Rhodium-Catalysed coupling of allylic, homoallylic, and bishomoallylic alcohols with aldehydes and n-tosylimines: Insights into the mechanism

Article abstract of DOI:10.1002/adsc.200900448

The isomerisation of alkenols followed by reaction with aldehydes or N-tosylimines catalysed by rhodium complexes has been studied. The catalytically active rhodium complex is formed in situ from commercially available (cyclooctadiene)rhodium(I) chloride dimer [Rh(COD)Cl]₂. The tandem process affords aldol and Mannich-type products in excellent yields. The key to the success of the coupling reaction is the activation of the catalysts by reaction with postassium tert-butoxide (t-BuOK), which promotes a catalytic cycle via alkoxides rather than rhodium hydrides. This mechanism minimises the formation of unwanted by-products. The mechanism has been studied by ¹H NMR spectroscopy and deuterium labelling experiments.

Full text of DOI:10.1002/adsc.200900448

FULL PAPERS
DOI: 10.1002/adsc.200900448
Rhodium-Catalysed Coupling of Allylic, Homoallylic, and
Bishomoallylic Alcohols with Aldehydes and N-Tosylimines:
Insights into the Mechanism
a
a,
Nanna Ahlsten and Belꢀn Martꢁn-Matute *
a
Department of Organic Chemistry, Arrhenius Laboratory, Stockholm University, 106 91 Stockholm, Sweden
Fax : (+46)-8-154-908; e-mail: belen@organ.su.se
Received: June 29, 2009; Revised: August 25, 2009; Published online: October 21, 2009
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200900448.
Abstract: The isomerisation of alkenols followed by postassium tert-butoxide (t-BuOK), which promotes
reaction with aldehydes or N-tosylimines catalysed a catalytic cycle via alkoxides rather than rhodium
by rhodium complexes has been studied. The catalyt- hydrides. This mechanism minimises the formation
ically active rhodium complex is formed in situ from of unwanted by-products. The mechanism has been
1
commercially available (cyclooctadiene)rhodium(I) studied by H NMR spectroscopy and deuterium la-
chloride dimer [Rh ACTHUNGTERNN(UNG COD)Cl] . The tandem process belling experiments.
2
affords aldol and Mannich-type products in excellent
yields. The key to the success of the coupling reac- Keywords: aldehydes; allylic alcohols; isomerization;
tion is the activation of the catalysts by reaction with rhodium; N-tosylmines
Introduction
Enolates are undoubtedly very important intermedi-
ates in synthetic organic chemistry for the formation
[1]
of CꢀC bonds. Despite the numerous applications
where enolates are key intermediates, the develop-
ment of new strategies to generate enolates in a regio-
selective manner is rare. The most common way to
form enolates is by treatment of carbonyl compounds
with base. Although regioselective deprotonation of
unsymmetrical ketones may be achieved by using ki-
netically or thermodynamically controlled reaction
conditions, in those cases where both groups at the a- Scheme 1. Transition metal-catalysed regioselective aldol
formation via (a) reductive aldol from enones, and (b)
tandem isomerisation/aldol reaction from allylic alcohols.
position of the ketone functionality have similar de-
grees of substitution, a mixture of enolates is usually
produced. In these instances, other approaches may
be used. For example, regioselective enolate forma-
tion can be achieved by stoichiometric reduction of very attractive method to regioselectivily produce
[
2]
enones using lithium and liquid ammonia. Much enolates is the isomerisation of allylic alcohols (1) cat-
[8]
more attractive is the in situ catalytic formation of alysed by transition metal complexes [Scheme 1(b)].
enolates mediated by transition metal complexes In this case, a stoichiometric reductant is not required
[
3,4]
using stoichiometric reductants, such as silanes,
hy- since the enolate is formed from the starting allylic al-
[5]
[6]
[7]
drogen, diethylzinc or aldehydes [Scheme 1(a)]. cohol in an internal redox process.
The enolate intermediates can be trapped by electro-
The two methods (Scheme 1) offer several advan-
philes present in the reaction mixture. In the case of tages over the classical approaches in terms of regio-
[
9]
using aldehydes, aldol products are produced through selectivity and atom-economy. While great achieve-
[
3–7]
an overall reductive aldol tranformation.
Another ments have been reached in the reductive aldol ap-
Adv. Synth. Catal. 2009, 351, 2657 – 2666
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2657

Nanna Ahlsten and Belꢀn Martꢁn-Matute
FULL PAPERS
Despite the excellent results obtained with Ru com-
plex 4, the current limitations of the system are that 4
is costly to make, and modification of the catalyst
structure, for example, changing the ligands, is not
straightforward. For these reasons, we planned to de-
velop a tandem transformation using transition metal
complexes which can be prepared in situ from com-
mercially available reagents. A key aspect to the suc-
cess of the transformation shown in Scheme 2 is the
activation of 4 using t-BuOK. As a result, a ruthenium
tert-butoxide is produced, which is the active catalytic
Scheme 2. Coupling of allylic alcohols and aldehydes cata-
5
lysed by [h -(Ph Cp)Ru(CO) Cl] (4).
5
2
proach where aldols can be produced with high species. We first tried to use commercially available
[
5,3d,h,j–l]
diastereo- and enantioselectivity,
isomerisation/aldolisation of
Scheme 1(b)] still remains in its infancy.
