]
[Ru(PPh3)3Cl2–
1 equiv. nBuLi]
[Ru(PPh3)3Cl2–
1 equiv. LDA]
[Ru(PPh3)3Cl2–
1 equiv. PhLi]
[Ru(PPh3)3Cl2–
1 equiv. MeLi]
[Ru(PPh3)3Cl2–
1 equiv. K2CO3]
MgBr2
LiBr
36 h
100
25
69
b
8
6
7 h
b
9
6
LiBF4
7 h
72
10
11
12
13
14
9
30 min
90 min
40 min
40 min
90 min
100
100
100
100
100
62
51
75
73
45
9
10
11
a Conversions are based on transformation of compound 12 to aldols 13 and have been calculated from 1H NMR of the crude mixture.
b The reaction could not be completed and an inseparable mixture of 12 and 13 was isolated. So, isolated yields for 13 could not be obtained and
are not given.

R. Uma et al. / Tetrahedron Letters 42 (2001) 3069–3072
3071
Catalyst 9, 5mol%
R
+
RCHO
+
OH
O
THF, reflux
OH
O
12
16-20
15 R = H, aryl
14
Scheme 3.
Table 2.
Entries
Aldehyde
Time (min)
15 (Yield %)
15 (syn:anti)
14 (Yield %)
1
2
3
4
5
16 p-ClPhCHO
35
120
30
45
30
72
46
31
27
26
56:44
59:41
57:43
60:40
–
9
31
57
55
52
17 p-NO2PhCHOa
18 p-OCH3PhCHO
19 5-Acetoxymethyl-2-furaldehyde
20 HCHO (anhyd.)
a 10 mol% of the catalyst was used.
R1
R1
R1
Preliminary screening with ruthenium catalyst 9 indi-
cated that the reaction could be extended to other
aldehydes 16–20 and in each case aldol products were
obtained in moderate yields (Scheme 3, Table 2). In the
case of aldehyde 17 the reaction was slow and did not
readily go to completion.
M
OH
O
OH
M
M
R1
R1
R2
R1
R2CHO
O
OM
OH
O
The mechanism of the reaction will require further
studies. However, by analogy with iron carbonyl-medi-
ated reactions, it is possible to consider a process via
p-oxa allyl intermediates and/or h1-oxygen bound
rhodium/ruthenium enolates, which can either be
trapped by an aldehyde or just undergo a tautomeriza-
tion to give the saturated ketone (Scheme 4). There is
ample literature precedence for p-oxa-h1 equilibria, the
later depending on the ligands and the metal.15 Further,
h1-oxygen bound rhodium enolates have already been
prepared by another route and have also been engaged
in catalytic aldol reactions.16
M: (PPh3)3RhR, (PPh3)3RuRCl; R = H, Ph, Me
Scheme 4.
sensitive towards oxygen and moisture. This leads to
the unavoidable presence of magnesium or lithium salts
in the reaction mixture. To evidence any eventual effect
of these salts on the aldol reaction we have conducted
some experiments in the presence of added salts (entries
7–9). The addition of 5% MgBr2 to a catalyst solution
prepared from Wilkinson’s catalyst and an organo-
lithium reagent has an influence on the kinetics (the
rate of the reaction being slower) and the conversion11
(entry 7). The addition of 5% LiBr has a critical effect
on the reactivity as the conversion is very low (only
25% after 7 hours) in this case (entry 8). It is notewor-
thy that the use of a less coordinating counter anion
In conclusion, though preliminary, our studies have
established that allylic alcohol 12 reacts with aldehydes
in the presence of rhodium and ruthenium catalysts
6–11 to give aldol adducts with complete atom econ-
omy. Though the stereoselectivity and the scope of this
reaction need to be improved, it is noteworthy that with
this family of catalysts no regioisomeric aldol could be
detected in the crude reaction mixture as opposed to the
iron carbonyl mediated reaction.3 Such catalysts are easy
to prepare and the reaction occurs under mild condi-
tions. Our route is distinctly different from the one
reported by Motherwell as no lithium alcoholate is
involved in the process.2 Extension to other systems as
well as asymmetric synthesis are under active study in
the laboratory.
(BF4 ) instead of Br− increases the conversion (entry 9).
−
We also explored the extension of this reaction to
ruthenium catalysts. By analogy, we propose that the
chlorine of (PPh3)3RuCl2 is replaced by a H, Ph or Me
group, when the organolithium reagent is n-BuLi, PhLi
or MeLi, respectively, to give new transition metal
complexes 912–11 (entries 10–13) of the general formula
(PPh3)3RuRCl, (R=H, Ph, CH3). In the presence of
these catalysts, the aldol products 13 were isolated in
51–75% yield with modest but reverse diastereoselectiv-
ity (syn/anti 1.2:1).13 When (PPh3)3RuCl2 was activated
according to the conditions described by Ba¨ckvall,4a
a
Acknowledgements
decrease in yield of the aldol along with a slight
increase of the diastereoselectivity was observed (entry
14). In each case the ketone 14 was also isolated (13–50
yield%).14
We thank the CNRS for the award of a research
associate position to R.U.