A major problem encountered in the transition hydes. However, these catalytic systems afforded
the tandem Ru chloride complexes, such as Ru
ACHTUNGTRNENU(G PPh ) Cl or
3 3 2
allylic10,11,12a] lcohols [Ru(p-cymene)Cl ] /PPh , in combination with t-
AHCTUNGTRENNUNG
2
2
3
[
[
BuOK in the coupling of allylic alcohols with alde-
metal-catalysed coupling between allylic alcohols (1) aldol products with very low efficiency, and the
and aldehydes (2) to produce aldols has been the low amount of ketone by-product (3) produced made this
[12]
efficiency of the reaction due to the formation of un- approach of no practical use in a total synthesis.
wanted ketone by-products (3 in Scheme 2) formed The Rh-catalysed reaction of allylic alcohols with al-
[
11a]
from the corresponding allylic alcohols in an internal dehydes was first reported by Gree et al.
redox process. Li et al. used Ru(PPh ) Cl in toluene/ total regioselectivity was obtained, the coupling pro-
H O mixtures at 1008C to obtain aldols in moderate ceeded in moderate yield and isomerisation of the al-
Although
AHCTUNGTRENNUNG
3
3
2
2
[10a]
yields (27–72%).
alcohols, such as a-vinylbenzyl alcohol (1a),
piophenone (3a) became the major product. The yield 19–37%. They used Rh
of the aldol products dramatically increased in the was generated in situ from Wilkinsonꢃs catalyst (6)
When they used aromatic allylic lylic alcohol (1-octen-3-ol) to the corresponding
[
10b]
pro- ketone (3-octanone) occurred in yields in the range of
(PPh ) H as the catalyst, which
AHCTUNGTRENNUNG
3
3
[
10b]
presence of In
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(OAc) .
They also found that in an upon treatment with n-BuLi. We envisioned that the
3
ionic liquid, the reactions proceeded well at activation of commercially available Rh chloride com-
[10c–d]
908C.
Grꢀe et al. used Ru ACHTNUGTNERNU(G PPh ) HCl to produce plexes could also be performed with t-BuOK, and
3 3
aldols in up to 72% yield and in short reaction times that if the precatalyst is a Rh alkoxide rather than a
[11a]
(
<2 h).
However, under such conditions, ketones 3 Rh hydride, the efficiency of the coupling reaction
were formed (5–52%). They have also used Fe and Ni could be improved. Here we report a very efficient
[11b–e]
complexes.
In the former case, small amounts of coupling of allylic alcohols and other alkenols with al-
regioisomeric aldols were produced. Very good results dehydes or N-tosylimines catalysed by commercially
were obtained with Ni complexes with Mg salts as co- available Rh chloride complexes activated by t-
catalysts; aldol products were formed in high yields BuOK. Mechanistic investigations are also presented.
together with small amounts of ketones 3 (2–15%).
During our search for a catalytic system that would
perform the isomerisation/aldol domino process with Results and Discussion
the highest possible atom economy, we found that the
5
ruthenium complex [h -(Ph C )Ru(CO) Cl] (4) cataly- First, we tested Wilkinsonꢃs complex (6) in combina-
5
5
2
ses the transformation of allylic alcohols into aldols in tion with t-BuOK as the catalyst system for the cou-
very high yields at room temperature, and the forma- pling between allylic alcohols and aldehydes
tion of unwanted ketone by-products (3) is completely (Scheme 3, Table 1). At room temperature and in tol-
[
12]
suppressed (Scheme 2).
uene, complex 6 afforded aldol 5, albeit in only 49%
Scheme 3. Rh-catalysed coupling of allylic alcohols with aldehydes.
2658
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2657 – 2666

Rhodium-Catalysed Coupling of Allylic, Homoallylic, and Bishomoallylic Alcohols
FULL PAPERS
[a]
Table 1. Screening of reaction conditions.
[b]
[b]
[b]
Entry Catalyst [5 mol% of Rh]
Phosphine [mol%]
T [8C] Solvent t [h] Conv. [%]
5 [%]
3a [%]
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
Rh
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(PPh ) Cl (6)
–
–
–
–
r.t.
50
rt
Toluene
Toluene 0.5
5
100
100
70
100
100
–
49
80
47
90
85
–
46
20
23
10
15
–
3
9
4–23
10
3
5
16
2
5
3
–
–
–
4
3
3
6
6
6
6
[Rh
7
7
7
7
7
7
7
7
7
7
7
[Rh
8
7
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
3
3
3
3
3
3
3
3
40
40
40
40
40
40
40
40
40
PPh (5)
–
3
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(COD)Cl] (7)
2
dppe (5)
dppp (5)
dppf (5)
(ꢁ)-BINAP (5)
9
21
6
12
77–96
60
97
95
84
52
14
7
–
–
–
96
100
70
100
100
100
54
19
10
–
0
1
2
3
4
5
6
7
8
9
0
PPh (8)
2.5
3
3
PPh (10)
P(p-OMeC H ) (7.5) 40
P
P
PMePh (7.5)
P
3
2.5
2.5
1.5
1.5
3
3
3
2
6
4 3
A
H
U
G
R
N
U
G
40
40
40
40
40
40
40
6
4 3
A
H
U
G
R
N
U
G
3
2
A
H
N
R
N
N
(t-Bu) (7.5)
3
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(MeCN)
A
T
N
T
E
N
N
(COD)] BF (8) PPh (7.5)
<3
<80
100
2
4
3
[
[
d]
e]
PPh (7.5)
PPh (8)
3
[c]
3
[
a]
Unless otherwise stated, t-BuOK (0.5M in THF, 52 mL) was added to a solution of the Rh complex (5 mol%) and the
ligand (entries 5, 7–20) in anhydrous THF or toluene (0.5 mL) (dry solvents were obtained using VAC solvent purifier
system). The mixture was stirred for 2–3 min before adding a solution of the allylic alcohol 1a (0.4 mmol) and aldehyde
2
a (0.48 mmol) in THF or toluene (1.5 mL) [allylic alcohol:Rh:t-BuOK=100:5:6.5]. The reaction mixture was introduced
in an oil bath at the temperature indicated.
Determined by H NMR spectroscopy.
[
[
[
[
b]
1
c]
d]
e]
Commercially available THF (puriss p.a., from Sigma–Aldrich). See Supporting Information.
t-BuOK was not used. A complex reaction mixture was obtained. Aldol 5 was not detected.
Average of 3 experiments.
yield, and the unwanted isomerisation by-product 3a tron-poor ones (entries 14 and 15), and other phos-
was produced in 46% (Table 1, entry 1). The results phines such as PMePh and P(t-Bu) suppressed the
AHCTUNGTRENNUNG
2
3
were improved when the temperature was increased reaction completely or considerably slowed it down
to 508C, and aldol 5 could be obtained in 80% yield (entries 16 and 17). Neither did cationic complexes
after only 30 min (Table 1, entry 2). THF (entries 3 yield any product in the presence or in the absence of
and 4) gave better results than toluene. When extra t-BuOK (entries 18 and 19, respectively). The reaction
PPh was used, similar results were obtained (entries 4 is not moisture sensitive and excellent results were
3
vs. 5). We turned our attention to the use of obtained in degassed commercially available THF,
[13]
[
Rh ACHTUNGTRENNUNG( COD)Cl] (7). In the absence of any ligand, com- (see supporting information and Table 1 footnotes).
2
plex 7 did not afford any product (entry 6). The use The results obtained were reproducible (entry 20
of bidentate phosphines such as dppe and dppp af- shows the average of 3 runs) and very similar to those
forded the product in very low yields (entries 7 and obtained when anhydrous THF was used (entry 20 vs.
8
). Dppf in combination with 7 gave aldol products in 11). The THF used in all the reactions tested negative
good yields (entry 9). However, variable yields, from for peroxides. Increasing the amount of water by run-
7% to 96%, were obtained. This variation was ascri- ning the reactions in 10:1 THF/H O mixtures resulted
7
2
bed to the uncontrolled formation of different Rh in a substantial decrease of the reaction rate and
complexes when mixing 7 with dppf [e.g., monomeric lower conversions. The use of other bases (e.g.,
Rh(I) complexes or Rh(I) coordination polymers]. NaOH) suppressed the reaction or slowed it down.
BINAP did not give better results (entry 10). We
Using the conditions in entry 20 (Table 1), a variety
were very pleased to see that PPh together with com- of allylic alcohols were coupled to different aldehydes
3
plex 7 gave excellent results. Thus, aldol 5 was pro- (Table 2). Allylic alcohol 1a was coupled to a p-
duced in 97% yield together with only 3% of 3a chloro
ACHTUNGTRENNUNGb enzaldehyde (2a), benzaldehyde (2b) and p-
(
Table 1, entry 11). Increasing the amount of PPh did fluorobenzaldehyde (2c) yielding aldols 5, 9, and 10 in
3
not have any effect (entry 11 vs. 12). Electron-rich excellent yields (Table 2, entries 1–3). When more
phosphines (entry 13) gave better results than elec- sterically hindered aldehydes were used, such as 2d,
Adv. Synth. Catal. 2009, 351, 2657 – 2666
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2659

Nanna Ahlsten and Belꢀn Martꢁn-Matute
FULL PAPERS
[a]
Table 2. Rh-catalysed coupling of alcohols with aldehydes.
[b]
[c]
[b]
[b]
Entry Alcohol
1
2
Aldol
t [h] Conv. [%]
Yield [%]
syn:anti
3
[%]
[
d]
[f]
[d]
2a
2
2
2
100
100
100
96 (92)
50:50
3a (4)
3a (4)
3a (4)
[
e]
2
3
2b
2c
96 (92)
46:54
50:50
[f]
97 (92)
85 (78)
94 (91)
81 (74)
[
g]
4
5
2d
2c
1.25 100
56:44
40:60
3a (15)
3b (6)
2
100
6
7
8
2b
2b
2b
2.5
16
18
100
92
50:50
56:44
36:63
3c (19)
3d (13)
3g (13)
[h]
79
86
73 (66)
[h]
[i]
9
2b
18
86
64
40:60
3h (22)
[
a]
Unless otherwise stated, t-BuOK (0.5M in THF, 45 mL) was added to a solution of dimeric Rh complex 7 (2.5 mol%) and
PPh (8 mol%) in THF (0.5 mL, puriss p.a., from Sigma–Aldrich). The mixture was stirred for 2–3 min before adding a
3
solution of the allylic alcohol (0.4 mmol) and aldehyde (0.48 mmol) in THF (1.5 mL) [allylic alcohol:Rh:t-BuOK:PPh =
3
1
00:5:6.5:8]. The reaction mixture was introduced in an oil bath at 408C and stirred for the time indicated.
[
[
[
[
[
[
[
[
b]
c]
d]
e]
f]
1
Determined by H NMR spectroscopy.
Determined by H NMR spectroscopy, and in parenthesis isolated yield.
Average of three experiments. For results in anhydrous THF see Table 1, entry 11.
In anhydrous THF, 100% conversion, aldol:ketone (%)=92/8.
Small amounts (5–7%) of regioisomeric aldols were detected. See Supporting Information.
In anhydrous THF, 100% conversion, aldol:ketone (%)=83/17.
1
g]
h]
i]
Could not be separated from traces of the starting alcohol. See Supporting Information.