3072
R. Uma et al. / Tetrahedron Letters 42 (2001) 3069–3072
densation probably due to the large excess of triethyl-
References
amine used in this preparation; (b) Hudson, B.; Webster,
D. E.; Wells, P. B. J. Chem. Soc., Dalton 1972, 1204–
1207.
1. Bergens, S. H.; Bosnich, B. J. Am. Chem. Soc. 1991, 113,
958–967.
2. (a) Gazzard, L. J.; Motherwell, W. B.; Sandham, D. A. J.
Chem. Soc., Perkin Trans. 1 1999, 979–993; (b) Mother-
well, W. B.; Sandham, D. A. Tetrahedron Lett. 1992, 33,
6187–6190.
13. Representative experimental procedure. To a stirred sus-
pension of (PPh3)3RuCl2 (48 mg, 0.05 mmol, 5 mol%) in
5–10 ml of dry THF, under an atmosphere of nitrogen,
was added one equivalent (32 mL, 0.05 mmol) of n-BuLi
(1.6 M solution in hexane) at room temperature. A wine
red solution was quickly formed and this mixture was
allowed to react for about 10–15 min before adding a
degassed mixture of the allylic alcohol 14 (154 mL, 1
mmol) and freshly distilled benzaldehyde (122 mL, 1.2
mmol) in dry THF (5–7 mL). The reaction mixture was
heated at reflux and the progress of the reaction was
monitored by thin-layer chromatography. Upon con-
sumption of the starting material, the reaction mixture
was cooled to ambient temperature and THF removed
under reduced pressure. The crude reaction mixture was
purified by column chromatography (silica gel, pen-
tane:ether, 9:1) to initially afford ketone 16 (22 mg, 17%).
Upon further elution (pentane:ether, 4:6), a diastereo-
meric mixture of aldol products 15 (145 mg, 62%) was
obtained. All the diastereoisomeric mixtures of aldol
products gave satisfactory 1H and 13C NMR spectra
and satisfactory high resolution mass measurements.
14. In some reactions we have observed that longer reaction
times can lead to increased amounts of ketone.
3. Cre´visy, C.; Wietrich, M.; Le Boulaire, V.; Uma, R.;
Gre´e, R. Tetrahedron Lett. 2001, 42, 395–398.
4. (a) Ba¨ckvall, J. E.; Andreasson, U. Tetrahedron Lett.
1993, 34, 5459–5462; (b) Trost, B. M.; Kulawiec, R. J. J.
Am. Chem. Soc. 1993, 115, 2027–2036 and references
cited therein; (c) Marko´, I. E.; Gautier, A.; Tsukazaki,
M.; Llobet, A.; Plantalech-Miz, E.; Urch, C. J.; Brown,
S. M. Angew. Chem., Int. Ed. 1999, 38, 1960–1962; (d)
Slugorc, C.; Ru¨ba, E.; Schmid, R.; Kircher, K.
Organometallics 1999, 18, 4230–4233.
5. Boons, G.-J.; Burton, A.; Isles, S. Chem. Commun. 1996,
141–142.
6. Uma, R.; Davies, M. K; Cre´visy, C.; Gre´e, R., submitted.
7. Diastereomeric ratios were obtained by analyses of 1H
NMR spectra of the crude reaction mixtures.
8. Diamond, S. E.; Mares, F. J. Organomet. Chem. 1977,
142, C55–C57.
9. (a) Keim, W. J. Organomet. Chem. 1967, 8, P25–P26; (b)
Michman, M.; Balog, M. J. Organomet. Chem. 1971, 31,
395–402; (c) Keim, W. J. Organomet. Chem. 1968, 14,
179–184.
15. See for instance: (a) Rasley, B. T.; Rapta, M.; Kulawiec,
R. J. Organometallics 1996, 15, 2852–2854. (b) Slough, G.
A.; Hayashi, R.; Ashbaugh, J. R.; Shamblin, S. L.;
Aukamp, A. M. Organometallics 1994, 13, 890–898. (c)
Hartwig, J. F.; Bergman, R. G.; Andersen, R. A.
Organometallics 1991, 10, 3326–3344 and references cited
therein.
10. Strauss, S. H.; Diamond, S. E.; Mares, F.; Shriver, D. F.
Inorg. Chem. 1978, 17, 3064–3068.
11. In the presence of 5% MgBr2, 61% conversion was
obtained after 8 hours with catalyst 6 and 72% conver-
sion after 10 hours with 7.
12. (a) Catalyst 9 can be prepared by the procedure described
by Wells,12b nevertheless the catalytic solution prepared
in this way did not perform the isomerization–aldol con-
16. Slough, G. A.; Bergman, R. G.; Heathcock, C. H. J. Am.
Chem. Soc. 1989, 111, 938–949.
.
.