Contains g,d-unsaturated ketone (5%).
2660
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2657 – 2666

Rhodium-Catalysed Coupling of Allylic, Homoallylic, and Bishomoallylic Alcohols
FULL PAPERS
Scheme 4. Rh-catalysed coupling of allylic alcohols with N-tosylimines.
the yields of the aldol products decreased slightly, and both the OꢀH and a-CꢀH hydrogen atoms of the al-
the yield of the unwanted ketone 3a increased lylic alcohol are transferred to the metal complex
(
Table 2, entry 4). We turned our attention to other yielding a metal dihydride and an a,b-unsaturated
[15]
allylic alcohols. Thus, alcohols 1b and 1c afforded ketone.
As a result, the hydrogens are scrambled
aldols 12 and 13, respectively, in excellent yield and lose their identity. To distinguish between these
(
Table 2, entries 5 and 6). More interesting is the use two possibilities, the coupling between deuterated al-
of more substituted allylic alcohols such as 1d, which lylic alcohol [D ]1a and benzaldehyde was performed.
1
afforded, in good yield and with total regioselectivity, Monodeuterated aldol [D ]5, with the deuterium label
1
the coupling product 14 upon reaction with benzalde- exclusively on the methyl group, was obtained
hyde (entry 7). In contrast to our previous results (Scheme 5). Thus, a mechanism via Rh dihydrides can
[
12]
with Ru complex 4, that the alcohol and the C=C be ruled out.
bond have an allylic relationship is not a requirement If metal monohydrides are involved, three general
[
8]
for the coupling to occur: the double bond can be mechanisms can be considered (Scheme 6). The first
placed further away from the alcohol, such as in 15 involves alkyl-metal intermediates [path (a) in
and 17, and they also afforded aldols 16 and 18 in Scheme 6]. In this case, the catalyst is a metal hy-
good yields (Table 2, entries 8 and 9).
dride, either isolated or generated in situ. Insertion of
The coupling gave complex reaction mixtures when the allylic alcohol double bond into the MꢀH bond,
using electron-rich aldehydes such as p-anisaldehyde followed by a-CꢀH bond cleavage leads to an enol
or 3-tolualdehyde. In these instances, a retroaldol re- and regenerates the metal hydride complex. In the
action takes place yielding a mixture of different second mechanism [path (b) in Scheme 6], an enol is
aldol products.
Having established the viability of this reaction, we by the oxidative addition of the allylic CꢀH bond,
turned our attention to the use of N-tosylimines 19a and would as such involve Rh(III) intermediates. In
and 19b as the electrophilic partner (Scheme 4, path (c), the allylic alcohol coordinates to the metal
generated via a p-allyl metal hydride complex formed
AHCTUNGTRENNUNG
[10c]
Table 3).
When we searched for optimal reaction centre forming a metal alkoxide (or a metal-alcohol
conditions, we found that the coupling with N-tosyl- complex). Following an elimination/insertion pathway,
[14]
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
imines requires longer times than that of aldehydes, the key metal-enolate intermediate is produced.
and that the reaction needs to be performed in dry We first tried to identify the catalytic species
THF to avoid the formation of aldol products arising formed when the system Rh (7)/PPh is treated with t-
3
from hydrolysis of the imines. A variety of allylic alco- BuOK. Reaction of dimeric complex 7 with PPh im-
3
hols bearing electron-withdrawing or electron-donat- mediately yields the monomeric complex 26
3
1
ing groups yielded Mannich-type products in good to (Scheme 7). The P NMR spectrum showed a doublet
1
31 103
excellent yields (Table 3, entries 1–4). Aliphatic (1e) at 30.9 ppm with a J( P, Rh) of 152 Hz. The
1
as well as homoallylic alcohols (15) were suitable sub-
H NMR spectrum was also in agreement with struc-
strates (Table 3, entries 5 and 6).
ture 26, showing the characteristic multiplets in the
aromatic region due to coordinated PPh , and at
3
higher field the peaks corresponding to coordinated
Mechanistic Investigations
cyclooctadiene (COD). When an excess of PPh was
used, the P NMR did not show any peak, probably
3
3
1
The first step in the coupling reaction is the isomeri- due to a fast exchange of free and coordinated phos-
1
sation of the allylic alcohol to an enol (enolate) inter- phine. The H NMR spectrum showed broad peaks
mediate, which in a second step reacts with the alde- corresponding to the aromatic protons of free and co-
hyde or N-tosylimine forming a new CꢀC bond. Al- ordinated PPh ligands. The resonances due to coordi-
3
ternatively, the enol (enolate) tautomerises to the cor- nated COD were identical to those observed before,
responding ketone (yielding the unwanted by-product when no excess of PPh was present. When t-BuOK
3
3
in the coupling reaction). The isomerisation of the was added to the NMR tube, complex 26 disappeared,
allylic alcohol into the enol(ate) can occur via a mon- and a new tert-butoxide complex was formed by sub-
ohydride or a dihydride mechanism. In the latter, stitution of the Cl ligand by tert-butoxide. The new
AHCTUNGTRENNUNG
Adv. Synth. Catal. 2009, 351, 2657 – 2666
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2661

Nanna Ahlsten and Belꢀn Martꢁn-Matute
FULL PAPERS
[a]
Table 3. Rh-Catalysed coupling of alcohols with N-tosylimines (Sheme 4).
[b]
[c]
[b]
[b]
Entry Alcohol
Imine Product
t [h] Conv. [%]
Yield [%]
91 (84)
syn:anti
3 [%]
1
19a
14
14
100
100
74:26
3a (9)
3a (5)
2
3
4
5
6
1a
19b
19a
19a
19b
19a
95 (85)
91 (75)
81 (73)
64 (56)
75 (72)
60:40
66:34
70:30
60:40
67:33
14
15
15
15
100
100
100
100
3b (9)
[
d]
3d (19)
3e (36)
3g (25)
[
d]
[
a]
Unless otherwise stated, t-BuOK (0.5M in THF, 45 mL) was added to a solution of dimeric Rh complex 7 (2.5 mol%) and
PPh (8 mol%) in dry THF (0.5 mL). The mixture was stirred for 2–3 min before adding a solution of the allylic alcohol
3
(
0.4 mmol) and N-tosylimine 19a or 19b (0.44 mmol) in dry THF (1.5 mL) [allylic alcohol:Rh:t-BuOK:PPh =
3
1
00:5:6.5:8]. The reaction mixture was introduced in an oil bath at 408C and stirred for the time indicated.
1
[
[
[
b]
c]
d]
Determined by H NMR spectroscopy.
Determined by H NMR spectroscopy and in parenthesis, isolated yield.
1
608C.
Scheme 5. Coupling of a deuterium-labelled allylic alcohol [D ]1a.
1
31
complex (27) showed a doublet in the P NMR spec- (see Supporting Information). This is consistent with
1
31 103
trum at 20.5 ppm with a J( P, Rh) of 162 Hz. The coordination of the chiral allylic alcohol 1a to the Rh
1
1
H NMR spectrum was also in agreement with the complex, forming intermediate 28. The H NMR
structure proposed for 27. When 1 equivalent of allyl- peaks corresponding to the allylic alkoxide were also
ic alcohol 1a was added to the NMR tube, complex in agreement with structure 28, and differed from
2
7 disappeared, and a new complex was formed. This those of allylic alcohol 1a.
new complex showed splitting of some of the signals Although the NMR studies (Scheme 7) indicate
corresponding to the olefinic hydrogens of the COD that the alkoxide mechanism [Scheme 6 (c)] may be
ligand, which indicates that they are diastereotopic involved, 28 could undergo b-hydride elimination pro-
2662
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2657 – 2666

Rhodium-Catalysed Coupling of Allylic, Homoallylic, and Bishomoallylic Alcohols
FULL PAPERS
Scheme 6. Simplified mechanism of isomerisation of allylic alcohols involving metal monohydride complexes via: (a) alkyl-
metal intermediates. (b) p-allyl metal hydride intermediates. (c) metal-akoxides.
formed at 708C afforded 5 (13%) and ketone 3a
(
80%). These results should be compared with the
much higher yields of aldol products that we obtained
using a [RhꢀCl]/t-BuOK catalytic systems (Table 1,
entries 4, 11 and 20). Thus, it is unlikely that RhꢀH
complexes act as the catalysts coupling directly allylic
alcohols with aldehydes under our catalytic reaction
conditions (Table 2 and Table 3).
The isomerisation of acyclic allylic alcohols into ke-
tones occurs in excellent yields when Rh hydride
[16]
complexes are used, as observed by us and others.
However, when Rh hydrides are used in the isomeri-
sation of cyclic allylic alcohols, such as 2-cyclohexen-
1-ol (1f), the corresponding ketone (cyclohexanone,
Scheme 7. In situ NMR studies.
[
16]
3
f) is obtained in very low yield.
A plausible ex-
planation is that the isomerisation is assisted by coor-
ducing a Rh hydride, which in turn could be the dination of the Rh hydride complex to the alcohol
actual catalytically active species coupling allylic alco- functionality, and in this way, the OH group directs
hols and aldehydes. To study this possibility, we tried the insertion of the double bond into the RhꢀH bond.
to generate a RhꢀH complex upon treatment of a Rh The intermediate generated cannot undergo syn b-hy-
(
7)/PPh mixture with n-BuLi. This catalytic system dride elimination to produce the enol (Scheme 9).
3
gave very low yields of aldol 5 after 16 h. However,
Under our reaction conditions, i.e., Rh (7)/PPh /t-
3
we were not sure that a RhꢀH complex was actually BuOK, isomerisation of cyclic allylic alcohol 1f takes
being produced from Rh complex 7. Hence, we per- place quantitatively (Scheme 10). This result also indi-
formed a similar experiment using the well known cates that the mechanism of our catalytic system is
(
PPh ) RhH complex (29), produced from Wilkinsonꢃs different from that followed when [RhꢀH] complexes
3
3
catalyst upon treatment with n-BuLi (Scheme 8). The are used as the catalyst precursor.
coupling of allylic alcohol 1a with p-chlorobenzalde-
Supported by the mechatisnic studies shown above,
hyde (2a) catalysed by (PPh ) RhH (29) in THF at we propose a mechanism via Rh-alkoxide intermedi-
3
3
408C afforded only 12% of aldol 5 together with 28% ates, as shown in Scheme 11. The rhodium tert-butox-
of ketone 3a after>1 day. The same reaction per- ide complex (27) can undergo alkoxide exchange with
Scheme 8. RhꢀH as catalyst in the coupling of allylic alcohols and aldehydes.
Adv. Synth. Catal. 2009, 351, 2657 – 2666
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2663

Nanna Ahlsten and Belꢀn Martꢁn-Matute
FULL PAPERS
Scheme 9. RhꢀH as catalyst in the isomerisation of cyclic allylic alcohol 1f.
allylic alcohol 1a to give the new alkoxide 28, as de- intermediate is impossible to adopt. In the absence of
1
tected by H NMR spectroscopy (see above and Sup- aldehydes, the corresponding Rh(H)!cyclohexenone
porting Information). After b-hydride elimination, complex evolves to yield cyclohexanone (vide supra,
a,b-unsaturated ketone 29 is formed. In the case of Scheme 10). However, if an aldehyde is present in the
acyclic allylic alcohols, such as 1a, a fast 1,4-addition reaction mixture, its reduction to benzylic alcohol
of the hydride to the s-cis conformation of the a,b-un- mediated by an [RhꢀH] species becomes a competing
[
4,5,17]
saturated ketone yields a Rh-enolate (30).
This is pathway, as it has been confirmed (see Supporting In-
supported by the fact that an added aldehyde (2) is formation). This experiment suggests that an s-cis
not reduced to benzyl alcohol by RhꢀH species (the conformation of the unsaturated ketone is required
1
H NMR spectra of the crude mixtures of the reac- for a fast 1,4-hydride addition producing Rh-enolate
tions shown in Table 2 showed only traces of benzylic intermediates (30). Enolate 30 then reacts with the al-
alcohols). On the other hand, for the cyclic allylic al- dehyde present in the reaction media affording the
cohol 1f, an s-cis conformation of the cyclohexenone aldol product after protonation of alkoxide 31 by an
incoming molecule of allylic alcohol.
For non-allylic alcohols such as 15 and 17, a mecha-
nism via enolates may also be operating. Thus, after
b-hydride elimination an enone and a RhꢀH are pro-
duced. The RhꢀH then coordinates to the olefin, and
after a number of insertion/b-hydride elimination
steps, a conjugated enone is produced. From here, a
Scheme 10. Isomerisation of cyclic allylic alcohol 1f by Rh mechanism similar to that shown in Scheme 11 can
(
7)/PPh /t-BuOK.
explain the formation of the products.
3
Scheme 11. Proposed mechanism for the Rh-catalysed coupling of acyclic aldehydes with allylic alcohols.
2664
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2657 – 2666

Rhodium-Catalysed Coupling of Allylic, Homoallylic, and Bishomoallylic Alcohols
FULL PAPERS
The diastereoselectivity obtained in the aldol reac- and 100 MHz, respectively, using a Bruker spectrometer.
Chemical shifts (d) are reported in ppm, using the residual
tion is rather low (see Table 2). In our previous work
[
12]
solvent peak in CDCl (d =7.26 and d =77.00) or in ben-
with ruthenium complexes,
we observed that at
3 H C
zene-d (d =7.16) as internal standards, and coupling con-
6
H
short reaction times, high diastereoselectivity was ob-
tained in favour of the syn aldol product (ca. 90:10).
As the reaction proceeded, the syn/anti ratio dropped
to ca. 80:20. We confirmed that this drop of selectivity
is due to a fast epimerisation of the aldol product cat-
alysed by the same Ru complex. In the case of Rh
complexes, at short reaction times, the diastereoselec-
tivity is already rather low (e.g., 5 was obtained with
syn/anti=51:49 at 25% conversion after 1 h). This low
diastereoselectivity can be explained by (i) a Rh-cata-
lysed epimerisation of the final products, (ii) a non-
Zimmerman–Traxler-type transition state or (iii) a
rapid interconversion of the s-allyl Rh-enolates 30.
Krische et al. have suggested that the use of a weakly
stants (J) are given in Hz. High resolution mass spectra
HR-MS) were recorded on a Bruker microTOF ESI-TOF
mass spectrometer. The products were characterised by
(
1
13
H NMR, C NMR, and HR-MS.
General Procedure for the Coupling of Allylic
Alcohols with Aldehydes
A freshly prepared solution of t-BuOK (45 mL, 0.5M in
THF, 5.5 mol%) was added to a stirred solution of
[Rh
ACHTUNGTENRUNNG( COD)Cl] (4.9 mg, 0.01 mmol) and PPh3 (8.4 mg,
2
0
.032 mmol) in degassed THF (0.5 mL, (Sigma Aldrich,
puriss. p.a.) via syringe. The mixture was stirred at room
temperature for 3 min before the addition of the aldehyde
(
0.48 mmol) and the allylic alcohol (0.4 mmol) in THF
coordinating p-acidic ligand, such as Fur P, may pro- (1.5 mL). The flask was immediately introduced in an oil
3
mote enhanced stereocontrol by “tightening” the bath at 408C, and on completion (monitored by TLC)
[
19]
Zimmerman–Traxler-type transition state. When we quenched by absorption on silica. Products were isolated by
column chromatography (pentane:EtOAc 10/1) as mixtures
of syn and anti diastereomers.
replaced PPh3 for Fur P, we were able to obtain
slightly better diastereoselectivity (e.g., 3 was ob-
3
tained with syn/anti=60:40 at 12% conversion after
General Procedure for the Coupling of Allylic
Alcohols with N-Tosylimines
1
h), although very low conversion (19%) was ob-
tained after long reaction times.
A freshly prepared solution of t-BuOK (45 mL, 0.5M in
THF, 5.5 mol%) was added to a stirred solution of
Conclusions
[Rh
A
H
U
G
E
U
G
2
(4.9 mg, 0.01 mmol) and PPh
3
(8.4 mg,
0
.032 mmol) in dry and degassed THF (0.5 mL) containing
MS 3 ꢄ (~10 mg) via syringe. The mixture was stirred at
room temperature for 3 min before the addition of a solu-
tion of the N-tosylimine (0.44 mmol) and the allylic alcohol
In conclusion, we have developed an efficient Rh-cat-
alysed coupling of allylic alcohols with aldehydes and
N-tosylimines under mild reaction conditions where
the formation of ketones (3) or other by-products
(
0.40 mmol) in THF (1.5 mL). The mixture was heated to
the appropriate temperature and stirred for the time indicat-
ed (see Table 3). Products were isolated by column chroma-
tography (pentane:EtOAc 8/1) as pure syn and anti diaste-
reomers.
(
benzyl alcohols or a,b-unsaturated ketones) is mini-
mised, and aldol or Mannich-type products are ob-
tained in very good yields (up to 97%). Homoallylic-
and bishomoallylic alcohols can also be used in the
tandem reactions. Commercially available Rh com-
plexes have been employed. Mechanistic investiga-
tions indicate that a mechanism via metal alkoxides
may be involved. We believe this is the key to the suc-
cess of the tandem process, since when preformed Rh
hydrides are used as catalysts, low yields are obtained.
Further investigations on the mechanism and future
applications of this efficient coupling are under devel-
opment.
Supporting Information
1
13
Characterisation and H/ C NMR spectra of 5, 5[D ], 9–14,
1
1
6, 18, 20–28 and spectra of the mechanistic studies are
available as Supporting Information.
Acknowledgements
Financial support from the Swedish Research Council (Veten-
skapsrꢀdet) and the Berzelius Center Exselent is gratefully
acknowledged.
Experimental Section
General Experimental Methods
All reactions were carried out under nitrogen atmosphere. References
Unless otherwise noted, THF was used as obtained from
supplier (Sigma Aldrich, puriss. p.a.). Anhydrous THF and
toluene were obtained using VAC solvent purifier system.
Flash chromatography was carried out on silica gel 60 (35–
[1] Review on the generation and utilisation of enolates: P.
Arya, H. Qin, Tetrahedron 2000, 56, 917.
[2] G. Stork, P. Rosen, N. Goldman, R. V. Coombs, J.
Tsuji, J. Am. Chem. Soc. 1965, 87, 275.
1
13
70 mm). H and C NMR spectra were recorded at 400 MHz
Adv. Synth. Catal. 2009, 351, 2657 – 2666
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2665

Nanna Ahlsten and Belꢀn Martꢁn-Matute
FULL PAPERS
[
3] For examples of Rh-catalysed reductive aldol reaction
using silanes as reducing agents, see: a) A. Revis, T. K.
Hilty, Tetrahedron Lett. 1987, 28, 4809; b) I. Matsuda,
K. Takahashi, S. Sato, Tetrahedron Lett. 1990, 31, 5331;
c) S. J. Taylor, J. P. Morken, J. Am. Chem. Soc. 1999,
M. J. Krische, Acc. Chem. Res. 2004, 37, 653; b) M.-Y.
Ngai, J.-R. Kong, M. J. Krische, J. Org. Chem. 2007, 72,
1063.
[6] P. M. Joensuu, G. J. Murray, E. A. F. Fordyce, T. Lueb-
bers, H. W. Lam, J. Am. Chem. Soc. 2008, 130, 7328.
[7] M. C. Willis, R. L. Woodward, J. Am. Chem. Soc. 2005,
1
21, 12202; d) S. J. Taylor, M. O. Duffey, J. P. Morken,
J. Am. Chem. Soc. 2000, 122, 4528; e) C.-X. Zhao, J.
Bass, J. P. Morken, Org. Lett. 2001, 3, 2839; f) D. Emia-
bata-Smith, A. McKillop, C. Mills, W. B. Motherwell,
A. J. Whitehead, A. J. Synlett 2001, 1302; g) M. Freirꢁa,
A. J. Whitehead, D. A. Tocher, W. B. Motherwell, Tet-
rahedron 2004, 60, 2673; h) H. Nishiyama, T. Shiomi, Y.
Tsuchiya, I. Matsuda, J. Am. Chem. Soc. 2005, 127,
1
27, 18012.
[
8] a) Ruthenium in Organic Synthesis; (Ed.: S.-I. Muraha-
shi), Wiley-VCH, Weinheim, 2004; b) R. Uma, C.
Crꢀvisy, R. Grꢀe, Chem. Rev. 2003, 103, 27; c) R. C. van
der Drift, E. Bouwman, E. Drent, J. Organomet. Chem.
2
002, 641-664, 1; d) S. Gladiali, E. Alberico, in: Transi-
tion Metals for Organic Synthesis, 2nd edn., Vol. 2,
Eds.: M. Beller, C. Bolm), Wiley-VCH: Weinheim,
004, pp 145–166.
9] B. M. Trost, Acc. Chem. Res. 2002, 35, 695.
6
972; i) N. O. Fuller, J. P. Morken, Synlett 2005, 1459;
j) J. I. Ito, T. Shiomi, H. Nishiyama, Adv. Synth. Catal.
006, 348, 1235; k) T. Shiomi, J.-I. Ito, Y. Yamamoto,
(
2
2
[
H. Nishiyama, Eur.J. Org. Chem. 2006, 5594; l) T.
Shiomi, H. Nishiyama, Org. Lett. 2007, 9, 1651; m) T.
Hashimoto, J-I. Ito, H. Nishiyama, Tetrahedron 2008,
[
10] a) M. Wang, C.-J. Li, Tetrahedron Lett. 2002, 43, 3589;
b) M. Wang, X.-F. Yang, C.-J. Li, Eur. J. Org. Chem.
2
003, 998; c) X.-F. Yang, M. Wang, R. S. Varma, C.-J.
6
4, 9408.
Li, Org. Lett. 2003, 5, 657; d) X.-F. Yang, M. Wang,
R. S. Varma, C.-J. Li, J. Mol. Catal. A: Chemical 2004,
2
[
4] For catalytic reductive aldol couplings using other
metals, see: Co: a) S. Isayama, T. Mukaiyama, Chem.
Lett. 1989, 2005; b) T.-G. Baik, A. L. Luis, L.-C. Wang,
M. J. Krische, J. Am. Chem. Soc. 2001, 123, 5112; c) L.-
C. Wang, H.-Y. Jang, Y. Roh, V. Lynch, A. J. Schultz,
X. Wang, M. J. Krische, J. Am. Chem. Soc. 2002, 124,
14, 147.
[
11] a) M. Uma, M. Davies, C. Crꢀvisy, R. Grꢀe, Tetrahe-
dron Lett. 2001, 42, 3069; b) C. Crꢀvisy, M. Wietrich, V.
Le Boulaire, R. Uma, R. Grꢀe, Tetrahedron Lett. 2001,
4
2, 395; c) R. Uma, N. Gouault, C. Crꢀvisy, R. Grꢀe,
9
448; d) H. W. Lam, P. M. Joensuu, G. J. Murray,
E. A. F. Fordyce, O. Prieto, T. Luebbers, Org. Lett.
006, 8, 3729; e) R. J. R. Lumby, P. M. Joensuu, H. W.
Tetrahedron Lett. 2003, 44, 6187; d) V. Branchadell, C.
Crꢀvisy, R. Grꢀe, Chem. Eur. J. 2004, 10, 5795; e) D.
Cuperly, C. Crꢀvisy, R. Grꢀe, Synlett 2004, 93; f) D. Cu-
perly, J. Petrignet, C. Crꢀvisy, R. Grꢀe, Chem. Eur. J.
2
Lam, Org. Lett. 2007, 9, 4367; Ir: f) C.-X. Zhao, M. O.
Duffey, S. J. Taylor, J. P. Morken, Org. Lett. 2001, 3,
2
006, 12, 3261.
1
829; Ru: g) T. Doi, T. Fukuyama, S. Minamino, I.
[
12] A. Bartoszewicz; M. Livendahl, B. Martꢁn-Matute,
Chem. Eur. J. 2008, 14, 10547.
[13] The water content in the THF varied between 300 and
Ryu, Synlett 2006, 3013; Pd: h) S.-I. Kiyooka, A. Shimi-
zu, S. Torii, Tetrahedron Lett. 1998, 39, 5237; Cu: i) T.
Ooi, K. Doda, D. Sakai, K. Maruoka, Tetrahedron Lett.
1
999, 40, 2133; j) H. W. Lam, P. M. Joensuu, P. M. A.
800 ppm.
Org. Lett. 2005, 7, 4225; k) H. W. Lam, G. J. Murray,
J. D. Firth, Org. Lett. 2005, 7, 5743; l) J. Deschamp, O.
Chuzel, J. Hannedouche, O. Riant, Angew. Chem. 2006,
[14] B. M. Trost, R. J. Kulawiec, J. Am. Chem. Soc. 1993,
115, 2027.
[15] RuCl2 ACHTUGNTRNEUNG( PPh ) reacts with isopropyl alcohol to produce
3 3
1
18, 1314; Angew. Chem. Int. Ed. 2006, 45, 1292; m) O.
the corresponding dihydride, see: A. Aranyos, G. Gsjer-
nyik, K. J. Szabꢆ, J.-E. Bꢇckvall, Chem. Commun. 1999,
351.
Chuzel, J. Deschamp, C. Chausteur, O. Riant, Org.
Lett. 2006, 8, 5943; n) D. Zhao, K. Oisaki, M. Kanai,
M. Shibasaki, Tetrahedron Lett. 2006, 47, 1403; o) D.
Zhao, K. Oisaki, M. Kanai, M. Shibasaki, J. Am. Chem.
Soc. 2006, 128, 14440; p) A. Welle, S. Dꢁez-Gonzꢅlez,
B. Tinant, S. P. Nolan, O. Riant, Org. Lett. 2006, 8,
[16] R. Uma, M. K. Davies, C. Crꢀvisy, R. Grꢀe, Eur. J.
Org. Chem. 2001, 3141.
[17] Synthesis of Rh enolates: a) G. A. Slough, R. G.
Berman, C. H. Heathcock, J. Am. Chem. Soc. 1989,
6
059; q) B. H. Lipshutz, B. Amorelli, J. B. Unger, J.
1
1
11, 938; b) J. Wu, R. G. Bergman, J. Am. Chem. Soc.
989, 111, 7628; c) G. A. Slough, R. Hayashi, J. R. Ash-
Am. Chem. Soc. 2008, 130, 14378; Ni: r) C. C. Chrovi-
an, J. Montgomery, Org. Lett. 2007, 9, 537; In: s) I. Shi-
bata, H. Kato, T. Ishida, M. Yasuda, A. Baba, Angew.
Chem. 2004, 116, 729; Angew. Chem. Int. Ed. 2004, 43,
baugh, S. L. Shamblin, A. M. Aukamp, Organometallics
994, 13, 890.
18] H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc.
957, 79, 1920.
19] C.-K. Jung, S. A. Garner, M. J. Krische, Org. Lett. 2006,
, 519.
1
[
[
7
11; t) K. Miura, Y. Yamada, M. Tomita, A. Hosomi,
1
Synlett 2004, 1985.
[
5] For examples of Rh-catalysed reductive aldol reaction
8
using hydrogen as reducing agent, see: a) H.-Y. Jang,
2666
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2657 – 2666